\section{Upgrade of the RICH-1 gaseous photon detectors} %\section{Hadron identification} \label{sec:upgrades.pid} %\subsection{RICH-1} The large majority of the topics of the \compass\ physics programme performed so far and presented in this proposal requires hadron identification. In the \compass\ spectrometer, it is provided by a large size Cherenkov imaging counter, RICH-1, operated in its initial version since 2001 \cite{Albrecht:05} and in its upgraded version characterised by a more powerful photon detection system since 2006 \cite{Abbon:2010,Abbon:2008zza}. In RICH-1 particles cross 3~m of gaseous radiator, $\mathrm{C_4F_{10}}$. %Image focusing is obtained thanks to a 21~m$^2$ wall %formed by a mosaic arrangement of 116 spherical UV mirror elements. It consists %of two (upper and lower) spherical surfaces (nominal radius of 6600~mm) with %different orientation. During the years 2001--2004, RICH-1 photodetection has been performed with Multi-Wire Proportional Chambers (MWPC) equipped with solid state CsI photocathodes \cite{RD26Collaboration:93,ALICEcollaboration:98,Piuz:2003vu}. One of the two cathode planes of the proportional chamber is a Printed Circuit Board (PCB) segmented into $8\times8~\mm^2$ pads coated with a CsI film. The Cherenkov photons enter the chamber via a fused silica window and hit the photocathode PCB. The photo-electrons produced by the converted photons are multiplied in the MWPC. The detectors are operated at low gain (below $5\times10^4$), as imposed by the presence of the CsI photocathode. %The first stage of the electronics read-out system in use till 2004, is %characterized by long integration time (0.6~$\mu$s) related to the reduced gain. %This results in an effective detector memory limiting the RICH-1 performance in %the \compass\ environment, where a high rate uncorrelated background is present %due to the large component of the muon beam halo. Also, the base-line %restoration time of the front-end in use till 2004 (about 3.5~$\mu$s) generates %data acquisition dead-time. To overcome these limitations and to face the %higher rates foreseen for the \compass\ data taking from 2006 on, the RICH-1 %detector has been upgraded. The RICH-1 upgrade is two-fold. %The peripheral regions (75\% of the surface) are populated by the images produced by lower %momentum hadrons, and experience a less severe level of the uncorrelated %background. Here the photon detectors are unchanged; they are read out by a new Since 2006 the peripheral region (75\% of the surface) with the images of lower-momentum particles is read out by a system \cite{Abbon:2006ba} based on the APV chip \cite{French:01a}. %with negligible dead-time and increased %time resolution, obtained by measuring three amplitude samples on the raising %edge of the signal. The Cherenkov images produced by the high-momentum particles are detected in the central photon detection area (25\% of the surface), a region highly populated by uncorrelated background images. %These aspects require very good resolution of the measured Cherenkov angle to %push towards momenta as high as possible the limit for hadron identification %and pushed time resolution to discriminate the uncorrelated background. This region is instrumented with a fast detection system based on MultiAnode PhotoMultiplier Tubes (MAPMT) \cite{Abbon:2010,Abbon:2008zza} coupled to individual telescopes of fused silica lenses (a prismatic field lens followed by a concentrator lens) to enlarge the effective active area of the photon detectors. %The system allows for %detection of about four-times more Cherenkov photons than in the peripheral detectors. The effective pad size, resulting from the MAPMT pixel size and the lens telescope magnification is about $12\times12~\mm^2$. %These detectors, intrinsically fast %and exhibiting sub-nanosecond time resolution, are coupled to a read-out system based %on high sensitivity amplifier/discriminators and fast TDCs, which fully %exploits the fast photon detection characteristics of the new detection system %\cite{Abbon:2008zza}. Presently the resolution on the measured Cherenkov angle for particles with $\beta\rightarrow 1$ is 0.3~mrad for the central photon-detection area and 0.9~mrad for the peripheral area. %\subsection{Upgrade of the RICH-1 gaseous photon detectors} The MWPCs in operation for RICH-1 are examples of the first generation