KQC test chips

These are KQCircuits’ own test structures: chips that test a process or a part rather than a quantum device. Two of them (simple, shaping) carry a qubit. The rest test airbridges, junctions, lithography, flip-chip crossings, through-silicon vias and the sample holder. All are 10 × 10 mm and rebuilt on the Quantcore PDK (GPL-3.0, © IQM Finland Oy).

Build them to see how a whole KQC chip cell is placed in a Designer and what each process test looks like. Each block below has the whole chip, a one-line purpose and how the notebook builds it. Most have no qubit or resonator, so their LOM section has nothing to estimate.

The notebooks

Notebook

What it tests

simple

a minimal connected chip: one Swissmon coupled to a through line; the smoke test for KQC parity

shaping

a readout design: one Swissmon with two Purcell-filtered quarter-wave readout chains

airbridge_crossings

airbridge crossings of lines, and a mechanical array of 608 short bridge spans (19 lengths, 8 widths, 4 distances)

airbridge_dc_test

four-point DC resistance of airbridge chains (1 to 22 bridges per test)

crossing_twoface

flip-chip line crossings (bridges on the top chip) and a crosstalk pair

quality_factor_twoface

resonator quality factor with a flip-chip lid: six capped spiral quarter-wave resonators on one probe line

sample_holder_test

a 40-pad sample holder: every pad pair is joined by a line

empty

electron-beam lithography with the ground removed in a cross-shaped area

junction_test

electron-beam lithography and junction process: a cross opening filled with an array of 80 × 80 µm squares

junction_test2

SQUID junction arrays on test pads, with stepped junction widths

lithography_test

optical lithography resolution: stripe blocks with 1 to 15 µm lines, spaced 1 and 5 times the width, and crosses

lithography_test_twoface

the same lithography test on both faces of a flip chip

stripes

ground-grid and lithography stripes: 80 µm metal stripes at 200 µm pitch and 80 µm squares in a cross opening

tsv_test

the through-silicon-via process: a TSV fence pattern and three metrology arrays for cross-sections

simple

Tests: a minimal connected chip: one Swissmon coupled to a through line; the smoke test for KQC parity.

The whole simple chip as the notebook builds it

Notebook

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

Source

KQCircuits chips/simple.py, © IQM Finland Oy, GPL-3.0

On the chip

3 SMA8 launchers (WN, EN, SE) · 1 Swissmon · 2 finger capacitors · 1 T · 1 flux line

Builds today

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

How it builds: Die, bond pads, signal line, qubit, qubit onto the line, flux line, whole chip.

The layout segments of Simple, one picture each

Numbers are from a run of this notebook on 2026-11-19 (Setup, Design, Check and the LOM cells, 7 s): QuCat normal modes with QuTiP dressed values.

mode

kind

f (GHz)

κ (MHz)

A (MHz)

participation

QB

qubit

6.193

0.425

198

J_QB 1.00

parameter

LOM (QuTiP)

target

QB.f_q_GHz

5.99

–

QB.alpha_MHz

-214

–

LOM plot of simple

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

shaping

Tests: a readout design: one Swissmon with two Purcell-filtered quarter-wave readout chains.

The whole shaping chip as the notebook builds it

Notebook

examples/designs/kqc/test_chips/shaping.ipynb (an older one-cell version is kept as shaping.v1.ipynb)

Source

KQCircuits chips/shaping.py, © IQM Finland Oy, GPL-3.0

On the chip

8 SMA8 launchers · 1 Swissmon · 2 readout chains (resonator, filter, capacitors) · charge and flux line

Builds today

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

How it builds: Die, bond pads, qubit, the two chains, control lines, whole chip.

The layout segments of Shaping, one picture each

Numbers are from a run of this notebook on 2026-11-19 (Setup, Design, Check and the LOM cells, 17 s): QuCat normal modes with QuTiP dressed values.

parameter

LOM (QuTiP)

target

QB.f_q_GHz

6.45

–

QB.alpha_MHz

-242

–

PB2.chi_MHz

-0.00569

–

PA2.chi_MHz

-0.00847

–

PB2.f_GHz

7.48

–

PB2.kappa_MHz

27.5

–

PA2.f_GHz

7.72

–

PA2.kappa_MHz

27.9

–

qubit

f_q (GHz)

A (MHz)

resonator

f_r (GHz)

κ_r (MHz)

χ (MHz)

QB

6.68

222

RA3

8.345

0.326

-6.46

resonator

LOM line (µm)

KQC target sum (µm)

T arms not in the LOM (µm)

RA3

2750

3060

306

PA2

3790

3950

153

RB3

2870

3170

306

PB2

3910

4070

153

LOM plot of shaping

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

airbridge_crossings

Tests: airbridge crossings of lines, and a mechanical array of 608 short bridge spans (19 lengths, 8 widths, 4 distances).

The whole airbridge_crossings chip as the notebook builds it

Notebook

examples/designs/kqc/test_chips/airbridge_crossings.ipynb (an older one-cell version is kept as airbridge_crossings.v1.ipynb)

Source

KQCircuits chips/airbridge_crossings.py, © IQM Finland Oy, GPL-3.0

On the chip

4 lines (PL-1 to PL-4) between SMA8 launchers · line crossings on airbridges · the mechanical array

Builds today

yes (layout built on 2026-11-19 in 73 s)

How it builds: Die, bond pads, the mechanical array, the four lines.

The layout segments of AirbridgeCrossings, one picture each

This chip has no junction or resonator (four 50 Ω through lines), so there is no lumped-model estimate. What it measures (the transmission across the crossings and the PL-1 to PL-3 crosstalk) needs a driven FEM solve.

airbridge_dc_test

Tests: four-point DC resistance of airbridge chains (1 to 22 bridges per test).

The whole airbridge_dc_test chip as the notebook builds it

Notebook

examples/designs/kqc/test_chips/airbridge_dc_test.ipynb (an older one-cell version is kept as airbridge_dc_test.v1.ipynb)

Source

KQCircuits chips/airbridge_dc_test.py, © IQM Finland Oy, GPL-3.0

On the chip

22 AirbridgeDC tests in three bands, no launchers

Builds today

yes (layout built on 2026-11-19 in 4 s)

How it builds: Die, the chip cell, a view of the longest chain.

