modeling of polymerization of urea and formaldehyde using functional group approach.pdf

7/29/2019 Modeling of Polymerization of Urea and Formaldehyde Using functional Group Approach.pdf

1/14

Modelingof Polymerizationof Ureaand Formaldehyde Using FunctionalGroup Approach

A. KUMAR* and A. SOOD,Department of Chemical Engineering,Indian Institute of Technology, Kanpur-208,016, India

SynopsisThe urea formaldehyde polymerization has been modeled using the functional group approach,

which accounts for the formation of higher oligomers. The kinetic model involves four molecularspecies and three rate constants, as opposed to six as proposed in earlier studies. The experimentaldata of Price et al. have been curve fitted using our model, which is found to describe them verywell in the entire range of temperature and urea formaldehyde ratio. Based on the activationenergies, it has been argued that the reverse reaction step must involve the condensation prod-uct, water.

INTRODUCTIONThe polymerization of urea and formaldehyde is normally carried out by

mixing water solution of formaldehyde (formalin) and urea and subjecting itto high temperature.'S2 Formaldehyde is highly reactive, and when i t ispresentin large concentration it polymerizes into white powder called polyformaldehydeas fol10ws~-~

nCHzO*OH(-CH20),H + ( n- 1)HZOwheren isusually of the order of 10.Under alkaline conditions polyformaldehydedepolymerizes into formaldehyde, which exists mostly as methylene glyc01~-~in aqueous solution.Formaldehyde can polymerize with phenol, cresol, urea, and melamine, anddepending upon the reaction conditions, it can either form a linear polymer ora network."-14 In this study, we have analyzed the formation of urea formal-dehyde polymer. The proposed kinetic model accounts for the addition as wellas the condensation reactions and is completely general. This involves onlythree rate constants, as opposed to six used in earlier studies.The urea-formaldehyde resin is mainly used as an adhesive in plywood andparticle board industries.'?'' The resin is prepared as prepolymer in the firststage. This issubsequently mixed with suitable hardener and catalyst and ap-plied on sheets to be jointed. There is a gelation of the polymer leading to theformation of a network in the second stage.

* To whom correspondence should be addressed.Journal of Applied Polymer Science, Vol. 40, 1473-1486 (1990)0 1990J ohn Wiley & Sons, Inc. CCC 0021-8995/90/ 9- l 01473-14$04.00


2/14

1474 KUMAR AND SOODThere have been several studies reported on the formation of prepolymer of

urea and f~rmaldehyde.l~-~~rea (NH2- O- H2 denoted by U ) has fourreactive hydrogens, and each of them has a different reactivity. Three of itshydrogens are highly reactive compared to the fourth On reaction ofurea with formaldehyde (existing as methylene glycol, OH- H2- H, inwater) there is a formation of U-CH20H.23-32 This is called the addition stepand a methylol bond is formed. If urea is in excess in the acidic medium,U- H20H reacts with another U preferentially to give diurea methyl ether,U- H2- - H2Uormethyl diurea U- H2- . In addition to this, lowmolecular weight polymer is also formed, and this is sometimes called the poly-condensation step. A s the pH increases, methylene linkages reduce, and theprimary products are methylols and ethers, and the chain length of the polymeris limited to a small value.

The gelation of resins occurs at low pH, and the process can be induced byany method that lowers the pH of the reaction mass. Due to the difficulty ofinstrumental analysis, the progress in the studies of gelation has been relativelyslow. The gelation has been followed using X-ray, differential thermal analysis,infrared spectroscopy, thermal gravimetric analysis, and nuclear magnetic res-onance.23,24,3%34 Mayer' has discussed at great length the effect of pH, variousadditives, and their concentrations upon the gelation time.

Most of the analysis of the urea-formaldehyde system has been limited tothe first stage. Price et aL3' have assumed the absenceof the polycondensationstep and only twoof the sitesof urea participate in the first stage as follows:

In this, U and F represent a urea and a formaldehyde molecule whereas UF,and UF2are two of their substitution products. Price et aL3' have experimentallymeasured the free formaldehyde in the reaction mass. They observed that thekinetic model of eq. (1 fitsthe experimental results for low temperature only.For high temperatures they proposed the following additional reaction:

k5UF2+F UF3The urea-formaldehyde polymer has been analyzed by Katuscak et al.,29 whoclearly show that itcontains higher oligomers. They have studied several sam-ples and characterized them with a molecular weight distribution, average mo-lecular weight, and polydispersity index.

