Accelerators to push the Envelope

Peter McIntyre

Texas A&M University

Accelerator-Based CosmologySince 1974, colliders have been the main frontiers for discovery in particle physics:

Hadron colliders:

Lepton colliders:

We must continually re-invent the means for discovery

As we complete the reach of each collider, we must improve the technology so that we can increase the energy x2-7, increase the luminosity x4-50, for the same cost as the last collider!

This Moore’s Law for accelerators requires new miracles from each generation.

Today some have viewed the limits from superconductors and synchrotron radiation to make LHC the last hadron collider;

And the limits from expensive structures and wakefields as making ILC the last lepton collider.

Good news: fresh miracles are on the way!

?

ILC

1) LHC Tripler

25 T dipole field; same tunnel, same detectors; new injector:

X3 collision energy; x10 luminosity

LHC is a tool for discovery in high energy physics

• Higgs sector• Supersymmetry / Supergravity• New gauge couplings

The Higgs boson and the spectrum of sparticles should be discovered at LHC, unless…

The flood of precise data from astrophysics suggests that the gauge fields of nature may be far more complex than the picture of the Standard Model + Supergravity

Can we extend the energy reach for direct discovery of new gauge fields?

Hadron colliders are the only tools that can directly discover gauge

particles beyond TeV • Predicting the energy for discovery is perilous.• Example: for a decade after discovery of the b

quark, we ‘knew’ there should be a companion t quark. But we couldn’t predict its mass. Predictions over that decade grew (with the limits) 2040 80 120 GeV

• 4 colliders were built with top discovery as a goal.• Finally top was discovered at Fermilab – 175 GeV!• In the search for Higgs and SUSY, will history

repeat?

Evolution of the gluon spectrum

Dutta 2004Triple the energy – double the mass reach

Assumptions:

•Luminosity grows x3 with adiabatic damping

•Luminosity needed to produce a given number of particles of mass m (assuming gauge couplings constant) scales with m2

•So twice the mass scale requires 4/3 the luminosity.

Discovery of sparticles

• Ellis et al have calculated the masses of the lightest 2 visible sparticles in minimum supersymmetric extension of the Standard Model (MSSM), constrained by the new results from astrophysics and cosmology.

= observable in WIMP searches (> 10-8 pb)

X = observable at LHC

= only observable at LHC Tripler

= constrained by WMAP, , and lab data

Higher field requires new superconductor, handling immense stress loads

NbTi Nb3Sn Bi-2212

Bi-2212

860

880

900

920

940

960

980

1000

0 20 40 60 80 100 120 140 160 180

Stress (MPa)

Cri

tica

l Cur

rent

(A

) Face Loading

Unload

4 T, 4.2 K

Nb3Sn

Cost today: NbTi $100/kg

Nb3Sn $1,000/kg

Bi-2212 $2,000/kg

Both Nb3Sn and Bi-2212 are brittle!

Texas A&M and LBNL are developing 16 T dipoles using Nb3Sn

First successful 16 T model dipoles in 2004. Both groups are building and testing a succession of short model dipoles to mature the technology.

LARP will build a first long Nb3Sn quadrupole by 2010.

Holik’s thesis

New tricks make Nb3Sn feasible

Stress in Pa for TAMU2

Stress Management Flux plate suppression of multipoles

New Nb3Sn dipole technology: stress management, flux plate, bladder preload

Extend to 24 Tesla:Bi-2212 in inner (high field) windings,

Nb3Sn in outer (low field) windings

Dual dipole (ala LHC)

Bore field 24 Tesla

Max stress in superconductor 130 MPa

Superconductor x-section:

Nb3Sn 26 cm2

Bi-2212 47 cm2

Cable current 25 kA

Beam tube dia. 50 mm

Beam separation 194 mm

Magnet issues• Nb3Sn windings must be reacted at 650 C in argon for a

week to form the superconducting phase.• Bi-2212 windings must be reacted at 850 C in oxygen,

~10 minute excursion to partial melt, T ~2 C• How to do both on one coil???

– Wind Bi-2212 inner windings, do heat treat.

– Control fast excursion to partial melt using ohmic heating in coil itself and/or modulation of pp O2.

