Inventing Temperature Page 4
9. For further details, see Cavendish et al. 1777, esp. 816-818, 853-855.
10. Robert Boyle had already noted this in the seventeenth century, and Daniel Gabriel Fahrenheit knew the quantitative relations well enough to make a barometer that inferred the atmospheric pressure from the boiling point of water. See Barnett 1956, 298.
11. The description of Adams's scale is from Chaldecott 1955, 7 (no. 20). For information about his status and work, see Morton and Wess 1993, 470, and passim.
end p.12
Figure 1.1. George Adams's thermometric scale, showing two boiling points (inventory no. 1927-1745). Science Museum/Science & Society Picture Library.
end p.13
De Luc, who was a key member of the Royal Society committee and perhaps the leading European authority in thermometry in the late eighteenth century.
As Jean-André De Luc (fig. 1.2) is a little-known figure today even among historians of science, a brief account of his life and work is in order.12 In his own day De Luc had a formidable reputation as a geologist, meteorologist, and physicist. He received his early education from his father, François De Luc, a clockmaker, radical politician, and author of pious religious tracts, who was once described by Jean-Jacques Rousseau as "an excellent friend, the most honest and boring of men" (Tunbridge 1971, 15). The younger De Luc maintained equally active interests in science, commerce, politics, and religion. To his credit were some very popular natural-theological explanations of geological findings, strenuous arguments against Lavoisier's new chemistry, and a controversial theory of rain postulating the transmutation of air into water.13 One of the early "scientific mountaineers," De Luc made pioneering excursions into the Alps (with his younger brother Guillaume-Antoine), which stimulated and integrated his scientific interests in natural history, geology, and meteorology. His decisive improvement of the method of measuring the heights of mountains by barometric pressure was a feat that some considered sufficient to qualify him as one of the most important physicists in all of Europe.14 More generally he was famous for his inventions and improvements of meteorological instruments and for the keen observations he made with them. Despite his willingness to theorize, his empiricist leanings were clearly encapsulated in statements such as the following: "[T]he progress made towards perfecting [measuring instruments] are the most effectual steps which have been made towards the knowledge of Nature; for it is they that have given us a disgust to the jargon of systems … spreading fast into metaphysics" (De Luc 1779, 69). In 1772 De Luc's business in Geneva collapsed, at which point he retired from commercial life and devoted himself entirely to scientific work. Soon thereafter he settled in England, where he was welcomed as a Fellow of the Royal Society (initially invited to the Society by Cavendish) and also given the prestigious position of "Reader" to Queen Charlotte. De Luc became an important member of George III's court and based himself in Windsor to his dying day, though he did much traveling and kept up his scientific connections particularly with the Lunar Society of Birmingham and a number of German scholars, especially in Göttingen.
De Luc's first major scientific work, the two-volume Inquiries on the Modifications of the Atmosphere, published in 1772, had been eagerly awaited for the promised discussion of the barometric measurement of heights. When it was finally published after a delay of ten years, it also contained a detailed discourse on the
12. The most convenient brief source on De Luc's life and work is the entry in the Dictionary of National Biography, 5:778-779. For more detail, see De Montet 1877-78, 2:79-82, and Tunbridge 1971. The entry in the Dictionary of Scientific Biography, 4:27-29, is also informative, though distracted by the contributor's own amazement at De Luc's seemingly unjustified renown. Similarly, W. E. K. Middleton's works contain a great deal of information about De Luc, but suffer from a strong bias against him.
13. On the controversy surrounding De Luc's theory of rain, which was also the cornerstone of his objections to Lavoisier's new chemistry, see Middleton 1964a and Middleton 1965.
14. For this appraisal, see Journal des Sçavans, 1773, 478.
end p.14
Figure 1.2. Jean-André De Luc. Geneva, Bibliothèque publique et universitaire, Collections iconographiques.
end p.15
construction and employment of thermometers, with an explanation that De Luc had originally become interested in thermometers because of the necessity to correct barometer readings for variations in temperature.15 I will have occasion to discuss other aspects of De Luc's work in thermometry in chapter 2, but for now let us return to the subject of possible variations in the boiling temperature according to the "degree of boiling." Initially De Luc asserted: When water begins to boil, it does not yet have the highest degree of heat it can attain. For that, the entire mass of the water needs to be in movement; that is to say, that the boiling should start at the bottom of the vessel, and spread all over the surface of the water, with the greatest impetuosity possible. From the commencement of ebullition to its most intense phase, the water experiences an increase in heat of more than a degree. (De Luc 1772, 1:351-352, §439)
In further experiments, De Luc showed that there was an interval of 76 to 80 degrees on his thermometer (95-100°C, or 203-212°F) corresponding to the spectrum of ebullition ranging from "hissing" to full boil, which is quite consistent with the range of 204-212°F indicated in Adams's thermometer discussed earlier. The weakest degree of genuine boiling started at 78.75° on De Luc's thermometer, in which the full-boiling point was set at 80°, so there was a range of 1.25° (over 1.5°C) from the commencement of boiling to the highest boiling temperature.16
The Royal Society committee investigated this issue carefully, which is not surprising given that its two leading members, Cavendish and De Luc, had been concerned by it previously. The committee's findings were somewhat reassuring for the stability of the boiling point: For the most part there was very little difference whether the water boiled fast or very gently; and what difference there was, was not always the same way, as the thermometer sometimes stood higher when the water boiled fast, and sometimes lower. The difference, however, seldom amounted to more than 1/10th of a degree. (Cavendish et al. 1777, 819-820)
Still, some doubts remained. The trials were made in metallic pots, and it seemed to matter whether the pots were heated only from the bottom or from the sides as well: In some trials which we made with the short thermometers in the short pot, with near four inches of the side of the vessel exposed to the fire, they constantly stood lower when the water boiled fast than when slow, and the height was in general greater than when only the bottom of the pot was exposed to the fire. (820)
Not only was that result in disagreement with the other trials made by the committee but also it was the direct opposite of the observations by Adams and De Luc,
