Manuel
Rivas
a,
Luís J.
del Valle
ab,
Pau
Turon
*ac,
Carlos
Alemán
*ab and
Jordi
Puiggalí
*ab
aDepartament d'Enginyeria Química, EEBE, Universitat Politècnica de Catalunya, C/Eduard Maristany, 10-14, 08019, Barcelona, Spain. E-mail: pau.turon@bbraun.com; carlos.aleman@upc.edu; jordi.puiggali@upc.edu
bBarcelona Research Center for Multiscale Science and Engineering, Universitat Politècnica de Catalunya, C/Eduard Maristany, 10-14, Ed. C, 08019, Barcelona, Spain
cB. Braun Surgical, S.A. Carretera de Terrasa 121, 08191 Rubí, Barcelona, Spain
First published on 18th December 2017
The industrial process of nitrogen fixation is complex and results in a huge economic and environmental impact. It requires a catalyst and high temperature and pressure to induce the rupture of the strong N–N bond and subsequent hydrogenation. On the other hand, carbon dioxide removal from the atmosphere has become a priority objective due to the high amount of global carbon dioxide emissions (i.e. 36200 million tons in 2015). In this work, we fix nitrogen from N2 and carbon from CO2 and CH4 to obtain both glycine and alanine (D/L racemic mixture), the two simplest amino acids. The synthesis, catalyzed by polarized hydroxyapatite under UV light irradiation and conducted in an inert reaction chamber, starts from a simple gas mixture containing N2, CO2, CH4 and H2O and uses mild reaction conditions. At atmospheric pressure and 95 °C, the glycine and alanine molar yields with respect to CH4 or CO2 are about 1.9% and 1.6%, respectively, but they grow to 3.4% and 2.4%, when the pressure increases to 6 bar and the temperature is maintained at 95 °C. Besides, the minimum temperature required for the successful production of detectable amounts of amino acids is 75 °C. Accordingly, an artificial photosynthetic process has been developed by using an electrophotocatalyst based on hydroxyapatite thermally and electrically stimulated and coated with zirconyl chloride and a phosphonate. The synthesis of amino acids by direct fixation of nitrogen and carbon from gas mixtures opens new avenues regarding the nitrogen fixation for industrial purposes and the recycling of carbon dioxide.
On the other hand, CO2 recycling is an absolute necessity for our society knowing that its accumulation in the atmosphere is now approaching 1 Tera ton.9 Significant steps have been taken towards the utilization of CO2 in order to convert it into valuable chemicals (150 Mt urea, 100 Mt methanol, 70 Mt salicylic acid, 9.7 Mt formaldehyde and 0.7 Mt formic acid are the most produced).10–12 In the last few years, catalytic reactions via carbon dioxide fixation have gained a prominent role as representative green processes with enhanced sustainability.13–16 However, by learning from nature, photosynthesis as performed by living organisms is the carbon fixation process par excellence. Efforts to mimic it synthetically have been elusive through time. The development of photosynthetic processes requires significant advances in new materials for light harvesting and the development of fast, stable, and efficient electrocatalysts.17,18
In this work, we introduce a new catalyst based on permanently polarized crystalline hydroxyapatite (p-cHAp) with enhanced electrical and photochemical properties that allows the coupling of nitrogen and carbon fixation processes. Such a catalyst family opens an interesting field of research where simple gas mixtures that usually do not react among themselves are combined to yield basic organic molecules, such as AAs, one of the main building blocks of life. More specifically, in this work we prove that glycine (Gly) and alanine (Ala) are produced at atmospheric pressure through an artificial photosynthetic nitrogen and carbon fixation reaction, starting from a weakly reducing atmosphere (N2, H2O, CO2 and CH4) and using UV radiation as a source of energy. This reaction represents a very simple alternative to the costly chemical and enzymatic processes used to produce commercially Gly and Ala.
