Making The Invisible Visible

IN 1948 JAMES HILLIER, THE RCA PHYSICIST LARGELY responsible for developing the first commercial electron microscope in America, summarized the progress that had keen made and the challenges that lay ahead in mapping the microscopic world using electrons. In a speech to the members of the American Association for the Advancement of Science, he asked his listeners to perform a thought experiment. Suppose, he said, a scientist wanted to examine an entire plant or animal under an electron microscope. “He would soon find that he would not live long enough to complete his work,” Hillier said. It would take 40 years to get a fast glance at the structures in one square inch of the surface of his sample; to photograph the same square inch at the highest magnification thrn available would require 6,000 years. Millier and a colleague had themselves made 40,000 photographic exposures over 9 years, at an average magnification of 10,000 times. “We now realize,” he told his audience, “that in this time we have photographed only one square millimeter of our world.”

But what a square millimeter it was. The electron microscope had allowed biologists to glimpse for the first time the inner structures of the cell, to measure the size of a virus, and to watch bactcriophages—single-cell organisms that attack invading bacteria—in operation. Metallurgists could now see the treacherous peaks and chasms of a smooth-to-the-touch metal surface, and chemists could measure the invisibly small carbon-black particles that mysteriously added strength to automobile rires. All this from an instrument that hadn’t existed before 1931 and that most scientists had believed would prove useless.

Dennis Gabor, a Hungarian physicist who contributed significantly to electron microscopy and later invented holography (for which he won the Nobel Prize in 1971), recalled a conversation he had with the physicist Leo Szilard in 1928 at Merlin’s Cafe Wien. Szilard suggested that Gabor go to work on an electron microscope; Gabor replied, “What is the use of it? Everything under the electron beam would burn to a cinder!” Szilard came up with his own electron-microscope design, which he patented in 1931. The physical chemist Ladislaus L. Marton, of the Université Libre in Brussels, summed up the spirit of the pioneers of electron microscopy when he later remarked. “Let it burn, but let us look at the cinder.”

The electron microscope was born to find a way out of a dead end. In 1873, after almost 300 years of continuous improvements, optical microscopy finally faced its limits. Ernst Abbe, a German physicist and microscope developer, discovered those limits. His mathematical analyses of lenses and compound microscope designs took optics out of the realm of craftsmanship and into that of physics, but they also revealed an insurmountable problem: The best resolution of a microscope cannot exceed about half the wavelength of the light it uses. Once the optics of a microscope have been optimized, nothing will resolve smaller objects except decreasing the wavelength of the light source.

NATURE HAD SET THE LOWEST WAVELENGTH OF VISIBLE light at about 400 nanometers (a nanometer is a ten-millionth of a centimeter), so the best resolution that could come from an optical microscope was about 200 nanometers. Biologists could see the nucleus of a cell, but very little beyond that. The basic shape of a bacterium such as that of tuberculosis or diphtheria could be observed, but none of its internal structure. Viruses, which could be as small as 10 nanometers, could not be seen at all. Abbe’s discouragement was evident when he wrote, “It is poor comfort to hope that human ingenuity will find means and ways to overcome this limit.” In 1904 Zeiss produced a microscope using shorter-wavelength ultraviolet light. The image could be viewed on a fluorescent screen and recorded on film. The ultraviolet microscope pushed back the barrier somewhat but ran into problems with the limitations of optical lenses. So it wasn’t a solution either. Yet despite the bleak outlook, a possible solution was soon found.

German physicists had been embedding electrodes in sealed, gas-filled glass tubes since 1858. When a potential difference was applied to the electrodes, a current developed between them, causing gases in the tubes to emit light. In 1897 the British physicist J. J. Thomson identified these cathode rays as streams of electrons, and investigators began experimenting with cathode-ray tubes as a possible basis for oscillographs for observing electrical phenomena in power lines, and for television, which would emerge several decades later.

