Nội dung chính
Chúng tôi rất vui được chia sẻ kiến thức sâu sắc về từ khóa Direct conversion of N 2 and O 2 : status, challenge and perspective. Bài viết n2 o2 tập trung giải thích ý nghĩa, vai trò và ứng dụng của từ khóa này trong tối ưu hóa nội dung web và chiến dịch tiếp thị. Chúng tôi cung cấp phương pháp tìm kiếm, phân tích từ khóa, kèm theo chiến lược và công cụ hữu ích. Hy vọng thông tin này sẽ giúp bạn xây dựng chiến lược thành công và thu hút người dùng.
Nitrogen and oxygen, as the key components of air (contributing to 78.1 and 20.9 vol.%, respectively), are the vital constituents of living organisms and the critical elements of proteins, DNA, RNA and amino acids [1,2]. Humans and animals require some particular amino acids every day for nutrition and survival . On the other hand, nitrogen compounds are essential for the growth of plants and they can even synthesize amino acids from these nitrogen sources. Regarding the non-biological aspect, nitrogen oxides (NOx) are a crucial building stock for chemistry and industry. Their further products, ranging from nitric acid (HNO3), carbamide (CH4N2O) to adiponitrile (NC(CH2)4CN), are highly desired and widely used in fertilizer and plastic manufacturing, etc. For example, the world’s annual production of HNO3 reached 67.8 million tons in 2017 , while the demand for HNO3 products is still increasing . By far, almost 80% of produced nitric acid is used in the manufacture of fertilizers, among those 96% is used to produce ammonium nitrate and calcium ammonium nitrate. Others can be used to produce intermediates in the polymer industry, particularly in the manufacture of adipic acid to produce polyamides, toluene diisocyanate to produce polyurethanes and nitrobenzene to produce dyes . However, being the most abundant source in the atmosphere, nitrogen is elusive for almost all living organisms (except diazotrophs) due to its inertness. The extremely strong triple bond (945 kJ mol-1) and nonpolar stable electron configuration lead to its chemical conversion being extensively energy demanding. Thus, the naturally abundant nitrogen must first be manually ‘fixed’ (so-called nitrogen fixation).
In nature, some micro-organisms can capture the atmospheric nitrogen and supply it to plants as biological nitrogen fixation  and it annually contributes to ∼297 million metric tons of fixed nitrogen . Apart from biological nitrogen fixation, oxidation of N2 to NOx can occur through lightning formed from the electric discharge between two clouds or between clouds and the earth if the conditions (pressure or temperature) are proper for the formation of NOx. This process, as a combination of heat shock with electric discharge (thermal and plasma process), represents the early natural nitrogen fixation into NOx in Earth’s atmosphere . The total natural non-biological nitrogen fixation is reported at ∼171 million metric tons per year . Besides, human activity, such as the combustion of fuels, can more or less contribute to nitrogen fixation. Yet, it takes responsibility for some environmental issues as unusable NOx is undoubtedly a pollutant in the atmosphere . The global population expansion has intensified the demand for usable nitrogen sources not only in terms of agriculture, where nitrogen from the soil is diminishing quickly, but also in terms of industry. The naturally fixed nitrogen cycle is too slow and uncontrollable to satisfy the growing demand for nitrogen sources.
At the beginning of the twentieth century, scientists devoted lots of effort to fixing atmospheric nitrogen as the natural process had no longer met the demands. In 1895-98, German scientists Frank and Caro developed the Frank-Caro cyanamide process. In this process, nitrogen was fixed in the form of calcium cyanamide by the reaction of calcium carbide with nitrogen . The first industrialized application of atmospheric nitrogen fixation mimicked the natural process of lightening, as we have mentioned previously. This process, known as the Birkeland-Eyde (B-E) process, directly produces NOx from atmospheric nitrogen and oxygen using electric arcs to induce high temperature, whereby the activation of N2 molecules can be triggered . The B-E process was successfully realized and reported in 1903, and so was the first successfully industrialized plasma process . The plasma arcs were generated in the B-E furnace, where the temperature was high due to the thermal plasma heating, then the air rapidly passed the furnace to ‘combust’ and produce NOx. However, the energy efficiency of this process was inconsiderable and later it was prevailed by a more practical process, known as the Haber-Bosch (H-B) process. The H-B process, which was first developed in 1908 and commercialized in 1913, produces ammonia by initiating the reaction of pure nitrogen and hydrogen under heating and pressurizing conditions with the presence of iron-based catalysts . It was soon extensively studied and widely applied in industries. Over recent decades, significant development in engineering and fundamental aspects has been achieved to marginally reduce the energy consumption of the H-B process and comprehend the reaction mechanism . Until now, the H-B process has contributed to nearly 50% of the nitrogen found in human tissues and feeds ∼40% of the world’s population . Nevertheless, this process suffers from extreme conditions (heating and high pressure). It also consumes a large amount of pure hydrogen, which was unavailable in nature and has to be obtained from a steam reforming process, generating ∼1.9 metric tons of CO2 per metric ton of NH3 production (3 CO2 per 8 NH3) . On the other hand, the produced NH3 is still an intermediate in the industrial nitrogen source cycle, used as the feedstock for further products such as HNO3 through an Ostwald process. Compared to the H-B process, a more desirable route of N2 fixation is the direct conversion of N2 into N-containing organic compounds under mild conditions . The timeline of milestones in the chemical nitrogen-fixation process is shown in Fig. 1 for convenience [9,17-19].
