INTRODUCTION
The escalating demand for 5th generation mobile communication technology necessitates the development of enclosure materials for electronic products, including cell phones and smart wearable devices, to satisfy stringent properties standards[1,2]. These include high thermal conductivity (k), with k > 20 W/(m·K), for effective equipment cooling; low dielectric loss (tanδ), with tanδ < 10-3, for optimal signal transmission, high mechanical properties, with fracture toughness (KIC) > 10 MPa·m1/2 and high flexural strength (σ) > 700 MPa, to ensure sufficient damage tolerance. Such stringent requirements render plastic-based, metal-based and glass-based materials and traditional structure ceramics such as zirconia (ZrO2) and alumina (Al2O3) unsuitable. Alternatively, silicon nitride (Si3N4) has emerged as a leading contender for use in 5G electronic products, due to its superior overall performance, as depicted in Supplementary Figure 1. However, the issue of single color (typically grey) coupled with control over mechanical properties remains a significant challenge for large-scale commercialization[3,4]. Therefore, there is an increasing urgency to address both color regulation and improvement of mechanical performance.
In the realm of color regulation, rare earth metal ions (Re3+) exhibit the capacity to absorb light in a manner that is triggered by the visible light excitation of a 4f→4f electron transition[5,6]. Notably, among these rare earth metal ions, Eu3+ ions have the ability to transmute ultraviolet (UV) radiation into a strong orange-red emission. They are also highly sensitive to their coordination environment due to the unique combination of intra-configurational 4f→4f transitions that are either magnetic or electric-dipole in nature[7,8]. These attributes position Eu3+ as a promising candidate for use as a colorant in the creation of orange, red, or similar orange-red Si3N4 ceramics. However, conventional coloring strategies such as dissolving Eu3+ ions within lattices to color ceramics, which has been successfully employed in zirconia[9-11], are not applicable to Si3N4 coloring. This is primarily due to the robust covalent nature of the Si-N bond in β-Si3N4[12], coupled with the substantial radius difference between Si4+ (0.41 Å[13]) and Eu3+ (1.06 Å[14]). In this case, the design of Eu3+-doped second crystal phases or new local structures as chromophores is the alternative way. Besides, it is reported that the introduction of second phases or novel local structures can also act as a feasible and practical way for improving mechanical properties[15-17].
The selection of sintering additives in the liquid phase sintering process of Si3N4 is critically important due to their significant influence on the formation of secondary phases and novel local structures. Compared to the conventional sintering additives [e.g., yttrium oxide-aluminum oxide (Y2O3-Al2O3)[18], ytterbium oxide-aluminum oxide (Yb2O3-Al2O3)[19], MgO[20], yttrium aluminum garnet (YAG, Y3Al5O12)[21], etc.], we develop YAG-MgO as a sintering additive for its lower eutectic temperature (< 1,613 K[21]) than YAG or MgO via the YAG-MgO-SiO2-Si3N4 reaction. The lower eutectic temperature offers more time for α-Si3N4 dissolution in the eutectic liquid phase, and then will promote the growth of high aspect ratio β-Si3N4 grains[22-24] and the formation of second crystal phases or novel local structures.
In this contribution, we propose using phase separation and crystallization in liquid phases supported by Eu2O3-YAG-MgO to engineer hollow structures for achieving the color change from yellow to orange-red, while ensuring excellent mechanical properties of Si3N4. The combined microstructural characterizations of scanning electron microscopy (SEM), high-resolution transmission electron microscopy (HRTEM), scanning transmission electron microscopy (STEM), energy dispersive X-ray spectroscopy (EDS), electron energy loss spectroscopy (EELS) and cathodoluminescence (CL) were utilized to elucidate the crystallographic characteristics and chemical composition of the hollow structures, including morphology, chemical composition, distribution and luminescent characteristic. Based on these characterizations, the formation mechanism of the hollow structure in Si3N4 ceramics, as well as its impact on their color and mechanical properties, was meticulously analyzed and discussed.
