3.1. Upgraded Oil API Gravity and Viscosity
API gravity is an indicator of crude oil lightness by comparing its density to that of water, it is used to classify oils into light, medium, heavy, or extra heavy, and also one of the scale that determines its market value; hence, the larger the API gravity, the lighter the oil. The effect of reaction environment (i.e., hydrogen rich and nitrogen rich) on the upgraded oil API gravity and viscosity as a function of time-on-stream is in shown
Figure 1. Under nitrogen environment, the API gravity increase started from 6°, decreased promptly to ~2.5° at 200 min, and afterward settled at about 2.2° for the rest of the reaction. In contrast, under hydrogen atmosphere the upgraded oil API gravity decreased from 7° to approximately 4° and stabilised at 3.7° throughout the rest of the experiment.
On average the upgraded oil API gravities are 3.1 ± 0.6° (nitrogen) and ~5 ± 0.5° (hydrogen) above 13° for the THAI feed oil (
Figure 1a). This represents approximately 2° increase attributable to the addition of hydrogen. Similar results have been reported in references [
12,
13,
14,
15].
The absolute viscosities of the upgraded oils are 0.093 Pa·s (nitrogen) and 0.075 Pa·s (hydrogen) compared to the 0.5 Pa·s (THAI feed oil). This represents approximately 81% and 85% viscosity reductions under nitrogen and hydrogen, respectively (
Figure 1b). Wang et al. [
33] ascribed the slightly further reduction of viscosity observed under hydrogen environment to HDS and hydrogenation reactions, which potentially increased the amount of light hydrocarbons produced compared to when nitrogen was used. It has been reported that significant amounts of hydrogen and hydrogen-rich gases such as H
2, CH
4 and C
2H
6 were stripped off from the heavy crude oil into the gas phase [
13,
15]; thus the upgrading achieved under a nitrogen environment can therefore be ascribed to carbon-rejection due to cracking of macromolecular weight species such as resins and asphaltenes. The radical fragments from the cracked macromolecules can readily regroup to form large hydrocarbon compounds, whilst under hydrogen environment these free radicals are readily scavenged by active hydrogen to form lower molecular weight hydrocarbons [
13].
At a significance level of 0.05 (5%), the one-tailed probability-value (p-value) of a Z-test for the experimental data at the hypothesized dataset mean of 4° API gravity and 0.08 Pa·s. under nitrogen environment, the p-value is 0.9997 and with hydrogen environment it is 0.0149. Since the p-value, 0.0149 (under H2), is less than 0.05, the null hypothesis (mean = 4° API) is rejected in favour of the alternative, that is the upgraded oil API gravity is greater than 4°. For the viscosity, while the p-value, 0.0348 (under N2) is less than 0.05, rejecting the null hypothesis of equal to 0.08 Pa·s, the p-value of 0.7696 (under H2) favours it. This implies the viscosity of the upgraded oil under H2 is either less than or equal to 0.08 Pa·s.
3.2. Effect of Hydrogen Pressure on Upgraded Oil API Gravity and Viscosity
The effect of hydrogen pressure on the level of upgrading was investigated at 425 °C for the range of pressures 20 to 40 bar. The upgraded oil samples API gravity for the range of pressures studied as a function of time-on-stream is shown in
Figure 2a. Although the API gravity of upgraded oils increased narrowly as hydrogen pressure increases from 20 to 40 bar with an average of 0.3°per 10 bar increase (
Figure 2a), this could be possibly due to improved availability of hydrogen for hydroprocessing reactions as increased hydrogen pressure could have lowered mass transfer limitation from the gas-phase to the oil-phase [
34,
35].
The viscosity of the upgraded oil samples when hydrogen pressures were increased from 20 to 40 bar is shown in
Figure 2b. Just like the API gravity, the upgraded oils showed marginal decrease in viscosity when the hydrogen pressure was increased from 20 to 40 bar. The mean degree of viscosity reduction (DVR: (μ
f − μ
u)/μ
f, where the subscript represents feed and upgraded oils) increase negligibly by 1% for every 10 bar increase in pressure. Since the changes in the API gravity and viscosity of the upgraded oils as hydrogen pressure increases fall within the standard deviations, it shows that hydrogen does not exert much influence on the physical properties of the upgraded oil as reaction temperature and catalyst loading [
17], but rather steers the chemistry of the reaction and its interactions with the catalyst surface to produce low-boiling hydrocarbons and suppress carbon-rejection.
In contrast, it has been reported that reaction temperature and WHSV exert significant influence on the level of upgrading achievable [
17]. It is well known that the catalytic upgrading reaction involves the breaking down of large molecular species into smaller ones; the cracked radical intermediates however can reassemble into larger molecular weight species in the absence of a hydrogen-donor to hydrogenate them. This hypothesis could be confirmed in the presence of nitrogen, as their API gravities were lower and their viscosities slightly higher than those obtained when the upgrading reaction was carried out with hydrogen (
Figure 1). The slight improvement in API gravity and viscosity of the upgraded oils when the hydrogen pressure was increased beyond 20 bar, reaffirmed the increased hydrocracking and hydrogenation functionalities of the CoMo/Alumina catalyst being enhanced as more hydrogen becomes available. At the significance level of 0.05 and null hypothesis of 5° API gravity increase, the
p-values are as follows 0.2543 (H-20 bar), 0.007 (H-30 bar) and 4.8 × 10
−8 (H-40 bar). In this respect, at 20 bar pressure the null hypothesis is accepted while at 30 and 40 bar pressures the API gravity increases are most likely to be greater than 5° API gravity increase. Conversely, with a null hypothesis of 0.06 Pa·s for the upgraded oil viscosity, the
p-values are 0.798 (H-20 bar), 0.998 (H-30 bar), and 0.9999 (H-40 bar), respectively.
Maipur et al. [
36] proposed an empirical equation for estimating hydrogen consumption, in this study however the outlet hydrogen flow was not measured; hence, the total hydrogen uptake in moles during the upgrading reactions was not calculated. However, hydrogen concentrations in the outlet gas phase decreased by 13.3% (20 bar), 17.33% (30 bar), and 19.83% (40 bar) relative to 99.99% pure hydrogen fed. This is indicative of hydrogen involvement in hydroprocessing reactions, which increases as pressure increases [
37,
38,
39,
40].
3.4. Upgraded Oil Asphaltene and Spent Catalyst Coke Contents
Asphaltenes are among the largest and heaviest polar component of heavy oil and thus are readily deposited upon catalyst surfaces as coke-precursors. Being a major contributor to coke formation, hydrogenating radicals formed when they are cracked could reduce the asphaltene content of the produced oil and potentially lead to a longer catalyst lifetime. The asphaltene content of the upgraded oil samples can be summarised as thus 8.6 ± 0.6 (N
2, 20 bar), 7.8 ± 0.4 (H
2, 20 bar), 5.7 ± 0.3 (H
2, 30 bar), and 5.8 ± 0.4 (H
2, 40 bar) compared to 11.2 wt % (THAI feed oil). Though the asphaltene contents of the upgraded oils were lower than that of the THAI oil, the presence of hydrogen further decreased it due to the hydrogenation of cracked fragments which is rarely experienced in the presence of nitrogen. This reaction involves hydrogen transfer from the gas-phase to the macromolecular radicals in the oil phase, which is possible under high reaction temperatures such as 425 °C and high hydrogen pressure [
42,
43].
The coke contents of the recovered catalyst after experiment with nitrogen (20 bar) and hydrogen (20 to 40 bar) as determined using TGA are presented in
Figure 4. It has been reported that the burn-off beyond 600 °C represents coke [
44]. The thermogram (TG), that is weight loss with temperature and its differential (DTG) curves show that the coke formation under nitrogen environment is higher (35.4 wt %) compared to hydrogen (27.2 wt %).
