3563
J. Phys. Chem. 1986, 90, 3563-3568
Dlssociation Rates of Energy-Selected Dichloro- and Dibromobenzene Ions Susan Olesik, Tomas Baer,* and J. C. Morrow Department of Chemistry, University of North Carolina, Chapel Hill, North Carolina 2751 4 (Received: January 16, 1986; In Final Form: March 3, 1986)
The dissociation rates of internal-energy-selected 0-, m-, and p-dichloro- and dibromobenzene ions have been measured by photoelectron photoion coincidence spectroscopy. The rate measurements indicate that these isomers rearrange to a common structure, the para isomer, prior to fragmentation to the C6H4X++ X products. The activation energies for the loss of a halogen atom, determined by fitting RRKM calculated rates to the experimental values, are 3.35 f 0.1 and 2.8 f 0.1 eV for the chloro and bromo compounds, respectively. The resulting heats of formation of the C6H4C1+and C6H4Br+fragments are 260 f 3 and 270 f 3 kcal/mol. From these values, the C-H bond energy in the chloro- and bromobenzene ions was calculated to be 3.88 f 0.10 eV. The dissociation limits to X2 C6H4+are 4.4 and 3.9 eV for the chloro and bromo compounds, respectively. However, even up to the limit of the hydrogen light source of 14.2 eV, no C6H4+signal was observed. This is consistent with the statistical theory that predicts the rate of X loss is more than lo4 times faster than loss of Xz.
+
Introduction Many of the substituted benzene ions are extemely stable with respect to dissociation. As a result, the dissociation rates of the these ions, when prepared above the dissociation limit, can be very slow. The dissociation rates of a number of such metastable substituted benzene ions have been measured.'-'* In particular, energy-selected chloro-, bromo-, and iodobenzene have been carefully investigated by a number of workers with a variety of technique^.^-^ The methods of energy selection have included photoelectron photoion coincidence (PEPICO), differentiation of the metastable signal in photoionization scans: and laser MPL9 The results of these investigations of the monohalobenzene ions have led to the conclusions that (a) the loss of the halogen atom proceeds without a reverse activation energy, (b) the reaction rate k ( E ) as a function of E can be well modeled with the statistical theory, (c) a t low energy the structure of the benzene ring is maintained in the C6H5'+ product ion, and (d) the heat of formation, AHfoo,of the C6H5'+is 271 f 1 kcal/mol.s The determination of this heat of formation has not been a trivial problem. Every known reaction that produces this product ion involves a large activation energy, so that the dissociation rate at the thermochemical onset is too slow to be measurable. Hence, the value is one based on an extrapolation of the measured decay rate, k ( E ) , by the RRKMJQET statistical theory.I5-l6 The dissociation of the dihalobenzene ions has recently been investigated by photoi~nization'~ and laser MPII8 techniques.
(1) Andlauer, B.; Ottinger, Ch. J . Chem. Phys. 1971, 55, 1471. (2) Baer, T.; Tsai, B. P.; Smith, D.; Murray, P. T. J . Chem. Phys. 1976, 64, 2460. (3) Rosenstock, H. M.; Stockbauer, R. L.; Parr, A. C. J . Chem. Phys. 1979, 71, 3708. (4) Rosenstock, H. M.; Stockbauer, R. L.; Parr, A. C. J . Chem. Phys. 1%0, 73, 773. (5) Dannacher, J.; Rosenstock, H. M.; Buff, R.; Parr, A. C.; Stockbauer, R. L.; Bombach, R.; Stadelmann, J. P. Chem. Phys. 1983, 75, 23. (6) Pratt, S . T.; Chupka, W. A. Chem. Phys. 1981,62, 153. (7) Baer, T.; Kury, R. Chem. Phys. Len. 1982, 92, 659. (8) Eland. J. H. D.: Schulte. H. J . Chem. Phvs. 1975. 62. 3835. (9) Durant, J. L.; Rider, D.'M.; Anderson, S: L.; Zare, R.N. J . Chem. Phys. 1984.80, 1871. (10) Rosenstock, H. M.; Stockbauer, R.; Parr, A. C. J . Chim. Phys. 1980, .77. , 145. . .- . (1 1) Brand, W. A,; Baer, T. Int. J . Mass Spectrom. Ion Phys. 1983, 49, 103. (12) Baer, T.; Carney, T. E. J . Chem. Phys. 1982, 76, 1304. (13) Panczel, M.; Baer, T. Int. J. Mass Spectrom. Ion Proc. 1984,58,43. (14) Fraser-Monteiro, M. L.; Fraser-Monteiro, L.; deWit, J.; Baer, T. J . Phys. Chem. 1984,88, 3622. (15) Marcus, R. A,; Rice, 0. K. J . Phys. Colloid chem. 1951, 55, 894. (16) Rosenstock, H. M.; Wallenstein, M. B.; Wahrhaftig, A. L.; Eyring, H. Proc. Natl. Acad. Sci. U.S.A. 1952, 38, 667.
