infrared Spectra of the Molecular Ions and Radicals Produced by

uum-uv photolysis, Milligan and Jacox subsequently pro- duced matrix-isolated NOz-. A large number of negatively charged species have been produced as...
6 downloads 0 Views 1MB Size
L. Andrews, J. M. Grzybowski, and R. 0. Allen

904 Milligan and M. E. Jacox, ibid., 53, 2034 (1970). Work in progress in this laboratory, B. S. Ault and L. Andrews, has confirmed the anion identification. (16) V. E. Bondybey a d G.c. PIfnentel, J. Chem. W S . , 56,3832 (1972); D.

E. Milligan and M. E. Jacox, J. Mol. Spectrosc., 46, 460 (1973); L. Andrews, B. S. Ault, J. M. Grzybowski, and R. 0. Allen, J. Chem. Phys., 62, 2461 (1975). (17) J. A. stone, Can. J. Chem., 46, 1267 (1968), and referencestherein.

infrared Spectra of the Molecular Ions and Radicals Produced by Proton Radiolysis of Carbon Tetrachloride in Argon during Condensation at 15'K Lester Andrews,' Joseph M. Grzybowskl, and Ralph 0. Allen Chemistry Department, University of Virginia, Charlottesville, Virginia 2290 I (Received August 5, 1974:Revised Manuscrlpt ReceivedFebruary 12, 1975)

Samples of CCl4 and its carbon-13- or bromine-substituted counterparts at high dilution in argon were deposited at 15'K during simultaneous 2.0-keV proton beam irradiation. New infrared absorptions at 1037, 1020,927, 898,746,502,374, 290, and 242 cm-l are attributed to CCl4 radiolysis products and ion-molecule reaction products. The 898- and 746-cm-' bands are due, respectively, to the CC13 and CClz chemical intermediates. The 1020-, 927-, 502-, 374-, and 290-cm-l bands, which disappear upon photobleaching and neutralization with thermal electrons, are assigned to the matrix-isolated positive molecular ions CC13+, CC12+, C13+, Cl~CClz+,and C14+, respectively. The 242-cm-' absorption is likely due to C13- perturbed by a large cation. The 1037-cm-l band is due to a CC13+ vibration in an electrically neutral species, presumably CCl3+CI-. The fate of the bombarding proton is demonstrated by the observation of CHCl3 and HClz-, and in deuteron experiments, by the appearance of CDCl3 and DCIz-.

Introduction Ionic molecular species are of considerable spectroscopic and chemical interest for examination of mechanistic and bonding principles. Negatively charged species such as the superoxide anion have been synthesized using matrix reactions of alkali atoms which produced an ion pair. The first example of this type of species, Li+Oz-, was reported by Andrewsl in 1968. Using electron bombardment and vacuum-uv photolysis, Milligan and Jacox subsequently produced matrix-isolated NOz-. A large number of negatively charged species have been produced as alkali cation-anion pairs using the matrix reaction t e ~ h n i q u e . ~ However, positively charged molecular ions are more difficult to produce and trap owing to the fact that ionization potentials are relatively high on a chemical energy scale. Lewis acid adducts provide one means of producing cations; Gillespie and Morton4 have prepared C13+AsF6- and obtained its vibrational spectrum from mixtures of ClF, Clz, and AsFj. Jacox and Milliganj produced CCl3+ and CHClz+ using vacuum-ultraviolet photolysis of chloroform. Electrical discharge techniques also provide positive species for matrix isolation; a discharged flow stream of argon and CC4 condensed at 15'K revealed the infrared spectrum of cc13+and the cc13 radical.6 In unpublished studies, Current' irradiated a CC14-argon gas sample with 200-eV electrons immediately prior to condensation yielding several intense new bands. These bands included the trichloromethyl radicals and several positively charged molecular ions which were produced and identified in this study. using the proton bombardment technique described The Journal of Physicai chemistry, Vol. 75, No. 5, 1575

in the previous paper: and in a preliminary communication.lO A detailed infrared spectroscopic study was conducted on the matrix-isolated products formed by proton irradiation of argon-CX4 (X = C1, Br) matrix samples during 15'K deposition. Since the neutral chlorocarbon* and bromocarbonll intermediate species are well known, the present study will be concerned only with identification of the several new molecular ions produced by radiolysis and any new ionic species synthesized by ion-molecule reactions during sample condensation. Experimental Section The proton beam apparatus, parameters, and experimental technique have been described in the previous paper.Q While being bombarded with protons, samples of carbon tetrahalide vapor a t high dilution in argon ArICX4 = 400/1 (X = C1, Br) were deposited on a CsI windo@ held at 15OK. The reagents CCld (Mallinckrodt, reagent), CCl3Br (Aldrich, spectroscopic), CClzBrz (Eastern Chemical), CClBr3 (synthesized),ll CBr4 (Eastman Organic Chemicals, reagent), carbon-13 enriched CC14 (51 and 90%, Merck Sharpe and Dohme), nitrogen (Air Products, ultra high purity), and argon (Air Products, 99.995%) were used without purification except for CBr4 which was recrystal1ized.ll The gases used for ion production in the rf source, Hz, He (Matheson, high purity), and Dz (Matheson, CP), were also used directly. Infrared spectra were recorded in the 200-3000-~m-~ spectral range on a Beckman IR-12 filter-grating spectro-

Proton Radiolysis of CC14 in Argon

-

'

1

1020

905

I

960

I

I

I

'

I

" A

900 840 WAVENUMBER

"

I

720

'

I

*

'

680'520

I

'

A

r

1

48Ov%O0

(ern-')

I

' 360

I

I

I

320

280

Flgure 1. Infrared spectra of the major radiolysis products trapped during 2.0-keV proton irradiation and 15'K condensation of an Ar/CCI4 = 400/1sample for 17 hr, trace a. Spectrum b illustrates the effect of temperature cycling to 41'K and trace c is for a cycle to 50'K.

photometer. Survey scans at 20 or 40 cm-l/min were recorded before, during, and after sample deposition. Expanded scale scans were taken at 8 and/or 1.6 cm-l/min after sample deposition. The approximately 3-cm-' resolution was limited by the observed band widths. Wave number accuracy was &1cm-l except where otherwise indicated. In several experiments the infrared source beam was blocked to prevent radiation from reaching the sample during deposition and spectra were recorded using a Kodak No. 220 filter (transmits 400-4000 cm-l) in front of the Nernst glower to eliminate visible light from the sample.