With increasing hydrogen pressure from 20 to 40 bar, it was observed that the catalyst coke content decreased from 27.2 to 17.3 wt %, signifying that coke formation under hydrogen environment was sensitive to pressure. Hence, the activity of the catalyst can be sustained long enough compared to about 90 h observed by Shah et al. [
10] with nitrogen environment, as the susceptibility of the catalyst to coke fouling has been decreased with hydrogen, following the lower coke formation observed compared to when nitrogen was used.
Zhang and Shaw [
45] and Matsumura et al. [
46] observed a similar trend in coke content of the catalyst as hydrogen pressure was increased. Thus, increasing the hydrogen pressure could have improved the transfer and the solubility of hydrogen in the oil-phase. Higher hydrogen pressure provided more hydrogen in the vicinity of the catalyst surface, which is thermodynamically favourable for hydroprocessing reactions considering the reaction temperature of 425 °C [
36,
37,
47].
The TGA only quantifies the total amount of coke deposited on the catalyst after 11 h of experiment; to comprehend the extent of pore plugging mercury porosimetry and nitrogen adsorption-desorption were used. The catalyst pore size distribution before and after 11 h of upgrading reactions under nitrogen and hydrogen is shown in
Figure 5.
Figure 6 shows the nitrogen adsorption-desorption isotherm for the fresh and spent CoMo/Alumina catalyst. It can be observed that less coke was formed with hydrogen-addition (
Figure 4 and
Figure 5), while the entire pores of the catalyst after reaction under nitrogen was utterly plugged (
Figure 5). A narrow pore size distribution can be observed after the upgrading reaction under hydrogen. Similar observation using the nitrogen adsorption-desorption isotherm after upgrading reactions under hydrogen and nitrogen has been reported by Hart et al. [
12,
13]. This observation reaffirmed the suppression of coke formation via hydroporcessing reactions such as hydrocracking, hydrotreating, and hydrogenation of intermediate radicals, olefins, and polynuclear aromatics once they are formed [
48,
49].
In
Figure 6, the isotherm of the fresh CoMo/Alumina revealed it is type IV which is characterised by meso-pores with specific surface area of 214 m
2·g
−1. Compared to the fresh catalyst, the spent catalyst showed a remarkable drop in nitrogen adsorbed-desorbed as the relative pressure approached 1. This is indicative of loss in pore volume and porosity due to coke deposition; hence, the specific surface areas were decreased to 59.4 m
2·g
−1 (after upgrading with hydrogen) and 2.03 m
2·g
−1 (after upgrading with nitrogen). While the spent catalyst after the upgrading reaction under nitrogen environment experienced almost total loss of catalyst surface area and pore volume due to high coke formation, which obtained with hydrogen experienced moderate loss of area and pore volume, consistent with the TGA (
Figure 4) and porosimetry (
Figure 5).
Since the catalyst is less prone to pore plugging and loss of surface due to lower coke formation when the upgrading reaction occurs under H-rich environment, the catalytic activity is prolonged significantly compared to N
2 environment. Hence, the catalyst achieved an additional increase in API gravity of 1–2° over that obtained when the experiment was carried out under N
2 (
Figure 1). The reality is that this level of in situ partial upgrading with H
2 is still valuable to oil industries as it is worth approximately
$0.5–
$1.5 per API point and up to
$9/barrel depending on oil price [
8].
Figure 5 also shows that significant upgrading would have occurred at the early hours of the experiment as observed in
Figure 1; before the catalyst pore channels were plugged, allowing only the low molecular weight hydrocarbons to access the pores and then get cracked. Once the surface and pores of the catalyst were covered by coke the level upgrading drops rapidly as noticed in
Figure 1, especially when the upgrading reaction was performed under nitrogen. Also, the polynuclear aromatics adsorbed onto the external surface of catalyst, could prevent the optimum utilisation of the internal surface of the catalyst when the upgrading reaction was carried out under N
2. The higher amount of coke formed under nitrogen environment compared to hydrogen (
Figure 4 and
Figure 5), reaffirmed that the level of upgrading achieved with nitrogen can be attributed mainly to carbon-rejection (
Figure 1).
