Fermion-qubit fault-tolerant quantum computing

Abstract: Simulating the dynamics of electrons and other fermionic particles in quantum chemistry, material science, and high-energy physics is one of the most promising applications of fault-tolerant quantum computers. However, the overhead in mapping time evolution under fermionic Hamiltonians to qubit gates renders this endeavor challenging. We introduce fermion-qubit fault-tolerant quantum computing, a framework which removes this overhead altogether. Using native fermionic operations we first construct a repetition code which corrects phase errors only. We then engineer a fermionic color code which corrects for both phase and loss errors. We show how to realize a universal fermionic gate set in this code, including transversal Clifford gates. Interfacing with qubit color codes we realize qubit-fermion fault-tolerant computation, which allows for qubit-controlled fermionic time evolution, a crucial subroutine in state-of-the-art quantum algorithms for simulating fermions. We show how our framework can be implemented in neutral atoms, overcoming the apparent inability of neutral atoms to implement non-number-conserving gates by introducing a neutral-atom braiding gate using photodissociation of bosonic molecules. As an application, we consider the fermionic fast Fourier transform, an important subroutine for simulating crystalline materials, finding an exponential improvement in circuit depth from $\mathcal{O}(N)$ to $\mathcal{O}(\log(N))$ with respect to lattice site number $N$ and a linear improvement from $\mathcal{O}(N^2)$ to $\mathcal{O}(N\log(N))$ in Clifford gate complexity compared to state-of-the-art qubit-only approaches. Our work opens the door to fermion-qubit fault-tolerant quantum computation in platforms with native fermions such as neutral atoms, quantum dots and donors in silicon, with applications in quantum chemistry, material science, and high-energy physics.