PHYS598500 · T2 Outlook · Week 2–3
From Artificial Atoms to Protected Qubits
Sequence, Interfaces, and Advanced-Learning Map for the Ten Questions
What Are These Two Weeks Answering Together?
W2 and W3 are not two unrelated introductions to qubits. W2 establishes how an artificial atom can exist; W3 uses the same language in reverse: once we know which operator the environment uses to create errors, can we design eigenstates, symmetries, and Hamiltonian terms that make those errors less likely?
T2 Overview of the Ten Questions
Five W2 Questions + Five W3 Questions
| No. | Week | Core Question | Capability Established by This Question |
|---|---|---|---|
| 1 | W2 | Why can a macroscopic circuit have quantum energy levels? | Identify collective circuit modes that can be quantized and isolated |
| 2 | What plays the roles of position and momentum in a superconducting circuit? | Use \((\Phi,Q)\) and \((\varphi,n)\) to describe quantum degrees of freedom | |
| 3 | Why is the simplest artificial atom still not a qubit? | Distinguish discrete energy levels from a selectively controllable two-level subspace | |
| 4 | How does a Josephson junction shape the energy levels of an artificial atom? | Explain anharmonicity through cosine nonlinearity | |
| 5 | Why is the artificial atom ultimately designed as a Transmon? | Use \(E_J/E_C\) to analyze the charge-dispersion/anharmonicity trade-off | |
| 6 | W3 | Under what conditions can an external perturbation cause a quantum transition? | Decompose decoherence into the noise spectrum and operator matrix elements |
| 7 | What quantum-state structures suppress noise while preserving control? | Distinguish sweet spots, selection rules, suppression, and controllability | |
| 8 | What does adding a superinductor contribute to the Hamiltonian? | Write the new energy term introduced by a circuit modification | |
| 9 | How does Fluxonium use \(E_C,E_J,E_L,\Phi_{\rm ext}\) to realize protection? | Map parameters to the potential, wavefunctions, and matrix elements | |
| 10 | Hamiltonian engineering: how can we design the Hamiltonian of a new qubit starting from the error operator we want to suppress? | Use four criteria to analyze \(0-\pi\), \(\cos2\varphi\), and the ferro-transmon |
Week 2: First Establish That a Circuit Can Become an Atom
Question 1 → Question 2: From Existence to Calculable Degrees of Freedom
Question 1 first explains why a macroscopic circuit can still behave quantum mechanically, so students do not initially mistake circuit quantization for a purely formal exercise. But saying only that “superconductors are quantum” is not enough to write a Hamiltonian, so Question 2 immediately identifies the conjugate variables: \((\Phi,Q)\) or \((\varphi,n)\).
Question 2 → Question 3: Successful Quantization Does Not Yet Mean a Successful Qubit
Once the conjugate variables are known, the LC mode can be quantized; Question 3 immediately exposes its limitation. This step deliberately separates a discrete spectrum from a selectively addressable qubit subspace. Otherwise students can easily assume that the two lowest states of any oscillator automatically form a good qubit.
Question 3 → Question 4: The Failure Condition Determines the Required Element
The LC oscillator fails because its levels are equally spaced. The next step is therefore not to add an arbitrary element, but to seek low-loss nonlinearity. The cosine potential of a Josephson junction addresses the need exposed by Question 3: it creates an anharmonic spectrum.
Question 4 → Question 5: After Building a Qubit, We Still Must Choose Its Parameter Regime
The Josephson junction makes an artificial atom possible, but a charge qubit remains sensitive to \(n_g\). The Transmon does not replace the basic Hamiltonian; it uses a larger shunt capacitance to enter the large-\(E_J/E_C\) regime, trading reduced charge dispersion for weaker anharmonicity.
W2 progression: Existence of a quantum mode → identify conjugate variables → discover the harmonic limitation → add Josephson nonlinearity → choose the Transmon parameter regime.
