The reaction H+ClCN to CN+HCl

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Request accessible transcript

Transcript (if available)

Content THE REACTION H + C1CN — » CN + HC1 by Susan Hansen Callister A Thesis Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree MASTER OF SCIENCE (Chemistry) August 1988 UMI Number: EP41675 All rights reserved INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted. In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if material had to be removed, a note will indicate the deletion. Dissertation .FublisMng UMI EP41675 Published by ProQuest LLC (2014). Copyright in the Dissertation held by the Author. Microform Edition © ProQuest LLC. All rights reserved. This work is protected against unauthorized copying under Title 17, United States Code ProQuest LLC. 789 East Eisenhower Parkway P.O. Box 1346 Ann Arbor, Ml 48106-1346 UNIVERSITY O F S O U T H E R N CALIFORNIA TH E GRADUATE SCH O O L UN IV ER SITY PARK LO S A N G ELES. CA LIFO R N IA 9 0 0 0 7 c C/62 ?3? n - v i This thesis, written by Susan Hknsen Callister under the direction of h...QJ£:Thesis Committee, and approved by all its members, has been pre sented to and accepted by the Dean of The Graduate School, in partial fulfillm ent of the requirements fo r the degree of Dean D ate..... TI ITTEE Chairman ACKNOW LEDGEM ENTS Many thanks go to Curt Wittig, my thesis advisor, for providing plentiful funding and advice; I am also indebted to Hanna Reisler, C.X.W. Qian, and the other members of the group for many helpful discussions. I want to express my most sincere appreciation to Sarah Novak for typing this thesis and other papers. Thanks go to Gerald Segal for calculations, and most importantly to all my coworkers while at USC - especially Julian de Juan without whose help this thesis could not have been written. ii TABLE OF CONTENTS Page ACKNOWLEDGEMENTS ii LIST OF TABLES iv LIST OF FIGURES v INTRODUCTION 1 EXPERIMENTAL 5 RESULTS AND ANALYSIS 8 DISCUSSION 22 Calculations Chlorine attack Carbon attack Nitrogen attack SUMMARY 29 REFERENCES 30 APPENDIX 32 LIST OF TABLES Page I. Parameters used in fitting CN(X2L) Doppler profiles. 16 II. Average energies and fractions of available energy associated with different degrees of freedom. 19 III. Results of the prior calculation compared to experiment. 23 IV. Energy disposal in several L+HH reactions. 23 V. Calculated vs. experimental energies for the dissociation of the molecules associated with the HC1CN potential surface. 24 LIST OF FIGURES Page 1. Schematic drawing of the experimental arrangement. 2. Spectrum of CN resulting from a flowing 1:1 mixture of C1CN and H2S. 9 3. UV absorption spectrum of C1CN. 10 4. Spectrum of CN from a flowing 1:1 mixture of C1CN and Helium. 11 5. Plot of CN(X2E) rotational populations vs. N(N+1). 13 6 . (a) Doppler profile for anisotropic distribution. 14 (b) Doppler profile for isotropic distribution. 14 7. Doppler profiles of three CN B2£<— X2X R-branch transitions. 20 8 . (a) Estimate of populations vs. center-of-mass kinetic energy. 21 (b) The 'reflection' of (a), showing HC1 population vs. internal energy. 21 9. (a) Reaction coordinate diagram for HC1CN. 27 (b) Stationary points on the HC1CN potential energy surface. 28 & v IN T R O D U C T IO N The field of molecular dynamics is a relatively new one to physical chemistry. In the past thirty years, technologies have become available which allow chemical reactions to be studied in detail never before imagined. These include molecular beam techniques which first allowed the study of single-collision events and the development of the laser which can provide an extremely narrow line width and high intensity for use in various forms of spectroscopy. Now, not only can the products of a reaction be ascertained, but their nascent rovibrational and translational distributions may be probed. It is also desirable to know the initial states of the reactants, ideally including their energies (and energy distributions) and their orientation and alignment relative to each other. Recently, sub-Doppler resolution spectroscopy has been used with laser-induced fluorescence (LIF) measurements.1 This allows the correlation of internal (mainly rovibrational) and translational degrees of freedom, because kinetic energy distributions are obtained for specific quantum states. Reactions of hydrogen atoms have been studied for many years. Only recently, however, have methods become available to very narrowly select the initial kinetic energy of the H atoms. By dissociating selected precursors with excimer lasers, H atoms with well-defined, narrow kinetic energy distributions may be produced. There have been several studies done of reactions with such atoms, and these have been recently reviewed by Flynn.2 In such systems, if the enthalpy change for the reaction is well-known, the energy available to the system is well- defined. For the title reaction, H atoms were produced by dissociating H2S with the output from a KrF excimer laser (248 nm). This results in a center-of-mass (c.m.) kinetic energy of 21.6 kcal moH.3 Two other reaction channels are also available to the system, and the complete set of energetically-allowed, chemically- distinct products is given by:4> 5 1 H + C1CN -> HC1 + CN -> Cl + HCN -» Cl + HNC AH298 = -6.1 kcal mol' 1 AH298 = -23.9 kcal mol' 1 AH298 = -8.7 kcal mol' 1 ( 1) (2) (3) For reaction (1), the average energy available to the system at 298 K is Et = Ec.m.