« Reading List: Fashion, Faith, and Fantasy | Main | Reading List: Time in Powers of Ten »

Thursday, October 13, 2016

Reading List: The Perfect Machine

Florence, Ronald. The Perfect Machine. New York: Harper Perennial, 1994. ISBN 978-0-06-092670-0.
George Ellery Hale was the son of a wealthy architect and engineer who made his fortune installing passenger elevators in the skyscrapers which began to define the skyline of Chicago as it rebuilt from the great fire of 1871. From early in his life, the young Hale was fascinated by astronomy, building his own telescope at age 14. Later he would study astronomy at MIT, the Harvard College Observatory, and in Berlin. Solar astronomy was his first interest, and he invented new instruments for observing the Sun and discovered the magnetic fields associated with sunspots.

His work led him into an academic career, culminating in his appointment as a full professor at the University of Chicago in 1897. He was co-founder and first editor of the Astrophysical Journal, published continuously since 1895. Hale's greatest goal was to move astronomy from its largely dry concentration on cataloguing stars and measuring planetary positions into the new science of astrophysics: using observational techniques such as spectroscopy to study the composition of stars and nebulæ and, by comparing them, begin to deduce their origin, evolution, and the mechanisms that made them shine. His own work on solar astronomy pointed the way to this, but the Sun was just one star. Imagine how much more could be learned when the Sun was compared in detail to the myriad stars visible through a telescope.

But observing the spectra of stars was a light-hungry process, especially with the insensitive photographic material available around the turn of the 20th century. Obtaining the spectrum of all but a few of the brightest stars would require exposure times so long they would exceed the endurance of observers to operate the small telescopes which then predominated, over multiple nights. Thus, Hale became interested in larger telescopes, and the quest for ever more light from the distant universe would occupy him for the rest of his life.

First, he promoted the construction of a 40 inch (102 cm) refractor telescope, accessible from Chicago at a dark sky site in Wisconsin. At the epoch, universities, government, and private foundations did not fund such instruments. Hale persuaded Chicago streetcar baron Charles T. Yerkes to pick up the tab, and Yerkes Observatory was born. Its 40 inch refractor remains the largest telescope of that kind used for astronomy to this day.

There are two principal types of astronomical telescopes. A refracting telescope has a convex lens at one end of a tube, which focuses incoming light to an eyepiece or photographic plate at the other end. A reflecting telescope has a concave mirror at the bottom of the tube, the top end of which is open. Light enters the tube and falls upon the mirror, which reflects and focuses it upward, where it can be picked off by another mirror, directly focused on a sensor, or bounced back down through a hole in the main mirror. There are a multitude of variations in the design of both types of telescopes, but the fundamental principles of refraction and reflection remain the same.

Refractors have the advantages of simplicity, a sealed tube assembly which keeps out dust and moisture and excludes air currents which might distort the image but, because light passes through the lens, must use clear glass free of bubbles, strain lines, or other irregularities that might interfere with forming a perfect focus. Further, refractors tend to focus different colours of light at different distances. This makes them less suitable for use in spectroscopy. Colour performance can be improved by making lenses of two or more different kinds of glass (an achromatic or apochromatic design), but this further increases the complexity, difficulty, and cost of manufacturing the lens. At the time of the construction of the Yerkes refractor, it was believed the limit had been reached for the refractor design and, indeed, no larger astronomical refractor has been built since.

In a reflector, the mirror (usually made of glass or some glass-like substance) serves only to support an extremely thin (on the order of a thousand atoms) layer of reflective material (originally silver, but now usually aluminium). The light never passes through the glass at all, so as long as it is sufficiently uniform to take on and hold the desired shape, and free of imperfections (such as cracks or bubbles) that would make the reflecting surface rough, the optical qualities of the glass don't matter at all. Best of all, a mirror reflects all colours of light in precisely the same way, so it is ideal for spectrometry (and, later, colour photography).

