Qubit test chip (QPDK)

Four floating double-pad transmons, each with a meandered λ/2 readout coupled to one U-shaped probeline, two series IDCs at the probeline ends and two flux lines. The layout comes from QPDK’s qubit_test_chip.pic.yml (MIT, © 2024 gdsfactory) and is rebuilt on the Quantcore PDK: sapphire, trilayer SQUIDs, the QCT die frame. The rebuild is GPL-3.0.

This notebook is an example of a probeline with hanger readouts, the PDK pad rule (launchers 100 µm from the die edge) and the centring of the whole layout in the die.

The whole qubit_test_chip chip as the v2 notebook builds it

At a glance

Notebook

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

Source

gdsfactory quantum-rf-pdk qpdk/samples/qubit_test_chip.pic.yml, MIT (© 2024 gdsfactory)

Chip

7.5 × 7.5 mm Quantcore die, the source’s 4250 × 6200 µm circuit centred in it, CPW 10 / 6 µm

On the chip

4 launchers · 4 double-pad transmons · 4 λ/2 readouts · 1 probeline with 2 series IDCs · 2 flux lines

Builds today

yes (layout built on 2026-11-19 in 7 s)

How the notebook builds it

The layout segments of qubit_test_chip, one picture each
  1. Die. The 7.5 mm Quantcore die with the frame and the device ID.

  2. Bond pads. Four launchers on the 100 µm pad rule.

  3. Probeline. The two series IDCs and the U-shaped probeline.

  4. Qubits. Four DoublePads transmons with QPDK’s geometry, Q3 and Q4 mirrored.

  5. Readouts. Four λ/2 readouts hanging on the probeline.

  6. Coupling. None: the qubits do not couple to each other.

  7. Flux lines. Two flux lines ending in a 6 µm open gap between the qubit pairs.

  8. Whole chip. Sim groups (one per qubit), the centring (+1660 µm in x), save.

LOM estimates

Numbers are from a run of this notebook on 2026-11-19 (Setup, Design, Check and the LOM cells, 8 s): QuCat normal modes with QuTiP dressed values.

group

qubit

f_q (GHz)

α (MHz)

resonator

f_r (GHz)

κ (MHz)

χ (MHz)

Q1R1

Q1

5.763

-196

R1

11.87

0.638

-1.36

Q2R2

Q2

5.763

-197

R2

12.86

0.206

-1.27

Q3R3

Q3

5.764

-197

R3

16.16

0.4

1.61

Q4R4

Q4

5.763

-197

R4

12.33

0.713

-1.31

LOM plot of qubit_test_chip

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: qpdk/samples/qubit_test_chip.pic.yml (gdsfactory YAML netlist; qubit_test_chip.py only loads it) · licence MIT (“Copyright (c) 2024 gdsfactory”, checked in LICENSE) · Read on 2026-11-02 from a depth-1 clone. What it is: 4 floating double-pad transmons, each with its own meandered readout line coupled to one probeline that runs east along y = +400, turns, and runs back west along y = −400 (source frame). Two series interdigital capacitors at the probeline input and output, two T-shaped flux lines, four launchers. Image: docs/_static/images/qubit_test_chip.png. Frame: the source has no chip edge. Its extent is x 0…4015, y −3000…+3000 (launchers included). We use the chip edge of the filled variant (see filled_qubit_test_chip): box (0, 0, 4250, 6200), so Designer = source + (100, 3100). Layers: QPDK M1_DRAW / M1_ETCH (negative etch), Si εr 11.45. CPW everywhere: centre 10 µm, gap 6 µm, bend radius 100 µm (cpw). Launcher CPW 200/110 µm. Every ref resolves (DoublePads, FingerCapacitorSquare, BondPad), and DoublePads has a lumped LOM model.

