to Scan agreement form
Do
you take in projects from outside researchers?
Yes; in fact, one of the primary purposes for this facility is
to provide accessible and top-quality high-resolution scanning
services for the geological and overall scientific community.
What
are your prices?
Scanning services cost $108/hour for academic projects, or $283/hour for commercial work. Scanning time includes not only image data acquisition, but also the time necessary to mount the specimen, and to calibrate and optimize the scanner. There are also associated charges relating to converting the raw CT data into slices, processing the data to minimize artifacts, and transferring and archiving the data. These associated charges are billed at a lower ‘image processing’ rate of $63/hour academic and $163/hour commercial, as above. We also offer additional image processing and analysis services at the image processing rate. In addition, our status as an NSF-supported multi-user facility entitles NSF-funded projects to be charged at reduced rates (50% reduction for projects funded by the NSF EAR directorate, and 25% for all other directorates), and preferred scheduling. We can also provide a limited number of free test scans (typically less than an hour of scanning) as proof-of-principle to justify or help to obtain funding for a complete study. We must have a Scan Agreement Form on file before we start any project.
It is downloadable as a Word
file, a RTF
file or a PDF
file.
How long
does it take to acquire an individual scan slice?
Scanning times are quite variable, depending on the scientific objectives of the researcher, the geometrical and material characteristics of the objects being scanned, and the scanning system we use. On the high-energy system, per-slice times for 3rd-generation (rotate-only) scans are generally on the order of 1-4 minutes, but have ranged from about 29 seconds to 10 minutes. 2nd-generation (translate-rotate) scans, which are used for larger objects or where greater resolution is desired, usually range from 8-20 minutes per slice. Our microfocal system is capable of acquiring data for multiple slices within a single rotation, which considerably speeds things up. There can be some cost in distortion when this mode is employed, which is linked to the number of slices acquired simultaneously. In general, we keep this distortion to negligible levels by aquiring data for only 20-30 slices per rotation, though up to 100 is possible. Per-rotation times generally range from 3 to 10 minutes, resulting in per-slice times from 6 to 30 seconds.
In general, if the details being inspected have a large density contrast
with the surrounding matrix,
or if they are fairly large in proportion to the overall object,
scan times are shorter; conversely, if the features being studied
are small or consist of subtle density variations, longer scan
times will probably be required to distinguish them with sufficient
clarity. Denser objects tend to take longer to scan than less-dense
objects, and very small objects (<2 cm) often require longer
scan times to achieve suitable noise reduction in the images.
On our high-energy system, a good, general base estimate is 2
minutes per slice, which will probably be within a factor of two
of the correct answer. On the microfocal system, 6-12 seconds per
slice is usually sufficient. We might
be able to provide a better estimate if you describe your sample
and imaging requirements to us.
What
resolution does the UT scanner have?
Again, this depends largely on sample characteristics, of which
size is the most directly influential. A useful benchmark is that
the images we collect are at most 1024x1024 pixels, and the entire
sample must fit within the scan field of view. Thus, insofar as
it takes a few pixels on a computer image to distinguish a feature,
our maximum resolution is correspondingly a few 1024th's of the
maximum dimension of the object in the scan plane. The optimal number
of pixels in an image roughly corresponds to the number of channels
in the detector being used, which in most high-energy scans will be
512, and 1024 on our micro-focal system.
How
large/small a sample can you scan?
Our smallest objects scanned to date have been a few mm in diameter.
The largest field of view we can image when scanning using 3rd-generation
(rotate-only) mode on our high-energy system is about 30 cm. We
have the capability of scanning objects up to 50 cm in maximum
in-plane dimension, but only using 2nd-generation (translate-rotate)
mode. Samples scanned on the high-energy system can be up to 1.5
m high. On our microfocal system, samples can be up to about 5
or 6 cm in diameter and 15 cm high.
How
thin can you make the slices?
This depends on the sample size and the scanning system we use to image it. For our microfocal system, achievable slice thickness is a direct function of the object maximum dimension in the scan plane. For a sample with a 50 mm maximum dimension, the thinnest slice we can collect is about 52.5 micrometers; for a sample with a 3 mm maximum dimension we have achieved a slice thickness of 3.7 micrometers, which is close to our lower limit of possible resolution. For samples larger than 50 or 60 mm, we generally use our high-energy system, in which slice thickness is independent of sample size and has a minimum of about 0.25 mm. Also keep in mind that image quality is influenced largely by how many X-rays we can send through the sample, so thin slices may take longer than thick slices for a given amount of noise reduction.
What
kind of sample preparation needs to be done?
It depends on what you're scanning; in many cases, no preparation
needs to be done at all, or whatever preparation that is necessary
is best handled by us. In general, the optimal sample geometry
is a cylinder; thus if your sample can be made cylindrical, it
would be beneficial for scanning efficiency and image quality.
If a cylindrical geometry is not possible, it would still be good
to have the sample be as equidimensional as possible in the scan
plane. One technique we often use is to pack the sample in a cylindrical
container surrounded by material of similar density.
Is
there any other information that you need for scanning?
In general no, but the more information you provide the better
we might be able to optimize the imaging conditions. One thing
that can sometimes be helpful is compositional information, such
as what materials or mineral phases are in the sample, and if
possible detailed chemical compositions of them. Another useful
item is often a thin section or some other means of "ground-truthing"
our scanning to compare what we see in the images to what is really
in the sample.
How
many slices will I need for my study?
