Binder jetting of reaction bonded silicon carbide: Effects of printing and siliconizing modifications
Isakhani Zakaria, Michael (2024)
Isakhani Zakaria, Michael
2024
Master's Programme in Materials Science and Engineering
Tekniikan ja luonnontieteiden tiedekunta - Faculty of Engineering and Natural Sciences
This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Hyväksymispäivämäärä
2024-05-15
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:tuni-202404254571
https://urn.fi/URN:NBN:fi:tuni-202404254571
Tiivistelmä
The main objective of this thesis is to study some aspects of processing and post processing in additively manufactured Reaction Bonded Silicon Carbide (RBSiC) to promote homogeneity of microstructure and to enhance mechanical properties. This is particularly important for achieving isotropic properties for engineering applications that require it. Efforts were made to minimize printing and layering imperfections, to reduce porosity, and to reach a more uniform distribution of phases, assessed in the final siliconized samples. Effective variables to study were selected from the printing and post processing stages of sample fabrication.
Green parts were fabricated with Binder Jetting (BJT) using SiC powder feedstock and phenolic resin binder as starting raw materials. This was followed by densification using Liquid Silicon Infiltration (LSI) technique, wherein Silicon melt infiltrates the structure via capillary force, reacting to form secondary SiC upon contact with available carbon in the debinded parts. Alternatively, to decrease the amount of residual Si and increase the amount of secondary SiC, extra carbon was also added to some samples by impregnation of green parts with phenolic resin and subsequent pyrolysis (I&P) prior to LSI.
For BJT, parameters were selected to be layer thickness (LT) of 35 and 70 µm and printing strategy (normal and shell), both of which presumably affect the pore architecture of the green parts. Normal printing bonds powder particles within the whole volume, while shell printing leaves loose powder entrapped in the core surrounded by a printed wall with specified thicknesses. Results showed a significant dependence of green density on printing variables. Low-er LT increased green density regardless of printing strategy, with shell-printed samples exhibiting lower density compared to normally printed ones. Samples with higher apparent porosity, i.e. shell printed and/or higher LT, benefitted from a higher amount of added carbon after I&P.
For LSI, two initiation sides (top and bottom surface of the samples, corresponding to the last and the first printed layers respectively) were selected on samples in each printing and carbonization condition. This was aimed to examine the possible effect of LSI relative direction to the build-up direction on microstructure and phase fractions. Abnormal infiltration was observed in debinded shell printed samples, containing loose powder within their core, while other samples had successful densification through LSI. Porosity measurements using image analysis of the microstructures demonstrated an overall lower final porosity when executing the LSI from the bottom surface of the samples (LSI/B). Also, shell printing and/or higher LT resulted in lower final porosities in LSI/B. Additionally, based on the amount of added carbon in each sample, various amounts of secondary SiC were detected and calculated. Carbonized normally printed samples had the highest overall SiC content, contributing to a maximum density (2.82 g/cm3) at LT of 35 µm. Additionally, siliconizing carbonized shell printed samples resulted in a more uniform microstructure of the core compared to normally printed ones, i.e. diminished layered microstructure and printing marks from binder deposition. It was demonstrated that a lower LT is beneficial for a homogenous microstructure, both in normal and shell printing.
Results from compression tests showed that a higher LT has the potential to enhance the average strength values, possibly due to a more robust phase composition. This average trend was also observed for the measured compressive stress until the occurrence of the first externally visible crack in each sample.
Green parts were fabricated with Binder Jetting (BJT) using SiC powder feedstock and phenolic resin binder as starting raw materials. This was followed by densification using Liquid Silicon Infiltration (LSI) technique, wherein Silicon melt infiltrates the structure via capillary force, reacting to form secondary SiC upon contact with available carbon in the debinded parts. Alternatively, to decrease the amount of residual Si and increase the amount of secondary SiC, extra carbon was also added to some samples by impregnation of green parts with phenolic resin and subsequent pyrolysis (I&P) prior to LSI.
For BJT, parameters were selected to be layer thickness (LT) of 35 and 70 µm and printing strategy (normal and shell), both of which presumably affect the pore architecture of the green parts. Normal printing bonds powder particles within the whole volume, while shell printing leaves loose powder entrapped in the core surrounded by a printed wall with specified thicknesses. Results showed a significant dependence of green density on printing variables. Low-er LT increased green density regardless of printing strategy, with shell-printed samples exhibiting lower density compared to normally printed ones. Samples with higher apparent porosity, i.e. shell printed and/or higher LT, benefitted from a higher amount of added carbon after I&P.
For LSI, two initiation sides (top and bottom surface of the samples, corresponding to the last and the first printed layers respectively) were selected on samples in each printing and carbonization condition. This was aimed to examine the possible effect of LSI relative direction to the build-up direction on microstructure and phase fractions. Abnormal infiltration was observed in debinded shell printed samples, containing loose powder within their core, while other samples had successful densification through LSI. Porosity measurements using image analysis of the microstructures demonstrated an overall lower final porosity when executing the LSI from the bottom surface of the samples (LSI/B). Also, shell printing and/or higher LT resulted in lower final porosities in LSI/B. Additionally, based on the amount of added carbon in each sample, various amounts of secondary SiC were detected and calculated. Carbonized normally printed samples had the highest overall SiC content, contributing to a maximum density (2.82 g/cm3) at LT of 35 µm. Additionally, siliconizing carbonized shell printed samples resulted in a more uniform microstructure of the core compared to normally printed ones, i.e. diminished layered microstructure and printing marks from binder deposition. It was demonstrated that a lower LT is beneficial for a homogenous microstructure, both in normal and shell printing.
Results from compression tests showed that a higher LT has the potential to enhance the average strength values, possibly due to a more robust phase composition. This average trend was also observed for the measured compressive stress until the occurrence of the first externally visible crack in each sample.