I had heard some gossip about this from a couple of my colleagues last week, and now it would appear that the news has broken in the English-language media. The paper at hand is this one, which reports that, in one of the iron pnictide superconductors, there can be phase separation between regions of one composition (K0.68Fe1.78Se2) and regions of another composition (K0.81Fe1.6Se2). This is important because in trying to understand the superconducting properties of samples with some nominal composition, it's a big deal to know whether you're looking at a homogeneous system or one where some other composition is actually dominating the properties.
The accusation is reported here and here, and comes from Prof. Mu Wang at Nanjing University (also the home institution of the accused, Prof. Hai-Hu Wen). There are at least two scientific ethics issues. First, there is a question about co-authorship (did all of the authors on the paper actually contribute, and did they even see the paper prior to submission). Second, from what I can gather, there are concerns about data selection in Fig. 4 of the paper. People who know more about this, please feel free to chime in, since it's a good idea to understand specifically what the concerns are regarding the validity of the scientific claims of the article. The added dimension to all of this is the claim that both the accuser and the accused were up for membership in the Chinese Academy of Sciences (though apparently the accuser has withdrawn his candidacy - see my second link in this paragraph). Interesting, particularly since accusations like this in the physical sciences remain relatively rare.