A photoluminescent covalent triazine framework: CO2 adsorption, light-driven hydrogen evolution and sensing of nitroaromatics

Department of Inorganic and Physical Ch Organometallics and Catalysis, Ghent Uni Belgium. E-mail: asamanjoy.bhunia@gmail Institut für Anorganische Chemie und Struk Düsseldorf, Germany. E-mail: janiak@uni-d Department of Organic Chemistry, Nanoc Institute (IUIQFN), Faculty of Sciences, Rabanales, Marie Curie Building, Ctra. Nal Department of Chemistry – Ångström Labo 751 20, Uppsala, Sweden Toyota Central R&D Laboratories, Inc., Na † Electronic supplementary information analysis, KFT, isosteric heat of adsorpti adsorption, IAST selectivity, emission nitroaromatic explosives. See DOI: 10.103 Cite this: J. Mater. Chem. A, 2016, 4, 13450


Introduction
Porous organic materials have attracted considerable interest because of their unique physicochemical properties in various technological and energy-related applications such as gas storage, separation, catalysis, etc. 1,2 In particular, p-conjugated porous organic polymers are an integral part of research for light emission, light harvesting and chemo-sensing. 3 Due to their rigidity at the molecular level, they have a high affinity to aggregate in solution and the solid state, resulting typically in aggregation caused quenching (ACQ) effects. 4 The aggregation of conjugated polymers can be prevented by molecular approaches based on site isolation with bulky polymeric matrices. 5 This approach can improve the luminescence properties of the polymeric materials. At the same time, p-conjugated materials are also a promising platform for CO 2 capture because of the interaction between the quadrupole moments of CO 2 and the p-clouds in the porous organic materials. 6 The incorporation of high amounts of nitrogen into the porous materials can facilitate CO 2 capture. Such multifunctionality has been a long term challenge for the development of porous photofunctional materials which have attracted great interest as chemical sensors, e.g. to detect explosive nitroaromatic compounds. 7 The sensitivity and selectivity in the detection of these compounds by using porous materials suffer from analyte-host interactions, chemical stability and so on. p-Conjugated triazine based porous materials may be attractive chemical sensors because of their high stability and high porosity. Due to the high electron mobilities and electron withdrawing character of the triazine ring, these materials have been widely used in synthetic chemistry and optoelectronics. 8 Currently, these types of materials have attracted great interest as organic photocatalysts for hydrogen production via photocatalytic water splitting under visible light. 9 Their chemical structures, analogous to that of g-C 3 N 4 , in combination with their tunable porosity make them attractive candidates for lightinduced hydrogen evolution. 10 Lostch et al. reported the rst example of a covalent organic framework (COF) which was active in visible light-induced hydrogen production from water using platinum as a cocatalyst. 9a More recently, a yellow coloured phenyl-triazine oligomer with high crystallinity exhibited the highest activity obtained for metal-free organic photocatalysts in the absence of noble metal catalysts. 9b Herein we report a new strategy for the construction of photoluminescent materials based on p-conjugated building blocks. We employed tetra(4-cyanophenyl)ethylene as a single component for the synthesis of a porous covalent triazinebased framework (PCTF) using triuoromethanesulfonic acid (TFMS) as the catalyst at room temperature, in which the tetraphenylethylene (TPE) units are directly connected to strong electron accepting triazine rings. Recently, some of us reported PCTFs with the TPE unit by using the ionothermal ZnCl 2 route. 11 The structures of CTFs from ionothermal reactions with ZnCl 2 are in-between well-dened COFs and porous carbon materials. 12 TFMS-catalyzed condensation frameworks 13,14 offer signicant advantages: (1) lower temperature and shorter reaction time, (2) avoiding undesired decomposition and condensation reactions such as C-H bond cleavage and carbonization which otherwise lead to defects in the framework and (3) absence of ZnCl 2 contamination in the porous materials. These advantages are attractive to obtain reasonably well-dened CTFs which can exhibit a broad variety of physical properties including chemical sensing, gas storage/separation and organic photocatalysis. So far, CTFs have been used mainly in gas storage and separation. 11,13 There are also few reports on the use of CTFs for heterogeneous catalysis including light-induced hydrogen evolution, 9,15 as catalytic supports in liquid phase reactions, 16 and as adsorbents in liquids, 17 due to their chemical and thermal stabilities. In spite of their many applications, their photoluminescence was only briey noted 14 to the best of our knowledge. Thus, we show in this paper that PCTF-8 can not only be used for gas adsorption/separation but also exhibits high selectivity for the detection of 2,4,6-trinitrophenol (TNP) over other nitroaromatic analytes such as 4-nitrotoluene (NT), nitrobenzene (NB), 2,6-dinitrotoluene (2,6-DNT), 2,4-dinitrotoluene (2,4-DNT), 2,4-dinitrophenol (DNP) together with light-induced hydrogen evolution.

