In situ catalytic conversion of tar using rice husk char/ash supported nickel–iron catalysts for biomass pyrolytic gasification combined with the mixing-simulation in fluidized-bed gasifier

In situ catalytic conversion of tar using rice husk char/ash supportednickel–iron catalysts for biomass pyrolytic gasiﬁcation combined with Yafei Shena,⇑, Peitao Zhaob,c, Qinfu Shaod, Fu

Trang 1

In situ catalytic conversion of tar using rice husk char/ash supported

nickel–iron catalysts for biomass pyrolytic gasiﬁcation combined with

Yafei Shena,⇑, Peitao Zhaob,c, Qinfu Shaod, Fumitake Takahashia, Kunio Yoshikawaa

a

Department of Environmental Science and Technology, Interdisciplinary Graduate School of Science and Engineering, Tokyo Institute of Technology, Yokohama 226-8502, Japan

b

Key Laboratory of Energy Thermal Conversion and Control of Ministry of Education, School of Energy and Environment, Southeast University, Nanjing 210096, PR China

c

School of Electric Power Engineering, China University of Mining and Technology, Xuzhou 221116, PR China

d Department of Environmental System, Graduate School of Frontier Sciences, The University of Tokyo, Chiba 277-8563, Japan

h i g h l i g h t s

Tar can be in situ converted by the RHC/RHA supported nickel–iron catalysts

The tar conversion efﬁciency could reach about 92.3% by the RHC Ni–Fe

Partial metal oxides in the carbon matrix could be in-situ reduced into the metallic state

Mixing with RHA Ni can also improve biomass ﬂuidization behavior

a r t i c l e i n f o

Article history:

Received 4 August 2014

Received in revised form 25 September

2014

Accepted 27 October 2014

Available online xxxx

Keywords:

Biomass gasiﬁcation

Tar conversion

Rice husk char

Catalysts

Mixing-simulation

a b s t r a c t

A catalytic gasiﬁcation technology has been proposed for tar in situ conversion using the rice husk char (RHC) or rice husk ash (RHA) supported nickel–iron catalysts Biomass tar could be converted effectively

by co-pyrolysis with the RHC/RHA supported nickel–iron catalysts at 800 °C, simplifying the follow-up tar removal process Under the optimized conditions, the tar conversion efficiency could reach about 92.3% by the RHC Ni–Fe, which exhibited more advantages of easy preparation and energy-saving In addition, the tar conversion efficiency could reach about 93% by the RHA Ni Significantly, partial metal oxides (e.g., NiO, Fe2O3) in the carbon matrix of RHC could be in-situ carbothermally reduced into the metallic state (e.g., Ni0) by reducing gases (e.g., CO) or carbon atom, thereby enhancing the catalytic per-formance of tar conversion Furthermore, mixing with other solid particles such as sand and RHA Ni, can also improve biomass (e.g., RH) fluidization behavior by optimizing the operation parameters (e.g., par-ticle size, mass fraction) in the mode of fluidized bed gasifier (FBG) After the solid–solid mixing simula-tion, the RH mass fraction of 0.5 and the particle diameter of 0.5 mm can be employed in the binary mixture of RH and RHA

1 Introduction

Biomass pyrolysis and/or gasiﬁcation is recognized as one of the

most promising technologies for the production of sustainable

fuels used for power generation system or syngas applications

[1–5] Biomass gasiﬁcation is a process in which biomass

experiences incomplete combustion to produce syngas that mainly

consists of H2, CO, CH4, and CO2 Biomass gasification has many advantages over direct combustion It can convert low-value feedstocks to high quality gas products directly burned or used for electricity generation Syngas is also synthesized into liquid transportation fuels [6] Condensable organics referred as tar are produced with syngas during biomass gasification and their contents of 0.5–100 g/m3depending on the type and design of gas-ifier, feedstock types, and processing conditions[7] Herein, tar is composed of all organic compounds in syngas [8] Tar can con-dense in pipes, filters, or downstream equipment and processes, thereby breaking down the entire system Tars may also deactivate catalysts in the refining process Tar removal by efficient adsorption and reforming to syngas would be important and

http://dx.doi.org/10.1016/j.apenergy.2014.10.074

q

This paper is included in the Special Issue of Energy innovations for a

sustainable world edited by Prof S.k Chou, Prof Umberto Desideri, Prof Duu-Jong

Lee and Prof Yan.

⇑ Corresponding author Tel.: +81 45 924 5507; fax: +81 45 924 5518.

E-mail addresses: yafeisjtu@gmail.com , shen.y.ad@m.titech.ac.jp (Y Shen).

