Reaction of ethylene over a typical Fischer‐Tropsch synthesis Co/TiO2 catalyst

In order to identify the potential reaction paths of C2H4 and their product distribution in Fischer‐Tropsch synthesis (FTS), a series of experiments were designed over a Co/TiO2 catalyst in the absence of CO. C2H4 did quickly react with H2 to produce C1‐6 products under Fischer‐Tropsch (FT) reaction conditions. Although the dominant reaction is C2H4 hydrogenation to ethane, changing the reaction conditions (temperature and partial pressure of reactants) can lead to the other reaction pathways being enhanced, resulting in varying product selectivity to both linear and branch olefins and paraffins. Possible reaction pathways had been summarized and discussed, which including C2H4 reaction to ethylidene followed by dimerization; C2H4 insertion into C2 surface species and dimerization and C4 decomposition and/or direct C2 hydrogenolysis. Furthermore, the products obtained from C2H4 reactions were fit to a typical FTS product distribution, which indicate that both the chain growth initiators and monomers are not necessarily only derived from hydrogenation of CO but also from the secondary reactions of olefins.


INTRODUCTION
Fischer-Tropsch synthesis (FTS) is an important technology used to convert syngas derived from coal/gas/biomass into clean transport fuels or other valuable organics. [1][2][3][4][5][6][7][8] FTS is generally regarded as a polymerization-like reaction. The products of FTS are a wide range of hydrocarbons, consisting of mainly olefins and paraffins, with a small amount of oxygenates. The Anderson-Schulz-Flory (ASF) equation is used to describe the FT product distribution; however, deviations from the ideal ASF distribution, such as a higher yield of C 1 , and a lower yield of C 2 , have been observed. 9, 10 Many theories have been proposed to explain the deviation, and secondary reactions of olefins is considered as a reasonable explanation. [11][12][13][14] Therefore, it is interesting to investigate the reaction of C 2 H 4 under typical FTS operating conditions (similar space velocity, temperatures, and pressures).
The olefin product produced by the FTS may re-adsorb on the catalyst surface and undergo secondary reactions. The effect of co-feeding low concentrations of C 2 H 4 to the FTS reaction system has been studied previously, in order to investigate the reactivity of the olefins in the FTS reaction. [14][15][16][17][18][19][20] Different secondary reactions of olefins have been reported, and in particular hydrogenation, 14,15 isomerization (bond shift reaction), 16 hydrogenolysis, 17 reinsertion into chain growth as monomers, 18 initiation of hydrocarbon chain, 19 and hydroformylation 20 have been observed.
The reactivity of C 2 H 4 under FTS conditions has been reported for both Fe-based and Co-based catalysts. [21][22][23] Schulz et al 24 found that the conversion of C 2 H 4 was less than 80% for iron-based catalysts, while almost all the C 2 H 4 was converted (conversions over 90%) when using cobalt-based catalysts. In addition, hydrogenation of C 2 H 4 is the dominant reaction under all reaction conditions. [21][22][23] Later, reabsorption and insertion of C 2 H 4 to form longer carbon chain products on Co-based catalysts were reported. 14,[25][26][27] Furthermore, a decrease in CH 4 selectivity was found when C 2 H 4 was co-fed, which was considered due to the competitive reaction between CO methanation and C 2 H 4 incorporation with the C 1 species. 26 Some studies focused on the hydro-polymerization of C 2 H 4 at very low partial pressures of CO over supported Co-based catalysts. [28][29][30][31][32] It is reported that the C 2 H 4 in the feed was completely consumed, and the CO partially hydrogenated to generate longer chain hydrocarbons. Kokes et al 31 found that a large number of C 2 H 4 dimerization products (C 4 ), especially olefins, were formed when hydrogen was deficient. They 31 proposed that an adsorbed intermediate of 1,1-σ-ethylidene, converted by 1,2-diadsorbed C 2 H 4 with hydrogen assistance, could polymerize to form C 2 H 4 dimers. However, most of the studies reported were carried out under typical FT reaction conditions, such as low conversion and/or atypical pressures. More recently, some studies on co-feeding C 2 H 4 to an FTS reaction system were carried out using a combined quantitative in-situ FTIR and online gas chromatography. 33,34 The researchers 33,34 found that co-feeding C 2 H 4 did not alter the selectivity of the product, but changed the chain length of the adsorbate on the catalyst surface.
In summary, the types of FTS catalyst, and the operating conditions (temperature, residence time, and partial pressure of the reactants and products) had a significant impact on the secondary reactions of C 2 H 4 . 12 Researchers have found that it was difficult to investigate certain aspects independently due to the system complexity and number of the reactions occurring in FTS. In addition, some reaction pathways may be obscured when many reactions are occurring simultaneously. In order to fully understand the reaction pathways of the olefin, we suggest simplifying the complex system and investigating one aspect at a time.
In this work, several groups of experiments were carried out over a FT Co-based catalyst, in the absence of CO, under typical FT reaction conditions. First, mixtures of C 2 H 4 and N 2 were fed to the reactor and later the feed was changed to mixtures of C 2 H 4 and H 2 . The effect of the feed gas ratio (C 2 H 4 /H 2 ) and the reaction temperatures on the reactivity of C 2 H 4 and resulting product selectivity could be investigated without either the reactant CO or the FT by-products influencing the system.

