Studies towards Dibenzothiophene-S-oxide Arrays and their photochemical Reactivity

Thies Thiemann,^a* Kazuya Kumazoe,^b Kazuya Arima,^b Shuntaro Mataka^a

^aInstitute of Advanced Material Study and ^bGraduate School of Engineering Sciences, Kyushu University, 6-1, Kasuga-koh-en, Kasuga-shi, Fukuoka 816-8580, Japan
e-mail: [email protected] (for TT)

Received: 20 August 2001 / Uploaded 21 August 2001

Keywords: Dibenzothiophene-S-oxides, Photoextrusion, Deoxygenation

INTRODUCTION

It is known that dibenzothiophene-S-oxide can lose oxygen when photoirradiated.^[1-3] The products are the corresponding dibenzothiophenes. The mechanism of this deoxygenation and the nature of the released oxygen is still under discussion.^[4,5] The authors have become interested in this field as their original study of the photochemical behaviour of thiophene-S-oxide showed it to be very dependent on the substituent pattern of the thiophene-S-oxides^[6,7] and that more than one mechanism (pathway) can operate for one compound. Thus, 2,5-dimethylthiophene-S-oxide?1 gives hydroxymethylthiophene 2 and bisthienylether 3 upon photoirradiation (Scheme 1).^[6] At that time, this seemed to be an indication of a release of atomic oxygen from the molecule. When substituting the methyl groups for more sterically demanding substituents such as tert-butyl as in 4, the reaction followed a different pattern and the oxygen was introduced into the heterocyclic ring system ultimately giving the furan 5 as a product (Scheme 2).^[7]
 
 

Other compounds, such as 2,3,4,5-tetraphenylthiophene-S-oxide deoxygenate to give the corresponding thiophenes. This seemed to point towards the possibility of a bimolecular process in the photo-deoxygenation process, where large substituents at C-2 and/or C-5 would hinder the formation of an excited thiophene-S-oxide dimer and would induce the molecule to follow a different pathway.
 
 

As a similar discussion was ongoing in the case of dibenzothiophene-S-oxides (is it a monomolecular or a bimolecular process?), the authors decided to investigate whether a steric hinderance in the neighborhood of the sulfoxy-unit in the dibenzothiophene-S-oxides (i.e., at C-4 and/or C-6) or other changes within the molecule that would be detrimental to a dimer formation would cause a change in the photodeoxygenation behaviour of these compounds. Preliminary results are reported below.

RESULTS AND DISCUSSION

1. C-C-coupling reactions of Dibenzothiophenes.^[8] The pursued preparative route to dimers, trimers and tetramers of dibenzothiophenes joined at positions C-4 and C-6 necessitated the functionalisation of dibenzothiophene at C-4/C-6 with the aim of carrying out at a later stage a Suzuki-Kumada type coupling reaction between halodibenzothiophenes and dibenzothiopheneboronic acids. Lithiation of dibenzothiophene with n-BuLi in THF leads to the 4-lithiated compound which can be reacted with 1,2-diodoethane to 4-iododibenzothiophene 7 (yield 27%). Alternatively 7 can be synthesized in a 2-step procedure by lithiation of dibenzothiophene (n-BuLi/THF) and reaction with chlorotrimethylsilane (TMSCl) to yield 4-TMS-dibenzothiophene 9 (yield 91%). Subsequent TMS-halo exchange using BTMAICl₂/ZnCl₂^[9,10] in acetic acid (AcOH) gives 7 (yield 60%) Analogous TMS-halo exchange with BTMABr₃ in the presence of ZnCl₂ yields 4-bromodibenzothiophene 11 (yield 61%). Lithiation of dibenzothiophene with n-BuLi in THF in the presence of tetramethylethylenediamine (TMEDA) leads to 4,6-diiodobenzothiophene 8 (yield 26%).

In order to differentiate the C-4 and C-6 positions for multiple coupling reactions it was necessary to functionalize these positions with halo-groups of different reactivity in the Suzuki coupling reaction. For this purpose 4-TMS-dibenzothiophene was lithiated (n-BuLi/THF) and reacted with 1,2-diiodoethane to furnish 4-iodo-6-trimethylsilyldibenzothiophene 12 (yield 67%). Subsequently, 12 was subjected to BTMABr₃ (ZnCl₂/AcOH) to give 6-bromo-4-iododibenzothiophene 13 (yield 30%, Scheme 3).

