Details from a Distance: Seeing Angstroms at 10,000 Kilometers

Auke van BALEN <avb@nl.feico.com>
FEI/Philips Electron Optics
Netherlands

Christopher L. MORGAN <cmorgan@csuhayward.edu>
California State University, Hayward
USA

Contents

Introduction

In recent years, the interest in remote microscopy has risen considerably. In this paper, we would like to define the term "remote microscopy" as microscopy at any distance from the instrument, including the possibility of enabling a user to view images and other results while not being seated in front of them. The distance to the instrument itself is unimportant; it might be that the observer is in the adjacent office. Moreover, it must also be possible to operate the instrument from the remote location. In this definition, "instrument" could be any microscope. For our purposes, we would like to limit the discussion to the remote operation of Scanning and Transmission Electron Microscopes (SEM and TEM instruments). However, much of what we say applies to networking a wide variety of scientific instruments. These instruments are highly complex pieces of equipment, traditionally requiring highly skilled operating personnel. These instruments use electron beams to obtain images of surface detail (in the case of a SEM) or internal structure (in the case of a TEM). Because they use electrons instead of ordinary light they are capable of a wide range of magnification (hence scale) from a factor of as little as 10 to over 1 million. Remote microscopy is one way of making this kind of equipment available and useful to a far greater audience.

We start with a survey of current applications and projects in remote microscopy, then analyze which factors have contributed to the increasing demand for remote microscopy and characterize the various application areas where remote microscopy can be used. Subsequently, we give an overview of a number of projects already in development in this area and try to formulate the important questions. In the last part, we reveal what we have done so far and indicate in which areas our research will be directed in the near future. The last section contains suggestions for network designers in order to make applications like remote microscopy even more practical.

Examples of current applications

Remote use of such equipment allows researchers, students, and technicians to operate instruments at their own convenience, comfort, and safety. For example, one researcher has used the XL 40 in Microscope and Graphics Imaging Center (MAGIC) [MAGIC] at Hayward to finish her thesis, manipulating samples while at Moss Landing a couple of hours away. Professors in classrooms at San Francisco State University have demonstrated electron microscopy to students by live operation of the microscope at Hayward across the bay.

Larger institutions such as universities, national laboratories, and companies can encourage potential scientists and workers by giving them a taste of real operations. For example, the Beckman Institute [bugscope] offers access to its ESEM for an hour once a week for schoolchildren to view their own samples.

National laboratories in the U.S.A. such as Lawrence Berkeley, Argonne, and Oakridge share equipment including electron microscopes over the Internet and are mandated to create a virtual "collaboratory" together, sponsored by the U.S. Department of Energy [MMC].

Why is there so much interest in remote microscopy?

The most important reason we can identify as to why the interest in remote microscopy has surged is the fact that science and technology have long since ceased to be an individual affair. Scientists have collaborated in scientific research projects for a very long time already, and the number of collaborative projects continues to climb. To make effective use of their valuable time, scientists not only need to be in daily direct contact (as is readily possible with e-mail and telephone) but also really want to work together without having the burden of (international) travel. Time can be saved even while working in the same facility or building (e.g., when conferring with a more experienced colleague or one who has special relevant expertise). Similarly, instruction in instrument operation for a group of students is easy [bugscope].

Second, even when travel time is not a factor it can be advantageous to operate the instrument from the next room. When high-resolution (to below 1 angstrom, 1 Å = 10-10 m) pictures are to be taken, the physical presence of a human operator close to the instrument can be the cause of disturbances. Temperature changes, air drafts, and acoustic noise can affect even a modern high-performance instrument such that the result obtained is less than optimal. By placing the instrument in a well-designed space where human access during operation is not required, operating conditions can be improved. Similarly, the clean rooms of the semiconductor industry, where human-generated dust particles should be kept to an absolute minimum, can benefit from remote instrument access. Inversely, sometimes it is not the instrument that should be protected against the presence of the operator, but the operator against the instrument and/or the sample. Some materials are so dangerous that the operator is not allowed to be in the same room as the microscope. An example is examination of material containing a virus so virulent that it has been shown to be biologically active even after radiation with a 200 kV electron beam. Radioactive material, as commonly used in the nuclear industry, is another case where operator protection can be of the utmost importance.

