1: \documentclass[a4paper,11pt]{article}
2: %\usepackage{babel}
3: \usepackage[T1]{fontenc}
4: \usepackage{a4wide}
5: %\usepackage{babel}
6: \usepackage{graphics}
7: \usepackage{epsfig}
8: \usepackage{wrapfig}
9: \usepackage{amssymb}
10: \usepackage{lscape}
11:
12: \makeatletter
13:
14: %\providecommand{\LyX}{L\kern-.1667em\lower.25em\hbox{Y}\kern-.125emX\@}
15: \newcommand{\noun}[1]{\textsc{#1}}
16:
17: \makeatother
18: \linespread{1.1}
19:
20: \begin{document}
21:
22: \title{Simulation of a fine grained GEM used in the PixiE Experiment
23: \footnote{http://glastserver.pi.infn.it/pixie/pixie.html} \\
24: {\it PixiE Internal Report}}
25: \author{M. Del Prete \\
26: {\it\small INFN and University, Pisa} }
27: \date{15 May 2005}
28: \maketitle
29:
30: \section*{Introduction}
31: We have simulated the performances of a GEM with a large density of
32: multiplication holes.
33: The elementary cell is an equilateral triangle whose side is 90$\mu m$. We
34: shall assume that this pattern extends in the (x,y) plane.
35: At each vertex of the equilateral triangle there is a GEM hole
36: with an external radius of 30$\mu
37: m$ and an internal radius of 20$\mu m$.
38: The reference frame used in this study has the origin of axis in the center of
39: the GEM hole with the z axis pointing to
40: the drift plane.
41: The geometry of the GEM is shown in figure \ref{fig:fori}.
42: \begin{figure}[ht]
43: \begin{center}
44: \mbox{\epsfig{file=figure/foro90.eps,width=0.5\linewidth,angle=0}}
45: \caption{\small A section of the hole of GEM. The red circles indicate the
46: starting points of electrons.}
47: \label{fig:fori}
48: \end{center}
49: \end{figure}
50:
51: In this short note we will describe the simulation of the GEM and, in
52: particular we will study the gain and diffusion of the charge for different
53: gas mixtures.
54: This study has been performed to finalize the design of the PixiE Imager
55: Detector.
56:
57: We have started the simulation by generating single electrons in different
58: positions in the (x,y) plane at fixed z-coordinate. This is
59: the most elementary element through which we can simulate tracks and the
60: imaging performance of the detector.
61: The process has been followed through multiplication in the large fields of the
62: GEM and diffusion of produced electrons reaching the readout
63: plane.
64: Due to the cylindrical symmetry of the GEM the first quadrant ($x>0,y>0$, see
65: figure \ref{fig:fori}) has been selected to produce the
66: coordinate ($r$,$\phi$) of the starting electrons at the quota of $40\;\mu
67: m$ (approximately $15\;\mu m$ over the top GEM):
68: \[ r=10\cdot{n}\;\mu m\quad (n\;=1,\;2,\;3)\qquad
69: \phi=\frac{\pi}{2}\cdot{\frac{(k-1)}{3}}\quad (k\;=\;1,\;2,\;3,\;4) \]
70: Where $\phi$ is the azimuthal angle.
71:
72: At each point we have generated 25 events.
73: The study has been performed for the following gas mixtures:
74: \begin{itemize}
75: \item 100\%$CO_2$ atm 0.5 ed 1 atm.
76: \item 20\%Ar/80\%DME, 50\%Ar/50\%DME, 80\%Ar/20\%DME ad 1 atm.
77: %\item 20\%Ne/80\%DME, 50\%Ne/50\%DME, 80\%Ne/20\%DME ad 1 atm.
78: \end{itemize}
79:
80: \section*{Gain Study}
81: We have defined as absolute gain the number of electrons which reach the
82: quota $z\;=\;-40 \mu m$ (below the plane the GEM, approximately $15 \mu m$
83: below the bottom GEM plane).
84: At this quota most of the multiplication processes at the GEM hole are done.
85:
86: However not all these electrons drift to the readout plane,
87: some recombine and many
88: stick to the lower GEM plane (re-attachment).
89: For this reason we have defined also an effective gain as the number
90: of electrons
91: which arrive at the quota $z\;=\;-200 \mu m$. The electrons reaching this
92: quota are considered to be collected by the read out plane.
93:
94: It's customary to describe the gain with a Polya:
95: \[
96: P_{n} =\frac{1}{b \cdot \overline{n}} \cdot
97: \frac{1}{\Gamma\left(\frac{1}{b}\right)}\cdot
98: \left(\frac{n}{b \cdot
99: \overline{n}}\right)^{\left(1/b \right) - 1}\cdot e^{-n/b \cdot \overline{n}}
100: \]
101:
102: Where $b$ is an adjustable parameter and $\overline{n}$ is the average gain.
