The influence of the support on the singlet oxygen quantum yields of porphyrin supported photosensitizers

A halogenated porphyrin was covalently supported on aminoalkylated silica and aminoalkylated Merrifield polymer through a chlorosulfonation reaction. The supported porphyrin was tested for the photoxidation of α-terpinene. The kinetics of photocatalysis were evaluated using 9,10-dimethylanthracene as target and singlet oxygen formation quantum yield was calculated. The influence of the support, silica, Merrifield and other polymers, on the quenching of singlet oxygen were also evaluated.


Introduction
Singlet oxygen mediated oxidations are very attractive reactions which make use of clean, accessible and economical reaction conditions. 1,2They can give products which can hardly be obtained by other methods, for example endoperoxides 3,4 and allylic hydroperoxides. 5,6The need for a photosensitizer molecule to transfer energy through oxygen to organic molecules, which sometimes have low stability and thus that must be separated from the products, can be viewed as a limitation of the process.][9][10] However, supported photosensitizers also have weaknesses that affect the performance of the oxidation reaction such as: (1)  leaching the photosensitizers into the solution phase if sensitive bonds exist between the support and photosensitizer, (2) diminished activity of the photosensensitizer when attached to a support and (3) the accessibility of oxygen or substrates to photosensitizers can be affected by the support structure.Another important factor that can limit the action of supported photosensitizers is the possibility of fast quenching of the generated singlet oxygen by the support structure due to the presence and proximity of quenching groups.Porphyrins are good photosensitizers to generate singlet oxygen due to their photochemical characteristics and stability in reaction media. 11A porphyrin macrocycle can incorporate several functional groups that can modulate solubility properties and thus it is possible to carry out reactions in different solvents.Particularly important is the efficiency of porphyrins to generate singlet oxygen due to the heavy atom effect 12 as well as their stability in reaction media. 13n this work we present the synthesis of new supported photosensitizers (SPS) based on chlorinated porphyrins, and their efficiency as photooxidation catalysts.We also studied the ability of these SPS in generating singlet oxygen relative to free chlorinated porphyrins, highlighting the quenching effects caused by the support that can influence the efficiency of these supported catalysts.

Results and Discussion
In our recent studies, we developed a straightforward method for supporting chlorinated porphyrins on a modified Merrifield polymer using different lengths of spacers between porphyrin and polymer backbone. 10It was concluded that the Merrifield supported photosensitizers with a C12-chain spacer SPS-1 showed better efficiency than those with a C6-chain spacer SPS-2 possibly because the photosensitizer experiences a more homogeneous environment, Chart 1. 10,14 Merrifield Polymer In the present study, we extended the aforementioned strategy and prepared a supported catalyst containing a halogenated porphyrin covalently bonded to silica through a long C12 carbon chain, SPS-3, Scheme 1.

Scheme 1
The silica surface was modified by incorporation of amino groups (Silica-NH2) through a glycidyl bridge resulting from reaction between (3glycidyloxypropyl)trimethoxysilane, 1,12-dodecyldiamine and activated silica. 15The covalent linkage between the porphyrin and the amino groups of the modified silica was achieved by reaction with the chlorosulfonated porphyrin P2, which was easily obtained by chlorosulfonation of porphyrin P1.The presence of porphyrin bonded to silica was confirmed by characteristics bands in the infrared spectra of SPS-3.In particularly, the weak bands at 1558 and 1429 cm -1 (for the free porphyrin, at 1557 and 1428 cm -1 respectively) can be assigned to the carbon-carbon double bond stretch and carbonhydrogen bending. 16hotocatalyst SPS-1 showed low values for loading of supported porphyrin (0.040 mmol/g).With the objective of preparing a catalyst with a higher loading, we followed the same procedure for supporting the chlorinated porphyrin but using a commercial Merrifield resin with higher initial loading of benzyl chloride groups.As expected, we obtained a supported photosensitizer SPS-4 (see experimental) with three times more supported porphyrin (loading of 0.11 mmol/g).Following the same methodology the chlorosulfonated porphyrin P2 was reacted with commercial aminated polystyrene divinylbenzene copolymer (PSDV-NH2) to obtain a supported photosensitizer with no spacer between the polymer and the porphyrin (SPS-5).All supported photosensitizers (SPS-1 to SPS-5) were tested in the photooxidation of α-terpinene (1) to produce ascaridole (2) as major product and p-cymene (3) from α-terpinene aromatization.For comparison purposes, the result of the photooxidation with the free porphyrin (TDCPPH2) are also presented (Table 1).a Catalyst/substrate molar ratio of 1/5000.b Determined by the NMR integration of the reaction mixture after removal of the polymer by filtration, and evaporation of solvent.The number in parentheses is the amount of p-cymene (3).
The results in Table 1 show that all the supported photosensitizers are less active than the free porphyrin.SPS-1, with the longer spacer between the porphyrin and the polymer structure, is the most active and selective supported photosensitizer.The absence of a spacer is not a favourable situation as indicated by the results obtained using SPS-5.On the other hand, the presence of a longer spacer alone is not sufficient to form an active catalyst.Silica supported porphyrin SPS-3, with a C12 spacer, does not give the same results as SPS-1 (which has the same spacer).This observation points to some influence of the support on catalyst efficiency.Moreover, a deactivating effect of the support is seen in the case of SPS-4 which, in spite of having a higher loading of porphyrin, is the least active catalyst.
In fact, supports give heterogeneity to the catalysts and help the catalyst recovery, for further use.However the support may interfere with the oxidation process particularly on the amount of 1 O2 generated.The incorporation of the photosensitizer on the support can bring some difficulties to the efficiency of the oxidation due to problems caused in 1 O2 photogeneration related to molecular distortions, 17,18 dye aggregation 19 or 1 O2 deactivation. 20To assess whether the support has any influence on singlet oxygen generation, singlet oxygen quantum yields for the different photocatalysts (SPS-1 to SPS-5) were measured by steady state photolysis using dimethylanthracene (DMA) as target and TDCPPH2 as reference, comparing the slopes of the semilogarithmic plots of the decay of the absorption of the quencher at 380 nm with time (ln(A0/A) vs time) (Figure 1), with the corresponding slope obtained for the reference.The solutions were prepared with the same concentration of DMA and the same amount of porphyrins taking into account the particular loading of each supported photosensitizers.To validate our methodology, a set of two experiments using the TDCPPH2 as reference were carried out daily.The analysis of all the data obtained on different days gave good agreement (slope 0.057 ±0.001, r 2 = 0.992).Considering the first 10 minutes, the rate of quenching of DMA is higher for TDCPPH2 and SPS-3 than for the other photocatalysts.Only in these two cases is the difference in the measured absorption of DMA sufficient to allow calculation of the rate constants.Extending the reaction time, significant differences were observed in the case of the other photocatalysts.The rate of DMA quenching using SPS-4 and SPS-5 could be obtained by linear fitting of the experimental data between 0 and 70 minutes.In the case of SPS-1 and SPS-2 the linear fit is not applicable to the complete reaction time, but clearly there are two different rates: in the first 30 min, the reaction rate is slow (slopes of 0.006 and 0.004 for the respective liner fitting) and after that there is a significant increase of reaction rate (slopes of 0.016 and 0.010 for the respective linear fitting).This indicates that these Merrifield supported photosensitizers have an induction time of ca. 30 minutes that could be related to the wetting process (diffusion of the solvent with the dissolved oxygen and the DMA into the polymer).After this period of time, the efficiency of production of singlet oxygen increases.The singlet oxygen formation quantum yields are calculated using the slopes after 30 minutes for SPS-1 and SPS-2 and the slopes of the total reaction time for the other supported catalysts (Table 2).All the supported photosensitizers present low yields for singlet oxygen formation and these values are significantly lower than the value for TDCPPH2 (ФΔ =1), ranging from 4 times less for SPS-1 to 30 times less for SPS-4.In a similar case, supported Rose Bengal on a poly(ethylene glycol) polymer presented a smaller decrease for the heterogeneous form relative to the homogeneous one (0.75 to 0.43, respectively). 21Thus, it seems reasonable to assume that in the present case the polymer has a great influence on singlet oxygen generation.An analysis of the results in Tables 1 and 2 shows that the values of ФΔ for the supported photosensitizers correlated well with the kinetics observed for the photooxidation of α-terpinene, the most efficient supported photosensitizers have the higher singlet oxygen quantum yields.Concerning the Merrifield supported photosensitizers, SPS-1 with the longer spacer has the highest  and SPS 5 with no spacer between polymer and porphyrin has the lowest value.The result with SPS-4 confirms the low activity of this sensitizer in the photooxidation of α-terpinene.Although having a C12 spacer SPS-4 is not able to generate singlet oxygen to then react with DMA.This polymer has the same spacer with SPS-1 and a significant increase in the porphyrin loading.As the porphyrin is the same, the reason for this sluggish activity can be related to quenching effects of formed singlet oxygen or the quenching of the triplet porphyrin state promoted by the proximity of porphyrin groups to the polymer surface.Porphyrin quenching effects by increasing the concentration are well known. 22olymer structure has influence on the lifetime of 1 O2 if it is generated inside the polymer structure.This lifetime is considerably lower due to deactivation by the polymer structure particularly if a large amount of C-H oscillators exists. 23,24In the present situation it is possible that, since the singlet oxygen is generated near the surface of the polymers, the polymer structure may also influence the deactivation processes.To evaluate if there is some effect of polymer structure on the singlet oxygen generation we carried out experiments to analyze the variation on  of TDCPPH2 in the presence of the supports used.We also tested other materials that can be used as supports for photosensitizers like poly(ethyleneglycol), 8 ion exchange resin, 25 sulfonated polystyrene, 26 chitosan, 27 poly(ethyleneimine) and a Merrifield polymer linked to a 4,4'diaminobibenzyl spacer.Solutions of DMA and TDCPPH2 were prepared in the same concentration that the solution used as reference and 1 mg of polymers or modified polymers were added.From the same plot as shown in Figure 1 the  values were calculated (Table 3).The singlet oxygen quantum yield formation of TDCPPH2 in the presence of Merrifield, Merrifield-C12 or PSDV-NH2 is the same, within the error interval, as for TDCPPH2 in the absence of the support.In the case of silica and silica linked to the spacer C12 there is a small decrease of the singlet oxygen quantum yield that can be attributed to a quenching due to the OH groups in the silica. 28Chitosan also acts as a quencher to singlet oxygen.Poly(ethyleneimine), although having a great number of amino groups, is not a suitable support for these photosensitizers since it acts as a strong quencher of singlet oxygen.The reason may be related to the fact that it forms a homogeneous solution and the great quenching ability of amino groups. 29The influence of the nature of the spacer could be evaluated by comparing the quantum yields of the Merrifield with the biphenyl groups and Merrifield-C12; the presence of aromatic groups in the spacer decreases the quantum yield making them undesirable as supports in photooxidation reactions.The influence of the amount of support could be significant when considering higher quantities of polymer to be used.For example, in the case of Merrifield polymers that do not affect the singlet oxygen formation (Table 3) we verified a decrease of the singlet oxygen quantum yield from 0.96 to 0.83 when the amount of polymer was increased from 1 mg to 5 mg.

Conclusions
Halogenated porphyrins can be covalently linked to aminoalkylated silica or aminoalkylated Merrifield polymer through a chlorosulfonation/amidation reaction.The supported porphyrins were tested in the photoxidation of α-terpinene and the results correlated with the singlet oxygen formation quantum yields.Increasing the loading of the porphyrin incorporation does not increase the catalytic efficiency or the singlet oxygen quantum yield.With TDCPPH2 as photosensitizer, quenching of singlet oxygen by some heterogeneous supports were observed.

Synthesis of SPS-3
To a stirred solution of 80 ml dry toluene containing 3.0 mmol of 1,12-diaminododecane, 1.0 mmol of (3-glycidyloxypropyl)-trimethoxysilane was added.The resulting mixture was allowed to react at 80 ºC for 24 h.To this solution 1.5 g of activated SiO2 and 5 ml of EtOH were added and the stirred solution was maintained at 80 ºC for 24 h.The amino functionalized silica was isolated by filtration and washed with MeOH and EtOH.Then, ARKAT USA, Inc.
it was refluxed with EtOH during 1 h, then filtrated again and dried at 40 ºC for several days.Elemental analysis showed a nitrogen content of 0.87%.This aminoalkylated silica derivative (Silica-NH2) (380 mg) was reacted with the chlorosulfonyl derivative P2 as described for SPS-1. 10Elemental analysis showed a nitrogen content of 1.52%, which corresponds to a porphyrin loading of 0.12 mmol/g.

Synthesis of SPS-4
The procedure is similar to that used for SPS-1.First the aminoalkylated Merrifield polymer was prepared by mixing 3.0 g of Merrifield polymer (higher loading) in 25 mL of DMF, 1.5 g of 1,12-diaminododecane and keeping the mixture at 70 ºC for 24 h.After cooling, the mixture was poured into 150 mL of water, filtered, and washed with water, methanol, dichloromethane, and methanol again.The residue was dried in an oven under vacuum.Elemental analysis of this product gives a nitrogen content of 2.64%.This aminoalkylated polymer (380 mg) was made to react with the chlorosulfonyl derivative P2 as described for SPS-1.Elemental analysis showed a nitrogen content of 3.24%, which corresponds to a porphyrin loading of 0.11 mmol/g.

Synthesis of SPS-5
At room temperature, 15 mL of chlorosulfonic acid was added to 213 mg of porphyrin P1.The solution was stirred for 2 h and then carefully poured onto ice in order to precipitate the porphyrins.The precipitate was filtered, dried, dissolved in dichloromethane, and the solution dried with sodium sulfate.The solution was concentrated to 30 mL, 1 mL of pyridine was added followed by 300 mg of PSDV-NH2.The mixture was stirred for 2 days at 30 ºC, filtered, and washed with dichloromethane, tetrahydrofuran, methanol, and dichloromethane again.Non-bonded porphyrin was eliminated with these washings.After drying the solid under vacuum, elemental analysis showed a nitrogen content of 2.0%, which corresponds to a porphyrin loading of 0.22 mmol/g.

Synthesis of Merrifield polymer linked to 4,4'-diaminobiphenyl
To a mixture of 3.0 g of Merrifield polymer (lower loading) in 25 mL of DMF, 1.5 g of 4,4'diaminobiphenyl was added.The mixture was placed at room temperature for 24 h.After this time, the mixture was poured in 150 mL of water, filtered, and washed with water, methanol, dichloromethane, and methanol again.The residue was dried in an oven under vacuum.Elemental analysis showed a nitrogen content of 1.18%.

Photosynthetic oxidation experiments. General photooxidation procedure
Photooxidation experiments were carried out at room temperature using a laboratory-built photoreactor consisting of three 50 W lamps.The reactions were carried out in a flask equipped with a water condenser and an inlet for air.The solutions were irradiated with a stream of air continuously flowing into the flask.The substrate (α-terpinene, 4.9 mmol) in 65 mL of chloroform was mixed with the appropriate amount of photosensitizer (9.8×10 −4 mmol of porphyrin or the equivalent molar amount of supported porphyrin) to produce a 1/5000 molar ratio of sensitizer to substrate, and 203 mg of sodium hydrogen carbonate was added.The progress of the reaction was monitored by monitoring the disappearance of the reagent by GC.The photosensitizer was collected by filtration, and the product was obtained by evaporation of the solvent and analysed by 1 H NMR measurements. 32,33

Photokinetic oxidation experiments. General procedure
The experiments were carried out at room temperature using a laboratory-built photoreactor consisting of one 50 W lamp equipped with water filter (radiation intensity 0.75 mW/cm 2 ).The reactions were carried out in a quartz cell equipped with a magnetic stirrer and an air inlet.The solutions were irradiated with a stream of air continuously flowing into the cell.
Experiments of photooxidation of 9,10-dimethylanthracene (DMA) with TDCPPH2 1.5 ml of DMA solution (1x10 -4 M) was mixed with 0.5 ml of a solution of TDCPPH2 (2x10 -5 M), 1 ml of chloroform.The progress of the reaction was monitored by UV-vis spectroscopy at 380 nm by the disappearance of the DMA absorbance.

Photooxidation experiments of DMA with supported photosensitizers (Figure 1 and table 2)
1.5ml of DMA solution (1x10 -4 M) was mixed with 1.5 ml of chloroform and the appropriate amount of supporting photosensitizer, taking into account the respective loading.The mixture was kept with stirring in the dark for 15 min.Then, the mixture was irradiated and the progress of the reaction was monitored by UV-vis spectroscopy at 380 nm, by the disappearance of DMA absorbance.
Photooxidation experiments of DMA with TDCPPH2 (as sensitizer) in presence of bulk polymers (Table 3) 1.5 ml of DMA solution (1x10 -4 M) was mixed with 0.5 ml of a solution of TDCPPH2 (2x10 -5 M), 1 ml of chloroform and 1 mg of different polymers.The mixture was kept in the dark for 1 min.Then, the mixture was irradiated and the progress of the reaction was monitored by UV-vis spectroscopy at 380 nm by disappearance of DMA absorbance.
Photocatalyst preparation

Figure 1 .
Figure 1.Semilogarithmic plots of DMA quenching using photocatalysts SPS-1 to SPS-5.The zoom for SPS-3 and TDCPPH2 is depicted in the inset.

Table 3 .
Values for singlet oxygen quantum yields for TDCPPH2 in the presence of 1 mg of different supports