Akhtar
Alam
,
Atikur
Hassan
,
Ranajit
Bera
and
Neeladri
Das
*
Department of Chemistry, Indian Institute of Technology Patna, Patna 801106, Bihar, India. E-mail: neeladri@iitp.ac.in; neeladri2002@yahoo.co.in; Tel: +91 9631624708
First published on 28th October 2020
A unique set of polymers were conveniently obtained by marriage of ‘cubic’ octavinylsilsesquioxane (OVS) and ‘paddle wheel’ shaped triptycene using the simple and economical Friedel–Crafts reaction. The resulting ‘hybrid inorganic–organic polymers’ (STNPs) are nanoporous with higher surface areas than several previously reported porous polymeric networks containing either one of the two structural motifs (OVS or triptycene). In addition, the STNP with the highest surface area (STNP3) also has better gas storage and dye capture abilities than several organic adsorbents derived from silsesquioxanes. The performance of STNP3 as a dye adsorbent suggests that it can be further explored for environmental remediation.
Silsesquioxanes are three dimensional oligomers and are generally represented as (RSiO1.5)2n. Depending on the R group (such as phenyl, vinyl, propyl – to name a few) these exhibit unique and interesting properties.14 Polyhedral oligomeric silsesquioxanes (POSSs) having the general formula (RSiO1.5)8 are a subclass of silsesquioxanes.15 In contemporary research, POSS moieties have received substantial research attention. These have been incorporated in various polymeric networks to yield hybrid inorganic–organic materials that are thermally stable, resistant to chemicals, non-toxic and biocompatible in nature.16–18 If the R group is vinyl, the resulting POSS is termed as octavinylsilsesquioxane (OVS) and is represented by the formula (C2H3)8Si8O12. OVS is commercially available and has been reported as one monomer/cross-linking agent in the design of several new hybrid (inorganic/organic) materials for versatile applications. Polymeric networks derived from OVS monomers are porous materials that have potential application in selective capture and storage of CO2 as well as adsorbents for toxic dyes present in industrial effluents.19–22 OVS derived porous materials may also be promising materials for the design of light-harvesting applications,23 catalysis, photo-voltaic or electronic devices,24,25 and sensors for explosive detection.26–28 Such silsesquioxane based hybrid porous materials can be conveniently synthesized using various synthetic strategies that include (but are not limited to) the Friedel–Crafts reaction,26,29 hydrosilylation,30 and coupling reactions such as the Heck reaction,17,31 Sonogashira32 and Yamamoto.33 While most coupling reactions employ expensive transition metal catalysts and rigorous reaction conditions, the use of Friedel–Crafts reaction conditions makes the polymerization facile, economical as well as versatile. Friedel–Crafts reaction conditions were also applied for the facile and ‘large-scale’ production of porous materials.34
Triptycene is an organic molecule with a bicyclic ring in its core. It has three peripheral phenyl rings that are suitably oriented to provide multiple crosslinking sites when used as one of the monomers in the Friedel–Crafts reaction.35–37 Network polymers derived from triptycene are porous materials.38 The presence of micropores and the associated high surface areas are attributed to the three-dimensional and contorted structure of triptycene motifs present in the polymeric backbone.38–40 Triptycene-based microporous organic polymers are also reported to demonstrate interesting applications in materials chemistry, such as small gas molecules (CO2, H2, CH4 and N2) capture and separation (CO2/CH4, CO2/N2 and CH4/N2), toxic dye removal and iodine adsorption for environmental remediation.46–51
Thus it is expected that the material incorporating triptycene and octa-functional OVS would yield a highly porous network. It is desirable that such materials should also have the potential to selectively capture carbon dioxide and toxic organic dyes for environmental remediation. These are in line with the objective to develop/obtain novel hyper-crosslinked polymers (with improved properties) synthesized from inexpensive monomers. It is also expected that such polymers have desirable properties such as high porosity and surface area, and high stability. Furthermore, it was anticipated that the ratio of the two monomers (i.e., triptycene and OVS) would be crucial in the fine tuning of the porosity of the resultant silsesquioxane based hybrid porous materials. With these premises in mind, the objective was to obtain a set of unique triptycene based porous polymers bearing OVS moieties using the Friedel–Crafts reaction.
