Characterization of a hybrid GEM-Micromegas detector with respect to its application in a continuously read out TPC

In the context of the upgrade of the LHC during the second long shutdown the interaction rate of the ALICE experiment will be increased up to 50 kHz for Pb-Pb collisions. As a consequence, a continuous read-out of the Time Projection Chamber (TPC) will be required. To keep the space-charge distortions at a manageable size, the ion backflow of the charge amplification system has to be significantly reduced. At the same time an excellent detector performance and stability of the system has to be maintained. A solution with four Gaseous Electron Multipliers (GEMs) has been adopted as baseline solution for the upgraded chambers. As an alternative approach a hybrid GEM-Micromegas detector consisting of one Micromegas (MM) and two GEMs has been investigated. The recent results of the study of the hybrid GEM-Micromegas detector will be presented and compared to measurements with four GEM foils.


Introduction
A Time Projection Chamber (TPC) [1] is used in many ongoing experiments like STAR [2] or ALICE [3] as central detector.With its low material budget and excellent pattern recognition capabilities it is an ideal device for three-dimensional tracking and identification of charged particles.The working principle of a TPC is based on ionization and gas amplification: Once a charged particle traverses the gas-filled detector the gas molecules will be ionized along the particle's trajectory.Owing to an electric field within the detector volume, the produced electrons will drift to the endplates where they get amplified and finally detected at a space-resolved readout.With a gas volume of 90 m 3 the ALICE TPC is the largest TPC on the world.Up to now the readout is done by 72 Multi-Wire Proportional Chambers (MWPC) [4] with a total of 550000 readout cathode pads.In order to prevent ions to drift back to the active drift volume, a bipolar Gating Grid (GG) is used which limits the theoretical readout rate to the order of 3 kHz.Together with the present readout system, the rate capability for central Pb-Pb collisions is limited to 300 Hz.As the interaction rate of the Large Hadron Collider (LHC) will be increased up to 50 kHz for RUN3, the expected amount of events can not be handled by the current readout technology which demands a complete redesign of the TPC.In Ne-CO 2 -N 2 (90-10-5) and at a total gain of 2000 the energy resolution must not exceed σ/E = 12 % for 55 Fe which translates to a dE/dx resolution of about 5.5 % and 7 % in pp and central Pb-Pb collisions.The ion backflow should be lower than 1 % in order to be able to completely correct all effects from space-charge distortions [3].Therefore the gated and MWPC-based readout technology will be replaced by a continuous read-out technology [5] based on Micro-Pattern Gaseous Detectors (MPGD).In previous investigations it has been shown that at least four GEM foils [6] with different pitches (see table 1) are required in order to cope with the aspired goal [7].An alternative approach is given by the hybrid GEM-Micromegas detector which consists of only two standard GEM foils and a bulk Micromegas [8].other.The investigated bulk Micromegas was manufactured at CERN and has a gap of 128 µm which separates a small mesh (160 lines / cm) from the readout padplane (10 cm x 10 cm).Finally a summation card has been designed to sum up all 128 pads as four concentric rings.All rings can be read out individually or in an arbitrary combination.The setup has been assembled within a gas tight vessel with a volume of 6 liters.A Kapton window on the lid allows an external irradiation of the drift region with an X-ray or an 55 Fe source.The currents at the cathode, the different GEM stages or the anode can be measured by picoammeters which were developed at the TU Munich [9] and which are sensitive for currents down to 20 pA before systematic errors become dominant.In order to reduce electromagnetic influences and to shield the system, the whole detector and the picoammeters are placed in a Faraday cage made of copper and lead.

Gain contributions
Depending on the field configuration above and below a GEM, electrons can get lost by drifting to the copper or the Kapton layer, i.e. the measured effective gain (equation 1) is a function of the field-dependent extraction-and collection efficiency coll and extr [10] of the GEM.
Finally the total gain of the detector can be expressed as the product of the effective gain of its individual amplification stages (equation 2).
In order to understand the contribution of the different amplification stages, the gain of each component has been measured.Figure 2 shows the absolute gain of the Micromegas as a function of the applied voltage: Since the 55 Fe source has a certain solid angle, the full padplane (all pads) needs to be read out in order to cover the avalanches from all converted photons.The gain rises exponentially and is in the order of several 10 2 -10 3 for common Micromegas voltages.Since the primary charge above the first amplification stage is below the sensitive range of the picoammeters, the gain contribution of the first GEM can only be estimated by the second GEM (see figure 3): A field configuration of TF1 0.4 kV/cm and TF2 3.0 kV is commonly used for the first amplification stage and thus represents the effective gain of the first GEM.Accordingly field configurations like TF1 3.0 kV/cm and TF2 0.4 kV/cm are common settings for the second GEM.Most of the total gain of the hybrid GEM-Micromegas detector is contributed by the Micromegas (order 10 2 -10 3 ), followed by the first GEM (order 10 1 -10 2 ).Commonly the second GEM is operated at an effective gain around 1 to maintain the energy resolution and to act as a further stage to block the ions from the Micromegas.