The first breakthrough in microscopy came in 1923, when the French physicist Louis de Broglie proposed that those electron particles in motion also behaved as waves. A simple calculation showed that high-voltage apparatus could give them wavelengths about five orders of magnitude shorter than optical waves. And since they consisted of charged particles, they could be deflected by magnetic or electrostatic fields. In 1926 Hans Busch of the University of Jena demonstrated mathematically that a doughnut-shaped ring of magnets could serve as an electron lens, focusing electrons just as a convex glass lens focused light. In 1927 Clinton Davisson and Lester Germer, at Bell Telephone Laboratories in New York City, showed that electrons traveling through crystals bent like light waves, proving experimentally that electron waves indeed existed. Similar work was performed almost simultaneously in England by G. P. Thomson, J. J. Thomson’s son. Davisson and G. P. Thomson later shared a Nobel Prize for their work.

It occurred to many physicists before long that a combination of electron waves and electron lenses could produce an electron microscope, but it is not known who had the idea first. Whoever it was, he or she no doubt met with skepticism like Gabor’s and dropped the idea immediately. Yet within five years of Gabor’s remarks, many of the initial objections had been overcome, and several primitive laboratory versions were in operation. They were crude, but they performed all the basic functions.

In an ordinary microscope, light shines up through a thin sample of the object under observation. A series of glass lenses magnifies and focuses the light’s rays on the viewer’s retina (or a piece of photographic film), giving an enlarged image of the sample being observed. In an electron microscope, the light source is replaced by an electron gun, usually located at the top of a vertical column. Electrons are shot from a cathode —a heated V-shaped tungsten wire—and are accelerated to very high velocities toward the receiving anode at the other end. The acceleration is achieved by maintaining an enormous potential difference between the cathode and the anode, up to several hundred thousand volts.

The speeding electrons pass through a small hole, or aperture, at the exit of the gun, which allows only those traveling close to the axis of the microscope tube to proceed. They then pass through the center of the first magnetic lens, called the condenser. A magnetic lens consists of a ring-shaped wire coil with an iron coating that focuses electrons in much the same way that a glass lens focuses light. The microscopist can adjust the lens by varying the current in the coils, and thus the strength of the magnetic field. The condenser lens produces a narrow, focused beam of electrons to “illuminate” an extremely thin film of the sample being observed.

Electrons passing through the sample are scattered by varying amounts, depending upon (among other things) the density of the sample. This differential scattering provides the contrast between detailed features the microscopist wants to see. Regions of the sample containing different chemical elements will scatter electrons differently as well. The scattered electrons exiting the sample pass through the objective lens next. It focuses a magnified image of the illuminated region of the sample farther down the microscope tube, typically enlarging it a hundred times. A portion of that image, selected by moving an aperture, is focused and magnified further by an intermediate lens, followed by a final magnification by a projector lens, near the bottom of the microscope column.

These sequential magnifications produce a highly magnified, high-resolution image of a portion of the sample on a phosphorescent white screen, typically coated with a substance such as zinc sulfide that emits light when electrons strike it. The microscopist looks at faint, hazy images on this screen. By lifting the screen to reveal a photographic plate underneath, he or she can record an “electron micrograph” of the image on high-density film. This creates a permanent record of the findings.

Many challenges stood in the way of a successful electron microscope. Extremely stable power sources would be needed, because fluctuations in power change the wavelengths of the electrons, yielding fuzzy images, but such stable sources were not yet available. The high voltages involved would put inventors’ lives at risk. Moreover, while visible light makes its way through the air virtually unimpeded, electron waves crash into air molecules and scatter widely, so an electron microscope tube would have to be kept under a higher vacuum than anybody could create at the time. Plus, there was that little problem of turning samples into cinders.

Not to be deterred, several groups around the world began to attack the challenge simultaneously. Max Knoll and his student Ernst Ruska, at the Technische Hochschule in Berlin, were the first with a working prototype. Knoll was a professor of engineering leading a group of students trying to make a cathode-ray oscillograph. To build one, they needed an electron lens. When searching for an undergraduate thesis topic, Ruska noted that Busch’s theory of an electron lens had never been verified experimentally. So he did the job himself, wrapping coils of wire around a ring-shaped iron core to produce a lens with the exact dimensions Busch had specified and then trying to focus the electrons from a cathode with it. He was able to take sharp pictures of a o. 3-millimeter aperture in front of the cathode-ray tube. He published these first electron micrographs in his undergraduate thesis, in 1929.