MATERIALS AND METHOD
Raw material
All raw materials were utilized in their analytical grade (99+%), with no additional purification required. Commercial α-Si3N4 powders (SN-E10, O 1.08 wt%, BET 9.64 m2/g, α > 95 wt%) were procured from UBE Industries Ltd, located in Yamaguchi, Japan. Additionally, other chemicals such as Eu2O3, Y2O3, Al2O3, and MgO were acquired from Haoxi Research Nanomaterials, Inc. situated in Shanghai, China.
Synthesis of YAG (Y3Al5O12) powders
YAG powders were synthesized by using purity Y2O3 and Al2O3 powders via a solid-state reaction method as reported[25,26]. The mole ratio of Y2O3 to Al2O3 powders was maintained at 3:5. Following this, the mixed powders were sintered at 1,400 °C for 3 h with the heating rate of 2 °C·min-1 in air using a muffle furnace. The resultant powders after sintering were sieved through a 100-mesh screen to spare. The sintered powder was subjected to X-ray diffraction analysis, which revealed the formation of only the YAG phase [Supplementary Figure 2].
Preparation of colored Si3N4 ceramic
Si3N4 Ceramic was prepared by gas pressure sintering using commercial α-Si3N4 powders as the primary raw materials, MgO powders and as-prepared YAG powders as sintering additives, Eu2O3 powders as colorants, Polyvinyl Butyral (PVB) as a binder, and C2H5OH as a solvent. The mass fractions of YAG and MgO powders were fixed at 4 and 2 wt%, respectively. The quantities of PVB and C2H5OH incorporated were 0.8 to 1 wt% and 200 wt% of the total powder, respectively. Eu2O3 powders, with varying contents of 2, 4, 5, 6, 7, 8, and 9 wt%, were uniformly blended with raw materials and other additives. This mixture was then subjected to a planetary ball mill for a duration of 2 h. The obtained slurry was subsequently dried and subjected to sieving through a 100-mesh screen. The fine powder is inserted into a stainless steel mold (50 mm × 50 mm) and then pre-pressed at 25 MPa using an oil press. Following demolding, the sample is encased in a multi-layer vacuum compression bag to prevent water ingress and subsequently vacuumed. The vacuum bag is then directly placed into the water chamber of the cold isostatic press and subjected to a pressure of 250 MPa for a duration of 90 s. Upon completion of the cold isostatic process, the vacuum bag is removed to extract the sample, which is then immediately subjected to debonding and sintering. Si3N4 ceramic samples were synthesized by sintering at 1,850 °C, with a heating rate of 3 °C·min-1, a holding time of 2 h, and a nitrogen pressure of 0.6 MPa. This process was preceded by the burnout of the binder at 900 °C for 1 h under vacuum.
Characterization
The bulk densities (ρ) of the sintered samples were ascertained utilizing the Archimedes method, conducted in distilled water. The Vickers hardness (H) was measured using a Vickers microhardness tester (FV-700, Future-Tech, Japan). This measurement was conducted three times on a polished surface, with the load and holding time set at 10 kg and 10 s, respectively. The flexural strength (σ) was determined using a three-point bending test (Model 5566, Instron Co., High Wycombe, UK). This involved a span of 30 mm and a press speed of 0.5 mm/min, utilizing machined rectangle bars (3 mm × 4 mm × 36 mm) with a polished surface. The data from each specimen were averaged across six tests. The fracture toughness (KIC) was measured by the single-edge notched beam method (SENB) at room temperature with a crosshead speed of 0.05 mm/min for a span of 24 mm. The thermal diffusivity (λ) was measured using a laser thermal conductivity meter (LFA-457, Netzsch, Germany). The dielectric loss (tanδ) was measured at a frequency of 1 GHz with an impedance meter (E4991B), in accordance with IPC-TM-650 2.5.5.9-1998. This standard stipulates that the specimens should have dimensions of 50 mm × 50 mm × 0.9 mm.
