The lore of mercury, especially its uses in science and engineering


  1. Properties of Mercury
  2. Chemistry of Mercury
  3. Mercury as a Poison
  4. Production of Mercury
  5. Amalgams
  6. The Barometer
  7. The Thermometer
  8. The Diffusion Pump
  9. The Mercury Battery
  10. Miscellaneous Uses of Mercury
  11. Experiments with Mercury
  12. References

Properties of Mercury

The symbol in the title is the alchemical symbol for mercury, the sign of Mercury (Hermes), fleet messenger of the gods. Mercury is the only metal for which the alchemical planetary name became the ordinary name. To a scientist, it seems more scientific than "quicksilver," which is sometimes still used in the trade. In classical times, mercury found native was distinguished in name from mercury produced by condensing the vapour when cinnabar was heated. The former was "argentum vivum" and the latter "hydrargyrum." The knowledge of cinnabar, HgS, far antedated the knowledge of the metal. It was a valuable red pigment, known as vermilion when artificially produced. Cinnabar is a rather recent name for this substance; classically, it was called "minium," which is now red lead. Cinnabar was then the common name of Dragon's Blood, a red gum exuded from certain palm fruits. Mercury was not considered a metal in ancient times. Arab alchemists considered all metals to be composed of mercury, sulphur and salt. Mercury has been of outstanding service to science; its use in the barometer, thermometer and diffusion pump are discussed below. It is one of the most interesting of elements.

Mercury is the only pure metal liquid at room temperature, as odd in its way as water. The eutectic alloy of 22.8% Na and 77.2% K melts at -12.3°C, so it also is a liquid at room temperature. The name hydrargyrum, "water silver," was given by Pliny from Greek roots for the common name, and is the source of its chemical symbol, Hg. In German, it is called Quecksilber, from its usual ancient name, and in French it is mercure, from which the English "mercury" is derived. This, no doubt, comes from the fancies of late medieval alchemy, where it was represented by the symbol for the god Mercury, seen in the heading of this article. Its atomic number is 80, and its atomic weight is 200.59, an average of isotopes of each mass number from 196 to 204, except for 203. The most abundant isotope (30%) has mass number 202. Mercury freezes at -38.9°C (234K) and boils at 356.58°C (630K). Professor Braune of St. Petersburg succeeded in freezing mercury in the winter of 1759-1760, convincing any doubters that it really was a metal. Its latent heat of fusion is 2.7 cal/g, and its latent heat of vaporization 65 cal/g. Its first ionization potential is 10.39V, and its second is 18.65V. Its specific heat is 0.03325 cal/g-K, and its heat conductivity is 0.0782 W/cm-K at 0°C, 0.0830 at 20°C and 0.0947 at 100°C. The electrical resistivity is 98.4 μΩ-cm. Its bulk modulus is 2.67 x 1010 Pa, or 2.65 x 105 atm. The thermal expansion of mercury is given by V(t) = V(1 + 1.82 x 10-4t + 7.8 x 10-9t2), where t is in °C, and V is the volume at 0°C. Its viscosity at 10°C is 0.5123 poise, at 93.3°C .4022 poise, at 149°C .3543 poise, at 204°C .3208 poise, and at 316°C .2777 poise. Its crystalline structure as a solid is rhombohedral, and the electrical resistivity is strongly anisotropic. In the liquid, it has a close-packed structure. The molecular diameter is about 0.314 nm. At 20°C, its density is 13.54562 g/cm3 and its vapor pressure is 1.2 μmHg. At 100°C, the density is 13.3522 and the vapor pressure 0.2729 mmHg. Mercury is denser than lead, and is the densest liquid known. Its surface tension in air at 20°C is 435.5 dyne/cm. For comparison, the surface tension of water under the same conditions is 73.05 dyne/cm. These values are for clean surfaces.

The reason why mercury is a liquid, unlike other metals, is not usually broached in the literature. Mott-Jones indicates that this might be the result of an unusually high Fermi level caused by a large electron density, since two electrons are available for each mercury atom. This would cause a low binding energy, and in fact, the binding energy is only 18.5 kcal/mol, which can be compared to 76 for Cu, 83 for Au, 32.5 for Zn and 28 for Cd. The low binding energy would permit thermal agitation to liquefy the crystal, in view of the large entropy increase that would occur, with a corresponding decrease of free energy.

The metallic shine of mercury is due to the presence of free electrons with a high plasma frequency, so that they reflect electromagnetic waves of lower frequency, aided by the perfect smoothness of the liquid surface.

Chemistry of Mercury

Mercury is associated with zinc (Zn) and cadmium (Cd), its lighter associates in the same column of the periodic table, which it resembles. It was first recognized as an element by Lavoisier. Previously, it was thought to be the principle in metals that made them white and shiny. It has two valence electrons in s-orbits, and displays valence +1 and +2. Zn and Cd show only +2. Its electrochemical potential is -0.85V, below hydrogen, so it will not displace hydrogen and acts like a noble metal, not corroding in air. Like Zn and Cd, its compounds are poisonous. Mercuric chloride, HgCl2, also called corrosive sublimate and mercury bichoride, is soluble and very poisonous. It is used as an antiseptic in an 0.1% solution. It is a covalent compound, Cl-Hg-Cl, instead of the salt that one would expect, and is little ionized in solution. Mercury is not toxic to plants, so has been used in insecticides. However, this use is to be deprecated because of its general toxicity. Mercurochrome is another mercury-based antiseptic, much less dangerous than corrosive sublimate, but not as powerful. Metaphen is the best of the mercury-based antiseptics. Mercurous chloride, Hg2Cl2, or calomel, is white and insoluble. Light oxidizes it to mercuric chloride and mercury. When treated with ammonia, mercury aminochloride, a white precipitate, is produced, as well as free mercury, which turns the precipitate black. Calomel means "beautiful black," and this reaction, which is used in qualitative chemical analysis, is the reason. Its structure is Cl-Hg-Hg-Cl, where the two Hg atoms are covalently bound. The mercurous ion Hg-Hg++ is stable in solution.

Mercury fulminate, Hg(ONC)2 is made by dissolving mercury in nitric acid and adding ethyl alcohol. It was discovered by Luke Howard in 1799, but was little used until applied to making copper percussion caps for firearms beginning in 1816 and especially until Alfred Nobel showed in 1865 that it could be used for detonating dynamite and other high explosives. It is rather dangerous to handle, and must be kept wet until used. It ignites at a low temperature, about 170°C. It is sensitive to friction and impact, producing a puff of flame that will ingnite black powder, as well as a detonation that will initiate high explosives. Black powder usually just burns if ignited in the open, but at higher temperatures will explode. (Confinement is not the reason it detonates, but the additional heat.) High explosives cannot be detonated by a simple fuse, as black powder can, but require an explosive detonator. A detonator, or blasting cap, is a copper tube filled with a mixture including the fulminate. Different versions can be detonated by a powder fuse or by electrically heating a fine wire. Lead and silver azide are also very sensitive and good detonators, but mercury fulminate is the best overall. Justus Liebig studied the fulminates and clarified their structure in 1822. It is the N≡C bond that is unstable, combined with the nearby oxygen. Mercuric azide, HgN6, which gives only nitrogen and mercury on explosion, is also a good initiator. Mercury fulminate is very heavy (d = 4.43 g/cc), has a sweet, metallic taste, and is poisonous. Fulminates were the second most important use of mercury after drugs and chemicals, but mercury fulminate is now used much less than formerly, replaced by lead styphnate.

The third most important use is for vermilion, mercuric sulphide, HgS, a red pigment. When HgS is precipitated in chemical reactions, it is a black powder, not red as in vermilion and cinnabar. This is simply an effect due to the very fine division. Next comes the oxide, HgO, used in anti-fouling paints for ships (since it is poisonous to barnacles and such). Mercuric cyanide, Hg(CN)2 gives cyanogen, C2N2, when heated. Mercuric thiocyanate, Hg(SCN)2 is used in Pharoah's Serpent eggs, which form huge snakes of light ash when burned. The alchemist Paracelsus (1493-1541) introduced the mercury treatment for syphilis, just in time for the sailors returned from the Americas who had acquired Montezuma's Revenge.

Mercuric oxide was the red powder heated by Joseph Priestley (1733-1804) with a burning-glass in his discovery of oxygen, or "dephlogisticated air," in 1774. The reaction is 2HgO → 2Hg + O2↑. The oxide was produced by more gentle heating of metallic mercury in air, and called mercurius calcinatus per se. Priestley introduced the practice of collecting gases over mercury or water. It was not until after Lavoisier's revolution in chemistry that mercury and oxygen were recognized as elements, whose combination made the red powder. Lavoisier interpreted Priestley's observations on the new understanding, naming oxygen and clarifying its role in combustion. Thus, mercury played an important role in the development of chemistry.

Mercury as a Poison

When considering the hazards of mercury, soluble mercury compounds and metallic mercury are very different. Mercury can enter the body through the lungs, through the skin, and via the digestive system. The absorption of mercury vapor by the lungs is an efficient process, but it is difficult to acquire dangerous amounts of metallic mercury by the other routes. Ingestion of soluble mercury compounds produces acute mercury poisoning, but is very easy to avoid and rare. Usually, mistaking mercuric chloride pills for something harmless is the cause (see below). The kidneys contain a protein, metallothionein, that binds mercury tightly until it is excreted. Mercury in the blood has a half-life of only three days, but tissue mercury has a half-life of perhaps 90 days. Small amounts of mercury, therefore, are efficiently excreted and cause no harm. This information is from the Handbook of Laboratory Safety (2000). Some earlier sources claim that mercury is a cumulative poison, but it apparently is not. The kidneys, however, can be overloaded and symptoms of mercury poisoning will then result. The mercuric ion Hg++ attacks the kidneys and can result in fatal kidney failure. Acute mercury poisoning is, however, a small risk and easily avoided. Mercury passes through the placenta in acute mercury poisioning, and so is a hazard to the fetus. Whether there is any danger in cases of the usual concentration of environmental mercury (as in fish) may be doubted.

One indication of acute mercury poisoning is the secretion of excess saliva, and the condition is called salivation. This often occurred among workers in mercury refining, and also in felt-making, where the soluble mercury compound mercuric nitrate, Hg(NO3)2·2H2O is (was) used. This was the reason for the "mad hatter." The Environmental Protection Agency named mercury a hazardous air pollutant in 1971, which it would be if there were any mercury in the air.

Mercury metal and insoluble compounds are little hazard, and can be handled occasionally with safety. Chronic exposure is a different matter. Mercury vapor is absorbed in the lungs, and chronic exposure even at low levels should be avoided. The toxic limit (PEL)is only 0.05 mg/m3, which is very low indeed, exceeded by the vapor pressure at room temperature. However, the accumulation of the equilibrium vapor pressure is extremely unlikely in most cases, where the mercury is present in visible droplets. Although much worry accompanines spills of liquid mercury, the hazard is probably negligible. T. G. Winter (see References) has recently demonstrated that mercury vapor due to small spills of metallic mercury is not hazardous. Notably, mercury from minor spills can be found in the cracks between vinyl tile flooring. Mercury should not be heated in the open; this can easily produce dangerous amounts of vapor.

A vacuum cleaner should not be used for mercury cleanup. Unless specially designed for this purpose, it only atomizes the mercury and makes it more hazardous than before the cleanup. Substances that amalgamate with the mercury can lock it up so it does no damage, and often these amalgam particles can be swept up easily.

However, the modern panic over environmental hazards coupled with faulty estimation of risks have led to something of a mindless crusade that has eliminated mercury in many common applications, such as in clinical and laboratory thermometers and in antiseptics (mercurochrome and merthiolate). At one time, the medical sciences erred in the other direction, administering mercurous chloride and liquid mercury as medicines. It was estimated that there were two broken clinical thermometers per bed per year, which would seem to indicate considerable clumsiness in nurses. Careless disposal of mercury batteries was said to represent 68% of mercury pollution.

Mercury itself, and most of its compounds, are very insoluble and so are not hazardous in themselves. It was discovered, however, that in anaerobic sediments where there was industrial mercury waste, slightly soluble dimethyl mercury, (CH3)2Hg, was produced. If fish containing dimethyl mercury are consumed, the often fatal Minamata Disease results. The disease is named after the Japanese bay in which the mercury was released that brought this to notice. Dimethyl mercury occurs in tuna fish in extremely tiny amounts that are probably only a testimony to how sensitive the detection methods can be. In larger amounts, it is reputed to cause fetal damage, but is very probably no general hazard at all, because it is so rare.

In 1942, the German submarine U-166 sunk a cargo ship at the mouth of the Missippi. The submarine itself was soon lost near the same location due to depth charges from an escort ship. There was an urban legend that this submarine used mercury in its ballast tanks. The wreck was examined in 2003, and no mercury was found. The legend was very unlikely, but had it been true the mercury would still mostly have been there, and of little danger to anyone.

Recently (December 2003), news reports stated that tuna fish contained "large amounts of mercury" in an alarming tone. What they really should have said is that tuna fish contains only extremely small, scarcely detectable, amounts of methyl mercury. The extrapolated threat appears to be to the nerves of developing fetuses. I wonder how many cases have been observed of fetal mercury damage by normal tuna fish. I suspect that the number is zero. The authorities probably have no idea of the magnitude of the hazard, only that a hazard could be possible. Perhaps we have homeopaths at work here! Generally, they deal in extrapolations of inexact data to minute concentrations, and probably do not know the rate at which organisms take up and excrete such tiny quantities. The problem is that methyl mercury could be a threat, but I have no confidence that the authorities know enough about it to protect the public, only to enjoy the creation of panic. Tuna fish, indeed, also contains many nutrients valuable to fetuses.

The level of knowledge of many experts can be judged from statements like: "mercury vapor is very heavy and collects at low levels." Well, the molecular weight is large, but mercury vapor is usually mixed with air and obeys Dalton's Law like any other gas. Even aside from turbulent mixing, the scale height at ordinary temperatures is a kilometer or so, so no considerable increase in concentration can be expected in a room. Large amounts of vapor might not mix with air quickly, and so would indeed initially sink to the floor. However, the vapor these people are talking about is slowy emitted by droplets. You might as well say that water vapor collects near the ceiling.