With the Yerkes refractor in operation, it was natural that Hale would turn to a reflector in his quest for ever more light. He persuaded his father to put up the money to order a 60 inch (1.5 metre) glass disc from France, and, when it arrived months later, set one of his co-workers at Yerkes, George W. Ritchey, to begin grinding the disc into a mirror. All of this was on speculation: there were no funds to build a telescope, an observatory to house it, nor to acquire a site for the observatory. The persistent and persuasive Hale approached the recently-founded Carnegie Institution, and eventually secured grants to build the telescope and observatory on Mount Wilson in California, along with an optical laboratory in nearby Pasadena. Components for the telescope had to be carried up the crude trail to the top of the mountain on the backs of mules, donkeys, or men until a new road allowing the use of tractors was built. In 1908 the sixty inch telescope began operation, and its optics and mechanics performed superbly. Astronomers could see much deeper into the heavens. But still, Hale was not satisfied.

Even before the sixty inch entered service, he approached John D. Hooker, a Los Angeles hardware merchant, for seed money to fund the casting of a mirror blank for an 84 inch telescope, requesting US$ 25,000 (around US$ 600,000 today). Discussing the project, Hooker and Hale agreed not to settle for 84, but rather to go for 100 inches (2.5 metres). Hooker pledged US$ 45,000 to the project, with Hale promising the telescope would be the largest in the world and bear Hooker's name. Once again, an order for the disc was placed with the Saint-Gobain glassworks in France, the only one with experience in such large glass castings. Problems began almost immediately. Saint-Gobain did not have the capacity to melt the quantity of glass required (four and a half tons) all at once: they would have to fill the mould in three successive pours. A massive piece of cast glass (101 inches in diameter and 13 inches thick) cannot simply be allowed to cool naturally after being poured. If that were to occur, shrinkage of the outer parts of the disc as it cooled while the inside still remained hot would almost certainly cause the disc to fracture and, even if it didn't, would create strains within the disc that would render it incapable of holding the precise figure (curvature) required by the mirror. Instead, the disc must be placed in an annealing oven, where the temperature is reduced slowly over a period of time, allowing the internal stresses to be released. So massive was the 100 inch disc that it took a full year to anneal.

When the disc finally arrived in Pasadena, Hale and Ritchey were dismayed by what they saw, There were sheets of bubbles between the three layers of poured glass, indicating they had not fused. There was evidence the process of annealing had caused the internal structure of the glass to begin to break down. It seemed unlikely a suitable mirror could be made from the disc. After extended negotiations, Saint-Gobain decided to try again, casting a replacement disc at no additional cost. Months later, they reported the second disc had broken during annealing, and it was likely no better disc could be produced. Hale decided to proceed with the original disc. Patiently, he made the case to the Carnegie Institution to fund the telescope and observatory on Mount Wilson. It would not be until November 1917, eleven years after the order was placed for the first disc, that the mirror was completed, installed in the massive new telescope, and ready for astronomers to gaze through the eyepiece for the first time. The telescope was aimed at brilliant Jupiter.

Observers were horrified. Rather than a sharp image, Jupiter was smeared out over multiple overlapping images, as if multiple mirrors had been poorly aimed into the eyepiece. Although the mirror had tested to specification in the optical shop, when placed in the telescope and aimed at the sky, it appeared to be useless for astronomical work. Recalling that the temperature had fallen rapidly from day to night, the observers adjourned until three in the morning in the hope that as the mirror continued to cool down to the nighttime temperature, it would perform better. Indeed, in the early morning hours, the images were superb. The mirror, made of ordinary plate glass, was subject to thermal expansion as its temperature changed. It was later determined that the massive disc took twenty-four hours to cool ten degrees Celsius. Rapid changes in temperature on the mountain could cause the mirror to misbehave until its temperature stabilised. Observers would have to cope with its temperamental nature throughout the decades it served astronomical research.

