Marion Weissenbergerab,
Adrien Vincentb,
Yves Champavierc,
Cristina Coelho Diogod,
Florence Babonneaue,
Nicolas Pradeillesa,
Alexandre Maîtrea and
Romain Lucas-Roper*a
aUniversity of Limoges, IRCER, UMR 7315, Limoges, F-87068, France. E-mail: romain.lucas@unilim.fr
bSaint-Gobain Research Provence, Cavaillon, F-84306, France
cUniv. Limoges, CNRS, Inserm, CHU Limoges, BISCEm, UAR 2015, US 42, Limoges, F-87068, France
dFCMat, Fédération de Chimie et Matériaux de Paris Centre, CNRS, Sorbonne Université, FR2482, Paris, F-75005, France
eLaboratoire de Chimie de la Matière Condensée de Paris (LCMCP), CNRS, Sorbonne Université, UMR 7574, Campus Jussieu, 4 Place Jussieu, Paris, F-75005, France
First published on 10th July 2024
The innovation introduced in this study consists of replacing toluene with safer solvents such as cyclopentane or diethyl ether in the processing of a preceramic polycarbosilane (allylhydridopolycarbosilane, AHPCS) and assessing its impact on the functionalisation of B4C powders to produce B4C/SiC composites. Fourier-transform infrared (FT-IR) with ATR and nuclear magnetic resonance (NMR) spectroscopy revealed no major modification in the polymer structure. SEC/MALS analysis showed a slight change in the number-average molar mass of the polymer regardless of the functionalisation solvent used in correlation with a slight decrease in the polymer ceramic yield due to oligomer loss. The thermal behaviour of the preceramic polymer investigated via mass spectrometry remained unaffected by the solvent change. The search for polymer residues after distillation highlighted the recyclability of both the functionalisation solvent and the polymer, despite a slight increase in the molar mass of the polymer. Finally, the sinterability of B4C/AHPCS samples was studied with the preparation of B4C/SiC composites via a polymer-derived ceramic (PDC) route and spark plasma sintering (SPS). The effect of the solvent on the microstructure and relative density of the specimens (>92%) is negligible. The specimens retain a fine and homogeneous phase distribution despite process modification. The results highlight the approach developed to use greener solvents for the chemical synthesis of functionalised ceramics and represent a step towards the generalisation of more environmentally friendly processes.
Fig. 1 AHPCS (SMP-10, Starfire Systems) structure confirmed by the literature.14 |
FT-IR (ATR) (cm−1) for AHPCS, AHPCS-T, D and C: ν(C–H): 2840–3000; 3078, ν(Si–H): 2130, ν(CC): 1631, δ(C–H): 1356, ν(Si–CH3): 1255, δ(Si–CH2–Si): 1046, δ(Si–H): 939, ν(Si–C): 856, δ(Si–CH3): 766. T-AHPCS: ν(C–H, aromatic): 2856–3102, δ(C–H): 1027; 1080. D-AHPCS: ν(C–H): 2870; 2976, δ(CH3): 1351; 1383. C-AHPCS: ν(C–H): 2874; 2954.
1H NMR (ppm) AHPCS, AHPCS-T, D and C (toluene-d8): δ: 4.96–5.75 (CHCH2), δ: 3.73–4.12 (Si–H), δ: 1.36–1.60 (Si–CH2), δ: −0.3–0.08 (Si–CH3). 13C NMR (ppm) AHPCS, AHPCS-T, D and C (toluene-d8): δ: 50–70, 114 and 130 (CHCH2 allyl carbons), δ: <0 (Si–CH3).
29Si liquid-state NMR spectra were recorded on a Bruker Advance 500 MHz spectrometer in chloroform-d.
29Si NMR (ppm) AHPCS, AHPCS-T, D and C (CDCl3): δ: −8 to −15 (C3SiH), δ: −30 to −40 (C2SiH2), δ: −60 to −67 (CSiH3).
This list was based on the most widely available and used solvents. The selection was made by excluding solvents bearing a “Serious health hazard” pictogram. A second criterion was taken into account to select solvents with HSP, permittivity, or dipole moment that was close to toluene. Limits have been set regarding the permittivity (<7), dipole moment (<2), and HSP (between 14 and 22). With these criteria, four candidates were selected to replace toluene (T): acetic acid (A), diethyl ether (D), cyclopentane (C) and tetramethyloxolane (TMO). The latter was withdrawn from the selection for economic reasons and due to its low availability (600 € for 250 mg).28 For the targeted application presented in Fig. 2, the polymer must be miscible with the solvent in order to enable a homogeneous dispersion of the polymer during the surface functionalisation of B4C particles during mixing at room temperature. The solvent was extracted with a rotary evaporator, which implies that a solvent with a low boiling temperature will be preferred. For the sake of recyclability, the polymer–solvent couple that allows for the maximum recovery of each species without degrading them will be chosen. After the preselection of the solvents as an alternative to T, miscibility tests were performed to testify to the miscibility between the polymer and the solvent. The polymer is miscible with D and C, unlike A. Fig. 3 illustrates the experiments carried out in this study to analyse the interaction between AHPCS and the solvents in contact with it.
