Pd loaded amphiphilic COF as catalyst for multi-fold Heck reactions, C-C couplings and CO oxidation

COFs represent a class of polymers with designable crystalline structures capable of interacting with active metal nanoparticles to form excellent heterogeneous catalysts. Many valuable ligands/monomers employed in making coordination/organic polymers are prepared via Heck and C-C couplings. Here, we report an amphiphilic triazine COF and the facile single-step loading of Pd0 nanoparticles into it. An 18–20% nano-Pd loading gives highly active composite working in open air at low concentrations (Conc. Pd(0) <0.05 mol%, average TON 1500) catalyzing simultaneous multiple site Heck couplings and C-C couplings using ‘non-boronic acid’ substrates, and exhibits good recyclability with no sign of catalyst leaching. As an oxidation catalyst, it shows 100% conversion of CO to CO2 at 150 °C with no loss of activity with time and between cycles. Both vapor sorptions and contact angle measurements confirm the amphiphilic character of the COF. DFT-TB studies showed the presence of Pd-triazine and Pd-Schiff bond interactions as being favorable.


Materials and methods
All the organic chemicals were purchased from Sigma Aldrich and were used without any further purification.
(s1). Deborah C. T and Tomikazu S;J. Org. Chem. 1994,59, 679-681. Synthesis of trzn-COF 2,4,6-Tris(p-formylphenoxy)-l,3,5-triazine (TRIPOD) (100 mg, 0.23mmol) 1,4-diaminobenzene (48 mg, 0.46mmol) were dissolved in 1,4-dioxane (5.0 mL) in a Pyrex tube, to this mixture mesitylene (5.0mL) was added and the contents were homogenized by stirring. Following this, about 0.5 mL of aqueous acetic acid (3M) was added. Then the Pyrex tube was flash frozen in a liquid nitrogen bath, the free space was evacuated and the tube was closed under a blanket of nitrogen. The tube was placed in an oven at 120 ºC for 3 days. A brown solid was obtained, which was washed with DMF , dioxane , acetone and THF. The yield was about 110mg (87%). Activation of the sample for gas adsorption was done by soaking the sample in THF for 3days with three time replenishment of the solvent. (CHN Analysis: Obsd. C = 66.9; H = 3.74; N = 17.38. Calc. 70.9; H = 5.41; N = 15.04, (Note: the CHN values have been calculated using a COF made with a TRIPOD monomeric unit to phenylenediamine ratio of 2:3. This is the ratio employed in the synthesis).
Immediately an orange-yellow precipitate started to appear. Following this, the reaction mixture was stirred for additional 5 mins. Contents were left standing at RT for 3 days. The precipitate, a light brown solid, was filtered and washed with copious amounts of 1,4-dioxane, MeOH DMF, THF and Acetone. The procedure yielded 106mg of the trzn-COF (∼83%). CHN analysis: C = 66.48;H = 3.73;N = 17.58. Calc. 70.9;H = 5.41;N = 15.04. This sample gave lower porosity compared to the sample obtained from 120 o C synthesis.
Based on porosity as the major aspect, we gave more priority to higher temperature synthesised COF and all preliminary studies were done for RT-COF expect catalytic applications. PXRD indicated the room temperature sample to be more crystalline compared to the 120 o C synthesis.

General Procedure for Heck reaction
Aryl halide (1.0 mmol), butyl acrylate or styrene (1.1 mmol), sodium acetate (1.2 mmol), and Pd-trzn-COF (1mg) were added to 3ml of N-Methyl-2-pyrrolidone (NMP). The reaction mixture was stirred at 120 o C for 1hr in open air, whereas in the case of di to hexa bromination, the reaction was carried out over 6-10 hrs and 3 to 5mg of catalyst was used. At the end of the reaction (as monitored from TLC), the reaction mixture was poured in to water and extracted with DCM. DCM was evaporated under reduced pressure. Products were purified by column chromatography and characterized by using NMR spectroscopy.
After each cycle, the reaction mixture was centrifuged and the catalyst was recovered, for the next cycle it was used directly without any further treatment.

General Procedure for Ulmann type (C-C) coupling
Aryl halide (0.5mmol), Potassium Carbonate (0.6mmol), and Pd-trzn-COF (2mg,) were added to 3ml of DMF. The reaction mixture was stirred at 120 o C for 6hrs in open atmosphere. After completion of reaction (as monitored by TLC), the reaction mixture was poured in to water and extracted with DCM. DCM was evaporated under reduced pressure. Products were purified by column chromatography and characterized by using NMR spectroscopy and some of them were isolated as single crystals and characterized using SCXRD.

Reaction conditions:
Aryl halide (1.0 mmol), phenylboronic acid (1.1 mmol), NaOH (1.2 mmol), Tetra-n-butylammonium bromide (TBAB) (1.2 equiv), 1mg of Pd-trzn-COF were added to 3mL of water. The reaction mixture was stirred at 65 o C for 4hrs in open air. After completion of reaction, the reaction mixture was poured in to water and extracted with DCM. DCM was evaporated under reduced pressure. Products were purified by column chromatography and characterized by using NMR spectroscopy and in some cases using SCXRD.