The SEM photomicrograph of the catalyst was studied over an area of 2 μm widths and magnification of 35,000×.
Figure 7 shows the photomicrographs of the fresh and coked catalyst after upgrading reactions in the presence of nitrogen and hydrogen. The surface morphology after upgrading reactions shows carbonaceous deposits such as precipitated asphaltene and coke on the surface of the catalyst.
The coke formed an amorphous encapsulate of the catalyst surface and revealed pore plugging. It is clear that the catalyst experienced severe pore plugging and coking when the upgrading reaction was carried out under nitrogen environment (
Figure 7b) compared to when hydrogen was used (
Figure 7c). This reaffirmed the observations in
Figure 4,
Figure 5 and
Figure 6. Notably, the coke formed on the surface of the catalyst after upgrading reactions under hydrogen environment was mostly spheroids of size ranging from nano to micro-meters globules, while when nitrogen was used; the coke was an amorphous compact ground mass with few globules.
The effect of connate water on the catalyst was not investigated as the focus is mainly deactivation due to coke deposition and catalyst pore plugging. Also, the heavy oil used in this study as received from Petrobank Energy and Resources Ltd. is approximately free of water. However, under field production at a temperature range of 400– 700 °C most of the connate water will be converted into steam and with the aid of the catalyst and the rock minerals more hydrogen can be produced through the water-gas shift reaction (CO + H
2O → H
2 + CO
2) and steam-methane reforming reaction (CH
4 + H
2O → CO + 3H
2) [
17,
18,
21]. Consequently, in addition to hydrogen being expensive, it could be challenging to introduce it across the CAPRI zone. Hence, water is a potential alternative source of hydrogen and its potential has been reported for heavy oil upgrading with supercritical water, steam cracking, and catalytic aquathermolysis [
50].The potential of the CoMo/Alumina catalyst to promote hydrogen production from water through the water-gas shift reaction and simultaneously support hydroprocessing reactions by stabilizing the oil cracked while suppressing coke formation has been reported in the literature [
18].
Figure 8 depicts the intimate contact between the oil and the catalyst, which resulted in the occurrence of the upgrading reaction as the hot oil and combustion gases, including hydrogen, would flow across the catalyst bed taking advantage of the well-bore pressure and temperature to further crack the heavy oil. The cracked heavy hydrocarbons into lighter molecular weight hydrocarbons led to the improved viscosity and API gravity of the upgraded oil (
Figure 1), and as a consequence coke was deposited on the catalyst as illustrated in
Figure 8 and observed in
Figure 4,
Figure 5,
Figure 6 and
Figure 7. The rapid decrease in the upgraded oil API gravities presented in
Figure 1a from 20–300 min can be attributed to catalyst deactivation due to coke deposit, as confirmed in
Figure 4,
Figure 5,
Figure 6 and
Figure 7 which is an indication of the heavy hydrocarbons being cracked into lower fraction hydrocarbons [
17,
27]. Heavy metals (e.g., V, Ni, and Fe) and coke deposits have been reported by Leyva et al. [
27] as major contributors to catalyst deactivation due to the cracking of macromolecular weight species containing heteroatom (e.g., S and N) and heavy metals such as V and Ni during hydroprocessing of heavy oil as illustrated in
Figure 8.
