Flammability of layered silicate epoxy Nanocomposites combined with low-melting inorganic Ceepree glass.
In the pursuit of eco-friendly and economical flame retardants for polymers, over the last decade and a half layered silicate polymer nanocomposites have been studied intensively in flame retardancy to improve fire proper- ties (1-6). With increasing nano-dispersion, layered silicate polymer nanocomposiies exhibit extraordinary reduction in the peak heat released rate (PHRR) (6-9), which is one of the most significant lire properties monitored by cone calorimeter testing (10), (11). Different flame retardancy mechanisms have been proposed in polymer nanocomposites (12-16). The main flame relardancy effect is caused by an inorganic-carbonaceous layer shielding against heat and mass transport during combustion (13), (17), (18).
According to our studies on different layered silicate epoxy (EP) nanocomposites, as indicated in Fig. 1, nonfilled EP burns almost completely, leaving little or no residue. The EP composite using unmodified layered silicate with microdispersion produced a fragmentary residue after burning, corresponding to a slight reduction in PHRR. The EP composites with the same amount of organic modified layered silicate with nano-dispersion exhibited a fire residue surface with a more integral structure. Compared with the different morphologies of fire residues, integral structures with fewer cracks and openings on the residue surface correspond to a greater reduction in PHRR, as shown in Fig. 1. A homogeneous residue with structural integrity is believed to be crucial for providing the most effective protective layer (12, 13). Nevertheless, a fully closed structure without fracture cracks on the surface of layer formation is hard to achieve in polymer nanocomposites.
[FIGURE 1 OMITTED]
Furthermore, the overall flame retardancy achieved by using layered silicate alone is generally not sufficient (9), (12), (16). The total heat evolved (THE), corresponding to fire load, ignitability (time to ignition), and flammability (reaction to a small flame: LOI, UL 94) usually shows only little or no improvement in nanocomposites. This result suggests that the layered silicate works mainly as inert filler during combustion. Thus, a combination with other (lame retardants is in high demand (3), (4), 19-30).
Among the halogen-free flame retardants available for polymers, low-melting inorganic glasses also have been promoted as a flame retardant/smoke suppressant for organic polymers. The (lame retardancy effect results from the glass forming a protective layer at a sufficiently low temperature during combustion (31-35). Such low-melting glasses for flame reiardancy have been reported for various common thermoplastic polymers such as PVC, PP. PA, PMMA, etc., but rarely for thermosets. In addition. LOI and UL 94 performance can be improved by incorporating low-melting glass in polymer materials. However, the relatively high loading required for significantly efficient flame retardancy is detrimental to other physical properties (36).
Therefore, layered silicate was combined with conventional low-melting inorganic glass in EP composite to optimize the protection layer. The aim is to provide a more integral layer, so that the inorganic glass may act as an agent to merge and strengthen the mixture of layered silicate and polymer/char during combustion. The combination of layered silicate and low melting glass may lead to synergistic effects. Such a combination is appropriate to the purpose of producing halogen-free flame retarded EP.
This study was designed to identify the actual effects on flame retardancy mechanisms of combining organic modified layered silicate with low melting inorganic glass through different methods. A comprehensive lire behavior characterization and the detailed features of the residue formation were right on target.
Materials and Sample Preparation
Ceepree (CP) is a commercial, near while powder manufactured by Chance and Hunt (UK) and was obtained from Nordmann, Rassmann (Germany), For this work, the grade "Ceepree B200U" was used, which has not received an organic coating. According to the information given by the manufacturer Chance & Hunt, it is a tailored blend of different unleaded glass frits with an average particle size D50 of 3-8 [micro]m and a maximum particle size of 30 [micro]m. The manufacturer claims that its lower melt temperature begins at 623 K (melt point of the soft frit component), and the setting action due to the devitrifying component at about 1123 K, so that CP remains hard at temperatures well over 1273 K. The suppliers point out that CP is not a (lame retardant with respect to preventing combustion or flame spread but instead produces glassy char that serves as a fire barrier. Our own elemental analysis proved that it was a mixture of different glasses, as a large variety of elements (O. Si, Ca, C, Na. W Al, F, P. S, Ti, K) were found. Proposing CP as fire barrier filler goes back to the late 80s (37-401).
The organically modified layered silicate was produced from a sodium montmorillonite (Nanofil 757, Sudchemie, Germany) with a cation exchange capacity of 0.8 meq [g.sup.-1] by reaction with teiraphenylphosphonium bromide (Evonik). A dispersion of 50 g Nanolil 757 (40 mmol exchangeable sodium ions) in 1 I cieionized water was prepared and heated to 342 K. A warm solution of 16.7 g (40 mmol) tetraphenylphenylphosphonium bromide in 600 ml deionized water was slowly added while the mix-line was vigorously stirred. Tetraphenylphosphonium montmorillonite (TPPMMT) instantly precipitated in the form of voluminous fiakes. The dispersion was diluted with 31 deionized water and stirred for I h at room temperature before the TPPMMT was Tillered off. By carefully washing with deionized water, the remaining sodium and bromide ions were removed from the product. The final filler cake, which had a solid content of about 10 wt%, was granulated and deep-frozen with liquid nitrogen. After freeze drying for I week at 0.16 mbar (alpha 1--4, Christ) TPPMMT was obtained as a light powder with a high specific surface area, typically of 150 m [m.sup.2] [g.sup.-1] as determined from BET isotherms.
The epoxy resin system is based on a bisphenol A diglycidyl ether (Araldite GY250, Huntsman) with an epoxy equivalent weight of 185 g e[q.sup.-1], which was cured with an equimolar amount of methyl hexahydrophthalic anhydride (MHHPA. Acros Organics). The curing reaction of the mixture of epoxy resin and anhydride hardener was catalyzed by 1 wt% of 1-methylimidazole (Sigma Aldrich).
The composition of the investigated composites is summarized in Table 1. Composite materials were prepared in a dissolver (CA40, Gelzmann) equipped with a 60 mm disc. Either CP glass or TPPMMT was dispersed for 1 h in the epoxy resin at room temperature and at a rotation speed of 1500 rpm. After mixing with the calculated amounts of hardener and catalyst, the dispersions were cast into preheated aluminum molds, which were treated with a release agent (Acmosan 82-6007, Acmos Chemie). Curing was achieved in 1 h at 393 K in a ventilated air oven. Plates with a size of 250 x 170 [mm.sup.2] were obtained, which were cut into the lest specimens for UL94, LOI and cone calorimetry.
TABLE 1. Compositions of the investigated materials. Abbreviations Epoxy rosin Layered silicate Inorganic glass EP GY250/MHHPA -- -- EP/5%TPPMMT GY250/MHHPA Tetraphenylphosphonium -- modified MMT (TPPMMT) EP/5%TPPM GY250/MHHPA Telraphenylphosphonium Silicate MT/10%CP modified MMT (TPPMMT) glass, Ceepree B200U EP/I0%CP GY250/MIIHPA -- Silicate glass, Ceepree B200U EP/I5%CP GY250/MHIIPA -- Silicate glass, Ceepree B200U
Preparation of the composite with 10 wt% CP and 5 wt% TPPMMT required a modification of the procedure described above. When both fillers were incorporated in the epoxy resin, the viscosity of the resulting mixture became too high for proper processing. Therefore, only the TPPMMT was dispersed in the epoxy resin. In a separate step, this was mixed with a dispersion of CP in MHHPA.
The morphology of EP/5%TPPMMT was investigated using X-ray, scanning electron microscope (SEM), and TEM and has been reported before (30). The morphology of EP/5%TPPMMT clearly differed from that of a common microcomposite, but also showed neither simple dispersion nor simple exfoliation, nor either of these phenomena in a perfect manner. It is a hierarchical mixture of a microcomposite and nanocomposite morphology. TPPMMT-rich EP/TPPMMT nanocomposite phases are observed within a TPPMMT-free EP matrix. Thus, the overall distribution is characterized by a blend demixed on the micrometer scale, with one phase being a nanocomposite characterized by intercalated and exfoliated structures. The interface between silicate plates and the polymer are determined by their delamination but not by their homogenous dispersion. Thus quite comparable with better-dispersed systems, this special kind of nanocomposite morphology shows an extremely large interphase between layered silicate sheets and EP.
To identify the effect of inorganic additive on the degree to which EP was cured, the samples were investigated by differential scanning calorimeter (DSC) (TA instruments DSC 2920) with a heating rate of 10 K min under nitrogen at a flow rate of 50 ml [min.sup.-1]. In DSC measurements, the weight of the granular sample was 5 [+ or -] 1 mg, with two measurements taken to ensure reproducibility. The pyrolysis of materials was investigated by thermogravimetry (TGA/SDTA 851, Mettler Toledo, Germany). Thermal stability and thermo oxidative stability were measured at a heating rate of 10 K [min.sup.-1] under nitrogen and air How (30 ml [min.sup.-1), respectively. The sample weight was about 10 [+ or -] 1 mg. For evolved gaseous product analysis, the thermogravimetry was coupled with Fourier transform infrared spectrometry (TG-FTIR, Nexus 470, Nicolet, Germany).
Fkmunahitity and Fire Tests
The flammabilily (reaction to small flame) was determined by limiting oxygen index (LOI) according to ISO 4589 and the UL 94 vertical and horizontal test according to IEC 60695-11-10. The sample size for LOI was 150 mm x 6.5 mm and 3.2 [+ or -] 0.1 mm thickness, for UL 94 125 mm X 13 mm and 3.2 [+ or -] 0.1 mm thickness.
The lire behavior (forced-naming condition) was investigated by a cone calorimeter (FTT, UK) as a bench scale fire test according to ISO 5660. Three different irradiations of 35, 50, and 70 kW m were applied. The samples with a size of 100 mm x 100 mm and thickness of 5.5 [+ or -] 0.2 mm were placed into a retainer frame sample holder. All measurements were performed in duplicate or triplicate. Instead of a visual determination by the operator during the test, flame-out was defined more accurately using smoke release data during data evaluation after the tests.
The lire residues obtained from cone calorimeter under an irradiation of 50 kW [m.sup.-2] were characterized by SEM (FEI XL30 ESEM, Eindhoven, the Netherlands). The fire residue samples with a representative structure were selected in the central area of residue formation and investigated with a sputtered golden coaling for conductivity.
