The CERN Large Hadron Collider LHC Abstract

Home , CERN

EUROPEAN ORGANIZATION FOR NUCLEAR RESEARCH

Europ ean Lab oratory for Particle Physics

Large Hadron Collider Pro ject LHC Project Rep ort 83

The CERN Large Hadron Collider LHC

Eb erhard Keil

Abstract

The Large Hadron Collider Pro ject LHC was approved by the CERN Council in December

1994. Commissioning with b eam will start in the second half of 2005. The pap er discusses

the most imp ortant parameters, the general layout of the LHC o ctants, and the status of the

exp eriments ALICE, ATLAS, CMS and LHC-B. Also describ ed are some asp ects of the b eam-

b eam interactions, the transverse coupled-bunch instability,aswell as the technical developments

of the cryogenics, the tests of half an arc cell, and the new civil engineering needed.

Invited talk at the 34-th Eloisatron Workshop, Erice, Italy, 4{13 Novemb er 1996

Administrative Secretariat

LHC Division

CERN

CH-1211 Geneva23

Switzerland

Geneva, Decemb er 6, 1996

1 INTRODUCTION

The LHC pro ject was approved by Council in December 1994. Big contracts start

b eing placed. In Septemb er 1996, the Finance Committee approved contracts for the whole

50000 t of steel supply, the sup ervision of the civil engineering work, and eight magnet

measuring b enches. I assume for the remainder of the milestones that enough funds are

available to construct the LHC in a single stage. By the end of 1999 most of the big

contracts will have b een placed, the prices will be known, and a nal decision on the

con guration can be taken. We assume that LEP will stop op eration at the end of 1999.

Continuing LEP op eration into 2000 must be justi ed on scienti c grounds and money

must be found. Dismantling LEP will start in Octob er 2000 when the civil engineering

work for LHC is advanced such that it b ecomes necessary to break into the tunnel, in

particular for the ATLAS and CMS caverns. Injection tests are foreseen from Octob er

2003. Commissioning with b eam will start in the second half of 2005. The latest conceptual

design rep ort [1] for the LHC was issued in Octob er 1995.

The remainder of this pap er is organised as follows: Chapter 2 discusses the most

imp ortant parameters, the overall layout, the state of the exp eriments, and the layout

and optical functions in the neighb ourho o d of Pit 5, Chapter 3 the b eam-b eam fo otprints

and the choice of the crossing angle, and Chapter 4 the cryogenics, the string tests of

half an arc cell, and the new civil engineering needed are presented in Chapter 5. Finally,

Chapter 6 contains the conclusions.

2 PARAMETERS, LAYOUT, EXPERIMENTS, OPTICAL

FUNCTIONS

This chapter is divided into three sections, discussing the most imp ortant parame-

ters, the overall layout of the LHC, and the detailed layout and optical functions in the

neighb ourho o d of the CMS exp eriment in Pit 5.

Table 1: LHC Parameters

Circumference C 26659 m

Energy E 7 TeV

Dip ole eld B 8.4 T

Bunch spacing s 25 ns

Bunch p opulation N 10

Bunch radius = 16 m

x y

Bunch length 75 mm

Beam-b eam parameter 0.0034

1 1

Luminosity L 10 nb s

Distance to nearest quadrup ole ` 23 m

Events/collision n 19

2.1 Parameters

Tab. 1 shows the LHC parameters. The circumference C is that of LEP, and known

to even more digits than shown. The maximum energy E is a round gure, achieved at the

dip ole eld B listed. The bunch spacing s corresp onds to 10 RF wavelengths. Together

with the distance from the interaction p oint IP to the separating dip oles it determines the

numb er of parasitic collisions, ab out 15 on either side of the IP. The bunch p opulation N 1

and the bunch radii and are shown at the interaction p oint IP. They are adjusted

x y

such that the b eam-b eam tune shift parameter falls into range b elieved to b e achievable

from exp erience with other hadron colliders which were or are in op eration [2], and that

the luminosity L is in a range which the exp eriments b elieve they can handle. The total

b eam-b eam tune spread from the nearly head-on collisions and from all parasitic collisions

should b e small enough to t b etween nonlinear resonances of order up to twelve. Not all

the space ` between the IP and the front face of the nearest quadrup ole is available to

the exp eriments. At the assumed inelastic non-di ractive cross section =60 mb, the

number of events in a single collision is n = 19. c

Low ß (pp) High Luminosity CMS RF & Future Expt. Dump

Octant 5 Octant 6

Octant 4 Octant 7 Cleaning Cleaning Octant 3

Octant 2

Octant 8 Octant 1

ALICE Injection LHC-B

Injection Low ß (Ions) ATLAS Low ß (B physics) Low ß (pp)