The bulk samples were subjected to phase identification via X-ray diffraction (XRD, D8 Advance, Bruker, Germany). The data collection for the diffraction was conducted within a range of 10°-80° 2θ, employing a scanning step of 10°/min. Prior to examination with SEM (Magellan 400, FEI, USA), the composites were ground using a resin-bonded diamond wheel (SD54R75B1/3) and polished with varying particle sizes (7, 5, 2.5, and 1 μm) of diamond slurry to achieve a surface finish of 0.02 μm. The microstructure of the samples was examined using transmission electron microscopy (TEM, JEM-2100, JEOL, Japan), encompassing STEM and HRTEM. Within the framework of STEM, both compositional and valence state analyses were conducted utilizing EDS and EELS, respectively. The valence state of the Eu element was also documented using X-ray photoelectron spectroscopy (XPS, ESCALAB 250, Thermo Fisher Scientific, USA).
The optical reflectance within the wavelength range of 380 to 750 nm, along with the Commission International del’Eclairage (CIE) chromaticity coordinates, was ascertained utilizing a Spectrophotometer, which was procured from X-rite in USA. Before color measurement, calibration was conducted utilizing the instrument’s standard plate, encompassing both white calibration and zero position calibration. The mean values of L*, a*, and b* were computed to denote the chromaticity value of each sample following three surface measurements. This was achieved by positioning the sample in front of the measuring aperture, which has a diameter of 2 mm. Photoluminescence (PL) emission spectra were performed by a fluorescence spectrometer (Hitachi F-4600, Japan). The luminescence position inside the sample was determined by Scanning Electron Microscopy-Cathodoluminescence (SEM-CL, Gemini450, ZEISS). The optical energy gaps, denoted as Eg, were determined using the Wood and Tauc equation[27,28]. This method involved a transformation of diffuse reflectance spectra to estimate the value of Eg:
where α and hν represent the absorption coefficient and photon energy, respectively, A is a constant, Eg denotes the optical band gap, and n takes on values of 1/2 or 2 for direct allowed and indirect allowed transitions, respectively. According to literature[29,30], silicon nitride displays an optical absorption spectrum that is dominated by the indirect absorption process (n = 2).
RESULTS AND DISCUSSION
During the liquid-phase sintering process of Si3N4 ceramics, the assembly of Eu ions, attributed to their high field strength and larger radius, leads to the segregation of the glass phase[31]. This phenomenon is depicted in Supplementary Figure 3. The liquid phase, enriched with Eu, encapsulates the pores, thereby forming a hollow structure at the grain boundary. Furthermore, the generation of these hollow structures within β grains transpires during the solution-reprecipitation process[12], which can be divided into two stages: initially, the α to β phase transformation in the liquid phase is facilitated by the homogeneous mixing of sintering additives (YAG-MgO) and colorants (Eu2O3) between 1,450-1,800 °C; all transformations into β are completed within this stage. Subsequently, the second stage involves the solution of small β grains and their reprecipitation on larger β grains at elevated temperatures (> 1,800 °C), concurrently introducing hollow structures into the larger β grains [Figure 1A]. A more comprehensive depiction of the second step is illustrated in Figure 1B and C. Two or more adjacent, small β grains with identical orientations are susceptible to dissolution, merging, and growth into larger β grains at elevated temperatures. Concurrently, the gas and liquid phases within the grain boundary are compressed into the β grains at the onset of the merger, and then persistently expelled to the grain boundary as the merger progresses. However, when sintering or grain growth is impeded, a limited amount of gas and liquid phase remains in the larger β grains. The residual liquid phase primarily consists of heavier Eu-clusters due to their lower diffusion rate and follows the growth of the β grain lattice, subsequently exhibiting a hexagonal morphology akin to that of the β grain. In contrast, the residual gas is enveloped by the hexagonal liquid phase, forming a hollow structure. Figure 1D illustrates the coloring process for Si3N4 ceramics, wherein Eu ions within the hollow structures absorb blue-green light and emit red-orange light due to electronic transitions under photon excitation.
