Machine Status after Bangalore N Walker ILC DESY Project

Machine Status after Bangalore N. Walker ILC@DESY Project Meeting 17. 03. 2006

A show and tell… • Basically I will show a few highlights from the plenary presentations from the Bangalore GDE meeting • Almost none of these slides are my own! • You are strongly encouraged to look at the presentations on the web-site: http: //ilcagenda. cern. ch/conference. Time. Table. py? conf. Id=6 • Only dealing with technical design here – cost issues and methodology for another day.

Basic Footprint

Electron Source SHB Buncher Laser |---- RT Pre-Accelerator----| 12 Me. V / m Diagnostics Tune-up dump (diagnostics section) n Gu Laser n Gu 120 ke. V Taken from presentation by T. Raubenheimer 12 Me. V 71 Me. V Klystron 10 MW Spare Klystron 10 MW

Electron Injector Linac |-Matching-| |--------- SC e- Injector Linac -----------| |---- Spin Rotation & transfer to DR ----| (one spare per 9 klystrons) 21 CM (w QD) ~ 5. 5 Ge. V Tune-up dump (diagnostics section) E, E diagnostics Taken from presentation by T. Raubenheimer Linac is followed by spin rotator, emittance & energy diagnostics, PPS gates, and DR injection

Positron Source Photon production at 150 Ge. V electron energy Two e+ production stations including a back up. Keep alive auxiliary source is e+ side. Beam Delivery System 150 Ge. V e- DR e- source IP 100 Ge. V Helical Undulator In By-Pass Line 250 Ge. V Photon Collimators e- Dump Photon Target Adiabatic Matching Device e- Dump Photon Dump Auxiliary e- Source e- Target Taken from presentation by T. Raubenheimer Positron Linac e+ pre-accelerator ~5 Ge. V Adiabatic Matching Device e+ DR

Taken from presentation by T. Raubenheimer

Positron Source Linacs • Normal conducting – 0 -125 Me. V: • 2 x 1. 27 m (SW NC + 0. 5 T Solenoid) driven by 2 10 MW klystrons • 3 x 4. 3 m (TW NC + 0. 5 T Solenoid) driven by 3 10 MW klystrons – 125 -400 Me. V: • 8 x 4. 3 m (TW NC + 0. 5 T Solenoid) driven by 8 10 MW klystrons • Superconducting – 400 – 1135 Me. V: • 4 x 11. 7 m (6 SC + 6 Q) driven by 1 10 MW klystron – 1135 -2605 Me. V: • 6 x 13. 6 m (8 SC + 2 Q) driven by 2 10 MW klystrons – 2605 – 5000 Me. V: • 12 x 12. 3 m (8 SC + Q) driven by 4 10 MW klystrons Taken from presentation by T. Raubenheimer

Positron Source Issues • Location of the positron return line in the linac beam line tunnel • Location of the positron return line through the BDS and interaction regions • Location of the positron linacs – Do we install two linacs in the main linac beamline tunnel? – Layout of the NC linacs and power sources – much higher density of rf components • Requirements for remote handling Taken from presentation by T. Raubenheimer

Positron Source Issues • • • Lattice development: 5 Me. V-5 Ge. V Magnets: definition and cost Vacuum system: specification and costing NC rf details SC rf details 0. 5 T, 0. 66 m id, ~165 m solenoid details Radiation issues: personnel protection Undulator details Target details Remote handling strategy and details • Costing of target and AMD will be done by sources group • Positron source should not be a cost driver – Biggest cost items will be SC linacs – Concern about the positron transport lines Taken from presentation by T. Raubenheimer

Taken from presentation by T. Raubenheimer R 2 ML two stage bunch compressor: 6 mm 150 mm 9 mm 300 mm stage 2 stage 1 15 Ge. V

Taken from presentation by T. Raubenheimer R 2 ML • Technical risks – Machine protection • Pulsed extraction kickers risetime ~100 nsec • Collimator protection relies on stopping DR extraction with very low latency – RF system phase stability • Relative (e+ vs e-) phase stability must be around 0. 2° of 1. 3 GHz – Pitch angle alignment of RF cavities • 300 urad RMS, if achievable, will (probably) already dominate vertical emittance growth in RTML • Cost Risks – RF components will be hardest to cost • Identical to main linac components – Other (likely) cost drivers are conventional: room temp magnets, vacuum system, tunnel

Damping Rings • Electron: 1× 6 km ring • Positron: 2× 6 km rings (stacked) • Frequency now 650 MHz (was 500 MHz) – greater flexibility in bunch patterns – faster in/ejection kicker needed – requires development of RF system • Rings located at ends of machine – may not be the cost optimimum

Damping ring working group well organised, but resources missing (many ? )

