TP B3 2-NRLF 3E7 TECHNICAL GAS ANALYSIS BATES LIBRARY UNIVERSITY OF CALIFORNIA OF MRS. MARTHA E. HALLIDIE. Class rro^ws.s^o*IF ttbe UnJmsttial as Series. VOLUME I. TECHNICAL GAS ANALYSIS, BY FRANK H. BATES. PHILADELPHIA : PHILADELPHIA BOOK COMPANY, PRACTICAL, SCIENTIFIC AND TECHNICAL BOOKS, 15 SOUTH NINTH STREET. 1901. \ COPYRIGHT BY FRANK H. BATES, 1901. s Printed at the WlCKERSHAM PRINTING HOUSE, 53 and 55 North Queen St., Lancaster, Pa., U. S. A. PREFACE TO THE INDUSTRIAL GAS SERIES. VOLUME I.-TECHNICAL GAS ANALYSIS. SOMK two years back the author of this work com- menced writing articles for the Journal of Electricity under the title of Industrial Gas. Since that time these have appeared almost continuously and have been the subject of favorable comment. Many requests for back numbers, now out of print, have encouraged the author to present the series in book form, hoping that they may prove of value to the engineering fraternity, by affording a guide to systematic investigation of economies in Power Plants. The volumes will probably be issued in the following order : i. Technical Gas Analysis. 2 . Calorimetry . 3. Fuel Economies. 4. The Gas Engine and its Economies. 5. The Economical Production of Gas for Power Uses. Technical Gas Analysis is intended to serve as a guide in the selection of the best method and proper apparatus for use in making analyses of gases of varying composi- tion. Gas analysis forms an important factor in the work of many industrial establishments, enabling the operator in many instances to keep in constant touch with the var- ious changes arising. The value of gas analysis is instanced in the case of iron, steel and glass works, the former having to deal with blast furnace and the latter iv PREFACE. with producer gases ; in the case of steam plants where flue gas analysis proves an aid in the maintenance of furnace and boiler efficiency ; in connection with gas engine test- ing, where efficiencies are determined and losses traced by the analysis of the gases admitted and exhausted ; and finally in . the gas works, where a uniformity of product is a necessary condition to successful manufacture. The book being written in a plain style, well illustrated by both sketches and examples and with a simple treat- ment of volume contractions and chemical reactions, make it valuable to the engineer who has no knowledge of chemistry as well as to the expert analyst. In the preparation of the following chapters the present literature on the subject has been freely consulted, and the author would acknowledge his indebtedness to ' ' Methods of Gas Analysis," by Hempel, in the treatment of the Hempel apparatus, in which the descriptive matter is largely as in the original, but made to conform to the general plan of this work. FRANK H. BATES. San Francisco, January 31, CONTENTS. CHAPTER I. COLLECTION OF SAMPLES. PAGE The plain tube with straight ends to be closed by a blow- piP e I The collection tube with glass stop-cocks; Keeping the glass stop-cocks lubricated; Cases for collection tubes; Apparatus for obtaining average samples of a gas, being generated, when the composition varies 2 Suitable piping for conducting gases from furnaces; Porcelain, glass, platinum, lead or rubber tubing; Currents in flue- ways; Getting a sample representing a fair average ........ 5 Collecting mine gases; Packing used iri making joints . . . 6 CHAPTER II. METHODS OF GAS ANALYSIS. Apparatus to be used for the analysis of furnace or chimney gas, producer gas, illuminating gas ; A make-shift apparatus. 7 Analysis with this apparatus n CHAPTER III. THE ORSAT APPARATUS. Examination of furnace or chimney gases by means of the Orsat apparatus; Description of the Orsat apparatus ... 17 Manipulation 19 Determination of carbon dioxide 21 Of oxygen 23 Of carbon monoxide; Of nitrogen; Special hints; precautions as to (order of) using the reagents, draining the walls of the burette, time required for an analysis, absorbing capacities of the reagents; limit of error with this apparatus . ...... 22 (v) Vi CONTENTS CHAPTER IV. THE ELLIOTT APPARATUS. PAGE Principle of this apparatus; Description 24 Determination of carbon dioxide 29 Of illuminants; Of oxygen 30 Of carbon monoxide; Treatment of the residue 31 Chemical reactions involved 33 Volume contraction . 34 Calculation of the amounts of methane and hydrogen from the combustion with oxygen; Determination of nitrogen by difference; Example of an analysis of coal gas . . 35 Table of analyses showing range of work to which this appar- atus is adapted 38 CHAPTER V. THE HEMPEN APPARATUS. When to use this apparatus; Description of the 'simple gas burette 39 Operation of the gas burette 41 The simple absorption pipette 43 The simple absorption pipette for solid and liquid reagents . 44 The double absorption pipette; The double absorption pipette for solid and liquid reagents 45 The ethylene pipette 48 The explosion pipette 49 The hydrogen generator 50 The explosion pipette with electrodes for the decomposition of water . .52 The combustion of methane and hydrogen in a gas mixture without explosion; Operation of the Hempel apparatus ... 53 Example of an analysis of coal gas 57 Special schemes; The fractional combustion of hydrogen .'..'. *6l The absorption of oxygen by phosphorus 66 Scheme for the analysis of coal gas by which the traces of carbon monoxide remaining from the cuprous chloride ab- sorption and any ethane present may be determined .... 67 CONTENTS Vil CHAPTER VI. MEASUREMENT OF GASES. PAGE Influence of pressure on volume .72 Boyle's law; Example illustrating Boyle's law 75 Influence of temperature on volume 76 Absolute zero; To obtain absolute temperature; Charles' law; Example of Charles' law 77 Vapor tension of liquids 79 Effect of vapor tension on volume 81 Example of combined corrections for pressure, temperature and vapor tension; Measurement of gases over liquids other than water; Measurement of dry gases over mercury; The expan- sion of mercury 82 CHAPTER VIL PROPERTIES OF GASES. PREPARATION OF REAGENTS EMPLOYED. Properties of carbon dioxide; Of ethylene . 84 Of oxygen 85 Of carbon monoxide 86 Of nitrogen; Of hydrogen 87 Of methane ... 88 Preparation of potassium hydroxide; Of sodium hydroxide; Of barium hydroxide; Of bromine water; Of potassium pyro- gallate * 89 Of cuprous chloride 90 Table I. Tension of water-vapor in millimeters of mercury for different temperatures, also the weights in grams of the vapor con- tained in a cubic meter of air when saturated 91 Table II. French measure 93 Table III. Tension of mercury vapor 93 INDEX. **, .... 94 CHAPTER I. COLLECTION OF SAMPLES. The method to adopt in collecting a gas sample is de- pendent upon the conditions under which one is operating. The first point to be observed is to use such means as will give a sample representing the average gas. Often- times in the gases to be sampled there exist currents whose composition vary considerably, this being especially true in the case of chemical processes and furnace gases. In such instances some ingenuity must be exercised to meet the special conditions. FIGURE 1. For the retention of a gas sampel, the collection tube shown in Figure i will serve. The ends are made quite long and drawn down to about two millimeters ( T V inch) internal diameter. They should be of about 200 cubic centimeters capacity. It will be found very convenient for transportation to have a box, provided with separate FlGURE 2. compartments, to hold about a dozen of these tubes. To take a sample it is simply necessary to connect, by a short rubber tube, at one end to the gas supply, allowing the gas to flow through (the gas being under pressure) till certain that all air has been expelled and that only pure 2 TECHNICAL GAS ANALYSIS. gas remains; then almost stopping the flow to relieve the tube of pressure, seal near the end by holding a lighted candle to it, or by means of an improvised blowpipe made by pointing and nearly closing the end of a glass tube, and then bending near the point, cutting off at a con- venient length. The collection tube will serve for a number of samples before becoming so short as to make it necessary to weld on new ends. A somewhat handier tube is shown in Figure 2, hav- ing glass stop-cocks, which obviates the necessity of seal- ing by heat. Although there is but little liability to leakage with these tubes when the ground glass stop-cocks fit properly and are kept well greased, yet the writer has made it a practice, when obliged to retain the sample some time before making the analyses, to cover all joints and the open ends with sealing wax, which renders leakage impossible. A case of these collection tubes is shown in Figure 3. Instead of water for the confining