Figure 11.2 outlines the main options available for production of H2 for use as an energy source. It should be noted that the hydrogen economy envisages combined aco2
H2 and electricity generation. Electricity generation from nuclear energy could also be included in the scheme, but this is not an option available everywhere. Clearly, the choice is quite varied. The most realistic short-term option in the transition to a sustainable hydrogen economy is catalytic production from H2 fossil fuels, although in principle this could be produced catalytically using any hydrogen-containing compound as a fuel. In general, for hydrocarbons or oxygenates, H2 can be produced by one of three chemical reactions, or combinations thereof, collectively known as reforming reactions: steam reforming (SR), dry (or CO2) reforming (DR), and partial oxidation (PO). These reactions can be represented as follows:
Steam reforming (SR)
CxHyOz + (x - z) H2O = x CO + (x-z + y/2) H2 (11.5)
CxHyOz + (2x - z) H2O = x CO2 + (2x-z + y/2) H2 (11.6)
Dry reforming (DR):
Partial oxidation (PO):
Direct catalytic decomposition is another option, but this is realistically a possibility only for methanol and will not be further discussed here. Partial oxidation can also proceed noncatalytically (often denoted POX, to distinguish it from PO). Equations 11.5 through 11.9 represent the idealized reactions, and other reactions may occur at the catalyst surface:
Water-gas-shift reaction (WGSR):
Reverse water-gas-shift reaction (rWGSR):
CxHyOz + (2x +y/2 - z) / 2 O2 = x CO2 + y/2 H2O (11.13)
Carbon formation reactions:
In reality, reforming reactions are mechanistically quite complicated and are composed of multiple steps and reactions. This makes it difficult to establish with certainty the exact mechanism. For example, while some reports claim a direct partial-oxidation mechanism (Equation 11.8), it has been established that, in the case of many Ni-based catalysts, the reaction takes places by complete combustion at the front of the catalytic bed (Equation 11.13), followed by steam and dry reforming (Equations 11.5 and 11.7) further along. In all cases, the mechanistic complexity of Equations 11.5 through 11.9 is indicated by the fact that the product stream produced invariably contains CO2 and H2O, along with unreacted fuel.
Before giving an overview of industrial-scale hydrogen production, a distinction should be made between the reforming reactions discussed above and industrial processes based on these reactions. SR, PO, and DR offer H2/CO mixtures of various compositions (3, 2, and 1, respectively, in the case of methane). Industrial processes often use multiple reformers and variations of conditions to tailor the final composition to the required use. Some processes, such as autothermal reforming (ATR) (discussed below) are combinations of the reforming reactions. Differences between processes may be defined on the basis of the composition of the feed or modality of reaction and may thus be considered as engineering distinctions. However, due to the previously discussed complexity of the surface processes, from the point of view of the basic chemistry/catalysis taking place, distinctions become less clear, often with the same reactions taking place under different conditions.
Hydrogen production is of enormous importance in the chemical and refining industries. In mixtures with nitrogen, H2 is used for ammonia synthesis. It is used in mixtures with CO for methanol and higher alcohol synthesis, for liquid hydrocarbons synthesis (Fischer Tropsch synthesis), and for synthesis of oxygenates (hydroformilation). It is also required for methanol-to-gasoline (MTG) processes. In fact, due to this importance, mixtures of CO and H2 are commonly known as synthesis gas or syngas. In refineries, H2 is used in a number of hydrotreating processes (hydrocracking, hydrotreating, hydroconversion, hydrodesulfurizing). The main source of H2 in an oil refinery is catalytic reforming of naphtha, with some contribution from H2 recovery from H2-rich off-gases. However, the ever-decreasing hydrogen content of crude reserves (especially U.S. reserves), combined with the growing demand for hydrogen in processes such as fuel reformulating and desulfurization (to combat air pollution), has given rise to the so-called refinery hydrogen balance problem, where demand is approaching or has in some cases surpassed supply. The shortfall is being met by plants that produce H2 by steam reforming of methane [29, 30].
