Toward a Diffraction Limited Light Source: The ESRF Storage
Ring Upgrade Varenna, July 9-11 2015 Pantaleo Raimondi On behalf of
the Accelerator Project Phase II Team
Accelerator Upgrade Phase II 3 Reduce the horizontal
equilibrium emittance from 4 nm to less than 140 pm Maintain the
existing ID straights beamlines Maintain the existing bending
magnet beamlines Preserve the time structure operation and a
multibunch current of 200 mA Keep the present injector complex
Reuse, as much as possible, existing hardware Minimize the energy
lost in synchrotron radiation Minimize operation costs,
particularly wall-plug power Limit the downtime for installation
and commissioning to less than 18 months. In the context of the
R&D on Ultimate Storage Ring, the ESRF has developed a
solution, based on the following requirements and constraints: Page
3 The Accelerator Upgrade Phase II aims to: - Substantially
decrease the Store Ring Equilibrium Horizontal Emittance - Increase
the source brilliance - Increase its coherent fraction Maintain
standard User-Mode Operations until the day of shut-down for
installation
Slide 4
LOW EMITTANCE RINGS TREND 4 Existing machines In Construction
Advanced Projects Concept stage APS SPring8 Sirius Page 4 Several
facilities will implement Low Horizontal Emittance Lattices by the
next decade Based on 1980 KnowHow Based on state of the art
technologies
Slide 5
Undulator : CPMU18, K=1.68 L=2m 5 Improved Coherence Hor.
Emittance [nm]40.135 Vert. Emittance [pm]45 Energy spread
[%]0.10.09 x [m]/ z [m] 37/36.9/2.6 Page 5 BRILLIANCE AND COHERENCE
INCREASE Coherence Source performances will improve by a factor 50
to100 Brilliance 18mm Undulator spectrum
Slide 6
BENDING MAGNETS SOURCE: 2-POLE, 3-POLE OR SHORT WIGGLERS Page 6
Half assembly Field Customized Large fan with flat top field 2 mrad
feasible for 1.1 T 3PW Mechanical length 150 mm Source shifts
longitudinally by ~3m Source shifts horizontally by ~1-2cm Flux
from 3PW All new projects of diffraction limited storage rings have
to deal with: Increased number of bending magnets / cell => BM
field reduction Conflict with hard X-ray demand from BM beamlines
ESRF will go from 0.85 T BM to 0.54 T BM The BM Sources will be
replaced by dedicated 2-Pole or 3-Pole Wigglers
Slide 7
UP PI: Preparing the UP Phase II ESRF Phase II Upgrade at the
Bone Present ESRF Arc Layout: Ex=4nm New Low Emittance Layout:
Ex=0.135nm The 844m Accelerator ring consists of: - 32 identical
Arcs 21.2m long - 32 straight sections 5.2m long equipped with
undulators and RF Each Arc is composed by a well defined sequence
of Magnets (dipoles, quadrupoles etc), Vacuum Components (vacuum
vessel, vacuum pumps etc), Diagnostic (Beam Position Monitors etc)
etc. All the Arcs will be replaced by a completely new Layout
21.2m
Slide 8
8 Double-Bend Achromat (DBA) Many 3 rd gen. SR sources Local
dispersion bump (originally closed) for chromaticity correction
Andrea Franchi Optimization of dynamic aperture for the ESRF
upgrade THE EVOLUTION TO MULTI-BEND LATTICE
Slide 9
9 Multi-Bend Achromat (MBA) MAX IV and other USRs No dispersion
bump, its value is a trade-off between emittance and sextupoles
(DA) Double-Bend Achromat (DBA) Many 3 rd gen. SR sources Local
dispersion bump (originally closed) for chromaticity correction
Andrea Franchi Optimization of dynamic aperture for the ESRF
upgrade
Slide 10
10 THE HYBRID MULTI-BEND (HMB) LATTICE ESRF existing (DBA) cell
Ex = 4 nm rad tunes (36.44,13.39) nat. chromaticity (-130, -58)
Proposed HMB cell Ex = 140 pm rad tunes (76.21, 27.34) nat.
