Biomass is "any material having recent biological origin, including plant materials, agricultural crops, and even animal manure" (2). Biomass is recognized as being an abundant carbon and energy source. It is renewable and therefore sustainable. Biomass also has the potential to produce less greenhouse gas emissions. In addition to the practical qualities biomass has to offer, it is important to recognize that it has "excess functionality" such as oxygen-containing groups, in addition to carbon and hydrogen. These qualities make it an excellent source to serve as fuels and chemicals (1, 2). Currently, only 5 % of chemicals are produced from renewable resources (2). The successful conversion of biomass compounds could lead to the production of a wide variety of products (1). It has been shown that there is an increased interest in bio-based products. In a matter of only eleven years, the sales of biotechnology products jumped 5,600 million dollars worldwide (3). By 2090 the United States bio-based industry aims to have bio-based production greater than 90 % for organic chemicals, up to 50 % for liquid fuels, and 99 % for materials such as bioplastics (3). These goals are realistic and within reach if the proper attention is paid to biomass.
1.2 Biorefineries and Lignin
We can make use of biomass as an energy and fuel source by way of a bio-refinery. A bio-refinery is used to produce fuels, power, and chemicals from biomass by incorporating conversion processes and equipment in one facility (3). There are a variety of biorefineries that process biomass in different ways. They range from a "whole-crop" biorefinery to an "integrated" biorefinery. Lignin, which is found in great abundance in biomass, is processed through a Lignocellulose Feedstock (LCF) biorefinery. With considerable amounts present in biomass, lignins have received a lot of attention due to their structural diversity and the potential of their conversion to useful chemicals. Lignin consists of hydroxyl- and methoxy- substituted aromatic units in a branched polyphenolic macromolecule as illustrated in Figure 1 (4).
Figure 1. Typical fragments of lignin molecules with multiple functional groups (5).
Because lignin is a plant material, it is broken down using enzymes, chemicals or temperature. At a LCF biorefinery, the raw plant material is initially cleaned. Using chemical digestion or enzymatic hydrolysis, the plant material is then broken down into lower molecular weight organic compounds. The current uses for lignin include adhesives and binders, as well as a fuel for direct combustion. Lignin has enormous potential to contribute to producing multiple substituted aromatic hydrocarbons. Accomplishing this in a cost-effective and ecologically aware manner would be very useful for generating interest in the LCF process. There are a variety of classes of organic compounds present in significant amounts in bio-oil. These compounds include acids, alcohols, phenols (guaiacols, syringols), aldehydes, ketones, and sugars. For example, the analysis of wood biomass shows that it commonly produces 30 % water, 30 % phenolic compounds, 20 % aldehydes and ketones, 15 % alcohols, and 10 % miscellaneous compounds; a mixture that is common, depending on the properties of the feedstock (3).
1.3 Platform Molecules
The conversion technologies for biomass are grouped into "platforms." Lignin gives rise to a very large group of organic molecules referred to as "platform molecules." There are five different platforms which include sugar, thermo chemical, biogas, carbon-rich chains, and plant products. The purpose of creating platforms is so that industry can make a variety of fuels, chemicals, and materials from platform chemicals, and generate power (3). Platform molecules can be used to produce many usable compounds. Ideally, platform molecules are multifunctional compounds that are inexpensive and produce little byproducts with high purity using conventional chemistry.
Guaiacols, a group of phenols typically found in bio-oil, include eugenol and isoeugenol. Guaiacol is derived from guaiacum or wood creosote plants. It is often used as a flavoring (6). Isoeugenol and eugenol stand out from the other compounds in this group because of their high concentration in bio-oil as compared to the other phenolic compounds (3). Isoeugenol is a monomer that is associated with the biosynthesis of lignin (7). Its abundance in bio-oil is greater than 7 % which makes it readily available and convenient to use (3). Isoeugenol, which was studied in the past, has been utilized as a source of different products through a variety of techniques. Isoeugenol, an isomer of eugenol, can be extracted as a component of essential oils from sources such as the ylang-ylang tree, gardenia, and nutmeg. It is often used in fragrances, but also in beverages, chewing gum, and baked goods (8). Eugenol is a component of natural oil that can be extracted from a variety of natural herbs including cloves, nutmeg, and cinnamon. It is often used in flavorings and essential oil (9, 10).
Vanillin and vanillic acid are natural products that can be extracted from vanilla. Vanillin is used as a food flavor and also in cosmetics and drugs. Vanillin has been shown to produce multiple products under oxidation, such as vanillic acid and divanillin (11). Salicylic acid can be found in several plants, particularly in wintergreen leaves and the bark of sweet birch trees (12). It can also absorb on TiO2 from solution (13). These compounds are all naturally derived from plants, and they are closely related in their origin and molecular structure. Many of them have been studied in the past, but not by way of heterogeneous photocatalysis. Although they can all be extracted from biomass lignin, they can also be made synthetically. In this research, all of the compounds were purchased from Sigma-Aldrich and used as they were received.
Chemical identity of these products was verified using common experimental techniques such as thin layer chromatography (TLC), high performance liquid chromatography (HPLC), infrared spectroscopy (IR), and mass spectrometry, (MS) (14). Other researchers have found that, under photooxidation, isoeugenol produces "dimeric" products that are structurally related to the parent lignin structure. Five different solvents were used that lead to a variety of products including dehydrodiisoeugenol (DHDIE) (15). Another technique other researchers used to produce vanillin from isoeugenol utilized atmospheric pressure, room temperature, a light source, and in contrast to our research, an organic photosensitizer dye. It was found that the yield of vanillin was related to the concentration of the substrate, isoeugenol (16). Eugenol and guaiacol at low concentrations have been removed from paper mill wastewater pollutants by photocatalytic oxidation (17). Another method of decomposition of lignin is using catalytic oxidation in alkaline media to yield vanillin (18).
1.4 Photocatalysis and Photochemistry
In order to contribute to the "green" chemistry community, the goal of this research is to use sustainable methods to produce usable products (lignin-based molecules) from biomass. It is clear that oxidation is what connects isoeugenol to vanillin. Therefore, it is important to understand the differences between photocatalytic oxidation and photochemical oxidation of those substances, and how each process produces different usable products from isoeugenol and vanillin. Isomerization and oxidation are frequently used with metal oxide catalysts to process biomass in the liquid phase at high temperatures. Catalytic oxidation of carbohydrate-derived feedstocks has been proven to produce multiple products (1). Thus appreciating chemistry fundamentals behind photocatalytic and photochemical oxidation processes is imperative for the further development of emerging technologies of "green chemistry".
1.4.1 Photocatalytic Oxidation
Photocatalysis can be applied to multiple processes including oxidation, polymerizations, and substitutions (19). Photocatalysis incorporates advanced oxidation processes with solar and chemical forms of energy to produce a very reactive radical species. Heterogeneous photocatalysis indicates that the catalyst and reactant are in a different phase, i.e. solid and liquid. Heterogeneous photocatalysis increases the rate of oxidation using a semiconductor catalyst (19), via excitation of an electron from valence to conduction bands. Titanium dioxide (TiO2) is known to be "the most widely used semiconductor photocatalyst" (20). Figure 2 is the energy-band diagram of titanium dioxide that includes the valence band and conduction band of the titanium dioxide photocatalyst. TiO2 has a band-gap energy of 3.1 eV (λ= 400 nm) (21).