liquid, mercury is often used when the previous saturation of the water is inconvenient or for other reasons. It is sometimes desirable to obtain a sample of a gas passing through a conduit of some kind, which may rep- resent the average gas for a certain period of time. For such purpose the apparatus illustrated by Figure 4 is well suited. The tanks A and B are constructed of gal- vanized iron, and measure some 12 inches in diameter and 1 8 inches in length, one being slightly longer than the other. Bottoms and tops of the tanks should be slightly curved downwards and upwards respectively, to render it possible to secure complete drainage. At about the center of both top and bottom is inserted a ^-inch gas pipe nipple, the j oint being strongly soldered, on which are screwed pet-cocks for the tops and pipe connections to the outside for the bottoms, there being a rim sufficient to OF SAMPLES. raise the tank above the piping. Pet-cocks are put on the lower pipes at a convenient distance from the side. On the tops there are also auxiliary nipples with pet-cocks attached, to allow of rapid filling. Operation, The upper (larger) tank is first filled with water and then connected to the gas source by suitable means, while on a shelf below, and with its auxiliary pet-cock connected by rubber tubing to the outlet from FIGURE 3. the upper tank, is placed the smaller one. By turning on the globe valve a (on the pipe leading to the gas source) and opening pet-cocks , d, e and f, and with pet-cocks c and g closed, allow the water to run from the tank A into TECHNICAL GAS ANALYSIS. FIGURE 4. COLLECTION OF SAMPLES. 5 tank B, displacing the air from the latter and causing it to issue from the open pet-cock f. As the water recedes from the tank A, a partial vacuum is created which causes the gas to rush in. The receding water, mingling with the inrushing gas, becomes thoroughly saturated and is used to fill the lower tank B. On the overflow of water from the pet-cock of .the tank B, pet-cocks f and e are closed, and the tank is disconnected from the upper one (which is removed), and connected up to the gas source by attaching the rubber tube h to the pet-cocky". With a long rubber tube attached to the outlet pet-cock g, to serve as a drain to the waste barrel, open the pet-cocks/ and g, causing the water to run from g and draw in the gas through f. By adjusting these pet-cocks, the time taken to fill the tank may be lengthened or shortened, so that an average sample of a run of a certain duration may be obtained. The previous saturation of the water used prevents the washing out of the soluble constituents of the gas. The kind of piping to use for running into the furnace or holder will depend upon the temperature and nature of the gases. With a high temperature iron tubes with ample water-jacketing should be employed.* Porcelain, glass or platinum tubes may also be used. Where the gases are strongly acid and the time occupied in making the collection is long, only glass tubes should be employed ; they will not be affected by temperatures within 600 C. At temperatures lower than 300 C. , lead piping will be found serviceable owing to its great flexibility. Since vulcanized rubber tends to absorb the gases, rubber tubing of any length should be avoided. In a single flue, way or conduit from which the sample is being drawn, there are often found currents of some *Iron piping, if raised to a temperature of redness, may alter the gas com- position, since the oxygen of the iron oxide (iron rust) will combine with the unsatisfied carbon monoxide to form carbon dioxide \ 6 TECHNICAL GAS ANALYSIS. strength and of quite different composition, necessitating, therefore, the use of several branches of piping to pene- trate equally all parts and thus afford a fair average. Where the current in the flue way is strong, a suction pump or siphon may be employed to advantage. For collecting gases from mines, etc., the form of tank shown in Figure 4 may also be used. For this purpose fill with water and place in an upright position where it is desired to sample the gas. By opening both lower and upper pet-cocks cause the gas to replace the water. As soon as the tank empties of water, close all pet-cocks en* closing the sample. For packing used in making tight joints, asbestos is probably the best, although, under certain conditions, putty, plaster of Paris, or wet waste may be used. CHAPTER II. METHODS OF GAS ANALYSIS. Apparatus. For furnace and chimney gases, consisting principally of carbon * dioxide, carbon monoxide, oxygen and nitrogen, by far the most convenient apparatus is the Orsat, (Fisher's modification) arranged in a portable case. For furnace or chimney gases, producer gas, and illu- minating gas, where great accuracy may be sacrificed to rapidity of operation and convenience, Elliot's latest mod- ification with explosion burette, half size, and arranged in a traveling case, is suitable. For accurate laboratory analysis, use Hempel's appa- ratus. It sometimes happens that one is so situated as to make it desirable to obtain a rough estimate, or an approxima- tion, of the constituents of a gas, there being no standard apparatus at hand. In such event, the tubes illustrated in Figure 5 can be quickly made up from plain glass tub- ing, of about 20 millimeters* (^ inch) internal diameter. First bend and draw out the projecting end d of the tube A, retaining the conical form and making the external diameter about 4.7 millimeters ( 3 / 16 inch). Make the end corrugated so as to permit a rubber tube to be se- curely fastened thereon by wire, and have it project from the main tube some 38 millimeters (1^2 inches) thus providing for the wooden stand e y which is to be slotted to allow of its receiving the tube, and of the latter being secured thereto by a clasp. This finished, cut the glass tube off to a length of about 500 millimeters (20 inches), and shape the end c down to an internal diameter * To reduce millimeters to inches, multiply by .03937, or divide by 25.4. TECHNICAL, GAS ANALYSIS. of about 0.8 millimeters (V 32 inch). Make this end 32 millimeters (1.25 inches) in length, and form its outside to correspond to the end d for a rubber tube connection. Next make up the tube A' of similar diameter but one- third longer. The projection d' is identical with d, but LEVEL BOTTLE MfASURING BUfiSTTE- \ f- PUBBER c CONNECTION} COLLECTION TUBE- FIGURE 5. at c f the tube, instead of being conically drawn out, as at c, is enlarged to a funnel shape, forming a mouth for the reception of liquids. Having this complete, fasten a piece of seamless rubber tubing on the end c of the tube A by wire ligatures, leav- METHODS OF GAS ANALYSIS. tf ing some 40 millimeters (1.5 inches) of the tube length above the glass for the purpose of inserting a connecting piece thereon. Fasten the end of a piece of rubber tub- ing one-half meter (19.6 inches) long to the projection d, and a similar piece to the end d\ connecting the two free ends by a piece of glass tubing, thereby obviating the necessity of removing the wired ends at d and d* when cleaning the tubes. The tube A, which is termed the measuring burette, is thus connected to tube A', termed the level-tube. To calibrate A, proceed as follows : First ascertain the amount of water contained by, say, 325 millimeters of the length of the tube A'. In this instance, as the internal diameter of the tubing is 20 millimeters, the quantity would be 102.102 cubic centimeters,* or each millimeter of the length of the tube would contain about 0.31416 cubic centimeters. Now fasten a thin strip of paper, scaled to millimeters, and about 400 millimeters in length, on the outside surface of the tube A', to serve temporarily. Place a pinch-cock, f y on the rubber tube close to the end c, and place another pinch-cock, g, in a similar manner on the rubber tubing close to the end d. With both pinch-cocks open, fill the two tubes with water, having the water overflow at the end of the rubber tube c, the capillary tubing B not being attached as yet ; then close the pinch-cocks f and g and partially empty the level- tube A' so that the level of the water is a little above the lowest reading of the paper scale. Open the pinch- cock g, then by raising the level-tube A' so that the water in it will be on the same level as the water in the burette A (which is at the very top or end c), the contents of both tubes will be under equal pressure ; /. e. , under equal at- mospheric pressure. Notice at this time the reading of the water level in the level-tube A' on the improvised * To reduce cubic centimeters to cubic inches, divide by 16.383. 