Improved

Urea Process

Developed

A new technique,, already incorporatedinto two commercial 'mstallations, has re-duced steam and watt* requirements, low-ered irtvettment and operating costs,,andprovided reliable plant operation.

I. Mavrovic, Consulting Engineer, New York, N.Y.

An improvement in urea processing that increases re-actor conversion to from 65% to 72% will provide for asubstantial reduction in investment cost for a completerecycle plant. In addition, a new method for heat re-covery is included which lowers requirements for steamand for cooling water.

Covered by a number of patents and patent applica-tions both here and abroad, Table 1, the new technologyis called the Heat Recycle Urea process. Its developmentcame about as a result of observations in urea manu-facturing over the past eight years of consulting engineer-ing practice. Peculiarities were noticed in medium andlarge urea plants in operation or in upset conditions.These led to new ideas and new concepts in the processdesign and to their practical application in commercialplants.

Right from the beginning it became apparent that inthe liquid carbamate recycle type urea process the keyfactor is the overall conversion of C02-to-urea per pass inthe reactor: the higher the conversion rate of C02-carba-mate to urea, the smaller the amount of unconvertedcarbamate in the reactor effluent and consequently thesmaller the size of the decomposition, absorption andrecirculation equipment. Thus, the importance of attain-ing a high conversion per pass in the reactor is quiteobvious.

In 1948, Frejacques determined values for the equi-librium constant of the NH3—H20—C02 and urea sys-tem with respect to temperature, on the basic assumptionthat the urea formed in the reactor is in direct equi-librium with the ammonium carbamate formed in thereactor from NH3 and C02. For many years these valueswere used to determine the maximum conversion onecould attain in a reactor, according to Frejacques' basicequilibrium constant formula:

K =x(b + x ) ( l + a + b - x)

(1 - x)(a - 2x)2 (1)

where: K = equilibrium constant,a = NH3-to-C02 molar ratio in reactor feed.b = H20-to-C02 molar ratio in reactor feed.x = C02 as formed urea to total C02 in reactor

feed, molar ratioAt 370°F, Frejacques' value for K is about 1.4. From

the observation of the performance of urea reactors incommercial plants operating under some upset condi-tions quite different from the original design conditions,it became apparent that at a given temperature a muchhigher conversion per pass could be attained than pre-dicted by equation 1.

These findings were reported in an earlier article (1)and a graph was presented, (see Figure 1) which is theexact graphical representation of the above cumbersomeequation 1. Handy scales No. 1 and No. 3 are given onthe graph for quick determination of the actual K valuedetermined in relation to temperature.

For example, by substituting into equation 1 a set ofactual values for a, b, and x obtained in a commercial

Table 1. U.S. patent coverage on the HeatRecycle Urea process.

Patent TitleU.S. Patent

No.

Heat Recycle Urea process 3,579,636Air Injection in Liquid Ammonia 3,574,738CO2 Injection in Absorbers & Decomposers.. 3,759,992

PatentApplication Title

U.S. PatentApplication

No.

Concentration Analyzer 88,272Isothermal Reactor 190,519Decomposer Improved Design 237,644Ammonia Recovery Stripper Not Yet IssuedPrilling Tower Not Yet IssuedPulsation Dampener for Carbamate Pump ..Not Yet IssuedUrea Process with Low-Pressure CO2 Feed ... Not Yet IssuedLiquid Distributor for Decomposers Not Yet Issued

Licensor: Ivo Mavrovic, New York City.Exclusive licensee: Technip, Rueil-Malmaison, France

Technip, Inc., New York City.

127

plant, respectively 4.0,0.70 and 0.72, the back calculatedK is equal to 2.9, instead of Frejacques' value of 1.4.

