1: \section{Silicon Vertex Tracker}

2: \label{sec:svt}

3: \subsection{ Charged Particle Tracking}

4: The principal purpose of the \babar\ charged particle tracking

5: systems, the SVT and the DCH, is the efficient detection of

6: charged particles and the measurement of their momentum and angles

7: with high precision.  Among many applications, these precision

8: measurements allow for the reconstruction of exclusive $B$- and

9: $D$-meson decays with high resolution and thus minimal background.

10: The reconstruction of multiple decay vertices of weakly decaying

11: $B$ and $D$ mesons is of prime importance to the physics goals.

12: 

13: Track measurements are also important for the extrapolation

14: to the DIRC, EMC, and IFR.

15: At lower momenta, the DCH measurements are more important,

16: while at higher momenta the SVT measurements dominate.

17: Most critical are the angles at the DIRC,

18: because the uncertainties in the charged

19: particle track parameters add to the uncertainty in the

20: measurement of the Cherenkov angle.

21: Thus, the track errors from the combined SVT and DCH measurements

22: should be small compared to the average DIRC Cherenkov angle measurements,

23: \ie\ of order of 1\mrad, particularly at the highest momenta.

24: 

25: \subsection{SVT Goals and Design Requirements}

26: 

27: The SVT has been designed to provide precise reconstruction of

28: charged particle trajectories and decay vertices near the

29: interaction region. The design choices were driven primarily by

30: direct requirements from physics measurements and constraints

31: imposed by the PEP-II interaction region and \babar\ experiment.

32: In this section  the mechanical and electronic design of the SVT

33: are discussed, with some discussion of the point resolution per

34: layer and \dedx\ performance. The tracking performance and

35: efficiency of the SVT alone and in combination with the DCH are

36: described in Section~\ref{sec:trk}.

37: 

38: 

39: \subsubsection{SVT Requirements and Constraints}

40: The SVT is critical for the measurement of the time-dependent \CP\

41: asymmetry. To avoid significant impact of the resolution on the

42: \CP\ asymmetry measurement the mean vertex resolution along the

43: $z$-axis for a fully reconstructed $B$ decay must be better than

44: 80\mum~\cite{R_LOI}. The required resolution in the $x$--$y$ plane

45: arises from the need to reconstruct final states in $B$ decays as

46: well as in \mtau\ and charm decays. For example, in decays of the

47: type \Bztodd, separating the two $D$ vertices is important. The

48: distance between the two $D$'s in the $x$--$y$ plane for this

49: decay is typically $\sim 275 $\mum. Hence, the SVT needs to

50: provide  resolution of order $\sim$100\mum\ in the plane

51: perpendicular to the beam line.

52: 

53: Many of the decay products of $B$ mesons have low \pt. The SVT

54: must provide standalone tracking for particles with transverse

55: momentum less than 120\mevc, the minimum that can be measured

56: reliably in the DCH alone.  This feature is fundamental for the

57: identification of slow pions from \Dstar-meson decays: a tracking

58: efficiency of 70\% or more is desirable for tracks with a

59: transverse momentum in the range 50--120\mevc. The standalone

60: tracking capability and the need to link SVT tracks to the DCH

61: were crucial in choosing the number of layers.

62: 

63: Beyond the standalone tracking capability, the SVT provides the

64: best measurement of track angles, which is required to achieve

65: design resolution for the Cherenkov angle for high momentum

66: tracks.

67: 

68: Additional constraints are imposed by the storage ring components.

69: The SVT is located inside the $\sim$4.5\m-long support tube, that

70: extends all the way through the detector.  To maximize the angular

71: coverage, the SVT must extend down to 350\mrad\ (20\degrees) in

72: polar angle from the beam line in the forward direction.  The

73: region at smaller polar angles is occupied by the B1 permanent

74: magnets.  In the backward direction, it is sufficient to extend

75: the SVT sensitive area down to 30\degrees.

76: 

77: The SVT must withstand 2 MRad of ionizing radiation. A radiation

78: monitoring system capable of aborting the beams is required. The

79: expected radiation dose is 1~Rad/day in the horizontal plane

80: immediately outside the beam pipe (where the highest radiation is

81: concentrated), and 0.1~Rad/day on average otherwise.

82: \label{req:radhard}

83: 

84: The SVT is inaccessible during normal detector operations. Hence,

85: reliability and robustness are essential: all components of the

86: SVT inside the support tube should have long mean-time-to-failure,

87: because the time needed for any replacement is estimated to be

88: 4--5 months. Redundancies are built in whenever possible and

89: practical.

90: 

91: The SVT is cooled to remove the heat generated by the electronics.

92: In addition, it operates in the 1.5\tesla\ magnetic field.

93: 

94: To achieve the position resolution necessary to carry out physics

95: analyses, the relative position of the individual silicon

96: \emph{sensors} should be stable over long time periods. The

97: assembly allows for relative motion of the support structures with

98: respect to the B1 magnets.

99: 

100: \begin{figure}

101: \includegraphics[width=7.5cm]{svt_overall_pic}

102: 

103: \vspace{-2pc}

104:  \caption{Fully assembled SVT. The silicon sensors of

105: the outer layer are visible, as is the carbon-fiber space frame

106: (black structure) that surrounds the silicon.} \label{\secname

107: fig:svt_overall_pic}

108: \end{figure}

109: 

110: 

111: \begin{figure*}[htb]

112: \includegraphics[width=16cm]{sideview}

113: \vspace{-2pc}

114: 

115: \caption{Schematic view of SVT: longitudinal section. The roman

116: numerals label the six different types of sensors.}

117: \label{\secname fig:sideview}

118: \end{figure*}

119: 

120: These requirements and constraints have led to the choice of a SVT

121: made of five layers of double-sided silicon strip sensors. To

122: fulfill the physics requirements, the spatial resolution, for

123: perpendicular tracks, must be 10--15\mum\ in the three inner

124: layers and about 40\mum\ in the two outer layers.  The inner three

125: layers perform the impact parameter measurements, while the outer

126: layers are necessary for pattern recognition and low \pt\

127: tracking.

128: 

129: 

130: \subsection{SVT Layout}

131: 

132: The five layers of double-sided silicon strip sensors, which form

133: the SVT detector, are organized in 6, 6, 6, 16, and 18 modules,

134: respectively; a photograph is shown in Figure~\ref{\secname

135: fig:svt_overall_pic}. The strips on the opposite sides of each

136: sensor are oriented orthogonally to each other: the \mphi\

137: measuring strips (\emph{\mphi\ strips}) run parallel to the beam

138: and the $z$ measuring strips (\emph{$z$ strips}) are oriented

139: transversely to the beam axis.  The modules of the inner three

140: layers are straight, while the modules of layers 4 and 5 are

141: \emph{arch}-shaped (Figures~\ref{\secname fig:sideview} and

142: \ref{\secname fig:endview}).

143: 

144: This arch design was chosen to minimize the amount of silicon required

145: to cover the solid angle, while increasing the crossing angle for

146: particles near the edges of acceptance.  A photograph of an outer

147: layer arch module is shown in Figure~\ref{\secname fig:arch}.  The

148: modules are divided electrically into two half-modules, which are read

149: out at the  ends.

150: 

151: \begin{figure}[htb]

152: \includegraphics[width=7.5cm]{endview}

153: \vspace{-2pc} \caption{Schematic view of SVT: tranverse section.}

154: \label{\secname fig:endview}

155: \end{figure}

156: 

157: 

158: 

159: To satisfy the different geometrical requirements of the five SVT

160: layers, five different sensor shapes are required to assemble the

161: planar sections of the layers. The smallest detectors are 43

162: $\times$ 42\mm$^2$ ($z\ \times\ \phi$), and the largest are 68

163: $\times$ 53\mm$^2$. Two identical trapezoidal sensors are added

164: (one each at the forward and backward ends) to form the arch

165: modules.  The half-modules are given mechanical stiffness by means

166: of two carbon fiber/kevlar ribs, which are visible in

167: Figure~\ref{\secname fig:arch}. The \mphi\ strips of sensors in

168: the same half-module are electrically connected with wire bonds to

169: form a single readout strip.  This results in a total strip length

170: up to 140\mm\ (240\mm) in the inner (outer) layers.

171: 

172: 

173: 

174: The signals from the $z$ strips are brought to the readout

175: electronics using fanout circuits consisting of conducting traces

176: on a thin (50 \mum) insulating Upilex~\cite{upilex} substrate. For

177: the innermost three layers, each $z$

178: \begin{figure}

179: \includegraphics[width=7.5cm]{arch}

180: \vspace{-2.1pc}

181: 

182:  \caption{Photograph of an SVT arch module in an

183: assembly jig.} \label{\secname fig:arch}

184: \end{figure}

185: strip is connected to its own preamplifier channel, while in layers 4

186: and 5 the number of $z$ strips on a half-module exceeds the number of

187: electronics channels available, requiring that two $z$ strips on

188: different sensors be

189: electrically connected (ganged) to a single electronics channel.  The

190: length of a $z$ strip is about 50\mm\ (no ganging) or 100\mm\ (two strips

191: connected).  The ganging introduces an ambiguity on the $z$ coordinate

192: measurement, which must be resolved by the pattern recognition

193: algorithms.  The total number of readout channels is approximately

194: 150,000.

195: 

196: The inner modules are tilted in \mphi\ by 5\degrees, allowing an

197: overlap region between adjacent modules, a feature that provides

198: full azimuthal coverage and is advantageous for alignment. The

199: outer modules cannot be tilted, because of the arch geometry. To

200: avoid gaps and to have a suitable overlap in the \mphi\

201: coordinate, layers 4 and 5 are divided into two sub-layers (4a,

202: 4b, 5a, 5b) and placed at slightly different radii (see

203: Figure~\ref{\secname fig:endview}). The relevant geometrical

204: parameters of each layer are summarized in Table~\ref{\secname

205: tab:geometrical}.

206: 

207: \begin{table}[!htb]

208: \caption{Geometric parameters for each layer and readout plane of

209: the SVT. Floating strips refers to the number of strips between

210: readout (R-O) strips. Note: parts of the $\phi$ sides of layers 1

211: and 2 are bonded at 100 \mum\ and 110 \mum\ pitch, respectively,

212: with one floating strip. Strip length of $z$-strips for layers 4

213: and 5  includes ganging. The radial range for layers 4 and 5

214: includes the radial extent of the arched sections.}

215: \medskip

216: \begin{tabular}{lcccr} \hline

217: \rule{0pt}{12pt}  &  &R-O  & & Strip\\

218: Layer/&Radius &pitch &Floating   &length\\

219: view \rule[-5pt]{0pt}{0pt}    &(mm) &(\mum) &strips  &(mm)\\ \hline

220: \rule{0pt}{12pt}1 z & 32       & 100

221: & 1 & 40 \\ 1 $\phi$ & 32 & 50-100& 0-1 & 82\\ 2 z        & 40

222: & 100 & 1 & 48  \\ 2 $\phi$   & 40       & 55-110 & 0-1 & 88\\ 3 z

223: & 54 & 100 & 1 & 70  \\ 3 $\phi$ & 54       & 110  & 1 & 128 \\ 4

224: z & 91-127 & 210 & 1 & 104 \\ 4 $\phi$   & 91-127  & 100 & 1 & 224

225: \\ 5 z        & 114-144 & 210 & 1 & 104 \\

226: 5\rule[-5pt]{0pt}{0pt}  $\phi$   & 114-144 &

227: 100 & 1 & 265 \\ \hline

228: \end{tabular}

229: \label{\secname tab:geometrical}

230: \end{table}

231: 

232: In order to minimize the material in the acceptance region, the

233: readout electronics are mounted entirely outside the active

234: detector volume.  The forward electronics must be mounted in the

235: 10\mm\ space between the 350\mrad\ stayclear space and B1 magnet.

236: This implies that the hybrids carrying the front-end chip must be

237: positioned at an angle of 350\mrad\ relative to the sensor for the

238: layers 3, 4, and 5 (Figure~\ref{\secname fig:sideview}).  In the

239: backward direction, the available space is larger and the inner

240: layer electronics can be placed in the sensor plane, allowing a

241: simplified assembly.

242: 

243: The module assembly and the mechanics are quite complicated,

244: especially for the arch modules, and are described in detail

245: elsewhere~\cite{module_assembly}. The SVT support structure

246: (Figure~\ref{\secname fig:svt_overall_pic}) is a rigid body made

247: from two carbon-fiber cones, connected by a \emph{space frame},

248: also made of carbon-fiber epoxy laminate.

249: 

250: An  optical survey of the SVT on its assembly jig indicated that

251: the global error in placement of the sensors with respect to

252: design was $\sim$200\mum, FWHM.  Subsequently, the detector was

253: disassembled and shipped to SLAC, where it was re-assembled on the

254: IR magnets. The SVT is attached to the B1 magnets by a set of

255: gimbal rings in such a way as to allow for relative motion of the

256: two B1 magnets while fixing the position of the SVT relative to

257: the forward B1 and the orientation relative to the axis of both B1

258: dipoles.  The support tube structure is mounted on the PEP-II

259: accelerator supports, independently of \babar, allowing for

260: movement between the SVT and the rest of \babar. Precise

261: monitoring of the beam interaction point is necessary, as is

262: described in Section \ref{sec:svtmonitoring}.

263: 

264: The total active silicon area is 0.96\m$^2$ and the material

265: traversed by particles is $\sim 4$\% of a radiation length (see

266: Section~\ref{sec:layout}). The geometrical acceptance of SVT is

267: 90\% of the solid angle in the c.m.~system, typically 86\% are

268: used in charged particle tracking.

269: 

270: \subsection{SVT Components}

271: 

272: A block diagram of SVT components is shown in Figure~\ref{\secname

273: fig:block}. The basic components of the detector are the silicon

274: sensors, the \emph{fanout} circuits, the \emph{Front End

275: Electronics} (FEE) and the data transmission system. Each of these

276: components is discussed below.

277: 

278: \begin{figure}[htb]

279: \centering

280: \includegraphics[width=7.5cm]{fig_block}

281: \vspace{-2.5pc}

282:  \caption{Schematic block diagram showing the

283: different components of the SVT.} \label{\secname fig:block}

284: \end{figure}

285: 

286: 

287: \subsubsection{Silicon Sensors}

288: 

289: The SVT sensors~\cite{silicon_detectors} are 300\mum\ thick

290: double-sided silicon strip devices.  They were designed at INFN

291: Pisa and Trieste (Italy) and fabricated

292: commercially~\cite{micron}. They are built on high-resistivity

293: (6--15\kohm-cm) n-type substrates with p$^+$ strips and n$^+$

294: strips on the two opposite sides.  The insulation of the n$^+$

295: strips is provided by individual p-stops, so as to achieve an

296: inter-strip resistance greater than 100\Mohm\ at operating bias

297: voltage, normally about 10 V above the depletion voltage.

298: 

299: Typical depletion voltages are in the range 25--35\volt. The

300: strips are biased on both sides with polysilicon resistors

301: (4--20\Mohm) to ensure the required radiation hardness, keeping

302: the voltage drop across resistors and the parallel noise as low as

303: possible. Strips are AC-coupled to the electronics via integrated

304: decoupling capacitors, the capacitance of which depends on the

305: sensor shape, but is always greater than 14\pf/cm.  The sensors

306: were designed to maximize the active area, which extends to within

307: 0.7\mm\ of the physical edges.  Another design goal was to control

308: the inter-strip capacitance: values between 0.7\pf/cm and

309: 1.1\pf/cm were obtained for the various sensor shapes.  To achieve

310: the required spatial resolution, while keeping the number of

311: readout channels as low as possible, most of the modules have a

312: \emph{floating strip} (\ie\ not read out) between two readout

313: strips.

314: 

315: \begin{table}

316: \centering

317:  \caption{Electrical parameters of the SVT, shown for

318: the different layers and views. $C_{input}$ refers to the total

319: input capacitance, $R_{series}$ is the series resistance. The

320: amplifier peaking time is 200\ns\ for layers 1--3 and 400\ns\ for

321: layers 4--5.}

322: 

323: \vspace{\baselineskip}

324: 

325: \begin{tabular}{lrrrrrr} \hline

326: \rule{0pt}{12pt} &  &  &

327: \multicolumn{2}{c}{Noise,}\\

328: Layer/  &$C_\mathrm{input}$   &$R_\mathrm{series}$ &calc.  &meas. \\

329: \rule[-5pt]{0pt}{0pt}view &(pF)  &($\Omega$)  &(elec)& (elec)\\ \hline

330: \rule{0pt}{12pt}1 z        &  6.0 &  40.  &  550 & 880  \\ 1

331: $\phi$ & 17.2 & 164.  & 990  & 1200 \\

332: 2 z   &  7.2 & 48.   &600  & 970

333: \\ 2 $\phi$   & 18.4 & 158.  & 1030 & 1240 \\ 3 z        & 10.5 &

334: 70. & 700  & 1180  \\ 3 $\phi$   & 26.8 & 230.  & 1470 & 1440 \\ 4

335: z & 16.6 & 104.  & 870  & 1210 \\ 4 $\phi$   & 33.6 & 224.  & 1380

336: & 1350 \\ 5 z        & 16.6 & 104.  & 870  & 1200 \\

337: \rule[-5pt]{0pt}{0pt}5 $\phi$   & 39.7 & 265.  & 1580 & 1600 \\

338: \hline

339: \end{tabular}

340: \label{\secname tab:electrical}

341: \end{table}

342: 

343: The leakage currents, because of the excellent performance of the

344: manufacturing process, were as low as 50\nA/cm$^2$ on average,

345: measured at 10\volt\ above depletion voltage.  The silicon sensor

346: parameters have been measured after irradiation with $^{60}$Co

347: sources. Apart from an increase in the inter-strip capacitance of

348: about 12\% during the first 100\krad, the main effect was an

349: increase of the leakage current by 0.7\muA/cm$^2$/MRad.  However,

350: in a radiation test performed in a 1\gevc\ electron beam, an

351: increase in leakage current of about 2\muA/cm$^2$/MRad and a

352: significant shift in the depletion voltage, dependent on the

353: initial dopant concentration, were observed.  A shift of about

354: 8--10\volt\ was seen for irradiation corresponding to a dose of

355: approximately 1~MRad. These observations indicate significant bulk

356: damage caused by energetic electrons. As indicated by the change

357: in depletion voltage, the SVT sensors could undergo type inversion

358: after about 1--3~MRad. Preliminary tests show that the sensors

359: continue to function after type inversion~\cite{rad_damage}.

360: Studies of the behavior of SVT modules as a function of radiation

361: dose continue.

362: 

363: \subsubsection{Fanout Circuits}

364: 

365: The fanout circuits, which route the signals from the strips to the

366: electronics, have been designed to minimize the series resistance and

367: the inter-strip capacitance.  As described in ref.~\cite{fanouts},

368: a trace on the fanout has a series resistance about 1.6~\ohm/cm,

369: an inter-strip resistance $>20$\Mohm, and an inter-strip

370: capacitance $<0.5$\pf/cm.  The electrical parameters of the final

371: assembly of sensors and fanouts (referred to as  \emph{Detector Fanout

372: Assemblies} or DFAs) are summarized in Table~\ref{\secname tab:electrical}. Due

373: to the different strip lengths, there are large differences between

374: the inner and the outer layers.  Smaller differences are also present

375: between the forward and backward halves of the module, that are of

376: different lengths.

377: 

378: \subsubsection{Front End Electronics}

379: 

380: The electrical parameters of a DFA and the general \babar\

381: requirements are the basic inputs that drove the design of the SVT

382: front-end custom IC; the ATOM (\emph{A Time-Over-Threshold

383: Machine}).  In particular, the front-end IC had to satisfy the

384: following requirements:

385: 

386: \begin{itemize}

387: 

388: \item signal to noise ratio greater than 15 for \emph{minimum ionizing

389: particle} (MIP) signals for all modules;

390: 

391: \item signals from all strips must be retained, in

392: order to improve the spatial resolution through interpolation,

393: while keeping the number of transmitted hits as low as possible. A

394: \emph{hit} refers to a deposited charge greater than $ 0.95$\fC,

395: corresponding to 0.25 MIP;

396: 

397: \item the amplifier must be sensitive to both negative and positive

398: charge;

399: 

400: \item the peaking time must be programmable, with a minimum of 100\ns\

401:   (in layers 1 and 2, because of the high occupancy), up to 400\ns\

402:   (outer layers, with high capacitance);

403: 

404: \item capability to accept random triggers with a latency up to 11.5\mus\ and

405: a programmable jitter up to $\pm 1$\mus, without dead

406:   time;

407: 

408: \item radiation hardness greater than 2.5~MRad;

409: 

410: \item small dimensions: 128 channels in a 6.2\mm-wide chip.

411: 

412: \end{itemize}

413: 

414: These requirements are fully satisfied by the ATOM IC~\cite{atom},

415: which is depicted schematically in Figure~\ref{\secname fig:atom}.

416: 

417: \begin{figure}[htb]

418: \centering

419: \includegraphics[width=7.5cm]{fig_atom}

420: \vspace{-2pc}

421: \caption{Schematic diagram of the ATOM front end

422: IC.} \label{\secname fig:atom}

423: \end{figure}

424: 

425: The linear analog section consists of a charge-sensitive

426: preamplifier followed by a shaper. Gains of 200\mv/fC (low) or

427: 300\mv/fC (high) may be selected. The channel gains on a IC are

428: uniform to 5\mv/fC.  Signals are presented to a

429: programmable-threshold comparator, designed so that the output

430: width of the pulse (\emph{Time over Threshold} or ToT) is a

431: quasi-logarithmic function of the collected charge. This output is

432: sampled at 30\mhz\ and stored in a 193 location buffer.  Upon

433: receipt of a \emph{Level 1} (L1) trigger, the time and ToT is

434: retrieved from this latency buffer, sparcified, and stored in a

435: four event buffer. Upon the receipt of an \emph{L1 Accept} command

436: from the data acquisition system, the output data (the 4 bits for

437: the ToT, 5 bits for the time stamp, and 7 bits for the strip

438: address) are formatted, serialized, and delivered to the ROM.  The

439: IC also contains a test charge injection circuit.  The typical

440: noise behavior of the ATOM, as described by the \emph{Equivalent

441: Noise Charge} (ENC) of the linear analog section is given in

442: Table~\ref{\secname tab:atom}.

443: 

444: \begin{table}

445: \caption{ATOM chip ENC parameters at different

446:  peaking times}

447: \vspace{\baselineskip}

448: \begin{center}

449: \begin{tabular}{ccc} \hline

450: Peaking\rule{0pt}{12pt}    &ENC   &Noise  \\

451: time\rule[-5pt]{0pt}{0pt}  &(0\pf)  & slope \\   \hline

452: \rule{0pt}{12pt}100\ns     &380 \en &40.9 \en /pF\\

453:    200\ns  &280 \en   &33.9 \en /pF \\

454: 400\ns\rule[-5pt]{0pt}{0pt}     &220 \en   &25.4\en /pF\\ \hline

455: \end{tabular}

456: \end{center}

457: \label{\secname tab:atom}

458: \end{table}

459: 

460: 

461: The average noise for the various module types is shown in

462: Table~\ref{\secname tab:electrical}. Given that shot noise due to

463: sensor leakage current is negligible, the expected noise may be

464: calculated from the parameters of Tables~\ref{\secname

465: tab:electrical} and \ref{\secname tab:atom}. The results of such a

466: calculation are also shown in Table~\ref{\secname tab:electrical}.

467: The maximum average noise is 1,600 electrons, leading to a

468: signal-to-noise ratio greater than 15.

469: 

470: The power consumption of the IC is about 4.5\mw/channel. Radiation

471: hardness was studied up to 2.4~MRad with a $^{60}$Co source. At

472: that dose, the gain decreased 20\%, and the noise increased less

473: than 15\%.

474: 

475: The ATOM ICs are mounted on thick-film double-sided hybrid

476: circuits (known as \emph{High Density Interconnects} or HDIs)

477: based on an  aluminum-nitride substrate with high thermal

478: conductivity. The electronics are powered through a floating power

479: supply system, in such a way as to guarantee a small voltage drop

480: ($<1$\volt) across the detector decoupling capacitors.

481: 

482: \subsubsection{Data Transmission}

483: 

484: The digitized signals are transmitted from the ATOM ICs through a

485: thin \emph{kapton tail} or cable to the \emph{matching cards},

486: from where they are routed to more conventional cables.  Just

487: outside the detector, signals are \emph{multiplexed} by the MUX

488: modules, converted into optical signals and transmitted to the

489: \emph{Readout Modules} (ROMs). The MUX modules also receive

490: digital signals from the DAQ via a fiber optical connection.  The

491: SVT is connected to the \babar\ online detector control and

492: monitoring system via the industry standard CAN bus.  Details on

493: SVT data transmission system and DAQ can be found in

494: references~\cite{dat,daq}. Power to SVT modules (silicon sensor

495: bias voltage and ATOM low voltages) is provided by a CAEN A522

496: power supply system~\cite{power_supply}.

497: 

498: 

499: \subsection{Monitoring and Calibration}

500: \label{sec:svtmonitoring}

501: 

502: To identify immediately any operational problems, the SVT is

503: integrated in the  control and monitoring system

504: (see Section~\ref{sec:online}).  Major

505: concerns for SVT monitoring are temperature and humidity, mechanical

506: position, and radiation dose.

507: 

508: \subsubsection{Temperature\\ and Humidity Monitors}

509: The total power dissipation of the SVT modules is about 350\watt,

510: mainly dissipated in the ATOM ICs. External cooling is provided by

511: chilled water at 8\degc. In addition, humidity is reduced by a

512: stream of dry air in the support tube.

513: 

514: Since condensation or excessive temperature can permanently

515: damage the FEE, temperature and humidity monitoring are very

516: important to the safe operation of the SVT.  Thermistors are

517: located on the HDIs (for the measurements of FEE temperature),

518: around the SVT, along the cooling systems, and in the electronics

519: (MUX) crates.  The absolute temperatures are monitored to

520: 0.2\degc\ and relative changes of 0.1\degc. Additionally, a series

521: of humidity sensors are employed to monitor the performance of the

522: dry air system.  The temperature and humidity monitors also serve

523: as an interlock to the HDI power supplies.

524: 

525: \subsubsection{Position Monitors}

526: A system of capacitive sensors was installed to identify and track

527: changes in the position of the SVT with respect to the PEP-II B1

528: magnets and the position of the support tube with respect to the

529: DCH. An example of the understanding that can be achieved by this

530: system is given in Figure~\ref{\secname fig:pos}, where the

531: measured changes in the horizontal position of the SVT relative to

532: the DCH are shown for a period of six day in the summer of 1999.