Week 3: Shift from “Building a Qubit” to “Making Errors Harder to Produce”
Question 5 → Question 6: Parameter Sensitivity Is Not All of Decoherence
The Transmon shows how to reduce one type of parameter sensitivity. W3 then asks a more general question: through which operator, and at what frequency, does the environment connect which pair of eigenstates? Question 6 therefore uses the transition rate to connect Hamiltonian structure to the environmental spectrum.
\[ \Gamma_{i\to f}\propto |\langle f|\hat O|i\rangle|^2S_F(\omega_{fi}). \]
Question 6 → Question 7: We Can Define Protection Only after Understanding How Errors Occur
Question 7 does not begin by introducing a particular qubit. It first establishes diagnostic tools: a sweet spot addresses low-frequency parameter fluctuations; a selection rule relies on symmetry; and small overlap produces matrix-element suppression. The control operator is introduced immediately so that “the environment cannot drive it” is not mistakenly rewritten as “nobody can control it.”
Question 7 → Question 8: Fluxonium as the First Concrete Realization
Only after the principle is established does Question 8 introduce the superinductor. Students then see more than the name of a new component: they see how a circuit modification adds the Hamiltonian term \(\tfrac12E_L(\varphi-\varphi_{\rm ext})^2\).
Question 8 → Question 9: After Adding a Term, Follow It through to the Eigenstate Structure
An additional Hamiltonian term does not automatically mean “protection.” Question 9 traces how \(E_C,E_J,E_L,\Phi_{\rm ext}\) changes the potential, parity, localization, tunneling, transition frequencies, and matrix elements, while examining dephasing and relaxation separately.
Question 9 → Question 10: Generalize the Fluxonium Case into Hamiltonian Engineering
If W3 ended with Fluxonium, students would merely have learned one more qubit type. Question 10 generalizes Fluxonium's design logic into Hamiltonian engineering: first specify the dominant noise operator and four design criteria, then survey \(0-\pi\), which adds degrees of freedom; \(\cos2\varphi\), which redesigns the Josephson potential; and the ferro-transmon, which uses a material state to modify \(E_J\).
W3 progression: How noise causes transitions → conditions for protection with retained control → add a Hamiltonian term → shape wavefunctions and matrix elements → design a new qubit backward from the target error.
Shared Formula Framework for the Ten Questions
1. LC oscillator: discrete but equally spaced
\[ \hat H_{LC}=\frac{\hat Q^2}{2C}+\frac{\hat\Phi^2}{2L} =\hbar\omega\left(\hat a^\dagger\hat a+\frac12\right) \]
2. Transmon: reduce charge dispersion by choosing a parameter regime
\[ \hat H_{\rm T}=4E_C(\hat n-n_g)^2-E_J\cos\hat\varphi \]
3. Transitions: both operator matrix elements and the environmental spectrum matter
\[ \hat V(t)=\hat O F(t),\qquad \Gamma_{i\to f}\propto|\langle f|\hat O|i\rangle|^2S_F(\omega_{fi}) \]
4. Fluxonium: add inductive energy and external-flux control
\[ \hat H_{\rm F}=4E_C\hat n^2-E_J\cos\hat\varphi +\frac12E_L(\hat\varphi-\varphi_{\rm ext})^2 \]
5. Four protected-qubit criteria
\[ \begin{aligned} &\frac{\partial\omega_{01}}{\partial\lambda}\approx0,\\ &\langle0|\hat O_{\rm noise}|1\rangle\approx0,\\ &\langle0|\hat O_{\rm drive}|1\rangle\neq0,\\ &\Delta_{\rm leakage}\ \text{sufficiently large}. \end{aligned} \]
Two Deliberately Retained Cross-Week Tasks
W2: Infer the Hamiltonian Backward from a Chip Image
The geometry → parameter → Hamiltonian → spectrum/control chain in Version 5 becomes a reverse-design assignment. Before they have full architecture knowledge, students need only propose one evidence-bounded hypothesis for a Transmon cell and identify which missing information requires simulation, fabrication data, or measurement.
W3: When Should Fluxonium Be Chosen over a Transmon?
This question is worth retaining, but W3 should not pretend to complete it. Students first leave a provisional answer; after adding control, coupling, readout, thermalization, fabrication, and yield in sequence, they revisit the architecture decision in W16. The difference between the two answers is evidence of cross-layer learning.
Toward W4: From Eigenstate Structure to Control
By the end of T2, students know that the effectiveness of a drive depends on \(\langle0|\hat O_{\rm drive}|1\rangle\), and that anharmonicity and the leakage gap constrain selective control. W4 can therefore proceed naturally to time-dependent drives, the rotating frame, Rabi rates, and DRAG without turning pulse control into a list of techniques disconnected from the device Hamiltonian.
Peir-Ru Wang