It is obvious from the above data that at a high K valuea high conversion can be attained in the reactor with lessexcess ammonia and more water. To achieve this result,however, the reactor must be properly designed withrespect to all the parameters involved. With this new

knowledge it became possible to design a urea plant ona much more economical basis.

isothermal reactor

It was observed that in a standard plug flow reactor,where the feed streams entered at the bottom and the

FIND EQUILIBRIUM UREA YIELD,SCALE N0.4

% CONVERSIONX.IOO

SCALE NO-5

SCALE NO 3

EQUILIBRIUMCONSTANT

K

4.0- -

3IO--

320--

330 - -

34O--

360-

370--

390--

3.0--

SCALE NO g0j FEED MOL RATIO

b

REFERENCE POINT O3P

K . X ( b * x ) [ | . g « b - x )(i-x)(a-2x)z

K=£QUILII3RIUM CONSTANT

X- - CONVERSION. ffi^ g*

a = g^ FEED MOL RATIO

b-S|5 PEEO MOL RATIO

05-*-

STEP NO. I. SELECT OPERATING TEMPERATURE (t) ON SCALE NO.I

STEP N0.2. DRAW A STRAIGHT LINE FROM (t) THROUGH THE FIXED REFERENCE POINT (P) TO 'INTERSECT OCALE NU3. READ THE VALUE FOR K AT INTERSECTION.

STEP NO 3. SELECT VALUE FOR (dl ON SCALE N0.5 AND VALUE FOR Ibl ON SCALE N0.2.CONNECT (0) AND (b) AND MARK THE POINT OF INTERSECTION OF THIS LINE WITH THE

APPROPRIATE REFERENCE CURVE CORRESPONDING TO THE SAME VALUE OF <D\

STEP NO.I. DRAW A STRAIGHT LINE FROM (K\ AS DETERMINED IN STEP N0.2 THROUGH THEREFERENCE POINT AS OETERMItsED IN STEP NO 3 AND INTERSECT SCALE N01READ THE VAI UE I-OR CONVERSION (X) AT THE INTERSECTION.

FOR A SPECIAL CASE WHEN (b)rO, (CM MAY. BE READ DIRECTLY ALONG THEREFEUCNCE CURVE FOR D = o THUS ELIMINATING STEP NO.Î.

-Pi—,_0'-

EXAMPLE:

FOR! I -- 365°Fa-3.6b--o.s

K =2.06X : 67%

Figure 1. If reactor temperature is known, this nomograph will provide equilibrium conversionfor recycled ammonia and carbon dioxide to urea.

128

reactor effluent was discharged from the top, the tem-perature of the reactor mixture was invariably higher atthe bottom than at the top. This is quite understandable,based on the fact that the reaction of NH3 and C02 toammonium carbamate in the bottom section is stronglyexothermic and almost instantaneous, and the subse-quent formation of urea from ammonium carbamate issomewhat endothermic. As a result, the exothermic heatwas being stored as sensible heat in the synthesis mix-ture, and it was later slowly released to supply endo-thermic heat required for the dehydration of carbamateto urea throughout the reactor.

As a consequence, carbamate formation in the bottomhot section of the reactor was hindered, and it was shiftedto higher zones of the reactor. At the same time, the rateof formation of urea was reduced in the upper reactorsection due to the gradual adiabatic cooling of the syn-thesis mixture.

It was discovered that by operating the reactor bottomsection relatively colder, and by supplying heat to theupper section of the reactor, the conversion increasedconsiderably. One method of attaining this, which elimin-ates the need of internal heating steam coils, is to installan open-ended coil in the reactor and to introduce theG02 and carbamate streams at the top of the reactorthrough the coil. As this combined stream flows down-wards through the coil, part of the heat of reaction ofNH3 and C02 to carbamate is released, thereby heatingthe urea synthesis mixture.

The ideal condition of an isothermal reactor, counter-currently heated, is approached, resulting in a relativeboost in conversion.