High Luminosity

Figure 1: Layout and Exp eriments

2.2 Overall Layout and Exp eriments

Fig. 1 shows the layout of the LHC. The pits are at the centres of the o ctants. The

two LHC rings cross in Pits 1, 2, 5, and 8. The circumferences of the two rings are the same,

since b oth rings have four inner and four outer arcs. The two large exp eriments, ATLAS

and CMS, are diametrically opp osite in Pits 1 and 5, resp ectively. Both are approved

exp eriments with a cost ceiling of 475 MCHF each. Technical prop osals were published

in 1994 [3, 4]. The LHCC exp ects technical prop osals for the subsystems. The heavy-ion

exp eriment ALICE will b e in Pit 2. The technical prop osal for the core exp eriment [5] was

published in 1995. The LHCC is waiting for the technical prop osal for the muon arm. The

LHC-B exp eriment is dedicated to the study of CP violation and other rare phenomena

in the decay of Beauty particles. It uses colliding b eams and a forward detector, contrary

to the HERA-B exp eriment which uses a single b eam and a gas jet target. A letter of

intent [6] has b een submitted to the LHCC. The LHCC wants an R&D programme for

the detector. The two b eams are injected into LHC into outer arcs upstream of Pits 2 and 2

8. The b eam cleaning insertions to steer the b eam halo into staggered sets of collimators

rather than the sup er-conducting magnets are in Pits 3 and 7. Pit 4 houses the RF system.

Pit 6 is reserved for the b eam dumping system.

2.3 Layout and Optical Functions near Pit 5

Fig. 2 shows the layout and optical functions near Pit 5. The mimic diagram shows

the LHC layout schematically over ab out 500 m in the neighb ourho o d of Pit 5. The

low- interaction p oint IP5 and CMS are close to the centre. On either side of IP5 is a

quadrup ole triplet, actually consisting of four quadrup oles. Because of the antisymmetry

designed into LHC, the rst quadrup ole of the triplet fo cuses horizontally on the left,

and defo cuses on the right, and similarly for all other quadrup oles. The b oxes b ehind the

rst triplet are the dip ole magnets which rst separate the two b eams, and then make

them parallel again at the correct distance of 194 mm at 1.9 K. Just b efore the disp ersion

suppressors there is a second triplet. The graph shows the optical functions in black,

in red. They are prop ortional to the horizontal and vertical b eam radii. The and

antisymmetry of LHC makes them swap values when passing through IP5. Their values

at IP5, = =0:5 m, are to o small to show clearly. Their maximum values in the rst

x x

triplet are quite large. The green curve shows the horizontal disp ersion D . It is matched

to D = D = 0 at IP5, and has asymmetric nonzero values b ehind the rst separating

dip oles. The layout and optical functions near Pit 1 are the same as those near Pit 5. The

layout and optical functions near Pits 2 and 8 are very similar.

LHC4.2 near IP5 HP/UX version 8.17/3 18/06/96 09.12.11

) 70. .40 1/ 2 63. .35 (m) x (m D 56. .30 1/ 2 49. .25 42. .20 35. .15 28. .10 21. .05 14. 0.0 7. -.05 0.0 -.10 13.05 13.25 13.45 s (m)

E /p0c= 0. Table name = TWISS

∗10∗∗( 3)

Figure 2: Layout and Optical Functions near Pit 5. The interaction p oint is close to the

centre of the abscissa. The diagram ab ove the graph schematically shows the horizontally

fo cusing (defo cusing) quadrup oles as rectangles ab ove (b elow) the axis, and the separating

dip oles as centred rectangles. The curves are the square ro ots of the horizontal and vertical

amplitude functions in black and red and the horizontal disp ersion in green. 3

3 BEAM-BEAM EFFECTS

The tune spread caused by the nonlinearityof the b eam-b eam force is usually dis-

played in the form of \fo otprints" which show the horizontal and vertical tune shifts Q

and Q for combinations of horizontal and vertical b etatron amplitudes a and a . Fig.3

y x y

shows the fo otprint for the head-on collisions of round b eams [7]. The tune shifts Q

and Q are expressed in units of the b eam-b eam tune shift parameter .Particles with

vanishing b etatron amplitudes have the highest tune shift, particles with large b etatron

amplitudes have the lowest tune shift. The graph is symmetrical around the diagonal b e-

cause of the assumed round b eams and equal -functions at the head-on collision p oint.