Fig. 1 Schemes describing (a) the preparation of the p-cHAp/Phos-ZC-Phos catalytic system and (b) the reaction medium used to produce AAs. The details of the reactor are provided in the ESI.† (c) Representative results of the ninhydrin test for a positive reaction before stirring (left) and after stirring (middle) and a negative reaction (right). |
The synthesis of AAs was carried out in an inert reaction chamber under a weakly reducing atmosphere constituted of N2 (0.33 bar), CO2 (0.33 bar), CH4 (0.33 bar) and liquid H2O, using an UV lamp directly irradiating the catalyst and gas mixture (Fig. 1b). The formation of primary amines adsorbed onto the solid substrate was shown by positive ninhydrin tests through the development of purple spots inside the catalyst recovered after the reaction (Fig. 1c, left). Amine compounds were well dissolved in an acetone solution after vigorous stirring (Fig. 1c, middle), contrasting with the uncolored solid and solution (Fig. 1c, right) observed under many other assayed reaction conditions and catalytic systems (see below).
After 72 h at 95 °C, the clean synthesis of both Gly and Ala is demonstrated by the NMR spectra displayed in Fig. 2. The 1H NMR spectrum of the samples obtained by dissolving the catalyst and products of the reaction (Fig. 2a) shows the presence of the catalyst signal corresponding to the Phos methylene group (doublet at 3.79–3.76 ppm) and the signals corresponding to the newly produced AAs such as methylene protons (singlet at 3.65 ppm) of Gly and both methine (quadruplet at 3.91–3.85 ppm) and methyl (doublet at 1.54–1.52 ppm) groups of Ala. The same compounds are also observed in the solid-state 13C NMR spectrum (Fig. 2b), where only peaks assigned to the Phos (54.34 and 53.00 ppm), Gly (171.95 and 41.26 ppm) and Ala (175.25, 50.25 and 16.01 ppm) units are detected. The 31P NMR spectrum (Fig. 2c) shows the presence of the p-HAp (7.21 ppm) and Phos (−0.03 ppm) peaks, but additional signals related to the products coming from the decomposition of the catalyst were not detected. With respect to CH4 or CO2, the Gly and Ala molar yields at 95 °C after 72 h are 1.9% and 1.6%, respectively (Table 1). Instead, after 24 h at 95 °C no trace of Gly and Ala is detected by NMR. Furthermore, the formation of AAs is unsuccessful without the sustained exposure to the UV radiation, which appears to be a fundamental issue to generate radicals (e.g. ˙CH3 and ˙OH) needed for further reaction intermediates towards the final yielding of Ala and Gly.
Experimental conditionsa | Yield | |||
---|---|---|---|---|
Mineral | P (atm) | Time | Gly | Ala |
a Temperature was kept fixed at 95 °C in all cases. b P N2 = 0.33 bar; PCH4 = 0.33 bar; PCO2 = 0.33 bar. c P N2 = 2 bar; PCH4 = 2 bar; PCO2 = 2 bar. d P N2 = 1 bar; PCH4 = 2 bar; PCO2 = 2 bar. e P N2 = 3 bar; PCH4 = 2 bar; PCO2 = 2 bar. | ||||
p-cHAp | 1b | 24 h | — | — |
p-cHAp | 1b | 72 h | 1.9% (6.7 mg cm−2) | 1.6% (3.3 mg cm−2) |
p-cHAp | 6c | 24 h | 1.9% (6.6 mg cm−2) | 1.3% (2.7 mg cm−2) |
p-cHAp | 6c | 72 h | 3.4% (11.9 mg cm−2) | 2.4% (5.0 mg cm−2) |
p-aHAp | 6c | 72 h | 0.9% (3.2 mg cm−2) | 0.7% (1.3 mg cm−2) |
p-cHAp | 5d | 24 h | 1.5% (5.3 mg cm−2) | 0.9% (1.9 mg cm−2) |
p-cHAp | 7e | 24 h | 2.4% (8.3 mg cm−2) | 1.6% (3.3 mg cm−2) |
Photoemission spectroscopy (XPS) analyses show that the amines in AAs come from the molecular nitrogen and not from a hypothetical decomposition of the Phos. The N 1s spectra registered for different representative samples (Fig. 3a) indicate that the peak at 399.5 eV, which is ascribed to the C–N of Phos, is observed with practically the same intensity when both negative and positive reactions (i.e. without and with sustained exposure to UV radiation) are monitored. However, only in the latter case bands at lower binding energies appear due to the formation of deprotonated and protonated amino groups (401.2 and 404.5 eV, respectively).19 Furthermore, the amount of atomic nitrogen increased from 0% to 2.75–2.97% when the Phos-ZC-Phos 3-layer coating was added to the p-cHAp substrate. The nitrogen percentage, after the positive reaction, was increased up to 6.2%, which corroborates the fact that the formed AAs remain adsorbed into the catalysts.