Catalyst preparation
The catalysts used in this study (15 wt% Co supported on TiO 2 ) were prepared using the incipient wetness method. Co(NO 3 ) 3 ⋅6H 2 O (Sigma Aldrich) was used as the metal precursor salt, and TiO 2 (Degussa P-25) was used as the catalyst support precursor. The TiO 2 paste was made by mixing TiO 2 with distilled water at a mass ratio of 1:1. After being dried at 120 • C for 2 hours, the paste was calcined in a Muffle oven that was heated from room temperature to 400 • C at a rate of 5 • C/min and then maintained at 400 • C for 6 hours. The support was crushed and sieved after calcination, and particles of which between 0.5 and 1 mm were selected for the impregnation step. 35 1 g of the support was measured out, and distilled water (0.8 mL) was gradually dropped onto the support until the surface was infiltrated. The pre-treated and selected TiO 2 support was impregnated with a Co(NO 3 ) 3 aqueous solution, with a Co metal loading of 15% by mass. Thereafter, the wet catalyst was dried at 120 • C for 2 hours, and then calcined at 400 • C for 6 hours. Catalyst information is summarized in Table 1.

Reaction procedure and product analysis
The fixed bed reactor used in this study had a 203.8 mm long stainless tube with an inner diameter of 8 mm. 1 g of catalyst was loaded into the reactor, and the catalyst was reduced at 350 • C in pure H 2 (AFROX [African Oxygen] Ltd., 99.999%), for 4 hours at 30 mL/min and 1 bar gauge. After catalyst reduction, the reactor was cooled to a temperature below 100 • C. Thereafter, three groups of experiments were conducted, as follows: • Group 1 (Run 1) A normal FT reaction was conducted with a feed gas of H 2 /CO/N 2 (H 2 /CO = 2:1), at 30 mL/min, 200 • C, and 20 bar (on gauge). This was used to test the catalyst performance during FTS. Thereafter, the reactor was purged with inert gas N 2 .

• Group 2 (Runs 2-4)
First, a feed of pure C 2 H 4 was introduced into the reactor at 20 bar and the temperature was varied between 180 • C and 220 • C, while the total flow rate was maintained at 50 mL/min. Then, N 2 was co-fed into the reactor using various C 2 H 4 /N 2 ratios. The operating temperature was also varied from 180 • C to 220 • C while keeping the total pressure at 20 bar.
• Group 3 (Runs 5-11) H 2 was co-fed with C 2 H 4 into the reactor at 20 bar and 180 • C, 60 mL/min. The ratio of C 2 H 4 to hydrogen was changed from 0.5 to 4.6. Thereafter, experiments were run at 2 different C 2 H 4 /H 2 ratios, namely, 0.8 (which we denote "Excess H 2 ") followed by 4.8 (denoted "H 2 limiting" ), and the operating temperature in the reactor was varied between 100 • C and 250 • C, to investigate the effect of the C 2 H 4 /H 2 ratio and temperature on C 2 H 4 reactivity.
The reaction conditions are summarized in Table 2. The feed and tail gas were analyzed using an online Gas Chromatograph (GC, Agilent 7890B). A flame ionization detector (FID) was used to analyze the hydrocarbons, and two thermal conductivity detectors (TCD) were used to analyze H 2 /N 2 . The experimental results monitored by the GC indicated that the normal FT reaction (Run 1) stabilized after 13 hours (see in Figure 2); and the C 2 H 4 hydrogenation reaction reached a steady state less than 2 hours for each run. The experimental results reported in the current work were the average values obtained between 17 and 23 hours for all the runs.
In the FTS experiment (Run 1), the feed gas flow rate was controlled by a mass flow controller (Brooks) and N 2 was used as an internal standard for the calculation of the results. The CO conversion ((%)CO) and hydrocarbon selectivity (S Cn_1 ) were calculated using the following equations: During the reaction of ethylene (Runs 5-11), the flow rate of feed gas was controlled by a mass flow controller (Brooks) and the flow rate of the tailgas was measured using a bubble flow meter. The C 2 H 4 conversion ((%)C 2 H 4 ) and hydrocarbons selectivity (S Cn_2 ) were estimated using the following equations: Where: F CO,in and F CO,out are the CO molar flow rates of the feed gas and the tailgas, respectively (mmol/min), F C2H4,in and F C2H4,out are the ethylene molar flow rates of the feed gas and the tailgas, respectively (mmol/min), n is the carbon number of product Cn, and F Cn,out is the molar flow rate of hydrocarbon product with carbon number n in tailgas (mmol/min).

