Structure and morphology analysis of LFPN
TEM and SEM-EDS were applied to characterize the structure and morphology of LFPN, which was shown in Fig. 2. As shown in Fig. 2A, there were many nanometer particles, which were irregular sphere shapes, assembled around a certain structure became a cluster. In order to further explore the internal form of the cluster, TEM was carried out to get a 2D image of LFPN (Fig. 2b, c). Some rod nanofibres can be clearly observed with the width of 50–105 nm, and were coated with a lot of nanometer particles. It can be seen from the Fig. 2c, irregular spherical particles were uniformly scattered on its surface, forming clusters. The phenomenon explains the aggregation of nanometer particles in the SEM image of LFPN. Figure 2d shows the chemical element distribution of LFPN by SEM-EDS image. The distribution of carbon elements showed a rod-like structure like TEM images, and phosphorus and nitrogen elements distributed evenly around carbon. Therefore, the possible structure of LFPN is the grafting of ammonium polyphosphate onto the surface of nanometer lignocellulosic fibres cut by shear force.
Chemical structure analysis of LFPN
The chemical composition and structure of LFPN were analyzed by FTIR, XRD, TGA, XPS, and dispersion stability test as shown in Fig. 3, respectively.
The FTIR spectra of LF, APP, and LFPN was shown in Fig. 3a. As for LF, the peaks at 1430 cm− 1 and 1032 cm− 1 were typical bands of -OH and C-O stretching, respectively (Hejna et al., 2020). The peaks at 1072 cm− 1 (P-O) and 1246 cm− 1 (P = O)were typical peaks of ammonium polyphosphate (Prabhakar et al., 2019). In the FTIR spectrum of LFPN, it can be seen not only characteristic peaks of APP at 1242 cm− 1 (P = O) and 1430 cm− 1 (N-H) (Prabhakar et al., 2019) but also the characteristic peak of LF at 1032 cm− 1 (C-O). More importantly, there were two new peaks at 1078 cm− 1 and 511 cm− 1 attributed to P-O-C (Zeng et al., 2020; Wan et al., 2020; T. L. Zhang et al., 2020c). Interestingly, compared to APP, the peak strength of LFPN at 1687 cm− 1 (N-H) and 1430 cm− 1 (N-H) is significantly reduced (Cai et al., 2018). The results can explain the phenomenon after ball milling: a strong smell of ammonia can be smelled when the lid of the ball mill jar was opened. Overall, the FTIR spectra revealed a possibility that crosslinking type APP was reacted with the hydroxyl group on LF to form the P-O-C bond.
The XRD patterns of LF, LFPN, and APP were shown in Fig. 3b. There were typical and sharp peaks of APP ad LF, indicating both APP and LF have good crystallinity. However, the XRD pattern of LFPN showed an amorphous crystal form, which was completely distinct from LF and APP. This characteristic of LFPN is consistent with the results of SEM and TEM images. The XRD images showed that amorphous morphology was due to the encapsulation of clusters of particles. And it is proved that LFPN has a smaller particle size than LF or APP. Moreover, only weak peaks at 15.7° and 29.8° were corresponding to the characteristic peaks of LF and APP, respectively. In conclusion, LFPN was obtained by the chemical reaction between lignocellulosic fibre and ammonium polyphosphate
The XPS survey spectra were also used to characterize the structure of LFPN as shown in Fig. 3e. And, the peaks ascribed to O1s, N1s, C1s, and P2p can be easily found. What needs illustration is that the peak at 346.0 eV was assigned to Ca2p, which was introduced during the production and treatment of LF (Dhyani and Bhaskar, 2018). The above results indicated that there were C, O, P, N elements for LFPN, which was the sum of the elements in both LF and APP.
The P2p spectra of LFPN can be divided into double peaks at 134.0 eV (P-O-C) and 132.9 eV (P = O) (T. L. Zhang et al., 2020c; T. L. Zhang et al., 2020b), as shown in Fig. 3f. The appearance of P = O and P-O-C bonds indicated that there was a new chemical bond formed by the reaction between ammonium polyphosphate and LF. For LFPN and APP, it should be figured out there were two subpeaks at 400.7 eV for NH4+ and at 398.5 eV for -NH- in the N1s spectra, respectively(P. K. Zhang et al., 2018). However, there were many differences for the two subpeaks between LFPN and APP. Firstly, the ratio of NH4+ to -NH- content in APP is 1.55, while the ratio in LFPN is 6.46. Secondly, the peak intensity of N1s for LFPN was 30% lower than that of APP. Lastly, the P/N atomic ratio of LFPN is about 1.00, which is two times higher than that of APP. The above three points indicated that the chains of cross-linked APP were broken by the action of shear force in the process of the mill, greatly reducing the -NH- and some of the NH4+ undergo hydrolysis, making APP with a small molecular chain react with the hydroxyl group in LF, and increasing the P/N atomic ratio in LFPN.
The transparent orange LFPN suspension showed good dispersion and stability in water (Fig. 3d). And there were no particles felled down after 24 hours. This phenomenon can be illustrated by two reasons: one is that LFPN has a smaller molecular weight than both APP and LF; the other is the introduction of hydrophilic groups such as phosphate groups into LF (Zabihi et al., 2020).
By combining SEM, TEM, FTIR, XRD, XPS, and dispersion stability test, the reaction mechanism can be speculated in Fig. 1. A long time and intermittent ball milling make the chains of LF and APP break into small molecules and release gases such as ammonia and water vapor, which results in a pressure increase in the jar. At the same time, the physical impact of the ball mill produces a lot of heat, and due to the heat dissipation ability of the stainless-steel jar is not as good as the agate jar, the temperature in the jar continuously increased. It can be easily verified by the boiling phenomenon after ball milling. It is easy that the multi-hydroxyl LF can react with APP to generate LFPN flame retardant with short-chain biobased phosphate ester in an environment of high temperature, high pressure, and strong shear force.
The thermal stability of LF, APP, and LFPN was analyzed by TGA. The TGA results of the samples under nitrogen atmosphere were carried out as Fig. 3c. The thermal degradation of APP proceeds by two stages: the temperature of the first stage is between 300–500 ℃, which is due to the release of NH3 accompanied by the formation of polyphosphoric acid; the other stage is between 500–700 ℃, corresponding to volatilization of metaphoric acid and P2O5 (Prabhakar et al., 2019; Sun et al., 2013; P. K. Zhang et al., 2018). There were three main degradations in 200–750 ℃ of LF: the first stage between 200–310 ℃ was due to pyrolysis of hemicellulose; the second stage between 310–400 ℃ was caused by pyrolysis of cellulose; the third stage between 200–750 ℃ was corresponding to the disintegration of lignin through a wide temperature range (P. P. Li et al., 2019; Dhyani and Bhaskar, 2018; Ramirez-Estrada et al., 2021). As for LFPN, the pyrolysis process has some similarities with LF, but it is completely different. A weight loss was observed between 150–250 ℃ of LFPN, indicating the release of initial phosphoric acid and escape of ammonia due to the introduction of low polymeric degree ammonium phosphate structure. And the degradation process between 220–550 ℃ was caused by the slow decomposition of cellulose, hemicellulose, and lignin. It is worth noting that, compared to the DTG curve of LF, the decomposition rate of LFPN was significantly reduced after 220℃. This phenomenon may be that phosphoric acid promotes dehydration and cross-linking of P-O-C bond linked lignocellulosic fibre to carbon formation. And the inference can be confirmed by the initial decomposition temperature (T5%). T5% of APP, LF and LFPN were 348.1 ℃, 319.4 ℃ and 236.9 ℃, respectively. T5% demonstrated that the introduction of the ammonium phosphate structure and the formation of P-O-C bond promote the early decomposition of ammonium polyphosphate and the cross-linking with LF to form carbon char. The formation of the carbon char layer significantly reduces the decomposition rate of LF, resulting in a higher final residual mass of LFPN.
In addition, it has been reported that the decomposition temperature for cotton fabric is about 250℃. In contrast, the decomposition temperature of LFPN is lower than 250℃, and LFPN has a good carbon formation effect, so it is very suitable for the flame retardant finishing of cotton fabric (Grancaric et al., 2017; Pakdel et al., 2020).
Characterization of Cotton-LFPN
SEM-EDS was used to survey the surface morphology of Cotton and Cotton-LFPN, as shown in Fig. 4. As displayed in Fig. 4, both Cotton and Cotton-LFPN showed typical smooth cotton fibre structure at low magnification. At the same time, many fine particles on the surface of cotton-LFPN at high magnification makes the surface rough. The EDS results of Cotton-LFPN (Fig. 4e) confirmed that the many small particles on the fabric surface were due to the formation of LFPN flame-retardant coating.
To understand the form of LFPN coating on the cotton surface, XRD and ATR-FTIR were used to study the fabric surface, which was shown in Fig. 5.
As is shown in Fig. 5a, XRD patterns of Cotton and Cotton-LFPN both showed the typical peaks of cotton fabric. The difference lies in the intensity of the diffraction peak. The peak strength of Cotton-LFPN is weaker than that of Cotton, indicating that there is a thin film composed of amorphous material on the surface of cotton fabric. The phenomenon is corresponding to Fig. 4 and Fig. 3b.
As shown by ATR-FTIR spectra in Fig. 5b, most peaks of Cotton-LFPN were similar to those of Cotton, except the new peaks of N-H, P = O, and P-O at 1430 cm− 1, 1242 cm− 1, and 878 cm− 1, respectively (Bai et al., 2020). In particular, the peak strength of P-O is much stronger than it of LFPN at this wavenumber. This indicates that under the action of heating, the LFPN bonds with the hydroxyl group on the fabric to form a new P-O-C bond.
Thermostability of cotton fabric
The TGA curves for Cotton and Cotton-LFPN under N2 were presented in Fig. 6, and the related data were summarized in Table 1. Both Cotton and Cotton-LFPN showed the evaporation of water before 110 ℃. For Cotton, there was one main degradation process between 250 and 400 ℃, which was due to the thermal decomposition of cellulose (Lin et al., 2019; Costes et al., 2017; Y. Zhang et al., 2019). However, the pyrolysis of Cotton-LFPN was distinct from that of Cotton. As for the Cotton-LFPN compared with that of Cotton., the initial decomposition began about 38.8 ℃ earlier, and the char residue was increased from 16.6 % to 34.7%, respectively. Moreover, the maximum decomposition rate of Cotton-LFPN decreased by 27%, and the temperature at the maximum decomposition rate of Cotton-LFPN decreased from 359 ℃ to 296 ℃. Indicating that LFPN decomposed and formed a char layer, and then the phosphorus products after decomposition dehydrated cotton fabric into carbon char to inhibit further decomposition.
To sum up, the pyrosis of cotton-LFPN consists of two carbonization processes: one process is the formation of LFPN char layer with thermal insulation and shielding effect; the other process is the cross-linking carbonization process of cotton fabrics promoted by LFPN.
Table 1
TGA data of Cotton and Cotton-LFPN.
Sample
|
T5%/ ℃
|
Tmax/ ℃
|
Dmax/ (%/℃)
|
Char residue at 700 ℃/ wt %
|
Cotton
|
271.9
|
359
|
1.26
|
16.6
|
Cotton-LFPN
|
233.1
|
296
|
0.95
|
34.7
|
Flame-retardant Performance Analysis
In corresponding to investigate the flame retardant efficiency of LFPN for cotton fabric, the improved vertical flammability test (IVFT) and cone calorimeter test (CCT) were carried out, and the results were shown in Fig. 7.
As shown in Fig. 7a and video in the support information, both Cotton and Cotton-LFPN were ignited at 3s, but showed a completely different combustion behavior. The flame height and fire spread speed of Cotton-LFPN were weaker than those of Cotton. Moreover, the flame was extinguished, which meant the end of combustion for Cotton-LFPN, while the smolder phenomena lasted for about 20 seconds of Cotton. Smoldering is extremely harmful to fire safety, especially for cotton fabrics. So, it is of great significance to eliminate the smoldering of cotton fabrics (Alongi and Malucelli, 2015; Hagen and Meyer, 2021; Wan et al., 2020). In addition, self-extinguishing of Cotton-LFPN occurred in the horizontal direction, and a dense carbon layer with less than 1/3 damaged area was formed. The IVFT results indicated the good flame-retardant performance of Cotton-LFPN and the excellent char residue-formation property of LFPN.
To further quantitatively analyze flame retardancy, CCT was carried out to study the combustion behavior of cotton fabrics (B. H. Yuan et al., 2017). And the results of CCT were shown in Fig. 7b, c, and Table 2, respectively. The peak heat release rate (PHRR) and total heat release (THR) values of Cotton-LFPN were significantly reduced with the gradual increasing load of LFPN, compared with Cotton. Especially, the PHRR and THR values of Cotton-LFPN-4 were reduced by 57.4 % and 42.2 % compared to those of the control cotton fabric, respectively. As for the smoke parameter, the smoke factor (SF, SF = PSPR × HRR) is an effective index to balance heat and smoke effect (X. L. Chen et al., 2017). From SF curves, it was obvious that a good effect on smoke suppression can be obtained for Cotton-LFPN. However, there was no obvious regularity of SF with different loading of LFPN. Especially for Coton-LFPN-2, the SF value was much higher than that of the control cotton fabric. This may be due to the crack of the surface char-forming fabric, leading to the rapid escape of the isolated flue gas into flame zoom. As shown in Fig. 7c, in the case of Cotton-LFPN, there was a dense char residue layer formed, which can play a role as a physical barrier. In addition, the CO2/CO ratio can reflect the combustion efficiency. With the increase of LFPN loads, the CO2/CO ratio decreased gradually, which means the combustion efficiency decreased gradually. This indicated that the formation of a cross-linking char layer made the burning more inadequate. Moreover, the mass loss curves correspond to thermal stability results and can be explained by digital images of char residue layer after CCT (Fig. 7c). Excellent char-formation performance can effectively reduce the generation of flammable volatiles gases to delay combustion and self-extinguishing behavior (P. P. Li et al., 2019).
The CCT results were closely related to IVFT and showed the same effect, which indicated that cotton-LFPN has excellent flame-retardant performance and flame retardant efficiency.
Table 2
CCT data of Cotton and Cotton-LFPN.
Sample
|
TTI/ s
|
PHRR/ (kW/m2)
|
THR/ (MJ/m2)
|
SF/ (MW/m2)
|
CO2/CO/ (g/g)
|
Residue/ wt%
|
Cotton
|
23
|
287.89
|
7.20
|
4.22
|
49.06
|
11.11
|
Cotton-LFPN-1
|
16
|
221.59
|
6.12
|
1.97
|
17.64
|
22.24
|
Cotton-LFPN-2
|
13
|
184.94
|
5.72
|
2.37
|
10.25
|
19.46
|
Cotton-LFPN-3
|
13
|
168.38
|
5.32
|
11.12
|
7.53
|
18.90
|
Cotton-LFPN-4
|
12
|
122.56
|
4.16
|
2.60
|
7.87
|
28.76
|
Flame Retardant Mechanism
Condensed phase analysis. In order to deduce the flame-retardant mechanism of Cotton-LFPN, SEM, LRS and ATR-FTIR were used to investigate the structural change of condensed phase, as shown in Fig. 8.
As shown in Fig. 8d, the SEM image of char residues after CCT qualitatively confirmed the results of CCT. It was easily found that the fibre structure of char for Cotton was incomplete, exhibited relatively loose and fragile. In contrast, a compact char residue with a complete and tight fibrous structure obtained from Cotton-LFPN can effectively insulate flammable gas into oxygen and delay heat transfer (Y. Yuan et al., 2020). Specifically, some carbon tumors with spherical structure can be observed, indicating ductile char surface from the LFPN coating acted effective protection for cotton fabric from further combustion(Liu et al., 2016).
The LRS was conducted to further quantitative evaluation performance of char residues of Cotton and Cotton-LFPN after CCT, as depicted in Fig. 8a, b. The two strongly overlapping peaks at around 1580 (G) and 1350 (D) cm− 1, which belonged to graphitized and amorphous carbon, respectively (Tao et al., 2021; Zhao et al., 2021). The proportion of the integral area of D band and G band (ID/IG) can reflect the degree of graphitization for char residue. It means the lower ID/IG signifies the higher graphitization and physical barrier performance(R. Chen et al., 2020). The ID/IG value of Cotton-LFPN (ID/IG =0.89) was lower than that of Cotton (ID/IG =1.27), which means the graphitization degree of Cotton-LFPN is higher than that of Cotton. The results of LRS is corresponding to that of SEM.
Moreover, the char residues of Cotton-LFPN were analyzed by ATR-FTIR to speculate about the possible course of change of condensed phase. As shown in Fig. 8c, peaks at 1101 cm− 1 assigned to the P-O-C bond, demonstrating the phosphor carbonaceous structure was formed by LFPN (Zeng et al., 2020). And characteristic absorption peak of C-O-C and C-H peaks were obtained at 1270 and 2950 cm− 1 (Zheng et al., 2019), respectively. This indicated that there were carbonized into the char with cross-linked aromatic ring components. The analysis of the condensed phase showed the phosphate groups in LFPN catalyzed dehydration to form a crosslinked char protection layer, which played a good flame retardant effect.
Gas phase analysis. In addition to the condensed phase, the gas phase analysis can further understand the combustion and pyrolysis process and infer the flame-retardant mechanism.
Effective heat of combustion (EHC), which means the ratio of heat release rate to the mass loss rate, is a valid measure to judge the flame-retardant mechanism. Because EHC can show the degree of involvement for flammable volatile gases in the combustion process (Tang et al., 2016; Dong et al., 2016). The EHC curves of Cotton and Cotton-LFPN was given in Fig. 9. Compared with the control cotton fabric, the EHC of treated cotton fabric was significantly reduced. This indicated Cotton-LFPN produced less-flammable volatile gases and gas phase flame-retardant effect existed.
It has been reported that the microscale-combustion calorimeter was an effective method to vertify the result of EHC (Xing et al., 2009; Qian et al., 2016; H. X. Yuan et al., 2012). From Fig. 9b, the PHRR value of Cotton-LFPN was reduced by 77% than that of Cotton, indicating the Cotton-LFPN pyrolized flammable volatiels less than the control cotton fabric.
Therefore, TG-IR-GC/MS was conducted to analyze the change among the gas phase during pyrolysis (Yu et al., 2021), as shown in Fig. 10. The peak in 3746 cm− 1 assigned to the O-H bond of H2O. And it was obvious that the peaks at 2359, 2182, 1732, and 1506 cm− 1 correspond to CO2, CO of the thermal cracking process, carbonyl compounds, and aromatic compounds, respectively. From the FTIR spectrum of gas products at maximum decomposition rate (Fig. 10a), combustible gas of Cotton-LFPN such as hydrocarbon and carbonyl compounds were decreased compared with control cotton fabrics. Nay, the TG-IR spectra at a specific temperature (Fig. 10b, c) of 100, 200, 300, 400, 500, 600 and 700 ℃ showed the intensity of pyrolysis gas for Cotton-LFPN is generally weaker than that for Cotton. Gas chromatogram (Fig. 10d) combined with mass spectrum can further confirm the infer of TG-IR. Compared with control fabrics, aldehydes, ketones, esters, and aromatic pyrolysis products of Cotton-LFPN were decreased except for the furans. This indicated that LFPN catalyzes the dehydration and carbonization of cotton fabrics, and inhibits the formation of L-glucose and the increase of pyrolysis products. It is also worth noting that the peaks at 3300, 1625, and 1500 cm− 1, corresponding to the NH3 absorptions, can be observed from Fig. 10c (Z. Z. Wang et al., 2009). And significant enhancement of the water absorption peak. This indicated there was gas phase flame-retardant mechanism in LFPN.
Based on the above analysis, the flame-retardant mechanism of LFPN could be speculated, and the possible process as illustrated in Fig. 11. The flame-retardant mechanism of LFPN can be divided into two parts: condensed phase and gas phase. The condensed phase flame-retardant mechanism could be separated into two aspects. On the one hand, LFPN flame-retardant coating was dehydrated and cyclized into graphitization LFPN char layer at high temperature due to the polyhydroxy structure and phosphate-rich group of LFPN. LFPN char layer was endowed with a dense protective layer with the carbon tumor structure, which can isolate the contact between oxygen and combustible gas, suppress released smoke, shield the transfer of heat. On the other hand, LFPN flame-retardant coating released phosphoric acid and derivatives, catalyzed the cross-linking carbonization of cotton fibres, and form a stable phosphorus-containing ring structure, to prevent the further decomposition of cellulose. In the gas phase, the concentration of combustible gas was diluted by non-flammable gases such as NH3 and H2O vapor which was released by LFPN flame retardant coating. Moreover, the self-extinguishing phenomenon proved that LFPN decomposed P· radicals which participated in the free radical competitive reaction and reduced the concentration of OH· in the flame region (Velencoso et al., 2018).