Sub and supercritical water reforming of n-hexadecane in a tubular flow reactor

J. of Supercritical Fluids 107 (2016) 723–732

Contents lists available at ScienceDirect

The Journal of Supercritical Fluids journal homepage: www.elsevier.com/locate/supflu

Sub and supercritical water reforming of n-hexadecane in a tubular flow reactor Y.M. Alshammari, K. Hellgardt ∗ Department of Chemical Engineering and Chemical Technology, Imperial College London, South Kensington Campus, London SW7 2AZ, UK

a r t i c l e

i n f o

Article history: Received 9 June 2015 Received in revised form 29 July 2015 Accepted 30 July 2015 Available online 31 July 2015 Keywords: Hydrothermal conversion Water gas shift reaction Hydrogen generation n-Alkane to 1-alkene ratio Tubular flow reactor Cracking of hydrocarbons

a b s t r a c t This work investigates the hydrothermal conversion of hexadecane, as a heavy hydrocarbon model for heavy oil upgrading, and syngas generation. This experimental analysis was carried out in tubular flow reactor where water and hexadecane react at 525–605 ◦ C and 150/220 bar under different residence times. Hydrogen and syngas formation was observed and quantified under the current conditions without significant formation of coke in the reactor. Residual hexadecane was also analysed for its contents of cracking products using GCMS. Investigating the temperature and residence time effects enabled determining the reaction kinetic data from which the activation energy, Ea , was determined to 263 kJ/mol (at 220 bar) and 202 kJ/mol (at 150 bar). The determined kinetic data were compared with previously reported results on high pressure pyrolysis of hexadecane and other hydrocarbons. The effects of increasing the water density by increasing the reactor pressure to (220 bar) was found in particular to enhance heat and mass transfer leading to a higher degree of conversion at lower temperatures, and increasing the ratio of n-alkane to 1-alkenes via in situ hydrogenation. Cracking was found to follow the free radical mechanism under both sub and supercritical conditions, producing nearly equi-molar distribution of nalkanes and 1-alkenes, C7 –C13 , under lower conversions. Increasing the reaction temperature enhances the formation of 1-alkene via ␤-scission, while increasing the pressure increases the formation n-alkanes via H-abstraction. In addition, it is found that the hydrothermal conditions have inhibited the formation of higher molecular weight hydrocarbons, C16+ , via addition reactions. Results show the potential for a continuous process for hydrogen generation from heavy hydrocarbons using sub and supercritical water with minimised carbon formation. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Hydrogen is a promising energy carrier which is commercially produced via steam reforming of methane as well as the partial oxidation and autothermal reforming of hydrocarbons [1]. The increase of heavy hydrocarbon resources requires new and more sustainable processes for hydrogen and syngas generation from hydrocarbon resources. Underground heavy hydrocarbon resources, in particular, require large quantities of hydrogen for hydrocracking and separation of sulphur and metals contents [2]. It would, hence, be desirable if hydrogen can be produced directly from heavy low-grade hydrocarbons as a clean energy carrier for future fuel cell power generation or other potential chemical and energy applications. Carbon rejection processes (pyrolysis,

∗ Corresponding author. E-mail addresses: [email protected] (Y.M. Alshammari), [email protected] (K. Hellgardt). http://dx.doi.org/10.1016/j.supflu.2015.07.037 0896-8446/© 2015 Elsevier B.V. All rights reserved.

coking) are well established routes for heavy oil upgrading, yielding hydrogen as well as lighter oil fractions [2]. Previous research on gas generation via high pressure pyrolysis of hydrocarbons underground was reported [3–11] aiming to understand the stability of oil under high pressure hydrothermal conditions over an extended period of time. Other more recent studies showed the potential for hydrogen production through the catalysis of steam reforming of heavy hydrocarbons, reporting different key kinetic data [12–15]. The gasification of hydrocarbons in supercritical water (SCW) has been proposed due to its unique physicochemical properties. In particular, the increase in water temperature to critical point causes a significant reduction of dielectric constant, while increasing the ionic product (Kw ) and the concentration of H+ and OH− ions [16]. This makes SCW particularly useful for many chemical applications as a green solvent for single phase reactions, and as a catalyst for acid-base organic reactions. Supercritical water gasification of nhexadecane, as a heavy oil model, was reported by Arai et al. [17] where hydrogenation of hydrocarbons via partial oxidation in SCW has been proposed as promising process for in situ upgrading of

724

Y.M. Alshammari, K. Hellgardt / J. of Supercritical Fluids 107 (2016) 723–732

heavy oil. They reported that in situ hydrogenation, which uses H2 generated via the water gas shift reaction (WGS), proceeds faster than conventional hydrogenation in SCW media. This study was followed by many other studies which showed the potential of nhexadecane gasification in SCW under catalytic partial oxidation conditions [10,18,19]. For instance, Watanabe et al. [18] showed that the catalytic effects of NaOH and ZrO2 include the significant enhancement of H2 yield, and decomposition of intermediates into CO. Although such studies are useful in understanding various parametric effects, they were all carried out in catalytic batch-type reactors under long reaction times, limiting their potential for commercial scale-up. Other studies on SCW gasification in tubular flow reactors have only made use of either biomass model compounds [20–28] or light hydrocarbons [29–31]. The generation of hydrogen via hydrothermal gasification of heavy hydrocarbons in tubular flow systems remains an underexplored area. Flow systems are useful in not only obtaining reproducible kinetic data but they also offer more potential for commercial applications as they reduce the processing time, control product distribution while enabling better understanding of the reactor behaviour under different flow regimes. Our previous work has reported the thermodynamic analysis of the hydrothermal gasification of hexadecane identifying optimal conditions for maximising the yield of hydrogen for both, hydrothermal gasification and partial oxidation scenarios [32]. This enabled understanding the effects of the reactor isothermal temperature, pressure, water to carbon ratio, and oxygen to carbon ratio on the theoretical molar yields of produced gaseous species at equilibrium with a complete conversion of the hydrocarbon feedstock. In this study we report a new set of experimental results on the gasification of n-hexadecane, under hydrothermal conditions where the effects of reactor temperature, pressure and residence time, are investigated in sub- and supercritical water, as a potential process for clean production of hydrogen from underground heavy hydrocarbons resources. Complete product yield and first order kinetic analysis are derived from continuous flow experiments.