Catalyst characterization
The reducibility of the catalyst was measured by hydrogen temperature programmed reduction (H 2 -TPR). Two experiments were conducted: 1. Experiment 1:50 mg of catalyst was loaded into the quartz reactor. A 5% H 2 /N 2 reducing mixture, at a flow rate of 30 mL/min, was passed through the reactor. The temperature of the reactor was programmed to increase from room temperature to 700 • C at a heating rate of 10 • C/min. 2. Experiment 2: the same experimental procedure was followed as with experiment 1, the difference being that the reaction temperature increased from room temperature to 350 • C at a rate of 10 • C/min, and it was then maintained at 350 • C for 1 hour. The H 2 concentration during the reduction was measured using a thermal conductivity detector (TCD).
X-ray diffraction (XRD) analysis was performed on a Rigaku D/max-2500 diffractometer, with Cu Kα radiation at 40 kV and 100 mA in a scanning range of 3 • -80 • (2θ). The diffraction peaks of the crystalline phase were compared with those of standard compounds reported in the JCPDS Date File.
The nitrogen adsorption-desorption experiment was measured using a Quantachrome Autosorb-1MP sorption analyzer at −196 • C. Before measurement, the samples were de-gassed at 200 • C for at least 6 hours. The specific surface area (SBET) was calculated using the Brunauer-Emmett-Teller (BET) method (P/P 0 < 0.1).

Catalyst characterization
The characteristic results of TPR and XRD are shown in Figure 1A-C. Figure 1A indicates that the catalyst starts to reduce at temperatures above 320 • C. Figure 1B shows that the catalyst can be activated at 350 • C. Based on the literature, there are two steps during the cobalt catalyst reduction process: (a) from Co 3 O 4 to CoO, around 300 • C; (b) from CoO to metallic F I G U R E 1 Characterization results for: A, H 2 -TPR profiles with the reduction temperature increasing from room temperature to 700 • C at a rate of 10 • C/min; B, H 2 -TPR profiles with the reduction temperature increasing from room temperature to 350 • C (10 • C/min) and then being maintained at 350 • C for 1 hour; C, XRD patterns for 15% Co/TiO 2 before and after reduction; and D, TEM images of the catalyst Co, around 500 • C. 36 As shown in Figure 1A,B, A and B represent these two reduction steps. It indicates at 350 • C, at least part of Co 3 O 4 reduce to metallic Co. The comparison of the XRD results for the fresh catalyst, before reduction and after reduction, at 350 • C, is shown in Figure 1C. The pattern produced by the catalyst before reduction shows some sharp peaks, which may be due to the cubic Co 3 O 4 (JCPDS65-3103). However, after reduction, the intensity of the Co 3 O 4 peaks either disappears or is reduced. At the same time, certain Co (JCPDS15-0806) peaks appear. Although the signal of metallic Co is weak and difficult to distinguish from the noise, combined with the TPR result, it confirms that metallic Co exists in reduced catalysts. According to the BET results, the catalyst surface area is about 43 m 2 /g, the pore volume is 0.24 cm 3 /g, and the average pore diameter is 16.7 nm. (See Table 2.) All the characterization results show that the catalyst used in this experiment is a typical Co-based FTS catalyst. Figure 1D shows the TEM images of catalysts.

Confirming FTS reactivity of the Co catalyst
When the syngas (H 2 /CO = 2:1) is introduced into the reactor, a typical product distribution of a Co-based FT catalyst 13 is obtained (see Table 3 and Figure 8A). The CH 4 selectivity is high (16%), while the C 2 product selectivity is lower than that expected from an ideal ASF distribution In addition, the C 3+ product distribution is consistent with an ASF distribution, with α-value = 0.82 ( Figure 8A). Moreover, the total C 5+ selectivity is 65.03%. Figure 2 shows the CO conversion and methane selectivity with respect to time on steam (TOS). As can be seen, FTS stabilized at around 13 hours TOS. To avoid errors, the average value of results obtained between 17 and 23 hours after the experimental conditions are changed are used in Table 3 and Figure 8A. The experimental results clearly confirm that the catalyst used in this experiment is a typical FTS Co-based catalyst. 37