In another vein, from the financial side, an electron microscopy facility is a capital-intensive venture and should be used as efficiently as possible. This can be done by cofunding and shared usage, where researchers and scientists from various collaborating institutions share the instruments at a distance. Once the remote access facility is in place, institutions that never could afford such an instrument on their own can utilize state-of-the-art equipment. Recently it has been shown that even high school students can safely use these instruments [bugscope], offering very good prospects for science education.

In a more production-oriented environment, such as in a clean room of a semiconductor manufacturer, sample handling is often restricted to doing a more or less standard set of measurements. The operator is required to follow a recipe. If something unusual happens, the operator has to call in a supervisor or other more knowledgeable person for consultation. It would certainly be advantageous if it were possible to do this supervision or assistance from a distance, or even to call for assistance automatically.

The last reason for interest in remote microscopy is service and diagnosis. When an instrument malfunctions, it saves the service engineer time if he or she can diagnose the system from his or her own office. In some cases it might even be possible to fix a problem from a distance. In other cases, when a visit is necessary, it is valuable that the field service engineer already knows the problem and therefore can bring the proper parts and tools. Access to some customer facilities like semiconductor clean rooms may be difficult at times, so all preparatory work that can be done outside the customer's premises helps in making the service operation more efficient. People responsible for supervising a number of instruments in a plant might also appreciate the efficiency of remotely monitoring the operation of the equipment.

Current projects

A number of researchers have already worked for a considerable time in the field of remote microscopy. Various people use different terms to indicate what we have defined above as "remote microscopy." Titles of articles and papers contain terms like "telepresence microscopy" and "interactive collaboration." This section is a brief list of people and institutions working in this area, though it is quite possible that we are not aware of all developments in this rapidly expanding field.

  1. Mark Ellisman et al. at U.C. San Diego and the San Diego Supercomputer facility [Fan1993, Ellisman1998]
  2. DOE2000 from the U.S. Department of Energy (DOE). "DOE2000 is a new initiative to fundamentally change the way scientists work together and how they address the major challenges of scientific computation. To accomplish this change, DOE2000 will develop and explore new computational tools and libraries that advance the concept of 'national collaboratories' and Advanced Computational Testing and Simulation (ACTS)." [DOE2000, MMC, ORNL, Argonne, Allard1998, Mabon1999, O'Keefe1998, Wright1998, Zaluzec1998]
  3. Nestor Zaluzec et al. at the Materials Science Division of Argonne National Laboratory [Argonne, Zaluzec1996a, Zaluzec1996b, Zaluzec1998]
  4. Michael O'Keefe et al. at the National Center for Electron Microscopy, Lawrence Berkeley National Laboratory [NCEM3, O'Keefe1996, O'Keefe1998, Parvin1996, Parvin1997]
  5. Christopher Morgan et al. at the Microscope and Graphic Imaging Center, California State University Hayward [MAGIC, IRSA, Morgan1998a, Morgan1998b, Morgan1999, Smith1996a, Smith1996b, Smith1996c]
  6. Steve Barlow at the Electron Microscope Facility, San Diego State University [SDSU]
  7. Bridget Carragher et al. at the Beckman Institute for Advanced Science and Technology, University of Illinois at Urbana-Champaign [bugscope]
  8. B.C. Breton et al. at the Department of Engineering, University of Cambridge (UK) in collaboration with LEO Electron Microscopy [Cambridge, Breton1996, Breton1997, Caldwell1998, Chand1996, Chand1997]
  9. M. De Graef et al. at the Department of Material Science and Engineering, Carnegie-Mellon University
  10. James Vesenka at the Department of Physics, California State University at Fresno [Vesenka1999]

Commercial vendors as well have shown interest in developing products for remote microscopy. LEO's NetSEM [NetSEM] was one of the first commercially advertised packages capable of running an SEM from a distance. More recently, JEOL's FasTEM [JEOL] was announced. Customer interest and acceptance of these products seem to be low for now, however, judged by the absence of reports of practical usage.

Are we solving the right problem?

Imagine a scientist working with concentration on a TEM to obtain a good image of his sample. Suddenly, he hears a loud tick and the image disappears. As every TEM user knows, sometimes the high tension in the electron gun causes a flash-over, often due to imperfect insulation. It is generally harmless however, and switching on the high-tension generator again and waiting for it to stabilize permits the work to be continued rapidly. At a distance though, the flash-over tick cannot be heard and the user only observes that the image suddenly has disappeared, which of course can have a great number of causes. Diagnosing why the image is no longer visible will certainly take more time than when the audible component was also present.

