Condensation kinetics of polyphthalamides. I. diamines and diacids or dimethylesters.

INTRODUCTION

The polyamides PA6 (poly[imino(1-oxohexane-1,6-diyl)]) and PA66 (poly[iminoadipoyliminohexane-1,6-diyl)]) have dominated the PA market ever since they were developed in World War II. These purely aliphatic polyamides owe their high melting points and excellent properties to NH-CO hydrogen bonds. They do not really need aromatic rings which impart similarly high melting points to polyesters like PET (poly(oxyethyleneoxyterephthaloyl)) or PBT (poly(oxybutyleneoxyterephthaloyl). Nonetheless, aromatic polyamides are conquering an increasing market segment. They are distinguished by a lower water uptake, better dimensional stability, and especially, higher softening temperatures and melting points than their purely aliphatic relatives. Under the hood of modern automobiles, where long-term stability at 150[degrees]C is desired, the semiaromatic polyamides, mostly polyphthalamides of aliphatic diamines and phthalic acids, compete with metals. The polyphthalamide PA6T, consisting of hexamethylene diamine and terephthalic acid (TPA) units, is almost too good: it melts at 370[degrees]C, which would be great in use but, unfortunately, prevents melt processing. Therefore, various copolymer derivatives of PA6T are on the market, made of TPA and other aromatic acids or aliphatic acids and melting at about 300[degrees]C (Ems-Chemie, BASF, Mitsui, DuPont, Solvay). A problem of the polyphthalamides is still the price.

In this study, various routes towards polyphthalamides were investigated kinetically (Scheme 1). The amidation of diacids and diesters is treated in this first part and that of polyesters, in particular of PET, in the second part.

The kinetics of the diamine-diacid polycondensation, which is the conventional route in industries, were monitored directly. Terephthalic and isophthalic acids were used and, for comparison, also adipic acid, which leads to the fairly well investigated PA66 [1-4].

The kinetics of the amine-ester condensation were studied more indirectly, using model systems. It is widely known that the diamine-diester polycondensation does not work properly, yielding only short-chained oligoamides due to side reactions. Some authors found special ways to deal with this situation: the esters were hydrolyzed [5, 6] or converted into monoamides [7] prior to the condensation, or activated esters were used that are amidated already at low temperatures [8-11]. But the direct condensation of simple diamines and simple diesters is addressed in only a few publications, all dealing with solutions, not with bulk systems [12-16]. Some mention alkylated amines in the products. Only one paper, on fully aromatic polyamides made from dimethylesters, offers an explanation for these byproducts: Flannigan and Mortimer [14] proposed a competition of amidation and alkylation as formulated in Scheme 2. The alkylation of the amine was believed to stop the polycondensation. This hypothesis was tested, quantitatively, in ester-amine model systems that were unable to polymerize (as in Scheme 2), which were allowed to react at high temperatures and were analyzed chromatographically.

[GRAPHIC OMITTED]

One aim of this work was to provide a basis for a process whereby polyethylene terephthalate (PET) can be amidated. PET waste is in rich supply from PET bottles, of which more than 10 megatons are produced annually. The postconsumer PET is still of high quality and accordingly much in demand. The bottles are either recycled into new bottles or, increasingly, processed into fibers. Amidating PET waste to produce polyphthalamides may become a new attractive alternative. It will be discussed in Part II as a process of real upcycling. The polyphthalamides are distinguished by outstanding properties that enable them to compete even with metals.

EXPERIMENTAL

Amines and Acids

Ammonium salts were prepared first by dissolving the amine in water and then the stoichiometric amount of acid was added, so an overall 50 wt% aqueous slurry resulted. This solution was heated to dissolve the acids and then diluted with ethanol (1:1 by weight) to precipitate the salt. After cooling and storing for I day, the salt was isolated by filtration, washed with ethanol. and dried at 100[degrees]C. Salts of different acids were prepared separately. The salts were then weighed into steel ampoules (cylindric, 5 x 30 [mm.sup.2]) together with the necessary water for the desired content.