(H+ClCN) + 3/2 kT + Eint(ClCN) - AH298 (4) AH298 is = -6.1 kcal mol' 1,4 Ec.m.(H+ClCN) is approximately 21.6 kcal m ol'1, 3 and 3/2 kT = 0.9 kcal m ol'1. The C1CN internal energy is calculated as follows:6 where vibrations are taken as harmonic, and g>i = 714.52 cm-1, 0 2 = 0)3 = 378.3 cm' 1 and 04 = 2215.5 cm' 1.7 Thus, Eint(ClCN) = 1.1 kcal m ol'1, and E t, the total energy available to reaction (1), is 29.7 kcal mol*1. The fate of the portion of a molecule not directly involved in a chemical reaction, the spectator, has not received much attention in the past. For some systems, the spectator portion is the only one amenable to facile detection. We are interested specifically in the fate of the cyanide radical in reaction (1). There are two extremes in the possible fate of the 'spectator' part of the molecule. First, during the reaction the available energy may flow throughout all degrees of freedom of the reaction complex and the 'spectator' may finish in quite a different range of rovibrational states than those in which it would end up if it did not fully sample the available energy. This is likely to happen if the reaction complex is long-lived. In this case, the predictions for the energy distribution in the products from statistical theories would be useful since such predictions are the result of assuming that the available energy is distributed statistically throughout the Ejnt(ClCN) — Erot + Evjb — RT + Eyjb (5) Eyib(ClCN) = I (6) 2 product states. A first level of approximation is given by a 'prior distribution' which conserves and statistically distributes the energy but does not include angular momentum conservation, which plays a minor role in large molecular systems. A prior calculation for reaction (1) is described in detail in the Appendix. At the other extreme, all of the energy available to the reaction may be confined to the reactive site and then the rovibrational levels of the 'spectator' would be much colder than predicted by statistical theories. In such a 'direct reactive encounter’ the reactants would collide and go directly to products. In the language of potential energy surfaces, in the first case the reactive system would at some point be trapped in a well, stay long enough so that the energy would flow throughout all the degrees of freedom and then find its way to products. In the second case, reactants would travel through the entrance channel, over the activation barrier and out the exit channel without becoming 'trapped' anywhere. In the case of CN in reaction (1), a direct reactive encounter is expected to be much more probable than a statistical one. CN resembles fluorine and chlorine atoms in size, mass and electronegativity. The reactions H+CI2 and H+ C1F have been studied extensively.8,940 The energy disposal in these reactions is dominated by their kinematics; these systems have been termed light plus heavy-heavy J (L+HH) mass combinations. 10 These reactions are found to occur on surfaces that j are largely repulsive in nature and the products mainly are the result of end-on attack by hydrogen. A repulsive surface is one in which the energy is released after the new bond is formed.11 Thus, H would approach X2 or XY to nearly the equilibrium bond distance of the product HX, before the diatom starts to separate. i For the case of L+HH on a repulsive surface, the repulsive energy release will j appear mostly in product translation if the attack is exactly collinear and may include : some product rotation if the attack is slightly off-center.11 Relative to other mass ' 1 combinations (e.g. Cl+HI, H+O3, etc.)11 much more energy is found in product J translation for the L+HH mass combinations. A system that is more relevant to ! H+C1CN is H+C1NO. The results from this system, too, could be explained using 1 L+HH arguments, allowing for the fact that NO could take some energy in internal ■ 3 ; I degrees of freedom (though NO took very little energy in rotation and vibration).12 The results from H+C1CN are compared with the results from these systems in the discussion section (see Table TV). Another consideration in the reaction H+C1CN is the effect of reactant orientation. It is obvious that for most collisions between molecules, the orientation in which the collision takes place will affect its outcome. However, experimental methods for knowing and/or controlling the orientations of collisions have been devised only in the past 15 years. Initially, reactants were oriented using electric fields,13 while recently, van