January 1982 / Vol. 7, No. 1 / OPTICS LETTERS 25

XeI(2 ir1/2-2 1 i/2) excimer emission at 1.3 ,um

E. Gerck and E. Fill

Max-Planck-Institut fur Quantenoptik, 8046 Garching, Federal Republic of Germany

Received September 25, 1981

The collisional enhancing of the magnetic dipole transition I(52 P1 /2-52 P3 /2 ) at 1.3152 pm by Xe is shown to proceedover an exciplex channel: I(52P1/2) + Xe - XeI(2 r1 /2) - I(52P3 /2 ) + Xe + hp (1.3 Am). The pseudo-first-orderrate coefficient for this process is (1.6 L 0.2) X 10-13 molecule-' cm3 sec-. The termolecular rate coefficient forXeI(2 r1 ,/2 ) excimer production is (2 + 1) X 10-35 molecule- 2 cm6 sece'. The radiative lifetime of the XeI excimerstate is of the order of 100 nsec. Similar effects are briefly reported for C2 F6 , SF6 , Kr, Ar, and i-C3 F7 I.

We report on the 2 7rl/2-22i/ 2 excimer transition ofXeI* at 1.3 gm that results from collisions of spin-orbit-excited I(52P1/2) iodine atoms with ground-statexenon atoms. The lower state of this excimer transitioncorrelates with ground-state I(52P3/2) iodine and Xeatoms, which leads to excimer emission around theI(52 P1 /2 -52 P3 /2 ) atomic line at 1.3152 Aim.

We also show that the collisional-enhancing effect ofthis I-forbidden transition by Xe, first reported by Er-shov and Zalesskii,1 is caused by the XeI emission. Thismeans that the Einstein A coefficient of the atomic lineitself (A = 8 sec-1) is not increased by collisions, assuggested by Ref. 1, but that a second, electric-dipole-allowed radiative channel is made available by the ex-cimer XeI* buildup. This applies also to the en-hancement effect caused by any other nonpolar mole-cule or atom, such as Ar, Kr, SF6 , and C2 F6.

This excimer model is in agreement with the pressurebehavior of the stimulated-emission cross section for the1(52 p1/2-52P3/ 2 ) transition of Galanti et al.2 and withtheir conclusion that the respective A coefficient does notincrease with pressure. The apparent contradictionbetween Refs. 1 and 2 is solved by the introduction ofthe excimer channel described in this Letter as the en-hancing mechanism of the I(52P1/2) fluorescence. Al-though our discussion is centered on Xe, similar resultsare obtained for Kr, Ar, SF6 , and C2 F6. Also, the weaklypolar molecule i-C3 F71 causes a collisional-enhancementeffect.

The experiment consisted of observing the relativeincrease of the IR fluorescence signal when i-C3 F71 wasphotolyzed in the presence of a buffer gas. The ex-periment used a XeCl laser to photolyze the i-C3F71molecule, thus producing I(52P1/2) with approximateunity quantum efficiency.3 The 308-nm pulse (300 mJin 20 nsec) was filtered by a H20 cell to suppress any IRlines. The i-C3F71 and the buffer gas under investiga-tion were mixed with a circulation pump connected tothe fluorescence cell. The fluorescence signal was fil-tered by a Si plate (transmitting above 1.2 ,in) anddetected by a Ge diode. The signal was amplified andsampled by an analog-to-digital sample hold at itsmaximum, T = 9 ,sec after its beginning. The XeCllaser energy was measured at each shot by means of abeam splitter and a calorimeter. Both the XeCl laser

energy and the sampled signal at 9 ,sec were availabledigitally. The experimental results are reproduciblewithin 110%. In Fig. 1 a typical fluorescence signal canbe seen. The conditions were 30 mbars of i-C3F7I and2 bars of Xe. The beginning of the intensified portionof the trace corresponds to a flat part of the signal, nearits maximum, at T = 9 ,sec. Since the kinetics of ex-cimer formation for similar effects involving O(1S) andS(US) in collisions with Xe, Kr, etc. are well known, 4' 5

we proceed directly to the analysis of the experiment.The dominant processes are described by the fol-

lowing set of reactions [M = Xe, Kr, etc., and I* isI(52P1/2*]

i-C3F7I + hv(308 nm) - i-C 3F7 + I*,k2

i-C 3F71 + I* -- i-C3F7 1 + I + hv(1.3 ,im),

I* + M + M --MI* + M,A4

MI* -*M + I + hv(1.3 ,um),

I* A51* I + hv(1.3152 hum),

(1)

(2)

(3)

(4)

(5)

Fig. 1. Infrared fluorescence signal for 30 mbars of i-C3F 7Iand 2 bars of Xe. The beginning of the intensified part of thetrace corresponds to the sampling time of 9 Asec. At this timethe signal voltage was 1.33 mV. Vertical sensitivity, 500,V/div. Horizontal scale, 2 ,sec/div.

0146-9592/82/010025-03$1.00/0 © 1981, Optical Society of America

26 OPTICS LETTERS / Vol. 7, No. 1 / January 1982

where k2 < 2.0 X 10-17 molecule- 1 cm3 sec-1 (Ref. 6)and A5 = 8 sec-1. 3 Equation (3) implies that MI* is(weakly) bound. This is expected for two reasons: (1)the similar exciplex molecule XeO* involving ground-state xenon and metastable O(WS) atoms is known to beweakly bound with a dissociation energy of ;0.06 eV,4and (2) the existence of a nonnull intercept in Fig. 2supports a three-body reaction M + M + I* MI* +M to form bound MI*. The equilibrium constant K3is the ratio of the forward to the backward reaction rate:K 3 = k3/k- 3.

We have observed that the slow quenching given byEq. (2) is an effective source of IR radiation even within20 nm of the atomic-iodine emission. This is also acollisional enhancement effect of the I* fluorescence,but with a weakly polar molecule. For low enoughpressures of i-C3F7I the contribution from this processcan be neglected, as is shown below.

The quenching of I* by the I2 that is produced is notimportant up to T = 9 ,usec because only two shots aremade for each mixture and also because the circulationpump dilutes the I2 in the whole gas volume. The Xegas used (Linde, reinst quality) contained no importantamount of any quenching species even at the highestpressure of 2 bars used. This can be checked from Fig.2.

The amount of MI* molecules in equilibrium with I*and M can be calculated by supposing that processes (3)and (4) are much faster than any of the others. Theexcimer MI* concentration is given by

(M* h 3(M)2 (I*) 6A 4 + k- 3(M) (6)

Since the MI* excimer should be weakly bound, thereaction rate k- 3 is expected to proceed at a gas-kineticrate of about 10-1o molecule-' cm3 sec-' at room tem-perature. By studying the MI* excimer emission as afunction of the M pressure, one may determine, fromEq. (6) and the estimate for k- 3, the reaction rate k3 andthe radiative lifetime of the MI* excimer. The IR signalSo without buffer gas M is given by

So - A5(I*) + k 2(C3F 7D)(I*). (7)

The value of the IR signal at T = 9 ,sec is proportionalto the true maximum value of the signal, i.e., for zero risetime of the detector, because quenching is not importantup to 9 usec. Since the photolysis under our experi-mental conditions involves only a small fraction (n10%)of the total iodide concentration and the iodide is op-tically thin up to 30 mbars, Eq. (7) gives in this case

So(9 gsec) = A5 + k 2X,x (8)

where X is proportional to the C3F7I pressure and A5is 8 sec-1. Equation (8) can be seen as a straight line onX. The slope of Eq. (8) for various iodide concentra-tions can be obtained from a fit to the fluorescencesignal data, yielding

k2 = (4.1 ± 0.8) X 10-18 molecule-' cm 3 sec-1. (9)

Next, with Xe as buffer gas, it was observed that inorder to have an acceptable signal level an iodide pres-sure of 8 + 0.5 mbars could be used. Since this corre-sponds to k2(C3F71) _ 0.8 sec-1, the contribution from

Eq. (2) can be neglected in this case, and Eq. (7)simplifies to

(10)

The IR signal for a nonpolar buffer gas M is increasedover Eq. (10) by the excimer radiation and is given by[(MI*) << (I*)]:

S A5(I*) + A4(MI*).