The direct benefit of an increase in conversion can beillustrated by the following example. The heat load due tocarbamate decomposition and absorption within the de-composition, absorption and recirculation sections of aliquid carbamate total recycle, 700 ton/day urea plant,operating at about 65% conversion per pass in the re-actor, is exactly the same as the corresponding heat loadin a 1,000 ton/day total recycle urea plant with liquidcarbamate recycle operating at about 72% conversion.

Because the relative size of the equipment is approxi-mately proportional to the amount of carbamatehandled, it is possible in practice to increase the capacityof a 700 ton/day urea plant, operating at 65% conversion,to a 1,000 ton/day urea plant by simply boosting thereactor conversion from 65% to 72%. Of course, addi-tional C02 and NH3 reactor feed capacity would berequired in this case.

In most urea plants, the reactor is lined with stainlesssteel, the most economical material for this use. However,the surface of the stainless steel liner in contact with theprocess mixture must remain passivated with oxygen tominimize corrosion. For this purpose, air was usuallyadded to the C02 before compression to reactor pressureso as to maintain a concentration of 02 in the reactor atabout 2,500 parts/million based on fresh C02 feed.

By observing the operation of several such urea plants,it was found that the air supplied to the urea synthesisreactor via the C02 gas was not readily dissolved in theurea synthesis mixture, and it had the tendency of escap-ing toward the upper section of the reactor together withC02 gas. As a consequence, the bottom part of the reactorliner was not adequately passivated, and corrosion of theliner often resulted. To be effective, an undesirably highamount of air was used, which it was noted became quitedetrimental to the conversion rate of C02 to urea. Thiswas due to improper mixing of C02 with NH3 to formcarbamate and to the relatively high carbamate vaporpressure in the reactor.

This problem was solved by adding air to the liquidammonia stream before it entered the reactor. Takingadvantage of the fact that air is moderately soluble inliquid ammonia, it is possible to attain much higher 02concentrations in the liquid phase with much less air, andtherefore prevent corrosion of the bottom part of thereactor liner. At the same time, it was noted that withrelatively less air in the reactor the conversion wasboosted considerably.

Electrical power consumptionis reduced

The fresh C02 makeup is usually compressed to reactorpressure and introduced into the urea reactor for con-version to urea. The compression of C02 to reactor pres-sure requires a great deal of power. Almost 70% of thetotal power consumption within a urea synthesis processis used by the C02 gas compressor.

A new system,' recently tested in a commercial ureaplant, makes it possible to feed almost 40% of the totalfresh C02 feed directly into the low-pressure stages ofthe decomposition and absorption stages of the synthesissection and to recycle it as carbamate solution into thereactor, whereas only the remaining 60% of the fresh C02feed is compressed to reactor pressure and directlyintroduced into the reactor.

The saving in compression power is quite substantial,because only a minor fraction of the electrical powerrequired to compress a given quantity of C02 gas to re-action pressure is used in pumping C02 in an ammoniumcarbamate solution.

The idea of feeding part of the fresh C02 feed to thelow-pressure section of the absorption system does notrepresent a novelty per se, because it has been triedbefore. This technique had the drawback of being diffi-cult to control the concentration of the carbamate solu-tion formed by reaction of C02 with ammonia. Frequentfreeze-ups of the carbamate recycle solution in the associ-ated equipment were the main cause of plant shutdownsand of consequent poor plant efficiency, because time-consuming analytical methods were normally used to de-termine the carbamate solution concentration.

Following the development and debugging of a con-tinuous carbamate solution concentration analyzer,which will be described later, the low-pressure C02 in-jection technique was very successfully used in a com-mercial scale urea plant, and a considerable saving inutility consumption was attained.

Improved decomposition withlower biuret

In urea synthesis processes operating above 3.5- to 1.0NH3-to-C02 molar ratio it was noted that another usualcause of relatively low reactor conversion was the factthat an excess of water was unavoidably recycled to thesynthesis reactor along with the unconverted ammoniaand C02. This was because in the first decompositionstep, too high a proportion of the water from the reactoreffluent was carried along with the recovered gases, NH3and CÛ2, which are usually condensed in a body of anammoniacal aqueous solution of ammonium carbamateand recycled to the synthesis reactor.

