J. CHEM. SOC. FARADAY TRANS., 1994, 90(14), 2003-2007 Kinetic Study in a Microwave-induced Plasma Afterglow of the CU(~~S)Atom Reaction with N20from 458 to 980 K and with NO, from 303 to 762 K Chris Vinckier,* Tom Verhaeghe and lnge Vanhees Laboratory for Analytical and Inorganic Chemistry, Department of Chemistry, K. U. Leuven Celestijnenlaan 200F 300 1 Heverlee , Belgium The rate constants for the reaction of ground-state copper atoms (4 ’S) with N,O and NO, have been derived in a fast-flow reactor. The microwave-induced plasma (MIP) afterglow technique was used for the generation of copper atoms in the gas phase. The rate constant for the reaction Cu(4’S) + N,O CuO + N, at temperatures between 458 and 980 K is found to be (2.4 k0.6) x lo-’’ exp -(48.6 +_ 1.4 kJ mol-’/RT) cm3 molecule-’ s-’.The reaction Cu(4’S) + NO, -+ products, was followed in the temperature range from 303 to 762 K and has a rate constant (1.3 0.6) x lo-’’ exp -(0.2 f1.3 kJ mol-’/RQ cm3 molecule-’ s-’. From the copper atom decays in argon at 528 K, a diffusion coefficient, DCu,Ar 543.2 f60.2 cm2 Torr s-’, could be derived. = The introduction of the high-temperature fast-flow reactor (HTFFR) technique by Fontijn has allowed the determi- nation of a large number of kinetic parameters of refractory metal atom-gas reactions over a temperature range of between 300 and 2000 K.’ Recently the high-temperature photochemistry (HTP) technique, has been developed. While in the former technique metal atoms were generated mostly by thermal heating of the metal, in the HTP technique, metal atoms are produced photolytically and are monitored in a real-time detection mode.An alternative method used to vaporize non-volatile metals is the plasma-afterglow atomization technique3v4 where the metals are generated in the gas phase by a reaction of a vola- tile metal salt, MeX(g), with hydrogen atoms. In this work the reaction between copper atoms, dinitrogen oxide and nitrogen dioxide has been investigated in a fast- flow reactor. The reaction with dinitrogen oxide Cu(4 ’S) + N,O -+ CuO + N, has been studied in the temperature range from 458 to 980 K and the Arrhenius expression for k, will be derived and com- pared with the recent measurements by Narayan et ul.’ on the same reaction in the somewhat broader temperature range of 470 to 1340 K.The other reaction with nitrogen dioxide Cu(4 ,S) + NO, -+ products was followed between 303 and 762 K. No other kinetic data on reaction (2) are yet available in the literature for compari- son. Experimental A schematic view of the slightly modified experimental technique3 is shown in Fig. 1. It is basically a quartz fast-flow reactor with an internal diameter of 5.7 cm and a length of 1 m. At the upstream end the water-cooled flange carries a carrier-gas inlet (GI)and a sample holder (SH) with an exter- nal diameter of 1.9 cm. A Kanthal resistance wire allows the solid CuCl pellet to heat up to a temperature, T,, of cu. 800 K independently of the reactor temperature.A shielded chromel-alumel thermocouple (TC in contact with the CuCl(s) pellet allows the temperature, T,, of the solid to be monitored. At a distance of 13.5 cm downstream of the sample holder, a second gas inlet (G,) is equipped with an air-cooled microwave cavity, type 216 L, powered by an Electro Medical Supplies (EMS) microwave generator oper- ating at 2450 & 25 MHz and with a maximum power.of 200 W. Between the cavity and the reactor a Wood’s horn traps the UV light from the microwave-induced plasma (MIP). Downstream of G, the reactor oven (RO) allows the gas tem- perature, T, to be varied in the range 300-1000 K. Con-stant temperature can be maintained within 2.5% over a distance of ca. 25 