Collisional quenching of K* (4p2p) and K* (5p2p by H2O, CF4, and CH4

Volume 22. number 1

COLLlSlONAL

15 September

CHEMICAL PHYSICS LETTERS

QUENCHING OF K’“(4p2P) AND K”(Sp*P) Bold L. EARL and Ronald

1973

BY H,O, CF,, AND CH,

R. HERM

horgmic Materials Research Divisiorl, Lawreme Berkeley Laborarory clnd Department cf Chemistry, Utliversity o(Califomia, Berkeley, California 94720,

USA

Reccivcd 11 May 1973

Rxte constants are reported for collisional quenching ot K’(4p’P) and K’(5p’P) by H20, CFq, and CA+. TRc X;‘(+p’P) Or K’(Sp*P) is produced by photodissociation of K1 vapor nt 2450 A or 1925 A. respectively. AS for Ns*t3pZP), H20. CF=. and CH4 arc wry inefficient at quenching K*(4p2P); howcvcr, they are very efficient at

quenching K*(Sp2P).

Although a number of workers have studied the gas-phase collisional quenching of the lowest excited configuration of the alkali atoms?, very little is known about the quenching of higher energy configurations. In contrast to Hg and rare gas atoms which have also been the subjects of many studies, the escitation energies of the alkali atoms are typically less than the bond energies of molecules which are efficient quenching agents. For example, the excitation energies (in ev) of the first siu excited configurations of K are: 1.Gl (4p2P) , 2.60 (js*S), 2.67 (3d?-D), 3.06 (5p2P), 3.40 (4d2D), and 3.41 (6~~s). Nevertheless, a variety of molecules are quite efficient at quenching the lowest excited configuration of the alkali metals, although certain “saturated” molecules, such as CH,, CF,, and, to a lesser extent., Hz0 are inefficient as quenching agents [ 1,2]. The work described here was undertaken in order to determine whether this inefficiency would persist f7r higher energy configura. tions. Measured rate constants are presented for quenching of K*(4p2P) and K*(5p2P) by H,O, CFq, and CH,. The apparatus and experimental procedure was similar to that described in ref. [2]. All measurements were made with ultraviolet radiation from a hydrogen arc which passed through a monochromaior with a triangular bandpass function, 60 A fwhm, and imt For Breview of the literature, see ref. [l].

pinged upon a quartz cell containing S[ vapor at 890 “K. The K*(4p*P) fluorescence intensities at 7665 A and at 7699 A, which were produced by photodissociaticn of the KI at 2450 #, were isolated by different interference filters. irradiation at 1925 .% permitted measurement of the K*(Sp’P) fluorescence through an interference filter which transmitted both the 4044 and 4047 A resonance lines. With the UV intensity held constant at one of these two wavelengths: the rate constant, k,, for quenching ofthe corresponding 4p or 5p potassium fluorescence was determined b> admitting a known number density, n, of quenching gas to the KI cell and fitting the measured fluorescence intensity, R,, , to the Stern-Volmer relation:

The radiative fluorescence iifetimes, 7, were assigned values [3]?? of 2.6 X IO-* set for K*(4p2P) and 1.4 X 10e7 set for K*(5p2P). Thus, k, is determined by assuming that the resonance fluorescence is produced by the direct photodissociation of KI to give the corresponding K*(4p2P) or K*(5p2P) configuration. This clearly holds for production of K*(4p’P) at 2450 A because threshold for production of K*(Ss*S) by photodissociation of KI in its ground vibrational

ti Corrections for radiztion irnpr+xnnent of K* resonance fluoresumce are negligible hecause potessium is produced il very low concentrations by photodissociation of KI.

95

Vdume

22,

number 1

K*(.!+*P) + CH, and CF4, 3 few torr of the quenching gas WDSpre-mixed with =Z 650 torr of Ar, and this mixture (at a reduced pressure) was admitted to the KI cell in order to ensure that the K*(SplP) speed distribu tion was thennalized by collisions with Ar. In separate experiments, it was ascertained that pure Ar (at z 600 torr) or pure Xe (at s 200 torr) failed to produce a measurable attenuation of the K*(5p2P) fluorescence intensity. This “Ar themlalizing” procedure was not used in measuring k, for K*(Sp2Y) + Hz0 due to experimental difficulties. Here, however, the distribution of relative collision speeds is primarily determined by the thermal speed distribution [2] of Hz0 by virtue of its light mass re.lative to K. This was verified with K*(5p2P) + CH4 where a value of (6.7 2 0.7) X I O-lo cm3/sec was measured for kq in the absence of the “AI thermalizing” procedure, in reasonable agreement with the thermal value given in table 1. We are not aware of any previous studies on the quenching of K resonance fluorescence by CH4 or CF,. Flame fluorescence studies of K*(4p2P) + Hz0 have determined quenching cross sections of (2.8 50.9) A* at 1400-1800°K [6] and (2.6 50.3) A2 at z 2000°K [7], in reasonable agreement with our result in view of the temperature differences. On G the other hand, the only previous study [7] of

to be 2060 A. Calculated thresholds for production of K*(5p2P) and K*(4d2D) [or K*(6s2S)J are at 1920 and 1820 A, respectively. The rneasurcd reduced K*(Sp2P) fluorescence efficiency (i.e., the ratio of fluorescence intensity to incident UV in tensity) peaks at =Z1900 A and decreases to 20% of its peak value at ~1760 and 2020 A, indicating that very little of the K*(jp’P) fluorescence produced by 1925 8, can be clue to ini tiaJ production of higher energy configurations. This was further verified in an level is calculated?

