Activation Parameters in the Photocycloaddition of Phenanthrene to

Caldwell, Creed / Photocycloaddition of Phenanthrene to Dimethyl Fumarate. 6961. Table 1. Quantum Yields" of Photoprocesses Originating in 'E. T , O C...
1 downloads 0 Views 803KB Size
Journal of the American Chemical Society

6960

/

101.23

/ Nouember 7, 1979

Activation Parameters in the Photocycloaddition of Phenanthrene to Dimethyl Fumarate R. A. Caldwell* and D. Creed' Contribution from the Department of Chemistry, The Uniuersity of Texas at Dallas, Richardson, Texas 75080. Receiued March 12, I979

+

Abstract: The [2 21 direct photocycloaddition of phenanthrene to dimethyl fumarate was studied a s a function of temperature in the range 21 -65 "C. From temperature dependencies of the quantum yields for the various photoprocesses, activation parameters were derived for the processes originating from the singlet exciplex I E. The cycloaddition and internal conversion processes are activated, while intersystem crossing is not. Fluorescence was assumed to be temperature independent. The benzophenone-sensitized reaction i s not sensitive to temperature in this range. Cycloaddition enthalpies of activation are found to be larger and entropies of activation less negative for I E when compared to previous results for cyanophenanthrene-styrene exciplexes. The difference is attributed to a combination of increased charge transfer and less favorable remote orbital overlap in 'E.

Introduction decreases its (chemical) reactivity. Second, remote orbital overlap in the former, expected to be important in the cyThe use of temperature dependency of reaction rate to study cloaddition and internal conversion processes, may lower the the mechanisms of thermal reactions via analysis of activation energy barrier, but probably at the cost of a more negative parameters is ubiquitous. For photochemical reactions, howAS+. ever, it is less common, in part because the relevant elementary rate constants are generally not easy to obtain and in part beExperimental Section cause temperature sensitivities might have been expected to Analytical procedures, product structures, and quantum-yield be rather small in view of the necessarily high reactivities of techniques have all been previously reported.I9 For studies a t variable photochemical intermediates. temperature, both fluorescence spectra and quantum-yield meaSome photochemical reactions which have been studied as burements were taken utilizing the thermostated cell compartment a function of temperature have indeed shown quite small enof the Farrand MK- I spectrofluorometer. Temperatures were meaergy barriers to reaction, e.g., trans-stilbene i s o m e r i ~ a t i o n ~ - ~sured with an iron-constantan thermocouple and are accurate to fO.l ( 1 -4 kcal/mol), the addition of alkenes to enone9 (ca. 1 "C. kcal/mol), and the [4 41 cyclodimerization of anthracenes For quantum-yield measurements a quartz plate was placed at an angle of about 45" to the light beam incident on the thermostated cell (ca. 1.5 kcal/mol).6-y Others are somewhat larger: an intraholder, diverting about 7% of the beam into an insulated 1-cm cuvette. molecular enone annelation'O (4.7 kcal/mol), intermolecular The ratio, R , of light entering the sample holder to light entering the hydrogen abstraction of acetophenone' I (3.3 kcal/mol), and cuvette was then determined by ferrioxalate actinometry at ambient Norrish I1 reactions of aryl (3.5-6.7 