Progress on Catalyst Development for the Steam Reforming of Biomass and Waste Plastics Pyrolysis Volatiles: A Review

In recent decades, the production of H2 from biomass, waste plastics, and their mixtures has attracted increasing attention in the literature in order to overcome the environmental problems associated with global warming and CO2 emissions caused by conventional H2 production processes. In this regard, the strategy based on pyrolysis and in-line catalytic reforming allows for obtaining high H2 production from a wide variety of feedstocks. In addition, it provides several advantages compared to other thermochemical routes such as steam gasification, making it suitable for its further industrial implementation. This review analyzes the fundamental aspects involving the process of pyrolysis-reforming of biomass and waste plastics. However, the optimum design of transition metal based reforming catalysts is the bottleneck in the development of the process and final H2 production. Accordingly, this review focuses especially on the influence the catalytic materials (support, promoters, and active phase), synthesis methods, and pyrolysis-reforming conditions have on the process performance. The results reported in the literature for the steam reforming of the volatiles derived from biomass, plastic wastes, and biomass/plastics mixtures on different metal based catalysts have been compared and analyzed in terms of H2 production.


INTRODUCTION
The growing energy demand and the increasing awareness of the dependency on fossil fuels are promoting the use of alternative routes for the production of clean energy from sustainable raw materials and consumer society wastes. Currently, almost 80% of the global primary energy demand is supplied from crude oil, natural gas, and coal. 1 Thus, the development of H 2 technologies can help to alleviate the problems associated with global warming and climate change, promote sustainable development, and help to reduce CO 2 emissions.
The current global H 2 production is about 8 EJ/year, with around 96% being produced from fossil fuels. Thus, the actual sources whereby H 2 is produced consist of natural gas (48%), oil (30%), coal (18%), and water electrolysis (4%) (see Figure  1a). This distribution shows that the current H 2 generation is associated with the consumption of fossil fuels and the emission of CO 2 in their production processes (mostly steam reforming processes). In addition, its consumption is largely carried out in the petrochemical industry for the generation of automotive fuels, which in turn are CO 2 generators.
The future of the H 2 market is driven by the need to reduce CO 2 emissions, which requires its production from alternative raw materials and renewable sources. Thus, in order to achieve the global target signed in the Paris Agreement in 2015, in which the aim was to limit the global temperature rise in this century to 2°C, dramatic changes should be done in order to decrease CO 2 emissions by 60% until 2050. 2,3 In this scenario, the short-term implementation of H 2 production processes from biomass and waste is gaining growing attention. Consequently, the development of technologies for obtaining H 2 from biomass and waste, especially waste plastics, has deserved remarkable attention in the recent literature. 4−10 Among them, thermochemical routes have proven a great potential and efficiency. These processes can be grouped into direct and indirect routes. Thus, the direct routes are classified as follows: (i) pyrolysis (catalytic pyrolysis and high temperature pyrolysis), (ii) steam gasification, (iii) supercritical water gasification, and (iv) the steam reforming of pyrolysis volatiles. The indirect routes are those involving intermediate steps to obtain liquid products (bio-oil or plastics pyrolysis oil) for their subsequent reforming (steam reforming and gasification).
Steam gasification has been widely studied in the literature for the valorization of biomass and waste plastics, since it can generate H 2 -rich syngas 9,11,12 and the gaseous products can be directly used as fuel or intermediate products in the large scale chemical and fuel production. 13,14 The main challenge of gasification processes involves the quality of the syngas and its tar content and, in fact, remarkable research efforts have been made to improve tar elimination processes. 15−17 Thus, it is well-established that an optimized gasifier and a highly active catalyst can efficiently contribute to biomass tar elimination. 18 Under suitable operating conditions and using appropriate catalysts, biomass steam gasification allows obtaining H 2 productions in the 5−7 wt % range. 19−22 The higher carbon and hydrogen content of waste plastics increases H 2 production to values above 10 wt % when they are valorized by steam gasification. 23−25 An alternative strategy to direct gasification proposed for syngas production is the one based on the gasification of pyrolysis oil, especially the one derived from biomass pyrolysis. This process involves the advantage of reducing the expensive transportation costs of the biomass and waste. 26 Although the differentiation in the literature between bio-oil gasification and reforming is sometimes unclear and rather confusing, it has herein considered that bio-oil gasification is the process that requires higher temperatures (around 800−1400°C) than reforming reactions and is carried out without catalysts or in the presence of primary mineral catalysts. 8 Thus, the composition of the syngas obtained in the bio-oil gasification is similar to the one obtained in the biomass gasification, with higher H 2 yields being attained in the gas product from bio-oil reforming. The production of H 2 from bio-oil gasification is greatly influenced by the type of bio-oil used, reactor configuration, and operating conditions. Accordingly, H 2 production values from 1.4 to 12.6 wt % have been reported in the literature. 27−30 Bio-oil steam reforming is an indirect thermochemical route for H 2 production from biomass. 31,32 It should be noted that this strategy has been scarcely studied as a route for the upgrading of waste plastics. The bio-oil or liquid product from biomass pyrolysis has a higher energy density compared to biomass, leading to lower transportation costs and therefore allowing bio-oil valorization in centralized large scale catalytic conversion units. In spite of the operational problems associated with bio-oil handling and fast deactivation rate of the reforming catalyst, high H 2 productions have been reported. Thus, values above 12 wt % by mass unit of the organic bio-oil have been obtained. 33−36 The selection of suitable catalysts for oxygenate reforming has been extensively studied in the literature by using model compounds, the bio-oil aqueous phase, and raw bio-oil. 37,38 Thus, although a wide range of base transition metals, such as Ni, Co, and Fe, and noble metals, such as Rh, Pt, Ir, and Ru, have been studied, Ni based catalysts are the most used ones because they strike a suitable balance between activity and cost. 39,40 In addition, different supports and their modification with promoters have also been widely analyzed. 12,41,42 An alternative and direct thermochemical conversion route for H 2 production from biomass and waste is the strategy based on pyrolysis and in-line reforming of the volatiles, which has several advantages in comparison with the aforementioned routes. Thus, the integration of both reactors in the same unit allows selecting the optimum conditions in the pyrolysis and in-line reforming steps. 43 Therefore, operation at lower temperatures than those in the gasification process and the use of highly active reforming catalysts lead to the avoidance of tar formation, reduce material costs, and prevent catalyst deactivation by sintering. 44−46 Moreover, the direct contact of the reforming catalyst with the biomass and its impurities is avoided, given that they are retained in the pyrolysis reactor. 47 In addition, this process has shown a remarkable capacity for H 2 production. Concerning biomass conversion, values of around 10 wt % were obtained operating under optimum conditions and catalysts. 40,48−53 The volatiles derived from biomass pyrolysis can be classified into two fractions: (i) biooil or liquid fraction (the condensable fraction at the outlet of the pyrolyzer), which is formed by a complex mixture of oxygenated compounds (phenols, ketones, saccharides, furans, acids, and alcohols) and water, and (ii) the non-condensable gaseous stream, which is mainly made up of CO and CO 2 and, to a minor extent, by light hydrocarbons and H 2 . In the case of waste plastics, a wide range of conversions was reported depending on the polymer nature. Thus, polyolefins and polystyrene allow obtaining H 2 production values higher than 30 wt %, 43,46,54,55 whereas other polymers with lower carbon content, such as polyethylene terephthalate, lead to productions below 20 wt %. 56 The volatile stream derived from polyolefins (PE and PP) pyrolysis under mild conditions was mainly formed by waxes (C 21 +) followed by hydrocarbons in Energy & Fuels pubs.acs.org/EF Review the diesel fraction (C 12 −C 20 ), with the yields of light oil (C 5 − C 11 ) and gases being rather low. The products obtained in the pyrolysis of polystyrene (PS) are made up of aromatic hydrocarbons (styrene recovery being higher than 70%) and a small yield of light gaseous compounds. It is to note that no solid residue is formed in the pyrolysis of polyolefins and polystyrene, and therefore all the products that make up the volatile stream are fed into the reforming step. Conversely, a solid residue is obtained in the pyrolysis of PET, with the volatile stream being mainly composed of non-condensable gases (CO and CO 2 ), oil (acetaldehyde), and a mixture of oxygenated compounds, which are solid at room temperature (mostly benzoic acid). Therefore, H 2 production by pyrolysis-reforming is highly dependent on the raw material. Furthermore, the volatile composition to be reformed affects the reaction mechanism as well as the catalyst deactivation rate. 57 However, the development of the combined process of pyrolysis and in-line reforming is conditioned by the efficiency of the reforming catalyst. Thus, highly selective and active catalysts are required to enhance H 2 production and ensure full conversion of pyrolysis volatiles and therefore avoid tar formation. Moreover, the complex nature of the volatile stream derived from the pyrolysis of biomass and waste plastics causes a fast deactivation rate. 58 Thus, a remarkable research effort has been made in recent years to develop new catalysts and overcome the mentioned challenges. Accordingly, a wide variety of catalysts based on different supports, promoters, and active phases have been proposed in the literature. This review analyzes the development and application of these catalysts for the reforming of biomass and waste plastics. Moreover, a general overview on the combined process of pyrolysis and inline reforming has also been included. The role played by pyrolysis and reforming conditions on process performance was briefly discussed, and technological aspects were also considered.

PYROLYSIS AND IN-LINE STEAM REFORMING
The pyrolysis-reforming strategy pursues H 2 production by combining two reaction steps in a single process (see Figure  2). The aim of the pyrolysis step is to convert the solid feedstock into a volatile stream suitable for further transformation into a H 2 rich gas in the subsequent catalytic reforming step. The composition and yields of the pyrolysis volatiles (gas and pyrolysis oil) depend on the pyrolysis conditions and feedstock characteristics. Moreover, a solid product or char can also be formed in the pyrolysis step, which is a valuable byproduct that is not fed into the reforming reactor. The pyrolysis step is usually performed at around 500°C , as this is the minimum temperature to ensure full devolatilization of biomass and waste plastics.
The step of catalytic steam reforming of the pyrolysis volatiles is commonly carried out between 600 and 800°C. This temperature range is conditioned by, on the one hand, the low reaction rate below 600°C and, on the other hand, the catalyst instability due to metal sintering when operation is carried out at high temperatures. Different supported metallic catalysts have been used in this process, with Ni based ones the most common due to their suitable activity and low price. Steam partial pressure and catalyst space time also play a significant role in the performance of the reforming step and must be carefully considered.
In the following sections, the most relevant aspects involving the steps of pyrolysis and reforming are briefly discussed. Moreover, the main reactor designs used in this process are also presented and analyzed.

