1. Introduction
Green hydrogen (H
2) could play a crucial role in the realization of carbon neutrality, since its production can achieve the clean hydrogenation of chemicals, and it can be directly used as a fuel for transportation and household cooking [[
1], [
2], [
3], [
4]]. The ideal source of green H
2 (
Fig. 1) is definitely water pyrolysis with renewable electricity (solar and wind energy) [
5,
6]. However, the current capacity to produce such green H
2 is far below that of grey H
2 (
Fig. 1) from fossil fuels (coal, natural gas, or liquid fuels such as diesel), which exceeds 50 million tonnes annually worldwide [
2,
3]. Industrially, the production of grey H
2 usually demands the auxiliary conversion of oxygen (in the forms of water (H
2O) and oxygen gas (O
2), resulting in significant carbon dioxide (CO
2) emissions [[
7], [
8], [
9]]. Although the direct transformation of hydrocarbons by means of chemical vapor deposition (CVD) would only produce H
2 as a gaseous product in theory (
Fig. 1), this process tends to generate large amounts of methane (CH
4)—a much more serious greenhouse gas than CO
2 [
1,
2]. Even worse, most hydrocarbons undergo endothermic reactions under high-temperature conditions and exhibit high energy consumption, leading to significant indirect CO
2 emissions (when burning fuel for energy).
Herein, we propose the production of “jadeite H
2” from the direct catalytic decomposition of cheap diesel (i.e., light cycle oil (LCO)) in a thermodynamically favorable and exothermic process. LCO, which mainly consists of naphthalene-series compounds, has a hydrogen/carbon (H/C) weight ratio similar to that of coal. Compared with most research in this field, which relies on a selection of noble metal catalysts (Rh, Pt, Pd, etc.) for the steam reforming of LCO to produce grey H
2 [[
10], [
11], [
12]], the proposed method very selectively converts LCO into H
2 and carbon nanotubes (CNTs) using a cheap but highly active nanosized iron (Fe)-based catalyst. We define the molar ratio of H
2/CH
4 and other gaseous hydrocarbons as the green index (GI), which can be used to compare the proposed process with the CVD transformation of other hydrocarbons. The proposed technology has the highest GI values among all reported CVD processes using hydrocarbons to produce H
2. In addition, the decomposition of LCO offers the greatest practical feasibility, based on an evaluation of the energy consumption and operating temperature range of this process, as compared with other industrial routes (e.g., turquoise H
2 from methane pyrolysis (MP) and green H
2 from water electrolysis with fossil energy). Considering that the emissions can be shared by the CNT products, this route has the lowest carbon footprint (CFP) of any CVD transformation of hydrocarbons; thus, it can be used to achieve a net-zero carbon or negative carbon effect. Since huge amounts of diesel (in this case, LCO) are available in practice, in addition to diesel’s advantages of high density and easy transport ability [
13,
14], the as-produced clean H
2 holds promise to serve as a clean distributed H
2 source to fill the gap before a sufficient amount of green H
2 can be obtained from water electrolysis.
2. Materials and methods
2.1. Preparation of Fe/Mo/Al2O3 catalyst
A nanosized Fe-based catalyst was prepared using a co-precipitation method. First, Fe(NO
3)
3·9H
2O and Al(NO
3)
3·9H
2O were dissolved in 500 mL of deionized water, while (NH
4)
6Mo
7O
24·4H
2O and (NH
4)
2CO
3 were dissolved in 200 mL of water. The two solutions were then stirred together at a temperature of 80-100 °C, with an addition rate of 1.5 g∙min
−1. Precipitants gradually formed and were ripened overnight. Then, NH
4+ was removed by washing with deionized water five times. Finally, the precipitants were filtrated to remove the solution, dried at 120 °C for 12 h, and calcinated at 450 °C for 6 h. The final catalyst obtained in this way had a metal ratio of 4:5:15 (Fe:Mo:Al
2O
3). A nickel (Ni)-based catalyst was also produced using a previously reported method [
15].
