Full chip (Metal A.4)

A six-qubit teaching chip with a central hub qubit, a tunable coupler between two flux-tunable qubits, three flux lines and a multiplexed feedline along the north edge. It is the Qiskit Metal “A.4” full-chip example (Apache-2.0, © IBM and contributors), whose resonator lengths are only roughly chosen. The Quantcore rebuild is distributed under GPL-3.0.

Build this one to see a whole processor-style chip assembled in the order a designer would draw it: feedline, qubits, readout, coupling, then flux lines. It mixes five component types and shows tunable coupling, series-capacitor readout and multiplexed readout in one layout.

The whole a4 chip as the v2 notebook builds it

At a glance

Notebook

examples/designs/metal/a4_full_chip.ipynb (an older one-cell version is kept as a4_full_chip.v1.ipynb)

Source

Qiskit Metal A.4-Full-chip-design.ipynb, © IBM and contributors, Apache-2.0

Chip

12.2 × 9.2 mm (the source’s 11 × 8 mm grown by 0.6 mm per side), CPW 10 / 6 µm, flux lines 5 / 3 µm

On the chip

6 qubits (hub Q_Main, 2 crossmons with flux lines, 3 pocket transmons) · 1 tunable coupler · 6 readouts · 7 buses · 3 flux lines · 9 launchpads

Builds today

yes (layout built on 2026-11-19 in 14 s)

How the notebook builds it

The layout segments of a4, one picture each
  1. Die. An empty chip with the QCT frame (dicing marks, alignment grids, chip boundary, device ID).

  2. Bond pads. Nine launchpads on the south, west and east edges, slid to the 100 µm pad-edge rule (the picture zooms on one pad).

  3. Feedline. The north feedline with three tees and the series capacitor of the hub qubit.

  4. Qubits. Six qubits, each with a Quantcore trilayer junction, and the tunable coupler with its SQUID.

  5. Readout resonators. Six fixed-length resonators, one λ/4 and five λ/2, to the feedline tees, the capacitor and the coupled pads.

  6. Coupling. The tunable coupler between the two crossmons and the seven buses around the hub.

  7. Flux lines. Three 5 / 3 µm lines to the two crossmons and the coupler.

  8. Whole chip. The target for the hub qubit, sim groups for the FEM jobs, the content centred in the die, the design saved.

LOM estimates

Numbers are from a run of this notebook on 2026-11-19 (Setup, Design, Check and the LOM cells, 31 s): QuCat normal modes with QuTiP dressed values.

group

qubit

f_q (GHz)

α (MHz)

resonator

f_r (GHz)

κ (MHz)

χ (MHz)

Q_Main+read

Q_Main

4.977

-376

Read_Q_Main

10.49

2.05

-0.181

Q1+read

Q1

5.787

-199

Read_Q1

8.864

5.58

-0.016

Q2+read

Q2

5.787

-199

Read_Q2

11.04

10.7

-0.0102

TC+read

Tune_C_Q12

8.334

-440

Read_TuneC

10.55

9.32

-6.78

Q3+read

Q3

7.197

-319

Q3_Read_T

5.665

0.0876

-0.98

Q4+read

Q4

7.093

-310

Read_Q4

10.23

4.21

-0.513

Q5+read

Q5

7.193

-320

Read_Q5

10.7

37.4

-0.429

group

pair

ZZ (kHz)

neighbours

hybridised

Q_Main+Bus_01+Q1

Q1-Q_Main

-0.108

yes

no

Q_Main+Bus_02+Q2

Q2-Q_Main

-0.133

yes

no

Q_Main+Bus_03+Q3

Q3-Q_Main

-1.43

yes

no

Q_Main+Bus_04+Q4

Q4-Q_Main

-1.52

yes

no

Q_Main+Bus_05+Q5

Q5-Q_Main

-6

yes

no

Q3+Bus_Q3_Q4+Q4

Q3-Q4

46.9

yes

no

Q4+Bus_Q4_Q5+Q5

Q4-Q5

75

yes

no

Q1+TC+Q2

Q1-Q2

–

no

yes

Q1+TC+Q2

Q1-Tune_C_Q12

–

yes

yes

Q1+TC+Q2

Q2-Tune_C_Q12

–

yes

yes

LOM plot of a4_full_chip

These are lumped-model estimates from 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: every component, port, route and length.

