Fuel 107 (2013) 162–169

Contents lists available at SciVerse ScienceDirect

Fuel

journal homepage: www.elsevier .com/locate / fuel

Autothermal reforming of Fischer–Tropsch diesel over alumina and ceria–zirconiasupported catalysts

Angélica V. González a,⇑, Xanthias Karatzas b, Lars J. Pettersson a

a KTH Royal Institute of Technology, Chemical Science and Engineering, Chemical Technology, Teknikringen 42, SE-100 44 Stockholm, Swedenb Scania CV AB, SE-151 87 Södertälje, Sweden

h i g h l i g h t s

" Full-scale Fischer-Tropsch diesel autothermal reforming for hydrogen generation." Two catalytic zones were used. Bimetallic RhPt catalysts supported on d-Al2O3 and CeO2-ZrO2.

" Catalytic monoliths were sequentially placed in the autothermal reformer." Higher reducibility of RhxOy species in CeO2-ZrO2 lead to 42% vol hydrogen after CeO2-ZrO2.

" Optimum operating conditions were determined at O2/C = 0.42 and H2O/C = 2.5.

a r t i c l e i n f o

Article history:Received 20 March 2012Received in revised form 20 December 2012Accepted 4 February 2013Available online 19 February 2013

Keywords:Ceria–zirconiaAluminaHydrogenFischer–Tropsch dieselAutothermal reforming

0016-2361/$ - see front matter � 2013 Elsevier Ltd. Ahttp://dx.doi.org/10.1016/j.fuel.2013.02.010

⇑ Corresponding author. Tel.: +46 (0)8 790 9150.E-mail address: avga@kth.se (A.V. González).

a b s t r a c t

Autothermal reforming (ATR) of synthetic Fischer–Tropsch diesel has been carried out to evaluate thefuel reformer and the catalyst performance at realistic operating conditions. Hydrogen was producedvia ATR in a full-scale reformer (ID = 84 mm, L = 400 mm) at 650–750 �C. The two monolithic catalystswere sequentially located in the reformer and simultaneously tested. The catalysts were composed of1:1 wt% Rh:Pt as active metals; CeO2, MgO, Y2O3, and La2O3 were used as promoters. The first catalyticmonolith was supported on d-Al2O3 and the second on CeO2–ZrO2. Fresh samples were characterizedby N2-BET, XRD and H2-TPR analyses. Catalyst activity was evaluated at O2/C � 0.34–0.45 and H2O/C � 2–3. Results show an increased catalyst activity after the second monolithic catalyst due to the effectof steam reforming, water–gas shift reaction (WGS) and higher catalyst reducibility of RhxOy species onthe CeO2–ZrO2 mixed oxide as a result of the improved redox properties. Hydrogen concentrations of42 vol% and fuel conversion of 98% after the CeO2–ZrO2-supported catalyst was obtained at O2/C = 0.42and H2O/C = 2.5.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Use of alternative fuels and integration of fuel cell auxiliarypower units (FC-APUs) in vehicles, have been considered for reduc-tion of issues such as pollutant emissions, oil dependence andgreenhouse emissions in the transport sector. FC-APUs can provideon-board auxiliary electricity in vehicles, in particular duringstand-still mode when 60% of engine operating time is used ineffi-ciently [1]. Commercialization of this technology is been promotedby several public and private pilot projects but also, due to thestricter legislation on emission limits in the US and Europe, suchas the EURO VI directive and the anti-idling regulation, respec-tively [2]. Besides, use of FC-APUs on-board will overcome limita-tions such as storage and transport of hydrogen [3,4].

ll rights reserved.

Commercialization of the FC-APU is linked with the viability ofthe system fuel flexibility, to cover a wider range of applications. Inactual commercial fuel qualities i.e. European standard diesel (DIN590), great challenges are found for hydrogen production. This isdue to its sulfur and aromatic content. Recently, researchers havemade significant progress in overcoming these limitations bydecreasing diesel sulfur content and promoting usage of alterna-tive fuels. For instance use of rapeseed methyl ester (RME) andFischer–Tropsch (FT) diesel in the transport sector for reductionof pollutant emissions [5,6]. Specifically, FT diesel characteristics,i.e., high cetane number, low aromatic content, high paraffinic con-tent, and low sulfur content represent the advantage of facilitatinghydrocarbon oxidation and decreasing autothermal reforming(ATR) catalyst deactivation [1,7–9]. FT diesel properties are shownin Table 1.

The integrated FC-APU contains a fuel reformer, where fuel isconverted over a catalyst to obtain a H2-rich gas via ATR. Studies

Table 1Fischer–Tropsch diesel properties.

