Energy deposition studies at IR7 M. Santana, M. Magistris, A. Ferrari, V. Vlachoudis Collimation collaboration meeting

Introduction

Motivation Large Hadron Collider : 27 km cryogenic installation LHC is a proton-proton collider LHC is a proton-proton collider 2 proton beams at 7 TeV of 3×10 14 p + each 2 proton beams at 7 TeV of 3×10 14 p + each stored for hours in collision stored for hours in collision total stored energy of 0.7 GJ (sufficient to melt 1 ton of Cu) total stored energy of 0.7 GJ (sufficient to melt 1 ton of Cu) ~5000 cold magnets ~5000 cold magnets Tiny fractions of the stored beam suffice to quench a superconducting LHC magnet or even to destroy parts of the accelerators. Tiny fractions of the stored beam suffice to quench a superconducting LHC magnet or even to destroy parts of the accelerators. The LHC collimation system will protect the accelerator against unavoidable regular and irregular beam loss. The LHC collimation system will protect the accelerator against unavoidable regular and irregular beam loss. from MC2005 V. Vlachoudis

Two Stage Cleaning Secondary halo p p e  Primary collimator Core Diffusion processes 1 nm/turn Shower Beam propagation Impact parameter ≤ 1  m Sensitive equipment Primary halo (p) e  Shower p Tertiary halo Secondary collimator Titanium Graphite Copper Beryllium Aluminum Escaping % Density g/cm 3 Material Example for 1 m long jaws! Secondary collimators intercept halo --> Shower energy escapes to downstream elements! so then... What happens downstream?

RADWG-RADMON Workshop Day, CERN 01/12/20045 E6C6 IP7 A6 C6E6 UJ76 RR77 RR73 IR7 layout LHC lattice and optics files V Primary and Secondary collimators, Scrapers, Absorbers Normal operation 0.2 hours beam lifetime  4×10 11 p/s for 10 s

RADWG-RADMON Workshop Day, CERN 01/12/ IR7 Geometry UJ76 RR73RR77

Geometry of the dipoles Tesla 14m long objects with a field of 8.3 Tesla: 5mrad bend  ~3cm sagitta. The superconducting dipoles(MB)are made out of 4 straight sections to accommodate the trajectory.

Magnetic field maps ● General routine for handling magnetic field maps (Analytic and/or 2D Interpolated) with the use of an external file with a special format ● Magnetic field type – CONSTConstant field – QUAD2D Analytic quadrupole field – QUADINT2D Analytic+Interpolated quadrupole field – INTER2D2D Interpolated field ● Symmetries: – NONENo symmetry – X, Y, ZSymmetrical on plane X, Y, Z (-x  x, …) – XYOn both planes XY – XYZOn all planes XYZ ● Table with interpolated data (Bx,By,Bz) ● Quadrupole analytic description – Origin of the magnetic field map origin – Limiting radius up to where to consider an analytic field ● Translation and Rotation of the field map ● Field intensity / gradient specified per region or lattice from MC2005 V. Vlachoudis

Magnetic field example MQW – Warm Quadrupole XY Symmetry Analytic 2D Interpolated from MC2005 V. Vlachoudis

Primary Inelastic collisions map ● Generated by the COLLTRACK V5.4 program – 3 scenarios: Vertical, Horizontal and Skew – Pencil beam of 7 TeV low-beta beam on primary collimators – 100 turns without diffusion – Impact parameter:  – Spread in the non-collimator plane: 200  m – Recording the position and direction of the inelastic interactions ● FLUKA source: Force an inelastic interactions on the previously recorded positions Beam Loss Map M. Brugger et al from MC2005 V. Vlachoudis

Execution time Biasing ● Importance biasing: radially decreasing ● Leading particles biasing ● High energy cuts on EMF on regions far away ● Weight Windows per region Statistics ● 30% on maximum ● Linux Cluster ● 64 3GHz ● 1 week run Improvements: ● Bias the diffractive/inelastic scattering ratio from MC2005 V. Vlachoudis

