N N piX 1 Ns Nd NR G2

This form of the S/N equation is essentially the same as that given above, but two additional terms have been added. The first term, (1 + npix/nB), provides a measure of the noise incurred as a result of any error introduced in the estimation of the background level on the CCD image. The term is the total number of background pixels used to estimate the mean background (sky) level. One can see that small values of will introduce the largest error as they will provide a poor estimate of the mean level of the background distribution. Thus, very large values of are to be preferred but clearly some trade-off must be made between providing a good estimate of the mean background level and the use of pixels from areas on the CCD image that are far from the source of interest or possibly of a different character.

The second new term added into the complete S/N equation accounts for the error introduced by the digitization noise within the A/D converter. From our discussion of the digitization noise in Chapter 3, we noted that the error introduced by this process can be considerable if the CCD gain has a large value. In this term, G2/ G is the gain of the CCD (in electrons/ADU) and u/ is an estimate of the 1 sigma error introduced within the A/D converter1 and has a value of approximately 0.289 (Merline & Howell, 1995).

In practice for most CCD systems in use and for most observational projects, the two additional terms in the complete S/N equation are often very small error contributors and can be ignored. In the instances for which they become important - for example, cases in which the CCD gain has a high value (e.g., 100 electrons/ADU), the background level can only be estimated with a few pixels (e.g., less than 200), or the CCD data are of poor pixel

1 The parameter u/ and its value depend on the actual internal electrical workings of a given A/D converter. We assume here that for a charge level that is half way in between two output ADU steps (that is, 1/2 of a gain step), there is an equal chance that it will be assigned to the lower or to the higher ADU value when converted to a digital number. See Merline & Howell (1995) for further details.

sampling (see Section 5.9) - ignoring these additional error terms will lead to an overestimation of the S/N value obtained from the CCD data.

Let us work through an example of a S/N calculation given the following conditions. A 300-second observation is made of an astronomical source with a CCD detector attached to a 1-m telescope. The CCD is a Thomson 1024 x 1024 device with 19-micron pixels and it happens that in this example the telescope has a fast /-ratio such that the plate scale is 2.6 arcsec/pixel.1 For this particular CCD, the read noise is 5 electrons/pixel/read, the dark current is 22 electrons/pixel/hour, and the gain (G) is 5 electrons/ADU. Using 200 background pixels surrounding our object of interest from which to estimate the mean background sky level, we take a mean value for NB of 620 ADU/pixel. We will further assume here (for simplicity) that the CCD image scale is such that our source of interest falls completely within 1 pixel (good seeing!) and that after background subtraction (see Section 5.1), we find a value for N* of 24 013 ADU. Ignoring the two additional minor error terms discussed above (as the gain is very small and = 200 is quite sufficient in this case), we can write the CCD Equation as

N = -24013(ADU) ■ G + (1) ■ (620(ADU) ■ G + 1.8 + 52(e-))'

Note that all of the values used in the calculation of the S/N are in electrons, not in ADUs. The S/N value calculated for this example is ~ 342, a very high S/N. With such a good S/N measurement, one might suspect that this is a bright source. If we compare VN with all the remaining error terms, we see that indeed this measurement has its noise properties dominated by the Poisson noise from the source itself and the expression S/N ~ VN = 346 works well here.

While the S/N of a measurement is a useful number to know, at times we would prefer to quote a standard error for the measurement as well. Using the fact that S/N = 1/a, where a is the standard deviation of the measurement, we can write

^magnitudes n .

In this expression, p is equal to npix(1 + npix/nB)(NS + ND + N| + G2a/), the same assumptions apply concerning the two "extra" error terms, and the value of 1.0857 is the correction term between an error in flux (electrons) and that same error in magnitudes (Howell, 1993). We again see that if the Poisson error of N* itself dominates, the term p can be ignored and this equation

1 Using the results from Section 4.1, what would be the /-ratio of this telescope?

reduces to that expected for a error estimate in the limiting case of a bright object.

Additionally, one may be interested in a prediction of the S/N value likely to be obtained for a given CCD system and integration time. N, is really N ■ t, where N is the count rate in electrons (photons) per second (for the source of interest) and t is the CCD integration time. Noting that the integration time is implicit in the other quantities as well, we can write the following (Massey, 1990):

S Nt

in which we have again ignored the two minor error terms. This equation illustrates a valuable rule of thumb concerning the S/N of an observation: S/NaVi, not to t itself. Solving the above expression for t we find

where A = N2' B = -(S/N)2(N + «^(N + ND)), and C = -(S/N)2 n^N*. Most instrument guides available at major observatories provide tables that list the count rate expected for an ideal star (usually 10th magnitude and of 0 color index) within each filter and CCD combination in use at each telescope. Similar tables provide the same type of information for the observatory spectrographs as well. The tabulated numeric values, based on actual CCD observations, allow the user, via magnitude, seeing, or filter width, to scale the numbers to a specific observation and predict the S/N expected as a function of integration time.

4.5 Basic CCD data reduction

The process of standard CCD image reduction makes use of a basic set of images that form the core of the calibration and reduction process (Gullixson, 1992). The types of images used are essentially the same (although possibly generated by different means) in imaging, photometric, and spectroscopic applications. This basic set of images consists of three calibration frames -bias, dark, and flat field - and the data frames of the object(s) of interest. Table 4.1 provides a brief description of each image type and Figures 4.2-4.5

CCD imaging Table 4.1. Types of CCD images

CCD Image Type Image Description

Bias This type of CCD image has an exposure time of zero seconds. The shutter remains closed and the CCD is simply read out. The purpose of a bias or zero frame is to allow the user to determine the underlying noise level within each data frame. The bias value in a CCD image is usually a low spatial frequency variation throughout the array, caused by the CCD on-chip amplifiers. This variation should remain constant with time. The rms value of the bias level is the CCD read noise. A bias frame contains both the DC offset level (overscan) and the variations on that level. The nature of the bias variations for a given CCD are usually column-wise variations, but may also have small row-wise components as well. Thus, a 2-D, pixel-by-pixel subtraction is often required. A single bias frame will not sample these variations well in a statistical fashion, so an average bias image of 10 or more single bias frames is recommended. *****

Dark CCD dark frames are images taken with the shutter closed but for some time period, usually equal to that of your object frames. That is, if one is planning to dark correct a 45 second exposure, a 45 second dark frame would be obtained. Longer dark frames can often be avoided using the assumption that the dark current increases linearly with time and a simple scaling can be applied. However, this is not always true. Dark frames are a method by which the thermal noise (dark current) in a CCD can be measured. They also can give you information about bad or "hot" pixels that exist as well as provide an estimate of the rate of cosmic ray strikes at your observing site. Observatory class CCD cameras are usually cooled with LN2 to temperatures at which the dark current is essentially zero. Many of these systems therefore do not require the use of dark exposure CCD frames in the calibration process. Thermoelectrically cooled systems are not cooled to low enough temperatures such that one may ignore the dark current. In addition, these less expensive models often have poor temperature stability allowing the dark current to wander a bit with time. Multiple darks

4.5 Basic CCD data reduction Table 4.1. (cont.)

CCD Image Type Image Description averaged together are the best way to produce the final dark calibration frame. Note that if dark frames are used, the bias level of the CCD is present in them as well, and therefore separate bias frames are not needed.