RESULTS AND DISCUSSION
In general, the formation of polymeric networks can be affected by the additives used or by the thermal history of the material . DSC curves of all of the investigated materials are shown in Fig. 2, The first healing showed a postcure reaction at between 400 and 470 K for all materials with a peak at 426 K, which is the typical behavior for EP cured at a temperature of 393 K. The reaction enthalpy was 20.9 J [g.sup.-1] for the EP and 24.2 J [g.sup.-1] for EP/5%TPPMMT, indicating that curing was nearly complete (around 94%), as the enthalpy for complete curing is 360 J [g.sup.-1]. The small reaction peak was hardly influenced by the addition of TPPMMT. The same glass transition temperature was observed at about 415 K (Table 2). EPs containing CP exhibited successive endothermic events in the course of postcuring, with the exothermic reaction enthalpies reduced to 4-11 J [g.sup.-1], indicating a slightly increased degree of curing. A slight decrease of 10 K in the glass transition temperatures seemed to indicate the opposite effect but may also be influenced by the filler through other phenomena such as increased free volume. In any case, only minor negligible effects on the degree of curing and the glass transition temperature of EP occurred through the use of various inorganic additives. There is no deterioration of the network influencing decomposition or fire behavior.
[FIGURE 2 OMITTED]
TABLE 2. DSC and ihermouravimeiry data in N" of all the investigated materials. DSC [T.sub.g] [T.sub.5]wt%/K [T.sub.max]/K /K[+ or -] [+ or -] 5 [+ or -] 5 s EP 415 637 685 EP/5%TPPMMT 415 637 685 H!75%TPPMMT/I0%CP 405 639 685 EP/10%CP 405 637 689 EP/I5%CP 405 639 689 Thermogiavimetry Mass loss Residue (up to (at 720 K)/wt% 1000 K)/wt% [+ or -] 3 [+ or -] 3 EP 88 7 EP/5%TPPMMT 80 13 H!75%TPPMMT/I0%CP 76 IS EP/10%CP 79 16 EP/I5%CP 72 22
The pyrolyses of TPPMMT, CP, and TPPMMT/CP (1:1) were compared with each other (Fig. 3). CP exhibited relatively high thermal stability, in which very little mass was lost up to a temperature of 1000 K ([DELTA] = 8 wt% considered as moisture and some weakly bonded substances). TPPMMT exhibited a multistep-decomposition between 640 K and 850 K. A total of 22 wt% mass loss occurred due to the decomposition of organic moiety and dchydroxylation of layered silicate as proposed in the literature (42-44). TPPMMT/CP displayed no dramatic change in the main decomposition of TPPMMT, except that the peak mass loss rate of dehydroxylation of aluminosilicate was somewhat reduced at high temperatures. However, these results implied that chemical interaction between the components of CP and TPPMMT played a minor role.
[FIGURE 3 OMITTED]
The thermal stability of the investigated EP composites in [N.sub.2] is shown in Fig. 4. AH of the materials investigated showed similar decomposition behavior, strongly related to that of EP. AH materials showed one main decomposition step between 500 and 800 K with a peak temperature of 685 K, resulting in a main mass loss of 72-80 wt% for the composites and 88 wt% for EP (Table 2). The residues remaining at high temperatures up to 1000 K were increased when various inorganic additives were added. The deviation in mass loss during the decomposition step ([DELTA] = 8-16 wt%) and the additional residues at high temperatures ([DELTA] = 6-15 wt% at 1150 K) corresponded to the amount of inorganic substances added.
[FIGURE 4 OMITTED]
Pyrolysis gas analysis showed the same decomposition products for all of the investigated materials. The main volatile products were C[O.sub.2] (2354 [cm.sup.-1]), CO (2174 [cm.sup.-1]), [H.sub.2]O (3853 [cm.sup.-1]), methane (3016 [cm.sup.-1]), and some other organic substances containing carboxylic acid (1729 [cm.sup.-1]), anhydride (1806 [cm.sup.-1]), phenolic derivatives (3647 [cm.sup.-1]), etc., as has been previously reported (45). The representative spectra at the peak mass loss rate are shown in Fig. 5 for all the investigated materials. However, when the spectra at the same decomposition step were compared, there was no change in volatile products. Furthermore, the product release rates were similar to those for EP. Thus, chemical reactions played a minor role during decomposition. Either the TPPMMT or CP functioned as inert filler; thereby, they did not significantly change the decomposition behavior of EP.
[FIGURE 5 OMITTED]
The thermo-oxidative stability of the investigated materials in air is shown in Fig. 6 and Table 3. All of the materials exhibited two main decomposition steps in the aerobic condition, which is the typical thermo-oxidative decomposition behavior of EP (46). The first main decomposition step under air exhibited similarity to the main thermal decomposition step for all samples. In particular, the same temperature range of 600-750 K was observed, with a peak temperature of 688 K. The mass loss was 67-74 wt% for the composites and 78 wt% for EP and thus only slightly less than the mass loss reported for the thermal decomposition step.