of gaseous photon detector with a solid state photoconverter. In spite of the remarkable success of proving that solid state photoconverters can operate in gaseous atmospheres, MWPCs with CsI photocathodes suffer because of some performance limitations: ageing, causing a severe decrease of the quantum efficiency after a collected charge of the order of some mC$/\Cm^2$ \cite{Braem:2005es,Hoedlmoser:2007mr} and long recovery time (about 1 day) after a detector discharge \cite{Albrecht:05}. Therefore, they must be operated at low gain, reducing the single photo-electron detection efficiency. These limitations are related to the bombardment of the CsI photocathode film by the positive ions and photons generated in the multiplication process. In \compass\ RICH-1, these features limit the number of detected photons and, at high beam rate, cause instabilities of the photon detector performance related to the long recovery time after a discharge. To overcome these limitations, a large-gain gaseous photon detector is developed, characterised by a closed geometry architecture and based on the use of THick GEM (THGEM) \cite{Chechik:2004wq,Chechik:2005ud,Shalem:2006iw,Bidault:2006kx} electron multipliers, coupled to a solid-state CsI photocathode. In fact, in a multilayer structure of electron multipliers, a good fraction of the ions is trapped in the intermediate layers, and no photons can reach the photocathode \cite{Chechik:2004wq,Chechik:2005ud,Shalem:2006iw,Kozlov:2003zr,Fraenkel:2005wx,Milov:2007xh}. The THGEM electron multiplier is derived from the GEM \cite{Sauli:97} one. The copper-coated kapton foil of the GEM multipliers is replaced by a standard PCB and the holes are produced by drilling. The conical shape of the GEM holes that forms uncoated polyamide rings around the holes themselves are replaced by a clearance ring, the rim, surrounding the hole and obtained by copper etching (Fig.~\ref{fig:upgrades.pid.thgem}). Typical values of the geometrical parameters are PCB thicknesses of 0.4--1~mm, hole diameters ranging between 0.3~mm and 1~mm, hole pitches of 0.7--1.2~mm and rim width values between 0~mm and 0.1~mm. The electron multiplication is obtained applying an appropriate voltage between the two conductive faces of the PCB, which are electrically insulated with respect to each other. Large gains have been reported for detectors with single or double THGEM layers as well as good rate capabilities. \begin{figure}[tbp] \begin{minipage}{0.48\hsize} \begin{center} \includegraphics[width=\hsize]{upgrades_pid_thgem} \caption{Detail of a THGEM PCB} \label{fig:upgrades.pid.thgem} \end{center} \end{minipage} \hfill \begin{minipage}{0.48\hsize} \begin{center} \includegraphics[width=\hsize]{upgrades_pid_thgem_diagram} \caption{Scheme of the basic architecture of a THGEM-based photon detector (not to scale).} \label{fig:upgrades.pid.thgem_diagram} \end{center} \end{minipage} \end{figure} %The basic architecture of a THGEM-based photon detector, schematically %illustrated in Fig.~\ref{fig:upgrades.pid.thgem_diagram}, %consists in a configuration with triple THGEM layers %forming a multiplication structure, where the first layer is coated with a CsI %film and acts as a reflective photocathode. The electrical field above and The basic architecture of a THGEM-based photon detector (Fig.~\ref{fig:upgrades.pid.thgem_diagram}) comprises a multiplication structure of three THGEM, where the first is coated with a CsI film and acts as a reflective photocathode. The electrical field above and below the multiple THGEM structure is defined by additional electrodes. A plane of metallic wires parallel to the THGEM PCB %is used in front of the entrance %face of the structure, namely the face from which the charge to be multiplied %enters the structure itself. Wires are used to keep side from which the particles enter as transparent as possible. A PCB segmented into pads, again parallel to the THGEM PCBs, is used both to define the electric field at the exit side of the structure and to collect the signals. Its segmentation makes it possible to preserve the space information. The particles enter through an UV-transparent window, closing the volume of the detector chamber. %, is used to let the UV photon reach the photocathode. %This configuration, where a reflective photocathode is present, is preferred to %architectures with a