State-of-the-art thermal analysis and thermophysical properties testing instruments can help manufacturers of Si3N4 ceramics optimize the sintering process to achieve maximum product quality.
High-quality Si3N4 wear components. Photo courtesy of Ceradyne Inc., Costa Mesa, Calif.
Silicon nitride (Si3
) plays a dominant role among nitride ceramics, primarily due to its good thermal shock resistance and high corrosion resistance. It also provides extremely high stability, high viscosity, excellent abrasion resistance, low thermal expansion, medium thermal conductivity and very good chemical resistance. This combination of properties yields a ceramic that withstands extreme application conditions and is ideally suited for machinery components with very high requirements for reliability and mechanical loads.
For the manufacture of dense Si3N4, diverse sintering additives (aluminum oxide [Al2O3], yttrium oxide [Y2O3], magnesium oxide [MgO], etc.) are generally added to the original powder. After the shaping process, the green body is sintered at temperatures between 1950 and 2200 K (3050 and 3500°F). In many production processes, an increased nitrogen pressure is frequently used in the sintering furnace above 2000 K to counteract the decomposition of the silicon nitride and considerably reduce the pore size in the ceramic. However, few manufacturers understand what happens to silicon nitride during the various stages of the sintering process; as a result, it can be difficult to make changes to further optimize the results.
State-of-the-art thermal analysis and thermophysical properties testing instruments can be used to overcome this challenge. By obtaining accurate measurements of properties such as thermal expansion, density change, phase transitions, specific heat, thermal diffusivity and thermal conductivity throughout the sintering process, such instruments enable manufacturers to confidently adjust their processes to maximize product quality.
Figure 1. Schematic design of the dilatometer used in the experiments (NETZSCH DIL 402 C/7, 2250 K version).
A Sample Analysis
To prove the instruments' capabilities, researchers at NETZSCH-Gerätebau GmbH analyzed samples of debinded silicon nitride throughout the sintering process. The sintering powder consisted of fine-grained silicon nitride, with an average grain size of 0.7 mm. Al2
were added at 5 mass% each as sintering additives, and the raw material still contained small amounts of silicon dioxide contamination on the Si3
powder surfaces. The green body particles were mainly from the a-phase (trigonal structure), which transfers into the b-phase (hexagonal structure) during sintering. In actual production processes, the ceramic material is sintered up to ~1973 K under normal pressure in a nitrogen atmosphere. To reduce the pore sizes, the pressure is increased to approximately 1 Mpa, and the temperature is increased to a maximum of 2023 K. The influence of the pressure increase was not examined in this project.
To determine the thermal expansion and density change of the material, the samples were analyzed with a high-temperature dilatometer1 in a nitrogen atmosphere. The system was equipped with a graphite furnace and sample holder system and allowed measurements to be carried out between room temperature and 2250 K. The schematic design of the instrument used for the tests is shown in Figure 1. To reduce the potential for the samples to react with the graphite, the samples were contained in protective aluminum nitride tubes. A W/Re thermocouple was used to measure the temperature at the sample.
A high-temperature differential scanning calorimetry (DSC) instrument2 was used to characterize the phase transitions and measure specific heat. The measurements were carried out at 20 K/min between room temperature and 1773 K in a nitrogen atmosphere.
To measure the thermal diffusivity of small sample dimensions at high temperatures, the laserflash method was used.3 In this technique, a plan-parallel sample is heated in a furnace to the required measuring temperature. After temperature equilibration, the front surface of the sample is heated by a short laser pulse (generally <1 ms in length). The energy/temperature rise penetrates the sample and leads to a temperature rise on the rear surface of the sample, which is measured with an infrared detector vs. time. From the temperature vs. time behavior, the thermal diffusivity can be determined.4
Figure 2. The length change and rate of length change of a silicon green body at a heating rate of 5 K/min measured with the DIL 402 C/7.
Thermal Expansion and Density Change
The results for the length change and rate of length change at a heating rate of 5 K/min are shown in Figure 2. The measurement was conducted from room temperature to 2150 K. At 1553 K (extrapolated onset), a first shrinkage step of 1.12% was obtained due to the formation of a new liquid phase within the sintering body. The liquid phase was formed by the sintering additives and the silicon dioxide that was present as contamination on the silicon nitride particles.
Figure 3. The volume and density change of a silicon nitride green body at a heating rate of 5 K/min (room-temperature density: 1.749 g/cm3).
