Fibonacci quasicrystals

A Fibonacci quasicrystal is a one-dimensional pattern with long-range order but no repeating unit cell. It is not random: every local choice follows a precise global rule. It is not periodic either: no finite block, slid along the line, reproduces the whole pattern.

The familiar construction is almost absurdly small:

L → LS
S → L

Start with L and apply the substitution everywhere at once:

L
LS
LSL
LSLLS
LSLLSLSL
…

Read L and S as long and short tiles. In the geometric version, give the short tile length 1 and the long tile length the golden ratio, φ = (1 + √5) / 2 ≈ 1.618. As the sequence grows, the ratio of long tiles to short tiles settles on that same number.

The small rule and the large order

There is no dice roll hiding in the substitution. Order shows up at every scale: recognisable patches recur, the long/short frequencies settle, and the pattern has sharp diffraction peaks. Yet it has no translation period.

That middle ground is the point. A conventional crystal repeats. A random sequence has no exact organizing rule. The Fibonacci chain keeps coherence as it grows while denying us the one free gift of a periodic crystal: a unit cell you can stamp forever.

An irrational step around a circle

There is a simple picture behind that refusal to repeat. Walk the integer orbit

e^(2πinα),    n ∈ ℤ

on the unit circle.

If α is rational, you hit only finitely many points and then loop. If α is irrational, the orbit is dense: it comes arbitrarily close to every point on the circle, but it never lands exactly on a previous point. Equivalently, the sequence

0, α, 2α, 3α, …  (mod 1)

eventually closes for rational α, and never closes for irrational α.

For Fibonacci, take

α = φ = (1 + √5) / 2 ≈ 1.618.

Only the fractional part matters in a rotation modulo 1, so the same walk is

α = φ − 1 = 1 / φ ≈ 0.618.

Both are irrational. They differ by the integer 1, so after “mod 1” they describe the same rotation.

The rule that turns the rotation into tiles

Here is the missing concrete step. Set β = 1 / φ² ≈ 0.382 and start at 0 on a circle of circumference 1. Repeatedly advance by β, reducing modulo 1. Mark one point on the circle. On a step that crosses the mark, write S; otherwise write L.

Without a picture, the same rule is

aₙ = floor((n + 1)β) − floor(nβ),    n = 1, 2, 3, …

aₙ = 0  →  L
aₙ = 1  →  S

The subtraction asks a yes/no question: did the nth irrational step pass an integer? Equivalently, did the walk cross the mark? The output begins

L S L L S L S L L S …

which is the Fibonacci word above. An irrational step never returns to its starting phase, so the rule cannot settle into a repeating tile block. But it is not random: every tile is decided by the same rigid rotation.

Treat the symbols as intervals of two lengths, or put a point at every tile boundary, and you already have a one-dimensional quasicrystal model. A quasiperiodic function can represent density, scattering strength, or occupancy on those points; the quasicrystal is the arrangement the rule specifies, not merely the graph of the function.

Why the torus is the better picture

For a continuous real variable x, the single phase

e^(2πiαx)

just covers the circle for any non-zero α. To see quasiperiodicity, keep two circular phases at once:

x ↦ (e^(2πix), e^(2πiαx)).

That path lives on a two-dimensional torus — think of a doughnut surface, two independent angles. If α is rational, the path eventually closes into a loop. If α is irrational, it is dense on the torus and never closes.

Now take any function that is periodic on that torus and restrict it to the path. What you observe in one dimension is a quasiperiodic function: built from periodic ingredients, yet with no period in the signal itself.

That is the clean bridge to quasicrystals. The order is real — it comes from a periodic object in a higher-dimensional space — but the one-dimensional slice does not repeat, because its irrational orbit cannot return to its starting phase.

From Fibonacci counting to quantum information

The word Fibonacci also shows up in the theory of non-Abelian anyons. In the Fibonacci model, the nontrivial charge τ obeys the fusion rule

τ × τ = 1 + τ.

The plus sign is not ordinary addition. It lists the two allowed fusion outcomes. As you add more anyons, the number of compatible fusion histories grows as Fibonacci numbers: 1, 2, 3, 5, 8, …. A constrained long/short tiling history has the same kind of counting structure, which is why people compare the two spaces.

With three τ anyons whose total charge is τ, the first pair may fuse through one of two intermediate channels. Those two paths form a two-dimensional space and can serve as a logical qubit. Braiding non-Abelian anyons acts on such fusion-path states; the order of the braids matters.

This is a mathematical correspondence and a research proposal. It is not evidence that a Fibonacci tiling by itself is a fault-tolerant quantum computer.

A hypothetical diamond architecture

One possible hardware story uses diamond colour centres as addressable spin qubits in a two-dimensional network. In a hypothetical Fibonacci string-net architecture, selected spins would label the edges of a trivalent graph by 1 or τ. Local measurements would test fusion constraints at vertices and loop constraints around faces; logical information would live in global patterns across the network.

Diamond platforms have demonstrated useful ingredients, including logical encoding and stabilizer measurements. A full Fibonacci Turaev–Viro code in diamond has not. It remains a proposed architecture that would combine diamond spin control with non-Abelian topological-code operations — not a present device.