Hydrogen is perhaps one of the most promising energy carriers of the future. In renewable energy systems with high fractions of intermittent supply (e.g. wind power and solar thermal energy), potential surplus electricity could be converted into hydrogen through water electrolysis. This hydrogen can be used in a wide variety of applications. The most often discussed option, the reconversion of hydrogen into electricity, be it by gas turbines or by fuel cells, appears to be rather unattractive, due to the low round-trip efficiencies. Electrolysis – based on the process scale – can be estimated to have an efficiency of about 60% (if higher efficiencies are given, they are typically relative to the cell level). A recent NREL analysis, based on questionnaires given to manufacturers, indicate a mean efficiency value of 53% for the system.¹  Considering that the fuel-cell efficiency on the systems’ level and gas turbines (not available for hydrogen yet) is estimated at about 50–60%, the overall round-trip efficiency is thus reduced to slightly above 30%. It will certainly be possible to improve this figure to some extent, but substantial losses in the round trip from electricity to electricity will invariably always be present. Therefore, the use of “renewable” hydrogen – not for the reconversion into electricity, but rather as a feedstock for the chemical industry, in oil refineries, or in biorefineries – appears to be promising. For biomass upgrading, a substantial need for hydrogen undoubtedly exists due to the high oxygen content present in biogenic molecules.

Figure 1.1 plots the chemical composition of different energy carriers in an O/C vs. H/C diagram.²  The typical biomass constituents contain much more oxygen than potential target molecules. In addition, the hydrogen content often needs to be increased. Decarboxylation and decarbonylation pathways are one of the possibilities for the reduction of the O/C ratio, but they alone are insufficient for this purpose. Accordingly, for further oxygen removal, hydrogen is often required as a reducing agent in order to convert biogenic molecules into less-oxygenated target compounds. In order to increase the H/C ratio, hydrogen is needed directly either in the hydrogenation and hydrogenolysis pathways, or indirectly after dehydration, since the dehydration leads to unsaturated compounds that are often undesired intermediates, as they are very reactive, and thus may undergo side reactions, decreasing overall product yields. The various process options that hydrogen is used for in order to convert biomass into chemical intermediates and end products will be briefly discussed at the end of this chapter.

The vast majority of hydrogen is currently produced from fossil fuels (estimated at 49% from natural gas, 29% from liquid hydrocarbons, either directly from naphtha or related feedstocks, or indirectly by converting residues in refineries or as off-gases from chemical or refinery processes, 18% from coal, and 4% from electrolysis).³  Most of today's production is intended for further processing in the chemical and refinery industries, and is thus not traded on the market. It is estimated that only ca. 10% of the produced hydrogen is traded (i.e. merchant hydrogen); the rest is produced and directly used onsite (i.e. captive hydrogen).⁴  Due to this fact, there are various figures available, as the production levels are difficult to assess. From the estimates published for different years and projected growth, current global production is about 60 million metric tons per year, with wide margins of error.

For the purposes of this chapter, we will consider processes rendering hydrogen as the main product (i.e. hydrogen made on purpose). Thus, typical refinery processes (e.g. coking and visbreaking) are not further discussed. Also disregarded are petrochemical processes, such as steam cracking for lower olefins production, since here the olefin is the main product, although the hydrogen produced contributes to the overall profitability of the process. Some processes can be considered as borderline cases, such as cracking, in which the amount of produced hydrogen can be adjusted by the processes conditions, and can thus be tuned to the hydrogen requirements of the refineries.

There are three main processes for the production of hydrogen from carbon-containing feedstocks: catalytic steam reforming (SR), autothermal reforming (AR) and partial oxidation (PO), as well as other configurations, which contain various aspects of any of the aforementioned processes. The selection of the reforming technology depends on many factors, such as the intended use of the hydrogen, acceptable impurity level, pressure level of downstream processes, price and availability of hydrocarbon and fuel, investment and operational costs, catalyst price, and several others secondary factors.⁵  Overall, the reforming technologies are intimately connected to the chemical reaction networks that govern hydrogen formation, making them perhaps best to be discussed altogether.

The basic reaction of steam reforming is given as eqn (1.1) for the example of methane,

however, any other carbon-containing feedstock can be converted by a similar process, although this may require changes in process technology. Moreover, the syngas composition is dependent on the elemental composition (i.e. H/C/O ratios) of the feedstock.

Partial oxidation can be described by eqn (1.2), again formulated for methane as feed:

However, it is difficult to achieve full selectivity to syngas as expressed in eqn (1.2), since partial oxidation is always competing against total oxidation:

Moreover, several other equilibria are relevant in all of these systems, such as:

In addition, methane can be cracked into hydrogen and carbon; for higher hydrocarbons, cracking reactions also come into play, and heteroatoms, which are almost invariably present in the feedstocks, react as well under the conditions of the hydrogen-generating reactions. While all of these reactions can occur in all of the systems, it is helpful to conceptually separate them, and discuss the three processes as prototypes. SR is described by eqn (1.1) and PO by eqn (1.2). As the overall reaction network contains exothermic and endothermic reactions, the feed can be composed in such a manner that the resulting enthalpy becomes nearly zero. This option is called autothermal reforming (AR).