chromaticity (-99, -82) Andrea Franchi Optimization of dynamic
aperture for the ESRF upgrade - Multi-bend for lower emittance
-Dispersion bump for efficient chromaticity correction => weak
sextupoles ( less SR - No need of large dispersion on the inner
dipoles => small Hx and Ex
Slide 11
LATTICE EVOLUTION: S28B OPTICAL FUNCTIONS Natural equilibrium
emittance: x0 = 134 pm Emittances with 5 pm coupled into the
vertical plane and 0.5 MV radiation losses from IDs: x = 107 pm z =
5 pm Page 11 1st MAC MEETING 14-15 April L. Farvacque
Slide 12
INJECTION CELL: S28B 2 quadrupoles in the straight section, The
global phase advance is still strictly identical to the standard
cell, However the tuning of the whole cell is slightly different,
giving more flexibility for the injection tuning. Page 12 1st MAC
MEETING 14-15 April L. Farvacque S28B Septum location
Slide 13
LINEAR AND NONLINEAR OPTIMIZATIONS Lifetime and dynamic
aperture are computed on 10 different errors seeds. Sextupoles:
from 6 to 3 families, weaker and shorter. Octupoles: from 2 to 1
family, weaker and shorter. Tunes: 76.21 27.34 Linear matching
parameters: x ID = 6.9m Chromaticities: 6, 4 1st MAC MEETING 14-15
April 2015 - N. Carmignani Linear and nonlinear optimizations have
been done with the multi-objective genetic algorithm NSGA-II, to
maximize Touschek lifetime and dynamic aperture. Page 13
Slide 14
LIFETIME OF S28B 1st MAC MEETING 14-15 April 2015 - N.
Carmignani S28B DA -10mm@S3 TLT ~ 21h S28A DA -8.1mm@S3 TLT ~ 13h.
e y =5pm ESRFUpgra de 7/8 multibunch 64 h21 h 16 bunch 6 h2.1 h 4
bunch 4 h1.4 h Page 14
Slide 15
COLLIMATION OPTIMIZATION TO DECREASE LOSSES IN THE IDS Page 15
1st MAC MEETING 14-15 April 2015 - N. Carmignani
Slide 16
COLLIMATOR SHIELDING50 CM LEAD LOCAL SHIELDING Page 16 1st MAC
MEETING 14-15 April 2015 - Paul Berkvens Sv/h 0.5 ambient dose
equivalent rate ( Sv/h) 0.5 cm total neutrons
Slide 17
17 The ESRF Low Emittance Lattice Proposed hybrid 7 bend
lattice Ex = 146 pm.rad Several iterations made between: Optics
optimization: general performances in terms of emittance, dynamic
aperture, energy spread etc Magnets requirements: felds, gradients
Vacuum system requirements: chambers, absorbers, pumping etc
Diagnostic requirements Bending beam lines source From Concepts to
a Real Machine Page 17
Slide 18
18 Technical challenge: Magnets System D1D1 D2D2 D6D6 D7D7 D3D3
D4D4 D5D5 Page 18 High gradient quadrupoles Mechanical design final
drawing phase Large positioning pins for opening repeatability
Tight tolerances on pole profiles Prototypes delivered in the
period September 2014-Spring 2015 Gradient: 90 T/m Bore radius:
12.5 mm Length: 390/490 mm Power: 1-2 kW Combined
Dipole-Quadrupoles 0.54 T / 34 Tm -1 & 0.43 T / 34 Tm -1
Permanent magnet (Sm 2 Co 17 ) dipoles longitudinal gradient 0.16
0.65 T, magnetic gap 25 mm 1.8 meters long, 5 modules Sextupoles
Length 200mm Gradient: 3500 Tm -2 Quadrupole Around 52Tm -1 Gael Le
Bec Octupoles
Slide 19
QUADRUPOLES DESIGN Pole shape optimization Imposed 11mm stay
clear from pole to pole for all magnets for optimal synchrotron
radiation handling Page 19 Low gradient pole profileHigh gradient
pole profile Vacuum chamber and magnets sections
Slide 20
QUADRUPOLES High Gradient 91 T/m gradient, 388 484 mm length
12.7 mm bore radius, 11 mm vertical gap 1.4 1.6 kW power
consumption HG Prototype +/-20um pole accuracy Moderate Gradient Up
