Partial Trace: First Encounter
The partial trace removes one subsystem’s indices from a joint density operator, leaving the state that predicts all measurements on the remaining subsystem. We work through the basic rule here; partial trace gives the basis-independence proof, block-matrix methods, and algorithms.
Helpful background. Reduced states explains why this operation is useful; bra–ket notation makes the calculation short.
Tracing One Factor of a Dyad
Section titled “Tracing One Factor of a Dyad”Tracing out means taking the ordinary trace only on its factor. Thus
The order in the overlap matters for complex vectors. Apply this rule linearly to all terms in . If the input is a density operator, the result is positive and has unit trace; no extra normalization factor is needed.
Bell-State Calculation
Section titled “Bell-State Calculation”Expand the projector onto . Its first diagonal term reduces as
One cross term gives
The other two terms give and zero. Retaining the common factor ,
The cross terms vanish because the two states are orthogonal. Partial trace does not delete every off-diagonal term indiscriminately.
Quick Check
Section titled “Quick Check”For the product state , what is ? Does its coherence survive?
Solution
The joint density operator is . Its partial trace is , because . The coherence survives: both terms carry the same ket, whose overlap is one, unlike the Bell example.
References
Section titled “References”- M. A. Nielsen and I. L. Chuang, Quantum Computation and Quantum Information, 10th anniversary ed., Cambridge University Press, 2010, sec. 2.4.3.
- J. Watrous, The Theory of Quantum Information, Cambridge University Press, 2018, ch. 2. Author’s manuscript.