Four-qubit ring (Metal 51)

Four transmon qubits sit in a diamond and are coupled to each other through a ring of four bus resonators. Each qubit has its own readout resonator, which ends on its own coupled bond pad (there is no shared feedline), and its own charge line for driving it. The layout comes from Qiskit Metal’s four-qubit example (tutorial 51 / 2.21, Apache-2.0), rebuilt on the Quantcore PDK: sapphire, Quantcore trilayer junctions, the QCT die frame and bond pads 100 µm from the edge.

Build this one to learn λ/2 readouts that end on a coupled pad, qubit–qubit coupling through λ/2 buses, and charge lines routed through corner points.

The whole m51 chip as the v2 notebook builds it

At a glance

Notebook

examples/designs/metal/m51_four_qubit_chip.ipynb (an older one-cell version is kept as m51_four_qubit_chip.v1.ipynb)

Source

Qiskit Metal 51-Four_qubit_chip.ipynb (same layout as tutorial 2.21), © IBM and contributors, Apache-2.0

Chip

9 × 6.5 mm (the source’s size, larger than the 7.5 mm Quantcore die), CPW 10 / 6 µm

On the chip

4 pocket transmons · 4 λ/2 readouts (8.6 mm) · 4 λ/2 buses (9 mm) · 4 charge lines · 8 bond pads

Builds today

yes (layout built on 2026-11-19 in 12 s)

How the notebook builds it

The eight layout segments of m51, one picture each
  1. Die. An empty 9 × 6.5 mm 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.

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

  4. Qubits. Four Metal pocket transmons with Quantcore junctions (L_J 6.73 nH), rotated so their pads face.

  5. Readout resonators. One 8.6 mm λ/2 serpentine from each qubit to its launchpad.

  6. Coupling. Four 9 mm λ/2 buses close the ring between neighbouring qubits.

  7. Charge lines. One line from each qubit’s charge pin to its corner pad.

  8. Whole chip. 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, 23 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

7.126

-437

ol1

7.012

2.74

86.4

Q2+ol2

Q2

7.126

-437

ol2

7.012

2.74

86.4

Q3+ol3

Q3

7.126

-437

ol3

7.012

2.74

86.4

Q4+ol4

Q4

7.126

-437

ol4

7.012

2.74

86.4

group

pair

ZZ (kHz)

neighbours

hybridised

Q1+cpw1+Q4

Q1-Q4

–

yes

yes

Q3+cpw2+Q4

Q3-Q4

–

yes

yes

Q3+cpw3+Q2

Q2-Q3

–

yes

yes

Q1+cpw4+Q2

Q1-Q2

–

yes

yes

LOM plot of m51_four_qubit_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.

Source: Quantum Metal docs/circuit-examples/F.Small-quantum-chips/51-Four_qubit_chip.ipynb. The same layout, cell for cell, is in docs/tut/2-From-components-to-chip/2.21-Design-a-4-qubit-full-chip.ipynb, so it only needs building once. Licence Apache-2.0 (© 2017 IBM and contributors). The chip has four TransmonPocketCL qubits in a diamond, joined by a ring of four 9 mm λ/2 bus meanders. Each qubit is read out through an 8.6 mm λ/2 meander that ends on a capacitively coupled launchpad, and has a charge line to a corner pad. Chip 9000 × 6500 µm. Notebook 51 sets size_y twice, so size_x keeps the 9 mm default; 2.21 sets 9 × 6.5 mm explicitly. The Metal origin is the chip centre, so Designer = Metal + (4500, 3250). CPW 10/6 µm everywhere, route fillet 99.99 µm. Metal layer 1 only. The junctions are FakeJunction_01 placeholders. The Metal substrate is the default silicon, 750 µm thick. The first build uses these substitutions: Xmon for TransmonPocketCL, BondPad + KQC FingerCapacitorSquare for LaunchpadWirebondCoupled, and free-end charge lines.

Components

Metal TransmonPocketCL options (same for all four qubits): pads 455 × 90 µm, pad_gap 30, pocket 650 × 650. The connection pads are a (loc_W +1, loc_H −1, pad_width 70), b (−1, −1, 125) and c (−1, +1, 110), all with cpw_extend 50. The charge line has cl_off_center −50 and cl_pocket_edge 180. First build: Xmon(cross_length=300, claws=["n","w","e"]), with claw_n as the readout claw facing outward. Faithful build: metal:TransmonPocketCL with the options above.

instance

catalog ref

key params

at (x, y)

rot

notes

Q1

Xmon (metal:TransmonPocketCL)

see above

(6920, 3319)

270 (Metal 0)

claw_n→E readout; claw_w→N bus to Q4 (Metal c); claw_e→S bus to Q2 (Metal b)

Q2

Xmon

(4500, 2392.4)

180 (Metal 270)

claw_n→S readout; claw_w→E bus to Q1 (c); claw_e→W bus to Q3 (b)

Q3

Xmon

(2080, 3319)

90 (Metal 180)

claw_n→W readout; claw_e→N bus to Q4 (b); claw_w→S bus to Q2 (c)

Q4

Xmon

(4500, 4107.6)

0 (Metal 90)

claw_n→N readout; claw_e→E bus to Q1 (b); claw_w→W bus to Q3 (c)

P1_C … P4_C

BondPad (metal:LaunchpadWirebond)

pad_width 80, pad_height 80, pad_gap 58, taper_length 122 (= Metal default pad)

(8045, 6062), (8045, 438), (955, 438), (955, 6062)

180, 0, 0, 180

charge-line pads, lead_length 0

P1_Q … P4_Q

BondPad (metal:LaunchpadWirebondCoupled)

same pad params

(8520, 3250), (3510, 438), (480, 3250), (5490, 6062)

