Velcro-Inspired SiC Fuzzy Fibers for Aerospace Applications - ACS

Mar 28, 2017 - *E-mail: [email protected]., *E-mail: [email protected]., *E-mail: [email protected]. Cite this:ACS Appl. Mater. Interfaces 9, 15, 13742...
1 downloads 0 Views 6MB Size
Research Article www.acsami.org

Velcro-Inspired SiC Fuzzy Fibers for Aerospace Applications Amelia H. C. Hart,†,‡ Ryota Koizumi,† John Hamel,† Peter Samora Owuor,† Yusuke Ito,† Sehmus Ozden,† Sanjit Bhowmick,§ Syed Asif Syed Amanulla,§ Thierry Tsafack,† Kunttal Keyshar,† Rahul Mital,‡ Janet Hurst,*,‡ Robert Vajtai,† Chandra Sekhar Tiwary,*,† and Pulickel M. Ajayan*,† †

Materials Science and NanoEngineering Department, Rice University, Houston, Texas 77005, United States NASA Glenn Research Center, Cleveland, Ohio 44135, United States § Hysitron Inc., Minneapolis, Minnesota 55344, United States ‡

S Supporting Information *

ABSTRACT: The most recent and innovative silicon carbide (SiC) fiber ceramic matrix composites, used for lightweight high-heat engine parts in aerospace applications, are woven, layered, and then surrounded by a SiC ceramic matrix composite (CMC). To further improve both the mechanical properties and thermal and oxidative resistance abilities of this material, SiC nanotubes and nanowires (SiCNT/NWs) are grown on the surface of the SiC fiber via carbon nanotube conversion. This conversion utilizes the shape memory synthesis (SMS) method, starting with carbon nanotube (CNT) growth on the SiC fiber surface, to capitalize on the ease of dense surface morphology optimization and the ability to effectively engineer the CNT−SiC fiber interface to create a secure nanotube−fiber attachment. Then, by converting the CNTs to SiCNT/NWs, the relative morphology, advantageous mechanical properties, and secure connection of the initial CNT−SiC fiber architecture are retained, with the addition of high temperature and oxidation resistance. The resultant SiCNT/NW−SiC fiber can be used inside the SiC ceramic matrix composite for a high-heat turbo engine part with longer fatigue life and higher temperature resistance. The differing sides of the woven SiCNT/NWs act as the “hook and loop” mechanism of Velcro but in much smaller scale. KEYWORDS: silicon carbide nanotubes and nanowires, carbon nanotubes, aerospace applications, in situ mechanical testing, thermal stability



INTRODUCTION

Specifically, the most current, cutting edge technology in aerospace applications is the use of woven, layered nearstoichiometric SiC fibers (SiCF) in SiC CMCs for high heat engine parts to enable the creation of turbo engines that are larger and operate at higher temperatures than any other engine previously constructed.12−16 To achieve its ultimate application temperature and life cycle, however, the SiCF−SiC matrix composite must overcome several problems, resulting from mechanical stress-induced cracks and/or oxidation. The three most significant problems include: fiber pullout from the matrix or fibers being pulled from the matrix resulting in a void where the fiber was previously; weakness in the z-direction, normal to the fiber weave, causing the fibers to slide in the x and y directions, parallel to the fiber weave, as seen relative to the schematics in Figure 1(a); and/or oxygen to enter the composite and break down the fiber.17 The most common method of protecting the fiber from oxygenation, and eventual breakdown, and to improve mechanical strength is the use of a two-dimensional boron