to 58 T/m gradient, 162 295 mm length 16.4 mm bore radius, 11 mm
vertical gap 0.7 1.0 kW power consumption Page 20 1st MAC MEETING
14-15 April 2015 - LE BEC
Slide 21
DIPOLE WITH LONGITUDINAL GRADIENT Specifications 0.17 0.67 T
field 5 modules of 357 mm each Larger gap for the low field module
Allows the installation of an absorber Engineering design Final
drawings produced Prototyping Final prototype to be build in the
coming months Page 21 1st MAC MEETING 14-15 April 2015 - LE BEC
Measured field integral homogeneity (one module)
Slide 22
DIPOLE QUADRUPOLES Page 22 1st MAC MEETING 14-15 April 2015 -
LE BEC DQ1 pole shape DQ1 gradient homogeneity: Integration of
trajectory along an arc DQ1: 1.028 m, 0.57 T, 37.1 T/m G/G < 1%
(GFR radius 7 mm) DQs are machined in 7 laminated iron plates
Slide 23
OCTUPOLES S28b specifications 48 kT/m 3 nominal strength (70
kT/m 3 maximum) 90 mm length 4 Water cooled coils at the
return-field yoke Allows for the required stay-clear for
Synchrotron Radiation fans Prototyping Air cooled prototype
measured Page 23 1st MAC MEETING 14-15 April 2015 - LE BEC
Slide 24
1000 LARGE POWER SUPPLIES AND 1000 SMALL POWER SUPPLIES Page 24
TypeName NOMINAL FIELD VALUESElectrical design PS nommaxWatt
quantityLengthdB/dxlatticePowerVoltageCurrentOVdesign Watts P total
per cell[m][T/m] [kW][V][A]factorImaxPnomPmaxcell Quadrupole, mod.
gradientQF120.34953.7 1.0612.187.51.21021167157623343152
Quadrupole, mod. gradientQD220.26651.5
0.869.887.51.2106966141819322836 Quadrupole, mod.
gradientQD320.21646.5 0.748.487.51.2117843151916873037 Quadrupole,
mod. gradientQF440.21651.5 0.748.487.51.2106843123833734952
Quadrupole, mod. gradientQD520.21252.5
0.869.887.51.2104966136419322729 Total 12 1125716705 Quadrupole,
high gradientQF620.3695.2 1.4215.790.41.1991535185730703714
Quadrupole, high
gradientQF820.4896.21.6618.6891.1981767213935354277 Total 4
66057992 Dipole-Quadrupole, high
fieldDQ121.1137.5433.91.5915.75100.71.21211729249034584980
Dipole-Quadrupole, mod
fieldDQ210.7737.0433.71.3817.081.01.2971469211614692116 Total3
49287096 Sextupole, longSD4 450043001.0111.7861.1951111134444445377
Sextupole, longSF2 1.0111.7861.1951111134422222689 Total 6 66668066
OctupoleOF1-220.1 0.303.2941.21134266138521226 Total 2 8521226 27
Total PS power for one cell for main electromagnets30.341.1 KWKVA
magnetcoils type corrector AC+DC (5 independent coils)35 AC+DC
Sextupole, short correctors66 DC Total number of coils/cell 51
About 1000 DC-DC low voltage converters: the average channel power
is around 1kW and a maximum of 2.3kW. The stability requested will
be 15ppm with a MTBF of more than 400 000 hours. The integration in
32 cabinets will be designed with the Computer Services for
redundancy and HOT-Swappability
GIRDER DESIGN, THE ORTHOGONAL HEPTAPOD Page 26 4th APAC MEETING
- 22-23 OCTOBER 2014 - Cianciosi 5900mm 800mm Mass: Magnets: ~ 5-6
T Magnet supports: ~ 1 T Girders: ~ 3-3.5 T Vacuum chamber, pumping
etc: ~ 0.5T Total weight: ~9-11T Z X 4 motorized adjustable
supports in Z direction 3 manual horizontal jacks (one in X, two in
Y) Technology: Girder material: carbon steel Typical tickness: 30mm
(20-50) Piece junction: full penetration and continous welding
Rails flatness: 30 micrometers Filippo Cianciosi Quantity : 128
Procurement time scale: 2016-2017
Slide 27
SYSTEM INTEGRATION MACHINE LAYOUT IN THE TUNNEL Courtesy :J
Guillemin Design and integration of all components very advanced
