Minutes of the 8th LHC Insertions Upgrade Working Group held on 24th January 2008

Present: V. Baglin, O. Brüning, F. Ceruti, S. Fartoukh, P. Fessia, J. Kerby, J.-P. Koutchouk, K.H. Meß, R. Ostojic, V. Parma, H. Prin, L. Tavian, E. Todesco, D. Tommasini, R. Van Weelderen, E. Wildner, F. Zimmermann

Excused: M. Giovannozzi

Invited: F. Butin, S. Gilardoni, M. Mauri, S. Roesler

    1.    News and approval of the minutes of the last meeting

The minutes of the last meeting are approved without any comments.

    2.    Energy deposition in the TAS and inner triplet (E. Wildner, F. Cerutti, M. Mauri, ppt file)

Elena presented first results concerning the energy deposition expected in the new inner triplet of the LHC. These studies were performed in collaboration with the FLUKA team of the AB/ATB group, in particular Francesco Cerutti and Marco Mauri.

Elena first reported on the situation expected for the nominal LHC. The shielding scheme of the current LHC triplets can be briefly described as follows. Only the first inner triplet quadrupole and its MCBX orbit corrector are equipped with a 6.5 mm thick stainless steel absorber which has been accommodated in the 2K helium bath of Q1. The TAS in front of Q1 has an opening of 41.5 mm and an additional mask, the TASB, has been implemented in between Q2b and Q3 in order to protect Q3. With this scheme and for the nominal peak luminosity L0 of 10³⁴ cm^-2s^-1, the peak power density deposited in the magnet coils remains below the  limit of 4.3 mW/cm³ recommended in the current MQX magnets. Those peaks are generally located at the entry of each quadrupole magnets with the exception of Q1, directly protected by the TAS, but still exhibiting at its exit the highest loss peak of the order of 2.5-3.5 mW/cm³. These results were produced in 2007 (report to be published) and benchmarked against earlier simulations performed at Fermilab in 2002 (LPR633) showing an rather good agreement both in terms of local and integrated losses expected in the magnet coils and in the iron.

Based on these simulations and others performed on the nominal LHC triplet, several qualitative conclusions were drawn in order to converge quickly towards an optimized shielding of the new inner triplet. First of all, it was clearly demonstrated that the main mechanism of losses is driven by  the high magnetic field of the inner triplet quadrupoles which affects strongly the transverse excursion of particles of non nominal energy. In this respect, the TAS only protects the entry of Q1 and has a minor impact for the energy deposited  at the exit of Q1 or in the other quadrupole magnets of the inner triplet (the TAS has no impact at all upstream the entrance of Q2b without magnetic field). The non-zero crossing angle has a rather small impact in terms of longitudinal density of losses (increase by 20% of the peaks at the entry of Q2b and Q3). Finally it was stated that a few centimeter thick mask outside the magnet aperture did not affect the energy deposited in the cables of the downstream magnet.

Elena then summarized the results obtained for the new LHC inner triplet proposed in LPR1000, based on 130 mm aperture NbTi quadrupoles  in the so-called symmetric layout. In a first approximation, mechanical and magnetic lengths are assumed to be identical, the coil ends are not modeled and the hard edge approximation is used in order to describe the end field of the magnets. On the other hand, for mechanical reasons, the thickness of the cold bore tube is increased from 1.8 mm (nominal) to 3.45 mm and that of the beam-screen from 0.6 mm to 2 mm in the MQXs (see also minutes of  LIUWG#3). The TAS aperture is increased to 55 mm, as requested by the aperture requirements of the primary beam and just compatible with the 60 mm outer diameter of the beam tube in this region. Finally, the half crossing angle is set to 220 murad (corresponding to a 10 sigma beam-beam separation for a beta* of 25 cm). For a luminosity increased by a factor of 2.5 w.r.t. nominal, and w/o specific shielding devices (but the TAS), the expected losses exceed the limit currently recommended (see above), with energy deposition  peaks at the exit of Q1 and the entry of Q2. The highest peak reaches about 10 mW/cm³ and is azimutahlly located in the crossing plane, depending on the polarity of the crossing angle (i.e. 90 degrees for a positive vertical crossing angle).  The energy density deposited in the magnet coils is then put back to an acceptable level if a 3 mm thick tungsten absorber runs through all quadrupole and corrector magnets equipping the triplet. If the liner is stopped right after the MCBX orbit corrector attached on the non-IP side of Q1, the situation is more or less back to acceptable. This liner would be attached to the Q1 beam-screen in order to extract the deposited power at 20 K rather than at 1.9K as in the present design of Q1 (see Elena's slide for more details on a possible design of a two-in-one  beam-screen/absorber). In this configuration, however, the density of losses  is still 50%  higher than the recommended limit at the entry of Q2a and an additional shielding might be needed in the region close to the beam-pipe between Q1 and Q2a. The total power absorbed by the TAS is 300 W (compared to about 140 W for the nominal LHC), with peaks of density reaching 130mW/cm³. ANSYS simulations, not yet done,  has to be performed to verify the mechanical stability of this device under such hard conditions.

