Excited vs. Hot Ground State Reactivity in the Photochemistry of Alkylcyclobutenes in the Gas Phase and Solution

William J. Leigh and Bruce H.O. Cook

Department of Chemistry, McMaster University, 1280 Main Street West, Hamilton, ON Canada L8S 4M1

INTRODUCTION

Direct irradiation of alkylcyclobutenes in solution results in competing ring opening to the isomeric dienes and cycloreversion to yield an alkyne and alkene.  Ring opening is non-stereospecific, while formation of alkene via cycloreversion is stereospecific. For example, Equation 1 shows the products obtained from 214-nm photolysis of cis- and trans-1,2,3,4-tetramethylcyclobutene in pentane solution. The relative yields of each of the products are indicated:

Leigh, W.J.; Zheng, K.; Clark, K.B. Can. J. Chem. 1990, 68, 1988.
Leigh, W.J.; Postigo, J.A. J. Am. Chem. Soc. 1994, 117, 1688.

 

The two types of products have been explained in terms of two distinct mechanisms arising from different excited states.  Ring opening is thought to occur in the p,p* excited singlet state to produce the formally photochemically-allowed diene isomer as the initial product, but proceeds adiabatically to form the excited diene which decays to form a mixture of diene isomers.  Cycloreversion is believed to proceed mainly from an excited singlet state which in solution appears in roughly the same spectral position as the gas phase p,R(3s) Rydberg absorption. Both the Rydberg state (in the gas phase) and the "Rydberg-like" state (in solution) are at lower energies than the p,p* excited singlet state, at least in 1,2-dialkylcyclobutenes.

The ring opening mechanism has been experimentally investigated quite thoroughly by our group.  One of the best pieces of evidence in support of the adiabatic mechanism came from a study where the ratio of diene isomers formed upon irradiation is compared to that formed upon irradiation of the formally allowed diene isomer. Note that this particular diene is constrained to be in the s-cis conformation, the same as should be produced (initially) in the electrocyclic ring opening of cyclobutene:

Leigh, W.J.; Postigo, J.A.; Venneri, P.C. J. Am. Chem. Soc. 1995, 116, 1593.

We have continued to worry about the involvement of the "Rydberg-like" excited singlet state in solution phase cyclobutene photochemistry. While the photochemistry of aliphatic alkenes in solution has often been explained in terms of the involvement of the p,R(3s) Rydberg excited state, there remain serious questions as to whether pure Rydberg states can exist in condensed phases. Solution phase alkene photochemistry can just as conveniently be explained as due to the population of a valence (p,s*) state, which either underlies or is mixed with the Rydberg state in the gas phase. In order to probe the true involvement of this excited state in solution phase alkylcyclobutene photochemistry and compare its behavior to the gas phase p,R(3s) state, we have carried out a comparison of the gas and solution phase photochemistry of three compounds: 1,2-dimethylcyclobutene (DMCB) and cis- and trans-1,2,3,4-tetramethylcyclobutene (c- and t-TMCB, respectively).  The gas phase uv absorption spectra (1 atm, SF6) are shown below, with the p,p* and p,R(3s) absorptions assigned. The solution phase spectra lack the fine structure observed in the gas phase spectra, but tail out to similar wavelengths.

 

 

 

W.J. Leigh, K.C. Zheng, and K.B. Clark, Can. J. Chem. 68, 1988 (1990).

These spectra suggest that by irradiating at 214-nm or (especially) 228-nm, it should be possible to selectively populate the "pure" Rydberg state in the gas phase, and compare its behavior to the photochemistry observed upon irradiation at the same wavelengths in solution.

RESULTS

Photolyses were carried out in the gas phase (~250 torr cyclobutene diluted to 1 atm with SF6) and in deoxygenated isooctane or cyclohexane solution (0.05-M) using 193, 214, and 228-nm light sources.  The data obtained are summarized as product ratios (ring opening / cycloreversion; EE:EZ:ZZ diene distributions; photochemically allowed / thermally allowed diene isomers) in Table 1. The table also contains the photostationary state (pss) ratios for the dimethylhexadienes formed from c- and t-TMCB, determined in both the gas phase and solution with 228-nm excitation.

Table 1.  Product ratios from photolysis of DMCB, c-TMCB and t-TMCB in the gas phase and in hydrocarbon solution (25 oC). "RO/CR" refers to the ratio of ring-opening (all isomers) to cycloreversion products, "EE:EZ:ZZ" is the distribution of the three isomeric dienes from c- and t-TMCB, and "Therm/Photo" is the ratio of thermally-allowed to photochemically-allowed diene isomers from c- and t-TMCB.