Results A large number of matrix experiments have been performed in order to identify the intermediate radiolysis products of carbon tetrachloride. The rzsults obtained by proton beam irradiation of CCLj will be presented first followed by carbon-13 and bromine substitution experiments. Carbon Tetrachloride. Protons, deuterons, and a particles ranging in energy from 0.5 to 2.0 keV were used in this extensive matrix-isolation study of the radiolysis of CCL. The relative yields of two of the principal radiolysis products were a function of particle energy. However, the type of charged particle used in the bombardment did not affect the relative amounts of most of the observed species. The only exceptions were the products formed by chemically binding the bombarding particle, as will be discussed later. Figure 1,trace a, illustrates the major radiolysis products trapped during simultaneous 2.0-keV proton irradiation and condensation of an Ar/CCb = 40011 sample for 17 hr. The most intense product band, 898 cm-l, labeled A in Figure l a to conform to the symbols used earlier, has been

assigned to v3 of the CC13 radical.s The naturally occurring carbon-13 counterpart was observed at 869 cm-l (0.1 OD). The intense A+ band at 1037 cm-l has been attributed5 to CC13+. Other known reactive species present include CClz (720 cm-l, labeled E, and 746 cm-l, not shown),lZ CC1 (866, 860 cm-l),13 and HC12- (956, 697 cm-l).14 Stable molecules of interest include CHC13 (1224 cm-l),15 CHzClz (1269 crn-l),l6 CzCl4 (915 cm-l), and CzC16 (684 cm-1).8 The usual trace impurities, HzO and COz, were also present and took part in the radiation chemistry forming ClCO (1879 cm-l and a weak 572-cm-' band, not shown),l' C12CO (1815 and 838 cm-l and a 582-cm-l band, not shown),18 and CO (2141 crn-I).l9 The hydrogen containing species resulted in part from the H20 but, as shown by later deuteron experiments, also from reaction with the bombarding particles. The bands of particular interest illustrated in Figure l a are A+ (1037 and 1020 cm-l), E+ (927 cm-l), the C13+ (502 cm-l) multiplet, a 374-cm-l absorption, and the 290-cm-l band. All of the new absorptions observed following CC4 radiolysis are listed in Table I. Figure l b shows the diffusion behavior of the trapped species following temperature cycling to 41'K and trace c is for a cycle to 50'K. Spectrum b shows the loss of CClz (E, 720 cm-l) and the partial decrease of CC13 (A, 898 cm-l) and the growth of CzC16, which agrees with earlier findingsU8The intensity of the sharp ClzCO band (838 crn-l) was reduced due to aggregation, the new A+ bands and the 374 and 290 cm-l bands were halved and the new 927- and 502-cmW1absorptions were reduced to one-third of their trace a intensities by the diffusion operation. An intense band appeared a t 867 cm-l. The final diffusion (c) shows only the most stable CzC14 and CzC16 species along with a The Journal of Physical Chemistry. Vol. 79, No. 9. 1975

L. Andrews, J. M. Grzybowski, and R . 0. Allen

906

TABLE I: New Absorptions (em-I) Produced by the Radiolysis of Argon-CCl4 Samples during 15°K Depositiona Absorptions

AbsorptiQns

2140 co 838 c1,co 1879 ClCO 746 cc1, 1815 C1,CO 720 cc 1, 1269 CHzC1, 697 HC1,' 1224 CHC1, 684 c2c16 1037 CCl,+(Cl') 582 c1,co 1020 CCl,' 572 ClCO 956 HC1,' 5 02 Cl,' 92 7 CCl,+ 498 Cl,' 915 c2c14 494 C13+ 898 CCl, 491 Cl,' 869 %c1, 374 CCl,Cl,+* 866 ~ 3 5 ~ 1 290 C14+* 860 ~ 3 7 ~ 1 242 cC13'cl3-* 850 broad ? a Traces of HzO, COZ, and air are unavoidably present. An asterisk denotes tentative identification. small amount of CC13 remaining and the growth of a broad aggregate band at 850-870 cm-l. An analogous experiment was run for 22 hr irradiating with 2.0-keV deutrons. The spectrum, which is shown in Figure 3a of the previous papery was essentially identical with that of Figure l a except for several new deuterium containing species. The deuteron irradiated sample was subsequently neutralized by applying 10 V to the exposed bulb filament for 25 min. Resulting changes in the spectrum were as follows (see Figure 3b of the previous paperg): the 1020-, 927-, 502-, and 374-cm-l bands disappeared almost completely and the 645- and 290-cm-l bands were reduced to 70 and 45%, respectively, of their original intensities. The A+ band (1037 cm-l) along with the C2Cl.1 (915 cm-l), CC13 (A, 898 cm-l), CCl2 (E, 746, 720 cm-l), Cl2CO (838 cm-l), and DC12- (465 cm-l) bands were not affected by photobleaching and adding thermal electrons. The electron titration and photobleaching of numerous samples irradiated with protons or a particles followed the same pattern? Removal of the 1 0 2 0 - ~ m -band ~ by photobleaching revealed a weaker counterpart a t 1010 cm-l which was also ultimately photobleached. In the most productive experiment, illustrated in Figure 4a of the previous papery a moderately intense band was observed a t 242 f 2 cm-l which was noted as a shoulder in most of the other experiments. This new band decreased with photobleaching of the matrix sample, Figure 4b in the previous paper, as did the five bands of interest here. A 2.0-keV proton experiment was conducted with a nitrogen matrix sample, N2/CCl4 = 400/1, for 18 hr. The chlorocarbon absorptions were broadened and shifted 5-10 cm-l higher relative to their argon matrix counterparts, including the CC14 precursor. The CC13 radical band was observed at 904 cm-l (0.38 OD); CClz was not detected. In the higher frequency region, new bands were observed at 1042 (0.19 OD, 18-cm-l half-width) and 942 cm-l (0.27 OD, 16cm-l half-width). The lower frequency region yielded a sharp triplet at 505.3, 501.0, and 496.7 cm-l (0.68, 0.63, 0.18 OD), a broad 385-cm-l (0.12 OD) feature, a sharp 292-cm-l band (0.35 OD), and a sharp 243-cm-l absorption (-0.3 OD). The exposed bulb filament was operated at The Journai of Physical Chemistry, Voi. 75, No. 5. 1575