auxiliary experiment in which no variation was observed in kq for CH, quenching of the K*(Sp”P) fluo-

rescence produced by irradiation of KI at 1925, 19.50, and 1975 A. Measured values of urements at 7665 and k, value for K*(4p2P) or CF4, where k, was

kq are listed in table 1. Meas-

7699 a determined the same + H20. For Ka(4p2P) + CH, too small to measure reliably,

some collisional mixing of the two fine structure levels was observed. In contrast to work jr-rref. [a] where the translational

15 September 1973

CHEMICAL PHYSICS LETTERS

energy dependence

of kq was

measured, k, values listed in table ! were determined for thermal distributions. Relatively large pressures of quenching gas (z 20 torr for H?O, e 50 torr for CF, and CH,) were necessary to determine the slow quenching rates of K*(Gp”P), so that the initial K*(4p2P) speed distribution should have collisionaliy relaxed before being quenched. In experiments on

K*(5p*P) determined (at 2000°K) a quenching cross section For I-l20 of (10-+4)A*, substantially different from the result obtained here. In ref. [?I, we chose to compare

t Bond dissocjation

energies for #I and KH were taken from ref. [4]; other values are from reT_IS].

our measured

cross

Table 1 hIe3surcd quenching K

configuration

4p2P 4pZP 4p2P 5p2P 5p2P 5p2P

rate constantgko,

Quenching gas

at 890°K k&,

C)

ADo (e\3 d)

1.3(KF + CF$

(34

<0.01

c

I

co.01

CH4

-co.01

c

1


-0.9(KH + CH3) O.O(KOH + H)

0.16 0.45

2.8(KF + CF$ 0.6(RH + CHJ)

0.71

1.S(KOH + HI

H20

0.024rG.008

CFQ

0.21

to.04

CH4

0.77

20.08

26 60

1.5

Hz0

1.0

to.1

85

a) The uncertainty given for each k value is simply the standard deviation prcwided by a least squares fit of the pressure dependence of the fluorescence intensi?y to eq. (1). The true error in k should b: somewhat larger. avsragc relarive speed. b) Quenching cross section. ulcuht=d a~ kq = Q (g), where (g> is th:thrrmal cl As discussed in the text, k /k~ is a meaare o‘f the quenching efficiency. d> A& is the exocrgiciry or t&c possible reaction; data from refs. 14,151.

96... -:

kk~me 22, ntimbcr 1

CHEMICAL PHYSICS LETTERS

sections, Q,, for quenching of Na*(3p2P) with CTOSS sections, QD, for “close encounters” wherein the incident trajectories surmount the centrifugal barriers in the effective potentials associated with the longrange attractive dispersion forces. In doing so, we did not propose that Qq should equal QD because (1) the description of the quenching collision must include some matrix element coupling different electronic states and (2) QD provides only an estimate of the cross section for close encounters because of neglect of shorter range forces. Nevertheless, this comparison proved useful in assessing the efficiency of differenr molecules for quenching Na*(3p2P). Molecules with appreciable electron aftinittis (e.g., 1, and SO,) are “super-efficient” with Qq > Q,, ;Insbt::ated molecules (e.g., N, and C2H4) as well as some potentially reactive compounds [e.g., CF3Cl) are efficient quenchers with Q, 5 Q,, , and certain saturated molecules (e.g., H,O, CF, and U-I,) are very inefficient wi-ih Q, +Z QD. A simi!ar comparison, shown in table 1 in the form of rate constants (kq/kD) rather than cross sections, indicates that H20, CF,, and CH4 are also very inefficient at quenching X*(4p’P). For K*(5p2P), however, H20, U-I,. and, to a lesser extent, CFJ prove to be quite efficient as quenching agents. The quenching collisions studied here refer to some unknown combination of inelastic processes producing lower K configurations and.reactive processes. For H20 and CF,, especially, the large efficiency for quenching of K*(5p2P) might simply be a consequence of the relatively large reaction exoergicities listed Lr table I; this point of view is advanced by Dowling et al. [8] in discussing their results on quenching of Na*(3p2P) and T1*(7szS). At any rate, the failure of Ar or Xe to quench K*(5p2P) suggests that H20, CF,, and CH, quench this configuration either Ly reaction or by an inelastic process wherein electronic energy is at least partially converted into vibrational and rotational excitation. Furthermore, the failure of CH, to efficiently quench H~*(~P~P~,~) [9], despite the larger excitation energies, suggests that the dramatic difference shown in table 1 is not simply a consequence of higher density of vibrational-rotational levels in the quenching molecules corresponding tion

