L07: Hardware Platforms & the Materials Physics of Qubits

Learning Objectives¶

• Identify the physical two-level system, control knob, and readout for each major platform.

• Interpret $T_1$, $T_2$, gate fidelity, and connectivity as the noise channels of L06.

• Connect device limitations to materials problems: interfaces, defects, TLS, purity.

• Match algorithm requirements (depth, connectivity, shots) to platform choice.

What Makes a Good Qubit? (DiVincenzo Criteria)¶

• Well-defined, scalable two-level systems.

• Initialization to a fiducial state (e.g. $|0\cdots0\rangle$).

• Coherence times $\gg$ gate times ($T_2 / t_{\text{gate}} \sim 10^3$--10⁵ today).

• A universal gate set (L04).

• Qubit-specific measurement.

Eureka! The tension: qubits must couple strongly to our controls but weakly to everything else. Materials engineering lives in that contradiction.

Platform Overview¶

(Orders of magnitude; check vendor datasheets — numbers improve yearly.)

Superconducting Transmons — a Materials Story¶

• Josephson junction (Al/AlO$_x$/Al) provides the anharmonicity that isolates $|0\rangle,|1\rangle$.

• Dominant decoherence: two-level systems (TLS) in amorphous oxides at interfaces (substrate--metal, metal--air).

• Progress has come from materials substitutions: tantalum films, sapphire substrates, surface treatments $\Rightarrow$ $T_1$ from $\sim1~\mu$s (2005) to $>300~\mu$s.

Eureka! For a materials scientist, transmon $T_1$ is a defect-spectroscopy problem: the qubit is the most sensitive probe of its own dielectric losses.

Trapped Ions & Neutral Atoms¶

• Ions (Yb$^+$, Ca$^+$, Ba$^+$): identical-by-nature qubits; gates via shared motional modes; all-to-all connectivity within a chain; slower gates ($\mu$s--ms).

• Neutral atoms (Rb, Sr): optical-tweezer arrays of 100-1,000+ atoms; reconfigurable geometry $\Rightarrow$ natural lattice-model simulators for condensed matter.

Eureka! Analog quantum simulation of spin models on atom arrays is already producing materials-relevant phase-diagram data — a preview of L11.

Spins, Photons, and Topological Approaches¶

• Si spin qubits: CMOS-compatible; isotopic purification ($^{28}$Si) removes nuclear-spin noise — purity as a coherence resource.

• Photonics: room temperature, networking-native; loss is the enemy; measurement-based computing model.

• Topological (Majorana): protection from local noise by encoding in nonlocal states; materials challenge: clean superconductor-semiconductor interfaces.

Metrics $\leftrightarrow$ Noise Channels (Closing the Loop with L06)¶

• $T_1$ (energy relaxation) $\leftrightarrow$ amplitude damping channel, $\gamma = 1 - e^{-t/T_1}$.

• $T_2$ (dephasing) $\leftrightarrow$ phase damping; always $T_2 \le 2T_1$.

• Gate infidelity $\leftrightarrow$ depolarizing channel per layer; readout error $\leftrightarrow$ classical confusion matrix.

• Benchmarks: randomized benchmarking, quantum volume, CLOPS.

Eureka! Every datasheet number is a Kraus map in disguise. You can now simulate any advertised device with L06 machinery.

Choosing Hardware for a Job¶

• Deep circuits, modest qubit count (phase estimation): ions (fidelity, connectivity).

• Wide shallow variational circuits (VQE, L12): superconducting (speed, shots/s).

• Lattice-model simulation: neutral-atom arrays (geometry).

• Long-distance QKD (L16): photonics.

Mini-Lab¶

• Pull calibration data from a public cloud backend; convert $T_1$, $T_2$, gate times into Kraus channels and predict GHZ-state fidelity vs. measured.

• Estimate the maximum useful circuit depth for two real devices from their datasheets.

Takeaways¶

• No platform wins on all axes: coherence, speed, connectivity, scalability trade off.

• Qubit progress is materials progress: interfaces, defects, and purity set $T_1, T_2$.

• Hardware metrics are noise channels — the L06 formalism is your translation layer.