Retraction of: Nature Communications 10.1038/ncomms6883, published online 19 December 2014.

During follow-up work to this Article, the authors discovered that its main claim—namely, that CO₂ was reduced to form oxalate—is incorrect. Rather, the detected oxalate is formed by the oxidation of ascorbate in the system, as discussed in a subsequent analysis¹. Although other results reported therein are valid, the authors wish to retract this Article.