3.1 Morphology and Structural characterization
In order to clearly observe the microscopic morphology of the R-AlN and M-AlN powders surfaces, the thickness of the coating layer was observed by TEM (as shown in Fig. 2). The R-AlN powders have an oxide layer on the surface due to long-term storage, with a thickness of 5.59 nm (as shown in Fig. 2b). In addition to the presence of Al, N elements, the R-AlN particles also uniformly distribute oxygen elements (as shown in Fig. 3). The oxygen element comes from the oxide layer on the surface, forming hydroxyl groups homogeneously distributed on the surface of the AlN powders, and these hydroxyl groups are the active sites for the reaction with NVP-IA. Figure 2c and 2d shows the surface of the modified AlN powders is coated with a layer of NVP-IA with a thickness of about 20 nm. And it is clearly observed that there is a silk-like substance implicated between the AlN powders. This is because NVP-IA is a macromolecular polymer, and the molecular chains are easily entangled with each other to stick multiple powders together. Figure 2d further shows the thickness of the local coating layer of the M-AlN powder, which ranged from 17.21 nm to 19.91 nm. From the M-AlN powders element distribution diagram in Fig. 4, it can be observed that in addition to Al and N, there are also C, O elements from the NVP-IA polymer, and P elements from the titanate coupling agent. It can be observed from Fig. 4d that the filaments connected between the spherical particles do not contain Al and N, but it can be clearly noticed that the main element is carbon. The C is derived from the carbon chain in NVP-IA, further confirming that this silky substance is NVP-IA. Figure 5 is a line scan of a partial magnification of R-AlN and M-AlN powders. In the R-AlN powder, the Al and N increased significantly at the edge of the powder, and the oxygen element began to increase slightly which originated from oxide layer (as shown in Fig. 5a). When M-AlN was line-scanned into the cladding layer, the carbon content in the cladding layer increased markedly, with a slightly increased amount of oxygen and nitrogen elements (as shown in Fig. 5b). This indicates that the main element of the cladding layer is carbon, as well as small amounts of nitrogen and oxygen, which corroborates with the elemental composition of the polymer NVP-IA. When continuing to scan into the interior of AlN particles, the weight percentage of carbon element decreases, and the content of Al and N increases obviously.
In order to explore the interaction between NVP-IA and the surface of AlN powders, the chemical structure of the powders was tested (as shown in Fig. 6). The NVP-IA, R-AlN and M-AlN were tested by using FT-IR spectroscopy in the range of 4000 − 400 cm− 1, respectively (as shown in Fig. 6a). In the infrared characteristic curve of NVP-IA, it can be observed that the -CH2- stretching vibration peak at 2975 cm− 1, the band at 1718 cm− 1 is assigned to the C = O stretching vibration of the carboxylic acid, 1659 cm− 1 is attributed to the C = O of NVP, 1290 cm− 1 is assigned to the C-N stretching vibration peak, and 1049 cm− 1 attributed to the bending mode of hydroxyl groups of the carboxylic acid in IA. [24]. Characteristic peaks of R-AlN were identified, a strong Al-N vibration was observed at 500-900cm− 1, a sharp Al-N peak at 1332cm− 1, and the absorption bands of –OH stretching at 3339cm− 1 appeared [25]. When R-AlN is wrapped with NVP-IA polymer, it is worth noting that the strong peaks at 1049 cm− 1 were completely unobservable and instead the weak peaks at 1018cm− 1 appeared of M-AlN, indicating a strong interaction between AlN particles and COOH of NVP-IA. These FT-IR spectral results confirm the successful grafting of polymers from the surface of AlN particles.
For more insight into the interactions between AlN and the NVP-IA, XPS measurements were performed on R-AlN and M-AlN to analyze the surface chemical structure and composition. Figure 6b shows the qualitative full spectrum analysis of R-AlN and M-AlN surfaces by XPS. In addition to Al 2P and N 1S, R-AlN also has C 1S and O 1S, among which the O 1S peak is the most noteworthy characteristic, indicating that the surface oxygen concentration of the R-AlN is relatively high, a large part of which is the result of atmospheric oxidation. A substantial amount of carbon is also being observed here, which is characteristic of uncleaned surfaces. M-AlN particles were performed on the areas around the peaks C 1S, O 1S, P 2P, Al 2P and N 1S. Upon the modification, the peak of C 1S increases remarkably from 33.12–61.30%. In contrast, the peak of Al 2P decreases substantially from 32.11–6.29%, and the peak of N 1S drop from 11.09–4.48% (Table 1). This is because XPS generally brings out photoelectron information within 10 nm of the surface, and the thickness of the M-AlN coating layer is mostly more than 10 nm, so the detected signals of Al and N are weakened, which come from the interior of the particles. Figure 6c shows the C 1s peak of R-AlN and M-AlN, fitted by a multipeak Lorentzian fitting program. The C 1S electron binding energy spectrum of R-AlN is derived from the contaminated carbon adsorbed in the air. High-resolution XPS scans of the C 1S peaks for the M-AlN can be deconvolved into four Gaussian-Lorentzian subpeaks, centered at 284.8 eV, 285.91 eV, 287.03 eV and 288.64 eV. Where 285.91 eV corresponds to the C-N chemical bond of the VP. The N 1S data and the fitted subpeaks from a sample of AlN are shown in Fig. 6c. The N 1S peaks of R-AlN can be determined as 396.48 eV and 398.25 eV, corresponding to Al-N and N-C, respectively. The intensity of the Al-N peak of M-AlN is significantly weaker compared to the R-AlN, and the 399.96 eV corresponds to the tertiary amine C-N bond of NVP-IA [22, 26, 27].