Fig. 3 (A) N 1s high-resolution XPS spectra for (1) p-cHAp, (2) p-cHAp/Phos-ZC-Phos, and (3) p-cHAp/Phos-ZC-Phos after the negative reaction (i.e. without exposure to UV radiation) and (4) p-cHAp/Phos-ZC-Phos after the positive reaction (24 h at 95 °C). SEM micrographs before (B) and after (C) successful reaction using the p-HAp/Phos-ZC-Phos catalytic system (24 h at 95 °C). The reactions were performed at 1 bar. Variation of the Gly/Phos (○), Ala/Phos (□) (Gly + Ala)/Phos (◆) and Gly/Ala (▲) ratios versus (D) time for reactions performed at 95 °C and (E) temperature for reactions performed for 24 h using the p-cHAp/Phos-ZC-Phos catalytic system prepared as displayed in Fig. 1a. The reactions were performed at 6 bar. |
The p-cHAp/Phos-ZC-Phos catalyst exhibits a rough and relatively irregular surface morphology (Fig. 3b), which changed after the reaction due to the sporadic formation of micrometric prismatic crystals with the hexagonal basal plane parallel to the disk surface (Fig. 3c). Although the ninhydrin test reflects the presence of AAs adsorbed inside the catalyst, micrographs also demonstrate the growth of AA crystals on the surface of the catalytic system. This behavior is consistent with the capacity of organophosphonate films for inducing the crystallization of oriented molecular sieves as proved by the growth of stable, vertically-oriented and one dimensional aluminum phosphate crystals.20,21 On the other hand, the formation of AAs has also been proven by FTIR spectroscopy, which shows intense absorption bands in the amine region of the corresponding spectra, and characteristic X-ray diffraction patterns. Finally, chiral high-performance liquid chromatography (HPLC) analyses were carried out to quantify the ratio of D- and L-Ala adsorbed into the catalysts. As was expected, after dissolution of the catalysts after the reaction in a 0.1 M HCl solution with 50 mM NaCl, a racemic D-Ala:L-Ala mixture was determined.
The effects of both the reaction time and the temperature in this catalytic process were studied considering a chamber pressure of 6 bar. The variation of the Gly/Phos and Ala/Phos ratios, which were determined from the areas of signals corresponding to CH2 (Gly and Phos) and CH (Ala) protons in the 1H NMR spectra, against the reaction time (from 2 to 96 h) reflects that Gly is produced first (Fig. 3a), while Ala is subsequently derived from this simple AA. Thus, the Gly/Ala ratio decreases from 5.4 to 2.2, while a continuous increase of the Gly/Phos ratio with the reaction time is nevertheless observed (i.e. from 0.8 to 4.5). On the other hand, the minimum temperature required for the successful production of detectable amounts of Gly and Ala after 24 hours is 75 °C (Fig. 3b). The Ala/Phos ratio increases progressively with the reaction temperature while the Gly/Phos ratio decreases at 105 °C due to the conversion of Gly into Ala, even though the (Gly + Ala)/Phos ratio still increases.
For the sake of completeness, the effectiveness of the substitution of the Phos-ZC-Phos 3-layer coating by two possible combinations of 2-layer coatings (Phos-ZC and ZC-Phos), by Phos monolayers or by a Phos/ZC mixture was also assayed. In all cases, negative results were obtained demonstrating the importance of the 3-layered architecture, and thus discarding a process based on the photodecomposition of Phos. In addition, both the capability of incorporating water molecules24 and the effective role of metal/phosphonate compounds as chelating ligands25 are noteworthy due to the presence of free phosphonic acid groups. The results obtained after introducing the above-mentioned changes in the 3-layered p-HAp catalyst are summarized in Fig. 4.