Materials
All chemicals were purchased from commercial suppliers (Sigma-Aldrich, Acros Organics, and Alfa Aesar chemical company) and used without further purication, unless stated otherwise. Tetra(4-cyanophenyl)ethylene was synthesized according to the reported procedures. 11 Chloroform was distilled from P 2 O 5 under a nitrogen atmosphere.

Instrumentation
Infrared (IR) spectra were obtained on a Bruker FT-IR Tensor 37 Spectrometer in the 4000-550 cm À1 region with 2 cm À1 resolution as KBr disks. UV-vis absorption and uorescence emission spectra were obtained using Jasco V-670 and FP-6500 spectrometers, respectively. Fluorescence quantum yields were evaluated using an absolute photoluminescence quantum yield measurement system equipped with a calibrated integrating sphere and a multi-channel spectrometer (C9920-02, Hamamatsu Photonics). The excitation wavelength was 395 nm. 1 H and 13 C spectra were recorded on a Bruker Avance DRX-500 instrument. 1 H and 13 C NMR chemical shis are given in ppm relative to SiMe 4 (d ¼ 0.0 ppm) with calibration against the residual protonated solvent signal (CDCl 3 : 7.26 ( 1 H) and 77.0 ( 13 C)). 13 C crosspolarization (CP) MAS NMR measurements were performed at 100.6 MHz at a sample spinning frequency of 5 or 6 kHz using a Bruker Avance 400 spectrometer with a 7 mm zirconia rotor. For the 13 C CP-MAS NMR measurements, the repetition delay was 5 s, the contact time was 1.75 ms, and the pulse width was 4.5 ms ( 1 H 90 pulse). Elemental (CNH) analyses were carried out with a PerkinElmer 2400 series 2 elemental analyser. Powder X-ray diffraction (PXRD) data were collected on a Bruker D2 Phaser diffractometer using a at sample holder (also a at silicon, low background sample holder) and Cu Ka 1 /a 2 radiation with l ¼ 1.5418Å at 30 kV covering 2theta angles 5-80 over a time period of 2 h, that is 0.01 s À1 . Diffractograms were obtained on at layer sample holders where at low angles, the beam spot is strongly broadened so that only a fraction of the reected radiation reaches the detector which leads to low relative intensities measured at 2q < 7 . For hygroscopic or air-sensitive samples, the sample holder can be sealed with a dome. Scanning electron microscopy (SEM) images were obtained using an ESEM Quanta 400 FEG SEM equipped with a secondary electron detector. Thermogravimetric analyses (TGA) were carried out at a ramp rate of 5 C min À1 under air with a Netzsch Thermo-Microbalance Apparatus TG 209 F3 Tarsus. A coulometric Karl-Fischer titration (KFT) for the determination of the water content was carried out with a Karl-Fischer titration apparatus AQUA 40.00 with a headspace module. The solid CTF sample was heated to 170 C in the head space module and the liberated water was transferred to the measurement cell.
Sorption isotherms were measured using a Micromeritics ASAP 2020 automatic gas sorption analyser equipped with oilfree vacuum pumps (ultimate vacuum <10 À8 mbar) and valves, which guaranteed contamination free measurements. The sample was connected to the preparation port of the sorption analyser and degassed under vacuum until the outgassing rate, i.e. the rate of pressure rise in the temporarily closed manifold with the connected sample tube, was less than 2 mTorr min À1 at the specied temperature of 200 C. Aer weighing, the sample tube was then transferred to the analysis port of the sorption analyser. All used gases (H 2 , He, N 2 , CO 2 , and CH 4 ) were of ultra-high purity (UHP, grade 5.0, 99.999%) and the STP volumes are given according to the NIST standards (293.15 K, 101.325 kPa). Helium gas was used for the determination of the cold and warm free spaces of the sample tubes. H 2 and N 2 sorption isotherms were measured at 77 K (liquid nitrogen bath), whereas CO 2 and CH 4 sorption isotherms were measured at 293 AE 1 K (passive thermostating) and 273.15 K (ice/deionized water bath). The heat of adsorption values and the DFT pore size distributions ('N 2 DFT slit pore' model) were calculated using the ASAP 2020 v3.05 soware.
Photocatalytic H 2 production experiments were performed with a Solar Box 3000 equipped with a Xe lamp (300 W). For each experiment, the polymer powder (10 mg) was suspended in a phosphate buffer solution (pH-7, 9 mL) containing triethanolamine (TEoA) or MeOH as a sacricial donor (1 mL). Dihydrogen hexachloroplatinate (5 mL, 8 wt% aqueous solution) was added as the precursor for the in situ formation of a platinum cocatalyst. The solution was ultrasonicated for 15 min before being degassed by nitrogen bubbling for 30 min. The ask was illuminated at a 90 angle at a distance of 22 cm with a 300 W Xe lamp for the time specied. Gas samples were taken with a gas-tight syringe and quantied by using a gas chromatograph (Shimadzu GC-2010 Plus) equipped with a barrier discharge ionization detector (BID).
Time-correlated single photon counting (TCSPC) measurements were performed using a pulsed diode laser source (Edinburgh Instruments EPL405) operating at 404.6 nm with a pulse FWHM of ca. 77.1 ps. A neutral density lter was used to attenuate the beam to obtain photon counts of approx. 1% or less of the incoming light intensity. The detector used was a Hamamatsu MCP-photomultiplier tube R3809U-51 (cooled to ca. À40 C), the signal was passed to a discriminator (Ortec 9307) and then into a TAC (Ortec 566, 50 ns time range used). The electrical trigger signal from the laser was also passed through a discriminator (Tennelec TC454) and onto the TAC (Ortec 566). The TAC output was read by a DAQ-1 MCA computer card using 4096 channels and collected with Horiba Jobin Yvon DataStation 2.3 soware. All measurements were done in the reverse mode at 10 MHz and under magic angle polarisation. A cut-off lter (GG3) was used to block scattered excitation light. The instrument response function (IRF) was collected by using a dilute solution of Ludox as a standard scattering sample. Fluorescence lifetimes were obtained by iterative reconvolution of the IRF and the collected decay curves, by tting to a multiexponential decay model using inhouse scripts.

Synthesis
A 50 mL Schlenk ask was charged with triuoromethanesulfonic acid (0.45 g, 3.0 mmol) in 8 mL dry CHCl 3 under a N 2 atmosphere. The reaction mixture was stirred and cooled to 0 C. Tetra-(4-cyanophenyl)ethylene (217 mg, 0.5 mmol) in 30 mL dry CHCl 3 was added dropwise into the acid solution with stirring over 30 min and the temperature of 0 C was maintained for another 2 h. Then, the resulting solution was stirred for 12 h at room temperature. An NH 3 /H 2 O solution (0.5 mol L À1 ) was added until the reaction mixture became neutral and the mixture was stirred for another 2 h. The yellow product was isolated by ltration and washed with water (3 Â 50 mL), THF (3 Â 30 mL), acetone (3 Â 30 mL), chloroform (3 Â 30 mL) and further puried by Soxhlet extraction for 24 h with acetone. The nal product was dried in a vacuum. The yield was 180 mg, 83% for (C 30 H 16 N 4 ) n .

Results and discussion
Synthesis and spectroscopic characterization PCTF-8 was synthesized by the TFMS catalyzed reaction from tetra(4-cyanophenyl)ethylene (1) at room temperature as a yellow powder in 83% yield (Scheme 1). In FT-IR, the carbonitrile band in the precursors around 2226 cm À1 has signicantly decreased in intensity which indicates only a few free nitrile groups in PCTF-8 (Fig. 1, top). The intense bands around 1503, 1362 and 800 cm À1 correspond to the formed triazine rings. 12 In the solidstate 13 C CP/MAS NMR spectrum (Fig. 1, bottom), the major resonances at 174, 141 and 134 ppm can be assigned to the triazine ring (-CAr]N-), ethylene carbon (-C]C-) and aromatic carbon (Ar) atoms, respectively. 12 The peaks around 124 and 114 ppm were associated with residual cyano end groups. 14 Powder X-ray diffraction of PCTF-8 conrmed its expected amorphous nature (Fig. S1, ESI †).
Thermogravimetric analysis (TGA) showed that PCTF-8 exhibited good thermal stability up to 420 C under air (Fig. S2, ESI †). Elemental analysis (Table S1, ESI †) of PCTF-8 nearly matches the calculated values corrected with the results of KFT analysis which determined the water content of the sample (Fig. S3, ESI †). It became apparent that the PCTF-8 is hygroscopic and adsorbs about 10 wt% of moisture from ambient air upon storage under ambient conditions aer vacuum drying at 120 C. A similar hygroscopic behaviour of related CTFs was reported recently by Janiak et al. 18 and may be a common phenomenon to CTFs which may also explain the frequently observed mismatch in theoretical and experimental elemental (CHN) analysis. Under SEM, the PCTF-8 particles are small and largely irregular as well as angular shaped with a wide range of sizes (greater than 1 mm) (Fig. 2).

Gas sorption
The N 2 sorption isotherm shows a steep slope at low P/P 0 values and a type I character typical for microporous materials (Fig. 3 and Table S2, ESI †). 19 The calculated BET surface area of 625 m 2 g À1 is higher than for similar tetrakis(4-carboxyphenyl)ethylene linker-based uorescent MOFs (244-317 m 2 g À1 ). 20 The surface area of PCTF-8 is in the range of other TFMS-catalysed CTFs with 2 to 1152 m 2 g À1 . 14 The ratio of micropore volume to total pore volume (V 0.1 /V tot ) of 0.78 suggests that the material contains a very high fraction of micropores. A narrow distribution of micropores centered at 5, 5.9, 6.8, 11.8 and 14.8Å was observed (insert in Fig. 3). PCTF-8 showed a H 2 sorption of 112 cm 3 g À1 at 77 K (with no hysteresis) that is very close to those of other CTFs. 13 The CO 2 adsorption capacities in activated PCTF-8 are 56 cm 3 g À1 at 273 K and 35 cm 3 g À1 at 293 K, respectively (Fig. 4). The CO 2 uptake of PCTF-8 was comparable to that of CTF-type polymers synthesized by TFMS, which showed CO 2 uptakes of 20-93 cm 3 g À1 at 273 K and 1 bar. 14 Such uptake capacity is also comparable to that of nanoporous 1,3,5triazine frameworks (NOP-1-6, 29-56 cm 3 g À1 ). 21 The CO 2 uptake of PCTF-8 is comparable to that of PCTF-1 which has about three times the BET surface area of PCTF-8. 11 This is due to the higher nitrogen content and intact triazine units of PCTF-8 from the milder synthesis conditions as CO 2 interacts with the triazine nitrogen due to its large polarizability and quadrupole moment.
The isosteric heat of adsorption from the CO 2 adsorption isotherms at 273 and 293 K (Fig. S4, ESI †) at zero loading is 37 kJ mol À1 and rapidly drops at adsorbate loadings of only a few cm 3 g À1 . The heat of adsorption stays around 29 kJ mol À1   for loadings between 3 and $30 cm 3 g À1 . When the loading is increased beyond 30 cm 3 g À1 the heat of adsorption value appears to increase slightly which should not be over-interpreted however in view of the error margin of readily AE3 kJ mol À1 . 13,22 This heat of CO 2 adsorption of $29 kJ mol À1 for PCTF-8 stays well above the heat of liquefaction of bulk CO 2 with 17 kJ mol À1 23 or around the isosteric enthalpy of adsorption for CO 2 on activated charcoals (e.g. BPL: 25.7 kJ mol À1 , A10: 21.6 kJ mol À1 , and Norit R1 Extra: 22.0 kJ mol À1 ). 24 The high Q st value can be attributed to the high polar framework and the pore size effect. The high adsorption enthalpy at zero coverage is explained by the initial lling of the small ultramicropores with 4Å diameter (Fig. S5, ESI †) with adsorbate-surface interactions to both sides or ends of the CO 2 molecules. 13 If the heat of adsorption would indeed exhibit an increase with CO 2 uptake, this means that a simultaneous, exothermic process must take place, such as possibly the rearrangement of already adsorbed CO 2 molecules towards a closer, energetically, more favorable conguration. CH 4 sorption values of PCTF-8 were 17 cm 3 g À1 at 273 K and 10 cm 3 g À1 at 293 K (Fig. 4) which are comparable with the values for CTFs from our previous work. 11,13 From the available single-gas adsorption isotherms, the CO 2 selectivity over N 2 of PCTF-8 was calculated at 273 K by using the Henry equation and the ideal adsorbed solution theory (IAST) in order to understand greater affinity toward CO 2 . The calculated adsorption selectivity of CO 2 /N 2 is 24 and 61 at 273 K using the Henry equation and the ideal adsorbed solution theory (IAST), respectively ( Fig. S6 and Section 8 in the ESI †), which are comparable with the values for PCTF-1 to À7 from our previous work as well as other CTFs. 11,13,14,25,26 This comparably large value of PCTF-8 for CO 2 may be attributed to the presence of more basic nitrogen moieties.

Photophysical properties
Absorption and photoluminescence spectra of PCTF-8 and its monomer (1) were measured at room temperature (Fig. 5). The monomer 1 dispersed in CH 2 Cl 2 shows several absorption bands below 400 nm attributed to the p-p* transitions of the TPE chromophore. 27,28 Upon excitation at 350 nm, the monomer emits a weak blue emission with l max ¼ 389 nm (Fig. 5a). This type of subunit is known to show an aggregation induced emission in the solid state or concentrated solutions. The weak emission is due to the non-restricted rotation of phenyl rings in dilute solutions. 29,30 On the other hand, a solid-state absorption spectrum of PCTF-8 exhibits a broad absorption band centred at around 400 nm ranging from 300 to 600 nm, which can be attributed to the TPE units (Fig. 5b). These different absorption proles are responsible for the different colours; the monomer shows a light yellowish colour owing to absorbing UV light but reecting all visible light, while PCTF-8 is a bright yellow powder due to absorbing blue light (Fig. 5).
To compare the emission behaviour of PCTF-8 synthesized under different conditions (Brønsted acid and ionothermal conditions), we have measured the emission spectrum of black coloured PCTF-1 that was synthesized under ionothermal conditions (Fig. S5, ESI †). PCTF-1 does not show appreciable emission because of partial carbonization in the framework. Therefore, highly emissive properties of CTFs seem only possible when avoiding carbonization by preparation at room temperature. In the solid-state, PCTF-8 shows a strong greenishyellow uorescence with an emission maximum at 562 nm when excited at 395 nm and a uorescence quantum yield (F f ) of 31% at room temperature (Fig. 5b). This high value, compared to the extremely low quantum yield observed for the monomer in solution, is primarily attributed to the immobilization and restricted rotation of the aromatic rings in the tetraphenylethene units of the framework. 30,31 From its ultraviolet-visible spectrum, the optical band gap of PCTF-8 was estimated to be 2.25 eV. In principle, this band gap is large enough to overcome the endothermic character of the water-splitting reaction. To investigate this property, photocatalytic hydrogen evolution experiments were performed with PTCF-8 and Pt-loaded PCTF-8 in the presence of a methanol or triethanolamine sacricial electron donor (10 mol%) in buffered aqueous solutions. Without any platinum co-catalyst, no hydrogen evolution was observed. Aer loading with 2.3 wt% Pt, PCTF-8 showed H 2 production from buffered (pH 7) aqueous solutions containing sacricial electron donors under simulated sunlight (Fig. S8, ESI †). The total amount of H 2 evolved of  Pt-loaded PCTF-8 (10 mg) in a buffered aqueous methanol or TEoA solutions when irradiated for 20 h was 1780 and 2370 mmol g À1 , respectively. These moderate values of hydrogen production could be due to the poor crystallinity of PCTF-8 which hinders the diffusion of light-induced charge carriers to the surface. 9 To examine the sensing behaviour of PCTF-8, luminescence spectrometric quenching titration experiments were performed by gradual addition of nitroaromatic analytes (TNP, NT or NB, 2,4-DNP, 2,6-DNT, 2,4-DNT, cf. Fig. 6) to PCTF-8 dispersed in acetonitrile at room temperature ( Fig. 6 and 7a, and Section 11 in the ESI †).