Collimators

Collimators Material Choice Not driven so much by the standard collimation but rather by the faulty operations or malfunctions ● Worst Accident scenarios: – Due to a spontaneous rise of one of the extraction kicker modules during the coast, part of the 7 TeV/c beam is spread across the front of a collimator jaw. – Faulty kick by the injection kicker where a full batch of protons hit the front of a collimator jaw at 450 GeV/c ● Very fast absorptions of part of the proton energy: – Instantaneous temperature rise – Thermally induced stresses (overheating/melting)  Limits material choice which can be used and still be compatible with other machine requirements. ● FLUKA A.Fasso, A.Ferrari, J.Ranft, P.R.Sala Proceedings of the Monte Carlo 2000 Conference, Lisbon, Oct , Springer-Verlag Berlin, p (2001) from MC2005 V. Vlachoudis

Collimators Criteria: ● Primary and secondary collimators are the closest elements to the beam ● Activating single scattering for thin layer on jaws ● Jaw halfgap / tilt variable during runtime Primaries: Gap:6  Jaws:C-C Length:20cm... may be changed to 60 Secondaries: Gap:7  Jaws:C-C Length:100cm Absorbers: Gap:10  Jaws:Cu or W Length:100cm from MC2005 V. Vlachoudis

Secondary Collimator Maximum Energy density in TCSGA6L1 carbon jaws

Simulations

Simulation Strategy ● Dynamic FLUKA input generation with several ad-hoc scripts ● Detailed description of 20 prototypes located in a virtual parking zone. ● Prototypes are replicated with the LATTICE card, rotated and translated. ● Magnetic field maps: Analytic + 2D Interpolated ● Dynamic generation of the ARC (curved section) ● Optics test: Tracking up to 5 , both vertical / horizontal, reproduce beta function Input Files FLUKA input template Twiss files Collimator summary Absorbers summary Prototype Info mklattic.r BRexx Script Fluka Input (.inp) LATTICE definitions Curved Tunnel creation Magnetic Fields Intensity Scoring cards Fluka Executable LATTICE transformations Dynamic adjustment of collimator gaps Fortran Files Source routine Si Damage 1 MeV n eq. Magnetic Field History Tracking

Automatic Geometry Creation 1. Initial input file template 2. Space Allocation & Geometry Creation 3. Lattice generation 4. Magnetic Fields mapping

Implementation of vertical and horizontal absorbers Beam profile z(m) Geometry Like secondary collimator, with Cu jaws and 10 sigma halfwidth

Steps to launch a simulation ● 1) Modify active absorbers : #icoll Name Material Length Rotation Tilt(jaw1) Tilt(jaw2) Halfgap N_Impacts N_InelInt Impact(av) Impact(sig) # [m] [rad] [rad] [rad] [m] (protons) (protons) [m] [m] 1 TCL.A4R7.B1 CU E E E E E E-04 # 1 TCL.A4R7.B1 CU E E E E E E-04 # 1 TCL.A6R7.B1 CU E E E E E E-04 # 1 TCL.A6R7.B1 CU E E E E E E twiss/absorber_summary.dat "RCOLLIMATOR" "TCL.A4R7.B1" "RCOLLIMATOR" "TCL.A6R7.B1" Absolute position V6.5_absorbers.b1.phase1.data Use twiss/tensigma.da t = active = inactive # = no line = Vertical = Horiz. - Each absorber must be defined in both files (inactive absorbers count but not hashed lines). - There cannot be two active replica of the same absorber.

Steps to launch a simulation ● 2) Modify geometry, activate relevant USRBIN in ir7.fluka – USRBIN must always be active. ● 3) Check for errors in prototypes etc.: – $ fluka.r -m ir7.fluka ir7.i – $ mcnpx ip i=ir7.i ● 4) Select appropiate beam file: – Check ir7.fluka: – Check run.sh to include that beam: BEAM PROTON SOURCE beam1 *SOURCE A7 *SOURCE B7 ir7.fluka SOURCES="Twiss/beam1.dat Twiss/beam2.dat Twiss/beam1V.dat Twiss/beam1H.dat Twiss/beam1S.dat Twiss/beam1VP.dat Twiss/beam1VPH.dat Twiss/beam1VPS.dat Twiss/beam1VPV.dat Twiss/beam1VS1.dat Twiss/A7.dat Twiss/B7.dat Twiss/nbeam1.dat" run.sh

