Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.sao.ru/drabek/CCDP/TAPER/HTM_resume/PRINSTON/Cameras/ccdcams.pdf Äàòà èçìåíåíèÿ: Fri Sep 11 02:25:30 1998 Äàòà èíäåêñèðîâàíèÿ: Sat Sep 11 22:05:09 2010 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: guide 8.0

Advantages of Slow Scan CCD Cameras
Professional quality scientific cameras are available in many forms, ranging from what are essentially modified surveillance cameras to high speed cameras to slow scan high sensitivity cameras. These can be used with "frame grabber " video boards or other, more sophisticated hardware to provide computer-based image manipulation capabilities. The superior capabilities of scientific grade slow scan CCD cameras are outlined below.
CCD Camera Categories
There are many excellent cameras based on charge coupled devices (CCDs) on the market today, which are applicable to scientific data gathering of one type or another. Indeed, the advantages of CCDs over older camera technologies are so extreme that even consumer grade CCD cameras can be used in some experiments. Scientific grade slow scan CCD cameras are distinguished from other imaging devices by their superior capabilities in the following parameters: u u u u u u u u u u Sensitivity & Quantum Efficiency Low Noise Low Dark Charge High Baseline Stability Linearity Spatial Resolution Intensity Resolution & Precision Dynamic Range Spectral Range Flexible Scan Pattern Definition Princeton Instruments thermoelectrically cooled CCD camera. Temperatures as low as -45°C are achieved with circulating air used to dissipate heat. eye. Images produced by these cameras will be perceived as more "lifelike" by the consumer. These devices are fabricated with partially transparent polysilicon structures covering the imaging area, virtually eliminating the response in the far blue and ultraviolet regions of the spectrum. The remainder of this section will compare the properties of scientific slow scan CCD cameras with high performance video-rate cameras in more detail. In the following comparison, the use of a standard 8 bit resolution frame grabber is assumed for the video rate camera. Princeton Instruments

CCD Cameras

Some of these characteristics, such as sensitivity, linearity, and dynamic range, are a direct result of the slow scan operation. There therefore exists a trade-off between scanning speed (readout rate) and performance in these areas. Often the trade-off is tied to fundamental physical principles; in other cases it is a limitation of the present state of the art of CCD technology. Other characteristics, such as spectral range, are not directly dependent on the scan rate, but are linked to the CCD designs by market forces. For example, low cost, small aperture CCDs for consumer use are designed to have a limited spectral response to more closely mimic the human Fax: 609-587-1970

Sensitivity and Quantum Efficiency
Sensitivity is possibly the most important reason to choose a slow scan camera. Many scientific applications are light starved. In many astronomical experiments, the only way to increase the light flux is to increase the telescope size, a very expensive proposition. For some biomedical applications it is not useful to increase the illumination of the sample because the light changes the biochemistry under examination. Of course, for a given sensitivity 11


the light flux (photons/unit area-sec) can sometimes be increased by extending the exposure time, increasing the signal proportionally. If a time exposure approach is used the frame rate is automatically reduced. Unfortunately, not all the photons that reach the CCD interact to generate electron-hole pairs in the CCD. The fraction which does interact is the quantum efficiency (QE). CCDs that use reflective or non-responsive structures to provide convenience features such as electronic shuttering or anti-blooming, have even lower QE and require even more light to achieve the same signal to noise ratio as a scientific camera. The impact can be quite pronounced, as shown below. Consumer grade CCDs often have QEs as low as 12% at their peak response wavelength (see Fig. 1), while back illuminated scientific CCDs have peak QEs of up to 80%, a sensitivity gain of more than 3:1 from the QE alone. In the blue and UV regions of the spectrum, which are very important to many scientific fields (biochemistry, combustion dynamics, etc.) the difference in QE can be over 100 times. The readout noise of a CCD device, an important factor in the overall sensitivity, is normally proportional to the square root of the readout rate. Therefore operation at slow scan rates reduces readout noise, often quite substantially. Princeton Instruments slow scan cameras have readout noise levels as low as 4 electrons/pixel. Consequently, over most of their dynamic range, the signal to noise ratio achieved with such a camera is limited by the statistical fluctuations of the incident light (photon shot noise), not by the readout noise of the camera. This of course assumes that the readout noise of the camera is small. Of course, there are limits to the increase in sensitivity achieved with a longer exposure. The primary limit is the "dark charge", the charge generated not by incident light but by leakage currents within the CCD. Since this charge accumulates even in the dark it is called dark charge and the leakage current generating it is called "dark current". Exposure 1/30 sec 1 sec 1 sec 1 minute 10 minutes

The lower the CCD temperature, the lower the dark current. For this reason, scientific slow scan cameras almost always cool the CCD, using either thermoelectric coolers (TE) or liquid nitrogen (LN). In addition, a new CCD technology called multi-pin phasing (MPP) can reduce dark current by a factor of 100 or more, by reducing the potential at the surface of the CCD during the exposure time. Both techniques are available (separately and in combination) in PI scientific cameras. With these techniques, exposure times of hours are feasible -- the practical limit set only by the fluence of cosmic rays, which produce signals in CCD cameras just as light does. Table 1 compares the scene illuminance required by a slow scan scientific and a video rate camera, for various signal gathering intervals. Note that to take advantage of the longer light gathering intervals, the video rate camera must be used with a frame grabber which includes the capability of frame averaging (or accumulation), which will increase the system cost dramatically. In preparing the table, the following assumptions have been used: Slow Scan Scientific Cooled Camera: 27 µm x 27 µm pixel, QE = 70%, 4 electrons readout noise, cooled Video Rate Camera: 11 µm x 13 µ m pixel, QE = 25%, 50 electrons readout noise, cooling irrelevant Other Parameters: Lens F/1.4, lossless both cases, signal to noise level required for each is 40 dB (100:1)

sion. This is simply because photon shot noise is equal to the square root of the number of photoelectrons (shot noise = 10). The total noise for a real detector is the square root of the sum of the square of the shot noise and the square of the readout noise. Thus, if we wish to find the required number of photons to produce a signal to noise ratio (precision) of P we require the number of photoelectrons to be:

CCD Cameras

(1)

where R is the readout noise in electrons. For a scientific camera with 4 electrons readout noise, a light signal equivalent to 114 photoelectrons is needed. A high quality standard video rate camera produces a readout noise of approximately 50 electrons per pixel. Such a camera needs 553 photoelectrons to make the same measurement. Clearly, from Equation (1), as the precision required increases, the importance of the readout noise decreases (it becomes negligible compared to the photon shot noise), but the total number of photoelectrons required increases as the square of the desired precision. Thus, for the most precise measurements, the CCD must be able to handle large numbers of photons per pixel, even though it may take a long exposure to collect them all. Here again, scientific grade CCDs differ greatly from CCDs designed for other purposes. The amount of charge (number of photoelectrons) a CCD can store in each pixel depends largely on the physical size of the pixel. For this reason scientific CCDs usually have relatively large pixels, as large as 24 to 27 µm on a side. Since the cost of producing integrated circuits is strongly area dependent, non-scientific CCDs are usually made as small as possible, with a correspondingly smaller "well capacity". Video Rate Camera 38 lux 38 lux w/o frame averaging 3.4 lux w/frame averaging 0.086 lux w/frame averaging 0.086 lux w/frame averaging

Linearity, Dynamic Range, Precision, and Signal to Noise Ratio
These are the properties that differentiate between a quantitative, publishable measurement and a qualitative observation. Suppose one wishes to measure the light falling on a particular pixel to a precision of 10%. Even with a noiseless detector, from the random nature of photons we would need an average of 100 photoelectrons incident on the pixel to achieve this preciSlow Scan CCD Camera .045 lux .0015 lux .0015 lux 2.5 x 10-5 lux 2.5 x 10-6 lux

Table 1. Comparison between a scientific slow scan CCD camera and a video rate camera with frame grabber. 12 Princeton Instruments Tel: 609-587-9797


# frames for S/N=500 photons/sq µm-sec S/N=500 # frames S/N=10 CCD, 7.5 e/cnt Video, 230 e/cnt Video, 50 e/cnt 1 4 23 860 50,000 10,750 200 1667 9337

4

photon s/sq µm-sec S/N=10 860 50,000 10,750

4

Table 2. Comparison of scientific CCDs and commercial video arrays. One may attempt to achieve high precision measurements by using a small capacity CCD, reading it many times and summing the frames. This allows the collection of a large number of photoelectrons and overcomes the photon shot noise problem, but reading out the CCD multiple times increases the effective readout noise by the square root of the number of frames. Because of this factor, the gain in signal to noise ratio with total exposure time will not be as great as with a scientific grade CCD, which is read only once. Depending on the exact type of video-rate CCD, the addition of frames is usually performed by a dedicated frame rate arithmetic board, which may cost more than the potential savings of using a small, cheap CCD. Dynamic range is perhaps the most difficult parameter of a camera to specify in a meaningful way for scientific applications. At Princeton Instruments, we define it as the ratio of the largest signal the CCD can handle (linearly) to the readout noise (in the dark) of the CCD camera system. Immediately apparent ambiguities include, "How linear?" and "Is the signal from one pixel or a sum of pixels?" Usually for PI systems the former is not a problem as the CCD is usually able to fill the analog to digital converter (A/D) before any substantial nonlinearity occurs. Since the noise is about one A/D unit, the dynamic range becomes essentially the range of the A/D. A Princeton Instruments camera with "16 bit dynamic range" therefore has a dynamic range Pixel Arrangement 512 x 512 1024 x 1024 1024 x 1024 1152 x 770 1152 x 1242 1317 x 1035 2044 x 2033 2044 x 3072 of approximately 65,000:1. When it is necessary to combine pixels to fill the A/D range (sometimes required when an 18 bit A/D is used) the term used is "binning mode dynamic range". The discussion below on scanning modes explains the summing of pixels in greater detail. Operation of CCDs at readout rates (and therefore A/D rates) greater than 500 kilopixels/second tends to eliminate the option to use a 16 or 18 bit A/D because video rate converters of this resolution are exceedingly expensive at the present state of the art. The higher noise associated with fast readout rates also considerably reduces the dynamic range. Note that the above method of specifying dynamic range, which is the standard for scientific CCD cameras, doesn't address all concerns. One of the most important issues not addressed is intrascenic dynamic range: the ratio of the largest and smallest optical signals which can be distinguished in the same frame. This depends not only on the camera electronics, but more importantly on the optical properties of the CCD, including light scattering from its surface and light wave guiding in its surface features, and on the optics used with the CCD. In most cases, the electronic dynamic range exceeds the intrascenic dynamic range of the optical system used to convey the light to the CCD. This is due to various sources of stray light such as imperfections in the optics and typical noise, electrons 7 5 9 6 6 8 10 10 reflections from optical surfaces. In some cases a larger size scientific grade CCD can help reduce stray light problems. A large dynamic range is very helpful in experiments where the experimenter does not know a priori the amount of light to be generated by the experiment and the experiment is difficult or costly to repeat. In this case, the larger the dynamic range, the higher the probability of getting useful data on the first try. The linearity of scientific grade CCDs operated with proper signal conditioning circuitry is almost always better than 1%, except at extremely high signal levels. In fact, it is so difficult to measure the nonlinearity of these CCDs that the 1% linearity specification usually means "the CCD is more linear than the test fixture". High speed operation of CCDs sometimes reduces the achieved linearity because of slew rate limitations in either the signal conditioning electronics or in the CCD output stage itself. Thus for the most linear operation the CCD should be operated at slow scan rates. Precision of intensity measurements is improved by the use of scientific CCD cameras because of a number of factors. The low noise and high dynamic range discussed above certainly contribute. In addition, the larger pixels reduce the effects of small changes in optical system alignment with temperature, pressure, and vibration. typ. full well, electrons 300,000 400,000; 250,000 MPP 175,000; 120,000 MPP 500,000; 200,000 MPP 500,000; 200,000 MPP 45,000 85,000 85,000

CCD Cameras

pixel size, µm 24 24 13.5 22.5 22.5 6.8 9 9

Table 3. Some typical formats available in Princeton Instruments scientific CCD cameras. Fax: 609-587-1970 Princeton Instruments 13


Contrast Resolution
In many applications, such as microscopy, there is very little contrast in the image. In order to visualize the image in these cases it is necessary to apply image enhancement techniques. Of course the final image is still ultimately limited by the original data, so the better the raw data the better the enhanced image.

Additional CCDs are constantly being added to the product line, so if you need something not listed here, contact PI for the latest information. While the spatial resolution measurable in an application depends on the optical system as well as the CCD, even here slow scan scientific cooled cameras have a few advantages. First, the larger number of pixels translates directly to higher resolution. A second, more subtle effect is that of "aliasing". Aliasing is a phenomenon in sampled data systems when signals at frequencies above one half the sampling frequency (the Nyquist frequency) appear as lower frequency signals (usually at equivalent amplitude). Since all CCDs sample in the spatial domain they are all subject to aliasing in the spatial frequency domain. Finally, CCDs designed for live video (interline transfer type) usually have non-responsive areas between pixels. The one dimensional spatial frequency response of a CCD with a rectangular pixel response function is

CCD Cameras

Consider an image whose contrast is only 0.1%, defined as change in intensity/average intensity. To measure this contrast with a modest 10% accuracy requires a S/N of 10,000. As shown in Table 2, this requires considerable signal averaging. The table also compares values for an S/N = 500 (0.2 % contrast detection). Finally, the table includes the required light fluxes in photons/µm2 -sec required to achieve these S/N ratios. Two cases of video rate cameras are considered, one at 50 electrons/"count", with the camera set for maximum sensitivity, and one at 230 electrons/ "count", with the camera set for maximum signal handling, assuming a typical full well capacity of 60,000 electrons for a video rate CCD. Clearly, the video rate camera requires substantially more light, a real drawback for sensitive samples or light starved applications.

where Xa is the active pixel width. For a pixel pitch equal to Xp, the Nyquist frequency is 1/(2Xp). Clearly, for Xa < X p the response function is greater beyond the Nyquist frequency than it is if Xa = Xp. Scientific slow scan cooled cameras use all the active area for light sensing, and so have a much lower response above the Nyquist spatial frequency, therefore there will be less aliasing.

Conclusion
Scientific slow scanned CCD cameras have numerous and significant advantages over video rate cameras. Particularly important are applications where there is a low light level or a need for precision measurement of light intensity as a function of position, high spatial resolution, or low contrast. Since slow scan cameras are made in smaller volume than video rate cameras and contain more costly components, such as high resolution A/Ds, low noise signal processors, and temperature regulating electronics, they are more expensive. In many scientific applications, the difference in performance can raise the signal level to the point where it can be quantitatively measured. For these cases, the slow scan camera is well worth the additional cost.

(2)

Spatial Resolution
The most popular video rate cameras have spatial resolutions commensurate with live television standards, NTSC for US and Japan and CCIR for Europe. Thus the sensors have typically 385 to 770 pixels in the horizontal direction and approximately 300 to 600 in the vertical direction (partly this depends on how the field integration is done: 300 physical pixels can be made electrically equivalent to 600 totaled over 2 fields of interlaced video). As HDTV becomes more popular, it is expected that higher pixel counts will become available, but this is not happening as rapidly as one might wish. Many of the scientific grade imagers are partially offshoots of HDTV research.

