flame safety lamp, is extinguished when the oxygen content falls below 17 percent, but the carbide lamp continues to burn until oxygen is below 13 percent. Although men seldom lose consciousness. until there is considerably less than 13 percent of oxygen present, no one should try to enter or remain in an atmosphere in which a flame safety lamp or a candle will not burn.

The use of an open light or flame in any place where combustible gases are present or suspected is very hazardous. Numerous disasters have resulted in mines, with great loss of life and heavy property damage, because of gas ignitions and explosions. Many mine. fires have also occurred because combustible material such as timber has been ignited by flame from candles, carbide lamps, or matches used by smokers.

ELECTRIC METHANE DETECTORS

Detectors of an electrical type determine the amount of explosive gas as well as indicate its presence. In general, they utilize the principle that the electrical resistance of a wire filament increases with an increase of temperature; the increase of temperature results from combustion of the flammable gas on the surface of the filament, which is heated by an electric current. The filament is connected with fixed and variable resistances in an electric circuit known as a Wheatstone bridge. Any change in the resistance of the filament unbalances the bridge circuit and causes the current to flow through a meter; the amount of movement of the meter's pointer over a graduated scale indicates directly the percentage of combustible gas (methane) in the air.

CARBON MONOXIDE DETECTORS

Because of the dangerous characteristics of carbon monoxide, means for its detection are vitally important, especially following mine fires and explosions. The methods or devices generally used are:

1. Iodine pentoxide (hoolamite) detector.

2. Canaries and small animals.

3. Pyrotannic acid method.

4. Sensitive catalyst (hopcalite) indicator or continuous recorder. 5. Analysis of air sample.

6. Ampoule-type detector.

The hoolamite detector has replaced, to a large extent, the use of canaries and small animals. Hoolamite is a mixture of iodine pentoxide and fuming sulfuric acid on granular pumice stone. In contact with carbon monoxide iodine is liberated, changing the originally white granules to bluish green of increasing depths, violet, brown, or black, depending upon the concentration of carbon monoxide. Air is drawn into the detector by squeezing an aspirator bulb and passes through a tube containing hoolamite. The resulting color, if any appears, is compared with the colors in a color-comparison tube to determine the concentration of carbon monoxide present. The comparison scale is devised for a range of 0.10 to 1.0 percent of carbon monoxide by making a test with 10 squeezes; however, if readings are made after 20 squeezes or 5 squeezes and corrections are made in the percentage indicated on the scale, the range may be extended from 0.05 to 2.00 percent.

Small birds and animals, especially canaries and mice, have long been recognized as a means of detecting carbon monoxide; however, small animals or birds are not safe indicators of oxygen-deficient atmosphere because some small animals thrive in an atmosphere whose oxygen content would kill human beings. Moreover, some animals are markedly more resistant to carbon monoxide than others; tests show also that there is wide variation in the susceptibility of individual canaries to carbon monoxide, and undue dependence cannot be placed on an individual bird unless its characteristics are known from previous tests in known concentrations of carbon monoxide gas. The use of small animals or birds to detect carbon monoxide has one distinct advantage over chemical or mechanical devices now available; the small animal or bird automatically and promptly gives warning upon entrance into a toxic carbon monoxide atmosphere, while present-day devices do not give warning until or unless the device is being operated by the person or persons who need the warning.

The pyrotannic acid method was developed to determine quantitatively and accurately low percentages of carbon monoxide in the blood. The basis of this method is that a light brownish-gray suspension is formed in a few minutes when normal blood, diluted with water, is treated with a solution of pyrogallic and tannic acids, whereas light carmine suspensions are formed in blood having carbon monoxide in combination with the haemoglobin (red coloring matter). When compared with color standards of known concentration, the intensity of the carmine suspension gives the saturation of the haemoglobin in the blood being tested. The range of the method is 0.01 to 0.20 percent carbon monoxide. Permanent standards prepared from pigments are supplied with the apparatus, as those made from blood deteriorate in time and would result in inaccurate determinations. The pyrotannic acid method is adaptable for determination of carbon monoxide in both blood and air by a slight difference in procedure (fig. 25).

42

A portable carbon monoxide indicator has been developed by which percentages of the gas may be read directly on a calibrated meter. The detectors are made in both hand-operated and motor-driven types, the latter using 6-volt current from storage batteries or power circuits. The range of the hand-operated type is 0.01 to 0.15 percent; the motor-driven type may have two scales, one ranging from 0.00 to 0.01 percent and the other from 0.00 to 0.10 percent. The air sample is drawn through the apparatus by a small pump and passes at a constant rate through a drying canister and then through a catalyst cell. The catalyst, hopcalite, is a specially prepared mixture of manganese dioxide and copper oxide which causes the carbon monoxide and oxygen to combine chemically. The reaction is accompanied by liberation of heat, the quantity liberated depending upon the concentration of carbon monoxide. The heat produced is measured by thermocouples and registered on a meter calibrated directly in percentages of CO. The continuous carbon monoxide recorder is a nonportable apparatus previously developed by the Bureau of Mines. It operates on the same principle as the portable types and is used for determining and recording the CO content of

42 Berger, L. B., and Schrenk, H. H., Methods for the Detection and Determination of Carbon Monoxide: Bureau of Mines Tech. Paper 582, 1938, p. 16.

the atmospheres in vehicular tunnels, industrial plants, and other places where there is danger of contamination. Recording indicators of this type have been installed on compressed-air lines at several

FIGURE 25.-Pyrotannic-type carbon monoxide detection equipment.

mines in the Lake Superior district and perhaps at some other mines, to avoid excessive carbon monoxide content in the mine air lines due to excessive oil and heat or other contributing causes. When the carbon monoxide present exceeds the set limit, power to the compressor is automatically cut off. The sensitivity of the carbon monoxide recorder may be varied to suit the conditions under which it is to be

used. A sensitivity to 0.0001 percent of carbon monoxide in air is said to be attainable. The recorder is adapted to the determination of relatively low concentrations of carbon monoxide in air; however, relatively high concentrations may be determined by diluting the sample with air and applying a suitable correction to the recorded concentration.

