There is no fundamental dd diquark in that model. The "diquark" here is a scalar with a diquark coupling, not a QCD diquark. If you follow the references back, the 2007 paper cites a 1998 paper which cites a 1980 paper which talks about Higgses made of bound states of fermions. It doesn't call them diquark Higgses. Actually I can't parse the figure in that paper; it seems the ΔR,44 is the scalar with a VEV, then it has an interaction with three other scalars, and then they interact with quarks and induce a ΔB=2 transition. So the "diquarkness" might be hiding in that diagram somewhere. But that's the best I can do, in the search for a diquark Higgs which is a genuine QCD diquark. Another consideration is that QCD diquarks are not gauge invariant. A diquark condensate breaks the gauge symmetry, it's involved with phenomena like color-flavor locking and color superconductivity. I can imagine that such exotic phenomena play a role in the appearance of QCD scales in the Koide triplets, e.g. maybe they help to hide a second confining SU(3) interaction, as in the amended version of Koide 1981 that I proposed. But I do think a chiral condensate (qqbar, not qq) is a more plausible way to get EWSB. In sbootstrap language, diquark -> squark and meson -> slepton. There's a small literature on sneutrino Higgses, but I can't see anything at all about a "squark Higgs". (There is some stuff out there, about squark condensates and CFL in holographic QCD.) But this difference of opinion shouldn't be too much of a problem, the big picture probably involves both chiral condensates and diquark condensates and we'll have to understand both.