Green photochemistry: solar photooxygenations with medium concentrated sunlight

Abstract

The rose bengal sensitized photooxygenations of citronellol and 1,5-dihydroxynaphthalene were performed successfully under solar irradiation conditions, and complete conversions (>95%) were achieved in almost all cases in relatively short illumination times. The selected reactions were easily performed on multigram to kilogram scales using cheap and commercially available starting materials, and yielded important key-intermediates for industrial applications.

---

Introduction

Over the last few decades, the growing demand for environmentally friendly technologies has led to an increasing interest in Green Chemistry.¹ Among the many greenchemical approaches, photochemistry can serve as a valuable application since light is regarded as a clean reagent.² Despite obvious advantages in terms of selectivity and sustainability, photochemical applications for the production of chemicals on large industrial scales remained, however, rare.³ To overcome this neglect, photochemical reactions have been recently subjected to moderately or highly concentrated sunlight,⁴ and selected examples for the production of specific fine chemicals using modern solar collectors have been reported.⁵

Results and discussion

All experiments were performed at the solar-chemical facility of the German Aerospace Center (DLR) close to Cologne, Germany (latitude 50°51′ N, 7°07′ E, 70 m above sea level).⁶ The research site receives about 1500 hours of direct sunshine (with a peak in July/August) and about 850 kWh m⁻² of direct insolation per year, and thus provides sufficient conditions for solar-chemical operations. Two different sunlight-collecting systems were used for the present investigation.⁷ The larger PROPHIS plant for multikilogram syntheses⁸ and a smaller parabolic trough collector designed for laboratory-scale applications.⁹ Both reactor types require direct sunlight since they can only concentrate the direct part of global radiation. The technical key-data of each reactor are given in Table 1.

Table 1 Technical data of the solar reactors^a

	PROPHIS	Small parabolic trough
a At the location of the DLR in Cologne, Germany. b Geometric concentration factor; ratio of the collector aperture area to the absorber area. c Dichromated gelatine (Holotec GmbH, Germany); reflectivity range 550 ± 140 nm.
Volume/l	35–120	0.2–1
CF^b/suns	32	15
Aperture
Total/m^2	32	0.188
Size per trough	189 cm × 4.5 m	20 cm × 94 cm
Troughs	4	1
Concentrator material	Ag (on glass)	DCG^c (between glass)
Position to sun	2 axis-tracking	Tracking (elevation)
		Manual (azimuth)
System	Circulating	Circulating

---

Dye sensitized photooxygenations are superior model reactions for solar-chemical applications.^5,10 For the present study, we have selected rose bengal as a sensitizer since it shows a favorable absorption up to 600 nm with a maximum at 555 nm (data in ethanol).†^5b

Photooxygenation of citronellol

The photosensitized oxygenation (or Schenck-ene-reaction) of citronellol 1 was studied as a first example (Scheme 1).¹¹ This reaction is currently performed on a ca. 100 t a⁻¹ scale by Symrise (former Dragoco and Haarmann & Reimer) in Germany using artificial light sources.^11a Further reduction and acid mediated cyclization of the regioisomer 2b gives the important fragrance rose oxide, which makes this photoreaction a prototype for solar photochemical comparison studies.¹² For the ‘outdoor’ reactions in the PROPHIS loop (Table 2), the solvent methanol from the industrial process was replaced by the less hazardous isopropanol.

Scheme 1 Solar photooxygenation of citronellol 1.

Table 2 Experimental data for the photooxygenation reactions of citronellol 1 using the PROPHIS loop

	Experiment I	Experiment II
a Time until conversion reaches an almost constant value. b Estimated amount of photons collected between 500–600 nm. c Estimated amount of photons (500–600 nm) for complete conversion. d Conversion of citronellol 1 as determined by GC-analysis (vs. tetradecane) after reduction with Na2SO3.
Date	11.8.1997	12.9.2002
Scale
Citronellol/l	5.8	8.0
Rose bengal/g	20	36
i-PrOH/l	40	72
O2-flow/l h^−1	600	200
Aperture/m^2	8	32
Fluid flow/l min^−1	30	30
Temperature/°C	20	20
Time (CEST)	10:55–14:30	10:20–18:00
Total (effective^a)/h	ca. 3 ½ (3)	ca. 7 ¾ (2 ¼)
Photons^b (effective^c)/mol	54.7 (47.1)	442.6 (133.4)
Conversion^d (%)	>95	>95

---

For the first experiment performed in August 1997, only one trough with an aperture of 8 m² was used. The reactor was loaded with a solution of 5.8 l (31.8 mol) of citronellol (7) and 20 g of rose bengal in 40 l of isopropanol. Upon illumination, the amount of citronellol was rapidly consumed and after ca. 3 h, an almost quantitative conversion of 1 was already achieved. GC-analysis, performed after reduction of the corresponding sample with Na₂SO₃, furthermore proved the high purity of the regioisomeric photoproducts which were formed in a ratio of about 45 ∶ 55 in favor of 2b (determined vs. tetradecane as internal standard). The latter finding is in good agreement with the reported isolated yields of 35% (2a) and 60% (2b) from laboratory experiments with artificial light.^10b During the experimental period the reactor collected 47.1 mol of photons in the important absorption range of rose bengal between 500–600 nm.‡

In the second run in September 2002, all 4 troughs were used, giving a total aperture of 32 m² (Fig. 1). In addition, the experiment was scaled-up to 8.0 l (43.9 mol) of citronellol (1) and 36 g of rose bengal in 72 l of isopropanol. The weather conditions were optimal and consequently, total conversion was readily observed after less than 3 h (Fig. 2). During this time, the PROPHIS plant received 133.4 mol of photons in the range of 500–600 nm—almost 3-times as much as during the experiment with only one trough.‡

Fig. 1 PROPHIS reactor during the photooxygenation of citronellol 1.

Fig. 2 Direct normal irradiance and conversion vs. illumination time for the photooxygenation of citronellol 1 (Experiment II).

Remarkably, the increase in aperture did not cause a better efficiency of the photooxygenation process. This becomes obvious when comparing the conversion based on time and area profiles for both experiments (Fig. 3). The main reason for this observation is the poor distribution of oxygen and thus the different saturation of it inside the reactor tubes, because the given technical set-up allows gas feeding only at the bottom of the first trough. Furthermore, the given feeding-nozzle generates relatively large gas bubbles with a small overall surface.

Fig. 3 Conversion based on time and area vs. illumination time for both photooxygenations of citronellol 1.

Photooxygenation of 1,5-dihydronaphthalene

An additionally interesting application of the solar-chemical concept was the synthesis of the important intermediate Juglone (5-hydroxy-1,4-naphthoquinone, 4)¹³ from 1,5-dihydroxynaphthalene (Scheme 2). With artificial light sources, the photosensitized oxygenation of 3 furnishes Juglone in yields of 70–75%, even on multigram scales.¹⁴ Noteworthy, most of the thermal alternative pathways suffer from several disadvantages concerning yield, selectivity, sustainability or reproducibility.¹⁵

Scheme 2 Solar photooxygenation of 1,5-dihydroxynaphthalene 3.

For the solar-chemical experiments, we have selected a small parabolic trough collector equipped with holographic mirrors (Table 1; Fig. 4). The given holographic concentrators (2 elements; 20 × 100 cm total) are especially designed to reduce warm-up effects (and thus the costs for process cooling) caused by infrared radiation and show a reflectivity range of 550 ± 140 nm—optimal for the usage of rose bengal.⁹ Oxygen is added via a simple Y-connector which does not allow a constant oxygen flow or a homogeneous distribution of it within the absorber tube.

Fig. 4 Small scale parabolic trough reactor equipped with holographic mirrors during the solar photooxygenation of 1,5-dihydroxynaphthalene 3.

In August 2003, two laboratory-scale experiments were conducted using rose bengal as sensitizer and isopropanol as solvent. In contrast, the laboratory procedure commonly uses methylene blue and acetonitrile or a mixture of methanol with dichloromethane,¹⁴ respectively. The experimental details and results from the solar-chemical studies are summarized in Table 3. In both cases, the progress of the reaction was followed via GC-analysis vs. tetradecane as internal standard.

Table 3 Experimental data for the photooxygenation reactions of 1,5-dihydroxynaphthalene 3 using the laboratory-scale reactor

	Experiment III	Experiment IV
a Time until conversion reaches an almost constant value. b Estimated amount of photons collected between 500–600 nm. c Estimated amount of photons (500–600 nm) for complete conversion. d Conversion of diol 3 as determined by GC-analysis (vs. tetradecane). e Isolated yield of Juglone 4. f Yield calculated based on conversion.
Date	4–5.8.2003	12–13.8.2003
Scale
Diol 3/g	2.0	1.0
Rose bengal/g	0.1	0.1
i-PrOH/ml	200	200
Aperture/m^2	0.188	0.188
Fluid flow/l min^−1	ca. 60	ca. 60
Temperature/°C	20	20
Time (CEST)
1st day	13:25–17:45	13:40–16:45
2nd day	08:50–16:20	09:10–15:40
Total (effective^a)/h	ca. 12 (8)	ca. 9 ½ (3)
Photons^b/mol
1st day	4.3	2.3
2nd day	6.3	4.2
Total (effective^c)/mol	10.5 (7.4)	6.5 (2.3)
Conversion^d (%)	83	>95
Isolated yield^e (%)	54 (65^f)	79

---

The first test run was performed with 2.0 g of diol (3) and 0.1 g of rose bengal in 200 ml of solvent. The starting material was readily consumed and after 8 h, the conversion has already reached a constant value of 83%. During that period the reactor collected 7.4 mol of photons between 500–600 nm.‡ After work-up, the desired product (4) was obtained in 54% yield (65% based on conversion of 3).

For the second experiment, the amount of diol (3) was reduced to 1.0 g in order to achieve complete conversion. After 3 h (Fig. 5), GC-analysis revealed that most of the starting diol (3) had already been consumed. At this stage the collector has received 2.3 mol of photons in the range of 500–600 nm.‡ After a total illumination period of ca. 9.5 h, Juglone was isolated in a compared to the laboratory experiments improved yield of 79%.

Fig. 5 Direct normal irradiance and product composition vs. illumination time for the photooxygenation of 1,5-dihydroxynaphthalene 3 (Experiment IV).

Although the photooxygenations of 3 required relatively long illumination times in comparison to the large scale experiments involving citronellol (1), it must be taken into account that solar-chemical experiments with non-concentrated sunlight often require several days or weeks to reach high conversion rates.¹⁶ Additionally, the given laboratory-scale reactor and especially the oxygen feeding equipment were not optimized. Thus, the preliminary results obtained from the present study clearly indicate that the photosensitized oxygenation of 1,5-dihydroxynaphthalene (3)—especially if performed with the PROPHIS system—opens a promising and mild alternative pathway to Juglone (4).

Conclusion

The present study on solar photooxygenations nicely demonstrates that the solar-chemical production of specific fine chemicals can serve as a powerful and environmentally friendly alternative to existing thermal processes. Cost estimates for specific fine chemicals (rose oxide¹² or ε-caprolactam¹⁷) furthermore revealed that an industrial solar-chemical production can indeed operate economically. Although not all conditions were optimal in the present study to realize Giacomo Ciamician's vision of the ‘Photochemistry of the Future’,¹⁸ a solar production plant in sunnier regions (e.g. in southern Spain) seems within range.

Acknowledgements

This research project was financially supported by the Arbeitsgemeinschaft Solar Nordrhein-Westfalen (Themenfeld 3: Solare Chemie und Solare Materialuntersuchungen). The authors would like to thank Prof. Axel G. Griesbeck and Dr. Christian Sattler for their support, and Dr. Jens Bunte for help in the preparation of this manuscript.