Low-level mercury poisoning is characterized by fatigue, headaches, lack of concentration and hair loss. These are hardly distinctive symptoms, being much like those caused by work, or a falling stock market. The body rids itself of excess mercury, as we have seen. There is recent agitation to reduce mercury in coal-burning power plant emissions. The amount of mercury concerned is extremely small, and probably no hazard compared with other chemical hazards, but it is a convenient basis for attacking power plants by enthusiasts for this crusade. The levels of mercury over which enthusiasts become irate are much less even than those causing low-level poisoning, and in fact cause no clinical symptoms at all; their hazard is an extrapolation. All known cases of environmental mercury poisoning have involved very high levels of pollution from specific sources, much larger than those actually encountered under normal conditions. Motor vehicles are, by far, much more dangerous sources of atmospheric pollution than all the mercury in coal and thermometers, emitting particulates, nitrogen oxides, carbon monoxide and sulphur. Catalytic converters not yet poisoned and rendered ineffective just get rid of hydrocarbons; all this other stuff is largely unaffected.

The Hg++ ion generally combines with proteins, so if a soluble mercuric compound is ingested, egg white or milk is an antidote. Recognized first aid for acute mercury poisoning is to administer three raw eggs in a quart of milk, followed by a soapy water emetic. The older instructions say to call a physician immediately: lots of luck these days! The sufferer should be taken to emergency medical care as quickly as possible, and 911 should be called. The general cause of acute mercury poisoning is the swallowing of a mercuric chloride tablet. These tablets are angular and marked with skull and crossbones. The anti-Lewisite (a poison gas) compound BAL is an antidote to mercury poisoning. Inorganic mercury poisoning is very rare.

The importance of the state of the mercury to the hazard presented is well illustrated in the case of dimethyl mercury, an extremely toxic compound. Professor Karen Wetterhahn of Dartmouth, a toxicologist, spilled a tiny drop on a hand protected by a latex glove in 1996. Unfortunately, methyl mercury penetrates latex, and the contact proved fatal within six months, the toxin destroying her brain. The positively charged mono-methylmercury ion binds to plasma proteins and can cross the blood-brain and placental barriers (which inorganic mercury cannot do), explaining its dangerous effects.

Dimethyl mercury, Hg(CH3)2, is very slightly soluble in water, but the mono ion can appear in fish, as mentioned above. Tuna and swordfish are known to contain it. A 125-pound person can safely consume one 6-oz. can of tuna per week, according to Washington State University. However, you could probably eat 10 pounds a week for life, if you could stand it, and not show enough mercury to cause any problems at all, unless you happen to be a fetus. I don't think anyone has ever been poisoned by canned tuna fish. If medical research were more trustworthy, we would have better guidelines. HgS, as in vermilion and cinnabar, is too insoluble to be poisonous. It is one of the most insoluble substances known.

The most famous case of mercury poisoning from industrial waste is the Minamata Disease, recognized in 1956, and with a second outbreak in a different place in Japan in 1964. This was severe poisoning by the methylmercury ion transmitted through fish living in the waters in which the waste was discharged. Its origin is in the catalytic production of acetaldehyde from acetylene and water, with HgSO4 catalyst. Small amounts of methylmercury ion are produced in this reaction, and discharged in the waste. The concentration of mercury in the hair of victims was as high as 705 ppm, while an average amount is 4 ppm. The methylmercury ion is extremely poisonous, as has already been said, and the Minamata Disease is the result of acute poisoning, not trace amounts. Any effects of the inorganic mercury are not noted in reports of this outbreak.

The latest mercury panic is outstanding in its ignorance and failure to assess hazards. The construction of crematoriums is being opposed because of the mercury from the dental fillings of the dead being emitted in the smoke! Only the questionable hazard to the developing brains of children is brought forward, which would be of concern only if there were enough mercury anyway.

Production of Mercury

Mercury is a rare metal. Its only commercial ore is cinnabar, HgS, a red mineral occurring in fine granular massive form, often mixed with other materials. It was used as a cosmetic from the earliest times, and good crystals have been faceted as gems. The name comes from India, where it applied to a red resin, via Persia. Its crystals are rhombohedral, transparent to translucent, but very rare. It has a scarlet streak and perfect prismatic cleavage. The hardness of pure cinnabar is 2.5, and its specific gravity is 8.10. In spite of its heaviness, cinnabar in its usual form cannot be purified by panning, but responds to flotation. However, it is most economical to roast the unenriched raw ore with lime, CaO or iron metal, Fe or air O2, which oxidizes the sulphur to SO2 or FeS, and volatilizes the mercury, which is then condensed and collected. This reaction begins at about 250°C and is complete by 800°C. Mercury has a very simple metallurgy, and can be purified by distillation, unlike most other metals. Mercury is sold in iron flasks holding 76 lb (metric, 34.5 kg), volume about 2.5 litres. The 76 lb flask was adopted in June 1927. It is a weight equal to the Spanish quintal, and was derived from Spanish practice. The mercury of commerce is 99.9% pure prime virgin mercury.

Spain and Italy are the traditional sources of mercury. The Spanish mines at Almadén (Arabic: "the mine") and the Italian (now Slovenian) mines at Idria have been worked continuously since Roman times. Some cinnabar also comes from China, southern Peru, and other places. Shortages have encouraged mining in the United States and Mexico. Cinnabar was mined in California at New Idria in San Benito county and New Almaden in Santa Clara county. The New Almaden mine opened in 1850, and the New Idria in 1853, just in time for the gold rush and the demand for mercury for amalgamation. Mercury has also been mined in Oregon, Nevada, Texas and other places. The Chisos Mine at Terlingua, Texas was worked from 1902 until 1945. Cinnabar is concentrated and deposited by active fluids in volcanic regions. Occasionally, mercury is found naturally in metallic form, as commonly at Idria. The Santa Bárbara mine in Peru was worked from 1566 to 1790, supplying mercury for amalgamation.

The world production of mercury in 1980 was 197,000 flasks, of which the United States produced 30,657 flasks. The price then was $78.29 per flask. By 1985, the world production was about the same, but the United States produced only 16,530 flasks, and 50% of the demand was satisfied by imports. In 1989, the price had soared to $285 per flask, but in 1998 it was $137. Price flucuation has long been characteristic of the market in mercury. The price ranges from about $100 to $300 per flask in constant dollars. Demand for mercury was high during World War II, when European supplies were interrupted, but afterwards the industry went into steep decline. In 1950, only 4500 flasks were produced in the United States, but production recovered later. There is currently only one producer in the United States, and production has probably continued to fall. The 2001 World Almanac did not mention mercury at all.

Mercury can best be purified by distillation. Triple-distilled mercury is used for dental and medical purposes. A dilute nitric acid wash removes base metals, and mercury can be filtered through chamois skin to remove impurities. If mercury is bright and clean-looking, base metal impurities are less than 1 ppm. Impurities tend to congregate on surfaces. Finely divided mercury with surface impurities (oils, etc) will not coalesce into larger drops. This is called "sick" or "floured" mercury. Distillation may be necessary to restore it.


