Well, here's another suggestion that could constitute prior art.
The authors of the cited paper make a point about discussing Oswald ripening. I've used Oswald ripening in the past in order to grow crystals such that I could meet a customer specification of a particular crystal size. This was done on a tonne scale in batches. The technique consists of forming a slurry of crystals in a suitable solvent and passing the mixture through a number of heating and cooling cycles. The difference between the maximum and minimum temperatures is usually small (eg, 10 °C); as the mixture heats up, some of the crystals dissolve, the smallest ones, as the mixture cools back, the size of the already present crystals increase as another layer of crystals form on their surfaces.
My idea in the previous post didn't allow for any Oswald ripening; concentrating, instead on the attrition effect that was mentioned.
In order to do the necessary Oswald ripening I would use an oscillatory flow reactor, where each of the portions of the reactor between each baffle had a temperature difference (eg, of 10°C).
The technique would run as follows: a slurry consisting of solvent (with racemic product), solid racemic conglomerate mixture, racemization agent and enantiomeric seed would be charged into the oscillatory flow reactor. It would be anticipated than any necessary attrition would come from the undidssolved crystals. If the first chamber was a cooling chamber, a greater amount of enantiomer would exist in the resulting slurry due to entrainment (see earlier post), entering the second (warmer) chamber would necessarily dissolve the racemic mixture due to its greater solubility properties (recall that this process is for conglomerates). Meanwhile, throughout these processes the racemization agent would ensure that the solution was a racemic mixture. The process would be repeated as it went through the next set of chambers in the reactor, and so on, until homochiral product was formed which could be collected by filtration.
Sand could even be added to the mixture in order to effect attrition, if necessary. The resulting enantiomeric product could be separated in a separate, dissolution filtration, crystallisation step.