10 TECHNICAL GAS ANALYSIS. scale, taking care to read the position of the lowest point of the sharply defined, crescent-shaped meniscus formed at the surface of the water. For the purpose of illustration, consider this reading to be 10 millimeters from the bottom of the scale. Now lower the level tube A', and open the pinch-cock f to admit air so as to allow the water in burette A to drop about 42 millimeters (1.6 inches) from its former level the top of the tube at c. After waiting three minutes for the walls of the tube to drain, take the reading of the water level on the scale on level-tube A', under atmo- spheric pressure, by bringing the level of the water in both tubes to the same height. Suppose the reading to be 20 millimeters, then 20 10 (the former reading) =10 millimeters, and as each millimeter was found to contain 0.31416 cubic centimeters, 10 millimeters would equal 10X0.31416 or 3.1416 cubic centimeters, the amount of air admitted. Keep the level of the liquid in both tubes at the same height and then mark the line of water level on the burette A with a file. Now, in a similar manner as before, admit sufficient air to make a total of 100 cubic centimeters, which in the present case, would correspond to 318.3 mil- limeters in the scale of , the level tube. In doing this no difficulty need be experienced if care is taken to admit very little air as the required amount (100 cubic centi- meters) is approached. The level-tube A' should be raised frequently to make the readings until about the amount is obtained when three minutes should be allowed to drain the water from the walls of the burette before making the final measurement. Now with the liquid in the level-tube A' and in the burette A at the same height, mark with a file the level of the water in burette A. This corresponds to 100 cubic centimeters, measuring from the top at c. Cut a strip of METHODS OF GAS ANALYSIS. 11 paper of a length equal to the distance between the two file marks, and on this paper lay off a scale in proportion, divisioning it into tenths of cubic centimeters, gluing it on to the measuring burette A and shellacking. The scale on the level-tube A', may now be discarded, and there has been thus improvised an apparatus that, in lieu of a bet- ter one, will suffice for rough work. Operation. The pinch-cocks f and g are first opened by simply slipping them off over the ends of the glass tubes c and d, and they remain so. Water is poured into the mouth c 9 of the level-tube A', until it overflows from the short rubber tube on , when the pinch-cocks f and g should both be closed. During the operation of filling with water, the level-tube A' should be held at such a height with respect to the burette A, that it may fill the former about two-thirds full. Now make connection with the collection tube c, containing the gas sample, by means of the capillary tube ,B, which should be as short as pos- sible and bent at right angles at a point about 30 milli- meters from each end. In joining the capillary tube B, first prepare it by filling with water and wiping a little vaseline ovi r the ends c" and h" where the rubber tube connections are to be attached. By means of a small dropping pipette also fill with water the short pieces of rubber tubing in the ends c and h of the burette and col- lection tube respectively, and by slipping one end of the capillary tube into the rubber tube at c and the other end into the end of the rubber tube at h , all air will be ex- cluded. The end k of the collection tube c should be connected by a piece of rubber tubing a meter (39.37 inches) long to the level- bottle D, after which slip a pinch- cock m on the rubber tube near k. The above connections having been made, next place the level-tube A' on the floor with burette A resting on a stand, and raise the level-bottle D to a support arranged 12 TECHNICAL, GAS ANALYSIS. above the collection tube c. Open the pinch-cocks g, / and m and the glass-cock j in the collection tube c, there- by putting the collection tube c and level-bottle D in con- nection. By opening the glass cock i the pressure of the column of water between the collection tube and the level- bottle will force the gas over into the burette A, and in this it will be assisted by the expansion of the water in A, due to the relatively low position of level-tube A'. Pass over about 50 cubic centimeters of gas, close the glass- cock z", and agitate the water in the burette vigorously in order that it may become saturated with the gas. Next open the glass-cock z*, raise the level-tube A' and lower level-bottle D, causing the gas to return to the collection tube. Agitate the collection tube, after which transfer the gas back and forth a few times, thoroughly saturat- ing the water in both the level-bottle and level-tube, to pre- vent the water washing out the soluble constituents, of the gas when making the analysis.* This preliminary operation complete, force all the gas from the burette A by raising the level-tube A' and lower- ing the level-bottle D. Close the pinch-cocks /"and m and the glass-cocks i and /, and disconnect the rubber tubing of the level-bottle D from the collection tube the pinch-cock m serving to prevent the out-flow of water ; also remove capillary B from both the collection tube and burette. Fill the short rubber tube at c and the capillary tube B with water, and after wiping the ends of the latter with vaseline, make connections with a new collection tube, containing sample gas. One is now ready to proceed with the analysis proper. Analysis. By lowering the level-tube and raising the level-bottle, opening pinch-cocks f and m and glass-cocks z" and j, about 103 cubic centimeters of gas is admitted to * With the present apparatus this is almost a useless precaution, but is given to familiarize the reader with principles to be followed in more exact analyses. METHODS OF GAS ANALYSIS. 13 the burette, when pinch-cock /and glass-cocks i and/ are closed, the capillary tube disconnected, and, together with the collection tube and level-bottle, laid aside. Close pinch-cock and place level-tube on the support above, thereby creating a pressure due to the weight of the water column above the water level in the burette. After wait- ing three minutes for the walls of the burette to drain, open pinch-cock g j ust sufficient to allow the lowest point of the meniscus of the water to rise to the 100 cubic cen- timeter mark in the burette, taking precaution to have the eyes on the same level. There is now contained in the burette 100 cubic centimeters of gas under pressure a trifle greater than atmospheric. Open pinch-cock f to the air for but a moment, thus allowing the excess pres- sure to escape. By now opening pinch-cock g and hold- ing the level-tube so that the surface of the liquid therein will be on the same level with the liquid in the burette, thus putting our gas under atmospheric pressure, the level in the burette will be found to coincide with the 100 cubic centimeter mark. In these illustrations, 100 cubic centimeters of gas is taken to simplify matters, as the readings thus give per- centages direct. Determination of Carbon Dioxide. We first determine the quantity of carbon dioxide (chemical formula CO 2 ) by absorption with a reagent.* The reagent used as an absorbent may be potassium hydroxide, sodium hydroxide, or barium hydroxide. The first is to be preferred, owing to its quick action, but the last is employed to advantage when the quantity of carbon dioxide is very small. To treat the gas with the reagent, lower the level-tube to expand the gas in the burette until the level of the water * Most of the constituents of coal gas may be removed or absorbed by cer- tain liquid reagents, there being exceptions, however, such as nydrogen, methane, nrtrogen, etc., which are removed by combustion. 