In this study, we have used the functional group approach in which we followthe polymerization by defining four reactive species. Based on the chemistryof polymer formation,wehave rewritten the mechanism involving these species.In addition, eqs. (1 and (2) seem to indicate that the reversed step is uni-molecular, and there is a bond scission that may not occur at moderate tem-


3/14

MODELING OF POLYMERIZATION 1475peratures. In view of this, we have considered an additional kinetic model inwhich water was assumed to participate in the reverse step of polymerization.Using both these kinetic schemes, we have shown in this study that it waspossible to describe the experimental data of Price et aL31 in the com-plete range.

KINETIC MODEL FOR THE FORMATIONOF UREA-FORMALDEHYDE POLYMERSince formaldehyde exists in the aqueous medium as methylene glycol,

OH-CH2-OH, it shows a functionality of two. In urea there are four hy-drogens that can undergo reaction with the hydroxyl group of methylene glycol.In view of this, the first methylene glycol molecule can react at four positionsof a urea with equal likelihood, thus urea exhibiting a functionality of four. Wehave attempted to model the polymerization by using functional group approach.In this we identify species in terms of which the formed polymer can be rep-resented. In Figure 1,we have definedfour species, A, B, C, andD. At a given

HI IB -N-C-N-H =0II

or

I I0 -N-C-N-H =II0

or

-N-C-N-H or OHCH2-N-C-N-HI1 II0 0

---N-C-N-H or OHCH2-k-N-HII8 0CH2OH HI I I I-N-C -N-H or OHCH2-N-C-N-H

CH2OH H

a 87 1 ? t

d-N-C-N-H or OHCH2-N-C-N-H

H 5HzOHCH2OH-N-c-N-H or C~+OH-A--C-N-H8 I;

\ 1 I !d

\ y 2 0 H I I6 8C.H 20H I

-N-C-N-H or OHCH2-N-C-N-HII0

CH2OH CH20H-N-C-N-H or -N-C-N-H

CI;120H C,H2OHor OHCHZ - A -c-N-H or O H C H ~N-c-N-HII6 0

Fig. 1. Reactive species used for modeling reversible urea-formaldehyde polymerization.


4/14

1476 KUMAR AND SOODtime within the reactor, we have unreacted CHzOH groups on the polymermolecule formed during reaction as well as the reacted CH2bond. The latterformed whenever two repeat units react, which also leads to formation ofhigher oligomers. In defining these species, no distinction has been madewhether the linkages at these reacted sites are a reacted- H2- ond or a- HzOH group.SpeciesA toD can be used to represent any polymer molecule.For example,

CHzOHICH~OH- NH- CO- NH- CH~- NH- CO- N- CH~- N- CO- NH~ICHZOH-N-CO -NH-CHz-NH-CO -NHz(3)

can be represented by

C-D-BD -AI ( 4 )

Instead of attempting to find out the concentration of different isomers ina polymer, an effort is made here to determine the conversion of urea andformaldehyde in the reaction mass as a function of time. When polymerizationis carried out for some time starting with a feed consisting of urea and form-aldehyde, polymers of various lengths and structures are formed. One plausibledescription of the progress of reaction would be to follow the concentration ofspeciesA toD in the reaction mass. The overall polymerization representedby reaction of functional groups can be written in terms of the following rateconstants:

k1=rate constant for the reaction of primary hydrogenof urea with theOHk2=rate constant for the secondary or tertiary hydrogen of urea with thek3=rate constant for the reverse reaction occurring between a reacted

groups.OH group.- H2- ond (denoted by Z ) and water molecule.