2.2. Preparation of jadeite hydrogen and CNTs
LCO was used as the carbon source, with the addition of nitrogen (N2) as the inert carrier gas. In general, 10-30 g of catalyst powder was placed into a fluidized-bed (FB) reactor made of stainless steel with an inner diameter of 50 mm. The reaction temperature was 700-800 °C, and the space velocity of the carbon source was 0.025-0.15 h−1. The operating pressure was ambient pressure at the exit of the reactor. The collected solid products were immersed in 200 mL of hydrochloric acid (10 wt%) for 24 h, followed by washing with 50 mL of deionized water three times. After drying at 120 °C for 4 h, CNTs were obtained as the final product. To verify the high GI value for LCO using the proposed technology, 1-methylnaphthalene, liquefied petroleum gas (LPG), and pentane were also used as carbon sources.
2.3. Characterization
Online gas chromatography (Agilent 7890B, Agilent, USA) was used for the gas analysis, where the gas products were sampled every 30 min. Scanning electron microscopy (SEM; GSM-7401; Japan Electronics (JEOL), Japan) and transmission electron microscopy (TEM; JEM-2010; JEOL) were used to observe the surface morphology of the solid products, while Raman spectroscopy (Labram-800; HORIBA JOBIN YVON, France) was used to confirm the existence of the CNTs and characterize their structure. A TGA/DSC 1 instrument (STARe; Mettler Toledo, Switzerland) was used to estimate the purity of the CNTs. X-ray diffraction (XRD; D8 Advance; Bruker, Germany) was also used to determine the configuration of the catalysts and solid products. The redox properties and acidity of the catalyst were measured by means of hydrogen temperature-programmed reduction (H2-TPR) and ammonia temperature-programmed desorption (NH3-TPD), respectively (Chembet PULSAR TPR/TPD, USA). X-ray photoelectron spectroscopy (250XI, Thermo Scientific, USA) was used to detect the surface structure of the catalyst.
3. Results and discussion
3.1. Technical issue in the catalytic decomposition of diesel
As shown in
Fig. 2(a), the jadeite H
2 was produced by means of a FB-CVD method. Feedstocks of LCO (a liquid mixture of C
10-C
20 paraffin, alkylbenzenes, naphthalenes, biphenyls, anthracenes, and phenanthrenes;
Fig. 2(b) and Table S1 in Appendix A), were fed from the bottom of the reactor. As the feedstock was decomposed via the nanosized Fe/Mo/Al
2O
3 catalyst (
Fig. 2(c) and Fig. S1 in Appendix A) at 700-800 °C, CNTs and gaseous products (mainly H
2 and CH
4) were produced with high efficiency (
Fig. 2(a)). The solid products were mostly entangled multi-walled CNTs with outer diameters of 8-15 nm and inner diameters of 6-10 nm (
Figs. 2(d) and
(e)). As shown in
Fig. 2(f), the purity of the CNTs was approximately 84%, and the derivative thermogravimetry (DTG) pattern displayed a burning peak at 470 °C, indicating that there was no amorphous carbon formation. As shown in
Fig. 2(g), the intensity ratio of the D band to G band in the Raman spectra was around 0.5-0.6. The peak of the two-dimensional (2D) band showed high intensity, and the 2D/G intensity ratio was also large (0.57). These results all indicate that high-quality CNTs (i.e., with few defects) were successfully prepared from the LCO [
16,
17].
As CNTs were produced from the LCO, the Fe catalyst exhibited very high activity to decompose the LCO completely (
Figs. 3(a) and
(b)). The distribution of gas products differed depending on the reaction time (
Fig. 3(a)). The CH
4 selectivity was 25% at the start of the reaction but dropped linearly to 2% within 1 h and then remained at that low value from 1 to 12 h. Moreover, the volume ratio of the other C
2-C
6 hydrocarbons was lower than 0.1% during the 12 h reaction. According to the law of conservation of mass, the mass balances of carbon (C), hydrogen (H), and sulfur (S) were 86.2%, 136.3%, and 94.4%, respectively (Table S2 in Appendix A). When 1-methylnaphthalene was used as the feedstock, the C
3-C
5 hydrocarbons were effectively transformed, since the LCO had a complex composition and it was difficult to convert the C
2-C
6 hydrocarbon byproducts into CNTs and H
2 (Fig. S2 in Appendix A). Such a stable product distribution also indicated the high activity of the catalyst, with no obvious deactivation trends.
It appears that the main reaction in the LCO conversion was the breaking of carbon-carbon bonds and gradual dehydrogenation into pure carbon products via a vapor-liquid-solid (VLS) mechanism during the CVD process [[
18], [
19], [
20]]. The side reaction was the hydrogenation of intermediates to form CH
4, C
2H
4, C
2H
6, and C
3H
6. Among these, C
2-C
6 hydrocarbons with high activity, once produced, can be further transformed into carbon and H
2 [[
21], [
22], [
23], [
24]]. Therefore, the key to obtaining highly selective H
2 lies in the suppression of CH
4 byproducts.