cm. At the downstream end the water-cooled flange carries two quartz probes : an along-the-axis movable thermocouple (TC,), to measure the gas temperature, and the additive inlet (AI) for the introduction of the co-reagent, N,O or NO,.An Alcatel-type 2033 double-stage oil rotary pump with a nominal pump capacity of 35 m3 h- ’ results in a flow veloc- ity for argon of 309 cm s-’at 295 K. Pressures are measured by means of a Datametrics-type 1018 electronic manometer powering a sensor (P) which measures pressures from to 13 Torr. In the kinetic zone the copper atoms are detected by atomic absorption spectroscopy at 324.7 nm. The light path of the hollow cathode lamp HCL is focussed perpendicular to the main reactor axis through a slit of 1 cm height in the reactor oven construction.The light exiting the reactor is then focussed on the entrance slit of a McPherson model 270 monochromator (M). A photomultiplier (PM) Hamamatsu R955 with a wavelength range 160-900 nm is used as a detector. While the detection system remains at a fixed posi- tion, the fast-flow reactor assembly is mounted on a carriage C 0HCL Fig. 1 Schematic view of the experimental set-up: G, and G,, gas inlets; AI, N,O or NO, inlet; SH, sample holder for the CuCl(s) pellet; TC, , thermocouple for the CuCl(s) pellet; C, microwave cavity; HW, horn of Wood; RO, reactor oven; P, pressure sensor; RP, rotary pump; TC, , thermocouple for gas temperature; HCL, hollow cathode lamp; M, monochromator; PM, photomultiplier; A, amplifier; R, recorder.which allows a horizontal displacement so that copper absorbances can be measured along the reactor axis. Typical initial absorbances in the kinetic zone are 50.3, corresponding to an upper limit for the copper atom concen- tration of 4.3 x lolo atoms cmP3. In this way the concentra- tion of copper atoms remains much lower than the additive N20 or NO, concentration so that pseudo-first order condi- tions for copper atom decays are fully established. The gases used are argon and dinitrogen oxide from UCAR with a purity in excess of 99.999% and 99.5%, respec-tively. Hydrogen is from L'Air Liquide with a quality of 99.9997%. Nitrogen dioxide was added as a 0.92% mixture in UHP helium (UCAR). Gas flows were regulated via Brook's precision needle valves of ELF type.Microwave-induced Plasma (MIP) Afterglow Atomization of CuCl A number of aspects of the MIP-afterglow atomization of CuCl have been discussed previ~usly.~.~ In principle, the copper chloride oligomers, Cu,Cl,(g), produced by the vapor- ization of the CuCl pellet are mixed with the reaction pro- ducts of the MIP-afterglow downstream of G,. A complex and unknown reaction sequence converts a fraction of Cu,Cl,(g) into Cu atoms through subsequent atomic hydro- gen reactions : Cu,CI,(g) + nH +Cu(g) + products (3) Owing to a loss process on the reactor wall, the copper con- centration reaches a maximum within a timescale of less than 30 ms which corresponds to a distance of about 10 cm down- stream of the MIP inlet, G,.Since one can assume that the sticking coefficient, ycu ,on the reactor wall approaches unity, the decay of Cu(g) is diffusion controlled and the observed decay constant k, is given by:778 where DCu,Aris the binary diffusion coefficient of copper atoms in argon and r the reactor radius. As an example, the values of k, averaged over four runs are given in Table 1 for various initial hydrogen concentrations. The initial hydrogen concentration has no systematic effect on the value of k,. As a weighted average for the five deter- minations, one obtains k, = 30.7 & 3.4 s-' at a reactor press- ure P, = 8 Torr and temperature = 528 _+ 2 K. From eqn. (I) one can calculate the binary diffusion