When 4-TMS-dibenzothiophene 9 is lithiated (n-BuLi/THF) and reacted with trimethylborate 4-TMS-dibenzothiophene-6-boronic acid 14 is obtained (yield 67%, Scheme 4). Here, the trimethylsilyl group at C-4 is used as positional protective group. As the trimethylsilyl group can exchanged to a halo-substituent (see above) it can also function as a directing group for a second coupling reaction. When boronic acid 14 is subjected to Suzuki-Kumada coupling conditions^[11,12] (Pd[PPh₃]₄, Na₂CO₃, DME) the bis-trimethylsilyl-substituted dimer 15 is isolated in 58% yield (Scheme 4).
 
 

The trimethylsilyl groups in dimer 15 can be exchanged either to bromo-substituents (BTMABr₃, ZnCl₂, AcOH) to give dimer 16 (yield 97%) or to iodo-substituents (BTMAICl₂, ZnCl₂, AcOH) to give dimer 17 (yield 97%). When dimer 17 is reacted with 4-TMS-dibenzothiophene-6-boronic acid 14 under Suzuki conditions, the tetramer 6,6’-bis-(trimethylsilyl)-4,4’-bis(dibenzothiophene) 18 is formed, albeit in low yield (8%, Scheme 5). The low solubility of the symmetric dibenzothiophene-dimer 17 may be the reason (see below).

Trimer 18 can be obtained by Suzuki-Kumada coupling of 8 and 10 in 13% yield.
 
 

B. Preparation of the Dibenzothiophene-S-oxides. In general, there are various methods for the preparation of dibenzothiophene-S-oxides to dibenzothiophenes.^[4] In the following, all dibenzothiophene-S-oxides have been prepared by oxidation with peracids. 4,6-Dimethyldibenzothiophene, which has been prepared by a known method from dibenzothiophene, readily gave 4,6-dimethyldibenzothiophene-S-oxide upon reaction with meta-chloroperbenzoic acid (m-CPBA) in the presence of boron trifluoride etherate (BF₃^.Et₂O) (yield 92%). 4,6-Bistrimethylsilyldibenzothiophene, prepared from 4-trimethylsilyldibenzothiophene 9 [i.) n-BuLi, THF; ii.) TMSCl], was treated with m-CPBA to give the corresponding dibenzothiophene-S-oxide 22 (yield 82%, Scheme 6). When there is a trimethylsilyl group present in the substrate, often it is not advisable to use BF₃^.Et₂O in the reaction. For the most part BF₃^.Et₂O is used in these reactions to activate the peracid and, more importantly, to avoid further oxidation of the S-oxide to the S,S-dioxide by complexation of the Lewis acid to the sulfoxy moiety. The use of BF₃^.Et₂O in trimethylsilyl containing substrates, however, often leads to a cleavage of the trimethylsilyl group (see below).

When bis(6-trimethylsilyldibenzothiophene) 16 is reacted with m-CPBA in the presence of BF₃^.Et₂O the mono-TMS-bis(dibenzothiophene-S-oxide) 20 can be isolated (20% yield). Oxidation of 16 with m-CPBA in the absence of BF₃^.Et₂O leads to a mixture of products in the course of the reaction which terminates upon addition of further m-CPBA with the bis(6-TMS-dibenzothiophene-S,S-dioxide) (yield 46%). A better approach to bis(6-TMS-dibenzothiophene-S-oxide) may be the oxidation of the non-symmetric dimeric product 26 or a direct coupling of a TMS-dibenzothiophene-S-oxideboronic acid such as 25 with 4-iododibenzothiophene-S-oxide 23 (for the preparation, see below).
 
 

While dibenzothiophene-S-oxides are known to be reducible with a number of metals, it nevertheless seemed expedient to find an access to x-mers of dibenzothiophene-S-oxides by Pd(0) mediated coupling of halo-substituted dibenzothiophene-S-oxides and of dibenzothiophene-S-oxide boronic acids. 4-Iododibenzothiophene 7 could be oxidized readily with m-CPBA to 4-iododibenzothiophene-S-oxide 23. Oxidation of TMS-dibenzothiopheneboronic acid 10 with m-CPBA proved to be more difficult; also, the corresponding dibenzothiophene-S-oxide 25 could not be obtained in very high purity.
 
 

Coupling of 10 and 23 and of 12 and 25 produced the non-symmetric bis(dibenzothiophene)-S-monoxides 24 and 26 (Scheme 7), respectively, in sufficient quantities for initial photodeoxygenation experiments.