Feeding C 2 H 4 /N 2 to the reactor
In order to test if pure C 2 H 4 could react under normal FT operating conditions, gas mixtures of C 2 H 4 /N 2 were fed to the reactor. N 2 is an inert gas with the function of adjusting the partial pressure of C 2 H 4 to the reactor system. When the reactor was operated at 20 bars (on gauge) and 180 • C, no product was detected by the online GC. The reaction temperature was then increased from 180 • C to 220 • C, but product was still not detected. Adjusting the partial pressure of C 2 H 4 , by changing the molar ratio of C 2 H 4 /N 2 in the feed mixture (

Effect of C 2 H 4 /H 2 molar feed ratio
The feed gas was switched to a mixture of C 2 H 4 /H 2 with the reactor operating at 180 • C and 20 bars gauge and the GC detected some short chain hydrocarbon products. The feed ratio was varied and the conversion of C 2 H 4 and H 2 and the product selectivity are plotted in Figure 3. Figure 3A shows that when C 2 H 4 /H 2 < 1, C 2 H 4 is completely converted. However, the conversion of H 2 reaches more than 98% when C 2 H 4 /H 2 ≥ 1, but the C 2 H 4 conversion drops from 91% to 22%. (See Figure 3A.) The ethane selectivity decreases from 98.98% to 95.75%, when the C 2 H 4 /H 2 ratio is increased from 0.5 to 1; thereafter the ethane selectivity does not change much as the C 2 H 4 /H 2 ratio is increased above 1. In all cases, hydrogenation of C 2 H 4 to ethane is the main reaction and this contributes to the high ethane selectivity (>95%). This is in line with published reports. 6 Figure 3B,C shows: 1. When the C 2 H 4 /H 2 ratio increases from 0.5 to 4.6, the paraffin product selectivity for C 1 , C 3 and C 4 increases, reaching a maximum at ratio of 0.8 and then decreases. 2. When C 2 H 4 /H 2 ≥ 1, both C 3 and C 4 olefins (including the corresponding isomers) are produced, and the total yield of olefins is much higher than that of paraffins.

F I G U R E 3
The effect of C 2 H 4 : H 2 molar ratio on the C 2 H 4 reaction at 180 • C and 20 bar gauge). A, Conversion of reactants and the selectivity of C 2 H 6 . B, Selectivity of CH 4 and C 3 products. C, Selectivity of C 4 products 3. When the C 2 H 4 /H 2 ratio is increased from 1 to 4.6: the selectivity of total C 4 product increases from 2.2% to 2.8%; the selectivity of CH 4 and total C 3 product decreases from 0.5% to 0.1% and 1.5% to 0.3%, respectively.
Because C 2 H 4 was the only carbon source in these experiments, the production of C 1 and C 3 indicates that the C-C bond ruptures to form odd carbon number products when co-feeding H 2 . The selectivity of both CH 4 and C 3 decreases with increasing partial pressure of C 2 H 4 (C 2 H 4 /H 2 > 1). This implies that few carbon chain products formed from the hydrocracking or demethylation reaction.
The major C 4 olefin under these reaction conditions was 2-butene. This phenomenon indicates that the production of 2-butene in FTS may come from the C 2 H 4 dimerization reaction. In addition, with an increase in the partial pressure of C 2 H 4 , the selectivity of cis-2-butene increased, while the trans-2-butene yield decreased.

Influence of reaction temperature when C 2 H 4 /H 2 < 1 in the feed gas (H 2 in excess)
As described earlier, the product distribution changes with C 2 H 4 /H 2 ratio. To investigate the effect of temperature on the reaction of C 2 H 4 with H 2 , two groups of experiments were carried out: one using the feed gas with a molar ratio of C 2 H 4 /H 2 = 0.8 (excess H 2 ); the other with C 2 H 4 /H 2 = 4.8 (H 2 limiting-see next section). The term "excess" of "limiting" refers to the hydrogenation reaction, where a ratio C 2 H 4 /H 2 = 1 would be the correct ratio if all the C 2 H 4 reacted to ethane. Figure 4 shows the conversions of reactants and the selectivity of C 1 to C 4 products, at 20 bar gauge, a total inlet flow rate of 52.5 mL/min, C 2 H 4 /H 2 = 0.8 and with the reaction temperature varying from 100 to 250 • C (H 2 excess). The C 2 H 4 was completely converted when there was H 2 in the feed gas for all operating temperatures between 100 • C and 160 • C. As temperature increased above 160 • C, C 2 H 4 conversion decreased to reach 97% at 250 • C. A similar trend in H 2 conversion was observed. This might be caused by acceleration of C 2 H 4 adsorption/ desorption with the increase in temperature. Figure 4A also shows the variation of the ethane selectivity with operating temperature, and it can be seen that the higher the temperature, the lower the ethane selectivity. The dominant reaction at all temperatures is C 2 H 4 hydrogenation, as the ethane selectivity is always higher than 94% as shown in Figure 4; however, the ethane selectivity drops with increase in temperature, which implies that the more C 2 H 4 is consumed in competing reactions, such as dimerization and hydrocracking.
The selectivity of CH 4 and C 3 and C 4 products (paraffins and olefins) are shown in Figure 4B,C. The selectivity of CH 4 and total C 3 and C 4 products all increased with increasing temperature. As shown in Figure 4, there was an obvious increase in selectivity between 160 • C and 180 • C. At temperature lower than this critical temperature range, only paraffins were produced and no olefins were formed. At very low temperatures (less than 120 • C), the C 2 H 4 dimer (butane) had a higher selectivity than either CH 4 or C 3 (odd number hydrocarbons) . This indicates that the reaction of C 2 H 4 oligomerization was faster than that of hydrogenolysis or demethylation at the low reaction temperatures. When the reaction temperature was increased to 180 • C, there was a marked increase in the selectivity of CH 4 , total C 3 , and total C 4 hydrocarbon products, with CH 4 selectivity increasing the most. The CH 4 selectivity increased from 1.80% to 2.83%, while the total C 3 selectivity only increased from 1.1% to approximately 1.5%; however, the selectivity of C 2 H 4 dimer (C 4 ) was fairly constant (∼1.2%) as the temperature was increased from 180 • C to 250 • C. (See Figure 4.) It is worth noting that the olefin products were only produced when the temperature was higher than 180 • C.
The product distribution for the reaction of C 2 H 4 in the presence of H 2 appears to occur via reactions that are different from those indicated in the theory on carbon chain growth in FTS. Liu et al 38 suggested two pathways to explain these product distributions. One is the oligomerization (dimerization) of the α-olefins in the feed and the subsequent cracking with H 2 assistance. Another is attributed to the addition of methylene species, which are formed via the demethylation reaction. 39 To investigate the different reactions, the yields of CH 4 , total C 3 , and C 4 products were calculated and the results are shown in Figure 5. In comparison to odd carbon number hydrocarbons, the yield of the C 4 products did not change much with temperatures over the range tested (100 • C to 250 • C). This suggests that the C 2 H 4 dimerization reaction is not sensitive to the reaction temperature. However, for the demethylation products, the yield of CH 4 and C 3 increased more than 20-and 10-fold, respectively. The formation of CH 4 and C 3 can be described by the following two equations: Equation (A) represents direct C 2 H 4 hydrogenolysis to produced CH 4 . Equation (B) shows the methylene species formed by demethylation in carbon chain growth, which produces CH 4 and propene with a mole ratio of CH 4 /C 3 equal to 1:1. Equation (B) could be considered an analogical disproportionation. The results in Figure 4 indicate that the CH 4 yield was much higher than that of the C 3 products in the case of H 2 rich feeds. This phenomenon could indicate that a F I G U R E 4 Effect of temperature on C 2 H 4 reaction when H 2 is in excess (C 2 H 4 /H 2 = 0.8, total flow rate: 52.5 mL/min; reaction pressure: 20 bars). A, Conversion of reactants and selectivity of C 2 H 6 . B, Selectivity of CH 4 and C 3 . C, Selectivity of C 4 high coverage of chemisorbed H 2 on the surface of catalysts inhibits the methylene species that take part in chain growth reaction.

