主题：关于CTD的论述

Advanced Array Detectors
In the 1960s, solid-state devices were introduced into the electronics industry. These
devices, such as transistors and diodes, were based on the properties of silicon but
were relegated to research and aerospace applications that could afford the
relatively high cost of these components. As their use expanded to the digital
electronics industry in the form of integrated circuits (ICs), not only did the cost of
the devices become affordable but the cost of systems using the ICs such as digital
computers were drastically reduced. Today, a small personal calculator can be
purchased for less than ten dollars.
It was also discovered that silicon-based sensors responded to light and were
quickly integrated into linear and two-dimensional arrays called solid-state imagers
or detectors. The commercial need for an affordable television camera drove the
market into reducing the cost of the technology. However, since the television
camera did not require the sensitivity and resolution that many scientific applications
demand, there was a need for higher-quality solid-state detectors.
Consequently, three generic, advanced solid-state detectors with high sensitivity
and resolution for spectroscopic applications have been developed - the photodiode
array (PDA), the charge-injection device (CID) and the charge-coupled device
(CCD). The CID and CCD devices are based on the light-sensitive properties of
solid-state silicon and belong to the broad class of silicon-based devices called
charge transfer devices (CTD).
To illustrate the principals associated with CTDs, a block of very high purity
crystalline silicon is considered (Figure 3-26). Onto this silicon substrate is grown
an insulating layer of silicon dioxide (SiO2). As shown in the pictorial Figure 3-27,
each silicon atom in the substrate is bonded to its adjacent silicon atom in a three
dimensional lattice. The silicon-silicon bond may be broken by energy of sufficient
strength such as photons with visible or ultraviolet wavelength. When the bond is
broken, an electron is released within the lattice structure and a subsequent hole in
the crystalline structure is formed. This is called an electron-hole pair.
If a voltage is applied across the block of silicon (Figures 3-26 and 3-27), the freed
electrons will move in the opposite direction of the applied electric field or towardthe silicon-silicon dioxide interface while the holes will move in the other direction
or in the same direction as the electric field and leave a region depleted of positive
charge. This electron and hole motion within the crystalline lattice creates a current
which is proportional to the amount of photons impinging on the structure. That is,
the more light absorbed by the silicon, the more electrons are captured at the siliconsilicon
oxide interface.
The CTDs elements, called pixels, may vary in size from 6 to 30 microns and
arranged generally in a two-dimensional silicon wafer configuration from 512 x 512
to 4096 x 4096 pixels. Each of these pixels is capable of storing photon- generated
charge. It is the manner in which the charge is obtained, interrogated and stored
that distinguish the various types of devices.
In general, each pixel of the two-dimensional Charge Injection Devices (CIDs) may
be randomly interrogated to determine the amount of charge that has been accumulated
during a measured time to which the device has been exposed to light
(called the integration time). With the advent of high speed microprocessors,
individual pixels may be examined even during the integration time to determine the
accumulated charge. This process of examining the contents does not destroy the
contents and, hence, is known as a non-destructive read-out mode. However, even
though the CID has a random access and non-destructive read-out, it has an
inherently higher noise level or dark current than, for example, a CCD, and requires
cooling to liquid nitrogen temperatures to effectively decrease this noise. The dark
current of any device is the electronic current that flows in a detector when operating
voltages are applied but no light is present.
An ICP spectrometer with a CID detector is commercially available. It utilizes an
echelle design and a CID detector with over 250,000 pixels capable of detecting
ICP spectral lines across a large wavelength region. The technique produces a
"snap shot" of all detectable lines very similar to that observed on a photographic
plate in optical emission spectroscopy. Such analytical improvements as background
corrections and spectral interference corrections may be accomplished in a
post processing mode rather than at the time of analysis.
The charge accumulated on the pixels of a Charge Coupled Device (CCD) must be
read sequentially and, in the process of reading the charge, it is destroyed. Hence,
the CCD is known to have a sequential, destructive read-out system. The manner
in which the CCD read-out system operates has been compared to "tiny" buckets
corresponding to the individual pixels of the CCD. The buckets collect the electrons
which have been triggered by the photons impinging on the silicon substrate. At the
end of the exposure, the pixels are commanded to "empty their buckets". The
electrons in the buckets are emptied one row at a time into bucket holders so that,
at the read-out electronics, there is a single line of buckets that are tagged with theirposition in the original array and the content of each bucket is measured. It is easily
understood that this bucket brigade operation is a sequential, destructive process
whereby the content of an individual pixel cannot be re-examined after the read-out
process begins. However, the CCD read-out is extremely fast and has very low noise
or dark current associated with the process. Thus, it is possible to cool CCDs by
thermoelectric devices such as Peltier units. Furthermore, since most commercial
television cameras are CCDs, there is considerable continuing research and
development devoted to these types of devices.