For his graduate thesis, in 1930, Ruska experimented with electrostatic lenses instead of magnetic ones. After obtaining unsatisfactory results, he returned to the magnetic method, hoping to see if higher magnifications could be obtained by placing a second electron lens behind the first. He sketched the two-stage instrument in his notebook on March 9, 1931, and took the first magnified images of a platinum grid with it the next month. The total magnification was only 17.4 times, but it was enough to earn the primitive instrument posterity’s designation as the world’s first electron microscope. (The 192.9 model, with only one lens, was considered more a magnifying glass than a microscope.)

Knoll announced the results at a lecture at the Technische Hochschule. “Not wishing to be accused of showmanship,” Ruska noted when he won a Nobel Prize for this breakthrough, in 1986, “Max Knoll and I agreed to avoid the term electron microscope.” Meanwhile, Gunther Reinhold Rudenberg, research director of the Siemens-Schuckert company in Germany, applied for a patent on his idea for an electron microscope on May 30, 1931. This led to patent battles that lasted until after World War II. In 1932 another German researcher claimed success using electrostatic lenses, with oppositely charged metal surfaces enclosing a static field of electricity. This required less power but had shortcomings that kept it from dominating in the long run.

Word of all this work spread quickly. In Brussels, Ladislaus Marton read about it and came up with the use of a hot tungsten wire for the electron gun. This proved more efficient and controllable than the cold cathode-ray tube and has remained the standard to this day. Marton also addressed the problem of samples burning up. In a letter to Nature magazine in 1934, he suggested three possible solutions: “1) Intense cooling of the object (for example, by contact with an extremely thin metal foil; 2) impregnating the object with a substance which makes the object less destructible; 3) impregnating the object in such a way that a framework of the object is preserved although the object itself is destroyed.” He investigated the first notion by placing samples on thin aluminum foil, but the foil scattered the electrons, making a fuzzy image. Next, he tried impregnating samples with heat-resistive osmium tetroxide and produced the first successful micrograph of a histological specimen, albeit one with no internal detail visible. Finally, he made the key observation that led to success in examining biological specimens: Only electrons that not only strike a sample but are absorbed heat it up. If the sample is cut thin enough to allow the electrons to pass through unabsorbed, almost no heat will be generated.

But the preparation of thin enough samples proved difficult. At the time, microtomes, instruments that slice specimens, could shave tissue samples as thin as 1,000 nanometers, but this was far too thick for the electron microscope. Structural details were lost in the fog of scattered electrons. It would take several more years and improved microtomes to obtain good micrographs of biological specimens; in the meantime, investigators had to be content with looking at carbonized fossils of their samples.

IN THE LATTER HALF OF 1933, RUSKA SPENT FIVE MONTHS building a second, improved electron microscope. It had a maximum accelerating power of 75,000 volts, up from 50,000, and three magnetic lenses: one condenser and two magnifiers. All three lenses were encased in soft iron and cooled by water. He was able to obtain micrographs of carbonized cotton fibers using this instrument, and the next year he took pictures of a housefly’s wings with a resolution of 40 nanometers, five times better than any optical microscope. But convinced that the improvements necessary to produce a truly practical and commercial electron microscope were years away, and discouraged by a lack of money to pursue them, Ruska took a job in industry in December 1933.

In 1937 a pair of graduate students at the University of Toronto, James Hillier and Albert Prebus, began their own investigations into electron microscopy under the direction of Eli F. Burton, the chairman of the department of physics. Their work—along with simultaneous efforts at Siemens, in Germany—would lead to the commercial electron microscope as we know it. Burton had worked under J. J. Thomson at Cambridge from 1904 to 1906 and had long been interested in the science of colloids, mixtures in which extremely fine particles of one substance are dispersed in another. In many cases, the particles are too small to be seen by an optical microscope. He first encountered an electron microscope when he visited Knoll in Germany in 1935 while attending a conference on electron microscopy, and he quickly grew interested in its possibilities.

Albert Prebus was a quiet man who had originally expected to study spectroscopy; Hillier was more outgoing. Just a few years earlier he had thought he would follow the family path into a career in art. But an enterprising high school teacher, noticing his facility with a ham radio, applied for a fellowship in physics on his behalf. He got the fellowship and four years later had just completed his undergraduate work in mathematics and physics at Toronto.