[GRAPHIC OMITTED]

W = [m.sub.water]/[[m.sub.salt] + [m.sub.water]]. (1)

After the ampoules were sealed under argon and heated to 150-180[degrees]C, the slurry turned transparent (as tested with glass ampoules). Then the ampoules were transferred into a metal block oven where the polycondensation proceeded isothermally. A salt-water mixture with the water content W has the following initial molal concentrations (in mol [kg.sup.-1], almost equal to mol [1.sup.-1]):

2[c.sub.salt](t = 0) = [c.sub.a min](t = 0) = [c.sub.acid](t = 0)

= [2/[M.sub.salt]][[m.sub.salt]/[[m.sub.salt] + [m.sub.water]]] = 2 [[1 - W]/[M.sub.salt]]

[c.sub.water](t = 0) = [1/[M.sub.water]][[m.sub.water]/[[m.sub.salt] + [m.sub.water]]] = W/[M.sub.water]. (2)

When the condensation is underway, the system is characterized by the conversion p(t) and the initial water content w, which are defined as:

p(t) = c/[2[c.sub.salt](t = 0)] = 1 - [[[c.sub.N[H.sub.2]](t)]/[[c.sub.N[H.sub.2]](t = 0)]]

w = [[c.sub.water]/[2[c.sub.salt]]](t = 0) = [[M.sub.salt]/[2[M.sub.water]]] X [W/[1 - W]]

[M.sub.salt] = 0.282 kg/mol, [M.sub.water] = 0.018 kg/mol. (3)

Since the molal water content w is a bit abstract, the salt-water feeds of all condensation runs will be characterized by the water mass fraction W.

The conversion p(t) was measured by dissolving the samples in cresol and some drops of isopropanol and then titrating the amino functions potentiometrically with perchloric acid (Titrino DMS 716, Metrohm).

Amines and Esters

Model Reactions. Methylbenzoate (BMe) was mixed with hexylamine (A) or hexamethylene diamine (DA), without a solvent, in glass ampoules (cylindric, 5 x 30 [mm.sup.2]) which were sealed under nitrogen. The ampoules were then annealed at constant temperature for various times in a metal block oven. The products were analyzed by gas chromatography, dissolved in N-methylpyrrolidone (GC, Carlo Erba HRGC 5160 Mega Series and Agilent Trace GC 2000) at a heating rate of 40 K [min.sup.-1] inside 120-320[degrees]C. Products were identified by [.sup.1.H]-NMR spectroscopy. Calibration factors for the GC peak areas were determined using model compounds and the laws of mass conservation. Always, care was taken to keep the mixtures homogeneous in all stages, without precipitates and without loss of volatile components.

Polyamides. Dimethylterephthalate was reacted with hexamethylene diamine in a glass reactor equipped with a stirrer, a reflux condenser, a temperature controller, and a thermometer. The molar mass of the product was measured by gel chromatography (GPC) in hexafluoropropanol.

RESULTS AND DISCUSSION

Amines and Acids

The polycondensation of hexamethylene diamine and a mixture of terephthalic and isophthalic acid (T:I = 7:3) to the polyphthalamide PA6T6I (poly[imino-(l,4-co-l, 3-dioxophenylene)-iminohexamethylenel]) was studied. PA6T6I is a product of Ems-Chemie (Grivory XE3733NK) and melts at 305[degrees]C. In a process similar to that used for PA66, PA6T6I is prepared first by homogenizing an ammonium salt slurry with some water in a reservoir tank, then oligocondensing the solution isothermally in a reactor at temperatures >200[degrees]C, where most of the water is removed, and finally polycondensing the oligoamide in a degassed extruder. Our investigation dealt with the oligocondensation which is controlled by the two parameters indicated in Scheme 3, the initial water content w in the feed and the amide conversion p (Eq. 3).

The main processes, indicated in Scheme 3, have been discussed at length in papers on the kinetics of the polycondensation for PA66 [1-4] and PA6 [17-21]. The ammonium salt is in equilibrium with the nonionic diamine and diacid. According to most authors, the nonionic monomers are reactive, forming the amide. The amidation is reversible, the rate constants [k.sub.1] and [k.sub.-1] describing the condensation and the decondensation. Most papers report second-order kinetics, with acid catalysis only in the late stages of the condensation [17, 18, 22-24]. However, a retarding effect of the water was observed.

[GRAPHIC OMITTED]

The condensation to PA6T61 turned out to proceed quite similarly but more slowly. The kinetics were investigated in salt-water mixtures at water contents in the range of 0 < W < 40 wt%, in the wide temperature range of 180-270[degrees]C. The whole set of measured conversion curves p(t) could well be fitted with the second-order rate equation described in Scheme 3.

dp/dt = [q.sub.1][1 - p(t)][.sup.2] - [q.sub.-1]p(t)[p(t) + w]. (4)

An underlying assumption is that the reactivity of the amino, acid, and amide functions does not change as the chains grow longer [25]. The rate parameters [q.sub.1] and [q.sub.-1] in Eq. 4 (in [s.sup.-1]) are related to the rate constants [k.sub.1] and [k.sub.-1] in Scheme 3 (in kg [mol.sup.-1] [s.sup.-1]) by Eq. 2.

[q.sub.i] = 2[c.sub.salt][k.sub.i] i = 1, -1. (5)

The equilibrium constant K is given by

K = [q.sub.1]/[q.sub.-1] = [k.sub.1]/[k.sub.-1] = [[p.sub.eq]([p.sub.eq] + w)]/[(1 - [p.sub.eq])[.sup.2]]. (6)

To obtain the rate parameter [q.sub.1] from conversion curves ([q.sub.-1] resulting from [q.sub.1] and Eq. 6), Eq. 4 can be solved analytically and integrated to

1/p(t) = 1 + [w/2K] + [D/2K] [[1 + exp(-[[D[q.sub.1]]/K]t)]/[1 - exp(-[[D[q.sub.1]]/K]t)]]

D = [square root of (4K(1 + w) + [w.sup.2])]. (7)

However, fitting Eq. 4 numerically to the data is more convenient.

Two sets of conversion curves p(t) are shown in Fig. 1. The polycondensation is faster at higher temperatures T and lower water contents W. The equilibrium conversion [p.sub.eq] decreases as more water is added (Fig. 1b). Tests with only isophthalic acid yielded the same curves, which means that the terephthalic and isophthalic acid react at equal rates. This was confirmed by a [.sup.1.H]-NMR analysis on the oligoamides. This means that the PA6T6I chains contain the two acids in a random sequence.

Basically, the curves in Fig. I behave as expected: the temperature must accelerate the reactions and the water must retard them since it favors the decondensation (Fig. 2). But these reactions are kinetically not simple. The evaluation of the conversion curves with Eqs. 4 and 6 yields the curves of the equilibrium constant K and the condensation rate constant [k.sub.1] in Fig. 2. Both "constants" depend on the water content.

[FIGURE 1 OMITTED]

The curves K(W) and [k.sub.1](W) in Fig. 2 can be described by almost equal exponentials:

K(W) = [K.sub.W[right arrow]1] + [DELTA]K[e.sup.-zW]

with [K.sub.W[right arrow]1] = 20, [DELTA]K = [K.sub.W=0] - [K.sub.W[right arrow]1] = 5[K.sub.W[right arrow]1], z = 6 (8)

[k.sub.1](W) = [k.sub.1,W[right arrow]1] + [DELTA][k.sub.1][e.sup.-[z.sub.1]W]

with [DELTA][k.sub.1] = [k.sub.1,W=0] - [k.sub.1,W[right arrow]1] = 4[k.sub.1,W[right arrow]1], z = 6 (9)

where [K.sub.w=0] and [k.sub.1,W=0] are the values in the bulk system without water and [K.sub.w[right arrow]1] and [k.sub.1,W[right arrow]1] are the extrapolated values for the dilute aqueous solution. The consequence of these equations is that the rate constant [k.sub.-1] for the decondensation is almost independent of water, since the effects of the water on K and [k.sub.1] are almost equal:

[k.sub.-1] [congruent to] [[k.sub.1](W)]/K(W) [congruent to] const. (10)

[FIGURE 2 OMITTED]

The fact that K and [k.sub.1] are not true constants is not surprising. When complex chemical processes comprising several elementary reactions are described with a single equation, the rate parameters are only "apparent constants." The amine-acid condensation is such a complex process, involving at least the ionic-nonionic equilibrium of the monomers (Scheme 3) and, in the amidation reaction, the steps of amine addition and water elimination. Therefore, [k.sub.1], K, and [k.sub.-1] are only apparent constants.

More difficult is the interpretation of the effects of water observed in Fig. 2. The condensation is fastest and goes furthest without water. Water slows it down and favors the decondensation. Similar phenomena have been reported for PA66 [1-4], PA6 [17-21], and polyesters [22, 26]. Actually, Eqs. 8 and 9 hold almost exactly also for PA66 [1, 2]. Various effects have been proposed to explain the behavior of K and [k.sub.1] [27-37]: water increases the dielectric constant [27, 28], it ionizes N[H.sub.2]--COOH pairs, and it hydrates them [27-31]. But a generally accepted model does not yet exist. This issue will be addressed again later.

The effects of the temperature are again similar to those observed with PA66 [1, 2]. The equilibrium constant K did not change noticeably with temperature (Fig. 2a), so Eq. 8 holds at all relevant temperatures. The rate constant [k.sub.1](T) follows an Arrhenius law. In Fig. 3, the [k.sub.1](T) data are displayed not in the usual Arrhenius fashion but, for better readability, in terms of the half-life time

[[tau].sub.1/2] = 1/[k.sub.1] (11)

as a function of degrees Celsius. The total equation for [k.sub.1](W,T), written for a reference temperature at T* = 220[degrees]C, is given by (with Eq. 9)

[k.sub.1](W, T) = ([k*.sub.1,W[right arrow]1] + [DELTA][k.sub.1][e.sup.-[z.sub.1]W]) exp[[[E.sup.[not equal to]]/R]([1/T*] - [1/T])]

[k*.sub.1,W[right arrow]1] = [k.sub.1,W[right arrow]1] (T*) = 1.5 X [10.sup.-5] kg/mol s,

[E.sup.[not equal to]] = 130 kJ/mol, T* = 220[degrees]C. (12)

Since the process is practically athermal, [E.sup.[not equal to]] holds approximately both for [k.sub.1] and [k.sub.-1].

To summarize: phthalic acids condense with an aliphatic diamine much like adipic acid. But the half-life times in Fig. 3 demonstrate that PA66 is formed much faster than PA6T6I. Conversion curves p(t) for PA66 and PA6T6I in Fig. 4a reflect the big difference. The PA66 curve coincides with data published in Refs. 1 and 2.

The condensation to PA66 is not only faster but it goes also further. Selected data in Table 1 demonstrate that both [k.sub.1] and K for PA66 exceed those for PA6T6I by an order of magnitude. This means that the rate constant [k.sub.-1] of the decondensation is practically equal in both systems:

[FIGURE 3 OMITTED]

[FIGURE 4 OMITTED]

[K.sub.66] [congruent to] 10[K.sub.6T6I] [k.sub.1.sup.66] [congruent to] 10[k.sub.1.sup.6T6I] [k.sub.-1.sup.66] [congruent to] [k.sub.-1.sup.6T6I]. (13)

To demonstrate these differences in the reactivity of aliphatic and aromatic acids, copolyamides PA6T6I66 and PA6T6I612 were prepared from the two phthalic acids and the aliphatic adipic or dodecandioic acid (T:I:6 = T:I:12 = 7:3:10). The condensation kinetics of these copolyamides are compared in Fig. 4 with those of PA66 and PA6T6I. The conversion curves of PA6T6I66 and P6T6I612 first come up fast, almost as fast as that of PA66, but then slow down markedly. This indicates that the aliphatic acids react first and the aromatic acids afterwards. To prove this, the curves in Fig. 4 were calculated with an appropriately extended version of Eq. 4, containing two separate conversions, [p.sub.arom] and [p.sub.aliph] (the total conversion being p = [p.sub.arom] + [p.sub.aliph]):

[d[p.sub.arom]]/dt = [[q.sub.1,arom]/2] (1 - 2[p.sub.arom])(1 - p) - [q.sub.-1,arom][p.sub.arom](p + w)

[d[p.sub.aliph]]/dt = [[q.sub.1,aliph]/2](1 - 2[p.sub.aliph])(1 - p) - [q.sub.-1,aliph][p.sub.aliph](p + w)

dp/dt = [[d[p.sub.arom]]/dt] + [[d[p.sub.aliph]]/dt]. (14)

At this point, an interpretation of the acid-amine condensation kinetics may be attempted. The special effects can be understood assuming that the condensation-decondensation reactions themselves are kinetically simple and the complications are introduced only by the ionic-nonionic equilibrium of the monomers, which is formulated in Scheme 3. Three observations support this view: firstly, the ionic ammonium salt is more stable in mixtures with more water so the concentration of the nonionic monomers is diminished. This slows the condensation down, which is reflected by a decrease of the apparent rate constant [k.sub.1] at higher water contents (Fig. 2). Secondly, the acidity of the aromatic acids (pK [congruent to] 4.2) exceeds that of the aliphatic acids (pK [congruent to] 4.9) so the ammonium salt of the former is more stable than that of the latter. This slows down the PA6T6I process more than the PA66 process ([k.sub.1.sup.6T6I] < [k.sub.1.sup.66], Eq. 13). Thirdly, the decondensation reaction does not involve an ionic species. Therefore, the rate constant [k.sub.-1] does not depend on the water (Eq. 10) and is similar for PA6T6I and PA66 ([k.sub.-1.sup.6T6I] [congruent to] [k.sub.-1.sup.66], Eq. 13).