der Waals complexes of one of the reactants and a precursor molecule for the other reactant have been used.1445 The precursor then dissociates (via absorption of a photon) and reaction occurs. Control of the orientation and alignment of reactants is then limited only by zero point fluctuations of the complex. In the present system, it appears that the observed product channel is limited to a narrow range of incident angles and impact parameters. It is a minor channel and most angles and impact parameters result in other reaction products, which cannot be seen with the present apparatus. Thus, reaction products are observed which result from well-defined initial conditions: the available energy is quite narrowly defined and the initial orientations which lead to reaction are qualitatively understood. In our experiment, the CN from reaction (1) was monitored via LIF, and rovibrational populations were obtained. Then three rotational lines were slowly scanned at sub-Doppler resolution. By modeling the Doppler profiles, an average value for the translational energy of the CN could be obtained from the best fits. The average energy available to the HC1 could then be found by difference. 4 EXPERIM ENTAL A diagram of the experimental apparatus is shown in Fig. 1. Rather monoenergetic H-atoms were produced by 248 nm H2S photolysis (> 90% of the SH is produced in v=0, the remainder is in v= l) 16 using the weakly focused output from an excimer laser (Lambda Physik EMG 50, ~ 80 mJ). HI was initially chosen as the hydrogen atom source (its absorption cross section at 248 nm is 5.1xl0-19 cm2 vs. 0.26xl0'19 cm2 for H2S),3 but it reacted with C1CN upon mixing. H2S and C1CN were flowed slowly from separate containers and allowed to mix just prior to entering the chamber in order to minimize contamination from dark reactions. The pressure in the chamber was monitored by a capacitance monometer and adjusted to remain constant throughout an experiment. CN was detected by laser induced fluorescence (LIF) using the B2X+« — X2E+ transition. In LIF, a tunable laser is scanned over the wavelength region of interest. The molecules absorb photons and are excited to an upper electronic state; fluorescence from this state is then detected. The Av=0 transition was used to excite the molecules and fluorescence from the same transition was subsequently detected. For CN B2Z+< — X2£ +, AN = ± l ;17 the P- and R-branches can be seen in Figs. 2 and 4. The 'probe' laser operated on BBQ dye and was aligned counter- propagating to the photolysis laser. The region from ~ 389 nm to ~ 383 nm was scanned using a nitrogen pumped dye laser (Molectron UV14/PL13) operating with the oscillator only to reduce ASE. By scanning this region all the rotational lines in v=0 and v=l that were populated could be detected. For the sub-Doppler measurements, an excimer pumped dye laser (Lambda Physik EMG=101, MSC/FL 2001) with an intracavity etalon was used. The etalon narrowed the laser linewidth to 0.05 cm-1 FWHM. The laser was pressure-tuned 5 Transient Digitizer Pump Laser r e s s u re auge PMT Vacuum Pump Laser Photodiode X B e am sp litter M aster O scillator Computer Chamber Photodiode Excimer Laser C1 CN-----F 3 ------- 3 ------- ---------'FLS I ____ r r e s s u r > 8 / -J' Beamsplitter Fig. 1. Schematic drawing of the experimental arrangement. using SF6. It is important that the frequency change be a linear function of pressure and this was monitored. The constant of proportionality was experimentally determined by scanning over v=0, J=17, 18, and 19 and calibrating the change in pressure to their well-known line positions. The laser intensities were monitored on a shot-to-shot basis by fast photodiodes with homemade sample and hold circuits. The fluorescence from Av=0 was monitored through an interference filter (Spectrofilm, 385 ± 20 nm) by a GaAs PMT mounted at right angles to the laser beams. Signals from the PMT were recorded using a digital oscilloscope (Nicolet Explorer III) interfaced to a computer (LSI 11/23) which controlled the experiment and normalized the signals to the probe and photolysis laser energies with each laser firing. A capacitance manometer was interfaced to the computer to monitor the 6 cavity pressure during pressure tuning. Then pressure vs. time could be checked for linearity. Both lasers were operated at 10 Hz; signals from 20-50 laser firings were averaged for each data point. Saturation must also be considered when working with fairly high laser intensities. Saturation occurs when the rate of stimulated emission equals the rate of absorption. Therefore, the molecules spend as much time in the excited state as they do in the ground state. Absorption no longer increases linearly with laser intensity, and the transition is said to be saturated. Even before the excited state density becomes constant, its variation with laser fluence begins to deviate from linearity at sufficiently high laser intensities. To avoid this problem, the LIF signal was checked for saturation and sufficient neutral density filters were placed in the beam to insure that the experiments were performed in the linear regime. In order to see nascent products, it was necessary to work at low gas densities and short delays. In a typical experiment, a total pressure of 80-100 mTorr (1:1 mix) and a delay of 200 ns between photolysis and probe lasers was used. As a check for nascent conditions, a spectrum was also taken using a 100 ns delay, without a significant change in the distribution. Both H2S (Matheson, > 99.5%) and C1CN (Synthatron Corporation, 99%) were used directly without further purification. 