Components

instance

catalog ref

key params (units)

at (x, y) Designer

rot

notes

PAD_IN

BondPad

default 400×450 pad; QPDK launcher is 200×200 + 100 taper

(100, 3500)

270 (port faces east)

source launcher_in (0, 400); port a ≈ (400, 3500)

PAD_OUT

BondPad

same

(100, 2700)

270

source launcher_out (0, −400)

PAD_FT

BondPad

same

(3100, 6100)

180 (faces south)

source launcher_top (3000, 3000) rot −90; flux top

PAD_FB

BondPad

same

(3100, 100)

0 (faces north)

source launcher_bot (3000, −3000) rot 90; flux bottom

C_IN

FingerCapacitorSquare

finger_number 4, finger_length 20, finger_gap 2, finger_width 5

(900, 3500)

0

series input capacitor (QPDK interdigital_capacitor defaults)

C_OUT

FingerCapacitorSquare

same

(900, 2700)

0

series output capacitor

Q1

DoublePads

island1/2_extent [400, 250], island_island_gap 15, island_r 0, ground_gap [800, 915], coupler_extent [394, 20], coupler_offset ≈35

(2485, 4725)

180 (coupler faces south)

source (2385, 1625); QPDK pads 250×400 rotated 90°, 200 µm etched pocket margin

Q2

DoublePads

same

(3715, 4725)

180

source (3615, 1625)

Q3

DoublePads

same

(3715, 1472.6)

0 (coupler faces north)

source (3615, −1627.421), mirrored about x

Q4

DoublePads

same

(2485, 1471.6)

0

source (2385, −1628.351), mirrored

Junction: QPDK squid_junction (2 junctions, loop area 4 µm²). No Lj is stated; use junction={"Lj_nH": 11, "Cj_fF": 2}. That value is assumed and gives about 5 GHz with C_Σ ≈ 80 fF. The QPDK coupler is a 394×20 µm plate (plate_capacitor_single) about 45 µm from the pad edge.

Routes

The feedline is a polyline. Hanger s is the coupler centre arclength from FL’s start (C_IN.b ≈ (930, 3500)), segments [0, 2485] east, [2485, 3285] south, [3285, 5770] west. Probeline ↔ resonator: QPDK coupler_straight gap 16 µm edge-to-edge, so centre spacing 26 µm and ground_gap = 26 − 5 − 6 − 5 − 6 = 4 µm.

route

from → to

type, length

waypoints / hanger

notes

FL_IN

PAD_IN.a → C_IN.a

cpw

straight, y 3500

probeline_input

FL

C_IN.b → C_OUT.b

cpw

nodes (3415, 3500), (3415, 2700)

U-turn of probeline_qubit2_to_qubit3

FL_OUT

C_OUT.a → PAD_OUT.a

cpw

straight, y 2700

probeline_output

R1

Q1.cplr → FL.hanger

meander 5000

s 1055, length 800, side left, ground_gap 4, direction backward

coupler x 1585…2385

R2

Q2.cplr → FL.hanger

meander 4600

s 2085, length 400, left, 4, backward

x 2815…3215; source resonator_meanders 3

R3

Q3.cplr → FL.hanger

meander 3600

s 3685, length 400, left, 4, forward

x 3215…2815 on the west-bound leg

R4

Q4.cplr → FL.hanger

meander 4800

s 4715, length 800, left, 4, forward

x 2385…1585

FLUX_T

PAD_FT.a → free_end((3100, 4875), dir 90, term 6)

cpw

straight down x 3100

ends in a QPDK tee with two 10 µm open stubs (east/west), between Q1 and Q2

FLUX_B

PAD_FB.a → free_end((3100, 1325), dir 270, term 6)

cpw

straight up x 3100

same tee between Q3 and Q4

side="left" is correct on both legs. Going east the resonators lie at +y; going west they lie at −y, which is also left. In QPDK the resonator line drops from the qubit to the coupler end nearest the qubit, runs along the probeline, then meanders away from it to an etched open end. The QPDK cell is a resonator_quarter_wave stitched to a coupler that is open at the qubit plate, so both ends are electrically open. The first spec used the standard λ/4 hanger (short at the coupler end) with the same total lengths; the notebook uses open λ/2 hangers (see Build below).