Fig. 4 Summary of the results obtained using different experimental conditions: green and red boxes refer to the successful and unsuccessful production of AAs, respectively. The details about the 16 different experimental conditions are provided in the ESI.† |
These observations support the idea that the nitrogen source for the AAs is N2 and not Phos. Furthermore, the experiments performed without incorporating the coating to the catalyst failed to produce AAs (Table S5†), whilst XPS measurements evidenced that N2 was adsorbed by the Phos layers of the coating (Table S3†). Altogether, these results support that the catalyst coating facilitates both the adsorption and further reactivity of N2. This point is particularly relevant since most of the reported AA syntheses are based on ammoniacal solutions as the nitrogen source. It is well known that N2 is fixed by means of some specific bacteria by the action of enzymes that have metal-sulfide active sites.26 Numerous efforts have been made to develop metal-based catalysts to coordinate and activate N2 through the formation of metal–dinitrogen complexes with limited success.27 The splitting of the N–N bond in bridging dinitrogen complexes was first achieved by Cummins, who showed how three-coordinate molybdenum amido-complexes react with dinitrogen at −35 °C to yield an unstable dinitrogen bridging complex that, subsequently, breaks down into two molecules of the nitride-complex.28 Although some additional progress has been made in this area, the fixation of atmospheric N2 remains a challenge because of the limited reactivity and the harsh conditions required to convert dinitrogen into useful nitrogen-containing organic compounds. The p-cHAp/Phos-ZC-Phos catalyst overcomes such drawbacks rendering nitrogen containing compounds under relatively mild reaction conditions.
Regarding carbon fixation, experiments performed without CH4 in the initial mixture failed to produce AAs and, therefore, prove its role as a carbon source for the methyl group of Ala and for the methylene group of Gly (i.e. it again discards a synthetic route based on the decomposition of Phos). On the other hand, the carboxylic groups of the two AAs would appear to come from CO2. The transformation of CO2 into organic moieties under UV irradiation, mimicking photosynthesis, represents a very interesting approach for the minimization of CO2 emissions and for developing sustainable industrial processes. Thus, the catalytic process reported in this work fulfils the very ambitious goals of artificial photosynthesis: to use earth-abundant and inexpensive materials converting them into highly appreciated organic compounds (i.e. AAs) of industrial and environmental significance. Although zirconium phosphonates are able to adsorb CO2,29 HAp has shown a noticeable capacity to incorporate carbonate and hydrogen carbonate from atmospheric CO2 when dissolved in water23,30 facilitating the formation of carboxylic groups. In this work, XPS experiments prove that both the p-cHAp substrate and the ZC layer are responsible for the CO2 adsorption onto the catalyst, while Phos layers do not contribute.
Finally, water also displays a fundamental role in the synthesis of AAs (Fig. 4). In fact, water molecules do not appear as an element source of AAs, but they have a significant influence on the mechanism contributing to the formation of ˙OH radicals and dissolving CO2. Furthermore, hydration of Phos layers is also expected to play a major role in ionic mobility through such Phos layers.31
Our first hypothesis for the reaction mechanism of the processes presented in this work is the occurrence of water-splitting photocatalysis on the surface of p-cHAp, in a way that resembles the one on TiO2 electrodes earlier reported by Fujishima and Honda.34 Thus, the reaction is not successful without water, which is consistent with the fact that water plays an important role in the mechanism. The hydroxyl radical (˙OH) photoproduction on those mineral surfaces is a well-known process that can be described as follows.35 The photon energy is adsorbed and electrons are excited to the conduction band, leaving holes in the valence band (h+). These electron–hole pairs travel to the surface of the substrate to participate in water splitting. Thus, the transfer of electrons from water molecules to valence band holes forms ˙OH (eqn (1)), whereas in the corresponding reductive reaction the transfer of electrons to O2 molecules forms superoxide radical anions (O2˙−).
H2O + h+ → {˙}OH + H+ | (1) |
Because of the electrical properties of p-cHAp, we have hypothesized a similar process for the water photolysis on this mineral, regardless of the possible variations in the intermediate stages of the photooxidation and photoreduction reactions. It is worth noting that p-cHAp presents vacancies that originate from the transformation of hydrogen phosphate into phosphate during the thermally stimulated polarization process, as proved by NMR experiments. Therefore, its catalytic role should be comparable to that of activated TiO2 surfaces. Furthermore, Nishikawa36 demonstrated that, after heat treatment, HAp becomes electro-conductive by UV irradiation, leading to the formation of stable but very reactive O2˙− by altering its own PO43− groups. These features together with the electrochemical activity of p-cHAp, in combination with the Phos-ZC-Phos layers could produce an electrochemical capacitor,37 explaining the electrophotocatalytic activation of the p-cHAp/Phos-ZC-Phos system. Accordingly, the mobility of O2˙−, which may produce ˙OH in an aqueous environment, H+ and vacancies could be the driving forces of this electrochemical mechanism. The high electrochemical activity of p-cHAp in comparison with the as-prepared and sintered cHAp is proved in Fig. 5, which presents the voltammetric charge per surface unit (Q) against the number of consecutive oxidation–reduction cycles. As can be seen, Q is significantly higher for p-HAp than for the two non-polarized samples, such a difference increasing with the number of redox processes.