80 70 60 50 40 30 20 10 0 300 400 500 600 700 800 900 1000 Figure 1. Back illuminated vs. interline CCD arrays. Quantum efficiency (%) shown as a function of wavelength (nm).
Commercial (interline) CCD Back illuminated CCD

14

Princeton Instruments

Tel: 609-587-9797


Selection of CCD Arrays
As more and more CCD arrays from different manufacturers become available, it becomes more and more difficult to determine which CCD is most appropriate for a specific application or range of applications. Features such as array size, spectral response, dynamic range, and scan rate must all be weighed carefully to determine the optimal array for your application.
This brochure describes general features of scientific CCD arrays. More specific information about CCD cameras available from Princeton Instruments can be found immediately following this section, in the CCD Selection Chart. Individual data sheets for all CCD arrays appear in a later section of the catalog. Additional sections of this catalog are devoted to specialized applications such as X-ray detection or multispectral imaging.

CCD Cameras

Summary of CCD Features
The following is a summary of features that you should consider in order to optimize the selection of a CCD camera for your specific application. u Format & Resolution: Number of pixels in the X and Y axes, size of pixels, and the aspect ratio of the CCD Spectral Response: QE, stability, availability of front and back illuminated modes, UV enhancement Dark Charge: Standard vs. MPP design Cooling Methodology: Thermoelectric (TE) vs. liquid nitrogen (LN), ease of maintenance Well Capacity: Signal capacity of the imaging section, shift register, and preamplifier Dynamic Range: A function of well capacity, readout noise, and readout rate Linearity: Response linearity over the specified dynamic range Readout Noise: Single pixel vs. superpixel, effect of scan rate Scan rate: Maximum scan rate, effect on noise and dynamic range

u

u u

The Kodak 3072 x 2048 array. This particular CCD offers maximum image resolution and high readout rates, but only average quantum efficiency and dynamic range. tant parameters include: pixel size, pixel density, aspect ratio, and overall package size. For example, optical microscopes have very high f/# values and therefore can utilize high density arrays with small pixels, e.g., 6.8 or 13.5 µ m. A square format, e.g. 512 x 512, is probably preferred over a TV format, e.g. 576 x 384 for this application. Princeton Instruments, Inc. offers a large variety of CCDs with formats from 576 x 384 to 3072 x 2048. Both the selection chart and the data sheets specify the number of pixels and the overall size. The effective resolution of a CCD can always be reduced by summing or "binning" the pixels in either direction. Binning results in increased dynamic range and reduced noise and readout time, but of course reduces the final image resolution. Princeton Instruments

Spectral Response
The spectral response of a CCD in the NIR is theoretically limited to approximately 1080 nm (band-gap limitation). However, in reality, it is strongly affected by the following parameters: u The polysilicon structures of frontilluminated CCDs reflect almost 60% of the incident NIR radiation. At long wavelengths most radiation passes through the silicon without being detected. Thick silicon devices (>50 µ m) improve responsivity, but usually at the expense of a reduction in resolution. Superior responsivity is obtained with thinned back illuminated CCDs (see below). QE of >20% at 1000 nm is achievable. 15

u

u

u u u

u

Format and Resolution
The ideal format for imaging largely depends on the intended application. ImporFax: 609-587-1970


u

QE decreases with temperature. For example, the peak QE of an AR coated back illuminated CCD goes from 85% at room temperature to 65% at -90°C.

At the shortest wavelengths the theoretical limit of response of a silicon CCD is 190 nm, due to SiO2 overcoat limitations. In reality, the following parameters govern the overall UV response:

2% at 400 nm, and 20% QE compared to 4% at 1000 nm. Also, the blue response is shifted to lower wavelengths, e.g., down to 260 nm (from 400 nm). Finally, these CCDs can be coated with Lumogen to provide superior UV response (30-40% compared to 15-20%).

Special Silicon Devices
By fabricating CCDs on alternative types of silicon substrates, the performance of the arrays can be enhanced. PI can provide certain CCD arrays manufactured on a proprietary substrate which enhances the quantum efficiency in the NIR and at midX-ray energies (8-20 keV). Contact Princeton Instruments for additional information.

that significantly lower dark charge. MPP CCDs have special implants under the electrodes and utilize a different phase clocking scheme. Consequently, surface charge is not collected during integration time. Because most of the dark charge is produced at the surface, the overall dark charge is reduced dramatically. The efficiency of the MPP effect depends on the characteristics of the silicon, the implant used, and on the particular design. With most EEV devices dark charge reduction is 300-500 fold. It is less for CCDs made by other manufacturers. The only adverse effect of MPP is reduction in well capacity of the imaging pixels. This results in a reduced dynamic range and a higher probability for blooming. Partially, this reduction can be alleviated when binning is performed, where the "well capacity" of the shift register and the node are used to extend the dynamic range. With EEV MPP devices the dark charge is approximately 0.003 e- /pixel-second at -60°C. Therefore, thermoelectrically cooled CCD cameras can be used for most imaging applications, even those requiring up to a few hours integration. Thermoelectric cooling is now available with either closed cycle water circulators or with integrated air fans, either method being easier to maintain than LN dewars. With cryogenic cooling, the dark charge of standard CCDs (at -120°C) is typically 0.2 to 0.5 e- /pixel-hour and therefore the use of MPP is superfluous and not recommended.

CCD Cameras

u

u

Polysilicon gates limit the lowest detectable wavelength of front illuminated CCDs to about 400 nm. Depth of rear depletion region and the consequent adverse electrical fields in the back surface limit the lowest detectable wavelength with AR coated back-illuminated CCDs to about 250 nm. Our special silicon devices overcome some of these limitations.

X-Ray Response
CCDs can be used for direct detection of Xray photons in the range of 0.2 to 30 keV. Special phosphor coatings extend this energy range to 70 keV, and a fiber optic window or taper with a special scintillator allows detection at up to 300 keV. See the X-ray section for more details.

UV-to-VIS Converters
To achieve UV response (190 to 400 nm) the CCD surface must be coated with a very thin layer of converter which absorbs UV radiation and re-emits at wavelengths in the visible region. The standard coating for UV applications is a material called lumogen. Lumogen coating is performed at the Princeton Instruments factory using very accurate vapor deposition. This is followed by an annealing process to generate a highly uniform coating capable of sustaining continuous thermal cycling. Lumogen, like any other organic UV coating, degrades over time due to photo-decomposition (bleaching). Under normal illumination this process is slow, but care must be taken to reduce unnecessary exposure to UV radiation. In-house recoating of the CCD is offered to PI customers. Some back illuminated CCDs are available with a permanent, non-organic UV/AR coating. This material achieves higher QE throughout most of the UV range, and does not degrade over time. See the CCD Selection Chart for more information.

Dark Charge
The terms "dark charge" and "dark current" refer to noise produced by thermally generated charges. Dark charge and signalgenerated charge are read out together, so there is no way to differentiate between them. Dark charge increases the background level, reducing dynamic range and increasing overall noise due to fluctuations in this dark charge signal. Although the background signal can be digitally subtracted, the dark charge noise cannot. To reduce or even eliminate these adverse effects the CCD must be cooled. The lower the temperature, the lower the dark charge. For long integration periods, especially when many pixels are added (binned) together, cryogenic dewars using liquid nitrogen (LN) as a coolant are preferred. Optimal dewar design allows LN to remain in the dewar overnight. Careful thermal control offers a wide range of operating temperatures, a user selected trade-off between QE and dark charge. The temperature is maintained with great precision, ±0.03°C in PI cameras, to ensure short and long-term baseline stability. To reduce maintenance time, an external LN dewar is available that can automatically refill the camera dewar. Dark charge is generated both in the bulk of the silicon and at the surface of the silicon, in the silicon to silicon-oxide interface regions. New CCDs, called MPP (multipinned phase) devices, are now available Princeton Instruments

Well Capacity
Well capacity is defined as the maximum number of electrons that can be contained in a single "potential" well without causing excessive spill-over to adjacent regions (blooming). Three different well capacity values are significant: a value for the imaging section, a value for the horizontal shift register, and a value representing the preamp node capacitance. The first value relates to the well capacity of an individual pixel in the imaging (light sensitive) detection region. CCD arrays with a pixel size >20 µm x 20 µ m have a typical well capacity of 400,000-750,000 electrons. For an MPP device these values are reduced to around 120-250,000 electrons. Devices with smaller pixels (<10 µm x 10 µm) can have a well capacity as small as 40-80,000 electrons. The imaging well capacity normally sets the dynamic range. The horizontal shift register has a typical well capacity of 250,000 to 900,000 Tel: 609-587-9797

Back Illuminated CCDs
To eliminate loss of response due to absorption by and reflection from the polysilicon gates on the front surface of the CCD, special CCD arrays that can detect through the rear surface of the CCD are used. These CCDs are usually thinned to a thickness of less than 15 µ m. When the rear surface is efficiently coated with antireflection (AR) material, great improvements in the overall responsivity of the CCD are achieved; for instance, 80% peak QE compared to 40%; 60% QE compared to 16


electrons. When binning is done along the column (and therefore in the shift register), the well capacity of the shift register sets the limit on dynamic range. For instance, if the readout noise is 5 electrons RMS and the well capacity is 900,000 electrons, then the maximum dynamic range is 900,000/5 or 180,000, about 17.5 bits. The well capacity of the shift register is unaffected by the MPP design. The preamp node capacitance is most important in a binning mode, since it has to read the charge integrated over the entire binned region. Typically this capacitance is only 600,000 to 750,000 electrons. Princeton Instruments proprietary technology enables an increase of this value to around 1,500,000 electrons. When binning is performed along the row (and therefore in the preamp capacitance), this extended node capacitance provides a true 18 bits dynamic range.

CCD Cameras

The EEV 1152 x 1242 array. The large pixels and low readout noise of this CCD allow a high dynamic range with average readout rates and quantum efficiency. u Readout noise is significantly reduced when the temperature is lowered below -60°C. Readout noise increases with readout rate. The rate of increase depends solely on the electrical design of the CCD and the associated external electronics. Princeton Instruments provides the lowest readout noise at any readout rate. Only Princeton Instruments camera controllers allow the user to select readout rates from 12.5 to 5 MHz, providing a user selected trade-off between scan rate and noise performance. Readout noise levels range from 4-6 electrons on such arrays as the EEV 1152 x 1242 and the SITe 1024 x 1024, to 9-10 electrons on arrays such as the Kodak 2044 x 2033. Performance can be expected to improve to a 2-3 electron level on some arrays in the near future. read out separately the overall scan rate, i.e., images/second, depends mostly on the efficient transfer of charge in the horizontal shift register and on the design of the internal pre-amplifier and external electronics, including the A/D converter circuitry. The Princeton Instruments high speed camera can sustain scan rates up to several MHz. This camera also contains a slower A/D, for high dynamic range measurements. Several cameras from the standard product line allow scan rates of up to 1 MHz, and many are now available for 500 kHz operation. For most cameras operating at this speed, a true 14 bit dynamic range is possible. Normally, only minimal thermoelectric cooling is advisable for fast-scanned CCD cameras because dark charge becomes less significant at these rates and because charge transfer efficiency is higher at higher temperatures. Transfer efficiency defines the fraction of charge which is transferred in each shift. Good quality CCDs have transfer efficiencies greater than 99.9999%.

Dynamic Range
High resolution, slow scan imaging applications normally require images at full resolution (without binning). For arrays with relatively large pixels (20-24 µm) the well capacity of a single pixel is usually sufficient to provide 17 bits of dynamic range. When binning is done in the shift register, the maximum dynamic range is 17.5 bits; when it is done in the preamp it can be as high as 18 bits. With symmetrical CCDs, e.g., 512 x 512, 1024 x 1024, or 2048 x 2048, the user can select either mode of binning without any adverse effects. With thermoelectrically cooled CCDs the camera can be rotated to the orientation necessary for a particular binning mode. With cryogenic dewars this is not possible unless a proprietary "all-direction" dewar is used, enabling operation at any orientation. The standard dewar requires an operating position within 30° of vertical.

u

u

Readout Noise
Unlike dark charge, which originates in the imaging section, readout noise originates in the output preamplifier of the CCD. This preamp measures very small variations in voltage produced on a small node capacitor each time the charge content of one or more pixels is transferred to it by the shift register. The various components of this noise are beyond the scope of this brochure. However, an ideal preamp should have a "white noise" behavior over the range of readout rates. In reality, the overall readout noise depends on the following parameters: Fax: 609-587-1970

Finally, it is important to note that all manufactures define their readout noise in electrons root-mean square (RMS, or standard deviation). The peak-to-peak variations of this noise are about five times higher. This peak-to-peak variation is more meaningful in determining the lowest detectable signal and is unfortunately often ignored.

"Cosmetics" Grades
CCD devices are usually imperfect and contain various defects. Included in this mÈlange of defects are: Dark point defects: Representing reduced response compared to adjacent pixels. Hot point defects: Representing higher dark current compared to the adjacent pixels. Hot point defects are therefore areas of high dark charge signal. 17

Scan Rate
CCD arrays can be read out at various rates depending on the design of the CCD and the external associated electronics. For imaging applications where each row is usually Princeton Instruments


Pixel traps: Pixels which interfere with the charge transfer process and usually result in either a partial or a whole bad column (either dark or white). Clusters: a cluster of adjacent pixels which is either dark or white with respect to the neighboring pixels. Unless otherwise stated, Princeton Instruments utilizes first grade CCD arrays with a very small number of defects and no partial or full defective columns. For more expensive CCDs, lower grade devices are often offered. Normally, the adverse effect of most defects can be easily compensated through computer data processing manipulations or simple averaging-out through the process of binning. Whenever available, perfect CCD arrays, i.e. without defects, are offered.

The signal, S, is determined by the quantum efficiency (QE) of the CCD, i.e., the efficiency at which photons are converted to electrons. Other important factors include the number of pixels binned together, the size of the pixels, and the signal integration time, either on the CCD or in memory. Quantum efficiency or spectral response was previously discussed. The overall noise NT associated with the acquisition of an image by the CCD is defined as: NT = [NR2 + ND2 + Nph 2] NR ND Nph the
1/2

= Readout noise of the CCD = Dark charge noise of the CCD = Photon shot noise associated with signal.

to a few thousand electrons. Usually these are in the form of very sharp spikes. Various methods are used to remove these "spikes", either with spatial or temporal digital filters. Often it is helpful to acquire a few short exposure images with less cosmic events, remove their cosmic interferences, and finally accumulate them in memory. This method is superior to that of acquiring a single very long exposure where the frequency of cosmic event occurrence is usually too high to remove them accurately. However, to do so requires superior electronics that ensure minimum 1/f noise effects. This leads to an important general comment. The difference in performance between cameras made by different manufacturers is not always obvious to the inexperienced user, not even during a limited demonstration session. Users often erroneously conclude that because two manufacturers use the "same" CCD array, the cameras made by them must also be "identical" in performance. Here are but a few areas of significance where PI products are far superior to any other: u Arrays are selected for best uniformity, lowest blemish, lowest preamp noise, lowest dark charge, minimum traps, and other factors. Post amplifier electronics provide the lowest overall noise, operate at various scan rates, do not contribute to CCD preamp noise, and have insignificant 1/f noise. Very high precision thermostating over a very wide temperature range guarantees short and long term baseline stability at all temperatures. The lowest dark charge is offered at any given temperature, using proprietary techniques. This explains why some other manufactures require the use of MPP devices with LN dewars whereas PI uses only standard arrays when cryogenic cooling is used. Camera can be mounted in any orientatation, or rotated 360° about either the x, y, or z axis. With cryogenic cooling this is achieved through the use of proprietary "all directional" LN dewars. The camera controller box and the camera cables are designed for minimal RF or magnetic interference. PI's unique dewar design has the highest LN hold time to LN volume ratio in the industry: 24 hours/500 ml of LN. This efficiency is important to maintain the compactness and ease of Tel: 609-587-9797

CCD Cameras

Special CCD Arrays
Two additional types of arrays are available in Princeton Instruments cameras. The first is a frame transfer CCD, where half of the image section is masked and used as a temporary storage region. This device allows rapid capture of two consecutive images (approximately 1 ms/image), essential for monitoring certain transient phenomena. It also allows continuous collection of images without a shutter, dramatically increasing the percentage of time the CCD is collecting data. The second device utilizes a similar concept for transient imaging. In this device most of the CCD is used as a temporary storage region. The small unmasked region, which detects the incoming signal, is shifted toward the horizontal shift register until the array is full. Following these shifts the shutter is closed and the data is digitized with the highest possible dynamic range. The temporal resolution achievable with these CCDs is on the order of 250 microseconds/image.

