Upgrading of anisole using in situ generated hydrogen in pin to plate pulsed corona discharge

Abstract

In this paper, the upgrading of anisole as a model compound for lignin derived bio-oils has been studied experimentally via a pin to plate pulsed corona discharge for the first time. The required hydrogen is generated in situ by plasma decomposition of methyl radicals which exist in bio oil chemical structure. In order to maximize the conversion and improving the selectivity of products, some operating parameters including pulsed repetition frequency, pin number, gap distance, plate electrode diameter and carrier gas flow rate have been evaluated. According to the results, the main products were BTX and phenol. Demethylation, transalkylation, hydrogenolysis, demethoxylation and methyl decomposition are the main chemical reactions. Results reveal that the selectivity of the desired products was enhanced through increasing the decomposition of generated methyl radicals which can be obtained by increasing the energy and number of collisions in the discharge zone. The maximum anisole conversion of 68%, phenol selectivity of 45.6% and BTX selectivity of 21.6% were obtained. Due to the low temperature and atmospheric pressure conditions no hydrogenation of the aromatic ring was observed. The results indicated that the corona discharge plasma reactor can overcome the hydrogen challenges of the hydrodeoxygenation reaction.

---

1. Introduction

Nowadays, the high demand for energy, the depleting fossil fuel resources and environmental concerns have aroused interest in utilizing alternative sources of hydrocarbons. Nonfood lignocellulosic biomass is one of the most available, abundant, low price¹ and renewable sources of hydrocarbons which still remains untouched.² Thermal processing of lignocellulosic biomass generate liquid products called bio-oil. Consumption of produced bio oil has short CO₂ cycle^3,4 without any SOx emission.^5,6 However, there is a significant challenge for utilization of this type of fuel which is high oxygen content (up to 45 wt% O) resulting in some defects which hinders its immiscibility with fossil fuels. High acidity and viscosity, limited heating value,⁷ poor stability, are some disadvantages of the produced bio oil with high oxygen content. It would be advantageous if bio-oil can be upgraded to an oil similar to conventional crude oil.

To date, achieving oxygen removal from bio oil has been mostly limited to hydrodeoxygenation (HDO) which involves treatment of oxygen-containing molecules with hydrogen in the presence of heterogeneous catalysts.^8,9 However, this method suffers from high pressure and costly consumption of H₂.¹⁰ This accompanies some drawbacks. Firstly, high hydrogen pressure leads to large consumption of hydrogen, which is explosive.¹¹ A special care is needed for production, transportation, consumption and storage of this gas. Moreover, the current approaches for production of hydrogen is energy intensive, leading to considerable amount of pollutants and greenhouse gases.

High hydrogen pressure also produced ring saturation products via unwanted hydrogenation, which causes a reduction in octane number and hinders its direct utilization as fuel or co-feeding in existing oil refineries.

Additionally, high reaction temperatures makes this process most energy-consuming and unfeasible for industrial application.¹ Many challenges remain for the development of HDO process that utilize minimum amounts of H₂ and perform at low temperature. Li et al.¹² introduced instead mild electrocatalytic hydrogenation and hydrodeoxygenation of bio-oil using ruthenium supported on activated carbon cloth. The required hydrogen formed on the catalytic electrode surface via reduction of protons from solution.

Recently, a detailed review paper has been published by our group to introduced catalytic hydrodeoxygenation of lignin derived bio oil.¹³ Moreover, several efforts have paid by our group to investigate the upgrading of different model compounds which represent the lignin derived bio-oils in thermal reactor.^14–16

Our intent was and still is to improving this process in terms of reducing reaction temperature and elimination of high hydrogen requirement. Recently, we disclosed the dielectric barrier discharge (DBD) plasma as an effective and sustainable alternative approach for upgrading of lignin-derived bio-oil.^11,17–20 Plasma is an ionized gas, which consists of reactive species, such as ions, electrons and excited atoms and have the ability of performing reactions.^21,22 In the NTP, although the electrons temperature rise up to 10⁴ to 10⁵ K, the gas temperature remains within the level of ambient temperature.^23,24 In NTP reactors used for upgrading process, free electrons with mean energy of 1–10 eV, lose their energy through collisions with bio oil molecules. This collision can break C–H, C–O and O–H bonds which have lower mean energy than electron energy and lead to the production of small activated radicals which could be recombined and produce new products.

We studied upgrading of anisole and 4-methylanisole as simple lignin-derived monomers by using a combination of DBD plasma and heterogeneous catalysis and plasma alone.^11,17–20 Results revealed that plasma technology was responsible for the progress of upgrading of lignin-derived monomers. Upgrading reaction using plasma discharge occurred at low temperature but the reactions of thermal catalysis performed above 573 K. The most striking difference between plasma and thermal methods is the elimination of high rate of H₂ consumption by the plasma which can suppress the requirement of high-pressure equipment. Despite these operational differences between plasma and thermal methods, the major products were almost similar in both methods. Similar main reactions were also observed for plasma and thermal methods (e.g., demethylation, transalkylation and hydrogenolysis). However, due to the low temperature and atmospheric pressure conditions no hydrogenation of aromatic ring was observed in DBD plasma. Our experimental results showed that the conversion and deoxygenation for lignin-derived monomers can reached to 99% and 47% by adjusting operating and design parameters of DBD reactor.¹¹

Considering the ability of plasma to decompose hydrocarbons,^25,26 in this study, the ability of pin to plate pulsed corona discharge plasma for decomposition of methyl and methoxyl radicals was used for providing the required hydrogen combined with hydrodeoxygenation reaction for the first time. This goal cannot be achieved in conventional HDO reactors due to high energy requirement for breaking the chemical bonds of methyl radicals which is much more than the required energy for HDO reaction.

Anisole was used as a model compound since anisole is one of the major component of bio oil which contains methoxyl groups. The effects of important operating parameters, namely, pulsed repetition frequency, pin number, gap distance, plate electrode diameter and carrier gas flow rate are investigated on anisole conversion, discharge power and selectivity of products.

2. Experimental section

A schematic diagram of the experimental system for bio oil upgrading process with a pulsed corona discharge is illustrated in Fig. 1. The experiments are performed in a pin to plate corona reactor. The reactor is made of a Pyrex glass tube with outer diameter of 4.5 cm. The pin (diameter of 0.2 mm) and plate electrodes (diameter of 20–35 mm) made of stainless steel placed in the center of the reactor. The high voltage is connected to the pin electrode and the plate electrode is connected to the ground. In the case of multi pin electrodes, pins were distributed uniformly with a circular shape. Anisole (purity: 99%, Merck, Darmstadt, Germany) was used as the liquid feed and injected on a circular pan before starting the reaction. The pan was made of polytetrafluoroethylene (PTFE) with the thickness of 2 mm and placed on the plate electrode. Ar as a carrier gas was introduce to the reactor. The flow rate of Ar is controlled using Alicat Scientific mass flow controller (MC-1SLPM-D).

Fig. 1 Schematic of corona plasma reactor and a photo of radial propagation of the streamers.

Voltage, rise time and fall time, pulse width and frequency of high voltage pulse generator are up to 10 kV, 100 ns, 50 ns and 20 kHz, respectively. Since the pulse duration is much smaller than time lag between pulses, the average power input is low.

The applied voltage and current are measured using a high voltage probe (Tektronix, P6015A, 1000 : 1) and a current probe (Pearson 150, 2A/V), respectively. The output signals are transmitted to a high-speed oscilloscope (Tektronix TDS 1012B-SC, bandwidth of 100 MHz and a time resolution of 1 GS s⁻¹).

The average power delivered to the reactor can be calculated using eqn (1):

where f, V_i, I_i and (t_i+1 − t_i) are pulse repetition frequency, instantaneous applied voltage, current and time interval between each pulse, respectively.

Analysis of the liquid products was performed with a gas chromatograph-mass spectrometer (Shimadzu QP-5050, Kyoto, Japan) equipped with a SGE-BPX5 capillary column with helium as the carrier gas.

A gas chromatograph (GC112A) equipped with FID detector and a SGE-BPX5 capillary column was employed for quantification of the compounds by injection of 2 μL of each sample into the GC. The major components were detected and quantified by GC-MS and GC, respectively; in contrast, some trace products were detected but not quantified.

For investigation and analysis of the experimental results, some parameters including conversion, product selectivity and deoxygenation were defined which calculated using eqn (2)–(4) respectively.

where C_in and C_out are the input and output concentrations of anisole in the reactor, respectively.

3. Result and discussion

In this experimental investigation, an effective approach was disclosed for upgrading of bio oil using in situ generated hydrogen. Pin to plate pulsed corona discharge was used and the influences of important parameters, namely, the pulsed repetition frequency (PRF), carrier gas flow rate, pin number, plate electrode diameter, gap distance were examined on the conversion of anisole and distribution of products. The purpose of this study is to use the ability of methane decomposition in corona plasma in order to generate the required hydrogen for hydrodeoxygenation reaction from methyl and methoxyl radicals which exist in bio oil chemical structure.

Table 1 summarize the operating conditions under which the tests were performed. It is worth mentioning that all of the experiments were performed at atmospheric pressure. Moreover, each test was performed two times. Therefore each presented datum is the mean of two data points with a standard deviation of 4–10%.

Table 1 Operating parameters of experiments

Experiments	Applied voltage (kV)	PRF (kHz)	Pin number	Gap distance (mm)	Ar flow rate (ml min^−1)	Plate electrode diameter (mm)
PRF	10	5–20	6	5	300	25
Gap distance	10	20	6	1–15	300	25
Pin number	10	20	1–6	5	300	25
Ar flow rate	10	20	6	5	100–900	25
Plate electrode diameter	10	20	6	5	500	20–35

---

In these series of experiments, firstly the carrier gas (Ar) was fed into the reactor to saturate the reactor volume and push the air out of the reactor. After that the high voltage power supply was connected to the inner electrode. The produced electric field leads to the formation of plasma, which generate a high flux of energetic electrons. These electrons impact the carrier gas atoms (Ar) and exciting them.

The liquid anisole with total volume of 5 ml cover approximately the entire area of the plate electrode surface. In the discharge zone, the liquid film of anisole may be evaporated or decomposed as a result of collision with high electric streamers, energetic electrons or metastables (Ar*), which depends on the energy transferred to the anisole molecules. If the transferred energy is high enough, the anisole molecules can be decomposed via breaking of chemical bonds and resulting in the formation of ions, atoms, and free radicals. Recombination of these free radicals led to the production of a range of liquid products which were collected from the reactor after 5 minutes.

3.1. Reaction mechanism and product distribution

There are a great number of energetic species in the discharge zone of corona plasma reactor. Because of this, the reaction mechanisms in this type of reactor are more complex than those in a conventional reactor.²⁷ Moreover, the reaction pathway in plasma reactor changes easily under different conditions, and the reactions are very sensitive to the population of active species.²⁸ Therefore, proposing a unique and decisive mechanism that can explain the reaction pathway in plasma reactor is so difficult. However, integration of the thermodynamic database of species with plasma discharge mechanisms enables us to propose the main initiation reaction steps.

Plasma reaction is based on collision of energetic species and feed molecules. Generally by applying a high voltage to needle electrode, the generated electric field produces a high flux of fast electrons which excite the neutral argon atoms and a pulsed corona develops at needle tip (R1).

e⁻ + Ar → Ar* + e⁻ (R1)

Then it is lengthened and reaches the liquid surface and forms discharge channels.²⁹ A great number of energetic electrons and excited species are formed during the generation of discharge channels. After that, energetic species in plasma medium collide to chemical bonds of feed molecules and release their energy. These attacked chemical bonds would break, if the transferred energy to chemical bonds is sufficient.

Generally, in the plasma reactor the mean energy of electrons and other reactive species is higher than dissociation energy of the chemical bonds in anisole molecules which makes the breaking of chemical bond possible. The possible initiation reactions are as follow:

C₆H₅OCH₃ → C₆H₅O + CH₃, (D_e = 267.78) (R2)

C₆H₅OCH₃ → C₆H₅OCH₂ + H, (D_e = 389.4) (R3)

C₆H₅OCH₃ → C₆H₄OCH₃ + H, (D_e ≤ 424.68) (R4)

C₆H₅OCH₃ → C₆H₅ + OCH₃, (D_e = 424.68) (R5)

where D_e is the bond dissociation energy with unit of kJ mol⁻¹.

Although these reactions are possible, the probability of each of them is not the same, and consequently the production rate of radicals in initiation step is not equal. It is clear that, breaking the weaker chemical bond is easier and more probable than stronger one. Data in Table 2 show that among the chemical bonds in anisole molecule, the weakest chemical bond is C₆H₅O–CH₃ bond which has a dissociation energy of 2.78 eV (267.78 kJ mol⁻¹). The reason is the resonance stabilization of the phenoxy radical (C₆H₅O).³⁰ Therefore, the rate of demethylation reaction (R2) through breaking the C₆H₅O–CH₃ bond seems to be much higher than other reactions, leading to the formation of phenoxy radicals.^30–38 Considering initiation reactions, a large number of free radicals are generated in the active volume of plasma as a result of subsequent decomposition of reactants by collision with electrons (e⁻) or M. Hereafter, M represents the temporary excited collision partner which has a higher energy level like Ar* and H*. Moreover, the produced radicals from reactions (R3) and (R4) (methylium,^30,32 phenoxy radical^30–38 and phenyl, x-methoxy radical³²) can be decomposed into smaller fragments, but there is a low possibility for them because of their relatively low concentrations within discharge zone of reactor.

Table 2 Bond dissociation energies in anisole and phenol (kJ mol⁻¹)

Bond	De	Ref.
C6H5O–CH3	267.78	30 and 39
C6H5OCH2–H	389.11	40
C6H5–OCH3	424.68	30 and 39
Caromatic–H in anisole	≤424.68	41
Caromatic–Caromatic	∼518	42
C6H5O–H	372.6	43 and 44
C6H5–OH	431.2	41
Caromatic–H in phenol	≤431.2	41

---

Finally recombination of generated free radicals led to the production of a range of products. In the present study, BTX (benzene, toluene, p-xylene and o-xylene), mono-oxygenated compounds, namely, phenol, methylphenols (2-methylphenol and 4-methylphenol), dimethylphenols (2,4- 2,3- 2,6- 3,4 and 3,5-dimethylphenol), trimethylphenols (2,3,6- 2,4,5- 2,4,6 and 3,4,5-trimethylphenol), tetramethylphenols (2,3,5,6 and 2,3,4,6-tetramethylphenol) and methylanisoles (2-methylanisole and 4-methylanisole) were detected as major products. A sample numerical details of products and their selectivity are presented in the ESI.†

The produced phenoxy free radicals, from reaction (R2), can react with the H free radicals, forming phenol as shown in reaction (R6).^32,45 Moreover, the produced radical from reaction (R4), especially p-methoxyphenyl (4-methoxyphenyl) radical (C₆H₄OCH₃), may react with methyl radical, producing 4-methylanisole (reaction (R7)).^46,47 In this study, the selectivity of phenol in products is greater than the selectivity of other components. This implies that there is a good agreement between the proposed mechanism of anisole dissociation and product distribution. Moreover, the total selectivity of phenol and phenol-derived products is much higher than the selectivity of 4-methylanisole in all experiments. This clearly implies that the rate of reaction (R2) is much higher than reaction (R4), and shows a good agreement between proposed anisole dissociation mechanism and experimental results.

The generated phenol molecules (C₆H₅OH) can be decomposed into phenyl radical (C₆H₅) via breaking the chemical bond between O and aromatic C (R8), and also into hydroxyphenyl radicals (C₆H₄OH) via breaking the bond between of aromatic C and H (R9) (ortho-, meta-, or para-position).^40,41

C₆H₅O + H → C₆H₅OH (R6)

C₆H₄OCH₃ + CH₃ → CH₃C₆H₄OCH₃ (R7)

C₆H₅OH → C₆H₅ + OH, (D_e = 431.2) (R8)

C₆H₅OH → C₆H₄OH + H, (D_e ≤ 431.2) (R9)

As the bond dissociation energy of C_aromatic–H is lower than C₆H₅–OH, the rate of dissociation of phenol into hydroxyphenyl radicals is greater than into phenyl radical. This leads to more available hydroxyphenyl radicals than phenyl radicals.⁴¹ Consequently, methyl substituted phenols were detected with higher selectivity than BTX. Comparing the selectivity of benzene with the total selectivity of methylphenols, i.e., methylphenols, and dimethylphenol, reveals that the rate of breaking the C_aromatic–H bond (reaction (R9)) is greater than the rate of breaking the C_aromatic–O bond (reaction (R8)) in dissociation of phenol. 4-Methylphenol may be formed by reaction between p-hydroxyphenyl and methyl radicals (reaction (R11)).⁴⁰ 2,6-Dimethylphenol is formed through decomposition of 2-methylphenol, by collision with electrons or M, and reaction with methyl radical as shown in reactions (R12).⁴⁸ Consecutive deoxygenation (R8) and hydrogenation (R13) of phenol (or hydrogenolysis of phenol) proceed to form benzene.³¹ Benzene could also be produced through demethoxylation of anisole (R5) + (R13).^48,49

C₆H₄OH + CH₃ → CH₃C₆H₄OH (ortho) (R10)

C₆H₄OH + CH₃ → CH₃C₆H₄OH (para) (R11)

CH₃C₆H₄OH + CH₃ → (CH₃)₂C₆H₃OH (R12)

C₆H₅ + H → C₆H₆ (R13)

C₆H₅ + CH₃ → CH₃C₆H₅ (R14)

CH₃C₆H₅ + CH₃ → (CH₃)₂C₆H₅ (R15)

In addition, the C–H bond energies of aromatic rings vary between ortho-, meta-, and para-position.⁵⁰ Squires and co-workers demonstrated that the strength of C–H bond increases from the ortho- to meta- to para-positions.^51,52 This fact clearly quantifies the increasing production rate of o-hydroxyphenyl radical (2-hydroxyphenyl radical), as represented by reaction (R10)-o, in comparison with meta-(reaction (R10)-m) or para-radicals (reaction (R10)-p).

The required hydrogen for the mentioned reactions can be generated in situ through decomposition of generated methyl radicals from breaking the O–CH₃ bond to CH₂, CH, and H free radicals.^53–58 Several efforts have paid for decomposition of methane and methanol to hydrogen in plasma reactors especially pin to plate corona discharge.^59–63 In a recent study, a pin to plate plasma reactor operating at atmospheric pressure has been used by Aleknaviciute et al.⁶¹ for converting methane to hydrogen. Their experimental results shows the highest hydrogen selectivity of 56% with methane conversion of 31% at 19 W discharge power.

Considering the ability of plasma to decompose methane, decomposition of methyl groups in plasma can alter the selectivity of products. The lower the decomposition of methyl group is, the more is the selectivity of methylphenols. Higher selectivity of phenol and benzene can be obtained in plasma conditions which more methyl group decomposition occurs. The remaining methyl radicals which are not decomposed transferred to the aromatic ring and produce methylphenols (R10)–(13) and methylanisoles (R7) by transalkylation reaction.⁶⁴ This is depend on the energy and number of collisions in the discharge zone. Energy and number of collisions can be adjusted via variation of operating and design parameters. If the energy of collisions increases, plasma will be able to break the stronger chemical bonds. Hence, phenoxy radicals, which is the result of breaking the weakest bond, is converted to other products which their production need more energy. Moreover more energetic collisions leads to increasing the converted methyl and methoxyl radicals to hydrogen. Subsequently, the amount of available methyl radicals decrease and causes a reduction in methyl substituted products which form by methyl transfer reaction. Number of collisions can also increase the probability of bond breakage. In the next section effects of operating and deign parameters will be discussed.

A simple graphical reaction mechanism which contains both decomposition of methyl radicals to hydrogen and hydrodeoxygenation reaction is shown in Fig. 2.

Fig. 2 A simple reaction mechanism of in situ hydrogen production and hydrodeoxygenation reaction.

3.2. Reaction network

The reaction network for anisole upgrading in the corona plasma reactor, based on the plasma reaction mechanisms and observed distribution of products, is proposed, as shown in Fig. 3. As may be seen from this figure, phenoxy radical is the primary intermediate radical which reacts with H radical, giving phenol as a primary intermediate product.^{30–32,34–38,45,65} Phenol is converted to benzene and methylphenols via electron-attack dissociation and free-radical reactions. During the upgrading of anisole in the DBD plasma reactor, three four reaction routes occur: demethylation, hydrogenolysis, and transalkylation (methyl transfer) and methyl decomposition. Since the breaking of C₆H₅O–CH₃ bond (reaction (R2)) is easier than the other bonds, demethylation of anisole is the primary reaction leading to formation of phenol (reaction route: (R2) + (R6)), while only a small fraction of anisole is converted to 4-methylanisole via transalkylation (reaction route: (R4) + (R7)). Further, phenol is converted to benzene via the direct hydrogenolysis (reaction route: (R8) + (R14)) and to methylphenols (especially 2-methylphenol) via the transalkylation. Toluene can be formed via methyl transfer to phenyl radicals (R14) and can proceed by methyl transfer for xylene production (R15). Therefore, we propose that during the upgrading of anisole in the DBD plasma reactor, the hydrodeoxygenation of anisole involves three main steps: (1) demethylation of anisole to phenol; (2) transalkylation of methyl with phenol resulting methylphenols; and (3) hydrogenolysis of phenol to benzene.

Fig. 3 Reaction network for the upgrading of anisole in corona plasma reactor.

3.3. Effect of operating and design parameters

3.3.1. Effect of pulse repetition frequency (PRF). As illustrated in Fig. 4, increasing the PRF generally increases the anisole conversion. Furthermore, rising pulse frequency at a constant voltage leads to the higher discharge power, according to eqn (1). As can be seen, with an increase in the PRF from 5 to 20 kHz, the conversion of anisole and discharge power increases from 3.5% to 37.8% and 37 to 60 W, respectively. The number of pulses in a second has been named as PRF. Changing of this parameter can alter the ascent and decent time of pulse voltage. Increasing the PRF leads to decrease in ascent and descent times. Consequently, the reaction would be improved due to promotion effect on each pulse by the generated energetic species in the former pulse discharge.⁶⁶ Therefore, as a result of stronger collisions between activated species and electrons with anisole molecules the conversion would increase.

Fig. 4 Anisole conversion, degree of deoxygenation and discharge power as a function of PRF.

Moreover, the amount of generated energetic species is affected by PRF. By increasing frequency the micro discharges become stronger and causes an enhancement in the energy and number of electrons and reactive species.^17,18,26 This can increase the probability of electron impact dissociation reactions involving ionization, excitation and dissociation of gas molecules. It can be concluded that, the PRF improves the performance of the corona discharge for upgrading of anisole. However, experiments using a higher PRF than 20 kHz is prevented due to the limitations of the pulse generator.

In Fig. 5, the selectivity of product are presented as a function of PRF. Phenol is the main product in all experimental conditions. Since the breaking of C₆H₅O–CH₃ bond is easier than the other bonds, demethylation of anisole is the primary reaction leading to formation of phenol. Anisole and phenol are converted to methylphenols (especially 2-methylphenol) via transalkylation reaction. It is also evident that the maximum selectivity of phenol (42.3%) and BTX (20.6%) is obtained at the frequency of 20 kHz. The variation trend of selectivity of methyl substituted products as a function of PRF is completely opposite to that of phenol and BTX. The reason is attributed to increasing demethylation and demethoxylation reaction by increasing frequency on one hand and decreasing transalkylation reaction on the other hand. The later effect is believed to be the result of less available methyl radicals due to increasing methyl decomposition by increasing PRF. Indeed, at higher frequencies, the formed methyl and methoxyl radicals from demethylation and demethoxylation reaction decompose more to hydrogen and results in the reduction of methyl transfer to aromatic ring.

Fig. 5 Products selectivity as a function of PRF.

These results are in accordance with findings made by other authors about methane decomposition in plasma reactors and the effect of PRF on its conversion.²⁶ As shown by Khalifeh et al.²⁶ increasing PRF caused an increase of methane decomposition to hydrogen which is due to more effective collision that occurred in the discharge zone.

3.3.2. Effect of gap distance. As discussed in previous section, the highest conversion of anisole is obtained at PRF of 20 kHz (Fig. 3). Therefore, to study the effect of gap distance in anisole upgrading, experiments were performed using 20 kHz of PRF with gap distances between pin electrode and liquid surface ranging from 1–15 mm. Fig. 6 shows the relationship between output parameters and gap distance. As the figure shows, by increasing gap distance, anisole conversion initially increases, reaches a maximum of 37.8%, and then decreases. For reactors with large electrode gaps (10–15 mm), the electric field becomes weak leading to generation of small discharge current and few streamers which causes a reduction of active species and consequently the effective collisions. For small gaps less than 5 mm, the spark forms which decreases the homogeneity and the area of treatment and increases the discharge power. Spark formation hinders the radial propagation of streamers on the liquid surface which reduces the conversion. Radial propagation of streamers and its effect on conversion is completely explain in the Section 3–4. This result is in consistent with the results in other literature.⁶⁷

Fig. 6 Anisole conversion, degree of deoxygenation and discharge power as a function of gap distance.

According to Fig. 7 variation of the gap distance can change the product distribution. Phenol selectivity increases form 29.7% to 45.5% when the gap distance rises from 1 to 15 mm. The highest BTX selectivity (21.6%) was achieved at 10 mm gap. However simultaneous decreases of methyl substituted products was recorded. This can be explained considering the results of previous studies which used pin to plate corona reactor for decomposition of methane to hydrogen.

Fig. 7 Products selectivity as a function of gap distance.

Aleknaviciute et al.⁶¹ carried out the experimental study of methane decomposition to hydrogen by using a pin to plate corona under different gap distances. These results showed that methane decomposition and subsequently hydrogen generation increased by increasing the gap distance. A similar result is also obtained by Kado and his team.⁶² It was found that in a pulsed corona reactor by increasing the electrode distance from 0.5 to 10 mm the conversion of methane increased from 30 to 90%. They mentioned that this can improve the collision efficiency between methane and electrons and increase the residence time of methane molecules. In the light of these premises, increasing gap distance reduces the transalkylation reaction due to increasing decomposition of available methyl radicals which causes a drop in methyl substituted products and enhancement of phenol and BTX. This combination of products leads to a sharp enhancement in deoxygenation degree by increasing the gap distance from 1 to 5 mm and after that it reaches a steady value when the gap distance is more than 5 mm.

3.3.3. Effect of pin number. There are a number of publications investigating the effect of pin number in corona plasma. Chen et al.⁶⁸ evaluate the degradation efficiency of pulsed corona discharge for treatment of contaminated water. They examined two configurations of needle and multi needle electrodes. Their experimental results showed higher disinfection efficiency of multi needle electrode compared to one needle. We observed similar effect of the number of pins on anisole conversion. As can be seen in Fig. 8 the more the pin number is, the more is the conversion of anisole and discharge power. By applying a high voltage to needle electrode, a pulsed corona develops at its tip. Then it is lengthened and reaches the liquid surface and forms discharge channels.²⁹ Increasing the number of pin electrodes leads to formation of more discharge channels⁶⁸ which enhance the rate of ionization and probability of collisions between anisole and formed radicals with reactive species. Furthermore, as it was reported by Wang et al.⁶⁹ the injected energy to the reactor was enhanced by increasing the number of pin numbers. Under their studied conditions reactor with 7 pins has 0.006 J per pulse higher energy than reactor with 1 pin. As can be seen in Fig. 8 the conversion increases from 12.5% to 37.8% by increasing the pin number from t to 6.

Fig. 8 Anisole conversion, degree of deoxygenation and discharge power as a function of pin number.

Formation of more streamers due to higher number of pin electrodes results in filling more space between pin electrodes and liquid surface. Indeed, the residence time of generated radicals or gaseous materials becomes higher which causes an increase in the number of collisions.^67,69 Subsequently the generated radicals from demethylation or demethoxylation reactions decomposed more to hydrogen or other radicals leading to alteration of product selectivity (Fig. 9). This reduces the available methyl radicals for transalkylation reaction. Therefore, the higher the pin number is, the lower is the methyl substituted products. Phenol selectivity increases from 28.8% to 42.3% by increasing pin number from 1 to 6. The highest BTX selectivity was also obtained when the reactor contain 6 pin electrodes. These mentioned effects of pin number on the selectivity of products result a direct effect of this parameter on the degree of deoxygenation. Deoxygenation degree rises from 13% to 32% by increasing the pin number from 1 to 6.

Fig. 9 Products selectivity as a function of pin number.

3.3.4. Carrier gas flow rate. The percentage conversion of anisole and other parameters as a function of the Ar flow rate is demonstrated in Fig. 10. The enhancement is observed for both conversion and discharge power as the Ar flow rate is increased from 100 to 900 ml min⁻¹. This effect is more significant when the flow increases from 100 to 500 ml min⁻¹ and causes a 270% enhancement in anisole conversion. This tendency can be clarified by considering easier ionization of gaseous mixture when more Ar is presented in the discharge zone. Ar has lower ionization energy than other gaseous products. Therefore increasing Ar flow rate enhances the rate of ionization in the reactor which increase discharge power and the amount of reactive species. This can facilitate transformation of Ar molecules into their metastable state of Ar* as consequence of effective collisions between Ar molecules and energetic electrons.^70,71 Moreover, As a consequence of more Ar gas molecules in the discharge the possibility of collisions between these activated Ar molecules (Ar*) and anisole molecules become higher, which causes higher anisole conversion.

Fig. 10 Anisole conversion, degree of deoxygenation and discharge power as a function of carrier gas flow rate.

The selectivity of phenol and BTX peaked at Ar flow rate of 300 ml min⁻¹. As shown in Fig. 11 phenol selectivity increases from 31.5% to a maximum of 42.3% and then decreases to 29.4%. This trend is due to two contrasting effects of increasing Ar flow rate on the rate of transalkylation reaction. On one hand, increasing Ar flow rate increases the rate of ionization and number of collisions in the reactor which leads to increase of demethylation and demethoxylation reactions. Moreover, it results more decomposition of methyl and methoxyl radical which are formed in the reactor. This can decrease the available methyl radicals to perform transalkylation reaction. On the other hand, when the Ar flow rate increases from 300 to 900 ml min⁻¹, the shorter residence time, which can arise from the higher Ar flow rate, can decrease the collision opportunities between methyl and methoxyl molecules and reactive species such as electrons, leading to the more available radicals to perform transalkylation reaction. The less significant effect of Ar flow rate on anisole conversion at higher flow rate can be ascribed to this mentioned negative effect. Increasing of Ar flow rate had no significant influence on the selectivity of tetramethylphenols. Additionally, the selectivity of methylanisoles almost remains unchanged.

Fig. 11 Products selectivity as a function of carrier gas flow rate.

3.3.5. Effect of plate electrode diameter. The effect of plate diameter on conversion, discharge power and deoxygenation are presented in Fig. 12. According to experimental results increasing the plate electrode diameter results in higher conversion and discharge power. This effect is rather steep at diameters between 20 and 30 mm, then it is roughly constant (67%) when it changes from 30 to 35 mm. The reason is attributed to increasing the leader area as a result of plate diameter enhancement.

Fig. 12 Anisole conversion, degree of deoxygenation and discharge power as a function of plate electrode diameter.

When the plasma channels reaches to the liquid surface, they propagate out radially over the surface of liquid. During the propagation many side arms and branches are formed which make a uniform and radial distribution of leaders on the liquid surface (see Fig. 1).⁷² Therefore, the probability of collisions between energetic species with feed molecules increases which caused as increase in the conversion and discharge power. However, as the distance from the core increases more, the energy density of streamers become lower which consequently decrease the rate of ionization and make the streamers narrower.⁷³

Plate diameter also has effect on the selectivity of product (Fig. 13). As mentioned, higher plate diameter results in increasing ionization rate which can increase the breaking of C₆H₅O–CH₃ bond as the weakest bond in anisole chemical structure. Moreover, the hydrogenolysis of phenol to benzene is improved by increasing the plate electrode diameter. This also has effect on the generated methyl and methoxyl radicals and leads to more decomposition of these radicals. Subsequently the methyl substituted products are reduced since there are less amount of methyl radicals in the reaction zone. As illustrated in figure the selectivity of methyl substituted products are higher when smaller plate diameter were used. The opposite trend is observed for phenol and BTX. Phenol selectivity increases from 34.5% to 47.3% as the plate diameter increased from 20 to 35 mm. Moreover, an increase of 48% in BTX selectivity is observed by changing the plate diameter.

Fig. 13 Products selectivity as a function of plate electrode diameter.

4. Conclusion

An experimental study was conducted on the upgrading of anisole using in situ generated hydrogen in the pin to plate pulsed corona plasma reactor at atmospheric pressure. The influences of various operating and design parameters such as the carrier gas flow rate, pulse repetition frequency, number of pin electrodes, gap distance and plate electrode diameter on the performance of corona plasma reactor were successfully investigated. It was found that:

(1) The highest anisole conversion of 68% was obtained with optimal pin number of 6, carrier gas flow rate, plate electrode diameter and gap distance of 500 ml min⁻¹, 35 mm and 5 mm, respectively.

(2) Phenol is the most predominant product of anisole upgrading in corona reactor.

(3) Demethylation, transalkylation, hydrogenolysis and demethoxylation are the main chemical reactions which lead to phenol, BTX and methylphenols as the major products.

(4) Corona discharge is able to decompose methyl radical and produce the required hydrogen for upgrading process in situ.

(5) Results show that increasing the decomposition of generated methyl radicals leads to increasing phenol and BTX selectivity.

(6) Due to the low temperature and atmospheric pressure conditions no hydrogenation of aromatic ring was observed.

(7) Increasing the plate electrode diameter, pulse repetition frequency and number of pin electrode results in higher ionization of argon and enhance effective collisions in the discharge zone. Subsequently higher anisole conversion is obtained.