Experimental work of de Jong" have shown that the reactivity of varioushydrogens on urea is dependent upon the number of sites already reacted. Alsoat moderate reaction conditions, there is little evidence of finding reacted tetra-substituted urea, whether it is part of a polymer chain or it is a freely existingmolecule. Based on this chemistry of polymerization of urea and formaldehyde,the forward reaction can be written in terms of speciesA to D. Since there isno formation of tetrasubstituted urea, polymerization can be represented by


5/14

MODELING OF POLY MERIZATION 1477U +F 2 A +CH20H+H20 (5a)

(5b)(5c)(5d)(5e)(5f)(5g)(5h)(5i)(5j)

U +CH20H2 A +H20+ZA +F 2 B +CH,OH +H20k2A +CHzOH+B +H2O +ZA +F2C +CH20H+H2O

A +CH20H2 C +H20+ZC +F2 D +CH20H+H20

C +CH20H2 D +H20+ZB +CHzOH3D +H20+Z

2k2B +F+D +CH2OH +H20

In writing these reactions it has been assumed that the overall reactivity ofa given reaction iscompletely governed by the site involved. Therefore whenurea consisting of four hydrogens reacts with formaldehyde (ormethylene gly-col) having two -OH groups, it forms speciesA as in eq. (5a),and the overallreactivity is8 k I .Lastly, species D does not have any reactive site left, thereforeit is assumed not to react anymore.

Since in species A to D, the- H,- ond and the CH20H inkage havenot been distinguished, it is not possible to write the mechanism of the reversereaction exactly. In view of this, two extreme possibilities have been propo~ed.~,~In the first one (model I ) the various linkageof speciesA to D have all beenassumed to be mainly CH20H groups. As opposed to this, in the second one(model 11)the linkage of species A to D have all been assumed to be mainlyreacted -CH2- bonds. I t is assumed that model I would be a better repre-sentation of the situation in the initial phases of polymerization whereas modelI1would give a better description in the final stages of polymerization. Forthese two models the mechanism of polymerization is given next.

Reverse Reaction for Model IAll linkage are assumed to be reacted- H20H groups.

A +H202U +F - CH20H)B +H202A +F - CH20H)C +H202A +F - CH20H)D +HzO2B+F - CH20H)D +H202C +F - CH20H)


6/14

1478 KUMAR AND SOODReverse Reaction for Model I 1

A +H202 U +(CHzOH)- ZB +H202A +(CHZOH) - ZD+H202 B +(CH20H)- ZD +H202C+(CH20H)- Z

All linkage are assumed to be- H2-(7a)(7b)(7c)(7d)(7e)

C+H202A +(CH2OH)- Z

The mole balance relation for various species in batch reactors can now beeasily written for both these mechanisms. It may be observed that the molebalances for speciesA to D and water are identical for both these models, andthese are considered next.

Model I and I1--[A 1- (2[F] +[CH,OH])dt


7/14

MODELING OF POLY MERIZATION 1479Model I1

+k,([A] +z[C])}+k3[H20]( [A ]+2[B] +2[C] +3[D]} (18)--[ Z 1- [CH20H](2kl(2[U]+[A ]+[BI)+k2([Al +2ICl))dt

RESULTS AND DISCUSSIONPrice et aL3' have carried out polymerization of urea and formalin (37%formaldehyde in water by weight) inaspecial shaped sealed reactor. A careful

examination of the experimental data reported in ref. 31reveals that the con-centrations of free formaldehyde in the reaction mass falls very rapidly forshort times. However, for large times, depending upon the temperature of thereaction mass, it attains an equilibrium. In the kinetic model proposed in thiswork, there are two equilibrium constants,K1andK 2defined as

W i t h o u t w a t e rW i t h w a t e r-_ _ _ -

Fig. 2.of this article.Experimental data of Price et al?' for U /F = 1:1.33curve fittedby the kinetic model


8/14

1480 KUMAR AND SOOD

This clearly means that the initial fall in F is largely affected by the choice ofkl andk2while for large times asymptotic level of F is controlled byKl andK2.With this background, we simulated eqs. (8) ( 16) for model I on a computer.We first kept k2=k3=0and determinedkl ,which would give the same initialfall as observed experimentally, and determined k2andk3from the asymptoticlevels of free formaldehyde in the reaction mass. The final simulation forU :F ratio as1:1.33and 1:2.2are shown in Figures 2and3, and on the sameplots the experimental results of Price et aL31 have been shown. I t isfound thatfor theU :F ratio of 1:2.20andatemperature of 100C,the initial rate of falland the eqililibrium value are lower than that for 16OoC.This is found to havethe opposite trend from that seen in the rest of the data. Even for these, it ispossible to find the set of rate constants to fitthe data in the entire range. Itis thus seen that the kinetic model proposed in this work describes the exper-imental data very well.We have already observed that de Jong et a1.l' and Price et al.31 assumedthat the reversed reaction was unimolecular. Assuming the breakage of bondsto occur this way, we modified the mechanism in eqs. (6)and (7)and the molebalance relations in eqs. (8)- 16).These were similarly simulated on computer,and it was possible to determine the rate constants similarly. We have sum-marized rate constants in Table I for comparison, while we have shown results

