Introduction

Soluble uranium (VI) compounds occur in sources of water supply (ground water, lakes, and rivers) as a results of leaching from natural deposits, release in mill tailings, emissions from nuclear power plants, combustion of coal or other fuels, and even due to use of the phosphate fertilizers (Alirezazadeh and Garshasbi 2003; Hoover et al. 2017; Orloff et al. 2004). Maximum permissible concentration for soluble compounds of uranium is 0.015 mg dm−3. Lower values are also recommended (Brine 2010), since this metal is not only radioactive, but also toxic.

To remove small amounts of U(VI) from water, adsorption methods, particularly ion exchange, are preferable (Awwad 2018; Everett 1998). As opposite to other techniques, such as chemical (Ling et al. 2015), photocatalytic (Liu et al. 2018), or biological (Gilson et al. 2015) reduction of U(VI) to U(IV) (formation of insoluble UO2), adsorption method leads to no formation of finely dispersed substance, and provides the most complete uranium recovery.

A number of commercial polymer ion exchangers are used for uranium sorption, namely ion-exchange resins (Gu et al. 2004; Wen et al. 2019; Outokesh et al. 2016) and fibers (Perlova et al. 2019; Sazonova et al. 2017). Improvement of selectivity towards U(VI) is achieved by additional functionalization with organic substances (Dragan et al. 2004; Metilda et al. 2005) or nanoparticles of ion exchangers (Dzyazko et al. 2017; Perlova et al. 2017, 2018). However, radionuclides provoke the degradation of polymer ion exchangers, and the products of polymer destruction form complexes with radioactive ions impeding their removal from water (Kao et al. 2016; Leybros et al. 2010; Loon and Hummel 1999). Moreover, polymer ion exchangers are fouled with organic species (Gönder et al. 2006) or inorganic colloidal particles, which are formed during equipment corrosion (Zvezdov and Ishigure 2003).

Inorganic ion exchangers, such as hydrated oxides or hydrophosphate of multivalent metals, look attractive for U(VI) sorption (Ali 2018; Amphlett 1964; Liu et al. 2016; Reinoso-Maset and Ly 2016; Wang et al. 2016; Zakutevskyy et al. 2012), since they are stable against ionizing radiation and fouling. They possess higher selectivity towards U(VI) compounds, since sorption is accompanied by the complex formation with phosphate (Drot et al. 1998; Almazan-Torres et al. 2008; Ordoñez-Regil et al. 2003) or hydroxide (Pan et al. 2017; Um et al. 2008) functional groups. Similar effect has been also found for sorption of d metals on phosphates (Dzyazko et al. 2013).

However, the inorganic materials are very sensitive to the solution acidity. Dissociation of hydrophosphate functional groups becomes sufficient at pH > 2; dissociation of—OH groups dominates in alkaline media (Amphlett 1964). Moreover, speciation of U(VI) ions is strongly dependent on pH and salt background (Cornelis et al. 2005): U(VI) ions are in cationic forms in sulphuric or nitric acid; anionic species dominate in carbonate or chloride media. It should be also noted that sorption of UO22+ on both natural (Bachmaf and Merkel 2011) and synthetic (Criscenti and Sverjensky 1999) inorganic materials is depressed by Cl− ions, which are usually contained in water.

At the same time, graphene oxide (GO) can be recommended as a modifier of inorganic materials for improvement their sorption properties, since it is characterized by high adsorption capacity caused by the developed surface, where carboxyl and phenolic functional groups are located (Dreyer et al. 2010). This two-dimensional carbon nanomaterial was applied to modifying hydrated zirconium dioxide, HZD (Dzyazko et al. 2018; Seredych and Bandosz 2010,2011), or zirconium hydrophosphate, ZHP (Pourbeyram, 2016). GO is an effective adsorbent of U(VI) cationic species (Yang et al. 2018). However, the GO production from graphite involves a number of aggressive reagents (KMnO4, HNO3, and so on) (Dreyer et al. 2010). Moreover, the synthesis process is dangerous from the point of view of safety. More safe and eco-friendly synthesis can be performed using multiwalled carbon nanotubed (MWCNTs); multilayered GO-like materials (GOlm), namely GO nanoribbons, are obtained by chemical or electrochemical methods (Danilov et al. 2016a). It is expected that this modifier will improve adsorption properties of inorganic ion exchangers towards U(VI) ions. The necessary requirements for ion exchanger are coarse mechanically durable granules: this would give a possibility to use the sorbent as a filler of sorption columns.

The aim of this work is to establish the effect of GOlm, which was obtained from MWCNTs, on the grade of the composites based on HZD and ZHP, as well as its influence on sorption of U(VI) ions on these materials. The tasks of the work also involve characterization of the composite materials.

Experimental

Materials

Multiwalled carbon nanotubes (MWCNTs) of standard characteristics were the initial material for the GOlm synthesis. They have been purchased from the "TM Spetzmash" LTD. Their outer diameter was 10–30 nm, the bulk density was 25–30 g dm−3, the wall amount was 8–15, and the specific surface area was 130 m2g−1.

Such reagents as ZrOCl2⋅8H2O, NH4OH, H3PO4, K2Cr2O7, H2SO4, HCl, HNO3, and NaOH were purchased from the Cherkassy KhimProm LTD (Ukraine). The UO2Ac2·2H2O salt was produced by Chemapol, Czech Republic.

Synthesis of GOlm and composites based on HZD and ZHP

Graphene-like carbon material, which consists of several grapheme sheets (8–15 according to the numbers of MWCNTs’ layers), was synthesized according to the method described in (Danilov et al. 2016a). In general, the methods that allow one to open carbon nanotubes are summarized in (Danilov et al. 2016b). Briefly, 1 g of MWCNTs was added to 30 cm3 of H2SO4 solution (98%), the suspension was stirred during 1 h. After this, 10 g of K2Cr2O7 was added, and the mixture was stirred during 72 h. Furthermore, the dispersion was filtered and washed subsequently with a HCl solution and distilled water. Then, the obtained powder was dispersed in water (2.8 mg cm−3).

Sol of insoluble zirconium hydroxocomplexes was obtained similarly to (Dzyazko et al. 2014; Dzyaz'ko et al. 2006) by means of gradual adding of a NH4OH solution to a 1 M ZrOCl2 solution under intensive stirring at 80 °C. Then, sol was boiled for 48 h and stored under room temperature. Relatively to anhydrous zirconium oxide, 100 cm3 of sol contained 12.3 g ZrO2.

Sol (100 cm3) and GOlm suspension (22.5, 45, 90, 135, 220, and 307 cm3) were mixed und activated with ultrasound for 5 min at 30 kHz using a Bandeline bath (Bandeline). Then, the GOlm-containing suspension was added to saturated NaOH solution, the beads were washed subsequently with a 1 M NH4OH solution and deionized water. Pure HZD hydrogel was obtained from sol by the same manner. A part of hydrogels, which were obtained by precipitation from sol or GOlm suspension in sol, was washed repeatedly with deionized water down to pH 7 of the effluent. Then, hydrogels were dried under room temperature down to constant mass.

Other part of hydrogels (pure and GOlm-containing HZD) was boiled in a 1 M H3PO4 for 1 h, washed with deionized water up to pH 7, and dried under room temperature. The samples HZD-GOlm and ZHP-GO were obtained by this manner. The samples contained 0.5, 1, 2, 3, 5, and 7 mass % of GOlm relatively anhydrous ZrO2.

Dry sorbents were sieved, and fractions of 0.8–1, 0.5–0.8, 0.25–0.5, 0.2–0.25, 0.1–0.2, and < 0.1 mm were taken and weighted. Before the investigations, the samples of HZD and HZD-GOlm were regenerated electrochemically at 90 V using a three-compartment cell supplied with ion-exchange membranes. This was necessary to remove NH4+ ions from the solids. The cell construction and equipment are described in (Dzyazko et al. 2019).

Characterization of samples

In this work, the samples based on HZD and ZHP composites containing 2% of GOlm were in a focus of attention, since they are characterized by optimal functional properties. Pure inorganic sorbents were also considered for comparison.

FTIR spectra were recorded using a Spectrum BX FT-IR spectrometer (PerkinElmer Instruments, USA). Before the measurements, the samples were grinded and compressed with KBr, which was preliminarily grinded and dried at 135 °C. The mixture was compressed at 10 bars.

The XRD patterns of the grinded samples applied to glass substrates were obtained by means of a DRON-3 spectrometer (Burevestnik, RF) supplied with tungsten cathode, copper anode, and nickel filter. The sample was inserted into a quartz capillary, a diameter of which was 0.5 mm. The measurements were within the interval of 2θ = 0–90°, the step was 0.01°. The current was 4 A.

Visualization of the sample structure was carried out using a JEOL JEM 1230 transmission electron microscope (Jeol, Japan). Preliminarily, the samples were grinded and fixed on copper substrates as thin films. Accelerating voltage was 120 kV.

Before sieving, the granule size of some sorbents was estimated using a Crystal-45 optical microscope (Konus, USA). 300 particles for each sample were analyzed. The particle-size distributions were plotted as dependencies of particle fraction (\(W\)) on grain diameter. The \(W\) value was calculated via:

$$W_{i} = \frac{{q_{i} }}{{\sum\limits_{i = 1} {q_{i} } }}.$$
(1)

The method of adsorption–desorption of nitrogen was used for porosimetric measurement. The Autosorb Nova-6B device (Quantochrome Instruments, USA) was applied to the investigations.

Potentiometric titration was carried out by means the solid addition method (Somasekhara Reddy et al. 2012) using a 0.1 M KNO3 solution. The solution pH was regulated with HCl and NaOH solutions. The solution pH was measured with I-160MI pH meter (Izmeritelnaya technika, Republic of Belarus).

Adsorption testing

All adsorption experiments that are described in this and next subsections were carried out under batch conditions at 25 °C. The samples of different composition were used for the preliminary test. 50 dm3 of a 0.1 M KOH solution was added to 2 g of sorbent. After 48 h, the alkali solution and solid were separated, and the sorbent was washed with deionized water down to pH 7 of the effluent. Then, the sample was treated three times with a 0.1 M HNO3 solution to provide the most complete regeneration. K+ ions in the effluent were analyzed with a PFM-U4.2 flame photometer (Analitpribor, Republic of Belarus).

Uranium adsorption and sorbent regeneration

The materials containing 2% of GOlm, as well as pure HZD and ZHP were researched. The sorbent mass was 0.1 g; the solution volume was 50 cm3. For the preparation of U(VI)-containing solution, HNO3 was added to tap water. Water contained initially 1.2 mmol dm−3 Ca2+ and 0.5 mmol dm−3 Mg2+; the resulting acid concentration was 0.02 mol dm−3. Uranyl acetate was dissolved in this solution; nitric acid provided cationic forms of U(VI). The solution pH was regulated by adding NaOH solution. The U(VI) content was determined using a Shimadzu UV-mini1240 spectrophotometer (Shimadzu, Japan) at 670 nm (Kadam et al. 1981). Preliminarily, U(VI) was transformed into a complex with Arsenazo III. The initial (Co) and equilibrium (C) concentration was obtained by this manner. The degree of uranium removal (S) was calculated as \(\frac{{C_{0} - C_{{}} }}{{C_{0} }} \times 100\%\); capacity (A) was determined as \(\frac{{V(C_{0} - C)}}{m}\); where \(V\) is the solution volume and \(m\) is the sorbent mass.

To investigate adsorption rate, the solution containing 25 mg dm−3 of U(VI) was used, and the initial pH was seven. When sorbent was added, the flask content was intensively stirred with a Water Bath Shaker Type 357 system (Elpan, Poland). After the predetermined time, the solution probe (1 cm3) was taken and analyzed later. The effect of pH on sorption was investigated using the solution containing 25 mg dm−3 of U(VI); the sorption time was 24 h. When sorption isotherms were obtained, the initial solution concentration of U(VI) was 5–25 mg dm−3, the pH was seven, and the adsorption time was 120 h.

Regeneration of loaded samples was carried out by means of deionized water or 0.1 M NaHCO3 or HNO3 solutions. The sorbent mass was 0.1 g. The volume of the solution for regeneration was 50 cm3. The regeneration degree was calculated as \(R_{{\text{d}}} = \frac{{C_{{\text{d}}} }}{{C_{0} - C}} \times 100\%\), where Cd is the effluent concentration.

Results and discussion

Effect of GOlm on grading sorbents

HZD and ZHP samples are coarse white grains; their shape is close to spherical. The granules of the composites containing GOlm are black. GOlm affect the grain size of the sorbents, as shown in Fig. 1a. The figure illustrates the dependence of the average granule diameter of the heaviest (dominating) fraction on the content of this carbon material. Increasing in the GOlm amount reduces a size of the granules, which are formed during synthesis.

Fig. 1
figure 1

Average granule diameter of the dominating fraction of sorbents as a function of GOlm content (a), grain size distribution of pure inorganic sorbents and their composites containing 2% of GOlm (b).

Full size image

Figure 1b illustrates detailed granule size distribution for both pure inorganic sorbents and their composites containing 2% of GOlm. As seen, ZHP forms larger granules than HZD due to cross-linkage of hydrogel with Zr–O–P–O–Zr bonds. Addition of GOlm to both ZHP and HZD results in a decrease of the granule size of dominating fraction in ≈2 times.

Visualization of primary particles

TEM image of MWCNTs is given in Fig. 2a. Both straight and stranded nanotubes are seen; the diameter of the straight ones is about 30 nm. Chemical treatment causes partial expansion of the nanotubes: GOlm sheets are seen (Fig. 2b). However, some of undisclosed nanotubes are present in the bulk of the carbon material.

Fig. 2
figure 2

TEM images of MWCNTs (a), GOlm (b), HZD (c), and ZHP (d)

Full size image

Nanoparticles, a size of which is 6 nm, dominates in sol of insoluble zirconium hydroxocomplexes (Dzyazko et al. 2014). During precipitation, the nanoparticles are merged due to the formation of Zr–O–Zr bonds. As a result, larger strongly aggregated particles (≈20–50 nm) are formed (Fig. 2c). Their shape is close to spherical. The particles of irregular shape are formed during ZHP deposition (Fig. 2d). In general, they are larger than 100 nm. These formations are evidently splices of smaller particles; coalescence is caused by Zr–O–P–O–Zr bonds.

Addition of GOlm causes loosening of inorganic nanoparticles (Fig. 3a, b) due to their adsorption on carbon sheets. GOlm forms more compact coating in the surface of ZHP, the primary particles of which cannot be recognized as opposed to HZD. High-resolution image of coated ZHP particles is typical for graphene materials (Fig. 3c). The stripes correspond to the expended graphite nanosheets.

Fig. 3
figure 3

Low-resolution TEM images of HZD-GOlm (a) and ZHP-GOlm (b), and high-resolution image of ZHP-GOlm (c). The GOlm content was 2%

Full size image

XRD and FTIR spectroscopy

XRD patterns of MWCNTs are shown in Fig. 4. The (002) peak corresponds to 2θ = 26.02°; it corresponds to the reflex from the interplanar distance between graphite layers. The position of this reflex depends on the MWCNTs thickness (Futaba et al. 2011; Zhao et al. 2015). The reflex is not narrow, since the wall numbers are within the interval of 8–15. The diffuse symmetric reflex about 2θ = 21° is due to SiO2 of a capillary.

Fig. 4
figure 4

XRD patterns for MWCNTs, GOlm, and ZHP-GOlm-containing 2% of the carbon material

Full size image

In the case of GOlm, the reflex at 2θ = 26.02° disappears; the peak attributed to SiO2 shows a diffuse shoulder from the side of smaller angles. This shoulder is related to GOlm. Regarding GO, the narrow reflex at 11.4° has to be present in the XRD pattern. However, the observed diffuse peak is located at larger angles; this indicates several walls of the carbon material obtained from MWCNTs.

The characteristic reflex for GOlm is masked by inorganic matrix: the XRD pattern for the ZHP-GOlm composite shows a wide galo, which is characteristic for amorphous materials. Similar results were obtained for the ZHP-GOlm composite.

The data of FTIR spectroscopy are given in Fig. 5. The spectra for all samples show very broad peaks at 3400 cm−1: these stripes are attributed to OH vibrations of water molecules. This peak is the widest for the inorganic ion exchangers and their composites. More narrow stripe is observed for the GOlm.

Fig. 5
figure 5

FTIR spectra of GOlm, inorganic matrices and composites containing 2% of GOlm

Full size image

The FTIR spectrum of GOlm shows a number of peaks: they are attributed to stretching vibration of C–O–C (epoxy groups, 1070 cm−1), C–OH (hydroxyl and carboxyl groups, 1257 cm−1), O = C–OH (carboxyl groups, 1401 cm−1), C–C (aromatics, i.e. carbon skeletal network, 1620 cm−1), and C = O (carbonyl of carboxyl groups, 1722 cm−1) (Zhao et al. 2015).

Characteristic peaks of HZD, which are attributed to metal–oxygen bonds, are as follows: stretching vibrations of Zr–O (Zr–OH surface groups, 490 cm−1), Zr–O–Zr (bulk of particles, 550–750 cm−1), Zr–OH (1380, 1553 cm−1), and O–Zr–OH bonds (1640 cm−1) (Kostrikin et al. 2010). Regarding ZHP, triply degenerates symmetric stretch vibration of PO4 is at 1050 cm−1. Addition of GOlm causes no shift of the peaks related to metal–oxygen bonds. It means no chemical bonding graphene sheets with functional groups of inorganic matrices. Vibrations, which are related to GOlm, are masked due to small amount of the carbon filler.

Porometric measurements

Isotherms of nitrogen adsorption–desorption are given in Fig. 6a, b. Regarding GOlm, the isotherm is related to II type of IUPAC classification (Gregg and Sing 1982), since it is convex relatively abscissa axis in the region of low relative pressure (P/Ps) and concave at high pressure. The hysteresis loop is attributed rather to B type according to de Boer classification. This is characteristic for slit-like pores. At the same time, the isotherms for both inorganic sorbents and their composites are Langmuir type (type I) indicating microporous structure of the materials.

Fig. 6
figure 6

Isotherms of adsorption–desorption of nitrogen for GOlm (a), inorganic matrices and composites containing 2% of GOlm (b), and differential pore-size distributions for some sorbents (c)

Full size image

Differential pore-size distribution for GOlm that were obtained according to the BJH method are plotted in Fig. 7. The narrow peak, which is related to smaller mesopores, indicates their ordering. Larger mesopores are disordered (the corresponding maximum is diffuse). Both HZD and its composite are practically non-porous within the interval of 10–100 nm. Considerable porosity has been found for lower r values. The maxima position is at r < 1.5 nm. The distributions for ZHP and ZHP-GOlm samples look similarly.

Fig. 7
figure 7

Total adsorption capacity of composites towards K+ ions as a function of GOlm content

Full size image

The results of porometric measurements are summarized in Table 1. The value of specific surface area for GOlm is much smaller than that calculated theoretically for isolated grapheme sheets (≈2600 m2g−1) as pointed in (Dreyer et al. 2010). This disagreement is due to agglomeration, overlapping and curling the sheets of graphene. Nitrogen molecules cannot penetrate between the graphene sheets under low temperature. The values of surface area that are an order of magnitude higher were obtained for liquid adsorbates, such as water and octane.

Table 1 Pore structure and PZC of inorganic matrixes and their composites containing 2% of GOlm
Full size table

As opposed to ZHP, HZD is characterized by more developed surface comparing with GOlm (Table 1). Lower value of surface area has been found for ZHP indicating coalescence of primary HZD particles during boiling in phosphoric acids. Indeed, this is confirmed by TEM images (see Figs. 2c, d) Additions of even small amount of GOlm to HZD slightly reduce its microporosity and specific surface area. At the same time, mesoporosity of the composite increases. When GOlm is added to ZHP, the composite shows higher micro- and mesoporosity comparing with pure ZHP.

Acid base properties of composites and adsorption testing

Regarding GO, several mechanisms of the surface charge formation are known: (i) dissociation of carboxyl and phenolic groups, (ii) proton complexation of the π-electron system of graphene planes acting as Lewis basic sites, and (iii) protonation of Bronsted basic oxygen-containing fragments, such as epoxy- and carbonyl groups (Szabo et al. 2018). Dissociation causes cation-exchange ability of GO, protonation leads to anion adsorption. As found, dissociation dominates at pH > 3.2, while protonation occurs under lower pH (see Table 1). ZHP contains a several amount of − OH groups; they provide anion-exchange ability below pH 2. The composite containing higher basic GOlm shows slightly higher pH of PZC. At last, GOlm is more acidic than HZD. Thus, the addition of this carbon material causes a shift of the PZC to acidic region.

Increasing in GOlm content in ZHP causes a growth of adsorption capacity towards K+ ions (Fig. 7). This growth is rather rapid up to the amount of the carbon constituent of two mass %. The increase of capacity is 14% comparing with pure ZHP. This is due to highly developed surface of GOlm. More loose structure of the composite provides the access for K+ ions to adsorption sites, which are unavailable in pure ZHP. Moreover, GOlm groups are also involved to ion exchange. Further increasing in GOlm content causes no significant effect to adsorption: the growth of capacity is only 6%. This is due to a decrease of ZHP content. Namely − OPO3H2 groups provide high capacity, since they contain two counter-ions. As opposed to ZHP, the groups of GOlm contain only one counter-ion. In general, the capacity of the composites is comparable with that for weakly acidic cation-exchange resins.

Carboxyl and phenolic groups of GOlm provide a growth of adsorption capacity of the composites based on HZD. However, the samples containing more than 2% of GOlm show a decrease of capacity probably due to compaction of the composites (specific surface area of GOlm is lower comparing with pure HZD).

Thus, the optimal GOlm content in the composites is 2%. Moreover, these materials form rather large granules (see Fig. 1a), a size of which is similar to the beads of ion-exchange resins. These composites were chosen for investigations of U(VI) adsorption.

U(VI) adsorption and desorption

Uranium (VI) exists in nitrate-containing solutions in a form of cationic complexes of UO22+. Nevertheless, cationic species are sorbed by HZD in neutral media about PZC, since some − OH groups are dissociated under these conditions. Figure 8a illustrates the dependence of adsorption capacity over time (t). More than 50% of ions are removed from water in 4 h. Among known approaches, the model of pseudo-second order (Ho and McKay 1999) describes the experimental data most adequately. Following equation was applied:

$$\frac{t}{A} = \frac{1}{{KA_{\infty }^{2} }} + \frac{t}{{A_{\infty } }},$$
(2)
Fig. 8
figure 8

Adsorption capacity of inorganic adsorbents and their composites, which contain 2% GOlm, as a function of time (a) and these dependencies were simulated according to the model of pseudo-second order (b)

Full size image

where A∞ is the capacity at t → ∞, K is the rate constant. The data are given in Fig. 8b and Table 2.

Table 2 Application of the models of pseudo-second order and DR to adsorption of U(VI) cations
Full size table

As seen from Table 2, GOlm depresses U(VI) adsorption on HZD. This is due to a decrease of the composite surface (lower A∞ value) and bonding with carboxyl groups (slowdown of adsorption). At the same time, the A∞ value for the ZHP-GOlm sample corresponds to maximal possible magnitudes, when the removal degree is 100%. The fastest adsorption has been found namely for this composite. Adsorption capacity A∞, which has been found during investigations of adsorption rate, increases in the order: ZHP-GOlm > HZD > HZD-GOlm > ZHP. In general, GOlm improves sorption on ZHP only within a wide interval of pH (2–7), where dissociation of carboxyl groups occurs (Fig. 9). High removal degree is achieved despite significant excess of hardness ions. In the case of HZD-GOlm, it demonstrates higher removal degree than pure HZD only at pH 2, 5–4.

Fig. 9
figure 9

Removal degree of U(VI) ions as a function of pH

Full size image

Isotherms of UO22+ adsorption are given in Fig. 10a. For the equilibrium conditions, the adsorption capacity increases in the order: ZHP-GOlm > HZD > ZHP > HZD-GOlm. The Langmuir, BET, Freundlich, and Dubinin–Radushkevich (DR) models (Rieman and Walton 1970) were applied to the isotherms. As found, the most suitable is the DR model (Fig. 10b):

$$\ln A = \ln A_{DR} - \frac{{R^{2} T^{2} }}{{E^{2} }}[\ln (1 + 1/c)]^{2},$$
(3)

which has been found to give linear approximation within a certain interval of concentration of equilibrium solution. Here, ADR is the constant and E is the adsorption energy. The DR model corresponds to full filling of micropores with large UO22+ ions. The energy values are in the interval of 8–16 kJ mol−1 indicating ion-exchange mechanism. However, the data of sorption kinetics show an additional interaction with functional groups. This is typical for inorganic matrices and their composites. Regarding nitrate solutions, the pH of which is 7, U(VI) is in a form of the mixture of neutral (UO2(OH)2) and cationic ((UO2)3(OH)5+, (UO2)4(OH)7+) hydroxocomplexes (Cornelis et al. 2005). It is possible to suppose that the interaction of uranyl complexes with the functional groups of the adsorbent is due to surface complexation. This is evidently more characteristic for the composite based on ZHP, since the maximal uranyl uptake and the highest adsorption energy. However, this assumption requires confirmations in a future. The best kinetic characteristics of the ZHP-GOlm sample could be explained by a decrease of the micropore volume and increase of mesoporosity. Increase of pore size accelerates ion transport (Volfkovich 1984; Volfkovich et al. 1999).

Fig. 10
figure 10

Isotherms of U(VI) sorption: initial data (a) and application of DR model (b)

Full size image

The indirect confirmation of chemosorption is the data obtained during regeneration of U(VI) loaded adsorbents with deionized water (Table 3). As follows from the table, the desorption degree is very low indicating a significant contribution of chemosorption to the adsorption mechanism. As found, GOlm deteriorates regeneration of the HZD-based composite comparing with pure HZD (Table 3). In the case of ZHP, regeneration with sodium hydrocarbonate is preferable. The highest regeneration degree of the ZHP-GOlm sample is achieved when the solution of nitric acid is used. The development of regeneration conditions is the task of further investigations.

Table 3 Regeneration of sorbents
Full size table

Concluding remarks

GOlm, which was obtained by treatment of MWCNTs, can be considered as an addition to sorbents for U(VI) removal. Comparing with the traditional GO synthesis techniques, the method of MWCNTs’ disclosure is more eco-friendly. This carbon material increases the specific surface area of ZHP and reduces it in the case of HZD. Comparing with pure inorganic materials, sorption of U(VI) cations is improved on the ZHP-based composite and deteriorated on the sorbent based on HZD. The mechanism of sorption is ion exchange, which is accompanied by an additional interaction of sorbed ions with functional groups. The composite based on ZHP shows also high regeneration degree in the solution of nitric acid.

Increasing in GOlm content in composites causes a decrease of their granules evidently due to screening of aggregates of primary particles with graphene sheets. The optimal content of the carbon addition is 2%, which provides the formation of large granules. Considerable removal degree of U(VI) is achieved even in the presence of significant excess of hardness ions. Thus, this material can be used as a filler of sorption columns. Higher sorption capacity towards U(VI) is possible under higher content of GOlm. In this case, the sorbents should be fixed in a support. The composites are expected to be effective towards other toxic cations.