1: \documentclass[11pt]{article}
2: \usepackage{moriond,epsfig,graphicx}
3:
4: \bibliographystyle{unsrt}
5: \newcommand {\um} {\, \mu \rm m}
6: \newcommand {\cm} {\rm \, cm}
7: \def \e {{\mathrm e}}
8: \def \Z {{\mathrm Z}}
9: \newcommand {\ee} {{\e ^+ \e ^-}}
10: \newcommand {\gtrsim}
11: {\,\raisebox{-0.6ex}{$\stackrel{\textstyle>}{\textstyle\sim}$}\,}
12: \newcommand {\nm} {\rm \, nm}
13: \newcommand {\ns} {\rm \, ns}
14: \begin{document}
15: \vspace*{4cm}
16: \title{AN UNEXPECTED EFFECT IN ALEPH: LONG-TERM DISPLACEMENT
17: OF THE SILICON VERTEX DETECTOR}
18:
19: \author{ Giacomo Sguazzoni }
20:
21: \address{European Laboratory for Particle Physics (CERN), CH--1211
22: Geneva 23, Switzerland}
23:
24: \maketitle\abstracts{
25: The ALEPH Silicon Vertex Detector for LEP2 featured a
26: laser survey system to monitor its mechanical stability. The analysis of
27: laser system data from 1997 to 2000 showed that VDET suffered a
28: time-dependent displacement. It resulted to be compatible with a
29: deformation of the support structure that made the device to slowly
30: rotate during the data-taking. A~maximal local displacement of
31: $\sim$20$\um$ was observed, corresponding to a rotation of
32: $\sim$$10^{-4}{\rm rad}$.~The implementation of a time-dependent
33: correction on the alignment by using the laser system data led to
34: sizeable improvements on the ALEPH data quality.}
35:
36: \section{The ALEPH Silicon Vertex Detector}
37:
38: Many physics studies of the high energy phase of LEP (LEP2) relied
39: on powerful heavy-flavour tags provided by high-resolution
40: vertex detectors.
41:
42: The ALEPH Silicon Vertex Detector$\,$\cite{vdetref} (VDET) had
43: an active length of $\sim$40$\cm$ and consisted of two concentric cylindrical
44: layers of 144 micro-strip silicon detectors of $\sim$5.3$\cm
45: \times$6.5$\cm$ with double-sided readout. The cylinders' axes
46: coincided with the nominal $\ee$ beam axes, i.e.~the $z$-axis in the
47: ALEPH reference system. Six silicon detectors were glued together and
48: instrumented with readout electronics to form the VDET
49: elementary unit ({\em face}). The inner layer ($\sim$6.3$\cm$ radius)
50: was formed by 9 faces, the outer layer ($\sim$10.5$\cm$ radius)
51: consisted of 15 faces.
52:
53: \section{Laser System}
54:
55: VDET featured a laser survey system$\,$\cite{laser1} to monitor its mechanical
56: stability with respect to the external tracking devices, since an
57: undetected large movement ($\gtrsim$20$\um$) of VDET during the
58: data-taking could have degraded significantly its performances.
59: The VDET alignment was normally performed at the beginning
60: of the annual data-taking with events collected during
61: a dedicated calibration run at the Z resonance peak. Afterwards the reduced
62: event rate of LEP2 energies did not allow a precise survey to be done by
63: using particle tracks, and the laser system thus played a crucial
64: role in monitoring the stability of the device.
65:
66: The system made use of infrared light ($\lambda = 904\nm$)
67: from two pulsed laser diodes with an output power of $6\,$W
68: each and a pulse length of $50\ns$.
69: The light was distributed via optical fibers to prisms and lenses
70: attached to the inner wall of the Inner Tracking Chamber (ITC, the closest
71: outer detector). The lenses focused several light beams
72: on 14 of the 15 VDET outer faces. Information on the VDET
73: displacements with respect to the ITC were obtained by monitoring the
74: laser beam impact position ({\em spot}\/) on the silicon wafers.
75: All laser beams were nominally parallel
76: to the ($xy$)-plane. As shown in Fig.~\ref{fig:vdetface}a
77: and in Fig.~\ref{fig:vdetface}b, there were normally three spots per face: the
78: ones close to the face ends had the beam direction normal to the silicon
79: surface; the one placed about in the middle had the beam direction
80: at $\sim$45$^\circ$ with respect to the silicon surface, in order to be
81: also sensitive to displacements normal to the face.