kcal/mol) tcmperature before each run. During a run the solution was mainand 4-methyI-2-pentan0ne'~(7.9 kcal/mol from the triplet). tained in the sample holder at the desired temperature (20 65 "C) Structural isomerizations which presumably go through quite using a Haake circulating bath while the actinometer solution restrained transition states can have even higher activation mained at ambient temperature (23-25 "C). In this way the light dose energies: the rearrangement of 4,4-diphenylcyclohexenone' for each sample could be determined from R and the actinometer (10.5 kcal/mol for major product, 11.3 for minor) and the reading. conversion of o-xylene to m-I6 (7.7 kcal/mol) have been Results studied. The role of exciplexes in cycloaddition reactions has become The photoreaction of phenanthrene (P) with dimethyl fuquite clear over the last few We have undertaken marate (F)Iyproceeds through both a singlet exciplex ('E) and to study a few examples of [ 2 21 cycloadditions involving a triplet exciplex (3E), affording as products the two isomeric exciplexes carefully in order to learn as much as possible about cyclobutane diester [2 21 adducts C and T, an oxetane X arising from addition of the phenanthrene 9,lO bond to an ester the excited-state potential energy surface. The photoreaction of phenanthrene with dimethyl fumarate in b e n ~ e n ise an ~ ~ ~carbonyl ~ ~ of F, and isomerization of F todimethyl maleate (M). interesting model on two counts. First, the involvement of a A triplet 1,4-biradical 3SS formed essentially quantitatively from 3E is the predominant precursor of C and T, although a singlet exciplex in the reaction has been prover^.'^^^^ Second, small amount of T arises as a product of stereospecific collapse a variety of processes originate from the exciplex, leading to of 'E. Both ' E and 3SS are precursors of M. Scheme I sumthe chance to study several different photochemical and phomarizes this mechanism. tophysical processes. Elementary rate constants for these processes a t room temperature are known.Iy W e here report Quantum Yields. The quantum yields of the various proa study of their temperature dependence. Cycloaddition and cesses are given in Table 1 for the temperature range 21-64 OC. internal conversion from the singlet exciplex are activated and The concentration of F was 0.2 M, which assures essentially show substantial negative entropies of ktivation. Intersystem complete scavenging of ' P by F. Consequently, these values crossing shows little or no temperature dependence. Fluoresrefer to the chemistry of E exclusively. As the temperature cence is assumed to be temperature independent by analogy increases, the distribution of products at low conversion ( 5 2 % ) with related systems. W e compare the present activation pachanges significantly. Most striking are the increase in oxetane rameters with those we obtained in a previous the X and the decreases in yield of the cis diester C and the olefin isomer M. As Scheme I shows, X arises purely from 'E, while addition of cyanophenanthrenes to styrenes, in which the same C is purely a triplet product. At F = 0.2 M, the yield of M general features were observed. Differences in the two systems originating from 'E is much greater than that arising from 3P are attributed to a combination of two factors. First, increased or 3E, and consequently its decrease with increasing tempercharge transfer in the present case stabilizes the exciplex and