Thus, H2 production is a mature industrial technology, and processes based on catalytic reforming have been commercialized. In all cases, natural gas (predominantly CH4) is the raw material. Two processes are mainly used to produce H2 from methane:
Steam reforming (SR)
Autothermal reforming (ATR)
Steam reforming of methane (SRM) is the most widely employed process and was commercialized as early as 1926:
Despite the fact that it is a highly endothermic and therefore an energy-intensive process, it is the preferred option for large-scale production due to the wide availability of natural gas and the fact that, unlike the other options, it does not require a costly O2 plant. The reaction produces the highest theoretical CO/H2 ratio, but the ratio can be tailored by changing the reaction conditions [30, 31]. Given its endothermicity, the reaction is favored at high temperature, and the increase in the total number of molecules means that it is also favored at low pressure.
A schematic illustration of a H2 production process based on reforming is shown in Figure 11.3. Steps commonly used in SRM processes are highlighted in bold.
Ni-based catalysts (Ni/NiO, ca. 15 wt%) are used for the reforming step(s), typically supported on ceramic materials such as a-Al2O3, MgO, MgMAlO, spinel, and Zr2O3 . A number of additives are included to enhance the overall performance characteristics of the basic formulation:
h2o co2+ h2o h2o co2+ h2o
Calcium aluminate (10 to 13 wt% CaO) as a binder to improve the mechanical strength of the material
Potassium oxide (up to 7 wt%) to inhibit coke formation
Silica (up to 7 wt%, as a silicate) to stabilize the potassium oxide
Such formulations have been developed to overcome the chief problem encountered under steam-reforming conditions: a strong tendency to deactivate through various mechanisms. The severe hydrothermal conditions normally employed strongly favor sintering and catalyst weakening (loss of mechanical strength), while supports based on Al2O3 or MgO form spinels with Ni, thus removing the active Ni phase. Calcium aluminate and the high Ni loadings can alleviate these problems. Deactivation through coke formation is an insurmountable problem. In addition to the potassium oxide additive, high steam/CH4 ratios are used to minimize coke deposits. However, high steam/CH4 ratios put a further strain on the mechanical integrity of the catalyst.
Autothermal reforming, which has also achieved widespread application, is a combination of partial oxidation and steam reforming. The process was originally developed by Haldor Topsoe in the late 1950s [30, 32] as a means of increasing the H2 content for ammonia plants. The process uses the exothermic partial oxidation reaction to supply heat to the steam-reforming process and involves cofeeding natural gas, steam, and oxygen. Original configurations were based on the idea of consecutive reaction zones. In the first, the thermal or combustion zone, noncatalytic partial oxidation took place. The reaction was ignited externally with a burner:
CH4 + 1.5 O2 = CO + 2 H2O; AH° = -35.7 kJmol-1 (11.19)
This was followed by a catalytic zone, in which steam reforming and the WGSR took place:
This technology has also seen considerable improvements. In later developments, the two segments were combined into a single unit [33-36]. Not surprisingly, given the similarity of the chemistry involved, Ni-based catalysts are also used in ATR processes.
Industrial-scale catalytic partial oxidation has not yet been developed; however, as it represents an attractive alternative to SR and ATR, it has been the subject of intensive academic and industrial research. The advantages include the exothermicity of the reaction, its greater selectivity to syngas, the requirement for high space velocity, which offers the possibility of more compact design, and the fact that it offers another option of syngas composition for industrial use. As mentioned earlier, there have been claims for direct partial oxidation over some catalysts. However, in most cases the mechanism involves total combustion of part of the methane at the front of the catalyst, followed by steam and CO2 reforming further along the bed, the partial oxidation reaction being a linear combination of these three reactions [37-41]. From this point of view, it has some similarities to ATR.
CO2 reforming supplies syngas in the theoretical H2/CO ratio of 1 and has been commercialized, for example, in the Calcor process [42, 43]. It is most widely used in secondary reforming processes to reduce the H2/CO ratio obtained from steam reforming. Finally, a number of other processes have been described, including combined reforming (CR), gas-heated reforming (GHR), and combined autothermal reforming (CAR) . These are various combinations of SR and ATR.
The WGSR serves two important purposes: the simultaneous reduction of the CO content of the primary reformate (up to 12% for steam reforming and 6 to 8% for ATR) and increase of the hydrogen content. A number of extensive reviews of the reaction have been published [45-49].