The metal: references in this spec build as written, through the metal: wrapper (ports are the Metal pin names, parameters the Metal options). The Quantcore and KQC substitutions are optional. The full chip builds with metal:TransmonPocket6, metal:TunableCoupler01 (pins Control / Flux) and metal:CapNInterdigitalTee; subset A below is the quickest first build.

Source: Quantum Metal docs/circuit-examples/full-design-flow-examples/A.4-Full-chip-design.ipynb. Licence Apache-2.0. It is a teaching layout (“resonator lengths only roughly chosen”) that is taken through Q3D/Elmer LOM, HFSS EPR and GDS. Q_Main is a TransmonPocket6 with 6 pads, bused to five qubits. Q1 and Q2 are TransmonCrossFL with flux lines, coupled through a TunableCoupler01. Q3, Q4 and Q5 are TransmonPocket6 in a bus chain and are multiplexed on a north feedline through a CoupledLineTee (λ/4) and two CapNInterdigitalTee (λ/2). Q_Main, Q1, Q2 and TC are each read out separately. Chip 11 000 × 8000 µm, centred at (0, −0.5) mm, so Designer = Metal + (5500, 4500). CPW 10/6 µm; flux lines 5/3 µm. Fillet 99 µm. The junctions are FakeJunction_01 cells. Q_Main needs 6 couplers and Q4 needs 4, and no Quantcore/KQC qubit has more than 3, so the full chip needs metal:TransmonPocket6 and metal:TunableCoupler01. The north block (subset A below) builds from Quantcore and KQC parts.

Components

TransmonPocket6 base: pads 455 × 90, pad_gap 30, pocket 650 × 650. Connection pads at loc_W 0 have vertical pins at (0, ±425); those at loc_W ±1 have pins at (±425, ±120).

instance

catalog ref

key params (µm)

at (x, y)

rot

pins (Designer frame)

Q_Main

metal:TransmonPocket6

pad_width 425, pocket_height 650; readout (0, −1, w 80, gap 50); bus_01 (−1, −1, 60, 10); bus_02 (−1, +1, 60, 10); bus_03 (0, +1, 90, 30); bus_04 (+1, +1, 60, 10); bus_05 (+1, −1, 60, 10); hfss Lj 14 nH

(5500, 3500)

0

readout (5500, 3075) S; bus_01 (5075, 3380) W; bus_02 (5075, 3620) W; bus_03 (5500, 3925) N; bus_04 (5925, 3620) E; bus_05 (5925, 3380) E

Q1

Xmon(cross_length=400, arm_width=20, gap=20, claws=["w","e"], flux_port=True) (metal:TransmonCrossFL)

bus_01 claw E, claw_length 95; readout claw W

(2750, 2700)

0

Metal bus_01 (3031, 2700) E, readout (2469, 2700) W, flux (2750, 2441.5) S

Q2

Xmon(... flux_port=True)

bus_02 claw (Metal loc 0 → E after 180°), claw_length 95

(2750, 3300)

180

Metal bus_02 (3031, 3300) E, readout (2469, 3300) W, flux (2750, 3558.5) N

Tune_C_Q12

metal:TunableCoupler01

c_width 500 (others default: l_width 20, l_gap 10, a_height 60, fl 5/3/10)

(2690, 3000)

90

Control (2596, 3048) W, Flux (2602.5, 2856) W

Q3

metal:TransmonPocket6, or Xmon(claws=["n","e"]) for subset A

bus_03 (0, −1, 80, 15), bus_q3_q4 (+1, −1, 80, 15), readout (0, +1, 80, 50); Lj 14 nH

(2500, 5000)

0

bus_03 (2500, 4575) S, bus_q3_q4 (2925, 4875) E, readout (2500, 5425) N

Q4

metal:TransmonPocket6, or Xmon(claws=["n","w","e"])

bus_04 (0, −1), bus_q3_q4 (−1, −1), bus_q4_q5 (+1, −1), readout (0, +1, 80, 50)

(5500, 5500)