Table 1
XPS chemical composition of samples
Sample
|
Al 2P(at%)
|
N 1S(at%)
|
C 1S(at%)
|
O 1S(at%)
|
P 2P(at%)
|
R-AlN
|
32.11
|
11.09
|
33.12
|
23.68
|
\
|
M-AlN
|
6.29
|
4.48
|
61.30
|
26.54
|
1.39
|
3.2 Characteristics of hydrolysis resistance
A certain amount of AlN powders was soaked in deionized water, dispersed uniformly to form a suspension, and the hydrolysis resistance was measured at 25°C. In Fig. 7a, it can be found that the pH value of R-AlN powder increased rapidly within 12h, from 7.25 to 9.43, and reached 10.28 after 36h. The R-AlN suspension generated Al(OH)3 and NH3 at 25°C, and ammonia gas was partly dissolved in water to produce OH−, leading to an increase in the pH of the suspension and partly released to produce an irritating odor, which was similar to the results of Knel's[20] study. The initial pH of the M-AlN suspension was 4.12, which was lower than R-AlN, due to the effect of two unsaturated carboxylic acids in itaconic acid. After soaking for 276 h, the pH of the M-AlN suspension maintained around 4.22, which was similar to the initial value. No irritating odor was produced during the whole process, indicating that M-AlN has good stability after long-term soaking in water. R-AlN, M-AlN, R-AlN soaked in water for 12 hours (R-AlN-12), and M-AlN soaked in water for 276 hours (M-AlN-276) were respectively subjected to XRD tests (as shown in Fig. 7b). Compared with the Al(OH)3 standard card, the diffraction peaks of Al(OH)3 can be observed in R-AlN-12, while the AlN diffraction peaks remain unchanged in M-AlN-276. This is the consistency with the pH change, where R-AlN is easily hydrolyzed to form Al(OH)3, while M-AlN has excellent resistance to hydrolysis.
Figure 8 is the SEM image and EDS mapping spectrum of the R-AlN powder. It can be found that the original powder is oval with a smooth surface, and some AlN particles agglomerate together to form large particles. After immersion in water for 12 h at 25°C, the R-AlN particles aggregated to form larger cluster sizes, which were hydrolyzed to form rod-like structures of about 2–5 µm in length (Fig. 8b). Compared to the Al element, the N distribution is substantially reduced, while the O element content is elevated. Table 2 shows that the Al, N, and O contents are 31.35%, 10.44%, and 58.11%, respectively. As shown in Fig. 9 of M-AlN particles, the NVP-IA macromolecular chains are easily cross-linked to form a network structure. After immersion in water for 276 h, M-AlN kept spherical morphology unchanged and did not produce a rod-like structure of Al(OH)3. The Al, N elements were uniformly distributed, the C, O and P elements from the modifier were also present, with the contents of 10.84%, 2.15% and 0.31%, respectively (Table 2). This indicates that M-AlN particles has not undergone hydrolysis reaction with its excellent core-shell structure.
Table 2
EDS elements composition of samples
Sample
|
Al (at%)
|
N (at%)
|
C (at%)
|
O (at%)
|
P (at%)
|
R-AlN-12
|
31.35
|
10.44
|
\
|
58.11
|
\
|
M-AlN-276
|
46.56
|
40.14
|
10.84
|
2.15
|
0.31
|
3.3 Dispersion performance and thermal stability of modified powders
The zeta potential can reflect the dispersion behavior of powder particles in the water, The higher the Zeta potential of the particle surface, the greater the surface charge density, the stronger the electrostatic repulsion between the particles, and the more conducive to the stable dispersion of the slurry. When the pH of R-AlN is 7.3, the Zeta potential is 0 (Fig. 10a), and it is easy to attract each other through van der waals attraction to form clusters. Due to the large number of hydroxyl groups on the surface of R-AlN powders, it is positively electrically charged by adsorbing H+ under acidic conditions, and negatively charged by adsorbing OH− under alkaline conditions. There is a substantial amount of COOH exists on M-AlN surface, which is negatively charged after dissociation in water, thus resulting in an increase in the negative value of M-AlN. This caused a variation in the zeta potential, with a shift of the equipotential point toward the acidic range, which is located at pH = 4.2. At the equivalent base pH, the M-AlN powders had a relatively larger absolute value of zata potential, the double electric layer repulsion between the powders was strengthened, which was beneficial to improve the dispersion stability of the slurry. Figure 10b presents the thermal weight loss curves of M-AlN powders under nitrogen atmosphere. The weight loss of the specimens was observed to be 1.61% below 183°C, it was mainly due to the detachment of adsorbed water. The weight loss rate of M-AlN powders was accelerated in the region of 183–573°C with a weight loss ratio of 7.03%. The sample mass remained stable at 90.01% at 1100°C, which is in accordance with our initial additive of 10%. The NVP-IA is completely vaporized and removed before 1100°C, which has no influence on the sintered AlN ceramics.