CO to CO 2 Oxidation experimental conditions:
The catalytic activity of supported Pd-trzn-COF catalyst for CO oxidation was measured in a fixed bed reactor under atmospheric pressure using 100 mg pelletized catalyst. The temperature was ramped between 30°C to 300°C at 2°/min ramping rate. The total flow rate was 50 ml/min with a ratio of (1:5:19 CO: O 2 : N 2 ).
The calculated GHSV was 30000 cm 3 / gcat/ hr. The temperature of tubular furnace where the reactor was mounted was controlled by Radix6400 temperature controller and the catalyst bed temperature was measured by a K-type thermocouple. The effluent gases were analysed online by gas chromatograph equipped with online gas sampling valve and a TCD detector. The activity was examined by looking at the CO conversion.

CO initial
Where CO T is amount of consumed at particular temperature T.

COF-PMMA composite preparation:
A SPEKTROSPIN spin coater was used for making the films of trzn-COF/PMMA and Pd-trzn-COF/PMMA composites on glass substrate. Before spin coating the substrate was cleaned with soap and distilled water. It was then soaked in IPA and acetone for half an hour.
In a typical process, a suspension of 50 mg trzn-COF/Pd-trzn-COF and 50 mg PMMA (50% weight loading of COF in PMMA) in 5 ml THF was heated at 65°C for 12 hours with vigorous stirring. The solution, while hot, was dropped using a dropper on to the substrate maintained at ~60 o C, in a quantity just enough to cover it. It was spun at room temperature at 500 rpm for 60s, 1000 rpm for 60 s, 1500rpm for 60s and 2000 rpm for 60 s. This was repeated as long as the desired thickness of film was obtained. To remove the extra solvent, the film was dried in open air for 30 minutes and then placed in a UV curer for 30 minutes after spin coating. To peel off the film from the glass substrate, it was soaked in distilled water for 2 minutes. The obtained film was again dried in an oven for half an hour.

Analytical characterizations
Powder X-ray diffraction: Powder XRDs were carried out using a Rigaku Miniflex-600 instrument and processed using PDXL software and for some cases Bruker Discover.
Thermogravimetric Analysis: Thermogravimetry was carried out on NETSZCH TGA-DSC system. The routine TGAs were done under N 2 gas flow (20ml/min) (purge + protective) and samples were heated from RT to 500ºC at 2K/min. FEI (model Tecnai F30) high resolution transmission electron microscope (HRTEM) equipped with field emission source operating at 300 KeV was used.
X-Ray photoelectron spectroscopic (XPS) measurements were carried out on a VG Micro Tech ESCA 3000 instrument at a pressure of >1 x 10 -9 Torr (pass energy of 50 eV, electron take-off angle of 60°, and overall resolution was 0.1 eV).

Ultra Plus Field Emission Scanning Electron Microscope with integral charge compensator and embedded
EsB and AsB detectors. Oxford X-max instruments 80mm 2 . (Carl Zeiss NTS, Gmbh), Imagin conditions: 2kV, WD=2mm, 200kX, Inlens detector.         . Left: A BET fit for as-synthesized trzn-COF, carried out using the 77K N 2 isotherm showing a surface area of 408 m 2 /g, exclusively for the micropore region (P/P 0 = 0.0001-0.15); the mesoporous region could not be fit satisfactorily with a BET model. Right: The BET fit for the Pd-trzn-COF carried out using the 77K N 2 isotherm and in the micropore region showing a surface area of 404m 2 /g; the mesopores could not be fit satisfactorily using BET. BJH model fitted to the mesoporous region presented a pore size of 23Å, which is in reasonable agreement with the structure. The Dubinin-Radushkevich (DR) model gave a pore volume of 0.21cc/g and a surface area of 563m 2 /g.

Discussion on bimodal pore distribution:
Based on our BET and DFT model based analyses of the 77K N 2 isotherm, we do see the presence of these micropores. We realized two possibilities that could create micropores in this material. Considering the very less dense powder character of the material and based on the microscopy images we anticipated some inter-particle spaces that are in the microporous regime. For this reason we prepared this sample differently using methods like mechanical grinding and/or sonication and carried out the N 2 adsorptions. All such studies gave a same isotherm profile with an isotherm showing microporous behavior at low P/P 0 . Another reason could be the high temperature (120 o C) synthesis resulting in stacking faults creating micropores, again this is unlikely as we see that the room temperature synthesis where such stacking faults are expected to be minimal also shows this typical micro-mesoporous isotherm profile. This indicates that these pores are inherent to the material. However, now we have used other models (BJH and DR) to describe the pore size distribution, which seem to yield a single mesopores, but they are not capable of fitting the micropores.

Supporting discussion on lack of pore size reduction with Pd loading:
It is quite unusual trzn-COF did not lose any N 2 uptake on Pd 0 loading unlike other Pd 2+ loaded COFs. This could suggest the presence of only Pd 0 in our COF as it occupies much lesser pore space than Pd(OAc) 2 . Adding the Pd 0 to the weight of the adsorbing sample should be expected to bringing down the total N 2 uptake? One possible explanation would be that the surface grafted Pd 0 also acts as a N 2 adsorption site, which would add to the overall N 2 uptake and purely by coincidence this matches up with the porosity of the pristine COF. Also, significant amount of the Pd nanoparticles are much larger than the pore size (2.7nm) and are located on the surface of the COF as observed from the FE-SEM and HR-TEM. These particles most likely do not seem to hinder the access to the pores, in which case, again no significant drop in porosity would be expected on Pd loading. Importantly, when a model independent BJH method is used, we obtain mesopores of size 23Å for the as-made form, trzn-COF and this drops down to 19Å for the Pd loaded phase, Pd-trzn-COF. This drop could certainly be attributed to the loading of some small (< 3nm) Pd nanoparticles into the pore. Amount of catalyst used in each reaction = 1mg for 1.0mmol of reactant, reaction time1hr at 120 o C, Aryl halide (1.0 mmol), butyl acrylate or styrene (1.1 mmol), sodium acetate (1.2 mmol), and Pd-trznCOF (~0.01 mol%) were added to 3ml of N-Methyl-2-pyrrolidone. % of yield = (Actual weight of the yield/ predicted weight of the product)100 ; TON (Turnover number) calculation = moles converted / moles of active sites; TOF ( Turnover frequency) = TON/Time hours. Note the wide range of substrates with aromatic and aliphatic substituents. Amount of catalyst used in each reaction = 4mg for 1.0mmol of reactant, Aryl halide (0.5mmol), Potassium Carbonate (0.6mmol),and Pd-trzn-COF (2mg,) were added to 3ml of DMF. The reaction mixture was stirred at 120 o C for 6hrs in open air.

Additional test to confirm lack of Pd leaching:
In a separate experiment, the Pd-trzn-COF was stirred in DMF solution at 120ºC for 1hr and when the supernatant was isolated and evaporated to dryness, and analyzed using EDAX it showed nil Pd redundant. Importantly, this extract was suspended back into DMF and it was employed for Rxn. 5 in the table s1 and no product was obtained.

Interactions of Pd-trzn-COF with H 2 and O 2 :
We carried out these measurements to screen for any type of chemisorption at these temperatures. H2 (77 and 303K) and O2 (273 and 303K) adsorptions were carried out on Pd-trzn-COF. No adsorption was observed for O 2 and for H 2 very little physisorptive uptake (~1 mmol/g) was observed only at 77K and none at RT. Also, the material showed no degradation upon exposure to these gases during the adsorption. Also, the sample exposed to oxygen during the CO oxidation experiments was intact, confirming the stability of the sample to oxygen even at elevated temperatures. Thus indicating an apparent stronger chemical interactions with CO over these gases. Note that the interaction with CO is claimed from the oxidation experiment.

trzn-COF-PMMA and Pd-trzn-COF-PMMA composites
A major concern with Schiff base COFs is the hydrolyzable nature of the Schiff bonds under both acidic and basic conditions or even when subjected to heating in aqueous solutions. Banerjee and co-workers developed a COF based on keto-enol tautomerism, 37 which has shown exceptional chemical stabilities to date.
However, implementing this chemistry in many monomeric units is not easy or would be expensive. In fact, the starting materials for their COF is quite expensive. This brings forth the need for development of other strategies for improving the stability of these Schiff based COFs. Since covalent organic frameworks are built up entirely from organic components they should poses inherent characteristics to blend with other organic materials.

PMMA-COF composite preparation:
In a typical synthesis of the composite, the trzn-COF (50mg) or Pdtrzn-COF (50gm) was dispersed in THF solution (5mL) containing PMMA (50mg). Membranes were made by spin coating these dilute dispersions at 60°C. A loading of up to 50% could be achieved. The resulting composite can be made into membranes of different shape and size. Contact angle of the membrane was measured to be in the range of 110-120°. The presence of the hydrophobic casing could favor longer life of the catalyst, and facilitate easy handling and recovery.
Unfortunately, the PMMA part of the composite is incompatible with DMF. To tackle this issue, we have managed to use the membrane tablets of the composite in methanol and carry out the catalysis for Rxn. 5 in Table s1. This yielded the same product with similar yields indicating no loss in activity of the catalyst membrane. The catalyst-PMMA membranes showed no sign of Pd leaching during catalysis. However, significant swelling of the membrane was observed in methanol.

S.No
Structure Name Table   1 Dibutyl 3 Note: The ORTEP plots for the products have been shown with a 50% probability.