Microsoft Word - 476hernandez.docx


 CHEMICAL ENGINEERING TRANSACTIONS  
 

VOL. 43, 2015 

A publication of 

The Italian Association 
of Chemical Engineering 
Online at www.aidic.it/cet 

Chief Editors: Sauro Pierucci, Jiří J. Klemeš 
Copyright © 2015, AIDIC Servizi S.r.l., 
ISBN 978-88-95608-34-1; ISSN 2283-9216                                                                               

 

Hydrogen Production from Biomass Pyrolysis and In-line 
Catalytic Steam Reforming 

Aitor Arregi, Itsaso Barbarias, Jon Alvarez, Aitziber Erkiaga, Maite Artetxe, Maider 
Amutio, Martin Olazar* 

University of the Basque Country UPV/EHU. Chemical Engineering Department. PO Box 644- 48080 Bilbao, Spain 
martin.olazar@ehu.es 

Hydrogen production from pyrolysis-catalytic steam reforming of pine sawdust has been investigated using 
two subsequent reactors: i) a conical spouted bed reactor for biomass pyrolysis at 500 ºC, and ii) a fluidized 
bed reactor for catalytic reforming of volatiles from the pyrolysis step. A commercial Ni reforming catalyst has 
been used for the reforming step (Reformax® 330). 99.7 % conversion and a H2 yield of 93.45 % are achieved 
at 600 ºC, 0.28 gcatalyst h gbiomass-1 and S/C ratio of 8.2, producing 11.2 g of hydrogen per 100 g of biomass fed 
into the process. Increasing reaction time, higher coke contents are deposited on the catalyst due to 
secondary reactions. This is the main cause of catalyst activity decrease, although deactivation is attenuated 
by the good performance of the fluidized bed reactor and the excess of steam in the reaction medium (high 
S/C ratio). 

1. Introduction 

Hydrogen is regarded as an alternative energy carrier and has received much attention due to its several 
advantages, such as near-zero carbon emissions at the point of use (Acar and Dincer, 2014). Moreover, it is a 
raw material which is widely used for oil refineries and ammonia and methanol production, and nowadays is 
still mainly produced by reforming of natural gas and oil derived feedstocks (Balat and Kirtay, 2010).  
In order to satisfy the growing demand of hydrogen and reduce greenhouse gas emissions to the atmosphere, 
other alternative routes from renewable sources are becoming essential for the near future. Among the 
different renewable sources available, biomass is accepted as the best route for hydrogen production in terms 
of environmental impact criteria, since it does not contribute to a net increase in atmospheric carbon dioxide 
(Gil et al., 2014).  
Gasification and pyrolysis are two effective methods to convert biomass into hydrogen (Alvarez et al., 2014a), 
although pyrolysis is energetically more efficient due to the lower temperature required for the process. Flash 
pyrolysis of biomass allows obtaining a gas yield of 10-20 %, bio-oil yield of 60-80 % and char yield of 15-25 
%, depending on the type of biomass and reactor used and operating conditions (Amutio et al., 2012). In order 
to obtain a hydrogen-rich gas, a catalytic steam reforming step is required subsequent to the pyrolysis 
process. Fixed bed reactors have been extensively used for reforming processes (Bimbela et al., 2013), 
although some authors have proven that fluidized beds are more suitable. The maximum hydrogen yield 
obtained by Lan et al. (2010) in a fluidized bed reactor has been 7 % higher than that in a fixed bed. 
Catalytic reforming of bio-oil aqueous fraction and catalytic reforming of raw bio-oil have been extensively 
investigated in the literature. Several modified Ni-Al catalysts have been studied by Yao et al. (2014) for the 
catalytic reforming of the bio-oil aqueous fraction, obtaining a maximum hydrogen yield of 56.46 % of the 
stoichiometric one by using a Ni-Mg-Al catalyst. Remiro et al. (2013) studied the steam reforming of the raw 
bio-oil in a fluidized bed reactor on a Ni/La2O3-αAl2O3 catalyst and have obtained a high bio-oil conversion 
(>80 %) and a maximum hydrogen yield of 95 % of the stoichiometric one in the 600-800 ºC range. Ni-based 
catalysts are the most common catalysts for reforming bio-oil aqueous fraction (Remón et al., 2014), raw bio-

                                

 
 

 

 
   

                                                  
DOI: 10.3303/CET1543092 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Please cite this article as: Arregi A., Barbarias I., Alvarez J., Erkiaga A., Artetxe M., Amutio M., Olazar M., 2015, Hydrogen production from 
biomass pyrolysis and in-line catalytic steam reforming, Chemical Engineering Transactions, 43, 547-552  DOI: 10.3303/CET1543092

547



oil (Remiro et al., 2013) or ethanol (Montero et al., 2014), although noble metals, such as Pt, Pd and Rh, can 
also be used for reforming oxygenated compounds (Rioche et al., 2005). 
Nevertheless, studies dealing with the catalytic reforming in-line of biomass pyrolysis volatiles are rather 
scarce. A two-stage continuous screw-kiln reactor has been used for this process by Efika et al. (2012), 
obtaining a maximum H2 production of 44.4 vol. % when a NiO/Al2O3 catalyst was used.  
In this study, continuous pyrolysis of biomass and in-line catalytic reforming of volatile products has been 
performed in a two-stage system, using a commercial Ni reforming catalyst in the second step. This process 
allows valorizing both the whole bio-oil and gases from the pyrolysis step, avoiding additional costs and bio-oil 
vaporization operational problems that occur during the bio-oil reforming process in one step. The aim of the 
work is to ensure the good performance of the two-stage system and study the product distribution obtained 
and catalyst deactivation. 

2. Experimental 

2.1 Raw material and catalyst 

The raw material used in this study is forest pinewood waste (Pinus insignis) with a particle size between 1 
and 2 mm and moisture content below 10 %. Its main characteristics are summarized in Table 1.  

Table 1: Pinewood sawdust characterization (on a wet basis) 

Ultimate analysis (wt. %) 
          Carbon 
          Hydrogen 
          Nitrogen 
          Oxygen 
Proximate analysis (wt. %) 
          Volatile matter 
          Fixed carbon 
          Ash 
          Moisture 
Calorific value (MJ/kg) 

 
49.33 
6.06 
0.04 
44.57 
 
73.4 
16.7 
0.5 
9.4 
19.8 

 
The catalyst (Reformax® 330) was supplied by Süd Chemie in the shape of perforated rings (19 x 16 mm) and 
consists of a Ni metal phase supported on Al2O3 which is doped with Ca. The chemical formulation of the 
catalyst is NiO, CaAl3O4 and Al2O3. The catalyst rings have been crushed and sieved to particles with a 
diameter of 0.4-0.8 mm and have been reduced by introducing a 10:90 % vol. mixture of H2 and N2 at 710 ºC 
for 4 hours. 

2.2 Experimental unit and procedure 

The experimental equipment used to perform the pyrolysis-catalytic steam reforming of biomass is shown in 
Figure 1. The system for solid feeding consists of a vessel equipped with a vertical shaft connected to a piston 
placed below the material bed. The biomass is fed into the reactor by raising the piston at the same time as 
the whole system is vibrated by an electric engine. The tube that connects the feeding system with the reactor 
is cooled with tap water and, in order to avoid the condensation of steam in the feeding system, a small 
nitrogen flow is introduced from the top of the feeding vessel. Water has been fed by means of a Gilson 307 
pump and vaporized by an electric cartridge. Other gases, such as nitrogen, air or hydrogen, can also be 
introduced from the bottom of the pyrolysis reactor, and their flows are controlled by mass flow controllers. 
Prior to entering the reactor, a gas preheater raises the gases and steam temperature to the reaction 
temperature. 
The main element of the pilot plant is a conical spouted bed reactor where the pyrolysis of biomass has been 
carried out. This reactor has been designed with the knowledge acquired in the pyrolysis of different waste 
materials, such as tyres (Lopez et al., 2009), plastics (Artetxe et al., 2010) and different types of biomass like 
pine sawdust (Amutio et al., 2012) or rice husk (Alvarez et al., 2014b). The volatile products leave the reactor 
and pass through a high-efficiency cyclone in order to remove the char particles.  
The reforming step has been carried out in a fluidized bed reactor. The gases leaving the reactor circulate 
through a volatile condensation system consisting of a condenser and a coalescence filter. The condenser is a 
double shell tube cooled by tap water and the coalescence filter removes completely the liquid from the 
gaseous stream.  