Influence of reaction temperature with a H 2 limited feed gas
In another group of experiments, a mixture of C 2 H 4 /H 2 where C 2 H 4 :H 2 = 4.8 corresponding a H 2 limited feed gas, was fed to the reactor. The experiments were conducted at 20 bar, a flow rate of 40 mL/min, and the reaction temperature was varied from 100 • C to 250 • C. Figure 6A shows that the conversion of both H 2  temperatures. H 2 reacted to completion and the reaction rates of both H 2 and C 2 H 4 were essentially the same, which is in agreement with the observed high selectivity of ethane. Thus, even in this H 2 limited situation, the dominant reaction was still C 2 H 4 hydrogenation. The experimental results in Figure 6 show that with a H 2 limited feed, the selectivity of ethane decreased from 98% (at 110 • C) to 95% (at 250 • C). (See Figure 6A.) This suggests that demethylation and oligomerization reactions are more likely to occur at a higher reaction temperature, which is the same as the result obtained for H 2 rich feeds. For H 2 limited feeds (or correspondingly excess C 2 H 4 ), we see from Figure 6B,C that olefins are the main product, even at a very low temperatures and that the selectivity of C 4 is much higher than that of C 3 , which indicates that high partial pressure of C 2 H 4 promotes the dimerization reaction. The selectivity of CH 4 , C 3 H 6 and C 3 H 8 all increase with increasing temperature. These results indicate that a high temperature is better for the production of odd carbon number hydrocarbons. However, in comparison to the results of the previous experiment ( Figure 4B for a feed with excess H 2 ), the product distribution of CH 4 and C 3 is different (see Figure 6B.) In this case, the selectivity to C 3 product is higher than that of CH 4 , and the selectivity of C 3 increases more rapidly with increasing temperature than the selectivity of CH 4 . This indicates that the formation of CH 4 could be limited by the availability of chemisorbed hydrogen. The methylene species formed by demethylation participate in chain growth more easily when C 2 H 4 in the feed is in excess, especially at high temperatures. Moreover, the paraffin to olefin ratio decreased with increasing temperature for C 3 also indicating that the availability of H 2 may be limiting.
The change in the selectivity of C 4 products with temperature is shown in Figure 6C and it can be seen that the selectivity of n-butane did not change much with temperature, while the total olefin selectivity increased initially, reached a maximum, and then decreased. The maximum selectivity of cis-2-butene was achieved at about 180 • C; for trans-2-butene, the maximum occurred at about 200 • C while for 1-butene it occurred at around 220 • C. The decrease in the selectivity of C 4 products at a high temperature suggests that the intermediates from C 2 H 4 dimerization on the catalyst surface may react to form other hydrocarbons. Similar to the product distribution of propene/propane, the selectivity of butene was significantly higher than butane. The selectivity of both cis-2-butene and trans-2-butene are higher than that of 1-butene. Figure 7 shows the effect of temperature on the molar formation rate of the C 1 , C 3 and C 4 products. These are different from the results seen with excess H 2 in the feed (Figure 5), as the total yield of C 4 product is much higher than that of CH 4 and C 3. This provides strong evidence that chemisorbed H 2 limits C 2 H 4 dimerization. In addition, the formation rates of CH 4 and the total C 3 product were different to those found for feeds with excess H 2 . At a low temperature (less than 180 • C), the mole ratio of CH 4 /C 3 was almost 1:1, which corresponds with Equation (B). Under these conditions, some of the C 2 H 4 "disproportionates" with H 2 assistance to form a 1:1 mol ratio of CH 4 and C 3 product. At temperature is equal to or above 180 • C, the increase in the yield of C 3 is significantly higher than that of CH 4 . This finding implies that another reaction occurs to produce C 3 . Combined with the decrease in yield of C 4 products, an analogical comproportionation could explain this phenomenon, that is: (C) F I G U R E 7 Effect of reaction temperature on the CH 4 yield, total C 3 and total C 4 , for a H 2 limited feed Figure 8 shows the product distribution plotted in an ASF plot for light hydrocarbons produced by (a) a feed of syngas, (b) feeds of C 2 H 4 /H 2 with excess H 2 , and (c) H 2 limited for different operating temperatures. When the reaction temperature is low, as shown in Figure 8B,C, the online GC cannot detect C 5+ products. However, the product distribution for feeds of C 2 H 4 /H 2 is similar to that of normal FTS at reaction temperature higher than 160 • C for feeds with H 2 in excess and higher than 200 • C, when the feed is H 2 limited When the carbon number n is greater than 2, the product distribution of light hydrocarbons approaches a straight line, which is consistent with the typical FTS product distribution, as shown in Figure 8A. However, when compared to typical FTS product distribution, several differences can be seen: 1. Compared to the positive deviation of CH 4 from the ideal ASF distribution in FTS, the CH 4 in the C 2 H 4 hydrogenation reaction product does not deviate from an ASF distribution in a H 2 rich feed ( Figure 8B) and even has a negative deviation for H 2 deficient (C 2 H 4 rich) feed ( Figure 8C). 2. Due to the dominant reaction of C 2 H 4 hydrogenation, the C 2 product distribution deviated positively from the ideal ASF distribution for n = 2. 3. The C 4 product distribution deviates slightly from the ASF distribution. This can be attributed to C 2 H 4 dimerization, which would enhance C 4 production. Figure 8 indicates that, in the absence of CO, the reaction of C 2 H 4 and H 2 could also produce the monomers required for chain growth. Moreover, products with both even and odd carbon numbers follow the ASF product distribution. This indicates that the chain growth monomer is not only C 2 and that some adsorbed intermediate CH x might be produced by hydrocracking of C 2 H 4 .

Carbon chain growth
Studies are still being done to determine whether the C 2 intermediate formed on the surface of a Co-based catalyst under FT reaction conditions is ethyl, ethylidene, or vinyl. In addition, it has not been determined whether the carbon chain-growth monomer of the C 1 intermediate is CH, CH 2 , or CH 3 in the FT reaction. However, the large amount of 2-butene obtained from these experiments indicates that adsorbed ethylidene (CH 3 -CH-*) probably formed on the catalyst surface. Kokes 31 reported that the 2-butene dimer monomer is ethylidene.
In order to understand the carbon chain growth of C 2 H 4 and H 2 reaction under typical FTS operating conditions, the chain growth probability (α-value) was calculated using Equation 5.