The relative increase of the IR radiation isS-So [ A 4k 3(M)2 1

sO [A 4 + k. 3 (M)J A 5

It is useful to rewrite Eq. (12) as

(11)

(12)

So (M) 2 = -5 + AA (M), (13)S- so k3 A4K3

where the right-hand side is linear on M. The slope andthe intercept from Eq. (13) determine the excimer ratesA4K3 and k3, respectively. For a Xe pressure rangefrom 10 mbars to 2 bars, the standard procedure was toobtain for each mixture the values of So (9 ,usec) and S(9 ,sec) with only two shots to avoid I2 buildup. Themixing time under circulation was always larger than15 min. The IR signal values were corrected for thesmall changes in the XeCl laser energy. The results forvarious sets of experiments in the total Xe range werealways within 110%. A typical (not averaged) experi-mental result is shown in Fig. 2 for six values of the Xepressure. the error bar corresponds to the 10% uncer-tainty. From Fig. 2 one can obtain the slope and theintercept.

The average value of the slope for the various sets ofexperiments yields, from Eq. (13),

A 4K3 = (1.6 1 0.2) X 10-18 molecule-' cm3 sec-1.(14)

This constant is also a lower bound for the quenchingrate of I* by Xe:

Xe + I* -*Xe + I, kq > A 4K 3. (15)

Because of the small value of the intercept it is impor-

rxX- 4EU

3J0W

E2

. . .

tn

0 0.4 0.8 1.2 1.6

Xe PRESSURE (bars)

2.0 2.4

Fig. 2. Graph to determine the rate coefficient A4K3 for XeI*excimer emission and the three-body-rate coefficient k 3 forproducing XeI* for one typical experimental run.

-Vu I I I I I

.0 /

0 ,

,0- ~~~~0

.0 P

n I I I I0

So -A5(I*).

January 1982 / Vol. 7, No. 1 / OPTICS LETTERS 27

C"

IE

(DwLLz

20k

101-

0

INTERNUCLEAR SEPARATION

Fig. 3. Estimated potential curves for XeI in the region ofinterest. From Ref. 7.

tant to obtain its value from the low-pressure data toreduce the effect of imperfect mixing and noise. Theaverage value of the intercept gives, from Eq. (13),

k3 = (2 i 1) X 10-35 molecule- 2 cm 6 seC 1. (16)

The large error arises from the small value of the in-tercept.

With the reasonable guess for k_3 of 10-10 molecule'cm3 sec-' already mentioned, one may obtain from Eqs.(14) and (16) the following estimate for A4 :

A4 - 107 sec', (17)which corresponds to a radiative lifetime for this XeI*excimer of the order of 100 nsec.

It should be noted that the combination Si plate/Gedetector should not influence the measured noble-gas-excimer XeI* rates because the XeI* emission islocated around the atomic line and is not expected togo beyond 1.2 Am since the iodine atoms are thermalizedafter the photodissociation, as is discussed below.

To assign the observed excimer transition in XeI* werefer to the approximate potential-energy diagram ofXei from Ewing and Brau.7 Similar curves have alsobeen derived by Hay and Dunning.8 Figure 3 showsthis diagram in our region of interest. As can be seenfrom Fig. 3, the observed IR excimer transition corre-sponds to the XeI(27r/2- 221 /2) transition. This tran-sition corresponds to an emitted photon with about theI(2P1/2)-I(2P3/2) energy, an excess being supplied by thekinetic energy of the colliding particles or a small deficitbeing caused by a dissociation energy around 0.06 eV,as for XeO*. Although the I* atoms leave the parentmolecule with suprathermal velocities, they are rapidlythermalized by collisions with the buffer gas. Thismeans that the emitted IR-excimer radiation shouldhave a wavelength around the atomic line at 1.3152 gim,but not shifted by much more than k T _ 0.03 eV.

