Selective Catalytic Conversion of Acetylene to Ethylene Powered by Water and Visible Light


 The production of polymer-grade ethylene requires the purification of ethylene feed from acetylene contaminant. Accomplishing this task by state-of-the-art thermal hydrogenation requires high temperature, external feed of H2 gas, and noble metal catalysts, and is not only expensive and energy-intensive but also prone to overhydrogenate to ethane. We report a photocatalytic system to reduce acetylene to ethylene with >99% selectivity for ethylene under both non-competitive (no ethylene co-feed) and competitive (ethylene co-feed) conditions, and near 100% conversion under the latter industrially relevant condition. Our system uses a molecular catalyst based on earth-abundant cobalt operating under ambient conditions and sensitized by either [Ru(bpy)3]2+ or an inexpensive organic semiconductor (mpg-CN) under visible light. These features and the use of water as a proton source offer substantial advantages over current hydrogenation technologies with respect to selectivity and sustainability.


Results and discussion
The energy footprint associated with commodity separation and purification accounts for 10-15% of the world's energy consumption and the demand for commodities could triple by 2050 1 .
A prime example is the commodity chemical ethylene (C2H4), an intermediate in the production of ~50-60% of all plastics 2 . Ethylene is obtained by steam cracking, and typically contains ~1 vol.% acetylene (C2H2) contaminant. Acetylene is poisonous to downstream polymerization catalysts and therefore is usually removed by catalytic thermal hydrogenation (Fig. 1A) with a precious metal (Pd)-based catalyst 3 . This process has ample room for improvement in terms of price (e.g., ~10,000$ per kg Pd 4 ) and selectivity (85% for >90% conversion at 200 °C 5 where the side product is ethane). Over-hydrogenation to ethane is an inherent limitation of precious metal catalysts that achieve high conversion using H2 as the hydrogen source 3,5,6 . Traditional thermal hydrogenation also has disadvantages with respect to safety and sustainability: it is an energyintensive (high-pressure, high-temperature) thermochemical reaction in an H2 atmosphere, where the H2 is usually needed in excess and this excess must be separated to avoid possible thermal runaway processes 3 . Physisorption-based acetylene purification (Fig. 1B) is energy-efficient 7,8 , but wastes the separated acetylene rather than converting it to valuable ethylene.
Ideally, we would transition from energy-intensive thermochemical routes for ethylene purification to sustainably driven, selective electrochemical and/or photochemical semihydrogenation of acetylene impurity. A scheme that does not require an H2 feed would mitigate greenhouse-gas emission since steam methane reforming, which co-produces 75.0 kg CO2 per MJ of H2, accounts for three quarters of all H2 global production 9 . Recently, the mild conditions of electrochemical transformations have been exploited to achieve near 100% conversion of acetylene in the presence of excess of ethylene (0.5 vol.% C2H2, 20 vol.% C2H4, Ar balance) with 21% energy efficiency, and ethylene selectivity of 90.1% with 0.08% residual hydrogen 10 . Direct photochemical hydrogenation could be at least as impactful as electrochemical approaches given that it has the potential to use abundant, sustainable solar irradiation, rather than the electrical grid, as its energy input [11][12][13][14][15] . With respect to light-powered hydrogenation, a photothermal system achieved 99% conversion in the presence of excess of ethylene and H2 gas (1 vol.% C2H2, 20 vol.% C2H4, 20 vol.% H2, Ar balance) but with a selectivity of only 93.5% 16 . A photocatalytic system achieved 97% selectivity but with a pure acetylene feed (5 vol.%, He balance) and H2 gas, and had a lower saturated acetylene hydrogenation yield (~5%) than the analogous thermal reaction (~18%) 17 . Both of these systems have the disadvantage of requiring the precious metal Pd as the catalyst and an external feed of flammable H2.
Here, we achieve >99% selective, visible light-powered, catalytic conversion of acetylene to ethylene with no Pd and no H2 gas, and at room temperature. Water is the proton source and an There is much precedent for chemical 21,22 and photo-reductions of cobalt compounds to catalytically active Co(I) states [23][24][25] to access a diverse set of catalytic reductions. Our choice of catalyst is bio-inspired. Co(I) complexes of vitamin B12 or porphyrin act as non-protein models for nitrogenase in the reduction of non-physiological nitrogen-(e.g. nitrous oxide) and carbon-based (e.g. alkyne) substrates 21,22,26 . Reduction of Co(III) porphyrin or Co(II) phthalocyanines by sodium borohydride enables conversion of acetylene to ethylene, but with significant ethane and hydrogen side products 21,22 . Light-driven strategies for catalytic acetylene to ethylene conversion have however not been reported.  Table 1). Upon increasing [Ru(bpy)3] 2+ by a factor of five to 250 M, the TON (C2H4) increased by ~35% with SC2H4 >99% after 6 h of illumination; further increase of [Ru(bpy)3] 2+ did not increase the amount of produced C2H4 ( Supplementary Fig. 2).
The concentration of C2H4 produced increased with the concentration of catalyst but saturated at Supplementary Fig. 3), probably because of parasitic light absorption by CoTPPS ( Supplementary Fig. 4). No C2H4 is produced in the absence of sensitizer, catalyst, sacrificial donor, light or C2H2 feedstock (Fig. 2B). When we perform the photoreduction reaction with cobalt nanoparticles (as opposed to CoTPPS), the TON decreases dramatically, and when we add mercury to the CoTPPS system, the TON is not meaningfully affected (Supplementary Table   1). These control experiments unequivocally prove that this reaction is primarily homogeneous and photocatalytic, powered by [Ru(bpy)3] 2+ -sensitized CoTPPS.
The photocatalytic activity of our acetylene reduction system is tolerant to presence of various gases, including O2, CO2, and CO ( Supplementary Figs. 6-8), the latter two of which are typically present as impurities 27 . CO absorbs to the active sites of conventional hydrogenation catalysts and acts as an inhibitor; tolerance to these adventitious adsorbates highlights our system's potential advantages over traditional catalysts in an industrial setting.
We can replace the [Ru(bpy)3] 2+ photosensitizer with the organic semiconductor mpg-CN, which is a broadband absorber and can be prepared at only a few dollars per kg from readily available starting material 19,20   A major advantage of the CoTPPS system with respect to sustainability is that it does not require an external feed of H2, so we investigated the source of protons/hydrogens further. Our three-component system did not evolve any detectable H2 (Supplementary Table 1), so it is extremely unlikely that in situ production of H2 is a source of hydrogen here. When we conducted experiments in H2O and using C2H2 (5 vol.%, He balance) as feedstock, gas chromatography/mass spectrometry (GC-MS) analysis identified C2H4 (m/z = 28) as the reaction product (Fig. 3D green).
When the photoreduction was instead performed in D2O solvent, we observed C2D4 (m/z = 32) ( Fig. 3D purple), produced by exchange between the feedstock C2H2 and D2O, which we preequilibrated before illumination 28,29 ; the supplementary materials contain details of sample preparation and GC-MS experiments. These two experiments prove unambiguously that not only is acetylene the precursor for the observed C2H4, but also the protons added to make the C2H4 reduction product originate from the water solvent.  Figs. 10 and 11). We determined the order of charge transfer reactions by measuring the bimolecular rate constants for quenching of the photoluminescence of PS* by Stern-Volmer analysis (Fig. 3C and the supplementary information). These data show that photoinduced hole transfer to NaAsc is a factor of 1,000 faster than photoinduced electron transfer to CoTPPS.
The nucleophilic attack by Co(I) species on one of the carbon atoms of the electrophilic C2H2, via a  complex, is followed by a rapid addition of a proton from water 23,30,31 . In the final segments of the photocatalytic cycle (Fig. 3A), a second protonation of the Co−C bond by water yields C2H4 and [Co II P(H2O)2] and re-starts the cycle.