2. Materials and methods Hexadecane (99% ALFA AESAR) was purchased from VWR and used as received. De-ionised water, which was de-aerated with Argon prior to pumping, was used as hydrothermal reaction medium. The reactions were carried in a continuous tubular flow reactor system under hydrothermal conditions. The sketch and details of our experimental flow reactor systems were previously reported in other work [33]. The reactor used was made of Stainless Steel 316 (6.35 mm od, 0.54 mm id, and 260 mm length). Fig. 1 shows a sketch of the reactor used in this work. No catalysts were used in this study, aside from the potential catalytic effects of the reactor wall. The maximum tolerable pressure for the reactor, at the maximum potential experimental temperature (700 ◦ C), was determined in order to identify safe operating limits. The water in this experiment was pumped using an HPLC pump (WATERS 510) while the same syringe pump (100DM ISCO) was used to pump the hydrocarbon fuel (hexadecane) [33]. In order to reduce temperature gradients along the reactor, the water feed is preheated to the reactor temperature using an 1/8 tubing (3.175 mm od) positioned also inside the copper block before it mixes with n-hexadecane at the mixing point where TI-1 is recorded [33]. Then, the mixed water-n-hexadecane fluid flows and reacts along the reactor tube. The thermocouple (TI-2) is position right at the exit of the reactor tube in order to measure the reactor peak temperature which is taken into consideration for this

Fig. 1. Reactor set-up sketch.

investigation. The mass spectrometer allows for analysis and quantification of produced gaseous species. The masses used to detect H2 , CO, CO2 , and CH4 , are 2, 28, 44, and 16, respectively, which are the molecular ions that each gas would be fragmented into through ionisation. The mass spectrometer was calibrated using a BOC calibration gas mixture containing known amounts of H2 , CO, CO2 , and CH4 . As the system does not contain any Nitrogen, facile detection and quantification of CO was possible. Using a three way valve, the gas flowrate may be measured using a bubble flowmeter connected to the phase separator through a secondary line. The GC–MS used, for analysing oil samples, was a Varian STAR 3400 CX Gas Chromatography, equipped with a fused silica capillary column (30 m × 0.32 mm) operated with a helium gas carrier at 13 psi, providing 1 mL/min. The GC is connected to a mass spectrometer (Varian SATURN 2000 GC/MS) through a heated transfer line at (300 ◦ C) in split mode. The GC oven is programmed to raise the temperature from 38 ◦ C by (5 ◦ C/min) to (100 ◦ C), followed by (15 ◦ C/min) to (320 ◦ C) which was held for (10 min). The mass spectrometer was operated with a mass range of (40–650 m/z). n-Alkane standard (C8 –C20 ), purchased from Sigma–Aldrich, was used to authenticate produced alkanes and consequently its alkenes based on the mass spectra. Measured amounts of naphthalene were added to each sample, as an external standard, in order to quantify each produced species based on its relative response. The response factor for 1-alkenes was assumed to be equal to its corresponding n-alkanes. 3. Results and discussion 3.1. Gasification reactions Gasification reactions are very complex but the major reactions involved in this process have been widely described elsewhere in the literature [30]. They include the steam reforming, water gas shift, methanation, partial oxidation, when oxidant is used,

Y.M. Alshammari, K. Hellgardt / J. of Supercritical Fluids 107 (2016) 723–732

CO + H2 O = CO2 + H2

(2)

H(25 ◦ C) = −41.39 kJ/mol

4.00 H2

11.60

(3)

H(25 ◦ C) = 1811 kJ/mol CO + 3H2 = CH4 + H2 O

11.20

1.00

11.00

(b)

550

570 590 Temperature [oC]

12.00

610

0.00

4.50

CO2

4.00

11.80

3.50 3.00

CH4

11.60

2.50

H2

2.00

11.40

1.50 1.00

11.20

0.50

CO 550

570 590 Temperature [oC]

610

0.00

(4)

(5)

H(25 ◦ C) = 373.590 kJ/mol CH 4 = C + 2H 2

2.00

Fig. 2. Equilibrium molar yield of hexadecane hydrothermal gasification (28 wt%) at (a) 150 bar and (b) 220 bar.

H(25 ◦ C) = −206.12 kJ/mol C16 H34 = 17H2 + 16C

3.00 CH4

11.40

11.00

C16 H34 + 32H2 O = 16CO2 + 49H2

5.00

CO2

CO

(1)

H(25 ◦ C) = 2473.17 kJ/mol

12.00 11.80

mol/mol(C16)

C16 H34 + 16H2 O = 16CO + 33H2

(a)

mol/mol(C16)

and pyrolysis reactions. Since the current study focuses on the high pressure steam reforming, partial oxidation reactions will not be discussed as a source of hydrogen generation. There are two primary reactions responsible for the production of Hydrogen, which are steam reforming (Eq. (1)) and water gas shift (Eq. (2)), and they may be combined under certain conditions as shown in Eq. (3). The heat of reactions was estimated using Aspen HYSYS and the Peng–Robinson equation of state as reported previously [32]. The steam reforming of hexadecane (Eq. (1)) is a highly endothermic reaction, whereas the water gas shift reaction (Eq. (2)) is slightly exothermic and when the two reactions are combined together they become less endothermic (Eq. (3)). Hydrogen may be consumed by side reactions which include the exothermic methanation reaction (Eq. (4)) favoured under low temperature conditions. In addition, hydrogen may be produced through the thermal pyrolysis of hexadecane (Eq. (5)), which is another highly endothermic reaction producing hydrogen, carbon and lighter hydrocarbons.

