Four homoallyl ortho-vinylaryl ketones (10a-d) -1,8dienes of sorts -were prepared by several approaches. In the presence of 1-2 mol-% Grubbs-II catalyst, they ring-closed to give 6,7-dihydrobenzocyclohepten-5-ones (11a-d) in 90-96 % yield. With SeO 2 the parent compound (11a) delivered benzocyclohepten-5-one (13a) and/or selenium-containing compounds (18)(19)(20)(21)(22) but no more than traces of 6,7-benzotropolone (5a). However, 5a was accessible from compound 11a via the sodium enolate and allowing it to react with a stream of oxygen [a] Scheme 1. Top: our previously established ring-closing olefin metathesis ("RCM")/ketal hydrolysis route to type-5 6,7-benzotropolones. [9] Underneath: regiocomplementary processings of a type-4 RCM product via the dibromide 7 and the bromoolefins 7 or iso-7 by cross-couplings and hydrolyses giving type-8 or type-iso-8 6,7-benzotropolones, respectively. [10] The Arican routes of Scheme 1 have a major drawback, though, namely the harsh conditions required for hydrolyzing a type-4 ketal in the last step: Liberating the enol required refluxing with 10 equiv. of tosic acid and 100 equiv. of water in acetonitrile for 1 h -2 d. [9,10] These conditions were applicable to certain type-9 or type-iso-9-ketals, as well ( Figure 2). When their substituent R was Et, Ph, CO 2 Me or CO 2 Et, such hydrolyses also proceeded well. [9] However, when their substituent R was CH=CH-CO 2 iPr (in 9a or iso-9a), HC=CH 2 (in 9c or iso-9c) or C≡C-SiMe 3 (in iso-9c) the substrate vanished under hydrolysis conditions without delivering any benzotropolone. [9,10] The C≡C-SiMe 3 -containing ketal 9b was an in-between-case: Its hydrolysis released a benzotropolone as expected but the sidechain R had been converted into C(=O)-CH 3 . [10] Eur.2930 Scheme 2. Can 6,7-benzotropolones 5 be reached via an RCM/oxidation route rather than via the RCM/hydrolysis route of Scheme 1? A 4-e 2 oxidation 11 → 5 would be required overall, but two 2-e 2 oxidations 11 → 12 and 12 → 5 might be used as an alternative. 5.We wondered whether step 2 might be disadvantaged vs. a dehydration delivering the benzotropone 13. While 13 looks like a dead-end at first, it might be re-routable towards 5. This is because unsubstituted benzotropone (13, all R = H) gave unsubstituted benzotropolone (5, all R = H) by endoperoxide formation and an ensuing reduction with thiourea. [11] Consequently, benzotropones 13 appeared as conceivable intermediates of our Scheme-2 strategy towards benzotropolones. A Modified Synthesis of 6,7-BenzotropoloneOur proof-of-principle benzotropolone synthesis by the approach of Scheme 1 delivered the parent compound 5a and is shown in Scheme 3 and Scheme 5. Scheme 3 advances to the RCM product and Scheme 5 supplements its oxidation. Scheme 3. Reaching the RCM product 11a, our synthetic precursor of unsubstituted benzotropolone (5a).Our synthesis began by a salt-free Wittig methylidenation of ortho-bromobenzaldehyde (14, Scheme 3). [12] The resulting ortho-bromostyrene (15) was treated sucessively with nBuLi and the...

The CsCO-mediated annulations ("Deslongchamps annulations") of three spirocyclic benzoquinone monoketals 5b-d with an ester or acyl substituent at C-2 to two tert-butyl esters of γ,δ-unsaturated β-ketocarboxyl acids ("Nazarov reagents"; 2a,b) were monitored H NMR spectroscopically. This revealed that a primary product, by all likelihood the Michael adduct, forms fast and prior to the appearance of the Deslongchamps adduct. These primary products form reversibly. This was proved by two crossover and four scavenging experiments. Therein, components already incorporated in one of the mentioned primary products ended up in Deslongchamps adducts different from the one, which would have resulted if the respective primary product had reacted alone. However, these experiments leave open whether our primary products are intermediates en route to Deslongchamps products or whether they represent dead ends.