Quantum Computing RF

Coupling Q

Q: How does coupling Q relate to the resonator linewidth kappa?

The coupling quality factor is inversely proportional to the external photon decay rate: Q_c = omega_r / kappa_ext, where omega_r is the resonator angular frequency and kappa_ext is the rate at which photons leak from the resonator into the feedline. For a readout resonator at 7 GHz with a target linewidth kappa/2pi of 1 MHz, the coupling Q is approximately 7000. Lowering Q_c widens kappa, which speeds readout because the cavity responds faster, but it also increases the Purcell decay channel for the qubit. Designers typically tune Q_c between 1000 and 20000 by adjusting the coupling capacitor between the resonator and the feedline.

Q: What is the difference between coupling Q, internal Q, and loaded Q?

Internal Q (Q_i) captures intrinsic losses such as dielectric two-level-system loss, conductor loss, and radiation. Coupling Q (Q_c), also called external Q, captures the deliberate energy leakage into the measurement feedline. They combine into the loaded Q through 1/Q_L = 1/Q_i + 1/Q_c. In a well-designed readout resonator you want Q_c much smaller than Q_i so the loaded response is dominated by the controlled coupling rather than uncontrolled loss; this is the overcoupled regime. Superconducting resonators routinely reach Q_i above 1 million, while Q_c is engineered to a few thousand for fast readout.

Q: How does coupling Q limit qubit lifetime through the Purcell effect?

When a qubit is dispersively coupled to a lossy resonator, it inherits a decay channel called the Purcell rate, gamma_Purcell = kappa times (g / Delta) squared, where g is the qubit-resonator coupling, Delta is the detuning, and kappa is set by the coupling Q. A smaller Q_c gives a larger kappa and faster readout but shortens the Purcell-limited T1. A typical transmon at 5 GHz detuned 1.5 GHz from a 7 GHz resonator with g/2pi of 100 MHz and kappa/2pi of 1 MHz sees a Purcell limit near 360 microseconds. Engineers add a Purcell filter to break this trade-off, allowing low Q_c for fast readout while protecting the qubit at its own frequency.

/ KUH-pling kyoo /

Often written Q_c and called the coupling quality factor or external Q, this figure of merit measures how strongly a readout resonator is coupled to the transmission line that carries the microwave probe signal in a circuit QED system. It fixes the external photon decay rate κ_ext = ω_r/Q_c, which in turn governs how fast a qubit state can be measured. A low Q_c (strong coupling) gives a wide resonator linewidth and fast dispersive readout, but it also opens a larger Purcell decay channel that shortens qubit lifetime. Together with the internal loss captured by external Q and intrinsic Q, it sets the loaded response that an engineer sees on a vector network analyzer at millikelvin temperatures.

Category: Quantum Computing RF

Typical Q_c: 1,000 to 20,000

κ/2π range: 0.3 to 5 MHz

How Coupling Q Shapes Qubit Readout

In a circuit quantum electrodynamics architecture, each superconducting qubit is interrogated through a dedicated microwave resonator that is capacitively tapped onto a shared feedline. The coupling quality factor describes the rate at which photons stored in that resonator leak out into the feedline where they can be amplified and digitized. Because the measured signal-to-noise ratio accumulates while photons remain in the cavity, but the measurement must finish before the qubit decays, Q_c sits at the center of a speed versus protection trade-off. A resonator near 7 GHz with Q_c of 7,000 has a linewidth κ/2π of about 1 MHz, giving a cavity ring-up and ring-down time near 160 ns, which is well matched to single-shot readout in roughly 200 to 500 ns.

The coupling is set physically by the value of the coupling capacitor or the gap between the resonator and the feedline. Increasing that capacitance lowers Q_c and broadens κ, accelerating readout. The penalty appears through the Purcell effect: the qubit, dispersively detuned from the resonator, inherits a decay path proportional to κ. Without mitigation, pushing Q_c below about 2,000 can drag the Purcell-limited T₁ down toward 100 μs, which is unacceptable for modern processors targeting coherence above 200 μs.

The standard remedy is a bandpass Purcell filter placed between the resonator and the feedline. It presents low impedance at the resonator frequency, so readout photons escape quickly, while presenting high impedance at the qubit frequency to block spontaneous emission. This decouples the readout coupling from the qubit protection, letting designers choose an aggressively low Q_c for fast measurement without sacrificing lifetime.

Governing Relations for Q_c

Coupling Q and external linewidth:
Q_c = ω_r / κ_ext → κ_ext = 2πf_r / Q_c

Loaded Q from internal and coupling Q:
1 / Q_L = 1 / Q_i + 1 / Q_c

Purcell decay rate (dispersive regime):
γ_Purcell ≈ κ × (g / Δ)² → T_1,Purcell = 1 / γ_Purcell

Where ω_r = 2πf_r resonator frequency, κ_ext = external photon decay rate, Q_i = internal (loss) Q, g = qubit-resonator coupling, Δ = qubit-resonator detuning. Example: f_r = 7 GHz, Q_c = 7000 → κ/2π ≈ 1 MHz; with g/2π = 100 MHz, Δ/2π = 1.5 GHz → T_1,Purcell ≈ 360 μs.

Coupling Q Design Regimes

Regime	Q_c range	κ/2π (7 GHz)	Readout time	Purcell T₁ impact	Typical use
Strongly overcoupled	500 to 2,000	3.5 to 14 MHz	< 150 ns	Severe without filter	Fast feedback, error correction
Overcoupled (typical)	2,000 to 8,000	0.9 to 3.5 MHz	200 to 500 ns	Moderate; filter advised	Standard single-shot readout
Near critical	Q_c ≈ Q_i	< 1 MHz	> 1 μs	Low	High-fidelity, slow readout
Undercoupled	> 50,000	< 0.14 MHz	Very slow	Negligible	Resonator loss characterization

Common Questions

Frequently Asked Questions

How does coupling Q relate to the resonator linewidth kappa?

They are inversely related through Q_c = ω_r/κ_ext. A 7 GHz readout resonator with a target κ/2π of 1 MHz needs Q_c ≈ 7,000. Lowering Q_c widens κ and speeds readout because the cavity responds faster, but it also increases the Purcell decay channel. Designers set Q_c between roughly 1,000 and 20,000 via the coupling capacitor to the feedline.

What is the difference between coupling Q, internal Q, and loaded Q?

Internal Q (Q_i) captures intrinsic loss such as two-level-system dielectric loss and conductor loss. Coupling Q (Q_c), or external Q, captures deliberate leakage into the feedline. They combine as 1/Q_L = 1/Q_i + 1/Q_c. For readout you want Q_c << Q_i, the overcoupled regime, so the controlled coupling dominates. Superconducting resonators reach Q_i above 1 million while Q_c is engineered to a few thousand.

How does coupling Q limit qubit lifetime through the Purcell effect?

A dispersively coupled qubit inherits a decay rate γ_Purcell ≈ κ(g/Δ)², and κ is set by Q_c. Smaller Q_c means faster readout but shorter Purcell-limited T₁. A transmon at 5 GHz, detuned 1.5 GHz from a 7 GHz resonator with g/2π = 100 MHz and κ/2π = 1 MHz, has a Purcell limit near 360 μs. A Purcell filter breaks the trade-off by blocking emission at the qubit frequency.

Related Terms

← Coupling Path Coupling Strength →

Quantum RF Hardware

Build Your Readout Chain

From millikelvin-rated cryogenic components to low-noise frequency converters, RF Essentials supplies the microwave hardware that supports high-fidelity qubit readout. Talk to our engineers about your circuit QED requirements.

Get in Touch