The layout segments of AirbridgeDcTest, one picture each

This chip has no junction, qubit or resonator, so there is no lumped-model estimate, and the notebook’s LOM section says there is nothing to simulate. The chip is measured in the process (optical or EBL inspection, DC probing) and not with microwaves.

crossing_twoface

Tests: flip-chip line crossings (bridges on the top chip) and a crosstalk pair.

The whole crossing_twoface chip as the notebook builds it

Notebook

examples/designs/kqc/test_chips/crossing_twoface.ipynb (an older one-cell version is kept as crossing_twoface.v1.ipynb)

Source

KQCircuits chips/crossing_twoface.py, © IQM Finland Oy, GPL-3.0

On the chip

4 routes between SMA8 launchers · 6 flip-chip crossings of PL-1 by PL-3

Builds today

yes (layout built on 2026-11-19 in 5 s)

How it builds: Die, bond pads, the two straight lines, the crossing lines.

The layout segments of CrossingTwoface, one picture each

This chip has no junction or resonator (four 50 Ω through lines), so there is no lumped-model estimate. What it measures (the transmission through the flip-chip crossings and the PL-1 to PL-3 and PL-2 to PL-4 crosstalk) needs a driven FEM solve.

quality_factor_twoface

Tests: resonator quality factor with a flip-chip lid: six capped spiral quarter-wave resonators on one probe line.

The whole quality_factor_twoface chip as the notebook builds it

Notebook

examples/designs/kqc/test_chips/quality_factor_twoface.ipynb (an older one-cell version is kept as quality_factor_twoface.v1.ipynb)

Source

KQCircuits chips/quality_factor_twoface.py, © IQM Finland Oy, GPL-3.0

On the chip

2 SMA8 launchers · 1 probe line · 6 spiral λ/4 resonators (5374 to 5589 µm) · 6 coupling capacitors

Builds today

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

How it builds: Die, bond pads, couplers, resonators, probe line.

The layout segments of QualityFactorTwoface, one picture each

Numbers are from a run of this notebook on 2026-11-19 (Setup, Design, Check and the LOM cells, 16 s): QuCat normal modes with QuTiP dressed values.

mode

kind

f (GHz)

κ (MHz)

A (MHz)

participation

R5

resonator

5.578

0.0652

0

R5 1.00

R0

resonator

5.652

1.42

0

R0 1.00

R4

resonator

5.682

0.0456

0

R4 1.00

R1

resonator

5.688

0.739

0

R1 1.00

R3

resonator

5.773

0.124

0

R3 1.00

R2

resonator

5.795

0.226

0

R2 1.00

mode

Q_ext (LOM)

R0

3980

R1

7700

R2

25600

R3

46500

R4

125000

R5

85600

LOM plot of quality_factor_twoface

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

sample_holder_test

Tests: a 40-pad sample holder: every pad pair is joined by a line.

The whole sample_holder_test chip as the notebook builds it

Notebook

examples/designs/kqc/test_chips/sample_holder_test.ipynb (an older one-cell version is kept as sample_holder_test.v1.ipynb)

Source

KQCircuits chips/sample_holder_test.py, © IQM Finland Oy, GPL-3.0

On the chip

40 RF launchers (10 per side) · 20 CPWs

Builds today

yes (layout built on 2026-11-19 in 6 s)

How it builds: Die, the 40 launchers, the 20 lines.

The layout segments of SampleHolderTest, one picture each

This chip has no junction or resonator (20 × 50 Ω through lines; the lengths are printed in Design 4), so there is no lumped-model estimate.

empty

Tests: electron-beam lithography with the ground removed in a cross-shaped area.

The whole empty chip as the notebook builds it

Notebook

examples/designs/kqc/test_chips/empty.ipynb (an older one-cell version is kept as empty.v1.ipynb)

Source

KQCircuits chips/empty.py, © IQM Finland Oy, GPL-3.0

On the chip

one chip-level cross-shaped opening, no components

Builds today

yes (layout built on 2026-11-19 in 4 s)

How it builds: Die, the chip cell, a view of the cross corner.

The layout segments of Empty, one picture each

This chip has no junction, qubit or resonator, so there is no lumped-model estimate, and the notebook’s LOM section says there is nothing to simulate. The chip is measured in the process (optical or EBL inspection, DC probing) and not with microwaves.

junction_test

Tests: electron-beam lithography and junction process: a cross opening filled with an array of 80 × 80 µm squares.

The whole junction_test chip as the notebook builds it

Notebook

examples/designs/kqc/test_chips/junction_test.ipynb (an older one-cell version is kept as junction_test.v1.ipynb)

Source

KQCircuits chips/junction_test.py, © IQM Finland Oy, GPL-3.0

On the chip

one chip-level cross opening with a square array, no components

Builds today

yes (layout built on 2026-11-19 in 3 s)

How it builds: Die, the chip cell, a view of the cross corner.

The layout segments of JunctionTest, one picture each

This chip has no junction, qubit or resonator, so there is no lumped-model estimate, and the notebook’s LOM section says there is nothing to simulate. The chip is measured in the process (optical or EBL inspection, DC probing) and not with microwaves.

junction_test2

Tests: SQUID junction arrays on test pads, with stepped junction widths.

The whole junction_test2 chip as the notebook builds it

Notebook

examples/designs/kqc/test_chips/junction_test2.ipynb (an older one-cell version is kept as junction_test2.v1.ipynb)

Source

KQCircuits chips/junction_test2.py, © IQM Finland Oy, GPL-3.0

On the chip

6 pad arrays, each with a Quantcore SQUID between its pad pairs

Builds today

yes (layout built on 2026-11-19 in 4 s)

How it builds: Die, the chip cell, the pad arrays, one test SQUID.

