Kinetic-inductance TWPA

A kinetic-inductance travelling-wave parametric amplifier: a very long thin-film line whose inductance depends on the current. A DC bias and a pump give three-wave mixing along the line, and periodic loaded cells open a stop band that blocks the pump’s third harmonic. The geometry is the NY CREATES layout as ported in twpa-sim (MIT), redrawn as a negative mask on the Quantcore PDK.

This chip is dominated by one huge component: a 16-row serpentine of about 50 000 cells with two launches, two pads and two short feeds. The PDK Nb film is far less nonlinear than the source’s NbN.

The whole ki_twpa_serpentine chip as the v2 notebook builds it

At a glance

Notebook

examples/designs/amplifiers/ki_twpa_serpentine.ipynb (an older one-cell version is kept as ki_twpa_serpentine.v1.ipynb)

Source

twpa-sim kit.py / KITLayout (MIT), NY CREATES research layout

Chip

7.5 × 7.5 mm Quantcore die (the source’s is 5 × 5 mm), sapphire 650 µm

On the chip

1 KiTwpaSerpentine (16 rows × 55 supercells, 49 950 cells) · 2 bond pads · 2 feeds

Builds today

yes (layout built on 2026-11-19 in 15 s)

How the notebook builds it

The amplifier notebooks use six Design segments (die, pads, feed, nonlinear element, routing, whole chip); see the notebook template.

The six layout segments of ki_twpa_serpentine
  1. Die. The empty die with the QCT frame and the device ID KITWPA.

  2. Bond pads. Input and output pads on the west edge, at the heights of the first and last rows.

  3. Feed lines. No cell: signal, pump and DC bias share the serpentine’s own line; the feeds are drawn in segment 5.

  4. Nonlinear element. The serpentine: rows of one supercell etch, radial U-turns, and the two launches.

  5. Routing. Two short 50 Ω feeds from the pads to the launches.

  6. Whole chip. The launch sim group around the serpentine and the saved design.

LOM estimates

Numbers are from a run of this notebook on 2026-11-19 (Setup, Design, Check and the LOM cells, 16 s). This is an amplifier, so the LOM section gives gain and phase matching, not qubit frequencies. The line’s dispersion gives a stop band, and the 3WM coupled-mode gain is read at the pump:

line

pump (GHz)

G at f_p/2 (dB)

max G (dB)

PDK Nb

no stop band below 12 GHz

-

-

NbN 20 pH/□

8.985

14.8

14.8

twpa-sim v2

9.113

15.1

15.1

Stop band of the serpentine line

These are lumped-model estimates (component formulas and calibration), not FEM results. The LOM page says how far to trust them.

Build spec

The rest of this page is the detailed spec the notebook was built from, with every component, port, route and length.

Source · licence: mandalsoumyajit/twpa-sim (MIT): src/twpa/layout/kit.py, KITLayout, process kit_nycreates_research.json (NbN 20 pH/□). Die: 7.5 × 7.5 mm PDK die, PDK frame, sapphire 650 µm, Nb trilayer junctions. Notebook and design: examples/designs/amplifiers/ki_twpa_serpentine.ipynb / .json.

Components

instance

catalog ref

key params

notes

P_IN, P_OUT

BondPad

west edge (rows 0 and 15)

KIT

KiTwpaSerpentine

16 rows × 55 supercells (24 + 6 + 24 cells, 1.5 µm pitch), turns r 154 µm, 3 supercells per turn: 49 950 cells

Routes

route

from → to

type

L_IN

P_IN.a → KIT.a

CPW

L_OUT

P_OUT.a → KIT.b

CPW

LOM node structure

one conductor (wire between the pads); ladder from qcc_ladder (L_geo + L_s·squares, C per cell)

Targets (from the source)

twpa-sim v2: L 53.8 pH, C 22.3 / 8.4 fF per cell, ≈ 50 Ω, 3WM with DC bias

Simplifications / PDK differences

PDK Nb film (0.116 pH/□): ≈ 6 Ω, weakly nonlinear; NbN shown by a KineticInductance override in the analysis only