The new magnetic field optimisation procedure of the nEDM experiment at PSI Elise Wursten On behalf of the nEDM collaboration Probing Fundamental Symmetries with UCN workshop, Mainz April 11, 2016
Contents • Introduction • Magnetic field knowledge • CsM • Field maps • Variometer • Optimisation procedure • Results • Homogeneity of 𝐶 𝑨 • Ramsey visibility of 2015 • Conclusion 1
Goal To optimise the magnetic field such that we gain in nEDM sensitivity by improving the visibility of the Ramsey curve − 𝑈 𝛽 0 : Initial polarisation (86%) 𝛽 = 𝛽 0 𝑓 𝑈2 𝑈 2 : Transverse depolarisation time while keeping systematic effects under control. 2
Introduction What do we know? The transverse depolarisation time 𝑈 2 is mainly dependent on the homogeneity of the longitudinal component 𝐶 𝑨 whereas systematic effects are related to the transverse 2 component squared of the magnetic field 𝐶 𝑈 2 small So we want to homogenise 𝐶 𝑨 while keeping 𝐶 𝑈 3
Introduction What do we need? Coils to trim the field 4
Introduction What do we need? Coils to trim the field ? Vector information about the magnetic field shape This would we easy if we had vector magnetometers, but we don’t have vector magnetometers… Or do we? 4
Magnetic field knowledge We have 16 Cs magnetometers installed above and below the precession chamber : See talk M. Kasprzak Probe the field locally Scalar sensors 5
Magnetic field knowledge We have 16 Cs magnetometers installed above and below the precession chamber : Probe the field locally Scalar sensors 2 = 𝐶 𝑨 + 𝐶 𝑈 𝐶 + ⋯ ≈ 𝐶 𝑨 2𝐶 𝑨 Homogenisation using only CsM was not really successful: No control over 𝐶 𝑈 5
Magnetic field knowledge Field maps from mapping campaign in 2014: Vector fluxgate magnetometer mounted on a mapping device Maps of main field 𝐶 0 , trimcoils ,… 6
Magnetic field knowledge Field maps from mapping campaign in 2014: Vector fluxgate magnetometer mounted on a mapping device Maps of main field 𝐶 0 , trimcoils ,… Homogenisation using only maps was not really successful: Accuracy is not high enough to get a reliable estimate of 𝐶 0𝑨 6
Magnetic field knowledge Combine CsMs and field maps: variometer mode Apply a known transverse field additional to the main field and measure the response in the magnitude of the CsMs 7
Magnetic field knowledge Variometer mode: B 0 8
Magnetic field knowledge Variometer mode: B 0 x-direction B DC 8
Magnetic field knowledge Variometer mode: B 0 x-direction B DC 8
Magnetic field knowledge Variometer mode: B 0 x-direction B DC 8
Magnetic field knowledge Variometer mode: B 0 x-direction B DC 8
Magnetic field knowledge Variometer mode: B 0 x-direction B DC 8
Magnetic field knowledge Variometer mode: B 0 y-direction B DC 8
Magnetic field knowledge Variometer mode: B 0 y-direction B DC 8
Magnetic field knowledge Variometer mode: B 0 y-direction B DC 8
Magnetic field knowledge Variometer mode: B 0 y-direction B DC 8
Magnetic field knowledge Variometer mode: B 0 y-direction B DC Fit function: Parabola! 8
Magnetic field knowledge Fit function is a simple parabola Extract 𝐶 0𝑦 and 𝐶 0𝑧 from 𝐶 0 ∙ 𝐶 1 and 𝐶 0 ∙ 𝐶 2 Using 𝐶 0𝑨 = 𝐶 0 from CsM • 𝐶 1 and 𝐶 2 from field maps • 9
Magnetic field knowledge Variometer mode Precision: • Ranges from 5pT to tens of pT for 30s measurement time Accuracy : • Limited by knowledge of the fields produced by the two coils • Absolute values could be off by tens of nT 𝜀𝐶 0𝑦 ≈ 𝐶 0𝑨 𝜀𝐶 1𝑨 ≈ 1000 𝜀𝐶 1𝑨 = 20 𝜀𝐶 1𝑨 𝐶 1𝑦 50 • But differences (with common 𝐶 0 ) are accurate up to a few % 10
Magnetic field knowledge Measure the field of the trimcoils: example BTC B x (nT) B y (nT) B z (nT) 11
Optimisation procedure Measure the field of each trimcoil with the variometer mode as in the previous example => response matrix 1. Measure the instantaneous field 12
Optimisation procedure Measure the field of each trimcoil with the variometer mode as in the previous example => response matrix 1. Measure the instantaneous field 2. Minimise least squares function such that 𝐶 𝑨 → 𝐶 a) 𝐶 𝑦 → 0 and 𝐶 𝑧 → 0 b) c) and the trimcoil currents are regulated. 2 based on the fluxgate maps of 2014 Calculate 𝐶 𝑈 12
Optimisation procedure Measure the field of each trimcoil with the variometer mode as in the previous example => response matrix 1. Measure the instantaneous field 2. Minimise least squares function such that 𝐶 𝑨 → 𝐶 a) 𝐶 𝑦 → 0 and 𝐶 𝑧 → 0 b) c) and the trimcoil currents are regulated. 2 based on the fluxgate maps of 2014 Calculate 𝐶 𝑈 3. Assign a large weight to the Bz term, fix the regularisation parameter and scan over (small) weights for the Bx and By 12
Optimisation procedure Measure the field of each trimcoil with the variometer mode as in the previous example => response matrix 1. Measure the instantaneous field 2. Minimise least squares function such that 𝐶 𝑨 → 𝐶 a) 𝐶 𝑦 → 0 and 𝐶 𝑧 → 0 b) c) and the trimcoil currents are regulated. 2 based on the fluxgate maps of 2014 Calculate 𝐶 𝑈 3. Assign a large weight to the Bz term, fix the regularisation parameter and scan over (small) weights for the Bx and By 2 and small predicted spread in Bz. Select a solution with low 𝐶 𝑈 4. 12
Results B z before B z after 13
Results Visibility of nEDM runs of 2015: 2014 Typically between 0.45 and 0.6 30% 2015 Typically between 0.7 and 0.8 14
Conclusion Vector information from CsM in variometer mode We have a successful routine to homogenise 𝐶 𝑨 while 2 under control keeping 𝐶 𝑈 We report an increase of 30% in nEDM sensitivity! 15
Recommend
More recommend