The layout segments of JunctionTest2, one picture each

This chip has no modes to simulate, so there are no qubit or resonator estimates. Numbers are from a run of this notebook on 2026-11-19 (4 s): what the chip measures is each test SQUID’s R_n at room temperature, so the notebook lists the R_n and L_J the PDK junction rule gives each stepped finger (two junctions in parallel per SQUID; Nb Δ = 1.4 meV, R_n·A = 13.5 kΩ·µm² assumed).

array

loop

SQUIDs

straight finger (µm)

R_n per SQUID (kΩ)

L_J per SQUID (nH)

in 40-50 kΩ

1

large (130 µm²)

8

0.10 to 0.31

112.5 to 36.3

16.84 to 5.43

2

2

large (130 µm²)

13

0.34 to 0.70

33.1 to 16.1

4.95 to 2.41

0

3

small (80 µm²)

13

0.15 to 0.27

75.0 to 41.7

11.22 to 6.24

5

4

small (80 µm²)

13

0.28 to 0.40

40.2 to 28.1

6.01 to 4.21

1

5

small (80 µm²)

13

0.41 to 0.53

27.4 to 21.2

4.11 to 3.18

0

6

large (130 µm²)

8

0.73 to 0.94

15.4 to 12.0

2.31 to 1.79

0

8 SQUIDs fall inside the PDK’s 40 to 50 kΩ window.

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

lithography_test

Tests: optical lithography resolution: stripe blocks with 1 to 15 µm lines, spaced 1 and 5 times the width, and crosses.

The whole lithography_test chip as the notebook builds it

Notebook

examples/designs/kqc/test_chips/lithography_test.ipynb (an older one-cell version is kept as lithography_test.v1.ipynb)

Source

KQCircuits chips/lithography_test.py, © IQM Finland Oy, GPL-3.0

On the chip

horizontal, vertical and diagonal line blocks, no components

Builds today

yes (layout built on 2026-11-19 in 3 s)

How it builds: Die, the chip cell, horizontal, vertical and diagonal blocks.

The layout segments of LithographyTest, one picture each

This chip has no junction, qubit or resonator, so there is no lumped-model estimate, and the notebook’s LOM section says there is nothing to simulate. The chip is measured in the process (optical or EBL inspection, DC probing) and not with microwaves.

lithography_test_twoface

Tests: the same lithography test on both faces of a flip chip.

The whole lithography_test_twoface chip as the notebook builds it

Notebook

examples/designs/kqc/test_chips/lithography_test_twoface.ipynb (an older one-cell version is kept as lithography_test_twoface.v1.ipynb)

Source

KQCircuits chips/lithography_test_twoface.py, © IQM Finland Oy, GPL-3.0

On the chip

the line blocks of LithographyTest on the bottom chip, six mirrored blocks on the top chip

Builds today

yes (layout built on 2026-11-19 in 3 s)

How it builds: Die, the chip cell, horizontal, vertical and diagonal blocks, top chip.

The layout segments of LithographyTestTwoface, one picture each

This chip has no junction, qubit or resonator, so there is no lumped-model estimate, and the notebook’s LOM section says there is nothing to simulate. The chip is measured in the process (optical or EBL inspection, DC probing) and not with microwaves.

stripes

Tests: ground-grid and lithography stripes: 80 µm metal stripes at 200 µm pitch and 80 µm squares in a cross opening.

The whole stripes chip as the notebook builds it

Notebook

examples/designs/kqc/test_chips/stripes.ipynb (an older one-cell version is kept as stripes.v1.ipynb)

Source

KQCircuits chips/stripes.py, © IQM Finland Oy, GPL-3.0

On the chip

one chip-level cross opening with stripes, no components

Builds today

yes (layout built on 2026-11-19 in 3 s)

How it builds: Die, the chip cell, a view of the cross corner.

The layout segments of Stripes, one picture each

This chip has no junction, qubit or resonator, so there is no lumped-model estimate, and the notebook’s LOM section says there is nothing to simulate. The chip is measured in the process (optical or EBL inspection, DC probing) and not with microwaves.

tsv_test

Tests: the through-silicon-via process: a TSV fence pattern and three metrology arrays for cross-sections.

The whole tsv_test chip as the notebook builds it

Notebook

examples/designs/kqc/test_chips/tsv_test.ipynb (an older one-cell version is kept as tsv_test.v1.ipynb)

Source

KQCircuits chips/tsv_test.py, © IQM Finland Oy, GPL-3.0

On the chip

1 TsvTestPattern · 3 arrays of single vias (10 × 10, 8 × 8, 6 × 6)

Builds today

yes (layout built on 2026-11-19 in 5 s)

How it builds: Die, the chip cell, the centre pattern, a metrology array.

The layout segments of TsvTest, one picture each

This chip has no junction, qubit or resonator, so there is no lumped-model estimate, and the notebook’s LOM section says there is nothing to simulate. The chip is measured in the process (optical or EBL inspection, DC probing) and not with microwaves.

Build spec

The rest of this page is the detailed spec the notebooks were built from: the sources, coordinates, routes and the build status of each chip.

Sources other-libs/KQCircuits-main/KQCircuits-main/klayout_package/python/kqcircuits/chips/<module>.py, GPL-3.0 (IQM Finland). Rebuilt from extracted parameters. All of them are 10 × 10 mm with box (0,0)–(10000,10000), and the KQC frame = Designer frame. Coordinates come from a scratch build of each KQC chip on 2026-11-02. The “buildable now?” column below is per chip. launchers={"type": T, "assign": {pad: name}} reproduces KQC launcher_assignments. Plain "SMA8" gives pads NW, NE, EN, ES, SE, SW, WS, WN at (2800,9200), (7200,9200), (9200,7200), (9200,2800), (7200,800), (2800,800), (800,2800), (800,7200).

Components

chip (module)

purpose

main elements (catalog ref, params, placement)

buildable now?