Energy resolution and ion backflow
For the planned upgrade of the ALICE TPC the ion backflow has to be less than 1 % while the energy resolution (σ/E) should not exceed 12 % for 55 Fe (at a total gain of 2000 in Ne-CO 2 -N 2 (90-10-5)) [3].The energy resolution and the ion backflow of the hybrid GEM-Micromegas detector have been determined within a GEM1-Micromegas scan, i.e. the applied voltage of the first GEM was successively increased or decreased while the total gain of the detector was kept at a constant level of 2000 by adjusting only the voltage of the Micromegas.While the electrons from avalanches drift to the readout padplane, some of the produced ions may drift back to the cathode.Thus the ion backflow can be defined as the fraction of the currents at the readout padplane (anode) and the cathode according to equation 3, where is the number of back drifting ions per incoming electron coming from the amplification region.For the ion backflow measurements an Amptek MiniX has been used to irradiate the detector (30 kV, 5 µA) and the whole padplane was read out.
All energy resolutions were determined with an 55 Fe source and the innermost pads (2x2) were read out.An example MCA spectrum is shown in figure 4: The prominent peak at the center belongs to K α/β photons which converted in the drift region while the lower energetic peaks appear due to conversion processes between the amplification stages (GEM1-GEM2 and GEM2-Micromegas).
To determine the energy resolution the K α/β peak was fitted by a simple Gaussian.Due to the non-negligible background under the peak a more refined fit model will be used in future investigations.Within the GEM1-Micromegas scan the energy resolution of the detector can be tuned by changing the amplification at the first GEM stage as the number of avalanche electrons will change and influence the energy resolution statistics.Figure 5 shows that the energy resolution is mostly dominated by the first amplification stage (GEM1) and only weakly depends on TF2 or the Micromegas.Only for low TF2 (80 V/cm) the energy resolution gets worse: Due to a small extraction efficiency of the second GEM electrons get lost within the transfer process to the Micromegas and hence its gain fluctuations become more dominant.Furthermore a change of the amplification at the first GEM strongly influences the ion backflow since the number of ions (which are immediately injected into the drift region) will be altered (figure 6).In contrast to the energy resolution the ion backflow is also influenced by TF2 and the Micromegas as most of the ions are produced within the last amplification stage (highest gain contribution).Thus the energy resolution and the ion backflow are not independent of each other: An optimization of the energy resolution leads to a degradation of the ion backflow and vice versa.Figure 7 illustrates this behavior for different TF2 and additionally indicates that specific field configurations exist which fulfill the requirements of the upgrade of the ALICE TPC.A comparison to the 4-GEM baseline configuration (S-LP-LP-S) reveals that the hybrid GEM-Micromegas detector can be considered as an alternative approach.

Summary and Outlook
A hybrid GEM-Micromegas detector consisting of two GEM foils (standard pitch) and a bulk Micromegas has been assembled at Bonn University.First measurements reveal that an energy resolution less than 12 % (σ/E with 55 Fe) and an ion backflow better than 1 % is possible in Ne-CO 2 (90-10) at a total gain of 2000.Within the detector most of the gain is contributed by the Micromegas  (order 10 2 -10 3 ), followed by the first GEM stage (order 10 1 -10 2 ).The second GEM stage is used in order to prevent ions from drifting back to the drift region while the energy resolution can be maintained.With respect to the ion backflow and the energy resolution the capability of the hybrid GEM-Micromegas detector can be compared to the 4-GEM baseline configuration and shows a similar performance.As Ne-CO 2 -N 2 (90-10-5) will be used for the ALICE TPC in RUN3 all previous measurements will be repeated in this gas mixture.Yet the full possible parameter space of the hybrid GEM-Micromegas detector has not been scanned in order to find an optimized configuration with respect to the ion backflow and the energy resolution.To optimize the energy resolution measurements, a veto setup for the innermost readout pads has been established and will be used in further investigations.Several measurements have shown that the quadruple S-LP-LP-S GEM configuration fulfills the specified requirements and thus has been chosen for the upgrade of the ALICE TPC.At the time of the decision the hybrid GEM-Micromegas detector was not studied as extensive as the multiple GEM setup and further stability measurements are still necessary to get a complete understanding of the hybrid GEM-Micromegas detector.

Table 1 .
Different GEM geometries (p pitch, d i inner diameter, d o outer diameter).