to the higher

K*(SpzP)

excita-

15 September 1973

ing of K*(5pZP) a charge-transfer K configurations

I ] in the related (not atoms. Appearante potentials of 5.0--5.7, z7.5, and 4.7 eV have been reported for dissociative electron attachment to Hz0 [ 121, CH, [ 121, and CF, [ 131, respectiveIy, whereas vertical electron affinities of == -3 to = --4cV wouId appear to be necessary in order to account for the observed contrasting efficiencies for quenching of K*(4p”P) and KY(Sp2P) in terms of this charge-transfer model. However, these dissociative electron attachment results may not refer to excitation of the lowest negative ion resonance states. Indeed, Claydon et al. [ 141 have reviewed the known data on negative ion states of Ha0 arld have concluded that the lowest

resonance state is not seen in dissociative electron attachment because it is bound relative to dissociation into OH- + H, suggesting a vertical electron affinity greater than -3.28 eV, probably between -7, and -3 eV. In terms of the simple charge transfer modeI: a value near -3 eV would seem to account for the large value of x-q for K*(Sp’P), the relatively small values ofkq for K*(4p2P) and Na*(3pzP), as weU as the trend [ I] of increasing kq on proceeding from K*(4p2P) to Rb!(SpzP) and Cs*(6p2P). Similarly, Christophorou and Stockdale [ 15J have suggested that the lowest energy dissociative electron attachment observed in CH4 proceeds-via excitation of a core-excited resonance level. Thus, it seems plausible to attribute the quenching behavior of H,O, CH4, and CF4 observed here to formation of charge-transfer intermediates involving the lowest energy resonance states of the negative ions. It should also be noted, however, that the increasing density of electronic terms with increasing excitation energy might be expected to lead to avoided crossings and efficient quenching even in the absence of formation of sucfr a charge-transfer intermediate. This work was supported

by the U.S. Atomic En-

ergy Commission through the Lawrence BerkeIey Laboratory.

Partial support

from the AIfred P. Sloan 97

Volume

22, number

1

CHEMICAL

Foundation, in the form of a 1970-72 RRH, is also gratefully acknowledged.

fellowship for

PHYSICS

References [II

P.L. Lijnse, Review of Literature tion and Mixing Collision Cross

on Quenching, ExcitaSections for the First Re-

sonance Doublets of the Alkalis, Report i-398, Fysisch Imborsrorium,

Rijks

,Lands (1972). 121B.L. Earl, RR.

Un+crsiteir

Utrcchr,

The

LETTERS

(111 K. Lacmann and D.R. Herschbach,Chem, Phys. Letters

6 (1970) 106.

Nether-

1121 L.C. Christophorou, Hcrm,

S.-51. Lin and C.A. Mims. J. Chem.

Phys. 56 (1972) 867. hi.W. Smith and B.M. Miles, Atomic transition probabilities, Vol. 2, Natl. Stand. Ref. Data Ser.,

I141

Natl. Bur. Std. US 21 (1964). (41 A.G. Gadson, Dissociation energies and spectra of djatomic molecules (Chapman and Hall, London, 1968).

98 .. _ -.

‘.

.,. ; ;

‘.

,.,. I

,,

‘. :,

: _

(Wiley.

Atomic and molecuhr New York, 1971).

radiation

MacNeil and J.C.J. Thynne. Intern. J. hlass Specfrom. Ion Phys. 3 (1970) 455.

[31 W.L. Wiese,

,.

1973

[5] B. de B. Denvent, Bond dissociation energies in timole molecu!es, Natl. Stand. Ref. Data Ser., NitI. Bur. Sid. us 31 (1970). [61 D.R. Jenkins, Proc. Roy. SOC. A303 (1968) 453. 171 H.P. Hooymayers and P.L. Lijnse, J. Quant. Spectry. hdiative Transfer 9 (1969) 995. 181 D.1. Dowling, G.R. H. Jones and E. Warhurst, Trans. Faraday Sot. 55 (1959) 537. H.C. Moser, J. Chcm. Phys. 53 (1970) t91 A.C. Vikicand 1491. 1101 E. Bauer, E.R. Fisher and F.R. Gilmore, 1. Chem. Phys. 51 (1969) 4173.

physics [I31 K.A.G.

,. . . ..

15 September

I151

C.R. Claydon, G.A. Segal and H.S. Taylor, Phys. 54 (1971) 3799. L.C. Christophorou and J.A.D. Stockdale, Phys. 48 (1968) 1956.

3. Chem. J. Chem.