15. See De Luc 1772, 1:219-221, §408.
16. See De Luc 1772, 2:358, §983. De Luc's own thermometer employed what came to be known as the "Réaumur" scale, which had 80 points between the freezing and the boiling points. R. A. F. Réaumur had used an 80-point scale, but his original design was considerably modified by De Luc.
end p.16
according to which water boiling vigorously had a higher temperature than water boiling gently.
There were other factors to worry about as well. One was the depth of the boiling water: "[I]f the ball be immersed deep in the water, it will be surrounded by water which will be compressed by more than the weight of the atmosphere, and on that account will be rather hotter than it ought to be" (817-818). Experiments did vindicate this worry, revealing a variation of about 0.06° per inch in the depth of the water above the ball of the thermometer. However, the committee was reluctant to advance that observation as a general rule. For one thing, though this effect clearly seemed to be caused by the changes of pressure, it was only half as large as the effect caused by changes in the atmospheric pressure. Even more baffling was the fact that "the boiling point was in some measure increased by having a great
depth of water below the ball … [T]his last effect, however, did not always take place" (821-822; emphasis added). Although the committee made fairly definite recommendations on how to fix the boiling point in the end, its report also revealed a lingering sense of uncertainty: Yet there was a very sensible difference between the trials made on different days, even when reduced to the same height of the barometer, though the observations were always made either with rain or distilled water. … We do not at all know what this difference could be owing to. … (826-827)
Superheating and the Mirage of True Ebullition
The work of the Royal Society committee on the boiling point is a lively testimony to the shakiness of the cutting-edge knowledge of the phenomenon of boiling in the late eighteenth century. No one was more clearly aware of the difficulties than De Luc, who had started worrying about them well before the Royal Society commission. Just as his book was going to the press in 1772, De Luc added a fifteen-chapter supplement to his discussion of thermometers, entitled "inquiries on the variations of the heat of boiling water." The logical starting point of this research was to give a precise definition of boiling, before disputing whether its temperature was fixed. What, then, is boiling? De Luc (1772, 2:369, §1008) conceived "true ebullition" ("la vraie ébullition") as the phenomenon in which the "first layer" of water in contact with the heat source became saturated with the maximum possible amount of heat ("fire" in his terminology), thereby turning into vapor and rising up through the water in the form of bubbles. He wanted to determine the temperature acquired by this first layer. That was a tall order experimentally, since the first layer was bound to be so thin that no thermometer could be immersed in it. Initial experiments revealed that there must indeed be a substantial difference between the temperature of the first layer and the rest of the water under normal conditions. For example, when De Luc heated water in a metallic vessel put into an oil bath, the thermometer in the middle of the water reached 100°C only when the oil temperature was 150°C or above. One could only surmise that the first layer of water must have been brought to a temperature somewhere between 100°C and
end p.17
150°C. De Luc's best estimate, from an experiment in which small drops of water introduced into hot oil exploded into vapor when the oil was hot enough, was that the first layer of water had to be at about 112°C, for true ebullition to occur.17
Was it really the case that water could be heated to 112°C before boiling? Perhaps incredulous about his own results, De Luc devised a different experiment (1772, 2:362-364, §§994-995). Thinking that the small drops of water suspended in oil may have been too much of an unusual circumstance, in the new experiment he sought to bring all of a sizeable body of water up to the temperature of the first layer. To curtail heat loss at the open surface of the water, he put the water in a glass flask with a long narrow neck (only about 1 cm wide) and heated it slowly in an oil bath as before. The water boiled in an unusual way, by producing very large occasional bubbles of vapor, sometimes explosive enough to throw off some of the liquid water out of the flask. While this strange boiling was going on, the temperature of the water fluctuated between 100°C and over 103°C. After some time, the water filled only part of the flask and settled into a steadier boil, at the temperature of 101.9°C. De Luc had observed what later came to be called "superheating," namely the heating of a liquid beyond its normal boiling point.18 It now seemed certain to De Luc that the temperature necessary for true ebullition was higher than the normally recognized boiling point of water. But how much higher?