Pyrolysis
Step. Pyrolysis is a thermochemical conversion process performed under inert conditions. Although pyrolysis is usually performed under a steam atmosphere, the mild conditions used for pyrolysis in the combined process of pyrolysis-reforming minimize the impact of steam in the pyrolysis mechanism 44,50 when it is used as an inert carrier gas. Pyrolysis processes are often divided into three groups depending on the heating rate and gas residence time, namely, fast, intermediate, and slow pyrolysis. 59,60 Thus, fast pyrolysis is usually conducted at moderate temperatures under very high heating rates (10 3 −10 4°C s −1 ) and short residence times (below 1 s). These conditions minimize secondary reactions leading to the non-condensable gas fraction and enhance the yield of primary products accounting for the liquid fraction. However, slow pyrolysis is characterized by long residence times and slow heating rates. It is to note that these conditions favor the formation of gas and solid products at the expense of the liquid one. Finally, intermediate pyrolysis conditions are halfway between those of fast and slow pyrolysis. In the processes involving pyrolysis and reforming, the most common fast pyrolysis reactors are fluidized and spouted beds, and their use is associated with the operation under a continuous regime. 54,56,61,62 However, fixed bed reactors are used under slow pyrolysis conditions (below 40°C min −1 ) and they operate in a batch regime. 53,63,64 Thermal pyrolysis of biomass and waste plastics leads to wide and complex product distributions. The features of the obtained products are briefly mentioned below due to their impact on the subsequent reforming step. Biomass pyrolysis leads to three fractions, namely, gas, bio-oil, and char, with their yields depending on the pyrolysis conditions and original biomass composition. Although biomass has a variable composition in terms of cellulose, hemicellulose, lignin, sugar, protein and ash contents, some general features apply in all cases. Thus, the gases are mainly made up of CO and CO 2 , with minor contents of light hydrocarbons and H 2 . Biooil is the main product obtained in biomass pyrolysis, with its yield reaching 75 wt % under suitable processing conditions. 59 The bio-oil is a complex mixture of oxygenated compounds, with phenols, ketones, saccharides, furans, acids, or alcohols being worth mentioning. Moreover, a significant amount of water is also formed in the biomass thermal degradation. The char is a carbonaceous material obtained as a byproduct in the pyrolysis-reforming process, which is not treated in the reforming step. However, it has several applications, as are those related to sorbents, fertilizers, catalyst supports, and soil amender. 65,66 The type of waste plastic fully conditions the pyrolysis mechanism and the obtained product distribution. Accordingly, the product distribution obtained in the pyrolysis of each type of plastic waste deserves an individual description. Polyolefins, such as polyethylene (PE) and polypropylene (PP), are the most used plastics, whose thermal degradation takes place via a random radical scission mechanism leading to a wide product distribution from light gaseous hydrocarbons to heavy waxes. Furthermore, under suitable operating conditions, the yield of solid residue from these plastics is negligible. 67,68 It should be noted that, under mild pyrolysis conditions, waxes and hydrocarbons in the diesel range are the prevailing products, with the yields of light oil and gases being low. 69−71 However, the thermal degradation of polystyrene (PS) is a highly selective process. In fact, styrene recovery is above 90%, with the remaining products being other aromatics. 72−74 Polyethylene terephthalate (PET) is a commodity plastic whose valorization by pyrolysis is challenging. Thus, the product distribution is a complex mixture of oxygenates (gaseous, liquid, and solid at room temperature) and a variable yield of solid residue. 75,76 The yield of gas is higher than 40% and consists of CO and CO 2 , whereas the main heavy compounds are oxygenates of aromatic nature, such as benzoic acid. The co-pyrolysis of waste plastics and biomass is attracting increasing attention in recent years due to the interaction between their pyrolysis products and therefore the synergistic effects observed in their joint treatment. 77 In spite of these interactions, the composition of the products is basically that corresponding to the contribution of the individual raw materials pyrolyzed.

Catalytic Steam Reforming
Step. In the steam reforming step, the volatile stream from the pyrolysis step reaches the catalytic bed and reacts with steam on the catalyst active sites to yield a H 2 rich gaseous product. The most relevant reactions in the reforming reactor are summarized in Table 1. These reactions can be divided into two groups, namely, heterogeneous catalytic reactions and secondary Interconversion: reactions in the gas phase. The first group includes the main desired reactions, which are promoted by the use of metallic catalysts. Steam reforming reactions convert oxygenates (eq 3) and hydrocarbons (eqs 4 and 5) into H 2 and CO. These reactions are highly endothermic, and they are therefore favored at high temperatures. Moreover, CO can be further oxidized to CO 2 by the water gas shift (WGS) reaction (eq 6), forming additional H 2 . This reaction is exothermic and is therefore hindered at high temperatures. Secondary reactions include thermal cracking reactions of oxygenates (eq 1) and hydrocarbons (eq 2) and a wide variety of interconversion reactions (eq 7). These reactions are favored by severe operating conditions, i.e., high temperatures and long residence times. However, their impact is in general limited due to the competence with the much faster reactions promoted by the catalysts, i.e., steam reforming reactions. Thus, they only influence the final product composition when the process is performed under low conversion conditions (low space times or severely deactivated catalysts).
Apart from the conventional reforming under a steam atmosphere, the use of CO 2 or dry reforming has been also proposed, especially when methane is used as the feedstock. 78 Nevertheless, the low hydrogen content of biomass limits the interest of its valorization by dry reforming. However, the dry reforming of the volatiles derived from waste plastics leads to a significant H 2 production. 79 The oxidative steam reforming or oxygen co-feeding is a common strategy applied in catalytic steam reforming for the attenuation of the endothermicity of the process. Although the presence of oxygen in the reforming reactor contributes to reducing H 2 production, it also has positive effects, such as the in situ combustion of the coke deposited on the catalysts and therefore, stability improvement. 80 2.3. Pyrolysis-Reforming Reactor Configurations. A wide variety of reactor configurations has been proposed in the literature for biomass and waste pyrolysis and in-line reforming. Figure 3 summarizes the main reactor designs used. However, the majority of the studies were carried out in batch laboratory units, and they are therefore of a preliminary nature. 9 Nevertheless, continuous operation regime is highly relevant in the pyrolysis-reforming process. Thus, continuous pyrolysis leads to a volatile stream of constant composition once steady state conditions have been attained. This fact not only allows for the extrapolation of the results to the conditions of industrial reactors but also eases the evaluation of catalyst performance as a pyrolysis stream with homogeneous composition throughout the time treated. Moreover, steady state conditions also allow for a better control of the process conditions and catalyst stability with time on stream.
A combination of two fixed bed reactors for the pyrolysis and reforming steps has been widely used due to its simple design and operation as well as limited investment cost. The most common approach is the operation in a batch regime with the pyrolysis step being performed under slow pyrolysis conditions by using heating rates below 50°C min −1 . 47,81−85 Thus, the pyrolysis volatiles formed throughout the heating process are transferred to the fixed bed reforming reactor, which operates under isothermal regime at the desired reforming temperature. The main shortcoming of this reactor configuration lies in its scaling up due to the poor heat transfer rate and complex control of operating conditions in the pyrolysis step. Besides, the use of a fixed bed reactor in the reforming step may involve operational problems related to bed blockage by coke deposition on the catalyst surface. Moreover, a two fixed bed configuration has also been used in the continuous regime for biomass 86−89 and waste plastics 43,90 valorization. It is to note that, in these studies, the continuous feed into the pyrolysis reactor leads to high heating rates. The use of fast pyrolysis reactors, such as fluidized and spouted beds, has also been proposed in the literature. These studies were carried out with a continuous biomass/plastic feed. The main advantages of these reactors are related to their suitable gas−solid contact features, high heat and mass transfer rates between phases and bed isothermicity. Fluidized bed reactors have limitations related to the feed and particle size in the bed, with defluidization being the main shortcoming, especially when plastic wastes are handled. The vigorous solid circulation movement in spouted beds allows handling coarse solids with irregular texture. 91,92 The suitability of the conical spouted bed reactor has been proven in the pyrolysis of different residues, such as biomass, sewage sludge, 93 tires, 94 and plastics. 75 In fact, a conical spouted bed reactor pilot plant (25 kg h −1 ) for biomass fast pyrolysis has been successfully operated. 95 Thus, the pyrolysis in bubbling fluidized bed reactors was combined with the reforming of the volatiles in fixed 49 and fluidized bed 54,96 reactors. In the same line, the use of spouted beds in the pyrolysis step was also combined with fixed 97,98 and fluidized bed 99−101 reforming reactors. Moreover, Efika et al. 102 used a screw kiln/fixed bed reactor configuration for the continuous biomass processing. Kuchonthara et al. 103 studied biomass pyrolysis-reforming over a K 2 CO 3 /NiO/γ-Al 2 O 3 catalyst in a batch reaction unit including a drop tube furnace and fixed bed for the pyrolysis and reforming steps, respectively.
It should be noted that the use of a fast pyrolysis reactor has practical advantages for the full scale operation in the pyrolysisreforming process. Thus, it ensures an efficient conversion of biomass into volatiles with low yields of char 59,104 and therefore increases the H 2 production potential of the process. Moreover, these reactors are easy to operate and control in the continuous regime and allow for continuous char removal. In the same line, fluidized beds have advantages for the steam reforming process in relation to fixed beds. On the one hand, a better control of process conditions can be attained, which is essential in the case of temperature in this highly endothermic process. On the other hand, the pyrolysis-reforming process is greatly conditioned by the fast deactivation rate of the catalysts, 8,9,105 and the use of a fluidized bed reactor allows implementing advanced catalyst regeneration strategies. Accordingly, a careful selection of the reactor design must be considered for full scale operation.

REFORMING CATALYSTS
Although great effort has been devoted to the development of catalysts for the reforming of bio-oil model compounds, the studies conducted by feeding crude bio-oil are scarce. Similarly, few studies deal with the assessment of catalyst performance in the pyrolysis-reforming of plastic wastes. Therefore, knowledge on the performance of reforming catalysts under real process conditions, i.e., with a real pyrolysis volatile composition, is still limited. In fact, most of the studies have been carried out in batch-scale plants, and therefore further research effort is required in order to establish a suitable relationship between catalyst properties and their activity and stability in continuous large-scale plants.
Energy & Fuels pubs.acs.org/EF Review Thus, this section deals with a brief introduction of heterogeneous catalysts, with special attention being paid to catalyst design (section 3.1) and the main causes and mechanisms of catalyst deactivation (section 3.2). Moreover, section 4 summarizes the aspects involving the catalysts used in the literature for biomass and plastic wastes pyrolysis and inline reforming.
3.1. Catalyst Design. The design of a suitable catalyst for a specific process involves several challenges, as a wide range of features should be considered and optimized in order to achieve the desired performance. Accordingly, a large number of studies have been conducted on the reforming of bio-oil oxygenated compounds on commercial catalysts. 50,106,107 However, the starting point for selecting possible materials and conditions for catalyst synthesis requires delving into the mechanisms of the process and the analysis and comparison of literature results.
3.1.1. Catalyst Components. The main components of a typical heterogeneous catalyst are as follows ( Figure 4): (i) active phase, metal that provides active sites for the chemical reaction; (ii) support or carrier, high specific surface area oxide or carbon on which the active phase is dispersed and stabilized; and (iii) promoter, additive that improves catalyst properties, i.e., activity, selectivity, and catalyst life.
Consequently, the selection of suitable catalytic materials is one of the most important factors for catalyst synthesis. Accordingly, given that active components are responsible for the main chemical reactions, the metal oxide selected as the active phase should promote reforming and WGS reactions in order to enhance H 2 production in the reforming step. Nibased catalysts have been widely used in the literature for CH 4 and naphtha reforming 108,109 as well as in the reforming of oxygenated compounds derived from biomass pyrolysis. 110−112 Besides, other base transition metals, such as Co or Fe, and especially noble metals, such as Rh, Pt, Ir, and Ru, have been widely studied in the literature. 113−115 Moreover, a suitable support must provide a high specific surface area, adequate pore distribution, mechanical strength, and good thermal stability. These two latter features are usually key factors in the steam reforming reactions. 116 In addition, an ideal support should not have catalytic activity promoting secondary reactions leading to catalyst deactivation, as happens with Al 2 O 3 support, which although is the most used support in reforming reactions, its acid properties promote coke deposition under reaction conditions and lead to catalyst deactivation. 117,118 In this way, the support should facilitate the dispersion of the active phase and modulate catalyst activity.
The addition of the right promoter is an interesting option, since its incorporation into the catalyst enhances activity, selectivity, and stability. The promoters can be used for modifying both the active phase and the support and may therefore contribute to the following aspects: (i) improving thermal stability, as is the case of small amounts of SiO 2 or ZrO 2 into γ-Al 2 O 3 , which shift the phase transition from γ-Al 2 O 3 to α-Al 2 O 3 toward higher temperatures; 119 (ii) hindering undesirable secondary reactions leading to catalyst deactivation by coke formation, which are attenuated by adding promoters with basic properties or capacity for gasifying the coke deposited during the reaction; 120,121 and (iii) improving the dispersion of the active phase metal. 12,122 3.1.2. Catalyst Properties. The design of reforming catalysts implies a compromise involving mechanical, physico-chemical, and catalytic properties. The balance of these interconnected properties is illustrated in Figure 5, with the relative significance of these features being greatly influenced by the process type, reactor design, process conditions, and economic factors. Thus, the selection of a suitable catalyst particle size leads to a good flow distribution and low pressure drop in the reactor for a catalyst having suitable mechanical strength and attrition resistance.
High activity and selectivity are required for the catalysts, which are mainly influenced by the chemical components making up the catalyst but also by the preparation method and synthesis conditions. Both factors (catalytic materials and synthesis method) play a key role for attaining a proper specific surface area and ensuring a good dispersion of the metal active phase onto the support. In addition, the balance between porosity and mechanical strength should be considered, since a highly porous catalyst leads to a high activity at the expense of decreasing the mechanical strength. 123 An active catalyst fulfilling the aforementioned requirements is not enough, but stability should also be considered. Thus, catalyst lifetime depends on its resistance to deactivation mechanisms, which in reforming reactions are mainly metal sintering and coke deposition. Therefore, although thermal stability and coke resistance are sensitive to process conditions, they are usually improved by adding a promoter. 117,124 3.1.3. Synthesis Method. This section deals with the description of the usual catalyst synthesis methods in both the laboratory and industry for the preparation of heterogeneous catalysts. Thus, the preparation of an active catalyst can be carried out by different synthesis methods, and the properties obtained are strongly affected by each step in the preparation method and the quality of the raw materials. 125 The significance of solid catalysts in large scale processes for the conversion of chemicals, fuels and pollutants is well-known. Moreover, the catalysts may be classified as follows: (i) unsupported (bulk) catalysts, (ii) supported catalysts, (iii) confined catalysts (ship-in-a-bottle catalysts), (iv) hybrid catalysts, and (v) polymerization catalysts, among others. 126 However, with the aim of simplifying these groups, some  Energy & Fuels pubs.acs.org/EF Review authors have classified them according to the preparation procedure, as follows: (i) bulk catalysts or supports, in which the catalytic active phase is generated as a new solid phase; and (ii) impregnated catalysts, in which the active phase is introduced or fixed on a pre-existing solid by a process depending on the support surface. 127 In addition, a third group was included by Hutchings and Vedrine, 128 i.e., (iii) mixedagglomerated catalysts, in which the catalyst is an agglomerated mixture of an active substance and the support. However, this type of catalyst has been less frequently used in the literature. Generally, the catalyst must have certain features for each specific process, such as suitable texture (specific surface area, pore structure, and bulk density), attrition resistance, and shape, with the final properties being highly dependent on the preparation steps ( Figure 6).
As mentioned above, the catalysts may be classified into non-supported or bulk, supported, and mixed-agglomerated ones. The different types of catalysts and their preparation methods are briefly explained below.
3.1.3.1. Bulk Catalysts. Bulk catalysts are mainly made up of an active substance, although some inert binder is frequently added to ease the forming and/or shaping operation. The addition of this binder only influences the mechanical strength and should therefore be considered when operating under certain conditions in specific reactors. Accordingly, bulk catalysts can be used without binder addition, with the catalysts being prepared by high-temperature fusion.
Among the methods for preparing bulk catalysts, the following ones are worth mentioning: (i) precipitation, which is the most widely used for its simplicity and cost; (ii) sol−gel, which allows the preparation of aerogels or xerogels; (iii) hydrothermal synthesis; and (iv) flame hydrolysis.
3.1.3.2. Supported Catalysts. Supported catalysts have been widely used in both laboratory and large scale plants, given that they have certain advantages, such as: (i) high active phase dispersion, (ii) high mechanical strength and thermal conductivity, (iii) suitable catalytic properties by means of different metal−support interactions, and (iv) suitable for preparing bifunctional catalysts.
Moreover, some of the bulk materials previously described as unsupported catalysts can also be used as supports, with the most widely used being Al 2 O 3 , SiO 2 , TiO 2 or ZrO 2 . 126 Thus, the support can be synthesized following the different steps displayed in Figure 7. However, commercial preformed supports are usually considered more convenient, efficient, and economic, 116 given that they already have the desired porous texture and mechanical strength.
Once the support with the optimum physical, chemical, and mechanical properties has been prepared or selected and stabilized with additives if required, the active phase is dispersed on the support by one of the following methods: (i) precipitation, (ii) adsorption, (iii) ion exchange, and (iv) impregnation. All these methods have certain advantages and disadvantages, and the selection of one or the other depends on the final features required for the catalyst. These methods are described below.
3.1.3.2.1. Precipitation. The aim of the precipitation method is to achieve the following reaction: metal salt solution support metal hydroxide on support base + ⎯→ ⎯⎯ (8) The choice of the salt precursor or the alkali depends on several factors, with the most influential being the possible detrimental effects on the final catalyst. 116 Thus, the support, frequently powder, is slurried with enough amount of salt solution to reach the desired metal loading. Subsequently, enough alkali solution is added to cause precipitation, and the powder is then filtered or separated and washed to remove  Energy & Fuels pubs.acs.org/EF Review alkali ions, reagent anions, and an excess deposit on the outside of the particles. After washing, the excess of moisture is removed from the pores by drying, which is not a problematic step given that the active component is firmly anchored to the surface. The calcination step is then carried out to decompose the metal salt precursors into their oxidized form. Moreover, the processes involved in the precipitation are as follows: (i) precipitation from bulk solutions or from pore fluids and (ii) interaction with the support surface. The precipitation takes place in three steps, supersaturation, nucleation, and growth, and depends on concentration, temperature, pH, and ripening. 125,127 This method is usually preferred when loadings higher than 10−20% are required.
3.1.3.2.2. Adsorption. In this process, the support is exposed to metal salt solutions in which equilibrium amounts of salt ions are adsorbed following adsorption isotherms. Depending on the properties of the support surface, the adsorption can be cationic or anionic. This process is greatly influenced by the adsorption conditions, especially by the pH of the solution. 126 The equilibrium reactions involved in the ionic adsorption (except for zeolites) are as follows: Moreover, the support (as powder or particle form) is dehydrated and soaked into the solution for a suitable period of time, leading to a uniform active phase distribution, with the pores being properly filled during the soaking time. Likewise, drying and calcination steps are required.
Therefore, although this is an excellent method when low metal loadings are required, the amount of metal loading allowed until the saturation point is generally small (e.g., loadings of 2−3 wt % of Ni on Al 2 O 3 are obtained). 116 3.1.3.2.3. Ion Exchange. The catalyst synthesis by ion exchange is very similar to ionic adsorption. However, it involves the exchange of an ion by electrostatic interaction with the support surface by another ion species. 125,127 The support containing ions A is plunged into an excess volume (compared to the pore volume) of a solution containing ions B, which gradually penetrate into the pores of the support, while ions A pass into the solution until equilibrium is reached, according to the following reaction: 130 with s and z being the solution and the support, respectively. This is an adequate method for removing harmful agents and adding promoters and is considered as a promising alternative for the modification of catalytic materials. 116 3.1.3.2.4. Impregnation. Impregnation is the simplest and most direct method to deposit the active phase onto the support. It implies the contact between a certain volume of active phase metal precursor and the solid support, with the solvent being removed in a subsequent drying step. This methodology allows improving the metal dispersion on the support, although the selection of the suitable metal salt precursor also plays a key role in the final dispersion. 131 Depending on the volume of the solution, the impregnation may be classified as follows: (i) wet impregnation, in which an excess of solution is used, and (ii) incipient wetness impregnation, in which the volume of a solution with appropriate concentration is equal or slightly lower than the pore volume of the support.
In this method, the solubility of the precursor in the solution limits the maximum metal loading, although another impregnation step may be conducted after the drying or calcination step when higher metal loadings are required.
Furthermore, the drying process, which is necessary to crystallize the salt on the pore surface, may lead to irregular and uneven concentration, and this step should therefore be slow enough to form uniform deposits. Moreover, the calcination step, in which the salt precursor is converted into its oxide form, also plays a key role in the final metal dispersion. 116 3.1.3.3. Physical Mixing. Mixed agglomerated catalysts are prepared by this method. These catalysts are prepared by physically mixing the active substances with a powdered support or support precursors in a ball mill. The final mixture is then agglomerated and activated.
The selection of a given preparation method and its synthesis steps condition the catalyst properties and therefore its overall performance in the pyrolysis-reforming step. In order to analyze their influence in detail, these studies should be performed under similar operating conditions. Although few research studies have investigated the influence the synthesis method and conditions have on the reforming step, the main studies are summarized and discussed in section 4.
3.2. Mechanism of Catalyst Deactivation. Over the last decades, the mechanisms and causes of catalyst activity decay have been widely analyzed in order to step further into the knowledge of catalysis, thereby establishing the basis for modeling deactivation processes, improving catalyst design, and preventing or slowing the degradation of the catalyst. 116,123,129,132,133 The economic feasibility of a given catalytic process largely depends on the activity and catalyst lifetime, and although the catalyst activity decay with time is unavoidable, certain strategies may be developed in order to improve its performance during the reaction step. Thus, possible causes of catalyst deactivation and their mechanisms are described below.
The causes of catalyst deactivation can be ascribed to three main factors: (i) mechanical, (ii) thermal, and (iii) chemical. However, deactivation is not the consequence of a single mechanism, but usually their combination is responsible for the catalyst degradation. Table 2 shows the main causes of catalyst activity decay. As observed, the mechanisms may be classified as follows: (i) poisoning, (ii) coking and fouling, (iii) sintering, (iv) component volatilization, (v) inactive com- Poisoning is the strong chemisorption of reactants, products, or impurities on sites otherwise available for catalysis. 134 The fact that a compound acts as a poison depends on the nature of the process and the adsorption strength of this compound to physically block the active sites. Besides, the poison can lead to the modification of the catalyst structure or formation of a compound.
Moreover, the decrease in catalytic activity as a consequence of poisoning can involve the following aspects: (i) physical blockage of the adsorption/reaction sites; (ii) electronic modification of the nearest metal atoms, thus hindering their capability to adsorb and dissociate the reactants; (iii) dramatic changes in the catalytic properties by rearranging the catalyst surface; (iv) blockage of reactant access; and (v) hindrance of surface diffusion of adsorbed reactants. 123 Most of the compounds considered as poisons are contained in the feed in small quantities and deactivate the catalyst following a different mechanism from the main reaction. However, poisons can also be generated either in parallel or in a series of reactions, resulting in catalyst activity decay. 116 The irreversibility of poisoning and catalyst regenerability are greatly influenced by the type of poison, the catalyst, and the process. Thus, the main poison of reforming catalysts is sulfur, which may be in the feed as an organic sulfur compound, at concentrations of up to 1500 ppm. 135 However, the catalyst tolerance against poisons depends on the materials that make up the catalyst, with the best performance against poisoning deactivation having been reported for noble metal based catalysts. 136,137 3.2.2. Fouling/Coking Deposition. The physical formation of species deposits from the fluid phase on the catalyst surface is the common mechanism of deactivation known as coking or fouling, although this latter term can also be associated with other types of deposition from the reactor material. 138 Fouling leads to a loss of activity due to the blockage of active sites and/or pores and may involve the breakup of catalyst particles and reactor plugging. 123 The consequences of carbon deposition are shown in Figure  8 and may be summarized as follows: (i) the coke deposited is strongly chemisorbed to the metal particle or physically adsorbed in a multilayer, hindering the access of reactants to the active sites; (ii) full deactivation of the metal particle due to its complete encapsulation by coke deposition; (iii) blockage of the support pores, hindering the access of the reactants to the active sites; and (iv) fracture of the support material due to the formation and growth of filaments, which may lead to reactor plugging. 134 The mechanisms of coke formation and its intrinsic nature are greatly influenced by the feed, the type of process, and its operating conditions. Moreover, the nature of this coke may evolve with time on stream. Thus, some authors have reported that the main coke precursors are aromatics and/or olefins, which by reactions of dehydrogenation, condensation, and oligomerization, lead to coke formation. 138 Several authors have studied the mechanisms of carbon deposition and coke formation on metal catalysts from carbon monoxide and hydrocarbons in steam reforming reactions. 133,134,139 Accordingly, the mechanisms proposed by Bartholomew 139 are presented in Figure 9.
As observed, the nature and morphology of the coke formed differs depending on the different reactions described above, with its formation and nature depending on the reaction conditions (Table 3).
