Tuesday, September 20, 2016

Flowing Forward: A Technological Appendix for Postblogging August 1946, I.

(Edited 24 September 2016).

Robert Solo was a Professor of Economics at Michigan State University, East Lansing, for many years. His solid research and long publishing record has made him a famous, widely loved figure, in spite of showing some signs of being an arrogant young man, and and a wise and self-reflective old one.

Just kidding!

Apart from the odd little autobiography with which he decides to begin The Philosophy of Science and Economics (1991), this blurb, cut from the top full-text entry on the American wartime synthetic rubber programme in a Google Books search* is the best the Internet will do for me. An edition of Alexis de Tocqueville's Old Regime and Revolution is a higher hit on Google than the requested obituary when I search for "professor robert solo east michigan state obituary." Try searching for his titles, and Google quickly gives up on "Professor Robert Solo" and moves on to "Professor Robert Solow."

EDIT: After reading a paper on competition in the digital economy that laid out a strained argument about how it might pay a monopolistic search engine provider, for example, to deliberately degrade search results, I fired up Bing on my Surface. (It turns out that this was the first time I'd used Edge on it, and the "welcome" page was a bit  needy. Just saying, Redmond. . . ) Anyway, the first result for "Robert Solo Michigan State" is his 2011 obituary. (Which information admittedly now, or perhaps always, shows up in the first Google search item, although much less prominently.)

In the end, I guess that we can blame Professor Solo for choosng not to be named Higginbotham or writing a book about Zamboanga, as opposed to frontiers of technological progress, or accepting a tenured position at a univesity with a music school that holds solo performances. (Or he could be still alive in his late 90s.) It is certainly not his strong connection with unspeakably obscene ideas of the most perverse regions of the netherworld, such as [TRIGGER ALERT STOP WITH THE PRETTY PICTURE AND DON'T CLICK ON THE JUMP IF YOU HAVE PTSD AND STUFF LIKE THAT]
Image Search  result: "Robert Solo synthetic  rubber" (NSFW source)

. . . That research and development spending does not drive productivity increases. Ph'nglui mglw'nafh cthulhu r'lyeh wgah'nagl fhtagn. fnord.

Astonishingly enough, it turns out that everyone was talking about this back in the early 60s, including Professor Solo. That was before Professor Robert Solo(w) arrived on the scene to save us from our doubts. The reason that I've spent so much time on him is that his example of a failure of a research-and-development-led production effort is the US wartime synthetic rubber programme, which used butadiene, which, as we've seen, was an early candidate for production by continuous flow methods. He goes on, of course, to add that if you want an example of a successful WWII research-and-development-led production effort, you want to look at the Manhattan Project, which is where we came in. So there you go: research and development, butadiene, synthetic rubber, process control, uranium enrichment, Manhattan Project. It's all a rich tapestry. 

As Professor Solo(no w) points out, the United States entered the war emergency period with serious doubts about whether enough natural rubber would be available to meet demand at wartime production levels. Synthetic rubber being a big thing in the increasingly autarkic 30s, Americans were both aware of the industry's potential capability, and in a position to exploit foreign technology. I. G. Farben had developed a method of stripping butadiene from ethylenes made by steam cracking hetergeneous aliphatic hydrocarbons (ie Silesian soft brown coal), although the same catalytic cracking method could be used on normal butane, present in petroleum feedstocks that would otherwise have gone to making gasoline. It was this technology which Standard Oil enthusiastically presented to a family of New Deal "alphabet soup" agencies in 1940/1, and which was adopted as the basis of the American synthetic rubber industry. (Along with stripping from acetylene, the method already used by duPont to produce its novel "neoprene."

