**3. Results and discussion. Acid-base equilibria of inorganic and organoinorganic materials and systems**

### **3.1. Inorganic oxide materials**

#### *3.1.1. Pure oxides*

220 Nitroxides – Theory, Experiment and Applications

The pH values of the equilibrium solutions (pHext) over the samples were measured using a Mettler Toledo pH meter (Switzerland) with an accuracy of 0.01 units. The samples separated from the solutions by centrifugation or ltration were placed into unsealed quartz ampules and their ESR spectra were recorded. After measuring the *a* distances in the ESR spectra of corresponding radicals located in the samples (Fig. 4), the pHint values of the

**Figure 3.** Titration curves for NR R1 in bulk aqueous solution (calibration curve) (1), PC (2), SiO2 (3) and

As can be seen from Fig. 4, the ESR spectra of NR in the samples studied represent the superpositions of three components of an isotropic signal of the probes in aqueous solutions inside pores and a spectrum of the probes immobilized on the surface of the objects studied. For determination of pHint values, only the isotropic signals in the ESR

TiO2 (4) xerogels. a, % = ((a-aNRH+)/(aNR-aNRH+))×100%

spectra were used.

studied materials were determined using the calibration curves (Fig. 2,3).

The comparison of the titration curves for NR R1 radical in **γ-Al2O3, -Al2O3, the BS-50 type SiO2 and TiO2 and SiO2** xerogels with the calibration curve of this NR (Fig. 2, 3) indicates that for γ-Al2O3, the curve is shifted to the left, and for other samples, to the right. As was shown previously for organic sorbents (Molochnikov et al., 2004), the left- side shift of the titration curves for the NRs used is characteristic of anionites in acidic solutions in the pH range, where amino groups bind hydrogen ions, for example,

$$\text{P-NH}\_2\text{+H}^+ \leftrightharpoons \text{P-NH}\_3^+\tag{1}$$

where *P* is the sorbent matrix. At the same time, the shift in the titration curves to the right is related to the dissociation of acidic functional groups of cationites, for example, a carboxylic one,

$$\text{P-COOH} \leftrightharpoons \text{P-COO}^{\cdot}\text{+H}^{+}\tag{2}$$

Thus, the shift of the titration curve of R1 to the left observed for γ-Al2O3 indicates the binding of H+ ions with the surface (pHintr is higher than solution pH) and to a positive charge of γ-Al2O3 surface, while the right-hand side shift of the titration curve of this NR in SiO2 gel suggests the release of H+ ions and, consequently, a negative surface charge.

According to the data reported in (Hubbard, 2002) the acidic dissociation of hydrated aluminum oxide occurs in two stages, which can be described by the following equations:

$$\text{-AlOH}\_2\text{}^+ + \text{H}\_2\text{O} \leftrightharpoons \text{-AlOH} + \text{H}\_3\text{O}^+\text{.}\tag{3}$$

$$\text{-AlOH} + \text{H}\_2\text{O} \leftrightharpoons \text{-AlO}^\cdot + \text{H}\_3\text{O}^+\text{.}\tag{4}$$

The pKavalues of these equilibria (pKa1= 5.87 and pKa2 = 7.50) are given in (Lidin et al., 1987). Using NR, we can determine only pH values relevant to the ascending part of its S-shaped calibration curve (Fig. 2, 3 curve 1), i.e., the range, in which variations in *a* parameter correspond to variations in pH. The range of pH values appropriate to the ascending part of the calibration curve for a given NR is referred to as the zone of its sensitivity. The sensitivity zone of NR R1 falls in the pH range from 2.5 to 6.5 (Fig. 2, 3). If the above data on pKa reect (even if qualitatively) the pKa of functional group dissociation for γ-Al2O3 studied, then the comparison of the sensitivity zone of NR R1 with pKa values indicates that the use of NR allows one to study only the pH range in which γ-Al2O3 either has a positive charge or is electrically neutral. On the other hand, there are data on the point of zero charge (PZC) for Al2O3 in the literature.

For example, as was shown previously, the PZC of an Al2O3 sample kept in water for a long time was equal to 9.2, and for Al2O3 treated at 1400°C, the PZC decreased to 6.7 (Robinson et al., 1964). The rst value corresponds to γ-Al2O3, and the second one, to -Al2O3. For hydrated Al2O3 the PZC was found to be equal to 8, and it decreases as a result of the modication of Al2O3 surface with alkyl phosphate-based surfactants (Jeon et al., 1996). Moreover, these data suggest that, in the studied pH range, the surface of γ-Al2O3 has a positive charge that is in accordance with the shift of its titration curve to the left.

The pHint of the sorbents differ from the pHext by 0.5–1.5 units (Table 3). Smaller pHintr values compared to pH of the external solutions for all the studied samples (except for γ-Al2O3) are indicative of the shift in titration curves of NR R1 to the right relative to the calibration curve, similarly to that demonstrated in Fig. 2, 3 all the samples excepting γ-Al2O3.

It should be noted that the pHint value determined in this study for the hydrated -Al2O3 is in good agreement with the PZC value found in (Robinson et al., 1964). From the data given in Table 3 we notice that in all the studied samples excepting γ-Al2O3 acidic centers predominate. However, the acidity of these centers is rather weak; their pKa of dissociation are presumably close to 6. Only when strongly acidic sulfate residues appear in Al2O3, a drastic decrease in pHint is observed as a result of their dissociation. It should be noted that the same decrease in pHint of TiO2 hydrogel results from the sorption of Cu2+ ions by the hydrogel, which leads to the ion-exchange displacement of H+ ions (Fig. 5).

222 Nitroxides – Theory, Experiment and Applications

(PZC) for Al2O3 in the literature.

one,

where *P* is the sorbent matrix. At the same time, the shift in the titration curves to the right is related to the dissociation of acidic functional groups of cationites, for example, a carboxylic

Thus, the shift of the titration curve of R1 to the left observed for γ-Al2O3 indicates the binding of H+ ions with the surface (pHintr is higher than solution pH) and to a positive charge of γ-Al2O3 surface, while the right-hand side shift of the titration curve of this NR in

According to the data reported in (Hubbard, 2002) the acidic dissociation of hydrated aluminum oxide occurs in two stages, which can be described by the following equations:

<sup>+</sup> -AlOH + H O -AlOH+H 2

The pKavalues of these equilibria (pKa1= 5.87 and pKa2 = 7.50) are given in (Lidin et al., 1987). Using NR, we can determine only pH values relevant to the ascending part of its S-shaped calibration curve (Fig. 2, 3 curve 1), i.e., the range, in which variations in *a* parameter correspond to variations in pH. The range of pH values appropriate to the ascending part of the calibration curve for a given NR is referred to as the zone of its sensitivity. The sensitivity zone of NR R1 falls in the pH range from 2.5 to 6.5 (Fig. 2, 3). If the above data on pKa reect (even if qualitatively) the pKa of functional group dissociation for γ-Al2O3 studied, then the comparison of the sensitivity zone of NR R1 with pKa values indicates that the use of NR allows one to study only the pH range in which γ-Al2O3 either has a positive charge or is electrically neutral. On the other hand, there are data on the point of zero charge

