Polymer Degradation and Stability 88 (2005) 114e122 www.elsevier.com/locate/polydegstab
The sensitisation of thermal decomposition of ammonium polyphosphate by selected metal ions and their potential for improved cotton fabric flame retardancy Philip J. Davies, A. Richard Horrocks*, Andrew Alderson Centre for Materials Research and Innovation, Bolton Institute, Bolton, BL3 5AB, UK Received 15 September 2003; accepted 25 January 2004 Available online 11 November 2004
Abstract A thermal analytical study of a series of metal ioneammonium polyphosphate (APP) combinations has been undertaken to examine possible interactions that might enhance subsequent flame retardant activity. It is shown that certain metal ions, particularly Mn2C and Zn2C, promote thermal degradation of APP at lower temperatures than in their absence, and that this enables flame retardant activity to commence at lower temperatures in the polymer matrix thereby enhancing flame retardant efficiency. This has significance where flame retardant formulations are required to become active at temperatures well below normal substrate ignition temperatures and especially in the case of back-coated, flame retardant treatments for textiles. Initial experiments show that metal ion-doped APP in the presence of cellulose not only indicates a further sensitisation of cellulose decomposition but also improved flame retardance determined by limiting oxygen index when applied as a back-coating to cotton fabric demonstrates. Ó 2004 Elsevier Ltd. All rights reserved. Keywords: Ammonium polyphosphate; Metal salts; Manganese; Zinc; Cellulose; Flame retardant; Textiles; Limiting oxygen index
1. Introduction Certain polymeric applications require that flame retardants should start to function at temperatures low enough for ignition of surrounding polymer matrices to be prevented and yet are high enough to withstand processing temperatures. Applications of relevance to fibres and textiles include back-coating formulations involving phosphorus-based systems where diffusion of active retardant species from the reverse of the fabric to the front face must occur efficiently before front face fibres can ignite [1,2]. Normally, however, most phos-
* Corresponding author. Tel.: C44 1204 528851; fax: C44 1204 399074. E-mail address:
[email protected] (A.R. Horrocks). 0141-3910/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.polymdegradstab.2004.01.029
phorus-based flame retardants decompose at temperatures well above 300 C when ignition temperatures of cotton, for example, are close to 350 C [2]. Any means of reducing flame retardant decomposition temperatures has relevance to improving efficiency in such applications. Work has been published recently that has demonstrated that the presence of manganese dioxide with ammonium polyphosphate appeared to improve the performance of the flame retardant within polyamide 6 [3]. Following on from this, a more analytical investigation demonstrated that the presence of the manganese dioxide, at loadings of 25% and 12.5% with respect to ammonium polyphosphate, was having an effect on the thermal behaviour of the ammonium polyphosphate. It was evident an interaction between the manganese dioxide and the ammonium polyphosphate was occurring [4]. However, it was interesting to
115
P.J. Davies et al. / Polymer Degradation and Stability 88 (2005) 114e122
note that the decomposition mechanism of the mixture did depend on the percentage loading of the manganese dioxide. More recently, Lewin and Endo have published work [5,6], which has investigated the addition of small percentages of heavy metal salts to ammonium polyphosphate in combination with pentaerythritol (PER) in polypropylene. They observed that metal salt addition appears to increase the flame retardant performance of the ammonium polyphosphate as part of a flame retardant system with PER in a polypropylene matrix. The loading required was quite specific and an optimum point was observed at loadings of metal ion/salt of 1.5% w/w or less. Loadings above the optimum resulted in a drop off in the performance in the flame retardant as determined by LOI. It is therefore proposed that this potential enhancement of activity of ammonium polyphosphate may also be exploited within back-coating formulations. Within such systems the potential is generated for the application of an ammonium polyphosphate with a high degree of polymerisation, which also fulfils the durability requirements via a sensitised flame retardant activity. Lewin and Endo [5,6] have discussed mechanisms of APPePER-MnC combinations in terms of MnC-accelerated phosphorylation of PER and dOH groups formed on PP molecules following MnC oxidation in parallel with cross-linking. It was suggested that the improved performance may be due to partial crosslinking (effectively increasing the degree of polymerisation) and increased stability of the ammonium polyphosphate, thus reducing the volatility of the phosphorus oxides formed during pyrolysis and making more phosphorus available for phosphorylation and char formation. In the presence of a polypropylene matrix, this cross-linking will also increase the viscosity of the melt thereby diminishing the rate of flow to the flaming surface and improve the barrier effect of the char. However, at higher MnC levels, excessive APP cross-linking may occur with accompanying lower levels of PER phosphorylation and hence reduced char. This hypothesis explains why an optimum level of MnC exists for maximum flame retardant effect. To date no studies have been undertaken to examine whether the presence of small concentrations of heavy metal ions will destabilise APP and hence reduce its thermal decomposition and liquefaction temperature thereby increasing its general reactivity. This paper extends our previously reported study [7] of the effect on the thermal behaviour of ammonium polyphosphate of introducing a series of metal salts as a possible means of sensitising flame retardant activity and in improving understanding of the previously reported work of Lewin and Endo [5,6].