Mercury has the property of dissolving nearly all metals, forming liquid or solid solutions called amalgams. It amalgamates well with gold, silver and tin, but does not dissolve iron or platinum. This is the reason iron flasks and iron vessels are used to refine mercury. Curiously, the electrical resistivity of an amalgam may be less than that of the pure metal, which is not usually the case for alloys. The amalgamation of the surface of impure zinc is used to eliminate local action in primary batteries. This replaces an inhomogeneous surface with localized impurities, usually iron, with a homogeneous surface offering no differences of potential, and so there is no local action to corrode the zinc uselessly.

Dental amalgam is mainly mercury and silver. One source says 65-70 Ag, 25-29 Sn, 3-6 Cu, and max 2 Zn, with the rest mercury. These figures give too little mercury, so possibly the Sn is really meant to be Hg. Another source says 70 Hg, 30 Cu, which is surely incorrect. The amalgam is mixed from powdered ingredients and liquid mercury, thoroughly shaken. The amalgam remains soft for a short time so it can be packed to fill any irregular volume, and then forms a hard compound that performs the arduous duty of being a tooth surface. The amount of mercury released to the system by fillings is extremely small and not hazardous, but some people react to the name alone, not to any credible hazard, as they do with radioactivity, radiation or nudity.

Amalgamation was once an important method of purifying gold or silver in refining. The gold or silver, in small particles in sand or mud, or the roasted ore, was dissolved in mercury, which was then easily separated and distilled to recover the precious metals. An important amalgamation process was the Patio process, invented at Pachuca, Mexico in 1557 by Bartolomé de Medina, which opened Mexico's vast silver resources to exploitation. The original inhabitants knew no metallurgy, and used only native metals in small amounts. The process involved roasting the ore, Ag2S, argentite, with salt and cupric sulphate, followed by amalgamation and distillation. The mercury came from Spain, so a severe problem arose when English depredations in the Atlantic cut off the supply, and mercury was sought in Mexico, largely in vain. Since the amount of silver produced was proportional to the mercury used, the government could estimate its revenue by controlling the mercury supply and discover defalcations. This implies that the mercury vapor was simply released in the atmosphere, with pernicious effect on the nearby inhabitants. Cyanidation has replaced amalgamation at the present day, which hardly seems less dangerous (but actually is rather benign).

Mercury is still used by primitive gold miners in the Amazon region of Brazil, to recover colloidal gold from placer deposits. This use requires no advanced equipment or procedures, just an amalgamation table, some pots, and the mercury. It is a wasteful procedure, and not only gold is lost, but an estimated 100 tons of metallic mercury is dispersed in the region every year. In ancient times, gold was recovered from wornout garments by amagamation, which left the cloth intact. These days, we simply burn the garments and blow away the ash.

The amalgam with sodium metal is used for electrodes in electrolyis cells for producing chlorine and sodium hydroxide by the Castner-Kellner process. A diagram of the cell is shown at the left. It is not difficult to make chlorine and sodium hydroxide by electrolysis of concentrated brine, but the trick is keeping the hydroxide separate from the salt in the product. The Castner-Kellner process uses the mobility of sodium amalgam to solve this problem. Sodium produced by electrolysis is dissolved in the mercury in the end compartments, and the amalgam is moved to the central compartment by tipping the cell, where the Na reacts with water to produce NaOH and H2. The process uses the fluidity of mercury, its ability to make amalgams, its conductivity, and its chemical inertness. This is a major modern use of mercury, and one for which there are few alternatives. Aluminium was first isolated by Hans Oersted in Denmark in 1825, who used potassium amalgam to displace it from aluminium oxide.

Silver amalgam occurs naturally as the very rare mineral Amalgam. It is found at Almadén, Kongsberg, Norway and a few other places. The composition is between AgHg and Ag2Hg3. The cubic crystals have a metallic lustre, hardness 3 to 3.5, are malleable, and have a high specific gravity, 13.7 to 14.1. Gold amalgam is also found in nature, but it is much rarer. It has been reported from Colombia and from California. A copper amalgam would also be possible, but has not been reported.

The Barometer

The barometer was invented by Evangelista Torricelli (1608-1647), a student of Galileo's and his secretary, in 1643. A diagram of a barometer is shown at the right. The barometer tube is first filled with mercury, then inverted and its end put in a reservoir, or cistern, of mercury. The column comes to equilibrium with its upper surface a distance h above the surface of the mercury in the reservoir. This distance can be measured along a scale beside the column that is moved until a reference pointer just touches the mercury surface in the reservoir. At equilibrium, p = ρgh, where ρ is the density of mercury, 13.5462 g/cm3 at 20°C, and g is the local acceleration of gravity, about 981 cm/s2. The actual acceleration of gravity at the location must be used here, not the standard value 980.665 cm/s2. Corrections can also be made if the temperature is different from 20°C to allow for the difference in the density of mercury. The atmospheric pressure can be measured quite accurately with the mercury barometer, to 0.1 mmHg, using a vernier scale. It is the high density of mercury that makes the barometer a convenient instrument.

It is easy now to understand how the barometer works. Above the column, the pressure is the vapor pressure of mercury, 1.2 x 10-3 mmHg, so small that it does not have to be considered. The pressure on the surface of the mercury in the reservoir is the atmospheric pressure p. When Torricelli first showed his barometer, it was met with astonishment. Descartes, wrong as usual, had said that a vacuum, an absence of matter, could not exist in the universe. If so, then what was above the mercury? Some, clinging to their beliefs, concluded that the universe would tolerate only so much vacuum, and held the column up at the limit of its tolerance. The failure of lift pumps to raise water more than 33 feet was quoted as proof. Torricelli knew well enough that the column was not held up, it was pushed up, and there was a vacuum above it. When the doubters had passed away, everyone knew it. Blaise Pascal (1623-1662) carried a barometer up the Puy de Dôme (1545 m) in 1648 to prove that it was barometric pressure that held up the column. It is really amazing that the pressure of the atmosphere is so large, although we are not aware of it.

It isn't difficult to make a barometer like Torricelli's. The one I made is shown at the left. You will need a barometer tube, mercury, a 50 ml beaker, medicine dropper and wood for the support. A barometer tube 850 mm long and 4 mm ID, 8 mm OD, is available from, as is the beaker. The mercury I obtained locally, for $16 per pound. This is "virgin pure" mercury, not reagent mercury, which is much more expensive. You will need a pound, although the full barometer tube takes no more than a third of a pound.

The support is made from clear pine 1x4, 1x2 and 2x2, available at good lumber suppliers. This is excellent wood, straight and well-planed, and clear of knots. A circular saw will make the cutting easy, but since nothing depends on the accuracy of the cuts, a hand saw would do if that is all that is available. The parts are shown in the figure at the right. The holes for the tube are 3/8" diameter, with centres 1" from the back. The upper support is 4" from the top, the lower 10" above the the table. The hole at the upper part of the back is for suspending the barometer with an 8d finish nail. It should not be permanently mounted, since it may require servicing. The table support is connected to the back with two 8x1-1/8 flat-head screws, and the table is connected to that with two 8x1-1/8 flat-head screws. Each tube support is connected to the back by two 8x2 flat-head screws. The scale support is not shown in the photograph, but should be placed so that the centre of the scale is at the normal height of the barometer at your location. I put its top 7" down from the top of the back, and 3/8" in from the edge. The millimeter scale should be directly behind the barometer tube, and attached with household cement. Be sure to drill for the screw bodies, with countersinking, and make pilot holes for the screws. I did not use carpenter's glue, since large stressess are not anticipated, but this, of course, would make a nice permanent connection. The wood was so nice that it should have been stained, but I had some dark green water-based latex paint that I used instead. Two coats are required. The whole thing went together in about two hours.