The energy expenditures for various reductive and oxidative N2 fixation pathways have been well concluded and illustrated by Chen et al. [2,20]. We simplified the completed data and illustrated them in Fig. 2. HNO3 is typically produced through the three-step process of steam reforming followed by the H-B process, then the Ostwald process for NO2 production. However, as aforementioned, the overall process is exceptionally energetically extensive and the emission of CO2 is also dramatic. The improvement in the reaction process may alleviate the energy consumption to some extent; for example, the development of catalysts could conduct the steam reforming or H-B process under milder conditions such as avoiding pressurizing. Nevertheless, the direct oxidative process is more attractive in terms of theoretical energy efficiency, enrichment of feedstock (two main air components) and the convenience for engineering design if it can be successfully realized. The B-E process has demonstrated the possibility for the direct oxidative conversion of main air components. However, the present energy consumption of this thermal plasma process is far more than that of the three-step process, not to mention the other drawbacks such as the high investment for the equipment and the decomposition of NOx. In principle, the oxidative conversion of N2 with O2 can undergo a lower energy input than the three-step process, which indicates great potential to replace the three-step process. The fundamental chemistry of the reaction merits extensive investigation. Meanwhile, since many C-N-O organics are highly required, introducing a carbon-containing molecule ‘X’—either a simple gas or complex organic in this oxidation and conversion of the N2 process to synthesize C-N-O products such as amino acids—is of great significance. The scheme of the conversion pathways is represented in Scheme 1.
Herein, we will elaborate on the status and the challenges on the conversion of N2 and O2, and address the future perspectives of the conversion of N2/O2/‘X’ into high-value products. Furthermore, as the direct conversion of N2 must overcome an enormous-amount-of-energy barrier, thus demanding extreme temperatures, this review will mainly focus on the works and possibilities to realize this process under soft conditions, such as non-thermal plasma, electrochemical, ultrasonic and photon-driven conversion. The representative works are summarized in Table 1. All these points will be carefully illustrated in the following sections.
N2 + O2
We calculated the equilibrium composition of NOx using HSC Chemistry v6.0. The results (Fig. 3) show that the equilibrium NO concentration reaches only 3.4% in the air at 2500°C while the other NOx products are negligible. Thus, the reaction is tough to operate via a typical thermochemical process due to the demand for high temperature and the difficulty in developing catalysts. The catalyst needs to remain stable and activate N2 with the presence of O2 at these extreme temperatures. Unless an efficient catalyst can initiate the reaction under milder temperatures with pressurizing, the process is, for now, not suitable to conduct under thermochemical conditions.
The analogous radical reactions also occur in the thermal plasma process, such as in the B-E process. The gas molecules are highly dissociated and ionized in thermal plasmas through the plasma arcs under high temperatures. Similarly, catalysts are also impossible to pack into a thermal plasma reactor due to the extreme conditions (several thousand K and plasma arcs), where the catalysts are probably out of work function or melted, corroded and decomposed. Moreover, the B-E process for conversion (3.4-4.1 MJ mol-1 HNO3 production)  is far more energy wasteful than the H-B process. The overall energy efficiency of N2 activation in thermal plasmas is inconsiderable and the dissociation of NOx products is competitive with the forward reaction when no thermal quenching exists. Due to these limitations, the research based on thermal conditions has not progressed much yet. Alternative methods are essential to compensate for the energy requirements for the direct oxidation of N2 by O2 under mild conditions. For these reasons, electro-, photo-, ultrasonic- or non-thermal plasma-assisted conversion routes probably prevail and offer more excellent opportunities for progress. Moreover, the reasonable combination of two or more of those techniques may bring out ‘magical’ synergies, achieving green and mild N2 fixation.
