The near ultraviolet spectra of benzene in inert solid solutions

JOURNAL

OF MOLECULAR

The

SPECTROSCOPY

Near

Y. Department

18,

158-169 (1965)

Ultraviolet Spectra of Benzene Inert Solid Solutions

DIAMANT,

R.

of Chemistry,

R4. HEXTER,* Israel Institute

AND 0.

in

SCHNEPP

of Technology,

Haifa,

Israel

The near ultraviolet absorption and phosphorescence spectra of benzene in solid solutions in argon, krypton, xenon, nitrogen, carbon dioxide, and methane have been recorded. The spectra exhibit multiplet structure in most solid solution systems but consist of sharp single lines in solid nitrogen as solvent. Experiments were carried out at liquid helium (4.2”K) and liquid hydrogen (20.4”K) temperatures in order to determine the physical origin of the multiplets. It was concluded that the multiplet components arise from solute molecules having sharply defined but different environments in the host lattices. The possibility of the coexistence of two crystalline phases, cubic and hexagonal, of the host solid in the presence of the benzene impurity is discussed as a possible cause for the observations. I. INTRODUCTION The absorption and emission spectra of benzene in solid solution in t,he rare gas crystals in the 2600-A region have been briefly described in two separate and independent communications (1). The spectra exhibit multiplet structure with separations of 40 and 80 cm-l and the problem of the inOerpretation of these multiplets has remained unresolved. Further investigation of these spectra and the determination of the origin of the multiplet structure are the subject’ of the present, paper. Multiplet structure in the absorption and phosphorescence spectra of benzene in solid solution in cyclohexane has been previously reported and in this case the components were assigned as arising from benzene molecules embedded in different phases of the cyclohexane solid (9). This assignment, has subsequently been proved to be correct (3’). One remarkable feature of the benzene spectra in inert solid solutions is their sharpness in some solvents and more generally, the change of line widt,h with solvent. This property has been studied in the emission spectra of larger aromatic systems in solid solutions in paraffins by Shpolskii, who correlat’ed the degree of sharpness of the solute spectra wit’h the similarity in dimensions between t’he solute and solvent molecules (4).

* Fellow of the John Simon Guggenheim Memorial Foundation, address: Mellon Institute, Pittsburgh, Pennsylvania, 15213. 158

1961-1962;

Permanent

L-V SPECTRA

OF BENZENE

IN INERT

SOLID

SOLUTIONS

159

The spectrum of the benzene molecule in the near ultraviolet has been extensively studied and is among t.he most. thoroughly understood in t.he field of molecular spectroscopy. The molecule is planar and hexagonal (point group D,,). The absorption and fluorescence involve transitions between the totally symmet,ric ground stat,e ‘A Ig, and the first’ excited st,ate, lBp, (5). The phos1~horescencc at, 3300 w has been assigned to the transition 9B1,, + lAll, (0). II. EXPERIMENTAL

The exljerimental arrangement was very similar to that’ described previous13 for similar experiments (1, 7). The inner dewar of a glass double cryostat ended ~II a cold finger to which a copper plate was attached by means of a kovar-pyrcx seal. A copper window mount was screwed to the plat,e. A sapphire cold window was c~lamped t.o the window mount. and good thermal cont.act was assured by the use of indium gaslie&. The glass cryostat’s outer container ended in a ground flange which rested on an O-ring in the corresponding flange of a brass outc.1 “can.” The lat’ter vessel was fitt#ed with quartz windows and a gas inlet which was directed at a 45” angle to the optical axis-a line normal to the two outc>t windows and the cold inner window. A11 additional quartz window allowccl illuminntio~~ of the cold window for emission experiments. The mixture of benzene in one of the matrix gases (argon, krypton, SCILOII, Ilitrogcn, carbon dioxide, and methane) was prepared by standard gas-handling techiliques. Molar concent~ration ratios varying from 1:300 to 1: 10 000 wf’re used. I,icluid hydrogen and liquid helium were used as refrigerants. In ROI~(~cases n-hw the temperature was to be varied in the course of an experiment, the imlel demar and sample were cooled by t,he flow of cold helium gas from the liquid helium st’ornge dewar. For such experiments the transfer tube was so adjusted that it’s lower tip was locatted 5-10 cm above the liquid helium level and the flolr rate was rcgulat’ed by varying the pressure applied to the storage den-ar. In this way t hc teml)erature of the sapphirtl window could be kept quite constant :it WI< or at 8°K or at any intermediat,c value. The gas-1jhase mixture of benzene and matrix gas was allowed to flow into the v:tc’uunI chamber of t.he cryoslat through the inlet tube nnd it. deposited on the xaplJhircb colt1 window in the form of a solid film. The rate of growth of the filnl was det~ermined by the observation of interference colors on the window in natural light. The film thickness increased at the rate of 0.2-1.0 micron Iber minute, corresponding to pressures of 500-1500 microns in the flow tube before ml ering into the cryost,at. The absorption and emission spectra were recorded photographically. For recording the absorption spectrum R Hanovia high pressure, 150-W xenon arcs lnml) wa-: used and t-he spectrograph was a larfc Hilger quartz prism instrunlent. The ~pmtral region covered was 3350-3200 A and Ilford ordinary I)lates wercb used for most experiments. The slit widths varied from 20-100 microns and the