Amines and Esters

In Fig. 5a, the degree of polymerisation [P.sub.n] of the diamine-diacid polycondensation is shown as a function of the conversion p. The characteristic curve, which remains low up to high conversions and rises then steeply, is given by Carothers equation:

[FIGURE 5 OMITTED]

[P.sub.n] = 1/[1 - p]. (15)

The late stage with the steep rise was actually not attained in the previous section where the oligocondensations, proceeding in closed ampoules, ended at fairly low equilibrium conversions [p.sub.eq] (Fig. 1b). To show that the chain length could further grow, some ampoules were opened and the water was removed to reactivate the condensation. As shown in Fig. 5a, [P.sub.n] length increased dramatically as predicted by Eq. 15.

This, however, did not happen when the diamine was condensed with the diester dimethylterephthalate. In an open system, where all methanol was distilled off, only an oligoamide end product was obtained (after 2 h, 270[degrees]C). It is characterized by the GPC curve in Fig. 5b. According to MALDI spectrometry, the short chains, on average only about five monomers long, carried methylated amine and acid end groups which apparently were unable to condense (Scheme 4).

Several compounds known to catalyze polycondensations and transcondensations, such as hypophosphites, triphenylphosphite, and dibutyltindilaurate, were tested without much success. The chains remained short.

Therefore, the amidation of esters was analyzed in all detail, using nonpolymeric model systems which could be analyzed much more accurately than polymer systems. BMe was amidated with hexylamine (A) and hexamethylene diamine (DA). These components should have yielded the amides BA, BDA, and [B.sub.2]DA in Scheme 5.

Conversion curves, obtained from GC analysis, are displayed in Fig. 6. The mixtures BMe-A and BMe-DA (molar 1:1) yielded indeed these amides. The educts and products in Fig. 6a and c seem to behave more or less as expected. However, the GC trace of a BMe-DA sample in Fig. 7 reveals that there are more products. In fact, the trace is littered with byproducts. Benzoic acid B is present and also all conceivable methylated derivatives DAM[e.sub.1-4] and BDAM[e.sub.1,2] of the diamine DA and of the monoamide BDA. The conversion curves of the major byproducts are shown in Fig. 6b and d (the GC peak of the acid B, easily recognized by its characteristic sawtooth form, was not evaluated because, not unusual with acids, it yielded unreliable results).

[GRAPHIC OMITTED]

As was already pointed out by Flannigan and Mortimer [14], the easiest explanation for these methylated amines and amides is not valid: the methanol which is produced in the amidation process could have alkylated the amine as shown in Scheme 6. But this definitely does not happen under the reaction conditions.

A consistent explanation for all products is formulated in Scheme 7 for the BMe-A system: the amine is alkylated directly by the ester. In principle, an amine and an ester can react in two ways: the nitrogen can attack the carboxyl group either at the carbonyl carbon, which results in an amidation to BA, or at the alkyl carbon, which results in an amine alkylation to AMe. Both processes are aided by favorable transition states. In the amidation process, methanol is eliminated from the ester, but in the alkylation process, only the methyl group leaves the ester.

This kind of parallel reactions is known as an acyl-alkyl competition. In the acidic hydrolysis of esters, for instance, methyl esters follow the acyl route, eliminating methanol, while tert-butyl esters follow the alkyl route, eliminating isobutene. But so far, an acyl-alkyl competition was not reported for the amidation, probably because it happens only at high temperatures.

Further reactions which do or do not occur are indicated in Scheme 7. Firstly, the acid B is amidated as well. This amine-acid condensation was discussed in the previous section on PA6T6I. Secondly, the methylated amine AMe is again methylated to the dimethylated amine AM[e.sub.2]. Thirdly and most importantly, the monomethylated amine AMe does not condense anymore with BMe: the methylated amide BAMe was never detected among the products (it is also not formed by methylation of the amide BA).

The data in Fig. 6 were fitted numerically with the appropriate sets of second-order rate equations. The less complex set for the system BMe-A reads:

-[[d[c.sub.BMe]]/dt] = [c.sub.BMe][(k + [k.sub.alk])[c.sub.A] - 2[k.sub.alk][c.sub.Ame]] - [[d[c.sub.A]]/dt] = [c.sub.A][(k + [k.sub.alk])[c.sub.BMe] + [k.sub.acid][c.sub.B]]

[d[c.sub.AMe]]/dt = [k.sub.alk][c.sub.BMe][[c.sub.A] - 2[c.sub.AMe]] [d[c.sub.AMe2]]/dt = 2[k.sub.alk][c.sub.BMe][c.sub.AMe]