Contents lists available atScienceDirect

Applied Energy

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 o m / l o c a t e / a p e n e r g y

Please cite this article in press as: Shen Y et al In situ catalytic conversion of tar using rice husk char/ash supported nickel–iron catalysts for biomass

Trang 2

pyro-indispensable for commercialization [9] It is essential to reduce

the level of tars to enable widespread utilization of syngas Up to

now, several approaches such as physical treatment[10–12],

ther-mal cracking[13], plasma-assisted cracking[14,15], and catalytic

reforming[16–20], have been widely applied for tar elimination

Among these, catalytic reforming is considered the most promising

in large-scale applications due to fast reaction rate and reliability

[20]to increase the additional syngas such as CO and H2

Various types of catalysts such as calcined rocks[21], zeolites

[22], iron ores[23], alkali metals[24], Ni-based catalysts[25,26],

and noble metals[27–31]have been employed for tar removal in

biomass pyrolysis/gasiﬁcation Due to catalytic activity and

economic reasons, the nickel-based catalysts show considerable

performances on tar reforming[32–38] Nickel catalysts are usually

supported by metal oxides or natural minerals (e.g., dolomite,

olivine)[39–44] These supports are relatively expensive, and the

catalyst preparation steps are complex and energy consuming As

an alternative, char has been employed as a low-cost adsorbent/

catalyst in tar removal[45–49] The deactivated char catalysts could

be simply burnt or catalytic gasiﬁed to recover the energy without

the frequent regeneration However, biomass char has unﬁxed

properties depending on biomass type and process conditions

Recently, Wang et al [50] studied char and char-supported

nickel catalysts for secondary syngas cleanup Ni-based catalysts

were fabricated by mechanically mixing NiO with char particles

More than 97% of tars can be removed by using the Ni/coal-char

and Ni/wood-char catalysts (15% NiO loading) at 800 °C Iron-based

catalyst and additive iron have also been attracted more attention

For instance, Nemanova et al.[51]used the Fe based catalysts for

biomass tar decomposition Liu et al.[52]studied in details, the

dif-ferent metal additives (i.e., Fe, Mg, Mn, Ce) on catalytic reforming

of biomass tar over Ni6/palygorskite It has been proved that iron

plays a better role in improving its reactivity Fe-based catalysts

are cheaper and environmental friendly than Ni-based catalysts

[53] Besides, mono- or bi-metallic catalysts, such as Fe/Al2O3,

Co/Al2O3, Fe–Co/A12O3 and Ni–Co/A12O3 were beneﬁt for steam

reforming of tar[54,55] In this work, we investigated in-situ tar

conversion by co-pyrolysis of biomass and the rice husk char

(RHC) and rice husk ash (RHA) supported nickel–iron catalysts

for biomass gasiﬁcation Subsequently, the RHA-supported

cata-lysts could be considered as bed materials by mixing-simulation

to optimize the mass fraction and particle size used for the

ﬂuid-ized-bed gasiﬁcation

2 Materials and methods

2.1 Biomass and char characterization

Biomass feedstock of RH was collected from Thailand.Table 1

shows the properties of RH, RHC and RHA including the ultimate

and proximate analysis, which were conducted by the elemental analyzer (EA, Vario MICRO Cube, Elementar, Germany) and the ther-mogravimetric analyzer (TGA, DTG-50, SHIMADZU, Nakagyo-ku, Kyoto, Japan), respectively In addition, the chemical composition

of RHA was determined by the X-ray ﬂuorescence (XRF, SHIMADZU, Rayny EDX 700, Japan) It can be found that RHA is composed of a mass of SiO2, up to 94.64%, and a small quantity of minerals, such

as alkali or earth alkali metal oxides If RHC or RHA is employed as

a catalyst support, it is necessary to introduce some metal species (e.g., Ni and Fe) with the aim of improving the catalytic perfor-mances Furthermore, the BET surface area of RHC was larger than those of RH and RHA It could be concluded that RHC is more suit-able to be used as a support for catalysts preparation

2.2 Catalysts preparation RHC was fabricated by slow pyrolysis from 25 to 700 °C in an inert gas (i.e., N2) atmosphere The general procedure of the RHC/ RHA catalysts preparation was illustrated inFig 1 Three RHA Ni (Ni2+: 0.2 mol/L), RHA Fe (Fe3+: 0.2 mol/L) and RHA Ni–Fe (Ni2+: 0.1 mol/L, Fe3+: 0.1 mol/L) catalysts were simply prepared by the incipient wetness impregnation and calcination using the Fe(NO3)39H2O and Ni(NO3)26H2O as iron and nickel precursors After impregnation and drying overnight at 105 °C, the metal spe-cies in RHC were calcined in air at 600 °C for 1 h before storage and further use However, RHC Ni–Fe (Ni2+: 0.1 mol/L, Fe3+: 0.1 mol/L) was prepared without calcination Furthermore, the fresh and used catalysts were characterized by the TGA and the X-ray diffraction (XRD, Rigaku, XRD-DSC II, Japan), respectively

2.3 Biomass gasification and tar conversion The main experimental parameters of operating condition are shown inTable 2.Fig 2A presents a schematic diagram of experi-mental apparatus, composed of a gas supplying system, a gas cleaning system and a pyrolysis-reforming facility First, the gasifi-cation temperature was heated to 800 °C Then, the carrier gas of nitrogen (N2) was continuously leaded into the entire system before adding the feedstock to ensure the gasification conducted

in the absence oxygen When the RH blended with the RHC or RHA catalysts was fed into the pyrolyzer (ﬁrst-zone), the volatile matters were released in the forms of gas and tar (Fig 2B) Conse-quently, biomass tar could be in-situ cracked and transformed by thermochemical reactions and catalytic conversion Finally, the residual tar was condensed and collected in the gas-cleaning unit 2.4 Sampling and analysis

The condensable tar can be determined by weighing[56] And the yield of producer gas including the non-condensable tar was Table 1

The properties of RH, RHC and RHA.