160

DIAMANT,

I

38

HEXTER,

AND

SCHNEPP

I

I

I

39

40

41

I

x LO%!

FIG. 1. Microdensitometer tracings of the absorption spectra 20°K and at 4.2 “K and in a solid mixture of Ar and NP at 20°K.

of benzene

in solid Ar at

exposure times were in the range 5 see-45 min. The emission was excited by a General Electric AH-6 high pressure mercury arc filtered by Kasha’s filter (8) combination “E” which passes the light of wavelengths 2300-2800 A. The emission spectrum was photographed on a quartz prism spectrograph constructed in this laboratory and similar to the small Hilger instrument, with slit widths of 20-50 microns and exposure t,imes of l-3 hr. Tracings of the plates mere prepared with a microphotometer. Spectral grade benzene was used without further purification. The rare gases argon, krypton, and xenon were of Linde manufacture. Carbon dioxide, methane, and nitrogen were C.P. grade Mat’heson Co. gases. The carbon dioxide was distilled once from a cold trap while t,he other gases were not further purified.

I;\’ Sl’li~:‘i’l:A 011’ BENZENE

Kr

I

I

38 FIG. 2. Microdensitometer CO2 2nd CiH4 at, 20°K.

I

I

33

40

IN INERT SOLID SOLUTIONS

I(il

2

I 41

I 42 X ldbcnr’,

tracings of the absorption spectra of benzene in solid Kr, Ne,

III. EXPERIMENTAL

RESULTS

The ~wults are presented in the form of microdensitometer tracings in J’ig. 1-4. Sote the mercury emission line in the source at, 2537 8 (39 417 cm-l). The analyses of the spectrum of benzene in a solid argon mat,rix at 20% ttre summxriz~d in Ta,bles I and II.

162

DIAMANT,

I 39

I 39

HESTER,

I 40

AND

I

I

41

SCHNEPP

I

42 x IO%!,

FIG. 3. Microdensitometer tracings of the absorption spectrum of benzene in solid Nz at 20°K.

A. ABSORPTION SPECTRA The present results for t#heabsorption spectrum of benzene in solid solutions are best described in terms of the known vapor phase analysis. The lowest, frequency intense band represents the transition from the zero vibrational state of the ground electronic state to one quantum of the ezs vibration which has a frequency of 516 cm-l in the excited electronic state. Superimposed on thisabsorption, which represents an electronically forbidden transition made allowed by vibrat’ional perturbation, there appears a progression of the totally symmetric vibration, v2 = 923 cm-‘. Other perturbing efg vibrat’ions are observed 1140 cm-l and 3070

U\’ SPECTRA

OF BENZENE

IN INERT

SOLID

SOLVTIONS

1 (ii3

Ar 20-K

Ar 4.2’K

I 24

I

I

25

I

I

26

I

,

27

,

,

26

,

,

29x Kfkfb

FIG. 4. Microdensitometer tracings of the phosphorescence Ar at 20°K and at, 4.2”K and in solid N? and CO, at 20°K.