103: We have used this formula and fit the data with the function:
104: \[
105: F(n;p_{0},p_{1},p_{2}) = p_{2} \cdot n^{p_{0}-1} \cdot e^{-n/p_{1}}
106: \]
107: where $p_{2}$ is a normalization factor and the gain is the Polya mean, $G\;=\;p_{0}\;\cdot\;p_{1}$.
108:
109: The gain distribution is different for different mixtures of gases. In
110: particular the absolute gain distribution is often wide with long tails
111: and a description with a single Polya is not always satisfactory. Hence, we
112: have described the distribution with the sum of two Polya, of which
113: the first one fits most of events and the second one accounts
114: for the long tails. An example is shown in figure \ref{fig:garco2}.
115: We have taken as the
116: average gain of the GEM the mean of the first Polya.
117:
118: \begin{figure}[ht]
119: \begin{center}
120: \mbox{\epsfig{file=figure/ga_90_480_4k_ARDME20-80_1atm.epsi,
121: width=0.3\linewidth,angle=270},
122: \epsfig{file=figure/ge_90_480_4k_ARDME20-80_1atm.epsi,
123: width=0.3\linewidth,angle=270}}
124: \caption{\small Distribution of absolute (left)
125: and effective (right) gain, the gas mixture
126: is 50\% Argon \%50 DME. The GEM is operated of $480V$ with a collection field
127: of $4KV/cm$.}
128: \label{fig:garco2}
129: \end{center}
130: \end{figure}
131:
132: Sometimes the distribution shows two clear maxima and a two Polya fit is
133: satisfactory. In this case the mean of the two Polya is the average gain of
134: the GEM under
135: analysis. Results are shown in the figure \ref{fig:geco2}, \ref{fig:geardme1}
136: for a collecting field $E_t\;=\;4KV/cm$ and two gas mixtures and in the table \ref{tab:co2051} for a collecting field of $E_t\;=\;5KV/cm$.
137:
138: \vspace{0.5 cm}
139: \begin{figure}[h]
140: \begin{center}
141: \mbox{\epsfig{file=figure/GvxHCO2100_0.5atm.epsi,
142: width=0.35\linewidth,angle=270}}
143: %{\resizebox*{8 cm}{!}
144: % {\includegraphics[angle=270]{figure/GvxHCO2100_0.5atm.epsi}}}
145: \caption{\small Dependence of effective gain on the GEM voltage. The simulation
146: concerns a gas of 100\% $CO_2$ and a collection field of $4KV/cm$.}
147: \label{fig:geco2}
148: \end{center}
149: \end{figure}
150:
151:
152: \begin{table}[h]
153: \begin{center}
154: \begin{tabular}{c || c | c }
155: %%%%%100
156: $ \Delta V_{GEM} = 560\;V\quad E_{Trasf} = 5KV/cm$ &
157: $p=0.5 atm$ & $p=1 atm $ \\
158: \hline \hline
159: $Gain_{eff}$ & 830 $\pm$ 100 & 70 $\pm$ 10 \\
160: \hline
161: $Gain_{ass}$ & 7420 $\pm$ 400 & 2400 $\pm$ 900 \\
162: \end{tabular}
163: \end{center}
164: \caption{\small Average of effective and absolute gain of 100\% $CO_2$ at
165: 0.5 and 1 atm and for $\Delta V_{GEM}$ of $560V$ and collection field of $5KV/cm$.}
166: \label{tab:co2051}
167: \end {table}
168:
169: \begin{figure}[h]
170: \begin{center}
171: \mbox{\epsfig{file=figure/GvxHARDMEa_1atm.epsi,
172: width=0.33\linewidth,angle=270},
173: \epsfig{file=figure/GvxHARDME80-20_1atm.epsi,
174: width=0.33\linewidth,angle=270}
175: }
176: \caption{\small Left: gas mixtures of 20\%Ar-80\%DME and 50\%Ar-50\%DME.
177: The plots show the dependence of the effective gain on the GEM voltage.
178: The collection field is $4KV/cm$.
179: Right: gas mixture 80\%Ar-20\%DME with a large production of secondary
180: avalanches. The figure shows the mean of effective gain as a function of the
181: GEM voltage and for a collection field of $4KV/cm$.
182: The values of gain are the mean of first (gain low) and second
183: (gain hight) Polya which fit the two peaks observed in the
184: gain distribution.}
185: \label{fig:geardme1}
186: \end{center}
187: \end{figure}
188:
189: The GEM gain increases with the voltage different GEM according to an exponential curve.
