present: Niels van Bakel, Themis Bowcock, Jo van den Brand, Henkjan Bulten, Paolo Chiggiato, Ian Collins, Paula Collins, Hans Dijkstra, Martin Doets, Massimiliano Ferro-Luzzi, Oswald Gröbner, Georg von Holtey, Tjeerd Ketel, Sander Klous, Juan Ramon Knaster, Jean-Michel Laurent, Maria Pilar Lozano, Tatsuya Nakada, Adriana Rossi, Vladimir Rouzinov, Thomas Ruf, Raymond Veness, Hans de Vries.

Purpose: discussion meeting on open questions concerning the vacuum system of the LHCb Vertex Detector (VD).

Minutes of 21/9/2000:  there was only a minor comment (Thomas Ruf) regarding the clean area around the VD.

Current layout of LHCb beam pipe mechanics and vacuum: (presented by Juan Ramon Knaster)

JRK presented the current layout, which is still under discussion. GvH mentioned that new background estimations by Monte-Carlo were carried out within LHCb and will be discussed in another meeting. JRK also presented the current view of installation (VD and RICH1) and beam pipe heating procedures, and some time estimates (by Daniel Lacarrère) associated with these operations (for instance, about 48 hours are needed to retract all detectors from the beam pipe to prepare for heating of the NEGs).

NEGs: recent R&D progress: (presented by Paolo Chiggiato)

PC presented the status of R&D on NEG coatings at CERN. Recently, a substantial increase of the "true" surface site density has been obtained by coating at 300 C instead of the former 100 C. Next, 200 C (important for Al chambers) will be investigated. PC also presented a routine procedure for venting NEG-coated vacuum chambers with an inert gas which preserves NEG pumping properties. Vacuum chambers can be coated at CERN (currently up to 6-7 m long, maybe more in the future). Coating thicknesses of up to 10 um (the peeling limit) can be obtained.
AR commented that a gas purity of a few tens of ppb is necessary for the ultrapure inert gas not to contaminate the NEG (10 ppb, or 0.06 mbar*l, corresponds to a loss of about 1% of the maximum adsorption site density). In the case of the LHCb experiment, if there is a large unbaked surface area in the VD tank or if the 1ary and 2ary vacua are in communication during venting (via f.i. the spurious conductance of the gravity-controlled valve), it is not worthwhile to inject ultrapure gas, as it will be rapidly contaminated.

Get his (PowerPoint) presentation here.

Dynamic vacuum simulation results: (presented by Adriana Rossi)

NEG saturation: AR presented an estimation of the time needed for gas flow from the unbaked VD vacuum tank to saturate the beam pipe NEGs. Her estimation results varies from 2 days to 3 weeks depending on the assumptions (residual gas spectrum with beam on, capacity of NEG surfaces, etc.).
It was remarked that this estimation is conservative, because it assumes a constant conductance (that of a an aperture of ~5 cm diameter) between VD tank and beam pipe. In reality, as the saturated surface extends more and more into the LHCb beam pipe, the geometrical flow resistance towards the unsaturated areas increases with time (longer and longer tube). It is however not possible to conclude with certainty that the NEG life time would be larger than the relevant "threshold" life time of 9 month (i.e. one LHC operational year). For this, a monte-carlo simulating gas trajectories would be needed which takes into account the different gas species, blocking factors on NEG surfaces, etc.