In SR, the hydrocarbon, methane for the majority of the produced hydrogen, and steam react over a nickel/γ-Al₂O₃ catalyst at 1073–1173 K under a pressure of 1.5 to 3 MPa in a tubular reactor, producing syngas. One of the key problems is the supply of the heat to the reactor, since the reaction is highly endothermic (206 kJ mol⁻¹). Typically, the heat is integrated into the system via a firebox, in which part of the feed gas is combusted. Methane and other fuel gases can be used in the firebox. The product gas after the steam-reforming reactor has the equilibrium composition at the exit temperature of the reformer, containing H₂, CO, CO₂, and CH₄ in the case of natural gas and naphtha as feedstock;⁶  in turn, with oil or coal, nitrogen and sulfur compounds can be present, as well. In order to maximize hydrogen yields, the product gas is first exposed to a high temperature shift reaction (1.4) over iron-based catalysts at 623–723 K. Depending on the plant configuration, the final hydrogen product is obtained after purification via pressure swing adsorption (standard in modern plants)⁷  or by a low-temperature shift stage at 493 K over Cu/ZnO/Al₂O₃ catalysts, followed by CO₂ scrubbing, with a final removal of CO_x by catalytic methanation. Instead of providing heat for the reforming reaction by an external heating of the reformer tubes, the required energy can be supplied to the process by partial combustion of the feed gas (autothermal reforming). In this case, the feed is mixed with oxygen before it enters the catalyst bed. In sequence, the gas is combusted in the entrance section of the bed, thus providing the heat for the reforming reaction in the later section of the unit. Such designs are advantageous if high temperatures are desired in order to minimize methane concentration, to allow high pressures, or to directly provide nitrogen for ammonia synthesis (in which air is thus used instead of oxygen).⁸ 

Conventional partial oxidation processes for hydrogen production are used for heavy feeds (e.g. heavy residues or oil fractions with high sulfur or metal contents). The PO reaction takes place in a flame in an empty furnace with substoichiometric amounts of oxygen. For the production of hydrogen, CO needs to be shifted to hydrogen, which is advantageously carried out in the presence of a catalyst at temperatures below 773 K after the steam addition to the feed. The purification of hydrogen proceeds by one of the many available scrubbing processes. Partial oxidation has also been reported to be suitable for syngas production from methane with very high selectivity and yields.⁹  At residence times of milliseconds, methane can be partially oxidized with oxygen rendering ca. 90% yields of CO and H₂. Key for this development is the very short contact time with noble-metal catalysts supported on ceramic foam monoliths. Despite the promising results, this technology does not seem to have been commercially implemented, probably due to the potential hazards of the gas mixture that is in the middle of the explosion regime.

Cost analysis for hydrogen production is not a simple task, and data provided in the literature have to be interpreted with caution. Should just production costs be analyzed, centralized facilities have clear advantages over distributed production, as the former can provide significant economies of scale. However, this advantage can be easily lost upon including the cost for delivery (i.e. liquefaction/pressurization and transport).¹⁰  An analysis for specific boundary conditions gave a cost for the dispensed hydrogen of 3.3 $/kg H₂ for a distributed production. Costs were only half of that in a centralized plant, but liquefaction and delivery cost of 3.5 $/kg H₂ have to be added to the pure production costs. However, in other studies, the additional cost for transportation has been estimated at only 1.02 $/kg H₂.¹¹  Transportation costs are obviously dependent on the mode of hydrogen transportation (pressurized, liquefied, truck, pipeline, etc.). In the following, if no other information is given, the hydrogen cost is referred to as the production costs alone. The studies quoted are also from different years, and thus for a direct comparison, inflation would need to be taken into account. However, most of the studies quoted are from the last five to ten years, and thus the error in neglecting inflation is minor when compared to strongly fluctuating prices of raw materials. Finally, the differences in exchange rates will also affect the figures. For the purpose of this contribution, the values are given in the currency that had been used in the original publication; when currencies are converted, an exchange rate of 1.38 $/€ is utilized.

Due to the broad range of feedstocks that can be used for production of hydrogen and the different process options available, the production costs greatly depend on local conditions. A detailed analysis from 1983 is available for hydrogen production in Germany.¹²  The situation, especially with respect to feedstock prices, has changed significantly since then, but some of the features, especially the split between feedstock and investment costs, are probably still valid. On the one hand, in the case of methane as a feedstock, investment costs for the reformer are the lowest, while feed/fuel costs are intermediate. On the other hand, a reformer operating on lignite shows the highest investment costs; however, due to the low price of lignite, the lowest feed cost. Most unfavorable are naphtha- and fuel-oil-based reformers. In the case of fuel oil, the feed cost is much higher than the investment costs so that the overall cost is one of the highest (Figure 1.2). Altogether, the pathway starting with natural gas is currently the most favored, as it is associated with the lowest overall costs.¹⁰ 

