The University of Manchester Research

Upgrading of heavy oil by dispersed biogenic magnetitecatalystsDOI:10.1016/j.fuel.2016.08.015

Document VersionAccepted author manuscript

Link to publication record in Manchester Research Explorer

Citation for published version (APA):Brown, A., Hart, A., Coker, V., Lloyd, J., & Wood, J. (2016). Upgrading of heavy oil by dispersed biogenicmagnetite catalysts. Fuel, 185, 442-448. https://doi.org/10.1016/j.fuel.2016.08.015

Published in:Fuel

Citing this paperPlease note that where the full-text provided on Manchester Research Explorer is the Author Accepted Manuscriptor Proof version this may differ from the final Published version. If citing, it is advised that you check and use thepublisher's definitive version.

General rightsCopyright and moral rights for the publications made accessible in the Research Explorer are retained by theauthors and/or other copyright owners and it is a condition of accessing publications that users recognise andabide by the legal requirements associated with these rights.

Takedown policyIf you believe that this document breaches copyright please refer to the University of Manchester’s TakedownProcedures [http://man.ac.uk/04Y6Bo] or contact uml.scholarlycommunications@manchester.ac.uk providingrelevant details, so we can investigate your claim.

Download date:08. Apr. 2020

Upgradingofheavyoilbydispersedbiogenicmagnetitecatalysts1

AshleyR.Browna1,AbarasiHartb,VictoriaS.Cokera,JonathanR.Lloyda*andJosephWoodb*2

aWilliamsonResearchCentreforMolecularEnvironmentalScience,SchoolofEarth,3

AtmosphericandEnvironmentalSciences,UniversityofManchester,Manchester,M139PL,4

UnitedKingdom.5

bSchoolofChemicalEngineering,UniversityofBirmingham,Edgbaston,Birmingham,B156

2TT,UnitedKingdom.7

1Presentaddress:Eawag,SwissFederalInstituteofAquaticScienceandTechnology,86008

Duebendorf,Switzerland.Email:ashley.brown@eawag.ch9

*Correspondingauthors:ashley.r.brown@manchester.ac.uk;jon.lloyd@manchester.ac.uk;10

j.wood@bham.ac.uk11

12

Keywords13

THAI,CAPRI,magnetite,palladium,catalyst,heavyoil14

15

Abstract16

Insitucatalyticupgradingofheavyoilofferssignificantcostsavingsandovercomeslogistical17

challenges associated with the high viscosity, low API gravity and highmolecular weight18

fractionsofunconventionalhydrocarbonresources.TheTHAI-CAPRIprocess(toe-to-heelair19

injection–catalyticupgradingprocess insitu)offersonesuchroutetoupgradingthrough20

theuseofhighsurfaceareatransitionmetalcrackingcatalystssurroundingtheproduction21

well.Here,wedescribethecatalyticupgradingofheavyoil inastirredbatchreactorbya22

biogenic nanoscalemagnetite (BnM; Fe3O4). A 97.8% decrease in viscosity relative to the23

feedoilwasachievedandcokingwas lowercomparedtothermalcrackingalone(6.9wt%24

versus10.2wt%). Theactivityof this catalystwas furtherenhancedbya simpleone-step25

additionofsurfaceassociatedPdtoachieveloadingsof4.3,7.1and9.5wt%Pd.Thisledto26

significant decreases in viscosity of up to 99.4% for BnM loaded with 9.5 wt% Pd. An27

incrementof7.8° inAPIgravitywithrespect tothefeedoilwasachievedfor9.5wt%Pd-28

BnM, compared with thermal cracking alone (5.3°). Whilst this level of upgrading was29

comparabletocommerciallyavailableandpreviouslytestedcatalysts,significantdecreases30

in thecokecontent (3wt% for9.5wt%Pd-BnMversus10wt% for thermal cracking)and31

associated increases in liquid content (~90wt% for 9.5wt% Pd-BnM versus ~79wt% for32

thermal cracking) demonstrate the potential for the use of Pd-augmented biogenic33

magnetiteasacatalystintheTHAICAPRIprocess.34

35

1.Introduction36

Continuingdepletionofglobaloilreserveshasledtounconventionaloilresources,suchas37

oil sands, heavy oil and bitumen becomingly increasingly attractive for exploitation.38

However,heavyoilandbitumenhavethedisadvantageofhighviscosity,lowAPI(American39

PetroleumInstitute)gravityandhighmolecularweightfractions.Thispresentsachallenge,40

as the transportation of these materials is both costly and energetically expensive, with41

heating of the pipeline or solvent addition often required to improve flow rates [1].42