NR, the readout noise, depends on many parameters, such as: u The inherent on-CCD preamplifier design, which differs greatly from model to model. Scan rate: How fast the data is transferred to and read out by the preamplifier. Usually the faster the readout rate the higher the noise. The design of the post-amplifier electronics, e.g., double correlation. The preamp (CCD) temperature: Noise is reduced rapidly as temperature is lowered to around -60°C.

u

u u

u

Signal-to-Noise Considerations
The following discussion defines some quantitative criteria for the "sensitivity" of a camera and the lowest signals it can detect. Two separate issues are discussed; the first, detectability, determines the smallest signal that can be discerned from the associated noise in a single readout. The second is the signal-to-noise (S/N) performance, where S refers to the incident incoming signal (image) and N refers to the noise associated with this measurement. S/N defines the measurement precision rather than merely the camera response. 18

ND, the dark charge noise, is a result of statistical fluctuations in the dark charge accumulated in the CCD. Whereas the dark charge singal, SD, can be subtracted digitally, its associated noise, ND, cannot. The best way to reduce ND or SD, for that matter, is to lower the array temperature as previously discussed. Even when dark charge is low, however, it is important to maintain the array temperature as precisely as possible. In Princeton Instruments cameras temperature is maintained to within ±40 millidegrees C. Otherwise baseline stability is compromised resulting in poor measurement precision. Dark charge ND = (SD)1/2. Nph, the photon shot noise, is a result of statistical fluctuations in the flux of the measured signal, Sph. For the sake of simplicity we will limit the discussion here to "white noise" fluctuations where Nph = (Sph )1/2 "Cosmic noise": This source of noise is a result of high energy particles (from outer space) which strike the silicon and produce spurious signals in the range of a few tens Princeton Instruments

u

u

u

u

u


u

u

interfacing of cryogenic cameras. The dewars utilize the safest valves available to prevent loss of vacuum and accidental damage to the CCD. PI offers a universal controller that guarantees the long term viability of the system, even when new CCD arrays are introduced. At the present time CCD array technology is often considered old within 12-18 months after its first introduction to the market. PI imaging software controls all CCDs available through PI. Real time display and manipulation are available for all arrays, and changing to a different CCD array is simple and straightforward.

noise (which will be very small when Sph is small), then NT = [12 + 0.77 2]
1/2

u

= 1.26 counts RMS

u u u

Roughly 1.26 counts RMS noise represents about 5 times higher peak-to-peak noise: 1.26 x 5 = 6.3 counts. The detectability will be defined here as the signal whose magnitude equals that of the noise fluctuations: Detectability = 63 = 63 electrons , 0.25 (QE)

Either LN cooled CCD or TE cooled MPP type CCD. In either case thermostating precision is important. Thinned back illuminated CCDs are preferred. Lowest possible readout noise is necessary. If the signal is very small, binning can be used to further improve S/N and reduce detectability.

CCD Cameras

= 250 photons The smallest photon flux that can be measured is 252/1200=0.21 photon/sec-pixel. Let us examine a sub-case where = 400 nm instead of 500 nm. At this wavelength QE @ 0.01 for a front illuminated CCD and QE = 0.6 for a back illuminated CCD. The following shows how these very different QE values affect the detectability and S/N of the CCD camera. Front Illuminated CCD Sph = 5 photons/sec-pixel x 0.01 x 1200 sec. = 60e- =6 counts. N
ph

For brevity, the two main cases of a) low light level imaging, and b) high light level, high precision imaging are discussed below.

As a final note, QE is reduced as temperature drops. Therefore, it is good to have a detector which can be operated over a wide range of temperatures so that the user can trade off dark charge noise with QE. Ony PI's cryogenic cameras can maintain a temperature range of -60°C to -140°C without compromising thermostating precision.

High Light Levels
Many imaging applications require measurement of medium to high light level signals with very high precision. An example is laser induced fluorescence imaging (LIF), often used in analysis of combustion processes. The LIF signal is superimposed on intense background signals. Another is transmission microscopy, where often the incident (Io ) and transmitted (I) signals are similar in magnitude. Let us examine this case in more detail. Because Io I it is obvious that a camera with high dynamic range is needed in order to discriminate between them. Also, the ideal camera should have a short scan time and the resulting high volume data stream should be storable in real time in the computer RAM or on the hard drive. CCD arrays with larger pixels and therefore larger well capacity are usually preferred for such applications. MPP type CCDs with smaller well capacity and lower dynamic range are not well suited. Good precision, however, depends not only on the dynamic range but also on the spectral response (QE) of the CCD. Therefore a thinned, back-illuminated CCD has a significant advantage, especially below 400 nm or above 800 nm. Cameras that are currently recommended for such applications are TE/CCD-1024TK or TKB and the TE/CCD-1242E. We hope that the few examples discussed above will somewhat facilitate the selection of the CCD camera most appropriate for your application. If not, assistance is always readily available from our sales representatives in the field or our application engineers at the factory. 19

Low Light Level Imaging
Examples of this application include images acquired from Raman and luminescence signals or images from astronomical telescopes. In such cases, long signal integrations are often possible to improve detection. Incoming signal is 5 photons/second-pixel at 500 nm. QE (@ 500 nm) = 0.25 Dark charge (MPP/CCD at -50°C) = 0.05e/second-pixel. Preamp gain is set @ 1 count = 10 electrons = NR. Integration time = 1200 seconds. Then: Total dark charge signal, D, is D = 0.05 x 1200 = 60e -=6 counts ND = (60)1/2 = 7.7 = 0.77 counts. Total integrated signal Sph , is: Sph = 5 photons/second-pixel x 0.25 x 1200 second = 1500 e- = 150 counts. Nph = (1500)1/2 = 38.7 e- = 3.87 counts. NT = [12 + 0.772 + 3.872]1/2 = 4.1 counts. Clearly, in this example the noise is predominated by the photon shot noise of the measured signal, which is the ideal situation in any image detection system. S/N Performance S/N = 150/4.1 = 36.6 @ 2.7% precision An important question here is what is the smallest detectable signal, i.e. detectability, of the camera. If we ignore photon shot Fax: 609-587-1970

= (60)1/2 = 7.7e- = 0.77 cts.
1/2

NT =[12 + 0.772+0.77 2]

= 1.5 cts.

S/N = 6/1.5 = 4 or roughly only 25% precision. Detectability will be approximately 63e-/ 0.01 = 6300 photons and the lowest detectable flux will be 6300/1200 = 5.25 photon/second-pixel. Back Illuminated CCD Sph = 5 x 0.6 x 1200 = 3600 e- = 360 counts. N
ph

= (3600)1/2 = 60 e- = 6 counts
1/2

NT = [12 + 0.772 + 62 ] S/N = 360/6.1 @ 60

= 6.1 counts

or roughly 1.7% precision. Detectability is 63/0.6 = 105 photons, and the lowest detectable flux will be 105/ 1200 = 0.09 photons/sec-pixel. Summary S/N: Back-illuminated CCDs are 15 times better than front-illuminated ones. Detectability: Back-illuminated is 60 times better. It is obvious that the AR-coated back-illuminated CCD offers a great advantage in this case. To summarize, for this type of application the ideal detector should be: Princeton Instruments


CCD Selection Chart
Specifications Model Less than 1 million pixels CCD-576E EEV 02-06 576 x 384 22 x 22 12.7 x 8.4
-

1-2 Mpixels, front illuminated CCD-768K Kodak KAF-0400 768 x 512 9.0 x 9.0 6.91 x 4.6 85 14, 1 MHz; 22, 5 MHz 12, 14 with binning CCD-770E EEV 05-20 1152 x 770 22.5 x 22.5 25.9 x 17.3 500 (non-MPP) 300 (AIMO) 4-6, 50 kHz 22, 500 kHz 14 to 17 <1 (16 bits) <2 (18 bits/binning) 1.8, 500 kHz CCD-1024SF SITe ST-003 1024 x 1024 24 x 24 24.6 x 24.6 300-350 4-6, 50 kHz 24, 1 MHz 16 1-2 (17 bits) 2-3 (18 bits/binning) 2.2,500 kHz 1.2, 1 MHz CCD-1242E EEV 05-30 1152 x 1242 22.5 x 22.5 27.5 x 25.9 500 (non-MPP) 300 (AIMO) 4-6, 50 kHz 25, 1 MHz 16 (non-MPP) 16 (AIMO) 1-2 (16 bits) <3 (18 bits/binning) 2.9,500 kHz 1.5, 1 MHz

CCD-512SF SITe SI502FA 512 x 512 24 x 24 12.3 x 12.3 300-350 4-7, 50 kHz 18-21, 1 MHz

CCD Cameras

CCD Manufacturer CCD Format Pixel Size, µ m Active Area, mm Full Well Capacity, ke Readout Noise, electrons RMS Dynamic Range, bits

500 (non-MPP) 300 (AIMO) 6-8, 50 kHz 28, 1 MHz

16 (non-MPP), 14 to 16, 17 15 (AIMO) with binning

Non-Linearity, % Min. scan time, seconds for full frame TE cooling temp., °C, vacuum/nitrogen backfill Forced air circulation Water circulation Coolant circulation LN cooling temp., °C Hold time at -120°C, hours Standard dewar Large dewar Thermostating stability, millidegrees C Dark charge at -50°C, electrons/pixel-second Dark charge at -120°C, electrons/pixel-hour Quantum Efficiency, % Peak At 250 nm with converter NIR, 1000 nm

<1 (17 bits) <1.5 (16 bits), <1 (12-13 <3 (18 bits <2.5 (17 bits) bits) <2 (14 with binning) bits/binning) 0.286, 1 MHz 0.29, 1 MHz 0.41, 1 MHz; 0.09, 5 MHz

-45/-35 -55/-40 -65/-45 -70 to -130

-40/-30 -55/-40 -60/-40 -80 to -140

-45/-35 -55/-40 -60/-40 N/A

-40/-30 -50/-40 -60/-45 -80 to -130

-40/-30 -50/-40 -60/-45 -80 to -130

-40/-30 -50/-40 -60/-45 -80 to -140

>14 >30 35 4-8, <0.1 MPP <1, <0.1 MPP 40 12 3 to 5

>12 >30 40 <1, MPP only 0.6

N/A N/A 40 <0.006 N/A

>10 >25 40 6-8, <0.1 MPP <1

>10 >25 40 0.2-0.6 MPP <1

>10 >25 35 6-8, <0.1 MPP <1

35 to 40 12 3 to 5

45 12 <5

40 12 3-5

40 12 3-5

40 12 3-5

Note: See page 83 for rectangular format, pages 86-87 for line scan, and pages 22-24 for frame transfer arrays.

20

Princeton Instruments

Tel: 609-587-9797


Back illuminated CCD-1280K Kodak KAF-1300L 1280 x 1024 16.0 x 16.0 20.5 x 16.4 150 6-8,500 kHz 10, 1 MHz 12 to 14 <1 (12-14 bits) 2.7, 500 kHz 1.4, 1 MHz CCD-1317K Kodak KAF-1400 1317 x 1035 6.8 x 6.8 8.98 x 7.04 45 10, 1 MHz; 20, 5 MHz 12 (14 with binning) <1 (12-13 bits) 1.4, 1 MHz; 0.3, 5 MHz CCD-1536K Kodak KAF-1600 1536 x 1024 9.0 x 9.0 13.8 x 9.2 85 CCD-1000PB Princeton Instruments 1000 x 800 15 x 15 15.0 x 12.0 60-80 CCD-512SB SITe SI502BA 512 x 512 24 x 24 12.3 x 12.3 300-350 4-7, 50 kHz 20, 1 MHz 14 to 17 CCD-1024SB SITe ST-003 1024 x 1024 24 x 24 24.6 x 24.6 300-350 4-8, 50 kHz 13, 500 kHz 16

More than 2 million pixels CCD-2033K Kodak KAF-4200 2044 x 2033 9.0 x 9.0 18.3 x 18.4 85 CCD-3072K Kodak KAF-6300 3072 x 2048 9.0 x 9.0 27.65 x 18.48 85

CCD Cameras

9-10,500 kHz 6-8, 100 kHz 20, 500 kHz 12, 1 MHz 12 (14 with binning) <2 3.4, 500 kHz 1.8, 1 MHz 14 (16 with binning) <1 (12 bits) contact factory

12, 1 MHz; 12, 1 MHz; 20-24, 5 MHz 20-24, 5 MHz 12 (14 with binning) <2 4.2, 1 MHz; 0.95, 5 MHz 12 (14 with binning) <2 6.3, 1 MHz; 1.38, 5 MHz

<1.5 ( 1 6 bits ), <2 (16 bits) <2.5 (17 bits) <3 (18 bits/ binning) 0.5, 1 MHz 2.32, 500 kHz

-40/-35 -50/-40 -60/-50 N/A

-45/-40 -50/-40 -60/-50 N/A

-45/-35 -50/-40 -60/-40 N/A

-40/-30 -45/-35 -55/-40 N/A

-40/-30 -55/-40 -60/-40 -80 to -140

-40/-35 -50/-40 -55/-45 -70 to -130

-40/-30 -45/-35 -55/-40 N/A

-40/-30 -45/-35 -55/-40 N/A

N/A N/A 40 <1 N/A

N/A N/A 40 <0.005 N/A

N/A N/A 40 <0.006 N/A

N/A N/A 40 <0.3 MPP N/A

>10 >25 40 <1, MPP only <1

>9 25 to 35 40 0.6-2 MPP <1

N/A N/A 40 <0.006 N/A

N/A N/A 40 <0.006 N/A

45 12 <5

45 10 <5

45 10 <5

<80 40 18-20

<80 40 18-20

<80 40 18-20

45 10 <5

45 10 <5

With permanent AR coating, no converter required.

Fax: 609-587-1970

Princeton Instruments

21


Frame Transfer Cameras
CCD Cameras

Frame transfer is a powerful mode of CCD operation available for many Princeton Instruments CCD cameras. By operating in a frame transfer mode a shutter is not necessary, and the exposure duty cycle jumps to nearly 100%.
Principles of Frame Transfer Operation
In full frame CCD operation, the CCD must alternate between exposure and readout. During readout the CCD is still sensitive to light, so a mechanical shutter is used to prevent additional exposure. Due to the time necessary to open and close the shutter, and the time needed to read out the CCD, the percentage of time available for exposures is limited. The maximum frame rate is also limited by the shutter. For a given pixel readout rate, the only way to achieve a faster frame rate is to lower the exposure time. As the exposure time is shortened, the CCD records less signal, and the camera becomes less sensitive. In addition, events that occur during readout are missed completely. Lastly, as the frame rate increases, the lifetime of the mechanical shutter becomes an issue. Frame transfer operation of a CCD attempts to resolve these problems. As shown in Figure 1, only half of the CCD (the image area) is exposed to light. The other half is masked and is used to temporarily store images. This schematic diagram of frame transfer operation shows how an image focused on the image area of the array is shifted in a few milliseconds to the storage area (panel b and c). Once this image (now in the form of electrical charge) is shifted, the image area of the CCD can start collecting the next frame while the storage area is read out and digitized.