Figure 3: Amplitude-dep endent b eam-b eam tune shifts Q = and Q = . Lines mark

x y

amplitudes of horizontal and vertical b etatron oscillations at 0 ...4 .

Fig. 4 shows the fo otprints in the case of two round b eams colliding at a full hori-

zontal crossing angle of 200 rad in a single interaction p oint, including the contributions

of the long-range collisions on either side of the nearly head-on collision p oint [8]. The

(colour) co ding is shown in the top right corner. The small (violet) fo otprint in the lower

right corner is the contribution of the nearly head-on collisions. The (red) fo otprint on

the left side is the contribution of all long-range collisions. Since the two b eams are hor-

izontally separated there, the horizontal tune shift is negative, while the vertical one is

p ositive as for head-on collisions. The (blue) fo otprint is the sum of the head-on and long-

range tune shifts. Bunches at the head and tail of the bunch trains meet bunches of the

opp osite b eam only on one side of the interaction p oint, are called PACMAN bunches,

receive only half the long-range tune shifts, and have the (green) fo otprint.

Fig. 5 show the fo otprints in the case of two round b eams colliding in two inter-

action regions and the neighb ouring long-range collision p oints, with full crossing angles

of 200 rad, once in the horizontal and once in the vertical plane [8]. The fo otprint of

the nearly head-on collisions is twice as big as in Fig. 4, as exp ected. The tune shifts

caused by the long-range collisions have the opp osite sign in the two interaction regions.

Therefore, their e ects cancel to a remarkable degree, and the total size of the fo otprints 4 Jul 6 1995 LHC V4.1 Tune Footprint : IR2, horizontal crossing, 200murad crossing angle, no offset ∆ 3.5 Qv Head On & Long Range (PACKMAN) ξ Head On & Long Range Long Range 3 Head On

2.5

1.5

0.5

0 ∆ -2 -1.5 -1 -0.5 0 0.5 Qh 1

john miles ξ

Figure 4: Amplitude-dep endent b eam-b eam tune shifts Q = and Q = for various

x y

collisions. Lines mark amplitudes of horizontal and vertical b etatron oscillations at 0 ...4

. The (colour) co ding is shown in the top right corner. The small (violet) fo otprint in the

lower right corner is the contribution of the nearly head-on collisions. The (red) fo otprint

on the left side is the contribution of all long-range collisions. The (blue) fo otprint is

the sum of the head-on and long-range tune shifts. The (green) fo otprint applies to the

PACMAN bunches.

Nov 2 1995 LHC V4.1 : Alternating Crossing ATLAS & CMS, Angle 200murad, No offset ∆ 3 Qv Head On & Long Range (PACMAN) ξ Head On & Long Range Long Range 2.5 Head On Footprints for 4 sigma

1.5

0.5

-0.5

-1 ∆ -1 -0.5 0 0.5 1 1.5 2 2.5 Qh 3

john miles ξ

Figure 5: Amplitude-dep endent b eam-b eam tune shifts Q = and Q = for alternating

x y

collisions. Lines mark amplitudes of horizontal and vertical b etatron oscillations at 0 ...4

. The (colour) co ding is shown in the top right corner. 5

is ab out the same as in Fig. 4, although there are twice as many collisions.

The choices of the amplitude functions = at the interaction p oint and of the

x y

crossing angle are correlated. The arguments entering into the choice [8] are related to

(i) the relative size of the fo otprints due to the head-on and parasitic collisions and their

p ossible comp ensation by alternating horizontal and vertical crossings, (ii) the drop of

the luminosity and of the nearly head-on b eam-b eam tune shift caused by the crossing

angle, (iii) synchro-b etatron resonances driven by the crossing angle, (iv) the ap erture and

gradient of the low- quadrup oles, (v) the so-called hourglass e ect when the amplitude

functions and bunch length b ecome comparable, (vi) the chromaticity caused by the low-

insertion and its comp ensation. A discussion of these arguments in the framework of the

Eloisatron workshop can b e found elsewhere in these pro ceedings [9].