82: %
83: \begin{figure}[t]
84: \begin{center}
85: \leavevmode
86: \begin{picture}(0,0)
87: \put(18,20){\mbox{\small (a)}}
88: \end{picture}
89: \epsfig{file=fig1a.eps, width=0.60\textwidth}
90: \begin{picture}(0,0)
91: \put(23,20){\mbox{\small (b)}}
92: \end{picture}
93: \epsfig{file=fig1b.eps, width=0.37\textwidth}
94: \caption{(a) Sketch of laser spots on two VDET faces; (b) VDET
95: ($r\phi$)-section with laser beams.}
96: \label{fig:vdetface}
97: \label{fig:vdetsec}
98: \end{center}
99: \end{figure}
100: %
101:
102: The system was operating during standard data-taking: the lasers were
103: fired approximately every $\sim$100 physics triggers (i.e.~once per one
104: to two minutes) and the spots were reconstructed as standard particle
105: hits. The spot position resolution was impressive, typically
106: $\sim$0.5--$1.5 \um$, thanks to the large signal and the cluster width
107: extending over more than one readout strip. After the installation five
108: spots out of 44 were missed, probably due to misalignment or breakage
109: of the optics. The number of collected laser events was 62k in 1997,
110: 129k in 1998, 125k in 1999, and 170k in 2000.
111:
112: \section{Analysis of Laser System Data}
113:
114: The analysis of the laser system data consisted in studying the time
115: dependence of the spot positions. Along with the expected short-term
116: displacements due to temperature effects$\,$\cite{laser2}, a long-term
117: dependence was discovered in ($r\phi$)-spot positions since 1997. As an
118: example, this behaviour over the entire data-taking
119: period in 2000 is shown in Fig.~\ref{fig:rawtc}a and in Fig.~\ref{fig:rawtc}b,
120: referring to endcap A ($z<0$) and endcap B ($z>0$) spots,
121: respectively. Some time-chart is not present because the corresponding
122: spot was missing or removed from the analysis because of low
123: quality$\,$\cite{laser2}.
124: The size of the ($r\phi$)-displacements was systematically dependent on
125: the face azimuthal position, in some case being as large as
126: $\sim$10--20$\um$, thus comparable to the charged track single hit
127: resolution ($\sim$10$\um$). On the contrary, the $z$-side deviations did not
128: show any critical~effect.
129:
130: \begin{figure}[t]
131: \begin{center}
132: \begin{picture}(0,0)
133: \put(12,12){\mbox{\small (a)}}
134: \put(146,12){\mbox{\small (b)}}
135: \put(283,8){\mbox{\small (c)}}
136: \end{picture}
137: \epsfig{file=fig2ab.eps, width=0.58\textwidth}
138: \hskip 3mm
139: \includegraphics[bb=24 200 521 764,width=0.38\textwidth]{fig2c.eps}
140: \caption{Time-charts of ($r\phi$)-deviations for (a) endcap A ($z<0$)
141: and (b) endcap B ($z>0$) spots in 2000 laser data. Each
142: time-chart is identified by the face number (as in
143: Fig.~\ref{fig:reconstruction}) and shifted for graphical
144: reasons. (c)~Two-days averaged correction parameters as
145: a function of time from the beginning of the data-taking in 2000.}
146: \label{fig:rawtc}
147: \label{fig:parameters}
148: \end{center}
149: \end{figure}
150: %\newpage
151:
152: \begin{figure}[t]
153: \begin{center}
154: \hskip -2mm
155: \begin{picture}(0,0)
156: \put(20,20){\mbox{\small (a)}}
157: \put(177,20){\mbox{\small (b)}}
158: \put(337,86){\mbox{\small (c)}}
159: \put(337,2){\mbox{\small (d)}}
160: \end{picture}
161: \includegraphics[bb=30 145 562 686, height=0.365\textwidth]{fig3a.eps}
162: \includegraphics[bb=60 145 562 686, height=0.365\textwidth]{fig3b.eps}
163: \hskip 2mm
164: \epsfig{file=fig3cd.eps, height=0.36\textwidth}
165: \caption{Pictorial view of the reconstructed displacements
166: for (a) endcap A and (b) endcap B spots: the dots represent their
167: positions during time. The scatter plots outside the outer layer
168: are the raw deviation time-charts of the corresponding spot. All
169: displacements are amplified by a factor 1000. The cross is the
170: probable rotation center.
171: (c) Distribution of sum of impact parameters of the two
172: tracks in $\Z\to\mu\mu$ events before the
173: laser correction to the alignment and (b) after the alignment correction.
174: The plots refers to a Z calibration run taken in year 2000,
175: approximately 150 days after the beginning of data-taking.}
176: \label{fig:reconstruction}
177: \label{fig:sumd0}
178: \end{center}
179: \end{figure}
180: %
181:
182: %\pagebreak
183:
184: The long-term deviations were likely to be due to a global rotation of
185: VDET in the ($xy$)-plane, as suggested by the distinctive geometrical
186: configuration.
187: The deviations were thus parametrized under the assumption that
188: VDET was moving as a rigid body, i.e.~by using three parameters to correct
189: the nominal alignment: the $x$-translation, the $y$-translation and the rotation
190: angle around the $z$-axis. These were extracted by a fit
191: procedure. As an example, Fig.~\ref{fig:parameters}c reports the
192: parameters relative to year 2000 plotted as a function of the time
193: from the beginning of the data-taking. A pictorial view of the corresponding VDET
194: displacement is shown in Fig.~\ref{fig:reconstruction}a and in
195: Fig.~\ref{fig:reconstruction}b. The reconstructed motion was
196: consistent with a rotation and the rigid body model was found to
197: reproduce the observed spot deviations within a few
198: microns$\,$\cite{laser2}. The maximum $z$-rotation value was
199: $\sim$7$\cdot10^{-5}$rad in 1997, $\sim$8$\cdot10^{-5}$rad in 1998,
200: $\sim$3$\cdot10^{-5}$rad in 1999 and $\sim$9$\cdot10^{-5}$rad in 2000.
201:
202: During the last three years of LEP running, several runs at the Z resonance peak
203: were also taken for calibration purpose, during and at the
204: end of the data-taking. The standard alignment performed by using
205: these calibration data allowed to independently confirm the motion of
206: VDET as extracted from the laser system, definitely proving its
207: reliability. A time-dependent correction of the alignment based on the
208: laser data was thus applied from 1998 on.
209:
210: The correction had a sensible
211: impact on the data quality. This was pointed out studying
212: the Z calibration runs, where a significant amount of data was
213: collected in a period of time much shorter than the time-scale of the
214: effect under observation. Figure~\ref{fig:sumd0}c and Fig.~\ref{fig:sumd0}d
215: shows the distribution of the sum of the impact parameters of the two
216: muons in $\Z\to\mu\mu$ events, before and after applying the laser
217: correction. The plots refer to the Z calibration run taken in 2000,
218: $\sim$150~days after the beginning of data-taking, when the
219: displacement averaged over all VDET faces was about
220: $\sim$10$\um$. Without the correction the systematic shift of the
221: distribution is of the same order of magnitude. Applying the laser
222: correction the mean of the distribution is again compatible with
223: zero.
224:
225: The rotation may be explained considering that VDET is supported by
226: flanges which slot into two long metal rails located at the top
227: and bottom of the ITC inner cylinder.
228: The top flanges were not metallic as the bottom ones, but made of
229: springy plastic to compensate for distance variations between
230: the rails along which VDET slided during the installation.
231: The VDET environmental conditions during running were quite different
232: with respect to the shutdown period, both in terms of temperature and
233: humidity. They probably had some long-term effect on the material of
234: the plastic flanges, causing a tiny deformation that made VDET to
235: slowly rotate around the lower rail (represented by the small crosses in
236: Fig. \ref{fig:reconstruction}a and in Fig.~\ref{fig:reconstruction}b).
237:
238: As a further indication of the environmental conditions as the possible cause,
239: the deformation was found to recover during the shutdown. In the
240: 1999--2000 shutdown VDET was not removed for maintenance allowing its
241: position to be monitored. At the beginning of the data-taking in 2000
242: the position resulted again the same as at the beginning of the
243: data-taking in 1999.
244:
245: Extrapolating the LEP experience to the next generation of
246: experiments, the design issues of the Silicon Tracker devices
247: for LHC turn out to be extremely challenging. The huge scale
248: involved forces the effects as the one here described to be carefully
249: taken into account. In fact, as an example, studies for the CMS Tracker
250: support structure have shown that a variation of environmental humidity
251: may lead to material deformation with very long
252: time-scale$\,$\cite{cms}.
253:
254: \section{Conclusions}
255:
256: The laser system of the ALEPH Vertex Detector for LEP2 was a simple but
257: extremely effective survey system. It allowed a reliable and high-precision
258: monitoring of the VDET position during the data-taking revealing a tiny
259: long-term rotation. An alignment correction based on the
260: information from this system was successfully applied.
261:
262: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
263: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
264: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
265:
266: \begin{thebibliography}{99}
267:
268: \bibitem{vdetref}
269: D.~Creanza
270: et al., {\em The new ALEPH Silicon Vertex Detector},
271: Nucl. Instrum. Methods {\bf A~409} (1998) 157; ALEPH
272: Collaboration, {\em The ALEPH Handbook 1995}, ed. C.~Bowdery, ISBN
273: 92-9083-073-5 (1995).
274:
275: \bibitem{laser1} A.~Wagner, H.~Dietl,
276: {\em The Laser Calibration System for VDETII},
277: ALEPH 97-006 (1997).
278:
279: \bibitem{laser2}
280: G.~Sguazzoni et al., {\em Monitoring the Stability of the ALEPH
281: Silicon Vertex Detector}, Nucl. Phys. {\bf B 78} (1999) 301.
282:
283: \bibitem{cms}
284: S.~Da~Mota~Silva et al., {\em Hygro-Thermal Transient Analysis for
285: Highly Stable Structures}, CMS CR 1999/014 (1999), published in
286: ``Proceedings of ICCM-12, 12th International Conference on Composite
287: Materials, 1999, Paris''.
288:
289: \end{thebibliography}
290:
291: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
292: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
293: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
294:
295: \end{document}
296: