Simple and green synthesis of calcium alginate/hydroxyapatite hybrid material without high temperature treatment and its flame retardancy

Aiming to improve the thermal stability and flame retardant properties of calcium alginate, calcium alginate (CaAlg)/hydroxyapatite (HAP) hybrid material was synthesized in situ by the sol–gel method at room temperature. It was found that the HAP particles were generated from nanospheres to urchin-like microspheres, finally to wrinkled sponge microspheres with the two-dimensional nanostructured surface. The results showed that the thermal stability and flame retardancy of the hybrid material were significantly improved, and its mechanism was proposed as HAP promoted the decarboxylation of alginate resulting in carbonization of CaAlg/HAP, and it was the function of solid phase before 350 °C, while the synergistic effects of gas phase and solid phase after 350 °C that led to the high flame retardancy of CaAlg/HAP.


Introduction
Alginate is a marine biomass polysaccharide, which is composed of b-D-mannuronic acid (M blocks) and a-L-guluronic acid (G blocks) in an unspecified sequence Kabir et al. 2020). On account of its biodegradability, recyclability, low price and accessibility (Mokhena et al. 2020), alginate has been used in textile (Yu et al. 2020), decorative materials and other fields (Ahmad et al. 2021). However, some applications of alginate are restrained for its limited flammability (Xu et al. 2021). It has been found that sodium alginate (NaAlg, the most common alginate) can form polyanionic electrolytes under neutral and alkaline conditions. Therefore, it can cross-link with divalent cations (Liu et al. 2016b;Liu et al. 2015b;Liu et al. 2015c;Liu et al. 2016c) and trivalent cations (Liu et al. 2015a;Liu et al. 2016a) to form a stable ''egg-box'' structure microstructurally, and a sol-gel reaction visually (Hu et al. 2021), resulting in better flame retardant property than NaAlg. G block is the main factor for the formation of the ''egg-box'' (Hecht and Srebnik 2016), and M block is related to the mechanical properties of alginate .
In order to expand the application of alginate, researchers have applied CaAlg more frequently which can be prepared by NaAlg and calcium salt as it exhibits excellent flame retardancy (Yu et al. 2021b). However, CaAlg cannot fulfill the ideal thermal stability and flame retardancy, thus people try to prepare other hybrids to solve these defects (Xu et al. 2021). Recently, the synergistic effect between natural biomacromolecules and inorganic compounds has been a hot research direction for scientists. Compared with the complex and even harmful doping process of organics (Sun et al. 2021b), inorganics doping is simpler and more environmentally friendly . In our previous works, the flame retardant properties of CaAlg doped with inorganic substances such as calcium borate (Liu et al. 2018), Cu 2 O , Ag 3 PO 4 (Zhang et al. 2020), and AgCl ) have been studied.
Hydroxyapatite (Ca 10 (OH) 2 (PO 4 ) 6 , HAP), as inorganic phosphate, may effectively improve the flame retardancy of alginate for it possesses a very high decomposition temperature (above 1000°C) (Yang et al. 2018). Since HAP is a major component of human teeth and bones, its excellent biocompatibility and bioactivity make it a reliable material for synthetic teeth, artificial bones and drug carriers (Pluta et al. 2018;Sokolova et al. 2020). Now HAP with different sizes and morphologies can be prepared by various methods, exhibiting multiple properties and applications (Sadat-Shojai et al. 2013;Sarker et al. 2015;Sun et al. 2017). Researchers have obtained applicable materials by combining CaAlg and HAP, such as porous HAP/CaAlg composite beads for improving the cycling performance of pectinase (Qi et al. 2020), and ultralong HAP nanowires/CaAlg hybrids with excellent mechanical properties (Jiang et al. 2017), while flame retardancy and mechanism of these hybrids have not been reported yet.
Among the existing synthesis methods of HAP, hydrothermal and sol-gel methods are the common ones, while both of them require a high-temperature treatment (Costa et al. 2012). For example, the linear HAP prepared by hydrothermal method needed to be treated at 180°C for at least 12 h (Ma et al. 2017). In the sol-gel method, the HAP must be calcined at above 1000°C, or else, resulting in the formation of amorphous HAP crystals (Ruban Kumar and Kalainathan 2010). Other methods are more complex or time-consuming (Mohd Pu'ad et al. 2020;Yelten and Yilmaz 2016). However, due to the pyrolysis of CaAlg when heated, it is impossible to prepare the required hybrid materials directly and simply with phosphate by high temperature calcination or hydrothermal method. Furthermore, it would be more consistent with the concept of green chemistry if the hightemperature treatment was eliminated. Hence, the synthesis of HAP with practical value at room temperature is meaningful, although challenging.
In this direction, a green and simple sol-gel method without a high-temperature treatment was implemented to prepare the CaAlg/HAP hybrid material, and the dynamic characteristics of the consumed Ca 2? and the morphology of HAP particles were studied during the in-situ reaction. Furthermore, the thermal stability, flame retardant properties and its pyrolysis mechanism of CaAlg/HAP were explored, aiming to provide a new approach and theoretical basis to prepare alginate hybrids with high flame retardant properties. Preparations of CaAlg and CaAlg/HAP hybrid material 14.1425 g Na 2 HPO 4 Á12H 2 O was dissolved in 470 mL deionized water, and the pH was adjusted to 10 with 1 mol/L NaOH. Then 15 g NaAlg was added to obtain 500 mL 3% treated sol. It was placed in a 45°C water bath for 4 h, stirred per 20 min until NaAlg was dissolved completely, and a colorless transparent and bubble-free sol was obtained after cooling for 12 h at room temperature of 20°C (Tr). The pH of 3% calcium acetate solution was adjusted to 10-11 by 1 mol/L NaOH. The treated sol was poured into a certain rectangular container and soaked in the calcium acetate solution until the shape was fixed for 24 h, and finally, it was washed and dried to a constant weight. The treated sample was labeled as CaAlg/ HAP, and all processes were carried out at room temperature (Tr) except for the dissolution of NaAlg.