Neoprene is very sexy, but makes lousy truck tyres. By Maegan Tintari - http://www.flickr.com/photos/lovemaegan/8870184619/, CC BY 2.0, https://commons.wikimedia.org/w/index.php?curid=28082741

Only the cracking methods from petroleum feedstocks didn't work at any reasonable cost, and assorted Iowa farm interests (or some such) began pushing for the much less high-tech and well-proven approach of making butadiene from grain alcohol, which did work. In spite of getting under way much later than the catalytic cracking method, the synthetic-rubber-from-butadiene-from-alcohol-from-grain approach worked just fine, although there was that thing with the postwar global famine, which Professor Solo doesn't mention. In fact, it worked more than fine, vastly out-producing its butadiene-production rating, and producing more butadiene in 1945 than the original synthetic industry cracking target. 

If you're wondering about the rubber plants in Akron, Ohio that produced the final product from butadiene and the role of new technology there, there wasn't much of one. Once a butadiene solution has been obtained, it can be treated in a way not too dissimilar from natural latex, so the technology was no great novelty. It's hard to learn this from the available literature on the industry, but there's a great 1970s-era report on wastewater treatment from the EPA that has somehow, unaccountably not been reduced to snippet view over at Google Books. Maybe it has something to do with the Cuyahoga River catching fire?

By the way, since we've been talking about flow chemistry, here's the flow diagram for rubber production from butadiene solution. It's not a terribly relevant flow diagramme, since I'm not talking about the final stage, but it is an illustration of the sort of thing I'm talking about. I think. I'm no chemical engineer, admittedly. 

What we are talking about is. . . hard to say. The Fortune article that provoked this is ostensibly about the Taylor Equipment Company's contribution to the gaseous diffusion plant located in Building K25 of the Oak Ridge National Laboratory

If you follow the first two links, you'll find yourself wandering into one of Wikipedia's remaining weak points. Essentially, there are two articles about the gaseous diffusion method of refining uranium-235. "Gaseous Diffusion" is significantly shorter and less interesting than the "K25," but they're both worth reading in that they give slightly different accounts, especially with regards to the role of the British Tube Alloy project in seeding the American work. I'll draw a veil over this, as while it is interesting from a "Correlli Barnett is a moron" angle, it otherwise looks like chauvinistic special-pleading that wears especially poorly on a Canadian. The exception here is the role of Sir Wallace Alan Akers, introduced as "one of fifteen members of the British gaseous diffusion project," the existence of which was news to me when I got that deep in the long and technical "K25" article. It turns out that he was actually the director of the Tube Alloys Project; but he was certainly well-qualified to give advice on matters of chemical engineering, having followed a career path at Brunner Mond/ICI leading from an early research laboratory; to petroleum refining; to ammonia production via the Haber process; to petroleum production by coal hydrogenation. 

So, as from the decision to proceed with gaseous diffusion at Oak Ridge in 1943, participants could draw upon a shared set of experiences with many similarities to those behind the would-be petroleum-crackers over at the synthetic rubber effort. Making rubber by high temperature catalytic cracking of hydrocarbon feedstocks of various kinds goes back, in some sense, to cokemaking in the British Industrial Revolution and the rise of byproducts industries that used the volatile wastes cooked off the coke ovens. Even making ammonia from atmospheric nitrogen looks a lot like this, since coal is both burnt to provide the necessary heat, and used as the source of the hydrogen. So here we have the commanding heights of the chemical engineering economy, succeeding with the completely unfamiliar uranium hexaflouride gas, and failing with the much more familiar petrochemical feedstock. 

It's a curious thing, and not surprising that Professor Solo calls our attention to it. On the other hand, he is also emphatic in identifying Volgograd as the peak of the synthetic rubber crisis. For, had the Russians been defeated there, the Allies would have had to attempt a cross-Channel invasion in 1943, and no American synthetic rubber would have been available. A potential crisis was averted, and the synthetic rubber programme can be seen as a success. World War II set a very different schedule for the atomic bomb, lingering on just long enough for the semi-experimental use of two devices. 

So there's that. Now recall that while the article is about Taylor Equipment's contribution to K25, the examples are drawn from butadiene production from alcohol in the nation's distilleries. It is a little perplexing that the main example is a failure in the complex interractions of the instruments. It's an illustrative example, in that the consequences of a faulty reading entered into one instrument controlling inputs into one retort eventually caused a relief valve several retorts over to open, leading to the undetected release of alcohol into the sewer system. But it's still a failure.

