Launch-video chip (Metal A.5)

A ring of four pocket transmons joined by four λ/2 buses. Each qubit has its own λ/2 readout to a coupled bond pad and its own charge line to a corner pad, and all lines use a wider 15 / 9 µm CPW. It is the chip from the Qiskit Metal launch video (Apache-2.0, © IBM and contributors). The Quantcore rebuild is distributed under GPL-3.0.

Build this one to learn how a ring of coupled qubits is laid out with no shared feedline. Every readout ends on its own pad, and the charge lines end on open-to-ground stubs next to each pocket. The bus pad heights differ (30 and 50 µm) to detune neighbouring qubits.

The whole a5 chip as the v2 notebook builds it

At a glance

Notebook

examples/designs/metal/a5_launch_video.ipynb (an older one-cell version is kept as a5_launch_video.v1.ipynb)

Source

Qiskit Metal A.5-Launch-video-example.ipynb, © IBM and contributors, Apache-2.0

Chip

10.2 × 7.4 mm (the source’s 9 × 6.2 mm grown by 0.6 mm per side), CPW 15 / 9 µm

On the chip

4 pocket transmons · 4 λ/2 readouts (5.5 and 13 mm) · 4 λ/2 buses (5.6 / 5.7 mm) · 4 charge lines · 8 bond pads

Builds today

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

How the notebook builds it

The layout segments of a5, one picture each
  1. Die. An empty chip with the QCT frame (dicing marks, alignment grids, chip boundary, device ID).

  2. Bond pads. Four corner pads for the charge lines, slid to the 100 µm pad-edge rule (the picture zooms on one pad).

  3. Readout pads. Four coupled launchpads, one per qubit: a pad with a series capacitor.

  4. Qubits. Four Metal pocket transmons with Quantcore junctions, in a ring.

  5. Readout resonators. Four λ/2 resonators, routed clear of the charge lines.

  6. Coupling. Four λ/2 buses closing the ring between neighbouring qubits.

  7. Charge lines. One line from each corner pad to an open-to-ground stub beside its qubit.

  8. Whole chip. Sim groups for the FEM jobs (the ring core and each qubit with its readout), 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, 19 s): QuCat normal modes with QuTiP dressed values.

group

qubit

f_q (GHz)

α (MHz)

resonator

f_r (GHz)

κ (MHz)

χ (MHz)

Q1+ol1

Q1

6.967

-297

ol1

10.59

9.07

-2.48

Q2+ol2

Q2

6.979

-298

ol2

4.676

0.818

-0.69

Q3+ol3

Q3

6.967

-297

ol3

10.59

9.07

-2.48

Q4+ol4

Q4

6.979

-298

ol4

4.676

0.818

-0.69

group

pair

ZZ (kHz)

neighbours

hybridised

Q1+cpw1+Q2

Q1-Q2

215

yes

no

Q3+cpw2+Q2

Q2-Q3

235

yes

no

Q3+cpw3+Q4

Q3-Q4

215

yes

no

Q1+cpw4+Q4

Q1-Q4

235

yes

no

LOM plot of a5_launch_video

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.

Source: Quantum Metal docs/circuit-examples/full-design-flow-examples/A.5-Launch-video-example.ipynb, first video. The second video re-creates the qubit core only, for an HFSS render. Licence Apache-2.0. This is the same family as 51_four_qubit_chip, with four differences:

  • TransmonPocket without charge lines, each with 3 pads.

  • Wider 15/9 µm CPW.

  • Buses of 5.6/5.7 mm.

  • Charge lines that start at OpenToGround ends 0.35 mm outside each pocket. Readout goes to coupled launchpads at 5.5 mm (E/W) and 13.0 mm (S/N). Chip: DesignPlanar default 9000 × 6000 µm, centred, so Designer = Metal + (4500, 3000). Fillet 90 µm, Metal layer 1, junctions from Fake_Junctions.GDS. A first build uses these substitutions: Xmon for TransmonPocket, BondPad + FingerCapacitorSquare, free ends.

Components

Metal TransmonPocket options: pad_width 425, pocket_height 650; pads a (+1, −1, pad_width 200), b (−1, +1, pad_height 30), c (−1, −1, pad_height 50). First build: Xmon(claws=["n","w","e"]) with claw_n = readout (a).

instance

catalog ref

at (x, y)

rot

claw mapping (Metal pin, Designer frame)

Q1

Xmon (metal:TransmonPocket)

(6922.5, 3000)

270 (Metal 0)

claw_n→E = a (7347.5, 2802.5); claw_w→N = b (6497.5, 3197.5) bus to Q4; claw_e→S = c (6497.5, 2802.5) bus to Q2

Q2

Xmon

(4500, 2050)

180 (Metal 270)

claw_n→S = a (4302.5, 1625); claw_w→E = b (4697.5, 2475) bus to Q1; claw_e→W = c (4302.5, 2475) bus to Q3

