Z. Physik 229, 1--13 (1969)
Reactions of Excited Atoms and Molecules , i
with Atoms and Molecules I I I . Relat ive Cross Sections for Penning- and Associat ive Ion izat ion
by He(2 IS)- and He(2 3S)-Metastables
H. HOTOP, A. NIEHAUS, and A. L. SCHMELTEKOPF* Physikalisches Institut der Universit/it Freiburg, GermanY
Received April 23, 1969
From a He-beam excited by electron impact we eliminated the He(2 1S) component to better than 0.5 % by irradiating light from a He discharge. The quenching process is hv(2 1P---~2 1S)+He(2 1S)--*He(2 1P)---rile(1 1S)+hv(2 1p_ 1 1S). By measuring the ions produced in collisions of the He-metastables with various target gases in a mass spectrometer, singlet to triplet Penning-cross section ratios were obtained. These ratios are without exception close to one, which is taken as evidence for the previously proposed electron exchange mechanism of the Penning ionization. In the case that more ions are produced in the collision of He(2 1S) and He(2 3S) with a target gas, separate relative production cross sections are obtained for the two metastables. For the rare gases the measurements are performed at two temperatures of the He- beam, 320 and 90 ~ It is found that the cross section ratio of associative -- to Penning ionization increases considerably as the temperature is decreased for both, He(2 1S) and He(2 3S), the effect being much more pronounced for He(2 IS). The results of this Work are found to confirm conclusions drawn from measured energy distributions of the electrons ejected in the Penning process.
Cross sections for ionizing collisions of He-atoms in the metastable states 23S and 21S with other atomic and molecular gases are of interest because of mainly two reasons, (1) because these metastables play an important role in the physics of the atmospheres of stars and planets, and (2) because the knowledge of the relative cross sections for the two states might give some insight into the mechanism of Penning reactions in general.
In our preceding publication 1 we proposed that ionizing reactions of metastables in thermal collisions with other gases occur by exchange of the two participating electrons, (1) and (2):
A*(1) + B(2)-+A(2) + B + + e- (1). (1)
* Present address: E.S.S.A. Laboratories, Boulder, Colorado (U.S.A.). 1 Hotop, H., and A. Niehaus: Submitted for publication, Z. Physik and VI. I.C.P.
E.A.C., Cambridge 1969.
1 Z . Phys ik , Bd . 229
H. Hotop, A. Niehaus, and A. L. Schmeltekopf:
If this is true, it is to be expected that the cross sections for the two He metastables, which differ only in the spin state, are approximately equal (see Ref. i). On the other hand, if the ionizing reactions are predominantly radiative transitions (2)
A*(1)+B(2) -~A(1)+B + +e- (2), (2)
in which, in the course of the collision, the perturbed metastable emits a photon, h v (A* ~A), which in turn is absorbed by the target particle B, leading to its photoionization, B~B + +e- , the two cross sections should in general be very different favoring the cross section for the singlet metastable because of its shorter radiative lifetime. The latter mechanism has been proposed by several authors 2,3 for the ease that the transition A* ~A is optically allowed.
Experimental absolute cross sections have been reported by Benton et al. 4 for the destruction of the He-metastables in collisions with various target gases, and by Sholette et al. 5 for the total ionization by the He-metastables. Benton et al. give separate values for He (21S) and He (23S), whereas the Penning ionization cross sections reported by Sholette et al. are obtained for an essentially unknown mixture of He (21S) and He (2aS). The destruction cross sections for He (2aS) are in fairly good agreement with the average Penning cross sections, indicating that the main destruction process is the Penning ionization. The singlet destruction cross sections reported are larger but very unreliable, as the authors 4 remark. In the work of Ref. 5 the average Penning cross sections are obtained by measuring, in a collision chamber, the ions produced by the He-metastables contained in a He-beam excited by electron impact. By varying the energy of the exciting electrons it was possible to change the relative numbers of singlet and triplet metastables in the beam. From the fact, that the average cross sections obtained did not depend on the composition of the beam, it was concluded that the cross sections are equal for He (21S) and He (23S) within the accuracy of the determination of 10-20 ~o. The only exception was H2, for which a singlet to triplet cross section ratio of cq/o-t~0.65 was esti- mated.