Ultimate analysis (wt.%, dry and ash free basis) Proximate analysis (wt.%, dry basis/as received) S BET (m 2 /g)

Chemical composition of RHA (wt.%)

a

Calculated by mass difference.

b

VM-volatile matters.

c

FC-ﬁxed carbon.

d

700 °C char.

Trang 3

pyro-calculated by mass balance The produced syngas mainly

com-posed of H2, CH4, CO, CO2, and C2 (i.e., C2H4, C2H6) was collected

by using an air bag at the outlet and measured by the micro gas

chromatograph (Agilent, Micro GC, 3000A, America), which was

ﬁt-ted with a thermal conductivity detector (TCD) Each trial was

maintained for 10 min to ensure the mass balance and the

repeat-ability experiments were performed Therefore, the collected tar

sample was the total amount of tars generated from the

repeatabil-ity experiments

3 Results and discussions 3.1 Catalysts characterization

Fig 3shows the thermogravimetric (TG) analysis of the RHC and RHA/RHC supported catalysts under the air atmosphere It can be observed that the RHA supported catalysts after calcination had higher thermal stability ascribed to the ash-basis Neverthe-less, the mass of RHC and RHC Ni–Fe was decreased with the increase of heat temperature under the air condition When the temperature was above 400 °C, the mass of RHC was decreased rapidly due to the char-basis After that, the mass kept constant after the temperature up to 600 °C It could be suggested that RHC Ni–Fe are appropriate for tar catalytic conversion in the oxy-gen-free or oxygen-less atmosphere to extend its use longevity It

is known that carbon in RHC could react with plenty of oxygen agents to generate carbon oxides (i.e., CO, CO2) at high tempera-tures The carbonaceous residue in the presence of high content

of silica had considerable thermal stability and abrasive resistance From TG curves of the RHC and RHC Ni–Fe, it indicated that the temperature range of 400–450 °C could be chosen for thermal regeneration of the RHC supported catalysts

Fig 1 Schematic diagram of the RHC/RHA Ni–Fe catalysts preparation procedure.

Table 2

Main experimental parameters of operating conditions.

Fig 2 Schematic of the lab-scale experimental setup (A) and biomass gasiﬁcation scheme (B).

Trang 4

pyro-The identiﬁcation of crystal phases was performed by XRD

using Rigaku D/Max 3400 powder diffraction with Cu Karadiation

(k = 0.1542) at 45 kV and 40 mA with a scanning rate of 5°/min

Fig 4 shows the XRD patterns of the RHC, the fresh and used

RHC/RHA supported nickel–iron catalysts Moreover, their

charac-teristics peaks are listed inTable 3 The typical amorphous silica

characteristic peak in RHC is observed at a broad peak centered

at 2h = 22.5°, which is attributed to the presence of the disordered

cristobalite (SiO2)[57] In the RHA supported metal (i.e., Fe, Ni)

cat-alysts, the main crystal phases were metal oxides As for the fresh

RHA Fe1, the iron crystal phases were in the forms of iron oxide

and magnesioferrite, which might be caused by iron oxide

sintering with the mineral MgO in RHA Nevertheless, the iron crystal phases in the used RHA Fe2 were both iron oxide and metallic iron During the pyrolysis, partial iron oxide in RHA Fe1 might be reduced into the metallic iron (i.e., Fe0) by the carbother-mal reduction(R1) and (R2)and hydrogen reduction(R3) In the same way, partial nickel oxides in the fresh RHA Ni were reduced into the metallic nickel (i.e., Ni0) after used for biomass catalytic pyrolysis In addition, the bimetallic catalysts of RHA Ni–Fe and RHC Ni–Fe can form the crystal structures of nickel iron oxides The nickel oxides in the carbon matrix of RHC could be much easier reduced into the metallic nickel, while the iron oxides (e.g., Fe2O3) were transformed into the magnetites(R4)

C ðsÞ þ NiO ðsÞ ! Ni ðsÞ þ CO ðgÞ ðR1Þ

CO ðgÞ þ NiO ðsÞ ! Ni ðsÞ þ CO2ðgÞ ðR2Þ

H2ðgÞ þ NiO ðsÞ ! Ni ðsÞ þ H2O ðgÞ ðR3Þ

Fe2O3! Fe3O4! FeO ! Fe0 ðR4Þ

3.2 Tar yield and conversion efﬁciency

Fig 5shows the condensable tar yield and conversion efﬁciency with the different char-supported catalysts From the tar instance graphs after in-situ conversion, it could be intuitively observed that tar yield was decreased through co-pyrolysis of RH with the RHC/ RHA supported catalysts The conversion efﬁciency of biomass tar was around 42% only mixed with RHC; whereas it can reach about 93% by co-pyrolysis with RHA Ni It can indicate that the Ni-based catalysts have higher tar cracking/reforming performances Metal-lic nickel (Ni0) catalyst possesses much higher reforming activity of hydrocarbons than metallic iron (Fe0) catalyst caused by the high activation ability of C–H and C–C bond in the hydrocarbon

Fig 3 TG curves of RHC and RHA/RHC supported catalysts under the air

atmosphere.

Fig 4 XRD patterns of the RHC, the fresh (1) and used (2) RHC/RHA supported catalysts.