14 TECHNICAL, GAS ANALYSIS. lowers almost to the top of foot-piece e ; then close pinch- cock gi and empty the level-tube and tubing of water and then partially replace it with the absorbent. Raise the level-tube with the left hand as high as the rubber tubing permits, when, by opening pinch-cock g> a considerable quantity, diluted with the water left in the rubber tubing, will be forced into the burette. Lowering and raising the level-tube will thoroughly mix the reagent and the water and assist in the absorption of the carbon dioxide. By closing pinch-cock g and agitating the burette, all the gas will be brought into intimate contact with the absorbent. Note the readings from time to time, and when no further diminution of volume takes place, open pinch-cock g, and after waiting three minutes for the walls of the burette to drain, measure the gas volume under atmospheric pressure by bringing the level of the liquid in both tubes to the same height. This reading, subtracted from 100, will give the per cent, by volume of carbon dioxide in the gas. Determination of Illuminants. Again lower the gas in the burette as much as possible, and in doing so, exer- cise the same care as before in order to prevent the liquid in the burette from dropping too low lest the gas gets into the tube and escapes. Then close the pinch-cock g, and empty the level-tube of the reagent, afterwards rinsing out with water. Next, add a little water in which a few drops of hydrochloric acid was previously mixed, to neu- tralize the remaining reagent in the lower portion of the burette. The next determination is that of the illuminants or heavy hydrocarbons. These consist principally of ethy- lene, (formula C 2 H 4 ) the absorbent for which is bro- mine water. The level tube is nearly filled with water and a little bromine is added with a dropping pipette. The gas residue is then treated with this reagent as in the previous case until no further action takes place. In the METHODS OF GAS ANALYSIS. l5 present instance, however, before measuring the amount of absorption, it will be necessary to empty the level-tube of the bromine water, first closing pinch-cock g, and then partially replacing, not with the next reagent, but with potassium hydroxide, with which treat the gas. This is necessary after using bromine, owing to its high tension, or tendency to vaporize. The potassium hydroxide ab- sorbs this vapor, after which the gas may be measured and the true percentage of ethylene may be ascertained by direct reading. Determination of Oxygen. Again lowering the gas in the burette and emptying the level-tube, rinse and parti- ally fill with potassium pyrogallate for the absorption of oxygen, (symbol O). The method of treatment in this case is similar to the preceding, except that no vapors are formed, and, in consequence, treatment with potassium pyrogallate alone suffices. Determination of Carbon Monoxide. The percentage of oxygen determined, empty the level-tube, rinse with water, and add the next absorbent, cuprous chloride, for the absorption of carbon monoxide, (formula CO). This is our last determination with this apparatus. An ex- ample may serve to illustrate the foregoing operations. Partial analysis of Dowson producer gas : The water in level-tube and level-bottle was saturated with the gas. From another collection tube 100 cubic centimeters of the sample gas was taken for analysis. After treating with potassium hydroxide until no fur- ther absorption took place, the gas volume measured 93.43 cubic centimeters; hence, 100 93.43=6.57 cubic centi- meters, the amount of carbon dioxide in 100 cubic cen- timeters of sample gas, or, carbon dioxide is 6.57 per cent. Level-tube emptied, rinsed with water, and a little hydrochloric acid added to neutralize the potassium hy- 16 TECHNICAL, GAS ANALYSIS. droxide. Gas is treated with bromine water until absorp- tion is complete, level-tube then emptied, rinsed, and partially filled with potassium hydroxide to absorb the bromine vapors. Gas volume then measured 93.12 cubic centimeters; hence, 93.43 (last reading) 93.12=0.31 cubic centimeters, the amount of ethylene present in 100 cubic centimeters of sampk gas, or, ethylene is 0.31 per cent. Ivevel-tube emptied and rinsed, and without adding hy- drochloric acid, gas is treated with potassium pyrogallate until no further absorption, and volume then measured 93.09 cubic centimeters; hence, 93.12 (last reading) 93.09=0.03 cubic centimeters, the amount of oxygen present in 100 cubic centimeters of sample gas, or, oxy- gen is 0.03 per cent. Level-tube emptied, rinsed, and gas treated with cup- rous chloride until no further absorption. Gas volume was 68.02 cubic centimeters; hence, 93.09 (last reading) 68.02=25.07 cubic centimeters, the amount of carbon monoxide present in 100 cubic centimeters of sample gas, or, carbon monoxide is 25.07 per cent. Other constituents of the gas could not be determined with this apparatus. They must be determined by com- bustion, as will be shown later. This apparatus would suffice, however, for furnace or chimney gases composed of carbon dioxide, oxygen, carbon monoxide, and nitro- gen. The nitrogen in this case would be found by differ- ence; i. e., by subtracting the sum of the other gases from 100, which would give its percentage. CHAPTER III.-THE ORSAT APPARATUS. Examination of Furnace or Chimney Gases by means of the Or sat Apparatus. With the apparatus already consid- ered, the gas was drawn into a measuring burette and suc- cessively treated by the absorbents, which (in the case of potassium pyrogallate and cuprous chloride), owing to exposure to the oxygen of the air during the treatment, could be used but a few times before their powers of ab- sorption became exhausted. Orsat devised an apparatus whereby the reagents are protected from the air, and in which the gas itself is brought in contact with the absorbents. It is only to be regretted that the scope of the apparatus is so limited. It offers, however, a most convenient means for the deter- mination of carbon dioxide, oxygen, and carbon monoxide by absorption, and nitrogen by difference, these being the gases found as constituents in furnace and chimney analyses. Description. Consists of three double pipettes, B, C, and D, Figure 2, made stationary in the case and con- nected by means of capillary tubing to a measuring burette A, enclosed by a water jacket. Glass stop-cocks K, F, and G close each pipette to the main capillary, and the glass stop-cock H, thereon, affords an inlet for air from the end j. A level-bottle i,, provides a means of transferring the gas. To the rear of the pipettes B, c, and D, and connected by glass tubing, are three similar pipettes, whose ends, s, /, and /', are connected by rubber tubing with a small, flexible rubber bag, which acts as a seal, preventing the absorption of oxygen from the air by the reagents in the pipettes. The water-jacket may, 18 TECHNICAL GAS ANALYSIS. under ordinary conditions, be simply filled with water, v/hich will tend to prevent slight changes of temperature affecting the gas volume. When the apparatus is so sit- uated as to be subject to draughts or sudden changes, FIGURE 6. although this should be avoided when possible con- nections should be made with the water supply, so as to have a circulation cf water in the jacket, which will se- cure a fairly even temperature. Figure 6 shows Fisher's modification of the Orsat, THE ORSAT APPARATUS. 19 which is of small size and arranged for convenience in traveling. Figure 7 shows a slight modification known as Petri- zilka's, with one universal stop-cock, dispensing with the four small ones. Manipulation. The level -bottle I, is filled with pure water, and with stop-cocks E, F, and G closed, and H FIGURE 7. opened, and the measuring burette is partly filled by rais- ing level-bottle i,, forcing air through exit J. By closing stop-cock H and lowering level-bottle i<, the air remaining in the burette and capillary is expanded, so that on opening stop-cock E, the reagent in pipette B will be drawn up to 20 TECHNICAL GAS ANALYSIS. a point just below the connecting rubber M, when stop- cock K is closed and the reagents in pipettes C and D simi- larly raised to corresponding positions. This done, open stop-cock H, and by raising level-bottle L, force all the air from both the burette A and the capillary tubing, displac- ing by water which will overflow from the end of capillary at J, when close stop-cock H. Preliminary to the analysis proper, make connections with a collection tube of sample gas and the end j, of cap- illary tubing of the apparatus, taking same precautions as before, with previous apparatus, to expel all air from the rubber tube and connecting tube, if one is used. Draw in about 50 cubic centimeters of gas (there is con- tained 100 cubic centimeters from the stop-cock H in the capillary tube, to a point marked 100 cubic centimeters near the bottom of the burette, and the graduations are in tenths of cubic centimeters), by lowering level-bottle L. As soon as the gas is admitted, close stop-cock H, and raise and lower the level-bottle to cause the gas to come in through contact with the water, thus saturating the latter with the absorbable constituents of the gas. Open stop- cock H and expel the gas, filling completely the burette and capillary with the saturated water, and on its overflowing at j close the stop-cock H. We are now ready for our sam- ple for analysis, so make connections with a new collection tube and draw in through end j, in a similar manner, a little more than 100 cubic centimeters, say, for instance, 101 cubic centimeters closing stop-cock H immediately on securing that amount. Wait one minute for the walls of the burette to drain, and then close the pinch-cock I on the rubber tube connecting the level-bottle and the burette, close to the latter. By raising the level-bottle L, a pressure is created, due to the height of the column of water, so that on gradually opening the pinch-cock I, the gas is forced up in the burette. With the eyes on the THE ORSAT APPARATUS. 