Regiospecific Photochemical Synthesis of Methylchrysenes

Methylated polycyclic aromatic hydrocarbons (PAHs) are suspected to be some of the toxic compounds in crude oil towards marine life and are needed as single compounds for environmental studies. 1-, 3- and 6-methylchrysene (3a,b,c) were prepared as single isomers by photochemical cyclization of the corresponding stilbenoids in the Mallory reaction using stoichiometric amounts of iodine in 82-88% yield. 2-methylchrysene (3d) was prepared by photochemical cyclization where the regioselectivity was controlled by elimination of an ortho-methoxy group under acidic oxygen free conditions in 72% yield. These conditions failed to form 4-methylchrysene from the corresponding stilbenoid. All stilbenoids were made from a common naphthyl Wittig salt and suitably substituted benzaldehydes. We have also demonstrated that methylchrysenes can be oxidized to the corresponding chrysenecarboxylic acids by KMnO4 in modest yields.


Introduction
Polycyclic aromatic hydrocarbons (PAHs) are a group of pollutants of great concern, particularly to the aquatic environment [1,2]. Petrogenic PAHs in nature typically originate from industrial or urban effluents, manmade accidents, and discharge of produced water from offshore oil production [1]. Contrary to pyrogenic PAHs, petrogenic PAHs have a large content of alkylation [3]. The concentration of monomethylated chrysenes is typically 10 times higher than chrysene in crude oil [3,4]. Some species, like Atlantic Haddock, subject to commercially important fisheries, are very sensitive to oil pollution at the egg stage [5]. The toxic effects of PAHs are often caused by their metabolites, and the position of alkylation have impact on these effects [1,6,7]. Alkylation on small PAHs makes them more potent agonists than the mother compounds toward aryl hydrocarbon receptors (AHR receptors) that regulate the PAH metabolism [8].
Further studies on the effect of alkylated PAHs require pure single compounds to elucidate these effects in exposure studies, and the compounds made in this work have already contributed to understanding some of the effects of methylation [7][8][9]. When the work described in this paper began, methylated chrysenes where only available as expensive analytical standards is small amounts, but not in the 0.1-0.5 g quantities desired for various environmental exposure studies.
Substituted chrysenes have been made in a variety of methods like the Diels-Alder reaction [10] and intra-molecular Pd-catalyzed C-H activation [11], but most common is photochemical oxidative cyclization, also known as the Mallory reaction [12,13]. The oxidative photocyclization of stilbenes is catalyzed by iodine, and typically air is bubbled through the solution during irradiation [14]. With the extensive studies of this reaction through the solution during irradiation [14]. With the extensive studies of this reaction in the 1960-1980′s one might expect all methylated chrysenes to have been made this way before.

Photochemical Cyclization Using Stoichiometric Amount of I 2
To obtain a single isomer of a substituted chrysene with the Mallory reaction [12,13] we need both ortho-positions to be identical by symmetry, like in the synthesis of 3b and 3c (Scheme 1c,d), or one ortho-position blocked like in the synthesis of 3a (Scheme 1d). A substituent in meta-position will give two isomers as both ortho-positions are available for reaction but giving different products [25].
The stilbenes needed for the photocyclization are readily available through a Wittig reaction. Wittig salt 1a (Scheme 2) was made by refluxing triphenylphosphine and 1-(chloromethyl)naphthalene in toluene. The product formed a precipitate that was washed with diethyl ether to obtain the pure product in 88% yield. The following Wittig reaction can be performed using an array of different bases. We preferred using a two-phase reaction with 50% aqueous NaOH in dichloromethane [26] at room temperature for practical reasons. Reaction with the suitable benzaldehydes gave stilbenoids 2a and 2b in high yields (Scheme 2). The E/Z-ratio of the stilbenoids have no consequence for the following photocyclization as the double bond isomerize in the process. Close inspection of NMR spectra sometimes allowed determination of the ratio which is then given in the experimental section. We found that the coupling constants for the double bond were about 12 Hz for Z-configuration and 15-16 Hz for E-configuration. This matches the reported coupling constants of 15.9 Hz (E) and 12.1 Hz (Z) for styrylnaphthalene [27]. The Wittig-reaction were less reactive with acetophenone giving stilbenoid 2c (Scheme 2). After 2 days we achieved only 52% yield. The more reactive Wittig-Horner reagent 1b gave 2c in 81% yield upon reflux in THF with potassium tert-butoxide as a base.