Filling the barometer is by far the most troublesome part of the exercise. I did not clean or bake the tube, since it came with a plastic boot and I trusted that it was clean. Heating the tube gently from end to end to drive off adsorbed water would be a good thing. Put the tube through its supports, and turn the barometer upside down, the top of the tube resting on some soft surface. Pour the pound of mercury in a 50 ml beaker. With a medicine dropper, carefully fill the tube. Put your finger over the end and shake it gently up and down to let any air bubbles escape. Do this several times while filling, to make sure there are no air bubbles. When the tube is full, very carefully top it off so there is a meniscus above the end. Slide your index finger over the top, closing it without introducing any air. Keeping your finger in place, carefully invert the barometer to its normal position, sliding the beaker with the rest of the mercury beneath it. This is helped by using a nearby wall as a support. Now lower the tube into the mercury, with your finger in place as long as possible, removing it only when the end of the tube is beneath the surface. The level of the mercury in the tube will then fall, and the barometer can be set upright. Making this possible is the reason for needing the full pound of mercury. If anything happens, the beaker is large enough to contain all the mercury present. It is probably impossible to do this without spilling some small drops of mercury, so work on a surface where any spillage can be collected. Do not return any spilled mercury to the beaker, since it will pick up all kinds of corruption rolling around on the tabletop, but consign it to scrap. This is all quite safe, since you do not fill barometers for a living, which would require better precautions. If you do not succeed the first time, just lift the tube above the surface of the mercury in the beaker and try again. I made a reasonable filling on the second try. The first time, I used the plastic boot to keep the end of the tube closed instead of a finger. This was clever, but let a bubble of air in that ruined the filling. The barometer tube can rest directly on the bottom of the reservoir; the mercury will find its way in and out well enough.

Rough measurements with a meter stick gave a column length of 622 mm ± 2 mm, good enough for a test but not accurate. When I mounted a scale, I found that 550 mm should be added to the observed height, by comparison with a good aneroid barometer. 622 mm is equivalent to 24.49 inHg, and is about the average height of the barometer at Denver. The temperature was 70°F. Conversion to sea level is given in the Handbook of Chemistry and Physics, p. E-43. I took a correction index of 160 from the table. This gave a correction of 4.85 inHg, so my sea level pressure would be 29.34 in Hg. My aneroid read 29.85 inHg, which is close enough to be sure the barometer is operating properly. When proper measurements of the column height can be made, a better comparison will be possible.

Although I supposed the barometer would be mainly a demonstration, it has shown itself to be an accurate and useful instrument, certainly as good as the aneroid barometer I have used for comparison. The mercury barometer can be used for absolute measurement of pressure, independently of any standard, but for this purpose careful attention must be paid to sources of inaccuracy. This barometer has a fixed cistern, so it is of the "Kew" type. When the mercury rises in the tube, it falls in the cistern, so the actual length of the column is greater than the observed rise. In this case, the ratio of areas of tube and cistern is about 100, so a rise of 10 mm is actually an increase of 10.1 mm in the mercury column. Similarly, a fall of 10 mm is a fall of 9.9 mm. This is too small to worry about with this barometer, but the correction can be made.

The top of the meniscus is read, which is a little too high. The surface tension of mercury is about 480 dyne/cm, which would shorted the column by about 0.5 mm if it acted with 180° contact angle (the precise effect is less than this, since the contact angle is smaller than 180°). These two errors are compensating, and have no important effect here. There is also an error due to residual pressure in the Torricellian vacuum. As noted above, I took no special pains to clean and bake the tube.

The weight of the mercury column depends jointly on the density of mercury and the acceleration of gravity. The acceleration of gravity is constant at one point. It varies with latitude and altitude. A standard formula will give the acceleration of gravity closely enough for barometry: g = 980.621(1 - 0.00259 cos 2φ)(1 - 3.14 x 10-7z) cm/s2, where z is the altitude in metres, and φ is the latitude. For Denver, at altitude 1633 m, I find g = 979.67 cm/-2. The standard value of gravity is 980.621 cm/s2. The density of mercury decreases as the temperature increases, because of the thermal expansion of mercury. From tabulated values, d = 13.5955(1 - 1.8057x10-4t), where t is the Celsius temperature. Barometric pressures in mmHg are at 0°C and standard gravity. Observed column length must be corrected by the ratio of gd. The temperature correction is the most important of all, and must be made for absolute measurements. There may also be a temperature correction for the expansion of the scale, which is often of brass.

My barometer is at practically constant temperature, so there are no variable errors. I have compared its readings with the aneroid barometer, and now the two give the same results as the atmospheric pressure varies. I made them agree at sea level pressure 30.00 inHg (scale reading 90.0 mm), with reduction to sea level of 123 mm. The aneroid was made to agree with TV barometric pressure reports.

Mercury is also convenient for manometers, U-shaped tubes that measure the difference of two pressures. Electronic pressure sensors have improved greatly in sensitivity and accuracy, but they are scarcely superior to a manometer, and much more subject to error.

The common laboratory unit of pressure, the mmHg, was named "torr" after Torricelli. The SI unit of pressure, N/m2 was named "pascal" after Pascal. Standard atmospheric pressure is 760 torr or 1.01325 x 106 dyne/cm2 or 30" Hg. A pascal is 10 dyne/cm2, and a "bar" is 106 dyne/cm2. The actual atmospheric pressure is not the standard atmosphere except by chance near sea level, and the atmospheric pressures quoted on TV are not the real pressures, but are adjusted for altitude above sea level.

The Thermometer

In 1714, Gabriel Fahrenheit (1686-1736) invented the mercury-in-glass thermometer, shown at the left. This instrument depends on the volume thermal expansion of mercury, or actually the expansion of the mercury relative to glass. If they both expanded at the same rate, the length of the mercury column would not change. Mercury not only expands much more rapidly than glass, but its expansion is fairly uniform, so it is a good thermometric substance. The sensitivity of the thermometer depends on the ratio of the reservoir volume to the square of the inside diameter of the stem. The formula is shown, where β is the coefficient of relative thermal expansion.

Thermometer design is a relatively simple matter. We measure the thermal expansion of a volume V of liquid by its expansion into a capillary of cross-sectional area A. Since the coefficient of cubical expansion is β = (1/V)(dV/dT), and dV = Adx, where x is the length of liquid in the capillary, the sensitivity of the thermometer, dx/dT = β(V/A). For β, we use the difference of the cubical expansion coefficients of the liquid and glass. Thermometers are not affected by vapor pressure above the capillary column, as a barometer would be. It is only necessary that the liquid be clearly distinguishable from the volume above the liquid. The glass capillary magnifies the column, and can be shaped to increase the magnification.