[d[c.sub.B]]/dt = [k.sub.alk][c.sub.BMe][[c.sub.A] + 2[c.sub.AMe]] - [k.sub.acid][c.sub.B][c.sub.A] [d[c.sub.BA]]/dt = k[c.sub.BMe][c.sub.A] + [k.sub.acid][c.sub.B][c.sub.A]. (16)

[GRAPHIC OMITTED]

The complex set of equations for the BMe-DA system is found in the Appendix (Eq. A1). Three rate constants parameterize Eqs. 16 and A1, k for the ester amidation, [k.sub.acid] for the acid amidation, and [k.sub.alk] for the methylation of the amine. Making allowance for the imperfections of a GC analysis on systems with amines and acids, the fit of all curves is satisfactory. The BMe-A and BMe-DA systems yielded the same constants:

[FIGURE 6 OMITTED]

[k.sub.170[degrees]C] = 8 x [10.sup.-5]kg/(mol s) [k.sub.alk] = 0.2k [k.sub.acid] [congruent to] 0.1k. (17)

The rate constant [k.sub.acid] for the acid amidation comes from Eq. 12. Only one complication had to be introduced, and only in the system BMe-A where [k.sub.alk] describes only the first methylation step, A[right arrow]AMe. The second step, AMe [right arrow] AM[e.sub.2], turned out to be twice as fast (2[k.sub.alk]), evidently because AMe is more basic than A.

It is very evident from [k.sub.alk] in Eq. 17 that the methylation cannot be neglected. The main and the side reaction proceed on the same order of magnitude. The kinetics were studied at various temperatures. As shown in Fig. 8, the half-life times [[tau].sub.1/2] (Eq. 11) of the acid and the ester amidation follow similar Arrhenius laws. The condensation of the amine with BMe is about an order of magnitude faster than that with TPA. By coincidence, therefore, the amidation of an aromatic methylester is as fast as the amidation of the aliphatic adipic acid (Fig. 3).

The kinetics of a stoichiometric mixture BMe:DA (molar 2:1) at 240[degrees]C are displayed in Fig. 9. The diagram shows the late stages better than Fig. 6 where the temperature was 170[degrees]C. Only components which are still present in the final product are shown in Fig. 9. If there had been no side reactions, only the diamide [B.sub.2]DA would have been there, which would have approached a concentration of [c.sub.B2DA] = 50% (which is the limit because [B.sub.2]DA contains two benzoate groups). Instead, there is only half as much [B.sub.2]DA, due to the byproducts. The alkylated amines DAM[e.sub.1-4] are less important than the alkylated amides BDAM[e.sub.1,2] which dominate together with the benzoic acid. The acid is actually the main product.

This diagram is the explanation for Scheme 4: in a polymerizing system, the acid units B and their counterparts, the methylated amines and amides, are permanent chain ends. The expected degree of polymerization of an oligoamide is, therefore, given by the acid concentration [c.sub.B](t [right arrow] [proportional]) [congruent to] 25% (Fig. 9):

[FIGURE 7 OMITTED]

[P.sub.n] = 1/[[c.sub.B](t [right arrow] [infinity])] [congruent to] 4. (18)

This prediction agrees with the observed [P.sub.n] value indicated in Fig. 5a.

A troubling observation in Fig. 9 is that the final product contains so much acid. One may believe that this is due to the fact that the acid amidation is so slow, one order of magnitude slower than the ester amidation (Eq. 17). But model calculations with Eqs. 16 and A1 proved otherwise. Even a much faster acid condensation would have changed little because the acid competes with the ester for the primary amine functions.

This left catalysts as the last remedy to save the ester amidation. As mentioned earlier, various catalysts were tested, in the hope that they would favor the amidation and suppress the methylation. This hope was not unfounded: according to Scheme 7, both reactions run in parallel, and so catalysts should have a chance, but their effects were weak. Only dibutyltindilaurate suppressed the amine alkylation markedly, at 170[degrees]C. At higher temperatures, even this catalytic effect disappeared and was entirely gone above 240[degrees]C where polyamides are normally produced.

[GRAPHIC OMITTED]

CONCLUSIONS

The polycondensations of diamines with diacids and diesters are very different. The condensation of phthalic acids with diamines proceeds kinetically almost exactly like the corresponding polycondensation of adipic acid. The only difference is that the aromatic acids react about one order of magnitude more slowly. While this diamine-diacid condensation led without problems to long-chained polyphthalamides, the diamine-diester condensation failed, coming to an early halt at the stage of oligoamides. Systematic investigations on model compounds revealed that the diamine is being alkylated by the ester in a side reaction of considerable intensity. The alkylated amino functions thus created cannot condense anymore. In polycondensations, they form permanent chain ends. No catalyst was found that could suppress this side reaction. It will be demonstrated in Part II that, despite this failure of the polycondensation of diamines and diesters, polyesters like PET can successfully be amidated to long-chained polyphthalamides.