For example, as was shown previously, the PZC of an Al2O3 sample kept in water for a long time was equal to 9.2, and for Al2O3 treated at 1400°C, the PZC decreased to 6.7 (Robinson et al., 1964). The rst value corresponds to γ-Al2O3, and the second one, to -Al2O3. For hydrated Al2O3 the PZC was found to be equal to 8, and it decreases as a result of the modication of Al2O3 surface with alkyl phosphate-based surfactants (Jeon et al., 1996). Moreover, these data suggest that, in the studied pH range, the surface of γ-Al2O3 has a

The pHint of the sorbents differ from the pHext by 0.5–1.5 units (Table 3). Smaller pHintr values compared to pH of the external solutions for all the studied samples (except for γ-Al2O3) are indicative of the shift in titration curves of NR R1 to the right relative to the calibration

It should be noted that the pHint value determined in this study for the hydrated -Al2O3 is in good agreement with the PZC value found in (Robinson et al., 1964). From the data given in Table 3 we notice that in all the studied samples excepting γ-Al2O3 acidic centers

positive charge that is in accordance with the shift of its titration curve to the left.

curve, similarly to that demonstrated in Fig. 2, 3 all the samples excepting γ-Al2O3.

SiO2 gel suggests the release of H+ ions and, consequently, a negative surface charge.


<sup>+</sup> O , 2 3 (3)



**Table 4.** pH values of the external bulk solution of NR R1 and in the pores and near a surface of the inorganic materials studied

**Figure 5.** The pH values of a CuCl2 equilibrium solution (pHext) (1) and a solution inside TiO2 hydrogel (pHint) (2) vs. the amount of sorbed Cu2+ (c**Cu**)

The titration curves of radicals occurring in organic sorbents are characterized by the presence of horizontal plateaus corresponding to constant pHint (parameter *a* remains unchanged), when the pH of the external solution increases (Molochnikov et al., 2004). The presence of the horizontal plateaus on the titration curves is related to the consumption of the titrant solution for the neutralization of functional groups of ionites. There was found no horizontal plateau on the titration curve of NR R1 in γ-Al2O3 within its zone of sensitivity (Fig.2). This fact can be explained either by the presence of a small number of acidic functional groups in γ-Al2O3 or by noncoincidence of pKa values of these acid groups with the zone of sensitivity of the radical used. Based on our experimental data, it is difcult to make a conclusion in favor of any of these assumptions.

It should be noted that the titration curve of NR R1 in the BS-50 SiO2 contains two horizontal plateaus corresponding two constant pHint determined near the SiO2 surface when the pH of the external solution is changed (pHext) (Fig.2, curve 3). These plateaus are referred to as the titration process of functional groups on the surface (Golovkina et al., 2008). The рКа values for these groups were determined from the titration curves, and were рКа1 = 4.4–4.7 and рКа2 = 6.5–6.8. According to (Long et al., 1999 ; Zhao et al., 1997, 1998) amorphous SiO2 contains groups of three types: silanol, silandiol, and siloxane groups in a ratio of 59.2**,** 14.7, and 26.1%, respectively. We can therefore assume that the lower and longer horizontal plateau reflects the titration process of silanol functional groups on the SiO2 surface, and the upper plateau corresponds to silandiol groups. According to (Méndez et al., 2003), the intervals of changes in the dissociation constants are рК1 = 3.51–4.65 and рК2 = 6.17–6.84, respectively. The same values of constants of dissociation for silanol and silandiol groups of silica were measured and published in (Nawrocki, 1997 ; Neue, 2000). The presence of silanol and silandiol groups on the surface of the studied SiO2 was thus identified.

**From two types of the xerogels studied (TiO2 and SiO2)** electrical potential of TiO2 xerogel is slightly more than that of SiO2 xerogel (Fig.3). There was found a small horizontal plateau in the range of pHintr between 3.75 and 4.25 on the titration curve for NR R1 near the SiO2 surface. From this plateau pKa values for silanol groups were determined. They were found to be equal to 3.95 0,07 within the zone of constant pHint values (Molochnikov et al., 1996). The measured values of pKa are in good agreement with those determined before for silicas using chromatography(Méndez et al., 2003) and the method of pH probe in different silicacontaining samples (Golovkina et al., 2008). At lesser pH values this curve goes near (but slightly righter) than the calibration curve of the above-mentioned NR. As seen from the figure, the surface of SiO2 xerogel carries a small negative charge which is characteristic for the sample before titrating silanol groups. The lack of horizontal plateaus on the titration curves of NR R1 for the TiO2 xerogel indicates the fact of titration of its functional groups at pH above 7. It can be explained by the presence of mainly terminal OH- -groups with basic properties on a surface of TiO2 xerogel. The range of their titration lies above the zone of sensitivity of the NR used.

When studying acidity inside organic sorbents or hydrogels, pH probes are placed into water, which penetrates into a solid due to sample swelling. In this case, the probes occur in water environment far away from the chains of organic (synthetic ion-exchange resins) or inorganic (TiO2 and SiO2 hydrogel, xerogel) polymers.

The particles of nanostructured oxides in diameter of several nm possess extremely large surface (the ratio of the number of atoms on the surface of a nanoparticle to the number of atoms located inside the particle is equal to 1/7). Therefore, a pH-sensitive NR surrounded by water molecules can approach the hydrated surface of nanoparticles rather closely. It was expected that they as probes and labels would appear to be sensitive to SEP, which would make it possible to estimate its value.

224 Nitroxides – Theory, Experiment and Applications

make a conclusion in favor of any of these assumptions.

presence of the horizontal plateaus on the titration curves is related to the consumption of the titrant solution for the neutralization of functional groups of ionites. There was found no horizontal plateau on the titration curve of NR R1 in γ-Al2O3 within its zone of sensitivity (Fig.2). This fact can be explained either by the presence of a small number of acidic functional groups in γ-Al2O3 or by noncoincidence of pKa values of these acid groups with the zone of sensitivity of the radical used. Based on our experimental data, it is difcult to

It should be noted that the titration curve of NR R1 in the BS-50 SiO2 contains two horizontal plateaus corresponding two constant pHint determined near the SiO2 surface when the pH of the external solution is changed (pHext) (Fig.2, curve 3). These plateaus are referred to as the titration process of functional groups on the surface (Golovkina et al., 2008). The рКа values for these groups were determined from the titration curves, and were рКа1 = 4.4–4.7 and рКа2 = 6.5–6.8. According to (Long et al., 1999 ; Zhao et al., 1997, 1998) amorphous SiO2 contains groups of three types: silanol, silandiol, and siloxane groups in a ratio of 59.2**,** 14.7, and 26.1%, respectively. We can therefore assume that the lower and longer horizontal plateau reflects the titration process of silanol functional groups on the SiO2 surface, and the upper plateau corresponds to silandiol groups. According to (Méndez et al., 2003), the intervals of changes in the dissociation constants are рК1 = 3.51–4.65 and рК2 = 6.17–6.84, respectively. The same values of constants of dissociation for silanol and silandiol groups of silica were measured and published in (Nawrocki, 1997 ; Neue, 2000). The presence of

silanol and silandiol groups on the surface of the studied SiO2 was thus identified.

pH above 7. It can be explained by the presence of mainly terminal OH-

inorganic (TiO2 and SiO2 hydrogel, xerogel) polymers.

sensitivity of the NR used.

**From two types of the xerogels studied (TiO2 and SiO2)** electrical potential of TiO2 xerogel is slightly more than that of SiO2 xerogel (Fig.3). There was found a small horizontal plateau in the range of pHintr between 3.75 and 4.25 on the titration curve for NR R1 near the SiO2 surface. From this plateau pKa values for silanol groups were determined. They were found to be equal to 3.95 0,07 within the zone of constant pHint values (Molochnikov et al., 1996). The measured values of pKa are in good agreement with those determined before for silicas using chromatography(Méndez et al., 2003) and the method of pH probe in different silicacontaining samples (Golovkina et al., 2008). At lesser pH values this curve goes near (but slightly righter) than the calibration curve of the above-mentioned NR. As seen from the figure, the surface of SiO2 xerogel carries a small negative charge which is characteristic for the sample before titrating silanol groups. The lack of horizontal plateaus on the titration curves of NR R1 for the TiO2 xerogel indicates the fact of titration of its functional groups at

properties on a surface of TiO2 xerogel. The range of their titration lies above the zone of

When studying acidity inside organic sorbents or hydrogels, pH probes are placed into water, which penetrates into a solid due to sample swelling. In this case, the probes occur in water environment far away from the chains of organic (synthetic ion-exchange resins) or


The dependences of NR ionization constants on the surface electric potentials and the polarity of media were reported by Fromherz (Fromherz , 1989) and Khramtsov et al. (Khramtsov et al., 1992) for uorescent pH indicators and for pH-sensitive NRs, respectively, as follows:

$$
\Delta \text{p} \text{K}\_{\text{a}} = \Delta \text{p} \text{K}\_{\text{a}}^{\text{el}} + \Delta \text{p} \text{K}\_{\text{a}}^{\text{pol}} \tag{5}
$$

where pKa is the total shift in the pKa of an NR occurring in a sample relative to that of the NR in an aqueous solution and el <sup>Δ</sup>pKa and pol <sup>Δ</sup>pKa are the shifts in the pKa of the NR occurring in the sample, which are due to th electric potential arising on the sample surface and to a change in the polarity of the medium relative to the polarity of the NR aqueous solution, respectively.

$$
\Delta \text{p} \text{K}^{\text{el}}\_{\text{\tiny\text{\tiny\text{\tiny}}}} = \text{-}e \times \text{p} / 2.3 \text{k} \times \text{T} \tag{6}
$$

$$\mathsf{ApK}\_{\mathsf{a}}^{\mathsf{Pol}} \mathsf{a}^{\mathsf{Pol}} \quad \mathsf{=b} \times (\mathsf{e} \cdot \mathsf{e}\_{\mathsf{HOH}}) \tag{7}$$

where φ is the electrostatic potential, e is the electron charge, k is the Boltzmann constant, T is absolute temperature, b is a proportionality coefcient dependent on the radical nature, ε is the medium permittivity, and εHOH is the permittivity of water.Figures 6 and 7 demonstrate the results of studying nano- structured TiO2. The measurements were performed using R3 probe at pH ≈ 5.5, which was close to the PZC of TiO2 (5.7) (Poznyak et al., 1999), and R2 probe at pH ≈ 4.3. Near the surface of TiO2 nanostructured particles, the titration curves of the radicals are shifted to the left, whereas, for TiO2 hydrogel and xerogel (Fig.3), a shift to the right was observed. At present, we can not find any explanation for such principal different shift in the titration curves for TiO2 –based samples. The TiO2 hydrogel obtained by hydrolysis was washed out with water, removing acidic components of the initial salts. Hereafter, the samples were prepared at different drying temperatures. The samples of xerogels and nanostructured oxides were dried at 100◦C and 200◦C, respectively.

As was mentioned above, the shift of the titration curve for γ-Al2O3 to the left corresponds to the positive charge of the particle surface that, in principle, correlates with the data obtained at pH below PZC. The left-shift (toward more acidic solutions) of the titration curve of the probe located near the surface of TiO2 nanoparticles indicates that the surface electric potential is also positive. No changes were found in the hyperne splitting constants *aN* for both deprotonated and protonated R3 radicals (Fig. 6). The curves of titration of NR

**Figure 6.** Titration curves for NR R3 in bulk aqueous solution (calibration curve) (1) and near the surface of nanostructured TiO2 (2) at 25°C and µ = 0.1

**Figure 7.** Titration curves for NR R2 (1) in bulk aqueous solution and (2) near the surface of nanostructured TiO2 at 25°C and µ = 0. 1 . The amount of Cu2+ ions sorbed in the pores of TiO2 is 0.36 mmol/g of the sample

R3 in the solution (curve 1) and near the surface of the nanostructured TiO2 (curve 2) begin and end at the same value of parameter *a* characterizing the hyperne splitting constant *aN* for the protonated and deprotonated forms of NR. This fact shows that there are no changes in the polarity of the medium on the surface of nanostructured TiO2 relative to the external solution, and the term ∆pKapol calculated by formula (7) is reduced to zero (Khramtsov et al., 1992). Thus, the shift in pKa is determined by the electric potential of nanoparticle surface alone (pKa = pKael). The calculation by formula (6) gives φ = 31.7 mV at the point of R3 probe location. It should be stressed that the calculated φ value is not the electric potential of TiO2 nanoparticle surface, but it only characterizes the electric eld generated by a nanoparticle at the site where the radical fragment –N–O• of R3 probe is located.

### *3.1.2. Cu2+-containing samples*

226 Nitroxides – Theory, Experiment and Applications

surface of nanostructured TiO2 (2) at 25°C and µ = 0.1

mmol/g of the sample

**Figure 6.** Titration curves for NR R3 in bulk aqueous solution (calibration curve) (1) and near the

**Figure 7.** Titration curves for NR R2 (1) in bulk aqueous solution and (2) near the surface of

nanostructured TiO2 at 25°C and µ = 0. 1 . The amount of Cu2+ ions sorbed in the pores of TiO2 is 0.36

**For TiO2 hydrogel,** pH values of the external solution, which is in contact with the hydrogel (pHext), and of those inside the hydrogel (pHint) during the sorption of Cu2+ ions from copper(II) chloride and nitrate solutions are found to be different (Fig. 5). It was observed that the anions of these copper(II) salts had no effect on the acidity of a medium. In the process of sorption Cu2+ ions are involved in the chemical interaction with the active centers of the hydrogel (Kharchuk et al., 2004) that manifests itself both in the parameters of the ESR spectra of Cu2+ in the hydrogel and in the impossibility of washing out sorbed Cu2+ ions from the hydrogel by Na+ and Ca2+ nitrate solutions**.** According to the data obtained, initially, pHint is lower by 1.5 units than pHext (Table 3). As the amount of Cu2+ ions in the hydrogel grows, pHint decreases (Fig. 5), and, beginning with an amount of 0.24 mmol/g of the gel, it coincides with pHext. Thus, the acidity inside the original hydrogel sample is considerably lower than that of the equilibrium external solution. Similar regularities were found for granulated organic ion-exchange resins (Kovaleva et al., 2000). For KB-2 × 4 cationite in a mixed H+–Na+ form, it was found that in the process of absorption of copper, the acidity of the external solution diminishes, while the acidity inside the sorbent remains unchanged until a certain amount of Cu2+ ions in the ionite is attained, with this content being dependent on the fraction of Na+ form of the cationite When this value is exceeded, the acidity inside the sorbent begins to decrease and pHext and pHext values for the ionite become equal at a certain degree of saturating with Cu2+ -ions. For TiO2 hydrogel, the equalization of pH values is also observed, but the horizontal plateau with a constant pHint value on the titration curve of NR is not observed. The constant pHint values for KB-2 × 4 cationite are explained by the buffer properties of its H+–Na+ mixed form. TiO2 hydrogel has no buffer properties, and the acidity increases smoothly with a rise in the amount of sorbed Cu2+ ions.

**For nanostructural TiO2**, in the presence of sorbed Cu2+ ions, at pH close to 4.3 (Fig. 7), like in the case of pure nano -TiO2 (Fig.6), a shift of the titration curve toward lower pH values is observed; however, the hyperne splitting constant aN decreases (the titration curve is shifted downward). The decrease in aN is probably related to a lower polarity (the effective permittivity of the medium) near the nanoparticle surface where the probe is located. The theory predicts that the permittivity of a medium must have the same effect on both electroneutral (the deprotonated form of the radical) and charged (the protonated form) species (Griffith et al., 1974). We found that the rise in the Cu2+ amount from 0 to 0.36 mmol/g. of the gel results in an increase in the ∆pKa value. The shift of this curve (∆pKa) to the left is significantly smaller that for NR R3, although the change from the PZC to pH 4.3 and the incorporation of Cu2+ ions must enhance the positive charge of the nanoparticle surface and, according to formula (6), increase in ∆pKa. In accordance with formula (7), a reduction in the permittivity near the surface of TiO2 particles compared to εHOH, which causes the observed decrease in aN, must also shift the titration curve to the left.

Hence, the explicit inconsistency between the determined ∆pKa values for NR R2 and NR R3 radicals and the theoretical predictions is established. We relate this difference to a larger distance of NR R2 from the surface compared to NR R3; this is possible because NR R2 has a substituent in position 4, which is presumably positioned toward the surface due to the tendency of its amino groups to protonation (or complexation with Cu2+ ions sorbed on the surface). Thus, the –N–O• fragment of the radical turns out to be removed from the surface of TiO2 by the chain length of this substituent. The electric potential induced on the −N–O• fragment will decrease with the distance from the surface and will be inversely proportional to the distance from this surface, when it is represented as, for example, a plane. The small NR R3 radical can approach the nanoparticle surface much closer and, therefore, it is affected more strongly by the electrostatic potential of the TiO2 surface. The sorption of Cu2+ ions on the surface of nanoparticles increases the charge of the latter. The observed shift of the titration curve (Fig. 7) and the broadening of the ESR spectra of these radicals attest to the orientation of the substituent in position 4 of NR R2 toward the surface and to the interaction between amino groups of the substituent and the Cu2+ ions sorbed on the nanoparticle surface.

### **3.2. Organic supports of composite and hybride materials**

For PC and MCC, the titration curves of NR R1 were found to be shifted to the left of the calibration curve (Fig. 2, 3 ), indicating a posiive charge on the surface (Kovaleva et al., 2000 ; Molochnikov et al., 2007). Hydroxyl groups of celulose most likely play the role of surface bases whose protonation gives the surface a positive charge.

## **3.3. Solid-phase composites based on SiO2, TiO2 and cellulose powder**

### *3.3.1. Pure systems*

As shown in Fig. 8 and 9, the titration curves of NR R1 in the CMs TiO2 – PC with 70 (not shown) and 53 % wt. TiO2 and in the SiO2 – PC with 68 and 35 % wt. SiO2 (unprotonated parts) are shifted to the right relative to the calibration curve, as in the case of TiO2 and SiO2 xerogels. For CMs based on SiO2, TiO2 and cellulose powder ΔpH = pHext – pHint decreased (the curves are shifted to the left relative to those of pure xerogels) as the PC content in the samples increased; this corresponds to a decrease in the negative charge of the CM surface. This is because of the lower acidity of cellulose compared to the acidity of the solution (cellulose has basic alcohol functional groups in its structure) and the positive charge of its surface (Parshina, 2011).

electroneutral (the deprotonated form of the radical) and charged (the protonated form) species (Griffith et al., 1974). We found that the rise in the Cu2+ amount from 0 to 0.36 mmol/g. of the gel results in an increase in the ∆pKa value. The shift of this curve (∆pKa) to the left is significantly smaller that for NR R3, although the change from the PZC to pH 4.3 and the incorporation of Cu2+ ions must enhance the positive charge of the nanoparticle surface and, according to formula (6), increase in ∆pKa. In accordance with formula (7), a reduction in the permittivity near the surface of TiO2 particles compared to εHOH, which

Hence, the explicit inconsistency between the determined ∆pKa values for NR R2 and NR R3 radicals and the theoretical predictions is established. We relate this difference to a larger distance of NR R2 from the surface compared to NR R3; this is possible because NR R2 has a substituent in position 4, which is presumably positioned toward the surface due to the tendency of its amino groups to protonation (or complexation with Cu2+ ions sorbed on the surface). Thus, the –N–O• fragment of the radical turns out to be removed from the surface of TiO2 by the chain length of this substituent. The electric potential induced on the −N–O• fragment will decrease with the distance from the surface and will be inversely proportional to the distance from this surface, when it is represented as, for example, a plane. The small NR R3 radical can approach the nanoparticle surface much closer and, therefore, it is affected more strongly by the electrostatic potential of the TiO2 surface. The sorption of Cu2+ ions on the surface of nanoparticles increases the charge of the latter. The observed shift of the titration curve (Fig. 7) and the broadening of the ESR spectra of these radicals attest to the orientation of the substituent in position 4 of NR R2 toward the surface and to the interaction between amino

causes the observed decrease in aN, must also shift the titration curve to the left.

groups of the substituent and the Cu2+ ions sorbed on the nanoparticle surface.

**3.3. Solid-phase composites based on SiO2, TiO2 and cellulose powder** 

For PC and MCC, the titration curves of NR R1 were found to be shifted to the left of the calibration curve (Fig. 2, 3 ), indicating a posiive charge on the surface (Kovaleva et al., 2000 ; Molochnikov et al., 2007). Hydroxyl groups of celulose most likely play the role of surface

As shown in Fig. 8 and 9, the titration curves of NR R1 in the CMs TiO2 – PC with 70 (not shown) and 53 % wt. TiO2 and in the SiO2 – PC with 68 and 35 % wt. SiO2 (unprotonated parts) are shifted to the right relative to the calibration curve, as in the case of TiO2 and SiO2 xerogels. For CMs based on SiO2, TiO2 and cellulose powder ΔpH = pHext – pHint decreased (the curves are shifted to the left relative to those of pure xerogels) as the PC content in the samples increased; this corresponds to a decrease in the negative charge of the CM surface. This is because of the lower acidity of cellulose compared to the acidity of the solution (cellulose has basic alcohol functional groups in its structure) and the positive charge of its

**3.2. Organic supports of composite and hybride materials** 

bases whose protonation gives the surface a positive charge.

*3.3.1. Pure systems* 

surface (Parshina, 2011).

**Figure 8.** Titration curves for NR R1 in bulk solution (1), PC (2), TiO2 xerogel (3) и CMs TiO2 : PC (53% wt.TiO2)(4) and TiO2 : PC (43% wt.TiO2) (5)

**Figure 9.** Titration curves for NR R1 in bulk solution (calibration curve) (1), PC (2), SiO2 xerogel (3) and CMs SiO2 : PC (58% wt. SiO2) (4) and SiO2 : PC (35% wt. SiO2) (5). a, % = ((a-aNRH+)/(aNR-aNRH+))×100%

This proves that a surface of the composites studied carries lesser negative charge as compared to that of pure xerogels. The left-shift of the curve of NR R1 in CM TiO2 – PC with 43% wt.TiO2 relative to the calibration curve can be explained by initial positive charge of the surface of this CM due to binding H+ ions. Hence, SEP of CMs based on TiO2 and PC is varied over a wide range and even changes its sign from negative to positive as an increase in PC content. From the curves plotted, we notice that a decrease in percentage of TiO2 xerogel in CMs from от 53% до 43% leads to a pHintr greater than a pHext. This fact can be explained by positive charge of a surface of cellulose due to the presence of alcohol groups in its structure.

For all the synthesized CMs based on SiO2 xerogel and PC after complete protonating silanol groups, the titration curves of NR R1 were shifted to the left relative to the calibration curve and, hence, a surface of the samples studied remained positively charged (Fig.9). Unlike the composites based on TiO2 and PC, an increase in a percentage of PC in the SiO2-PC composites leads to changing a surface charge from positive to negative with increasing pHext (above the horizontal plateau on the titration curves) due to dissociation of functional groups. Thus, the incorporation of PC into the samples doesn't change acidity of silanol groups and doesn't make a polarizing effect on the SiO-H bond. The length of the horizontal plateau slightly increases in the accordance with the amount of silanol groups and the percentage of PC in the CMs based on SiO2. This can be caused by the increased dispersity of SiO2 due to rising in Ssp of the CMs (Table 1).

By varying the cellulose percentage in the composites and pHext values**,** pHint values and SEP can be selected over a wide range. This information is needed for optimization of the conditions for pH-dependent adsorption and catalytic processes through a choice of CM with a certain pHint and SEP as catalyst substrate and adsorbent.

### *3.3.2. Cu2+-containing composites*

The sorption of Cu2+ on the TiO2 and SiO2 xerogels and on the related CMs is accompanied by a change in the pH in their phases. However, the dependences of pHext(int) vs c**Cu** (amount of sorbed Сu2+ ions) for the above- mentioned types of xerogels and the CMs based have some differences (Fig.10, 11).

From these figures we notice that :


in its structure.

(Table 1).

This proves that a surface of the composites studied carries lesser negative charge as compared to that of pure xerogels. The left-shift of the curve of NR R1 in CM TiO2 – PC with 43% wt.TiO2 relative to the calibration curve can be explained by initial positive charge of the surface of this CM due to binding H+ ions. Hence, SEP of CMs based on TiO2 and PC is varied over a wide range and even changes its sign from negative to positive as an increase in PC content. From the curves plotted, we notice that a decrease in percentage of TiO2 xerogel in CMs from от 53% до 43% leads to a pHintr greater than a pHext. This fact can be explained by positive charge of a surface of cellulose due to the presence of alcohol groups

For all the synthesized CMs based on SiO2 xerogel and PC after complete protonating silanol groups, the titration curves of NR R1 were shifted to the left relative to the calibration curve and, hence, a surface of the samples studied remained positively charged (Fig.9). Unlike the composites based on TiO2 and PC, an increase in a percentage of PC in the SiO2-PC composites leads to changing a surface charge from positive to negative with increasing pHext (above the horizontal plateau on the titration curves) due to dissociation of functional groups. Thus, the incorporation of PC into the samples doesn't change acidity of silanol groups and doesn't make a polarizing effect on the SiO-H bond. The length of the horizontal plateau slightly increases in the accordance with the amount of silanol groups and the percentage of PC in the CMs based on SiO2. This can be caused by the increased dispersity of SiO2 due to rising in Ssp of the CMs

By varying the cellulose percentage in the composites and pHext values**,** pHint values and SEP can be selected over a wide range. This information is needed for optimization of the conditions for pH-dependent adsorption and catalytic processes through a choice of CM

The sorption of Cu2+ on the TiO2 and SiO2 xerogels and on the related CMs is accompanied by a change in the pH in their phases. However, the dependences of pHext(int) vs c**Cu** (amount of sorbed Сu2+ ions) for the above- mentioned types of xerogels and the CMs based have

pHint is different from pHext, and with no Cu2+ in an external solution (c**Cu =** 0), pHext is

 With an increase in the amount of Сu2+ in the CMs, both the pHext and pHint descrease. There is a horizontal plateau on the dependences of pHint on c**Cu** in the range of sorbed Cu2+ from 0.15 up to 0.3 mmol Cu2+/g. TiO2, SiO2 xerogel, within which pHint remains

At greater c**Cu** pHint and pHext come closer and become almost equal.

with a certain pHint and SEP as catalyst substrate and adsorbent.

*3.3.2. Cu2+-containing composites* 

some differences (Fig.10, 11).

greater than pHint.

From these figures we notice that :

constant with increasing in c**Cu.**

**Figure 10.** pH of external solution (pHext) (curves 1,3) and near the surface (pHint) (curves 2,4) of TiO2 xerogel (curves 1, 2) and the TiO2 : PC (47% wt. TiO2) composite (curves 3,4) vs.cCu

**Figure 11.** pH of external solution (pHext) (curves 1,3) and near the surface (pHint) (curves 2,4) of SiO2 xerogel (curves 1,2) and the SiO2 : PC (68% wt. SiO2) composite (curves 3,4) vs.cCu

The increase in acidity of external solution, first of all, is caused by the hydrolysis of Cu2+ at the initial pHext equal to 6-7 and is accompanied by intense release of H+ ions. The lesser pHint values as compared to pHext for both types of xerogels and the related CMs can be explained by a surface negative charge. A surface of the samples studied attracts H+ ions in its double electrical layer to compensate this charge. Also an increase in acidity of both external and internal (near the surface) solutions is caused by the fact that the sorption of Cu2+ ions is accompanied by intense release of H+ ions into them. The increase in the acidity of a medium due to the competition of H+ ions almost completely suppresses the ion exchange sorption of Cu2+. An analogous tendency was previously observed for organic ion-exchange resins (as grains) and TiO2 hydrogel (Kovaleva et al., 2000 ; Molochnikov et al., 2004).

The presence of the horizontal plateaus on the curves pHint vs c**Cu** (Fig. 9) indicates the buffer properties of the TiO2 xerogel and the TiO2 : PC samples by analogy with the same properties of the mixed H+–Na+ form of the KB-2 × 4 cationite (Kovaleva et al., 2000). The greater is percentage of PC in a sample, the shorter is the horizontal plateau on the abovementioned curves. The buffer effect of the TiO2 xerogel and the related CMs is caused by the existence of some amount of deprotonated functional groups in the samples before the sorption of Cu2+. The complexation of Cu2+ ions with these groups occurs first of all and does not lead to the release of H+ ions, which would reduce both pHext and pHint. In addition, releasing of hydroxide groups as a result of dissociation of molecules of Cu(OH)2 responsible for complexation with functional groups of a surface of xerogel can retard decreasing pH with increasing c**Cu**.

Unlike TiO2 systems, for SiO2 and the related CMs pHint values were found to be more than pHext ones (Fig. 11). Apparently, this differance can be explained by the significant differences in pKa of functional groups of SiO2 and TiO2 oxides. pKa values of silanol groups (3.95 0.07) are significantly lower than pHext from which sorption of Cu2+ has been conducted. Therefore, silanol groups are completely dissociated and H+ ions don't exchange with Cu2+ ions during complexation. Hence, no functional groups can neutralize OH—groups forming during breakdown of Cu(OH)2 molecules. As a result, pHint values (near a surface of samples) have high values.

As the titratuon curves for TiO2-based CMs don't contain the horizontal plateaus as well as pure systems, hence, pKa values of active acidic centers of a surface of these materials are out of zone of sensitivity of the NR used (greater than 7) (Fig. 8). Therefore, these centers are mainly in the protonated form before sorption of Cu2+-ions. During complexation Cu2+-ions exchange with H+ ions of functional groups releasing them into external solution. Hence, pHint became less than pHext. It shows up in a general decrease pH (both pHint and pHext) with increasing c**Cu** (Fig. 10 ).

## **3.4. Hybrid organo-inorganic materials based on the chitosan– Al2O3, chitosan– SiO2, and cellulose –chitosan–systems**

Hybrid systems were obtained in accordance with the scheme (Fig.12)

**Figure 12.** Scheme for obtaining hybrid organo-inorganic materials. s -support

The compositions of the obtained hybrid systems were characterized by elemental analysis (Table 2). According to the obtained data, the reactivity of substrates was different and declined in the order MCC >Al2O3 > SiO2, since the compositions of the obtained hybrids in the case of inorganic oxides do not correspond to the molar ratios of the initial components.

The presence of chitosan molecules on the surface of the obtained systems was confirmed by the IR spectra, which contain characteristic absorption bands at 1652 and 1555 cm–1 corresponding to residual acetoamide groups of the polymer.

Since the obtained composites have to be electroneutral, we must assume that in the case of SiO2, the substrate, being a weak acid, forms an ionic bond with the chitosan's amino group. In case of Al2O3, the same situation is possible, but some of the chitosan's amino and hydroxyl groups are involved in complex-forming with aluminum. In the case of MCC, the interaction occurs through hydrogen bonds.

NR were used as pH-probes to obtain a more detailed characterization of the surface structure of hybrid systems, their acid–base properties.

A comparison of the ESR spectra of NR R1 in solution and in the samples studied showed that in all cases, there was an isotropic signal which indicated the lack of covalent bonds between the NR and the sample surface and the presence of this radical near the surfaces of particles (Fig. 4, spectrum 1).

### *3.4.1. Pure hybride systems*

232 Nitroxides – Theory, Experiment and Applications

decreasing pH with increasing c**Cu**.

with increasing c**Cu** (Fig. 10 ).

**SiO2, and cellulose –chitosan–systems** 

high values.

The increase in acidity of external solution, first of all, is caused by the hydrolysis of Cu2+ at the initial pHext equal to 6-7 and is accompanied by intense release of H+ ions. The lesser pHint values as compared to pHext for both types of xerogels and the related CMs can be explained by a surface negative charge. A surface of the samples studied attracts H+ ions in its double electrical layer to compensate this charge. Also an increase in acidity of both external and internal (near the surface) solutions is caused by the fact that the sorption of Cu2+ ions is accompanied by intense release of H+ ions into them. The increase in the acidity of a medium due to the competition of H+ ions almost completely suppresses the ion exchange sorption of Cu2+. An analogous tendency was previously observed for organic ion-exchange

resins (as grains) and TiO2 hydrogel (Kovaleva et al., 2000 ; Molochnikov et al., 2004).

The presence of the horizontal plateaus on the curves pHint vs c**Cu** (Fig. 9) indicates the buffer properties of the TiO2 xerogel and the TiO2 : PC samples by analogy with the same properties of the mixed H+–Na+ form of the KB-2 × 4 cationite (Kovaleva et al., 2000). The greater is percentage of PC in a sample, the shorter is the horizontal plateau on the abovementioned curves. The buffer effect of the TiO2 xerogel and the related CMs is caused by the existence of some amount of deprotonated functional groups in the samples before the sorption of Cu2+. The complexation of Cu2+ ions with these groups occurs first of all and does not lead to the release of H+ ions, which would reduce both pHext and pHint. In addition, releasing of hydroxide groups as a result of dissociation of molecules of Cu(OH)2 responsible for complexation with functional groups of a surface of xerogel can retard

Unlike TiO2 systems, for SiO2 and the related CMs pHint values were found to be more than pHext ones (Fig. 11). Apparently, this differance can be explained by the significant differences in pKa of functional groups of SiO2 and TiO2 oxides. pKa values of silanol groups (3.95 0.07) are significantly lower than pHext from which sorption of Cu2+ has been conducted. Therefore, silanol groups are completely dissociated and H+ ions don't exchange with Cu2+ ions during complexation. Hence, no functional groups can neutralize OH—groups forming during breakdown of Cu(OH)2 molecules. As a result, pHint values (near a surface of samples) have

As the titratuon curves for TiO2-based CMs don't contain the horizontal plateaus as well as pure systems, hence, pKa values of active acidic centers of a surface of these materials are out of zone of sensitivity of the NR used (greater than 7) (Fig. 8). Therefore, these centers are mainly in the protonated form before sorption of Cu2+-ions. During complexation Cu2+-ions exchange with H+ ions of functional groups releasing them into external solution. Hence, pHint became less than pHext. It shows up in a general decrease pH (both pHint and pHext)

**3.4. Hybrid organo-inorganic materials based on the chitosan– Al2O3, chitosan–**

Hybrid systems were obtained in accordance with the scheme (Fig.12)

Analysis of the structure of the Al2O3–chitosan system according to the adsorbed probe molecules of the nitroxyl radical showed that coating the initial -Al2O3 substrate with chitosan leads to a slight increase in the surface negative charge, as was confirmed by the shift of the NR titration curve to the right (Fig. 13). The titration curves of -Al2O3 and Al2O3–chitosan are not parallel to the calibration curve at low pH values due to the gradual dissolving of -Al2O3 particles in weakly acidic media, as has been noted during investigations of samples containing aluminate ions (Iller, R., 1979 ; Golovkina, 2009).

**Figure 13.** Titration curve for NR R1 in aqueous bulk solution ( calibration curve) (1), near the surface of Al2O3 (2), hybride Al2O3– chitosan (3) and Al2O3– chitosan-Co2+ (4) systems . a, % = ((a-aNRH+)/(aNRaNRH+))×100%

For the SiO2–chitosan system, the NR titration curve overlaps the corresponding curve for SiO2 values within the horizontal section of the titration curve for silandiol groups in the range of high pH (Fig. 14). We can correspondingly claim that in this pH region, the surfaces of the samples all have the same charge, demonstrating that chitosan has no influence on the SiO2 surface. If pH is decreased below 6, the titration curve for the SiO2–chitosan system shifts to the right of the corresponding curve for SiO2, indicating the negative charge of the surface. The horizontal pleteau of the titration curve for silanol groups is lower in the case of the SiO2– chitosan system (pH 4.5–5), and the рКа value for silanol groups in the presence of chitosan falls slightly. A similar reduction in this parameter has been observed during modification of the surface of mesoporous molecular sieves based on SiO2 with aluminate and borate ions (Golovkina et al., 2009). Below the titration curve plateau for silanol groups (pH < 4.2), the titration curves of SiO2 and SiO2–chitosan overlap again, indicating a similar surface charge.

The deposition of chitosan on MCC causes the titration curve to shift to the right of to both the calibration curve and the NR titration curve of the initial MCC sample (Fig. 15), demonstrating the negative charge of the surface, as was observed for the Al2O3–chitosan and SiO2–chitosan systems. A peculiarity of this substrate is the recharging of the surface, from positive for MCC to negative for the MCC–chitosan system during chitosan deposition. In this case, the type of change in the surface charge occurring upon a reduction in pH is the same for both the MCC and the MCC–chitosan systems. Changing symbatically with the curve typical of the initial MCC sample, the titration curve for the MCC–chitosan system almost overlaps the calibration curve in the acidic region (pH ≤4.5). In this pH range, the potential of the MCC–chitosan system surface is therefore close to zero.

234 Nitroxides – Theory, Experiment and Applications

aNRH+))×100%

**Figure 13.** Titration curve for NR R1 in aqueous bulk solution ( calibration curve) (1), near the surface of Al2O3 (2), hybride Al2O3– chitosan (3) and Al2O3– chitosan-Co2+ (4) systems . a, % = ((a-aNRH+)/(aNR-

For the SiO2–chitosan system, the NR titration curve overlaps the corresponding curve for SiO2 values within the horizontal section of the titration curve for silandiol groups in the range of high pH (Fig. 14). We can correspondingly claim that in this pH region, the surfaces of the samples all have the same charge, demonstrating that chitosan has no influence on the SiO2 surface. If pH is decreased below 6, the titration curve for the SiO2–chitosan system shifts to the right of the corresponding curve for SiO2, indicating the negative charge of the surface. The horizontal pleteau of the titration curve for silanol groups is lower in the case of the SiO2– chitosan system (pH 4.5–5), and the рКа value for silanol groups in the presence of chitosan falls slightly. A similar reduction in this parameter has been observed during modification of the surface of mesoporous molecular sieves based on SiO2 with aluminate and borate ions (Golovkina et al., 2009). Below the titration curve plateau for silanol groups (pH < 4.2), the titration curves of SiO2 and SiO2–chitosan overlap again, indicating a similar surface charge.

The deposition of chitosan on MCC causes the titration curve to shift to the right of to both the calibration curve and the NR titration curve of the initial MCC sample (Fig. 15), demonstrating the negative charge of the surface, as was observed for the Al2O3–chitosan and SiO2–chitosan systems. A peculiarity of this substrate is the recharging of the surface, from positive for MCC to negative for the MCC–chitosan system during chitosan

**Figure 14.** Titration curve for NR R1 in aqueous bulk solution ( calibration curve) (1), near the surface of SiO2 (2), of the hybride SiO2– chitosan (3), of the SiO2– chitosan-Co2+ (4) systems. a, % = ((a-aNRH+)/(aNRaNRH+))×100%

The deposition of chitosan on the substrate always causes the titration curve of the radical near the surface of hybrid material to shift to the right; i.e., it leads to a negative charge on the surface. While the deposition of chitosan leads to relatively slight changes in the surface potential in the case of inorganic substrates, these changes are so great in the case of MCC that they even lead to changes in the surface charge.

Some of differences in the behavior of titration curves for chitosan-containing materials on the inorganic substrates that occur in the range of high pH values could be due to the close рКа values of the silandiol groups of the SiO2 substrate (рКа = 6.5–6.8) and of the chitosan (рКа = 6.42 (Skorik et al., 2003)). When the pH falls, amino groups of chitosan or silandiol groups become protonated and form hydrogen bonds with one another. As a result, the charge of the initial substrate surface and that of the hybrid material become similar. During further protonation of silandiol and amino groups at the same pH range, this interaction becomes impossible, and the surface charge of the SiO2–chitosan hybrid material becomes more negative than that of the SiO2.

**Figure 15.** Titration curve for NR R1 in aqueous bulk solution ( calibration curve) (1), near the surface of MCC(2), of the hybride MCC– chitosan (3), of the MCC– chitosan-Co2+ (4) systems . a, % = ((a-aNRH+)/(aNRaNRH+))×100%

This is in complete agreement with the relative positions of the radical titration curves for Al2O3 and Al2O3–chitosan. It must be emphasized that the рКа dissociation value of hydroxyl groups of Al2O3 is higher, and lies outside the radical sensitivity range.

Deposits of chitosan on the MCC substrate breaks it into smaller particles (Table 2), due most likely to the recharging of the surface from positive to negative. Chitosan molecules are appreciably smaller in size than MCC molecules, that is why a negative charge leads to increased repulsion inside large MCC particles, to their disintegration, and finally to an increase in the total surface area of the material. Chitosan thus plays the role of a disaggregating surfactant. In case of inorganic substrates, the deposition of chitosan leads to an increase in the particle size, and consequently to a decrease in their surface area (Table 1). Chitosan is therefore a weakly aggregating surfactant as to inorganic substrates.

The рКа value for amino groups of chitosan is 6.42 (Skorik et.al, 2003). In the investigated pH range (pH ≤ 6), polymer molecules must consequently be positively charged due to the formation of NH3+ ammonium groups**.** If chitosan molecules were adsorbed on the surface of inorganic oxides so that the protonated amino groups were on the material surface, this kind of adsorption would be reflected in the titration curves of the modified samples and would cause the titration curve to shift to the left of the titration curves of the initial oxides. Since this did not happen, we can assume that during the formation of hybrid material, chitosan amino groups form weak hydrogen bonds or Van der Waals bonds with functional groups of substrates (organic or inorganic), leading to an arrangement of chitosan molecules in which glucosamine rings are turned in the direction of the substrate.

### *3.4.2. Co2+ -containing hybride systems*

236 Nitroxides – Theory, Experiment and Applications

aNRH+))×100%

**Figure 15.** Titration curve for NR R1 in aqueous bulk solution ( calibration curve) (1), near the surface of MCC(2), of the hybride MCC– chitosan (3), of the MCC– chitosan-Co2+ (4) systems . a, % = ((a-aNRH+)/(aNR-

This is in complete agreement with the relative positions of the radical titration curves for Al2O3 and Al2O3–chitosan. It must be emphasized that the рКа dissociation value of

Deposits of chitosan on the MCC substrate breaks it into smaller particles (Table 2), due most likely to the recharging of the surface from positive to negative. Chitosan molecules are appreciably smaller in size than MCC molecules, that is why a negative charge leads to increased repulsion inside large MCC particles, to their disintegration, and finally to an increase in the total surface area of the material. Chitosan thus plays the role of a disaggregating surfactant. In case of inorganic substrates, the deposition of chitosan leads to an increase in the particle size, and consequently to a decrease in their surface area (Table 1).

The рКа value for amino groups of chitosan is 6.42 (Skorik et.al, 2003). In the investigated pH range (pH ≤ 6), polymer molecules must consequently be positively charged due to the formation of NH3+ ammonium groups**.** If chitosan molecules were adsorbed on the surface of inorganic oxides so that the protonated amino groups were on the material surface, this kind of adsorption would be reflected in the titration curves of the modified samples and

hydroxyl groups of Al2O3 is higher, and lies outside the radical sensitivity range.

Chitosan is therefore a weakly aggregating surfactant as to inorganic substrates.

The presence on the support surface of a polymer that is capable of acquiring metal ions of high coordination numbers is needed to fix these ions more firmly on metal-containing hybride materials. This role was played by chitosan on the surfaces of the inorganic (SiO2, Al2O3) and organic (MCC) supports. Co2+ ions were sorbed from an aqueous ethanol solution (Fig. 16).

The composition of the studied systems was characterized by elemental analysis. According to the data, the sorption capacity of the hybrid systems relative to Co2+ ions is different and declines in the following order: SiO2 >Al2O3 > MCC. We should note that in this case, there is a difference between the mechanisms of binding ions for different supports, since their surfaces have the same coating influencing the process.

As follows from the elemental analysis, chitosan plays the role of a complexing agent. In this case, all amino groups are involved in the coordination by cobalt ions (ratio Co:NH2 = 1:1-1.5).

In the case of the Al2O3–chitosan system, some of the amino groups do not participate in the complexation, but the Co : Cl ratio corresponds to the composition of the initial salt that provides the coordination mechanism for binding cobalt ions (Fig.17). In the case of the SiO2–chitosan system, the Co : Cl ratio is ~1. The need to obey the law of electroneutrality for the obtained hybrid systems requires assuming that the support, in addition to being a weak acid, also plays a coordinating role with respect to Co2+ ions, which accords with the covalent mechanism of binding Сo2+ ions (Fig. 17). In the case of the MCC–chitosan system, metal ions are sorbed as a basic salt, and hydroxyl groups on the surface provide weaker coordination binding than Al2O3.

**Figure 16.** Scheme for synthesis of cobalt-containing hybrid systems. s - support

**Figure 17.** Scheme for the mechanism of binding Co2+ by hybrid systems

The elemental composition of the surface shows that chitosan does not cover it completely and some part of the functional groups remains on the support surface in free form. The more complex structure of the surface of hybrid system was characterized by investigating its acid–base properties via ESR spectroscopy of the NR used as pH probes. The titration curves of NR R1 on the surface of the Co2+-containing system are shown in Figs. 13-15. The shift in these curves relative to the calibration curve left or right allows us to determine whether the surface has a positive or negative charge, respectively (Kovaleva et al., 2000 ; Molochnikov et al., 2007).

The technique for introducing Co2+ ions into the hybrid materials from the ethanol solutions could lead to the formation of sediments of basic cobalt chloride or chloride–alcoxide micelles on the surface of these materials. In the case of the MCC–chitosan (Fig. 15) and Al2O3–chitosan (Fig. 13) systems, the titration curves of the cobalt-containing materials at pH > 5 are shifted slightly to the left relative to the titration curves of the initial samples. Consequently, the occurrence of basic cobalt chloride on the surface of these materials does not lead to a change in the surface charge, but changes its value slightly. This is due to the effective neutralization of a positive charge of Co2+ ions by negative chloride and hydroxide ions. In order to confirm our hypothesis, we present the published data on studying the sorption of Co2+ ions from aqueous solutions and on the nature of the interaction between the metal center and amino groups of chitosan. According to (Minimisawa et al., 1999) maximum adsorption starts to decrease with increasing pH due to the formation of cobalt hydroxocomplexes. The maximum sorption of cobalt ions by chitosan found at pH 6–8 in (Silva et.al, 2008) is in good agreement with the data from (Minimisawa et al., 1999) and is determined by the formation of Со(ОН)2 phase or slightly soluble basic salts. No chemical interaction with amino groups of chitosan occurs in this case (Zhao et al., 1998).

The SiO2–chitosan system behaves differently. Even at the highest pH values, the titration curve of the sample modified with cobalt (Fig. 17) was appreciably shifted to the left relative to the NR curve of the initial sample, although it remains to the right of the calibration curve. Accordingly, the formation of the chloride hydroxyl cobalt micelles immediately leads to a considerable reduction in the negative surface charge of the SiO2–chitosan system. The different behavior of the titration curve for system III (Table 2) is most likely associated with the initially lower amount of Cl– ions in the cobalt micelles (Co : Cl = 1 : 1), i.e., to its higher amount of ОН– ions. Acid sites on the SiO2 surface (silanol groups) also likely interact with the basic cobalt chloride particles precipitating on the surface, thereby replacing the Cl– ions. In both cases, the titration curve must shift left due to the neutralization of the negative surface charge. In the study of the samples in the aqueous medium at high pH values, the transformation for systems I–III (Table 2) thus occurs; hydration of the surface, followed by hydrolysis leading to the formation of colloidal particles based on chloro-hydroxo complexes of Co2+ ions, is observed.

These micelles of Co2+ chloro-hydroxocomplexes begin to dissolve at pH ≤ 5 for all investigated systems. The dependence of sorption of cobalt ions on pH (Minimisawa et al., 1999) shows that the maximum adsorption starts to decline at pH < 5, and Co2+ ions in the solution are in the form of aqua complexes (Zhao et al., 1998). The latter are sorbed by the chitosan primary amino groups, thereby charging the surface positively; in this case, the titration curves of systems I–III (Table 2) (Figs.13-15) are shifted left relative to the corresponding curves for the initial samples and to the calibration curve. A further reduction in pH leads to the neutralization of ОН– groups, and no further changes in pH near the particle surface are observed upon a change in pHext.

This indicates the presence of a horizontal plateau on the titration curve. The value of the horizontal section is lower for system III than for systems I and II (Table 2), since some of the ОН– groups are replaced by residues of silicic acid. At high pH values, Co2+ ions initially cause a substantial decrease in the negative surface charge for the case of system III; their transformation into the form coordinated by the chitosan - NH2 groups thus influences the surface charge to a lesser degree upon declining pH. Our results indicate the participation of chitosan amino groups in the complexation with Co2+ ions. During the interaction of the hybrid materials with the Co2+-containing solution, at least a part of glucosamine rings are consequently turned outward and are capable of becoming ligands.
