Porous Resorcinol-Formaldehyde Resins

Porous ion-exchanging resorcinol-formaldehyde resins were prepared by removing an inorganic template (CaCO3) and by polymerization of a highly concentrated emulsion with toluene as a dispersing phase. As compared to original resorcinol-formaldehyde resins, the porous one is characterized with higher values of internal diffusion, providing exceptional purification coefficients. However, the amounts of the introduced CaCO3 and toluene have to be below 10 wt.% and 25 wt.%, respectively, to avoid fast degradation of the ion-exchanger. Under dynamic conditions, average purification coefficients in the model solution of Cs-137 are twice higher than those of the non-porous sample. The prepared porous ion-exchange resin can be used in cases when high Cs-137 uptake from liquid media should be realized at increased rates of solution percolation.


Introduction
One of the main issues of advanced radiochemistry is related to "nuclear legacy" problems.Particularly, plutonium production in the USA has yielded huge amounts of heterogeneous liquid radioactive wastes (LRW) (around 200 million L) [1], stored at the Savannah River Site and Hanford Site.Such LRWs contain mainly Na + (5-6 M), NO 3 − (1.5-3.0M), and OH − (0.5-2.4 M) [2][3][4], while their activity is caused by Cs-137 and Sr-90 [3].FSUE "PO Mayak" stores similar to LRW of the following composition: Na + -1-3 M, OH − -1-3 M, with the presence of Fe, Ni, Cr, and Al in the form of hydroxides and suspended matter [1,5].LRW are usually immobilized into solid matrices, generally, concrete ones that, however, lead to increased amounts of dumped wastes.Cs-137 can be preliminary removed from evaporator concentrates using selective sorption to reduce the amount of processed wastes with the small volume of the formed radionuclide concentrate [6].
For selective removal of Cs-137, two main groups of materials are widely used: Granulated and composite transition metal (Ni, Co, Cu) ferrocyanides [7][8][9] and resorcinol-formaldehyde resins (RF-resins) [10,11] related to the phenol-formaldehyde thermoreactive group of resol-type polymers.The main disadvantage of transition metal ferrocyanide sorbents is a low chemical stability in alkaline media [12,13], which results in fast degradation of the materials.Resorcinol-formaldehyde resins are characterized with greater chemical stability to alkaline solutions and can selectively remove Rb and Cs at high background concentrations of Na + [14].RF-resins are used mainly under dynamic conditions, when ion-exchanger is put into the sorption column with consequent feeding of LRW [15].RF-resins were also used in the form of spheres to treat highly mineralized LRW followed by radionuclide desorption using HNO 3 solution [2,[16][17][18].
Unfortunately, RF-resins are oxidized easily, which negatively impacts on their sorption properties [19].Arm et al. [20] reported on the effect of RF-resins storage on their sorption characteristics.Dry Na-form resin storage for 6 months led to a 1.8-fold reduction of the distribution coefficient [20].Oxidation occurs mainly on methylene bridges and resorcinol, leading to quinoid structures that results in a reduction of the total amount of functional groups [21].Shelkovnikova et al. [22] assumed that oxymethylene bridges and methylol groups were oxidized to carboxyl groups that are non-specific to Cs + ions, but increase the overall capacity of the ion-exchanger.Brown el al. [16] reported the data on gradual degradation of the RF-resin under dynamic sorption conditions in the repeated sorption-regeneration cycle.Arm et al. [23] provide information on the resin selectivity reduction after multiple treatments with NaOH and HNO 3 .
The RF-resin lifetime for a high pH LRW treatment can be prolonged by the increase of intragranular mass-transfer.The latter will provide faster percolation of purified solution with purification coefficients remaining on a high level that will reduce processing times and ion-exchanger degradation.One of the ways to increase intragranular diffusion is to develop porosity by removing the inorganic template in the form of CaCO 3 [24] or by polymerizing the emulsion followed by dispersing phase removal [25].
There have been attempts to obtain porous spherical RF-resins by adding olygomers containing CaCO 3 particles to hot condenser oil followed by solidification at 120 • C, yet the final product was chemically unstable in alkaline media due to low polymer cross-linking [26].In addition, there are no works related to successful synthesis of porous RF-resins selective to Cs-137 using polymerization of dispersed phase followed by the dispersion phase removal.
The present paper is devoted to studies of the preparation of porous RF-resins via two routes based on solidification at high temperatures, yielding alkali-resistant materials with exceptional kinetic characteristics of ion-exchange process.

