Thermogravimetric studies of benzoylthiourea-modified MCM-41 after adsorption of mercury ions from aqueous solutions

Abstract

High-resolution thermogravimetry (HR TG) was used to study the adsorption of mercury(II) ions by modified MCM-41 material and regeneration of the loaded adsorbent with mercury ions by using different eluents. The weight change curves were measured for MCM-41 samples modified with 1-benzoyl-3-propylthiourea ligand loaded with mercury ions. The differential thermogravimetric (DTG) curves were analyzed to investigate the adsorption of mercury ions by the aforementioned multifunctional ligand and to monitor the decomposition of the metal–ligand complexes. A series of experiments performed for different Hg²⁺ : ligand ratios allowed us to correlate the adsorption data for mercury ions measured by means of UV spectrophotometry with those obtained by HR TG analysis. The DTG results provided additional information about mercury-ligand interactions as well as the thermal stability of mercury–ligand complexes. This study shows that HR TG is a very attractive technique for studying the adsorption of mercury ions on modified nanoporous silicas and monitoring their regeneration. Since the samples used are small, this method seems to be promising for studying adsorption systems of environmental significance.

---

1 Introduction

Thermal analysis has been extensively used by materials chemists for the characterization of various materials. It was especially shown to be a very effective technique for studying the thermal stability, surface coverage, and modification process of alkyl-modified mesoporous silicas, such as MCM-41, a member of ordered mesoporous materials discovered in the early 1990s.^1,2 High resolution thermogravimetry (HR TG) has been extensively used to investigate adsorption properties such as specific surface area, pore volume and surface heterogeneity of various nanoporous materials.^3,4 The adsorption and surface properties of nanoporous materials decide their prospective applications as adsorbents, catalysts and separation media. This work demonstrates that HR TG is an extremely attractive technique to study the adsorption affinity of organically modified MCM-41 material towards mercury(II) ions, and consequently shows the great promise of this technique for studying adsorption systems of environmental significance.

Thermogravimetric analysis, the principle of which is to monitor the change in sample weight (loss or gain) as a function of temperature and/or time,⁵ provides information not only about the thermal stability and composition of the initial sample but also about the thermal stability and composition of any formed intermediate. This method was successfully used to evaluate the mesopore volume and specific surface area of various adsorbents which were in a good agreement with those evaluated from low-temperature nitrogen adsorption, which is used as the standard technique for characterization of the adsorption properties of nanoporous materials.⁶ It was also reported that the highly precise monitoring of the weight change under controlled conditions appeared to be a useful method for environmental-type studies which involved the calibration of permeation tubes that were affected by a trace concentration of volatile organic compounds.⁷ Recently, a great interest in environmental chemistry has been focused on the design of novel adsorbents by chemical modification of ordered mesoporous silicas with specific organic and organometallic ligands of high selectivity toward specific ions. The most important characteristics of those adsorbents are their very high adsorption capacity, high ligand coverage, suitable thermal stability, tailored pore size and pore structure, as well as their ease of regeneration without substantial loss of adsorption capacity. As regards their thermal stability, it is desired that the attached ligands should not decompose in the temperature range in which these adsorbents are planned to be used.

Thermogravimetric methods are widely used to monitor almost every step of the preparation of adsorbents, including the study of their thermal stability, surface coverage of attached ligands and adsorption properties such as specific surface area and pore volume. More often porous materials are subjected to adsorption of gases and vapors to evaluate their interactions with probes of different polarity. In this work we employed HR TG analysis under controlled conditions to study the mercury adsorption by 1-benzoyl-3-propylthiourea modified MCM-41 silica and to monitor the regeneration of this material after mercury loading by using different eluents, namely concentrated HCl, solution of 10% thiourea in 0.05 M HCl, 10% cystein in 0.05 M HCl, and 10% thiourea in 0.05 M AcOH. The thermal events observed in the systems studied are in a good correlation with the proposed scheme of coordination of mercury ions by attached multifunctional ligand. The amount of loaded mercury increased proportionally with increasing Hg²⁺ : ligand ratio, which was observed both on the TGA weight change curve and on the corresponding DTG curves. The HR TG method can also be employed to characterize adsorption and regeneration characteristics of the surface-modified porous adsorbents with other multifunctional ligands showing high affinity towards mercury and related heavy metal ions.

2 Experimental

2.1 Synthesis of ordered mesoporous silica

The MCM-41 material was synthesized according to a previously published procedure by using the mixture of alkyltrimethylammonium and alkyltriethylammonium bromide surfactants.^8,9 The as-synthesized material was washed with EtOH–HCl mixture and calcined as described elsewhere.⁹ The calcined sample was designed as MCM41-C.

2.2 Post-synthesis functionalization

The calcined material was subjected to modification with 3-aminopropylthriethoxysilane in toluene solution (refluxing for 24 hours) according to the previously reported procedure.² The multifunctional 1-benzoyl-3-propylthiourea ligand was attached onto the mesopore walls of MCM-41 in the second step of modification via a reaction between the aminopropyl functionality of the MCM-41NH₂ material and benzoyl isothiocyanate. The modified sample with grafted thiourea functionality inside the mesopores had a light yellow color and was designated as MCM-41BTU. The specific surface area, pore volume and pore size of MCM-41BTU were equal to 380 m² g⁻¹, 0.45 cm³ g⁻¹ and 3 nm, respectively. Other details related to the synthesis and characterization of this material can be found in our previous article.¹⁰

2.3 Mercury adsorption study

Mercury adsorption was studied from aqueous solutions. A series of experiments for the different Hg²⁺ : ligand ratios (0.2 ∶ 1, 0.5 ∶ 1, 1 ∶ 1, 3 ∶ 1, 4 ∶ 1, 5 ∶ 1) was performed using solutions of different Hg²⁺ concentrations which were prepared by dilution of a specified volume of the mercury(II) nitrate volumetric standard, 0.145 N aqueous solution, to a total volume of 10 ml. In a typical experiment 0.05 g of the sample (MCM-41BTU) was equilibrated for 40 min with 10 ml of aqueous mercury solution of known concentration. The filtrate was collected after adsorption and the sample was washed with 5 ml of DI water. The collected solution was diluted to 25 ml. All samples, loaded with mercury at different mercury–ligand ratios, were dried under vacuum at room temperature for 24 hours and subsequently subjected to thermal analysis under controlled conditions. Also, all the samples loaded with mercury were subjected to further regeneration and second mercury adsorption. Similar conditions for mercury adsorption experiments were repeated for the regenerated samples.

2.4 Determination of the adsorbent capacity

The mercury(II) concentration in the filtrate after adsorption was measured spectrophotometrically with dithizone (diphenylthiocarbazone) as a complexing agent.¹¹ Mercury photometric determinations were performed on a Shimadzu-1601 spectrometer in 1 cm quartz photocell (volume ca. 5 ml). The amount of mercury was determined from the calibration curve that was prepared for the mercury concentration interval of 0–50 µg with desirable correlation to the complexing agent concentration.¹² The background correction was performed against pure chloroform. Each sample was analyzed at 490 nm.

2.5 Regeneration of the adsorbent

In the regeneration experiments, the samples loaded with mercury were treated with 10% thiourea solution in aqueous 0.05 M HCl, concentrated HCl, 10% cystein in 0.05 M HCl, and 10% thiourea in 0.05 M AcOH. Desorption was performed under static conditions by soaking 0.05 g of the metal-loaded adsorbent for 15 min in 10 ml of the eluent solution, followed by filtration and rinsing of the sample with an additional 10 ml of the regeneration agent and 10 ml of deionized water. The regenerated samples were finally dried overnight under vacuum at 65–70 °C and were subjected again to the mercury adsorption.

2.6 Thermogravimetric measurements

The unmodified, modified, loaded with mercury and regenerated samples were analyzed by using high-resolution thermogravimetry. All measurements were carried out in a nitrogen atmosphere using a TA Instruments Inc. (New Castle, DE, USA) Model TGA 2950 high-resolution thermogravimetric analyzer. This instrument was equipped with an open platinum pan and an automatically programmed temperature controller. The weight change (TG) curves were recorded over a temperature range from 20 °C to 1000 °C. In order to maintain the temperature constant during a given thermal event the heating rate was regulated automatically by the instrument software. The maximum heating rate between thermal events was set to be 5 °C per minute. The resolution and sensitivity parameters were 4 and 6, respectively. The flow rate of nitrogen gas in the system was 100 cm³ and 50 cm³ per minute on the furnace and balance, respectively.

3 Results and discussion

The HR TG analysis of the unmodified and modified materials was done as described in the previous work.¹⁰ This analysis confirmed the two-step functionalization of the MCM-41 sample and demonstrated that the 1-benzoyl-3-propylthiourea functionalized MCM-41 is thermally stable up to ca. 170 °C. The modified MCM-41 sample showed two perfectly distinguishable thermal events on the DTG curve at ∼200 °C and ∼400 °C which corresponded to the decomposition of organic ligand and the residual aminopropyl groups. The qualitative estimation of the ligand concentration and the concentration of the residual NH₂ groups was estimated from high-resolution TG analysis under controlled conditions. The quantitative estimation of the ligand coverage was done by the CHNS analysis. A monomeric-type attachment of organosilane with unhydrolyzed RO-groups as well as monomeric- and/or polymeric-type attachments of organosilane with hydrolyzed RO-groups was taken into account in calculations as described previously.¹³ The surface concentrations of benzoylthiourea ligands and residual aminopropyl groups were equal to 1.50 and 0.65 mmoles per gram, respectively. The ligand concentration was evaluated with the error ±0.05 mmol g⁻¹.