spectra

of benzene

in solid

cm-l removed from the O-O frequency. The pure electronic transition (O-O) has been observed as a very weak band wit,h very thick deposits in all but one of t’he matrices used. A progression of the totally symmetric breathing frequency of 923 cm-l based on O-O has also been recorded for these cases. Solid argon is the except,ion not’ed above; only the barest, indicat,ion of the O-0 band has hecn detected in this matrix. The most important feature of the absorption spectrum, and the one characteristic of the solid solution system, is the fine structure of the individual bands. For benzene in argon, depositfed at 20’5, each band consists of three lines, the extreme lines of the triplet being sharp (half int(ensit,y widths 11 cm+ and 6 cni-‘, respect)ively) and t’he middle line is much weaker and broadened. This fine structure persisted unchanged when t,he molar concentra-

164

DIAMANT, Table

I

-

HEXTER,

AND

SCHNEPP

Analysis of the absorption spectrum of benzene in solid argon at 20°K.

No.

Wavelength 0 A

Analysis

Frequency Inv-vm -1 -1 cm tensity cm

Deviation -1 C"

1

2596.6

38500

516

C+516(ezg)

2

2594.1

38538

554

W4M-516

3

2591.1

38582

598

o+82+516

4

2588.3

38623

639

5

2584.7

38678

694

6

2567.4

38938

954

-12(31)

7

2562.8

39008

1024

-8

a

2554.1

39141

1157

0t1140(e2g); (H5a5(elg)x2

9

2549.4

39213

1229

0+82+1140; o+516+365(b2g)x2

7;-17

10

2543.4

39306

1322

ot585+755(bzg)

-18

11

2535.9

39421

1437

o+51&923

-2

12

2533.3

39462

1478

ot40t516+923(W1480(e

13

2530.7

39502

1518

ot82+516+923

14

2527.9

39546

1562

o+a2+1480

15

2524.5

39599

1615

16

2520.8

39658

vvw

1674

0+51&t-585x2

17

2507.8

39860

w

1876

W516+225x2+923;

18

2503.4

39935

VVW

1951

o+516x2+923

19

2497.7

40026

VVW

2042

20

2495.4

40062

w

2078

0+114o+923; o+585x2+923

21

2490.7

40138

w

2154

o+516+365(b2g)x2+923 C+82+11413+923

22

2485.4

40222

VW

-2 0

2g))

-17;-13

.l (-2) -3 0

-12 (@+923x2)

-13(30) -4

15;-15 -15 9

2238

tion ratio was varied from 1: 300 to 1: 10 000. The spacings are 40 cm-l and 82 cm-l. When a deposit, which had been prepared at 20°K and which exhibited such a triplet structure, was subsequently cooled to helium temperature, the spectrum remained unchanged. On the other hand, when the deposit is prepared

UV SPECTRA Table I -

Wavelength

OF BENZENE

IN INEP,T

SOLID

16:

SOLCTIONS

(Contd.)

Analysis

Deviation -1 c;il

NO.

;

Frequency -1 cm

23

2477.9

40345

VS

2361

c-k516+923x2

24

2475.1

40391

w

2407

Ot_4Ot51G~923x2; (0+1480+923)

25

2472.9

40426

“S

2442

0+82+51&923x2

26

2470.0

40473

ww

2489

Ot82+148ot923

27

2467.2

40520

VVW

2536

28

2463.5

40580

“VW

2596

Ot516+585x2+923

-13

29

2453.3

40749

VW

2765

C+923n3

-4

30

2450.9

40788

“VW

2804

0+516+225x2+923x?

-8

31

2446.6

40860

w

2876

0+516x2+923x2

-2

32

2439.1

40987

“VW

3003

m114Ot923;

33

2435.2

41052

w

3068

0+3070(e

-I

@+585x2+923x2 )

5 ;A: -7 ii

171-13 ^

28 34

2430.0

41140

w

3156

0+82+3070

35

2425.4

41218

“VW

3234

o+240(e2")X2+923x3

36

2422.2

41272

VS

3288

Ot516+923x3

1

37

2419.8

41313

VW

3329

O+401_516+923x3

2

38

2417.4

41354

s

3370

C+32+516+923x3

1

39

2408.0

41516

VW

3532

0+516+585X2+923x2

0

40

2404.1

41583

VW

3599

516+3100(alg)

-17

41

2392.3

41787

“VW

3803

ot516x2+923x3

2

42

2385.1

41914

vvw

3930

43

2381.1

41984

w

4000

ot3070+923

7

4 -15

44

2377.3

42052

w

4068

Ot82+307ot923

45

2369.2

42196

w

4212

Ot516t923x4

4

46

2364.7

42275

w

4291

Ot82+51&t923x4

1

-7

at liquid helium temperature, the band structure is quite different, as seen in Fig. 1. The lowest energy component, which was the more intense one for the 20°K deposit, is now the predominant peak and the high energy component only manifests itself as a relatively weak shoulder.