1000

500 ) Ω

(M 0 ⊥ Z 〉 y 〈β Real Part Imaginary Part

−500

−1000 0246810

Frequency (GHz)

Figure 6: Real and imaginary part of the total pro duct of the average amplitude function

and transverse broad band imp edance in the LHC

4 TRANSVERSE COUPLED-BUNCH INSTABILITY

The transverse coupled-bunch instability, driven by the imp edance of various com-

p onents, is the most imp ortant instability in the LHC[10]. The following comp onents

contribute to the broad-band imp edance (i) direct space charge, (ii) shielded b ellows and

monitor tanks, (iii) b eam p osition monitors, (iv) ab ort kicker magnets, having a ceramic

chamb er coated with Ti on the inside. Over some 90% of the LHC circumference the b eam

sees the b eam screen with an inner Cu layer and an outer stainless steel layer, pumping 6

slots and wall resistivityat5to20K.Over the remaining 10% of the LHC circumference,

the b eam sees a Cu chamb er at ro om temp erature. The contributions of the b eam screen

and Cu chamb er to the resistive-wall imp edance are roughly equal. Fig. 6 shows the total

pro duct of the average -function and the transverse broad band imp edance [11].

The RF cavities, the longitudinal feedback cavities, and the chamb ers of the LHC

exp eriments contribute to the narrow-band imp edance. Their imp edances are computed

by programs. Both the broad-band and the narrow-band transverse imp edances are used

in a computer program which assumes linear b etatron and synchrotron oscillations, and

nds the growth rates of the coupled-bunch mo des, with or without feedback systems,

but in the absence of Landau damping [11]. Fig. 7 shows the results at 450 GeV without

feedback, where the growth rates are highest. The abscissa is the coupled bunch mo de

numb er. The internal bunch mo des are distinguished by a colour co de. The m = 0 mo de is

shown in red (dark grey in the printed version). The broad curve on the right edge of the

graph is caused by the wall resistivity. The spikes are driven by narrow-band imp edances.

The large spike on the left side is driven by two higher-order mo des of the RF cavities.

Insp ecting Fig. 7, one asks oneself what growth rates are acceptable, b ecause they

can be overcome by either feedback or Landau damping. A feedback system with the

assumed half-bandwidth of 730 kHz [11] damps many of the coupled-bunch mo des driven

by the wall resistivity, but has rather little e ect on most of the spikes. Because of the

emittance blow-up caused by noise, the feedback system will only be used for lling and

ramping, but not while the b eams are in collision. Landau damping must be invoked for

stabilising the mo des not damp ed by the feedback system during lling and ramping, and

all mo des in collision. The tune spreads in the b eam due to the naturally present and

delib erately added nonlinearities of the lattice should be large enough to Landau damp

the transverse coupled-bunch instability at the growth rates shown in Fig. 7.

The analysis of the imp edances, their consequences for the stability of the bunches,

and the ways for avoiding instabilities is an iterative pro cess of which I describ ed the rst

step. An imp edance database has b een installed [12], the damping of the higher-order

mo des and the parameters of the feedback system are under discussion.

5 ENGINEERING

In this section I discuss the cryogenics, the string tests, and the new civil engineering.

The dip oles are presented elsewhere in these pro ceedings [13].

5.1 Cryogenics

The sup er-conducting magnets are immersed in a bath of sup er- uid He at 1.9 K,

pressurised at ab out 1 bar. The heat is transferred by heat exchangers, consisting of a

tub e containing owing saturated He I I, and running through a half cell of the LHC arcs.

It was checked in the string tests, that the liquid and gaseous He may ow in the same

or in the opp osite direction in this tub e. The LHC tunnel is inclined with resp ect to the

average vertical by at most 14.2 mrad. Allowing He ow in either direction avoids the

complication of changing the orientation of the co oling lo op whenever the tunnel slop e

changes sign. Contrary to earlier designs, the cryogenic uids are supplied by a separate

cryogenics line which runs along the tunnel walls and is connected to the cryostats every

half cell. The cryogenics supply line is fed from the four even pits where much equipment

is available from LEP2 [14].