725

(6)

H(25 ◦ C) = 74.9 kJ/mol 3.2. Equilibrium calculations Using the method of direct minimisation of Gibbs free energy and the Peng–Robinson equation of state, explained in our previous work [32], equilibrium analysis of hydrothermal gasification of hexadecane was undertaken. This was carried out under the desired experimental conditions (565–605 ◦ C, 150–220 bar, 33 wt% C16 H34 ) in order to estimate the maximum theoretical yields of hydrogen and other gaseous species (CH4 , CO, CO2 ) in sub (150 bar) and supercritical (220 bar) water. The equilibrium composition of the gases resulting from subcritical water gasification of hexadecane is shown in Fig. 2a where the yield of hydrogen starts to rise from 2.54 mol/mol, at 565 ◦ C, to reach 3.40 mol/mol, at 605 ◦ C. The yield of CO is considerably lower compared with the yields of other gaseous species, and remains less than 0.26 mol/mol. Furthermore, the yield of methane shows a slight decrease with increasing temperature, and this is associated with a slight increase in the yield of CO2 . These observations suggest that increasing the temperature inhibits methanation reactions (Eq. (4)), while they promote the steam reforming and, to a certain extent, water gas shift reactions.

In contrast, increasing the pressure to 220 bar would shift the gasification process into the supercritical water region, which shows a slight effect on the yield profile of gaseous species. This is illustrated in Fig. 2b, where the yield of hydrogen slightly decreases from 2.54 mol/mol (at 565 ◦ C and 150 bar) to 2.02 mol/mol at (565 ◦ C and 220 bar). In addition, methane, CO2 , and CO, also show slightly decreased levels compared with subcritical conditions. This shows that increasing the pressure would inhibit the gasification reaction and reduce the overall process gasification efficiency. Hence, it is concluded that increasing the temperature promotes the steam reforming and reversed methanation reactions, and it inhibits forward methanation reaction which increases the overall gasification efficiency. On the other hand, increasing the pressure inhibits the gasification reaction and reduces the overall process efficiency. It is also found that it is possible to achieve, under the investigated experimental conditions, a complete conversion of hexadecane to hydrogen and methane (hythane), with very low CO fraction, which may be supplied to fuel cell or internal combustion applications if CO2 can be efficiently captured for alternative applications.

3.3. Experimental gas analysis The effects of temperature, and residence time on the yield of different gaseous species (H2 , CO, CH4 , CO2 ) were investigated in sub and supercritical water media (150 bar and 220 bar). The experimental yield and composition of gaseous species is compared with the equilibrium analysis, explained above, in order to establish how close to equilibrium the system can be driven at relatively short residence times.

726

Y.M. Alshammari, K. Hellgardt / J. of Supercritical Fluids 107 (2016) 723–732

(a) 1.20E-03

mol/mol(C16)

1.00E-03 8.00E-04

CH4 CO

6.00E-04

H2

4.00E-04 2.00E-04

CO2 0

2

4

6

8

10

12

Residence time [s]

(b) 2.00E-03 CH4

mol/mol (C16)

1.60E-03

CO

1.20E-03 8.00E-04

H2

4.00E-04 0.00E+00

CO2 0

2

4

6

8

10

12

Residence time [s]

(c)

6.E-03

mol/mol(C16)

5.E-03 4.E-03 3.E-03 2.E-03

CH4

1.E-03 0.E+00

CO2 0

2

4

CO H2

6 8 10 Residence time [s]

12

Fig. 3. Temporal variation of molar production of gaseous species under subcritical water conditions (150 bar) at (a) 565 ◦ C, (b) 585 ◦ C, and (c) 605 ◦ C.

3.3.1. Temperature-residence time effects at 150 bar The maximum gasification temperature achieved in this study was 605 ◦ C above which coke formation would result in blocking the reactor system at the maximum residence time (11.09 s). Increasing both the reactor residence time, at different temperatures, has shown significant effects on the gaseous molar production rate and relative yield as shown in Fig. 3. Hydrogen was the most sensitive gas to the variation in reactor temperature. By looking at Fig. 3, it is found that the gaseous yield is in the order of CH4 > CO > H2 > CO2 at all residence times (11.09–3.80 s). Comparing the experimental data in Fig. 3 with the thermodynamic equilibrium data in Fig. 2 shows that the consumption of CO increases to reach closer to equilibrium as the residence time increases. The thermodynamic equilibrium data shows that the consumption of CO is the lowest when looking at 565 ◦ C due to its consumption by steam in the water gas shift reaction which causes

the higher yield of hydrogen and CO2 . The experimental data shows that increasing the residence time tends to support this hypothesis. The hydrogen yield shows continuous increase as the residence time increases from 3.80 to 11.09 s. This increase also becomes sharper at higher temperatures reaching 3.16 × 10−3 mol/mol (26.04 mol%) at 605 ◦ C and 11.09 s, as shown in Fig. 3c. It is also found that the yield of hydrogen diminishes as the residence time is shortened (3.8 s) for all reactor temperatures selected (605–565 ◦ C). The production of hydrogen shows the occurrence of the reforming reaction between hexadecane and steam as well as the water gas shift reaction, Eqs. (1)–(3) which offers a new and promising route for the application of sub and supercritical water for hydrogen generation. Nonetheless, the hydrogen yield reported in other experimental studies, is much higher compared with the currently obtained yield. For example, Susanti et al. [30] showed that hydrogen yield, during the noncatalytic supercritical water gasification of iso-octane, was maximised up to a value of 12 mol/mol, which is very close to the equilibrium value (13 mol/mol) and this was obtained at a temperature of 763 ± 2 ◦ C, a pressure of 250 bar, and a residence time of 91 s. This is not surprising because of they have a factor of 10 residence time compared with the residence time in our study, and they also have a higher temperature. Additionally, Pinkwart et al. [34] showed that hydrogen yield during the supercritical water gasification of n-decane (20 vol%) approaches 0.05 mol/mol, at a temperature of 550 ◦ C, a pressure of 250 bar, and a residence time of 31 s. These results show that the main parameters affecting the yield of hydrogen are the feed carbon number, residence time, feed concentration, oxygen concentration, and more importantly, the reactor temperature. In the current study, the yield of hydrogen yield did not exceed 3.16 × 10−3 mol/mol, which equates to 26.04 mol% in the gas phase, at 605 ◦ C and 150 bar. This corresponds to only 0.10% of the equilibrium yield (3.40 mol/mol). This may be justified by the very low residence time (11.09 s) compared with reported literature, in addition, to the use of a hydrocarbon with a significantly higher carbon number (C16 ). The formation of CO results from the steam reforming reaction of hexadecane. At 150 bar, the production of CO shows initial increase as the residence time increases reaching a maximum of 9.94 × 10−4 mol/mol(C16 ) and 1.17 × 10−3 mol/mol(C16 ) at 565 ◦ C and 585 ◦ C respectively, as shown in Fig. 3a and c. The production of CO, then, declines as the residence time increases to 11.09 s, which suggests its enhanced consumption by the water gas shift reaction as conversion increases at higher residence times. This also agrees with the molar production of hydrogen explained above which continues to increase at higher residence time, with decreasing levels of CO. It was reported that increasing the residence time would decrease the molar yield of CO due to the enhancement in the water gas shift reaction consuming CO and maximising the yield of hydrogen as shown in Eq. (2) [30]. This is illustrated by looking into the relative yield of CO, shown in Fig. 3, where the CO yield shows continuous decrease with increasing residence times. Combining the observation of hydrogen and CO yields it is suggested that the role of water gas shift reaction is key in maximising the yield of hydrogen during this type of gasification processes. The maximum yield of CO is 3.10 × 10−3 mol/mol obtained at 605 ◦ C, 150 bar, and 11.09 s, which corresponds to 1.19% of the equilibrium value 0.26 mol/mol. The production of methane is found to be the largest compared with the production of other gaseous species. The yield of methane also shows a comparable trend to the yield of CO as the residence time increases. Methane may be formed by the endothermic cracking of hydrocarbons as well as the exothermic hydrogenation of CO and C, as shown in Eqs. (4) and (6). Previous reports have attributed the increase in methane yield with increasing temperatures to the non-equilibrium methanation reactions at very low residence times [31]. In this study, however, the rise of methane yield as the

Y.M. Alshammari, K. Hellgardt / J. of Supercritical Fluids 107 (2016) 723–732

(a)

2.00E-03

mol/mol(C16)

1.60E-03

CH4

1.20E-03 8.00E-04

CO

4.00E-04

H2

0.00E+00

0

2

4

CO2

6 8 10 Residence time [s]

12

14

16

(b) 2.00E-03 CH4

mol/mol(C16)

1.60E-03 1.20E-03

CO

8.00E-04

H2

4.00E-04 0.00E+00

CO2 0

2

4

6 8 10 Residence time [s]

12

14

16

(c) 2.50E-03 CH4

mol/mol(C16)

2.00E-03 1.50E-03

CO

1.00E-03 H2

5.00E-04 0.00E+00

CO2 0

2

4

6

8

10

12

14

16

Residence time [s] Fig. 4. Temporal variation of molar production of gaseous species under supercritical conditions at (a) 565 ◦ C, (b) 585 ◦ C, and (c) 595 ◦ C.

727

7.49 × 10−4 mol/mol(C16 ) at 585 ◦ C, and 8.14 × 10−4 mol/mol(C16 ) at 595 ◦ C and 13.89 s, as shown in Fig. 4b and c. The trend of the CO relative yield remains similar to its trend in the subcritical region, though at a much higher yield 41.93–36.01 mol%, and it decreases with increasing residence time as a result of the enhanced forward water gas shift reaction. However, its molar production continues to increase, without reaching a maximum, for the investigated current range of residence times (4.70 s–13.89 s) which shows the low activity of the water gas shift reaction compared with the subcritical state where a maximum production of CO was observed at 6.57 s at 585 ◦ C and 565 ◦ C. The higher yield of CO at 220 bar shows that the syngas produced under these conditions has a lower H2 to CO ratio. Resende et al. [27] investigated the effects of water density, which increases with increasing pressure, on the hydrogen yield during noncatalytic gasification of cellulose and lignin. They concluded that the effect of the water density on the water gas shift reaction is unclear and it depends in the range of water density being studied. They also stated that higher water densities would lead to more water molecules per unit volume, leading to more collisions, and thus increasing the reaction rate. The increase in pressure to 220 bar, also caused the molar yield and production of methane to decrease slightly compared with 150 bar. Increasing the residence time showed more observable effects on increasing the methane yield between 7.35 × 10−4 mol/mol(C16 ) (37.32 mol%) to 1.95 × 10−3 mol/mol (46.57 mol%), at 595 ◦ C. The decrease in temperature, however, showed less effect on the yield of methane, compared with its effects at 150 bar, that the methane reaches nearly an equal maximum of around 1.53 × 10−3 (nearly 43 mol%) at 565 and 585 ◦ C when the residence time is 13.89 s. The decrease in the yield of methane with increasing residence time, may be attributed to the enhanced reverse methanation reaction. This reverse methanation is also expected to contribute to the enhanced yield in hydrogen production. The yield of CO2 shows a slight rise with increasing the pressure to 220 bar which agrees with the thermodynamic prediction in Fig. 2. However, its yield continues to remain almost constant at varying residence times and the lowest among all other gaseous species, <4 mol%. The yield of CO2 rises from 5.10 × 10−5 to 8.10 × 10−5 mol/mol(C16 ) as the pressure is raised from 150 bar ( = 5.09 s) to 220 bar ( = 5.69 s) at 585 ◦ C. 3.4. Oil phase analysis

temperature increases may be attributed to the non-equilibrium endothermic pyrolysis of heavier hydrocarbon gases C2 –C4 . The maximum yield achieved in this study is 5.45 × 10−4 mol/mol at 605 ◦ C, 150 bar, and 11.09 s. This value corresponds to 0.01% of the equilibrium yield of methane under the same conditions. The high methane yield obtained in this study shows that it is a very promising route to achieve commercial production of hythane providing cleaner and more efficient combustion. 3.3.2. Effects of supercritical pressure on gaseous yield Looking at Fig. 4, where the effects of the residence time on the yields of gaseous species are shown under supercritical water condition (220 bar), it is evident that the increase in the yield of hydrogen with time is not as significant as it is in the subcritical region. This minor increase is in agreement with the thermodynamic analysis shown in Fig. 2 which shows that the yield of hydrogen decreases as the pressure increases from 150 bar to 220 bar, and this is related to the activity of water gas shift reaction. Increasing the temperature to 585 ◦ C and 595 ◦ C, at 220 bar, has not significantly improved the yield of hydrogen which reaches

3.4.1. GC–MS qualitative analysis In addition to the gas analysis explained above, residual oil samples were analysed using GCMS as shown in the chromatograms in Fig. 5. The chromatograms show different product species resulting from the cracking of hexadecane at different retention times. The mass spectrometer analysis was used to initially identify each formed species before final authentication through their retention times using an n-alkanes GC standard solution. For both the supercritical and subcritical systems, it was found that hexadecane cracks into smaller n-alkane and 1-alkene hydrocarbons ranging from C15 to C1 suggesting that the reaction followed the free radical chain mechanism, shown in Scheme 1, as proposed earlier by Rice and Herzfeld [35]. The free radical pyrolysis mechanism assumes that decomposition of single carbon bond, C–C, by the removal of one hydrogen atom producing a highly reactive radical that undergoes a series of reactions, ␤-scission, hydrogen abstraction, isomerisation, and addition reactions, before it terminates by combining with another free radical to a form a stable n-alkane. The presence of supercritical water may induce further reactions as shown in Scheme 2.

728

Y.M. Alshammari, K. Hellgardt / J. of Supercritical Fluids 107 (2016) 723–732

(a) Decomposition

n − C16 H 34 → R • + Rn• (b) Hydrogen Abstraction:

R• + n − C16H34 → R − H +•C16H33 (c) ß-Scission

pri − • C16 H 33 → pri − • C14 H 29 + C 2 H 4 sec − • C16 H 33 → C m H 2 m + •C n H 2 n +1 (d) Addition

C m H 2 m + •C16 H 33 → nC16+ (e) Termination

R • + R • → n − alkane

affects the n-alkane to 1-alkene ratio as discussed in Section 3.4.2. This also shows that hydrogen saturation is promoted under supercritical water conditions which increase the yield of n-alkanes while this is not observed under subcritical water conditions as shown in the chromatograms in Fig. 5. Formation of higher molecular weight hydrocarbons (>C16+ ) was reported earlier to occur through the addition step where the hexadecyl radical combines with an olefin to yield higher n-alkanes [4,37]. However, no formation of higher n-alkanes (>C16+ ), through the reaction of the hexadecyl radical with a 1-alkene, was observed in this study under both conditions, as this may have been inhibited by the presence of sub and supercritical water, which was also previously observed during steam reforming of n-hexadecane [38].

Scheme 1. Free radical pyrolysis of hexadecane, Tsuzuki [6].

(a) OH-Dissociation

H 2 O → 2H • + O • (b) Hydrogen-Saturation

Ri• , R•j + 2H • → R − H , Rn (c) Carbon Oxidation

Rn• + 2O • → Rn −1 + CO2 + H 2 Scheme 2. possible reactions for participation of oxygen during SCW pyrolysis of hydrocarbons.

Under the subcritical conditions in this study, the formation of 1-alkenes is found to be favoured over the formation of n-alkanes which suggests the low hydrogen availability for saturation. This is clearly shown by the low n-alkane peaks in Fig. 5a which suggests that the ␤-scission step occurs faster than the hydrogen abstraction step, and this agrees with the R-K mechanism [36]. During the ␤scission step, primary, and secondary, hexadecyl radicals undergo decomposition reactions to form smaller alkyl radicals and the corresponding 1-alkene. On the other hand, it is found that there are sharper peaks for nalkanes under the supercritical water conditions with significantly lower n-alkanes peaks under the subcritical water conditions. This

3.4.2. n-Alkane to 1-alkene ratio The molar ratio of n-alkanes to 1-alkenes (C C/C C) was determined at different reaction conditions as shown in Fig. 6. This ratio remains mostly less than unity showing the low hydrogen selectivity required to increase the yield of n-alkanes by saturating the double bond of their respective 1-alkenes. The increase in conversion, at higher temperatures and residence times, increases the C C/C C ratio as may be seen in Fig. 6a–c. This may be attributed to the increase in hydrogen selectivity at higher temperatures, which increases the yield of n-alkanes over their respective alkenes-1 by enhancing saturation reactions. In fact, a ratio higher than unity was achieved for C13 at a temperature of 585 ◦ C, and a residence time of 13.89 s, at which a conversion of around 55% was achieved under supercritical conditions (220 bar), as seen in Fig. 6a. This may be compared with previously reported work where the C C/C C ratio exceeded unity in the presence of injected oxygen which increases the hydrogen yield through partial oxidation in supercritical water compared with the lower ratio, less than 1, which was obtained under non-oxidative pyrolysis of n-hexadecane in SCW [17]. Furthermore, a high selectivity towards producing lower weight hydrocarbons (
Fig. 5. GCMS analysis of hexadecane reforming products at (a) 565 ◦ C; 150 bar; 11.09 s and (b) 565 ◦ C; 220 bar; 13.89 s.

Y.M. Alshammari, K. Hellgardt / J. of Supercritical Fluids 107 (2016) 723–732

729

Fig. 6. Variation of n-alkane to 1-alkene ratio for supercritical conditions (a) 13.89 s, (b) 7.09 s, (c) 4.7 s, and subcritical conditions (d) 11.09 s.

greater tendency for large radicals to decompose into smaller radicals [7,8]. By comparing Fig. 6a and d, it is observed that increasing pressure from 150 bar to 220 bar increases the C C/C C ratio due to the increase in conversion. In addition, while increasing the temperature enhances the decomposition and ␤-scission reaction forming smaller 1-alkenes, increasing the pressure enhances the hydrogen abstraction and radical addition to form stable n-alkanes. Higher pressure increases conversion by accelerating the rate of chain transfer reactions which enhances the selectivity of liquid phase products while it reduces gas generation as observed in this study [4].

The tendency to form an equi-molar distribution of n-alkanes and 1-alkenes, i.e., equal ratio of n-alkane to 1-alkene at different carbon numbers, is observed at lower reaction temperatures. For instance, it is observed that the C C/C C ratio is almost constant at 0.25 between C7 –C13 when the temperature is 525 ◦ C, and the residence time is 13.89 s, at 220 bar, Fig. 6a. A similar trend is also observed between C7 –C12 at 565 ◦ C and 7.09 s, Fig. 6b, and C7 –C13 at temperatures of 565 ◦ C and 585 ◦ C and a residence time of 4.70 s, Fig. 6c. This suggests that reducing the conversion by decreasing the temperature and residence time enhances the tendency to follow the single step F-S-S free radical cracking mechanism suggested earlier by Fabuss, Satterfield, and Smith [39,40]. Similar trends for

730

Y.M. Alshammari, K. Hellgardt / J. of Supercritical Fluids 107 (2016) 723–732

Fig. 8. Arrhenius plot for sub and supercritical water gasification of n-hexadecane.

Fig. 7. Conversion profile (1st order) and Arrhenius plot at (a) 220 and (b) 150 bar.

the C C/C C ratio were also reported by Watanabe et al. [10] who observed nearly equi-molar distribution between C12 –C7 during anhydrous high pressure pyrolysis of n-hexadecane at 420–450 ◦ C. However, another work by Arai et al. [17] did not show an equimolar distribution of the C C/C C ratio at a temperature of 400 ◦ C and a long reaction time of 60 min, in a batch system. The deviation from equi-molar distribution of produced cracking products is also observed as the reactor pressure decreases to 150 bar at which it operates under subcritical conditions, Fig. 6d. The decrease in pressure to 150 bar causes a major reduction in the attainable conversion while reducing the selectivity of n-alkane over 1-alkene. This suggests that the subcrtical water conditions enhance the tendency for cracking to occur in a number of steps rather than the single-step F-S-S free radical mechanism [39,40]. It is, hence, concluded that under supercritical conditions, increasing the conversion, at higher temperatures and residence times, leads to a significant increase in the n-alkane to 1-alkene ratio. However, the formation of an equi-molar distribution of cracking products through a single step mechanism is enhanced at lower temperatures and residence times. 3.5. Conversion profile and kinetic data 3.5.1. Conversion As shown in Fig. 7, the hexadecane fractional conversion (X) increases with increasing temperature and residence time (), as fitted by a first order rate Eq. (7).

 ln

1 (1 − X)



= kg 

(7)

This correlation assumes that the reaction follows first order kinetics and the reactor obeys plug flow behaviour.

The conversion profile shows a reasonable agreement with the first order kinetic model depicted as observed by the straight lines passing through the origin, which also agrees with previous reports on cracking of heavy alkanes [6,8]. The increase in the hexadecane conversion showed a significant observable increase in the gaseous flowrate, as explained above. Furthermore, the fractional conversion, shown in Fig. 7, shows that a high conversion of around 75% may be achieved when the residence time is increased to 13.89 s, at 595 ◦ C and 220 bar. It is also observed that increasing the pressure to supercritical conditions (220 bar) not only increases the conversion but it also enhances the temperature effects. This may be observed by comparing the conversion profiles in Fig. 7a and b for the sub and supercritical conditions. It is found that conversion occurs at 525 ◦ C when operating under supercritical water conditions. However, this did not occur under subcritical condition, Fig. 7b where conversion was not observed at a temperature less than 565 ◦ C. The increase in oil conversion was observed further by the change in oil colour and volumetric quantity collected from the phase separator. The volumetric fraction of the oil collected decreased with increasing conversion at higher temperatures and residence times, with a significant change in oil colour. Higher conversions also resulted in the decrease of boiling point, density and viscosity compared with the n-hexadecane feedstock. This confirms the increase of lighter hydrocarbons (
Y.M. Alshammari, K. Hellgardt / J. of Supercritical Fluids 107 (2016) 723–732

731

Table 1 Comparison of kinetic data. Reference

E [kJ/mol]

A [s−1 ]

Remarks

Tsuzuki et al. [6] Khorasheh and Gray [37] Watanabe et al. [10] Jackson et al. [4]

318.20 280.06 263/196 309.32

1.08 × 1019 11.06 × 108 – 3.00 × 1019

Ford [11] Depeyre et al. [38]

3.16 × 1014 9.6 × 1013

Bartekova and Bajus [43]

249.13 57 (quartz) 238.26 162

370–410 ◦ C, 250–490 bar hydrothermal pyrolysis 380–450 ◦ C, 139 bar, 1.5–10% conversion, Anhydrous Anhydrous pyrolysis; 0.22/0.07 mol/L Anhydrous 300–370 ◦ C 150–600 bar Anhydrous 330–370 ◦ C Steam reforming (1 atm) 600–850 ◦ C

Yong et al. [44] This work This work

300 263.68 ± 1.40 202.20 ± 1.31

– 7.36 × 1014 2.29 × 1012

3.5 × 109

reported for other model compounds/conditions [42]. To illustrate, Promdej and Matsumura [42] found that the hydrothermal conversion of glucose is an ionic reaction when operating under subcritical conditions (573 K and 250 bar) and it becomes radical in the supercritical region (733 K and 250 bar). This was confirmed by the observation that radical reactions, in the supercritical region, showed good agreement with the Arrhenius behaviour, while ionic reactions, in subcritical conditions deviated from the Arrhenius behaviour. In our work, however, reducing the pressure to 150 bar resulted in a gas-phase, high-pressure steam reforming reaction observed previously to follow the free radical mechanism [38]. This shows operating in a subcritical water regime, by decreasing the pressure to less than 220 bar at a constant supercritical temperature, >374 ◦ C, follows a different reaction mechanism (radical) when compared to creating subcritical conditions by decreasing the temperature to less than 374 ◦ C, at a constant supercritical pressure, >220 bar. In the current analysis, Fig. 8 shows the Arrhenius plot showing the average values of ln [kg ], obtained at different residence times, against the reciprocal of the reactor temperature, 1000/T [K−1 ]. Using the least square method the reaction activation energies were determined to be 263.68 kJ/mol and 202 kJ/mol for both super and subcritical conditions, respectively. Furthermore, the preexponential frequency factor was determined to be 7.36 × 1014 s−1 , and 2.29 × 1012 s−1 at 220 bar and 150 bar, respectively, as shown in the resulting Eqs. (8) and (9).