The cracking of the adsorbed macromolecular weight hydrocarbons on the catalyst surface is aided by the support acid sites while the dissociated hydrogen proton (H•) reacts with a heteroatom such as sulphur due to the hydrotreating functionality of the impregnated Co and Mo on the support, and is removed as H
2S [
17,
27]. The other fragments of active hydrocarbon intermediates (i.e., R
1 and R
2) are hydrogenated to stable hydrocarbon molecules and released into the oil phase while the coke and metallic sulphides are deposited on the surface of the catalyst [
27]. The organometallic hydrocarbons in the heavy oil deposits metals (e.g., V, Ni) on the surface of the catalyst as metallic sulphides (M
xS
y) as illustrated in
Figure 8 [
27]. Hence, this reaction is dependent on hydrogen availability and catalyst activity. However, in a hydrogen limited medium such as a nitrogen environment, the intermediate fragments can readily aggregate to form bigger hydrocarbon molecules, which explains why the upgraded oils under this environment have lower API gravities and lower fuel distillate fractions compared to those obtained when the upgrading reaction was carried out with hydrogen (
Figure 1 and
Figure 3). As a consequence of starved hydrogen when the reaction was carried out under nitrogen, the crack hydrocarbon radicals polymerised and condensed into higher molecular weight species leading to higher coke formation and catalyst pore plugging observed in
Figure 5,
Figure 6 and
Figure 7. Whilst under hydrogen the cracked radicals are stabilised into the oil-phase hence the upgraded oil has lower asphaltene component [
13,
17].
Heavy oil is a complex mixture of different classes of hydrocarbons grouped into paraffins, olefins, naphthenes, aromatics, hetero-atomic compounds, and poly-nuclear aromatics (e.g., resins and asphaltenes).
Figure 9 summarises generic pathways the different classes of hydrocarbons found in the heavy oil would undergo depending on the reaction environment. Under nitrogen environment, a higher amount of coke was observed as confirmed in
Figure 4,
Figure 5,
Figure 6 and
Figure 7 as well as higher catalyst deactivation can be noticed from 20–300 min in
Figure 1 compared to when hydrogen was used. These observations are possible if the reactions that result in coke formation are amplified under nitrogen environment such as dehydrogenation, polymerisation, and condensation of aromatics into polynuclear aromatics, including the cracking of macromolecular weight hydrocarbons into lighter fractions. While cracking and ring opening are mostly temperature dependent, hydrogenation, dehydrogenation and polymerisation/condensation are largely promoted by the reaction environment under favourable pressure. As a consequence, olefins produced by dehydrogenation of paraffins, thermal or catalytic cracking of the oil and the aromatics fractions under nitrogen environment would have readily polymerised into larger molecular weight compounds compared to when hydrogen was used.
On the other hand, the presence of hydrogen supported and promoted hydrocracking, hydrogenation, and hydrotreating (HDS, HDN, and HDM) reactions, while nitrogen environment favoured cracking, dehydrogenation, polymerisation, and condensation of aromatic ring reactions which resulted in high carbon-rejection as is observed in
Figure 4,
Figure 5,
Figure 6 and
Figure 7. We have previously shown that more aliphatic hydrocarbon and less olefinic gases were produced when the upgrading reaction was performed under hydrogen environment compared to nitrogen [
13]. This proves that in the presence of nitrogen, hydrogenation is very limited and carbon-rejection was dominant during the upgrading process. This was affirmed by the presence of more hydrogen and H-rich gases such as methane and ethane in the gas-phase [
13,
15], and as a consequence higher coke formation and catalyst pore plugging was observed when nitrogen was used as was confirmed from the analysis of the spent catalyst coke content and its pore plugging presented in
Figure 4,
Figure 5,
Figure 6 and
Figure 7. However, with hydrogen environment, hydroprocessing (i.e., hydrocracking, hydrogenation, and hydrotreating) reactions were amplified as shown in
Figure 9, including hydrogenation of free radicals, unsaturated hydrocarbons, and aromatics which inhibited polymerisation reactions and increased the lighter fractions of the upgraded oil. These promoted reaction pathways in
Figure 9 due to the presence of hydrogen contributed towards the additional 0.5–2° API increase observed in
Figure 1 and the more naphtha and middle distillate fuel fractions obtained upon simulated distillation presented in
Figure 3 over those attained when nitrogen was used. Hence, the upgraded oils obtained with hydrogen environment were lighter and richer in lower molecular weight hydrocarbons than those obtained when nitrogen was used.