Furthermore, as mentioned above, the mechanism of catalyst deactivation and coke formation is greatly influenced by the feed. Thus, in the steam reforming of the volatiles derived from the pyrolysis of biomass and plastic wastes, the cracking of oxygenates (eq 1) and hydrocarbons, HCs (eq 2), respectively, should be taken into account.
Accordingly, Ochoa et al. 141 analyzed the role oxygenates play in the catalyst deactivation in the steam reforming of biooil and distinguished between two types of coke: (i) encapsulating coke, with mainly aliphatic nature and oxygenates in its composition, since its precursors are mostly bio-oil  oxygenates adsorbed on the metallic sites, and (ii) filamentous coke, with a higher ordering and carbonization in its structure, i.e., more polyaromatic and olefinic compounds, whose possible precursors are CO and CH 4 , largely contributing to a more severe catalyst deactivation. Likewise, the same authors evaluated the deactivation mechanisms of a Ni supported catalyst in the pyrolysis-steam reforming using HDPE as raw material. 142 They reported that the mechanisms of coke formation follow two consecutive steps: (i) the formation of amorphous and encapsulating coke by condensation of promoters, and (ii) carbonization of adsorbed coke promoters leading to the formation of filamentous coke.
Thus, the formation of carbon deposits on metal oxides involves cracking reactions of the coke precursors, which are catalyzed by the acid sites, and may evolve to a more condensed coke by dehydrogenation and cyclization reactions. 123 Several characterization techniques have been used in the literature in order to analyze catalyst and coke properties during the catalyst deactivation in the reforming step. 140 Thus, Ochoa et al. 143 conducted a detailed analysis of the evolution of catalyst deactivation in the biomass pyrolysis and in-line steam reforming. Thus, they determined coke content by TG-MS/TPO and analyzed its morphology by scanning and transmission electron microscopy (SEM and TEM), reporting its encapsulating nature as the main cause of catalyst decay.
SEM and TEM images have also been used in the pyrolysisreforming of wastes plastics for analyzing the nature of the coke deposited. Thus, Barbarias et al. 46 reported the presence of filamentous carbon in the reforming of HDPE pyrolysis volatiles, although no filamentous carbon was observed in the coke deposited when PS was used as the feedstock.
3.2.3. Sintering. The agglomeration and growth of the catalyst metal crystallites is the process known as sintering, which is influenced by several factors, such as temperature and the reaction medium, among others. 135 Thus, high temperatures, typically above 500°C, and the presence of steam promote the sintering phenomenon, leading to irreversible or difficult to reverse deactivation of the catalyst. Besides, the metal type and its dispersion on the support, the presence of promoters or impurities on the surface, and the textural properties of the support greatly influence the sintering rates. 134 The influence of temperature on metal sintering has been correlated as follows: T T Tamman temperature: where T m is the melting point temperature in K. Accordingly, atomic particles on the surface, especially surface atoms weakly bound to defect sites, become mobile at temperatures above the Huẗtig temperature, and bulk metal atoms acquire enough thermal energy to migrate within the crystallite at temperatures above the Tamman one. 116,138 The mechanisms whereby metal particles increase in size may occur as illustrated in Figure 10, and they are summarized as follows: (i) atomic migration, which implies metal atom detachment from crystallites and their migration to the support surface, where they are finally captured by larger metal particles; (ii) migration of entire crystallites, forming larger particles by collision and coalescence; and (iii) spreading and splitting. 123,134 The catalyst activity decay by sintering may lead to: (i) crystallite growth on the catalytic phase, decreasing the catalytic surface area, and (ii) collapse of the support and/or support pores, involving a loss of catalytic surface area and loss of support area.
Moreover, the support may also be subject to sintering involving the following processes: (i) surface diffusion, (ii) solid-state diffusion, (iii) grain boundary diffusion, (iv) evaporation/condensation of volatile atoms or molecules, and (v) phase transformations. 123 These two latter processes are explained in the following sections.
3.2.4. Component Volatilization. The catalyst deactivation by direct volatilization of catalytic metals hardly plays a significant role in the catalyst activity decay, given that extremely high temperatures are required (normally, up to 1000°C) for this process. On the contrary, the formation of volatile compounds, such as metal carbonyls, oxides, sulfides, and halides, which lead to a loss of catalytic material, can take place at moderate temperatures. 123 Some examples of catalyst deactivation by component volatilization are found in the regeneration of molybdenumcontaining catalysts in the hydrotreating process when temperatures of up to 800°C are attained (Mo vaporizes at this temperature). It also applies in the naphtha steam reforming, although coke formation from heavier hydrocarbons is controlled in this process with potassium. Nevertheless, volatile KOH compound may be formed in the presence of steam, which accelerates coke formation. 138 3.2.5. Inactive Compound Formation. A further chemical route leading to catalyst deactivation is the reaction between the vapor phase and the catalyst surface to produce inactive phases instead of strongly adsorbed species, which lead to a decrease in catalyst activity. Moreover, these reactions are usually promoted at high temperatures and can be easily detected by the X-ray diffraction (XRD) technique.
Accordingly, the reaction between the active phase of nickel with common supports, such as Al 2 O 3 or SiO 2 to form nickel aluminates or silicates, the formation of oxides with steam when cobalt or iron are used, 116 or the oxidation, sulfidation,  Energy & Fuels pubs.acs.org/EF Review or carbidation of the metal active phase involve the loss of catalytic activity. 134 3.2.6. Phase Transformation. Phase transformation is the consequence of the thermal degradation of the support and adversely affects physical or chemical catalyst properties. Thus, the support can be induced by thermal treatment to attain a phase modification, as in the case of the Al 2 O 3 support, which can change from γ-Al 2 O 3 to a transitional phase with lower specific surface area (δ-Al 2 O 3 , θ-Al 2 O 3 ...) until reaching the most stable phase (α-Al 2 O 3 ). Consequently, the supported catalyst surface area decreases, and therefore activity drops. Several authors have called this process support sintering. 123 Moreover, the addition of small amounts of silica may contribute to control this phase change. 131 3.2.7. Particle Attrition. The catalyst mechanical failure can arise from (i) granule, pellet, or monolithic crush when loading; (ii) attrition or reduction/breakup of the catalyst particle size forming fine particles, especially in fluidized bed reactors; and (iii) erosion of the catalyst particles when high fluid velocities are used. 134 As a result, reactor plugging or channeling may occur, leading to an increase in pressure drop and non-uniform bed performance. Thus, thermal and coking effects promote other catalyst deactivation mechanisms.
In order to prevent catalyst attrition, the following alternatives have been proposed in the literature: (i) increasing strength by advanced catalyst synthesis methods, (ii) adding binders to improve strength and toughness, (iii) coating aggregates with porous but very strong materials, and (iv) inducing compressive catalyst stress by chemical or thermal tempering. 123,134

APPLICATION OF METALLIC CATALYSTS IN THE PYROLYSIS-REFORMING PROCESS
In recent decades, the development and study of a wide range of different metal catalysts has been approached in the literature for the reforming of oxygenates derived from biomass pyrolysis. However, most of these studies have been conducted by using model compounds 144−146 or the aqueous fraction of the bio-oil. 34,36,107 Similarly, the valorization of different waste plastics, namely, polyolefins (HDPE, LDPE, and PP), polystyrene (PS), polyethylene terephtalate (PET), and real world plastics recovered from municipal solid waste (MSW), or even the joint co-feeding of biomass and plastics mixtures, has attracted increasing attention in the two-step pyrolysis-reforming process in recent years due to the great versatility of this strategy. 54,56,147,148 Furthermore, special attention has been paid to the design of a suitable reforming catalyst for the production of H 2 . This section reviews the catalysts used in the literature for the two-step processing of biomass, plastic wastes, and biomass/plastics mixtures, in which pyrolysis is carried out in the first step and in-line steam reforming of the volatiles in the second one. Accordingly, and with the aim of easing the comparison of the different catalysts, the following sections have been considered: (i) Ni/Al 2 O 3 catalysts, which are the most widely used because Al 2 O 3 support provides a high specific surface area, eases Ni dispersion, and confers mechanical strength and stability upon the catalyst 149 (section 4.1); (ii) Ni supported on other materials, either on several metal oxides, such as SiO 2 , MgO, ZrO 2 , or TiO 2 or on other types of supports, such as chars obtained from the fast pyrolysis of different biomasses, Figure 11. Main elements studied in the catalyst design for H 2 production from the pyrolysis and in-line steam reforming process using biomass, plastic wastes, and biomass/plastic mixtures as feedstock. Adapted from ref 12 Figure 11 shows the main components used as catalytic materials, namely, the active phase, promoter, or support, in the catalyst design for H 2 production from the steam reforming of the volatiles derived from the pyrolysis of biomass, plastic wastes, and their mixtures. Thus, the suitability of these elements has been classified according to a color code, with green, yellow, and red being good, moderate, and poor performances, respectively.
4.1. Ni/Al 2 O 3 Catalysts. Ni/Al 2 O 3 catalysts (both commercial and prepared) have been widely studied in the literature on the reforming reactions. 50,110,150,151 Thus, Ni has proven to be highly active on the reforming of the volatiles derived from biomass and plastic wastes pyrolysis as well as on methane steam reforming, WGS reaction, and ammonia decomposition. 11,152,153 Moreover, its moderate cost made this metal an interesting option compared to other active phases, such as noble metals (Rh, Ru, Pd, Pt...). Similarly, Al 2 O 3 is the most widely used support in reforming reactions, and its suitable properties (e.g., high specific surface area) iwi, Incipient wetness impregnation; wi, wetness impregnation. b H 2 production defined as g H2 /100 g biomass, daf (dry and ash free). c T P = pyrolysis temperature; T R = reforming temperature.  Regarding the crystalline phases of the Al 2 O 3 support, significant differences in the catalyst properties can be observed. Thus, the transition from γ-Al 2 O 3 to α-Al 2 O 3 , which is usually formed at extremely high temperatures (>1200°C ) through intermediate crystal phases such as θand δ-Al 2 O 3 , leads to a remarkable decrease in the specific surface area of the support and therefore to a poorer active phase dispersion on the catalyst. Furthermore, the metal−support interaction is also influenced by the type of Al 2 O 3 selected, and the reducibility of the metal catalyst is therefore modified, i.e., weaker interaction between metal and α-Al 2 O 3 is obtained in comparison with γ-Al 2 O 3 , which is more likely to form a spinel phase. 154 Besides, α-Al 2 O 3 has proven to be attrition resistant when operating in fluidized bed reactors. 155 Moreover, the acidity of the support is related to the crystalline phase of the Al 2 O 3 support. Thus, the acid sites of γ-Al 2 O 3 promote carbon deposition on the catalyst surface, which is the main cause of catalyst deactivation in reforming reactions.
As mentioned before, the operating conditions, type of reactor configuration, and the catalyst selected in the pyrolysisreforming strategy greatly influence the overall H 2 production. Accordingly, Tables 4−6 summarize the main results reported in the literature for Ni/Al 2 O 3 catalyst for the pyrolysis-steam reforming process of biomass, plastic wastes, and biomass/ plastics mixtures. In regards to the use of biomass as raw material (Table 4), an extensive use of commercial catalysts has been reported in the literature. Xiao et al. 61 carried out a parametric study in a two-step fluidized bed/fixed bed configuration, using wood chips and pig manure compost as feedstock and Ni/Al 2 O 3 and Ni/BCC (brown coal char) as catalysts. Although Ni/Al 2 O 3 undergoes gradual catalyst deactivation, mainly by carbon deposition, wood chip biomass leads to higher H 2 productions compared to pig manure compost (7.2 and 5.0 wt %, respectively, on dry and ash free, daf).
Moreover, Arregi et al. 50 studied H 2 production by continuous fast pyrolysis (500°C) of pine wood sawdust in a conical spouted bed reactor (CSBR) followed by in-line steam reforming of the pyrolysis vapors in a fluidized bed reactor. A commercial Ni/Al 2 O 3 catalyst was highly active, with a H 2 production of 11.2 g H2 100 g biomass −1 at a reforming temperature of 600°C, a space time of 20 g cat min g volatiles −1 , and a S/B ratio of 4. A similar H 2 production was obtained by Fernandez et al. 98 under the same operating conditions but using a CSBR and a fixed bed reactor for the pyrolysis and reforming steps, respectively. Despite this latter configuration led to lower catalyst deactivation, especially when low space times were used, the advantages of fluidized beds should be taken into account for further scalability of this strategy.
Furthermore, Cao et al. 156 evaluated the performance of a commercial Ni/Al 2 O 3 catalyst on the low-temperature catalytic reforming of volatiles and nitrogen compounds from sewage sludge (SS) pyrolysis in a two-step fixed bed reactor. They observed that although the results significantly depend on the operating conditions, a H 2 -rich gas stream with a content of 68.0 vol % and a H 2 production of 11.6 wt % (daf) were obtained at 650°C.
Besides, the synthesis of the Ni/Al 2 O 3 catalysts has been analyzed in detail in the literature. Thus, Akubo et al. 157 prepared a 10 wt % Ni based Al 2 O 3 catalyst by incipient wetness impregnation and investigated the pyrolysis-catalytic steam reforming of six agricultural biomass waste samples (rice husk, coconut shell, sugar cane bagasse, palm kernel shell, cotton stalk, and wheat straw) in a two-step fixed bed reactor. The results for the different types of biomasses revealed that H 2 production ranged from 3.3 wt % for wheat straw to 5.1 wt % for palm shell kernel. Efika et al. 102 analyzed the performance of different prepared catalysts (Ni/Al 2 O 3 , Ni/ CeO 2 -Al 2 O 3 and Ni/SiO 2 ) in a two-step continuous screw-kiln reactor, in which biomass pyrolysis (500°C) and the subsequent catalytic steam reforming of the pyrolysis oils and gases was conducted (760°C). They found that Ni/Al 2 O 3 catalyst was the most effective one for H 2 production, i.e., it allowed obtaining the highest H 2 concentration of 44.4 vol %. Moreover, they characterized the coke deposited on this deactivated catalyst by SEM images and reported the presence of filamentous carbon. Santamaria et al. 51,100 obtained a H 2 production of 10.2 wt % in a CSBR-fluidized bed reactor configuration with pinewood sawdust being continuously fed at a rate of 0.75 g min −1 and using a 10Ni/Al 2 O 3 catalyst prepared by the wet impregnation method.
Likewise, Ni/Al 2 O 3 has been widely selected for waste plastic valorization by pyrolysis-reforming runs due to the aforementioned advantages, i.e., suitable activity, moderate cost, high specific surface area, good metal dispersion, and appropriate mechanical strength. Thus, Table 5 shows the main studies dealing with the pyrolysis-reforming of different waste plastics using Ni/Al 2 O 3 as a reforming catalyst. The initial studies reported in the literature were performed by Czernik and French, 54 who developed a continuous process consisting of two fluidized bed reactors. A commercial Ni reforming catalyst (C11-NK) used in industry for naphtha reforming was selected and PP was continuously fed at a rate of 1 g min −1 . Operating at pyrolysis and reforming temperatures of 650 and 800°C, respectively, these authors reported a H 2 production of 34 wt %. A similar H 2 production was reported by Erkiaga et al., 97 who conducted the continuous HDPE pyrolysis and in-line steam reforming in a bench scale plant consisting of a conical spouted bed and a fixed bed reactor for the pyrolysis and reforming steps, respectively, and using a commercial Ni based catalyst (G90-LDP). Under the optimum operating conditions, i.e., pyrolysis and reforming temperatures of 500 and 700°C, respectively, and a S/C ratio of 3.1, they reported a H 2 production of 34.5 wt %. However, due to the operational problems related to coke formation, Barbarias et al. 44 conducted a parametric study in the same experimental unit but using a fluidized bed reactor instead of a fixed one. Thus, they reported a H 2 production of 38.1 wt % at a reforming temperature of 700°C and with a S/C ratio of 3.89. The good performance of this technology was further demonstrated in a later study on the valorization of different plastic wastes (PP, PET, and PS) and their mixtures. 56 In this latter study, the highest H 2 production was obtained with polyolefin plastics (37.3 and 34.8 wt % for HDPE and PP, respectively). However, a considerably lower H 2 production was reported with PET (18.2 wt %), which is evidence of the high influence of the volatile composition and therefore of the plastic type selected.
A two-step fixed bed configuration operating in batch mode has been widely used in the literature. Thus, the group headed by Prof. Williams investigated the effect of operating conditions, type of plastic, and catalyst synthesis method using a wide range of Ni-based catalysts. 63 150 in the pyrolysis−reforming of PP at a pyrolysis temperature of 500°C and a reforming one of 800°C. They reported an increase in the potential H 2 production from 48.8 to 57.7 wt % when the Ni/Al molar ratio was increased from 1:4 to 1:1, which corresponds to a rise in the H 2 production from 20.9 to 24.8 g H 2 per 100 g PP . Yao et al. 63 investigated the influence the synthesis method has on the physicochemical properties and therefore on the catalyst activity for the production of H 2 from pyrolysis-steam reforming of waste plastics (HDPE, PP, and PS). Accordingly,  co-precipitation, impregnation, and sol−gel. These authors observed that the catalyst prepared by the sol−gel method had higher specific surface area and fine nickel particle size with uniform dispersion, which led to higher H 2 productions (12.1, 13.4, and 12.5 wt % for HDPE, PP, and PS samples, respectively). Figure 12 compares the main results obtained in the literature for the steam reforming of the volatiles derived from the pyrolysis of biomass (Figure 12a) and different plastic wastes (Figure 12b) using Ni/Al 2 O 3 as the reforming catalyst. iwi, Incipient wetness impregnation; wi, wet impregnation; cp, co-precipitation; sg, sol−gel. b T P = Pyrolysis temperature; T R = reforming temperature; T G = gasification temperature.
Energy & Fuels pubs.acs.org/EF Review As observed in Figure 12, this catalyst shows a great potential for the production of H 2 in the two-step pyrolysis reforming process. Thus, a H 2 production in the range of 5.9− 11.6 wt % is obtained when biomass is used as the raw material. 50,51,98,156 Similarly, this catalyst shows an excellent performance when different types of plastic waste materials are used, being able to reach a H 2 production of around 34 wt % when PP and PE are selected. 54,56,97 However, the acid nature of the Al 2 O 3 support promotes the coke formation on the catalyst, leading to a fast catalyst deactivation in the reforming step. In this regard, studies dealing with the main mechanisms of catalyst deactivation and coke formation are attracting increasing attention in the literature 140,142,143 in order to take a step further in the scalability of the process.
A comparison of the different results reported in the literature studies is not straightforward, since they have been conducted under different reaction conditions. Thus, the reactor configuration, operating conditions (pyrolysis and reforming temperatures, steam/feedstock ratio, space time) and the catalysts synthesis conditions (preparation method, Ni loading, calcination temperature) should be taken into account for the final optimization of this process.
Moreover, the main differences observed in Figure 12 among the research studies are ascribed to the operation in continuous/discontinuous mode. Thus, the preliminary character of some of these studies conducted in laboratoryscale batch reactors led to lower H 2 productions. In this regard, a great effort has been made for the implementation of a continuous feeding system in the two-step pyrolysis-reforming process in order to advance in the scaling up of this strategy.
The joint valorization of biomass and waste plastics on a Ni/ Al 2 O 3 catalyst in the two-step pyrolysis-reforming process has also been assessed in the literature. Thus, the co-feeding of plastics along with biomass can provide the following advantages to the process: 57,147 (i) improvement of the overall H 2 production by incorporating higher amount of H 2 in the feed; (ii) limitations derived of seasonal availability of biomass are avoided; (iii) reduction of environmental problems related to the management of plastic wastes; and (iv) attenuation of catalyst deactivation in the reforming step. Accordingly, Table  6 summarizes the main studies concerning the pyrolysis and inline catalytic reforming of biomass and plastic waste mixtures.
As observed, Arregi et al. 57 analyzed the continuous pyrolysis and in-line catalytic reforming of pine wood waste and HDPE mixtures on a commercial Ni/Al 2 O 3 catalyst (G90-LDP). The pyrolysis step was conducted in a CSBR at 500°C, whereas the reforming one was performed in a fluidized bed reactor at a reforming temperature of 700°C. In the mentioned study, the influence biomass/plastic (B/P) feeding ratio (25, 50, and 75 wt % HDPE) has on the product yields and catalyst deactivation was evaluated. As a result, these authors reported a marked attenuation of catalyst deactivation as well as a linear improvement of H 2 production when HDPE feed was increased (from 17.5 wt % when a B/P weight ratio of 75/25 was used to 31.4 wt % when a B/P ratio of 25/75 was used). They also reported significant differences in the coke structure and nature (analyzed by TEM images), which were ascribed to the different composition of the volatiles fed into the reforming step. Moreover, a two-stage fixed bed reactor was used by Alvarez et al. 147 in the co-pyrolysis-reforming of pine wood sawdust and different plastics (HDPE, PP, PS, and a real plastic mixture (RP)), with a biomass/plastic weight ratio of 80/20. The highest H 2 production was observed when PP was co-fed with biomass (5.5 wt %), followed by HDPE in the feed (5.1 wt %).
More recently, Xu et al. 148 analyzed the influence biomass/ plastic ratio (in this case, rice husk (RH) and PE were selected as feedstock) has on the quality of the gaseous products in the catalytic steam reforming of the volatiles derived from copyrolysis of biomass and plastics. They reported a synergistic effect on gas and tar yields when PE was co-fed with rice husk, especially in the runs conducted using a RH/PE ratio of 50:50, wherein the H 2 production reached 4.4 wt %.
4.2. Ni Supported Catalysts. The support significantly influences the catalyst activity and stability during the reaction. Thus, high specific surface area, adequate pore distribution and mechanical strength, and good thermal stability are required in order to obtain a suitable catalyst performance. Moreover, the acidity/basicity of the support may promote/hinder the deactivation by carbon deposition. Accordingly, metal oxide supports have been extensively analyzed in the steam reforming of oxygenates, using the aqueous fraction of biooil 107,167 or model compounds 145,168 as feedstocks. Similarly, studies have been conducted in a two-step process of biomass pyrolysis-steam reforming and they are summarized in Table 7. Hence, Miyazawa et al. 89 carried out an activity test of different Ni supported catalysts (Ni/Al 2 O 3 , Ni/ZrO 2 , Ni/TiO 2 , Ni/ CeO 2 and Ni/MgO) in the steam reforming of the tar derived from the pyrolysis of cedar wood and observed that tar conversion at 650°C decreased as follows: Ni/Al 2 O 3 > Ni/ ZrO 2 > Ni/TiO 2 > Ni/CeO 2 > Ni/MgO. Besides, they concluded that the support only has influence on Ni dispersion, whereas Ni metal is responsible for tar conversion.
Efika et al. 102 investigated the production of synthesis gas in a two-step continuous screw-kiln reactor using wood pellets as a biomass feedstock. Thus, they analyzed the influence of the synthesis method for a Ni/SiO 2 catalyst, which was prepared by incipient wetness impregnation and by a sol−gel method. The results showed that the catalyst prepared by the sol−gel method had a higher specific surface area (765 m 2 g −1 ) and so led to a higher gas yield (54 wt %), whereas the gas yield obtained on the Ni/SiO 2 catalyst prepared by the incipient wetness method was lower (49.8 wt %), which is associated with its lower specific surface area of 136 m 2 g −1 . The influence the support has on the activity and stability of Ni catalysts used in the reforming of pinewood sawdust fast pyrolysis volatiles was assessed by Santamaria et al., 51,100 with the selected supports being as follows: Al 2 O 3 , SiO 2 , MgO, TiO 2 and ZrO 2 . They observed that despite Ni/Al 2 O 3 , Ni/ZrO 2 , and Ni/MgO catalysts were overall more active and stable over time on stream, the Ni/Al 2 O 3 catalyst led to a remarkable coke deposition. However, a lower deactivation rate was evidenced for Ni supported on MgO and ZrO 2 due to the properties conferred by these supports upon the catalysts, as are basicity and capacity for gasifying the coke precursors, respectively. The highest H 2 production was achieved with Ni/ZrO 2 catalyst (10.7 wt %), followed by Ni/Al 2 O 3 (10.2 wt %) and Ni/MgO (9.1 wt %).
Moreover, Ma et al. 169 investigated H 2 production in a novel process integrating biomass pyrolysis, gas−solid simultaneous gasification, and the catalytic reforming process. Timber wood sawdust was the biomass selected, and a commercial Ni/MgO catalyst was used. H 2 production significantly increased from 4.4 (using the two-step pyrolysis-catalytic reforming process) to 7.6 wt %. Likewise, Chen et al. 170 evaluated the production of H 2 -rich gas from cotton stalks on a commercial Ni/MgO Energy & Fuels pubs.acs.org/EF Review catalyst in the mentioned integrated process, and H 2 production increased from 3.9 to 6.5 wt %, since the condensable gas and char is more efficiently used through simultaneous conversion. Although the use of metal oxides supports have been scarcely investigated in the literature dealing with pyrolysisreforming of plastics, details about the main studies have been summarized in Table 7. Thus, the suitability of the two-step pyrolysis-reforming process for H 2 production from polypropylene (PP) was analyzed by Wu and Williams, 164 who synthesized several nickel based catalysts. Among them, different metal oxides were selected: Al 2 O 3 , MgO, and CeO 2 . However, these Ni supported catalysts showed lower activity for H 2 production compared to other supported or promoted catalysts tested (Ni/ZSM-5, Ni/CeO 2 /Al 2 O 3 and NiMgAl).
Recently, several Ni supported catalysts were analyzed in the steam reforming of LDPE pyrolysis volatiles for syngas production by Huang et al. 171 In this study, the great influence the support selected has on catalytic activity, selectivity, and coke formation was evidenced, with Ni/CeO 2 catalyst being Furthermore, the use of alternative supports, such as dolomite, char, or active carbon, are gaining increasing attention due to their lower cost. Moreover, the concentration of alkaline metals, such as K and Ca, in the biochar produced from biomass pyrolysis may promote the decomposition of hydrocarbons and the water gas shift reaction during the reforming process, 173 making this material suitable as catalyst support. The suitable properties of biochar and coal char, i.e., high specific surface area, optimized pore volume, excellent thermal stability, and abundant surface functional groups, have promoted their use in the reforming of biomass tar. Recently, Ren et al. 174 discussed in detail the preparation, modification, and characteristics of biochar and coal char as well as their application in biomass tar reforming. Table 8 compares the studies reported in the literature for pyrolysis and in-line reforming of biomass. Thus, some studies have been conducted with continuous biomass feed using a fluidized or entrained flow reactor in the pyrolysis step and a fixed bed reactor in the reforming one. Thus, Xiao et al. 61 compared a commercial Ni/Al 2 O 3 and a Ni/BCC (brown coal char) catalyst in the continuous pyrolysis-reforming of two different feedstocks (wood chips and pig manure compost). They obtained lower tar content, higher coking resistance, and higher H 2 production (9.3 wt % daf) when Ni/BCC was used. Liu et al. 160 investigated the syngas production in the catalytic reforming of pyrolysis volatiles from pine sawdust with a continuous feed of 4 g min −1 . Five innovative slag carriers were selected as catalytic supports ((magnesium slag (MS), steel slag (SS), blast furnace slag (BFS), pyrite cinder (PyC), and calcium silicate slag (CSS)) and Ni as metal active phase. Under the operating conditions selected (pyrolysis temperature of 900°C and reforming one of 800°C), the H 2 production decreased as follows: Ni/MS > Ni/γ-Al 2 O 3 > Ni/ SS > Ni/BFS > Ni/CSS > Ni/PbyC. Nevertheless, most of these studies have been performed in batch operation using a two-step fixed bed reactor. Thus, Kannari et al. 159 used chicken droppings (CD) and chicken dropping ash (CDA) as catalytic support on a Ni based catalyst for decomposing tar derived from Japanese cypress pyrolysis. The results revealed that Ni/CDA leads to higher H 2 production than the commercial Ni/Al 2 O 3 catalyst (5.8 wt % vs 5.0 wt % daf) and a lower amount of carbon deposition. Ren et al. 175 synthesized a layered Ni/modified lignite char (Ni/ HSL) and analyzed its performance in a two stage fixed bed reactor using corncob as the raw material. They reported optimum activity of Ni/HSL compared to Ni/Al 2 O 3 , with a H 2 production of 10.0 wt % daf. The suitability of this lignite char support was also evident in previous studies conducted by this research group. 176−178 Moreover, Waheed et al. 179 investigated the conditions for H 2 production from rice husk in a two-step pyrolysis/catalytic reforming process containing Ni supported on dolomite catalyst. The highest H 2 production was obtained when the temperature was increased to 1050°C in the reforming step (6.1 wt %). Besides, different biomasses (rice husk, sugar cane bagasse, and wheat straw) were analyzed by these authors 180 on a 10 wt % Ni/dolomite catalyst, obtaining the highest H 2 production when rice husk was used as the feedstock (5.1 wt %).
The use of biochar as catalyst support is gaining increasing interest in the literature. Yao et al. 173 used three types of biochar (obtained from the pyrolysis of wheat straw (WC), rice husk (RC), and cotton stalk (CC)) as a support of a 15 wt % Ni catalyst in the pyrolysis/reforming of wheat straw. They concluded that biochar is a promising catalytic support for this process, with the cotton-char supported Ni leading to the best results in terms of H 2 concentration and production (64.0 vol % and 9.2 wt %, respectively). Moreover, Gai et al. 181 developed a mild one-step hydrothermal synthesis method for the production of well-dispersed metallic nickel nanoparticles on hydrothermal carbons derived from waste biomass (hydrochar). Under the optimum synthesis conditions (calcination temperature of 700°C), this catalyst led to attenuation of coke deposition and resistance of nickel agglomeration during the catalytic reforming process. Thus, the formation of hydrogen-rich syngas was promoted, obtaining a H 2 production of 10.9 wt %. Energy & Fuels pubs.acs.org/EF Review Figure 13 compares the main supports used in the steam reforming of biomass pyrolysis volatiles. Among the different supports selected, metal oxides alternative to Al 2 O 3 have been extensively analyzed (Figure 13a) in order to attenuate the fast catalyst deactivation by coke deposition promoted by the acid properties of the Al 2 O 3 support. In this regard, supports with basic properties, namely SiO 2 , MgO, ZrO 2 , or CeO 2 are known to delay coke formation. 100 It is to note that the results presented in Figure 13 only account for their initial performance, and catalyst stability should therefore be monitored throughout the reaction for the selection of the more suitable catalyst. However, this aspect has been scarcely studied in the literature.
As mentioned before, the metal dispersion on the catalytic support plays a key role in the initial catalyst activity. Besides, other factors such as adequate metal−support interactions, catalyst reducibility, mechanical strength and thermal stability greatly influence the performance of the reforming catalyst.
As observed in Figure 13a, supported catalysts are in general active for the reforming of the biomass pyrolysis volatiles. In the case of SiO 2 , although this support has high specific surface area and allows for a high Ni dispersion, its fine porous structure hinders the accessibility of oxygenate bulky molecules, leading to a considerable reduction in activity. 51 The good performance of a commercial MgO supported catalyst was evidenced in a three-step process (pyrolysisgasification-reforming) by Ma et al. 169 and Chen et al., 170 as they obtained H 2 productions in the 6.5−7.6 wt % range. Santamaria et al. 100 reported that, in spite of the poor porous structure of the MgO support (specific surface area of 1 m 2 g −1 ) and its low reducibility (strong metal supportinteraction), a suitable activity and stability was observed (with a H 2 production of 9.0 wt %) due to the external location of Ni particles, which enhanced the accessibility of the bulky oxygenate molecules. Nevertheless, Miyazawa et al., 89 ascribed the poor activity of the MgO supported catalyst to the low metal dispersion obtained when this catalyst was tested.
The suitable features of ZrO 2 (redox properties, mechanical strength and thermal stability) and TiO 2 supports (good reducibility and attenuation of coke formation) have promoted their use in the pyrolysis-reforming strategy. Figure 13b shows the main biomass pyrolysis-reforming studies for other supports, such as biomass derived char, coal derived char, metal slag, dolomite, zeolite or biomass derived ashes. The use of these alternative supports is motivated by their lower cost compared to the conventional metal oxide supports. As regards the biochar derived supports, the presence of alkaline metals make these materials suitable for this purpose. 173 Moreover, the high surface provided by these supports ensures suitable metal dispersion. Nevertheless, these supported catalysts are not suitable for reaction-regeneration cycles due to their limited stability under oxygen atmosphere, which hinders their use in large-scale units.
As regards the use of metal slag as catalytic support, it provides the additional advantage of recycling an industrial waste residue. Besides, the metal additives present in the Nibased slag catalysts resulted in a better Ni dispersion and promoted synergistic catalysis effects, which resulted in an enhancement of catalyst activity and coke deposition resistance. 160 However, to provide reliable conclusions concerning the suitability of each support, further research must be conducted in terms of catalyst activity, stability, and regeneration. Generally, the use of alternative low-cost supports ( Figure  13b) leads to lower H 2 productions compared to metal oxide supports. Among these latter ones, ZrO 2 is regarded as a promising catalyst support for use in reaction-regeneration cycles. 100 Besides, as mentioned before, any comparison of the studies showed in Figure 13 should consider the catalyst design Energy & Fuels pubs.acs.org/EF Review (synthesis method and conditions), the operating parameters used in the runs and the type of biomass used as feedstock.
The use of alternative supports in the pyrolysis-reforming of plastic wastes has been analyzed in the literature, and details about the main studies are shown Table 9. In these studies, zeolite supports have been preferably selected for the production of H 2 or syngas. Yao et al. 186 analyzed different zeolite supported nickel catalysts (Ni/ZSM5−30, Ni/β-zeolite-25 and the Ni/Y-zeolite-30 catalysts) in a two-step fixed bed unit wherein HDPE was pyrolyzed at 500°C and the volatiles produced were reformed at a temperature range of 650−850°C . At the highest catalytic temperature and using a steam feeding rate of 6 g h −1 , the Ni/ZSM5−30 catalyst revealed the best performance, with a H 2 production of 13.2 wt %. Moreover, the Ni/Y-zeolite shows the worst performance in terms of syngas production and highest coke formation, which was ascribed to the ultramicropores in this zeolite support. In a similar experimental unit, but using PP as feedstock, Wu and Williams 164 reported the effectiveness of Ni/ZSM-5 catalyst for the production of hydrogen (19.0 wt %), as mainly filamentous coke was deposited on this catalyst, which has little influence on the catalytic activity.
4.3. Promoter Incorporation into Ni-Supported Catalysts. The incorporation of a promoter contributes to improving the activity, selectivity, and stability, as it modifies the active phase and/or the support. Thus, the promoter can positively enhance the thermal stability, mechanical strength, reducibility, metal dispersion, and coke resistance.
The elements that may act as promoters may be grouped as follows: (i) alkali metals, such as Li, Na, K, Rb, or Cs, which modify the reducibility of the metal active phase and enhance the initial catalyst activity; 187 (ii) alkaline-earth metals, such as Mg, Ca, Sr, and Ba, which reduce the acidity of the catalyst and enhance water adsorption and OH surface mobility, thereby reducing coke deposition rate; 188 (iii) rare earth oxides, such as La 2 O 3 , CeO 2 , or Pr 6 O 11 , which modify the metal−support interaction, improve the dispersion of the active phase, provide redox properties, reduce metal sintering, and so hinder carbon deposition; 117,189,190 and (iv) transition metal oxides, such as ZrO 2 , MnO x or ZnO, among others, which influence the metal−support interaction, decrease the acidity of the catalyst and enhance the coke deposition resistance. 88,117 Several authors have considered the addition of secondary metals (transition metals, such as Fe, Co, Mn, or Cu, and noble metals, such as Ru, Rh, Pd, or Pt) as promoters to form bimetallic catalysts. These types of catalysts are described in the following section (section 4.4).  iwi, Incipient wetness impregnation; swi, sequential wetness impregnation; cwi, co-impregnation; wi, wet impregnation; cp, co-precipitation. b H 2 production defined as g H2 /100 g biomass, daf (dry and ash free). c T P = Pyrolysis temperature; T R = reforming temperature.
