Quantum Computing RF

Coupling Strength

Q: How is coupling strength g related to the vacuum Rabi splitting?

When a qubit is tuned into resonance with a cavity, the coupling lifts the degeneracy of the dressed states and splits the single transmission peak into two peaks separated by 2g in frequency. This observable separation is the vacuum Rabi splitting, and measuring it directly yields the coupling strength. For a typical transmon with g divided by 2 pi of 100 MHz, the two normal-mode peaks sit 200 MHz apart. The splitting is resolvable, meaning the system is in the strong-coupling regime, only when 2g exceeds the linewidths set by the cavity decay rate kappa and the qubit decoherence rate gamma. A vacuum Rabi splitting of 200 MHz against a cavity linewidth of a few MHz gives a cooperativity well above unity.

Q: What sets the value of g for a transmon coupled to a coplanar waveguide resonator?

The coupling strength scales as g equals the qubit transition dipole times the resonator vacuum voltage divided by h-bar. In capacitive coupling, g is proportional to the ratio of the coupling capacitance Cg to the total qubit capacitance, multiplied by the square root of the resonator frequency divided by the resonator capacitance. Practically, the designer sets g by sizing the coupling pad geometry between the transmon island and the resonator. Increasing Cg from about 2 fF to 10 fF raises g divided by 2 pi from roughly 40 MHz to over 150 MHz. Values are kept below a few hundred MHz so that g stays small compared to the qubit anharmonicity, preserving the two-level approximation.

Q: Why must g stay much smaller than the qubit-resonator detuning during readout?

Dispersive readout relies on operating with a large detuning Delta between the qubit and resonator frequencies so that no direct photon exchange occurs. The dispersive shift chi is approximately g squared divided by Delta, and the validity of this regime requires the ratio g divided by Delta to be well below 1, typically around 0.1. If g approaches Delta the qubit and resonator hybridize, the state becomes entangled with cavity photons, and measurement-induced dephasing and Purcell decay degrade qubit coherence. A common design point is g divided by 2 pi near 100 MHz with Delta divided by 2 pi near 1 GHz, giving a chi divided by 2 pi of several MHz that is large enough for fast readout yet safely dispersive.

/kuhp-ling strengkth/

Often written as g (or g/2π when quoted in frequency units), this is the coherent rate at which a qubit and a microwave resonator mode exchange a single photon. In circuit QED it is the central parameter of the Jaynes-Cummings model, fixing how fast energy oscillates between the artificial atom and the cavity field. When 2g exceeds both the cavity decay rate κ and the qubit decoherence rate γ, the system enters the strong-coupling regime and shows a resolvable vacuum Rabi splitting. Typical superconducting transmon designs target g/2π between 50 and 300 MHz; operating far off resonance turns this same g into the dispersive shift that enables dispersive readout.

Category: Quantum Computing RF

Typical g/2π: 50 to 300 MHz

Strong coupling: 2g > κ, γ

How Coupling Strength Governs Qubit-Resonator Interaction

In a superconducting quantum processor, a qubit is rarely isolated; it is deliberately tied to a microwave resonator that serves as a control bus and a readout sensor. Coupling strength quantifies that connection. Physically, the qubit presents an electric dipole moment to the resonator's vacuum electric field, and the product of the two, divided by Planck's reduced constant, sets g. A larger coupling capacitance between the transmon island and the resonator center conductor produces a larger g, which is why the coupling pad geometry is one of the most carefully laid-out features on the chip.

The behavior of the coupled system splits cleanly into two regimes governed by the detuning Δ between the qubit and resonator frequencies. On resonance (Δ = 0), the qubit and cavity hybridize into two dressed states separated by 2g, and a single photon sloshes back and forth at the vacuum Rabi frequency. Far off resonance (|Δ| much larger than g), no real photon exchange happens, but the qubit pulls the resonator frequency by the dispersive shift χ ≈ g²/Δ. The same physical g therefore enables two very different functions: fast iSWAP-style energy exchange near resonance, and quantum-non-demolition state measurement in the dispersive limit.