The 512 x 512 x 2 frame transfer CCD. On-chip aluminization masks half the array. is the ratio of the amount of time spent shifting divided by the exposure time between frames. Faster shifting and/or longer exposure times will minimize this effect (see Table 1). Note that while 1% smearing is insignificant in an 8 bit camera (256 gray levels), in a 16 bit camera (65,000 gray levels) 1% smearing is over 600 counts, enough to obscure faint features in a high dynamic range image.

Advantages and Disadvantages of Frame Transfer Operation
There are many advantages of frame transfer operation, including the following: u Light collection is continuous, so the system is more sensitive and will not miss transient events. Readout and light collection are simultaneous, so for a given resolution and pixel rate, frame rates are higher. A mechanical shutter is not required, so reliability is higher, particularly at the high frame rates that result from subimaging or binning. The ability to take two images in rapid succession. This is helpful in multispectral imaging where images at two Tel: 609-587-9797

Smearing Issues
When operating without a shutter image smearing may occur, depending on the exact nature of the experiment. This effect, caused by light falling on the CCD as the charge is shifted to the masked area, occurs only if the CCD is illuminated during shifting. In the case of intensified cameras (ICCDs), this effect can be eliminated by using a fast phosphor and gating the intensifier at the same frame rate as the CCD. The fraction of total signal due to smearing 22

u

Exposure Time Limitations
The minimum exposure time (without a shutter) is the time required to read out the storage section of the CCD array. To shorten the readout time, PI's flexible hardware binning and/or subimaging modes are available. Full resolution readout times for all frame transfer arrays appear in Table 2. Princeton Instruments u

u


wavelengths need to be taken with as little time in between them as possible (e.g., cell calcium experiments). When used this way, frame transfer is a simple form of kinetics imaging (in which typically 5-20 images are taken in rapid succession). This type of operation can be done with any CCD array, as it does not require independent control of the upper and lower regions. The main disadvantages of frame transfer are: u u Detection area half the size of There can be frame transfer continuous. is smaller (typically the standard array) smearing during the if the illumination is

Manufacturer EEV EEV EEV EEV SITe PI PI EEV

Total Number of Pixels on array 576 x 384 1152 x 298 1152 x 770 1152 x 1242 1024 x 1024 1100 x 330 1752 x 532 512 x 1024

Number of imaging pixels 288 x 384 576 x 298 576 x 770 576 x 1242 512 x 1024 1100 x 165 1752 x 266 512 x 512

Availability of on-array masking (aluminization) Yes No Yes Yes No No No Yes

CCD Cameras

Table 1. CCD arrays available for frame transfer cameras. A number of arrays do not provide independent clocking of the top and bottom halves of the array and therefore cannot be operated in frame transfer mode. These arrays include the EEV 1024 x 256, the SITe 512 x 512, and all of the Kodak arrays. Customers should note that frame transfer operation is a special option on a CCD detector head. It requires additional circuitry in the camera and mechanical modifications. Thus even if a CCD array is itself capable of frame transfer operation, a camera head built with that CCD array will not necessarily be able to operate in frame transfer mode unless these additional mechanical and electrical changes are made at the time of purchase. with half of the array masked by a layer of aluminum. In measurements in our laboratory, we find that no detectable signal is generated in this masked region, except near the edges. At the interface between the two halves, there can be 1 or 2 rows in the masked area with significant signal (>10% of the signal in the open area) and 3-5 more rows with slight light leakage. Around the other edges light leakage under the aluminum is low (<1%), and it extends at most 5-10 pixels in from the edge. No pin holes have been observed.

CCDs Available for Frame Transfer Operation
Arrays for operation in frame transfer mode are shown in Table 1. Although there are many scientific CCD arrays available for frame transfer operation, most are not available with masking directly on the array. It therefore becomes the responsibility of the system optics to prevent light from falling on the storage area of the array. CCD arrays supported by Princeton Instruments and available with on-array masking are the EEV 576 x 384, 512 x 1024, 1152 x 770, and 1152 x 1242 CCDs. An advantage of cameras built with nonmasked arrays is that they can also be used for full-frame (non-frame transfer) operation. Addition of the frame transfer circuitry to the detector head does not preclude this.

Mechanical Masking of Arrays
Since most frame transfer CCD arrays are not available with masking directly on the silicon, mechanical masking provides a practical alternative. In this approach, a mechanical mask is mounted just above the surface of the silicon, shielding the storage area of the array. This type of

Masking on the CCD Array
The EEV 576 x 384, 512 x 1024, 1152 x 770, and 1152 x 1242 CCDs are available

(a)

(b)

(c)

Figure 1. The above drawing illustrates frame transfer operation. An initial exposure (a) illuminates the upper half of the CCD. This charge is shifted to the masked section of the array (b). Once this shift is complete this charge is read while another exposure is taking place (c). Fax: 609-587-1970 Princeton Instruments 23


Array

Size

Shift time Frame per row, transfer time, microseconds milliseconds 3.0 3.0 4.8 4.8 15.0 153.6 3.0 0.87 1.73 2.76 2.76 7.68 78.64 1.54

A/D rate

Full resolution readout time, milliseconds 234/133 369 900 1458/811 1248/656 1396 271

Smearing when exposure time is minimum read time 0.7% 0.5% 0.3% 0.3% 1.2% 5.6% 0.6%

EEV EEV EEV

576 x 384, non-MPP 1152 x 298, non-MPP 1152 x 770, non-MPP 1152 x 1242, non-MPP 1024 x 1024 front, MPP 1024 x 1024 back, MPP 512 x 1024, non-MPP

500 kHz/1 MHz 500 kHz 500 kHz 500 kHz/1 MHz 430 kHz/1 MHz 430 kHz 1 MHz

CCD Cameras

EEV SITe SITe EEV

Table 2. Frame rates for various CCDs calculated using the maximum A/D speeds. Values are for TE/CCD cameras only. masking causes a soft edge (0.1-1 pixel wide) and thus somewhat fewer usable pixels. It is possible to align this mask to within one millimeter of the center of the CCD array, and the mask to within one millimeter of the surface of the silicon. Mechanical masking in this way is not as good as masking on the array, as some light can leak in under the mask. To minimize the reflections between the mask and the CCD the masks are black anodized. Below are two examples of the typical performance of mechanical masks. Example #1 If f/5 optics are used, a mask 1 mm from the CCD surface causes a shadow with an edge 0.2 mm wide. On a CCD with 24 µm pixels (e.g., SITe 1024 x 1024), this results in a "soft edge" 8 pixel wide. Our ability to center the mask only within 1 mm (fine adjustment provided) adds another 40 pixels. The resulting useful area is therefore at least 1024 x (512-48) = 1024 x 464, i.e. 90% of what it would be if on-array masking were available. For many applications, this performance is acceptable. Example #2 As a second example, assume that f/1.4 optics are used. The shadow of a mask 1 mm from the silicon surface in this instance has an edge 0.7 mm wide. With 13 µm pixels (e.g., Reticon 1024 x 1024) this results in a 54 pixel soft edge. Our ability to center the mask only within 1 mm adds another 74 pixels. This reduces the useful area from 1024 x 512 to 1024 x 384, i.e. 75% of what it would be if on-array masking were available. The acceptability of this performance depends on the user's application. 24 Note that masking can occur anywhere in the optical system and there are advantages to masking in places other than directly above the array. One possibility is to mask the illumination, so that only the portion of the subject seen by the top half of the array is illuminated. Alternatively, a black mask can be placed between the subject and the lens (or other optics), limiting the field of view and hence the portion of the array which is illuminated. These approaches may allow use of a higher fraction of the array, closer to the level achieved by on-array masking. They also retain the option of imaging with the entire CCD illuminated. as higher and higher vertical shifting rates are used. Note that the vertical shift time is user selectable through software, and is independent of the A/D speed. Note: all of these times in Table 2 are for TE/CCD cameras. LN/CCD models often shift much more slowly. For the most up-to-date information on maximum performance for a particular array, please contact your Princeton Instruments sales representative.

Availability
All of the arrays listed in Table 1 are currently available in a Princeton Instruments frame transfer camera. These are offered as thermoelectrically cooled (TE/ CCD) models, with air or liquid used for heat dissipation. Liquid nitrogen cooled (LN/CCD) cameras, which generally have slower shift rates and are used primarily for long exposures, are not normally available with frame transfer operation. A frame transfer 576 x 384 ICCD is available from Princeton Instruments, with a special moveable intensifier masking. In this type of detector the intensifier can be used as a shutter, preventing smearing due to continued exposure during the frame shift time. Intensified CCD cameras for use in frame transfer mode should also use a high speed phosphor in the intensifier, to minimize smearing due to phosphor lag during the frame shifting. Since these high speed phosphors have 3-10 times less light output per photoelectron, fiber optic coupling between the intensifier and CCD is even more important than in normal operation. Without it, single photoelectron detection is nearly impossible. Tel: 609-587-9797

Performance of Specific CCD Arrays
The amount of time required to shift the image into the storage area depends on the number of rows to be shifted and the vertical shift time per row. These values vary from array to array and between front and back illuminated configurations. Generally, Princeton Instruments systems use standard vertical shift rates that are known to be "safe", i.e., rates that maintain full vertical well capacity, charge transfer efficiency, and linearity. For some arrays these "safe" values may be very conservative. Contact the factory for the most current information on readout rates. Table 2 reflects speeds that have been confirmed experimentally. If these arrays must be read at higher rates, they will probably have lower signal levels (lower full well capacities). At the highest transfer rates, the charge transfer efficiency suffers, resulting in blurred images. Performance degradation begins at the center of a CCD Princeton Instruments


Some Limitations of CCD Arrays
Modern scientific CCD arrays are not, unfortunately, perfect. To achieve maximum performance, it is important to understand the limitations of the arrays and how camera design can minimize their impact.
Blemishes in CCD arrays are generally classified into point, cluster and column defects. What is often misunderstood is that these defects are not just regions of reduced sensitivity. They can also be regions of dramatically increased dark current. In these pixels, the dark current can be high enough to swamp the true signal. It may seem that this is not any worse than a dark defect (in which the photoresponse is low or even zero) because in either case, the data from those pixels is lost. Hot pixels can be worse however: In a long exposure, a hot pixel can bloom into the adjacent pixels, particularly those above and below in the same column. Thus a hot defect which only occupied a single pixel when the CCD was being tested by the manufacturer (typically with a fairly short exposure time) can become a hot column when a longer exposure is used. Princeton Instruments works with CCD manufacturers to exclude this type of defect from arrays whenever possible. This extra level of screening leads to a separate grade level, called PI Grade 1. Unfortunately, it is not yet available from all CCD manufacturers. There is also some potential confusion about the definition of a hot defect. It is not a pixel whose dark current is above the dark current specification. That specification is for the average of all pixels except the blemishes. Individual pixels are not held to this standard. The threshold for a pixel to be considered a hot defect is generally much higher. On non-MPP CCD arrays, most array manufacturers use a threshold which is ten times the maximum average dark current. On MPP CCD arrays (where the average dark current is much lower) the threshold is as much as 100x the maximum allowable average dark current. The reason for these high thresholds is that the histogram of dark current values on a CCD has a tail, with a substantial number of pixels having dark current which is significantly above average. If the CCD manufacturers set their thresholds for hot defect status at only 2x average dark current, it would be difficult for them to Fax: 609-587-1970
Number of pixels this dark current level

Hot Defects and Warm Pixels

10,000,000

CCD Cameras

1,000,000

100,000

10,000

1,000 0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

9.0

Dark current / average dark current

Dark current histogram measured on a 1.3 million pixel CCD.
1,000,000
dark current, electrons/pixel-second

100,000 10,000 1,000 100 10 1 0.1 0.01 0.001 -60 -40 -20 0 20 40 60 Normal MPP axel "Warm" pixel

Temperature, °C

Comparison of dark current (electron/pixel-second) vs. temperature (°C) for a "warm" pixel vs. a normal pixel on a MPP array. The dark current of the warm pixel reduces with temperature, but more slowly than the normal pixel. In this regard, the warm pixel behaves more like a pixel on a non-MPP CCD. ship anything. Thus the high threshold is a practical consequence of the current state of the art. The first figure shown here is a dark current histogram for a 1.3 million pixel CCD. From it one can see that thousands of pixels (this is still less than 1% of the CCD area) have "excess" dark current. These pixels with increased but subthreshold dark current are sometimes called "warm" pixels. The physical mechanisms that cause dark current in these pixels are different from those in the other pixels. As a result, Princeton Instruments the dark current in these pixels reduces more slowly with cooling. Thus the difference is even more prominent in a camera where the CCD is cooled. The second figure shows typical dark current reduction with cooling for a warm pixel vs a normal pixel on an MPP CCD. At room temperature (where many CCD manufacturers do their testing), the hot pixel has 50 times the dark current of the normal pixel. It would thus not be classified as a blemish, because it is less than 100 times the specification for the average dark current. At -50°C however, the ratio has become almost 1,000:1. 25


For many applications, the warm pixels still have low enough dark current to be negligible. For others, they may require background subtraction. In general, they cannot be avoided: even Grade 0 CCD arrays have "warm" pixels. For applications which are affected by this problem, the best solution is to continue lowering the CCD temperature, to a point where both the normal pixels and the warm pixels have acceptably low dark current. This requires more cooling than one might expect from interpretation of the (average) dark current specification. In some cases, this requires as much as 30-40 degrees more cooling. Princeton Instruments cameras have been designed to achieve the lowest possible temperatures for their configurations, and they routinely achieve much lower temperatures than similar products from our competitors. In some cases, we achieve lower temperatures with air cooling than competitors achieve with water cooling.

by them and classified as a partial dark column defect. By operating the vertical shifting of a CCD more slowly, some charge traps may be reduced in magnitude or eliminated. Princeton Instruments' Model ST-138 Camera Controllers support software definition of the vertical shift rate, and this may be used to improve the performance of an individual CCD if traps are a problem.

bination of better cooling and better analog electronic design.

Cosmic Rays
All CCD arrays are sensitive to cosmic rays. These events, which typically occur about once per second per square centimeter of CCD area, can generate a narrow spike of charge in one or a few adjacent pixels. Each spike can contain several thousand electrons of charge. This is small enough to be almost invisible in the noise of a video CCD, but it is a very noticeable signal on a quiet scientific CCD. For short exposures on small CCD arrays, many images will be free of cosmic ray "hits". For long exposures on a big CCD array, it will be almost impossible to completely avoid these. If the image that you are trying to acquire is fainter than the cosmic ray signals, then autoscaling the image display may incorrectly scale on the spikes caused by the cosmic hits instead of the actual image data. This can be avoided by autoscaling on a mouse selected region of interest. The traditional method to eliminate cosmic ray corruption from long exposure images taken with cooled CCD cameras is to take a pair of identical exposures. The chance of a cosmic ray hit occuring in the same pixel in two successive images is small. Thus if the images are compared in software, the pixels which have been hit in each image can be identified. The images can then be merged in software, keeping just the uncorrupted values. Princeton Instruments software can be used to perform this elimination.