548



The volatile stream has been analysed on-line by means of a GC (Agilent 6890). The sample has been 
injected into a GC through a line thermostated at 280 ºC in order to avoid the condensation of heavy 
oxygenated compounds. Moreover, the non-condensable gases have been analysed in a micro GC (Varian 
4900). 

 

Figure 1: Scheme of the pyrolysis-catalytic steam reforming unit 

The biomass and steam have been continuously fed into the reactor with rates of 0.75 g min-1 and 3 mL min-1, 
respectively, corresponding to a steam/biomass ratio of 4 (S/C=8.2). The pyrolysis step has been carried out 
at 500 ºC and the bed was made up of 50 g of sand with a particle diameter in the 0.3-0.355 mm range in 
order to guarantee a vigorous movement and bed isothermicity. The reforming step has been carried out at 
600 ºC and with a space-time of 0.28 gcatalyst h gbiomass-1, using 12.5 g of catalyst (particle diameter in the 0.4-
0.8 mm range) and 12.5 g of sand (particle diameter in the 0.3-0.355 mm range).  

3.  Results 

3.1 Reforming indices 

In order to quantify the products of the reforming process, conversion and product yields have been defined as 
reaction indices. The conversion is calculated as the ratio of molar flow of C from the biomass pyrolysis 
volatiles converted to the gas fraction in the reforming step (Eq (1)).  

x100
feed the in C offlow  molar

fraction gas the in C offlow  molar
X =  (1) 

Moreover, hydrogen yield is calculated as percentage of the maximum allowed by stoichiometry (Eq (2)) and 
the yields of CO, CO2, CH4 and light HCs (ethylene, ethane, propylene, propane) are calculated as the ratio 
between the molar flow of C in each one of these components and the molar flow of C in the feed (Eq (3)). 

x100
trystoichiomeby  allowed maximum

obtained H offlow  molar
Y

2

2H
=  (2) 

x100
feed the in C offlow  molar

obtained HCs) ,CO CO, ,(CO i in C offlow  molar
Y 42i =  (3) 

549



3.2 Catalytic steam reforming of volatiles from biomass pyrolysis 

Firstly, the limited reactivity steam has at low temperatures (500 ºC) has been proven in the pyrolysis step, 
i.e., the yields for the different product fractions are similar to those obtained by Amutio et al. (2012) in a 
previous paper in which nitrogen was used as fluidizing agent. A gas yield of 7.33 %, bio-oil yield of 75.33 % 
and char yield of 17.34 % have been obtained at 500 ºC. The gas is mainly composed of carbon monoxide 
and carbon dioxide and water is the main compound in the liquid fraction (33.7 %). Among the various types of 
oxygenated compounds in the bio-oil, phenols are the main products, which are formed in the decomposition 
of lignin macromolecules above 350 ºC (Alvarez et al., 2014b).  
Table 2 shows the conversion, the initial product yields and the initial gas composition obtained in the 
reforming step. As observed, almost all volatiles are reformed into gases and a hydrogen yield of 93.45 % of 
the stoichiometric one is achieved under the conditions studied. The high conversion obtained evidences the 
good performance of the conical spouted bed and fluidized bed reactors (high heat and mass transfer rates). 
Therefore, the two-step system has proven to be suitable for pyrolysis-catalytic steam reforming of biomass, 
obtaining 11.2 g of hydrogen per 100 g of biomass fed into the process. Full conversion and similar hydrogen 
yield (95 %) have been obtained by Remiro et al. (2013) in the steam reforming of the raw bio-oil at 700 ºC, 
with 0.14 gcatalyst h gbio-oil-1 (Ni/La2O3-αAl2O3) and a S/C ratio of 9 with prior separation of pyrolytic lignin.  

Table 2: Results of conversion (%), product yields (%) and gas composition (% vol.) at zero time on stream. 
Reforming conditions: T=600 ºC, space-time= 0.28 gcatalyst h gbiomass-1, S/C=8.2 

X (%) 
Yields (%) 
          YH2 
          YCO2 
          YCO 
          YCH4 
          YHCs 
Gas composition (% vol.) 
          H2 
          CO2 
          CO 
          CH4 
          HCs 

99.72

93.45
88.55
10.58
0.56
0.04

66.06
30.15
3.60
0.19
0.01

 
Figure 2 shows the evolution of conversion with time on stream. It is noteworthy that catalyst deactivation is 
only significant for times on stream longer than 60 min, with the decrease in activity being pronounced 
subsequent to 75 min. The deactivation is due to blocking of active centres by carbon deposition, which has 
been partially gasified favoured by the reactor hydrodynamics and operating conditions (high S/C ratio). Lan et 
al. (2010) have proven that carbon deposition in fluidized beds is lower than in fixed beds at the same reaction 
temperature. 

 

Figure 2: Evolution of conversion with time on stream. Reforming conditions: T=600 ºC, space-time= 0.28 
gcatalyst h g biomass-1, S/C=8.2 

550



Regarding gas composition, it is mainly composed of H2 and CO2 (reforming products) and small amounts of 
CO, CH4 and light HCs (Table 2). Similar values of H2 concentration have been obtained by Hou et al. (2009) 
and Seyedeyn-Azad et al. (2012) in the reforming of raw bio-oil. In addition, gas composition is slightly 
affected by coke formation on the catalyst, as shown in Figure 3. H2 concentration decreases slightly from 
66.06 % to 63.51 % after 106 min time on stream due to the decrease in the activity of the catalyst for 
reforming (Eq (4)) and WGS reactions (Eq (5)), whereas there is no significant change in CO2 concentration 
(around 30 %). 

2Hk
2

m
nnCOO2k)H(nkOmHnC 








−++→−+  (4) 

2H2COO2HCO +↔+  (5) 
When the catalyst is severely deactivated, the cracking reactions of oxygenated compounds become more 
significant, increasing CO, CH4 and light hydrocarbon (ethylene, ethane, propylene and propane) 
concentrations. CO concentration increases from 3.60 % at zero time on stream to 4.93 % at around 106 min, 
while CH4 concentration increases from 0.19 % to 0.64 %. Even when the catalyst is deactivated, light 
hydrocarbon concentration does not exceed 0.13 %. Ethylene is the major compound, followed by ethane and 
propylene. Propane concentration is negligible throughout 106 min time on stream (<0.01 %). A similar trend 
has been observed by Remiro et al. (2013) in the reforming of raw bio-oil, obtaining higher yields of CO, CH4 
and light hydrocarbons when the catalyst is deactivated. 
 

 

Figure 3: Evolution of gas composition (vol. %) with time on stream: a) H2, CO2, CO, CH4 and b) ethylene, 
ethane, propylene, propane. Reforming conditions: T=600 ºC, space-time= 0.28 gcatalyst h g biomass-1, S/C=8.2 

4. Conclusions  

The two-step reaction system has proven to be suitable for the reforming of volatiles from biomass pyrolysis. 
The conical spouted bed reactor performs well in the continuous pyrolysis of pine sawdust, given its capacity 

551



for operating in isothermal regime. In addition, the use of a fluidized bed reactor and high S/C ratios in the 
process helps mitigating coke formation on the catalyst, given that gasification is enhanced under these 
conditions. 
The catalyst has a high activity, which allows a high initial transformation of the volatiles from biomass 
pyrolysis into H2 and CO2 at 600 ºC, a space-time of 0.28 gcatalyst h gbiomass-1 and a S/C ratio of 8.2. The catalyst 
maintains its activity for 75 minutes and high conversion and hydrogen yield are obtained. After this time on 
stream, reforming and WGS reactions follow a decreasing trend and secondary reactions take place, 
producing higher yields of CO, CH4 and light hydrocarbons. 
 

References 

Acar C., Dincer I., 2014, Comparative assessment of hydrogen production methods from renewable and non-
renewable sources, Int. J. Hydrogen Energy, 39, 1-12. 

Alvarez J., Kumagai S., Wu C., Yoshioka T., Bilbao J., Olazar M., Williams P.T., 2014a, Hydrogen production 
from biomass and plastic mixtures by pyrolysis-gasification, Int. J. Hydrogen Energy, 39, 10883-10891. 

Alvarez J., Lopez G., Amutio M., Bilbao J., Olazar M., 2014b, Bio-oil production from rice husk fast pyrolysis in 
a conical spouted bed reactor, Fuel, 128, 162-169. 

Amutio M., Lopez G., Artetxe M., Elordi G., Olazar M., Bilbao J., 2012, Influence of temperature on biomass 
pyrolysis in a conical spouted bed reactor, Resour. Conserv. Recycl., 59, 23-31. 

Artetxe M., Lopez G., Amutio M., Elordi G., Olazar M., Bilbao J., 2010, Operating conditions for the pyrolysis 
of poly-(ethylene terephthalate) in a conical spouted bed reactor, Ind. Eng. Chem. Res., 49, 2064-2069. 

Balat H., Kirtay E., 2010, Hydrogen from biomass – Present scenario and future prospects, Int. J. Hydrogen 
Energy, 35, 7416-7426. 

Bimbela F., Oliva M., Ruiz J., García L., Arauzo J., 2013, Hydrogen production via catalytic steam reforming of 
the aqueous fraction of bio-oil using nickel-based catalysts, Int. J. Hydrogen Energy, 38, 14476-14487. 

Efika C.E., Wu C., Williams P.T., 2012, Syngas production from pyrolysis-catalytic steam reforming of waste 
biomass in a continuous screw kiln reactor, J. Anal. Appl. Pyrolysis, 95, 87-94. 

Gil M.V., Fermoso J., Rubiera F., Chen D., 2014, H2 production by sorption enhanced steam reforming of 
biomass-derived bio-oil in a fluidized bed reactor: An assessment of the effect of operation variables using 
response surface methodology, Catal. Today, DOI: 10.1016/j.cattod.2014.04.018. 

Hou T., Yuan L., Ye T., Gong L., Tu J., Yamamoto M., Torimoto Y., Li Q., 2009, Hydrogen production by low-
temperature reforming of organic compounds in bio-oil over a CNT-promoting Ni catalyst, Int. J. Hydrogen 
Energy, 34, 9095-9107. 

Lan P., Xu Q., Zhou M., Lan L., Zhang S., Yan Y., 2010, Catalytic steam reforming of fast pyrolysis bio-oil in 
fixed bed and fluidized bed reactors, Chem. Eng. Technol., 33, 2021-2028. 

Lopez G., Olazar M., Amutio M., Aguado R., Bilbao J., 2009, Influence of tire formulation on the products of 
continuous pyrolysis in a conical spouted bed reactor, Energy Fuels, 23, 5423-5431. 

Montero C., Valle B., Bilbao J., Gayubo A., 2014, Analysis of Ni/La2O3-αAl2O3 catalyst deactivation by coke 
deposition in the ethanol steam reforming, Chemical Engineering Transactions, 37, 481-486. 

Remiro A., Valle B., Aguayo A.T., Bilbao J., Gayubo A.G., 2013, Steam reforming of raw bio-oil in a fluidized 
bed reactor with prior separation of pyrolytic lignin, Energy Fuels, 27, 7549-7559. 

Remón J., Broust F., Valette J., Chhiti Y., Alava I., Fernandez-Akarregi A.R., Arauzo J., Garcia L., 2014, 
Production of hydrogen-rich gas from fast pyrolysis bio-oils: Comparison between homogeneous and 
catalytic steam reforming routes, Int. J. Hydrogen Energy, 39, 171-182. 

Rioche C., Kulkarni S., Meunier F.C., Breen J.P., Burch R., 2005, Steam reforming of model compounds and 
fast pyrolysis bio-oil on supported noble metal catalysts, Appl. Catal., B, 61, 130-139. 

Seyedeyn-Azad F., Abedi J., Salehi E., Harding T., 2012, Production of hydrogen via steam reforming of bio-
oil over Ni-based catalysts: Effect of support, Chem. Eng. J., 180, 145-150. 

Yao D., Wu C., Yang H., Hu Q., Nahil M.A., Chen H., Williams P.T., 2014, Hydrogen production from catalytic 
reforming of the aqueous fraction of pyrolysis bio-oil with modified Ni-Al catalysts, Int. J. Hydrogen Energy, 
39, 14642-14652. 

552