This example illustrates the fact that simply reproducing the "normal" user interface of an instrument so that it can be operated at a distance is most likely not a good solution for designing a remote user interface. Another aspect of the remote operation is that there will be in all cases problems with the bandwidth and the delay time; that is, transport of information (e.g., images) from the instrument to the user or vice versa (e.g., control commands) is restricted by the transport capacity of the medium (e.g., Ethernet or a telephone line) and by the inherent delay that is caused by the transport. Think of an instrument where focusing must be done manually by looking at an image that is updated only once per second; the instrument would be next to unusable (unless you are very patient).

We think therefore that remote operation of a scientific instrument can be fully successful only when the operator is not in the "inner loop" -- that is, when it is not necessary anymore that a human operator observe the image, adjust parameters, and check the result. The role of the user must be the role of director, instructing the instrument to perform an operation like "obtain an image from that interesting-looking spot, do an analysis on the particle, report the result." The user is the final judge of the results, directing the actions on a far higher level by deciding what the overall flow of the experiment must be rather than dealing with all details. This supposes that the instrument has a great deal of intelligence and autonomy -- that it can partially analyze results like images and take corrective actions where required. It should also compensate for imperfections like drift. In this way, it becomes possible to let the instrument run experiments autonomously. This possibility will be very welcome in the application areas where instruments are used for doing routine measurements, as in the semiconductor industry.

Our strong conviction is that the research efforts in the field of remote microscopy should be directed towards three goals:

  1. To make the instrument and its associated equipment easily computer controllable and as autonomous as possible.
  2. To provide a user interface for remote operation that is so easy to use that remote operation becomes the preferred way of working.
  3. To provide an environment that makes it possible for nonspecialists in electron microscopy to set up and conduct their experiments and standard procedures.

Classes of users

This figure shows the classes of users that can be distinguished by their physical distance from the instrument. The instrument is shown as the center disk.

Instrument requirements

In our view, an instrument capable of remote operation must meet the following requirements:

  1. Everything on the instrument is controlled by software. Due to the long development history of electron microscopes, some parts of the system are still only controllable by a human hand. Example: hand-operated vacuum valves, panel switches.
  2. All parameters and values of the instrument must be readable. This is not so obvious as it sounds, as the physical construction can prevent readability. Example: the vacuum system, where it might not be possible without opening or closing valves to check the vacuum condition of some containers in the system.
  3. The system must be designed to act in a predictable and reproducible way. It is very cumbersome when the result of an action depends partially on an action performed some time ago. The control actions should be sufficiently fast so that more complex routines can be assembled.
  4. The interface of the system must be as orthogonal as possible; that is, the various system parts should operate independently.
  5. The system must have functions available that can judge the quality of an obtained result. Example: a function that gives a quality rating to an image indicating how "fuzzy" the image is. The system must have the ability to calibrate these functions.
  6. The system must have procedures in place that use the above-outlined judgment and apply a calculated correction to an instrument setting in order to improve the results. Example: if two subsequent images show that the sample has drifted, apply a drift compensation algorithm.

Image transmission

Sending commands over a network to do such operations as turning the beam on and off or changing the magnification is relatively straightforward. However, transmitting images is more of a challenge. Because images require much more data, the bandwidth of the remote connection is likely to be the bottleneck. Fortunately, some situations such as viewing a fixed sample require image updating only as the user changes magnification, moves the stage to a new position, or performs another adjustment. Even then, a high-quality image is usually not needed until the user has navigated to a particular choice of stage position and magnification and has made several adjustments such as focus and astigmatism.

Images can be compressed to reduce bandwidth requirements. Compression can be lossless or lossy. With lossy compression, image quality is degraded to increase performance, which is often needed during navigation. Image compression standards such as JPEG degrade image quality by removing detail in the form of high-frequency components of a virtual image signal. JPEG uses a rather arbitrary division of the image into fixed-size blocks and uses discrete cosine approximations that assume or presume periodic behavior. Wavelets offer a more natural approach because they adaptively adjust to and measure different scales of detail. They also support the use of progressive image transmission. Wavelets are currently an active area of research in digital signal processing [Mallat1998].