Mercury has β = 0.181 x 10-3 per °C, while ordinary soda-lime glass has β = 0.0276 x 10-3 per °C. The β of most liquids is on the order of 10-3, while that of most solids is about 10-5, so the solid expansion is only about 1% of that of the liquid. For mercury, the difference is β' = 0.153 x 10-3 per °C. Suppose our thermometer has V = 250 mm3, with a capillary bore of 0.2 mm. The sensitivity will be dx/dT = 1.22 mm/°C, so a scale reading from -10°C to 110°C will be 158 mm long. This is actually fairly typical of small mercury thermometers. Mercury melts at -38.87°C, and boils at 356.7°C, so it is useful over a wide range. A mercury column is also very easily seen.

Objections, which I regard as specious, have been made to mercury thermometers because of the danger of mercury spillage. A popular alternative fluid is an alcohol. Ethyl alcohol boils at 78.4°C (173°F), so it would be all right for room thermometers, and has been widely used for that purpose for many years. The alcohol is colored red (usually) so it can be seen easily. Amyl alcohol (1-pentanol) melts at -78.9°C and boils at 138.1°C, so it can be used to replace mercury in laboratory thermometers that must read to 110°C. Its coefficient of cubical expansion is 0.902 x 10-3 per °C, so β' = 0.874 x 10-3. If we want the same sensitivity as before, 1.22 mm/°C, then if V is to be the same, A = 0.179 mm2, or the capillary bore should be 0.7 mm. This larger column is more easily seen, which is an advantage.

When Fahrenheit made his thermometer, the significance of fixed points such as the freezing and boiling points of water was not appreciated, and Fahrenheit calibrated his thermometer rather arbitrarily, trying to encompass the full range of temperatures met with in practice. The coldest temperature he could get, that of a freezing mixture, he called 0°. Body temperature, which he knew to be constant, he called 24°, and noted that water froze at 8° on his scale. These were "old" degrees, which were multiplied by 4 to get a finer scale. Now the ice point became 32°, and body temperature 96°. When the boiling point of water was taken as a new fixed point, it fell at 212°. Meanwhile, it seemed body temperature was a little hotter (depending on how you measure it) and became 98°, and finally 98.6°. The idea of having a really accurate thermometer was a surprising one, and it took a while for it to sink in. It might seem that he could have aimed at an even 100° for body temperature, but this was not the case.

Fahrenheit, a German, was made a fellow of the Royal Society and his thermometer became the English thermometer. Réaumur, a Frenchman, devised a scale that became the German thermometer. Anders Celsius (1701-1744), a Swede, created a scale based on the freezing and boiling points of water which became the French thermometer. On the Réaumur scale, ice melts at 0° and water boils at 80°. A Réaumur degree was represented by an expansion of 1/10000 of his fluid, alcohol with 1/5 its volume of water added, at 0°R. 20°C, a comfortable room temperature, is 16°R and 68°F. Mercury solves most of the problems of a thermometric substance quite neatly, since it neither freezes nor boils in extremes of weather.

Any thermometer based on the expansion of a substance is essentially arbitrary, except one based on an ideal gas, for which p = nkT. (n is the number density of molecules, and k is Boltzmann's constant.) This gives us the absolute temperature T, now measured in kelvin, K. Experiments are difficult, but thermometer scales have be calibrated absolutely by this and other means. 0°C corresponds to 273.15K. The absolute temperature based on the Fahrenheit scale is in rankine, R. 32°F corresponds to 491.67R, or 0°F to 459.67R. All thermometer scales are equally good, none is superior to any other, and the concept of "metric" is meaningless, since all are divided decimally.

Any skilled glassblower was once able to make a thermometer from a length of capillary tube. The bulb is blown, then filled with water which is heated to make steam. The tube is put in some mercury, the correct amount of which is pushed into the vacuum (or sucked in by the vacuum, as some would say) when the steam condenses. Then the mercury is heated and vapor expelled until the tube is clean, and the tube is finally fused closed.

The Diffusion Pump

The third major laboratory application of mercury came much later than the barometer and thermometer. Vacuum techniques have led to great advances in physics and chemistry, originating in the study of electrical discharges in gases. A mechanical pump can create a vacuum of 5 to 20μHg, a pretty good value, but this is a long way from a real vacuum. To create a pure environment for an experiment, or to remove the effects of residual gases, requires a still greater vacuum. High vacuum is a pressure from 1 μHg down to 10-3 μHg. Even at the lowest of these pressures, a cubic centimeter is still swarming with molecules, 3.54 x 1010 of them. However, the mean free path is up to something like 7.3 x 106 cm, or 73 km, so the molecules are moving around fairly freely, and this will do in most cases. To produce such a vacuum requires a new kind of pump, and the diffusion pump was the solution. The diffusion pump was invented by Irving Langmuir (1881-1957) early in the 20th century to help in his studies of electrical discharges in gases.

Before the diffusion pump, the best vacuum pump was the Sprengel pump, which used drops of mercury as pistons to evacuate the air. The vacuum obtained could not exceed the vapor pressure of mercury, which is 0.001201 mmHg at 20°C, or about 1 μm. The pump consisted of a mercury reservoir, which supplied mercury to one or up to five capillary tubes that would drop mercury slowly into the tops of one to five fall tubes, trapping small intervals of gas between the drops. The drops would fall into a collector at the bottom, which could be raised periodically to refill the reservoir at the top. This pump was well-adapted to electrical discharge studies.

A diffusion pump is represented schematically at the right. Its job is to evacuate the volume at the top to a pressure of 1 μHg. It contains a boiler that produces mercury vapor and sends it through an annular jet. The jet is directed towards the cooled walls, where it condenses, and the mercury drips down to return to the boiler reservoir. Very few mercury atoms get into the vacuum receiver. In fact, there are usually extra cooled baffles between the jet and the vacuum receiver that are not shown in the diagram for simplicity to make sure of this. The vapor pressure of mercury at 20°C is only 1.2 μHg anyway, and this is the lowest pressure that the pump can create. The mechanical forepump sucks on the output, which it maintains at the lowest pressure it can reach, which is 40μHg here. The purpose of the diffusion pump is to maintain the difference in pressure.

A typical molecule, such as "a" may, in its random motion, pass through the jet as shown. If it is pointed in the correct direction, this is easy to do. It is much more difficult for the molecule to get out again, as any collision with a mercury atom in the jet will give it considerable downward momentum. Of course, some molecules do get out again, but most do not, and are eventually pumped out by the forepump. There is, in essence, a screen of downward momentum in the jet that keeps the molecules below it where they are, and lets new ones in. This can be made even more efficient by using two diffusion jets in series, one to maintain a difference of pressure from 1 μHg to 30 μHg, and the second, the booster pump, from 30 μHg to 100μHg. These pumps work in a pressure range where the mean free path is so long there is no pressure stress in the gas, only molecules moving freely.

Although these pumps originally used mercury, which works very well, most diffusion pumps now use special oils with high molecular weights (200 and over) and low vapor pressures. This eliminates the mercury vapor hazard, which is very real with mercury diffusion pumps (but not overwhelming). A typical oil, Litton Molecular C, goes down to 2 x 10-3 μHg, which is much better than mercury can do.