Steps to launch a simulation ● 5) IF a new prototype has been designed, include it in prototype.pos: ● 6) Compile geometry --> ir7.inp: – $ make proper – $ make ● 7) Check correctness of ir7.inp: – Check number and type of absorbers ir7.inp. – Make plots of newly introduced elements: ● $ flukaplot.r ir7.inp flukair7 ● 8) Make a test run: – Check results and speed. – Check lattice tables: ● $ EnLattice.pl < 'ir7.inp' 'usrbinf_39' '1' ● 9) RUN and analyze. prototype.pos # Absorber (like Hybrid) TCL 0 0 TCL # One or two beams (normalization) File with formatted results $ usbsuw to summarize $ usbfuf to format 1 or 2 beams

Correction of beam direction. x x z y Only primaries are affected 85% inelastic scattering (minor consequences) 15% diffractive scattering (deviated and partially lost) 40% more dose in MQTLH, but still below limit Much higher dose in the curved section, but still well below the limit 0.5 mrad rotation

Results: warm elements

Preliminary results for the straight section (corrected beam) Total energy deposited in the MBWB6L: Corrected beam: 28.4 kW (Uncorrected beam: 37 kW) Energy deposited in the TCSGA6L1: Total energy: 20 kW (Uncorrected beam: 22.6 kW ) Energy in both jaws: 5.1 kW (Uncorrected beam: 1.02 kW ) Hot spot with no physical meaning, due to the beam error

Heat in the finger collar of the TCSGA6L W 80.4 W

Energy deposition in flanges W/cm 3 TCSGA6L1 <-front back-> 70 W 22 W 457 W 85 W

Passive absorber ● Most of the radiation deposited in the MBW insulator comes from inside the beam pipe. ● The efficiency of the absorber strongly depends on the inner radius. Ideal absorber. Pipe size. Fe absorber. 2 cm radius (pipe is 4 cm) 5-10 MGy/y ~ 1 MGy/y ● Need for smaller radius. Cu absorber? Ideas?

Results: cold elements

Implementation of vertical and horizontal absorbers 1 Finally selected. TCL.A6R7.B1 s= TCL.C6R7.B1 s= TCL.E6R7.B1 s= candidate absorbers in straight section 2 candidate absorbers before curved section 4 Finally selected TCL.B7R7.B1 s= Beam 1 Beam 2 TCL.A7R7.B1 s= TCL.A4R7.B1 TCL.F6R7.B1

4+1: A6 v C6 h E6 v F6 h -A7 h Number of simulations: 442 ******* Straight Section ************* ** * MQTLHA6R ******************* * max heat in coil: mW/cm3 ( %) * Total heat in the coil: W ( %) * heat in MQ: W ( %) ** * MQ6 group ****************** MQTLHA6R 1.22 ( %) W MQTLHB6R 0.40 ( %) W MQTLHC6R 0.26 ( %) W MQTLHD6R 0.19 ( %) W MQTLHE6R 0.14 ( %) W MQTLHF6R 0.13 ( %) W TOTAL 2.07 ( %) W ******* Curved Section **************** Total energy in coils and magnets of MQ[7-11]R. MQ7 | max: (+-99.0%) | 1.796e % | Total: W % MQ8 | max: (+-82.8%) | 1.193e % | Total: W % MQ9 | max: (+-65.5%) | 1.474e % | Total: W % MQ10 | max: (+-99.0%) | 3.124e % | Total: W % MQ11 | max: (+-99.0%) | 1.566e % | Total: W % Total energy in coils and magnets of MB[A-B][8-11]R. MBA8R | 1:inner_coil 1.120e % | 1:outer_coil 5.637e % | max: (+-98.2%) | MBB8R | 2:inner_coil 8.321e % | 2:outer_coil 4.208e % | max: (+-85.5%) | MBA9R | 3:inner_coil 3.937e % | 3:outer_coil 2.060e % | MBB9R | 4:inner_coil 3.069e % | 4:outer_coil 1.721e % | MBA10R | 5:inner_coil 3.439e % | 5:outer_coil 1.343e % | MBB10R | 6:inner_coil 2.132e % | 6:outer_coil 4.892e % | MBA11R | 7:inner_coil 2.003e % | 7:outer_coil 1.135e % | MBB11R | 8:inner_coil 1.085e % | 8:outer_coil 6.023e % |

Radiation in the MBA8

Heat spikes in MB's

Radiation in the MQ's 1W W 0.01W W mW/cm 3

Comparison between A7 h and B7 h Part of the beam halo will interact with the absorbers and generate a hadronic shower => energy deposition in the cold magnets The contribution from B7 h will be 15% higher than A7 h, but still at an acceptable level. Peak values in MQ7: A7h => 0.22 mW/cmc (*) B7h => 0.26 mW/cmc (*) (*) values refer to 1 proton interacting out of 10,000 lost in TCP. Error is below 6%. Tertiary halo

Comparison between A7 h and B7 h ● Simulations were run with corrected beam. ● The accuracy of the magnetic field in the MB was improved. ● Low energy photons were fully simulated. from PAC2005 M. Santana et al.