Furthermore,conventionaloilrefineriesrequireheavyoilandbitumentobefirstupgraded43

toalightcrudeoilbeforebeingdistilled[2].Again,thiscarriesassociatedcostsandhence44

previousexploitationoftheseresourceshasbeenlimited.45

This demand for enhanced oil recovery (EOR) has led to the development of in situ46

upgradingtechnologies[3].Thesepresentthekeyadvantagethatupgradingoccursdown-47

hole, thus, enhancing both production and oil quality whilst precluding the need for a48

secondarysurfaceupgradingfacility.Inaddition,severalenvironmentalbenefitsareoffered49

as unwanted contaminants or by-products are retained in the reservoir, including heavy50

metals,sulphurandcarbondioxide[4].51

TheTHAI-CAPRIprocess(toe-to-heelairinjection;catalyticupgradingprocessinsitu)isone52

suchtechniquethatcombinesthermallyenhancedoilrecoverywithinsitucatalysisofthe53

heavy oil being mobilised [5–7]. First, in situ combustion is advanced by the continuous54

injection of air through a vertical well (THAI), which allows the combustion front to be55

sustained. Second, the CAPRI add-on involves the addition of a catalyst to the horizontal56

productionwell,whichpromotesupgradingofthemobilisedoil,e.g.viahydrocrackingand57

hydrodesulfurization.However,challengesarisefromtheuseoffixedbedcatalystsdueto58

severedepositionofcoke,asphaltenesandmetalsleadingtocatalystdeactivation[8,9].59

Alternatively, dispersed submicron and nano-particulate catalystsmay circumvent this by60

presentingalargersurfaceareadistributedthroughoutthesiteofcombustion,andhence,61

overcoming diffusion limitations and pore plugging associated with fixed bed catalysts,62

which may obstruct production lines [10–12]. Previous work has shown that dispersed63

micrometersizedCo-Mo/Al2O3showedsuperiorupgradingcomparedtoitsmillimetresized64

pelleted counterpart [10]. Significant upgrading has also been described for dispersed65

nanoparticulatehematiteandseveralun-supportedtransitionmetalnanoparticles,namely66

MoS2,NiOandFe2O3 [11,13].Although theextentofupgradingwasmodest compared to67

thermal cracking, these dispersedmaterials significantly suppressed coke formation,with68

theremainingcokehavingasponge-typecharacterwhichmayfinduseasanindustrialfuel69

compared with typical coke materials generated from thermal cracking. These studies70

highlightthepotentialforthecombinationofdispersednano-sizedcatalystswiththeinsitu71

upgrading offered by the CAPRI process, as an alternative to previously used pelleted72

hydroprocessingcatalysts.73

Noblemetals,suchaspalladiumandplatinumalsoexhibitcatalyticupgradingofoilthrough74

thepromotionofhydrogenationandhydrogenolysis reactionswhichaddhydrogen to the75

oilmolecule (hydrocracking) [14–16]. Supported-Pd catalysts are particularly effective, as76

the surface area of the palladium is maximised by a nanoparticulate support structure.77

Typically, this has been done with carbon, silica, alumina and recently bacterial biomass78

[14,16,17]. Indeed, the use of bacteria in nanocatalyst production represents a more79

environmentallybenignsynthesisroute.80

Further work using bacteria has demonstrated the synthesis of magnetic iron oxide81

nanoparticles using Fe(III)-reducing bacteria such as Geobacter sulfurreducens [18,19].82

Biogenic nanoscale magnetite (bionanomagnetite; BnM) has a large surface area, high83

chemicalreactivityandexhibitseffectivereductionofarangeoforganicsandmetals,being84

notably more efficient at reducing Cr(VI) than a commercially available synthetic Fe3O485

[20,21].Importantly,theparticlesizeofthispotentialcatalystsupportcanbemanipulated86

during its microbial synthesis, permitting its optimization for novel uses during scalable87

production[22].88

Furthermore,biogenicmagnetite is veryamenable to surface functionalizationwithother89

transitionmetalssuchaspalladium,mediatedbyreactivesurfaceFe(II)andanorganiclayer90

facilitatingreductiveprecipitationandattachmentofPd(0)[19]. Pd-biomagnetiteishighly91

reactivetowardCr(VI)andisefficientinhydrogenationofnitroaromatichydrocarbonsand92

halogenated solvents [21,23,24]. Additionally, when tested in a Heck reaction, coupling93

iodobenzene to ethyl acrylate or styrene, reaction rates using Pd-biomagnetite were94

superior or equal to an equimolar amount of commercially available colloidal Pd catalyst95