C. Photochemical Experiments with Dibenzothiophene-S-oxides.

4,6-Dimethyldibenzothiophene-S-oxide 27 (54mM) in CD₂Cl₂ was photoirradiated at 20°C for 12 h. The NMR spectral analysis of the reaction mixture showed 4,6-dimethyldibenzothiophene as the exclusive product. No hydroxylated compound could be observed. Upon purification by column chromatography on silica gel 4,6-dimethyldibenzothiophene 28 could be isolated in 99% yield. The reaction was monitored at 20 min., 1h, 2h and 3h. The experiment was repeated with 4,6-dimethyldibenzothiophene-S-oxide 27 at another concentration (4.3 mM in CD₂Cl₂). Also in this case, no hydroxylated products could be observed.

4,6-Bis(trimethylsilyl)dibenzothiophene-S-oxide 22 was photoirradiated in CD₂Cl₂ (c = 4.3 mM) at 20°C for 20h. Here, no photo-product could be obtained. The reaction was monitored at 5 min., 10 min., 20 min., 30 min., 1h, 2h, 3h and 6h. The NMR spectral analysis of the reaction mixture showed exclusively 4,6-bis(trimethylsilyl)dibenzo-thiophene-S-oxide 22.

Also compounds 20 and 26 were photoirradiated in CD₂Cl₂ at 20°C. The authors supposed that in the case of a possible bimolecular deoxygenation process any structural change that would be detrimental to the formation of a reactive dimer should retard the deoxygenation process.

Similarly substituted monomeric dibenzothiophene-S-oxides should form dimeric intermediates with greater ease than the more complex benzothiophene-S-oxide dimers such as 20 or 26.
 
 

While the photodeoxygenation of 20 was only complete after 14 h (c = 2.1 mM in CD₂Cl₂), the complete photodeoxygenation of 26 (c = 2.1 mM in CD₂Cl₂) only took 4 h. 24 has not yet been photoirradiated.

In recent years there has been a discussion of whether in the deoxygenation of dibenzothiophene-S-oxides atomic oxygen O(³P) is produced. While the behaviour of O(³P) is well-known in the gas phase,^[13,14] irrevocable evidence of O(³P) in solution chemistry has been scarce. It has been suggested, however, that O(³P) is formed when pyridine-N-oxide is irradiated at 308 nm in acetonitrile.^[15] The photodeoxygenation of dibenzothiophene-S-oxide has been used in the hydroxylation of interior sites within polyolefinic films.^[16] Nevertheless, all authors stress that it is not yet possible to absolutely prove the existance of O(³P) in solution.

In the above experiments the dibenzothiophene-S-oxides were chosen in such a way that their substituents would affect the formation of dimer intermediates due to their steric demand or that the formation of dimer intermediates would be affected by the shapes of the molecules. No optical filter was used in the photoirradiation Thus, the compounds were photoirradiated with light of a broad spectral range (l > 315 nm). All experiments were carried out in CD₂Cl₂. CD₂Cl₂ itself can also react with O(³P).

Nevertheless, the following observations were made from the above experiments: A.) 4,6-Dimethyldibenzothiophene-S-oxide 27 cleanly deoxygenates to 4,6-dimethyldibenzothiophene 28. No hydroxylated product can be found. If O(³P) is present in the reaction, then CD₂Cl₂ competes effectively for it with two well-placed methyl groups. B.) 4,6-Trimethylsilyldibenzothiophene-S-oxide 22 does not deoxygenate at all under the conditions. This can be an indication that in this case a reaction from an excited dimer is necessary for the deoxygenation to take place. C.) The dimeric dibenzothiophene-S-oxides on the whole react more slowly than either dibenzothiophene-S-oxide itself or 4,6-dimethyldibenzothiophene-S-oxide 27. The presence of a dibenzothiophene-unit in the dimeric S-oxide as in 26 enhances the deoxygenation of the compound.

At the moment the photoirradiation experiments are repeated in deuterated benzene and careful kinetic measurements of the deoxygenation process are taken. Furthermore kinetic experiments on the co-photoirradiation of dibenzothiophenes and dibenzothiophene-S-oxides are carried out.

EXPERIMENTAL SETUP

General. For the irradiation a Rikoh-Kagaku-Sangyo RIKO 1KW high-pressure mercury lamp was used. The lamp was in a glass-jacket cooled by water. The sample tubes (pyrex, glass thickness 0.37 mm) were immersed in methanol 7 cm from the lamp (non-rotating). During the photoirradiation the samples were measured by ¹H NMR (JEOL EX-270, JEOL JNM-LA 395, or JEOL JNM-LA 600) at the times given in the text. After the photoirradiation, the reaction mixtures were concentrated in vacuo and the residue was subjected to column chromatography on silica gel.