As the 1920s progressed, driven in large part by work done on the 100 inch Hooker telescope on Mount Wilson, astronomical research became increasingly focused on the “nebulæ”, many of which the great telescope had revealed were “island universes”, equal in size to our own Milky Way and immensely distant. Many were so far away and faint that they appeared as only the barest smudges of light even in long exposures through the 100 inch. Clearly, a larger telescope was in order. As always, Hale was interested in the challenge. As early as 1921, he had requested a preliminary design for a three hundred inch (7.6 metre) instrument. Even based on early sketches, it was clear the magnitude of the project would surpass any scientific instrument previously contemplated: estimates came to around US$ 12 million (US$ 165 million today). This was before the era of “big science”. In the mid 1920s, when Hale produced this estimate, one of the most prestigious scientific institutions in the world, the Cavendish Laboratory at Cambridge, had an annual research budget of less than £ 1000 (around US$ 66,500 today). Sums in the millions and academic science simply didn't fit into the same mind, unless it happened to be that of George Ellery Hale. Using his connections, he approached people involved with foundations endowed by the Rockefeller fortune. Rockefeller and Carnegie were competitors in philanthropy: perhaps a Rockefeller institution might be interested in outdoing the renown Carnegie had obtained by funding the largest telescope in the world. Slowly, and with an informality which seems unimaginable today, Hale negotiated with the Rockefeller foundation, with the brash new university in Pasadena which now called itself Caltech, and with a prickly Carnegie foundation who saw the new telescope as trying to poach its painfully-assembled technical and scientific staff on Mount Wilson. By mid-1928 a deal was in hand: a Rockefeller grant for US$ 6 million (US$ 85 million today) to design and build a 200 inch (5 metre) telescope. Caltech was to raise the funds for an endowment to maintain and operate the instrument once it was completed. Big science had arrived.

In discussions with the Rockefeller foundation, Hale had agreed on a 200 inch aperture, deciding the leap to an instrument three times the size of the largest existing telescope and the budget that would require was too great. Even so, there were tremendous technical challenges to be overcome. The 100 inch demonstrated that plate glass had reached or exceeded its limits. The problems of distortion due to temperature changes only increase with the size of a mirror, and while the 100 inch was difficult to cope with, a 200 inch would be unusable, even if it could be somehow cast and annealed (with the latter process probably taking several years). Two promising alternatives were fused quartz and Pyrex borosilicate glass. Fused quartz has hardly any thermal expansion at all. Pyrex has about three times greater expansion than quartz, but still far less than plate glass.

Hale contracted with General Electric Company to produce a series of mirror blanks from fused quartz. GE's legendary inventor Elihu Thomson, second only in reputation to Thomas Edison, agreed to undertake the project. Troubles began almost immediately. Every attempt to get rid of bubbles in quartz, which was still very viscous even at extreme temperatures, failed. A new process, which involved spraying the surface of cast discs with silica passed through an oxy-hydrogen torch was developed. It required machinery which, in operation, seemed to surpass visions of hellfire. To build up the coating on a 200 inch disc would require enough hydrogen to fill two Graf Zeppelins. And still, not a single suitable smaller disc had been produced from fused quartz.

In October 1929, just a year after the public announcement of the 200 inch telescope project, the U.S. stock market crashed and the economy began to slow into the great depression. Fortunately, the Rockefeller foundation invested very conservatively, and lost little in the market chaos, so the grant for the telescope project remained secure. The deepening depression and the accompanying deflation was a benefit to the effort because raw material and manufactured goods prices fell in terms of the grant's dollars, and industrial companies which might not have been interested in a one-off job like the telescope were hungry for any work that would help them meet their payroll and keep their workforce employed.

In 1931, after three years of failures, expenditures billed at manufacturing cost by GE which had consumed more than one tenth the entire budget of the project, and estimates far beyond that for the final mirror, Hale and the project directors decided to pull the plug on GE and fused quartz. Turning to the alternative of Pyrex, Corning glassworks bid between US$ 150,000 and 300,000 for the main disc and five smaller auxiliary discs. Pyrex was already in production at industrial scale and used to make household goods and laboratory glassware in the millions, so Corning foresaw few problems casting the telescope discs. Scaling things up is never a simple process, however, and Corning encountered problems with failures in the moulds, glass contamination, and even a flood during the annealing process before the big disc was ready for delivery.

Getting it from the factory in New York to the optical shop in California was an epic event and media circus. Schools let out so students could go down to the railroad tracks and watch the “giant eye” on its special train make its way across the country. On April 10, 1936, the disc arrived at the optical shop and work began to turn it into a mirror.