coefficient D,,, Ar = 67.9 f7.5 cm2 s-' which can be reduced to a reference pres- sure of 1 Torr : Dcu,Ar = 543.2 f60.2 cm2 Torr s-'.With our earlier value,4 D,,, = 219.9 _+ 9.2 cm2 Torr s-' determined at Tp = 300 K, one can calculate the temperature dependence of the diffusion coefficient from the relation D,,, = D3,,(528/300)". From the measurements at 300 and 528 K one obtains n = 1.6 & 0.25 which is somewhat larger than n = 1.5 calculated on the basis of pure kinetic gas consider- Table 1 Decay constants, k,, of the Cu atoms as a function of the initial hydrogen concentration" [H ,]/mTorr T,/K TJK k&-' 5 530 533 26.5 f1.5 15 527 534 33.3 f0.7 50 527 532 23.6 11.5 100 527 534 28.9 f3.8 200 530 534 21.4 f6.1 a Reactor pressure P, = 8 Torr and the microwave power P, = 40 W. J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 ations, but which is in qualitative agreement with the average temperature dependence, derived from the more elabo- rate Lennard- Jones f~rmalisrn.~ Results and Discussion Cu + N,O Reaction The rate constant for reaction (1) can now be determined from the copper atom decay as a function of the reaction time under various amounts of added N,O. The mathemati- cal treatment is the same as used in our previous work on the kinetics of the Mg + Cl,, N,O reactions:"." ~ +In A,, = -{k1c:201 7*34DCu, Ar}i + B (11)2r2 in which A,, is the copper absorbance, t, the reaction time, B, an integration constant and q, a correction factor. A complete discussion on the mathematics behind eqn.(11) and the influ- ence of the various flow characteristics on the magnitude of q is presented in the earlier work of Talcott et aL7 and Fontijn and Felder,' respectively. The value of k, may be determined by following the copper absorbance A,, as a function of the reaction time, t, at various amounts of added N,O. When the slopes, S, derived from eqn. (11) are plotted against [N,O], straight lines are obtained with an intercept of 7.34Dc,, Ar/2r2 and a slope equal to kl/q. For the measurements in argon the correction factor, q, is set equal to 1.3 with an associated sys- tematic error of lo%, and the overall accuracy is estimated to be in the range 40-70%.' All plots and calculations are made using the SAS-607 statistical package.', The uncertainties are given as standard deviations.As an example to illustrate the procedure used the natural logarithm of the copper absorbance In A,, is plotted against the reaction time for various N,O concentrations, Fig. 2, at a reaction temperature = 609 K and a pressure P, = 10 Torr and with the MIP-afterglow parameters set at T, = 539 K and [H2] = 200 mTorr. When the slope, S, of these lines is plotted us. the N,O concentration, Fig. 3, a straight line is obtained from which a weighted linear regression yields a value for k, = (1.1 & 0.2) x cm3 molecule-' s-'. -2.5 -3.0 -3.5 TI V 7 E -4.0 -4.5 -5.0 -5.5 0 10 20 30 4b 50 E I reaction time/ms Fig.2 Natural logarithm of the Cu absorbance as a function of the reaction time. The experimental conditions are: P, = 10 Torr; T, = 609 K; [HJ = 200 mTorr; T, = 539 K and P, = 40 W. The N,O concentrations are (0)0; (+) 1.6; (0)3.2; (A) 4.0; (x) 4.8 and (V) 6.4 each in units of 10l5molecules crnp3. J. CHEM. SOC. FARADAY TRANS., 1994,VOL. 90 80 60701 $ 4050j 30v *O/10 0 2 4 6 [N,Ol/l Owl5 molecules ~rn-~ Fig. 3 Observed slope, S, of eqn. (11) plotted against the N,O con-centration. The experimental conditions are the same as in Fig. 2. The five points shown at the ordinate are the observed copper decays in the five blank experiments in the absence of the co-reagent N20. Influence of the MIP-afterglow Parameters In order to verify that the MIP-afterglow conditions have no effect on