[GRAPHIC OMITTED]

[FIGURE 8 OMITTED]

[FIGURE 9 OMITTED]

APPENDIX

The set of second-order rate equations for the mixtures BMe-DA is given by

-[[d[c.sub.BMe]]/dt] = k[c.sub.BMe][2(1 + 2r)[c.sub.DA] + (1 + 3r)[c.sub.DAMe] + (0.333 + 2r)[c.sub.DAMe2] + r[c.sub.DAMe3] + (1 + 2r)[c.sub.BDA] + r[c.sub.BDAMe]]

-[[d[c.sub.DA]]/dt] = 2k[c.sub.BMe](1 + 2r)[c.sub.DA] + 2qk[c.sub.B][c.sub.DA]

[d[c.sub.B]]/dt = rk[c.sub.BMe][4[c.sub.DA] + 3[c.sub.DAMe] + 2[c.sub.DAMe2] + [c.sub.DAMe3] + 2[c.sub.BDA] + [c.sub.BDAMe]] - qk[c.sub.B][2[c.sub.DA] + [c.sub.DAMe] + 0.333[c.sub.DAMe2] + [c.sub.BDA]]

[d[c.sub.DAMe]]/dt = k[c.sub.BMe][4r[c.sub.DA] - (1 + 3r)[c.sub.DAMe] - q[c.sub.B][c.sub.DAMe]]

[d[c.sub.DAMe2]]/dt = k[c.sub.BMe][3r[c.sub.DAMe] - (0.333 + 2r)[c.sub.DAMe2] - 0.333q[c.sub.B][c.sub.DAMe2]]

[d[c.sub.DAMe3]]/dt = rk[c.sub.BMe][2[c.sub.DAMe2] - [c.sub.DAMe3]]

[d[c.sub.DAMe4]]/dt = rk[c.sub.BMe][c.sub.DAMe3]

[d[c.sub.BDA]]/dt = k[c.sub.BMe][2[c.sub.DA] - (1 + 2r)[c.sub.BDA]] + qk[c.sub.B][2[c.sub.DA] + [c.sub.BDA]]

[d[c.sub.BDAMe]]/dt = k[c.sub.BMe][[c.sub.DAMe] + 2r[c.sub.BDA] - r[c.sub.BDAMe]] + qk[c.sub.B][c.sub.DAMe]

[d[c.sub.BDAMe2]]/dt = k[c.sub.BMe][0.333[c.sub.DAMe2] + r[c.sub.BDAMe]] + 0.333qk[c.sub.B][c.sub.DAMe2]

[d[c.sub.B2DA]]/dt = k[c.sub.BMe][c.sub.BDA] + qk[c.sub.B][c.sub.BDA]. (A1)

Frequency factors are taken into account (for instance, the diamine DA can be amidated twice but ethylated four times), and two diethylated amines DAE[t.sub.2] were distinguished, one symmetric and one asymmetric, of which only the asymmetric (probability 0.333) can still amidate the ester. The rate constants are k (in the differentials), [k.sub.alk], and [k.sub.acid], represented by the two ratios

r = [k.sub.alk]/k q = [k.sub.acid]/k. (A2)