Non-thermal plasma conversion
Non-thermal plasmas (NTPs) are a feasible solution to alleviate the limitations owing to their unique non-equilibrium property. A non-thermal plasma consists of high-energy electrons, excited species and ions. The temperature of high-energy electrons can exceed several thousand K while the bulk temperature can remain ambient. As a result, the thermally inaccessible chemical reactions can be triggered under mild conditions, thus potentially coupling with catalysis. The direct dissociation of N2 is difficult in NTPs as the dissociation energy is 9.8 eV for its triple bond, while the mean electron energy is typically ∼1-3 eV in NTPs. Therefore, the vibrationally excited N2, as the most atmospherically enriched species, plays an essential role in the NTP process. On the one hand, the accumulation of vibrational quanta from several vibrationally excited N2 molecules into one could lead to high vibrational quanta, and thus direct dissociation of the acceptor. On the other hand, a more efficient dissociative adsorption can occur on the catalysts, as shown in Fig. 4 . The dissociation of a ground-state N2 molecule undergoes the path with activation energy Et in the energy-potential diagram. For a vibrationally excited N2 molecule, the energy of the initial state is increased by the energy of vibration Ev. The vibrationally excited N2 molecule can either undergo an ideal path, where the enhancement of the initial state energy decreases the dissociation barrier by a value equivalent to (blue curve) or more than Ev (red curve), or undergo a general path, where the vibrational coordinate does not project on the reaction coordinate (yellow curve) . The single, extended vibrational ladder and a relatively large vibrational spacing (∼0.3 eV) of the N2 molecule make the enhancements in the dissociation rates possible through vibrational excitation . According to current approximates, theoretical energy consumption for the conversion of N2 with O2 into NO production via Equation (1) in a non-thermal plasma (∼400 kJ mol-1 N2) is >2.5 times lower than that for the H-B process with a methane-derived H2 source. Moreover, the energy efficiency already achieved in the laboratory scale, which ranges from 200 to 1200 kJ mol-1 N2 assuming 700%-efficient plasma generation , is far better than the H-B process (∼3000 kJ mol-1 N2 consuming H2 from water electrolysis) . If the energy consumption can be reduced to 1000 kJ mol-1, the NTP process will prevail over the three-step processes.
Indeed, from the 1980s, intensive studies on NTPs assisted conversion of N2 have been conducted, including different NTP types such as radiofrequency discharges , DC discharge [28,29], microwave discharge [30,31], glow discharge [32,33], pulsed arc and gliding arc discharge [34-36], dielectric barrier discharge (DBD) [37-41] and some related modeling [36,42-44]. Among those, DBD has drawn great interest due to its simple design and convenience for industrialization and integration with catalysts. For example, Patil et al. studied the direct synthesis of NOx by packing different catalyst support materials in a DBD reactor . Different surface areas, relative dielectric constants and particle shapes due to the different properties of support materials and their particle sizes had a significant effect on the formation of NOx . However, the author also proposed that the non-catalytic route via direct gas-phase interaction of excited N2 with O2 species was dominant . Other information can be found in the publications mentioned above.
Electronic excitation can also lead to the acceleration of NOx production while being limited by high energy costs and low efficiency . MoO3 and WO3 seem to be effective catalysts in this process . However, experimental results indicate that plasma atmospheric reactions and plasma-surface reactions occur in parallel, leading to a more complex work function of catalysts in plasma . The understanding of complex plasma-catalyst synergistic effects is indeed of great importance.
Other types of NTPs and the related engineering optimization have also been widely studied for NOx productions. For example, Jardali et al.  developed an atmospheric-pressure rotating gliding arc plasma reactor for highly efficient NOx production (concentration of ≤5.5%). The authors also studied the behavior of the plasma arc using various numerical modeling patterns. Their experimental and modeling results indicated that both the vibrationally promoted and the Zeldovich mechanisms dominated in the plasma zone as the gas and vibrational temperatures are in equilibrium at ∼2600 K . It should be noticed that catalysis is inconvenient to couple in these plasma processes, thus the selectivity towards one product, such as NO2, is beyond control. In fact, even for a plasma-catalytical process in a DBD reactor, the selectivity to NO2 is unsatisfactory. There is a pressing need to comprehend the fundamental mechanisms that unite plasma physical chemistry, gas-surface chemistry and catalysis to guide the rational design of a plasma reactor, and thus the optimization of plasma type and the development of packed-catalytic materials.
Direct oxidation of nitrogen to nitrogen-containing compounds such as oxynitride, nitrite and nitrate is one of the most challenging reactions in electrochemistry. According to the standard electrode potentials for the conceivable half-reactions calculated from their thermodynamic properties [47,48], useful oxidation state diagrams plotted with the volt equivalent of the half-reaction of a particular nitrogen compound to nitrogen versus its oxidation state were presented in the literature . Dinitrogen and ammonia are suggested to be the most stable under standard conditions and a steep climb for the oxidation of nitrogen indicates that extraordinarily high energy is essential .
The electrochemical reaction process involves reactant dissolution, mass transportation, adsorption, reaction and desorption steps, and is always accompanied by the decomposition of a solvent such as hydrogen evolution, oxygen reduction and oxygen evolution reactions in aqueous solution aroused by the competitive adsorption. However, except for the inherent inertia characteristic of nitrogen, the lower solubility, weaker adsorption and higher activation energy, which are strongly associated with the properties of reactants and solutions, electrode potential and the catalytic activity of the electrode material also make it extremely difficult for the electrochemical oxidation reaction of nitrogen to be achieved. As a consequence, only a few studies on this topic have been reported so far.