90, 0, 270, 180

readout pads; Metal pos = taper tip

C1Q … C4Q

KQC FingerCapacitorSquare

defaults (port a at −x, b at +x)

(8420, 3250), (3510, 538), (580, 3250), (5490, 5962)

0, 90, 0, 90

stands in for the coupled pad’s gap capacitor; sits at the Metal tie pin

All four junctions point to the chip centre (south arm after rotation). Metal pin positions, for reference, in the Designer frame: Q1 a (7295, 3124) E, b (6545, 3124) W, c (6545, 3514) W, CL (7337, 3364) E. Q2 a (4305, 2017) S, b (4305, 2767) N, c (4695, 2767) N, CL (4545, 1975) S. Q3 a (1705, 3514) W, b (2455, 3514) E, c (2455, 3124) E, CL (1663, 3274) W. Q4 a (4695, 4483) N, b (4695, 3733) S, c (4305, 3733) S, CL (4455, 4525) N.

Routes

route

from → to

type, length

waypoints / end

notes

cpw1

Q1.claw_w → Q4.claw_e

meander 9000

auto

Metal Q1.c→Q4.b, lead 100/250, asymmetry −413.75

cpw2

Q3.claw_e → Q4.claw_w

meander 9000

auto

Metal Q3.b→Q4.c, asymmetry +413.75

cpw3

Q3.claw_w → Q2.claw_e

meander 9000

auto

Metal Q3.c→Q2.b, asymmetry −551.75

cpw4

Q1.claw_e → Q2.claw_w

meander 9000

auto

Metal Q1.b→Q2.c, asymmetry +551.75

ol1

Q1.claw_n → C1Q.a

meander 8600

auto

λ/2 readout; Metal lead 430/0, asym +150

ol2

Q2.claw_n → C2Q.b

meander 8600

auto

lead 535/0, asym +200

ol3

Q3.claw_n → C3Q.b

meander 8600

auto

lead 430/0, asym +150

ol4

Q4.claw_n → C4Q.a

meander 8600

auto

lead 535/0, asym +200

t1 … t4

C1Q.b→P1_Q.a, C2Q.a→P2_Q.a, C3Q.a→P3_Q.a, C4Q.b→P4_Q.a

cpw ≈ 60–100

—

cap-to-pad stubs

CL1

P1_C.a → free_end((7337, 3364), dir=0, term=6)

cpw 3277

nodes [(8045, 5837), (7457, 5837), (7457, 3364)]

Metal RouteAnchors, reversed

CL2

P2_C.a → free_end((4545, 1975), dir=-90, term=6)

cpw 4866

[(8045, 663), (6500, 663), (6500, 1775), (4545, 1775)]

Metal anchor (2, −2.5) mm

CL3

P3_C.a → free_end((1663, 3274), dir=180, term=6)

cpw 3415

[(955, 663), (1543, 663), (1543, 3274)]

CL4

P4_C.a → free_end((4455, 4525), dir=90, term=6)

cpw 4866

[(955, 5837), (2500, 5837), (2500, 4725), (4455, 4725)]

Metal anchor (−2, 2.5) mm

The meander lengths are Metal total_length, the full route length from pin to pin. A free end’s dir equals the Metal Charge_Line pin normal, and term=6 gives an open end (Metal charge lines end open in their own etch pocket).

Launchers, pads, test structures, markers, labels

There are 8 pads: 4 corner charge-line pads (P*_C) and 4 edge readout pads (P*_Q, east/south/west/north). The source has no markers, labels or test structures. Use the chip defaults (with_grid False).

LOM node structure

  • Q1–Q4: each has an island with C to ground and a JJ. Each claw is a coupler node (C_g).

  • Bus ring: B1 Q1–Q4, B2 Q3–Q4, B3 Q3–Q2, B4 Q1–Q2. Each is a λ/2 CPW of 9 mm with C_g at both claws.

  • Readout: R1–R4 are λ/2 CPWs of 8.6 mm. Each has C_g at the qubit claw and a series C (FingerCapacitorSquare) to a 50 Ω port P*_Q. There is no shared feedline.

  • Control: 4 open charge lines, each a small C_d to the island and a 50 Ω port P*_C. There are no flux lines.

Targets (if stated)

None are stated. No Lj is set, so take the Metal renderer default, Lj 10 nH and Cj 0: junction={"Lj_nH": 10, "Cj_fF": 2} (Cj 2 fF as in the Metal LOM examples). Rough λ/2 estimates, not targets: 9 mm bus ≈ 7.0 GHz on sapphire (6.6 on Si), 8.6 mm readout ≈ 7.3 GHz (6.9 on Si).

Simplifications for a first rebuild

  1. Use Xmon for TransmonPocketCL, rotated as in the table, so claw_n faces the readout pad and the two bus claws face their neighbours. Ignore the pocket’s two-pads-per-side geometry.

  2. Replace each coupled launchpad with a BondPad plus a FingerCapacitorSquare at the Metal tie pin. This builds a λ/2 between the claw and the series cap.

  3. Make each charge line a free end (term=6) at the Metal pin. The Xmon readout claw sits on the qubit axis, so if the DRC flags CL1–CL4 against the readout meander, shift those free ends 150 µm away from the readout axis.

  4. Drop the meander asymmetries; set only lengths. Add nodes only if a meander collides.

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/m51_four_qubit_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), drawn on the source’s 9 × 6.5 mm chip (the design does not fit the 7.5 mm die); no KQC chip frame.

  • 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 FakeJunction_01 has no value). Expected f_q ≈ 7.1 GHz (an estimate), above the ≈ 6.85 GHz λ/2 buses.

  • 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 9000 µm, readouts 8600 µm, the four charge lines).