In a centralized methane reforming, the natural gas price dominates the cost for hydrogen production. Natural gas prices are influenced by regional economic conjunctures, since it is traded globally only to a limited extent, despite the availability of liquefied natural gas terminals in addition to the steady expansion of their capacity over the last decades. Considering the natural gas price at 5 $/GJ, the cost for hydrogen has been estimated at approximately 1 $/kg H₂ gate price for a large-scale plant with a production capacity of 427 tons per day.¹³  In the United States, due to the shale gas boom, natural gas prices are currently below 5 $/GJ (4.3 $/MMBTU,¹⁴  corresponding to about 4.1 $/GJ). Accordingly, the cost for hydrogen should be ca. 0.9 $/kg H₂ or 0.65 €/kg. In Europe, natural gas prices are about 22.5 €/MWh (i.e. corresponding to approximately 8.5 $/GJ) at the European Energy Exchange (March 24th 2014). The EU feed cost leads to a hydrogen production cost of about 1.2 €/kg H₂ (calculation based on the figures given in Ref. 13). Nonetheless, with the increasing supply of natural gas traded globally, it can be expected that also the European price of natural gas will eventually fall within the next few years, and thus hydrogen production costs at around 1 €/kg H₂ may also be estimated as a base-case scenario, against which other hydrogen production technologies will have to compete. In addition to the raw materials costs, it is worth mentioning that such estimates are dependent on other boundary conditions, e.g. the process scale (hydrogen production is at substantially higher cost on small-scale plants), price of possible coproducts (e.g. oxygen in electrolyzers), to mention just a few.

Importantly, hydrogen production from natural gas is very CO₂ intensive. This problem is even more serious for hydrogen production from feedstocks rich in carbon (e.g. lignite). Therefore, the price of CO₂ emission certificates – by either traded emission certificates or taxes on emissions – comes at the costs to other parts of the world. Depending on plant size and specific conditions, CO₂ equivalent emissions for hydrogen produced from natural gas are given as 11.88 kg CO₂/kg H₂.¹⁵  Considering the current European market prices for the CO₂ emission certificates (ca. 5 €/t CO₂), there is no considerable change in the economics of the process yet. However, when the European Emission Trading System was implemented, target prices were at 40 €/t (this amount had to be paid during the first trading period between 2005 and 2007, if an emitter did not have the corresponding certificate), this price corresponds to additional costs of ca. 0.5 €/kg H₂, if the full certificate price would have to be paid.

The cost situation is not much different for hydrogen production by coal gasification in centralized plants.^10,16  In 2008, an analysis of different studies revealed that the cost of hydrogen from methane reforming and coal gasification differed by less than 10% – again, this is strongly dependent on the regional context, since prices for gas and coal differ worldwide. However, coal would have a strong disadvantage, if CO₂ emission certificate prices would rise substantially in the future. Coal gasification has about twice the CO₂ footprint (i.e. ca. 25 kg CO₂/kg H₂) compared to centralized methane reforming. Accordingly, for the coal-based hydrogen, a CO₂ emission certificate at a cost of 40 €/t would lead to a cost increase of more than 1 €/kg H₂.

Biomass can be used for the production of hydrogen via reforming reactions. As for the conversion of fossil carbon-containing materials, there are also different general methods for the production of hydrogen from biomass.¹⁷  For instance, biomass can directly be gasified by partial oxidation at high temperatures of 1073 to 1173 K. This process resembles the gasification of solid fossil feedstocks. Gasification is possible with oxygen, air, or steam, with differences in the process layout and the heating value of the produced gas. In addition to hydrogen, the product gas contains CO, CO₂, water, and hydrocarbons. Water and carbon monoxide can be converted into hydrogen and CO₂ by the water-gas-shift (WGS) reaction (1.4).

Alternatively, plant biomass can be pyrolyzed in the absence of air. The standard process operates at around 773 K, and results in the formation of gases, pyrolysis oil, and solid char residue (as presented in Chapter 7). In turn, the pyrolysis oil can be converted into hydrogen by steam reforming, in a similar manner to that for the reforming of liquid fossil feedstocks. The major advantage of first producing a pyrolysis oil and then reforming it, as compared to direct gasification, is the marked reduction in volume provided by decentralized, small pyrolysis plants.¹⁸  Compared to plant biomass, pyrolysis oil is more efficiently transported to a centralized, large facility for the gasification. In this case, such plants are characterized by substantial economies of scale.

The pyrolysis can be carried out at higher temperatures of 973–1073 K in the presence of a catalyst, directly rendering a hydrogen-rich gas. Such biomass utilization schemes typically target liquid transportation fuels, which can be produced from syngas. When hydrogen production is the target, the advantage of the relatively easier transportation of pyrolysis oil might be lost, since the transportation of hydrogen is costly, and thus hydrogen production facilities are advantageous when strategically located directly at the point of use.

The conversion of biomass into hydrogen, be it by direct gasification or via pyrolysis schemes, seems to be the most cost-competitive pathway of all hydrogen-production routes from renewable energy. As in hydrogen production of natural gas and other fossil resources, the cost depends on many factors, such as local availability of biomass, transportation cost, exact plant layout, nature of the biomass, to mention just a few. Thus, there is a substantial spread of cost estimates for the production of hydrogen from biomass.

