An introduction by  F.Eisele

HERA-B [1] is an experiment designed primarily to search for CP violation in decays of B-mesons. The B-mesons are produced in interactions of 820 GeV protons in HERA proton beam with an internal wire target in the beam halo. Due to the small B-meson cross section, the experiment has to run at high interaction rates - between 30 and 50 MHz - to accumulate sufficient statistics to detect the rare CP-violating B decays within reasonable time. Among the strong background of "ordinary" proton-nucleus interactions, B-decay candidates have to be selected already at the first triggering level. This requires fast track detectors with sufficiently high granularity and high rate capabilities. For radial distances from the beam axis larger than about 25 cm the expected particle densities are small enough that wire drift tubes (honeycomb chambers) of 5 and 10 mm diameter can be used with an acceptable level of occupancy (outer tracker). For the inner region, covering the angular range larger than 10 mrad up to radial distances of about 30 cm and comprising about half of all tracks of a typical HERA-B event, particle rates of up to 20000  mm**-2s**-1 are expected . This corresponds to a  radiation load of up to 1 Mrad per year.
The high particle rate requires  fine granularity of the detector. It is determined by the maximum occupancy of the detector which is compatible with efficient track reconstruction. The area of one detector channel should not exceed 75 mm**2. The solution chosen for HERAB  is a microstrip detector where one readout element corresponds to a  narrow strip of up to 25 cm length and a pitch (distance from one strip to the next) of .3 mm. Such a strip detector satisfies at the same time the resoltion requirement of <.1 mm.

Typical event in the HERAB detector at a target  interaction rate of 40 MHz. In average about 120 charged tracks will be visible in the
 detector every 100 ns. The innner tracker covers the area near the beam pipe where the particle density is highest.

It should be noted that  high energy physics experiment did so far not have such severe requirements for tracking and there had been no field proven technology .
The solution to the problem chosen for the experiment HERAB is the 'micro strip gaseous chamber' ( MSGC) which was invented by Oed 10 years ago. The  principle of the MSGC is shown below: The readout structure consists of a microelectrode structure of anodes and cathodes  which are lithographically put to a substrate. The anodes in our case have a width of 10 microns  , the cathodes a width of 170 microns separated by a 60 micron gap. The pitch of anodes is therefore 300 microns. The electrodes are made of gold 500 nm thick. Above the electrodes is the chamber gas volume, 3 mm high which is closed by the drift cover.  The MSGC works like a multiwire proportional chamber with gas amplification. The anodes are put to ground, a drift field is applied between cover and anodes of typically 4 kV/cm. The gas gain at the anodes is mostly dependent on the cathode voltage. For a typical voltage of 650 V the gas gain is about 4000 and it depends exponentially  on the cathode voltage. Counting gases used are Ar:DME 50:50 or Ar:CO2 70:30.  A minimum jonising particle frees about  30 electrons in the 3 mm drift gap. These drift to the anodes and are amplified there . In average the signal of one particle is seen on 1.6 anodes (strip multiplicity).


3D view of an MSGC/GEM detector                                                                        cross section through an MSGC/GEM detector

 It had been shown already by Oed, that such a detector can tolerate very high particle rates (up to 1 million particles per square mm and second.). What was unclear however if these detectors can survive several years of operation under high particle fluxes. Gaseous detectors have always the problem of 'gas ageing' e.g. they have a tendency to deposit polymer layers on the anodes which are formed in the plasma from organic substances e.g. impurities in the gas. The detectors also work very near to the discharge limit and are therefore vulnerable to  gas discharges which could destroy the thin electrodes.
          Extensive tests of MSGC detectors have been done in the institute over the last 5 years. The results are documented in detail in a series of diploma theses and theses (  publications on MSGC's  ). We have done extensive ageing tests with X-ray irradiation and we have performed several beam tests in intense hadron beams both at PSI with pion and proton beams and within HERAB at HERA.  Here is a summary of the main results:
         1) the substrates need a well defined electronic conductivity. We use glass substrates which are coated by a  'diamond' layer with a conductivity of about 10**14  Ohms per
square (coated wafers).
          2) MSGC's with coated wafers and carefully selected gases and materials are able to survive several years of intense X-ray radiation. If they are exposed to intense hadron beams however they show frequent gas discharges at gains above about 1500. These gas discharges are induced by heavily ionising particles -- most likely spallation products - which deposit a very large charge in the neighbourhood of the anodes. Gas discharges damage the electrodes and can therefore not be tolerated. Our studies showed that the spark rate depends to very good approximation  only on the cathode voltage. A reduction of cathode voltage below the onset of discharges results in too low gas gain however.
         3) The solution to this problem is to do the gas amplification in two stages by the introduction of a second gas amplification device , the ' gas electron multiplier' (GEM). This is a capton foil, 70 micron thick covered with Cu on both sides. A large number of small holes is etched into this foil -- e.g.  holes with 60 micron diameter and a distance of 140 micron between the holes. A photograhy of a GEM is shown below.


Cut through a GEM foil under microscope       GEM foil seen from top under microscope           gold electrode structure. Left: anode bondpads.
Visible are 2 holes 60 micron wide.                  Distance between holes is 140 microns.                Anode width 10 microns, cathodes 140 microns.

If a voltage of about 400 V is applied across the GEM then all electrons passing through the holes are multiplied by a factor of about 25. A GEM foil combined with a MSGC allows then to limit the gas gain of the MSGC to about 200 which is reached by cathode voltages below the
set on of gas discharges.

      Fully assembled MSGC/GEM with readout electronics

MSGC/GEM detectors are presently constructed and installed in the HERAB experiment. The  Inner Tracker will consist of  46 detector layers, each layer being composed of four rectangular MSGC's of about 30 x 30 cm2 outer dimensions. With an electrode pitch of 300 micron a total number of 140 000 channels has to be read out.
This readout is performed by a highly integrated readout chip specially developed by the Heidelberg ASIC laboratory in collaboration with a group from the MPI. This chip
Helix 128  makes an analog readout of 128 strips with a sampling rate of 10 MHz and stores the information in a pipeline for 12 microseconds.
            The detector should be ready for operation by the end of 1999.

[1] Th. Lohse et al., An Experiment to study CP Violation in the B System Using an Internal Target at the HERA Proton Ring, Proposal, DESY-PRC 94/02 (1994)