Properties Unit FT

Density kg/m3 800Viscosity at 40 �C mm2/s 2.5–4Flash point �C 75Boiling point �C 180–360Self-ignition temperature �C Above 200Lower heating value MJ/kg 47Sulfur content ppm 0C mass fraction wt% 85H mass fraction wt% 15

A.V. González et al. / Fuel 107 (2013) 162–169 163

on ATR of diesel have described it as the most viable process forhydrogen production in mobile applications such as trucks [10].The ATR combines endothermic steam reforming (SR), see Eq. (1),and partial oxidation (POX), Eq. (2); where ATR, Eq. (3), is the glo-bal reaction of the process [1,11,12]

CxHyOz þ ðx� zÞH2O! x� zþ y2

� �H2 þ xCO

DH > 0 kJ=mol ð1Þ

CxHyOz þðx� zÞ

2ðO2 þ 3:76N2Þ !

y2

H2 þ xCO

þ 3:76ðx� zÞ

2N2 DH < 0 kJ=mol ð2Þ

CxHyOz þ nðO2 þ 3:76N2Þ þ ðx� 2n� zÞH2O

! x� 2n� zþ y2

� �H2 þ xCOþ 3:76nN2 DH � 0 kJ=mol ð3Þ

In ideal ATR adiabatic operation heat generated by the exother-mic reaction is used to supply the endothermic steam reformingreaction. Which takes place together with reactions such as thewater–gas-shift reaction (WGS) in the catalytic zone [13]. How-ever, in realistic conditions, as is the objective of the present study,the system presents heat losses resulting in incomplete adiabaticoperation; thus excess of air is used [14] for practical operationof the FC-APU.

Despite of many advances in catalyst optimization and processperformance i.e., introduction of zone-coating catalysts, new fuelevaporation systems such as injectors and nozzles [14–17], issuessuch as fuel flexibility, use of alternative fuels for hydrogen pro-duction and the amount of noble metal loading required for com-plete fuel conversion still remain to be solved [1,7]. As acontinuation of previous research at our group, in this study is re-ported the ATR of FT diesel in a 5 kWe-scale reformer as well as theeffect of using two different catalyst compositions sequentially lo-cated and simultaneously tested.

Noble metal catalysts (Rh:Pt) were used as active metals forboth catalytic monoliths. Active metals and promoters were sup-ported on Al2O3 and CeO2–ZrO2, for the first and second monoliths,respectively. The monolith location was chosen depending on thereaction conditions to which the catalyst was subjected. That isoxidation reactions (exothermic) and steam reforming (endother-mic) [13]. Results show an increased catalyst activity after the sec-ond monolithic catalyst due to the effect of steam reforming, thewater gas shift reaction (WGS) and higher catalyst reducibility onCeO2–ZrO2 mixed oxide. Hydrogen concentrations of 42 vol% andfuel conversion of 98% after the CeO2–ZrO2-supported catalystwas obtained at O2/C = 0.42 and H2O/C = 2.5.

1.1. ATR catalyst selection

Catalyst optimization for autothermal reforming conditions hasbeen reported previously, describing systems with different sup-

ports, which demonstrate the superiority of Rh over Pt and Pd,for reforming of high sulfur-content fuels [9,18,19]. Similar resultswere also shown by Nilsson et al. [20] and Ferrandon et al. [21]. Asexample, Sigarov et al. [22] studied the performance of Co, Mn, Rh,BaO, La2O3/Al2O3 and SiO2 in diesel ATR in a reactor with two cylin-drical metallic monoliths. Maximum hydrogen yield of 18 mol/mol(fuel) was achieved at O2/C = 0.5, H2O/C = 1.5–1.7, GHSV = 2000–10.000 h�1 and at temperatures of 840–880 �C. The authorsshowed that the most active catalysts were the combination ofCo3O4/MnO4–BaO in the first zone of the reactor and Rh/MnO4–BaO in the second zone, with a product composition of H2 = 32%,CH4 = 1%, CO2 = 12%, CO = 11%, N2 = 44%.

Kaila and Krause [23] studied ATR of simulated gasoline anddiesel fuels using catalysts composed of Rh and Pt as active compo-nents in monometallic and bimetallic form with zirconia as sup-port. They found that Rh/ZrO2 catalyst showed high ATR activityand selectivity. Additionally, it was confirmed less coke formationwith Rh/ZrO2 than for a nickel catalyst in the same application. Inaddition, Kaila et al. [24] found that the bimetallic prepared cata-lyst presented better performance than the monometallic catalyst.

As reported by Kang et al. [25] the role of the support in areforming catalyst is not only determined by its mechanical orstructural characteristics, but it can also affect the catalyst activity,selectivity, thermal stability, carbon and sulfur tolerance with achemical role in formation of oxygen vacancies. This is possibledue to the oxygen storage capacity (OSC) of materials such asCeO2. The addition of oxygen, facilitates the reaction by regenerat-ing the oxide species of the surface belonging to the promoters andby oxidizing surface carbon species and carbon-containing prod-ucts as is reported by Srisiriwat et al. [26]

Ceria presents a fluorite structure in which it is possible to sub-stitute Ce4+ ions for Zr3+ or Gd3+ producing oxygen vacancies andmetal sites on the surface of the substrate. Other example isCeO2–ZrO2 or ZrO2 stabilized with yttrium [24,26]. Research onthe oxygen storage behavior of CeO2–ZrO2 mixed oxides has revealthat CeO2, ZrO2 and Mn are sulfur adsorbents and decrease catalystdeactivation by sulfur poisoning and carbon deposition [27]. Sulfuradsorption on CeO2–ZrO2 is accomplish through formation of a sta-ble sulfide at the temperature range of ATR [28]. This species canpartially being removed after vacuum treatment at high tempera-tures. However, under reducing conditions the CeO2–ZrO2 facili-tates desorption of sulfur species as H2S [29].