Better functioning, easier to produce, lightweight, mechanically/thermally/electrically robust, chemically inert materials with a highly active surface area are constantly sought after to improve materials currently used in the extreme environments encountered in many applications, specifically aerospace applications.1,2 The National Aeronautics and Space Administration (NASA), the leader in aerospace technology, is interested in the investigation of ultralight-weight materials that function in a wide variety of capacities, such as thermal and mechanical. Silicon carbide has been long known and used for its high hardness, low density, high-temperature mechanical strength, corrosion, radiation, oxidation resistance, chemical inertness, wide band gap (2.2−3.3 eV), and high decomposition temperature (2545 °C).3−5 In the form of fiber, silicon carbide can utilize these advantageous mechanical and thermal properties with a diameter (8.5−15 μm) small enough to have increased flexibility but large enough to be able to add continuous strength to a composite material.6−10 The synthesis, composition, crystallinity, and mechanical and thermal properties of SiC fibers have been optimized since the 1970s, primarily for structural reinforcements in ceramic matrix composites (CMCs) for use in air at temperatures above 1000 °C.7,8,11 © XXXX American Chemical Society

Received: January 28, 2017 Accepted: March 28, 2017 Published: March 28, 2017 A

DOI: 10.1021/acsami.7b01378 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Figure 1. Schematic of nanotubes on SiC fiber. (a) Microscopic views illustrating the interlocking of the nanowire/nanotubes of the SiC fuzzy fiber when multiple fibers are woven together. (b) Schematic of the steps of the CNT growth on the SiC fiber: (i) bare SiCF, (ii) iron catalyst particles adhering to the surface of the SiCF, and (iii) growth of CNTs on the SiCF; and conversion of the CNT−SiCF to SiCNT/NW−SiCF: (iv) immersion of the CNT−SiCF in silicon particles, (v) the conversion of CNTs to SiCNT/NWs. (c) Schematic of: (i)−(iii) BN-coated SiCF in the SiC matrix, (i) under 700 °C, (ii) BN coating evaporating and leaving void between the fiber and matrix at 700−800 °C, and (iii) SiO2 filling the void above 800 °C, and (iv)−(v) SiCNT/NW-coated SiCF in the SiC Matrix (iv) under 800 °C and (v) coated with SiO2 above 800 °C.

properties than the larger SiCFs.10,23−25 However, adding dispersed SiCNT/NWs alone to a composite does not significantly enhance the mechanical strength compared to woven SiCF.26 By embedding SiCNT/NWs in the surface of the SiCFs, via carbon nanotube (CNT) template conversion, the SiCNT/NWs act as a high-surface-area coating to prevent the oxygen being pushed through the cracking matrix from affecting the fiber and form a SiO/SiO2 coating at high temperatures, as shown in Figure 1(c) (iv, v). Furthermore, the SiCF surface-anchored SiCNT/NWs entwine with the matrix and each other in a Velcro-like fashion, preventing pull-out from the matrix as well as preventing fiber−fiber sliding and increasing the interlaminar strength of the woven SiCF layers in the CMC, depicted in Figure 1(a), all which increase the overall strength of the structure. Where actual Velcro contains multiple stiff, curled “hooks” which grab or catch onto “loops” of thread, here the “fuzzy” “hooks” of SiCNT/NWs on the fiber intertwine with each other and the matrix. The SiCNT/

nitride (BN) coating, which can be deposited via chemical vapor infiltration (CVI).18−20 Hexagonal BN, deposited at or above 1000 °C, provides high fiber toughness.17,21,22 In dry air, the BN produces a boron oxide which reacts with the silica to become borosilicate glass, sealing the void between the matrix and fiber. However, in humid environments, as shown in Figure 1(c) (i−iii), the boron oxide liquid turns into boron hydroxides, which evaporate away leaving a void between the matrix and fiber. Though at 900−1200 °C the SiC fiber and matrix produce a thick SiO2 layer to fill the void, between 700 and 800 °C this unfilled void causes oxygen to affect the fiber, causing embrittlement and eventual fiber break down.9,17,21 To prevent this matrix−fiber void formation and to increase the CMC’s overall mechanical strength, SiC nanotubes and nanowires (SiCNT/NWs) are synthesized on the surface of the SiCF, dubbed a “fuzzy fiber”. These nanocrystalline forms of SiC, due to nanosize effects, are comparatively able to withstand higher temperatures and have better mechanical B