Flat Field Flat field exposures are used to correct for pixel-to-pixel variations in the CCD response as well as any nonuniform illumination of the detector itself. Flat fields expose the CCD to light from either a dome screen, the twilight sky, the nighttime sky, or a projector lamp in an attempt to provide a high S/N, uniformly illuminated calibration image. For narrow-band imaging, flats are very helpful in removing fringing, which may occur in object frames. Flat field calibration frames are needed for each color, wavelength region, or different instrumental setup used in which object frames are to be taken. A good flat should remain constant to about 1%, with 2% or larger changes being indicators of a possible problem. As with the other calibration frames, at least 5 or more flat fields should be taken and averaged to produce the final flat used for image calibration. *****

Object These are the frames containing the astronomical objects of interest. They are of some exposure length from 1 second or less up to many hours, varying for reasons of type of science, brightness of object, desired temporal sampling, etc. Within each object image pixel is contained contributions from the object and/or sky, read noise, thermally generated electrons, and possibly contributions from cosmic rays. Each pixel responds similarly but not exactly to the incident light, so nonuniformities must be removed. All of the noise and spatial factors are correctable to very low levels via standard CCD reductions as described in the text.

show examples of typical bias, dark, flat field, and object CCD images. Note that a CCD dark frame contains not only information on the level and extent of the dark current but also includes bias level information.

The use of the basic set of calibration images in the reduction of CCD object frames is as follows. First, subtract a mean bias frame (or dark frame

Fig. 4.2. Shown is a typical CCD bias frame. The histrogram of this image was shown in Figure 3.8. Note the overall uniform structure of the bias frame.

if needed1) from your object frame. Then, divide the resulting image by a (bias subtracted) mean flat field image. That's all there is to it! These two simple steps have corrected your object frame for bias level, dark current (if needed), and nonuniformity within each image pixel. During the analysis of your object frames, it is likely that the background or sky contribution to the image will need to be removed or accounted for in some manner. This correction for the background sky level in your image frames is performed

1 The need for dark frames instead of simply bias frames depends entirely on the level of dark current expected during an integration or the stability of the dark current from integration to integration. The first situation depends on the operating temperature of the CCD. LN2 systems have essentially zero dark current, and thus bias frames are all that is needed. Inexpensive and thermoelectrically cooled CCD systems fall into the category of generally always needing dark frames as part of the calibration process.

Fig. 4.3. Shown is a typical CCD dark frame. This figure shows a dark frame for a Kodak CCD operating in MPP mode and thermoelectrically cooled. Notice the nonuniform dark level across the CCD, being darker (greater ADU values) on the top. Also notice the two prominent partial columns with higher dark counts, which extend from the top toward the middle of the CCD frame. These are likely to be column defects in the CCD that occurred during manufacture, but with proper dark subtraction they are of little consequence. The continuation of the figure shows the histogram of the dark frame. Most of the dark current in this 180 second exposure is uniformly distributed near a mean value of 180 ADU with a secondary maximum near 350 ADU. The secondary maximum represents a small number of CCD pixels that have nearly twice the dark current of the rest, again most likely due to defects in the silicon lattice. As long as these increased dark current pixels remain constant, they are easily removed during image calibration.

Fig. 4.3. Shown is a typical CCD dark frame. This figure shows a dark frame for a Kodak CCD operating in MPP mode and thermoelectrically cooled. Notice the nonuniform dark level across the CCD, being darker (greater ADU values) on the top. Also notice the two prominent partial columns with higher dark counts, which extend from the top toward the middle of the CCD frame. These are likely to be column defects in the CCD that occurred during manufacture, but with proper dark subtraction they are of little consequence. The continuation of the figure shows the histogram of the dark frame. Most of the dark current in this 180 second exposure is uniformly distributed near a mean value of 180 ADU with a secondary maximum near 350 ADU. The secondary maximum represents a small number of CCD pixels that have nearly twice the dark current of the rest, again most likely due to defects in the silicon lattice. As long as these increased dark current pixels remain constant, they are easily removed during image calibration.

30000

25000

20000

15000

10000

5000

30000

25000

20000

15000

10000

5000

100 200

100 200

as part of each specific analysis step using "sky" regions in the object frame itself and is not removed or corrected for with some sort of separate "sky" frame. In equational form, the calibration process can be written as

Raw Object Frame - Bias Frame

Final Reduced Object Frame =-,

Flat Field Frame where, again, the flat field image has already been bias subtracted and the bias frame would be replaced by a dark frame when appropriate.

