Halloysite Nanotubes: Controlled Access and Release by Smart Gates

Hollow halloysite nanotubes have been used as nanocontainers for loading and for the triggered release of calcium hydroxide for paper preservation. A strategy for placing end-stoppers into the tubular nanocontainer is proposed and the sustained release from the cavity is reported. The incorporation of Ca(OH)2 into the nanotube lumen, as demonstrated using transmission electron microscopy (TEM) imaging and Energy Dispersive X-ray (EDX) mapping, retards the carbonatation, delaying the reaction with CO2 gas. This effect can be further controlled by placing the end-stoppers. The obtained material is tested for paper deacidification. We prove that adding halloysite filled with Ca(OH)2 to paper can reduce the impact of acid exposure on both the mechanical performance and pH alteration. The end-stoppers have a double effect: they preserve the calcium hydroxide from carbonation, and they prevent from the formation of highly basic pH and trigger the response to acid exposure minimizing the pH drop-down. These features are promising for a composite nanoadditive in the smart protection of cellulose-based materials.


Introduction
Halloysite clay (HNT) is a natural and abundantly available nanoparticle formed by rolled kaolin sheets. The main deposits of HNT are from Dragon Mine and Matauri Bay, which are in Utah (USA) and Northland (New Zealand), respectively. Due to its biocompatibility [1,2] HNT was recently studied for the development of innovative nanomaterials useful for biotechnological applications, such as the controlled release of drugs [3][4][5][6], tissue engineering [7][8][9], oil recovery [10], and eco-compatible packaging [11][12][13]. Furthermore, several studies proved that HNT is a suitable catalytic support [14,15], as well as an efficient removal agent [16], because of its geometrical and surface properties (large specific area, hollow tubular shape, and tunable surface chemistry). Both the sizes and polydispersity are influenced by the HNT geological deposit [17]. Typically, the HNTs lengths range between 0.1 and 3.0 µm, while their external and inner diameters are ca. 50-200 and 15-70 nm, respectively [18]. The HNT surfaces are oppositely charged within a large pH range (between 2 and 8) because of their different chemical compositions [19]. Particularly, the internal surface consists of gibbsite octahedral sheet (Al-OH) groups with a positive surface charge, whereas the outer surface is composed of siloxane groups (Si-O-Si) with a negative electrical potential. Accordingly, the selective HNT functionalization can be easily achieved through electrostatic interactions between the nanoparticle surfaces and ionic Thermogravimetric analysis (TGA) data on HNTs-Ca(OH)2 with calcium phosphate endstoppers did not show any significant difference from the HNTs-Ca(OH)2 sample as proof that the end-stopper treatment did not alter the general composition of the material to a large extent ( Figure  S1). To investigate the end-stopper formation, TEM experiments were carried out on HNTs-Ca(OH)2 with calcium phosphate end-stoppers. Literature reports on TEM images for HNTs samples were able to identify the lumen filling especially for high electron density materials, such as metals [51,52]. The images for HNTs-Ca(OH)2 based system show that the lumen of HNTs is filled (Figure 2, additional images are in Figure S2).  Thermogravimetric analysis (TGA) data on HNTs-Ca(OH) 2 with calcium phosphate end-stoppers did not show any significant difference from the HNTs-Ca(OH) 2 sample as proof that the end-stopper treatment did not alter the general composition of the material to a large extent ( Figure S1). To investigate the end-stopper formation, TEM experiments were carried out on HNTs-Ca(OH) 2 with calcium phosphate end-stoppers. Literature reports on TEM images for HNTs samples were able to identify the lumen filling especially for high electron density materials, such as metals [51,52]. The images for HNTs-Ca(OH) 2 based system show that the lumen of HNTs is filled ( Figure 2, additional images are in Figure S2). Nanomaterials 2017, 7,199 3 of 12 at 900 °C for pristine components and assuming the rule of mixtures. On this basis, one can calculate a value of 3.9 ± 0.2% w/w (corresponding to 4.5 ± 0.3% v/v) for the loading. Given that the full geometrical filling would provide ca. 10% v/v of loaded material [19], one may conclude that ca. half of the lumen is filled by the calcium hydroxide. The presence of Ca(OH)2 in the HNTs-Ca(OH)2 composite was confirmed by Fourier transform infrared spectroscopy (FTIR) spectra. As evidenced in Figure 1, the composite material presents the characteristic bands of both components, proving that during the loading procedure the Ca(OH)2 is preserved and incorporated in the composite. Thermogravimetric analysis (TGA) data on HNTs-Ca(OH)2 with calcium phosphate endstoppers did not show any significant difference from the HNTs-Ca(OH)2 sample as proof that the end-stopper treatment did not alter the general composition of the material to a large extent ( Figure  S1). To investigate the end-stopper formation, TEM experiments were carried out on HNTs-Ca(OH)2 with calcium phosphate end-stoppers. Literature reports on TEM images for HNTs samples were able to identify the lumen filling especially for high electron density materials, such as metals [51,52]. The images for HNTs-Ca(OH)2 based system show that the lumen of HNTs is filled ( Figure 2, additional images are in Figure S2).   EDX mapping allowed us to make a proper identification of the filling; as Figure 3 shows, the Ca signals come from the same spots where tubular-like nanostructures are imaged. As a confirmation, this is also the case for Al and Si, which are HNT components, and the Ca signal is absent in pristine HNTs. Going further, a phosphorus signal was detected, proving that phosphate was, by some means, kept in the sample during the treatment. Its concentration is relatively small and far below that stoichiometrically expected for Ca 3 (PO 4 ) 2 . On the other hand, P is not phase separated within the observed sample. By a close look at the nanotube ends (Figure 1), it is revealed that they are closed by what appears to be a stopper, moreover the lumen cavity nearby the HNTs' termination appears empty. Such a morphological observation is in agreement with a mechanism of end stopper formation based on the reaction between partially-released Ca(OH) 2 and Na 3 PO 4 in proximity of the nanotube ends forming Ca 3 (PO 4 ) 2 due to a high local concentration. It should be noted that a flow of Na 3 PO 4 aqueous solution is used and that a short solution-HNT/Ca(OH) 2 contact time is ensured by vacuum filtration in order to avoid a complete HNT unload. A schematic representation of end stoppers' formation is depicted in Figure 4.
confirmation, this is also the case for Al and Si, which are HNT components, and the Ca signal is absent in pristine HNTs. Going further, a phosphorus signal was detected, proving that phosphate was, by some means, kept in the sample during the treatment. Its concentration is relatively small and far below that stoichiometrically expected for Ca3(PO4)2. On the other hand, P is not phase separated within the observed sample. By a close look at the nanotube ends (Figure 1), it is revealed that they are closed by what appears to be a stopper, moreover the lumen cavity nearby the HNTs' termination appears empty. Such a morphological observation is in agreement with a mechanism of end stopper formation based on the reaction between partially-released Ca(OH)2 and Na3PO4 in proximity of the nanotube ends forming Ca3(PO4)2 due to a high local concentration. It should be noted that a flow of Na3PO4 aqueous solution is used and that a short solution-HNT/Ca(OH)2 contact time is ensured by vacuum filtration in order to avoid a complete HNT unload. A schematic representation of end stoppers' formation is depicted in Figure 4.
Additional dark field optical images were taken from the aqueous dispersion HNTs-Ca(OH)2 with calcium phosphate end-stoppers. Figure 5 shows that the nanotubes generate a uniform dispersion as they are not aggregated in water. Therefore, the preparation protocols avoid any clustering of nanoparticles. The literature reports that aggregation and dispersion behaviours of halloysite nanotubes (HNTs) can be influenced by pH [53]. In particular, it is reported that the pH variation could be used as a strategy for blocking and opening the halloysite cavity. In our system, based on the observed morphology by TEM and dark field (DF) microscopy, we can exclude a clustering of HNTs-Ca(OH)2 nanoparticles and, therefore, the controlled access/release due to aggregative phenomena.   confirmation, this is also the case for Al and Si, which are HNT components, and the Ca signal is absent in pristine HNTs. Going further, a phosphorus signal was detected, proving that phosphate was, by some means, kept in the sample during the treatment. Its concentration is relatively small and far below that stoichiometrically expected for Ca3(PO4)2. On the other hand, P is not phase separated within the observed sample. By a close look at the nanotube ends ( Figure 1), it is revealed that they are closed by what appears to be a stopper, moreover the lumen cavity nearby the HNTs' termination appears empty. Such a morphological observation is in agreement with a mechanism of end stopper formation based on the reaction between partially-released Ca(OH)2 and Na3PO4 in proximity of the nanotube ends forming Ca3(PO4)2 due to a high local concentration. It should be noted that a flow of Na3PO4 aqueous solution is used and that a short solution-HNT/Ca(OH)2 contact time is ensured by vacuum filtration in order to avoid a complete HNT unload. A schematic representation of end stoppers' formation is depicted in Figure 4.
Additional dark field optical images were taken from the aqueous dispersion HNTs-Ca(OH)2 with calcium phosphate end-stoppers. Figure 5 shows that the nanotubes generate a uniform dispersion as they are not aggregated in water. Therefore, the preparation protocols avoid any clustering of nanoparticles. The literature reports that aggregation and dispersion behaviours of halloysite nanotubes (HNTs) can be influenced by pH [53]. In particular, it is reported that the pH variation could be used as a strategy for blocking and opening the halloysite cavity. In our system, based on the observed morphology by TEM and dark field (DF) microscopy, we can exclude a clustering of HNTs-Ca(OH)2 nanoparticles and, therefore, the controlled access/release due to aggregative phenomena.   Additional dark field optical images were taken from the aqueous dispersion HNTs-Ca(OH) 2 with calcium phosphate end-stoppers. Figure 5 shows that the nanotubes generate a uniform dispersion as they are not aggregated in water. Therefore, the preparation protocols avoid any clustering of nanoparticles. The literature reports that aggregation and dispersion behaviours of halloysite nanotubes (HNTs) can be influenced by pH [53]. In particular, it is reported that the pH variation could be used as a strategy for blocking and opening the halloysite cavity. In our system, based on the observed Nanomaterials 2017, 7, 199 5 of 12 morphology by TEM and dark field (DF) microscopy, we can exclude a clustering of HNTs-Ca(OH) 2 nanoparticles and, therefore, the controlled access/release due to aggregative phenomena.

Kinetics Study on Carbonatation and Release of Ca(OH)2 from HNT Lumen
In addition to the interesting molecular architecture, we investigated the functionality of the end stopper in playing any barrier role for gas or to control the release of Ca(OH)2 from the lumen. Calcium hydroxide typically undergoes CO2 capture with CaCO3 formation. This process has been widely investigated due to the relevant applicative interest [54]. To explore the ability of HNT lumen in controlling such a process, we used thermogravimetric analysis under a CO2 atmosphere. The degree of Ca(OH)2 conversion to CaCO3, based on measured mass gain and initial Ca(OH)2 content in the measured sample is provided in Figure 6 as a function of time. It is worth noting that confining Ca(OH)2 within the HNTs lumen cavity significantly retards the carbonation reaction. Furthermore, the end-stoppers prevent the CO2 contact and less than 10% of the calcium hydroxide is converted to carbonate after 1 h under the experimental conditions. Although the time frame is relatively short (one hour), the experiment proves that encapsulated Ca(OH)2 is still in its original form when bare Ca(OH)2 undergoes complete conversion. This result is very promising for applications as it shows the possibility to keep Ca(OH)2 preserved from carbonation during the treatment.
The release kinetics of Ca(OH)2 in water were investigated by measuring the pH of the dispersion over time. To this aim, the aqueous dispersions of HNT and HNT/Ca(OH)2 with and without end-stoppers (0.1 wt %) were left to equilibrate under static conditions while a glassy

Kinetics Study on Carbonatation and Release of Ca(OH) 2 from HNT Lumen
In addition to the interesting molecular architecture, we investigated the functionality of the end stopper in playing any barrier role for gas or to control the release of Ca(OH) 2 from the lumen. Calcium hydroxide typically undergoes CO 2 capture with CaCO 3 formation. This process has been widely investigated due to the relevant applicative interest [54]. To explore the ability of HNT lumen in controlling such a process, we used thermogravimetric analysis under a CO 2 atmosphere. The degree of Ca(OH) 2 conversion to CaCO 3 , based on measured mass gain and initial Ca(OH) 2 content in the measured sample is provided in Figure 6 as a function of time.

Kinetics Study on Carbonatation and Release of Ca(OH)2 from HNT Lumen
In addition to the interesting molecular architecture, we investigated the functionality of the end stopper in playing any barrier role for gas or to control the release of Ca(OH)2 from the lumen. Calcium hydroxide typically undergoes CO2 capture with CaCO3 formation. This process has been widely investigated due to the relevant applicative interest [54]. To explore the ability of HNT lumen in controlling such a process, we used thermogravimetric analysis under a CO2 atmosphere. The degree of Ca(OH)2 conversion to CaCO3, based on measured mass gain and initial Ca(OH)2 content in the measured sample is provided in Figure 6 as a function of time. It is worth noting that confining Ca(OH)2 within the HNTs lumen cavity significantly retards the carbonation reaction. Furthermore, the end-stoppers prevent the CO2 contact and less than 10% of the calcium hydroxide is converted to carbonate after 1 h under the experimental conditions. Although the time frame is relatively short (one hour), the experiment proves that encapsulated Ca(OH)2 is still in its original form when bare Ca(OH)2 undergoes complete conversion. This result is very promising for applications as it shows the possibility to keep Ca(OH)2 preserved from carbonation during the treatment.
The release kinetics of Ca(OH)2 in water were investigated by measuring the pH of the dispersion over time. To this aim, the aqueous dispersions of HNT and HNT/Ca(OH)2 with and without end-stoppers (0.1 wt %) were left to equilibrate under static conditions while a glassy It is worth noting that confining Ca(OH) 2 within the HNTs lumen cavity significantly retards the carbonation reaction. Furthermore, the end-stoppers prevent the CO 2 contact and less than 10% of the calcium hydroxide is converted to carbonate after 1 h under the experimental conditions. Although the time frame is relatively short (one hour), the experiment proves that encapsulated Ca(OH) 2 is still in its original form when bare Ca(OH) 2 undergoes complete conversion. This result is very promising for applications as it shows the possibility to keep Ca(OH) 2 preserved from carbonation during the treatment.
The release kinetics of Ca(OH) 2 in water were investigated by measuring the pH of the dispersion over time. To this aim, the aqueous dispersions of HNT and HNT/Ca(OH) 2 with and without end-stoppers (0.1 wt %) were left to equilibrate under static conditions while a glassy electrode was used to monitor the pH. A blank experiment reporting the kinetics for pure Ca(OH) 2 , in the same amount as the loaded value in HNTs, revealed a quick dissolution of the hydroxide that occurs within 5 min. After that a constant pH value was approached, 0.15 cm 3 of HCl (0.1 M) was added to the dispersion, and the pH response was measured for 18 h. With respect to the release in water, the HNT/Ca(OH) 2 composite showed a sustained increase of pH (Figure 7). Even slower is the pH increase for the composites with the phosphate end-stoppers being the most efficient in retarding the hydroxide solubilization in water. It is reported that the dissolution kinetics of nanosized materials is influenced by the grain size due to high specific area and surface energy effects. It should be noted that even if the net Ca(OH) 2 amount was similar for all samples, a higher pH is approached at the plateau for the HNT/Ca(OH) 2 composite compared to the system with end-stoppers. This result reflects the ability of the end-stopping strategy to retain the hydroxide in the HNT lumen even in water media for a certain extent.
The HCl addition generates a sudden drop-down of the pH that slowly returns toward higher values due to a further release of the calcium hydroxide from the lumen. The pH increasing trend is significantly slowed by the end-stopper presence. Moreover, the ∆pH, due to the HCl addition and after equilibration, is 0.35 and 0.85 for HNT/Ca(OH) 2 with and without end-stoppers, respectively. From the stoichiometric calculation a pH change of 1.35 is expected if all of the calcium hydroxide would have been dissolved prior to HCl addition. Therefore, we might conclude that confining Ca(OH) 2 into the HNT lumen generates an alkaline reservoir which is released in response to acid addition. Nanomaterials 2017, 7, 199 6 of 12 electrode was used to monitor the pH. A blank experiment reporting the kinetics for pure Ca(OH)2, in the same amount as the loaded value in HNTs, revealed a quick dissolution of the hydroxide that occurs within 5 min. After that a constant pH value was approached, 0.15 cm 3 of HCl (0.1 M) was added to the dispersion, and the pH response was measured for 18 h. With respect to the release in water, the HNT/Ca(OH)2 composite showed a sustained increase of pH (Figure 7). Even slower is the pH increase for the composites with the phosphate end-stoppers being the most efficient in retarding the hydroxide solubilization in water. It is reported that the dissolution kinetics of nanosized materials is influenced by the grain size due to high specific area and surface energy effects. It should be noted that even if the net Ca(OH)2 amount was similar for all samples, a higher pH is approached at the plateau for the HNT/Ca(OH)2 composite compared to the system with end-stoppers. This result reflects the ability of the end-stopping strategy to retain the hydroxide in the HNT lumen even in water media for a certain extent. The HCl addition generates a sudden drop-down of the pH that slowly returns toward higher values due to a further release of the calcium hydroxide from the lumen. The pH increasing trend is significantly slowed by the end-stopper presence. Moreover, the ∆pH, due to the HCl addition and after equilibration, is 0.35 and 0.85 for HNT/Ca(OH)2 with and without end-stoppers, respectively. From the stoichiometric calculation a pH change of 1.35 is expected if all of the calcium hydroxide would have been dissolved prior to HCl addition. Therefore, we might conclude that confining Ca(OH)2 into the HNT lumen generates an alkaline reservoir which is released in response to acid addition.

Effect of HNT/Ca(OH)2 on Paper Deacidification and Consolidation
The efficacy of the prepared nanomaterials on paper deacidification was monitored by cycling the aging protocols and controlling the paper conditions and its damage by pH measurements and tensile experiments.
The paper sample without a Ca(OH)2 basic reservoir reaches acid pH values after the first aging cycle and it remains constant, not being able to compensate for the effect of acid gas presence ( Table  1). The HNTs/Ca(OH)2 system generated a paper alkaline pH which systematically decreases with aging approaching the value for the paper sample without the basic reservoir. Keeping in mind the strong acidic environment used in this investigation compared to the actual situations that might be experienced in the conventional conservation for books, the obtained results are already promising. On the other hand, the end-stopped system could be considered even more performant as the starting

Effect of HNT/Ca(OH) 2 on Paper Deacidification and Consolidation
The efficacy of the prepared nanomaterials on paper deacidification was monitored by cycling the aging protocols and controlling the paper conditions and its damage by pH measurements and tensile experiments.
The paper sample without a Ca(OH) 2 basic reservoir reaches acid pH values after the first aging cycle and it remains constant, not being able to compensate for the effect of acid gas presence ( Table 1). The HNTs/Ca(OH) 2 system generated a paper alkaline pH which systematically decreases with aging approaching the value for the paper sample without the basic reservoir. Keeping in mind the strong acidic environment used in this investigation compared to the actual situations that might be experienced in the conventional conservation for books, the obtained results are already promising. On the other hand, the end-stopped system could be considered even more performant as the starting value for pH is only slightly basic and the pH change is kept within one unit even after two aging cycles when a still alkaline/neutral pH is maintained.
Tensile measurements provided information on the alteration of mechanical performance for paper samples after exposure to acidic gas. Stress at breaking point (σ r ) data showed that no treatments significantly altered the paper property, while only the samples with the alkaline reservoir were able to minimize the σ r reduction upon aging ( Table 1). The mechanical performance might also be described by the storage energy parameter (SE) that is obtained from the stress vs. strain curve integral and provides an idea on the maximum energy that can be adsorbed by the paper sample until it breaks down. Results in Figure 8 demonstrate that paper aging reduced the SE if the alkaline reservoir is not introduced within the paper. On the other hand, the end-stopped system is more efficient in strengthening the paper, maintaining a relatively high SE value even after the aging protocol. value for pH is only slightly basic and the pH change is kept within one unit even after two aging cycles when a still alkaline/neutral pH is maintained. Tensile measurements provided information on the alteration of mechanical performance for paper samples after exposure to acidic gas. Stress at breaking point (σr) data showed that no treatments significantly altered the paper property, while only the samples with the alkaline reservoir were able to minimize the σr reduction upon aging ( Table 1). The mechanical performance might also be described by the storage energy parameter (SE) that is obtained from the stress vs. strain curve integral and provides an idea on the maximum energy that can be adsorbed by the paper sample until it breaks down. Results in Figure 8 demonstrate that paper aging reduced the SE if the alkaline reservoir is not introduced within the paper. On the other hand, the end-stopped system is more efficient in strengthening the paper, maintaining a relatively high SE value even after the aging protocol.
Thermogravimetry analysis (TGA): Experiments were done using the Q5000 IR (TA Instruments, Milan, Italy) under nitrogen flow (25 cm 3 ·min −1 ) by heating the samples from room temperature to 900 • C. Each sample (ca. 5 mg) was placed in a platinum pan and heated under the temperature program of 10 • C·min −1 . Loading was calculated according to the procedure in the literature and errors were evaluated from standard deviations of three measurements [11]. The CO 2 capturing experiments were carried out by quickly heating the sample (200 • C·min −1 ) to 600 • C in a N 2 flow (25 cm 3 ·min −1 ). Afterwards, the gas flow was switched to CO 2 with 99.995% chemical purity (25 cm 3 ·min −1 ). The mass gain was monitored for 60 min. The high temperature was chosen to accelerate the CO 2 capture based on literature reports [55]. Calibration was carried out as reported elsewhere [56].
Tensile Analysis: Tensile properties on paper samples were determined by means of a DMA Q800 instrument (TA Instruments, Milan, Italy). Tensile tests were performed on rectangular paper samples (10 mm × 4 mm) under a stress ramp of 1 MPa min −1 at 26.0 ± 0.5 • C. We determined the stress at which the material undergoes fractures (σ r ) and stored energy up to sample breaking by integrating the stress vs. strain curves. The reproducibility was checked by repeating the experiment three times.
pH measurements: The pH curves were obtained by using a PCD650 pH meter (Eutech Instruments, Landsmeer, The Netherlands) immersed in an aqueous dispersion of loaded nanoclay under stirring conditions. For all of the tested nanomaterials, dispersions were kept under a controlled environment, magnetic stirring, and measured continuously. Degassed water was used and the concentration of the dispersions was 0.1 wt %. The pH values of paper was measured by using a HI 1413B/50 portable pH meter with a flat-tip electrode (Hanna Instruments, Milan, Italy) in accordance with a non-destructive test that may be used to measure the hydrogen ion concentration (pH) on the surface of the paper in books and documents that constitute the collections of libraries and government archives (working procedure: TAPPI T 529 om-04).
TEM-EDX: For electron microscopy imaging and energy-dispersive X-ray analysis (EDX) a Hitachi HT7700 Exalens transmission electron microscope (Tokyo, Japan) was used. The samples were prepared by placing 10 µL of the suspension on a carbon-coated lace 3 mm copper grid, then dried at room temperature. TEM imaging was performed at a 100 kV accelerating voltage in TEM mode. EDX analysis was carried out in scanning transmission electron microscope (STEM) mode using an Oxford Instruments (High Wycomb, UK) X-Max™ 80T detector.
Enhanced dark-field imaging: During enhanced dark field microscopy experiments the images were obtained using a CytoViva ® enhanced dark-field condenser attached to an Olympus BX51 upright microscope equipped with fluorite 100× objectives and a DAGE CCD camera. Extra-clean dust-free Nexterion ® glass slides and coverslips (Schott, Mainz, Germany) were used for EDF microscopy imaging to minimise dust interference.
Loading of Ca(OH) 2 onto HNTs: Degassed aqueous solution of Ca(OH) 2 (1.5 g·dm −3 ) was mixed with halloysite powder (5 g·dm −3 ) and sonicated for 15 min. Then, the obtained suspension was stirred and kept under vacuum for 5 min resulting in light fizzling and the loaded compound condensated within the tube. This procedure was repeated three times to improve the loading efficiency. Successively, the nanotubes were separated from the aqueous phase by centrifugation and dried under vacuum at 70 • C overnight.
End-stopper formation: Aqueous phosphate solution was prepared by dissolving 40 g of trisodium phosphate dodecahydrate in 250 cm 3 of water. This solution was poured onto the HNT/Ca(OH) 2 powder placed in a Buechner funnel with filter paper placed on the perforated plate. Vacuum was applied during the pouring. The filtered material was dried by using a side-arm flask connected to a vacuum pump.
Paper treatments: For the paper treatment we prepared a 2 wt % HPC solution in ethanol. A certain amount of HNT (1 wt %) was added to the polymer solution and kept under stirring over night at 25 • C. The same procedure was followed by using HNT/Ca(OH) 2 with and without end-stoppers. It should be noted that ethanol was used as the solvent to avoid calcium hydroxide solubilization during the paper treatment. The paper samples were cut in a rectangular shape (40 mm × 8 mm) and they were deeply immersed into the well-dispersed aqueous mixtures for 24 h at 20 • C. The treated samples were dried at 35 • C.
Paper aging under acidic conditions: Paper specimens were placed in a closed desiccator. The vapours were saturated with HNO 3 by equilibrating the system with 30% acid solution. One aging cycle was three days. Before any characterization, the paper samples were re-equilibrated with air for 20 days.

Conclusions
We developed a novel strategy for sustained release and controlled access to the halloysite nanotubes lumen. Calcium hydroxide was loaded into the HNTs lumen and imaged by TEM and EDX mapping. End-stoppers were created when calcium hydroxide was partially released in the presence of phosphate anions. The obtained end-stoppers prevent CO 2 from entering the tube lumen and preserving the calcium hydroxide from carbonation. Moreover, they slow the release in water, minimizing the pH jumps if an acid is added to the dispersion. These features are very promising for paper preservation, as was demonstrated by aging experiments on treated and pristine cellulose paper samples. This composite nanomaterial would allow adding an alkaline reservoir to the paper and minimizing the pH changes, as well as the aging impact on the mechanical performance of the sample. The proposed strategy could be interesting in designing and building up nanocontainers with nanogates that are sensitive to external stimuli.