With the disc in hand, work on the telescope structure and observatory could begin in earnest. After an extended period of investigation, Palomar Mountain had been selected as the site for the great telescope. A rustic construction camp was built to begin preliminary work. Meanwhile, Westinghouse began to fabricate components of the telescope mounting, which would include the largest bearing ever manufactured.

But everything depended on the mirror. Without it, there would be no telescope, and things were not going well in the optical shop. As the disc was ground flat preliminary to being shaped into the mirror profile, flaws continued to appear on its surface. None of the earlier smaller discs had contained such defects. Could it be possible that, eight years into the project, the disc would be found defective and everything would have to start over? The analysis concluded that the glass had become contaminated as it was poured, and that the deeper the mirror was ground down the fewer flaws would be discovered. There was nothing to do but hope for the best and begin.

Few jobs demand the patience of the optical craftsman. The great disc was not ready for its first optical test until September 1938. Then began a process of polishing and figuring, with weekly tests of the mirror. In August 1941, the mirror was judged to have the proper focal length and spherical profile. But the mirror needed to be a parabola, not a sphere, so this was just the start of an even more exacting process of deepening the curve. In January 1942, the mirror reached the desired parabola to within one wavelength of light. But it needed to be much better than that. The U.S. was now at war. The uncompleted mirror was packed away “for the duration”. The optical shop turned to war work.

In December 1945, work resumed on the mirror. In October 1947, it was pronounced finished and ready to install in the telescope. Eleven and a half years had elapsed since the grinding machine started to work on the disc. Shipping the mirror from Pasadena to the mountain was another epic journey, this time by highway. Finally, all the pieces were in place. Now the hard part began.

The glass disc was the correct shape, but it wouldn't be a mirror until coated with a thin layer of aluminium. This was a process which had been done many times before with smaller mirrors, but as always size matters, and a host of problems had to be solved before a suitable coating was obtained. Now the mirror could be installed in the telescope and tested further. Problem after problem with the mounting system, suspension, and telescope drive had to be found and fixed. Testing a mirror in its telescope against a star is much more demanding than any optical shop test, and from the start of 1949, an iterative process of testing, tweaking, and re-testing began. A problem with astigmatism in the mirror was fixed by attaching four fisherman's scales from a hardware store to its back (they are still there). In October 1949, the telescope was declared finished and ready for use by astronomers. Twenty-one years had elapsed since the project began. George Ellery Hale died in 1938, less than ten years into the great work. But it was recognised as his monument, and at its dedication was named the “Hale Telescope.”

The inauguration of the Hale Telescope marked the end of the rapid increase in the aperture of observatory telescopes which had characterised the first half of the twentieth century, largely through the efforts of Hale. It would remain the largest telescope in operation until 1975, when the Soviet six metre BTA-6 went into operation. That instrument, however, was essentially an exercise in Cold War one-upmanship, and never achieved its scientific objectives. The Hale would not truly be surpassed before the ten metre Keck I telescope began observations in 1993, 44 years after the Hale. The Hale Telescope remains in active use today, performing observations impossible when it was inaugurated thanks to electronics undreamt of in 1949.

This is an epic recounting of a grand project, the dawn of “big science”, and the construction of instruments which revolutionised how we see our place in the cosmos. There is far more detail than I have recounted even in this long essay, and much insight into how a large, complicated project, undertaken with little grasp of the technical challenges to be overcome, can be achieved through patient toil sustained by belief in the objective.

A PBS documentary, The Journey to Palomar, is based upon this book. It is available on DVD or a variety of streaming services.

In the Kindle edition, footnotes which appear in the text are just asterisks, which are almost impossible to select on touch screen devices without missing and accidentally turning the page. In addition, the index is just a useless list of terms and page numbers which have nothing to do with the Kindle document, which lacks real page numbers. Disastrously, the illustrations which appear in the print edition are omitted: for a project which was extensively documented in photographs, drawings, and motion pictures, this is inexcusable.

Posted at October 13, 2016 00:30