Samples with various Eu2O3 contents (2, 4, 5, 6, 7, 8, and 9 wt%) were designated as SEu-2, SEu-4, SEu-5, SEu-6, SEu-7, SEu-8, and SEu-9, respectively. SEM revealed that only rod-shaped grains and grain boundary liquid phases with distinct contrasts could be observed in the samples SEu-4, SEu-5, and SEu-6 [Supplementary Figure 4], aligning with prior research[32-34]. Interestingly, STEM detected the hollow structures ranging from ~5 to ~110 nm in both grains and grain boundaries, as depicted in Figure 2A-D and Supplementary Figures 5-7. Additionally, HRTEM analyses demonstrated that the hollow structure in β-grains exhibits a hexahedral morphology identical to that of the β-grains [Figure 2E-H and Supplementary Figure 7]. The corresponding Fast Fourier Transformation (FFT) patterns of region 1 [Figure 2E and Supplementary Figure 7G] and region 2 [Supplementary Figure 7H] suggest two hexagonal phases, but the lattice spacing, daverage = 0.648, calculated from the FFT1 pattern, aligns with the zone axis [001], which falls in that of β-Si3N4 phase[35,36], whereas the lattice spacing, daverage = 0.654, calculated from FFT2 pattern, is slightly different from that of the β grain. These findings suggest that the hollow structure is epitaxially grown on original β grains, albeit with minor distortions to the original crystal structure. Notably, the corresponding FFT4 pattern of the hollow structures at grain boundaries displays a halo indicative of an amorphous nature [Figure 2H], thus preventing the detection of the hollow structure by XRD patterns [Supplementary Figure 8] due to its amorphous nature. The chemical composition of the hollow structures is illustrated in Figure 2I and Supplementary Figure 9. Elemental maps reveal that the core, characterized by its element-free and rounded morphology, was classified as a pore. Conversely, the shell, which demonstrated a pronounced contrast difference with the β grains due to its enrichment in Eu elements, was identified as an Eu-rich secondary phase.
The electron valence states and luminescence of Eu ions in Si3N4 ceramics were examined using XPS, STEM-EELS, and SEM-CL techniques, as depicted in Figure 3A-C and Supplementary Figure 10. The XPS analysis of the SEu-5 sample revealed that both Eu3+ and Eu2+ signals, identified as Eu3+:3d5/2,3/2 and Eu2+:3d5/2,3/2, respectively, were readily found [Supplementary Figure 10]. This suggests a reduction of some Eu3+ to Eu2+ in an N2 atmosphere. The STEM-EELS results indicated that the grain boundary glass phase (region 1) and the hollow structure (region 2) predominantly consisted of EELS signal pairs at 1,135.8/1,164.0 eV, attributed to Eu3+[37], but no EELS peaks were observed in the β grains (region 3). The SEM-CL spectra of the β-grain and grain boundary phases in the SEu-5 displayed a broad emission peak and four narrow emission lines [Figure 3C]. The broad emission peak at 530 nm corresponds to the 5d-4f transition of Eu2+. Additionally, several distinct lines at 593, 616, 656, and 697 nm are associated with the 5D0→7F1,5D0→7F2, 5D0→7F3 and 5D0→7F4 transitions of Eu3+[38,39], respectively. These CL emission peaks likely arise from the Eu-rich hollow structure present in grain boundary and β-grain. Consequently, the Eu2+/Eu3+-doped hollow structures present within grain boundaries and β-grains serve as a primary chromophore in Si3N4 ceramics. Furthermore, the distribution density of these hollow structures could be regulated via control of Eu2O3 content for color regulation of Si3N4 ceramics. The distribution density is defined as the average number of hollow structures per unit area (nm-2) in the context of STEM, where 70 regions were randomly selected in each sample at magnification photography for counting. As shown in Figure 3D-G, the distribution density of the hollow structures progressively increases with Eu2O3 content. A higher distribution density implies more chromophores, leading to increased absorption of blue-green light. Further details will be explored in Figure 4.