All relevant current information available at Cornell WIKI site

Current general parameters Parameters and fill patterns in 6. 6 km damping ring with 650 MHz RF frequency (General Parameters). Circumference Energy RF frequency Harmonic number Transverse damping time, e+ DR (e. DR) Normalized natural emittance Equilibrium bunch length 6642. 4784 m 5 Ge. V 650 MHz 14402 <25 ms (<50 ms) 5 m 6 mm Equilibrium energy spread <0. 13% Momentum compaction ~ 4 10 -4 Damping wiggler peak field 1. 67 T Damping wiggler period 0. 4 m Energy acceptance | |<0. 5% Dynamic aperture Ax+Ay<0. 09 m-rad (up to | |=0. 5%)

650 MHz RF Parameters D. R. MAIN RF PARAMETERS Damping rings parameters e- RING e+ RINGS 5 5 Number of bunches per train 2700 1350 Number of particles per bunch 2. 2 x 1010 Average current (amps) 0. 43 0. 22 Energy loss per turn (Me. V) 9. 3 Beam power (MW) 4. 0 2. 0 Bunch current (m. A) 0. 16 Total RF voltage (MV) 19. 3 6. 642 Energy (Ge. V) Circumference (km)

Summary of Fill pattern with 650 MHz RF Ø Optimized Beam Patterns: 4320 bunches, 4500 bunches Ø The difference between different fill patters is small: Low-Q has similar ion density as high-Q fill pattern Major concern of EC was that the DR WG had not taken full consideration of the so-called low -Q parameter set during their choice of baseline DR design. Ion reduction factor for different fill patterns (ref: http: //www. desy. de/~awolski/ILCDR/)

Fast Ion Instability (e- ring) Impact of low-Q (large bunch #) scenario Ø There are more beam fill patters with 600 MHz RF, and they have similar FII (low-Q, high Q), the growth time is about 28 turns (6 km ring) with 1 n. Torr CO ion Ø FII becomes worse when the total number of bunches is more than 4050 (a factor of 2) Ø Feedback can be used to damp the dipole motion due to ions, a small kick (10 Volts) is enough in the example of Low –Q due to the characters of FII Note: bunch spacing ~3 ns kicks!

Main Linac • RF components still basically TESLA/XFEL – modifications to cryomodule (so-called type 4) – 10 MW MBK klystron and Bouncer Modulator still baseline – basic RF distribution w/ circulators • Linac layout different – cyrogenic distribution halls – RF distribution layout – inclusion of diagnostics, MPS dumps etc • still under discussion

• International Effort between the three regions • Design changes are towards nailing down slot length of components – Costing should be straight-forward from TTF (and possibly XFEL) experience Type IV (ILC) Slides from Talks by Don Mitchell, Tom Peterson and Others at Jan 16 -17 CERN Meeting

Type IV Cryomodule • Move quad package to middle of cryomodule to achieve better support and alignment. • Shorten cavity-to-cavity interconnect and simplify for ease of fabrication and cost reduction. • Overall improved packing factor. • Simplify the assembly procedure. • MLI redesign to reduce hands-on labor costs. • More robust design for shipping. • Reliability of tuner motors in cold operation. • Revaluate cryogenic pipe sizes – partially done for the XFEL already

Cavity with Frequency Tuner • No BCD Tuner, some designs are very close to requirement – Generic issue to all designs: motor and piezo reliability – Deemed to be feasible, but some R&D needed • E. g Bladetuner – Issue with cavity’s magnetic shielding • Could be also another tuner that does not need intercavity space – Just watch out for the cryo-lines…

ACD: Seperate Quad Cryo-section Baseline is TDR like with Quadrupole at centre of cryomodule 1530 mm

BPM / Quad / Corrector Package 887 66 BPM 77 666 QUAD and Correctors ILC Preliminary 78 TDR

Examples of RF three-way split Leibfritz, FNAL Fukuda, KEK

Need more compact design (Each Cavity Fed 350 k. W, 1. 5 msec Pulses at 5 Hz) Two of ~ 16, 000 Feeds

Installation Systems Concept Development Step Installation Phases & Sequencing Tunnel cross-sections still under discussion: currently 4. 5 m and 5. 5 m diameter

Crossovers between the tunnels

Tunnel Rad calc w/ FLUKA by Fasso • Uses full tunnel geometry • Loss is 1 m upstream of 35 cm diameter penetration • Support tunnel below 3. 5 m has 0. 003 m. Sv/hr/k. W which is < 0. 014 OK • Conclusion: < 5 m between tunnels is definitely not OK. 5 m is OK, if willing to fence off area near penetrations. One penetration per ~36 m? Or one every 12 m? ? ?

Cryoplant Layout in e- Linac Tom Peterson For ILC 500, total of ten 25 k. W @ 4 K plants requiring 52 MW of AC power.