Searching for a project for his graduate studies, Hillier learned of the department chairman’s interest in electron microscopes. “I had never heard those two words put together before,” he recalls. “What did electrons have to do with glass lenses and microscope tubes?” Off he went to the library to read the small body of existing literature on the subject.

Cecil Hall, another of Burton’s graduate students, had already built a primitive electron microscope that made no provision for a sample and could take pictures only of the emitting cathode itself. Hall had gone on to work at Eastman Kodak. Hillier and Prebus studied Hall’s design, figured out what could be improved upon, and decided to build an electron microscope themselves.

They spent their Christmas holidays in 1937 working day and night on it. Prebus was thorough and meticulous, while Hillier was more of a risk taker, and the combination of personalities worked well. They were able to show Burton their first image, of the edge of a razor blade, when he returned from the holidays. He authorized them to try to make an improved microscope, but with money almost nonexistent in the Depression, he forbade them to spend a nickel on the project. They could scrounge through used equipment all they wanted and draw upon any resources in the building, but they couldn’t buy a thing.

The restriction proved to be a boon rather than a hindrance. Forced by circumstances, the two young researchers found themselves spending all night in the machine shop, making the small parts that they couldn’t buy. Hillier attributed his later success at RCA to his ability to make things himself in a machine shop. They made grease for the microscope’s vacuum seals by melting Vaseline together with gum rubber or rubber bands in the laboratory. They found a 60-hertz transformer to power the instrument, but they couldn’t run it from Ontario’s 25-hertz electric lines, so they searched some more and found a 60-hertz generator in the basement. They experimented with commercial photographic film but found that it gave off vapors in the vacuum for hours and then shattered when removed, so they used a glass photographic plate instead.

In four months, they completed what has since been known as the 1938 Model Toronto electron microscope. It consisted of a six-foot-tall vertical tube in six sections. Each section was joined to the next using plane-lapped, vacuum-greased seals and could be slid horizontally to align the electron beam. The electron gun, with an acceleration energy of 45,000 volts, used a tungsten filament as the electron source. Three water-cooled, magnetic electron lenses—condenser, objective, and projector —produced a magnified image of the sample on a screen at the bottom of the tube. By lifting the screen and exposing a photographic glass plate, they could record micrographs. Experience soon showed that the camera had to be isolated from the rest of the apparatus with an air lock, to make pumping-down easier after changing photographic plates.

The high voltages involved required constant caution. Hillier recalls that one time Prebus went down to the basement to turn off the generator, was distracted by a conversation, and returned to the microscope lab without having shut off the power. He touched the electron microscope with his hand and was “laid out flat” by the current. He had medical problems for several months.

The 1938 Model Toronto had many bugs, including contamination from the grease, which caused carbon deposits to build up inside. Despite the water cooling, the electron lenses would heat up enough to change the resistance of the coils and the strength of the magnetic fields, making everything blurry. The high-voltage power supply suffered fluctuations and magnetic interference when streetcars passed by, and the vibration caused by a truck coming up the driveway could throw the image out of focus. Nonetheless, Hillier and Prebus were able to obtain satisfying results from the instrument.

In June 1938 the American Physical Society held its annual meeting in Toronto, and Hillier and Prebus showed some of the pictures they had taken. Soon they were inundated with requests from biologists, chemists, and metallurgists wanting to look at things. Investigators for the Canadian government went to Toronto to examine asbestos fibers, suspecting that they were causing lung disease in asbestos miners.

WHILE HILLIER AND PREBUS WERE PUTTING TO gether their 1938 Model Toronto, Ernst Ruska was finally persuading Siemens to attempt a practical, commercial instrument. His younger brother Helmut, who had recently completed his medical studies, had kept him interested. Even before it was clear that biological materials could survive an electron beam, Helmut was convinced of the value of the electron microscope in fighting diseases and had the utmost confidence that Ernst could solve any problems along the way. He encouraged his brother to keep seeking funding, and in 1936 they got Dr. Richard Siebeck, director of a leading Berlin medical clinic, to back their cause. That October, Siebeck wrote, “What seems attainable now, I consider to be so important, and success seems to me so close, that I am ready and willing to advise on medical research work and to collaborate by making available the resources of my Institute.” This won over the people at Siemens, so Ruska set up an electron-optics laboratory there.