On the other hand, this is still the second time we've heard about the distillery industry. Taylor Equipment is still around today, as Taylor Precision Products, standing somewhere between Minneapolis Honeywell, one of the "seven dwarves" that competed with IBM for computing business in the 1960s and also rans like the once-Foxboro Measurement  (taken over by Schneider, interestingly enough) and  Brown and Sharpe, if that's the "Brown" to which Fortune is referring. The difference, if there is one, is that Taylor started with industrial thermometers, and basically stayed with industrial thermometers. There's actually a lot more space in this industrial sector than you'd think. 

Combined temperature, barometer and --some other gauge. It's for airports rather than distilleries, so it detracts from my point, but it is a vintage '50s bit of Taylor Instruments work. 
And who needed industrial thermometers in the 1850s, back when Taylor was starting out? Distilleries, that's who! Distilleries have an interesting history in the United States. Premium brand Jack Daniels tells us that their "No.7" brand gets its name from the Jack Daniels' distillery being the seventh on an excise list in one county, and makes heavy use of the fact that it is located in a "dry" county in its advertising. The Appalachian states had a great many distilleries in the Nineteenth Century, weren't very happy to have them, and did their best to legislate them out of existence. Notoriously, they succeeded in an out-of-scale way with the Eighteenth Amendment, which made the production and sale of alcoholic beverages illegal in the United States, as from 30 June 1919. This colossal experiment in social engineering was recognised as a mistake, and was abolished with the Twenty-First Amendment, carried through in a hurricane campaign in the spring of 1933, immediately after the inauguration of President Roosevelt. (Not to tire out the theme of Herbert Hoover being America's most incompetent modern President, but he managed to get on the wrong side of this issue, too.)

At this point, distillery firms had two choices: bring the old facility back into production, the choice taken by Jack Daniels; or build a new one, the choice taken, for example, by Jim Beam. From the sounds of things, more firms took the second approach than the first.

I would be willing to bet that paper was just whizzing through the Telex at Taylor Instruments in the summer of 1933. Look: Taylor is not promising a revolutionary new product. Nice whiskey (allowing for taste) is not a thing that had to be invented. It was offering instruments to new distilleries for exactly one reason.

As Fortune puts it its discussion of K-25, it's to reduce the number of operators needed per foot of process line. It's automating to reduce labour costs. Stop me if you've heard this one, etc, etc. You know, reducing the number of human brains required replacing them with automatic relays. It can't get rid of all the humans, or the alcohol will just flow out into the river, which will catch fire. 

You can, however, can take these little electronic boxes and use them to replace some human operators. This very summer, 28 invitees are attending a summer-long seminar on the electronic computer at the Moore School of Engineering of the University of Pennsylvania. They will "invent" the computer. Also, if they have time, they can pick up the August number of Fortune and be told about how a company that makes thermometers for distilleries is inventing a littlel electronic box that can replace the human element in everything from chemical process control to anti-aircraft directors. 

(Also, presumably, in autopilots, so that atomic-armed drone bombers can eradicate the global scoure of Communism, but that's an aspect of this month's postblogging that probably doesn't  need to be covered at the same length. "Can recover from as much as a 50-degree bank." They're even automated in the sense that there's a human operator in the next plane over, waiting to take over control. Leaving aside the feedback issues that lead to venting all that alcohol into the river, a fine idea! Perhaps it can be applied to taxis?)

Is there a conclusion? Yes: that advances in computing science have been driven by the demand side, and that "advances" in computing science come to us from grifters. Probably mostly cognitively biased grifters, but grifters. On the other hand, setting the Cuyahoga River on fire seems to have mainly led to some fine music, so what's the problem with giving the utopians their head, anyway? I mean, what's the worst that can happen?

*Remind me again why there's no full text access to GPO-printed wartime reports to Congress?

No comments:

Post a Comment