[FIGURE 6 OMITTED]
TABLE 3. Thermogravimctry data in air of all investigated materials. [T.sub.5]wt%/K lst Mass 2nd [+ or -] 5 [T.sub.max]/K loss [T.sub.max]/K up to [+ or -]5 720 K)/% [+ or -] 3 EP 635 688 78 892 EP/5%TPPMMT 631 688 74 883 EP/5%TPPMMT/10%CP 632 688 71 800 EP/10%CP 630 688 71 800 EP/15%CP 635 690 67 809 Residue (at 1150 K)/% [+ or -] 3 EP 2 EP/5%TPPMMT 5 EP/5%TPPMMT/10%CP 15 EP/10%CP 11 EP/15%CP 18
Surprisingly, oxidations of the transitory charring networks were facilitated when CP was used at high temperatures in the range of 750-950 K, as the peak temperature during the second main mass loss step occurred about 90 K earlier than for the non-filled EP and the EP/5%TPPMMT. Thus, the thermo-oxidative stability of the char deteriorated in the presence of CP at high temperatures. The oxidative two-step decomposition led to greater material consumption than anaerobic decomposition. The additional residues at high temperatures ([DELTA] = 3-16 wt% at 1150 K) remaining for the composites after complete oxidation were in good agreement with the amount of additives in the EP composites.
Flammability (Reaction to Small Flame)
The values for flammability in terms of reaction lo small flame, characterized by LOI and UL 94, are listed in Table 4. Using either TPPMMT or CP in the EP composites increased the LOI by 3-4% to 24-25%. This improvement through the use of small amounts of inert filler was caused mainly by a physical residual protection layer mechanism. However, an effect clearly less than expected occurred for EP/5%TPPMMT/10%CP (15 wt% filler), as the LOI was limited to 25%, which is the same as for EP/5%TPPMMT or EP/10%CP (5 or 10 wt% filler).
TABLE 4. Flummability (LOI and UL 94) and ignitability (cone calorimeter). Cone calorimeter 35 kW 50 kW [m.sup.-2] [m.sup.-2] LOI/% UL [t.sub.ig]/s [t.sub.fo]/s [t.sub.ig]/s [+ 94 [+ or -] 5 [+ or -] 41 [+ or -] 3 or -] 1 EP 21 HI 100 547 47 EP/5%TPPMMT 25 HB 110 617 53 EP/5%TPPMMT/10%CP 25 HB 101 771 48 EP/10%CP 25 HB 101 800 44 EP/15%CP 24 HB 89 900 46 70 kW [m.sup.-2] [t.sub.fo]/s [t.sub.ig]/s [t.sub.fo]/s [+ or -] 25 [+ or -] 2 [+ or -] 18 EP 399 22 352 EP/5%TPPMMT 537 25 406 EP/5%TPPMMT/10%CP 573 22 415 EP/10%CP 619 20 463 EP/15%CP 685 19 413 [t.sub.ig]: time to ignition; [t.sub.fo] time to flame-out.
The residual protective layer that formed was not sufficient lo achieve self-extinguishing behavior in UL 94 tests. All of the investigated materials were classified as only HB in the UL 94. After burning for a while, melt flow and slight dripping behavior occurred for EP, indicating that liquid decomposition products of the thermoset accumulated to a melt. Melt flow and dripping were prevented when the various inorganic additives were applied, as has been reported for comparable systems before (9), (16), (47), (48).
The ignitability of all of the materials was determined using the time to ignition ([t.sub.ig]), obtained from cone calorimeter tests for constant irradiations and spark ignition (Table 4). The [t.sub.ig] of the different materials was around 89-110 s for an irradiation of 35 kW [m.sup.-2] 46-53 s for an irradiation of 50 kW [m.sup.-2] and 19-22 s for an irradiation of 70 kW [m.sup.-2]. In general, higher irradiation levels give better reproducibility (11), (49). The [t.sub.ig] and [t.sub.fo] of all male-rials were shortened, and the burning time decreased with increasing irradiation. All materials displayed similar results for the [t.sub.ig]; CP worsened the [t.sub.ig] a little bit, whereas EP/5%TPPMMT showed a slight delay. This improvement in [t.sub.ig] is superior to most other layered silicate nano-composites, particularly those using ammonium-based modifications, in which the earlier decomposition of the modifier triggers the mass release. The impact of CP and TPPMMT appeared to counterbalance each other in EP/5%TPPMMT/10%CP.
Fire Behavior (Forced Flaming Conditions)
The results of cone calorimeter investigations in various irradiations are summarized in Tables 5 and 6.