semi-transparent photocathode, where the photo-converter %layer is applied onto the internal face of the window: the reflective %arrangement results in a larger photo-electron collection. In fact, a %semi-transparent photocathode requires the application of a thin metallic film, %which absorbs photons, to keep the entrance window at a fix potential; also, %the probability of photo-electron absorption is lower in a reflective %photocathode than in a semitransparent one as the conversion probability is the %highest at the entrance surface of the photo-converter. Moreover, the thickness %of the photo-converter layer is non-critical in the reflective configuration, %contrary to the semitransparent one. We have performed a systematic study \cite{Alexeev:2009zz,Alexeev:2009a,Alexeev:2010a, Rocco:10a} of the properties of the THGEM electron multipliers, in particular disentangling the role of the various geometrical, electrical and production-related parameters to determine the optimal geometry for photon detection applications. In particular, we have shown that stable gain can be obtained for THGEM with no or small (10~$\mu$m) rim, while a large rim ($\sim 0.1$~mm) causes severe gain variations versus time (Fig.~\ref{fig:upgrades.pid.thgem_gain}). \begin{figure}[tbp] \begin{minipage}[t]{0.48\hsize} \begin{center} \includegraphics[width=\hsize]{upgrades_pid_thgem_gain} \end{center} \caption{Gain versus time behaviour for two THGEMs with the following geometry: thickness 0.4~mm, pitch 0.8~mm and hole diameter 0.4~mm (common parameters); different parameter: 0.1~mm rim for (a), no rim for (b). Continuous detector irradiation; $\Delta V=1750$~V for (a) and 1330~V for (b).} \label{fig:upgrades.pid.thgem_gain} \end{minipage} \hfill \begin{minipage}[t]{0.48\hsize} \begin{center} \includegraphics[width=\hsize]{upgrades_pid_thgem_spec} \end{center} \caption{Single photo-electron amplitude spectrum measured with the THGEM detector described in the text.} \label{fig:upgrades.pid.thgem_spec} \end{minipage} \end{figure} For single photon detection, both simulation studies and laboratory measurements indicate that THGEM geometries for which the electric field at the CsI photocathode surface is large enough to guarantee a good photo-electron extraction are those for which the ratio between the hole diameter and the hole pitch is large ($\ge~0.5$). At the same time the larger this ratio is the larger is the detector dead zone. Selecting a diameter of 0.4~mm and a pitch of 0.8~mm a good compromise can be obtained. Gains up to 106 have been obtained with a triple layer detector employing THGEM PCBs with these parameters (Fig.~\ref{fig:upgrades.pid.thgem_spec}), to be compared with the gain of the present gaseous photon detectors of $5\times10^4$. The measured ion feedback rate is reduced to about 25\%. Further reduction can be obtained introducing one more wire plane electrode between the first and the second THGEM layer. This improved geometry is presently under test. The measured time resolution is about 10~ns. The intrinsic space resolution is of the order of 1~mm. Small-size prototypes of THGEM-based photon detectors have been employed so far. The engineering aspects related to the extension of the size to detectors of about $60\times60~\Cm^2$ is ongoing. Good quality THGEM PCB of this size have already been produced. % For these studies the digital read-out system presently used to read-out the MAPMTs employed in the centre of the RICH-1 detection area \cite{Abbon:2008zza} has been successfully used to detect with good efficiency the photoelectron signals, thanks to the large detector gain. % % %Thanks to the large detector gain, the digital %read-out system presently used to read-out of the MAPMTs employed in the centre %of the RICH-1 detection area can be successfully used to detect with good %efficiency the photo-electron signals. The proposed upgrade of the RICH-1 gaseous photon detectors consists in replacing the MWPCs with CsI photocathode with THGEM-based detectors of large size, coupled to CsI photocathodes. The pad size will remain the same of the MWPCs, namely, $8\times8~\mm^2$. The major improvements of the new gaseous photon detectors are the larger gain, which results in a larger efficiency of the single photo-electron detection, the reduced ion feed-back allowing for more stable detector operation and better time resolution.