I'm not sure which anthraquinone you're talking about. But it it's 9,10-anthraquinone then the C=O keto groups are within the plane of the entire molecule, which is basically completely flat.
I'm not sure though what the three member ring...
There is definitely no 3-membered ring in the structure in your .log file. The aromatic substituents are not co-planar withe the anthracene ring system because of steric interference with the carbonyls, The slight twisting of the anthracene ring system is also probably just to relieve steric interference. The carbonyls of the anthraquinone are not tilted from the (approximate) ring plane. The substituent carbonyls, on the other hand, are probably almost free rotating and the conformation shown is very near a local minimum. (I think I answered more than you asked!)
Dear Bartosz Trzaskowski,i meant the anthracene ring in the 1,4-bis(4-formylphenyl)anthraquinone. Gaussian09 log file is attached.
Sorry, I didn't notice the file.
I agree with Allen: looks like a raesonable geometry. The two benzene rings with CHO groups are not in place with the rest of the molecule due to the steric hindrance caused by the C=O groups of anthracene. The slight twist of the whole molecule is also not a cause of concern.
Bartosz is correct - there is no concern with either the rotating of the CHO groups out of the phenyl plane or the overall slight twisting of the anthracene ring.
Why you did not employed a 6-31G(d,p) basis for opt? Only a question.
Dear Renjith!
Tilting of C=O group is may be due to occurrence of C=O...π interaction with phenyl rings.