DualYbSim

Parameter reference

All parameters live on NoiseModelParameters and are named after the symbols in the paper’s noise-model appendix. See docs/channel_reference.md for the notation-to-identifier convention and for what each channel means.

Layout

Per-encoding parameters are stored flat with a trailing encoding tag:

Tag	Encoding
_c	174Yb optical clock qubit (isotope="174", qubit_type="gm")
_g	171Yb ground-manifold nuclear-spin qubit ("171", "g")
_m	171Yb metastable-manifold nuclear-spin qubit ("171", "m")

Use for_qubit() to read them grouped, with the tag dropped:

params = NoiseModelParameters()
v = params.for_qubit("171", "m")
v.p_1  # == params.p_1_m
v.gate_time  # == params.gate_time_m

The view is built on demand, so it always reflects the current values. A field that reads 0.0 means the channel does not fire for that encoding — either it is physically absent (p_flip_m on the 174Yb clock qubit) or it has been calibrated away — so the usual if p > 0: guard works unchanged.

Rates that are properties of the atom rather than of the encoding are stored once, as shared constants.

Presets

Constructor	Meaning
NoiseModelParameters()	the paper’s tabulated values
NoiseModelParameters.paper_defaults()	identical to the above, explicit
NoiseModelParameters.legacy_defaults()	the legacy parameter set that predates this library

legacy_defaults() exists so that numbers produced before this library existed can be reproduced. It restores parameters only; formula-level corrections stay in place, in particular the factor of two in t_move = 2 sqrt(l / a).

Shared physical constants

Parameter	Paper	Default	Meaning
gamma_Ryd	Γ_Ryd	2.0e4 s⁻¹	total Rydberg decay rate, 1/50 µs
ryd_branching	—	{LOSS_r: 0.51, DECAY_rg: 0.42, DECAY_rm: 0.07}	branching of Γ_Ryd
gamma_gL	Γ_gL	1/30 s⁻¹	ground-manifold trap loss
gamma_mL	Γ_mL	1/30 s⁻¹	metastable-manifold trap loss
gamma_Z_gm	1/T_2^(gm)	1/5 s⁻¹	171Yb clock-transition coherence. Currently inert — see below

Γ_mg is not shared: it is per encoding, because the effective metastable lifetime depends on the trap depth the encoding operates in. See the coherent-control table below.

gamma_Z_gm is set to its tabulated value but no circuit path reads it, so ZERR_gm never reaches a Stim circuit. The channel exists in the Kraus layer (Kraus1Q_171m.ZERR_gm and friends) and can be used directly from there. What is unsettled is when it should fire in a circuit: an idling 171Yb qubit sits wholly in one manifold, so there is no ground–metastable superposition to dephase, and the windows where such a superposition does exist are the clock pulses, which already carry DEP1_gm.

Transport geometry and schedule

Parameter	Paper	Default	Meaning
l_site	l_site	3e-6 m	site spacing
l_zone	l_zone	100e-6 m	computation-to-readout zone separation
t_hand	t_hand	200e-6 s	handover time
a	a	5500 m s⁻²	transport acceleration

Transport time is t_move(l) = 2 sqrt(l / a), and the one-way readout distance is l = d · l_site + l_zone for code distance d.

Per-encoding parameters

— marks a parameter the encoding does not define. Where the legacy column is blank the two presets agree.

Coherent-control

Parameter	Paper	_c	_g	_m	legacy
p_1_*	p_1^(c/g/m)	1e-4	1e-4	1e-4	2e-3 / 1e-3 / 1e-3
p_2_*	p_2^(c/m)	1e-3	1e-3	1e-3	2e-3 (all)
p_2_dual	p_2^(dual)	1e-3 (cross-encoding)			2e-3
p_1_gm	p_1^(gm)	—	1e-4 (shared with _m)	1e-4	2e-3
p_m_g_gate	p_(m→g)^(gate)	—	1e-3 (shared)	1e-3
gamma_mg_*	Γ_mg	1/20	1	1	_c was 1

p_1_gm and p_m_g_gate are 171Yb clock-transition channels stored once and exposed on both 171Yb encodings, because the g qubit reaches the Rydberg state through the same clock pulse. DEP1_gm reaches a g qubit only while that pulse is driving it into the metastable manifold, so it is emitted with the metastable rates for both encodings.

gamma_mg_* is per encoding because the effective metastable lifetime depends on the trap depth. The clock qubit idles in a shallow trap and so uses the natural ³P₀ lifetime of 20 s; the 171Yb channels are reached only through the clock pulse and the readout shelving step, both of which run in the deep imaging trap, and keep the scattering-dominated 1 s⁻¹. The clock qubit’s value must also satisfy T_2 ≤ 2 T_1, since amplitude damping at T_1 alone already decoheres at 2 T_1; this is the one parameter legacy_defaults() does not restore, because the legacy 1 s⁻¹ against T_2^(c) = 5 s does not.