Energy & Fuels pubs.acs.org/EF Review Table 10 summarizes the main researches reported in the literature on Ni-promoted Al 2 O 3 catalyst for biomass pyrolysis and in-line steam reforming. Thus, CeO 2 is regarded as a promising promoter for reforming catalysts due to its redox characteristics and oxygen storage capacity. Accordingly, Kimura et al. 124 prepared two Ni/CeO 2 -Al 2 O 3 catalysts by co-impregnation and sequential impregnation, respectively, obtaining better catalytic performance in terms of tar conversion and coke deposition with the catalyst prepared by co-impregnation, as this synthesis method leads to a stronger interaction between Ni and CeO 2 . Efika et al. 102 compared different Ni based catalysts, obtaining the lowest H 2 concentration and production (43.1% and 2.2 wt %, respectively) with a Ni/CeO 2 -Al 2 O 3 one containing a high CeO 2 load (20 wt %). Furthermore, the enhancement of the Ni/Al 2 O 3 catalyst performance by the addition of different promoters (La 2 O 3 , CeO 2 , and MgO) in the continuous biomass pyrolysis-reforming was analyzed by Santamaria et al. 191,192 In these studies, all the promoted catalysts revealed a similar initial activity, with the highest H 2 production by mass unit being for the Ni/CeO 2 -Al 2 O 3 catalyst (11.5 wt %), followed by Ni/La 2 O 3 -Al 2 O 3 (10.0 wt %) and Ni/MgO-Al 2 O 3 (9.7 wt %) catalysts. The stability was greatly improved by the incorporation of CeO 2 and La 2 O 3 promoters due to the characteristic features conferred upon the catalysts. Thus, CeO 2 provides redox properties, high oxygen storage capacity, and capability to favor water adsorption, whereas the basicity and water adsorption capability of La 2 O 3 promoter inhibits the formation of coke and leads to the gasification of the coke deposited. Despite the low reducibility of Ni/MgO-Al 2 O 3 due to MgAl 2 O 4 spinel phase formation, which leads to a poorer stability of this catalysts compared to Ni/Al 2 O 3 , these authors greatly improved the performance of MgO promoted catalyst Energy & Fuels pubs.acs.org/EF Review by the modification of the calcination temperature in the synthesis step. 193 The promotion of Ni/Al 2 O 3 catalyst with MgO has also been analyzed by Li et al. 194 in the pyrolysis-reforming of cedar wood. Thus, the influence the composition and reduction conditions have on the catalytic performance was determined, observing that the Ni/MgO-Al 2 O 3 catalyst with Ni/Mg/Al weight ratio of 9/66/25 exhibited much higher activity, resistance to coke deposition, and stability than Ni based catalysts supported on Al 2 O 3 and MgO. Likewise, Wu et al. 195 analyzed the influence of the reforming temperature on the pyrolysis/reforming of lignin on a Ni-Ca-Al catalyst, obtaining an improvement in H 2 production from 2.1 wt % at 700°C to 3.6 wt % at 900°C. Moreover, they tested a Ni-Mg-Al catalyst with three different biomass components (lignin, cellulose, and xylan) and reported the highest H 2 production for the cellulose feedstock (4.5 wt %).
The influence of adding alkali metals (K 2 CO 3 ) on a Ni/ Al 2 O 3 catalyst was analyzed by Kuchonthara et al. 103 in the steam reforming of rice husk-derived tar. They obtained better results with the promoted Ni/K 2 CO 3 -Al 2 O 3 catalyst in terms of carbon conversion and H 2 production, attributing this fact to the effect of this promoter by attenuating the thermal sintering of Ni-catalysts.
Similarly, Koike et al. 88 used a MnO x promoted Ni/Al 2 O 3 catalyst in the steam reforming of the volatiles from cedar wood pyrolysis and proved that an optimum catalyst composition, in which the interaction between Ni metal and MnO x is positively modified, enhances catalytic activity, and minimizes coke deposition. However, an excess of MnO x involves a decrease in the number of surface Ni atoms, reducing catalytic activity.
The influence of adding promoters to Ni/Al 2 O 3 catalysts on the catalytic steam reforming of the volatiles from plastic waste pyrolysis has also been assessed in the literature, and the main studies are presented in Table 11. Thus, Arregi et al. 55 evaluated the performance of Ni/Al 2 O 3 and two promoted catalysts (Ni/CeO 2 -Al 2 O 3 and Ni/La 2 O 3 -Al 2 O 3 ) in a con-tinuous bench scale pyrolysis-reforming plant using PP as the feedstock. The results evidenced a suitable performance of all the synthesized catalysts, with the best results in terms of conversion, hydrogen production, and coke deposition being obtained when the La 2 O 3 promoted catalyst was tested (H 2 production of 34.9 wt %).
Moreover, Nahil et al. 197 analyzed the influence of promoter incorporation into a nickel based catalyst for the co-production of hydrogen and carbon nanotubes in the pyrolysis−catalytic reforming of PP. The metal oxides selected as promoters were Zn, Mg, Ca, Ce, and Mn, and all the catalysts were prepared by the co-precipitation method. The Ni-Mn-Al was reported as the optimum catalyst, with H 2 production being 14.3 wt %.
Besides, a wide range of studies have been performed using Mg as the promoter on Ni/Al 2 O 3 catalysts. 47,150,164 Namely, the production of H 2 in a two-step fixed bed reactor was investigated by Wu and Williams, 47 who analyzed the influence the plastic type (PP, PS, HDPE, their mixtures, and real-world plastics) has on the product distribution. The highest H 2 production was obtained when PP was fed (26.6 wt %), followed by HDPE (26.0 wt %), whereas PS presented the lowest one (18.5 wt %). The use of real-world plastics (waste plastic processing) allowed a H 2 production of 23.6 wt %.
It is to note that, although increasing attention has been paid in recent years to the joint valorization of biomass and plastic mixtures by the two-step pyrolysis-reforming process, the studies performed are still scarce. Table 12 shows the main studies reported in the literature wherein promoted Ni based catalysts were selected for the pyrolysis and in-line steam reforming of biomass and plastic waste mixtures. Thus, Kumagai et al. 198 prepared a Ni-Mg-Al catalyst by the coprecipitation method and analyzed the influence Ca loading and catalyst calcination temperature have on the production of H 2 in pyrolysis−reforming runs using a mixture of wood sawdust and PP. The highest hydrogen production (6.2 wt %) was obtained in the presence of a Ca containing catalyst having a molar ratio of Ni/Mg/Al/Ca = 1:1:1:4, calcined at 500°C. More recently, a new dual-support catalyst Ni-CaO-C was used by Chai et al. 166 with the aim of enhancing the H 2 production in the catalytic reforming of the volatiles from the pyrolysis of different plastics (HDPE, PP, and PS) and biomass (pine sawdust) using a feedstock with a biomass/plastic ratio of 5:5. The results revealed that H 2 production decreased depending on the type of plastic mixed with the biomass, as follows: HDPE (14.0 wt %) > PP (13.8 wt %) > PS (13.2 wt %).
A comparison of the main Ni/Al 2 O 3 promoters used in the literature for the two-step pyrolysis-reforming strategy is shown in Figure 14. Thus, Figure 14a summarizes the studies dealing  Energy & Fuels pubs.acs.org/EF Review with biomass used as raw material, whereas Figure 14b displays the ones concerning to the use of polypropylene (PP) as feedstock (the use of other different types of plastics has been scarcely studied, and therefore the influence the promoter has on the reforming of other plastic wastes pyrolysis volatiles cannot be assessed). As observed in Figure 14, CeO 2 , La 2 O 3 , and MgO are the most widely used promoters in the literature for the pyrolysisreforming of biomass and plastic wastes. It is to note that, although the incorporation of a promoter can positively contribute to improving catalyst activity, the design of promoted catalysts is usually focused on enhancing catalyst stability by attenuating the fast deactivation caused by coke deposition. In this regard, the results provided in Figure 14 only give an idea of the initial catalyst activity, and therefore the stability of these catalysts must also be considered in further studies.
Moreover, the use of rare earth oxides, such as La 2 O 3 and CeO 2 as promoters, has led to encouraging results in the continuous steam reforming of the pyrolysis volatiles derived from biomass 191,192 and plastic wastes (PP). 55 Thus, the addition of CeO 2 as promoter improves considerably the Ni/ Al 2 O 3 catalyst stability, which is related to the CeO 2 redox properties that increase the surface available for oxygen as well as to its higher water adsorption capacity that enhances coke precursor gasification and attenuates catalyst deactivation. The incorporation of La 2 O 3 reduces the acidity of the Al 2 O 3 support, so coke formation and its water adsorption capability promotes the gasification of coke deposits. Thus, under suitable operating conditions, catalyst stability is improved and H 2 productions of around 10.5 wt % and 35 wt % are obtained when biomass and PP are used as the feedstock, respectively. 55,191,192 The use of alkaline-earth metals (Mg and Ca) and transition metal oxides (MnO x , ZnO) as promoters has also been analyzed in the literature, since these basic materials can reduce catalyst acidity and so coke formation. 196 However, the results shown in Figure 14 are also influenced by the following factors: (i) synthesis conditions (preparation method, calcination temperature, metal loading); (ii) reactor configuration; (iii) the operating variables in the pyrolysis (which greatly conditions the composition of the volatile stream to be reformed) and reforming steps; and (iv) process operation (batch or continuous mode). Thus, a more detailed research must be conducted to evaluate catalysts stability and regenerability and therefore step further toward the scalability of this two-step pyrolysis-reforming process for the production of H 2 from biomass and plastic wastes.
4.4. Bimetallic and Non-Nickel Based Catalysts. The catalytic performance highly depends on the active phase. As shown before, Ni is the most widely used active phase in the literature due to its high activity in the reforming reactions and moderate cost. However, other metals have been evaluated for the pyrolysis and in-line steam reforming of biomass, plastic wastes, and their mixtures. Among these metals, transition metals, such as Fe, Co, or Cu, and noble metals, such as Rh, Pt, Pd, and Ru are worth mentioning. Moreover, these metals have also been added as secondary metals forming bimetallic catalysts, which may improve catalytic activity and coke resistance. Thus, Tables 13 and 14 summarize the studies reported in the literature for non-nickel based catalysts and bimetallic catalysts, respectively, for the biomass derived feedstock, and Table 15 shows the ones when plastic wastes are used as feedstock.
Consequently, Li et al. 86 studied different Co based catalysts supported on Al 2 O 3 , ZrO 2 , SiO 2 , MgO, TiO 2 , and BaAl 12 O 19 (BA) in the steam reforming of the tar from the pyrolysis of wood biomass. The highest catalytic activity was obtained when Co/BA was used, which was attributed to the high dispersion obtained on this strongly basic support. Moreover, Kaewpanha et al. 184 tested different metal catalysts, i.e., Ni, Fe, and Rh supported on commercial zeolite, in the steam reforming of the tar derived from the steam pyrolysis of seaweed in a fixed bed reactor. They found that the highest H 2 production was obtained at a reaction temperature above 610°C , with Rh/zeolite catalyst being the most effective for tar removal, with a H 2 production of 5.3 wt %. Tomishige et al. 96 investigated the steam reforming of biomass-derived tars on several noble metal based catalysts (Rh, Pt, Pd, and Ru) supported on CeO 2 -SiO 2 . The experiments were conducted in a laboratory scale continuous feeding device, which was made up of a primary bed for pyrolysis of biomass and a secondary fluidized catalytic bed for the reforming step. They reported that the activity order at 550°C was as follows: Rh > Pt > Pd > Ru.