Photochemical Cyclization under Eliminative Conditions
A synthesis route like in Scheme 2 with a meta-substituted stilbenoids would make a mixture of 2-and 4-methylchrysene (3d,f) that would be demanding to separate. Olsen and Pruett [29] attempted to control the regioselectivity with a bromine substituent in one ortho-position. This worked as a blocking group under regular I2, O2 conditions (Scheme 3a). Another approach was to eliminate Br in a basic environment without I2 nor O2 and control the regioselectivity this way. This basic elimination (KOtBu or KOMe in the Finally, the stilbenoids were subjected to the photochemical oxidation with stoichiometric amounts of iodine (Scheme 2). The reactions were followed by TLC, but the disappearing color of iodine in the reaction were also a good indication on completion of the reaction. The photochemical reactions were performed in a 400 W medium pressure mercury lamp in a quartz glass immersion well fitted with a Pyrex filter. The reactions were made in 3-13 mM solution depending on the amount of starting material. Purification with flash chromatography gave 3a-c in 82-88% yield, better than the results reported for 3b [20] and 5-methylchrysene [18] using catalytic amounts of I 2 in a batch reactor. Synthesis of 3a in a flow reactor gave a similar yield [21,22]. Recrystallization were performed to get melting points and ensure that the compounds intended for toxicology studies were as pure as possible. We intentionally did not synthesize 5-methylchrysene [18] because it was already commercially available, and it is known to be very carcinogenic [28], requiring more strict safety precautions than was available to us.

Photochemical Cyclization under Eliminative Conditions
A synthesis route like in Scheme 2 with a meta-substituted stilbenoids would make a mixture of 2-and 4-methylchrysene (3d,f) that would be demanding to separate. Olsen and Pruett [29] attempted to control the regioselectivity with a bromine substituent in one orthoposition. This worked as a blocking group under regular I 2 , O 2 conditions (Scheme 3a). Another approach was to eliminate Br in a basic environment without I 2 nor O 2 and control the regioselectivity this way. This basic elimination (KOtBu or KOMe in the corresponding alcohol) changed the regioselectivity some but were hampered with significant amounts of regular photocyclization on the unsubstituted ortho-position. Regular photochemical cyclization worked better and 1-bromo-4-methylphenanthrene was obtained, but with a significant amount of dehalogenation occurring after the photocyclization. Treating this mixture with LiAlH 4 gave however 4-methylphenanthrene in 65% yield. Another approach by Mallory and coworkers [30] used a methoxy group as a controlling group that is eliminated under acidic conditions (A few drops of H2SO4 in t-BuOH/benzene) in the I2-and O2-free photoreaction (Scheme 3b). They succeed to eliminate ortho-methoxy and form 2-methylphenanthrene from the corresponding stilbene in 74% yield. The 4-methylphenanthrene was formed in 53% yield together with 9% 1-methoxy-2-methylphenanthrene. The eliminative reaction was 2-4 times slower than the oxidative reaction, giving 30-175 h irradiation time.
Considering these options, we decided to try the elimination of a methoxy group, and follow the route outlined in Scheme 3b. Both approaches lacked commercially available starting materials, but the methoxy aldehydes were readily available by the Skattebøl ortho-formylation [31,32].
Formylation of 4-methylfenol, benefiting from the symmetry, gave only aldehyde 4a (Scheme 4). Aldehyde 4b was made from 2-methylfenol where only the desired position was available for formylation. After a simple methylation of the hydroxy group these compounds were subjected to the same Wittig reaction (Scheme 5) as the previous aldehydes. Substituting benzene in the original conditions with toluene, the solvent mixture was degassed by ultrasound under N2, and kept under a stream of N2 during Another approach by Mallory and coworkers [30] used a methoxy group as a controlling group that is eliminated under acidic conditions (A few drops of H 2 SO 4 in t-BuOH/benzene) in the I 2 -and O 2 -free photoreaction (Scheme 3b). They succeed to eliminate ortho-methoxy and form 2-methylphenanthrene from the corresponding stilbene in 74% yield. The 4-methylphenanthrene was formed in 53% yield together with 9% 1-methoxy-2-methylphenanthrene. The eliminative reaction was 2-4 times slower than the oxidative reaction, giving 30-175 h irradiation time.
Considering these options, we decided to try the elimination of a methoxy group, and follow the route outlined in Scheme 3b. Both approaches lacked commercially available starting materials, but the methoxy aldehydes were readily available by the Skattebøl ortho-formylation [31,32].
Formylation of 4-methylfenol, benefiting from the symmetry, gave only aldehyde 4a (Scheme 4). Aldehyde 4b was made from 2-methylfenol where only the desired position was available for formylation. After a simple methylation of the hydroxy group these compounds were subjected to the same Wittig reaction (Scheme 5) as the previous aldehydes. Substituting benzene in the original conditions with toluene, the solvent mixture was degassed by ultrasound under N 2 , and kept under a stream of N 2 during irradiation. After 40 h the starting material was consumed, and 2-methylchrysene (3d) could be isolated in 72% yield. Stilbenoid 2e (Scheme 5) was subjected to the same conditions. Here, the reaction was even slower and was stopped after 134 h with some remaining starting material. The product was isolated in 49% yield but turned out to be the oxidative product 3e. There was no trace of 4-methylchrysene (3f). As 4-methylphenanthrene could be formed this way, although in less amounts than 2-methylphenanthrene, this came as a surprise. Repeated photocyclization of 2e gave the same result. To make sure nothing was wrong with the procedure nor equipment we made the corresponding stilbene and obtained 4-methylphenanthrene in the same yield as reported [30]. Apparently, the steric hindrance in this reaction [30] increases so much from phenanthrene to chrysene that no 3f is formed. As mesyl groups are easily eliminated to form double bonds by base, an attempt on elimination by basic conditions was made. Aldehyde 4b was protected with a mesyl group (Scheme 4), and stilbenoids 2f made in the Wittig reaction (Scheme 5). Applying the same base system used for eliminative photocyclization with Br [29], 3 eq. KOtBu in tBuOH/toluene, 2f were irradiated under oxygen free conditions. Unfortunately, the starting material decomposed rather than forming a cyclized product, making a greenish color to the reaction. After 5 h the starting material was decomposed without forming any isolable products. As mesyl groups are easily eliminated to form double bonds by base, an attempt on elimination by basic conditions was made. Aldehyde 4b was protected with a mesyl group (Scheme 4), and stilbenoids 2f made in the Wittig reaction (Scheme 5). Applying the same base system used for eliminative photocyclization with Br [29], 3 eq. KOtBu in tBuOH/toluene, 2f were irradiated under oxygen free conditions. Unfortunately, the starting material decomposed rather than forming a cyclized product, making a greenish color to the reaction. After 5 h the starting material was decomposed without forming any isolable products. elimination by basic conditions was made. Aldehyde 4b was protected with a mesyl group (Scheme 4), and stilbenoids 2f made in the Wittig reaction (Scheme 5). Applying the same base system used for eliminative photocyclization with Br [29], 3 eq. KOtBu in tBuOH/toluene, 2f were irradiated under oxygen free conditions. Unfortunately, the starting material decomposed rather than forming a cyclized product, making a greenish color to the reaction. After 5 h the starting material was decomposed without forming any isolable products.

Direct Oxidation of Methylchrysene 3b
One expected metabolite from methylchrysenes is the corresponding carboxylic acids [9]. To provide reference material, experiments were conducted to oxidize 3b to the corresponding acid. Although toluene has been oxidized to benzoic acid in a wide range of ways, we were unable to find any description of direct oxidation of a methyl group on PAHs larger than naphthalene. Vogel [33] describes oxidation of several substituted toluenes with KMnO 4 , but also describes the oxidation of phenanthrene to biphenyl-2,2'dicarboxylic acid with hydrogen peroxide. A study on degradation of PAHs by KMnO 4 found the order of reactivity as benzo[a]pyrene > pyrene > phenanthrene > anthracene > fluoranthene > chrysene [34]. Chrysene being the more stable PAH towards oxidation of the ring system motivated us to attempt oxidation by KMnO 4 . KMnO 4 slowly degrades in water and several equivalents of reagent are needed [35]. Based on an oxidation procedure in pyridine/water [36], we were after several attempts able to isolate chrysene-3-carboxylic acid (6) in 25% yield, when all starting material was consumed (Scheme 6).