0

1dOG 860 ' 660 400 WAVE N UMBE R (crn-' )

'

~

I

Figure 2. infrared spectra of the proton beam radiolysis products of carbon tetrachloride and the carbon-13 enriched molecules: spectrum a, natural CCI4, 0.5-keV proton beam: spectrum b, 51 % carbon-13 enriched CC4, 0.5-keV proton beam; spectrum c, 90% carbon-13 enriched CC14, 1.0-keV proton beam. 10.5 V for 0.5 hr and the 942-, 500-cm-l triplet, and 385-, 292-, and 243-cm-l bands disappeared completely. Carbon-1 3 Tetrachloride. Experiments were conducted with carbon-13 enriched CC& using 0.5-, 1.0-, and 2.0-keV proton beam irradiations. Two of these experiments are contrasted with the natural isotopic material in Figure 2 using the lower beam energies which produced smaller yields of the trichloromethyl radical. Trace a illustrates a 0.5-keV experiment with natural isotopic CC&; note the reduced A band intensity. In a similar 51% l3Cc14 experiment, shown in trace b, the A+, E+, A, and 374-cm-l bands appeared as equal intensity doublets, which indicates that these species contain a single carbon atom. The 90% l3Cc14 experiment, illustrated in trace c, shows the dominant carbon-13 counterparts of the observed bands; Table I1 lists the observed wave numbers. In a particularly productive 2.0-keV 90% l3Cc14 experiment, the E+ band was observed at 927.0 f 0.2 cm-l (0,080 OD) and the 13E+ band was measured at 897.7 f 0.2 cm-1 (0.84 OD) on several highresolution scans, Photobleaching of the sample markedly reduced the 988-, 898-, 502-, 361-, and 290-cm-l band intensities without affecting the 1003- and 869-cm-l absorptions, A 980-cm-l side band emerged upon reduction of the 988-cm-l feature. The 810-cm-l 13COC12 and 723-cm-1 WC12 bands were not affected by the bulb filament. The 502-cm-l multiplet, the 290-cm-l band, and the 242-cm-l

Proton Radiolysis of CC14 in Argon

907

TABLE 11: Effect of Bromine Substitution on the New Absorptions (cm-I) Produced by Radiolysis of Argon-CX4 Samples' Precursor

CX,'

(x-)

CClh

(A') 1037.0

i3CC1,

(13A+)1003.0

CC1,Br

cx,+ 1020 (1010) 988 (980)

cx,+

cx3

x3+

502, 498, 494, 491 s h 502, 498, 494, 491 s h 492, 488, 484 s h 437, 420 432, 416 362, 339

(E+)927.0

(A) 898

(E) 746, 720

('3E*) 897.7

(I3A) 869

(13E) 723

(A) 898

(E) 746

(B) 889, 836 (B) 891 un, 836 (C) 856

(F)612 (F)612

(C) 856, 783 s h

(F) 612

(D) 773 (D) 773

(GI 641, 596 (GI 641, 596

(A') 1036.0, (E') 925 1034.0 (B') 1019, 957 (F*)870 CCl,Br, (B') 1019, 958 (E') 925 (C') 976, (F') 868 891 un CClBr, (C+)978, 975; (F') 870 s h 894, 891 (D+) 874,880 (G') 778 CBr, (D') 874, 880 (G') 778 a sh denotes shoulder, un denotes unresolved CT, B band.

feature showed no measurable carbon-13 shift in the 90% l3Cc14 experiments. Bromine Substitution. Bromine substitution in the CX, precursor molecule provided a pseudo-isotope for chlorine. Figure 3 contrasts the 750-1050-~m-~regions of the infrared spectra for 0.5-keV proton irradiations of CC14, CC13Br (1.0 keV in this case), CCIzBi-2, CClBr3, and CBr4 samples. Figure 4 illustrates the 200-700-~m-~regions for tlese same precursors following 2.0-keV proton bombardment. The new absorptions found upon bromine substitution are tabulated in Table 11. For all reagent molecules, the analogous 2.0-keV experiments produced more intense bands; however, the primary cation/radical ratio was greater for the 0.5-keV experiment, and for this reason these spectra were used for Figure 3. The C c 4 spectrum revealed the very intense A+ band at 1037 cm-l with a high-wave number shoulder at 1041 cm-l and a satellite band a t 1020 cm-l along with the intense E+ band a t 927 cm-l and A band a t 898 cm-l. These same bands were observed in the CC13Br spectrum with lower intensities but identical wave numbers, except E+ which appeared a t 925 cm-l. In addition CC13Br gave rise to bands at 1019 cm-l and a 957.0-, 9 5 3 . 5 - ~ m -doublet ~ labeled B+, B ( C c l ~ B rabsorptions )~ at 889 and 835 cm-l, and a band a t 870 cm-l designated F+. Both the F+ and the E+ bands were almost completely removed by a 30-min titration with thermal electrons while the A, A+, B, and B+ bands were not affected. In the CClzBr2 spectrum the B+ bands a t 1019 cm-l and at 958, 953 cm-l (doublet), and the B bands a t 891 (unresolved from C+ at 893 cm-l) and 836 cm-l were more intense. New C+ bands were observed at 976 and 891 cm-l (unresolved from B a t 889 cm-l), and E+ a t 925 cm-l, F+ at 868 cm-l, and a weak C band a t 856 cm-l also resulted. Continuing in the series, the CCIBr3 spectrum produced two C+ doublets 978, 975 cm-l and 894, 891 cm-l along with weak C bands (CCIBr2) a t 856 and 783 cm-l (shoulder). An intense new doublet at 880, 874 cm-l is labeled D+ and the sharp D (CBr3)11band is noted a t 773 cm-l along with another new band G f , a t 778 cm-'. The label I denotes impurity bands from the precursor sample. The CBr4 spectrum contains these last three bands which are labeled D+ (880, 874 cm-l, 0.65 OD), G+ (778 cm-l 0.12 OD), and D (773 cm-l, 0.08 OD). In CBr4 experiments using 2.0-keV proton and a particle beams, the D+ doublet became a single sharp

cx,x,+ x,i x,

CXZ

(GI 640

374

290

242

361

290

242

251

225

356 355

221 230

355 283 283

326 326

Y l

t

,

1000 WAVENUMBER

,

,

l

900

Del ,

,

,

,

l

*

800

kin-')