In this publication we report on the comparatively accurate (1 -5 ~) measurement of the quantity ~=NscrJNt(rt, where NJNt=R(Ee~ ) is the ratio of singlet to triplet metastables contained in the He-beam excited by electron impact at the energy Eel. Cross section ratios are obtained from ~ by setting cr~/at(N2)=l. 5 We took N2 as reference target for 2 Watanabe, T., and K. Katsuura: J. Chem. Phys. 47, 800 (1967). 3 Smirnov, B. M., and O. B. Firsov: J. Exptl. Theoret. Phys. 2, 478 (1965). 4 Benton, E. E., E.E. Ferguson, F.A. Matson, and W. W. Robertson: Phys. Rev.
128, 206 (1962). 5 Sholette, W. P., and E. E. Muschlitz Jr. : J. Chem. Phys. 36, 3368 (1962).
Reactions of Excited Atoms and Molecules with Atoms and Molecules. III 3
reasons outlined below. A renormalization of the as~at-values can be made, of course, if a different more reliable ajar-value is known.
In the case that more than one ion species are formed in the collision of the metastable atom with the target gas, relative cross sections for their formation by both He (21S) and He (23S) are given. In addition, for the rare gases At, Kr, Xe being the target particle, the dependence of the relative cross sections for Penning- and associative ionization on the collision energy is investigated separately for the two metastable He-states.
The measurements are facilitated by a new and simple method to produce a "pure" He(23S) beam from a beam containing an arbitrary mixture of He(2aS) and He(23S) metastables. This method consists of irradiating the beam of metastables with light from a He-discharge containing photons which correspond to the transition He(2XP~2~S). Because of the very high cross section for absorption of these photons by the He(2aS)-metastables, it is possible to eliminate the singlet com- ponent almost quantitatively (to better than 0.5 ~) by the process
hv(21P~21S)+He(21S)~He(21P)~He(11S)+h ~(584A). (3)
The corresponding quenching process for the triplet metastable does not occur, because the I-Ie(23S), if excited to the (23P)-state, returns to the (23S)-state.
II. Experimental Fig. 1 shows the main parts of the apparatus used. A He-beam formed
by passage through a multichannel tube (M) is excited by electron impact in an excitation chamber (EC) and passes through an area where charged particles are removed by potentials applied to the grids (G1 - G4), and where the light from a He-discharge lamp is irradiated. The beam finally enters a reaction chamber (RC) where it crosses a beam of target particles without hitting any metal surfaces in RC. Ions produced in collisions of the excited He-atoms with the target gas are mass analysed in a mass spectrometer, and electrons can be energy analysed in an electron spectrometer consisting of an electrostatic Einzellens and a retarding electric field, described elsewhere 1. The He-discharge lamp is a closed glass spiral (pyrex glass) surrounding the metastable beam path for ca. 3 cm and filled with 6 Torr He. It is operated at currents (i) up to 50 mA and a voltage of ca. 1 kV. Since no cooling was applied, the time of operation was limited to ca. 30 sec, the time needed for cooling down being a few minutes. Under these conditions the intensity of the (21P~21S)-photons was found to depend linearly on the current (i) up to 50 mA. The efficiency of the lamp to quench He(21S)-atoms
H. Hotop, A. Niehaus, and A. L. Schmeltekopf."
"il E chamber G1
Electron - '~ - - - - - G4
Spectrometer ~ ~ct ion chamber
" " Target Gas
Mu-Meta[ DI ~1 to Mass Si~ectrometer
Fig. 1. Schematic drawing showing the part of the apparatus used for the production of a pure He(23S)-beam
was tested in the following way: In the Penning process for each ionic state of the target, electrons of energies approximately equal to the difference between the excitation energy of the metastable and the ionization potential of the target gas are liberated 6,1. In the retarding electric field we therefore measure, for each target ion state, two electron groups separated in energy by about 800 meV, the difference between the excitation energies of He(2aS) and He(23S). The corresponding "'steps" of the integral retarding curve are shown in Fig. 2 for the ionization of Xe. The step hights Ss and St are proportional to Nsa~ and Ntat.* If only metastables with a discrete velocity v are present in the beam, we expect that the singlet step hight S~(i) depends on the
(') .discharge current (i) as S~(i),,~N~(O). as .exp -c v . A plot of log S~(i) against (i) is shown in Fig. 3. The deviation from linearity is expected because (1) the metastables have a velocity distribution, (2) ~r~ depends probably on v, and (3) the angular distribution of the emitted electrons might change with v. The presently important fact however is, that N~(40 mA)/Ns(O),,~S~(40 mA)/Ss(0)
Reactions of Excited Atoms and Molecules with Atoms and Molecules. III 5
Xe~(2p3/2) He(2sS) I He(2'S)
75 8 0 8.5eV Time Electron energy
Fig. 2. Stopping curve for electrons resulting from Penning ionization of Xe by He(21S)- and He(23S)-metastables. Also shown is the decrease of the singlet Signal S s and of the total signal S~+ S t caused by irradiation from the He-discharge lamp