TABLE 5, Cone calorimeter results under various irradiations of 35, 50, 70 kW [m.sup.2]. Irradiation Residue PHRR t-PHRR kW % kW s [m.sup.-2] [+ or [m.sup.-2] [+ or -] [+ or -] -] 2 20 10 EP 35 4 733 248 50 4 891 182 70 1 1196 134 EP/5%TPPMMT 35 10 482 306 50 7 571 228 70 6 694 195 EP/5%TPPMMT/10%CP 35 19 353 350 50 16 474 272 70 16 617 224 EP/10%CP 35 15 315 120 50 14 408 308 70 12 565 252 EP/15%CP 35 20 26S 371 50 16 346 344 70 17 585 269 FIGRA kW [m.sup.-2] [s.sup.-1] EP 3.0 4.9 8.9 EP/5%TPPMMT 1.6 2.5 3.6 EP/5%TPPMMT/10%CP 1.0 1.8 2.8 EP/10%CP 2.6 1.3 2.2 EP/15%CP 0.7 1 2 FIGRA: PHRR/time to PHRR. TABLE 6. Total heal evolved (THE), the effective smoke, and CO yield obtained from cone calorimeter under various irradiations of 35, 50. 70 kW nr. Irradiation THE THE/ML TSR/ML kW MJ MJ [g.sup.-1] [m.sup.-2] (m.sup.-2) [m.sup.-2] [+ or -] 4 [g.sup.-1] EP 35 141 2.5 82 50 151 2.5 82 70 147 2.4 85 EP/5%TPPMMT 35 140 2.4 85 50 138 2.4 89 70 140 2.3 94 EP/5%TPPMMT/10%CP 35 131 2.4 83 50 130 2.3 88 70 130 2.3 89 ep/10%CP 35 139 2.5 76 50 136 2.4 89 70 137 2.3 93 ep/15%CP 35 132 2.4 76 50 134 2.3 S3 70 129 2.3 92 CO-yield g (g.sup.-1) EP 0.046 0.044 0.047 EP/5%TPPMMT 0.044 0.040 0.041 EP/5%TPPMMT/10%CP 0.049 0.049 0.049 ep/10%CP 0.040 0.012 0.044 ep/15%CP 0.039 0.041 0.046 ML; mass loss; TSR: total smoke release.
After ignition, all of the EP composites showed a smoothly burning flame rather than EP burning with a vigorous sputtering flame. Figure 7a presents characteristic heat release rate (HRR) curves for all of the materials at an irradiation of 50 kW [m.sup.2], which simulates the developing fire scenario (11), (49). Figure 7b shows the corresponding fire residues. After a very similar initial increase in HRR up to a shoulder for all the materials, EP reached a rather sharp peak, which is the typical behavior of non-charring material--also indicated by the almost complete consumption of BP. All of the EPs containing the different additives displayed more plateau-like curves before approaching the second PHRR at the end of burning. None of the HRR curves showed a clear decrease in the HRR after the PHRR to zero, but they all exhibited a subsequent extension in heat release with prolonged burning before flame-out. It turned out that material below the retainer frame was protected, and most of it consumed only after the PHRR. This phenomenon results in uncertainties for the determination of when flame-out occurred. What is more, this afterburning of the edges shifts the overall fire scenario to small surface flames accompanied by higher surface temperatures and probably thermo-oxidative pyrolysis, both of which reduce the residue.
[FIGURE 7 OMITTED]
The dominant PHRR at the end of burning was significantly decreased by using the different inorganic additives. Under various irradiations, the reduction in PHRR was around 32-42% when TPPMMT was used, around 57-53% when 10%CP was added, and even about 63-51% for the higher loading of 15%CP (Fig. 8a). Similar to other EP nanocomposites studied before (29), (30), (50), the efficiency of flame retardancy tended to increase with increasing irradiation when TPPMMT was used. In contrast, a better performance occurred at lower external heat flux when CP was added.
[FIGURE 8 OMITTED]
These remarkable reductions in the PHRRs did not correspond to the amount of residue (Fig. 8b). They were attributed to the protection layer and shielding effects of the more integral inorganic-organic residual layer (Fig. 7b). The observed integral characteristics explained well the difference between all of the composites and the non-residue forming EP, as well as between EP/10%CP and EP/5%TPPMMT, and between EP/10%CP and EP/5%TPPMMT/10%CP. In contrast, the macroscopic appearance of the lire residue of EP/5%TPPMMT/10%CP showed larger cracks than EP/5%TPPMMT but a greater reduction in PHRR.
A flame retardancy less than expected for a superposition was observed for EP/5%TPPMMT/10%CP, with only 47-52% reduction in the PHRR for all of the irradiations applied. The flame retardancy was better than that achieved by EP/5%TPPMMT, but clearly less than for EP/10%CP and less than expected if the flame retardancy effects of TPPMMT and CP were superimposed. The fire growth rate (FIGRA = PHRR/time to PHRR) displayed the order: EP > EP/5%TPPMMT > EP/5%TPPMMT/10%CP > EP/10%CP > EP/15%CP.
The declined flame retardancy effect in EP/5%TPPMMT/10%CP corresponded to some extent with the visual morphologies of the lire residues. The integral structure of EP/CP with a fully closed surface is not obtained in EP/5%TPPMMT/10%CP. Consequently, EP/5%TPPMMT/10%CP performed worse. However, as discussed before, the relation between the visual appearance of the residue and the reduction in PHRR did not hold for EP/5%TPPMMT compared with EP/5%TPPMMT/ 10%CP. This contradiction raised the demand for addressing (ire residue morphology more intensively to clarify the structure-property relationships governing flame retardancy.
Figures 9 and 10 show representative SEM images of the different lire residues for the cross section and for the top view, respectively. EP/CP, which exhibited the most effective (lame retardancy of all of the materials, showed a porous but interconnected, multilayer structure parallel to the surface (in Fig. 9b), and thus perpendicular to the release direction of the volatile pyrolysis products. The pores were considered as pathways for the release of pyrolysis products during combustion. In principle, a glass suitable for micro-structuring must be as pure and homogeneous as possible (51). Owing to the release of decomposition products from EP, the molten CP did not form a perfectly closed film on the micrometer scale. However, the fire residues of EP/10%CP and EP/15%CP consisted of a glassy and fully closed surface on a macroscopic scale, with a porous multilayer structure on the micrometer scale, yielding an impressive protection mechanism for flame retardancy.