Ln
( Wn Where: Wn is the weight fraction of hydrocarbon product containing n atoms; n is the carbon number. The results are summarized in Table 4. Under conditions of excess H 2 in the feed, the product is almost entirely paraffinic, and as the temperature increases, the α-value decreases. Similarly, the α-value of the total product (including α-olefins and n-paraffins) decreases with increasing temperature for H 2 limited feeds. The trend in the α-value of the individual n-paraffins and α-olefins with temperature is also consistent. This indicates that high temperature is more conducive to the formation of low carbon chain products. In addition, the a-value of the alkane higher for H 2 limited feeds that for feeds with excess H 2 . This indicates that adsorbed H 2 limits the chain growth, as if the chain-growth intermediate is hydrogenated by the surface adsorbed H 2 it forms a paraffin and results in the termination of the carbon chain.
The α-value of total n-paraffins + α-olefins, individual paraffins and olefins under FTS with a syngas feed were calculated and are shown in Table 4. The α-value of total n-paraffins + α-olefins is 0.78, while the α-value of the paraffin is 0.82 and olefins is 0.76. A comparison of α -values of the two feed gases (CO/H 2 and C 2 H 4 :H 2 ) shows that a typical FTS has a much higher carbon chain growth factor than that of C 2 H 4 hydrogenation. This result indicates that: the either less chain-growth monomer is formed by the reaction of C 2 H 4 and H 2 than the reaction of CO with H 2 or that the monomer is less reactive and that the presence of CO probably has a strong inhibitory effect on hydrogenation of the chain-growth precursor and monomer (chain termination reaction). The CO essentially competes with H 2 for adsorption sites on the catalyst surface; thus, the presence of CO reduces the amount of H 2 adsorbed on the catalyst surface. As this H 2 reacts with chain-growth precursor and monomer (chain termination reaction), it has a significant limiting effect on the chain growth reaction. Based on this idea, the question arises: is it possible to control the α-value of the hydrogenation-type chain growth product by controlling the amount of H 2 in the adsorbed state?

