":J""'~TEMPERATURE

By H. ThAN, X. MAO, J. CAMERON AND C.M. BAIR

J HE USE OF SODIUM BORATES to causticizesmelt directly in soda and kraft recov-

-ery boilers was first suggested by Jan-son in the late 19705 [1-3].Janson pro-

posed that the "autocausticizing" process isattained by reacting sodium carbonate (Na2CO3)in the molten smelt with sodium meta-borate(NaBO2) to form disodium borate (Na4B205) ,which subsequently is hydrolysed to form sodiumhydroxide (NaOH) and to regenerate NaBO2 inthe dissolving tank:

2 NaB°2 + Na2CO3 ~ Na4B205 + CO2Reaction 1

Na4B205 + H20 --+ 2 NaOH + 2 NaB°2

Reaction 2

J. CAMERON,Western Michigan

University,Kalamazoo, MI, USA

C.M. HAIR,US Borax Inc.,

Valencia, CAUSA

quires a proportionately smaller borate load.However, it was suggested by Janson [2] that Re-action 1 would be severely hindered if the Na:Bmolar ratio ( or Na/B) of the reactants is greaterthan 1.5:1 (or 1.5); and that it would not evenoccur if Na/B is greater than 3. This means thatpartial autocausticizing is technically not feasiblebecause Na/B is likely to be much greater than 3.The development work involving borates became H THANdormant by the end of the :980s. .P~lp & p'aper Centre,

In 1990, US Borax Inc., m search of new appll- University of Torontocations for borate products, began to re-examine Toronto ON ,technologies that involve the use ofborates in the ,

pulp and paper industry. Previous research workon borate-autocausticizing was reviewed, andnumerous in-house experiments were performed.This, along with the positive results of recent stud-ies on the effect ofborates on kraft, kraft-AQandsoda-AQ cooking of black spruce [9,10] , led to theinitiation of a large research program to examinethe effects ofborates on various processes in kraftpulping and chemical recovery. The program in-volved several independent research groups in- X. MAD,cluding: Econotech Services Inc., Vancouver, BC, Pul.p & ~aper Centre,Western Michigan University, Kalamazoo, MI, University of Toronto,Institute of Paper Science and Technology, Toronto, ONAtlanta, GA, Oregon State University, Corvallis,OR, and the University of Toronto. At the sametime, Sandwell Inc., Vancouver, B.C. was asked toconduct an economic analysis of the borate-auto-

causticizing process.Work by Econotech Services suggests that the

use of borate may produce better quality pulp,and pulp which shrinks less during bleaching.The study at the University of Toronto, as will bediscussed in detail below, shows that in air, sodiumborates can react with molten Na2CO3 at anyNa/B value, and that the reaction product is likelyto be trisodium borate, 3Na2O.B2O3 ( or Na3BO3) .These findings imply that partial autocausticizingwith borate may now be technically feasible.

This paper discusses the results and practicalimplications of the recent study at the Universityof Toronto, and those of the unpublished workconducted at the Institute of Paper Chemistry in1987 on autocausticizing reactions betweenNa2CO3 and sodium borates.

AUTOCAUSTICIZING REACTIONSIn the study at the University of Toronto, dehydratedborax (Na2B4O,) and dehydrated sodium meta-borate (NaBO2) were used as a source of borate.Mixtures ofNa2CO3 and Na2B4O, ( or NaBO2) with

The process is attractive in that it producesNaOH directly in green liquor, thus potentiallyeliminating the need for slakers, causticizers andthe lime kiln.

Besides sodium borates, there are several com-pounds that may be used as autocausticizingagents: for example, alumina (AI203) , silica(SiO2) and disodium phosphate (Na4P207) [2],ferric oxide (Fe203), titanium dioxide (TiO2) andilmenite (FeTiO3) [ 4-6]. Two basic differencesbetWeen borates and most other autocausticizingcompounds are: i) borates are water-soluble whilethe others (except for Na4P207) are not; and ii)the decarbonization of Na2CO3 (Reaction 1) andthe hydrolysis of the resulting higher-Na/B-ratioborates (Reaction 2) occur rapidly in the boratecase because they happen in solution and are notthe result of solid-solution interactions. Further-more, borates do not react with sulphide in smelt,and thus can be used in kraft process [3].