Figure 3. Infrared spectra in the 750-1050-cm-' region for 0.5-keV proton radiolysis products of CX4 molecules (X = CI, Br). Ar/CX4 = 40011. CCI3Br experiment used 1.O-keV proton beam.

band a t 874.0 (0.43 OD) cm-l with a sharp 878-cm-l shoulder, the D radical band at 773 cm-l (1.5 OD) and the G+ (0.75 OD) absorption increased markedly in intensity. Other species of interest observed in these CBr4 experiments were CHBr3 (1154 cm-l), CHzBrz (651 cm-l), The Journal of Physical Chemistry, Vol. 79, No. 9, 1975

L. Andrews, J. M. Grzybowski, and R. 0. Allen

908

520 480 WAVE N UMBER

(c m-'1 Figure 5. Expanded-scale infrared spectra of the multiplets near 500 cm-' in CCI4 and CC13Br experiments. ly from 0.05 to 0.01 OD. The D (0.28 OD) and D+ (0.27 OD) bands were unchanged.

800 600 400 WAVE N U M BE R

200 ( c m" )

Figure 4. Infrared spectra from 200 to 700 cm-' for a 2.0-keV proton beam irradiation of CX4 matrix samples during 15'K deposition.

HBr2- (729, 894 cm-1),20 and CBrz (640.5 cm-l, 0.54 OD and 595.7 cm-l, 0.06 OD).ll In the low-frequency region, Figure 4 shows CClz bands a t 746 and 720 cm-l, the HC12- band at 697 cm-l, and previously unidentified bands a t 502, 374, 290, and 242 cm-l in the CC14 spectrum. The CCl3Br parent gave a new band a t 490 cm-l, two bands a t 437,420 cm-l, a sharp 356-cm-l band, an absorption a t 251 cm-1, and a doublet a t 221, 225 cm-l. These bands are listed in Table 11. In the 1.0-keV experiment, all but the latter doublet were observed, and all of these bands almost completely disappeared upon addition of thermal electrons for 3 hr. The 502-cm-l multiplet in CC14 experiments consisted of four bands: 502.0 (0.12 OD), 498.0 (0.14 OD), 494.5 (0.07 OD), and 491 cm-l (shoulder, 0.02 OD). The 490-cm-l band in CC13Br experiments was resolved into absorptions a t 492.4 (0.09 OD), 488.4 (0.07 OD), and 484 cm-1 (shoulder, 0.02 OD). Expanded-scale spectra of these multiplets are contrasted in Figure 5. For these and the other halogen containing species in Table 11, a gradual shift was observed to lower wave numbers with increasing bromine substitution. In the 1.0-keV proton radiolysis study of CBrd, the final sample was neutralized with thermal electrons for 5 hr with the filament a t 6 V and for an additional 5 hr a t 10 V. The 778-cm-l G+ band was reduced from 0.36 to 0.25 OD and the 283-cm-l absorption decreased from 0.06 to 0.04 OD while the 326-cm-l absorption decreased almost completeThe Journai of Physical Chemistry, Vol. 79, No. 9. 1975

Discussion Chemical identification of the new ionic species produced here will each be discussed along with some postulates of the mechanisms responsible for their formation. In the previous paper: five infrared absorptions, 1020, 927, 502, 374 and 290 cm-l, were attributed to matrix-isolated cations owing to their marked decrease when the sample was photolyzed with near-infrared and visible light which releases trapped electrons and when thermal electrons were added to the sample. Since the counterparts of the new argon matrix bands were also observed in solid nitrogen (assuming the broad 1042-cm-l nitrogen matrix band incorporates both of the 1037- and 1 0 2 0 - ~ m -argon ~ matrix features), the matrix host is not chemically combined in any of the new molecular species derived from CCL. The cations responsible for each of the five infrared absorptions will be identified from the available spectroscopic evidence in the following discussion. CC13+. The trichloromethyl carbonium ion was first identified by Jacox and Milligan (heceafter JM) following hydrogen resonance vacuum-uv photolysis of CHC13 samples without any mention of the possibility of a counterion? The J M experiment was similar to the present technique; both irradiations were performed during sample deposition but the JM ionizing radiation was 10-eV photons while 2000-eV protons were utilized in this study. In both cases the major molecular species were the same. The 1036.6-cm-l band observed by J M was seen as sharp band peaks a t 1036.5 to 1037.0 cm-1 in this work. This band was not affected when the sample was neutralized with thermal electrons or photobleached while the new band observed nearby a t 1020 cm-l was completely neutralized, revealing a weak band a t 1010 cm-l which itself was ultimately destroyed. This indicates that the 1 0 3 7 - ~ m -absorber ~ is part of an electrically neutral species and the 1 0 2 0 - ~ m -band ~ is a matrix-isolated cation. The unusually high frequency for

