Addition of hematite as a sintering aid in alumina: effect of concentration on physical, microstructural and mechanical properties

Hematite (Fe 2 O 3 ), or ferric oxide, is a ceramic oxide that, despite being recognized in the field of science and materials engineering, finds limited use as a sintering additive. Therefore, the purpose of this study was to explore the incorporation of hematite as a sintering additive in alumina (Al 2 O 3 ), employing conventional sintering at 1400 °C. Seven compositions were processed, with Fe 2 O 3 content varying from 0 to 8 wt.%. The samples underwent conventional ceramic processing steps (homogenization, drying, deagglomeration, sieving, and cold uniaxial pressing), followed by sintering at 1400 °C. The physical and mechanical properties of the produced samples were assessed through dilatometry, density measurement using the Archimedes’ method, scanning electron microscopy with energy-dispersive spectroscopy (SEM/ EDS), and flexural and compression tests. The results revealed that the addition of 4 and 6 wt.% of Fe 2 O 3 resulted in reduced shrinkage of the ceramics, leading to low densification, highly porous surfaces, and diminished flexural strength. On the other hand, lower additions (0.5, 1, and 2 wt.%) of Fe 2 O 3 improved the sintering of Al 2 O 3 , yielding samples with increased flexural and compressive strength, linear shrinkage, and densification.


Introduction
Ceramics based on alumina (Al 2 O 3 ) have been the subject of numerous studies in structural applications, owing to their remarkable properties such as dimensional stability, high-temperature resistance, and high hardness [1].Al 2 O 3 has characteristics similar to other polycrystalline ceramics, such as moderate tensile and bending resistance, and brittle fracture behavior, which is the main disadvantage of the mechanical properties of alumina.Al 2 O 3 is an ionic-covalent solid that does not yield under load like metals and alloys.The strong chemical bonds in Al 2 O 3 are reflected in its characteristics, such as low electrical and thermal conductivity, the high melting point that makes it practically impossible to shape alumina by casting, and the high hardness that characterizes this material and makes its machining complex and costly [2].The brittleness of Al 2 O 3 is another property that concerns engineers when designing components using this material.In metals, crack energy is dissipated by deformation at the crack tip, while alumina components can fail without any previous plastic deformation at the location of high tensile stresses, such as surface defects, notches, internal flaws, or in the occurrence of thermal shocks [3].With the advancement of materials science and engineering, the use of new techniques and/or materials is employed in ceramics so that their properties can be improved, and their disadvantages reduced, making these materials applicable, either as structural material or as functional material [4,5].
The addition of sintering additives with a low melting point creates a layer of a liquid phase, making it possible to sinter and bond ceramic particles at lower temperatures [6].The liquid phase formed by the additives "wets" the ceramic particles, and flows between them, causing the rearrangement of the particles, increasing the densification of the material.The surface tension of the liquid phase assists in the densification and removal of ceramic pores [7,8].From Muan's initial studies, other studies were published in order to extend the limits of knowledge on this subject: Dayal et al. [29] performed a crystal structure analysis, where they showed that FeAlO 3 has an orthorhombic structure with a Pc21n space group.Feenstra et al. [30] disclosed the synthesis of the orthorhombic (Fe 0.53 Al0.47) 2 O 3 phase coexisting with the corundum (Al 0.88 Fe 0.12 ) 2 O 3 at 1300 °C, a temperature lower than that required for the stabilization of FeAlO 3 (1318 °C).Dreval et al. [31] presented ternary phase diagrams calculated in the Al 2 O 3 -Fe 2 O 3 -FeO system based on experimental data for relevant phases.The results obtained by the authors are similar to those of Muan [28].Majzlan et al. [32] determined the enthalpy of FeAlO 3 formation from hematite and corundum and estimated the entropy of formation, indicating that FeAlO 3 is unstable under high-temperature P-T metamorphic conditions.The Al 2 O 3 -Fe 2 O 3 system has interesting characteristics, however, concerning the properties of alumina when utilizing hematite as a sintering aid, only the study by Wang et al. [33] addresses aspects such as ceramic densification with varying Fe 2 O 3 compositions, in addition to flexural strength.
This work aims to expand the knowledge about the properties of the Al 2 O 3 -Fe 2 O 3 compound by analysing its behaviour during sintering, morphology, and mechanical properties, especially through flexural and compressive testing.

Materials
The materials used in the fabrication of ceramic bodies are: α-Al 2 O 3 (ρ = 3.96 g/cm 3 ; Alcoa P-913, Brazil), with a particle size of 1μm, glycine (Sigma Aldrich), and iron nitrate (Neon Química, Brazil), employed for the preparation of hematite (Fe 2 O 3 ) powders (ρ = 5.26 g/cm 3 ).Polyethylene glycol (PEG) (Isofar, Brazil) was added to provide mechanical strength to the green bodies.The theoretical density of the samples was determined using the Rule of Mixtures, excluding PEG which is eliminated during sintering.Table 1 presents the theoretical density (TD) values and nomenclature of each sample.
The Al 2 O 3 -Fe 2 O 3 compound is a mixed oxide system, commonly used as a sorbent and catalyst.This compound is the subject of several investigations regarding its physicochemical properties [20][21][22][23].Alumina and hematite have applications in catalysis and adsorption [22,24].The main characteristics of alumina powders are surface acidity, mechanical strength, high surface area, and stability against sintering, insulating and diamagnetic properties [25].On the other hand, hematite powders present as main characteristics the redox behavior, medium to low surface area, easy sintering, semiconductor, and antiferromagnetic properties derived from iron [26].In theory, the chemical mixing of Fe 2 O 3 with Al 2 O 3 could be beneficial, enabling the production of oxidation catalysts with improved stability or mechanically and thermally stable powders with redox functionality.Furthermore, the treatment of two oxides at high temperatures produces new crystal structures with interesting properties, especially in heterogeneous catalysis, different from the properties of the initial oxides [27].
Considering the Fe 2 O 3 -Al 2 O 3 system, some studies have been crucial for understanding phase formation from this compound.Muan and Gee [27] and Muan [28] verified phase formation from this compound between 1085 and 1725 °C with oxygen pressure (PO 2 ) variation, where the formation of FeAlO 3 was verified at pressures higher than 0.03 atm.The FeAlO 3 phase was also found to be unstable, to decompose into hematite and corundum at temperatures below 1318 °C, where the decomposition temperature remains unchanged with respect to PO 2 .
The stability of FeAlO 3 increases with increasing pressure, where the stabilization temperature of FeAlO 3 ranges from 1318 to 1410 °C in air, changing to 1318-1495 °C at PO 2 = 1 atm.

Fe 2 O 3 powder synthesis
The Fe 2 O 3 powders were obtained through the spontaneous combustion synthesis method.In this process, ferric nitrate [Fe(NO 3 ) 3 •9H 2 O] was used as the oxidizer, and glycine [NH 2 CH 2 COOH] as the fuel.This method was similar to the one performed in the study by Cao et al. [34].To prepare the powders, Fe(NO 3 ) 3 •9H 2 O and NH 2 CH 2 COOH were dissolved in 150 mL of distilled water and the solution was stirred to achieve a homogeneous mixture.Subsequently, the mixture was poured into a 500 mL beaker and heated to 100 °C.
During the heating process, the water in the solution evaporated, transforming the mixture into a gelatinous mass.After a few minutes of heating, the mixture expanded, releasing a significant amount of gases.This was followed by a non-explosive exothermic reaction, resulting in spontaneous combustion.After the burn, a foam composed of iron oxide was obtained, which was crushed to produce the Fe 2 O 3 powders used as sintering additives.

Ceramic processing
To obtain the desired materials, Al 2 O 3 , Fe 2 O 3 , and PEG were homogenized in a ball mill for 8h, using distilled water in a 1:1 ratio to facilitate the mixture.After this process, the powders were dried for 48h at a temperature of 120 °C, manually deagglomerated, and sieved through a 60-mesh screen.Using these materials, discs with a diameter of 15 mm and a mass of 0.5 g were produced through cold uniaxial pressing with a load of 30 MPa.Sintering was conducted in a JUNG furnace, without the application of pressure, with a maximum temperature of 1400 °C, followed by inertial cooling.The heating ramp utilized is illustrated in Figure 1.

Dilatometric analysis
Dilatometric analyses were conducted using the Netzsch DIL402 PC dilatometer in the Laboratory of Advanced Materials (LAMAV) at the Universidade Estadual do Norte Fluminense (UENF).The thermal cycle for this analysis commenced at room temperature, where the samples were heated at a rate of 5 °C/min up to 1400 °C, holding at this temperature for 1 hour, followed by cooling at a rate of 10 °C/min.The samples used in the dilatometry test have a diameter of 12mm and a length of 6mm.The samples were heated in an Al 2 O 3 crucible, which can withstand temperatures up to 1600 °C.To rectify the obtained results and eliminate any thermal expansion curves that might emerge due to dilatometer variations, the thermal expansion constant (α) value is derived using Equation 1.This value serves as a basis for determining the technological shrinkage (ε tech ), as presented in Equation 2.
Technological shrinkage employs the thermal expansion constant value to exclude expansion readings during the analysis, thereby representing only the accurate material shrinkage value [35].
Where: ε amb represents the material shrinkage after cooling; εmax signifies the material shrinkage at the maximum temperature; T amb stands for the final temperature after cooling, and T max denotes the maximum temperature of the test.
Where: ε(t,T) represents the shrinkage at a specific point; t stands for the time at which the formula was applied, and T denotes the temperature at that particular point.

Green density calculation
Prior to sintering, the density of the material in its green form was determined by measuring the mass and calculating the volume of the green bodies using Equation 3. Furthermore, green state densification (Equation 4) was calculated as the percentage difference between the measured density and the theoretical density obtained through the Rule of Mixtures.Where:  g -density of the green body (g/cm 3 ); m -mass of the body (g); v -volume of the body (cm 3 );  -theoretical density (g/cm 3 ); d g -densification of the green body (%).

Density and linear shrinkage of sintered samples
To calculate the densification of the sintered bodies, the method described in NBR 16667:2017 [36], based on the Archimedes technique, was employed.Using the immersed mass (m i ), wet mass (m w ), and dry mass (m d ) data, it was possible to determine the apparent density through Equation 5.With this result, the densification of the ceramic body was then calculated by the difference between the apparent density and the relative density of the sintered body (Equation 6).
• water md ap mw mi ( ) To calculate the linear shrinkage (L S ) of the sintered bodies, the difference between the diameter of the green body (D GB ) and the sintered body (D SB ) was employed, as shown in Equation 7.

Scanning Electron Microscopy (SEM)
The sintered samples were morphologically analyzed using the QUANTA FEG 250 microscope, manufactured by FEI.The samples underwent ceramic preparation, being embedded and then ground successively with #220, #400, #600, #1200, and #2000 grit sandpapers, followed by polishing with 6µm diamond paste.Subsequent to preparation, the samples were coated with gold utilizing a Leica ACE600 high-vacuum coating chamber.SEM analysis was carried out under the following parameters: electron beam power of 20kV, working distance ranging between 10.5 to 13 mm, spot size of 5, and image magnification at 20,000x, utilizing the secondary electron detector.For EDX analysis and compositional mapping, a detector from the manufacturer Bruker was employed, coupled to the microscope column.

Flexural tests
Flexural strength testing of the studied materials was performed on ceramic discs with a diameter of 14 ± 2 mm and a thickness of 1.2 ± 0.2 mm, following the ABNT NBR ISO 6872:2016 Standard [37].Six samples were tested for each group, positioned atop three metallic spheres, and subjected to a rate of 0.5 mm/min until fracture occurred.The flexural strength was automatically calculated by the software based on the data collected by the equipment.The test was conducted at the Mechanical Testing Laboratory of IME, using a DL-1000 EMIC machine.

Compressive tests
The compressive test of the samples was carried out on a DL-1000 EMIC machine at a testing speed of 0.5 mm/min.For each group, three samples were used, prepared in a cylindrical shape following the method described by Awaji, Watanabe, and Nagano for compressive testing [38].The samples used have a diameter of 12.5 mm and a thickness of 5 mm.

Dilatometric analysis
In Figure 2 and Table 2, the results of dilatometry for the studied samples are shown.The inclusion of Fe 2 O 3 as a sintering aid in Al 2 O 3 resulted in significant changes in behavior during the sintering process.The addition of 0.5 and 1 wt.% of Fe 2 O 3 , represented by samples ALFE05 and ALFE10, respectively, did not influence the onset retraction temperature (indicated as T 1 in the table).The final retraction temperatures (T 2 ) showed no variation in the compound samples, while the maximum retraction temperature (T 3 ) of samples ALFE05 and ALFE10 exhibited a reduction.Compared to the AL group, sample ALFE05 showed a lower retraction, with a value of ϵ = 9.04%, while sample ALFE10 presented a value of ϵ = 11.59%.The addition of 2 wt.%Fe 2 O 3 in the ALFE20 sample resulted in a reduction of T1, with shrinkage starting at 1010 °C, and an increase in T 3 , reaching 1286 °C.The samples containing 4 and 6 wt.% of Fe 2 O 3 (ALFE40 and ALFE60) exhibited a delayed onset of shrinkage compared to other compound samples, with T 1 values of 1060 and 1047 °C, respectively, and an increase in T 2 .The higher T 1 values indicate that higher temperatures, or in other words, higher activation energy values, are necessary for the shrinkage to begin [39].Depending on the sintering temperature and the time employed, good densification of the material may not occur.If the sintering temperature is low, remaining below the maximum shrinkage point (T 3 ), and the sintering time is short, proper grain rearrangement and growth may not occur, thus compromising the final densification of the ceramic [8].The shrinkage values obtained for these samples were the lowest among all groups, with ϵ equal to 5.90% and 6.89%, respectively.The ALFE80 sample present a T 1 temperature similar to the ALFE60 sample.Its T 2 was recorded at 1386 °C, close to the maximum test temperature, while it was not possible to record T 2 , as it is above 1400 °C.Its shrinkage was similar to that of group ALFE05, with a value of 9.05%.
When observing the temperatures T 1 , T 2 , and T 3 , it is noticeable that there is a tendency for the ε value of the samples to increase as the temperatures T 1 and T 3 decrease, due to the greater activation energy required to initiate the shrinkage during sintering.This is evidenced by the value of alpha.This trend is observed in the ALFE10 and ALFE20 samples, which showed the highest ε values, with ALFE10 exhibiting a lower T 3 value and the ALFE20 sample presenting a lower temperature at T 1 .However, the ALFE80 sample exhibited an abnormal behavior, with an ε value similar to that of the ALFE05 sample, but with much higher T 1 and T 3 values.Low ε values can lead to materials with low densification and high porosity.
Residual porosity in ceramics leads to a decrease in the mechanical strength of the final product.As a result, porous ceramics generally exhibit lower mechanical resistance when compared to dense ceramics.Consequently, their applications are distinct and cannot be used as structural materials [40].
From the calculation of the thermal expansion constant (α), as shown in Equation 1, it was observed that the value of α is inversely proportional to the material shrinkage.
It can be observed that as the shrinkage value increases, α decreases.This observation is supported by the work of Zygmuntowicz et al. [41], where the authors studied the sintering behavior of the Al 2 O 3 -Ni compound and, through dilatometry analysis, found that the α of samples with a higher Ni content increased due to lower shrinkage values of the compound compared to pure Al 2 O 3 .Some examples in the literature demonstrate the influence of sintering additives on the shrinkage curves of alumina.García et al. [42] processed Al 2 O 3 -Nb compounds in fractions of 10, 30, and 50 vol.%.Through dilatometry tests, the authors observed that the onset of shrinkage began at 850 °C due to the addition of Nb.However, to achieve a 12% shrinkage value, it was necessary to add 50 vol.% of Nb.Jesus et al. [43] reports the production of functional gradient ceramics (FGM) using the spark plasma sintering (SPS) method, employing a compound of Al 2 O 3 -Nb 2 O 5 -LiF -ZrO 2 .In order to investigate which compositions would present similar shrinkage curves for the formation of FGM layers, dilatometry technique was used.The addition of ZrO 2 to Al 2 O 3 resulted in an ε value of 15%, while compositions with different levels of LiF added to the Al 2 O 3 -Nb 2 O 5 compound maintained lower ε values, close to 13%.The significant difference in ε values resulted in a large amount of residual stress, leading to the formation of cracks during the processing of FGMs by SPS, as well as reducing the mechanical strength and densification of the material.

Density and linear shrinkage of samples
Figure 3 shows the density results of the green bodies, density of sintered samples, and linear shrinkage are illustrated.
The results presented in Figure 3(a) indicate that the addition of Fe 2 O 3 influenced the densification of the green bodies during compaction.A slight increase in the density of the ALFE05, ALFE10, and ALFE20 bodies was observed compared to pure alumina after the addition of 0.5, 1, and 2 wt.% of Fe 2 O 3 .The AL group exhibited a relative green density of 57.79 ± 2.40% of the theoretical density, while the samples from the ALFE05, ALFE10, and ALFE20 groups exhibited relative green densities of 60.89 ± 1.92%, 61.14 ± 2.53%, and 59.00 ± 1.37%, respectively.On the other hand, the addition of higher percentages of Fe 2 O 3 resulted in a reduction in the average density of the green bodies.The samples with 4 and 6 wt.% Fe 2 O 3 additions (ALFE40 and ALFE60) exhibited relative densities of 52.78 ± 2.20% and 55.87 ± 3.87%, respectively.The sample with 8 wt.% of Fe 2 O 3 (ALFE80) presented an average density value of 53.29 ± 3.64%.The high Fe 2 O 3 content in the samples may have hindered homogenization, causing the reduction in density of the ALFE40, ALFE60, and ALFE80 groups.
Regarding densification after sintering, as illustrated in Figure 3(b), the AL group exhibited a relative densification of 82.99 ± 0.82%.Adding 0.5 wt.% Fe 2 O 3 resulted in a slight reduction in the densification of the ALFE05 group, which reached 79.06 ± 1.89%.On the other hand, the addition of 1 and 2 wt.% of Fe 2 O 3 led to a slight increase in the relative density of the ALFE10 and ALFE20 groups, which had values of 83.64 ± 0.75% and 85.69 ± 1.16%, respectively.The addition of Fe 2 O 3 at concentrations of 4 and 6 wt.% resulted in a significant reduction in the densification of the ALFE40 and ALFE60 groups, which exhibited values of 66.96 ± 1.83% and 62.69 ± 0.99% of TD, respectively.The ALFE80 group exhibited a relative densification of 88.24 ± 2.91%, the highest value among all compositions.
The results of linear shrinkage of the sintered samples shown in Figure 3(b) corroborate the results of the dilatometric analysis.The addition of Fe 2 O 3 up to 2 wt.% causes an increase in shrinkage, with a consequent increase in densification, except for the ALFE05 sample, which showed a reduction in the ε value in dilatometry compared to AL.In the calculation of linear shrinkage, the AL group exhibited a shrinkage of 11.24 ± 0.23%, while ALFE05 exhibited a shrinkage of 13.56 ± 1.31%.The ALFE10 and ALFE20 samples exhibited shrinkage of 16.28 ± 0.39% and 16.00 ± 0.37%, respectively.The ALFE40 and ALFE60 groups, which exhibited low densification values, showed linear shrinkage of 5.28 ± 0.77% and 3.89 ± 0.50%, respectively.The ALFE80 group exhibited a linear shrinkage value of 12.89 ± 0.27%.The sample presented an unexpected result, given that the shrinkage value from the dilatometry test of the sample was similar to that of the ALFE05 sample, which exhibited lower densification after sintering.The difference in the shrinkage values obtained with dilatometry and calculated after sintering is due to the difference in the heating rates of the test and the sintering ramp, as well as the dwell time at the sintering temperature.The longer dwell time in sintering allowed for grain growth in the sample, increasing densification.

Scanning Electron Microscopy (SEM)
In Figure 4, the surface micrographs of the samples captured by SEM with a magnification of 5000x are illustrated.The presence of Fe 2 O 3 in Al 2 O 3 influenced densification, as shown in the dilatometry and density results, but the morphology changed with an increasing amount of sintering additive.Pure alumina (group AL), shown in Figure 4(a), exhibits a microstructure characterized by a series of small interconnected pores, indicated by yellow arrows in the figure, as well as isolated and loosely cohesive grains.These characteristics are attributed to low densification, confirmed by the obtained densification results and Rahaman's study (2017), reporting that the initial stage of sintering occurs at up to 65% of theoretical density.Although sample AL has a higher density than mentioned by the author, it is still in the early stage of sintering, even with an average densification above 65%.One contributing factor to pure alumina not achieving a microstructure with higher densification was the sintering temperature.Pure alumina, to obtain good densification values, should be sintered at 1600 °C, approximately 80% of its melting temperature [44].The ALFE05 sample, Figure 4(b), shows a microstructure with fewer pores and larger spacings between them.Moreover, the presence of the Fe 2 O 3 additive among the grains results in a greater number of grains interconnected by necks.These necks formed due to the additive's presence, promoting lower porosity in the material.
The ALFE10 sample, Figure 4(c), demonstrates that adding 1 wt.% of the additive resulted in a slight increase in grain size compared to the ALFE05 sample.This grain growth allows for a clearer visualization of neck formation during the sintering process.Like the ALFE05 sample, the ALFE10 sample exhibits characteristics of an intermediate stage of sintering, with grain growth and a microstructure full of interconnected pores.As the necks form, these pores tend to isolate [45].
On the other hand, increasing the Fe 2 O 3 concentration to 2 wt.%, the ALFE20 sample shows a similar effect to the ALFE05 and ALFE10 samples, with the difference that, due to the higher average densification of these samples, there is a lower quantity and size of pores in the material.Increasing the Fe 2 O 3 content to 4 and 6 wt.% resulted in an excessively porous microstructure with low density.The ALFE40 and ALFE60 samples, Figures 4(e) and 4(f), respectively, exhibit a highly porous structure.The images reveal that the sintering process was not efficient in these samples, with almost complete absence of neck formation, only a mild adhesion between grains due to diffusion during heating.This results in materials with extremely low density, implying inferior mechanical properties [46].
The ALFE80 sample, Figure 4(g), stands out from the other groups due to its remarkably dense microstructure and considerable grain growth.Additionally, it exhibits completely isolated pores, tending to assume a spherical shape.This dense microstructure with high grain growth is the result of the elevated densification values observed for the addition of 8 wt.% of Fe 2 O 3 , where the sample reached the final stage of sintering.Figure 5 illustrates the surface images of the sintered samples and the compositional map obtained through EDS analysis.Meanwhile, Table 3 presents the weight percentages of each element present in the samples.
When analyzing the elemental mapping images, it is observed that the distribution of elements is related to the composition of oxides, and the variation in the quantity of each element occurs according to the increase in the additive content in alumina.In the AL sample, the presence of iron is not detected in its structure, as it consists solely of Al 2 O 3 .In the ALFE05 sample, small iron points are identified scattered throughout the sample, as can be seen in the image of overlapped elements, concentrated in regions near grain boundaries and pores.An iron content of 0.26% was found, which may be attributed to a region with a lower concentration of Fe 2 O 3 in the material.Addition of hematite as a sintering aid in alumina: effect of concentration on physical, microstructural and mechanical properties 9/14 Tecnol Metal Mater Min.2024;21:e2978 When analyzing samples with higher additions of Fe 2 O 3 , a gradual increase in the Fe content is noticeable.As the analysis progresses from the ALFE10 sample to the ALFE80 sample, concentration points of Fe are observed along the samples, indicating agglomeration of Fe 2 O 3 in Al 2 O 3 .Despite the presence of these agglomeration zones in the ALFE05, ALFE10, ALFE20, and ALFE80 samples, the distribution of Fe in the material is quite uniform in these samples, highlighting good homogenization during the manufacturing process.However, the ALFE40 and ALFE60 samples exhibited highly agglomerated regions, especially in the porous areas of the material.The difference in density between Al 2 O 3 and Fe 2 O 3 (3.96and 5.26 g/cm 3 , respectively) may be a likely explanation for this phenomenon.Due to its higher density, Fe 2 O 3 deposited in the pores during homogenization, resulting in heterogeneous distribution, with consequent poor densification in these two samples.
Observing Table 3, it is verified that the ALFE60 sample presented a significantly lower value than 6% of Fe, exhibiting only 4.21% due to poor material homogenization, resulting in a low concentration of Fe in the analyzed region.
In addition to the compositional mapping performed on the samples, a point analysis of EDS was conducted on the sample with the highest additive content (ALFE80) to investigate the regions where Fe 2 O 3 tends to deposit.Three regions were considered: the grain center, grain boundary, and grain interior.Figure 6 presents images indicating the analysis points, while the values obtained at each point are described in Table 4.
The point EDX analysis of the ALFE80 sample indicated that the sintering additive distributed itself in all grain regions, as the presence of Fe was detected at all analyzed points.However, a tendency for Fe to deposit in the grain interior was identified, where an Fe content of 8.21% was observed.The analysis conducted at the grain boundary revealed an Fe value of 6.64%, while at the grain center, a value of 5.88% was recorded.This indicates that, in contrast to other sintering additives, such as Nb 2 O 5 , which typically accumulates preferentially at grain boundaries, as reported by Gomes et al. [47], Fe 2 O 3 tends to deposit in the grain interior.

Mechanical properties
The Figure 7 presents the results of the bending and compression tests of analyzed samples.
From the results presented in Figure 7(a), it is possible to observe that the flexural strength of alumina was significantly improved with the addition of Fe 2 O 3 at low concentrations.The AL group exhibited a flexural strength value of 160.07 ± 34.78 MPa.With the addition of Fe 2 O 3 in fractions of 0.5, 1, and 2 wt.%, there was a considerable increase in the flexural strength of ceramics, with the ALFE05, ALFE10, and ALFE20 samples exhibiting flexural strength values of 299.84 ± 73.64 MPa, 270.48 ± 37.23 MPa, and 357.67 ± 43.30 MPa, respectively.This is directly related to the higher densification values obtained for the samples with Fe 2 O 3 additions up to 2 wt.%.Furthermore, as observed through SEM images, the grains are more coalescent, favoring an increase in mechanical properties.However, the low densification of the ALFE40 and ALFE60 samples contributed to the low flexural strength of these groups, even worse than the group without the presence of Fe 2 O 3 .The ALFE40 and ALFE60 samples exhibited flexural strength of 63.07 ± 4.28 MPa and 62.76 ± 18 MPa, respectively.This may be associated with larger grain sizes, as seen in electron microscopy, coupled with high porosity after sintering.When comparing the flexural strength results obtained in this work with the results from Wang et al. [33], where Al 2 O 3 -Fe 2 O 3 ceramics were produced with molar fractions of 1, 3, and 5, through uniaxial pressing at 100 MPa and conventional sintering at 1550 °C, it is evident that the values obtained by the authors were lower than the values found in this work, ranging from 48 to 68 MPa.Another example is the work of Ren et al. [48], where Al 2 O 3 -CaO glass-ceramics were produced with additions of Fe 2 O 3 ranging from 0 to 10 wt.%, resulting in a significant increase in flexural strength, ranging from 25 MPa in the sample without Fe 2 O 3 to approximately 220 MPa in the sample with 10 wt.% of the oxide.This demonstrates that through the optimization of parameters, such as applied pressure and the appropriate amount of sintering additive, the mechanical properties of alumina can be considerably improved, enabling the use of these materials in structural applications, such as refractories and abrasion-resistant ceramics, or in military applications, such as armor plates and projectile tips [49].
The results of the compressive test, displayed in Figure 5  In this study, ceramics of the Al 2 O 3 -Fe 2 O 3 compound were produced with the aim of enhancing the understanding of their physical, mechanical, and microstructural properties.To achieve this goal, various analyses were conducted, such as dilatometry, density measurement, scanning electron microscopy (SEM), energy-dispersive spectroscopy (EDX), and flexural and compression tests.The obtained results led to the following conclusions: After dilatometric analysis, it was observed that the addition of Fe 2 O 3 in fractions of 1 and 2 wt.% results in an increase in linear shrinkage, while additions of 4 and 6 wt.% lead to very low shrinkage, and a slight contraction with 8 wt.%.These findings are reinforced by density analysis using the Archimedes method, which highlights the low linear shrinkage of samples with 4 and 6% of Fe 2 O 3 , indicating a deficiency in their densification.
SEM analyses revealed the microstructure of the sintered samples, allowing for the observation of a porous aspect in all of them.This porous aspect was reduced in the ALFE05 and ALFE10 samples but accentuated in the ALFE40 and ALFE60 samples due to their low densification, resulting in high porosity.In the ALFE80 sample, the material showed low porosity with pronounced grain growth.
EDX analyses allowed for the quantification of the elements present and their distribution in the sample.An increase in the Fe content was observed as a result of the addition of Fe 2 O 3 in the samples.Furthermore, it was found that Fe was mainly located in the pores and grain boundaries of the Al 2 O 3 .
The results of flexural tests demonstrated a significant increase in the flexural strength of Al 2 O 3 with the addition of Fe 2 O 3 up to 2%.However, additions of 4% and 6% resulted in a drastic reduction in the flexural strength of 12/14 Tecnol Metal Mater Min.2024;21:e2978 the ceramic, even worse than pure Al 2 O 3 .The addition of 8 wt.% of Fe 2 O 3 promoted a slight increase in strength, but the marginal improvement did not justify the use of sintering additive in this quantity.
The results of compression tests demonstrated an increase in strength similar to the flexural test results; however, the ALFE10 group exhibited the highest compression strength among all groups.Similar to the density and flexural strength results, samples from the ALFE40 and ALFE60 groups presented the worst values among all groups.The addition of 8 wt.% of Fe 2 O 3 increased compression strength, ranking just below the ALFE10 group.

Figure 1 .
Figure 1.Sintering ramp adopted in this study.

Figure 3 .
Figure 3. (a) Green body density results; (b) Relative density results of sintered samples and linear shrinkage.

Figure 6 .
Figure 6.EDX analysis of the ALFE80 sample in different regions to determine the preferential deposition of Fe 2 O 3 : (a) Grain center; (b) Grain boundary; (c) Grain interior.
(b), show that the addition of Fe 2 O 3 to Al 2 O 3 followed the same trend of increase and decrease in mechanical strength values, similar to the results of densification and flexural strength.The samples from group AL had an average value of 369.27 ± 4.08 MPa.The elastic modulus was 2.01 ± 0.23 GPa.The increase in compressive strength caused by Fe 2 O 3 resulted in an increase in the elastic modulus.The ALFE05, ALFE10, and ALFE20 samples had average compression strength values of 430.48 ± 113.13, 621.86 ± 60.29, and 467.43 ± 91.53 MPa, representing an increase of 16.58, 68.40, and 26.58%, respectively.The elastic modulus for ALFE05, ALFE10, and ALFE20 groups was E = 1.86 ± 0.26 GPa, 2.79 ± 0.05 GPa, and 2.19 ± 0.29 GPa, respectively.For this compound, the addition of 1 wt.%Fe 2 O 3 yielded the best compressive strength result.As a result, it was almost impossible to perform the test on this group of samples since the applied load approached the testing machine's capacity (100 kN), requiring careful attention to avoid damaging the machine.

Table 1 .
Composition, nomenclature and theoretical density of sample groups of this paper

Table 2 .
Shrinkage temperatures, thermal expansion coefficient, and shrinkage values obtained in the dilatometric analysis

Table 3 .
Mass concentrations of the elements present in the sintered samples

Table 4 .
Variation of iron content in different regions of the grain present in the EDX analysis of ALFE80 sample