Moreover, although noble metal catalysts enhance catalyst activity and reduce coke deposition, their high cost limits their use as a single active phase. Therefore, their joint used with cheaper metals, such as Ni, in order to form bimetallic catalysts, is an interesting option followed by many authors. Similarly, with the aim of improving the overall activity of the catalysts, the addition of various transition metals to form different alloys has been approached in the literature.
The performance of Ni-Fe/Al 2 O 3 catalysts in the steam reforming of the tar from the pyrolysis of cedar wood was analyzed by Wang et al. 201 obtaining higher activity than that corresponding to monometallic Ni and Fe catalysts. The alloy formed between Ni and Fe improved the reaction involving the tar and hindered coke formation, since oxygen atoms are supplied by Fe species. Similarly, this research group analyzed the performance of Fe-Co/Al 2 O 3 , reporting its higher activity and stability compared to Fe/Al 2 O 3 and Co/Al 2 O 3 catalysts. Thus, the H 2 production obtained with the bimetallic catalyst was 6.0 wt %. 202 Moreover, Santamaria et al. 199 studied the effect the active phase has on the performance of the reforming catalysts. The metals selected as active phase were Ni, Co, and two bimetallic Ni-Co catalysts with different loading ratios. The runs were conducted in a continuous lab scale unit provided with a conical spouted bed reactor for the pyrolysis step and a fluidized bed reactor for the reforming of the volatiles formed in the first step. The H 2 production on these catalysts decreased as follows: Ni/Al 2 O 3 (10.17 wt %) > 7.5Ni-2.5Co/ Al 2 O 3 (9.94 wt %) > 5Ni-5Co/Al 2 O 3 (9.82 wt %) ≫ Co/ Al 2 O 3 (2.3 wt %). Moreover, the poor initial performance of Co/Al 2 O 3 catalyst at zero time on stream was attributed to the oxidizing nature of the steam, which favored the conversion of Co 0 into inactive CoO phase. In addition, the influence of incorporating noble metals into Ni based catalysts was investigated by Nishikawa et al. 203 Thus, Pt, Pd, Rh, and Ru were incorporated into a Ni/CeO 2 -Al 2 O 3 catalyst, and Pt was the most effective catalyst, even when its loading amount was low (0.01 wt %), which is due to the strong interaction between Pt and Ni to form the alloy. Moreover, the activity order attained was as follows: Pt > Rh > Ru > Pd.
The H 2 productions reported in the literature using different active phases in the biomass pyrolysis and in-line steam reforming are summarized in Figure 15. As previously mentioned, most of these studies have been conducted using transition metals, such as Fe and Co, due to their lower cost compared to noble metals. As observed, a maximum H 2 production of 8.5 wt % was obtained in a dual fixed bed reactor, using Co/BaAl 12 O 19 as reforming catalyst. 86 The lower H 2 productions obtained with metal phases alternative to Ni active phase have led to the use of bimetallic catalysts. However, it should be noted that the results provided in Figure  15 are greatly conditioned by the type of the support used (apart from the reaction conditions). Therefore, any comparison of the results obtained in the literature involves great difficulties. In fact, few studies deal with the influence the metal active phase and bimetallic catalysts have on H 2 production in the two-step pyrolysis-reforming process.
Similarly, the use of bimetallic and non-nickel based catalysts has also been assessed in the pyrolysis and in-line steam reforming of plastic wastes. Thus, Park et al. 43 conducted a parametric study of the two step pyrolysis-reforming of PP using two fixed bed reactors operating in the continuous regime (1 g min −1 ) and two commercial Ru/Al 2 O 3 catalysts with different Ru loadings (0.5 and 5 wt %). The highest H 2 production (36.5 wt %) was reported when the 5Ru/Al 2 O 3 catalyst was used under the following operating conditions: pyrolysis temperature of 400°C, reforming temperature of 680°C , and S/C ratio of 3.7. The use of PS as feedstock was analyzed by the same authors in the same experimental unit at pyrolysis and reforming temperatures of 400 and 630°C, respectively. 90 Under these conditions, they reported a slightly lower H 2 production when PS was fed (33.0 wt %) compared to the previous work, wherein they obtained a H 2 production of 34.2 wt % using PP as the feedstock.
Recently, the production of high H 2 /CO gas in the steam reforming of PS volatiles was analyzed by Zhou et al. 172 in a fixed bed reactor provided with two-zone horizontal furnaces. Ni-Fe bimetallic catalysts supported on ZrO 2 were synthesized by a co-impregnation method. The results showed that, although Fe/ZrO 2 was not an effective catalyst for this process, the bimetallic Ni-Fe/ZrO 2 catalyst could catalyze the reaction efficiently and produce a much higher H 2 compared to the monometallic catalyst (8.6 wt % for 15Ni5Fe/ZrO 2 and 2.6 wt % for 20Fe/ZrO 2 catalyst).

CHALLENGES AND PERSPECTIVES
In recent decades, the dependency of fossil fuels for energy production has triggered an increasing concern for the environmental problems associated with CO 2 emissions and global warming. Within this scenario, H 2 is regarded as one of the future energy carriers, and the development of new sustainable routes for H 2 production from biomass and plastic wastes is therefore a pressing need for the market. Among the different thermochemical routes for H 2 production, the twostep pyrolysis and in-line steam reforming strategy has gained increasing attention. Thus, this strategy provides several advantages compared to the conventional steam gasification and steam reforming of the bio-oil or plastic pyrolysis oil, as are: (i) operation under optimum conditions due to the integration of the two reactors in the same unit (the reforming temperature is lower than the one used in gasification process, thus reducing possible sintering problems of the catalyst); (ii) avoidance of tar formation; (iii) direct contact between the feedstock and its impurities with the reforming catalyst is avoided; and (iv) higher H 2 productions are attained (above 10 and 30 wt % when biomass and plastic wastes are used, respectively). Besides, the versatility of this process allows cofeeding biomass and plastic mixtures, reducing the limitations derived from seasonal biomass, and decreasing the environmental problems related to the management of plastic wastes.
Although great effort has been made to analyze in detail the optimum operating conditions in both steps, (i.e., pyrolysis and reforming temperatures, S/B or S/P ratio, space time, and so on), most of these studies have been conducted in laboratory reactors, which operate in batch mode. In fact, the studies performed in the continuos regime are restricted to bench scale units. Thus, the degree of development of pyrolysis-reforming technologies must be improved in order to ensure their scale up and implementation. Accordingly, special attention should be paid to the reactor design. Fluidized and spouted beds are those with better perspectives for the pyrolysis step. Thus, fast pyrolysis conditions allow obtaining high selectivity toward volatile compounds, especially bio-oil and plastic derived oil, with low solid byproduct formation. Accordingly, H 2 production in the overall pyrolysis-refoming process can be enhanced. Regarding the selection of the reforming reactor, two configurations have been analyzed in the previous literature, fixed and fluidized beds. Fixed beds are of easy design and operation, with catalyst attrition problems being avoided. However, fluidized beds have significant operational advantages over fixed beds for the full-scale operation. Thus, the solid circulation in fluidized beds ensures a better temperature control and avoids temperature gradients in the catalytic bed, which involves a remarkable challenge in highly endothermic steam reforming reactions. Moreover, the Energy & Fuels pubs.acs.org/EF Review reforming of biomass and plastics derived pyrolysis volatiles faces a fast deactivation rate, mainly associated with coke deposition. In this respect, a suitable reaction-regeneration strategy must be considered for the process scale up. Fluidized bed reactors are more flexible than fixed beds, as reactionregeneration strategies with continuous catalyst circulation between the reforming reactor and catalyst regenerator can be developed. Furthermore, the development of this process is greatly conditioned by the catalyst design, in which the catalytic materials and synthesis method and conditions are crucial facts. Accordingly, the main challenges to overcome to step further in the scaling-up of this two-step process are as follows: (i) the implementation of a continuous process in any reactor configuration, andconfiguration; and, (ii) the fast catalyst deactivation in the reforming step. It is worthy to note that, although the scaling-up of this strategy has not been implemented yet, certain research studies have proposed a continuous pyrolysis-reforming operation in bench scale units. Future research is needed to acquire further knowledge on reforming catalysts in order to progress toward the industrial scale. Special attention should be paid to the studies dealing with the mechanisms of catalyst deactivation and the role played by the catalyst design in the attenuation of catalyst deactivation. Thus, the influence catalytic materials, the synthesis method, and conditions have on the catalyst stability in the reforming step must be addressed. Alternative strategies may also be proposed for improving the fast catalyst deactivation, as are those involving the modification of the volatile stream to be reformed, which may be conducted in the pyrolysis reactor itself or downstream the pyrolysis reactor.
Besides, studies dealing with the regenerability of the reforming catalysts are essential for a suitable scaling of the pyrolysis-reforming process, since large scale plants entail working under reaction-regeneration cycles. Thus, prior to proceed with the optimum discrimination of catalysts, regeneration studies should be conducted.

CONCLUSIONS
This review analyzes the background and state-of-the-art of metal catalysts for the steam reforming of the volatile stream derived from the pyrolysis of biomass and waste plastics. In recent years, increasing research studies have focused on this strategy, since it has been demonstrated to be a suitable process for direct H 2 production.
The development of suitable reforming catalysts is essential for the viability of this process. Thus, a wide range of materials has been analyzed in the literature for improving the catalyst activity and stability in H 2 production by pyrolysis-reforming of biomass and plastic wastes. It should be noted that nonmetal based catalysts have not been included in this review due to their lower activity in the reforming reaction. The Ni/Al 2 O 3 catalyst has been extensively used due to its moderate cost, high activity of Ni metal, and suitable properties conferred by the Al 2 O 3 support (high specific surface area, mechanical strength, and stability) upon the catalyst. However, the acid nature of the Al 2 O 3 support, which promotes coke formation on the catalyst, has led to the substitution of Al 2 O 3 by other supports, both metal oxides, such as MgO, ZrO 2 , TiO 2 or CeO 2 , and low-cost materials, such as biomass and coal derived char, active carbon, dolomite, and zeolites. Among them, ZrO 2 is regarded as a promising catalyst support for its use in reaction-regeneration cycles. Conversely, the problems asso-ciated with the regeneration of the char and active carbon supported catalysts may condition its use in large scale units.
Incorporation of promoters has also been assessed in order to enhance catalyst stability in the steam reforming of the volatiles derived from the pyrolysis of biomass and plastic wastes. Thus, La 2 O 3 and CeO 2 have led to promising results for the attenuation of catalyst deactivation. The features of these promoters, i.e., basicity and water adsorption capability of La 2 O 3 and redox properties of CeO 2 promoter, involve a reduction in the formation of coke deposits and therefore improve the stability of Ni/Al 2 O 3 catalysts. Besides, the basic properties of other metal oxides, such as MgO, CaO, and MnO x have boosted their use as catalytic promoters.
Similarly, transition metals (mainly Co and Fe), noble metals (Rh, Ru, Pt, and Pd) and bimetallic catalysts have been tested in this two-step strategy, although these studies are still scarce.
A comparison of the diverse results reported in the literature involves great difficulty. Thus, apart from the catalyst selected in the reforming step, the overall H 2 production in the pyrolysis-reforming process from biomass, plastic wastes, and their mixtures greatly depends on the following factors: (i) type of feedstock, which must consider biomass/plastic features, particle size, moisture content, and ash composition; (ii) feeding regime, with continuous feed allowing for reaching the steady state and controlled operating conditions; (iii) pyrolysis step, in which the type of reactor, heating rate, gas residence time, and temperature greatly condition the volatile stream to be reformed; (iv) reforming conditions, as are type of reactor, S/B or S/P ratio, space time, and temperature; and (v) catalyst design, i.e., preparation method, catalytic materials, and synthesis conditions (calcination temperature, metal loading, reduction temperature).
Although encouraging results have been obtained based on this strategy, further knowledge of reforming catalysts is required in order to progress toward the industrial scale. Thus, special attention should be paid to the studies dealing with the mechanisms of catalyst deactivation and regeneration in order to proceed with the optimum discrimination of catalysts.
Aitor Arregi − Department of Chemical Engineering,