[FIGURE 9 OMITTED]
[FIGURE 10 OMITTED]
The fire residues of EP/5%TPPMMT and EP/ 5%TPPMMT/10%CP showed extensive agglomeration of layered silicate during burning, enhanced by ablation and probably the release of pyrolysis products. In Fig. 9a and c, the lire residues displayed a columnar structure. Taking into account the large tendency of the silicate layers to form stacks, and the high silicate content of the fire residues, as well as the great reduction in volume through polymer ablation accompanied by the need to minimize surface energies, columnar structures are believed to be a probably class of morphology for the residue material. The columns were perpendicular to the surface, forming wide, defined channels for the release of volatile pyrolysis products. Thus, not only on the macroscopic visual level but also on the micrometer range, a lack of integral structure was observed for materials containing TPPMMT. The dominant columnar feature in residue formation turned out to be a key factor contributing to the deterioration in EP/5%TPPMMT/10CP with respect to flame retardancy.
Comparing EP/5%TPPMMT and EP/5%TPPMMT/10%CP, particularly using the lop view on the micrometer scale (Fig. 10), explained why the EP/5%TPPMMT/10%CP is characterized by better performance. In contrast to the visual macroscopic view, EP/5%TPPMMT consisted only of detached craggy islands in a loose texture, whereas the EP/5%TPPMMT/10%CP displayed a cauliflower-like morphology at the top of the residue. However, the cross section is dominated by columns, the lop layer is clearly a combination of layered silicate and CP features. The islands seemed to be coaled and merged together into larger units. In contrast to its appearance on the visual, macroscopic scale, the resulting micrometer-scale structure was more integral, explaining the improved flame retardancy effect.
As mentioned above, apart from the formation of a protective layer, both TPPMMT and CP functioned as inert fillers within the EP matrices. Even though the flame retardancy worked mainly in the condensed phase, there is hardly any relevant additional carbonaceous charring. Concurrently the additional lire residues under the same irradiations corresponded to the presence of additives remaining (for instance [DELTA] = 3-12 wt% at external heal flux of 50 kW [m.sup.-2] shown in Fig. 8b, which was in good agreement with the pyrolysis residues under different conditions. The THE/mass loss (THE/ML), attributed to the combustion efficiency multiplied by the effective heat of combustion, hardly changed through the use of the various inorganic additives (Table 6). No flame inhibition or fuel dilution occurred. Thus for all of the investigated materials the THE corresponding to the fire load did not show any reduction due to carbonaceous charring or flame inhibition, but only due to the limited replacement of EP by inert filler.
Other (ire hazard assessments, such as the smoke and CO release rates, occurred in a strong correlation with the heat or mass release curve. Thus, the peak of smoke and CO release rates were also reduced during combustion by adding the different inorganic additives to the EP compo-sites. Furthermore, the smoke and CO yield (Table 6), determined by total smoke and CO produced divided by total mass loss, respectively, were hardly changed by the presence of inert additives. This result also supports the conclusion that no flame inhibition or fuel dilution occurred in the gas phase as a result of the different inorganic additives being added.
The pyrolysis and fire behavior of EP composites using organic modified layered silicate TPPMMT, low-melting inorganic glass based on silicate glass of CP, and the combination thereof, respectively, were investigated in comparison with EP.
Insignificant influence was observed on the degree of curing and glass transition temperature for EP through the use of different inorganic additives. The thermal and thermo-oxidative stability of the different EP composites were hardly changed, as both the layered silicate and CP functioned as inert filler. The only exception was reduced oxidative char stability at high temperatures for EP composites containing CP.
The flammability results revealed improvement through the use of the different inorganic additives. LOI increased from 21 to 25% through the addition of either TPPMMT or CP, respectively. However, the LOI was limited to 25% even for the higher-loaded material when the combined system was added. Furthermore, the efficiency of the protection layer was not sufficient to achieve self-extinguishing behavior, as all the materials were classified HB in UL 94 test. A slight improvement in ignitability under forced ignition was found when TPPMMT was used, but counterbalanced by adding CP.
The fire behavior (cone calorimeter) tests revealed that the PHRRs were clearly reduced through the use of the different inorganic additives. The more the fire residue exhibited a fully closed integral structure, the greater the flame retardancy efficiency. A pronounced deteriorating effect on flame retardancy occurred when the combination of TPPMMT and CP was added, in correspondence with the deteriorated fire residue structure. The best results were obtained in EP/10%CP and EP/15%CP, respectively, showing a glassy integral structure on the visual macroscopic length scale and a porous multilayer structure parallel to the surface on the micrometer length scale.
The efficiency of flame retardancy was indeed strongly influenced by the morphology of fire residue on the different length scales. On the micrometer scale, TPPMMTs yield a dominant columnar structure, which is attributed to the accumulation of layered, silicate-rich residue when ablation and volatile product release occur. Cracks and openings on the visual macroscopic length scale, and the columnar structure perpendicular to the surface on the micrometer length scale, thus define release channels. The flame retardancy effect obtained for the samples containing TPPMMT was inferior to that of the CP systems. Deterioration was observed for EP/5%TPPMMT/10%CP. A more closed surface on the micrometer length scale, not the wider openings on the visual macroscopic length scale, explains the superior performance of EP/5%TPPMMT/10%CP compared with EP/5%TPPMMT.