Proton Radiolysis of CC14 in Argon

9 09

these carbon-chlorine vibrational modes suggests cationic species. Since the most abundant anion produced by either type of radiation, other than electrons, must be C1- owing to the fact that atomic chlorine is an excellent electron trap, it is proposed that the 1037-cm-l absorber is the CCl3+C1- ion pair, which results from the combination of CC13+ and C1- ions during sample condensation. Chloride ion is also produced by the dissociative electron capture of CC&. The 1020-cm-l absorption is assigned to v3 of the isolated CC13+ species; the 1010-cm-l band is presumably due to CC13+ isolated in a matrix site more resistant to photobleaching. The 1037-cm-l band is assigned to the v3 mode of CC13+in the CC13+Cl- species. Support for this assignment comes from the bromine substitution data which indicate that the 1037-cm-l band is doubly degenerate and that two bands result upon partial bromine substitution. That four, and only four, unique species can be produced from the five tetrahalide precursors in Figure 3 defines the new species as containing three equivalent halogen atoms. Similar data has been presented for the CX3 radicah8 The mixed-carbon isotopic experiment indicates a single carbon atom species, and the carbon-13 shift is appropriate for the v3 mode. In agreement with JM, the 1037-cm-l absorption must be caused by a CC13+ vibration; however, the failure to neutralize the species with electrons indicates the immediate presence of a counterion. The unique and perhaps unexpected stability of CCb+ in the presence of a counterion is probably due to its planar geometry. CC13+ must be a D3h species since it is isoelectronic with BC13. For CC13+C1- to charge transfer to give cc14, the planar CC13+ group must deform to the tetrahedral geometry and this process requires an activation energy. Interestingly, the CCb+Cl- ion pair is not very unstable thermodynamically.21

cc14

D = 6 8 i 3 kcal /mol

cc1,

+ c1

(1)

k = 2 0 2 JBz-83

CCl,' C1' The gaseous ion pair requires 187 kcal/mol for its production, but electrostatic attraction between the ions can provide some of this energy. If the interionic distance is 2.0 %., the conversion CC14 CC13+Cl- is endothermic by only 2 1 kcal/mol. It appears that a coplanar ion-pair arrangement 1 might

-

c1-C'.

/cl \

. -c1-

c1 1 resist conversion to CC4 more than a CsOstructure with C1- above the positive carbon center. Thus, if CCh+ and C1- produced by radiolysis approach each other during condensation of the matrix sample, the ion pair could be stabilized in the matrix unless the kinetic energy of the ions is greater than the activation energy necessary to deform the cc13+group and charge transfer to yield CC14. Resistance of the sharp 874-cm-l band in the CBr4 experiments to photobleaching and neutralization with thermal electrons indicates the formation of the analogous CBr3+Br- ion pair which is presumably stabilized by the planar geometry of CBr3+. In fact all of the B+ and C+ species must also be CX3+X- ion pairs; the B+ bands are

assigned to CClZBr+ modes and the C+ absorptions are due to CClBr2+ vibrations. The 2- (two) cm-l variation in splittings and band positions might be due to a change in counterion between C1- and Br-. In the ?r molecular orbital scheme for a planar CC13 radical, the free-radical electron is in an antibonding molecular orbital and its removal to form planar CC13+ clearly leads to stronger C-Cl bonds as indicated by the increase in v3 frequencies from 898 to 1020 or 1037 cm-l. This is, of course, expected since the binding in the cation will be greater due to increased nuclear attractions for the remaining electrons and decreased electron-electron repulsions as compared to the radical. CCl2+. An electron-impact mass spectrum of C C 4 showed CC13+ to be the major ion produced; however, the CC12+ ion signal was one-fourth as intense. The sharp E+ band at 927.0 cm-l exhibited a carbon-13 counterpart a t 897.7 cm-l. It was markedly decreased by sample diffusion, by bombardment with protons after sample deposition, and i t could be completely eliminated with thermal electrons and by photobleaching (see Figures 2 and 3 of the previous paperg) all of which indicates that it is due to an isolated cation. The bromine substitution data show three different species, E+ (927 cm-l), F+ (869 cm-l), and G+ (778 cm-l), all of which could be neutralized by thermal electrons. The observation of three different bands indicates two equivalent halogen atoms in the absorber. In fact, the observation of the E+ band in C C 4 , CCIsBr, and CClzBrz experiments is indicative of a two chlorine species. Confirmation of this stoichiometry was found in the high-resolution spectrum of the 927-cm-l band recorded by Current' which showed resolved chlorine isotopic splittings with a 9/6/1 relative intensity. This observation requires two equivalent chlorine atoms for the absorbing species. The 927.0 f 0.2-cm-l band and its 897.7 f 0.2-cm-l carbon-isotopic counterpart are therefore assigned to 113 of the 12CC12+ and 13CC12+ species, respectively. This isotopic data can be used to predict a lower limit of 115' for the (Cl-C-Cl)+ valence angle which is larger than the 100' value found for CClzl2 and is consistent with the known relationship between 17 and 18 valence electron molecules (NO2, 134O and 03,117O, for example). The true bond angle for CC12+ is probably 125 f 10 '. The bromine substitution work supports these assignments. An absorption in the CBr4 experiments at 778 cm-l is assigned to u3 of CBr2+. The F+ bands at 869 f 1 cm-l are assigned to CClBr+. Here the electrons removed from CClz and CBr2 to form CC12' and CBr2+ are probably nonbonding as far as the carbon-halogen bond is concerned. Nevertheless, the v3 frequencies increase substantially. This is in part due to expanding the valence angle which incorporates more s character into the bonding and the fact that n nuclear charges can bind n - 1 electrons in a halogen system more tightly and give rise to a stronger bonding. The infrared spectrum was closely examined for double positively charged species, CC122+ and CC132+, which were present in the mass spectrum with 2-3% of the CC12+ and CC13+ intensities. CC122+, isoelectronic with C02, would be expected to have an intense u3 absorption above 1050 cm-1. No additional bands were observed which could be assigned to double positively charged species. Such species are produced by radiolysis in lower yields than single charged cations, and are presumably more reactive during The Journal of Physical Chemistry, Vol. 79, No. 9, 7975

L. Andrews, J . M. Grzybowski, and R. 0. Allen

91 0

sample condensation. l'hus, 11 present, the yields were too low to be observed. Furthermore, if the second ionization potentials of the CClz and CC13 species exceed the ionization potential of argon, charge transfer would produce Ar+ and the singly charged cation. I t can be noted from Figure 3 that bromine detachment occurs more readily upon radiolysis than chlorine separation. This is consistent with mass spectra for the CCl3Br and CC12Br2 compounds. In the former, the Br+/Cl+ ratio was 23/10 while the latter gave 16/4. Accordingly, the CCl+/CBr+ ratios were 27/10 and 11/8 for the two precursors, respectively. The CC12+/CClBr+ ratio from CC13Br was 52/12 and the CC12+/CC1Br+/CBrzt yield from CClzBrz was 18/16/1. The carbonium ion yields in Figure 3 and in the mass spectra also matched nicely. The CCbBr mass spectrum yielded a CC13+/CClzBr+ ratio of 75/33 and the CC12Brz mass spectrum gave CClZBr+/CClBrz+ = 48/ 10. The CX3 radical absorptions in Figure 3 followed the intensities of their CX3+ analogs with more intense radical bands for the departed bromine atom. Since the prominent 1037-cm-l band has been assigned to the CC13+C1- ion pair, the possibility of a CClZ+Cl- ion pair should be explored. However, there is no geometry barrier to resist the formation of planar CC13 from CClZ+Cl- when the latter ions approach each other during sample condensation. So the CC12+ and C1- ion association increases the yield of CC13 radical. In the case of CC13+C1-, the different planar CC13+ geometry in the ion and in the tetrahedral CC14 neutral makes it possible to trap some CC&+Cl- as a stable ion pair in the solid matrix. C/3+. The quartet near 500 cm-l in Figure 1 exhibits chlorine isotopic splittings which requires three chlorine atoms, and since no carbon-13 shift could be detected, the absorber probably contains no carbon. Statistical weights of the six naturally occurring chlorine isotopic species with three chlorine atoms are as follows: 35-35-35 27/64 27 35-35-37 35-37-35

18/64) 27 9/64

3 7-3 7-3 5 37-35-37

6/64} 3/64

37-37-37 1/64 1 If the position of the different isotopes within the molecule cause splittings which can not be resolved, a 27/27/9/1 quartet should result. The observed bands are in excellent agreement with this possibility. The complete and rapid neutralization of the 500-cm-l band by thermal electrons indicates that this absorber is also an isolated cation, which must therefore be C13+. Strong support for assigning this infrared band to u3 of C13+ is found in the vibrational spectrum of C13+AsF6-, where the antisymmetric stretch of C13+ has been attributed to a 508-cm-' Raman band.4 The agreement between the argon matrix and Lewis acid adduct Cis+ frequencies is excellent. The frequency analogy to the isoelectronic SClz (v3 = 535 cm-1)22a molecule is also supportive. The bromine substitution data add strength to this assignment. The first band a t 492, 488, 484 cm-l in the CCl3Br experiment arises from a ClZBr+ species. The isotopic triplet shown in Figure 5 is more consistent with a (ClBr-C1)+ species with two equivalent chlorine atoms. The 437-, 420-cmV1bands are candidates for a (Cl-Cl-Br)+ arThe Journal of Physical Chemistry, Vol. 79, No. 9, 1975

rangemenLZzb The CClzBrz experiment produced four bands at 432,416,362, and 339 cm-l which could be due to two isomeric BrzC1+ species. Finally, the perbromo species, Br3+, was observed at 283 cm-l in the CClBr3 and CBr4 experiments. The failure to observe C13+ in the CC13Br experiment is due to the fact that bromine detachment is favored over chlorine separation by an approximate factor of 3 (see Figure 3). Accordingly, ion-molecule reaction products containing halogen species will be preferentially bromine substituted using mixed chloro-bromo precursors. The chlorine-bromine shift in the fundamental ratio of 502/283 = 1.77 is appropriate for a pure halogen vibrational mode and is in excellent agreement with the C12/Brz fundamental ratio23of 554/318 = 1.74. These trihalo cations are presumably synthesized by the ion-molecule reaction 2 during sample condensation rather c1, + C1' ---t Cl,' (2) than by a direct ionization process since no C13+ was detected in the mass spectrum of cc14. Gillespie and Morton4 have proposed an analogous mechanism for the formation of C13+AsFs- from an AsF5 solution of C1F and Cl2 where Cl+AsFg- is first formed followed by the addition of Cl2 to give C13+AsFe-. The bonding in trihalide ions has been discussed by Pimente124 in simple molecular orbital terms using only pr atomic orbitals. Very recently, matrix isolated Cl3- has been observed at 374 cm-l in proton bombardment experiments with C12.25The two electrons removed in forming C13+ from C13- are nonbonding. However, chemical binding is strengthened owing to the greater attraction of 51 nuclear charges for 50 electrons in the trichloride cation as compared to 52 electrons in the trichloride anion. Accordingly, U S increases from 374 cm-l for Cl3- to 502 cm-l for C13+ as the number of nonbonding electrons is decreased and the effective nuclear charge for the remaining electrons is increased. Evidence for CCl2(Cl2) +, The 374-cm-l band observed here exhibits a large carbon-13 shift consistent with the vibration of carbon against two chlorine atoms. The bromine substitution data provide two counterpart bands a t 355 and 326 cm-l which form a uniform progression and indicate that this vibrational mode involves two equivalent halogen atoms. The rapid and complete disappearance of these three bands following the addition of thermal electrons indicates that the absorber is an isolated cation. The chlorine-bromine shift from 374 to 326 cm-l (a ratio of 1.15) is consistent with a carbon-halogen stretching mode (ratios for v3 of CClZ/CBrz = 1.16 and for u 1 of CClz/CBrz = 1.21)8J1 and not an X-C-X bending mode (ratios for v4 of CCld/CBr4 = 314/183 = 1.72 and for v2 of CC12/CBr2 = 326/196 = 1.66).26aHowever, 374 cm-l is about half of the wave number of a normal C-C1 stretching fundamental, so the 374-cm-l absorber involves particularly weak carbonhalogen bonds. This suggests another ion-molecule reaction product. The data limit the possible sources for this C--C12 stretching mode to the ions CC13+ or CC12+ bonded to molecular Clz or the molecules CC13 or CC12 bonded to the ion C12+, the latter of which could be formed by reaction 3. Ste-

-

c 1 + C1' Cl,' (3) ric arguments suggest a preference for the (CClZ--Clz)+ species, which satisfies tetravalent carbon, over the more hindered (CC13--C12)+ possibility. Thus, a reasonable explanation for the 374-cm-l absorption is the product of

Proton Radiolysis of CC14 in Argon

91 1

reactions 4 which occur during sample condensation. This cc1, + Cl,+ Cl,C:Clz+ (4) CC1,' + c1,