As a slightly exothermic process, the WGSR is favored at low temperature. The temperature at which activity is observed is therefore determined by the kinetics, which are slow, and the dew point . In addition, the exothermicity is an important process consideration: the higher the shift conversion, the higher the bed temperature generated. This may result in catalyst sintering. In industrial applications, a compromise between high CO conversion and high temperature is achieved by dividing the water-gas-shift step into two adiabatic stages using separate high-temperature shift (HTS) and low-temperature shift (LTS) catalysts. Fe-based catalysts are used for HTS and Cu-Zn for LTS. On exiting from the reformer section, the gases are hot, which can be exploited for HT-WGSR. Further downstream, the again cooler gases can be treated in a second LTS unit .
The final stage shown in Figure 11.3, selective CO removal, is usually accomplished by selective CO methanation or pressure-swing adsorption. As will be discussed in more detail below, preferential CO oxidation (PrOx) is a more suitable option for automotive application.
Of the H2 production methods shown in Figure 11.2, water electrolysis is also an established technique for H2 production at an industrial level. However, because of its high cost, electrolysis is used only in specialized processes, where extremely high purity is required, and currently it represents <5% of total industrial production . Of the other techniques, pyrolysis of coal is the closest to wide-scale industrial application.
11.2.4 Onboard Hydrogen Production and Purification 22.214.171.124 Requirements for Onboard Production
The process of conversion of a carbon-based fuel to H2 for use in fuel cells is commonly called fuel processing. One option to supply H2 for automotive applications is that of onboard reforming, in which the H2 is obtained from a hydrocarbon or oxygenate fuel by reforming reactions onboard the vehicle. Because of the absence of a H2 infrastructure, this was initially considered to be the most accessible option, and much early work was devoted to this approach, with a view to utilizing the existing gasoline network or another liquid fuel that could be adapted (usually methanol).
In principle, the industrial H2 production methods discussed above can be applied to onboard reforming, with suitable modification. Thus, three options are described in most reports on onboard reforming: steam reforming (SR), autothermal reforming (ATR), and partial oxidation (PO).
It should also be noted that some reports, perhaps more correctly and more or less explicitly, divide the options into two: steam reforming and partial oxidation. Here, a note of caution should be sounded on the definitions involved in onboard reforming processes. As outlined by Ahmed, each process can be defined by the manner in which the reactants are introduced to the catalyst . Thus, in steam reforming, a mixture of fuel and H2O is fed to the reforming catalyst, while fuel and oxygen are used in partial oxidation (with H2O added later for the WGSR), and in autothermal reforming fuel, H2O and O2 are co-fed. However, strictly speaking, the autothermal reforming process should only be considered so if the feed composition is such that the overall reaction is heat balanced, even though, in practice, an excess of air is often used (to compensate for heat losses and to obtain a reformer with a rapid response):
CxHyOz + a H2O + b O2 = c H2 + d CO2; AH° = 0 (11.22)
On the other hand, many applications use various amounts of H2O and O2 in the feed. As pointed out by Pettersson and Westerholm, such processes are better described as "combinatorial reforming" . As discussed previously, the surface reactions are quite complex and mechanistic analysis is often not conducted. For these reasons, it is also common for reports to describe these systems as being based on one type of reforming reaction or another.
Given the established history of industrial reforming, it is not surprising that research into H2 production for small-scale or mobile applications has taken an example from this knowledge. The concept of onboard H2 production follows a strategy similar to that outlined in Figure 11.3, with the same basic aim: to produce H2 with minimal CO content. As will be discussed below, the latter is particularly important, as CO is a poison for the polymer electrolyte membrane fuel cell (PEMFC) located downstream. If maximum H2 content were the only consideration, then steam reforming would be the only option. However, onboard reformers must meet a different set of criteria than their industrial counterparts. These criteria include :
An H2 production capacity that is significantly lower than those in stationary plants
Significantly smaller size and weight
The ability to undergo multiple start-up/shutdown cycles
Variable and rapid response
High reliability and durability (although lower than stationary reformers)
In addition, current industrial technology suffers from a number of drawbacks that must be overcome if reformers are to be candidates for onboard H2 production. These include :
The pyrophoric nature of Ni-based catalysts used for reforming and Cu- and Fe-based catalysts used for WGSR. This represents a potential risk (sintering and fire hazard) for onboard application.
The unsuitability of hydrodesulfurization and pressure-swing adsorption (PSA) for onboard application because they require high pressure. The absence of sulfur is necessary not only because Ni catalysts are extremely sulfur sensitive, but also because sulfur is a poison for PEMFCs (discussed below).
