Parasitic Capacitance

Parasitic capacitance checks reveal unwanted coupling paths, loading, frequency shifts, and edge-field concentration.

#	Parameter	Symbol / Unit	Extraction Method	Typical Q3D Value	Ideal / Optimal	Good Range	Worst Case	Why It Matters	Key Design Note
29	Trace-to-Ground Parasitic Cap	C_trace / fF/μm	Q3D electrostatic solve per-unit-length	0.1 – 0.4 fF/μm	0.1 – 0.4 fF/μm (50 Ω CPW)	0.1 – 1 fF/μm	> 3 fF/μm	Distributed capacitance sets CPW characteristic impedance Z_0 = √(L/C); target 50 Ω	C’ depends on gap width and substrate thickness; narrowing the gap increases C’ (lowers Z_0)
30	Pad-to-Substrate Parasitic Cap	C_sub / fF	Q3D C matrix with substrate dielectric stack	< 10 fF	< 10 fF (small contact pads)	5 – 30 fF	> 100 fF	Substrate capacitance creates parasitic shunt path reducing resonator Q and shifting qubit freq	Thinning substrate from 500 μm to 200 μm reduces C_sub by ~2.5×; used in 3D chip stacks
31	Inter-Layer Cap (3D integration)	C_layer / fF	Q3D 3D stack model; bump height geometry sweep	< 5 fF per crossing	< 5 fF (flip-chip bump)	1 – 20 fF	> 50 fF	Capacitance between chip layers through indium/SnAg bumps must be included in Hamiltonian	Bump height variation (σ ~ 1–2 μm) causes C_layer spread of ~0.5–1 fF; matters at scale
32	Fringe Capacitance (gap edges)	C_fringe / fF/μm	Q3D conformal mesh at conductor edges	0.02 – 0.1 fF/μm	0.02 – 0.1 fF/μm (5–20 μm gap)	0.05 – 0.5 fF/μm	> 1 fF/μm	Fringe fields at edges add to intended coupling capacitance; must be in C_g design model	~30–50% of C_g in typical transmon designs comes from fringe; underestimating shifts f by 50+ MHz
33	Wirebond Pad Parasitic Cap	C_pad_wb / fF	Q3D parallel-plate + fringe approximation; or analytical C = ε₀ εr A/d	< 50 fF (100×100 μm Al pad)	< 50 fF	20 – 150 fF	> 300 fF	Bond-pad capacitance loads the signal line; degrades bandwidth and causes reflection at bond	Reducing pad size from 150×150 μm to 80×80 μm cuts C_pad by ~2.5× with no bond yield penalty