Fig. 1 shows the TG curves recorded for the MCM-41BTU material after adsorption at different mercury concentrations by using Hg²⁺ : ligand ratios of 5∶1, 4∶1, 3∶1, 2∶1, 1∶1, 0.5∶1, 0.2∶1. A similar profile of the TG curves over the entire temperature range (from 20 to 1000 °C) was observed for all measured samples after adsorption of mercury. The organic ligand decomposition profile remained unchanged for the samples studied, which indicates that the organic moiety is not significantly affected by mercury ion adsorption. This confirms that no structure changes occurred during mercury adsorption and regeneration.

Fig. 1 The TG curves recorded for the MCM-41BTU material after mercury adsorption by using different Hg²⁺ : ligand ratios: 5∶1; 4∶1; 3∶1; 2∶1; 1∶1; 0.5∶1; 0.2∶1.

The DTG curves plotted in Fig. 2 for the mercury-loaded samples showed distinguishable changes in the thermal behavior in comparison with the modified sample without adsorbed mercury. The first derivative curve was used to specify the temperatures of the thermal events and to determine the characteristic temperatures at the beginning, maximum and the end of a given TG weight change. The derivative curves showed that the main thermal events related to the metal–ion adsorption occurred in the temperature range 180–350 °C. The intensities of the DTG peaks depend on the mercury : ligand ratio. There are two distinct thermal events that each sample underwent before ligand decomposition temperature (380–400 °C).

Fig. 2 The DTG curves recorded for the MCM-41BTU material after adsorption at different Hg²⁺ : ligand ratios equal to ∶1; 1∶1 and 0.5∶1.

The first event occurred at ∼200 °C and was observed much more intensively for the samples that were subjected to the mercury adsorption at the higher range of mercury concentrations (larger than 1 ∶ 1 Hg²⁺ : ligand ratio). The second event, which occurred at 260–290 °C on the DTG curve, was observed for all samples and its intensity remained similar for all samples studied.

Fig. 3 shows the difference between the thermal events in these two temperature ranges. As can be seen from this figure the decomposition mechanism of the samples studied is different in these ranges. For the samples that underwent mercury adsorption at the mercury–ligand ratios lower than 1 ∶ 1 a small peak is observed on the weight loss curve for the thermal event at 250–600 °C. The weight losses related to the mercury adsorbed on the samples are much higher in the first temperature range (100–250 °C) than in the second temperature range (250–600 °C). In the first range the weight loss related to mercury increases gradually from ∼1% to ∼17%. On the contrary, the samples that were subjected to the mercury adsorption at higher mercury concentrations showed an extremely high characteristic peak in the lower temperature range with more than 17% of the total weight loss. This peak becomes almost negligible on the DTG curves for the samples subjected to the mercury adsorption at low concentrations. The reason for this behavior is explained below.

Fig. 3 Comparison of the TG weight change for the MCM-41BTU material at different distinguishable temperature ranges after mercury adsorption. The weight change in the mercury loaded material is normalized to the unloaded MCM-41BTU. In order to perform this normalization the 100–600 °C range was used to make a proper correction for the amount of organic groups present.

As was reported previously,¹⁰ the mercury adsorption isotherm on the 1-benzoyl-3-propylthiourea modified MCM-41 material can be fitted by a two-term Langmuir–Freundlich equation, which suggested a two-step adsorption mechanism characterized by different adsorption constants. The stronger and weaker binding was reflected by a large difference in the adsorption constants, K₁ = 1.41·10⁵ and K₂ = 1.08·10² L mol⁻¹. The equilibrium adsorption isotherm reported in the previous work¹⁰ for mercury, obtained by spectrophotometrical analysis with dithizone (diphenylthiocarbazone) as a complexing agent, was used to characterize the adsorption capacity of the material and to correlate these adsorption data with those evaluated from HR TG analysis in the current study (see Fig. 4). As can be seen from this figure the TG weight loss for mercury adsorption correlates well with spectrophotometric measurements of the mercury adsorbed at low concentrations. Thus, in this concentration range the HR TG analysis can be successfully used for estimation of the amount of mercury adsorbed.

Fig. 4 Correlation between the TG weight change (in %) and the amount of adsorbed mercury determined spectrophotometrically. The weight change in the mercury loaded material is normalized to the unloaded MCM-41BTU.

It was shown elsewhere¹⁰ that the MCM-41BTU material has high adsorption capacity due to the multifunctionality of the attached ligand and fast diffusion of the targeted ion inside the mesopores. The successful choice of the ligand not only afforded material of extremely high adsorption capacity (as high as 5.0 mmol Hg²⁺ per gram of the adsorbent), but also introduced the multiple mercury–ligand interactions through stronger (sulfur) and weaker (nitrogen and oxygen) containing groups that were previously observed in liquid systems.¹⁴ The difference in the decomposition behavior of the adsorbent with 1-benzoyl-3-thiourea functionality without and with loaded mercury ions supports also the involvement of different adsorption sites (NH, CO and sulfur groups) in the mercury–ligand complex formation.

The multifunctional ligand used assured relatively weak mercury–ligand interactions that facilitated adsorbent regeneration. The regeneration of the mercury loaded adsorbent was monitored by high-resolution thermogravimetry (Fig. 5). Four different eluents were studied for regeneration of the adsorbent loaded with 0.65 g Hg²⁺ g⁻¹ material: 10% thiourea (TU) solution in 0.05 M HCl, 10% thiourea solution in 0.05 M AcOH, 10% cystein solution in 0.05 M HCl, and 12 M HCl. Fig. 5 shows that the TG profile for the mercury-loaded sample regenerated with concentrated hydrochloric acid is analogous to that obtained for the MCM-41BTU material, which indicates that almost all mercury ions were removed during sample regeneration. The most suitable candidate for the regeneration under milder conditions is thiourea in aqueous hydrochloric acid solution (0.05 M). The repeated adsorption data suggested that the sample regenerated by this eluent can adsorb almost the same amount of mercury in the lower range of the Hg²⁺ : ligand ratio (up to 1 ∶ 1). As regards the total adsorption capacity, the sample regeneration under the aforementioned mild conditions restored over 70% of the initial adsorption capacity.

Fig. 5 Comparison of the TG weight change curves recorded in nitrogen for the MCM-41BTU material regenerated at different conditions. The symbol RS refers to the sample regenerated by using specific regenerating agents.

4 Conclusions

This work demonstrated that high resolution thermogravimetry is a specific technique for characterization of novel mesoporous mercury-specific adsorbents and monitoring the process of their synthesis and post-synthesis modification. At low mercury concentrations the amount of loaded mercury on the chemically modified MCM-41 silica estimated by the TG analysis was shown to be linearly proportional to the amount of adsorbed mercury determined spectrophotometrically. The thermal events observed for the decomposition of the mercury–ligand complexes suggested a multifunctional binding mechanism of mercury ions which involves weaker and stronger binding sites of the 1-benzoyl-3-propylthiourea ligand. Also, HR TG was shown to be useful, quick and informative method to monitor the regeneration process of the adsorbent and to select the optimal eluent for regeneration under mild conditions. This technique seems to be especially valuable for preliminary analysis of numerous adsorbate–adsorbent systems because its duration is reatively short and it requiers a small amount of the sample.

Acknowledgements

The authors acknowledge the National Science Foundation grant CTS-0086512 for the support of this research and Dr. Ryong Ryoo, Department of Chemistry and School of Molecular Science (BK21), KAIST, Korea, for providing the MCM-41 material.

References

1. C. T. Kresge, M. E. Leonowicz, W. J. Roth, J. C. Vartuli and J. S. Beck, Nature, 1992, 359, 710.
2. V. Antoshchuk and M. Jaroniec, J. Phys. Chem. B, 1999, 103, 6252.
3. A. S. Araujo and M. Jaroniec, Thermochim. Acta, 2000, 345, 173.
4. M. Jaroniec, R. K. Gilpin, J. Ramler and J. Choma, Thermochim. Acta, 1996, 272, 65.
5. W. Wendlandt, Thermal Analysis, Wiley, New York, 3rd edn., 1986.
6. L. P. Mercuri, J. R. Matos and M. Jaroniec, Thermochim. Acta, 2002, 383, 79.
7. P.-C. Maria, J.-F. Gal, M. Balza, E. Pere-Trepart and S. Tumbiolo, Anal. Chem., 2002, 74, 305.
8. R. Ryoo, C. H. Ko and I.-S. Park, Chem. Commun., 1999, 1413.
9. M. Jaroniec, M. Kruk, H. J. Shin, R. Ryoo, Y. Sakamoto and O. Terasaki, Microporous Mesoporous Mater., 2001, 48, 127.
10. V. Antochshuk, O. Olkhovyk, M. Jaroniec, H. J. Shin and R. Ryoo, Langmuir, 2003, 19, 3031.
11. Comprehensive Analytical Chemistry, ed. G. Svehla, Elsevier, Amsterdam, 1985, vol. XX, ch. 6.
12. AOAC Official Methods of Analysis, ed. K. Helrich, The Association of Official Analytical Chemists, Arlington, 1990.
13. V. Antoshchuk and M. Jaroniec, Chem. Mater., 2000, 12, 2496.
14. A. C. Hiremath and A. S. R. Murthy, Ind. J. Chem., 1977, 15A, 55.

---