Cryogenic unit segmentation and other cryogenic boxes • Segmentation issue is ultimately tied to reliability • RDR should include features for vacuum segmentation • Assume 4 cryo strings (48 modules, 563 meters) per segmentation unit • Cryogenic string supply and end boxes (cryogenic service modules), which may (should!) be separate from modules, are also required within the ILC linac

Full segmentation concept (ACD) • A box of slot length equal to one module • Can pass through cryogens or act as “turnaround” box from either side – Does not pass through 2 -phase flow, so must act as a supply or end of a cryogenic string • Includes vacuum breaks • May contain bayonet/U-tube connections between upstream and downstream for positive isolation • May contain warm section of beam pipe • May also want external transfer line for 4 K “standby” operation (4 K only, no pumping line)

Segmenting the cold vacuum • Optional: Include 6 -12 m long warm sections after every 48 cryomodules (560 m) • Use for beam line and insulating vacuum isolation. • Each would [could] contain a laser wire: with 21 wires, have 7 independent measurements of emittance along each linac. • Could contain other instrumentation such as beam halo and dark current monitors. • Could contain spoilers for short-train beam abort. • Could be used for cryo-segmentation as discussed earlier • Penalties Cost; MPS issues, contamination Many ‘coulds’ but where are the ‘musts’ ?

Cryogenic architecture For shaft depth above 30 m, the hydrostatic head in the 2 K pumping line becomes prohibitive and active cryogenics (e. g. cold compressor system) has to be installed in caverns (LBC), i. e. additional cost for cryogenics and civil engineering.

Slope of ~5 km cryo unit OK Cryogenics can support a laserstraight machine with this segmentation! Current baseline follows the Earth’s curvature

Baseline layout 20 mrad IR and 2 mrad IR e+, low E • Grid size: 100 m * 5 m • (Beamline is not placed near external walls, as suggested above)

Shafts and path of access in BDS • One possible criteria: Beamlines are close to external walls, to allow access from IR halls without crossing the extraction lines • This seem to contradict to present linac layout, where service tunnel is at the side of low angle IR • It may also be less optimal for gg upgrade Low E e+ Where this shaft is located and can it provide access to all three branches?

20 mr => 25 mr 1. 8 m 5 m • Group beginning to consider g-g option (upgrade)

14 mr => 25 mr 7 m 4. 2 m • additional angle is 5. 5 mrad and detector need to move by about 4. 2 m

Hall with shielding wall 18 MW loss on Cu target 9 r. l at s=-8 m. No Pacman, no detector. Concrete wall at 10 m. Dose rate in mrem/hr. • For 36 MW MCI, the concrete wall at 10 m from beamline should be ~3. 1 m Alberto Fasso et al Wall 25 rem/hr 10 m

Discussion of requirements to enable push-pull • Detailed discussion will happen on MDI panel/LCWS side (will be presented by Tom Markiewicz) • Technically, rapid (1 -2 day every ~3 month) switch may require – – – Detailed engineering for push-pull from inception Part of Final Doublet carried with detector during move Break point in optics with double valves & cold warm transitions Sufficient embedded steel in the floor Reference alignment network in the hall to monitor detector and floor deformation; jacks under detector minimize detector deformation during move. Vertex and tracker may be on its own independent alignment system – Detector is self shielded and no shielding wall – SC solenoid of detector may need to stay cold and energized – etc.

MPS • Very little done since BCD • Starting failure analysis + simulation in linac. With various failure modes (phasing, magnet shorts, magnet settings) what will beam hit and will it destroy it. • Answers may determine if pilot bunch and 1 dump/km in linac (ACD) needed. from Tom Himel’s presentation

Dumps – the reasons • Some needed to allow fast recovery from probable frequent MPS trips – Full power dumps needed downstream of systems with significant beam heating. Allows them to stay warm during an MPS trip and speeds the recovery. • Others needed to allow a system to be tuned without potentially damaging beam going through downstream systems – Only need to take ~100 bunches per train at 5 Hz. Enough for intra-train feedback and LLRF. – Above assumes we handle beam loading compensation OK. • Some need added redundant stoppers and shielding walls to allow beam in one region with people in another. from Tom Himel’s presentation

Dumps – the LIST • We have made a draft list of all the dumps and the reason each is needed • Will be distributed shortly • Some dumps may be controversial as high power dumps are expensive • Need a decision making process: – We (operations) dictate? – Negotiate with each region? from Tom Himel’s presentation – Exec board mediates?

Many Outstanding Issues • Global timing still needs to be resolved – impact of e+ source may add 1 -2 km to the ‘path length’ [future talk from H. Ehrlichmann] • Many questions from technical systems – component counts there but… – where to place equipment, racks, RF etc. still needs to be resolved CF&S requirements! • Need to prudently identify cost-relevant issues for July deadline – should not waste time worrying about details under the cost radar! – This is not a full Technical Design!

Summary (My Opinion) • Very large amount of technical work and progress being made by a truly international group – although RDR design work heavily biased to US groups due to resource availability in EU and Japan! • Level of detail varies between groups/machine sub-systems – some more advance than others • CF&S is a driving term – cost driver: need to understand quickly what constraints are

Summary (My Opinion) cont. • Current ‘bottom-up’ design is gold plated …and probably expensive! • Primary goal now is to consolidate what we have enough to produce rough cost estimate for Vancouver in July (4 months!) • After Vancouver, we will almost certainly need to push back on the design and component costs • Many 1% effects will be difficult to remove! • We are on a very ambitious time-scale!