From: Michal KOZIEL Date: Tue, 3 Feb 2015 20:14:15 +0000 (+0100) Subject: update 03.03.2015 X-Git-Url: https://jspc29.x-matter.uni-frankfurt.de/git/?a=commitdiff_plain;h=5a4693c14a64ed9611e3fb2bde71e95d4ff390e2;p=reports.git update 03.03.2015 --- diff --git a/GSI_2015_MK_TT_Progress_mechanical_integration/integration.tex b/GSI_2015_MK_TT_Progress_mechanical_integration/integration.tex index d67b40b..bbad52f 100644 --- a/GSI_2015_MK_TT_Progress_mechanical_integration/integration.tex +++ b/GSI_2015_MK_TT_Progress_mechanical_integration/integration.tex @@ -11,7 +11,7 @@ \setlength{\titleblockheight}{35mm} \begin{document} -\title{The CBM-MVD: Progress on Mechanical Integration \thanks{This work has been supported by BMBF (05P12RFFC7), EU-FP7 HadronPhysics3, HGS-HIRe, GSI and HIC for FAIR.}} +\title{The CBM-MVD: Progress in Mechanical Integration \thanks{This work has been supported by BMBF (05P12RFFC7), EU-FP7 HadronPhysics3, HGS-HIRe, GSI and HIC for FAIR.}} \author[]{M. Koziel} \author[]{T. Tischler} @@ -30,35 +30,15 @@ This report summarizes the activities undertaken towards the construction of the %\section{Sensor integration} -\section{Quality assurance} -%\textbf{Quality assurance:$\;\;$} -The concept of quality assurance for the ultra-thin sensors to be integrated into the MVD has been discussed in \cite{QA}. Since the final sensors for the MVD do not exist yet, $18$ MIMOSA-$26$ AHR sensors thinned to 50 $\upmu$m were used to establish the required test procedures. The number of sensors which passed the tests has been compared with the know yields after thinning \cite{LG}. $12$ sensors were found without a significant number of dead/noisy pixels; they were qualified as fully operational. Four sensors exhibiting some dead rows/columns were marked as faulty. The two remaining sensors were not operational due to a power supply short (one sensor) and problems while powering one out of the four MIMOSA-$26$ sub-matrices. It was concluded that the observed yield is in agreement with the expectations. +\textbf{Quality assurance of 50~$\upmu$m thin PRESTO sensors:$\;\;$} Thinned MIMOSA-26 sensors will be used for assembly the so called PRESTO module. PRESTO addresses the double-sided integration of $15$ MIMOSA-$26$ sensors (dummies and working sensors) onto a $8\times 8~cm^{2}$ CVD diamond support (see \cite{PRESTO} for more details). Sensors will be connected with the R/O system by means of a newly designed ultra-low material budget flex cable employing commercially available processes based on copper traces \cite{FPC}. Constructing the PRESTO allows to estimate the integration yield providing that the employed sensors are tested prior to assembly. Up to now, $18$~MIMOSA-$26$ AHR sensors thinned to 50 $\upmu$m were probe tested using the setup described in \cite{QA}. The setup allows testing the standard operation modes of the sensor as well as measure the fixed pattern and temporal noise by the means of so called s-curves. $12$ sensors were found without a significant number of dead/noisy pixels; they were qualified as fully operational. Four sensors exhibiting some dead rows/columns were marked as faulty. The two remaining sensors were not operational due to a power supply short (one sensor) and problems while powering one out of the four MIMOSA-$26$ sub-matrices. The estimated yield was then of about 65$\%$ which is in agreement with expectations for this type of sensors \cite{LG}. The temporal noise was found to be of about 1.6-1.8~mV and the fixed patter noise of about 0.5-1.0~mV. This is by factor of 2-3 higher than the noise specified by a sensor provider. This was nevertheless as expected since the sensor power signals were generated outside the probe card. The addressed probe tests allowed also to establish test procedures required for non-destructive tests of thinned CMOS sensors and can be applied for testing the final MVD sensors. +\textbf{Development of a custom made glue:$\;\;$} +An "ideal" adhesive for the integration of the sensors onto their supports should be easy to dispense in a thin and uniform layer\textemdash calling for a low viscosity\textemdash, radiation hard as well as flexible (to compensate for the thermal expansion mismatches between the sensor and their support material) within the temperature range foreseen for the operation of the MVD sensors. Since there are none "on-shelf" products that meet these requirements, a custom-made, two compound adhesive with a working name RAL-$247$ was manufactured at the Rutherford Appleton Laboratory (RAL), Composites and Materials Testing Group, UK. The glue features a glass temperature of -$45~^{\circ}$C, a viscosity of below $100$ mPa$\cdot$s and a curing time of $48$ h at +$50~^{\circ}$C. To investigate its radiation hardness, RAL-$247$ samples were irradiated with X-rays to 100~Mrad and to a proton dose of about 10$^{15}$~n$_{eq}$/cm~$^{2}$. The irradiated samples were sent to RAL for further Dynamic Mechanical Analysis tests which unraveled no significant change of properties that confirms the expected radiation hardness at the range of radiation doses expected at the MVD. -\section{Development of an appropiate glue} -%\textbf{Development of an appropiate glue:$\;\;$} -An "ideal" adhesive for the integration of the sensors onto their supports should be easy to dispense in a thin and uniform layer\textemdash calling for a low viscosity\textemdash, radiation hard as well as elastic within the temperature range foreseen for the operation of the MVD sensors. Since there are none "on-shelf" products that meet these requirements, a