Launchers, pads, test structures, markers, labels

4 launchers only. No markers or labels. The filled variant adds the chip edge and vortex fill.

LOM node structure

  • Q_i: two floating pad nodes, a SQUID between them, C_pad–pad, C_pad–gnd; the coupler plate gives C_g to R_i’s open end.

  • R_i: CPW resonator (5000 / 4600 / 3600 / 4800 µm) coupled to FL by a 800/400/400/800 µm coupled-line section (4 µm ground strip).

  • FL: one conductor, PAD_IN → C_IN (series C) → … → C_OUT (series C) → PAD_OUT, two 50 Ω ports.

  • FLUX_T / FLUX_B: flux-bias controls. Each tee lies between two qubits and shares mutual inductance with both SQUIDs (weak, ~600 µm away).

  • No qubit–qubit bus.

Targets

None stated in the YAML. As λ/4 on Si (ε_eff ≈ 6.2) the lengths give about 6.0 / 6.5 / 8.3 / 6.2 GHz for R1–R4.

Simplifications for a first rebuild

  1. Drop C_IN/C_OUT and route FL straight from PAD_IN.a to PAD_OUT.a through the same nodes. Subtract ~30 µm from every s.

  2. Flux lines: plain CPW free ends (no tee). KQC WaveguideCoplanarSplitter can add the tee later.

  3. Stand-in qubit: use Xmon (claw_n toward FL) at the same positions.

Build

The notebook is examples/designs/qpdk/qubit_test_chip.ipynb. The readouts are open hangers (λ/2), not shorted λ/4: QPDK’s line runs from the coupler plate to an etched open end, so both ends are open. Meanders sit between the qubit and the coupled section (r = 50 µm). Launcher ports 110 µm further in (BondPad’s gap runs behind the pad). Lj 11 nH assumed. Flux lines end open (no tee).

Quantcore PDK rebuild

The notebook is a Quantcore PDK rebuild, not a replica: the netlist, LOM and imported components are kept, the process is Quantcore’s.

Notebook: examples/designs/qpdk/qubit_test_chip.ipynb.

  • Chip: the Quantcore die frame (L dicing markers 1000/0, 3 × 3 alignment grids 1001/0, chip boundary 100/0, device ID where an edge is free; any brand text is QCT), on the 7.5 × 7.5 mm Quantcore die; the source’s 4.25 × 6.2 mm chip area is centred in it (box=QP.die_box(4250, 6200)), so every coordinate stays the source’s.

  • Stack: sapphire 650 µm, εr 10.25 (the PDK default stack, used by the LOM estimates and every FEM export); QPDK’s process is Al on Si (εr 11.45), so frequencies are not the QPDK ones.

  • Junctions: Quantcore trilayer SQUIDs on the PDK rule (_qpdk.JJ_PARAMS): two 0.25 × 0.6 µm junctions (KQC’s junction_width = the straight finger) → R_n 45 kΩ in parallel, L_J 6.73 nH at zero flux, C_J 15 fF; fingers on 2/0 + 4/0, spacer 50/0 with the 51–53/0 gradient (Nb Δ = 1.4 meV, R_n·A 13.5 kΩ·µm² and C_J 50 fF/µm² assumed), in place of the assumed 11 nH + 2 fF (QPDK states none; its evaporated Al junctions are another process). Expected f_q ≈ 5.8 GHz (an estimate from the earlier LOM C_Σ; the 11 nH junction gave 4.75 GHz).

  • Kept: the topology and every component with the source’s parameters and positions, the readout / resonator lengths and coupled sections, the sim groups (readouts 5000 / 4600 / 3600 / 4800 µm as open-hanger λ/2, coupled 800 / 400 / 400 / 800 µm).