The Net Advance of Physics RETRO:

The Thermometer Freezes

Nikolaus von Himsel's Artificial Cold

2013 October 21

Detail from the frontispiece of Linnæus's Hortus Cliffortianus, 1738.

Thermometers are today so ubiquitous that one can hardly imagine a world in which noöne knew the exact temperature of anything. Such however was the world for most of history. Designing a thermometer from scratch is difficult (and that is true even for people who, unlike the ancients, already possess in their mental arsenals the concept of "temperature" as something which can be quantitatively measured). The creation of reliable thermometers, and a standard scale of temperature, was one of the great achievements of early modern science.

One obvious problem faced by the would-be thermometrist is the choice of a material which changes its properties in some fairly uniform way as it becomes hotter or colder. Water, for instance, will clearly not do; apart from its nonlinearities, it boils and freezes far too easily. Alcohol has a somewhat wider range, but the ideal "working fluid", according to Eighteenth Century thinking, was liquid mercury.

As a metal -- so the argument went -- mercury cannot possibly boil or freeze. It will retain its fluidity no matter how much heat or cold may be poured into it (whether cold was the absence of heat or a separate mysterious substance was still debated).

Indeed, mercury did work very well -- the simple glass tube with quicksilver became the universal standard, the device which not only measured but even defined temperature. Such thermometers were taken by explorers to remote locales and used to document the climate of the world; they were a key technology of the great Enlightenment cosmographical project.

Nevertheless, and quite soon, disturbing rumours began to circulate. Their original source was Siberia, the proverbial home of exotic cold, and noöne took them too seriously. Russia, though, is also very cold, and St. Petersburg was no remote outpost but a bastion of Western science and culture. Imagine, then, the alarm of European thermometrists when Nikolaus von Himsel, a distinguished physician and philosophe of Empress Catherine's modern Russia, sent the following report about some experiments conducted in the dead of winter ...

Plaque in Riga, Latvia, commemorating Nikolaus von Himsel. [Photo by Alma Pater.]

From Philosophical Transactions of the Royal Society of London 51, 670 (1760). Original text of article in bold; my notes in [bracketed light italics].

An Account of artificial Cold produced at Petersburg : By Dr. Himsel. In a Letter to Dr. De Castro, F. R. S. Translated from the French by James Parsons, M. D., F. R. S.

On the 14th of December [1759] we had, at Petersburg, the most excessive cold weather that ever was known, even to 205 degrees of De Lisle's thermometer [-37°C or -34°F].

[The Delisle scale, created in 1732, was specifically designed for use with mercury thermometers. Unlike modern temperature scales, it measured "coldness"; water boils at 0°D and freezes at +150°D. Absolute zero is 559.725°D.]

At that time Professor [Joseph Adam] Braun repeated Fahrenheit's experiments, in order to produce excessive cold by means of spirit of nitre combined with snow.

[The formation of brine by dissolving nitre in snow is an extremely endothermic process, one already known to the alchemists and used by Blasio Villafranca to create the world's first ice-cream in 1550. Daniel Fahrenheit investigated this and similar reactions systematically; in fact, the zero he chose for his scale was the equilibrium temperature spontaneously approached by a mixture of equal parts snow, salt, and ammonium chloride.]

He saw, with surprize, the quicksilver fall considerably in the thermometer, and descend even to 470 degrees [-210°C or 60°K] at last : there the quicksilver remained fixed in the open air for the space of a quarter of an hour, and did not begin to rise till it was carried into a warm room.

He repeated the same experiment, first with the same, and then with another thermometer, with the same success. The immobility of the quicksilver made him conjecture, that it might be frozen, or become a solid body.

But as Mr. Braun had not broken the glasses, he could only at that time form a conjecture. On the 17th he produced, again, cold equal to that of the 14th, but on that day there was no experiment made ; and Mr. Braun communicated his discovery at a meeting of the [Imperial Russian] academy. On the 25th of December in the morning, between nine and ten, De Lisle's thermometer was at the 199th degree of cold [-32°C or -27°F] and Mr. Braun, as well as Professor [Franz Ulrich Theodor] Æpinus, then repeated this experiment. As soon as the former had observed the quicksilver immoveable in the thermometer, he broke the glass ; and he found, to his amazement, the quicksilver frozen, but not intirely ; for in the middle of the glass ball there was a small portion yet remaining fluid. Mr. Æpinus's thermometer fell, with extreme rapidity, almost to the 500th degree [40°K], and in breaking the glass from below, he found the quicksilver contained in it absolutely frozen.

Both the gentlemen found, that the quicksilver, thus rendered solid, bore hammering and extension, like other metals ; but being aftwards exposed to the open air, it recovered its former fluidity in a little time.