Typical Preparative Example. 20 (5 mg) in deaerated CD₂Cl₂ (1 mL) was photoirradiated for 19h at rt. A pyrex tube was used. Thereafter, the solvent was evaporated and the residue was subjected to column chromatography on silica gel (ethyl acetate/hexane 1:3) to give 28 (3 mg, 64%); ¹H NMR (600 MHz, CDCl₃) d 0.62 (s, 9H, SiMe3), 7.36 – 7.81 (m, 10H), 8.07 – 8.21 (m, 3H); MS (70 eV) m/z (%) 438 (100). HRMS Found: 438.0933. Calcd. for C₂₇H₂₂SiS₂: 438.0932.

REFERENCES

1. G. M. Gurria, G. H. Posner, J. Org. Chem. 1973, 38, 2419.
2. Z. Wan, W. S. Jenks, J. Am. Chem. Soc. 1995, 117, 2667.
3. D. D. Gregory, Z. Wan, W. S. Jenks, J. Am. Chem. Soc. 1997, 119, 94.
4. For a review on dibenzothiophene-S-oxides, see: T. Thiemann, K. Arima, S. Mataka, Rep. Inst. Adv. Mat. Study Kyushu Univ. 2000, 14(1), 37.
5. E. Lucien, A. Greer, J. Org. Chem. 2001, 66, 4576.
6. T. Thiemann, D. Ohira, K. Arima, T. Sawada, S. Mataka, F. Marken, R. G. Compton, S. D. Bull, S. G. Davies, J. Phys. Org. Chem. 2000, 13, 648.
7. See also: T. Thiemann, K. Arima, D. Ohira, K. Kumazoe, S. Mataka, Preparation and Photochemistry of Thiophene-S-oxides, ECSOC-4 (Sept. 1 - 30, 2000); http://reprints.net/ecsoc-4/a0090/a0090.htm; ECSOC-4 Proceedings (T. Wirth, C. E. Kappe, E. Felder, U. Diedrichsen, S.-K. Lin, eds., ISBN 3-906980-05-7).
8. T. Thiemann, K. Kumazoe, K. Arima, S. Mataka, Rep. Inst. Adv. Mat. Study Kyushu Univ. 2001, 15(1), 63.
9. For the preparation and the general reactivity and preparative utility of quaternary ammonium polyhalides, see: S. Kajigaeshi, T. Kakinami, T. Okamoto, S. Fujisaki, Halogenation and Oxidation with Quaternary Ammonium Polyhalides, Colony Press 1996.
10. Silyl-halo exchanges are also know to proceed with the following reagents: (for bromo) Br₂: [ref. 10a] C. Earborn, A. A. Najam, D. R. M. Walton, J. Chem. Soc., Perkin Trans. 1972, 2481; NaBr, NCS: [ref. 10b] D. S. Wilbur, K. W. Anderson, W. E. Stone, H. A. O’Brien, Jr., J. Label. Cpds. Radiopharm. 1982, 19, 1171; Br-/t-BuOCl: ref. 10b; (for iodo) I₂; ICl ref. 10a,b; NaI/NCS: ref. 10b, and D. S. Wilbur, W. E. Stone, K. W. Anderson, J. Org. Chem. 1983, 48, 1542; NaI/t-BuOCl: ref. 10b.
11. For similar conditions in Suzuki reactions, see: T. Thiemann, K. Umeno, D. Ohira, E. Inohae, T. Sawada, S. Mataka, New J. Chem. 1999, 23, 1067 and ref. cited.
12. A. Suzuki In Metal Catalyzed Cross-Coupling Reactions (F. Diederich, P. J. Stang, eds.), Wiley-VCH, Weinheim 1998, pp. 70.
13. R. J. Cvetanovic, J. Phys. Chem. Ref. Data 1987, 16, 261.
14. N. Taniguchi, K. Takahashi, Y. Matsumi, J. Phys. Chem. A 2000, 104, 8936 and ref. cited.
15. G. Bucher, J. C. Scaiano, J. Phys. Chem. 1994, 98, 12471 and ref. cited.
16. C. Wang, R. G. Weiss, Macromolecules 1999, 32, 7032.