IN HIS best-selling novel Something of Value Robert Ruark described a test of truth in an African village. Two men, one who was telling the truth and one who was lying, had "undergone the ordeal of the licking of the hot knife, so that he whose guilt was most apparent would scorch his tongue when the saliva dried up from the knowledge of fear." My grandmother does something similar when she tests a hot iron by touching it briefly with a wet finger. Touching hot metal was developed into a stunt in turn-of-the-century carnival sideshows where daredevils would briefly plunge their wet fingers into some molten metal, probably lead. Some candymakers similarly plunge wet fingers into hot, melted candy in order to test the temperature.

These examples of touching hot material with wet skin, which suffers no harm, are all related to the same physics. Jules Verne cleverly employed the physics in one of his novels, Michael Strogoff. Sent by the Czar into Siberia with a secret message for the Czar's brother, Strogoff was captured by the invading Tartars, identified as a courier for the Czar and sentenced to be blinded. The Tartar method of blinding was to pass a red-hot blade just in front of the victim's eyes; the intense heat radiated by the blade would damage the eyes beyond repair. Just as Strogoff was to be blinded, however, his mother fell before him, bringing tears to his eyes, since it would be his last glimpse of her. The watery layer of tears protected his eyes from the heat radiated by the blade, and he ultimately completed his journey.

The most remarkable exhibition of this protection afforded by a thin watery layer appears when people walk barefooted over hot coals or solidified hot lava. Probably you think there is some trick involved, the walker having surreptitiously applied a protective coating to his feet before the walk. As we shall see in the following experiments, however, no special coating is needed and no trick is involved, although perhaps a certain willingness to tolerate pain is needed.

A phenomenon of this kind is familiar to anyone who is skilled at making pancakes. Such a person will have the griddle very hot so that the batter solidifies quickly after being poured. Not having a thermostat on the griddle, one tests the griddle's temperature by sprinkling water on it. If the griddle is hot but not hot enough, the water drops spread out, wet the surface and evaporate within about two seconds. If the griddle is ready for the batter, the sprinkled drops dance, vibrate and skim over the griddle surface for from 30 to 100 seconds. This result seems all wrong. How can drops last longer on a hotter griddle?

The persistence of water drops on hot surfaces was first described by a German physician, Johann Gottlieb Leidenfrost, in "A Tract about Some Qualities of Common Water." Although the paper appeared in 1756, it was translated from the Latin only in 1965 and therefore has not been widely read. Leidenfrost placed small water drops in a red-hot iron spoon taken from the fireplace and noted how long they lasted by counting the swings of a pendulum Among the several things he noticed about the suspended drops was that they appeared to suck light and fire from the spoon because under the drops the spoon immediately turned black. With the spoon initially red-hot, his first drop lasted for about 30 seconds. A second drop then placed on the spoon lasted for only 10 seconds. Further drops on the spoon lasted for only one or two seconds. Leidenfrost did not regard the longer-lasting drops as boiling, but modern science classifies them as a type of boiling different from the more common type (called nucleate boiling) that the quickly vanishing drops underwent.

Leidenfrost was unable to predict how long the drops would last for two main reasons. First, he was unable to measure the high temperatures of his spoon. (Indeed, he suggested using the length of time that drops persisted as a calibration of the higher temperatures.) Second, he could not calculate theoretically the heat supplied to a drop from the spoon because the concept of the latent heat of vaporization (the heat needed to transform liquid to vapor) did not exist at that time. (It was introduced 14 years later.) Leidenfrost apparently was also still captive to one of the Aristotelian errors, namely that fire and water may produce a solid. "These observations," he wrote, "[suggest] that water is changed into earth by a large fire, because always after the complete evaporation of the drop some terrestrial matter remains in the heated vessel." Finally, Leidenfrost thought the hotter spoon sustained drops longer because the heat increased the surface tension of the water, which is not true.