[19].AlthoughPdisexpensivetoapplyinanoilupgradingprocess,thecostcouldpotentially96

bereducedbyutilisingmetal recovered fromsecondarysourcessuchaselectronicwaste,97

scrapcatalytic convertersoreven lowgrade roaddust,whichcontains tracesofplatinum98

groupmetalsexhaustedfromautomotivecatalysts[25,26].Inaddition,theuseofmagnetite99

asasupportoffers thepotential formagnetic recoveryandreuseof thecatalyst;amajor100

drawbackforotherpreciousmetalcatalystsupports[27].101

In light of the previously reported catalytic properties of Pd-bionanomagnetite, and its102

amenability for optimization for a range of novel uses, this study aimed to assess the103

catalyticupgradingofoilbybiogenicmagnetitesupportedPd.Theupgradingpropertiesof104

biogenicmagnetitealone,asapotentialnon-functionalisednano-dispersedtransitionmetal105

catalyst,werealsoexamined.Extentsofoilupgradingwereassessed in termsofviscosity106

and API gravity, whilst true boiling point distributions and liquid, gas and coke mass107

distributionsprovidedimportantinformationonoilyieldsandcatalystefficacy.108

109

1. MaterialandMethods110

2.1.Bionanomagnetite(BnM)catalystpreparations.111

BiogenicmagnetitewassynthesisedbythedissimilatoryreductionofFe(III)inaferrihydrite112

suspension,bylate-logphaseculturesofGeobactersulfurreducensasdescribedpreviously113

[19,23,24,28,29].114

First,ferrihydritewassynthesisedbytherapidhydrolysisofa0.6MFe(III)chloridesolution115

viathedropwiseadditionof1MNaOHtopH7,whilststirringvigorously[30].Theresultant116

precipitatewaswashedsixtimesbycentrifugationat5000gfor20min,followedbyremoval117

ofthesupernatantandresuspensionin18MΩde-ionisedwater.118

AcultureofG.sulfurreducenswasgrowninamodifiedfreshwaterbasalmediumcontaining119

20 mM acetate and 40 mM fumarate as electron donor and acceptor respectively [31].120

Cultures were grown to late log-early stationary phase at 30°C under an 80:20 N2-CO2121

atmosphere [31]. Cells were then harvested by centrifugation at 5000g for 20 min and122

washedthreetimesinsterile30mMNaHCO3atpH7.123

For magnetite production, 1 L of autoclaved medium was prepared containing 20 mM124

acetate, 100 mM Fe (as ferrihydrite), 30 mM NaHCO3 and 10 µM anthraquinone-2,6-125

disulfonate (AQDS)asanelectronmediator. Asbefore, themediumwasspargedwithan126

80:20N2-CO2gasmixandadjustedtopH7.AsuspensionoftheharvestedG.sulfurreducens127

culturewasadded toa finalopticaldensity (at600nm)of0.6.Thepreparationwas then128

incubatedat30°Cfortwodaysinthedark,duringwhichablackmagneticprecipitatewas129

produced.Thisprecipitatewasthenmagneticallyseparatedandwashed3timesin18MΩ130

de-ionisedwatertoremovebacterialcells.Theproductionandpurityofthemagnetitewas131

confirmedbypowderX-raydiffraction(XRD)usingaBrukerD8AdvanceinstrumentwithCu132

Kα1 radiation. Transmission electron microscopy (TEM) analysis was performed using a133

Philipsmicroscope equipped with a 200 keV field emission gun and Gatan imaging filter134

(GIF200).Sampleswerepreparedbysuspendinginethanolpriortodrop-castingontoholey135

carbonsupportfilms(AgarScientific).136

2.2.Functionalizationofbiogenicmagnetitewithpalladium.137

Thesurfaceofbiogenicmagnetitecrystalliteswasfunctionalizedwithvariouswt%Pd138

loadings(%PdbymassofBnM)accordingtomethodsdescribedpreviously[19,23].139

PrecipitationofPdontomagnetitewasachievedviasurfaceFe(II)-mediatedreductive140

precipitationfromaN2spargedsolutionofNa2Pd(II)Cl4.Bottlescontainingthemagnetite141

suspendedinPd(II)-containingsolutionswereagitatedonrollersfor12hinthedark.Excess142

ionswerethenremovedbywashingthreetimesin18MΩde-ionisedwaterunderanN2143

atmosphere.Suspensionswerethenre-suspendedinacetonesothatthesuspensionswould144

bemisciblewiththeoilforreactortests.Priortotesting,1mLaliquotswereremoved,145

centrifugedat5000gfor5minandthesupernatantaliquotswereanalysedbyICP-AESto146

confirmtheremovalofPdfromsolution.Theremainingsolidmaterialwasdigestedin3M147

HClandagainanalysedbyICP-AEStodeterminethefinalFeconcentrationsandPdwt%148

loadingsachieved.149

2.3.OptimizationofPdloadingandreactivityonBnM.150

To optimize the wt% Pd loading and reactivity on magnetite surfaces, materials were151

producedasbeforebutwiththeinclusionofH2spargingsteps,followedbytheadditionof152

Fe(II)aqueoussolutionsinordertoincreasesurface-sorbedFe(II).Thesetwostepsaimedto153

provide excess electron donors for the enhanced reductive precipitation of Pd on to the154

magnetite surfaces. First,magnetite suspensionswere dispersed by sonication in awater155

bathfor5minandthenspargedwithH2for1min.A1mLsolutionofN2sparged100mM156

Fe(II)SO4·7H2Owas thenadded to the suspensions, followedby sonication for5minsand157

sparging with H2 for 1 min. An N2 sparged solution of Na2Pd(II)Cl4 was added, providing158

sufficientPdtoachievealoadingof20wt%Pd.Asbefore,bottleswereagitatedonrollers159

for12hours in thedark. Finally, the suspensionswerewashed three times in18MΩde-160

ionisedwater,andthreetimesinacetonetosuspendthePd-BnMinanoilmisciblesolvent161

andtoremoveexcessions.Pdloadingswereconfirmedasdescribedabove.162

2.4.Reactortests.163

TheheavyoilusedinthisstudywassuppliedbyPetrobankEnergyandResourcesLtd(now164

Touchstone Exploration Inc.), Canada. The initial API gravity and viscositywere 13.9˚ and165

952mPa·s.PriortoadditionofthePd-BnMpreparationstothebatchcatalyticreactor,the166

suspensionswereshaken,vortexedfor30s,sonicatedinawaterbathfor5minafterwhich167

analiquotwasrapidlyaddedtothereactor.168

2.5.Reactorconditions.169

Upgrading experiments were carried out in a 100 mL capacity stirred batch reactor170

(Baskerville, United Kingdom) using an experimental procedure described in detail171

previously[10,32].Theexperimentalconditions,chosentomatchtypicalinsitucombustion172

conditions for the THAIprocess,were as follows: 425 °C, 20bar initial pressure (N2), 500173

rpmstirringspeed,15goilwitha1mgg-1catalyst-to-oilratioanda30minreactiontime.174

Theexperimentswerealsocarriedout for thermalupgradingalone(withoutcatalyst)and175

withadditionofbiomagnetitewithoutpalladiumascontrolconditions.176

2.6.Analysisofoil.177

Before and after upgrading tests, an advanced rheometer AR 1000 (TA Instruments Ltd,178

UnitedKingdom)wasusedtomeasureoilviscosity,anAntonPaarDMA35densitymeter179

was used to determine the density and API gravity, and the True Boiling Point (TBP)180

distributionwas obtained using an Agilent 6850N gas chromatograph (GC) in accordance181

with the ASTM D2887 standard test method. Macromolecules, such as resins and182

asphaltenes, cannotbe accounted for in thismethod, since the calibrationmix contained183

only hydrocarbons from C5 to C40). A detailed description of these instruments and184

techniquescanbefoundin[9,32,33].185

The portion of the carbonaceous deposit after reaction that represented coke was186

determinedusingthermogravimetricanalysis (TGA).Thecarbonaceousdeposits left inthe187

reactor,whichareacompositeofcoke,residualoil,asphaltenesandcatalystnanoparticles,188

werecollectedandburntoffinaThermogravimetricAnalyser(NETZSCH-GeratebauGmbH,189

TG209F1Iris®).Theanalysiswasperformedusingaramptemperatureincreasefrom25to190

900 °C with an air flow rate of 50 mL min-1. The amount of coke in the deposit was191

determined from the temperature-weight loss curve. Previous analyses under the same192

experimentalconditionshaveestablishedthetemperaturesatwhichresidualoil(25to410193

°C),resinsandasphaltenes(410to620°C)andcoke(>620°C)areburnt-off[32–34].Upon194

heating, it is possible that the remaining catalyst may undergo modification, e.g. by195

oxidationofPd(0)oroxidationofmagnetitetohematite[35].However,thecatalyst/oilratio196

of0.001usedheremeansthatanymassdifferencesduetocatalystphasechangeswould197

resultinonlynegligibleeffectsontherecordedmassdistributionsofreactionproducts.198