+

-

+

0002-7863/79/150l-6960$01 .OO/O

+

0 1979 American Chemical Society

Caldwell, Creed

/

6961

Photocycloaddition of Phenanthrene to Dimethyl Fumarate

Table 1. Quantum Yields" of Photoprocesses Originating in 'E T,O

6 F (TIEc)

6X

4.5 x 10-3 (1.2) 3.9 x 10-3 (1.041 3.3 x 10-3 (0.88) (0.69) 2.6 X

0.026 0.030 0.037 0.044

C

21.0 33.8 48.8 63.5

6T

7.0 x 6.2 x 6.0 x 5.7 x

10-3 10-3 10-3 10-3

6C

6M

6D

3.0 x 10-3 2.4 x 10-3 1.6 x 10-3 1.3 X

0.042 0.038 0.034 0.026

0.81 0.82 0.85 0.86

f 5 % except as noted. Since $JDis calculated as 1 - Z@(observables), its error is determined by errors in the observables, the sum total of which is a modest fraction of @D. The $JDerror is thus quite modest. We estimate f 1.5%. T I E ( T = ) @ F ( TX) 1.2 t 4.5 X I n nanoseconds.

Table 11. Fluorescence Quantum Yields and Lifetimes of the Phenanthrene- Fumaronitrile and 9-Cyanophenanthrene- Anisylisobutylene Exci~lexesa s Functions of TemDerature phenanthrene-fumaronitrile 7, nsb

T , OC

@Fa

20.4 34.6 49.6 64.6

0.049d 0.049d 0.049d 0.049d

18.6 19.3 16.9 16.3

T , "C

kF, S - I 2.6 X 2.5 X 2.9 X 3.0 X

IO6 IO6 IO6 IO6

23.0 35.6 50.0 64.8

9-cyanophenanthrene-anisylisobutylene @Fa 7,nsb kF, S-' 0.14 0.13 0.1 1 0.088

33.3 28.0 26.2 21.2

4.2 X IO6 4.6 X106 4.1 X I O 6 4.1 X IO6

s of temperature ( f 2 % ) over the range f 5 % . f l - 1 . 5 ns. Error estimated at &8%. d I n the present case, we show that 6 ~ i independent 16-65 " C . The quoted quantum yield is that reported p r e v i o ~ s l y . ~ ' Scheme I

3 F -+F+Ra

plest case among the elementary processes) as a function of temperature, via the simple equation

I

Because other photoproducts of 'E either involve subsequent intermediates (C, T, and M) or are derived from two different pathways, some further manipulation of the results is required. For M, presuming 3F to be the initial product from 'E, the decay ratio of 3F (i.e., its partitioning between M and F) must P + F +heat be known as a function of temperature. From benzophenonesensitized isomerization a t room temperature, we have deter3ss P+F+hVF mined it as 0.50, i.e., 5 0 5 0 partitioning between M and F. We have not measured the temperature dependence of our decay ratio directly; however, from a study of the benzophenoneC+T+P+F+M sensitized P-F isomerization and cycloaddition a t 65.4 "C, we ature must be explained on the basis of the chemistry of 'E. extrapolate a value of &=-M = 0.47 f 0.03. The few other Fluorescence quantum yields for ' E decrease as the temstudies of the effect of temperature on olefin triplet decay ratios perature increases, showing that either k~ varies inversely with also have shown that it is very slight.32 We therefore assign k3F temperature or the sum of all rate processes competing with = 2.04M/TIE. k~ increases with temperature, or both. Direct measurement It is necessary to know a t least something about the temof T I E as a function of temperature would settle the question, perature dependence of the triplet pathway to C and T in order since k~ = ( ~ F / Tbut , for such short lifetimes (1.2 ns at room to interpret @C and 4~ as a function of temperature. W e have temperature) is not feasible. studied the temperature dependence of the quantum yields of Fortunately, there is good evidence that radiative lifetimes the benzophenone-sensitized cycloaddition over this temperof exciplexes are essentially constant with t e r n p e r a t ~ r e . ~ ~ - ~ O ature range, and find that the total quantum yield @C 4~ is Ware has reported that kF for the I-naphthonitrile-l,2-di- nearly temperature invariant. Most importantly, the ratio methylcyclopentene exciple^*^ and for the anthracene@T/@c is constant within experimental error as 1.85 f 0.05, N,N-dimethylaniline exciple^*^ vary at most slightly with over the range 21-63 "C. Since C arises only from 3E, the yield temperature. W e have examined exciplexes longer lived than of 3E from 'E can be measured as ' E as a function of temperature, measuring both fluorescence yields and lifetimes, and find the same result. Table I1 presents results for the phenanthrene-fumaronitrile exciplex ( ~ 2 1 0=~ 19 ns) and the 9-cyanophenanthrene-P-anisylisowhere & ~ ( 2 1"C) is known.19 By inspection of Table I, then, butylene exciplex (T23oC = 33 ns). Here too 4~ decreases with it clearly follows that 4 3 E decreases with temperature. temperature, and T decreases almost in direct proportion, The temperature dependence of & is complicated in that demonstrating that k~ is essentially (&lo%) temperature inT arises by two pathways. The temperature dependence via the dependent in these systems (Table 11). The phenanthrene'E route, based on the temperature studies of the benzophefumaronitrile exciplex especially is obviously a good model for none-sensitized reaction above, is dT(via 3E) = 1.85+c(T). ' E , and we thus presume that kF for ' E is also temperature This follows, of course, from the constancy of the T/C ratio independent. We may thus conclude that activated alternative with temperature. Since the total yield of T is the summation processes decrease 4 ~ . of the yields of the 'E and 3E pathways, the yield via 'E can be The knowledge of 7 l E a t room t e r n p e r a t ~ r e and ' ~ C$F as a determined by subtraction of the 3Ederived part from the total, function of temperature permits calculation of T I E as a function i.e. of temperature, under the presumption of constant k ~and , thus also calculation of the elementary rate constant k x (the sim&-(via 'E) = & - 1 . 8 5 4 ~ 'P

