physical chemistry - ACS Publications - American Chemical Society

physical chemistry - ACS Publications - American Chemical WH Br...

0 downloads 83 Views 958KB Size


PHYSICAL CHEMISTRY Registered in U S . Patent Office

0 Copyright, 197!?,by the American Chemical Society


FEBRUARY 8, 1'979

Reaction of Excited Cadmium('P,) Atoms with Alkane Hydrocarbons. Quenching Cross Sections and Primary CdH Yields W. H. Breckenrldge" and Anita ld. Renlundt Depatfment of Chemishy, University of Utah, Salt Lake City, Utah 841 12 (Rsceivad May 8, 1978; Revised Manuscript Received August 2, 1978) Publication costs assisted by the Petroleum Research Fund

Absolute cross sections for the quenching of Cd(3PJ) by several alkane hydrocarbons have been determined using the resonance-radiation flash photolysis technique. Yields of CdH from the alkanes as compared t o the CdH yield in the reaction of Cd(3P~) with Hz were also determined. The Cd(3P~) quenching cross sections for the normal alkanes are four orders of magnitude smaller than the cross section for quenching by HB,even though CdH formation is exothermic for the alkanes and only thermoneutral for Hz. The CdH yields from Cd(3PJ quenching for propane and isobutane are equal within experimental error to the CdH yield for H2, and it is suggested that this yield is probably unity. In marked contrast to the quenching of the valence isoelectronic species Hg(3Po)and Hg(3P1)by the alkanes, there is virtually no dependence of the Cd(3P~) quenching rate on C-H bond strength, and none of the theories advanced to explain the Hg(3P~) interaction with C-H bonds can account for the Cd(3P~) results. A mechanism is tentatively suggested in which a loose complex is formed between Cd(3PJ) and alkane molecules, followed by a low-probability potential surface crossing which has little dependence on C-H bond strength, or at least can occur with primary C-H bonds. It is also suggested that differences between Cd(3P~) and Hg(3P~) quenching by the alkanes can be understood if the mechanistic pathways which account for the Hg(3PJ) results are merely energetically inaccessible to the lower-energy Cd(3PJ) state at moderate temperatures.

Introduction The reactions of excited Hg(3P1)and Hg(3Po)atoms with alkane hydrocarbons are probably the most well studied of the chemical reactions of excited atoms wiith molecules.1-6 The reactions are interesting in that the chemical attack rseems to be localized on individual C-H bonds, since the rates of reaction correlate very strongly with C:-H bond strengths and there is no evidence for any primary C-C bond scission. On the other hand, little definitive work has been published on the reactivity of analogous atomic states :such as Cd(3PJ) or Zn(3PJ) with the allranesa7-15

* Camille

and Henry Dreyfus Foundation Teacher-Scholar,


'Department of Physical Chemistry, Cambridge [Jniversity, Cambridge, England. 0022-3654/79/2083-0303$01 .OO/O

Some classical experiments on Cd(3PJ Sensitized decomposition of alkanes and cycloalkanes have indicated a similar mechanism of C-H bond cleavage, but the lower reactivity of Cd(3PJ) toward alkanes results in severe experimental difficulties because of quenching by secondary products, such as Hz or alkenes.lZJ3 In a previous flash photolysis study of Cd(3PJ) quenching in our laboratories7 we were able to show that the reactivity of Cd(3P,) with the higher alkanes was indeed much les,s than that of Hg(3Pl), but only rough estimates could be made of the absolute quenching cross sections. We report here accurate determinations of the extremely low cross sections for quenching of Cd(3PJ) by some representative alkanes, as well as primary yields of CdH as compared to the CdH yield from the Cd(3PJ)-H2 reaction. The results cannot readily be explained in terms 0 1979 American Chemical Society


W. H. Breckenridge and A.

The Journal of Physical Chemistry, Vol. 83, No. 3, 1979

of previous theories of Hg(3Pl)-alkane interactions, and we propose a mechanism which is consistent with our findings.