A Leidenfrost drop is supported by a layer of its own vapor approximately .09 millimeter thick if the metal plate on which it is suspended is sufficiently hot. On a metal plate at a relatively low temperature, although the plate is still hot, the drops fall on it, wet it and evaporate within several seconds. On a hotter plate the bottom of a falling drop vaporizes almost immediately as it nears the plate, leaving a layer of vapor to support the remaining portion of the drop. The drop dances and skims about in reasonable safety from the hot plate because of this protective vapor layer, which is constantly supplied by evaporation of the bottom of the drop. Eventually enough heat is conducted and radiated through the vapor layer to vaporize the remaining drop, and then the dancing is over. The primary mechanism of heat transfer to the drop appears to be conduction through the vapor layer, although at higher temperatures radiation is progressively more important. Water vapor is a relatively poor conductor of heat, typically an order of magnitude poorer than liquid water and several orders of magnitude poorer than a solid such as copper. The drops therefore have a surprisingly long duration.

A plot of the drop durations v. the temperature of the metal plate shows a fairly sharp transition between the drops that last a short time (not becoming suspended) and the drops that last a long time (with suspension). The temperature at which the drops last longest is the Leidenfrost point. The actual value of this temperature depends on, among other things, the saturation temperature of the fluid, that is, the temperature at which the fluid will boil in the normal way. If the saturation temperature is lowered, say by moving to a higher altitude, the Leidenfrost point will be lower. With plate temperatures somewhat higher than the Leidenfrost point, the drops last longer than the one or two seconds for the nonsuspension case but not as long as those right on the Leidenfrost point. In the experiments described here one can determine the Leidenfrost point and the maximum duration for drops of water and of several other liquids. Not all the various aspects of the phenomenon have been explored, and I offer several of them that you might want to investigate on your own.

You can examine the duration of drops on a hot surface with some relatively simple arrangements in your kitchen. In obtaining my data I worked with a flat piece of 1 /16-inch aluminum on a tripod straddling a Bunsen burner. You could use instead the burner on either a gas stove or an electric stove; for more control over the temperature of the surface you could plug a hot plate into a Variac so that by varying the voltage to the hot plate you can smoothly and easily vary the plate's temperature. I dented the metal plate with a ball-peen hammer so that the Leidenfrost drops would stay in one place.

The drops are formed on the end of a hypodermic needle. A glass syringe is better, since the plastic type may react a with some of the fluids. (If you have trouble getting a hypodermic needle, an eyedropper is a tolerable substitute.) Several needles of different diameter should be obtained so that drops of different initial radii can be formed. Hold the syringe just over the hot metal plate and slowly depress the plunger until a drop forms on the end of the needle and then falls of its own accord. This technique produces drops of just about the same size each time. To calculate the volume of each drop you can either count the number of drops needed to reduce the syringe contents by one or two cubic centimeters or, if the syringe does not have cubic centimeters marked off, count the number of drops needed to fill a known volume.

Each time I changed the plate temperature I allowed at least five minutes for the temperature to stabilize. The temperature can be roughly monitored up to about 430 degrees Fahrenheit by a candy thermometer. The readings are not accurate, however, because the end of the thermometer does not make good contact with the hot plate and the thermometer therefore reads low.

A better monitor and one that can be used at higher temperatures is a thermocouple [see illustration at right]. One end of the thermocouple is submerged in ice water (which, assuming it stays at a constant temperature, is at 0 degrees Celsius) and the other end is placed on the metal plate near the area in which the Leidenfrost drops are formed. Because the two wires in the thermocouple are of different metals an electric potential develops between the two ends according to their temperature difference. Having measured this voltage difference, you can convert it to a temperature difference by using a calibration table. Since one end is at 0 degrees C. the temperature difference is the temperature of the other end in degrees Celsius. I used a thermocouple with copper for one wire and the copper-nickel alloy constantan for the other and found the calibration table in the Chemical Rubber Company Handbook. The electric potential was measured on a potentiometer by comparing it with a reference voltage. Other types of metal can serve in the thermocouple: one is chromel-alumel wire, available from the Edmund Scientific Company (555 Edscorp Building, Barrington, N.J. 08007) for about $8 and probably from distributors in your area. (Look under thermocouples in the yellow pages.) A complete package consisting of thermocouple wire and a meter can be bought from Edmund Scientific for about $50.