the magnitude of the derived rate constants, various parameters such as the hydrogen content, temperature of the CuCl pellet and reactor pressure have been varied and their influence on k, checked.The experimental conditions are summarized in Table 2. While part of the scatter in the values of k, is due to the fact that neither < nor T, could be kept rigorously constant, one cannot see a systematic effect of the hydrogen content on the derived value of k, . Also a variation of T, between 530 and 551 K does not show an effect on k, . It is known that the sublimation of CuCl occurs via the for- mation of oligomers Cu,Cl,.'3 On the basis of their subli- mation heat one can calculate that the gas-phase concentration increases by a factor of three in the narrow temperature range 530-551 K.3-6 Variation in both the hydrogen concentration and T, seriously affects the MIP- afterglow composition but this does not seem critical for the kinetic measurements.Table 2 Influence of the MIP-afterglow parameters on the value of k, for the Cu +N,O reactiona T,/K T,/K [H,]/mTorr P,/Torr k1/cm3 molecule-' s-' hydrogen content 530 527 527 527 530 533 534 532 534 534 5 100 200 15 50 8 8 8 8 8 (3.4 0.5) x 10-15 (2.1f0.6)x 10-15 (4.4f 1.1) x 10-15 (3.0f0.5)x (2.8f0.5)x temperature, T,, of the solid 682 682 682 68 2 530 539 544 551 20 10 10 10 8 8 8 8 (4.3k 0.9)x 10-14 (3.7f0.8)x 10-14 (4.0k 0.8)x (5.4f 1.3)x reactor pressure, P, 483 485 486 485 529 532 532 529 2.5 2.5 2.5 2.5 6 6.5 8 11 (2.0 f0.6) x 10-15 (2.1k 0.5)x 10-15 (2.4f0.4)x (1.7k 0.4)x Microwave power, P, = 40W.Another MIP-afterglow parameter is the reactor pressure, P,, which was varied between 6 and 11 Torr. Even in this narrow range all MIP-afterglow phenomena related to the diffusion of the reagents and the copper atom wall loss are enhanced at a lower pressure. The initial absorbance Aocu at 11 Torr was indeed a factor of two larger than at 6 Torr. Table 2 shows that P, has no effect on the derived value of kl . Temperature Dependence of k A summary of the experimental conditions and derived values for k, is given in Table 3 at temperatures between 458 and 980 K. A weighted non-linear regression yields the expression k, = (2.4 & 0.6) x lo-'' exp -(48.6 & 1.4 kJ mol-'/RT) cm3 molecule-' s-' (111) The only other determination of k, has recently been carried out by Narayan et al.' using the metals-HTP technique in which copper atoms are photochemically produced at 248 nm by excimer laser dissociation of thermally generated Cu2C12 or CuF, molecules.The copper atom decays were monitored in real time by means of resonance fluorescence. The overall fit in the entire temperature range 470-1340 K yielded k, = 3.04 x T2.97 exp-(3087 K/T) cm3 molecule-' s-'. In view of the observed non-Arrhenius behaviour at temperatures above 1190 K, the classical Arrhe- nius expression was only derived at temperatures below 1190 K and was given by: k, = 1.70 x lo-'' exp -(5129 K/T) cm3 molecule-' s-' (IV) An Arrhenius plot of our results as In k, us.1/T is shown in Fig. 4 and gives the relation k, = 1.8tg:z x lo-'' exp -(46.9 & 1.2 kJ mol-'/RT) cm3 molecule-' s-'. In view of the symmetric error on the value of the pre-exponential factor, eqn. (111)will be used in the further discussion. When these data are compared with the work of Narayan et ~l.,~ Table 3 k, for the Cu +N,O reaction as a function of temperature T,/K T,/K [H,]/mTorr P,,Torr k1/cm3molecule-' s-l 458 468 483 485 485 486 527 527 527 530 530 546 609 621 627 682 682 682 682 704 728 774 797 818 858 919 980 524 542 529 532 529 532 532 534 534 534 533 529 539 543 539 5 30 544 539 551 529 539 532 544 532 527 529 532 2.5 5 2.5 2.5 2.5 2.5 50 100 15 200 5 2.75 200 20 20 20 10 10 10 4.5 4 20 20 5 5 5 5 8 8 6 6.5 11 8 8 8 8 8 8 8 10 8 8 8 8 8 8 8 8 8 8 8 8 8 8 (9.1f3.4)x (1.6k 0.3)x (2.2f0.5)x 10-' (1.7k0.4)x (2.4f0.4)x (2.8 0.5) x (3.0f0.5)x (2.0_+ 0.6) x 10-15 (2.1 0.6) x 10-15 (4.4f 1.1) x 10-15 (3.4f0.5)x 10-15 (1.1 f0.2) x 10-14(6.3f 1.4)x (1.9f0.4)x (1.9 f0.4)x (4.3f0.9)x (4.1f0.8)x (3.7 0.8)x 10-14 (5.4 1.3) x 10-14 (4.7f0.7) x 10-14 (7.0f 1.1) x 10-14 (1.4f0.2)x 10-13 (1.8f0.3)x (1.8f0.3)x (2.8f0.5)x (6.6f1.3)x (5.3 1.1) x 10-13 2006 -27 '. '.