Determination of Oxides of Carbon in Hydrogen-Nitrogen Mixtures Continuous Determination of M i n u t e Concentrations by Eiectroconductivity H. BROWN
EARL
AND
MAURICE M. FELGER coiitinuou,.. ~t'pai':itc., or riiiiiilt:i:itv)ii+ clc.terniinat iori of i,:d)otl dioxi(l(~ :ind cr:irhoti ~ n o ~ ~ r ) iuni (wnwntr:lt d(~ ions of 0 t u I O U p.p.~ii.
A n electroconductometric analyzer for the continuous, separate, or simultaneous determination of carbon dioxide and carbon monoxide in concentrations of 0 to 100 p.p.m. in hydrogen-nitrogen mixtures and the application of the analyzer to the analytical control of the TVA synthetic ammonia plant are described. A n absorptionconductivity cell of new design and a recording alternating current Wheatstone bridge are employed. Carbon monoxide, after oxidation by an improved iodine pentoxide reagent, i s determined as carbon dioxide. Full operating details are given.
h ~ r g ( a~i rd Schrcnk ( - $ I clcxrihcil cc.,nipletely variouh rric.1 t i o t l . tor tho det,ection and dcterininotiori of ( w h o r l monoxide. o n i y
T"
gas-purifio:ition s y s t c m of thr 'l'\,.:A synthet,ic aniiiionia plant h a j bcen described by lliller and Junkins (IS). The present paper describes in drtail the niethod by which the purity o f the synthesia n1istur.e is cwi(iullrd with respect to tlir osiclcs of rarbon. Tire 1iydropc.11-~iitrogctri iiiixture protlric~edby thv c:ttalyt ic I,('wtion of semi-water ~ ; t swith steam contains a large coiictmtraricin of carbon riioxitk. :ind a slnali conccnt ration of unconvertc~i carbon monoxith.. Thcst: oxitlrs of carbon arc removed by p:asing the convrrttd rcnii-\r.atc~rgar through two scrubbing tow(aiv iinder pressur(&. In tlit. firht t o w r , water reiiioves about 98% o f the carbon dioxitlc, arid in tlic second, aniinoiliacal copper soliition rcducea the concent,rationof each oxide to 5 p.p.m. or 1( Sinccs both carbon dioxide and carbon inonoxide poison thc synthesis catalyst), it is 1iccc~ssitIyto have a continuous anti rapid iiitlication of their conccntrations, so t,hat gas containing Iiarinf'iil eoncent'rations niay be vented and corr ve measures takeii. For :ind.vtic~alcontrol purposes, mi :mal> {vas tlesirctl f o r t l i c t
Figure 1 . A endcoil E endcoil C m . Measurlns electrode chamber C,. Reference electrode chamber D. HzSO, =rubber Em. Measuring electrodes C. Reference electrodes A. E.
t\vo of these metliods appear plscticrtl for the continuous determination of Ion conrt presence of the w t a1y.t Hopcalit e. Tlie analyzer w-as foiiritl witable for tlic tletc,i.iriiri:ttioii of ( x r t ) o i i inonoxide in hydrogvii (/?), provided thc saiiipl(- [vas tlilutctl ivith a sniall, fixed proportion of air prior to its iiitiudu~~ioii i,ntii t,he apparatus. Thi.: inc+hotl was not suita!)lti for :tpplic.:itloll to aninionia-rynthesigw, hoxever, bocausc, i t w:i* not : i i l : i l i ( t l l ) l ( ~ t o thta iic:t~c,r.riiiu;itioii of carbon tliositli~. \lethods b a d 011 thv oxi(l;itioii r i t cartioti nionoxiile by io(linc. pwtoxide :ire numcrous ( 1 , .i, 9, 10, 1 1 , /.C, 1 2 7 , I ? , 18). The conrentration of c a r h i inoiiouiclc is calculated from tlic quantity 0 1 c a r h n dioxide formed o r iotline 1il)t:ritted by a givc:n volume, of gw. Tha quantity of iotliiic. liberated is dt:termiiied by absorption in potassium iodide solution and titration with sodium tliiosulfate. The carbon dioxide generally is absorbed i n :til R I M i n e solution xnd the quaiititmydetc?rmined by titrat,ioii or ot h i . incthods. I n B;tltlewyn~'method (2) tlit: pas sampie is pttsseti tlirougli ii i~ificatioiirtntl oxidation train in which the carhon monositlib oxidized with iodine pentoxide. The resulting carbon dioxiiic. then pashotl int.0 an electrolytic cell containing a definite voliirnt, of 0.005 S barium hydroxide solution. This cell and a refertmc*c~ ecll containing barium hydroxide solution of the same con tion are made the arnis of it \Vheatstcme bridge with a Kohl .;lide-wire as the intlicating in?(rument. The co~icent,rationo f ( m h n iiio~ioxidais deterniinc.tl from the bridge readings and t h(, voliini(> of gas nsc~l. Thc~i I i s t i ~ i r n etr ~ is cali hrnt i ~ 1 r mi Iii ric.:i I I.!