A full-scale plant trial was carried out at theEnso Gutseit liner board mill in Kotka, Finland, in1982. While the results were inconclusive, the trialprovided some insights into the feasibility of theprocess. A review of various autocausticizing pro-cesses conducted by Grace [7] suggested that full-scale implementation of the borate process istechnically difficult and economically unattrac-tive, due primarily to the large amount of borate"deadload" in the liquor cycle, and to the lowheating value and the high viscosity of the result-ing borate<ontaining black liquor [8].

For mills which require incremental causticiz-ing and lime kiln capacities, partial autocausticiz-ing may be an attractive alternative since it re-

100:9 (1999) 111 35Pulp & Paper Canada T 283

~1.="' TEMPERATURE

N~B407 + 3 Na2CO3 ~ 2 N~O5 + 3 CO2

Reaction 4

various Na/B values were prepared and thereaction was examined by means of simulta-neous differential and thermogravimetricanalysis (DTA/TGA). A small amount,about 20 mg, of the sample, was placed in aplatinum crucible and heated in air up to925°C at a constant heating rate of10°C/min. The sample weight was continu-ously monitored. The total weight losscaused by the release of CO2 as a result ofthe reaction was used to determine thedegree of decarbonization ofNa2CO3. It wasassumed that no weight loss occurred due toevaporation. Experiments in air that con-tained 15% CO2 were also carried out toexamine the effect of CO2 on the reactions.Reactions between Na2CO3 and dehy-drated borax: Figure I shows the thermo-gravimetric (TG) curves of Na2B407,Na2CO3, and a mixture Na2CO3 andNa2B407 which had a Na/B of 1. Na2B407shows no significant weight loss at temper-atures up to 925°C. Na2CO3 begins tovaporize markedly at its melting tempera-ture of about 850°C. In the case of theNa2CO3 and Na2B407 mixture (Na/B=I),the weight loss began at about 600°C,reached 15 wt% at about 740°C and re-

It is interesting to note that the weightloss data in Fig. 2 at Na/B>3 was above thesolid curve. This high weight loss is possi-bly due to the vaporisation of excessNa2CO3 in the Na2CO~Na3BO3 mixture,since molten Na2CO3 is much morevolatile than sodium borates, Fig. 3.Reactions between Na2CO3 and dehy-drated sodium meta-borate: Similar resultswere obtained for Na2CO3 and NaBO2mixtures, Fig. 4. The measured weight losswas highest at Na/B= 3. The data was,again, more consistent with the solid curve(Na3BO3 is assumed to be the reactionproduct, Reaction 6), than with the bro-ken curve (Na4B2O5 is assumed to be thereaction product, Reaction 1, as was sug-

gested by Jansen [2]).

mained constant at higher temperatures.The maximum weight loss of this mixtureis consistent with the theoretical weightloss for the complete decarbonization ofthe Na2CO3 in the sample. This suggeststhe reaction was complete under the testconditions. For mixtures with higherNa/B values, the maximum weight lossincreased markedly with an increases inNa/B values, up to 3; then decreasedthereafter, Fig. 2.

The solid curve in Fig. 2 represents thetheoretical weight loss of the sample cal-culated on the assumption that the re-action product was trisodium borate(3 Na20.B203 or Na3BO3): Reaction 3.The broken curve is the theoretical weightloss of the sample if the reaction productare disodium borate (2 Na20.B203 orNa4B2O5) : Reaction 4. The weight lossdata obtained follow more closely the solidcurve than the broken curve, suggestingthat the decarbonization of Na2CO3 oc-curs according to Reaction 3, with Na3BO3as the reaction product.

NaBO2 + Na2CO3 ~ Na3BO3 + CO2Reaction 5

Note that the weight loss data in Fig. 4fitted better with the solid curve atNa/B>3 than that in Fig. 2. This is because

Na2B407 + 5 Na2CO3 ~ 4 Na3BO3 + 5 CO2

Reaction 3

T 284 Pulp & Paper Canada36 III 100:9 (1999)

~1.:-"TEMPERATURE

To Hood+

CO/CO2

~al~

Mass Flow Meter

~Particulate

-Filter .

,1::jj tI

--...