In some urea processes, this problem is avoided byrefluxing the total amount of relatively colder reactoreffluent counter-currently to the decomposed gases con-taining excess moisture, in order to condense some

129

CONDENSER H.P CARBAMATE

RECYCLE PUMP

Figure 2. Flow diagram of the Heat Recycle Urea process

moisture out from the gas. This system has the disad-vantage of considerable biuret build-up in the reboilertype decomposer which must be used in this case. Thus itaffects the product quality and the cost of the additionalequipment required to remove part of the biuret from theproduct before prilling.

A novel decomposition system that was devised, andwhich will be described later in this article, substantiallylowers this water content in the decomposer off-gas andat the same time yields a urea product with a relativelymuch lower biuret content. This is accomplished by by-passing a minor portion of the reactor effluent around theforced-feed, high-velocity, vertical decomposer, and byintroducing this minor portion into the upper section ofthe decomposer separator in counter-current contact withthe moist decomposer off-gas.

A substantial portion of the water vapor contained inthe decomposer off-gas is condensed, and the urea solu-

tion product can be used to produce low biuret urea prillsby the very inexpensive method of direct evaporation,instead of the relatively expensive crystal-meltingmethod. For instance, it is possible to produce 0.6-0.7wt. % biuret product by direct evaporation, thus elimi-nating the need for costly crystallizers, centrifuges, andassociated equipment. From experience it appears that aurea prill product containing 0.6-0.7 wt. % biuret is verywell suited for the production of synthetic glues. Forhigh-grade technical urea, with 0.25 wt. % urea, the con-ventional crystal melting route must be used.

Concentration analyzer forcarbamate solutions

In reducing the water moisture content in the de-composer off-gas and in feeding fresh C02 gas to theabsorption and recirculation stages, it became apparentthat a reliable method of monitoring and controlling the

130

TO VACUUMCONDENSER

CONCENTRATOR

SEPARATOR

UREA

CONCEN-TRATOR

INERTS

FROMRECOVERYSTRIPPER

-TOSCRUBBINGANDRECOVERY

_j. 86-88%-v^UREA

SOLUTION

UREA RECYCLE

PUMP

water content of the carbamate recycle solution wasrequired to prevent crystallization of the carbamaterecycle solution. This problem usually leads to obviousconsequent mechanical problems in the high pressurecarbamate pumps and associated equipment handlingthe recycle solution.

Occasional laboratory analyses of the stream are notsatisfactory because of the usual six to seven hours timelapse between solution sampling and final analyticalresults. A very reliable method of obtaining continuousand instantaneous concentration readings and freezingpoint readings of the carbamate recycle stream wasdeveloped, and an apparatus based on a novel principlehas been in continuous, trouble-free operation for thepast two years in a commercial urea plant. It is now inuse in two urea plants, and it has been recently installedin two more.

The concentration analyzer consists of a very simpledevice of standard construction, with extreme sensitivity

and accuracy. It has no mechanical moving parts in con-tact with the process fluid; and its obstructionless sensinghead is mounted directly on the carbamate piping carry-ing the solution to be analyzed.

The electrical signal from the sensing head is amplifiedin a panel-board mounted console and transformed into aproportional pneumatic signal, which activates a stand-ard recorder/controller. The scale on the recordinginstrument is directly calibrated in terms of temperatureunits. The reading on the instrument is proportional tothe concentration of carbamate in solution. It indicatesthe salting-out temperature of the carbamate solutionanalyzed.

The output air signal from the concentration recorder-controller is used to automatically control the waterbalance in the system by varying the pressure in thesecond separator and, consequently, by varying the watervapor content in the second decomposer NHs—C02off-gas. As a result, the operator can maintain an opti-mum water balance in the synthesis, decomposition andabsorption system without fear of carbamate salting-outproblems and equipment stoppage.