JunctionTest (junction_test)

EBL / junction process test

One chip-level polygon in base_metal_gap_wo_grid: a cross-shaped opening (arms 2000–8000, reaching 100 + dice_width from the edges) minus an array of 80 × 80 µm squares at 100 µm pitch (edge_len=80, inter_space=20). No elements.

yes (placed as the whole KQC chip PCell, see the rebuild below)

JunctionTest2 (junction_test2)

SQUID test-pad arrays

6 × JunctionTestPadsSimple, margin=20, pad_width=500, pad_spacing=200, area_width=1600: #1 (300,2000) h 6000, loop_area 130, junction_width_steps=[0.08,0.03]; #2 (2000,300) h 9400, 130, [0.32,0.03]; #3 (3400,300) h 9400, 80, [0.15,0.01]; #4 (4800,300) 80, [0.28,0.01]; #5 (6200,300) 80, [0.41,0.01]; #6 (8000,2000) h 6000, 130, [0.71,0.03]. KQC also passes only_arms=True, which is not in the Simple schema and is ignored.

yes

AirbridgeCrossings (airbridge_crossings)

airbridge crossing + mechanical test

SMA8 assign {1:PL-1-IN, 2:PL-2-IN, 3:PL-4-IN, 4:PL-4-OUT, 5:PL-2-OUT, 6:PL-1-OUT, 7:PL-3-OUT, 8:PL-3-IN}. Mechanical array: 19 lengths × 8 widths × 4 distances = 608 straight CPWs, WaveguideCoplanarStraight(l = 2·5·(dist+width)+4), centred at (3600,9650) + (0,−121)·i + (400,0)·j + (0,−2350)·k. On rows with (i+k) odd and 1000 < y < 9000 there are 5 AirbridgeRectangular (defaults) at x offsets 50·(n−2): 1280 bridges.

built element by element (test_chips/airbridge_crossings.ipynb): XOR 0 on all mask layers; LOM netlist test blocked on the airbridge ground-strap model (open request)

AirbridgeDcTest (airbridge_dc_test)

4-point DC airbridge resistance

22 × AirbridgeDC(width=2000, n_ab=1…22) at test_0…21: (3000,8850), (5000,8850), (7000,8850); rows y = 6875 / 5625 / 4375 / 3125 at x = 2000, 4000, 6000, 8000; then (3000,1150), (5000,1150), (7000,1150). n_ab = index+1.

yes

CrossingTwoface (crossing_twoface)

flip-chip crossing / crosstalk

frames_enabled=[0,1], SMA8 with the same assign as AirbridgeCrossings. 4 routes only (see Routes).

built element by element (test_chips/crossing_twoface.ipynb): face-change waypoints, XOR 0 on both faces, DRC clean

LithographyTest (lithography_test)

optical-litho resolution

For each width w = 1…15 µm and spacing factor s ∈ {1, 5}: StripesTest(num_stripes=20, stripe_width=w, stripe_length=100, stripe_spacing=s·w) and CrossTest(num_crosses=n, cross_spacing=100), grouped as horizontal / vertical (rot 90; crosses rot 90 + mirror) / diagonal (rot −45, no crosses) blocks. Placement: horizontal s=1 at (2000,6500), (3000,6500); s=5 at (4000,6500), (6000,6500); vertical s=1 at (1500,5700), (5500,5700); s=5 at (1500,3800), (5500,3800); diagonal s=1 at (400,1500), (1500,1500); s=5 at (3100,1500), (6100,1500). Inside a block, item i sits at (0, 200i+40) / (200i+140, 0) / (500+100i, 500+100i).

yes (script loops, ≈ 300 instances; KQC uses plain container cells)

LithographyTestTwoface (lithography_test_twoface)

same, both faces

The above on 1t1, plus 6 blocks on face 2b1 (face_ids=["2b1"]) mirrored with DCplxTrans.M90 (= mirror=True, rot=180): h s=1 (5900,4000), h s=5 (8000,4000), v s=1 (5000,6200), v s=5 (5000,4000), d s=1 (7500,1800), d s=5 (6200,1800). frames_enabled=[0,1].

yes (compose transforms per item)

TsvTest (tsv_test)

through-silicon-via process

TsvTestPattern(tsv_array_form=[2,2,2,6,14,2,14,14,2,2,14,14,2,14,6,2,2,2], tsv_diameter=10, tsv_elliptical_width=10) at (5000,5000). Metrology arrays of TsvStandard(tsv_diameter=10): 10 × 10 from (950,950) pitch 60; 8 × 8 from (970,8470) pitch 70; 6 × 6 from (8525,1025) pitch 75. Use d.array.

yes

Stripes (stripes)

ground-grid / litho stripes

Chip-level polygon (cross opening 2000–8000) minus 80 µm-wide stripes at 200 µm pitch and 80 µm squares (axis="Vertical").

yes (placed as the whole KQC chip PCell, see the rebuild below)

Shaping (shaping)

readout with Purcell filter (1 qubit)

Swissmon at (4000,5000) rot 270: arm_width=[30,23,30,23], arm_length=[190,96,160,96], gap_width=[29.5,33,29.5,33], island_r=2, cpl_length=[235,0,205], cpl_width=[60,42,60], cpl_gap=[110,112,110], cl_offset=[150,150]. Two mirrored chains, from cplr0 (4000,5296) up and from cplr2 (4000,4734) down. Each chain has 3 × WaveguideCoplanarSplitter(lengths=[61,61,31]), a J gap cap (FingerCapacitorSquare gap 37.5 / 36.8), a κ interdigital cap (4 × 67.9 / 71.5), a drive gap cap (36.2), a readout resonator (segments 611.586 / 1834.76 / 611.586 µm qubit → short) and a Purcell resonator (3158.32 / 789.581 µm output → short). tunable=True adds Manhattan SQUIDs at the shorted stubs.

built element by element (test_chips/shaping.ipynb): meanders as cpw routes on KQC’s Meander centre line, XOR 0 on mask layers; LOM: one summed λ/4 per chain (see note)

DcTest (dc_test)

DC sample-holder pads

launchers="DC24": 24 × LauncherDC(width=500), gap 300, indent 680, pitch 850. These have no chip ports (DC pads have no corner refpoint).

yes

Empty (empty)

EBL test, ground removed

Chip-level polygon in base_metal_gap_wo_grid + ground_grid_avoidance: a cross from d1 = 300 (dice 200 + margin 100) to d2 = 2000 from the edges.

yes (placed as the whole KQC chip PCell, see the rebuild below)

Launchers (launchers)

bare sample-holder pads

launchers=sampleholder_type (default SMA8; ARD24/DC24 possible).