Experimental Section The apparatus and general technique have been described previ~usly.~ Briefly, high concentrations of Cd(3PJ) are generated in Cd vapor a t 280 “C by an intense 40-ps pulse of 3261-A cadmium resonance radiation. The excited states Cd(3Po) and Cd(3P1), as well as certain direct products of their chemical reactions (such as CdH), can be detected by kinetic absorption spectroscopy using a conventional Kr-filled continuum flash lamp. Relative concentrations of Cd(3Po),Cd(3P1),CdH, and CdD are determined by plate photometry. Ethane, propane, and isobutane (all Matheson Research Grade) were freeze-pumped several times. Cyclohexane (Eastman Spectrochemical Grade) was refluxed over sodium, passed through an alumina-packed column, then freeze-pumped several times. Carbon dioxide (Coleman Instrument Grade) was distilled from -78 to 196 “C and freeze-pumped several times. Hydrogen (Matheson Prep. Grade) was passed through a trap packed with 5A molecular sieve a t -196 “C. Argon (Matheson Ultra High Purity) and nitrogen (Matheson Prep. Grade) were passed through a glass-wool-packed trap a t -196 “C. Methane (Phillips Research Grade) was used without further purification. Results It has previously been shown7 that under conditions of greater than 100 torr pressure of Ar, N2, or CH4, the ratio of concentrations of Cd(3Pl) and Cd(3Po)generated by the resonance radiation flash photolysis method is constant, and that the concentration-time profile of the excited atoms is essentially identical with the intensity-time profile of the 3261-A excitation pulse. A steady state in collisionally equilibrated Cd(3Po,l)is thus established very rapidly because of the short radiative lifetime of the Cd(3P1)state (2.3 ps) compared to the length of the excitation pulse (-40 ps). The quenching and reactivity of Cd(3Po,l)in this study can thus be described by the following set of equations where Q is any quencher molecule: Cd(’S0) Cd(3PJ

+ hv


Cd(3P1) + M Cd(3Po) + M






+ Cd(3Pl) + Q Cd(3Po) Q

+ hv


Cd(3Po) + M


Cd(3P1) + M







A steady-state analysis results in the following expressions:

These equations are similar to the classical SternVolmer expression for fluorescence quenching, except that a single quenching rate constant is replaced by a weighted sum of rate constants (hdhO/h-d + kl), representing

M. Renlund

TABLE I: Radiation Imprisonment Ratios for Cd( 3P,) under Various Conditionsa gas Ar CH,

press., torr 400






250 400




750 1000

1.16 1.12


a T = 280 “ C . [ C d ] = 1.9 X lo-* torr. Cylindrical reaction vessel, 80 cm length, 0.25 cm radius. T is the effective (“imprisonment”) lifetime of Cd( ”,), T~ the natural radiative lifetime of Cd(3PI)at very low [Cd] ,

quenching by an equilibrated pair of excited states. A t 280 “C, a statistical mechanical calculation yields (kdko/kd k,) = kl = (1.3i’ho kl). For simplicity, the composite quenching constant (1.37k0 k,) will hereafter be designated k,. If cadmium monohydride, CdH, is produced in a constant fraction of the quenching collisions of Cd(3PJ) with a particular quencher Q, the CdH concentration near the end of the excitation pulse may also be used as a “marker” for Cd(3Po,l)reaction kinetics because of the relatively long life of CdH:7




[ C ~ H ] [ Q ] = = . / [ C ~=H1I -l (kr/h,[Q1)


where Q is any quencher which produces CdH, and [CdH][Q]=,is the extrapolated maximum value of [CdH] for that particular quencher when l/[CdH] is plotted vs. l/[Q]. For the efficient quenching gas Hz, for example, small pressures of H2can be added to 400 torr of an N2/Ar mixture. The excess N2/Ar mixture assures rapid equilibration of Cd(3Po,l),prevents errors due to line broadening by the added quenching gas, and prevents possible temperature rise.7 Equations A, B, and C have been used to determine kH2 with entirely consistent results.7 As has been discussed previously,7 at the vapor pressure of Cd at 280 “C (even with the 5.0-mm diameter reaction vessel) the effective lifetime of Cd(3Pl) is slightly greater than the natural radiative lifetime because “imprisonment” of resonance radiation cannot be totally ignored. Using the diffusion theory of radiation imprisonment7 (which has recently been given theoretical justification,16 and should be particularly accurate in this case because the vessel, 80 cm long and 0.5 cm in diameter, is effectively an infinite cylinder with uniform radial illumination), the ratios T / T ~ , where T is the actual “imprisonment” lifetime and T~ the natural radiative lifetime, have been calculated for the conditions of interest, and are shown in Table I. The lifetimes for CH4 were also assumed to be valid for the same pressures of the other alkanes. Since only small ./TO corrections are involved, such an assumption will lead to less than 5% overall error. For the gases studied here, only cyclohexane quenched Cd(3PJ) a t a rate sufficiently high to use eq A and B directly to determine kQ, and a plot showing the validity of these equations in the cyclohexane case is shown in Figure 1. Because the values of hQfor the other alkanes are three to four orders of magnitude smaller than kHz, the procedures outlined above cannot be utilized directly. Since pressures of several hundred torr of alkane are required for appreciable quenching of Cd(3PJ), serious linebroadening errors could result. However, two kinds of experiments can be performed which do allow determination of kQ for the alkanes and also provide interesting information about the relative yield of CdH in the quenching of alkanes as compared to Hz:

The Journal of Physical Chemistry, Vol. 83, No. 3, 1979 305

Reaction of Excited Cd(3P,) Atoms with Alkanes



/' 00 I











l/c [Isobutane]

Flgure 1. A plot of the relative concentration of CCI(~P,) and Cd(3P,) vs. the pressure of cyclohexane. See ecl A and B. The conditions were as follows: 8 torr of N, total pressure always macle up to 400 torr with Ar; T = 280 'C; delay time 10 ps. A

/ [Cd(3P,)I,2



IO 0





(torr-' 'Io3)

Figure 3. A plot oi the ratio of the CdH yield with high pressure! of HP present to the CdH yield with only isobutane present vs. the inverse of the isobutane piressure. See sq E and the text. ( C = TIT,;see Table I.) The conditions were as follows: T = 280 "C;delay time, 40 ys.