SO FAR, EVERY ELECTRON MICROSCOPE HAD REQUIRED delicate fingers and an insider’s knowledge of its quirks to produce a viable image. If electron microscopy was ever to become an everyday tool of the scientist, a far more robust, easy-to-use machine would be needed. The two major competitors in the race to build one were Siemens and the Radio Corporation of America (RCA). Most other companies opted out, believing that the world market would absorb only a small number of the devices.

RCA began its effort in 1938, a year after Siemens. Vladimir Zworykin, a Russian-born physicist whose work for Westinghouse and RCA had already made him one of the fathers of television, began by bringing Marton over from Europe. Within a year, Marton produced the RCA Type A, essentially a repackaging of a microscope he had made in 1935. Now it was encased in a second evacuated metallic shell to provide further magnetic shielding and vibration stability, but the increased volume that had to be pumped down strained the vacuum pumps, and air got into the system, leading to contamination. It could produce good results with constant attention, but the Type A was hardly the sturdy, easy-to-use commercial instrument Zworykin had in mind.

In Germany, meanwhile, Ruska’s extensive experience moved the Siemens effort along quickly. The firm had two prototypes built by 1938 and delivered the first commercial instrument to the Hoechst group of IG Farbenindustrie late in 1939. It had separate air locks for the sample and photographic ports to make pumping-down easier, but it needed bulky, high-maintenance batteries to maintain a constant power supply. Siemens soon set up a lab where visiting scientists could bring samples for study. By the end of World War II, the company had delivered more than 30 electron microscopes.

RCA forged ahead with its own effort. In 1940 Hillier and Prebus were invited there to interview for a position. Zworykin asked just one basic question: “How fast can you build one of these things?” According to Hillier, “I crossed my fingers behind my back and said ‘six months.’” He was hired. Prebus, to his surprise, was not; he signed on at Ohio State University instead, where he later built his own electron microscope and made significant contributions to the field.

Hillier was wrong in his estimate. His team took only four months to turn out its first commercial electron microscope, the RCA Type B. Essentially an improved version of the 1938 Model Toronto, it incorporated a new 60,000-volt power supply that constantly checked its output against a reference battery and corrected it to keep the output stable. Neoprene replaced seals made of vacuum grease, reducing contamination and maintaining a better vacuum. RCA sold the prototype in late 1940 to American Cyanamid for $10,000, and the money paid for the initial cost of the project.

It turned out that Zworykin hadn’t had funding for Hillier and had been playing a game against the clock. “Zworykin figured it would take nine months for the accountants to figure out what was going on,” Hillier says. “We built the thing in four months, so we had five months to become famous.”

The triumph of Hillier’s microscope over Marion’s contributed to oersonal differences that auicklv led to tension between the two. Hillier told Zworykin one of them would have to go, and before long Marton was on his way to Stanford University to lead an electron-microscopy program there. Hillier began working closely with leading biologists and bacteriologists. One of them, Wendell Stanley of the Rockefeller Institute, a future Nobel Prize winner, helped prove the worth of the electron microscope with a single experiment. Having isolated the tobacco mosaic virus as the cause of a destructive tobacco disease, he had spent n years using numerous physical and chemical techniques, including ultracentrifuging and x-ray diffraction, to obtain an estimate of the virus’s size and shape. On September zo, 1940, Hillier placed a sample of the virus in his electron microscope and “in 2.0 minutes we showed him what it had taken n years to do by other means.” The electron microscope confirmed Stanley’s findings.

Even more exciting was catching a bacteriophage, a virus that destroys bacteria, in the act. Between 1940 and 1942., three different teams did so. Salvador Luria, a bacteriologist at Columbia University, and Thomas Anderson, a National Research Council fellow at RCA’s laboratories, started by placing a drop of Esckerichia coli bacteria on a thin membrane and adding a drop containing bacteriophage anti-coli PC. An electron micrograph taken five minutes later with the RCA Type B showed bacteriophage cells, each looking like a tadpole with a head 80 nanometers across and a tail 130 nanometers long, swimming toward a bacterium. A few of the bacteriophage cells had already attached to the wall of the bacterium. After 30 minutes, another micrograph showed the bacterium completely destroyed, while the bacteriophage cells had multiplied a hundredfold in feeding on it.

