Production of Grained Ammonium Nitrate Fertilizer

2 Hygroscopicity is used hero to designate the rate of moisture abaorption. The conditioning .... factory seal against moisture penetration, that mois...
0 downloads 0 Views 1MB Size
Production of Grained Ammonium Nitrate Fertilizer CONDITIONING TREATMENTS AND MOISTUREPROOF BAGS A

PHILIP MILLER, G. A. LENAEUS, W. c. SAEMAN, AND M. N. DOKKEK'

rolatum, and paraffin a s coating agents for the grained salt for fertlh e r use, and concluded t h a t petroammonium nitrate produced in Tennessee Valley Authority, Wilson D a n , Ala. latum la as the best of those batch graining kettles. The I n 1940 a rosin-paraffin- coating wnditioning treatments contreatment, developed by the Hercules Powder Company for use si-ted either of 1% of a coating agent, 4% of a dust, or a with explosives, n-as patented (1); the treatment, in combination comhination of hoth. The behavior of the conditioned with a dusting agent, was later applied by this company to aninionium nitrate during prolonged bag storage and in ammonium nitrate fertilizer production ( I O ) . fertilizer distributors indicated that the combination The coating and dusting type of conditioning treatment g1vc.s treatment gibes a satisfactory product for direct use, a product only slightly lower in nitrogen content than the pure whereas a dust used alone is much less satisfactory and a salt ; this kind of product was therefore selected for use by TI'A coating treatment used alone is quite unsatisfactory. in preference to a material similar to Cal-Nitro, because of the The multiwall moistureproof paper bags were more effecsavings in bags, labor, and transportation. .4 recent paper by tive than were asphalted burlap hags in protecting amRoss and eo-workers ( 7 ) describes the resulp of small-scale studmonium nitrate from atmospheric moisturecl uring storage. ies of coating and dusting treatments applied t o spray-granulated ammonium nitrate produced in Canada, and of large-scale storage tests of Canadian and TVA ammonium nitrate, carried out during S April, 1913, when the War Department sLiddenly curtailed the same period as the present work. Ross and T e e (8) studied its demand for ammonium nitrate then being produced in the the suitability of various types of paper and burlap bags for storgraining plant of t h e Tennessee Valley Authority, tho plant outing ammonium nitrate. put became available for u.e as fertilizcr. I n view of the n.ell Up t o a certain point, both the storage and drillability prop(^knoivn tendency of ammoiiiutn nitrate to cake anti to absorb ties of ammonium nitrate improve with increased particle size, moisture from t h c attno*p!ierc, it wa* ctvident tliar its physical and material produced in graining kctt1,es is somewhat smallcr properties v,-ould have to he improved to aqsiire a satisfactory than optimum. Ross and eo-workers (7) concluded from thcir fertilizer material. Studies \wre iirirlcrtnlicn immediately t o detests that, because of its smaller particle size, grained ammonium velop a conditioning treatment that would make popsible prompt nitrate is less drillable than spray-granulated materiiil. Spray diversion of the ammonium nitrate plant output to help meet the grznulation is carried out continuously and, like other continuous pressing demand for