fire retardant mechanism in intumescent ethylene vinyl acetate compositions
TRANSCRIPT
Fire retardant mechanism in intumescent ethylene vinylacetate compositions
A. Rivaa, G. Caminoa,b,*, L. Fomperiec, P. Amigouetc
aCentro di Cultura per l’Ingegneria delle Materie Plastiche, V. T. Michel 5, 15100 Alessandria, ItalybPolitecnico di Torino, Sede di Alessandria, V. T. Michel 5, 15100 Alessandria, Italy
cNEXANS-NRC, 170, Avenue Jean Jaures, 69353 Lyon Cedex 7, France
Received 11 November 2002; accepted 6 January 2003
Abstract
The thermal and combustion behaviour of an intumescent fire retardant system based on Polyamide 6 (PA6) and AmmoniumPolyphosphate (APP), used to improve flame retardant properties of poly(ethylene-co-vinyl acetate) (EVA), loaded with Mg(OH)2(MH) was examined. The study of the interactions between the additives introduced in EVA was focused in particular on the MH-
APP interaction. The evolution of water from MH takes place at about 400 �C, with a fair overlap with ammonia and water evo-lution from APP degradation. Ammonia evolution from APP is facilitated by the presence of MH, in their mixture heated alone orin the polymer matrix. UL94 test shows that the interaction between MH and APP modifies the combustion behaviour of the
intumescent mixture.# 2003 Elsevier Ltd. All rights reserved.
Keywords: EVA; Intumescence; Ammonium polyphosphate; Polyamide 6; Magnesium hydroxide; Flame retardancy
1. Introduction
Ethylene polymers and co-polymers are widely usedin many fields, particularly in electrical engineeringapplications. Due to their chemical compositions, thesepolymers are easily flammable, and because of this,flame retardancy of these materials is widely studied.The main approach used up to now to impart flameretardant properties to this class of polymeric materialshas been the incorporation of additives, specifically ofhalogen compounds. The combustion products comingfrom these materials have a number of negative char-acteristics (corrosiveness, toxicity. . .) that pushed theindustry and the legislation to improve some newapproaches to flame retardance [1–3].One of these developing approaches is that of intumes-
cence. The intumescence mechanism consists in creatingon the polymer surface an expanded shield, able to reduceboth the heat flux from the flame to the polymer matrix,responsible for the fuel production, and the transfer offuel to the flame, limiting the spread of fire [4,5].
Generally intumescent formulations consist of threeingredients: an acid source (phosphates, borates etc.),a carbonising compound (polyols, polyamides, poly-urethanes etc.), and a blowing agent (melamine andmelamine compounds etc.). On heating, the acidsource gives out a mineral acid, that takes part in thedehydration of the carbonising compound, that forms acellular structure when the blowing agent decomposes[4–7]. The association of PA6 or other char formingpolymers and APP as flame retardants for EVA andother thermoplastic polymers has already been reported[8–11]. In this work we have studied the effect of com-bination of the intumescent system APP-PA6 with MHwhich is a widely used fire retardant in electrical cablesheeting materials.
2. Experimental
2.1. Materials
The following products were used: ethylene–vinyl ace-tate 24% copolymer (Elvax 265, DuPont, EVA), Poly-amide 6 (UltramidB4 BASF, PA6), magnesium hydroxide
0141-3910/03/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0141-3910(03)00191-5
Polymer Degradation and Stability 82 (2003) 341–346
www.elsevier.com/locate/polydegstab
* Corresponding author. Fax: +39-0131-229-331.
E-mail address: [email protected] (G. Camino).
(Magnifin H10, Martinswerk, MH) and ammoniumpolyphosphate (Exolit AP 422, Clariant, APP).The compositions were prepared via a two step pro-
cess to avoid APP degradation: in the first step EVA,PA6 and MH were mixed at a temperature of 235 �Cusing a Brabender PLE Mixer, with roller blades, with arotation speed of 60 rpm, for 5 min. In the second step,APP was added to the mixture and mixed at a tem-perature of 180 �C at 60 rpm for 5 min.Samples for all the performed tests were prepared by
pressing the material with an ATSFAAR hydraulic pressat a pressure of 200 bar and a temperature of 230 �C.