INDUSTRIAL CHEMISTS AT THE TIME WERE EXTREMELY interested in the structure and behavior of colloids, whose particles can be too small to be seen by an optical microscope. These included carbon black, gelatin, starch, latex, dye pigments, and fog and smoke particles, among other things. The first of them to be studied was carbon black. Its main use was in automobile tires, where it greatly increased strength and wear-resistance. Why carbon black added such durability was not known. One theory proposed that the particles were notched like tiny gears and interlocked to strengthen the rubber. Another held that they were smooth and worked some undetermined other way.

In 1939 William B. Wiegand, of the Columbian Carbon Company, decided to solve the mystery. He established a fellowship at the University of Toronto (his alma mater) to work on it using a newly built microscope. After several months and hundreds of micrographs, the particles were determined to be round, completely smooth, and so small they had a total surface area of five million square inches per pound. This cumulative expanse of surface made possible extensive bonding with the rubber, which was what provided the strengthening.

Metallurgists still faced a huge problem with electron microscopy: How could they examine the surface of a thick metal sample, for instance, to study the wear patterns of an airplane part? The density of the metal would prevent electrons from passing through the sample, and thin samples of bulk metals couldn’t simply be “shaved off” as could tissue samples. The answer came in the form of surface “replicas.” Zworykin and Edward G. Ramberg, a colleague of his at RCA, discovered they could coat the surface of a metal sample with collodion or various plastics to form a negative that could be removed without distortion. Coating this negative surface with a very thin film of metal and then dissolving the collodion or plastic with an acid solution would yield a positive impression of the metal surface. This positive, low-density replica, through which electrons could be transmitted, reproduced the original surface down to fine detail. When placed in an electron microscope, it would show the soaring peaks and plunging chasms of what had seemed to be smooth on the original.

A better solution for metallurgists, one that enabled them to view a metal surface directly, would soon arrive in the form of the scanning electron microscope. In 1934 Max Knoll had gone to work on television technology at Fernseh AG in Berlin and observed that when a high-energy electron beam strikes the surface of an object, low-energy “secondary” electrons are emitted from its surface. It later emerged that these secondary electrons can be collected by a detector above the surface and give a picture of the spot where the beam is striking. By scanning the electron beam systematically across the sample, much the way an electron beam scans a television picture tube, row by row, a complete picture of a scanned area can be obtained. Synchronizing one electron beam scanning a sample surface with another scanning a television picture tube gives an immediate, magnified picture of the surface being scanned.

Baron Manfred von Ardenne, a prolific German pioneer in the use of cathode-ray tubes for television, built a primitive scanning electron microscope in 1937. (This was not a scanning electron microscope in the modern sense; it scanned the sample but used transmitted electrons rather than secondary electrons for imaging.) Zworykin, Hillier, and R. L. Snyder, at RCA, built a refined version in 1940 that used secondary electrons, but it would take some time for the technique to be perfected. Not until 1965 was a commercial version available.

An additional advantage of secondary electrons is that their energy can indicate what chemical element they come from. Hillier was the first person to identify electrons according to their energies and thereby obtain a chemical microanalysis of the spot under the electron beam. Now researchers could map the chemical composition of a sample at high resolution.

The market for electron microscopes was not, as some had feared, exhausted by a few instruments. Siemens and RCA continued to develop and improve their commercial electron microscopes over the following decades, and other companies joined them. Advances included the addition of a third magnifying lens, introduced by Marton at MIT and separately by the team of Ernst Ruska and his long-time colleague Bodo von Borries in 1942-43; the development of the “astigmator” by Hillier in 1945 to correct an astigmatism problem in electron lenses similar to that in human vision; and further improvements to microtomes. More recent refinements, such as the scanning-tunneling electron microscope and the atomic-force microscope, have made the atom-by-atom assembly of new materials a possibility. (Scanning-tunneling microscopes use a probe with a nanometer-size tip from which electrons jump across a vacuum gap to the sample, allowing the manipulation of atoms one by one.)

Ruska and Hillier jointly received the Albert Lasker Award in 1960 for the contributions their work had made to medicine, and Hillier was elected to the National Inventors Hall of Fame in 1980. Today, the hundreds of electron microscopes in laboratories around the world are indispensable for the study of medicine, chemistry, materials, and solid-state physics. They have opened a world that was simply invisible before the taming of the electron.