7 RESULTS AND ANALYSIS The spectrum of CN obtained from a flowing mixture of both C1CN and H2S is shown in Fig. 2. Because the CN signal from reaction (1) is very small compared to CN signals from similar reactions studied using the same apparatus, it is thought that HC1+CN is a minor channel. At 248 nm, C1CN has a small absorption coefficient, as shown in Fig. 3.18 Nevertheless, the sensitivity of our detection system, together with the small yield of CN from reaction (1) under single-collision conditions, allowed the observation of the small amount of CN from C1CN photodissociation. A spectrum of CN from C1CN photodissociation is shown in Fig. 4. Its vibrational distribution is different than that from reaction (1), since it has observable population in v=l. To obtain the distribution from reaction (1) above, the spectrum in Fig. 4 was taken with the same C1CN density and a helium buffer to bring the chamber to the same total pressure that was used for the spectrum in Fig. 2. All other experimental conditions were the same. The peak heights of the second spectrum were then subtracted from the first. Ideally, peak areas should be used. The rotational distribution was obtained by normalizing the peak heights to each other and then dividing them by the appropriate linestrength factors and filter transmission factors. Ideally, peak areas should be used. However, if the width of the peaks does not change with N, then peak heights may be used. In general, for CN, the widths will change with N because of spin-rotation splitting.17 However, in this case, the N's and the laser resolution were both sufficiently low to justify taking heights. The size of the peaks are also proportional to the laser intensity. Signal a (linestrength) • (ground state population) • (laser intensity) • (filter transmission) (7) 8 Fig. 2. o • s . e o * o a. 8 O U * » u o * — 8 OS G O o c e a o a c Xirr& NSITT ( t i B . mt.fi [ ___________________________________________ Spectrum of CN resulting from a flowing 1:1 mixture of C1CN and H2S. Total pressure is 80 mTorr. Delay between the photolysis and probe lasers is 200 ns. 9 ex o O o in CM o o o CM o in C D O Fig. 3. UV absorption spectrum of C1CN.18 10 Fig. 4 C J iL i GZ (3-xun -qjrs) X^tsus^ui Spectrum of CN from a flowing mixture of C1CN and Helium (1:1). Total pressure is 80 mTorr. Delay between the photolysis and probe J lasers is 200 ns. 1 1 i To correct for this, the intensities of both the probe and photolysis lasers were monitored on a shot-to-shot basis and fluorescence signals were normalized to these intensities. The nascent rotational distribution is shown in Fig. 5.19 A plot of ln{population/(2N+l)} vs. Er0t could be fit to a straight line (see Fig. 5), with slope Bv-'QcTr)-1,17 yielding a rotational temperature, T r, equal to 950±50 K. An absorption line may be broadened by several factors. One of these, the Doppler shift, may be used to establish a translational energy distribution for the state in question, if a laser with a sufficiendy narrow linewidth is available. The maximum Doppler broadening for a given fragment velocity is where Vo is the unshifted frequency of the transition, v is the speed of the fragment and c is the speed of light.20 For a situation in which the fragments have velocity components only along (or counter to) the propagation direction of the laser, this would result in the situation shown in Fig. 6a. However, for an isotropic distribution of fragments all having a single velocity, the situation shown in Fig. 6b would result20 This is due to the fact that the velocity components along the direction of the laser would range from 0 to v. In this experiment, an isotropic distribution was assumed. To attempt to reproduce the experimental profiles, the Doppler lineshape must be convoluted with the laser profile. The laser was assumed to be Gaussian and its linewidth was measured by taking a Doppler profile of a resolved spin- rotation component from thermalized CN, formed by the photolysis of BrCN. The velocity distribution is given by a Boltzmann distribution: Av = vo • v/c (8) P(vz) = exp[-m vz2/2kT] (9) This results in a Gaussian line whose FWHM20 is given by (10) 12 CN(X2E+) r o t a t i o n a l p o p u l a t i o n s o 0 200 400 600 800 N(N+1) Fig. 5. Plot of CN(X2X) rotational populations vs. N(N+1). The straight line corresponds to a temperature of 950 ± 50 K.19 which must then be convoluted with the laser. Since the convolution of two Gaussians is another Gaussian, the resulting FWHM, is where wiand W 2 are the FWHM’ s of the velocity distribution and the laser.2® The parameter w is measured experimentally; wi is calculated using equation (10) and thus W 2, the linewidth of the laser, may be determined. It was found to be 0.05 cm*1. A complication that must be introduced in the Doppler profile for CN is spin-rotation splitting. Each rotational line will be split into two components, Fi and F2. The splitting of those components is given by17 (11) 13 Fig. 6 « A V - -A V______ » Vo AV •AV Vo (a) Doppler profile (before convolution with laser) for all molecules at velocity v, traveling along (or counter to) the k vector of the exciting light. (b) Doppler profile (before convolution with laser) for all molecules at velocity v, distributed isotropically. 14 AF(K) = y(K+l/2) (12) For a CN B2L+— »X2X+ transition, the splittings are thus17 A v r = (Yb - Yx) (K+l/2) + Y b R-branch (13) Avp = (Y B - Yx) (K+l/2) - Y B P-branch (14) It was found by others20 that the actual spin-rotation splitting differed from the calculated one for CN lines they measured. Therefore the spin-rotation splitting was experimentally determined for each of the lines in question. To make this measurement, Doppler profiles were taken for N = 17 and 25 using thermalized CN from the photolysis of BrCN, which gives a strong signal. The splitting for N=2 was too small to be seen, so it was calculated using equation (15) from published values of Yx and YB-21 The values used for the splittings are shown in the first column of Table I. In order to determine the translational distributions associated with the measured Doppler profiles of the rotational lines N = 2, 17 and 25, a translational distribution was guessed; a Doppler profile was then simulated and this was compared to experiment. As a starting point, single velocity profiles (as in Fig. 6b) were convoluted with the laser lineshape. This did not result in a good fit for any of the rotational lines measured. Next a Maxwell-Boltzmann speed distribution was used. This was truncated in order to conserve energy. The maximum energy available for translation on a given rotational line, N, is P(u) = u2 exp [ -mu2/2kT ] (15) E = E t - BN(N+1) (16) 15 TABLE I. Parameters used in fitting CN(X2E) Doppler profiles. See text for details. N Spin-Rotation Splitting (cm*1 ) Maximum Possible CN Kinetic Energy (cm*1 )0 Maximum Possible CN Velocity (m s* 1)0 Average CN Kinetic Energy (cm*1 ) Function 4a Function 4b n = 1 n =2 n = 3 (nj=l, mj=6, n2=m2=10) 2 0.04a 6030 2360 2100±400 1950±400 1500+400 1500+400 17 0.20b 5680 2290 2250±200 1980+200 1800+200 1800+200 25 0.26b 5280 2200 235G±200 2180±200 2150+200 2200+200 a) From reference 17 and 21. b) Experimentally determined. c) From conservation of energy and linear momentum. In our case, E t is 10,350 cm-1 and the maximum possible CN kinetic energies and their corresponding velocities are shown in Table I. The Maxwell-Boltzmann distribution resulted in a good fit for the Doppler profiles. Note that since the distribution is truncated, T is not a temperature, but a fitting parameter. The average translational energies for the T's which produced the best fits were calculated and are shown in Table I. To test for the uniqueness of the translation energy distributions which could fit the observed Doppler profiles, the following convenient forms were used for speed distributions.22’ 23 P(u) = un exp [ -mu2/2kT ] (17) (u/u*)nl exp { (nl/m l) [ 1 - (u/u*)ml ] } u < u* P(u) = (18) (u/u*)n2 exp { (n2/m2) [ 1 - (u/u*)m2 ] } u > u* where u* is the most probable speed (i.e., the value where the distribution peaks); nl and m l are parameters that fit the rising part of the distribution, while n2 and m2 correspond to the falling part. When using (17), conservation of energy was assured by truncating the distributions at the proper limit (see Table I). Again, T is a fititng parameter, not a temperature. Acceptable fits were obtained using n = 1,2, or 3 and varying T. Simulations with n > 3 did not fit the data because they were too narrow in the wings and too broad near the tops of the peaks. Function (18) falls almost to zero before the energy conservation limit when using large n2 and m2. This allows energy to be conserved without sharply truncating the speed distribution, as was done for distribution (17). nl and m2 were chosen so that the shape of the peaks matched the data, and u* was then fine-tuned until the peaks and valleys matched the data. The m's and n's are not unique, but m2 must be large and nl must be small in order to fit the experimental profiles. In all of the simulations, the relative heights of the 'valleys’ for N=17 and 25 were very sensitive to the average kinetic 17 energy, whereas the widths of the peaks were not as sensitive. Thus, N=2 could be fit to a broader range of average translational energies than the other profiles. Finally, the distribution of energy among the product states was tabulated (see Table II). All of the observed CN was in v=0. The CN rotational population could be fit to a temperature of 950 ± 50 K which corresponds to an average rotational energy of 660 ± 30 cm*1. The fraction of the available energy in CN rotation is therefore 0.06. The average values of the CN translational energy vary some with the two functional forms shown in equations (17) and (18) and the different N's, but all can be described by <Ej(CN)> = 2000 ± 500 cm*1. Three fittings for each N are shown in Fig. 7 and the average translational energies with estimated uncertainties are given in Table I. The total c.m. kinetic energy was obtained from momentum conservation: E c.rn. (HC1 + CN) = < E t > c n (M c n / M Hc i) = 3400 ± 800 cm*1 (19) The internal energy of HC1 can be found by energy conservation: Eint(HCl) = E t - Eint(CN) - Ec.m.(HCl+CN) - 3/2 RT (20) The translational energy distributions corresponding to the velocity distributions used to fit CN(N=17) are shown in Fig. 8a. The broad limits shown by the shaded regions indicate a qualitative estimate of the latitude of the fits (i.e. the velocity distributions used to generate the solid curves shown in Fig. 7 lie comfortably within the shaded regions in Fig. 8a). The HC1 internal energy distribution is obtained using equation (20), and is shown in Fig. 8b. Arrows indicate the energies which correspond to v = 1,2, and 3 of HC1(J=0). A summary of the average internal and kinetic energies is given in Table n. 18 Table II. Average energies and fractions of available energy associated with different degrees of freedom. Total Available Energy = 10,350 cm-1 Average Energy (cm'1 ) Fraction of Available Energy CN (vib) 0 0 CN (rot) 660 ± 30 0.06 ± 0.01 c.m. kinetic energy 3430 ± 830 0.33 ± 0.08 HC1 (internal) 6260 ± 830 0.61 ± 0.08 19 o t W w X z o 1 0 ■« CM > O 1 % > x < m > CM . > O ui > X I ] « c O J CM > ■ ■ > C _ Fig. 7. Doppler profiles of three CN B2I*-X 2X R-branch transitions. The spin- I rotation splitting increases almost linearly with N. Each experimental profile is fit using three velocity distributions (see text). The corresponding speed distributions are indicated at the top of the figure. I The solid lines are fits and the dots are the data. The average CN kinetic energy corresponding to each fit is shown in the upper right-hand comers.19 20 E int < CN> = 660 cm .o u a 9690 S 2000 4000 c.m. kinetic energy (c m 1 ) -----► 6000 8000 10000 • m m c s (B) « p O U rn C 3 e o •mm + m i v=2 v=3 rt 53 a o a. 2000 4000 6000 HC1 internal energy (cm ;1 ) 8000 10000 u a Fig. 8. (a) Estimate of population vs. center-of-mass kinetic energy. The total available energy is 10,350 cm-1, but the distribution is truncated at 9690 cm’1 to allow for the CN internal energy.19 (b) The 'reflection' of (a), showing HC1 population vs. internal energy.19 21 D IS C U S SIO N The energy disposal in the reaction products does not fit the statistical distribution from the prior (see Table HI). Therefore, it can be concluded that the reaction is much closer to a direct process than to a statistical one. It is probable that kinematics heavily influence the reaction; the energy distribution in the products closely resembles that of the L+HH systems discussed earlier (see Table IV). These reactions are known to occur via direct collinear encounters.^ '12’ 29 Since the HC1+CN products are so similar to these systems, it is assumed that the observed products accrue only from collisions that are mainly collinear. To further test these inferences, the pathways available to reaction ought to be examined further. There are three possible atoms at which hydrogen can attack in C1CN: chlorine, carbon and nitrogen. C alculations To shed further light on the reaction mechanism, Professor Gerald Segal agreed to perform ab initio calculations on the system. He calculated the energies and geometries of the stationary points for the HC1CN surface. All of the calculations were performed without zero point energies, so that values given do not include zero point energies and are in the Bom-Oppenheimer approximation. It was desired that the computational errors in the energies of the various dissociation channels be roughly equal so that the computed surface did not have a built in bias. Table V shows the calculated and experimental 'enthalphies' (based on De's).17’24* 26 The largest discrepancy is for the product channel HC1 + CN, which is calculated to be 11.35 kcal mol-1 less stable than the reactants, while the experimental value indicates that it should be 8.1 kcal mol' 1 more stable, making a 22 Table III. Results of the prior calculation compared to experiment. Total Available Energy = 10,350 cm*1 Fraction of Available Energy Experiment Prior Calculation CN (vib) 0 0.13 CN (rot) 0.06 ± 0.01 0.21 Center-of-mass kinetic energy 0.33 ± 0.08 0.36 HC1 (internal) 0.61 ± 0.08 0.10 in vibration 0.30 0.20 in rotation Table IV. Energy disposal in several L+HH reactions. The values from reference 8 are the result of averaging over 17 different experiments. L+HH -+ LH+H <fy> <fR> <fx> <fv,r(NO)> Reference H+C12 -» HC1+C1 0.30 0.08 0.53 8 H+C1F -> HC1+F 0.47 0.11 0.42 9 H+C1NO HC1+NO 0.50 0.07 0.30 0.13 12 23 Table V. Calculated vs. experimental energies for the dissociation of the molecules associated with the HC1CN potential surface. Also, calculated vs. experimental energies for the reaction channels. De (in kcal mol-1) Calculated Experimental17.24-26 HC1 -» H+Cl 89.3 106.4 HCN -+ H+CN 136.5 126.4 C1CN -> Cl+CN 100.7 98.3 HNC H+CN 118.8 111.5 AH (without zero point energy) (in kcal mol"1) Calculated Experimental4* 5 H+C1CN -+ HC1+CN 11.35 -8.1 H+C1CN -+ HNC+C1 -18.1 -13.2 H+C1CN -> HNC+C1 -35.8 -28.1 difference of 19.45 kcal mol-1. The absolute values of the computed energies should therefore be treated cautiously.19 The results of the calculation are shown in Fig. 9a. The optimized geometries, corresponding to the labelled parts of Fig. 9a are shown in Fig. 9b. 24 Chlorine attack Attack on the chlorine results in transition state G. Even though it looks as if G is not energetically accessible from Fig. 9a, recall that the calculated HC1+CN asymptote is 19.45 kcal mol-1 higher than the experimental value. If the HC1+CN channel is lowered by this amount, transition state G must drop considerably, making it accessible for reaction. Also, tunneling should further enhance the G pathway. Carbon attack Attack on the carbon in a direction approximately perpendicular to the C1CN axis, goes through transition state A, and on to B, which is the most stable isomer of HC1CN. H2CN27 and H2CO+ 28 have well-studied, analogous forms, which are also their most stable isomers. Structure B can then dissociate to either HCN+C1 or HC1+CN. Dissociation to HCN+C1, the most thermodynamically stable product channel, occurs via transition state D which lies 6.1 kcal mol' 1 above intermediate B. The transition state, C, leading to the observed products, HC1+CN, is much less exothermic. It lies 38.6 kcal mol' 1 above structure B, though considering the calculated error in the HC1+CN energy, C is more likely to be around 19.2 kcal mol' 1 above B. C may be formed through a highly excited HCC1 scissors motion, which must result in a hydrogen insertion into the C-Cl bond. Nitrogen attack Direct collinear attack at nitrogen is repulsive up to 50 kcal mol'1. This is unaccessible at the energies available in the experiment. However, approximately perpendicular attack at nitrogen results in structure F via transiton state E. Structure 25 F has both a cis and a trans form, both of which are stable isomers of HC1CN and have analogs in H2CN and H2CO+.27* 28 The observed products, HC1+CN, could be a result of either attack on chlorine or attack on carbon followed by the formation of structure B and subsequently, HC1+CN via transition state C. The observed products could not result from a long-lived intermediate because the results are far from being statistical (see Table IH). Chlorine attack results in no intermediate but carbon attack is also a possibility, if structure B is not long-lived. However, the fact that the HC1 internal energy distribuiton is very similar to analogous systems (H+CI2, H+C1F, H+C1NO) where end-on attacks are known to be strongly favored8' 12* 29 indicates that attack on chlorine is the major pathway for formation of the observed product. Further, it is highly unlikely that the observed population inversion in HC1 would result from intermediate B. On examining the results of the calculations it can be concluded that HC1+CN is probably a minor channel. Just by assuming all angles of approach to be equally likely, only a small percentage will result in end-on attack at chlorine. Further, reaction to form HC1+CN via end-on attack at chlorine is the least thermodynamically favored channel. This conclusion is further bolstered by the fact that the experimental CN signal was very small compared to the CN signal seen from other reactions using the same apparatus. From looking at Fig. 9a, it can be concluded that HCN+C1 is probably the major reaction pathway. In this experiment, the orientation,as well as the energies of the reactants was fortuitously well-known. In retrospect, the orientations resulting in the observed products most likely were limited to those close to end-on approaches to the chlorine. This advantage came from the fact that a minor channel was observed and more favorable pathways 'drained away1 the other orientations of attack. 26 , kcal mol 40 30 31.3 20 E(cis) 17.8 6.7 ■ HCI+CN 11.35 (-8 .1) -18.9 (cis) -18.2 (trans) -16.4 -20 HNC+CI -18.1 (-13.2) -22.5 -30 HCN+CI -35.8 -40 (-28.1) Fig. 9. (a) Reaction coordinate diagram for HC1CN. Energies are listed relative to H + C1CN; experimental values are in parentheses. Zero-point energies are not included.19 I 27 ______ (A) .