>

C12C:C12+ species might be a nonclassical carbonium ion which has the CzUstructure suggested in 2. Two carbon-

%, ,c1 CldC'=c, 2

chlorine bonds involve normal electron pair bonding while the other two involve electron-deficient bonding with presumably delocalized electrons around the cation center. This structure nicely accounts for the low 374-cm-l C-Cl2 vibrational frequency and the large carbon-13 isotopic shift. The structure suggested is a more stable form of CC14+than the tetrahedral structure formed upon electron impact since the latter decomposes before detection in the mass spectrometer. The formation of a C1-C1 bond about the cation center provides a mechanism for stabilization of Cc14'. Support for this nonclassical ion structure is found in very recent studies26bon CHj+ which indicate a C, structure containing three normal C-H bonds and two weaker C-H bonds involved in three-center electron deficient bonding around the cation. It should be pointed out that the 4800-A optical band observed by Hamill and coworkers27 following radiolysis of CC14 in hydrocarbon glasses and attributed to the positive ion CC14+ could, in fact, be due to the species 2 suggested here. Evidence for C14+. The 290-cm-l absorber is probably a halogen species since no carbon-13 shift was detected in the 90% carbon-13 experiment. Bromine substitution products were observed at 251 cm-l in CC13Br and at 230 cm-l in CCl2Br2 experiments. The more completely bromine substituted species were not observed due to the 200-cm-l spectrophotometer limit; however, if the Br2/C12 wave number ratio is multiplied by 290 cm-l, a 167-cm-l absorption is predicted for the perbromo species. Although we cannot be certain, the space between the observed 230cm-l band and the predicted perbromo band at 167 cm-l suggests another unobserved band around 200 cm-l. This possible interpretation provides five bands which suggests a four-halogen atom species for the cation absorbing at 290 cm-l. Interestingly, the 290-cm-l band appeared to photobleach faster and neutralize with thermal electrons slower than the other cation bands. On the basis of the present data, the 290-cm-l band is tentatively assigned to the weak intermolecular bond stretching mode of the Cl2. C12+ species. It is proposed that a three-electron bond is formed between Cl2 and C12+ using the ( T * - T * ) ~scheme of Spratley and Pimente128 following ion-molecule reaction 5. In support of this assign-

-

c1, +

Cl,'

e

-

Cl,.' *c1,+

(5 )

ment is the half-order bond stretching frequency of Cl2observed near 250 cm-l for the M+C12- species.29 While the Ck+ stoichiometry and assignment cannot be regarded as definitive, it is the best present explanation for the 290cm-l absorber. Evidence for C13-. The 242-cm-l band appeared as a shoulder in samples exposed to the Nernst glower during deposition; however, in the most productive experiment

where the sample was protected from the glower, the 242cm-l band was observed to be moderately intense (0.15 OD) on the first infrared scan. It was reduced to 0.03 OD following prolonged exposure to the full light of the glower. In fact, this absorption was the most sensitive to photobleaching of all the observed bands. No carbon-13 shift was observed for the 242-cm-l band so it is presumed to arise from a halogen species. The only bromine-substituted counterpart that could be observed was the 221-, 225-cm-l doublet in CCl3Br experiments. Although it is not possible to eliminate positively charged species from consideration, the agreement between the 242-cm-l argon matrix band and the broad 242-cm-l absorption assigned to v g of C13- in the R4N+C13- species dissolved in acetonitrile30 suggests the tentative assignment of the former to C13-. Likewise, the 221-, 225-cm-l doublet is in very good agreement with the 227-cm-l band of C12Br- in R4N+C12Br-. However, matrix-isolated C13has been observed at 374 cm-l in recent proton bombardment experiments with C12.25If the 242-cm-l band is due to C13- as its agreement with the earlier assignment suggests, it must likewise be perturbed by a large cation. It is conceivable that CC13+ could be a possible counterion, and its own absorption is obscured by the bands already discussed between 1000 and 1040 cm-l. Reaction 6 suggests CCl,+Cl- + c1, CCl,+Cl,' (6 a mechanism for the formation of this species which could readily photolyze to give the initial reactants or CC14 and Cl2. This reaction is consistent with the formation of Br2C1- by the reaction of R4N+C1- with Br2.30 Finally, observation of the isolated ions HC12- and DC12- in these e ~ p e r i m e n t s l indicates ~.~~ the participation of chloride ion in the matrix reactions and supports the presence of the related trichloride ion. Bichloride ion is formed by reaction 7 , and it contributes to the charge balance in the matrix sample. HCl + ClHC1,(7 1 Matrix-Isolated Cations. This work has shown that isolated cations can exist in an argon matrix which also must contain isolated anions including C1-, Clz-, and HC12-. The subject of isolated ions in rare gas matrices has been discussed by Kasai, who presented ESR evidence for cations and anions isolated in argon matrices at distances on the order of 10-%,apart.31 The existence of matrix-isolated ions indicates an argon-cation "solvation" energy arising from ion-induced dipole forces (probably somewhat larger than van der Waals energies) which is instrumental in holding unlike charged ions apart. In order to determine the average separation of reactive species in these experiments, it is necessary to determine the concentration of products. Not all of the 1 part of CCb in 400 argon atoms are decomposed; however, more than one product may be formed from a given CCl, molecule. If the concentration of charged species were as high as 1 in 500 argon atoms, the average separation between these species would be approximately eight argon atoms or approximately 30 A. The electrostatic attractive energy between a cation and anion separated by this distance in solid argon of dielectric constant 1.53 is approximately 7 kcal/mol. The rigidity of the matrix which provides a barrier to diffusion and the "solvation" of cations by the matrix are sufficient to overcome this attractive energy and hold cations and anions apart at these low temperatures and dilute concentrations. It is therefore reasonable for several of the charged

-

-

The Journalof Physical Chemistry, Vol. 79, No. 9, 7975

L. Andrews, J. M. Grzybowski, and R. 0. Allen

91 2

species observed here to be "effectively isolated" within the matrix without adjacent counterions.

tron impact work and for permission to quote his experimental results here.