The unsuitability of the other method of CO removal commonly used. For methanation of CO, CO2 must first be removed, as it can be also methanated consuming H2. The inadaptability of both methods makes preferential oxidation (PrOx) the most viable option for this important step, although other options do exist (discussed below).
The endothermic nature of steam reforming, which requires complicated engineering for heat management.
The fact that industrial steam reforming is optimized for steady-state operation.
When all factors are taken into account, ATR/PO have considerable advantages over SR, although the hydrogen content of the primary reformate is lower. These processes can be achieved with more compact systems, and they are less energy intensive and have a more rapid response, thus making them much more suitable to the multiple start-up/shutdown operations expected in a mobile source. Despite this, it should be noted that steam reforming has also been extensively investigated, and indeed demonstrated, for onboard application . The main thrust of current research is to meet the criteria by overcoming the limitations of current industrial methods.
Precious metals (Rh, Pt, Ru, Ir) supported on oxides have been shown to be highly active for reforming reactions [37, 39-41, 53]. They also generally show a higher tolerance to sulfur than Ni . Their main drawback is their prohibitive cost, which has thus far precluded their industrial use. However, as these metals are intrinsically more active than Ni, their loading can be lowered by more than an order of magnitude. This high activity means that these metals have received attention as possible replacements for Ni.
Another significant field of research has been to find ways to overcome the limitations of Ni and the other first-row transition metals Fe through Zn. The effect of the support on activity has been investigated. With nonnoble metal catalysts (Ni, Co, Fe), supports with a low concentration of Lewis acid sites or the presence of basic sites (ZrO2, MgO, La2O3) have been found to promote activity and offer a higher resistance to coking [54-56]. ZrO2 and CeO2-ZrO2 mixed oxides have been shown to be particularly effective supports [57-59]. This approach is also used in the case of precious metal catalysts. For example, the higher activity of Pd supported on CeO2 with respect to SiO2 or Al2O3 was attributed to the ability of ceria to supply lattice oxygen to the steam-reforming reaction [60, 61]. Doped ceria supports (e.g., Ce08Sm0.15Gd005O2) have also been shown to be effective .
Due to the high thermal stability of hexa-aluminates , their use as supports has been investigated, especially for the partial oxidation reaction, which can give rise to very high bed temperatures [64-66].
The use of the support to promote reaction is often used in conjunction with novel preparation methods. Controlled synthesis of highly active metallic nanopar-ticles is an area of strong topicality in heterogeneous catalysis, and the synthesis of reforming catalysts is no exception. In the case of Ni, it has been demonstrated that coking is less of a problem with very small particles [67, 68]. In the case of precious metals, very active particles allow a decrease of the loading and therefore of cost. If sintering of these nanoparticles could be prevented, then such precious metal catalysts could become viable alternatives to Ni-based catalysts. One approach that has been successfully applied to achieve better and more-reproducible metal dispersion with respect to traditional synthesis methods is incorporation of the active phase within precursors such as perovskite oxides or hydrotalcite-type clays [42, 69-75]. As the active phase is present homogeneously within the precursor, very finely dispersed and highly active materials are formed upon heating the starting material to high temperature.
Another area that is being actively investigated is that of bimetallic catalyst, for which there is a considerable scientific literature. A common approach has been to promote the activity of Ni with small amounts of a precious metal.
Catalysts for so-called oxidative methanol reforming have also been widely investigated. Great attention has been devoted to the well-established commercial Cu/ZnO/Al2O3 systems [76-79]. Inclusion of zirconia in the formulation has been found to be beneficial for the methanol reforming reaction [80-82]. Noble, transition, and base metals (using various supports) have also been considered as the active phase [83-87]. More recently, there has been a growing interest in the reforming of ethanol to produce hydrogen [88-94]. This might be due to the fact that it is becoming generally accepted that ethanol, as a bioproduct of fermentation, can be one of the large-scale renewable hydrogen sources. It should be noted, however, that, with respect to the reforming of methanol, the presence of the C-C bond in the ethanol might favor coke formation on the surface of the catalyst and, therefore, catalyst deactivation. Indications that Co-based catalysts show good activity have made them a subject of particular attention for ethanol reforming [95, 96].