Q3

Xmon

(2077.5, 3000)

90 (Metal 180)

claw_n→W = a (1652.5, 3197.5); claw_e→N = c (2502.5, 3197.5) bus to Q4; claw_w→S = b (2502.5, 2802.5) bus to Q2

Q4

Xmon

(4500, 3950)

0 (Metal 90)

claw_n→N = a (4697.5, 4375); claw_e→E = c (4697.5, 3525) bus to Q1; claw_w→W = b (4302.5, 3525) bus to Q3

OTG1 … OTG4

free ends (metal:OpenToGround, width 15, gap 9)

(7271, 3000), (4500, 1699), (1729, 3000), (4500, 4301)

—

charge-line tips; dir 0, −90, 180, 90

P1_C … P4_C

BondPad(80, 80, gap 58, taper 122), trace 15/9

(8500, 5812), (8500, 188), (500, 188), (500, 5812)

180, 0, 0, 180

corner charge pads

P1_Q … P4_Q

BondPad + FingerCapacitorSquare (metal:LaunchpadWirebondCoupled)

pads (8520, 3000), (3510, 188), (480, 3000), (5490, 5812); caps at the tie pins (8450, 3000), (3510, 258), (550, 3000), (5490, 5742)

pads 90, 0, 270, 180; caps 0, 90, 0, 90

readout pads

Routes (all 15/9 µm)

route

from → to

type, length

waypoints

notes

cpw1

Q1.claw_e → Q2.claw_w

meander 5600

—

Metal Q1.c→Q2.b, lead 200/200, asym +140

cpw2

Q3.claw_w → Q2.claw_e

meander 5700

—

Metal Q3.b→Q2.c, asym −140, lead flipped

cpw3

Q3.claw_e → Q4.claw_w

meander 5600

—

Metal Q3.c→Q4.b, asym +140

cpw4

Q1.claw_w → Q4.claw_e

meander 5700

—

Metal Q1.b→Q4.c, asym −140, flipped

ol1, ol3

Q1.claw_n → cap(P1_Q); Q3.claw_n → cap(P3_Q)

meander 5500

—

λ/2 readout; lead 350/150, asym +140, flipped

ol2, ol4

Q2.claw_n → cap(P2_Q); Q4.claw_n → cap(P4_Q)

meander 13 000

—

long λ/2, about 4.5 GHz on Si (estimate)

cap stubs

cap → P*_Q.a

cpw ≈ 70

—

line_cl1

P1_C.a → free_end((7271, 3000), dir=0, term=6)

cpw 3925

nodes [(8500, 5500), (7500, 5500), (7500, 3000)]

Metal RouteAnchors, anchors (3, 0), (3, 2.5), (4, 2.5) mm, reversed

line_cl2

P2_C.a → free_end((4500, 1699), dir=-90, term=6)

cpw 5356

[(8500, 500), (6500, 500), (6500, 1500), (4500, 1500)]

line_cl3

P3_C.a → free_end((1729, 3000), dir=180, term=6)

cpw 3925

[(500, 500), (1500, 500), (1500, 3000)]

line_cl4

P4_C.a → free_end((4500, 4301), dir=90, term=6)

cpw 5356

[(500, 5500), (2500, 5500), (2500, 4500), (4500, 4500)]

Launchers, pads, test structures, markers, labels

8 pads (4 corner charge pads, 4 edge readout pads). The corner pads at ±4000 µm fit inside the 9 mm chip. There are no markers.

LOM node structure

The structure is the same as 51_four_qubit_chip: 4 islands, a ring of 4 λ/2 buses (5.6/5.7 mm), 4 λ/2 readouts with series C to the ports, and 4 open charge lines (C_d).

Targets (if stated)

None. The junctions are Fake_Junctions (Metal default Lj 10 nH). The second half only renders the core to HFSS eigenmode (“FourQ”), with no numbers.

Simplifications for a first rebuild

  1. Reuse the 51 builder and change CPW (a=15, b=9 per route), lengths and pad positions.

  2. The OTG charge tips sit outside the original 650 µm pocket. With the smaller Xmon they are far from the island, so move them 100–150 µm inward if the drive coupling matters.

  3. Ignore lead_direction_inverted (flip); the Designer places its own meanders.

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/a5_launch_video.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 chip (9 × 6.2 mm, see below) grown by 0.6 mm on every side (box (-600, -700, 9600, 6700)) so that the four corner charge-line pads 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); all four qubits 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) (the source’s Fake_Junctions carry no value). Expected f_q ≈ 7.0 GHz (an estimate). The bus pads still detune each bus-coupled pair (30 µm pad b at one qubit, 50 µm pad c at the other), so the pairs do not hybridise.

  • 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 5.6 / 5.7 mm, readouts 5.5 / 13.0 mm, the four charge lines, the 15/9 µm CPW).