Amorphous polyamide coating resins from sugar-derived monomers

In this article, the synthesis of bio-based polyamides for powder coating applications and their evaluation in a solventborne coating system are reported. The Mn values of the resins were between 3000 and 4000 g mol−1 and the resins displayed Tg values from 60 to 80°C. Both amine and carboxylic acid functionalities (total ~0.6 mmol g−1) were introduced for curing purposes. The resins were cured with triglycidyl isocyanurate (TGIC) or N,N,N′,N′-tetrakis(2-hydroxyethyl)adipamide (Primid XL-552). The curing reaction was followed using rheology which indicated that TGIC achieved higher reaction rates and higher gel contents. The DSC analysis of the cured disks showed that all cured samples were amorphous as is desired for the targeted coating application. The resins required a curing temperature higher than 150°C. Aluminum panels were coated using a solventborne approach and the coatings were cured at 180°C during 1 h. Dewetting was observed on all panels. Network formation was adequate for an amine-functional resin cured with TGIC as indicated by solvent resistance testing. In conclusion, the developed bio-based polyamide resins are promising materials to be used as binder resins in powder coating applications.


U N C O R R E C T E D P R O O F Introduction
As emission regulations are becoming more stringent, powder coatings are very interesting solvent-free coating systems.Coating of metal substrates with this technique is the largest market but many other materials like wood can also be coated, because nowadays curing can be successfully performed at lower temperatures.[3][4][5] The dry paint which constitutes a powder coating, consists of a binder resin, a curing agent, pigments, and additives to enhance properties such as flow and to reduce film imperfections. 6,7The different components are dispersed into the resin using an extruder.The residence time in the extruder is very short ($ 30 s) to prevent premature cross-linking, which may occur due to the relatively high temperature needed to facilitate flow of the resin.For coating appearance reasons, the final coating needs to be amorphous, so the binders should have a low degree of crystallinity or should be completely amorphous.
Several binder-curing agent combinations exist.Often, the binders in the powder paint are polyesters with a relatively low molecular weight (M n = 2000-6000 g mol À1 ).These polyesters are typically based on bifunctional monomers, e.g., terephthalic acid, isophthalic acid, adipic acid, neopentyl glycol, and ethylene glycol, with the addition of a small percentage of trifunctional monomers, e.g., trimellitic anhydride and ][10] These compounds are derived from petrochemicals.[13][14][15][16] Storage of the powder paint imposes requirements on the glass transition temperature (T g ) value of the used polymer resin, which has to be significantly higher than ambient temperature.Storage above the T g of the powder will cause fusion of the particles, decreasing the ability of the coating to level well or even completely hampering paint application.For these reasons, the binder T g value should exceed 60°C.As a consequence, the resulting coating will typically have a T g value of more than 70°C due to the curing.
The polyester binders can have either hydroxyl-or carboxylic acid-functionality.The curing agents typically used for bifunctional polyesters must have a functionality of more than two.][19] Besides reaction with carboxylic acids, TGIC can also react with hydroxyl groups and amines.However, its use in Europe is declining because of its toxicity.For carboxylic acid-functional resins, b-hydroxyalkylamides have become popular under the trade name Primid. 1,2,7,20This type of curing agents are non-toxic but can only react with carboxylic acid groups via an oxazolinium intermediate.This reaction cannot be catalyzed. 21,22e curing reaction of thermosets can be followed by molecular changes, the disappearance of end groups, the energy required for the reaction, and by measuring material properties.4][25][26][27] Sufficient heat will allow the powder paint to flow, after which the viscosity will decrease with increasing temperature.When the curing reaction starts, the viscosity will increase again due to the formed cross-links.This makes rheological measurements convenient to monitor the curing as well as to assess the flow of the powder.
This article describes the synthesis of amorphous polyamide resins from sugar-derived monomers (Scheme 1) and their performance is tested in curing and coating experiments.Polyamides have been 115 selected for their good stability against hydrolysis and 116 solvents.Before applying the resins to aluminum 117 panels, rheology will be used to analyze the binder-118 curing agent combination.As mentioned, powder 119 coatings are formulated using additives.To make a 120 fair assessment of the properties of the binders 121 investigated, no formulation with additives such as 122 flow agents will be applied.

123
Experimental section F impeller, two argon inlets, and a distillation setup.Approximately 5 mL of methanol was added and the mixture was stirred under an inert atmosphere without gas flow at a set temperature of 100°C, to allow for oligomerization of the reactants during the initial phase of the reaction.After a day, the inert gas flow through the reactor was allowed to remove the methanol from the reaction mixture.The temperature was slowly increased to 170°C in the course of several hours, while gas flow through the reactor was prevented to retain the volatile BDA.After 2 h, the temperature was increased to 230°C and condensate was removed from the system.The condensation setup was removed 2 h later and the pressure was decreased to 3-4 hPa in the reactor.After another 2 h, the products were discharged from the reactor in the liquid state.
Compound A was obtained as a brittle, light colored, transparent material with a yield of 3.12 g (50.0 wt% of maximum achievable weight), and two separate fractions with a different composition of monomers (20.6 and 13.9 wt%).B1 was obtained directly from the reactor as a brittle, greenish, transparent compound (B1) with a yield of 1.949 g (39.8%).B2 was dissolved in 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) and was subsequently precipitated from a large excess of acetone and was dried after decantation of the solvent mixture.It was obtained as a brittle, graycolored, opaque solid with a yield of 2.729 g (55.8%).N.B. samples B1 and B2 originate from the same polymerization run, so that the overall yield of the reaction was 95.6%. A: The acid value (AV) and amine value (AmV) were calculated from NMR data rather than potentiometric end group titration due to the limited amount of resins available for analysis.(see Supplementary Information).With optimized NMR measurement settings and data analysis, the resulting error is around 10%.The end groups of the polymers were controlled by the feed ratio between PA, BDA, and IIDA.Polyamide A was carboxylic acid functional, while polyamides B1 and B2 were predominantly amine functional as determined by NMR spectroscopy.The lower AV and AmV of B2 compared to B1 were expected due to the precipitation step.This is caused by the fact that the shortest chains in B2 could remain in solution in the mixture of HFIP (solvent) and acetone (non-solvent, large excess) used for precipitation.The total functionality for these polyamides is in a suitable range for powder coating resins (target: $0.50-0.70 mmol g À1 ).
The molar ratio between IIDA and BDA was targeted at 60:40 to obtain amorphous polymers with an expected T g value of 70°C.The obtained polymers have a ratio of 66:34 between the respective diamines, which is fairly close to the aim.Evaporation of part of the relatively volatile BDA explains the observed discrepancy.The polyamide resins have been thermally characterized with DSC.The results are reported in Fig. SI1 and Table 1.During the first heating run, all three polyamides showed a moderate melting endotherm between 100 and 200°C with DH m,1 ranging from 10 to 17 J g À1 .This indicated that some ordered phases were present in the polymers after uncontrolled cooling due to collection of the material from the melt.Furthermore, B2 showed a sharp melting peak at 238°C.This can be explained by crystallized short PA-4,7 segments, which together with the observation of an increased ratio of IIDA chain ends in 1 H NMR suggests that the molecular composition was not completely random.The T g values determined from the first heating trace were between 45 and 58°C, which is well above room temperature.However, the 45°C for B2 was unexpectedly low in this first run, which is probably caused by the presence of some residual solvent from the work-up procedure.
The cooling runs showed no crystallization phenomena for these polyamides, which therefore remained 357 amorphous.In the second heating run, the T g values 358 showed some spread despite that the polymers had 359 similar M n values.The T g value of A was similar to the 360 value expected, while B1 was about 6°C lower and B2 361 much higher at a value of 89°C.The large difference in 362 the T g values for B2 and B1 is caused by the higher M n 363 value for B2 after precipitation, and the evaporation of 364 the solvent that effected its first heating run.All values 365 exceeded the requirement of 60°C for a powder 366 coating resin.
367 Rheology 368 Ultimately, the polyamides described in this paper are 369 intended to replace petrochemistry-based resins used 370 in powder coatings.Therefore, the synthesized resins 371 were analyzed in terms of their curing behavior with 372 two well-known curing agents: triglycidylisocyanurate 373 (TGIC, a trifunctional epoxy compound) and 374 N,N,N¢,N¢-tetrakis(2-hydroxyethyl)adipamide (Primid 375 XL-552, a tetrafunctional b-hydroxy alkamide).TGIC is capable of reacting with both acid-and aminefunctional resins, while the activated OH-groups of Primid will only react with acid functionalities.Crosslinking using Primid results in the release of water molecules as condensate.Resin A is expected to perform slightly better with Primid due to its higher functionality, while the resins B1 and B2 should in principle form a more dense network with TGIC than with Primid because of the presence of amine groups.
