The following text was automatically extracted from the image on this page using optical character recognition software:
580 mm Space Frame
Bkwd.
support
520 mrad
~ ~ Q I n II _ Fwd. support350m
cone
Front end e +
electronics
Beam Pipe
Fig. 1. Longitudinal cross section of the SVT.
rise over the next five years. Considerable effort have gone into
understanding the future performance and radiation limitations
of the existing system and whether any replacement will be
needed. We will report on these studies here.
II. THE BABAR SILICON VERTEX TRACKER
The SVT has two tasks. First, B and D decay vertices have to
be accurately reconstructed in order to measure the difference in
decay length, Az, between the two B mesons. With the design
Az resolution of 130 pm, the precision of a CP violation mea-
surement is only 10% worse than with a perfect Az measure-
ment. Achieving this Az resolution requires a single vertex res-
olution better than 80 pm. Second, low transverse-momentum
(PT < 120 MeV/c) particles, such as pions from D* decays, do
not fully traverse the main BaBar tracking chamber and there-
fore have to be reconstructed in the SVT alone.
The SVT design [2], [3] accomplishes these goals by having
five layers of double-sided, ac-coupled silicon micro-strip de-
tectors in a concentric configuration around the beam pipe as
shown in Fig. 1. Each layer is built from 6 to 18 modules, where
each module itself consists of several 300 p1 thick silicon de-
tectors glued together. The modules are mechanically supported
by Kevlar/carbon-fiber along the modules, while the detector
signals are carried by flexible fanout circuits to the front-end
electronics located at both ends of each module. The readout
of a single module is divided into four so-called readout sec-
tions. The strips on the two sides of a modules are perpendicular
and provide a z and r-o5 measurement. The three inner layers
are planar, located very close to the interaction point and have
a readout pitch of 50-110 /cm. This provides the resolution in
track angle and impact parameter measurements needed to get
a precise vertex determination. The two outer layers are of most
importance to the track finding, particularly for low momentum
particles. The silicon detectors at the end of layer 4 and 5 mod-
ules are at an angle to the beam pipe in order to provide the
largest possible angular acceptance, while reducing the amount
of material traversed by low-angle tracks. The polar angle ac-
ceptance is 20 < 0 < 150 and is mainly constrained by the
design of the PEP-II interaction region.
More than 150 000 silicon strips are read out using a fully
custom-designed chip, known as AToM [4], located at both ends
and sides of each module. Each AToM chip is capable of si-
multaneous acquisition, digitization, and sparsified readout of
128 channels. The linear analog section consists of a charge sen-
sitive preamplifier, followed by a shaper. The output signal is
compared to a programmable threshold producing a logic pulse
whose width [time over threshold (ToT)] is approximately pro-
portional to the logarithm of the collected charge. The logic
pulse is digitized at 15 MHz and stored into a latency buffer.
Upon a level-1 trigger signal, a 1 ps window in the buffer is
searched for the first hit. For channels with a hit, the ToT and hit
time in the buffer is transmitted to the data acquisition system.
III. PERFORMANCE AND ALIGNMENT
Of the 208 readout sections in the SVT, nine failed during
or right after their installation in 1999. During a shutdown in
2002, four of the failures were identified as slipped connectors
and fixed. No readout section has failed during running due to
radiation damage or otherwise.
The hit efficiency is measured by counting hits on tracks
passing through the active area of the silicon sensors. Excluding
the five broken readout sections, the average hit efficiency is
97%, which includes inefficiencies due to broken capacitors and
dead channels in the AToM chips. The SVT hit resolution is
measured using high momentum tracks in two-prong events. For
tracks with a 90 incidence angle to the silicon modules, the hit
resolution is 10-15 m in the inner three layers and 30-40 pm
for the outer two layers. With this, the SVT achieves its design
of an average vertex resolution of better than 80 pm. The Az
resolution is 190 m, dominated by uncertainties in inclusively
reconstructing one of the two B vertices.
Tracking efficiencies for low momentum particles, where the
SVT contribution is essential, have been studied in data using
slow pions from D* decays. These studies give an tracking ef-
ficiency above 75% for momentum above 100 MeV/c and above
90% for momentum above 200 MeV/c.
The deposited charge measured by the ToT mechanism gives
a reasonable ionization loss measurement, dE/dx, when aver-
aged over hits in all five layers. Using a 60% truncated mean,
the dE/dx resolution on minimum ionizing particles is approx-
imately 14%. For low momentum tracks this can be used for
particle identification. We achieve better than 26- separation be-
tween kaons and pions with momentum below 500 MeV/c and
between kaons and protons below 1 GeV/c.
To achieve the SVT performance given above, the SVT needs
to be carefully aligned. In BaBar this involves two separate pro-
cedures. First, the internal alignment of the SVT and second, the
rad
