+ All Categories
Home > Documents > Hydrogen Production by Methane Decomposition - A Review 2010

Hydrogen Production by Methane Decomposition - A Review 2010

Date post: 05-Jul-2018
Category:
Upload: cristian
View: 218 times
Download: 0 times
Share this document with a friend

of 31

Transcript
  • 8/16/2019 Hydrogen Production by Methane Decomposition - A Review 2010

    1/31

    Review

    Hydrogen production by methane decomposition: A review

    Hazzim F. Abbas*, W.M.A. Wan Daud

    Department of Chemical Engineering, University of Malaya, 50603 Kuala Lumpur, Malaysia

    a r t i c l e i n f o

    Article history:

    Received 19 September 2009

    Received in revised form

    9 November 2009

    Accepted 9 November 2009

    Available online 27 November 2009

    Keywords:

    Methane decomposition

    Hydrogen production

    Metal catalysts

    Carbonaceous catalysts

    Deactivation

    Regeneration

    a b s t r a c t

    Methane decomposition can be utilized to produce COX-free hydrogen for PEM fuel cells, oil

    refineries, ammonia and methanol production. Recent research has focused on enhancing 

    theproduction of hydrogen by thedirect thermocatalytic decomposition of methaneto form

    elemental carbon and hydrogen as an attractive alternative to the conventional steam-

    reforming process. In this context, we review a comprehensive body of work focused on the

    development of metal or carbonaceous catalysts for enhanced methane conversion and on

    the improvement of long-term catalyst stability. This review also evaluates the roles played

    by various parameters, such as temperature and flow rate, on the rate of hydrogen

    production and the characteristics of the carbon produced. The heating source, type of 

    reactor, operating conditions, catalyst type and its preparation, deactivation and regener-

    ation and the formation and utilization of the carbon by-product are discussed and classi-

    fied in this paper. While other hydrogen production methods, economic aspects and

    thermal methane decomposition methods using alternative heating sources such as solar

    and plasma are briefly presented in this work where relevant, the review focuses mainly on

    the thermocatalytic decomposition of methane using metal and carbonaceous catalysts.

    ª 2009 Professor T. Nejat Veziroglu. Published by Elsevier Ltd. All rights reserved.

    1. Introduction

    One of the major challenges posed by the continuous increase

    in global population and economic development is providing 

    more energy while limiting greenhouse-gas (GHG) emissions.

    The dramatic increase in the concentrations of carbon

    dioxide, methane and nitrous oxides, the pH decrease of theocean surface and atmospheric temperature increase show

    that human activity affects geochemical systems on a global

    scale. The main sources of GHG emissions are due to the

    combustion of natural gas (NG), coal and oil for heating,

    electricity production, transportation and industrial purposes.

    For example, oil accounts for 39% of hydrocarbon-related CO2emissions and NG for 20%, with coal accounting for the

    remaining. The more efficient use of fuel has focused current

    research on new energy sources that do not emit GHGs as well

    as with the capture GHGs from the burning of hydrocarbon

    fuels; both might be effective ways to gradually decrease the

    quantity of GHG emissions [1,2].

    Hydrogen appears to be one of the most promising energy

    vectors as it is considered to be environmentally benign. The

    amount of energy produced during hydrogen combustion ishigher than that evolved by any other fuel on a mass basis,

    with a low heating value that is 2.4, 2.8 or 4 times higher than

    that of methane, gasoline or coal, respectively. Currently, the

    annualproduction of hydrogen is about 0.1 Gton, 98% of which

    is from the reforming of fossil fuels; it is used mainly in oil

    refineries, ammonia and methanol production   [3]. The fuel

    cell (a device transforming chemical energy into electricity

    and heat) is a rapidly emerging technology. One area of great

    *   Corresponding author. Tel.: þ60 172907256; fax:  þ60 103 79675371.E-mail address: [email protected] (H.F. Abbas).

    A v a i l a b l e a t w w w . s c i e n c e d i r e c t . c o m

    j o u r n a l h o m e p a g e :   w w w . e l s e v i e r . c om / l o c a t e / h e

    i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 1 1 6 0 – 1 1 9 0

    0360-3199/$ – see front matter  ª  2009 Professor T. Nejat Veziroglu. Published by Elsevier Ltd. All rights reserved.

    doi:10.1016/j.ijhydene.2009.11.036

    mailto:[email protected]://www.elsevier.com/locate/hehttp://www.elsevier.com/locate/hemailto:[email protected]

  • 8/16/2019 Hydrogen Production by Methane Decomposition - A Review 2010

    2/31

    potential for fuel cells is in powering automobiles; the

    advantages of fuel cells are that they are quiet, do not contain

    any hazardous material, and therefore can be classified as an

    environmentally benign power source   [4]. Research iscurrently being conducted for the development of safe, cost-

    effective hydrogen production, storage and use technologies

    that support and foster the use of hydrogen. Hydrogen

    production from fossil fuels will continue for the foreseeable

    future, given the large resource and the established industrial

    base, and a number of papers have address advanced

    hydrogen production technologies that reduce or eliminate

    CO2 emissions in the production process [5].

    In this review of the literature, reports relating to the direct

    methane decomposition process for hydrogen production are

    classified and summarized according to heating sources,

    reactor types, catalyst (metal or carbonaceous) types, their

    deactivation and regeneration and the form of the carbon by-product. It should be noted that there were also studies

    focused on methane decomposition reactions using different

    metal catalysts with the aim of produce carbon filaments (CF),

    which is favored by operating conditions yielding low rates of 

    methane decomposition. These works were excluded from

    the review because the focus here was on hydrogen

    production.

    2. Methods of hydrogen production:economic and environmental aspects

    The current methods of producing hydrogen are based onsteam methane reforming (SMR), coal gasification, electrol-

    ysis, biomass gasification and thermochemical processes.

    Table 1 lists a summary of the energy efficiencies and costs of 

    hydrogen production for several methods [6].

    A comparison between the cost of hydrogen production

    from renewable sources, such as from water electrolysis using 

    renewable energy, and fossil sources, currently mainly

    natural gas (NG), shows that the cost of renewably sourced

    hydrogen must be considerably reduced from present levels

    before this type of hydrogen becomes economically competi-

    tive  [7]. Currently, the process based on SMR and methane

    partial oxidation generates large quantities of CO2; the esti-

    mated GHG potential of hydrogen production by the SMR

    process is approximately 13.7 kg CO2 (equiv.)/kg of hydrogen

    produced [8].

    Methane can be thermally or thermocatalytically decom-

    posed into carbon and hydrogen without producing CO2, andthis hydrogen production method has recently attracted the

    attentions of researchers. Lane and Spath [9]  estimated that

    hydrogen could be produced by the thermocatalytic decom-

    position (TCD) of methane at a selling price of (7–21) $/GJ

    (Note: 1 GJ ¼ 1.05461 MMBtu) depending on the cost of NG and

    the selling price of carbon. In 1988, a techno-economic

    assessment showed that the cost of hydrogen produced by

    the thermal decomposition of NG, at $58/1000 m3 H2   (with

    carbon credit; based on the low heating value of hydrogen,

    this is equivalent to $5.38/GJ), was somewhat lower than that

    for SMR, at $67/1000 m3 H2   (equivalent to $6.21/GJ)   [10].

    Steinberg   [11]   made a comparison between the SMR and

    methane TCD processes for the decarbonization of NG andshowed that the TCD of methane appears to have several

    advantages over a well developed SMR process, as the major

    one being that it is much easier to sequester carbon in form

    of the stable solid produced by methane TCD rather than as

    the CO2 produced as a reactive gas or low-temperature liquid

    from the SMR process. Dufour et al.   [12] compared different

    processes for hydrogen production (SMR, SMR with CO2capture and storage, methane autocatalytic decomposition

    and methane thermal cracking) using life-cycle assessment

    tools to evaluate their relative environmental feasibilities and

    CO2   emissions. They reported that autocatalytic decomposi-

    tion is the most environmentally friendly process for

    hydrogen production as it presented the lowest total envi-ronmental impact and CO2 emission; although SMR with CO2capture and storage actually led to lower CO2  emissions, it

    still had a higher total environmental impact than conven-

    tional SMR.

    Fig. 1 shows a comparison between the TCD of methane

    and the commercial conventional SMR process. The energy

    input requirements per mole of hydrogen for TCD is signifi-

    cantly less than that of SMR (37.8 and 63.3 kJ/mol H2, respec-

    tively). In terms of net hydrogen yield, although the

    theoretical hydrogen yield for SMR is twice that of TCD (four

    vs.2 mol of H2 per mole of CH4, respectively), the high reaction

    endothermicity and required CO2 sequestration process, both

    consuming significant amount of energy in methane

    Table 1 – List of hydrogen production technologies and costs  [6].

    Production technology Energy efficiencya Hydrogen selling priceb ($/MMBtu)

    Hydrogen selling price ($/kg)

    SMR 83% 5.54 0.75

    Partial oxidation

    of methane

    70–80% 7.32 0.98

    Autothermal reforming 71–74% 16.88 1.93Coal gasification 63% 6.83 0.92

    Direct biomass gasification 40–50% 9–18 1.21–2.42

    Electrolysis (nuclear fission-powered) 45–55% 14.5 1.95

    Photocatalytic water splitting 10–14% 37 4.98

    c Assumes an NG price of $3.15/MMBtu [6].

    a Energy efficiency is defined as the energy value of the hydrogen produced divided by the energy input required to produce the hydrogen.

    b The hydrogen selling price for SMR and coal gasification does not consider CO2 sequestration costs.

    i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 1 1 6 0 – 1 1 9 0   1161

    http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-

  • 8/16/2019 Hydrogen Production by Methane Decomposition - A Review 2010

    3/31

    equivalents, would considerably reduce the net yield of 

    hydrogen produced by a ‘‘CO2-neutral’’ SMR process [13].