To follow the curing reactions between the crosslinkers and the resins, a rheometer was used.Sample disks were prepared by cold pressing the resins A and B1 and one of the cross-linkers at high pressure (1600 bar) and room temperature.The compositions of the samples are reported in Table 2.The higher amine value of B1 compared to that of A (0.47 vs 0 mmol g À1 , respectively) results in a significantly lower point of gelation (p c ) when TGIC is used.The presence of amine groups effectively increased the functionality of the polymer chain (f e0 ) from 2.0 (with Primid) to 3.7 (with TGIC).This is likely to result in differences in the final cross-link density and hence the 398 toughness and hardness of the final coating.When 399 Primid XL-552 was used, a different phenomenon 400 occurred.While for resin A all the end groups (i.e., 401 carboxylic acid groups) can be fully cross-linked, for 402 polyamide B1 only the carboxylic acid groups can 403 react, while the amine end groups remain as dangling 404 chains.Because the AmV is higher than the AV, the 405 largest fraction of the polymers will bear two amine 406 groups, i.e., one at both ends.Most of the remaining 407 chains will have one amine group and one carboxylic 408 acid group.This results in polymers that will have 409 reacted only once, and chains that were not cross-410 linked (i.e., bearing two amine groups) and remained 411 as free chains in the network.412 Sample disks were loaded in the rheometer at 70°C 413 and the curing reaction was followed during a temper-414 ature program (see Fig. 2).Unfortunately, some sam-415 ples (A + TGIC, A + Primid) had defects at the edges 416 which will affect the measurement.The measured 417 viscosities and moduli will be lower than their actual 418 values.An insufficient amount of material remained to 419 prepare new samples, so the results should be consid-420 ered in a qualitative and not in a quantitative manner.421 The samples were heated gradually to 150°C and kept 422 at this temperature for 60 min to cure.Subsequently, the temperature was increased to 200°C to assess the cross-linking in the system.
During heating, all samples behaved as solids until roughly 20 min (T = 110°C), at which point the graphs smoothened out as the material started to flow well above T g .For A + TGIC and A + Primid, fluctuations were observed in the first fifty minutes of the experiments.This is attributed to trapped air pockets that originate from the more coarse samples and macroscopic inhomogeneity resulting from sample preparation.Both samples were submitted to a trial experiment with a similar but different temperature profile in which these unexpected fluctuations were not observed.
Figure 2 shows the G¢, G †, and tan d of the rheological experiments.When G † is larger than G', the system behaves predominantly as a viscous liquid.At G¢ > G †, the system is behaving mostly like an elastic solid.Therefore, at G¢ = G †, or tan d = 1, the system transfers from liquid to solid behavior.This crossover point is regarded as the gel point.
After the rheological measurements, the sample disks were broken and the resulting pieces were analyzed using DSC (see Fig. 3).The first heating run showed two ranges for the T g values (49-61 and 72-87°C) for all four samples.The presence of two separate T g values in one material indicates that two phases were present.The first transition at lower temperatures was attributed to domains of pure resin with unreacted cross-linker.The cross-linker will act as a plasticizer and hence reduce the T g of the unreacted resin (see Table 1). 32The second transition was 21 to 35°C higher and is attributed to the T g value of the formed network.The T g of the network is usually higher than that of the resin.
All samples showed an exothermic reaction in the first heating run.For A + TGIC (11 J g À1 , 178 to 250°C) and B1 + TGIC (27 J g À1 , 164 to 250°C), these were rather clear, while for A + Primid and B1 + Primid the curing was observed mostly while the samples were kept isothermal at 250°C (t = 30-35 min).This indicated that for both polyamides, the reaction with TGIC proceeded more readily than with Primid.
During the cooling run, none of the samples showed crystallization, indicating that the networks were amorphous.A neat glass transition was observed.The second heating run showed only one T g value in the range of 78 to 81°C for all samples, indicating that the distinction between the domains in the samples had disappeared.This verifies that the T g value at 49-61°C was caused by unreacted species.Furthermore, no additional curing is observed in this heating run.The second cooling run is a copy of the first cooling run.The DSC thermograms of the A surface with a low surface energy will not be wetted properly, and contamination with hydrophobic particles (e.g., dust particles) can cause cratering.With Primid, the coating was smooth but a large amount of material flowed to the edges.The large difference between B1 and B2 is due to removal of low molecular weight chains during the precipitation step to isolate B2.These low molecular weight chains have a higher mobility than higher molecular weight polymers.