533: These position changes can be attributed to local temperature

534: variations.  The sensor data are compared to measurements of the

535: mean position of the interaction point (in the horizontal plane)

536: determined with $e^+e^-$ and $\mu^+\mu^-$ events recorded over

537: this period.  While the amplitude of motion at the time was

538: uncharacteristically large, the strong correlation between these

539: independent measurements is quite evident.  Alignment with charged

540: particle tracks is now performed routinely to correct for relative

541: motion of the tracking systems, as described in Section 5.6.2.

542: 

543: \begin{figure}

544: \centering

545: \includegraphics[width=.9\columnwidth]{8583A55}

546: \vspace{-2pc}

547: \caption{Horizontal motion between the DCH and the support tube

548: measured with the capacitive sensors (curve) compared to the mean

549: $x$ coordinate of the interaction point (circles) measured with

550: $e^+e^-$ and $\mu^+\mu^-$ events over a three-day period in July

551: 1999. An arbitrary offset and scale has been applied to the beam

552: position data.} \label{\secname fig:pos}

553: \end{figure}

554: 

555: 

556: \subsubsection{Radiation Monitors}

557: 

558: Radiation monitoring is extremely important to ensure the SVT does not

559: exceed its radiation budget, which could cause permanent damage to the

560: device. To date, the measured radiation absorbed by the SVT is well

561: within the allowed budget.

562: 

563: The monitoring of radiation dose to the SVT is discussed in detail in

564: Section~\ref{sec:ir}.

565: 

566: \subsubsection{Calibrations}

567: 

568: Once a day, and each time the SVT configuration has changed,

569: calibrations are performed in absence of circulating beams. All

570: electronic channels are tested with pulses through test

571: capacitors, for different values of the injected charge.  Gains,

572: thresholds, and electronic noise are measured, and defective

573: channels are identified. The calibration results have proven very

574: stable and repeatable. The main variation in time is the

575: occasional discovery of a new defective channel. The calibration

576: procedures have also been very useful for monitoring noise sources

577: external to the SVT.

578: 

579: \subsubsection{Defects}

580: 

581: Due to a series of minor mishaps incurred during the installation of

582: the SVT, nine out of 208 readout sections (each corresponding

583: to one side of a half-module) were damaged and are currently not

584: functioning. There is no single failure mode, but several

585: causes: defective connectors, mishandling during installation, and

586: not-fully-understood problems on the FEE

587: hybrid. There has been no module failure due to radiation damage.

588: It should be noted that

589: due to the redundancy afforded by the five layers of the SVT, the

590: presence of the defective modules has minimal impact on physics analyses.

591: 

592: In addition, there are individual channel

593: defects, of various types, at a level of about 1\%. Calibrations

594: have revealed an increase in the number of defective channels at a rate of

595: less than 0.2\%/year.

596: 

597: \subsection{Data Analysis and Performance}

598: 

599: This section describes the reconstruction of space points from

600: signals in adjacent strips on both sides of the sensors, the SVT

601: internal and global alignment, single hit efficiency, and

602: resolution  and \dedx\ performance of the SVT.

603: 

604: 

605: \subsubsection{Cluster and Hit Reconstruction}

606: Under normal running conditions, the average occupancy of the SVT

607: in a time window of 1\mus\ is about 3\% for the inner layers, with a

608: significant azimuthal variation due to beam-induced backgrounds, and

609: less than 1\% for the outer layers, where noise hits dominate.

610: Figure~\ref{\secname fig:occupancy} shows the typical occupancy as a

611: function of IC index (equivalent to azimuthal angle, in this case)

612: for layer 1, \mphi\ side. In the inner layers, the occupancy is

613: dominated by machine backgrounds, which are significantly higher in

614: the horizontal plane, seen in the plot as the peaks near IC indices

615: 3 and 25.

616: 

617: \begin{figure}[htb]

618: \centering

619: \includegraphics[width=6.5cm]{8583A39}

620: \vspace{-1.5pc}

621: 

622: \caption{Typical occupancy in percent as a function of IC index in layer

623: 1, \mphi\ side for a) forward half-modules and b) backward

624: half-modules. The IC index increases with azimuthal angle and

625: the higher occupancy in the horizontal plane is visible near

626: chip indices 3 and 25.}

627: \label{\secname fig:occupancy}

628: \end{figure}

629: 

630: The first step of the reconstruction program consists in

631: discarding out-of-time channels. A time correction, \ie\ the time

632: between the passage of the particle and the time the shaper

633: exceeds threshold, is performed, after which hits with times more

634: than 200\ns\ from the event time (determined by the DCH) are

635: discarded. The loss of real hits from this procedure is

636: negligible.  The resulting in-time hits are then passed to the

637: cluster finding algorithm.  First, the charge pulse height (Q) of

638: a single pulse is calculated from the ToT value, and clusters are

639: formed grouping adjacent strips with consistent times. In a second

640: pass, clusters separated by just one strip are merged into one

641: cluster.  The two original clusters plus the merged cluster are

642: made available to the pattern recognition algorithm, which chooses

643: among the three.

644: 

645: The position $x$ of a cluster formed by $n$ strips is determined,

646: with the ``head-to-tail'' algorithm:

647: \begin{displaymath}

648:    x = \frac{(x_1+x_n)}{2} + \frac{p}{2}

649:    \frac{(Q_n-Q_1)}{(Q_n+Q_1)},

650: \end{displaymath}

651: where $x_i$ and $Q_i$ are the position and collected charge of $i$-th

652: strip, respectively, and $p$ is the readout pitch.  This formula

653: results in a cluster position that is always within $p/2$ of the

654: geometrical center of the cluster. The cluster pulse height is simply

655: the sum of the strip charges, while the cluster time is the average of

656: the signal times.