Carbon M o n o x i d e Analyzer
F. Flowmeter
M. KI scrubber N. Electrolyte inlet
C. Galvanometer H. KOH-Ascarite tube 1. 1 2 0 5 tube J. Pz05tube K. 1 2 0 5 furnae L . KOH-Ascarite tube
0. Temperature stabilizing coil P. Pressure stabilizer R . Electrolyte Bow regulator SI S3 S + By-passdopcocks S,: Crbsr-bore stopcock
277
S W . Slide-wire
T. Gar absorption coil U. Gas inlet to cell V. Gas flow regulating valve W. Gas exit
X . Constant-level electrolyte overflow
Y. Electrode adjusting screw
INDUSTRIAL AND ENGINEERING CHEMISTRY
278
This electroconductivity method was adapted to the continuous determination of carbon monoxide in hydrogen-nitrogen mixtures in the production of synthetic ammonia ( 8 ) . The train was modified by the addition of a by-pass arrangement, so that carbon monoxide and carbon dioxide could be determined singly or together. A recording alternating current Wheatstone bridge was used for indicating and recording the ratios of the resistance of the measuring cell to the resistance of the reference cell. This method has the obvious disadvantage that it does not give a direct indication of the concentration, because the continuous use ?f the same electrolyte gives a progressively higher bridge reading. The concentration of carbon monoxide a t any given time is calculated from the slope of the curve drawn by the recorder. An eleetroconductometric analyzer for the continuous determination of carbon dioxide in gas mixturea, which employs a continuous flow of both gas and electrolyte through the absorp-
Vol. 17, No. 5
tion-conductivity cell as well as a recording Wheatstone bridge, is described by Smith (16). The White (19) aspirator method is used to control the ratio of gas flow to electrolyte flow. This cell has two disadvantages: (1) it is neither rugged nor compact, and (2) the accuracy of the gas-electrolyte ratio is seriously impaired if the apparatus is subjected to vibration (16).
As none of these methods appeared entirely satisfactory for continuous control of the purity of the synthesis mixture in the ammonia plant, the method described below was developed and applied to the control of the TVA ammonia plant. O P E R A T I N G PRINCIPLE
The analyzer developed consists of a preparation train, an absorption-conductivity cell, and an indicating and recording alternating current Wheatstone bridge. The preparation train removes traces of interfering gases and its construction is such that, by the proper adjustment of stopcocks, the analyzer may be used to determine carbon dioxide, carbon monoxide, or total carbon dioxide and carbon monoxide. The carbon monoxide is oxidized by an iodine pentoxide reagent and determined as carbon dioxide. In the absorption-conductivity cell the electrolyte, 0.001 S barium hydroxide, flows past a pair of reference electrodes, meets and reacts with the carbon dioxide in the gas, and then flows past a pair of measuring electrodes, The rates of gas and electrolyte flow, as well as the temperature of the absorption-conductivity cell, are maintained constant. Thus, the resistance of the electrolyte between the reference electrodes, R,, remains constant, while the resistance of the solution between the measuring electrodes, R,, varies with the concentration of carbon dioxide. The electrode pairs are approximately the same size and shape. The distance between the measuring electrodes is adjusted so that the ratio R J R , is 1.0 when the concentration of carbon dioxide is 0. The electrode pairs serve as the arms of a recording alternating current Wheab stone bridge, rn shown in Figure 1. I n this circuit the A and B end coils have the same resistance, and a jumper is connected between the alternating current source and the B end-coil end of the slide-wire. Thus the bridge gives a linear indication of the ratio R,/R, with the ratio 1.0 a t the zero end of the recorder scale. The upper limit of the recorder indication depends on the relative resistances of the end coils and the slide-wire. This limit is given by the formula: Ratio
Figure 9. Aboue.
Calibration of Cell
Firrl run. Below. Check run. Tort gas, mixlures of air and nitrogen. Elestroiyle, 0.001 N barium hydroxide
-
resistance of A end coil + resistance of slide-wire resistance of B end coil
Should the recorder fail to operate properly, the absorptionconductivity cell is connected to a manually operated Kohlrausch slide-wire and galvanometer by means of a triple-pole, double-throw switch. Since the conductivity of the barium hydroxide electrolyte is less after reaction with carbon dioxide, the resistance of the solution between the measuring electrodes, and hence the value of the ratio R,/R,, increases BS the concentration of carbon dioxide increases. The analyzer is calibrated empirically with gas mixtures of known carbon dioxide content, and the recorder scale is graduated to show the ratio value directly as parts per million of carbon monoxide or carbon dioxide. Figure 2 shows a calibration curve and a subsequent check thereof. The several concentrations are represented by vertical lines. It should be noted that the recorder pen returned to its original position when the carbon dioxide of the test gas was reduced to zero. The accuracy of the method depends on the accuracy of the accepted value for the carbon dioxide content of air which was employed for the calibration. Frequent checks indicate a precision of a t least 3 p.p.m. over the range indicated.
A N A L Y T I C A L E1
May. 1945
Figure 3.
279
Eloctroconducto
&ne lag in the sample line to i 5 seconds or less. Valve V and ell
temperature
Gonstant: between z8" and SL" 6.
The pota$sium hvdGxide-Ascrarite tube. H . rernoge's a n y &bon pentoxide reagent
100" to 105' C
The analyzer is placed in operation as follows: dioxide'is being determined. The carbin moniiide is oxidized to carbon dioxide in I , and the liberated. iodine is absorbed in the potassium iodide scrubber, M. Stopcock S1 allows the gas flow to he diverted throueh potassium hydroxide-Ascarite tube, L, and the resulting carbon dioxide-free gas is used in the adjustment of the zero point of the analyzer. From M the gas goes to the absorption-conductivity cell.
The absorption-conductivity cell is both rugged and compact and is not affected by vibration. It is construoted of Pyrex except for the electrodes and the wire used to regulate the flow of electrolyte and is operated in a constant-temperature bath. The electrodes are platinized platinum rings, 15 mm. in diameter, 5 mm. wide, and 0.2 mm. thick. The electrolyte, 0.001 N barium hydroxide, flows from a constant-head feed tank through N to coil I where it comes to a constant temperature, up through the capillary flow regulator, R, and down into the bottom of the electrode chamber, C., where i t flows past the reference eleotrodes, E,. As the electrolyte flows from C, to the lower end oi the absorption coil, T,i t is joined through U by the gas stream from the prepmation train. The carbon dioxide and electrolyte reaot while rising through T and pass into the electrode chamber, C,. The gas leaves a t the exit tube, W , and the electrolyte flows past the measuring electrodes, E,, and leaves the cell through the overtlow tip, X . OPERAIING PROCEDURE
Operating conditions for tbe analyzer for determination of minute concentrations of carbon monoxide and carbon dioxide me:
The electrolyte is stmted flowing through the absorption-conductivity cell a t approximately the correct rate, and gas is started' through the analyaer a t 8 liters per hour. The flow of the electrolyte is regulated to 15 ml. per minute by adjusting the wire in the capillary flow regulator, R (Fignure 1). The gas is passed through absorption tube L,and the measuring electrodes of the absorption-conductivity cell are adjusted by screw Y , so that the recorder indicates zero. Stopcock 9,is then turned M O O , plaoing the analyzer in operation. I n the determination of carbon monoxide done. stopcocks SI , "
~~~~~~~~
electric furnace c&trolled by a variade t"ransformer. If carbon dioxide alone is to be determined, stopcocks S,and Sa are turned 180", 80 that H and I are by-passed. Only H is bypassed in the determination of total carbon dioxide and carbon monoxide. Daily checks are made of the rate of electrolyte flow and the recorder zero, and adjustment? are made when necessary. The activity of the iodine pentoxide reagent is checked daily by dotermiding the response of the recorder t o injections of known volumes of carbon monoxide; exit gas from the water scrubbers (about 2% carbon monoxide) is introduced into tha analyzer by means of the special cross-bore stopcock, S, (5). The 10% potassium iodide solution requires replacement ahout every third day, but the reagents in the other units of the train retain their activity 3 to 4 weeks. Thc 0.001 N barium hydroxide electrolyte is standardized by electroconductance.
INDUSTRIAL AND ENGINEERING CHEMISTRY
280
PREPARATION OF IODINE PENTOXtDE REAGENT
The iodine pentoxide reagent must completely oxidize low concentrations of carbon monoxide a t comparatively low temperatures in the presence of about 75% hydrogen and in addition must maintain its activity for a long period. The following procedure, using reagent-grade chemicals, gives a reagent that has an active life of 30 to 4 0 days. Seventy grams of calcined, crushed insulating brick (Armitrong A-25; -9 +14 mesh) are added to a solution containing 100 grams of iodine pentoxide and 4 grams of vanadium pentoxide. The mixture is evaporated to dryness over a steam bath, with frequent stirring during the latter part of the evaporation. This material is then oven-dried for several hours a t 125’ to 130” C. The oxidation tube, which holds about 50 prams, is filled with the oven-dried reagent and placed in the activation train. Activation is accomplished by passing purified air through the sample ns the temperature is gradually raised to 220’ C. Tlie air is purified by passage through three scrubbers containing concentiated sulfuric acid, solid potassium hydroxide pellets? and phosphorus pentoxide, respectively. The reagent is maintained a t 220’ C. for about 8 hours and is allowed to cool in a current of purified air, after which it is ready for use or for temporary *torage. The tube must be tightly sealed during storage. The activated reagent has a bright orange color and the spent reagent a brown color. The spent reagent is prepared for re-use by suspending it in water, evaporating to dryness, oven-drying, and iictivating as previously described. APPLICATION TO PLANT CONTROL
Two complete analyzers are required to serve the two production trains in the TVA ammonia plant. Figure 3 shows the control laboratory installation of analyzers (except for control desk and recorders) for the determination of low concentrations of carbon dioxide, carbon monoxide, or mixtures of the two in the purified hydrogen-nitrogen mixture from one production train. Z t the left is the constant-terriperature bath containing the absorption-conductivity cells. To the right, a t the end of the liath, are the preparation trains. The short (front) train is used in the determination of carbon dioxide in the pas from the caustic vxubber, sample point PS-17 (see flow sheet in 7 for specified sample points), and the second train is used to determine total c.urhon dioxide and carhot1 nionouide i n the makr-up gai to the
Vol. 17, No. 5
synthesis system, sample point SS-19. The third and fourth trains are used in the determination of carbon monoxide in the gas leaving the copper scrubbers, sample point PS-15; while one train is in service, the other is kept in stand-by condition. This system of analytical control has been in operation for more than 2 years in the TVA ammonia plant and has proved very effective in preventing poisoning of the synthesis catalyst by oxides of carbon. ACKNOWLEDGMENT
The authors express their appreciation to the staR of the control laboratory, especially to R. Bowen Howard, Jr., for the cooperation received in the installation and initial operation of these analyzers. LITERATURE CITED
(1) Am. Gas. Assoc., “Gas Chemists’ Handbook”, pp. 289-95, New
York, Chemical Publishing Co., 1929. (2) Baldewyns, Joseph, Congr. pharm. Lidge, 1935, 183-5 (1934). (31 Bech. A.. Ann. huo. vubl. i n d . sociale. 1933. 376-80. (4) Berger, L. B., a-Gd‘Schrenk, H. H:, U. S. Bur. Mines, Tech. Paper 582 (1938). (5) Brown, E. H., I X D . ENG.CHEM., ANAL. ED., 14, 551 (1942). (6) Brown, E. H., Cline, J. E.. Felger, M. M.. and Howard, R. B., Jr., Ibid., 17, 280 (1945) (7) Brown, E. H., and Felger, M.M., Ibid., 17, 273 (1945). (8) Dely, J. G., private communication. (9) Edell, G. M., ISD.ENG.CHEM.,20,275 (19%). (lo) Graham, J. I., and Winmill, T. F., J. Chem. SOC.,105, 199h2003 (1914). (11) Haldane, J. S., and Graham, J. J., “Methods of Air Bnalysis”. pp. 116-29, London, J. B. Lippincott Co., 1935. (12) Katz, S. H., Reynolds, D. A., Frevert, H. W., and Bloornfield J. J., U. S. Bur. Mines, Tech. Paper 355 (1926). (13) Miller, A. M., and Junkins, J. N., Chem. & M e t . Eng.,50, No. 11. 119-25, 152-5 (1943). (14) ltotnashchenico, T. A., Zavodskayu Lob., 9, 912-13 (1940) Chem. Zentr., 1942, I, 2957. (15) Seidell, J. A., J. IXD. ENG.CHEM.,6, 321-3 (1914). (16) Stnith, A. S., IND. Exo. CHEM.,AXAL.ED.,6, 293-5 (1934). (17) Teague, M. C., J. IND.ENG.CHEM.,12, 964-8 (1920). (18) Tandaveer, F. E., and Gregg, R. C . , IND.ENG.CHEM..AXAL ED., 1, 129-33 (1929). (19) JVhite, E . C., J . Am. Chem. SOC.,50, 2148-54 (1928)
Continuous Determination of Ammonia Activity in Ammoniacal Solutions EARL
H. BROWN, JAMES E. CLINE, MAURICE M. FELGER, AND R. BOWEN HOWARD, JR.
S T H E T V A syiith(8tic :uiniionia platit, an animoniacal “copper solution” (3) is employed for the removal of harmful inipurit,ies, such as carbon inonoxide, carbon dioxidc, and oxygen, from the synthesis gas. The dolutiori contains cupric ammino, cuprous ammino, ammonium, formate, and carbonate ions, as \r.c\ll tis uncombined ammonia. There is a r:tnge of concentratioils of unrunibined nmmouia below which thcb removal of carboil dioxide froin the hynthc gas is inconiplete and above which the loss of aniirionia is excessi\,e during regeneration of the copper solution. The uncombined ammonia is an important factor also iii niaintnining the coppcr complexes necessary for tlic complete rriiioval of carbon monoxide. The concentration of uncombiiird antmonia cannot be calculated accurat,ely from chemical tleterminations of the componeiits in the copper solution, how~ Y ( ’ I ’ . for there are several cquilihrimns involved n-1iic.h have not, yt’t I)(YII rlvtermined prrcisely.
I
.in empirical formula t)risc’d on five analytical determirintionstotal animonia, formic acid, carbon dioxide, and monovalent and bivalent copper-was employed, prior to the development of the present niethod, t o give a value related to uncombined ammonia for plant control. The utilization of thr. empirical formula had several disadvantages. Values n Ith 110 strong theoretical foundation were obtained, many analytiral determinations were necessaiy, aiid considerable time was required before the results could be reported to the plant operators. This paper describes the drvelopment and application of an instrument that gives rapid and continuous information regarding the unconibined ammonia 111 the coppcr wlution. THEORETICAL CONSIDERATIONS
The thermodynamic activity of ammonia was selected as the function of the uncombined ammonia to hc> tlctcrniined, ticc:iu-e
