Social Network is omnipresent in today’s world .Co-authorship network is a subset of Social Network. Interactions in Co-authorship network are based on common interests or similar profiles. The formation of groups is unavoidable in this setting. The identification of these groups in this setting may help in harnessing the social dynamics that exist in the institution. This will in turn assist in understanding the informal research groups evolution in an institution. It should also taken to consideration that Co-authorship networks usually evolve over time. It is a complex task to efficiently identify k-cliques from dynamic social networks. To address this challenge, this paper proposes an efficient k-clique detection method based on Temporal Concept Analysis (TCA). Experimental results illustrate that the proposed detection method is efficient for extracting the k-cliques from the Co-authorship networks.