Engineering g is a balancing act. It must be large enough that 2g beats the linewidths to reach strong coupling, and large enough that g²/Δ gives a measurable dispersive shift. Yet it must stay small compared to the transmon anharmonicity (typically 200 to 300 MHz) so the device remains a clean two-level system, and small compared to Δ so the dispersive approximation holds and Purcell-limited decay through the resonator stays manageable.

Vacuum Rabi Rate and the Cooperativity Figure of Merit

The cleanest experimental signature of g is the vacuum Rabi splitting observed in cavity transmission. To compare devices regardless of their loss, designers fold g, κ, and γ into a single dimensionless cooperativity C = g²/(κγ). Strong coupling requires C greater than 1, and modern transmon-cavity systems routinely reach C in the hundreds to thousands because g/2π near 100 MHz dwarfs cavity linewidths of a few MHz and qubit decoherence rates below 100 kHz.

Coupling Strength Equations

Jaynes-Cummings interaction (single mode):
H_int = ℏg (σ₊a + σ_-a^†)

Capacitive coupling rate (transmon):
g ≈ (C_g / C_Σ) × ω_r × √(π × Z_r / R_Q)

Dispersive shift (off resonance):
χ ≈ g² / Δ, with Δ = ω_q − ω_r

Cooperativity:
C = g² / (κ × γ)

Where g = coupling strength (rad/s), C_g = coupling capacitance, C_Σ = total qubit capacitance, ω_r = resonator frequency, Z_r = resonator impedance, R_Q = h/(2e)² ≈ 6.45 kΩ, κ = cavity decay rate, γ = qubit decoherence rate. Example: g/2π = 100 MHz, Δ/2π = 1 GHz → χ/2π ≈ 10 MHz.

Coupling Regimes Compared

Regime	Condition	g/2π (typical)	Key signature	Primary use
Weak coupling	2g < κ, γ	< 1 MHz	Purcell-modified decay only	Spontaneous-emission control
Strong (resonant)	2g > κ, γ; Δ ≈ 0	50 to 300 MHz	Vacuum Rabi splitting 2g	Photon exchange, swap gates
Dispersive	g << Δ	50 to 300 MHz	State-dependent shift χ	QND readout, ZZ gates
Ultrastrong	g / ω_r > 0.1	> 500 MHz	Counter-rotating terms matter	Fundamental light-matter studies

Common Questions

Frequently Asked Questions

How is coupling strength g related to the vacuum Rabi splitting?

On resonance, the coupling lifts the degeneracy of the dressed states and splits a single cavity transmission peak into two peaks separated by 2g. Measuring that separation directly gives g. For g/2π = 100 MHz the two normal-mode peaks sit 200 MHz apart. The splitting is only resolvable, meaning strong coupling, when 2g exceeds the cavity linewidth κ and the qubit decoherence rate γ.

What sets the value of g for a transmon coupled to a coplanar waveguide resonator?

g scales with the qubit dipole and the resonator's vacuum voltage, which for capacitive coupling means g is proportional to C_g/C_Σ. Sizing the coupling pad sets it directly: raising C_g from about 2 fF to 10 fF moves g/2π from roughly 40 MHz to over 150 MHz. Designers cap g well below the anharmonicity so the transmon stays a clean two-level system.

Why must g stay much smaller than the qubit-resonator detuning during readout?

Dispersive readout needs a large detuning Δ so no real photon exchange occurs, with the shift χ ≈ g²/Δ. The approximation holds only when g/Δ is well below 1, around 0.1. If g approaches Δ the qubit and cavity hybridize, causing measurement-induced dephasing and Purcell decay. A common design point is g/2π near 100 MHz and Δ/2π near 1 GHz, giving χ/2π of several MHz.

Related Terms

← Coupling Q Coupling (Waveguide) →