yes

Simple (simple)

minimal connected chip

SMA8 enabled=["WN","EN","SE"]. Swissmon(cpl_length=[0,150,0]) at (2500,7200) rot 270 (flux W at (2254.67,7200), cplr1 E at (2719,7200)). C1 FingerCapacitorSquare(finger_width=20, finger_length=100, finger_number=7) at (5000,7200); C2 same at (7200,5000) rot 270; T WaveguideCoplanarSplitter(lengths=[11,11,11], port_names=[right,left,bottom]) at (7200,7200).

yes (verified: 6 route lengths = KQC, LOM fully connected)

SampleHolderTest (sample_holder_test)

40-pad holder test

40 RF pads (width 160, gap 96, indent 520, pitch 635), 10 per side, paired by 20 CPWs. Designer launchers= only knows SMA8/ARD24/DC24, so place 40 × Launcher(s=160, l=160, a_launcher=160, b_launcher=96, launcher_frame_gap=96): N (2142.5+635i, 9480) rot 90; E (9480, 7857.5−635i) rot 0; S (7857.5−635i, 520) rot 270; W (520, 2142.5+635i) rot 180.

built element by element (test_chips/sample_holder_test.ipynb): 40 Launcher instances, XOR 0

QualityFactorTwoface (quality_factor_twoface)

flip-chip (capped) spiral λ/4 Q test

frames_enabled=[0,1], SMA8 assign {8: PL-1-IN, 3: PL-1-OUT}. QualityFactor’s lists (res_lengths 5434/5429/5374/5412/5493/5589, n_fingers, l_fingers, type_coupler) with res_a=[10,10,10,20,10,5], res_b=[6,6,6,12,6,3], resonator_types=['capped']*6, resonator_faces=[0,1] (all lines on 1t1, the top chip 2b1 is a lid with its frame only). 6 × WaveguideCoplanarSplitter (lengths 15/15/65, a = b = 10) at x = 2800 + (i+0.5)·816.67, y = 5000, side arm down (even) / up (odd); per T a FingerCapacitorSquare (KQC cap_params, a/b 10/10 → res a/b) aligned port a to the side arm, then a SpiralResonatorPolygon (box 2000 × 500, cap_res_distance 200, bridge_spacing 0, mirrored for odd i) at its port b.

built element by element (test_chips/quality_factor_twoface.ipynb): spirals as routes on KQC’s spiral centre line (DPath params do not fit design.json), XOR < 1 nm slivers only

Routes

  • AirbridgeCrossings: PL-1 chip.PL-1-IN → chip.PL-1-OUT, straight x = 2800. PL-2 is straight at x = 7200. PL-3 runs chip.PL-3-IN (800,7200) → chip.PL-3-OUT (800,2800) and weaves across x = 2800 20 times, with step = 4400 / 9.5 / 2 = 231.58 and wiggle 250. Per pair: (2550,y), wp((2800,y), airbridge=True), (3050,y); y −= step; (3050,y), wp((2800,y), airbridge=True), (2550,y); y −= step. PL-4 runs EN → ES weaving between x = 7500 and 8000 (no crossing).

  • CrossingTwoface (a_capped = b_capped = 10 on the top-chip span): PL-1 NW → SW: (2800,8900), (4300,8900), wp((4300,8500), a=10, b=10), wp((4300,1600), a=10, b=6), (4300,1100), (2800,1100). PL-2 likewise at x = 5700 from NE → SE. PL-3 runs WN → WS: wp((1500,7200), a=10, b=10), then 3 × [(3850,y), wp((4100,y), face_id=”2b1”), wp((4500,y), face_id=”1t1”), (4750,y); y −= 880; mirrored back], then wp((1600,2800), a=10, b=6), (900,2800). This crosses PL-1 on the top chip. PL-4 runs EN → ES: (9100,7200), wp((8500,7200), a=10, b=10), 3 × wiggle 6000 ↔ 6900 (step 880), wp((8400,2800), a=10, b=6), (9100,2800).

  • Simple: WN → Swissmon.flux; Swissmon.cplr1 → C1.a; C1.b → T.left; T.right → EN; T.bottom → C2.a; C2.b → SE. All are straight cpw.

  • Shaping: chargeline WN → drive via (900,7200), (900,5150), (3750,5150), term2=6. Fluxline WS → flux via (900,2800), (900,5000), (3687.67,5000). Per chain: qubit → T_a (meander 1681.76 to T_b) → T_b stub shorted (580.6) → J cap → T_c (stub shorted 758.6) → Purcell meander 2479.24 → κ cap → output to EN (top) / ES (bottom). The drive cap goes from T_a.bottom to NW / SW.

  • SampleHolderTest: PL-i (i = 1…10) N pad (x_i = 2142.5+635(i−1)) → (x_i, 8180) → (8180, y_i = x_i) → E pad. PL-(11+k) (k = 0…9) S pad (x = 7857.5−635k) → (x, 1820) → (1820, 7857.5−635k) → W pad. The middle segments are diagonal and nested.

  • QualityFactorTwoface: probe line PL-1-IN (800,7200) → taper to a/b = 10/10 at (1500,7200) → (2000,7200) → (2000,5000) → T0 … T5 → (8000,5000) → (8000,7200) → taper back to 10/6 at (8400,7200) → PL-1-OUT. KQC’s face nodes at (2300,5000) and (7700,5000) go to face 0, i.e. no face change with the defaults. Each spiral: capacitor port b → KQC’s polygon spiral (one and a half turns in the 2000 × 500 box) → shorted end (term 0).

  • AirbridgeDcTest, JunctionTest2, TsvTest, LithographyTest*, DcTest, Launchers: no routes.

Launchers, pads, test structures, markers, labels

Covered per chip above. Every chip gets the KQC chip frame and markers from the DesignChip base. DcTest/Launchers are the pads themselves.