2.2. TGA-FTIR analyses
The TGA-FTIR analyses were performed using aPerkin Elmer Pyris 1 TGA, coupled with a Perkin ElmerSpectrum GX Infrared Spectrometer equipped with anIR gas cell. The TGA and the FTIR spectrometer werecoupled by a Perkin Elmer TG-IR Interface.The transfer line was heated to 220 �C, while the IR
gas cell was heated to 230 �C to avoid condensation ofdegradation products inside the gas cell. The samplessize was between 25 and 30 mg. The samples wereheated from 50 to 600 �C with an heating rate of 10 �C/min under a nitrogen flow of 30 ml/min. The nitrogenflow was switched on 10 min before the beginning of theanalysis, keeping the furnace closed, to get a stable IRbackground. 1 IR spectrum/�C was collected to draw agood evolution profile.
2.3. UL-94 tests
The flammability behaviour of the intumescent mix-tures was investigated by the UL-94 test on 3 mm thickspecimens ignited from bottom in the vertical configur-ation. The best ranking is V-0 when burning time isshort and there is no dripping of flaming particles,whereas the worst corresponds to ‘‘not classified’’ whenthe sample burns for more than 30 seconds or up to theholding clamp at 125 mm from the ignition point (seeASTM D 3801/00 for detailed procedure). Themeasurement of the burning rate of the materials wasbased on the time of combustion of the bottom 8 cm ofthe specimen, and was measured to integrate evaluationof those specimens that did burn up to the holdingclamp and could not be classified. The higher is t8, thelower the burning rate and hence the fire retardantbehaviour is better for samples showing higher t8 values.
2.4. FTIR analyses
The FTIR analyses were performed with a PerkinElmer Spectrum GX FTIR spectrometer, equippedwith a Multiscope FTIR microscope and a Micro-ATRgermanium crystal.
3. Results and discussion
3.1. TGA-FTIR studies
TGA-FTIR analyses were carried out on all the mix-ture components to identify the IR signals that could beused to monitor the degradation of the different com-pounds on heating the samples under nitrogen. Wefound that for all the mixture components except thatfor MH a typical IR signal was present.For EVA, the acetic acid evolution at about 350 �C
could be monitored, by the C¼O stretching vibration ofthe carboxylic group (1797 cm�1), while for PA6,caprolactam evolution at about 470 �C was monitored,by the C¼O stretching vibration of the amidic group(1710 cm�1). Finally for APP NH3 evolution at about380 �C is detected by the absorbance at 966 cm�1, wherethe typical ammonia gas phase doublet is found.The analysis of all these evolving products in the dif-
ferent mixtures allowed us to investigate the mutualinfluences between the various components.As MH degradation takes place with evolution of
water alone, it was not possible to find an IR absor-bance useful for monitoring its degradation. The TGAcurve of MH heated to 600 �C showed two main weightloss steps with two weight loss rate maxima at about400 and 510 �C respectively. The first degradation stepis well known to be due to water release, that is respon-sible for the FR mechanism of inorganic hydroxides[12]. The second might be due to the thermal degrada-tion of an impurity, as analyses carried out on analyticalgrade MH did not show this degradation step.Fig. 1 reports the results of the TG-FTIR carried out
on pure APP at 10 �C/min under nitrogen flow. In theupper part of the figure the weight loss curve (TG) and itsderivative curve (DTG) are reported (solid and dashedline respectively), while the ammonia evolution profile(dotted line, Absorbance), that represents the intensity ofthe IR absorbance peak found at 966 cm�1 is reported inthe bottom part of Fig. 1. Ammonia evolution as a func-tion of temperature, in terms of IR absorption at 966cm�1, is shown by the dotted curve in Fig. 1, upper part.It can be seen that a fair overlap of the DTG and
NH3 evolution curves exists, meaning that ammoniarepresents a considerable part of the evolved gases fromAPP in the temperature range 300–450 �C. In the lowerpart of the fig. is reported the IR spectrum of the gasesevolving at 385 �C, where the maximum ammonia evo-lution is taking place. It can be noticed that after theammonia evolution peak the DTG curve doesn’tdecrease to zero, meaning that other products are stillevolving, while the NH3 signal is disappearing. This factcan be related to evolution of water, which is anotherdegradation product of APP in this temperature range,as reported in the literature [13,14], but could not bemonitored by means of FTIR spectroscopy as H2O
342 A. Riva et al. / Polymer Degradation and Stability 82 (2003) 341–346
doesn’t give a sharp IR signal. After water and ammo-nia release the degradation of the polyphosphate net-work takes place, beginning at about 550 �C [14], asshown by the DTG increase. The IR spectra of theproducts evolved during this phase were not collected toavoid damaging the IR gas cell.Ammonia evolution from APP is related to acidic site
formation involved in the intumescence phenomena, asalready reported in the literature [13]. As MH is a base,we were interested in analysing the interaction betweenthese two components, to investigate whether it couldlead to suppression of the FR behaviour of the intumes-cent mixture or not. Our first approach was to carry outTG-FTIR on mixtures of APP andMH, characterised bydifferent APP/MH ratios, comparing experimental curveswith those expected from the behaviour of APP and MHheated separately. For example, if no interaction waspresent between the two compounds, only one weight lossstep should take place, as shown by the dotted line inFig. 2 for a 50–50%mixture. Only one DTG peak shouldbe found at a temperature of about 385 �C, as both thewater evolution from MH and the ammonia and waterevolution from APP take place in the temperature range280–460 �C. The second DTG peak expected at 496 �C isrelated to the degradation of MH, while the DTG curveincrease above 500 �C is related to the phosphate net-work degradation.