CCD Cameras

Readout Noise and Dynamic Range
The readout noise of a CCD array generally increases with the pixel readout rate and the CCD temperature. The actual relationship of noise to pixel rate is a complex function of many variables in the design of the on-chip amplifier of a CCD. It is not unusual for the readout noise to increase by 2 - 4 â from 50 kHz to 1 MHz. This is the reason that Princeton Instruments has introduced dual speed CCD controllers. The high speed can be used for operations where minimum noise is not crucial, such as camera alignment, and the lower speed can be used for precision data taking. The readout noise of a cooled CCD can be up to 2x lower than the same CCD at room temperature, and better cooling generally results in lower noise. This is another reason that we push our designs to achieve the lowest possible temperatures. Even below the temperature at which dark current becomes negligible, a CCD's performance can still be improved by lowering its temperature (and thus lowering the read noise). Princeton Instruments can often offer lower read noise than competitors specifically because of this better cooling. Dynamic range is the ratio of the maximum signal that can be accomodated in the linear range of a CCD to the RMS readout noise. Thus as noise is reduced, the dynamic range which can be achieved becomes higher. Princeton Instruments offers the highest dynamic range (at any given pixel rate) of all the major manufacturers. We are the only manufacturer to offer: u u u u 12 bits at 5 MHz 14 bits at 1 MHz 16 bits at 430 kHz 18 bits at 100 kHz

Charge Traps
When a CCD is being read out, charge packets are clocked down the columns to the horizontal register, and from there to the amplifier. Charge transfer along the columns must be essentially perfect for clean imaging. If there is an obstruction in a column that impedes charge transfer, the result can be a partial dark column. Lesser obstructions can lead to what are called traps. In a low level trap, a minimum level of charge must be "poured into" the trap pixel before the pixel becomes filled and charge packets can be passed correctly "over the top". At low light levels, this defect will appear as a string of pixels in a column which are dark. As the light level is increased, the number of pixels affected decreases. At high enough light levels, many traps dissapear. Thus a defect definition which uses high signal levels may hide some defects that will show up at lower levels. We applaud the CCD manufacturers that are beginning to include low level charge traps in their defect list for CCD grading purposes. Another type of trap can occur when there is the equivalent of a constriction in the charge path at a pixel. Small amounts of charge may transfer past these pixels unaffected, but large amounts of charge may not be able to pass. Thus these traps only show up at high light levels. Since most CCD manufacturers use moderate to high light levels in screening their CCD arrays for defects, this type of trap is generally caught 26

Summary
The limitations of performance of modern CCD arrays are often subtle and easy to misunderstand if they have not been explained. In addition, many of them can be overcome if a CCD camera system is used that provides extra cooling, flexible clock timing and powerful software. Many of these advanced features have been designed into Princeton Instruments camera systems and operate automatically. These are some of the engineering features which make the performance difference between our systems and "more economical" systems that use scientific CCD arrays from the same manufacturers. A cooled CCD camera's performance is determined by a combination of the CCD itself and the camera that drives it. To get the best data that is possible and avoid these CCD limitations, one should use a grade one CCD in a "grade one camera". Tel: 609-587-9797

These are not just A/D converter specifications, but real CCD dynamic range. This higher performance is achieved by a comPrinceton Instruments


Cooled CCD Camera Data Sheets
Princeton Instruments has the broadest line of cooled CCD cameras of any manufacturer in the world. Rather than attempt to address ever y imaging application with one or two camera models, Princeton Instruments offers over 35 different cameras for imaging applications, not including variations such as UV coating or MPP operation.
Cooling and Vacuum
All Princeton Instruments unintensified CCD cameras are cooled for optimal performance. Two methods of cooling are available: thermoelectric, models TE/CCD, TEA/ CCD, RTE/CCD, and PentaMAX, or liquid nitrogen cooling, model LN/CCD. Princeton Instruments TE/CCD, TEA/CCD, RTE/CCD, and PentaMAX cameras use a four-stage peltier device to cool the CCD. To prevent condensation and contamination from occurring, these cameras can either be evacuated or nitrogen backfilled. As indicated by the following pages (values given for TE/CC D and TEA/CCD cameras only), cameras under vacuum reach lower temperatures, while nitrogen backfilled detectors are relatively maintenance free. There are three means of dispersing the heat of the peltier: passive air, forced air, or circulating coolant. The following pages give values for air circulation, available only on TEA/CCD, RTE/CCD, and PentaMAX cameras, and tap water/chilled coolant circulation, available on all TE/CCD and TEA/CCD models. Chilled coolant must be provided by a closed circulator, available from companies such as Neslab. The RTE/ CCD model also offers passive cooling. Princeton Instruments LN/CCD cameras use liquid nitrogen to cool the CCD. Several types of dewars are available, including large capacity (1.5 liters) and all-directional (for operation in any orientation). Because of the extremely low operating temperatures, these cameras should only be operated under vacuum.

CCD Cameras

Window Specifications
All Princeton Instruments camera windows are required to meet strict specifications. For applications involving highly collimated light or coherent radiation, special grade windows may be required. See page 122 for more detailed window specifications. Coated windows are also available to further optimize performance.

CCD Blemish Defects
CCDs are generally available in several cosmetic grades, as specified by the manufacturer. For every system, Princeton Instruments independently confirms that the CCD meets or exceeds these specifications. In many cases, Princeton Instruments offers cameras with superior performance (fewer or no defects). Contact the factory or your sales representative for information on a specific array.

The model TEA/CCD cooled camera.

Individual Specifications
The following pages are specifications for individual Princeton Instruments imaging cameras. PentaMAX, Small Cooled (model RTE/CCD), rectangular format, and line scan cameras are described in the sections discussing these specialized cameras. Some cameras are available with or without MPP, as discussed in Selection of CCD Arrays. This feature affects dark charge. Readout noise, another important but confusing camera specification is given as values for total readout noise RMS. Peakto-peak values are generally 5 times higher. All of these cameras are operated by Princeton Instruments universal controllers with interfaces to various computers. These cameras represent the state-of-theart of CCD camera technology. As such, the information on the following pages is subject to change without notice. Please contact your sales representative for the most current information. 27

Two dewar options for the model LN/CCD camera. Fax: 609-587-1970 Princeton Instruments


CCD Cameras
based on EEV 512 â 512 frame transfer CCD
These new frame transfer ideal for imaging applications frame rates, vibration free acquisition of two frames in sion. cameras are requiring high operation, or rapid succes-

TE/CCD-512EFT RTE/CCD-512EFT PentaMAX-512EFT

100 90
Quantum Efficiency, %

CCD Cameras

80 70 60 50 40 30 20 10 0 200 300 400 500 600 700 800 900 1000 1100 With UV Coating Front Illuminated

When comparing slow scan CCD cameras, these are noted for the following features. u u u u u u Binary pixel format, 512 x 512 High scan rate, up to 5 MHz High sensitivity, due to 100% detection duty cycle Relatively low cost Shutterless operation Frame transfer with on-array masking

Wavelength, nm

Performance Characteristics
CCD Arrays: Format: EEV CCD-37; Note that at the time this catalog was written, EEV unfortunately does not offer an MPP version of this CCD. Therefore cooling is particularly important for this CCD. 512 x 512 x 2; 7.7 x 7.7 mm overall; 15 x 15 µm pixels 200,000 electrons 10 electrons at 50 kHz; 18 electrons at 1 MHz; For performance at 2.5 and 5 MHz contact your Princeton Instruments representative 400-1080 nm; 190-1080 nm with UV-to-visible converter 14-15 bits at 50 kHz; 13-14 bits at 1 MHz; 12 bits at 2.5 and 5 MHz < 1% < ±4% over entire CCD area, except blemish regions EEV measures blemishes with 20 msec integration time and video rate readout at room temperature. Under these conditions a dark defect is made of pixels with <50% of the normal sensitivity and a hot defect would exceed 250,000 electrons/pixel-second at room temperature (i.e., ~25% normal) No full or partial dark columns; 2 white spots; 7 dark clusters containing three or fewer pixels, no dark clusters containing more than 3 pixels; Traps are counted as dark clusters TE/CCD, -45°C with air circulation, -60°C with tap water circulation, -35°C with air circulation and nitrogen backfill ±0.040°C over entire temperature range 11 electrons/pixel-second at -40°C using PI mode 1; 0.75 electrons/pixel-second at -40°C using PI mode 2 TE/CCD and RTE/CCD maximum scan rate is 1 MHz; PentaMAX models are available with maximum scan rates of 2.5 and 5 MHz 0.273 seconds at 1 MHz; 0.063 seconds at 5 MHz At 5 MHz, 15 frames/second without binning, 29 frames/second with 2 â 2 binning, 52 frames/second with 4 â 4 binning, 85 frames/second with 8 â 8 binning, 125 frames/second with 16 â 16 binning, 160 frames/second with 32 â 32 binning < 1.6 milliseconds/frame Princeton Instruments Tel: 609-587-9797

Full Well Capacity: Readout Noise: Spectral Range: Dynamic Range: Response Nonlinearity: Response Nonuniformity: Blemish Definitions:

Blemish Specifications: Operating Temperature:

Thermostating Precision: Typical Dark Charge: Scan Rate: Full Frame Readout: Maximum Frame Rates:

Vertical Shift Time: 28


CCD Cameras
based on EEV 1152 â 1242 CCD
These cameras utilize large format, high pixel density (close to 1.5 million pixels) CCD arrays specifically designed for high resolution imaging. The large size pixels with their large well capacity ensures maximum light collection (with low f/# optics) and wide linear dynamic range. When comparing slow scan CCD cameras, these are noted for the following features. u u u u Combination of high number of pixels and large well capacity Available non-MPP for maximum well capacity Low dark charge (MPP version) Full frame or frame transfer operation

TE/CCD-1242E LN/CCD -1242E

100 90
Quantum Efficiency, %

80 70 60 50 40 30 20 10 0 200 300 400 500 600 700 800 900 1000 1100 With UV Coating Front Illuminated

CCD Cameras

Wavelength, nm

Performance Characteristics
CCD Arrays: Format: EEV model 05-30, standard or AIMO (MPP) 1152 x 1242, 25.9 x 27.5 mm overall, 22.5 x 22.5 µm pixels; 576 x 1242 for frame transfer, 12.9 x 27.5 mm imaging area Note that the MPP version is not recommended for most frame transfer applications due to slower vertical shifting Standard, 500,000 electrons; AIMO (MPP), 300,000 electrons 4-6 electrons at 50 kHz; 25 electrons at 1 MHz 400-1080 nm; 190-1080 nm with UV-to-visible converter 16 bits 1-2% for 14-16 bits; < 3% for 18 bits ±3% over entire CCD area, except blemish regions Dark defects, pixels with less than 50% of the signal of surrounding pixels; Non-MPP hot defects, >10â specified average dark current; AIMO (MPP) hot defects, >100â specified average dark current; Traps are counted as dark clusters 43 or fewer point defects; 6 or fewer cluster defects; 2 or fewer partial column defects or 1 or fewer full column defects; Higher and lower grade devices are available on request, please call the factory for details TE/CCD, -50°C with tap water circulation, -60°C with coolant circulation, -40°C with air circulation; LN/CCD, -80°C to -140°C ±0.035°C over entire temperature range At -50°C, 6-8 electron/pixel-second for standard, <0.1 electron/pixel-second for MPP; At -120°C, < 1 electron/pixel-hour for standard TE/CCD maximum scan rate is 1 MHz; LN/CCD maximum scan rate is 100 to 150 kHz 1.5 seconds at 1 MHz > 10 hours for standard dewar; > 25 hours for large capacity dewar Princeton Instruments 29

Full Well Capacity: Readout Noise: Spectral Range: Dynamic Range: Response Nonlinearity: Response Nonuniformity: Blemish Definitions:

Blemish Specifications:

Operating Temperature:

Thermostating Precision: Typical Dark Charge:

Scan Rate: Full Frame Readout: Liquid Nitrogen Hold Time:

Fax: 609-587-1970


CCD Cameras
based on EEV 1152 â 770 CCD
These cameras use CCD arrays with the same large pixels and high well capacity of the EEV 1152 â 1242, but with a somewhat smaller pixel count and thus lower cost. In frame transfer mode, they offer a less oblong format than the 1152 â 1242. When comparing slow scan CCD cameras, these are noted for the following features. u u u u u Combination of high number of pixels and large well capacity Available non-MPP, for maximum well capacity Low dark charge (MPP version) Full frame or frame transfer operation (masking available on the array) Moderate cost

TE/CCD-770E TE/CCD-770EFT LN/CCD -770E

100 90
Quantum Efficiency, %

CCD Cameras

80 70 60 50 40 30 20 10 0 200 300 400 500 600 700 800 900 1000 1100 With UV Coating Front Illuminated

Wavelength, nm

Performance Characteristics
CCD Arrays: Format: EEV model 05-20, standard or AIMO (MPP) 1152 x 770, 25.9 x 17.3 mm overall (3:2 aspect ratio), 22.5 x 22.5 µm pixels; 576 x 770 for frame transfer (4:3 aspect ratio), 12.9 x 17.3 mm imaging area; Note that the MPP/AIMO version is not generally recommended for frame transfer operation due to its slower vertical shift rate Standard, 500,000 electrons; AIMO (MPP), 300,000 electrons 4-6 electrons at 50 kHz; 22 electrons at 500 kHz 400-1080 nm; 190-1080 nm with UV-to-visible converter 14-17 bits < 1% for 14-16 bits; < 2% for 18 bits ±3% over entire CCD area, except blemish regions Dark defects, pixels with less than 50% of the signal of surrounding pixels; Non-MPP hot defects, >10â specified average dark current; AIMO (MPP) hot defects, >100â specified average dark current; Traps are counted as dark clusters 27 or fewer point defects; 4 or fewer cluster defects; 2 or fewer partial column defects or 1 or fewer full column defects; Higher and lower grade devices are available on request, please call the factory for details TE/CCD maximum scan rate is 1 MHz; LN/CCD, -80°C to -130°C ±0.040°C over entire temperature range At -50°C, 6-8 electron/pixel-second for standard, <0.1 electron/pixel-second for MPP; At -120°C, < 1 electron/pixel-hour for standard TE/CCD maximum scan rate is 1 MHz; LN/CCD maximum scan rate is 100 to 150 kHz < 1 second at 1 MHz > 10 hours for standard dewar; > 25 hours for large capacity dewar Princeton Instruments Tel: 609-587-9797

Full Well Capacity: Readout Noise: Spectral Range: Dynamic Range: Response Nonlinearity: Response Nonuniformity: Blemish Definitions:

Blemish Specifications:

Operating Temperature: Thermostating Precision: Typical Dark Charge:

Scan Rate: Full Frame Readout: Liquid Nitrogen Hold Time:

30


CCD Cameras
based on EEV 576 â 384 CCD
These cameras are based on CCD arrays that have moderate resolution, but high dynamic range. Quantum Efficiency, % When comparing slow scan CCD cameras, these are noted for the following features. q q q q q q Large pixels, for efficient light collection Available non-MPP, for maximum well capacity Moderate resolution but high dynamic range Low dark charge (MPP version) Highest frame rates with frame transfer devices On-array masking for frame transfer operation is available

TE/CCD-576E LN/CCD-576E PentaMAX-576EFT

100 90 80 70 60 50 40 30 20 10 0 200 300 400 500 600 700 800 900 1000 1100 Front Illuminated with UV Coating Front Illuminated