657: 

658: \subsubsection{Alignment}

659: 

660: The alignment of the SVT is performed in two steps. The first step

661: consists of determining the relative positions of the 340 silicon

662: sensors. Once this is accomplished, the next step is to align the

663: SVT as a whole within the global coordinate system defined by the

664: DCH. The primary reason for breaking the alignment procedure into

665: these two steps is that the local positions are relatively stable

666: in time compared to the global position. Also, the local alignment

667: procedure is considerably more complex than the global alignment

668: procedure. Thus, the global alignment can be updated on a

669: run-by-run basis, while the local alignment constants are changed

670: as needed, typically after magnet quenches or detector access.

671: 

672: The local alignment procedure is performed with tracks from

673: $e^+e^- \rightarrow \mu^+\mu^-$ events and cosmic rays. Well

674: isolated, high momentum tracks from hadronic events are also used

675: to supplement di-muon and cosmic data. Data samples sufficient to

676: perform the local alignment are collected in one to two days of

677: typical running conditions.

678: 

679: In $\mu^+\mu^-$  events, the two tracks are simultaneously fit

680: using a Kalman filter technique and the known beam momentum. The

681: use of tracks from cosmic rays reduces any systematic distortion

682: that may be introduced  due to imprecise knowledge of the beam

683: momenta. No information from the DCH is used, effectively

684: decoupling the SVT and DCH alignment.

685: 

686: In addition to the information from tracks, data from an optical

687: survey performed during the assembly of the SVT are included in

688: the alignment procedure. The typical precision of these optical

689: measurements is 4\mum. This survey information is only used to

690: constrain sensors relative to other sensors in the same module,

691: but not one module to another or one layer to another.

692: Furthermore, only degrees of freedom in the plane of the sensor

693: are constrained as they are expected to be the most stable, given

694: the assembly procedure.

695: 

696: Using the hit residuals from the aforementioned set of tracks and

697: the optical survey information, a $\chi^2$ is formed for each

698: sensor and minimized with respect to the sensor's six local

699: parameters.  The constraints coming from the overlapping regions

700: of the silicon sensors, the di-muon fit, the cosmic rays, and the

701: optical survey result in internally consistent local alignment

702: constants.

703: 

704: \begin{figure}[htb]

705: \centering

706: \includegraphics[width=7.5cm]{SvtAlign_1}

707: \vspace{-2pc}

708:  \caption{Comparison of a local alignment of all the

709: sensors in the SVT using data from January 2000 with the optical

710: survey of the SVT made during assembly in February 1999 in the (a)

711: $r\Delta\phi$, (b) $\Delta z$ and (c) $\Delta r$ coordinates.

712: Plots (d), (e), and (f) show the difference between two local

713: alignments using data from January 15-19 and March 6-7, 2000 for

714: the  $r\Delta\phi$, $\Delta z$, and $\Delta r$ coordinates,

715: respectively. In all the plots, the shaded regions correspond to

716: the sensors in the first three layers. In comparing the different

717: alignments and optical survey, a six parameter fit (three global

718: translations and three global rotations) has been applied between

719: the data sets.} \label{\secname fig:SvtAlign1}

720: \end{figure}

721: 

722: Figure \ref{\secname fig:SvtAlign1} shows a comparison between the

723: optical alignment made during the SVT assembly in February 1999

724: and a local alignment using data taken during January 2000.  The

725: alignment from tracking data was made without using cosmics or

726: constraints from the optical survey. The width of the

727: distributions in these  plots has four contributions: 1)

728: displacement during the transfer of the SVT from the assembly jig

729: to the IR magnets,  2) time dependent motion of the SVT after

730: mounting, 3) statistical errors, and 4) systematic errors. The

731: second set of plots shows the difference in two alignment sets for

732: data taken in January 2000 as compared to March 2000.  In general,

733: the stability of the inner three layers is excellent, with

734: slightly larger tails in the outer two layers. The radial

735: coordinate is less tightly constrained in all measurements because

736: the radial location of the charge deposition is not well known,

737: and most of the information about the radial locations comes only

738: from constraints in the overlap region of the sensors.

739: 

740: 

741: 

742: After the internal alignment, the SVT is considered as a rigid

743: body. The second alignment step consists in determining the

744: position of the SVT with respect to the DCH. Tracks with

745: sufficient numbers of SVT and DCH hits are fit two times: once

746: using only the DCH information and again using only the SVT hits.

747: The six global alignment parameters, three translations and three

748: rotations, are determined by minimizing the difference between

749: track parameters obtained with the SVT-only and the DCH-only fits.

750: As reported above, because of the diurnal movement of the SVT with

751: respect to the DCH, this global alignment needs to be performed

752: once per run ($\sim$ every 2--3 hours). The alignment constants

753: obtained in a given run are then used to reconstruct the data in

754: the subsequent run.  This procedure, known as \emph{rolling

755: calibration}, ensures that track reconstruction is always performed

756: with up-to-date global alignment constants.

757: 

758: 

759: 

760: A record of the changes in the relative position of the SVT as

761: determined by {\emph rolling calibrations} is shown in

762: Figure~\ref{\secname fig:8583A53}.  The position is stable to

763: better than $\pm 100 \mum$ over several weeks, but changes

764: abruptly from time to time, for instance, during access to the

765: detector. The calibrations track diurnal variations of typically

766: $\pm 50 \mum$ that have been correlated with local changes in

767: temperature of about $\pm 2 \degc$. Movements within a single run

768: are small compared to the size of the beam.