21 level of the 100 cubic centimeter mark on the burette, close the pinch-cock i just as the lowest point of the me- niscus reaches the mark, when we will have ibo cubic centimeters of gas at little more than atmospheric pres- sure. By opening the stop-cock H for but a moment, this excess will escape to the air, leaving us exactly 100 cubic centimeters of gas. By opening pinch-cock .1 and bring- ing the level;bottle I, to such position that the level of the liquid contained corresponds to the level of the liquid in the burette, w r e will find that the level in the burette will be at the 100 cubic centimeter mark, proving the amount of gas contained to be 100 cubic centimeters at atmospheric pressure. Determination of Carbon Dioxide. We may now treat our gas to the first absorbent, potassium l^droxide, which is contained in pipette B, Opening stop-cock E and rais- ing level-bottle i,, forces the reagent down the front and up into the rear pipette, laying bare the contained tubes, wet with the reagent, thus exposing a large absorbing surface. This reagent quickly absorbs the carbon dioxide present in the gas, one passage of the gas into the pipette gener- ally proving sufficient. By raising and lowering the level-bottle a few times all the gas is brought into thor- ough contact with the absorbent, and can then be drawn back into the burette for measurement by lowering the level-bottle I,, closing stop-cock K, when the reagent has ascended to its former position close to the rubber connec- tion. After waiting one minute for the walls of the bu- rette to drain, bring the level of the liquid in level-bottle and burette to the same height, and read the position of the lowest point of the meniscus on the scale, giving the quantity of carbon dioxide absorbed. Again pass the gas into the pipette B, return to burette, (closing stop- cock K as before) and measure, after waiting one minute for burette to drain. See if this reading corresponds to 22 TECHNICAL GAS ANALYSIS. the former, to make certain that the absorption is complete. This reading, subtracted from 100, the total volume of sample gas, gives the percentage of carbon dioxide. Determination of Oxygen. The residue, or gas remain- ing from the last absorption, is now passed into the second pipette c, containing potassium pyrogallate, which absorbs the oxygen. . Before making the final measurement, the operation should be repeated as in the previous case, to make certain that all the oxygen is absorbed. The time required will not exceed three minutes. In making the reading first allow one minute for walls of burette to drain, this being done whenever a measurement is to be made. The reading in this case subtracted from the previous one, gives the percentage of oxygen. Determination of Carbon Monoxide. The residue is now passed into the third pipette D, containing cuprous chlo- ride, which absorbs the carbon monoxide. A little more time should be given to this absorption, which is some- what uncertain ; say, at least five minutes. The resulting reading, subtracted from the previous one (with potassium pyrogallate), gives the percentage of carbon monoxide. Determination of Nitrogen. Adding the percentages of carbon dioxide, oxygen, and carbon monoxide together, and subtracting the total from 100, gives the percentage of nitrogen by difference. Special Hints. In the foregoing operations always pass the gas into the pipettes in the order named, since potas- sium pyrogallate, contained by the second pipette, will absorb both oxygen and carbon dioxide ; and cuprous chloride, of the third pipette, both carbon monoxide and oxygen. Before measuring and after each absorption, wait some stated interval, making it of the same duration through- out, for the walls of the burette to drain of liquid, since otherwise the readings would be in error. THE ORSAT APPARATUS. 23 The total time required will usually be about twenty minutes for the entire analysis. An accurate account of the cubic centimeters of absorp- tion with each reagent should be kept, and by comparing with the absorbing capacities, one will constantly know the power of the reagent, so that he may renew it as its strength becomes taxed. The capillary tubing of the Orsat apparatus, shown in Figure 6, being two millimeters internal diameter, it re- quires 31.8 millimeters of length (1.25 inches), to give one-tenth of a cubic centimeter.. Thus, error due to the reagents not being exactly on the mark, will be slight. To prevent the glass stop-cocks from becoming fast in their sockets, which often causes fracture on attempts at freeing, keep well lubricated with a mixture of one part tallow and three parts vaseline. No trouble need be experienced in the use of this ap- paratus, even by one unfamiliar with chemical work or analysis. CHAPTER IV. THE ELLIOTT APPARATUS. By means of the apparatus already described, it was only possible to analyze a mixture of gases containing carbon dioxide, ethylene (illuminants), oxygen, and car- bon monoxide. When it is desirable to ascertain the composition of a mixture containing hydrogen or methane as found in producer gas, illuminating gas, etc., recourse must be had to apparatus which affords a means of burning these gases with oxygen, a measurement of the products of combustion, together with the contraction, enabling one to compute the volumes burned. With Dr. Elliott's apparatus such a determination may be made very rapidly, although at the expense of great accuracy. Owing to the explosion, over water, of the residual gas (the portion remaining after the absorbable constituents are removed), the error of results may be as great as two per cent., although the total error may generally be kept within one per cent., which will meet the requirements of technical work. In iron and steel works, metallurgical establishments, for glass-ovens, and regenerative furnaces using producer gas, and for the analysis of chimney or flue gases, the * l Elliott ' ' makes possible sufficiently rapid work to enable the operator to keep in constant touch with the various changes taking place. The principal innovation with this apparatus lies in the treatment of the gases by permitting the reagents to spread and run down the sides of a long tube containing the gas mixture, bringing the absorbent in intimate contact with loss of but little time. Description. It consists of an absorption tube A (Figure 4), measuring burette B, and explosion burette c, all grad- uated to 100 cubic centimeters, although of but one-half ( 24- \ THE ELUOTT APPARATUS. 25 the indicated capacity. The two burettes, B and c, are divisioned in tenths of cubic centimeters. The bulbs, at the junction of the tubes A and B with the capillary tub- ing, serve to render the apparatus more compact by lessen- ing the height, and, in the case of the absorption tube A, provides a means of spreading the liquid reagents used. At the lower end of the absorption tube A is inserted, by means of a cork c',* an arm of capillary tubing, which communicates by the three-way stop-cock D' with the level-bottle M, through the arm E' and rubber tubing fast- ened thereon, and affords an exit for the discharge of reagents by way of the arm v. Above the horizontal capillary on the tube A is a stop-cock* B' communicating with the open end A', ground tapering to receive the funnel F' by which the reagents are introduced. The horizontal capillary just below is cut off square at the ends i' and j' to permit of a piece of glass capillary tubing G, being inserted to form a close joint with a short piece of rubber tubing H holding it in place. A similar arrangement is provided for the ends K' and I/, thus connecting the three tubes. The lower ends of the measuring and explosion burettes are drawn out to receive rubber tubing with which to connect with the level-bottles N and P. It is ad- visable to fasten the rubber tubing on the projecting ends by wire ligatures, necessitating the use of three level- bottles, which is preferable, in the writer's opinion, to loose connections and frequent interchange, as is the case when but two are used, as supplied by the makers. Near the lower end of the explosion burette, below the rubber stopper R' of the water-jacket w, projects an arm of capillary tubing u' with a stop-cock T', through which to introduce the air and oxygen (or hydrogen) used in the ccmbustion of the residual gas. Above the rubber *A recent modification, making the absorption tube and capillary of one piece, does away with this cork. 2G TECHNICAL GAS ANALYSIS. stopper Q of the water-jacket, are fused two platinum wires p' for an ignition spark. The water-jacket w serves to prevent undue heating of the burette during combustion, saving time which would otherwise be required for its cooling. As directed in connection with the Orsat, when SEf FIGURE 8. the measuring burette is subject to changes of temperature during an analysis, it should also be water-jacketed. The explosion burette is made of heavy glass and there is con- sequently but little liability to fracture with the exercise of proper care during the explosion. Pperation. Before commencing an analysis, about two THE ELUOTT APPARATUS. 