Direct Oxidation of Methylchrysene 3b
One expected metabolite from methylchrysenes is the corresponding carboxylic acids [9]. To provide reference material, experiments were conducted to oxidize 3b to the corresponding acid. Although toluene has been oxidized to benzoic acid in a wide range of ways, we were unable to find any description of direct oxidation of a methyl group on PAHs larger than naphthalene. Vogel [33] describes oxidation of several substituted toluenes with KMnO4, but also describes the oxidation of phenanthrene to biphenyl-2,2'dicarboxylic acid with hydrogen peroxide. A study on degradation of PAHs by KMnO4 found the order of reactivity as benzo[a]pyrene > pyrene > phenanthrene > anthracene > fluoranthene > chrysene [34]. Chrysene being the more stable PAH towards oxidation of the ring system motivated us to attempt oxidation by KMnO4. KMnO4 slowly degrades in water and several equivalents of reagent are needed [35]. Based on an oxidation procedure in pyridine/water [36], we were after several attempts able to isolate chrysene-3-carboxylic acid (6) in 25% yield, when all starting material was consumed (Scheme 6).

General Information
The photochemical reactions were performed in a 400 W medium pressure mercurylamp in a 2 L quartz immersion well reactor (reaction volume 1.2 L) fitted with a no. 3408 Pyrex glass filter sleeve supplied by Photochemical Reactors Ltd. Silica gel 60A C.C. 40-43 μ m from SDS were used for flash chromatography. Melting points were obtained in sealed capillary tubes on a Stuart Scientific melting point apparatus SMP3. NMR-spectra were measured on a Varian Mercury 300 MHz instrument with tetramethylsilane or solvent peak as internal reference (CDCl3: 0.0 ppm, 77.0 ppm; CD3OD: 3.31 ppm, 49.0 ppm). HRMS analyses were performed on an JMS T100 GC-AccuTOF TM EI-TOF from Jeol.

General Information
The photochemical reactions were performed in a 400 W medium pressure mercury-lamp in a 2 L quartz immersion well reactor (reaction volume 1.2 L) fitted with a no. 3408 Pyrex glass filter sleeve supplied by Photochemical Reactors Ltd. Silica gel 60A C.C. 40-43 µm from SDS were used for flash chromatography. Melting points were obtained in sealed capillary tubes on a Stuart Scientific melting point apparatus SMP3. NMR-spectra were measured on a Varian Mercury 300 MHz instrument with tetramethylsilane or solvent peak as internal reference (CDCl 3 : 0.0 ppm, 77.0 ppm; CD 3 OD: 3.31 ppm, 49.0 ppm). HRMS analyses were performed on an JMS T100 GC-AccuTOF TM EI-TOF from Jeol.
The NMR data were in accordance with those reported [39].

Synthesis of Stilbenes
General procedure for Wittig-reaction to Stilbenes Wittig-salt 1a (1.2 eq.) and desired aldehyde (1 eq.) in DCM (120 mL) and 50% aq. NaOH (12 mL) was vigorously stirred under N 2 -atmosphere at room temperature until the aldehyde was consumed (1-3 days). The mixture was washed with water (300 mL) and the water phase extracted with DCM (100 mL). The combined DCM-phases were dried with anhydrous MgSO 4 , concentrated under reduced pressure, and purified by flash chromatography (Eluent: Petroleum ether/Ethyl acetate: 19/1) to obtain a mixture of E/Zisomers as a viscous oil (94-99% yield). The oil was used in the following photo-cyclization without further purification.
The NMR-spectra of the mixtures were complicated and sometimes also containing a rotamer giving a mix of three NMR-species. NMR-spectra are provided in the supplementary materials. Spectra are partly tabulated for compounds when one isomer can be separated from the NMR-spectra.