Fig. 2 General protocol for the functionalisation of the B4C particle with AHPCS, followed by sintering (T = toluene; D = diethyl ether; C = cyclopentane). |
Fig. 3 Experimental plan for analysing the interactions between AHPCS and the solvents in contact with it. |
This set of experiments aims to investigate the behaviour of the polymer in different solvents, mainly in terms of miscibility and whether the solvent can affect the structure or composition of the polymer. Furthermore, the extraction ability of the polymer will be studied, as well as the potential recyclability of the solvent after polymer extraction. Before and after mixing of the preceramic polymer with the solvent, mass measurements were performed to evaluate the possible losses of polymer and solvent during the process. The amount of AHPCS collected compared to that introduced is lower for each solvent, which reflects a loss of polymer. Polymer yields after extraction are as follows: T (92 wt%) < D (97 wt%) < C (99 wt%). This could be explained by the stronger interactions between T and the polymer. Moreover, the heating provided during the evaporation of T could also lead to the evaporation of oligomers present in the polymer, a result confirmed by SEC-MALS analysis. On the other hand, regarding the solvent collected, the mass losses are as follows: D (44 wt%) > C (19 wt%) > T (10 wt%). After collecting the polymer, some structural investigations were performed by infrared spectroscopy to see whether the contact with the solvent affected its structure. (a) The AHPCS spectrum shows bands characteristic of the vibrations of polycarbosilane bonds in agreement with the literature.13,29–31
After extraction of the solvent, it was interesting to see whether polymer traces could be detected in the solvent. An absence of polymer residues in the solvent could lead to a potential recyclability of the solvent throughout the process. Fig. 4 presents the IR spectra of the solvent before and after contact with the polymer. The vibration bands typical of each solvent can be observed. However, none of them were attributed to the AHPCS. This indicates that there is no detectable amount of polymer in the solvent, and that the solvent could therefore be recycled.
1H, 13C and 29Si NMR spectra were also recorded on AHPCS, AHPCS-T, D and C (Fig. 5). The 1H peaks (Fig. 5a) correspond to what is expected for AHPCS, according to the literature.13,31,32 Some additional small peaks are present (see arrows in Fig. 5b), which are due to some residual solvent (increase of the peak at 2.12 ppm due to the T trace; singlet at 1.46 ppm due to the C residue and triplet at 1.11 ppm due to the D residue). The 13C NMR spectra were also recorded (Fig. 5c). The signals at 114 and 134 ppm are assigned to the CHCH2 allyl carbons. Peaks below 0 ppm correspond to Si–CH3 groups, and could be responsible for the presence of free carbon after the polymer pyrolysis.13,31,33 These spectra are similar to what is expected for AHPCS, except for some small peaks between 25 and 35 ppm for AHPCS-C due to C residue. As for the 29Si NMR spectrum of AHPCS in Fig. 5d, three chemical shift ranges can be observed due to C3SiH sites from −15 to −8 ppm, due to C2SiH2 sites from −40 to −30 ppm, and due to CSiH3 sites from −67 to 60.13,34,35 The spectra of AHPCS-D, AHPCS-T and AHPCS-C are very similar to the previous ones, with a global lower intensity for AHPCS-T, which is certainly due to the lower sample concentration. As a conclusion and in line with the IR study, this NMR study (1H, 13C and 29Si) confirms that the polymer structure is not affected by the presence of any of the three solvents tested.
Fig. 5 1H (a and b), 13C (c), and 29Si (d) NMR analysis of the AHPCS before and after being dissolved in T, D and C. |
The thermal behaviour and the ceramic yield of a polymer are of great interest when it comes to sintering. It highlights the suitability of the polymer for use in the sintering process. The thermal behaviour of AHPCS was studied before and after contact with the solvent. Results are presented in Fig. 6a. The as-received AHPCS presents a ceramic yield of 75% and three distinct weight loss domains. The first weight loss is observed from 100 °C, the second starts at 200 °C, and the last one at 450 °C. Above 800 °C, no significant weight loss was observed. Mass losses and associated phenomena are in line with expectations according to the literature.11 MS combined with TGA analysis revealed the departure of different species (Fig. 6b). The first stage of weight loss (4%) was attributed to the departure of oligomers. During the major weight loss (15%) which starts at 200 °C, the reaction of dehydrocoupling took place with the release of hydrogen (m/z = 2), and organic fragments (m/z = 14 for CH2, m/z = 15 for CH3) were also detected.
Fig. 6 (a) TGA of the polymer before and after contact with the solvent. (b) Mass spectrometry analysis. |
In addition, the departure of organic fragments with higher molecular weight (m/z = 29 for CH3–CH2, m/z = 31 for CH3–OH/SiH3, m/z = 43 for C3H7 and m/z = 44 for C3H8) was observed, as confirmed by the work of Li et al.30 This main stage was due to the conversion from an organic to inorganic material. Finally, the last weight loss domain (7%) was correlated to the loss of hydrogen (m/z = 2) and some organic fragments (m/z = 44 for C3H8). After contact with solvents, the ceramic yields of the polymer slightly decreased compared to the initial value of 75%: 72% for T, D and 73% for C. The second weight loss starting at 200 °C was also different for the polymers in contact with solvents. These characteristics could be due to a loss of oligomers, leading to slightly modified crosslinking.