For the first time, herein we report the marriage of triptycene and polyhedral silsesquioxanes (in the form of OVS) to yield a set of four unique silsesquioxane-based and triptycene-linked hybrid nanoporous polymers (STNPs). By varying the molar ratio of triptycene and OVS, STNP1, STNP2, STNP3, and STNP4 were obtained with multimodal porosity. Thus using this strategy, it was easily possible to tune the porous properties and surface area of the resulting polymeric network. Subsequently, the STNPs’ ability to capture small gases was studied. Finally, the STNPs were tested as adsorbents for common organic dyes that are often used in the textile industry and are present in industrial effluents as toxic pollutants above restricted levels.
STNPs (1–4) thus obtained were structurally characterized using FTIR and 13C CP-MAS NMR spectroscopies. In the respective FTIR spectrum, the incorporation of OVS units in STNPs was evident from the presence of a strong peak at 1103 cm−1 which is observed due to Si–O–Si stretching vibrations (Fig. 1). Additionally, the signal at 780 cm−1 is due to Si–C bond vibration.52 These signals are signature peaks observed in the FTIR spectrum of materials having silsesquioxane units.53,54 A comparison of FTIR spectra of triptycene and OVS with those of STNPs shows that the intensities of peaks between 1292 cm−1 and 1608 cm−1 (prominently seen in monomers) had decreased considerably in the STNPs. This suggested the successful Friedel–Crafts alkylation reaction between vinyl groups (in OVS) and phenyl rings (in triptycene motifs). Consequently, a decrease in intensity and sharpness of C–H stretching (in triptycene) at around 3000 cm−1 was also observed in the resulting STNPs with the simultaneous increase in methylene signal at around 2923 cm−1. Thus FTIR spectral analysis suggested the successful cross-linking reaction between triptycene and OVS as depicted in Scheme 1.
The formation of the desired STNPs was also confirmed from the analysis of their corresponding 13C CP-MAS NMR spectrum. The 13C solid-state CP-MAS spectrum of STNP3 is shown in Fig. 2 as a representative example. The signal appearing at around 52 ppm was due to the bridgehead carbons present in the triptycene unit and this has been observed in NMR spectra of polymers having triptycene motifs.55–57 In addition, signals due to the carbons present in the arene rings were observed in between 127 and 147 ppm. The resonance peaks around 27 and 17 ppm are assigned to the ethylene carbons connecting the Si atom on one side and a triptycene-based arene ring on the other side. Thus, solid 13C CP-MAS NMR (Fig. 2) data also confirmed the successful crosslinking between triptycene and OVS (Scheme 1).
The morphologies of STNPs were characterized by field emission scanning electron microscopy (FESEM). Fig. 3 shows the FESEM image of STNP3 as a representative example and images of the other three STNPs are included in the ESI† (Fig. S1). The morphologies observed for all STNPs have similar characteristics such as non-uniform aggregates of various sizes. Such a wide distribution in size was also observed in the previously reported hybrid polymers incorporating OVS units. The presence of the Si element due to the inclusion of OVS units in STNP3 was further confirmed from the EDX (Energy Dispersive X-ray) spectral data (Fig. 3).