The presentation triggered several questions and comments.

SF commented that these results are very promising in the sense that no specific absorbers might be needed in Q2 and Q3 for which the demand on aperture by the beam is the highest. On the other hand, the Q1 quadrupole shows a very large aperture margin of the order of 30-35 mm in diameter compared to Q2/Q3. SF therefore suggested to sensibly increase the thickness of the liner proposed for Q1 to see if this could not resolve the energy deposition peak observed at the entry of Q2a. Elena agreed to have a look, e.g. with a 13 mm thick liner (i.e. rescaling by a factor 2 w.r.t. to the nominal LHC).

=> ACTION: E. Wildner

SF mentioned that the effect of the MCBX corrector in front of Q1 should to be added in the simulations. This corrector contributes directly  to the generation of the crossing scheme  (integrated strength of about 1 Tm with a polarity of opposite sign with respect to that of the crossing angle). At a later stage, JPK  suggested  to study also the possible effects induced by the triplet misalignments.

RO asked if the TAS length, positioning or material has been reoptimized in this study. EW replied not but arguing that in any case, the TAS only protects the entry and, up to some extent, the body of Q1, and therefore will not solve the problems encountered in the other parts of the inner triplet quadrupoles. RO then asked Elena to comment on the longitudinal density of losses and the total losses  integrated per quadrupole magnets (i.e. not only in the coils). Unfortunately Elena did not have the relevant document at hand during the meeting. This important piece of information will be reported back  at the next meeting.

=> ACTION: E. Wildner

Rather than using a specific material for the absorbers in the simulations (i.e. tungsten presently), SG suggested to start with a so-called "black hole material" (i.e. an ideal material with an infinite absorption capability, available in FLUKA) in order to define  the appropriate locations and the transverse opening of the different shielding devices which should equip the new inner triplet. This approach would have the benefit to considerably reduce the CPU time needed to simulate a given configuration (no low energy particle produced) and, for a given layout, to derive quickly the matrix relating the locations of the liners and the regions of the triplet which are "geometrically" protected by the latter. Then will come the choice of the material and of the  thickness of the proposed absorbers. Several participants agreed that this kind of approach might present some interest. Elena further commented that this approach was already used in some specific cases (e.g. impact of a thin mask at the entrance of Q2a). On the other hand, she showed some reluctance to use this approach for the liner themselves arguing that a black hole material is much too far from any real material and may at the end not help in converging quickly towards an optimized shielding scheme for the new inner triplet.

SG then commented that tungsten is well-known in fixed target physics to produce high flux of neutrons and low energy particles in general. He then asked what was the energy cut-off used in the simulations. FC replied the thermal energy for the neutrons, 1 MeV for electrons and positrons, and 100 KeV for photons and other particles

LT strongly encouraged  to push forward this idea of decoupling the Q1 absorber  from the 1.9K cryogenic system, i.e. to extract the deposited power at 5-20K (or 40-60 K if the operational temperature range of the beam-screen is changed for phase I, see minutes of the LIUWG#4). He then further commented that even if the 4.3 mW/cm³ limit is not exceeded in the coils of Q2 and Q3, it might be preferable anyway to implement similar liners in these quadrupoles in order to absorb the losses at the level of the beam-screen itself. Indeed, knowing that 1W @ 1.9K  is equivalent to 5W @ 5-20 K in terms of power needed for refrigeration, and working under the assumption that the cryogenic capability will not be upgraded for phase I (400 W @ 1.9 K reserved for the triplet, see minutes of the LIUWG#4), this may help in pushing the LHC up to the ultimate intensity and gain in luminosity (despite of a small re-in crease of beta* to compensate for the loss of aperture in Q2/Q3 and assuming no limitation coming from the collimator impedance).

    3.     A.O.B. and follow up of actions

Date and topics for the next meeting not yet decided.

S. Fartoukh and R. Ostojic