During ATR, the catalyst is subjected to oxidizing, partial oxida-tion (POX) and reducing conditions, steam reforming (SR). There-fore, CeO2–ZrO2 with its oxygen storage capacity (OSC) couldincrease oxidation reactions under reducing conditions (SR) whilepresenting superior sulfur resistance than d-Al2O3 at the sameconditions.

Alumina is the most common support material for environmen-tal catalysis due to its high surface area, low cost and thermal sta-bility. The Al2O3 is thermally stabilized by CeO2 or La2O3 to avoidsintering of the active metal [30].

2. Experimental

2.1. Catalyst preparation

Catalyst samples were prepared by the incipient wetnessimpregnation procedure. Powders of d-Al2O3 (prepared by calcin-ing c-Al2O3 supplied by Sasol GmbH Germany with a specific sur-face area of 105 m2/g) and CeO2–ZrO2 (supplied by MEL Chemicalswith a specific surface area of 241 m2/g) were used as supports.The d-Al2O3 was impregnated with an aqueous solution ofCe(NO3)3�6H2O (Alfa Aesar, 99.99%),and La(NO3)3�6H2O (Alfa Aesar,99.99%), and finally impregnated with an aqueous solution ofRh(NO3)3 (Sigma Aldrich) and (NH3)4Pt(NO3)2 (Alfa Aesar), dried

Table 2Catalyst properties.

Samples Designation Surface area(m2/g)

Pore volume(cm3/g)

d-Al2O3 Al 105 0.92Rh1.0Pt1.0Ce10La10/d-Al2O3

a CAT 1 82 0.5CeO2–ZrO2 CZ 241 0.3Rh1.0Pt1.0Mg4.0Y5.0/CeO2–ZrO2

a CAT 2 57 0.21

a The numbers denote wt% of individual species.

Table 3Operating conditions of FT diesel autothermal reforming with a fuel flow of 19 g/min.

O2:C H2O:C GHSV (103 h�1) k

0.34 2.5 10.1 0.230.38 2.5 10.5 0.250.42 2.5 10.8 0.280.45 2.5 11.2 0.30

164 A.V. González et al. / Fuel 107 (2013) 162–169

at 110 �C after each impregnation The CeO2–ZrO2 was impregnatedwith an aqueous solution of Mg(NO3)2�6H2O (Alfa Aesar, 98%),Y(NO3)3�6H2O (Sigma–Aldrich, 99.9%), dried at 110 �C and finallyimpregnated with an aqueous solution of Rh(NO3)3 (Sigma Aldrich)and (NH3)4Pt(NO3)2 (Alfa Aesar). The two impregnated powderswere calcined at 800 �C for 3 h, since the reaction temperature inthe gas phase will be around 650–800 �C, the prepared powderscomposition are shown in Table 2. Ethanol slurries of the catalystpowders were ball-milled for 24 h. Ceramic monoliths were cutout from Corning synthetic cordierite (2MgO�2Al2O3�5SiO2) blockswith a cell density of 400 cpsi, with dimensions OD = 80 mm andl = 70.6 mm. The cordierite monoliths were dip-coated into theethanol slurry followed by drying at 110 �C for 30 min. This proce-dure was repeated until an intended loading of 20 wt% of the totalweight (monolith and washcoat) was reached. The final coatedsamples were calcined at 800 �C for 3 h.

2.2. Catalyst characterization

The specific surface area and pore size distribution of the cata-lysts were measured by the Brunauer, Emmet and Teller (BET) andBJH methods, respectively, using nitrogen adsorption at liquidnitrogen temperature by a Micromeritics ASAP 2010 instrument.Samples were degassed at 250 �C for 3 h prior to analysis.

The morphology of the samples was determined by X-ray dif-fraction (XRD), on a Siemens Diffractometer D5000 using Ni fil-tered Cu Ka radiation and scanning 2h from 10� to 90� in thescan mode (0.02�, 1 s). Crystal phases were determined by compar-ing sample diffractograms with powder diffraction database files(ICDD/JPCDS).

Fig. 1. Thermocouples and reformer configur

Temperature programmed reduction of catalyst samples wasconducted applying a pre-treatment of the sample in flowing Hewith 5 vol% of O2 at a ramp temperature of 10 �C/min up to800 �C, maintaining the sample constant at that temperature for30 min. Thereafter the sample was cooled to room temperaturein order to initiate reduction with 5 vol% of H2 in Ar with a ramptemperature of 10 �C/min up to 800 �C. The temperature was keptat 800 �C for 30 min. The analysis was carried out using a Microm-eritics Autochem 2019; H2 consumption was measured by a ther-mal conductivity detector.

2.3. Full-scale FT-diesel reforming

In this study, autothermal reforming of synthetic Fischer–Trop-sch diesel has been carried out in a full-scale reformer (ID = 84 mm,L = 400 mm); over bimetallic (1 wt% Rh, 1 wt% Pt) catalysts sup-ported on Al2O3 and CeO2–ZrO2 at 650–750 �C. Both catalysts weresequentially located in the reformer and simultaneously tested.Catalyst activity was evaluated at O2/C � 0.34–0.45 and H2O/C � 2–3 as is shown in Table 3.