There was one major problem in answering that question. The presence of dissolved air in water induced an ebullition-like phenomenon before the temperature of true ebullition was reached. De Luc knew that ordinarily water contained a good deal of dissolved air, some of which was forced out by heating and formed small bubbles (often seen sticking to the inner surface of vessels), before the boiling point was reached. He was also well aware that evaporation from the surface of water happened at a good rate at temperatures well below boiling. Putting the two points together, De Luc concluded that significant evaporation must happen at the inner surfaces of the small air bubbles at temperatures much lower than that of true ebullition. Then the air bubbles would swell up with vapor, rise, and escape, releasing a mixture of air and water vapor. Does that count as boiling? It surely has the appearance of boiling, but it is not true ebullition as De Luc defined it. He identified this action of dissolved air as "the greatest obstacle" that he had to overcome in his research: "that is, the production of internal vapors, which is occasioned by this emergence of air, before there is true ebullition."19
De Luc was determined to study true ebullition, and that meant obtaining water that was completely purged of dissolved air. He tried everything. Luckily,
17. For further details on these experiments, see De Luc 1772, 2:356-362, §§980-993.
18. The term "superheating" was first used by John Aitken in the 1870s, as far as I can ascertain; see Aitken 1878, 282. The French term surchauffer was in use quite a bit earlier.
19. For the discussion of the role of air in boiling, see De Luc 1772, 2:364-368, §§996-1005; the quoted passage is from p. 364.
end p.18
sustained boiling actually tended to get much of the air out of the water.20 And then he filled a glass tube with hot boiled water and sealed the tube; upon cooling, the contraction of the water created a vacuum within the sealed tube, and further air escaped into that vacuum.21 This process could be repeated as often as desired. De Luc also found that shaking the tube (in the manner of rinsing a bottle, as he put it) facilitated the release of air; this is a familiar fact known to anyone who has made the mistake of shaking a can of carbonated beverage. After these operations, De Luc obtained water that entered a steady boil only in an oil bath as hot as 140°C.22 But as before, he could not be sure that the water had really taken the temperature of the oil bath, though this time the water was in a thin tube. Sticking a thermometer into the water in order to verify its temperature had the maddening side effect of introducing some fresh air into the carefully purified water. There was no alternative except to go through the purging process with the thermometer already enclosed in the water, which made the already delicate purging operation incredibly frustrating and painful. He reported: This operation lasted four weeks, during which I hardly ever put down my flask, except to sleep, to do business in town, and to do things that required both hands. I ate, I read, I wrote, I saw my friends, I took my walks, all the while shaking my water. … (De Luc 1772, 2:387, §§1046-1049)
Four mad weeks of shaking had its rewards. The precious airless water he obtained could stand the heat of 97.5°C even in a vacuum, and under normal atmospheric pressure it reached 112.2°C before boiling off explosively (2:396-397, §§1071-1072). The superheating of pure water was now confirmed beyond any reasonable doubt, and the temperature reached in this experiment was very much in agreement with De Luc's initial estimate of the temperature reached by the "first layer" of water in ebullition.
Superheating was an experimental triumph for De Luc. However, it placed him into a theoretical dilemma, if not outright confusion. Ordinary water was full of air
20. Compare this observation with the much later account by John Aitken (1923, 10), whose work will be discussed in some detail in "A Dusty Epilogue": "After the water has been boiling some time there is less and less gas in it. A higher temperature is therefore necessary before the gas will separate from the water. Accordingly, we find that the water rises in temperature after boiling some time. The boiling-point depends in fact not on the temperature at which steam is formed, but on the temperature at which a free surface is formed."
21. De Luc had used this technique earlier in preparing alcohol for use in thermometers. Alcohol boils at a lower temperature than water (the exact temperature depending on its concentration), so there was an obvious problem in graduating alcohol thermometers at the boiling point of water. De Luc (1772, 1:314-318, §4
23) found that purging the alcohol of dissolved air made it capable of withstanding the temperature of boiling water. If W. E. Knowles Middleton (1966, 126) had read De Luc's discussion of superheating, he would have thought twice about issuing the following harsh judgment: "If there was any more in this than self-deception, Deluc must have removed nearly all the alcohol by this process. Nevertheless, this idea gained currency on the authority of Deluc."
22. For a description of the purging process, see De Luc 1772, 2:372-380, §§1016-1031. The boiling experiment made with the airless water (the "sixth experiment") is described in 2:382-384, §§1037-1041.
end p.19
and not capable of attaining true ebullition, but his pure airless water was not capable of normal boiling at all, only explosive puffing with an unsteady temperature. To complicate matters further, the latter type of boiling also happened in a narrow-necked flask even when the water had not been purged of air. De Luc had started his inquiry on boiling by wanting to know the temperature of true boiling; by the time he was done, he no longer knew what true boiling was. At least he deserves credit for realizing that boiling was not a simple, homogeneous phenomenon. The following is the phenomenology of what can happen to water near its boiling point, which I have gathered from various parts of De Luc's 1772 treatise. It is not a very neat classification, despite my best efforts to impose some order.