Ex-Situ NMR: Design Approach to High Field Quality G. Sabbi, LDRD Progress Report 2/25/03

Ex-Situ NMR: Design Approach to High Field Quality

G. Sabbi, LDRD Progress Report 2/25/03

Field Quality Requirements

Ex-situ NMR technique aims at reducing FQ requirements with respect to standard NMR

But - for the moment ~10-5 homogeneity is still needed (!)

Minimal requirement for high resolution spectroscopy:<10-4 in 3 mm cube

We are setting a design goal of 10-4 in 10 mm radius (then correct to 10-5 using trim coils or magnetic shims)

Field measurement (to set correctors) will be an issue, use of NMR techniques may be best strategy

Best magnet design to date: ~3*10-4 in 5 mm radius

Design strategy

Presently following “accelerator magnet” approach:

- coil has a “long” straight section, field is optimized in 2D - 3D design to eliminate end effects with minimal coil length

Motivations:

- our main expertise is in accelerator coils- appears to be the most promising strategy for high field quality- potential for stretching the good field volume along the axis

Alternative design approaches:

Nested solenoids (a design based on normal conducting coils is being pursued in parallel to SC magnet effort). “2.5d” field optimization.

Conductor in groove for more freedom in conductor placement (may also be good for correctors)

Others….

Design “algorithm”

1. Optimize coil cross section (efficiency, 2D field quality)

2. 2D iron design (shielding, saturation, stray field)

3. Iterate on cross-section to adjust systematics

4. Find minimal coil length for no significant 3D effects

5. Try to further reduce coil length (3D coil, iron geometry)

6. Estimate random errors (tolerances, sweet spot distance)

7. Design corrector package (trim coils or magnetic shims)

Main challenge: 2D cross-section optimization for

- High efficiency (sweet spot field vs. coil peak field)

- Low systematic harmonics (esp. a4, b5)

2D Field Quality Optimization

- I+ I

ap (r,

x

y

nra

nIrA

nra

nIrB

nnn

nnn

sinsin

,

coscos

,

10

10

- I

xy

+ Idt

Cross-section design

We looked at designs with 4, 6, 8 blocks/layer.

Field quality depends on geometric center of block. However:

• Narrow blocks allow more design space• Narrow blocks allow better efficiency

Trying to work with aligned, identical blocks for design simplicity(however, relative y-shifts & spacers are needed for best results)

Assuming RD configuration for best field quality(find one solution first, then will have another look at HD)

• Based on SM coil design

• Optimize upper layer, then correct for lower layer

• 2D issues:- blocks are too wide – limits on 2D harmonics- narrow island decreases efficiency- narrow island can result in vertical forces

• 3D issues:- insufficient ratio of coil length to sweet spot distance- decreasing sweet spot distance makes design less attractive

& makes 2D optimization more difficult (high order systematics, random errors)

Coil cross-sections (Phase I)

Phase I – 2D cross-sections

Four blocks/layer (2D):

B1 = 2600 Gauss (11 kA)B1/Bpk = 0.025

For Rref = 1.5 mma2=0.1, b3=19.4, a4=1.1

Six blocks/layer (2D):

B1 = 3320 Gauss (11 kA)B1/Bpk = 0.037

For Rref = 1.5 mma2 <0.1, b3 <0.1, a4=-0.4

Coil cross-sections (Phase II)

Still based on RD coils, but:

- RD3c type design with central spacer- drop restrictions on length & top/bottom distance - NbTi assumed, with same cable geometry as SM

Advantages:

- doubled degrees of freedom for same # coils & splices- better efficiency (main field to peak field ratio) - better control on conductor positioning- standard approach to 3D effects

Phase II – 2D cross-sections (1 layer)

RDopt3

RDopt5

RDopt4

Phase II Performance Parameters (2D)

Parameter Symbol RDOPT3 RDOPT4 RDOPT5

Ref. Current [kA] Iref 10.0 10.0 10.0

Ref. Radius [mm] Rref 5.0 5.0 5.0

Main Field [Gauss] B1 2780 2800 2570

Skew Quadrupole a2 -0.2 0.3 0.2

Normal Sextupole b3 -0.2 0.3 0.3

Skew Octupole a4 -9.8 -5.5 -2.9

Normal Decapole b5 2.1 0.8 0.2

Main/Peak Field B1/Bpk 0.073 0.066 0.053

Case study: RDOPT4

Starting from the single layer RDOPT4 cross-section, thefollowing calculations were performed:

•Requirements on layer-to-layer separation (18 cm)•Field enhancement due to iron shield (10%)•Iron saturation effect (none)•Cross section iteration to correct harmonics (ok)•Coil length (upper bound) for no end effect (80 cm)

Optimized Design Parameters (NbTi)

Param. RDOPT4_V4

Iref Iss [kA] 10.0

Rref [mm] 5.0

B1 [Gauss] 3052

a2 -0.3

b3 0.2

a4 -5.3

b5 0.8

B1/Bpk 0.061

Title

Title