nitrogen fertilizer. As a result of these processes proposed for ammonium nitrate, will almost cvrtainly s t u d i e , production of Conditioned ammonium nitrate was give a cheaper product than does batch graining. I3ec.ausc of started; meanwhile, research was continued on developiw and the higher cost aiid the relatively small particlr size of the grained evaluating improved conditioning methods. This paper is based product, it is likely that future installations for ammoniuni nion the,= studies. trate fertilizer production will utilize some continuous process Ammonium nitrate has long been recognizrd as a n excellent t h a t gives a coarser product. Howevcr, at present virtually all soiiree of nitrogen for all crop?, and it is also one of the most of the solid ammonium nitrate produced in this country is made concentrated forms of nitrogen fertilizer (35Yc K for the pure in graining kettles in a batchwise operation. Effective condisalt). Prior to World K a r I1 it was employed extensively as a n tioning is more important for grained ammonium nitrate than for ingredient of solutions used t o ammoniate superphosphate; subother types that have larger particle size. Furthermore, the restantial quantities of the solid salt a1.o we're uscd rommcrcially sults of conditioningstudies carried out with other types cannot be in ftsrtilizers but gcnerally i n adnLixturewith considerable proapplied directly to grained material. It is therefore believed t h a t portions of other matcrials that modified its physical properties. a report of plant-scale tests on the conditioning of grained amIn Europe granulated mixtures of animonium nitrate with limemonium nitrate, along with information on the application of constone or dolomite, known as Cal-Sitro or Kitrn-Chalk and conditioning treatments in plant operation, will serve a practical taining 16 t o 2057, nitrogen, have been marketed for some time, need. ani1 considrrable amounts were exported t o the United States I n the present investigation a number of conditioning treat(91. R i w n t l y a similar muterial has bccn produced in this ments employing a variety of coating and dusting agents xvere country (9). . h m o n i u m nitrate and ammonium sulfate in the applied to grained ammonium nitrate i n plant-scale graining f o r m of mechanical mixtures or of the double salt have also been kettles, and the products were evaluated for caking under comwid(zly usrd abroad. mercial storage conditions and for drillability under simulated C'mting of ammonium nitrate with petrolatum or paraffin field condition-, using a conimt~rcinl&$tributor. to redure moisture absorption, wit'h or without addition of small amounts of various dusts t o minimize caking, has long been GRAINIKC AMMONIUM NITRATE practiced in the explosives industry and has been proposed for fertilizer conditioning. Following World War I, the Fixed XiThe ammonium nitrate plant at Musc.1~Shoals was described trogcn Laboratory (6') made tests of rosin-paraffin mixtures, petby Fairlie (3) in 1919; it has since been modified considerably ' Present address, Office of the Quartermaster General, Military Planning t o comply with revived Ordnance safety regulations, but the process is essentially unchanged. The graining process will be tleDivision. Washington, D. C. plant-scale study was made