166

DIAMANT,

HEXTER,

AND

TABLE

SCHNEPP

II

ANALYSIS OF THEPHOSPHORESCENCE SPECTRUM OF BENZENE IN SOLIDARGONAT 20% No.

Wave-0 length A

Frequency cm-l

Intensity

3377.4 3446.9

29 600 29 003

VW vvw

69 666

0 0 -

3497.3

28 585

VW

1084

3508.9 3519.6 3561.1 3573.1

28 28 28 27

s s s s

1178 1265 1596 lG90

0 (0 0 0 0 0 -

VW-

y

91 91 -

606(e&

(0 - 703(&J) 491 404 073 979

Deviation cm-’

Analysis

cm-1

91 - 992&J; 91 - 985(&)) 1178(e& 91 - 1178 1596(e*,) 91 - 1596

;

-22 -31 t-37) ii) 0 -4 3

The absorption spectrum of benzene in solid krypton at 20°K exhibits the same fine structure of the bands as in solid argon. In a carbon dioxide matrix a similar triplet structure was observed but in this case the more intense component is at higher frequency. It is of importance to note that for this matrix the multiplet structure is clearly observed also for the pure electronic transition as well as for two more members of the totally symmetric vibrational progression 0 + 923 err-‘. For xenon and methane matrices the absorptions are broad with some remnants of fine structure. The absorption spectrum in solid nitrogen, cnz the othe? hand, consists of narrow single lines. B. EMISSIONSPECTHA We were unsuccessful in recording the fluorescence spectrum of benzene in solid matrices at 20°K. The observed phosphorescence spectra are represented in Fig. 4. In solid argon the spectrum was sharp (half intensity width 10 err-r) but because of its low intensity and (because of) the small spectrograph used, the accuracy of the results is appreciably lower than those obtained in absorption. Frequency measurements are accurate to &5 cm-‘. The analysis of the phosphorescence spectrunr of benzene in solid argon is summarized in Table II. The lowest frequency emission line is assigned as O-91 cm-’ since its assignment as the pure electronic transition leads to great difficulties. The most intense lines belong to the two progressions 0- 1596 -992n and O91- 1596 - 922n. Another pair of analogous progressions is based on the perturbing 1178 cm-l eZovibration. The multiplet structure of the phosphorescence lines parallels that of the absorption spectrum at 20°K, but the spacing is somewhat larger. The middle component was not observed in the emission, probably because of the diminished sensitivit,y of the experiment. It is important to note that the lowest energy component of the multiplet is the more intense one here as in the absorption spectrum at the same temperature. It should also be noted that in Fig.

1‘V SPECTRA

OF BENZENE

IN INERT

SOLID SOLUTIONS

167

4 (top tracing) the line at 28 191 cm-l appears with greater intensit;y than the line at 28 304 cm-‘. This is due to a fault in the plat’e and the real relative itttensities of the two components in the progression 0-1178-992n is correctly represented by the pair of lines at 27 504 and 27 303 cn-l. The multiplet struct’ure in the phosphorescence of benzene in an argon matrix deposited at liquid helium teml)erature is again quite analogous to that of thti absorption spectrum of such a deposit. As in the absorption spectrum, the Ion-et, frequency component of the doublet, retains its predominan1 intensity in t lw phosl)horescence spectrum, whereas the higher frequency comltonent is gtwt 1~ reduced in intensity and appears as a weak shoulder only. The peak at 28 200 IX-~ is due to a fault, in the plate emulsion nrld is not significant. III nitrogen the lthosphorescettce sltectrum at 20°K is somewhatJ broader (20.30 cnrl half inlcnsity width) and no multjil)let strucktrc has been observed. This again parallcls the absorption spectrum for the smlc system. In solid meth:nw the I)hosI)horesence spectrum al, 20°K is again broader (10 ct+ half intensity width) and no fine slrurturc could be discerned. In :L carbon dioxide ma1 rix 21 20°K the I)hosphoresccnce of benzctte consists of lines of half intensity widths 30 cn-‘. :1 doublet structuw \vas observed for some of i hc lines with :L spacittg of 150 m-1. IV.