It should be pointed out that the excimer channel forthe collisional enhancement of the metastable iodinetransiton is theoretically expected for a nonpolar buffergas, such as Xe, Kr, Ar, SF6, or C2F6. In this case theinteraction proceeds only through short-range inter-atomic forces and involves the formation of a molecularcomplex. On the other hand, for a polar buffer gas suchas DF, as studied by Rogovin and Avizonis,9 the inter-action is caused by long-range electrostatic dipole forces

and involves a change of the Einstein A coefficient ofthe atomic line itself. The difference between the twomechanisms is considerable. First, the excimer channelproduces a very large change of the A coefficient (from10 to 107 sec 1), but just for a small number of particles,whereas the dipole channel involves a smaller changeof the A coefficient but for a large fraction of the iodineatoms. Second, the excimer formation should produceby its nature a shift of the IR emission, whereas, bycontrast, the dipole channel emission should be sym-metrical around the atomic line.9

We have studied the XeI(2 7r1/2- 2 ;1 /2 ) excimer emis-sion at 1.3 Am that results from the collisions of spin-orbit excited I(52P1 /2) atoms with ground-state Xeatoms. The process is characterized by a rate coeffi-cient of (1.6 ± 0.2) X 10-18 molecule- cm3 sect and anI(2P1 /2) + Xe + Xe - XeI(2ir1 /2) + Xe rate of (2 ± 1) X10-35 molecule-2 cm6 sece. The collisional enhancingof the iodine I(2P1 /2-2P3/2) atomic transition by colli-sions with a nonpolar particle such as Xe is shown to bedue to the excimer channel and not to a change of theA coefficient of the atomic line itself. This is importantto the high-power iodine laser3 because the stored en-ergy can be increased by pressure broadening, also ata high pressure. The pressure behavior of the stimu-lated-emission cross section for the atomic-iodine linewith Xe, Ar, or SF6 as a buffer gas from Ref. 2 is con-firmed, and the apparent contradiction with Ref. 1 issolved. The parent molecule i-C 3F7I is also shown toenhance the 1.3-pm fluorescence signal.

E. Gerck acknowledges support from CTA/CNEN,Brasil. This research was supported by the Bundes-ministerium fur Forschung und Technologie and Eu-ratom.

References

1. L. S. Ershov and V. Yu. Zalesskii, "Collision-inducedJ(52p1/2-52 P3 12) radiative transition," Sov. J. QuantumElectron. 8, 649 (1978).

2. M. Galanti, W. Thieme, and K. J. Witte, "Saturation en-ergy density and line profile of the atomic iodine lasertransition at high pressure," IEEE J. Quantum Electron.QE-17, 1817 (1981).

3. K. Hohla and K. L. Kompa, in Handbook of ChemicalLasers, R. W. F. Gross and J. F. Bott, eds. (Wiley, NewYork, 1976).

4. G. Black, R. L. Sharpless, and T. G. Slanger, "Collision-induced emission from 0(1S) by He, Ar, N2, H2, Kr, andXe," J. Chem. Phys. 63, 4546 (1975).

5. G. Black, R. L. Sharpless, and T. G. Slanger, "Collision-induced emission from S('S) by He, Ar, N2 , H2 , Kr, andXe," J. Chem. Phys. 63, 4551 (1975).

6. This process was not known to proceed, at least partially,over a radiative channel. The rate should be equal to orless than the total quenching rate by i-C3F 7I: S. L. Do-bychin et al., Kvantovaya Elektron. (Moscow) 5, 2461(1978).

7. J. J. Ewing and C. A. Brau, "Emission spectrum of XeI*in electron-beam-excited Xe/I 2 mixtures," Phys. Rev. A12, 129 (1975).

8. P. J. Hay and T. H. Dunning, Jr, "The covalent and ionicstates of the xenon halides," J. Chem. Phys. 69, 2209(1978).

9. D. Rogovin and P. Avizonis, "Collision-induced gain,"Appl. Phys. Lett. 38, 666 (1981).

Xe+ I(5 2 P3 1 2 )