Figure 2. Energy-band diagram of the photocatalyst, titanium dioxide (TiO2) (22).
Specifically, when a semiconductor photocatalyst such as TiO2 is exposed to light, the photon is absorbed by the photocatalyst, when energy of the photon is equal or greater than the band gap of the photocatalyst. Upon absorption of the photon, the photoexcitation of an electron from the valence band to the conduction band occurs in the semiconductor photocatalyst TiO2, Figure 3 (19). Specifically, an electron is found in the conduction band, and a hole, h+ is formed in the valence band (17).
Photoexcitation: TiO2 + hv e- + h+
Oxygen ionosorption: (O2)ads e- + O2 •_
Figure 3. An illustration of how photocatalysis works to form hole-pairs and radicals where ads stands for "adsorbed on the surface of the photocatalyst" (19).
In water media, oxidation of a water molecule by a hole, h+ yields a solvated proton, H+ and a hydroxyl radical, HO.. Titanium dioxide is known to promote the degradation of phenols through oxidation that includes the radical, HO. as the major reactive intermediate (23, 24). Choosing a catalyst is a crucial part of this process because much of the reaction relies on the performance of the catalyst. Titanium dioxide (TiO2) is frequently used because, in contrast to other catalysts, it is has a higher photocatalytic activity (19). In addition to having higher photocatalytic activity, TiO2 is readily available, inexpensive, and chemically stable (25). Using oxygen/air as the oxidant in photocatalytic oxidation is substantial and cost-effective in the chemical industry and in cleaning up industrial waste (17, 25).
The non-selective oxidation of organic substrates in water media is due to oxidation by holes that have a very high value of standard reduction potential at about 2.53 V versus the normal hydrogen electrode (NHE) (26). On the other hand, the oxygen radical anion, O2-. has a moderate reduction potential of the half-reaction O2 + e- > O2-. at 0.87 V versus the standard calomel electrode (SCE) in CH3CN (27). That would cause mild and selective oxidation of the organic substrate with O2 via the O2-. reactive intermediate (28) in CH3CN as the solvent, as well as in other non-protic solvents.
1.4.2 Photochemical Oxidation
Photochemical oxidation reactions involve a photon of light and an oxidant in the form of oxygen, either from ambient air or pure O2. Photochemical reactions rely on this photon for their activation energy. The molecule of the reactant, aka the substrate, directly absorbs the photon, and photoexcitation of the reactant molecule occurs (22). Figure 4 illustrates a general scheme of a photochemical reaction: M indicates the organic substrate that absorbs the photon, M* indicates the photoexcited form of the substrate M (usually, first excited singlet state S1), the oxidant could be, for example, a molecule of dissolved oxygen O2, and the products indicate molecular products of photooxidation reactions.
Figure 4. A general example of a photochemical reaction (22).
Chapter 2: Experimental
To perform the following experimental procedures, there are some safety precautions that should be followed. When using the 450 W quartz mercury vapor lamp it is required that one wears, at all times, light-protective goggles that are rated for >99 % UV protection. During high-pressure lab experiments, a protective face shield should be worn when working with pressurized glass vessels. In addition, light absorbing protective shields and implosion-resistant Plexiglass panels should be used for experiments under pressure.
Organic compounds from the plant-based platform group investigated in the Samokhvalov laboratory include vanillin, vanillic acid, salicylic acid, salicylaldehyde, salicyl alcohol, guaiacol, eugenol, and isoeugenol; structures of these compounds and their molecular weights are given in Figure 5.
Figure 5. Selected phenolic compounds commonly found in bio-oil.
A manageable list of nine organic molecules out of twelve compounds, were selected by trial and error. Such narrowing down of the research agenda allowed the project to focus on three main categories of phenolic compounds: i) phenols, such as isoeugenol and eugenol, ii) phenols with a carboxylic group, such as vanillic acid and salicylic acid, and iii) phenols with an aldehyde group, such as vanillin and salicylaldehyde.
2.3 Solvent Selection
Each of the selected phenolic compounds was dissolved in one of a variety of solvents, including acetonitrile, dibutyl ether, dimethylformamide (DMF), dimethylsulfoxide (DMSO), or perfluorodecalin (PFD). These solvents were selected and tested because of their differences in dipole moments or being listed as a "green solvent." To categorize solvents as green, yellow, or red, they are analyzed in the following areas: worker safety, process safety, and environmental and regulatory considerations (29). The least "green" solvents fall under the "red" category, those that are milder than "red" solvents are deemed "yellow," and the environmentally greenest solvents are labeled, obviously, "green."
Acetonitrile was the initial solvent choice, because it has a high dipole moment of 3.92 D and a high solubility of molecular oxygen under standard conditions. It is also considered a "yellow" solvent, and is a good alternative for "red" solvents such as dimethylformamide (29). Dibutyl ether was also tested; it has a rather low dipole moment of 1.18 D, but is considered a "green" solvent, as it is a good replacement for diethyl ether. Dimethylformamide (DMF) was used because it has a high dipole moment of 3.86 D; however it is considered a "red" solvent, as it has a high toxicity. Dimethyl sulfoxide (DMSO) was chosen because it has a high dipole moment of 3.96 D, and falls on the solvent selection guide as a "yellow" solvent. Lastly, in an attempt to try a solvent with no dipole moment, Perfluorodecalin (PFD) was chosen. Although it is not rated on the solvent selection scale, an approximate rating may be "yellow" or "green", based on published material safety data sheets.
The solvents were purchased from Sigma, Fisher, and Acros Organics. Each selected phenolic compound was dissolved in one of the solvents, discussed above, to make 0.5 M solutions ("stock solutions"). For the solid compounds, dissolution was facilitated using a sonicator (SRA Digital Ultrasonic Cleaner, UC-20D).
The photocatalytic experiments conducted utilized titanium dioxide Degussa P25 (recently re-branded as Evonik) from Acros Organics, which is a standard oxidation photocatalyst. It is an inexpensive white powder that performs well under UV light and has a low toxicity. Determining the amount of TiO2 to use is based on the fact that it has been reported that "higher titanium dioxide concentration could cause lower light penetration resulting in lower oxidation rates" (30). In these lab experiments 50 mg of titanium dioxide powder was added to every 10 ml of stock solution used in the photoreactor test tube.
The custom-built photo-reactor contains a light source, and reaction vessels in a constant temperature bath. The light source is a commercial 450 W medium pressure, quartz mercury-vapor lamp from Ace Glass, Inc. It is in the center of a commercial well platform set up for parallel reactions, with the capacity of eight test tubes to fit around it at once (31). The constant temperature bath is a commercial liquid thermostat from Neslab, EX221, and it is able to regulate the temperature between 30 and 100 °C. In these experiments, a mixture of water and ethylene glycol was used to maintain the temperature between 30 and 35 °C. The reaction vessels are either quartz or Pyrex test tubes from Ace Glass, Inc. Quartz vessels transmit light with λ ≥ 254 nm, while Pyrex vessels transmit light with λ ≥ 355 nm. Each reaction vessel is set up to receive ambient air with ca. 100 % humidity, at a flow rate of 25 ml per min. The size of the vessel is 20 cm in length and 32 mm in diameter.
Some photochemical experiments were conducted outside of the temperature bath at ambient temperature under pressure. In this experimental set-up, the mercury-vapor lamp was placed approximately three inches away from the pressurized vessel on a ring stand. The pressurized vessel was from Ace Glass, Inc. This vessel was secured on a ring stand over a stir plate. A pressure regulator was used to pressurize the vessel at approximately 100 psi of either air or dry oxygen. A stir bar was placed inside the reaction vessel.
2.6 Photoreaction Procedures
Concentrated solutions (0.5 M) of phenolic compounds in a chosen solvent were prepared. Either Pyrex or quartz test tubes were used in the photoreactor. Each reaction vessel was filled with 10 ml of the 0.5 M solution. For the experiments conducted without solvent, 10 g of the selected liquid phenolic compound was used for each reaction vessel. Usually, 50 mg of catalyst was added to each photocatalytic reaction vessel. This ratio was optimized by trial and error. It has been reported that "at high-substrate concentrations, the photonic efficiency (of organic substances) decreases and the titanium dioxide surface becomes saturated, thus leading to catalyst deactivation" (19). The reaction took place at a temperature of about 30 °C for the duration of 4 hours. In most cases, ambient air at ca. 100 % humidity was bubbled as the oxidant at a flow rate of 25 ml per minute. Up to eight, but usually four, parallel reactions took place simultaneously in the photoreactor, so multiple experiments could be run at once. Photochemical and photocatalytic experiments were often run together to compare the different reaction conditions. After four hours, the samples were collected for analysis. The samples containing titanium dioxide were centrifuged to settle the catalyst. A 13,000 rpm centrifuge (Fisher Scientific AccuSpin Micro) was used to isolate the clear supernatant. When all of the clear liquid products were isolated, a pipette was used to collect 0.25 ml from each sample. The sample was placed into a 25 ml volumetric flask, diluted with acetonitrile and prepared for analysis.
2.7 Product Analysis
A variety of techniques were used to obtain proof of reaction. Initially, Thin Layer Chromatography (TLC) was used to identify products of chemical reactions. Silica gel plates from Sigma Aldrich and a 60/40 mixture of hexane/ethyl acetate as the eluent were used to analyze the products. A UV light was passed over the plate to visualize the spots of compounds that fluoresce. Some spots were difficult to see with the UV light only, so vapor of iodine was used to develop the spots on the silica gel plates.
The retention times for HPLC of the stock solutions of reagents and reaction products were measured and compared. UV-visible spectroscopy (UV-Vis) was used as well to show the progress of chemical reactions. The UV-Vis spectrophotometer used for this analysis was a Cary Bio-Rad 50 in absorbance mode. Usually, 3 micro liters (μL) of liquid product were diluted with acetonitrile up to 3 ml for analysis. A quartz cuvette with l =1 cm was used in the instrument to analyze diluted solutions of products.
Fourier Transform Infrared (FTIR) spectroscopy was also used to find reaction products. Polytetrafluoroethylene (PTFE) slides, NaCl salt plates, and KBr pellets were used in the Bio-Rad Excalibur Series FTIR instrument for analysis. One method used to identify molecular products was high performance liquid chromatography with UV-Vis detection (HPLC-UV), with a C18 reverse phase (RP) column, with a 5 micron stationary phase. For each specimen at least three replicates were obtained. The HPLC-UV was a Beckman Coulter, System Gold 126 solvent module and 168-detector. The mobile phase was a mixture of methanol, water, and acetonitrile at different volume ratios. Another method used to identify molecular products was high performance liquid chromatography with UV-Vis detection and mass spectrometry with electro-spray ionization (HPLC-UV-MS-ESI) from Agilent Technologies, 6410 Triple Quad LC/MS, 1200 Series. The method designed for the instrument was a full mass/ion (m/z) scan in order to identify the products. A C18 RP, 3.5 micron column was used, as well as the instrument’s auto sampler with a capacity of up to 100 samples. All reaction products were diluted by a factor of one thousand by volumetric dilution prior to injection into the HPLC-UV-MS-ESI for analysis. The mobile phase was a mixture of 0.1 vol. % formic acid in water and 0.1 vol. % formic acid in acetonitrile.
For HPLC-UV-MS analysis, the reference compound, dehydrodiisoeugenol (DHDIE) was synthesized and purified. It was used to verify that under heterogeneous photocatalytic oxidation of isoeugenol, DHDIE was produced. A stock solution of the DHDIE standard was used to compare the retention time and mass spectra with those of the reaction products. In order to synthesize DHDIE, the method reported by Bengt Leopold was followed (32). The synthesis required 50 g isoeugenol dissolved in 450 ml ethyl alcohol and 200 ml water. Then, 70 g ferric chloride in 200 ml water were added drop-wise using a burette, until the solution turned dark green and cloudy (32). The original synthesis requires "a few crystals of dehydrodiisoeugenol," to initiate crystallization, however this was not available. The solution was kept in the freezer for about 1 week, having produced a reddish white crystalline product in a dark yellow solution. The crystals were isolated via vacuum filtration and washed with hexane. Lastly, the crystals were re-crystallized from hexane. The identity of DHDIE was established by the melting point, the retention time on the chromatogram, and the mass spectrometry data when compared to literature values (33, 34).
Chapter 3: Photocatalytic Oxidation Reactions
In photooxidation experiments, a standard mercury lamp that emits light of a certain spectrum was used. Figure 6 shows the optical spectrum of light that penetrates the Pyrex reaction vessel (top) and the quartz reaction vessel (bottom). This is the spectrum of light that is absorbed by the sample, and that would cause photocatalytic or photochemical reaction in the respective vessel. In the spectrum of light transmitted through the Pyrex vessel, there are two major lines: a strong line at 366 nm and a weaker one at 313 nm. Therefore, in photocatalytic experiments which take place in the Pyrex vessel, it is expected that the reaction is due to the band gap excitation of TiO2 with either of these wavelengths.
If a quartz vessel is used, then the radiation penetrating the vessel has the spectrum as shown in Figure 6 (bottom). It is expected that any wavelength including, and shorter than, 366 nm would cause a photocatalytic oxidation reaction due to the absorption and photoexcitation of the TiO2 photocatalyst with an energy band diagram seen in Figure 2. Therefore, one can expect higher yields of the same reaction products, or the formation of different products, as compared to a reaction in a Pyrex vessel.
Figure 6. Optical spectrum of light that penetrates the Pyrex reaction vessel (top) and the quartz vessel (bottom).
3.1 Screening of Phenolic Compounds
Experiments began with the phenolic compounds commonly found in bio-oil (Figure 5) in photocatalytic oxidation reactions. TLC was used to screen products after initial photocatalytic oxidation reactions. Phenolic compounds that reacted under photocatalytic oxidation, such as isoeugenol and vanillin in Figure 9, show more than one spot on the TLC plate with a retention factor (Rf) greater than zero. Phenolic compounds that did not react under photocatalytic oxidation, such as salicylic acid and salicylaldehyde in Figure 7, show one spot on the TLC plate with an Rf value greater than zero. Compounds that did not react have not been investigated further.
Figure 7. TLC plate of phenolic compounds that do not react under photocatalytic oxidation.
3.2 Chemical reactions involving the C=C bond
Isoeugenol and eugenol (Figure 5) are constitutional isomers that differ in the location of the carbon-carbon double bond in the substituent on the benzene ring. Although these constitutional isomers are similar, this difference in the location of the carbon-carbon double bond was proven to be significant. In general, it has been determined that isoeugenol is reactive under heterogeneous photocatalysis and photochemical oxidation, while eugenol is not. Because isoeugenol has an internal double bond in the substituent on the ring, which is in conjugation with the ring, it is more stable than the terminal double bond in the substituent on the ring of eugenol. The locations of these double bonds, internal versus terminal, may be the cause of the difference in reactivity between these constitutional isomers.
In isoeugenol, the carbon-carbon double bond is internal on the substituent group attached to the aromatic ring (Figure 8). The location of this carbon-carbon double bond makes a significant difference in the products from heterogeneous photocatalysis.
Figure 8. Isoeugenol with an internal carbon-carbon double bond.
Initially, products of photocatalytic reactions were analyzed on silica gel TLC plates to determine if a chemical reaction had occurred or not. Each reactant was expected to produce one spot on the TLC plate, while the products were expected to show more than one spot. When looking at the spots on the TLC plate in Figure 9, it is observed that when the wavelength of light used in the reaction (λ, nm) is greater or equal to 254 nm, three spots were produced from isoeugenol. On the other hand with λ ≥ 355 nm, only two spots were observed. Therefore, with light λ ≥ 254 nm, more reaction products are produced from isoeugenol in acetonitrile under photocatalysis. Under the same conditions eugenol yields no product.
Figure 9. TLC silica gel plate of vanillin and isoeugenol photoreaction products from both Pyrex and quartz reaction vessels.
Comparing the retention factors (Rf) of the reactants versus products allowed for the identification of the chemical compound from each spot. The retention factors and assignments for the products of this reaction are shown in Table 1.
Table 1. The TLC Rf values and assignments of photocatalytic oxidation products from silica gel plates.
Because isoeugenol yields more products than eugenol as shown by TLC, the HPLC-UV method was employed further to identify and quantify the compounds observed as three spots on the TLC plates.
The UV spectra of stock solutions were compared to the spectra of the reaction products (Figure 10). UV-Vis spectroscopy was able to show that upon reaction, there is a change in the relative amplitude/area of absorption spectral bands at 240-280 nm and 290-320 nm as seen in Figure 10. These changes indicate that isoeugenol undergoes a chemical reaction.
UV-Vis of Isoeugenol
Figure 10. UV-Vis absorption spectra of isoeugenol in acetonitrile. Before reaction (A); after reaction (B).
The HPLC-UV chromatogram of the stock solution of isoeugenol, before using the HPLC-UV method for its quantification, is shown in Figure 11. The wavelength of 275 nm was chosen for HPLC-UV analysis, since it corresponds to the absorption maximum of the spectral band of isoeugenol in Figure 10. This allowed for the identification of the UV absorbance band of isoeugenol at 275 nm, as well as its retention time on the chromatogram. Identifying the peak of pure isoeugenol was necessary in order to identify what remained in the solutions after a reaction.
Isoeugenol Stock UV
Figure 11. Chromatogram of isoeugenol stock solution (optical absorption at 275 nm). X axis: retention time, min. Y axis: optical absorbance at 275 nm, a.u.
The chromatogram of the isoeugenol stock solution (Figure 11) illustrates that this compound absorbs at wavelength 275 nm, and it elutes off the chromatographic column at about 10 minutes. The width of the single peak is 0. 253 minutes at half height.
Under heterogeneous photocatalytic oxidation, isoeugenol reacts to form two products as shown by HPLC-UV: vanillin and the oxidative "dimer", dehydrodiisoeugenol (DHDIE). Photocatalytic oxidation of isoeugenol to form vanillin apparently occurs through an oxidative C=C bond cleavage reaction (Figure 12).
Figure 12. Reaction of isoeugenol to form vanillin under photocatalytic oxidation.
This photoreaction was done in both a quartz (λ ≥ 254 nm) and Pyrex (λ ≥ 355 nm) vessel. By injecting vanillin stock solution into the HPLC-UV-MS, the retention time for vanillin was determined at 8.8 minutes. By the retention time on the chromatogram, the production of vanillin from isoeugenol could be confirmed. HPLC-UV demonstrates that there is a much higher absorbance at the retention time of vanillin when the product obtained in the quartz vessel was analyzed as compared to the one in the Pyrex vessel. Specifically, the absorbance of the vanillin product at 8.81 minutes is at approximately 36 mau. The vanillin product from the Pyrex vessel had a retention time of 8.81 minutes, but the absorbance is much less, being at only about 10 mau. Thus, absorbance of the vanillin product is more than 3.5-fold higher in the quartz vessel as compared to the Pyrex vessel. It was calculated that there was a 10 % yield of the vanillin product (yield is calculated on a molar basis) as obtained under photocatalysis in the quartz vessel.
In addition to the vanillin product, isoeugenol also undergoes oxidative "dimerization" to form DHDIE as shown in Figure 13. The reaction mechanism of the formation of lignans, including DHDIE, from isoeugenol via photooxidation, involves both a radical of isoeugenol (14) and a radical-cation of isoeugenol (14, 35). It is believed that in non-water media, formation of lignans from isoeugenol proceeds via a radical-cation as the major reactive intermediate (35).
Figure 13. Reaction of isoeugenol to form dehydrodiisoeugenol under photocatalytic oxidation.
The chromatogram of the photocatalytic product revealed the oxidative "dimer," which had a retention time of 11.04 minutes in both the quartz vessel (Figure 14) and the Pyrex vessel (Figure 15). The products from the quartz vessel had the same retention times as those from the Pyrex vessel. In the quartz vessel, products include: vanillin at 8.81 minutes, unreacted isoeugenol at 10.14 minutes, and DHDIE at 11.04 minutes. In the Pyrex vessel, the products include: vanillin at 8.81 minutes, unreacted isoeugenol at 10.14 minutes, and DHDIE at 11.04 minutes. Although these compounds have the same retention times, the absorbance of the products from quartz and Pyrex, as observed in the chromatogram, is significantly different. The absorbance of both products (vanillin and DHDIE) from the quartz vessel is approximately three times that of the products in the Pyrex vessel. It was expected that the quartz vessel (λ ≥ 254 nm) would yield more products than the Pyrex vessel (λ ≥ 355 nm) due to the higher energy of the photons used.
