Flame-Retardant Wood Composites Based on Immobilizing with Chitosan/Sodium Phytate/Nano-TiO2-ZnO Coatings via Layer-by-Layer Self-Assembly

Composite coatings of inorganic nanomaterials with polyelectrolytes are promising materials for wood modification. Endowing wood with flame retardancy behavior can not only broaden the range of applications of wood, but also improve the safety of wood products. In this work, chitosan/sodium phytate/TiO2-ZnO nanoparticle (CH/SP/nano-TiO2-ZnO) composite coatings were coated on wood surface through layer-by-layer self-assembly. The morphology and chemical composition of the modified wood samples were analyzed using scanning electron microscopy and energy dispersive spectrometry. The thermal degradation properties and flame retardancy of the samples treated with different assembly structures were observed by thermogravimetric analysis, limiting oxygen test, and combustion test. Due to the presence of an effective intumescent flame retardant system and a physical barrier, the CH/SP/nano-TiO2-ZnO coatings exhibited the best flame retardant performance and required only approximately six seconds for self-extinguishing. The coated samples had a limiting oxygen index of 8.4% greater than the original wood.


Introduction
In recent years, the modification of wood surfaces with inorganic nanomaterials in order to improve or create new functional features has attracted intense research interests [1,2]. With the development of surface modification techniques based on layer-by-layer (LbL) self-assembly, the introduction of metal nanoparticles into polyelectrolyte multilayer coatings has proven to be feasible and promising [3][4][5]. The LbL self-assembly approach has been widely adopted in many fields for surface functionalization due to its advantages such as low cost, simple process, controllable coating thickness, and absence of limitations on the substrate size and shape [6,7]. Moreover, since Renneckar first found that polycations can be effectively attached to a wood surface by LbL self-assembly [8], it has been applied successfully in wood surface modifications, with many recent reports on the modification of wood with different materials via the LbL self-assembly process [9][10][11].
While many studies have investigated a polyelectrolyte/nanomaterials composite membrane, most of these have focused either on the fabrication of nanomaterials or on the performance of modified materials; therefore, a systematic study of the functional mechanism or the interaction between the different phases of the composite coatings has been lacking. In particular, due to its surface porosity and chemical heterogeneity, there has been relatively little research in this area for wood [12]. Furthermore, most of the studies performed to date have focused on nanocomposite coatings while of the wood. The mechanism for these changes related to the individual components is discussed in detail below.

Wood Substrate and Solution Preparation
Wood samples were cut into specimens with dimensions of 10 × 10 × 1 mm 3 and 120 × 10 × 4 mm 3 for characterization and evaluation of flame retardancy. Following that, the samples were rinsed twice with distilled water and dried in an oven at 60 • C for 1 h. They were then sorted into several groups and were weighed and measured prior to processing. The entire coating fabrication process was carried out at room temperature and pressure.
For the preparation of a mixed nano-TiO 2 and nano-ZnO solution, zinc oxide nanoparticles (1 g) and titanium oxide nanoparticles (1 g) at a mass ratio of 1:1 were added to distilled water (200 mL) under magnetic stirring, and the pH of the solution was adjusted to 2-3 using hydrochloric acid. The nano-TiO 2 -ZnO solution was then positively charged by adjusting the pH below its isoelectric point. Solutions of pure nano-TiO 2 (positively-charged), pure nano-ZnO (positively-charged), chitosan (positively-charged), and sodium phytate (negatively-charged) were prepared, respectively, using the same procedure as that used for the preparation of the mixed nano-TiO 2 and nano-ZnO solutions. The concentration of each solution was set at 1%. Finally, the solutions were divided into groups for coating fabrication as described below.