When we ran the TG experiment on the mixtures ofMH and APP we found some considerable differenceswith the calculated curves. For example, for the 50–50%mixture the expected single peak was substituted bythree peaks in the range 250–500 �C, as shown in Fig. 2by the solid line. A further difference is represented bythe disappearance of the peak at 496 �C, related to MH,and of the polyphosphate network degradation(>500 �C), related to APP. These differences show thatan interaction between MH and APP occurs on heating.Only ammonia is shown by FTIR to be evolved onheating the mixture with an absorption profile, as afunction of temperature, revealing that only the firstweight loss step is related to NH3 release, as shown inFig. 3 by the dotted line (IR absorbance at 966 cm�1)compared to the solid line that represents the DTGcurve of the mixture. The peak of ammonia evolution inthe mixture (330 �C) appears to be earlier if comparedto that of pure APP in Fig. 1 (385 �C). The two follow-ing DTG peaks can be attributed to water evolution dueto the interaction between MH and the hydroxy groupsleft after ammonia release from APP (Scheme 1a and b),as no gas phase IR signal was associated with thesedegradation steps, showing the evolution of other pro-ducts. The reactions suggested in scheme 1 are related tomagnesium phosphate formation during the thermaldegradation of the MH/APP mixture, which is inagreement with the thermal stability of the residue at600 �C.A number of TGA-FTIR experiments was carried out
on mixtures containing APP, MH and the EVA-PA6
Fig. 1. TGA-FTIR of pure APP under nitrogen, heating rate 10 �C/
min. Upper part: solid line: TG curve; dotted line: NH3 evolution
(A.U.); dashed line: DTG curve. Lower part: FTIR spectrum of gases
evolving at 385 �C.
Fig. 2. Calculated (dotted line) and experimental (solid line) TGA
results for a 50% w/w physical mixture of APP and MH. TG (bold
lines) and DTG (thin lines) curves are reported.
A. Riva et al. / Polymer Degradation and Stability 82 (2003) 341–346 343
polymer matrix. The results are reported in Fig. 4,where a comparison between different mixtures char-acterized by different APP/MH ratios and differentoverall filler concentrations is reported. The compar-ison is focused on the ammonia evolution profile,recorded as absorbance at 966 cm�1. All the ammoniadetected on heating derived from APP, as NH3 evol-ving from PA6 was found not to be detectable inthese conditions, PA6 being about 7% of the polymermatrix. The absorbance curves here reported in Fig. 4were obtained from samples containing MH/APPratios 4, 1, 0.25 and 0, respectively dash–dot, dashed,dotted and solid line. The overall filler content forthese samples is 50, 60, 50 and 60% respectively.Looking at the figure it can be noticed that the pre-sence of MH causes an earlier of the ammonia evolu-tion that is not related to the MH percentage into thecomposition, as demonstrated by the peaks of the B,C and D curves, all found at the same temperature,about 30 �C lower than that of curve A. The tem-perature of the maximum ammonia evolution foundfor the mixtures containing both APP and MH dis-persed in the polymer matrix is the same found forthe mixtures of the additives alone. Whereas in theabsence of MH APP evolves ammonia at a tempera-ture 20 �C lower in the polymer matrix (comparisonbetween Figs. 1 and 4, with APP ammonia evolutionpeaks at 380 and 360 �C).