Possible C 2 H 4 reactions
When combining our research results with the theories in the literature, the reactions of C 2 H 4 with H 2 could be classified as follows: In order to compare the effect of the different experimental conditions on the product distribution, Table 5 lists the product selectivity for C 1 to C 4 .
In this study, C 2 H 4 was the sole source of carbon. Thus all products, both odd and even carbon number hydrocarbons, all form from C 2 H 4 , which makes it relatively simple to recognize the possible pathways or secondary reactions of C 2 H 4 in FTS; these reactions may be limited by CO and/or the products of classical FTS; however, these effects do not occur in this study. As the experimental results show, the selectivity of ethane was more than 94% under all the reaction conditions, thus confirming that C 2 H 4 hydrogenation is the dominant reaction.
When C 2 H 4 hydrogenates to produce ethane in FTS, it changes the ratio of paraffin to olefin, but has no direct impact on the chain growth probability. As shown, high reaction temperatures and a high ratio of C 2 H 4 /H 2 could inhibit C 2 H 4 hydrogenation to some extent, which may be attributed to activation of the C-C bond at a high temperature. However, this change in the reaction kinetics was small and had no significant effect on the product distribution.
The C 2 H 4 dimerization is a very interesting reaction, since only a small amount of the main dimer (2-butene) is generated when in FTS over a Co catalyst. The generation of 2-olefins is routinely attributed to bond-shift isomerization of the corresponding α-olefins in FTS. 9 Based on this theory, it is surmised that 2-σ-alkyl species generated from partial hydrogenation of α-olefins, and then dehydrogenated to form 2-olefins. However, Kokes established a model of C 2 H 4 dimerization based on the results of a study that used hydrogenation of C 2 H 4 over bulk Co. 31 In this model, C 2 H 4 was reversibly bridge adsorbed on the surface of Co, and transformed to the surface intermediate 1,1-σ-ethylidene with the assistance of the dissociated hydrogen atoms. Two molecules of this intermediate encountered head-to-head to generate cis-2-butene, while 1-butene was produced by a head-to-tail encounter. As C 2 H 4 could react both as an initiator and monomer under FTS conditions, it could be proposed that C 2 H 4 partially hydrogenates to form an ethyl surface species as the chain initiator, and another adsorbed C 2 H 4 inserts to generate 1-butene. In our view, 1-butene might form from both C 2 H 4 direct dimerization and H 2 assistant carbon chain growth. We also believe that these results support that trans-2-butene is a secondary product isomerized by cis-2-butene and that cis-butene, trans-butene, and that 1-butene could be interconverted through internal hydrogen transformation.
C 2 H 4 hydrogenolysis used to form CH 4 is sensitive to reaction temperature and the quantities of chemisorbed hydrogen. Shi and Davis argue that co-fed C 2 H 4 could decrease CH 4 selectivity to some extent. 26 The current experiments show extremely low selectivity to CH 4 for H 2 limited feeds and is in agreement with this conclusion. This might be because C 2 H 4 reacts as a monomer in carbon chain growth rather than producing CH 4 from hydrogenolysis. C 2 H 4 either reacted directly as a monomer, or demethylased to form methylene, which participated in the chain growth reaction and could be expressed to analogical comproportionation and disproportionation. These reactions TA B L E 5 The selectivity of products C 1 to C 4 on ethylene hydrogenation condition generated odd carbon number products, which affected the product distribution. As the results show, the comproportionation and disproportionation were favored in H 2 limited feeds where there is less dissociated H 2 on the catalyst surface, which is in agreement with the accepted FTS models on Co catalysts. In previous work done by the researchers, Lu et al reported an olefin quasi-equilibrium reaction existed in FTS when using the same catalyst. 13 This proposed quasi-equilibrium equation is similar to the analogical comproportionation and disproportionation reaction proposed in this work. In brief, we attempt to isolate the C 2 H 4 reaction system in this work, and find various C 2 H 4 reaction occurs. This is the first step in investigating and understanding the C 2 H 4 reactions in normal FTS. In normal FTS, the strong adsorption of CO on the catalyst surface results in competitive adsorption between CO and olefins (C 2 H 4 in this work). This competition directly suppresses some C 2 H 4 reaction (like dimerization). Moreover, because of the presence of CO, some important phenomenon is easy to be ignored. As we find in this work, in the absence of CO, the ethylene reacted with H 2 to produce products of C 3 , C 5 and C 6 , which fitted the typical FTS product distribution. This phenomenon means FTS-type chain growth is not necessarily only derived from the hydrogenation of CO. The chain growth initiators and monomers required for the FTS reaction can also be formed by olefin hydrocracking, although the chain growth probability in these reactions is much lower than in normal FTS. It may give us a new understanding about the FTS mechanism when viewed from this different angle.

CONCLUSION
With the aim of investigating the feasible reactions paths of C 2 H 4 and their product distribution in FTS, a series of experiments were designed over a Co/TiO 2 catalyst in the absence of CO. When using C 2 H 4 /N 2 as the feed gas, no products could be detected under any of the reaction conditions tested. However, when co-feeding H 2 with C 2 H 4 over a range of operating temperatures, C 2 H 4 reacted, even at 100 • C. Short chain hydrocarbons (including CH 4 , C 3 H 6 , C 3 H 8 , 1-C 4 H 8 , cis-2-C 4 H 8 , trans-2-C 4 H 8 , n-C 4 H 10 , etc.) were formed.
In the presence of hydrogen, although the dominant product was ethane, varying the temperature and partial pressure of the reactants could lead to the other reaction pathways being enhanced, resulting in varying product selectivity to CH 4 and C 3-6 olefins and paraffins. C 2 H 4 hydrogenated to form ethane was slightly inhibited by higher reaction temperatures. The main C 4 product produced from C 2 H 4 /H 2 feeds was 2-butene, from C 2 H 4 dimerization. C 2 H 4 hydrogenolysis and demethylation reaction rates were affected by the H 2 partial pressure and therefore the quantity of chemisorbed hydrogen. The selectivity of the products of CH 4 , C 3 , and C 5 with odd carbon numbers increased with increasing temperature. These results indicated that a high temperature preferred the production of odd carbon number hydrocarbons.
In the absence of CO, the products obtained from the C 2 H 4 reacted with H 2 fitted a typical ASF product distribution. This indicated that both the chain-growth monomers and initiators were not necessarily only derived from CO hydrogenation, but also from C 2 H 4 reactions. In FTS, C 1 normally lay above the ideal ASF distribution, while when using C 2 H 4 /H 2 feeds, the C 1 selectively lay on or even slightly below the ideal ASF plot. The α-values for the C 2 H 4 hydrogenation products was lower (0.32) than that obtained for FTS (0.82) indicating that the rate of termination of the chain growth was higher. Although the feeds used in this work were not the same as that used in normal FTS, the results obtained suggested that C 2 H 4 followed various reaction pathways to form different products, and that it acted as a monomer and as an initiator in the chain growth reactions, thus might affect the product distributions during FTS. These results could have implications for the reaction mechanisms of FTS.