Surface characterization was performed at 500x magnification for each of the significantly stronger CNF microstructures (Figure 20). The surface morphology is extensively flawed with macropores and cracking, and the microstructure appears to be open-celled.
(a) (b)
(c) (d)
Figure 20:Surface morphology at 500x maginification of a) isothermally sintered CNFs, b) the 750oC
and c) 875oC 2 hour TSS CNFs, and d) the 4 hour 875oC TSS CNFs.
Higher magnification microscopy reveals fiber morphology (Figure 21). The isothermally sintered CNFs inFigure 21aare relatively thin and disperse, and grains are not visibly resolved.
CONTAINS LLNL PROPRIETARY INFORMATION 5.2 Microstructural Characteristics
The TSS CNFs are more densely packed and grains can be seen to span the fiber diameter. Slight de-sintering is observed in the 875oC micrographs (Figure 21c and d).
(a) (b)
(c) (d)
Figure 21:Representative SEM images at 50,000x maginification of a) isothermally sintered CNFs, b) the 750oC and c) 875oC 2 hour TSS CNFs, and d) the 4 hour 875oC TSS CNFs.
The CNF diffraction pattern is the same shape as the reference with highly broadened peaks due to the nanocrystalline microstructure; however, the shift in diffraction angle suggests there are other Ga2O3polymorphs, significant lattice strain, or displacement errors present (Figure 22). The CNF (0 0 2) peak, located around∼29.5o
CONTAINS LLNL PROPRIETARY INFORMATION 5.2 Microstructural Characteristics
2o
. Higher 2θangles, such as the peak corresponding to the (5 1 2) plane typically located at
∼64o, are shifted even further. This may be attributed either to displacement error during
measurement or inherent porosity causing lattice spacing to expand by∼10% larger than the reference.
Figure 22:CNF diffraction pattern with significant peak displacement errors, masking peak matching withβ-Ga2O3reference pattern.
Crystallite size is determined by the (0 0 2) peak (Table VII). The isothermally sintered crystallite size is 32.2 nm, while most TSS profiles produced smaller CNF crystallites with the exception of the 750o
C 4 hour profile. Although the peaks are shifted, the qualitative analysis and comparison between peak widths reveals that extended sintering times implemented through TSS suppresses grain growth and produces nanocrystalline microstructures.
CONTAINS LLNL PROPRIETARY INFORMATION 5.2 Microstructural Characteristics
Table VII:β-Ga2O3CNF (0 0 2) Crystallite Size
Profile Temperature Time 2θ(o
) β(rads) τ(nm) TSS 625oC 1 hr 29.515 6.98E-3 20.5 2 hr 29.46 6.72E-3 21.3 4 hr 29.415 5.88E-3 24.4 750oC 1 hr 29.445 6.37E-3 22.5 2 hr 29.605 5.15E-3 27.9 4 hr 29.39 4.36E-3 32.9 875oC 1 hr 29.515 5.67E-3 25.3 2 hr 29.55 4.71E-3 30.4 4 hr 29.355 5.32E-3 26.9 Iso, Own 1,000oC 1 hr 29.31 4.45E-3 32.2
Using the images in Figure 21, porosity, the pore size distribution, and the fiber size distribution is determined. Porosity is directly correlated with CNF average yield stress (Figure 23). Of the significantly strengthening TSS profiles, the CNF tube with the lowest porosity or highest density was the 875o
C 2 hour profile at 49%. Compared to the isothermally sintered tubes, the TSS CNFs are much denser.
Figure 23:CNF characteristic strength, with 90% confidence, is inversely proportional to porosity.
The pore size distributions are exponentially distributed and highly skewed right (Figure 24). The pore sizes are much smaller than 0.3µm; this is advantageous because, by definition, HEPA-grade filters must be able to filter 99.97% of particulates larger than 0.3µm.
CONTAINS LLNL PROPRIETARY INFORMATION 5.2 Microstructural Characteristics
(a) (b)
(c) (d)
Figure 24:Pore size distribution for the a) isothermally sintered CNFs, b) the 750oC and c) 875oC 2
hour TSS CNFs, and d) the 4 hour 875oC TSS CNFs.
The fiber diameter distribution is best fit by a lognormal distribution (Figure 25). At low magnifications, there are many large fibers on the tube surface, but as magnification is increased, much smaller fibers appear within the microstructure.
CONTAINS LLNL PROPRIETARY INFORMATION 5.2 Microstructural Characteristics
The microstructural characteristics are summarized inTables VIII and IX. Compared to the isothermally sintered CNFs, the TSS tubes are more dense with nearly the same pore sizes (Pr). The fiber diameters (fd) are larger at longer sintering times, similar to the crystallite size.
Table VIII:β-Ga2O3CNF Porosity Characteristics
Profile Temperature Time Pr(nm) s (nm) Pr, max(nm) Porosity (%) TSS 750oC 2 hr 42.8 37.7 247 55.6 875oC 2 hr 42.6 34.1 251 49.1 4 hr 60.1 45.2 308 51.1 Iso, Own 1,000oC 1 hr 52.1 41.7 248 64.3
Table IX:β-Ga2O3CNF Fiber Diameters
Profile Temperature Time fd(nm) s (nm) TSS 750oC 2 hr 87.9 20.0 875o C 2 hr 82.4 26.1 4 hr 104.2 25.1 Iso, Own 1,000oC 1 hr 78.4 20.8
CONTAINS LLNL PROPRIETARY INFORMATION
6
DISCUSSION
β-Ga2O3CNF mechanical properties are influenced by geometrical constraints and mi- crostructural consequences of two-step sintering. The post-sintering specimens were ex- tremely distorted from their original tubular shape. The contact area between the compres- sion platens and the specimen was variable as most tubes were nearly tetrahedral shaped. This undesirable geometry does not distribute loads evenly and may lead to much lower strength values than expected. The tetrahedral geometry is also undesirable for LLNL’s pur- pose, as the cross-sections at the tube ends were elliptic rather than circular. An elliptic cross-section requires more power to force air through the tube, increasing the pressure drop and operating cost. In order to produce reliable results, tube geometry control requires refinement.
The severe volumetric shrinkage undergone during the high-temperature step determines the overall effectiveness of TSS. The high-temperature step determines the degree to which grain boundary mobility is limited and densification is maximized during the sintering dwell. Densification during the sintering dwell is the main driving force for enhancing porous ceramic strength [53]. Further investigation into the high-temperature step is required to fur- ther optimize grain growth suppression, densification, and strength. If the high-temperature step were optimized, there would ideally be no grain growth between the 2 hour and 4 hour sinterings. Because there is a slight amount of grain growth as sintering time and temperature increases, other microstructural characteristics like fiber diameter are affected. Unsuppressed grain growth may lead to fiber enlargement, de-sintering, and reduced strength. Conversely, if the sintering time is too short, the slow kinetics of densification will not produce a structure dense enough to handle any significant loading. The sintering temperature should also be high enough to densify the microstructure in a timely manner, while reducing de-sintering effects from constrained fibers.
A common theme among many of the observed distributions is their lognormal fit. Be- cause ceramic strength is based on a distribution of flaw sizes, it is reasonable to assume the type of flaw size distribution determines the type of statistics used to evaluate the strength distribution. Weibull statistics may not be the most representative type of statistics for un- derstanding microstructural effects on CNF strength due to their lognormal fiber diameter distribution and exponential pore size distribution. Fiber diameter is inversely proportional with strength, except in the case of the isothermally sintered tubes which were much more porous. A smaller fiber diameter can indicate a smaller volume and probability for large flaws to be present. However, this relationship between fiber diameter is not as strongly evident as
CONTAINS LLNL PROPRIETARY INFORMATION
the relationship between porosity and strength.
The multiple modes of fracture and the multiple Weibull trendlines suggest there are multiple flaw distributions. The CNF tubes fracture due to defects; however, there are both surface defects and volume defects present. Large surface defects could have caused speci- mens to prematurely fail, skewing the strength distribution towards lower values. This can account for the lognormal fit.
The significant peak shift observed in the XRD data do not provide accurate quantitative crystallite sizes, but peak broadness may still comparable across different sintering profiles. CNF strength increased with crystallite size, except for the case of the isothermally sintered tubes which had a much higher porosity. The 875o
C 4 hour sintered tubes had smaller crystallite size but lower strength. One explanation for the TSS strength and TSS crystallite size proportionality is the inverse Hall-Petch relationship; the critical crystallite size where grain boundary strengthening reaches its limit may be in the range∼30 nm.
CONTAINS LLNL PROPRIETARY INFORMATION
7
CONCLUSIONS
1. With a high-temperature step of 1,000o
C, two-step sintering strengthensβ-Ga2O3CNFs optimally with a 2 hour, 875oC sintering dwell, enhancing the average yield stress from
3.36 kPa to 5.72 kPa. However, the components are considered unreliable using Weibull statistics, yielding a Weibull modulus of 2.20.
2. Two-step sintering successfully supresses grain growth at extended sintering times and densifies CNF microstructure at higher sintering temperatures. The minimal grain growth observed points to subprime high-temperature step parameters, where a threshold densification responsible for freezing grain boundary mobility is not reached. 3. Porosity is the major flaw type and microstructural characteristic influencing the aver-
CONTAINS LLNL PROPRIETARY INFORMATION
8
ACKNOWLEDGEMENTS
Wholeheartedly, I would like to thank Dr. James Kelly and Dr. Jeff Haslam at Lawrence Livermore National Laboratory for providing me with a meaningful research topic in elec- trospun nanomaterials. I’d also like to thank my advisor and the department chair of the Materials Engineering Department at Cal Poly, Dr. Trevor Harding, for guiding me throughout my senior year. I’d like to thank Dr. Blair London and Eric Beaton for providing me with the necessary instrumentation and work space to complete this project. Finally, I’d like to thank Cassi Goldsmith for being a great department administrative coordinator.
CONTAINS LLNL PROPRIETARY INFORMATION
9
REFERENCES
[1] P. Mallick,Fiber-reinforced composites. New York: Marcel Dekker, 1993.
[2] C. Zetterling, "Integrated circuits in silicon carbide for high-temperature applications", MRS Bulletin, vol. 40, no. 05, pp. 431-438, 2015.
[3] R. Brown and B. Bellman, "Survey of the Biological Effects of Refractory Ceramic Fibres: Overload and Its Possible Consequences", The Annals of Occupational Hygiene, vol. 49, no. 4, pp. 295-307, 2005.
[4] H. Tan, X. Ma and M. Fu, "Preparation of continuous alumina gel fibres by aqueous solâ ˘A¸Sgel process", Bulletin of Materials Science, vol. 36, no. 1, pp. 153-156, 2013.
[5] D. Li and Y. Xia, "Electrospinning of Nanofibers: Reinventing the Wheel?", Advanced Materials, vol. 16, no. 14, pp. 1151-1170, 2004.
[6] H. Wu, W. Pan, D. Lin and H. Li, "Electrospinning of ceramic nanofibers: Fabrication, assembly and ap- plications", Journal of Advanced Ceramics, vol. 1, no. 1, pp. 2-23, 2012.
[7] P. Panda, "Ceramic Nanofibers by Electrospinning Technique - A Review", Transactions of the Indian Ceramic Society, vol. 66, no. 2, pp. 65-76, 2007.
[8] G. Hussein, "Rare earth metal oxides: formation, characterization and catalytic activity Thermoanalyti- cal and applied pyrolysis review", Journal of Analytical and Applied Pyrolysis, vol. 37, no. 2, pp. 111-149, 1996.
[9] M. Rahaman,Ceramic Processing and Sintering, 2nd ed: Taylor and Francis, 2003.
[10] H. Esfahani, R. Jose and S. Ramakrishna, "Electrospun Ceramic Nanofiber Mats Today: Synthesis, Properties, and Applications", Materials, vol. 10, no. 11, p. 1238, 2017.
[11] J. Stitzel, J. Liu, S. Lee, M. Komura, J. Berry, S. Soker, G. Lim, M. Van Dyke, R. Czerw, J. Yoo and A. Atala, "Controlled fabrication of a biological vascular substitute", Biomaterials, vol. 27, no. 7, pp. 1088-1094, 2006.
[12] Ramakrishna, S., Fujihara, K., Teo, W., Yong, T., Ma, Z. and Ramaseshan, R. "Electrospun nanofibers: solving global issues, Materials Today, 9(3), pp.40-50, 2006.
[13] K. Mondal, "Recent Advances in the Synthesis of Metal Oxide Nanofibers and Their Environmental Re- mediation Applications", Inventions, vol. 2, no. 2, p. 9, 2017.
[14] A. Carvalho, F. Raupp-Pereira, J. Rodrigues Neto and A. Novaes de Oliveira, "A new source for produc- tion of ceramic filters", Materials Letters, vol. 145, pp. 250-252, 2015.
CONTAINS LLNL PROPRIETARY INFORMATION
[15] S. Meille, M. Lombardi, J. Chevalier and L. Montanaro, "Mechanical properties of porous ceramics in compression: On the transition between elastic, brittle, and cellular behavior", Journal of the European Ceramic Society, vol. 32, no. 15, pp. 3959-3967, 2012.
[16] C. Barry Carter and M. Grant Norton,Ceramic Materials: Science and Engineering, New York: Springer- Verlag, 2013.
[17] V. Salvini, V. Pandolfelli and D. Spinelli, "Mechanical Properties of Porous Ceramics", 2018.
[18] N. Savchenko, T. Sablina, I. Sevostyanova, S. Buyakova and S. Kulkov, "Deformation and Fracture of Porous Brittle Materials Under Different Loading Schemes", Russian Physics Journal, vol. 58, no. 11, pp. 1544-1548, 2016.
[19] A. Butcher, C. Koh and M. Oyen, "Systematic mechanical evaluation of electrospun gelatin meshes", Journal of the Mechanical Behavior of Biomedical Materials, vol. 69, pp. 412-419, 2017.
[20] A. Mahdi, M. Mustapa, A. Tobi and I. Zaman, "Micro-Hardness and Compression Strength of Particle Sizes Recycling Aluminium Alloy AA6061 Using Powder Metallurgy Method", Materials Science Forum, vol. 887, pp. 74-82, 2017.
[21] K. Miyake, Y. Hirata, T. Shimonosono and S. Sameshima, "The Effect of Particle Shape on Sintering Behavior and Compressive Strength of Porous Alumina", Materials, vol. 11, no. 7, p. 1137, 2018.
[22] H. Ryou et al., "Below the Hall-Petch Limit in Nanocrystalline Ceramics", ACS Nano, vol. 12, no. 4, pp. 3083-3094, 2018.
[23] R. Chaim, M. Levin, A. Shlayer and C. Estournes, "Sintering and densification of nanocrystalline ceramic oxide powders: a review", Advances in Applied Ceramics, vol. 107, no. 3, pp. 159-169, 2008.
[24] A. Kumar, "Ceramics: Sintering and Microstructure", Ceramics and Colloids: TP3 Sintering.
[25] S. Kang,Sintering: Densification, Grain Growth, & Microstructure, Amsterdam: Elsevier, 2005.
[26] U. Sutharsini, M. Thanihaichelvan and R. Singh, "Two-Step Sintering of Ceramics", 2018.
[27] I. Chen and X. Wang, "Sintering dense nanocrystalline ceramics without final-stage grain growth", Na- ture, vol. 404, no. 6774, pp. 168-171, 2000.
[28] X. Wang, P. Chen and I. Chen, "Two-Step Sintering of Ceramics with Constant Grain-Size, I. Y2O3", Journal of the American Ceramic Society, vol. 89, no. 2, pp. 431-437, 2006.
CONTAINS LLNL PROPRIETARY INFORMATION
2018.
[30] F. Lange, "De-Sintering, A Phenomena Concurrent with Densification Within Powder Compacts: A Re- view", Sintering Technology, pp. 1-12, 1996.
[31] K. Momma and F. Izumi, "VESTA 3 for three-dimensional visualization of crystal, volumetric and mor- phology data", J. Appl. Crystallogr., 44, 1272-1276, 2011.
[32] S. Stepanov and V. Nikolaev, "Gallium Oxide: Properties and Applications - A Review", Review of Ad- vanced Materials Science, vol. 44, pp. 63-86, 2016.
[33] M. Santia, N. Tandon and J. Albrecht, "Lattice thermal conductivity inβ-Ga2O3from first principles",
Applied Physics Letters, vol. 107, no. 4, p. 041907, 2015.
[34] L. Reddy, Y. Ko and J. Yu, "Hydrothermal Synthesis and Photocatalytic Property ofβ-Ga2O3Nanorods",
Nanoscale Research Letters, vol. 10, no. 1, 2015.
[35] S. Arnold, S. Prokes, F. Perkins and M. Zaghloul, "Design and performance of a simple, room-temperature Ga2O3nanowire gas sensor", Applied Physics Letters, vol. 95, no. 10, p. 103102, 2009.
[36] S. Kang, B. Kang, Y. Song and D. Yoon, "Synthesis and structural properties ofβ-Ga2O3nanoparticles
by liquid phase precursor method", Materials Science and Engineering: B, vol. 173, no. 1-3, pp. 105-108, 2010.
[37] C. Sun, J. Deng, L. Kong, L. Chen, Z. Shen, Y. Cao, H. Zhang and X. Wang, "Structure and Photolumi- nescence Properties ofβ-Ga2O3Nanofibres Synthesized via Electrospinning Method", IOP Conference Series:
Materials Science and Engineering, vol. 275, p. 012046, 2017.
[38] T. Zhou, P. Chen, S. Hu, Y. Yan, W. Pan and H. Li, "Enhanced yellow luminescence of amorphous Ga2O3
nanofibers with tunable crystallinity", Ceramics International, vol. 42, no. 5, pp. 6467-6474, 2016.
[39] S. Park, N. Barakat, K. Jeong and H. Kim, "Polymer nanofiber-templated fabrication and characteriza- tion of gallium oxide nanofibers consisting of granular nanoparticles", Polymer International, vol. 60, no. 2, pp. 322-326, 2010.
[40] J. Zhao, W. Zhang, E. Xie, Z. Ma, A. Zhao and Z. Liu, "Structure and photoluminescence ofβ-Ga2O3:Eu3+
nanofibers prepared by electrospinning", Applied Surface Science, vol. 257, no. 11, pp. 4968-4972, 2011.
[41] J. Zhao, W. Zhang, E. Xie, Z. Liu, J. Feng and Z. Liu, "Photoluminescence properties ofβ-Ga2O3:Tb3+nanofibers
prepared by electrospinning", Materials Science and Engineering: B, vol. 176, no. 12, pp. 932-936, 2011.
[42] K. Girija, S. Thirumalairajan, G. Avadhani, D. Mangalaraj, N. Ponpandian and C. Viswanathan, "Syn- thesis, morphology, optical and photocatalytic performance of nanostructuredβ-Ga2O3", Materials Research
CONTAINS LLNL PROPRIETARY INFORMATION
Bulletin, vol. 48, no. 6, pp. 2296-2303, 2013.
[43] Y. Zhao, R. Frost and W. Martens, "Synthesis and Characterization of Gallium Oxide Nanostructures via a Soft-Chemistry Route", The Journal of Physical Chemistry C, vol. 111, no. 44, pp. 16290-16299, 2007.
[44] J. Du, J. Zhao, Q. Luo, Z. Yu and J. Xia, "Growth of One-dimensional Gallium Nitride Nano- and Mi- crostructures in an Alumina Matrix Containing Gallium Oxide", Bulletin of the Korean Chemical Society, vol. 27, no. 8, pp. 1235-1238, 2006.
[45] J. Du, J. Zhao, Q. Luo, Z. Yu, J. Xia and M. Yang, "Growth and Characteristics Analysis of the Thermal Oxide Grown on Gallium Nitride", 2009 Symposium on Photonics and Optoelectronics, 2009.
[46] C. Tang, Y. Bando and Z. Liu, "Thermal oxidation of gallium nitride nanowires", Applied Physics Let- ters, vol. 83, no. 15, pp. 3177-3179, 2003.
[47] J. Kelly, J. Haslam, L. Finkenauer and M. Mitchell, "NSRD-12, Novel Mini-Tubular HEPA Media for Nuclear Facility Ventilation Systems", 2018.
[48] R. Roy, V. Hill and E. Osborn, "Polymorphism of Ga2O3and the System Ga2O3-H2O", Journal of the American
Chemical Society, vol. 74, no. 3, pp. 719-722, 1952.
[49] V. Berbenni, C. Milanese, G. Bruni and A. Marini, "Thermal decomposition of gallium nitrate hydrate Ga(NO3)3xXH2O", Journal of Thermal Analysis and Calorimetry, vol. 82, no. 2, pp. 401-407, 2005.
[50] J. Shan, C. Li, J. Wu, J. Liu, A. Chen and Y. Shi, "Sintering behavior and microstructural evolution of the monodispersedβ-gallium oxide micro-particles with different morphology and size", Ceramics International, vol. 43, no. 18, pp. 16843-16850, 2017.
[51] A. Rabbani and S. Salehi, "Dynamic modeling of the formation damage and mud cake deposition us- ing filtration theories coupled with SEM image processing", Journal of Natural Gas Science and Engineering, 42, 157-168, 2017.
[52] J. Seuba, S. Deville, C. Guizard and A. Stevenson, "The effect of wall thickness distribution on mechanical reliability and strength in unidirectional porous ceramics", Science and Technology of Advanced Materials, vol. 17, no. 1, pp. 128-135, 2016.
[53] P. de Wit, F. van Daalen and N. Benes, "The mechanical strength of a ceramic porous hollow fiber", Journal of Membrane Science, vol. 524, pp. 721-728, 2017.
[54] B. Basu, D. Tiwari, D. Kundu and R. Prasad, "Is Weibull distribution the most appropriate statistical strength distribution for brittle materials?", Ceramics International, vol. 35, no. 1, pp. 237-246, 2009.
CONTAINS LLNL PROPRIETARY INFORMATION
[55] A. 13(2018), "ASTM C1239 - 13(2018) Standard Practice for Reporting Uniaxial Strength Data and Esti- mating Weibull Distribution Parameters for Advanced Ceramics", Astm.org, 2019.