of the conditioning of granular

I

709

710

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. 38, No. 7

I n the studies reported here the particle size, within these limits, byas a constant factor. T h e bulk denrity of this material is about 62 pounds per cubic ioot; it is approximately t h e hame for the unconditioned material. Three t y p e of conditioning treatmenti. \vcrc invtbatigated: a water repell(bnt coating material, a water repi!llent coating plus a dusting agent, and a dusting agent used alone. I n general, the coating has the effect of reducing the rate a t which ammonjum nitrate ahsorbs mokture and of improving its drillability under humid conditions, whereas the dust reduces the caking tendency. T h e two-step trcatment is inherently more c:ffective than the coating or the dust alone, sinct each improves the product in B F i g u r e 1. A d d i n g P R P C o a t i n g M i x t u r e to G r a i n e d Ammonium Titrate different reipcct ; however, in view of the lower cort and greater simplicity of the one-step treatments, itwas desirable scribed here only in sufficient detail t o make cvitlcnt the mcihotl t o t l t ~ t ~ ~ i ~ nwi ihnc(t-h ( ~ either r step would he satisfactory by itself. of applying the conditioning trcatment. CONDITIONING MATERIALS A 1-ton batch of molten ammonium nitrate, a t about 320' F. and containing about 1.5% water, is charged to a jacketed grain-4,large number of materials were considered for use as coating ing kettle in which it cools slowly while being agitated by .slon.and dusting agents. L h n y of them were eliminated because of moving plows. As cooling progresses and the water is evapounfavorablc results in exploratory tests or because of excessive rated, crystallization begins and the material changes from a cost. Among the materials eliminated an coatings were lubricatpasty mixture to large soft lumps and finally into grains. T11e.e ing oil, oleic acid, petrolatum-rosin-clay mixture, alumina gel, grains can bc scen under the microscope as large numbers of fine and silica gel; among those eliminated as dusts wcre limestone, crystals cemented together t o form a rounded granule; they are whiting, calcium silicate slag from phosphorus electric furnaces, fully formed by the time the tcmpcrature falls t o 260" F . At tripoli, common clay, gypsum asphalt rock, cyanamide, and seathis point the moisture content has been reduced t o a constant x a t e r magnesia. Because of its availability, low cost, and wide value of 0.05yc,and the cooling rate is increawd by admitting use in agriculture, limestone^ was studied rather extensively before water to the kettle jacket. T h e batch is discharged from the being climinatcd. In addition t o providing poor protection kettle a t about 200" F. onto a belt conveyor; each graining cycle against caking, it caused some loss of ammonia when added t o requires 2.5 t o 3 hours. T h e ammonium nitrate is screened t o the hot ammonium nitrate. After exploratory tests, the followremove the 12 mesh material and is bagged in 100-pound bags: ing materials were tested thoroughly on a plant scale. the ammonium nitrate temperature at the time of bagging is Coating Ingredienta Dust about 190' F. T h e + I 2 mesh material, mostly lumps and irPetrolatum Kaolin regular fragments, is crushed and recyclcd to another kettle Rosin Kieselguhr Paraffin Plaster of Paris (CaSO4.l/aH20) batch. Gilsonite Soapstone T h e conditioning trcatment was incorporated into the above 11)tlir t e s t these rn:iterials are given the symbols P , R , P , and 0,a i d are always referred t o i n the order given here. For example. P R means petcycle in the following manner: Molten coating agent was poured rolatuni-rosin mixtures a n d R P rosin-paraffin mixtures. in at a point behind the plows near the center post (Figure 1) T h e rosin generally used in this work was FF grade wood rosin. when the batch was a t 230-240" F. Tests showed t h a t addition of the coating material by pouring was as effective as spraying in The suitability of the various grades of rosin as coating ingredients obtaining satisfactory distribution. After 10 t o 15 minute>, appeared to be governed by the "oxidized rosin" content (determined analytically as thc fraction inmluble in pctroleum when the batch tempt,rature was 210-220" F., the dusting agent c.ther). Excessive contents of oxidized rosin rcsultctl in unstable was clumped into the kettle. Samples were taken from tcn point.* in the kettle just before the dust was added and wei'e composited; PRP mixtures and also decrcascd the re4stance of the coating t o water vapor penctration. The oxidized rosin content of FF a similar compoaite was obtaincd just before tlic licttlc \vas dikgradc wood rosin (about 4%) \vas sufficiently low so t h a t stable charged. T h e proportion of fines in ammonium nitrate produced in a mixturcs Fvere obtaincd. .Uthough the niorc highly refined grades of wood rosin and thc various grades of gum ro.-in congraining kettle can be varied over a wide range by changes in tht, tained even less oxidized rosin, they did not offer suficicmt adgraining cycle ancl technique; a discussion of these changes is vantage to compensate for their higher price. On the other beyond the x o p e of this paper. For fertilizer use the product hand, B grade wood rosin (the iowest grade), which is much is made as coarse anti as f r w from fines as is practical. TVA cheaper than the FF grade, contained 40 t o GOY' of material inammonium nitrate contains a small amount (less than 0,194 soluble in petroleum ether and a o u l d not form homogeneous total) of ferric and aluminum hydroxides (from the ceramic niPRP mixtures. Because of the occasional shortage of FF grade tric acid absorption toivers) which has been found t o aid in the wood rosin, experiments me-re made on the compatibility of PRP production of coarse grains. The usual particle size range of the (1:3-1) mixtures in which one third or one half of the rosin w&s grained conditioned product follows: R grade wood rosin and the remainder was K grade wood rosin. Screen Size Per Cent U grade rosin obtained from different vendors, unlike the higher - 1 2 +35 mesh 60 t o 85 grades, varied widely- in properties ; however, all three brands - 3 5 +to0 mesh 10 to 35 testrd in t h e manner described proved incompatible. - 100 mesh 1 to 5

+

''