In comparison, LCO was also effectively transformed into CNTs with the Ni-based catalyst (Figs. S3-S5 in Appendix A). As demonstrated in
Figs. 3(c) and
(d), the decomposition of LCO with the Ni-based catalyst gave similar gaseous products but with different selectivity. The CH
4 selectivity was 10%, slightly better than that of the Fe-based catalyst. However, the CH
4 selectivity increased continuously from 10% at 1 h to 20% at 10 h and 30% at 12 h. Meanwhile, the initial H
2 selectivity was only 87% and gradually dropped to 80% at 10 h and 60% at 12 h. Strangely, the LCO was always completely converted within 12 h, with the H
2 selectivity declining dramatically. As shown in Fig. S5, the CNTs prepared using the Ni-based catalyst had a larger diameter, 20-60 nm greater than that of the Fe-based catalyst (8-15 nm). The Raman spectra also showed that the intensity ratio of D-band to G-band (the
ID/
IG ratio) of the CNTs prepared with the Ni-based catalyst was higher than that of the Fe-based catalyst indicating a higher degree of defect (Fig. S6 in Appendix A).
Given that the CH4 and other hydrocarbon products were components of volatile organic compounds and much more serious greenhouse gases than CO2, we summarized the gross molar ratio of the hydrocarbon product as the molar ratio of CH4 equivalent (CH4 eq) for a total amount control. Then, the molar ratio of H2 to CH4 eq was defined as the total amount control. Then, the of H2. Thus, the GI value not only represents the catalyst or process efficiency from the viewpoint of the atomic economy but also assesses the carbon emissions of different hydrogen-production processes.
More specifically, in an experiment of varying space velocities of LCO (
Fig. 3(e)), the maximum GI value of the Fe-based catalyst approached 42 at a space velocity of 0.05 h
−1. Increasing the space velocity to 0.15 h
−1 resulted in a decrease in the GI value to 20. Reasonably, increasing the space velocity resulted in the excess carbon forming metal carbides with the metal catalyst. In turn, the insufficient carbon diffusion and precipitation of the CNTs increased the competitive hydrogenation to generate undesirable CH
4 [
17,
18,
20]. Likewise, prolonging the reaction time to 12 h changed the state of the catalyst, which gave a GI value of 25 at a space velocity of 0.025 h
−1. It is clear that the carbon diffusion and carbon precipitation ability of the deactivated catalyst gradually became weak, exacerbating the competitive hydrogenation and producing more CH
4.
As shown in
Fig. 3(e), the GI values with Ni-based catalyst hardly exceeded 6 and declined continuously with the reaction time, finally falling to 2.7 at 12 h. Among all reported CVD processes using a gram- or larger-scale catalyst, where the hydrocarbon feedstock included CH
4 [
25,
26], ethane [
27,
28], propane [
28,
29], and so on, it was extremely challenging to obtain highly selective H
2 during a CVD process to produce CNTs. As shown in
Fig. 3(f) and Table S3 in Appendix A, the GI values of these previously reported processes were mostly in the range of 2-7. In the present work, the GI value with a Ni-based catalyst fell within the same range as those of the previously reported processes, but the GI value with a Fe-based catalyst far exceeded them (
Fig. 3(f)) [[
25], [
26], [
27], [
28], [
29], [
30], [
31], [
32]]. The higher the GI value is, the closer the route is to achieving the desired pure clean H
2. Compared with other hydrogen-production routes (Fig. S7 in Appendix A), the diesel decomposition reported herein exhibited an optimal GI value (5-42), which was higher than those of grey H
2 (1-3) [[
33], [
34], [
35], [
36], [
37]], turquoise H
2 (5-7) [
5], and blue H
2 (3-13) [
5,
33], and just lower than that of water electrolysis (1-∞) [
5,
6,
33]. Theoretically, the GI value would be infinitely large for pure green H
2 prepared by green electricity. Therefore, to stress the relatively high greenness of the present H
2 product, the hydrogen produced by this method was named “jadeite hydrogen,” with a GI value of 5-42.