769: 

770: \begin{figure}

771: \centering

772: \includegraphics[scale=.8]{8583A54}

773: 

774: \caption{ Global alignment of the SVT relative to the DCH based on

775: $e^+e^-$ and $\mu^+\mu^-$ events: changes in the relative vertical

776: placement measured a) over the entire ten-month run in the year

777: 2000, and b) a ten-day period, illustrating diurnal variations.  }

778: 

779: \label{\secname fig:8583A53}

780: \end{figure}

781: 

782: 

783: 

784: \subsubsection{Performance}

785: The SVT  efficiency can be calculated for each half-module by

786: comparing the number of associated hits to the number of tracks

787: crossing the active area of the  module. As can be seen in

788: Figure~\ref{\secname fig:efficiency}, a combined hardware and

789: software efficiencies of 97\% is measured, excluding defective

790: readout sections (9 out of 208), but employing no special

791: treatment for other defects, such as broken AC coupling capacitors

792: or dead channels on front-end chips.

793: %

794: \begin{figure}[htb]

795: \centering

796: \includegraphics[width=7.5cm]{eff_both_14558}

797: \vspace{-2pc} \caption{SVT hit reconstruction efficiency, as

798: measured on $\mu^+\mu^-$ events for a) forward half-modules and b)

799: backward half-modules.  The plots show the probability of

800: associating both a $\phi$ and $z$ hit to a track passing through

801: the active part of the detector. The horizontal axis corresponds

802: to the different modules, with the vertical lines separating the

803: different layers as numbered.  Missing values correspond to

804: non-functioning half-modules.} \label{\secname fig:efficiency}

805: \end{figure}

806: %

807: Actually, since most of the defects affect a single channel, they

808: do not contribute to the inefficiency, because most tracks deposit

809: charge in two or more strips due to track crossing angles, and

810: charge diffusion.

811: 

812: The spatial resolution of SVT hits is determined by measuring the

813: distance (in the plane of the sensor) between the track trajectory and the

814: hit, using high-momentum tracks in two prong events.  The uncertainty

815: due to the track trajectory is subtracted from the width of the

816: residual distribution to obtain the hit resolution.

817: Figure~\ref{\secname fig:resolution} shows the SVT hit resolution for

818: $z$  and $\phi$ side hits as a function of track incident angle, for each of the

819: five layers. The measured resolutions are in excellent agreement

820: with expectations from Monte Carlo simulations.

821: 

822: 

823: \begin{figure}[htb]

824: \includegraphics[width=7.5cm]{res_zedphi_14558}

825: %\includegraphics[width=7.5cm,height=7.0cm]{res_phi_14558}

826: 

827: \vspace{-2pc} \caption{SVT hit resolution in the a) $z$ and b)

828: $\phi$ coordinate in microns, plotted as a function of track

829: incident angle in degrees. Each plot shows a different layer of

830: the SVT. The plots in the $\phi$ coordinate for layers 1-3 are

831: asymmetric around $\phi=0$ because of the ``pinwheel'' design of

832: the inner layers. There are fewer points in the $\phi$ resolution

833: plots for the outer layers as they subtend smaller angles than the

834: inner layers. } \label{\secname fig:resolution}

835: \end{figure}

836: 

837: 

838: Initial studies have shown that hit reconstruction efficiency and

839: spatial resolution are effectively independent of occupancy for the

840: occupancy levels observed so far.

841: 

842: 

843: Measurement of the ToT value by the ATOM ICs enables one to obtain

844: the pulse height, and hence the ionization \dedx\ in the SVT

845: sensor. The values of ToT are converted to pulse height using a

846: lookup table computed from the pulse shapes obtained in the bench

847: measurements. The pulse height is corrected for track length

848: variation. The double-sided sensors provide up to ten measurements

849: of \dedx\ per track. For every track with signals from at least

850: four sensors in the SVT, a 60\% truncated mean \dedx\ is

851: calculated. The cluster with the smallest \dedx\ energy is also

852: removed to reduce sensitivity to electronics noise. For MIPs, the

853: resolution  on the truncated mean \dedx\ is approximately 14\%. A

854: $2\sigma$ separation between the kaons and pions can be achieved

855: up to momentum of 500\mevc, and between kaons and protons beyond

856: 1\gevc.

857: 

858: \subsection{Summary and Outlook}

859: The SVT has been operating efficiently since its installation in

860: the \babar\ experiment in May 1999. The five layer device, based

861: on double-sided silicon sensors, has satisfied the original design

862: goals, in particular the targets specified for efficiency, hit

863: resolution, and low transverse momentum track reconstruction. The

864: radiation dose during the first 25\invfb\ of integrated luminosity

865: is within the planned budget, and  no modules have failed due to

866: radiation damage. The performance of the SVT modules at high

867: radiation dose is currently being studied. Early results indicate

868: that the sensors will continue to function after type inversion

869: (at 1--3~MRad), but further tests with irradiated sensors and ATOM

870: ICs need to be performed. A program of spare module production has

871: commenced, with the goal of replacing modules that are expected to

872: fail due to radiation damage. Beam-generated backgrounds are

873: expected to rise with increasing luminosity.  Physics studies at

874: five times the current backgrounds levels indicate no change in

875: mass or vertex resolution for the mode \bpsiks\ and a $\sim 20$\%

876: loss of resolution in the $D^{*+}-D^0$ mass difference. In this

877: study the detector efficiency for both decay modes was lower by

878: 15--20\%.

879: