Hydrogen-rich gas production by continuous pyrolysis and in-line catalytic reforming of pine wood waste and HDPE mixtures

1 Hydrogen-rich gas production by continuous pyrolysis and in1 line catalytic reforming of pine wood waste and HDPE 2 mixtures 3 Aitor Arregi, Maider Amutio, Gartzen Lopez*, Maite Artetxe, Jon Alvarez, Javier 4 Bilbao and Martin Olazar 5 Department of Chemical Engineering, University of the Basque Country UPV/EHU, 6 P.O. Box 644 - E48080 Bilbao (Spain). [email protected]7 Abstract 8 The continuous pyrolysis-catalytic steam reforming of different mixtures of biomass 9 and high density polyethylene (25, 50 and 75 wt % HDPE) has been carried out in a 10 two-step reaction system, provided with a conical spouted bed reactor (CSBR) and 11 fluidized bed. The influence of HDPE co-feeding on the conversion, reforming products 12 yields and composition and catalyst deactivation has been studied at a reforming 13 temperature of 700 ºC, with 16.7 gcat min gfeeding-1 and steam/(biomass+HDPE) mass 14 ratio of 4, comparing the results with those obtained by feeding pure biomass and 15 HDPE. The co-feeding of plastics enhances the production of hydrogen, which 16 increases from 10.9 g of H2 per 100 g of biomass to 37.3 g of H2 per 100 g of HDPE 17 fed. Catalyst deactivation by coke is attenuated when HDPE is co-fed due to the lower 18 content of oxygenated compounds in the reaction environment. The higher yield of 19 hydrogen achieved with this two-step (pyrolysis-reforming) strategy, its flexibility to 20 jointly valorise biomass and plastic mixtures and the lower temperatures required in 21 relation to gasification, makes this process promising to produce H2 from renewable raw 22 materials and wastes. 23 This is the accepted manuscript of the article that appeared in final form in Energy Conversion and Management 136 : 192-201 (2017), which has been published in final form at https://doi.org/10.1016/j.enconman.2017.01.008. © 2017 Elsevier under CC BY-NC-ND license (http:// creativecommons.org/licenses/by-nc-nd/4.0/) 2 Keywords: hydrogen; pyrolysis; reforming; biomass; waste plastics; conical spouted 24 bed, catalyst deactivation 25 1. Introduction 26 The environmental awareness associated with the use of traditional resources (natural 27 gas, petroleum and coal) gives way to the development of new routes for sustainable 28 hydrogen production, whose demand is growing due to its interest as energy carrier and 29 reactant in refinery hydroprocessing units [1]. In this scenario, biomass can play an 30 important role as an alternative feedstock, given that is a CO2 neutral renewable source 31 and chemicals or fuels produced from it are considered sustainable [2]. 32 Amongst the different thermochemical routes, direct steam gasification [3-6] and the 33 indirect route of bio-oil reforming [7-10] are the most studied routes for hydrogen 34 production from biomass. Nevertheless, the gasification process is directed to produce 35 syngas and the tar formation is an issue for its industrial applications [11,12]. On the 36 other hand, the indirect route of bio-oil reforming has several problems related to bio-oil 37 properties and its vaporization and re-polymerization [13,14]. Therefore, the two-step 38 pyrolysis-catalytic steam reforming process, in which is not necessary to condensate 39 and re-vaporized the bio-oil, is gaining attention last years [15-19]. This process, in 40 which each step is carried out in different reactors, involves some advantages in relation 41 to one-step pyrolysis process with a reforming catalyst in-situ. On the one hand, the 42 temperature of each step can be optimized in order to maximize the production of 43 hydrogen [20] and on the other hand, the catalyst is more effective for volatiles 44 transformation and the process is more versatile in order to establish the desired 45 catalyst/feeding ratio. Therefore, a more homogeneous product stream will be obtained, 46 due to the higher efficiency of the catalyst in order to attenuate secondary reactions. 47 3 Nevertheless, the low content of hydrogen and high content of oxygen of the biomass 48 feedstock is a drawback to obtain high production of hydrogen. Moreover, the catalyst 49 has a considerable deactivation by coke [17]. In this work, the improvement of H2 50 production and the attenuation of catalyst deactivation by the valorisation of biomass 51 and HDPE mixtures has been studied. 52 The improvement on hydrogen production has been reported by several authors in the 53 co-gasification of biomass and HDPE mixtures [21-24]. The co-feeding solves the 54 seasonal limitations of biomass availability and contributes to attenuate the 55 environmental problems associated to the waste plastics management. Even though the 56 pyrolysis is considered a suitable route for the valorisation of waste plastics on a large 57 scale, and particularly for polyolefins [25-27], the studies concerning pyrolysis and in-58 line catalytic steam reforming of biomass and plastic mixtures are very scarce. Alvarez 59 et al. [20] studied the co-feeding of polypropylene in pyrolysis-reforming of biomass in 60 batch laboratory scale reactor, obtaining higher gas yield and higher hydrogen 61 production in relation to feeding pure biomass. In the same experimental unit Kumagai 62 et al. [28] tested a Ni-Mg-Al-Ca catalyst synthesized by a co-precipitation method for 63 pyrolysis-reforming of a biomass/polypropylene mixture, obtaining a maximum 64 hydrogen production of 6.0 g of H2 per 100 g of feeding when the catalyst was calcined 65 at low temperatures, 500 ºC. 66 The aim of this work is to increase the production of hydrogen by plastics co-feeding, 67 using a continuous two-step process (Figure 1). The equipment combines the excellent 68 performance of the conical spouted bed reactor (CSBR) for the pyrolysis of biomass 69 [29] and plastics [30] with the suitability of the fluidized bed reactor for the steam 70 reforming process [31,32]. The cyclic vigorous movement of the sawdust and sand 71 particles coated with melted plastic in the CSBR minimizes the segregation problems 72 4 and avoids the defluidization of the bed. On the other hand, the fluidized bed catalytic 73 reactor allows controlling the temperature of the endothermic reforming reaction and 74 delays the blocking of the bed by coke formation. This two step configuration has been 75 described in previous papers for the pyrolysis-catalytic steam reforming of biomass [17] 76 and plastics [33], in which the good performance of the process without operational 77 problems and high hydrogen yields were reported. 78 Figure 1 (falta este haremos un esquema del proceso) 79 2. Materials and Methods 80 2.1. Materials 81 Pine sawdust (pinus insignis) waste has been crushed, ground and sieved to a particle 82 size between 1 and 2 mm, which is a suitable particle diameter in order to guarantee the 83 good performance of the solid feeding system, and dried at room temperature to a 84 moisture content below 10 wt %. The high density polyethylene (HDPE) was provided 85 by Dow Chemical (Tarragona, Spain) in the form of chippings (4 mm), with the 86 following properties: average molecular weight, 46.2 kg mol-1; polydispersity, 2.89 and 87 density, 940 kg m-3. The higher heating value (HHV) of both feedstocks has been 88 measured in a Parr 1356 isoperibolic bomb calorimetry. Moreover, the ultimate and 89 proximate analyses have been determined in a LECO CHNS-932 elemental analyzer 90 and in a TGA Q5000IR thermogravimetric analyzer, respectively and the results of the 91 characterization of biomass and HDPE used in this study are summarized in Table 1. 92 Table 1. Characterization of the biomass and HDPE used. 93 Ultimate analysis (wt %) Biomass HDPE Carbon 49.33 85.71 5 Hydrogen 6.06 14.29 Nitrogen 0.04 0 Oxygen 44.57 0 Proximate analysis (wt %) Volatile matter 73.4 99.7 Fixed carbon 16.7 0.3 Ash 0.5 - Moisture 9.4 - HHV (MJ kg-1) 19.8 43.1 94 A commercial Ni reforming catalyst (G90-LDP) provided by Süd Chemie (Germany) 95 has been used for the reforming step. The original catalyst (in the form of perforated 96 rings 19 x 16 mm) has been ground and sieved between 0.4-0.8 mm, which is the 97 suitable particle size in order to guarantee the fluid dynamic conditions of the fluidized 98 bed. The metal content (provided by the supplier) and physical properties of the catalyst 99 are summarized in Table 2. The adsorption-desorption isotherm of the catalyst has been 100 measured by N2 adsorption-desorption (Micromeritics ASAP 2010). As observed, the 101 catalyst shows low BET surface area and low porosity. 102 Table 2. Metal content and physical properties of the catalyst. 103 Catalyst NiO content (wt %) SBET (m2 g-1) Vporous (cm3 g-1) dporous (Å) G90-LDP 14 19 0.04 122 104 Moreover, the catalyst has been reduced in-situ in order to ensure its activity. The 105 reduction has been conducted for 4 h under 10 vol % H2 at 710 ºC according to the 106 results obtained by temperature programmed reduction. Both the adsorption-desorption 107 isotherm and the TPR profile of this catalyst can be found elsewhere [34,35]. 108 6 2.2. Equipment and reactors 109 Figure 2 shows the scheme of the experimental equipment. The plant is provided with 110 two reactors in-line: (i) a CSBR for pyrolysis step and (ii) a fluidized bed reactor for the 111 reforming step of pyrolysis volatiles. 112 113 Figure 2. Scheme of the bench scale unit. 114 The plant is provided with two independent feeders for biomass and HDPE as 115 segregation problems took place when both materials were mixed in a single unit. Each 116 feeder consists of a vessel equipped with a vertical shaft connected to a piston placed 117 below the material bed. The material is fed into the reactor by raising the piston at the 118 same time as the whole system is vibrated by an electric engine. The pipe that connects 119 7 the feeders with the reactor is cooled with tap water. Moreover, a very small nitrogen 120 flow is introduced into the vessel, which avoids the condensation of steam in the 121 feeding vessel. 122 A pump (Gilson 307) has been used in order to feed the water into the pyrolysis reactor, 123 which has been previously vaporized by an electric cartridge placed inside the forced 124 convection oven. Nitrogen, air or hydrogen can also be introduced to the CSBR reactor 125 and their flows are controlled by mass flow controllers, which allow feeding up to 20 L 126 min-1 of nitrogen and air, and up to 5 L min-1 of hydrogen. The temperature of the steam 127 and the gases is increased up to reaction conditions in a gas preheater located in the 128 lower section of the reactor, which is filled with stainless steel pipes that increase the 129 surface area for heat transfer. 130 The pyrolysis step has been carried out in a CSBR. This reactor has been successfully 131 used in the pyrolysis and gasification of different waste materials, such as biomass 132 [5,36], plastics [37,38] and tyres [39,40]. The detailed design and main dimensions of 133 the CSBR can be found elsewhere [17,33]. The temperature of the reactor is controlled 134 by two thermocouples located inside the reactor, one in the bed annulus and the other 135 one close to the wall. Prior entering the reforming reactor the product stream is cleaned 136 by a high-efficiency cyclone, for retaining the fine sand and char particles entrained 137 from the CSBR. 138 In order to avoid the blocking of the flow due to the coke deposition, which has been 139 observed in a fixed bed reactor [34], the reforming step has been carried out in a 140 fluidized bed reactor, whose dimensions are 38.1 mm of diameter and 440 mm of 141 length. The temperature of the fluidized bed reactor is controlled by a thermocouple 142 placed inside the catalyst bed. The volatiles from fluidized bed reactor circulate through 143 8 a sintered steel filter (5 µm) to retain catalyst fines elutriated from the fluidized bed, 144 with this amount being bellow 5% of the catalyst used in the experiments performed. 145 All the interconnection pipes, cyclone, filter and both reactors are located inside an oven 146 kept at 270 ºC, which ensures that the steam and products are not condensed in the 147 connections between the reactors. 