To elucidate the impact of the process (contact with solvent followed by solvent evaporation) on the molar masses of AHPCS, a SEC-MALS analysis was carried out. Fig. 7 displays the elution time of AHPCS, and of the different preceramic polymers after the evaporation of the three solvents (AHPCS-C, -D and -T), at a flow rate of 1 mL min−1. It can be seen that the broad elution profiles changed from AHPCS with one maximum of the Rayleigh ratio at roughly 7.0 mL, to AHPCS-i, with two peaks at approximately 6.0 and 6.6 mL. The distillation of the solvents led to an increase of the average molar mass of AHPCS-i due to the evaporation of oligomers present in AHPCS (Table 2). Thus, even if the molar-mass dispersity remained in the range of 1.7–2.0, AHPCS-D and AHPCS-C had the highest mass-average molar masses at 128600 and 126500 g mol−1, respectively, while AHPCS-T had a mass-average molar mass that was intermediate between AHPCS and AHPCS-C.
To link these results to the thermal behaviour of the polymers, it seems that the presence of small polymer chains had a slight impact on the final ceramic yield, as the absence of them reduced the ceramic yield by only 2–3%. In summary, the polymer structure did not seem to be affected by the change of solvent, and AHPCS recycling could also be considered feasible at the end of the process.
The preparation of a composite via the PDC route is an alternative that allows for overcoming these drawbacks,11 and was used here to make B4C/SiC (AHPCS) composites. The influence of the nature of the solvent on the PDC route was studied. Fig. 2 illustrates the protocol used to produce the B4C/SiC specimens by spark plasma sintering (SPS).
B4C powders were studied by transmission electron microscopy (TEM) before and after mixing with AHPCS in solvent, as presented in Fig. 8. Before mixing, B4C particles present a native amorphous layer of boron hydroxide (H3BO3),36 identified with dotted line, that is less than 5 nm thick. After mixing with the polymer in the presence of a solvent, a thicker amorphous layer can be observed that is up to 10 nm thick, depending on the particle geometry. X-ray photoelectron spectroscopy (XPS) experiments (cf. ESI†) reveal the presence of Si in this amorphous layer, which confirms the presence of AHPCS at the particle surface. Owing to the interactions (chemical or physical) that can be developed between the polymer and the particles in the presence of the solvent, the polymer is dispersed and deposited homogeneously, leading to the surface functionalisation of the B4C particles.
Fig. 8 TEM images of boron carbide powder before (raw) and after functionalisation in different solvents: T, D or C. |
The sinterability of B4C/AHPCS was investigated using SPS, a non-conventional sintering technique in which densification is assisted by uniaxial pressure and heating by Joule effect with a current flowing through the sample. The specimens were subjected to the same thermomechanical sintering cycle that had been optimised in previous studies.24 The sintering steps were adapted in order to achieve an in situ conversion of the polymer from organic to inorganic, followed by the sintering step. The relative density of the specimens was slightly different depending on the functionalisation solvent (92% for T, 95% for D and 97% for C).11 This result could be explained by a modification of the surface potential depending on solvent, resulting in an agglomeration state of grains. The relative densities obtained are above 90%, which was the lower limit targeted in this study. For refractory or abrasive applications, a minimum relative density is required to achieve the mechanical properties required for the application. Moreover, the comparison of the microstructures of the samples, shown in Fig. 9, did not reveal any major differences regardless of what solvent was used. According to XRD analysis, the phases present in each case are B4C (dark grey) and β-SiC (light grey), as well as residual carbon (no visible) due to the non-stoichiometric Si/C ratio of the starting polymer. The samples can be described as a fine dispersion of β-SiC within the B4C matrix, with the presence of residual porosity.
Fig. 9 SEM (back-scattered electrons) observations of the sintered specimens after functionalisation in T, D or C. The dark grey phase corresponds to B4C and the light grey to β-SiC. |
On the other hand, FT-IR (ATR) characterisation of the solvent, before and after contact with the polymer, revealed no major polymer residue. In the context of the process industrialisation, this result means that the solvents can be recycled.
With regard to the targeted application, the solvent used for the functionalisation of B4C particles did not affect the final result. The type of interactions which are involved in the deposition of the polymer at the surface of B4C particles is not yet identified, and is still under investigation with complementary NMR and XPS investigations. The functionalisation using C gives the best relative density (97%) of the sintered B4C/SiC composites compared to D-(95%) and T-(92%). Indeed, the solvent did not affect the functionalisation step. The variation in relative density was attributed to the sintering step.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4ra04431b |
This journal is © The Royal Society of Chemistry 2024 |