Both TPPMMT and CP worked mainly in the condensed phase as inert fillers and through the formation of a physical protective surface. Changes in fire properties such as fire load, the effective heat of combustion, smoke, and CO production were rather insignificant, whereas the PHRR and FIGRA were reduced impressively.
Overall, the performance of combining TPPMMT and CP was below expectations. Deterioration dominated residue formation in EP/5%TPPMMT/10%CP; thus, the best flame retardancy was obtained by using EP/10%CP. However, the outermost top layer in EP/5%TPPMMT/10%CP on the microscale yielded improvement over EP/5%TPPMMT, thus showing that two additives can work together synergistically to form a protective layer.
The authors thank the working group of Prof. R. X. Fischer (University of Bremen, Germany) for its assistance. The authors are also grateful to Prof. H. Sturm for the SEM investigations and discussion. Dr. U. Braun for the fruitful discussions, and Mr. H. Bahr for assisting in the measurements.
(1.) A.B. Morgan and C.A. Wilkie, Flame Retardant Polymer Nanocomposites, Wiley, Hoboken (2007).
(2.) J.W. Oilman, T. Kashiwagi, and J.D. Liehtenham SAMPE J., 33,40 (1997).
(3.) ML Bodiger, T. Eckel, D. Wittmann, and H. Alberts. German Patent DE 195,30,200 AI (1997).
(4.) J.W. Oilman and T. Kashiwagi, "Polymer-layered Silicate Nanoeomposites with Conventional Flame Reuvrdants," in Polymer-Clay Nanoeomposites, Chapter 10, T.J, Pinnavaia, and G.W. BealL Eds., Wiley, Chichester, 193 (2000).
(5.) J.W. Gilman, Appl Clay Sci., 15, 31 (1999)
(6.) M. Zanetti, G. Camino, and R. Mulhaupt, Polym. Degrad. Stab., 74, 413 (2001).
(7.) J.W. Gilman, C.L. Jackson, A.B. Morgan, R. Hams, E. Manias, E.P. Giannelis, M. Wuthenow, D. Hilton, and S.H. Phillips, Chan. Mater., 12, 1866 (2000).
(8.) S. Duquesne, C. Jama, M. Lc Bras, R. Dclobel, P. Recourt, and J.M. Gloaguen, Compos. Set. Technol., 63, 1141 (2003).
(9.) M, Bartholmai and B. Schartel, Polym. Adv. Technol., 15, 355 (2004).
(10.) V. Babrauskas and R.D. Peacock, Fire Safely.J., 18, 255 (1992).
(11.) B. Schartel and T.R. Hull, Fire Mater., 31, 327 (2007).
(12.) J.W. Gilman, "Flame Retardanl Mechanism of Polymer-clay Nanocomposilcs," in Flame Retardanl Polymer Nanocomposites. Chapter 3, A.B. Morgan and C.A. Wilkic, Eds., Wiley, Hoboken, 67 (2007).
(13.) T. Kashiwagi, R.H. Harris Jr., X. Zhang, R.M. Briber. B.H. Cipriano, S.R. Raghavan, W.H. Awad, and J.R. Shields, Polymer, 45, 881 (2004).
(14.) J.W. Gillman. R.H. Harris Jr, J.R. Shields, T. Kashiwagi, and A.B. Morgan, Polym, Adv. Teclmol., 17, 263 (2006).
(15.) M. Zanetti, T. Kashiwagi, L. Falqui, and G. Camino, Chem. Mater., 14, 881 (2002).
(16.) B. Schartel, M. Bartholmai, and U. Knoll, Polym. Adv. Technol., 17. 772 (2006).
(17.) J.W. Gilman, T. Kashiwagi, E.P. Giannelis, E. Manias, S. Lomakin, J.D. Lichtenham, and P. Jones, "Nanocomposites: Radiative Gasification and Vinyl Polymer Flammability," in Fire Retardancy of Polymers: The Use of Intumescence, M. Le Bras, G. Camino, S. Bourbigot, and R. Delobel, Eds., Royal Society of" Chemistry, London, 203 (1998).
(18.) B. Schartel and A. Wei[beta], Fire Mater., 34, 217 (2010).
(19.) S. Bourbigot, S. Duquesne, and C. Jama, Macromol. Symp., 233, 180(2006).
(20.) A.B. Morgan. Polym. Adv. Technol., 17, 206 (2006).
(21.) G. Chigwada and C.A. Wilkie, Polym. Degrad. Stab., 81, 551 (2003).
(22.) G. Chigwada, P. Jash, D.D. Jiang, and C.A. Wilkic, Polym. Degrad. Stab., 89, 85 (2005).
(23.) N. Schall, T, Engelhardl, H. Simmler-Hubenthal, and G. Beyer, German Patent DE 199,21,472 Al (2000).
(24.) S. Bourbigot, M. Le Bras, F. Dabrowski, J.W. Gilman, and T. Kashiwagi, Fire Mater., 24, 201 (2000).