27 liters (approximately two quarts) of water should be sat- urated with the gas under examination, to prevent, as far as possible, the washing out of the soluble constituents of the gas mixture, previous to their absorption by the re- agents. With this apparatus the gas is successively treated with the absorbents, as was done with the impro- vised tubes shown in Figure 5, differing in this respect from the Orsat. The absorptions are made in the absorp- tion tube A (Figure 8), and the subsequent measurements in the burette B. In preparing for the analysis the level- bottles M, N and P are nearly filled with the saturated water, the connections on the horizontal capillary tubing between the ends j' and i', K' and i/ being made, stop- cocks D', H' and M' are opened to give a passage between the level-bottle M and absorption tube, and both through the horizontal capillary and into the measuring and ex- plosion burettes. A rubber tube should be fastened on the end v of the projecting arm from the stop-cock D', to carry off the waste liquids to the drain. Open stop-cock B' and raise all three level-bottles, filling the entire apparatus with water, thereby expelling all air. Close three-way cock M' so that the run is just closed to the capillary, bringing the outlet up, the reverse of the former position. (In speaking of three-way stop-cocks the long or through passage will be termed the "run," and the branch at right angles to this the " outlet.") Close stop-cock H' so that the run and outlet are closed to the capillary and measuring burette respectively, with the outlet down. Close stop-cock B', taking care that the water overflows at the end A'. Open stop-cock T' on the explo- sion burette to allow the air, caught in the capillary portion, to escape ; close when water overflows at the end u'. We are now ready to admit the gas sample. Connection is made at end A' of the absorption tube with the gas source. The level-bottle M is lowered and 28 TECHNICAL, GAS ANALYSIS. stop-cock B' opened to admit a little more than 100 cubic centimeters of gas. (The actual amount admitted will be a little more than 50 cubic centimeters, inasmuch as all three tubes are graduated to read double the capacity to avoid computation in deriving the percentages.) With stop-cock B' closed, open stop-cock H' bringing the outlet up to the main capillary, and the run up and down (the position shown in Figure 4), when connection will be made between the absorption tube and measuring burette. Raise level-bottle M and lower level-bottle N, causing the gas to pass into the burette. Now raise the level-bottle N to a height that will bring the level of the water therein to the zero or lowest graduation on the scale of the burette B. Retain in this position, and after waiting one minute for the walls of the burette to drain, lower level-bottle M a little, causing the water in the burette B to rise to the zero mark; then close stop-cock H' with the outlet up, when the burette will contain exactly 100 cubic centimeters of gas under atmospheric pressure. The error due to the capillary tubing between the 100 cubic centimeter mark and the slop-cock H' is of no consequence, since the tubing in this half-sized apparatus is but one millimeter internal diameter, requiring 127.3 millimeters (5 inches) of length to give one-tenth of one cubic centimeter. It might be remarked here that any error consequent upon the retention of air in the capillary tubing, will be slight, but may be readily estimated by measuring the length of the air bubble. The gas remaining in the capillary and absorption tube may now be expelled by raising level-bottle M, opening stop- cocks H' and M', to make connections with exit N' (having the outlets of the three-way cocks up) and stop-cock B', closing all three, on the overflow of w r ater, to their former positions. Transfer the 100 cubic centimeters of gas by raising the level-bottle N, lowering the level-bottle M, and opening stop-cock H', so that it communicates with the THE EIvUOTr APPARATUS. 29 absorption tube. Have the water rise in the measuring burette until it reaches the absorption tube end of the hori- zontal capillary, thus causing all the gas to pass over, when stop-cock H' is to be closed, with outlet up. Place the funnel F' on the end A', noticing whether there is any air in the vertical capillary ; if so, it can be dislodged by inserting a copper wire, with a little water in the funnel. We are now ready to treat the gas with the reagents. Determination of Carbon Dioxide. The first determi- nation is, as in former instances, that of the carbon dioxide. The reagent, potassium hydroxide, is poured into the funnel F', filling the latter about two-thirds full, before opening the stop-cock B' just enough to allow of the re- agent spreading evenly around the bulbed portion of the absorption tube A, and running slowly down the sides. Never allow the reagent in the funnel to quite reach the top of the vertical tube, lest air be admitted. Sufficient of the potassium hydroxide should be added to suit the requirements of the gas under examination, a very little sufficing as a rule. On the admission of the proper amount, close stop-cock B' and transfer the gas to the measuring burette by opening stop-cock H' and raising and lowering level-bottles M and N respectively, closing stop-cock H' as soon as the water from the absorption tube reaches the vertical capillary of the measuring burette. Wait one minute before making the measurement, for the walls of the burette to drain ; then, with the level of the water in the measuring burette and level 'bottle N at the same height, note the position of the lowest point of the meniscus at the surface ot the water, with refererence to the graduations of the scale. This gives the reading under atmospheric pressure, and subtracted from 100, gives the amount of absorption. Transfer the gas to the ab- sorption tube by opening stop-cock H', raising level-bottle N and lowering level-bottle M, and again treat with potas- 30 TECHNICAL GAS ANALYSIS. slum hydroxide. Transfer to measuring burette and take the reading under atmospheric pressure. If this corres- ponds to the former reading, the absorption is complete ; if not, more reagent must be added until two consecutive readings agree. It is seldom that one treatment does not suffice, but this precaution should be taken after each ab- sorption. With the gas in the measuring burette, turn the stop-cock D' so that it communicates with the drain v (position shown in Figure 4); open stop-cock B' and add water through the funnel F', rinsing all reagent from the tube. With the tube clean, turn stop-cock D' to shut off the drain and connect the level-bottle and absorption tube; open stop-cocks H' and M' to the horizontal capillary, with outlets up, and raising level-bottle M, fill the absorption tube and capillary with water, closing stop-cocks M', H' and B' as the water overflows from the exit N' and the end A' of the vertical tube. Transfer the gas to the absorption tube. Determination of Illuminants. Fill the funnel F' two- thirds full with water, and with a small dropping pipette add a few drops of bromine. Partly open stop-cock B', allowing the bromine water to spread around the bulb and flow down the sides of the absorption tube. The tension of bromine being high, the tube will quickly fill with vapor, when the stop-cock B' should be closed. Before transferring to the burette for measurement, add potassium hydroxide (emptying funnel of any bromine water remain- ing) to absorb the bromine vapors, and then transfer the gas to the measuring burette. Note the reading, which, subtracted from the previous one, gives the percentage of illuminants. Determination of Oxygen. With the gas in the meas- uring burette, rinse the absorption tube and capillary well out with water and refill. Transfer the gas to the absorp- tion tube