-35 \1 --, I. I, 1. I. I I I-%! I I I I 1.0 1.2 1.4 1.6 1.8 2.0 2.2 103 KIT Fig. 4 Arrhenius plot of In k, us. 1/T. (-) Non-linear regression of k, us. 1/T; (---) linear regression of In k, us.1/T; (-. -) Narayan et al.' one sees that our values of k, are systematically lower over the entire temperature range. This is mainly due to the lower Arrhenius activation energy of 42.6 kJ mol- ' compared with 48.6 kJ mol-' derived from our work. In the middle of the temperature range at 720 K, the value of k, determined by Narayan et ~l.,~is a factor of two higher, which falls outside the systematic uncertainty level of the correction factor, q, in eqn. (11). Therefore, there seems to be no direct explanation for the observed discrepancy between both data sets. It is interesting to compare the values of the observed acti- vation energy of k, with the barrier heights for a number of alkali-metal atom-N,O reactions.For these reactions much smaller activation energies, in the range of 5.1 and 14.6 kJ mol-',have been mea~ured.'~.'~ This has been explained by semi-empirical correlation between the so-called promotion energy of the metal atom and the observed energy barrier of the rea~tions.~,'~"' The former has been defined in two ways, being either the excitation energy to the lowest excited state of the metal atom or the excitation energy plus its ionization energy. On the basis of this latter definition one would expect a barrier for the Cu + N,O reaction at around 35 kJ mol-' which is about 20-40% lower than the experimental values. Note that by extrapolating to room temperature (300 K) k, = 8.3 x lo-'' cm3 molecule-' s-', which is almost an order of magnitude lower than 6.4 x cm3 molecule-' s-' calculated from eqn.(IV). In view of the large value of the activation energy, an uncertainty of only 5% leads to a spread of a factor of seven for k, when extrapolated to 300 K. This reinforces the need for kinetic measurements of atom reactions over a wide tem- perature range. In this way correlations between the metal atom-N,O rate constants and the promotion energy can be established at e.g. 600 K, avoiding the enormous uncer-tainties when extrapolating to room temperature. Concerning the products of reaction (1) there is no doubt that CuO is formed in its electronic ground state X21-I. With an exothermicity18 of 113.2 kJ mol-l for reaction (l), the lowest electronically excited state A' 2X+at 181.1 kJ mol-' above the ground statelg will not be accessible.Cu + NO, Reaction The experimental procedure described above is also followed to determine the Arrhenius expression of reaction (2) in the J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 temperature range 303 to 762 K. It was again checked that no systematic effect on k, of the molecular hydrogen concen- tration, the temperature, T,, of the solid and the reactor pres- sure, P,, could be observed. A summary of the experimental conditions and the derived values for k, are shown in Table 4. At 305 K the average value of k, was found to be (1.8 0.8) x lo-'' cm3 molecule-' s-'. With helium as carrier gas, k, was found to be (1.9 f0.6) x lo-'' cm3 molecule-' s-l and hence the nature of the carrier gas also has no influence on the value