Quantitatively, the mass balance calculation indicated that the conversion of LCO with the Fe-based catalyst resulted in 91 wt% CNTs, 7 wt% H
2, and 2 wt% CH
4 (
Fig. 3(b)), equivalent to 67 mol% CNTs, 32 mol% H
2, and 1 mol% CH
4. In comparison, the conversion of LCO with the Ni-based catalyst resulted in 82 wt% CNTs, 6 wt% H
2, and 12 wt% CH
4 (
Fig. 3(d)), equivalent to 64 mol% CNTs, 29 mol% H
2, and 7 mol% CH
4. Thus, the Fe-based catalyst was clearly superior to the Ni-based catalyst in obtaining large proportions of CNTs and H
2 while inhibiting the formation of CH
4.
3.2. Comparison of the CFPs of different H2-production technologies
CFP (t CO
2eq∙t
−1 product) values were calculated, considering that the different H
2-production processes had various products. As shown in
Fig. 4(a) [[
4], [
5], [
6],[
33], [
34], [
35], [
36], [
37], [
38], [
39], [
40], [
41], [
42], [
43], [
44], [
45]] and Table S4 in Appendix A, the CFP of coal gasification was the highest (29.33 t CO
2eq∙t
−1 product) due to less hydrogen (H) being present in coal and the high energy consumption of the gasification process [
33,
38]. Next, the CFP of steam methane reforming (SMR) was around 7.33-21.86 t CO
2eq∙t
−1 product, determined via a life-cycle assessment [
37,
40,
46]. In comparison, the CFP values of propane dehydrogenation (PDH) and ethane dehydrogenation (EDH) were in the range of 1.4-3.01 t CO
2eq∙t
−1 product, since the C
2-C
3 olefins shared the product weight ratio [[
42], [
43], [
44], [
45]]. Turquoise H
2 from CH
4 had a CFP value of 0.83-5.54 t CO
2eq∙t
−1 product [
5,
40] because the single-pass conversion of CH
4 was low and the recycled conversion consumed energy [
47,
48]. Interestingly, blue H
2 from CH
4 and coal power was not as clean as expected; its CFP value varied significantly between 0.92 and 19.26, assuming that 56%-90% of CO
2 was captured and stored by carbon capture, utilization, and storage (CCUS) technology [
6,
37,
39,
40]. Similarly, the use of grey electricity (#1) made the H
2 from water electrolysis not clean enough, and the CFP value varied from 7 to 39.52 t CO
2eq∙t
−1 product [
34,
39]. Only using green electricity (#2) allowed the CFP value of H
2 from water electrolysis to drastically decrease to 0.1. The CFP value of the H
2 from the proposed technology was 0.27-0.30, since the production of CNTs in a higher weight ratio shared the contribution of emissions. It should be noted that the present work was carried out with an LCO containing sulfur (6960 µg∙g
−1). The energy consumption of desulfurization was not considered in the CFP calculation, since it was easily offset by the much higher GI from the transformation of the desulfurized LCO. Different carbon products resulted in different CFP values for the direct decomposition of LCO and coal gasification, regardless of their similar H/C ratio. In addition, the present technology could exhibit a carbon-negative effect, considering its further use.
3.3. Practical feasibility evaluation regarding energy consumption and reactor scale up
The practical feasibility of the proposed technology was evaluated from the two perspectives of operating temperature and energy consumption (per mole of H
2). First, the operating temperatures of the hydrogen-production technologies were correlated (
Fig. 4(a)). In general, higher temperatures were associated with greater energy consumption and more CO
2 emissions, given the unavoidable energy loss during heat recovery and utilization. The proposed technology operates at lower temperatures (700-800 °C) and is easily realized, in comparison with MP, SMR, and coal gasification. Second, the Gibbs free energy (△
G) and enthalpies (△
H) of these technologies were calculated using HSC chemistry software (
Figs. 4(b) and
(c)). In fact, the LCO consisted of four lumping components: alkanes (19.2%), benzene homologues (18.4%), naphthalene compounds (58.3%), and phenanthrenes (4.1%). The △
G of each lumping component was negative when considering its separated transformation into carbon and H
2. However, the transformation of the first two lumping components was endothermic (△
H > 0), while the transformation of the latter two was exothermic (△
H < 0) (
Fig. 4(b)). The gross △
H of the LCO conversion was slightly negative (-4.8 to -8.5 kJ∙mol
−1 H
2), and the associated △
G was -98 to -107 kJ∙mol
−1 H
2, considering the weight ratios of the four lumping components in the GI range of 6-42 (
Fig. 4(c) and Table S5 in Appendix A). As shown in
Fig. 4(c), the proposed route was the only exothermic reaction among the compared routes. For the other endothermic H
2 production routes, the △
H sequences were as follows: water electrolysis (> 300 kJ∙mol
−1 H
2) ≫ PDH, EDH, and coal gasification (120-130 kJ∙mol
−1 H
2) > SMR (< 60 kJ∙mol
−1 H
2). The △
G values were negative but very small for SMR, MP, EDH, and coal gasification. The △
G values of water electrolysis and PDH were positive, indicating that these processes are non-spontaneous under the operating conditions. The spontaneous and exothermic decomposition of LCO can be attributed to the low H/C ratio of diesel, whose chemical equilibrium is not seriously influenced by H
2, in comparison with the MP process. In addition, the lower H
2 weight ratio of LCO is probably a factor in the high GI value of the proposed process, compared with the various hydrocarbons in other CVD processes (
Fig. 3(f)). This evaluation reveals the incredible feasibility and spontaneity of the proposed technology.
The key reactor apparatus will always be the bottleneck for the scale-up of clean H
2. Here, we compared the decomposition of LCO with water electrolysis, since they produce a similar weight ratio (9%-11%) of H
2 based on the feedstock. As shown in
Fig. 4(d) and Table S6 Appendix A, an alkaline water electrolyzer (AWE) for water electrolysis was compared with an FB reactor for LCO conversion. The space-time yield of the AWE was 22.65-31.36 m
3 H
2∙m
−3 (the volume of hydrogen can be produced per cubic meter of reactor volume) and its capacity was 500-1500 Nm
3∙h
−1 H
2, while the space-time yield of the FB was 29.49-40.44 m
3 H
2∙m
−3 and its capacity was 500-10 000 Nm
3∙h
−1 H
2. Based on the similar space-time yield of H
2, FB would be more attractive in large-scale preparation, since the absolute capacity of a single reactor could be 60-100 times greater than that of an AWE.
Fig. 4(d) summarizes the four aspects of energy consumption, single reactor capacity, space-time yield of H
2, and product cost, confirming the great potential of our technology. Taking China as an example, the yield of LCO exceeded 40 million tonnes per year, contributing to a total of 3.0 million tonnes per year H
2 if completely decomposed. Thus the H
2 capacity could reach nearly 9%-10% of China’s current H
2 gross production (30-35 million tonnes per year). Considering the current distribution of oil refinery factories and the easy transportability of diesel, jadeite H
2 could probably serve the local demand for hydrogenation in factories without costly storage and distribution or could serve for the construction of a portable supply network of clean H
2.
4. Conclusions
This research developed a new route to produce clean hydrogen from diesel with high selectivity. We also proposed the universal concept of the GI (the molar ratio of H2/CH4 eq, also equal to H2/CO2), which can be used to assess the greenness of any hydrogen-production technology. In this work, the optimal GI values of the process using a nanosized Fe catalyst were as high as 42, and the average GI value exceeded 25 within 12 h. A comparative study suggested that the H2 selectivity was influenced by the catalyst type, space velocity, and deactivation trend of the catalyst. The proposed technology has the highest GI value and the lowest CFP value, in comparison with industrial grey H2, blue H2, PDH, EDH, and even H2 from water pyrolysis with grey electricity. Energy-consumption (per mole of H2) and reactor scale-up evaluations validated our technology’s capability for easy scale up and competitive economic attraction.
CRediT authorship contribution statement
Bofan Li: Investigation, Writing - original draft, Data curation, Methodology, Formal analysis. Ruijing Jiao: Investigation, Methodology. Chaojie Cui: Funding acquisition, Conceptualization, Writing - review & editing. Xiang Yu: Methodology, Investigation. Jian Wang: Methodology, Investigation. Yunhai Ma: Methodology, Investigation. Weizhong Qian: Supervision, Conceptualization, Project administration, Writing - review & editing, Funding acquisition. Yong Jin: Supervision, Conceptualization.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
This work was financially supported by the National Natural Science Foundation of China (22109085 and 21975142) and the Tsinghua-Sinopec joint project (123068 and 20232930013).
Appendix A. Supplementary data
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.eng.2025.03.041.
PDF
(1651KB)