In collab oration with CEA in France, key technologies for high-capacity refrigeration

at 1.8 K are b eing develop ed. [15]. This includes very low pressure heat exchangers, cold 7 10−3

10−4 0 ω / 10−5 Ω Im

10−6

10−7 0.0 1000.0 2000.0 3000.0 Multibunch Mode Number 880 1118 1269 1104 1254 1325 719 517 749 824 942 612 809 363 1028 1301 1059 1369 1368 1243 754 589 507 666 539 738 777 655 610 599 717 635 673 631 711 629 588 506 705 661 538 616 731 609 688 603

508

Figure 7: The upp er part shows the transverse growth rates as a function of the coupled-

bunchmodenumb er. The red (dark grey in the printed version) curve shows the internal

bunchmodem= 0. The other colours show higher internal bunch mo des. The lower part

asso ciates the p eaks of the growth rate with particular narrow-band resonances. 8

volumetric and hydro dynamic compressors to be used as comp onents of practical and

ecient thermo dynamic cycles. Prototyp es of such machines from Europ ean industry

were tested in the lab oratory.

Figure 8: Pressure rise in dip ole string tests during a quench. The right axis is the time in

seconds. The left axis shows several p oints along the string. The ordinate is the pressure

in bar.

5.2 String Tests

The LHC test string [16] consists of a quadrup ole of 3 m length and three dip oles of

10 m length each, the cryogenic equipment needed to supply the magnets with cryogenic

uids and gases, the cryogenic valves and short circuits for the electrical bus-bars, the

power converters for the magnets, and control and diagnostic equipment. The test string

has b een in op eration for twoyears, with more than 7500 hours at 1.8 K. It was cycled 2150

times, simulating several years of routine LHC op eration. Its purp ose is the exp erimental

validation of the cryogenic co oling scheme and the development of the quench detection

and magnet protection systems. The 1.8 K co oling with sup er- uid helium was tested

in steady state conditions and during transients. Much was learned on quench detection

and magnet protection from the 20 natural and 64 provoked quenches so far; 35 of them

o ccurred at or ab ove the nominal eld. In addition, there were ab out 15 quenches in

the magnets b efore they were installed in the string. The temp erature increases during 9

ramping upwards at 10 A/s and downwards at 130 A/s were 6 mK and 50 mK, resp ectively,

small enough not to quench the magnets. Simulating a heat load due to particle losses at

1 W/m caused temp erature increases less than 30 mK.

The pressure rise during quenches was measured for various con gurations of pres-

sure relief valves, and for various delays in op ening them, and is shown in Fig. 8. This

made us reduce the number of pressure relief valves from four to two in a half cell. The

delays in op ening the pressure relief valves are long enough that commercially available

valves can be used. The string is installed on a slop e simulating the slop e of the tunnel.

By reversing the ow of liquid He, the ow of liquid and gaseous He in the same and

opp osite direction was tested, and b oth were found to b e p ossible. This makes the layout

of the co oling lo ops indep endent of the lo cal slop e of the tunnel.

5.3 New Civil Engineering

The civil engineering is shown in Fig. 9. It is concentrated around the interaction

p oints. New caverns and access shafts are needed for the ATLAS and CMS exp eriments

in Pits 1 and 5. The ATLAS cavern is so large that the CERN Main Building would t

into it. Less work is needed in Pits 2 and 8 where caverns and shafts already exist. New

transfer tunnels are needed from the SPS to the LHC; TI2 for the clo ckwise and TI8

for the anti-clo ckwise b eam, resp ectively. The transfer lines will be equipp ed with ro om

temp erature magnets. The tunnels for the b eam dumps near Pit 6 are also new.

RF CMS 4 5 6 Cleaning I Dump 3

LEP/LHC

SPS 2 Cleaning II PGC2 PM18 ALICE BA4 TI 2 8 Injection BA6 TI 8 1 LHC-B ATLAS Injection

