## Purpose

**Tracking:**
The TPC is discussed as a solution for the outer tracking within the target spectrometer. The required momentum resolution is ~1%, the required vertex resolution ~150um in the xy plane and < 1cm in z direction.

**PID:**
In the momentum range below ~1GeV and above ~2GeV the TPC provides information for particle identification within the target spectrometer. Especially for particles with momenta below ~1GeV this is of great help for the overall PID performance and to supplement the information from the barrel DIRC.

## Working principle (in short)

**General:**
3D tracking device - charged particles ionize detector gas - electric field along cylinder axis separates positive gas ions from electrons - primary electrons drift towards readout anode - gas amplification done by several GEM foils - ungated, continuous operation mode due to HESR beam properties - intrinsic ion feedback suppression by GEM foils - continuous data readout within PANDA DAQ - parallel online data reduction and processing (including tracking)

**PID:**
performed via measurement of mean energy loss per track length (dE/dx), described by Bethe-Bloch-formula, in combination with (obligatory) momentum measurement - PANDA TPC offers to do ~50-100 (fluctuating) energy loss measurements per track - truncated mean algorithm used to get rid off Landau tail and to calculate mean

## Important values

**Geometry:**
inner radius: 15cm, outer radius: 42cm, length: 150cm, gas volume: 700l, 2 separate chambers (due to beam pipe)

**Material budget:**
~1.5% X/X0

**Detector gas:**
Ne/CO2 (90/10, maybe admixture of CH4), gas gain: several 1000

**Operation:**
drift field: 400V/cm, 2x2mm pads (100000)

## First estimates and simulations (obtained from old PANDA framework and preliminary)

Data were generated based on an event generator which shoots p, K, pi, mu and e (plus antiparticles) isotropically through the TPC. All tracks come from the IP, with momenta between 0.2 and 4GeV/c. Tracks are divided into 6mm pieces, for each the energy loss is calculated resulting in 50-100 measurements depending on track length. Upper 40% are discarded and mean dE/dx calculated (truncated mean). The spread of the these dE/dx values for certain p bins is fitted with a Gaussian and the dE/dx resolution is defined as the corresponding sigma.

The separation power between two particles is defined as: 2*|I1 - I2| / (sigmaI1/I1 + sigmaI2/I2), where I stands for the dE/dx of the respective particle. A constant dE/dx resolution was assumed.

Note: For all the simulation results shown here the gas density value was a factor of 1.5 to high. Therefore we expect the performance to be a bit worse. For example the dE/dx resolution will change from ~5% to ~7%. Simulations will be repeated with the new PANDA framework as soon as possible.

--

QuirinWeitzel - 10 May 2006