LOM node structure (qubit ↔ resonator ↔ feedline ↔ bus ↔ control lines)

Only Shaping and Simple have a qubit. In Simple, the Swissmon flux port takes the WN line (a control), and cplr1 feeds C1 → T → EN with C2 to SE. Shaping’s resonators are tapped (T’s inside the λ/4). As built, every piece of a chain is a cpw route (the meanders too, on KQC’s Meander centre line), so the LOM builder makes each chain one λ/4 TLres of the summed route length (the T arms, 306 / 153 µm, are not counted) with the drive and J capacitors on its coupler-end node: right mode structure (qubit, 2 readouts, 2 Purcell filters), rough couplings. A Meander instance or a meander route between two T’s would instead become a λ/2 bus (wrong). QualityFactorTwoface: six λ/4 spiral routes behind finger/gap capacitors on one probe-line conductor between two 50 Ω launchers. The other chips have no circuit LOM: transmission lines only, or DC/process structures (AirbridgeCrossings’ 1280 mechanical-test bridges are ground straps; the LOM Airbridge model is a wire, so their unconnected ports have no node yet).

Targets (if stated)

Shaping: segment_length_target_rr=[611.586, 1834.76, 611.586], segment_length_target_pr=[3158.32, 789.581], caps_length=[37.5, 67.9, 36.2] (J, κ, drive). The others give no targets.

Simplifications for a first rebuild

Build Simple first: it is the smoke test for KQC parity. AirbridgeDcTest, JunctionTest2, TsvTest and DcTest are pure placement. The polygon-only chips (JunctionTest, Stripes, Empty) are placed as the whole KQC chip PCell.

Quantcore PDK rebuild

Each notebook is a Quantcore-PDK rebuild of the source chip. The netlist, LOM and imported components are kept, and the process is Quantcore’s.

simple

  • 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), drawn on the source’s 10 × 10 mm chip (its launchers sit on the chip edges, so the design does not fit the 7.5 mm die). KQC’s ChipFrame (dicing edge, corner markers, labels) is not drawn. KQC’s original grids its whole ground plane (with_grid, 5 µm squares on a 10 µm step). The rebuild keeps that coverage at the PDK density instead (chip parameter cheese: "chip", with_grid: false): 5 µm squares with 20 µm gaps over the whole die, at least 4 µm from the 3/0 keep-outs, inside the design area and clear of the frame markers and the device ID.

  • Stack: sapphire 650 µm, εr 10.25 (the PDK default stack, used by the LOM estimates and every FEM export); KQC’s own simulation default is silicon (εr 11.45), so LOM capacitances and resonator frequencies here are for sapphire.

  • Junctions: Quantcore Nb/AlOx/Nb trilayer junctions (junction_type from layouts.registry.trilayer_junction): fingers on 2/0 + 4/0, spacer 50/0 with the 51–53/0 gradient, sized on the PDK junction rule (Nb Δ = 1.4 meV, I_c·R_n = πΔ/2e, R_n·A 13.5 kΩ·µm² and C_J 50 fF/µm² assumed); the Swissmon’s SQUID becomes a PDK SQUID of two 0.25 × 0.6 µm junctions → R_n = 45 kΩ in parallel, L_J = 6.73 nH at zero flux, C_J = 15 fF (in the 40–50 kΩ window). KQC’s junction_width sets the straight finger. This replaces the assumed Lj 12 nH + 2 fF. Expected f_q ≈ 6.0 GHz (was 4.74 GHz; estimate from the pre-rebuild LOM C_Σ).

  • Kept: the topology and every KQC element with the source’s parameters, positions and orientations, the launcher assignment, every route’s path and length, the sim groups.

shaping

  • 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), drawn on the source’s 10 × 10 mm chip (its launchers sit on the chip edges, so the design does not fit the 7.5 mm die). KQC’s ChipFrame (dicing edge, corner markers, labels) is not drawn. KQC’s original grids its whole ground plane (with_grid, 5 µm squares on a 10 µm step). The rebuild keeps that coverage at the PDK density instead (chip parameter cheese: "chip", with_grid: false): 5 µm squares with 20 µm gaps over the whole die, at least 4 µm from the 3/0 keep-outs, inside the design area and clear of the frame markers and the device ID.

  • Stack: sapphire 650 µm, εr 10.25 (the PDK default stack, used by the LOM estimates and every FEM export); KQC’s own simulation default is silicon (εr 11.45), so LOM capacitances and resonator frequencies here are for sapphire.

  • Junctions: Quantcore Nb/AlOx/Nb trilayer junctions (junction_type from layouts.registry.trilayer_junction): fingers on 2/0 + 4/0, spacer 50/0 with the 51–53/0 gradient, sized on the PDK junction rule (Nb Δ = 1.4 meV, I_c·R_n = πΔ/2e, R_n·A 13.5 kΩ·µm² and C_J 50 fF/µm² assumed); the Swissmon’s SQUID becomes a PDK SQUID of two 0.25 × 0.6 µm junctions → R_n = 45 kΩ in parallel, L_J = 6.73 nH at zero flux, C_J = 15 fF (in the 40–50 kΩ window). KQC’s junction_width sets the straight finger. This replaces the assumed Lj 12 nH + 2 fF. Expected f_q ≈ 6.4 GHz (was 5.17 GHz), below the ≈ 8.5 GHz readout (estimate from the pre-rebuild LOM C_Σ).

  • Kept: the topology and every KQC element with the source’s parameters, positions and orientations, the launcher assignment, every route’s path and length, the sim groups (KQC’s cap_params capacitors, every straight with KQC’s length formula).

airbridge_crossings

  • 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), drawn on the source’s 10 × 10 mm chip (its launchers sit on the chip edges, so the design does not fit the 7.5 mm die). KQC’s ChipFrame (dicing edge, corner markers, labels) is not drawn. KQC’s original grids its whole ground plane (with_grid, 5 µm squares on a 10 µm step). The rebuild keeps that coverage at the PDK density instead (chip parameter cheese: "chip", with_grid: false): 5 µm squares with 20 µm gaps over the whole die, at least 4 µm from the 3/0 keep-outs, inside the design area and clear of the frame markers and the device ID.

  • Stack: sapphire 650 µm, εr 10.25 (the PDK default stack, used by the LOM estimates and every FEM export); KQC’s own simulation default is silicon (εr 11.45), so LOM capacitances and resonator frequencies here are for sapphire.