custom-made, two compound adhesive with a working name RAL-$247$ was manufactured at the Rutherford Appleton Laboratory (RAL), Composites and Materials Testing Group, UK. The glue features a glass temperature of -$45~^{\circ}$C, a viscosity of below $100$ mPa$\cdot$s and a curing time of $48$ h at +$50~^{\circ}$C. To investigate its radiation hardness, RAL-$247$ samples were irradiated with X-rays to 100~Mrad and to a proton dose of about 10$^{15}$~n$_{eq}$/cm~$^{2}$. The irradiated samples were sent to RAL for further Dynamic Mechanical Analysis tests which unraveled a very similar performance of glue samples before/after irradiation \cite{glue}, see figure \ref{fig:picture}. -\section{Development of the heat sinks for the MVD} -%\textbf{Development of the heat sinks for the MVD:$\;\;$} -The operation of the MVD in vacuum requires a continuous cooling of the sensors to limit radiation induced defects as well as noise. To keep the material budget of the individual MVD station as low as possible, the cooling approach of the MVD employs highly thermal conductive sensor support materials (CVD diamond \cite{CVD} and encapsulated graphite) in the acceptance of the MVD and actively cooled aluminum-based heat sinks outside of this area. To evaluate the cooling concept and its vacuum compatibility, half-station heat sinks of the first three MVD stations were manufactured. The heat sinks incorporate a buried cooling pipe and have thermally been simulated prior their manufacturing using a worst case scenario for the sensor power dissipation plus an additional safety factor of four. These heat sinks are currently being evaluated under laboratory conditions focusing on their vacuum compatibility \cite{vacuum}. The heat dissipation of the MVD sensors is provided by electric heaters. +\textbf{Development of the heat sinks for the MVD:$\;\;$} +The operation of the MVD in vacuum requires a continuous cooling of the sensors to limit radiation induced defects as well as noise. To keep the material budget of the individual MVD station as low as possible, the cooling approach of the MVD employs highly thermal conductive sensor support materials (CVD diamond \cite{CVD} and encapsulated graphite) in the acceptance of the MVD and actively cooled aluminum-based heat sinks outside of this area. To evaluate the cooling concept and its vacuum compatibility, half-station heat sinks of the first three MVD stations were manufactured at COOLTEK GmbH. The heat sinks incorporate a buried cooling pipe and have thermally been simulated prior their manufacturing using a worst case scenario for the sensor power dissipation plus an additional safety factor of four. These heat sinks are currently being evaluated under laboratory conditions focusing on their vacuum compatibility \cite{vacuum}. The heat dissipation of the MVD sensors is provided by kapton insulated flexible heaters from OMEGA Engineering, INC. -\section{PRESTO - PREcursor of Station TwO} -%\textbf{PRESTO - PREcursor of Station TwO:$\;\;$} -There are numerous activities related to the mechanical integration of the MVD currently addressed by the MVD group including a design of ultra-low material budget flex cable employing commercially available processes based on copper traces \cite{FPC}. Those cables will be used within the PRESTO module which is currently constructed. PRESTO addresses the double-sided integration of $15$ MIMOSA-$26$ sensors (dummies and working sensors) onto a $8\times 8~cm^{2}$ CVD diamond support. The reader is referred to the report \cite{PRESTO} for more details. - -%The PRESTO module will employ the new flex cables \cite{FPC} providing all signals needed to operate and read out the sensors. To construct this PRESTO module, new sensor positioning jigs aiming for a sensor positioning precision w.r.t. to support and neighboring sensors of below 100~$\upmu$m were manufactured. To evaluate the integration concept, the RAL-$247$ adhesive and new jigs, a dummy PRESTO module based on $200~\upmu$m glass plate was assembled, employing $50~\upmu$m MIMOSA-$26$ dummies, see fig.~\ref{fig:picture}. The horizontal sensor-to-sensor distances were measured to be below $5~\upmu$m. The vertical shifts between sensors were measured to be of about 20~$\upmu$m. The achieved precision is significantly below the envisioned one. Next steps comprise the establishing of procedures for the integration of the flex cables, the exercise of double-sided bonding and the verification of the vacuum compatibility. - -%\begin{figure}[htb] -%\centering -%\includegraphics*[width=65mm]{glue.eps} -%\caption{RAL-$247$ adhesive modulus as a function of temperature for different irradiation types and doses.} -%\label{fig:picture} -%\end{figure} - -%For gluing the sensors we worked with RAL247 adhesive. We learned that 3-5~$\upmu$l of glue is required to dispense and uniform and thin (about 30~$\upmu$m) layer under the sensors. Evacuation of air bubbles introduced into the glue during mixing was done at an exicator (4$\cdot$10$^{-1}$~mbar) for about 1 hour. However, this did not prevent the air bubbles to appear after sensor gluing on a carrier. The introduced air bubbles featured a size of about 100-300~$\upmu$m diameter. To verify their impact on the 50 um sensor dummies, we put the cured module inside the vacuum for about 48 hours (4$\cdot$10$^{-1}$~mbar). Visual inspection did not show any mechanical damage. Further studies will be addressed with working sensors to check on-fly any possible correlation between sensor performance, pressure and bubble sizes. - -%\section{The MVD heat sinks development}