The influence of hollow structures on the optical properties of Si3N4 ceramics was examined, as depicted in Figure 4. Yellow and orange Si3N4 ceramics were synthesized using varying Eu2O3 concentrations. As the Eu2O3 concentration escalated from 2 to 9 wt%, the color transitioned from yellow to orange-red, with a heightened chroma [Figure 4A]. Figure 4B presents that the L* value diminishes while a* and b* values rise in response to the diminished brightness and a color shift towards reddish-yellow. The emission spectra of all samples by monitoring 465 nm excitation are shown in Figure 4C, where the intense broadband emission peaks for samples SEu-2, SEu-4, SEu-5, SEu-6, SEu-7, SEu-8, and SEu-9 occurred at 587, 605, 609, 612, 622, 624 and 626 nm, respectively, corresponding to the 5d→4f transition of Eu2+[40]. A relatively faint yet broad emission is discernible at 530 nm, which can also be attributed to Eu2+ ion emissions. A more pronounced narrow emission peak at 700 nm is evident later in the spectrum. The 5D0→7F4 transitions of Eu3+ ions account for this narrowband emission[39]. Additionally, samples SEu-2, SEu-4, SEu-5, SEu-6, SEu-7, SEu-8, and SEu-9 absorb blue or green light and exhibit strong reflectance in the yellow or orange wavelengths at 587, 588, 589, 593, 599, 600, and 601 nm, respectively [Figure 4D]. The band gap of Eu-doped Si3N4 ceramics falls within the yellow and orange light region (Eg: 2.03-2.4), as depicted in Figure 4E, and an increase in Eu2O3 content, a decrease from 2.24 to 2.04 in Eg values correspondingly [Figure 4F]. In conjunction with the findings presented in Figure 3D-G and the subsequent analysis of the optical properties of Si3N4 ceramics varying in Eu2O3 content, it can be deduced that the orange hue of Si3N4 ceramics exhibits a positive correlation with the distribution density of hollow structures. Specifically, as the distribution density of these structures increases, there is a corresponding rise in chromophore content. This leads to an enhanced absorption of blue-green light and a consequent increase in reflection of yellow-red light, resulting in a deeper orange coloration of the samples. Consequently, the coloration of Si3N4 ceramics can be effectively modulated by manipulating the distribution density of the hollow structures.
The impact of the hollow structures on the mechanical properties of Si3N4 ceramics was examined, as depicted in Figure 5 and Supplementary Figure 11. The KIC and σ initially increased, subsequently decreasing in line with the trend of relative density [Supplementary Figure 11]. The microstructures of the sample SEu-5, as revealed by SEM and TEM micrographs, exhibit elongated grains, grain pull-out, curved crack growth paths, and pronounced intergranular fracture [Supplementary Figures 12 and 13]. These attributes promote the utilization of crack energy and augment fracture resistance[41-44]. Consequently, the sample SEu-5 demonstrates superior mechanical properties, boasting a flexural strength of 915.5 ± 43.1 MPa and a fracture toughness of 11.1 ± 0.3 MPa·m1/2 [Supplementary Figure 11]. The strain distribution depicted in Figure 5A is derived from geometric phase analysis (GPA), with the corresponding strain mapping illustrated in Figure 5B. Around the hollow structure, symmetrical compression-tension strain pairs are induced due to lattice mismatch [Figure 5C], which presumably leads to significant local internal stresses and a stress shielding effect[45], both of which contribute to enhanced friction resistance against motion or microcrack accumulation. Consequently, the strain couples associated with compressive-tensile forces surrounding hollow structures may present an additional avenue for improving the toughness of bulk ceramics. In this case, we evaluated the physical