Trichromatic Open Digraphs for Understanding Qubits

We introduce a trichromatic graphical calculus for quantum computing. The generators represent three complementary observables that are treated on equal footing, hence reflecting the symmetries of the Bloch sphere. We derive the Euler angle decomposition of the Hadamard gate within it as well as the so-called supplementary relationships, which are valid equations for qubits that were not derivable within Z/X-calculus of Coecke and Duncan. More specifically, we have: dichromatic Z/X-calculus + Euler angle decomposition of the Hadamard gate = trichromatic calculus.

💡 Research Summary

The paper “Trichromatic Open Digraphs for Understanding Qubits” proposes a new graphical calculus, called the RGB calculus, that extends the well‑known red‑green (RG) or Z/X calculus by incorporating a third complementary observable. The authors begin by recalling the categorical setting: symmetric monoidal †‑categories (SM†‑categories), the category of finite‑dimensional Hilbert spaces (FdHilb), its projective version (FdHilb_w_p), the subcategory generated by qubits (FdHilb_Q), and the stabilizer subcategory (Stab). Within this framework the RG calculus is formalised as a SM†‑category whose objects are tensor powers of a distinguished object * and whose morphisms are open digraphs built from red and green “spiders” (copy‑delete maps) together with phase gates taking values in the cyclic group C₄ = ℤ/4ℤ. The Hadamard gate H is the only colour‑changing primitive; it swaps red and green structures while preserving the compact structure. The RG relations (spider fusion, bialgebra, Hopf law, etc.) are listed, and it is shown that the interpretation functor J_RG : RG → Stab is a symmetric monoidal †‑functor. However, J_RG is not faithful: certain distinct RG diagrams (e.g., the supplementary rule diagrams) are mapped to the same stabilizer morphism. This non‑faithfulness reflects the known incompleteness of the Z/X calculus with respect to stabilizer quantum mechanics.

To overcome this limitation the authors introduce the RGB calculus. RGB retains the same basic graphical syntax (open digraphs, tensor product as horizontal juxtaposition, composition as vertical plugging) but now includes three colours—red, green, and blue—corresponding to the three mutually unbiased bases that can exist for a qubit. Crucially, there is no primitive Hadamard; instead two colour‑changing gates, denoted ⨂ and ⨁, are defined in terms of the existing generators. These gates allow any node of one colour to be expressed as a composition of nodes of the other two colours, thereby achieving full symmetry among the three observables.

The RGB relations (9)–(30) mirror those of RG but are extended to accommodate the third colour. Each colour’s spiders form a special commutative Frobenius algebra; pairs of colours satisfy bialgebra laws; and together they obey Hopf‑type equations. The authors prove that the hom‑set RGB(,) contains a faithful copy of the octahedral group O ≅ S₄, generated by three 90° rotations σ_r, σ_g, σ_b. Explicit group homomorphisms f : O → G and g : G → O are constructed, showing that the graphical rotations faithfully represent the rotational symmetries of the Bloch sphere. Moreover, the authors demonstrate that the RGB calculus is equivalent to the RG calculus enriched with the Euler‑angle decomposition of the Hadamard gate; this equivalence is essentially Van den Nest’s theorem for the dichromatic setting.

A †‑structure is defined on RGB by sending each generator to its adjoint (phase inversion, swapping inputs/outputs), preserving the SM†‑axioms. The interpretation functor J_RGB : RGB → Stab maps red, green, and blue spiders to the X, Z, and Y stabilizer structures respectively, while the colour‑changing gates correspond to the Hadamard and its variants. Consequently, every diagram in RGB has a well‑defined stabilizer semantics, and the functor is again a symmetric monoidal †‑functor.

The paper concludes by noting that RGB subsumes all known extensions of the Z/X calculus (such as the supplementary rule) in a natural, symmetric way, and that recent advances in the Quantomatic rewriting system make automated reasoning within RGB feasible. Nonetheless, the authors acknowledge that the question of whether RGB is complete for stabilizer quantum mechanics remains open, presenting an interesting direction for future work.