Ready to start large scale procurement process in 2015 Construction
joint
Slide 28
SYSTEM INTEGRATION: GIRDER 1 AND 2 (& 3 AND 4) E-beam
Courtesy M Soulier All girders will be fully assembled before
starting the shutdown for installation
Slide 29
ASSEMBLY BUILDING LAYOUT 29 front view side view top view
Slide 30
PREPARING THE UPGRADE PHASE II Page 30 Technical Design Study
(TDS) Completed on May-2014 and submitted to: Science Advisory
Committee (SAC) Accelerator Project Advisory Committee (APAC) Cost
Review Panel (CRP) ESRF Council All committees very positive
Project Approved and Funded Official Start: Jan 1 st 2015
Slide 31
PLANNING AND BUDGET 31 Global Schedule Nov 2012 White paper Nov
2012- Nov 2014 Technical Design Study / Orange Book Jun 2014
Council Approval Jan 2015 Start of the project Jan 2015 July 2016
Engineering Design Jul 2015 Jun 2018 Procurement Sep 2017 Oct 2018
Girder Assembly Dec 2018 Start of the shutdown Dec 2018 Sep 2019
Dismantling and Installation Oct 2019 Jun 2020 Accelerator and
Beamlines Commissioning Jun 2020 Start of User Mode Operations
Budget: 100 M Construction of the new storage ring lattice 3.5
MBuilding programme for storage and pre-assembly 8. Concl.
Slide 32
CONCLUSION (1) Project officially started on Jan 1 st 2015
Optics and layout finalized Prototypes & manufacturers contacts
in progress Accelerator Engineering well advanced Tender process
for serial production foreseen started Mid 2015 All ESRF Divisions
greatly supporting the Accelerator Upgrade Project Significant step
toward a DLSR: about a factor 40 Another factor 10 reduction in the
horizontal emittance is still needed to reach the diffraction
limit
Slide 33
GREEN FIELD DLSR: GENERAL CONSIDERATIONS Current Upgrades
forced to maintain existing layout General Ring Structure: Straight
Sections interleaved with Arcs How to build the DLSR? Lets start
just listing the what is needed (not exhaustive): 0) 2pi total bend
angle 1) Straight Sections 2) Matching SS-Arc 3) Chromaticity
Correction 4) Low Emittance Lattice Arc 5) Injection and RF The
requirements for 1-to-5 are in general conflicting!
Slide 34
GREEN FIELD DLSR: GENERAL CONSIDERATIONS 1)Straight Sections:
How many?: 20-40 This number is mostly dictated by the optimal cost
of the infrastructure w.r.t. the number of beamlines. Low energy
ring costs much less, so in general fewer SS are found on them. How
long?: 3.0-5.0-10.0m From last 20 years of IDs progress (e.g:
revolver IDs, short periods, low gaps) can we define the optimal
length? Two IDs in a SS (say 5m long) already cost a lot: -
Beta-Mismatch equivalent to a factor two in emittance (for DLSR) -
Minimum gap and minimum period larger (=> higher ring energy
needed) Minimum Angle between SS?: 50-100mrad?
Slide 35
GREEN FIELD DLSR: GENERAL CONSIDERATIONS 2) Matching SS-Arc
Matching is detrimental to: Emittance (Dipoles with Curl-H not
ideal), Chromaticity Canted Long SS require less Matching Section
3) Chromaticity Correction Requires large betas, large dispersion
4) Low Emittance Lattice Arc Requires short dipoles, small
dispersion, small beta_functions 5) Injection and RF Requires large
x_beta and long Straight Section Could it make sense to not
integrate 2-3-4-5 in one Arc Lattice Structure but have different
optimized sections: - Straight Sections Section - Chromatic
Correction Section - Low Emittance Arc Section - Injection and RF
Section