« * ' ^ 1 0 5 .2 ° 2.49 / 156.8 ( B ) h < 124.9° 2.40 123.6° (C) 1.33 1.63 1.14 Cl H N (E) Cl 1 6 6 3 1 .6 5 ^ C 127.2° A _ L N / l - 4 3 1.16 " 144.8° HC 1.07 2 .5 6 , 113.8° (F) ( cis) 136.7 152.5° c t i ^ i (trans) 139 ?0 N 1.02 126.9° 90 1 - l f H 152.9° 172.0° Fig. 9. (b) Stationary points on the HC1CN potential energy surface. The letters denoting the different structures correspond to the energies listed in (a). Distances are in A's; all structures are planar.19 28 SUM M ARY Energy disposal in the products of H+C1CN — > HC1+CN was determined. The results are displayed in Table II. The reaction products are thought to accrue via approximately end-on attack by hydrogen on the chlorine, which leads directly to products. The HC1 is vibrationally inverted. The reaction channel observed is a minor one; other products are HNC+C1 and HCN+C1, with the latter likely to be the major pathway. ] I 29 R EFER EN C ES 1. J.F. Cordova, C.T. Rettner, and J.L. Kinsey, J. Chem. Phys. 75, 2742 (1981). 2. G.W. Flynn and R.E. Weston, Jr., Ann. Rev. Phys. Chem. 37, 551 (1986). 3. C.A. Wight and S.R. Leone, J. Chem. Phys. 79, 4823 (1983). 4. S.W. Benson, "Thermochemical Kinetics," Wiley-Interscience, New York, 1976, 2nd ed. 5. (a) P.K. Pearson, H.F. Schaefer, and U. Wahlgren, J. Chem. Phys. 62, 350 (1975); (b) P.K. Pearson, G.L. Blackman, H.F. Schaefer, B. Roos, and U. Wahlgren, Astrophys. J. 184, L19 (1973); (c) L.T. Redmon, G.D. Pruvis III, and R J. Bartlett, J. Chem. Phys. 72, 986 (1980). 6 . D.A. McQuarrie, "Statistical Mechanics," Harper & Row, New York, 1976. 7. National Bureau of Standards, "JANAF Thermochemical Tables," 1971, 2nd ed. 8 . W. Jakubetz, Chem. Phys. 88, 271 (1984) and references cited therein. 9. D. Brandt and J.C. Polanyi, Chem. Phys. 35, 23 (1978). 10. M.R. Levy, Prog, in React. Kin. 10, 1 (1979). 11. J.C. Polanyi, Science 236, 680 (1987). 12. M.A. Nazar, J.C. Polanyi, and W.R. Skrlac, Chem. Phys. Lett. 29, 473 (1974). 13. P.R. Brooks, Science 193, 11 (1976). 14. S. Buelow, G. Radhakrishnan, J. Catanzarite, and C. Wittig, J. Chem. Phys. 83, 444 (1985). 15. G. Radhakrishnan, S. Buelow, and C. Wittig, J. Chem. Phys. 84, 727 (1986). 16. G.N.A. VanVeen, K.A. Mohamed, T. Bailer, and A.E. DeVries, Chem. Phys. 74, 261 (1983). 17. G. Herzberg, "Spectra of Diatomic Molecules," D. Van Nostrand, Princeton (1950). 30 18. G.W. King and A.W. Richardson, J. Mol. Spectrosc. 21, 339 (1966). 19. J. de Juan, S. Callister, H. Reisler, G.A. Segal, and C. Wittig, J. Chem. Phys., submitted (1988). 20. C.X.W. Qian, private communication. 21. K.P. Huber and G. Herzberg, "Molecular Spectra and Molecular Structure," Van Nostrand Reinhold Co., New York (1979). 22. D.S. King and J.C. Stephenson, Chem. Phys. Lett. 114, 461 (1985). 23. J.D. McDonald, P.R. LeBreton, Y.T. Lee, and D.R. Herschbach, J. Chem. Phys. 567, 769 (1972). 24. G. Herzberg, "Electronic Spectra of Polyatomic Molecules," D. Van Nostrand, Princeton (1966). 25. R.A. Creswell and A.G. Robiette, Mol. Phys. 36, 869 (1978). 26. R.B. Heiart and G.B. Carpenter, Acata Crystallographica 9, 889 (19xx). 27. R.A. Bair and T.H. Dunning, Jr., J. Chem. Phys. 82, 2280 (1985). 28. M. Vaz Pires, C. Galloy, and J.C. Lorquet, J. Chem. Phys. 69 3242 (1978). 29. B.A. Hodgson and J.C. Polanyi, J. Chem. Phys. 55, 4745 (1971). 31 A PPE N D IX A prior calculation distributes the available energy to the products statistically while conserving energy; for two diatomic products, the distribution is1 P (vi Jl V2 J2 /E ) = pvl pv2 P jl Pj2 PT a (2Jl+l) (2J2+1) E t 1/2 (1) For a specific v,J state of one of the diatomics , this becomes1 v2=v2max p (v iJi/E )a2 Ji + l X (E -E vi-E v2-E j i )3/2 (2 ) v2=0 where the rotational levels of the other diatomic are treated as a continuum and are integrated over. To calculate the rotational and vibrational energy for one of the diatomic products, AB, v imax, the maximum vibrational level possible is first determined. Then for each vibrational level, the maximum possible rotational level, Jimax, is determined. Then, for each v,J the maximum vibrational level of the other diatomic is determined. Then equation (2) is used to find the statistical weight for each v,J state. All of the v,J states are normalized and the average vibrational and rotational energies may then be found by summing. The prior distribution is compared to the experimental distribution in Table III. 32

Asset Metadata

Creator Callister, Susan Hansen (author)

Core Title The reaction H+ClCN to CN+HCl

Contributor Digitized by ProQuest (provenance)

Degree Master of Science

Degree Program Chemistry

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag chemistry, organic,OAI-PMH Harvest

Language English

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c17-797601

Unique identifier UC11348278

Identifier EP41675.pdf (filename),usctheses-c17-797601 (legacy record id)

Legacy Identifier EP41675.pdf

Dmrecord 797601

Document Type Thesis

Rights Callister, Susan Hansen

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA