3.1. Structural characterization and performance testing of P/N-CNCs
The structure and the thermal stability of P/N-CNCs were further characterized by FTIR, XPS, XRD and TGA. Fig. 3a is the infrared spectrum of P/N-CNCs. By comparing with CNC curves, it was found that the peak strength ratio of O-H bond to C-H bond in P/N-CNC curve decreased (Ling et al. 2019). On the one hand, some hydroxyl groups in CNCs dehydrated with the hydrolyzed modification agent, resulting in the reduction of the number of hydroxyl groups. On the other hand, the modification agent with phosphazene group contains C-H bond, which made the increase of C-H bond in the system. As a result, the peak strength ratio of O-H bond to C-H bond decreased. It was also observed that the P/N-CNC curve showed a dentate peak in the wave number range of 1100 cm-1, which was due to the typical absorption peak of Si-O bond contained in the modification agent (Chen et al. 2018). Based on the above analysis, it suggests that the modification agent has been successfully grafted onto CNCs.
The chemical compositions of CNCs and P/N-CNCs were analyzed by X-ray photoelectron spectroscopy. Fig. 3b shows the XPS curves of CNCs and P/N-CNCs, and the corresponding data are listed in Table 2. As shown in Fig. 3b, the peaks at 101 eV represented Si2p, 153 eV represented P2p, 285 eV represented C1s, 398 eV represented N1s and 532 eV represented O1s respectively. In addition, it can be seen from Table 2 that the contents of C element and O element in CNCs were 46.8 wt% and 53.2 wt% respectively, while in P/N-CNCs, there were 3.1 wt% N, 2.5 wt% P and 5.4 wt% Si in addition to 41.2 wt% C and 47.8 wt% O. It indicates that the modified CNCs contain nitrogen, phosphorus and silicon elements.
The crystal structures of CNCs after chemical grafting were studied by X-ray diffraction (Fig. 3c). There were three typical diffraction peaks at 2θ=14.6°, 16.5°and 22.6° in CNC, corresponding to the (100), (010) and (110) crystal faces of type I cellulose, respectively (French AD. 2014). It could be seen clearly from the XRD curves of P/N-CNCs that its peaks at 2θ=14.6°, 16.5° and 22.6° were almost the same as the typical diffraction peaks of CNC, and the relative intensity does not change significantly. In addition, a new diffraction peak appears at 2θ=11.5°, which was caused by the crystal structure of the modification agent with phosphazene group (Feng et al. 2017), indicating that the modification agent was successfully grafted onto the surface of CNCs and had no effect on the crystal structures of CNCs.
Thermal stability of CNCs before and after modification was evaluated by TGA. The initial decomposition temperature (the temperature at 5% mass loss, T5%), the maximum decomposition temperature (Tmax1 and Tmax2) and the residual char at 700 ℃ are shown in Table 3. Both CNCs and P/N-CNCs had two stages of decomposition in N2 atmosphere. As shown in Fig. 3d and Table 3, the T5% (272.1℃) and Tmax1 and Tmax2 (340.9 ℃ and 413.2 ℃) of P/N-CNCs were all much higher than those of CNCs (202.4℃, 272.2 ℃ and 373.5 ℃), indicating that the thermal stability of CNCs could be greatly improved by grafting the modification agent onto the surface of CNCs. This is because the modification agents grafted on the surface of CNCs are polycondensation structures, rather than small molecules. This polycondensation structures have better thermal stability, which is conducive to the improvement of the thermal stability of CNCs. In addition, it can be seen from Table 3 that the residual char of P/N-CNCs at 700 ℃ was 55.8 wt%, which was more than double of CNCs (24.9 wt%), indicating that the modified CNCs (P/N-CNCs) had better char-forming ability.
Table 2. Surface elemental analysis of CNCs and P/N-CNCs obtained from XPS
Samples
|
C
(wt%)
|
O
(wt%)
|
N
(wt%)
|
P
(wt%)
|
Si
(wt%)
|
modification agent
|
36.3
|
25.4
|
9.0
|
7.8
|
21.5
|
CNCs
|
46.8
|
53.2
|
-
|
-
|
-
|
P/N-CNCs
|
41.2
|
47.8
|
3.1
|
2.5
|
5.4
|
Table 3. Thermal behavior and charring parameters of CNCs and P/N-CNCs under N2 atmosphere
Samples
|
Td,5%
|
Tmax (℃)
|
Residuals at 700 ℃
(wt%)
|
(℃)
|
Tmax1 (℃)
|
Tmax2 (℃)
|
CNCs
|
202.4
|
272.2
|
373.5
|
24.9
|
P/N-CNCs
|
272.1
|
340.9
|
413.2
|
55.8
|
The micro morphologies of CNCs were studied by transmission electron microscope to reveal the change caused by modification, and the images are shown in Fig. 4. Fig. 4a1 and 4a2 show that the CNCs present rod-like whiskers. After the modification, the sizes and the morphologies of CNCs were both changed. As shown in Fig. 4a2, the diameters of the CNCs before coating are between 10-40 nm, and as shown in Fig. 4b1, the diameters of the CNCs after modification are about 500 nm. In TEM images, the elements with high relative atomic mass show dark color, while the elements with low relative atomic mass show light color. Thus, the dark spherical aggregates in Fig. 4b1 and 4b1 are modification agents because the P and Si elements had high relative atomic mass. The light whiskers are CNCs. The TEM images of P/N-CNCs also indicate that the modification agents with phosphazene group have been grafted to the CNC whiskers.
3.2. Flame retardancy of PLA composites
The flame retardant properties of PLA composites were evaluated by limiting oxygen index and vertical combustion tests (Chen et al. 2018; Ma et al. 2021; Zhou et al. 2021; Gao et al. 2020). The test results are shown in Table 4. In Table 4, the LOI value of pure PLA was only 20.1%, while the LOI values of PLA composites containing flame retardants were significantly increased. It is worth noting that the LOI value of PLA/7APP/3P/N-CNCs was the highest, reaching 28.1%, which was 11.5% higher than that of PLA/7APP/3CNCs (25.2%) and 14.2% higher than that of PLA/10APP (24.6%). The vertical burning test results showed that the combustion time of PLA/7APP/3P/N-CNCs was very short, and the total combustion time was only 2.1 s, which was the shortest among all the samples. PLA/7APP/3P/N-CNCs reached UL 94 V-0 rating, while PLA/7APP/3CNCs only reached UL 94 V-2 rating. Although PLA/10APP also reached UL 94 V-0 rating, the combustion time was longer than that of PLA/7APP/3P/N-CNCs. These results implied that after the grafting of modification agent with phosphazene groups, the char-forming performance of P/N-CNCs was improved. Thus, P/N-CNCs could play a better synergistic flame retardant effect with APP in the combustion process.
From Table 4, it can also be seen that, the flame retardancy of the composite prepared by adding APP and P/N-CNCs into PLA was obviously better than that of the other three flame retardants (MPP, AHP and PPAP). When MPP or AHP or PPAP were added to PLA/P/N-CNCs system, it was found that the LOI values of these composites were all lower than that of PLA/7APP/3CNCs, and the t1 and t2 of the material were significantly increased, resulting the UL 94 V-2 rating. in vertical burning test. In addition, it can be clearly seen from Fig. 5 that the drops of PLA/7APP/3P/N-CNCs did not ignite the absorbent cotton (Fig. 5c) and passed the UL 94 V-0 rating. However, the droplets of the other three samples in the combustion process could ignite the absorbent cotton (Fig. 5d, 5e and 5f), and all of them were rated as UL94 V-2 rating. Contrast Fig. 5b with Fig. 5c, we found that CNCs were carbonized, indicating that the unmodified CNCs could not meet the processing temperature requirements of PLA. The thermal stability of the modified CNCs were significantly improved, so they were not carbonized during the processing of PLA composites. The above results indicated that APP combined with P/N-CNCs were beneficial to improve the LOI value and UL 94 rating of PLA compared with the other three phosphorus-containing flame retardants. Therefore, we only selected APP combined with P/N-CNCs in the later research.
Table 4. The LOI and UL-94 test results of PLA composites
Samples
|
LOI (%)
|
vertical burning test
|
after-flame time
|
UL-94
|
dripping
|
t 1 (s)
|
t2 (s)
|
Pure PLA
|
20.1
|
48.3
|
-
|
NR
|
Y
|
PLA/10APP
|
24.6
|
4.8
|
0.9
|
V-0
|
Y
|
PLA/7APP/3CNCs
|
25.2
|
4.3
|
0.7
|
V-2
|
Y
|
PLA/7APP/3P/N-CNCs
|
28.1
|
1.8
|
0.3
|
V-0
|
Y
|
PLA/7MPP/3P/N-CNCs
|
22.4
|
12.4
|
2.2
|
V-2
|
Y
|
PLA/7AHP/3P/N-CNCs
|
22.7
|
14.5
|
3.1
|
V-2
|
Y
|
PLA/7PPAP/3P/N-CNCs
|
23.7
|
11.7
|
5.8
|
V-2
|
Y
|
Cone calorimeter could simulate the real combustion behavior of materials and was widely used to study the combustion behavior of polymer materials (Li et al. 2019). In order to further explore the influence of APP combined with P/N-CNCs on the flame retardancy of PLA, cone calorimeter tests were carried out. In Fig. 6, we could see the heat release rate (HRR) curves and the char yield curves of the samples. The test data are listed in Table 5. As shown in Fig. 6a, pure PLA burned rapidly after ignition and reached the peak within a relatively short time, with a pk-HRR value of 561 kW/m2 and the total heat release (THR) value of 84.7 MJ/m2. The pk-HRR and THR of PLA were decreased after APP was added to PLA. Compared with pure PLA, the pk-HRR and THR values of PLA/10APP were decreased by 28.9% and 10.8%, respectively, which indicated that APP could inhibit the combustion of PLA. When 3 wt% CNCs or P/N-CNCs were added into the PLA/APP system, both of the pk-HRR values for PLA/7APP/3CNCs and PLA/7APP/ 3P/N-CNCs were lower than that of PLA/10APP. The pk-HRR value of PLA/7APP/ 3P/N-CNCs was 266 kW/m2, which was the lowest among all the composites.
As shown in Table 5, the total smoke release (TSR) of pure PLA was only 50 m2/m2. After the addition of 10% APP, it improved significantly with a value of 280 m2/m2. Compared to PLA/10APP, the TSR value of the PLA/7APP/3CNCs decreased by 15.0%, while the TSR value of the PLA/7APP/3P/N-CNCs increased by 15.7%. It suggested that the modified CNCs could not inhibit the generation of smoke. In addition, compared with PLA/7APP/3CNCs, the average amount of CO for PLA/7APP/3P/N-CNCs decreased, while the average amount of CO2 increased, indicating the increase of incomplete combustion substances, which was consistent with the trend of the increase of final char yield.
As shown in Fig. 6b, there was almost no residual char generated after pure PLA combustion. After 10% APP was added, the final char yield of the PLA composites was 6.5 wt%. The carbon residue of the material decreased after replacing 3% APP with 3% CNCs, indicating that the char forming ability of the unmodified CNCs are poor. Compared with PLA/7APP/3CNCs, PLA/7APP/3P/N-CNCs has more residual char, even more than PLA/10APP. The final char yield of PLA/7APP/3P/N-CNCs was 6.8 wt%. This result was consistent with the curves in Fig. 6b, in which PLA/7APP/3P/N-CNCs shows the lowest mass lost rate and the highest char residue. It indicates that P/N-CNCs play a flame retardant role in the condensed phase.
From Table 5, we also find that the time to ignition (TTI) of PLA composites containing flame retardants are shorter than that of pure PLA. This is because APP, CNCs and P/N-CNCs, as flame retardants, usually decompose before PLA, leading to shortening of the TTI of the material. In addition, the average effective heat of combustion (av-EHC) of PLA was significantly decreased after adding flame retardant, but the av-EHC of PLA/10APP, PLA/7APP/3CNCs and PLA/7APP/3P/N-CNCs had little difference. It showed that CNC and P/N-CNCs had little flame retardant effect in the gas phase. APP released large amount of inert gas during the combustion process, which played the main role of dilution effect in gas phase.
In summary, the synergistic flame retardant effect of APP and P/N-CNCs in inhibiting the peak heat release rate and mass loss of PLA matrix was better than that of APP and CNCs. That is because the synergistic flame retardant effect of P/N-CNCs and APP in the condensed phase is enhanced after being treated with modification agent containing phosphazene groups.
Table 5. Date of PLA composites obtained from cone calorimeter test
Samples
|
TTI
(s)
|
pk-HRR
(kW/m2)
|
THR
(MJ/m2)
|
av-EHC
(MJ/kg)
|
TSR
(m2/m2)
|
av-CO
(kg/kg)
|
av-CO2
(kg/kg)
|
Char yield
(wt%)
|
Pure PLA
|
44
|
561
|
84.7
|
20.0
|
50
|
0.0148
|
2.0325
|
0.1
|
PLA/10APP
|
31
|
399
|
77.1
|
18.4
|
280
|
0.0285
|
1.8192
|
6.5
|
PLA/7APP/3CNCs
|
27
|
282
|
70.0
|
18.3
|
238
|
0.0279
|
1.8422
|
5.2
|
PLA/7APP/3P/N-CNCs
|
28
|
266
|
77.6
|
18.1
|
324
|
0.0307
|
1.7792
|
6.8
|
3.3. The analysis of residual char after cone calorimeter test
In order to further study the flame-retardant mechanism of P/N-CNCs and APP in PLA, the surface morphology of the samples after cone calorimeter tests were analyzed in detail. As shown in Fig. 7, the char layer of PLA/10APP was thin and had obvious cracks. PLA/7APP/CNCs exhibited uneven distribution of carbon residue, while that of PLA/7APP/3P/N-CNCs was uniform and complete. In addition, it could be seen that PLA/7APP/3P/N-CNCs had more residual char than PLA/7APP/CNCs after burning and there was no obvious hole on the surface of the char layer. It demonstrated that APP with P/N-CNCs could effectively promote the formation of char layer.
The micromorphology of residual char for flame retardant PLA composites was further studied by SEM. And the relative element content in the char residue was determined by energy-dispersive X-ray (EDX). Fig. 8a1, b1 and c1 are the SEM images for the outer surface of the char residue, and a2, b2 and c2 are for the inner surface. As shown in Fig. 8a1, the outer surface of the char residue for PLA/10APP showed a typical morphology of intumescent char layer, but there were obvious holes. On the inner surface of the char residue in Fig. 8a2, it could also be seen that there were many voids or holes left, which was due to the release of the dilution gas decomposed by APP. As can be clearly observed from Fig. 8b2, the inner surface of the char residue of PLA/7APP/3CNCs presented a fibrous char (marked with yellow arrow). This structure could maintain the structural integrity and provide the mechanical strength for the char layer. It means that CNCs not only provide carbon source for char layer, but also play the role of carbon layer skeleton. As can be clearly observed from Fig. 8c2, PLA/7APP/3P/N-CNCs also showed fibrous char, which were inherited from CNCs. However, unlike PLA/7APP/CNCs, there were many white particles on the carbon skeleton for PLA/7APP/P/N-CNCs, which were compounds containing silicon elements (Fig. 8c4). Meanwhile, there were many small-sized char foams with closed pores on the fibrous carbon skeleton. In addition, the outer surface of PLA/7APP/P/N-CNCs was more compact and complete than that of PLA/7APP/3CNCs. The results of elemental analysis showed that not only silicon elements but also more nitrogen elements were locked in the carbon residue, which indicated that the phosphazene containing modification agent remained in the carbon residue and enhanced the flame retardancy of the condensed phase.
3.4. The thermal stability of PLA composites
The thermogravimetric tests of PLA composites were carried out in two different atmospheres. Its test curves are shown in Fig. 9, and the corresponding test data are shown in Table 6. It can be observed from Fig. 9 that the thermal degradation process of the samples in air atmosphere were similar to that in N2 atmosphere, with one-stage thermal degradation process. In N2 atmosphere, pure PLA exhibited high thermal stability, and its initial decomposition temperature was 355 ℃, and only a small amount of residual char was generated at 700 ℃. After the addition of APP, the initial decomposition temperature of PLA/10APP was significantly lower than that of pure PLA, which was due to the partial decomposition of APP at low temperature. After further addition of CNCs, the initial decomposition temperature of the material was further reduced, because the thermal stability of CNCs was poor, which begins to decompose at 202.4 ℃. Comparing with CNCs, the addition of P/N-CNCs can increase the initial decomposition temperature of flame retardant PLA composites from 337 ℃ to 340 ℃, because the thermal stability of CNCs can be improved by grafting modification agent containing phosphazene groups onto CNC. According to Table 6, Tmax of PLA/7APP/3P/N-CNCs was the same as that of PLA/7APP/3CNCs, but the maximum decomposition rate was slower. The maximum mass loss rate (MLR) of PLA/7APP/3P/N-CNCs was 2.6 wt%/℃, which was lower than those of PLA/7APP/3CNCs (2.9 wt%/℃). The decrease of MLR indicated that the modified CNC could further inhibit the combustion of PLA/APP system. In addition, the residual char amount at 700 ℃ of PLA/7APP/3P/N-CNCs was 10.2 wt%, which was slightly higher than that of PLA/7APP/3CNCs(9.2%). This is because the carbonization capacity of P/N-CNCs is improved after the modification.
In air atmosphere, the initial decomposition temperature and the final carbon residue of PLA/7APP/3P/N-CNCs were higher than those of PLA/7APP/3CNCs. They had the same maximum decomposition temperature and smaller maximum decomposition rate. Compared with PLA/7APP/3CNCs (3.2 wt%/℃), the maximum MLR of the PLA/7APP/3P/N-CNCs (2.8 wt%/℃) was reduced by 12.5%. These results indicated that the modification of CNCs by grafting flame retardant agent containing phosphazene groups could promote the thermal stability and the char-forming ability of the PLA composite.
Table 6. Thermal behavior and charring parameters of PLA composites
Samples
|
Td,5%
(℃)
|
Tmax
(℃)
|
MLR
(wt%/℃)
|
Residuals at 700 ℃ (wt%)
|
N2
|
Pure PLA
|
355
|
378
|
3.7
|
2.9
|
PLA/10APP
|
344
|
375
|
2.7
|
9.2
|
PLA/7APP/3CNCs
|
337
|
374
|
2.9
|
9.2
|
PLA/7APP/3P/N-CNCs
|
340
|
374
|
2.6
|
10.2
|
Air
|
Pure PLA
|
349
|
386
|
4.3
|
2.2
|
PLA/10APP
|
344
|
375
|
3.3
|
9.5
|
PLA/7APP/3CNCs
|
339
|
374
|
3.2
|
7.1
|
PLA/7APP/3P/N-CNCs
|
341
|
374
|
2.8
|
9.2
|
3.5. The mechanical properties of PLA composites
It was found that more and more applications require PLA composites to have good mechanical properties as well as excellent flame retardant properties (Chen et al. 2019). However, most of the flame retardants will reduce the mechanical properties of the materials (Huang et al. 2021; Singh et al. 2021; Yin et al.2018). As shown in Fig. 10, adding APP to PLA can significantly reduce the tensile strength and impact strength of the material. In Fig. 10, compared with PLA/10APP, the tensile strength of PLA/7APP/3CNCs was 35.8 MPa, which decreased slightly. The impact strength was 8.6 kJ/m2, 24.6% higher than that of PLA/10APP. As shown in Fig. 10a and 10b, comparing with PLA/7APP/3CNCs, the tensile strength of PLA/7APP/3P/N-CNCs increased from 35.8 MPa to 38.4 MPa, the impact strength increased from 8.6 kJ/m2 to 10.2 kJ/m2, and the elongation at break increased from 1.7% to 2.2%. It was worth noting that the PLA/7APP/3P/N-CNCs had the highest elongation at break. In general, PLA/7APP/3P/N-CNCs had the best mechanical properties among the flame retardant PLA composites. This is because the thermal stability of CNCs is improved by grafting flame retardant, which can mitigate the effect of decomposition during processing. Moreover, CNCs take advantage of large aspect ratio and benefit the mechanical properties of the material (Abraham E et al. 2016). These results indicate that the grafting treatment of CNCs give the material excellent flame retardancy and improve the mechanical properties at the same time.
The mechanical properties test results showed that the mechanical properties of the PLA composites were enhanced after the grafting of modification agent with phosphazene on CNCs. In order to further study the mechanism of improving the mechanical properties, the microstructure of the cross sections of the material after impact test were observed by scanning electron microscope. As shown in Fig. 11, after adding flame retardants, the impact cross section showed sea-island structures, and the flame retardant existed as dispersed phases. The cross section of PLA/10APP with uneven and filamented structure (yellow arrow) is presented in Fig. 11b. It indicated that PLA/10APP had good toughness. By comparing Fig. 11c and Fig. 11d, it was found that the flame retardant in PLA/7APP/3CNCs were not uniformly dispersed, while the section of PLA/7APP/3P/N-CNCs was rough and the dispersed phases were uniform. It is further proved that PLA/7APP/3P/N-CNCs has better toughness than PLA/7APP/3CNCs.