When I measured the temperature of molten lead, as I shall describe below, I used a resistance temperature detector from Engelhard Industries Division (700 Blair Road, Carteret, N.J. 01008). Encased in either glass or ceramic, the detector has a length of chemically pure platinum wire of temperature-dependent resistance. By measuring the resistance of the instrument on an ohmmeter (or a Multimeter) temperatures up to 750 degrees C. (corresponding to a resistance of 350 ohms) can be determined.

The persistence of the drops depends on the plate temperature For distilled water there is a sharp transition from durations of 10 seconds or less to durations of 65 seconds or longer. The temperature of the maximum duration, the Leidenfrost point, was found to be between 210 and 240 degrees C., although other investigators find the maximum duration for water drops nearer 290 degrees C. under similar conditions. I do not know why my results were lower. At temperatures higher than the Leidenfrost point the duration slowly dropped off, but even for temperatures of from 400 to 500 degrees C: the drops lasted much longer than drops at temperatures below the Leidenfrost point. The Leidenfrost point that you find will depend on the temperature at which water boils where you live. At higher altitudes the boiling temperature is lower and so is the Leidenfrost point.

I found that the Leidenfrost point for smaller drops was slightly lower than the point for larger drops. More careful control over the experiment may produce the result found by others, namely that the Leidenfrost point is independent of the drop size.

Tap water produces no sharp transition at the Leidenfrost point. The difference might be that the particulate matter deposited by a fading drop of tap water bridges the vapor layer from the drop to the plate or at least prevents the necessary evaporation from the bottom of the drop. Either the bridge or the diminished vapor production would significantly shorten the duration of the drop and smooth out the Leidenfrost transition.

Taking data at or just below the Leidenfrost point is difficult because of the greater tendency for the drop to vibrate up and down at those temperatures. In variably a vibrating drop lasts a short time because it periodically break through the vapor layer to touch the hot metal and is thereby warmed faster. Th additional warming might decrease the lifetime of the drop by as much as 2 seconds. During the vibrations the drop jump as high as five millimeters above the hot surface.

In addition to jumping up and down the Leidenfrost drops may also oscillate radially in a variety of shapes, called normal modes, much as standing wave can be generated on a guitar string. Some of these radial oscillations are easily visible, but their visibility is in creased by first dyeing the water black dark blue or dark red and then viewing the drop under a stroboscope with flash frequencies between about four and 10 hertz. A well-behaved drop slowly shifts from one pattern (from one normal mode of radial oscillation) to another. If you have equipment for closeup photography, you can photograph the drop in one of these normal modes. The cause of the normal-mode oscillations is not understood but probably has to do with the uneven heating and uneven vaporization of the drop.

The vapor flow below a Leidenfrost drop can be demonstrated in two ways. A light powder, such as baby powder, initially sprinkled onto the metal surface will be blown by the vapor radially outward from under a drop. A similar flow can be seen on the surface of a molten-lead bath on which a Leidenfrost drop is floating, particularly if the drop is spinning.

Instead of depositing calibrated water drops on the metal plate, you can continuously feed water to the top of a Leidenfrost drop by inserting the needle into the top of the drop. Large drops can be built up, skimming and oscillating about the surface like some wild amoeba. If the drops become too large, however, they collapse under their own weight, suddenly vaporizing large portions of themselves with a sizzle.

To show that the Leidenfrost drop is really separated from the metal plate you can place a wire in the top of a large drop, attach another wire to the metal plate and connect the two leads through a battery and a small light bulb. The vapor layer below the drop is not electrically conductive, and the circuit remains open (light bulb unlighted) unless the drop happens to break through the vapor layer and momentarily touch the metal plate.

I did not investigate how the Leidenfrost point for distilled water varied with the type and nature of the metal. How the type of metal affects the Leidenfrost point is not clear in the writing on the subject. You might want to investigate that aspect yourself. The surface roughness of the metal plate is definitely a factor in the persistence of the drop; a rough surface generally has drops of shorter duration. Spraying the metal surface with a vegetable shortening such as Pam (a commercially available cooking spray used on pots and pans to decrease the sticking of food) has no apparent effect on either the Leidenfrost point or the persistence of the drops. You might try other surface coatings such as Crisco (another vegetable shortening) or a Teflon spray.

You may use many common household liquids in place of water in investigating the Leidenfrost phenomenon, but first you should eliminate any that are flammable or likely to explode near an open fire or on a hot surface. Karo corn syrup beads into Leidenfrost drops at temperatures near 300 degrees C.

Instead of floating quietly until it vanishes, as a water drop does, the syrup drop puffs up into a large hollow sphere and then cooks into a mess on the metal. Save this liquid for last or you will have to clean the metal. White vinegar has a Leidenfrost point near 225 degrees C. My sample of a commercial white vinegar had been reduced with water to an acetic value of 4 percent. Nearly all the Leidenfrost vinegar drops popped out of existence at the end of their dance, perhaps because only the water was left then and they suddenly heated. I think it would be fun to try other fluids that are combinations of liquids with different boiling points. Does the Leidenfrost point become less pronounced with such combination liquids? Do the drops last longer than drops of the individual pure liquids?

Trying to float Leidenfrost drops on surfaces of a superheated fluid would also be fun because the vapor released by the superheated fluid would help to support the drop. Work has already been done on floating water drops on superheated water, the drops in such cases being called globules or boules [see "The Amateur Scientist," SCIENTIFIC AMERICAN, April, 1974]. In the Leidenfrost case you would have to superheat another fluid so that its temperature was then at or above the Leidenfrost point for the drops.

The most difficult fluid I worked with was vodka. The Leidenfrost point was fairly low, about 160 degrees C., and thus easy to reach. Although the duration of the drops was typically less than that of water drops because of the increased evaporation rate, the durations were still measurable. The trouble was that as I waited for the metal plate to stabilize in temperature between each set of drops I kept sampling the fluid.

Leidenfrost drops are not limited to fluids that must be heated above room temperature to bead up and skim about. Liquid nitrogen sprinkled on the floor behaves in the same way that drops of water do on a hot skillet. The drops of liquid nitrogen last from five to 15 seconds. Again, the surface on which the drop is suspended provides heat to vaporize the bottom of the drop, and that vapor layer then supports the remaining drop.

An inverse Leidenfrost effect occurs if you heat small bits of metal a millimeter or less in diameter to a temperature near 1,000 degrees C. in a flame and then sprinkle them on a water surface. The metal pieces float over the surface while they are supported on a vapor layer emitted by the water. You can use small metal chips from a machine shop, but the irregular shapes limit the time of the skimming because as the chips tumble the vapor layer is periodically spoiled. If you use small metallic spheres (I cannot find any), the pieces should last longer. Since only one brief paper has been published on the inverse Leidenfrost effect, you might want to investigate the duration of the skimmers as a function of the temperature of the main body of the water and the size of the skimmer.

Still another demonstration of the same principle, that a thin layer of water vaporizes when it is suddenly heated and then briefly provides protection against further heating, is the old sideshow stunt where the performer dips his wet fingers into a molten metal. I have done the demonstration with molten lead, dipping from one to five fingers to a depth of about two centimeters into lead at approximately 400 to 500 degrees C. (You can cook meat at 100 degrees C., and I stuck a piece of meat- my finger-into a fluid four or five times hotter.) Let me warn you, however, about this demonstration before I explain it. It can be dangerous. If the lead is near its melting point of 328 degrees C., it might suddenly solidify around the finger with disastrous results. If the lead is spilled or splashed, it can cause serious burns. The demonstration with hot lead is clearly not one to be performed by youngsters.

The protection to the finger in the molten lead comes from the water on the initially wet skin. On sudden contact with molten lead, the water on the skin vaporizes to create a brief protective sheath around the finger. One or two seconds are then needed to conduct and radiate sufficient heat through the vapor layer to the skin to bring the skin to a noticeable temperature. I find little sensation of heat in a very quick plunge. Of course I do not dillydally while my finger is in the lead. With an initially dry finger I cannot do more than quickly touch the molten lead.