The chromatogram of the quartz vessel (Figure 14) and the Pyrex vessel (Figure 15) illustrate the amounts of the DHDIE product in each case. The molar concentrations of the products, from each case were decreased by a factor of one thousand by volumetric dilution prior to injection into the HPLC-UV. In the quartz vessel, an absorbance of 50 mau for DHDIE was detected. The DHDIE eluted off the chromatographic column at 11.05 minutes, which is the same retention time as the stock solution of the synthesized DHDIE. When analyzing the chromatogram of the sample from the Pyrex vessel, it is clear that a much smaller amount of DHDIE was produced. The absorbance was only 17 mau for the product from this vessel; however, the presence of DHDIE product is still confirmed at a retention time of 11.04 minutes, which again, matches the data obtained with the stock solution of the DHDIE standard. Thus, the absorbance of the DHDIE product is three times higher for the content of the quartz vessel, as compared to the of the Pyrex vessel. Because the amount of DHDIE produced in the quartz vessel varied so much from that of the Pyrex vessel, it was concluded that the wavelength of photoexcitation strongly affects the yield of reaction product. Namely, at the shorter wavelengths (λ ≥ 254 nm), yield of the DHDIE products is ca. 3-fold higher than at the shorter wavelengths (λ ≥ 355 nm) used.
Isoeugenol Quartz UV
Figure 14. Chromatogram of the isoeugenol product in the Quartz vessel (λ ≥ 254 nm) after photocatalytic oxidation. X axis: retention time, min. Y axis: optical absorbance at 275 nm, a.u.
Isoeugenol Pyrex UV
Figure 15. Chromatogram of the isoeugenol product in the Pyrex vessel (λ ≥ 355 nm) after photocatalytic oxidation. X axis: retention time, min. Y axis: optical absorbance at 275 nm, a.u.
The idea that DHDIE could be a reaction product of isoeugenol (36) has been verified in the past. Synthesis of DHDIE from isoeugenol was done using the stable 2,2-diphenyl-1-picrylhydrazyl radical (34), and DHDIE was detected and identified as a product by HPLC–MS. Other researchers have produced DHDIE through photo-sensitized oxidation of isoeugenol, and verified the synthesis using chromatography (15).
The DHDIE that was synthesized and purified in the Samokhvalov laboratory was injected into the HPLC-UV-MS for chromatographic analysis. The stock solution of DHDIE produced a HPLC-UV chromatogram and mass spectrum, Figures 16 and 17, to verify whether the synthesis and purification of DHDIE was successful.
DHDIE stock UV
Figure 16. Chromatogram of the synthesized dehydrodiisoeugenol in acetonitrile. X axis: retention time, min. Y axis: optical absorbance at 275 nm, a.u.
DHDIE stock MS
Figure 17. The mass spectrum of DHDIE.
The chromatogram of the synthesized DHDIE, at wavelength 275 nm, is shown in Figure 16. The DHDIE eluted off of the chromatographic column at 11.05 minutes with an absorbance of 100 mau. The width of the single peak is 0.527 minutes at the full width at half maximum (FWHM). That FWHM confirms that the peak on the chromatogram of DHDIE is likely to be due to one substance. Further, the mass spectrum of the DHDIE reveals three characteristic peaks which include m/z 327, 349, and 675 atomic mass units (a.m.u.) which are proven to be characteristic peaks of DHDIE (34). The masses of those peaks are 327 for [M+H]+, 349 for [M+Na]+, and 675 for [2M+Na]+ (34), with M being the molecular weight of the DHDIE (The peak of 509 a.m.u is attributed to an impurity). Three characteristic mass spectra peaks of 327, 349, and 675 a.m.u. confirm the successful synthesis of DHDIE that can be used to verify the molecular identity of the product of photocatalytic oxidation of isoeugenol.
A calibration curve (Figure 18) of the DHDIE stock was created using the HPLC-UV method. A stock solution of 5.0 mM of DHDIE in acetonitrile was injected into the HPLC-MS at different volumes. The injection volumes ranged from 0 to 0.5 μL. Five different injection volumes, in typical increments of 0.1 μL were used to create the UV-Vis absorption calibration curve. As the injection volume of the solution was increased, there was a corresponding increase in the absorbance at 275 nm. The relationship between the injection volume and the integrated area of the corresponding peaks are illustrated in Figure 18. This calibration curve was used to calculate the molar concentrations of DHDIE produced in both photocatalytic and photochemical oxidation experiments.
DDIE Calibration Curve
Figure 18. Calibration curve of DHDIE stock solution.
After analyzing the major product of photocatalytic oxidation of isoeugenol, DHDIE, the next question to ask was whether different solvents would yield different amounts of DHDIE product. The solvents that were tested under photocatalytic oxidation included acetonitrile, dimethyl sulfoxide, dimethylformamide, and pure isoeugenol (no solvent). Table 2 shows the amounts of DHDIE that were produced in these solvents.
In eugenol, the carbon-carbon double bond is terminal on the substituent group attached to the aromatic ring. Figure 19 illustrates the location of the terminal double bond in eugenol.
Figure 19. Eugenol with a terminal carbon-carbon double bond.
It was found that eugenol does not react under heterogeneous photocatalytic oxidation (37), although it is very similar in structure to isoeugenol. Thin layer chromatography data provided evidence that eugenol was non-reactive under photocatalytic oxidation, because only one spot on the silica gel plate appeared. This TLC spot matched the Rf (0.69) of the eugenol stock solution that was also deposited on the plate. Given that the Rf (0.69) of the eluted spots matched, it was concluded that eugenol did not react. Other compounds such as vanillin and isoeugenol showed two or more spots on the TLC plates, which indicated that a reaction occurred under photocatalytic oxidation. The terminal location of the double bond in eugenol is shown to inhibit photocatalytic oxidation of this compound, as compared to its isomer isoeugenol with an internal double bond, Figure 8.
3.2.3 Origin of reactivity of the radical intermediate
Under photocatalytic oxidation it was proven that isoeugenol reacts to form products, while eugenol does not. It was hypothesized that the differences in reactivity is due to the location of the carbon-carbon double bond in each substituent group on the ring. During photocatalytic oxidation a radical can be formed on the hydroxyl (-OH) oxygen on the substituent. The stability of the radical intermediate is what affects the reactivity of these two isomers. The radical that forms on the hydroxyl oxygen of isoeugenol has extra delocalization (compared to the eugenol analog) because it is in conjugation with the pi-bond in the substituent group on the ring which is illustrated in Figure 20. The extra delocalization of the radical, due to the location of the pi-bond, provides extra stability for the radical ((38). Because the radical is more stable on isoeugenol it is expected to be more reactive than eugenol.
Figure 20. Resonance structures which illustrate the delocalization of the likely radical intermediate of isoeugenol.
When isoeugenol undergoes photocatalytic oxidation, the radical could be formed in conjugation with the ring. Because the carbon-carbon double bond on the substituent of isoeugenol is internal, the radical intermediate has extra resonance stabilization. Eugenol does not lead to a radical that is delocalized because of the location of the pi-bond on the substituent. The radical intermediate of eugenol is less stable because it is not in conjugation with the pi-bond on the substituent group. Resonance delocalization on the substituent chain, shown in structure C of Figure 20, is not possible for eugenol; therefore the radical formed on eugenol is less stable than the radical formed on isoeugenol.
Recognizing that isoeugenol reacts under photocatalytic oxidation and eugenol does not may have some practical applications. This understanding can be used, for example, to separate these two constitutional isomers. Under photocatalytic oxidation isoeugenol reacts to form the solid products vanillin and DHDIE, and these products can be precipitated out of solution. Eugenol, however, does not react to form products and will therefore remain in the liquid phase.
3.3 Chemical reactions involving the C=O bond
Under photocatalytic oxidation it was shown that vanillin does not react. When the product from the vanillin reaction was spotted on a silica gel TLC plate, only one spot appeared on the plate with the same Rf (0.30) as the vanillin stock solution. The HPLC-UV analysis revealed the presence of vanillic acid in a very small amount that was considered negligible. This finding indicates that in non-water solvents, TiO2 does not cause vanillin to react under photocatalytic oxidation; the reason could be, for example, deactivation of the catalytic action of TiO2 due to adsorption of vanillin onto its surface (13). Therefore, under photocatalytic oxidation conditions in non-water solvents, vanillin was not studied any further.
Chapter 4: Solvent Dependence for Photooxidation of Isoeugenol
Isoeugenol was chosen to analyze the dependence of photooxidation reaction upon solvents, because it is abundant in bio-oil and thus inexpensive and practical to use. Isoeugenol also initially formed DHDIE under photocatalytic oxidation in acetonitrile, so it was selected for further tests concerning production of DHDIE under both photocatalytic and photochemical oxidation.
The first goal was to observe how the amount of produced DHDIE product would vary under photocatalysis in various solvents. The second goal was to determine if varying solvents would produce different yields of DHDIE under photochemical oxidation. Solvent selection was based on two criteria: 1) dipole moment, and 2) position on the scale of being a "green" solvent. The results varied for each solvent under both photocatalytic and photochemical oxidation conditions.
4.1 Effect of Dipole Moment of the Solvent
It was hypothesized that the higher the dipole moment, the higher yield of DHDIE product would be produced. This hypothesis was made assuming the formation of a highly polar reactive intermediate of isoeugenol in these experiments. Indeed, such intermediate was reported to be a radical-cation of isoeugenol (35) formed upon flash photolysis at excitation wavelength of 308 nm in CH3CN. Acetonitrile was selected as the initial solvent of choice because of its high dipole moment, 3.92 D. When CH3CN was used as the solvent under photocatalytic oxidation, the most DHDIE was produced, among other solvents tested. Dibutyl ether has a lower dipole moment, 1.18 D, but was chosen because it is considered a "green" solvent. The product of photocatalytic oxidation of isoeugenol in dibutyl ether was analyzed using the HPLC-UV and HPLC-MS analysis which provided evidence of the incorporation of the butyl group to form the product seen in Figure 21, with m/z = 383 amu: 326 amu (DHDIE) plus 57 amu (C4H9).
Dibutyl Ether Product UV
Figure 21. Chromatogram of "butyl substituted" DHDIE formed in dibutyl ether at a retention time of 10.56 minutes.
Dimethylformamide (DMF) was used because of its high dipole moment, 3.86 D. Upon analysis it was found that there was about 5 % (yield is calculated on a molar basis) yield of DHDIE formed under photocatalytic conditions. Perfluorodecalin (PFD) was tested because it has zero dipole moment. Out of all of the solvents tested isoeugenol was only insoluble in PFD which is thought to be the reason why no significant amount of product was detected in this solvent. Dimethyl sulfoxide (DMSO) was used because of its high dipole moment, 3.96 D; however no products were found at a significant yield under photocatalytic oxidation.
Although the dipole moments of the solvents are very similar, the amount of DHDIE that was produced in each solvent was very different. The molar concentration of DHDIE was calculated for each solvent under photocatalytic oxidation (Table 2). Figure 22 shows a comparison of each solvent’s dipole moment to the concentration of DHDIE that was produced under photocatalytic oxidation. There is no simple dependence of the yield of DHDIE upon dipole moment of the solvent used in photocatalytic oxidation. This indicates that there is some other factor that affects the yield of the DHDIE product in photocatalytic oxidation. On the other hand, our data (Figure 22) can be used to speculate about the possible mechanism of photocatalytic oxidation of isoeugenol in non-water solvents (Table 2).
It is known that when water is used as the solvent, the major reactive intermediate of photocatalytic oxidation of organic compounds (39) is the hydroxyl radical, HO.. Specifically, the hydroxyl radical is produced from water via oxidation of the latter by the hole, h+ formed upon photoexcitation of the photocatalyst, including the TiO2 photocatalyst (Figure 3). On the other hand, in polar non-water solvents, the major reactive intermediate is assumed to be the superoxide anion-radical (28). Even when non-water media is used in photocatalytic oxidation with TiO2, water can still be present in the adsorbed form on the surface of the TiO2 photocatalyst (40). Then, that adsorbed water may be the source of the hydroxyl radicals that could participate in the reaction of photocatalytic oxidation in non-water media and lead to products such as DHDIE. This hypothesis can be tested by considering the following. First, if such hydroxyl radicals were present in significant amounts during photocatalytic oxidation, as performed in this research, then products of photocatalytic oxidation would have been significantly different from those obtained during photochemical oxidation. Instead, in both photocatalytic and photochemical oxidation reactions of isoeugenol, as judged by the HPLC-UV chromatograms, major products are the same, namely, DHDIE and vanillin. Therefore, hydroxyl radicals, even if present in the photocatalytic oxidation reactions, performed in this research, have no significant effect on the chemical identity of the major reaction products. Second, if such hydroxyl radicals were present in significant amounts during photocatalytic oxidation, the yield of the major product, DHDIE would have been higher as compared to the yields obtained in photochemical oxidation in various organic solvents. Then, one would expect that concentrations of DHDIE in photocatalytic oxidation reactions in the solvents, performed in this research (Figure 22, with solvents ranked by dipole moment) would have been higher than those obtained in the photochemical oxidation reactions (Figure 23). However, such effect has not been observed; the yield of DHDIE in DMSO as the solvent is significantly lower in photocatalytic oxidation vs. photochemical oxidation. Therefore, TiO2 acts as an apparent inhibitor of photooxidation of isoeugenol to DHDIE, rather than the "true" photocatalyst as one would expect for TiO2, with participation of the hydroxyl radicals as the major reactive intermediate.
Table 2. The concentration of dehydrodiisoeugenol (DHDIE) produced from each solvent tested under photocatalytic oxidation (λ ≥ 355 nm).
Molar Concentration of dehydrodiisoeugenol (DHDIE) in Photocatalytic Reactions
Retention Time (min.)
Molar Concentration [mM]
Figure 22. The concentration of DHDIE (mM) compared to the dipole moment of each solvent under photocatalytic oxidation (λ ≥ 355 nm).
Under photochemical oxidation, the solvents used behaved very differently as compared to photocatalytic oxidation. Although DMSO, DMF, and CH3CN have very close dipole moments, they produced significantly different yields of DHDIE under photochemical oxidation. DMSO has the highest dipole moment (3.96 D) and had the highest yield of DHDIE under photochemical oxidation (Table 3 and Figure 23). It is speculated that the reaction depends on the solvation of a radical cation intermediate (35), thus it is different for different solvents.
Table 3. The concentration of dehydrodiisoeugenol (DHDIE) produced from each solvent tested under photochemical oxidation (λ ≥ 355 nm).
Molar Concentration of dehydrodiisoeugenol (DHDIE) in Photochemical Reactions
Retention Time (min.)
Molar Concentration [mM]
Figure 23. The concentration of DHDIE (mM) versus dipole moment of each solvent used under photochemical oxidation (λ ≥ 355 nm).
4.2 Effect of Oxygen Solubility of the Solvent
Another property compared among the solvents was their solubility of oxygen. It was hypothesized that the higher the oxygen solubility of the solvent, the more product that would be produced. Oxygen solubility in the solvent was compared for both photocatalytic and photochemical oxidation experiments. The result of these two different reaction environments differs greatly.
Under photocatalytic oxidation (Figure 24), low oxygen solubility of the solvent resulted in low yields of DHDIE. When no solvent was used, only pure isoeugenol, which has low oxygen solubility, (solubility in benzene was taken (41), since the solubility of oxygen in isoeugenol was not available) no DHDIE was produced. DMSO, which has a low oxygen solubility of 0.33 mM, had a low yield of DHDIE. DMF, with an oxygen solubility of 1.31 mM, and CH3CN, with an oxygen solubility of 2.42 mM (42), have high oxygen solubility, and produced the highest yields of DHDIE. This trend is evidence that the formation of DHDIE may depend on the presence of oxygen. As the oxygen solubility of the solvent increases, the concentration of the DHDIE product increases as well, which is illustrated in Figure 24.
Figure 24. The concentration of DHDIE (mM) compared to the solubility of each solvent under photocatalytic oxidation (λ ≥ 355 nm).
Under photochemical oxidation (without TiO2), in Figure 25, the oxygen solubility of each solvent compared to the yield of DHDIE produced, is very different from the results of the photocatalytic oxidation reactions. DMSO has the lowest solubility of oxygen at 0.33 mM (42), yet it has the highest yield of DHDIE. DMF (1.31 mM), and CH3CN (2.42 mM), have higher oxygen solubilities, yet they produced very little DHDIE or none under photochemical conditions (Figure 25). This is evidence that the reaction is not dependent on the concentration of oxygen. Based on this information there are two possible cases for the reaction pathway for the production of the oxidative "dimer". In case 1, the DMSO solvates the substrate. In case 2, the DMSO solvates the intermediate. The energy of DMSO and the other solvents is the same; however the energy of the reaction intermediates differs. The energy of the intermediates in DMSO is significantly lower than the energy of the intermediates in the other solvents.
Figure 25. The concentration of DHDIE (mM) compared to the solubility of each solvent under photochemical oxidation (λ ≥ 355 nm).
The high yield of DHDIE in dimethyl sulfoxide (DMSO) was expected. It is known that DMSO is an electron donor and a highly polar solvent. This makes DMSO a selective solvent for a variety of organic compounds and polymeric compounds. DMSO can enter into hydrogen bonding and dipole-dipole association with molecules. DMSO is also well suited to initiate polymerization and to strongly solvate cations (43). It was speculated that DMSO strongly solvates one of the reactive intermediates, a radical cation of isoeugenol, compared to other solvents. Such solvation would lower the energy of the reactive intermediate and accelerate the rate of formation of the DHDIE product.
Chapter 5: Photochemical Oxidation Reactions
Photochemical oxidation experiments have been conducted in a Pyrex vessel with the spectrum of light illustrated in Figure 6. This is the light at 313 nm (Figure 6, top) that would photoexcite the molecules that have an optical band at ca. 310-325 nm, such as isoeugenol, Figure 10A.
5.1 Chemical reactions involving the C=O bond
A molecule of vanillin (Figure 26) has an aromatic benzene ring with three different substituent groups around the ring. The aldehyde group is in the para position to the hydroxyl group, while the methoxy group is in the meta position to the aldehyde group as Figure 26 illustrates.
Figure 26. Vanillin with a hydroxyl substituent group, methoxy substituent group, and aldehyde substituent group on the aromatic ring.
Vanillic acid is a common oxidation product of vanillin (11). Studies have shown that upon standing at room temperature in moist air over long periods of time, vanillin oxidized in a water solution to vanillic acid. Others have directly oxidized vanillin to vanillic acid by means of silver oxide (44).
In this experiment, the liquid phase from the reaction vessel after photochemical oxidation with air was diluted with CH3CN and injected into the HPLC-UV-MS. The product had a peak that eluted on the chromatogram at 9 minutes. This product eluted at the same time, under the same conditions as vanillic acid that was injected as the standard. Figure 27 shows the reaction of vanillin to form vanillic acid (37). Because vanillin is reactive under photochemical oxidation, while salicylaldehyde which lacks the methoxy group (Figure 5) is not (no peak for salicylic acid was detected on the chromatogram), it is hypothesized that the presence of the methoxy substituent on the ring plays a significant role in the reactivity of this molecule.
Figure 27. Photochemical oxidation of vanillin to vanillic acid.
5.2 Dependence of Reaction upon Pressure of Oxygen
The photoreactions that were conducted in the majority of lab experiments took place at atmospheric pressure. To increase the solubility of oxygen in the liquid solutions during photoreactions, the pressure of the oxygen in the reaction vessel was increased. Under photochemical oxidation at atmospheric pressure, vanillin produced only small amounts of solid product. The hypothesis was that if more oxygen was able to dissolve in the solution during the photochemical reaction this would produce a higher yield of solid product from vanillin.
Considering Henry’s law of pressure and gas solubility, this hypothesis was tested. Henry’s law states that "the solubility of a gas is directly proportional to its partial pressure, P" (45). The way Henry’s law works is that when the pressure of a gas in equilibrium with its solution increases, the rate of the gas molecules striking the solution increases. Therefore one would expect that the solubility of the gas would increase as its pressure is increased (45). The solubility of oxygen in acetonitrile is 2.42 mM at atmospheric pressure (42). In order to increase oxygen’s solubility in acetonitrile, the reaction vessel was pressurized under approximately 100 psi of either air or oxygen.
In a photochemical oxidation experiment under approximately 100 psi of pure oxygen, a 0.5 M stock solution of vanillin in acetonitrile was allowed to react in a pressurized reaction vessel. After 4 hours, a 7 % yield by mass of solid product from vanillin was obtained (yield by mass is calculated as mass of "polyvanillin" product obtained divided by the mass of the vanillin reactant multiplied by 100 %). In order to further increase the yield of product, the same experiment was conducted for a longer period of time. The second experiment, conducted for a period of 24 hours at a constant 100 psi O2, produced a yield of 20 % solid product. This is evidence of a relationship between the time of the reaction and the mass of precipitate formed during the reaction.
Under photochemical oxidation, vanillin forms a white powder-like precipitate in the reaction vessel. The solid precipitate was isolated from the liquid solution by vacuum filtration and washed with CH3CN in preparation for analysis. It was intended that a solution of this precipitate was injected into the HPLC-UV-MS for analysis; however, the solid was insoluble in common solvents such as chloroform, dichloromethane, and water.
Because the powder remained intact after attempted dissolution, it was dried, and a melting point determination was attempted. Using a Mel-Temp apparatus the solid powder would not melt, and it thermally decomposed (with darkening) at above 300 ○C. At this point it was clear that the product that had been formed via photochemical oxidation was a compound that would not have a distinct melting point. Formation of the dimeric product, divanillin (Figure 28), had been reported in the past "after oxidation of vanillin in aqueous solution with peroxidase in the presence of hydrogen peroxide" (11). Divanillin has also been observed in fermented vanilla beans via NMR (46).