- i t h o u t waterW i t h water_ _ _ _

-100%0-160'C

Fig. 3. Experimental dataof Price et al?' for U/F =1 :2.2curve fitted by thi s kinetic model.


9/14

MODELING OF POLY MERIZATION 1481TABLE IRateConstantsNeeded to Fit theExperimentalDat a of Price etal?'

(a) K inetic Model Assuming the Reverse Step as UnimolecularRateConstantTemperature k, for U/F k2 for U /F kl for U/F kZ for U /F k3 for U/F k3 for U/F("C) = 1/1.33 = 1/ 22 = 1/1.33 = 1/2.22 =1/1.33 = 1/2.22

25 3.5x 10-6 - 1.75X lo-' - 5.0X lo-' -40 8.5X 8.5 X lo-' 4.25X lo-' 4.25X lo' 8.3X lo-' 8.3X lo-'60 3.5X lo-' 3.0X lo' 1.75X 15X 1.25X lod 1.25X lo-'80 1.0X lo-' 2.0 X lo-' 8.0X 3.0X 4.0X 10" 8.0X

100 - 4.0X lo-' - 4.0X - 9.0 x 10-6120 2.3X 4.0X lo-' 3.0X lo' 4.0X lo-' 80X lo' 50X lo-'160 2.9 X 4.0X lo-' 4.0X 5.5X 1.25X lo-' 1.0X

( b)Kinetic Model Assuming the Reverse Step Involving Water

25 35x 10- 6 - 1.75X l o-' - 7.0X lo-' -40 8.5X 85X 4.25X l o- " 4.25X l o-' 8.0X lo-' 80X lo-'60 35X lo-' 3. 0X lo-' 1.75X 1.0X 9.0 X lo-' 9.0 X lo'80 10x lo-' 20x 3.0x 3.0x lo-' 10x lo-' 2.0x lo-'

100 - 4.0x - 4.0 X - 9.0x lo-'120 23x 4. 0x 30 x lo-' 4.0x lo-' 2.0 x lo-' 50 x l o-*160 26x 10-3 40x 10-3 40x 10-6 40x 10-6 3.0x lo-' 3.0x l o-'Indicates that rate constants are for U/F =1/1.33.Indicates that rate constants are for U/F =1/2.2.

by dotted lines on Figures 2and 3. It is seen that the two simulations almostoverlap for all cases, indicating a relative insensitivity of these models to theconcentration of water in the reaction mass. This is found because in formalinthere is already a large amount of water present and its concentration changesonly by a small amount due to polymerization. We have prepared Arrheniusplots for these in Figure 4, and we find that the choice of inclusion of water inthe reverse step has an effect only on kio ( i = 1, 2,3) while activation energiesare affected only negligibly.Price etaL31have presented experimental data for two differentU :F ratios.We have curve-fitted experimental data for both these sets, and results havebeen plotted in Figures 4 and 5. In Figure4 we assume that water affects thedepropagation reaction, while in generating Figure5, we assume that it doesnot. In both these figures, the filled legends have been used forU :F as 1:2.20.We have already observed that for 100 and 120C and U :F ratio as 1:2.2,the experimental data have different trends from those observed for the restof the data. Consequently, the rate constants needed to fit these are higherthan those for 160C as the reaction temperature. If these are ignored, it ispossible to draw a decent single straight line for kl , 2,andk3as seen, whichis independent of the U/F ratio. These figures give rate constants that aretemperature dependent only, in contrast to that found in earlier kinetic models


10/14

1482

-

- 7-

- 9-._xc--1 1

-13

-15

-16

K UMAR AND SOODO r45-

-

-

--

---

-

-

C a s e 1 C a s e 2kl Ok2 A 6k3 O

EJ 2000

I$O*C 120 100 80 60 4p 2;' 2 . 2 24 26 28 30 32 3 418l T (%-'I

Fig.4. Arrhenius plotsof rate constantsk, tok3assuming water havingnoeffect on the reversestep.