4.6 CCD imaging

This section details issues related to the application of using CCDs to produce images of an extended area of the sky. Examples of this type of CCD observation are multi-color photometry of star clusters, galaxy imaging to isolate star-forming regions within spiral arm structures, deep wide-field searches for quasars, and extended low surface brightness mapping of diffuse nebulae. Use of the areal nature of a CCD introduces some additional issues related to the calibration procedures and the overall cosmetic appearance as any spurious spatial effects will have implications on the output result. We briefly discuss here a few new items that are of moderate concern in two-dimensional imaging and then move on to the topic of wide-field imaging with CCD mosaic cameras.

Fig. 4.4. Shown is a typical CCD flat field image. This is an R-band flat field image for a 1024 x 1024 Loral CCD. The numerous "doughnuts" are out of focus dust specks present on the dewar window and the filter. The varying brightness level and structures are common in flat field images. As seen in the histogram of this image (Figure 4.1) this flat field has a mean level near 6950 ADU, with an approximate dispersion of (FWHM) 400 ADU.

Fig. 4.4. Shown is a typical CCD flat field image. This is an R-band flat field image for a 1024 x 1024 Loral CCD. The numerous "doughnuts" are out of focus dust specks present on the dewar window and the filter. The varying brightness level and structures are common in flat field images. As seen in the histogram of this image (Figure 4.1) this flat field has a mean level near 6950 ADU, with an approximate dispersion of (FWHM) 400 ADU.

4.6.1 CCD fringing and other cosmetic effects

We mentioned earlier that observations of monochromatic (or nearly so) light can cause a pattern of fringes to occur on a CCD image. These fringes, which are essentially Newton's rings, are caused by interference between light waves that reflect within the CCD or long wavelength light that passes through the array and reflects back into the array. Fringing may occur for CCD observations in the red part of the optical spectrum, when narrowband filters are used, or if observations are made of a spectral regime (e.g., the I-band) that contains strong narrow emission lines. For a given fringe cause (e.g., a specific wavelength set of emission lines) the fringe pattern on the CCD remains constant. Figure 4.6 shows a Gemini North GMOS image obtained in a z' filter (central wavelength is near 8800 A) on a photometric night with no moon but plenty of OH emission. The GMOS detector consists of three EEV red 13.5 micron 6144 x 4608 CCDs placed next to each other

Fig. 4.5. Shown is a typical CCD object frame showing a star field. This image has been properly reduced using bias frame subtraction and division by a flat field image. Note how the background is of a uniform level and distribution; all pixel-to-pixel nonuniformities have been removed in the reduction process. The stars are shown as black in this image and represent R magnitudes of 15th (brightest) to 20th (faintest). The histogram shown in the remainder of the figure is typical for a CCD object frame after reduction. The large grouping of output values on the left (values less than about 125 ADU) are an approximate Gaussian distribution of the background sky. The remaining histogram values (up to 1500 ADU) are the pixels that contain signal levels above the background (i.e., the pixels within the stars themselves!).

Fig. 4.5. Shown is a typical CCD object frame showing a star field. This image has been properly reduced using bias frame subtraction and division by a flat field image. Note how the background is of a uniform level and distribution; all pixel-to-pixel nonuniformities have been removed in the reduction process. The stars are shown as black in this image and represent R magnitudes of 15th (brightest) to 20th (faintest). The histogram shown in the remainder of the figure is typical for a CCD object frame after reduction. The large grouping of output values on the left (values less than about 125 ADU) are an approximate Gaussian distribution of the background sky. The remaining histogram values (up to 1500 ADU) are the pixels that contain signal levels above the background (i.e., the pixels within the stars themselves!).

105 104 103 102 101 0

0 250 500 750 1000 1250 1500

vertically. The frame on the left shows a typical CCD fringe pattern caused by the night sky emission lines while the frame on the right has been defringed. Figure 4.7 presents line plots across typical fringing i' and z' GMOS frames. The typical level of fringing is near ±0.7% in i' and ±2.5% in z'.

The troubling aspect with fringing in CCD data is that it is often the case that the fringe pattern does not occur in the flat field frames (flats contain no emission lines!) or the level of fringing is highly variable throughout the night. Without a pattern match between the flats and the image data, fringe removal will not robustly occur during image calibration, and residual fringes will remain in the final object images. One of the major causes of CCD fringing is the night sky emission lines, which occur in the Earth's upper atmosphere (Pecker, 1970). These night sky lines are mainly attributed to OH transitions in the atmosphere, which are powered by (UV) sunlight during the day. Since they are forbidden transitions they have long decay lifetimes and are very narrow spectrally. In addition, due to upper atmosphere motions, OH concentrations, and their long decay times, these emission lines are highly variable in time and strength, even within a given night. Dealing with fringes that occur in CCD data can be a difficult problem but one for which solutions exist (Broadfoot & Kendall, 1968; Wagner, 1992). Observations with newly defined moderate-band filters that lie between the OH transitions is one such example.

Additionally, cosmetic effects such as bad pixels, hot pixels (LEDs), stuck bits, or dead columns can be present and can mar a CCD image. Not only do these

0 250 500 750 1000 1250 1500

Fig. 4.6. Gemini North GMOS CCD fringe frame (left) and reduced, defringed, frame (right). The night was photometric and near new moon but had OH emission present. Notice that the 1-2% fringes can cross over objects of interest but are mostly fully removed during reduction.

GMOS-N red EEV CCDs Aug2001

GMOS-N red EEV CCDs Aug2001

Column [Pixel]

Fig. 4.7. Line plots across the unprocessed CMOS i' and z' images. The plots have been normalized such that the mean image level is zero and the fringe level can be seen to be both positive and negative deviations from this level. The i' fringing is about 0.7% while the z' fringing is near 2.5%.

Column [Pixel]

Fig. 4.7. Line plots across the unprocessed CMOS i' and z' images. The plots have been normalized such that the mean image level is zero and the fringe level can be seen to be both positive and negative deviations from this level. The i' fringing is about 0.7% while the z' fringing is near 2.5%.

flaws spoil the beauty of the two-dimensional data, they can cause problems during calibration and analysis by hindering software processes and not allowing correct flux estimates to be made for the pixels that they affect. Procedures for the identification and removal of, or correction for these types of problems can be applied during image calibration and reduction. They are specialized reduction tasks, which depend on the desired output science goals and generally are specific to a particular CCD, instrument, or type of observation being made. Most observatories provide solutions to such fixed flaws. One example is a bad pixel map, which consists of an "image" of 0s and 1s with 0s at the locations of bad columns or other regions of bad pixels. These maps are used by software in the reduction process to eliminate and fix offending CCD problems. A complete discussion of all of these topics lies beyond our space limitations but the interested reader will find discussions of such corrections in Djorgovski (1984), Janesick et al. (1987a), Cilliland (1992), Cullixson (1992), Massey & Jacoby (1992), and numerous specific instrument manuals and reference papers concerning the finer points of specific CCD related issues (see Appendix A).

4.6.2 Tip-tilt corrections

The Earth's atmosphere causes a blurring of an image and thus a reduction in image quality during an observation. A solution that often eliminates nearly 70-80% of this effect is the use of adaptive optics to perform low order tip-tilt corrections. Mechanical tip-tilt systems exist today at many observatories and consist of a guide star sensor of some type (avalanche photodiodes (APDs) or a small CCD) and a small optical mirror that can tip and tilt rapidly. The sensor receives light from a bright guide star in the field of view (or a laser guide star) during an observation and the quality (mainly the x,y position) of the image observed by the sensor is assessed. A fast feedback is established by which any movement in the guide star is measured and a correction tip-tilt signal is sent to the moveable mirror.

Systems of this type have small fields of view (~2-4 arcminutes) and can only work well if a bright guide star is present of if the telescope is equipped with a laser beacon. Orthogonal transfer CCDs were developed to provide nonmechanical tip-tilt corrections. The OTCCD camera OPTIC (Tonry et al., 1997) has four guide regions (at the ends of the CCDs) and four associated science regions. Up to four stars that fall in the guide regions are used for tip-tilt correction. These stars are read out fast (typically 10-20 Hz), assessed, and a tip-tilt correction signal is fed back to the science regions of the CCDs during the integration. OTCCDs can shift charge on the array in both x and y directions and use this property to provide fast tip-tilt correction in the science image. This same type of feedback can also simultaneously correct for telescope drive errors and wind shake (see Tonry et al., 1997, Howell et al., 2003).

The new generation of OTCCD, the OTAs (see Figures 2.7 and 2.8), will extend the tip-tilt correction ability. They allow use of any of the individual OTCCDs within the 8 x 8 array to be used as a guide region. Additionally, the ability to tip-tilt correct an image can be extended to an arbitrarily large field of view as each part of the OTCCD array corrects itself locally. The WIYN observatory is building a one-degree imager that will provide tip-tilt corrections across the entire 1° field of view. The Pan-STARRS project is developing a similar imager that will cover a 3° field (Jacoby etal., 2002, Tonry etal., 2002).

4.6.3 Wide-field CCD imaging

With the advent of large-footprint CCDs and the construction of CCD mosaic arrays containing many chips, wide-field imaging is becoming one of the major applications in astronomy today. One of the major efforts in observational astronomy today is large field of view, multi-color imaging of the sky. Large surveys such as the Sloan digital sky survey (SDSS) and the two-micron all sky survey (2MASS) are complete and their contribution to astronomy has been amazing. New types of objects, large, very complete samples, and follow-up spectroscopy have shown that imaging surveys can provide tremendous new information.

Temporal variation of the objects in the night sky (both known and unknown) is an additional parameter becoming an integral part of modern imaging surveys. At least six very ambitious wide-field imaging projects are well underway to complement the ten or so, 0.5-1.0° field of view imagers already in action. Table 4.2 lists a few examples of modern wide-field CCD imaging cameras available to the astronomer today as well as those planned to be built and on-line in the next decade. Wide-field imagers of even five years ago consisted of four large format CCDs and required minutes for readout and often days for data reduction. Modern wide-field cameras consist of dozens of CCDs, readout very fast (and will get faster with estimates of 2-4 s), and pass through automated software pipelines in a matter of hours. Figures 4.8-4.10 show two currently working large CCD mosaic cameras plus the planned Pan-STARRS OTCCD camera.

MegaCam on the CFHT was the first operational wide-field, megapixel CCD imager (Boulade et al., 1998) starting science operation in 2002. Today, the Large Synoptic Survey Telescope (LSST) project is the most ambitious of the currently planned wide-field imagers. This special purpose imaging telescope will have a camera containing gigapixels of CCD real estate and image an area of nearly ten square degrees at a time. The LSST camera (see Table 4.1) will likely use an array of 1K or 2K CMOS or CCD ASIC devices. ASIC (Application Specific Integrated Circuits) devices are special purpose production circuits made with a number of non-changeable specific modes built directly into the chip. As such, ASIC devices are often higher in efficiency but somewhat limited in expandability for use other than what they were designed for. An example of a common ASIC device is the computer chip residing under the hood of most modern automobiles. The LSST will image the entire sky every few nights and the amount of data produced will run into the petabytes.

