It's all very well predicting which forms of crystal an inorganic solid can adopt, but how can proof be obtained if these structures aren't thermodynamically stable? The answer is to build them up ...