0

bus_04 (5500, 5075) S, q3_q4 (5075, 5375) W, q4_q5 (5925, 5375) E, readout (5500, 5925) N

Q5

metal:TransmonPocket6, or Xmon(claws=["n","w"])

bus_05 (0, −1), bus_q4_q5 (−1, −1), readout (0, +1)

(8500, 5000)

0

bus_05 (8500, 4575) S, q4_q5 (8075, 4875) W, readout (8500, 5425) N

Read_Q_Main_Cap

KQC FingerCapacitorSquare (metal:CapNInterdigital, default 5 × 20 µm fingers)

(7500, 927)

90

Metal north_end (7500, 1000), south_end (7500, 854)

Q3_Read_T

hanger (metal:CoupledLineTee)

coupling_length 200, space 3, down 100, short

centre (2500, 7500)

—

λ/4

Q4_Read_T

hanger end="open" (metal:CapNInterdigitalTee)

finger_length 0, 8 fingers (gap cap, ~122 µm wide)

centre (5500, 7500)

—

λ/2

Q5_Read_T

hanger end="open" (metal:CapNInterdigitalTee)

finger_length 50, 11 fingers (~170 µm)

centre (8500, 7500)

—

λ/2

pads

BondPad(80, 80, gap 58, taper 122); coupled pads add a FingerCapacitorSquare at the Metal tie

QMain_Read (7500, 525) r0; Q1_FL (5500, 525) r0 (5/3); Q1_Read* (3500, 500) r0, cap (3500, 595); TuneC_FL (1500, 525) r0 (5/3); TuneC_Read* (500, 1500) r270, cap (595, 1500); Q2_Read* (500, 3500) r270, cap (595, 3500); Q2_FL (525, 5500) r270 (5/3); NW (525, 7500) r270; NE (10475, 7500) r90

* = LaunchpadWirebondCoupled

Routes

route

from → to

type, length

waypoints / hanger

notes

Bus_01 … Bus_05

Q_Main.bus_0k → Q1.bus_01 / Q2.bus_02 / Q3.bus_03 / Q4.bus_04 / Q5.bus_05

meander 6000, 6400, 6800, 7200, 7600

—

Metal asym 1305 (spacing 250), 450, 50, 150 (start jogs L250, R200), 50. EPR retunes Bus_05 to 7500

Bus_Q3_Q4, Bus_Q4_Q5

Q3→Q4, Q4→Q5

meander 6400, 6800

—

lead 125/125, 125/25

Read_Q_Main

Q_Main.readout → Read_Q_Main_Cap (north port)

meander 5600

—

λ/2; Metal lead 725/625, jogs L250 / end L600

Read_Q_Main_Cap_Launch

cap (south port) → Launch_QMain_Read.a

cpw ≈ 330

—

Read_Q1, Read_Q2

Q1 readout claw → cap at Launch_Q1_Read; Q2 → cap at Launch_Q2_Read

meander 6800, 5400

—

λ/2; Read_Q2 asym −450

Read_TuneC

TC.Control → cap at Launch_TuneC_Read

meander 5800

—

start lead 1525

Flux_Line_Q1

Launch_Q1_FL.a → Q1.flux

cpw 4582, a 5 b 3

nodes [(5500, 1483), (2750, 1483)]

Metal RoutePathfinder, reversed

Flux_Line_TuneC

Launch_TuneC_FL.a → TC.Flux

cpw 3306, 5/3

[(1500, 875), (1728, 875), (1728, 2856)]

Flux_Line_Q2

Launch_Q2_FL.a → Q2.flux

cpw 3954, 5/3

[(675, 5500), (675, 4358), (2225, 4358), (2225, 3734), (2750, 3734)]

FL_N

Launch_NW.a → Launch_NE.a

cpw 9950, straight y = 7500

—

north mux feedline (Metal ML_TL_01…04 through the three tees)

Read_Q3

Q3 readout → FL_N.hanger(s=1975, length=200, side="right", ground_gap=3, direction="backward", end="short")

meander 5300

—

Metal 5000 + 300 tee

Read_Q4

Q4 readout → FL_N.hanger(s=4975, length=120, side="right", end="open")

meander ≈ 5900

—

Metal 5800 + tee

Read_Q5

Q5 readout → FL_N.hanger(s=7975, length=170, side="right", end="open")

meander ≈ 5600

—

Metal 5400 + tee

Launchers, pads, test structures, markers, labels

11 pads: 4 on the south edge (QMain_Read, Q1_FL, Q1_Read*, TuneC_FL), 3 on the west (TuneC_Read*, Q2_Read*, Q2_FL), and NW/NE on the north feedline. There are no markers.

