Controlled Release from Core-Shell Nanoporous Silica Particles for Corrosion Inhibition of Aluminum Alloys

1 Department of Chemical and Nuclear Engineering and Center for Micro-Engineered Materials, University of New Mexico, Albuquerque, NM 87106, USA 2 Aerosol and Respiratory Dosimetry Program, Lovelace Respiratory Research Institute, 2425 Ridgecrest Dr. SE, Albuquerque, NM 87108, USA 3 Transmission Electron Microscopy Labs, The University of New Mexico, Albuquerque, NM 87131, USA 4 SCB Innovation Accelerator, Dow Corning Corporation, 2200 W. Salzburg Road, P.O. Box 0994, Midland, MI 48686, USA 5 Department of Geoscience, University of Wisconsin-Madison, Madison, WI 53706, USA 6 Advanced Material Laboratory, Sandia National Laboratories, Albuquerque, NM 87185, USA


Introduction
Controlled release is of particular interest for a broad spectrum of studies dealing with corrosion inhibition, fertilizers, and medical applications such as gene delivery, tissue engineering, and diagnostic agents.Long-term cost-effective corrosion inhibition of steels and alloys requires a controlled release of corrosion inhibitors especially for protecting the aluminum alloy AA2024-T3 used in aeronautic applications [1].AA 2024 is composed of (in weight percentage) 93.5% Aluminum (Al), 4.4% Copper (Cu), 0.6% Manganese (Mn), and 1.5% Magnesium (Mg).Currently, carcinogenic chromate compounds have been widely used as corrosion inhibitors [2,3], but these are expensive to process and maintain.The most potent and less toxic candidate inhibitors for aluminum alloys, according to the study by Scully's group and Tailor's group, are lanthanide, metavanadate, molybdenum, and cobalt ions [1,4].The inhibitors not only suppress cathodic reactions in the alloys that are essential for corrosion inhibition and copper replating, but they provide self-healing for the alloys as well.Much effort has been made to investigate corrosion inhibition, ranging from barrier to conversion coatings or pigmentation of the inhibitors [5][6][7][8].Layered clays have been tried for encapsulating the inhibitors by ion exchange [9].A multifunctional organic coating containing silica can be formed by the in situ sol-gel process.However, this method lacks appropriate storage and slow release of corrosion inhibitors [10][11][12][13].Nanoporous silica is a mechanically and thermally stable, cheap, and an environmentally benign ceramic, and it has found various applications in adsorption, separation, and catalysis [14,15].The well-controlled pore size, narrow pore size distribution, and controllable functionality and pore surface chemistries of self-assembled nanoporous silica materials suggest their use as the encapsulating material for controlled release of corrosion inhibitors.The promising nanocapsules can be 50 50 nm dispersed as pigments into a protective epoxy coating for a long-term cost-effective protection of the alloys.
In this paper, we report a novel single-step method for encapsulating cerium (Ce) corrosion inhibitors into nanoporous silica particles by aerosol-assisted evaporationinduced self-assembly (EISA).Nanostructured particles with a core of Ce compounds and a shell of ordered nanoporous silica were prepared.These core-shell nanostructured particles are effective for inhibiting corrosion of aluminum alloy AA2024-T3.We investigated the controlled release of inhibitors and found that long-term corrosion inhibition can be realized using hydrophobic microporous nanocapsules.

Results and Discussion
The core-shell nanostructured particles were synthesized via aerosol-assisted EISA using cetyltrimethylammonium bromide (CTAB) as a structure-directing agent.Figure 1 shows representative transmission electron microscopy (TEM) images of calcined silica particles.At low Ce loading the particles keep a hexagonally ordered mesoporous silica structure featuring parallel aligned stripes and hexagonally arranged pores (Figure 1(a)), reflecting differing orientations of a 2D hexagonal silica mesophase confined within a spherical shell.By increasing the loading of cerium chloride (CeCl 3 ), the mesostructure turns into a core-shell structure (Figures 1(b) and 1(c)).At high loading (Figure 1(d)), the particles became worm-like or even less ordered, especially near the particle center.
Core-shell structured sample c has been characterized by X-ray diffraction (XRD) and nitrogen isothermal sorption.The nitrogen sorption isotherm of the nanostructured core-shell Ce/silica particles shown in Figure 2 was measured using liquid nitrogen at 77 K and yielded a type IV isotherm with a very narrow hysteresis loop, typical of hexagonal MCM-41-like mesoporous silicas.The pore size was determined to be 1.95 nm with a narrow distribution using   the Barrett-Joyner-Halenda (BJH) model.The Brunauer-Emmett-Teller (BET) surface area is 615 m 2 /g.The low-angle peaks are associated with the 2D hexagonal silica mesophase.
As can be seen from the XRD patterns shown in Figure 3, compared to pure mesoporous silica the mesostructural order for sample c (Figure 1) is reduced as measured by less intense diffraction peaks.
The TEM image and the energy-filtered elemental TEM maps of Ce and silica for particle sample c are shown in Figure 4 and clearly reveal that Ce is located mainly at the core.The core is composed of mostly Ce compounds combined with a little amount of silica.The surrounding crust contains less Ce, and it is dominantly composed of silica.
A high resolution TEM image (Figure 5) for the core of a nanostructured particle shows that the core is composed of multiple crystallites ∼3 nm in size.The inset selected area electron diffraction pattern (SAED) evidences the nanocrystalline cubic CeO 2 structure.CeCl 3 is easily hydrolyzed and further oxidized into Ce dioxide at high temperature.Figure 6 shows the energy dispersive X-ray (EDX) spectrum for the particle sample, indicating that most of the CeCl 3 has been hydrolyzed and the Cl/Ce ratio is much less than the stoichiometric value of 3. The existence of various forms of Ce compounds and the cocondensation of hydrolyzed Ce species with silica acid limit the growth rate of crystallites and favor the formation of multicrystals instead of single crystals within the particles.Hence, the particle samples showed different colors, varying from white, purple, to yellow, with the increasing oxidation.As shown in Figure 7, the electron paramagnetic resonance experiment (EPR) for sample c (Figure 1) proved that only 32% of the Ce is in the state of Ce 3+ .Care should be taken to keep the system from air during aerosol generation and succeeding calcination.
The evaporation of solvents, condensation kinetics, and the mass and heat transport together determine the final structure and the morphology.To better understand the formation process of the core-shell structure, the HCl concentration in precursor sol was intentionally increased 10 times for sample c, and the as-synthesized particles have an inner hollow structure composed of Ce compounds as shown in Figure 8.The reduced Ce salt solubility and quicker condensation/gelation due to accelerated hydrolysis and condensation under more acidic condition impede the ions' diffusion and seize the formed Ce salt crystallite before the ions diffuse inward along with the receding aqueous phase (effective aqueous diffusion coefficient for CeCl 3 in CTAB-templated mesoporous silica is estimated at ∼1.0 × 10 −14 m 2 /s at 25 • C).For comparison, a silica-free dilute ethanol solution of Ce(NO 3 ) 3 •6H 2 O with a molar ratio of 664 EtOH : 1 Ce(NO 3 ) 3 •6H 2 O was also sprayed into the particle-generating system as the precursor sol.Even though no silica species existed, the much lower solubility of Ce(NO 3 ) 3 •6H 2 O in ethanol than in water and the much quicker evaporation of the more volatile ethanol in the absence of a silica network accounted for the early nucleation and succeeding crystallization.Decomposition of the salt and Oswald ripening during aerosol particle generation and succeeding calcination rendered the hollow particles spherical and uniform in thickness (Figure 8(a)).The hollow spherical particle size and the thickness depend on the precursor composition, salt type, and the solvent evaporation rate.Further calcination of the particles makes hollow spherical Ce oxide or Ce dioxide, depending on the calcination environment.A less volatile good solvent and lower heating temperature help form smaller hollow spherical particles of a thicker Ce species layer or even form solid Ce species particles.Figure 9 shows a snapshot of an aerosol particle collected at the end of a 200 • C heating glass tube by a cryofinger kept at 77 K.With the evaporation of the solvents and a growing silica/surfactant gel layer, liquid sol recedes toward the droplet's center, and the silica/surfactant crystal phase starts the EISA into the ordered mesostructures.
A mechanism for the formation of the core-shell nanostructure is proposed here.Starting from an homogeneous solution with the initial CTAB concentration much less than critical micelle concentration (CMC), the evaporation of solvents within each droplet creates growing radial gradients in surfactant, solvent, and salt concentrations that keep maximum concentrations for the surfactant and Ce salt at the droplet surface, resulting in an inward diffusion of the salt and surfactant and outward diffusion of solvents.The relatively lower mobility of surfactant due to the high molecular weight and the strong interaction (electrostatic and/or hydrogen bonding) with the silica network causes an inward drift of Ce species relative to the silica/surfactant phase of the droplet's core along with the receding aqueous phase.CMC is first exceeded at the droplet surface after a certain evaporation time.Once that happens, the CMC location continuously moves to the core of the droplet throughout the course of the EISA process.The surfactant enrichment induces silica-surfactant self-assembly into  state.The evaporation and resulting inward moving liquidvapor interface boundary concentrate the salt solution.
The partitioning of CeCl 3 in the CTAB-templated liquidcrystalline state is quite lower than that in the aqueous phase.As evaporation proceeds, part of the Ce salt is hydrolyzed into a less soluble Ce hydroxide due to the increased temperature and steep temperature gradient.Hydrolyzed Ce species may take part in the cocondensation along with the silica species.have a higher Ce/Cl ratio, and the ratio is lower at the center than at the crust.At the final period of evaporation, the vapor-liquid boundary shifts to the core.With the increasing temperature and salt concentration, Ce salt nucleation and crystal growth continues and causes phase separation of the mixture.Bigger Ce chloride/Ce hydroxide crystallites are formed at the center.The crystallites distort the assembled liquid-crystal phase and make the mesostructured silica less ordered at the center.solution.As shown in Scheme 1, the particle has a radius of r p and the outer boundary of the solution surrounding the particle is r L .The concentration of solute in the particle and solution are C p and C s , respectively.The general diffusion equation ( 1) in spherical coordinates was used to begin the evaluation process:

Mathematical Modeling on
( A numerical method is used to evaluate the effective diffusion coefficient, D eff , by comparing the calculated C s -t curve with the measured release profiles based on initial conditions (2) and boundary conditions (3).Using the Crank-Nicolson method and assuming a constant D eff , a matrix of the concentration profile at any time is obtained based on the initial conditions and boundary conditions.There is no limitation of D eff δ/(δr) 2 ≤ 1/2 for this method.The C s value is averaged by an integration method and plotted as a function of time:

D eff ∂C p ∂r r=rp
= D S ∂C s ∂r r=rp , C p r p , t = C s r p , t . ( A few factors such as the effective diffusion coefficients in the particles (D eff ) and in the bulk solution (D S ), particle size effect, dilution level, and distribution of the active corrosion inhibitor have been taken into consideration.
Figure 10 shows how external diffusion, D S (the diffusion in deionized (DI) water), affects the predicted release of 2.5 μm particles with an uniformly distributed inhibitor.The dilution level, DI water/particle (weight), is kept as 100, and   the internal diffusion coefficient, D P , is set as 1 × 10 −15 m 2 /s.We can see that external transport resistance can be negligible for a D S /D P greater than 100.The release curves are the same for the D S of a value of 1 × 10 −9 , 1 × 10 −10 , and 1 × 10 −12 m 2 /s.The diffusion coefficients of ions in water are within this range.For the release process, stirring does not speed the release as also evidenced by the experiment.There is no need to obtain the exact value for D S .We can select a D S with a value 10 −10 or 10 −9 m 2 /s without compromising the accuracy of D P in the simulation.The internal diffusion has also been examined.Figure 11 shows the effect of D P on the release with a D S fixed as 1 × 10 −10 m 2 /s.The characteristic release time is reverse to D P .The particles with a higher D P show a quicker release.The release time also depends on the particle size.As shown in Figure 12, bigger particles show delayed release.Figure 13 shows simulated release results for 2.5 μm particles using the same inhibitor loading.The D S was set as 1 × 10 −10 m 2 /s, D P as 1 × 10 −15 m 2 /s, and a dilution level of 100.Compared to uniformly distributed particles, core-shell structured particles demonstrated delayed release, giving them a much longer service time.
To elucidate, actual effective diffusion coefficients for monodisperse Ce/silica and NaCl/silica particles of 10 μm were made by a vibrating orifice aerosol generator (VOAG), TSI Atomizer 3076.The particles were dispersed in DI water.The solution concentrations were characterized by measuring the conductivity or absorbance of the solution at different release times.The release profiles are shown in Figures 14 and  15 for Ce/silica and NaCl/silica particles, respectively.The effective diffusion coefficient is 2.5 × 10 −13 m 2 /s for NaCl and 1.0 × 10 −14 m 2 /s for Ce 3+ .The release behavior has been checked for the release rate in different n-alcohols.As shown in Figure 16, the constant release rate in various solvents over a long time suggests a solubility control for the release.The longer the solvent chain, the slower the release.The lower solubility and lower diffusion coefficient due to high viscosity account for the slower and straight release in high-carbon alcohols.For longterm corrosion inhibition, the inhibitors can be selected or converted to the form of suitable solubility in the corrosion medium.Highly soluble inhibitors are not a judicious choice from this point of view, but they may be if the inhibitor concentration is higher than the critical level required for inhibition.
Figure 17 shows release profiles for various Ce/silica particles.Methyltrimethoxysilane (MTES)/TEOS-derived hydrophobic particles and the microporous particles templated by triethylene glycol (TEG) demonstrated slow release.The estimated effective corrosion inhibition time for TEGtemplated particles is greater than 3000 h.
The nanostructured core-shell Ce/silica particles were evaluated for corrosion inhibition of aluminum alloy AA2024-T3.As shown in Figure 18, without the nanostructured Ce/silica particles the alloy was easily corroded, and many pits were formed after 100 h of immersion in a 0.05 M NaCl solution; see Figure 18(b).However, dispersing 1 g of nanostructured silica particle b (Figure 1(b)) mixed into 100 mL 0.05 M of NaCl solution efficiently inhibited corrosion (Figure 18(c)).A superhydrophobic coating layer containing the particles was also checked for corrosion protection.As can be seen in Figure 19, the coating protected the alloy well (Figure 19(b)).Without the core-shell particle coating, the alloy turned black and many pits appeared on the surface (Figure 19(c)).For comparison, the superhydrophobic coating without addition of the particles was checked for protection.Even though there was a lag time to sufficiently wet the sample, after being immersed in the solution for 447 h, the sample containing only the superhydrophobic turned black as well.The controlled release of the particles' corrosion inhibitors inhibited the corrosion and selfhealed the corroded surface.This preliminary data clearly demonstrate the effectiveness of Ce 3+ as corrosion inhibitor for protecting alloy AA2024-T3, because the inhibitor is releasable from the nanostructured particles.
Mathematical modeling for Ce 3+ diffusion in the nanoporous particles demonstrated that the core-shell structure helps delay the release process compared to uniformly distributed Ce/silica particles.The evaluated effective diffusion coefficient in water is 1.0 × 10 −14 m 2 /s at 25 • C. The experiment also proved that the release can be controlled effectively by pore size and pore surface chemistry.Longterm corrosion inhibition can be realized by particles with small and hydrophobic pores.But a denser and thicker coating is needed for a water/ion transport barrier.Further corrosion test is being conducted for the particles that are dispersed into a protective hydrophobic or epoxy coating layer for alloy AA2024-T3.

Conclusion
In conclusion, the nanostructured particles with a cerium core and mesoporous silica shell were synthesized via aerosol-assisted EISA and characterized using TEM, XRD, EPR, and so forth.We also demonstrated that the coreshell nanostructured particles were effective for corrosion inhibition of aluminum alloy AA2024-T3.

Particles Synthesis.
Silica/surfactant aerosols were generated using a commercial atomizer (Model 9302A; TSI, Inc., St. Paul, MN) operated with nitrogen as a carrier/atomization gas.The procedure was the same as described by Lu et al. [14,15].It starts with a homogeneous solution of soluble silica, acid, and surfactant prepared in an ethanol/water solvent with an initial surfactant concentration much less than the CMC.The pressure drop at the pinhole was 20 psi.The temperature for three heating zones was kept at 400 • C. Particles were collected on a Durapore Membrane Filter maintained at 80 • C. The precursor sol molar ratio was 1 TEOS : 7.7 EtOH : 24.8 water : 0.0125 HCl : 0.18 CTAB : 0-0.24CeCl 3 .All calcination treatments were performed in a Lindberg tube furnace at 500 • C for 5 h in forming gas (7% H 2 and 93% N 2 ) at a heating rate of 1 • C/min.

Characterization. XRD spectra were recorded on a
Siemens D-500 diffractometer, using Ni-filtered Cu Kα radiation with λ = 1.5406Å, a graphite monochromator, and a scintillation detector; data were refined using the JADE software package (Materials Data Inc., Livermore, CA).
TEM was performed using a JEOL 2010 HREM equipped with Oxford-Link X-ray energy-dispersive spectroscopy (EDS).A Gatan slow scan CCD camera was used for recording TEM images and electron diffraction patterns, operating at an accelerating voltage of 200 kV.4.4.N 2 Isothermal Sorption.Nitrogen isothermal adsorption was measured at 77 K on a Micromeritics ASAP 2010 porosimeter.Surface area was estimated by using the BET equation.The pore size was calculated using the BJH model.

EPR Spectroscopy.
Electron paramagnetic resonance (EPR) analysis was operated using a Bruker Elexsys spectrometer with a microwave frequency of 9.8 GHz, a modulation amplitude of 10 GHz (100 kHz), a receiver gain of 10 4 , and a time constant of 82 ms.Samples were prepared by mixing ground pure CeCl 3 or sample c with boron nitride.The standard sample made from pure CeCl 3 had the same Ce loading as the sample c mixtures.The Ce 3+ /(Ce 3+ + Ce 4+ ) was calculated by the ratio of integrated EPR signals of sample c to the standard sample.
Release of NaCl and Ce 3+ from the particles was characterized by measuring the conductivity or the absorption using a conductivity meter or an UV-Vis spectrometer.

Corrosion Inhibition Test.
The corrosion inhibition of particles was checked by dispersing the particles in 0.05 M NaCl aqueous solution.AA2024-T3 alloy coupons were immersed in the solution over 100 h followed by optical microscopy analysis.4.6.1.Superhydrophobic Coating on AA20240-T3.Tetramethylorthosilicate (TMOS), methanol, ammonium hydroxide (35% wt), 3,3,3-trifluoropropyltrimethyoxysilane (TFPT-MOS), and DI water were mixed in a molar ratio of 1 : 41.56 : 0.003 : 0.33 : 5.85.The mixture was stirred for 10 min and aged for 48 h at 50 • C. The formed gel was washed with hexane and ethanol for 2 h.The gel was diluted with 1 : 1 ethanol and kept for sonification for reliquefaction.The redispersed sol was filtered through a 1 μm glass fiber filter.The filtered sol was mixed with core-shell Ce/silica particles (100 : 1 wt) and was spin-coated at 200 rpm for 20 s on a AA20240-T3 alloy substrate.AA2024-T3 alloy coupons with superhydrophobic coating containing coreshell particles were dipped in a 0.05 M NaCl aqueous solution for 447 h before optical microscopy analysis.

Figure 6 :Figure 7 :Figure 8 :Figure 9 :Scheme 1 :
Figure 6: EDX spectrum of the core-shell particles observed on the surface of a carbon-coated copper grid.

15 Figure 10 :
Figure 10: Effect of external diffusion (D S ) on the release process.

Figure 11 :
Figure 11: Effect on internal diffusion (D P ) on the release process.

Figure 12 :
Figure 12: Effect of particle size on the release process at a dilution level of 100 and D S = 1 × 10 −10 m 2 /s, D P = 1 × 10 −15 m 2 /s.

Figure 13 :
Figure 13: Inhibitor distribution of particles during the release process.

Figure 14 :
Figure 14: Release of Ce/silica particles in DI water.

Figure 15 :
Figure 15: Release of NaCl/silica particles in DI water.
To a point of evaporation, less soluble Ce species nucleate and precipitate.The salt crystallites cannot grow into a big single crystal in the presence of other Ce species as impurity.The oversaturated Ce chloride and Ce hydroxide nucleate and precipitate are collected/filtered by increasingly dense liquid-crystalline mesophases.The held- up Ce species undergo continued hydrolysis and have a higher tendency to be hydrolyzed than the inner salt does due to the steep temperature gradient.However, they cannot continue to grow because of the limited Ce source and existence of the well-assembled bi-continuous CTAB/silica liquid-crystal phase.EDX data prove that smaller particles 4.1.Materials.All the chemicals were used as received.TEOS, MTES, cetyltrimethylammonium bromide (CTAB), Ce(NO 3 ) 3 •6H 2 O, and CeCl 3 •7H 2 O were obtained from Sigma-Aldrich (St. Louis, MO).Hydrochloric acid was VWR brand (VWR International, West Chester, PA), while absolute ethanol was obtained from Aaper (Brookfield, CT).DI water (Millipore, Billerica, MA) was used for all experiments.Aluminum alloy AA20240-T3 was obtained as a 3.0-mm-thick sheet from Davidson Aluminum & Metal Corporation (Deer Park, NY).