Ion-desorption simulations: her calculations are based on the following assumptions: VD surfaces are always unbaked, the NEGs in beam pipe are fully saturated (see above), i.e. they have zero pumping speed, whereas for the desorption yield two cases are considered: (a) NEGs have a desorption yield like for an in-situ baked metal surface (i.e. low yield, eta_CO = 0.5), and  (b) NEGs have a desorption yield like for unbaked Al surface (i.e. relatively large yield, eta_CO = 2.8). The most critical gas is CO. The values of eta are an average of available data for stainless steel. The calculations were done with the optimized geometry of the corrugated foil (july 2000): gap of 2 mm between top/bottom foils, and long flat surfaces where no Si sensors are present which gives a large lateral pumping duct.
Results: case (a) --> beam is stable, I_crit_CO = 6.5 A; case (b) -->   unstable with I_crit_CO = 1.3 A (should be >3.4A, which includes a safety factor 2). Note: the instability is now in the beam pipe section, not in the VD section.
It was remarked that the eta values used for the above calculations are rather on the high side of the results yielded by MPL's first ion-desorption measurements (see MPL's presentation below). Lower values of eta imply higher values for I_crit. This should be further discussed.

Electron cloud: AR recalled the simulation results of Giovanni Rumolo and Frank Zimmermann, see the previous meeting minutes. Some multipacting occurs in the VD section, and only when it is in the "out" position (6 cm gap seems to be an unlucky choice). AR presented a prelimary quantification of the local pressure increase due to electron multipacting (the only potential nuisance due to eletron multipacting in the VD). A pressure of a few times 1e-7 mbar could be expected. This does not seem to be a problem for LHC operation (VD is open, thus ion-induced desorption phenomena are subcritical). One side effect for LHCb: one could expect much increased background rates, however, with the VD open (i.e. not in physics data taking mode). Note: beam cleaning might be effective.

Get the scanned transparencies (Word document) of  AR's presentation here.

R&D progress on ion-desorption measurements: (presented by Maria Pilar Lozano)

First results of ion-desorption yield measurements were presented by MPL. These were obtained for unbaked Al and Cu surfaces, at 3 different impinging ion energies (3, 5 and 7 keV Ar ions), and as a function of  ion doses.
It was remarked that measurements on activated and saturated NEGs would be very valuable for drawing conclusions on stability issues in the LHCb section (see above, AR's presentation on the critical current). Performing such measurements with MPL's setup seems possible.

Get MPL's presentation here (Word document) of  AR's presentation.

Vertex detector vacuum test results: (presented by Sander Klous)

SK presented the first results obtained at NIKHEF on the gravity-controlled safety valve. The main goal was first to measure the spurious conductance of the tandem valves under normal operating conditions (molecular flow regime). The result is 1e-5 liter/s (1e-3 liter/s) for H2O when the valves are (not) differentially pumped. Therefore, one can expect a total leak rate (in normal mode) of  1e-9 mbar*liter/s H2O ( = 1e-4 mbar * 1e-5 liter/s) from the 2ary to the 1ary vacuum. This is negligible compared to the total outgassing of the unbaked 1ary vacuum (~1e-5 mbar*liter/s).
Further measurements will be performed to characterize the dynamic response of the valves in case of an abrupt leak and to test the pump down and venting procedures of the actual VD setup.

Get SK's presentation (PowerPoint document) here.

Silicon detector cooling studies: (presented by Themis Bowcock)

TB presented ANSYS calculational results for a thermal model of a Si module. To study the possibility of baking out the foil while cooling the Si detector, calculations were made with the simulated electronics turned off and by enclosing the whole module in a blackbody cavity at 150 C. The temperature in the Si sensor did not increase by more than 2 degrees (-5 to -7 C). Furthermore, a dummy module was constructed and will be used in the coming weeks to perform measurements of temperature distributions with a thermal camera (first in air, then in vacuum, with and without a hot box around).

Get TB's presentation (PowerPoint document) here.

Presentation of the Monte-Carlo optimization  of the VD geometry: (presented by Thomas Ruf)

TR presented briefly the physics reasons that limit the thickness of the Si detector encapsulation. Intensive Monte-Carlo simulations were performed this year to study (among others) the effect of the Al vacuum/RF shield on the LHCb physics performance. The increase of the foil thickness from 0.1 mm to 0.25 mm is expected to reduce the trigger efficiency by more than 20 %. Background rates (and occupancies) are also increased, which degrades track finding and reconstruction.