– Then wind Nb3Sn outer windings, stress management structure isolates the ventilation of the two regions

– React the Nb3Sn with Ar purge, hold O2purge on Bi-2212.

• Quench protection – need to investigate microquench stability of Bi-2212, very different quench strategy from that with all-Nb3Sn dipoles.

Accelerator Issues• Synchrotron radiation: power/length

critical energy

– Use photon stop:Instead of intercepting photons at ~10 K along

dipole beam tube, intercept between dipoles on room-temperature finger.

– Soft X-rays actually easier to trap that hard UV

24 /~ IEP

/3EEc

LHC: E = 7 TeV P = 0.22 W/m Ec = 44 eV (hard UV) scatters, desorbs

LHC Tripler: E = 20 TeV P = 14 W/m Ec = 1.2 keV (soft X-ray) absorbs!

-10

-5

0

5

10

15

0 100 200 300 400 500 600 700z(mm)m

ult

ipo

lesx

104

b2 @ 24T

b2 @ 3 T

Photon Stop

• Photoemission yields vanish for E > 100 eV

Vertical penetration through flux return (coils have clearance)

Effect on <b3> ~10-5 cm-2

Photon stop swings:clears aperture at injection energy,

collects light at collision energy

Injection Collision

150 W/stop collected @ 1 W/cm2

heat transfer to Liquid Xe (160 K)

Same refrigeration power for Tripler as for LHC!

Rapid-cycling Injector for LHC

• For luminosity upgrade of LHC, one option is to replace the SPS/PS with a rapid-cycling superconducting injector chain.

• 1 TeV in SPS tunnel 1.25 T in 25T hybrid dipole: flux plate is unsaturated, suppression of snap-back multipoles at injection.

• SuperSPS needs 5 T field, ~10 s cycle time for filling Tripler > 1 T/s ramp rate

• A pacing issue for design is AC loss during ramp

Block-coil dipole:

Cables are oriented vertically:

Result: minimum induced current loop, minimum AC losses

Again block-coil geometry is optimum!

nB ˆ cos dipole:

Cables are oriented azimuthally:

Result: maximum induced current loop, maximum AC losses

nB ˆ

We demonstrated this suppression of AC losses in TAMU2 test!

Nb3Sn Super-SPS dipole for rapid-cylcing TeV injector to LHC

6 T block-coil suppresses extrinsic losses

-flux plate suppresses snap-back

Bronze-process fine-filament wire suppresses intrinsic losses

- Lowest cost Nb3Sn wire

Efforts until now have concentrated on NbTi cos dipoles – misses on both counts.

This is an unexpected bonus from high-field magnet development.

Liquid He channel

05

101520253035404550

0 5 10 15 20 25field strength (T)

coil

are

a (

cm2 )

quadratic B dependence

RHIC (7 cm)

Tevatron (5 cm)

Pipe (2 cm)

SSC (5 cm)

LHC (7 cm)

microbore (3x2 cm)

TAMU4 (3 cm)

LHC Tripler(5.6 cm)

Magnets are getting more efficient!

NbTi

Nb3Sn

Bi-2212

SuperSPS

HD2

2) TeV e+e- linac collider

• ILC utilizes 1.3 GHz superconducting cavities to accelerate the beams.

• It takes ~17 km of these structures, all driven to 35 MV/m gradient, to produce a collision energy of 500 GeV.

• The cavities must be driven ~1% of the time to provide adequate luminosity – a huge operating cost for electricity and for cryogenics.

Pacing challenges for cost/performance

• The linac cavities and associated cryogenics dominate the capital cost.– How to attain pristine surfaces on inside when you can’t

reach them after each cavity string is welded?– Can we push gradient to reduce length?– Each module must be immersed in superfluid He.– Lorentz detuning requires that every cavity be deformed in a

feed-forward control to keep it on resonance.

• The on-time of cavities dominates the operating cost.– The bunch spacing is limited by long-range wake fields.– How to kill deflecting modes so bunches could be closer?

circumferencialelectron beam welds

TESLA superconducting cavities are made by forming Nb foil, then e-beam welding, then cleaning inside the 9-cell string.