of k, .A weighted non-linear regression gives the Arrhenius expression : k, = (1.3 f0.6) x lo-'' exp -(0.2 & 1.3 kJ mol-'/RT) cm3 molecule- ' s-' (V) The Arrhenius plot of In k, vs. 1/T is shown in Fig. 5. From the slope the very slightly different expression k, = 1.0:;:: x lo-'' exp(l.2 f1.2 kJ mol-'/RT) cm3 molecule-' s-' (VI) can be derived. Eqn. (V) shows a slightly positive, and eqn. (VI), a negative activation energy, but taking into account the rather large uncertainty on k, , k, is virtually temperature independent between 303 and 762 K. This is in sharp contrast to reaction (1) which has a large energy barrier of 48.6 kJ mol-'. In view of the large value of 2.36 eV for the electron affinity20*2' of NO,, the metal atom-NO, reaction is likely to occur via the 'electron jump' mechanism,, similar to a number of other metal atom-halogen reactions.The distance rc, where the potential-energy curve of the two approaching Cu and NO, Table 4 k, for the Cu + N,O reaction as a function of temperature ~~ ~ ~~ T,/K T,/K [H,]/mTorr P,/Torr k1/cm3 molecule-' s-' 305 551 200 7.5 (0.7 f0.1) x lo-" 305 559 25 7.5 (1.0 f0.1) x lo-" 305 552 6 7 (1.2+_ 0.2)x lo-" 305 557 3.5 7.5 (2.9 f0.6) x lo-" 303 516 10 9 (0.8 f0.2) x lo-" 303 528 10 9 (1.4 f0.3) x lo-" 303 537 15 9 (2.0 f0.4) x lo-" 303 541 30 5 (2.9 f0.4) x lo-" 303 535 200 6 (1.8 +_ 0.3) x lo-" 303 529 100 7 (2.0 f0.2) x lo-" 303 535 200 7.5 (2.7 f0.4) x lo-" 303 521 200 8 (1.6 f0.3) x lo-'' 303 515 30 10 (1.3 f0.4) x lo-" 305 573 10 6 (3.3 f0.5) x lo-" 332 539 3 7 (1.6 f0.4) x lo-" 335 549 5 7.6 (2.5 & 0.5)x lo-" 353 542 4 12 (2.1 f0.3) x lo-" 355 545 3 8.5 (1.3 f0.2) x lo-" 357 542 3 10 (1.7 f0.3) x lo-" 357 522 5 6.5 (1.5 f0.3) x lo-'' 362 539 3 5.5 (0.8 f0.1) x lo-" 397 527 3.5 7 (1.8 & 0.3)x lo-" 413 554 6 6 (1.6 f0.3) x lo-" 513 532 2 5.5 (1.0 f0.2) x lo-" 624 532 4.5 7 (1.1 f0.2) x lo-" 694 529 3 10 (1.4f0.2)x lo-" 762 55 1 9.5 7 (1.7 f0.4) x lo-" 303" 530 5 9 (1.3 & 0.1) x lo-" 303" 53 1 30 9 (2.6 f0.2) x lo-" 303" 526 10 9 (2.6 & 0.5) x lo-" 303" 545 170 10 (1.6 f0.1) x lo-" 303" 526 10 12 (1.6 f0.3) x lo-" 303" 543 10 13 (1.9 f0.3) x lo-" Measurements in helium carrier gas.J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 -20 -21 --22 --23 --37 I I 2007 conditions. It is indeed very unusual that a reaction involving only four atoms has already reached its high-pressure limit at around 10 Torr. Finally, it should be mentioned that the error quoted for the heat of formation,18 Af H" (298 K), of CuO is 41.8 kJ mol-l. This leaves the possibility that reac- tion (2) may be slightly exothermic in which case a common second-order reaction is dealt with. The authors are grateful to the Joint Fund for Basic Research (FKFO) for funding this project.T.V. acknowledges a docto- ral fellowship from the Institute for Scientific Research in Agriculture and Industry. I.V. and C.V. are, respectively, Research Assistant and Research Director of the National Fund for Scientific Research (NFWO), Belgium. References-2gd-30 1.3 1.5 1.7 1.9 2.1 2.3 2.5 2.7 2.9 3.1 3.3 1 A. Fontijn and W. Felder, in Reactiue Intermediates in the Gas 103 KIT Fig. 5 Arrhenius plot of In k, us. 1/T. (-) Non-linear regression of k, us. 1/T; (---) linear regression of In k, us. 1/T. species crosses the Coulomb interaction curve of the Cu' and NO,-ions, is given by: rc = 14.35/[Ei(Cu) -E,,(NO,)] The distance rc is given in A, the ionization energy, Ei, and electron affinity, E,,, are expressed in eV and 14.35 is a con- stant.Inserting an ionization energy of 7.73 eV for Cu atoms, eqn. (VII) yields the value of rc = 2.7 A. With this value of rc , the rate constant for the reaction, nr:E, where C is the average thermal speed of the reaction