Measurement

Parameter	Paper	_c	_g	_m	legacy
p_meas_*	p_meas	1e-4	1e-4	1e-4	1e-3 / 1e-3 / 1/300
q_BB_*	q_BB	0.5	0.5	0.5
p_g_L_meas_*	p_(g→L)^(meas)	1e-3	1e-3	1e-3
p_flip_g_*	p_flip^(g)	—	1e-3	1e-3
p_depol_meas_idling_*	—	—	0.0	0.0

q_BB_* is the probability of assigning the ambiguous bright–bright readout record to |0>. The twirled p_Z of the MERR channel scales as (2 q_BB − 1)², so it vanishes at the default even split.

p_depol_meas_idling_* is not part of the paper’s channel set. It depolarises an idle 171Yb qubit while 174Yb is measured in place, and is held at zero.

Reset

Parameter	Paper	_c	_g	_m	legacy
p_g_L_reset_*	p_(g→L)^(reset)	1e-3	1e-3	—	_g was 0.0
p_m_L_reset_m	p_(m→L)^(reset)	—	—	6e-3	1e-3
p_flip_m_m	p_flip^(m)	—	—	1e-3

Idling

Parameter	Paper	_c	_g	_m	legacy
gamma_Z_*	1/T_2^(c/g/m)	1/5	1/10	1/10
gamma_X_*	1/T_1^(g/m)	—	1/200	1/200	0.0

Note the dephasing convention for the 171Yb encodings. ZERR_g and ZERR_m are emitted as Z_ERROR(1 − exp(−t/T_2)), following the paper’s channel definition K_1 = √(p_Z) Z. A Pauli-Z channel of probability p maps the transverse Bloch component by 1 − 2p, so the effective coherence time is about T_2/2 rather than T_2: at T_2^(m) = 10 s and t = 1 ms the emitted channel implies T_2,eff = 5.0 s. Set gamma_Z_g / gamma_Z_m accordingly if you want the nominal value to be the effective one.

ZERR_c is not affected. It acts on the same two levels as DECAY_mg, so the two are twirled together into a single PAULI_CHANNEL_1 whose p_Z = p_2(t)/2 − p_1(t)/4 already carries the factor of a half, reproducing exp(−t/T_2^(c)) exactly.

Transportation

Parameter	Paper	_c	_g	_m
p_hand_*	p_(g→L)^(hand), p_(m→L)^(hand)	1e-3	1e-3	1e-3

The paper splits handover loss by manifold; this library splits it by encoding, which is the finer granularity.

Operation times

Seconds. Keys of the gate_time_c / gate_time_g / gate_time_m dicts.

Key	Paper	_c	_g	_m	legacy
t_1Q	t_1Q^(c/g/m)	100e-6	1e-6	1e-6	100e-6 for _g / _m
t_1Q_gm	t_1Q^(gm)	—	10e-6	—	100e-6
t_2Q	t_2Q	300e-9	300e-9	300e-9
t_reset	t_reset	2e-3	2e-3	2e-3
t_read	t_read	1e-3	1e-3	1e-3

Reading and writing parameter files

save_noise_params_to_file / load_noise_params_from_file use a flat name: value text format, with slash notation for the nested dicts:

# 174Yb clock qubit
p_1_c: 0.0001
gate_time_c/t_1Q: 0.0001

set_parameters accepts the same slash notation and rejects unknown names:

params.set_parameters(p_1_c=3e-4, gamma_mg=1.5)
params.set_parameters(**{"gate_time_c/t_1Q": 200e-6})

rescale_error_params(scale) multiplies every parameter whose name starts with p_ or gamma_, leaving ratios (q_BB_*), geometry (l_site, l_zone, a) and times (t_hand, gate_time_*) untouched.

Mapping from the legacy parameter names

Legacy	Here
p_gm_174	p_1_c
p_g_171	p_1_g
p_m_171	p_1_m
p_gm_171	p_1_gm
p_2_dep	p_2_c, p_2_g, p_2_m, p_2_dual
p_MG_gate_171m	p_m_g_gate
p_MERR_174	p_meas_c
p_MERR	p_meas_g
p_MERR_171m	p_meas_m
q_MERR_174, q_MERR_171m	q_BB_c, q_BB_m
p_GLOSSR_174/171g/171m	p_g_L_meas_c/_g/_m
p_GLOSSP_174/171g	p_g_L_reset_c/_g
p_MLOSSP_171	p_m_L_reset_m
p_X_g_171	p_flip_g_g
p_X_meas_171m	p_flip_g_m
p_X_motional_171	p_flip_m_m
p_handover_loss_174/171g/171m	p_hand_c/_g/_m
gamma_Zidl_174/171g/171m	gamma_Z_c/_g/_m
gamma_MG_174	gamma_mg_c
gamma_MG_idl_171m	gamma_mg_g, gamma_mg_m
gamma_RL	gamma_Ryd
RL_ratio	ryd_branching
site_separation, zone_separation	l_site, l_zone
handover_time, acceleration	t_hand, a
gate_time_174/171g/171m	gate_time_c/_g/_m
gate_1q, gate_1q_clock, gate_2q, reset, measure	t_1Q, t_1Q_gm, t_2Q, t_reset, t_read
p_X_mr_174, p_X_mr_171	removed with the XERR_mr channel
alpha	removed; MERR p_Z uses ((2 q_BB − 1)²/16) p_meas²
T_2_174	removed; nothing read it

This site is open source. Improve this page.