Mr. Æpinus went somewhat farther, in order to examine the quicksilver, when it was made solid. He poured quicksilver into a glass tube as thick as one's finger, closed at bottom, but open at top. The quicksilver in this cylinder, which was about one inch and half long, froze in three quarters of a minute ; and he observed, that it became solid, perfectly resembling other metals, except iron : it continually contracted, and its surface, which was at first pretty high, soon sunk very low. This cylinder of frozen quicksilver sunk to the bottom of fluid quicksilver, in the same manner, as is observed of other metals, except iron. We know the contrary happens with regard to water frozen and other fluids, which extend as they become solid, and their ice swims in the fluid matter, of which they were produced.

On the 26th of December in the morning, between nine and ten, the cold became extremely sharp at 211 degrees, and such as exceeded the greatest degree of artificial cold fixed by Fahrenheit ; for 40 degrees below zero, in Fahrenheit's thermometer, is equal to 210 degrees of that of De Lisle.

Mr. Braun repeated this experiment again, exactly with the same success with that of the day before. The counsellor and professor Lomonossow made the same experiment on the same day ; and by means of aqua fortis the cold came to 495 degrees.

["Lomonossow" is of course the universal genius M. V. Lomonosov, a towering figure in Russian cultural history.]

He then poured in spirit of common or sea salt, and the quicksilver fell down in the thermometer to 554 degrees [3.82°K] and in taking the thermometer from the mixture, the quicksilver continued to fall in the open air to the 552d degree. He threw yet into the glass a little more snow, pouring on it some oil of vitriol, and suddenly the quicksilver fell to 1260 degrees.

[That is, to about 500° below absolute zero! But modern physicists will be relieved to find that, a few paragraphs below, Lomonosov himself notes a possible source of error in this measurement.]

Then he broke the ball, and found the mercury changed to a solid body. The quicksilver, which yet remained in the tube, was also become solid, and appeared like a loose silver wire, attached to the ball, which was flexible every way. He gave the ball of quicksilver several blows with a turned ax, and it became flat like a half-ruble, or English half-crown; but receiving thereby some cracks, it dissolved in about 20 minutes. These experiments were made when the air was at about 208 degrees of cold.

Mr. Krase, Mr. Zeicher, and the first apothecary Model, and again Mr. Æpinus, repeated the same experiment with the very same success. It is to be observed, that at the second experiment by Professor Zeicher, on the 31st of December, as the air was then only at the 183d degree of cold, in taking the thermometer from the mixture, in which the quicksilver was at 300 degrees, it fell yet 100 degrees more in the open air.

Casting Solid Mercury Figurines
Boing Boing Video [4 min]

The testimony of so many philosophers, each of whom had respectively made the experiment, will, no doubt, be sufficient to prove the truth of it. But in order to remove all doubt about it, it must be remarked, that distilled quicksilver only was made use of in every experiment ; nay, in some, the quicksilver was revivified from sublimation. There can therefore be no suspicion, that what they used was impure, or mixed with any heterogeneous matter. (This appears to have happened to Mr. De Lisle de la Croyere, when he says, that in Siberia he found the quicksilver congealed in the barometer : and even his papers, which are in the academy, shew that he made a mistake in his remarks; for, according to them, the mercury became solid as soon as it fell to about 195 or 200 degrees : but the mercury, which is pure, does not congeal at that degree ; for otherwise it would not be very extraordinary with us to see it take a solid form, because it is not rare to find the cold at this degree here [in Russia]. We may believe, that the quicksilver used by Mr. De la Croyere was impure, and therefore might sooner become an amalgama than pure mercury.)

Now there are two things we cannot reason upon with any certainty : As to the hardness of the quicksilver congealed, it appeared to have had, in some essays, less hardness than lead, and in others more : also we cannot be very sure of the degree of cold, by which the mercury is consolidated. The greatest part of the experiments agree in this, that the quicksilver becomes solid, when it falls in the thermometer to 500 degrees, more or less. Nevertheless, they do not so sufficiently agree as to deduce any thing certain about it.

Although in the experiment made by Mr. Lomonossow the quicksilver fell to 1260 degrees, this philosopher nevertheless says, that he could not sufficiently observe, in his hurry, whether the ball might not have received some crack, and the quicksilver thereby perhaps might have had liberty to fall the lower, which otherwise would not have happened ; for the same thing happened to Mess. Braun, Zeicher, and Æpinus, that the balls of their thermometers were cracked and broken. By the experiment of Mr. Æpinus, made on the 25th of December, in which the quicksilver fell suddenly in the thermometer, and the cylinder of quicksilver of the thickness of one's finger becoming solid so quickly, it may be observed, that the degree of cold then produced ought to exceed 300 degrees.