These initial studies have raised a number of interesting questions, among them the question of isomerization of the various isomeric forms. Hydrogen atom shifts are easily induced photochemically in neutral substituted benzenesI9 and thiophenes.20 Evidence for complete hydrogen scrambling has also been obtained for ions with sufficient energy to dissociate.21 In particular, the hydrogen atoms in the benzene molecular ion scramble prior to fragmentation.22 One question of interest is, will the much heavier halogen atoms also scramble in excited dihalobenzene ions, thereby converting the higher energy isomers into the most stable isomeric configuration prior to dissociation? Comparison of metastable ion abundances of the para and meta dichlorobenzene isomers led Brown to suggest that isomerization might take place.23 Such a hypothesis can be tested by measuring the dissociation rate for the three isomers. The observation of identical rates is strong circumstantial evidence that such isomerization reactions have taken place. The loss of a halogen atom from a dihalobenzene ion provides an interesting connection with the reaction of the monohalobenzene ions. The H atom loss in reaction 1 C&SX+ C&X2+
-
+
C6H&+ C&X+
+H +x
(1)
(2) leads to the same product as X loss in reaction 2. Because the X loss from the monohalobenzene is of lower energy, the C-H bond strength in these compounds is not known. However, determination of the onset for reaction 2 can lead to a determination of the C6H4X+heat of formation, from which the AH for reaction 1 can be determined. One of the four reactions occurring at low energies in the dissociation of benzene ions is the loss of H2, leading to the C6H4+ product ion. Its structure is not well established, although it is suspected to be the benzyne ion.22 Such conclusions can be established by a comparison of the measured heat of formation, obtained from the activation energy, with the known ionic heats of formation of some of the h e a r C6H4+ions.1° The latter are either estimated by the use of group equivalence methods or (17) Moini, M.; Leroi, G. E. Proceedings of the 33rd Annual Conference on Mass Spectrometry and Allied Topics May 26-31, San Diego, CA; American Society for Mass Spectrometry: 1985; p 35. (18) Szaflarski, D. M.; Simon, J. D.; El-Sayed, M. A. J . Phys. Chem., in press. (19) Wilzbach, K. E.; Kaplan, L. J . Am. Chem. SOC.1964, 86, 2307. (20) Wynberg, H.; van Driel, H. J . Am. Chem. SOC.1967, 87, 3998. (21) Beynon, J. H.; Caprioli, R. M.; Perry, W. 0.;Baitinger, W. E. J. Am. Chem. SOC.1972, 94, 6828. Yeo, A. N. H.; Cooks, R. G.; Williams, D. H. J . Chem. SOC.B: 1969, 149. (22) Rosenstock, H. M.; Dannacher, J.; Liebman, J. F. Radiat. Phys. Chem. 1982,40, 7. (23) Brown, P. Org. Mass Specrrom. 1970, 3, 639.
0022-3654/86/2090-3563$01.50/00 1986 American Chemical Society
3564
The Journal of Physical Chemistry, Vol. 90, No. 16, 1986
obtained from measured ionization energies of stable C6H4 molecules. It is thus of interest to determine the energies of the C6H4+ions from as many different sources as possible in order to learn more about the various structures and energies of these isomeric ions. If the dihalobenzene ions were to dissociate via X2 or 2X loss, this might offer another route for determining the C6H4+ion structure and energy. The recent MPI'* and photoionization" studies have indicated that Xz loss might take place.
