Article pubs.acs.org/Langmuir

Chemical Bath Deposition of Aluminum Oxide Buffer on Curved Surfaces for Growing Aligned Carbon Nanotube Arrays Haitao Wang† and Chongzheng Na*,†,‡ †

Department of Civil and Environmental Engineering and Earth Sciences, University of Notre Dame, 156 Fitzpatrick Hall, Notre Dame, Indiana 46556, United States ‡ Department of Civil, Environmental, and Construction Engineering, Texas Tech University, 10th Street and Akron Avenue, Lubbock, Texas 79409, United States S Supporting Information *

ABSTRACT: Direct growth of vertically aligned carbon nanotube (CNT) arrays on substrates requires the deposition of an aluminum oxide buffer (AOB) layer to prevent the diffusion and coalescence of catalyst nanoparticles. Although AOB layers can be readily created on flat substrates using a variety of physical and chemical methods, the preparation of AOB layers on substrates with highly curved surfaces remains challenging. Here, we report a new solution-based method for preparing uniform layers of AOB on highly curved surfaces by the chemical bath deposition of basic aluminum sulfate and annealing. We show that the thickness of AOB layer can be increased by extending the immersion time of a substrate in the chemical bath, following the classical Johnson−Mehl−Avrami−Kolmogorov crystallization kinetics. The increase of AOB thickness in turn leads to the increase of CNT length and the reduction of CNT curviness. Using this method, we have successfully synthesized dense aligned CNT arrays of micrometers in length on substrates with highly curved surfaces including glass fibers, stainless steel mesh, and porous ceramic foam.

1. INTRODUCTION Multiwalled carbon nanotubes (CNTs) made of concentric rolls of graphene sheets often have diameters on the order of nanometers but can grow up to centimeters in length.1 The quasi-one-dimensional structure gives CNTs remarkable electrical conductivity,2−4 heat conductivity,5,6 and tensile strength7,8 as well as a high surface-to-volume ratio,9,10 making them excellent anisotropic materials for manufacturing microelectronics, reinforcing polymers, and supporting nanocatalysts.11−13 In these applications and devices, a large number of CNTs are often used together to create complex macroscopic structures. To preserve anisotropy during scale-up, individual CNTs can be aligned along their longitudinal directions, forming CNT arrays. Previously, aligned CNT arrays have been synthesized on relatively flat substrates using thermal chemical vapor deposition (CVD),14,15 one of the standard methods for large-scale CNT production.15,16 Compared to flat substrates, synthesizing CNT arrays on fibers, meshes, and foams that have highly curved surfaces is still challenging. A main technical hurdle is how to create an even layer of aluminum oxide buffer (AOB, γ-Al2O3) before growing CNT arrays using CVD. A buffer layer is a required component of the CVD operation, in which CNTs grow from metallic catalyst nanoparticles at high temperatures (usually 600−800 °C) at the expense of a hydrocarbon compound.17 Without the buffer layer, catalyst nanoparticles would coalesce with one another or diffuse into the substrate at the high © 2015 American Chemical Society

temperatures in CVD, leading to rapid deactivation of the catalyst and cessation of CNT growth.17−20 A variety of materials have been investigated as buffer, including aluminum oxide (Al2O3), titanium nitride (TiN), titanium dioxide (TiO2), calcite (CaCO3), and silicon dioxide (SiO2);21−23 among which, γ-Al2O3 is often considered as the most effective buffer material for growing CNT arrays.24 For flat substrates with fully open and accessible surfaces, AOB can be prepared by both physical and chemical methods, including atomic layer deposition (ALD), electron beam physical vapor deposition (a.k.a. e-beam evaporation), thermal evaporation, spin coating, dip coating, the sol−gel process, and the layer-by-layer (LBL) assembly.25−39 When the substrates have spatially complex structures such as highly curved surfaces, particularly those hidden in torturous pores, the creation of a uniform AOB layer to support the growth of CNT arrays is no longer feasible by using most of the existing methods. To our knowledge, aligned CNT arrays have only been synthesized on highly curved surfaces either treated using the expensive ALD method40,41 or with preexisting oxide coatings that can act as buffer, for example, ceramic fibers made of aluminum oxide42,43 and stainless steel with chromium oxide.44,45 Received: March 18, 2015 Revised: June 4, 2015 Published: June 8, 2015 7401

DOI: 10.1021/acs.langmuir.5b01002 Langmuir 2015, 31, 7401−7409

Article

Langmuir

Figure 1. Synthesis of aligned carbon nanotube arrays on highly curved surfaces. Main steps: (1) chemical bath deposition of basic aluminum sulfate; (2) annealing at 750 °C for 1 h; and (3) chemical vapor deposition after deposition of iron oxide nanoparticles. Colors: gray, substrate; blue, AOB layer; orange, iron catalyst; black, carbon nanotubes. magnetite nanoparticle suspension for 30 min. At the end of deposition, the substrate was rinsed with water to remove loosely adsorbed nanoparticles and blown dry using nitrogen. Prior to CNT growth, the substrate was treated with oxygen plasma for 5 min to remove organic contaminants on the substrate’s surface. To grow CNTs, the magnetite-coated substrate was placed in a quartz boat at the center of a 2-in. quartz tubing, which was housed in a tube furnace. The furnace was ramped to 750 °C in the flow of 500 sccm argon. Once the temperature reached 750 °C, 100 sccm hydrogen was introduced to transform iron oxide nanoparticles to metallic iron catalysts. The growth of CNTs was initiated by introducing 100 sccm ethylene. A trace amount of water vapor was added to the reactor by bubbling a small amount of argon through a water tank. After 20 min of growth, the furnace was cooled to the ambient temperature under argon protection. The morphology and quality of AOB layer and CNT arrays were examined using scanning electron microscopy (SEM; FEI Magellan 400), transmission electron microscopy (TEM; FEI 80−300 at 300 kV), atomic force microscopy (AFM; Park System XE-70), powder Xray diffraction (XRD; Advance 8 at 40.0 kV and 120 mA with Cu Kα radiation), Raman spectroscopy (Renishaw RM2000 with He−Ne laser excitation line 532 nm), and ellipsometry (Gaertner Scientific). Sample preparation and data interpretation were performed following standard protocols.

Here, we report a general and facile method for preparing AOB on substrates with highly curved surfaces for growing CNT arrays using CVD. We show that after a substrate is immersed in a chemical bath containing aluminum sulfate (Al2(SO4)3) and sodium bicarbonate (NaHCO3), a thin layer of AOB can be formed after annealing. The thickness of the AOB layer can be controlled by adjusting the immersion time, thereby allowing the process to be optimized for efficiency. Using silicon chip, glass fiber, stainless steel mesh, and ceramic foam as model substrates, we further show that aligned CNT arrays up to hundreds of micrometers in length can be readily synthesized using conventional CVD after the substrates are treated by this chemical bath deposition (CBD)46 method. The key feature of CBD is the controlled crystallization of basic aluminum sulfate (BAS; Al3(H3O)(SO4)2(OH)6), which has been previously used to synthesize aluminum oxide and oxyhydroxide particles.47,48

2. EXPERIMENTAL SECTION All chemicals were of analytical grade and purchased from SigmaAldrich. Deionized (DI) water was generated on site using a Millipore system. Glass fibers (diameter: 4 μm), stainless steel mesh (AISI 304; 400 × 400 openings per square inch), and porous cordierite ((Mg,Fe)2Al3(Si5AlO18)) monolith (14 × 14 openings per square inch) were purchased from Technical Glass (Painesville Twp., OH), Grainger (Lake Forest, IL), and Gaoke Ceramic Co. LTD (Jiangxi, China), respectively. The chemical bath solution was prepared by dissolving sodium bicarbonate powder in a saturated aqueous solution of aluminum sulfate. Once all NaHCO3 is dissolved, DI water was immediately added to increase the solution pH to 3.8. The solution was filtered through a 0.22 μm filter to remove any residual powder. The filtered solution was transparent with final concentrations of Al2(SO4)3 and NaHCO3 at 80 and 220 mM, respectively. A piece of substrate was first cleaned by sonication sequentially in water, ethanol, and acetone and then blown dry with pure nitrogen gas. The substrate was further cleaned under oxygen plasma for 10 min and then immersed in the chemical bath. The bath was incubated in a 30 °C oven (Fisher Scientific) under vigorous shaking. After a preset period of time, the substrate was taken out of the solution, rinsed with copious water, and blown dry with nitrogen. Finally, the substrate was heated to 750 °C at a heating rate of 1 °C min−1 and held for 30 min in air, leading to the formation of the AOB layer. The magnetite nanoparticles used as the precursor of CNT growth catalyst were synthesized as follows. Two mmol tris(acetylacetonato) iron (99%, Alfa Aesar) was dissolved in 25 mL of triethylene glycol (99%) under sonication. Under the protection of argon, the mixture was slowly heated to reflux (∼280 °C) and kept at reflux for 30 min under vigorous stirring. After the mixture was cooled to the ambient temperature, a black homogeneous colloidal suspension containing magnetite nanoparticles was obtained. The colloidal suspension was diluted 10 times using ethanol. To deposit magnetite nanoparticles, the AOB-coated substrate was first immersed in an aqueous solution containing organic linker polyethylenimine (molecular weight: ∼25 000; concentration: 0.1% by weight)44 for 30 min. The substrate was then rinsed with water, dried in air, and then immersed in the

3. RESULTS We synthesize aligned carbon nanotube arrays on highly curved surfaces using a method involving three main steps as illustrated in Figure 1. The first step is the chemical bath deposition of basic aluminum sulfate.47,48 This step is carried out by immersing a substrate in an acidic solution of aluminum sulfate and sodium bicarbonate for at least 15 min. In this process, BAS crystallizes into a thin film on the surface of the substrate: 3Al 2(SO4 )3 + 10NaHCO3 + 4H 2O → 2(H3O)Al3(SO4 )2 (OH)6 + 5Na 2SO4 + 10CO2

(1)

Second, BAS is converted to γ-Al2O3 by annealing at 750 °C for 1 h, releasing sulfur trioxide gas and water vapor: 2Al3(H3O)(SO4 )2 (OH)6 → 3Al 2O3 + 4SO3 + 9H 2O (2)

Last, the substrate is further decorated with magnetite nanoparticles, which are reduced to iron nanoparticles by hydrogen in CVD and catalyze CNT growth in the presence of ethylene.37 The first two steps together produce the aluminum oxide buffer layer, which is the focus of this study. The catalyzed growth of CNTs using magnetite nanoparticles and CVD is a well-established technique; therefore, the third step is performed under preset standard conditions. In the remainder of this article, we describe our investigation on the mechanism and control of γ-Al2O3 formation and the quality of the subsequently grown CNT arrays. Before applying the CBD-annealing method to fibers, meshes, and foams, we first use a flat silicon chip to validate 7402

DOI: 10.1021/acs.langmuir.5b01002 Langmuir 2015, 31, 7401−7409

Article

Langmuir

Figure 2. Aluminum oxide buffer (AOB) on a silicon chip prepared by the chemical bath deposition (CBD) of basic aluminum sulfate (BAS). (a−c) Atomic force micrographs of the silicon chip, the BAS film, and the AOB film. (d) Powder X-ray diffraction patterns of (i) silicon chip, (ii) BAS film (rhombohedral (H3O)Al3(SO4)2(OH)6; JCPDS 16−409), and (iii) AOB (γ-Al2O3; JCPDS 29−0063), respectively. (e, f) Crystal structures of BAS and γ-Al2O3. Scale bars: 100 nm. Immersion time in the chemical bath: 1.5 h.

the deposition mechanism. Figure 2a−c compares the silicon chip before BAS deposition, after BAS deposition for 1.5 h, and after annealing, respectively, using atomic force microscopy. The AFM images show that a film of nanoparticles is formed on top of the chip’s surface after the chip has been immersed in the chemical bath, consistent with the polycrystalline nature of BAS deposition.47 According to the height of the nanoparticles, the deposition of nanoparticles increases the root-mean-square (RMS) roughness of the surface from 0.5 to 2 nm. After annealing, the nanoparticles shrink with a reduction of RMS roughness from 2 to 1.4 nm, suggesting the formation of γAl2O3 by dehydration. The structural identities of BAS and γAl2O3 are provided by powder X-ray diffraction, as shown in Figure 2d along with the spectrum of the silicon chip. BAS and γ-Al2O3 show distinctively different diffraction patterns from their unique rhombohedral48 and cubic49 phases, as illustrated in Figure 2f and g, respectively. Figure 3 shows an aligned CNT array grown on the AOBcoated silicon chip after being immersed in the chemical bath for 1.5 h. Scanning electron microscopy reveals that the CNT array grows vertically and uniformly, forming a dense forest (Figure 3a). Microscopically, individual CNTs are curled into a repeating wave with an array length of l on the order of 100 nm (Figure 3b). A curviness index is defined to characterize this morphology: φ = lc/l, where lc is the true length of the CNT. For perfectly straight nanotubes, we would expect φ = 1. For the sample shown in Figure 3b, we estimate φ = 1.17(±0.05), suggesting that CNTs are relatively straight. Although aligned CNT arrays grown on substrates have excellent mechanical strength, a few CNTs can be removed by extended sonication for an hour, permitting the examination of the quality of individual CNTs using transmission electron microscopy.45 Little amorphous carbon deposit is visible under TEM (Figure 3c), suggesting that the conversion of ethylene to CNT is highly efficient. High-resolution TEM reveals that CNTs have uniform diameters with a relatively narrow size distribution

Figure 3. Carbon nanotube (CNT) arrays grown on aluminum oxide buffer (AOB) prepared with chemical bath deposition and annealing. (a) Scanning electron micrograph (SEM) of the entire length of the CNT array. (b) SEM showing the organization of the array. (c) Transmission electron micrographs of individual carbon nanotubes. (d) Raman spectrum. Scale bars: a, 100 μm; b, 2 μm; c, 100 nm (inset: 10 nm). Immersion time: 1.5 h.

around 9.4(±1.9) nm and an average wall number of 4.6(±1.7) (insets of Figure 3c). Raman spectroscopy shows that CNTs have a high degree of graphitization with an in-plane graphene crystallite size Lg of 10.7 nm, as estimated using the empirical equation Lg = 8.28(IG/ID) with IG/ID = 1.3 (Figure 3d).50 The critical parameter controlling the outcome of AOB preparation and thus the quality of CNT array is the time for a substrate to be immersed in the chemical bath. To maximize the length of aligned CNT arrays, we vary the immersion time t from 15 min to 3 h and examine AOB thickness τ, the length l 7403

DOI: 10.1021/acs.langmuir.5b01002 Langmuir 2015, 31, 7401−7409

Article

Langmuir of CNT arrays, and curviness index φ. The correlations of these parameters are shown in Figure 4. As t is extended, τ increases

diameter of 9.2 nm and 5 walls, they have a density of 1.53 g 2 54 cm−3 according to ρ = 3.04[n/d − (0.34∑i n−1 = 0 i)/d ]. Using the immersion time of t = 1.5 h, glass fiber, stainless steel mesh, and ceramic foam are prepared by the CBDannealing method for growing aligned CNT arrays. As shown in Figure 5, CNT arrays up to several hundred micrometers in

Figure 4. Optimization of chemical bath deposition (CBD) immersion time for growing aligned carbon nanotube (CNT) arrays. (a) Increase of aluminum oxide buffer (AOB) thickness τ with increasing of immersion time t. (b) Asymptotic increase of CNT array length l with τ. (c) Asymptotic decrease of CNTs’ curviness index φ with τ. (d) Invariance of CNT diameter ϕ with τ. Fit functions: a, eq 1; b and c, exponential (l = 333[1−2.6 exp(−0.13t)], φ = 1.2[1 + exp(−0.15t)], R2 = 0.98); d, average (τ = 9.2(±0.2) nm).

exponentially (Figure 4a). The two parameters show a relationship conforming to the classical Johnson−Mehl− Avrami−Kolmogorov rate law:51−53 τ = τ0[1 − exp(−kt n)]

(3)

where τ0 is the maximal thickness of AOB that the chemical bath can create, k is the rate constant of BAS deposition, n is a dimensional parameter. Least-squares regression gives τ0 = 211(±13) nm, k = 0.063(±0.031) h−1, and n = 1.9(±0.6) (R2 = 0.96). The estimate of k is consistent with a crystallization process sufficiently slow to produce an even layer of BAS. The estimate of n ≈ 2 is consistent with the fact that the crystallization of BAS on a surface is a two-dimensional nucleation process. As the AOB thickness increases, the length and curviness of CNT arrays vary systematically, as shown in Figure 4b,c. l increases exponentially with τ, giving a maximum length of lm = 333(±15) nm. In comparison, φ decreases exponentially with τ, giving a minimal value of φm = 1.14(±0.01). The correlations of l and φ with τ indicate that increasing the thickness of AOB layer promotes the growth of longer and straighter CNTs. At τ = 35 nm, one can expect l = 0.97lm and φ = 1.02φm, indicating that the optimal conditions are sufficiently achieved with an immersion time of t = 1.5 h. Different from length and curviness, the diameter of CNTs exhibits no dependence on AOB thickness. As shown in Figure 4d, the average CNT diameter is estimated at ϕ = 9.2(±0.2) nm, consistent with the diameters of typical multiwalled carbon nanotubes.13 The independence of CNT diameter on AOB thickness confirms that the properties of individual CNTs such as diameter and wall number are controlled by the deposition of iron oxide nanoparticles, which is the same for all AOB layers deposited on silicon chips in our experiments. For CNTs with a

Figure 5. Scanning electron micrographs of vertically aligned carbon nanotube (CNT) arrays grown on (a, b) glass fibers (GFs), (c, d) stainless steel mesh (SSM), and (e, f) porous cordierite monolith (PCM), all of which are coated with aluminum oxide films using the chemical bath deposition method. Insets in a, c, and e are GFs, SSM, and CPM before the CNT growth. Scale bars: a, 200 μm; b, 20 μm; c, 200 μm; d, 20 μm; e, 500 μm; inset of e, 500 mm; f, 5 μm. Time for chemical bath deposition: 1.5 h (resulting buffer thickness, ca. 35 nm).

length are successfully grown on all three substrates using conventional CVD. When glass fibers are used as the supporting substrate, CNTs grow perpendicular to the fiber length and in the radial direction (Figure 5a). After 1.5 h of CVD growth, the length of CNTs can reach 100 μm, approximately 25 times the diameter of the glass fiber support. At this length, CNTs are split into two arrays, with one on each side of the fiber and consisting of approximately half of the total amount of CNTs. Similar to CNTs grown on the silicon chip, the arrays on glass fibers are formed by a high density of wellaligned CNTs (Figure 5b). CNTs grown on an AOB-coated stainless steel mesh are split in 4 arrays, each pointing to an orthogonal direction around a stainless steel wire (Figure 5c,d). The mesh-supported CNTs are only ca. 30 μm, which is 7404

DOI: 10.1021/acs.langmuir.5b01002 Langmuir 2015, 31, 7401−7409

Article

Langmuir

the redistribution of precursor solution and thus an uneven AOB layer after annealing. Different from spinning and dip coating methods that coat substrates with aluminum hydroxide solution before annealing, sol−gel and layer-by-layer assembly use solid precursors. The sol−gel process uses γ-AlOOH (boehmite) nanoparticles as the AOB precursor.55 Typically, the oxyhydroxide gel is first created by the hydrolysis of aluminum alkoxides and then delivered by immersing a substrate into the gel. Similar to dip coating using solution, the imbalance between gravity and capillarity on highly curved surfaces can cause an uneven distribution of the precursor gel before annealing is complete. To prevent the potential interference of gravity, the layer-bylayer assembly method attaches premade γ-AlOOH nanodisks on substrates using organic linkers.37 However, the use of premade nanodisks of tens of nanometers in diameter can lead to cracks in the AOB layer deposited on highly curved surfaces upon annealing. In addition, both gels and suspensions containing γ-AlOOH nanoparticles and nanodisks are difficult to pass through tortuous pores. The only chemical method that is sufficiently versatile to deposit oxide coatings on fibers and foams is atomic layer deposition.40,41 In this method, the substrate is placed in a vacuum chamber and alternately exposed to trimethylaluminum (Al2(CH3)6) and water vapor. The alternating process allows trimethylaluminum first absorbs on the substrate surface and then transforms to aluminum oxide in situ by reacting with water. In spite of the success of ALD, the requirement of vacuum limits its potential for scale-up and the use of pyrophoric trimethylaluminum poses safety concerns. The analyses of the existing AOB preparation methods suggest that a solution-based chemical method is economically and operationally attractive if an AOB precursor can be deposited on curved surfaces by an in situ reaction similar to that in ALD, through which the precursor is firmly attached to the surface while the substrate is still immersed in the precursor solution. Previous studies of AOB precursors have been mostly focused on aluminum (oxy)hydroxide compounds; however, the chemistry of aluminum (oxy)hydroxide is complicated by the formation of stable Keggin cations.56 Attempts to precipitate aluminum (oxy)hydroxide from an hydroxide solution often end up either in failure or without kinetic control.57 A successful deposition of a new AOB precursor can be rationally designed by reconsidering the balancing anion used to initiate precursor crystallization. Because polyaluminum cations are positively charged, a balancing anion is required to facilitate its crystallization from a stable aqueous solution. Recent studies have established that sulfate is an effective balancing anion for crystallizing polyaluminum cations, possibly due to both its divalence according to the Schultz-Hardy rule58 and the unique size of the anion.59 Sulfate has been extensively used in both scientific experiments such as the crystallization of Keggin structures59 as well as engineering applications such as the promotion of coagulation in water treatment.60 BAS is also well recognized as an important intermediate for preparing aluminum oxide particles with controllable sizes and morphologies.47,48 The chemistry of BAS crystallization has not been utilized to create AOB surface coatings until now. 4.2. Length and Curviness of CNTs. In the process of creating AOB by CBD and annealing, we have identified the time that a substrate is immersed in the chemical bath as an important parameter for controlling AOB thickness and thus

approximately half the size of the mesh openings. Compared to glass fiber and stainless steel mesh, CNT arrays grown on a ceramic foam have much reduced long-range organization. The arrays form localized patches of ca. 50 μm in size (Figure 5e). Accompanied with the lack of organization is the reduced length of CNTs, which are only 5 μm long (Figure 5f). The differences in CNT length and morphology suggests that the underlying substrate, in addition to the coated AOB layer, exerts significant controls over CNT growth, which requires investigation in future studies. In spite of the differences, however, the new CBD-annealing method is successful to facilitate the growth of micrometer-long CNT arrays on these substrates with highly curved surfaces, particularly permitting the growth of CNT arrays on porous substrates for the first time without using expensive atomic layer deposition.

4. DISCUSSION The successful growth of CNT arrays on glass fiber, stainless steel mesh, and ceramic foam relies on the chemical bath deposition of basic aluminum sulfate, which is converted to aluminum oxide buffer upon annealing. Compared to existing AOB preparation methods, CBD succeeds in preparing AOB on these structurally challenging substrates because it uses the in situ crystallization of BAS as the mechanism for delivering the AOB precursor. In this section, we discuss the rationale for the design of the CBD-annealing method as well as the possible mechanisms responsible for the formation and morphology of the CNT arrays grown on substrates that have been treated using this method. 4.1. Chemical Bath Deposition. To understand the importance of the in situ crystallization of BAS in the design of the CBD-annealing method, we first summarize the mechanisms for delivering AOB and its precursors used by existing methods. In e-beam evaporation, metallic aluminum or aluminum oxide is evaporated under the bombardment of a beam of high-energy electrons under vacuum. The metal and oxide vapors are then delivered to the surface of a substrate by electrostatic attraction, which requires a straight and clear pathway between the source and the surface and thus is not suitable for substrates with highly curved surfaces and particularly those with surfaces hidden in tortuous pores. In thermal evaporation, the vapor is generated by heating a solid source while the substrate is placed underneath the source to receive the vapor through diffusion. The amount of material deposited at a specific location on the surface is proportional to its distance to the source, which is variable for a substrate with highly curved surfaces. Compared to physical methods such as e-beam and thermal evaporation, solution-based chemical methods operated under the ambient pressure always involve two steps, including the formation of a layer of AOB precursor and the conversion of the precursor to AOB by annealing. Spin coating and dip coating use the aqueous solution of aluminum hydroxide (Al(OH)3) as precursor.38,39 The solution is delivered by either casting or immersion, which do not require a clear pathway to the substrate. However, because the substrate must be separated from the solution before annealing can be performed, a delicate balance between gravity and capillarity is required to maintain an even distribution of the adsorbed solution once the substrate is removed from the solution in order to form an even AOB layer. When the surface is curved, gravity acts on the solution differently at different parts of the surface, leading to 7405

DOI: 10.1021/acs.langmuir.5b01002 Langmuir 2015, 31, 7401−7409

Article

Langmuir the length and curviness of CNT arrays. The positive correlation of AOB thickness and immersion time can be readily understood since the amount of BAS deposition increases with time. The mechanisms through which AOB thickness regulates the length and curviness of CNTs are less obvious. Here, we discuss them in detail with the aid of two physical models. The growth of CNTs by chemical vapor deposition is a catalyzed process. As illustrated in Figure 6a, ethylene reacts at

Figure 7. Formation of a curvy carbon nanotube (CNT) under the influence of inter-CNT attraction. (a) Initial tilted growth under attraction of a neighboring nanotube. (b) Curved growth due to the pinning at the top of the nanotube. (c) Sliding growth after the nanotube is unpinned. (d) Angled growth leading the top of the nanotube to the neighbor on the other side. ls and la are the length and amplitude of half of a bending cycle, respectively. For simplification of presentation, the neighboring nanotubes are depicted as straight tubes, which are also curvy in actual CNT arrays. Colors: gray, substrate; blue, aluminum oxide buffer; orange, iron catalyst.

interaction42,43 from the neighboring nanotube creates frictional resistance in the vertical direction. As a result, the top of the nanotube is pinned in place. Further growth of CNTs produces a new stretch of nanotube, which has to be accommodated by bending (Figure 7b). The bending of CNT generates an elastic force, pushing upward against frictional resistance, which eventually allows the newly grown CNTs slide through the pinning point (Figure 7c). Once passing through the pinning point, the extension of CNTs follows an angle heading toward another neighboring nanotube on the opposite side (Figure 7d). Repeating these steps will create wave-like curviness, as shown in Figure 3b. According to the model depicted in Figure 7d, the curviness index, defined as the ratio of the length of individual CNTs to the length of CNT array, can be computed using parameters within half of a wave cycle. Here, φ = 2la/ls, where la is the amplitude and ls is half of the wavelength. The scaling factor 2 is obtained by approximating the length of curvy CNTs using the sinusoidal function. In this equation, la is approximately the distance between catalyst nanoparticles, which is controlled by their surface density; therefore, la is constant in our experiments. ls is a function of CNT growth rate, which increases with the increase of AOB thickness. As a result, φ decreases as τ increases, as observed in Figure 4c. 4.3. Splitting of CNT Arrays. The splitting of CNT arrays into six groups has been reported previously for spherical substrates, each pointing to one of the six orthogonal directions in three-dimensional Cartesian coordinate system.64,65 This is consistent with the splitting of CNT arrays into 4 orthogonal groups on the wires of stainless steel mesh, as shown in Figure 5c,d, which may be considered as two-dimensional cylindrical substrates. The splitting of CNT arrays is likely a result of balancing inter-CNT attraction and the resistance originated from CNT’s rigidity.66 In the absence of inter-CNT attraction, the growth of CNTs should follow the radial directions around a cylindrical substrate, as shown in Figure 8a. For relatively short CNTs, we have neglected the influence of gravity by considering CNTs are rigid nanotubes. When inter-CNT attraction is considered, CNTs will group with close neighbors. The splitting into 4 groups produces the least stress for CNTs at the outer boundaries of each group, as shown in Figure 8b. The increase of CNT length permits longer deflection under the same rigidity, leading to bundling of CNTs from different

Figure 6. Controlling factors of carbon nanotube growth rate. (a) Rate-limiting supply of precursor(s). (b) Diffusion of iron catalyst through aluminum oxide buffer (AOB) and its prevention by increasing AOB thickness. (c) Roughness-controlled barrier for nanoparticle coalescence. Colors: gray, substrate; blue, AOB layer; orange, iron catalyst.

the surface of iron nanoparticles to provide carbon precursor(s), which dissolves into the melted nanoparticles. The continuous supply of precursor(s) creates supersaturation in the liquid nanoparticles, resulting in the continuous growth of crystalline carbon in the form of multiwalled CNTs. The rate of CNT growth is regulated by the rate of precursor formation at the nanoparticle surface, which is proportional to the area of exposed nanoparticle surface. The surface area of catalyst nanoparticles can be reduced by the diffusion of iron through AOB into the substrate and the coalescence of nanoparticles. As shown in Figure 6b, increasing AOB thickness can greatly reduce the loss of iron by diffusion. As shown in Figure 6c, increasing AOB thickness can also increase its surface roughness (error bars in Figure 4a; cf. Figure S1 in the Supporting Information) and thus the barrier for coalescence.61 Eventually, an AOB layer sufficiently thick and rough can prevent diffusion and coalescence nearly completely, leaving most of the catalyst surface exposed. As a result, the growth rate of CNT reaches a maximum value. Both the increase of CNT growth rate with increasing AOB thickness and its eventual stabilization are observed our experiments, as shown in Figure 4b. The minimum immersion time required to reach stabilization of 1.5 h is selected in our experiments to prepare glass fibers, stainless steel mesh, and ceramic foam. The formation of curviness in CNT arrays is believed to be a result of the attraction exerted by neighboring nanotubes although the detailed mechanism remains to be clarified.62,63 Here, we propose a four-step mechanism based on this idea. As illustrated in Figure 7, the initial growth of a nanotube from a catalyst nanoparticle can be considered to be straight until the top of the nanotube touches a neighboring nanotube due to inter-CNT attraction or the tilt of underlying surface (Figure 7a). The attraction exerted by van der Waals force and π−π 7406

DOI: 10.1021/acs.langmuir.5b01002 Langmuir 2015, 31, 7401−7409

Article

Langmuir

new BAS-CBD method may provide a cost-effective alternative for electronic applications.

■

ASSOCIATED CONTENT

S Supporting Information *

Figure S1: Increase of surface roughness with the increase of the thickness of aluminum oxide buffer layer. The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.langmuir.5b01002.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank the USDOE Office of Nuclear Energy’s Nuclear Energy University Program, the U.S. National Science Foundation Environmental Engineering Program, and the Notre Dame Sustainable Energy Initiative for generous financial support. We thank Prof. Xiuling Jiao and Dr. Yuguo Xia for providing the crystallographic information file of basic aluminum hydroxide.

Figure 8. Splitting of carbon nanotube arrays as a balance of internanotube attraction and nanotube rigidity.

groups. As shown in Figure 5a and 8c, only two groups exist for the long CNT arrays grown on glass fibers. The relationship between CNT length and the number of array groups on fibers has been investigated previously by other researchers.43

■

5. CONCLUSIONS We have successfully created AOB layers on structurally challenging substrates using chemical bath deposition of basic aluminum sulfate and annealing. The CBD-annealing method can provide new opportunities for growing aligned CNT arrays on substrates with curved surfaces such as fibers, meshes, and foams. These substrates are important materials used to strengthen structures, build electronic devices, and support catalyst nanoparticles.67−69 Although these materials have relatively large geometric surface areas, even greater surface areas are preferred to improve physical contact, facilitate electron collection, and increase catalyst loading.70,71 Conductive CNTs with high specific surface areas are ideal materials for adding surface areas to engineering materials without adding much weight. CNTs can also provide the engineering materials with new functionalities with their excellent electronic, absorptive, and mechanical properties.72−74 CNT arrays grown on stainless steel mesh and ceramic foam have already been investigated as advanced electrodes, filters, and catalyst supports.44,45,70,75 Further understanding of the formation, property, and morphology of CNT arrays grown on these substrates may lead to new insights for better incorporating CNTs into hierarchical and composite structures used in future applications. In addition to preparing substrates for CNT growth, the aluminum oxide layer created by CBD and annealing can be directly used as the corrosion resistance barrier for alloys,76,77 the electrical insulator in microelectronics,78,79 and the passivation layer for organic electronic devices and solar cells.80,81 Traditionally, aluminum oxide layers for corrosion protection are created by dip coating where the control of thickness is not stringent.76 Similar to the case of CNT growth, the use of BAS-CBD may greatly simplify the coating process for substrates with curved surfaces. The creation of pinhole-free aluminum oxide layers with a thickness of a few nanometers requires the use of the expensive atomic layer deposition. The

REFERENCES

(1) Cho, W.; Schulz, M.; Shanov, V. Growth and characterization of vertically aligned centimeter long CNT arrays. Carbon 2014, 72, 264− 273. (2) Wu, Z.; Chen, Z.; Du, X.; Logan, J. M.; Sippel, J.; Nikolou, M.; Kamaras, K.; Reynolds, J. R.; Tanner, D. B.; Hebard, A. F.; Rinzler, A. G. Transparent, conductive carbon nanotube films. Science 2004, 305, 1273−1276. (3) Behabtu, N.; Young, C. C.; Tsentalovich, D. E.; Kleinerman, O.; Wang, X.; Ma, A. W. K.; Bengio, E. A.; ter Waarbeek, R. F.; de Jong, J. J.; Hoogerwerf, R. E.; Fairchild, S. B.; Ferguson, J. B.; Maruyama, B.; Kono, J.; Talmon, Y.; Cohen, Y.; Otto, M. J.; Pasquali, M. Strong, light, multifunctional fibers of carbon nanotubes with ultrahigh conductivity. Science 2013, 339, 182−186. (4) Dai, H.; Wong, E. W.; Lieber, C. M. Probing electrical transport in nanomaterials: Conductivity of individual carbon nanotubes. Science 1996, 272, 523−526. (5) Berber, S.; Kwon, Y.-K.; Tomanek, D. Unusually high thermal conductivity of carbon nanotubes. Phys. Rev. Lett. 2000, 84, 4613. (6) Che, J.; Cagin, T.; Goddard, W. A., III Thermal conductivity of carbon nanotubes. Nanotechnology 2000, 11, 65. (7) Yu, M.-F.; Lourie, O.; Dyer, M. J.; Moloni, K.; Kelly, T. F.; Ruoff, R. S. Strength and breaking mechanism of multiwalled carbon nanotubes under tensile load. Science 2000, 287, 637−640. (8) Demczyk, B.; Wang, Y.; Cumings, J.; Hetman, M.; Han, W.; Zettl, A.; Ritchie, R. Direct mechanical measurement of the tensile strength and elastic modulus of multiwalled carbon nanotubes. Mater. Sci. Eng., A 2002, 334, 173−178. (9) Zhang, M.; Atkinson, K. R.; Baughman, R. H. Multifunctional carbon nanotube yarns by downsizing an ancient technology. Science 2004, 306, 1358−1361. (10) Pan, H.; Feng, Y. P. Semiconductor nanowires and nanotubes: Effects of size and surface-to-volume ratio. ACS Nano 2008, 2, 2410− 2414. (11) Baughman, R. H.; Zakhidov, A. A.; de Heer, W. A. Carbon nanotubes - The route toward applications. Science 2002, 297, 787− 792. (12) Chen, H.; Roy, A.; Baek, J.-B.; Zhu, L.; Qu, J.; Dai, L. Controlled growth and modification of vertically-aligned carbon nanotubes for multifunctional applications. Mater. Sci. Eng. R 2010, 70, 63−91.

7407

DOI: 10.1021/acs.langmuir.5b01002 Langmuir 2015, 31, 7401−7409

Article

Langmuir

(33) Zhang, H.; Cao, G.; Wang, Z.; Yang, Y.; Shi, Z.; Gu, Z. Influence of ethylene and hydrogen flow rates on the wall number, crystallinity, and length of millimeter-long carbon nanotube array. J. Phys. Chem. C 2008, 112, 12706−12709. (34) Yun, Y.; Shanov, V.; Tu, Y.; Subramaniam, S.; Schulz, M. J. Growth mechanism of long aligned multiwall carbon nanotube arrays by water-assisted chemical vapor deposition. J. Phys. Chem. B 2006, 110, 23920−23925. (35) Kim, Y.; Song, W.; Lee, S. Y.; Shrestha, S.; Jeon, C.; Choi, W. C.; Kim, M.; Park, C.-Y. Growth of millimeter-scale vertically aligned carbon nanotubes by microwave plasma chemical vapor deposition. Jpn. J. Appl. Phys. 2010, 49, 085101. (36) Li, J.; Srivastava, S.; Ok, J. G.; Zhang, Y.; Bedewy, M.; Kotov, N. A.; Hart, A. J. Multidirectional hierarchical nanocomposites made by carbon nanotube growth within layer-by-layer-assembled films. Chem. Mater. 2011, 23, 1023−1031. (37) Wang, H.; Na, C. Synthesis of millimeter-long vertically aligned carbon nanotube arrays on aluminum oxide buffer prepared by layerby-layer assembly of boehmite nanoplates. Carbon 2014, 66, 727−729. (38) Alvarez, N. T.; Hamilton, C. E.; Pint, C. L.; Orbaek, A.; Yao, J.; Frosinini, A. L.; Barron, A. R.; Tour, J. M.; Hauge, R. H. Wet catalystsupport films for production of vertically aligned carbon nanotubes. ACS Appl. Mater. Interfaces 2010, 2, 1851−1856. (39) Dörfler, S.; Meier, A.; Thieme, S.; Németh, P.; Althues, H.; Kaskel, S. Wet-chemical catalyst deposition for scalable synthesis of vertical aligned carbon nanotubes on metal substrates. Chem. Phys. Lett. 2011, 511, 288−293. (40) Herrmann, C. F.; Fabreguette, F. H.; Finch, D. S.; Geiss, R.; George, S. M. Multilayer and functional coatings on carbon nanotubes using atomic layer deposition. Appl. Phys. Lett. 2005, 87, 123110. (41) Brieland-Shoultz, A.; Tawfick, S.; Park, S. J.; Bedewy, M.; Maschmann, M. R.; Baur, J. W.; Hart, A. J. Scaling the stiffness, strength, and toughness of ceramic-coated nanotube foams into the structural regime. Adv. Funct. Mater. 2014, 24, 5728−5735. (42) Ci, L. J.; Zhao, Z. G.; Bai, J. B. Direct growth of carbon nanotubes on the surface of ceramic fibers. Carbon 2005, 43, 883−886. (43) Yamamoto, N.; John Hart, A.; Garcia, E. J.; Wicks, S. S.; Duong, H. M.; Slocum, A. H.; Wardle, B. L. High-yield growth and morphology control of aligned carbon nanotubes on ceramic fibers for multifunctional enhancement of structural composites. Carbon 2009, 47, 551−560. (44) Wang, H.; Dong, Z.; Na, C. Hierarchical carbon nanotube membrane-supported gold nanoparticles for rapid catalytic reduction of p-nitrophenol. ACS Sustainable Chem. Eng. 2013, 1, 746−752. (45) Wang, H.; Na, C. Binder-free carbon nanotube electrode for electrochemical removal of chromium. ACS Appl. Mater. Interfaces 2014, 6, 20309−20316. (46) Hodes, G. Semiconductor and ceramic nanoparticle films deposited by chemical bath deposition. Phys. Chem. Chem. Phys. 2007, 9, 2181−2196. (47) Bhattacharya, I. N.; Gochhayat, P. K.; Mukherjee, P. S.; Paul, S.; Mitra, P. K. Thermal decomposition of precipitated low bulk density basic aluminium sulfate. Mater. Sci. Eng. 2004, 88, 32−40. (48) Xia, Y.; Chen, B.; Jiao, X.; Chen, D. Large-scale synthesis and formation mechanism study of basic aluminium sulfate microcubic crystals. Phys. Chem. Chem. Phys. 2014, 16, 5866−5874. (49) Verwey, E. J. W. The structure of the electrolytical oxide layer on aluminium. Z. Kristallogr. 1935, 91, 317−320. (50) Delhaes, P.; Couzi, M.; Trinquecoste, M.; Dentzer, J.; Hamidou, H.; Vix-Guterl, C. A comparison between Raman spectroscopy and surface characterizations of multiwall carbon nanotubes. Carbon 2006, 44, 3005−3013. (51) Avrami, M. Kinetics of phase change. I General theory. J. Chem. Phys. 1939, 7, 1103−1112. (52) Avrami, M. Kinetics of phase change. II Transformation-time relations for random distribution of nuclei. J. Chem. Phys. 1940, 8, 212−224. (53) Avrami, M. Granulation, phase change, and microstructure kinetics of phase change. III. J. Chem. Phys. 1941, 9, 177−184.

(13) Wang, H.; Ma, H.; Zheng, W.; An, D.; Na, C. Multifunctional and recollectable carbon nanotube ponytails for water purification. ACS Appl. Mater. Interfaces 2014, 6, 9426−9434. (14) Li, W. Z.; Xie, S. S.; Qian, L. X.; Chang, B. H.; Zou, B. S.; Zhou, W. Y.; Zhao, R. A.; Wang, G. Large-scale synthesis of aligned carbon nanotubes. Science 1996, 274, 1701−1703. (15) Hata, K.; Futaba, D. N.; Mizuno, K.; Namai, T.; Yumura, M.; Iijima, S. Water-assisted highly efficient synthesis of impurity-free single-walled carbon nanotubes. Science 2004, 306, 1362−1364. (16) Yasuda, S.; Futaba, D. N.; Yamada, T.; Satou, J.; Shibuya, A.; Takai, H.; Arakawa, K.; Yumura, M.; Hata, K. Improved and large area single-walled carbon nanotube forest growth by controlling the gas flow direction. ACS Nano 2009, 3, 4164−4170. (17) Kaneko, A.; Yamada, K.; Kumahara, R.; Kato, H.; Homma, Y. Comparative study of catalytic activity of iron and cobalt for growing carbon nanotubes on alumina and silicon oxide. J. Phys. Chem. C 2012, 116, 26060−26065. (18) Noda, S.; Hasegawa, K.; Sugime, H.; Kakehi, K.; Zhang, Z.; Maruyama, S.; Yamaguchi, Y. Millimeter-thick single-walled carbon nanotube forests: Hidden role of catalyst support. Jpn. J. Appl. Phys. 2007, 46, L399. (19) de Los Arcos, T.; Gunnar Garnier, M.; Oelhafen, P.; Mathys, D.; Won Seo, J.; Domingo, C.; Vicente García-Ramos, J.; Sánchez-Cortés, S. Strong influence of buffer layer type on carbon nanotube characteristics. Carbon 2004, 42, 187−190. (20) Yamada, T.; Namai, T.; Hata, K.; Futaba, D. N.; Mizuno, K.; Fan, J.; Yudasaka, M.; Yumura, M.; Iijima, S. Size-selective growth of double-walled carbon nanotube forests from engineered iron catalysts. Nat. Nano. 2006, 1, 131−136. (21) de los Arcos, T.; Gunnar Garnier, M.; Oelhafen, P.; Mathys, D.; Won Seo, J.; Domingo, C.; Vicente García-Ramos, J.; Sánchez-Cortés, S. Strong influence of buffer layer type on carbon nanotube characteristics. Carbon 2004, 42, 187−190. (22) Magrez, A.; Smajda, R.; Seo, J. W.; Horváth, E.; Ribic̆, P. R.; Andresen, J. C.; Acquaviva, D.; Olariu, A.; Laurenczy, G.; Forró, L. Striking influence of the catalyst support and its acid−base properties: New insight into the growth mechanism of carbon nanotubes. ACS Nano 2011, 5, 3428−3437. (23) Magrez, A.; Seo, J. W.; Smajda, R.; Korbely, B.; Andresen, J. C.; Mionić, M.; Casimirius, S.; Forró, L. Low-temperature, highly efficient growth of carbon nanotubes on functional materials by an oxidative dehydrogenation reaction. ACS Nano 2010, 4, 3702−3708. (24) Noda, S.; Hasegawa, K.; Sugime, H.; Kakehi, K.; Zhang, Z.; Maruyama, S.; Yamaguchi, Y. Millimeter-thick single-walled carbon nanotube forests: Hidden role of catalyst support. Jpn. J. Appl. Phys. 2007, 46, L399. (25) Noda, S.; Tsuji, Y.; Murakami, Y.; Maruyama, S. Combinatorial method to prepare metal nanoparticles that catalyze the growth of single-walled carbon nanotubes. Appl. Phys. Lett. 2005, 86, 173106. (26) Nessim, G. D. Properties, synthesis, and growth mechanisms of carbon nanotubes with special focus on thermal chemical vapor deposition. Nanoscale 2010, 2, 1306−1323. (27) Melechko, A. V.; Merkulov, V. I.; McKnight, T. E.; Guillorn, M. A.; Klein, K. L.; Lowndes, D. H.; Simpson, M. L. Vertically aligned carbon nanofibers and related structures: Controlled synthesis and directed assembly. J. Appl. Phys. 2005, 97, 041301. (28) See ref 24. (29) Hasegawa, K.; Noda, S. Millimeter-tall single-walled carbon nanotubes rapidly grown with and without water. ACS Nano 2011, 5, 975−984. (30) Xiong, G.-Y.; Wang, D. Z.; Ren, Z. F. Aligned millimeter-long carbon nanotube arrays grown on single crystal magnesia. Carbon 2006, 44, 969−973. (31) Patole, S. P.; Alegaonkar, P. S.; Lee, H.-C.; Yoo, J.-B. Optimization of water assisted chemical vapor deposition parameters for super growth of carbon nanotubes. Carbon 2008, 46, 1987−1993. (32) Jayasinghe, C.; Chakrabarti, S.; Schulz, M. J.; Shanov, V. Spinning yarn from long carbon nanotube arrays. J. Mater. Res. 2011, 26, 645−651. 7408

DOI: 10.1021/acs.langmuir.5b01002 Langmuir 2015, 31, 7401−7409

Article

Langmuir (54) Laurent, C.; Flahaut, E.; Peigney, A. The weight and density of carbon nanotubes versus the number of walls and diameter. Carbon 2010, 48, 2994−2996. (55) Zhang, X.; Wu, Y.; Liu, G.; He, S.; Yang, D. Investigation on sol−gel boehmite-AlOOH films on Kapton and their erosion resistance to atomic oxygen. Thin Solid Films 2008, 516, 5020−5026. (56) Kloprogge, J. T.; Frost, R. L. Raman microscopy study of basic aluminum sulfate. J. Mater. Sci. 1999, 34, 4199−4202. (57) Vioux, A. Nonhydrolytic sol−gel routes to oxides. Chem. Mater. 1997, 9, 2292−2299. (58) Waite, T. D.; Cleaver, J. K.; Beattie, J. K. Aggregation kinetics and fractal structure of γ-alumina assemblages. J. Colloid Interface Sci. 2001, 241, 333−339. (59) Abeysinghe, S.; Unruh, D. K.; Forbes, T. Z. Crystallization of Keggin-type polyaluminum species by supramolecular interactions with disulfonate anions. Cryst. Growth Des. 2012, 12, 2044−2051. (60) Jiang, J.-Q. The role of coagulation in water treatment. Curr. Opin. Chem. Eng. 2015, 8, 36−44. (61) Burt, D. P.; Whyte, W. M.; Weaver, J. M. R.; Glidle, A.; Edgeworth, J. P.; Macpherson, J. V.; Dobson, P. S. Effects of metal underlayer grain size on carbon nanotube growth. J. Phys. Chem. C 2009, 113, 15133−15139. (62) Vinten, P.; Bond, J.; Marshall, P.; Lefebvre, J.; Finnie, P. Origin of periodic rippling during chemical vapor deposition growth of carbon nanotube forests. Carbon 2011, 49, 4972−4981. (63) Cui, X.; Wei, W.; Harrower, C.; Chen, W. Effect of catalyst particle interspacing on the growth of millimeter-scale carbon nanotube arrays by catalytic chemical vapor deposition. Carbon 2009, 47, 3441−3451. (64) He, D.; Li, H.; Li, W.; Haghi-Ashtiani, P.; Lejay, P.; Bai, J. Growth of carbon nanotubes in six orthogonal directions on spherical alumina microparticles. Carbon 2011, 49, 2273−2286. (65) He, D.; Bai, J. Acetylene-enhanced growth of carbon nanotubes on ceramic microparticles for multi-scale hybrid structures. Chem. Vap. Deposition 2011, 17, 98−106. (66) He, D.; Bozlar, M.; Genestoux, M.; Bai, J. Diameter- and lengthdependent self-organizations of multi-walled carbon nanotubes on spherical alumina microparticles. Carbon 2010, 48, 1159−1170. (67) Martínez-Hansen, V.; Latorre, N.; Royo, C.; Romeo, E.; GarcíaBordejé, E.; Monzón, A. Development of aligned carbon nanotubes layers over stainless steel mesh monoliths. Catal. Today 2009, 147 (Supplement), S71−S75. (68) Gamze Karsli, N.; Yesil, S.; Aytac, A. Effect of hybrid carbon nanotube/short glass fiber reinforcement on the properties of polypropylene composites. Compos. Part B 2014, 63, 154−160. (69) Liu, L.; Ma, P.-C.; Xu, M.; Khan, S. U.; Kim, J.-K. Strainsensitive Raman spectroscopy and electrical resistance of carbon nanotube-coated glass fibre sensors. Compos. Sci. Technol. 2012, 72, 1548−1555. (70) Wenmakers, P. W. A. M.; van der Schaaf, J.; Kuster, B. F. M.; Schouten, J. C. ″Hairy Foam″: carbon nanofibers grown on solid carbon foam. A fully accessible, high surface area, graphitic catalyst support. J. Mater. Chem. 2008, 18, 2426−2436. (71) Ishigami, N.; Ago, H.; Motoyama, Y.; Takasaki, M.; Shinagawa, M.; Takahashi, K.; Ikuta, T.; Tsuji, M. Microreactor utilizing a vertically-aligned carbon nanotube array grown inside the channels. Chem. Commun. 2007, 1626−1628. (72) Lin, C.-L.; Chen, C.-F.; Shi, S.-C. Field emission properties of aligned carbon nanotubes grown on stainless steel using CH4/CO2 reactant gas. Diamond Relat. Mater. 2004, 13, 1026−1031. (73) Parham, H.; Bates, S.; Xia, Y.; Zhu, Y. A highly efficient and versatile carbon nanotube/ceramic composite filter. Carbon 2013, 54, 215−223. (74) Boroujeni, A. Y.; Tehrani, M.; Nelson, A. J.; Al-Haik, M. Hybrid carbon nanotube−carbon fiber composites with improved in-plane mechanical properties. Compos. Part B 2014, 66, 475−483. (75) Lee, C. H.; Johnson, N.; Drelich, J.; Yap, Y. K. The performance of superhydrophobic and superoleophilic carbon nanotube meshes in water−oil filtration. Carbon 2011, 49, 669−676.

(76) Wang, D.; Bierwagen, G. P. Sol−gel coatings on metals for corrosion protection. Prog. Org. Coat. 2009, 64, 327−338. (77) Zhong, X.; Li, Q.; Chen, B.; Wang, J.; Hu, J.; Hu, W. Effect of sintering temperature on corrosion properties of sol−gel based Al2O3 coatings on pre-treated AZ91D magnesium alloy. Corros. Sci. 2009, 51, 2950−2958. (78) Wilk, G. D.; Wallace, R. M.; Anthony, J. M. High-κ gate dielectrics: Current status and materials properties considerations. J. Appl. Phys. 2001, 89, 5243−5275. (79) Chae, S. H.; Yu, W. J.; Bae, J. J.; Duong, D. L.; Perello, D.; Jeong, H. Y.; Ta, Q. H.; Ly, T. H.; Vu, Q. A.; Yun, M.; Duan, X.; Lee, Y. H. Transferred wrinkled Al2O3 for highly stretchable and transparent graphene−carbon nanotube transistors. Nat. Mater. 2013, 12, 403−409. (80) Kim, L. H.; Kim, K.; Park, S.; Jeong, Y. J.; Kim, H.; Chung, D. S.; Kim, S. H.; Park, C. E. Al2O3/TiO2 nanolaminate thin film encapsulation for organic thin film transistors via plasma-enhanced atomic layer deposition. ACS Appl. Mater. Interfaces 2014, 6, 6731− 6738. (81) Schmidt, J.; Merkle, A.; Brendel, R.; Hoex, B.; de Sanden, M. C. M. v.; Kessels, W. M. M. Surface passivation of high-efficiency silicon solar cells by atomic-layer-deposited Al2O3. Prog. Photovoltaics 2008, 16, 461−466.

7409

DOI: 10.1021/acs.langmuir.5b01002 Langmuir 2015, 31, 7401−7409