Requirements of a rational catalyst design for NOR should be far stricter than those for OER, oxygen reduction reaction and hydrogen evolution reaction, because it needs to be sufficiently active and accurately balance the competitive adsorption, activation and dissociation of N2 and H2O/OH- and the desorption of products processes. In principle, the reaction rate is measured by the activation energy for N2 dissociation, which determines the rate of dissociation and/or the effective adsorption of nitrogen and is limited by the activity and number of active sites on the surface. For reductive nitrogen fixation to ammonia, the dissociation of N2 as the rate-determining step on the most active catalysts was proposed by both theoretical calculation and experiments [52-54]. Analogously for NOR, the rate-determining step was also determined to be the dissociation of N2. The challenging task should be undertaken by developing highly efficient catalysts.
This outstanding catalytic performance could be attributed to the electronic effect and spillover effect. Specifically, the upshift of the d-band center of the Ru site in TiO2 promoted the rate of N2 activation and electrochemical conversion into NO* intermediates, which can combine with the active oxygen species spilled over from RuO2 to RuxTiyO2 to form nitrate, releasing active sites for producing more active oxygen species and accelerating the second step of NOR.
The synergetic catalytic effect for NOR was also performed on ZnFexCo2-xO4 spinel oxides in N2-saturated 1 M KOH . After correction, the highest yield rate of nitrate was observed to be 130 ± 12 μmol h-1 gMO-1 on a ZnFe0.4Co1.6O4 catalyst at a potential of 1.6 V (versus RHE). Nevertheless, the Faradaic efficiency is lower than that on the ZnFe2O4 catalyst, which shows the highest value of 10.1 ± 0.9% at 1.5 V (versus RHE). The higher catalytic activity towards OER will cause the lower Faradaic efficiency of NOR, meaning that with the increase in the applied potential, the rate of OER increases rapidly while the rate of NOR increases slowly. The roles of Fe and Co in ZnFexCo2-xO4 spinel oxides on the synergetic catalytic effect were explained by DFT, suggesting that Fe could facilitate the first N-O bond formation and Co could stabilize the adsorbed OH- for the further formation of the second and third N-O bonds.
For two completing electrochemical reactions with almost the same activation energy, the adsorption strength of the reactant on the electrode surface or adsorption energy plays a vital role because the reaction probability is primarily dependent on the amount of reactant adsorbed on the surface. However, when the reaction takes place by two different species adsorbed on the surface and one of them takes part in a competing reaction, the process becomes more complicated. In this case, the catalyst has to be well designed and synthesized to balance the adsorption and activation process for each species. For the NOR mechanism as shown in Fig. 6, it is convincing that the first N-O bond formation is a critical step. Once it is formed, further oxidation will occur relatively easily.
In acid media, Fe-SnO2 catalyst performed a bifunctional catalytic activity towards NOR and NRR (nitrogen reduction reaction) and NOR started from 1.6 to 2.1 V (versus RHE) . For NOR, on the optimized Fe (3%)-SnO2 catalyst surface, the yield of nitrate and Faradaic efficiency were 42.9 μg h-1 mgcat.-1 and 0.84% at an applied potential of 1.96 V, respectively. The yield rate was much higher than those of Ru/TiO2 at 2.2 V in neutral media and ZnFe0.4Co1.6O4 at 1.6 V in alkaline media, and the Faradaic efficiency was much lower than those of Ru/TiO2 at 1.8 V in neutral media and ZnFe2O4 at 1.5 V in alkaline media. The enhancement of catalytic activity for NOR on Fe-SnO2 was attributed to the oxygen vacancy-anchored single-atom Fe, where the energy barrier for the breakage of N≡N is lower, resulting in favorable adsorption and activation of the N2 molecule. It is also the reason for the improvement in NRR activity. Here, it should be considered that the catalyst, which can adsorb and activate N2 molecules efficiently, probably has a bifunctional catalytic activity for nitrogen fixation (NRR and NOR) in the case of the influence of the potential being negligible.
MXene, a series of novel 2D materials, was used extensively in the fields of electrochemical energy storage and catalysis. To our knowledge, the first work on using MXene-based material as an electrocatalyst for NOR was done by Yan et al. . They found that the well-dispersed Pd on MXene performed excellent catalytic activity towards NOR in neutral media. The highest yield rate of NO3- and Faradaic efficiency was obtained to be 2.80 μg h-1 mgcat.-1 (or 45.16 μmol h-1 gcat.-1) and 11.34%, respectively, at a potential of 2.03 V versus RHE corresponding to a current density of 0.4 mA cm-2. The process of NOR on a Pd-MXene catalyst surface was also concluded into two main steps that consist of the electrocatalytic conversion of nitrogen molecules to NO* intermediates as the initial step and the non-electrochemical oxidation of NO* into NO3- as the final step. It was mentioned that OER is a competing side reaction for NOR, hindering the initial step of NOR, while appropriate O2 produced by OER was regarded as the reactant in the final step of NOR.