Thenon-condensablegasyieldsaftertheupgradingreactionwasdeterminedbysubtracting199

the mass of the autoclave contents after reaction from the mass of heavy oil prior to200

reaction.201

202

3.ResultsandDiscussion203

3.1.Characterisationofbionanomagnetite.204

The microbial production of nano-scale magnetite was achieved via the respiratory205

reductionofFe(III),presentasaferrihydrite“gel”,byanoxicwashedcellsuspensionsofthe206

subsurfaceFe(III)-reducingbacteriumG.sulfurreducens.Completeconversiontomagnetite207

occurredinlessthan24h.Thebiosynthesisofapuremagnetiteproductwasconfirmedby208

both TEM and powder XRD (Figure 1). TEM further confirmed that the morphology and209

particlesize(15to30nm)wasverysimilartothatreportedpreviously[19,20,23,29]. The210

surface area of biomagnetite, generated using the same production methods, has been211

reportedpreviously by our laboratory as 17.1m2 g-1 [29].However, this figure should be212

regardedasanestimateduetoparticleaggregationthatroutinelyarisesduringpreparation213

andmeasurementusingtheBETN2adsorptionmethod.214

Previous studies have demonstrated that the surface of biogenic magnetite can be215

functionalised with Pd(0) via surface Fe(II)-mediated reductive precipitation from a216

N2spargedsolutionofNa2Pd(II)Cl4[19,23].Initially,twoloadingsof4.3and7.1wt%Pdwere217

achievedbysupplyingmagnetitewithPd(II)concentrationsthatshouldhavebeensufficient218

for5and20wt%Pd,respectively,ifallthesuppliedPd(II)wasreduced.Thissuggeststhat219

7.1wt%Pdwas themaximumPd loadingachievablevia this firstmethod.XRDconfirmed220

thatsolidphasePdwasassociatedwiththemagnetite(Figure2).Usingthesamesynthesis221

methods as used here, Coker et al. (2010) report that for bionanomagnetite augmented222

with 5mol.%Pd, particles of palladium (approx. 5 nm)wereprecipitatedonto the larger223

magnetitestructures(approx.20nm),asobservedviaTEM.224

Afterfurtheroptimization,involvingH2spargingstepsfollowedbyadditionofaqueousFe(II)225

to magnetite suspensions (described in detail above), a Pd loading of 9.5 wt% Pd was226

achieved.This increased loading is likelya resultof theadditionof thesereductants,with227

both H2 and Fe(II) enhancing the reductive precipitation of Pd onto magnetite surfaces.228

Despite an excess of Pd(II) supplied, the final loading of 9.5wt%Pd likely represents the229

upperlimitofPd-loadingachievableusingtheprotocolsdescribedhere.230

3.2.Catalyticupgradingofoilbypalladisedbiogenicmagnetite.231

Theyieldbalancebetweenliquid(upgradedoil),gasandcokeaswellastheupgradedoilAPI232

gravity and viscosity relative to the feed oil after upgrading experiments are displayed in233

Table 1. It can be seen that the liquid and coke yields after upgrading are inversely234

correlated, as an increase in liquid yield correspond to a decrease in coke yield. Whilst235

thermalcrackingproducedthelowestyieldofupgradedoil(78.8wt%)andthehighestcoke236

yield (10.2 wt%), catalytic upgrading with bionanomagnetite (BnM) yielded 82.3 wt%237

upgraded oil and 6.9wt% coke. Increasing the loadings of palladium (Pd) onto the BnM238

from4.3to9.5wt%,ledtoacorrespondingincreaseintheupgradedoilyieldfrom82.3to239

89.6 wt%, and a decrease in coke from 6.9 to 3.0 wt%. A similar observation has been240

reported for thermal cracking relative to catalytic upgrading by Hart et al. [9] and Al-241

Marshedetal. [13]. Ithasbeenreportedthat thermalcrackingproceedsbya freeradical242

mechanism that promotes the precipitation of coke precursors, such as resins and243

asphaltenes.Consequently,polymerisationandcondensationoccurs,as theabstractionof244

hydrogen,methyl,ethylandothermaterialfromdepositedpolyaromaticspeciestothegas245

phase occurs, resulting in significant coke formation. However, previous work has246

demonstrated that catalysts bearing various transitionmetals display reduced coke yields247

andcorrespondingincreasesinupgradedoilyields.Assuch,theadditionofBnMaugmented248

withPdmayalsoexhibit theability to transferhydrogen,methylandothermaterial from249

thegasintotheoilphase[13].Thisisconsistentwiththeobservationthatthermalcracking250

aloneproducedmoregas(11wt%)relativetoupgradingbyBnM(10.8wt%)andBnM+9.5251

wt%Pd(7.4wt%).252

253

Table1.Massdistributionsofliquid,gasandcoke,andAPIgravityandviscosityofoilbefore254

andafterthermalcrackingorreactionwithBnM,withandwithoutaddedPd.Resultsfor255

CoMo/Al2O3[34]andNiMo/Al2O3[36],astypicalrefinerycatalystsareshownfor256

comparison.257

Sample

Liquid

(wt%)