The Mercury Battery

Batteries currently available for portable electrical apparatus are carbon-zinc and alkaline cells, lithium cells and silver oxide button cells, all primary cells, and nickel-metal-hydride (NMH) and lead-acid secondary (rechargable) cells. The carbon-zinc and alkaline cells give 1.5V on open circuit, and are suited to most normal duties. Lithium cells give 3.0V, but are restricted to small drains. Silver oxide cells are small and rather expensive, used for hearing aids and similar uses. The lead-acid cell gives 2.0V and is very satisfactory except for weight and the liquid electrolyte. NMH cells give 1.2V and have largely replaced the popular nickel-cadmium cell.

In World War II, the only commonly available dry battery was the carbon-zinc cell, which suffered greatly in the high temperatures and high humidity of some theatres of operation. The search for a better battery succeeded in 1944, with the invention of the mercury dry battery. This battery could not only resist the high temperatures and high humidity, but also had better discharge characteristics, longer shelf life, and greater efficiency.

The excellent mercury battery, and to a large extent the nickel-cadmium cell, have been suppressed because of the mercury and cadmium they contain, which are a hazard with unwise disposal of used batteries. Since it is easier to legislate the batteries out of existence than to control disposal in the face of a careless and ignorant population, this has been done.

The mercury battery has a cathode of HgO depolarizer, a KOH electrolyte, and an anode of amalgamated zinc powder. It is essentially an alkaline cell with a different and more efficient cathode. Like the alkaline cell, it is sealed in a steel case to prevent leakage of the corrosive electrolyte. The open-circuit voltage is 1.35V or 1.40V, and the polarity is reversed compared to the carbon-zinc and alkaline cells. That is, the central contact is negative, not positive, and the case is positive, not negative. The mercury battery has an excellent efficiency, consuming more than 90% of its chemicals before its terminal voltage drops substantially, when used at a current of 100mA per square inch of depolarizer surface. A typical AA size cell can take loads of up to 200 mA, either intermittent or continuous, and has an output of 2400 mA-hr, about twice that of a comparable carbon-zinc cell, and considerably more than an alkaline cell. The output voltage is quite constant, and the shelf life is long, so the mercury battery is superior. The silver oxide cell is similar, using an Ag2O depolarizer, a KOH or NaOH electrolyte, and the usual zinc anode. Its output voltage is higher, 1.5V, and so is its cost. Silver cells come only in the button size.

These cells all burn zinc to zincate, and get their energy from this reaction. Fuel cells are another chemical source of electrical energy, and not a new one. They have been around for over a century, but have always proved noncompetitive because of their size, weight and expense. They burn things like alcohol or natural gas, to water and carbon dioxide, like any fire. Fuel cells are now being developed for military use, where cost is no object. These seem rather larger and heavier than batteries, but of course they can be used continuously and replace lots of batteries if you keep feeding them alcohol. It is difficult to believe that they will ever be cheap and convenient enough to replace zinc burners.

Miscellaneous Uses of Mercury

Since mercury has a considerably higher boiling point than water, the use of mercury vapor in turbines is more efficient than the use of steam. The plants are usually dual-cycle, the exhaust from the mercury turbines used to make steam for a second stage of steam turbines. Of course, the mercury cycle is closed. The first mercury-vapor plant was constructed in 1923, but none have been constructed since 1947, it appears. It was remarked that if they ever became really popular, the consequent demand for mercury would greatly exceed the world supply. It is said that 35 pounds of mercury, nearly a half-flask, is required for every horsepower. A 10MW turbine in Hartford, CT used 13 lb per kilowatt.

Mercury was found very useful by early electrical experimenters, who used it for making electrical connections, as we now use relay contacts. A mercury column in a capillary tube was the first standard of resistance. Both the Weston and Clark standard cells, once important laboratory standards, used mercury electrodes and mercurous sulphate. It is still used in tilt switches, where two electrodes are put in and out of contact by a small mass of liquid mercury that is moved by tilting the switch. Mercury could be used for collecting gases that reacted with water. In a McLeod vacuum gauge, gas in a large volume is compressed into a small volume by a mercury piston so that its pressure can more easily be measured by a manometer. Mercury is also used for reflecting planes that must be accurately horizontal, in astronomical and surveying measurements.

Glass mirrors were invented with the aid of mercury amalgams to produce a reflecting surface, protected by the glass. Paint protected the amalgam on the back. About 1300, a convex mirror was first made in Nürnberg by "silvering" the inside of a spherical flask with an amalgam of Hg, Bi, Pb and Sn. Soon afterwards, plane mirrors were made in Venice, whence the art spread generally. Bright, superior glass mirrors soon replaced speculum mirrors that had served since antiquity. Tin amalgam was the usual reflective surface, made with tinfoil and mercury. Silvered mirrors were introduced by Justus von Liebig in 1836. Evaporated Al and Ag mirrors came only in the 20th century as high-vacuum technique was perfected.

Mercury vapours were used to develop the positive image on exposed AgI plates in the first Daguerrotypes. Fortunately, this dangerous procedure was soon superseded. Mercury halides are photosensitive, like silver halides, but seem not to have been used in photography.

Mercury is used in electrical discharges, usually arcs, to lower the voltages required, and to make the voltage drop less dependent on the current. The mercury may be in the form of a pool on which the cathode forms. The mercury evaporates at the cathode spot and condenses on the cool walls of the envelope, controlling the pressure of the mercury, which should be in the neighborhood of 1 mmHg. Alternatively, the mercury may be present as small drops that evaporate when the discharge reaches operating temperature. This is a very frequent use of mercury, but does not require large amounts. A mercury discharge shows the characteristic lines of its visible spectrum, the cyan line at 436 nm, the green line at 546 nm and the two yellow lines at 577 and 579 nm. A fluorescent lamp is excited by the UV resonance line at 254 nm, and also shows the visible spectrum. The short-wave ultraviolet output makes the mercury arc a good sterilization lamp.

Mercury was used in temperature-compensated pendulums. Part of the pendulum bob was a column of mercury. When the pendulum lengthened due to thermal expansion, the mercury expanded upwards and much more than the metal, so that the pendulum remained of the same effective length if the mercury column was properly designed. The mercurial pendulum was invented by Thomas Graham (1675-1750), the famous clockmaker who also invented the dead-beat and cylinder escapements.

Mercury has recently been used in the determination of the Newtonian constant of gravitation, G, in the equation F = GmM/r2. Henry Cavendish (1731-1810) used a torsion balance for the first determination of this constant, and published his results in 1798. A measurement in 1995 was discordant with previous ones, and in addition to this, a possible systematic error in using torsion fibres put the whole subject in doubt. This stoked up interest in new measurements.

The recent measurement of G using mercury was made in Zürich, and instead of a torsion balance a commercial beam balance intended for the accurate comparison of masses in standards laboratories was used. The principle of the measurement was the same as used by Cavendish, in which large field masses were moved to opposite sides of the test masses. Here, the field masses are two cylindrical tanks of mercury, each weighing M = 7000 kg. The test masses are m = 1.1 kg slugs of copper or tantalum. Different materials are used because it is nice to verify that the gravitational force depends only on mass, not on the substance. In one configuration, the field masses are between the test masses, and in the other they are just on the other side of each test mass, as shown in the figure.

The force exerted on a single test mass by the whole 14000 kg of mercury in the first configuration was about 0.2 mg, or 0.2 dyne, from the average of the difference in weight of the two test masses with the field masses in the two different positions. If it is assumed that the field mass attracts as if its mass were concentrated at its center, 75 cm from the test mass, then a simple calculation gives G = 7.2 x 10-8 cm3/g-s2. The precise value from this experiment was (6.67407 ± 0.00022) x 10-8 cm3/g-s2. Combining this with the observed acceleration of gravity at the earth's surface, 980.665 cm/s2, you should be able to find the mass of the earth in grams.

Mercury was used for the field masses not only because it is dense, but because the liquid state made its mass distribution very uniform. We can estimate that the tanks required 405 flasks of mercury, or $115,560 worth at the 1989 price.

A curious technical use of mercury was in the production of frozen mercury patterns for making molds for casting. The patterns could be made in steel dies, and easily assembled by fusing them together to make complicated forms. The completed pattern was then coated with a mold material, and when the mold had been formed, the mercury was warmed and ran out. It was a kind of lost-wax process using mercury instead of wax.

It is said that the first emperor of China, Qin Shihuangdi, was buried in 210 BCE in a bronze coffin floating on mercury. Unusual concentrations of mercury have been found in the area. Caliph Abd-er-Rahman III of Cordoba (ruled 912-961) was reported to have maintained a large mercury bath in a porphyry basin. By disturbing the surface, he reflected sunbeams to the amazement and temporary blindness of his guests.

Experiments with Mercury

It is interesting to become familiar with some of the properties of mercury by working with a small amount in your home lab. Pure, analytical grade mercury, which is expensive, is not necessary. I found "virgin pure" mercury for $16.50 a pound at Colorado Science Co. It should be equally available in any large city from a local supplier. UPS charges extra for transporting anything over minute quantities, so web sources, even if low-priced mercury could be found, are not convenient. A pound of mercury occupies a volume of 33.48 cc. Even a quarter-pound (8 cc) is enough for basic experiments.

As mentioned above, mercury metal is not a hazardous chemical. It is not poisonous, explosive, inflammable, unstable, irritating or reactive. Mercury compounds are, however, very damaging to kidneys and nervous systems and should definitely be avoided. Being afraid of mercury metal is like being afraid of salt, because salt contains the poison gas chlorine. The only hazard from mercury metal is chronic exposure to its vapor, which requires bad ventilation, long exposure, and a lot of mercury scattered about in warm places. Don't drink mercury, though it was once prescribed as a medicine. There is no danger in occasional external contact. These days, it is good not to dispose of it casually. Keep a bottle for scrap (dirty) mercury, and be very careful that all you use is properly accounted for. Mercury is best cleaned up manually. It coalesces in little drops that can be coaxed onto a piece of paper and put in a safe place. Work with it in a place where it will not get away from you, and, of course, work only with small amounts.

It's a good idea to work on a piece of paper toweling, which will prevent small drops from wandering, and allowing any spilled mercury to be collected. Pour some mercury into a 50 cc beaker, so you can feel how it pours and look at it. It forms a perfectly horizontal reflecting surface, a property that is sometimes of use. Note that the surface is depressed at the contact with glass, showing that mercury does not wet glass. Put the end of a piece of glass tubing in the metal, and note that the mercury in the glass tubing is depressed. Suck up some mercury with a medicine dropper. Note that it avoids the restricted tip. It is easy to examine the meniscus now, and see that the angle of contact is about 120°, not the 0° of water and glass. Practice making small drops fall from the end of the dropper. It is easy to get small amounts of mercury in this way. When you empty the dropper, if any small drops are visible, tap it so that all the mercury is gone. It is easy to work with mercury and not leave any bits around. Finally, pour the mercury back into the bottle.

Using a digital multimeter, measure the resistance between two wires immersed in the mercury at their tips. I found a resistance less than 0.5Ω with very little immersion and about 2 cm spacing. Mercury is certainly electrically conducting.

Obtain some small pieces of zinc and copper, perhaps 2 cm x 1 cm. Put as small a drop of mercury as you can get on each one. They will stand up as spherical drops, as if put on glass. Moisten a cotton swab with 2M (dilute) hydrochloric acid and spread the mercury around with the wet swab. It should very quickly cling to the surface, forming a shiny layer of amalgam. The zinc was covered with a layer of carbonate, the copper of oxide, so the drops initially could not reach the metal. The acid dissolved this layer, allowing the mercury to come into contact with the metal, which it wets in both cases. Note how little mercury is necessary to make a shiny coating. If you try this with iron, you will find that the mercury will not wet it: iron does not amalgamate. This is the reason why mercury is handled in iron flasks. Amalgamated zinc is used in electrolytic cells to prevent local action due to impurities--the zinc is lifted free of the impurities when it is in the amalgam. A coating of amalgam should not be used for decorative purposes, of course. Mercury used for these experiments is contaminated, and any left over should not be put back into the supply bottle of clean mercury. You'll note that bits of dross are attracted to the surface of the mercury, where the surface energy is high.


Properties of mercury are scattered in handbooks such as the Handbook of Chemistry and Physics (56th ed.) and Lange's Handbook of Chemistry (10th ed.). Lange gives linear and cubical expansion coefficients of various substances on pp. 1675-78. The HC&P has no information on cubical expansion coefficients at all, except for tables of the densities of water and mercury.

The best source for the history of the thermometer and barometer is J. C. Poggendorff, Geschichte der Physik (Berlin: Zentral-Antiquariat der DDR, 1964 reprint of 1874 edition).

G. Agricola, De Re Metallica (1556). H. C. Hoover and L. H. Hoover, transl. (New York: Dover, 1950).

J. L. Bray, Non-Ferrous Production Metallurgy, 2nd ed. (New York: John Wiley & Sons, 1947), Chapter 17.

A. Guthrie and R. K. Wakerling, Vacuum Equipment and Techniques (New York: McGraw-Hill, 1949). Any of the many good references on vacuum technique will explain the diffusion pump and its use.

W. N. Jones, Inorganic Chemistry (Philadelphia: Blakiston, 1949), Chapter 34.

B. Schwarzschild, Physics Today, November 2002, pp. 19-21. Report on the measurement of G.

J. W. Pennington, Mercury--A Materials Survey (Washington, DC: United States Department of the Interior, Bureau of Mines Information Circular 7941, 1959). Principally on the occurrence and supply of mercury.

T. G. Winter, The Evaporation of a Drop of Mercury, American Journal of Physics, 71, 783-786 (2003).

A. K. Furr, CRC Handbook of Laboratory Safety, 5th ed. (Boca Raton, FL: CRC Press, 2000). pp. 300-305.

F. X. M. Zippe, Geschichte der Metalle (Wien: W. Braumüller, 1857). pp. 205-217.

National Geographic Magazine, May 2005. The U-166 story and the fatal accident with dimethyl mercury.

Return to Physics Index

Composed by J. B. Calvert
Created 5 November 2002
Last revised 17 Januray 2007