    3. Thermocatalytic methane decompositionprocesses

    Methane can be decomposed into carbon and hydrogen

    according to the following reaction:

    CH4/CðSÞ þ 2H2   DH

    ¼ 75:6kJ=mol

    Because the process does not produce CO or CO2   as by-products, the need for the water-gas shift and CO2-removal

    stages, as required in conventional hydrogen production

    methods (e.g., SMR, coal gasification and partial oxidation), is

    eliminated. Additionally, it is possible to reduce the cost of 

    hydrogen produced by methane TCD by marketing the carbon

    as a filler or construction material [14].

    3.1. Heating sources

    Methane decomposition is a moderately endothermic reac-

    tion. Due to the strong C–H bonds, non-catalytic thermal

    cracking of methane requires temperatures higher than

    1200   C obtain a reasonable yield. By using a catalyst, thetemperature can be significantly reduced, depending on the

    type of catalyst used. There are many reports on the use of an

    electrical furnace as a heating source for the TCD reactor,

    while there are rather few focusing on the use of concentrated

    solar energy, plasma or a molten-metal bath as alternative

    heating sources. A recent report [15] pointed out that solar is

    most likely to be the only source of energy capable of 

    producing very large volumes of hydrogen. Using solar

    thermal power, methane thermal dissociation into hydrogen

    and carbon black (CB), two valuable products, has been

    investigated   [16–21]. Concentrated solar energy is a clean

    source of high temperature process heat and direct solar

    irradiation of the reactants provides for a very efficient heat

    transfer directly to the reaction site   [22]. However, solar

    heating cannot be achieved directly because hydrocarbons

    absorb radiation in the visible spectrum only poorly. To

    overcome this problem, the use of a transparent window that

    permits direct heating of particulate material by radiation or

    opaque heat-transferring reactor walls that absorb the solar

    radiation on one side and then heat the gas through convec-

    tion on the other side have been employed.Beside a solar source, plasma may also be employed as an

    environmentally friendly heating source. To be effective, it

    needs to be carried out at a very high reaction temperature,

    which, with recent improvements in plasma technology, is

    now accessible [23]. Microwave (MW) plasma (generated using 

    an MW-frequency signal; typically 2.45 GHz), commonly used

    in MW ovens, diamond vapor deposition and IC

    manufacturing, has the advantages of easy operation, an

    electrodeless reactor, high plasma density and high electron

    mean energy [24]. The major advantage of this new process is

    the total conversion of the hydrocarbon into CB and hydrogen

    (100% carbon yields and the production of new carbon grades

    at higher temperature)   [25]. Recently, using carbon aerosolsgenerated by a non-thermal plasma, Muradov et al.   [26]

    showed that TCD can be accomplished over catalytically

    active carbon aerosol particles at temperatures comparable to

    that of the conventional SMR process, i.e., 850–900   C,

    a temperature range about 400–500 C lower than that for non-

    catalytic methane decomposition. Application of pulsed MW

    power was shown to be a promising heating source. However,

    the choice of the catalyst for an MW-driven processappears to

    be of great importance as it must meet some specific

    requirements in addition to the common demands of high

    activity, selectivity and stability. Such a catalyst must be

    a good receptor of MW energy and able to retain its structure

    and properties under intense MW radiation [27]. MW energyoffers a number of advantages over conventional heating:

    (i) noncontact heating; (ii) energy transfer without heat

    transfer; (iii) rapid heating; (v) volumetric heating; (vi) quick

    start-up and shutdown; (vii) the heating starts from the inte-

    rior of the material body; (viii) a higher level of safety and

    automation; and (ix) heating energy can be transported from

    the source through a hollow, nonmagnetic metal tube.

    Muradov   [28]   discussed three heating-arrangement

    options for the decomposition reaction (Fig. 2). In the first

    option, the heat source is located inside the reaction zone

    using a heat pipe, such as a heat exchanger or a catalytic

    burner that uses NG or a portion of the hydrogen product as

    a fuel. In the second option, externally heated catalyst parti-cles are used as the heat carrier, which is similar to the

    fluidized catalytic-cracking process currently used in many

    refineries. In the third option, a relatively small amount of 

    oxygen is added to the methane feedstock to generate the

    necessary heat, in a process termed autothermal pyrolysis.

    Table 2   lists and compared the different types of heating 

    sources employed to date for methane TCD.

    3.2. Reactor types

    Besides the heating source, another important factor is the

    type of reactor equipment used. Fluidized-bed reactors (FBR)

    and packed-bed reactors (PBR) are the most commonly used

    Fig. 1 – Comparative assessment of net hydrogen yields for

    SMR and TCD processes [13].

    i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 1 1 6 0 – 1 1 9 01162

  • 8/16/2019 Hydrogen Production by Methane Decomposition - A Review 2010

    4/31

    reactors for methane TCD. Since the main products of 

    methane TCD are solid carbon and hydrogen gas, the problem

    with using a PBR is carbon deposition over the external

    surface of the catalyst particles. As the reaction proceeds, the

    pressure drop increases, the catalyst particle size increases

    and particle shape and density are altered. For experiments

    conducted over a long duration, a PBR will be gradually filled

    with solid carbon, eventually blocking reactant gas flow [29–

    31].To maintain catalyst activity and to avoid plugging of the

    reactor, the deposited carbon must be removed periodically.

    The development of a reactor type that is efficient for

    methane TCD with continuous withdrawal of carbon product

    had been conducted by Muradov [13,32]. He studied different

    types of reactors (PBR, FBR, free-volume reactor, spouted-bed

    and tubular reactors) and found that the FBR was the most

    promising reactor for large-scale operation since it provides

    a constant flow of solids through the reaction zone, making it

    suitable for continuous addition and withdrawal of catalyst

    particles from the reactor. Additionally, the vigorous particle

    motion caused by fluidization increases both heat- and mass-

    transfer rates. The FBR superficial velocity used was in the

    range of one to four times the minimum fluidization velocity

    (Umf ). Fig. 3 shows some types of reactor used for catalytic and

    non-catalytic thermal decomposition of methane.

    One method of studying TCD parameters consists of using 

    a thermobalance to obtain rates of methane decomposition by

    directly measuring the mass gain (due to carbon deposition)

    with time. The advantages of using a thermobalance to study

    methane TCD are as follows: (1) the ability to directly measure

    the change of mass with time; as a result, the initial rate of methane decomposition (rO) can be used to determine the

    intrinsic kinetic parameters and the change in catalytic

    activity as the mass of deposited carbon increases; (2) a small

    quantity of catalyst can be used to ensure that the endo-

    thermicity of the decomposition reaction does not cause large

    temperature fluctuations in the catalyst and negligible pres-

    sure drop. The main drawbacks of using a thermobalance are

    diffusion limitations and the difficulties of controlling space-

    time due to the changes in sample volumes during the

    experiments [37–40].

    In most experimental work carried out for methane TCD

    quartz was used as a reactor-construction material to avoid

    any possible catalytic activity of the reactor wall on methane

    Fig. 2 – Schematic diagram of hydrogen and carbon production via catalytic decomposition of NG with an internal (A) or

    external (B) heat supply. 1. Fluidized-bed reactor (FBR); 2. Fluidized-bed heater-regenerator; 3. Internal heaters; 4. Cyclones;

    5. Heat exchangers; 6. Gas-separation unit; 7. Carbon/metal catalyst separation unit; 8. Catalyst regeneration unit; 9. Carbonproduct separation and conditioning unit; 10. CO2 scrubber and gas-purification system; 11. Combustion chamber  [28].

    i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 1 1 6 0 – 1 1 9 0   1163

  • 8/16/2019 Hydrogen Production by Methane Decomposition - A Review 2010

    5/31

    Table 2 – Recent methane TCD studies using different heating sources and reactor types.

    Researcher Heating source Reactor type

    Abanades and

    Flamant [43,44]

    Concentrated solar energy. Fluid-wall reactor, Stainless-steel reactor

    body and with internal graphite tube.

    * No catalyst used.

    * Temperature measured with IR solar-blind optical pyrometer.

    * Vertical-axis solar furnace (200-cm diameter concentrator) used.* Co-products are hydrogen-rich gas and high-grade CB.

    * Methane conversions of 88% and 42% were measured for a 2000-kW/m2 solar-flux density.

    Hirsch and Steinfeld [33]   Concentrated solar energy Vortex-flow reactor, 10-cm diameter

    and 20-cm length, made from steel

    alloy with a 6-cm diameter aperture.

    * Vortex-flow of methane confined to a cavity receiver and laden with carbon particles that served simultaneously

    as radiant absorbers and nucleation sites for the heterogeneous decomposition reaction.

    * A 5-kW reactor prototype with solar-flux intensities exceeding 3500 kW/m2 was used.

    * Conversion of methane was 67% at 1327  C and 1 atm.

    Dahl et al. [45]   Concentrated solar energy Quartz-tube reactor, 2.5-cm diameter,

    consisting of a feed mechanism with

    internal graphite ‘‘target’’ feed tube.

    * Series of mirrors used to concentrate sunlight into a focused beam at a maximum level of 10 kW onto anapproximate diameter of 10 cm.

    * Secondary concentrator used to obtain higher temperature.

    * Quartz reactor tube illuminated with a solar flux of 2400 kW/m2 resulted in a methane conversion of 90%.

    Dahl et al. [34]   Concentrated solar energy Quartz fluid-w all aerosol-flow reactor

    composed of three concentric vertical tubes.

    The innermost tube was 1.2-cm i.d., long,

    porous graphite, the center tube was a

    2.1-cm i.d., solid graphite tube 360 cm long.

    * No evidence of corrosion or erosion of the reactor tube reported.

    * 90% conversion of methane was obtained at 1860  C and a residence time of 0.01 s.

    * Due to the small size of reactor, CB co-feed did not enhance the heat transfer.

    Maag et al. [46]   Concentrated solar energy 5-kW solar chemical reactor consisting 

    of a 20-cm-long, 10-cm-diameter cylindricalcavity receiver containing a 6-cm-diameter

    circular aperture to allow entry of 

    concentrated solar radiation.

    * Reactor tested at temperature range of 1027–1327  C.

    * Methane conversion and hydrogen yield exceeding 95% were obtained at residence times of less than 2 s.

    * A solar-to-chemical energy conversion efficiency of 16% was experimentally reached.

    * SEM images revealed the formation of filamentous agglomerations on the surface of the seed particles.

    Muradov et al. [47]   Electric furnace FBR, quartz microreactor.

    TCD process could be arranged in continuous process similar industrial to fluidized catalytic-cracking or coking processes.

    Dunker et al. [36]   Three zone electrical

    tubular furnace.

    FBR made from quartz tube, 4.2 cm i.d.

    and 97 cm height

    * Reactor constructed from quartz and the thermocouple sheathed with quartz tube.

    * Fine carbon particles deposited on the reactor walls and polycyclic aromatic hydrocarbon condensed in the lines downstreamof the reactor.

    Aiello et al. [29]   Electric furnace Fixed-bed, quartz microreactor,

    1.8-cm o.d.

    The flow through the reactor was substantially hindered during the eighth and ninth production cycles.

    Fidalgo et al. [48]   Electric furnace and MW oven. Quartz reactor, 45 cm long and 2.2-cm i .d.

    * Activated carbon (AC) was used as both MW receptor and catalyst.

    * Under electric furnace heating, nitrogen distributed the methane molecules within the AC bed.

    * Under MW heating, the nitrogen, as well as distributing the methane molecules, favored the generation of energetic

    microplasmas, leading to higher conversions, with the highest obtained at low CH4 /N2 ratios.

    * The formation of carbon nanofibers (CNFs) was reported when a combination of nitrogen and MW heating was used.

    i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 1 1 6 0 – 1 1 9 01164

  • 8/16/2019 Hydrogen Production by Methane Decomposition - A Review 2010

    6/31

    decomposition (e.g., transition metals have catalytic activityfor methane decomposition). Additionally, the thermocouple

    used was sheathed with a quartz tube. However, there have

    been a few researchers who used kanthal or stainless-steel as

    reactor-construction materials   [30,41,42].   Table 2   presents

    examples of recent studies using different reactor types and

    heating sources along with major findings and notes.

    Different methods can be used to separate the carbon by-

    product from the hydrogen stream. For the CB process, the

    particulate carbon elutriates out of the furnace with the

    hydrogen gas stream and is separated by bag filters. A similar

    system is used in the plasma carbon -arc process. In the

    molten-metal reactor it may be possible to skim the carbon off 

    from the top of the metal surface due to its density difference[11]. Fig. 4 is a simplified diagram showing the types of heating 

    source, reactor, catalyst and technique used to analyze the

    reactor effluent stream and characterize the fresh and deac-

    tivated catalysts in methane TCD studies to date.

    4. Metal catalysts for thermocatalyticdecomposition of methane

    4.1. Operating parameters and kinetics

    As mentioned previously, several types of catalysts have been

    used to reduce the decomposition temperature; it has been

    reported that in the presence of freshly reduced Ni catalyst,hydrogen was detected in the effluent gas at a temperature as

    low as 200  C [5]. Ermakova [53] stated that the methane TCD

    process can hardly be of practical interest unless highly effi-

    cient catalysts are developed. Catalyst efficiency includes not

    only its specific activity but its useful operational lifetime

    given the large amount of carbon accumulation at a practical

    conversion. The use of metal catalysts such as Ni and Ni–Cu

    was first reported by Dudina et al.   [54], Muradov   [55–57],

    Chesnokov et al.   [58,59] and Parmon et al.   [60]. There is no

    general agreement among researchers regarding the relative

    catalytic activities of metals in methane TCD. It has been

    reported that the rate of methane decomposition activity by

    the transition metals follows the order: Co, Ru, Ni, Rh  > Pt, Re,Ir  >  Pd, Cu, W, Fe, Mo  [40]. However, other researchers have

    found that Ni or Ni/alumina and Fe/alumina exhibited the

    highest activities [61].

    The most important factors influencing carbon deposition

    during metal-catalyzed methane decomposition are the

    particle size, dispersion and stabilization of the metallic

    catalyst particles, which are controlled by selecting an

    appropriate support. Takenaka et al.   [62]   reported that

    a typical 40% Ni/SiO2 catalyst with a nickel-particle size of 60–

    100 nm could give carbon yield of as high as 491 g C (gNi) 1

    during methane decomposition at 500   C. Furthermore,

    Ermakova et al.   [63] reported that a 90 wt% Ni/SiO2  catalyst

    with nickel particles of 10–40 nm size provided carbon yields

    Table 2 ( continued ).

    Researcher Heating source Reactor type

    Domı́nguez et al. [49]   MW oven and electrical heating. Fixed-bed, quartz tube, 2.2-cm i.d.

    and 45-cm length.

    * Granular AC was used as a catalyst.

    * Methane conversions were higher under MW conditions than with electrical heating when the temperature was 800  C.

    * The formation of CNFs in one of the MW heating experiments was reported.

    Fulcheri et al. [25]   Plasma power supply (three-phase)

    supplied electricity to three graphite

    electrodes located at the top

    of the reactor.

    Three graphite electrodes located at top

    of reactor, and a graphite nozzle to mix

    feedstock with plasma gas flow.

    Reactor height 250 cm.

    * The reactor internal temperatures were measured at four locations using an optical pyrometer sighted

    on a graphite tube immersed in the plasma flow.

    * Experiment was carried out with electric power varying between 50 and 100 kW.

    * Presented a new alternative to normal furnace processes for CB production.

    Serban et al. [50]   Heat generated in generation

    IV nuclear reactor.

    Stainless-steel reactor, 2.54-cm diameter

    and 35-cm height.

    * Methane was bubbled through a bed of molten metal or granular or catalytic material or a mixture of molten metal and solid media.

    * Bubbling methane through porous metal filters was the most efficient.

    * Buoyant separation of generated carbon from the liquid heat-transfer media was reported.

    Pinilla et al. [51]   Electric furnace. Cylindrical drum rotating around its

    horizontal axis. Diameter 0.065 m,

    length 0.8 m and rotation speed 1–20 rpm.

    * In comparison to FBR, the rotary reactor showed higher hydrogen yields and more sustainable catalyst.

    * Rotation speed did not show any significant influence on the evolution of hydrogen concentration.

    Chesnokov and

    Chichkan [52]

    Electric furnace. Rotary reactor: volume 250 cm3 and

    rotation speed of 1 rpm.

    * Rotary reactor was developed for production of hydrogen and CNF from natural gas.

    * Compared to the reactor with a McBain balance, catalyst-operation time in the rotary reactor was very long.

    i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 1 1 6 0 – 1 1 9 0   1165

  • 8/16/2019 Hydrogen Production by Methane Decomposition - A Review 2010

    7/31

    on the order of 385 g C (gNi)1 at 550  C. Other supports, suchas TiO2, MgO, ZrO2  and Al2O3  gave relatively lower carbon

    yields   [64]. For the production of highly concentrated

    hydrogen during methane decomposition, higher one-pass

    conversion of methane is desirable. Ogihara et al.  [65] showed

    that the equilibrium conversion of methane is a function of 

    the reaction temperature for methane decomposition. The

    equilibrium conversion of methane was estimated, assuming 

    that the carbon was formed as graphite (CH4   4   2H2   þ   C

    (graphite)). Theyreported that methane conversion was not as

    high when methane decomposition was performed at

    temperatures below 600   C. Using a commercial Ni-based

    catalyst, Suelves et al. [66] reported that, at a temperature of 

    700  C, the concentration of hydrogen was around 80%, which

    corresponds to a methane conversion close to the theoreticalequilibrium value.

    The TCD of methane on various Ni-containing catalysts

    (Mn, Fe, Co or Cu) was investigated by Zabidi et al.   [30],

    showing that Ni/Mn based catalysts gave better results for

    hydrogen production with carbon nanotubes (CNTs) as the by-

    product. Choudhary et al.   [67]   studied the effect of using 

    different Ni-containing metal oxide (ZrO2, MgO, ThO2, CeO2,

    UO3, B2O3   or MoO3) and zeolite (HZSM-5, Hb, HM, NaY,

    Ce(72)NaY or Si-MCM-41) catalysts and concluded that

    Ni/ZrO2 and Ni/Ce(72)NaY showed the most promising results

    for a cyclic process. The effects on the catalytic activity of Ni

    using different kinds of zeolites as supports were investigated

    by Inaba et al.   [68], who found that Ni-supported on USY

    Fig. 3 – Schematic diagram of several types of reactor used for catalytic and non-catalytic thermal decomposition of 

    methane. a) Vortex-flow reactor confined to a cavity receiver  [33]. b) Solar-thermal fluid-wall reactor [34]. c) Multilayer

    reactor [35]. d) FBR  [36].

    i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 1 1 6 0 – 1 1 9 01166

  • 8/16/2019 Hydrogen Production by Methane Decomposition - A Review 2010

    8/31

    zeolite (Si/Al2¼ 14,360) showed a longer catalytic lifetime, and

    considered it to be the best catalyst for hydrogen production

    by methane TCD. In another study, methane TCD at 550   C

    over Ni-supported on HY, USY, SiO2  and SBA-15 was con-

    ducted and among all the catalysts tested the Ni/HY catalyst

    was found to have the highest activity, as shown in  Fig. 5.

    Additionally, the total run time increased with the various

    catalyst supports in the order: HY   w SiO2  >  USY  >  SBA-15.

    However, the overall accumulation of carbon was in the order:HY  >  USY  >  SiO2 >  SBA-15 [69].

    Catalysts containing Ni and Fe have been widely tested in

    the past. It was reported that the Ni-based catalysts have

    a maximum operating temperature of 600   C; thus, as

    methane conversion is thermodynamically limited at this

    temperature, concentrated hydrogen streams (H2>60%)

    cannot be obtained using a nickel-based catalyst [70]. On the

    contrary, Fe-based catalysts are more stable at higher

    temperatures (700–1000   C), but deactivation occurs during 

    repeated cycles, resulting in a shorterlifetime.Chesnokov and

    Chichkan   [52]   reported that the modification of a 75%Ni–

    12%Cu/Al2O3  catalyst with Fe made it possible to increase

    optimal operating temperatures to 700–750   C while

    maintaining excellent catalyst stability, and the hydrogen

    concentration at the reactor outlet exceeded 70 mol%. The

    structure of the filamentous carbon formed upon catalytic

    decomposition of hydrocarbons on iron-subgroup metals and

    their alloys, was reported by Chesnokov and Buyanov, along 

    with generalizations on the regularities of carbon deposition

    on these metals and a discussion of the growth models of 

    some morphological modifications of filamentous carbon [71].

    Fe catalysts can decompose methane at a temperature rangeof 700–1000  C, but Fe catalysts have a very short lifetime [72].

    For example, Ermakova et al.   [73] reported that Fe catalysts

    were effective for methane decomposition in the temperature

    range of 650–800 C; however, the hydrogen yield (H2 / M:molof 

    H2 formedper mol ofmetalscontainedin the catalyst)for Fe at

    800   C catalysts (H2 /Fe  ¼   418) was more than an order of 

    magnitude lower than that for Ni catalysts at 500   C (H2 /

    Ni ¼ 4802) [62].

    Nuernberg et al. [74] investigated the influence of operating 

    conditions on the catalytic performance of Co supported on

    Al2O3. They studied the effect of different Co contents at

    a temperature range of 600–800   C and different methane

    partial pressures. It was found that the best conditions for

    Fig. 4 – Simplified diagram of types of heating source, reactor, catalyst and analysis apparatus used in methane TCD studies.

    i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 1 1 6 0 – 1 1 9 0   1167

  • 8/16/2019 Hydrogen Production by Methane Decomposition - A Review 2010

    9/31

    hydrogen production by methane TCD were at 800   C with

    a nitrogen-to-methane molar ratio of 6:1 and a Co loading of 

    20 wt% (Fig. 6). Muradov and Veziroglu [75] summarized the

    bulk of the available literature data on a number of these

    catalysts; their preferred temperature ranges and carbon

    products in methane TCD are presented in Fig. 7.

    Ni–Cu/Al2O3 catalysts have been shown to have a number

    of advantages over Ni/Al2O3   catalysts   [76,77]. Ni–Cu/Al2O3catalysts with high metal loading had optimum operating 

    temperatures in the range of 600–675  C, making it possible to

    achieve higher methane conversion. Similar results were alsoreported for methane TCD over a Ni–Cu–MgO catalyst to

    produce hydrogen and CNF; Wang and Baker [78] found that

    a Ni–Cu–MgO catalyst maintained its activity at high levels for

    substantially long periods of time at 665–725   C, and the

    catalyst was capable of generating large amounts of CO-free

    H2 and solid carbon. In another report, 75% Ni-15% Cu–Al2O3catalysts with a nickel-particle size of 27 nm gave a carbon

    yield of 700 g C (gNi)1 at 625  C [79].

    To produce highly concentrated hydrogen and CNF bymethane TCD in a temperature range of 700–850  C, Ogihara

    et al.   [65]   investigated M/Al2O3   (M]Fe, Co, Ni and Pd) and

    Pd-based alloys containing Ni, Co, Rh or Fe. They found that

    the conversion of methane at 700  C for Fe, Co and Ni/Al2O3catalysts was very low (

  • 8/16/2019 Hydrogen Production by Methane Decomposition - A Review 2010

    10/31

  • 8/16/2019 Hydrogen Production by Methane Decomposition - A Review 2010

    11/31

    Table 3 – Recent studies on metal catalysts for methane TCD, preparation methods and major findings.

    Researcher Catalyst composition Catalyst preparationtechnique

    Operating conditions T/FlowPhysical properties W/SA/MPS

    Aiello et al. [29]   15wt% Ni/SiO2. Wet Impregnation. 650/30,000 h1 0.2/—/20–35 mesh

    * 3 h on stream.

    * Methane initial conversion (MIC) was about 24–35%, depending on the number of cycles.

    * Deactivated catalyst was regenerated with steam.* Ten cycles without significant loss of catalytic activity.

    * A small amount of carbon, concentrated in small pockets, resisted gasification.

    Suelves et al. [66]   65 Wt%Ni/Silica and alumina. – 550–700/20, 50 and 100 mL/min

    2 and 0.3/190/100and 2000

    * At 700  C, hydrogen concentration was 80%, catalyst activity did not decay after 8 h on stream at a space-time of 1 s.

    * Range of sustainability factor was 0.14–1.

    Venugopal et al. [82]   (5–90)Ni/SiO2. Wet Impregnation. 600/24,360 h1

    0.15/———/———

    * Fixed-bed vertical quartz reactor was used (diameter 0.8 cm and length 46 cm).

    * The 30 wt% Ni/SiO2 demonstrated superior activity and longevity and produced CNFs at 303 mol/molNi.

    * XRD result showed that only NiO phase was present in the fresh catalysts while the deactivated catalysts displayed both metallic nickel and

    graphitic carbon phases.

    Ermakova et al. [53]   90Ni–10Al2O3,

    90Ni–5ZrO2–5SiO2, and other catalyst

    structures.

    Impregnation, fusing followed by

    decomposition.

    550/120,000 mL/(g $h)

    0.01/–/–

    * Depending on calcination temperatures (400–800  C), the range of SA was 5–145 m2 /g.

    * Depending on textural parameters of the catalytic systems, the ranges of carbon yield and methane conversion were 287–384 g/g Ni and 12.9–

    15.3%, respectively.

    Monnerat et al. [83]   Ni-gauze – 410–550/75 mL/min

    0.207/26.7/—

    * Catalyst deactivated due to intensive coke deposition in the form of CF.

    * Maximum cycle time was 4 min for better hydrogen performance.

    * 10% of catalyst mass lost after 70 h on stream.

    Zabidi et al. [30]. Ni/M-based (M]Mn, Fe, Co or Cu) – 550–900/3200, 9600,

    16,000 h1

    –/8.79 to 4.79/—

    * Ni/Mn maintained more activity (59% methane conversion) for 120 min than the other catalysts.

    * For Ni/Mn at 900  C, an increase of space velocity from 3200 to 16,000 h1 resulted in a decrease in hydrogen yield from 96% to 54%.

    * Reactor blocking was reported for Ni/Co after 120 min on stream.

    Choudhary et al. [42]   Ni-containing different metal

    oxides and zeolites.

    Co-precipitating 500/7000 mL/(g  $h).

    0.4/266–786/30–52 mesh size

    * Two parallel reactors operated in cyclic manner.

    * Deactivated catalyst was regenerated by steam.

    * Ni/ZrO2 and Ni/Ce(72)NaY showed promising results for the cyclic process.

    * Methane conversion (w28%), hydrogen productivity (w31%) and CO2 produced passed through a maximum at a switchover time of 10 min.

    Rahman et al. [38]   5 wt%Ni/g- Al2O3   Wet impregnation 500–650/——

    0.1–1/——/—

    * The study was conducted using a thermobalance.* Partial catalyst regeneration with oxygen was recommended because full regeneration destroyed its activity.

    * Activation energy (Ea ) was  46 kJ/mol.

    Zhang and Amiridis

    [84]

    16.4 wt%Ni/SiO2   Wet impregnation 550/15,000–37,500 h1

    0.2/250/25–35 mesh

    * When silica was used, no measurable methane conversion was observed at temperatures below 800  C even at a low space velocity (6000 h1).

    * Methane initial conversion (MIC ) was 35% at 550  C and 30,000 h1.

    * Nickel supported on silica was found to be active, producing stoichiometric amounts of hydrogen and carbon.

    * The deactivated catalyst could be fully regenerated by either oxidation in air or steam gasification of the deposited carbon.

    Bai et al. [85]   6.7 wt%Ni/AC from coal. Impregnation 550–850/50 mL/min

    0.2/——/——

    i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 1 1 6 0 – 1 1 9 01170

  • 8/16/2019 Hydrogen Production by Methane Decomposition - A Review 2010

    12/31

    Table 3 ( continued ).

    Researcher Catalyst composition Catalyst preparationtechnique

    Operating conditions T/FlowPhysical properties W/SA/MPS

    * Very small amounts of hydrocarbons other than methane (40%) was reached at w800  C.

    * At temperature over 900 

    C, the homogeneous decomposition of methane was significant and the conversion of methane rapidly increasedafter complete deactivation of the catalyst.

    * The attempted catalyst regeneration by the removal of the carbon deposited on the external surface of catalytic particles by attrition failed.

     Jang and Cha [88]   Fe (iron powder) and (5–20) wt%

    Fe/Al2O3 catalyst.

    Fe/Al2O3 catalyst prepared by

    impregnation.

    600–1000/up to 50,000 h1

    —/———/359

    * MIC was w100% at 1000  C and 5000 h1.

    * PBR and FBR reactors were used.

    * The methane conversion rate was maintained by attrition of the produced carbon on Fe catalyst surface in an FBR.

    Qian et al. [80]   Reduced and unreduced

    Co/Mo/Al2O3 Ni/Cu/Al2O3

    Co-precipitation 550–850/50 h1

    —/——/———

    * For the reduced Co/Mo/Al2O3, the methane conversion was 

  • 8/16/2019 Hydrogen Production by Methane Decomposition - A Review 2010

    13/31

    temperature was increased to 740   C with a ramp of  

    10  C min1 (a pre-induction reaction; PIR). The morphologies

    and the sizes of the CNFs formed are similar to those formed

    during CTR at 543   C. However, the shapes of the metal

    particles are different. Some of them have the features of theparticles in the CTR at 543   C, but with a smoother surface.

    The others are more pear-shaped, but their tails are not as

    sharp as those in Fig. 8 D–F. They also showed that at constant

    temperatures the catalyst showed quasi–stable activity below

    640   C for a notable time period, while its performance

    decayed rapidly above this temperature, and a quick deacti-

    vation was observed at 740   C. Pinilla et al.   [97]  studied the

    activity of a Ni–Cu–Al catalyst using a thermobalance and

    reported that the methane and hydrogen reduced the effect of 

    thermal sintering on Ni and lowered the surface energy by

    chemisorption. In addition, they showed that the carbon

    deposited on the Ni leading face probably diffused to the Ni

    trailing face superficially, and not through the bulk Ni.Nuernberg et al.   [74]  attributed the deactivation of a Co-

    alumina catalyst to the methane molecules which were

    initially adsorbed (by dissociative adsorption) and decom-

    posed on the metal surface, resulting in the formation of 

    chemi-sorbed carbon species; the carbon species then dis-

    solved and diffused through the bulk of the metal particle.

    They concluded that catalyst deactivation occurred when the

    rate of carbon diffusion through the metal catalyst particle

    was slower than that of the formation of carbon at the surface

    of the CoO sites. Under these circumstances, carbon builds up

    at the catalyst surface and eventually encapsulates the metal

    particle causing a loss of activity. In an evaluation of co-

    precipitated Ni-alumina and Ni–Cu–alumina catalysts,

    Avdeeva et al. [98] revealed that the nickel in deactivated Ni

    catalysts was found in a metallic state, while in the Ni–Cu

    samples about half of the nickel was atomically dispersed in

    carbon. The catalyst deactivation mechanisms suggested

    included the fragmentation of the metal particles as well asatomic erosion in the Ni–Cu samples.

    With regard to the deactivation of metal catalysts, Aiello

    et al. [29] reported that decomposition of the hydrocarbon at

    the gas–metal interface was followed by dissolution of carbon

    into the metal and diffusion through the particle. The carbon

    then precipitated at the metal–support interface, detaching 

    the metal particle from the support and forming a filament

    with an exposed metal particle at its tip. The rate-determining 

    step of this process is believed to be the diffusion of carbon

    through the metal particle. This mode of carbon accumulation

    allows the catalyst to maintain its activity for an extended

    period of time without deactivation. They also showed that

    the cracking of methane over Ni/SiO2   has indicated thatthousands of carbon atoms can be deposited via this process

    on the catalyst per surface nickel atom. Eventuallythough, the

    catalyst is deactivated due to the limitations imposed by the

    available free space in the reactor.

    5. Carbonaceous catalysts forthermocatalytic decomposition of methane

    Many industrial catalysts consist of metals or metal

    compounds supported on an appropriate support. The main

    reason is to maintain the catalytically active phase in a highly

    dispersed state. Alumina, silica and carbon (mainly AC) are

    Table 3 ( continued ).

    Researcher Catalyst composition Catalyst preparationtechnique

    Operating conditions T/FlowPhysical properties W/SA/MPS

    * Temperatures used for calcination were 450, 600, 800 and 1000  C.

    * Ni–Cu–Mg catalysts showed a high and almost constant hydrogen production yield.

    * The structural properties of the obtained CNF were highly dependent on the presence of Cu and on the calcination temperature.

    Suelves et al. [91]   Ni–Al and Ni–Cu–Al Co-precipitation, impregnation and

    fusion.

    700/20 mL/min

    0.3/——/——

    * Hydrogen production was not highly dependent on the preparation method.

    * The presence of Cu as a dopant in Ni–Cu–Al catalysts enhanced the catalytic activity substantially.

    * Ni–Cu–Al catalysts enhanced the formation of a well-ordered graphitic carbon, while Ni–Al catalysts promoted the deposition of a turbostratic

    carbon.

    Lázaro et al. [92]   Ni–TiO2 and Ni–Cu–TiO2   Impregnation and fusion 700/20 mL/min

    0.3/——/——

    * Those catalysts prepared by fusion and those including copper in their composition both showed enhanced catalytic activity.

    * The calcination temperature only slightly affected the hydrogen concentration of the outlet gas from the reactor.

    * The deposited carbon appeared as long nanofilaments or uniform coatings on the catalyst particles, depending on the nickel-particle size.

    Echegoyen et al. [93]   Ni–Mg and Ni–Cu–Mg Co-precipitation, impregnation and

    fusion.

    700/20 mL/min

    0.3/——/——* The presence of copper enhanced hydrogen production and the best results were obtained for Ni–Cu–Mg catalysts prepared by the fusion

    method.

    * For the Ni–Mg catalysts, the nickel-crystal size influenced catalysts performance in that the highest crystallite size gave the lowest the

    hydrogen yields.

    * In the Ni–Cu–Mg catalyst the Ni particle size was not a significant influence on hydrogen yields.

    W, weight of catalyst (g); SA, surface area (m 2 /g); MPS, mean particle size (mm, unless other units are stated); T, temperature (C); F, either flow

    rate (mL/min) or space velocity (h1) unless otherwise stated; —, not mentioned in the original paper.

    i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 1 1 6 0 – 1 1 9 01172

  • 8/16/2019 Hydrogen Production by Methane Decomposition - A Review 2010

    14/31

  • 8/16/2019 Hydrogen Production by Methane Decomposition - A Review 2010

    15/31

    and no sulfur poisoning; and (3) the carbon formed can be

    used as a catalyst precursor [103–105]. In the case of methane

    TCD, furtheradvantages of using carbonaceous catalysts cited

    include: (1) it can be catalyzed by carbon produced in the

    process, so an external catalyst would not be required (except

    for the start-up operation); and (2) the separation of carbon

    product from the carbon catalyst is not necessary  [75].

    Various carbon materials have been used for methane TCD

    such as AC, CB, glassy carbon, acetylene black, graphite, dia-mond powder, CNT and fullerenes [47]. Among these carbon

    materials, most of the studies have focused on AC (manu-

    factured from different carbon-based sources) and CB due to

    their activities and good stabilities. The most important

    characteristics of an industrial catalyst are the chemical

    composition, surface area, porosity and pore-size distribution

    (if the catalyst is porous), stability and mechanical properties.

    5.1. Activated carbon

    A number of different lignocellulosic precursors, e.g., coconut,

    almond, peach, plum, olive, palm and cherry, are used as rawmaterials for manufacturing AC. It is widely assumed that the

    porosity of AC depends not only on the raw material used as

    the precursor but also on the manufacturing process (which

    involves carbonization and activation) [106]. According to the

    IUPAC classification, pores can be classified into three cate-

    gories, namely, micropores (50 nm).

    Daud and Ali   [107] studied the effects of burn-off on the

    micropore, mesopore and macropore volumes with palm and

    coconut shells as raw materials. The carbonization and acti-

    vation processes were carried out at 850  C using an FBR. They

    concluded that within the burn-off range studied (up to 70%),

    the micropore and mesopore volume created in palm-shell-

    based AC were always higher than those of coconut-shell-

    based AC. Achaw and Afrane [108] studied the development of 

    pore structure in AC produced from coconut shells and found

    that the modification of the pore structure was observed to

    begin at the drying stage and continued into the activation

    stage. The effects of using different activating agents (CO2,

    H2O, KOH, H3PO4 and ZnCl2) on the physical properties (bulk

    density, porosity, pore-size distribution and surface area) of the AC produced have been investigated   [109–112]. The

    difference between using H2O and CO2 as an activating agent

    is that CO2  produces an opening of narrow micropores fol-

    lowed by their widening, whereas water vapor widens the

    microporosity from the early stages of the process, leading to

    the production of a lower micropore volume in the AC product

    [113]. The different methods used for physical and chemical

    modifications of AC surfaces have been reviewed by Yin et al.

    [114].

    When using AC as a catalyst for methane TCD, it has been

    reported that the actual value of the highest ultimate mass of 

    carbon that AC can accumulate is significantly smaller than

    the theoretical value (ifone assumesthe entire pore volume inACPS were completely filled with deposited carbon), which

    indicatesthat only a fraction of the pore volume is filled by the

    deposited carbon and the lower capacity could be attributedto

    the blockage of the pore-mouth, which would hinder the

    internal diffusion of methane molecules  [39,40]. It has been

    reported that the capacity of AC to accumulate deposited

    carbon before deactivation is related to the surface area and

    pore-size distribution. Additionally, microporous carbons

    with high contents of oxygenated surface groups exhibited

    high initial rates of methane decomposition (rO) but rapidly

    became deactivated, while the use of mesoporous carbons

    with high surface area resulted in a more stable and sustain-

    able process [103]. Suelves et al. [39] reported that the amountof carbon deposited at deactivation showed a linear relation-

    ship with the total pore volume of the fresh catalysts.

    Most ACs posses carbon-oxygen functional surface groups

    such as R–COOH, R–OCO, R–OH and R]O. Studies have been

    carried out to investigate the effects of these surface func-

    tional groups on methane TCD. Muradov et al.  [47] compared

    the catalytic activity of virgin AC (lignite) samples with that of 

    AC (lignite) pretreated with pure hydrogen at 850 C to remove

    carbon-oxygen groups from the carbon surface. They found

    that although carbon-oxygen surface groups may play a role

    in methane dissociation, particularly at the initial stage of the

    process, the catalytic activity of carbons, in general, could not

    be solely attributed to the presence of such surface functionalgroups. Using ACs with different textural properties and

    surface chemistries, Moliner et al.   [103]   showed that no

    correlation could be found between the CO and CO2 evolved

    and the long-term behavior andthat rO seemedto be relatedto

    the content of oxygenated groups (Fig. 9). Suelves et al.   [39]

    found a good correlation between  rO and the concentration of 

    oxygenated groups desorbed as CO in a temperature-pro-

    grammed-desorption experiment.

    ACs usually contains 1–12% ash (inorganic constituents).

    Kim et al. [115] studied the effect of ash content on the activity

    of AC in methane decomposition. They found that the ash

    content in the ACs from coconut shell was about 1.8–4.4 wt%

    while that in the ACs from coal ranged from 6.3 to 10.2 wt%,

    Fig. 9 – Initial activity for methane conversion and carbonaccumulated as a function of the concentration of 

    oxygenated groups in the fresh catalyst;  T [ 850   8C [103].

    i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 1 1 6 0 – 1 1 9 01174

  • 8/16/2019 Hydrogen Production by Methane Decomposition - A Review 2010

    16/31

    and the major elements in the former were K, Na, Si and Mg,

    while in the latter they were Al and Si. These elements are

    known to be almost completely inactive for the decomposi-

    tion of methane (although Fe, found in trace amounts, is

    known to be active). They also reported that the surface area

    of ACs were reduced and the initial activities decreased byabout 20% when the ash was removed from AC. Muradov et al.

    [47] conducted controlled experiments with the inorganic ash

    produced from the lignite-derived AC and showed that these

    metal impurities had no particular catalytic activity in carbon-

    catalyzed methane decomposition.

    5.2. Carbon black

    Commercial CBs such as black pearls (BP) 110, 120, 1300, 2000,

    Regal 330, Vulcan XC72, Fluka 03866 and Fluka 05120 have

    been used as carbonaceous catalysts for methane TCD

    [39,47,116,117]. In contrast to the ACs, which showed accept-

    able   rO  but were rapidly deactivated due to the blocking of 

    pores by growing crystallites, which hinder the internal

    diffusion of methane molecules, the majority of CB surfaces

    are relatively easily accessible to methane molecules during 

    methane decomposition. CBs differ in particle size, average

    aggregate mass, morphology, etc. (for example, the oil-

    furnace process produces CBs with particle diameters in therange 10–250 nm, and surface areas of 25–1500 m2 /g) [118]. It

    has been reported that the catalytic activity of different CB

    samples remains almost unchanged despite the significant

    reduction in surface area during the reaction  [119]. Using CB,

    Lázaro et al.   [104]   showed that the amount of surface

    complexes desorbed as CO, as well as the surface area ( Fig. 10

    A), decreased gradually during the stable period of catalyst

    activity. In addition, it was found that the total pore volume

    decreased (Fig. 10B) as the carbon was deposited and the

    degree of graphitization increased as the reaction progressed.

    In a kinetic and deactivation study of carbon catalysts,

    Serrano et al.   [120]  showed that the most active catalyst at

    short reaction times was AC, but it underwent a fast

    Fig. 10 – Change of the BET area (a) and total pore volume (b) versus reaction time at   T [ 950   8C and GHSV [ 360 hL1 [104].

    i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 1 1 6 0 – 1 1 9 0   1175

  • 8/16/2019 Hydrogen Production by Methane Decomposition - A Review 2010

    17/31

    deactivation due to the deposition of carbon in methane TCD;

    contrastingly, CBs presented high reaction rates at both short

    and long reaction times. The authors showed that methane

    TCD using CB could go on for several hours until most of the

    surface was covered by carbon crystallites. A quasi-steady

    state quickly reached with CB and could remain constant for

    more than 8 h regardless of the reaction temperature due to

    the continuous changes in carbon morphology [117,121].

    5.3. Operating parameters and kinetics

    As mentioned previously, among carbon materials, the focus

    has been on using AC and CB as carbonaceous catalysts for

    methane decomposition. The operating parameters that have

    been studied for the activity and long-term stability of these

    catalysts are temperature range (700–1000   C), pressure

    (usually atmospheric) and methane flow rate. These studies

    also included physical properties (surface area, mean particle

    size, composition, porosity and pore-size distribution) of the

    virgin and deactivated catalysts. The weight of catalysts used

    ranged from few milligrams to 25 g, and the most employedreactor types were either PBR or FBR, generally constructed

    from quartz. Kinetic studies were also conducted to determine

    the reaction order and activation energy (Ea) of methane

    decomposition and the sustainability factor of the catalyst,

    which is defined as reaction rate after1 h on stream divided by

    ro. The effect of operating conditions on carbonaceous cata-

    lysts and their changing physical properties will be discussed

    in the next subsection.   Table 4   lists recent methane TCD

    studies carried out with carbonaceous catalysts, including the

    catalyst types, operating conditions and the major findings.

    5.4. Deactivation of carbonaceous catalysts

    The reason for the gradual deactivation of catalysts is due to

    the deposition of carbon (produced from the decomposition of 

    methane) on the catalyst surface, which results in blocking of 

    active sites and a reduction in catalyst surface area. The

    deposit has lower surface area and activity compared to the

    original carbon catalyst. Additionally, catalyst activity is

    influence by its structure. The carbon produced by decompo-

    sition of methane has a more ordered structure than amor-

    phous carbons, but they are less structurally ordered than

    graphite. The catalytic activity of carbons for methane

    decomposition varies according to their structure in the

    following order: amorphous  >   turbostratic  >   graphite. It is

    assumed that the process of new carbon build-up can besimply divided into two steps: formation of carbon nuclei

    (characterized by high catalytic activity) and carbon-crystal-

    lite growth (these have a structure close to that ofgraphite and

    thus the lowest catalytic activity). The formation of carbon

    nuclei has an Ea of 316.8kJ/mol, which is higher than the Ea for

    carbon-crystallite growth, at 227.1 kJ/mol. The total deacti-

    vation rate is the sum of both steps, and, in general, the rate of 

    carbon-crystallite growth tends to be higher than the rate of 

    nucleus formation [5,47].

    In a deactivation study of AC, Kim et al.  [124] showed that

    carbon-nucleus formation appeared to occur initially but was

    quickly terminated and then the carbon-crystallite growth

    become dominant. They explained the possibility of a linear

    relationshipbetweencatalystactivityandtheamountofcarbon

    deposited by reasoning that a pore would not have a uniform

    diameter but instead havea complicated structure withnarrow

    passages, wide passages, large cavities and interconnections.

    During nucleation, a uniform deposition may occur since the

    nucleiare notyet largeenough to block thepore,but nucleation

    is terminated when crystallite growth becomes dominant. As

    the crystallites grow, narrow passages are blocked sooner andthe inner active surface may not be accessible by the reactant.

    Using a thermobalance to study thedeactivation kinetics of AC,

    AbbasandDaud [40] developed a model to describethe decrease

    of catalytic activity with time and reported a deactivation-

    reaction order of 0.5 and an  Ea of 194 kJ mol1; in addition, the

    weight of deposited carbon was fitted with time using the

    Voorhies equation, which is commonly used for reactions

    involved in hydrocarbon cracking.

    There have been studies carried out to investigate the

    effect of carbon deposition on catalyst surface area. It was

    believed that the deposition results in the blockage of the pore

    mouths by growing carbon crystallites which leads to rapid

    deactivation and a sharp decrease in the overall surface area[102,117,121]. Moliner et al. [103] showed that the spent cata-

    lysts underwent a substantial reduction in specific surface

    area compared to the fresh catalysts and found that the

    mechanism for catalyst deactivation could be based on

    progressive pore blocking by carbon deposition; carbon

    deposition in a quantity of less than 25% of the catalyst mass

    reduced the surface area from 1300 to 39 m2 /g. Lee et al. [31]

    showed that after4 h of deposition most AC had a surface area

    of 20–60 m2 /g compared to 860–978 m2 /g for fresh AC catalyst.

    Ashok et al.   [122]   studied methane TCD using a variety of 

    commercialcoal-derivedcarbons and showed that the surface

    areas of the catalysts decreased from 117 to 1478 m2 /g in the

    fresh catalysts to 8–27 m2 /g. For CB, it was reported that afterreaction at 900 Cfor1handat1000 C for 2 h,the surface area

    of the CB (BP 2000) decreased from 1500 m2 /g to 310 and

    100 m2 /g, respectively [119]. Moradov et al.   [47] showed that

    the surface area vs. time curve closely followed the methane

    conversion curve, with an initial drop followed by a shallow

    decline. They obtained empirical decay-law equations for the

    surface area of AC and CB with time, and the two equations

    showed that the surface area of AC diminishes much faster

    than that of CB. Krzy _zyński and Koz1owski [125] attributed the

    controversial nature of the problem of the relationship

    between the surface area of the catalyst and its activity to the

    use of different carbon materials obtained from different

    precursors and applying different preparation methods, asthen the activity of the samples could depend not only on the

    surface area but also on their structure. They used ACs all

    obtained from the same precursor (brown coal) and activated

    with KOH, with different weight ratios of KOH to precursor(4:1

    to 1:2)and at different temperatures (550, 700 and850 C); their

    results on the changes in catalyst activity with time are shown

    in Fig. 11. Again, no discernible correlation wasfound between

    the surface area of the catalysts and their  rO, however, they

    found that with increasing surface area, and particularly with

    increasing pore volume, the resistance of the catalysts to

    deactivation increased.

    The pore-size distribution is a key factor in defining the

    long-term behavior of the catalyst. It has been reported that

    i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 1 1 6 0 – 1 1 9 01176

  • 8/16/2019 Hydrogen Production by Methane Decomposition - A Review 2010

    18/31

    Table 4 – Recent studies using carbonaceous catalysts and major findings.

    Researcher Catalysts Operating conditionsT/Flow

    Catalyst propertiesW/SA/Porosity/MPS

    Muradov et al. [47] ACs (different sources), CBs (different sources), Glassy carbon,

    Natural graphite, Polycrystalline graphite, Diamond powder, CNT,

    Fullerene soot and fullerenes C60/70

    850/0.1 s (residence time)

    0.03–0.1/4–2570/—/—

    * AC had the highest rO among all the carbon materials tested.

    * Structurally ordered carbon (graphite and diamond) with very low surface area

    showed negligible catalytic activity.

    * Nanostructured carbon (CNT and fullerenes) showed relatively low catalytic activity.

    * Fullerene soot was especially active.

    * Metal impurities played a negligible role in methane TCD.

    * Reaction order for methane was determined to be 0.6 0.1 for AC (lignite)

    and 0.5 0.1 for CB (BP2000) catalysts.

    * Ranges of  Ea for AC and CB were 160–201 and 205–236 kJ/mol, respectively.

    Muradov [118] ACs from different origins activated by steam or KOH. 850/1 and 10 s (residence time)

    0.03/650–3370/—/—

    * High rO was followed by a rapid drop in catalytic activity, and finally, a quasi-steady reaction rate was

    reached after 1–1.5 h on stream.

    * Range of  rO was 1.63–2.04 mmol/(g $min.).

    * The origin and method of AC activation had no significant effect on AC catalytic activity.

    * Hydrogen initial concentration (HIC) in the effluent gas was 40–46 vol%.

    * HIC reached up to 90 vol% when residence time was 10 s (comparable to Fe- and Ni-based catalysts at the identical conditions).

    Pinilla et al. [105]   CB (BP2000) and AC (CG Norit). 800–950/3000 mL/min

    0.03/1300–1337/1.11–3.06/—

    * Reaction order for AC and CB was 0.48 and 0.6, respectively.

    * Ea for AC and CB were 141 and 238 kJ/mol, respectively.

    * The higher the temperature used, the faster the catalyst reached the maximum amount of carbon that could be accumulated.

    Suelves et al. [117]   Two graphitized CBs (Carbopack B and C), Three CBs (Fluka 03866,

    Fluka 05120 and Black Pearls, 2000), CB (HS-50) and AC (CG NORIT).

    850/20–100 mL/min.

    ——/10–1337/——/——

    * The samples released tar and water during the first stage of the runs.

    * HIC was 10% for Carbopack B and up to 70% for Fluka 03866 and AC (CG).

    * CB B and C showed the highest sustainability factor (1), while Fluka 03866 showed the lowest (0.23).

    * AC showed the highest rO (0.69) while CB C showed the lowest (0.01 mmol/(g $min)).

    Moliner et al. [103]   Three different commercial ACs (char pyrolysis), Two

    activated chars (steam activation at 600  C and 750  C)

    850–950/20 mL/min

    2/52–1300/——/

  • 8/16/2019 Hydrogen Production by Methane Decomposition - A Review 2010

    19/31

    Table 4 ( continued ).

    Researcher Catalysts Operating conditionsT/Flow

    Catalyst propertiesW/SA/Porosity/MPS

    * All the carbons deactivated quite rapidly and the sustainability factors were 0.1–0.25.

    * The range of  rO was 0.7–1.5 mmol/(g $min.) at space-time of 0.6 s.* For certain ACs, the reaction order was 0.51 and 0.49 and  Eas were 194, 198 and 186 kJ/mol.

    Bai et al. [101]   One AC from hardwood, two ACs from coal

    (steam activated) and an active alumina.

    750–900/3000–18,000 mL/(g. h)

    1/152–783/0.36–0.73/246–833

    * Very small amounts of hydrocarbons other than methane (

  • 8/16/2019 Hydrogen Production by Methane Decomposition - A Review 2010

    20/31

    the amount of carbon deposited shows a linear relationship

    with the total pore volume of the fresh catalysts [39]. Dunker

    et al.   [116]  showed that carbon deposition eliminated all or93% of the total micropore area during certain runs and

    concluded that the reduction in catalyst activity may be

    associated with the reduction in microporosity and probably

    not associated with the change in mesopore area, which was

    actually increased in some runs. Ashok et al.   [122] reported

    a decrease in pore volume from 0.68 cm3 /g for fresh coal-

    derived catalyst to 0.042 cm3 /g, and the micropore volume

    decreased from 0.56 cm3 /g to 0.0038 cm3 /g. Bai et al. [101] used

    three types of AC for methane TCD and showed that the pore

    volume and micropore volume decreased greatly, and the

    decrease of micropore volume was especially great (for

    example, from 0.26 to 0.000112 cm3 /g). These changes were

    attributed to the carbon deposition, especially in the micro-pores of ACs; thus, they concluded that the adsorption and

    decomposition of methane occur mainly in the micropores.

    A survey of the effect of catalyst particle size indicated that

    thesmallerparticlesshowedthehigherrO.However,afteralong 

    deposition duration, the effect of particle size on methane

    conversion was insignificant. Among a range of particle sizes

    studied (108–300mm) used for methane TCD in an FBR, Lee et al.

    [31] reported that the 108-mm particle size resulted in higher

    methane conversion due to its higher surface area and smooth

    fluidization. Kim et al. [115] showed that while the decomposi-

    tion rate was higher over the smaller particles, the rate of 

    deactivation was lower, and this indicates that carbon deposi-

    tion occurs preferentially at the outer shell of the carbon parti-cles, resulting in pore-mouth blocking while the inner core was

    left intact, especially for larger particles. They also showed that

    pore-mouth blocking was alleviated as the particle size

    decreased and the inner surface could be utilized to a much

    greater extent. Among the range of particles sizes studied (137–

    1140mm),the137-mm particlesize gavethe highest conversion of 

    methane.In another study, Abbas andDaud [37] showed thatas

    theparticle size was decreased theultimate mass that AC could

    accumulate before complete deactivation increased (Fig. 12).

    On the effect of temperature, Muradov[13] showed thatwith

    a rise in temperature, the mean size of carbon crystallites ten-

    ded to decrease, resulting in an increased methane decompo-

    sition rate, and this explains the experimental fact that at

    higher temperatures (e.g.,  >850   C) carbon catalysts tend to

    deactivate at a slower rate compared to lower temperatures.

    With regard to the effect of temperature on catalyst deactiva-tion, Moliner et al. [103] showedthat two different effects could

    be identified: (1) a molecular-sieve effect, which would be

    associated with pore-mouth blocking. As carbon deposition

    progresses, the pore-mouth area decreases and the inner pore

    surface becomes unavailable for methane adsorption; and (2)

    an activated-diffusion effect, which is related to the rate of 

    diffusion of the methane molecules inside the smallest pores.

    As temperatureis increased, the rate of diffusion increases and

    deposition inside the pores is enhanced. In this way, at high

    temperatures the decompositioncould takeplace mainly inside

    the pores, where the majority of the AC surface is located.

    6. Catalyst regeneration

    As mentioned previously, the catalyst used for methane TCD

    is rapidly deactivated due to intensive carbon deposition

    Table 4 ( continued ).

    Researcher Catalysts Operating conditionsT/Flow

    Catalyst propertiesW/SA/Porosity/MPS

    Abbas and Daud [37]   AC from palm-shell and commercial AC 800–950/200 mL/min

    0.01/1027–1100/0.4501/117

    * Study conducted in a thermobalance.

    * Reaction order and Ea were 0.5 and 210 kJ/mol, respectively.

    * Ultimate mass gain increased with decreased mean particle size.

    * Carbon deposition occurred uniformly at the early stage of the process, while the diffusion effect was significant at the end of the process.

    Umf  , minimum fluidization velocity; T, temperature (C); F, either flow rate (mL/min) or space velocity (h1) unless otherwise stated; W, weight

    of catalyst (g); SA, surface area (m2 /g); Porosity, (cm3 /g); MPS, mean particle size (mm unless otherwise stated); —, not mentioned in the original

    paper.

    Fig. 11 – Catalytic activity of carbon samples in methane

    decomposition reactions at 750   8C [125].

    i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 1 1 6 0 – 1 1 9 0   1179

  • 8/16/2019 Hydrogen Production by Methane Decomposition - A Review 2010

    21/31

    because carbon is one of the two products. Different activating 

    agents such as CO2, H2O and O2   can be used to regenerate

    a deactivated carbonaceous catalyst; CO2   and O2  have also

    been used in studying the regeneration of deactivated metal

    catalysts used for methane TCD. Using different oxidizing 

    agents, Muradov [13]  reported that the treatment of deacti-

    vated AC samples with steam and steam-CO2 mixtures (1:1 by

    volume) resulted in a significant increase in the methane

    decomposition rate, while air exhibited a relatively lowcarbon-activating efficiency (Fig. 13).

    Using a thermobalance, Abbas and Daud [126] studied the

    regeneration of an AC catalyst for six methane decomposition

    runs at temperatures of 850 and 950  C with five regeneration

    cycles using CO2  at temperatures of 900, 950 and 1000   C to

    evaluate the stability of the catalyst. They reported that rO and

    the ultimate mass gain of the catalyst decreased after each

    regeneration step at both decomposition temperatures, but

    the decrease was slower under severe regenerating conditions

    (Fig. 14). Additionally, a slower decrease in ultimate mass gain

    was obtained when the decomposition was carried out at

    950   C compared to the one carried out at 850   C, while no

    significant differences were observed for the decrease of   rOusing reaction temperatures of 950 or 850  C. Pinilla et al. [127]

    studied methane TCD using 4 g of AC as a catalyst at 850 C for

    three decomposition cycles, and the deactivated AC was

    gasified using CO2 at 900 and 925  C; in comparison with the

    first reaction cycle, in the second and third cycles it was

    observed that rO decreased by 65 and 30%, respectively, while

    the ultimate mass gain decreased by 88 and 27%, respectively.

    Although metal catalysts have higher activity for acceler-

    ating the methane TCD reaction and thus reduce the required

    reaction temperature compared with carbon-based catalysts,

    unfortunately, the catalyst activity is gradually lost as the

    reaction proceeds due to the covering of active sites with the

    carbon by-product. The regeneration of deactivated catalyst is

    always done by a burning off or gasification process which

    leads to CO2 production in amounts nearly comparable to the

    quantity of CO2   emitted by the SMR process   [100]. Another

    serious problem arising from oxidative regeneration of metal

    catalysts is related to the unavoidable contamination of 

    hydrogen with carbon oxides, which would require an addi-

    tional purification step   [31]. Takenaka et al.   [128]   studied

    methane TCD using 40 mg of Ni/AL2O3, Ni/SiO2 or Ni/TiO2 asthe catalyst at 550   C for five decomposition–regeneration

    cycles with CO2 used as the regeneration agent at 650  C. The

    results showed that the catalytic activity of Ni/TiO2 remained

    high during the repeated reactions, while it gradually

    increased for Ni/AL2O3. For the ultimate mass gain, it was

    reported that Ni/SiO2  showed the highest value, however, it

    decreased significantly after repeated cycles, while for Ni/TiO2it increased from the first to the third cycle and was then

    unchanged after the third cycle. As for Ni/AL2O3, the ultimate

    mass gain was very low; however, it increased gradually with

    the number of reaction cycles. The catalytic performance of 

    Ni/SiO2, Ni/TiO2, Ni/Al2O3   and Pd–Ni/SiO2   in the repeated

    decomposition of methane and oxidation of the CNFs formedwith O2 was studied by Otsuka et al. [129]. They reported that

    theoretically, the catalytic decomposition of methane (CH4/

    C  þ  2H2) followed by the oxidations of the deposited carbon

    with CO2 and O2 ((1/2)Cþ (1/2)CO2/CO, (1/2)Cþ (1/2)O2/ (1/2)CO2)

    gives an over–all reaction CH4   þ   (1/2)O2   /   2H2   þ   CO

    (DH1073¼20 kJ/mol), requiring no energy input and with zero

    CO2  emission. They also showed that Ni/TiO2, Ni/Al2O3  and

    Pd–Ni(1:3)/SiO2 (Fig. 15) were found to be promising catalysts

    because the catalytic activities completely lost during the

    preceding methane decomposition could be repeatedly

    recovered by the oxidation of the CNFs formed with O2.

    Other researchers have also evaluated the possibility of 

    catalyst regeneration by the removal of the carbon deposited

    Fig. 12 – Mass gain vs. time for different particle sizes at PCH4 [ 0.19 atm and 850   8C [37].

    Fig. 13 – Effect of carbon catalyst activation by different activating agents on the methane decomposition rate at 

    850   8C. Activation temperature: 950   8C [13].

    i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 1 1 6 0 – 1 1 9 01180

  • 8/16/2019 Hydrogen Production by Methane Decomposition - A Review 2010

    22/31

    Fig. 14 – Mass gain versus time for six reaction cycles at 950   8C and PCH4  of 0.63 atm using pure CO2 at 50 mL minL1 a) 900   8C

     b) 950  8

    C. c) 1000  8

    C [126].

    i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 1 1 6 0 – 1 1 9 0   1181

  • 8/16/2019 Hydrogen Production by Methane Decomposition - A Review 2010

    23/31

    on the external surface of the catalytic particles by attrition.However, the results showed that attrition phenomena are

    effective only on the carbon deposited on the external surface

    of the catalytic particle, and this carbon represents only

    a fraction of the total carbon produced by the decomposition

    process   [70,86,130]. In separate study, Jang and Cha   [88]

    investigated the effect of PBR and FBR reactors on methane

    TCD using Fe and Fe/Al2O3   catalysis, and found that the

    conversion rate of methane was maintained by attrition of the

    by-product carbon from the Fe catalyst surface. In modeling 

    methane TCD in an FBR, Ammendola et al.   [131] considered

    the attrition phenomena as a novel catalyst regeneration

    strategy, in which the carbon-attrition rate,   Ec, is able to

    balance carbon deposition rate. Attrition of carbon deposited

    on the external surface of catalyst particles results in carbonemission as carbon fines transported in the exit gases as well

    as in the renewal of a part of external active surface of the

    catalyst. According to the scheme shown in  Fig. 16A, they

    assumed that constant values of the H2-production and

    carbon-elutriation rates can be obtained. Fig. 16A reduces to

    Fig. 16B when attrition is not present.

    7. Characteristics of carbon produced fromthermocatalytic decomposition of methane

    As mentioned previously, carbons can be classified into

    different types according to their crystallinity or the degree of 

    Fig. 15 – Decomposition of methane at 823 K and successive oxidation of the CNFs formed with oxygen at 480   8C (furnace

    temperature) over Ni (5 wt.%)/SiO2 for repeated cycles. a) Kinetic curves of methane conversion in methane decomposition.

     b) Kinetic curves of oxygen conversion in the oxidation of carbon. c) Yields of hydrogen (H2 /Ni) at complete catalyst 

    deactivation. Flow rate of methane: 60 mL/min (101.3 kPa). Flow rate of Ar and O2 mixture: 75 mL/min (partial pressure of O2,

    20 kPa) [129].

    i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 1 1 6 0 – 1 1 9 01182

  • 8/16/2019 Hydrogen Production by Methane Decomposition - A Review 2010

    24/31

    order, i.e., from highly ordered carbons, such as graphite and

    diamond, to less ordered (turbostratic and pyrolytic carbons)

    and, finally, to disordered (amorphous and microcrystalline)

    carbons, such as AC, charcoal and CB. It has been reported

    that, depending on the operating conditions of the methane

    TCD process, carbon can be produced in several types:

    amorphous, turbostratic and CFs. The prices of turbostratic,

    pyrolytic and filamentous carbons are 0.3, (1–18) and more

    than 1000 $/kg, respectively. This wide difference in the prices

    of different carbon modifications shows the importance of the

    type(s) of carbon(s) produced in methane TCD in terms of reducing the cost of hydrogen production. Amorphous

    carbons are more active in methane TCD than well-ordered

    carbons such as graphite, diamond and CNT because the

    surface concentration of high-energy sites (where a regular

    array of carbon bonds is disrupted, forming free valences,

    discontinuities and other energetic abnormalities such as

    surface defects and dislocations) increases with the decrease

    in carbon-crystallite size and conversely decreases as carbon

    becomes more ordered. Accordingly, the catalytic activity of 

    carbons towards methane decomposition is in the following 

    order: amorphous >  turbostratic >  graphite [13,28,47].

    The formation of CF on some carbonaceous catalysts,

    similar to those formed when methane TCD is performed on

    metal-supported catalysts, has been reported (Fig. 17B). Thisphenomenon attributed to the presence of trace amounts of 

    metal components such as K, Na and Fe in the ash of the

    carbonaceous catalysts   [31,101,117,122]. The non-catalytic

    (homogenous) decomposition of methane when conducted at

    temperatures in excess of 1000–1100   C has led to the

    production of various forms of amorphous carbon, e.g., CB and

    thermal black [75].

    For metal catalysts, a generally accepted mechanism of 

    filamentgrowthconsists of the formation of carbon specieson

    the surface of the metals, dissolution in the metal, diffusion

    through the metal and precipitation from the metal at a point

    on the surface, resulting in the formation of the filament body.

    The rate-determining step of this process is the bulk diffusionof the carbon through the metal particle, while excess carbon

    migrates along the metal surface to form the skin component

    of the filaments [45,91]. Regarding the types of carbon depos-

    ited on a catalyst when methane TCD is conducted with metal

    catalysts, Ni catalyst have attracted more attention due to

    their high catalytic activity and the capability of producing CF

    at moderate temperatures (500–700  C). Bai et al. [85] showed

    that CF formation was observed at moderate conditions with

    a low Ni loading and that temperature had a significant effect

    on CF formation and morphology.

    When Fe-based catalysts were used, a higher temperature

    range was required for efficient operation; also, Fe-based cata-

    lysts were able to catalyze the formation of CNT. Reshetenkoet al. [132] showed that catalyst compositions and preparation

    methods influencedtheir properties, andthe introduction of Co

    or Ni in small amounts (3–10% mass) resulted in significant

    increasein carbonyield (two tothreetimesmorethanFe–Al2O3)

    and in the formation of multi-wall CNTs. The influence of 

    nickel-crystal domain size on the behavior of Ni and Ni–Cu

    Fig. 16 – A conceptual representation of thermocatalytic

    decomposition of methane on a single catalyst particle.

    A) Presence of attrition phenomena; B) absence of attrition

    phenomena [131].

    Fig. 17 – SEM micrographs of the used catalysts after activity tests: (a) AC; (b) CB,  T : 850   8C, methane flow: 20 mL/min [117].

    i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 1 1 6 0 – 1 1 9 0   1183

  • 8/16/2019 Hydrogen Production by Methane Decomposition - A Review 2010

    25/31

    catalysts has been studied by Pinilla et al. [133]. They reported

    a ‘‘herringbone’’ CNF structure for Ni–Cu/MgO catalysts, with

    diameters that correlated with the nickel-particle size, while

    carbon deposited with the Ni/Al2O3 catalyst was usually in the

    form of ribbon nanofibers. Fig. 18A shows a low-magnification

    TEM micrograph with some metal particles, mostly withrounded or diamond-like shapes, and emerging CNFs on the

    tips. Fig. 18B shows the detail of a diamond-like nickel-particle,

    whereas in Fig. 18C, the inner arrangement of the graphene

    layers forms a small angle of w22 with respectto thefiber axis.

    Thus, the carbon deposited is in the form of herringbone

    nanofibers with a hollow core. Fig. 18D is a high-magnification

    TEM micrograph showing graphene structures extending from

    one wall of the fiber to the other.

    Shah et al. [95] conducted methane TCD using nanoscale,

    binaryFe–M(M]Pd,Mo, or Ni)catalysts supportedon alumina.

    High-resolution SEM and TEM characterization indicated that

    almost all the carbon produced in the temperature range of 

    700–800  C was in the form of potentially useful multi-walled

    CNTs. At higher t


Recommended