Gas

(wt%)

Coke

(wt%)

APIgravity

(˚)

Viscosity

(mPa·s)

Feedoil 13.2 1031

Thermalcracking 78.8±0.2 11.0±0.3 10.2±0.6 18.5±0.5 28.3±2.1

Co-Mo/Al2O3 80.8±0.1 7.1±0.5 12.2±0.1 23.6±0.2 4.4±1.3

Ni-Mo/Al2O3 87.3±0.5 7.7±0.7 5.0±0.8 24.9±0.3 3.7±0.6

BnM 82.3±0.3 10.8±0.2 6.9±0.4 18.7±0.3 22.6±1.7

BnM+4.3wt%Pd 85.2±0.2 8.9±0.1 5.9±0.3 20.6±0.3 17.8±1.1

BnM+7.1wt%Pd 83.7±0.4 10.9±0.6 5.4±0.2 19.9±0.4 15.6±2.3

BnM+9.5wt%Pd 89.6±0.7 7.4±0.3 3.0±0.4 21.0±0.2 5.9±0.8

258

The viscosity and API gravity of crude oil are properties that determine its value, refine-259

ability and ease of pipeline transportation. The API gravity of the upgraded oil improved260

from13.2o (feedstock)to18.5o (thermalcracking),18.7o (BnMonly)and21.0o (BnM+9.5261

wt%Pd),a159%increaseinthecaseofthelatter.Conversely,theviscosityofupgradedoil262

decreasedsignificantlyfrom1031mPa·s(feedstock)to28.3mPa·s(thermalcracking),22.6263

mPa·s(BnMonly)and5.9mPa·s(BnM+9.5wt%Pd),a99.4%decrease.Thismagnitudeof264

viscosityreductioniscapableofimprovingproductionandpipelinetransportationowingto265

theleveloffluidityexperiencedatlowerviscosity[37].Similarimprovementsincatalytically266

upgradedoilversusthermallycrackedheavyoilusingconventional,chemicallysynthesised267

catalysts has been reported byHart et al. [9,32,33]. In the absence of a transitionmetal268

oxide catalyst, free radicals lead to adduct formation as a result of the addition reaction269

between two ormore active hydrocarbon chains. This, in turn, leads to the formation of270

larger molecular weight hydrocarbons, which limits any increase in the API gravity or271

decreases in viscosity of the produced oil [37,38]. However, the presence of a transition272

metal oxide catalyst, such as BnM, likely promotes hydrogen-transfer reactions between273

hydrocarbons in thegasand liquidphases,whichhelps tocap free radicalsonce theyare274

formed, thus moderating the adduct formation. The augmentation of BnM with surface275

associatedPd likelypromotes furtherhydrogenation reactions,whichproduces significant276

improvementsinAPIgravityandviscosityoftheupgradedoilasthepercentageloadingof277

PdonBnMincreasesfrom4.3to9.5wt%.278

TheextentoftheimprovementinAPIgravityandviscositywithincreasingPdloadingmay279

belimitedduetothesulfurcontentoftheheavyoil(3.52wt%).Ithasbeenreportedthat280

sulfur and sulfur-containing species (e.g. H2S, RSH, RSSR) may lead to poisoning of Pd-281

surfaces [39], which may have impeded the activity and performance of the particles.282

Indeedasimilareffectwasobservedinthetestingofpalladizedironforthedechlorination283

of groundwater [40]. The experiments demonstrated that the dechlorination reaction284

occurredefficientlyuntilthesurfaceofthePd/Febecamefouled.Itwasfoundthatreduced285

sulfurspeciescouldeventually leadtopermanentpoisoningof thepalladium.Contrary to286

this, other studies have shown that sulfurmay prolong the life of platinum groupmetal287

catalysts and even promote activity and enhance selectivity [39,41,42]. However this288

phenomenon is typically noted for sulfur concentrations on catalyst surfaces that are289

significantlylowerthanthatreportedforthebulksulfurcontentoftheheavyoilusedinthis290

study(e.g.<0.5wt.%).Despitethis,theauthorsofapreviousstudyutilisingthesamenative291

oil and reactor conditions as used here, assumed that although sulfur is present in292

potentially inhibitory concentrations, itmay be present in non-available forms due to its293

associationwith othermetals present in the native feed oil [34]. Although the extent of294

sulphur deactivation was not evaluated in this study, the data suggest that the Pd-BnM295

catalystperformedcomparablywithtypicallyrefinerycatalysts(table1).296

Thetrueboilingpoint(TBP)distributioncurvesobtainedfromsimulateddistillation(SIMDIS)297

of the feedoilandoilsupgradedby thermalcrackingandBnMarepresented inFigure3.298

TheTBPcurvesshowthecumulativevolumedistilledasafunctionoftemperature.Theshift299

oftheTBPcurvesfromreactortestscontainingaddedcatalysttotheleftoftheTBPcurveof300

the feed oil indicate that the upgraded oil contains lighter hydrocarbon components301

comparedtothefeedoil.302

Anoticeableimprovementintheamountofdistillateobtainablecanbeobservedbetween303

the temperaturerangesof230to430oC, forall reactor tests relative to the feedoil.The304

TBP curves of the upgraded oils by thermal cracking and BnM only are approximately305

identical,whichisinlinewiththeirsimilarAPIgravitiesandviscositiesreportedinTable1.306

However,theadditionofPdtoBnMproducedafurtherincreaseintheamountofdistillate307

fractionsdistilledbetweentheboilingtemperaturerangeof200to450oC,withsignificant308

increases recorded when the Pd loading of the BnM was increased to 9.5 wt% Pd. This309

observedincreaseinnaphthaandmiddledistillatefractionsupontheincreasingadditionof310

Pd to BnM can be attributed to the enhanced ability of Pd to facilitate hydrogenation311

reactions, compared with Fe. Hence, the synergistic effect of Fe plus Pd enhanced the312

conversion of high boiling hydrocarbon components into lighter hydrocarbon fractions,313

relativetothatobservedforBnMalone.ThesetrendsinTBPcurvesareinlinewiththeAPI314

gravitiesandviscositiesoftheupgradedoils,relativetothefeedoil,showninTable1.315

3.3.Outlookandcomparisonwithcommerciallyavailablecatalysts.316

In summary, the experiments reported here have demonstrated that both biogenic317

magnetite augmented with Pd and biogenic magnetite alone are capable of significant318

catalytic upgrading of oil, as measured by increases in liquid yield, API gravity and319

correspondingdecreasesinviscosity.Areductionincokecontentandanincreaseinlighter320

hydrocarbonfractionsfurtherdemonstratethepotentialforitsuseasacatalystintheTHAI321

CAPRIprocess.322

Dispersed nanoparticulatemagnetite offers increased oil contact through a large surface-323

area-to-volumeratioanddecreaseddiffusionpath lengthscomparedwithusing fixed-bed324

reactors [10,43]. Indeed, the use of biomagnetite as the catalyst support not only offers325

catalytic properties itself but also provides a large surface area over which the Pd is326

precipitated, facilitated by the large pool of surface associated Fe(II) that serves as the327

reductant. As Pd also promotes hydrogenation and hydrogenolysis reactions [14–16], Pd-328

augmentedbionanomagnetiterepresentsabi-functionalcatalystwithincreasedactivityand329

performance.330

Indeed, previous studies have shown that commercially available Ni-Mo/Al2O3 and331

dispersed ultrafine Co-Mo/Al2O3 achieved a higher API gravity increase than the tested332

catalystsinthisstudyunderthesameconditionsasusedhere(23.6and22.5°respectively)333

[34,36].However,lowercokecontentsandahigherliquidyieldof89.6wt%using9.5wt%334

Pd-BnMwereachievedinthisstudy,comparedwith87.3and80.8wt%forNi-Mo/Al2O3and335

dispersed ultrafine Co-Mo/Al2O3, respectively. The coke content of 12.15 wt% for Co-336

Mo/Al2O3was higher than achieved for all the testedmaterials reported here, and lower337

cokingwasachievedusing9.5wt%Pd-BnM(3.0wt%) thanwas reported forNi-Mo/Al2O3338

(5.0wt%).Ascokeproductionreducesboththeliquidyieldandthelifetimeofthecatalyst339

[44],decreasesinthecokelevelsachievedbyPd-BnMareafavourableresult.However,itis340

not clearwhether suchgains in coke reductionareable tooff-set the lower levelsofAPI341

gravityandviscosityachieved.342

Similar comparisons can bemadewith dispersed transitionmetal catalysts. Fe2O3,MoS2,343

MoO3,FeSandNiOhaveallshownslightlyhigherlevelsofupgradingintermsofAPIgravity344

(typically~21°)[11,13].However,viscositieswereallsignificantlyhigherforthesematerials345

(typically 70 to 140 mPa·s; though this may be due to the slightly different reactor346

conditionsused)andcokelevels(>4.3wt%coke)wereagain,notaslowasfor9.5wt%Pd-347

BnM(3.0wt%coke).348

Recent work has investigated upgrading using nanoparticulate Pd and Pt supported on349

bacterialcellscaffolds,underthesamereactorconditionsusedhere[34,36].Usingsimilar350

