CaCl 2in-Mesoporous Silica Grown on Superabsorbent Polymer to Enhance Water Uptake †

Super absorbent polymer (SAP) sorbs copious amount of liquid water but its sorption power for water vapor is quite low. On the other hand, mesoporous silica loaded with CaCl2 (MPS-CC) has high sorption capacity for water vapor. However, this is determined by the salt loading, which is limited due to its corrosiveness. Even by simple powder blending of 75 mass% of SAP and the balance of MPS-CC (SAP75/MPS-CC25), the 3-hour maximum specific water uptake (g H2O/g sorbent) reached the maximum, 3 times the equilibrium uptake of MPS-CC. The sorption property was further enhanced by developing a new sorbent in which mesoporous silica was grown on the surface of swelling SAP and then impregnated by CC in ethanol. In the new sorbent (SAP-iMPS-CC), MPS grew organized, straight and lengthwise by the help of SAP. Owing to such intimate contact between MPS and SAP, the 3-hour specific water uptake of the new sorbents grown for more than 24 h (SAP-iMPS-CC24) was kept constant at the value of 2 times that of SAP75/MPS25. The SAP mass% of SAP-iMPS-144 could reach 25 without the sacrifice of sorption capacity.


Introduction
It is attractive to develop the adsorbents having high capacity and high rate of water vapor uptake for gas drying, water production from atmosphere, seasonal energy storage and adsorption cooling cycle (Restuccia et al., 2004;Saha et al., 2009;Gong et al., 2011).So far the mesoporous silica impregnated with calcium chloride (MPS-CC) has been exclusively studied (Aristov et al., 2002;Aristov, 2007;Zhang and Qiu, 2007;Tokarev et al., 2002;Xin et al., 2010;Glaznev et al., 2011;Ponomarenko et al., 2010;Espinosa et al., 2008;Glaznev and Aristov, 2008;Aristov et al., 2006;Aristov et al., 2008;Ovoshchnikov et al., 2011).This is because its sorption properties can be adjusted in a wide range by varying the porous structure of the host matrix, chemical nature of the impregnated salt, and the amount of the salt inside pores.However, the water uptake of such composite sorbent is limited by the amount of salt loading, which cannot be so high as to keep the corrosive salt inside the pores.On the other hand, superabsorbent polymer (SAP) has the ability to absorb the liquid water up to 1000 times its mass (Jin et al., 2000) but its sorption for water vapor is quite low.
This study was motivated with the expectation of the synergistic effect on the water vapor uptake when the two constituents were well blended.For such blended sorbents, water vapor would be taken by MPS-CC and then the water would migrate to SAP, until the SAP saturates with water.The study started with the preparation of such sorbents by simply mixing the MPS-CC as prepared with the SAP as received.The water sorption capacity and rate of such simple blends were investigated.Based on the results, new sorbents having the intimate contact between MPS-CC and SAP were developed to keep both the portion of SAP as low as possible and the water vapor sorption much higher than that of MPS-CC alone.The preparation was based on the controlled growth of MPS on swelling SAP.The mechanism of the SAP-assisted MPS formation was pursued by probing the changes in the composition and morphology of the blends prepared under various conditions.The CC impregnation and water uptake of the new sorbent was then investigated.

Preparation of mesoporous silica
The precursor solution was prepared for pure MPS with the molar ratio of TEOS:H 2 O:HCl:P123 = 1:195.8:6.09:0.0169.It was magnetically stirred for 24 hours at 35 °C for 24 h, aging at 90 °C for 24 h, filtering and finally calcining at 600 °C for 20 min to remove P123.

Impregnation of calcium chloride and simple mixing of MPS and SAP
One gram of MPS was then soaked in 100 mL of aqueous solution having 30 wt% calcium chloride (CC) with magnetic agitation for 24 h at room temperature.The impregnated MPS was filtered, washed three times with water and dried at 90 °C for 2 h.The average CC loading was 0.186 ± 0.12 g CC/g MPS.The CC-loaded MPS are called MPS-CC.For simple blending, SAP as received and MPS-pCC as prepared were simply mixed in the different mass ratios.These were called SAP/MPS-CC.The mass fraction of each constituent was often given in the parenthesis in the notation.For instance, SAP50/ MPS-CC50 represents the simply mixture having the equal mass of SAP and MPS-CC.

Sorbent preparation by the controlled growth of MPS on SAP and impregnation of calcium chloride
A s ol u t io n h av i n g H 2 O:HC l:P123 = 195.8:0.00609:0.0169was prepared.SAP was then added to the solution at 1 wt% with respect to the water in it.It took hours for the SAP to be well slurred.TEOS was then added to the slurry at TEOS:H 2 O molar ratio of 1:195.8.The molar composition of the final slurry was TEOS: H 2 O:HCl:P123 = 1:195.8:0.00609:0.0169plus SAP at 1 wt% of the water in it.The resulting solution was the same as that of the MPS precursor solution except both adding SAP and reducing the amount of the acid by 1/1000 (reference condition).The solution was stirred for 48 h (reference condition) at room temperature.In preliminary experiment, the amount of HCl varied as 1/10, 1/100 and 1/1000 times 6.09, the concentration used for the pure MPS preparation, both to control the rates of silica formation and make the SAP swelling.The time of preparation was also varied from 6 h to 144 h to search for the preparation mechanism.The specimen prepared so was called SAP-iMPS.Here, 'i' represents in-situ preparation.The preparation time in hours was often included in the abbreviation such as SAP-iMPS24.After some pre-experiments the CC impregnation to the blends was carried out with ethanol solutions to prevent the salt from washing out.The impregnation took place in 30 wt% CC ethanolic solution with magnetic stirring for 24 h at room temperature.The sorbents were filtered, washed three times with ethanol and dried at 90 °C for 24 h.The specimen was then called SAP-iMPS-CC.

Characterization
The porous structures were observed by Transmission Electron Microscope (TEM, JEM-2000EX, JEOL) and Scanning Electron Microscope (SEM, Hitachi, S-3400N) with Energy Dispersive X-ray Spectroscopy (EDX), and analyzed by Small Angle X-ray Spectroscopy (SAXS, GADDS, Brucker).The silica formation and CC impregnation were measured by Inductively Coupled Plasma-Atomic Emission Spectrometer (ICP-AES, Jobin Yvon, JY-Ultima-2) and the EDX.In order to see the effect of the CC impregnation precisely, it is measured every time the CC impregnation was carried out, instead of using a calibration chart.X-ray Photoelectron Spectroscopy (XPS, PHI 5000 VersaProbe, ULVAC-PHI) was used to analyze the surface chemical state of the sorbents.Surface silica complex was identified by 29 Si Solid-State Nuclear Magnetic Resonance Spectrometer (NMR, Bruker, Bruker Avance II).The relative proportions of each species were obtained by Magic Angle Spinning (MAS) associated with 1 H→ 29 Si cross-polarization (CP/MAS).

Water sorption in porous structures
Fifty milligram of the sorbents was spread on petri dish in 1-L chamber at 17 °C.The humid air at 30 °C and 28.9 % relative humidity entered the chamber at 1 L/min.These conditions are usually obtained in the adsorber belonging to a typical adsorption chiller (Restuccia et al., 2004;Saha et al., 2009).The water uptake of the sorbents was measured by microbalance up to 3 h.

Sorbent prepared by simple mixing of SAP and MPS-CC powders
Fig. 1 shows the FE-SEM image of the sorbent made by simply mixing ~100-μm SAP powder as received with ~1-μm MPS-CC powder as prepared in the same mass.Lumps of the small MPS-CC particles were dispersed among large SAP powders with some segregation.Fig. 2 shows the isotherms of both pure MPS-CC and SAP in humid air at 17 °C.It is confirmed that SAP has much lower sorption for water vapor than MPS-CC particularly in the low to moderate humidities.It is noted that MPS-CC used did not have unique CC loading but this could not affect the comparison between the two materials.In Fig. 3 the specific water uptake of the simply mixed sorbents was plotted with respect to sorption time at 17 °C and 75 % relative humidity, as a parameter of the mass fraction of SAP in the sorbent.The water uptakes of pure SAP and MPS-CC were already saturated at 0.0285 and 0.184 g H 2 O/g sorbent, respectively in the first 1 hour.The water uptakes of the sorbents, SAP/MPS-CC, were still increasing even at 3 h, irrespective of SAP content.The water uptake recorded the maximum at 75 mass% of SAP.The water uptakes of the sorbent for 3 h were 0.658 g H 2 O/g sorbent, respectively, while those of pure MPS-CC were 0.123, 0.184 and 0.184 g H 2 O.The water uptake of the mixed sorbent reached 3.5 times the equilibrium water uptake of the pure MPS-CC in the first 3 h and the difference will increase with the time.For these sorbents the segregation between MPS-CC and SAP would become minimized in which it was expected that small MPS-pCC particles would well surround each polymer particle in a thin layer.These clearly supported that water vapor would be taken up by MPS-CC and the water would then migrate to SAP.The sorption-migration process would continue until SAP saturates with water.However, as shown in Fig. 3, there were some sorbents whose water uptakes for 3 h were lower than the corresponding values of pure MPS-CC.They were the mixtures having a small amount of either SAP or MPS-CC, which suffered the migration resistance due to the poor contact between MPS-CC and SAP.For example, in case of SAP10/ MPS-CC90 its early water uptakes were lower than those of pure MPS-CC.The sorption would momentarily reduce since, in addition to the decrease in the portion of MPS-CC, the water sorbed in MPS-CC is hard to find the SAP for migration.Of course, it would be a matter of time that the sorption of the sorbents surpasses that of pure MPS-CC.Nevertheless, it is required to enhance the early water uptakes since the usual adsorption period in a typical adsorption chiller is no longer than 20 min.Therefore, the additional mass transfer resistance caused by the migration from MPS-CC to SAP should be reduced.The use of SAP can give a rise to the environmental problems.Since the conventional adsorption chiller has been using silica gel or zeolites as adsorbents, it is not so desirable to replace these inorganic materials with the polymer.More than these, it is highly recommended to enhance the efficiency of SAP usage in short period of adsorber operation.Therefore, it is better to include SAP as small as possible.

Sorbent preparation by assisted growth of MPS on SAP swelling
The simply mixed sorbents could not afford to satisfy such demands of keeping both the water uptake as high and the SAP fraction as low as possible.Therefore, it is  necessary to develop a new structure of the sorbents in which MPS-CC is well exposed outside to water vapor, and the intimate contact is maintained between MPS and a minimal amount of SAP.The desirable structure was aimed by the controlled growth of MPS on the surface of the swelling SAP.We will discuss separately the effect of hydrochloric acid concentration and growth time on the morphology of the new hybrid sorbents.

Effect of HCl concentration
Hydrochloric acid plays an important role as a catalyst to speed up the formation of the hydrolysis of TEOS.However, in the low pH the sodium carboxylate group on the polymer network was protonated.This in turn decreased the degree of ionization and the charge density on the network hence decreasing the swelling ratio (Elliott, 2004).In order to swell SAP and then grow MPS on it, it is necessary to keep pH low at the expense of reducing the rate of MPS formation.At the HCl/TEOS of 6.09, the reference ratio for the pure MPS, SAP did not swell at all, but both the SAP precipitation and silica phase formation immediately took place.As the acid-to-TEOS molar ratio decreased to 0.000609, 1/1000 of the reference value, SAP began to swell (be slurried) in water while the silica was formed slowly throughout the slurry in milky appearance.Fig. 4 shows the SAXS data on the SAP-iMPS prepared at the different HCl concentrations.A peak representing mesophase first appeared for SAP-iMPS-1/100.For SAP-iMPS-1/1000, a peak 2θ = 1.04° clearly shows the evidence of a weak mesophase.At low pH, the sodium ions are liberated by protonation of the carboxyl group (Elliott, 2004).The effect of NaCl on the formation of silica has been studied widely before (Pan et al., 2007;Lin et al., 2001;Yu et al., 2001).The hydrothermal stability of mesoporous materials after calcination can be directly enhanced considerably by adding a small amount of the additional salt.However, The mesophase transformations occur in the sequence MCM-41→KIT-1→MCM41→KIT-1→amorphous phase as the salt concentration increases.At high concentration of NaCl the regularity of the mesoporous materials was lost due to the perturbation of the double layer potential (Yu et al., 2001).When SAP was fully combusted with oxygen, the combustion residue was obtained at 40 wt% of the initial SAP mass.By EDX, the half of the residue belonged to Na.This strongly justified the amorphous nature of the silica obtained at HCl/TEOS greater than 0.0609.On the other hand, the sodium ions become less free and more bound to the polymer network and the SAP begins to swell at low acid concentrations.Mesoporous silica would grow under this condition.

Effect of preparation time
Our reference SAP-iMPS was therefore prepared at HCl/TEOS = 0.00609.Now it was necessary to know deeply how the silica formed on SAP and how the morphology of SAP-iMPS evolved with the time of preparation.The time evolution of their morphology is shown in Fig. 5. Silica was gradually being deposited on SAP as time went by.Fig. 6 shows the magnified FE-SEM and TEM images of MPS in SAP-iMPS-144.The mesophase was not observed in the SEM image until 48 h, after which the mesophase grew organizing, straight and lengthwise.The extremely organized, straight and long mesophases appeared in 144 h as shown in Fig. 5(d) and Fig. 6.Each long mesophase was composed of many units of the MPS segment glued at the ends like bamboo joint.It has been reported that the salts such as NaCl act as a bridging ion between two colloidal silica particles (Pan et al., 2007;Barrabino, 2011).The small Na + , usually surrounded by the oxygen atoms of six water molecules of hydration, is adsorbed on the silica surface.One or more of the oxygen atoms of hydration water are then displaced by the oxygen of the surface silanol groups.This directly links the two silica particles via Na + ions.Therefore, it is highly probable the sodium from the SAP glued the ends of the mesoporous cylindrical segments to help them grow in length.The wall-to-wall distance is about 10 nm as shown in Fig. 6.Fig. 7 shows the mass of silica deposited on SAP for the different times of preparation.The silica deposition took place rapidly up to 6 h, then slowed down to 72 h and finally accelerated.Fig. 8 shows the SAXS data on SAP-iMPS growing with the different ages.SAP-iMPS younger than 48 h showed no meaningful peaks since the one at 0.5° was considered as that inherent to the machine itself.Clear mesophase peak appeared for SAP-iMPS-96.The mesophase grew further thereafter at the d value of 11 nm unchanged.The d value was coincided with the wall-to-wall distance observed in the TEM image as described before but larger than that of MPS-CC (~8.9 nm).This reflects the two mesophase were generated in the different environments.Fig. 9 shows the isotherms of SAP-iMPS with the different ages obtained from BET analysis.Table 1 shows the results of BET analysis on them.The youngest, SAP-iMPS-6, belonged to a typical H4 hysteresis loop classified by IUPAC (Thommes, 2010), indicating complex materials containing both micropores and mesopores.SAP-iMPS-48 was classified in between H1 and H3 but biased to H3.In other words, it was composed of mostly non-rigid aggregates of plate-like particles or assemblages of slit-shaped pores while it had a trace of big cylindrical pores.Young SAP-iMPS had much lower BET surface area than the pure MPS (~375 m 2 /g) since they were low in silica deposition and in the early stage of mesophase formation.SAP-iMPS-144 was also in between of H1 and H3, but now had a lot of smaller cylindrical pores and some plate-like particles.The appearance of H3 type would be related with the layered growth of silica on the surface of SAP.It is noted that SAP-iMPS-48 had higher pore volume and  larger average pore diameter than SAP-iMPS-144.It is said that H3 types in principle should not be expected to provide a reliable assessment of either pore size distribution or the total pore volume (Thommes, 2010).However, these findings indicate that the mesophase became matured in the period from 48 to 144 h.This was reflected by the slope of silica deposition with respect to the synthesis time as shown in Fig. 7.After the active deposition of silica to 12 h, its deposition slowed down during the creation of mesophase to 72 h and then the mesophase grew rapidly lengthwise accompanying with the accelerated silica deposition.Table 2 shows the proportions of Q2 (Si(OSi) 2 (OH) 2 ), Q3 (Si(OSi) 3 (OH)) and Q4 (Si(OSi) 4 ) on the silica surface groups obtained from 29 Si NMR spectroscopy.Surface silanol (-SiOH) reduced gradually as the preparation time increased.This would be in relevancy to the exhaustion of silanol groups by sodium ions, described before.

Calcium chloride loading
In preliminary experiments the loading of calcium chloride on SAP-iMPS-48 was tried from its aqueous solution.However, the salt loading was found quite low.When the SAP-iMPS were immersed in the CC aqueous solution, the bulk water would direct to SAP through the   mesoporous phase, which would bring about high convectional flow of the salt solution.Therefore, the salt could not help moving to SAP with the water, having little chance to diffuse into mesoporous silica.Therefore, the salt would only smear on the SAP surface and be easily removed by the subsequent water washing.Since SAP was confirmed not to sorb ethanol in our separate experiment, water was replaced by ethanol as an alternative solvent for the calcium chloride.Fig. 10 shows two curves on CC loadings of the different aged SAP-iMPS in the ethanol: one is the loading per unit mass of SAP-iMPS itself and the other is the loading per unit mass of MPS deposited on SAP.The salt loading increased with the preparation time up to 6 h accordingly with the silica deposit, and then varied very slowly.The specific salt loading calculated based on the silica deposit increased up to 12 h and then even slightly decreased.Such an almost invariant salt loading would be related with the decrease in Q2 and Q3 of silanol groups and thus the formation a bamboo joints between mesoporous segments.In glass the addition of 25 % sodium oxide (soda, Na 2 O) to silica reduces the viscosity and lowers the melting point from 1723 °C to 850 °C.Sodium oxide also increases the tendency of silicon dioxide to form networks rather than crystals.The sodium-oxygen bridges may interrupt the regular silicon-oxygen bonding and/or sodium ions may intersperse among the silica molecules to prevent the formation of regular crystals (Best, 1990).Like the glass, the inclusion of the sodium ions was believed to lower the percentage of the silanol groups available to the loading of calcium chloride.In these SAP-iMPS, the increase in the surface area made up for this silanol group loss.Therefore, the salt loading resulted in a slight increase with the time of preparation

Water uptake
Fig. 11 shows the dynamic sorption curves for SAP, SAP75/MPS-CC25 and SAP-iMPS-CC aged for 48 h at 17 °C and 75 % RH.It is noted that the sorptions of the latter two were still increased at 2 and 3 hours, respectively.As shown in Fig. 11, SAP-iMPS-CC48, consisting of 44 % SAP and 56 % SiO 2 , had much more sorption for water vapor than SAP75/MPS-CC25 for given sorption time.Water uptake was also measured under adsorption chiller condition.Fig. 12 shows the variation in the 3-hour specific water uptakes of SAP-iMPS-CC with respect to preparation time in the adsorption chiller condition (30 °C and 28.9 % relative humidity).Since mesophase of silica was about to form for SAP-iMPS-CC grown for 24 h, the specific water uptake of the sorbent remained almost constant from 24 h to 144 h.It is realized that SAP-iMPS-CC-144 has 75 % SiO 2 and 25 % SAP.This is contrasted with the result of SAP/MPS-CC where the maximum water uptake occurs with SAP75/ MPS-CC25.As also shown in Fig. 12, all the specific water sorptions of SAP-iMPS-CC grown for more than 24 h, based on CaCl 2 , were higher than 3.5 g H 2 O/g CaCl 2 ,  in it with respect to preparation time under adsorption chiller condition (30 °C and 28.9 % relative humidity).
which was much higher than the water uptake of bulk CaCl 2 of 1.22 g H 2 O/g CaCl 2 (Elliott, 2004).This also strongly supported that the sorption of SAP-iMPS-CC was not limited by the CaCl 2 loading in MPS since water vapor was sorbed by the salt and subsequently migrated to SAP.In addition, the template growth of MPS on SAP kept the contact between MPS and SAP in good condition for water migration, even though the SAP mass% decreased up to 25 %.As described above, CaCl 2 loading was almost constant, irrespective of the preparation time, so was the initial sorption, since the SAP in the sorbent could never be saturated for 3-hour sorption.In other words, the initial water uptake of the sorbent was enhanced, due to the intimate SAP-MPS contact.

Conclusion
MPS-CC sorbs water vapor via CC in its pores, but its water uptake is determined by the amount of the salt loading, which has to be limited due to its corrosiveness.On the other hand, SAP is extremely high sorbent for liquid water but its uptake for water vapor is very low.Maximum 3-hour water uptake of the sorbents obtained by simple powder blending of the two reached more than 3 times that of the pure MPS-CC at 75 mass% of SAP.It is believed that the water vapor was taken by calcium chloride in MPS-CC and then migrated to SAP.In both the technological and environmental viewpoints, SAP addition should be as small as possible by its efficient use.Therefore, it is required to reduce the resistance of migration from MPS-CC to SAP through the intimate contact between MPS.New preparation method was developed in which the MPS grew slowly at low acid concentration on the surface of the swelling SAP and the salt was subsequently loaded from its ethanol solution.Silica deposited rapidly in the first 6 h and then linearly with the preparation time.Mesophase began to grow after 48 h.The mesoporous silica grew well ordered, straight and lengthwise by the help of sodium ions.However, salt loading was not so improved due to the unfavorable surface condition.Therefore, the MPS grown on SAP template for more than 24 h had almost constant specific water uptake.Nevertheless, the sorption reached two times that of SAP75/ SAP25.This enhancement of SAP utilization was achieved by the considerable improvement of the contact between MPS and SAP.

Fig. 4
Fig.4SAXS data on SAP-iMPS prepared with different HCl concentrations.

Fig. 8
Fig. 8 SAXS data on reference SAP-iMPS-pCC prepared in the different reaction times.

Fig. 10
Fig. 10 Specific salt loadings based on g sorbent and g silica in it with respect to time of preparation.

Table 2
Proportions of Q2, Q3, and Q4 (estimated from 29 Si NMR) for SAP-iMPS-CC prepared for different times