References

1. G. Yao, G. Wu, W. Dai, N. Guan and L. Li, Fuel, 2015, 150, 175–183.
2. M. Ishikawa, M. Tamura, Y. Nakagawa and K. Tomishige, Appl. Catal., B, 2016, 182, 193–203.
3. R. Ding, Y. Wu, Y. Chen, J. Liang, J. Liu and M. Yang, Chem. Eng. Sci., 2015, 135, 517–525.
4. Q. Guan, F. Wan, F. Han, Z. Liu and W. Li, Catal. Today, 2016, 259, 467–473.
5. M. Zhou, Y. Wang, Y. Wang and G. Xiao, J. Energy Chem., 2015, 24, 425–431.
6. W. Wang, K. Zhang, L. Li, K. Wu, P. Liu and Y. Yang, Ind. Eng. Chem. Res., 2014, 53, 19001–19009.
7. M. S. Jang, R.-S. Park, I.-G. Lee, J. M. Kwak, Y.-K. Park and C. H. Ko, Res. Chem. Intermed., 2016, 42, 3–17.
8. C. Chen, G. Chen, F. Yang, H. Wang, J. Han, Q. Ge and X. Zhu, Chem. Eng. Sci., 2015, 135, 145–154.
9. B. Yoosuk, D. Tumnantong and P. Prasassarakich, Chem. Eng. Sci., 2012, 79, 1–7.
10. N. Joshi and A. Lawal, Ind. Eng. Chem. Res., 2013, 52, 4049–4058.
11. H. Taghvaei, M. B. Hosseinzadeh, S. Rezazadeh, M. R. Rahimpour and A. Shariati, Chem. Eng. J., 2015, 281, 227–235.
12. Z. Li, M. Garedew, C. H. Lam, J. E. Jackson, D. J. Miller and C. M. Saffron, Green Chem., 2012, 14, 2540.
13. M. Saidi, F. Samimi, D. Karimipourfard, T. Nimmanwudipong, B. C. Gates and M. R. Rahimpour, Energy Environ. Sci., 2014, 7, 103.
14. M. Saidi, P. Rostami, M. R. Rahimpour, B. C. Gates and S. Raeissi, Energy Fuels, 2015, 29, 191–199.
15. M. Saidi, P. Rostami, H. R. Rahimpour, M. A. Roshanfekr Fallah, M. R. Rahimpour, B. C. Gates and S. Raeissi, Energy Fuels, 2015, 29, 4990–4997.
16. M. Saidi, M. R. Rahimpour and S. Raeissi, Energy Fuels, 2015, 29, 3335–3344.
17. M. B. Hosseinzadeh, S. Rezazadeh, H. R. Rahimpour, H. Taghvaei and M. R. Rahimpour, Chem. Eng. Res. Des., 2015, 104, 296–305.
18. H. Taghvaei, M. Kheirollahivash, M. Ghasemi, P. Rostami and M. R. Rahimpour, Energy Fuels, 2014, 28, 2535–2543.
19. H. Taghvaei, M. Kheirollahivash, M. Ghasemi, P. Rostami, B. C. Gates and M. R. Rahimpour, Energy Fuels, 2014, 28, 4545–4553.
20. M. R. Rahimpour, A. Jahanmiri, P. Rostami, H. Taghvaei and B. C. Gates, Energy Fuels, 2013, 27, 7424–7431.
21. Y. Xing, Z. Liu, R. Couttenye, W. Willis, S. Suib, P. Fanson, H. Hirata and M. Ibe, J. Catal., 2008, 253, 28–36.
22. Y. Xing, Z. Liu, R. Couttenye, W. Willis, S. Suib, P. Fanson, H. Hirata and M. Ibe, J. Catal., 2007, 250, 67–74.
23. H. Taghvaei, M. M. Shirazi, N. Hooshmand, M. R. Rahimpour and A. Jahanmiri, Appl. Energy, 2012, 98, 3–10.
24. A. Jahanmiri, M. R. Rahimpour, M. Mohamadzadeh Shirazi, N. Hooshmand and H. Taghvaei, Chem. Eng. J., 2012, 191, 416–425.
25. O. Khalifeh, H. Taghvaei, A. Mosallanejad, M. R. Rahimpour and A. Shariati, Chem. Eng. J., 2016, 294, 132–145.
26. O. Khalifeh, A. Mosallanejad, H. Taghvaei, M. R. Rahimpour and A. Shariati, Appl. Energy, 2016, 169, 585–596.
27. C.-H. Tsai, W.-J. Lee, C.-Y. Chen and W.-T. Liao, Ind. Eng. Chem. Res., 2001, 40, 2384–2395.
28. H. Taghvaei, A. Jahanmiri, M. R. Rahimpour, M. M. Shirazi and N. Hooshmand, Chem. Eng. J., 2013, 226, 384–392.
29. V. P. Belosheev, Tech. Phys., 1999, 44, 4.
30. A. V. Friderichsen, E.-J. Shin, R. J. Evans, M. R. Nimlos, D. C. Dayton and G. B. Ellison, Fuel, 2001, 80, 1747–1755.
31. H.-M. Yang, S. Appari, S. Kudo, J.-I. Hayashi and K. Norinaga, Ind. Eng. Chem. Res., 2015, 54, 6855–6864.
32. M. Nowakowska, O. Herbinet, A. Dufour and P.-A. Glaude, Combust. Flame, 2014, 161, 1474–1488.
33. K. Brezinsky, M. Pecullan and I. Glassman, J. Phys. Chem. A, 1998, 102, 8614–8619.
34. I. W. C. E. Arends, R. Louw and P. Mulder, J. Phys. Chem., 1993, 97, 7914–7925.
35. M. M. Suryan, S. A. Kafafi and S. E. Stein, J. Am. Chem. Soc., 1989, 111, 4594–4600.
36. J. C. Mackie, K. R. Doolan and P. F. Nelson, J. Phys. Chem., 1989, 93, 664–670.
37. M. H. Back, J. Phys. Chem., 1989, 93, 6880–6881.
38. R. H. Schlosberg, P. F. Szajowski, G. D. Dupre, J. A. Danik, A. Kurs, T. R. Ashe and W. I. Olmstead, Fuel, 1983, 62, 690–694.
39. I. W. C. E. Arends, R. Louw and P. Mulder, J. Phys. Chem., 1993, 97, 7914–7925.
40. M. S. Pecullan, Ph.D. thesis, Princeton University, 1997.
41. C. P. Kubiak, Preconversion catalytic deoxygenation of phenolic functional groups. Quarterly technical progress report, October 1-December 31, 1995, Purdue Univ. Dept. of Chemistry, Lafayette, IN (United States), 1996.
42. Y.-R. Luo, Comprehensive Handbook of Chemical Bond Energies, CRC Press, Boca Raton, 2007.
43. G. da Silva, C.-C. Chen and J. W. Bozzelli, Chem. Phys. Lett., 2006, 424, 42–45.
44. F. G. Bordwell, J. Cheng, G. Z. Ji, A. V. Satish and X. Zhang, J. Am. Chem. Soc., 1991, 113, 9790–9795.
45. B. Husson, M. Ferrari, O. Herbinet, S. S. Ahmed, P.-A. Glaude and F. Battin-Leclerc, Proc. Combust. Inst., 2013, 34, 325–333.
46. R. C. Runnebaum, T. Nimmanwudipong, D. E. Block and B. C. Gates, Catal. Sci. Technol., 2012, 2, 113.
47. R. C. Runnebaum, T. Nimmanwudipong, D. E. Block and B. C. Gates, Catal. Lett., 2010, 141, 817–820.
48. R. C. Runnebaum, R. J. Lobo-Lapidus, T. Nimmanwudipong, D. E. Block and B. C. Gates, Energy Fuels, 2011, 25, 4776–4785.
49. T. Nimmanwudipong, R. C. Runnebaum, D. E. Block and B. C. Gates, Energy Fuels, 2011, 25, 3417–3427.
50. S. J. Blanksby and G. B. Ellison, Acc. Chem. Res., 2003, 36, 255–263.
51. P. G. Wenthold and R. R. Squires, Int. J. Mass Spectrom. Ion Processes, 1998, 175, 215–224.
52. P. G. Wenthold, R. R. Squires and W. C. Lineberger, J. Am. Chem. Soc., 1998, 120, 5279–5290.
53. A. Indarto, J.-W. Cho, H. Lee, H. K. Song and J. Palgunadi, J. Rare Earths, 2006, 24, 513–518.
54. A. Indarto, D. R. Yang, J. Palgunadi, J.-W. Choi, H. Lee and H. K. Song, Chem. Eng. Process., 2008, 47, 780–786.
55. A. Indarto, J.-W. Choi, H. Lee and H. K. Song, Chin. Sci. Bull., 2008, 53, 2783–2792.
56. A. Indarto, N. Coowanitwong, J.-W. Choi, H. Lee and H. K. Song, Fuel Process. Technol., 2008, 89, 214–219.
57. A. Indarto, J. Chin. Inst. Chem. Eng., 2008, 39, 23–28.
58. A. Huang, G. Xia, J. Wang, S. L. Suib, Y. Hayashi and H. Matsumoto, J. Catal., 2000, 189, 349–359.
59. S. Delavari, N. A. S. Amin and H. Mazaheri, React. Kinet., Mech. Catal., 2014, 113, 557–573.
60. M.-W. Li, C.-P. Liu, Y.-L. Tian, G.-H. Xu, F.-C. Zhang and Y.-Q. Wang, Energy Fuels, 2006, 20, 1033–1038.
61. I. Aleknaviciute, T. G. Karayiannis, M. W. Collins and C. Xanthos, Energy, 2013, 59, 432–439.
62. S. Kado, Fuel, 2003, 82, 1377–1385.
63. A. Aziznia, H. R. Bozorgzadeh, N. Seyed-Matin, M. Baghalha and A. Mohamadalizadeh, J. Nat. Gas Chem., 2012, 21, 466–475.
64. A. B. Dongil, I. T. Ghampson, R. García, J. L. G. Fierro and N. Escalona, RSC Adv., 2016, 6, 2611–2623.
65. M. Pecullan, K. Brezinsky and I. Glassman, J. Phys. Chem. A, 1997, 101, 3305–3316.
66. S. L. Yao, E. Suzuki, N. Meng and A. Nakayama, Energy Fuels, 2001, 15, 1300–1303.
67. Y. Zhang, J. Zheng, X. Qu and H. Chen, Chemosphere, 2008, 70, 1518–1524.
68. C. W. Chen, H.-M. Lee and M. B. Chang, IEEE Trans. Plasma Sci., 2008, 36, 215–219.
69. T. C. Wang, G. Qu, J. Li and D. Liang, Vacuum, 2014, 103, 72–77.
70. L. T. Hsieh, W. J. Lee, C. Y. Chen, M. B. Chang and H. C. Chang, Plasma Chem. Plasma Process., 1998, 18, 215–239.
71. N. Rueangjitt, T. Sreethawong, S. Chavadej and H. Sekiguchi, Plasma Chem. Plasma Process., 2011, 31, 517–534.
72. G. R. Stratton, C. L. Bellona, F. Dai, T. M. Holsen and S. M. Thagard, Chem. Eng. J., 2015, 273, 543–550.
73. V. P. Belosheev, Tech. Phys., 2000, 45, 109–114.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra21417g

‡ Present address. Hamed Taghvaei: Department of Mechanical Engineering, College of Science and Engineering, University of Minnesota, 111 Church Street SE, Minneapolis, MN 55455, USA.

---