Photochemical cyclization of stilbenes
The photoreactor was flushed with N 2 and loaded with stilbene 2a (3.660g, 14.98 mmol), I 2 (4.190 g, 16.49 mmol), 1,2-epoxybutane (21 mL, 247 mmol) and degassed toluene (1200 mL). The reaction was stirred under N 2 atmosphere until all was dissolved, and stirred under UV-irradiation for 11 hrs. The reaction mixture was reduced to half volume under reduced pressure and washed with sat. aq. Na 2 S 2 O 3 (400 mL). The water phase was extracted with ethyl acetate (250 mL) and the combined organic phases washed with brine (250 mL) and dried over anhydrous MgSO 4 . The product was isolated by flash chromatography (Petroleum ether/ethyl acetate: 19/1) to yield 3.09 g (85%) of 3a.
This product was recrystallized from heptane/chloroform (100/70 mL), and the remains once more with heptane/chloroform (50/15 mL) to yield together 2. The NMR spectra were in accordance with those reported by Lutnaes and Johansen [43].
Product from this batch and 2 pilot batches were recrystallized together from methanol/ chloroform, and the remains once more with methanol/chloroform to yield a total of 1. The NMR spectra were in accordance with those reported by Lutnaes and Johansen [43].
The product was recrystallized from methanol/chloroform in two rounds to yield 0. The NMR-spectra were in accordance with those reported by Lutnaes and Johansen [43]. 2-Methylchrysene (3d) Stilbene 2d (1.303 g, 4.750 mmol) was dissolved in a degassed 9:1 mixture of tbutanol/toluene (1200 mL) under N 2 in the photoreactor and added 4 drops of conc. H 2 SO 4 . The mixture was irradiated for 40 hrs (followed by TLC) and concentrated under reduced pressure. The crude was purified by column chromatography (Petroleum ether/Ethyl acetate 9/1) to yield 0.828 g (72%) of 3d as a pale-yellow solid.
Melting Attempted photocyclization of stilbene 2f Stilbene 2f (0.532 g, 1.57 mmol) was dissolved in degassed t-Butanol/toluene (9:1, 300 mL) together with potassium t-butoxide (0.490 g, 4.37 mmol) under N 2 atmosphere. The mixture was irradiated for 5 hrs. The solution got a greenish cast to it, and starting material was decomposing without any product being formed.

Formylation of phenoles
2-Hydroxy-5-methylbenzaldehyde (4a) Following the description of Hansen and Skattebøl [31], water free MgCl 2 (2.973 g, 31.22 mmol) and paraformaldehyde (4.701 g, 156.5 mmol) was dissolved in dry THF (100 mL) under N 2 atmosphere. Triethylamine (10.5 mL) was added dropwise under stirring. After 10 min. 4-methylfenol (2.170 g, 20.06 mmol) was added dropwise. The mixture was refluxed in an oil bath at 75 • C for 1.5 h. After reaching room temperature the mixture was transferred to a separating funnel with diethyl ether (35 mL), and the organic phase washed with 1 M HCl (3 × 35 mL), water (2 × 35 mL) and brine (35 mL). The organic phase was dried over anhydrous MgSO 4 and concentrated under reduced pressure to afford 2.226 g (82%) of 4a as a tick oil that was used without further purification. 1 H NMR and 13 C NMR were in accordance with the description of Batt and Nayak [46]. 2-Hydroxy-3-methylbenzaldehyde (4b) Following the description of Hofsløkken and Skattebøl [32], water free MgCl 2 (2.956 g, 31.04 mmol) and paraformaldehyde (4.622 g, 155.0 mmol) was dissolved in dry THF (100 mL) under N 2 atmosphere. Triethylamine (10.5 mL) was added dropwise under stirring. After 10 min. 2-methylphenol (1.995 g, 18.44 mmol) was added. The mixture was refluxed in an oil bath at 75 • C for 1.5 h. After reaching room temperature the mixture was transferred to a separation funnel with diethyl ether (35 mL), and the organic phase was washed with 1M HCl (3 × 35 mL), water (2 × 35 mL) and brine (35 mL) The organic phase was dried over anhydrous MgSO 4 and concentrated under reduced pressure to afford 2.456 g (98%) of 4b as a tick oil that was used without further purification. 1 H NMR and 13 C NMR were in accordance with the description by Aspinall et al. [47].