I)ISCUSSION

Both ihc absorption and ~)hosl)horrscettce spectra of bcnzenr in solid solutions observed in t,his work arc in agreement with previous obscrva.tions in the gns I)hase and in c*ondettsed phase (1). The analyses arc wholly compatible with Tulane, regular hexagonal struct#ures for both the ground and excited states. We shall be l)rincipafly concerned here with the origin of the multiplet struc*tuw observed in 1he absorption and phosphorescence hands nt, 2O”T<. The fornt:ttiott of dimers or aggregates as a possible source of 1his skucture is elimit~ated on t hc bask of the observation that, the multiplets pcwisted unchanged with cahanges itt c*ottcentration from molar ratios of 1 :300 to 1 :10 000.In another I:ossiblc ittt cqmtnt ion mulliplet components might he assigned as occurring as :I result of the removal of space degeneracy. Th(l vibrottic states arc of symmelry B,,, X (‘2”= lC1,,. However, such a splitting has heett ohservcd to amount to only 9 cn? itI the bwzcnc carystal (9) where interactions are expected to he stronger than itt t hc systems tnwstigated here. Moreover, this intwpretatiotr is excluded by t hcbot,servation of the multiplet s;trwture in I he nondegenerate f)urc clectrottic~ t lattsiCon in solid CO,. The inteqwetation of the doublet structure as a manifestation of :LII itttertral molccwlar degree of freedom is excluded by the occwrrettw of single lines in solitl nitrogen and by the absence of relative itttjensity changes on cooling IO 1°K of ;I sample prepared at 20°K. In view of the absence of the multiplet, st,ructurr it] solidttit.rogcw it must he concluded at this stage that the structure is caused b?

168

DIAMANT,

HEXTER,

4ND

SCHNEPP

environmental factors. This conclusion is supported by t,he change in t,he relative int’ensity of the component lines upon deposition at liquid helium temperature as coml’ared to liquid hydrogen t)emperature (k’ig. 1). It is known that solid argon deliosited from the vapor on a surface kept at 4°K consists of very small crystallites, or has a very faulty crystal structure as evidenced by the broad x-ray diffraction lines observed (10). On the other hand, when Ar is deposited on a surface cooled to 2O”K, good crystallit)es are obtained. The occurrence of external degrees of freedom or quant,ized librations of the benzene molecule in the solvent cage as a possible cause for the observation of the multiplet structure of t,he lines will now be considered briefly. This interpret,ation is excluded by two observations. First, as in the case of internal degrees of freedom, the absence of t,emperature effects after the sample has been prepared is evidence against it. Second, comparison between the absorption and phosphorescence of samples prepared at different temperatures shows that t,he lower energy component, is clearly the more int,ense for both spectra at 42°K. This behavior is not as expected if the interpretation under discussion were correct. The occurrence of multiple kapping sites of t,he benzene molecule in the solid solut,ions as the cause for the observation of multiplet structure of the spectral lines of benzene in all cases except for solid nitrogen is consistent with all observat,ions made in the present work. This interpretation is supported in particular by the predominance of the low energy component, in both absorption and emission spectra in solid argon at 4.2”1<, by the sensitivity of the relative intensities of the multiplet components to the deposition temperature, and by the insensitivity of the spectrum to temperature change after t,he sample has been prepared. It is also supported by the observation that the absorption spectrum in a solid solution consisting of a mixture of nitrogen and argon is different from that in each of the pure solids of the components (Fig. 1 and 3). It is t’herefore concluded t,hat the observed multiplet skucture is caused by the occurrence of multiple sites for the benzene molecule in the solid solutions. It seems to us that the conclusion which is forced upon us by the experimental observations is remarkable in several respects. The total shift of the spect)rum from the gas phase to the solid solution in argon is 110 cnr-I. The multiplet splitt,ing amounts of 80 to 90 cm-i and the absorption lines in argon have half intensity widths of only 5 cm-r. In other words, we have to conclude that the differences in interartion with the solvent molecules at two sites cause a difference in transition energy of 60 cm-l or 73 o/oof the difference in transition energy between gas phase and solid solution. Of course, these transition energy differences represent differences between quantities which may be considerably larger, i.e., the solvent perturbation energies of excited and ground states, and t’hereforc direct comparison between them is not entirely relevant’. Nevertheless, it is clear t’hnt the solvent perturbation energies are large compared to the observed line widths of the individual multiplet components. This fact. requires the conclusion t)hat, the