In a comparative study of different hydrogen-production technologies, biomass-based routes were identified as the only ones that are able to economically compete against natural gas reforming and coal gasification under the current market conditions.¹⁰  The same conclusion was reached in a recent publication, where again natural gas reforming, coal gasification and biomass gasification were identified as the most cost competitive methods at a level of approximately 1.50 $/kg H₂, based on an average of a number of reported figures.¹⁶ 

For two selected sources of biomass, waste forest residues and straw, the costs were analyzed for a 2000 t_{dry biomass}/d biomass plant (resulting in a hydrogen production capacity of approximately 170 t/d). Depending on the exact feedstock and the design of the plant (atmospheric or pressurized operation), the costs were estimated at 1.17 $/kg H₂ and 1.33 $/kg H₂.¹⁹  Again, these figures are in the same range as those found for the aforementioned reports. This publication also provides a sensitivity analysis, i.e. the costs are broken down into different components in a transparent manner. Interestingly, also the transportation costs, which are substantial for biomass, are carefully discussed. This information is useful in order to assess the optimum plant size, where the cost benefit due to the economy of scale is offset by the additional transportation distance – and thus cost – for a larger plant. Although the optimum plant size depends again on feedstock and gasifier technology, the optimum volume production is estimated at about 5000 t_{dry biomass}/d. Nonetheless, it is important to point out that the curve becomes rather flat beyond 3000 t_{dry biomass}/d.²⁴ 

As will be discussed later in this chapter, many of the routes for transforming biomass into fuels and chemicals require hydrogen as a reducing agent. Since the infrastructure for transportation and handling of biomass is already well established in a biorefinery, synergies can be expected between the biomass exploitation to give the starting materials for fuel and chemical production, and biomass as a feedstock for hydrogen production. On a biorefinery site, gasification of biomass could thus become economically the most attractive method for hydrogen production.

Currently, a small fraction (around 4%)¹⁷  of the world demand of hydrogen is supplied by water electrolysis. However, under most circumstances, electrolysis for the production of hydrogen is economically not optimal. It is often argued that this situation will change in systems with a high fraction of intermittent renewable electricity, such as from wind power or solar energy, since the price for electricity goes down substantially at times of high availability. However, one has to keep in mind that the capital expenditure of electrolyzers is substantial. Consequently, electrolysis plants running only during the (currently) relatively short periods of low electricity prices are not economical yet. Our analyses will be discussed further below, after the technologies available for hydrogen production by electrolysis have been introduced.

There are two technologies, alkaline electrolyzers and PEM (polymer electrolyte membrane or proton exchange membrane) electrolyzers, which are commercially available, albeit on different scales. High-temperature electrolysis is a third option. Such systems, however, are still being developed and heavily studied, and thus, no commercial units are yet available. A very useful survey on the state of the art of such technologies is given by Smolinka et al.²⁰  or in a book edited by Stolten.²¹ 

When considering efficiency of electrolyzers, which is one important criterion to judge their performance, one has to be careful in checking how the efficiency is defined. Generally, efficiency is defined as energy output (as given by the energy content of the hydrogen produced) divided by energy input. Typically, electrolyzers are just driven by the electricity supplied, without additional heat supply. Thus, the energy input is directly provided by the electrical power consumed by the electrolyzer.

As for an energy output in the form of hydrogen, it is important to consider whether the higher heating value (HHV; condensation enthalpy of water included, 39.4 kWh/kg H₂ or 3.52 kWh/m³ H₂) or the lower heating value (LHV; combustion product gaseous water 33.3 kWh/kg H₂ or 2.96 kWh/m³ H₂) should be used. On the one hand, should hydrogen be converted into another form of energy, it is appropriate to use the lower heating value, but on the other hand, should hydrogen be utilized as a chemical raw material, the higher heating value is more suitable for the analysis.

One can also assess the energy efficiency by considering the thermodynamic minimum voltage required for water splitting divided by the voltage applied under operation conditions. The equilibrium (reversible) potential for the water splitting reaction is 1.23 V, based on ΔG_R⁰. However, under adiabatic conditions, in order to provide the required heat for the reaction, the potential has to be higher at 1.48 V (thermoneutral voltage at 298 K). Due to the overpotentials and the Ohmic losses in electrolyzers, this heat is typically available in any case, and heat supply is not the reason to use higher potential than 1.23 V. Thus, 1.23 V is probably the most useful value for analyzing efficiency based on voltage. Finally, one has to check whether efficiency is based on the cell itself or on the electrolyzer plant, i.e. including balance-of-plant energy needs.

In order to compare technical systems, efficiencies are often not used, but absolute values, i.e. total energy requirement for the hydrogen production in kWh/kg or m³ of the H₂ produced. In 2008, based on company questionnaires and a literature search, a study considered an energy requirement value of 62.8 kWh/kg H₂ as the base case. This estimate corresponds to an efficiency of 53% (LHV) or 63% (HHV).¹ In another survey, energy requirements between 4 and 5 kWh/m³ H₂ are quoted.⁴ In turn, in Ref. 13, a value of 4.3 kWh/m³ was proposed (based on information from one manufacturer). In Ref. 20, values between 4.1 and 6.3 kWh/m³ are given for alkaline electrolyzers, where the lower values correspond to large atmospheric units and the higher values to very small pressurized units. However, in most of these publications, it is not fully clear which balance-of-plant components are included in the analyses. Such a differentiation is available for PEM electrolyzers.²⁰  Stack productivities of 4–5 kWh/m³ H₂ are quoted, rather independent of size (it has to be noted that several of the systems considered were precommercial). On the systems level, the smallest units (less than 1 m³/h) had an energy consumption of up to 8 kWh/m³, whereas the large systems (of several 10 m³/h) showed values ranging from 4.5 to 6 kWh/m³. One should note, however, that the PEM technology is not as mature as alkaline electrolyzer technology. Alkaline electrolyzers, which are about one order of magnitude larger than PEM systems, are commercially available.

While alkaline electrolysis is already well advanced, there is still a high level of development work going on (for a recent survey, see Ref. 22). The electrolyte is a concentrated (20–40%) KOH solution, cathodes (where hydrogen is evolved) are typically made of corrugated steel sheets on which catalysts are typically deposited in order to improve efficiency. Unlike PEM electrolyzers, the alkaline electrolysis does not utilize noble metals as catalysts, but instead, nickel or nickel compounds are typically employed. Nickel, an active catalyst for electrolysis, or nickel alloys are used as electrodes on the anode side in the form of sheets, which are structured in a way to maximize surface area. There are also electrolyzers where both electrodes are made of nickel or nickel alloys. Most commercial electrolyzers are operated with electrodes mounted in series (bipolar plates) so that overall high voltages can be employed.

For most applications, be it for direct use or for intermediate storage, hydrogen would be required under high pressure. Compression is energy intensive, and it would thus be advantageous to carry out electrolysis at higher pressure. Thermodynamically, this would require higher reversible and thermoneutral potentials, but this is often offset by reduction of the overpotential caused by small gas bubbles formed under high pressure.²³  As a result, almost identical efficiencies are found, on the cell level, for both atmospheric and pressure electrolyzers, especially for large-scale units.^20,24 

On the one hand, high-pressure electrolyzers save the investment and energy required for postelectrolysis compression of the hydrogen produced, but on the other hand, this advantage is partly offset by the more expensive and complex layout. Most alkaline pressure electrolyzers operate under moderate pressures between 0.2 and 1.0 MPa, but systems are available also in the range of 1 MPa up to several MPa.²⁴  Alkaline electrolyzers have one disadvantage when combined with intermittent renewable energy. They typically have a lower operation threshold of approximately 20% of the maximum capacity rating and a slower dynamic behavior.¹³  They thus cannot be shut down completely without significant repercussions for the overall performance.

In contrast, PEM electrolyzers (the state-of-the-art has recently been surveyed in Ref. 25) have a low operation threshold of 0% of the power rating. Practically, however, one would require approximately 5% of the rating to supply the electricity to the auxiliary components.²⁰  On the high-power side, such systems can also be operated appreciably above the nominal rating. For a Siemens prototype, short-term operation at three-fold nominal rating is reported.²⁶  In addition, power can be ramped up and down with much higher transients than for alkaline electrolyzers.

PEM electrolyzers became possible when proton conducting polymer membranes were available as solid electrolytes. Compared to alkaline electrolyzers, they have the main advantages of higher possible current densities, lower hydrogen crossover, and better dynamic behavior. Moreover, they are apt to operate under high pressure, since hydrogen crossover is less problematic. PEM electrolyzers operate under acidic conditions (around pH 2), which poses substantial corrosion problems for the system components operating at high potentials. This holds true especially for the catalysts, which are, in most cases, iridium on the anode side (for oxygen evolution) and platinum at the cathode. Since these elements are scarce and expensive, research efforts are underway to replace them by cheaper and more abundant alternatives. However, currently the noble metals still have to be considered the state-of-the-art, contributing appreciably to the investment costs for the stack.

Hydrogen production via electrolysis, be it by using alkaline or PEM electrolyzers, is currently not cost competitive. Two extreme cases can be discerned with respect to the mode of operation: (i) hydrogen could only be produced by electrolysis during times of very low electricity prices, which would lead to a high effect of the capital expenditure (capex) on hydrogen price, or (ii) the electrolyzers could be operated basically all the time in order to compensate for the high capex. The latter process mode would translate into high operation costs, since electricity would also be bought at peak costs. Altogether, both strategies are certainly inappropriate choices. Accordingly, an optimized operation scenario has to be found that minimizes total production costs by operating the electrolyzer as much as possible at moderate overall electricity costs.