We have used wavelets for progressive image transmission in which gross information with slowly varying virtual signals is sent first. Then more information is sent in a progressive manner to gradually fill in finer and finer detail. Our approach uses what is called embedded encoding in which transmission can stop at any time to give some sort of approximation of the image. This allows a user to interrupt the flow and begin a new image with a stage movement or magnification change command whenever desired, giving the user very good responsiveness and reducing the required bandwidth. Sending as few as 64 bytes gives the user immediate feedback. The progression of gradually improving images is very effective in informing the user about the state of the instrument and the network connection. The final image can of course have the quality that the instrument can provide or can be limited by considerations such as bandwidth until the user wants a really good image. At that point the user can switch the instrument to a less interactive "slow scan" mode to collect a high-quality image. This can be saved in a file for later examination or sent over a network in the background for remote viewing later in the session or after the session has finished.

We have had success with even the simplest form of wavelets, called Haar wavelets. With this approach average and detail information is computed in a hierarchy of blocks of pixels. Alternatively, Daubechies wavelets provide a more graceful and subtle progression from coarse to fine detail at the expense of performance. We implemented the Haar wavelets in fixed point and integer arithmetic since computations consist of averages, differences, and sums. We use an object-oriented approach in which all wavelet operations are encapsulated in an object that contains a single buffer. The same buffer is used for both the image domain and the scaling function (averages) and wavelet (differences) domain. We use a pair of wavelet objects, one at the image source and one at the remote destination where the user wants to view the image. Starting with the original image, our code performs a series of in-place transformations through an arrangement commonly called filter banks. For a 384 by 384 image, we have encoded six levels of Haar wavelets within 30 milliseconds on a typical Pentium III NT workstation. This is much less time than it takes to move the specimen to a new position or to actually capture the image on the microscope that we were using, and it is definitely faster than is needed for animation. Once the image is transformed to the desired level of averages and wavelets, it is transmitted in pieces (of increasing detail) to the wavelet object at the destination. At each level of detail, the wavelet is inserted into the destination wavelet object and transformed one level back to recover image information for that level. This is extracted from the wavelet buffer and displayed as even more detail is arriving in the buffer. The forward and backward filters are reversible. We use standard TCP/IP to move these image pieces.

High compression ratios (30 or 50 to 1) for lossy images have been reported using schemes called zero trees, bit-plane slicing, and arithmetic encoding schemes. With a 384 by 384 pixel image, this can reduce bandwidth requirements from over 1 million bits per image to a mere 24K bits per image.

Wavelets are useful for more than just progressive image transmission. For example, they can provide a measure of roughness that could be used for automatic algorithms such as focusing and adjusting for astigmatism, and they are used in motion detection and analysis. Because they are multiresolutional, they can be used to analyze and measure structural details at multiple scales.

What it takes and what is needed to network

Remote microscopy is possible with the current Internet. Since ordinary NT workstations are used to control the instruments locally, interfacing packages can link a remote user to an instrument for both control and image viewing. At Hayward we have implemented such an experimental system directly with C++ programs and with Java through a Web interface. We have had this in place for several years [Smith1996a, Smith1996b, Smith1996c, Morgan1998a, Morgan1998b, Morgan1999]. We have used ATM channels to send compressed video and have sent images via ordinary TCP/IP connections at various bandwidths from 56K modem to T3.

Remote microscopy can benefit from advanced networks. Some features like dedicated channels with guaranteed bandwidth and latency and with advanced security are needed for production use and for installations with geographically dispersed instruments. Since these are needs that are common to many applications, it is far better for the network to provide these features than to place this burden on the instrument control application programs.

Asymmetry is fine for connections from instrument to user. Normally, commands from the user to the instrument take very little bandwidth, but images coming from the instrument to the user take considerable bandwidth. This is consistent with the design of current DSL telephony and cable modem protocols.

Multicasting will provide economical ways to distribute images for such applications as teaching, outreach, and mass communication of results.

System architecture

The software design of the FEI/Philips XL series SEMs was a pioneering effort that brought the modern Windows-based user interface to microscopy. Because the software was integrated in the standard way that Windows is organized, a side benefit was the ability to do remote control. This eventually led to our collaboration.