F

1

+

6962

Journal of the American ChemicalSociety

c

4

/ 101:23 /

November 7, I979

was significantly less polar than the exciplex, a conclusion reached independently for a different system by Lewis andH 0 y 1 e . ~The ~ present system resembles the cyanophenanthrene-styrene one in that the reacting excited chromophore is a phenanthrene, yet differs from it in that the present one represents a much greater difference in the donor-acceptor nature of the reacting partners and a presumably much smaller opportunity for H O M O - H O M O and LUMO-LUMO interactions. It is instructive to compare activation properties in the two systems. Photophysical Processes from 'E. Fluorescence and Intersystem Crossing in Exciplexes. As discussed above, it now seems likely that fluorescence of exciplexes is nearly a temperature-independent kinetic process, not surprising if it is essentially vertical. The present results reinforce the conclusion, a t least for phenanthrenes and alkenes, that intersystem crossing of exciplexes is also negligibly temperature dependent. The ' E 3E process, k 3 ~ has , been recognized clearly i n previous work.14 The present results would suggest a small inverse temperature dependence, but this may be an artifact. The ' E P 3Fprocess. k 3 ~has , been p o ~ t u l a t e dwithout ~~ satisfactory supporting evidence as a reasonable hypothesis. W e now feel that the absence of an activation energy for k3F is much better i n accord with an intersystem crossing mechanism than with conceivable alternatives. If, for example, a I ,4-biradical intermediate were involved in the isomerization, the newly formed bond would require much closer approach of the two partners than is thought to obtain in the stable exciplex geometry,35and would thus presumably require activation energy. The magnitude of the fluorescence rate is quite similar in the two systems. The total intersystem crossing rate, 1.2 X IO8 s-l, is somewhat larger (factor of ca. 4) in the present case than in the cyanophenanthrene-styrene For the present, it seems best to focus on the rough similarity of the values rather than the small difference. The anthracene-diethylaniline exciplex shows a somewhat smaller ki,,, -3.3 X I O 6 s - I at 20 0C,30but again nearly of the same order of magnitude; as in our case, the temperature dependence of kist is small. We believe that it is justified to conclude that both fluorescence and intersystem crossing in exciplexes are temperature insensitive and structurally insensitive. A good rule of thumb is kF 106-107 s-I, with considerable documentati or^,'^-^'^^^ and kist lo6-los s-l, with less documentation. Assuming that aromatic exciplexes can be reasonably well represented by a picture of D+' and A-' entities with K systems overlaid, and a roughly constant interplanar spacing, one might expect roughly a structurally insensitive interaction between the two electrons and thus structually insensitive values for kF and kist. Photocycloaddition from 'E. The two photoadducts X and T arise directly from ' E with enthalpies of activation of 4.4 f 0.65 and 7.3 f 1 .O kcal/mol, respectively. It is clear that one must introduce energy barriers into the electronic hypersurface between singlet exciplex and cycloadduct, as previously disc ~ s s e d , and ~ ~ that . ~ ~the height of each barrier contributes to determining the yield of product to which it leads. The barrier to X is in the range of those for the cyanophenanthrene-styrene exciplexes, but rather toward the higher energy end of the series; the barrier to T is noticeably higher. It is apparent, as mentioned before," that increased C T in an exciplex should increase the activation enthalpy for cycloaddition, and ' E certainly should possess a higher degree of CT than the phenanthrene-styrene cases.27 We calculate E,,(D) - E,,d(A) - E(0,O) = -0.53 V (- 12.2 kcal/mol) for the process ' P F P+' F-., but only -0.28 V (-6.5 kcal/mol) for l 9 c P t-An 9CP-. t-An+' (9CP = 9-cyanophenanthrene, t-An = trans-anethole) from published redox value^.^'-^^ Also, the possibility of remote orbital overlap in the 9CP-. .t-An +

