46
INDUSTRIAL AND ENGINEERING CHEMISTRY
Conclusions When the furnace user or designer is faced with a problem such as this one involving the disintegration of firebrick in a protective gas atmosphere, i t will be well for him to consider if there is, or ever has been, an analogy of his problem in some other field. It will be well for him to take the materials s u p plier more completely into his confidence, weighing with the utmost care all suggestions and recommendations. T o date, very little thought has been given to the choice between firebrick and light refractories in high-carbon-steel protective atmosphere furnaces. The light-weight materials have seldom been challenged in this particular field. Firebrick manufacturere, however, tell us that they are now be-
Vol. 33, No, 1
coming awakened to this new market for their twoduct and that they intend to acquaint the industry with i l l comparative facts. It is not a large market, but it is growing and its cultivation in the protective atmosphere furnace field will work more to the advantage of the brick and furnace user than to the brick supplier.
Literature Cited (1) Furnas, C. C., J . Am. Ceram. SOC.,19, No. 6 (1936). ( 2 ) Nesbit, C . E., and Bell, SI. L., Brick & Clay Record, 62, 1042-3 (1923). (3) Swain, M. L., North Am. Refractories Co., 1940. (4) Trostel, L. J . , Trans. Am. Inst. Chem. Engrs., 31, 477 (1935). (5) Unger, J. S.,Trans. A m . Inst. Mining Met. Engrs., 125, 88-90 (1937).
City Gas for Special Atmospheres CHALMER R. CLINE American Gas Association Testing Laboratories, Cleveland, Ohio C. GEORGE SEGELER American Gas Association, New York, N. Y . Composition and source of several typical protective furnace atmospheres produced from gases and used i n such operations as metal annealing without scaling or decarburizing, bright hardening, clean heat treating, carburizing, fire protection, and purging of combustible gases from storage vessels are presented. General methods followed and equipment required i n producing and purifying these special atmospheres are discussed. Original data are presented showing the effect o n flue gas composition of amount of air supplied for combustion, fuel gas com-
I T Y gas, or utility gas, is so familiar that we think of i t only as the fuel which keeps our Bunsen burners or our kitchen stoves going. Actually i t is not so simple a fluid, nor is it identical in composition in each of the burners from which it flows. 1More than thirty different types of fuel gases are commercially sold and used throughout the United States, but in each locality they are just “city gas”. Thus, to the individual there seems but little unusual about his own gas, although it may really be a n extraordinary chemical agent. Typical gas analyses are given in Table I. They vary in composition from 100 per cent hydrocarbons to mixtures containing principally carbon monoxide and hydrogen; still others are air-gas mixtures with only 17 per cent combustible by volume. The application of these many gases to the problems of the chemist and metallurgist has progressed far beyond their use as fuels. They have attracted attention from potential users with many different requirements for special atmospheres with which to surround their products, and even as raw mate-
C
position, combustion chamber wall temperature, and flue gas temperature when three different fuel gases were burned with insufficient air for complete combustion. Composition of flue products which will result when a fuel gas of given composition is burned with a predetermined air-gas ratio may be accurately estimated from data presented. Minimum aerations supporting combustion of different fuel gases are indicated, and factors i n furnace and burner design influencing combustion at low aerations are set forth.
rials for the modification of metals and the starting point for organic syntheses. Generally speaking, when a gas has been located to meet some special need, i t may also be successfully applied for heating the process, furnace, or oven in which the special atmosphere was also a requisite. This flexibility, providing both fuel and atmosphere, has become a predominant industrial factor only within the last decade. Formerly the limit of special atmospheres was the degree of adjustment obtainable by manipulating burner control valves, and these results were expressed by the old textbook terms “oxidizing”, “reducing” and “neutral” atmospheres, Their definition, so simple t o those who coined the terms, now eludes us as completely as a buzzing mosquito on a summer night. For convenience, the words are still used to classify types of atmospheres, but they have lost much of their significance as metallurgical or chemical factors. We define them as follows: An oxidizing atmosphere is one in which oxygen concentration exceeds 0.05 per cent, and the sum of carbon monoxide and hydrogen is below 0.05 per cent.
January, 1941
INDUSTRIAL A N D ENGINEERING CHEMISTRY
A neutral atmosphere is one in which the sum of carbon monoxide and hydrogen is not more than 0.05 per cent, and the oxygen is less than 0.05 per cent. A reducing atmosphere is one in which the sum of carbon monoxide and hydrogen exceeds 0.05 per cent, and the oxygen does not exceed 0.05 per cent.
Oxidizing atmospheres produced by large amounts of excessive air in the combustion products were long known to be required for the baking of cores, for the drying of synthetic enamels, and for many phases of ceramic work. Reducing atmospheres containing free hydrogen and carbon monoxide, as well as some of the hydrocarbons, were well established years ago in a limited way for the reduction of metals from their ores and were known to be helpful in minimizing scaling of steel, oxidation of copper, etc. Such atmospheres were also used to produce the crinkle effect of certain enamels. ‘The results were satisfactory within the limits of older manufacturing specifications but they were inadequate in terms of modern demands. Even the older nomenclature has given way and the terms “oxidizing”, “neutral”, and “reducing” are being replaced by more descriptive names which indicate the effect upon the products-for example, scaling, carburizing, decarburizing, nitriding, bright annealing atmospheres, etc. This process has been one of gradual evolution, developing from the use of individual gases such as hydrogen or ammonia which had been i n use for many years. The real strides in the application of special atmospheres seem to date from about 1932.
TABLE I.
PERCENTAGE OF COMBUSTIBLE CONSTITUENTS IN UNBURNED GASES‘ CarAnthraNatural Natural Blue bureted Cokecite RefinGas, Gas Water Water Oven Producer ery Wet SweLt Gas Gas Gas Gas Gas Type Type 1.0 41.6 87.0 73.1 1.3 8.1 29.6 19.2 23.8 1.3 0.0 4.1 1.3 0.0 2.6 0.0 17.0 0.0 0.1 0.3 0.0 8.6 2.0 0.0 0.0 0.02 0.0 0.0 4.0 0.7 2.5 0.0 0.0 0.0 0.0 5.3 0.0 0.0 0.0 0.3 0.0 1.5 0.0 0.0 0.0 0.0 1.3 0.6 0.0 17.7 1.4 0.0 47.3 37.4 66.7 0.0
Methane Ethane Propane Butane Ethylene Propylene Benzene Hvdroeen *C&-boE monoxide 3 7 . 0 35.0 5.7 23.2 1.0 0.0 0.0 a BaIFnce,to make 100%. pfincipally nitrogen. ,carbon dioxide, .and minor ampurities, including light oils and their constituents, aromatice, higher paraffine, cycloparaffins, unsaturates, mercaptans, and nitrogen and oxygen compounds.
Fostered by the demands from the automobile industry for a high degree of perfection of sheet steel for deep drawing of one-piece bodies, the special protective atmospheres assumed new importance. The rapidly growing interest in bright finishes added impetus to the spread of their use. Before long there was a rush to use special atmosphere machines which sometimes mere unwarranted for the particular application, but the setbacks only served to clarify the methods, and their history has been one of rapid spread and improvement made possible by cooperation between equipment manufacturers and industrial gas engineers with the metallurgists and production men in industry. As far as metal heat treating is concerned, the special atmospheres have made possible three accomplishments which have all but revolutionized metal manufacturing: (a) clean heat treatment of low-carbon steels (and nonferrous metals as well), (b) bright hardening of medium- and high-carbon steels without any surface oxidation or decarburization, and (c) bright treatments for high-carbon and alloy steels. These classifications overlap, and a given steel may be successfully treated in more than one type of special atmosphere, depending on the time and temperature of the process. Protection given by the gases surrounding metals is a function of several variables, and what may be inert in one instance may be ac-
47
tive in another. Nitrogen is typical of such action and may be unsatisfactory for certain metals in spite of its so-called chemical inertness. In these applications of special atmospheres we are dealing with attempts to prevent the interaction of the metal with any of the gases surrounding it, except possibly the reduction of scale to the free metal. On the other hand, there is an equally important application of special atmospheres in which direct interaction with the metal is desired. Perhaps the best example of such work is the application of special atmospheres for gaseous carburizing. I n a similar way it is possible to classify these special atmospheres for chemical applications such as the inert atmospheres for fire protection or for purging of combustible gases for storage vessels of various sizes.
General Methods for Special Atmospheres The particular use will indicate the required individual characteristics in the atmosphere which will produce the desired result. There are nine general methods in current use; two of them are not strictly special atmospheres but they have been included for the sake of completeness. There are more than twenty manufacturers of national reputation building the machines for producing these various gases. The nine general methods of producing these protective heat-treating atmospheres are as follows: PURIFIED PARTIALLY BURNEDGASES. The principal application of this type of atmosphere is for annealing low-carbon steels. It is also applied to the annealing of copper tubing and of wire and for hardening some of the s. A. E. steels. Within its limitations a protective atmosphere of this type is better than direct heat, in most instances. It is hard to discuss this subject without having these reservations constantly in mind; time and temperature are vital points to consider in generalizing on the effect of an atmosphere. It must also be remembered that the atmosphere gas entering a furnace and leaving a furnace may be quite different. CRACKED GASES. These gases are much higher in B. t. u. content and in reducing constituents. They have special application for carburizing and for heat treating certain highcarbon steels. This is the type of atmosphere which would undoubtedly have an important application for chemical reduction processes. CRACKED LIQUIDAMMONIA.The principal application, up to the present, seems t o be in the stainless steel field and for fabricated parts which can justify this relatively high-cost atmosphere. It can be modified in various ways to produce any desired nitrogen concentration. PURENITROGEN. A great deal can be said about atmospheres of this character which, theoretically, are neutral to most materials. Practically, it is hard to keep the nitrogen from becoming contaminated in one way or another. The most promising commercial possibilities lie in producing nitrogen in a generator from gas as a raw material, removing all oxygen, carbon dioxide, and water, and leaving some carbon monoxide, hydrogen, and methane behind. Broadly speaking, nitrogen atmospheres are best where neutrality is desired. They are not so good where a little favoritism must be shown to one side or another. ATMOSPHERES HIGHI N CARBON MONOXIDE, HYDROGEN, AND NITROGEN. Natural gas is treated with steam and air, and then carbon dioxide and water are removed. The chemistry of producing this type of atmosphere can work in reverse. Hydrogen easily reacts with carbon dioxide to produce water. This can ruin the effectiveness of expensive dryers. CRACKED METHANOL. The gas is protective for the shorter treatments but is decarburizing in its action for the highcarbon steels. This action has been noted by many observers and is one of the principal reasons why pure hydrogen, which
4a
I N D U S T R I A L A N D E N G I N E E R I N G CHEMISTRY
Vol. 33, No. 1
The washer and cooler (c) may also remove some carbon dioxide, sulfur dioxide, and sulfur Furtrioxide, although the low partial pressure of Air : CfIl nace Retort these constituents makes the complete removal Temp., Temp., Gas Source Ratio H2 CO COS C k N, O F . ' F. doubtful unless the water wash contains approNatural gas priate chemical absorbents. An oxide box is Partial combustion 6:l 9.6 8.7 5 . 8 l . B 74 .. .. .. Processed after combustion 9:l 1.1 2 . 8 0 . 0 0 . 5 96 .. often of advantage with manufactured gases, Manufactured gas 2 . 8 : l 13.9 5 . 8 9 . 5 0.0 70.8 .. *. since i t is not expensive and offers a factor of Natural gas5 as applied to the safety if the mixture is richer than normal. The following operations: 1150 2000 8.5:l 1.5 1.5 10.7 .. .. Annealing copper tubing removal of carbon dioxide and addition of some 6.5:1 13.6 9.6 5 . 7 .. .. 1300 1600 Annealing steel sheets 8.5:l 1.5 1 . 7 1 0 . 7 .. .. 800 1800 Annealing copper wire methane or its equivalent is of benefit in over3.0.1 11.1 3 . 0 .. .. 1550 1800 Hardening axle shafts coming any decarburizing tendency. Care must 1550 1600 6.011 1410 1 1 . 2 4.6 .. .. Hardening light springs 5 . 5 : l 13.2 1 0 . 4 4 . 8 .. .. 2150 1600 be taken that oxygen is not reintroduced through 9.O:l 0 0 . 5 10.8 * . .. Cold 2000 the water or by leakage. Cold 2000 9.O:l 0 0 . 5 10.8 The effect of even small quantities of water ... 2 29 3 6.b 66 .. vapor may be very marked, and various dryers a This natural gas contains 1006 B t u. per cubic foot and is corn osed of. C H I 70% CnHa17%, NE 11%. The results show'that these protective atrnospieres ar'e established (d) are used to reduce the moisture content to for each operation in accordance with the best results. Atmosphere generator practice is levels below those at which the work will be not yet standardized. affected. The principal types of equipment employed - - for this purpose include refrigeration and adsorption b y hlica gel or b y activated alumina. Washing eliminates carbon and dust. can and is being effectively used for certain work, is not recommended for the higher carbon steels. The resultant gas is nonexplosive if the heating value is kept below 50 B. t. u. per cubic foot. Precise analyses of the comI n all of these methods of producing special atmospheres, it is hard to keep out traces of moisture. It is too easy for position of the hot gases in a furnace are difficult to make, and oxygen is likely to be present in small quantities (0.05 per cent) oxygen to enter, and if hydrogen is present, then water is also. Oxygen enters in many ways. It creeps in through wash even in the presence of large excesses of combustible. water, through extracting solutions, by leakage, etc. The When richer gases are required, the unburned gas is heated in closed containers maintained a t temperatures higher than water so produced is practically impossible to eliminate and that at which the gas is ultimately to be used. The operation is generally troublesome. is carried out in the presence of a catalyst such as course iron CHARCOAL PRODUCER GAS. This type of atmosphere has turnings and steam and/or air added to assist in the removal been popular in Europe and has now been introduced to this country for the tool steels, molybdenum steels, etc. As in the of free carbon produced from the pyrolysis of hydrocarbons in the original gas. The principal reaction in the case of previous types, the question of water vapor is important, and methane is: some of the latest units preheat the charcoal to purify it before CHI e 2Hz C producing the atmosphere gas. C HIO CO HI PROTECTION BY COATING.Various methods have been tried, such as copper plating, coating with lithium carbonate, The carbon precipitated may be removed by filtering or coating with various porcelain frits and glazes. All of the washing if necessary. It has been noted that the life of an methods have certain objections but the idea should be eniron tube is shortened when temperatures over 1500' F. are couraged, as i t is a satisfactory procedure and permits the use used, but that substitute alloys do not yield quite such good of direct-fired gas furnaces. Possibly these coatings would results. Heathcoate ( I ) is credited with the following forhave their greatest value in plants where it would only be ocmula for determining the size of tubes needed: casionally necessary to heat treat with protection. Under such circumstances a prepared atmosphere furnace might not be justified and the coating would be sufficient. Practically unknown in this country, VACUUMFURNACES. The cracked gases are higher in B. t. u. content than those vacuum furnaces are used in Germany and, to some extent, from partial combustion, may thus be in the explosive range, in England, but they do not give the perfect results that and so require care in handling. might be expected. It is difficult to keep out traces of air, and Since some of the special gases are expensive to produce, i t the vacuum seems to accentuate the action of oil films left is necessary to employ tightly constructed furnaces so as to on the work, a point that must not be overlooked in any spereduce loss by leakage. It is also important to provide for cial atmosphere furnace practice. continuity of service and for the removal of the various adI n order to apply these methods, it was first felt that scisorbing, purifying, and washing agents without interfering entific work would lead to a definite statement regarding the with successful production of the atmosphere gas. Table I11 composition of a given atmosphere, but in practice other varishows the analysis of cracked gases. ables require consideration, particularly time and temperature; therefore it is impossible to lay down hard and fast rules RESULTS FROM CRACKING OF GASES for the composition of a given process atmosphere. Table I1 TABLE 111. TYPICAL is illustrative of the typical protective atmospheres in actual Source Hs CO COt 01 P*Tz use. These are largely partial combustion atmospheres 1.6 5.0 30.4 7.0 1.8 Mfd. gas 54.2 1 7 .8 27.3 0 . 5 1 0 . 0 1 . 7 Mfd. gas 42.7 supplemented by purification, removal of carbon dioxide and 0.0 48.0 7.0 14.0 3.0 Natural gas 28.0 0.0 40.7 0.3 12.0 dehydration, as required. 15.0 Xatural gas 32.0 The essential part of machines for this purpose are: (a) a gas burner permitting the control of the air-gas ratio a t SAFETYNOTE.Since a great many atmospheres contain fixed levels, (a) suitable combustion space with catalyst when large quantities of toxic constituents such as carbon monoxide, required, (c) a washer and cooler intended for removal of some every precaution should be taken to ensure the removal of of the water vapor produced and for cooling the gas, and (d) such gases around the furnace. They must not be allowed t o a dryer. PR6TECTIVE ATMOSPHERES PRODUCED TABLE 11. TYPICAL
FROX
GAS
..
+
+ +
"ZH?
January, 1941 contaminate the air in the room. This may require ventilating hoods a t loading and unloading ends, if not over the entire furnace. Extreme care is also necessary to avoid explosive mixtures. Leaky furnaces may be dangerous as explosion hazards as well as health hazards.
49
CAST IRON
Purification of Special Atmospheres CERAMIC 01% The very term “special atmospheres” which has been applied in this paper denotes CERCMIC CONE purification of the normal gaseous products, both burned and FIQWRB 1. WATER-COOLED FURNACE B unburned, to produce certain specific results. I n general, five If these we not applied, the analyses will be on the dry basis. Dew processes are employed to purify partially burned and cracked point methods are given by Stack and Hotchkiss (4). atmospheres as follows: Oa. Down to 0.05 per cent; easily reintroduced. Traces are removable by hot cast-iron borings. CONDENSATION OR WASHING.Water is generally used in a NO. Controllable but not present in quantities which appear to be suitable washer-scrubber tower to condense the atmosphere and troublesome for f ~ n a c operation. e cool it. If necessary, solutions to remove carbon dioxide, sulfur Dust. Not generally present but removable by filtration to pardioxide, and sulfur trioxide may also be employed. Condensation tides of around 1 0 size. ~ removes moisture down to the vapor pressure of the cooling water, but for certain processes this is insufficient and further waterremoval methods are applied. Experimental Limits of Partial Combustion CHEMICAL PURIFICATION. Chemical removal of the acidic constituents is applied when necessary-for example, the removal literature on combustion reveals that, ~~~i~~ of the of carbon dioxide and sulfur dioxide by the ethanolamines (GirbiOnce the experimenters reached perfect combustion condito1 process). The removal of hydrogen sulfide is a most importions, they practically ceased further investigation. A few tant consideration and is briefly outlined as follows: The removal of hydrogen sulfide, produced during gas prepare bolder workers attempted to predict ratios of hydrogen and tion, follows similar methods in regular gas company purification carbon monoxide which might be expected if the air supply practice. The gas is passed through a mixture of iron oxide and wood shavin or other b a y inert material. After the oxide were restricted, but their information was incomplete. Conhas become fouled (i. e., converted to iron sulfides), it can no trariwise, the expansion of inert atmosphere generators has longer absorb hydrogen sulfide and must be replaced or revivified yielded many data on the actual results obtained for a parby moistening and s reading in the open air. ticular furnace operating in a specific manner, with little efIn special atmosptere units the amount of hydrogen sdfide to be removed is substantially smaller than in the purification of fort being made to show the effect of each of the various factors involved. raw gas. Consequently, the oxide boxes for special atmospheres have a lon useful life and are inexpensive to maintain. ReContinuing its policy of establishing basic combustion removin fouyed oxide is obviously disagreeable, and proper prolations, the American Gas Association, through its Commitvision for ventilation is necessary to avoid hazards from hydrogen size to facilitate tee on Industrial Gas Research, sponsored a series of studies sulfide. Openings to the boxes should be to determine the effect of burning fuel gas with insufficient air this work. The formula of the Steere Engineering Company (g) for sizing these boxes is often used although it will result in rather for complete combustion. As a result of these studies, reliberal capacity: cently completed a t the American Gas Association Testing 0.16 (cu. ft. of gas per hr.) Laboratories, i t is now possible to predict accurately the comArea of box in sq. f t . = depth of layers in ft. 4 position of flue gases (as determined by ordinary Orsat analyThe factor4 is added in the denominator for a two-box series, SiS) resulting from combustion Of fuel gases with limited air 8 for three-box, 10 for four-box. supply. ABSORPTIONAND ADSORPTION.Methods for absorption and adsorption are principally applied to the removal of water vapor Three types of test furnaces were employed in these studies to below cooling water levels. The removal of organic sulfur, simulate burner tunnels used on industrial furnaces. One, furpresent in the form of carbon disulfide, mercaptans, or similar nace B , consisted of a water-jacketed brass combustion tube apcompounds, by activated carbon has been tried with moderate roximately 2 inches i. d. and 47 inches lon as illustrated in success. hgure 1. A ceramic disk drilled with twenty-fve No. 50 D. M. S. REFRIGERATION. This is principally applied as an alternate ports placed in the tube near one end served as a burner port method for the removal of water. and flame retainer. This arrangement kept the flames spread more or less uniformly over the cross-sectional area of the comThe degree to which various constituents are removed with bustion tube. Thoroughly remixed air-gas mixtures were concommercial equipment is shown in the following list: veyed to this drilled burner i s k through a ceramic tunnel diverging from a cast-iron nozzle. Four side sample ports, starting 3 HzS. Completely removed. Traces can be removed by hot copinches from the burner disk and spaced thereafter at 12-inch inper turnings. tervals along the combustion tube length, were provided to perCot. Completely removed. mit sampling and temperature measurement during the combusSOZ. Completely removed. tion process and subsequent cooling of flue products. A second SOa. Completely removed. test furnace, C, consisted of a 17/&-inchi. d. refractory wall comHxO. Down to vapor pressure of cooling water. Activated alubustion tube, 49 inches long with walls 1 inch thick. Side sammina will reduce to 0.01% or -4OO F. dew point. plin ports, drilled burner disk, and premixing arrangements HeO. Removal (by refrigeration) down to dew point corresponding simfar to those of furnace B were employed as shown by the to refrigeration temperature. upper sketch of Figure 2. The third test furnace, D, was similar HzO. Removal (by adsorption) down to vapor pressure of adsorbto furnace C except that a specially designed, combination preent. Relatively simple dew point methods are available which will mixing and diffusion type of burner was employed in place of the permit determination of the moisture content of furnace atmospheres. total premixing type. With this special burner an aerated gas
+
so
INDUSTRIAL AND ENGINEERING CHEMISTRY
Vol. 33, No. 1
aeration, or amount of air supplied for Combustion, on concentrations of various flue gas constituents produced b y combustion with limited air supply, is shown in Figure 4. Throughout this paper, air supplied for combustion will ERAMIC CONE be expressed as a percentage AIR-6PS MlXTU CERAMIC WALLS of that theoretically required for complete combustion, or 2 5 - NO. 5 0 D.M.S. PORTS more simply, as per cent aeration. Such a term simAIR-GAS plifies comparison of results MIXTURE obtained with fuel gases having widely different combustion air requirements and avoids confusion due to minor changes in combustion air r e q u i r e m e n t s , as heating FURNACE D value of each fuel gas varies within limits from day to day. Concentrations of the various flue gas constituents, including water vapor formed FIGURE 2. CERAMIC FURNACES C AND D by combustion, are expressed as Dercentaaes of the total mixture was su plied through the center opening of two concendry flue products; this basis is the-one commonly employed tric orifices whiye unburned gas was admitted through the larger in Orsat analytical work. outside orifice as shown by the lower sketch of Figure 2. The Although small variations in concentration of different flue air-gas mixture passed from the orifice t o the burner disk, burned on its outer face, and thus caused a hot spot at the center. The gas constituents from values shown in Figure 4 were observed unburned gas was consumed around the edges of this hot center, in many instances, average values are well represented by the its combustion being supported by diffusion of air from the center curves of these figures. air-gas stream. Test furnace D was employed only for a few RESULTS WITH NATURAL Gas. When burning natural gas special tests, and any references here to a ceramic furnace apply t o furnace C with the premixing burner, unless otherwise desigwith an air supply equal to that theoretically required for comnated. plete combustion (100 per cent aeration), concentrations of Natural gas with an average gross heating value of 1108 B. t. u. carbon dioxide and water vapor closely approached ultimate per cubic foot and coke-oven gas (manufactured gas A ) with an limits as indicated on Figure 4 (upper graph). Deviation in average gross heating value of 544 B. t. u. per cubic foot were ,burned in the water-cooled furnace; the same natural gas and a concentrations of these gases below their ultimate limits is synthetic coke-oven gas (manufactured gas B ) , made by pyrolyadequately accounted for by presence of small quantities of sis of natural gas and having a gross heating value of 534 B. t. u. oxygen, carbon monoxide, and hydrogen in the flue products per cubic foot, were burned in the ceramic-walled furnaces C and D. a t this aeration. Decreasing air supplied for combustion reIn operating these test furnaces, predetermined fuel inputs were burned with various proportions of air. Test furnaces were sulted in decreasing concentrations of carbon dioxide and operated under each condition for a sufficient length of time to water vapor, and proportional increases in carbon monoxide establish thermal equilibrium. Samples of flue gases were then and hydrogen concentrations in flue products. taken, and temperatures were measured at each sampling port. Suitable equipment was employed to assure accurate sampling and freedom from subsequent con%amhationof flue gases. Analyses of the flue gas samples were made with a modified Orsat type of analytical apparatus. Flue gas temperatures were measured by a highvelocity thermocouple method, the chrome-alumel and platinum-platinum rhodium couples employed being checked at frequent intervals t o avoid errors from possible contamination of the wires by reducing gases. A general view of sampling equipment and highvelocity thermocouple being used with test furnace B is shown in Figure 3.
h(FJ
Effect of Quantity of Air on Composition of Combustion Products When a fuel gas was burned in any given test furnace, material changes in composition of flue products were found to occur as the amount of air premixed with fuel gas prior to com.bustion was varied. This effect of
FIGURE 3. FURNACE B AND SUPPLEMEXTARY EQUIPMENT
,
January, 1941
INDUSTRIAL
22 20 I8
16 14
12
10
8
6 4 2
0 26
24 22
g
20
0 W
8 P
18
I
5
16
W
I-
E
14
8
12
w A
10
z
In
6 4
2 0
PERCENT AERATION
FIGURE4. FLUE PRODUCTS RESULTING FROM COMBUSTION OF NATURAL AND OF MANUFACTURED GASESA AND B WITH LIMITED AIR SUPPLY
.
i
HEMISTRY When natural gas was burned in the water-cooled furnace a t the same aeration as in the ceramic furnace, no noticeable change in concentrations of carbon dioxide, oxygen, carbon monoxide, hydrogen, or water vapor in products of combustion were noted. These results justify the conclusion that wall temperatures have no appreciable effect on the amount of these flue gas constituents formed a t any given degree of aeration. Although concentrations of oxygen and methane in the products of combustion were small a t all aerations, change i n concentration of these constituents, as air supplied for combustion was decreased from 100 per cent aeration, had certain peculiarities worthy of mention. Oxygen concentration i n flue products reached a minimum value of 0.05 per cent when air supplied for combustion was decreased to approximately 95 per cent, but increased slightly as aeration was further decreased. Actual presence of small concentrations of oxygen in the flue products a t all aerations was indicated by checking results obtained with potassium pyrogallate absorbing solution in Orsat analytical equipment against the more sensitive manganese hydroxide titration method. It is likely that these small amounts of free oxygen in flue products were in equilibrium with the other flue gas constituents, and that even if all oxygen were removed by some means, more would be formed a t elevated temperatures by dissociation of carbon dioxide or water vapor. Methane was found to occur in traces only or to be entirely absent from flue products of natural gas for combustion air supplies ranging downward from 100 to 80 per cent aeration. As the combustible limit a t lower aerations was approached and the flames became unsteady, methane content of the flue gases increased rapidly and reached its maximum concentration of approximately 0.5 per cent at the lowest aeration supporting combustion. Conditions such as heat loss characteristics of combustion tube and type of burner employed influenced the minimum air supply which would support combustion. For this reason, aeration could be decreased to a lower value without producing appreciable quantities of methane in the flue products when employing the ceramic furnace than for the water-cooled one. No illuminants (ethylene, benzene, etc.) or ethane were found in the products of combustion of natural gas a t any aeration which would support combustion. RPSULTSWITH MANUFACTURED GAS. Results obtained by combustion of manufactured gas A with limited air supply in the water-cooled test furnace are indicated in Figure 4 (center graph). As in the case of natural gas combustion, when carbon dioxide and water vapor concentrations approached ultimate values of 100 per cent aeration, only small amounts of hydrogen, carbon monoxide, and oxygen appeared in the products of combustion. As air supplied for combustion was decreased, amounts of carbon dioxide and water vapor in the flue products decreased while amounts of carbon monoxide and hydrogen increased, these concentration changes being proportional to aeration decreases. Although concentration of free oxygen in flue products again reached a minimum a t approximately 95 per cent aeration, minimum values were 0.15 per cent oxygen for this gas, and concentrations gradually increased as combustion air was decreased, 0.20 per cent oxygen being noted a t 70 per cent aeration. Only small concentrations of hydrocarbons were found in flue products of manufactured gas A for aerations above 75 or 80 per cent. As combustion air was decreased below this point, however, methane concentration became increasingly greater and reached approximately 1.0 per cent a t the lowest aeration supporting combustion. Similarly, illurninant content in flue products increased from zero to 0.5 per cent when aeration was decreased from 75 per cent to the lowest figure supporting combustion. It should be noted that a t 58 per
INDUSTRIAL AND ENGINEERING CHEMISTRY
52
cent aeration, which was the lowest aeration attainable in the water-cooled furnace, methane concentrations in the flue products represented only about one tenth of that originally in the fuel, while approximately half of the illuminants in the fuel appeared in the flue products. These results lead to the inference that methane is more readily burned or cracked than illuminants under conditions of minimum aeration. Absence of ethane in flue products a t any aeration demonstrates that this constituent is more readily combustible than other hydrocarbons in the fuel gas.
5
100
Y
b 90
z
' 8 Y
80 70
3 z
60
I"
50
K 0
0 4 0
se
Vol. 33, No. 1
carbon of each fuel gas between carbon monoxide and carbon dioxide was very nearly the same. For example, a t 90 per cent aeration, total carbon of natural gas was distributed 20 per cent to carbon monoxide and 80 per cent to carbon dioxide. For manufactured gases A and B , 20 and 18 per cent, respectively, of total carbon appeared in flue products as carbon monoxide and 80 and 82 per cent as carbon dioxide. At 70 per cent aeration, 52, 51, and 50 per cent of total carbon in natural gas, manufactured gas A , and manufactured gas B , respectively, appeared in flue gases as carbon monoxide. Amount of carbon found in flue products as carbon dioxide a t this aeration was 46, 46, and 48 per cent, respectively, of total carbon in the three fuel gases. At this aeration the small amount of total carbon not accounted for by that in carbon dioxide and carbon monoxide was contained in hydrocarbons, concentrations of which are not shown in Figure 5 . Division of total hydrogen in fuel gases between free hydrogen and water vapor in flue products was similarly consistent (Figure 5). Use of these relations should provide a simple means of estimating composition of flue products from any fuel gas within limits of composition of those employed in these studies when burned with limited air supply.
_1
8
w
30
Variation of Composition of Flue Products with Temperature
20
$
IO
L
o
0
60
70
80
90
100
PERCENT AERATION
FIQURE5. PERCENTAGES OF TOTAL CARBON TOTALHYDROGEN IN THE THREETEST GASESFOUND IN THE FLUECONSTITUENTS AT VARIOUSAERATIONS
AND
Manufactured gas B was found generally similar to the other two test gases in so far as effect of aeration on concentrations of carbon dioxide, water vapor, carbon monoxide, and hydrogen in flue gases was concerned. Relative concentrations of the different flue gas constituents found by combustion of this gas with varying amounts of air are shown in Figure 4 (lower graph). Changes in concentrations of oxygen were found similar to those noted for natural gas, minimum concentrations of 0.05 per cent appearing in the flue products a t 95 per cent and slowly increasing as aeration was further decreased, but never exceeding 0.10 per cent oxygen. As pointed out for the other two gases, methane content of flue products remained relatively small as air supplied for combustion was decreased until a figure was approached a t which combustion was not self-supporting. Maximum methane concentration observed a t the lowest aeration was 0.8 per cent.
Effect of Fuel Gas Composition on Composition of Combustion Products Although combustion characteristics of the synthetic cokeoven gas (gas B ) were similar to the regular coke-oven gas (gas A ) , in many respects the composition of flue products of this gas, which was made by pyrolysis of natural gas, more closely resembled those of natural gas. Further analysis of the data disclosed that, although the kind of constituents in the fuel gas was responsible for its combustion characteristics, total amount of carbon and hydrogen (either free or combined) in the fuel gas has the greater influence on volume of each flue gas constituent formed. This is illustrated in Figure 5 , which shows that for each aeration, proportional division of
Variations in composition of flue products with change in Rue gas temperature were found to be small under the conditions of this investigation. Although no change in concentration of any one flue gas constituent greater than 0.5 per cent of the total dry flue products was noted for any temperature change imposed, it does not follow that temperature may be considered negligible for all industrial applications. Flue products were cooled so rapidly in the test furnaces, from over 3000" to less than 1800" F. in approximately 0.2 second, that the element of time may have prevented greater composition change. Moreover, combustion tubes employed were constructed of brass and ceramic materials, and it is possible that the presence of other materials, such as iron oxides, might catalyze reactions and thus cause greater flue gas composition changes. The only flue gas constituents showing noticeable concentration change with decrease in temperature were carbon dioxide, carbon monoxide, hydrogen, and water vapor, and change in concentrations of these components was not observed below approximately 1600" F. At higher temperatures concentrations of carbon dioxide and hydrogen in flue products generally increased, and those of carbon monoxide and water vapor usually decreased as temRerature of flue products was decreased. These composition changes apparently were influenced by amount of temperature decrease, original temperature level and original amounts of each flue gas constituent. Temperature decreases from approximately 2000" to 1600" F. produced little effect on flue gas composition, changes in concentration of carbon dioxide, carbon monoxide, hydrogen, or water vapor being not more than 0.1 or 0.2 per cent of the total dry flue products. Concentration of any one of these flue gas constituents usually did not change more than 0.3 per cent for temperature decreases in the temperature range of 3000" to 2000" F., although a few instances were noted where hydrogen and carbon monoxide Concentrations changed as much as 0.5 per cent for such temperature decreases. As previously noted, furnace wall temperature had no noticeable effect on flue gas composition other than its influence on minimum aeration supporting combustion, which was responsible for appreciable quantities of unburned methane in flue products at aerations approaching this combustible limit.
INDUSTRIAL AND ENGINEERING CHEMISTRY
January, 1941
Effect of Furnace Design on Combustion I n the preceding discussion it was pointed out that each test gas could be burned in either test furnace a t any degree of aeration down to 80 per cent without producing more than traces of unburned hydrocarbons in the flue gases. As aeration was decreased below this figure, a point was reached for each gas and each furnace where heat developed by combustion wm lost to the furnace walls rapidly enough so that the gas temperature was only a little above the ignition temperature of the air-fuel gas mixture. As the lowest aeration supporting combustion was approached, less and less heat was available to crack the remaining hydrocarbons in the fuel, and methane concentrations in flue products became increasingly larger until that aeration was reached where heat generated would not sustain combustion. Effect of combustion chamber material on combustibility of natural gas is shown in Figure 6. Thus it may be seen that heat loss characteristics of the combustion chamber had an important effect on the minimum aeration which will support combustion of a given fuel gas. Although only two materials of construction were used in this investigation, the ceramic combustion tubes proved to be much better than the water-cooled metal tubes for combustion a t very low aerations, and results indicated that combustion tubes having better insulation could be employed to burn gas a t still lower aerations. For example, the minimum air supply which would support combustion of natural gas in the ceramic and water-cooled experimental furnaces was 67 and 77 per cent, respectively, of the total air required for theoretically complete combustion. Similarly, manufactured gases could not be burned below 53 and 58 per cent aeration in the ceramic and water-cooled test furnaces, respectively. 0.6
0.5 0.4
u)
9
flue products, leading to premature extinction of flames. Multiple-drilled ports in ceramic burner disks overcame such characteristics by providing an essentially vertical flame front, as well as a more uniform flow distribution of gaseous mixtures over the cross section of the combustion tubes. It was also noted that ceramic material placed around or close to the burner ports, which became incandescent by its proximity, was helpful in keeping the flame burning smoothly a t low aerations. Still further improvement in combustibility was made by subsequent changes in burner design. A double-port arrangement was substituted for the single port through which the combustible mixture was transmitted to the ceramic burner disk. A mixture containing approximately 100 per cent of the air theoretically required for complete combustion was supplied through a central port, while raw gas was admitted through an annular port surrounding this, as shown in the lower sketch of Figure 2. Although some mixing undoubtedly occurred prior to combustion, a relatively airrich mixture was maintained a t the center of the ceramic burner disk and a gas-rich mixture a t its extremities. The air-rich mixture provided a high-temperature ignition source, while absorption of heat by the outer layers reduced heat losses in proportion to average temperatures. Under these conditions natural gas was found combustible with 62 per cent aeration, or a t 5 per cent lower aeration than with the total premixing type of burner illustrated in the upper sketch of Figure 2. These results indicate that other burner designs may be developed which will utilize the heat generated by combustion to greater advantage in maintaining combustion a t very low aerations.
Future of Special Atmospheres In the preceding sections data have been presented on the results of the investigation of the limits of partial combustion. It is worth remembering that these limits are not the limits of special atmospheres in general because of modfications which can be produced by other methods such as cracking, removal of unwanted gases, catalysis, etc. Nevertheless, the rwults of the American Gas Association's current research appear useful because they have charted a previously unexplored area in our knowledge of combustion. They may be summarized as follows:
I-
37
53
0.3
W
2 0.2 3
I' 0.1 0
0 60
70 80 PERCENT AERBTION
FIGURE 6 METHANE I N FLUE PRODUCTS RESULTING FROM COMBUSTION OF NATURAL GASWITH LIMITED AIR SUPPLY
Although not investigated in this study, it seems reasonable from the foregoing to assume that for production of larger quantities of carbon rdonoxide and hydrogen lower aeration mixtures could be burned than were attained in this study, by external heating of the combustion chamber walls or by use of preheated air. Also, it is not unlikely that cross-sectional areas different from those employed in this investigation might permit combustion at lower aerations.
Effect of Burner Design on Combustion I n the early stages of these studies it was found desirable to employ drilled-port ceramic disks of the type illustrated in Figures 1 and 2. Single-port burners were found to promote flame fronts having a large horizontal component as air supply was reduced toward lower limits of combustion. Combustion on a long horizontal flame front usually resulted in a puff of flame and contamination of air-fuel gas mixture with
1. For a given fuel gas, composition of flue products may be varied over a wide range by limiting the amount of air supplied for combustion. 2. Other variables, such as flue gaa temperature, input rate, or combustion-chamber wall temperature, were found to have a relatively minor effect on concentrations of flue gas constituents formed at any given degree of aeration. 3. Oxygen may be practically eliminated from flue gases by reducing the air supplied for combustion to 95 per cent or less of that theoretically required for complete combustion. 4. Fuel gas can be burned at aerations close to the lower oombustible limit without more than traces of unburned hydrocarbons ap earing in the flue products. 5. d u e gas composition which will be produced by combustion of a given fuel gas at a given aeration may be estimated from the fuel gas analysis, since it was found that amounts of carbon dioxide, carbon monoxide, hydrogen, and water vapor formed by combustion were a function of the amounts of total carbon and of total hydrogen (either free or combined) in the fuel gas. 6 . Where it is desirable to burn fuel gas with least possible air supply to produce maximum quantities of reduoin gases, carbon monoxide and hydrogen, careful attention sho3d be given to both heat loss characteristics of combustion tubes or chambers and to the design of burner employed.
Throughout this paper the use of prepared atmospheres in metal treating has been stressed, mainly because this has been the largest field in which they have been applied, but it has by no means been the only one. Space will not permit more than a hint at the things t o come although so useful a tool is
I N D U S T R I A L A N D E N G I N E E R I N G CHEMISTRY
54
bound t o expand as manufacturers become familiar with its possibilities. The metallurgical industries are only beginning to apply these methods for the reduction of ores. Some of the typical examples are the decarburizing of pig iron t o make steel by the R. K. process, recently developed in Sweden (3). At least two and possibly three mining companies are experimenting with the direct reduction of iron ore with special atmospheres made from gas which, if successful, will be revolutionary in character. I n the chemical field it may be wrong to claim hydrogenation as an application of special atmospheres, but it does no harm t o remind ourselves that the special atmosphere is an excellent source of high hydrogen concentrations. Chemical manufacturers handling flammable liquids have used inert atmospheres for fire protection and for purging. Theoretically, boiler flue gases from any fuel might be applied for this work, but practice has shown the desirability of using controllable, independent, special atmosphere generation in order t o avoid the hazard of introducing air, which may happen so easily in the case of boiler gases. I n the related paint and varnish field, applications of inert atmospheres for agitating linseed oil and blanketing it against fire hazards are growing rapidly.
Vol. 33, No. II
Because the results of applied special atmospheres improve the quality and finish in so many products, there is certain to be a definite change in specification for many different classes of merchandise. Not only is it possible to make better looking products, but the bright finishes and hard surfaces make new applications practical. I n no field is this likely to become of greater significance than in our program of national defense. Twenty years ago during the first World War the demand for certain munitions almost exceeded our ability t o turn them out. It was a mechanical development (the continuous-conveyor oven) which received its start in those years and which proved to be one of the vital factors in doing the job. Today it seems as though a second opportunity may arise, and the gas industry is ready to play its part by concentrating on the inherent possibilities of speed developed by gas heat and quality protected by gas atmospheres.
Literature Cited (1) Fisher, A. J., “Furnace Atmospheres”, Baltimore, C. M. Kemp
Manufacturing Co. (2) Gas Engineers’ Handbook, D. 414, New York, McGraw-Hill Book Co., 1 9 3 L (3) Iron A g e , 144, 40 (Nov. 2, 1939). (4) Stack and Hotchkiss, Metal Progress, 31, 375-9 (1937): Gen. Elec. Rev., 41, 106-8 (1938).
Gas Carburizing by the Hypercarb Process W. A. DARRAH Continental Industrial Engineers, Inc., Chicago, Ill.
F
OR a long time it has been known
The Hypercarb process serves to apply a controlled highcarbon case on steel articles, which greatly increases the hardness and wear resistance of the surface. The process depends on treating steel articles at controlled temperatures for a controlled time by an atmosphere containing hydrocarbons and carbon monoxide in definite percentages. The articles leave the furnace clean and free from carbon scale or soot so that n o cleaning is necessary. The process is controllable and economical, and has a wide commercial application.
that heating steel articles in an atmosphere containing methane or other carbon-rich hydrocarbons causes them t o absorb a oortion of carbon on the outer surfaces. This led to several inferior processes which produced carbon-rich coatings on the surface of steel articles but usually formed carbon scale on these surfaces, which involved a costly cleaning operation, usually sandblasting, scratch brushing, or similar applications. Attempts were made to overcome this difficulty by reducing the amount of carbon available. The results were wide variations in the carbon content and usually a deficiency of carbon in the case. Further attempts made to overcome this problem by diluting the hydrocarbons with other gases led to similar difficulties. A study of the factors involved indicated that it is possible to carry on carburizing operations to good advantage in a container made of refractory material. Heat losses are greatly reduced, the first cost of equipment is reduced, and the maintenance and trouble incidental to the muffle are eliminated. Following extensive tests on the carburizing process and the muffle of refractory material, a study was made of the possibility of pretreating and preactivating the gases used for
carburizing. By heating some of the carburizing gases to approximately 1600” to 1700” F. prior to permitting them to come into contact with the steel being treated, a rapid carburizing action was obtained and the gases were much more active than unheated gases. The gases thus pretreated did not deposit carbon scale or soot on the steel being treated, and it was possible by the proper pretreatment of the gas to bring out steel articles which would be entirely clean and bright and still have decided carburizing effect on the surface of the article. A further study of this process seemed to indicate that the preheating of the carburizing gases developed especially active compounds, most of which appeared to contain carbon monoxide in some form. The action of free carbon such as soot or coke depositing from the carburizing gases was very slight in combining with the steel. The general results of this investigation tended t o indicate that the carbon which combines with steel under ordinary