The PXRD data of the STNPs were also recorded and are shown in Fig. 4. While it is well known that OVS is crystalline, the polymers (STNPs) resulting due to crosslinking of triptycene and OVS are amorphous in nature. This shows that polymerization results from the loss of the crystalline nature of POSS. The observed amorphous nature of STNPs is due to the randomness in the distribution of OVS units and the presence of triptycene units in the matrix of STNPs. The presence of Si–O–Si linkages due to the presence of OVS motifs in STNPs was clearly indicated from the presence of the broad peak centered at 2θ equal to 22°.58,59
Fig. 5 N2 adsorption–desorption isotherms at 77 K (left) and pore size distribution (right) of STNPs. |
The Brunauer–Emmett–Teller (BET) model is most appropriate to estimate the surface area of a porous material if it shows a type IV isotherm. Thus the BET model was applied to estimate the surface areas (SABET) associated with STNPs and these were found to be 1256 m2 g−1 (STNP1), 1421 m2 g−1 (STNP2), 1462 m2 g−1 (STNP3) and 1271 m2 g−1 (STNP4). The corresponding Langmuir surface areas for STNP1, STNP2, STNP3 and STNP4 are 1956 m2 g−1, 2266 m2 g−1, 2300 m2 g−1 and 1941 m2 g−1, respectively (Table 1 and Fig. S2, S3, ESI†). Among the four polymeric networks reported herein, STNP3 shows the highest surface area. The magnitude of surface area is better than that reported for most of the other silsesquioxane-based porous hybrid materials reported to date, such as the silsesquioxane-based thiophene-bridged network (915 m2 g−1 for THPP),17 the ferrocene-functionalized silsesquioxane-based polymer (1015 m2 g−1 for Fc-HPP),62 luminescent porous organosilicon polymers (1003 m2 g−1 for LPOP-2),26 hybrid polymers constructed from octavinylsilsesquioxane and benzene (904 m2 g−1 for HPP-3),29 octavinylsilsesquioxane-based luminescent hybrid polymers (685 m2 g−1 for PS-3)23 and polyhedral oligomeric silsesquioxane-based polymers (778 m2 g−1 for HPP-2).63 However, the SABET of STNP3 is lower than silsesquioxane-based tetraphenylethene-linked polymers (1910 m2 g−1) which to our knowledge is the highest value observed for a silsesquioxane-based POP.20 The surface area of STNP3 is also better than various triptycene based hyper-cross-linked polymers reported in the literature such as nanoporous organic polymers (1246 m2 g−1 for NOP-47),64 porous organic copolymers with triptycene and crown ether (848 m2 g−1 for POP-TCE-15),48 triptycene based hyper-cross-linked polymer sponge (1426 m2 g−1 for THPS),35 and triptycene based microporous polymers (1372 m2 g−1 for TMP3).36
Polymers | Molar ratio OVS:Trip | SABET (m2 g−1) | SALanga (m2 g−1) | V total b [cm3 g−1] |
---|---|---|---|---|
Surface area of STNPs calculated based on the BET model and the Langmuir modela from the N2 adsorption isotherms (P/P0 = 0.05–0.35). The total pore volumeb of STNPs calculated at P/P0 = 0.99. | ||||
STNP1 | 1.0:0.53 | 1256 | 1956 | 1.377 |
STNP2 | 1.0:0.86 | 1421 | 2266 | 1.167 |
STNP3 | 1.0:1.0 | 1462 | 2300 | 1.096 |
STNP4 | 1.0:1.5 | 1271 | 1941 | 0.855 |
Next, the pore size distribution (PSD) plots for STNPs were obtained from the N2 sorption isotherms by using the density functional theory (DFT) method. These curves (Fig. 5) indicated multimodal distribution with peaks centered in both microporous (around 1.5 nm) and mesoporous regions (between 2.5–6 nm). The pore-size distribution is relatively thinner in the microporous region (below 2 nm) than that in the mesoporous region. The difference in the shape of the PSD plots for STNPs implies that the molar ratio of triptycene and OVS is an important parameter that governs the pore structure of the resultant silsesquioxane-based and triptycene-linked hybrid polymers. Considering the presence of both micropores and mesopores, STNPs can be classified as nanoporous materials. The porosity data of STNPs are compiled in Table 1 for comparison. The total pore volume of each STNP was calculated from the volume of N2 adsorbed at P/P0 = 0.99 and this parameter was measured to be 1.377 cc g−1 for STNP1, 1.167 cc g−1 for STNP2, 1.096 cc g−1 for STNP3 and 0.855 cc g−1 for STNP4.
The data presented in Table 1 indicate that the magnitude of various porosity parameters of STNPs can be easily altered by changing the molar ratio of OVS and triptycene. Interestingly, the highest surface area (SABET = 1462 m2 g−1) was observed in the product (STNP3) that utilized equimolar quantities of OVS and triptycene in the polymerization reaction. A lower surface area was recorded when the OVS:triptycene molar ratio was either less or more than unity. These observations may be justified by considering that each triptycene molecule has three phenyl rings that connect multiple OVS to form STNPs. At a lower concentration of triptycene, it might be possible that all the vinyl groups present in OVS units are not linked to the phenyl rings of triptycene leading to a relatively lower BET surface area. As the concentration of triptycene units increases (STNP1 < STNP2 < STNP3 < STNP4), the vinyl rings (of OVS) covalently crosslink with the phenyl rings of triptycene to a greater extent. This results in higher surface area of the resulting polymeric network due to greater crosslinking. Thus the surface area increases as the OVS:triptycene ratio decreases from 1.89 (STNP1) to 1.0 (STNP3). Further decrease in the OVS:triptycene ratio to 0.67 (STNP4) results in a decrease in SABET. This implies that at such higher concentrations of triptycene, all phenyl rings of triptycene might not act as connecting sites and hence local crosslinking densities will not keep increasing. This effect may be also related to steric hindrance preventing further increase in SABET in the case of excess triptycene units used in STNP4. Hence, relative to STNP3, a further increment in the surface area is not observed in STNP4. The overall decrease in the total pore volume with a decrease in the OVS:triptycene ratio might be due to more interpenetration of the framework since all triptycene phenyl rings might not participate in crosslinking leading to the pore-filling of the network to some extent. To sum up, the flexible nature of Si–CH2–CH2–Ar and the rigid bulky structure of triptycene contribute to the observed trend in the porous properties of STNPs (Table 1).
Fig. 6 H2 and (left) and CO2 (right) sorption isotherms of STNP3 (filled triangle/star = adsorption and empty triangle/star = desorption) |
Next, the ability of the triptycene-linked hybrid and porous STNP3 to selectively capture CO2 was tested. Considering the harmful effects of global warming due to CO2 gas emitted from anthropogenic sources such as thermal power plants, assessing CO2 storage properties of porous materials has assumed industrial importance. A major technological challenge is to develop efficient thermally stable adsorbents to reduce atmospheric CO2 emissions from coal powered plants. Several recently published articles have highlighted the importance of hybrid porous materials as an adsorbent for CO2.20 With this background, it was our interest to evaluate STNPs as materials for CO2 uptake and compare their performance with other OVS derived polymeric networks. Thus, CO2 sorption isotherms were recorded at two temperatures (273 K and 298 K) and at pressures up to 1 bar (Fig. 6) for STNP3. The gravimetric uptake of CO2 for STNP3 was 86 mg g−1 (8.60 wt%) at 273 K and 50 mg g−1 (5.0 wt%) at 298 K. Compared to some previously reported silsesquioxane based hybrid porous materials, the performance of STNP3 for CO2 uptake had improved. Representative examples of polymers with a lower CO2 uptake include (but are not limited to) silsesquioxane-based tetraphenylethene-linked polymers (HPP-3, 6.25 wt% at 273 K),20 silsesquioxane-based triphenylamine functionalized polymer, (HLPP-OTS, 8.04 wt% at 273 K),19 ferrocene-functionalized silsesquioxane-based polymer (Fc-HPP, 4.87 wt% at 273 K),62 hybrid polymers derived from octavinylsilsesquioxane and polystyrene (HCP-3, 4.93 wt% at 298 K),66 hybrid polymers constructed from octavinylsilsesquioxane and benzene (HPP-3, 2.73 wt% at 298 K),29 hybrid porous polymers derived from octavinylsilsesquioxane and tetraphenylsilane (HPP-5, 3.31 wt% at 298 K).65 We attribute such marked improvement in our OVS based hybrid porous polymer (STNP3) to the concurrent incorporation of rigid and three-dimensional contorted triptycene units having internal free volume along with OVS.
The Qst (isosteric heats of adsorption) value of STNP3 for CO2 uptake was directly calculated from the experimental adsorption isotherms collected at 273 and 298 K using the Clausius–Clapeyron relation. The Qst values for CO2 were observed in the range 22.0 kJ mol−1 at zero coverage (Fig. S4, ESI†). Magnitude of Qst provides valuable insight regarding the nature and extent of interaction (strength) between the adsorbent and the adsorbate. The Qst value recorded for CO2 capture by STNP3 suggests uptake via a physisorption process because the magnitude (of Qst) is less than 40 kJ mol−1.68 This Qst value of CO2 capture by STNP3 is less than that reported for a previous literature reported porous polymer (HPP-1c) obtained using OVS and 2,2′,7,7′-tetrabromo-9,9′-spirobifluorene.25 This suggests that STNP3 has better potential for use as a porous adsorbent with ‘reversible’ CO2 uptake. The ability of STNP3 to selectively capture CO2 over N2 was also examined. The CO2/N2 selectivity was calculated in the pressure range of 0–0.1 bar, using the Henry law and on the basis of initial slope calculations (Fig. S5, ESI†). STNP3 showed CO2/N2 selectivity of 20 at 273 K and 11 at 298 K.
(1) |
(2) |
The dye adsorption data (Fig. 7) show that Qe for STNP3 increases significantly as the initial dye concentration is increased. However, with further increase in dye concentrations, adsorption saturation is reached for STNP3. The maximum value of Qe obtained was 198 mg g−1 for CR, 1261 mg g−1 for CV, 135 mg g−1 for MB, 106 mg g−1 for MO and 752 mg g−1 for RB. In order to describe the mechanism of dye adsorption by STNP3, the Langmuir and Freundlich adsorption models were applied. The Langmuir isotherm model is described by eqn (3) and assumes energetically identical and non-interacting adsorption sites as well as monolayer adsorbate coverage on the surface. The Freundlich isotherm model is defined by eqn (4), in which case, the adsorption is non-ideal since adsorption sites are non-identical with varied affinities for adsorbate species. Unlike the Langmuir model, the Freundlich model assumes non-homogenized and multilayer adsorption.77,78
Fig. 7 Equilibrium adsorption isotherms (left) and linear fitting of the equilibrium data by using the Langmuir Equation (right) of dyes on STNP3. |
The equations corresponding to these two classical adsorption models were employed to fit the equilibrium data of dye adsorption by STNP3 (Fig. 7 and Fig. S6, ESI†). The respective magnitudes of adsorption constants and the correlation coefficients (R2), calculated for both isotherm models are presented in Table 2. The correlation coefficients of the Langmuir isotherm model (RL2, 0.999 for CR, 0.997 for CV, 0.994 for MB, 0.999 for MO and 0.999 for RB) are higher than those calculated for the Freundlich isotherm model (RF2, 0.918 for CR, 0.981 for CV, 0.957 for MB, 0.975 for MO and 0.938 for RB). This implies that the Langmuir isotherm model has a better fit with the experimental data and adsorption of dyes on the surface of STNP3 is monolayer adsorption. It also suggests that the dye adsorption sites in STNP3 are more or less uniform and identical:
(3) |
(4) |
Dyes | Nature | Langmuir constants | Freundlich constants | ||||
---|---|---|---|---|---|---|---|
Q m (mg g−1) | K L (L mg−1) | R L 2 | K F (L mg−1) | n | R F 2 | ||
CR | Anion | 250 | 0.182 | 0.999 | 69.40 | 4.06 | 0.918 |
CV | Cationic | 1428 | 0.087 | 0.997 | 368.70 | 4.54 | 0.981 |
MB | Cationic | 166 | 0.009 | 0.994 | 17.28 | 3.51 | 0.957 |
MO | Anion | 111 | 0.125 | 0.999 | 63.43 | 12.82 | 0.975 |
RB | Cationic | 1000 | 0.20 | 0.999 | 372.41 | 8.77 | 0.938 |
From the Langmuir plot, the maximum adsorption capacity (Qm) can be computed and was found to be 1428 mg g−1 for CV, 1000 mg g−1 for RB, 250 mg g−1 for CR, 166 mg g−1 for MB and 111 mg g−1 for MO. These calculated values of Qm are slightly higher than the corresponding experimental Qe values. The values of Qm obtained for CV and RB are significantly higher than those for the other dyes used in this study. Furthermore, from these Qm data, STNP3 is a better adsorbent for CV and RB than several other porous polymeric networks reported previously. The comparison of dye adsorption data of STNP3 with those of other materials is provided in Table S2 (ESI†).
From the results of these experiments, it follows that the adsorption capacity of STNP3 for various dyes decreases in the order CV > RB > CR > MB > MO. These dye molecules have different molecular sizes.35,75,76 Thus, it can be concluded that STNP3 shows size-selective dye adsorption. The PSD profile of STNP3 shows that this polymer has micropores as well as mesopores. As far as the charge associated with these dyes is concerned, CV, RB and MB are cationic, while CR and MO are anionic. Among the five dyes used in this study, CR has the largest molecular size while MO has the smallest. RB and CV have similar size and have a diameter greater than 1.5 nm.62 On the other hand, MB and MO are relatively smaller in size and their diameters are less than 1.5 nm.
In the present study, STNP3 has much higher affinities for the two larger sized cationic dyes (CV and RB) relative to the smaller cationic dye (MB). As far as the anionic dyes (CR and MO) are concerned, STNP3 has a rather poor affinity for the anionic CR molecules even though CR has the biggest molecular size. From their structural composition, POSS cores are rich in electronegative oxygen atoms and thus they have a natural tendency to capture cationic dyes such as CV and RB via favorable electrostatic attractive interactions. The observed lower adsorption capacity in the case of anionic dyes is due to repulsive ion–dipole interaction between these dyes (CR and MO) and the electronegative inorganic silsesquioxane cores present in STNP3. Additionally, the correlation between the pore structure of STNP3 and the size of the dyes also supports the observed absorption capacity (CV > RB > CR > MB > MO). The larger dye molecules (CR, RB and CV) either clog the small micropores upon binding or interact well with the mesopores. The relatively smaller dyes (MB and MO) enter and exit the larger sized micro- and mesopores easily due to their smaller size (diameter < 1.5 nm) and hence these are adsorbed to a much smaller extent even though MB is cationic. Overall, the trend in absorption capacity of various dyes depends on the simultaneous interplay of several factors such as the porous structure and surface area of the adsorbent, the ionic nature of the dye (adsorbate) and groups present in the polymeric framework of the adsorbent.
Considering the high Qm value of RB and CV, it was assumed that STNP3 may be a potential adsorbent for these two dyes for practical applications. Therefore, it was our interest to further explore their adsorption kinetics using STNP3 as an adsorbent. While using RB as an adsorbate, the experiment was done by adding 10 mg STNP3 into 10 mL aqueous RB solution with an initial concentration of 40 ppm. The CV adsorption kinetics was checked by adding 10 mg STNP3 into 10 mL aqueous CV solution with an initial concentration of 100 ppm. Dye absorption kinetics was monitored by recording the visible spectrum at different time intervals. The adsorption spectra were recorded at different time intervals and are shown in Fig. 8. The gradual adsorption of dyes with time could be also visualized by the naked eyes (inset of Fig. 8). It was observed that with the onset of the experiment, the adsorption rate is fast and the uptake rate slows down with the progress of time. For STNP3, within 10 min, the removal efficiency reached 70% for RB and 90% for CV. The time taken for nearly complete removal of RB is around 30 min and that for CV is even less, at around 20 min.
STNP1, STNP2, and STNP4 were synthesized similarly, with OVS to triptycene molar ratios of 1.0:0.53 (OVS 1.0 mmol), 1.0:0.86 (OVS 1.0 mmol), and 1.0:1.5 (OVS 1.0 mmol), respectively. The final products were collected as a yellow powder with yields of 86% for STNP1, 85% for STNP2 and 89% for STNP4.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/d0ma00672f |
This journal is © The Royal Society of Chemistry 2020 |