\section{Pixelised Micromegas detectors}
%\subsection{Pixelised Micromegas detector with resistive boards}
\label{sec:upgrades.trackers}
The physics programmes of this proposal deal with low cross section processes,
which require high intensity of the incident muons or hadrons beams. 
As little as possible material should be introduced by the detectors which
should also stand high hadron flux. This poses new constraints on the tracking
system. Since 2007/2008 five GEM stations have been upgraded with new detectors
offering a pixelised readout structure in the centre, using the APV-based readout
system. They have been successfully operated in the hadron beam runs 2008/2009.
However, due to their overall size of $10~\Cm\times10~\Cm$, they do not fulfil the 
tracking requirements right downstream of the target region, presently equipped
with the large size $40~\Cm\times40~\Cm$ micro-pattern gaseous Micromegas detectors.
     The Micromegas detectors are read by strips which cover the full length 
of the active area and are prolonged in order to move the material of the 
front-end electronics (SFE16 amplifier cards and F1 TDC cards) out of the 
spectrometer acceptance. To avoid a too high occupancy in the vicinity of 
the beam and to minimise the material in the beam region, the detectors were
designed with a 5~cm diameter central dead zone. 
     In order to cope with the new requirements, we plan to replace the 
Micromegas built ten years ago by new ones of similar size and designed to:
\begin{itemize}
\item
stand higher beam intensities and up to a factor of five higher rates 
of hadrons, which are highly ionising particles,
and
\item
detect particles in the centre of the detector
without any extra cost in terms of material budget. 
\end{itemize}
The new detectors will also satisfy additional requirements: 
they will have improved robustness, and will be read out with light 
and integrated electronics. 
The central area  will be read by pixels. Since the present detectors
measure a single coordinate per plane, a rectangular shape of the 
pixels was chosen thus preserving the spatial resolution 
(Fig.~\ref{fig:trackers_MM_board}).

Tracking of particles from the beam or scattered at very low angles 
will become possible with high resolution. The non-central region of the detector 
will be read out via strips and the large size of the detector will be preserved. 
Special care is taken to reduce the frequency of the discharges, which limit 
the performance of the detector for high intensity hadron beams. A specific R\&D 
program was initiated exploring two lines:
\begin{itemize}
\item
use resistive layers in order to spread out the electrical charge,
\item 
preamplify the signal using a GEM foil thus 
allowing for a reduction of the gain in the amplification gap.
\end{itemize}
Two prototypes with 1~mm$^2$ pixels in the centre and 30~cm long strips in
the periphery were operated during several months in the hottest region of the \compass\ 
spectrometer with both muon and hadron beams. The first prototype was built with the new 
robust ``bulk'' technology, where a 20~$\mu$m woven stainless steel mesh is assembled with 
the board. This prototype is the first Micromegas detector, in which a large size thin 
bulk structure was glued to a honeycomb board. A second prototype was built with the standard 
techniques using a thin 5~$\mu$m copper mesh for comparison. Pixels and strips were read 
out by a compact and integrated electronics using APV chips (128 channels per chip) 
installed on new cards with a specific protection circuit. The gas mixture was the 
one used for all standard \compass\ Micromegas, 80\%/10\%/10\% of neon/ethane/CF$_4$.
 
The prototypes operated well, validating both the use of the APV front-end readout and
the new technology bulk structure on a thin board. However, the issue of reducing the
discharge rate in presence of hadrons was not addressed here.

Several small ($10~\Cm\times10~\Cm$) prototypes featuring different techniques of 
resistive coating were tested in a dedicated set-up. 
Standard bulk Micromegas and a Micromegas equipped with a GEM foil were tested 
in the same conditions for comparison. The discharge rates of all the prototypes were monitored for 
various intensities of hadron and muon beams in the SPS H4 test beam line. 
The preliminary results obtained with the Micromegas equipped with a GEM foil show that 
the discharge rate can be reduced by more than an order of magnitude 
(Fig.~\ref{fig:trackers_MM_disch}). For the resistive Micromegas prototypes, the discharge amplitudes were
so much reduced that the discharge rates could not be measured as a signal on the mesh.
Additional beam tests are ongoing to understand the impact of this technology on the 
reduction of the discharge rate. All other performances (efficiency, spatial 
resolution, cluster size) of the prototypes 
were similar to the standard Micromegas ones. A final decision on the technology 
to be implemented will be taken in the beginning of 2011.
Before the production of the twelve new detectors a final prototype 
will be built and validated in the beam.

\begin{figure}[tbp]
\begin{center}	
\includegraphics[width=110mm,clip=true]{trackers_PixelMM_fig_plancher}
\end{center}
\caption{Board design of the pixelised Micromegas.}
\label{fig:trackers_MM_board}
\end{figure} 


\begin{figure}[tbp]
\begin{center}	
\includegraphics[width=110mm,clip=true]{trackers_PixelMM_fig_spark}
\end{center}
\caption{Discharge rates per incident hadron for standard Micromegas and Micromegas with an added GEM foil.}
\label{fig:trackers_MM_disch}
\end{figure}