CCD Cameras

Wavelength, nm

Performance Characteristics
CCD Arrays: Format: EEV model CCD02, standard or AIMO (MPP) 576 x 384, 12.7 x 8.4 mm overall (3:2 aspect ratio), 22 x 22 µm pixels; 288 x 384 for frame transfer (4:3 aspect ratio), 6.3 x 8.4 mm imaging area; Note that for frame transfer operation, the non-MPP version of this CCD offers faster frame shift operation and hence somewhat less smearing Standard, 400,000 electrons; AIMO (MPP), 300,000 electrons 6-8 electrons at 50 kHz; 28-30 electrons at 1 MHz; 50-54 electrons at 5 MHz (PentaMAX only) 400-1060 nm; 190-1060 nm with UV-to-visible converter 16 bits standard, 15 bits MPP/AIMO at 50 kHz, up to 14 bits at 1 MHz < 1% for up to 17 bits; < 3% for 18 bits with binning ±3% over entire CCD area, except blemish regions Dark defects, pixels with less than 50% of the signal of surrounding pixels; Non-MPP hot defects, >10â specified average dark current; AIMO (MPP) hot defects, >100â specified average dark current; Traps are counted as dark clusters 16 or fewer point defects; 2 or fewer cluster defects; no column defects; Higher and lower grade devices are available on request, please call the factory for details TE/CCD, -55°C with tap water circulation, -65°C with coolant circulation; -45°C with air circulation; LN/CCD, -70°C to -130°C ±0.035°C over entire temperature range 4-8 electrons/pixel-second at -50°C standard; 35-40 electrons/pixel-second at -40°C standard; < 0.1 electrons/pixel-second at -50°C MPP TE/CCD maximum scan rate is 1 MHz; LN/CCD maximum scan rate is 100 to 150 kHz 0.286 seconds at 1 MHz; 0.026 seconds at 5 MHz At 1 MHz, 8 frames/second without binning, 75 frames/second with 4 â 4 binning, 263 frames/second with 16 â 16 binning At 5 MHz, 39 frames/second without binning, 126 frames/second with 4 â 4 binning, 286 frames/second with 16 â 16 binning > 14 hours for standard dewar > 30 hours for large capacity dewar Princeton Instruments 31

Full Well Capacity: Readout Noise: Spectral Range: Dynamic Range: Response Nonlinearity: Response Nonuniformity: Blemish Definitions:

Blemish Specifications:

Operating Temperature:

Thermostating Precision: Typical Dark Charge: Scan Rate: Full Frame Readout: Maximum Frame Rates:

Liquid Nitrogen Hold Time:

Fax: 609-587-1970


CCD Cameras
based on Kodak 768 â 512 CCD
These new economical cameras are ideal for microscopy or other imaging applications requiring medium resolution and 12 bit dynamic range. Because this CCD can maintain relatively low noise at speeds of 1 MHz, it can deliver over 2 frames/second with 12 bit dynamic range. q q q q q High scan rates, up to 1 MHz Low noise, even at high pixel rates Extremely low dark charge Low cost TE/CCD model Antiblooming available

TE/CCD-768-K RTE/CCD-768-K

100 90
Quantum Efficiency, %

CCD Cameras

80 70 60 50 40 30 20 10 0 200 300 400 500 Front Illuminated with UV Coating Antiblooming 600 700 800 900 1000 1100 Front Illuminated

Contact the factory for specifications on the antiblooming version of this array.

Wavelength, nm

Performance Characteristics
CCD Array: Format: Kodak model KAF-0400 768 (H) x 512 (V); 6.91 x 4.6 mm overall; 9.0 x 9.0 µm pixels Imaging section, 85,000 electrons; horizontal shift register, 170,000 electrons; preamp node, 340,000 electrons 9-10 electrons at 500 kHz; 14 electrons at 1 MHz 400-1080 nm; 190-1080 nm with UV-to-visible converter 500 kHz, 12 bits; 430 kHz, 13 bits, 14 bits with binning; 1 MHz, 12 bits < 1.5% for 12-13 bits; < 2% for 14 bits with binning ±4% over entire CCD area, except blemish regions Point defect, a pixel which deviates more than 6% (about 3,500 electrons) from adjacent pixels when illuminated to 70% of saturation level; Cluster defect, a grouping of not more than 5 adjacent point defects; Column/row defect, a grouping of point defects along a single column or row; Adjacent pixels, the surrounding 100 x 100 pixels or ±50 columns/rows Grade 0, no point, cluster, or column defects; Grade 1, 5 or fewer point defects, no cluster or column defects; Grade 2, 10 or fewer point defects, 4 or fewer cluster defects, 2 or fewer column defects -20°C for RTE/CCD with passive cooling; -45°C for TE/CCD, -40°C for RTE/CCD with air circulation; -60°C for TE/CCD with coolant circulation; -35°C for TE/CCD, -30°C for RTE/CCD with air circulation and nitrogen backfill ±0.040°C over entire temperature range <0.006 electrons/pixel-second at -50°C Up to 1 MHz 0.407 seconds at 1 MHz

Full Well Capacity:

Readout Noise: Spectral Range: Dynamic Range:

Response Nonlinearity: Response Nonuniformity: Blemish Definitions:

Blemish Specifications:

Operating Temperature:

Thermostating Precision: Typical Dark Charge: Scan Rate: Full Frame Readout:

32

Princeton Instruments

Tel: 609-587-9797


CCD Cameras
based on Kodak 1280 â 1024 antiblooming CCD
These cameras are ideal for imaging applications requiring over 1K â 1K resolution, but with anti-blooming control. This control is accomplished using on-chip lateral overflow drain and off-chip electronics. When comparing various slow scan CCD cameras, these are noted for the following features. u u u High resolution, 1.3 million pixels 1 MHz scan rates Moderate pixel size (16 µm) for extra well capacity Antiblooming control

TE/CCD-1280-K

100 90
Quantum Efficiency, %

80 70 60 50 40 30 20 10 0 200 300 400 500 600 700 800 900 1000 1100 Front Illuminated with UV Coating Front Illuminated

CCD Cameras

u

Wavelength, nm

Performance Characteristics
CCD Array: Format: Full Well Capacity: Readout Noise: Spectral Range: Dynamic Range: Antiblooming: Response Nonlinearity: Response Nonuniformity: Blemish Definitions: Kodak model KAF-1300L 1280 (H) x 1024 (V); 20.5 x 16.4 mm overall; 16.0 x 16.0 µm pixels Imaging section, 150,000 electrons; horizontal shift register, 300,000 electrons; preamp node, 300,000 electrons 6-8 electrons at 500 kHz; 10 electrons at 1 MHz 400-1080 nm; 190-1080 nm with UV-to-visible converter 500 kHz, 12 bits; 430 kHz, 13 bits, 14 bits with binning; 1 MHz, 12 bits 1000â protection < 1% for 12-14 bits ±4% over entire CCD area, except blemish regions Point defect, a pixel which deviates more than 6% (about 6,000 electrons) from adjacent pixels when illuminated to 70% of saturation level; Cluster defect, a grouping of not more than 5 adjacent point defects; Column/row defect, a grouping of point defects along a single column or row; Adjacent pixels, the surrounding 100 x 100 pixels or ±50 columns/rows; brackets below indicate the maximum number of defects for the central 800 x 600 region Grade 0, no point, cluster, or column defects; Grade 1, 5 [2] or fewer point defects, no cluster or column defects; Grade 2, 10 [5] or fewer point defects, 4 [2] or fewer cluster defects, 2 [0] or fewer column defects Antiblooming on this CCD will generally prevent hot pixels from blooming and becoming hot columns in long exposures. Hot pixels can, however, still contribute additional dark current to the other pixels in the same column during the short time that charge packets from the other pixels pass through the hot pixel during readout. This extra dark current may appear as a "warm column", i.e., a column of pixels with extra dark current, but not enough to be considered a column defect by Kodak. Defects in the barrier between the active light sensing region and the antiblooming drain can cause charge traps. These traps may or may not be large enough to meet the threshold in Kodak's definition of a dark blemish. -40°C with air circulation; -60°C with coolant circulation; -35°C with air circulation and nitrogen backfill ±0.040°C over entire temperature range < 1 electron/pixel-second at -40°C Up to 1 MHz 1.4 seconds at 1 MHz Princeton Instruments 33

Blemish Specifications:

Operating Temperature:

Thermostating Precision: Typical Dark Charge: Scan Rate: Full Frame Readout: Fax: 609-587-1970


CCD Cameras
based on Kodak 1317 â 1035 CCD (6.8 µm pixels)
These new cameras are ideal for microscopy or other imaging applications requiring a small format, high-pixel density camera with high frame rates. This CCD has 30% more pixels than other 1K x 1K arrays. When comparing slow scan CCD cameras, these are noted for the following features. q q q q q Very small pixels, 6.8 µm High resolution, 1.3 million pixels Highest scan rates, up to 5 MHz Low noise; even at high pixel rates Extremely low dark charge

TE/CCD-1317-K RTE/CCD-1317-K PentaMAX-1317-K

100 90
Quantum Efficiency, %

CCD Cameras

80 70 60 50 40 30 20 10 0 200 300 400 500 600 700 800 900 1000 1100 Front Illuminated with UV Coating Front Illuminated

See page 46 for a non-cooled camera based on this array.

Wavelength, nm

Performance Characteristics
CCD Array: Format: Full Well Capacity: Kodak model KAF-1400 1317 (H) x 1035 (V); 8.98 x 7.04 mm overall; 6.8 x 6.8 µm pixels Imaging section, 45,000 electrons; horizontal shift register, 90,000 electrons; preamp node, 180,000 electrons 6-8 electrons at 500 kHz; 10 electrons at 1 MHz; 20 electrons at 5 MHz (PentaMAX only) 400-1080 nm; 190-1080 nm with UV-to-visible converter 500 kHz, 12 bits; 430 kHz, 13 bits, 14 bits with binning; 1 MHz, 12 bits; 5 MHz, 12 bits (PentaMAX only) <1% for 12-13 bits ±4% over entire CCD area, except blemish regions Point defect, a pixel which deviates more than 6% (about 2,000 electrons) from adjacent pixels when illuminated to 70% of saturation level; Cluster defect, a grouping of not more than 5 adjacent point defects; Low level charge trap, a charge trap that fills with less than 2,000 electrons; Column/row defect, a grouping of point defects along a single column or row; Adjacent pixels, the surrounding 100 x 100 pixels or ±50 columns/rows; brackets below indicate the maximum number of defects for the central 800 x 600 region Grade 0, no point, cluster, low level charge traps, or column defects; Grade 1, 5 [2] or fewer point defects, 1 [0] or fewer low level charge traps, no cluster or column defects; Grade 2, 10 [5] or fewer point defects, 4 [2] or fewer cluster defects, 2 [1] or fewer low level charge traps, 2 [0] or fewer column defects; Grade 3, 20 [10] or fewer point defects, 8 [4] or fewer cluster defects, 6 [3] or fewer low level charge traps, 4 [2] or fewer column defects -20°C for RTE/CCD with passive cooling; -45°C for TE/CCD, -40°C for RTE/CCD, -35°C for PentaMAX with air circulation; -60°C for TE/CCD with coolant circulation; -40°C for TE/CCD, -30°C for RTE/CCD, -25°C for PentaMAX with air circulation and nitrogen backfill ±0.040°C over entire temperature range 10-15 electrons/pixel-hour at -40°C TE/CCD, up to 1 MHz; PentaMAX, up to 5 MHz 1.4 seconds at 1 MHz; 0.30 seconds at 5 MHz (PentaMAX only) Princeton Instruments Tel: 609-587-9797

Readout Noise:

Spectral Range: Dynamic Range: Response Nonlinearity: Response Nonuniformity: Blemish Definitions:

Blemish Specifications:

Operating Temperature:

Thermostating Precision: Typical Dark Charge: Scan Rate: Full Frame Readout: 34


CCD Cameras
based on Kodak 1536 â 1024 CCD (9 µm pixels)
These cameras are based on a CCD with 15% more pixels than the Kodak 1317 â 1035, and twice the well capacity. This results in more dynamic range and better imaging of low contrast subjects. The pixels (and cost) are still smaller than the SITe 1K â 1K or the EEV 1152 â 1242. When comparing slow scan CCD cameras, these are noted for the following features. u u u u u Small pixels, 9 µm High resolution, 1.6 million pixels 1 MHz scan rates Moderate well capacity (85,000 electrons), good for low contrast imaging Extremely low dark charge

TE/CCD-1536-K

100 90
Quantum Efficiency, %

80 70 60 50 40 30 20 10 0 200 300 400 500 600 700 800 900 1000 1100 Front Illuminated with UV Coating Front Illuminated

CCD Cameras

Wavelength, nm

Performance Characteristics
CCD Array: Format: Kodak model KAF-1600 1536 (H) x 1024 (V); 13.8 x 9.2 mm overall; 9 x 9 µm pixels Imaging section, 85,000 electrons; horizontal shift register, 170,000 electrons; preamp node, 340,000 electrons 9-10 electrons at 500 kHz; 12 electrons at 1 MHz 400-1080 nm; 190-1080 nm with UV-to-visible converter 430 kHz, 13 bits, 14 bits with binning; 500 kHz, 12 bits; 1 MHz, 12 bits < 2% for 12-14 bits < ±4% over entire CCD area, except blemish regions Point defect, a pixel which deviates more than 6% (about 3,500 electrons) from adjacent pixels when illuminated to 70% of saturation level; Cluster defect, a grouping of not more than 5 adjacent point defects; Low level charge trap, a charge trap that fills with less than 2,000 electrons; Column/row defect, a grouping of point defects along a single column or row; Adjacent pixels, the surrounding 100 x 100 pixels or ±50 columns/rows; brackets below indicate the maximum number of defects for the central 800 x 600 region Grade 0, no point, cluster, or column defects; Grade 1, 5 [2] or fewer point defects, no cluster or column defects; Grade 2, 10 [5] or fewer point defects, 4 [2] or fewer cluster defects, 2 [0] or fewer column defects; For charge trap specifications, contact the factory for the latest information -45°C with air circulation; -60°C with coolant circulation; -35°C with air circulation and nitrogen backfill ±0.040°C over entire temperature range 10-20 electrons/pixel-hour at -40°C Up to 1 MHz 1.8 seconds at 1 MHz Princeton Instruments 35

Full Well Capacity:

Readout Noise: Spectral Range: Dynamic Range:

Response Nonlinearity: Response Nonuniformity: Blemish Definitions:

Blemish Specifications:

Operating Temperature:

Thermostating Precision: Typical Dark Charge: Scan Rate: Full Frame Readout: Fax: 609-587-1970


CCD Cameras
based on Kodak 2K â 2K CCD
These new cameras offer high spatial resolution and performance, combined with an excellent blemish specification, ultra low dark charge, high sensitivity, moderate dynamic range (12 bits), and high readout rates. When comparing various slow scan CCD cameras, these are noted for the following features. u u u u u Small pixels, 9 µm High resolution, 4.1 million pixels Highest scan rates, up to 5 MHz Low noise; high sensitivity Extremely low dark charge

TE/CCD-2033-K PentaMAX-2033-K

100 90
Quantum Efficiency, %

CCD Cameras

80 70 60 50 40 30 20 10 0 200 300 400 500 600 700 800 900 1000 1100 Front Illuminated with UV Coating Front Illuminated

Wavelength, nm

Performance Characteristics
CCD Array: Format: Full Well Capacity: Kodak model KAF-4200 2044 (H) x 2033 (V); 18.4 x 18.3 mm overall; 9 x 9 µm pixels Imaging section, 85,000 electrons; horizontal shift register, 170,000 electrons; preamp node, 340,000 electrons 9-10 electrons at 500 kHz; 12 electrons at 1 MHz; 20-24 electrons at 5 MHz (PentaMAX only) 400-1080 nm; 190-1080 nm with UV-to-visible converter 430 kHz, 13 bits, 14 bits with binning; 500 kHz, 12 bits; 1 MHz, 12 bits; 5 MHz, 12 bits (PentaMAX only) < 2% for 12-14 bits, TE/CCD; contact the factory for PentaMAX information < ±4% over entire CCD area, except blemish regions Point defect, a pixel which deviates more than 6% (about 3,500 electrons) from adjacent pixels when illuminated to 70% of saturation level; Cluster defect, a grouping of not more than 5 adjacent point defects; Low level charge trap, a charge trap that fills with less than 4,000 electrons; Column/row defect, a grouping of point defects along a single column or row; Adjacent pixels, the surrounding 100 x 100 pixels or ±50 columns/rows; brackets below indicate the maximum number of defects for the central 1024 x 1024 region Grade 0, no point, cluster, low level charge traps, or column defects; Grade 1, 15 [6] or fewer point defects, 3 [0] or fewer low level charge traps, no cluster or column defects; Grade 2, 30 [15] or fewer point defects, 12 [6] or fewer cluster defects, 6 [3] or fewer low level charge traps, 6 [0] or fewer column defects; Grade 3, 60 [30] or fewer point defects, 24 [12] or fewer cluster defects, 18 [9] or fewer low level charge traps, 12 [6] or fewer column defects -40°C for TE/CCD with air circulation; -55°C for TE/CCD with coolant circulation; -30°C for TE/CCD with air circulation and nitrogen backfill; For PentaMAX cooling information, contact the factory ±0.040°C over entire temperature range 10-20 electrons/pixel-hour at -40°C TE/CCD, up to 1 MHz; PentaMAX, up to 5 MHz 4.2 seconds at 1 MHz; 0.95 seconds at 5 MHz (PentaMAX only) Princeton Instruments Tel: 609-587-9797