Fischer–Tropsch diesel (Ecopar AB), see Table 1, was used asfeedstock for autothermal reforming. The fuel flow was kept con-stant at 19 g/min, resulting in a thermal power input of 14 kWth.In previous studies has been reported the reactor set-up andinstrumentation used to performed the experiments; detail infor-mation can be found elsewhere [31]. The fuel was injected througha stainless steel spray nozzle (0.58 mm orifice diameter, Mistjet�,STEINEN) and blended with a 300 �C superheated air–steam mix-ture. An alumina foam disc (RQ-3085, Selee Corp.) and two reform-ing monolithic catalysts designated as CAT 1 and CAT 2, seeTable 2, were placed sequentially in a stainless steel tubular

ation. Adapted from Karatzas et al. [31].

A.V. González et al. / Fuel 107 (2013) 162–169 165

reformer with an inner diameter of 84 mm and a length of400 mm. The reformer has two sampling points one at the end ofCAT 1 and another at the end of CAT 2. Temperature measurementswere performed with 14 thermocouples (Pentronic, type K) locatedinside the reactor, two thermocouples inside each monolith wereplaced at reactor length 180 and 290 mm, see Fig. 1.

Samples were taken after each catalytic monolith; the concen-trations of dry reformate were analyzed using a gas chromatograph(GC) Varian CP-3800 equipped with a thermal conductivity detec-tor (TCD) and a flame ionization detector (FID), for detection ofCH4, CO, CO2, H2, N2 and O2. Concentrations of CO2, CO, CH4,C2H4, C2H6, C3H8, C3H6, C6H6, H2O, SO2 and diesel were continu-ously analyzed using a Fourier transform infrared spectrometer(FTIR), a MKS Multigas™ 2030 HS, equipped with a high speed0.5 cm�1 spectrometer. CO and CO2 contents for both methodswere compared to ensure reliability. The FTIR diesel response fac-tor was configured, designed and provided by the company MKSInstruments. A total hydrocarbon analysis from a wide range oflong hydrocarbons (>C12) typically found in diesel was introducedin a single FTIR method file. In this study the FTIR response factorwas assumed to apply for FT diesel. Hydrogen produced was mea-sured by a hydrogen gas analyzer (H-SENSE), equipped with anelectron pulse ionization mass spectrometer (EIMS).

3. Results and discussion

3.1. Catalyst characterization

The catalyst characterization was studied in detail previouslywhere catalyst activity was evaluated in bench scale experimentsto identify the most promising catalyst for ATR [13,28]. In this sec-tion will be given a summary of the findings related to the catalystcomposition used in this work. In Table 2 are presented the BETsurface area of the supports, d-Al2O3 and CeO2–ZrO2 mixed oxides.It can be seen that the surface area decreases as promoters and ac-tive materials are added into each support crystalline structureafter calcination at 800 �C for 3 h. The d-Al2O3-supported catalystpresented a characteristic small loss in surface area after calcina-tion, 82 m2/g, related to the initial unpromoted d-Al2O3 samplewith a surface area of 105 m2/g. Regarding the CeO2–ZrO2, the pro-moted sample presents a decrease by 75% with respect to the ini-tial CeO2–ZrO2 surface area of 241 m2/g. This results are inagreement with other studies [32].

XRD diffractograms for CeO2–ZrO2 supported catalysts are pre-sented in Fig. 2. The main peaks correspond to ZrO2 tetragonal

Fig. 2. CeO2–ZrO2 supported catalyst diffractograms. Crystalline phases, CeO2 cubic(a), tetragonal ZrO2 (b).

phases. However, according to Varez et al. [33] cubic and tetrago-nal phases can be present at 2h = 28� and 2h = 30�, respectively,which can be correlated with the obtained results. The intensityof the tetragonal phase decreases by sequential addition of activemetals Rh and Pt. Additionally, a second tetragonal phase (T0)was present when only Rh and Pt were added and broader peakswere obtained. This means that even though Rh, Pt and RhxPt1�x al-loy peaks were not possible to detect due to overlapping withthose corresponding to ZrO2, they were either well dispersed onthe support surface or present in smaller particle size. Higherintensity of tetragonal (T) peaks were seen when MgO and Y2O3

were added, due to the high stabilization of the ZrO2 tetragonalphase, by decreasing the energy strain with the addition of tetrava-lent Ce4+ ions, which is in good agreement with other reports [34].According to Kakuta et al. [35] MgO is present in the solid–solidsolution in smaller particle sizes than in unpromoted CeO2–ZrO2

and is a barrier material, decreasing the OSC loss at hightemperatures.

Fig. 3 shows XRD patterns for the fresh alumina-supported cat-alysts. Characteristic peaks for d-Al2O3 are present. No Rh phasecould be detected by XRD, due to the low metal loading (1 wt%)and probably due to high dispersion degree.

Metallic Pt was detected at 2h = 40�, 45�, 80� and 85�. CeO2 fluo-rite structure was present at 2h = 28� and 55�. La2O3 could not bedetected by XRD. It is observed that by adding promoters and ac-tive metals onto the support the peak intensity decreases. Theseresults are in agreement with other studies [31,36].