INDUSTRIAL AND ENGINEEHING CHEMISTRY

July, 1946

71 1

good distribution could still be obtained. These results can be explained by the fact’ that this proceUil . i t > dure would be expected t o give the least degree of nulk Screen Analysis. ‘ j . mixing of the two coatings. Density. C;. Oil - 200 l,t>,/ prr100 +2OO +325 -323 Microscopic examination of coatcd grains indinust Source (’u Ft. (;, Du-t ir,esh iiiesli nivsli rnted that the bulk of r,onting nintcrial was utilized Kaolin Thonias Kaolin Co.. Hackleburp, Ala 48 J!I 2 : .55 ‘320 in forming a protcctivc slic~il around the grains; Kieselguhr Johns-AIanville Corp., hut with the softer material., such as petrolatum (Celite 3 i V Xea Tork, N. T. 1i 1i 0 I 8 3 0 !I4 3 Plaster of I n t e r n a t i o n a l Alinarals a n d or PRP mixtures of low rosin content, there wa8 Parish Chemical Corp., Wales. T e n n . , and Armour Fersome penetration into the cracks arid fissures in tilizer Co., Colunihia, the grain. When the amnioniuni nitrate rras disTeiin. 60 40 22 3 13.7 6l.O 8oapstone Southern Tule Co., Atlanta. solved from grains that, had been coated with a 40 43 1.3 7 4 91 3 Ga. I’RP (1-1-1)mixture hut not dustcd, a continuous Garcliicr-Coleiiiali method (41. b I’urciiascd as wet hj--product calcium sulfate, ohtaiiied in the net-process rriaiiufacture -hell was left behind which enclosed a varying of phosphoric acid; dehydrated and ground a t Wilson Dam. amount of coating material of a spongclike cellular structure. If a dust had been added also. particles of dust were found enibtddcd in the coating, lowering its cohesive strcngth, with the rewlt that the coating Tn.0 grades of paraffin, spec4ietl as “light yellow crude scnale” fell into fragments whcm the ammohium nitrate was di-solved. (melting point, 124-12(i” F.) and “iymirefined 128 AMP white Harder coatings such as P R P G m i x t u r e rc~-ii.tcdt l i e abrasive accrude scale”, rcspc’rtivrtly, proved equally satisfactory. tion of the tlrist t o a cmnsiderahl(- tlc~gi‘wt i t i t f‘ailod to penetrate Petro1:ituin fwni a number of sources and varying \iidely in the interior of t h r grain. specification,*,appcaranccs, and cost was used with no appreciable This bchavioi~\ v a ~wflc-rtcd i i i t tic, r a t w at whir11 ammonium differenrc, in results. T h e material used included both “comnit rat(: absort)cd nioisturo \\.tien contlitionc~lby various methods. mercial anibel,” aiid “dark rrude” grade, .varied in color from ambvr to black, and mngcd in melting point from 118-155” F. The rate of moisture absorption \\.as dctc-rniinid by a simplified modification of tlie mrthod described by I < c ~ n e n( b ) ,in which air Somc ronsidcration was given t o the possibility of substituting a singlcj pctrolrum product for the petroleum constituents of the of mnt rollod temperature and humidity flowed over 2-gram P R P mixtuw or for the entire mixture. However, after cxplorahamplcs exposed on ~ v a t r hgla. This rate provcd to be of tory tests \\.it11 a few high melting waxes and consultation with limited signifirarice in the final evaluation of the effectiveness of the research departmcrit of a leading petroleum refiner, it was conditioning; it proved helpful, however, in obtaining some insight into the specific behavior of the conditioning agents when concluded t h a t neither possibility was promising. Gilsonite, one of the asphaltites of petroleum origin, is a natudifferent procedures were used. The rate a t which the grains rally occurring hard black material of high melting point, mined absorbed moisture was reduced to about 8-1076 of its original in C t a h . The gilsonite used in the present work was all of the value (relative hygroscopirityz %lo%) by applying a coating of grade known as selects and had a melting point of about 330” F. 1% by weight of pt,trolatum. When 4% kaolin dust was added Except for plaster of Paris, which was prepared by TVA from on top of the petrolatum coating, the rclativc rate increased to by-product calcium sulfate, the dusts were purchaytl under about 20%. A PRP coating (1%) used alone was more effective specifications of 97YC -200 mesh particle size, no extranvous than petrolatum. giving a relative hygroscopicity of 25’7”; gritty material, and not more than 1% moisture. The particle ho\rever, addition of the dust increased the value t o 40-60%’,.’ size \vas clctermined by a wet screen method. T h e grit test was Evidently the semifluid nature of the petrolatum maintained a made by rubbing a linseed oil-dust paste across a smooth glass continuous coating even after penetration by the dust particles, plate \\it11 a spatula. Measurements of th: apparent bulk den\\hereas the PRP coating, with a more skinlike structure, was sity and ttic CIII absorptivity were made t o obtain a preliminary ruptured by the dust, which provided gaps for the entrance of evaluation of the cffcctiveness of a d u s t ; usually, although not moisture. Varying the amount of dust from 1 to 4yo made no alwayy, lo\\. bulk density and high oil absofptivity indicated good difference in the extent t o which the moisture absorption rate conditioning propvrtiee. Tablc I contains t h f w d a t a for some was increased by dust addition. Ammonium nitrate grains with dusts. a PRPG coating, which was much liarder than the PRP, had a relative hygroscopicity of about 20y0. This comparatively high BEHAVlOR OF MATERIALS initial rate was attributed to failure of tkie coating t o penetrate the interior of the grain; thereby a large internal area of ammoFollowing are some qualitative and senii-quantitative observanium nitrate was left exposed to water vapor, which penetrated tions on tlie behavior of the conditioning materials during and the outer shell. On the other hand, the greater resistance of the after application. While t h e molten PRP coating mixture was outer coating t o abrasion reduced somewhat the increase in moisnormally added a t ammonium nitrate temperatures of 230ture absorption rate which was caused by the addition of dust. 240” F., satisfactory application was alio obtained when the mixOn the basis of this behavior a two-step coating treatment was turc (at 200” F.) was added t o the graining kettle a t a n amtried. A small amount of a soft material (0.2 t o ‘o,!% petroiamonium nitrate temperature m low as 120” F. .lddition of thc: tum or P R P ) mas added first t o penetrate the interior of the grains coating mixture a t ammonium nitrate temperatures abovc th(3 and t o form a thin surface coating. This was followed by addicrystal transformation temperature of 258’ F. resulted in a less tion of a larger amount (0.6 t o 1.0%) of a harder coating (RPG or effective coating. T h e effectiveness of the coating and its uniB gradnnFiiiiiuiri S i t 1'51 t e , C h a T Ipioil ~ 13rn nd. so. 1 s o . 2 .uo. 3 N o . 4 A v . 1 4 . 4 1 3 . 2 1 2 . i 15 2 1 3 . 9 9 3, 8 9 10.8 1 0 . 0 11.2 9.8 8 , 7 6 8.2 6.8 2 0 8.1 3.1 0.0 3.3 2.3 2.6 4.7 3 2 20 6 8 . 8 2; 3 . 4 13 9 100 100 100 100 100 i8 io 70 71 72 47 67 GO 64 60 0 61 0 . 8 3 0 13 0 . 3 8 0 . 5 7 2.05 2 . 6 0 1 . 6 8 1.20 1 . 8 8 3 47 3 . 2 3 2 94 1 . 7 i 2 . 8 6

Exposed at 79' t o 8 l 0 F. a n d i 9 to 8 1 5 relative hunlidity. Relative drillabilitj. is the ratio 100 X r ~ t ~ ~ e

(

. ~ \ l h l O S I C . l l KITRATE A S D C H I L E A S

~~

f

~

~

~

e

e

~

~

~

~

~

~

~

e

July, 1946

INDUSTRIAL AND ENGINEERING CHEMISTRY

i l u q t cauw.1 a greater cliwcascx than dit1 ZFc. .I d w r e a w i i i amount of coating used Ivith a dust ai-o c a u ~ e da tlccrcaw i n rttlntive drilhbility. XIatcriah conditioned with a d w t alone slio\~-cclr:itlier j ~ o o rdrillability, not niucli different from that of rinconditioiiecl nnnnonium nitrate. 0 1 1 tlic other haiitl, Ross and co-workers ( 7 ) found that spray-granuhteil amnionium nltrate of large particle size had sxtisfactory tirillability \vlien conditioned with a dust only, wliicli inilicatc,.s tlie important cffect of particle size on drilhbility. Four trestments in which 1'1 PRI' and 4% dust uw-e u v t l appeared quite satiifactory. Of these, tlie treatment using soapstone showed up somewhat better than the other three (with plaster of Paris, kaolin, and kieselgulir) which were very close together. On the other hand, when bentonite or Diluex ivas the dust, moisture absorption \vas rapid and drillability was very poor. Lse of 17, of a PGP mixture (gilsonite substituted for roTin'! re-ulted in somewhat poorer drillability than corresponding treatments using PRP. T h e effect of the particle size of ammonium nitrate on drillability is shonn by data for the last two tests in Table V. Ammonium nirrate ]vas screened to give +20 and - 20 mesh portions, and tlicee myre conditioned with 1% PRP and 4yo kaolin. Tlie +20 niesh Inaterial showed the best drillability of any of the ammoniiinl nitrate materials tested; t h e -20 mesh material was not far from the poorest and gave lower relative drillability than unconditioned and unscreened ammonium nitrate. These results indicate the brncficial effect on drillability of large particle size and freedom from fines. .-Ispecial test was made Jvhich t h r e n some light on the significance of the relative drillability value. When Pome of the f 2 0 rial, exposed 48 hours and giving a relative drillability of 30$, \vas ruii through tlie drill again after tlie drill speed had been tripled, the actual drilling rate was nearly tripled. Wlicn tlie sanio tczt \vas made with the -20 mesh material, which had a relntive clrillability of 15% after 48 hours, the actual drilling rate incrca-c{l only about 50%. These results suggest that relative dril1:iliihty values above 40% show the exposed material t o be still ~1rillablea t controlled rates, whereas values under 20$, show tliv opposite. Iii aildition to actual and relative drilling rates, Table V contain< data o n the moisture content of the material, as drilled, and on tht: relative hygroscopicity. One of the chief points is the poor correlation between relative hygroscopicity and drillability. I n the early stages of this Tvorlr and in work by others, considerable emphasis was placed on relative hygroscopicity as an indication both of storage and of drillability properties to be expected. T h e final results of the present work showed that thi? measurement can often be misleading. Thus, material treated with 1% petrolatum plus 4% kaolin (data not shown in Table V) had a relative hygroscopicity of only 20%; it was thought for a time t h a t this treatment would be effective because of its simplicity and becauqe of promising results in preliminary storage tects. I t was found, however, t h a t the drillability of material conditioned by this treatment decreases sharply with absorption of moiyture. T h e treatment was concluded to be unsatisfactory. Tied in Ivith this behavior is the fact t h a t t h e amounts of moisture absorbed by the 50-pound samples do not agree \vel1 n i t h the relative hygroscopicity values determined in the hboratoxy. JThile the causes for the difference w e r e not investigated, it appearcd t h a t there pvere significant differences in the Tvettability of the differently conditioned materials, which depended chiefly 011 tlie coating material used b u t also on t h e nature of the dust. -Apparently, in the ease of less wettable materials the moistuw remained largely on the surface and actually formed pools of clear liquid in some cases, whereas v i t h more wettable materials the moisture penetrated readily into the interior of the 50-pound samples. This was particularly notewcithy in the case of the petrolatum-kaolin treatment, wherein the entire sample appeared t o become wet and sticky after the absorption of a relatively small

4Fc

tlle

117

amount of inoi>tiire. This bc,liavior also influenced stor:lgi, tc'ht r e d t s , in ivliich tlie petrolatum-kaolin treatment gave relarivt~ly poor protection against caking after prolonged storage. This reversed tlic short-timc storage re.;ults. IVith certain trcatmcnts tlie relative hygroscopicity of the freshly cuiiditioned matcrial ~ v a qlower t!ian that of material which had been itored xvcral clays. Laboratory determinations of hygro%copicity on small samples are uaeful chicfly as an indication of ( 0 ) tlie effectiveness ivith wliich a treatment is applied to a given batch and ( b ) the uniformity of the batch. Comparison of the relative drillability valucs and the moisture content of the materials a t time of drilling, after 24 hours of exposure, s h o a s reasonably good concordance. Tlie results should be viewed with reservationr since, in practice, fertilizer materials will be subjected to fluctuating rather than uniform humidity conditions, and the degree of exposure is unpredictable. Severtheless, tlie behavior of the 50-pound samples may b,e taken as a fair indication of lvliat to expect in practice. CONCLUSIONS

On the basis of storage and drillability test results, supplemented by experience with commercial shipmmts, i t {vas eoneluded that ammonium nitrate produced in graining ket tles can be effectively conditioned n i t h ICo of a petrolatum-ro4n-paraffin mixture plus 4yoof onc of several dusts; Ivith this treatment it should remain in satisfactory condition lor direct application when stored i n wirable moistureproof papc'r hxgs in staclk.5 twelve bags high for a year, uniler thcx fairly uiif:ivorable climatic collditiom prevalcmt at Kilson D a m . The trcatmcnt adopted for plant uie con-ijtetl of 1c7, PRP (1-3-1) plus 47, kaolin. This treatment {vas u-ecl also, under tlie technical kupervision of TV;Z, in four Ordnance graining plants t h a t produced animoniuni nitrate fertilizer in 1043 and 1844. Several hundred thousand tons of grained ammonium nitrate conditioned by this treatment have been used successfully both in mixed fertilizer and for direct npplication during the past two years. Water repellent coatings, without a dust, do not afford adequate protection against caking. Tlie use of dusts alone for conditioning grained ammonium nitrate i;: not recommentled; although caking i n storage is effectively reduced, the dri1ht)iIity of the material under humid conditions is not good. It seems likely that a dusting treatment would be satisfactory for grained ammonium nitrate from nhich the -20 mesh material had been screened. I n the tlvo-step treatment kaolin, plaster of Paris, and soapstone are about equally effective as dusts. I in keeping ammonium nitrate fertilizer in good condition during prolonged storage in humid climates. Of a number of commcrcially available bags tested, multiwall paper bags with one or tlvo moistureproof plies gave better protection than asphalted biirlap hags. Sewed or valve-type closure was superior t o Tvire-tiod cloirire.

718

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

Storage tests of bagged material for 6 to 12 months under factory conditions are recommended for final evaluation of t h e storage properties of ammonium nitrate fertilizer. For material t o be applied directly, d a t i v e drillability as determined in the present work is believed t o be a useful guide. Laboratory determinations of rate of moisture absorption, in small samples of conditioned ammonium nitrate, are of limited significance in intlicating the'potential value of the conditioning treatment used ACKNOULEDGMENT

The authors are indebted to .I.31. AIiller, R.I;.Copson, J. H. Waltliall, a n d ' s . A. Harvey for helpful criticism, advice, and encouragement. hcknoivledgment is made t o J. E. Chenevey, W. L. narrow, J. Stalcy, H . J. Strauss, D. E. Bordelon, and S. P. Clark, d l of wliorn participated i n the experimental work.

Vol. 38, No. 7

LITERATURE CITED

Cairns, It. W., G. S.Patent 2,211.738 (1940). Chenevey, J. E., Chem. R. N e t . Eng., 52, No. 8, 115 (1945). I b i d . , 20, N o . 1, 8-17 (1919). (3) Fairlie, Ai.W., (4) Gardner, H. A , , "Phy.?ical and Chemical Examination of Paints, Varnishes, Lacquer and Colors", 7th ed., pp. 541-5, Washingt o n , Inst. of Paint and \-arnish Research, 1935. 31, 903-8 (1939). (5) Kc,enen. F. G . , ISD.ESC. CHEM., Chem. & M e t . E ~ Q . , (6) Kmse, H. J., l-ee. J . Y.. and Braham, J. M., 32, KO. 2, 241-3 (1925). ( 7 ) t l o s s , W. H . , Adanib, J. H.. Yee, J. Y . , and Whittaker, C. IT,, I s m ENO.CHEM.,36,1088-95 (1944). ( 8 ) Roy;;, \I7, H., and Tee, J. Y., A m . Fertilizer, 102, No. 7 (;ipril 7, 1945). (9) Titlestad, N.ii., I b i d . , 101, X o . 10, 9 (1944). (10) War Production Board. Chemicals Div., private commuiiication, .4pril 26, 1943. (1) (2)

Uncatalvzed Reaction of Natural Gas and Steam J

.ALVIN S. GORDON Central Experiment Station, U . S . Bureau of Mines, Pittsburgh, P a .

T h e uncatalyzed reaction of natural gas with steam has been studied between 1225' and 1516" C. with steamnatural gas ratios of 1.5 and 5, and time of contact between 0.21 and 4.6 seconds. Even at high temperatures, long contact times, and steam-natural gas ratio of 5 there is some carbonization of the natural gas. The extent of carbonization is shown to be a fuiiction of the ratio of steam to natural gas, and to Fary erratically with temperature and time of contact. As the temperature is increased, the percentages of natural gas undecomposed in one pass through the furnace become less dependent on contact time (in the range of contact times studied), until at about 1500"C. there is little or no effect. At any tempeiature the percentage of undecomposed gas shows little if any dependence on the ratio of steam to natural gas

T

HE catalyzed reaction bc.tn.een steam and natural gas has

been studied by a number of investigators (1, Z), and industrial processes arc>being used t o produce a mixture of hydrogen and carbon monoxide by this reaction a t 750' t o 900" C. Sickel seems to be the best cata1y.t; it i,? usually supported on some refractory with a promoter wcli a i magnesium oxidr, aluminum oxide, or chromium oxide. Cheap and efficient heat transfer in gas systems, obtainable by the use of pebble furnaces, furnishes a n attractive industrial posgibility for carrying out the uncatalyzed natural gas-steam reaction. Temperatures u p to 1500' C , may be used with Alundum pebbles. The uncatalyzed reaction has been studied by Karzhavin (3) who states: "The rate of reaction of methane with steam is very low mid only at temperatures above 1300" C. does it become sufficiently high for industrial use." K O data are supplied t o support this statement. T o obtain d a t a which can btl used for design of pilot plant installations, a laboratory investigation of this reaction at high temperatures was undertaken, and the reaction conditions w r r e limited to those significant for iritlustrial conditions. Figure 1 is a sch(.matic drawing of the apparatus. S a t u r a l gas i.\ admitted to the system through flowmerer A , and then goes iiito wattir saturator R , which consists of a glass distilling flaqk

with a heating coil resting inside and a t the bottom of the bull, The current t o the coil is controlled by a variable voltage transformer (T'ariac). T h e gas is admitted to the base of the asbestos lagged column, Khich is packed with a tightly rolled coil of stainless steel screening t o increase the contact' surface of water with gas. From temperatures read on a thermometer placerl with thc bulb a t the height of the exit tube, the water vapor pressure of the reactant gases is determined. The total pressure is one atmosphere. The difference between atmospheric pressure arid the vapor pressure of the water gives the pressure of the natural gas. In operation the exit temperature is controlled by the current supplied t o the heating coil. T h e saturator is similar to the one used by Hawk, Golden, Storch, and Fieldner ( I ) . The degree of saturation of the gas by water vapor was determined by trapping a known volume of the natural gas-steam mixture in a boiling-water thermostat, displacing the water vapor into a drying tube, and determining its weight. This method rhoiwd t h a t the gas was s a h r a t e d with water vapor nrithinexperimental error, even at the highest flow rates of natural gas u ~ r t i in this research. Some of the data are included in Table I . The flow rates of the gase* were incawTed lvith a capillary flolvmeter.

nllip. Bt

llxi: Tuhe. C. 85 ti 85 8 85 8 94 2 94.2

'pornl I'ressurr. AIni.

lqou.R~~~ of Nntural (>ah, r c . :set.

740 743 745 739 739

1 14 1 30 1 30

1 23 1 18

Ratio, Steitrn/Natural Gas lheoretical Observed 1.50 1.4i 1,50 1 20 1 50 1.25 5.00 4 31 5 00 5 3:

From the saturator t h r gases are let1 into the reaction vessel through tubing heated by resistanw wire. I n the earliest experiments this vessel was quartz, with quartz-Pyrex graded seals in the cold portion at each m d . Quartz \vas not used because i t devitrifiw rapidly at temperatures o w r 1000" C. and is permeable to hydrogen : consequc~ntlya porcelain brssel \vas substi-