Readout Noise:

Spectral Range: Dynamic Range: Response Nonlinearity: Response Nonuniformity: Blemish Definitions:

Blemish Specifications:

Operating Temperature:

Thermostating Precision: Typical Dark Charge: Scan Rate: Full Frame Readout:

36


CCD Cameras
based on Kodak 3K â 2K and 4K â 4K CCDs
These new cameras offer the highest spatial resolution and performance, combined with an excellent blemish specification, ultra low dark charge, high sensitivity, moderate dynamic range (12 bits), and high readout rates. When comparing various slow scan CCD cameras, these are noted for the following features. q q q q q Small pixels, 9 µm Highest resolution, 6.3 or 16.8 million pixels Highest scan rates, up to 5 MHz Low noise; high sensitivity Extremely low dark charge

PentaMAX-3072-K PentaMAX-4096-K

100 90
Quantum Efficiency, %

80 70 60 50 40 30 20 10 0 200 300 400 500 600 700 800 900 1000 1100 Front Illuminated with UV Coating Front Illuminated

CCD Cameras

Wavelength, nm

Performance Characteristics
CCD Array: Format: Full Well Capacity: Readout Noise: Spectral Range: Dynamic Range: Response Nonlinearity: Response Nonuniformity: Blemish Definitions: Kodak model KAF-6300 or KAF-16800 3072 â 2048 or 4096 â 4096; 27.65 x 18.43 mm or 36.86 â 36.86 mm overall; 9.0 â 9.0 µm pixels Imaging section, 85,000 electrons 12 electrons at 1 MHz, 20-24 electrons at 5 MHz for PentaMAX-3072-K; for PentaMAX-4096-K contact the factory 400-1080 nm; 190-1080 nm with UV-to-visible converter 12 bits for PentaMAX-3072-K, for PentaMAX-4096-K contact the factory Contact the factory ±4% over entire CCD area, except blemish regions Point defect, a pixel which deviates more than 6% (about 3,500 electrons) from adjacent pixels when illuminated to 70% of saturation level; Cluster defect, a grouping of not more than 5 adjacent point defects; Low level charge trap, a charge trap that fills with less than 2,000 electrons; Column/row defect, a grouping of point defects along a single column or row; Adjacent pixels, the surrounding 100 x 100 pixels or ±50 columns/rows; brackets below indicate the maximum number of defects for the central 1024 x 1024 region For PentaMAX-3072-K, Grade 0, no point, cluster, low level charge traps, or column defects; Grade 1, 22 [9] or fewer point defects, no cluster or column defects; Grade 2, 45 [22] or fewer point defects, 18 [9] or fewer cluster defects, 9 [0] or fewer column defects; Grade 3, 90 [45] or fewer point defects, 36 [16] or fewer cluster defects, 18 [9] or fewer column defects; For charge trap specifications, contact the factory for the latest information; Custom blemish specifications to meet individual user requirements can be negotiated with Kodak for this CCD; For PentaMAX-4096 contact the factory For the PentaMAX-4096-K specified by Kodak to be at least 0.99997 per shift at 2 MHz. Contact the factory for more details on the ramifications of the CTE. Contact the factory ±0.040°C over entire temperature range 10-20 electrons/pixel-hour at -40°C for PentaMAX-3072-K; for PentaMAX-4096-K contact the factory 5 MHz 1.4 seconds at 5 MHz for PentaMAX-3072-K; 3.9 seconds at 5 MHz for PentaMAX-4096-K Princeton Instruments 37

Blemish Specifications:

Charge Transfer Efficiency: Operating Temperature: Thermostating Precision: Typical Dark Charge: Scan Rate: Full Frame Readout:

Fax: 609-587-1970


CCD Cameras
based on PI 1000 â 800 CCD, back illuminated
This camera is ideal for microscopy imaging where high sensitivity and small pixels are required. The QE in the green is the highest of any imaging CCD available, making this camera the new standard for the low light fluorescent microscopy. When comparing various slow scan CCD cameras, these are noted for the following features. q Small pixels, 15 µm q High quantum efficiency (back illuminated) q Highest resolution, 800,000 pixels q MPP for low dark charge q Video output and front panel control support available using ST-133 controller

TE/CCD-1000PB

100
Estimated Quantum Efficiency, %

90 80 70 60 50 40 30 20 10 0 200 300 400 500 600 700 VISAR

Lumogen & VISAR

CCD Cameras

800

900

1000

1100

Wavelength, nm

Performance Characteristics
CCD Array: Format: Full Well Capacity: Readout Noise: Spectral Range: Dynamic Range: Response Nonlinearity: Response Nonuniformity: Blemish Definitions: Princeton Instruments CCD-1000PB, MPP and back illuminated only 1000 â 800, 15 â 12 mm overall, 15 x 15 µm pixels Imaging section, 60,000-80,000 electrons; horizontal shift register, 1,000,000 electrons 6-8 electrons at 100 kHz; 20 electrons at 500 kHz 260-1080 nm; 190-1080 nm with UV-to-visible converter 12 bits at 500 kHz; 13-14 bits at 100 kHz without binning, 16 bits with binning <1% for 12 bits Contact the factory Point defect, a pixel with significantly more or significantly less signal than adjacent pixels; Cluster defect, a grouping of adjacent point defects; Column defect, a grouping of point defects along a single column; Dark defects (point, cluster, or column), >50% reduction in response relative to adjacent pixels; Hot defects (point, cluster, or column), dark current >10â the maximum allowable average dark current. Contact the factory -40°C with air circulation, -55°C with coolant circulation, -30°C with air circulation and nitrogen backfill ±0.040°C over entire temperature range 0.70 electrons/pixel-second at -40°C up to 500 kHz Contact the factory Princeton Instruments Tel: 609-587-9797

Blemish Specifications: Operating Temperature:

Thermostating Precision: Typical Dark Charge: Scan Rate: Full Frame Readout: 38


CCD Cameras
based on SITe 512 â 512 CCD, front or back illum.
These cameras are ideal for light starved, long integration applications such as astronomy, where high quantum efficiency is a must. Permanent coatings extend the range of the camera into the UV, still with the highest QE of any camera in this wavelength range. When comparing slow scan CCD cameras, these are noted for the following features. u Large pixels (24 µm), front or back illuminated, maximum light collection u MPP operation for low dark charge u Square format, 512 x 512 u Exceptionally high quantum efficiency (back illuminated) u High dynamic range (16 bits) u Moderate cost, compared to 1K â 1K

TE/CCD-512SF&SB LN/CCD-512SF&SB

100 90
Quantum Efficiency, %

80 70 60 50 40 30 20 10

Back Illuminated with UV/AR Coating

Back Illuminated with VIS/AR Coating

CCD Cameras

Front Illuminated

0 200

300

400

500

600

700

800

900

1000

1100

Wavelength, nm

Performance Characteristics
CCD Arrays: Format: Full Well Capacity: SITe (formerly Tektronix) model SI502FA (front illuminated) or SI502BA (back illuminated), MPP only 512 x 512; 12.3 x 12.3 mm overall; 24 x 24 µm pixels Standard array image section, 300,000-350,000 electrons; horizontal shift register, 750,000 electrons; preamp node, 900,000 electrons 4-7 electrons at 50 kHz; 18-21 electrons at 1 MHz 512SF, 400-1080 nm, 190-1080 nm with UV-to-visible converter; 512SB, 260-1080 nm, 190-1080 nm with UV-to-visible converter 14-17 bits < 1.5% for 16 bits; < 2.5% for 17 bits ±4% over entire CCD area, except blemish regions Point defect, a pixel with significantly more or significantly less signal than adjacent pixels; Cluster defect, a grouping of adjacent point defects; Column defect, a grouping of point defects along a single column; Dark defects (point, cluster, or column), >50% reduction in response relative to adjacent pixels; Hot defects (point, cluster, or column), dark current >10â the maximum allowable average dark current. 10 or fewer point defects, no partial or full column defects; Except column 512 may not meet full specifications Lower cost, SITe grade 1 CCDs are also available, please consult the factory for details TE/CCD, -40°C with air circulation, -60°C with coolant circulation, -30°C with air circulation and nitrogen backfill; LN/CCD, -80°C to -140°C ±0.040°C over entire temperature range 512SF, < 1 electrons/pixel-second at -50°C; LN/CCD at -120°C, < 0.6 electron/pixel-hour; Somewhat higher levels may be experienced with the UV/AR coating TE/CCD, up to 1 MHz; LN/CCD maximum scan rate is 100 to 150 kHz 512SF, 0.29 seconds at 1 MHz; 512SB, 0.5 seconds at 1 MHz > 10 hours for standard dewar; > 25 hours for large capacity dewar Princeton Instruments 39

Readout Noise: Spectral Range: Dynamic Range: Response Nonlinearity: Response Nonuniformity: Blemish Definitions:

Blemish Specifications:

Operating Temperature: Thermostating Precision: Typical Dark Charge:

Scan Rate: Full Frame Readout: Liquid Nitrogen Hold Time:

Fax: 609-587-1970


CCD Cameras
based on SITe 1K â 1K CCD, front or back illum.
These cameras are ideal for light starved, long integration applications such as astronomy, where high quantum efficiency is a must. Permanent coatings extend the range of the camera into the UV. When comparing slow scan CCD cameras, these are noted for the following features. u Exceptionally high quantum efficiency (back illuminated) u Large pixels, 24 µm, front or back illuminated, maximum light collection u Large well capacity u High resolution, particularly for back illuminated operation, 1024 x 1024 u High dynamic range, 16 bits See pages 74, 75, and 78-79 for other cameras based on this CCD.

TE/CCD-1024SF&SB LN/CCD-1024SF&SB

100 90
Quantum Efficiency, %

CCD Cameras

80 70 60 50 40 30 20 10

Back Illuminated with UV/AR Coating

Back Illuminated with VIS/AR Coating

Front Illuminated

0 200

300

400

500

600

700

800

900

1000

1100

Wavelength, nm

Performance Characteristics
CCD Arrays: Format: Full Well Capacity: Readout Noise: Spectral Range: Dynamic Range: Response Nonlinearity: Response Nonuniformity: Blemish Definitions: SITe (formerly Tektronix) model ST-003, front or back illuminated, MPP only 1024 x 1024; 24.6 x 24.6 mm overall; 24 x 24 µm pixels 300,000-350,000 electrons Front illuminated, 4-6 electrons at 50 kHz, 24 electrons at 1 MHz; Back illuminated, 4-8 electrons at 50 kHz, 13 electrons at 500 kHz 1024SF, 400-1080 nm, 190-1080 with UV-to-visible converter 1024SB, 260-1080 nm, 190-1080 with UV-to-visible converter 14-16 bits < 2% for 17 bits; < 3% for 18 bits with binning ±6% over entire CCD area, except blemish regions Point defect, a pixel with significantly more or significantly less signal than adjacent pixels; Cluster defect, a grouping of adjacent point defects; Column defect, a grouping of point defects along a single column; Dark defects (point, cluster, or column), >50% reduction in response relative to adjacent pixels; Hot defects (point, cluster, or column), dark current >10â the maximum allowable average dark current; brackets below indicate the maximum number of defects for the central 750 x 750 region Grade 1, 40 [10] or fewer point defects, 4 [2] or fewer cluster defects, 2 [0] or fewer column defects; Grade 2, 60 [30] or fewer point defects, 20 [5] or fewer cluster defects, 6 [4] or fewer column defects TE/CCD, -40°C with air circulation, -50°C with tap water circulation, -60°C with coolant circulation; LN/CCD, -80°C to -130°C ±0.040°C over entire temperature range TE/CCD, 0.2-0.6 electrons/pixel-second for 1024SF at -50°C; 0.6-2 electrons/pixel-second for 1024SB at -50°C; LN/CCD, < 1 electron/pixel-hour at -120°C; Somewhat higher levels may be experienced with the UV/AR coating 25 to 100 kHz user selectable, 14 to 17 bits; 100 to 430 kHz, 13 to 14 bits; For 1 MHz operation, contact the factory; LN/CCD maximum scan rate is 100 to 150 kHz 1024SF, 1.2 seconds at 1 MHz; 1024SB, 2.32 seconds at 500 kHz > 9 hours for standard dewar; > 25 hours for large capacity dewar Princeton Instruments Tel: 609-587-9797

Blemish Specifications:

Operating Temperature: Thermostating Precision: Typical Dark Charge:

Scan Rate:

Full Frame Readout: Liquid Nitrogen Hold Time:

40


The PentaMAX Camera
The Princeton Instruments PentaMAX System is a high performance 5 MHz 12 bit cooled CCD camera. By incorporating ver y good cooling and state of the art electronics, this camera achieves record low noise levels. Images can be transferred to the computer via a high speed serial link for display and storage, while simultaneously being displayed on a standard video monitor.
The Princeton Instruments PentaMAX Camera combines both high speed and high precision readout in a single camera. The high speed mode collects 12 bit images at a readout rate of 2.5 or 5 million pixels per second (2.5 or 5 MHz). Instantly switch to precision mode to collect lower noise data at 1 MHz. CCD arrays up to 4K â 4K are supported. Frame transfer CCDs in this camera can sustain up to 150 frames per second with binning. Data is transferred directly to the memory of the computer via a high speed link and DMA. A frame buffer with standard video (RS-170 or CCIR) output is also incorporated. The PentaMAX Camera offers high performance thermoelectric cooling, to support long exposure times and to minimize readout noise. The PentaMAX Camera is a fully integrated package, containing analog and digital electronics, thermoelectric cooler with air and water options, and timing hardware in a single, shielded housing. The latest advances in surface-mount technology allow a small package without compromising many of the advanced hardware programming features found in other Princeton Instruments systems.