Experimental Approach The method of state selection by photoelectron photoion coincidence (PEPICO) is well e s t a b l i ~ h e d . *In ~ ~this ~ ~ study we have used two different PEPICO apparatuses. Briefly, light from a low-pressure hydrogen-discharge lamp is dispersed by a vacuum UV monochromator. The two experiments use a Jarrell Ash 1-m normal incidence and a McPherson 0.5-m Seya-Namioka monochromator. In both experiments ions and electrons are extracted in opposite directions by an applied electric field of about 10 V/cm. Initially, zero-energy electrons are selected by a steradiancy analyzer26 in which electrons with velocity vectors significantly different from the direction of the applied electric field are stopped by apertures. The combined photon and electron energy resolution of the experiments with the 1-m and 0.5-m monochromators are 35 and 80 meV, respectively. The major difference in the two experiments is in the method of measuring the dissociation rate. In the 1-m monochromator, the ions are accelerated in a single, 4.5-cm-long, acceleration region after which they pass through a single 9-cm-long drift region. Slowly dissociating ions fragment during the course of acceleration, resulting in asymmetric fragment ion TOF distributions. The rate is then extracted by fitting the experimental TOF distribution with a calculated one in which the dissociation rate is an adjustable parameter. The method used with the 0.5-m monochromator employs two acceleration and drift distances. Ions that dissociate in the first acceleration region ( F J can be differentiated from those that dissociate in the first drift region (F2)by their different times of flight. The ratio of the signals from these two regions can be expressed as
Olesik et al. TABLE I: Heats of Formation (kcal/mol) and Ionization Energies (eV) ion molecule mokcuk/ion AH,', AHr0298 IP AHrOo AHf029ed 1,2-C6H4C12 10.1" 7.2' 9.06' 219.1' 216.1 113-C6H4C12 9.0" 6.1b 9.1OC 218.9c 215.9 1,4-C6H4C12 8.3" 5.4' 8.92c 214.1' 211.1 8.98( 245 f 2' 239 1,2-C6H4Br2 38" 32 f 2' 9.01' 246 f 2c 240 1,3-C6H4Br2 38" 32 f 2' lr4-C6H4Br2 36" 30 f 2e 8.82' 239 f 2c 233 C~HSCI 15.74' 12.26 9.07g 224.9 221.4 C6H5Br 30.2h 24.9h 8.979 237.1 231.7 260 f 3 C6H4CI 263 f 3' C6H4Br 275 f 3' 270 f 3 23.9' 9.25' 231.3' C6H6 C6HS 271, C6H4 (benzyne) 3 15.7' C6H4 (linear) 339.6' Br2(g) 10.92' 7.39' CI 28.6Sk 29.08' Br 28.18' 26.74' H 5 1 .62' 52.09' "Converted from 298 K value by using frequencies from ref 27. *Pedley and Rylance.28 'This work. Unless otherwise stated, the uncertainties in IP's and AHfo's are f0.02 eV and f 0 . 5 kcal/mol, respectively. d Based on Rosenstock convention29in which AH,O,,,(e-) = 0. To convert to J A N A F convention, add 1.5 kcal/mol. eBased on Bensonso additivity scheme (see Table 11). fRosenstock et aL3 gBaer et aL2 hThis value was obtained by using the standard states of Br2 which are the liquid at 298 K and the crystal at 0 K. Rosenstock et aL4 assume the nonstandard gaseous state at both temperatures. 'Rosenstock et a1.22 JRosenstock et aLs 'Wagman et aL3I
TABLE II: Comparison of Experimental" and Estimated* PHfozB (kcal/moi) of Halobenzene Molecules C6H6-nFn 19.8 (exptl) -26.3 (est) -27.7 (exptl) -72.4 (est) -73.3 (exptl)
n 0 1
2
C6H6-nC1n
19.8 (exptl) 12.7 (est) 12.3 (exptl) 5.6 (est) 5.4 (exptl)
C6H6-nBrn
19.8 (exptl) 25.0 (est) 24.9 (exptl) 30.2 (est)
OValues taken from Pedley and Rylance.28 bValues estimated according to
The times, t, and tz, refer to the ion flight time to the end of the first acceleration region and to the end of the first drift region, respectively. The dissociation rate constant can be calculated from the measured ratio, F,/F,, and the known ion flight times. All the dichlorobenzene data were taken on the I-m PEPICO apparatus, while the dibromobenzene data were collected on both instruments. The samples were obtained from Aldrich Chemical Co. The dichlorobenzenes were used without purification. However, the m-dibromobenzene contained significant amounts (not more than 1%) of monobromobenzene, which interfered with the studies, and had to be purified by preparative G C separation. The p-dibromobenzene contained an unidentified impurity that did not interfere with the data analysis. A final problem was the low vapor pressure of these compounds, which made them difficult to remove from the sample inlet system. For this reason, even small amounts of samples from previous studies produced significant signals unless the sample inlet system was thoroughly baked out.
I2.O
- I ".O 0.0
-2
t I
t
3.r
0-0
X
a I I
0.YL
I
9.10 I
9.06
0.390
0.438
1.0 -
0.360
t t t Figure 1. Thermochemistry of the dichlorobenzene molecules and ions at 0 K. All values are based on experimental determinations.
Results A . C6H4C12and c6H4Br2Neutral and Ion Thermochemistry. The heats of formation of the dichlorobenzene molecules are well-known. These, along with o t h e r relevant thermochemical
information, are listed in Table I. The heats of formation were converted to the 0 K values by using known vibrational frequencies of the dichlorobenzene molecules.27 The adiabatic ionization
(24) h e r , T. In Gas Phase Ion Chemistry; Bowers, M. T., Ed.; Vol. I , Ch. 5, Academic: New York, 1979. (25) Baer, T. Adv. Chem. Phys. 1986, 64, 11 1. (26) Baer, T.; Peatman, W. B.; Schlag, E. W. Chem. Phys. Lett. 1969, 4 , 243.
(27) Green, J. H. S.Spectrochim. Acta 1970, 26A. 1503, 1523, 1913. Stojiljkovic, A,; Whiffen, D. H. Spectrochim. Acra 1958, 12, 47. (28) Pedley, J. B.; Rylance, J. "Sussex-N.P.L. Computer Analysed Thermochemical Data: Organic and Organo Metallic Compounds", University of Sussex, 1977.