Comments

Simulation accuracy. Sources of error - Physics modeling: – Uncertainty in the inelastic p-A extrapolation cross section at 7 TeV lab – Uncertainty in the modeling used Factor ~1.3 on integral quantities like energy deposition (peak included)while for multi differential quantities the uncertainty can be much worse. - Layout and geometry assumptions: I t is difficult to quantify, experience has shown that a factor of 2 can be a safe limit. - Beam grazing at small angles on the surface of the collimators: Including that the surface roughness is not taken into account a factor of 2 can be a safe choice. - Safety factor from the tracking program COLLTRACK is not included! from MC2005 V. Vlachoudis

Some facts... - Challenge: – 'Filter twice 450 kJ' in such a way that superconducting elements get less than 5 mW/cm3! – Track showers along 1.5 km of tunnel and build up statistics with rare occurring events. - Resources: Over 7 years of equivalent CPU-(2.8 GHz) over 15 months in a 3 man-year effort. - Models and scripts: one of the most complex simulations in FLUKA.

Related works: UJ & RR's electronics protection K. Tsoulou

RADWG-RADMON Workshop Day, CERN 01/12/ NoAbsorbers A6vC6Eh6v Absorbers RR77 RR73 Dose (Gy/y) UJ76 Flux (cm -2 /y) A6v C6h E6v beam1 beam2 E6v C6h A6v No Absorber vs. Absorber (tunnel) Mean values ± 2m horizontally and ± 1m vertically.

RADWG-RADMON Workshop Day, CERN 01/12/ Three Absorber Case for UJ76 Dose (Gy/y) Doses in racks ≤ 5 Gy Similar to NoAbsorber case !

Related works: Ozone production in IR7. Accident case studies,... A. Pressland

Introduction ● Radiation induced production of O 3 around IP7. – dose estimates provided by Fluka – assumed 4.1  lost protons per year. – assumed all Fluka energy loss in air is ionizing. ● Enclosures around regions of high dose (O 3 concentration) – enclosures seal the tunnel in areas where the ozone – voided independently of the main tunnel – air corridor to allow passage of tunnel air towards TU76

Tunnel Section

Energy scorings ● Annual dose (GeV/cm 3 ) based 4800 beam-hours ● Complient with the standard 10 4 – 10 5 Gy/year

Calculation (1) Fasso et el (1982) LEP Note 379 gives the following differential equation I = ionizing energy deposited in air per unit time, in eVcm -3 s -1 G = number of ozone molecules formed, in eV -1 (7.4  eV -1 )  = dissociation constant for ozone, in s -1 (2.3  s -1 ) N = concentration of ozone at time t, in cm -3 k = decomposition constant, in eV -1 cm 3 (1.4  eV -1 cm 3 ) Q = ventilation rate, in cm 3 s -1 V= irradiated volume, in cm 3 Integration leads to the following concentration kinetics: formation dissociation decomposition ventilation

Calculation (2) More useful steady state formulation in a tunnel –average energy, I ave, is deposited per unit time –air circulates with speed v ms -1 –length z of tunnel is irradiated This is a special case of the previous equation where the concentration N cm -3 increases with distance z traversed and air traverses a length z meters of tunnel in z/v seconds accumulating a concentration N(z) molecules of ozone

Results ● Steady state results for air exiting regions ● Assumed ventilation rates – 10 m 3 s -1 for the main tunnel – 0.2 m 3 s -1 for the enclosures ● Parts per million conversion requires – air density of kg m -3 – molecular weight of g mol -1 – Avogadro constant N A =  N O3 (ppm)4.3    Encl. 2Encl. 1Tunnel

Results ● Concentration kinetics using averaged dose – assumes ‘magic ventilation’ where air is not considered to travel to the ventilation point through a radiation environment. – only useful to compare growth rates etc TunnelEnclosure 1Enclosure  ppm 5.1  ppm2.5  ppm 25 mins 300 mins