Amorphous and ordered organosilicas functionalized with amine groups as sorbents of platinum (II) ions

The following compounds were used: tetraethoxysilane (TEOS, 99 %, ABCR), 3-aminopropyltriethoxysilane (APTS, 98 %, ABCR), N-[3-(trimethoxysilyl)propyl]-ethylenediamine (TMPD, 97 %, ABCR), NH₄F (98 %, POCH), HCl (36 %, POCH), NaOH (POCH), ethanol (EtOH, 99.8 %, POCH), Pluronic P123 (P123, Sigma-Aldrich), potassium tetrachloroplatinate(II) (>99.9 %, Sigma-Aldrich), sodium chloride (POCH), thiourea (POCH), NaCl (POCH), KNO₃ (POCH). All chemicals were used as received, without further purification.

The synthesis of these samples was described by us before (Barczak et al. 2009, 2010). In the model synthesis, 2 g of P123 was dissolved in 60 mL of 2 mol/L HCl and 11 mL of deionized water under vigorous stirring at 40 °C. After 8 h of stirring TEOS was added dropwise to this solution. After 15 min the APTS (samples M1 and M2) or TMPD (samples M3 and M4) was added dropwise. The resulting mixture was stirred at 40 °C for 24 h and aged at 100 °C for next 48 h. The white solid was washed with deionized water, filtered and dried at 70 °C. The template was removed by three-time extraction with the acidified EtOH at 70 °C. The numeration of the samples was as follows: M0 (TEOS = 20 mmol = 20/0), M1 (TEOS/APTS = 19/1), M2 (TEOS/APTS = 18/2), M3 (TEOS/TMPD = 19/1), M4 (TEOS/TMPD = 18/2).

The syntheses of these samples were similar to those described previously (Zub et al. 2010; Barczak and Borowski 2009). In the model synthesis, a solution of NH₄F was added to the solution of TEOS in 15 mL of EtOH. After 5 min the solution of APTS (samples X1 and X2) or TMPD (samples X3 and X4) in 15 mL of EtOH was added to this mixture. Opalescating gels were formed in a few minutes. After two days, the gel was pounded and dried in vacuum at 30/65/105 °C—2 h at each temperature.

The CHN elemental analysis was carried out using the Perkin Elmer CHN 2400 analyzer. Powder X-ray diffraction (XRD) patterns were recorded using a Seifert RTG DRON-3 diffractometer (CuKα radiation) with 0.02º size step and 10 s time step covering the range of 0.5º < 2θ < 5.0º at room temperature. The SEM analysis of randomly selected parts of the surface was performed by means of LEO SEM 1430 VP microscope. The accelerating voltage was set to 20 keV, and the beam current to 80 mA.

Nitrogen adsorption/desorption isotherms were measured at −196 °C using an ASAP-2405 N analyzer. All samples were degassed at 110 °C in vacuum prior to measurements. BET specific surface areas (S_BET) were evaluated in the range of relative pressures p/p_o of 0.05–0.25 (Brunauer et al. 1938). Total pore volumes (V_p) were calculated by converting the amount adsorbed at p/p_o ~0.99 to the volume of liquid adsorbate. Average pore sizes (d_BJH) were estimated using the BJH method (Barrett et al. 1951), while pore size distributions (PSD) from the desorption branches of isotherms using the same approach.

XPS spectra were obtained using a Multi-chamber analytical system (Prevac, Poland) with monochromated Kα-Al radiation (1486.6 eV) (Gammadata Scienta, Sweden) and an X-ray power of 450 W. The binding energy (BE) scale was referenced against C1 s = 284.7 eV line. The vacuum in the analysis chamber was 8 × 10⁻⁹ Pa. The high-resolution scans were performed over the 69–81, 99–110, 192–206, 282–294, 395–407, 527–538 eV ranges for Pt 4f, Si 2p, Cl 2p, C1 s, N1 s and O1 s spectra, respectively.

The batch mode was used to examine the sorption of Pt(II) ions. In every experiment, 50 mg of adsorbent and 50 mL of platinum solution were stirred at 25 °C. After the adsorption equilibrium had been attained, the solution was separated from the sorbent by centrifugation. The equilibrium adsorption quantity was calculated according to the following formula:where c _i is the initial Pt(II) concentration (in mg/L), c is the equilibrium Pt(II) concentration (in mg/L), V is the volume of the solution (in L) and m is the mass of the adsorbent (in g). Measurements were carried out using the atomic absorption spectrometer AAS-3 (Carl Zeiss) equipped with a deuterium-lamp background corrector, an EA-3 electrothermal atomizer, and an MPE autosampler. Measurements were performed using the Pt hollow cathode lamp (Varian) at a wavelength of 265.9 nm; 0.2 nm slit width was used. The time–temperature program applied in Pt(II) determination was as follows: drying: 120 °C for 10 s; pyrolysis: 1200 °C for 5 s; atomization: 2600 °C for 2 s. Pure argon was used as the inert gas with a flow rate of 280 mL/min except the atomization stage.