proposed in the literature. Comparison of Figures4 and5shows smaller scatterof data in the latter. I t is very difficult to conclude from the limited data ofPrice et aL31 as to whether the condensation product, water, participates in thereverse reaction. The activation energy of rate constant k 3 is found to be smallcomparedtothe degradation reaction, which suggests that water must be playingsome role in the reverse reaction.With the rate constantssofound, we have simulated the urea-formaldehydepolymerization in batch reactors. Figures 6 and7give some results; the formergives the unreacted urea in the reaction mass as a function of time at differenttemperatures. The unreacted urea decreases rapidly in the initial region, ulti-mately leveling at the equilibrium value. With increasing reaction temperature,it is found that the rate of fall in the initial region increases while the equilibrium


11/14

MODELING OF POLY MERIZATION

-4

1483

- Case 1 Case2kl 0k2 A Ak3 CI

m

Fig.5. Arrhenius plotsof rate constantsk, tok3assuming bimolecular reaction between waterand reacted bonds.

value falls. In Figure 7, we have plotted the concentration of species A as afunction of time with temperature. The behavior observed is similar to thoseof intermediate species in a setof consecutive reactions. At a given temperature,the concentration of speciesA first rises, and after passing through a maximum,it falls to the equilibrium value. A s the temperature is increased, the peakheight falls, and its position shifts to smaller times as seen.Earlier studies of de J ong and de Jong12were carried out at low temperatures,and the kinetic model proposed by them completely neglected the formationof higher oligomers. Assuming monosubstitution of formaldehyde on urea, theyevaluated the rate constants. Price et aL3' have performed their experimentsat much higher temperatures when the formation of higher oligomers cannotbe ignored. In fact in their kinetic model, they have relaxed the earlier as-


12/14

1484 KU MA R AND SOOD

\Wi thou t w a t e rWith w a t e r

-_ ._

Fig. 6. Concentrationof unreacted urea versus reaction time.

- i t h o u t w a t erW i t h w a t e r.___

-25'C

- at- - _ _1 6 0 t

I 1 -bU 1LU 1a0 3$0t (min)Concentrationof speciesA versus reaction time.ig. 7.


13/14

MODELING OF POLY MERIZATION 1485sumption of monosubstitution and included reactions given in eq. (2).Theyobserve the insufficiency of their kinetic model in fitting the experimental dataat high temperatures. The kinetic model proposed in this work involves fewerrate constants (only three as opposed to six) and accounts for the formationof higher oligomers. Polymerization of urea-formaldehyde has been reviewedby Meyer,' and he observed that it was possible to divide the polymer formationinto two stages for low temperatures only. In the first stage, according to Meyer,addition is the predominent reaction, even though small amount of condensationreaction also occurs. As opposed to this, in the second stage most of the chaingrowth occurs through the condensation mechanism. Industrially, the polymerformed in the first stage is available as a syrup. Katuscak et al." have subjectedthis to GPC studies and shown that it has a polydispersity index more thanunity, thus indicating the formation of higher oligomers. The kinetic modelproposed in this work serves as a basis for determining the MWD of the polymerformed using the kinetic approach.

CONCLUSIONSA kinetic model for reversible urea-formaldehyde polymerization based upon

functional group analysis has been proposed. This accounts for the formationof higher oligomer and is found to be extremely suitable for the correlation ofexperimental data at high temperatures. The model proposed in this work in-volves only three rate constants, as opposed to other models reported in lit-erature that use at least six rate constants.The kinetic model proposed in this work has been computer simulated andexperimental data of Price et al?l have been curve fitted using rate constantsas the parameters. I t is found that the model fits in the entire range and therate constants, so determined, canbe represented by a suitable Arrhenius re-lation. In the earlier kinetic models of de Jong12 and Price et al.,31 he reversereaction was shown to be a unimolecular reaction. It was argued in this articlethat this way of representation is similar to chain degradation and thereforeshould have high activation energy. We have also examined a kinetic model inwhich the reverse step was assumed to be a bimolecular reaction involving thecondensation product. On comparison of experimental data of Price et aL3'with computed results, one finds that it is not possible to distinguish themfrom each other. This is due to the fact that polymerization has been carriedout with formalin, which already has large amount of water. However, it wasfelt that the activation energy of rate constant k3 is small, and hence watermust have some role to play, and a second-order kinetic for the reverse step isa better representation.

References1. B. Meyer, Urea Formldehyde Resins, 1sted., Addison-Wesley, London, 1979.2. J .F. Walker, inKirk-Othmer Encyclopediaof Chem. Tech.,2nd ed., Interscience, New York,3. I. Anderson, G. R. Lindquist, and L. Molhave, Ugeskr,Laeg. , 141,6287 (1979).4. A. Kumar, S.K. Gupta, and B. Kumar, Polymer, 23, 1929 (1982).

1966. Vol. 10, p. 77.


14/14

1486 KUMAR AND SOOD5. M. Imoto, Formaldehyde, Asahnla Shoten Publisher, Tokyo, 1965.6. J . M. Lehn, G.Wipff, and H. B. Burgi, Helu. Chem. Acta, 57, 493 (1974).7. F. Auebach and H. Barschal, Die Festern Polymerender Formaldehyde, Springer, Berlin8. A. V. Rudnev, E.P.Kalyazin, K. S.Kalugin, and G. V . Kavalev,Zhm. Fiz. Khim.,61 l o),9. W. Dankelman and J . M. H. Daemen, Analyt. Chem.,48,401 (1976).

1907.2603 (1977).

10. S.K. Gupta andA. Kumar, Reaction Engineeringof Step Growth Polymerization,Plenum,11. S. K. Gupta and A. Kumar, Chem.Eng. Commun.,20, l (1983).12. J . J . de Jong and J . de Jong,Rec. Trau.Chim..72, 497(1953).13. L. M. Y eddanapalli and A. K. Kuriakose, J . Sci. Ind. Res. (I ndia). 18B, 467 (1959).14. M. F. Drumm and J . R. LeBlane, in Step Growth Polymerization, D. H. Solomon (Ed.)15. J . H. Updegraff andT. J .Suen, inPolymerization Process, C.E. Schildknecht and I. Skeist16 J . I. de Jong and J . de Jong, Recueil, 72, 1027(1953).17. J . I. de J ong and J . de Jong,Recueil, 72, 139 (1953).18. J . I. de Jong and J . de Jong, Recueil, 72, 88 1953).19. J . I . de J ong andJ .de Jong, Recueil, 71, 890 (1952).20. J . I. de J ong and J . de Jong, Recueil, 71, 643 (1952).21. J . I. de Jong andJ . de Jong, Recueil, 71, 661 (1952).22. T.Suge, T. Miyabayashi, and S.Tanaka, Burtseki Kuguku, 23, 11446 (1974).23. B. Tomita, J . Polym. Sci., 16, 2509C (1978).24. B. Tomita and Y. Hirose, J . Polym. Sci., 14, 387 (1976).25. G. Zigeuner and R. Pitter, Monotsch,86, 57 (1955).26. N. Landquist, Acta Chem. Scand., 11, 776 (1957).27. H. Kodawaki,Bull. Chem. SOC. pn., 11. 248 (1936).28. H. Kodawaki and Y. Hashimoto,Rept. Znd. Res. Inst. Osaku (J apan), 7(6),2 (1926).29. S. Katuscak, M. Tomas, and0.Sciessl, J . Appl. Polym. Sci.,26,381 (1981).30. M. Chiavarini, N. Del Fanti, andR. Bigatto, Angew. Makromol. Chem.,46,151 (1976).31. A. F. Price, A. R. Cooper, andA. S.Meskin, J . Appl. Polym. Sci.,26.2597-1611 (1980).32. P. M. Frontini, J . Borrajo, andR. J. J . Williams, Polym. Eng. Sci.,24,1245 (1984).33. S. Chow and P.R.Steiner, Holzforschung,29, 4 1975).34. P.R. Steiner, Wood Sci.,7,99 (1974).

New York, 1987.

Marcel1 Dekker, New York, 1972.(Eds.) Wiley-Interscience, New York, 1977.

Received J uly 10, 1989Accepted October 3, 1989

modeling of polymerization of urea and formaldehyde using functional group approach.pdf

Documents