2. Experimental 2.1. Ammonium polyphosphateemetal salt combinations Ammonium polyphosphate (APP) (Antiblaze MCM, supplied by Rhodia Consumer Specialities) having an approximate degree of polymerisation (DP) of 100 was mixed with a series of metal salts. Table 1 lists the heavy metal salts chosen for this study based on the previously reported observations of Lewin and Endo [5,6]. Two sets of experiments were carried out, the first in which a range of simple metal salt-APP mixtures was examined and the second set in which APP crystalline particles were doped with those selected metal salts that appeared to produce greatest effect on thermal decomposition behaviour of APP from the former study. It was considered that surface doping of APP with metal ions would maximise any sensitising effect and so a means of surface-treating ammonium polyphosphate particles was established by ensuring that their dissolution in any metal solution was minimised. For the preliminary mixture study, a series of dry mixtures was made of ammonium polyphosphate (Antiblaze MCM) with each metal salt at 2% w/w concentration. Ferric sulphateeAPP mixtures were also prepared with salt concentrations varying from 1 to 5% w/w. For the preparation of surface-doped APP, it was considered that its low solubility would enable the soluble metal salts to be added as small quantities of solutions sufficient to wet APP crystallite surfaces. Addition of such small volumes would create a paste, which on drying, would have metal salt ions present only on crystallite surfaces. It was found experimentally that, for the purposes of mixing, a paste could be made from 5 g of APP with 2 g of each salt solution. Based on
Table 1 Metal salts used for mixing with or doping ammonium polyphosphate Metal salt
Chemical formula
Aluminium sulphate Copper(II) sulphate Iron(III) (ferric) chloride Iron(III) (ferric) nitrate Iron(II) (ferrous) sulphate Magnesium acetate Magnesium chloride Magnesium sulphate Manganese(II) acetate Manganese(II) sulphate Sodium acetate Sodium sulphate Sodium tungstate Zinc acetate Zinc chloride Zinc sulphate
Al2(SO4)3xH2O (xZ14e18) CuSO45H2O FeCl36H2O Fe(NO3)39H2O FeSO47H2O Mg(CH3COO)24H2O MgCl26H2O MgSO47H2O Mn(CH3COO)24H2O MnSO4H2O Na (CH3COO) Na2SO410H2O Na2WO42H2O Zn(CH3COO)22H2O ZnCl2 ZnSO47H2O
116
P.J. Davies et al. / Polymer Degradation and Stability 88 (2005) 114e122
A series of coating formulations were produced containing ammonium polyphosphate and 2% metal salt by weight of ammonium polyphosphate. Table 2 shows the general formulation used for coating a bleached, plain woven cotton fabric with area density 220 g mÿ2. The metal salt was added by first dissolving in the water prior to adding APP. The resin formulation comprised Vycar 460x46, a PVC-acrylic latex manufactured by Noveon Inc (formerly BF Goodrich) and S5000, a proprietary de-foamer supplied by Noveon Inc. One hundred percent cotton fabric was coated using the k-bar technique with a No. 4 k-bar giving approximately 30% dry formulation add-on by weight of fabric [1,2]. The fabrics were dried at 100 C and cured at 150 C both for 3 min. 2.3. Thermal analysis
2.4. Limiting oxygen index Limiting oxygen index (LOI) measurements were undertaken using a Stanton Redcroft FTA Oxygen Index instrument and 150!50 mm fabric samples were tested according to ASTM D2863-77 (revised 1990).
3. Results 3.1. Decomposition of APP Fig. 1 shows a typical TG/DTG trace for the decomposition of ammonium polyphosphate under nitrogen. Fig. 2 shows the corresponding DTA trace overlaid with the DTG trace. The decomposition of ammonium polyphosphate has been studied in depth and Camino et al. [8] have reported that it decomposes
APP
0.5 80
APP-1 Mn+
0.3
20
Wet fraction
Weight required (g)
Vycar 460x46 S5000 Antiblaze MCM Metal salt H2O Total
100 2 250 5 0 357
204.08 4.65 250 5 131.27 595
0.3430 0.0078 0.4202 0.0084 0.2206 1
85.75 1.95 105.05 2.1 55.15 250
0.4
60
Table 2 Coating formulation for ammonium polyphosphate/metal salt combinations comprising 2% w/w salt with respect to APP Wet units
APP
0.6
40
Dry units
APP-1 Mn
+
100
Thermal analysis for all samples was carried out using simultaneous thermogravimetric and differential thermal analysis using a TA Instruments SDT 2960. Samples (10G1 mg) were contained in platinum
Component
0.7
120
Derivative Weight (%/°C)
2.2. Coated fabrics
crucibles and analysed from ambient temperature to 1000 C under flowing nitrogen (100 ml/min) at a heating rate of 20 C/min for preliminary experiments using APPemetal salt mixtures and at 10 C/min for metal salt-doped APP samples.
Weight (%)
the outcome of the APPemixed salt experiments (see below), the most APP-interactive salts were selected along with others, which enabled the effect of different cations and ions to be investigated. Nominal metal ion concentrations of 0.5, 0.75, 1.0, 1.25, 1.50 and 1.75 mol% were prepared with respect to ammonium polyphosphate. Pastes were made for each specified molar concentration of metal ion within the salts, which were then placed in an oven at 70 C for 12 h to dry. Once dry, the ‘doped’ ammonium polyphosphate samples were each ground down in a pestle and mortar to produce a series of powders for thermal analysis. The effect of the presence of APP (in both standard and doped forms) on the decomposition of cellulose was also investigated using manganese sulphate as the heavy metal salt dopant. Pulverised cotton was intimately mixed with APP samples containing 0, 0.96, 1.19 and 1.41 mol% Mn2C ion as manganese sulphate in a 1:1 mass ratio. The resultant powders were then investigated by thermal analysis.
0.2
0.1
0
0
200
400
600
800
0 1000
Temperature (°C) Fig. 1. TG-DTG responses of ammonium polyphosphate and ammonium polyphosphate doped with 1 mol% manganese (as manganese sulphate).
117
P.J. Davies et al. / Polymer Degradation and Stability 88 (2005) 114e122 0.15
0.7 APP-1 Mn+
APP
0.1
APP-1 Mn+
0.6
0.05
0 0.4 -0.05 0.3 -0.1 APP
Temperature Difference (°C/mg)
Derivative Weight (%/°C)
0.5
0.2 -0.15
0.1
-0.2
0 0
200
400
600
800
-0.25 1000
Temperature (°C) Fig. 2. DTG-DTA responses of ammonium polyphosphate and ammonium polyphosphate doped with 1 mol% manganese (as manganese sulphate).
in three stages. It is suggested that the first two stages, shown in Fig. 1 between 200 and 400 C as the minor DTG peak with a maximum at 330 C, are due to evolution of ammonia and water as gaseous products. The first stepdcorresponding to a small weight loss seen
as the small shoulder at about 200 Cdis probably due to partial degradation of the crystalline crystalline form I as it changes to the more stable form II accompanied by cross-linking and loss of some NH3. An endothermic peak likely to correspond to this may be observed in the DTA trace within Fig. 2 with a maximum at 180 C. Coincident with the evolution of the water and ammonia during the first two steps, acidic hydroxy groups will be formed, and cross-linking of the residue will also occur through the formation of PdOdP or PdNHdP bonds resulting in ultraphosphate structures. Before complete cross-linking can be achieved, phosphate chain fragments are evolved and ammonia is completely eliminated from the ultraphosphate structures; this occurs during the final decomposition stage with a maximum at 660 C. It is thought that the residue will be a complex phosphorus/nitrogen compound [8]. Figs. 1 and 2 also show the effect of doping the APP with 1 mol% manganese ion as manganese sulphate and it is seen that there is a general shift to lower temperatures in both respective DTA and TGA traces. In order to more effectively quantify the possible effect of added metal ions, three transition temperatures have been identified. These are the temperature at the minimum of the melting endotherm for the mixture (designated 1st DTA in Table 3), the temperature of the minimum for the endotherm relating to the first stage of decomposition (2nd DTA in Table 3) and the temperatures of onset and peak for the second derivative of the TGA signal during the first stage of decomposition (1st TGA onset and 1st TGA maximum in Table 3). 3.2. Thermal decomposition behaviour of metal salteAPP mixtures Table 3 summarises the data from the thermal analysis of the metal salteAPP mixtures. It is evident
Table 3 Transition temperatures from DTA/TGA analyses of mixed samples Sample APP APPC2% APPC2% APPC2% APPC2% APPC2% APPC2% APPC2% APPC2% APPC2% APPC2% APPC2% APPC1% APPC2% APPC3% APPC4% APPC5%
Na(CH3COO) Na2SO4 Mg(CH3COO)2 MgCl2 MgSO47H2O Mn(CH3COO)2 MnSO4H2O Zn(CH3COO)2 ZnCl2 ZnSO47H2O FeCl3 FeSO4
1st DTA ( C)
2nd DTA ( C)
1st TGA onset ( C)
1st TGA maximum ( C)
180.7 179.2 179.6 180.5 179.5 179.2 179.9 178.2 180.3 174.3 177.8 194.6 177.7 175.4 173.8 172.1 170.8
333.9 329.9 326.9 339.3 328.4 332.0 335.3 293.5 332.2 327.7 297.2 330.1 300.3 308.4 333.3 328.3 329.1
303.8 305.0 300.6 300.5 308.0 301.9 305.1 282.9 299.1 307.3 288.9 301.9 287.8 299.5 312.1 308.4 302.4
327.9 330.6 329.0 335.3 329.7 329.3 327.4 336.0 330.9 325.0 327.8 328.1 323.5 322.2 334.0 326.3 328.4
118
P.J. Davies et al. / Polymer Degradation and Stability 88 (2005) 114e122
that the presence of metal salts has little effect on the melting transition of APP but, depending on the salt, there are slight to significant reductions in the decomposition DTA minimum temperatures (2nd DTA) with sodium acetate and sulphate, magnesium chloride, zinc chloride and ferric chloride producing slight depressions. Significant reductions are observed with manganese sulphate, zinc sulphate and ferric sulphate, this last depending on concentration. Shifts to lower temperature of the onset of mass loss (1st TGA onset in Table 4) show slight depressions for sodium sulphate, magnesium acetate and sulphate and zinc acetate, while significant reductions occur with 2% w/w manganese and zinc sulphates and 1% w/w ferric sulphate. These respective shifts are not reflected in similar reductions for the respective TGA first peak maxima except for ferrous sulphate present at 1 and 2% w/w. The TGA onset