The Tecnai series of TEMs has carried this idea to the next logical level. The figure shows the system architecture of Tecnai, the FEI/Philips series of TEMs. An important characteristic of these instruments is that it is very easy to integrate equipment from third-party suppliers into the software and hardware environments. In the figure, a prototypical case where two third-party products are deployed alongside the TEM software is depicted. The software is built as a client/server system. The bottom side shows the low-level hardware control parts, some of which are supplied by the third-party vendors. The microscope hardware is typically controlled by a number of embedded processor boards, connected by an Ethernet link to the PC of the instrument. The TEM server contains the bulk of the software that is responsible for driving, supervising, and controlling the instrument. The functions in the TEM server can be called directly by the standard Tecnai User Interface. The user interface is notified when conditions in the instrument change, so the various parameters displayed on the screen can be correctly updated. To allow functions in a TEM server to be utilized by a variety of applications, the Scripting Interface provides a subset of all TEM server functionality in such a way that the functions are callable by programs written in languages such as C++, Visual Basic, and Delphi and by scripts written in languages like VBScript, JavaScript, or PERL. In a similar way, some third-party suppliers offer scripting interfaces for their products so that application program writers can control all functions available in the instrument, irrespective of which vendor has supplied them. When it comes to high-level functionality, it is important to make as much use as possible of the functions provided by the lower layers in the architecture. Additionally, as the high-level functions can be application specific, it is important that they can be expanded and changed rapidly. Therefore, these functions are typically implemented as scripting components. Scripting components can again be called from a program written in an arbitrary programming language. With all interfaces and functionality in the whole instrument available for usage in a very general way, it should not come as a great surprise that it is readily possible to embed calls to the various scripting interfaces in a Web page. In this way, user interfaces for specific application areas or target users can be constructed in a short time. By exploiting the security measures inherent in the Web environment, the access to the instrument can be safeguarded while allowing more privileged users rapid access to the desired results. In the instrument itself, it may for commercial reasons be desirable to block certain interfaces from being used without authorization. This is handled in the scripting interfaces.

Results obtained so far

The first results of our own efforts in the field of remote microscopy have been encouraging. We have shown that it is possible to control a Philips XL 40 SEM and a Philips Tecnai TEM from a distance, both over a LAN as well as over an Internet connection. There are no inherent technical limitations to the approach. When we tried to do real work on the instruments, however, the limitations of the setup became clearly visible. We noticed that real operation of an instrument, as opposed to operation during a "proof of concept" demo, requires far more careful design in terms of at which point in the total chain an operation should be done. Hence, we stepped back from the experimental setup and tried to understand better what problem we actually are trying to solve. Remote microscopy, in our view, forces us to concentrate even more than before on the question of what a user is actually trying to accomplish with the instrument and therefore what elementary and high-level functionality must be present. We have come to the conclusion that we really should push remote microscopy forward as the preferred way of working, as development of the associated technology and techniques will have enormous benefit for the user community as a whole.

Already now, users and third parties alike have taken advantage of the complete flexibility of the scripting interfaces in Tecnai and in XL to build novel applications and experimental environments in order to speed up and enhance their research projects. We are encouraged by seeing them using the present facilities in this way. However, to make these instruments fully useful to a greater public, the process of setting up experiments must be made considerably simpler than it is now. We think that we can expect only a minority of the user community to be so experienced in programming that they can write their own applications and/or scripts. In more production-oriented environments, the way of operating the instrument is generally strictly prescribed by a supervisory person because the operators themselves often have a low skill level.

Future directions of research

As may be clear from the results above, we should concentrate our efforts in essentially three areas:

Background of the authors

The authors collaborated extensively on the subject of remote microscopy while Christopher Morgan was, during his 1999-2000 sabbatical leave from California State University Hayward, working as a visiting scientist in the Research and Development group of FEI/Philips Electron Optics in Eindhoven, The Netherlands. Prof. Morgan heads the development of remote microscopy software at the Microscope and Graphics Imaging Center (MAGIC). Auke van Balen is chief technologist at FEI/Philips Electron Optics and was responsible for the design of the software for the Philips XL Scanning Microscope series as well as the Philips Tecnai Transmission Microscope series of instruments.

The authors' e-mail addresses are avb@nl.feico.com and cmorgan@csuhayward.edu.

References

[Allard1998] L. F. Allard, E. Völkl, T. A. Nolan, C. K. Narula, C. Montreuil, W. C. Bigelow, J. E. Mansfield, A Telepresence Microscopy Research Session in the DOE Materials Microcharacterization Collaboratory, Microsc. Microanal. Proceedings 1998, p. 20.