126.96.36.199 High-Temperature and Low-Temperature WGSR Catalysts
As an industrial process, the WGSR is efficiently achieved in two steps. Experimental and early-generation fuel processors use this type of arrangement with interstage cooling and industrial catalyst formulations. However, addition of an extra step increases the overall size, weight, and cost of the system. Thus, a single step (and nonpyrophoric) WGSR catalyst would be more desirable, although it should be noted that when methanol is used as a fuel, only one WGSR step is necessary. As this catalyst should ideally reduce the CO concentration to 1% or less, it must show high activity at a relatively low temperature (<600 K). In addition, it must satisfy the requirements common to all catalysts used in hydrogen production: high stability and durability, mechanical integrity and resistance to shock or temperature excursions, stability toward poisons such as H2S and chlorine, and the absence of side reactions (particularly methanation activity). WGS catalysis is an area in which a breakthrough is necessary, and research into more suitable shift catalysts has mainly considered two categories of material: base metal catalysts and precious metal catalysts or gold catalysts [34, 35].
Molybdenum carbide catalysts have been reported to show higher WGSR activity than commercial Cu-Zn-Al LTS catalysts. In addition, the conditions employed were suitable for mobile applications (493 to 568 K and atmospheric pressure); there was no apparent deactivation or modification of the structure during 48 h on-stream; and there was negligible methanation activity .
A base metal catalyst with activity comparable with commercial LTS catalyst (and a wider operating temperature) has also been reported. The authors stress the nonpyrophoricity and the ability to withstand situations common in fuel processing, such as water condensation during start-up and shutdown [98, 99].
Cu- and Ni-doped (La)CeO2 have been reported to show high WGS activity and stability at temperatures up to 900 K. A co-operative redox reaction mechanism was reported, involving oxidation of CO adsorbed on the metal cluster by oxygen supplied to the metal interface by ceria . On the other hand, high-temperature stability of Cu was reported to be a problem, and even the most promising material investigated promoted CO methanation .
In the case of precious metals, ceria or promoted ceria is usually used as the support, and Pt is the most widely studied metal. The WGS activity of Pt/CeO2 was recognized in the early 1980s, in connection with the use of CeO2 as an additive in TWC formulations , and ceria-supported precious metal catalysts are considered potential candidates for fuel processing [34, 35, 103-106]. Strategies aimed at improving these materials include the use of promoters , nanostructuring of the support , and alloying of the metal . Although the suitability of PGM catalysts has been questioned , Johnson Matthey (JM) has recently developed "non-pyrophoric PGM formulations" with suitable durability and no methanation activity across a wide temperature range (500 to 800 K) that can be incorporated into its own fuel processor in one WGS stage .
Following the report of the extraordinary low-temperature CO oxidation activity of gold-based catalysts , there has been an enormous interest in the use of supported gold as potential WGS catalysts. Au/CeO2 [103, 109-113], Au/TiO2 [114-118], and Au/Fe2O3 [113, 117, 119-123] have all been investigated. However, a general criticism of these catalysts is that reproducible catalyst synthesis has been a problem, with many papers stressing the sensitivity to synthesis. In addition, thermal stability is an issue: gold catalysts sinter too easily. For these reasons, gold-based catalysts are widely believed to be more suitable as PrOx candidates (discussed below), for which the lower operating temperatures allow them to maintain their sometimes remarkable activity.
188.8.131.52 Preferential Oxidation of CO (PrOx)
After the WGSR step(s), the CO content of the reformate is on the order of 1%. CO must be reduced, as it is a poison for the electrodes of PEMFCs. For onboard application, there are a number of approaches that can, in principle, be considered to lower the final CO content of the stream to levels compatible for use in a fuel cell. These include preferential oxidation of CO (PrOx), CO methanation, Pd-membrane separation, and the development of electrodes with greater CO tolerance . PrOx (also called selective oxidation) involves preferentially or selectively oxidizing the CO in the presence of the large excess of H2, and is the most studied method, as it is a relatively advanced technology. The oxygen needed to oxidize the CO is added after the shift step(s). For onboard application, the catalyst should ideally operate between the temperature of the shift reactor and the operating temperature of the PEMFC, usually between ca. 473 and 353 K.