Mansilla et al.¹³  have analyzed the cost of hydrogen from electricity for France, Germany and Spain, taking the real spot price electricity data into account for the years 2010, 2011 and 2012. The authors assumed continuous electrolyzer operation, and compared this with an operation mode in which the electrolyzer was only operated when the electricity price was below a certain threshold. Calculated prices for all three countries and years were – with one exception – between 3 and 3.5 €/kg H₂, with only very little difference between the two cases discussed. Should a reduction of electrolyzer investment of 50% be assumed, prices drop by approximately 0.5 €/kg H₂, and operation at low electricity cost only become somewhat more favorable, although the effect upon it was still below 5%.

Similar costs for hydrogen from off-peak power by electrolysis (for a capacity of 10 tons per day) were estimated for the United States.²⁷  The hydrogen costs calculated by Saur and Ainscough²⁸  for hydrogen from wind electricity ranged between 3.74 and 5.86 $/kg H₂ (dollars at purchasing power for the year 2007), which is in a similar range as the values calculated for the European countries quoted above. An interactive online tool is available to perform cost analyses for different sites in the USA using different assumptions.²⁹ 

Smolinka et al.²⁰  estimate hydrogen-production costs for several cases, concerning electrolyzer technology, size of the plant, and yearly operation hours. In most cases for hydrogen production in the renewable energy systems, costs were estimated between approximately 3 and 4 €/kg H₂. High costs of about 9 €/kg H₂ were calculated for the case of a small PEM electrolyzer with a capacity of 30 m³/h, operation under 2.5 MPa, plant utilization of 75% at electricity costs of 0.09 €/kWh. At high utilization, fixed costs do not impair the overall hydrogen cost. However, at utilization below 50%, fixed costs dominate the cost of the hydrogen produced. In this study, it was also estimated that costs of about 2 €/kg H₂ should be possible at high plant utilization, assuming additional development of technology. Investment cost reduction probably needs to concentrate on the stack cost, since the stack contributes about 46% to the total cost for PEM electrolyzers and 57%, for alkaline electrolyzers.¹

In summary, the available data suggest that costs of hydrogen produced by electrolysis are not competitive with other methods for hydrogen production. However, electrolyzers can relatively easily be decentralized and deliver hydrogen at the point of use. These factors speak in favor of hydrogen production by electrolysis. Nevertheless, it is important to consider the high investment costs for the construction of small-sized plants, which accounts for the high price of hydrogen from a small PEM electrolyzer, as reported in Ref. 20.

In addition to the more conventional routes for hydrogen production, the gas can also be synthesized via thermochemical cycles. These cycles were originally developed for use in connection with high-temperature nuclear power plants, but several of these processes are also suitable for use in connection with concentrating solar power plants.¹⁶  In total, more than 200 such cycles are claimed.²⁷  However, commercial processes have not yet been implemented. For nuclear heat, the sulfur–iodine (S–I) and the copper–chlorine (Cu–Cl) cycles appear to be among the systems most intensively explored. The S–I cycle relies on the following three prototypical reactions, the temperature levels and specifics may vary to some extent:²⁷ 

The Cu–Cl cycle is again available in several embodiments, but also here prototypical reactions describe the system:²⁷ 

In both cases, the net reaction is the splitting of water into hydrogen and oxygen.

Thermal efficiencies for these thermochemical cycles are reported to be as high as 43% for the Cu–Cl cycle and 52% for the S–I cycle.²⁷  In the various cost estimates, the Cu-Cl and the S–I cycle are rather similar, with the S–I cycle estimated at about 1.5 $/kg H₂ (2003 basis) and about 2 $/kg H₂ for the Cu–Cl cycle.²⁷  In another study, the cost was estimated depending on plant size, where the hydrogen price from a 200 t/d plant is given as 2 $/kg H₂, while the price increases to 3.49 $/kg H₂ for a 2 t/d plant.³⁰  The cost is also compared to that of steam gasification of methane, which is given as 2.67 $/kg H₂ for a 10 t/d plant. This value, however, takes CO₂ costs into account, and relatively high costs for distribution; in turn, for the thermochemical cycle, only about half of the costs for distribution are assumed. Therefore, these results should be considered with caution.

The published data suggest that hydrogen can be produced by thermochemical cycles at comparable costs to those published for the steam-reforming processes. Solar heat from concentrating solar power plants could be used to drive such cycles. For solarthermal plants, the Cu–Cl or the S–I systems, however, are not the center of the attention. Research focus is on the Zn/ZnO system,³¹  for which costs of 0.14–0.15 $/kWh is estimated (4.66 to 5 $/kg H₂), or based on solarthermal reduction of CeO₂ and reoxidation with water vapor, by which hydrogen is released.³² 

Regarding the utilization of hydrogen in upgrading processes performed on biomass, the hydrogen should be supplied at the biorefinery site, or at least at a hydrogenation plant, where e.g. pyrolysis oil would be hydrogenated to stabilize it prior to its transportation and storage. Ideally, hydrogen is provided onsite, reducing the considerable costs associated with its transportation, and thus, being conducive to the economics of the entire chain of process for biomass conversion.