3.2. FTIR analyses
The formation of magnesium phosphate is in agree-ment with comparison of the IR spectra of MH, APPand the residue of their mixture after heating at 600 �Cunder nitrogen flow, reported in Fig. 5. The solid linerepresents the pure APP FITR–ATR spectrum, thedashed line is related to pure MH spectrum and thedotted line is obtained from the residue. It can benoticed that the –OH stretching vibration of MH at3688 cm�1 disappears, as does the broad band due toNH4
+ stretching vibration from 2600 to 3300 cm�1 andthe NH4
+ bending vibration at 1415 cm�1 because ofevolution of water and ammonia on heating. The shiftof P¼O, P–O and P–O–P phosphate vibrations at 1246,1062 and 1012 cm�1 should be due to the formation ofmagnesium phosphate.
3.3. Flammability behaviour
In table 1 the UL-94 test results divide the samplesinto two groups, one reaching the V-0 rating (shortburning time, no dripping), the other being ‘‘not classi-fied’’ (n.c.). Looking at table 1 it can be noticed that thebest fire retardant behaviour (V-0 rating) is obtainedwith the mixtures containing at least 30% of APP (seeNo. 2, 4) or 60% of MH (see No. 3) used alone. Addi-tion of MH to APP reduces its fire retardant activity as
Fig. 3. TGA-FTIR results for experiments carried out on 50% w/w
APP/MH physical mixture under nitrogen, heating rate 10 �C/min.
Upper part: solid line: TG curve; dotted line: NH3 evolution (A.U.);
dashed line: DTG curve. Lower part: FTIR spectrum of gases evolving
at 330 �C.
Fig. 4. Ammonia evolution profiles (recorded as IR absorbance at 966
cm�1) for different intumescent mixtures. Solid line: APP 60%,
EVA+PA6 40%; dashed line: APP 30%, EVA+PA6 40%, MH 30%;
dotted line: APP 40%, EVA+PA6 50%, MH 10%; dash–dot line:
APP 10%, EVA+PA6 50%, MH 40%.
344 A. Riva et al. / Polymer Degradation and Stability 82 (2003) 341–346
Scheme 1. Mechanism of degradation of APP in absence (a) [14] and in presence of MH (b).
Table 1
Intumescent mixtures. Composition and UL-94 performances. Polymer matrix: EVA (93%)+PA6 (7%)
Sample
Filler t1a sec t2b sec
t1+2c sec t8d sec
Flamingdripping
UL94
ranking
Total filler
Content (%)
APP/MH
ratio
1
30 Y n.c.f 02
60% APP 0 0 0 n.m.e N V-0 603
60% MH 0 2.2 2.2 n.m.e N V-0 604
30% APP 0 2.6 2.6 n.m.e N V-0 305
30% MH 36 Y n.c.f 306
30% APP 3 83 86 n.m.e N n.c.f 60 130% MH
7
20% APP 66 Y n.c.f 40 120% MH
8
10% APP 46 Y n.c.f 20 110% MH
9
40% APP 0 2.3 2.3 n.m.e N V-0 50 410% MH
10
10% APP 0 60 60 n.m.e Y n.c.f 50 1/440% MH
a t1=burning time after first ignition.b t2=burning time after second ignition.c t1+2=overall combustion time.d t8=time for 80 mm combustion if the specimen burns until the clamp.e n.m. (not measured)=t8 was not measured if the sample did self extinguish.f n.c.=not classified.