148 Finally, the condensation system of the plant ensures total condensation and retention of 149 non-reacted steam and biomass and HDPE derived products, which consists of a 150 condenser and a coalescence filter. 151 2.3. Experimental conditions 152 The fluid dynamic requirements of the two reactors in-line with a common gas flow 153 have conditioned the steam flow and the particle size of the sand in the CSBR and the 154 particle size of both catalyst and sand in the fluidized bed reactor. Thus, 3 mL min-1 of 155 water flow has been established, which corresponds to a steam flow of 3.73 NL min-1, 156 and the bed consists of 50 g of sand in the pyrolysis step, with particle size being 157 between 0.3-0.35 mm. The runs have been carried out in continuous regime by feeding 158 0.75 g min-1 of biomass and HDPE mixtures. Moreover, the pyrolysis step has been 159 carried out at 500 ºC, which has been proved in previous studies to be a suitable 160 temperature for biomass [29] and HDPE pyrolysis [30] in a CSBR. 161 In the same way, after fluid dynamic tests, a bed of 25 g of catalyst and sand mixture 162 has been established for the fluidized bed reactor, with particle size being between 0.4-163 0.8 mm for the catalyst and 0.3-0.35 mm for the sand, in order to work with a relative 164 velocity 3 or 4 times higher than minimum fluidization velocity. These conditions 165 guarantee the complete fluidization of the bed, even when the coke content of the 166 catalyst is high. 167 9 The effect of feeding different HDPE/biomass mass ratios in the pyrolysis-reforming 168 process has been studied. Thus, HDPE/biomass mixtures of 25/75, 50/50 and 75/25 wt 169 % have been tested and the results have been compared with those of pure biomass and 170 plastic feeds. The temperature of the reforming step was 700 ºC, given that is the 171 minimum temperature needed for the complete conversion of volatiles from HDPE 172 pyrolysis [33]. The other operating conditions of the process are the following: 16.7 gcat 173 min gfeeding-1 (corresponding to 12.5 g of catalyst) and steam/(biomass+HDPE) mass 174 ratio of 4. The runs have been repeated several times (at least 3) under the same 175 conditions in order to guarantee reproducibility of the results. 176 2.4. Product analysis 177 The volatile products of the reforming step have been analysed on-line by means of a 178 GC Agilent 6890 provided with a HP-Pona column and a flame ionization detector 179 (FID). The sample has been transferred from the reactor to the GC by means of a 180 thermostated line at 280 ºC, in order to avoid the condensation of heavy compounds. 181 Moreover, the non-condensable gases have been analyzed on-line in a micro GC 182 (Varian 4900) once the gases were completely free of steam and non-reacted liquid 183 products. 184 The coke content deposited on the reforming catalyst has been determined at the end of 185 continuous experiments by temperature programmed oxidation (TPO) in a 186 thermobalance TGA Q5000 (TA Instruments), which was connected on-line to a mass 187 spectrometer Thermostar (Balzers Instruments), given that the Ni of the catalyst is 188 oxidized together with the carbonaceous coke, and accordingly, the carbon dioxide 189 formation must be monitored to determine TPO curves. The following procedure has 190 been carried out: (i) signal stabilization with He stream (10 mL min-1) at 100 ºC, (ii) 191 16 content of both feedstocks. Given the relation between the results of the reforming step 313 and the composition of the inlet stream to the reactor, the lineal increase of the reaction 314 indexes plotted in Figure 3 shows that there is not a significant synergetic effect of the 315 HDPE co-feeding to the pyrolysis reactor on reforming products composition. 316 Nevertheless, this effect has been observed by other authors, who have verified that the 317 co-pyrolysis of plastics and biomass have a noticeable effect on bio-oil composition 318 [50,51]. 