temperature results of Lewin and Endo [6] report data only for the effect of manganese oxide, MnO on the decomposition temperature of a range of MnO (0.5e5% w/w)epentaerythritol (PER) (8.4% w/w)eAPP (16.6% w/w)epolypropylene (PP) mixtures. These appear to reflect the onset of decomposition of APPePER in the main, since temperatures in the range 233e270 C are recorded relative to the value of 259 C in absence of MnO. Here a significant temperature depression to 229 C is observed when 1% w/w MnO is present. This same paper reports that the observed catalytic activity of metal salts on the overall flame retardancy of PPeAPPePEResalt mixtures, as determined by limiting oxygen index is in the order: manganese sulphate manganese acetate O zinc sulphate O zinc acetate. From Table 3, the depression of the first DTA decomposition minima are in the order: 2% w/w manganese sulphate O 2% w/w zinc sulphate O 1% w/w ferric sulphate [2%
w/w zinc acetate and the depression of onset of TGA temperatures are in the order: 2% w/w manganese sulphate O 1% w/w ferric sulphate O 2% w/w zinc sulphate O 2% w/w zinc acetate. Our results indicate that manganese acetate has no effect on either the DTA or TGA responses of APP. 3.3. Effect of metal salt doping on decomposition of ammonium polyphosphate Table 4 lists the actual metal ion concentrations for the six selected salts and Fig. 3 shows the typical TGA data for ammonium polyphosphate doped with 1 mol% metal ion of each salt. Although the individual responses for each of the salts cannot be clearly distinguished, it is generally evident that in all cases, again the presence of metal salts may have significant effects on the decomposition of the ammonium polyphosphate. In terms of discriminating between the TGA responses more easily, the relative order of decreasing residues at 800 C are: sodium tungstate (8.8%) O aluminium sulphate (7.5%) O magnesium chloride (4.9%) O manganese(II) sulphate (3.7%) O zinc sulphate (3.0%) O copper(II) chloride (2.72%). The two last APPesalt combinations produce almost coincident TGA responses as shown in Fig. 3. Fig. 4 shows the effect of metal ion presence on the APP peak DTG temperature relating to the initial evolution of ammonia and water. Again it is apparent that the presence of the metal salt results in a general reduction in the decomposition temperature of the APP. It can be seen that, whilst there is some variation from a clear trend, the reduction in the decomposition temperature does appear to show a minimum at specific and low levels of metal salt concentration as previously noted by Lewin and Endo [5,6].
Table 4 Residual-after-transition (RAT) weight (%) after first APP decomposition stage as defined by the end temperature of the 1st major DTA endotherm Salt concentration with respect to APP (mol%)
Aluminium sulphate
Copper(II) chloride
Magnesium chloride
Manganese(II) sulphate
Sodium tungstate
Zinc sulphate
0 0.25 0.49 0.5 0.73 0.75 0.96 1 1.19 1.25 1.41 1.5 1.63 1.75 2
77.8 76.7
77.8 77.2
77.8 76.3
77.8
77.8 78.3
77.8
76.8
76.8
77.1
76.3
77.3
77.5 77.0 75.8 75.2
77.3
76.3
74.9
77.1
76.6
74.1
76.1
74.8
75.6 75.8 76.5 76.3 75.3 75.9 74.7
77.7
75.9 77.1
76.6 76.5 76.2
119
P.J. Davies et al. / Polymer Degradation and Stability 88 (2005) 114e122 120
330
328 100
326
324
Temperature (°C)
Weight (%)
80
APP 60
322
320
318
316 APP/Al2(SO4)3
40
314 APP/Na2WO4 312
20
APP/MgCl2 310 0.00
APP/ZnSO4 APP/CuCl2 APP/MnSO4 0 0
200
400
0.50
1.00
1.50
2.00
Metal Ion Concentration (mole %) 600
800
1000
Temperature (°C)
Aluminium Sulphate Magnesium Chloride Sodium Tungstate
Copper II Chloride Manganese II Sulphate Zinc Sulphate
Fig. 3. TGA responses of ammonium polyphosphate before and after doping with metal salts.
Fig. 4. Change in DTG peak temperature relating to first stage of ammonium polyphosphate decomposition.
Fig. 5 shows the peak DTG temperatures relating to the mass loss arising from the complete cross-linking of the ultraphosphate structures. For both transitions, generally the addition of a metal salt reduces each below respective transition temperature values for the APP alone. Here it can be seen that respective minima are not observable and after an initial sharp decrease of up to 0.5 mol%, increasing metal ion concentrations appear to produce a gradually reducing decomposition temperature peak value. However, with regard to promoting APP salt decomposition and in correlating our data with that of Lewin and Endo [5,6], it is the effect of metal ions on the first DTA/DTG peak that appears to be of greater interest, especially if such shifts may relate to improved flame retardant properties. Of specific interest here is our observed maximum temperature shifts at about 0.6 mol% for MnSO4 and 1.2 mol% for ZnSO4 which are equivalent to 0.35% and 0.41% (w/w) with respect to each metal ion present. Lewin and Endo [5,6] report maximum flame retardant activities (as LOI) for MnSO4 at 0.33% w/w Mn and for ZnSO4, 0.45% w/w Zn ion presence. They also report a maximum depression of both onset and maximum temperatures of thermogravimetric responses in PPeAPPePEReMnO mixtures at 1% w/w MnO as previously mentioned. It should be
noted that all the concentrations cited in their work are with respect to total mass of mixture, whereas our metal ion concentrations are with respect to APP. Since the concentration of APP in the polypropylene APPe PEReMnC compositions was constant at 16.6% w/w [6], metal ion and salt concentrations with respect to APP present will be nominally higher by a factor of about 6. Only if during compounding of these samples, all metal ions concentrate within the APP present will this increased concentration be realised, however. Probably, this will not be the case and so metal ion concentrations in contact with dispersed APP particles may be closer to the values in our work. Given that this is the case, then there is a remarkable similarity between the two sets of results in spite of the different preparation methods used. Whilst considering the maximum rates of the two decomposition stages gives an indication of the effect of the metal salt presence on the decomposition of the ammonium polyphosphate, further analysis may shed light on whether the metal salt produces increased crosslinking within the ammonium polyphosphate during the decomposition. This possibility was proposed by Lewin and Endo [5,6] at higher salt concentrations. Considering Figs. 1 and 2, it is possible to measure the residueafter-transition (RAT) for the first APP decomposition
120
P.J. Davies et al. / Polymer Degradation and Stability 88 (2005) 114e122
stage where the TG-derived residual masses are determined at the end temperature of each DTA endotherm (see Fig. 2). These are listed in Table 4 and it can be seen that there is little change in the residue after the first stage of decomposition irrespective of the absence or presence of the metal salt or the type of metal salt present. This suggests that the presence of the metal salt may not increase the level of cross-linking within the ammonium polyphosphate as proposed by Lewin and Endo [5,6]. However, as it has been seen in Fig. 4, the presence of the metal salt effectively reduces the temperature at which the first stage of decomposition takes place. We conclude, therefore, that at the small metal ion doping levels used, the effectiveness of the ammonium polyphosphate is increased, not by crosslinking but by reducing the temperature of decomposition and release of polyphosphoric acids. If we apply this argument to the PPeAPPePEReMnC systems studied by Endo and Lewin, then it is possible that the addition of certain metal ions will sensitise APP decomposition and hence phosphorylation of PER at lower temperatures than normal, thus enabling flame retardant activity to occur sooner than in the absence of the metal ions. We suggest, therefore, that this mechanism should be added to those proposed by these authors [6].
For APP-based flame retardant formulations, where low temperatures of activation are required such as in flame retardant back-coatings for textiles and especially those containing easily ignitable cellulosic fibres [1,2], this effect may be of significant importance. Since such coatings are present on the reverse face of the fabric and the igniting flame source is applied to the unretarded front face, it is possible that the presence of small amounts of metal ions would enable release of the mobile polyphosphoric acids to occur well below ignition temperature of front face (e.g. cellulose fibres) and so enable more time for flame retardant diffusion to the front face. 3.4. Effect of presence of doped APP on decomposition of cellulose Fig. 6 shows the TGA curves for the pulverised cotton alone and in the 1:1 mass ratio mixes with APP. As would be expected the decomposition of the APP into the liquid polyphosphoric acids results in the phosphorylation of the cellulose thus favouring dehydration and char formation reactions. The effect of the doped APP is interesting in that the temperature of the onset of decomposition of cellulose is reduced. 120
660
650
100
640
Cell. + APP
Weight (%)
Temperature (°C)
80 630
620
Mn+)
Cell. + APP (1.41
Cell. + APP (0.96
60
Mn+)
610 40 600
Cellulose
590
580 0.00
20
0.50
1.00
1.50
Mn+)
2.00
Metal Ion Concentration (mole %) Aluminium Sulphate Magnesium Chloride Sodium Tungstate
Cell. + APP (1.19
Copper II Chloride Manganese II Sulphate Zinc Sulphate
Fig. 5. Change in DTG peak temperature relating to second major stage of ammonium polyphosphate decomposition.
0 0
200
400
600
800
1000
Temperature (°C) Fig. 6. TGA responses of cellulose and 1:1 (w/w) cellulose: ammonium polyphosphate mixes in absence and presence of manganese ion dopant (as manganese sulphate).
P.J. Davies et al. / Polymer Degradation and Stability 88 (2005) 114e122
Analysis of the DTG traces shows a peak of mass loss rate at 376 C, 315 C, 302 C, 307 C and 313 C for the pulverised cotton, pulverised cotton and APP and the pulverised cotton and APP mix doped with 0.96 mol%, 1.19 mol% and 1.41 mol% MnSO4, respectively. However, visual analysis of the TGA data after the initial mass loss shows that there is a reduced char formation in the presence of the heavy metal salt compared to the undoped APP sample. The general observed order of the retained mass above 400 C is APP O APP (1.41 mol% MnSO4) O APP (0.96 mol% MnSO4) O APP (1.19 mol% MnSO4). A cursory analysis of the residue-after-transition (RAT) data for the APP samples given in Table 4 shows similar residue values after the first decomposition transition, suggesting that there may be some interaction between the heavy metal salt and the cellulose. However, in the absence of residue or volatile analysis, the ability to investigate this is limited. 3.5. Flammability of coated cotton fabrics Table 5 shows the LOI results for the backcoated cotton fabric samples containing the selected APPesalt combinations. The effect of APP alone is considerable in elevating the LOI of the base fabric to a value of 25.1 and addition of each of the salts raises the values further. Listed also are these respective increase in LOI, designated DLOI(APPCsalt) and DLOI(salt), respectively. Interestingly, the sodium and magnesium salts produce the highest increases with DLOI(salt) R 1.8 although as Table 3 shows, these have minimal effects on the first DTG transition maximum temperature of APP. Conversely, salts like manganese and zinc sulphates, which produce the greatest depressions in the same DTG transition temperatures, show lower values of DLOI(salt) in the range 0.9e1.1. However, it must be stated that the LOI test is not testing the ability of a flame retardant to Table 5 LOI results of coated fabric samples Sample and APP and salt content Cotton APP APPC2% APPC2% APPC2% APPC2% APPC2% APPC2% APPC2% APPC2% APPC2% APPC2% APPC2% APPC2%
Add-on LOI DLOI DLOI (% w/w) (APPCsalt) (salt)
0 29 sodium acetate 30 sodium sulphate 30 magnesium acetate 33 magnesium chloride 30 magnesium sulphate 32 manganese acetate 33 manganese sulphate 29 zinc acetate 26 zinc chloride 28 zinc sulphate 30 iron(III) chloride 28 iron(II) sulphate 35
18.7 25.1 26.7 26.9 27.2 26.9 26.9 26.6 26.2 26.1 25.4 26.0 26.3 26.6
e 6.4 8.0 8.2 8.5 8.2 8.2 7.9 7.5 7.4 6.7 7.3 7.6 7.9
e e 1.6 1.8 2.1 1.8 1.8 1.5 1.1 1.0 0.3 0.9 1.2 1.5
121
become active at a lower temperature with resulting increased mobility within or across a substrate such as a backcoated fabric. Clearly, however, the results are interesting and the potential effectiveness of metaldoped APP requires further attention.