References

1. (a) P. Tundo, P. Anastas, D. StC. Black, J. Breen, T. Collins, S. Memoli, J. Miyamoto, M. Poliakoff and W. Tumas, Pure Appl. Chem., 2000, 72, 1207; (b) P. T. Anastas and J. C. Wagner, Green Chemistry. Theory and Practice, Oxford University Press, Oxford, 1998.
2. (a) A. Albini and M. Fagnoni, Green Chem., 2004, 6, 1; (b) J. Mattay, Chem. unserer Zeit, 2002, 36, 98; (c) A. Albini, M. Fagnoni and M. Mella, Pure Appl. Chem., 2000, 72, 1321; (d) K.-H. Funken, Sol. Energy Mater., 1991, 24, 370.
3. (a) K. Gollnick, Chim. Ind. (Milan), 1982, 64, 156; (b) A. D. Clements, Chem. Br. (London), 1980, 16, 464; (c) M. Fischer, Angew. Chem., Int. Ed. Engl., 1978, 17, 16.
4. (a) This concept leads back to the beginnings of organic photochemistry in the late 19th century where sunlight was the only available source of radiation. See: H. D. Roth, Pure Appl. Chem., 2001, 73, 395; (b) H. D. Roth, Angew. Chem., Int. Ed. Engl., 1989, 28, 1193.
5. (a) K.-H. Funken, G. Schneider, E. P. Esser, H.-D. Scharf, P. Esser and I. Wöhrle, in Proc., 6th Int. Symp. on Solar Thermal Concentrating Technologies, Ser. Ponencias, Editorial Ciemat, Madrid, 1993, vol. 2, p. 1027; (b) P. Esser, B. Pohlmann and H.-D. Scharf, Angew. Chem., Int. Ed. Engl., 1994, 33, 2009; (c) B. Pohlmann, H.-D. Scharf, U. Jarolimek and P. Mauermann, Sol. Energy, 1997, 61, 159; (d) C. Schiel, M. Oelgemöller, J. Ortner and J. Mattay, Green Chem., 2001, 3, 224.
6. K.-H. Funken and M. Becker, Renewable Energy, 2001, 24, 469.
7. (a) For overviews on different reactor types see: S. Malato, J. Blanco, A. Vidal and C. Richter, Appl. Catal. B: Environ., 2002, 37, 1; (b) K.-H. Funken and J. Ortner, Z. Phys. Chem., 1999, 213, 99.
8. J. Ortner, K.-H. Funken, E. Lüpfert, F. Plötz and K.-J. Riffelmann, in Proc., 9. Internationales Sonnenforum 1994 – Stuttgart, DGS-Sonnenenergie Verlags-GmbH, München, 1994, p. 1336.
9. (a) J. Ortner, D. Faust, K.-H. Funken, T. Lindner, J. Schulat, C. G. Stojanoff and P. Fröning, J. Phys. IV (France), 1999, 9, Pr3-379; (b) J. Ortner, D. Faust and K.-H. Funken, in Solare Chemie und solare Materialforschung, ed. M. Becker and K.-H. Funken, C. F. Müller Verlag, Heidelberg, 1997, p. 254.
10. (a) Singlet Oxygen, ed. A. A. Frimer, CRC Press, Boca Raton, FL, 1985; (b) Organic Chemistry, Singlet Oxygen, ed. H. H. Wasserman and R. W. Murray, Academic Press, New York, vol. 40, 1979.
11. (a) W. Rojahn and H.-U. Warnecke, DRAGOCO-Report, 1980, 27, 159; (b) G. Ohloff, E. Klein and G. O. Schenck, Angew. Chem., 1961, 73, 578; (c) H. Henning, A. Kokorakis and H.-D. Scharf, J. Inf. Rec., 1996, 23, 139.
12. N. Monnerie and J. Ortner, J. Sol. Energy Eng., 2001, 123, 171.
13. (a) For rac-Juglomycin A, see: G. A. Kraus and P. Liu, Synth. Commun., 1996, 26, 4501; (b) For Urdamycinone B, see: G. Matsuo, Y. Miki, M. Nakata, S. Matsumura and K. Toshima, Chem. Commun., 1996, 225; (c) For rac-Frenolicin B, see: P. Contant, M. Haess, J. Riegl, M. Scalone and M. Visnick, Synthesis, 1999, 821.
14. (a) G. Wurm and U. Geres, Arch. Pharm. (Weinheim), 1985, 318, 931; (b) H.-J. Durchstein and G. Wurm, Arch. Pharm. (Weinheim), 1984, 317, 809; (c) J. Griffiths, K.-Y. Chu and C. Hawkins, J. Chem. Soc., Chem. Commun., 1976, 676.
15. (a) For overviews on thermal preparations of Juglone see: D. J. Crouse, M. M. Wheeler, M. Goemann, P. S. Tobin, S. K. Basu and D. M. S. Wheeler, J. Org. Chem., 1981, 46, 1814; (b) C. Grundmann, Synthesis, 1977, 644.
16. (a) For some selected examples see: J.-T. Li, J.-H. Yang, J.-F. Han and T.-S. Li, Green Chem., 2003, 5, 433; (b) T. Rüther, A. M. Bond and W. R. Jackson, Green Chem., 2003, 5, 364; (c) Y. A. Jackson and K.-S. C. Marriott, Heterocycles, 2002, 57, 897; (d) H. Mallesha, N. D. Dinesh, K. S. Rangappa, S. Shashikanth, M. K. Lokanath, M. A. Sridhar and J. Shashidhara Prasad, Ind. J. Chem., 2002, 41B, 196; (e) N. W. A. Geraghty and J. J. Hannan, Tetrahedron Lett., 2001, 42, 3211.
17. (a) K.-H. Funken, F.-J. Müller, J. Ortner, K.-J. Riffelmann and C. Sattler, Energy, 1999, 24, 681; (b) C. Sattler, F.-J. Müller, K.-J. Riffelmann, J. Ortner and K.-H. Funken, J. Phys. IV (France), 1999, 9, Pr3-723.
18. G. Ciamician, Science, 1912, 36, 385.
19. T. Walsch and K.-J. Riffelmann, SEDES for Windows, v. 2.5, DLR, Cologne, 1997.

---

Footnotes

† Scharf and co-workers have studied the photooxygenation of furfural and reported that methylene blue is preferable to rose bengal.^5b However, this is mainly due to the weakly acidic condition in that particular case. Although we observed a partial decomposition of the sensitizer during our studies, the excess amount of rose bengal and the relatively short illumination time did not require additional feeding.

‡ The amounts of collected photons were calculated using SEDES for Windows.¹⁹

---