Heat recovery is major concern

In the development of this urea process, another areaof prime concern has been the conservation of heat withinthe system. In virtually all urea plants, decompositionof the unreacted carbamate contained in the reactor ef-fluent was accomplished by indirect heating with steam,and the resulting NH3—C02—H20 stream was thencondensed in indirect heat exchange with cooling water.

From an overall heat balance calculation on these ureaprocesses, it became obvious that the heat supplied to thesystem was very uneconomically rejected to the coolingwater. Bearing this in mind, with a given overall tem-perature rise it can be said that the specific cooling waterconsumption of a urea synthesis process per ton of ureaproduced is the gauge of its thermic efficiency and heatrecovery.

Over the years, field data were collected and analyzed,from which it was possible to develop practical con-densing curves for the NH3—C02—H20-urea system andto design a novel process in which most of the heat ofcondensation of carbamate could be used to decomposethe unconverted carbamate and to preheat the reactorfeed streams for thermic equilibrium. The major part ofthe heat of condensation of carbamate was recycled backinto the system, thus the name "Heat Recycle" Ureaprocess.

By this technique, the cooling water consumption waslowered from the usual 17,000-gal./ton level to a muchlower value, of about 13,000 gal./ton. As a result, thesteam consumption was also reduced, almost porpor-tionally from the usual figure of 1,800-2,000 Ib./ton toabout 1,200 Ib./ton. Due to the relatively low tempera-ture at which the carbamate decomposer is operated inthe Heat Recycle Urea process, 150 lb./sq. in. gauge levelsteam is used instead of valuable 300-600 lb./sq. in.gauge level steam.

Contrary to prior urea synthesis processes, in the HeatRecycle Process the amount of excess ammonia requiredfor optimum conversion in the reactor is not separatedfrom the carbamate recycle solution, condensed to liquid,and recycled into the reactor. Rather, it is totally con-densed and dissolved in the carbamate recycle solution.Thus, the whole system for excess ammonia condensationand recycle has been eliminated at considerable savingin equipment cost.

131

Stainless steel used in construction

Standard 316 L, and 304 L, stainless steel is used forthe equipment of a urea plant designed on the basis of theHeat Recycle Process. The relatively high excess am-monia used appears to be minimizing corrosion problems,whereas actual data gathered so far indicates that thelow NHs-to-COa mole ratio urea synthesis processes havea higher rate of corrosion. The iron content in the un-treated urea product solution from the urea synthesissection is usually taken as a relative measure of the cor-rosion rate throughout the synthesis and decompositionsystem. The Heat Recycle Process rates well with only0.4-0.6 parts/million iron content in the urea productsolution.

In the synthesis of urea from NH3 and C02, 18 Ib. ofwater are formed in the urea synthesis reactor per 60 Ib.of urea produced. Quite often the common problem in allurea processes is the disposal of this amount of processwater which is contaminated with residual NH3 andnitrogen compounds, whether it is released to theatmosphere as water vapor or rejected to the sewer in theliquid phase.

A novel and relatively economical recovery stripperwas developed for processing the waste effluent streamfrom a urea plant. It reduces the content of residual NH3in the waste liquid stream to about 20 parts/million. Theoverhead gas containing NH3 is totally recovered in theurea synthesis section. All gaseous vents from the ureasynthesis section are scrubbed with water, and the weakaqua solution is sent to the recovery stripper for NH3recovery. The novel design of the prilling tower in-corporates a wet dust abatement system to remove thesolids from the exhausted air to acceptable standards tomeet Environmental Protection Agency (EPA) require-ments.

The improvements described above, and many otherrefinements not mentioned in this article, have beenincorporated into the design of this new process, resultingin a considerable overall reduction in production costs.The main contributing factors are 1) lower investmentcost, 2) lower operating cost, 3) reliable plant operation,4) simplicity of equipment, 5) conventional plant oper-ating conditions, proven in many years of experience,and, 6) very economical waste effluent treatment system.

Process description

Preheated carbamate recycle solution and gaseous C02are introduced into the reactor at its top through theinner reactor coil with open end (see Figure 2). They aremixed with preheated liquid ammonia introduced intothe bottom section of the reactor.

About 73% of the total C02 present in the reactor isconverted to urea. The major portion of the reactor ef-fluent is reduced in pressure and heated in a conventionalcarbamate decomposer. The lower section of the first de-composer is heated by condensing NH3—C02—H20 off-gas, and the upper section is heated with steam. The firstdecomposer gas separated from the urea product solutionis reduced in water content by a wash of a minor portionof the reactor effluent, and it is directly used to heat thelower section of the first decomposer. C02 is used forstripping the residual excess ammonia from the ureasolution in the first decomposer separator.

The residual carbamate in the urea product solutionis further decomposed in the second decomposer at re-duced pressure in indirect heat exchange with the streamof condensing NH3—C02—H20 off-gas originating fromthe first decomposer.

Addition of urea and fresh C02 to the condensing off-gas ensures proper heat utilization. The NH3—CC^—HaOoff-gas from the second decomposer separator is con-densed in the second condenser, to which urea recycle isadded. The resulting solution is pumped into the firstdecomposer off-gas to ensure proper condensation of gasand heating of the lower section of the first decomposer.

The mixture of first decomposer NH3—COa—HaOoff-gas and second condenser liquid gives off heat con-secutively to the first decomposer, the second decom-poser, the urea concentrator, the carbamate heater andthe ammonia heater. Fresh C02 gas, required to produceurea in the reactor, is added to each of the above heatexchanger shells for proper temperature control and heatutilization. The mixture of off-gas and second condenserliquid is totally condensed in the first condenser andpumped into the reactor.

The continuous concentration analyzer installed atthe suction of the high-pressure carbamate recycle pumpmaintains the recycle solution at a constant concentra-tion by varying the pressure in the second separator, andconsequently by varying the rate of water evaporationfrom the urea solution in the second decomposer. The75 wt. % urea product solution from the second separatoris further concentrated to 86-88 wt. % solution by utiliz-ing the heat given off by the first decomposer off-gas con-densing in the concentrator shell side.

The 86-88 wt. % urea solution from the concentratoris further evaporated to pure melt and then processed tosolid urea prills in a prilling tower, which is providedwith a highly efficient wet dust abatement system for airpollution control.

Condensate from the hot well of the urea evaporator,containing 2-5 wt. % of NH3, is processed in a novelammonia recovery stripper and rejected to sewer withonly 20-30 parts/million of residual NH3. The overheadgas containing all the NH3 recovered from the solutionis sent to the second condenser for total utilization.

A plant producing urea in the form of 0.6-0.7 wt. %biuret prills by the evaporative route has the followingexpected raw material and utility consumption per shortton of prilled urea produced: NH3, 0.575 ton/ton; C02,0.74 ton/ton; steam at 150 lb./sq. in. gauge, saturated,1,260 Ib./ton; electric power, 110 kw. hr. and coolingwater, 86°F, 20°F temperature rise, 13,300 gal./ton.(Note: the 110 kw. hr. is reduced to 30 kw. hr./ton when600 lb./sq. in. gauge steam is used for the C02 com-pressor drive. The exhausted 150 lb./sq. in. gauge steamfrom the drive is used in the process at no overall steamconsumption increase.)

Features of the Heat Recycle Process described in thisarticle have been successfully incorporated in two total-recycle urea plants of an aggregate capacity of 460 ton/day, presently in operation. #

Literature cited1. Mavrovic, I., Hydrocarbon Proc., 50 (April, 1971).

I. Mavrovic is a consulting engineer spe-cializing in the fertilizer industry, particu-larly in urea manufacturing. He graduatedfrom Liceo Scientifico of Fiume, Italy and re-ceived his B.S. and M.S. in chemical engi-neering from the University of Zagreb,Yugoslavia. He also also worked as a pro-cess engineer.

132

DISCUSSION

E.R. JOHNSON, Allied Chemical Corp.: It's pretty wellestablished- that Frejacques is in considerable error in hisequilibrium constants, and we found that he seems to runpretty high on the required volume one would consideryou'd need to reach equilibrium. Have you been able todevelop any correlation of the volume against time wherethe conversion would fall off due to insufficient reactorvolume? In other words, the retention time required?MAVROVIC: Yes; the main problem is that there is notonly one variable (parameter) affecting the conversion.There are about five or six such variables. Unless youoptimize all of these variables, the variation in one variablewill not appreciably affect the conversion. It means that,unless you attain the optimum operating point for all thevariables involved, the optimization of one single variablewill not yield the right effect on the conversion.JOHNSON: That's why I was thinking it is rather compli-cated. Have you been able to develop a nomograph similarto the one for the equilibrium constants for the reactorvolume under those conditions?MAVROVIC: No. At the present time we have only empiri-cal formulas, based on actual data taken from operatingurea plants. It's extremely difficult I believe to try even todescribe in exact formulas or mathematical expressionswhat is actually happening inside the reactor. I believe,Frejacques equation determines just an overall reaction, itdoes not determine the single step reactions occurring inthe reactor.JOHNSON: I am glad to see that you observed the samephenomena. We found this same thing. We have one plantwe've debottlenecked twice and now we're down to aboutone third of the reaction volume that you'd predict needed,and we're about to go a little further. I guess we'll developempirically where it starts to fall off.MAVROVIC: Absolutely correct.Q. How do you get passivation of equipment in the re-cycle area?MAVROVIC: It is best to treat the critical parts with a 5%dichromate solution containing about 5% nitric acid forabout five minutes, for instance by dipping the lens ringsinto the solution or by simply applying a small amountof solution directly to the flanges or to the part to bepassivated.JOHNSON: Did you say dichromate?MAVROVIC: Dichromate, right. You don't have to passi-vate the equipment of the whole system with solution. It'simportant that at the very first start up of the urea plantmilder conditions and excess of oxygen feed to the reactorare used. Within three or four hours from the start up thesame result is attained as by pumping large amounts ofpickling solution throughout the whole system. This canbecome a rather expensive and complicated proposition. Bypassivating the individual and most critical parts of theequipment the need for passivation of the whole system iseliminated.JOHNSON: We have used the mixture of nitric and hydro-fluoric acids. Your dichromate sound like a much simplersolution. I think we'll try it.MAVROVIC: We usually find that a five percent nitric acidand five percent dichromate solution is very effective. Thedichromate is used to degrease (take the grease away from)the surface of the alloy and to facilitate the oxidation ofthe alloy surface.JOHNSON: Right. Lastly, what design do you favor for thebottom of the prilling tower for the removal of the prillsand to minimize caking, attrition, etc.?

MAVROVIC: There are many designs, such as horizontalfluidized bed bottom and flat bottom with collecting rake,but we adopted a very simple design which is actually afluidized cone, it's an air slide cone. There are no othermoving parts. Opposite to the. standard bottom cone (withsolid walls), our cone is fluidized with air, and it acts as adiverter for the (free falling) prills. The prills never hit thecone, they are simply diverted from their path. It takesmuch less air to simply divert the prills from their path,than to stop them totally (in their free fall, as it is done inthe horizontal fluidized bed).JOHNSON: Yes, it sounds like it should be less subject tocaking.MAVROVIC: Absolutely. Less subject to caking and youhave a better chance of operating it for a longer timebetween cleaning.ANON: One of the principles of your process is that youhave observed that a reactor temperature in the bottom ishigher than at the top. There is one thing that bothers mebecause in all our urea plants, both stripping and conven-tional, we see the opposite. We have the highest tempera-ture at the top (of the reactor). And we think that the factis that the composition of the reactor liquid phase followsthe so-called top ridge line. You know, the top ridge line isthe line of the maximum possible temperatures at a givenpressure, and this top ridge line says that if you have thehighest possible urea concentration at that" point you musthave automatically predicted by the equilibrium the highestpossible temperature.

One thing you must prevent in a urea reactor is theso-called back-mixing. And you can prevent the back-mix-ing by installing a certain number of trays in it.MAVROVIC: I know of the same problem — not only inyour urea plants, regarding the problem you mentionedthat you have hotter reactor tops and colder bottoms. Iknow of this problem in other urea processes too — I don'twant to mention names — the reactor top was hot and thereactor bottom was cold. We found out by analyzing thesystem that carbamate was simply not being formed in thebottom, but it was being formed at the top of the reactor.

We proved this point with a very simple test. Urea acts asa very good solvent for ammonia and ammonia formscarbamate (with CO2) whenever you have urea present in acertain region of the reactor. As soon as we fed a recyclestream of urea into such a reactor where the top was hotand the bottom was cold, we simply shifted the wholetemperature profile upside down: the hot spot moved tothe bottom of the reactor and the cold spot moved to thetop of the reactor. It seems that the carbamate was thenforming in the bottom section of the reactor — the correctplace — and not at the top of the reactor.Q. Is it true that you must have the carbamate formation atthat point where urea is formed, because the urea forma-tion is endothermic — it needs some heat? So what is easierthan to make the carbamate on the spot where urea isformed? This is automatically accomplished in a normalvertical reactor with upward flow.MAVROVIC: That is true, but in such vertical reactor withup flow, if the exothermic reaction of carbamate formationis to supply the endothermic heat of formation of urea, ithappens that carbamate remains in the reactor for 10-12minutes, instead of 30 minute, for instance. So that thispart of carbamate normally is not retained in the reactorlong enough to form urea and ends up exiting the reactorwith the reactor effluent mixture. This fact compel Is you tooverdesign the reactor volume, to provide a longer retention

133

time of carbamate in the reactor.Q. One other question: you said that for technical urea youneed product with low biuret content, and well, I thinkthat it is confusing because some people say they use lowbiuret product for technical grade urea. Some other peoplesay that it doesn't matter what the biuret content is be-cause the first step (in the production of) resins from ureais to heat urea. And if you heat urea, you have automati-cally biuret, triuret and cyanuric acid and so on, you see.MAVROVIC: I investigated this problem of low biuretproduct because I believe that urea manufacturers are com-mitting themself to high expenditures (for the productionof low biuret product) for no good reasons. For instance, aurea product containing 0.8-0.9 wt% of biuret is absolutelyacceptable for the production of glues used in the plywoodindustry. On the contrary, for the production of high graderesins, for instance, whenever transparency is required, thebiuret content of the urea product (used as raw material)

affects the color of the resin. Whenever the biuret contentexceeded about 0.25 wt%, the resin manufacturers foundout that the resin ages sooner — within six months itbecomes yellowish. And this is the only reason why peoplewant low biuret product.ANON: Well, I do not disagree with this. In our DSM plantin Holland we use 60% of our urea production for mela-mine, and the urea which is used for the melamine produc-tion has about 1.2-1.3 wt% of biuret, and we have anexcellent (melamine) product. So we meet the internationalrequirements, (specifications).MAVROVIC: I did not refer to urea product used for theproduction of melamine, if I did I made a mistake. 1 wasreferring to the process of mixing urea product with an-other agent to produce a resin, not to the process offorming a resin via the melamine route. The former isdefinitely affected by the biuret content in the urea prod-uct, as mentioned before.

134