Above all, the experimental and theoretical studies showed that the onset potential of NOR is highly close to that of OER and the linear sweep voltammetry or cyclic voltammetry curves obtained in N2 and Ar atmospheres almost overlap, resulting in lower Faradaic efficiency. Accordingly, an efficient electrocatalyst requires sufficient active sites where the chemically inert N2 molecules can be selectively adsorbed and activated, and the serious parasitic OER can be effectively inhibited during N2 oxidation. Little progress has yet been made in developing such catalysts because there is no doubt that most of catalyst surfaces prefer to adsorb OH- rather than N2, even if N2 molecules are adsorbed preferably but activated with difficulty at active sites under the specific potential. Some of them can be pushed aside by the fast adsorption of OH- and the formed O2. Furthermore, if the appropriate OER activity may accelerate the NOR process, it is advisable to design the catalysts with two different adjacent active sites for OER and NOR, respectively. It should be noted that not all active oxygen species can combine with adsorbed N2 to form NO. However, most of them combine with another oxygen species to form O2. These suggest that the rate-limiting step for NOR is probably the effective adsorption and activation of N2.
To design efficient catalysts, it becomes crucial to clarify the preferential adsorption mode of N2 and OH- at which an effective electrochemical reaction can further take place to form the first N-O bond on the catalyst surface, based on the consideration of all sorts of effects (such as electron effect, size effect, strain effect, ligand effect, boundary effect, etc.) of a well-designed catalyst. At present, bi-/multi-metallic and their oxide catalysts are attractive and promising for NOR due to the high catalytic activity for electrocatalytic reactions resulting from the synergistic effect between the different components. Consequently, it is expected that these catalysts of high activity, selectivity and stability for NOR can eventually be discovered.
Ultrasound is a kind of acoustic wave at frequencies above the audible range (above ∼20 kHz) used in cleaning, echo sounding and chemical reactions due to its good directivity and strong reflectivity. When the ultrasonic wave propagates in a medium, it undergoes physical and chemical changes due to the interaction between the ultrasound and the medium, resulting in a series of ultrasonic effects including mechanical, thermal, cavitation and chemical effects.
The ultrasonic cavitation can generate high local instantaneous temperatures and pressures and sonoluminescence [60,61]. Moreover, radicals generated during the cavitation can induce chemical reactions—the so-called chemical effects of ultrasound. These complex effects are not yet thoroughly clarified. Nevertheless, some theoretical models have been established to describe the origins of molecular activation. It was found that the molecules at the interior of the bubble of cavitation filled with vapor and gas are excited and further dissociated [62-67]. Inside the bubbles or at the interface of the two phases, the generated radicals can combine with gas to form products. Nitrogen oxidized to nitrite and nitrate directly have been achieved in aqueous medium saturated with air under an ultrasonic field.
As early as 1936, Gohr et al. found that H2O2, HNO2 and HNO3 were generated in water saturated with air under an ultrasonic field at a frequency of 540 kHz and the HNO3 formation is due to the further oxidation of HNO2 when oxygen is sufficient in water . Subsequently, further investigations of nitrogen fixation in the ultrasonic field were launched. In 1950, Ellfolk et al. [69,70] explored the factors affecting the oxidative nitrogen fixation in the ultrasonic field and found that the ratio of nitrite to nitrate was determined by the hydrogen ion concentration (or pH) of the solution and the formed H2O2 was lessened rapidly at a pH of <4. The authors considered that it is due to the consumption of H2O2 to oxidize nitrite to nitrate rather than the diminishing of the formation of H2O2. Furthermore, it was found that the process of nitrogen fixation in the ultrasonic field was inhibited in the presence of hydrogen and carbon monoxide, probably as a result of the competition of hydrogen and nitrogen for oxygen. Finally, based on the point of view of ionization potential, the first activation step of the aerobic fixation of nitrogen in the ultrasonic field was discussed and the same possible reaction pathways were summarized, as follows:
Verrall et al. studied the variety and yield of ultrasonic products in water in the presence of dissolved gases . The results indicated that the amounts of the formed hydrogen peroxide and total nitrogen fixation depend on the nature of the dissolved gases. For hydrogen peroxide, the formation follows the order O2 > air > Ar > N2. However, it follows the order air > N2 > Ar > O2 for the total amount of nitrogen fixation. In the case of using 447 kHz at 50 W irradiating, the initial formation rates of nitrite and nitrate in water saturated with air at 298 K were 22 × 10-9 and 6 × 10-9 mol min-1 W-1, respectively. The authors proposed that the aerobic fixation of N2 undergoes the dissociation of nitrogen molecular and oxygen molecular to atoms and then atomic nitrogen and atomic oxygen combined to form nitric oxide. In the absence of oxygen, the formed atomic nitrogen is reacted with hydroxyl radicals to produce NOH intermediates, which can further combine with hydroxyl or HO2 radicals to form nitric oxide and water or hydrogen peroxide.
