Materials engineering for surface-confined flame retardancy

Materials Science and Engineering R 84 (2014) 1–20

Contents lists available at ScienceDirect

Materials Science and Engineering R journal homepage: www.elsevier.com/locate/mser

Materials engineering for surface-confined flame retardancy Giulio Malucelli, Federico Carosio, Jenny Alongi, Alberto Fina, Alberto Frache, Giovanni Camino * Department of Applied Science and Technology and INSTM Local Unit, Politecnico di Torino, Alessandria Campus, Viale Teresa Michel 5, 15121 Alessandria, Italy

A R T I C L E I N F O

Article history: Available online Keywords: Fire retardants Polymer combustion Intumescent coatings Layer by layer Sol–gel processes Cone calorimeter

A B S T R A C T

Polymer materials flammability represents a major limitation to their use and hence to the development of most polymer-based advanced technologies. Environmental and safety concerns are leading to progressive phasing out of versatile and effective halogen-based fire retardants which, so far, ensured a satisfactory polymer fire hazard control. Among the intensive efforts which are being made to develop new, environmentally safe, polymer fire protection approaches, the recognition of the paramount role played by the polymer surface during combustion and the exploitation of the new nanotechnologies developed for polymer surface engineering offer a promising perspective for polymer fire retardance. Indeed, heat transfer to the polymer and diffusion to the gas phase of polymer degradation combustible volatiles, which both fuel the combustion, occur across the polymer surface which characteristics regulate the polymer combustion process. It is shown that by engineering the polymer material surface by intumescent coatings or layer by layer nano-deposition or by oxidic nanostructures sol–gel synthesis, polymer combustion can be conveniently slowed down to extinguishment, complying fire safety rules of specific applications, through the creation of a surface barrier to heat and mass transfer across the polymer surface. ß 2014 Elsevier B.V. All rights reserved.

Contents 1. 2. 3.

4.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Intumescent coatings. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Layer by layer assemblies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Deposition methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Layer by layer architectures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Inorganic barrier coatings on fabrics . . . . . . . . . . . . . . . . . . . . 3.2.1. Coatings with intumescent composition deposited on fabrics 3.2.2. Hybrid coatings on fabrics . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.3. Other substrates: Foams and thin films . . . . . . . . . . . . . . . . . . 3.2.4. Sol–gel treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sol–gel chemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. Sol–gel derived coatings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. Flame retardant sol–gel treatments on textile substrates. . . . . . . . . . . . 4.3.

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . .

2 2 6 6 7 8 8 10 10 11 11 12 13

Abbreviations: APP, ammonium poly(phosphate); APTES, 3-aminopropyl triethoxysilane; BL, bilayer; bTESB, 1,4-bis(triethoxysilyl)benzene; bTESE, 1,2-bis(triethoxysilyl)ethane; DBTA, dibutiltindiacetate; DEMPhS, diethoxy(methyl)phenylsilane; DNA, deoxyribose nucleic acid; DPTES, diethylphosphatoethyltriethoxysilane; EG, expandable graphite; HRR, heat release rate; LbL, layer by layer; LOI, limiting oxygen index; MF, N,N,N0 ,N0 ,N00 ,N00 -hexakis-methoxymethyl-[1,3,5] triazine-2,4,6-triamine; MMT, montmorillonite; PA 6,6 GF, glass fibres reinforced polyamide 6,6 (PA6,6); PAN, poly(acrylonitrile); PC, polycarbonate; PET, polyester; PHRRp, eak of heat release rate; POSS, polyhedral oligomeric silsesquioxane; PP, polypropylene; PU, polyurethane; SEM, scanning electron microscopy; TBOS, tetrabuthylorthosilicate; TEES, triethoxy(ethyl)silane; TEOS, tetraethylorthosilicate; THR, total heat release; TMOSt, etramethylorthosilicate; TSR, total smoke release; TTI, time to ignition; XPS, X-ray photoelectron spectroscopy; ZrP, a-zirconium phosphate. * Corresponding author. Tel.: +39 0131229361; fax: +39 0131229399. E-mail address: [email protected] (G. Camino). http://dx.doi.org/10.1016/j.mser.2014.08.001 0927-796X/ß 2014 Elsevier B.V. All rights reserved.

G. Malucelli et al. / Materials Science and Engineering R 84 (2014) 1–20

2

4.3.1. 4.3.2. 4.3.3. 4.3.4.

5.

Inorganic structures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Phosphorus-doped silica structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Smoke suppressant coatings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Interpenetrated hybrid organic–inorganic network structures derived from matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Other substrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1. Introduction Since industrial production of polymers begun in the sixties, the necessity arose of modifying their properties to produce materials meeting different physical, mechanical, chemical requirements, depending on their application. The simplest technology then available was compounding the polymer with specific additives or fillers, independently of whether bulk properties such as stiffness or surface properties such as hydrophobicity were required. Now, about fifty years later, new technologies are available for materials surface engineering that allow to bestow on polymers desired surface properties, avoiding bulk addition drawbacks such as high loading owing to low effectiveness due to bulk dilution, detrimental effects on polymer mechanical properties, etc. Polymer flammability is a typical surface property, in which traditional bulk addition of so-called fire retardant additives, shows a quite poor effectiveness, in some cases requiring up to 60% loading, with impairing effects on polymer properties and giving rise to processing difficulties [1]. On the other hand, the polymer surface is the critical zone in the polymer combustion scenario schematized in Fig. 1 because, being the interface between gas and condensed phase, it controls mass and heat transfers, which are the processes responsible for flame fuelling. Indeed, heat reaching the polymer surface is transmitted to the polymer bulk, from which volatile products of thermal degradation diffuse towards the surface and the gas phase, feeding the flame. The polymer surface plays thus a key role in polymer ignition and combustion because it is its chemical and physical characteristics that affect the combustible volatiles flux towards the gas phase. The concern about negative impact on environment and on health by well established, versatile and effective halogenated fire retardants has driven the enforcement of new European regulations, progressively restricting their use and created the need for

Fig. 1. Polymer combustion cycle.

............................... ............................... ............................... sol–gel processes performed within ............................... ............................... ............................... ...............................

. . . . . 13 . . . . . 16 . . . . . 17 a polymer . . . . . 18 . . . . . 18 . . . . . 18 . . . . . 18

environmentally safe alternatives. This event is of great concern for the development of polymer materials because although fire exposure is an accidental event, flammability of polymers still represents a major limitation to their use and to the ensuing beneficial effect on industrial development. Thus, in a number of applications such as in the electrical–electronic, transportation, building and furniture sectors polymer materials can be used only if provided by a satisfactory fire retardant behaviour. One of the most valuable fire retardant strategy pursued by bulk addition, proved to be the production or accumulation of a thermally stable surface layer able to act as a barrier to mass and/or heat exchange. Such a layer is built during the early stage of combustion as a consequence of polymer surface layer decomposition, in the presence of different kinds of fire retardants, including inorganic nanoparticles. However, the time required for build-up of the surface barrier is straightforwardly connected to the development of the fire in the early stage, consequently adversely affecting the protective barrier action (Fig. 1). Here it is shown how the combination of advancements in polymer surface engineering and development of nanotechnologies, supplies an innovative environmentally friendly approach to fire retardance, based on providing polymer material products with a surface barrier, which either reradiates heat and/or slows down heat transmission and volatiles diffusion, without affecting the product bulk properties. To this purpose, technologies available include intumescent coatings, layer by layer assembly, sol–gel inorganic nanoparticles synthesis, that will be described below. By building the fire protection onto the original polymer surface, its effectiveness will be larger than in the case of protection created when polymer combustion is already started as it usually happens with bulk addition of traditional fire retardants. 2. Intumescent coatings Intumescent coatings technology has been widely applied since the 1970s in fire protection of different substrates, mainly on wood boards and steel structures, as an alternative to inorganic protections such as rigid mineral-based boards and fibre blankets. Despite metals and woods exhibit obviously different reaction and resistance to fire, the coating with an organic intumescent layer aims in both cases to the thermal protection of the substrate, in particular to extend the time to reach its critical temperature, i.e. the thermal decomposition temperature of wood or the temperature corresponding to an elastic modulus drop for steel. Paints and varnishes technology for these substrates is well established and is based on blowing of the coating layer to a thick insulating charred foam when heated above a certain temperature, as a consequence of fire exposure. The thermal resistance for a coating is the parameter quantifying the heat shield from a protective foam and it depends on several factors, including foam thickness, bubble size, defects and intrinsic heat transport characteristics of the foam material. Despite the thermal resistance cannot be measured directly, computational models have been proposed for its calculation [2].

G. Malucelli et al. / Materials Science and Engineering R 84 (2014) 1–20

3

Fig. 2. Scanning electron microscope picture for the interface between untreated PP/CaCO3 composite and the coating, showing no adhesion. Reprinted with permission from S. Duquesne, N. Renaut, P. Bardollet, C. Jama, M. Traisnel, R. Delobel. Fire Retardancy of Polypropylene Composites Using Intumescent Coatings. In: Fire and Polymers V ACS Symposium Series, vol. 1013. Chapter 12, pp. 192–204. Copyright (2009) American Chemical Society.

The basics of traditional used intumescent coatings can be found in classical reviews of intumescent coatings [3,4] in which the early history of intumescent coatings is presented. Organic intumescent formulations are typically based on a mixture of an inorganic acid or its thermal precursor such as ammonium polyphosphate (APP), a source of carbon such as a polyol, a ‘‘spumific’’ compound such as melamine, able to release gases to foam the charring coating and a binder resin. Since then, intumescent formulations have been continuously developed and optimized, in terms of chemical formulation (ratio between components, reactivity and catalysis) as well as for the optimization of physical properties of the char (viscosity during blowing, bubble nucleation, thermal and mechanical stability of the foamed char). Detailed reviews are available [5] covering intumescent coatings formulations and developments since the pioneer paper by Vandersall [3]. Here, the literature is reviewed concerning intumescent coatings for polymer materials fire protection, in analogy with the protection of wood boards, which was somewhat dismissed in the past in favour of bulk addition of intumescent fire retardant additives and is now resumed for some applications. Duquesne et al. first reported on the use of a commercial intumescent coating on polypropylene (PP) filled with talc or calcium carbonate [6]. Deposition of 100–150 mm thick layers was obtained by brush on specimens suitable for standard UL94 classification and glow wire test. Coating on untreated substrates evidenced very poor adhesion between the intumescent paint and

the PP composite, owing to the poor wettability of PP by the paint (Fig. 2). Adhesion of organic coatings onto polymers is indeed wellknown to be critical, especially on polyolefins, which exhibit a low surface tension and therefore poor affinity with typically polar binders used in paints (acrylics, polyurethanes, epoxies). To improve surface wettability, an oxygen/argon cold radiofrequency plasma treatment was used to generate oxidized species on the surface, evidenced by X-ray photoelectron spectroscopy (XPS) measurements and to increase its surface roughness. Plasma modification of the surface was demonstrated to provide good adhesion of the coating to the substrate (5B classification accordingly with ASTM D3359-02 tape test) and to affect the performance in flammability tests. Indeed, rapid self-extinguishing and no dripping, leading to V0 classification in UL94, were obtained only for well adhered coatings. Furthermore, plasma-treated and intumescent-coated PP/CaCO3 also passed glow wire flammability index test at 960 8C, therefore satisfying the two main flame retardancy requirements for applications in the electrical sector (Table 1). Later on, the same research group extended this approach to pristine polypropylene and polycarbonate (PC), using an acrylicbased transparent intumescent varnish or a polyvinyl alcohol based intumescent coating, applied by dipping to a thickness of 200 mm [7]. Instead of a plasma treatment, the simpler and inexpensive flame treatment was used to activate the polymer surfaces before intumescent depositions. Indeed, depending on the

Table 1 Performance of intumescent coated PP composites, untreated and plasma treated for adhesion. Data from Ref. [6]. Sample

PP + 15% talc PP + 30%CaCO3

Untreated substrates

Plasma-treated substrates

UL94,1.6 mm (ASTM D3801)

Glow wire 960 8C (IEC 60695-2)

Adhesion (ASTM D3359-02)

UL94,1.6 mm (ASTM D3801)

Glow wire 960 8C (IEC 60695-2)

Adhesion (ASTM D3359-02)

NC NC

Fail Fail

0B 0B

V0 V0

Fail Pass

5B 5B

G. Malucelli et al. / Materials Science and Engineering R 84 (2014) 1–20

4

Table 2 Flammability test results for pristine and intumescent coated PP and PC. Data from Ref. [7]. Polymer

LOI [vol.%]

UL94, 1.6 mm

Non coated PP PP/PVA coating PP/acrylic varnish Non coated PC PC/PVA coating PC/acrylic varnish

19 32 33 25 40 58

NR (not rated) V0 V0 V2 V0 V0

type of coating deposited and the number of flame passes, effective adhesion improvements to 4B or 5B ranking were reached. Results obtained for flammability tests on intumescent-coated PP and PC are reported in Table 2, showing a very significant increase in limiting oxygen index value and VO classification obtained for all the coated materials. Forced combustion test in cone calorimeter for intumescent coated PP and PC also showed very interesting improvements, both in terms of heat release rate (HRR) reduction and delay in the ignition time (Fig. 3). When using the acrylic varnish, no ignition was observed during testing at 50 kW/m2, for neither PP nor PC, confirming the high effectiveness of the coating in protecting the substrate from extensive and rapid thermal decomposition. A similar transparent varnish usually applied on wood was also used on glass fibre-reinforced polyamide 6,6 (PA6,6 GF) [8]. A 100 mm thick coating, applied on flame activated surface, was shown to provide LOI values above 50 and V0 rating at both 0.8 and 0.4 mm thickness of the substrate, while GWFI on 1 mm thick substrate was passed at 750 8C. GWFI test passed at 960 8C, typically required in electrical applications, was obtained increasing the coating thickness to 200 mm, which may however represent a significant add-on to the specimens, at least for thin substrates. To reduce the coating thickness while maintaining the same fire performances, the surface coating was combined with fire retardants in bulk, using a reduced concentration. The combination in PA6,6 GF of aluminium diethyl phosphinate (Exolit1OP1230 by Clariant) at 5 wt.% loading with an intumescent varnish (100 mm), was shown to reach UL94 V0 and GWFI 960 8C, thus improving GW compared to the intumescent coating alone. It is worthy to note that 5 wt.% of OP1230 is a quite low loading, given that as much as 23 wt.% is needed to obtain UL94 V0 and GWFI 960 8C when used alone in PA6,6 GF. In forced combustion, compared to PA6,6 GF, a dramatic reduction (about 80%) in the peak of HRR was observed for PA6,6 GF+coating and even slightly lower HRR was obtained with PA6,6 GF 5 wt.% OP1230+ coating. To investigate the mechanism of interaction between aluminium

phosphinate and the intumescent coating, the same authors studied in details the decomposition products obtained during interrupted cone calorimeter tests at different times. Based on the nuclear magnetic resonance characterization results of residues, the authors proposed that the intumescent coating may act as a barrier for the gaseous products produced from the degradation of aluminium phosphinate. Indeed, aluminophosphates formed during burning may be trapped inside the coating and act as stabilizers for the char, thus allowing a better protection of the underlying substrate [9]. Intumescent coatings were also proposed to protect foams; expandable graphite (EG) or APP in polyurethane (PU) were used to obtain an intumescent coating to be applied to rigid polyurethane foams [10]. The two products act in different ways: while APP is a conventional intumescent product when combined to polyurethane, EG is a filler which expands to produce highly expanded worm-like carbon structures. Both types of intumescent coatings were shown to limit the temperature increase of the foam, depending on the loading of EG or APP in the coating, as presented in Fig. 4. From the above discussed results, the effectiveness of intumescent coating is proven in both flammability and forced combustion tests. In particular, rapid self-extinguishing and no dripping, which are the main requirements for electrical and electronic applications, were achieved in flammability tests, with no need of high fire retardants loading. Although this may be seen as a new strategy for the fire retardancy of materials, limitations to the exploitation of this technology should be carefully considered. Substrate adhesion problems seem to be overcome by proper plasma or flame treatments, but the durability of the coating remains to be proven, in terms of both adhesion and retaining of fire protection properties during ageing, which may be a serious issue in case of exposure to moisture. Furthermore, for applications, in which the control of the parts dimensions is crucial (e.g. electrical connectors), the deposition of well-controlled thickness coating is required. Finally, in all the coated applications, additional processing time and costs have to be considered for the surface activation and deposition treatments, which are clearly important, especially on low added value products. The aforementioned limitations will likely limit the application of conventional intumescent coatings to large productions of polymer goods for the electrical and electronic sector. However, there are some other application fields where intumescent coatings are becoming very important. For instance, in structural polymer composites, the use of intumescent coatings is being developed for the thermal protection and thermo-mechanical stabilization of the composites.

Fig. 3. HRR plots for PP and intumescent-coated PP (left) as well as PC and coated PC (right). Reprinted with permissions from M. Jimenez, S. Duquesne, S. Bourbigot, Polym. Adv. Technol. 23 (2012) 130–135.

G. Malucelli et al. / Materials Science and Engineering R 84 (2014) 1–20

5

Fig. 4. Temperature vs. time under torch flame exposure for foams when coated with APP based formulations (a) and EG formulations (b), as a function of the loading of APP or EG in the intumescent coating. Reprinted with permission from S. Duquesne, M. Jimenez, S. Bourbigot. Fire retardancy and fire protection of materials using intumescent coatings—a versatile solution? In: Fire Retardancy of Polymers: New Strategies and Mechanisms. Royal Society of Chemistry, 2009, pp. 240–252.

Fig. 5. Normalized flexural strength after cone tests in different heat flux and exposure time for pristine composite (a) and intumescent coated composite (b). Reprinted with permission from E. Kandare, B. Kandola, P. Myler. Fire Safety J. 58 (2013) 112–120.

An early example of the use of several commercial intumescent coatings for the protection of glass-reinforced vinyl ester resin/ balsa wood core sandwich composites was reported by Sorathia et al. for military applications [11]. In this paper, small scale fire tests demonstrated a significant reduction of flame spread and smoke density when applied over composites substrates. Furthermore, flaming ignition of composites was prevented in the cone calorimeter test up to 50 kW/m2 imposed heat flux. More recently, different types of commercial intumescent coatings were applied to glass fibre epoxy/PET foam sandwich composites and tested in fire tunnels, evidencing a significant reduction of the substrate temperature [12], which is clearly correlated with the fire resistance of the composite panels. The post-heat flexural moduli of glass/epoxy laminates with/ without intumescent coatings (0.5 mm thick) after heat exposure in cone calorimeter tests were also assessed [13]. The time to ignition, time to peak HRR (PHRR) and time to reach 150 8C, i.e. the temperature corresponding to the flexural stiffness drop, were all extended by a factor of about 2 or more, as reported in Table 3. Generally, the degradation in the post-test flexural stiffness increased with applied heat fluxes and exposure time. However,

the degradation in the post-heat flexural moduli of glass/epoxy laminates with intumescent coatings was significantly delayed as compared to the pristine composite (Fig. 5). Thermo-mechanical modelling and experimental validation of fibreglass laminates protected with an intumescent coating were also recently reported [14,15]. More specifically, a thick (5 mm) layer of protective coating was applied to one surface of the laminate and the panel was tested for failure using a tensile load

Table 3 Cone calorimeter data for pristine and intumescent coated composite at variable heat fluxes. Data from Ref. [13]. Heat flux [kw/m2] Pristine composite

Intumescent coated composite

25 50 65 25 50 65

Ignition time [s]

Time to peak HRR [s]

Time to reach 150 8C [s]

72 52 17 204 81 32

105 95 75 400 195 140

64 54 45 203 122 124

6

G. Malucelli et al. / Materials Science and Engineering R 84 (2014) 1–20

Fig. 6. Schematic experimental setup (a) and results for pristine and intumescent coated composites (b) under compression load and 50 kW/m2 imposed heat flux. Reprinted with permission from E. Kandare, G.J. Griffin, S. Feih, A.G. Gibson, B.Y. Lattimer, A.P. Mouritz. Compos. Part A 43 (2012) 793–802.

(Fig. 6). Experimentally obtained time to failures for the coated composite was observed to be extended by a factor of 2 to 3 with respect to the pristine composites, evidencing for an effective thermal protection of the substrate. Significant research efforts are currently focused on coatings for structural composites field, aiming at maintaining the temperature of the composite well below the thermal decomposition and possibly below the glass transition temperature of the polymer, to maintain sufficient mechanical resistance during and after the fire event. Success in the development of suitable protective coating will have a strong impact on the use of polymer composites in applications where lightweight and fire resistant materials are needed, as primarily in trains, ships and aircrafts. 3. Layer by layer assemblies Nowadays, the layer-by-layer (LbL) assembly can be considered as a new standard for the molecularly-controlled fabrication of surface-confined nanostructured materials. Although the principles of this technique were published in 1966 by Iler [16], the potentialities of the LbL technique remained undisclosed until the rediscovery by Decher and coworkers, who developed a practical method for the assembly of oppositely charged polyelectrolytes [17]. Since then, the LbL technique has been widely exploited for the production of nanostructured materials, the complex functionality of which can be classified into one of the following categories:

driving force for the multilayer build-up. Up to now, the electrostatic interaction has been mostly investigated; however, this is not a prerequisite, as many other interactions are available and can be successfully exploited (like donor/acceptor [21,22], hydrogen bonding [23,24], covalent bonds [25,26], stereo-complex formation [27,28] or specific recognition [29,30]). In the particular field covered by this review, the LbL assembly has been performed through electrostatic interactions; the deposition process requires the alternate immersion of the substrate into oppositely-charged baths containing either a polyelectrolyte water solution or a nanoparticle water suspension. After each adsorption step, a total surface charge reversal can be achieved [31]; thus, by repeating the adsorption steps, it is possible to create a structure of positively and negatively charged layers piled up on the substrate surface. Fig. 7 schematizes the LbL process through electrostatic attractions: the positively charged moiety is a cationic polyelectrolyte and the negatively charged objects are inorganic ‘‘nanoplatlets’’. The substances suitable for the deposition can be chosen from a wide range of materials [32–36], cationic or anionic polyelectrolytes [37–40], metallic or oxidic colloids [41–44], layered silicates [45–47]. The whole process can be considered independent from the substrate size and topology as it is possible to deposit the selected chemicals on almost any solvent-accessible surface, playing, as an example, with micron-scale areas [48] up to surfaces of several square metres. 3.1. Deposition methods

(1) Specific modification of surface physical and chemical interactions: every object interacts with the environment via its surface; as a consequence, all the properties that depend on this interaction can be tailored by controlling the surface functionality. (2) Fabrication of surface-confined devices: the selected deposition sequence eventually defines the final coating architecture, as well as the device properties (some examples refer to the LbL deposition of membrane reactors [18] or light emitting diodes [19,20]). One of the key points of the LbL technique is certainly its simplicity; indeed, in a raw description, the LbL simply consists in an alternate adsorption of chemical species on a selected substrate. This process is based on the exploitation of one specific interaction that takes place between the selected species and acts as the

The LbL technique was invented as a solution dipping process which is highly affected by such deposition parameters as the chemistry of used polyelectrolyte [49], the molecular weight [50], the types of counterions [51], the ionic strength [52], the pH [53,54], the temperature [55], the adsorption time [56] and the drying step [57]. By controlling these parameters it is possible to fine tune the final structure as well as the final morphology of the deposited coating. Within the very first deposition steps, the substrate roughness, the nature and density of charged groups (as a consequence of surface activation e.g. plasma, corona . . .), their local mobility or simply the presence of impurities play a key role in the multilayer growth and strongly influence its final stability. After a few layers have been deposited (i.e. those needed to completely cover the

G. Malucelli et al. / Materials Science and Engineering R 84 (2014) 1–20

7

Fig. 7. Layer by layer assembly through electrostatic interaction (simplified schematization of the first two adsorption steps depicted as if it started on a negatively charged substrate).

substrate) and the transition from island growth to homogeneous growth has been reached, each polyanion adsorbs onto a polycation-covered surface and vice versa, thus completely losing the influence of the substrate. The dipping process represents, by itself, one of the major drawbacks for this technology since the time span between two consecutive deposition steps is of the order of minutes and, subsequently, the whole process might last up to several hours, thus limiting its industrial application. The spray-assisted layer by layer deposition has been proposed as a solution to overcome this drawback; indeed, spray can be more advantageous for its efficiency and feasibility at an industrial scale, as recently thoroughly reviewed by Schaaf and coworkers [58]. The first work by Schlenoff demonstrated the feasibility of the spray variation of the process [59]; later on, Decher exploited the spraying deposition conditions investigating the influence of various parameters like spraying time, polyelectrolyte concentration, and effect of film drying during multilayer construction [60]. It was demonstrated that the quality achievable using sprayed multilayers is equal or even superior to that reached by exploiting classical dipping processes. In addition, in order to further speed up the process, Decher and Schaaf proposed a deposition method, in which each solution is sprayed simultaneously onto the substrate [61]; this approach certainly represents the quickest one as it can reach in tens of seconds thicknesses normally achieved after hours of dipping. An automated process for the layer by layer spray deposition has been also reported in the literature and patented [62,63]. Another fast method for LbL deposition involve spin coating processes, the application of which has been demonstrated by Hong [64,65] and also by Wang [66]: nevertheless, this method is limited with respect to substrate size and planarity. 3.2. Layer by layer architectures With the possibility to prepare organic, inorganic and hybrid organic–inorganic coatings, the application fields of the LbL technique are growing year by year [67]. In particular, quite recently, the LbL technique has been exploited for the building-up of protective coatings directed towards the fire safety and fire protection fields, clearly proving

that this technique can be successfully adopted as a versatile tool to confer flame retardancy properties. From an overall point of view, the LbL represents a potential solution capable of satisfying the need for innovatory solutions of the flame retardant market. Indeed, as compared with the traditional treatments, the LbL shows some advantages: the easy incorporation and large amount of functional materials, the feasibility of the process under ambient conditions (room temperature, atmospheric pressure) and environmental-friendly characteristics since water is adopted as the LbL solvent, the concentrations of the baths are below 1 wt.% and can be recycled after use. The first work that presented the concept of surface protection achieved through layer by layer can be dated back to 2006 [68]. Since this work, the LbL assembly has been continuously developed in order to achieve flame retardant effects by different mechanisms that are depicted in Fig. 8 in comparison with bulk addition. The two categories shown in Fig. 8 can be explained as follows:  Inorganic barrier coatings: the coatings that fall in this category are mainly based on the exploitation of nanoparticles assembled in completely inorganic or hybrid organic–inorganic structures. The mechanism of flame retardancy can be considered derived from that of bulk nanocomposites where, during combustion, the accumulation on the surface of nanoparticles creates an inorganic barrier that both protects the underlying polymer and favours the char formation [69–72]. This effect can be mimicked by the layer by layer deposition if these nanoparticles are assembled directly on the surface.  Coatings with intumescent composition: the LbL coatings that fall in this category are intumescent or intumescent-like systems. Indeed, by granting the simultaneous presence of an acid source, a carbon source and a blowing agent it is possible to obtain a similar heat or flame reaction as it is well-known in the literature (i.e. the formation of a protective blown charred structure) [73,74]. The main difference when the concept is applied to the LbL refers to the scale of the intumescence feature that, being developed by a submicronic coating, can be considered at a microscopic scale. Due to the extreme versatility of the LbL technique, the aforementioned two categories can be considered as a raw

8

G. Malucelli et al. / Materials Science and Engineering R 84 (2014) 1–20

Fig. 8. Schematic representation of nanocomposite and macroscopic intumescence FR action in comparison with inorganic barrier and miniaturized intumescence LbL.

description of the main coating typologies found in the LbL deposition of flame retardant coatings; indeed, the reader can certainly find hybrid coatings that combine an intumescent composition with nanoparticles with an overall improvement in the thermal stability of the generated charred structure. Up to now, most of the LbL coatings have been deposited on substrates characterized by high surface to bulk ratio; this means that these substrates possess a high surface available for the coating with respect to the bulk that needs fire protection. Among this kind of substrates, textiles certainly represent the most used ones, recently followed by foams and thin films; for this reason, the following paragraphs will aim to provide the reader with a detailed presentation on textiles LbL protective coatings (thermal, intumescent and hybrid) and, subsequently, an insight into the recently developed substrates (foams and thin films). 3.2.1. Inorganic barrier coatings on fabrics The first studies that developed the LbL assembly of nanoparticle-based coatings were focused on the deposition of hybrid organic–inorganic structures made of anionic clay (montmorillonite or laponite) coupled to a cationic polymer (branched polyethylenimine) [75,76]. The results, obtained on cotton as a substrate, showed an increased thermal stability and improved flame retardant properties. Following this research path, the coating structure has been adjusted by removing the organic composition in order to deposit a completely inorganic coating; from the LbL point of view, the deposition of all-nanoparticle coatings brings extra constraints and thus can be considered more challenging with respect to the polyelectrolyte/nanoparticle counterparts [77,78]. Notwithstanding this, the all-nanoparticle coatings built-up with positive alumina-coated silica (10 nm) and negative silica nanoparticles (10 or 40 nm) were able to achieve good flame retardant effects when applied to cotton or polyester (PET) fabrics [79,80]. In particular, for this latter substrate, the coatings were also able to suppress the incandescent melt-dripping phenomenon during flammability tests as well as to greatly enhance the time to ignition during cone calorimetry tests. Interestingly, the limitations of such coatings have been related to the number of deposited layers; indeed, high numbers of deposited layers did not yield to an increased fire performance but an overall worsening due to the lack of the coating physical stability.

Silica/silica architectures have been deposited also by sprayassisted layer by layer deposition, focusing on the direct comparison of the properties achieved by dipping, vertical or horizontal spray on both cotton and PET fabrics [81,82]. The horizontal spray turned out to be the most efficient method for achieving a homogeneous surface deposition and, subsequently, for obtaining the most effective thermal barrier effect during combustion. Alumina nanoparticles of 50 nm have also been employed for the construction of an all-nanoparticle multifunctional coating on cotton; this has been performed by exploiting the amphoteric nature of alumina, the surface charge of which can be tuned by controlling the pH of the nanoparticle suspension (positive-acid, negative-basic) [83]. The assembled coatings improved not only the cotton flame retardancy but also its tensile strength and UVtransmittance. In an effort to improve the achieved flame retardant properties, the size of the adopted nanoparticles was reduced by using Polyhedral Oligomeric Silsesquioxane (POSS1) salts bearing either positive or negative charges (typically, POSS molecule has a diameter that ranges within 1–3 nm, including the organic substituents) [84]. However, the results obtained were not able to match those achieved by the silica/silica assemblies previously reported. a-Zirconium phosphate nanoplatelets have also been adopted in combination with octapropylammonium-functionalized POSS1 or alumina-coated silica nanoparticles [85]. These assemblies deposited on PET fabrics were able to affect the fabric combustion by increasing the time to ignition and reducing the peak of heat release; in addition, a reduction of the smoke production rate as well as of CO and CO2 yields was observed. Recently, the problem of coating durability has been addressed by LbL assemblies made of sodium montmorillonite and poly(Nbenzyloxycarbonyl-3,4-dihydroxyphenylalanine) deposited on polyimide fabrics [86]. These coatings improved the fabric flammability properties and were able to maintain these performances after 20 standard washing cycles. Table 4 summarizes the adopted coating constituents and the main results achieved by the LbL inorganic barrier coatings. 3.2.2. Coatings with intumescent composition deposited on fabrics After the first proof of concept demonstrated by the inorganic barrier coatings, the coating composition and fire proof action have been redirected towards the intumescent field.

G. Malucelli et al. / Materials Science and Engineering R 84 (2014) 1–20

9

Table 4 Coating constituents and main results of LbL deposited inorganic barrier coatings. Coating constituents

Main results

Reference

Branched polyethylenimine/laponite

Afterglow of 10BL (bilayer)-coated cotton fabrics occurred 10s earlier with respect to the uncoated fabrics

[75]

Branched polyethylenimine/sodium montmorillonite Branched polyethylenimine or alumina-coated silica/silica

Vertical flame spread final residue of 20BL-coated cotton fabrics after tests is coherent 10BL-coated fabrics exhibit a PHRR reduction of 20%, as assessed by using pyrolysiscombustion flow calorimetry (PCFC) 5BL-coated fabrics exhibit a PHRR reduction of 20% and a TTI (time to ignition) increase of 45%, assessed by cone calorimetry. By vertical flame test the same coatings reduces the burning time by 95% and eliminates melt dripping phenomena of PET Deposited by spray-assisted LbL on cotton, a significant increase of TTI (ca. +40%) and decrease of PHRR and TSR -total smoke release- (30 and 20%, respectively), as assessed by cone calorimetry Deposited by spray-assisted LbL on PET, increase of TTI (ca. +22%) and decrease of PHRR and TSR (34 and 30%, respectively), as well as the disappearance of melt dripping, as assessed by cone calorimetry LOI values increased from 18 to 21vol.% for untreated and 16BL-treated cotton, respectively Afterglow time is reduced, the fabric weave structure and shape of the individual fibres are preserved, formation of hollow fibres after combustion

[76] [79]

Alumina-coated silica/silica

Alumina nanoparticles Octa-3-ammoniumpropyl chloride POSS1/ Octakis(tetramethylammonium) pentacyclo[9.5.1.13.9,15,15.17,13]octasiloxane 1,3,5,7,9,11,13,15-octakis(cyloxide)hydrate POSS1 Polydiallyldimethylammonium chloride (PDAC) or Octa-3-ammoniumpropyl chloride POSS1 or alumina-coated silica nanoparticles/a-Zr phosphate nano-platelets Poly(N-benzyloxycarbonyl-3,4-dihydroxyphenylalanine)/ Sodium montmorillonite

PDAC-based assemblies increase TTI up to 86%, as assessed by cone calorimetry POSS-based assemblies yielded a 26% PHRR decrease, as assessed by cone calorimetry. Silica nanoparticles promote a significant reduction in the smoke production rate (25%), together with a strong CO and CO2 peak decrease (30 and 45%, respectively), as assessed by cone calorimetry Deposited on polyimide fabrics, reduced burning length in vertical flame tests, coating durability to 20 standard washing cycles

The first example of intumescent LbL coating was prepared by coupling a poly(allylamine) (carbon source and blowing agent) with sodium phosphates (acid source) and deposited on cotton [87]; when applied to cotton, the coating was found to impart selfextinguishment during vertical flame tests as well as nonignitability during cone calorimetry measurements. SEM investigation on the residues after flammability tests revealed the intumescent features of this LbL assembly. A coating with a similar composition has been recently deposited on polyamide 6,6 fabrics, achieving a decrease of the heat release rate when tested by pyrolysis combustion flow calorimeter [88]. Chitosan (carbon source) has been adopted in combination with ammonium polyphosphate (acid source and blowing agent) in order to build intumescent coatings on cotton-polyester blends [89]. The deposited coatings were capable to suppress the afterglow phenomenon and to leave a coherent residue after flammability tests as well as reduce the combustion kinetics. In order to obtain a coating capable of protecting both synthetic and natural fibres, a LbL architecture characterized by a four layer repetitive unit (namely quad-layer) consisting of poly(diallydimethylammonium chloride)/poly(acrylic acid)/poly(diallydimethylammonium chloride)/ammonium polyphosphate has been developed by our research group [90,91]. Such a coating was capable of greatly enhancing the char forming characteristics of both cotton and polyester fabrics when exposed to high temperatures (namely 300, 400 and 500 8C). Furthermore, a detailed study on the flammability and combustion properties of this LbL architecture revealed its ability in imparting self-extinguishment and in suppressing incandescent melt dripping, as well as in reducing heat related parameters, regardless of the coated substrate. Recently, the concept of intumescent-LbL has been applied to ramie fibres [92]. This fabric has been coated with polyethylenimine (carbon source) and ammonium polyphosphate (acid source and blowing agent) aiming to assess the effect of different polyelectrolyte concentrations on the composition of the coating and, subsequently, on the achieved flame retardant properties. Coatings deposited at high ammonium polyphosphate concentration were capable of achieving self-extinguishing behaviour during vertical flame tests.

[80]

[81]

[82]

[83] [84]

[85]

[86]

Ramie fabrics have been also coated by LbL coatings containing branched polyethylenimine as carbon source and blowing agent and poly(vinylphosphonic acid) as the acid source [93]. Transitionmetal-ions (cupric and zinc ions) have been added to the above coatings in order to enhance the coating flame retardant properties [94]; although coated fabrics did not reach the self-extinguishment, the presence of Cu/Zn ions in the coatings enhanced the overall flame retardant performance by increasing the final residue and reducing the combustion time. Finally, for what concerns ramie fabrics, the LbL deposition of amino-functionalized carbon nanotubes and ammonium polyphosphate has been recently attempted with promising results [95]. Lately, the synthesis and layer by layer assembly of oligoallylamine and phosphonated oligoallylamine have been proposed [96]. Preliminary tests showed that the two oligomers assembled together can form a stable residue in both inert and oxidative environments. The LbL has been also adopted for depositing coatings characterized by an intumescent composition combined with an environmentally-friendly nature. Few examples of these coatings are reported in the open literature; the first refers to the LbL assembly of chitosan as carbon source and phytic acid as acid source on cotton [97]. By changing the pH of each aqueous solution, it was possible to tune the coating composition in order to achieve the self-extinguishing behaviour in vertical flame tests. Our research group has recently proposed the use of deoxyribonucleic acid (DNA) in combination with chitosan in order to deliver novel and environmentally sustainable LbL coatings [98]. In particular, the molecule of DNA has been demonstrated to represent an all-in-one intumescent system [99]: indeed, the phosphate groups can act as acid source, the deoxyribose units can serve as a carbon source and the nitrogen-containing bases may release ammonia as a blowing agent. When applied to cotton, this biomacromolecule was capable of extinguishing the flame during horizontal flammability tests as well as to strongly reduce the combustion kinetics [100]. Table 5 summarizes the coating constituents and the main results achieved by the LbL intumescent coatings described in this section.

10

G. Malucelli et al. / Materials Science and Engineering R 84 (2014) 1–20

Table 5 Coating constituents and main results of LbL deposited intumescent coatings. Coating constituents

Main results

References

Poly(allylamine)/poly(sodium phosphate)

On cotton, self-extinguishing behaviour during vertical flame tests is achieved, 10BLs induce a significant decrease of the THR (total heat release) and PHRR (80 and 60%, respectively), as assessed by PCFC On polyamide 6,6, a decreased PHRR (36%) after 40BLs, as assessed by PCFC Suppression of the afterglow phenomenon for cotton-rich (70%)-polyester blends; 20BLs are able to reduce the THR (22%) and PHRR (25%), as assessed by cone calorimetry Deposited on both natural and synthetic fibres. Char formation enhanced at different temperatures (namely, 300, 400 and 500 8C). Self-extinguishing behaviour during horizontal flame tests, melt dripping of synthetic fibres suppressed. PHRR and THR remarkably reduced for both natural and synthetic fibres when tested by cone calorimetry under different irradiative heat fluxes (25, 35, 50 kW/m2) Self-extinguishment during vertical flame tests achieved after 20BL deposition at high ammonium polyphosphate concentration Reduced after-flame time and increased final residue Reduced production of CO and CO2 for Cu ions containing coatings

[87]

Chitosan/ammonium polyphosphate Poly(diallydimethylammonium chloride)/ poly(acrylic acid)/ammonium polyphosphate

Polyethylenimine/Ammonium polyphosphate Branched polyethylenimine/poly(vinylphosphonic acid) Branched polyethylenimine and cupric or zinc ions/poly (vinylphosphonic acid) Amino functionalized carbon nanotubes/ammonium polyphosphate Chitosan/phytic acid

Chitosan/deoxyribonucleic acid

20BLs were capable of increasing the final residue during vertical flame test as well as reduce the PHRR (36%), as assessed by PCFC Renewable resources composition, 30BLs applied to cotton were able to reach selfextinguishing behaviour during vertical flame tests and to reduce PHRR of 50%, as assessed by PCFC Renewable resources composition, self-extinguishing behaviour during horizontal flame tests, LOI increased from 18% up to 24%, PHRR reduced by 40%, as assessed by cone calorimetry

3.2.3. Hybrid coatings on fabrics As mentioned in the previous paragraph, it is possible to use the LbL assembly in order to deposit coatings that combine the nanoparticle barrier effect and the intumescent features into a single hybrid coating. This strategy aims at enhancing the protection provided by the charred carbon structures produced by the intumescent composition, preventing the destructive effects of both flame exposure and mechanical stresses due to air drafts. The first attempt has been made by depositing complex architectures made of chitosan, ammonium polyphosphate and silica nanoparticles bearing positive or negative charge on cotton-PET blends [101]. Such coating constituents have been selected on the basis of the results achieved by silica/silica coatings and chitosan/ ammonium polyphosphate coatings described in the previous paragraphs. The flame retardant properties have been found to be highly influenced by the morphology and physical stability of the deposited coatings; indeed, only those that yielded to the formation of a continuous and homogeneous coating were capable of improving both the flammability and combustion properties. Another attempt has been made on cotton with assemblies of poly(acrylic acid) coupled with amino-functionalized montmorillonite nanoplatelets (MMT) [102]. Compared with the untreated fabrics, the coated substrates showed reduced combustion kinetics; in addition, SEM images performed on the residue after combustion tests revealed the formation of a blown charred structure containing nanoplatelets. Recently, graphene oxide nanosheets have been combined with a synthesized intumescent flame retardant-polyacrylamide [103]. The coating, assembled through hydrogen bonding interactions, was not able to reach the self-extinguishment, although the combustion time was reduced and the final residues increased. By

[88] [89] [90,91]

[92] [93] [94] [95] [96]

[97]

cone calorimetry, an increase of time to ignition and a reduction of combustion kinetics have been observed. Table 6 summarizes the coating constituents and the main results achieved by the LbL assembled hybrid coatings. 3.2.4. Other substrates: Foams and thin films Along with fabrics, open cell foams are substrates characterized by high surface to bulk ratio and thus they represent a valid substrate to be coated by the layer by layer assembly. Indeed, due to the open cell nature of foams, the solution or suspension adopted in the LbL process can penetrate inside the threedimensional structure of the foam and coat every surface available. Furthermore, from the fire-safety point of view, foams may represent a severe treat as they burn very quickly, releasing toxic gases and exhibiting melt dripping phenomena that can easily spread the fire to other ignitable materials [104]. The attempt to deposit a fire protection coating on foams has been done by employing polyethyleneimine stabilized carbon nanofibres (cationic layer) coupled with poly(acrylic acid) (anionic layer) as coating constituents. The LbL coating completely covered both the external and internal surfaces of the foam achieving a consistent reduction in the combustion kinetics as assessed by cone calorimetry [105]. Amino-functionalized multi-walled carbon nanotubes have been also employed in unconventional tri-layer architecture along with polyacrylic acid and branched-polyethylenimine; the carbon nanotube network generated within the polyelectrolyte coating was found to be responsible of a 35% heat release rate reduction [106]. Subsequently, sodium montmorillonite has been combined with chitosan in order to deposit environmental-friendly coatings

Table 6 Coating constituent and main results of LbL deposited hybrid coatings. Coating constituents

Main results

Reference

Chitosan/ammonium polyphosphate/alumina-coated silica/silica Synthesized flame retardant: poly(acrylic acid)/aminofunctionalized sodium montmorillonite Synthesized flame retardant polyacrylamide/graphene oxide

Suppression of the afterglow phenomenon and increase of the final residue in vertical flame tests. TTI increase (+60%) and PHRR reduction (22%), as assessed by cone calorimetry 20BLs favour an increase of TTI (+40%) and a reduction of THR and PHRR (50 and 18%, respectively), as assessed by cone calorimetry 20BLs increase TTI (+56%) and lower PHRR (50%), as assessed by cone calorimetry

[101] [102] [103]

G. Malucelli et al. / Materials Science and Engineering R 84 (2014) 1–20

with renewable resources composition [107]; when subjected to a butane flame torch, the 10BL assembly was capable of suppressing the foam melt dripping. A consistent reduction in combustion kinetics (52% PHRR) was also achieved. The use of sodium montmorillonite has been then widely exploited in LbL assembly containing different kind of polyelectrolytes; indeed, these nanoplatelets have been employed either in a tri-layer structure with anionic polyacrylic acid and cationic branched-polyethylenimine or bilayer structures with branchedpolyethylenimine [108–110]. The results reported for the above structures showed a reduction of average heat release rate (71%) during cone calorimetry combustion tests and were also confirmed by real scale furniture mockups tests. Sodium montmorillonite has been also employed in hybrid intumescent compositions by depositing coatings containing polyphosphates; this hybrid architecture completely suppressed the melt dripping of the substrates during flammability tests and achieved a 54% reduction in the peak of heat release rate [111]. A different flame retardant mechanism has been addressed by LbL coatings containing positively charged chitosan and anionic poly(vinyl sulfonic acid sodium salt) [112]: these coatings, that can be considered as the first LbL architecture without nanoparticles capable of enhancing the polyurethane flame retardancy, achieved great performances thanks to the release of gases exhibiting flame diluting effects during combustion. Recently our research group has developed a LbL architecture containing poly(acrylic acid), chitosan and poly(phosphoric acid); the deposited coatings were able to adapt to flame or heat exposure and to evolve into thermally stable carbon-based structures. When tested by cone calorimetry under different irradiative heat fluxes (from 35 up to 75 kW/m2), a 55% reduction in heat release rate was achieved regardless to the adopted heat flux [113]. Furthermore, when subjected to a flame torch penetration test (Tflame  1300 8C), the LbL-coated foam was capable of maintaining its three-dimensional structure, thus successfully insulating the unexposed side which detected temperature was below 100 8C after two flame torch applications. Up to now, most of the LbL coatings for fire protection have been deposited on substrates characterized by high surface to bulk ratios, such as textiles and open cell foams. The use of substrates, such as bulk plastic films (100 to 1000 mm thick), represents a challenging issue (according to the few articles published in the literature), as the flame retardant properties imparted by the LbL coatings have to be enhanced in order to protect an increased amount of material. However, this research field is potentially of great interest, as bulk plastic films are often employed in buildings and transportation and represent a potential fire source. The first attempt has been performed by depositing poly(allylamine)/montmorillonite coatings on 1 mm polylactic acid films [114]. After the deposition, the coating was subjected to a post-treatment in order to favour the diffusion of sodium polyphosphate within the layers; the resulting assembly, the flame retardant action of which combined both thermal barrier and intumescent mechanism, yielded to an increased time to ignition and reduced the peak of heat release rate (37%) during cone calorimetry tests. A similar coating has been developed for polyamide 6 films (500 mm), without including the polyphosphate post diffusional treatment, achieving an efficient decrease of the combustion kinetics during cone calorimetry tests (60% reduction of PHRR) [115]. Recently, a thermal barrier coating made of silica nanoparticles has been successfully assembled for the surface protection of polycarbonate thin and thick films: 200 or 1000 mm. On thinner substrates, the deposited LbL architecture suppressed the

11

incandescent melt dripping during flammability tests and achieved a reduction in both heat release rate peak and total heat release (20 and 30%, respectively) during cone calorimetry. Conversely, in the case of 1000 mm thick samples, the substrate thickness was found to be detrimental as far as both the flammability and combustion behaviour are considered [116]. 4. Sol–gel treatments 4.1. Sol–gel chemistry Nowadays, the most widely used technique for synthesizing bulk metal oxides is the ceramic method, which is based on direct reactions of powder mixtures [117]. These reactions are controlled by the diffusion of atomic or ionic species through reactants and products and require very high temperatures [118]. Although the harsh reaction conditions lead to thermodynamically stable phases only, preventing the formation of metastable solids, they allow the synthesis of a large number of new solid compounds, enabling the development of structure-properties relationships. However, such method is a rather crude approach; therefore, it is no surprising that worthy alternatives such as liquid-phase routes for the sizeand shape-controlled synthesis of nanoparticles have been further developed. Among the various soft-chemistry routes, sol–gel reactions have proven to be particularly successful in the preparation of metal oxides (e.g., ceramics, glasses, films and fibres) [119]. Indeed, such technique has shown its exceptional potential regarding the synthesis of new materials with a high degree of homogeneity at molecular level and with extraordinary physical and chemical features. More specifically, sol–gel is a versatile synthetic route based on a two-step reaction (hydrolysis and condensation), starting from (semi)metal alkoxides (usually tetraethoxysilane, tetramethoxysilane, titanium tetraisopropoxide, aluminiumisopropoxide,. . .) that leads to the formation of completely inorganic or hybrid organic–inorganic chemical structure at or near room temperature. In order to reach the highest performances by this approach, several process parameters have to be considered, like nature of (semi)metal atom and alkyl/alkoxide groups, structure of the (semi)metal alkoxide, water/alkoxide ratio, pH (acidic or basic conditions), temperature, reaction time and presence of co-solvents. All these parameters determine the structure/morphology of the resulting oxidic networks [120]. A schematic representation of the process is reported in Fig. 9. Despite great efforts of industrial and academic researchers, the number of oxidic nanoparticles obtained by sol–gel chemistry is still rather small compared to the variety of compounds obtained via powder routes. Indeed, aqueous sol–gel chemistry is quite complex, on one hand due to the high reactivity of the metal oxide precursors towards water, and to the double role of water as reactant and solvent. On the other hand, a high number of reaction parameters have to be strictly controlled (hydrolysis and condensation rate of the metal oxide precursors, pH, temperature, method of mixing, rate of oxidation, the nature and concentration of anions, . . .) [119]. In addition, through sol–gel reactions, only amorphous structures can be synthesized, and thus a postannealing step able to induce their crystallization is required for some applications. In the preparation of bulk metal oxides these limitations play only a minor role, whereas, in the case of nanoparticle synthesis, they can constitute a major issue. Non-aqueous (or non-hydrolytic) sol–gel processes in organic solvents, generally with the complete exclusion of water, are able to overcome some of the major limitations of aqueous systems, and thus represent a powerful alternative [121–124]. The advantages are a direct consequence of the multi-role of the organic components in the reaction system (e.g., solvent, organic ligand

12

G. Malucelli et al. / Materials Science and Engineering R 84 (2014) 1–20

Fig. 9. Schematic representation of the sol–gel process.

of the precursor molecule, surfactants, or in situ formed organic condensation products). Indeed, they can act as oxygen-supplier for the oxide formation, strongly affecting the resulting particle size and shape; as a consequence, non-aqueous sol–gel processes generally yield metal oxide nanoparticles with uniform, yet complex crystal morphologies, crystallite sizes in the range of a few nanometres, and good dispersibility in organic solvents. 4.2. Sol–gel derived coatings Generally, during sol–gel processes, nanosols consisting of colloidal solutions of nanometre sized metal oxide particles are formed, regardless of the reaction ambient (aqueous or organic solvents) [125]. Due to the very high surface area of such particles, the nanosols turn out to be metastable, thus, during the formation of a coating, some particles aggregate due to the evaporation of the solvent, easily forming a three-dimensional network. The main steps for preparing sol–gel derived coatings are depicted in Fig. 10. Nanosol particles exhibit diameters in the range from a few nanometres up to 100 nm, while coatings formed by nanosols can reach a thickness of up to several hundred nanometres [117]. Therefore, the length of a nanosol coating can cover a broad range of the structural elements, starting from molecules up to

three-dimensional, large-scaled objects such as fibres forming a textile. In relation to the process parameters, the inorganic metal oxide forms mainly amorphous networks after moderate heat treatment (so-called xerogels); on the other hand, if a treatment at high temperatures (>500 8C) is carried out, the resulting networks give increasingly crystalline structures. The basic nanosols can be modified in a wide range, leading to numerous new functionalities that can be applied to various surfaces. Thus, this type of coating is a suitable tool for modifying a large number of materials, such as glass, paper, synthetic polymers, wood, metal and textiles [125]. In conclusion, sol–gel technology promises the possibility to tailor surface properties to a certain extent, and to combine different functionalities in a single material. Without a doubt, coatings derived from sol–gel processes have found a wide number of uses in the textile industry, where the application of sols can be carried out with techniques commonly employed. As an example, finishing of textiles can be performed in a simple dip or padding process followed by a thermal treatment in an oven. Furthermore, nanosols can be modified in many ways to achieve new or additional functional properties. By adding new properties to a sol, the sol coated surface will be suitably provided with the corresponding functionalities. The modifications either chemical or physical can be carried out by adding particular

Fig. 10. Main steps for preparing sol–gel derived coatings.

G. Malucelli et al. / Materials Science and Engineering R 84 (2014) 1–20

compounds, either to the precursors before hydrolysis, or to the pre-fabricated nanosols, respectively. The chemical modification is performed with additives which are able to form covalent bonds with or within the metal oxide particles during the preparation process [126], or could be the result of a co-condensation of different types of metal alkoxides [125]. By this way, metal oxide compounds composed of different metals in a variable ratio can be produced. The most frequently used metals are silicon, aluminium, titanium, zinc and zirconium [127], and the resulting particles can consist of mixed oxides or core–shell structures might be formed, where core and shell are based on different oxide compositions [128]. Another approach for the chemical modification is the co-hydrolysis and co-condensation of trialkoxysilanes R0 -Si(OR)3 that are modified with an organic R0 . These alkoxysilanes can be introduced covalently into an inorganic metal oxide network. The organic group can act simultaneously as a network modifier that changes properties of the resulting coatings such as the porosity [129]. Hydrophobic or even oleophobic surfaces can be designed by introducing alkyl- or fluoroalkyl groups into the coating material [130,131]. If the organically modified alkoxysilanes used are modified with functional groups based on epoxy or methacrylic compounds these groups can also be used to achieve cross-linking of organic domains parallel to the development of the inorganic network [132,133]. Due to the mentioned combination of inorganic and organic functionalities the resulting materials are referred to hybrid inorganic–organic polymers. The modification of nanosols can also be carried out with additives that are homogeneously incorporated and immobilized/ entrapped into the metal oxide matrix without forming covalent bonds (physical modification). These additives are usually larger molecules such as polymers, pigments, dyestuffs, active substances or biomolecules [134–136]. The incorporation/entrapping can occur by adding the additives either before, or after hydrolysis of the precursors. Both routes lead to composite structures and immobilization of additives within the inorganic matrix is very efficient since it is assumed that encapsulation actually occurs during the formation of the network. A huge number of additives can be employed leading to manifold functions: some examples are reported in Table 7, with the exclusion of the flame retardancy that will be detailed in the following paragraph and in the following tables. 4.3. Flame retardant sol–gel treatments on textile substrates In the field of polymeric materials, the use of sol–gel processes for obtaining silica nanoparticles able to reduce the flammability of some bulk polymers is well documented in the literature. Indeed, several papers have studied the possibility to reduce the flammability of epoxy [194–198] and phenolic resins [199,200], polymethylmethacrylates [201] and polyesters [202], by exploiting silica phases derived from sol–gel processes. Furthermore, fumed or fused silica have shown less efficiency than that of sol– gel derived silica in the flame retardancy of polypropylene and polyethylene oxide, as reported by Kashiwagi et al. [203]. Although sol–gel processes have been known from the 1950s, their application in the flame retardancy of textiles is very recent [204,205]: in the last years, our group has clearly demonstrated that sol–gel derived hybrid architectures are able to protect the polymer surface exerting a thermal shielding effect, thus improving the flame retardancy of the treated fabrics. Indeed, by absorbing the heat from the surrounding atmosphere, these architectures can create a physical barrier to oxygen, heat and mass transfer, hindering the evolution of volatile species that fuel the flame and, at the same time, favouring the formation of a surface protective thermally stable inorganic residue [206,207].

13

Table 7 Some examples of nanosol modification for conferring new functionalities to textiles. Functionality

Additive

Reference

Anti-wrinkle effect Antistatic properties

Dimethylolethylene urea Aminosilane SnO2 Organic biocides TiO2 (anatase) Silver Copper Zinc oxide Benzoic acid Fragrances Dopamine Organic dye stuff Infrared dye Oxidase Lipase Papain Alkyltrialkoxysilanes Hydrophobic polysiloxanes Urea/polysiloxane Zirconium/aluminium oxide Epoxyalkyltrialkoxysilane Methyltriethoxysilane Aluminium oxide Zirconium oxide Titania oxide Epoxysilanes and aminosilanes Fe3O4 Fluorocarbon polymers Fluorinated alkyltrialkoxysilane TiO2 (anatase) Organic UV-absorber Organic UV-absorber/TiO2 TiO2 Zinc oxide Methyl red Barium sulphate Fe3O4

[137] [138] [139] [140,141] [142,143] [144,145] [146] [147] [148] [149] [150] [151–158] [159] [134] [135] [136] [160,161] [130,131] [162] [163] [164,165] [166] [167,168] [168] [168] [169,170] [171] [172–174] [175] [176–182] [183] [184–186] [187] [188,189] [190,191] [192] [193]

Antimicrobial properties

Controlled release

Dyeing

Enzyme immobilization Hydrophobicity

Heat resistance Improved abrasion resistance

Luminescence Magnetism Oleophobicity Photocatalytic properties UV protection

Sensor features Shielding of X/a-rays Super-paramagnetism

Furthermore, the flame retardancy/suppression promoted by the sol-gel derived coatings is really effective only when they operate in synergistic or joint effects with flame retardants, such as phosphorus and/or nitrogen-containing compounds. In parallel, the evolution of the sol–gel strategy has led to the setup of dualcure processes, which can be exploited for preparing hybrid organic–inorganic protective coatings through a photo-polymerization reaction followed by a thermal treatment for promoting the formation of silica phases. These developments will be described in the following paragraphs on the basis of the type of synthesized architecture. 4.3.1. Inorganic structures The simplest architectures derived from hydrolysis and condensation reactions of a (semi)metal alkoxide are completely inorganic and the most common ones consist of pure silica particles or coatings, derived from tetraethylorthosilicate (TEOS). This latter is usually the most employed sol–gel precursor. Table 8 summarizes the results published in the literature about the synthesis of inorganic coatings for conferring flame retardancy properties to fabrics. Precursor with number and type of hydrolysable groups, the resulting synthesized oxidic phase, textile and corresponding reference paper are detailed. Hribernik and co-workers [208] have deposited a silica coating (350 nm thickness) starting from TEOS in order to reduce the flammability of regenerated cellulose fibres (i.e. viscose). With respect to untreated fibres, the temperatures of their first degradation step and of flame combustion of volatile products were found increased by 20 8C, while the temperature rise of

14

G. Malucelli et al. / Materials Science and Engineering R 84 (2014) 1–20

Table 8 Inorganic architectures designed for conferring flame retardancy to textiles employing sol–gel treatments. Precursor

Oxidic phase

Textile

Reference

TEOS (4 ethoxy groups) TEOS (4 ethoxy groups) TEOS (4 ethoxy groups) TMOS (4 methoxy groups) TEOS vs. TMOS vs. TBOS (ethoxy vs. methoxy vs. buthoxy groups) TEOS vs. DEMPhS, APTES, TEES, bTESB, bTESE (4 vs. 2, 3 and 6 ethoxy groups) Tetraethylorthosilicate (4 ethoxy groups) Tetraethylorthotitanate (4 ethoxy groups) Tetraethylorthozirconate (4 ethoxy groups) Aluminium isopropylate (3 isopropoxy groups)

Silica Silica Silica Silica Silica Silica Silica Titania Zirconia alumina

Viscose (regenerated cellulose) Cotton, polyester, cotton-polyester blends Cotton Cotton Cotton Cotton Cotton

[208] [209] [210] [211,212] [213] [213] [167,168]

glowing combustion of the residue was higher than 40 8C. Similar results were collected by our group applying the same strategy to cotton, polyester and their blends [209]: indeed, the formation of a continuous silica film deposited on the fibres (regardless of the type) exerted a protective role towards their thermal degradation in air. In addition, these films reduced the PHRR during combustion, as assessed by cone calorimetry: as an example, under a 35 kW/m2 heat flux, the sol–gel-treated 35/65 cotton/polyester blend showed a remarkable TTI increase (up to 98%) and a strong decrease of PHRR (around 34%) with respect to the untreated fabrics. In such study, dibutyltindiacetate (DBTA) was used as a condensation catalyst in order to promote the oxidic phase formation. Indeed, in comparison with other smooth substrates like glass or polymeric films, the presence of a texture in the fabric can affect the kinetics of silica formation, and thus determine the resulting performances of the deposited coating. Analogously, Colleoni et al. [210] have found that the presence of DBTA favours the coating formation, and as a consequence a higher thermal stability of fabrics treated by sol–gel if compared with the untreated cotton. The scarce solubility of TEOS in water and its high reactivity have to be taken into account for its use in sol–gel technology. The former aspect can easily be addressed by employing some co-solvents such as ethanol or other alcohols, even though this limits a possible industrial exploitation, whereas the latter aspect can strongly affect the formation of the resulting coating. Indeed, the higher the hydrolysis/condensation rate, the lower is the conversion degree of TEOS on fabric surface. This finding does not agree with the wellestablished know-how about sol–gel process [120], but in such case the presence of fabric texture affects the diffusion process of precursor. It should also be taken into account that a part of precursor may remain entrapped within the warp and weft of the fabric, and thus it becomes less available to give hydrolysis/ condensation reactions. Alternatively to TEOS, other silane precursors or other oxidic phases as well as different process aspects directly connected with the resulting morphology of the coating deposited on the fabrics have been investigated (Table 8). The following aspects have been taken in consideration:  Process parameters in silica formation (namely, precursor to water molar ratio, temperature and time of the thermal treatment [211], moisture [212]),  Chemical structure of the silane precursors (namely, chain length and number of hydrolysable groups) [213],  Other oxidic phases: precursor type [167,168]. 4.3.1.1. Process parameters in silica formation. Owing to above limitations, TEOS was replaced with the analogous silane having 4 methoxy groups, TMOS (tetramethylorthosilicate). In order to reach the best fire performances by using this ‘‘new’’ precursor, TMOS to water molar ratio, temperature and time of the thermal treatment have been optimized. Cone calorimetry tests showed that the highest achievements for cotton (+56% TTI and 15%

PHRR) are obtained when the sol–gel process is carried out at 80 8C for 15 h, using a 1:1 TMOS:H2O molar ratio. This finding strictly depends on the level of silica distribution and dispersion on and within the fabrics. Indeed, the more homogeneous the morphology of the coating, the better are the flame retardancy performances, as observed correlating the results collected by cone calorimeter with those of 29Si solid state nuclear magnetic resonance [211]. The results collected in the literature show that silica coatings derived from sol–gel reactions are able to protect the polymer surface exerting a thermal shielding effect, improving the flame retardancy of the treated fabrics. Hence, the coating acting as a thermal insulator is capable to shift towards higher values the temperature at which degradation starts, favouring the char formation of cellulose. In this contest, a key factor that should be taken into account is the role of moisture and heat transfer coupled with moisture transfer in cotton under a simulated fire. These phenomena have been studied by our group employing an optimized procedure based on the use of a cone calorimeter as a heating source [214]. These data are compared with those collected following the ISO6942 standard, specific for the heat transmission in the case of protective garments. The presence of the inorganic coating turned out to significantly influence both heat and moisture transfer within the sol–gel treated cotton fabrics. Indeed, their thermal conductivity was found to be strongly affected by the porosity of coatings as well as method of conditioning of the samples, and consequently by the moisture uptake [212]. 4.3.1.2. Chemical structure of the silane precursors. Varying the chain length as well as the number of hydrolysable groups of the silane precursor, a different combustion behaviour of the sol–gel treated cotton fabrics has been detected [213]. More specifically, the flame retardant properties of cotton treated with TMOS, TEOS and TBOS (tetrabuthylorthosilicate), having four methoxy, ethoxy and buthoxy groups, respectively, have been investigated and compared. Vertical flame spread tests showed that, when a propane flame was directly applied to the cotton fabrics for 5 s, the presence of even a low amount of silica was enough to slow down the burning rate (Table 9): indeed, two flame applications were necessary to ignite the sol–gel treated specimens. The final residue significantly increased from 10 up to 48 wt.% upon cotton treatment with TMOS whereas the obtained residues for TEOS and TBOS treatments were 35 and 33 wt.%, respectively. These results have shown that the shorter is the precursor chain length, the lower is the cotton flammability. The same trend has been demonstrated for the resistance towards an irradiating heat flux of 35 kW/m2, as assessed by cone calorimetry. Indeed, cotton treated with TMOS exhibits the lowest values of PHRR and THR, as evidenced in Table 9. The hydrolysable group number can strongly affect the resulting performances of silica coatings deposited on cotton fabrics. For this reason, the behaviour of TEOS was compared with that of diethoxy(methyl)phenylsilane (DEMPhS), 3-aminopropyl

G. Malucelli et al. / Materials Science and Engineering R 84 (2014) 1–20

15

Table 9 Combustion tests of cotton fabrics treated with TMOS, TEOS and TBOS sol–gel precursors. Data from Ref. [213]. Vertical flame spread tests

Sample

Cotton Cotton_TMOS (4 methoxy groups) Cotton_TEOS (4 ethoxy groups) Cotton_TBOS (4 buthoxy groups)

Cone calorimeter tests

Number of flame applications

Burning rate [mm/s]

Residue [%]

TTI [s]

PHRR [kW/m2]

THR [MJ/m2]

1 2 2 2

2.50 1.45 1.75 1.75

10 48 35 33

20 30 30 28

91 82 85 90

2.8 1.8 2.3 2.2

Table 10 Vertical flame spread tests on cotton fabrics treated with silica precursors having a different number of hydrolysable groups. Data from Ref. [213]. Sample

Number of flame applications

Burning rate [mm/s]

Residue [%]

Cotton Cotton_DEMPhS (2 ethoxy groups) Cotton_TEES (3 ethoxy groups) Cotton_APTES (3 ethoxy groups) Cotton_TEOS (4 ethoxy groups) Cotton_bTESB (6 ethoxy groups) Cotton_bTESE (6 ethoxy groups)

1 2 2 2 2 10 2

2.50 1.81 1.81 2.00 1.75 1.70 1.45

10 30 31 20 35 96 20

triethoxysilane (APTES), triethoxy(ethyl)silane (TEES), 1,4-bis(triethoxysilyl)benzene (bTESB) and 1,2-bis(triethoxysilyl)ethane (bTESE), for which the number of the ethoxy groups ranges is in between 2 and 6 (Table 10). On the basis of the vertical flame spread tests, it was possible to conclude that [213]: (i) The alkoxysilanes with a low number of hydrolysable groups (i.e. 2 or 3) behave like TEOS, regardless of their large smoke release and the nature/morphology of their final residues (very thin and not dense/compact products). (ii) The replacement of an amino group (APTES) with an alkyl chain (TEES) promotes the formation of a compact and thicker residue. (iii) bTESB and bTESE, which bear the highest number of alkoxy functionalities, significantly change the flammability of the fabrics: in particular, bTESB-treated fabrics do not burn even after ten applications of 5 s flame. 4.3.1.3. Other oxidic phases: Precursor type. Alternatively to silica, other oxidic phases capable to enhance the flame retardancy of cotton can be obtained by sol–gel processes, as recently reported by our group [168]. Indeed, the fire performances of titania, zirconia and alumina (from tetraethylorthotitanate, tetraethylorthozirconate or aluminium isopropylate) were compared with those of silica when applied to cotton fabrics (Table 11). The collected data by vertical flame spread and cone calorimeter tests can be summarized as follows: (i) From the vertical flame spread tests, all coatings are responsible of a remarkable decrease of cotton burning rate and they

have an analogous behaviour when a flame is applied; indeed, the coating acting as thermal insulator protects cotton from the flame application. The only difference consists in a lower residue left by zirconia at the end of the tests, in comparison to the other coatings. (ii) From the cone calorimeter tests, the best flame retardancy performances were achieved by depositing a silica coating on cotton fibres (a TTI increase of 56% and a PHRR decrease of 20%). Regardless of the coating type, in all cases, a significant reduction of TSR has been found. These results are very important because they show that these ceramic architectures can act also as smoke suppressants coatings as discussed in more details below. A worthy aspect of these coatings is their abrasion resistance exhibited during Martindale tests, as shown in Fig. 11, where the maximum number of cycles before the formation of holes in the fabric is reported for each coating. These results together with those collected by cone calorimetry suggest that a combination of silica and alumina performances could be useful in order to deposit multi-functional coatings on cotton fabrics. To this aim, new silica coatings containing alumina micro- or nano-particles have been tested in order to reach an optimal formulation for conferring flame retardancy and abrasion resistance to the fabrics, exploiting the features of both silica and alumina. The approach used in this work [167] was the physical modification of the silica phases by entrapping alumina particles within its network. As a result, very small amounts of alumina particles (regardless of their micrometric or nanometric size) within the silica network are able to decrease the total burning rate, increase the final residue and increase the abrasion resistance of the treated fabrics with respect to pure silica or alumina, as reported in Table 12. These results agree with those published by Brzezinski et al. [215], who deposited SiO2–Al2O3 xerogel coatings on cotton fabrics, that provided a very good and durable protection against abrasion under both use and care conditions, thus achieving a significant enhancement of the service durability. On the basis of these results, it is important to highlight that the thermal protection on cotton exerted by the sol–gel derived oxidic species can be referred not only to the thermal insulating effect of a ceramic layer, but also to the presence of the metal cations in the precursor itself. Indeed, these cations may affect the thermooxidation of cellulose favouring its dehydration, and thus the formation of high amounts of char [216]. Usually, this effect is

Table 11 Combustion tests of cotton fabrics treated with silica, titania, zirconia and alumina. Data from Ref. [168]. Sample

Cotton Cotton_silica Cotton_titania Cotton_zirconia Cotton_alumina

Vertical flame spread tests

Cone calorimeter tests

Burning rate [mm/s]

Residue [%]

TTI [s]

PHRR [kW/m2]

THR [MJ/m2]

TSR [m2/m2]

20 11 10 12 12

10 30 31 21 32

18 28 22 22 20

88 70 84 82 106

2.8 2.0 2.3 2.7 1.9

48.0 11.5 11.0 23.0 22.0

G. Malucelli et al. / Materials Science and Engineering R 84 (2014) 1–20

16

(i) mixing the alkoxysilane precursor with a phosphoric-acid source, (ii) using an alkoxysilane precursor bearing both silane and phosphate functionalities and (iii) mixing an alkoxysilane precursor bearing both silane and phosphate functionalities with P- and N-containing chemicals.

Fig. 11. Abrasion resistance of cotton fabrics treated with silica, titania, zirconia and alumina precursors.

Table 12 Flammability and abrasion resistance results for cotton treated with pure and alumina-doped silica. Data from Ref. [167]. Sample

Cotton Cotton_silica Cotton_alumina Cotton_micro alumina Cotton_nano alumina

Vertical flame spread tests

Abrasion resistance tests

Burning rate [mm/s]

D [%]

Residue [%]

N8 cycles before hole formation

7.5 5.8 12.0 4.4 4.2

– 23 +60 41 44

0 23.0 32.0 46.0 46.0

2100 4730 5780 4500 4400

observed either when the metal ions are used singularly or in combination with phosphorus-based flame retardants [217–219]. 4.3.2. Phosphorus-doped silica structures As mentioned above, in some cases, sol–gel derived architectures cannot be fully considered effective flame retardant systems unless operating in synergistic or joint effects with other flame retardants [220–228]. Three strategies have been adopted for cotton in order to reach this goal:

Table 13 summarizes the pure phosphorus and phosphorus/ nitrogen-doped architectures designed for conferring flame retardancy to textiles employing sol–gel treatments, by using all three approaches described above. Silane precursor, doping agent, the resulting synthesized doped oxidic phase, textile and corresponding reference paper are detailed. As far as the first approach is considered (i.e. mixing the alkoxysilane precursor with a phosphoric-acid source, Table 13) synergistic effects in sol–gel derived architectures doped with aluminium phosphinate, or with a mixture of aluminium phosphinate, melamine poly(phosphate) and zinc and boron oxide, or with a-zirconium phosphate (ZrP) nano-platelets have been demonstrated [220]. As an example, the presence of at least 5 wt.% phosphorus-based compounds with respect to the silica precursor strongly improves the flame retardancy of cotton: a TTI increase from 14 (untreated cotton) to 40s was observed. In addition, the presence of 5 wt.% ZrP within the silica network increased the LOI of the treated fabrics from 19 (pure cotton) to 30%. Analogously, Cireli and co-workers [221] have prepared phosphorus-doped silica thin films mixing TEOS and H3PO4 or ethyldichlorophosphate. Flammability tests have shown that cotton does not burn when phosphoric acid acts synergistically with the silica coating; in addition, the surface treatment is stable up to ten washing cycles (according to TS EN ISO 105-C06-A1S). The synergistic effect between phosphoric acid and silica has been also successfully applied to poly(acrylonitrile) (PAN) fibres: indeed, phosphorus-doped silica films turned out to make PAN fibres not flammable when a flame was applied for 15 s; this treatment was found to be resistant up to 10 washing cycles according to TS EN ISO 105-C06-A1S [222]. Referring to the second approach (i.e. using an alkoxysilane precursor bearing both silane and phosphate functionalities, Table 13), the concurrent presence of P and Si elements in the same precursor can be exploited for preparing hybrid organic– inorganic coatings that behave, at the same time, like char promoters (because of the phosphoric-acid source) and thermal shields (due to the inorganic ceramer). To this aim, diethylphosphatoethyltriethoxysilane (DPTES) has been employed as a monomer to synthesize a hybrid phosphorus-silicon organic–inorganic material. First attempts have been carried out by our group depositing a different

Table 13 Phosphorus and phosphorus/nitrogen-doped architectures designed for conferring flame retardancy to textiles employing sol–gel treatments. Silane precursor

Doping agent

(i) Mixing the alkoxysilane precursor with a phosphoric-acid source TMOS Aluminium phosphinate A mixture of aluminium phosphinate, melamine poly(phosphate) and zinc and boron oxide a-Zirconium phosphate nano-platelets TEOS H3PO4 or ethyldichlorophosphate TEOS H3PO4

Doped oxidic phase

Textile

Reference

P-doped silica P-, P- and N-, Znand B-doped silica P-doped silica P-doped silica P-doped silica

Cotton Cotton

[220] [220]

Cotton Cotton Poly acrylonitrile

[220] [221] [222]

Cotton

[223–225]

Cotton Cotton Cotton Cotton

[226] [227] [228] [228]

(ii) Using an alkoxysilane precursor bearing both silane and phosphate functionalities DPTES 1, 3 or 6 layers P-doped silica (iii) Mixing an alkoxysilane precursor bearing both silane and phosphate functionalities with P- and N-containing chemicals DPTES APTES or APTES and melamine-based resin P- and N-doped silica DPTES APTES or N,N,N0 ,N0 ,N00 ,N00 -hexakis-methoxymethyl-[1,3,5] triazine-2,4,6-triamine P- and N-doped silica DPTES 1-Hydroxyethane 1,1-diphosphonic acid P-doped silica N,N,N0 ,N0 ,N00 ,N00 -Hexakis-methoxymethyl-[1,3,5] triazine-2,4,6-triamine urea P- and N-doped silica

G. Malucelli et al. / Materials Science and Engineering R 84 (2014) 1–20

number of DPTES layers (namely, 1, 3 and 6) on cotton [223]. The presence of the coating turned out to be responsible for a strong sensitization of the cellulose decomposition but, at the same time, promoted a significant increase of the residues obtained at high temperatures. Despite the strong reduction of TTI (assessed by cone calorimeter), the coatings were able to protect the cotton fabrics by decreasing the combustion duration, as shown by the obtained flame out values (66, 62 and 80 vs. 116 s for 1, 3 and 6 layers and untreated cotton, respectively). As a consequence, the formation of volatile species was hindered: indeed, the TSR was significantly lowered for the treated samples (20, 15 and 6 vs. 26 m2/m2 for 1, 3 and 6 layers and untreated cotton, respectively). Better results were achieved after an optimization process consisting of the prehydrolysis of the precursor [224]. These coatings also showed good durability when subjected up to five washing cycles according to the ISO 6330 standard. In order to establish the real efficiency of DPTES, and thus a possible synergism between phosphorus and silicon within the network of these hybrids architectures, we compared the results collected by thermogravimetry, vertical flame spread tests and cone calorimetry when cotton is treated with TEOS or DPTES as precursors [225], and we concluded that: (i) both pure and phosphorus-doped silica phases are able to enhance the thermal and thermo-oxidative stability of cotton, exploiting the joint effect of thermal shielding (exerted by silica phases) and char-forming (exerted by the phosphoric acid source present in the alkoxysilane precursor); (ii) pure silica phases, even at very low add-ons, are able to protect cotton better than phosphorus-doped silica, from the application of a methane flame for 5s, favouring the formation of a thermally stable residue; indeed, the presence of a phosphoric acid precursor is negligible; (iii) phosphorus-doped silica phases are able to protect cotton better than pure silica when the fabric is exposed to an irradiative heat flux of 35 kW/m2. Recently, a new route has been carried out in order to exploit the synergism between silica coatings and P- and N-based chemicals (i.e. mixing an alkoxysilane precursor bearing both silane and phosphate functionalities with P- and N-containing chemicals, Table 13). Once again, DPTES was used for preparing hybrid organic–inorganic architectures, but in the presence of a nitrogen compound, taking into consideration that the most well-known synergistic combination refers to the area of phosphorus-nitrogen flame retardants for cellulosics [229,230]. Indeed, Proban1 (an organophosphorus- and nitrogen-containing product based on tetrakis (hydroxymethyl) phosphonium salt-urea condensates) and Pyrovatex1 (based on Nmethylol dimethyl phosphonamide derivatives) are commerciallyavailable durable finishes for cotton that develop an optimal flame retarding effectiveness when N/P molar ratios are about 2–2.5 and 1.5–2, respectively. More specifically, DPTES was combined with APTES or APTES and melamine-based resin (M) [226] or with APTES or N,N,N0 ,N0 ,N00 ,N00 -hexakis-methoxymethyl-[1,3,5] triazine-2,4,6triamine (MF) [227] or with 1-hydroxyethane 1,1-diphosphonic acid, MF or urea [228]. Treating cotton with DPTES in combination with APTES or APTES and melamine-based resin promotes the char-formation; indeed, char yields of 42 and 38 wt.% for APTES/DPTES- and APTES/DPTES/MF-treated samples, respectively, were found by thermogravimetric analysis in air [226]. Similar results have been observed by changing the nitrogen source and optimising the relative ratios between DPTES and the other counterpart [227]. In particular, the char-forming character of such coatings turned out to protect cotton and promote high

17

residues (around 50 and 70 wt.% for APTES/DPTES and MF/ DPTES, respectively). When two or more flame retardant systems are concurrently used, it becomes important to quantify their synergistic effects. To this aim, the concept of ‘‘synergistic effectiveness’’ (reported by Lewin [231] and Horrocks et al. [232]) has been exploited to evaluate the concomitant presence of phosphorus- and/or nitrogen-based compounds on the flame retardancy of sol–gel treated cotton fabrics. More specifically, it was demonstrated that synergism occurs when hybrid phosphorus-doped silica coatings are coupled with 1-hydroxyethane 1,1-diphosphonic acid; on the other hand, a simple additive effect takes place when the hybrid phosphorus-doped silica coatings are further doped with Ncontaining molecules such as melamine or urea [228]. 4.3.3. Smoke suppressant coatings Silica nanoparticles or silica-based coatings easily synthesized through sol–gel processes can be considered smoke suppressant architectures. Indeed, when cotton was treated with pure silica, a strong decrease of TSR (55%), as well as of CO and CO2 yield (50% and 23%, respectively), has been assessed by cone calorimeter [233]. Furthermore, in comparison with a common phosphorusflame retardant for cellulosics like Proban1, the optical density of smokes released in the presence of silica coating is much lower. This is shown in Fig. 12, where the smoke transmittance of pure cotton, cotton treated with silica from sol–gel processes and Proban1 is reported as a function of time during combustion in a smoke density chamber, under a heat flux of 35 kW/m2, following the standard ISO5659. Comparing the data reported in Table 14, it is worthy to observe that all the ceramic coatings previously investigated [168] are responsible of a TSR significant decrease: about 80% for silica and titania, and 55% for zirconia and alumina. This finding is ascribed to the char-former effect exerted by such architectures that are able to reduce the volatile production of cotton. The presence of a phosphate group in the alkoxysilane precursor does not affect the smoke suppressant behaviour of silica. More specifically, also when DPTES is employed for depositing hybrid inorganic–organic coating on cotton, the TSR measured by cone calorimeter is strongly reduced, reaching a reduction of about 80% when 3 layers are deposited [223,224]. By adding different smoke suppressants (namely, zinc oxide, zinc acetate dihydrate and zinc borate) or flame retardants (i.e. ammonium pentaborate octahydrate, boron phosphate and

Fig. 12. Smoke transmission as a function of time for untreated and silica- and Proban1-treated cotton fabrics.

18

G. Malucelli et al. / Materials Science and Engineering R 84 (2014) 1–20

Table 14 Combustion data by cone calorimetry of cotton fabrics treated with silica, smoke suppressants (SS) and flame retardants (FR). Data from Ref. [233]. Sample

TSR [m2/m2]

CO2 peak [%]

Cotton Cotton_silica Cotton_silica+zinc oxide (SS) Cotton_silica+zinc acetate dihydrate (SS) Cotton_silica+zinc borate (SS) Cotton_silica+ammonium pentaborate octahydrate (FR) Cotton_silica+boron phosphate (FR) Cotton_silica+ammonium polyphosphate (FR)

24 11 9 18 17 21

0.146 0.112 0.095 0.110 0.108 0.108

25 19

0.092 0.090

ammonium polyphosphate) within the sol solution, TSR and CO2 yield values were significantly reduced (Table 14) [233]. In the presence of zinc-based smoke suppressants, the release of smoke and CO2 has been significantly further reduced with respect to the fabric treated with the silica coating alone. In particular, the joint effect of ZnO and silica has promoted the most significant decrease of TSR and CO2 yields (62% and 35%, respectively), while the combination of silica with phosphorus- or boron-based flame retardants did not promote any remarkable decrease in the smoke production, but reduce the CO2 release [233]. 4.3.4. Interpenetrated hybrid organic–inorganic network structures derived from sol–gel processes performed within a polymer matrix The exploitation of dual-cure processes that involve a photopolymerization reaction and a subsequent sol–gel process represents a versatile route, by which it is possible to synthesize hybrid organic–inorganic architectures. These latter are able to enhance the flame retardancy of cotton [234,235]. To this aim, different amounts of TMOS precursor (ranging from 30 up to 80 wt.%) have been added to a UV-curable acrylic formulation in the presence of methacryloyloxypropyltrimethoxysilane as coupling agent and applied to cotton [234]. The hybrid organic– inorganic coating acted as an efficient thermal insulator and inhibited the cotton combustion. In particular, the formulation containing 60 wt.% TMOS extended the total burning time, increased the final residue upon flammability tests in horizontal configuration and delayed the ignition under an irradiating heat flux of 35 kW/m2. Xing et al. [235] prepared UV-cured flame retardant coatings using tri(acryloyloxyethyl)phosphate and triglycidyl isocyanurate acrylate. Once again, the obtained hybrid coating turned out to be responsible of the ignition temperature, THR and PHRR decrease (about 60 and 30%), as assessed by the pyrolysis combustion flow calorimeter, as well as of a significant increase of LOI value (from 21 to 24% for untreated and treated cotton fabrics). 4.4. Other substrates Beside textiles, sol–gel treatments have been rarely used for enhancing the flame retardancy of other substrates apart from wood which modification applying the sol–gel process of silicon alkoxides has been reported in the literature. The application of the sol–gel process was most intensively studied by Saka et al. [236,237]. Their impregnation technique was aimed at using the bound water in the cell wall in order to direct the sol–gel process to the cell wall and to achieve a deposition of the silicate therein. The results collected by LOI tests [236] and thermogravimetric analyses [238] have demonstrated that the presence of silica coatings negatively affects the wood behaviour from the thermal and combustion point of view. On the other hand, its fire resistance was enhanced by combination of TEOS with trimethylphosphite

or/and trimethylborate in order to produce SiO2–P2O5–B2O3 wood–inorganic composites [239]. Further fire resistant silicon oligomers with ethylphosphite and/or boric hydroxide residues were produced that revealed higher resistance to leaching than the SiO2–P2O5–B2O3 compounds. Very recently, titania [240] and silica [241] derived from sol–gel method was synthesized by Schartel and co-workers in order to enhance wood flame retardancy. As far as titania coatings are concerned [240], the results obtained from cone calorimeter revealed that heat release rates after the initial PHRR and in particular, the heat release rate of the second PHRR in a developing fire were moderately reduced as well as the burning time was significantly increased. Furthermore, parameters indicating fire hazards such as CO and smoke production were reduced in a remarkable way. Beside these improvements, an impressive improvement in LOI values up to 64% was registered, showing better flame retardant properties of these materials when reacting to small flame. However, the Authors observed that fire behaviour and the flammability of these materials readily depended on the fire scenarios as well as on some drawbacks to overcome, such as fractures and cracks in gel layers and organic residues. Analogous improvements and limitations in the flame retardancy of sol–gel treated wood have been found when titania was replaced with silica or with a mixture of titania and silica [241]. 5. Conclusions Engineering the polymer surface is shown to provide a potential promising, environmentally-friendly and effective approach to polymer fire retardance, particularly when combined with nanostructurating technologies. Feasibility is demonstrated for textiles, films and foams while present efforts are directed towards composites with possible future extension to thick polymer materials. A major interest in this approach to surface polymer properties is the possibility to simultaneously confer multifunctional features that, besides fire retardance, may involve gas barrier, hydrophobicity, biocide activity, surface electrical conductivity, etc. References [1] (a) F. Laoutid, L. Bonnaud, M. Alexandre, J.-M. Lopez-Cuesta, Ph. Dubois, Mater. Sci. Eng. R. 63 (2009) 100–125; (b) A. Dasari, Z.-Z. Yu, G.-P. Cai, Y.-W. Mai, Prog. Polym. Sci. 38 (2013) 1357–1387. [2] (a) M. Bartholmai, R. Schriever, B. Schartel, Fire Mater. 27 (2003) 151–162; (b) M. Bartholmai, B. Schartel, Fire Mater. 31 (2007) 187–205. [3] H.L. Vandersall, J. Fire Flam. 2 (1971) 97–140. [4] J.A. Ellard, Div. Org. Coat. Plast. Chem. 33 (1973) 531–545. [5] (a) M. Le Bras, G. Camino, S. Bourbigot, R. Delobel (Eds.), Fire Retardancy of Polymers: The Use of Intumescence, The Royal Society of Chemistry, Cambridge, UK, 1998; (b) E.D. Weil, J. Fire Sci. 29 (2011) 259–296; (c) S. Bourbigot, S. Duquesne, J. Mater. Chem. 17 (2007) 2283–2300. [6] S. Duquesne, N. Renaut, P. Bardollet, C. Jama, M. Traisnel, R. Delobel, Fire retardancy of polypropylene composites using intumescent coatings, Fire and Polymers V ACS Symposium Series, vol. 1013, American Chemical Society, Washington, 2009, pp. 192–204 (Chapter 12). [7] M. Jimenez, S. Duquesne, S. Bourbigot, Polym. Adv. Technol. 23 (2012) 130–135. [8] M. Jimenez, H. Gallou, S. Duquesne, C. Jama, S. Bourbigot, X. Couillens, F. Speroni, J. Fire Sci. 30 (2012) 535–551. [9] M. Jimenez, S. Duquesne, S. Bourbigot, Polym. Degrad. Stab. 98 (2013) 1378–1388. [10] S. Duquesne, M. Jimenez, S. Bourbigot, Fire retardancy and fire protection of materials using intumescent coatings—a versatile solution? in: Fire Retardancy of Polymers: New Strategies and Mechanisms, Royal Society of Chemistry, 2009, pp. 240–252. [11] U. Sorathia, T. Gracik, J. Ness, A. Durkin, F. Williams, M. Hunstad, F. Berry, J. Fire Sci. 21 (2003) 423–450. [12] S. Bourbigot, P. Bachelet, F. Samyn, M. Jimenez, S. Duquesne, Compos. Interfaces 20 (2013) 269–277. [13] E. Kandare, B. Kandola, P. Myler, Fire Safety J. 58 (2013) 112–120. [14] E. Kandare, G.J. Griffin, S. Feih, A.G. Gibson, B.Y. Lattimer, A.P. Mouritz, Compos. Part A-Appl. Sci. 43 (2012) 793–802.

G. Malucelli et al. / Materials Science and Engineering R 84 (2014) 1–20 [15] A.P. Mouritz, S. Feih, E. Kandare, A.G. Gibson, Compos. Part B-Eng. 48 (2013) 68–78. [16] R.K. Iler, J. Colloid Interface Sci. 21 (1966) 569–594. [17] G. Decher, J.D. Hong, Makromol. Chem. Macromol. Symp. 46 (1991) 321–327. [18] S. Mecking, R. Thomann, Adv. Mater. 12 (2000) 953–956. [19] M. Ferreira, M.F. Rubner, B.R. Hsieh, Mater. Res. Soc. Symp. Proc. 328 (1994) 119– 124. [20] A.C. Fou, O. Onitsuka, M. Ferreira, M.F. Rubner, B.R. Hsieh, Mater. Res. Soc. Symp. Proc. 369 (1995) 575–580. [21] Y. Shimazaki, M. Mitsuishi, S. Ito, M. Yamamoto, Langmuir 13 (1997) 1385–1390. [22] Y. Shimazaki, M. Mitsuishi, S. Ito, M. Yamamoto, Langmuir 14 (1998) 2768–2773. [23] W.B. Stockton, M.F. Rubner, Macromolecules 30 (1997) 2717–2725. [24] L.Y. Wang, S.X. Cui, Z.Q. Wang, X. Zhang, M. Jiang, L.F. Chi, H. Fuchs, Langmuir 16 (2000) 10490–10494. [25] M.M. Fang, D.M. Kaschak, A.C. Sutorik, T.E. Mallouk, J. Am. Chem. Soc. 119 (1997) 12184–12191. [26] I. Ichinose, T. Kawakami, T. Kunitake, Adv. Mater. 10 (1998) 535–539. [27] T. Serizawa, K. Hamada, T. Kitayama, K. Katsukawa, K. Hatada, M. Akashi, Langmuir 16 (2000) 7112–7115. [28] T. Serizawa, H. Yamashita, T. Fujiwara, Y. Kimura, M. Akashi, Macromolecules 34 (2001) 1996–2001. [29] K. Spaeth, A. Brecht, G. Gauglitz, J. Colloid Interface Sci. 196 (1997) 128–135. [30] J. Anzai, Y. Kobayashi, Y. Suzuki, H. Takeshita, Q. Chen, T. Osa, T. Hoshi, X.Y. Du, Sens. Actuators, B: Chem. 52 (1998) 3–9. [31] P. Bernt, K. Kurihara, T. Kunitake, Langmuir 8 (1992) 2486–2490. [32] J.C. Love, L.A. Estroff, J.K. Kriebel, R.G. Nuzzo, G.M. Whitesides, Chem. Rev. 105 (2005) 1103–1169. [33] J.H. Fendler, Chem. Mater. 8 (1996) 1616–1624. [34] W. Knoll, Curr. Opin. Colloid. Interface Sci. 1 (1996) 137–143. [35] P. Bertrand, A. Jonas, A. Laschewsky, R. Legras, Macromol. Rapid Commun. 21 (2000) 319–348. [36] P.T. Hammond, Adv. Mater. 16 (2004) 1271–1293. [37] G. Decher, J.D. Hong, J. Schmitt, Thin Solid Films 210 (1992) 831–835. [38] G. Decher, Y. Lvov, J. Schmitt, Thin Solid Films 244 (1994) 772–777. [39] D. Korneev, Y. Lvov, G. Decher, J. Schmitt, S. Yaradaikin, Physica B: Condens. Matter 213 (1995) 954–956. [40] W. Chen, T.J. McCarthy, Macromolecules 30 (1997) 78–86. [41] N.A. Kotov, I. Dekany, J.H. Fendler, J. Phys. Chem. 99 (1995) 13065–13069. [42] K. Ariga, Y. Lvov, M. Onda, I. Ichinose, T. Kunitake, Chem. Lett. 26 (1997) 125–128. [43] Y.M. Lvov, K. Ariga, M. Onda, I. Ichinose, T. Kunitake, Langmuir 13 (1997) 6195–6203. [44] H.Y. Koo, D.K. Yi, S.J. Yoo, D.Y. Kim, Adv. Mater. 16 (2004) 274–277. [45] Y. Lvov, K. Ariga, I. Ichinose, T. Kunitake, Langmuir 12 (1996) 3038–3044. [46] A. Mamedov, J. Ostrander, F. Aliev, N.A. Kotov, Langmuir 16 (2000) 3941–3949. [47] Y. Zhou, N. Hu, Y. Zeng, J.F. Rusling, Langmuir 18 (2002) 211–219. [48] B. Neu, A. Voigt, R. Mitlohner, S. Leporatti, C.Y. Gao, E. Donath, H. Kiesewetter, H. Mo¨hwald, H.J. Meiselman, H. Baumler, J. Microencapsulation 18 (2001) 385–395. [49] O. Mermut, C.J. Barrett, J. Phys. Chem. B 107 (2003) 2525–2530. [50] Z.J. Sui, D. Salloum, J.B. Schlenoff, Langmuir 19 (2003) 2491–2495. [51] H.N. Zhang, J. Ruhe, Macromolecules 36 (2003) 6593–6598. [52] R.A. McAloney, M. Sinyor, V. Dudnik, M.C. Goh, Langmuir 17 (2001) 6655–6663. [53] L. Chang, X. Kong, F. Wang, L. Wang, J. Shen, Thin Solid Films 516 (2008) 2125–2129. [54] S. Shiratori, M.F. Rubner, Macromolecules 33 (2000) 4213–4219. [55] H.L. Tan, M.J. McMurdo, G.Q. Pan, P.G. Van Patten, Langmuir 19 (2003) 9311–9314. [56] S. Fujita, S. Shiratori, Thin Solid Films 499 (2006) 54–60. [57] J.M.C. Lourenc¸o, P.A. Ribeiro, A.M. Botelho do Rego, M. Raposo, J. Colloid Interface Sci. 313 (2007) 26–33. [58] P. Schaaf, J.C. Voegel, L. Jierry, F. Boulmedais, Adv. Mater. 24 (2012) 1001–1021. [59] J.B. Schlenoff, S.T. Dubas, T. Farhat, Langmuir 16 (2000) 9968–9969. [60] A. Izquierdo, S.S. Ono, J.C. Voegel, P. Schaaf, G. Decher, Langmuir 21 (2005) 7558– 7567. [61] C.H. Porcel, A. Izquierdo, V. Ball, G. Decher, J.C. Voegel, P. Schaaf, Langmuir 21 (2005) 800–802. [62] K.C. Krogman, N.S. Zacharia, S. Schroeder, P.T. Hammond, Langmuir 23 (2007) 3137–3141. [63] P.T. Hammond, K.C. Krogman, N.S. Zacharia, WO 2008030474 (A2), 2009. [64] J. Cho, K. Char, J.D. Hong, K.B. Lee, Adv. Mater. 13 (2001) 1076–1080. [65] S.S. Lee, J.D. Hong, C.H. Kim, K. Kim, J.P. Koo, K.B. Lee, Macromolecules 34 (2001) 5727–5730. [66] P.A. Chiarelli, M.S. Johal, J.L. Casson, J.B. Roberts, J.M. Robinson, H.L. Wang, Adv. Mater. 13 (2001) 1167–1170. [67] G. Decher, in: G. Decher, J.B. Schlenoff (Eds.), Polyelectrolite Multilayers, An Overview, Multilayer Thin Films, Sequential Assembly of Nanocomposite Materials, Wiley VCH, Winheim, 2003, p. p. 1. [68] K. Srikulkit, C. Iamsamai, S.T. Dubas, J. Jom, Min. Met. Mats 16 (2006) 41–45. [69] J. Hao, M. Lewin, C.A. Wilkie, J. Wang, Polym. Degrad. Stab. 91 (2006) 2482–2485. [70] Y. Tang, M. Lewin, Polym. Degrad. Stab. 92 (2007) 53–60. [71] M. Lewin, Y. Tang, Macromolecules 41 (2008) 13–17. [72] A.B. Morgan, J.W. Gilman, Fire Mater. 37 (2013) 259–279. [73] S. Bourbigot, M. Le Bras, S. Duquesne, M. Rochery, Macromol. Mat. Eng. 289 (2004) 499–511. [74] S. Bourbigot, S. Duquesne, J. Mater. Chem. 17 (2007) 2283–2290. [75] Y.C. Li, J. Schulz, J.C. Grunlan, Appl. Mater. Int. 1 (2009) 2338–2340.

19

[76] Y.C. Li, J. Schulz, S. Mannen, C. Delhom, B. Condon, S. Chang, M. Zammarano, J.C. Grunlan, ACS Nano 4 (2010) 3325–3330. [77] D. Lee, M.F. Rubner, R.E. Cohen, Nano Lett. 6 (2006) 2305–2310. [78] D. Lee, Z. Gemici, M.F. Rubner, R.E. Cohen, Langmuir 23 (2007) 8833–8840. [79] G. Laufer, F. Carosio, R. Martinez, G. Camino, J.C. Grunlan, J. Colloid Interface Sci. 356 (2011) 69–77. [80] F. Carosio, G. Laufer, J. Alongi, G. Camino, J.C. Grunlan, Polym. Degrad. Stab. 96 (2011) 745–754. [81] J. Alongi, F. Carosio, A. Frache, G. Malucelli, Carbohydr. Polym. 92 (2013) 114–119. [82] F. Carosio, A. Di Blasio, F. Cuttica, J. Alongi, A. Frache, M. Malucelli, Ind. Eng. Chem. Res. 52 (2013) 9544–9550. [83] S.S. Ug˘ur, M. Sarııs¸ık, A.H. Aktas¸, Mat. Res. Bull. 46 (2011) 1202–1210. [84] Y.C. Li, S. Mannen, J. Schulz, J.C. Grunlan, J. Mater. Chem. 21 (2011) 3060–3070. [85] F. Carosio, J. Alongi, G. Malucelli, J. Mater. Chem. 21 (2011) 10370–10376. [86] T. Xu, L. Zhang, Y. Zhong, Z. Mao, J. Appl. Polym. Sci. 131 (2014), http://dx.doi.org/ 10.1002/APP. 39608 (Article no. 39608). [87] Y.C. Li, S. Mannen, A.B. Morgan, S.C. Chang, Y.H. Yang, B. Condon, J.C. Grunlan, Adv. Mater. 23 (2011) 3926–3930. [88] K. Apaydin, A. Laachachi, V. Ball, M. Jimenez, S. Bourbigot, V. Toniazzo, D. Ruch, Polym. Degrad. Stab. 98 (2013) 627–634. [89] F. Carosio, J. Alongi, G. Malucelli, Carbohydr. Polym. 88 (2012) 1460–1470. [90] J. Alongi, F. Carosio, G. Malucelli, Polym. Degrad. Stab. 97 (2012) 1644–1650. [91] F. Carosio, J. Alongi, G. Malucelli, Polym. Degrad. Stab. 98 (2013) 1626–1630. [92] T. Zhang, H. Yan, L. Wang, Z. Fang, Ind. Eng. Chem. Res. 52 (2013) 6138–6140. [93] L. Wang, T. Zhang, H.Q. Yan, M. Peng, Z. Fang, Y. Li, H. Wang, Chin. J. Polym. Sci. 8 (2013). [94] L. Wang, T. Zhang, H. Yan, M. Peng, Z. Fang, J. Appl. Polym. Sci (2013), http:// dx.doi.org/10.1002/APP 39015. [95] T. Zhang, H. Yan, M. Peng, L. Wang, Z. Fang, Nanoscale 5 (2013) 3013–3020. [96] C. Guirao, F. Carosio, B. Boutevin, H. Cottet, C. Loubat, J. Polym. Sci. B 51 (2013) 1244–1250. [97] G. Laufer, C. Kirkland, A. Morgan, J.C. Grunlan, Biomacromolecules (2012) 2843– 2848. [98] F. Carosio, A. Di Blasio, J. Alongi, G. Malucelli, Polymer 54 (2013) 5148–5150. [99] J. Alongi, R.A. Carletto, A. Di Blasio, F. Carosio, F. Bosco, G. Malucelli, J. Mater. Chem. A 1 (2013) 4779–4785. [100] J. Alongi, R.A. Carletto, A. Di Blasio, F. Carosio, F. Bosco, G. Malucelli, Carbohydr. Polym. 96 (2013) 296–304. [101] J. Alongi, F. Carosio, G. Malucelli, Cellulose 19 (2012) 1041–1050. [102] G. Huang, H. Liang, X. Wang, J. Gao, Ind. Eng. Chem. Res. 51 (2012) 12299–12309. [103] G. Huang, J. Yang, J. Gao, X. Wang, Ind. Eng. Chem. Res. 51 (2012) 12355–12366. [104] M.M. Hirschler, Polym. Adv. Technol. 19 (2008) 521–529. [105] Y.S. Kim, R. Davis, A.A. Cain, J.C. Grunlan, Polymer 52 (2011) 2847–2850. [106] Y.S. Kim, R. Davis, Thin Solid Films 550 (2014) 184–189. [107] G. Laufer, C. Kirkland, A. Cain, J.C. Grunlan, Appl. Mater. Interfaces 4 (2012) 1643–1650. [108] Y.S. Kim, R. Harris, R. Davis, ACS Macro Lett. 1 (2012) 820–824. [109] Y.C. Li, Y.S. Kim, J. Shields, R. Davis, J. Mater. Chem. A 1 (2013) 12987–12997. [110] Y.S. Kim, Y.C. Li, W.M. Pitts, M. Werrel, R. Davis, ACS Appl. Mater. Interfaces 6 (2014) 2146–2150. [111] A.A. Cain, C.R. Nolen, Y.C. Li, R. Davis, J.C. Grunlan, Polym. Degrad. Stab. 98 (2013) 2645–2650. [112] G. Laufer, C. Kirkland, A.B. Morgan, J.C. Grunlan, ACS Macro Lett. 2 (2013) 361–365. [113] F. Carosio, A. Di Blasio, F. Cuttica, J. Alongi, G. Malucelli, RSC Adv. 4 (2014) 16674–16680. [114] A. Laachachi, V. Ball, K. Apaydin, V. Toniazzo, D. Ruch, Langmuir 27 (2011) 13879–13880. [115] K. Apaydin, A. Laachachi, V. Ball, M. Jimenez, S. Bourbigot, V. Toniazzo, D. Ruch, Polym. Degrad. Stab. 98 (2013) 627–634. [116] F. Carosio, A. Di Blasio, J. Alongi, G. Malucelli, Eur. Polym. J. 49 (2013) 397–404. [117] M. Niederberger, N. Pinna, Metal Oxide Nanoparticles in Organic Solvents Synthesis, Formation, Assembly and Application, Springer, London, UK, 2009. [118] L.L. Hench, J.K. West, Chem. Rev. 90 (1990) 33–72. [119] J. Livage, M. Henry, C. Sanchez, Prog. Solid State Chem. 18 (1988) 259–341. [120] S. Sakka, Sol–gel Science and Technology. Topics and Fundamental Research and Applications, Kluwer Academic Publishers, Norwell, 2003. [121] Y.W. Jun, J.S. Choi, J. Cheon, Angew. Chem. Int. Ed. 45 (2006) 3414–3420. [122] J. Park, J. Joo, S.G. Kwon, Y. Jang, T. Hyeon, Angew. Chem. Int. Ed. 46 (2007) 4630– 4640. [123] M. Niederberger, Acc. Chem. Res. 40 (2007) 793–800. [124] M. Niederberger, M. Antonietti, in: C.N.R. Rao, A. Mu¨ller, A.K. Cheetam (Eds.), Nanomaterials Chemistry: Recent Developments and New Directions, WileyVCH, Weinheim, 2007, pp. 119–138. [125] B. Mahltig, T. Textore, Nanosols and Textiles, World Scientific Publishing, Singapore, 2008. [126] U. Schubert, N. Hu¨sing, A. Lorenz, Chem. Mater. 7 (1995) 2010–2020. [127] S.-Z. Chu, S. Inoue, K. Wada, D. Li, H. Haneda, S. Awatsu, J. Phys. Chem. B 107 (2003) 6586–6590. [128] D.Y. Kong, M. Yu, C.K. Lin, X.M. Liu, J. Lin, J. Fang, J. Electrochem. Soc. 152 (2005) 146–151. [129] B. Tan, S.E. Rankin, J. Non-Cryst. Solids 352 (2006) 5453–5460. [130] B. Mahltig, H. Bo¨ttcher, J Sol-Gel Sci. Technol. 27 (2003) 43–52. [131] T. Textor, T. Bahners, E. Schollmeyer, Prog. Colloid Polym. Sci. 117 (2001) 76–79. [132] R. Nass, E. Arpac, W. Glaubitt, H. Schmidt, J. Non-Cryst. Solids 121 (1990) 370–374.

20

G. Malucelli et al. / Materials Science and Engineering R 84 (2014) 1–20

[133] J. Liu, J.C. Berg, J. Mater. Chem. 17 (2007) 4430–4440. [134] L. Betancor, F. Lo´pez-Gallego, A. Hidalgo, M. Fuentes, O. Podrasky, G. Kuncova, J.M. Guisa´n, R. Ferna´ndez-Lafuente, Biomacromolecules 6 (2005) 1027–1030. [135] F.Y. Li, Y.J. Xing, X. Ding, Y. Zu, Enzyme Microb. Technol. 40 (2007) 1692–1700. [136] J.-P. Chen, W.-S. Lin, M.F. Chang, J. Am. Oil Chem. Soc. 79 (2002) 309–314. [137] K.S. Huang, Y.H. Nien, K.C. Hsiao, Y.S. Chang, J. Appl. Polym. Sci. 102 (2006) 4136– 4140. [138] P. Xu, W. Wang, S.-L. Chen, Melliand Int. 11 (2005) 56–59. [139] C.-C. Wu, H. Yang, B. Feng, W.-W. Lu, Electron. Compon. Mater. 23 (2004) 22–24. [140] H. Bo¨ttcher, C. Jagota, J. Trepte, K.-H. Kallies, H. Haufe, J. Controlled Release 60 (1999) 57–65. [141] H. Haufe, A. Thron, D. Fiedler, B. Mahltig, H. Bo¨ttcher, Surf. Coat. Int. B: Coat. Trans. 88 (2005) 55–60. [142] Y.J. Xing, X.J. Yang, J.J. Dai, J. Sol-Gel Sci. Technol. 43 (2007) 187–192. [143] W.A. Daoud, J.H. Xin, Y.-H. Zhang, Surf. Sci. 599 (2005) 69–75. [144] B. Mahltig, D. Fiedler, H. Bo¨ttcher, J. Sol-Gel Sci. Technol. 32 (2004) 219–222. [145] S. Tarimala, N. Kothari, N. Abidi, E. Hequet, J. Fralick, L.L. Dai, J. Appl. Polym. Sci. 101 (2006) 2938–2940. [146] C.C. Trapalis, M. Kokkoris, G. Perdikakis, G. Kordas, J. Sol-Gel Sci. Technol. 26 (2003) 1213–1220. [147] N. Vigneshwaran, S. Kumar, A.A. Kathe, P.V. Varadarajan, V. Prasad, Nanotechnology 17 (2006) 5087–5090. [148] H. Bo¨ttcher, K.-H. Kallies, H. Haufe, J. Sol-Gel Sci. Technol. 8 (1997) 654–661. [149] G. Carturan, E. Pagani, R. Campostrini, R. Ceccato, J. Sol-Gel Sci. Technol. 8 (1997) 1115–1120. [150] T. Lo´pez, P. Quintana, J.M. Martı´nez, D. Esquivel, J. Non-Cryst. Solids 353 (2007) 987–989. [151] B. Mahltig, T. Textor, J. Sol-Gel Sci. Technol. 39 (2006) 111–118. [152] A. Cireli, N. Onar, J. Appl. Polym. Sci. 109 (2008) 97–105. [153] A. Dubois, M. Canva, A. Brun, F. Chaput, J.-P. Boilot, Synth. Met. 81 (1996) 305–308. [154] J. Trepte, H. Bo¨ttcher, J. Sol-Gel Sci. Technol. 19 (2000) 691–694. [155] L. Min, Z. Xiaoli, C. Shuilin, Color. Technol. 119 (2003) 297–300. [156] B. Mahltig, D. Knittel, E. Schollmeyer, H. Bo¨ttcher, J. Sol-Gel Sci. Technol. 31 (2004) 293–297. [157] B. Mahltig, D. Knittel, E. Schollmeyer, H. Bo¨ttcher, Text. Res. J. 74 (2004) 521–527. [158] D. Juan, Z. Li, C. Shuilin, Color. Technol. 121 (2005) 29–36. [159] P. Prosposito, M. Casalboni, F. De Matteis, R. Pizzoferrato, Thin Solid Films 373 (2000) 150–154. [160] M. Yu, G. Gu, W.D. Meng, F.L. Qing, Appl. Surf. Sci. 253 (2007) 3669–3670. [161] C.H. Xue, S.T. Ji, H.Z. Chen, M. Wang, Sci. Technol. Adv. Mater. 9 (2008) 035001 (1)–035001 (5). [162] J. Vince, B. Orel, A. Vilcnik, M. Fir, A.S. Vuk, V. Jovanovski, B. Simoncic, Langmuir 22 (2006) 6489–6490. [163] S.Benfer, S. Bo¨hme, R. Hu¨bner, E. Schmalz, G. Tomardl, A method for surface modification of textile filter media by sol–gel processing. German Patent DE10209667, 2002. [164] T. Textor, T. Bahners, E. Schollmeyer, J. Ind. Text. 32 (2003) 279–289. [165] C. Schramm, B. Rinderer, W.H. Binder, R. Tessadri, H. Duelli, J. Mater. Sci. 40 (2005) 1883–1890. [166] M.S. Lee, N.J. Jo, J. Sol-Gel Sci. Technol. 24 (2002) 175–180. [167] J. Alongi, G. Malucelli, Polym. Degrad. Stab. 98 (2013) 1428–1430. [168] J. Alongi, M. Ciobanu, G. Malucelli, Carbohydr. Polym. 87 (2012) 2093–2100. [169] Y. Han, J. Lin, H. Zhang, Mater. Lett. 54 (2002) 389–396. [170] G. Ramos, F. del Monte, B. Zurro, K.J. McCarthy, A. Baciero, D. Levy, Langmuir 18 (2002) 984–986. [171] S. Thomas, D. Sakthikumar, P.A. Joy, Y. Yoshida, M.R. Anantharaman, Nanotechnology 17 (2006) 5565–5570. [172] L. Cerne, B. Simoncic, Text. Res. J. 74 (2004) 426–432. [173] J.-T. Yeh, C.-L. Chen, K.-S. Huang, J. Appl. Polym. Sci. 103 (2007) 3019–3020. [174] J.-T. Yeh, C.-L. Chen, K.-S. Huang, J. Appl. Polym. Sci. 103 (2007) 1140–1150. [175] K. Satoh, H. Nakazumi, M. Morita, Text. Res. J. 74 (2004) 1079–1080. [176] H.F. Moafi, A.F. Shojaie, M.A. Zanjanchi, J. Therm. Anal. Calorim. 104 (2011) 717–724. [177] C. Colleoni, M.R. Massafra, G. Rosace, Surf. Coat. Technol. 207 (2012) 79–88. [178] W.A. Daoud, J. Xin, Y.-H. Zhang, K. Qi, J. Non-Cryst. Solids 351 (2005) 1486–1490. [179] W.A. Daoud, J.H. Xin, G.K.H. Pang, J. Am. Ceram. Soc. 88 (2005) 443–446. [180] Z. Liuxue, L. Peng, S. Zhixing, Mater. Chem. Phys. 98 (2006) 111–115. [181] Z. Liuxue, W. Xiulian, L. Peng, S. Zhixing, Appl. Surf. Sci. 254 (2008) 1771–1780. [182] T. Textor, F. Schro¨ter, E. Schollmeyer, Macromol. Symp. 254 (2007) 196–202. [183] M. Zayat, P. Garcia-Parejo, D. Levy, Chem. Soc. Rev. 36 (2007) 1270–1280. [184] B. Mahltig, F. Haufe, H. Bo¨ttcher, J. Mater. Chem. 15 (2005) 4385–4390. [185] N. Abidi, E. Hequet, S. Tarimala, L.L. Dai, J. Appl. Polym. Sci. 104 (2007) 111–117. [186] X.J. Xing, J. Ding, J. Appl. Polym. Sci. 103 (2007) 3113–3120. [187] B. Mahltig, H. Bo¨ttcher, H. Rauch, U. Dieckman, R. Nitsche, T. Fritz, Thin Solid Films 485 (2005) 108–114.

[188] J.H. Xin, W.A. Daoud, Y.Y. Kong, Text. Res. J. 74 (2004) 97–100. [189] P. Xu, X. Liu, W. Wang, S. Chen, Appl. Polym. Sci. 102 (2006) 1478–1480. [190] M. Caldara, C. Colleoni, E. Guido, V. Re, G. Rosace, Sens. Actuators, B: Chem. 171– 172 (2012) 1013–1021. [191] L. Van der Schueren, K. De Clerck, G. Brancatelli, G. Rosace, E. Van Damme, W. De Vos, Sens. Actuators, B: Chem. 162 (2012) 27–34. [192] M. Qu, A. Liu, H. Sun, Barium sulfate sol for modifying polyester and its preparing process. Chinese Patent, CN1365993, 2002. [193] Z.Z. Xu, C.C. Wang, W.L. Yang, S.K. Fu, J. Mater. Sci. 40 (2005) 4667–4670. [194] C.L. Chiang, C.C.M. Ma, Eur. Polym. J. 38 (2002) 2219–2220. [195] C.L. Chiang, F.Y. Wang, C.C.M. Ma, H.R. Chang, Polym. Degrad. Stab. 77 (2002) 273–278. [196] C.L. Chiang, R.C. Chang, Compos. Sci. Technol. 68 (2008) 2849–2850. [197] Y.L. Liu, C.S. Wu, Y.S. Chiu, W.H. Ho, J. Polym. Sci. A: Polym. Chem. 41 (2003) 2354–2360. [198] Y. Liu, C. Chou, Polym. Degrad. Stab. 90 (2005) 515–522. [199] D. Yu, W. Liu, Y. Liu, Polym. Compos. 31 (2010) 334–339. [200] C.L. Chiang, C.C.M. Ma, D.L. Wu, H.C. Kuan, J. Polym. Sci. A: Polym. Chem. 41 (2003) 905–913. [201] M. Messori, M. Toselli, F. Pilati, P. Fabbri, S. Busoli, L. Pasquali, S. Nannarone, Polymer 44 (2003) 4463–4470. [202] Q. Ji, X. Wang, Y. Zhang, Q. Kong, Y. Xi, Compos., A: Appl. Sci. Manuf. 40 (2009) 878–882. [203] T. Kashiwagi, J.W. Gilman, K.M. Butler, R.H. Harris, J.R. Schields, A. Asano, Fire Mater. 24 (2000) 277–289. [204] J. Alongi, A. Frache, G. Malucelli, G. Camino, in: F. Selcen Kilinc (Ed.), Handbook of Fire Resistant Textiles, Woodhead Publishing, Cambridge, 2013, pp. 68–93. [205] J. Alongi, F. Carosio, A.R. Horrocks, G. Malucelli, Update on Flame Retardant textiles: State of the art, Environmental Issues and Innovative Solutions, Smithers RAPRA Publishing, Shawbury, Shrewsbury, Shropshire, 2013. [206] J. Alongi, G. Malucelli, J. Mater. Chem. 22 (2012) 21805–21809. [207] J. Alongi, F. Carosio, G. Malucelli, Polym. Degrad. Stab. 106 (2014) 138–149. [208] S. Hribernik, M.S. Smole, K.S. Kleinschek, M. Bele, J. Jamink, M. Gaberscek, Polym. Degrad. Stab. 92 (2007) 1957–1960. [209] J. Alongi, M. Ciobanu, F. Carosio, J. Tata, G. Malucelli, J. Appl. Polym. Sci. 119 (2011) 1961–1970. [210] C. Colleoni, I. Donelli, G. Freddi, E. Guido, V. Migani, G. Rosace, Surf. Coat. Technol. 235 (2013) 192–203. [211] J. Alongi, M. Ciobanu, G. Malucelli, Cellulose 18 (2011) 167–177. [212] J. Alongi, G. Malucelli, J. Therm. Anal. Calorim. 11 (2013) 459–465. [213] J. Alongi, M. Ciobanu, G. Malucelli, Carbohydr. Polym. 87 (2012) 627–635. [214] J. Tata, J. Alongi, F. Carosio, A. Frache, Fire Mater. 35 (2011) 397–409. [215] S. Brzezinski, D. Kowalczyk, B. Borak, M. Jasiorski, A. Tracz, J. Appl. Polym. Sci. 125 (2012) 3058–3060. [216] S. Soares, G. Camino, S. Levchik, Polym. Degrad. Stab. 62 (1998) 25–31. [217] P.J. Davies, R.A. Horrocks, A. Alderson, Polym. Degrad. Stab. 88 (2005) 114–122. [218] C.M. Tian, J.X. Xie, H.Z. Guo, J.Z. Xu, J. Therm. Anal. Calorim. 73 (2003) 827–834. [219] C.M. Tian, H.Z. Guo, H.Y. Zhang, J.Z. Xu, J.R. Shi, Thermochim. Acta 253 (1995) 243–251. [220] J. Alongi, M. Ciobanu, G. Malucelli, Carbohydr. Polym. 85 (2011) 599–608. [221] A. Cireli, N. Onar, M.F. Ebeoglugil, I. Kayatekin, B. Kutlu, O. Culha, E. Celik, J. Appl. Polym. Sci. 105 (2007) 3747–3750. [222] N. Yaman, Fibers Polym. 10 (2009) 413–418. [223] J. Alongi, C. Colleoni, G. Malucelli, G. Rosace, Polym. Degrad. Stab. 97 (2012) 1334–1340. [224] J. Alongi, C. Colleoni, G. Rosace, G. Malucelli, Cellulose 20 (2013) 525–535. [225] J. Alongi, C. Colleoni, G. Rosace, G. Malucelli, Polym. Degrad. Stab. 99 (2014) 92–98. [226] G. Brancatelli, C. Colleoni, M.R. Massafra, G. Rosace, Polym. Degrad. Stab. 96 (2011) 483–490. [227] J. Alongi, C. Colleoni, G. Rosace, G. Malucelli, J. Therm. Anal. Calorim. 110 (2012) 1207–1210. [228] J. Alongi, C. Colleoni, G. Rosace, G. Malucelli, Polym. Degrad. Stab. 98 (2013) 579–589. [229] A.R. Horrocks, Polym. Degrad. Stab. 96 (2011) 377–392. [230] A.R. Horrocks, Rev. Prog. Color. Relat. Top. 16 (1986) 62–101. [231] M. Lewin, Polym. Adv. Technol. 12 (2001) 215–222. [232] A.R. Horrocks, G. Smart, S. Nazare´, B. Kandola, D. Price, J. Fire Sci. 28 (2010) 217–248. [233] J. Alongi, G. Malucelli, Carbohydr. Polym. 90 (2012) 251–260. [234] J. Alongi, M. Ciobanu, G. Malucelli, Cellulose 18 (2011) 1335–1340. [235] W. Xing, G. Jie, L. Song, S. Hu, X. Lu, X. Wang, Y. Hu, Thermochim. Acta 513 (2011) 75–82. [236] S. Saka, H. Miyafuji, F. Tanno, J. Sol-Gel Sci. Technol. 20 (2011) 213–217. [237] S. Saka, M. Sasaki, M. Tanahashi, Mokuzai Gakkaishi 38 (1992) 1043–1050. [238] K. Ogiso, S. Saka, Mokuzai Gakkaishi 39 (1993) 301–307. [239] H. Miyafuji, S. Saka, A. Yamamoto, Holzforschung 52 (1998) 410–416. [240] M. Shabir Mahr, T. Hubert, M. Sabel, B. Schartel, H. Bahr, H. Militz, J. Mater. Sci. 47 (2012) 6849–6850. [241] M. Shabir Mahr, T. Hubert, B. Schartel, H. Bahr, M. Sabel, H. Militz, J. Sol-Gel Sci. Technol. 64 (2012) 452–464.