Materials and Methods
Resorcinol, toluene, sodium hydroxide, nitric acid, formaldehyde, calcium carbonate, and sodium nitrate of chemically pure grade were purchased from Neva Reactiv LLC and used without purification.Cs-137 radionuclide in the solution of 1 M HNO 3 was purchased from A.I. Leypunsky Institute for Physics and Power Engineering.
Initial RF-resin denoted as RFR-i was obtained according to the method described in [26].RFR-i (initial resorcinol formaldehyde resin) sample was prepared at a molar resorcinol-formaldehyde ratio of 1:1.Resorcinol was dissolved in 6 M KOH, the obtained solution was added with 2.2 M formaldehyde solution under continuous stirring and cooling to 70 • C. Polymerization occurred in 10-15 min with the formation of a dense ruby-colored gel, which was ground to 0.6-2.0mm particles and washed with deionized water under vacuum.Resin solidification was carried out at 200-210 • C for 6 h in air, then it was ground and washed with 1 M solution of HNO 3 .
Samples denoted as "porous" were prepared by various routes.The first synthesis route included the following procedures.15 mL of 2.2 M formaldehyde solution was added to 10 mL of water solution containing 5.5 g of resorcinol and 2.8 of KOH under cooling.3 and 7 g of CaCO 3 powder was introduced into the obtained olygomeric mixture to obtain RFR-10 and RFR-25, respectively.Then, the mixture was heated till the polymerization started.The following day, the resins were annealed at 210 • C for 6 h in air.Thereafter, resins were ground and the fraction, 0.5-1.0mm, was washed with 1 M HNO 3 solution.
The second synthesis route encompassed the following steps.15 mL of 2.2 M formaldehyde solution was added to 10 mL of aqueous solution containing 5.5 g of resorcinol and 2.8 of KOH under cooling.Thereafter, 15 mg of sodium dodecylsulfate was added along with 7.5, 10.5, or 19.5 mL of toluene to obtain RFR-25t, RFR-35t, and RFR-65t resins, respectively.The mixture was heated till the polymerization started.The following day, resins were solidified at 210 • C for 6 h in air.After solidification, resins were ground and the fraction, 0.5-1.0mm, was washed with 1 M HNO 3 solution.
Sorption properties were assessed using the model solution of the following composition: NaNO 3 -1.25 M, NaOH-0.75M. Kinetic characteristics of the ion-exchange process were determined using the model solution of Cs-137 radiolabel (500-1000 Bq mL −1 ).The Cs-137 activity in solution was measured by the direct radiometric method using an RKG-AT1320 γ-radiometer with an NaI(Tl) detector 63 × 63 mm (NPP Atomtech, Minsk, Belorussia).
Ion-exchange resins were put in contact with model solution under continuous stirring using an orbital shaker with an amplitude of 10 mm and a rotation speed of 120 rpm, V/m ratio = 1000 mL g −1 .Solution samples were taken after certain time periods and the residual activity was measured.
The Cs-137 sorption (in %) from the model solution under static conditions was evaluated according to Formula (1): where A t is the activity of the model solution at time t (Bq mL −1 ), and A 0 is the initial activity of the model solution (Bq mL −1 ).
To determine the limiting stage and to quantify the kinetic parameters of the ion-exchange process, we used the Boyd-Adamson Equation ( 2) [27] that allows calculation of the effective diffusion coefficient: where Q t is the residual sorbent's activity with respect to radionuclide at time t (Bq mL −1 ), Q max is the residual sorbent's activity with respect to Cs-137 at maximal sorption (Bq mL −1 ), and B is the diffusion rate constant exerted of Equation ( 3): where D i is the effective diffusion coefficient (cm 2 min −1 ), and r is the ion-exchanger's grain radius (cm).
The dependence of Bt on F follows Equation ( 4) and is presented in table values by Richenberg in [28]: where Bt is the Fourier's homochronicity criterion.
B was found as a tangent of the line slope obtained at approximation of the first values on the kinetic curve (at F < 0.5) in coordinates, Bt = f (t).Using B values, we calculated D i using Equation (3).
It is noteworthy that porous resins exhibit desorption due to chemical instability followed by coloring of the model solution into yellow of a dark-orange color.Therefore, in each case, the Q max value was taken as a maximal sorption of Cs-137 ( max ), after which the sorbent decomposed.The half-exchange period (t 1/2 ) was calculated by Formula (5) [29]: The Cs-137 sorption under dynamic conditions was performed as follows. 1 mL of sample fraction, 0.5-1.0mm, was put into the column with an inner diameter of 1 cm.Model solution was fed into the column at a rate of 50 mL h −1 .The filtrate was collected by fractions and the residual activity was identified.After 300-350 mL of model solution were fed to the column, the experiment was stopped and the resin was removed and washed with 1 M HNO 3 solution under static conditions to remove Cs-137, while the completeness was assessed by the residual eluate's activity.Then, the resin was washed with distilled water and the sorption was repeated.The Cs-137 removal efficiency under dynamic conditions was assessed using values of the purification coefficient (K pur ) calculated as follows: where: A 0 is the initial solution activity (Bq mL −1 ), and A i is the filtrate's activity (Bq mL −1 ).Surface morphology was investigated by the method of scanning electron microscopy (SEM) using a Carl Zeiss Crossbeam 1540-XB device (Oberkochen, Germany).

Characterization of Ion-Exchange Resins
SEM images of RF-resins are presented in Figure 1.Porous samples have a pronounced sponge morphology as compared to a smooth surface of the initial resin (Figure 1a).Visually, porous structure is comprised of regular-shape macropores of average sizes of 10-30 µm.Samples obtained by CaCO 3 removal contain pores of irregular shapes, while particles have a lot of defects and cracks formed probably by CO 2 liberation during acid treatment.The pore number increased with the amount of added template, while the pore size did not change (Figure 1b,c).Samples prepared by route 2 have no visible defects, pores have a smooth surface, and their number increases with the amount of toluene introduced during synthesis (Figure 1d-f).Addition of toluene increases the average pore size: For RFR-25t and RFR-35t, it is 10-25 um, where for RFR-65t, it is 20-40 um.Despite defects and pores, all resin samples manifest good mechanical stability comparable with RFR-i.Surface morphology was investigated by the method of scanning electron microscopy (SEM) using a Carl Zeiss Crossbeam 1540-XB device (Oberkochen, Germany).

Characterization of Ion-Exchange Resins
SEM images of RF-resins are presented in Figure 1.Porous samples have a pronounced sponge morphology as compared to a smooth surface of the initial resin (Figure 1a).Visually, porous structure is comprised of regular-shape macropores of average sizes of 10-30 µm.Samples obtained by CaCO3 removal contain pores of irregular shapes, while particles have a lot of defects and cracks formed probably by CO2 liberation during acid treatment.The pore number increased with the amount of added template, while the pore size did not change (Figure 1b,c).Samples prepared by route 2 have no visible defects, pores have a smooth surface, and their number increases with the amount of toluene introduced during synthesis (Figure 1d-f).Addition of toluene increases the average pore size: For RFR-25t and RFR-35t, it is 10-25 um, where for RFR-65t, it is 20-40 um.Despite defects and pores, all resin samples manifest good mechanical stability comparable with RFR-i.

Sorption Kinetics
Figure 2 provides the initial kinetic curves of the Cs-137 sorption from model solution in semilogarithmic coordinates, which demonstrate that at the beginning stage (10-120 min), the ionexchange process occurs faster on porous resins as compared to RFR-i.RFR-25, RFR-35t, and RFR-65t samples, characterized by increased porosity, while the best sorption performance is realized during the first 120 min.However, two days of permanent contact of RFR-25 and RFR-65t with model solution leads to coloring into yellow-brown and considerable degradation of sorption (S%) related to gradual oxidation and polymer dissolution.There is a bend at the maximum point on the kinetic sorption curves, the curvature of which and, therefore, the degradation rate increase along with the increase of the porosity of the ion-exchanger.
The calculated values of Di and t1/2, as well as correlation coefficients of linear regression, are given in Table 1.As compared to RFR-i, porous resins show higher Di values and lower t1/2 ones, indicating higher rates of mass transfer.CaCO3 removal from the polymer matrix increases Di more than 1.5-fold, while ion-exchangers obtained via the second route exhibit only a 1.2-1.4-foldincrease

Sorption Kinetics
Figure 2 provides the initial kinetic curves of the Cs-137 sorption from model solution in semi-logarithmic coordinates, which demonstrate that at the beginning stage (10-120 min), the ion-exchange process occurs faster on porous resins as compared to RFR-i.RFR-25, RFR-35t, and RFR-65t samples, characterized by increased porosity, while the best sorption performance is realized during the first 120 min.However, two days of permanent contact of RFR-25 and RFR-65t with model solution leads to coloring into yellow-brown and considerable degradation of sorption (S%) related to gradual oxidation and polymer dissolution.There is a bend at the maximum point on the kinetic sorption curves, the curvature of which and, therefore, the degradation rate increase along with the increase of the porosity of the ion-exchanger.
The calculated values of D i and t 1/2 , as well as correlation coefficients of linear regression, are given in Table 1.As compared to RFR-i, porous resins show higher D i values and lower t 1/2 ones, indicating higher rates of mass transfer.CaCO 3 removal from the polymer matrix increases D i more than 1.5-fold, while ion-exchangers obtained via the second route exhibit only a 1.2-1.4-foldincrease of D i .Higher mass of the introduced calcium carbonate or toluene expectedly reduces the time of half-exchange.Based on the obtained kinetic curves, the Bt = f(t) values were plotted at F < 0.5 in Figure 3.The obtained experimental data were fitted by linear functions using the SigmaPlot software.One can see that the slope of the linear regression within the 60 min of sorption increases along with the introduced amounts of CaCO3 and toluene.Ion-exchange resins were tested under dynamic sorption conditions to prove their efficiency.Figure 4 demonstrates the curves of Cs-137 sorption from the model solution.During sorption, when  Based on the obtained kinetic curves, the Bt = f (t) values were plotted at F < 0.5 in Figure 3.The obtained experimental data were fitted by linear functions using the SigmaPlot software.One can see that the slope of the linear regression within the 60 min of sorption increases along with the introduced amounts of CaCO 3 and toluene.Based on the obtained kinetic curves, the Bt = f(t) values were plotted at F < 0.5 in Figure 3.The obtained experimental data were fitted by linear functions using the SigmaPlot software.One can see that the slope of the linear regression within the 60 min of sorption increases along with the introduced amounts of CaCO3 and toluene.

Sorption of Cs-137 under Dynamic Conditions
Ion-exchange resins were tested under dynamic sorption conditions to prove their efficiency.Figure 4 demonstrates the curves of Cs-137 sorption from the model solution.During sorption, when ion-exchangers transform into the salt form, their volume increases by 20%-30% on average.When resins are transferred into the H-form by washing with HNO 3 , their volume reduces to the original value.The purification coefficient, except Figure 3d, passes through the maximum followed by exponential decay, which was not caused by resin degradation, but arose from specially chosen conditions reducing the experimental time.Gradual increase of the purification coefficient with the number of the sorption cycles is caused by ion-exchanger transition into the working regime, which is peculiar for cesium removal under dynamic conditions.This effect can be attributed to the removal of K + ions having a competing effect on cesium sorption [23].
Colloids Interfaces 2019, 3, x FOR PEER REVIEW 6 of 9 ion-exchangers transform into the salt form, their volume increases by 20%-30% on average.When resins are transferred into the H-form by washing with HNO3, their volume reduces to the original value.The purification coefficient, except Figure 3d, passes through the maximum followed by exponential decay, which was not caused by resin degradation, but arose from specially chosen conditions reducing the experimental time.Gradual increase of the purification coefficient with the number of the sorption cycles is caused by ion-exchanger transition into the working regime, which is peculiar for cesium removal under dynamic conditions.This effect can be attributed to the removal of K + ions having a competing effect on cesium sorption [23].

Discussion
Kinetic sorption curves (Figure 2) indicate that when the experimental time increases to 2 days under static conditions, RFR-i has the highest chemical stability to alkaline media, while RFR-10 and RFR-25 resins are less stable.The RFR-25 and RFR-65t resins characterized with an enhanced porosity due to high amount of inorganic template or toluene are unstable and depolymerize rapidly with the sorption efficiency decrease that is implicitly proved by an intense coloration of the model solution with oligomer products of degradation.
Experimental data in Figure 3 obtained by processing kinetic sorption curves are well fitted with linear functions, with R 2 > 0.95, indicating that the kinetics of the ion-exchange process on RF-resin is dominated by mass-transfer inside the resin grain.Therefore, the porosity increase yields the increase of Di and the decrease of t1/2 values as compared to the non-porous sample (RFR-i) (Table 1).The decreased values of diffusion coefficients for resins obtained using toluene can be explained by closed macropores inside grains that have no contact with the adsorptive, while preparation of samples using CaCO3 removal with acid leads to CO2 liberation, with the defects and caverns reaching the surface of the grains formed.

Discussion
Kinetic sorption curves (Figure 2) indicate that when the experimental time increases to 2 days under static conditions, RFR-i has the highest chemical stability to alkaline media, while RFR-10 and RFR-25 resins are less stable.The RFR-25 and RFR-65t resins characterized with an enhanced porosity due to high amount of inorganic template or toluene are unstable and depolymerize rapidly with the sorption efficiency decrease that is implicitly proved by an intense coloration of the model solution with oligomer products of degradation.
Experimental data in Figure 3 obtained by processing kinetic sorption curves are well fitted with linear functions, with R 2 > 0.95, indicating that the kinetics of the ion-exchange process on RF-resin is dominated by mass-transfer inside the resin grain.Therefore, the porosity increase yields the increase of D i and the decrease of t 1/2 values as compared to the non-porous sample (RFR-i) (Table 1).The decreased values of diffusion coefficients for resins obtained using toluene can be explained by closed macropores inside grains that have no contact with the adsorptive, while preparation of samples using CaCO 3 removal with acid leads to CO 2 liberation, with the defects and caverns reaching the surface of the grains formed.
Sorption characteristics of ion-exchangers under dynamic conditions depends greatly on the synthesis conditions (Figure 4).Purification coefficients should increase with kinetic characteristics and porosity increase due to peculiarities of the ion-exchange process; however, this does not hold true in our case.As under static conditions, this phenomenon is related to the resin degradation despite the short time of contact with the model solution.Besides, the radionuclide removal efficiency depends on the way the porosity is formed.
The RFR-65t sample is characterized with high porosity, exceptional D i , and low values of t 1/2 , yet it does not retain Cs-137 after the first cycle due to degradation and it cannot be applied further as an ion-exchanger in highly-mineralized alkaline media (Figure 4d,e).The radionuclide sorption on RFR-25 and RFR-35t in second and third cycles greatly reduces due to sorbent's degradation and becomes comparable to RFR-i (Figure 4b,c,e,f), whereas the latter is more chemically stable at low mass transfer rates inside grains (Table 1).
Among all prepared ion-exchange resins, the highest purification coefficients in three consecutive cycles were demonstrated by RFR-10 and RFR-25t (Figure 4).The latter proves that there is an optimal amount of inorganic template or disperse phase, which have a positive effect on kinetic characteristics of ion exchange, overcoming the negative effect of polymer dissolution.Such resins can be successfully used for Cs-137 removal under dynamic conditions without risks of losing the sorption performance.

Conclusions
Porous RF-resins were prepared by introduction of inorganic (CaCO 3 ) and organic (toluene-based emulsion) templates into the resorcinol-formaldehyde mixture.Amounts of introduced CaCO 3 and toluene were optimized to provide high sorption characteristics in alkaline medium and low rates of polymer degradation.The synthesized resins can be used in cases when high efficiency of Cs-137 removal from liquid medium should be maintained at high linear rates of solution feeding.

Figure 3 .
Figure 3. Bt = f(t) functions at F < 0.5; (a)-samples obtained via the first route; (b)-samples obtained via the second route.

Figure 2 .
Figure 2. Kinetic curves of Cs-137 sorption in semi-logarithmic coordinates; (a)-samples obtained via the first route; (b)-samples obtained via the second route.

ColloidsFigure 2 .
Figure 2. Kinetic curves of Cs-137 sorption in semi-logarithmic coordinates; (a)-samples obtained via the first route; (b)-samples obtained via the second route.

Figure 3 .
Figure 3. Bt = f(t) functions at F < 0.5; (a)-samples obtained via the first route; (b)-samples obtained via the second route.

Figure 3 .
Figure 3. Bt = f (t) functions at F < 0.5; (a)-samples obtained via the first route; (b)-samples obtained via the second route.

Figure 4 .
Figure 4. Cs-137 removal from the model solution under dynamic conditions; (a-c)-first, second, and third sorption cycle of the resins obtained via the first route; (d-f)-first, second, and third sorption cycle of the resins obtained via the second route.

Figure 4 .
Figure 4. Cs-137 removal from the model solution under dynamic conditions; (a-c)-first, second, and third sorption cycle of the resins obtained via the first route; (d-f)-first, second, and third sorption cycle of the resins obtained via the second route.
Sample R

Table 1 .
D i and t 1/2 coefficients calculated using Boyd's equation.