On the contrary, it was also reported that the •OH radicals arising secondarily from water are evidently unable to oxidize nitrogen .
Ultrasonic frequency dependence of the yields of nitrite and nitrate in air-saturated water has been investigated by Tiehm et al. . In the range of 41-3217 kHz, the maximum yields were obtained at 360 kHz, which gives the formation rates of 7.1 mg (as N) L-1 for nitrate and 0.6 mg (as N) L-1 for nitrite, corresponding to 42 × 10-9 and 4 × 10-9 mol min-1 W-1, respectively. In the same report, the total nitrate + + + + Plus nitrite formation rate of 33 × 10-9 mol min-1 W-1 was obtained by using 30 W of 100 kHz ultrasound in 700 mL air-saturated water at 293 K by Petrier et al. . Besides, it was found that the yield of products changed with the irradiating time. The rate of nitrate formation increased steadily, while nitrite decreased after 500 min. The hydrogen peroxide formation rate was initially about the same as the total nitrate plus nitrite but decreased after 200 min. Therefore, it was considered that the primary products are hydrogen peroxide and nitrite, and then nitrate was formed via a pH-dependent oxidation reaction of nitrite by hydrogen peroxide.
Subsequently, Kruus et al.  investigated the effect of time, temperature and gas composition on the nitrite- and nitrate-formation rate in 300 mL air-saturated water at 278 ± 2 K under 27 W of 900 kHz ultrasound irradiation. The total formation rate increased with the decrease in temperature and with the increase in O2 fraction up to between 0.4 to 0.5 and then decreased. The highest formation rate of total nitrate plus nitrite over 20 min was 16 × 10-5 M, which is equivalent to 30 × 10-9 mol min-1 W-1 and close to the values presented above. Recently, Kobayashi et al. compared the yields of nitrite and nitrate in air, O2, N2 and Ar-saturated ultrapure water under a 23-, 28- and 43-kHz ultrasound field with 200-1200 W of output power. The optimum frequency was found to be 28 kHz and the higher the power supplied, the higher the yields produced. In these optimized conditions, the formation rate of nitrite and nitrate was deduced to be ∼0.60 and 0.44 μM min-1, respectively. However, in the presence of N2, O2 and Ar, the yields of these products are very low compared to those in the atmosphere, as Mead observed before.
To improve the nitrogen-conversion efficiency in gas-saturated aqueous solution under the ultrasound field, the idea should be guided to enhance the cavitation effect, the yield and stability of active radicals, gas solubility and dissolved fraction, and active sites of a catalyst. There are several effect factors behind this, such as pH, ionic strength, frequency, power density, gas composition, temperature, catalyst, etc. In the future, the influence rules and mechanisms of each effect factor among them should be elucidated. Based on this knowledge, it is easy to find an optimal solution for nitrogen fixation by ultrasound. For the catalyst, thermoelectric and piezoelectric materials should be given priority because these materials may exhibit excellent polarization performance under high pressure and high temperature produced by ultrasound.
Photocatalytic conversion of N2 with O2 represents a practically viable oxidative conversion of air not driven by fossil fuels. The photocatalytic nitrogen oxidation process can be divided into several steps: first, excitation of electrons from the valence band (VB) to the conduction band is initiated by light absorption, leaving holes in the VB. Then the photo-generated h+ oxidizes N2 to NO with water, while O2 is reduced to H2O by photoexcited electrons, and NO is further oxidized to nitrates evolving O2 and H2O, as shown in Fig. 7A. Overall, nitrate acid is synthesized from water, O2 and N2 under ambient conditions using sunlight as an energy source .
Photocatalysis as a green, renewable and sustainable technique has attracted massive attention on activating N-N bonds using various photocatalysts such as Diamond , BrOX [79,80], Mo-doped W18O49 , Bi5O7Br [82,83], TiO2 , LDH [85,86] and g-C3N4 . In this field, studies have mainly focused on nitrogen fixation to NH3 products, while a reaction involving N2 with O2 has been rarely attempted. In 2013, Yu et al. first reported a direct nitrate-formation process from atmospheric nitrogen and oxygen on nano-sized TiO2 surfaces under UV or sunlight irradiation [76,77]. This work is of significance as the group demonstrated that a continuous nitrate-producing reaction was observed over time. They detected an intermediate gaseous product at a retention time of 1.25 min by comparing with the results of the gas chromatogram before and after the photocatalytic reaction. According to the results of Fourier transform infrared (FTIR) differential spectrum and theoretical calculations (as shown in Fig. 7B and C) for NO formation, it was suggested that NO is an intermediate product. Afterward, Zhang et al.  successfully used Z scheme heterojunction TiO2/WO3 nanorods as a photocatalyst to synthesize NO, which is an intermediate product in photocatalytic nitrogen oxidation, and its production rate was determined to be 0.16 mmol g-1 h-1 associated with thermal energy (at 400°C) and quantum efficiency of 0.31% at 365 nm.
