Excitation of H2O at the plasma/water interface by UV irradiation for the elevation of ammonia production

Tatsuya Sakakura a, Shintaro Uemura a, Mutsuki Hino a, Shotaro Kiyomatsu a, Yoshiyuki Takatsuji a, Ryota Yamasaki ab, Masayuki Morimoto a and Tetsuya Haruyama *ab
aDivision of Functional Interface Engineering, Department of Biological Functions Engineering, Kyusyu Institute of Technology, Kitakyusyu Science and Research Park, Fukuoka, 808-0196, Japan. E-mail: haruyama@life.kyutech.ac.jp
bResearch Center for Advanced Eco-fitting Technology, Kyusyu Institute of Technology, Kitakyusyu Science and Research Park, Fukuoka, 808-0196, Japan

Received 5th October 2017 , Accepted 5th December 2017

First published on 6th December 2017


Ammonia is well known to be a very important chemical substance for human life. Simultaneously, the conventional ammonia production process needs pure nitrogen and pure hydrogen. Hydrogen has been produced from either liquid natural gas (LNG) or coal. In this study, water is used as a direct hydrogen source for ammonia production, thereby obviating the need for catalysts or water electrolysis. We have studied and developed a plasma/liquid interfacial reaction (P/L reaction) that can be used to produce ammonia from air (nitrogen) and water at ambient temperature and pressure, without any catalysts. In this study, the P/L reaction entails enhanced ultraviolet (UV) irradiation of the surface of the water phase. The nitrogen plasma/water interface reaction locus can produce ammonia. In contrast, the vacuum ultraviolet (VUV) irradiated interface reaction locus produces increased amounts of ammonia. In a spin trap electron spin resonance (st-ESR) experiment, large amounts of atomic H (H˙) were produced by UV irradiation, especially by VUV irradiation. The derived H˙ effectively enhanced the P/L reaction rate.


Introduction

Ammonia (NH3) has been widely used as an agricultural fertilizer,1,2 a reductant in NOx removal catalysts,3 and cuprammonium rayon (cupro),4 and also in other applications. More recently, it has additionally been used as a hydrogen carrier.5 The amount of ammonia being produced is 160 million tons/year and about 80% of this is used for fertilizers. It is said that the increase in the population results in the increasing use of ammonia as fertilizers. Ammonia is typically mass-produced by the Haber–Bosch (H–B) process from H2 and N2 on a large scale. Thus, the H–B process is suitable for producing high-purity ammonia in a large-scale chemical plant, which includes processes for the production of hydrogen and various other products. This means that the H–B process is required to be a combination of various processes. If the demand for ammonia increases, a technology that enables onsite production is necessary, apart from the mass synthesis technology. Clearly, there is a need for a new process to produce ammonia employing one-step production from air and water, with the aim of achieving its possible transport and onsite production.

An ammonia production process starting from N2 is tasked with breaking the N[triple bond, length as m-dash]N bond, which is known to be a very strong bond and requires the use of high temperature and pressure to break.6 The H–B process overcomes the N[triple bond, length as m-dash]N bond to expend any amount of energy at high temperature (>673 K) and high pressure (>200 bar).7–10 However, for on-site production, normal temperature and pressure would be desirable. Therefore, on-site production would need to break the N[triple bond, length as m-dash]N bond at ambient temperature and pressure.

Methods other than the H–B method have been developed for processing near room temperature, such as catalytic processes to produce ammonia starting from N2 and H2 or a nitrogen source and hydrogen source. Ammonia synthesis using Ru has been widely researched in the field of supported catalysts.11,12 Ru-based catalysts are known for their high activity and high rate of reaction for ammonia synthesis. Although these catalytic reactions require low energy compared with the H–B process, they have not yet achieved ammonia synthesis under normal temperature and pressure. Nitric acid (NO3) is an alternative source of N2 and is contained in agricultural waste produced by the bacterial decomposition of fertilizers.13 The synthesis of ammonia from nitric acid achieved at atmospheric pressure is not adequate for large-scale conversion to ammonia by way of selective reactions on catalysts.

A method using plasma for disconnection of the N[triple bond, length as m-dash]N bond in N2 has been developed.14–16 In ammonia synthesis using plasma it is possible to break the N[triple bond, length as m-dash]N bond required for catalysts because of nitrogen activation by plasma. In this case, N2 forms activated nitrogen, which is atomic nitrogen (N or N˙) and excited nitrogen molecules (N2*) in the plasma phase. However, the activated nitrogen needs a possible high-efficiency reaction and hydrogen supply. In contrast to the above methods, our method is distinctive that it presents a reaction system at normal temperature and atmospheric pressure where water is directly used as a hydrogen source. At the same time, it seems to overcome the problems associated with on-site production. In our recent study, a one-step ammonia synthesis from air (or N2) and water through a plasma/liquid interfacial reaction (P/L reaction) was successfully performed.17 Hydrogen production is unnecessary because the P/L reaction can be performed by direct hydrogen abstraction from water with a simple process at ambient temperature and pressure without any catalyst. In the P/L reaction, the hydrogen atoms of the outer-most surface water molecules in the liquid phase form special reaction fields that contribute to the gas phase side.18 In our previous reports, we also described the irradiation of the surface of the water phase with ultraviolet (UV) light in the P/L reaction locus to enhance ammonia production.

In this study, we have investigated the effect of UV irradiation by the P/L reaction to enhance ammonia production. This is a conspicuous effect because the water molecules in the outermost layer of the liquid phase surface can be excited efficiently by UV irradiation from the plasma gas phase side (as conceptually illustrated in Fig. 1). We report the advantage of ammonia synthesis with increasing hydrogen abstraction by atomic nitrogen from the excited vacuum ultraviolet (VUV)/UV irradiated water phase surface of the P/L reaction locus.


image file: c7gc03007j-f1.tif
Fig. 1 Schematic illustration of ammonia production through the plasma/liquid interfacial reaction (P/L reaction). Pure N2 was introduced into the discharger to produce N-plasma (activated nitrogen), which was introduced into the reactor. The P/L interface (reaction locus) was formed with the water phase with 2.5 mL pure water in a flat glass dish. The reaction locus was irradiated by UV light of arbitrary wavelength.

Results and discussion

Ammonia synthesis with the modulation of vacuum ultraviolet irradiation

The water molecules dissociate into hydrogen atoms and hydroxyl radicals under vacuum ultraviolet irradiation (VUV, 185 nm):19,20
 
H2O → OH˙ + H˙(1)

In addition, even under UV irradiation (254 nm), water molecules containing hydrogen peroxide dissociate into hydroxyl radicals, hydrogen ions, hydroxyl ions, and the dioxidanide.

 
2H2O2 → OH + OH˙ + HO2 + H+(2)

Increasing the source of hydrogen by VUV/UV irradiation was confirmed to increase the amount of ammonia. Actually, the activated air (plasma phase)/water (liquid phase) interfacial reaction (P/L reaction) is considered to consist of many reactions to form ammonia.21 Activated N2, mainly atomic nitrogen, has a very short lifetime, and readily recombines to form N2. However, atomic nitrogen could react with hydrogen generated by VUV/UV irradiation or extract hydrogen from H2O. Thus, there are numerous forms of hydrogen in water (liquid phase) that could bind with atomic nitrogen.

Fig. 4(a) shows the production of ammonia as a function of wavelength. We confirmed the synthesis of ammonia from water and N2 or N-plasma. In the cases of the reaction with UV irradiation at certain wavelengths (172, 185 + 254, 248, 280, 310, and 340 nm), small amounts of ammonia are produced even if the gas phase is N2 (not plasma gas). These results suggest that ammonia is derived from UV-excited water (which may produce H˙) and N2. The production of ammonia by water and N-plasma is higher than that from water and N2. In particular, the result shows that the largest amount of ammonia produced is from N-plasma under UV irradiation (185 nm and 254 nm). We confirmed that the ammonium concentration generated only by N-plasma and by N-plasma with irradiation at other wavelengths (248 nm, 280 nm, 310 nm, and 340 nm) is not much different. We considered that the wavelengths of 200 nm or less most effectively contribute to the synthesis of ammonia. This is because we found that the largest amount of ammonia is generated at wavelengths of 185 nm and 254 nm whereas not much is generated at 248 nm. Fig. 4(b) shows the relationship between the spin concentration and the wavelength change. The largest spin concentration is generated at 200 nm or less. Eqn (1) shows the water dissociation wavelength at 185 nm. Eqn (1) is apparent from the relationship between the spin concentration and wavelength. Since both the spin concentrations and ammonium ion concentrations are high at the dissociation wavelength of water of 185 nm, it is suggested that H˙ and N-plasma reacted. The spin concentration is the highest at 172 nm, but we obtained a low ammonium ion concentration at 172 nm. This suggests that ammonia is decomposed at the wavelength of 172 nm.22


image file: c7gc03007j-f2.tif
Fig. 2 Illustration of dielectric barrier discharge device (discharger) for plasma gas production.

image file: c7gc03007j-f3.tif
Fig. 3 Illustration of each reaction field by the plasma/liquid interfacial reaction (P/L reaction). The interfacial reaction at the N-plasma/water phase (a) without and (b) with UV irradiation. (c) Water irradiated with UV light after which the interfacial reaction between the N-plasma and water phase was carried out for specified reaction times.

image file: c7gc03007j-f4.tif
Fig. 4 Ammonia and spin production on the water phase surface under UV irradiation of arbitrary wavelength. (a) Ammonia production under UV irradiation of various wavelengths (172, 185 + 254, 248, 280, 310, and 340 nm). Water (2.5 mL) irradiated at various UV wavelengths (open bars). N-plasma reacted only with water (striped bar). N-plasma reacted with water under UV irradiation (filled bars). (b) Spin production under UV irradiation of various wavelengths (172, 185 + 254, 248, 280, 310, and 340 nm). Water (2.5 mL) irradiated at various UV wavelengths (dotted line open bars). N-plasma reacted only with water (the seventh bar from the left in b). N-plasma reacted with water under UV irradiation (dotted line filled bars).

The spin concentration in the water phase

Fig. 5 shows the relationship between the spin concentration and time changes. We confirmed that H˙ and OH˙ are generated because of exposing the outermost surface of water to UV radiation. The spin concentration increases with time. However, although the spin concentration increased as a result of UV irradiation, the production rate of spins slowed with the lapse of time. We considered the reason for the decrease in the production rate of spins be that the UV irradiation depth was such that the radiation only reaches water on the surface and radicals are not generated in the bulk liquid. The radicals on the water surface are trapped by 5,5-dimethyl-1-pyrroline-N-oxide (DMPO) and become DMPO – H and DMPO – OH. Therefore, less free DMPO occurs on the water surface.
image file: c7gc03007j-f5.tif
Fig. 5 Spin production as a function of duration of UV irradiation. The water phase surface was irradiated by UV light (185 nm + 254 nm). Production of spins was trapped by 0.1 M 5,5-dimethyl-1-pyrroline-N-oxide (DMPO) in the water phase.

This result suggested that the amount of spins trapped by the spin-trapping method, in which DMPO captures the radicals formed by UV irradiation because of the water molecules changing to H˙ and OH˙, depends on the surface area of the water phase and the depth of UV.

A relationship between the spin concentration and the ammonia concentration

We investigated the relationship between the spin concentration and ammonia (NH4+) (Fig. 6). The ammonia concentration due to plasma only (triangles) and UV (185 nm + 254 nm) + plasma (squares) increases with time. The result obtained by combining UV + plasma is more than twice the increase compared with plasma alone. We confirmed that N-plasma (activated nitrogen) directly reacts with water molecules, as ammonia is generated by only N-plasma.
 
N + nH2O → NH3 + byproducts(3)
where the byproducts include NOH, NO3, NO2, H2O, H2O2, H2, OH, and so on.

image file: c7gc03007j-f6.tif
Fig. 6 A triadic relationship among the number of spins, ammonia concentration, and duration of UV irradiation. The reaction time is within 300 s. N-plasma was introduced to the reactor in which the Petri dish (8.39 cm2) containing water (2.5 mL) was set. N-plasma reacted with water (triangles). N-plasma reacted with water under 185 nm + 254 nm UV irradiation (squares). A change in the spin concentration as a function of time within 5 min (circles).

The amount of produced ammonia was more than doubled by UV irradiation. This result was expected in that activated nitrogen and H˙ was expected to easily bond. However, it is clear from Fig. 6 that in the short time (within 30 s) there is not a much difference between ammonia formation by plasma and UV + plasma. This result is attributed to the increasing spin concentration (circles) as a function of time. Since the spin concentration is initially low, the amount of H˙ and OH˙ is also considered to be low in the beginning. We predicted that H˙ and activated nitrogen under UV irradiation can be more easily coupled than the direct reaction of water and activated nitrogen. Therefore, it is conceivable that slight H˙ formed by the excitation of water due to UV did not react with activated N-plasma in the initial short period of time. The amount of ammonia produced is the same irrespective of whether plasma or UV + plasma is used. The main reaction in a short time is considered to be the production of ammonia by N-plasma directly extracting H from water molecules. Since the difference between plasma and UV + plasma in terms of producing ammonia occurs for over 1 minute, the reaction is mainly with H˙ generated by UV irradiation. At the same time, ammonia production by N-plasma directly extracting H from water molecules proceeds. It is noteworthy that the ammonia concentration produced by UV + plasma is close to the value obtained by adding the ammonia concentration produced by plasma and the spin concentration. It is clear that the H˙ amount produced by UV irradiation is used for ammonia production. In addition, it is indicated that activated nitrogen easily binds with H˙. We considered that activated nitrogen requires easily reacting molecules, because activated nitrogen has a very short life. If H˙ is present in the reaction field, activated nitrogen reacts with H˙ to become NH. NH is expected to dissolve in water. Since ammonia has the composition of NH3 (with three times as much H as N), it is necessary to have a H˙ concentration of at least three times in the reaction between N and H˙ (3).

The result of our observation clearly shows that the spin concentration (OH˙, H˙) is related to the composition of ammonia. We confirmed the amount of spins generated by UV irradiation mainly by detecting the signal of OH˙ and the small H˙ signal. H˙ radicals produced by UV irradiation have a very short lifetime and recombine to form hydrogen.23

 
2H˙ → H2(4)

It is known that the hydroxyl radical (OH˙) is reduced by H2:24

 
H2 + OH˙ → H2O + H˙(5)

The relationship in eqn (4) and (5) suggests that sufficient H˙ is present to react with activated nitrogen. We obtained the result in Fig. 6 that shows the amount of ammonia produced by using UV + plasma is more than twice as much as that obtained with plasma alone. Under UV irradiation, it is considered that besides the direct reaction of activated nitrogen with H2O, NH reacts with H2O or H˙ to produce NH3.

 
N + 3H˙ → NH3(6)
 
NH + nH2O → NH3 + byproducts(7)
 
NH + 2H˙ → NH3(8)

Since H˙ generated by UV irradiation reacts with activated nitrogen, it is important to form a reaction field in which short lifetime activated nitrogen can immediately react. Therefore, a reaction field in which larger amounts of H˙ is present in the water phase could possibly produce larger amounts of ammonia.

The influence of increasing ammonia production by UV irradiation

Fig. 7 shows that the reaction was initiated after sufficient generation of H˙ by UV irradiation. We prepared UV-irradiated water 20 minutes before the synthesis reaction of ammonia (pre-UV irradiation). In this way, the concentration of ammonia that was produced increased due to the creation of a reaction field in which the activated nitrogen was enabled to fully react with H˙. In the initial stage of the reaction (10 s of Fig. 7), there was no difference in the ammonia concentration produced between the plasma-only reaction and UV + plasma reaction, but a marked increase was observed when pre-UV irradiation was employed. From Fig. 5, the spin concentration was approximately 50 μM (at 1200 seconds), suggesting that it was sufficiently consumed for increasing the ammonia production. The actual increase in the concentration was approximately 25 μM, which means that half of the spin concentration was consumed. The activated nitrogen is expected to react with the entire spin concentration, and the amount of ammonia produced is increased by increasing the spin concentration by extending the UV irradiation time. The reason why activated nitrogen did not react sufficiently within a short time despite the increased spin concentration is that the amount of activated nitrogen produced by plasma may affect the concentration of ammonia. In the 300 s shown in Fig. 7, an increase in the ammonia concentration included the spin concentration. Within the reaction time of 300 s, the concentration of ammonia is only about 30 μM when plasma only is used. The spin concentration generated by irradiation with UV for 20 minutes was approximately 50 μM, and the increment of this ammonia concentration was 80 μM (Fig. 7, striped bar), including the spin concentration 50 μM and the ammonia concentration 30 μM generated by plasma alone. Here, we suggested that the spin concentration generated by UV irradiation is largely related to the production of ammonia. In the P/L reaction, activated nitrogen reacts with most of the generated H˙ to become NH, which in turn dissolves in the water to form ammonia. At the same time, since the reaction proceeds only with plasma, a reaction of activated nitrogen obtaining H directly from H2O is also carried out. In summary, the formation of ammonia is rate limiting in the process in which activated nitrogen reacts with H˙ to become NH in water, and it is important to form a reaction field in which activated nitrogen would be enabled to react quickly. In order to increase the ammonia concentration, it is most important that a reaction field in which a large amount of activated nitrogen is generated and additional H˙ is present is formed. The ammonia concentration would be expected to increase by forming a reaction field between the gas phase of N-plasma and the water phase forming H on the outermost surface (P/L reaction).
image file: c7gc03007j-f7.tif
Fig. 7 A comparison of triadic conditions of the P/L reaction. The ammonia concentration at a reaction time of 10 s and 300 s. (a) The interfacial reaction at the N-plasma/water phase (2.5 mL) without (open bars) and (b) with 185 nm + 254 nm UV irradiation (filled bars). (c) Water UV irradiated for 1200 s, after which the interfacial reaction between the N-plasma and water phase was performed for specified reaction times (striped bars).

Experimental

The interfacial reaction (P/L reaction) between the UV irradiated water surface (liquid phase) and activated nitrogen (plasma phase)

The interfacial reaction mechanism is shown in Fig. 1. The interfacial reaction occurs between the UV irradiated liquid phase and the plasma gas phase (P/L reaction). UV light sources of 185 nm and 254 nm (6 W) were used (Sankyo Denki Co., Ltd, Japan). Other wavelengths (172 nm, 248 nm, 280 nm, 310 nm, and 340 nm) were used for comparison (248 nm, 280 nm, 310 nm, and 340 nm were generated by a MAX-303 instrument, Asahi Spectra Co., Ltd, Japan, and 172 nm by USHIO INC, Japan). The plasma gas was produced by dielectric barrier discharge (Fig. 2) using pure nitrogen (99.99%), which was introduced into the reactor at 4 L min−1. For the preparation of the plasma gas, the applied primary voltage was 6 Vp–p and the frequency was 18 kHz. Discharge voltage was determined with a high-voltage probe (attenuation factor: 1[thin space (1/6-em)]:[thin space (1/6-em)]1000) and a current probe (output rate: 1 V A−1) that were connected to an oscilloscope (TDS2024C, Tektronix Inc. USA). The discharge voltage was 5.09 kV. Pure deionized water was used for the pure water phase (specific electric resistance of deionized water ≥18.2 MΩ cm−1), which was contained in a round flat glass dish (8.39 cm2) placed inside the reactor. The reaction locus (water phase) is located 114 mm from the end of the discharge electrode. In the liquid phase, water molecules are excited by UV irradiation. Then, the nitrogen in the plasma phase and water molecules dissociated by UV irradiation were reacted. Fig. 3 illustrates three types of reaction locus and reaction procedures. Fig. 3(a) shows the reaction locus between the plasma gas phase and water phase without UV irradiation. Fig. 3(b) shows the reaction locus between the plasma gas phase and water phase with UV irradiation. Fig. 3(c) shows a two-step reaction procedure with pre-excitation of the water phase with UV irradiation. During the pre-excitation, the H˙ that was generated appeared at the water phase surface, after which the water surface forms the reaction locus with the plasma gas phase.

Quantitative analysis of produced ammonia and spins

The ammonia that was produced was detected by ion chromatography (Prominence HIC-NS, SHIMADZU Co, Japan) with a Shim-pack IC-C4 column (SHIMADZU).

Electron spin resonance (ESR) can be used to quantify the concentration of spins (JES-X310, JEOL RESONANCE Inc., Japan). It is known that 5,5-dimethyl-1-pyrroline-N-oxide (DMPO) is a spin-trapping agent for the OH radical and H radical. Thus, the spin concentration was determined by a spin-trapping method using DMPO. The magnetic field for ESR measurements was 335 ± 10 [mT], the time constant was 0.1 [s], and the modulation width was 0.2 [mT].

Conclusions

Ammonia was successfully produced by using a P/L reaction. A reaction field containing activated nitrogen required for molecules that are able to readily bind because of their short lifetime was formed. We formed a suitable reaction field by irradiating the water phase with UV radiation. This ensures the presence of a sufficient supply of H˙ in the water phase for the reaction with activated nitrogen. As a result, the amount of ammonia that was produced could be increased. It became clear that the spin concentration greatly contributes to the production of ammonia. In addition, prior irradiation of the water phase with UV radiation for the reaction with activated nitrogen dramatically increased the amount of ammonia that was produced. In conclusion, increasing the amount of activated nitrogen in the gas phase and increasing H˙ in the liquid phase by a large amount enhance the production of ammonia in the liquid phase.

Abbreviations

P/L reactionPlasma/liquid interfacial reaction (P/L reaction)
VUVVacuum ultraviolet
UVUltraviolet

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This research study was partially supported by a research grant from the New Energy and Industrial Technology Development Organization (NEDO), Japan.

References

  1. V. Smil, Nature, 1999, 400, 415–415 CrossRef CAS.
  2. H. G. Oswin, Platinum Met. Rev., 1964, 8, 42–48 CAS.
  3. M. Kang, E. D. Park, J. M. Kim and J. E. Yie, Appl. Catal., A, 2007, 327, 261–269 CrossRef CAS.
  4. G. B. Kauffman, J. Chem. Educ., 1992, 69, 311–314 CrossRef.
  5. F. Vitse, M. Cooper and G. G. Botte, J. Power Sources, 2005, 142, 18–26 CrossRef CAS.
  6. T. Shima, S. Hu, G. Luo, X. Kang, Y. Luo and Z. Hou, Science, 2013, 340, 1549–1552 CrossRef CAS PubMed.
  7. Ammonia: Catalysis and Manufacture, ed. A. Nielsen, Springer, Heidelberg, 1995 Search PubMed.
  8. J. R. Jennings, Catalytic Ammonia Synthesis. Fundamentals and Practice, 1991 Search PubMed.
  9. A. Vojvodic, A. James, F. Studt, F. Abild-pedersen, T. Suvra, T. Bligaard and J. K. Nørskov, Chem. Phys. Lett., 2014, 598, 108–112 CrossRef CAS.
  10. R. Lan, K. A. Alkhazmi, I. A. Amar and S. Tao, Appl. Catal., B, 2014, 152–153, 212–217 CrossRef CAS.
  11. Y. Inoue, M. Kitano, K. Kishida, H. Abe, Y. Niwa, M. Sasase, Y. Fujita, H. Ishikawa, T. Yokoyama, M. Hara and H. Hosono, ACS Catal., 2016, 6, 7577–7584 CrossRef CAS.
  12. C. Liang, Z. Li, J. Qiu and C. Li, J. Catal., 2002, 211, 278–282 CrossRef CAS.
  13. H. Hirakawa, M. Hashimoto, Y. Shiraishi and T. Hirai, ACS Catal., 2017, 3713–3720 CrossRef CAS.
  14. P. Peng, Y. Li, Y. Cheng, S. Deng, P. Chen and R. Ruan, Plasma Chem. Plasma Process., 2016, 36, 1201–1210 CrossRef CAS.
  15. A. Gómez-Ramírez, J. Cotrino, R. M. Lambert and A. R. González-Elipe, Plasma Sources Sci. Technol., 2015, 24, 65011 CrossRef.
  16. J. Hong, M. Aramesh, O. Shimoni, D. H. Seo, S. Yick, A. Greig, C. Charles, S. Prawer and A. B. Murphy, Plasma Chem. Plasma Process., 2016, 1–24 Search PubMed.
  17. T. Haruyama, T. Namise, N. Shimoshimizu, S. Uemura, Y. Takatsuji, M. Hino, R. Yamasaki, T. Kamachi and M. Kohno, Green Chem., 2016, 18, 4536–4541 RSC.
  18. D. Marx, Science, 2004, 303, 634–636 CrossRef CAS PubMed.
  19. K. Kutschera, H. Börnick and E. Worch, Water Res., 2009, 43, 2224–2232 CrossRef CAS PubMed.
  20. M. Li, Z. Qiang, P. Hou, J. R. Bolton, J. Qu, P. Li and C. Wang, Environ. Sci. Technol., 2016, 50, 5849–5856 CrossRef CAS PubMed.
  21. B. Y. Cyril, P. Richard, R. Mariner and M. Calvin, Proc. Natl. Acad. Sci. U. S. A., 1963, 49, 737–740 CrossRef.
  22. R. L. Lilly, R. E. Rebbert and P. Ausloos, J. Photochem., 1973, 2, 49–61 CrossRef CAS.
  23. I. V. Tokmakov and M. C. Lin, Int. J. Chem. Kinet., 2001, 33, 633–653 CrossRef CAS.
  24. I. Ohsawa, M. Ishikawa, K. Takahashi, M. Watanabe, K. Nishimaki, K. Yamagata, K. Katsura, Y. Katayama, S. Asoh and S. Ohta, Nat. Med., 2007, 13, 688–694 CrossRef CAS PubMed.

This journal is © The Royal Society of Chemistry 2018