Formation of Flame Retardant Coatings on Wood Surfaces via Layer-by-Layer Self-Assembly
In this experiment, the layer-by-layer self-assembly technique was used with mutual adsorption of polycation, polyanion, and metal oxide nanoparticles leading to the formation of composite coatings. In the deposition process, firstly, each set of negatively charged wood specimens was immersed in the corresponding positively charged chitosan solution for 90 min. The surface would then have a positive charge as the polycation ionized by the chitosan adsorbed onto the wood surface by electrostatic forces. Next, alternating negative and positive layers of sodium phytate and (or) nano-TiO 2 -ZnO were deposited on the wood surface by immersing the specimens in the corresponding groups of solutions for 90 min, followed by rinsing twice with distilled water and 1-hour oven-drying between each layer. One layer of sodium phytate and one layer of nano-TiO 2 -ZnO were considered to comprise a single deposition cycle. To ensure uniform coverage of the substrate, ten deposition cycles were used for each specimen ( Figure 1). Thus, the as-prepared samples named CH/SP/nano-TiO 2 -ZnO were obtained. For other samples as listed in Table 1, the nano-TiO 2 -ZnO layer was replaced by pure nano-TiO 2 or pure nano-ZnO, and the obtained samples were named CH/SP/nano-TiO 2 sample and CH/SP/nano-ZnO sample, respectively. The wood treated with chitosan only, while the other conditions remain unchanged was called the CH sample. The wood treated with CH and SP was named the CH/SP sample.

Characterization
The surface morphology of the CH/SP/nano-TiO2-ZnO coatings before and after wear, as well as their chemical compositions were characterized by scanning electron microscopy (SEM, Quanta 400 FEG, FEI Inc., Eindhoven, The Netherlands) equipped with an energy-dispersive X-ray spectroscopy (EDS) system. Each sample was coated with a 5-10 nm Au layer prior to SEM imaging. The mass change of the samples before and after the treatment was also recorded.

Thermogravimetric Analysis
Thermogravimetric analysis (TG209, Netzsch, Selb, Germany) under nitrogen atmosphere was performed to measure the thermal stability of the wood samples (0.5 mm thickness) before and after different kinds of treatment. The heating rate was 10 °C/min, and the temperature range was 30-800 °C.

Limiting Oxygen Index Test
The limiting oxygen index testing was conducted by oxygen index testing apparatus (PX-01-005, Phinix, Suzhou, China) according to standard GB2406. 2-2009 [38] with 150 × (6.5 ± 0.5) × (3.0 ± 0.5) mm 3 wood samples. The apparatus was set to standard conditions and the sample was placed vertically in the mixture of oxygen and nitrogen that had been adjusted to a specific value. After the airflow was stabilized, the sample was ignited. The minimum oxygen concentration required for the sample to maintain steady combustion for three minutes at 5 cm of the sample was identified as its oxygen index, expressed as a volume percent. Scheme of fabricating flame-retardant coating on a wood surface by layer-by-layer self-assembly.

Characterization
The surface morphology of the CH/SP/nano-TiO 2 -ZnO coatings before and after wear, as well as their chemical compositions were characterized by scanning electron microscopy (SEM, Quanta 400 FEG, FEI Inc., Eindhoven, The Netherlands) equipped with an energy-dispersive X-ray spectroscopy (EDS) system. Each sample was coated with a 5-10 nm Au layer prior to SEM imaging. The mass change of the samples before and after the treatment was also recorded.

Thermogravimetric Analysis
Thermogravimetric analysis (TG209, Netzsch, Selb, Germany) under nitrogen atmosphere was performed to measure the thermal stability of the wood samples (0.5 mm thickness) before and after different kinds of treatment. The heating rate was 10 • C/min, and the temperature range was 30-800 • C.

Limiting Oxygen Index Test
The limiting oxygen index testing was conducted by oxygen index testing apparatus (PX-01-005, Phinix, Suzhou, China) according to standard GB2406. 2-2009 [38] with 150 × (6.5 ± 0.5) × (3.0 ± 0.5) mm 3 wood samples. The apparatus was set to standard conditions and the sample was placed vertically in the mixture of oxygen and nitrogen that had been adjusted to a specific value. After the airflow was stabilized, the sample was ignited. The minimum oxygen concentration required for the sample to maintain steady combustion for three minutes at 5 cm of the sample was identified as its oxygen index, expressed as a volume percent.

Combustion Test
Combustion testing was carried out by direct exposure to an alcohol lamp flame. Specifically, flame retardant trends were observed by exposing the wood samples (120 × 10 × 4 mm 3 ) to the outer flame of the alcohol lamp, recording the combustion state and the time required for the flame to be completely extinguished. The entire burning process of every wood sample after ignition was videotaped for backup observation. After observing and comparing the combustion of all samples, we selected the most typical sample from each group as the representative to be discussed below. Each group had five replicates.

Surface Morphologies Analysis
In order to verify that the loading of films would not affect the wood's macroscopic characteristics, the weight of the coated wood samples before and after modification was recorded in Table 2. It is presented that the weight of all specimens after coating was hardly changed or only increased slightly. Figure

Combustion Test
Combustion testing was carried out by direct exposure to an alcohol lamp flame. Specifically, flame retardant trends were observed by exposing the wood samples (120 × 10 × 4 mm 3 ) to the outer flame of the alcohol lamp, recording the combustion state and the time required for the flame to be completely extinguished. The entire burning process of every wood sample after ignition was videotaped for backup observation. After observing and comparing the combustion of all samples, we selected the most typical sample from each group as the representative to be discussed below. Each group had five replicates.

Surface Morphologies Analysis
In order to verify that the loading of films would not affect the wood's macroscopic characteristics, the weight of the coated wood samples before and after modification was recorded in Table 2. It is presented that the weight of all specimens after coating was hardly changed or only increased slightly.

Thermal Stability and Fire Resistance Tests
To reveal the thermal degradation of uncoated and coated wood, TGA and DTG (Derivative Thermogravimetry) were carried out, and the curves are shown in Figure 3. With the introduction of the coatings, the thermal decomposition of wood changed significantly. The pyrolysis of the samples can be characterized by three stages. Before 250 • C, it is the drying and pre-decomposition stage. In this stage, the formed phosphoric acid and polyphosphoric acid produced by sodium phytate had a catalyzed effect on the thermal degradation of the wood. It resulted that the weight loss rates of the treated samples were higher than that of the uncoated wood, indicating that the coatings promoted the carbonization and dehydration of the material [39]. In addition, the treated samples came into the second stage at a lower temperature than the uncoated one, that is, holocellulose (cellulose and remaining hemicellulose) underwent violent pyrolysis at about 250 to 400 • C. In the third stage, at 450-800 • C, the residual weight of all samples remained stable. The residual weight rate of treated samples were higher than that of the uncoated one, and the samples in order of descending residual weight were: CH/SP/nano-TiO 2 -ZnO (red line), CH/SP/TiO 2 (blue line), CH/SP/nano-ZnO (yellow line), CH/SP (purple line), CH (green line), and uncoated wood (black line).
Coatings 2020, 10, x FOR PEER REVIEW 6 of 11 Table 2. Initial mass and mass change after treatment of coated wood sample.

Thermal Stability and Fire Resistance Tests
To reveal the thermal degradation of uncoated and coated wood, TGA and DTG (Derivative Thermogravimetry) were carried out, and the curves are shown in Figure 3. With the introduction of the coatings, the thermal decomposition of wood changed significantly. The pyrolysis of the samples can be characterized by three stages. Before 250 °C, it is the drying and pre-decomposition stage. In this stage, the formed phosphoric acid and polyphosphoric acid produced by sodium phytate had a catalyzed effect on the thermal degradation of the wood. It resulted that the weight loss rates of the treated samples were higher than that of the uncoated wood, indicating that the coatings promoted the carbonization and dehydration of the material [39]. In addition, the treated samples came into the second stage at a lower temperature than the uncoated one, that is, holocellulose (cellulose and remaining hemicellulose) underwent violent pyrolysis at about 250 to 400 °C. In the third stage, at 450-800 °C, the residual weight of all samples remained stable. The residual weight rate of treated samples were higher than that of the uncoated one, and the samples in order of descending residual weight were: CH/SP/nano-TiO2-ZnO (red line), CH/SP/TiO2 (blue line), CH/SP/nano-ZnO (yellow line), CH/SP (purple line), CH (green line), and uncoated wood (black line). The flame retardant properties of all the samples were evaluated by limiting oxygen index (LOI) tests. A high LOI value corresponded to a high oxygen content required for burning to occur, which meant a high flame retardancy. According to the Figure 4, the LOI values increased as the composition increased. It is increased dramatically to 33% after being coated with CH/SP/nano-TiO2-ZnO films, which suggested significant improvement in the sample's flame retardancy. In order to investigate the effect of individual components, LOI results of different assembly structure samples can be clearly seen from the bars below. The CH/SP/nano-TiO2 coated sample possessed a higher LOI value of 32.8% compared with the CH/SP/nano-ZnO coated sample's LOI of only 29.5%, which indicated that nano-TiO2 played a major role in flame retardancy. Overall, the results demonstrated the fire resistance of the coatings. The flame retardant properties of all the samples were evaluated by limiting oxygen index (LOI) tests. A high LOI value corresponded to a high oxygen content required for burning to occur, which meant a high flame retardancy. According to the Figure 4, the LOI values increased as the composition increased. It is increased dramatically to 33% after being coated with CH/SP/nano-TiO 2 -ZnO films, which suggested significant improvement in the sample's flame retardancy. In order to investigate the effect of individual components, LOI results of different assembly structure samples can be clearly seen from the bars below. The CH/SP/nano-TiO 2 coated sample possessed a higher LOI value of 32.8% compared with the CH/SP/nano-ZnO coated sample's LOI Coatings 2020, 10, 296 7 of 12 of only 29.5%, which indicated that nano-TiO 2 played a major role in flame retardancy. Overall, the results demonstrated the fire resistance of the coatings. The general flame retardancy of the original wood and five groups of samples (Table 1) was assessed by the direct exposure to an alcohol lamp flame. The process of flame initiation, spreading, and self-extinguishing time were recorded for the samples treated with different components and are shown in Figure 5. It is clear from Figure 5(a1,a2) that once the original wood was ignited, it burned quickly and fiercely with no indication of dying out, which is very reasonable since wood is known to be a flammable material. In particular, the wood used in this experiment was loose and had large pores, making it more flammable. Thus, eventually the whole strip of wood was burned out. By contrast, the coated wood burned slowly and was extinguished naturally after a time of 6-30 s, which was consistent with the results achieved above ( Figure 5(b2,c2,d2,e2,f2)), that is, with a shorter extinguishing time indicating a better flame retardant property. Table 3 lists the data description of each group and the results of statistical analysis (one-way ANOVA). It is obvious that P-value < 0.05, indicating that different assembly structures did have a great impact on wood flame retardancy.  The general flame retardancy of the original wood and five groups of samples (Table 1) was assessed by the direct exposure to an alcohol lamp flame. The process of flame initiation, spreading, and self-extinguishing time were recorded for the samples treated with different components and are shown in Figure 5. It is clear from Figure 5(a1,a2) that once the original wood was ignited, it burned quickly and fiercely with no indication of dying out, which is very reasonable since wood is known to be a flammable material. In particular, the wood used in this experiment was loose and had large pores, making it more flammable. Thus, eventually the whole strip of wood was burned out. By contrast, the coated wood burned slowly and was extinguished naturally after a time of 6-30 s, which was consistent with the results achieved above ( Figure 5(b2,c2,d2,e2,f2)), that is, with a shorter extinguishing time indicating a better flame retardant property. Table 3 lists the data description of each group and the results of statistical analysis (one-way ANOVA). It is obvious that p-value < 0.05, indicating that different assembly structures did have a great impact on wood flame retardancy.
As observed from Figure 5(a1,b1,c1,d1,e1,f1), the burning flame of the coated wood samples is much smaller than that of the original wood within 6 seconds of ignition. The fire of the sample coated with chitosan extended only slightly from the ignition point, as observed from Figure 5(b1,b2). It was revealed that the chitosan layer was barely effective for obtaining flame retardancy, which is consistent with the limiting oxygen index in Figure 4. The flame-retardant properties improved after the addition of the sodium phytate layer. This was because the nitrogen in chitosan and the phosphorus in sodium phytate constitute a 'nitrogen-phosphorus' flame retardant system. During thermal degradation, the generated phosphoric acid and metaphosphoric acid dehydrated strongly into charcoal, leading to the appearance of a carbon protection layer quickly after burning. These two materials have been proven to likely form an intumescent flame retardant system that can help trap air and heat [40]. In this system, cellulose and chitosan serve as carbon sources, while sodium phytate acts as an acid source and -NH contained in chitosan makes it act as a blowing agent. The char formation in this system not only prevents continued burning but also provides a certain mechanical support, which is of vital importance in wooden parts due to the severe dangers of collapse in a burning wood structure.
pores, making it more flammable. Thus, eventually the whole strip of wood was burned out. By contrast, the coated wood burned slowly and was extinguished naturally after a time of 6-30 s, which was consistent with the results achieved above ( Figure 5(b2,c2,d2,e2,f2)), that is, with a shorter extinguishing time indicating a better flame retardant property. Table 3 lists the data description of each group and the results of statistical analysis (one-way ANOVA). It is obvious that P-value < 0.05, indicating that different assembly structures did have a great impact on wood flame retardancy.   As shown in Figure 5(d2,e2,f2), upon the introduction of nanoparticles, the wood required a relatively shorter time of 6-20 seconds to extinguish the flame, demonstrating better flame retardant performance. It is well-known that both nano-TiO 2 and nano-ZnO provide high stability, high melting point, and good thermal stability, and their flame retardant effects have already been discussed and applied. For this multilayer composite, the immobilized nanoparticle layer further improved the flame retardant properties [17,41]. The metal oxides can significantly affect the pyrolysis process of the intumescent system and improve the fire resistance of the film. Physically, the dense cover of the nanoparticles on the wood surface also hindered the ignition and spread of combustion. The flame retardant mechanism of this nanocomposite film is illustrated in Figure 6.
Due to both physical and chemical interactions, the coated wood obtained in this work showed flame retardant performance. Among all of the samples examined in our experiments, the CH/SP/nano-TiO 2 -ZnO sample showed the best performance. As observed from Figure 5f, only approximately 6 seconds were required for the flame to be extinguished, the burning was difficult to maintain, and even ignition was difficult. It had a maximum limiting oxygen index of 8.4% greater than the original wood. Markedly, the nano-TiO 2 provided a stronger flame retardant behavior than nano-ZnO. With the introduction of nano-TiO 2 , both of the self-extinguishing time and ultimate oxygen index deviated greatly from before. According to all the results obtained above, the CH/SP/nano-TiO 2 -ZnO system had an excellent flame retardant property, and the wood showed an enhanced performance with the immobilization of the coatings. Nevertheless, it is still necessary to further optimize the experimental protocol and film parameters, because nowadays it is of commercial interest to develop natural, sustainable, renewable, and functional films. It is also important to point out that the materials investigated in our work have many other superior properties, and may have thus endowed the composite film with UV resistance, antibacterial behavior, and wettability.
Coatings 2020, 10, x FOR PEER REVIEW  9 of 11 properties, and may have thus endowed the composite film with UV resistance, antibacterial behavior, and wettability.

Conclusions
A novel chitosan/sodium phytate/nano-TiO2-ZnO flame retardant system was coated on the wood surface using the layer-by-layer assembly method. Considering all of the above-described results, we conclude that the composite films had a dramatic flame retardancy effect on the wood, leading to higher thermal stability. The higher limiting oxygen index of coated wood showed fire resistant property with a self-extinguishing time of around 6 seconds in combustion tests. According to the research of different component assembly structures, it has been proven that an effective combined effect between the constituents of the coating gave rise to the superior flame retardancy of the coatings, that is, the presence of an effective intumescent flame retardant system and a physical barrier. The use of these coatings was a promising approach for wood modification and for endowing other materials such as fabrics with flame retardancy. Since the impact of material modification on the environment has been paid more and more attention, this new environmentally friendly flame retardant composite film on wood surface could have significant importance in academic research and practical application.

Conclusions
A novel chitosan/sodium phytate/nano-TiO 2 -ZnO flame retardant system was coated on the wood surface using the layer-by-layer assembly method. Considering all of the above-described results, we conclude that the composite films had a dramatic flame retardancy effect on the wood, leading to higher thermal stability. The higher limiting oxygen index of coated wood showed fire resistant property with a self-extinguishing time of around 6 seconds in combustion tests. According to the research of different component assembly structures, it has been proven that an effective combined effect between the constituents of the coating gave rise to the superior flame retardancy of the coatings, that is, the presence of an effective intumescent flame retardant system and a physical barrier. The use of these coatings was a promising approach for wood modification and for endowing other materials such as fabrics with flame retardancy. Since the impact of material modification on the environment has been paid more and more attention, this new environmentally friendly flame retardant composite film on wood surface could have significant importance in academic research and practical application.