319 In order to compare the results in Figure 3 with the literature, it should be pointed out 320 that the pyrolysis and in-line steam reforming of biomass and plastic mixtures is limited 321 to the studies of the research group headed by prof. Williams. Thus, the influence of 322 biomass/polypropylene ratio (feeding between 5 and 20 wt % of PP) has been studied 323 by Alvarez et al. [20] in the batch pyrolysis-reforming process on a Ni/Al2O3 catalyst, 324 obtaining a maximum hydrogen production of 5.5 wt %, when 20 wt % of PP was used. 325 Kumagai et al. [28] studied a Ni-Mg-Al-Ca catalyst with different Ca contents and 326 calcination temperatures and the highest hydrogen production of 6.0 wt % was obtained 327 using a calcination temperature of 500 ºC. These results are significantly lower in 328 relation to those obtained in this study, which is a consequence of the continuous mode 329 used in this work. On the other hand, the results reported by other authors for the 330 individual valorization of biomass and plastics by continuous pyrolysis and in-line 331 reforming are in the same range of those obtained in the present study [15,16,52,53]. 332 The steam co-gasification of biomass and polyolefins studies reveal the existence of 333 positive and even synergetic effects over the hydrogen and gas production and tar 334 content in the gas product [21-24,54]. However, the hydrogen production obtained in 335 this work is higher than those reported in the steam gasification processes, between 4 336 17 and 7 wt % for biomass gasification [5,55,56] and in the 6-15 wt % range in the 337 gasification of polyolefins [37,57]. 338 339 Figure 3. Effect of HDPE co-feeding in the biomass pyrolysis and in-line 340 reforming over gas and hydrogen productions and reacted steam. 341 Figure 4 shows the effect of HDPE content in the feed on individual product yields 342 (graph a) and gas composition (graph b). As it can be observed, there are notable 343 differences in H2, CO2 and CO yields distribution. In this way, H2 and CO2 yields 344 decrease when HDPE content is increased from 0 to 100 wt %, from 93.2 to 85.7 % and 345 87.2 to 67.4 %, respectively. Nevertheless, CO yield increases from 12.5 to 29.3 %. 346 These results evidence the effect of the higher carbon content of HDPE and therefore, 347 the higher amount of carbon to be reformed. Moreover, as the same space time is used 348 for all experiments, higher yield of CO and lower yield of CO2 are obtained when 349 18 HDPE content in the feed is increased, due to the lower extent of WGS reaction (eq. 350 10). It can also be observed that CH4 and C2-C4 yields are very low in all cases studied, 351 although there is a slight increase of these yields when HDPE content is increased from 352 0 to 100 wt %, from 0.2 to 0.8 % for CH4 and from 0.1 to 0.7 % for C2-C4 fraction. 353 It can be pointed out that H2 concentration increases with HDPE content in the feed 354 until 72 vol % when 75 wt % of HDPE is used (Figure 4b), due to the higher content of 355 hydrogen and lack of oxygen in the plastics composition. Nevertheless, the lower 356 effective space time when HDPE is co-fed gives way to lower CO2 and higher CO 357 concentrations in the gaseous fraction, which change from 30.2 to 20.9 % and from 4.3 358 to 9.1 %, respectively, in the range of HDPE content studied. The hydrogen 359 concentrations reported by Alvarez et al. [20] in the pyrolysis-reforming of biomass/PP 360 mixtures were below those obtained in the present study, with the maximum value 361 being of 52.1 vol. % for 20 wt % of PP in the feed. The hydrogen concentrations 362 obtained in the steam co-gasification of these feedstocks are also below, in the 40 to 55 363 % range [21,23,58]. 364 19 365 366 Figure 4. Effect of HDPE co-feeding in the HDPE/biomass mixture fed over the 367 individual products yields (a) and gaseous product concentrations (b), in the pyrolysis 368 and in-line reforming process. 