I!, ..'00

N,

Ceramic -Purge Tube

~

,Steel Retortin RF Field

Ceramic Crucible

Smelt

Thermocouple

steel retort, Fig. 5. The retort was heated ina radio-frequency induction furnace, whichallowed the mixture to reach the desiredreaction temperature within a few minutes.In most of the experiments, a large excessamount of Na2CO3 was used relative toNaBO2' with an Na/B>50. The reactiontemperature was varied from 1000 to1200°K (727 to 927°C). During the experi-ment, the extent of the reaction was fol-lowed by purging the reactor with N2' andmonitoring the CO2 generated in the off-gases using an infrared spectrophotometer.

For each mole of NaBO2 used, onemole of CO2 was obtained. This suggeststhat Reaction 5 is the main reaction, andthat Na3BO3 is the reaction product.With the CO2 stripped from the reactor,the reaction could accurately bedescribed as a second order irreversiblereaction; first order with respect to boththe borate and carbonate.

wherek is the reaction rate constant =

2580:t400 L/mol-sec;M is the activation energy =

35:t2.5 Kcal/mol;R is the universal gas constant; andT is the absolute temperature of the

reaction, in the range of 1000 to1200°K.

In experiments where controlled levelsof CO2 were added to the purge gas, therate of decarbonization reaction was lower.

The above studies suggest that Reac-tions 3 and 5 can occur at any Na:B molarratio, and that the reaction product isNa3BO3. The conclusions contradict theoriginal findings by Janson [2] , which sug-gested that the reaction product isNa4B205 (Reaction 1); that the reactionwill be severely hindered if the Na/B valueof the reactants is greater than 1.5; andthat it will not even occur if Na/B isgreater than 3. The contradiction is likelydue to the following:

the Na:B molar ratio in NaBO2 is 1,whereas that in Na2B40, is 0.5. At a givenNa/B value, the amount of Na2CO3 usedin the Na2CO3-NaBO2 mixture wassmaller than that used in the Na2CO3-Na2B40, mixture. Therefore, the effect ofvaporization of excess Na2CO3 in Fig. 4 isless than that in Fig. 2.

The formation of trisodium borate,Na3BO3' rather than disodium borate,Na4B205' from the reaction betweenNa2CO3 and NaBO2 was also found inwork perfonned at the Institute of PaperChemistry in Appleton, WI, in 1987. Thepurpose of the work was to generate CO2in the smelt as a means of reducing thetendency of smelt to explode on contactwith water. Although this method ofreducing smelt-water explosions did notprove to be effective, the kinetics of thereaction between Na2CO3 and NaBO2were examined.

Mixtures of Na2CO3 and NaBO2 werecontained in an alumina crucible inside a d [ CO2] / dt = k ef.EIRT [NaBO2] [Na2CO3]

P..ln & P~npr C~n~rl~ T 2?5 100:9 (1999) III 37

~~ II TEMPERATURE

.The larger sample size, and the lack of apurge to remove CO2 from the relativelylarge crucible used in janson's experi-ments. Since CO2 was not removed fromthe system, high levels could accumulateand prevent the reaction from going to

completion..janson's samples were prepared byadding an aqueous solution of NaOH toa constant amount of boric acid (H3BO3'80 mmoles) and Na2CO3 ( 40 mmoles) togive the desired Na/B value. The additionof NaOH, instead of Na2CO3' to adjustNa/B in janson 's work was not realisticbecause: i) there is little or no NaOH insmelt; and ii) since NaOH is more alkalineand melts at a much lower temperaturethan Na2CO3, it would react rapidly withH3BO3 to form Na4B205 (Reaction 6).Thus at Na/B>3, there would be noH3BO3 left to react with Na2CO3.

2 H3BO3 + 4 NaOH ~ Na4B205 + 5 H20Reaction 6

Janson suggested that Na3BO3 wouldnot form from Reaction 7, because of thestrongly basic character of the B2054-anion in the melt [2] .The reaction, how-ever, was not examined in Janson's work:

Na4B205 and Na3BO3. For mixtures withan initial Na/B(3, the resulting Na4B205would react further with Na2CO3 to formNa3BO3 (Reaction 7).

EFFECT OF CO2Sodium borates, particularly those withhigh Na/B values, such as Na4B205 andNa3BO3' might be expected to be carbon-ated if they are exposed to CO2. This isbecause the Na20 component of the com-pounds might react with CO2 to formNa2CO3. Since the kinetics of Reactions I,3, 4 and 5 is slower in CO2 than in air. ThetWo main reactions in this study, Reactions3 and 5. should be rewritten as follows:

Na4B205 + Na2CO3 -? 2 Na3BO3 + CO2Reaction 7

Na2B40, + 5 Na2CO3 H 4 Na~O3 + 5CO2

Reaction 8

A) , the sample regained about 3% of itsweight immediately as the gas atmospherewas switched from air to air plus 15% CO2.The total weight loss was reduced to about28 wt% and remained constant at this leveluntil CO2 was turned off (Point B) .At thatpoint, the sample began losing weightagain until CO2 was re-introduced (PointC). A small weight loss was observed againwhen CO2 was off (Point D) .

Based on the above results, it can beconcluded that the kinetics ofNa3BO3 for-mation reaction are a strong function oftemperature, and are hindered by thepresence of CO2. However, once formed,Na3BO3 is relatively stable, and is not becarbonated. Its vaporization rate at 925°Cis about 1/50 of the vaporization rate ofNa2CO3, Fig. 3.

IMPLICATIONSThese above studies indicate that bothdehydrated borax (Na2B407) and sodiummeta-borate (NaBO2) can readily reactwith Na2CO3 in air -even at Na/B>3 -

and that the reaction product is Na3BO3.In water, Na3BO3 hydrolyzes to formNaOH and to regenerate NaB°2, accord-ing to Reaction 11. Thus, regardless of thestarting material, NaBO2 is the mainsodium borate compound in the solution.Since borax (Na2B407.5H20) is less ex-pensive and has a low Na/B value, itschoice as a make-up chemical is abetterone than sodium meta-borate. The over-allautocausticizing reactions can thereforebe summarized as follows:

NaBO2 + Na2CO3 H Na3BO3 + CO2Reaction 9

STABILITY OF Na3BO3While trisodium borate (Na3B03) is rarelymentioned in the borate-autocausticizingliterature, this compound is known toexist. Phase equilibrium diagrams of theNa20-B203 system by Milman and Bouaziz[11] show that Na20 and B203 form manycompounds with different Na/B values.They are, in an order of increasingNa/B value: Na20.9B203, Na20.5B203,

Na20.4B203' Na20.3B203' 2Na20.5B203,Na20.2B203, 2Na20.3B203' Na20.B203'3Na202~03' 2Na20.~03, 5Na202~03'and 3Na20.B203. The large number ofsodium borates implies the high affinity ofborate to sodium compounds.

The subsystem Na20-Na20 B203 in theNa20-B203 system is shown in Fig. 6 [11].Note that the dashed lines in the diagramrefer to metastable equilibrium, whilethe compounds in quotation marks,"3Na20.2B203"and "5Na20.2B203'" arenot certain and might correspond to5Na20.3B203 and 7Na20.3B203' respec-tively. Disodium borate (2Na20.B203or Na4B205) and trisodium borate(3Na20.B203 or Na3B03)' respectively,melt at 640°C and 675°C, and they form aeutectic at 570°C. This eutectic tempera-ture is consistent with the temperature atwhich the weight loss of carbonate-boratemixtures began in most of tests conductedin air at the University of Toronto, whenNa/B>2. This implies that the decarbon-ization of Na2C03 by borate in .air wouldinitially result in the formation of both

NaBO2 + Na2CO3 H Na3BO3 + CO2Reaction 10

Na3BO3 + H20 -7 2 NaOH + NaBO2Reaction 11

Figure 7 shows TG curves of a Na2CO~Na2B407 mixture (Na/B=3) in air and inair plus 15% CO2. The total weight loss ofthe mixture in air that contained 15%CO2 after the test was only 60% of that inair. This adverse effect of CO2 on decar-bonization is a strong function of temper-ature. As shown in Fig. 8, during theperiod when the reaction temperature waskept constant at 825°C (from Point A toPoint B) , the weight loss of the melt sta-bilised at about 18 wt%. The weight lossincreased to 22 wt% as the temperaturewas increased to 925°C (Point C). How-ever, as the temperature was graduallydecreased from 925°C (Point D) to 725°C(Point E), the sample gradually regainedweight to a final value of 13 wt%, andremained at this value as the temperaturewas maintained at 725°C.

In the experiment shown in Fig. 9, theNa2CO~Na2B407 mixture was first heatedin air from room temperature to 925°C,and then kept at that temperature for sixhours. The weight loss stabilized at about30 wt%, which is consistent with the theo-retical weight loss caused by the formationof Na3BO3, Fig. 2. At 120 minutes (Point

This reaction sequence suggests thatonly half a mole of borate is needed toproduce one mole of NaOH in the liquorsystem. This is important because theamount of borate required would be halfof that needed for the reaction sequencesuggested by Janson [2] , Reactions land

T 286 Pulp & Paper Canada38 III 100:9 (1999)

~'-=--'

cooking and Bleaching -Part 1. A New Approach toLiquor Generation and Alkalinity. Papt:riJa Puu 59(6-7): 425-430 (1977).2.lANSON,l. The Use of Unconventional Alkali Incooking and Bleaching- Part 5. AutocausticizingReactions. Papt:riJa Puu 61 (1): 20-30 (1979).3.lANSON,l. The Use of Unconventional Alkali Incooking and Bleaching -Part 6. Autocausticizing ofSulphur-containing Model Mixtures and SpentLiquors. PaperiiaPuu61(1): 20-30 (1979).4. KliSKlLA, E. Recovery of Sodium Hydroxide fromAlkaline Puling liquors by Smelt Causticizing -PartII. Causticizing of Molten Sodium Carbonate with Tita-nium Dioxide. Papt:riia Puu 61 (5): 394-401 (1979).5. KliSKlLA, E. Recovery of Sodium Hydroxide fromAlkaline Puling liquors by Smelt Causticizing -PartIV. Causticizing ofMolten Sodium Carbonate with Fer-ric Oxide. Paperiia Puu 61 (8): 505-510 (1979).6. KliSKlLA, E. Recovery of Sodium Hydroxide fromAlkaline Puling Liquors by Smelt Causticizing -PartV. Causticizing of Molten Sodium Carbonate withIlmenite. PaperiiaPuu61(9): 564-577 (1979).7. GRACE, T.M. An Evolution of Non.ConventionalCausticizing Technology for Kraft Chemical Recovery.Report One, A Progress Report to Members of theInstitute of Paper Chemistry,lanuary 30, 1991.8.lANSON,l. and SODERHlELM, L. The Viscosity ofBorate-Containing Black liquor. Nordic Pulp and Paper&.=hi. No.2: 107-110 (1988).9. KORAN, C., WANTDELT, P. and KUBES, G..j. TheEffect of Temperature on Borate-Based Kraft Cookingof Black Spruce. Paperiia Puu -Paper and Timber78(9): 541-544 (1996).10. PRIHODA, S., WANTDELT, P. and KUBES, G.,l.The Effect of Borates on Kraft, Kraft-AQAnd Soda-AQCooking Of Black Spruce. !,apt:ri fa Puu -Paper andTimber78(8): 456-460 (1996).11. MILMAN, T. and BOUAZIZ, R. Phase Diagrams forCeramists. Ed. Levin, E.M. and McMurdie, H.F. Ann.Chim.3 (4): 313-317 (1968).12. HUPA, M. Recovery Boiler Chemistry. Kraft Recov-try Boilers. Ed. Adams et al. Atlanta: TAPPI Press, 44(1997).

recovery furnace, the concentration ofCO2 in the char bed is low due to its con-version to CO; consequently, the decar-bonization reaction should proceed as anirreversible reaction within the furnace..The decarbonization reaction, as mea-sured through CO2 generation, can bedescribed as a second overall reaction, firstorder with respect to both reactants. Inthe temperature range of 1000 to 1200°K,the reaction constant k and the activationenergy ~E are respectively 2580~400L/mol-sec, and 35 ~ 2.5 Kcal/mole..The formation of Na3BO3 as a reactionproduct implies that only 0.5 mole ofNaBO2 is needed per mole ofNaOH pro-duced in the liquor system. The low boraterequirement substantially reduces con-cerns over high black liquor viscosity andlarge deadload in the liquor cycle..The amount of fume formed is ex-pected to be significantly less in boilersburning borate-containing black liquorthan boilers burning normal black liquor..Partial borate-autocausticizing appearsto be technically feasible and may beattractive economically.

ACKNOWLEDGEMENTSThe work at the University of Toronto wassupported by U.S. Borax Inc. and the Nat-ural Science and Engineering ResearchCouncil of Canada.

LITERATUREI.JANSON,J. The Use of Unconventional Alkali In

Abstract: The autocausticizing technology involving the use of borate to "causticize" sodiumcarbonate (Na2CO3) in the soda or kraft smelt directly in recovery boilers was reviewed by re-exam-ining the reactions between Na2CO3 and sodium borates (NaBO2 and Na2B4O7) at high tempera-tures. Both NaBO2 and N~B4O7 were found to react readily with molten N~CO3 at any Na:B molarlarger than 3:1 to form trisodium borate (Na3BO3) .The formation of Na3BO3 as a reaction productmeans that the amount of borate required for the autocausticizing reactions can be substantiallylowered than previously suggested.

Reference: TRAN, H., MAO, X., CAMERON,].. BAIR, C.M. Autocausticizing of smelt withsodium borates. Pulp Paper Can. 100(9): T283-287 (September 1999). Paper presented at the 1998TAPPI International Chemical Recovery Conference. co-sponsored by the Technical Section. CPPA,in Tampa. FL. on]une 1 to 4, 1998. Not to be reproduced without permission of Pulp and PaperTechnical Association of Canada. Manuscript received on March 15. 1998. Revised manuscriptapproved for publication by the Review Panel. March 12, 1999.

Keywords: CAUSTICIZING. SMELT. SODIUM BORATE. SODIUM CARBONATE. REcov-ERY FURNACES. TEMPERATURE.

II TEMPERATURE

2. Therefore, if the Reactions 9 and 11 canbe fully attained, the amount of boratedead load in the system would be reducedby half. This, in turn, would significantlylower the black liquor viscosity and reduceother potential effects caused by borate.

Although CO2 hinders the formationof Na3BO3 (Reactions 8 and 9), the prac-tical implications of this are minimal.Equilibrium calculations [12] show thatwhile the CO2 content in the flue gas inthe recovery boiler lower furnace may beas high as 15%, the CO2 content in thechar is very small, <1000 ppm, due to thepresence of a large amount of carbon, andthe equilibrium reaction, C + CO2 = 2 CO

that consumes CO2 at high temperatures.The decarbonization reac~ion of Na2CO3to form Na3BO3 may proceed as an irre-versible reaction within the furnace. OnceNa3BO3 has been formed in the smelt, theeffect ofCO2 becomes less important, as isshown in Fig. 9.

The occurrence of the decarbonizationreaction (Reaction 9) at any Na:B molarratio suggests that partial autocausticizingis technically feasible since it requires pro-portionally less borate. This may be anattractive alternative for the incrementalcapacity of the causticizing plant. How-ever, the effect of borate in various causti-cizing processes needs to be thoroughlyexamined before such a technology canbe implemented.

The much lower volatility of sodiumborate compounds compared to Na2CO3,Fig. 3 suggests that the amount fumeformed will be significantly less in boilersburning borate-containing black liquorthan boilers burning normal black liquor.

CONCLUSIONSThe basic reactions between sodium car-bonate (Na2CO3) and sodium borates(NaBO2 and Na2B4O7) at high tempera-tures were examined. The results show that:.In air, both NaBO2 and Na2B4O7 reactreadily with molten Na2CO3 at any Na/Bvalue; and the reaction product is likely tobe trisodium borate (Na3BO3) , not dis-odium borate (Na4B2O5) , as was previ-

ously suggested..The reaction kinetics ofNa3BO3 forma-tion depends strongly on the sample size,temperature, the initial Na/B, and CO2concentration in the gas. Once formed,Na3BO3 is relatively stable, and the effectof CO2 becomes insignificant-.The kinetics of the decarbonizationreaction can be increased by removing theCO2 generated, and by using an excessiveamount of sodium carbonate. In the

T 287 Pulp & Paper Canada40 III 100:9 (1999)