I can simulate this demonstration by wrapping a thermocouple end with Teflon tape and then dipping the end into the molten lead while I monitor the temperature rise in the thermocouple. The analogy with a human finger is not too close, but the effect of water on the thermocouple can be appreciated. With the lead at 430 degrees C., the thermocouple takes about six seconds to come to the temperature of the lead if the tape is dry and eight or nine seconds if the tape is wet. This result means that in the first second there would be a difference of some 18 to 24 degrees C. in the temperature of the thermocouple in the wet and dry states. I attempted to simulate the demonstration more closely by burying a thermocouple end just under the skin of a fresh hot dog before dipping the hot dog into the molten lead. The thermocouple hardly rose in temperature, but within four or five seconds the skin of the hot dog was a black mess. Apparently the burning is confined to the first half millimeter on the surface of the hot dog (and would be for a finger if it were left n and the conduction of heat to the next several millimeters is relatively slow.

When a candymaker dips his wet finger into the melted candy to test the temperature, the same kind of vapor protection forms around his finger. If a finger is dipped into liquid nitrogen, the protecting vapor is created by the nitrogen bath rather than by any moisture on the skin, but the principle is the same.

Of all the demonstrations of the Leidenfrost effect, the most remarkable to me is when people walk on hot coals or lava. The Guinness Book of World Records describes a 25-foot walk over coals measured to be at 1,200 degrees F. Unverified stories of longer walks constantly emerge from the South Pacific and the Far East. In none of these cases is there any reason to invoke unusual powers. Although somehow shutting off pain information to the brain, having a deep-seated religious faith or just being dumb enough to try the stunt might help, the primary protection to the walker's feet comes from the natural moisture on them. Each step places parts of a foot in contact with the coals, and moisture at those places partially vaporizes to give momentary protection. Sweating between footfalls can replenish some of the moisture, but eventually most of it is depleted and the foot begins to warm up perceptibly. The walk usually ends then unless the walker has an unusual tolerance for pain. A thick layer of ash and heavily calloused feet might also lengthen the walk somewhat, but running does not help because the feet are then slammed down into the coals.

Having always been amazed by stories about people walking on hot coals, and having now become a firm believer in the Leidenfrost effect, I set up a five foot bed of hot wood coals for such a walk. I suddenly found it is remarkably easy to believe in physics when it is on paper but remarkably hard to believe in it when the safety of one's own feet is at stake. As a matter of fact, walking on hot coals would be such a supreme test of one's true belief in what one had learned that I have suggested graduate schools might substitute it for the Ph.D. examination in physics. On one side of a pit of red-hot coals would be a line of fresh Ph.D. candidates. On the other side would be the department chairman with a handful of certificates. If a graduate student really believed in physics, he would stride across the coals without hesitation. If, however, he had any serious doubt, he would not be able to bring himself to do it.

I had to try it myself. Clutching my faded copy of Halliday and Resnick's Physics in one hand, I strode over the five feet of hot coals. Apparently I am a true believer in physics. I have to report, however, that my feet did get a bit hot. Oh well, I am almost a true believer.

Bibliography

VIBRATIONS OF EVAPORATTNG LIQUID DROPS. Norman J. Holter and Wilford R. Glasscock in The Journal of the Acoustical Society of America, Vol. 24, No.6, pages 682-686; November, 1952.

THE LEIDENFROST PHENOMENON: FILM BOILING OF LIQUID DROPLETS ON A FLAT PLATE. B. S. Gottfried, C. J. Lee and K. J. Bell in International Journal of Heat and Mass Transfer, Vol. 9, No.11, pages 1167- 1187; November, 1966.

ON THE FIXATION OF WATER IN DIVERSE FIRE. Johann G. Leidenfrost, translated by Carolyn Wares in International Journal of Heat and Mass Transfer, Vol. 9, No. 11, pages 1153-1166; November, 1966.

The Society for Amateur Scientists (SAS) is a nonprofit research and educational organization dedicated to helping people enrich their lives by following their passion to take part in scientific adventures of all kinds.