Quantum Chemistry Encoding (Second Quantization to Qubits)

Second quantization reformulates quantum chemistry in terms of occupation number representation: each spin-orbital in the molecular basis gets a fermionic mode, and the electronic state is described by specifying which modes are occupied. Creation operators a†_i and annihilation operators a_i add and remove an electron from mode i, satisfying the canonical anticommutation relations {a_i, a†_j} = delta_ij. The molecular electronic Hamiltonian in second quantization is H = sum_ij h_ij a†_i a_j + (1/2) sum_ijkl g_ijkl a†_i a†_j a_k a_l, where the one-body integrals h_ij and two-body integrals g_ijkl are computed from the basis functions using classical quantum chemistry codes like PySCF or Psi4.

The Jordan-Wigner (JW) transformation maps each fermionic mode to a qubit by encoding the occupation number in the qubit’s computational basis state: mode i occupied maps to |1>_i, unoccupied to |0>_i. Creation and annihilation operators become a†i = (Z_0 x Z_1 x … x Z{i-1}) x ((X_i - iY_i)/2), where the string of Z operators enforces fermionic anticommutation. The Z-string length scales linearly with the mode index, so JW-encoded operators can have weight O(n) in the worst case. This is manageable for small molecules but becomes expensive for larger systems where many long-range interaction terms appear.

The Bravyi-Kitaev (BK) transformation achieves logarithmic locality by encoding occupation numbers in a binary tree structure. Instead of direct occupation, each qubit stores parity information about a subset of modes, chosen so that creation and annihilation operators involve at most O(log n) qubits. A related approach, the parity mapping, encodes the cumulative parity of occupations and allows two of the n qubits to be tapered off using molecular symmetries (particle number and spin), reducing the qubit count by 2. For near-term devices where qubit count and gate depth are both limited, BK and parity mappings offer concrete advantages over JW for systems with more than a few dozen spin-orbitals.

The qubit count for a full-valence calculation scales as roughly 2x the number of electrons in the active space (one qubit per spin-orbital), which reaches hundreds to thousands of qubits for pharmaceutically relevant molecules. Active space approximations reduce this by selecting only the orbitals most important to the chemistry, typically those near the frontier between occupied and virtual, and treating the rest with classical mean-field methods. A molecule requiring 500 active spin-orbitals needs 1,000 logical qubits after encoding, and fault-tolerant estimates place the physical qubit requirement in the millions for surface-code protected circuits, illustrating the gap between current hardware and practical quantum chemistry advantage.