Part II · Two Qubits — 6b

Mixed States

So far every qubit has been a sharp arrow on the Bloch surface. But what if you’re also just unsure which one you were handed? That ordinary ignorance, stacked on top of the quantum kind, pulls the state inside the ball — and needs a richer description than a single arrow: the density matrix.

↩ before you start · keep these handy
·From Ch. 4: a pure qubit is a point on the Bloch surface — arrow length r = 1.
·From Ch. 2: |+⟩ = (|0⟩+|1⟩)/√2 is a definite state, not "we don't know which."
·From 0.4: a 2×2 matrix is a grid; the diagonal runs top-left to bottom-right.
🔑 symbol decoder · every new mark, in plain words
ρ"rho" — the density matrix, the richer description that also handles plain ignorance. rthe Bloch arrow's length: 1 = pure (on the rim), 0 = maximally mixed (dead center). Tr(ρ)the trace — add the diagonal entries; it's the total probability, always 1. Tr(ρ²)the purity: 1 for a pure state, ½ for a fully mixed one. I, σthe identity matrix and the Pauli set (X,Y,Z); r·σ packs the arrow into ρ. coherencethe off-diagonal entries — the live phase a classical mixture has lost.
feel

Two different kinds of “don’t know”

A qubit in superposition |+⟩ is in a definite state — you know exactly what it is, you just can’t predict a single measurement. That’s the quantum kind of not-knowing, and it’s already on the surface of the ball.

Now imagine a bag of qubits: half were prepared as |0⟩, half as |1⟩, and someone hands you one without saying which. That’s plain classical ignorance — a coin flip about which pure state you hold. Pile it on top of the quantum uncertainty and the state is no longer one arrow; it’s a blur of arrows. The blur sinks toward the center of the ball.

📷 everyday picture

Two blurry photos. One is a long-exposure of a spinning propeller — every blade really is there at once; the blur is true, complete information about the motion. The other is a stack of ordinary snapshots printed on top of each other — the blur is just you not knowing which single photo you hold. Superposition is the first kind of blur; a classical mixture is the second. The density matrix is the one ledger that keeps them apart.

recapSuperposition is sharp uncertainty on the surface; plain ignorance is a separate fog that sinks the state inward.
play

Sink the arrow into the ball

Here’s a slice through the Bloch ball. Drag the point. Its direction is still which qubit; but now its distance from the center — the length r — is how pure it is. On the rim: a crisp pure state. At the murky center: a total coin flip, all certainty gone. Watch the density matrix ρ rewrite itself as you drag.

▸ Bloch ball · sliceρ = ½(I + r·σ)
|0⟩ |1⟩ |+⟩ |−⟩ mixed
ρ = [
{{ r00 }}{{ r01 }} {{ r10 }}{{ r11 }}
]
trace{{ trace }}
P(0) / P(1){{ p0 }} / {{ p1 }}
coherence{{ coh }}
purity Tr(ρ²){{ purity }}
0.5 mixedpure 1.0
{{ tag }}
recapThe arrow's length r is its purity — on the rim it's a pure state, at the center a total coin flip.
math

One matrix carries both kinds of doubt

A pure state needs a vector; a mixed one needs a matrix. Build ρ by averaging the pure pieces you might have — each |ψ⟩⟨ψ| weighted by how likely it is. Its diagonal holds the outcome probabilities; its off-diagonal holds the surviving quantum phase.

ρ = ∑i pii⟩⟨ψi| = ½(I + r·σ)
always:  Tr(ρ) = 1  ·  purity  Tr(ρ²) = ½(1 + r²)
pure  r = 1 ⇒ Tr(ρ²) = 1  ·  fully mixed  r = 0 ⇒ Tr(ρ²) = ½

Purity is just the squared length of the Bloch arrow, rescaled. Drag to the rim and Tr(ρ²) reads 1; drag to the center and it bottoms out at ½ — the most ignorant a qubit can be.

✎ worked example · build ρ for a half-tilted arrow (r_z = 0.5)
1.ρ = ½(I + r·σ) = ½[[1+r_z, r_x],[r_x, 1−r_z]] with r_x = 0, r_z = 0.5
2.= ½[[1.5, 0],[0, 0.5]] = [[0.75, 0],[0, 0.25]]
3.read the diagonal: P(0) = 0.75, P(1) = 0.25
4.purity = ½(1 + 0.5²) = ½(1.25) = 0.625 — off the rim, so mixed.
recapρ = ½(I + r·σ): the diagonal gives outcome odds, the off-diagonal gives coherence, and purity = ½(1 + r²).
⚠ common misconception

“A mixed state is just another name for a superposition.” They couldn’t be more different. Superposition |+⟩ and the 50/50 mixture of |0⟩ and |1⟩ give the same 50/50 odds in the up/down basis — yet one is on the surface and one is dead center.

The tell is in ρ. |+⟩ has fat off-diagonal entries — live phase, ready to interfere — while the mixture is purely diagonal, those terms zeroed out. Measure in the |+⟩/|−⟩ basis and the difference screams: the superposition is certain, the mixture is still a coin. Coherence is the thing a mixture has lost.

✓ you can now
write a qubit's density matrix ρ and read its diagonal as outcome probabilities
tell a pure state (r = 1, purity 1) from a mixed one (r < 1) at a glance
explain why |+⟩ and a 50/50 mixture differ despite identical up/down odds
← 6a Bell’s Inequality next · 07 Teleportation