Welds alter grain structure, affects Imax @ waist, Emax at neck.

Difficult to clean, QC inside 9-cell module. ILC needs 17,000 9-cell modules!

Qa ~ 1010 Ga ~ 35 MV/m

Cost and performance are driven by the technology of the superconducting cavities

If a cavity string has transverse misalignment, each bunch drives dipole modes which are resonant with Qd ~ Qa.

Higher-order mode (HOM) couplers are used to extract the HOM fields from each module to a termination, spoiling Qd to reduce wake fields on following bunches.

Transverse wake fields cause emittance growth, instabilities

Suppose the same 9-cell module is assembled from polyhedral slices

Accelerating mode: Current flows in r/z, Qa unaffected by normal slits.

Deflecting modes: Current flows in , Qd spoiled by normal slits.

normal joints

sheet currentsr-z sheet currents

Each segment is fabricated, cleaned, polished, QC before assembly

a) flat s.c. strip

b) copper bar drilled with cooling channels

f) weld seams, HIP to bond g) EDM cut to 30o wedge

d) EDM cut contour

c) bend to contour

e) fit s.c. foil to Cu

weld seams on outside simple assembly/ alignment

cooling channels no pool-boiling cryostat

Explosion-bond Nb to Cu

Explosion-bond 1 mm Nb (400 RRR) to 1 cm Cu (OFHC)

Interface is excellent metallurgical bond.

Cold roll to 1 mm final thickness with intermediate heat treats

Anneal to make malleable foil

at joints between hedra E

2.0

12.1

1

/

E :12

6sin6

)/sin(),(

Rm

mmaE

maV

mmm

m

mm

E

reduced to half-value at distance 0.6 mm from joint.

Joint is shielded against breakdown from micro-irregularities.

Numerical modeling of accelerating mode

Qa = 0.8 1010, slot aperture is tapered so that fields damp exponentially into slot.

Numerical modeling of deflecting mode

Unloaded Qd = 105. Couple mode fields through slot into dielectric-loaded cylindrical

waveguide, out to room-temp load

105 lower Qd 300 x lower wake fields bunches closer, less drive power

less emittance growth, higher luminosity

Assemble the polyhedron on an alignment fixture, E-beam weld on outside Cu seams.

weld seams on outside

cooling channels

No potential for damage to Nb surfaces.

Cu provides accessible reference for alignment in cryostat, interconnection.

He refrigeration is provided by closed-circuit flow in cooling channels – No pool cryostat!

Cu provides rigid structure -

No Lorentz detuning!

Polyhedral cavity opens the way for advanced rf superconductorsGurevic proposed a heterostructure of thin films (< penetration depth )of Type II superconductor (NbN) and insulator (Nb2O5) to triple the surface fields compared to pure Nb:

.

Nb

- - - 3Sn (S) and insulator, e.g. Nb2O5 (I)

If film thickness is comparable to (65 nm), each Type II layer conducts sheet current to its limit then passes the rest to the next layer (Analagous to multi-layer magnetic shield).

. -The potential: twice the gradient, same power/length

Pogue’s thesis

So what might the future hold?

• Hadron collider: LHC tripler in same tunnel– 40 TeV collision energy, 1035 cm-2s-1 luminosity

• Lepton collider: Doubler with same length, Quadrupler with same r.f. power

– 2 TeV collision energy• 80 MV/m, 100 ns bunch spacing, 35 km length

• And from there?– I have a tunnel to sell you:

• 54 km circumference, 25 T bend field, s = 160 TeV• And there is a renewed effort towards a multi-TeV

Muon collider (Cline)

LHC x 3

ILC x 4

Texatron

We can continue to invent miracles to push the bounds for discovery if we have the guts and imagination.

All of this could happen, orNone of this could happen

• The core technologies are sound, but the device technology must be developed, proven, and industrialized and the accelerator physics to use it.

• That takes a decade if the effort is supported as a priority.

• Present focus on near-term R&D for today’s machines marginalizes support for the AARD.

• The Marx Panel endorsed the need and opportunity.• It will only happen if the HEP community (you!)

demand it and support its priority to the agencies and laboratories.