The thickness of the coatings was found to be around 10 lm which was roughly half of the calculated value.This can partly be explained by the application of the paint at 80°C.This decreased the viscosity and it was observed that all solutions flowed between the edge of the doctor blade and the panels.This produced a wider coating than intended, reducing its thickness.The color of the coatings is subject to thickness and substrate, but the color is reasonably light.
Several tests have been performed on the coated panels.The results are collected in Table 3.As mentioned, the thicknesses of the coatings were low.Despite that, all the coatings passed the solvent resistance double-rub test with acetone without showing damage.To further examine the solvent resistance of these films, ethanol was applied during the doublerub test.Ethanol is a better solvent for amorphous polyamides than acetone.The use of ethanol revealed large differences between the coatings.Except for B1 cured with TGIC, all coatings failed well before reaching 100 double rubs.B1 cured with TGIC did lose its gloss during the rubbing, probably due to swelling of the coating, but it remained undamaged and regained gloss upon evaporation of the ethanol.The failure of the other formulations indicates that network formation was insufficient.For the combinations of B1 and B2 cured with Primid this was expected as the curing chemistry should yield unreacted chains in the final coatings.These results are in agreement with the observations from the rheological and DSC experiments described in the previous two sections.These experiments showed insufficient cross-linking of the resins, especially when Primid was used as the curing agent.The network formation can be improved by introducing monomers with higher functionality into the polymers.This will give a higher cross-link density in the coating.
607 Pencil hardness values of the coatings were between 608 F and 4H.This is similar to the literature, and the value 609 for B1 with TGIC is even higher than common values. 1 610 The pencil hardness test could not be performed on the 611 coatings made from B2 as the highly irregular surface is 612 unsuitable.The reverse impact test, in which a 1 kg 613 weight is dropped from 1.00 m on the back of the 614 panels, was passed for all compositions.The thinness of 615 the coatings may have helped prevent cracks upon 616 impact, although even the thicker areas, especially in 617 B2-Primid, were free of fractures.
618 Conclusions 619 In this paper, the synthesis, characterization, and 620 testing of polyamides based on pimelic acid, butane-621 1,4-diamine, and isoidide diamine have been described.622 These resins have been developed for powder coating 623 applications.The M n values of these polyamides were 624 between 3000 and 3800 g mol À1 and they had both 625 amines as carboxylic acid end groups.The total values 626 of the AV and AmV combined were 0.54 to 0.64 mmol 627 g À1 .The T g values exceeded 60°C and the crystallinity 628 was low with DH m £ 20 J g À1 in the first heating run.629 Controlled heating and curing during the DSC exper-630 iment produced amorphous polyamides.631 The curing of the polyamide resins with TGIC or 632 Primid XL-552 (see Fig. 2) was investigated using a 633 rheometer and subsequent DSC analysis of the same 634 samples.The data indicated that a curing temperature 635 of 150°C is insufficient to produce a completely cross-636 linked product.Furthermore, TGIC performs signifi-637 cantly better than Primid for both carboxylic acid and 638 amine-terminated resins.After the experiment, the gel 639 content of systems with Primid was less than 25%.640 However, for TGIC it was over 66% (up to 97%).The 641 results of DSC analysis of the cured disks demon-642 strated that additional curing occurred at higher 643 temperatures.Therefore, a higher temperature than 644 150°C is necessary for curing the resins.Furthermore, 645 all the cured samples were amorphous.646 Curing at 180°C on standard aluminum Q-panels 647 was done with a solution of resin and curing agent in 648 DMAc.Dewetting of the coatings on the aluminum 649 panels was observed.Network formation was shown to 650 be inadequate as ethanol double rubs damaged most of 651 the coatings.Only B1 + TGIC showed excellent resis-652 tance against ethanol.Pencil hardness values were 653 between F and 4H which is similar to literature values.654 The reverse impact tests showed the coatings to be 655 flexible as no fractures were observed.A remark has to 656 be made that the thickness of all the coating layers was 657 rather low, i.e., around 10 lm.