In general, hydrogen can be transported and stored either under high pressure or in liquefied form.³³  In addition, hydrogen can also be transported via pipelines. Other means are possible, but have not reached the same level of maturity and shall not be discussed here; a survey is given in Ref. 34. For large amounts of hydrogen, the transport and storage of liquefied hydrogen is the best approach. Usually, the higher the amount transported or stored, the lower the problems with reducing boil-off losses are, since with higher amounts the surface-to-volume ratio of the cryogenic vessels becomes more favorable. However, liquefaction typically entails a high loss of energy. It is estimated that at up to 30% of the energy content of the hydrogen is lost upon H₂ liquefaction.³⁵  Compression is less energy intensive, although also here appreciable losses are encountered. Typical pressures for transport are 20 MPa; for onboard storage for fuel cell powered vehicles, 70 MPa are the state-of-the-art. For pressures of 70 MPa, the energy requirement for compression is about 15% of the energy content of the hydrogen.³⁴  Compressed hydrogen is transported in steel cylinders or bundles of cylinders. In fuel-cell-powered cars, the pressure vessel is a fiber-reinforced polymer composite. Since the force scales with the area of the container, the scaling behavior is the opposite of that for the liquefied hydrogen: the larger the container, the more serious the problems. Thus, the strength of the container material has to be markedly increased.

Geological formations, often salt domes, are the best storage option for large amounts of hydrogen. Such domes are in operation at several sites with a high concentration of refineries and chemical production plants. In Texas, there are two large salt domes (the Clemens dome and the Moss Bluff dome) connected by a pipeline system in a refinery region close to Houston. The Clemens dome has a volume of 580 000 m³, and a storage capacity of 5400 t. Leak rates as low as 0.01% per year are reported.³⁶  Three caverns of 70 000 m³ each are operated in Teesside, UK. Above ground, pressurized hydrogen is the best option for short times and small amounts; for larger volumes, liquefied hydrogen is the preferred alternative.³⁷ 

Considering the demand of hydrogen for the conversion of biomass, several conclusions with respect to the production and infrastructure technology can be drawn. Nonetheless, a detailed analysis is certainly required for each potential site. In order to avoid the transportation cost, onsite production is favored. In effect, the cheapest alternative seems to be biomass reforming, since the infrastructure for transporting and handling of biomass is available already for the biomass itself, since the cost of hydrogen from biomass reforming does not appear to be substantially higher than natural-gas reforming. Moreover, carbon-containing waste streams from biomass processing may be used for gasification, which could potentially further decrease the costs. Should fossil carbon resources be available at the site where the hydrogen is needed, reforming or co-reforming of fossil feedstocks could also be an attractive option. If this is not the case, electrolysis might be suitable, even if the production cost is still relatively high, because electrolytic processes can respond to the demand for hydrogen highly dynamically. For electrolysis, the production cost is at least partially offset by the fact that the expensive transportation of hydrogen is not required. In addition, hydrogen could be supplied directly at the pressure level needed. Thermochemical cycles using renewable energy do not seem to be suitable to provide hydrogen for biomass conversion, since solarthermal plants are typically located in arid regions with a high fraction of direct irradiation and little rainfall. Logically, these regions are not conducive to energy crops, and so high costs for the hydrogen transportation would result.

Depending on the kinds of feedstocks and the initial conversion processes, the details of the processes used for upgrading of the initial products change. Gasification of biomass typically results in a syngas with a H₂/CO ratio of approximately one. Such a syngas is highly suitable for the direct one-step production of dimethylether (DME),³⁸  an interesting fuel molecule for compression ignition engines.³⁹  However, any other syngas conversion process requires the addition of hydrogen to the biomass-derived syngas. Methanol production needs a H₂/CO ratio of two, the Fischer–Tropsch process requires essentially the same ratio, and methanation even requires a ratio of three. Thus, either gasification needs to be coupled with a water-gas-shift stage in order to increase the hydrogen content of the syngas, or hydrogen has to be supplied externally. Since hydrogen production from biomass seems to be almost competitive with hydrogen production from fossil sources, changing the CO/H₂ ratio by the shift reaction is probably the economic option, since the gasification unit is available in any case.

The situation could be somewhat different in two-step thermochemical processes, where first a pyrolysis oil is generated in a pyrolysis step, which is later converted into syngas in a centralized plant. The pyrolysis oil has a number of limitations, including poor chemical stability, and thus it typically needs to be subjected to upgrading treatments.⁴⁰  The upgrading serves the purpose to reduce the oxygen content and increase the H/C ratio. Hydrodeoxygenation is an approach to accomplish this. However, the removal of the high oxygen content of crude pyrolysis oil should consume very high amounts of hydrogen. Combination with other processes, which stabilize the pyrolysis oil, is thus advisable. Nevertheless, facilities for production of pyrolysis oil will probably always require some hydrogen, and the question of how this is best produced must be answered.