We confirmed that the Co(I) species is the active site to which C2H2 binds by adding NaBH4, which is known to reduce the Co(III) of the CoTPPS to Co(I), to a CoTPPS solution without photosensitizer or light 21,22 . We convert acetylene to ethylene under these conditions, although H2 is the major product 21,22 (Supplementary Table 1).
We rule out a mechanism in which [Co I P] − reacts with water to yield the cobalt hydride intermediate [Co III −H], which then coordinates C2H2 and releases C2H4 after a second protonation step 23,32,33,37 because our three-component system did not evolve any detectable H2, even when the sample was purged with Ar instead of C2H2 (Supplementary Table 1). Based on a previous report of H2 evolution using a CoTPPS catalyst at pH <8 in a phosphate buffer 35 , we believe we disfavor formation of [Co III −H] due to the pH of our bicarbonate system (8.4, whereas the acidbase equilibrium constant between Co(I) and Co(III)−H is pKa = 7.7). We also rule out the presence of radical intermediates because C2H4 formation is not affected by adding the radical trap TEMPO to the reaction mixture (Supplementary Table 1) [32][33][34]36 .
We specified further the geometry of the interaction between C2H2 and the CoTPPS catalyst and trans-C2H2D2 at 943 and 987 cm −1 , respectively) 38 (Fig. 3E). This result shows that the proton additions from water occur predominantly in a syn-manner (on one side of the triple bond of the substrate), which is consistent with the formation of a cis-alkene product from the cobalt(I)catalyzed reduction of propylene in alkaline media 30 and consistent with our proposed mechanism.
We suspect that the high selectivity of our three-component photocatalytic system originates in the better ability of the nucleophilic Co(I) species of the CoTPPS catalyst to coordinate highly electrophilic alkynes than less electrophilic alkenes 23,31,39,40 .
Importantly for potential industrial application of this process, our [Ru(bpy)3] 2+ -sensitized CoTPPS system selectively photoreduces acetylene even in the presence of excess of ethylene (1 vol.% C2H2, 30 vol.% C2H4, He balance). This ethylene/acetylene mixture is a typical industrial ethylene feed and requires a highly selective catalyst to eliminate ethane production. Our system achieved near 100% conversion of C2H2 from this mixture with >99% selectivity for ethylene (no detectable C2H6) after 18 h of illumination ( Fig. 4 and Supplementary Fig. 12).
This result demonstrates the suitability of this system for the isolation of a pure industrial propylene stream from propyne 41 , a task not accomplished by steam cracking but necessary for the production of polymer-grade propylene, which together with ethylene accounts for ~80% of global plastic demand 2 .
Finally, CoTPPS can be replaced with a tetracarboxyphenyl cobalt porphyrin (CoTPPC) (Supplementary Table 1), which can be anchored to an electrode to exploit this catalytic cycle in a photoelectrochemical cell 42 .
Until photosynthetic cascade reactions are developed enough to produce ethylene from CO2 feedstock, we need to produce this important commodity chemical with the lowest energy footprint possible. Our selective photocatalytic strategy is a major step toward that goal. Our system reduces acetylene into ethylene with several advantages over the present hydrogenation technology, including (i) operation with near 100% conversion in an ethylene-rich gas feed and >99% selectivity under both non-competitive (no ethylene co-feed) and competitive (ethylene co-feed) conditions, the latter being industrially relevant; (ii) operation at room temperature using light and water in place of high temperature and an external H2 feed, and (iii) use of earth-abundant cobalt in the catalyst, which works not only in combination with the benchmark photosensitizer [Ru(bpy)3] 2+ , but also with inexpensive and organic mpg-CN. The photoreduction strategy is sustainable, robust, and selective enough to complement, if not eventually replace, thermal hydrogenation to create polymer-grade ethylene.