[Breton1996] B. C. Breton, G. Chand, N. H. M. Caldwell, D. M. Holburn, A Java Powered SEM, 11th European Congress on Microscopy (EUREM '96).

[Breton1997] B. C. Breton, G. Chand, N. H. M. Caldwell, D. M. Holburn, Remote Microscopy in the Real World, Microscopy & Analysis, September 1997, pp. 19-21.

[Burton1999] K. Burton, D. L. Farkas, Net Progress, Nature, Vol. 39, Feb 1998, pp. 540-541.

[Caldwell1998] N. H. M. Caldwell, B. C. Breton, D. M. Holburn, Remote Instrument Diagnosis on the Internet, IEEE Intelligent Systems, May/June 1998, pp. 70-76.

[Chand1996] G. Chand, B. C. Breton, N. H. M. Caldwell, D. M. Holburn, World Wide Web Controlled SEM, Scanning: The Journal of Scanning Microscopies, Vol. 18(3), 1996, pp. 201-202.

[Chand1997] G. Chand, B. C. Breton, N. H. M. Caldwell, D. M. Holburn, World Wide Web Controlled Scanning Electron Microscope, Scanning: The Journal of Scanning Microscopies, Vol. 19(4), 1997, pp. 292-296.

[Chumbley1995] L. S. Chumbley, M. Meyer, K. Fredrickson, F. C. Laabs, Computer Networked Scanning Electron Microscope for Teaching, Research, and Industrial Applications, Microsc. Res. Technique, 32, 1995, p. 330.

[Chumbley1996] L. S. Chumbley, M. Meyer, K. Fredrickson, F. C. Laabs, The Instructional SEM Laboratory at Iowa State University, Microsc. Microanal. Proceedings 1996, pp. 396-397.

[Chumbly1997] L. S. Chumbley, M. Meyer, K. Fredrickson, F. C. Laabs, Development of a Multi-User, Networked Scanning Electron Microscope, Scanning: The Journal of Scanning Microscopies, Vol. 18(3), 1997, p. 202.

[Daberkow1998] I. Daberkow, M. Schierjott, Possibilities and Examples for Remote Microscopy Including Digital Image Acquisition, Transfer, and Archiving, Microsc. Microanal. Proceedings 1998, p. 2.

[Ellisman1998] M. H. Ellisman, M. Hadida, D. Greer, M. Wong, S. Lindsey, S. Peltier, S. J. Young, Telemicroscopy: Development of a Collaboratory for Microscopic Digital Anatomy, Microsc. Microanal. Proceedings 1998, p. 12.

[Fan1993] G. Fan, P. Mercurio, S. Young, M. Ellisman, Telemicroscopy, Ultramicroscopy, 52, 1993, pp. 499-503.

[Hunt1998] J. A. Hunt, C. Meyer, Network Control and Tools for Electron Microscopy, Microsc. Microanal. Proceedings 1998, p. 16.

[Koshinski1998] L. S. Koshinski and L. T. Weinrich, Teletutoring -- From Learning to Earning -- West Greene School District Team, RJ Lee Group Team, Microsc. Microanal. Proceedings 1998, p. 26.

[Mabon1999] J. Mabon, G. Metze, Remote Access to Microanalytical Instrumentation at UIUC-MRL/CMM, Scanning: The Journal of Scanning Microscopies, Vol. 21(2), 1999, p. 54.

[Mallat1998] S. Mallat, A Wavelet Tour of Signal Processing, Second Edition, Academic Press, 1998, 1999.

[Mansfield1996] J. Mansfield, The Teaching SEM -- An Example of Real-Time Remote Control SEM, Microsc. Microanal. Proceedings 1996, pp. 394-395.

[Mansfield1998] J. F. Mansfield, A. Adamson, K. Coffman, Real-Time Remote Control of a Scanning Electron Microscope Across the vBNS/Internet 2, Microsc. Microanal. Proceedings 1998, p. 44.

[Miller1999] M. K. Miller, D. J. Larson, Remote Atom Probe Field Ion Microscopy, Ultramicroscopy, 79, 1999, pp. 259-264.

[Morgan1998a] C. L. Morgan, R. A. Hillyard, G. M. Jones, D. L. Pardoe, N. R. Smith, A Model for Internet Access to Remote Visual Scientific Instruments, Microsc. Microanal. Proceedings 1998, p. 10.