Trang 5

pyro-Table 3

XRD characteristic peak lists of the fresh and used RHC/RHA supported catalysts.

(cps)

FWHM (deg)

Int I (cps) Phase name RHA Fe 1

3 35.639(12) 2.5172(8) 274(17) 0.27(2) 120(3) Magnesioferrite, syn(3, 1, 1), Iron oxide(1, 1, 0)

12 71.3758 1.3204 202.9877 0.4004 2.9033 Magnesioferrite, syn(6, 2, 0), Iron oxide(1, 0, 10)

RHA Fe 2

RHA Ni 1

RHA Ni 2

3 44.461(10) 2.0360(4) 846(29) 0.189(18) 261(5) Nickel, syn(1, 1, 1)

RHA Ni–Fe 1

3 35.63(2) 2.5178(14) 167(13) 0.35(4) 110(3) Nickel iron oxide(3, 1, 1), Hematite, syn(1, 1, 0)

4 37.246(14) 2.4121(9) 176(13) 0.24(3) 78(3) Nickel oxide(0, 0, 3), Nickel iron oxide(2, 2, 2)

5 43.262(15) 2.0897(7) 293(17) 0.28(3) 140(3) Nickel oxide(0, 1, 2), Nickel iron oxide(4, 0, 0), Hematite, syn(2, 0, 2)

8 57.58(4) 1.5995(9) 21(5) 0.60(10) 15(2) Nickel iron oxide(5, 1, 1), Hematite, syn(1, 2, 2)

9 62.85(2) 1.4773(5) 151(12) 0.41(5) 112(3) Nickel oxide(1, 1, 0), Nickel Iron Oxide(4, 4, 0)

10 75.36(2) 1.2602(3) 29(5) 0.39(7) 12(2) Nickel oxide(0, 2, 1), Nickel iron oxide(6, 2, 2), Hematite, syn(2, 1, 7)

11 79.37(8) 1.2062(10) 26(5) 0.42(15) 19(2) Nickel oxide(0, 0, 6), Nickel Iron oxide(4, 4, 4), Hematite, syn(1, 3, 1)

RHA Ni–Fe 2

1 35.93(5) 2.497(3) 16(4) 0.71(15) 13(4) Tetrataenite(1, 0, 1), Maghemite-Q, syn(2, 2, 5), Magnetite, syn(3, 1, 1)

2 43.591(19) 2.0746(9) 180(13) 0.44(4) 163(7) Tetrataenite(1, 1, 1), Maghemite-Q, syn(0, 0, 12), Magnetite, syn(4, 0, 0), Nickel

oxide(2, 0, 0)

3 44.412(18) 2.0382(8) 137(12) 0.22(3) 45(6) Nickel, syn(1, 1, 1), Maghemite-Q, syn(4, 0, 2)

4 50.87(3) 1.7935(9) 59(8) 0.29(4) 26(4) Tetrataenite(0, 0, 2), Maghemite-Q, syn(2, 2, 11)

5 51.75(8) 1.765(2) 51(7) 0.60(10) 49(6) Nickel, syn(2, 0, 0), Maghemite-Q, syn(4, 2, 4)

6 63.1114 1.4719 4.6439 0.5974 13.9212 Tetrataenite(1, 1, 2), Maghemite-Q, syn(3, 3, 11), Magnetite, syn(4, 4, 0), Nickel

oxide(2, 2, 0)

7 74.7538 1.2689 43.4666 0.5974 23.0236 Tetrataenite(2, 0, 2), Maghemite-Q, syn(5, 4, 3), Magnetite, syn(5, 3, 3)

8 76.165 1.2489 37.5747 0.5974 27.5757 Nickel, syn(2, 2, 0), Maghemite-Q, syn(5, 0, 13), Magnetite, syn(6, 2, 2), Nickel oxide(3, 1, 1) RHC Ni–Fe 1

5 32.9989 2.7123 19.4265 0.3163 10.2001 Nickel nitrate hydroxide hydrate(3, 1, 0), Hematite, syn(1, 0, 4)

RHC Ni–Fe 2

(continued on next page)