+

I

1

I

29

30

31

1

1

I

32

33

34

'0% Figure 1. Arrhenius plot for k~ ( X ) , (0).

k3F

(m), k 3 (O), ~ k x (O),and k~

The temperature dependence of @r(via 'E) is the greatest of any exciplex process in this system. The process accounts for only about 15% of total T a t 21 OC,but more than 55% a t 63.5 OC. The original evidence of Farid24 from triplet quenching studies that there is a stereospecific singlet component to T formation is strongly supported by our higher temperature results, in which the process is magnified. Rate Constants and Activation Parameters. As exemplified above for k x , elementary rate constants for process i originating from ' E can be obtained from the temperature-dependent quantities 4;(T) and TI,=(T). The analyses above therefore afford k x , k 3 ~k, 3 ~and , k r a t each temperature (see Scheme I). We take a temperature-independent k~ ofI9 2.5 X lo6 s-I. It is then possible to calculate kD( T ) , the rate constant for internal conversion of the exciplex, as a function of temperature from the equation kD(T) = T I E - '

-

k, I

where all other processes i must be included. All k's are precise to about lo%, except for the 21 "C value for kT, which is somewhat less precise. The rate constants are given in Table 111 and displayed graphically as an Arrhenius plot in Figure 1. Activation parameters AH* and AS* derived from the equations AH* = E , - R T and AS+ = R In (hA/ek( T ) ) = 4.60 log A - 60.95 are readily derived and are given in Table IV.

Discussion We have previously reported27a study of the temperature dependence of the reactions of substituted 9-cyanophenanthrenes with trans-p-methylstyrene and trans-anethole. In these cases, based on reaction from the (emissive) exciplex, small activation enthalpies (2-5 kcal/mol) and substantial negative activation entropies (-1 1 to -16 eu) were observed for cycloaddition, strikingly similar values were observed for internal conversion, and both fluorescence and intersystem crossing from the exciplex were found to be temperature independent. Substituent effects confirmed that the transition states for cycloaddition and internal conversion were certainly similar and quite conceivably identical. The transition state

+

- -

- -+ +

+

+

Caldwell, Creed

/

6963

Photocycloaddition of Phenanthrene to Dimethyl Fumarate

Table 111. Rate Constantsa for Processes Originating from ' E 21.0 33.8 48.8 63.5

9.16 X 1.54x 3.41 X 4.62 X

IO' 106

IO6 IO6

2.17 x 2.88 x 4.20x 6.35 x

107 107 107 107

5.49 x 5.07 x 3.99 x 4.12x

" Calculated as described in text from the quantum yields given in Table Table IV. Activation Parameters for Processes Originating in ' E A , s-I

6.0 X 10'' 1.0 X IO" k 3 ~ 4.0 X lo6 k3F 1.3 x lo8 kn 7.7 X I O ' O

kT

kx

E,, kcal/mol

AS*, eu

AH*, kcal/mol

7.3 f 1 7.9 f 1 -6.8 f 3 4.4 f 0.65 5.0 f 0.65 -10.3 f 2 -1.5 f 0.65 -30.6 f 2" -2.1 f 0.65a -0.2 f 0.65' 0.4 f 0.65 -23.6 i 2' 2.2 f 0.45 2.8 & 0.45 -10.9 f 1.5

a Calculated for comparison purposes only. AS* and AH* have no meaning for processes involving transitions between electronic surfaces.

transition state is greater, owing to the more extended conjugation and the especially good match of the orbital coefficients, as we have previously We have previously pointed out the low reactivity of the phenanthrene-fumaronitrile exciplex toward [2 21 cycloadduct f ~ r m a t i o n . This ~ ' cycloaddition proceeds with very low rate constant (