   

Gas Phase

 

Hydrocarbon Solution

Cmpd.

lex (nm)

RO/CR

EE:EZ:ZZ

Therm / Photo

 

RO/CR

EE:EZ:ZZ

Therm / Photo

 

193

3.0

-

-

 

1.3

-

-

DMCB

214

2.4

-

-

 

0.92

-

-

 

228

2.0

-

-

 

0.79

-

-

 

193

1.6

1 : 2 : 0

2.0

 

1.9

1.6 : 1 : 0

0.63

c-TMCB

214

1.0

1 : 6 : 0

6.0

 

1.1

1 : 1.5 : 0

1.5

 

228

0.82

<1 : 99 : <1

90

 

0.69

1 : 5.2 : 0

5.2

 

193

1.6

6 : 2 : 1

4.0

 

1.5

5.2 : 8.5 : 1

0.74

t-TMCB

214

1.3

3.8 : 1 : 1

4.9

 

1.0

2 : 4.1 : 1

0.74

 

228

0.74

2.8 : 1.5 : 1

2.5

 

0.32

1 : 4.3 : 2.7

0.85

pss

228

-

1 : 7 : 2

-

 

-

1 : 7.1 : 4.4

-

The following conclusions can be made from the data:
the yields of ring opening relative to cycloreversion products are similar or higher in the gas phase than in solution
the yields of ring opening relative to cycloreversion products decrease with increasing excitation wavelength in both the gas and solution phases

This suggests that both the p,p* and the Rydberg states lead to ring opening in the gas phase. In solution, the lower energy ("Rydberg-like") state also contributes to ring opening, but to a lesser extent than in the gas phase.
irradiation of c-TMCB at 228-nm produces mainly E,Z-3,4-dimethyl-2,4-hexadiene, the thermally-allowed diene isomer, in both the gas phase and solution.  In the gas phase, this isomer is formed with >95% stereospecificity!
irradiation of t-TMCB at 228-nm produces all three diene isomers. In solution, the diene distribution is very close to the diene pss, suggesting that secondary diene photolysis is skewing the results in this case. Note that in the gas phase, the thermally-allowed (EE and ZZ) diene isomers are again the major ring opening products.

This indicates that excitation of the lower energy (Rydberg or Rydberg-like) state leads to ring opening to yield predominantly the thermally-allowed diene isomer(s).

In the gas phase, this can most likely be attributed to a hot ground state reaction, due to internal conversion to upper vibrational levels of the ground state of the cyclobutene. Ground state ring opening occurs because the activation barrier for thermal ring opening is too low (Ea = 30-35 kcal/mol) to be quenched by collisions with the buffer. This is common in gas phase photochemistry; e.g. irradiation of cyclohexene in the gas phase at near atmospheric pressure yields 1,3-butadiene and ethylene (by retro Diels-Alder addition). In this case, the thermal cycloreversion of cyclohexene is known to have Ea ~ 65 kcal/mol. This reaction is completely quenched in solution phase photolyses.

An alternate possibility is that conrotatory ring opening occurs from the Rydberg state as a true excited state reaction. Since formation of the Rydberg state involves "semi-ionization" (a valence electron is promoted to a large, spatially diffuse carbon 3s orbital), one might expect it to have radical cation character. Cyclobutene radical cations undergo ring opening in the gas phase, and theory indicates that conrotatory stereochemistry is preferred. However, ring opening of cyclobutene radical cations is also believed to require thermal activation, and for this reason, we do not believe that Rydberg state ring opening is a very likely possibility.

So what is happening in solution?

While bona fide hot ground state reactions in solution phase photochemistry are not common, the thermal ring opening of cyclobutene is a fairly low activation energy process, and so we believe that this possible mechanism must be taken seriously.  In any event, it is now clear that there is a second mechanism for the photochemical ring opening of alkylcyclobutenes, which may or may not contribute to the p,p* process that appears to dominate with shorter wavelength excitation. The possibility remains that p,p* ring opening proceeds stereospecifically (yielding "disrotatory diene(s)") and the other isomer(s) are formed due to this second pathway; if this is the case, then the 3-4 cases in which the adiabatic p,p* mechanism explains the results perfectly must be serendipitous. These compounds need to be reinvestigated in the gas phase with longer wavelength excitation.

Allowing for these uncertainties, the photochemistry of alkylcyclobutenes is summarized below:

 

Future Research

Further investigations of the extent of the contribution of this new mechanism to cyclobutene ring opening are in progress.  These include study of:
gas and solution phase 228-nm photochemistry of bicyclic cyclobutenes (A) whose thermal ring opening proceed with higher activation energies than those studied here.
steric effects on long wavelength (monocyclic) cyclobutene ring opening (compounds B; see Leigh & Postigo, JACS 117, 1688 (1995)).
gas and solution phase 228-nm photochemistry of bicyclic cyclobutenes (C) whose behavior are consistent with the adiabatic p,p* ring opening mechanism.

 

 

Acknowledgements

Dr. William J. Leigh        Nick Toltl         Ed Lathioor

Tracy Morkin                 Xien Fu Zhang    Tom Owens