Get TR's presentation here (PowerPoint).

VD mechanical design update: (presented by Jo van den Brand)

JvdB presented a new design option elaborated at NIKHEF. In this design, the detectors can be (dis)mounted without exposing the primary vacuum to ambient air.
In particular, it allows to bake out the primary vacuum surfaces, then vent with a dry inert gas (ultrapure inert gas if the NEGs were activated), install the Si sensors into the secondary vacuum housing, and pump down again. It however involves the difficult construction of a large and rectangular membrane, which separates the two vacua and allows for displacing the Si housing. Investigations are ongoing to see where/how such a bellow could be fabricated. The solution is certainly costly but offers important advantages (bake out of primary vacuum, easiness of acces to Si detectors, preservation of NEGs when venting).
He also reported on the status of the beryllium option for the Si detector thousing. In addition to the very high price quoted by Brush Wellman, there are still strong doubts that such a corrugated box can be fabricated with 1 mm thick Be walls.

Get JvdB's presentation (PowerPoint document) here.

Progress of first risk analysis: (presented by Massimiliano Ferro-Luzzi)

MFL reported on a first risk analysis for the LHCb VD system. The framework used is identical to the one used for the CERN Safety Alarms and Monitoring Systems (CSAMS). Given a design for the VD, this risk analysis allows to define objectively a set of requirements/recommendations to be fulfilled in order to make the system acceptable for LHC. A draft document will be distrtibuted by the middle of december.

Get MFL's presentation (PowerPoint document) here.

Concluding discussions and remarks:

The LHC-VAC group expressed that their main worries concern the details of the vacuum system implementation and of the associated procedures used e.g. to vent and pump down. The differential pressure sustainable by the current Al housing (15 mbar for plastic deformation) is also a delicate item. A "critical" pressure of about 100 mbar would relax their worries.
The possibility of baking out the VD (which is guaranteed in the new design, and seems even possible in the previous design, see above) constitutes a major improvement. With this added feature it is likely that all dynamic vacuum and NEG saturation issues are solved (though this needs to be checked).

The LHCb participants insisted on the fact that a formal "green light" from the LHC-VAC group is essential in order for LHCb to continue the work (construction of prototypes, test setups, etc.), as this involves spending important amounts of resources (manpower, funding). An official agreement on the fundamental design concepts is required (use of a secondary vacuum with a foil that does not withstand atmospheric pressure, use of CO2 as a coolant). Obviously, such an agreement would not discard the need for further tests and developements (defined by LHC representants, e.g. via the risk analysis document) to make the design eventually comply with all LHC safety standards. For instance, the amount of risk involved with a foil of a given pressure resistance, or with vacuum manipulations on the system, should become clear from the risk analysis. This risk can be minimized by imposing well-defined safety requirements. In connection to this, JvdB suggested to have an  LHCb-VELO/LHC-VAC meeting on vacuum control systems in the first quarter of 2001.

It was agreed to have an extensive presentation of the LHCb vertex detector design in a LEMIC session in january 2001, with the goal to obtain a written approval (e.g. in the minutes of the LEMIC meeting) provided  the LHC-VAC group has no objections to fundamental design choices.

Next meetings: presentation to LEMIC in january 2001 and meeting on controls in 1st quarter of 2001 (dates to be defined).

New actions: (to be completed before the end of the year)
(a) Confirm LHCb beam stability assuming the VD is baked out (AR).
(b) Try to calculate more accurately the expected life time of the NEGs for an unbaked and a baked system (MFL).
(c) If possible, measure eta for NEGs activated and saturated in vacuum (MPL).
(d) Complete and distribute a draft of the first risk analysis (MFL, JRK).
(e) Define meeting on vacuum controls: date, place, agenda (JvdB).

Massimiliano Ferro-Luzzi, 7th of December 2000.