369 3.2.2. Catalyst deactivation 370 In order to study the effect of feed composition on the reforming catalyst deactivation, 371 the evolution with time on stream of conversion (Figure 5) and gas composition (Figure 372 6) in the reforming step has been analyzed. Figure 5 shows that the deactivation 373 behaviour strongly depends on the feed composition. Thus, the conversion in reforming 374 step is below 60 % for pure biomass after 120 min of continuous operation, whereas is 375 higher than 90 % with pure HDPE for the same time on stream. Moreover, it is 376 noteworthy the linear decay of catalyst activity for HDPE, while in the case of biomass, 377 the activity is maintained for the first 60 minutes and follows an acute decreasing trend 378 above 75 minutes. 379 The initial stable conversion period observed with pure biomass cannot be related to the 380 lower deactivation with this feed, and can be explained due to the space time value in 381 excess with respect to the equilibrium one. Consequently, the higher decrease of the 382 activity observed in Figure 5 for biomass is especially relevant, taking into account the 383 20 higher effective space time (around 2.5 times higher) for this feeding. Thus, these 384 results clearly shows that the oxygenated compounds and aromatic rings containing 385 compounds (as phenols) formed in biomass pyrolysis provoked a much faster 386 deactivation than that caused by long chain hydrocarbons from HDPE pyrolysis. In the 387 same line, Czernik et al. [49] remarked that oxygenated compounds have more marked 388 tendency than that of hydrocarbons to form carbonaceous deposits on the catalyst 389 surface and accordingly, provoke a faster catalyst deactivation. In fact, the severe 390 reforming catalyst deactivation has been previously reported by other authors in the 391 reforming of biomass derived oxygenates [59-61]. 392 When different HDPE/biomass mass ratios are used, the conversions evolutions 393 observed are between those of two pure feedings, which confirms that the plastic co-394 feeding has a notable effect on attenuation of catalyst deactivation. 395 396 21 Figure 5. Effect of HDPE content in HDPE/biomass mixture fed over the evolution 397 with time on stream of conversion in the reforming step in pyrolysis and in-line 398 reforming process. 399 Figure 6 shows the evolution of gas composition with time on stream for three different 400 feedings: pure biomass (a and b), a HDPE/biomass mixture of 50/50 wt % (c), and pure 401 HDPE (d). When pure biomass is used, H2 concentration decreases from 65 to 55 vol % 402 in 120 min on stream (Figure 6a), whereas it is maintained around 70 vol % when pure 403 HDPE is fed (Figure 6d). Moreover, for pure biomass valorization CO concentration 404 increases from 5 to 16 vol % and the opposite occurs for that of CO2, which decreases 405 from 30 to 24 vol % after 120 min on stream. This evolution of CO and CO2 406 concentrations reveal a significant deactivation of the catalyst towards the WGS 407 reaction (eq. 10). On the other hand, the concentrations of CO and CO2 remain constant 408 for pure HDPE (Figure 6d) and only change above 75 min on stream for different 409 HDPE/biomass mixtures (Figure 6c). Therefore, when HDPE is co-fed, the deactivation 410 of the WGS reaction is considerably attenuated. 411 As discussed previously, the concentration of main gaseous products formed by 412 secondary cracking reactions, i.e. CH4 and C2-C4 fraction (ethylene, ethane, propylene 413 and propane, mainly), are very low for different feeds studied due to the initial catalytic 414 activity for both oxygenated compounds and hydrocarbons reforming (eq. (8-9)) and 415 WGS reaction (eq. (10)). However, when the catalyst is deactivated, CH4 and C2-C4 416 fraction concentrations increase slightly, which is shown in detail in Figure 6b for 417 biomass valorization. A similar although less marked trend can be seen when HDPE is 418 co-fed (results not shown). 419 22 420 421 422 423 Figure 6. Evolution with time on stream of gas composition in the pyrolysis-424 reforming of pure biomass (a and b), HDPE/biomass mixture of 50/50 wt % (c) and 425 pure HDPE (d). 426 3.3.4. Characterization of the coke deposited 427 In order to explain the effect of feeding composition on the evolution of conversion 428 with time on stream, the coke deposited on the catalyst has been characterized by 429 temperature programmed oxidation (TPO) and transmission electron microscopy (TEM) 430 images. Figure 7 shows the TPO profiles of deactivated catalyst for three feedings: pure 431 biomass, a mixture of HDPE/biomass of 50/50 wt % and pure HDPE. In the 432 valorization of pure biomass, a main peak at around 600 ºC is observed, which 433 23 corresponds to a polyaromatic and structured coke, with a small shoulder at around 425 434 ºC related to coke whose combustion is activated by the Ni metallic sites. This coke can 435 be related to the carbon whiskers reported by Trane-Restrup and Jensen [2] in the steam 436 reforming of furfural and guaiacol at 600 ºC. On the other hand, in the steam reforming 437 of the pyrolysis products of pure HDPE, a main peak at 580 ºC with a shoulder at 450 438 ºC was observed. The main peak corresponds to a structured and filamentous coke 439 similar to that obtained by Wu and Williams [62] and Acomb et al. [63] in the 440 reforming of polypropylene (PP), as it could be observed in TEM images which will be 441 discussed later. The slight difference of maximum temperature (605 ºC for these 442 authors) can be attributed to the higher porosity of the catalyst used by them, which 443 complicates the combustion of the coke fraction which blocks the pores of the catalyst. 444 The peaks for the mixture of biomass and HDPE are between the TPO profiles obtained 445 for pure biomass and HDPE. 446 447 24 Figure 7. Comparison of TPO profiles of coke deposited in the catalyst for pure 448 biomass, HDPE/biomass mixture of 50/50 wt % and pure HDPE valorization. 449 Figure 8 shows TEM images of the deactivated catalyst for different feedings: pure 450 biomass (a), HDPE content in the feed of 25 wt % (b), 50 wt % (c), 75 wt % (d) and 451 pure HDPE (e). As observed, different structure and nature of the coke can be 452 distinguished, which is a consequence of the different composition of the volatiles fed 453 into the reforming step. In the images, Ni active sites can be identified as darker areas 454 and Figure 8a shows that the coke deposited is mainly non-structured for pure biomass 455 valorization, covering completely the Ni crystals (encapsulating coke). The presence of 456 amorphous and non-structured coke has also been observed in the catalytic steam 457 reforming of methane [64] different hydrocarbons [62,65] and oxygenated compounds 458 [2,66]. The high combustion temperature of this coke observed in TPO profile (600 ºC) 459 shows that it is a very condensed coke. 460 However, the structure of the coke changes when HDPE is co-fed, with its nature being 461 more filamentous as HDPE content in the feed is increased. This nature of filamentous 462 coke has been observed previously in the reforming of polyolefins pyrolysis products 463 [62,63]. 464 Consequently, the faster deactivation observed for biomass and the attenuation of the 465 deactivation when HDPE is co-fed can be attributed to the different nature of the coke. 466 The amorphous coke formed in the reforming of oxygenated compounds derived from 467 biomass pyrolysis encapsulates the Ni centres, causing fast deactivation of the catalyst, 468 whereas the structured and filamentous coke formed mainly in the reforming of 469 hydrocarbons derived from HDPE pyrolysis do not block the Ni active centres, although 470 its progressive deposition complicates the gas flow of the reactants into Ni particles 471 25 [65,67]. This interpretation of the deactivation results is consistent with the fast 472 deactivation of the catalyst in the reforming of oxygenated compounds (DME, ethanol 473 and bio-oil), attributed to the encapsulation of Ni centres by the amorphous coke formed 474 by condensation of intermediate oxygenates [8,66,68,69]. The deactivation is lower in 475 the reforming of hydrocarbons produced in the pyrolysis of polyolefins, where the coke 476 is mainly structured [62,63]. 477 32 [51] Xue Y, Zhou S, Brown RC, Kelkar A, Bai X Fast pyrolysis of biomass and waste plastic in a 649 fluidized bed reactor. Fuel 2015;156:40-6. doi:10.1016/j.fuel.2015.04.033. 650 [52] Xiao X, Meng X, Le DD, Takarada T Two-stage steam gasification of waste biomass in 651 fluidized bed at low temperature: Parametric investigations and performance optimization. 652 Bioresour Technol 2011;102:1975-81. 653 [53] Czernik S, French RJ Production of hydrogen from plastics by pyrolysis and catalytic steam 654 reform. Energy and Fuels 2006;20:754-8. 655 [54] Ahmed II, Nipattummakul N, Gupta AK Characteristics of syngas from co-gasification of 656 polyethylene and woodchips. Applied Energy 2011;88:165-74. 657 [55] Koppatz S, Pfeifer C, Hofbauer H Comparison of the performance behaviour of silica sand 658 and olivine in a dual fluidised bed reactor system for steam gasification of biomass at pilot 659 plant scale. Chem Eng J 2011;175:468-83. 660 [56] Michel R, Rapagn S, Di Marcello M, et al. Catalytic steam gasification of Miscanthus X 661 giganteus in fluidised bed reactor on olivine based catalysts. Fuel Process Technol 662 2011;92:1169-77. 663 [57] He M, Xiao B, Hu Z, Liu S, Guo X, Luo S Syngas production from catalytic gasification of 664 waste polyethylene: Influence of temperature on gas yield and composition. International 665 Journal of Hydrogen Energy 2009;34:1342-8. 666 [58] Narobe M, Golob J, Klinar D, Francetic V, Likozar B Co-gasification of biomass and plastics: 667 Pyrolysis kinetics studies, experiments on 100kW dual fluidized bed pilot plant and 668 development of thermodynamic equilibrium model and balances. Bioresour Technol 669 2014;162:21-9. doi:10.1016/j.biortech.2014.03.121. 670 [59] Trane R, Dahl S, Skjoth-Rasmussen MS, Jensen AD Catalytic steam reforming of bio-oil. 671 International Journal of Hydrogen Energy 2012;37:6447-72. 672 [60] Rioche C, Kulkarni S, Meunier FC, Breen JP, Burch R Steam reforming of model compounds 673 and fast pyrolysis bio-oil on supported noble metal catalysts. Applied Catalysis B: 674 Environmental 2005;61:130-9. doi:http://dx.doi.org/10.1016/j.apcatb.2005.04.015. 675 [61] Lemonidou AA, Kechagiopoulos P, Heracleous E, Voutetakis S Steam Reforming of Bio-oils 676 to Hydrogen. The Role of Catal for the Sustain Prod of Bio-Fuels and Bio-Chem 2013:467-93. 677 doi:10.1016/B978-0-444-56330-9.00014-0. 678 [62] Wu C, Williams PT Investigation of coke formation on Ni-Mg-Al catalyst for hydrogen 679 production from the catalytic steam pyrolysis-gasification of polypropylene. Applied Catalysis 680 B: Environmental 2010;96:198-207. 681 [63] Acomb JC, Wu C, Williams PT Control of steam input to the pyrolysis-gasification of waste 682 plastics for improved production of hydrogen or carbon nanotubes. Applied Catalysis B: 683 Environmental 2014;147:571-84. 684 33 [64] Angeli SD, Pilitsis FG, Lemonidou AA Methane steam reforming at low temperature: Effect 685 of light alkanes' presence on coke formation. Catal Today 2015:119-28. 686 doi:10.1016/j.cattod.2014.05.043. 687 [65] Barbarias I, Lopez G, Amutio M, et al. <br />Steam reforming of plastic pyrolysis model 688 hydrocarbons and catalyst deactivation. Chem Eng J 2016;Submitted for publication. 689 doi:10.1016/j.cej.2016.03.091. 690 [66] Montero C, Ochoa A, Castaño P, Bilbao J, Gayubo AG Monitoring Ni0 and coke evolution 691 during the deactivation of a Ni/La2O3-aAl2O3 catalyst in ethanol steam reforming in a fluidized 692 bed. J Catal 2015;331:181-92. doi:10.1016/j.jcat.2015.08.005. 693 [67] Latorre N, Cazaña F, Martínez-Hansen V, Royo C, Romeo E, Monzón A Ni-Co-Mg-Al 694 catalysts for hydrogen and carbonaceous nanomaterials production by CCVD of methane. Catal 695 Today 2011;172:143-51. doi:10.1016/j.cattod.2011.02.038. 696 [68] Vicente J, Ereña J, Montero C, Azkoiti MJ, Bilbao J, Gayubo AG Reaction pathway for 697 ethanol steam reforming on a Ni/SiO2 catalyst including coke formation. Int J Hydrogen Energy 698 2014;39:18820-34. 699 [69] Vicente J, Montero C, Ereña J, Azkoiti MJ, Bilbao J, Gayubo AG Coke deactivation of Ni and 700 Co catalysts in ethanol steam reforming at mild temperatures in a fluidized bed reactor. Int J 701 Hydrogen Energy 2014;39:12586-96. 702 703