As the hydroxyl radicals are very oxidative in nature, we suggest several possible reaction steps that could lead to the production of carbonaceous intermediates required for the production of Gly and Ala:
CH4 + {˙}OH → {˙}CH3 + H2O | (2) |
˙CH3 + {˙}OH → {:}CH2 + H2O | (3) |
On the other hand, the one electron reduction of CO2 into its radical anion, CO2˙−, is proposed as a net electron transfer process possible at the interface between p-cHAp and Phos-ZC-Phos:
CO2 + e− → CO2˙− | (4) |
Although this process is expected to be disfavored, because of both the thermodynamic stability of CO2 and the energy required for a stereochemical change in the geometry from a linear to a bent configuration, it has frequently been observed on HAp surfaces.23,30 This step is consistent with the analysis of volatile compounds, which revealed the formation of small amounts of aldehydes. Thus, around 600 μg m−3 of acetaldehyde and 90 μg m−3 of formaldehyde were detected after 16 h of reaction at 95 °C.
Besides, a number of mechanisms have been reported to explain the electrochemical transformation of N2 into NH3 through different steps that involve the N–N cleavage and the N protonation onto the catalyst surface.38 Within this context, such a transformation could follow associative or dissociative mechanisms. In the associative mechanism, N2 molecules adsorbed onto the catalysts are protonated sequentially without breaking the N–N bond until the first NH3 molecule is produced, as occurs in the reaction of N2-fixing by enzymes;39 while in the dissociative mechanism, the N2 molecules dissociate immediately and nitrogen radicals are protonated as they are in the Haber–Bosch process.40 After this the resulting NH3 may be easily oxidized by ˙OH to form ˙NH2.41 Then, Gly may be produced in the successive free-radical reactions involving ˙NH2, CO2˙− and ˙CH3 and the intermediate species. Finally, the transformation of Gly into Ala is possibly due to the α-H–(CαH) electrochemical removal and subsequent reaction with ˙CH3.
In order to get thermally and electrically stimulated minerals, preliminary assays were performed with s-cHAp discs using DC voltages that ranged from 250 to 2000 V. For this purpose, s-cHAp discs were sandwiched between stainless steel (AISI 304) plates, heated in a furnace to 1000 °C in air and, simultaneously, polarized for 1 h under the application of a constant DC voltage of 500 V (3 kV cm−1). Subsequently, the samples were cooled to room temperature, maintaining the DC voltage. The best results, in terms of both the mechanical consistence and maximum adsorption capacity of phosphates and phosphonates, were obtained at 500 V. After this, the discs of s-cHAp, s-aHAp, s-Nanofil 757 or s-LM were polarized using the same procedure for 1 h under the application of a constant DC voltage of 500 V, and the resultant polarized systems are denoted as p-aHAp, p-cHAp, p-N757 and p-LM, respectively.
After different trials to evaluate the influence of the content of ZC (detailed in the ESI†) all the results displayed in this work, including the main text, correspond to the ZC layer deposited from a 5 mM solution (unless another concentration is explicitly indicated).
Regarding the temperature range, the reactions were performed within 75–105 °C for reaction times between 2 and 96 h. The catalyst samples weighed approximately 150 mg and 0.5 mL of de-ionized liquid water were initially incorporated into the reaction chamber, except for assessing the water effect. The chamber was extensively purged with the first selected gas in order to eliminate the initial air content (i.e. N2 or CO2). After this, in order to reach the target pressure with the right mixture, each selected gas was introduced to increase the reaction chamber pressure (measured at room temperature) as described in the ESI.† The yields of AAs were determined by using the areas of 1H NMR signals corresponding to CH2 (Gly) and CH3 (Ala) protons, as is described in the ESI.†
Footnote |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c7gc02911j |
This journal is © The Royal Society of Chemistry 2018 |