[The accepted modern value for the freezing point of mercury is -38.35°C, corresponding to 208.3°D. This suggests that Delisle de la Croyère in Siberia was correct, and the much colder temperatures consistently reported in St. Petersburg were due to experimental error. There is no doubt, however, about the validity of von Himsel's main point: there exist regimes of cold which the mercury thermometer cannot measure.]

Nevertheless, whatever the degree be, we cannot determine how the common thermometer ceases to be of use as soon as the mercury begins to become solid.

[Photo by "Greenhorn 1"]

Here follows an account of the manner, in which these experiments may be made, that other philosophers may be capable of repeating them. It is therefore to be observed, that it is necessary to use fuming spirit of nitre, or of such as is evaporated till the fumes become red ; for the common aqua fortis, which is used, had not the desired effect.

["Aqua fortis" is regular nitric acid; "fuming spirit of nitre" -- the rocket fuel RFNA -- contains additional nitrogen dioxide.]

Mr. Æpinus has found, that this experiment is very easily and speedily made in the following manner. Take spirit of nitre, cooled as much as possible, and with it half fill a wine-glass, throwing in as much snow at the same time, and stirring it till it becomes of the consistence of pap : then you have almost in an instant the necessary degree for the congelation of the quicksilver : not only Mr. Æpinus used this method, but also Mess. Krase, Zeicher, and Model, who followed Mr. Æpinus, and found it the most convenient method.

Now, in reflecting upon the procedure of other philosophers, especially of Mess. Muschenbroek and Reaumur, for producing artificial cold, by the commixtion of snow with aqua fortis, as the former has mentioned in his edition of the experiments of the academy of Florence, tom. i. p. 174, and Mr. Reaumur, in the memoirs of the academy of sciences of Paris for the year 1734, it is astonishing how it happens, that these learned men have not obtained, by a great deal, the degree of cold produced by the gentlemen of the academy of Petersburg ; for their manner of making the experiments does not seem to differ much from that of Mr. Braun, as to what relates to any essential circumstances, nor from the manner mentioned before, so as to hinder them from producing effects nearly equal. Perhaps they may, in time, and by continued researches, be able to find out the circumstances, that prevented their success in the experiments of these great men : it may be, because the spirit of nitre was not endowed with its proper quality.

In fine, I must further observe, that a certain degree of external cold appears absolutely necessary to the experiment. Mr. Æpinus, who made it the 28th of December, in a room where De Lisle's thermometer shewed 122 degrees of cold, cooled the spirit of nitre in liquifying snow to 150 degrees, and the snow, which they used, came to the same degree; in making the mixture, the result was an augmentation of cold to 300 degrees. It must then happen, that they had obtained the surprizing degree necessary to congeal the mercury ; which Mr. Zeicher also at length obtained ; the degree of cold of the air being the 175th degree of De Lisle's thermometer, or the 30th of that of Fahrenheit.

[Photo by Pieter Kuiper]

Of course, solid mercury did not mean the end of thermometers! New working materials were found for the low-temperature regime; some of them (alloys of mercury and thallium and, more recently, of gallium, indium and tin) were even liquid metals, although electrical methods would eventually predominate.

The real importance of this story has little to do with thermometry per se, but with the broader theory of measurement. Before the discovery of frozen mercury, temperature was defined operationally: the temperature is whatever a mercury thermometer reads. If some materials are too cold for quicksilver, this definition is clearly inadequate. One must either find a new standard material with an infinite range, as mercury had been wrongly supposed to have, or one must define temperature in a fundamentally different way. With the advent of thermodynamics, the latter course prevailed.

This problem arises endlessly in science: one modern variant involves the decision to define the metre as the distance travelled by light in a certain number of seconds. This seemed extremely logical to most later-Twentieth Century physicists: certainly better than referring to a metal bar in Paris or to the meridians of an idealised spherical Earth. Hardly any physicists, even today, seem to realise that it makes some rather large assumptions: not only the truth of the (well-established) special theory of relativity, but perhaps more dramatically the exact equivalence of the relativistic causal-speed c which cannot be exceeded with the measured speed of light. What if photons really travel at almost, but not quite, the causal speed? Or what if they did so in the past, or will in the future? When João Magueijo proposed something along those lines in 1998, a great many people rejected it out of hand, because "the speed of light by definition cannot vary" -- and indeed, with the current definition of the metre, it cannot. But mercury freezes, and light, too, may yet surprise people!

Solid Mercury Giant Frozen Gummi Bear.
Video by "TaoFlederMaus" [4 min]