Trang 6

pyro-molecules on the nickel metal surface [58] Comparing the two

bimetallic catalysts of RHA Ni–Fe and RHC Ni–Fe, the RHC Ni–Fe

exhibited higher tar conversion efﬁciency (92.3% vs 86%) Based

on the catalysts characterization, RHA Ni–Fe exists mainly in form

of the silica-based catalyst, whereas RHC Ni–Fe is the carbon–silica

hybrid-based catalyst Carbon in the RHC Ni–Fe plays a signiﬁcant

role for tar conversion On one hand, porous carbon could increase

the BET surface areas of catalysts contributing to the sorption

effect; one the other hand, carbon itself can work as a medium

decreasing metal oxides and tar at high temperature by reductive

reactions Guan et al.[9]proposed that partial metal oxides might

be reduced into a metallic state by the reducing gases in syngas

(i.e., H2 and CO) produced from the biomass pyrolysis without

the aid of the catalyst Therefore, Fe and Ni in their metallic forms

rather than oxide forms were considered the main active sites for

the tar reforming Besides, it is possible that amorphous NiO in

the RHC Ni–Fe catalyst might easier to be reduced into the metallic

state of Ni (Ni0) than the crystalline NiO in RHA Ni–Fe RHC Ni–Fe

without calcination could be used for tar conversion as well The

synergy effect between the activation of tar on the Ni species

and the oxygen atom supplied to the carbonaceous intermediate

from neighboring Fe atoms was not displayed due to the low

cat-alytic activity of iron oxide at lower temperature and pressure

Moreover, probably in the absence of steaming water, the Fe

distri-bution in the samples after pyrolysis exhibits an imbalance

between the phases FeO and Fe3O4providing for tar conversion

Those Fe species always occur in the redox equations of the water

gas shift reactions(R5) and (R6) [59] Therefore, steam reforming

by using the char-supported monometallic Fe and bimetallic

Ni–Fe catalysts should be further studied to enhance the tar

conversion efﬁciency and H2yield

Fe3O4ðsÞ þ CO ðgÞ $ 3FeO ðsÞ þ CO2ðgÞ ðR5Þ

3FeO ðsÞ þ H2O ðgÞ $ H2ðgÞ þ Fe3O4ðgÞ ðR6Þ

3.3 Gas yield and composition Synthesis gas is the prime product of biomass pyrolysis Their properties can reﬂect the pyrolysis and tar conversion efﬁciency with the different catalysts, because tar is cracked or transformed into the gas molecules by catalytic reforming The gas yield could

be estimated by expression(E1) In addition, the volume concen-tration of the dominate syngas components (i.e., H2, CO, CO2,

CH4, C2H4and C2H6) can be calculated by expressions(E2)–(E6)

Gas yield ðL=gÞ ¼Exit gas ðLÞ N2flow rate ðL=minÞ Pyrolysis time ðminÞ

Feedstock weight ðgÞ

ðE1Þ

H2ð%Þ þ CO ð%Þ þ CO2ð%Þ þ CH4ð%Þ þ C2H4ð%Þ þ C2H6ð%Þ

ðE2Þ

ðE3Þ

ðE4Þ

ðE5Þ

C2 ðvol:%Þ ¼ C2H4ð%Þ þ C2H6ð%Þ

ðE6Þ

Fig 6 shows the producer gas yield and syngas composition when different catalysts were used It can be found that the amount of gas yield increased when the RHC and RHC-supported catalysts were applied The increase of gas yield may be attributed

to the thermochemical reactions between char, tar and catalysts at higher temperatures On one hand, char can react with syngas (i.e.,

CO2and H2) to produce more other syngas components (i.e., CO,

CH4); on the other hand, tar can be cracked/converted into gas components by dry reforming over RHC and RHC-supported cata-lysts More importantly, the further devolatilization of char could also contribute to the increase of gas yield In particular, the gas yield can reach approximately 2.11 L/g when co-pyrolysis of RH and RHC Ni–Fe at 800 °C When char was blended with RH, the

CO volume concentration increased from 44.8% to 52.0%; whereas, the CO2volume concentration decreased from 24.0% to 15.8% It suggested that CO2 most likely reacted with char by Boudouard reaction in the presence of nickel catalysts, leading to the increase

of CO volume concentration Meanwhile, the methane (CH4) vol-ume concentration slightly increased from 7.5% to 8.0%, possibly

Table 3 (continued)

(cps)

FWHM (deg)

Int I (cps) Phase name

3 43.43(3) 2.0818(12) 121(11) 0.6597 157.4386 Iron nitride(1, 1, 1), Maghemite, syn(4, 0, 0)

4 44.4788 2.0353 63.8036 0.3719 25.2644 Maghemite, syn(4, 1, 0), Nickel, syn(1, 1, 1)

5 50.58(7) 1.803(2) 33(6) 0.5454 38.7336 Iron nitride(2, 0, 0), Maghemite, syn(4, 2, 1)

10 74.42(8) 1.2738(12) 18(4) 1.8(2) 72(4) Iron nitride(2, 2, 0), Maghemite, syn(5, 4, 1), Nickel, syn(2, 2, 0)

Fig 5 Heavy tar yield and conversion efﬁciency using char and different char

catalysts.