A. Riva et al. / Polymer Degradation and Stability 82 (2003) 341–346 345
shown by increase of combustion times in sample 6 ascompared to sample 4. An excess of APP is required ifused with MH as compared to samples in which is usedalone (see Nos. 4, 6, 9)By considering also the t8 values (related to the com-
bustion rate), it is possible to distinguish between samplesnot classified because they burned for more than 30 s butable to give self- extinguishing of the flame (only t1 and t2recorded), and samples not classified because they burnedto the clamp (only t8 recorded). The fire behaviour of themixtures appears to be better when the APP/MH ratio isdifferent from 1, and gets better with increasing the ratio.The comparison between samples Nos. 6, 7, 8, (APP/
MH=1) shows that the flame retardant behaviour getsbetter as the total filler amount rises (No. 6 self extin-guished, N�8 burned to the clamp faster than No. 7).Nos. 9 and 10 show that with the same overall amountof APP+MH (50%) when APP/MH is higher than 1(No. 9) the sample is V-0 classified, while if APP/MH islower than 1 (No. 10) we have a n.c. sample, thoughwith better behaviour than Nos. 6, 7, 8 (shorter selfextinguishing time). A reason for this behaviour couldbe found in the salification effect caused by MH on theacidic sites of APP, able to inhibit its flame retardanteffect, and on APP being the most efficient componentas shown by comparison of mixtures 2–3 and 4–5.
4. Conclusions
This study has shown that APP and MH interact onheating when introduced into an intumescent mixture
based upon EVA 24% VA, PA6 and the effects on fireretardant behaviour and thermal degradation of thepolymer matrix of their interaction.APP and MH react on heating forming salts at the
acidic sites left on APP after NH3 release, and this redu-ces the flame retardant effectiveness of the additives.The TGA-FTIR analyses carried out on the intumes-
cent mixtures showed that a facilitation of the forma-tion of the acidic sites on APP can be caused even by alow MH quantity. This sort of catalytic effect might beused to impart flame retardancy to materials withdecomposition temperatures lower than that of EVA,by shifting the action of the acid source to match thecharring and blowing effects of the intumescent additiveto the decomposition of the polymer matrix.
References
[1] Nelson GL. In: Grand AF, Wilkie CA, editors. Fire retardancy
of polymeric materials. New York: Marcel Dekker Inc; 2000.
p. 1–24.
[2] O’Neill TJ. Proceedings of the 10th international Flame retar-
dants 2002 Conference 2002;33–42.
[3] Troitzsch JH. Proceedings of the 10th international Flame retar-
dants 2002 Conference 2002;249-257.
[4] Lewin M. In: Le Bras M, Camino G, Boubigot S, Delobel R,
editors. Fire retardancy of polymers: the use of intumescence.
Cambridge: The Royal society of Chemistry; 1998. p. 3–32.
[5] Le Bras M, Bourbigot S. In: Le Bras M, Camino G, Boubigot S,
Delobel R, editors. Fire retardancy of polymers: the use of intu-
mescence. Cambridge: The Royal Society of Chemistry; 1998.
p. 64–75.
[6] Camino G, Luda MP. In: Le Bras M, Camino G, Boubigot S,
Delobel R, editors. Fire retardancy of polymers: the use of intu-
mescence. Cambridge: The Royal Society of Chemistry; 1998.
p. 48–63.
[7] Camino G, Delobel R. In: Grand AF, Wilkie CA, editors. Fire
retardancy of polymeric materials. New York: Marcel Dekker
Inc; 2000. p. 217–43.
[8] Le Bras M, Bourbigot S. In: Grand AF, Wilkie CA, editors. Fire
retardancy of polymeric materials. New York: Marcel Dekker
Inc; 2000. p. 136–49.
[9] Almeras X, Dabrowski F, Le Bras M, Poutch F, Bourbigot S,
Marosi G, Anna P. Pol Deg Stab 2002;77:305.
[10] Le Bras M, Bourbigot S, Revel B. J Mat Sci 1999;34:5777.
[11] Bugajny M, Le Bras M, Noel A, Bourbigot S. J Fire Sci 1999;
17:494.
[12] Horn WE. In: Grand AF, Wilkie CA, editors. Fire retardancy of
polymeric materials. New York: Marcel Dekker Inc; 2000. p. 285–
352.
[13] Levchik SV, Camino G, Costa L. Fire Mat 1995;19:1.
[14] Camino G, Luda MP. In: Le Bras M, Camino G, Bourbigot S,
Delobel R, editors. Fire retardancy of polymers: the use of intu-
mescence. Cambridge: The Royal Society of Chemistry; 1998.
p. 48–63.
Fig. 5. FTIR spectra of pure APP (solid line), pure MH (dashed line)
and the residue left after heating their 50% w/w physical mixture at
600 �C under nitrogen flux.
346 A. Riva et al. / Polymer Degradation and Stability 82 (2003) 341–346