Carbon monoxide in mine air can be determined accurately by collecting and analyzing an air sample in one of several types of volumetric gas-analysis apparatus. This is done by taking a measured amount of an air sample, absorbing the carbon monoxide in a chemical, and measuring the resulting contraction of the sample. It may also be determined by combustion in a copper oxide tube heated to 300° C. The carbon monoxide is oxidized to carbon dioxide, which is absorbed in an alkaline solution and the contraction measured.

The ampoule-type detector is a cotton-covered, light glass tube filled with a solution of palladium chloride in a water-acetone mixture. When crushed, the ampoule is exposed for 10 minutes in the atmosphere to be tested; if carbon monoxide is present the color of the solution changes from yellowish black to black. Comparison with a color chart gives an estimate of the concentration of the gas present. The range is 0.02 to 0.10 percent.

DETERMINATION OF OXIDES OF NITROGEN

Oxides of nitrogen may be detected by the characteristic "burnedpowder" odor or by its reddish color in high concentration. For detection of low concentrations and determination of percentages of oxides of nitrogen, samples must be collected in specially prepared vacuum sampling bottles. These bottles contain 10 milliliters of of dilute sulfuric acid plus a few drops of hydrogen peroxide. The samples are taken by breaking the glass tip from the bottle and resealing with a rubber cap. The sample is analyzed in a laboratory such as the gas-analysis laboratory of the Bureau of Mines at Pittsburgh (fig. 27).

DETERMINATION OF CARBON DIOXIDE

Determination of carbon dioxide and low concentrations of oxygen in air may be made by analyzing samples on a gas-analysis apparatus such as the Orsat discussed in a following page under that heading (fig. 26). Extinction of the flame of an oil-fed lamp or a candle or a flame safety lamp will, of course, indicate low oxygen.

HYDROGEN SULFIDE DETECTORS

Hydrogen sulfide is detected at very low concentrations by the odor of rotten eggs, but this test will not protect against dangerous exposure, since the sense of smell is lost in 2 to 5 minutes of exposure to 0.010 to 0.015 percent of the gas. A chemical detector has been devised which indicates the existence of hydrogen sulfide in the range of 0.0025 to 0.05 percent; it is based on the reaction of silver cyanide in the presence of hydrogen sulfide in turning white granules dark gray.43

48 Forbes, J. J., and Grove, G. W., Mine Gases and Methods for Their Detection: Bureau of Mines Miners' Circ. 33, 1938, pp. 78-80.

AIR TESTING AND SAMPLING

The only available method of determining the quality of air quickly is to make a number of tests for the individual constituents; a candle, a flame safety lamp, or any oil-fed flame is used to indicate whether the oxygen is greater or less than 16 or 17 percent.

Tests for possible contaminating gases can be made as previously described. Accurate knowledge of the quality of mine air can be gained best through samples taken underground and analyzed in a laboratory.

A sample can be taken by filling a bottle with water and emptying it at the place where the sample is desired; the air rushes in to displace the water. The bottle is then sealed tightly and sent to the laboratory. For greater convenience, a metal or glass container with two stopcocks may be used. The water-displacement method involves small inaccuracies due to the solubility of some gases in water; samples of dry gas can be secured by simply aspirating air through a bulb aspirator into a bottle, continuing the aspirating long enough to displace all of the original air in the bottle. The Federal Bureau of Mines uses glass bulbs that have been evacuated of air by a pump and the tip of the sample container sealed in a gas flame. By taking a vacuum tube into the place where a sample of the air is desired and breaking the tip of the tube, the air fills the vacuum and the small aperture is sealed with wax. To sample behind a mine seal, a pipe or glass tube is inserted through a pipe built into the seal or through some other opening in the seals and connected by hose to the sample container so the gas may flow or be aspirated into the sample container; precautions must be taken to obtain a true or representative sample. A device has been perfected by the Bureau for taking samples in deep shafts or boreholes. A vacuum tube is placed in a metal shell fitted with electrically operated attachments for breaking off the tip and sealing the tube. The shell is lowered or pushed to the sampling point and there operated by electric circuit through the three-wire cable to which it is attached.44

45 ORSAT APPARATUS

The Orsat is a gas-analysis apparatus for determining percentages of carbon dioxide, oxygen, carbon monoxide, and methane in air samples. Briefly, a known volume of the gases to be tested is run through potassium or sodium hydroxide solution, which absorbs the carbon dioxide in the sample. From the difference in the resulting volume of the gases the percentage of carbon dioxide can be calculated. Similarly, the oxygen and carbon monoxide content of the gases are measured by passing the gas sample through pyrogallic acid solution and then through cuprous chloride solution. The combustible gases, such as methare, are determined by igniting the gases with an electrically heated platinum filament and calculating the percentage of methane from the change in volume. The field-type Orsat apparatus is accurate ordinarily within about 0.2 percent (fig. 26); for more

44 Fene, W. J., Novel Device for Collecting Air Samples in Inaccessible Places: Bureau of Mines Inf. Circ. 7122, 1940, 2 pp.

45 Yant, W. P., and Berger, L. B., Sampling Mine Gases and Use of the Bureau of Mines Portable Orsat Apparatus in Their Analysis: Bureau of Mines Miners' Circ. 34, 1936, 90 pp.