190:
191: \section*{Diffusion Study}
192:
193: The study of the diffusion of the charge in the
194: collecting region of the detector is important for two different issues.
195:
196: \begin{wrapfigure}[16]{r}{8 cm}
197: \begin{center}
198: \mbox{
199: \epsfig{file=figure/diff8.epsi,width=7 cm}}
200: ù\caption{\small $\Delta R$ in function of $R_{in}$.}
201: \label{fig:diff8}
202: \end{center}
203: \end{wrapfigure}
204:
205:
206: Firstly to establish if the GEM keeps memory of the starting point of
207: the electrons, both in azimuth and radius (with respect
208: to the center of the hole
209: where the avalanche occurs) with a better resolution than the granularity of
210: GEM's hole.
211:
212: For this, we have studied the position of the barycentre of charge
213: arrived on readout plane
214: (barycentre of the avalanche) as a
215: function of the position of the starting point.
216:
217: The second point is the RMS of Gaussian
218: distribution of charge in the collection gap which is related with the spatial
219: resolution of the detector.
220:
221: The average position of the collected charge indicates where the multiplication
222: occurs at the GEM hole.
223:
224: We have considered
225: \[
226: \Delta R\;=\;R_{fin}\;-\;R_{in}
227: \]
228: where $R_{in}\;=\;r$ and $R_{fin}$ is the radius of the average charge at the
229: quota $z\;=\;-200 \mu m$.
230: First of all we have studied the dependence of $\Delta R$ on $R_{in}$
231:
232: The figure \ref{fig:diff8} shows an example of such a dependence.
233: $ \Delta R$ is a linear function of $ R_{in}$ and a parameterization:
234: $ \Delta R\;=P_2 R_{in}\;+\;P_1 $
235: with $P_2 \approx -1$ and $P_1 \approx 20
236: \mu m$ is a good fit for all simulations (figure \ref{fig:p1}).
237: For a ideal GEM $\Delta R = 0$ in fact this means that the position (in radius)
238: of the multiplied charge is the same as the initial electron (the GEM does
239: not disturb the image). The results of the fit shows instead that the average
240: collected charge position is independent of the starting position ($P_2
241: \approx -1$) and that the multiplication occurs at the radius $r \approx 20-30
242: \mu m$ (the lower external radius).
243:
244: \begin{figure}[ht]
245: \begin{center}
246: \mbox{\epsfig{file=figure/AR-Ne_p1.epsi,
247: width=0.33\linewidth,angle=270},
248: \epsfig{file=figure/AR-Ne_p2.epsi,
249: width=0.33\linewidth,angle=270}}
250: \caption{\small The values of $p1$ ($\mu m$) and $p2$ in function of percentage
251: of Argon and Neon.}
252: \label{fig:p1}
253: \end{center}
254: \end{figure}
255:
256: The last step is the study of the dispersion of the avalanche's charge after
257: it drifted to the collection electrodes.
258:
259: We have averaged the RMS of the events produced at each point and verified
260: that its value is independent of the position of primary electron. Hence we
261: have averaged the RMS of all events at all points, to improve the statistics
262: and studied its dependence on the GEM voltage.
263: Since, again, we have found no dependence, we have average on all events for a
264: defined gas composition.
265:
266: The results for 100\%$CO_{2}$ gas at $0.5\;atm$ are:
267: \begin{itemize}
268: \item $RMS_{p}\;=\;12.5\;\mu m$
269: \item $RMS_{sp}\;=\;17.67\;\mu m$
270: \end{itemize}
271:
272: The results for Argon and Neon mixtures are shown in figure
273: \ref{fig:AR-Ne_RMSp}, the RMS decreases mildly with increasing percentage of
274: Argon and Neon.
275:
276: \begin{figure}[ht]
277: \begin{center}
278: \mbox{\epsfig{file=figure/AR-Ne_RMSp.epsi,
279: width=0.33\linewidth,angle=270},
280: \epsfig{file=figure/AR-Ne_RMSsp.epsi,
281: width=0.33\linewidth,angle=270}}
282: \caption{\small Left we show the projected $RMS$ ($\mu m$) averaged on all events for
283: two gas mixtures while on the right the spatial $RMS$ ($\mu m$).}
284: \label{fig:AR-Ne_RMSp}
285: \end{center}
286: \end{figure}
287: \section*{Acknowledgments}
288: I would like ti thank G.Spandre for the help and continuous advice on my work
289: and R.Veenhof for his support in the use of the simulation program Garfield
290: and also for many advice on how to perform reliable simulations.
291: \end{document}