and treat with the next reagent, potassium py- THE EId after waiting one minute for the walls to drain, raise level-bottle P so that the level of the contained water will correspond to that in the explosion burette, and note the reading, giving the volume of the sample of residual gas under atmospheric pressure. Since the determination of the constituents of the residual gas is by means of com- bustion or chemical union with oxygen, sufficient of the latter must be admitted to supply the hydrogen and meth- ane, the nitrogen remaining inert. Hydrogen requires one-half its volume of oxygen for complete combustion, and methane requires twice its volume, consequently, with an approximate knowledge of the amounts of these con- stituents present, it is a simple matter to decide upon the quantity of oxygen required taking care to admit in excess. With but a knowledge of the kind of gas under exam- ination, whether coal, water, or producer gas, etc., one will be sufficiently guided after a little experience. The oxy- gen is admitted partly as free air, (air is about 22 per cent. oxygen by volume), and partly as pure oxygen. Gener- ally speaking, for illuminating gas the air used should equal the volume of the gas sample. This may be admitted through the stop-cock T' at the lower end of the explosion burette, but more conveniently through the capillary N' and stop-cock M' by expanding the gas in the explosion burette by lowering level-bottle p. After admitting the air, make connection between the end u' of the projecting capillary and the supply of oxygen.* Open stop-cock T', admitting sufficient oxygen to make a total gas volume of *The oxygen may be prepared by heating a mixture of four parts, by weight, of potassium chlorate and one part of manganese dioxide, in a generator sup- plied for this purpose, and collecting in a gas holder in which the gas should be put under a slight pressure. THE ELLIOTT APPARATUS. 33 about 75 cubic centimeters, when close stop-cock T' and raise and lower level-bottle p, causing the gas to become thoroughly mixed, also dislodging any oxygen caught in the capillary above the stop-cock T'. Wait one minute for the burette walls to drain, then measure the total gas volume under atmospheric pressure. Dislodge any water adhering to the ends of the platinum wires by tapping the tube slightly. Lower the level-bottle p as far as possible to expand the gas, thus tending to render the explosion less severe. With. all precautions there is a possibility of fracture at this time ; it is therefore best not to expose the eyes. Make connection to the induction coil and battery, closing the circuit, causing a spark to pass between the points of the wires, thereby igniting the gas, which gives a sharp click. Allow the heat of the combustion to be absorbed by the circulating water in the jacket and then take the reading of the gas volume under atmospheric pressure. This, subtracted from the reading of total gas volume taken before the explosion, gives the contraction, which, for future reference, we will call C. Now raise level-bottle p, lower level-bottle M, and noting that the horizontal capil- lary and absorption tube are filled with water, transfer the products of combustion from the explosion burette to the absorption tube, where treatment with potassium hy- droxide will absorb the carbon dioxide formed. After treatment with the absorbent, transfer back to the explo- sion burette, and, after waiting one minute for the walls of the burette to drain, measure, under atmospheric pres- sure, the amount of absorption. Let the amount, which equals the carbon dioxide formed, be called D. We now have sufficient data to determine the percentage of hydro- gen and methane directly, and the nitrogen by difference. The chemical reaction of hydrogen with, oxygen is ex- pressed by the following equation : 34 TECHNICAL GAS ANALYSIS. H 2 + (OJj = H,O (Equation i) Hydrogen Oxygen Water One volume of hydrogen unites with one-half volume of oxygen, forming water, which, in changing from a gaseous to a liquid form, contracts so as to be of no ap- preciable volume, hence its volume may be neglected ; we have then, i volume + y 2 volume, or ij4 volumes con- tracting to zero volumes ; or, the amount of contraction equals ij^ volumes, which is i^ times the amount of hydrogen burned . ( i J4 H ) . The reaction of methane with oxygen is expressed by the equation : CH 4 + (O 2 ) a = CCX + 2H 2 (Equation 2) Methane Oxygen Carbon Water Dioxide One volume of methane uniting with two volumes of oxygen, forming one volume of carbon dioxide and water. As before, the water volume is neglected, hence we have i volume + 2 volumes, or 3 volumes ; contracting to i volume, or the amount of contraction equals 2 volumes, which is 2 times the volume of methane burned. (2CH 4 ). There is one volume of carbon dioxide formed for each volume of methane burned, or CO 2 CH,. Collecting from equations i and 2, we have C = i^H + 2CH, (Equations) D = iCH 4 (Equation 4) in which C = the total contraction resulting from the combustion of the residual gas, and D = the carbon di- oxide formed. C, the contraction, is known as measured in the explo- sion burette, also D, the carbon dioxide formed, the amount of which was determined by absorption with potassium hydroxide ; substituting for CH 4 in equation 3, its value D, from equation 4, we ha^e THE; EXUOTT APPARATUS. 35 C 2D 2C 4 D C = i^H + 2D, or, H = = (Eq. 5) _3_ 3 2 By substituting for D in equation 4, and for D and C in equation 5, their values, the per cents of methane and hydrogen become known. The sum total of the per cents so far determined, sub- tracted from 100, gives the percentage of nitrogen. An example will serve to illustrate the entire proceeding : Analysis of Coal Gas. A two-liter bottle of water was first saturated with the gas to be examined. The level- bottles and apparatus were then filled with the prepared water, expelling all air. A little more than 100 cubic centimeters of sample gas was drawn into the absorption tube, and afterward transferred to the measuring burette, where exactly 100 cubic centimeters, measured under at- mospheric pressure, was retained, the excess being expelled by filling the apparatus with water from the level-bottles. Gas then transferred to the absorption tube, where it was treated with potassium hydroxide, the first reagent. Gas returned to the burette, and, after waiting one minute for drainage of walls, measured under atmospheric pressure, giving a reading of 99.5. This, subtracted from 100, gave the percentage of carbon dioxide : 100 99.5 = 0.5, or, CO 2 = J< of i per cent. The absorption tube was drained of the reagent, rinsed, and refilled with water. Gas then transferred to the ab- sorption tube and treated with bromine water, the second reagent. As soon as the absorption was complete, potas- sium hydroxide was added to absorb the bromine vapors formed. Gas then returned to the burette and measured under atmospheric pressure, giving a reading of 95.5. This, subtracted from the previous one, 99.5, gave the per cent, of illuminants : 36 TECHNICAL GAS ANALYSIS. 99.5 95.5 = 4, or, illuminants, 4 per cent. The absorption tube was then drained of the reagent, rinsed and refilled with water. Gas then transferred to the absorption tube and treated with potassium pyrogallate, the third reagent. Gas returned to burette and measured under atmospheric pressure, givirg a reading of 95. This, subtracted from the previous one, 95.5, gave the percentage of oxygen : 95-5 95 = o-5) or > O = *4 of i percent. The absorption tube was drained of the last reagent, rinsed, and refilled with water. Gas then transferred to the absorption tube and treated with cuprous chloride, the fourth reagent. Afterwards water was added to absorb the acid vapors and wash down the curdy, white precipitate of cuprous chloride. Gas returned to burette and measured under atmospheric pressure, giving a reading of 89. This, subtracted from the previous one, 95, gave the percentage of carbon monoxide : 95 8 9 = 6, or, CO = 6 per cent. The absorption tube was drained of the last reagent, rinsed, and refilled with water. A portion of the gas now remaining, termed the resid- ual gas, was transferred to the explosion burette, and on measuring was found to be 18 cubic centimeters. To this was added about 15 cubic centimeters of air and then oxygen sufficient to bring the total volume to 80 cubic centimeters, all .measured under atmospheric pressure. By raising and lowering the level-bottle these gases were thoroughly mixed. A slight tapping dislodged the water adhering to the points of the platinum wires. The gas mixture was expanded by lowering the level-bottle as far as the rubber tubing would permit, thereby lessening the intensity of the explosion. On closing the circuit a sharp THE EIJJOTT APPARATUS. 3*7 click was heard, giving assurance that the explosion oc- curred. After a lapse of three minutes, the gas was measured under atmospheric pressure, giving a reading of 49.53. This, subtracted from 80, the amount burned, gave the contraction: 80 49.53 = 3O-47 The products of the combustion, or gas remaining in the explosion burette, was transferred to the absorption tube and treated with potassium hydroxide. Gas returned to the explosion burette and measured under atmospheric pressure, giving a reading of 41.43. This, subtracted from the previous one, 49.53, gave the amount of carbon dioxide formed by the combustion of the gas residue : 49.53 41.43 = 8.1, or, the CO 2 = 8. i cubic centi- meters, and CH A present in 18 cubic centimeters of resid- ual gas equals 8.1 cubic centimeters, since CH 4 = D = CO 2 , by equation 4. For the total amount of residual gas, 89 cubic centime- ters, the methane would equal, by simple proportion : 1 8 : 89 :: 8.1 : X ; X = 40.05, or, CH 4 = 40.05 per cent. The hydrogen is computed from equation 5 : H = ?c_iL_45 = 2(30.47) - 4(3.1) =Q5I or the hydrogen present in 1 8 cubic centimeters of residual gas equals 9.51 cubic centimeters; hence for the total volume of residual gas, 89 cubic centimeters, we have by simple proportion 18 : 89 :: 9.51 : Y; Y = 47.02, or H = 47.02 per cent. All the constituents of the sample gas have now been determined, with the exception of nitrogen, which may be found by difference as follows : 38 TECHNICAL GAS ANALYSIS. CO 2 = 0.50 C 2 H 4 (illuminants) = 4.00 O = 0.50 CO = 6.00 CH 4 = 40.05 H = 47.02 Total - - - =98.07 and 100 98.07 = 1.93, or, N = 1.93 per cent. The following table of analyses serves to illustrate the wide range of work to which this apparatus is adapted : GASES. C0 2 CO N C 2 H 4 CH 4 H Flue Gas 9.6s 8.SS o.oo 81.80 O.OO o.oo O OO (Bituminous coal) Hoffman Oven Gas. Producer Gas ... . 1.41 2. 5O o-43 o. ^o 6.49 27.00 0.00 S^.^o 2.04 0.40 36.31 2. SO 53-32 12 OO (Bituminous coal) Producer Gas . 2 SO o ^o 27 OO S7 .00 O OO 1 .20 I2.OO (Anthracite coal) Water Gas 4..OO O. If Cft Degrees Centigrade Tension in Millimeters Weight in Grams Degrees Centigrade Tension in Millimeters Weight in. Grams 2O. 0.927 5.4 6.717 13.2 11.309 10. 2.093 5-6 6.810 13-4 11.456 2. 3-955 5-8 6.904 13-6 11.605 r.8 4.016 6. 6.998 7-3 13.8 11-757 1.6 4.078 6.2 7-095 14. 11.908 12. 1.4 4.140 6.4 7-193 14.2 12.064 1.2 4.203 6.6 7.292 14.4 12.220 I. 4.267 6.8 7-392 14.6 12.378 0.8 4.331 7- 7-492 7-7 14.8 12.538 0.6 4-397 7-2 7-595 15- 12.699 12.8 0.4 4-463 7-4 7.699 15.2 12.864 0-2 4o3i 7-6 7.840 15-4 13.029 0. 4.600 7.8 7.910 15.6 13.197 +0.2 4.667 8. 8.017 8.1 15-8 I3.366 +0.4 4.733 8.2 8.126 16. 13.536 13-6 -fa 6 4.801 8.4 8.236 16.2 13.710 -f-o.8 4.871 8.6 8.347 16.4 13.885 -4-1. 4.940 8.8 8.461 16.6 14.062 -fl.2 5.011 9- 8-574 8.8 16.8 14.241 + 1.4 5.082 9-2 8.690 17- 14.421 H.5 + 1.6 5.155 9.4 8.807 17.2 14.605 + 1.8 5.228 9-6 8.925 17.4 14.790 + 2. 5.302 9.8 9.045 17.6 14-977 + 2.2 5.378 10. 9.165 9-4 17.8 15.167 + 2.4 5-454 10.2 9.288 18. 15-357 I5.I + 2.6 5.530 10.4 9.412 18.2 15.552 + 2.8 5.608 10.6 9-537 18.4 15-747 +3. 5.687 10.8 9-665 18.6 15-945 +3-2 5-767 ii. 9-792 10. i&i 16.145 +3.4 5.848 II. 2 9-923 19- 16.346 16.2 +3-6 5-930 II.4 10.054 19.2 16.552 3-8 6.014 ir. 6 10.187 19.4 16.758 4- 6.097 ii. 8 10.322 19.6 16.967 4.2 6.183 12. 10.457 10.6 19.8 17.179 4.4 6.270 12.2 10.596 20. 17.391 17.2 4.6 6.350 12.4 io.734 20.2 17.608 4.8 6.445 12.6 10.875 20-4 17.826 5- 6-534 6.8 12.8 10.019 20. 6 18.047 5-2 6.625 13. 11.162 "3 ! 20.8 18.271 (91) TABL,E I. CONTINUED. Tension of water vapor in millimeters of mercury for different temperatures. Also the weight in grams of the vapor contained in a cubic meter of air when saturated. Degrees Centigrade Tension in Millimeters |l 11 t/j Degrees Centigrade Tension in Millimeters Weight in Crams Decrees Centigrade Tension in Millimeters Si 3 3% m 21. 18.495 lS.2 27.6 27.455 99-3 741.16 21.2 18.724 27.8 27.778 99-4 743- S3 21.4 18.954 28. 28.101 27. 99-5 746.5 21.6 19.187 28.2 28.433 9> 6 749.18 21.8 19.423 28.4 28.765 99-7 751.87 22. I9-659 19-3 28.6 29.101 99-8 754.57 22.2 19.901 28.8 29.441 99.9 757.28 22.4 20.143 29. 29.782 28.6 100. 760. 22.6 20.389 29.2 30.I3I 1 00. 1 762.73 22.8 20.639 29.4 3-479 TOO. 2 765.46 23. 20.888 20.4 29.6 30-833 100.3 768.20 23.2 21.144 29.8 31.190 100.4 771.95 23-4 21.400 30. 3L548 29.2 100.5 773.71 23.6 21.659 3f. 33.405 100.6 776.48 23.8 21.921 32. 35-359 100.7 779.26 24. 22.184 21.5 33- 37.4io 100.8 782.04 24.2 22.453 34. 39.565 100.9 784.83 24.4 22.723 35- 41.827 101. 787.63 24.6 22.996 40. 54.906 105. 960.41 24.8 23-273 45. 7L39I no. 1075-37 25. 23-550 22.9 50. 91.982 120. 1491.28 25.2 23-834 55- 117.478 I 3 0. 2030.28 25-4 24.119 to. 148.791 I4O. 2717.63 25.6 24.406 65- 186.945 150. 3581.23 25.8 24.697 70. 233-C93 160. 4651.62 26. 24.988 24.2 75- 288.517 170. 5961.66 26.2 25.288 80. 354.643 1 80. 7546.39 26.4 25.588 So- 433-04I 190. 9442.70 26.6 25.891 90. 525-450 200. 11688.96 26.8 26.198 95- 633.778 2 2O. 17390. 27. 27.2 26.505 26.820 25.6 99- 99.1 733-21 735.85 224.7 25 atmos- pheres 27.4 27.136 992 738.5 (92) TABLE II. FRENCH MEASURE. One Millimetre ( r oV^ of a metre) 0.039370 inches One Centimetre ( T JQ of a metre) . . . f . . 0.393704 inches One Decimetre (^ of a metre j 3-937O43 inches One Metre (unit of length) 39.370432 inches One Decametre (10 metres) 393.704320 inches One Hectometre (100 metres) 3937.043196 inches One Kilometre (1000 metres) 39370.431960 inches One Myriametre (10,000 metres) 393704.319600 inches or, 6 miles, 376 yards, o feet and 8j 5 g inches TABLE III. TENSION OF MERCURY VAPOR. Degrees Centi- grade Tension in Millimetres Degrees Centi- grade Tension in Millimetres Degrees Centi- grade Tension in Millimetres 100 0-75 I So 11.00 260 96.73 110 1.07 190 14.84 270 123.01 120 1-53 200 19.90 280 155.17 I 3 2.18 210 2635 290 194.46 140 3.06 220 34-70 300 242.15 150 4.27 230 45-35 310 299.69 1 60 5-9 240 58.82 320 368.73 170 8.09 250 75-75 330 450.91 (93) INDEX. Absorption pipette, manipulation of the, 54-56 After-damp, 84 Analysis of a gas, 12-16 coal gas, 35-38, 57-6i scheme for the,67~72 Barium hydroxide, preparation of, 89 use of, as a reagent, 13 Benzene or benzene vapors, absorption of, 68 Boiling, definition of, 79 point, definition of, 80 Boyle's law, 75 Bromine water. 14 preparation of, 89 Burette, measuring, 9 to calibrate a, 9, 10 Carbon dioxide, absorption of, 68 * with a reagent, 13 amount of, formed by the combustion of the gas residue, 37 determination of, 13, 14, 21, 22, 29, 30 percentage of in coal gas, 57,58 properties and preparation of, 84 monoxide, absorbent for, 15 absorption of, 65, 66, 68,, 69 determination of, 15, 16, 22, 3 1 percentage of, in coal gas. 36,58 properties and preparation of, 86, 87 Carbonic acid gas, 84 oxide, 86, 87 > Carbouyl, 86, 87 * Charles' law, 77-79 . Chimney gases, apparatus for, 7 examination of, by means of the Orsat apparatus, 17-23 Chloride, cuprous, 15 Choke dump, 84 Coal gas, analysis of, 35-38, 57-61 percentage of carbon dioxide in, 57, 58 monoxide ^ A 36 ' 58 heavy hydrocar- carbons in, 58 illumiiiants in, 35, 36 . oxygen in, 36,58 scheme for the analysis of, 67-72 Collection tubes, i, 2 Compound pipettes, manipulation of ,the, 57 Conduit, currents in a, 5, 6 Correction for pressure, temperature and vapor tension, 73-75 Cuprous chloride, 15 absorption method of mak- ing the, 68, 69 preparation of, 90 Double absorption pipette, 45 lor solid and li- quid reagents, 45-47 Dowson producer gas, partial analysis of, 15, 16 Elliott apparatus. 24-38 description of, 24-26 errors in the results ob- tained with the, 24 operation with the, 26-29 principal inuovatiou with the, 24 table of analyses made with the, 38 Ethene, 84, 85 Ethylene, 14 absorbent for, 14 pipette, 48 properties and preparation of, 84, 85 Explosion pipette, 49, 50 with electrodes for the de- composition of water, Fire damp. 88. 89 Fisher's modification of the Orsat ap- paratus, 18, 19 Flue, currents in a, 5, 6 gas. analysis of, 38 Fractional combustion of hydrogen, 6 1-66 French measures, table of, 93 Furnace gases, apparatus for, 7 examination of, by means of the Orsat apparatus, 17-23 piping for running into the, 5 Gas, analysis of a 12-16 technical, 7-16 apparatus for, 7 apparatus for obtaining a rough estimate of the constituents of a, 7-1 i collection of samples of, 1-6 composition, alteration of the, by iron piping raised to a tempera- ture of redness, 5 (94) UNIVERSITY OF CALIFORNIA LIBRARY BERKELEY Return to desk from which borrowed. This book is DUE on the last date stamped below. iD 21-95m-ll l '50(2877sl6)476