10 20 )~ I I l I I
30 40 i I I
10 20 3 4 Discharge current i [rnA]
Fig, 3. Plot showing the dependence of the intensity of electrons arising from Penning ionization of Xe by He(21S) on the discharge current in the He lamp
For the determination of ratios Ns~Ts/Ntt7 t it is essential to establish that the number of triplet metastables reaching the reaction chamber is not influenced by the light f rom the discharge. A partial quenching is possible in principle if He(23S) is excited to a higher state of the triplet system from where a transition to the singlet system by spin
H. Hotop, A. Niehaus, and A. L. Schmeltekopf:
orbit coupling can occur. From Fig. 2 it can be seen that such a quenching does not occur to a measurable extent: the reduction AS(E1) of the singlet signal Ss by irradiation of the discharge light is equal to the reduction A S(E2) of the sum of singlet and triplet signal, S~ + St. From this a quenching of the triplet can be excluded to better than 3 700-
The conditions under which the measurements described in the next section were carried out may be summarized as follows:
1. The exciting electron current was 1 mA at 50 eV. In order to keep Ns/Nt constant, these conditions were not changed. In addition, to correct for small variations of N~/Nt, the ratio N~aJN~cr t for N2 was measured periodically. We chose N 2 as a normal because as/a t for N 2 is very insensitive to pressure and temperature variations, mostly due to the fact that no associative ionization with He metastables Occurs .
2. A contribution of photons, emitted by excited He-atoms in the excitation chamber, to the ion production in the reaction chamber, can be excluded to better than 1 70 since no photoelectrons could be detected in the electron spectrometer (see Ref. 1).
3. A small contribution to the signal may arise from ionization of the target gas by He atoms in highly excited long-lived Rydberg states. As was tested, He atoms in such states are not influenced by the He lamp. It was found that their number density is so low (mainly because of the long distance from EC to RC and the electric fields in between) that their contribution to the signal does very probably not exceed 1 ~o.
4. Since no differential pumping was applied, the total pressure in the mass spectrometer during the measurements had to be kept low enough as to avoid scattering of the ions as well as collisional dissociation of molecular ions. Especially in the case of molecular ions produced by associative ionization, for instance HeAr +, collisional dissociation was important down to very low pressures (~ 10 -6 Torr) in the mass spectrometer. In each case the pressure was chosen in such a way that the effects mentioned did not influence the measurements.
5. In general the energy distribution of the excited He-atoms is different from the distribution of the ground state atoms due to the exciting electron impact. Although mainly He-atoms with small-angle deflections remain in the beam, the corresponding change in energy may be rather high. In the case that the He beam and the electron beam are perpendicular to each other, the energy may be increased or decreased. As an example, the effect of the momentum transfer of exciting electrons of 50 eV on He atoms of 40 meV kinetic energy shall be outlined: the deflection angle of the excited atoms ranges from 5 to 38 ~ For deflec- tions into these limiting angles the change in energy is 2.7 meV. The
Reactions of Excited Atoms and Molecules with Atoms and Molecules. III 7
extreme energies possible for excited He atoms are 23 and 79 meV at the middle deflection angle of 21.5 ~ .
The general effect of the momentum transfer will be that the energy distribution for the metastables is broader than the one for the He beam from the multichannel jet M. The average energy of the metastable beam, however, is surely considerably lowered when cooling of M down to about 100 ~ is applied. What is important in the present study is to obtain qualitative ideas as to what happens when the relative energy of the collision system is changed.
III. Measurements and Results
The quantity ~=Ns/N t x ajar for a certain ion is obtained in the following way: (1) The ion intensity I at the output of the mass spectro- meter is measured with the He discharge lamp off (i=0). For the ion intensity I(0) we then have the relation I (0)= (N s a s + Nt at). C, where C is a constant. (2) The ion intensity is measured at i=40 mA. Since at this current the singlet metastables are essentially all quenched (Fig. 3) while the triplets are not influenced, we have I (40)=Ntat . C, where C, of course, is identical with the constant C for the foregoing measurement. For ~ it follows
1(0)-1(40) Ns as = (4)
I(40) Nt at
Using the result of Ref. 5 that as/at= 1 for the total ionization of Na, cross section ratios are obtained from the measured quantities ~. The results are summarized in Table 1 (for the rare gases, see Table 3).
If ~ is measured as a function of t...