[Morgan1998b] C. Morgan, D. Pardoe, N. Smith, Toward a Standard for Remote Microscope Control Systems, Scanning: The Journal of Scanning Microscopies, Vol. 20(2), 1998, pp. 110-116.

[Morgan1999] C. Morgan, D. Sarmah, S. Alexander, D. McQueen, D. Pardoe, N. Smith, A Software Framework for Remote Access to Visual Scientific Equipment, Scanning: The Journal of Scanning Microscopies, Vol. 21(2), 1999, p. 55.

[O'Keefe1996] M. A. O'Keefe, J. Taylor, D. Owen, B. Crowley, K.H. Westmacott, W. Johnston, U. Dahmen, Remote On-Line Control of a High-Voltage In-Situ Transmission Electron Microscope With a Rational User Interface, Microsc. Microanal. Proceedings 1996, pp. 384-385.

[O'Keefe1998] M. A. O'Keefe, J. R. Taylor, D. K. Owen, LBNL and the Materials Microcharacterization Collaboratory, Microsc. Microanal. Proceedings 1998, p. 42.

[Parvin1996] B. Parvin, J. Taylor, B. Crowley, L. Wu, W. Johnston, D. Owen, M. O'Keefe, U. Dahmen, Telepresence for In-Situ Microscopy, IEEE International Conference on Multimedia Systems and Computers, June 1996.

[Parvin1997] B. Parvin, Distributed Virtual Microscopy and Microcharacterization, A white paper, May 1997, http://www-itg.lbl.gov/MMC/white_paper.html.

[Postek1998] M. T. Postek, M. Bennett, N. J. Zaluzec, T. Wheatley, S. Jones, National Institute of Standards and Technology: Texas Instruments Industrial Collaboratory Testbed, Microsc. Microanal. Proceedings 1998, p. 22.

[Ramakrishna1998] B. L. Ramakrishna, A. Razdan, J. Sun, E. Ong, A. A. Garcia, Interactive Nanovisualization for Science and Engineering Education, Microsc. Microanal. Proceedings 1998, p. 46.

[Ritchie1998] N. W. M. Ritchie, P. V. Woods, A Generic Protocol for Controlling an SEM Over a TCP/IP Link, Microsc. Microanal. Proceedings 1998, p. 24.

[Ruijter1998] W. J. de Ruijter, J. K. Weis, Integrated Data Acquisition and Instrument Control for Local and Remote Operation of Electron Microscopes Equipped with Imaging Detectors and Spectrometers, Microsc. Microanal. Proceedings 1998, p. 48.

[Schmid1998] M. F. Schmid, P. Matsudaira, M. T. Dougherty, M. B. Sherman, C. Henn, W. Chiu, Televisualization: An Aid to Collaborative Research in Molecular Structure Biology, Microsc. Microanal. Proceedings 1998, p. 32.

[Smith1996a] N. R. Smith, N. Fegan, R. E. Tullis, C. L. Morgan, Remote Operation of an SEM From Distant Classrooms, Microsc. Microanal. Proceedings 1996, pp. 414-415.

[Smith1996b] N. Smith, D. Pardoe, C. Pelote, M. Honer, C. Morgan, Remote Control of a Scanning Electron Microscope Using Asynchronous Transfer Mode, Microscopy Today, 96(2), 1996, p. 18.

[Smith1996c] N. R. Smith, R. E. Tullis, D. Pardoe, C. Pelote, N. Fegan, C. L. Morgan, Remote Access to a Computer Operated Scanning Electron Microscope (SEM), Proc. 4th CSU EM Microscopy Colloquium, San Diego, May 4-5, 1996.

[Thorne1998] B. B. Thorne, E. B. Steel, A. J. Fahey, A Multi-Instrument Virtual Collaborative Environment via the World Wide Web at NIST, Microsc. Microanal. Proceedings 1998, p. 30.

[Vesenka1999] J. Vesennka, Remote Microscopy Operation at CSU Fresno, Scanning: The Journal of Scanning Microscopies, Vol. 21(2), 1999, p. 53.

[Völkl1996a] E. Völkl, Imaging Needs for Electron Microscope Distance Learning and Remote Operation, Adv. Imaging, 11, 1996, p. 31.