Commercial PrOx catalysts were first introduced in the 1960s, with the trade name Selectoxo™, to remove CO from H2 in ammonia synthesis processes. The catalyst consists of Pt supported on alumina, promoted by a base metal. The platinum content is usually rather high, on the order of 5 wt%. These commercial catalysts are indeed capable of reducing CO levels to <10 ppm, but only under very carefully controlled conditions. Various parameters — selectivity, temperature control, space velocity — are of fundamental importance to the overall efficiency. The variations in conditions encountered with onboard applications are often incompatible with the ideal PrOx conditions. For example, a problem of existing Pt-based catalysts is that low space velocity favors CO production by the reverse WGSR. If a fuel reformer must operate with variable loads, this could be a problem.
Selectivity, which can be gauged from the O2/CO ratio necessary to observe adequate removal of CO, is another important issue. If the catalyst is not 100% selective toward CO oxidation, which implies simultaneous conversion of H2, an excess of air must be added. This results not only in a loss of H2, but also in an overall increase of the heat generated through two highly exothermic reactions. This can have knock-on effects, as increased reaction temperature can lead to promotion of the reverse WGSR and therefore an increase in the outlet CO concentration.
Such problems can be overcome by: two-stage PrOx with interstage cooling; low inlet temperature to second stage; minimum air injection; fixed-flow operation; and second-stage preferential oxidation with catalyst immune to rWGSR. However, this would make an already complicated fuel reformer system even more complicated [35, 36].
Following the report by Haruta et al., the general properties of catalysts based on Au nanoparticles have become widely investigated . Because they show higher activity toward CO oxidation than H2 oxidation, they inevitably have attracted great interest as potential PrOx candidates [125-134]. Au-based catalysts show very high selectivity toward CO oxidation, even in the presence an excess of H2 typical of reformate streams. However, deactivation has been reported to be a problem when, more realistically, CO2 and H2O are added to the feed . This is a problem that must be overcome if such catalysts are to become viable alternatives to Pt-based catalysts.
Finally, Cu-based materials have recently emerged as promising candidates, in particular those containing ceria [135, 136].
In recent years, incorporation of membrane technology into hydrogen production processes has become an area of interest. To date, separation membranes, usually based on Pd-alloys, have been used most extensively . Separation membranes can be incorporated into various steps of the H2 production process. At the reforming step, oxygen-selective membranes can be used to deliver highly active oxygen to the catalyst and eliminate the need for gas-phase oxygen. At the WGS stage, hydrogen-selective membranes can be used to remove hydrogen from the stream and therefore overcome thermodynamic limitations. There have also been reports of CO or CO2-selective membranes, which can be used to purify hydrogen and eliminate the need for PrOx . Theses technologies are as yet at a preliminary stage, but they do offer interesting possibilities for the whole hydrogen production process.
Research has shown that methanol, ethanol, compressed natural gas, methane, biogas, naphtha, gasoline, diesel, kerosene, aviation fuels, marine fuels, hydrocarbons (pentane, hexane, octane, acetylene), propane, butane, dimethyl ether, and ammonia can all be used as feedstocks for onboard fuel reformers . Thus, there is a high degree of flexibility from the point of view of fuels. Of these, methanol and gasoline are the favored options. Technically, methanol is the easiest option, and indeed methanol appears to be the most commonly investigated liquid fuel for onboard applications . A liquid fuel also considerably simplifies handling problems. However, a methanol distribution network does not exist. A network for gasoline, on the other hand, does exist, but extremely high reforming ability is necessary for gasoline, especially in relation to coking of the catalyst. From the point of view of overall efficiency and CO2 emissions, there is very little difference between methanol and gasoline if the synthesis of methanol is included in the comparison. The best approach would most likely be the development of fuel-flexible reformers, but in that case the reformer is unlikely to be optimized for any one fuel [29, 51, 137].
To meet the particular requirements of onboard hydrogen generation, highly complex, integrated systems are required. Despite this, many reformers have been produced commercially . However, it should be noted that onboard reforming is now widely viewed as a first step toward the ultimate solution of onboard hydrogen storage.
Was this article helpful?
Quick start guide to skyrocket your offline and online business success with mobile apps. If you know anything about mobile devices, you’ve probably heard that famous phrase coined by one of the mobile device’s most prolific creators proclaiming that there’s an app for pretty much everything.