TABLE 11: C d ( 3 P ~Quenching ) Rate Constants (L/molls) at T = 2 8 0 ° C _ _ method competitive Cd(3P~) quenching CdH formation gasby Hz (1.7 * 0.3) X (2.1 f 0.3) x H*7 10" lo'] (2.2 t 0.3) X CH, 107 (H,) (2.4 t 0.6) x 107 KO,) CH,CH, (6.8 * 0.7j x ( 9 t 7) X 107 10' CH,CH,CH, (3,O t 0.2) X (2.9 i 1.2) X 10' 107 (CH,),CHCH, (3.6 ?r 0.2) x (2.3 t 1.2) X ~



I 2


100 torr




0 05

0 IO


(torr) 0 15

0 20 I

0 25 ,

Figure 2. A plot of the relative concentration of Cd(3P,) and Cd(3P,) vs. the pressure of H,. See eq D. The conditions were as follows: 400 torr of isobutane; T = 280 " C ; delay time 10 ys.

(i) At a constant pressure of alkane (RH) of', e.g., 500 torr the following steady-state equation can bie derived:

where Q is any added quencher besides RH, and kRH is defined in the same way as kQ, with Q = RH. A similar equation may be written for Cd(3PJ. A t constant [RH], vs. [Q] should yield a a plot of [Cd(3po)](Q=O)/[Cd(3PO)] straight h e , with a slope from which k R H can be determined if k, and kQ are known. A representative plot indicating the validity of eq D is shown in Figure 2 for isobutane with Q = Hz. (ii) I f experiments are performed a t varying pressures of pure RH, and then repeated at each pressure of RH with sufficient H2 present to react with all the Cd(3Po,l)atoms, the following equation can be derived: [CdHl [Hz]=m -___

Rkr (E) [CdHlRH RH [RH] where R = [CdH] , / [ c d H ] [ ~ ~ l . [CdH],, ~; is the relative CdH yield?gen all Cd(3Po,l)is quencfned by H, only, and [CdH][,,,=, is the relative CdH yield when all Cd(3Po,l)is quenched by RH only. Relative yields of CdH with and without 5 torr of H z present were measured for various pressures of ethane, propane, isobutane, and cyclohexane. The yield of CdH from CH4 was insufficient for accurate measurements. The CdH yield with only RH



107 -_ (5.8 ?: 0.9) x



( 4 4 f 0.5) X


present was equated to [ C ~ H I R H and , the CdH yield with 5 torr of Hz present was equated to [CdH],Hzl=m (this pressure of Hz is sufficient to quench 296% of aU Cd('P0,J under all conditions studied). A representative pllot of [ C ~ H ] [ H ~ ~ = ~ / [ C ~vs.Hr0/7[RH] ] [ R H I for isobutane is shown in Figure 3. Note that for any particular [RH], k, = 4.35 x 1 0 ~ ( ~ ~ in / 7 sir1 ) , (see Table I). Values of kRH can thus be calculated from the slope-to-intercept ratio of plots such as the one in Figure 3. Shown in Table I1 are values of kQ determined using both methods. The amounts of CdH from the reaction of Cd(3PJ) with CH4 were insufficient for reliable measurements, but experiments with Q = COz gave virtually an identical values of kCH4.The large uncertainty in one of the kQ values for ethane is also due to a very low apparent yield of CdH. The experiments with cyclohexane were conducted with the total pressure made up to 4001 torr with a 2% N 2 / k mixture? Quoted error limits were taken from calculated slope and/or intercept standard deviations of a least-syuarles data-fit routine. The rate constants obtained by the two different methods are in reasonable agreement. In Table I11 are shown cross sections for the quenching of equilibrated Cd(3P0,,)a t 280 "C, calculated using an average of the two values of k, in Table 11, weighted by the inverse of each listed uncertainty (Note that we have

W. H. Breckenridge and A. M. Renlund

The Journal of Physical Chemistry, Vol. 83, No. 3, 1979


TABLE 111: Quenching Cross Sections and Relative CdH Yields in the Quenching of Cd(3PJ)a AH composite (CdH Cd(3P~) forma- quenching tion), cross kcal/ section, relative gas mol A2 CdH yield H2’ CH, CH,CH, CH,CH,CH, (CH,),CHCH, c-C,H,, CH,=CHCH,’ a UQ

tO.0 -0

12.3 1.0 0.0041