kg [s−1 ] = 7.36 × 1014 s−1 exp

 kg [s−1 ] = 2.29 × 1010 s−1 exp

263.68 ± 1.40 kJ/mol Rg T 202 ± 1.31 kJ/mol Rg T



(8)



(9)

The effect of pressure is obvious by the higher activation energy obtained under supercritical conditions. Tsuzuki et al. [6] related the increase in activation energy at higher pressures to the suppression effect of a higher density on the radical chain mechanism due to the faster recombination rates. Table 1 compares the kinetic data reported in this work with different previously reported kinetic results on n-hexadecane cracking both in hydrous and anhydrous media. The high activation energy obtained in this work with increasing pressure is associated with a frequency factor (A) of 7.36 × 1014 s−1 , which is significantly higher than previously reported values for systems under either low pressure steam cracking or anhydrous pyrolysis of n-hexadecane. The lower activation energy determined for subcritical conditions in this work, shows a significantly lower frequency factor of 2.29 × 1012 s−1 . Previous work on low pressure steam reforming of nhexadecane [43] reported a value 162 kJ/mol for the activation energy, which is lower than the value that we obtained in this work for the subcritical region. However, their steam to carbon

700–780 ◦ C Steam to hydrocarbon 3:1 420–460 ◦ C and 23–27 MPa in SCW Supercritical (220 bar) Subcritical (150 bar)

ratio is relatively higher, 3:1, than our ratio which may point towards an effect of the feed concentration on the value of activation energy compared with pressure and temperature. Another observation is that the values obtained for the activation energy are very close to those reported for high pressure anhydrous pyrolysis of n-hexadecane. Yu and Eser [8] reported that pressure and temperature do not have significant effects on the values of activation energy while comparing different results on cracking of hydrocarbons. However, our results show that increasing the pressure, at constant temperature, shows an observable increase in the activation energy. 4. Conclusion This work reports new results on investigating the effects of temperature, pressure, and reactor residence time on the hydrogen production via hydrothermal gasification of n-hexadecane in tubular flow reactor. It is found that as the temperature increases, the yields of H2 and CO2 increase, which suggests the enhancement of the WGS reaction. This shows the need for the use of catalyst to drive the WGS closer to equilibrium at relatively short residence times. The gasification in hydrothermal media carried out in this work led to minimising carbon formation which presents a major challenge to conventional hydrogen production processes. The decrease in pressure from 220 bar to 150 bar enhanced the yields of gaseous species and reduced the yields of liquid species. The yields of liquid species was found to follow the free radical pyrolysis mechanism through which the yield of 1-alkenes has always been higher than the yield of n-alkanes suggesting the low hydrogen selectivity the dominance of ␤-scission over hydrogen abstraction. New kinetic data have been determined for uncatalysed hydrothermal gasification of hexadecane which may be used for reactor and process modelling for production of hydrogen and syngas. Future work will include investigating the effects of partial oxidation on the conversion of hexadecane and reactor modelling using produced kinetic data. Acknowledgement Authors would like to thank KAUST and the Saudi Royal Commission for Jubail and Yanbu for sponsoring this project. References [1] R.M. Navarro, M.A. Pena, J.L.G. Fierro, Hydrogen production reactions from carbon feedstocks: fossil fuels and biomass, Chem. Rev. 107 (10) (2007) 3952–3991. [2] J.G. Weissman, Review of processes for downhole catalytic upgrading of heavy crude oil, Fuel Process. Technol. 50 (2–3) (1997) 199–213.

732

Y.M. Alshammari, K. Hellgardt / J. of Supercritical Fluids 107 (2016) 723–732

[3] O. Weres, A.S. Newton, L. Tsao, Hydrous pyrolysis of alkanes, alkenes, alcohols and ethers, Org. Geochem. 12 (5) (1988) 433–444. [4] K.J. Jackson, et al., Temperature and pressure dependence of n-hexadecane cracking, Org. Geochem. 23 (10) (1995) 941–953. [5] F. Domine, Kinetics of hexane pyrolysis at very high pressures. 1. Experimental study, Energy Fuels 3 (1) (1989) 89–96. [6] N. Tsuzuki, et al., The kinetic modeling of oil cracking by hydrothermal pyrolysis experiments, Int. J. Coal Geol. 39 (1–3) (1999) 227–250. [7] J. Yu, S. Eser, Thermal decomposition of C10 –C14 normal alkanes in near-critical and supercritical regions: product distributions and reaction mechanisms, Ind. Eng. Chem. Res. 36 (3) (1997) 574–584. [8] J. Yu, S. Eser, Kinetics of supercritical-phase thermal decomposition of C10 –C14 normal alkanes and their mixtures, Ind. Eng. Chem. Res. 36 (3) (1997) 585–591. [9] J.A. Fairburn, L.A. Behie, W.Y. Svrcek, Ultrapyrolysis of n-hexadecane in a novel micro-reactor, Fuel 69 (12) (1990) 1537–1545. [10] M. Watanabe, et al., Kinetics and product distribution of n-hexadecane pyrolysis, AIChE J. 46 (4) (2000) 843–856. [11] T.J. Ford, Liquid-phase thermal decomposition of hexadecane: reaction mechanisms, Ind. Eng. Chem. Fundam. 25 (2) (1986) 240–243. [12] C. Fauteux-Lefebvre, et al., Steam reforming of liquid hydrocarbons over a nickel–alumina spinel catalyst, J. Power Sources 195 (10) (2010) 3275–3283. [13] S.K. Goud, et al., Steam reforming of hexadecane using a catalyst: kinetics of catalyst deactivation, Int. J. Hydrogen Energy 32 (14) (2007) 2868–2874. [14] D.H. Kim, et al., Steam reforming of n-hexadecane over noble metal-modified Ni-based catalysts, Catal. Today 136 (3–4) (2008) 228–234. [15] J. Thormann, et al., Steam reforming of hexadecane over a Rh/CeO2 catalyst in microchannels: experimental and numerical investigation, Int. J. Hydrogen Energy 34 (12) (2009) 5108–5120. [16] P.E. Savage, Organic chemical reactions in supercritical water, Chem. Rev. 99 (2) (1999) 603–622. [17] K. Arai, T. Adschiri, M. Watanabe, Hydrogenation of hydrocarbons through partial oxidation in supercritical water, Ind. Eng. Chem. Res. 39 (12) (2000) 4697–4701. [18] M. Watanabe, et al., Catalytic effects of NaOH and ZrO2 for partial oxidative gasification of n-hexadecane and lignin in supercritical water, Fuel 82 (5) (2003) 545–552. [19] M. Watanabe, et al., Partial oxidation of n-hexadecane and polyethylene in supercritical water, J. Supercrit. Fluids 20 (3) (2001) 257–266. [20] Y.J. Lu, et al., Hydrogen production by biomass gasification in supercritical water: a parametric study, Int. J. Hydrogen Energy 31 (7) (2006) 822–831. [21] A. Kruse, Hydrothermal biomass gasification, J. Supercrit. Fluids 47 (3) (2009) 391–399. [22] M.J. Antal, et al., Biomass gasification in supercritical water, Ind. Eng. Chem. Res. 39 (11) (2000) 4040–4053. [23] T. Yoshida, Y. Oshima, Partial oxidative and catalytic biomass gasification in supercritical water: a promising flow reactor system, Ind. Eng. Chem. Res. 43 (15) (2004) 4097–4104. [24] S. Guo, et al., Hydrogen production from glycerol by supercritical water gasification in a continuous flow tubular reactor, Int. J. Hydrogen Energy 37 (7) (2012) 5559–5568.

[25] P.T. Williams, J. Onwudili, Composition of products from the supercritical water gasification of glucose: a model biomass compound, Ind. Eng. Chem. Res. 44 (23) (2005) 8739–8749. [26] Q. Guan, P.E. Savage, C. Wei, Gasification of alga Nannochloropsis sp. in supercritical water, J. Supercrit. Fluids 61 (0) (2012) 139–145. [27] F.L.P. Resende, et al., Noncatalytic gasification of lignin in supercritical water, Energy Fuels 22 (2) (2008) 1328–1334. [28] R.F. Susanti, et al., High-yield hydrogen production from glucose by supercritical water gasification without added catalyst, Int. J. Hydrogen Energy 37 (16) (2012) 11677–11690. [29] T. Sato, et al., Analysis of the density effect on partial oxidation of methane in supercritical water, J. Supercrit. Fluids 28 (1) (2004) 69–77. [30] R.F. Susanti, et al., Noncatalytic gasification of isooctane in supercritical water: a strategy for high-yield hydrogen production, Int. J. Hydrogen Energy 36 (6) (2011) 3895–3906. [31] R.F. Susanti, et al., Continuous supercritical water gasification of isooctane: a promising reactor design, Int. J. Hydrogen Energy 35 (5) (2010) 1957–1970. [32] Y.M. Alshammari, K. Hellgardt, Thermodynamic analysis of hydrogen production via hydrothermal gasification of hexadecane, Int. J. Hydrogen Energy 37 (7) (2012) 5656–5664. [33] Y.M. Alshammari, K. Hellgardt, Partial oxidation of n-hexadecane through decomposition of hydrogen peroxide in supercritical water, Chemical Engineering Research and Design 25 (2015) 565–575. [34] K. Pinkwart, et al., Gasification of diesel oil in supercritical water for fuel cells, J. Power Sources 136 (2) (2004) 211–214. [35] F.O. Rice, K.F. Herzfeld, The thermal decomposition of organic compounds from the standpoint of free radicals. VI. The mechanism of some chain reactions, J. Am. Chem. Soc. 56 (2) (1934) 284–289. [36] A. Kossiakoff, F.O. Rice, Thermal decomposition of hydrocarbons, resonance stabilization and isomerization of free radicals1, J. Am. Chem. Soc. 65 (4) (1943) 590–595. [37] F. Khorasheh, M.R. Gray, High-pressure thermal cracking of n-hexadecane, Ind. Eng. Chem. Res. 32 (9) (1993) 1853–1863. [38] D. Depeyre, C. Flicoteaux, C. Chardaire, Pure n-hexadecane thermal steam cracking, Ind. Eng. Chem. Process Des. Dev. 24 (4) (1985) 1251–1258. [39] B.M. Fabuss, et al., Thermal decomposition rates of saturated cyclic hydrocarbons, Ind. Eng. Chem. Process Des. Dev. 3 (3) (1964) 248–254. [40] B.M. Fabuss, et al., Rapid thermal cracking of n-hexadecane at elevated pressures, Ind. Eng. Chem. Process Des. Dev. 1 (4) (1962) 293–299. [41] G. Wu, et al., Comparison of liquid-phase and gas-phase pure thermal cracking of n-hexadecane, Ind. Eng. Chem. Res. 35 (12) (1996) 4747–4754. [42] C. Promdej, Y. Matsumura, Temperature effect on hydrothermal decomposition of glucose in sub- and supercritical water, Ind. Eng. Chem. Res. 50 (14) (2011) 8492–8497. [43] E. Bartekova, M. Bajus, Pyrolysis of hexadecane, Collect. Czech. Chem. Commun. 62 (7) (1997) 1057–1069. [44] D. Yong, et al., Pyrolysis of Hexadecane as Model Compound in Supercritical Water, Petrochemical Technology, 2007.