Figure 28. Vanillin reacts to form divanillin in the presence of hydrogen peroxide (11).
Like any "monomeric" organic compound, divanillin would also show a distinct melting point below 300 ○C, so it is concluded that the solid product is not divanillin. In order to gain further insight into the identity of the solid product that had formed, FTIR was employed. KBr pellets containing pure vanillin or the solid product "polyvanillin" were made to obtain IR spectra of both substances. The IR spectra of the "polyvanillin" product and a sample of pure vanillin (47) provided insight to the structure of the "polyvanillin".
Oxidation of vanillin into "polyvanillin" has been reported. Catalytic oxidation of vanillin using oxygen to form "polyvanillin" in water media, in the presence of nano-sized Co3O4 and Mn3O4 and Pd/alumina, yields the "polyvanillin" that contains significant amounts of "free" and bonded COOH carboxylic groups (48), which was found by FTIR. Therefore, in addition to polymerization on the phenolic ring, oxidation of the substituent aldehyde group also occurs, so polymerization is non-selective. In addition, the "polyvanillin", by the above research group, contains three different fractions: CH2Cl2 soluble, water soluble and insoluble. Therefore, the reported catalytically produced "polyvanillin" features heterogeneous, structurally irregular molecules as shown by wide FTIR bands (48) as a result of non-selective transformation of aromatic rings, CHO and other substituents. "Polyvanillin" has been also reported to be synthesized from vanillin via electrochemical oxidation in alkaline media (49). This "polyvanillin" has been obtained with a high yield of 91 %; however, it does not contain the aldehyde group, which is the obvious result of non-selective polymerization on the aromatic ring, with the loss of a CHO substituent.
FTIR analysis was performed on both the pure vanillin and "polyvanillin" product produced in the Samokhvalov laboratory using KBr pellets. The spectral bands of the pure vanillin (Figure 29) and "polyvanillin" powder (Figure 30) that were obtained in absorbance mode of the FTIR spectrophotometer were compared. In the spectrum of vanillin, there is a broad intramolecular hydrogen bonded O-H stretch centered at 3184 cm-1 as expected, due to presence of the phenolic OH group in the molecule. There are overtone or combination bands with lower intensities between 2000 and 1670 cm-1. The aldehyde group shows a spectral band at 1668 cm-1 due to the stretch of the C=O bond. Because the aldehyde group is in conjugation with the phenyl group, a lower frequency is observed, as compared to an aldehyde group that is not attached to an aromatic ring. In the range of 1670-1400 cm-1 there are several ring vibration modes attributed to the carbon-carbon ring stretching of the phenyl group (50).
Between 1260 and 1100 cm-1 there are bands which result from the interactions between the O-H bending and C-O stretching of the hydroxyl group with the aromatic ring (50). In this range, a spectral band at 1201 cm-1 (Figure 29) can be assigned to the carbon-oxygen stretch of the hydroxyl (phenolic) group attached to the aromatic ring. The band at 1268 cm-1 could be assigned to the aryl C-O stretch of the methoxy group (51). At 1029 cm-1 there is a band that could be assigned to the O-C stretch of the methoxy substituent directly attached to the ring (50). Between 600 and 900 cm-1, there are peaks at 860 and 813 cm-1 that are assigned to the out-of-plane C-H bending mode (51). The spectral bands, as described above, are indicative of the 1,2,4 ring substitution pattern of vanillin (50), as expected.
In general, the FTIR spectrum of "polyvanillin" (Figure 30) significantly resembles that of vanillin (Figure 29). A few important changes can be noted. First, there is a broad spectral band at 3254 cm-1 due to the intramolecular hydrogen bonded O-H stretch in "polyvanillin" that is shifted from 3184 cm-1 in vanillin. Such spectral shift indicates a different kind of hydrogen bonding in "polyvanillin" as compared to vanillin. This could be due to 1) the change in the number of OH groups in "polyvanillin" compared to vanillin, 2) the change of the molecular environment of those groups, or both. Second, the presence of an aldehyde group in "polyvanillin" is indicated by the strong band at 1673 cm-1 due to the stretch of the C=O bond, that is close to spectral band at 1668 cm-1 in the spectrum of vanillin. Given spectral resolution of FTIR at 4 cm-1, that was used this indicates that the aldehyde group in "polyvanillin" is not significantly different from the aldehyde group in vanillin. There are several vibration ring modes that produce bands at 1422, 1457, 1506, and 1586 cm-1 that are attributed to the carbon-carbon ring stretching of the phenyl group in "polyvanillin". These stretches are also observed as bands in the range of 1670-1400 cm-1 in the FTIR spectrum of vanillin (50).
In "polyvanillin" the band at 1258 cm-1 could be assigned to the aryl C-O stretch of the methoxy group (51). This band could also be assigned to the C-C stretch of the aryl aldehyde group (51, 52). Compared to the band in vanillin at 1268 cm-1 there is a decrease in spectral frequency. This shift is due to the difference in ring substitution between vanillin and "polyvanillin." The decrease in frequency is a result of an increase of effective mass of the quantum oscillator of this bond, due to extra substituents on the ring in "polyvanillin".
An important difference between "polyvanillin" and vanillin is that the spectral band at 1201 cm-1 which is present in vanillin, is absent in "polyvanillin". This band is due to the carbon-oxygen stretch of the hydroxyl (phenolic) group attached to the aromatic ring. However, there is a weak band at 1182 cm-1 that is absent in the spectrum of vanillin. Therefore, in "polyvanillin" this OH phenolic functional group is not present, significantly reduced in intensity, or shifted due to the different molecular environment. On the other hand, the presence of the broad spectral band at 3254 cm-1 due to the intramolecular hydrogen bonded O-H stretch in "polyvanillin," proves that there are some OH groups present in "polyvanillin". Therefore, in "polyvanillin" this OH phenolic group is not missing entirely, but rather is decreased in intensity and shifted to lower wavenumbers. These changes could be due to the reaction of vanillin to "polyvanillin" via a chemical reaction or interaction between two phenyl rings of vanillin. Indeed, such a reaction pathway is well-known and reported for electrochemical oxidation of vanillin to "polyvanillin" in aqueous sodium hydroxide (49), as well as the formation of divanillin from vanillin when water is the solvent (53). On the other hand, the disappearance of the band of the OH phenolic group at 1201 cm-1 indicates that the formation of the reaction product via an alternative pathway is also possible, namely, through linking of two phenolic rings by the oxygen atom. This possibility is indicated for the formation of "polyvanillin" via catalytic oxidation of vanillin with molecular oxygen (48).
In "polyvanillin", there is one band at 847 cm-1 in the region where there are two bands in spectrum of vanillin at 859 and 813 cm-1. These bands are due to out-of-plane C-H bending (50). These differences indicate a change in the substitution pattern on the phenolic ring. Specifically, the tetrasubstituted phenyl ring in "polyvanillin" is observed (51), compared to the trisubstituted phenyl ring in vanillin with 1,2,4 substitution pattern (50). This finding confirms that "polyvanillin" is formed due to a substitution reaction on the phenyl ring of vanillin. Further work is needed to better understand the molecular structure of this product.