Future constructions Existing underground buildings

AC/HF 238

Figure 9: New Civil Engineering. The existing LEP tunnel, exp erimental halls and access

shafts are shown in bright grey. The new exp erimental halls for ATLAS and CMS with

their access shafts, the new transfer tunnels and the new b eam dump tunnels are shown

in red (or dark grey in the printed version). 10

6 CONCLUSIONS

Since the publication of the Yellow Bo ok [1] progress was made in many areas of

the LHC design. There still are many ongoing studies of which I only mention a few. My

colleagues in accelerator physics study the e ects of the errors of the magnetic elds in

the sup er-conducting magnets on the dynamic ap erture, mostly by computer simulation,

i.e. tracking [17]. These errors are caused by the arrangement of the coils, by fabrication

tolerances, ampli ed by the 2:1 design, and by p ersistent currents at the injection eld

of only ab out 0.5 T, etc. We are also concerned ab out designing an LHC lattice which is

robust enough to be op erated with ease. Apart from the dip oles, other magnets need to

be built. The insertion quadrup oles are particularly challenging, b ecause the large b eam

size and the high sensitivity of the b eam to their errors [18, 19, 20]. Studies of the sup er-

conducting cable continue to nd a cable which is mechanically stable and has a high and

uniform inter-strand contact resistance [21].

Acknowledgement

This status rep ort is based on the work of the very many p eople who are now

working on the LHC pro ject.

References

[1] The LHC Study Group, LHC { The Large Hadron Collider { Conceptual Design,

CERN/AC/95-05 (LHC) (1995).

[2] E. Keil, Beam-Beam E ects in Electron and Proton Colliders, Particle Accelerators

27 (1990) 165.

[3] ATLAS: Technical Prop osal for a General-Purp ose pp Exp eriment at the Large

Hadron Collider at CERN, CERN/LHCC/94{43 (1994).

[4] CMS: The Compact Muon Solenoid { Technical Prop osal, CERN/LHCC/94{38

(1994).

[5] ALICE { Technical Prop osal for A Large Ion Collider Exp eriment at the CERN LHC,

CERN/LHCC/95-71 (1995).

[6] LHC-B: A Dedicated LHC Collider Beauty Exp eriment for Precision Measurements

of CP-Violation, CERN/LHCC/95-5 (1995).

[7] W. Herr, LHC Seminar on 29 Feb 96.

[8] W. Herr and J. Miles, Optimising the LHC Interaction Region to Obtain the Highest

Possible Luminosity,EPAC 1996 (CERN LHC Pro ject Rep ort 40, July 1996).

[9] E. Keil, Summary of Performance Working Group, these pro ceedings.

[10] F. Ruggiero, Particle Accelerators 50 (1995) 83.

[11] J.S. Berg, Transverse Instabilities in the LHC, LHC Pro ject Rep ort 16 (1996).

[12] O. Bruning, ZBASE User's Guide Version 1.1 { An Imp edance Data Base Program,

CERN SL 96-69 (AP) (1996).

[13] M. Bona and R. Perin, Sup erconducting Magnets for the LHC, these pro ceedings.

[14] V. Benda et al., Conceptual Design of the Cryogenic System for the Large Hadron

Collider (LHC), EPAC 1996 (LHC Pro ject Rep ort 12, July 1996).

[15] Ph. Lebrun, L. Tavian, G. Claudet, Development of Large-Capacity Refrigeration

at 1.8 K for the Large Hadron Collider at CERN, Pro c. KRYOGENIKA'96, Praha

(1996) 55 (LHC Pro ject Rep ort 6, May 1996).

[16] A. Bezaguet et al., The LHC Test String: First Op erational Exp erience, EPAC 1996

(LHC Pro ject Rep ort 32, July 1996). 11

[17] J. Gareyte, Accelerator Physics Issues of the LHC, EPAC 1996 (LHC Pro ject Rep ort

10, June 1996).

[18] G.A. Kirby et al., Progress in the development of the 1 m mo del of the 70 mm ap erture

quadrup ole for the LHC low- insertions, ASC Pittsburgh '96 (LHC Pro ject Rep ort

68, Octob er 1996).

[19] G.A. Kirby et al., Design study of a sup er-conducting insertion quadrup ole magnet

for the large hadron collider, ASC Pittsburgh '96 (LHC Pro ject Rep ort 75, November

1996).

[20] A. Zlobin et al., Development of High Gradient Quadrup oles for the LHC Interaction

Regions, ASC Pittsburgh '96.

[21] D. Richter et al., DC Measurements of Electrical Contacts b etween Strands in Sup er-

Conducting Cables for the LHC Main Magnets, ASC Pittsburgh '96 (LHC Pro ject

Rep ort 67, Octob er 1996). 12