Searching for realistic 4d string models with a Pati-Salam symmetry -- Orbifold grand unified theories from heterotic string compactification on a Z6 orbifold

Motivated by orbifold grand unified theories, we construct a class of three-family Pati-Salam models in a Z6 abelian symmetric orbifold with two discrete Wilson lines. These models have marked differences from previously-constructed three-family models in prime-order orbifolds. In the limit where one of the six compactified dimensions (which lies in a Z2 sub-orbifold) is large compared to the string length scale, our models reproduce the supersymmetry and gauge symmetry breaking pattern of 5d orbifold grand unified theories on an S1/Z2 orbicircle. We find a horizontal 2+1 splitting in the chiral matter spectra -- 2 families of matter are localized on the Z2 orbifold fixed points, and 1 family propagates in the 5d bulk -- and identify them as the first-two and third families. Remarkably, the first two families enjoy a non-abelian dihedral D4 family symmetry, due to the geometric setup of the compactified space. In all our models there are always some color triplets, i.e. (6,1,1) representations of the Pati-Salam group, survive orbifold projections. They could be utilized to spontaneously break the Pati-Salam symmetry to that of the Standard Model. One model, with a 5d E6 symmetry, may give rise to interesting low energy phenomenology. We study gauge coupling unification, allowed Yukawa couplings and some of their phenomenological consequences. The E6 model has a renormalizable Yukawa coupling only for the third family. It predicts a gauge-Yukawa unification relation at the 5d compactification scale, and is capable of generating reasonable quark/lepton masses and mixings. Potential problems are also addressed, they may point to the direction for refining our models.