Trang 7

pyro-ascribed to the methanation reactions between CO2, C and H2 In

addition, char itself could play the role of an adsorption-type

cata-lyst for tar and CO2conversion

Because of char further decomposition and catalytic effect, the

gas yield by mixing with the RHC Ni–Fe was slight higher than

the gas yield by mixing with the RHA Ni–Fe (2.11 L/g vs 1.96 L/

g) Regarding to the volume concentrations of CO (55.2% vs

41.8%) and H2(22.7% vs 31.5%), it was concluded that H2could

be slightly consumed in the presence of char and metal oxides at

high temperatures Herein, high concentration of H2 was not

achieved, but the volume concentrations of CO and CH4 got

improved It should be noted that compared to raw yield (24.0%),

CO2volume concentrations were greatly decreased (the yields as

follows: 15.8%, 11.9%, and 8.2%) by mixing with RHC, RHC Ni–Fe

and RHA Ni, respectively Furthermore, it is accorded with the

previous result that NiO can play a critical role in decreasing the

carbon deposit and increasing the amount of CO in the gaseous

product[60] The lower heating value (LHV) and the higher heating

value (HHV) of the produced syngas is calculated by the empirical

expressions of(E7) [61]and(E8) [62,63], respectively, where (CO),

(H2), and (CH4) are molar fraction of the produced syngas inFig 6

The theoretical results are presented inTable 4 The LHV and HHV

could reach about 10.64–12.80 MJ/m3, and 13.02–14.55 MJ/m3,

respectively, at 800 °C, indicating a good quality of the syngas

Moreover, the producer gas showed a higher HHV by using the

RHC supported catalysts

LHV ðkJ=m3Þ ¼ ½30ðCOÞ þ 25:7ðH2Þ þ 85:4ðCH4Þ 4:2 ðE7Þ

HHV ðkJ=N m3Þ ¼12:63ðCOÞ þ 12:75ðH2Þ þ 39:82ðCH4Þ þ 63:43ðCnHmÞ

100

ðE8Þ

3.4 Tar catalytic conversion mechanisms

In this work, monometallic Ni catalysts exhibited much higher

reforming activity of hydrocarbons than monometallic Fe catalysts

This property is caused by the high activation ability of C–H and

C–C bond in the hydrocarbon molecules on the metal (Ni) surface Consequently, it seems that the additive effect of Fe is the increase

of the number of active Ni surface, while the characterization results in the particle size and surface enrichment of Fe on the Ni–Fe bimetallic particles did not support the increase of the sur-face Ni atoms Another reason might be the co-catalytic function

of Fe Since Fe has high oxygen afﬁnity than Ni, the addition of

Fe to Ni catalysts can increase the coverage of oxygen atoms during the reforming reactions The catalytic activity of the RHC/RHA sup-ported Ni–Fe catalysts for tar conversion can be concluded as the following order: RHA Ni > RHC Ni–Fe > RHA Ni–Fe > RHA Fe > RHC

In summary, biomass was initially decomposed into the small pieces of gas, tar and char by thermochemical reactions in the pyrolyzer The produced tar could be further cracked and reformed simultaneously through the RHC/RHA supported Ni–Fe catalysts at high temperatures Herein, RHC most likely plays two signiﬁcant roles in the process of biomass pyrolysis On one hand, it works

as an intermediate carbon source to reduce the metal oxides by carbothermal reduction; on the other hand, it works as a carbon adsorbent to insert metal cations and tar

3.5 Mixing-simulation in fluidized bed gasifier (FBG) Gasification of biomass and wastes in FBG has advantages, since FBGs are capable of being used in the pilot and large scales, over-coming limitations found in smaller scale, fixed-bed gasifiers On the other hand, the bed temperature is limited to avoid the bed agglomeration and the gasification efficiency of a fluidized bed (FB) may be limited if part of the fuel energy remains in uncon-verted char Meanwhile, if the temperature is not high enough in the gasifier, tar in the producer gas can make the process unsuit-able from a technical and economical point of view Models can

be helpful for design of gasiﬁer, for prediction of operation behav-ior and emissions during normal conditions, startup, shutdown, changes of fuel and load The modeling could be carried out from preliminary design of an industrial process to complex simulation

of a unit Experiments, especially at large scale, are usually expen-sive and complicated Nevertheless, modeling is economy and con-venient and it can support the preparation and optimization of experiments to be conducted in a real system[64] The tools avail-able for modeling of the FBG reactors are the more or less simpli-ﬁed equations for conservation of mass, energy and momentum, which complemented by boundary conditions, constitutive rela-tionships, and terms expressing the sources and sinks of the sys-tem To determine the latter, rate laws for the chemical or physical conversion processes are required Thermodynamic data are considerable to estimate properties and thermal data as well

as reaction products by equilibrium assumptions

Fig 7 schematically presents each process occurs in an FBG including the bed level with bubble and emulsion phases, the par-ticle level with gases release and char gasiﬁcation, and the gas phase reactions where water gas shift reaction plays a signiﬁcant role Some processes strongly interact between one level and another For instance, the heat and mass transport to a particle takes place on the particle level, while their rates are determined

by the ﬂuid dynamics of the bed (reactor level) and by fuel reactiv-ity, both in case of devolatilization and char conversion Moreover, these processes are included in source terms of the conservation equations treated by the sub-models during execution of numeri-cal numeri-calculations The description on the particle level is composed

of the particle size and biomass properties, such as density and thermal conductivity, which affect the devolatilization time and the volatiles composition On the reactor level various factors are considered: residence time (mass inventory in an FB), boundary conditions, just like fuel feed points and feed rate, freeboard size,

Fig 6 Gas yield and composition with the RHC and RHA/RHC supported catalysts.

Table 4

The LHV and HHV of the produced syngas using RHC and RHC/RHA supported

catalysts.

No

catalyst

Fe RHA Ni–Fe RHA Ni RHC Ni–Fe LHV (MJ/m 3

HHV (MJ/m 3

Trang 8

pyro-ﬂuidization velocity, and their effects on solids elutriation, solids

and gas mixing, segregation, etc.[64]

The performance of FBG, including the carbon conversion

efﬁ-ciency and the tar content, may signiﬁcantly depend on the

move-ment of solids and gas in bed and freeboard For instance, in a pilot

circulating FBG (CFBG), reactor conditions resulting in lack of

tact between char/oxygen and tar/catalyst and an unfavorable

con-sumption of oxygen by the devolatilization gases were identiﬁed,

resulting in lower char conversion and higher tar content in the

product gas Modeling of solids and gas mixing can identify the

best arrangement for design and operation of the gasiﬁer

Particu-larly, the solids mixing simulation could optimize the operation

parameter of solid particles (i.e., biomass, sorbent, catalyst, bed

material) in FBG as well

In theory, the solid–solid mixing of species i could be

deter-mined by the expression(E9), applied for transport of solid (index

s) in an isothermal system The closest representation of the real

process is a balance on the transported variables formulated and

solved for each phase k (gas and solids and their i components):

density, momentum, and enthalpy (qk,qk,i,lk, and hk, in general

terms,u) The balance of the conserved variablesuover a ﬁxed

element (eulerian formulation) of reactor volume can be expressed

in the following form (somewhat simpliﬁed, in the case of

momen-tum), applicable to any reactor type: the accumulation ofuis due

to the net difference between the rates of change by convection

and dispersion or by generation and consumption, S, per unit

volume

@uk

@t þ divðukukÞ ¼ divðD/kgradukÞ þ Su ;k ðE9Þ

@csi

@t þ divðuscsiÞ ¼ divðDsV=H;igrad csiÞ þ Ssi ðE10Þ

The solids movement in an FB is often described by the simple

version of the expression(E9), in which diffusion and convection

are lumped into one term, called dispersion and expressed by the

ﬁne particles and deep beds, i.e., chemical reactors, where the

small-scale mixing mechanisms are dominant In the FB, besides

the inert bed material or solid catalyst, there is a distribution of fuel and char particles in the bed Therefore, the movement of sol-ids in the vessel should be described by accounting for three or more solids types Nevertheless, qualitatively the motion of biofuel particles could be visualized as the movement of ﬂotsam particles

in a jetsam-rich bubbling FB (BFB) High superficial velocity might improve the mixing behavior, but biomass particles with lower density and larger size than bed material particles are still non-uniformly distributed At a given fluidization velocity, char parti-cles are most likely to be elutriated from the bed or to sink from the bed surface than devolatilizing particles The reason is that the jet force from escaping volatile matters tends to keep them floating A key issue in modeling FBG is whether the fuel particles keep floating, once they have reached the bed surface or if they are compelled to descend This depends much on the segregation behavior of a few flotsam particles in a bed of many jetsam parti-cles Thereby, segregation should be avoided to preserve the bed from sintering or excessive tar emission in an FBG Segregation is most likely to occur if the ratio is equal to or below 0.5 A mixing ratio between 0.5 and 1 is desirable to be out of segregation prob-lems[64] Nevertheless, Fermoso et al.[55]investigated a sorption enhanced catalytic steam gasification of biomass in a combined downdraft FB and fixed bed reactor The solids feed was composed

of 15 wt.% of raw biomass and 85 wt.% of a mixture of sorbent and catalyst particles (sorbent/catalyst = 9 g/g) with the aim of produc-ing high purity hydrogen (>99.9 vol.%) It can be indicated that the mixing ratio of biomass and catalyst is, in a great extent, deter-mined by practical demands

Fluidization mainly depends on the bed pressure drop and flu-idization velocity As shown inFig 8, when the fluid velocity is too small, the solid particles will remain on the bed The phase between A and B is the fixed bed phase When the gas velocity exceeds to B phase, the bed pressure drop decreases slightly attrib-uted to the loose arrangement of solid particles After that, the gas velocity continues increasing, whilst the bed pressure drop keeps constant and the bed height increases gradually In this moment, the solid particles can float in the fluid and roll up and down with the gas movement, which is called the fluidized bed phase When Fig 7 Description of processes in an FBG.

Trang 9

pyro-the gas velocity is much higher than E point, pyro-the ﬂuid can taken

away the whole bed; consequently, the solid particles form the

dilute phase in the state of suspension and are blown out After E

phase, the normal ﬂuidization state is broken, and the bed pressure

drop decreases rapidly Thus, the gas velocity in point E is referred

as the terminal or maximum ﬂuidization velocity ut

Qualitatively judging the tendency of a particle to be carried

away can be made by calculating the terminal velocity of a single

particle ut

/¼s

where / is the particle sphericity to account for particle shape, s is

the surface of a sphere having the same volume as the particle and S

is the actual surface area of the particle In the previous study, a

simple general correlation for predicting terminal velocities for

iso-metric particles, given information on the particles and physical

properties of the ﬂuid, can be deﬁned as the expressions(E12)–

(E15) [65]

u

t¼ 18

ðdp Þ2þ

ð2:335 1:744/Þ

ðdp Þ0:5

ð0:5 6 / 6 1Þ ðE12Þ

u

t¼ ut

q2

g

glgðqsqgÞ

" #1=3

¼Rep;t

dp¼ dp

gqgðqsqgÞ

l2

g

" #1=3

Rep;t¼qgdput

where utis the terminal velocity of particle in ﬂuid (m/s), ut⁄is

the dimensionless particle velocity (m/s), dpis the solid equivalent

particle diameter (m), dp⁄is the dimensionless particle diameter

(m), Rep,tis the Reynolds number based on the equivalent spherical

diameter of particle,qgis the density of gas (kg/m3),qsis the

den-sity of solid particle (kg/m3),lgis the viscosity of gas [kg/(m s)], g is

the acceleration due to gravity (9.81 m/s2), and Ar is the Archimedes

number

If the ﬂuidizing medium was dry air and the measurements

were conducted under a temperature of 20 ± 1 °C and the ambient

pressure, the density and dynamic viscosity of the air could be

esti-mated from the expressions of(E16) and (E17) [66]

qg¼ 3:485P

lg¼ 1:81 105 T

293

0:66

ðE17Þ

In general, the terminal ﬂuidized velocity utcould be calculated

by the semi-empirical expressions of(E18)–(E20), if the Reynolds number is estimated Additionally, the superﬁcial gas velocity umf

(m/s) at minimum ﬂuidization can be calculated by the empirical expression of(E21)

ut¼gd

2

pðqsqgÞ

18l ; Rep;t60:4 ðE18Þ

ut¼ 4 225

gðqsqgÞ2

qgl

" #1

dp; 0:4 6 Rep;t6500 ðE19Þ

ut¼ 3:1gdpðqsqgÞ

qg

" #1

; Rep;tP500 ðE20Þ

umf ¼ 0:695d

1:82

p ðqsqgÞ0:94

l0:88q0:06 g

ðE21Þ

Mixtures of solid particles of different size and density tend to separate in vertical direction under ﬂuidized conditions The non-uniform distribution of the different solid components is caused

by a competitive action of mixing and segregation mechanisms

RH has very low bulk density (96–160 kg/m3) with a very low ter-minal velocity ut (1.0–1.4 m/s based on its physical properties)

[67] The fluidization characteristics of single RH could be observed from Fig 9 When the superficial gas velocity ugis smaller, the pressure drop increases with the increase of superficial gas velocity

ug However, the pressure drop decreases due to gas block caused

by the formation of channeling and cavitas, when the superficial gas velocity ug is increased to 0.27 m/s Although the gas flow increases, most of bed materials keep stationary Meanwhile, small amounts of RH particles are entrained from channeling to bed by gas, and thus the cavitas and channeling continue caving and form-ing Even if the pressure drop trends to be keeping constant, the RH particles cannot perform the fluidization behavior in the whole process

Rao and Ram[68]also proved that it is difﬁcult to ﬂuidize single

RH, and its fluidization behavior is improved by mixing with other solid particles (i.e., sand) The biomass constituted 2%, 5%, 10% and 15% determined by weight of the mixtures The minimum fluidiza-tion velocity umfincreased with the increase of biomass mass frac-tion, as well as with increasing sand density and particle size It also can be observed inFig 10A that the bed pressure drop versus superficial gas velocities is plotted with the aim of determining the minimum fluidization velocity umf of mixture, which is obtained Fig 8 Effect of gas velocity on the pressure drop in FBG.

Fig 9 Characteristic curve of ﬂuidization of single RH with the particle density of

950 kg/m 3

and the particle diameter of 2 mm.

Trang 10

pyro-from the intersecting point of the curve of ﬁxed bed at

defluidiza-tion with the constant pressure line at the flow condidefluidiza-tion

More-over, the minimum ﬂuidization velocity umfcan increase with the

increase of the averaged mass fraction of RH, and decreases with

the decrease of the sand particle size (Fig 10B) The equivalent

diameter of a sphere RH particle dr,avcould be obtained by

calculat-ing the Ergun Eq.(E22)at a given superﬁcial gas velocity ug[69]

Consequently, the appropriate RH particle size could also be

obtained by measuring the pressure drop, in which the calculated

equivalent diameter of RH particles is 1.54 mm

DP

H ¼ 150

ð1 egÞ2lgug

e3

gd2r;av þ 1:75

1 eg

e3 g

qg

dr;avu2

As shown inTable 5, although the bulk densityqbof the mixed RHC/RHA is still low, its particle densityqpand particle diameter dp

are comparable to the sand particles Moreover, the particle size of the RHA-supported catalysts could be modified easier than sand particles It is most likely that mixing with the RHA-supported cat-alysts could improve RH fluidization behavior in FBG Thus, with the aim of implementing a fluidized state, it is necessary to simu-late the particle sizes of RH and RHA in a static system of binary mixture In this case, the fluidizing medium is assumed dry air, car-rying out at a temperature of 20 ± 1 °C under the ambient pressure, the gas densityqgand the dynamic viscositylgcan be estimated as 1.2 kg/m3and 4.26 105kg/(m s), respectively In the BFB, when biomass and inert particles are blended relatively homogeneous in

Fig 10 (A) Pressure drop of RH-sand and RH-silica sand binary mixture as a function of superﬁcial gas velocity; (B) Minimum ﬂuidization velocity as a function of mass fraction of RH particles The density sand particles is 2600 kg/m 3

with the average diameters of 360 and 440lm; the density of silica sand particles is 2700 kg/m 3

with the average diameters of 360 and 710lm; the density of RH particles is 950 kg/m 3 with the average dimension of 10 2 1 mm.

Table 5

The properties of RH, RHA and sand.

Density (kg/m 3

Fig 11 Effect of (A) mass fraction (RH particle diameter: 2 mm) and (B) particle diameter (RH mass fraction: 0.5) on the minimum ﬂuidization velocity in the modeled binary mixture of RH and RHC.

Định dạng
Số trang	12
Dung lượng	2,4 MB