In most cases there is no general answer to this question; it depends on the scientific requirements of the research.. One rule of thumb for paleontological investigations is that a vertebrate skull imaged in the coronal plane does well with about 400-600 slices, regardless of the size of the skull. For object identification (such as locating and measuring crystals in matrix), it is usually advisable to have the smallest object you want to identify appear in at least two or three slices, so your inter-slice spacing should be less than half the minimum size of the object you want to study. Another thing to keep in mind is that we often slightly overlap the scan slices to improve continuity for 3D reconstruction.
How
much would it cost to scan my sample?
To estimate the duration (and thus price) for a scan, start with an hour or two for setup, calibration, and optimization of the scanner for the particular requirements of your project (billed at the scanning rate). Then figure out how many slices you think you'll need, estimate the per-slice time, and multiply them together to get the total number of scanning hours. Remember that this estimate is probably only accurate within a factor of two or so. A "typical" scanning application takes from 2-8 hours per sample, but of course in any specific case it can take much shorter or much longer. We also typically charge image processing time for converting the raw CT data into slices and using custom software to minimize artifacts in data (one hour for the first 500 Mb of data, and then 0.5 hours for each additional Gigabyte), and subsequently for transferring and archiving the data onto CD (one hour) or DVD (three hours). Hourly rates are listed above. The size of the data set can be estimated by multiplying the number of slices times the size per slice (0.5 Mb per slice for a 512 X 512 pixel reconstruction; 2.0 Mb per slice for a 1024 X 1024 pixel reconstruction).
Can
CT scanning provide direct density measurements?
Yes, but it would take some preliminary work geared to a specific
study objective. CT images generally reflect relative density
variations in the object being scanned, but specific gray scale
values reflect a wide range of factors determined by the scanning
conditions. In order to obtain density data, calibrations must
be performed based on the specific material being scanned at the
exact scanning conditions that will be used.
What
file format do you produce?
The native format is a computer TIFF file in unsigned 16-bit gray scale (i.e.,
intensity values range from 0 to 65535). Exporting to an 8-bit format entails some loss of image
information, but in many cases it is not very significant, and
most present-day software packages for image processing or analysis
(Adobe Photoshop, ImageJ) tend to work much better with 8-bit
images. If necessary, for a charge we can also convert the data
to other formats. We generally distribute the data by creating
a CD or DVD and mailing it or via ftp. We also archive all of the data that we acquire, and can provide extra copies for a fee.
What
image processing services do you provide?
Right now we principally provide services to aid in visualization
of the scan data in 2D and 3D by creating images and animations.
The image folio on this web site provides good examples of many
of the things we can do. We are also developing capabilities for
more sophisticated image analysis, such as 3D object identification
and characterization and STL file creation. Our facility includes
a multi-platform (Windows, Mac, Silicon Graphics) computer lab
that can be made available to outside researchers wishing to come
here to work with their data.
I
want to try something out. What should I do next?
Two things: first, download and read the scan
agreement form, as we'll need a signed
copy of it before we can scan anything. Second, contact us, either
by email at ctlab@maestro.geo.utexas.edu (preferred), or by phone at (512) 471-0260.
What's
your turnaround time?
It varies depending on our workload at the time, so consult with
us to find out what the current situation is. In general, test
scans are the lowest priority, so turnaround time for them tends
to be longer. If you have a special deadline by which you need
scans, such as a grant application deadline, let us know and we'll
do our best to accommodate it.
How and where should I ship specimens, and how will specimens be handled?
We require that either Rich Ketcham or Matt Colbert be notified before samples are shipped to us for scanning. Samples should be securely packed in a sturdy container. We will generally use the same shipping material to send the specimen back to you. The CT lab has secure storage in locked specimen cabinets. Let us know if there are any special requirements for specimen storage (e.g., whether the specimen is pickled in ethyl alcohol, isopropyl alcohol, or formalin; whether the specimen is to remain frozen, etc.).
For all natural history samples (e.g., rocks, fossils, or biological material) we also require that you provide specimen identification and basic locality data. This data is entered into our internal database, and does not leave the lab without your consent. Such data should include, but not necessarily be limited to: taxon name (to smallest identified level) or rock type; specimen number (if accessioned into a collection); general locality information; age (in the case of rocks and fossils); and any other ancillary information that is deemed appropriate (i.e., sex, maturity, etc.). Specimen loan forms from museums will often include all this information. Specimen loan forms from museums should indicate that we have permission to scan the specimen.
In most cases we require that biological and fossil specimens either be accessioned into a recognized museum or natural history collection, or have a guarantee that they will be accessioned. Exceptions may be made, such as when there are plans to destructively examine experimental specimens after scanning.
Let us know if there are special requirements or restrictions in preparing specimens for scanning. We will need permission to remove specimen labels from biological samples if necessary to improve scan quality. In a similar vein, fossil samples occasionally are damaged in shipping, and may benefit from preparation prior to scanning. With permission we can use the services of the professional preparators at the Vertebrate Paleontology Laboratory of the Texas Memorial Museum to mend damaged specimens.
Most of the samples that we scan are shipped to us rather than hand-carried. The University of Texas has different mailing addresses depending on how the sample is shipped.
If you are shipping your sample using U.S. Mail, the address is as follows:
Either Rich Ketcham or Matt Colbert
High-Resolution X-ray CT Facility
The University of Texas at Austin
Department of Geological Sciences
1 University Station C1100
Austin, Texas 78712-0254
If you are using FedEx, UPS or another carrier that requires a street address, use:
Either Rich Ketcham or Matt Colbert
High-Resolution X-ray CT Facility
The University of Texas at Austin
Department of Geological Sciences
23rd at San Jacinto St.
Austin, Texas 78705
|