(25.) S. Bourbigot, M. Le Bras, S. Duquesne, and M. Rochery, Macromol. Mater Eng., 289, 499 (2004).
(26.) G. Beyer, Fire Mater., 25, 193 (2001).
(27.) G. Beyer, Firs Mater., 29, 61 (2005).
(28.) J. Zhang and C.A. Wilkie, Polym. Adv. Technol., 16, 549 (2005).
(29.) B. Schartel, U. Knoll, A. Hartwig, and D. Putz, Polym. Adv. Technol., 17, 281 (2006).
(30.) B. Schartel, A. Wei[beta], F. Mohr, M. Kleemeier, A. Hartwig, and U. Braun, J. Appl. Polym. Sci., 118, 1134 (2010).
(31.) S. Bourbigot and M. Le Bras, "Flame Retardant Plastics," in Plastics Flammability Handbooks, Chapter 5, 3rd ed., J. Troitzsch, Ed., Carl Hanser Verlag, Munich, 148 (2004).
(32.) W.J. Kroenke, J. Mater. Sci., 21, 1123 (1986).
(33.) R.E. Myers and E. Licursi, J. Fire Sci., 3: 415 (1985).
(34.) T. Kashiwagi, J.W. Oilman, K.M. Butler, R.H. Harris, J.R. Shields, and A. Asano, Fire Mater., 24, 277 (2000).
(35.) G. Camino, L. Costa, and G. Martinasso, Polym. Degrad. Stab., 23, 359 (1989).
(36.) C.J. Quinn and G.H. Beall, "Low-Temperature Glass-Polymer blends," in Chemical Processing of Advanced Materials, L.L. Hench and J.K. West, Eds., Wiley, New York, 799 (1992).
(37.) G. Crompton, European Patent EP 0,287,293 Bl (1993).
(38.) G. Crompton, U.K. Patent GB 2,234,754 A (1991).
(39.) G. Crompton, U.K. Patent GB 2,203,157 A (1998).
(40.) D. J. Parry and R.A. Carter, Cell. Polym., 9, 195 (1990).
(41.) M. Zammarano, "Thermoset Fire Retardant Nanocomposites," in Flame Retardant Polymer Nanocomposites, Chapter 9, A.B. Morgan and C.A. Wilkie, Eds., Wiley, Hoboken, 235 (2007).
(42.) W. Xie, Z.M. Gao, W.P. Pan, D. Hunter, A. Singh, and R. m&, Ohm. Mater, 13, 2979 (2001).
(43.) J.W. Gilman, W.H. Awad, R.D. Davis, J. Shields, R.H. Harris, C. Davis, A.B. Morgan, T.E. Sutto, J. Callahan, P.C. Trulove, and H.C. Delong, Chem. Mater., 14, 3776 (2002).
(44.) W. Xie, R.C. Xie, W.P. Pan, D. Hunter, B. Koene, L.S. Tan, and R. Vaia, Chem. Mater., 14, 4837 (2002).
(45.) S.V. Levchik and E.D. Weil, Polym. Int.; 53, 1901 (2004).
(46.) G. Camino, G. Tartaglione, A. Frache, C. Manferti, and G. Costa, Polym. Degrad. Stab., 90, 354 (2005).
(47.) B. Schartel, P. Potschke, U. Knoll, and M. Abdel-Goad, Eur. Polym. J., 41, 1061 (2005).
(48.) B. Schartel, U. Braun, U. Knoll, M. Bartholmai, H. Goering, D. Neubert, and P. Potschke, Polym. Eng. Sci., 48, 149 (2008).
(49.) V. Babrauskas, Fire Mater., 19, 243 (1995).
(50.) A. Hartwig, D. Putz, B. Schartel, M. Bartholmai, and M. Wendschuh Josties, Macromol. Chem. Phys., 204, 2247 (2003).
(51.) D. Hulsenherg, A. Harnisch, and A. Bismarck, Microstructuring of Glasses, Springer Series in Materials Science, Vol. 87, Springer, Berlin Heidelberg, 75 (2008).
Correspondence to: Bcmhard Schanel; e-mail: firstname.lastname@example.org
Contract gram sponsor: German Research Foundation (DFG); contract gram numbers: SCHA 730/8-1, 8-2, HA 2420/6-1, 6-2.
Published online in Wiley Online Library (wileyonlinelibrary.com).
[c] 2011 Society of Plastics Engineers
Guang Mei Wu, (1) Bernhard Schartel, (1) Malte Kleemeier, (2) Andreas Hartwig (2)
(1) BAM Federal Institute for Materials Research and Testing, Unter den Eichen 87, 12205 Berlin, Germany
(2) Fraunhofer Institute for Manufacturing Technology and Applied Materials Research, Wiener Str. 12, 28359 Bremen, Germany
|Printer friendly Cite/link Email Feedback|
|Author:||Wu, Guang Mei; Schartel, Bernhard; Kleemeier, Malte; Hartwig, Andreas|
|Publication:||Polymer Engineering and Science|
|Date:||Mar 1, 2012|
|Previous Article:||A Study on the fabrication of rounded patterns by spin coating of photoresist on silicon substrate with microstructures.|
|Next Article:||Ultrasound initiated maleic anhydride grafted onto a novel polypropylene copolymer.|