The production of hydrogen from biomass appears to be less favorable than in the case of direct biomass gasification, since the gasifier would need to be built up separately. Separation of decentralized generation of pyrolysis oil and centralized gasification and downstream processes is a key advantage of pyrolysis/processing routes starting with biomass. Integration of other complex equipment, such as a gasifier for hydrogen production, at the pyrolysis site may jeopardize the attractiveness of the overall scheme. Of some advantage is the fact that the full infrastructure for biomass handling (transportation, storage, preprocessing) is available at the pyrolysis plant. Whether this makes the gasification onsite more attractive is questionable, though. Overall, the viability of hydrogen generation from biomass at the pyrolysis site is most probably a question of scale on which the hydrogen would be needed. For low demand, one would purchase hydrogen at market price; if large amounts of hydrogen are required, a separate gasifier unit might become attractive.

Yet another situation is encountered, if biomass enters into the biorefinery stream via the hydrolytic pathway.⁴¹  In addition, the oxygen content of the primary products (sugars from cellulose and hemicellulose, aromatic alcohols from the lignin fraction) has to be reduced and the C/H ratio needs to be increased in order to obtain valuable chemical products or fuels. Most pathways described in the literature starting from sugars use externally supplied hydrogen to optimize C/H and C/O ratios. This holds, for instance, for the pathways leading to furan derivatives (Figure 1.3).⁴² 

Glucose is first isomerized to fructose and then by three dehydration reactions to 5-hydroxymethylfurfural, which is in turn hydrogenated to 2,5-dimethylfuran (DMF), a compound with excellent fuel properties.⁴³  This reaction is ideal, since it requires the minimum amount of hydrogen – however, this is still three moles of hydrogen per mole of DMF; further hydrogenation of the ring and/or hydrogenolysis of the methyl groups should be avoided. Thus, there are highly attractive pathways in which the hydrogen is generated in situ by partial catalytic aqueous phase reforming of the sugars. The formed hydrogen is used in the conversion of the sugars into monofunctional alcohols, ketones, or carboxylic acids, which undergo further conversion rendering alkanes.⁴⁴  If such processes could be sufficiently well controlled on a commercial scale, they would most probably be advantageous over the hydrogen production in a dedicated separate plant or externally sourced. However, the partial catalytic aqueous phase reforming of the sugars needs to be assessed in detail for specific process layouts.

Finally, hydrogen is also required, if the lignin fraction is to be upgraded to liquid compounds – nowadays the lignin is mostly burnt in order to generate process heat or electrical energy in biomass processing units. Due to the complexity of lignin and the relative inertness of phenols against dehydroxylation, conversion of the aromatic alcohols into aromatic or aliphatic hydrocarbons is rather difficult. This reaction is possible, for instance, at temperatures of 473 K at hydrogen pressure of 5 MPa over mixed catalysts consisting of H-ZSM-5 and Pd/C.⁴⁵  This method leads to deoxygenation and ring hydrogenation, so that substantial amounts of hydrogen are consumed. For the generation of this hydrogen, probably gasification of biomass onsite would be the economic option, since it is close to competitive in any case, and the facilities for biomass handling and storage are already in place at a biorefinery site.

An alternative to hydrogenation by externally supplied hydrogen is transfer hydrogenation from an alcohol or olefin to yield deoxygenated products. This approach has been proven successfully in a recent study by Wang and Rinaldi.⁴⁶  In this study, a mixture of Raney nickel and H-beta-zeolite was used as catalysts, and 2-propanol as an H-donor, for the low-severity hydrodeoxygenation of phenol into benzene. Also with real feeds, such as lignin and pyrolysis oil, high hydrogenation activity and good selectivity to arenes were observed in this study, which reduces hydrogen consumption, as the ring is not hydrogenated. This transfer hydrogenation would be especially favorable, if the aliphatic alcohols – such as cyclohexanols obtained from full saturation of lignin – are used as the H-donors. This process option does not only overcome the need for hydrogenation of acetone and its recycle in the process, but also works as an effective H-recovery process. Accordingly, the molecular hydrogen stored in the cyclohexanols could be used for the conversion of the phenolics into arenes through catalytic transfer hydrogenation performed at a second-stage process.

Due to the high O/C and low H/C ratio of biomass, as compared to traditional fuels and chemicals, large amounts of hydrogen are required in biorefinery schemes, regardless of whether the thermochemical or hydrolytic pathway is chosen for the conversion of lignocellulose biomass. There are several routes available for hydrogen production. Currently, the available data from technoeconomic analyses suggest that onsite hydrogen production is more advantageous over externally sourced hydrogen due to savings in H₂ transportation, but this economy will certainly depend on the amounts used at a site. Of the different hydrogen production routes, gasification methods seem to be economically advantageous, regardless of the feedstock used (i.e. gas, coal and biomass). Since a site for biomass conversion, irrespective of the detailed processes implemented, will need all the installations for transportation and storage of biomass in any case, there are most probably synergies between biomass gasification for hydrogen production and the other processes operated at a specific site. Nevertheless, in order to find the most economical solution, a detailed analysis will be necessary for each site in question. This analysis is certainly complicated by the fact that the boundary conditions in this field (e.g. CO₂ emissions cost, feedstock prices, availability of nonexpensive electricity, electrolyzer costs, and legal boundary conditions, to name only a few) are constantly changing.

Work in this field is funded by the Cluster of Excellence “Tailor-made fuels from biomass”, in addition to the basic funding provided by the MPG.