Upon successive additions of nitroaromatic analytes, the uorescence emission intensity was increasingly quenched, with TNP being the most effective quencher, for a maximum quenching of ca. 71% of the uorescence emission intensity. The other analytes show similar quenching behaviour in the following order: TNP [ NT > NB > DNP > 2,6 DNT > 2,4 DNT (Fig. 6). Following the quenching experiment over time up to sixty minutes, it is evident that the quenching occurs very rapidly and does not increase over time (Fig. S16, ESI †).
To quantify the sensitivity towards the nitroaromatic analytes, the Stern-Volmer (SV) quenching constants (K SV ) were calculated by using the standard linear curve-tting of the uorescence intensity (I 0 /I) in the presence of different concentrations of the analytes (1 Â 10 À5 to 5 Â 10 À5 mol L À1 ) (Fig. 7b, S17-S21 in ESI †). The obtained quenching constants (K SV ) are in the order TNP [ DNP > NB > 2,6-DNT > 2,4-DNT > NT. The K SV of TNP was 1.32 Â 10 5 L mol À1 (Fig. 7b), which is comparable with those of previously reported frameworks. 32 To further understand the nature of the quenching process involved in "turn-off" uorescence sensing by PCTF-8, we measured the emission decays by time-correlated single photon counting at different concentrations of TNP (added as aliquots of a stock solution in acetonitrile), as it has the highest quenching efficiency of the studied nitroaromatic compounds. From the decays shown in Fig. 7c, it can be observed that the lifetime is independent of the concentration of the quencher in the studied range, which strongly suggests that a static quenching mechanism takes place. The static quenching behaviour of PCTF-8 with TNP can be explained in terms of the formation of a ground-state complex due to a strong interaction with the analyte (see the ESI † for discussion). 33 The lifetime of PCTF-8 was tted to three exponential decay components (c 2 ¼ 1.324, Fig. S22 in the ESI †): 0.5 ns (49%), 2.1 ns (42%) and 5 ns (1%). This decay behaviour can be rationalised in terms of the size distribution of the polymer, and due to the fact that emission comes from a heterogeneous suspension, where different aggregation behaviours can be observed.

Conclusions
In conclusion, we have used TPE units in the preparation of a porous and photoluminescent covalent triazine-based framework (PCTF) by using TFMS as the catalyst. This PCTF-8 showed photoluminescence behaviour under UV light, whereas the material PCTF-1 prepared by the high-temperature ZnCl 2 route was completely non-emissive. PCTF-8 shows sensitive detection towards nitroaromatic analytes with trinitrophenol (TNP) quenching ca. 71% of the PCTF-8 uorescence emission intensity. The quenching mechanism was found to be static in nature because the emission lifetimes did not change with increasing addition of the analytes. From optical transmission measurements, we estimated a band gap of about 2.25 eV which determines its efficiency as a photocatalyst for hydrogen evolution. The total amount of H 2 evolved of Pt-loaded PCTF-8 in aqueous methanol or TEoA was 1780 and 2370 mmol g À1 , when irradiated for 20 h. Hence, CTFs are not only interesting for gas adsorption/separation but also hold great potential for hydrogen production and for the development of new luminescent chemosensors.