MAG*I*CAL CALIBRATION SAMPLE FOR TEM


Certificate of Traceability


The MAG*I*CAL calibration reference standard is manufactured from a single crystal silicon wafer, and so incorporates a fundamental constant of nature into the standard itself - the crystal lattice constant of silicon. All calibration markings on the standard are directly referenced to this natural constant. The calibrated values can be verified by the user by observing the crystal lattice image of the silicon substrate and validating the calibrated values.

National Metrology Institutes (NMI's) such as NIST in the USA, NPL in Great Britain, etc., currently do not certify any measurements less than approximately 0.3 micrometers. NMI's certify measurements as traceable to fundamental constants of nature, but they do not certify natural constants. The MAG*I*CAL calibration reference standard does not require NMI certification since the calibration is directly traceable to a natural constant available on the standard itself (the crystal lattice spacing of silicon) and all calibrated values can be verified by the user. The crystalline lattice spacing is an intrinsic property of a material. For pure silicon, the spacing has been well characterized and documented by the scientific community. It is known to better than 8 decimal places (0.313 560 136 (8) nm [1].). The current, best-known value can be obtained from a recent CRC Handbook of Chemistry and Physics [1], and many other references.

Documents that request a reference to traceability for the MAG*I*CAL reference standard should be filled in with "natural constant" or "directly traceable to a constant of nature [1]".


[1] CRC Handbook of Chemistry and Physics, CRC Press, Inc., Boca Raton, Florida 33431

Certificate of Calibration

The MAG*I*CAL calibration reference standard is a cross-sectional TEM sample made from of a semiconductor multilayer consisting of four sets of five nominally 10 nm thick Si0.81Ge0.19 alloy layers, alternating with nominally 13 nm thick pure silicon layers. The device-quality epitaxial layers were grown by Molecular Beam Epitaxy (MBE) as strained layers on a single crystal silicon <001> substrate. All calibrated values incorporate the strain affects.

The four sets of alternating layers (superlattices) provide dark and light contrast in the TEM, and were directly calibrated by high resolution transmission electron microscopy (HREM) in reference to the {111} lattice spacing of silicon, as measured on the single crystal silicon substrate. This spacing is known to be 0.313 560 136 (8) nm [1]. The uncertainty in measurement across each of the full superlattices is less than one atomic layer at both the top and bottom interfaces: dt (superlattice) < 0.5%.

The variation in layer thicknesses across the wafer material used for the current series of MAG*I*CAL calibration reference standards was measured by double crystal x-ray diffraction (DCXRD) mapping as < 1.0%. The estimated uncertainty in all sets of calibrated values (other than the individual thin SiGe and Si layers) assumes a normal distribution, and is the combination of all uncertainties added in quadrature. The overall uncertainty on the calibrated values listed on the 'Layer Thickness Values' sheet is:

dt < 1.0%

Multiplying the overall uncertainty by a coverage factor of k=2 gives a confidence level of approximately 95%. Any user can fine-tune the calibrated values and decrease the uncertainty of their individual calibration reference standard by verifying the calibrated values with reference to the Si {111} lattice spacing in the substrate.

One year limited warranty on defects in manufacture. Replacement or refund provided by the distributor.

[1] CRC Handbook of Chemistry and Physics, CRC Press, Inc., Boca Raton, Florida 33431




********************************************************
There have been some helpful discussions of MAG*I*CAL and traceability of TEM calibration samples held on the Microscopy Listserver (sponsored by the Microscopy Society of America; http://www.msa.microscopy.com/MicroscopyListserver/FAQ.html). Excerpts from several threads are given below:
********************************************************


Hi Listers: 6 April 2002

Renate asked what would appear to be a very simple question about the availability of traceable standards for TEM. There really are a couple of valid reasons underlying the problem, and there is a solution.

Actually, the rules of the game, which are based on ISO 17025, the international standard for laboratory accreditation, require that measurements be traceable to either a national standards laboratory (NIST in the U.S.) or to a fundamental natural constant. The American Association for Laboratory Accreditation adds the wrinkle that the traceability must be through an unbroken chain of accredited laboratories. That's the basis for the problem; there are no laboratories accredited to make the necessary measurements in the U.S.

Now, the rules allow you to do any sort of calibration within your own operation, provided that you establish traceability. So if you have something that is NIST traceable or traceable to a fundamental natural constant, you can trace your internal measurements to whatever that is. NIST has offered microspheres, although a search of their website a couple of minutes ago couldn't come up with a SRM (standard reference material) number. A number of vendors, including SPI Supplies, offer a range of microspheres which are sized based on those NIST samples.

The MAG*I*CAL specimen (SPI #02218-AB), however, bypasses the whole question of accreditation by being based on the fundamental properties of the element silicon. As a result, you can bypass NIST completely. More information on this sample can be found at

http://www.2spi.com/catalog/standards/magical.html

A note about NIST: I have nothing but the highest respect for the microscopy group at NIST. I don't know what their plans are, but I suspect that they are caught between a rock (the high requirements placed on standard reference materials) and a hard place (the realities of the world of microscopy). The kinds of uncertainty that we routinely accept in our work are simply not acceptable in the world of metrology (the art and science of measurement). I also suspect that their priorities are directed in directions other than TEM magnification.