properties of reference ZrO2 ceramics, Al2O3 ceramics, Si3N4 ceramics and the sample SEu-5. These properties included dielectric loss (tanδ), thermal conductivity (k), density (ρ), hardness (H), flexural strength (σ) and fracture toughness (KIC) [Figure 5D and E, Supplementary Figure 1]. Our findings indicate that Si3N4 ceramics can compensate for the mechanical deficiencies of Al2O3 ceramics and the thermal and dielectric shortcomings of ZrO2 ceramics, thereby exhibiting superior overall performance [Figure 5D]. In particular, the sample SEu-5 demonstrates superior physical properties compared to the referenced Si3N4 ceramics, with its tanδ, k, ρ, KIC, σ, and H reaching 8.24 × 10-4, 27 W/(m·K), 3.24 g·cm-3, 11.1 MPa·m1/2, 915.5 MPa, and 14.1 GPa, respectively [Figure 5E]. The tanδ and ρ of the SEu-5 are ~100 and ~2 times lower than those of ZrO2 ceramics, reducing signal transmission absorbance and weight. The k of the SEu-5 is ~10 times higher than that of ZrO2 ceramics, accelerating equipment cooling. The KIC and σ of the SEu-5 are ~2 times that of Al2O3 ceramics, enhancing damage tolerance and yield improvement.
CONCLUSIONS
In summary, we introduce a microstructure design that involves the construction of Eu-doped hollow structures to synthesize Si3N4 ceramics. These ceramics exhibit a color transition from yellow to orange-red and possess superior mechanical properties. The hollow structure is characterized by its amorphous nature, with the core corresponding to the pore phase and the shell to the Eu-rich liquid phase. This structure functions as one of the dominant chromophores due to the 5d→4f transition of Eu2+, which is coupled with the 5D0→7FJ transition of Eu3+ under photon excitation. Specially, the distribution density of this chromophore correlates positively with the depth of orange. Our sample, SEu-5, demonstrates optimized mechanical properties, with a flexural strength of 915.5 ± 43.1 MPa and a fracture toughness of 11.1 ± 0.3 MPa·m1/2. Factors such as grain pull-out, bridging and deflection, pronounced intergranular fractures, and the stress field surrounding the hollow structure collectively enhance the mechanical properties by reducing crack energy consumption.
DECLARATIONS
Acknowledgments
The authors would like to express their gratitude to Mr. Jingwei Feng for his invaluable support and guidance on the use of the transmission electron microscope.
Authors’ contributions
Conceiving the idea and designing the experiment: Liu N
Performing sample prep, data acquisition, and data analysis: Liu N, Hu T
Advising the scientific discussion on this research: Fu Z, Wang Z, Xu F, Dong S
Financial support: Duan Y
Supervising this research: Hu T, Zhang J
All authors participated in the writing of the manuscript.
Availability of data and materials
Not applicable.
Financial support and sponsorship
This work was supported by the National Natural Science Foundation of China (grant number: 52102082, Program Manager: Dr. Duan Y), Shanghai Science and Technology Committee (grant number: 21YF1454500, Program Manager: Dr. Duan Y), the State Key Laboratory of High Performance Ceramics and Superfine Microstructure of Shanghai Institute Ceramics, Chinese Academy of Sciences.
Conflicts of interest
All authors declared that there are no conflicts of interest.
Ethical approval and consent to participate
Not applicable.
Consent for publication
Not applicable.
Copyright
© The Author(s) 2024.
Supplementary Materials
Comments
Comments must be written in English. Spam, offensive content, impersonation, and private information will not be permitted. If any comment is reported and identified as inappropriate content by OAE staff, the comment will be removed without notice. If you have any queries or need any help, please contact us at support@oaepublish.com.