partners, can be calculated. At 300 K k, is calculated to be 1.12 x lo-'' cm3 molecule-' s-'. Table 4 shows that this calculated rate constant is about a factor of five larger than the experimentally determined value so that the pure 'electron jump' mechanism for reac- tion (2) seems unlikely. Instead it would be better to call this a 'close range' charge-transfer mechanism. An interesting point concerns the possible products of reaction (2).With a formation enthalpy A,H" (298 K) of 306.27 kJ mol-' for CuO in the gas phase, the reaction Cu + NO, -+ CuO + NO is endothermic" by 25.86 kJ mol-'. If reaction (2) was to yield CuO one should expect an energy barrier of at least 25.86 kJ mol-' which is not observed. In view of the absence of neither a pressure nor a carrier gas effect on the observed decay constant, a third-order reaction can be excluded. A remaining possibility is the formation of a so called 'sticky Cu+ NO,- collision ~omplex'.~~.~~ While most of the metal-NO, reactions lead to the formation of the metal o~ide,,~-'~ the only major exception seems to be the Cs + NO, reaction. This reaction, which is endothermic by 43.45 kJ mol-', has been shown to have a reaction cross-section exceeding23 100 A2, which is substantially larger than 3.89 A2 determined experimentally for reaction (2).This again confirms that the 'electron jump' mechanism for the reaction Cu + NO, certainly does not prevail. However, from a fun- damental kinetic point of view it is not clear how the complex Cu/NO, can be stabilized under our experimental Phasc., Generation and Monitoring, ed. D. W. Setser, Academic Press, New York, 1979, ch. 2. 2 P. Marshall, A. Narayan and A. Fontijn, J. Phys. Chem., 1990, 94, 2998. 3 C. Vinckier, A. Dumoulin, J. Corthouts and S. De Jaegere, J. Chem. SOC., Faraday Trans. 2, 1988,84, 1725. 4 C. Vinckier, J. Corthouts and S. de Jaegere, J. Chem. SOC., Faraday Trans. 2, 1988,84, 1951.5 A. Narayan, P. Futerko and A. Fontijn, J. Phys. Chem., 1992,%, 290. 6 C. Vinckier, P. Christiaens and M. Hendrickx, in Gas-Phase Metal Reactions, ed. A. Fontijn, Elsevier, Amsterdam, 1992, ch. 4. 7 C. Talcott, J. Ager I11 and C. Howard, J. Chem. Phys., 1986, 84, 6161. 8 J. Silver, J. Chem. Phys., 1984,81, 5125. 9 J. Hirschfelder, C. Curtiss and R. Bird, Molecular Theory of Gases and Liquids, Wiley, New York, 1954, ch. 8. 10 C. Vinckier and P. Christiaens J. Phys. Chem., 1992,%, 2146. 11 C. Vinckier and P. Christiaens, J. Phys. Chem., 1992,%, 8423. 12 SAS statistical package, SAS Institute Inc., Cary, NC, 1992. 13 M. Guido, G. Balducci, G. Gigli and M. Spoliti, J. Chem. Phys., 1971,55,4566. 14 D. Husain and Ji Bing, Combust. Flame, 1990,79,250. 15 J. M. Plane, in Gas-Phase Metal Reactions, ed. A. Fontijn, Else- vier, Amsterdam, 1992, ch. 3. 16 J. M. Plane, C-F. Nien and B. Rajasekhar, J. Phys. Chem., 1992, %, 1296. 17 P. Futerko and A. Fontijn, J. Chem. Phys., 1991,95,8065. 18 M. W. Chase, C. H. Davis, J. R. Downey and D. Y. Frurip, J. Phys. Chem. Ref. Data, 1985, 14, Suppl. No. 1. 19 J. M. Delaval, F. David, Y. Lefebre, P. Bernage, P. Niay and J. Schamps, J. Mol. Spectrosc., 1983, 101, 358. 20 E. Herbst, T. A. Patterson and W. C. Lineberger, J. Chem. Phys., 1974,61,1300. 21 E. P. Grimsrud, G. Caldwell, S. Chowdhury and P. Kebarle, J. Am. Chem. Soc., 1985,107,4627. 22 D. R. Herschbach, Adu. Chem. Phys., 1966,10, 319. 23 D. 0. Ham, J. L. Kinsey and F. S. Klein, Discuss. Faraday Soc., 1967,44,174. 24 R. R. Herm and D. R. Herschbach, J. Chem. Phys., 1970, 52, 5783. 25 D. D. Parrish and R. R. Herm, J. Chem. Phys., 1971,54,2518. 26 H. F. Davis, A. G. Suits, H. Hou and Y. T. Lee, Ber. Bunsenges. Phys. Chem., 1990,94, 1193. Paper 4/01443J; Received 1 lth March, 1994