Chemisorption analyses were performed on monometallic Rhcatalyst in previous studies [37]. Resulting in a dispersion degreeof �54% for Rh/CeO2–ZrO2 followed by �47% for the Rh/d-Al2O3,which are related to the XRD results previously described.

Temperature-programmed reduction (TPR) experiments showthat CAT 2 has higher H2 uptake and therefore the active particlesare more easily reduced. Fig. 4 shows TPR results where H2 con-sumption is detected at 180 �C and 240 �C for CAT 2, where RhxOy

species are reduced. PtOx species could not be detected maybe dueto its instability at high temperatures [23]. A second peak is pres-ent at around 400–500 �C attributed to H2 spillover effect on thesupport [32]. For CAT 1 supported on Al2O3, RhxOy species are re-duced at 290 �C and H2 spillover is also present around 400–500 �C [36].

From the catalyst characterization we identified higher reduc-ibility of RhxOy species on the surface of CeO2–ZrO2, compared withd-Al2O3, as is revealed by the TPR experiments. The increased avail-able RhxOy species has been explained as a result of Rh interactions

Fig. 3. XRD patterns of CAT 1 and precursors. Peaks representing metallic Pt (a) andthe fluorite structure of CeO2 (b).

Fig. 4. TPR profiles of the fresh powders of alumina-and CeO2–ZrO2-supportedsamples. Fig. 5. Temperature along the reformer during FT diesel ATR at O2/C = 0.42 and

H2O/C = 2.5.

166 A.V. González et al. / Fuel 107 (2013) 162–169

with more available oxygen species in the matrix of CeO2–ZrO2.The oxygen storage capacity, responsible for oxygen availability,is stabilized with tetragonal phases of ZrO2 [38], which were veri-fied with the XRD analysis. These results are consistent with earlierH2-TPR and TEM analyses [37].

3.2. Fischer–Tropsch diesel ATR experiments

Two monolithic catalysts, supported on d-Al2O3 and CeO2–ZrO2,were placed sequentially in the reactor and tested simultaneouslyin each experiment. Results are presented as volumetric concentra-tions of hydrogen (H2), carbon dioxide (CO2) and carbon monoxide(CO). Catalyst activity was evaluated by calculation of fuel conver-sion (Xfuel), defined as the amount of diesel fed with respect to theamount of diesel detected by the FTIR in the wet reformate and di-vided by the amount of diesel fed. The COx/(CO2 + CO) product ratiowas used to evaluate CO2 selectivity relative to CO. Hydrogenselectivity (H2 Sel) is defined as the moles of hydrogen in the prod-uct gas obtained per mole of fuel divided by the theoretical maxi-mum amount of hydrogen at the specific condition (assuming thatall carbon reacts to CO2). The reformer efficiency (gRef) was definedas shown in the following equation:

gRef ¼FH2 � LHVH2

FFuel � LVHFuel� 100 ð4Þ

Fig. 6. Analysis for O2/C = 0.42 and H2O/C = 2.5. Fuel flow of 19 g/min.

3.2.1. Steady state operationA temperature profile for steady state operation is shown in

Fig. 5 optimal operation conditions at O2/C = 0.42 and H2O/C = 2.5(k = 0.28) were obtained with GHSV of 10.8 � 103 h�1. The optimaloperating parameters were found by step wise increments of O2/Cand H2O/C ratios. Every point was kept constant until stable condi-tions were reached. In the mixing zone, upstream the first mono-lithic catalyst, homogeneous gas phase reactions take place,causing an increment in temperature from 430 �C to 580 �C, asthe O2/C ratio was increased during successive increments, thishas been reported to occur in the mixing zone by Creaser et al.[39] and similar conclusions were reported by Lindström et al.[12]. The increase in temperature may also be attributable to bettermixing of the reactant mixture after the foam and to back radiationfrom the initial monolith. The homogeneous gas mixture continuesalong the reformer axial direction, to the first catalytic monolith.

Coated monoliths with a washcoat composition of Rh1.0Pt1.0-

Ce10La10/Al2O3 and Rh1.0Pt1.0Mg4.0Y5.0/CeO2–ZrO2 are designatedas CAT 1 and CAT 2, respectively. At the inlet of the CAT 1 the exo-thermic partial oxidation takes place, hence an increase in temper-

ature up to 655 �C is observed, see Eq. (1). A decrease intemperature to 640 �C at the outlet of the CAT 1 is shown due toendothermic reactions taking place. Sequentially, the inlet flow atCAT 2 presents a slight temperature increase of 5 �C, which maybe attributable to the oxidation reactions taking place with theoxygen stored in the support matrix. However, mainly endother-mic steam reforming, see Eq. (2) and the water–gas shift reaction(WGS) take place in CAT 2.

A comparison between CAT 1 and CAT 2 is presented in Fig. 6,for the operating conditions O2/C = 0.42 and H2O/C = 2.5, wherefuel conversion reached 95% and 98%, respectively. A high H2/COratio is observed at the outlet of CAT 2, related to the H2 selectivitybeing slightly higher in CAT 2 and the low CO selectivity, sincemost of the hydrogen production takes place under steam reform-ing conditions.