4. Conclusions The effect of addition of metal salt addition and in particular certain heavy metal salts, on the decomposition of APP has been studied by thermal analysis (simultaneous TGA and DTA). It has been demonstrated that at small levels, the presence of the metal salt appears to catalyse the decomposition of the APP as shown by a reduction in the temperature of the first major endothermic peak in the decomposition of APP. Whilst it is difficult to see a relationship between the metal salt used to dope the APP and decomposition, there appears to be a concentration dependence on the reduction in the decomposition temperature yielding an optimum concentration at which maximum decomposition temperature depressions occurs. It was seen that the optimised values observed for MnSO4 and ZnSO4 are similar to those seen by Lewin and Endo for maximum enhancement of LOI in PPeAPPePEResalt combinations [5,6]. Intimate mixes of cellulose with APP show changes in the decomposition of the mix in the presence of the metal salt, however, it is interesting to note that, whilst the APP-containing samples show increased char formation, those containing metal ion-doped APP show slightly less of an increase. This suggests that the metal salt may also have an effect on the cellulose char formation and its stability above 400 C. Of course, it is well known that metal salts with Lewis acid characteristics typified by many of the salts chosen in this work, sensitise char formation in cellulose. This effect would be expected to augment any accompanying APP charpromoting effect when in fact the converse is seen, suggesting a degree of antagonism although this is not seen to have a noticeably adverse on flame retardancy as studied by LOI. While the mechanism of interaction is unclear, it is possible that the surface doping of APP by heavy metal ions or salts may sensitise early release of polyphosphoric acid. Whether or not such sensitisation is a consequence of APP crystal destabilisation by adsorbed MnC ions, Lewis acidic effects associated with the metal salt that promotes APP decomposition or some other effect merits further investigation. However, the reduced decomposition temperature of the ammonium polyphosphate within a back-coating formulation should encourage migration of the effective flame retardant species, polyphosphoric acid, to the front of the fabric before the face fibres are able to fully ignite. While the results from backcoated cotton fabrics suggest
122
P.J. Davies et al. / Polymer Degradation and Stability 88 (2005) 114e122
that the limiting oxygen index is increased in the presence of APP treated with metal salt over that of untreated APP, the respective elevations in LOI do not simply relate to a particular salt’s ability to reduce the temperature of the first major DTG transition in APP. The data presented here, however, corroborates previous work [5,6] that clearly shows that there is potential advantage in the addition of metal salts for catalysing APP decomposition and thus potentially increasing its performance as a flame retardant. Further work is still required to more fully elucidate the mechanism for the catalysing effect of the metal salts and the potential for metal salt-doped APP as effective flame retardants.
Acknowledgements
[2]
[3]
[4]
[5]
[6]
The authors wish to acknowledge the support of the UK DTI Link Secretariat and the ENFIRTEX consortium in these investigations.
[7]
References
[8]
[1] Wang M, Horrocks AR, Horrocks S, Hall ME, Pearson JS, Clegg S. Flame retardant textile back-coatings. Part 1. Antimony e halogen
system interactions and the effect of replacement by phosphorus containing agents. J Fire Sci 2000;18:265e94. Horrocks AR, Wang MY, Hall ME, Sunmonu F, Pearson JS. Flame retardant textile back-coatings. Part 2. Effectiveness of phosphorus-containing flame retardants in textile back-coating formulations. Polym Int 2000;49:1079e91. Levchik SV, Levchik GF, Camino G, Costa L, Lesnikovich AI. Mechanism of action of phosphorus-based flame retardants in nylon 6. III. Ammonium polyphosphate/manganese dioxide. Fire Mater 1996;20:183e90. Levchik GF, Levchik SV, Sachok PD, Selevich AF, Lyakhov AS, Lesnikoovich AI. Thermal behaviour of ammonium polyphosphate e inorganic compound mixtures. Part 2. Manganese dioxide. Thermochim Acta 1995;257:117e25. Lewin M, Endo M. Catalysis of intumescent flame retardancy of polypropylene by metallic compounds. In: Proceedings of the American Chemical Society Division of Polymeric Materials: Science and Engineering, Fall Meeting, August 20e24, Washington DC 2000;83:38e9. Lewin M, Endo M. Catalysis of intumescent flame retardancy of polypropylene by metallic compounds. Polym Adv Technol 2003; 14:3e11. Davies PJ, Horrocks AR, Alderson A. Do heavy metal ions sensitise flame retardant activity in ammonium polyphosphate? Proceedings of 14th Conference on Recent Advances in Flame Retardancy of Polymeric Materials, Norwalk, CT, USA: Business Communications Inc; 2003. Camino G, Costa L, Trossatelli L. Study of the mechanism on intumescence in fire retardant polymers: Part V e Mechanism of formation of gaseous products in the thermal degradation of ammonium polyphosphate. Polym Degrad Stab 1985;12:203e11.
