Hybrid phosphorus-doped silica architectures derived from a multistep sol–gel process for improving thermal stability and flame retardancy of cotton fabrics

Polymer Degradation and Stability 97 (2012) 1334e1344

Contents lists available at SciVerse ScienceDirect

Polymer Degradation and Stability journal homepage: www.elsevier.com/locate/polydegstab

Hybrid phosphorus-doped silica architectures derived from a multistep solegel process for improving thermal stability and flame retardancy of cotton fabrics Jenny Alongi a, Claudio Colleoni b, Giulio Malucelli a, Giuseppe Rosace b, * a b

Dipartimento di Scienza Applicata e Tecnologia, Politecnico di Torino, sede di Alessandria, Viale Teresa Michel 5, 15121 Alessandria, Italy Dipartimento di Ingegneria Industriale, Università di Bergamo, Viale Marconi 5, 24044 Dalmine, Bergamo, Italy

a r t i c l e i n f o

a b s t r a c t

Article history: Received 30 March 2012 Received in revised form 7 May 2012 Accepted 21 May 2012 Available online 4 June 2012

Hybrid phosphorus-doped silica architectures have been prepared through solegel processes in order to enhance the thermal stability and flame retardancy of cotton. To this aim, diethylphosphatoethyltriethoxysilane has been used as a functional phosphate alkoxysilane in a multistep process, consisting of consecutive depositions for obtaining architectures with a different number of layers (namely, 1, 3 or 6 layers). The role of such architectures has been deeply investigated and correlated with the final properties of the treated fabrics. FT-IR ATR spectroscopy has been exploited for assessing the formation of the silica skeleton on the cotton surface and for evaluating the interactions between the cellulosic fibres and the doped film. The solegel treatments have proved to play a protective role on the degradation of the cotton fibres, hindering the formation of volatile species that fuel the further degradation and favouring the formation of char. Ó 2012 Elsevier Ltd. All rights reserved.

Keywords: Flame retardancy Cotton fabric Solegel Thermal properties Combustion behaviour

1. Introduction In the last years, the solegel technique has remarkably proved its exceptional potential regarding the synthesis of new materials with a high degree of homogeneity at molecular level and with outstanding physico-chemical properties. The solegel represents a versatile synthetic route based on a two-step reaction (hydrolysis and condensation), starting from (semi) metal alkoxides (e.g. tetraethoxysilane, tetramethoxysilane, titanium tetraisopropoxide), that leads to the formation of completely hybrid inorganic or organiceinorganic coatings at or near room temperature [1,2]. These coatings are capable to protect the polymer surface by creating a physical barrier acting as insulator, thus improving the flame retardancy and combustion behaviour of the treated materials. The use of solegel processes for obtaining silica domains to blend with polymeric matrices in bulk is well documented in the literature: indeed, several papers have recently investigated the possibility to reduce the flammability of different polymers, such as epoxy [3e7] and phenolic resins [8,9], polymethylmethacrylates [10,11] and polyesters [12], by exploiting silica phases derived from solegel processes.

* Corresponding author. Tel.: þ39 35 2052021; fax: þ39 35 2052077. E-mail address: [email protected] (G. Rosace). 0141-3910/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. 10.1016/j.polymdegradstab.2012.05.030

As far as textile applications are considered, usually the solegel technique has been proposed for conferring new functional properties to fabrics, such as antimicrobial or UV radiation protection [13e16], dye fastness [17,18], anti-wrinkle finishing [19] and superhydrophobicity [13,20]; its application for imparting flame retardancy to textiles is very recent and has been documented only in the last five years by few research groups [21e27]. Recently, we have published some papers on the possibility to reduce the flammability of fabric substrates by using solegel treatments on synthetic [21] and natural fabrics [22e25], eventually trying to assess phosphorus-nitrogen synergisms [24,26,27]. In the present work, diethylphosphatoethyltriethoxysilane (DPTS) has been exploited for preparing via solegel hybrid phosphorus-doped silica architectures, able to enhance the thermal stability and flame retardancy properties of cotton. This precursor has been chosen because of some of its peculiar features: no cleavage of the SieC and PeC bonds is observed under hydrolytic, basic, or acidic conditions [28]. Furthermore, this compound overcomes the instability of the PeOeSi bonds in the presence of water, i.e. when a mixture of phosphorus and silicon alkoxides is hydrolysed. This organophosphorus compound, which contains a phosphoryl group, may be also an efficient extractant for many metal ions. DPTS has been employed in a novel multistep process consisting of 1e6 consecutive depositions in order to form architectures

J. Alongi et al. / Polymer Degradation and Stability 97 (2012) 1334e1344

differing for the number of layers. The role of such architectures, which combine the thermal insulation capability of the silica coating with the in situ generation of phosphoric acid at high temperatures (this acid is a charring agent for cotton) has been investigated and correlated with the final properties (thermal and thermo-oxidative stability, flammability and combustion behaviour) of the treated cotton. First of all, FT-IR ATR spectroscopy has been exploited for investigating the chemical structure of thin films derived from three DPTS-based xerogels applied on glass slides (model substrate). Then, the thermal and thermo-oxidative stability, the flammability and the combustion behaviour of the solegel treated cotton fabrics have been assessed by using thermogravimetric analyses, vertical flame tests and cone calorimetry, respectively. 2. Experimental part

1335

Two series of treated fabrics characterized by a different layer number were prepared with and without condensation catalyst: each sample was coded as nL (nL_D if the sample was prepared in the presence of DBTA), where n represents the number of layers. As an example, 6L_D indicates a fabric treated with 6 solegel layers in the presence of DBTA. The total dry solids add-on on cotton samples (A, wt.-%) was determined by weighting each sample before (Wi) and after the impregnation with the sol solution and the subsequent thermal treatment (Wf), using a Sartorius balance (104 g). The precursor uptake (reported in Table 1) was calculated according to the following equation:

A ¼

Wf  Wi 100 Wi

2.3. Characterization techniques

2.1. Materials Scoured and bleached 100% plain-weave cotton fabric (237 g/m2) was supplied by Mascioni Spa, Varese, Italy. The fabrics were washed in 2% non-ionic detergent at 40  C for 20 min, and then rinsed several times with deionized water, dried and put into drier for storage. The cleaned samples were conditioned under standard atmospheric pressure at 65  4% relative humidity and 20  2  C for at least 24 h prior to all the experiments. The solegel precursor (diethylphosphatoethyltriethoxysilane, DPTS, purity grade 95%) was purchased from Gelest and used as received; Dibutyl tindiacetate (DBTA, condensation catalyst), hydrochloric acid and ethanol were purchased from SigmaeAldrich and used without further purification. 2.2. solegel treatments on cotton fabrics The sol solutions for the impregnation of the cotton fabrics as well as for the production of the xerogels were synthesized according to the molar ratios listed in Table 1. In greater detail, 10.07 mL (0.03 mol) of DPTS were hydrolysed with 8 mL (0.0008 mol) of HCl (0.1 N) and 5 mL of ethanol, with or without DBTA; subsequently, the sol solution was made up with distilled water to achieve a final volume of 100 mL. In the sol prepared using dibutyl tindiacetate, the precursor:DBTA molar ratio was 1:0.09 and the resulting pH value was ca. 4.5. For all the prepared sols, the DPTS:HCl molar ratio was set at 1:0.026. Finally, the sol solutions were stirred for 4 h at room temperature. In order to produce the xerogels and investigate their chemical structure, a small amount of each sol was applied on glass slides, the solvent was removed at 80  C for 10 min and the thin films were subjected to a thermal treatment at 170  C for 4 min. The cotton fabrics (20  30 cm2) were impregnated with the hybrid sols and afterward were passed through a two-roll laboratory padder (Werner Mathis, Zurich, Switzerland) working with 3 bar nip pressure and obtaining 70% of wet pick-up. After drying at 80  C for 10 min, the fabrics were thermal treated at 170  C for 4 min in a gravity convection oven. Table 1 Composition and total dry solids add-on on cotton samples of hybrid sols. Sample

DBTA

DPTS:DBTA molar ratio

Add-on % (A)

1L 1L_D 3L 3L_D 6L 6L_D

No Yes No Yes No Yes

/ 1:0.09 / 1:0.09 / 1:0.09

5.5 4.6 14.1 12.8 23.9 21.0

The formation of bonds between cotton samples and the different formulations applied onto fabric was investigated by FT-IR spectroscopy. The spectra were recorded at room temperature in the range 650e4000 cm1 (64 scans and 4 cm1 resolution), using a Thermo Avatar 370 spectrophotometer, equipped with attenuated total reflection accessory and diamond crystal. The surface morphology of the treated samples was studied using a LEO1450VP Scanning Electron Microscope (beam voltage: 20 kV), equipped with an X-ray probe (INCA Energy Oxford, Cu-Ka X-ray source, k ¼ 1.540562 Å), which was used to perform elemental analysis. Fabric pieces (0.5  0.5 cm2) were cut and fixed to conductive adhesive tapes and gold-metallized. The thermal stability of the fabrics was evaluated by thermogravimetric (TG) analyses from 50 to 800  C with a heating rate of 10  C/min, both in nitrogen and in air (60 mL/min). To this aim, a TAQ500 thermogravimetric balance was used, placing the samples in open alumina pans (ca. 10 mg). The experimental error was 0.5% on the weight and 1  C on the temperature. The flame retardancy properties of the prepared samples were measured using a vertical fabric flammability test, applying a methane flame for 5 s at the bottom of a fabric specimen (50  100 mm2) and repeating the test 3 times for each formulation in order to get reproducible data; burning time, rate and the final residue were measured. The combustion behaviour of square fabric samples (50  50  0.5 mm3) was investigated using cone calorimetry (Fire Testing Technology, FTT). The measurements were carried out under a 35 kW/m2 irradiative heat flow in horizontal configuration, following the procedure described elsewhere [29]. Such parameters as Time To Ignition (TTI, s), Flame Out time (FO, s), Total Heat Release (THR, kW/m2), peak of Heat Release Rate (PHRR, kW/m2) were measured. Total Smoke Release (TSR, m2/m2), peak of Rate of Smoke Release (PRSR, 1/s), Smoke Factor (SF, calculated as PHRR  TSR, MW/m2) and CO and CO2 release (ppm and %, respectively) were evaluated, as well. The experiments were repeated four times for each material investigated to ensure reproducible and significant data; the experimental error was within 2%. Prior to flammability and combustion tests, all the specimens were conditioned at 23  1  C, for 48 h at 50% R.H. in a climatic chamber. 3. Results and discussion 3.1. Characterization of hybrid silica xerogels on glass slides Pure xerogels applied on optical quality glass slides have been prepared in order to perform analytical measurements, since the infrared absorption bands characteristics of the hybrid thin films

1336

J. Alongi et al. / Polymer Degradation and Stability 97 (2012) 1334e1344

applied onto the fabric surface are covered by the strong vibrational peaks of the cellulosic substrate. The FT-IR ATR spectrum of the xerogel applied and annealed on glass slides and the frequencies of major absorption bands are shown in Fig. 1 and Table 2. In the xerogel spectrum, the bands at 1020 and 724 cm1 are attributed to the SieOeSi asymmetrical and symmetrical stretching, respectively and confirm the formation of a silica phase. The band with moderate intensity at 1413 cm1 refers to d (SieCH2) vibration of ethyl fragments. The group of weakly intense bands in the range 1360e1480 cm1 and 2800e3000 cm1 confirms the presence of diethyl phosphate groups. These latter are characteristic of ns,as (CH) vibrations, whereas the absorption bands at 1368, 1393, 1445, and 1480 cm1 can be assigned to u(CH2), ds (CH3), das (CH3), and d (CH2) vibrations of ethoxy groups, respectively. The n (P]O) absorption band is identified in the FT-IR spectrum of pure DPTS at 1235 cm1. Furthermore, an absorption band at 778 cm1, attributable to the PeO vibration, is observed. It is noteworthy that the absorption bands at 2350e2400 and around 800 cm1 due to the PeOH stretching and PeOeP bending modes, respectively, are not present in the spectrum of the xerogel. This suggests that hydrolysis reaction on P-OEt units does not occur: indeed, their typical hydrolysis conditions involve refluxing overnight in strong acidic conditions [30]. Since all xerogels contain water, the medium intensity absorption band characteristic of d (H2O) vibrations appears at 1646 cm1, while a broad intense absorption band caused by the n (OH) vibrations of adsorbed water is observed above 3460 cm1. No significant differences can be observed between the spectra of the xerogel synthesized with and without DBTA catalyst (Fig. 2). 3.2. Characterization of hybrid silica xerogels obtained on treated cotton fabrics The FT-IR ATR spectra of the untreated and treated cotton fabrics are shown in Fig. 3, while the frequencies of the main absorption bands are collected in Table 2. The hybrid SieP xerogels applied onto cotton samples change some absorption bands because of the interactions between silica and cellulose. Indeed, an overall slight decrease of the intensities at 3500e3000 cm1, characteristic of cellulose hydrogen bonded OeH stretching vibrations is observed in the spectra of the treated fabrics, as well as the presence of SieOeSi asymmetrical stretching absorption band at 1020e1027 cm1. The signal related to PeO is present at 783 cm1.

Fig. 1. FT-IR ATR spectrum of the synthesized hybrid organiceinorganic xerogel.

Table 2 Main vibration modes ascribable to DPTS in xerogel thin film on glass slides and on cotton fabrics. Frequencies [cm1] On glass substrate

On fabric substrate

From literature

2800e3000 1368 1393 1445 1480 1413

2800e3000a 1370a 1392a – e 1414a

2800e3000 1364e1372 1392 [31] 1442e1446 1478 [31] 1409e1417

778 1235 1020 724

783 1210e1223 1020e1027 e

784 [27] 1241 [32] 1001 [26] 749e786 [26]

a

Vibrational modes

[26] ns,as (CH) [31] u (CH2) ds (CH3) [31] das (CH3) ds (CH2) [31] d (SieCH2)

n (PeO) n (P]O) n (SieOeSi) d (SieOeSi)

Absorption band ID

: Ethoxy groups

6

C B -

Ethyl , fragment A A

; 7

Only for 6L.

Such peaks increase with increasing the number of layers applied onto the textile samples. The n (P]O) absorption band, located at 1235 cm1 in the FT-IR spectrum of the xerogel, is shifted towards low frequencies in the spectra of the treated fabrics by about 10e25 cm1. This finding could indicate the participation of the phosphoryl group in the formation of hydrogen bonds. As for the xerogels coated onto the glass slides, the FT-IR ATR spectra of the treated fabrics do not show any difference when the solegel is carried out with or without DBTA. A schematic representation of the possible crosslinking interaction between cellulose and silica is shown in Fig. 4. 3.3. Morphology SEM observations have been performed in order to establish the morphology of the hybrid phosphorus-doped silica architectures deposited on cotton fibres through solegel processes. The typical morphology of a natural fibre is reported in Fig. 5a: as expected, the surface of raw cotton fibres shows a certain level of heterogeneity and irregularity, which disappear when the fibres are solegel treated. Indeed, their surface becomes smooth and flawless. All the treated fibres appear homogeneously covered by the oxidic phase, irrespective of the number of layers deposited or the use of DBTA as condensation catalyst. As an example, some typical SEM

Fig. 2. FT-IR ATR spectra of the xerogel synthesized with or without the condensation catalyst.

J. Alongi et al. / Polymer Degradation and Stability 97 (2012) 1334e1344

1337

micrographs of 1L, 3L and 6L samples are reported in Fig. 5b, c and d, respectively: a thin coating completely covers the cotton surface; furthermore, 6L sample shows the presence of some aggregates. The elemental analysis evidences the presence of Si and P elements, the main constituents of the obtained architectures: their distribution is homogeneous and regular in all the samples, as shown in Fig. 6 for 6L_D sample. The qualitative observation by electron microscopy has not been useful to distinguish significant differences among the hybrid architectures, either characterized by a different layer number (namely, 1, 3 or 6 layers) or prepared with or without condensation catalyst. Otherwise, in the latter case, the evaluation of the uptake after the solegel treatment has not revealed remarkable differences comparing the samples prepared with or without condensation catalyst (see Table 1). 3.4. Thermal stability The thermal and thermo-oxidative stability of the solegel treated samples has been assessed by thermogravimetric analysis and compared with that of pure cotton. Tables 3 and 4 collect the

Fig. 3. FT-IR spectra of untreated and DPTS treated cotton fabrics.

O C2H5O

O OH

P

O OC2H5 C2H5O

P

O O

P

CH2

CH2

CH2

CH2

CH2

CH2

Si O

O

Si O

O

Si O

O

Si

O O

O

O OH

Si O

OC2H5 H2 C

H2 C

P

C2H5O

Cotton fabric Fig. 4. Schematic representation of possible DPTS grafting on the cotton fabric surface.

Fig. 5. SEM magnifications of untreated (a) and 1L (b), 3L (c) and 6L (d) solegel treated samples.

O OH

1338

J. Alongi et al. / Polymer Degradation and Stability 97 (2012) 1334e1344

Fig. 6. Elemental analysis of 6L_D sample.

data in nitrogen and air, respectively; Fig. 7a and b plot the TG and dTG curves of the samples. As already demonstrated [24e27], the thermal degradation of cotton in nitrogen proceeds by only one step, during which the maximum weight loss is registered (Fig. 7a). Indeed, cotton usually pyrolyses in nitrogen according to two alternative pathways, which involve the decomposition of the glycosyl units to char at lower temperature and the depolymerization of such units to volatile products containing levoglucosan at higher temperature; such behaviour is well described in the literature ([33] and references quoted in). As these pathways are concurrent, it is not possible to distinguish two weight losses as a function of temperature by using thermogravimetric analysis. The presence of the coating is responsible of a strong anticipation of the cellulose decomposition, as revealed by the Tonset5% and Tmax values reported in Table 3 and well depicted in TG and dTG curves (Fig. 7a). This effect has been observed for all the samples under investigation, regardless of the number of deposited layers or the eventual use of the condensation catalyst. Indeed, up to 340  C (cotton Tmax) all the curves of the treated samples are shifted to lower temperatures if compared with that of pure cotton. In spite of this, the char left by the treated samples is much higher than that of cotton, as evident by the residues measured at Tmax and reported in Table 3. These carbonaceous structures derived from the decomposition of the treated cotton turn out to be thermally stable up to 700  C and justify the increased residues of the solegel treated samples with respect to the pure fabric. In addition, the residues increase by increasing the number of layers, irrespective of the use of the condensation catalyst. Indeed, comparing the homologue pairs (namely, 1L with 1L_D, 3L with 3L_D and 6L with 6L_D) negligible differences can be observed in terms of Tonset5%, Tmax and residues at Tmax and 700  C.

As far as the thermo-oxidative stability is considered, cotton degradation usually occurs by three steps. The first one (300e400  C) involves two competitive pathways, which yield aliphatic char and volatile products; during the second step (400e800  C) some aliphatic char converts to an aromatic form, yielding CO and CO2 as a consequence of simultaneous carbonization and char oxidation. Within the last step (at ca. 800  C), the char (and any remaining hydrocarbon species) are further oxidized mainly to CO and CO2. In the present work, two decomposition peaks are observable in between 300 and 500  C for cotton (namely at 340 and 470  C, Table 4 and Fig. 7b). Once again, the presence of the coating induces the anticipation of the cotton decomposition (Tonset5%) and of the maximum weight loss (Tmax1), but favours the formation of a thermally stable char that evolves at high temperatures (Tmax2 and 700  C) and leaves a residue significantly higher with respect to that of pure cotton (Table 4). Furthermore, as already observed in nitrogen, the residue increases by increasing the number of layers also in air. From an overall consideration, it is possible to observe that the samples prepared with DBTA as condensation catalyst (namely, 1L_D, 3L_D and 6L_D) form lower residues during the first (at Tmax1) Table 3 TGA data of cotton fabrics in nitrogen atmosphere. Sample

Tonset5% [ C]

Tmaxa [ C]

Residue at Tmax [%]

Residue at 700  C [%]

CO 1L 1L_D 3L 3L_D 6L 6L_D

309 280 282 277 278 280 275

360 316 318 309 310 306 305

44.3 62.8 62.7 73.0 72.9 75.3 76.1

9.1 26.6 25.9 35.8 36.2 40.2 38.9

a

From derivative curves.

J. Alongi et al. / Polymer Degradation and Stability 97 (2012) 1334e1344 Table 4 TGA data of cotton fabrics in air atmosphere. Sample

Tonset5% [ C]

Tmax1a [ C]

Residue at Tmax1 [%]

Tmax2a [ C]

Residue at Tmax2 [%]

Residue at 700  C [%]

CO 1L 1L_D 3L 3L_D 6L 6L_D

295 279 279 278 271 278 270

340 310 311 303 303 303 301

53.0 63.0 61.8 73.9 71.3 83.1 76.4

470 505 510 524 631 531 622

6.0 18.0 17.0 34.3 32.0 36.3 32.1

1.8 6.9 6.8 18.2 16.3 23.0 20.5

a

From derivative curves.

1339

and third (at 700  C) steps, if compared with the residues of the counterparts prepared without DBTA (namely, 1L, 3L and 6L); however, their Tmax2 are higher. The collected results prove that the deposited architectures on the cotton fibres play a protective role against the thermal decomposition of cotton, favouring the char formation both in nitrogen and air. We have already demonstrated that a compact silica coating (able to act as thermal insulator both in nitrogen and in air) can be deposited on the cotton fabrics using tetraethylorthosilicate or tetramethylorthosilicate as precursor for the solegel processes [24e26]. In addition, the joint effect between silica and phosphoric acid (produced by the hybrid architectures synthesized from diethylphosphatoethyltriethoxysilane) in favouring the char formation has been recently explored [27]. In the present study, the

Fig. 7. TG and dTG curves of untreated and solegel treated cotton fabrics in nitrogen (a) and air (b).

1340

J. Alongi et al. / Polymer Degradation and Stability 97 (2012) 1334e1344

observed during the thermal degradation of the samples in air (see Tonset5% in Table 4). After the ignition, the combustion proceeds: the heat release rate increases with a maximum peak (namely, HRR peak) and then decreases down to the flame out (FO). The solegel treatment reduces the HRR peak of cotton, irrespective of the number of layers or the use of the condensation catalyst; furthermore, the best flame retardancy performances have been achieved when the condensation catalyst is employed. Indeed, comparing the homologues pairs, the highest HRR peak decrease is observed for the samples prepared using DBTA (124 vs. 115 kW/m2, 110 vs. 79 kW/m2 and 113 vs. 87 kW/m2 for 1L, 1L_D, 3L, 3L_D, 6L and 6L_D, respectively). It is also noteworthy that the coating protects the cotton fabric decreasing the duration of the combustion, as shown by the FO values, and hindering the formation of volatile species, as evidenced by the strong THR decrease (Fig. 9 and Table 6). With regard to evolved smokes (Table 6 and Fig. 10), the average TSR (which indicatively refers to the smoke amount generated in a fullscale fire) is strongly affected by the coating as well as the SF (a useful parameter able to compensate the incomplete combustion of the solegel treated samples, according to Horrocks et al. [34]). Indeed, the deposited architectures, protecting the fabric, favour the char formation and hinder the evolution of volatile species that fuel the further degradation: as a consequence, the smoke production is significantly lowered. The best results are achieved when 6 layers are deposited (0.4, 0.3 vs. 5.4 MW/m2 for 6L, 6L_D and cotton, respectively). In addition, the smoke production rate is remarkably reduced by the deposited architectures on the cotton fibres (Fig. 11): indeed, after the ignition, cotton burns vigorously, giving rise to an RSR peak of 1.9 s1 followed by a sharp decrease of the smoke release as the sample is completely consumed. Regardless of the formulations under study, although the anticipation of the ignition in the presence of the coating induces the formation of smokes in advance with respect to pure cotton, as pointed out by the time of RSR peak, at the same time this peak is remarkably decreased since the total smoke release is reduced. The best performances have been achieved when 6 layers are deposited on the cotton fibres, regardless of the use of DBTA (0.4 and 0.3 vs. 1.9 s1 for 6L, 6L_D and cotton, respectively), as depicted in Fig. 11c. As clearly reported by Horrocks and co-workers, the evaluation of CO and CO2 species in conjunction with smoke is very important for two reasons: first of all, CO and CO2 are the main constituents of fire gases and high CO concentrations can be lethal; second, the analysis of these species can provide useful information on the mechanism of decomposition of such polymer, as cotton [34]. Low CO2/ CO ratios suggest inefficiency of combustion inhibiting the conversion of CO to CO2. Generally, flame retardant systems working through flame inhibition result in significantly increased CO yields in the forced flaming combustion of a cone calorimeter test. Table 6 collates the peaks of CO and CO2 yields, their ratios and the corresponding times. Pure cotton produces CO during two steps (after 56 and 90 s) due to the oxidation of the char formed during the combustion. At the same time, CO2 is released after the first

Table 5 Flammability data by vertical flame test. Sample

Total burning time [s]

Total burning rate [mm/s]

Residue [%]

CO 1L 1L_D 3L 3L_D 6L 6L_D

35 30 27 24 20 23 22

7.5 10.0 10.0 10.0 10.0 10.0 10.0

e 13 9 30 25 35 34

proposed multistep deposition process seems to be extremely efficient: the thermal and thermo-oxidative stability of cotton can be strictly related to the number of the layers deposited on the fibres. 3.5. Flame retardancy The flame retardant properties of the treated fabrics have been evaluated in terms of flammability and combustion behaviour by vertical flame tests and cone calorimetry, respectively. As far as flammability is considered, pure cotton and the solegel treated samples have been ignited directly by a flame (in vertical configuration), measuring total burning time, total burning rate and final residue. The flammability data, collected in Table 5, show that all the formulations are able to significantly reduce the total burning time of cotton (in spite of a slight increase of the burning rate) and to protect the treated fabric from the flame, favouring the char formation. More specifically, the architectures prepared without the condensation catalyst are capable to enhance the total burning time very efficiently and to form the highest residues if compared with the homologues prepared with DBTA. This finding can be ascribed to the protective role of the hybrid architectures deposited on the cotton surface that favour the char formation instead of the production of volatile species that could promote further combustion processes. Such results confirm the TGA data and also evidence that the coatings prepared without the condensation catalyst achieve the greater flame retardant effects. The most important result is the formation of high coherent and consistent residues at the end of the test, although all treated fabrics burnt their entire lengths at different burning rates as shown in Table 5. As far as the combustion behaviour is considered, the collected data by cone calorimetry are summarised in Table 6 and plotted in Figs. 8e13. More specifically, Fig. 8 shows the HRR curves of pure cotton and solegel treated samples as a function of time: in particular, Fig. 8a, b and c compare 1L and 1L_D, 3L and 3L_D, and 6L and 6L_D, respectively. Taking in consideration of the experimental error, TTI is substantially unchanged when only 1 layer is deposited on the cotton fibres; on the contrary, in the presence of 3 or 6 layers, TTI is strongly reduced (Table 6). This trend is similar to that Table 6 Combustion data by cone calorimetry.

HRR TSR SFa RSR CO CO2 CO2/CO Residue [%] Sample TTI [s] FO [s] THR [m2/m2] [MW/m2] peak [MJ/m2] 2 peak [1/s] Time [s] peak [ppm] Time [s] Peak [%] Time [s] Peak [kW/m ] Time [s] CO 1L 1L_D 3L 3L_D 6L 6L_D a

18 20 20 12 12 10 12

116 66 70 62 70 80 80

4.6 2.6 2.8 2.8 2.6 2.8 1.7

Corresponding to the HRR peak.

143 124 115 110 79 113 87

58 34 34 32 34 32 28

26 20 12 15 6 6 6

5.4 0.8 1.4 1.0 0.4 0.4 0.2

1.9 0.6 1.2 1.1 0.5 0.4 0.3

46 26 26 26 24 26 22

4.6/14.6 23.5 22.7 27.7 15.3 11.3 12.4

56/90 62 66 82 72 80 78

0.11 0.10 0.10 0.09 0.10 0.10 0.10

74 50 50 52 42 44 38

0.024 0.007 0.004 0.003 0.006 0.009 0.008

4 17 13 23 26 31 32

J. Alongi et al. / Polymer Degradation and Stability 97 (2012) 1334e1344

Fig. 8. HRR curves of pure cotton, 1L and 1L_D (a), 3L and 3L_D (b), 6L and 6L_D (c).

Fig. 9. THR curves of pure cotton, 1L and 1L_D (a), 3L and 3L_D (b), 6L and 6L_D (c).

1341

1342

J. Alongi et al. / Polymer Degradation and Stability 97 (2012) 1334e1344

Fig. 10. SF curves of pure cotton, 1L and 1L_D (a), 3L and 3L_D (b), 6L and 6L_D (c).

Fig. 11. RSR curves of pure cotton, 1L and 1L_D (a), 3L and 3L_D (b), 6L and 6L_D (c).

J. Alongi et al. / Polymer Degradation and Stability 97 (2012) 1334e1344

1343

spite of a slight decrease of the CO peak. Finally, the protective role of the solegel treatment has been further confirmed by cone calorimetry: indeed, the residues after the test appear thick, compact and dense (Fig. 13); in addition, the deposition of 6 layers on the cotton fabrics allows achieving the highest residue (Table 6). 4. Conclusions

Fig. 12. CO2/CO ratio of untreated and solegel treated cotton fabrics.

step, during which CO evolves (74 s). When any coating is deposited on the cotton fibres, the production of CO occurs by a single step, whereas the CO2 production is unchanged only as far as the maximum reached peak is concerned. When 1 or 3 layers are deposited, the CO peak strongly increases, whereas its time is reduced: this is due to the anticipated oxidation of the char to CO. Obviously, since the presence of the coating is responsible of a higher char formation with respect to the pure cotton, the CO amount increases. All the formulations do not change the CO2 peak, but are able to strongly decrease the CO and thus their ratio in a remarkable way, regardless of the use of DBTA, as well depicted in Fig. 12. This suggests that relatively larger amounts of CO are produced due to the incomplete combustion of the cotton in the presence of the coating. In the case of 6 layered samples, it is noteworthy that both the systems (namely, 6L and 6L_D) show lower CO2/CO ratios with respect to pure cotton, in

In the present work, the solegel process has been successfully applied for obtaining hybrid phosphorus-doped silica architectures, derived from diethylphosphatoethyltriethoxysilane, which have been shown to enhance the thermal stability and flame retardancy properties of cotton. To this aim, a novel multistep process consisting of 1e6 consecutive depositions has been set, obtaining architectures with a different number of layers. The occurrence of the formation of the hybrid phosphorus-doped silica phase has been assessed by means of FT-IR ATR spectroscopy. As clearly shown by SEM analyses, all the treated fabrics have been homogeneously covered by the formed oxidic phase, irrespective of the number of layers deposited or the use of a condensation catalyst. The presence of the coating turned out to be responsible of a strong anticipation of the cellulose decomposition but, at the same time, promoted a significant increase of the residues at high temperatures. As far as flammability is concerned, the architectures prepared without the condensation catalyst have been able to enhance the total burning time in a very efficient way and to form the highest residues after the test. Despite the strong reduction of TTI, the hybrid coatings proved to protect the cotton fabrics decreasing the duration of the combustion, as shown by the FO values found during cone calorimetry tests and hindering the formation of volatile species. Finally, the smoke production has been significantly lowered in the presence of the hybrid architectures. Acknowledgements The authors would like to thank Fabio Cuttica and Alessandro Di Blasio for the technical support in the flammability and combustion tests, and Daniel Capelli for the support in preparing the solegel treated fabrics. References

Fig. 13. Residues of untreated and solegel treated cotton fabrics at the end of the cone calorimetry test.

[1] Sakka S. Sol-gel science and technology. Topics and fundamental research and applications. Norwell: Kluwer Academic Publishers; 2003. [2] Mahltig B, Textor T. Nanosols and textiles. World Scientific Publishing Co; 2008. [3] Chiang CL, Ma CCM. Eur Polym J 2002;38:2219. [4] Chiang CL, Wang FY, Ma CCM, Chang HR. Polym Degrad Stab 2002;77:273. [5] Chiang CL, Chang RC. Compos Sci Technol 2008;68:2849. [6] Liu YL, Wu CS, Chiu YS, Ho WH. J Polym Sci Pol Chem 2003;41:2354. [7] Liu YL, Chou CI. Polym Degrad Stab 2005;90:515. [8] Chiang CL, Ma CCM, Wu DL, Kuan HC. J Polym Sci Pol Chem 2003;41:905. [9] Chiang CL, Ma CCM. Polym Degrad Stab 2004;83:207. [10] Messori M, Toselli M, Pilati F, Fabbri E, Busoli S, Pasquali L, et al. Polymer 2003; 44:4463. [11] Chiang CL, Chiu SL. J Polym Res 2009;16:637. [12] Ji Q, Wang X, Zhang Y, Kong Q, Xia Y. Compos Part A-Appl S 2009;40:878. [13] Mahltig B, Böttcher H. J Sol-Gel Sci Technol 2003;27:43. [14] Mahltig B, Fiedler D, Böttcher H. J Sol-Gel Sci Technol 2004;32:219. [15] Mahltig B, Haufe F, Böttcher H. J Mater Chem 2005;15:4385. [16] Mahltig B, Böttcher H, Rauch H, Dieckman U, Nitsche R, Fritz T. Thin Solid Films 2005;485:108. [17] Mahltig B, Textor T. J Sol-Gel Sci Technol 2006;39:111. [18] Cireli AC, Onar N. J Appl Polym Sci 2008;109:97. [19] Huang KS, Nien YH, Hsiao KC, Chang YS. J Appl Polym Sci 2006;102:4136. [20] Xue CH, Ji ST, Chen HZ, Wang M. Sci Tech Adv Mat 2008;9:1. [21] Cireli AC, Onar N, Ebeoglugil MF, Kayatekin I, Kutlu N, Culha O, et al. J Appl Polym Sci 2007;105:3747. [22] Hribernik S, Smole MS, Kleinschek KS, Bele M, Jamink J, Gaberscek M. Polym Degrad Stab 2007;92:1957.

1344 [23] [24] [25] [26]

J. Alongi et al. / Polymer Degradation and Stability 97 (2012) 1334e1344

Yaman N. Fibers Polym 2009;10:413. Alongi J, Ciobanu M, Malucelli G. Carbohyd Polym 2011;85:599. Alongi J, Ciobanu M, Malucelli G. Carbohyd Polym 2012;87:2093. Alongi J, Colleoni C, Rosace G, Malucelli G. J Therm Anal Calorim. doi:10.1007/ s10973-011-2142-0. [27] Brancatelli G, Colleoni C, Massafra MR, Rosace G. Polym Degrad Stab 2011;96:483. [28] Cardenas A, Hovnanian N, Smalhl M. J Appl Polym Sci 1996;60:2279. [29] Tata J, Alongi J, Carosio G, Frache A. Fire Mater 2011;35:397.

[30] Guerrero G, Mutin PH, Vioux A. Chem Mater 2001;13:4367. [31] Lin-Vien D, Colthup NB, Fateley WG, Grasselli JG. The handbook of infrared and Raman characteristic frequencies of organic molecules. London: Academic; 1991. 263. [32] Dudarko OA, Melnik IV, Zu YL, Chuiko AA, Dabrowski A. Colloid J 2005;67:683. [33] Price D, Horrocks AR, Akalin M, Faroq AA. J Anal Appl Pyrolysis 1997;40e41: 511. [34] Nazarè S, Kandola B, Horrocks AR. J Fire Sci 2008;26:215.