The main shrinkage step occurred at approximately 1681 K. At about the same point, the a-silicon nitride dissolved in the liquid phase and segregated again as b-silicon nitride. This means that the high amount of a-silicon nitride in the green body slowly disappeared during the sintering process, while the amount of b-silicon nitride increased. Connected to this effect was a significant length and density change. A total shrinkage of 14.93% was determined in this range, and a further shrinkage rate of 1.10% was measured above 2009 K-probably due to the reactions between the silicon nitride and the additives (reduction processes). Many manufacturers believe that decomposition of the silicon nitride can also occur in this range. However, the sample length stabilized again above 2080 K, and no significant length change can be seen-indicating that decomposition did not occur.
Figure 4. The length change and coefficient of thermal expansion (reference temperature: 298 K) of sintered silicon nitride at a heating rate of 5 K/min.
The volume and density changes, which were calculated from the measured length change and room temperature density of 1.749 gm/cm3
, are shown in Figure 3. The volume during sintering was reduced by approximately 40%. The typical maximum sintering temperature for the material is between 1873 and 2023 K.5
In the corresponding temperature range, a value of 2.93 g/cm3
was determined for the density of the sintered specimen.
Figure 4 shows the length change (linear thermal expansion) and average expansion coefficient (reference temperature: 298 K) of a silicon nitride specimen sintered at 1975 K. The measurement was carried out at 5 K/min to 1650 K. It can be seen that the sample length increased with an increasing rate of expansion vs. temperature; therefore, the coefficient of thermal expansion also increased with temperature. A change in the expansion rate was determined above 1200 K. Since pure silicon nitride should show no transitions in this range, this effect may be due to the influence of additives. The decrease in the rate of length change above 1600 K might be caused by post-sintering.
Figure 5. The specific heat flow rate of a silicon nitride green body between 700 and 1750 K.
Specific Heat and Transformation Energetics
The measured specific heat flow rate between 700 and 1750 K on the green material is shown in Figure 5. At 1285 K, an exothermal reaction with an enthalpy of -16.4 J/g was detected. The peak temperature was at 1419 K. The explanation for this effect is the formation of new phases from the sintering additives and silicon dioxide impurities. Above 1500 K, an endothermal reaction with an enthalpy of 18.4 J/g was detected. The peak temperature was 1591 K, and a further endothermal reaction started at 1710 K, possibly due to phase transitions in the additives and silicon nitride.
Figure 6. The apparent specific heat of a silicon nitride green body between room temperature and 1700 K.
Figure 6 shows the apparent specific heat (specific heat and overlapping energetic effects) of the silicon nitride green body. Between room temperature and 1250 K, the measured results continuously increased with temperature and corresponded to the typical behavior of the specific heat for such materials. The results agree within 5% of corresponding literature values.6
Above 1250 K, the specific heat was overlapped by the previously discussed effects. In this temperature range there was a tendency toward a further small increase. The analysis of the true specific heat (i.e., a separation of the specific heat and the overlapping effects) is possible by extrapolating the values at lower temperatures. The slight increase achieved by the extrapolation agrees well with the Debye Theory.7
Figure 7. The apparent specific heat of the sintered silicon nitride ceramic between room temperature and 1700 K.
Figure 7 shows the apparent specific heat of a silicon nitride sample sintered at 1973 K. The measurement was conducted between room temperature and 1700 K in a nitrogen atmosphere. Up to 1150 K, the values for the specific heat increased. The results were in the same range as for the corresponding green body. Between 1194 and 1222 K, a glass transition (step) was detected. This transition was due to a glass phase that was formed by the additives during the first thermal treatment. Beyond the glass transition, the specific heat was nearly constant up to 1396 K, at which point an exothermal effect was detected. This reaction can be explained by the crystallization of the amorphous contents (glass phase). The low crystallization enthalpy of -15.1 J/g is a clear indication that only a small amount of the sample material was amorphous at low temperatures.
Figure 8. The measured thermal diffusivity of a silicon nitride green body between room temperature and 1973 K.
The laser flash measurements were carried out between room temperature and 1973 K in steps of approximately 100 K. Five individual measurements were conducted at each temperature, again in a nitrogen atmosphere. The investigated green body had a thickness of 3.28 mm, a diameter of 18 mm and an initial density of 1.816 g/cm3
. The green body therefore had a slightly higher (>3.7%) initial density than the dilatometer sample. At 1973 K, the temperature was kept constant for one hour. The sample was then cooled and again measured up to 1973 K. The changes in the sample thickness were taken into consideration for the analysis of the results.
Figure 9. The measured thermal conductivity of a silicon nitride green body and the corresponding sintered ceramic between room temperature and 1973 K.
The thermal diffusivity results for the Si3
green body are shown in Figure 8. At room temperature, the thermal diffusivity was 0.34 mm2
/s. With an increasing temperature, the values first decreased slightly due to the decreasing thermal diffusivity within the silicon nitride particles. Between 800 and 1400 K, the thermal diffusivity values were approximately constant. Between 1400 and 1700 K, a considerable increase in the measured data was determined due to the formation of new phases (DSC test results). The new (liquid) phase led to a reduction of the contact resistances between the powder particles and an improvement of the heat transfer in the entire system. A further increase in the measured data can be seen in the main sintering range above 1700 K.
Figure 9 shows the thermal diffusivity for the green body, along with the values of the Si3
material sintered at 1973 K. As expected, the sintered material shows a considerably higher thermal diffusivity. The measured data continuously decrease with an increasing temperature. Due to the phase transitions (DSC results), a few (weak) steps can also be seen in the sintered ceramic above 1200 K. Approximately the same values for the sintered material were determined in the range of the maximum temperature as for the green body.
Figure 10. The calculated thermal conductivity of a silicon nitride green body and the corresponding sintered ceramic between room temperature and 1973 K.
Figure 10 shows the thermal conductivity values for the silicon nitride (both the green body and sintered material), which were calculated by multiplying the density, specific heat and thermal diffusivity. In the temperature range up to 1000 K, only slight changes in the thermal conductivity were obtained for the green material. Between 1000 and 1500 K, the values increased with the temperature due to the increasing density and thermal diffusivity, while a considerable increase in the thermal conductivity was determined in the sintering range. However, the measured values of the sintered material continuously decreased with the increasing temperature.
A thermal conductivity value of 12.8 W/(m•K) was determined at room temperature. While this is below the values (16 W/(m•K)) that are typically found in literature for corresponding materials,8 it can be explained by the low sintering temperature, as well as the fact that the sample was not sintered at high temperatures under pressure (1 MPa), as is usually the case in a production process. The temperature dependence of the thermal conductivity followed the course as expected according to the phonon model. Between 1100 and 1700 K, smaller overlapping steps occurred due to the phase transitions mentioned.
With the help of the measured data, a simulation of the temperature gradient in a silicon nitride green body was carried out through finite element calculation using ANSYS‘ simulation software.9 An infinitely long cylinder with an initial diameter of 30 mm was used for the product geometry, and a constant heating rate of 5 K/min for the outer surface area was used for simulation. The heat input was adjusted in such a way that this heating rate was constant over the simulation range between room temperature and 2000 K.
Figure 11 shows the results for the temperature differences between 374 and 1951 K vs. the actual radius. Up to 1500 K, nearly no changes in the diameter of the cylinder occurred. However, a large change in diameter (from 30 mm to 25.7 mm) appears in the sintering range above 1500 K.
Upon initial inspection, it seems unusual that the temperature differences between the center and surface increase from 374 K to 749 K. However, this can be explained with the decrease in the thermal diffusivity and conductivity of the material vs. the temperature in the lower temperature range. In the sintering range above 1400 K, a considerable change in the temperature gradients was found. Above 1850 K, the temperature gradients were reduced to values around 1 K.
The experiments described in this article show how state-of-the-art thermal analysis and thermophysical properties testing instruments can be used to analyze silicon nitride during sintering. The resulting values for the density change, specific heat, thermal diffusivity and thermal conductivity allow detailed insights into the sintering process and related phase transitions. The results can also be used to determine temperature gradients occurring in a sintering part during the production process.
By using these instruments in production operations, manufacturers of Si3N4 ceramics can truly optimize the sintering process to achieve higher quality final products.
For more information about thermal analysis and thermophysical properties testing, contact
NETZSCH-Gerätebau GmbH, Wittelsbacherstr. 42 95100 Selb/Bavaria, Germany; (49) 9287-881-49; fax (49) 9287-881-44; e-mail email@example.com ; or visit www.ngb.netzsch.com. (U.S. contact: NETZSCH Instruments, Inc., 37 North Ave., Burlington, MA 01803; (781) 272-5353; fax (781) 272-5225; e-mail firstname.lastname@example.org ; www.e-Thermal.com.)