U\’ SPECTRA

OF BENZENE

IN INERT

SOLID

SOLTTIOXS

16’)

solute sites are very sharply defined. One model for multiple sites which would sat.isfy this requirement is based on the occurrence of vacancies in the wellordered solvent sphere surrounding the benzene molecule. Such a mode1 has been discussed in connection with the observation of multiplet structure in the :thsoq)t ion specka of atoms dispersed in solids as used here (11). In this case, howevc~r, removal of orbital degeneracy is the cause of the multiple&. Another possible interpretation of the observations reported here would be based on recent studies of the crystal structure of argon containing an impurity (12). It appear:: t hnt :wgo11 containing a small concentration of oxygen crystallizes in the hrsngotl:~l vlosc Iwked lattice. The two principal components of the benzene mult il)lct s would then be assigned to molecules embedded in two different latticesnnc hexagonal, and the other cubic:-close-pacl;ed.’ ACXSOWLEDGMEiXT This

research was partially supported by the rrlited States Pulnlic Health Service.

RECEIVED: July

26, 1965

f s. L~ac~r, paper presented at, the “Free Radicals Symposillm,” 8eptemher (1957); C;. WILSE ROBINSON, J. Mol. Spectry. 6, 58 (1961).

Washington,

I). (‘.,

2. H. SPONER, J. KAND~, ANDL. A. BLACKWELL, Spectrochim. Acta 16,1135 (1960) ; J. I). SPAXGLER AND H. SPONER, Spectmchim. Llcta 17,1298 (1961); 19,169 (1963). 3. S. LEXH iurlpublished). 4. E. T. S:H~OI,SKIIAND I,. A. KLISOVII, Soviet Phys.---Doklarly 1,782 (1956); 141.i.. SHPOLSKII AND It. 1. PERSOXOV, Opt. i Spektroskopiya 8, 172 (1960). 5. H. SIJONEIL, (:. NORDHEIM, A. I,. SI~LAI~,AND E. TELLER, J. Chem. Phys. 7, 207 11930); F. M. GAHFOII. K.4SEI.4, J. Opt. ,%c. ~1~2.38, 928 (1948). 9. \-. L. BROUDE, T-. R. MEDVEDEV, AND A. F. PIUI~HOT~O, Zh. Eksperim. i Tear. Fiz. 21, 665 (1951); Opt. i Spektroskopiya 2, 317 (1957); Soviet Phys. I’sp. 4, 584 (1962). IO. H. S. PEISEK, in. “Formatiorl arld Trapping of Free Radicals,” A. M. Bass :ttld H. 1’. Broida teds.). p. 318. Academic Press, New York, 1960. 11. &‘I. BHITI~ .AND 0. SCHNEPP, J. Chew. Phys. 39, 2714 (1963). 13. c’. S. B.41111~~~ASD L. MEYER, J. C’hem. Phys. 41, 1078 (1964); I,. ME~EI~. (‘. S. BARRETT, .~ND I’. H.412~~s, ./. Chem. Phys. 40, 2744 (1964). 13. T. H. JORIMN, H. W. %ww, W. Z. STMXB, AXD W. N. LIPSVOMB, .J. Chenl. Plr?l,q. 41, 756 (1964) ; H. H. IA.\xoo~~ AND R. BijKixsiTEIs. “Zahletlwerte llrld Futlktiollerl,” \-ol. I, Parr 1. ,?priogrr, Berlill, 1955. 1 All the host solids used ill this work crystallize

iti the rllhic~close-l,acked

lattice

(IO, 1.7).