During partial oxidation conditions, taking place at CAT 1, theoxygen is used in the favored oxidation reactions but also in sidereactions such as hydrogen oxidation, Eq. (5), and carbon monox-ide oxidation, Eq. (6) [1], which leads to a lower hydrogenproduction.

H2 þ12

O2 ! H2O ð5Þ

COþ 12

O2 ! CO2 ð6Þ

Fig. 7. Product gas composition after CAT 1 (a) and CAT 2 (b) for H2O/C = 2.5.

Fig. 8. Analysis of the effect of O2/C ratio on catalyst activity. Compositions measured at the outlet of CAT 1 and CAT 2, H2O/C was kept constant at 2.5.

A.V. González et al. / Fuel 107 (2013) 162–169 167

An interesting trend is the higher CO2 selectivity for CAT 2 com-pared to CAT 1, due to enhancement of the WGS reaction,

COþH2O$ H2 þ CO2 ð7Þ

The CeO2–ZrO2 OSC effect is also involved in the higher hydro-gen concentration. Oxygen is released to the system from theCeO2–ZrO2 crystalline structure. As can be deduced from the TPRexperiments (see Section 3.1), a significant reducibility improve-ment for RhPt oxide species was observed. Results are well in linewith previous results from bench scale experiments on metallicmonoliths and cordierite monoliths at the same operating condi-tions [37].

3.2.2. Effect of the oxygen–carbon ratioEquilibrium gas compositions as function of O2/C ratio for FT

diesel reforming at H2O/C = 2.5 are presented in Fig. 7, were reac-tion temperature is also displayed. Analyses were taken fromCAT 1 and CAT 2 outlets. Results for increased O2/C show, for bothcatalysts, a decrease in hydrogen production, as more O2 is intro-duced to the system. Since Eqs. (1), (2), and (6) are favored, hydro-gen is consumed to produce more H2O and CO2, which arecharacteristic products of stoichiometric conditions [1]. The hydro-gen concentration increased in the range of �2 vol%, from CAT 1,see Fig. 7a, to CAT 2, Fig. 7b. Maximum hydrogen formation, 42%,

was obtained at 630 �C at the outlet of CAT 2 in which mainlythe steam reforming process is believed to take place. CO concen-tration decreases from CAT 1 to CAT 2, meaning that more hydro-gen is produced through the water gas shift reaction, Eq. (5).Similar trends have been reported by Karatzas et al. [31] andCheekatamarla and Thomson [14].

According to the results described above, the location of themonoliths in sequential order has considerably improved thehydrogen production and the reformer efficiency as shown inFig. 6 where an increase of 10% is shown.

In Fig. 8 both catalysts (CAT 1 and CAT 2) are compared in termsof fuel conversion and selectivity to hydrogen and carbon dioxide.As the catalysts were placed sequentially inside the reformer, thefuel conversion increased as the gases flowed through the catalyticmonoliths. Fuel conversion increases as the O2/C ratio increases forboth catalysts. A decreasing trend in H2 selectivity is shown for CAT1, while for CAT 2 an increasing tendency is observed. The lowhydrogen selectivity present in CAT 1 may be attributable to oxida-tion reactions being favored by the increase of oxygen concentra-tion in the system, increasing the CO formation, following Eqs.(1) and (2). This can also be understood from the decrease in CO2

selectivity. The higher hydrogen selectivity in CAT 2 is the resultof the steam reforming reaction, Eq. (2), and the effect of thewater–gas shift reaction taking place, Eq. (5).

Fig. 9. ATR analysis results at the outlet of CAT 2 for FT diesel reforming with O2/C constant at 0.42, of a constant fuel flow of 19 g/min.

Fig. 10. Ethylene concentration at the outlet of CAT 2. Variation of H2O/C ratio from 2 to 3.

168 A.V. González et al. / Fuel 107 (2013) 162–169

3.2.3. Effect on the steam-to-carbon ratioThe effect of the H2O/C ratio was studied keeping the O2/C ratio

fixed at 0.42. Fig. 9a) shows increments of H2O/C from 2 to 3. As isobserved, the fuel conversion increases as H2O/C increases. It canbe correlated with what is seen in Fig. 10 where ethylene forma-tion decreases as O2/C and H2O/C increase. Reduction of CO is alsoseen as the H2O/C increases due to the water–gas shift reaction,which further increases the CO2 formation, as seen in Fig. 9b).

In Fig. 10 ethylene concentration is shown to decrease with thevariation of the O2/C ratio for three different H2O/C ratios. Duringthe inspection of the catalyst after each experiment no carbon for-mation was observed during the FT-diesel experiments.

4. Conclusion

From the catalyst characterization we identified higher reduc-ibility if RhxOy species on the surface of the CeO2–ZrO2, comparedwith d-Al2O3, as is revealed by the TPR experiments. The increasedavailable RhxOy species have been explained as a result of Rh inter-actions with more available oxygenated species in the crystallinestructure of CeO2–ZrO2. The oxygen storage capacity, responsiblefor oxygen availability, is stabilized with tetragonal phases ofZrO2, which were verified by the XRD analysis. However, in theXRD analysis no presence of Rh and Pt were detected in the catalystbulk, due to the low concentration of the active metals.

The influence of operating parameters, O2/C and H2O/C ratios,showed a reduction in the amount of hydrogen produced as the

O2/C increased. This is due to the oxidation reactions being favoredas more oxygen is present in the system. With regard to the steam-to-carbon ratio, an increment in fuel conversion and hydrogenconcentration is required. The effect of operating conditions onethylene concentrations was evaluated. A decreasing ethyleneconcentration as the oxygen-to-carbon and steam-to-carbon ratiosincreased was found.

Catalyst activity was analyzed in terms of fuel conversion, CO2

and H2 selectivities. ATR full-scale experiments show a maximumhydrogen concentration of 42 vol% and a fuel conversion of 98%after CAT 2 at O2/C ratio of 0.42 and H2O/C ratio of 2.5. Reformerefficiency was found to be higher after CAT 2 as more hydrogenwas produced per mole of fuel introduced.

Thus, enhancement on catalyst activity is related to the im-proved redox properties on the catalyst surface, and promotionof the WGS reaction by the CeO2–ZrO2 support. Furthermore,the H/C ratio of Fischer–Tropsch diesel, increased the selectiv-ity towards H2 formation in the reformate gas. Taking into ac-count the complete consumption of oxygen at the entrance ofthe first monolith, beneficial effects are observed for the Rhand Pt interactions with stored oxygen in the crystalline struc-ture of the downstream CeO2–ZrO2-supported catalyst. For in-stance, process efficiency and the hydrogen generation wereimproved. This suggests that not only the metallic phases ofthe active materials represent the most active phases for die-sel reforming, but also the RhxOy species present catalystactivity.

A.V. González et al. / Fuel 107 (2013) 162–169 169

Acknowledgements

The Swedish Energy Agency is acknowledged for financial sup-port. Thanks also to Corning Inc. for supplying cordierite sub-strates, to Sasol Germany GmbH for providing the aluminasupport and MEL chemicals for providing the CeO2–ZrO2.

References

[1] Shekhawat D, Berry DA, Haynes DJ, Spivey JJ. Fuel constituent effects on fuelreforming properties for fuel cell applications. Fuel 2009;88:817–25.

[2] Contestabile M. Analysis of the market for diesel PEM fuel cell auxiliary powerunits onboard long-haul trucks and of its implications for the large-scaleadoption of PEM FCs. Energy Policy 2010;38:5320–34.

[3] Kang I, Bae J. Autothermal reforming study of diesel for fuel cell application. JPower Sources 2006;159:1283–90.

[4] Specchia S, Specchia V. Modeling study on the performance of an integratedAPU fed with hydrocarbon fuels. Ind Eng Chem Res 2010;49:6803–9.

[5] Pastore A, Mastorakos E. Syngas production from liquid fuels in a non-catalyticporous burner. Fuel 2011;90:64–76.

[6] Munack A, Krahl J. Generation and utilization of bio-fuels – National andInternational trends. Clean – Soil, Air, Water 2007;35:413–6.

[7] Kang IY, Carstensen HH, Dean AM. Impact of gas-phase reactions in the mixingregion upstream of a diesel fuel autothermal reformer. J Power Sources2011;196:2020–6.

[8] Gill SS, Tsolakis A, Dearn KD, Rodríguez-Fernández J. Combustioncharacteristics and emissions of Fischer–Tropsch diesel fuels in IC engines.Progr Energy Combust Sci 2011;37:503–23.

[9] Karatzas X. Rhodium diesel-reforming catalysts for fuel cell applications.Chemical Engineering and Technology. Stockholm: KTH Royal Institute ofTechnology; 2011.

[10] Krumpelt M, Krause TR, Kopasz JP. ASME, Fuel processing for mobile fuel cellsystems. In: Shah RK, Kanslikar SG, editors. Fuel cell science, engineering andtechnology, 2003, pp. 55–59.

[11] Ghenciu AF. Review of fuel processing catalysts for hydrogen production inPEM fuel cell systems. Curr Opin Solid State Mater Sci 2002;6:389–99.

[12] Lindström B, Karlsson JAJ, Ekdunge P, De Verdier L, Häggendal B, Dawody J,et al. Diesel fuel reformer for automotive fuel cell applications. Int J HydrogenEnergy 2009;34:3367–81.

[13] Bitsch-Larsen A, Horn R, Schmidt LD. Catalytic partial oxidation of methane onrhodium and platinum: Spatial profiles at elevated pressure. Appl Catal, A2008;348:165–72.

[14] Cheekatamarla PK, Thomson WJ. Poisoning effect of thiophene on the catalyticactivity of molybdenum carbide during tri-methyl pentane reforming forhydrogen generation. Appl Catal, A 2005;287:176–82.

[15] Karatzas X, Dawody J, Grant A, Svensson EE, Pettersson LJ. Zone-coated Rh-based monolithic catalyst for autothermal reforming of diesel. Appl Catal, B2011;101:226–38.

[16] Kang I, Bae J, Yoon S, Yoo Y. Performance improvement of diesel autothermalreformer by applying ultrasonic injector for effective fuel delivery. J PowerSources 2007;172:845–52.