Dissociation Rates of Energy-Selected Ions
The Journal of Physical Chemistry, Vol. 90, No. 16, 1986 3565
Br
hv = 13.75 cV
hu = 13.55 eV
hu
3
13.93 eV
u)
D
Figure 3. Product ion T O F distributions at various C6H4CI2+internal energies. The solid lines are calculated TOF distributions assuming a single exponential decay with the indicated rates. TOTAL ION ENERGY ( e V ) 14.0 14.5
I Figure 2. Thermochemistry of the dibromobenzene molecules and ions at 0 K. The heats of formation of the neutral molecules are calculated by using the additivity scheme of Benson.
CsH4CI$
+ CsH4CIt
10':
-
I
energies of the dichlorobenzenes were measured a number of years ago by Watanabe et al.32 In addition, several PES studies have determined vertical IP's of the dichloro- as well as the dibromobenzene isomers.33 In order to have a consistent and complete set of adiabatic IP's, we measured these again by collecting photoionization efficiency (PIE) scans in the vicinity of the ionization onsets. These values, listed in Table I, are about 0.1 eV but agree with the dichlorolower than the vertical PES IPS, benzene values obtained by Watanabe et a1.32within 0.02 eV. The energies of the neutral and ionic dichlorobenzene isomers are shown in Figure 1. No experimental values for the AHfo of the dibromobenzene molecules are known. For this reason, we have estimated the AHfo298of dibromobenzene using the additivity scheme of BenIn order to check the validity of the method for this of some molecule, the calculations were carried out for the AHfo298 known dihalobenzenes. These are shown in Table 11. The starting point in this procedure is the benzene molecule in which H atoms are successively replaced by the halogen atoms. This method gives values within 1 kcal/mol of the experimental values for the p difluoro- and p-dichlorobenzenes. In addition, the estimate for the monobromobenzene is quite good. For this reason, the estimated AHf0298of p-dibromobenzene of 30.2 kcal/mol should be accurate to at least i2 kcal/mol. Based on the relative heats of formation of the various isomers of difluoro- and dichlorobenzenes, the o and m-dibromobenzenes are probably 2 kcal/mol less stable than the p-dibromobenzene. The energies of the ions are shown in Figure 2. B. Ca4C12+Dissociation Rates. The dissociation rates of the dichlorobenzene ions were determined by measuring the asymmetric time-of-flight (TOF) distributions of the daughter ions as a function of the parent ion internal energy. An example of such a set for the case of o-dichlorobenzene is shown in Figure 3. The fit of the calculated solid line to the experimental points was achieved by assuming a single exponential decay with the indicated dissociation rate. Data collection is limited at high energy by the (29) Rosenstock, H. M.; Draxl, K.;Steiner, B. W.; Herron, J. T. J . Phys. Chem. Ref. Data, Suppl. 1977.6, 1. Rosenstock, H. M.In Kinetics of Ion Molecule Reactions; Ausloos, P., Ed.; Plenum: New York, 1979. (30) Benson, S. W. Thermochemical Kinetics, 2nd ed.;Wiley: New York, 1976. (31) Wagman, D. D.; Evans, W. H.; Parker, V. B.; Schumm, R. H.; Halow, I.; Bailey, S. M.;Churney, K. L.; Nuttall, R. L. J . Phys. Chem. Ref Data, Suppl. 1982, 11. (32) Watanabe, K.; Nakayama, T.; Mottl, J. J. Quant. Spectrosc. Radial. Transfer 1962, 2, 369. (33) Kimura, K.;Katsumata, S.; Achiba, Y.; Yamazaki, Y . ; Iwata, S .
Handbook of He1 Photoelectron Spectra of Fundamental Organic Molecules; Japan Scientific Society: Tokyo, 1980.
I
-: : w
-
ta
-
2
0
-
E
-
t u0 ;lo5
15.0
" J " " 1 " " 1 l
f f E,= 3.35 eV
AS*=l . 8 e u
f,, 0
;, 4.4
t CI
I
,
4.6 4.8
0
Poro Meta Ortho I ,
5.0 5.2
5.4
I
5.6
-
5.8
Figure 4. Derived dissociation rates as a function of the total ion internal
energy [ h v + AHfoo(C6H4C12) + &], where E,,, is the thermal energy of 0.144 eV. The solid line is a statistical theory calculation of the rates assuming an activation energy of 3.35 eV and an entropy of activation of 1.8 cal/(mol deg) at 1000 K.
disappearance of the asymmetry in the daughter ion TOF distribution. At low energies, the limit is not the dissociation energy but rather the fact that fewer and fewer parent ions dissociate in the time scale of the experiment (1-20 ps), so that the signal disappears. Similar data were collected for the other two isomers. The derived rates, k ( E ) , for all three isomers are plotted as a function of the total absolute energy in Figure 4. This absolute energy, E , is hv (the photon energy) AHfoo(molecule) Eth, where Eth is the initial thermal energy of the dichlorobenzene molecules (0.144 eV) . It is evident that the three sets of data fall on the same k ( E ) curve. Yet, the energies in Figure 1 indicate that the ortho and meta compounds have nearly the same ground-state ion energies, while the p-dichlorobenzene ion is lower in energy by about 0.2 eV. Because the dissociation rate is a strong function of the ion internal energy, the rates for the para ion are expected to be slower than the dissociation rates of the other two isomers. The observed equality of the rates, over the whole range of energies investigated, strongly suggests that the three ions rearrange to a common precursor ion prior to dissociation. This lowest energy structure must be either the pdichlorobenzene or a still more stable isomer. In the absence of any evidence that lower energy isomers exist, we will assume that the isomerized ion has the para structure. The dissociation limit for the loss of one chlorine atom from dichlorobenzene ions is not known. Even a high resolution PIE scan of the C6H4C1+ fragment ion as a function of the photon
+
+
Olesik et al.
3566 The Journal of Physical Chemistry, Vol. 90, No. 16, 1986
TABLE 111: Vibrational Frequencies Used in RRKM/QET Calculations" C6H4CI2+ 3076 (4). 1505 (4), 1220 ( I ) , 1145 (4), 1093 (2), 880 (4), 686 (3), 550 ( l ) , b 445 (2), 350 ( l ) , 328 ( l ) , 298 ( I ) , 226 ( l ) , 122 (1) molecular 1011
C6H4C12+ transition state C6H4Br2+ molecular ion C,H,Br,transit ion state
3076 (4), 1505 (4), 1220 ( l ) , 1145 (4), 1093 (2), 880 (4), 686 (3), 445 (2). 245 ( l ) , 230 ( l ) , 209 (I), 158 ( l ) , 85 ( 1 ) 3060 (4), 1495 (4), 1214 (3), 1060 (4), 940 (2), 810 (21, 660 (3): 430 (3), 295 (2), 222 ( l ) , 170 ( l ) , 100 (1) 3060 (4), 1495 (4), 1214 (3), 1060 (4), 840 (2), 710 (2), 560 (2), 360 (3), 235 (2), 180 ( l ) , 140 ( l ) , 80 (1)
The neutral o-dichloro- and D-dibromobenzene fre~uencies~' were rounded off for calculational purposes. The degeneracies are in parentheses. bAssumed reaction coordinate.
energy would not help because the rate of dissociation at the onset is too slow for fragment ion signals to be observable. One method of determining the dissociation limit is to fit the rate data of Figure 4 with the statistical theory (RRKM/QET).15116The input required for these calculations are the molecular ion and transition-state vibrational frequencies and the activation energy. An entropy of activation, S t ,can be calculated from the two sets of vibrational frequencies by A S * = k In ( Q * / Q ) = k In [nq,*/nq,]
hu = 12.52 eV
hu-
(4)
in which the total partition function Q is the product of the individual vibrational partition functions, q, = [ 1 - exp(-hu,/ k 7 9 - ' . The advantage of calculating an entropy of activation at some arbitrary temperature is that it provides a convenient method for reducing the total of 59 vibrational frequencies of the dichlorobenzene ion and its transition state into a single parameter. In addition, it demonstrates that the calculated rate, &E), in Figure 4 is a function of just two parameters, the entropy of activation and the activation energy. The entropy of activation also provides a means for comparing the microcanonical rates with the thermal, or canonical, rates of similar neutral reactions. The vibrational frequencies of the parent ions are not known. This lack of knowledge concerning the vibrational frequencies of large ions is unfortunately the norm, rather than the exception. The assumption that the ionic frequencies are similar to those of the neutral molecule in cases where the PES shows a sharp onset [good Franck-Condon overlap for the (0, 0, 0, 0, ...) -,(0, 0, 0, 0, ...) band] is probably a good one, although there is evidence that even in such cases some of the ionic vibrational frequencies are lower than their neutral counterpart^.^^ In view of the rather sharp PES onsets and the lack of any information about the ionic frequencies, we have chosen to use those of the neutral molecule for the molecular ion. The frequencies of the transition state and the activation energy were then varied until the fit shown in Figure 4 (solid line) was obtained. The resulting activation energy is 3.35 f 0.1 eV. The frequencies, listed in Table 111, yield an entropy of activation of 1.8 cal/(mol deg) at 1000 K. This is somewhat lower than the AS*%of about 7 cal/(mol deg) for the X loss from the monohalobenzene ions, C6HSCI+ and C6H5Br+.4The origin of this difference is not clear. C. C6H,Br2+Dissociation Rates. The dissociation rates of the dibromobenzene ions were determined by both the 1-m and the 0.5-m monochromator instruments. Examples of the results from the 0.5-m instrument are shown in Figure 5. The daughter ions produced in the first acceleration region appear between 44 and 50 ys of the TOF distribution, while those that are formed in the drift region appear between 5 1 and 53 p s . The corresponding decay times for these two regions are 0-10 and 11-23 p s , respectively. If the first region consisted of a uniform (constant) acceleration, the first peaks in Figure 5 would look similar to those in Figure 3 . However, the voltage of one of the elements was varied for maximum signal, thereby creating a nonuniform acceleration region. The resulting T O F distribution has the extra (34) Anderson, S . L.; Goodman, L.; Krogh-Jespersen,K.; Ozkubak, A. G.; Zare, R. N.: Zheng, C.F J . Chenz. Phys. 1985, 82, 5329.
I
u hu =
"*
12.65 e V
45
12.92 e V
50
T I M E - OF-FLIGHT ( p s e c )
Figure 5. Product ion TOF distributions at various C6H4Br2' internal energies. Ions appearing between 45 and 50 ps decayed in the acceleration region, while products appearing at 52.5 ps decayed in the drift region.
hump at 48.5 MS and a large space (50-51 MUS) between it and the ions formed in the drift region. A comparison of the derived rates with the two methods indicates that the rate converge to the same value at low ion energies, but that at high energies the method of Figure 5 gave rates that rise much less steeply with ion energy than those obtained by using the method of Figure 3. This difference is a consequence of the poorer energy selection and the method of data analysis used in the smaller apparatus. Energetic electrons directed toward the electron detector cannot be stopped by the steradiancy analyzer. Thus, as the photon energy increases, the ion signal from lower energy ions becomes progressively more important. These ions tend to increase the signal in the second peak at 52.5 I.LS relative to the first peak. Thus, the derived rate using eq 3 will be too low. On the other hand, the method of analysis used in Figure 3, in which the first part of the TOF distribution is fit at the expense of the latter part, tends to ignore these low-energy ions that contaminate the energy selection. The other interesting aspect brought out in this comparison is that the method utilized in the data of Figure 5 can be in principle extended to lower energies and lower decay rates. Work in improving the energetic electron discrimination is presently in progress in order to take full advantange of these two techniques. Figure 6 shows the decay rates of the dibromobenzene ions as a function of the total ion energy, hv AHfoo Eth,where Eth is 0.156 eV for the dibromobenzenes. Although the data overlap is reasonably good, the agreement among the rates for the three isomers is not as good as it is for the case of the dichlorobenzene isomers in Figure 4. The uncertainty in the neutral heats of
+
+
The Journal of Physical Chemistry, Vol. 90, No. 16, 1986 3567
Dissociation Rates of Energy-Selected Ions
+
hv t E t h AH,: TOTAL ION ENERGY (eV) 14.0 14.2 14.4 14.6 148
'
I I
[ 2?
'
1
CH ,B ,r:
'
I
C6H,Br++
--f
'
I
Br
TABLE I V 0 K C-H and C-X Bond Energies in C6HbnXn+Ions 15.0
Y ' 11
/ l
W
ion C6H6+
C~HSC~' C6H5Brf
C-H
c-CI
85.3" 90 89
74.8'
C-Br
62.0d 11.26
o - 2.80 eV A S * =2.6 eu
106
(kcal/mol)
656
"Rosenstock et aL2, bThis work. cRosenstock et al.' dRosenstock et aL4corrected for the AH:, of C & k .
formation of about 2 kcal/mol (0.09 eV) may be a contributing factor for the imperfect overlap of the three sets of rates. If we assume that the ions do not isomerize, the rate curves should overlap when the energy scale is hv - IP. This manner of plotting the rates shifts the ortho and meta data to lower energies by 0.21 and 0.18 eV, respectively, a shift that makes the agreement among the three isomers far worse than the one shown in Figure 6. On the basis of these results, we conclude that the dibromobenzene ions rearrange to a common precursor structure prior to dissociation. However, this conclusion is not as well founded in the case of the dibromobenzene ion dissociation as it was for the dichlorobenzene ion fragmentation. The rate data of Figure 6 were modeled with the statistical theory as described for the case of the dichlorobenzene data. The vibrational frequencies that gave the best fit are shown in Table 111. An activation energy of 2.8 eV gave the fit shown in Figure 6. This activation energy is, within the experimental error, identical with that found for the loss of Br from monobromobenzene ion^.^,^ D . On the Loss of X 2 from c6H&2+. Figure I shows the PEPICO T O F data for one of the isomers of the dichloro- and dibromobenzene ions. An arrow points to the expected position of the C6H4+product ion. It is abundantly evident that this product ion is not formed even up to a photon energy of 14.2 eV (the limit of our H2 light source).
energy, this value can be identified with the bond dissociation energy. Voyksner et al.35 have recently measured the kinetic energy released upon the dissociation of metastable dichlorobenzene ions. They determined that the energy release at full width half maximum (fwhm) is 18, 24, and 29 meV for the ortho, meta, and para isomers, respectively. As pointed out by Holmes and T e r l ~ u w these , ~ ~ fwhm values can be converted in an approximate manner to average release energies by multiplying them by the factor 2.4, thereby yielding average release energies between 45 and 70 meV. The interpretation of these average energy releases can be considerably aided by the use of the statistical theory from which one can calculate the mean energy of the parent ions that give rise to these release energies. The rate data of Figure 4 show that metastable ions that live for about 20 p s have internal energies of about 4.8 eV or mean excess energies above the dissociation limit of 1.5 eV. K 1 0 t s ~showed ~ that the energy above the dissociation limit, E*, can be partitioned among the dissociation products according to E* = ( R + l)kT* + ( E ) , , where kT* is the translational energy, and ( E ) ,is the average vibrational energy. The fictitious temperature, T*, is introduced by equating the microcanonical energy E* with a canonical energy at the temperature T*. When estimated vibrational frequencies of the product ion, C6H4Clf,are used, the statistically expected energy release at an excess energy of 1.5 eV is 85 meV. Considering the approximations made in estimating the ion internal energy and in extracting the average experimental energy, this calculated release energy is reasonably close to the measured average release energy of 45-70 meV. We therefore conclude that this reaction distributes its excess energy statistically and that there is no reverse activation barrier. The above discussion justifies the association of the 3.35-eV activation energy with the C-Cl bond dissociation energy. Furthermore, this leads to a calculated AHfoo(C6H4Cl+)of 263 kcal/mol. Combining this with the AHfoo(H)makes possible a determination of the C-H bond energy in C6H5C1+. A similar analysis for the dibromobenzene results yields a fiHfoo(C6H4Br+) of 275 kcal/mol. Table IV summarizes the various C-H, C-C1, and C-Br bond energies that can be determined from these data. There is an encouraging self-consistency in these bond energies, which lends credence to the estimated dibromobenzene heats of formation. The results indicate that the bonds in the halogenated benzene ions are slightly stronger than they are in the mono, or nonhalogenated, compounds. Second, they illustrate why the H atom loss reactions are not observed at low internal energy in the C6H5X+dissociations. A difference of 15 kcal/mol (0.65 eV) in the activation energies for C1 and H atom loss from chlorobenzene ions leads to statistically predicted rates of H atom loss that are several orders of magnitude lower than the C1 atom loss rates. The difference is even greater for bromobenzene where the activation energies differ by more than 1 eV. The onset for X2 loss, if it were to occur with no reverse activation energy, can be calculated from the known C6H4+heat of formation of 315.7 kcal/mo1.22 The onset for C1, loss from C6H4C12+lies at a total energy of 13.69 ev. This is 1.06 eV above the onset for C1 loss (see Figure 1). Thus, there is no mystery associated with the lack of C6H4+signal. The ratio of the rates
Discussion A . C-H and C-X Bond Energies in Halobenzene Ions. The activation energy for the loss of C1 from dichlorobenzene ions is 3.35 0.1 eV. If we can assume that there is no reverse activation
(35) Voyksner, R. D.; Hass, J. R.;Bursey, M. M. Anal. Chem. 1983,55, 914. (36) Holmes, J. L.; Terlouw, J. K.Org. Muss Spectrom. 1980, IS, 383. (37) Klots, C. E. J . Chem. Phys. 1973, 58, 5364 1976, 64, 4269.
Para 0 Meta 0 Ortho
b
I
3.8
4.0 P-C,H.,Br:
/
I
I
1
1
1
4.2 4.4 4.6 4.8 INTERNAL ENERGY (eV)
hv + E f h - I P Figure 6. Derived dissociation rates as a function of the total ion internal energy [hu + AHfo,(C6H4Br2)+ &,I, where Eth is the thermal energy of 0.156 eV. The solid line is a statistical theory calculation of the rates assuming an activation energy of 2.80 eV and an entropy of activation of 2.6 cal/(mol deg) at 1000 K. I CI
l
l
1
y7
z L1
0 W
u
w z
e
C6H, CI'.
-z
s '6
14
.. ..
H4'
. .. 16
I8
. .
C6H:
.,
20 TIME OF FLIGHT (psecl
Figure 7. Ion TOF distributions near 14-eV photon energy for two of the six molecules investigated. No C6H4+signal was observed for any of the isomers.
*
3568 The Journal of Physical Chemistry, Vol. 90, No. 16, 1986 of CI to Cl, loss is predicted to be over IO4, even at a photon energy of 14 eV (14.36 eV total energy). A similar conclusion can be reached with regard to Br, loss from C6H4Br2+.The difference in activation energies between Br and Br, loss is 1.1 eV, thereby making the Br, loss noncompetitive at all energies. It is expected to be noncompetitive even at high energies because the entropy of activation is certain to be less for X, loss than it is for X loss. Thus, if the reaction is statistical. this channel can never overtake the X-loss channel. The major source of C6H4+in the mass spectrum of C6H4X2+must come from the sequential reaction C6H4X,+ C6H4X+ C6H4+. The second reaction has in fact been observed as a metastable ( m / z 1 1 1 76) transition by Brown.23 This lack of X, loss is in sharp contrast to the H2 loss from benzene ions.,, The latter dissociation limit is in fact below that of H loss. The difference arises from the large H2 bond energy relative to that of Clz and Br,. B. Dissociation Rates and the Statistical Theory. The excellent fit of the RRKM/QET rates to the experimentally observed rates was achieved by assuming that the ion, which is initially formed in an excited electronic state, internally converts to the ground electronic state. This assumption has been justified in numerous slow ionic dissociation reactions.25 A consequence of rapid internal conversion in ions such as those of benzene and the monohalobenzenes is that these ions do not fluoresce.3* The lack of fluorescence places a lower limit of about 10" s-I for the internal conversion rate, which is clearly much faster than the dissociation rates of metastable ions. it is interesting that a very weak fluorescence from the B 2B, to X ,A2 states in dichlorobenzene ions has been observed by Maier and Marthaler.39 However, this emitting state lies below the energy necessary for dissociation. On the basis of the low quantum yield, Maier and Marthaler estimate an internal conversion rate of 10" s-'. The higher electronic states that are excited in our PEPICO experiment are likely to internally convert at least as fast as the B state, so that we do not expect fluorescence to compete with internal conversion followed by dissociation. In the study of the dissociation rates of bromobenzene ions, Rosenstock et al.4 raised the interesting problem of the two spin-orbit states of the bromobenzene ground state, which are split by about 0.4 eV. If internal conversion produces both components, then the rates from these two states would differ considerably. The experimentally observed T O F distribution
-
-
-
(38) Klapstein, D.; Maier, J. P.; Misev, L. In Molecular Ions: Specrroscopy, Structure and Chemistry; Miller. T . A., Bondebey, V . E., Eds.; North-Holland: Amsterdam, 1983; p 175. Maier, J. P. In Kinetics of lonMolecule Reactions; Ausloos, P., Ed.; Plenum: New York, 1979; p 437. (39) Maier, J. P.; Marthaler, 0. Chem. Phys. 1978. 32. 419.
Olesik et al. would be two-component, an effect that would be clearly evident in the T O F distributions such as those in Figure 3. Such twocomponent decay rates have been observed in a number of ionic dissociations involving competition between direct dissociation and isomerizatior~.~~ The absence of two-component decay in the dibromobenzene data indicates that the internal conversion proceeds exclusively to one or the other of the two spin-orbit components. In the absence of evidence to the contrary, it seems reasonable to assume that the lower energy state is populated because this maximizes the density of vibrational states.
Conclusion The dissociation rates of the dichloro- and dibromobenzene ions have been measured by PEPICO. A comparison of the rates for the three isomers suggests that these ions isomerize to the lowest energy structure, which is most likely the para isomer. The measured rates are consistent with the rates calculated by the statistical theory (RRKM/QET). From a fitting of the rates with the theory, we calculate activation energies (C-CI and C-Br bond energies) of 3.35 and 2.8 eV for the dichloro and dibromo compounds, respectively. The results also make possible the determination of the C-H bond energy in the C6H5X+ions of 3.88 eV. This is 0.19 eV higher than the C-H bond energy in the benzene ion. Acknowledgment. We are grateful for helpful discussions with George Leroi and Mostafa El-Sayed and to Tom Bunn for help with some of the experiments. It is evident that all of the quantitative information such as bond dissociation energies derived in this work depend on the interpretation of the rate data with the RRKM theory. Beyond that, the whole conceptual framework is based on the statistical theory. The contributions of Rudi Marcus to this theory have been profound, and we are grateful for his keen insights. Finally, we acknowledge the support of the National Science Foundation and the Department of Energy for financial support of this work. Registry No. l,2-C6H4ClZ, 95-50-1; 1,3-C6H4CIz,541-73-1; 1,4C6H4C12,106-46-7; 1,2-C6H4Br2,583-53-9; l,3-C6H4Br2,108-36-1; 1,4C6H4Brz, 106-37-6; 1,2-C6H4CI2 (radical cation), 96244-41-6; 1,3C6H4C12 (radical cation), 68128-49-4; 1,4-C6H4C12(radical cation), 681 70-44-5; 1,2-C6H4Br2 (radical cation), 74350-74-6; 1,3-C6H4Br2 (radical cation), 74336-44-0; 1,4-C,H4Br2 (radical cation), 74365-38-1; C6H4Cl (radical cation), 55450-32-3; C6H& (radical cation), 5545033-4. (40) Brand, W. A,; Baer, T. J . A m . Chem. SOC.1984, 106, 3154. Baer, T.; Brand, W. A,; Bunn, T. L.; Butler, J. J. Faraday Discuss. Chem. SOC. 1983, 75. 45.