Astronomers are beginning to become different types of observers. Virtual on-line databases, such as the National Virtual Observatory (NVO) will soon sponsor the ability for world-wide access to a tremendous amount of data. Preliminary versions of the web tools and software exist today and with many new CCD imagers available and ever larger ones coming along, there promises to be no shortage of data to sift through and extract scientific research projects from. One downside to this type of observational work is

Table 4.2. Some Present and Planned Large Telescope CCD Imagers (* marks planned imagers)

Name

Telescope

Field of View

Focal Plane CCDs

Pixel Size Pixel Scale (microns) ("/pix) Website

MageCam

QUEST

SUPRIME

Mosaic

MegaCam

Dark Energy Camera* OmegaCam*

CFHT

Palomar 48" Schmidt Subaru

KPNO/CTIO 4-m

CTIO 4-m

WIYN

One Degree Imager*

Pan-STARRS* 1.8-m

Kepler* 1.0-m Schmidt

(Spacecraft)

LSST Camera* LSST

16 sq. degrees 112 - 600 x 2400 Sarnoff 13

0.24 sq. degrees 10-2048 x 4096 MIT/LL 15

0.36 sq. degrees 8 - 2048 x 4096 SITe 15

0.16 sq. degrees 32 - 2048 x 4096 E2V 15

1.0 sq. degrees (OTA) 60 - 3840 x 3952 12 STA/Dalsa

105 sq. degrees 42 - 2200 x 1024 E2V 27

9.6 sq. degrees 3 Gigapixels LBL 10

0.185 http://www.cfht.hawaii.edu/

Instruments/Imaging/Megacam/ - 1.5 http://www.astro.caltech.edu/

~george/pq/ 0.2 http://www.naoj.org/

Observing/Instruments/SCam/ 0.26 http://www.noao.edu/

kpno/mosaic 0.087 http://cfa-www.harvard.edu/

~bmcleod/Megacam/ 0.28 http://www.fnal.gov/ pub/

0.21 http://www.eso.org/

instruments/omegacam/ 0.11 http://www.noao.edu/

wiyn/ODI/ 0.3 http://pan-starrs.ifa.hawaii.edu/

public/index.html 4 http://www.kepler.arc.nasa.gov/

0.2 http://www.lsst.org/

lsst home.shtml

Fig. 4.8. Photograph of the CCDs used in the CFHT MegaCam. Forty large format E2V CCDs are used in this camera, which can image a field of view of nearly 1 x 1 degree on the sky.

that the virtual observer will only be able to get the data that were taken and they may or may not suit their needs. So don't stop thinking of your own observational projects or planning to go to a telescope to collect your own data just yet.

Wide-field CCD mosaic imagers provide a tremendous amount of information (and data) to be collected in one exposure. The soon-to-be-operating OmegaCam on the VLT survey telescope (VST), for example, will produce

Fig. 4.9. A view of the Subaru SuPrime CCDs mounted in the camera dewar. This camera images a field of view of ~ 0.5° on a side using ten 2048 x 4096 MIT/LL CCDs.

over 4200 Mb of data in one exposure! CCD mosaic arrays are pioneering new scientific advances and driving astronomical technology such as readout, CCD controllers, and data storage. These larger CCD arrays are also being enabled by faster computational ability and increased effort in software and hardware development. Astronomers have been making a transition from internal View of Gigapixel Camera Cryostat

PtrtrARftS CMmera Cobft - X

Fig. 4.10. Engineering drawing of one of the four Pan-STARRS imagers currently under construction. This camera will use sixty OTAs to cover a 7 square degree field of view.

single researchers and few night runs at a telescope, to large collaborations that build instruments and telescopes, to the production of extensive non-proprietary databases. Financial constraints and enormous complexity are the prime drivers of this new research model. Physicists went down this road many years ago and we often joke about their papers having less text in the science portion then the two pages that list the 200 authors. Astronomy is going in this direction and the new generation of large, expensive CCD imagers are leading the way.

The efficiency of a large-area survey can be estimated by the metric

Was this article helpful?

0 0
Telescopes Mastery

Telescopes Mastery

Through this ebook, you are going to learn what you will need to know all about the telescopes that can provide a fun and rewarding hobby for you and your family!

Get My Free Ebook


Post a comment