The scheme of the co-condensation of TEOS with functionalized silanes (APTS or TMPD) is presented in Fig. 1. The molar ratios of TEOS/APTS and TEOS/TMPD in the reacting systems were 19:1 and 18:2. Such ratios were selected to achieve the high content of functional groups in the final materials and, in addition, to avoid fast gelation, which is often observed (even in the case of OMO samples) when the relative amount of amine-functionalized silane increases (Dąbrowski et al. 2005). All synthesized samples (OMO and AFX) are non-swelling in common non-aqueous solvents.

The content of nitrogen (and thus amine groups) in the synthesized samples is presented in Table 1. For all APX samples the nitrogen content is around 70–80 % of the theoretical values (calculated from the initial amounts of co-condensing alkoxysilanes). In the case of OMO samples, the nitrogen content is around 50–60 % of the values estimated on the basis of the initial molar composition, indicating less effective incorporation of amine groups than in the case of APX. Apparently, a part of the amine groups is lost during ageing, drying and extraction steps. It can be also seen that in the case of OMO samples there is higher carbon content due to incomplete removal of the template. This is confirmed by the FTIR spectra (not given here) where a group of absorption bands in the region of 2860–3030 cm⁻¹, is attributed to the stretching modes of CH₂ groups of poly(oxyethylene) and poly(oxypropylene) blocks of Pluronic P123 (Barczak et al. 2010).

The nitrogen adsorption–desorption isotherms of the APX and OMO samples are shown in Fig. 2. All the OMO samples exhibit the type IV isotherms according to the IUPAC classification (Rouquerol et al. 1999). The most striking feature is the presence of sharp capillary condensation step, reflecting capillary condensation of adsorbate in the uniform mesopore channels, and evaporation step related to the evacuation of adsorbate from the pores. The isotherms of the samples M1‐M4 exhibit hysteresis loops which is attributed to the presence of mesopores in the material. Those loops can be classified as belonging to the type H1 (Rouquerol et al. 1999). Thus, a framework of these materials has a uniform array of mesopores with the same diameter, which is consistent with the XRD and TEM data (vide infra).

Totally different shapes of the isotherms are observed for the APX samples and their attribution to any IUPAC group is ambiguous. The isotherms of samples X1 and X2 exhibit neither capillary condensation, nor hysteresis loop step which means that there are no mesopores, while in the case of X3 a small hysteresis loop can be observed indicating the presence of some mesopores. In the case of the sample X4 a steep increase in the adsorption volume of N₂ is observed at p/p _o close to 0.9. It can be assigned to the presence of intermolecular mesopores between the silica nanospheres (Sing et al. 1985). The hysteresis loop of X4 is attributed to the presence of mesopores in the material and is typical of the samples consisting of agglomerates of spheroidal particles of fairly uniform size and array.

The structure-adsorption characteristics are presented in Table 1. The OMO samples have much higher values of the specific surface area (S_BET) than corresponding the APX. In the case of OMO, the values of S_BET are in the range 740–840 m²/g while in the case of APX—280–520 m²/g. The sizes of the ordered mesopores of OMO are in the range 6.5–6.8 nm while for APX −3.7–12.1 nm. Pore volumes are much higher in the case of the OMO samples (1.29–1.43 cm³/g) when compared with the corresponding APX samples (0.06–0.48 cm³/g). Only the sample X4 has a relatively high pore volume (0.90 cm³), not as high as the corresponding M4 sample though (1.30 cm³). The analysis of the values of all parameters confirms deep structural differences between both groups of samples. When comparing the samples within the series, it can be seen that the differences between the values of S_BET, V_p and d_BJH are much lower, larger differences are observed for the APX samples—particularly the sample X4 differs from the remaining APX samples due to the presence of mesopores.

The pore size distributions of all samples are presented as inlets in Fig. 2. For the OMO samples two maxima are present: the first one, a very narrow maximum at ~10 nm corresponding to the primary ordered mesopores, and the second one located at ~2.5 nm, reflecting the presence of a network of smaller interconnecting pores. For the APX samples X1-X3 wide pore size distributions are observed with one small maximum at ~3.5–4 nm, while for the X4 sample—at 25, nm indicating totally different porous structure of the latter sample when compared with the other samples from the APX series.

The XRD patterns (not given here) show typical patterns for both series i.e. ordered and amorphous silicas. In the case of OMO samples three well-resolved peaks are usually observed in the range 2θ ≈ 0.8–2°. These peaks can be indexed according to the hexagonal p6m symmetry, indicating a SBA-15 mesostructure: one sharp reflection at 2θ ≈ 0.8 is indexed as (100) and two minor but distinct reflections at 2θ ≈ 1.4 and 1.6, are indexed as (110) and (200), respectively. All three signals disappear gradually with the increasing content of amine groups, which means that the ordering of the samples deteriorates. In the case of APX no signals in the low-angle range are observed demonstrating the lack of any type of ordering and amorphous character of the APX samples.

The SEM micrographs of the samples M1, X1 and X4 are presented in Fig. 3. All the OMO samples are composed of irregular “needle-like” particles having worm-like structure, i.e. parallelly oriented hexagonal motifs. The APX samples are composed of aggregated spherical particles. In the case of the sample X4 empty spaces between the aggregates make this sample less compact, when compared to the other samples from the APX series.

First, the effect of pH on Pt(II) adsorption onto the synthesized materials was studied. The series of solutions of the same platinum concentration were prepared and proper initial pH (pH_in) values were set up by adding the hydrochloric acid or sodium hydroxide. Then the Pt(II) sorption measurements were performed for each solution. In the case of APTS-functionalized samples, the highest adsorption capacity is obtained for the equilibrium pH (pH_eq) of 2.5 and 3.3 for M2 and X2, respectively (Fig. 4). However, the shapes of both curves are different. In the case of sample M2 the changes of the Pt(II) uptake are strongly pH-dependent, showing a narrow maximum. In contrast, X2 curve exhibits a wide maximum covering the range ~1.8–4.4. It is worth noting that in the case of sample M2 insignificant differences in pH_eq are observed for pH_in between 3.5 and 7. It means that in this range a buffering effect occurs. For the sample X2 analogus phenomenon is observed for pH_eq ≈ 5.

The effect of pH on Pt(II) uptake by the TMPD-functionalized samples is shown in Fig. 4. The maximum uptake for M4 is achieved at pH_eq ≈ 2.5, and for X4 the plateau is observed in the range of pH_eq ~ 0.8–5.5 and is slightly pH-dependent for both materials. Outside this range a significant decrease in the adsorbed Pt(II) species is observed. Similarly to the materials functionalized by APTS, the buffering effect in the range of pH_eq ~ 3.5 for M4 and ~5.5 for X4, is observed.

The effect of time on adsorption of Pt(II) ions onto the synthesized materials was studied and is presented in Fig. 5. For the APTS-functionalized samples the rate of the Pt(II) adsorption is significantly slower for APXs than for OMOs. The adsorption equilibrium for the samples M1 and M2 is achieved after 5 and 10 h, respectively, whereas for the samples X1 and X2 after 72 and 20 h, respectively. For X3 only a 45 min interval is required for reaching the equilibrium, and 18 h are necessary for M3. Very slow kinetics for X1 and X2 compared to M1 and M2 (cf. Fig. 5) is probably associated with a considerable content of the amine groups localized in the micropores for APX materials. In the case of OMO samples, the existing micropores follow from the penetration of the silica framework by hydrophilic parts of P123, thus there should be no considerable number of nitrogen centres located in the micropores. Moreover, in the case of X1 and X2 samples there are no mesopores facilitating the transport of Pt(II) ions while in the case of all OMO samples mesopores are predominant in the porous structure. As a result, there is a slow transport rate of Pt(II) ions to the micropores of the adsorbent. In the case of X3 and X4 the amount of mesopores is higher than for X1 and X2 (cf. Fig. 2) Therefore the diffusion into the micropores is facilitated and the adsorption kinetics is much faster. However, despite the relevant mesoporosity in the case of X4 sample as compared with X3, the time needed for reaching equilibrium state is comparable. This suggests that the chemical reactions occurring during the adsorption process are relatively slow constituting the limiting factor for the TMPD-functionalized APX samples.

In order to determine the selectivity of the Pt(II) sorption from aqueous solutions onto the amine-functionalized silicas, the effect of chloride ions on Pt(II) adsorption ability was studied in detail. In Fig. 6 the effect of chlorides on the adsorption of Pt(II) is presented. In the case of M2, M4 and X2 the presence of chlorides concentration higher than 0.01 mol/L causes drastic decrease of the adsorption by about 80–90 %. In turn, in the case of X4 the presence of chlorides at the concentration below 0.5 mol/L does not cause the decrease of Pt(II) adsorption. For higher concentrations of chlorides the adsorption of Pt(II) on X4 decreases almost proportionally to the chlorides content up to 70 % for 3 mol/L chlorides solution. Different behaviour of X4 seems to be the effect of higher adsorption pH compared to M2, M4 and X2. The content of amine groups for X4 is the highest comparing with other studied sorbents. Therefore, it appears that pH in the vicinity of sorbent surface is higher than in the bulk solution and as a result, the hydrolysis of PtCl₄ ²⁻ complex to more stable aqueous and hydroxy platinum complexes takes place. For the higher pH value deprotonation of surface amine groups facilitates the adsorption of platinum complexes.

The adsorption ability of APTS-functionalized OMO and APX has been determined by measuring the Pt(II) uptake from its aqueous solutions. The adsorption isotherms are presented in Fig. 7. As it can be noticed for the silicas functionalized by APTS, the adsorbed amount obtained for the OMO samples M1 and M2 is about 1.7 times higher than the corresponding values for the APX samples X1 and X2. However, the amounts of nitrogen in the X1 and X2 samples are significantly higher than for M1 and M2, respectively, which suggests that the adsorption of Pt(II) ions in the amorphous APTS-functionalized samples is hindered due to limited adsorbate diffusion into the micropores.

For the pure silica samples M0 and X0 no Pt(II) uptake was observed testifying the fact that the composition of the surface layer is the factor influencing the adsorption properties of the studied systems. The measured values of static sorption capacities (SSC) are 32, 20, 65 and 36 mg_Pt(II)/g for M1, X1, M2 and X2, respectively. For the TMPD-functionalized silicas the opposite trend is observed, i.e., higher values of SSC are obtained for the APX than for OMO samples. The measured values of static sorption capacities (SSC) are 50, 87, 102 and 139 mg_Pt(II)/g for M3, X3, M4 and X4, respectively. Moreover, in the case of the TMPD-functionalized silicas, the adsorption capacities are correlated with the nitrogen content in these sorbents. High SSC values and relatively fast kinetics observed for X3 and X4 suggest that due to the presence of numerous mesopores in the structure of these samples the transport of Pt(II) ions to the adsorption centres localized in micropores is easier than in the case of the samples X1 and X2. Moreover, a fraction of amine groups is probably inaccessible for Pt(II) ions in the case of the latter one.

The isotherms presented in Fig. 7 exhibit the Langmuirian behaviour and may be described by the Langmuir–Freundlich equation given as:where, a is the amount of adsorbed platinum at the adsorption equilibrium state [mg/g], a _m is the maximum adsorption capacity [mg/g], K is the Langmuir constant, c is the equilibrium concentration of Pt(II) [mg/L]. The parameter n is the factor characterizing the quasi-Guassian energetic heterogeneity of the adsorption system. As the value of this parameter decreases, the energetic heterogeneity of the adsorbent surface increases. In the case when n → 0 there is high heterogeneity. For n → 1, equation (1) simplifies to the classical Langmuir adsorption isotherm equation describing the homogeneous adsorbents. On the other hand, in the case of very small equilibrium concentrations, the factor (Kc)ⁿ in equation (1) is much lower than 1, so that equation (1) reduces to the form:which is known as the Freundlich isotherm equation, very often used for explanation of adsorption data because of its simplicity (Jaroniec 1983). The linear form of the Freundlich equation can be expressed as:where the Langmuir constant K is related to the Freundlich constant C ₀ by the relation:which, when inserted into equation (3), leads to the following equation:

Equation 5 was used to calculate the values of K, C _o , a _m , n, and the correlation coefficients R ² which are shown in Table 2. The isotherms presented in Fig. 7 show that the adsorption of Pt(II) ions satisfies well Eq. 5 (i.e. the Freundlich adsorption isotherm). Comparing the values of parameter n some additional relations can be found. For the ordered silica the decrease in the n values along with the nitrogen content increase is observed. It means that energetic heterogeneity of the functionalized ordered silica increases along with the increase of the nitrogen content. For the disordered functionalized silica, the parameter n shows an opposite tendency, along with the increase of the nitrogen content the increase of the n values can be noted. Energetic heterogeneity decreases along with the increase of the nitrogen content, and for X3 and X4 these materials become energetically homogeneous for the studied adsorption system.

In Table 3 the static sorption capacities of the samples studied here are compared with other studies for Pt(II) sorption found in the literature (Vassileva et al. 2008; Cox et al. 2005). The materials studied here have a superior properties and the values of static sorption capacity for Pt(II) ions are the highest when compared with those presented in the former works (Yin et al. 2010; Vassileva et al. 2008; Cox et al. 2005).

In order to explain the mechanism of Pt(II) ions adsorption onto the amine-functionalized silicas and their chemical behavior on their surfaces, the XPS studies were undertaken. In Fig. 8a the XPS spectrum of the Pt 4f region for the Pt loaded TMPD-functionalized OMO, being a cut of the full spectrum is shown. For M4 platinum species are characterized by two peaks in the line with the Pt 4f7/2 and Pt 4f5/2 transitions. This doublet with the peaks (4f7/2 and 4f5/2) at 73.2 and 76.5 eV is assigned to Pt(II) (Drelinkiewicz et al. 2009). Therefore, the platinum in the second oxidation state is the dominant form of platinum species present on the TMPD-functionalized surface. Additionally, it was stated that for unloaded M4 the XPS spectra of N1s region are characterized by two peaks at 399.1 and 401.4 eV, which corresponds to –NH– and –NH₂ groups (Li et al. 2009). On the other hand, for Pt loaded M4 (Fig. 8b) triplet of N1s is observed with the peaks at 400.1, 401.3 and 402.9 eV. The appearance of the third peak with a slight intensity at 402.9 may be explained by oxidation mechanism of –NH– groups to –N⁺=O. It worth to emphasizing that partial transformation of Pt(II) to Pt(IV) takes place in the acidic liquid medium (Nischwitz et al. 2003). It means that part of Pt(II) in a bulk solution could be oxidized to Pt(IV) during the adsorption process (24 h). Simultaneously, Pt(IV) ions were reduced to Pt(II) onto the TMPD-functionalized OMO surface. In the case of Pt loaded APTS-functionalized OMO, the two doublets with the peaks (4f7/2) at 72.5 and 71.4 and (4f5/2) at 75.8 and 74.7 eV, respectively, are observed on the XPS spectra of the Pt 4f region (Fig. 8c). The first pair of peaks corresponds to the Pt(II) presence, whereas the second doublet proves the presence of metallic Pt. Therefore it is probable that in the case of M2 the partial reduction of Pt(II) to Pt(0) takes place. In the case of Pt loaded M2 triplet for N1s region with the peaks at 399.1, 400.3 and 402.2 eV is also observed (Fig. 8d). The first two peaks correspond to –NH₂ group in the presence of [PtCl₄]²⁻, forming surface complex. Whereas the peak at 402.2 eV corresponds to the unprotonated –N=O group, coming from oxidation of amine groups. The mechanism of Pt(II) sorption proposed above should be confirmed by another independent method (e.g. Mössbauer spectroscopy) to exclude the influence of X-ray beam on the platinum species during the XPS measurements.

High adsorption capacities of the studied organosilicas create opportunity for their use in the platinum preconcentration and separation from the environmental samples. In order to find how many sorption/desorption cycles can be performed with the prepared sorbents desorption studies were carried out. As the desorption medium HNO₃, HCl and thiourea in 1.0 mol/L HCl solutions with respect to their concentrations were proposed.

The application of HNO₃ as the desorption medium was chosen due to the fact that the nitrate does not form a stable complex with Pt(II) and Pt(IV) ions, so its desorption activity is connected only with the competition effect to adsorption centres. In the case of HCl beyond the competition effect, chlorides can replace OH– groups in the coordination zone of platinum ions. Whereas thiourea forms a very stable complex with platinum ions so it can compete with –NH₂ surface groups.

The desorption study has been presented in Fig. 9. The application of HNO₃ causes a partial desorption of platinum species from OMO, whereas for APX especially for X3 and X4 desorption was negligible. In the case of HCl, the application of 2 mol/L solution cause the total desorption of platinum species from M1 and M2. While for APX the desorption was lower and does not exceed 10 %. When thiourea in 1.0 M HCl was applied as the desorption medium, the total desorption of platinum species from M1, M2, M3, M4 and X2 materials was reached. Only for X1, X3 and X4 the platinum species desorption was not quantitative. To summarize, the application of OMOs modified by APTS and TMPD for the Pt(II) ions preconcentration and desorption by thiourea in 1 mol/L HCl creates the possibility of multiple use of sorbents.