REFERENCES

1. N. Ogata, Macromol. Chem., 42, 52 (1960).

2. N. Ogata, Macromol. Chem., 43, 117 (1961).

3. F. Wiloth, Macromol. Chem., 144, 329 (1971).

4. A. Kumar, S. Kuruville, A.R. Raman, and S.K. Gupta, Polymer, 22, 387 (1981).

5. G. Schade and F. Blaschke, Dynamit Nobel, DE 1495393B2, Eur. Patent 1495393 (1979).

6. G.E. Kosmin, Dynamit Nobel, DE 1251520, Eur. Patent 1251520 (1968).

7. F. Jouffret and P.J. Madec, J. Polym. Sci. Part A: Polym. Chem., 34, 2363 (1996).

8. N. Ogata, K. Sanui, and K. Iijima, J. Polym. Sci. Polym. Chem. Ed., 11, 1095 (1973).

9. N. Ogata, K. Sanui, H. Tanaka, and T. Suzuki, J. Polym. Sci. Polym. Chem. Ed., 15, 2531 (1977).

10. N. Ogata, K. Sanui, T. Ohtake, and H. Nakamura, Polym. J., 11, 827 (1979).

11. K. Weisskopf and G. Meyerhoff, Makromol. Chem., 187, 401, 411, (1986).

12. S.L. Kwolek and P.W. Morgan, J. Polym. Sci. Part A: Gen. Pap., 2, 2693 (1964).

13. S.W. Shalaby, E.A. Turi, and P.J. Harget, J. Polym. Sci. Polym. Chem. Ed., 14, 2407 (1976).

14. J.E. Flannigan and G.A. Mortimer, J. Polym. Sci. Polym. Chem. Ed., 16, 1221 (1978).

15. S. Nakano and T. Kato, J. Polym. Sci. Part A: Polym. Chem., 37, 1413 (1999).

16. F. Jouffret and P.J. Madec, J. Polym. Sci. Part A: Polym. Chem., 34, 2363 (1996).

17. C. Giori and R.T. Hayes, J. Polym. Sci. Part A-I: Polym. Chem., 8, 351 (1970).

18. K. Thai, H. Teranishi, Y. Arai, and T. Tagawa, J. Appl. Polym. Sci., 24, 211 (1979).

19. F. Wiloth, Z. Phys. Chem. (Munich), N.F.4, 66 (1955).

20. K. Thai and T. Tagawa, Ind. Eng. Chem. Prod. Res. Dev., 22, 192 (1983).

21. H.K. Reimschussel and K. Nagasubramanian, Chem. Eng. Sci., 27, 1119 (1972).

22. P.J. Flory, Chem. Rev., 39, 137 (1946).

23. J. Colonge and E. Fickett, Bull. Soc. Chim., 412 (1955).

24. G. Champetier and R. Vorgos, Recl. Trav. Chim. Pays-Bas, 69, 85 (1950).

25. O.B. Edgar and R.J. Hill, J. Polym. Sci., 8, 1 (1952).

26. (a) A.J. Yu and R.D. Evans, J. Am. Chem. Soc., 81, 5361 (1959); (b) AJ. Yu and R.D. Evans, J. Polym. Sci., 42, 249 (1960).

27. T.C. Tranter, J. Polym. Sci. Part A: Gen. Pap., 2, 4289 (1964).

28. J.S. Ridgway, J. Polym. Sci. Part A-I: Polym. Chem., 8, 3089 (1970).

29. R.J. Gaymans and S. Aalto, J. Polym. Sci. Part A: Polym. Chem., 27, 423 (1989).

30. F.B. Cramer and R.G. Beaman, J. Polym. Sci., 21, 237 (1956).

31. D. Heikens, J. Polym. Sci., 22, 65 (1956).

32. P.J. Flory, J. Am. Chem. Soc., 61, 3334 (1939).

33. D.D. Steppan, M.F. Doherty, and M.F. Malone, J. Appl. Polym. Sci., 33, 2333 (1987).

34. F.K. Mallon and W.H. Ray, J. Appl. Polym. Sci., 69, 1213 (1998).

35. R. Giudici, C.A.O. Nascimento, A. Tresmondi, A. Domingues, and R. Pellicciotta, Chem. Eng. Sci., 54, 3243 (1999).

36. H. Morawetz and P.S. Otaki, J. Am. Chem. Soc., 85, 463 (1963).

37. M. Roses, Anal. Chim. Acta, 276, 223 (1993).

Jan Malluche, (1) Goetz P. Hellmann, (1) Manfred Hewel, (2) Hanns-Jorg Liedloff (2)

(1) German Institute for Polymers (Deutsches Kunststoff-Institut, OKI), Schlossgartenstrasse 6, D-64289 Darmstadt, Germany

(2) Ems-Chemie AG, Reichenauerstrasse, CH-7013 Domat/Ems, Switzerland

Correspondence to: G.P. Hellmann; e-mail: [e-mail]

Contract grant sponsor: Federal Ministry of Economics and Labour (Bundesministerium fur Wirtschaft und Arbeit, BMWA); contract grant numbers: AiF-No 12944, 13563.

TABLE 1. Rate constant [k.sub.1] and equilibrium constant K values for
polyamides PA66 and PA6T6I prepared with water W = 21 wt%.

           K ([10.sup.-5] kg         [k.sub.1] ([10.sup.-5] kg
           [mol.sup.-1] [s.sup.-1])  [mol.sup.-1] [s.sup.-1])
Polyamide  220 (a)  250              220  250

PA66       570      430              45   185
PA66 (b)   350                       45
PA6T6I      46       38               4    16

(a) The numbers 220 and 250 indicate temperatures T ([degrees]C).
(b) From Refs. 1 and 2.