CCD Cameras

The Princeton Instruments PentaMAX Camera. Dual A/D channels, thermoelectric cooling, and timing hardware are all found in this integrated package. between the two channels is completely under software control, for total experiment automation. salesperson for the latest information on this list). High data rates require high speed data transfer. The PentaMAX Camera provides data transfer at full speed and resolution direct to the memory of the computer operating the camera. This high speed DMA is available for IBM AT compatible computers. Support for Sun and Silicon Graphics, Inc computers is provided by adapters, as found on pages 101 and 102. Camera focusing and alignment require the highest possible readout rate. To support this, the digital data must be displayed in real time with minimum delay, a feat that becomes increasingly difficult through the computer at these high pixel rates. The PentaMAX is therefore available with a built-in frame buffer and standard video output circuitry. This allows data to be seen as soon it is read from the CCD. This is a major advance over earlier cooled CCD camera designs, and it is a part of our approach of providing a whole system compatible with high speed. 41

High Speed Supported Through the Whole System
The performance of a high speed CCD camera is a function of many factors in addition to the A/D rate These include the ability of the CCD to be read out at high speeds and the ability to transfer the data to computer RAM in real time. This camera system provides high speed in each of these categories, so the total system provides high throughput. To achieve high camera speed, the CCD array itself must support high pixel rates. This generally requires a dual stage output amplifier on the array. Since not all scientific CCD arrays have this, not all scientific CCD arrays can be read out at MHz rates. For the PentaMAX Camera, only CCD arrays that can pixel rates (see the CCD following page or contact PI has selected maintain high chart on the your regional

Dual Speed Operation
Read noise of CCD arrays always increases with pixel rate, so it is often necessary to trade off temporal resolution for high dynamic range. This is sometimes accomplished by operating a high speed A/D at slower readout rates, but the readout noise of these A/Ds limits the overall dynamic range. The PentaMAX Camera has been designed specifically to solve these experimental problems without compromise. The PentaMAX Camera incorporates two complete analog channels including separate A/D converters to provide optimum signal to noise ratios for both 1 MHz and either 2.5 or 5 MHz readout. Switching Fax: 609-587-1970

Princeton Instruments


Computer Interface
The PentaMAX Camera can be interfaced to Sun, Silicon Graphics, Inc, MacIntosh and IBM AT compatible computers (requires an EISA bus for high speed data transfer). An EISA computer can support a sustained readout rate of over 5 million pixels/sec. Contact the factory to ensure computer compatibility, or for a list of compatible computer models. All PentaMAX Cameras are available with interfaces to EISA bus IBM AT compatible computers. Because this interface transfers data immediately into the memory of the computer, information is available immediately for analysis. Data transfer is accomplished through a high speed serial or fiber optic cable. With the optional fiber optic output, digital data can be transferred to a computer up to 2 kilometers away. Since the output signal is already digitized, no signal degradation occurs.

CCD Manufacturer Kodak Kodak Kodak Kodak Kodak EEV EEV

Pixels 768 1317 2033 3072 4096 512 â 288 â â 512 â 1035 â 2044 â 2048 â 4096 512 (FT) 384 (FT)

Image Area, mm 6.91 8.98 18.4 27.6 36.9 7.7 6.3 â â â â â â â 4.6 7.04 18.3 18.4 36.9 7.7 8.4

See Page 3 3 3 3 3 2 3 2 4 6 7 7 8 1

CCD Cameras

These CCD arrays are available in the Princeton Instruments PentaMAX Camera. Contact your sales representative for information on other CCDs.
tested without having to digitize all the pixels of the array. Completely flexible exposure, also set through software, is supported as described below. is required to achieve the flattest possible background level. An LCD display, unique to this system, shows either the actual or the target temperature. You can now monitor the precise temperature of the CCD array, and begin work as soon as you have reached the target temperature. This thermal stability is most important for the repeatability of experiments, when, for example, a background image is taken for subtraction from later data. Once cooling has been achieved the CCD temperature is thermally regulated to within ±0.040°C. Heat dissipation for the thermoelectric cooling system is provided by either forced air or water circulation. For maximum flexibility, all models come equipped for both of these options. Air cooling provides maximum convenience and is sufficient for many applications requiring only short exposures. Water cooling allows the CCD to be operated at lower temperatures, for minimum dark current when acquring images with longer exposure times.

Temperature Control
As with all Princeton Instruments cameras, the PentaMAX cools the CCD to minimize thermal noise and maximize sensitivity. By advanced mechanical design, we are able to achieve lower CCD temperatures with air cooling than other manufacturers achieve with water cooling. With water cooling, we can go even lower. Because about 1% of pixels on a CCD have anomalously high dark current, extra cooling power

Sophisticated CCD Readout
Extremely flexible readout of the CCD is supported through a new electronic design. Readout modes supported include full resolution, simultaneous multiple subimages, and nonuniform binning. Arbitrary, software defined regions of interest can also be

Standard Video Output
The PentaMAX Camera is available with a digital scan converter for standard composite video output (choice of RS-170 or CCIR standards). Data can be transferred to the computer, to the video subsystem, or to both simultaneously. Because the video system includes a digital frame store, low frame rates can be used to achieve high sensitivity without incurring image flicker. Higher frame rates can also be accomodated easily by this system. Not all pixels digitized need to be displayed by the video monitor. Thus on a large CCD, where the CCD resolution exceeds the capability of the video output, selected pixels can be written to the video buffer while all the pixels are written to the computer. Thus the video buffer can show the whole Tel: 609-587-9797

All CCD arrays available in the PentaMAX Camera can be read at rates of several million pixels per second. Kodak KAF-1400 array shown.
42 Princeton Instruments


The mechanical shutter built into the camera can also be synchronized in a number of ways depending on the ambient light of the experiment.

Shuttering
PentaMAX Cameras incorporate a fast mechanical shutter to keep light from the CCD array during readout. The small size and mechanical isolation of the shutter allows it to be operated at several Hz for rapid sequences of images. If necessary, the shutter can be temporarily disabled by the user.

CCD Cameras

Advanced System Design
To meet all its high speed requirements, this camera uses a somewhat different system configuration from Princeton Instruments camera systems based on Model ST-130, ST-133, or ST-138 Camera Controllers. In the PentaMAX System all signal processing occurs in the camera head, and digital data is then sent directly to the computer. An available fiber optic interface for this camera allows separation of camera and computer by as much as 2 kilometers. Power for the system is provided by a separate module which can be located up to 3 meters from the camera head (further with a special cable). This module also displays the current CCD temperature, to facilitate immediate image collection as soon as the CCD reaches the desired operating temperature.

The PentaMAX system is ideal for applications requiring a high frame rate and a high dynamic range. Nikon Diophot 300 courtesy of Micron Optics, Parsippany, NJ. CCD array at partial resolution, or part of the CCD array at full resolution. Charge binning in the camera hardware can also be used to match the resolution of a CCD to that of the video buffer, with a corresponding increase in sensitivity of the CCD array. Video output also allows data from this extremely sensitive cooled camera to be recorded on video tape and other media, when a very large number of frames must be stored. It also provides easy integration of images from this camera with those taken by other cameras. The video display provides 8 bit gray scale resolution, which exceeds the capability of most monitors. Since the CCD data have a dynamic range of 12 bits or more, a downloadable look up table (LUT) is available to map the desired portion of the 12 bit range from the camera to the video visible range. The mapping can be nonlinear (for compression or gamma correction) or even reversed (for negative video) if the user so desires.

Triggering
The PentaMAX Camera can run with or without external triggering. In an external trigger mode, the camera hardware is programmed to collect images once a trigger is received. Data collection is therefore not dependent on software speed limitations. Any number of frames can be programmed to be collected by the camera without software intervention. This camera can also be used to trigger other experimental equipment, for full experimental control via software. A signal from the camera as the exposure begins starts the experiment. Programs created under Princeton Instruments WinView software package can process images as they are collected or store them for later processing.

Information Updates
This camera system, one of the newest from Princeton Instruments, is subject to continuous improvements. All data in this description should therefore be considered preliminary. Contact your local sales representative or the factory for the latest information on specific PentaMAX Camera parameters, including available CCDs.

Fax: 609-587-1970

Princeton Instruments

43


The Small Cooled Camera
CCD Cameras

The Princeton Instruments line of Small Cooled CCD Cameras, designed for use with the ST-133 Camera Controller, provide high performance imaging in a compact, lightweight camera. These cameras provide the same advanced preamplifier circuitry as other Princeton Instruments cameras, with cooling sufficient for exposures as long as one minute. The new cameras are operated through Princeton Instruments WinView software package.
Princeton Instruments complete line of cooled CCD cameras now includes compact, thermoelectrically cooled cameras available in several formats and with several options for heat dissipation. These cameras offer the same low-noise readout as the TE/CCD product line, and with sufficient cooling to render dark charge negligible for exposures of several seconds or more, depending on the exact model purchased. These cameras are easily mounted on a fluorescence microscope or other optical equipment. They provide high resolution digital imaging with incredible sensitivity in a compact, economical package.

Cooling Options
The small cooled camera models are available with three options for dissipating the heat generated by thermoelectric cooling. For only minimal cooling, the RTE (round, thermoelectric, passive) Model shown in the photograph below uses only radiating fins to dissipate heat. A second RTE model uses a small fan and a camera shroud for forced air cooling. For maximum cooling and therefore minimal dark charge, Princeton Instruments recommends use of the forced air model. These options give you the flexibility to choose a system that is optimized for your needs. The performance of these cooling options varies with the size of the CCD and the temperature of the air surrounding the camera. See the data sheets earlier in this catalog section for cooling specifics on these options.

Mechanically Compact
The small cooled line of cameras is designed in a cylindrical format, approximately 13 cm in diameter. The length of the camera depends both on the cooling option (passive air or forced air) and the optical interface. Available optical interfaces are listed on the following page. Camera length is also shown in the table at the end of this section. The distance shown includes camera body, camera back, and lens mount.

CCD Arrays Supported
This line of cameras supports a growing list of high performance CCD arrays. As of this writing, models available are shown in the table on the following page. More CCDs are being added to this list in the near future, as the CCDs become available. The Princeton Instruments Small Cooled Camera. Its lightweight design and compact size make it the perfect camera for general imaging applications. 44 Princeton Instruments Frame transfer cameras have the advantage that they do not require a mechanical Tel: 609-587-9797


Model Number RTE/CCD-1536-K RTE/CCD-1317-K RTE/CCD-768-K RTE/CCD-512EFT

Array Kodak 1536 â 1024 Kodak 1317 x 1035 Kodak 768 x 512 EEV 512 â 512

Pixel Size, microns 9.0 â 9.0 6.8 x 6.8 9.0 x 9.0 15.0 â 15.0

Imaging Area, mm Illumination 13.8 â 9.2 8.98 x 7.04 6.91 x 4.6 7.7 â 7.7 Front Front Front Front

MPP Yes Yes Yes No

Small Cooled Cameras currently available from Princeton Instruments. Contact the factory for the most up to date information on new cameras in this format, including frame transfer arrays for shutterless operation.
shutter, eliminating all moving parts in the passive cooled version of the camera. The EEV 512 â 512 device, listed above, offers a larger pixel than the Kodak devices, and is a masked frame transfer device. Most performance features of these cameras are described in the corresponding data sheet for the TE/CCD camera, including features such as readout rate, spectral range, response nonlinearity, and response nonuniformity. Since these cameras provide less cooling that the corresponding TE/CCD models, the dark current will be higher, typically a factor of two for each 6-7°C. Readout noise, also a function of temperature, may be a bit higher, typically 20-50% above the levels achieved at -60°C. See the appropriate data sheet for cooling specifications for the model of interest. For many applications the benefits of compact size, low weight, passive cooling, and economy outweigh these small differences in performance.

CCD Cameras

Compatibility with PI Controllers and Software
The Small Cooled Cameras operate with the Model ST-133 Camera Controller. This compact, high performance controller provides dual A/D converters optimized for high speed and high dynamic range. The Small Cooled Camera and ST-133 combination support readout rates of up to 1 million pixels per second, with a dynamic range of 12 bits (4,000:1) at this speed. A slower A/D and separate analog channel provide the lowest possible noise readout. Dynamic range is a function of CCD as well as operating speed, so the ST-133 provides up to 16 bits dynamic range with its slow channel, depending on the device. Control of the Small Cooled Camera is provided under Microsoft Windows 3.1 by the WinView image acquisition and processing software. For Macintosh computers they are also compatible with the IPLab software through a customized image acquisition extension. For OEM customers, libraries are also available for linking with customer written software. See the Software section of this catalog for detailed descriptions.

Shuttering
With most slow scan CCD cameras except frame transfer devices, a shutter is required to prevent smearing due to exposure during readout of the array. All models except frame transfer models are provided with a mechanical shutter. This includes the C mount cameras. This mechanical shutter is well isolated from incurring any vibrations in the camera head. For special applications the user can easily disable the shutter, keeping it open throughout an experiment. For experiments requiring special shutters such as focal plane devices, a connector supports operation of an external shutter elsewhere in the system. This shutter is fully compatible with all Princeton Instruments hardware and software, and its open and closing times are automatically compensated for during acquisition and readout of the CCD. Fax: 609-587-1970

The Small Cooled Camera is fully compatible with microscopes from all major manufacturers, including Nikon, Olympus, Zeiss, and the Leica group. Equipment shown above courtesy of Carl Zeiss, Inc.

Optical Interfacing
This line of cameras is available with either Canon, Nikon F mount or standard C mount lens adapters. Contact the factory for more information, including specifications on microscope and other adapters. Canon Mount 109 mm 144 mm 44.14 mm C-Mount 79 mm 112 mm 17.5 mm

Nikon F Mount Length, no fan Length, with fan Focal depth 111 mm 146 mm 46.5 mm

Length and focal depth of Princeton Instruments Small Cooled Cameras. Note that the C-mount cameras can now be offered with a mechanical shutter.
Princeton Instruments 45


CCD Cameras
based on Kodak 1317 â 1035 CCD (6.8 µm pixels)
The MicroView system is designed as a high resolution electronic still photography system. Although designed mostly for qualitative imaging, the MicroView can be used for quantitative analysis when short (< 500 msec) exposures are all that will ever be required. When comparing CCD systems, the MicroView is noted for the following. q q q q Very small pixels, 6.8 µm High resolution, 1.3 million pixels Minimal cooling MicroView package includes software for basic imaging collection

MicroView-1317

100 90
Quantum Efficiency, %

CCD Cameras

80 70 60 50 40 30 20 10 0 200 300 400 500 600 700 800 900 1000 1100 Front Illuminated with UV Coating Front Illuminated

Wavelength, nm

Performance Characteristics
CCD Array: Format: Full Well Capacity: Kodak model KAF-1400 1317 (H) x 1035 (V); 8.98 x 7.04 mm overall; 6.8 x 6.8 µm pixels Imaging section, 45,000 electrons; horizontal shift register, 90,000 electrons; preamp node, 180,000 electrons 10 electrons at 1 MHz 400-1080 nm; 190-1080 nm with UV-to-visible converter 12 bits <1% for 12 bits ±4% over entire CCD area, except blemish regions Point defect, a pixel which deviates more than 6% (about 2,000 electrons) from adjacent pixels when illuminated to 70% of saturation level; Cluster defect, a grouping of not more than 5 adjacent point defects; Low level charge trap, a charge trap that fills with less than 2,000 electrons; Column/row defect, a grouping of point defects along a single column or row; Adjacent pixels, the surrounding 100 x 100 pixels or ±50 columns/rows; brackets below indicate the maximum number of defects for the central 800 x 600 region Grade 0, no point, cluster, low level charge traps, or column defects; Grade 1, 5 [2] or fewer point defects, 1 [0] or fewer low level charge traps, no cluster or column defects; Grade 2, 10 [5] or fewer point defects, 4 [2] or fewer cluster defects, 2 [1] or fewer low level charge traps, 2 [0] or fewer column defects; Grade 3, 20 [10] or fewer point defects, 8 [4] or fewer cluster defects, 6 [3] or fewer low level charge traps, 4 [2] or fewer column defects +8°C ±0.040°C over entire temperature range Contact the factory 1 MHz only; no video option available 1.4 seconds MicroView software package offers basic image collection only Princeton Instruments Tel: 609-587-9797

Readout Noise: Spectral Range: Dynamic Range: Response Nonlinearity: Response Nonuniformity: Blemish Definitions:

Blemish Specifications:

Operating Temperature: Thermostating Precision: Typical Dark Charge: Scan Rate: Full Frame Readout: Software: 46