Photocatalytic nitrogen activation and oxidation were achieved at the photo-generated holes on the VB of semiconductors. Therefore, photocatalysts containing abundant potholes were beneficial to nitrogen fixation to nitrate. Recently, Xie et al. reported that pothole-rich WO3 nanosheets can activate the N≡N bond and synthesize nitrate directly under ambient conditions . Pothole-rich WO3 exhibited an efficient photocatalytic performance. The average rate of nitrate production is as high as 1.92 mg g-1 h-1 under ambient conditions, without any sacrificial agent or precious-metal co-catalysts under UV/Vis irradiation at 380 nm. The apparent quantum efficiency (AQE) was calculated to be 0.11%, which is better than pothole-free nanosheets and bulk WO3.
Photoconversion of nitrogen to nitrate under ambient conditions is expected as an alternative cost-effective approach for producing nitrate. However, the rate and AQE of nitrate production are too slow to meet the demand of industrial production. Great endeavors are essential to increase the efficiency of nitrate production. Moreover, the existing studies demonstrated that a good synergy between photon energy and thermal energy is more beneficial to nitrogen-conversion reactions. The development of an innovative catalytic process associated with thermal energy, or even another energy input, exhibits great potential to achieve efficient nitrogen conversion.
N2 + O2 + X
As discussed above, the direct oxidation of N2 is rather challenging due to the high energy threshold. Introducing another chemically active molecule ‘X’ alone with air to modify the reaction pathway, thus undergoing a relatively lower energy potential, could be another method to achieve the conversion of air. Moreover, the selective synthesis of high-value C-N-O organics from key components of air and C-containing ‘X’ is a holy grail in chemistry.
Inert gas (active to plasma) has been widely applied in plasma conversion of N2 and O2, mainly participating as a third-body molecule and stabilizing the discharge. Therefore, we do not include inert gas in the category of X gas in the NTPs process. Some other studies have investigated the plasma-driven reactions between CH4 and air  and simulation on plasma conversion of a gas mixture of CH4/CO2/N2/O2 . Even though these studies took a brief glimpse into the plasma chemistry of the conversion of N2, the main goal is still the conversion of CH4 and/or CO2, and needless to say the underlying reaction pathway and the possibility to soften the high energy barrier. H2O has also been involved in the NTP conversion of N2/O2 . The results indicate that the atomic oxygen and hydroxyl radical (OH) generated from O2 and H2O significantly affect the formation of NOx, proving that the presence of H2O enhanced the conversion of N2 and formation of NOx rather than N2/O2 . The addition of NO was also studied by some groups [39,40]. However, these works emphasized the oxidation of NO into NO2 and little information on the conversion of N2 with O2 can be found. Based on these existing works, further investigation should be made on how the third molecule ‘X’ mitigates the energy consumption for the conversion of N2 with O2. We believe that involving a chemically active molecule ‘X’ can be a feasible pathway in the NTP conversion of N2 with O2 and this pathway needs tremendous endeavors on both exploring an appropriate ‘X’ with appropriate catalysts and the underlying chemistry.
While for electrochemical conversion, it seems a great challenge to achieve the electrochemical conversion of N2, O2 with ‘X’ (such as CH4, CO2 or organic molecules) into organic compounds consisting of C-N-O (such as CH4N2O, RNH2, RCHNH2COOH, RNO2) because of the simultaneous adsorption and activation of N2, O2 with ‘X’ at the same or neighbor active sites under the same conditions, is almost impossible not to mention the bonding that includes the oxidation and reduction reactions simultaneously. However, it can become possible when the whole process sequentially passes through electron-transfer steps and chemical steps. More precisely speaking, the active species produced electrochemically at the active sites on the catalyst surface can induce further electron transfer or chemical reaction with other reactants nearby to form another active intermediate, which will take a further electrochemical or chemical reaction with the third reactants at the active site or in the electrolyte to form the final products, mimicking a chain reaction. According to the existing studies, the coupling reaction of N-O-X should be started with the activation of O2 to form superoxides and peroxides, which are expected to trigger this chain reaction.
Ultrasound induction could also be a possible method to achieve the coupling reaction of N2/O2/‘X’ due to its cavitation effect and chemical effect, which could dissociate N2, O2, X and H2O into atoms and/or active species, leading to the occurrence of some chemical reactions between multiple species. Particularly if the introduced ‘X’ molecule is chemically active under ultrasound conditions in the presence of catalysts, it probably enhances the conversion of N2 and O2 with lower energy potential. Even though no studies on this strategy have been reported, ultrasound induction is adopted widely in the field of organic synthesis involving N2 bonds with various components. Similar to plasma-driven or ultrasonic conversion, some active species and radicals can be produced during the photocatalytic process and the category and number of these species depend mainly on the solvent and the surface performance of a photocatalyst.