Disclaimer: I am laboratory director of Structure Probe, Inc., an independent analytical research laboratory which offers electron microscopy services to clients. Structure Probe, Inc. is the parent company of SPI Supplies. We have an obvious interest in promoting microscopy in general and the use of the products we sell in particular. Structure Probe has been involved with the American Association for Laboratory Accreditation since it was founded in the 1970s. For the last several years, I have served as an assessor, a member of the Accreditation Council (the group that makes the actual decision whether a laboratory is or is not granted accreditation) and a member of the Measurement Advisory Committee (the group that considers the specific problems of metrology laboratories) for A2LA.

Andy

Andrew W. Blackwood, Ph.D.
Structure Probe, Inc.
P.O. Box 656
West Chester, PA 19381-0656
Ph: 1 610 436 5400 X108
FAX: 1 610 436 5755
e-mail: ablackwood@2spi.com
WWW: http://www.2spi.com

Does anyone know if this John McCaffrey calibration standard is NIST traceable or if it has been NIST tested. Russ

Hi Russ,

The timing of this question is pretty good! We have been trying for quite some time to find the right person in NIST to arrange for some form of certification for this sample, because of the frequent requests for ISO-9000-type traceability. Three days ago we finally received a usable letter - the text is included below.

The biggest problem for both us and NIST has been that the MAG*I*CAL sample is made from a single crystal of silicon, with the calibration marks being directly traceable to the silicon (111) lattice spacing. This spacing is directly observable on the sample itself through lattice imaging; i.e.: the sample is internally calibrated to a "God-traceable" standard, and NIST generally doesn't bother confirming intrinsic properties of materials. Certification auditors requiring NIST-traceable samples are generally most concerned about adequate paperwork. This recent communication with NIST gives us that paperwork. I believe, as NIST acknowledges the lattice spacings of silicon as being well characterized and documented, making the MAG*I*CAL sample a NIST-recognized sample. We'll be discussing the proper NIST term to use with the MAG*I*CAL sample when Rob Gettings gets back from a business trip in a couple of weeks. In the meantime, a copy of the NIST letter will be available for users of the sample for their certification requirements. If anyone requires a copy of the NIST letter right now, they should contact South Bay Technology, the world-wide master distributor of the sample (fastest), or the vendor from which the sample was purchased (will take longer). The South Bay contact is:

David Henriks
South Bay Technology, Inc.
1120 Via Callejon FAX: 714-492-1499
San Clemente, CA 92673 USA
TEL: 800-728-2233 (toll-free in USA)
TEL: 714-492-2600 (outside USA)
e-mail: henriks@southbaytech.com

*****************************************************************

TEXT OF LETTER FROM ROBERT GETTINGS, NIST:

NIST
U.S. Department of Commerce
NATIONAL INSTITUTE FOR STANDARDS & TECHNOLOGY

Robert Gettings
Standard reference Materials Program
Bldg. 202, Rm. 212, Gaithersburg, MD 20899

June 23, 1998

Dear Mr. McCaffrey:

I am sending this communication in regards to your question concerning the calibration of Transmission Electron Microscopes (TEM) using silicon lattice spacings. NIST does not currently supply a standard for calibrating the magnification scale of TEM's, nor for the crystalline lattice spacings of silicon.

The crystalline lattice spacing is an intrinsic property of a material. For pure silicon, the spacing has been well characterized and documented by the scientific community. It is known to 6 decimal places, and can be obtained from the CRC handbook of Chemistry and Physics among other references.

Related work by NIST:

NIST is currently working on SRM 640c Silicon X-ray Diffraction Powder. This material, when finished will provide certified line positions traceable to the SI definition of length.

NIST is also working on a new standard, SRM1990, for the lattice spacing of single crystal ruby. The material will be in the form of a 0.15 mm sphere, and is expected to be complete by the end of this year.

Thank you for your interest in NIST standards.

Sincerely,
Robert J. Gettings
Project Manager


Does anybody have a new or tried and true suggestion for a TEM magnification calibration standard for a mag of 30KX using a CCD camera? I would like to have better than +/- 3% accuracy. Have tried catalase-not happy with it. Latex spheres seem to shrink about 20% (according to a 1968 paper). Diffraction gratings not acceptable--only two to three lines in the image.

John C. Wheatley
Lab Manager
Arizona State University
Center for Solid State Science
PSA-213
BOX 871704
Tempe, AZ 85287-1704

Phone: (602) 965-3831
FAX: (602) 965-9004
John.Wheatley@ASU.Edu

I have used the MAG*I*CAL sample that John McCaffrey and crew came up with and which South Bay Technology distributes. This sample covers the full range of Mags available in the TEM. It is done very quickly and simply and it can be used to measure the rotation calibration of the diffraction pattern and the camera constant as well. I have even used it in a High resolution SEM but its use there isn't quite as easy.

Incidentally, they have the Guinness World Record for the World's Smallest Ruler.

-Scott Walck

Scott D. Walck, Ph.D.
PPG Industries, Inc.
Guys Run Rd. (packages)
P.O. Box 11472 (letters)
Pittsburgh, PA 15238-0472

Walck@PPG.com

(412) 820-8651 (office) (412) 820-8161 (fax)

"The opinions expressed are those of Scott D. Walck and not of PPG Industries, Inc. nor of any PPG-associated companies."

MAG*I*CAL is manufactured and distributed by Norrox Scientific Ltd.