Fire retardant sol–gel coatings for flexible polyurethane foams

Abstract

Flexible polyurethane foam is one of the most versatile materials used in upholstered products; however, untreated polyurethane flexible foams are prone to rapid fire growth. In order to overcome issues encountered by the addition of flame retardant additives during the manufacturing process, the sol–gel process was evaluated to make flexible polyurethane foams flame retardant. Various formulations using different catalysts and monomers were prepared and deposited on flexible polyurethane foams by an impregnation process. Coating morphology was assessed by scanning electron microscopy analyses to optimize the formulation and to obtain a homogenous crack-free coating. Flame retardant properties were measured by mass loss calorimeter and UL 94, and thermal degradation was evaluated by thermogravimetric analysis. Chemical and structural analyses were carried out using attenuated total reflectance Fourier transformed infra-red spectroscopy and electron probe microanalysis. It appears that when a mixture of an appropriate ratio of 3-amino propyl triethoxysilane and diethyl phosphate is prepared in association with tetraethoxysilicate and methyltriethoxysilicate and deposited on flexible polyurethane foam, the coating shows an intumescent behavior when exposed to a flame. The foam also self-extinguishes after 30 seconds of flame application during UL94 test and a 60% reduction of the peak of heat release rate is obtained under mass loss calorimeter conditions.

---

Introduction

Flexible polyurethane foams (FPUF) are mainly used in furniture, mattresses or vehicle seat cushions. Because of the high flammability of FPUF, Europe, the USA and other developed countries and regions have formulated a series of mandatory regulations and standardized testing methods to promote the use of flame retardants in the foams. Indeed, untreated polyurethane flexible foams are prone to rapid fire growth and to the production of highly opaque and toxic fumes (in particular carbon monoxide and hydrogen cyanide). Furthermore, the low viscosity of the decomposition products generates smoldering and severe dripping that increases the fire hazard related to the combustion of such foams. In fact, this flow of flaming liquid often results in a pool fire that promotes flame propagation and boosts the rate of heat release (HRR).¹ As a result, FPUF in furniture and bedding is capable of setting a room to flashover in as little as 5–10 minutes once ignited.²

The flame retardant (FR) additives used initially were mainly aliphatic chlorinated phosphate esters, compatible with foam processing. The choice of FR materials suitable for use in FPUFs is reduced due to the combined requirements of (i) combustion modifying performance, (ii) compatibility with the foaming process and (iii) effect on the physical properties and aging performance of the foam.³

When a low fire requirement is needed (e.g. self-extinguishment after small flame exposition), even if many materials have been tested over several years,^4–6 the main technology still involves the use of halogenated compounds (polybrominateddiphenyl ethers, for example) in conjunction with organic phosphorus compounds (tris(2-chloroethyl) phosphate),⁷ despite their tendency to increase the ratio of toxic fumes during burning. Some special high-performance foam products (able to withstand an intense flame source) include the use of expandable graphite, large amounts of melamine, or ammonium polyphosphate.^8,9 Other FR organic or mineral additives that have been evaluated individually are generally not able to provide the performance provided by the combination of halogenated and phosphorus compounds. For example, some additives (such as borates or alumina trihydrate), known to be used in other types of plastics, have so far not been found effective in flexible foams. Nanotechnologies have also been evaluated and found completely unsuitable in the foam production process. As a result, high-performance foam products generally require a high level of additives, which results in substantially greater costs as well as processing issues (abrasive effect of fillers on machines, decantation) and reduced physical properties (increase of density, faster aging).^10–12 Researchers are now focused on the treatment of the end product to fire retard FPUF. Grunlan et al. used a polyelectrolyte multilayer (PEM) coating, comprised of polyethylenimine and polyacrylic acid (PAA)-stabilized aluminum hydroxide (ATH), deposited onto foam using layer-by-layer (LbL) assembly. With only three bilayers of PEI/PAA-ATH, the coating retained the shape of foam after exposure to a butane torch flame for 10 s. With six bilayers, the flame self-extinguished. They also used a bilayer system of chitosan and vermiculite to form a nanobrick wall exoskeleton, protecting the polyurethane long enough for an intumescing system of chitosan and ammonium polyphosphate to activate and form a bubbled char layer.^13,14 Borreguero et al. grafted flame retardant organophosphonate ester into the FPUF using copper(I)-catalyzed alkyne–azide cycloaddition (CuAAC) ‘click’ reaction of an alkyne-polyol and azidoalkylmonophosphonate, which demonstrated a well-formed polyhedral cell structure and an increase in the fire resistance.¹⁵ Davis et al. used a one pot FR protective layer made of a polysaccharide binder (starch or agar), a boron fire retardant (boric acid or derivative), and a dirt char former (montmorillonite clay).¹⁶ In our lab, we also managed, using a low vacuum argon plasma treatment, to fire retard open cell FPUF by inducing a graft-polymerization of vinylphosphonates.¹⁷ This treatment led to the suppression of dripping and very rapid self-extinguishment of the foams.

In this article, the sol–gel process was selected as the surface treatment as it has already proven its great potential in making coatings that can penetrate an entire network. The sol–gel technique is based on a two-step reaction (hydrolysis and condensation), starting from (semi-)metal alkoxides (usually tetraethoxysilane, tetramethoxysilane, titanium tetraisopropoxide, or aluminum isopropoxide) that leads to the formation of completely inorganic or hybrid organic–inorganic coatings at or near room temperature. It has recently been found to be a wet-chemical technique for the fabrication of environmentally friendly flame retardant coatings on textiles;¹⁸ however, this process has not been evaluated for foams yet. It has been clearly demonstrated for textiles that sol–gel-derived hybrid architectures are capable of protecting the polymer surface acting as a thermal insulator, thus improving the flame retardant (FR) properties of the treated substrates. However, sol–gel-derived architectures cannot be fully considered effective flame retardant systems, whatever the flaming test considered. Indeed, they operate only in the condensed phase during the combustion of a polymeric material and not in the vapor phase, since they are not able to interact with the released gases and smokes. To overcome this limitation, synergistic or joint effects achieved by combining the sol–gel oxidic phases with active species (phosphorus and/or nitrogen) can be employed.

In this paper, various sol–gel formulations using different catalysts and monomers were applied on open cell FPUFs. The FR properties of the treated foams were assessed by mass loss calorimeter (MLC) and UL94 tests to demonstrate the potential of the sol–gel process for FPUF substrates.

Experimental part

Raw materials

Samples (10 cm × 10 cm × 1.5 cm for mass loss calorimeter (MLC) and 10 cm × 5 cm × 1.5 cm for UL94) were cut from open-cell flexible molded polyurethane foams provided by Saira Seats, France and were used as received without further cleaning. The Saira Seats foams are composed of more than 98 wt% polyurethane and less than 2 wt% bis-chloromethylenebis(bis-2-chloroethyl)phosphate (Amgard V6, CAS no. 38051-10-4). They are obtained by (i) polymerization of a polyol on an isocyanate and (ii) release of carbon dioxide resulting from the polycondensation of an isocyanate on a water molecule. Both reactions occur simultaneously, and the components are added in stoichiometric amounts, in order to guarantee the total polymerization and neutrality of each reactive function (hydroxyl, amine, isocyanate), resulting in an inert polymer without any free monomer.

Chemical products tetraethoxysilicate (TEOS, 98% purity), methyltriethoxysilicate (MTES, 95% purity), 3-amino propyl triethoxysilane (APTES, 97% purity), diethyl phosphite (DEP, 99% purity), dibutyltindiacetate (DBTA, technical grade), titanium diisopropoxydebis(acetylacetonate) (TDIPA, 75% purity in isopropanol) and tin(II) 2 ethylhexanoate (TEH, 92.5–100% purity) were purchased from Sigma Aldrich, France. Diethylphosphatoethyltriethoxysilane (DPTES, 92% purity) was purchased from Abcr, GmbH & Co, Germany.

Sol–gel process

Different sol–gel formulations were tested (Table 1). A mixture of TEOS and MTES was used as the basis of all formulations.

Table 1 The different sol–gel coatings formulations

Formulations	TEOS (ml)	MTES (ml)	DPTES (ml)	APTES (ml)	DEP (ml)	Catalysts (ml)
						TEH	DBTA	TDIPA
Sol/TEH	5.8	1.44				0.3
Sol/DBTA	5.8	1.44					0.3
Sol/TDIPA	5.8	1.44						0.3
Sol/DPTES	5.8	1.44	3.6			0.3
Sol/DPTES/APTES	5.8	1.44	3.6	25		0.3
Sol/APTES2/DEP1	5.8	1.44		6	3.2	0.3
Sol/APTES4/DEP2	5.8	1.44		12	6.5	0.3
Sol/APTES8/DEP4	5.8	1.44		25	13	0.3
APTES8/DEP4				25	13	0.3
APTES8/DEP4 without catalyst				25	13
Sol/DEP4	5.8	1.44			13	0.3

---

MTES was added to TEOS in order to limit shrinkage and cracks of the coating during drying. However, since MTES is also known to increase gelation time, its amount was kept low.¹⁹

In all formulations, 10.8 ml ethanol with 5.8 ml TEOS and 1.44 ml MTES were mixed for 10 min in a 250 ml beaker. Then, flame retardant precursors were added to the mixture and stirred for 5 min before the addition of 216 ml of deionized water. DPTES has to be added in a formulation containing TEOS, the presence of a crosslinking agent like TEOS being indeed necessary to build a three dimensional network.²⁰ The highest DPTES concentration that can be introduced in such mixed organic–inorganic systems in order to obtain a monolithic and transparent gel is 30% molar ratio, which is the ratio used in this paper. After 5 minutes' stirring, 0.3 ml of the catalyst was added. The first sol–gel solutions were made in an acidic medium to catalyze the reaction; however, FPUFs degrade in acidic solutions and the foams showed major discoloration (turning from yellow to pink). To avoid any damage to the foams, catalysts were employed. Different catalysts have been studied in the literature.^21,22 Following the state of the art, three different catalysts were tested (DBTA, TDIPA and TEH). Foams were immersed after 3 min stirring of the catalyzed formulation and pressed several times for solution absorption without stirring. The foams were finally left to drip-dry for 1 hour at 70 °C in a convection air oven and 2 days at room temperature before characterization.

Fire testing

A Fire Testing Technology (FTT) Mass Loss Calorimeter (MLC) was used to perform measurements on samples following the procedure ASTM E 906. The equipment is identical to that used in oxygen consumption cone calorimetry (ISO 5660), except that a thermopile in the chimney is used to obtain the heat release rate (HRR) rather than employing the oxygen consumption principle. Foam samples (10 cm × 10 cm × 1.5 cm) were tested in horizontal orientation. Samples were then wrapped in aluminum foil leaving the upper surface exposed to the heater (external heat flux = 50 kW m⁻²) and placed on a ceramic backing board at a distance of 35 mm from cone base. MLC was used to determine the following fire properties: heat release rate (HRR) as a function of time, peak of heat release rate (pHRR) and total heat release (THR). All values were calculated using FTT's MLC calculation software. Tests were performed three times to ensure reproducibility.

Foam samples (10 cm × 5 cm × 1.5 cm) were also tested using the UL94 test (ASTM D3801, 1996 standard). In a UL94 test, the tested sample specimen is supported in the vertical position and a flame is applied to the end of the specimen. The flame is applied for ten seconds and then removed until flaming stops. The flame is then re-applied for another ten seconds and then removed (in order to be more drastic and rank the samples, it was necessary to re-apply the flame for 20 s in our case). Specimens are ranked from V-0 to V-2 following their burning behavior. A specimen is ranked (i) V-2 when burning stops within 30 seconds, but drips of flaming particles are allowed; (ii) V-1 when burning stops within 30 seconds, but drips of particles are allowed as long as they are not flaming; and (iii) V-0 when burning stops within 10 seconds, and drips of particles are allowed as long as they are not flaming.

Microscopy and chemical analyses

The morphologies of the virgin foam and of the different coated FPUF were observed using a field emission gun scanning electron microscope (FEG-SEM), Hitachi S4700, at an accelerating voltage of 6 kV and a current of 15 μA. Images were taken at various magnifications to ensure representative imaging of the samples. After mechanical stress–strain tests a JEOL JSM 7800F was used at 6 kV to analyze the deformed samples. To analyze the chemical composition of samples, Electron Probe Microscopy Analyses (EPMA) were carried out on a CAMECA SX100. The samples were mainly analyzed in cross sections: they were embedded into epoxy resin, polished (up to 1/4 μm) and carbon coated with a Bal-Tec SCD005 sputter coater. Phosphorus and tin X-ray mappings were carried out at 15 kV and 40 nA. Images are color coded; from red for a high concentration of the analyzed element, to black for a low concentration of the analyzed element.

Attenuated total reflection-Fourier transform infrared (ATR-FTIR) spectra of the designed coatings were measured on a Thermo Fischer Scientific Nicolet™ iS™50 FT-IR spectrometer, using an ATR device with a diamond crystal plate. Analyzed coatings were obtained in an aluminum plate after 1 hour of drying in a furnace at 70 °C and for 2 days at room temperature.

Spectra were recorded at 4 cm⁻¹ nominal resolution and 36 measurements were averaged. OMNIC 9.3 (Thermo Fisher Scientific Inc.) software analytical procedures were used to convert the spectra and to perform peak search.

Thermal analyses

Thermogravimetric analyses (TGA) were performed on a Q-5000 TA instrument apparatus. The thermograms were recorded in the 40–800 °C temperature range with a heating rate of 10C min⁻¹ under nitrogen flow, Air Liquide grade (100 ml min⁻¹). For each experiment, samples of 10 mg material were positioned on a gold sheet in order to avoid possible reactions of the phosphorus species of the coating with aluminum open pans.

Mechanical tests

An Instron 4466 universal testing machine was used to carry out stress/strain experiments from uniaxial tensile tests on virgin FPUF and coated FPUF. It was equipped with a 100 N load cell and experiments were carried out at ambient temperature with a 10 mm min⁻¹ speed on 40 × 10 × 3 mm³ samples.

Results and discussion

The FPUFs coated with the different sol–gel formulations were first observed by SEM to check the aspect of the coating (cracks and homogeneity). Some preliminary tests (not shown) revealed that using a sol–gel solution using only TEOS monomers – whatever the catalyst – was not able to provide a crack-free coating; this is why MTES was added to TEOS. It has been shown in the literature¹⁹ that a high content of MTES reduces the shrinkage during thermal treatments. Si–CH₃ groups reduce the network connectivity, thus increasing the compliance of the gel and allowing a better densification during drying. As mentioned in the experimental part, the amount of MTES was kept low to avoid an increase of the gelation time.

The influence on the coating morphology of the three different catalysts (DBTA, TDIPA and TEH) selected to prepare the sol–gel solutions was then analyzed by SEM. The parameter used to differentiate the catalysts was their ability to form a homogeneous crack-free coating. It is noteworthy that this is the first time that TEH and TDIPA catalysts were tested in sol–gel processes. Indeed, these catalysts were initially used to reticulate silicons.²² However, the reticulation reaction of silicon is similar to the condensation sol–gel reaction as it is a chemical reaction between the hydroxyl functions from curable silicone resin and the alkoxy groups of the cross linker. Fig. 1 shows the SEM pictures of the coatings obtained at the surface of the FPUFs. The TEH catalyst gives the best results (Fig. 1c) in terms of coating homogeneity: barely any cracks are observed compared to coatings obtained with the catalysts TDIPA and DBTA (Fig. 1a and b).

Fig. 1 SEM pictures of FPUFs coated with (a) Sol/TDIPA, (b) Sol/DBTA and (c) Sol/TEH.

As the TEH catalyst showed the most promising results in obtaining a crack-free sol–gel coating on FPUF, only the TEH catalyst is used from now on, in this article.

It is reported in the literature that silica sol–gel coatings are capable of protecting polymer surfaces, acting as thermal insulators, and thus improving the flame retardancy of the treated substrate;¹⁸ however, it is also known that such architectures operate only in the condensed phase during the combustion of a polymeric material and not in the gas phase, providing limited FR properties to the substrate. Thus, in this paper, potential additional FR monomers were incorporated into the sol–gel formulation.

Nitrogen or phosphorus monomers can enhance the fire properties of the silica coating. The monomers studied in this paper are DPTES and APTES with or without DEP at different ratios. The first experiment chosen to differentiate the formulations was the MLC at a high heat flux (50 kW m⁻²) to simulate a fully developed fire. Different formulations were tested and the HRR and THR curves obtained are shown in Fig. 2 (each complete formulation is summarized in Table 1).

Fig. 2 (a) HRR curves of virgin FPUF, FPUF treated with Sol/DPTES, Sol/DPTES/APTES, Sol/APTES2/DEP1, Sol/APTES4/DEP2 and Sol/APTES8/DEP4 and (b) corresponding THR curves.

The virgin FPUF shows a sharp HRR curve and high pHRR (425 kW m⁻²). Under MLC conditions, virgin FPUF starts to melt and burns very quickly without leaving any char at the end of the experiment. When the phosphorus silica coating, Sol/DPTES, is applied on the FPUF, the HRR curve is wider and pHRR is smaller (280 kW m⁻² compared to 425 kW m⁻², a 40% reduction). As already seen in the literature for textile substrates, pure silica coating acts as a thermal insulator and slows down the heat transfers from the surface to the inside of the material, and DPTES builds a good organosilicate coating with improved flame retardant properties.¹⁸ It is also probably the same behavior in the present case for FPUF, which explains the reduction of pHHR. However, in order to avoid the destabilization of the silica network, the DPTES ratio cannot be increased in the sol formulation to improve the flame retardant properties (30% molar ratio DPTES maximum). As a consequence, DPTES was tested in association with a nitrogen monomer (APTES) to improve the FR properties of the coating.²³ In terms of pHRR, there is no difference between Sol/DPTES/APTES and Sol/DPTES (pHRR around 280 kW m⁻²). However, more interestingly, the pHRR is shifted to a higher time when APTES is added. Some hypotheses to explain this behavior are that the release of decomposition products is slowed down with the addition of APTES in addition of DPTES, or that the decomposition products are less flammable in those conditions.

Instead of DPTES, DEP can be used in combination with APTES to improve FR properties. Thus, DPTES was replaced by DEP.²⁴ When FPUF is coated with Sol/APTES4/DEP2, the MLC tests show a 60% reduction of pHRR compared to virgin FPUF. The limit amount of APTES and DEP potentially used in such sol is not documented, and thus different amounts of APTES and DEP were tested. However, no better results were obtained. When less FR monomer is used (Sol/APTES2/DEP1), the obtained pHRR is higher (not enough FR properties) as well as when too much is added (extra fuel is given to the flame). The ratio needs to be adjusted to optimize protection of the FPUF. This is confirmed by the THR curves (Fig. 2b), which show that Sol/APTES4/DEP2 is the only formulation that induces a reduction of the THR compared to virgin FPUF.

Even if the FPUF seems to keep its color and hand touch softness, SEM imaging was carried out to check the coating homogeneity and morphology. Fig. 3a and b show the open cell structure of the virgin FPUF. When FPUF is coated with the basic TEOS/MTES (Sol) formulation without FR monomers, it is possible to see in Fig. 2c and d that the FPUF keeps its open cell structure and that the obtained coating is wrinkled but remains crack free. In the same manner, Fig. 3e and f show the FPUF coated with formulation Sol/APTES4/DEP2: at low magnification the FPUF does not exhibit any difference compared to virgin FPUF; the open cell structure is still visible. At higher magnification, a crack-free coating is visible, however, the coating is not perfectly homogenous, some parts without coating being still visible (red circle). When increasing FR monomer concentrations (formulation Sol/APTES8/DEP4) the coating is even less homogeneous but remains crack-free (Fig. 3g and h).

Fig. 3 SEM images at low and high magnification of virgin FPUF (a and b), FPUF treated with Sol (c and d), Sol/APTES4/DEP2 (e and f) and Sol/APTES8/DEP4 (g and h). Inhomogeneities of the Sol/APTES4/DEP2 coating are shown in the red circle.

EPMA results confirm the fact that the coating is quite homogenous on the surface of the FPUF. Fig. 4 shows X-ray mappings of a cross section of the sample Sol/APTES4/DEP2. The P and Sn elements were analyzed and found to be well distributed on the surface of the FPUF. Phosphorus EPMA mapping shows that the coating remains on the surface of the whole 3D FPUF; it does not penetrate inside the PU polymer chains, as can be seen in the literature for other kinds of formulations.¹⁷

Fig. 4 EPMA X-ray mappings of P and Sn elements on a part of the cross section of the sample coated with Sol/APTES4/DEP2.

In order to better understand the role of each additive, further MLC experiments were carried out (Fig. 5). Coatings were made with Sol/DEP4 (without APTES), with APTES8/DEP4 solution without the sol mixture and with APTES8/DEP4 without a catalyst. It appears that when DEP is not associated with APTES, APTES has almost no FR effect (HRR peak reduced from 425 kW m⁻² to 360 kW m⁻²). APTES can potentially form an organosilicate network by itself; however, the results show that it is better to add TEOS/MTES (pHRR at 170 kW m⁻² with TEOS/MTES versus 220 kW m⁻² without TEOS/MTES). When no TEOS/MTES is added to APTES/DEP the coating remains sticky even after several days of drying – the silica network is probably not sufficiently cross-linked or stable to protect the FPUF. Indeed, TEOS/MTES sol even at low concentration creates a stronger silicate coating. In order to confirm this hypothesis, the experiment was carried out with APTES/DEP without any catalyst to have a very low reticulation ratio. There is no FR effect of the coating when the catalyst is not added to the formulation. The coating is not reticulated enough to be stable in the MLC conditions and show FR properties.

Fig. 5 HRR curves of virgin FPUF, FPUF treated with Sol/APTES4/DEP2, Sol/DEP4, APTES8/DEP4 and APTES8/DEP4 without catalyst.

SEM analyses were carried out, after burning, on the most promising sample, i.e. the Sol/APTES4/DEP2 sample (Fig. 6). A black char remains (whereas no remaining char is obtained for virgin FPUF) and big bubbles are visible even at low magnification. Even if the same kind of chemical systems have already been studied in the literature for cotton,^23,24 this intumescent behavior has not been reported for foams yet. At higher magnification, it is possible to see that the bubbles are in fact composed of a network of small particles. Chemical analysis (Fig. 7a) shows the presence of a high amount of Si and O and the presence of C, P, Sn and N. The small particles are probably SiO₂ particles. This hypothesis was confirmed by ATR FTIR analyses. FTIR spectrum (Fig. 7b) exhibits a peak at 1070 cm⁻¹ ascribed to the Si–O–Si stretching mode (arrow 1) and a band at 800 cm⁻¹ due to the Si–O–Si bending vibration (arrow 2), which confirm the presence of the inorganic SiO₂ matrix.²³ Thus, the sol–gel coating is homogenous enough on the whole surface of the FPUF to be able to build a 3D silica structure during burning, preventing the char from collapsing.

Fig. 6 (a) Numerical picture of the char of the sample coated with Sol/APTES4/DEP2; (b) and (c) SEM pictures of the char of the sample coated with Sol/APTES4/DEP2 at low magnification, and (d) at high magnification.

Fig. 7 (a) Chemical analysis of the FPUF char coated with Sol/APTES4/DEP2 and (b) its FTIR spectrum.

In order to test the coated FPUF in a different fire scenario than the MLC, UL94 experiments were carried out in the vertical position. Four formulations were tested: virgin, Sol/DPTES, Sol/APTES4/DEP2 and Sol/APTES8/DEP4on FPUFs. Virgin FPUF melts and drips as soon as the flame is applied and Sol/DPTES FPUF burns entirely; only a small black and brittle char remains after the experiment. FPUFs coated with Sol/APTES4/DEP2 and Sol/APTES8/DEP4 do not burn, even after 30 s exposition to the flame. The surface of the FPUF coated with Sol/APTES4/DEP2 becomes black at the flame position and above (Fig. 8a) but the flame does not spread and self-extinguishes after removing the burner. The Sol/APTES8/DEP4 coated FPUF does not burn at all (Fig. 8b).

Fig. 8 UL94 test results after 30 second flame exposition in vertical position of sample coated with (a) Sol/APTES4/DEP2 and (b) Sol/APTES8/DEP4. Arrows A and B show the area where SEM analyses were performed.

SEM analyses were carried out after the UL94 test on the burnt part of the Sol/APTES4/DEP2 coated FPUF. As in MLC experiments, large bubbles are visible on the surface of the FPUF when exposed to the flame. However, Fig. 9 shows that these bubbles do not have the same structure in the center of the black burned part of the sample (arrow A) and next to the not degraded part of the sample (arrow B). In the black part (Fig. 9a–c), the bubbles correspond to those observed after the MLC tests. Next to the unburnt part of the sample (Fig. 9d and e) the bubbles look smoother; the coating has not been degraded yet. Between the bubbles (arrows in Fig. 9d), the open cell structure of the unburned sample is still visible. The bubbles form a layer protecting the underlying part of the sample from the heat and flame.

Fig. 9 SEM pictures at low and high magnification of (a–c) the black burned part of the sample coated with Sol/APTES4/DEP2 and (d and e) next to the unburned part of the sample. The red arrows show the unburned part of the FPUF.

MLC and UL94 experiments demonstrate that a significant residue is built during burning when Sol/APTES/DEP formulations are applied on the FPUF. In terms of MLC and UL94 results, the formulation Sol/APTES4/DEP2 gives the best compromise. Thus, the study will focus on this formulation from now on. TGA experiments were carried out under nitrogen gas flow to quantify the amount of residue and to determine how the FR sol–gel formulations influence the thermal degradation of the FPUF (Fig. 10a). Different formulations were tested: virgin, Sol and Sol/APTES4/DEP2 coated FPUF. The virgin FPUF degrades in two steps at 230 °C and 380 °C, as well as all the other treated samples (see Table 2). The thermal decomposition of virgin FPUF and the formation of its various decomposition products are well described in the literature.^25,26 The first stage of decomposition leads to depolymerization reactions, which are characteristic of urethane and substituted urea bond cleavage, to form isocyanate, polyol, primary or secondary amine. The following stages of decomposition are mainly due to subsequent degradation of the remaining polyol chains and dimerization and trimerization of isocyanates. The more significant difference is the weight loss of the samples at the end of the experiment. The virgin FPUF does not exhibit any residue at the end of the TGA and the sol sample exhibits only 5 wt% residue at 800 °C. The Sol/APTES4/DEP2 sample shows increased amounts of residue, i.e. 18 wt% (see Table 2). This residue is correlated with the weight gain of the coated FTUF after drying.

Fig. 10 (a) TGA analyses of virgin, Sol and Sol/APTES4/DEP2 coated FPUFs and (b) respective weight difference curves between Sol and Sol/APTES4/DEP2 coated FPUFs and virgin FPUF TGA curves (10 °C min⁻¹, nitrogen flow).

Table 2 Weight gain of coated FPUF after drying, and degradation temperature of each sample and the corresponding percentage of residue under TGA conditions

Sample	Weight gain after drying (wt%)	Degradation temperature 1 (°C)	Degradation temperature 2 (°C)	Residue (wt%)
Virgin FPUF	—	230	380	0
Sol	14	230	380	5
Sol/APTES4/DEP2	28	190	380	18

---

If we look at the weight difference curves (Fig. 10b), it is noticeable that the destabilization is less significant for the Sol coating; indeed, the two TGA thermograms are almost overlapping. The Sol/APTES4/DEP2FR coating starts to degrade at around 190 °C, sooner than the virgin FPUF (230 °C). However, after 290 °C the stabilization is more significant for the Sol/APTES4/DEP2FR coating. Thus, the protective intumescent behavior probably occurs between 190 °C and 290 °C.

Fig. 11 reports the FTIR spectra of Sol and Sol/APTES4/DEP2 coatings. One may notice the presence of absorption bands associated with hydrolyzed APTES, illustrated by different vibrational modes related to NH₂, CH₂ and SiO bonds. The spectral region (3700–2800 cm⁻¹) is characterized by two peaks at 3300 cm⁻¹ and 3290 cm⁻¹ related to the asymmetric and symmetric stretching modes of NH₂, respectively. The broad band ranging from 3700 to 3200 cm⁻¹ is associated with the stretching mode of hydrogen in OH bonds, where it is possible to point out the stretching modes of CH₂ at 2932 cm⁻¹ and 2883 cm⁻¹. The 1562 cm⁻¹ and 1484 cm⁻¹ peaks are assigned to the NH₂ deformation modes of the amine groups.

Fig. 11 FTIR spectra of Sol and Sol/APTES4/DEP2 films and as received DEP and a scheme of hydrogen bonding between APTES and DEP.

It is also possible to notice the deformation mode of the Si–CH₂ peaking at 1410 cm⁻¹ and the asymmetric stretching modes of the Si–O–Si bond peaking at 1130 cm⁻¹ and 1044 cm⁻¹. Diethyl phosphite revealed intense absorptions at 980, 1042, 1258 and 2355 cm⁻¹. The peaks at 980 and 1042, 1258 and 2355 cm⁻¹ are due to νP–O–C, P=O and P–H stretching vibrations, respectively.

The disappearance of the typical DEP P=O stretching band²⁷ at 1250 cm⁻¹ suggests a coordination of all the phosphoryl oxygens to amino-protonated nitrogen via hydrogen bonding (Fig. 11) as reported in the literature.^28,29

As seen by FTIR, DEP is not covalently linked in the network and can potentially be unstable. That is why the UL94 test was carried out again on treated FPUF samples (Sol/APTES4/DEP2) stored for three months at room temperature in the lab. Similar FR behavior was observed compared to the unaged samples. The FPUF FR foam still self-extinguishes even after a 30 second flame exposition. This means that the DEP is well integrated in the Si–O–Si network.

Moreover, in order to check how the coating affects the mechanical properties of the FPUF some stress/strain experiments were carried out on FPUF. Results (Fig. 12) showed that the elastic mechanical property of the FPUF is not lost. Indeed, the characteristic shape of the elastic behavior curve is kept when the coating is applied on the FPUF. However, since the initial slope of the curve is proportional to the elastic modulus of the tested material, there is a slight increase of the elastic modulus of the coated FPUF compared to virgin FPUF.

Fig. 12 Stress/strain curves of virgin FPUF and Sol/APTES4/DEP2 coated FPUF.

SEM analyses were carried out on the broken samples used for the stress/strain experiments in order to verify the stability of the coating after a strong deformation. Fig. 13 shows that the cells are still well-defined and that the coating still adheres well to the PU foam; only a few cracks are visible on some parts of the sample. Thus, the Sol/APTES4/DEP2 coated FPUF seems to keep the mechanical properties of a virgin FPUF even if more tests are necessary to fully characterize its mechanical properties.

Fig. 13 SEM images of virgin FPUF (a and b) and FPUF coated with Sol/APTES4/DEP2 (c and d) after deformation. A few cracks can be observed on the FPUF coated with Sol/APTES4/DEP2 (red arrows).

Conclusion

For the first time, sol–gel coatings were applied on FPUF to confer FR properties to FPUF. Several sol gel formulations with different catalysts and FR monomers were selected. Anon-acidic route was preferred in order to avoid foam discoloration, and a mixture of TEOS and MTES was applied to minimize crack formation in the coating. Instead of an acidic catalyst, tin II 2 ethylhexanoate catalyst was chosen to reticulate the network. It was noticed that DPTES is not sufficient enough to obtain good FR properties. An association of a proper ratio of APTES with DEP in a TEOS/MTES solution is necessary to achieve self-extinguishment in the UL94 test and 60% HRR peak reduction in the MLC test. The mechanical properties of the foam were checked and seemed not to be altered by the coating. The stable coating has an intumescent behavior and builds a 3D silicon network during burning, slows down the release of degradation products and protects the underlying FPUF from burning. The full mechanism of action is currently under investigation in our lab.

References

1. M. Zammarano, S. Matko, W. M. Pitts, D. M. Fox and R. D. Davis, Polym. Degrad. Stab., 2014, 106, 97–107.
2. R. H. Krämer, M. Zammarano, G. T. Linteris, U. W. Gedde and J. W. Gilman, Polym. Degrad. Stab., 2010, 95, 1115–1122.
3. Z. Lan, R. Daga, R. Whitehouse, S. McCarthy and D. Schmidt, Polymer, 2014, 55(11), 2635–2644.
4. S. V. Levchik and E. D. Weil, Polym. Int., 2004, 53, 1585–1610.
5. M. Modesti and A. Lorenzetti, Polym. Degrad. Stab., 2002, 78, 167–173.
6. H. Singh and A. K. Jain, J. Appl. Polym. Sci., 2009, 111, 1115–1143.
7. S. D. Shaw, A. Blum, R. Weber, K. Kannan, D. Rich, D. Lucas, C. P. Koshland, D. Dobraca, S. Hanson and L. S. Birnbaum, Rev. Environ. Health, 2010, 25(4), 261–305.
8. R. Bashirzadeh and A. Gharehbaghi, J. Cell. Plast., 2010, 46(2), 129–158.
9. L. Gao, G. Zheng, Y. Zhou, L. Hu, G. Feng and Y. Xie, Ind. Crops Prod., 2013, 50, 638–647.
10. A. König, A. Malek, U. Fehrenbacher, G. Brunklaus, M. Wilhem and T. Hirth, J. Cell. Plast., 2010, 46, 395–413.
11. J. Jang, H. Chung, M. Kim and H. Sung, Polym. Test., 2000, 19, 269–279.
12. M. Demirel, V. Pamuk and N. Dilsiz, J. Appl. Polym. Sci., 2010, 115, 2550–2555.
13. M. Haile, S. Fomete, I. D. Lopez and J. C. Grunlan, J. Mater. Sci., 2016, 51, 375–381.
14. K. M. Holder, M. E. Huff, M. N. Cosio and J. C. Grunlan, J. Mater. Sci., 2015, 50, 2451–2458.
15. A. M. Borreguero, P. Sharma, C. Spiteri, M. M. Velencoso, M. S. Carmona, J. E. Moses and J. F. Rodríguez, React. Funct. Polym., 2013, 73, 1207–1212.
16. R. Davis, Y.-C. Li, M. Gervasio, J. Luu and Y. S. Kim, ACS Appl. Mater. Interfaces, 2015, 7(11), 6082–6092.
17. M. Jimenez, N. Lesaffre, S. Bellayer, R. Dupretz, M. Vandenbossche, S. Duquesne and S. Bourbigot, RSC Adv., 2015, 5, 63853.
18. J. Alongi and G. Malucelli, J. Mater. Chem., 2012, 22(41), 21805–21809.
19. P. Innocenzi, M. O. Abdirashid and M. Guglielmi, J. Sol-Gel Sci. Technol., 1994, 3, 47–55.
20. A. Cardenas, N. Hovnanian and M. Smaihi, J. Appl. Polym. Sci., 1996, 60, 2279–2288.
21. J. Alongi and G. Malucelli, Polym. Degrad. Stab., 2013, 98, 1428–1438.
22. B. Gardelle, S. Duquesne, P. Vandereecken, S. Bellayer and S. Bourbigot, Eur. Polym. J., 2013, 49(8), 2031–2041.
23. J. Alongi, C. Colleoni, G. Rosace and G. Malucelli, J. Therm. Anal. Calorim., 2012, 110, 1207–1216.
24. G. Brancatelli, C. Colleoni, M. R. Massafra and G. Rosace, Polym. Degrad. Stab., 2011, 96, 483–490.
25. A. W. Benbow and C. F. Cullis, Combust. Flame, 1975, 24, 217–230.
26. S. Liang, M. Neisius, H. Mispreuve, R. Naescher and S. Gaan, Polym. Degrad. Stab., 2012, 97, 2428–2440.
27. L. Jiang, X. Mao, J. Yu and A. Lin, Anti-Corros. Methods Mater., 2009, 56, 13.
28. G. Guerrero, P. H. Mutin and A. Vioux, Chem. Mater., 2001, 13(11), 4367–4373.
29. M. B. Mitchell, V. N. Sheinker and E. A. Mintz, J. Phys. Chem. B, 1997, 101, 11192–11203.

---