Reasonably coupling two or more aforementioned processes (as shown in Scheme 2) can provide great potential to effectively transform N2, O2 and ‘X’ into high-value C-N-O organics as the reaction coordinates may differ greatly from one to another. For example, thermal catalysis undergoes the chemical reaction via a translational mode, while plasma catalysis initiates the chemical reaction by electron impact and vibrational and electronic excitation. It is worth exploring both the fundamental science of the possible synergies and the engineering improvements to amplify the synergies.
Since the beginning of the twentieth century, the demand for nitrogen fixation has been dramatically increasing. Researchers used to commercialize the B-E process to produce NOx from the air, yet it was soon surpassed by the H-B process due to the higher energy efficiency. However, this process still suffers from extreme conditions, consuming pure hydrogen sources and generating ∼1.9 metric tons of CO2 per metric ton of NH3 production. The one-step direct conversion of air for NOx products rather than the H-B process coupled with the Ostwald process is more attractive in terms of theoretical energy efficiency, enrichment of feedstock and the convenience for engineering design if it can be successfully realized. The most severe challenge for this process is to overcome the tremendous-amount-of-energy barrier. Therefore, a sole thermochemical or thermal plasma process is not favorable as few catalysts are stable in such high temperatures. Other processes, such as non-thermal plasma, electrochemical, ultrasonic and photon-driven conversion, could be appropriate to convert air into desired NOx products under soft conditions. However, not many efforts have been dedicated to this challenging field compared with ammonia synthesis. To make a brief conclusion of existing works, great opportunities exist in the direct conversion of N2 with O2, while progress will require far more improved energy efficiency (at least <3000 kJ mol-1 N2) from a macro perspective, a molecular-level understanding of nitrogen transformation reactions, as well as mechanistic insights into the discovery of new catalytic systems and multiple means of delivering the energy needed to drive those reactions from a micro perspective.
What is important for the future? This review has addressed the technical and scientific challenges of the direct conversion of air into NOx products. The authors also would like to provide some perspectives and strategies, some of which have been briefly discussed, for further investigation on the direct conversion of N2 and O2: (i) coupling multiple processes, (ii) introducing another gas molecule to undergo a softer reaction pathway and (iii) development of new catalysts.
Multiple processes of coupling allow multiple forms of energy input, thus providing different pathways to activate N2 and trigger its reaction. A sole thermochemical process suffers from the extreme conditions, high requirements for equipment and low energy efficiency. Coupling thermochemical process with the other energy inputs is simply designed and has been applied in thermo-electro, thermo-photo and thermo-plasma conversion as we discussed above. The enhancement on the conversion of N2 and O2 may be significantly amplified with the other energy inputs by tuning different reaction coordination. For example, a conversion of 3.8% and energy efficiency of 2 ngàn kJ mol-1 for N2 can be achieved in a microwave plasma reactor by coupling thermal and plasma processes . Once the benefit of energy efficiency surpasses the cost for reaching those conditions, it will be feasible for future scalability.
On the other hand, the conversion of N2 and O2 under mild conditions is the goal for the scientific community, which is also the focus of this review. The coupling of two or more processes among plasma-chemical, electrochemical, ultrasonic and photon-driven conversion can be one of the promising solutions to achieving this goal due to their special methods of energy inputs. However, no attempts at systematic studies have been made to find an effective method or process for coupling multiple energy inputs and investigating the possible synergy between those. Tremendous fundamental studies are required to discover how the synergies between different processes and catalysis work, and to realize sophisticated engineering improvements to maximize the synergies.
Direct conversion of N2, O2 and C-containing ‘X’ into high-value C-N-O organics is the long-standing and final pursuit of key components of air transformation. Although no studies have been reported to our knowledge until now, introducing a chemically active ‘X’ can alter the reaction pathways, thus probably undergoing lower energy thresholds for which some works have demonstrated the possibility. Exploration has been conducted on seeking an appropriate ‘X’ for different processes. It also requires extensive studies and in-depth comprehension of fundamental chemical coordination. The development of a new catalytic system and the exploration of the catalytic mechanism is also one of the permanent research cores for all these processes, particularly for coupling processes. The understanding of catalysis under this complex system is of great importance for catalyst design. All these advances will emerge through collective understanding and insights to be comprehended from fundamental research that associates experiments and theory in catalysis and different processes.
We would like to thank Prof. Shengming Ma (Department of Chemistry, Fudan University) for his valuable advice on the review.
Profs. Zhenfeng Xi, Shengming Ma, Ping Chen, Zhang-Jie Shi and Dehui Deng are working together as a team in the National Natural Science Foundation of China, Center of ‘Transformation Chemistry of Key Components of Air’ (21988101). Activation and conversion of N2 and O2 directly is one of the major aims of this center.
Conflict of interest statement. None declared.