[Völkl1996b] E. Völkl, L. F. Allard, T. A. Nolen, D. Hill, M. Lehmann, A Simple and Inexpensive Route to Remote Electron Microscopy, Microsc. Microanal. Proceedings 1996, pp. 386-387.

[Völkl1997a] E. Völkl, L. F. Allard, J. Bruley, D. B. Williams, Undergraduate TEM Instruction by Telepresence Microscopy over the Internet, J. Micros., 187, 1997, pp. 139-143.

[Völkl1997b] E. Völkl, L. F. Allard, T. A. Nolen, D. Hill, M. Lehmann, Remote Operation of Electron Microscopes, Scanning: The Journal of Scanning Microscopies, Vol. 19(4), 1997, pp. 286-291.

[Wolf1998] G. Wolf, D. Petersen, M. Dietel, I. Peterson, Telemicroscopy via the Internet, Nature, Vol. 39, Feb 1998, pp. 613-614.

[Wright1998] M. C. Wright, The Materials Microcharacterization Laboratory, Microsc. Microanal. Proceedings 1998, p. 6.

[Youngblom1999] J. H. Youngblom, J. J. Youngblom, J. Wilkinson, Remote Access Confocal Microscopy, Scanning: The Journal of Scanning Microscopies, Vol. 21(2), 1999, p. 53.

[Zaluzec1996a] N. J. Zaluzec, Tele-Presence Microscopy: An Interactive Multi-User Environment for Collaborative Research Using High Speed Networks and the Internet, Microsc. Microanal. Proceedings 1996, p. 14.

[Zaluzec1996b] N. J. Zaluzec, R. Stevens, R. Evard, T. Disz, R. Olson, T. Kuhfuss, Tele-Presence Microscopy & the ANL LabSpace (eLab) Project, PowerGrid, 1.01, http://www.powergrid.com/1.01/telemicro-wp.html (1996).

[Zaluzec1998] N. J. Zaluzec, Tele-Presence Microscopy: A Progress Report, Microsc. Microanal. Proceedings 1998, p. 18.

Some Web Sites

[Argonne] N. J. Zaluzec, AAEM TelePresence Microscopy Site, Argonne National Laboratory
http://tpm.amc.anl.gov/

[bugscope] B. Carragher, C. Potter, Bugscope
http://bugscope.beckman.uiuc.edu/

[Cambridge] G. Chand et al., Telemicroscopy and Remote Microscopy
http://www2.eng.cam.ac.uk/~bcb/remote1.htm/

[DOE2000] Collaboration, DOE2000
http://tpm.amc.anl.gov/MMC/HomePage.html/

[Fabbri] P. L. Fabbri, Centro Interdisciplinaire Grandi Instrumenti, University of Modena, Italy
http://www.cigs.unimo.it/

[IRSA] C. Morgan, Interactive Remote Shared Access to MAGIC
http://www.magic.csuhayward.edu/

[JEOL] JEOL, Company Web Page
http://www.jeol.co.jp/

[MAGIC] C. Lauzon, Microscope and Graphic Imaging Center
http://www.csuhayward.edu/sci/sem/

[Mems-exchange] A. Kuckling, Mems-exchange
http://www.mems-exchange.org/exchange/software/microscope/

[Michigan] J. Mansfield, Teaching SEM
http://emalwww.engin.umich.edu/people/jfmjfm/tsem/imp.html/

[MMC] A. Razdan, Remote Scientific Visualization and Operation Using Scanning Probe, Microscope, MMC Architectural Workshop, Oak Ridge National Laboratory
http://www.ornl.gov/doe2k/workshop/html/talks.html/

[MTL] MIT, Microsystems Technology Laboratories
http://www-mtl.mit.edu/research/microscope.html/

[NAMT] M. T. Postek, National Advanced Manufacturing Testbed
http://www.mel.nist.gov/namt/

[NCEM] M. O'Keefe, National Center for Electron Microscopy
http://ncem.lbl.gov/ncem.html/

[NetSEM] LEO, NetSEM
http://www2.leo-em.co.uk:8080/NetSEM.htm/

[ORNL] E. Völkl, ORNL-MMC
http://www.ornl.gov/doe 2k/

[SDSU] S. Barlow, Remote Access to the Scanning Electron Microscope
http://www.chemistry.sdsu.edu/EM_Facility/CUCMEoutreach.html/