Today there is an increasing interest in application of natural nanoscale materials like nanoclays, due to their biocompatibility and abundant availability. Halloysite nanotubes (also called HNTs) are one of the promising naturally occurring materials having unique multiwall tubular structure with mesoporous lumens of ca. 15 nm in diameter. The halloysite nanotubes vary in length from 0.5 to 1.0 μm, in external diameter from 40 to 60 nm, and in internal diameter from 10 to 30 nm (Fig. 1). Halloysite is chemically similar to kaolinite, but 0.72-nm thick aluminosilicate layers in its walls are separated by water molecules: Al2(OH)4Si2O5·nH2O . The layer unit consists of a corner-shared tetrahedral SiO4 sheet stacked with an edge-shared octahedral AlO6 sheet with internal aluminol group Al-OH. The halloysite is stable up to 590°C when de-hydroxylation occurs with disruption of the tube wall multilayer packing and 12 wt.% mass decrease; however, the tubular shape of the clay is preserved up to 1200°C . The outermost layer of halloysite tubes is silica, so its surface electrical ξ-potential is ca. −30 mV at pH 4–9. Halloysite has a positively charged Al(OH)3 inner lumen and negatively charged SiO2 outer surface. This gives the nanotubes a rare structure with different chemical composition of inner and outer surfaces.
The low price and biocompatibility of halloysite make its products easily scalable for industry, and its different inner/outer tube chemistry opens numerous applications , . The halloysite was used in sustained delivery of active chemical agents: anticorrosion, flame retardant, drugs , as a nanoreactor for enzymatic biocatalysts , and in environmental protection (for water purification and removal of heavy metals ions) .
One of the promising applications of the halloysite is its use as a template for metal particles formation. We formulated the following concepts of selective metal formation onto/into the HNTs:
- “Seeding” metal particles on the tubes outer-surface by simple metal ion reduction in presence of halloysite nanotubes. Various catalysts have been made by “seeding” metal nanoparticles (Au, Pt, Pd, Ag) on the halloysite surface , , , , .
- Loading metal ions selectively into 10–30-nm diameter halloysite lumen followed by a reduction reaction restricted in the tube’s lumen. Only two of our papers ,  are dedicated to the decoration of halloysite inner surfaces with metals, though this synthesis is very interesting for mesoporous catalysts because it allows avoiding nanoparticle aggregation at high temperatures (400–500°C). The meso-size diameter of the halloysite lumen restricts the growth of metals. With a proper modification of the inner surface, core-particles of 1–15 nm could be produced.
The main difficulty in loading metal ions inside halloysite nanotubes is that its positively charged internal surface composed of AlOx limits cation loading due to electrostatic repulsion. Some metal salts like silver acetate were loaded into the lumens but the yield of such core-shell synthesis was rather low, 5–6% . One could also estimate that to produce a high metal loading inside the nanotubes, we have to obtain a kind of nano-pump delivery system for reagents as it was proposed for gold “peapod” structures . To make the procedure more efficient, a ligand-assisted loading was applied for Cu/Ni catalysts preparation . It was proven that Cu/Ni@HNTs core-shell composite was more efficient in CO oxidation than Cu/Ni nanoparticles externally deposited on halloysite. Ni-nanowires were also grown inside halloysite lumen via electroless deposition of nickel .
Here we focus on the ligand-based strategy. An intercalation of metal ions in the lumen and between the multilayer wall slits in this case is promoted with Schiff-base complexation. Displacement of halloysite intercalated water molecules is possible with small organic molecules. Examples include urea and furfural; these small molecules can be controlled by increasing the wall multilayer spacing from 0.7 up to ~1.1 nm . We elaborated formation of metal nanocores inside the nanotube’s lumens using the ligand-assisted intercalation with the goal of designing efficient mesoporous catalytic material. Ru-based catalysts are effective in redox-reactions such as hydrocarbon hydrogenation, methanol oxidation in fuel cells, and hydrogen generation by hydrolysis of borohydrates , , .
The high price of particulate Ru catalysts and its tendency to aggregate make its usage difficult and not cost-effective without a porous support. A number of mesoporous substrates were proposed for Ru deposition , , , but no study was made on natural clay nanotubes. Here we report a high yield method for 2–3 nm Ru-particles synthesis inside halloysite lumen via the ligand-assisted intercalation/reduction procedure.
Materials and methods
Halloysite clay was obtained from Applied Minerals Inc., NY, USA. The white halloysite powder consists of cylindrical hollow tubes with length of 0.5–1.0 μm and an external diameter of 50–60 nm. Ruthenium chloride, (RuCl3), furfural (C5H4O2), hydrazine hydrate (NH2NH2×H2O), acetone (C3H6O), urea (CH4N2O), salicylaldehyde (C7H6O2), and ethanol (C2H6O) were purchased from Sigma-Aldrich.
Synthesis of HNTs@Schiff base composite
Schiff bases of different chemical composition (Fig. 2) were used as organic ligands to enhance Ru3+ intercalative binding. Urea and hydrazine hydrate was also used as a ligand for the complex formation but showed poor results. Halloysite nanotubes (1 g) were dispersed in hydrazine hydrate (20 mL) (samples 1, 2) or urea (saturated solution, 20 mL) (samples 3, 4), stirred for 20 min, precipitated with centrifugation and washed with ethanol 3×. Furfural (5 mL per 20 mL of ethanol), salicylaldehyde (20 mL) or acetone (20 mL) were then added. The mixture was stirred for 30 min and heated to 80°C to form the Schiff base. The resulted dispersion underwent washing/centrifugations with ethanol 3×. The HNT@Schiff base composites were dried at 50°C overnight.
Synthesis of Ru-nanoparticles inside the HNTs lumen
Thirty milliliter of RuCl3 ethanol solution (0.66–1.2 mg/mL) was mixed with 1 g of HNT@Schiff base composites. This mixture was then stirred and heated till 80°C for 30 min to form the ruthenium complex. The ruthenium complex which was linked into the nanotube interior was reduced with 100 mL of NaBH4 aqueous solution (2 mg/mL). The HNTs/Ru composite was washed with distilled water 3×, centrifuged at 7500 rpm for 3 min and dried at 50°C. The sample shown in Fig. 5c was prepared using a repeated loading step. Thirty milliliter of 0.66 mg/mL RuCl3 ethanol solution was loaded into an already reduced sample prepared with the same loading concentration. The loading procedure was repeated and followed by reducing step as was mentioned above.
The on-line adsorption experiment was performed in a quartz cell of the UV-VIS spectrophotometer Thermo Scientific Evolution. One milliliter of RuCl3 ethanol solution (1 mg/mL) was added into the quartz cell and the spectrum of the initial solution was measured. A tablet of HNTs@Schiff base (0.08 g) or pristine HNTs (for comparison) was placed on the bottom of the quartz sell. Changes in concentration of RuCl3 with time were recorded. The initial concentration and concentration of RuCl3 remaining in the solutions after 5, 10, 15, 20, 25 min exposure were determined. Absorbance measurements were made in the range of wavelength from 400 to 600 nm (the absorbance pick of RuCl3 is at 500 nm). The amount of RuCl3 adsorbed at time t (qt, mg/g) was calculated as (1):
where C0, Ct are the concentrations of RuCl3 initially and at time t, respectively, (mg/mL); m is the weight of the HNT@Schiff base composite (g); and V is the volume of the solution (mL). Initial concentration (C0) was 1 mg/mL, time dependent concentrations (Ct) were calculated from the optical density of solution at the maximum absorbance wavelength. Maximum adsorption of RuCl3 per 1 g of HNTs-Schiff base composite was reached after 25 min.
HNTs@Ru composites characterization
The samples morphology and elemental composition were analyzed with a transmission electron microscope (TEM), energy-dispersive X-ray spectrometry analysis (EDX) using a JEM-2100, and a JEOL instrument equipped with an energy EDX, voltage of 200 kV. Elemental analysis was also performed on X-ray flourometer ARL Perform’X (Thermo Fisher Scientific, New Wave).
Results and discussion
Halloysite is environmentally friendly material with low toxicity, and it has showed the best biocompatibility of all the studied clays, including kaolin and montmorillonite . The halloysite was used for ruthenium nanoparticle formation inside the tubes lumens. The Schiff base-assisted technique was used to avoid Ru3+ repulsion by the positively charged lumens. The pristine halloysite structure is shown on Fig. 1.
HNTs@Schiff base metal complexation
Schiff bases are compounds containing the azomethine group (-HC=N-). They are condensation products of ketones or aldehydes with primary amines . Schiff bases form stable metal complexes. These complexes can be used for nanoparticle formation , . In order to synthesize metal nanoparticles encapsulated into the clay, we used Schiff base assisted loading of metal ions into the nanotubes’ lumens (Scheme 1). For Schiff base formation exclusively in the tube interior, we loaded small guest molecules (hydrazine hydrate or urea) into the tubes and then washed the external surface. This was followed by synthesis of Schiff bases. Efficient Schiff bases for metal linkages were synthesized from the following combinations: /furfural+hydrazine/acetone+urea/furfural+urea/urea+salicilaldehyde (Fig. 2). The usage of only urea formed weak ligands resulting in loosely bound Ru-nanoparticle and their fast removal from the tubes’ lumens in aqueous solution.
To optimize the synthesis, Schiff bases derived from hydrazine hydrate and furfural was chosen (Scheme 1). Hydrazine hydrate as a first loading reagent was better than furfural for the following reasons: no temperature is needed to lower the solution viscosity for faster loading, no competing reaction of Ru-(N2H4)2 complex formation takes place, and no traces of hydrazine hydrate is left after the reaction.
Adsorption of RuCl3 from ethanol solution by HNTs@Schiff base composite
Figure 3 shows an adsorption isotherm for Ru3+ ions adsorption by HNTs@Schiff base composite. The pristine untreated HNTs showed very little adsorption from 1 mg/mL RuCl3 ethanol solution (less than 0.1 mg/g). HNTs-Schiff base composite adsorbed RuCl3 with saturation in 30 min. The total amount of RuCl3 adsorbed at time t (qt, mg/g) was calculated with Equation (1) and equals 3.2 mg of RuCl3 per 1 g of halloysite.
From Fig. 4a,b one can see that without organic linkages, Ru nanoparticles were not confined into the nanotubes interior but rather accumulated outside the tube. By utilizing the Schiff base linkage technique, small 3–5 nm Ru-nanoparticles were formed exclusively inside the tubes.
Without linking, Ru3+ did not saturate the pristine halloysite lumen and mostly adsorbed onto its outer surface forming metal aggregates after reduction. However, the Schiff base linking technique has been proven to be a more efficient method to selectively load nanoparticles only in the lumen and interlayer spacing. After the Schiff bases are formed, they then bind with metal cations and are reduced to form 3–4 nm particles with a narrow size distribution and without aggregation. EDX analysis and the corresponding elemental mapping proved the formation of Ru-particles (Fig. 5). The method of Schiff base assisted loading of metal ions inside halloysite lumen resulted in nearly a 100% yield of halloysite loaded with ruthenium nanoparticles.
Effect of RuCl3 concentration and additional cycle of intercalation-reduction
EDX spectra of HNTs@Ru (Fig. 5) demonstrate strong Ru signals in all the samples. The increase in the initial concentration of RuCl3 does not necessary leads to increase in number of Ru-particles and its distribution alone the lumen. At higher RuCl3 concentration in precursor solution (1.2 mg/mL) some nanotubes were not intercalated with metal. In case of RuCl3 concentration of 0.66 mg/mL nanoparticles were evenly distributed alone the lumen. So the lower concentration of RuCl3 gives better core-shell structures than higher ones.
Figure 5c shows that an additional-second cycle of 0.66 mg/mL RuCl3 intercalation leads to dramatic increase in the nanoparticle population within the tubes. EDX elemental analysis showed that after first cycle of intercalation/reduction 2 wt.% of Ru was encapsulated, and after second cycle the loading increased to 8 wt.% in a single tube. The elemental analysis performed on an X-ray fluorimeter showed that the loading after one cycle was 1.25 wt.% and after second cycle was 2.4 wt.%. This could be used in process optimization and for creating bimetallic halloysite loading.
Effect of Schiff base on Ru-particles size distribution
In Fig. 6 particle size distributions inside the halloysite lumen are shown depending on initial RuCl3 concentrations and composition of the ligand. By the ligand selection, we can produce less populated narrow size distribution Ru nanoparticles of 1.6±0.7 nm and well populated 3.3±1.0 nm or 3.0±1.0 nm core-shell systems. The largest Ru island-like clusters of about 15.3 nm were formed when Schiff base derived from urea and salicylaldehyde was used as a ligand.
It is interesting that we can produce very small and nearly monodispersed Ru-particles (<2 nm). It is expected that such materials can bridge the gap between homogenous and heterogeneous catalytic systems , .
According to the XPS data , Ru in the samples is present both in zero-valent and oxidized states. The obtained nanosystems have shown good result in reactions of catalytic hydrogenation of phenol and cresols. Phenol conversion 100%, selectivity to cyclohexanol – 100%, PhOH/Ru=2685 mol/mol and cresols conversion up to 33% at substrate/Ru ratio of 2600–2800 and methylcyclohexanol selectivity: o-cresol (61%) <m-cresol (90%) <p-cresol (99%) .
An effective noble metal intercalation into tubule nanoclay via ligand assisted method was demonstrated. 2–3 nm Ru-nanoparticles were synthesized inside halloysite clay lumen using a Schiff base assisted intercalation technique followed by reduction. The method does not need high temperature, halloysite pre-treatment, solvent purification, nor high concentrations of metals. This provides a perspective not only for functional core-shell nanosystems, but also as mesoporous material for adsorption of heavy metal ions. The protocol involved immobilization of ruthenium complex over the inner surface of halloysite lumen, and then its reduction to metal particles. The results are metal-ceramic nanosystems with different Ru size distributions depending on the chemical structure of the Schiff base used, initial RuCl3 concentration, and number of intercalation/reduction cycles. The method enabled the synthesis of ultra-small 1.8 nm Ru-particles with nearly monodispersed size distribution.
We presented here results on ruthenium nanoparticle synthesis, although our approach was generalized for production of Rh, Pt, Cu, Fe core-shell nanosystems. We assume that similar results could be obtained for Au, Ag, Pd, Co, Ni and other transition metals which are used as industrial catalysts.
This work was supported by the Ministry of Education and Science of the Russian Federation (Grant № 14.Z50.31.0035). We are thankful to Y. Darrat, Louisiana Tech, for his collaboration.
V. Vergaro, E. Abdullayev, Y. M. Lvov, A. Zeitoun, R. Cingolani, R. Rinaldi, S. Leporatti. Biomacromolecules 11, 820 (2017).
R. Kamble, M. Ghag, S. Gaikawad, B. K. Panda. JASR 3, 25 (2012).
Y. Zhang, X. He, J. Ouyang, H. Yang. Sci. Rep. 3 (2013).
N. M. Sanchez-Ballester, G. V. Ramesh, T. Tanabe, E. Koudelkova, J. Liu, L. K. Shrestha, Y. Lvov, J. P. Hill, K. Ariga, H. Abe. J. Mater. Chem. A 3, 6614 (2015).
Y. Fu, L. Zhanga. J. Solid State Chem. 178, 3593 (2005).
K. Kusada, H. Kobayashi, T. Yamamoto, S. Matsumura, N. Sumi, K. Sato, K. Nagaoka, Y. Kubota, H. Kitagawa. J. Am. Chem. Soc. 135, 5493 (2013).
A. Xavier, N. Srividhya. J. Appl. Chem. 7, 6 (2014).
C. R. Bhattacharje, P. Goswami, S. Neogi, S. Dhibar. Assam Univ. J. Sci. Technol. Phys. Sci. Technol. 5, 81 (2010).
F. A. Abdlseed, M. M. El-ajaily. Int. J. PharmTech Res. 1, 1097 (2009).
E. A. Karakhanov, A. P. Glotov, A. G. Nikiforova, A. V. Vutolkina, A. O. Ivanov, S. V. Kardashev, A. L. Maksimov, S. V. Lysenko. Fuel Process. Technol. 153, 50 (2016).
V. Vinokurov, A. Glotov, Y. Chudakov, A. Stavitskaya, E. Ivanov, P. Gushchin, A. Zolotukhina, A. Maximov, E. Karakhanov, Y. Lvov. Ind. Eng. Chem. Res. 56, 14043 (2017).