  • Junctions: none on this chip (no qubits or junctions).

  • Kept: the topology and every KQC element with the source’s parameters, positions and orientations, the launcher assignment, every route’s path and length, the sim groups (the 20 crossings, the 608-CPW / 1280-bridge array; airbridges on 300/0 + 301/0).

crossing_twoface

  • 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), drawn on the source’s 10 × 10 mm chip (its launchers sit on the chip edges, so the design does not fit the 7.5 mm die). KQC’s ChipFrame (dicing edge, corner markers, labels) is not drawn. KQC’s original grids its whole ground plane (with_grid, 5 µm squares on a 10 µm step). The rebuild keeps that coverage at the PDK density instead (chip parameter cheese: "chip", with_grid: false): 5 µm squares with 20 µm gaps over the whole die, at least 4 µm from the 3/0 keep-outs, inside the design area and clear of the frame markers and the device ID. The top chip (face 2b1) gets no frame, because PDK v3 defines none for the flip-chip top chip (its layers export to 800–804/0).

  • Stack: sapphire 650 µm, εr 10.25 (the PDK default stack, used by the LOM estimates and every FEM export); KQC’s own simulation default is silicon (εr 11.45), so LOM capacitances and resonator frequencies here are for sapphire.

  • Junctions: none on this chip (no qubits or junctions).

  • Kept: the topology and every KQC element with the source’s parameters, positions and orientations, the launcher assignment, every route’s path and length, the sim groups (both faces, the face changes and flip-chip connectors).

quality_factor_twoface

  • 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), drawn on the source’s 10 × 10 mm chip (its launchers sit on the chip edges, so the design does not fit the 7.5 mm die). KQC’s ChipFrame (dicing edge, corner markers, labels) is not drawn. KQC’s original grids its whole ground plane (with_grid, 5 µm squares on a 10 µm step). The rebuild keeps that coverage at the PDK density instead (chip parameter cheese: "chip", with_grid: false): 5 µm squares with 20 µm gaps over the whole die, at least 4 µm from the 3/0 keep-outs, inside the design area and clear of the frame markers and the device ID. The top chip (face 2b1) gets no frame, because PDK v3 defines none for the flip-chip top chip (its layers export to 800–804/0).

  • Stack: sapphire 650 µm, εr 10.25 (the PDK default stack, used by the LOM estimates and every FEM export); KQC’s own simulation default is silicon (εr 11.45), so LOM capacitances and resonator frequencies here are for sapphire.

  • Junctions: none on this chip (no qubits or junctions).

  • Kept: the topology and every KQC element with the source’s parameters, positions and orientations, the launcher assignment, every route’s path and length, the sim groups (resonator lengths 5374–5589 µm, the capped-CPW probe line, the top-chip lid).

sample_holder_test

  • 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), drawn on the source’s 10 × 10 mm chip (its launchers sit on the chip edges, so the design does not fit the 7.5 mm die). KQC’s ChipFrame (dicing edge, corner markers, labels) is not drawn. KQC’s original grids its whole ground plane (with_grid, 5 µm squares on a 10 µm step). The rebuild keeps that coverage at the PDK density instead (chip parameter cheese: "chip", with_grid: false): 5 µm squares with 20 µm gaps over the whole die, at least 4 µm from the 3/0 keep-outs, inside the design area and clear of the frame markers and the device ID.

  • Stack: sapphire 650 µm, εr 10.25 (the PDK default stack, used by the LOM estimates and every FEM export); KQC’s own simulation default is silicon (εr 11.45), so LOM capacitances and resonator frequencies here are for sapphire.

  • Junctions: none on this chip (no qubits or junctions).

  • Kept: the topology and every KQC element with the source’s parameters, positions and orientations, the launcher assignment, every route’s path and length, the sim groups (KQC’s produce_n_launchers(40, …) and the 20 lines).

airbridge_dc_test

  • 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), drawn on the source’s 10 × 10 mm chip. The KQC chip PCell is placed with frames_enabled=[], so its own ChipFrame (dicing edge, markers, labels) is not drawn. KQC’s original grids its whole ground plane (with_grid, 5 µm squares on a 10 µm step). The rebuild keeps that coverage at the PDK density instead (chip parameter cheese: "chip", with_grid: false): 5 µm squares with 20 µm gaps over the whole die, at least 4 µm from the 3/0 keep-outs, inside the design area and clear of the frame markers and the device ID.

  • Stack: sapphire 650 µm, εr 10.25 (the PDK default stack, used by the LOM estimates and every FEM export); KQC’s own simulation default is silicon (εr 11.45), so LOM capacitances and resonator frequencies here are for sapphire.

  • Junctions: none on this chip (no qubits or junctions).

  • Kept: the whole KQC chip PCell with the source’s parameters (KQC’s own build()).

empty

  • 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), drawn on the source’s 10 × 10 mm chip. The KQC chip PCell is placed with frames_enabled=[], so its own ChipFrame (dicing edge, markers, labels) is not drawn. KQC’s original grids its whole ground plane (with_grid, 5 µm squares on a 10 µm step). The rebuild keeps that coverage at the PDK density instead (chip parameter cheese: "chip", with_grid: false): 5 µm squares with 20 µm gaps over the whole die, at least 4 µm from the 3/0 keep-outs, inside the design area and clear of the frame markers and the device ID: here the four ground corners outside the cross opening and the band along the die edge (the opening itself has no metal).

  • Stack: sapphire 650 µm, εr 10.25 (the PDK default stack, used by the LOM estimates and every FEM export); KQC’s own simulation default is silicon (εr 11.45), so LOM capacitances and resonator frequencies here are for sapphire.

  • Junctions: none on this chip (no qubits or junctions).

  • Kept: the whole KQC chip PCell with the source’s parameters (KQC’s own build()).

junction_test

  • 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), drawn on the source’s 10 × 10 mm chip. The KQC chip PCell is placed with frames_enabled=[], so its own ChipFrame (dicing edge, markers, labels) is not drawn. The PDK cheese (5 µm squares with 20 µm gaps, at least 4 µm from the 3/0 keep-outs) covers the whole die (chip parameter cheese: "chip"): the ground round the cross, its corners included; the 80 µm squares inside the cross stay solid.