LOM node structure

  • Q_Main is the hub. It has a C_g to each of the 5 λ/2 buses: Q1 (5.6 GHz), Q2 (5.7), Q3 (5.8), Q4 (5.9), Q5 (6.0 per the LOM setup).

  • Q3–Q4–Q5 form a bus chain. Q1 and Q2 are coupled through TC (a SQUID island).

  • Readouts: R_QMain is λ/2 with a series C to its port. R_Q1, R_Q2 and R_TC are λ/2 with coupled-pad C to their ports. R_Q3 is λ/4 (hanger short) on FL_N. R_Q4 and R_Q5 are λ/2 with tee C on FL_N.

  • Flux bias on Q1, Q2 and TC.

Targets (if stated)

  • Q_Main LOM setup: Lj 14 nH, Cj 2 fF, readout 7.0 GHz, buses [5.6, 5.7, 5.8, 5.9, 6.0] GHz. The Q3D + LOM result quoted in the text is f_Q 4.96 GHz, α −351 MHz, g 44–52 MHz.

  • Q3D C reference: readout pad 42.1 fF, pad_top 97.5 fF, pad_top–pad_bot −31.2 fF.

  • EPR run: Lj(Q_Main) 13 nH, Lj(Q5) 15 nH, Bus_05 7.5 mm. Q3–Q5 Lj 14 nH; Q1/Q2 Metal default 10 nH.

Simplifications for a first rebuild

  1. Subset A: FL_N with Q3/Q4/Q5 as Xmon (claw_n = readout; e/w = chain buses), Bus_Q3_Q4, Bus_Q4_Q5, and Read_Q3–Q5. This is a 3-qubit mux test of hanger readout with both short and open ends.

  2. Full chip: use metal:TransmonPocket6 and metal:TunableCoupler01. For the open tees, use a hanger end="open" or, more faithfully, a KQC WaveguideCoplanarSplitter plus FingerCapacitorSquare.

  3. Every coupled launchpad becomes a BondPad plus a FingerCapacitorSquare at the Metal tie pin.

Quantcore PDK rebuild

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

Notebook: examples/designs/metal/a4_full_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 source’s 11 × 8 mm chip grown by 0.6 mm on every side (box (-600, -600, 11600, 8600)) so that the corner launchpads Launch_NW / Launch_NE and the feedline clear the PDK marker areas; 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); Metal’s DesignPlanar default is silicon (εr 11.45, 750 µm), so the LOM capacitances, resonator frequencies and FEM jobs here are for sapphire rather than the source’s substrate.

  • 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). Q_Main is sized for the tutorial’s f_q 4.96 GHz (its C_Σ ≈ 55 fF from the LOM): straight finger 0.23 µm × 0.6 µm, L_J 14.6 nH, C_J 6.9 fF, so R_n ≈ 98 kΩ, above the 40–50 kΩ window (a 45 kΩ junction would put this small pocket at ≈ 7.1 GHz). Q3–Q5 get the PDK nominal junction, 0.5 × 0.6 µm fingers → R_n = 45 kΩ, L_J = 6.73 nH, C_J = 15 fF (in the 40–50 kΩ window) (expected ≈ 7.1–7.6 GHz). Q1, Q2 and the coupler (flux lines) get 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); expected Q1 / Q2 ≈ 5.8 GHz, the coupler ≈ 8.3 GHz at zero flux (above the 4–8 GHz band; it is flux-tuned down).

  • Kept: the topology and every Metal component with the source’s options (through the metal: wrapper), positions and orientations, every route’s end pins and total length, the sim groups (buses 6.0–7.6 mm, readouts 5.0–6.8 mm, the three flux-line paths, the d.couple coupler estimates).