[17] Dias JAC, Assaf JM. Autothermal reforming of methane over Ni/c-Al2O3

catalysts: the enhancement effect of small quantities of noble metals. JPower Sources 2004;130:106–10.

[18] Au C-T, Ng C-F, Liao M-S. Methane dissociation and syngas formation on Ru,Os, Rh, Ir, Pd, Pt, Cu, Ag, and Au: a theoretical study. J Catal 1999;185:12–22.

[19] Vaidya PD, Rodrigues AE. Insight into steam reforming of ethanol to producehydrogen for fuel cells. Chem Eng J 2006;117:39–49.

[20] Nilsson M, Karatzas X, Lindström B, Pettersson LJ. Assessing the adaptabilityto varying fuel supply of an autothermal reformer. Chem Eng J 2008;142:309–17.

[21] Ferrandon M, Mawdsley J, Krause T. Effect of temperature, steam-to-carbonratio, and alkali metal additives on improving the sulfur tolerance of a Rh/La-Al2O3 catalyst reforming gasoline for fuel cell applications. Appl Catal, A2008;342:69–77.

[22] Shigarov AB, Kireenkov VV, Kuzmin VA, Kuzin NA, Kirillov VA. Autothermalreforming of diesel fuel in a structured porous metal catalyst: both kineticallyand transport controlled reaction. Catal Today 2009;144:341–9.

[23] Kaila RK, Krause AOI. Autothermal reforming of simulated gasoline and dieselfuels. Int J Hydrogen Energy 2006;31:1934–41.

[24] Kaila RK, Gutierrez A, Krause AOI. Autothermal reforming of simulated andcommercial diesel: The performance of zirconia-supported RhPt catalyst in thepresence of sulfur. Appl Catal, B 2008;84:324–31.

[25] Kang I, Yoon S, Bae G, Kim JH, Bae J, Lee D, et al. The micro-reactor testing ofcatalysts and fuel delivery apparatuses for diesel autothermal reforming. CatalToday 2008;136:249–57.

[26] Srisiriwat N, Therdthianwong S, Therdthianwong A. Oxidative steamreforming of ethanol over Ni/Al2O3 catalysts promoted by CeO2, ZrO2 andCeO2–ZrO2. Int J Hydrogen Energy 2009;34:2224–34.

[27] Ahmed S, Krumpelt M. Hydrogen from hydrocarbon fuels for fuel cells. Int JHydrogen Energy 2001;26:291–301.

[28] Hennings U, Reimert R. Noble metal catalysts supported on gadolinium dopedceria used for natural gas reforming in fuel cell applications. Appl Catal, B2007;70:498–508.

[29] Boaro M, De Leitenburg C, Dolcetti G, Trovarelli A, Graziani M. Oxygen storagebehavior of ceria-zirconia-based catalysts in the presence of SO2. Top Catal2001;16–17:299–306.

[30] Chorkendorff I, Niemantsverdriet JW. Concepts of modern catalysis andkinetics. 2nd ed. Weinheim: Wiley-VCH; 2007.

[31] Karatzas X, Nilsson M, Dawody J, Lindström B, Pettersson LJ. Characterizationand optimization of an autothermal diesel and jet fuel reformer for 5 kWemobile fuel cell applications. Chem Eng J 2010;156:366–79.

[32] Boaro M, Vicario M, Llorca J, de Leitenburg C, Dolcetti G, Trovarelli A. Acomparative study of water gas shift reaction over gold and platinumsupported on ZrO2 and CeO2–ZrO2. Appl Catal, B 2009;88:272–82.

[33] Varez A, Garcia-Gonzalez E, Jolly J, Sanz J. Structural characterization ofCe1�xZrxO2 (0 6 x 6 1) samples prepared at 1650 �C by solid state reaction. Acombined TEM and XRD study. J Eur Ceram Soc 2007;27:3677–82.

[34] Mastelaro VR, Briois V, de Souza DPF, Silva CL. Structural studies of a ZrO2–CeO2 doped system. J Eur Ceram Soc 2003;23:273–82.

[35] Kakuta N, Kudo Y, Eguchi T, Ohkita H, Mizushima T, Yamomoto T, Yasuda M.Preparation of CeO2/ZrO2 modified by MgO: roles of MgO for upgrading redoxproperty, vol. 162; 2006 p. 777–84.

[36] Eriksson S. Development of catalysts for natural gas–fired gas turbinecombustors. Stockholm: KTH; 2006. pp. 67.

[37] Karatzas X, Jansson K, González AV, Dawody J, Pettersson LJ. Autothermalreforming of low-sulfur diesel over bimetallic RhPt supported on Al2O3, CeO2–ZrO2, SiO2 and TiO2. Appl Catal, B 2011;106:476–87.

[38] Fornasiero P, Dimonte R, Rao GR, Kaspar J, Meriani S, Trovarelli A, et al. Rh-Loaded CeO2–ZrO2 solid-solutions as highly efficient oxygen exchangers:dependence of the reduction behavior and the oxygen storage capacity on thestructural-properties. J Catal 1995;151:168–77.

[39] Creaser D, Karatzas X, Lundberg B, Pettersson LJ, Dawody J. Modeling study of5 kWe-scale autothermal diesel fuel reformer. Appl Catal, A 2011;404:129–40.