  • Stack: sapphire 650 µm, εr 10.25 (the PDK default stack, used by the LOM estimates and every FEM export); KQC’s own simulation default is silicon (εr 11.45), so LOM capacitances and resonator frequencies here are for sapphire.

  • Junctions: none on this chip (the KQC cell draws EBL test squares, no junctions).

  • Kept: the whole KQC chip PCell with the source’s parameters (KQC’s own build()).

lithography_test

  • 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), drawn on the source’s 10 × 10 mm chip. The KQC chip PCell is placed with frames_enabled=[], so its own ChipFrame (dicing edge, markers, labels) is not drawn. The PDK cheese (5 µm squares with 20 µm gaps, at least 4 µm from the 3/0 keep-outs) covers the whole die (chip parameter cheese: "chip").

  • Stack: sapphire 650 µm, εr 10.25 (the PDK default stack, used by the LOM estimates and every FEM export); KQC’s own simulation default is silicon (εr 11.45), so LOM capacitances and resonator frequencies here are for sapphire.

  • Junctions: none on this chip (no qubits or junctions).

  • Kept: the whole KQC chip PCell with the source’s parameters (KQC’s own build()).

lithography_test_twoface

  • 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), drawn on the source’s 10 × 10 mm chip. The KQC chip PCell is placed with frames_enabled=[], so its own ChipFrame (dicing edge, markers, labels) is not drawn. The PDK cheese (5 µm squares with 20 µm gaps, at least 4 µm from the 3/0 keep-outs) covers the whole die (chip parameter cheese: "chip"). The top chip (face 2b1) gets no frame, because PDK v3 defines none for the flip-chip top chip (its layers export to 800–804/0).

  • Stack: sapphire 650 µm, εr 10.25 (the PDK default stack, used by the LOM estimates and every FEM export); KQC’s own simulation default is silicon (εr 11.45), so LOM capacitances and resonator frequencies here are for sapphire.

  • Junctions: none on this chip (no qubits or junctions).

  • Kept: the whole KQC chip PCell with the source’s parameters (KQC’s own build()).

stripes

  • 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), drawn on the source’s 10 × 10 mm chip. The KQC chip PCell is placed with frames_enabled=[], so its own ChipFrame (dicing edge, markers, labels) is not drawn. The PDK cheese (5 µm squares with 20 µm gaps, at least 4 µm from the 3/0 keep-outs) covers the whole die (chip parameter cheese: "chip"): the ground round the cross, its corners included, and a row of holes down the middle of each wider metal stripe inside the cross.

  • Stack: sapphire 650 µm, εr 10.25 (the PDK default stack, used by the LOM estimates and every FEM export); KQC’s own simulation default is silicon (εr 11.45), so LOM capacitances and resonator frequencies here are for sapphire.

  • Junctions: none on this chip (no qubits or junctions).

  • Kept: the whole KQC chip PCell with the source’s parameters (KQC’s own build()).

tsv_test

  • 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), drawn on the source’s 10 × 10 mm chip. The KQC chip PCell is placed with frames_enabled=[], so its own ChipFrame (dicing edge, markers, labels) is not drawn. KQC’s original grids its whole ground plane (with_grid, 5 µm squares on a 10 µm step). The rebuild keeps that coverage at the PDK density instead (chip parameter cheese: "chip", with_grid: false): 5 µm squares with 20 µm gaps over the whole die, at least 4 µm from the 3/0 keep-outs, inside the design area and clear of the frame markers and the device ID. The top chip (face 2b1) gets no frame, because PDK v3 defines none for the flip-chip top chip (its layers export to 800–804/0).

  • Stack: sapphire 650 µm, εr 10.25 (the PDK default stack, used by the LOM estimates and every FEM export); KQC’s own simulation default is silicon (εr 11.45), so LOM capacitances and resonator frequencies here are for sapphire.

  • Junctions: none on this chip (no qubits or junctions).

  • Kept: the whole KQC chip PCell with the source’s parameters (KQC’s own build()).

junction_test2

  • 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), drawn on the source’s 10 × 10 mm chip. The KQC chip PCell is placed with frames_enabled=[], so its own ChipFrame (dicing edge, markers, labels) is not drawn. KQC’s original grids its whole ground plane (with_grid, 5 µm squares on a 10 µm step). The rebuild keeps that coverage at the PDK density instead (chip parameter cheese: "chip", with_grid: false): 5 µm squares with 20 µm gaps over the whole die, at least 4 µm from the 3/0 keep-outs, inside the design area and clear of the frame markers and the device ID.

  • Stack: sapphire 650 µm, εr 10.25 (the PDK default stack, used by the LOM estimates and every FEM export); KQC’s own simulation default is silicon (εr 11.45), so LOM capacitances and resonator frequencies here are for sapphire.

  • Junctions: Quantcore Nb/AlOx/Nb trilayer junctions (junction_type from layouts.registry.trilayer_junction): fingers on 2/0 + 4/0, spacer 50/0 with the 51–53/0 gradient, sized on the PDK junction rule (Nb Δ = 1.4 meV, I_c·R_n = πΔ/2e, R_n·A 13.5 kΩ·µm² and C_J 50 fF/µm² assumed); every test SQUID is a Quantcore SQUID (junction_type="Quantcore Squid" on the chip cell): KQC’s stepped junction_width is the straight finger, the bent finger stays 0.6 µm. KQC’s large-loop sweep starts at 0.08 µm, below the PDK minimum finger (0.1 µm), so it starts at 0.10 µm here (step 0.03 µm kept); the small-loop sweep keeps 0.15 µm + 0.01 µm steps. Per SQUID, R_n then spans ≈ 75 → 21 kΩ (small loops, 0.15–0.53 µm) and ≈ 113 → 12 kΩ (large loops, 0.10–0.94 µm), across the 40–50 kΩ window (section 4).

  • Kept: the whole KQC chip PCell with the source’s parameters (KQC’s own build()) (the six arrays, pad sizes, loop areas, the step sizes).