DOI: 10.1021/acsami.7b01378 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Figure 2. Characterization and mechanical testing of CNT−SiCF. SEM images: (a) as-received SiC fiber (scale bar: 5 μm), (b) melted iron particles on iron-dipped SiC fiber heated under growth conditions (scale bar: 2 μm), (c) CNT−SiCF (scale bar: 10 μm), (d) high-resolution CNT−SiCF (scale bar: 3 μm). TEM images: (e) lower-resolution CNT−SiCF (scale bar: 1 μm), (f) end of CNTs curled around iron particles (scale bar: 50 nm), (g) end-cap of CNT (scale bar: 50 nm), (h) CNT with multiple walls (scale bar 10 nm), (i) lower-resolution curly CNT bundle protruding from SiC fiber (scale bar: 200 nm); (j) Raman of bare SiCNTF (black) and CNT−SiCF (blue); (k) schematic of the PicoIndenter mechanical test, (l) friction−time graph for SiCF (black) and CNT−SiCF (blue), (m) graph of lateral force vs lateral displacement for bare SiCF (black) and CNT− SiCF (blue), (n) load−depth curve for the CNT−SiCF compression test; (o) SEM image of the compression test; (p) stress−strain curve for the tensile test of the PDMS/SiCF (black) and PDMS/CNT−SiC fiber composites (blue) with illustration of the composite; (q) SEM image of the bare fiber pulling out of the PDMS; (r) SEM image of the CNT−SiCF pulling out of the PDMS; and (s) SEM image of entangled CNTs between SiCFs.

NW−SiCFs, which contain very curly nanotubes/nanowires radially protruding outward from a larger fiber, are ultimately woven, creating a fuzzy fiber−fuzzy fiber “hook and loop” mechanism, as well as a fuzzy fiber “hook” and matrix (polymer or ceramic) “loop” mechanism. In the fiber−fiber scenario, the first “hook” fuzzy fiber grabs and secures onto the second “loop” fuzzy fiber, allowing the curly nanotube/wires to intertwine. In the fiber−matrix scenario, the fuzzy fiber “hook” is securely embedded in the “loop” matrix and holds tight when being pulled. To create the strongest fiber_-nanotube/wire interfacial bonding necessary to mimic the mechanisms of Velcro and possess the desired morphology, carbon nanotubes (CNTs) are used as a template to synthesize the SiCNT/NWs. CNTs, known for their superb mechanical properties and easy synthesis methods, can be grown directly on the fiber surface

using catalyst dip-coating and water-assisted chemical vapor deposition (WACVD) to enable a strong nanotube−substrate adhesion.27−34 By optimizing the catalyst type and deposition method and CNT growth time, the desired fuzzy fiber morphology of densely packed curly, curvy CNTs is achieved to maximize the fiber’s intertwining and grasping capabilities. This ability to specifically and easily optimize the growth of CNTs on fiber has been well documented.35−41 The CNT− SiCF is characterized and mechanically tested to show the ability of this “fuzzy fiber” morphology to significantly enhance the mechanical properties of the fiber and composite as well as to exhibit the superior nanotube−fiber adhesion. However, due to their susceptibility to oxidation and high temperatures, it is necessary for their intended application to convert the CNTs to SiCNT/NWs via immersion in Si nanopowder at high temperature.42,43 This conversion technique enables the C

DOI: 10.1021/acsami.7b01378 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

also the density of the CNTs on the SiCF. The optimal CNT growth time was 5 h, the morphology of which is seen in the SEM images in Figure 2(a)−(d). At this growth time, the CNTs were long enough to completely cover the SiC fiber, as well as entangle with each other with minimal contact, as can be seen in Figure 2(s). The CNTs are seen to not only grow outward but also curl around other CNTs and around the fiber creating a large mass of entangled curvy, curly CNTs. This morphology enables the fiber to anchor itself more securely to other fibers, increasing the interlaminar strength of the composite. From the TEM images in Figure 2(e)−(i), the closely situated CNTs are about 1−3 μm long, have about 14 walls, an inner diameter of 10−13 nm, an outer diameter of about 20 nm, and can contain residual iron particles. The Raman spectra in Figure 2(j) show the comparison of bare SiCF (black) and 5 h CNT−SiCF (blue). After CNT growth, the SiC peaks at 700−1000 cm−1 are seen to decrease, due to the thickness of the CNT coating and the Raman laser penetration length. Before CNT growth, the bare SiCF shows large carbon peaks at 1300−3100 cm−1 from the intergranular free C, even though there is very little free C in this type of SiC fiber.48−51 The higher intensity D-band, or disordered band, is related to the number of defects in the crystal structure. The lower intensity G-band is in all graphitic structures and shows the sp2 stretching modes of the C bonds. The ratio of the intensities of the D-band over the G-band is inversely proportional to the in-plane graphitic crystallite size and is also a good reference for the amount of disorder in the sample.52 Here, the carbon D-band/G-band intensity ratio decreases from the bare SiC fiber to the 5 h CNT−SiC fiber, showing an increased amount of ordered carbon, or CNTs. CNTs are initially grown on the SiCF to act as both an anchor and template, to create the morphological shape that the SiCNTs will take after conversion. To test how well the CNTs are embedded in the SiCF, the fuzzy fiber underwent in situ friction and compression testing, as shown in Figure 2(k)− (o). During the friction testing, illustrated in Figure 2(k), a probe moves laterally over the fiber, with and without CNTs, in one direction 5 μm, in the opposite direction 10 μm, then back to the starting point. This test measures the lateral force necessary to overcome the static friction on the surface of the sample to continue to move the probe in two directions. The friction between the sample and the probe as a function of time was calculated using normal and lateral forces. The graph of friction vs time in Figure 2(l) shows a larger CNT− SiC fiber friction coefficient compared to the bare fiber, demonstrating the strong CNT−SiCF interface adhesion as well as the retention of the structural integrity of the CNTs. In the plot of the lateral force vs lateral displacement in Figure 2(m), the lateral force necessary to move over the CNTs was consistently greater than the SiC fiber alone; i.e., the CNTs were not pulled out of the fiber and were thoroughly embedded. If the CNTs are either pulled from the fiber or crushed with the lateral force of the probe, the friction and the lateral force necessary to move the probe would decrease dramatically due to the smoothness of the fiber and the lamellar slippage of the resulting graphite from the crushed nanotubes. During individual nanotube compression testing, to further measure the ability of the CNTs in the SiC fiber to withstand outside compressive pressure and avoid breakage, the CNT completely recovered its shape after being compressed under a 4 μN load. In Figure 2(n), this compression test was plotted as load vs depth, with an SEM image of the test in Figure 2(o).

SiCNT/NWs to retain the relative morphology and secure attachment of the initial CNTs but with high thermal stability in air. To support this claim, and to compare their mechanical and thermal properties, both the CNT−SiCF and SiCNT/ NW−SiCF were characterized and mechanically and thermally tested. Compared to BN, this CNT template-assisted synthesis of SiCNT/NW−SiCF simplifies the CMC fabrication process and provides better control of interface engineering. The nanosize, one-dimensional structure with high surface area and mechanical properties can provide structure rigidity to the CMC. The addition of a nanostructure with the same elemental makeup as the fiber and matrix can also help improve the atomic interface without any additional steps. Using our two-step process, strongly adhered, long, dense CNTs are grown on the surface of the SiC fiber, then converted to SiCNT/NWs to create a novel SiCNT/NW-coated SiC fiber that increases the strength, toughness, and recoverability of SiC ceramic matrix composites used in high temperatures and oxidative environments. Compared to the direct synthesis of SiCNT/NWs, the template synthesis method uses the versatility and interface strength of CNT growth to easily and economically optimize the surface morphology of the fuzzy fiber by increasing CNT lengths and SiCF surface density. Though both the growth of CNTs on SiC fibers and CNT-toSiCNT/NW conversion techniques have been reported separately, this specific CNT growth on the SiC fiber method as well as their subsequent conversion to SiCNT/NWs and their application for mechanical use in high-temperature and oxidative environments are novel. To demonstrate the mechanical strength and resiliency of both the NT/NWs and NT/NW−fiber interfaces, the CNT− SiCF and SiCNT/NW−SiCF undergo in situ SEM indention and friction testing as well as tensile testing within a polydimethylsiloxane (PDMS) matrix. Because PDMS is transparent and its mechanical properties are well documented, particularly in composites, it was selected to be the better matrix for the demonstration of the Velcro-like adherence abilities of the nanotubes/wires.44−47 Additionally, as the process of infiltrating the woven fibers with the SiC matrix used in the intended application involves high heat, the resulting degradation of the CNTs would prevent a comparison of the SiCNT/NWs to the CNTs. For these reasons, PDMS was selected to act as the matrix for tensile mechanical testing in which to compare the three SiC fibers, bare and with either CNTs and SiCNT/NWs. Additionally, thermal gravimetric analysis (TGA) and butane lighter burning tests are performed on both types of fuzzy fibers to compare their resiliencies in temperatures between 800 and 1000 °C, in air. The schematic in Figure 1(b) illustrates the two parts necessary to create the fully converted SiCNT/NW−SiCF: (i)−(iii) growing securely attached, long, SiCF−surface dense CNTs via iron catalyst and (iv)−(v) converting the CNTs to SiCNT/NWs with silicon and high heat. A representative SEM image of an individual as-received fiber is depicted in Figure 2(a), with a smooth surface devoid of craters or other macroscopic defects. After soaking the fiber in the iron solution and heating under CNT growth conditions, the size-similar iron particles can be seen uniformly distributed over the surface, as shown in the high-magnification SEM image in Figure 2(b). This iron particle morphology initiates more uniform and closepacked 20−90 nm diameter CNTs on the SiC fiber surface, as can be seen in Figure 2(c) and (d). As the CNT growth times increase, not only does the average CNT length increase, but D

DOI: 10.1021/acsami.7b01378 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Figure 3. Characterization and mechanical testing of SiCNT/NW−SiCF: (a)−(c) SEM images of SiCNT/NW−SiCF with scale bar: (a) 10 μm, (b) 0.5 μm, and (c) 2.5 μm; (d)−(h) TEM images of SiCNT/NW−SiCF with diffraction pattern (f) and with scale bar: (d) 5 μm, (e) 10 μm, (g) and (h) 20 nm; (i) full XPS spectrum and (j) Si 2p XPS spectra of SiCNT/NW−SiCF and CNT−SiCF; (k) Raman of SiCNT/NW−SiCF and CNT− SiCF; (l) plot of lateral force vs lateral displacement for bare SiCF (black), CNT−SiCF (blue), and SiCNT/NW−SiCF (red); (m) friction−time graph for CNT−SiCF (blue) and SiCNT/NW−SiCF (red); (n) load−depth curve for CNT−SiCF (red) and SiCNT/NW−SiCF (blue) compression test; (o) SEM images of the compression test; (p) stress−strain curve for the tensile test of the PDMS/SiCF (black) and PDMS/ SiCNT/NW-SiCF (red) composites with illustration of the composite; (q)−(s) SEM images of SiCNT/NW−SiCF pulling out of the PDMS matrix with scale bar: (q) 25 μm, (r) 10 μm, and (h) 5 μm.

To illustrate the mechanical improvement of the composite due to the nanotubes/nanowires, the fibers are cut and dispersed in PDMS, which acts as the SiC matrix. Plotted in Figure 2(p), the stress−strain curve of PDMS/CNT−SiCF and PDMS/SiCF composites, illustrated in the inset, shows that the CNTs on the surface of the SiC fiber greatly enhance the tensile strength of the composite due to the CNTs’ ability to entangle with each other and the PDMS matrix, providing increased resistance to tearing. The SiC fiber alone initially aids the composite in resisting fracture but is eventually pulled from the matrix causing catastrophic failure. The void between the fiber and PDMS matrix depicted in the SEM image in Figure 2(q) demonstrates the low interfacial strength of the composite compared to the CNT−SiCF/PDMS composite in Figure 2(r). The ability of the CNTs to readily entangle with each other is shown in the SEM image in Figure 2(s). This image was taken with no outside force on the fibers or nanotubes, as they were simply near each other.

Once the CNT−SiCF is optimized and its interface verified, the CNTs are converted to SiCNT/NWs. In the SEM image in Figure 3(a), the converted SiCNT/NWs emulate the CNT morphology on the SiC fiber; the SiCNTs completely cover the surface of the SiC fiber and are long enough to create a network with other fuzzy fibers. The higher-resolution SEM image (Figure 3(b) and (c)) shows the presence of nanotubes as well as nanowires and excess silicon and silicon carbide. Both amorphous silicon or silicon carbide nanowires on the surface of the SiCNT/NWs are inevitable due to the conversion conditions and the structural stability of NWs which also contribute to the advantageous mechanical and thermal capabilities of the resulting fuzzy fiber.53 The lowermagnification TEM image of the 5 h converted SiCNT/ NW−SiCF in Figure 3(d) shows a nice distribution of both nanowires (straighter) and nanotubes (curlier), as well as the presence of junctions between nanowires and/or nanotubes. In the higher-resolution TEM images in Figures 3(e), (g), and E

DOI: 10.1021/acsami.7b01378 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Figure 4. Thermal testing of CNT−SiCF and SiCNT/NW−SiCF: burning of CNT−SiCF with a butane lighter for 1 min: SEM images (a) before burning (scale bar 5 μm), (b) after burning (scale bar 5 μm), (c) Raman spectra of before (blue) and after burning (orange) with inset ID/IG for comparison; (d) TGA plotted as weight gain/loss (%)−temperature of CNT−SiC fiber and SiCNT/NW−SiCF in air, up to 900 °C. Burning of SiCNT/NW−SiCF with a butane lighter for 1 min: SEM images: (e) before burning (scale bar 5 μm) and (f) and (g) after burning ((f) scale bar 5 μm and (g) scale bar 1 μm). (h) Raman spectra of before (red) and after burning (orange); optical images of: (i) CNT−SiCF with burned and not burned ends and SiCNT/NW−SiCF (j) before burning and (k) after burning.

plotted as load vs depth in Figure 3(n) with SEM inset in Figure 3(o), show that the SiCNT/NW−SiCF (red) can withstand higher loads at greater depths than the CNT−SiCF (blue). It was also found that the compressive elastic modulus of the SiCNT/NW−SiCF is 56 GPa. The addition of SiCNT/NWs to the surface of the SiC fiber improves the mechanical properties of the currently used SiC fiber−ceramic matrix composite by entangling with each other as the fibers slide against each other, which prevents excessive movement and the eventual SiC fiber pullout from the ceramic matrix. To demonstrate the adhesion of SiCNT/NWs to the matrix and compare them to the SiC fiber alone, SiCNT/NW− SiC fiber and bare fiber were each immersed in PDMS and tensile tested via DMA. It should be noted that the overall Young’s moduli of these samples are low compared to the literature, due to the method in which the sample is made. These values are strictly for comparison. The stress−strain curves, plotted in Figure 3(p), show the addition of the SiCNT/NWs improves the mechanical properties of the SiCF/ PDMS composite. Here, the bare SiCF/PDMS composite breaks at 47% strain and has a Young’s modulus of 2.4 kPa, whereas the SiCNT/NW−SiCF/PDMS composite breaks at 98% strain and has a Young’s modulus of 5.1 kPa. This increased Young’s modulus is due to the ability of the SiCNT/ NWs to entangle with the PDMS matrix, as shown in SEM images in Figure 3(q)−(s), preventing SiC fiber pullout from the matrix. To demonstrate the high heat and oxidation resistance of the SiCNT/NW−SiC fiber compared to the CNT−SiC fiber, TGA (in air up to 950 °C) and manual burning (bic butane lighter, for 1 min in lab air, up to 1000 °C) were performed on both. During TGA, seen in Figure 4(d), the CNTs burned off the fiber, significantly reducing the weight of the CNT−fiber, whereas the weight of the SiCNT−fiber increased, due to the accumulation of SiO/SiO2. Similarly, in the SEM images (Figure 4(a), (b), (e), (f), and (g)) and Raman spectra (Figure 4(c) and (h)) of each fuzzy fiber type before and after the butane lighter burning, the CNTs were completely burned off

(h), a Si/SiO2/SiC coating covers the nanotube, where the actual nanowire/nanotube diameter is ∼20 nm, and the coating is ∼3 nm. The SiCNWs seen in Figure 3(e) grow during the conversion from excess silicon and carbon atoms and are longer and straighter than the filled SiCNTs, with stacking faults and a cubic [111] crystal structure, as seen in Figure 3(f). Additionally, as seen in Figure 3(h), junctions and couplings form between the SiCNT/NWs during the conversion. The XPS data of the CNT−SiCF (blue) and SiCNT/NW−SiCF (red) in Figure 3(i) show the evolution and accumulation of the elements in the nanotubes/nanowires as they transition from carbon to silicon carbide. As expected, the carbon peak is higher, and as seen in Figure 3(j), the Si 2p peak is lower for the CNT−SiCF compared to the SiCNT/NW−SiCF. The Raman spectra, seen in Figure 3(k), of the CNT−SiCF (blue) and SiCNT/NW−SiCF (red) show the change in the silicon carbide and carbon peaks. Before the conversion, when the nanotubes are all carbon, there is no SiC peak, and a low Dband−G-band intensity peak ratio confirms the presence of CNTs. After 5 h of conversion, the SiC peaks are very high and have a high D-band−G-band intensity peak ratio, confirming conversion. To test how well the SiCNT/NWs are embedded into the SiC fiber, this fuzzy fiber underwent the same friction and nanotube compression testing as with the CNT−SiCF. The lateral force versus lateral displacement plot in Figure 3(l) demonstrates that the lateral force necessary to move over the SiCNT/NWs (red) was as much or higher than the CNTs (blue) and SiC fiber (black), showing the SiCNT/NWs were thoroughly embedded in the SiC fiber. The tight adhesion between the SiCNT/NWs and SiCF can be seen in the friction coefficient vs time plot in Figure 3(m), where the friction coefficient for the SiCNT/NW−SiCF (red) is similar to the CNT−SiCF (blue). To further measure the ability of the SiCNT/NW−SiCF to withstand uniaxial compressive pressure and avoid breakage, individual nanotube compression was performed with the PicoIndenter. The same tests were performed with the CNT−SiCF. The compression tests, F

DOI: 10.1021/acsami.7b01378 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces the SiC fiber, while the SiCNT/NWs remained intact. In the optical image in Figure 4(i) it is clear which side is burned (left, gray/orange from excess iron) and unburned (right, black from CNTs). The SEM images show that before burning (Figure 4(a)) the CNTs are very curly, long, and densely situated around the SiC fiber and that after burning (Figure 4(b)) the CNTs are completely burned off, and only ash or excess burned carbon remains on the surface of the fiber. The after-burning Raman spectra in Figure 4(c) are like the spectra of the bare SiC fiber, with larger D-band and SiC peak. This may be due to the burning of the outer layer of the SiC fiber, revealing more of the internal structure. The SEM and Raman spectra of the SiCNT/NW−SiC fiber in Figures 4(e)−(h) look the same before and after burning, meaning the SiCNT/NWs can withstand high temperatures and oxidation.

to 15% Hydrogen−85% Argon (Ar−H2) as the carrier gas; and the furnace is heated to 775 °C. Once the growth temperature is reached, the 5 h growth is carried out under 100 standard cubic centimeters (sccm) Ar−H2 through a water bubbler, 100 sccm of ethylene (carbon source), and 1.3 slm Ar−H2 (carrier gas). After growth, the system is cooled under 100 sccm Ar−H2. Once CNTs are grown on the SiCF, they are converted to SiC nanotubes by an additional high-temperature CVD process. During this process the CNT−SiCF is immersed in Si nanopowder (SigmaAldrich, product number 633097,