658
In conclusion, the developed resins can be used with 659 standard cross-linkers of which the epoxy-based cross-660 linker showed better results than the b-hydroxyalky-661 lamide-based curing agent.To obtain well-performing 662 coatings, a higher cross-link density of the network is necessary and therefore the functionality of the resins needs to be increased, for example, by introducing branching in the resins with tri-or tetrafunctional monomers.Furthermore, the wetting of the substrates is poor.Therefore, either the resins have to be modified to reduce their surface energy, or additives have to be added to reduce the surface tension of the paint.

Fig. 3 :
Fig.3: DSC thermograms as functions of time for samples experiments A + TGIC (black), A + Primid (red), B1 + TGIC (blue), and B1 + Primid (green) after the rheology experiment.The temperature profile is depicted in orange and as top Xaxis to help the reader.The exotherm is directed upwards

performed with a TA Instruments 146 AR-G2 using plate-plate geometry, angular frequency 147 x = 6.283 rad s À1 , strain c = 1%. Samples were pre- 148 pared by mixing 250 mg of resin and cross-linker in a 149 mortar, followed by cold pressing of the mixture to a 150 disk (D = 25 mm, H = 500 lm) at a pressure of
2Scheme 1: Synthesis of polyamide resins based on pimelic acid (PA), butane-1,4-diamine (BDA), and isoidide diamine (IIDA)

Table 1 :
SEC, 1 H NMR, and DSC results for polyamide resins A, B1, and B2 a Number-average molecular weight and Ð measured by SEC in HFI1P against PMMA standards b 1 H NMR performed in DMSO-d 6 c DSC recorded at 10°C min À1 , data extracted from the second heating run d PA/IIDA/BDA Journal : 11998 Dispatch : 21-1-2016 Pages : 10 Article No. : 9783 h LE h TYPESET MS Code : 9783 h CP h DISK 4 4

Table 2 :
Properties of rheology samples: stoichiometric balance r, critical point of gelation p c , and gel content after the experiment Journal : 11998 Dispatch : 21-1-2016 Pages : 10 Article No. : 9783 h LE h TYPESET MS Code : 9783 h CP h DISK 4 4

Table 3 :
Ratio between reactive groups r, gel point p c , and test results of the coatings from A, B1, and B2 Journal : 11998 Dispatch : 21-1-2016 Pages : 10 Article No. : 9783 h LE h TYPESET MS Code : 9783 h CP h DISK 4 4