References and Notes

Conclusions The simultaneous radiolysis of an argon-CC14 sample with 2.0-keV protons during condensation at 15OK provides a fruitful method for studying the chemical intermediates (radicals and cations) of CCld radiolysis, and the ionmolecule reactions of these products. Photobleaching of the sample releases trapped electrons to neutralize isolated cations. The residual positive charge on the matrix sample allows the isolated cations to be neutralized by thermal electrons and, therefore, identified as cations. The CC13 and CClz intermediate species were observed in good yields. A significant amount of the principal charged radiolysis product CC13+ undergoes reaction with an anion, presumably C1-, to give an electrically neutral ion pair CCl3+Cl- which was not photobleached or neutralized by thermal electrons; some isolated CC13+ was also trapped. The CC12+ species was observed to be isolated in the matrix, based upon its complete and rapid neutralization with the addition of thermal electrons to the sample and upon photolysis. Three absorptions were identified as the cation-molecule reaction products C13+, C12C:C12+, and C14+, which were photobleached and neutralized by thermal electrons. Spectroscopic evidence was also presented for the anion-molecule reaction product HClz- and the species CC13+C13-. This work demonstrates that cations can be trapped in an argon matrix at 15OK. Solvation of the cation by argon arising from ion-induced dipole interactions and the activation energy for diffusion through the dielectric solid argon prevent neutralization and allow their preservation as isolated ions. Acknowledgments. The authors gratefully acknowledge financial support for this research by a National Science Foundation Institutional Subgrant from the University of Virginia Research Policy Council and an Alfred P. Sloan Fellowship for L.A. We thank Dr. J.W. Nibler for the tesla coil discharge spectrum and Dr. D.F. Hunt for the mass spectra. Special acknowledgment is due Dr. J.H. Current for providing an unpublished report on his significant elec-

The Journal of Physical Chemistry, Voi. 79, No. 9, 1975

(1) L. Andrews, J. Chem. Phys., 50, 4288 (1969). (2) D. E. Milligan and M. E. Jacox, J. Chem. fhys., 52, 3664 (1970). (3) L. Andrews, "Infrared and Raman Spectroscopic Studies of Alkali-Metal Atom Matrix-Reaction Products", in "Cryogenlc Chemistry", Wiley-lnterscience, New York, N.Y., 1975. (4) R. J. Gillespie and M. J. Morton, Inorg. Chem., 9, 81 1 (1970). (5) M. E. Jacox and D. E. Miillgan, J. Chem. Phys., 54, 3935 (1971). (6) J. W. Nibler, unpublished results. (7) J. H. Current, unpublished results. (8) L. Andrews, J. Chem. Phys., 48, 972 (1968). (9) R. 0. Allen, J. M. Grzybowski, and L. Andrews, J. Phys. Chem., previous paper. (IO) L. Andrews, R. 0. Allen, and J. M. Grzybowski, J. Chem. fhys., 61, . 2156 (1974). (1 1) L. Andrews and T. G. Carver, J. Chem. Phys., 49, 896 (1968). (12) L. Andrews, J. Chem. Phys., 48, 979 (1968). (13) M. E. Jacox and D. E. Milllgan, J. Chem. fhys., 53,2688 (1970). (14) P. N. Noble and G. C. Pimentel, J. Chem. Phys., 49, 3165 (1968); D. E. Milllgan and M. E. Jacox, /bid., 53, 2034 (1970). Work In progress in this laboratory, 8. S. Ault and L. Andrews, has confirmed the anion identification. T. G. Carver and L. Andrews, J. Chem. Phys., 50, 4235 (1969). L. Andrews and D. W. Smith, J. Chem. Phys., 53, 2956 (1970). M. E. Jacox and D. E. Milligan, J. Chem. Phys., 43, 866 (1965). A phosgene (Matheson) sample, Ar/CI&O = 400/1 was examined after 20 min of sample deposition to accurately measure the 582.0-, 838.0-, and 1816-cm-' CipCO bands. 2.0-keV proton irradiation and sample deposition for 15 hr produced an excellent spectrum of ClCO In solid argon; the bands were observed at 1878 (0.50 OD), 572 (0.14 OD), and 285 cm-' (0.04 OD). G. E. Leroi, G. E. Ewing, and G. C. Plmentel, J. Chem. Phys., 40, 2298 (1964). V. E. Bondybey, G. C. Pimentel, and P. N. Noble, J. Chem. Phys., 55, 540 (1971); D. E. Miiligan and M. E. Jacox, bid., 55, 2550 (1971). Work in progress in this laboratory has confirmed the anion identification. V. I. Vedeneyev et al., "Bond Energies, Ionization Potentials and Electron Affinities", Edward Arnold, London, 1966. (a) H. Stamrneriech, R. Forneris, and K. Sone, J. Chem. Phys., 23, 972 (1955). (b) BrCl in solid argon absorbs at 436 and 428 cm-'. The 437cm-' band may Include absorption due to BrC1. W. Holzer, W. F. Murphy, and H. J. Bernstein, J. Chem. Phys., 52, 399 \119701 (24) G. C. Pirnentel, J. Chem. Phys., 19, 446 (1951). (25) L. Andrews, B. S. Auit, J. M. Grzybowski, and R. 0. Allen, J. Chem. fhys., 62, 2461 (1975). (26) (a) D. E. Tevault and L. Andrews, J. Mol. Spectrosc., 54, 110 (1975); J. Am. Chem. Soc., in press; (b) M. D. Sefcik, J. M. S. Henis, and P. P. Gaspar, J. Chem. Phys., 61, 4321 (1974), and references therein. (27) J. P. Guarino and W. H. Hamill, J. Am. Chem. Soc., 86, 777 (1964); D. W. Skelly and W. H. Hamill, J. Phys. Chem., 70, 1630 (1966); T. Shida and W. H. Hamill, J. Chem. Phys., 44, 2369 (1966). (28) R. J. Spratley and G. C. Pimentel, J. Am. Chem. Soc., 88, 2394 (1966). (29) W. F. Howard, Jr., and L. Andrews, J. Am. Chem. Soc., 95, 2056 (1973). (30) J. C. Evans and G. Y-S. Lo, J. Chem. Phys., 44, 3638 (1966): 45, 1069 (1966). (31) P. H. Kasai, Acc. Chem. Res., 4, 329 (1971). - I