Materials
In order to explore the changes of HAP during its generative process, a series of investigations were carried out on the reaction of 20 g treated sol in 200 mL calcium acetate solution with the concentration of 0.6 wt.%, 1.0 wt.%, and 3.0 wt.% as time went on.
For comparison, the sample denoted as CaAlg was prepared as the same process as CaAlg/HAP, only excluding the steps related to Na 2 HPO 4 Á12H 2 O.
During the synthesis process of CaAlg/HAP hybrid material, sol-gel reaction and ion exchange reaction mentioned in Scheme 1 would occur simultaneously between the treated homogeneous sol and calcium acetate solution. Sol-gel reaction is a reaction between NaAlg and Ca 2? , in which NaAlg sol is cross-linked with Ca 2? to generate CaAlg gel. Ion exchange reaction refers to the ionic interaction between Ca 2? , OHand PO 4 3under a specific reaction condition, and generated HAP for its small solubility.

Characterization
The morphology of samples was characterized by scanning electron microscopy (SEM, FEI Quanta FEG 250, USA). The voltage was 15 kV and the samples were coated with gold before the test, in order to improve their conductivity.
The valence states of elements were analyzed by X-ray photoelectron spectroscopy (XPS, Thermo Fisher K-Alpha 250Xi, USA), with the operating voltage of 10 kV and current of 10 mA.
Fourier Transform infrared spectroscopy (FT-IR, Thermo Fisher-Nicolet 6700, USA) was recorded in the range of 4000-500 cm -1 , cooperating with a smart orbit attenuated total reflectance (ATR) accessory at room temperature and dry environment.
The samples were continuously scanned with Cu Ka filtered radiation (k = 0.15418 nm) at 2h with the range of 0 * 90°at 2°/min by an X-Ray Diffractomer (XRD, Bruker D8 Advance, Germany).
Inductively coupled plasma optical emission spectrometer (ICP, Agilent ICP-OES 730) was employed to quantitatively determine the content of Ca 2? and Na ? in samples, a series of work needed to be done before ion concentration was measured. 0.0328 ± 0.0003 g of samples were accurately weighed, and were placed into a Teflon beaker. Aqua regal of 6 mL, 0.5 mL hydrofluoric acid and 0.5 mL hydrogen peroxide were added and after 15 min, dissolved the samples by heating at 300°C for 45 min. After natural cooling, the samples were filtered, and diluted with deionized water to the desired concentration for use.
EDTA titration method was applied to determine the content of Ca 2? in calcium acetate solutions treated with different calcifications. One milliliter of calcification solution was diluted to 250 mL, and 25 mL solution from it was placed into a conical flask and the pH was adjusted to 9-10 by adding 1 mol/L sodium hydroxide. A small amount of chrome black T indicator was added to the conical flask and the solution was turned into wine red. EDTA solution of 0.01 mol/L was used to titrate, and when the wine red solution changed into pure blue, the titration end-point has arrived. The concentration of Ca 2? in the calcium acetate solutions and the reciprocal of Ca 2? consumption rate at a certain time point can be calculated according to the following formulas: c Ca 2þ ;t and c EDTA meant the concentration of Ca 2? and EDTA, V Ca 2þ and V EDTA means the volume of diluted calcium acetate and EDTA consumed, and t meant the certain point in time of the calcification reaction. Atmospheric water and CO 2 were corrected for more reliable results.

Flame retardant performance test
Thermogravimetric analysis (TGA, NETZS 209F3, Germany) of samples was measured from 30 to 900°C with a heating rate of 10°C/min in nitrogen and air respectively. The samples were pretreated for 12 h at the required temperature and humidity. The samples (size of 150 9 60 9 2 mm 3 ) were tested in a limiting oxygen index meter (LOI, LFY-606B, Shandong, China) according to ASTM D2863-1997.
The samples (size of 130 9 13 9 2 mm 3 ) were tested by a vertical burning tester (LFY-601A, Shandong, China) in vertical burning test, and the response to an open flame of a sample was determined by UL-94 in accordance with ASTM D3801-19a.
In a microscale combustion calorimeter (MCC, Govmark, Farmingdale, NY, USA), samples were heated to 750°C with 1°C/s in a nitrogen flow of 80 cm 3 /min. The result of HRR curve was processed a ''smoothing'' treatment for better visual convenience.
Cone calorimeter (CONE, FTT0242, East Grinstead, UK) operated with a heat flux of 50 kW/m 2 and samples in size of 100 9 100 9 2 mm 3 according to ISO 5660.
Thermogravimetry coupled with Fourier transform infrared spectroscopy (TG-FTIR) was conducted and samples were heated to 750°C at a heating rate of Scheme 1 Synthesis process of CaAlg/HAP hybrid material 20°C/min under N 2 atmosphere, with a flow rate of 50 mL/min by a simultaneous thermal analyzer (STA8000, PerkinElmer, England) and a Fourier transform infrared spectroscopy (SQ8 rontier, Perk-inElmer, England), in the range of 4000 to 450 cm -1 with a resolution of 2 cm -1 . The infrared absorption peak curves of the selected functional groups were smoothed.
Pyrolysis gas chromatography-mass spectrometry (Py-GC-MS) was performed by a thermal cracker (EGA/PY-3030D, Frontier, Korishan, Japan) and a gas chromatography/mass spectrometry (TRACE 1310-ISQLT, Thermo Fisher Scientific, USA). The pyrolysis temperatures in the furnace were set to 250°C, 450°C, and 750°C respectively. The samples were heated to 280°C at a heating rate of 10°C/ min, kept in the injector for 15 s, and then the temperature was rose at 20°C/ms.

Statistical analysis
Results were showed as the pattern of a mean value with standard deviation (mean ± S.D.), while each experiment and analysis were repeated three times.

Analysis of in situ formation of HAP
The leveling effect was considered to investigate the kinetic characteristics of Ca 2? , thus 3.0 wt.%, 1.0 wt.% and 0.6 wt.% of calcium acetate were used to explore the effect of the formation of HAP on Ca 2? consumption. The reciprocal of the rate of calcium consumption 1/v Ca 2þ were plotted intuitively over time, and the linear relations are fitted in Fig. 1a, in which y is for 1/v Ca 2þ and x is for time. The fitted lines overlapped in 3.0 wt.% and 1.0 wt.% CA, and four linear formulas are showed in Fig. 1a. The difference between formulas À and`indicated the treated sol consumed more Ca 2? with a faster rate in 0.6 wt.% CA, which agrees with the expectation that not only the cross-linking reaction between NaAlg and Ca 2? , but also the ion exchange reaction among Ca 2? , PO 4 3and OHto form HAP occurred. In addition, 0.81 g HAP was produced in 0.6 wt.% CA by calculation, and 0.89 g in 3.0 wt.% CA according to ICP in Table S1.
Faster reaction rate and higher yield implied that 3.0 wt.% CA is the best adoption in subsequent experiments.
The in-situ generation of hexagonal HAP was monitored by XRD (Fig. 1b) and SEM (Fig. 1c-h). CaAlg and HAP began to form as soon as calcification occurred, and the crystallinity of HAP might be confirmed after 10 min (Sinulingga et al. 2021). CaAlg has a diffraction peak around 23°in Fig. 1b and exhibits a smooth layered microstructure in Fig. 1c, while the treated sol generated CaAlg and HAP after calcification for 1 min, with a dense arrangement of spherical nanoparticles of a diameter about 2 nm on the surface of CaAlg (Fig. 1d). Then the nanosheets were grown in the HAP spherical particles in all directions, forming hierarchical microspheres with one-dimensional nanowires and twodimensional nanosheets. The urchin-like microspheres with average diameters of 3 lm, 4 lm and 6 lm were generated, when calcified for 5 min, 10 min and 30 min ( Fig. 1e-g), respectively. However, when calcified for more than 30 min, the size of HAP microspheres was narrowed. HAP microspheres became wrinkled sponge microspheres with an average diameter of 2 lm after 90 min (Fig. 1h). According to Ostwald's ''rule of stages'', HAP may tend to form unstable and more soluble phases in the early stage of reaction, and would transform into more stable phases in the later stage (Daryan et al. 2020). CaAlg/HAP semi-interpenetrating polymer network may more tightly contact for wrinkled sponge microspheres with a larger surface area than smooth microspheres, which was beneficial for HAP to adhere to the layered structure of CaAlg (Fong et al. 2017).

Flame retardant properties of CaAlg and CaAlg/ HAP
Thermal stability TGA was conducted under N 2 to explore the pyrolysis process without oxidation reaction, and both of the samples show a similar tendency with mainly three intense weight loss processes in Fig. 2a, b. The first stage of the weight loss below 200°C was mostly due to the removal of free water and bound water in the samples. The first similar weight loss processes suggested that the content of water and the combination with water in both samples are alike. The second stage (200-300°C) and third stage (300-400°C) were the primary and complex pyrolysis processes, including the breaking of glycosidic bonds, dehydration, decarboxylation, decarbonylation, etc ). According to Table. S2, CaAlg/HAP proceeded the second stage of pyrolysis at the lower temperature and less weight loss after 200°C occurred, resulting in more 11.58% residue than that of CaAlg within 900°C. Possibly, HAP promoted the formation of more stable char and less flammable small molecules during the thermal decomposition in N 2 , granting the hybrid material better flame retardancy.
As shown in Table. S1, HAP accounts for 0.974 wt.% in CaAlg/HAP, and no mass change of HAP could be assumed in TGA for it can exist stably below 1000°C (György et al. 2019). If HAP did not affect the thermal decomposition of CaAlg, the weight loss of CaAlg doped HAP calculated by the theoretical formula should be the same as the experimental CaAlg/HAP. The theoretical residue was calculated according to the following formula ): where m cal , m CaAlg and m HAP represented the theoretical char of CaAlg/HAP, experimental char of CaAlg, and char of HAP (calculated as 100%), respectively, and wt HAP was the mass proportion of HAP in CaAlg/ HAP. The obtained theoretical TG curve (blue curve in Fig. 2a) was just a little bit above that of CaAlg. It showed that HAP promoted the carbonization of CaAlg during thermal degradation, possibly because of the thermal insulation of HAP (Liu et al. 2018). The pyrolysis of samples in the air is supposed to be more complex -as oxidation reactions occur between O 2 and the products of CaAlg, leading to a further loss of weight . As seen in Fig. 2, CaAlg before 402°C and CaAlg/HAP before 384°C in both atmospheres were in consistence. However, the oxygen subsequently reacted with the combustible gas born from previous stage, causing one more obvious pyrolysis in Fig. 2c-d. As shown in Table. S2, CaAlg/ HAP possessed a lower R max and more 15% residue than that of CaAlg in air, and its residual mass remained unchanging earlier than CaAlg. It indicated that HAP can prevent oxygen from participating in the thermal decomposition of alginate at high temperature, possibly because the hybrid material can generate more stable char at a lower temperature, resulting in a markedly improved thermal stability of CaAlg/HAP, compared with that of CaAlg.

SEM and XRD of CaAlg and CaAlg/HAP after calcination in air
According to TGA, CaAlg and CaAlg/HAP were calcined in air at 250°C, 450°C and 750°C for 1 h to explore the morphological and compositional changes of both samples. Figure 3a-c show that the morphology of CaAlg was slightly irregular deformation when heated at 250°C and 450°C, while at 750°C, a densely arranged residue structure appeared. However, CaAlg/HAP was significantly different from CaAlg. When heated at 250°C, the diameter of HAP microparticles started to narrow to about 1-2 lm, and the characteristic wrinkled sponge was retained, as shown in Fig. 3d. While it turned to less than 1 lm at 450°C, there were many nanopores on the compact surface of the residue which can be observed in Fig. 3e. The XRD in Fig. 3h demonstrated that these nanopores formed due to the generation and discharge of CO 2 from the decomposition of CaCO 3 . As seen in Fig. 3f, when CaAlg/HAP was heated at 750°C, the size of HAP particles continued to shrink to the diameter of 20 nm and the nanoparticles were closely arranged on the surface of CaAlg, indicating that the compact surface structure prevented the thermal degradation of CaAlg (György et al. 2019).
As observed in Fig. 3g, there was no CaCO 3 formed in CaAlg until about 450°C, and a little CaCO 3 was decomposed into CO 2 and Ca(OH) 2 at 750°C. While as seen in Fig. 3h, the generation of CaCO 3 occurred in CaAlg/HAP when heated at 250°C, and possibly the inner layer of alginate was covered and protected by CaCO 3 . At 450°C, CaCO 3 cracked to produce CaO and CO 2 , and when the temperature was raised to 750°C, the crystallinity of CaO increased, indicating that more CaO formed. It showed that HAP can Fig. 3 SEM images of (a-c) CaAlg and (d-f) CaAlg/HAP after being treated in air at 250°C, 450°C, and 750°C for 1 h, respectively. XRD patterns of (g) CaAlg and (h) CaAlg/HAP after treated in air at 250°C, 450°C, and 750°C for 1 h, respectively effectively promote the carbonization of CaAlg during combustion, while HAP crystal was undecomposed and a higher crystallinity emerged at a higher temperature.
In terms of the above, HAP with ultra-high temperature resistance may absorb heat to form smaller particles that were tightly packed onto the surface of CaAlg. What's more, HAP promoted faster carbonization of CaAlg and produced stable CaCO 3 at around 250°C, and CaO at around 450°C or even higher temperature, which prevented further pyrolysis, and thus improved the flame retardant performance of the hybrid material.

Combustion properties
As it can be seen from Table 1, the LOI of CaAlg/HAP was 67%, which was 20.5% higher than that of CaAlg . Both samples cannot be ignited in the UL-94 test in 10 s and maintained their initial shape without dripping, which are classified as V-0. The EHC of CaAlg was 5.98 MJ/kg, while that of CaAlg/ HAP was 12.3 MJ/kg. The EHC of the hybrid material was higher. In general, lower EHC indicates better flame retardancy, while it may be based on the consistent mass loss. EHC represents the ratio of the heat release rate to the mass loss rate at a certain time point. The results of TGA show that the mass loss of CaAlg/HAP was about 20% less than that of CaAlg, and the heat release rate in the CONE test was also lower. Therefore, the results of EHC may indicate that the mass loss of CaAlg/HAP was much less than that of CaAlg. PHRR and THR in MCC were significantly reduced, and the residue in CONE and MCC of the samples showed the same tendency with TGA, as displayed in Fig. 2. All these results indicated that HAP was beneficial for the improvement of the thermal stability and flame retardant properties of CaAlg. Figure 4a shows the variation of the total smoke release (TSR) from the samples during the CONE test over time. The TSR of CaAlg/HAP was remarkably less than that of CaAlg, possibly because the more char the hybrid material produced, the less smoke containing toxic components was released. Importantly, less smoke release can reduce casualties in fire (Kong et al. 2021;Yu et al. 2021a).
Since most of the oxygen may be consumed in real fire cases, the pyrolysis of samples under an anaerobic condition can be studied via MCC. MCC simulates heat release throughout the decomposition of samples in nitrogen (Mensah et al. 2018), thus it may correspond to the TGA in N 2 theoretically. As seen in Figs. 2b and 4b, the third weight loss stage in TGA at 300-400°C corresponded to the first heat release stage in MCC, while the TGA curves at 400-600°C was associated with the second heat release stage with little weight change, suggesting that the chemical energy was converted into heat at 300-600°C. The less weight-loss and HRR of CaAlg/HAP implied that the pyrolysis of hybrid was effectively prevented. There were two potential reasons for the decrease of HRR (Liu et al. 2016b). One was that HAP may catalyze the formation of stable char from CaAlg and contribute to a reduction of combustible gas release, as confirmed by the results of TGA (Fig. 2) and XRD measurements ( Fig. 3g-h). The other one was that HAP could facilitate the formation of non-flammable gases from alginate such as CO 2 and water, which would dilute the concentration of flammable products.

TG-FTIR
In order to investigate whether the improvement of flame retardant properties is related to the synergistic effect of both aforementioned reasons, TG-FTIR was conducted to explore the generation of CO 2 and H 2 O during the whole pyrolysis process. The main functional groups containing carbonyl group C = O at 1795 cm -1 , carbon monoxide CO at 2182 cm -1 , carbon dioxide CO 2 at 2362 cm -1 , -CH at 2962 cm -1 , = CH at 3018 cm -1 , and H 2 O at 3732 cm -1 from the 3D images of TG-FTIR were selected to explore the pyrolysis of CaAlg (Fig. 5a) and CaAlg/HAP (Fig. 5b). It was found that C = O, CO and CO 2 showed a similar generation trend, and all of them were produced apparently above 250°C, indicating that the pyrolysis of glycosidic bonds started and a variety of complex reactions occurred ( Fig. 5c-e). What's more, CaAlg/HAP generated more C = O, CO and CO 2 than CaAlg, especially over 350°C, reflecting that the carboxyl groups or similar groups in CaAlg/HAP were more reactive when heated, and more decarboxylation and decarbonylation could be activated. The -CH and = CH were markedly generated around 450°C, as shown in Fig. 5f, g, suggesting that the macromolecular residues from the samples went on a further crack into smaller chains. As seen in Fig. 2h, the water was formed throughout the pyrolysis process periodically. Not only the amount of -CH and = CH but also H 2 O were decreased in CaAlg/HAP, implying that its degree of pyrolysis was significantly reduced. The above functional groups of CaAlg/HAP exhibited a similar trend as CaAlg, however, CaAlg/HAP produced more CO 2 above 350°C and the degree of pyrolysis in the later stage was alleviated. Therefore, it can be inferred that, for the carbonization mechanism of CaAlg/HAP, HAP promoted the decarboxylation of alginate, probably because the interfacial sites between HAP and alginate were conducive to the occurrence of decarboxylation (Fu and Mei 2021). Since HAP would not produce phosphoric acid when heated, its flame-retardant mechanism is not similar to most of the phosphates that generate phosphoric acid during heating (Sun et al. 2021a).
Collectively, it can be confirmed that the increasing release of CO 2 diluted the combustible gas in the CaAlg/HAP combustion system, and turned to isolation among air, fuel and heat. In general, it was the combination of the production of more stable char earlier and generation of more CO 2 over 350°C together that improved the flame retardancy of the hybrid material. The high flame retardancy of CaAlg/ HAP can be mainly attributed to the synergistic effects of the gas phase and condensed phase.

Py-GC-MS
According to the curves in Fig. 6 and the data in Table S3-4, when the temperature rose from room temperature to 250°C, CaAlg/HAP produced less CO 2 than the CaAlg, indicating that the pyrolysis degree of CaAlg/HAP was lower in the early stage of heating, and the alginate was effectively protected before 250°C. However, when the temperature rose to 450°C, CaAlg/HAP generated more CO 2 than CaAlg, and as the temperature continued to rise to 750°C, there was a further increase of CO 2 , which was in line with more generation of CO 2 from 350°C in Fig. 5d. The types and amount of pyrolytic products produced by CaAlg/HAP were less than CaAlg. Such CaAlg/ HAP produced less acetic acid than CaAlg significantly at 450°C, and it produced fewer kinds of products than CaAlg at 750°C. It was proved that the flame retardant mechanism of CaAlg/HAP was various at different stages of pyrolysis. The function of the solid-phase accounted for the main flame retardant mechanism from room temperature to 350°C, while the synergistic effect of the solid phase and the gas phase worked at 350-750°C.

Flame-retardant mechanism
The speculative pyrolysis mechanism of CaAlg/HAP with no oxygen involved is presented in Scheme 2, which was extrapolated from the pyrolytic products listed in Table S3 (Xu et al. 2021). According to the morphology and XRD patterns of CaAlg/HAP calcined at 250°C, 450°C, and 750°C (Fig. 2), micronsized HAP particles with low energy changed into nano-sized particles with high energy, because of gaining heat. Therefore, during the whole pyrolysis process, part of the heat was received by CaAlg/HAP Fig. 5 3D images (a, b) and main functional groups as a function of temperature (c-h) of CaAlg and CaAlg/HAP obtained from TG-FTIR to obtain a compact structure on the surface of CaAlg, which can delay he thermal decomposition to some extent.
As shown in Fig. S3, hydrogen bonds may be formed between HAP and M blocks. The structure may be ''O À H Á Á Á O'', which is formed between À COOH or À OH in CaAlg and electronegative oxygen atoms of PO 4 3in HAP, and ''Á Á Á'' indicates the newly formed hydrogen bond, which could improve the thermal stability of M blocks. The interfacial site between HAP and alginate activated the decarboxylation of ''egg-box'' structure consisting of G blocks and Ca 2? , which resulted in the generation of CaCO 3 , CaO and CO 2 to protect alginate when heated. Dehydration and slight decarboxylation occurred in alginate below 250°C, thus CO 2 and H 2 O were the main products. Less CO 2 was produced from CaAlg/ HAP below 250°C, indicating that the formation of CaCO 3 protected CaAlg and reduced pyrolysis. The glycosidic bonds inside CaAlg can be broken into M and G blocks above 250°C. M blocks further cracked into butanedione, acetone and acetic acid in CaAlg as a general rule, while the amount of these products was significantly reduced in CaAlg/HAP, indicating that less pyrolysis occurred in M blocks. However, although CO 2 accounted for 92.65% and other components were relatively rare at 750°C, a few pyrolytic products from G blocks such as acetaldehyde, 2-methylfuran and furfural can be found from CaAlg/HAP. These results suggested that HAP protected M blocks by forming hydrogen bonds with them, and promoted the partial decomposition of the ''egg-box'' structure to produce CaCO 3 and CaO to enhance the thickness and strength of the carbon layer as well. Moreover, CO 2 from that was to dilute the flammable gas and prevented further decomposition of alginate at high temperature. The combustion process of CaAlg/HAP in the air is exhibited in Fig. 7, and more CO 2 was produced than in N 2 due to the oxidation reactions.

Conclusion
In this report, CaAlg/HAP hybrid material was in situ synthesized by a green and simple sol-gel method at room temperature. During the reaction, the HAP spherical particles were grown from nanospheres to urchin-like microspheres, and finally to wrinkled sponge microspheres which were favorable as binding with layered CaAlg for their large surface area. CaAlg/ HAP presented remarkably better thermal stability and flame retardant properties, in comparison with CaAlg, as it had more residue (above 12%) in TGA at 900°C, and its LOI reached 67%, while HRR and TSR were significantly decreased. Furthermore, the flame retardant mechanism of CaAlg/HAP was proposed based on experimental data and reliable theories. The wrinkled sponge microspheres absorbed energy and transformed into nanoparticles after heating, which alleviated pyrolysis of CaAlg for sharing heat from the external environment. The HAP particles with ultrahigh temperature resistance covered the surface of CaAlg, which also hindered the further decomposition of alginate. What's more, the hydrogen bonds formed by the M blocks and HAP enhanced the stability of the M blocks. The interfacial sites between HAP and alginate may facilitate the decarboxylation of the ''egg-box'' structure made up of G blocks and Ca 2? , resulting in the production of CaCO 3 early at 250°C, and the decarboxylation of CaCO 3 at higher temperature to produce more CaO and CO 2 . In summary, HAP promoted decarboxylation of alginate causing carbonization of CaAlg/HAP, and mainly CaCO 3 , HAP, and char were generated below 350°C in the solid phases, while synergistic effects of gas-phase CO 2 and solid phases CaO, HAP, and char over 350°C that improved thermal stability and flame retardancy of CaAlg/HAP by restraining the transfer among oxygen, fuel and heat.
Scheme 2 Proposed pyrolysis mechanisms of CaAlg/HAP