Pdloadings,theauthorsachievedverysimilarimprovementsinyields,cokingandviscosity,351

supportingourfindingthatbiotechnologicalapproachestocatalystsynthesiscangenerate352

commerciallycompetitivecatalystsusing“green”biosynthesisroutes.353

Whilst such improvements inoilupgradingbyPd-BnM,compared to thoseof commercial354

andother testedmaterials,mayappearmodest at the laboratory scale testedhere, such355

incremental changes may lead to significant benefits in process economics at full scale.356

Recent work has demonstrated that biogenic magnetite production can be successfully357

scaled up from laboratory to pilot plant-scale whilst controlling its reactivity, magnetic358

properties and particle size, suggesting that its exploitation in commercial settings are359

achievable [45]. In addition, the production process may offer a more environmentally360

benign route to catalyst production than established chemical methods and thus, such361

environmental benefits may offset the cost of using expensive precious metals, such as362

palladium [27]. Indeed, recent work suggests that this cost may be further reduced by363

sourcing metals from electronic waste, scrap catalytic converters or road dust, which364

contains platinum group metals exhausted from automotive catalysts [25,26]. In this365

context,thecompetitiveresultsreportedheresuggestthataneconomiccostanalysismay366

bewarranted.367

Finally,ourexperimentshavereliedonthelaboratoryproductionofmagnetiteusingapure368

culture of a model Fe(III)-reducing bacterium. However, the natural occurrence of such369

speciesinsub-surfacesediments,alongsidenativeFe(III)-bearingoxides,mayrepresentthe370

potentialfortheinsitugenerationofacatalyst.Futureexperimentswill,therefore,focuson371

achieving the stimulation of such organisms in relevant geological formations and, thus,372

eliminating the need for delivery of a catalyst to the reaction front of the THAI process.373

Enhanced oil recovery by suchmeansmay further reduce financial overheads associated374

withtheTHAI-CAPRIprocess.375

376

4.Acknowledgements.377

Funding: Thisworkwas supported by theUK Engineering and Physical Sciences Research378

Council(EPSRC;grantnumberEP/J008303/1).Thedatafromthisstudyareavailableonline379

viaepapers.bham.ac.uk.Theauthorswould liketothankPetrobankEnergyandResources380

Ltd. and Touchstone Exploration Inc., Canada for supplying the heavy oil used in these381

experiments.WealsothankProf.LynneMacaskie,UniversityofBirmingham,foradviceand382

helpful discussions; Paul Lythgoe, Manchester Analytical Geochemistry Unit, for ICP-AES383

analyses; Dr. John Waters for XRD and Dr. Mike Ward, University of Leeds EPSRC384

NanoscienceandNanotechnologyFacility,forTEM.385

386

5.References387

[1] Martinez-PalouR,MosqueiraMdeL,Zapata-RendonB,Mar-JuarezE,Bernal-388

HuicocheaC,delaCruzClavel-LopezJ,etal.Transportationofheavyandextra-heavy389

crudeoilbypipeline:Areview.JPetSciEng2011;75:274–82.390

doi:10.1016/j.petrol.2010.11.020.391

[2] CarrilloJA,CorredorLM.Upgradingofheavycrudeoils:Castilla.FuelProcessTechnol392

2013;109:156–62.doi:10.1016/j.fuproc.2012.09.059.393

[3] WeissmanJG,KesslerRV,SawickiRA,BelgraveJDM,LaureshenCJ,MehtaSA,etal.394

Down-HoleCatalyticUpgradingofHeavyCrudeOil.Energy&Fuels1996;10:883–9.395

doi:10.1021/ef9501814.396

[4] ShahA,FishwickR,WoodJ,LeekeG,RigbyS,GreavesM.Areviewofnovel397

techniquesforheavyoilandbitumenextractionandupgrading.EnergyEnvironSci398

2010;3:700–14.doi:10.1039/b918960b.399

[5] XiaTX,GreavesM.UpgradingAthabascatarsandusingtoe-to-heelairinjection.JCan400

PetTechnol2002;41.401

[6] XiaTX,GreavesM,TurtaAT,AyasseC.THAI-A“short-distanceDisplacement”insitu402

combustionprocessfortherecoveryandupgradingofheavyoil.ChemEngResDes403

2003;81:295–304.doi:10.1205/02638760360596847.404

[7] GreavesM,El-SaghrA,XiaTX.CAPRIhorizontalwellreactorforcatalyticupgradingof405

heavyoil:Advancesinoilfieldchemistry:Downholeupgrading.PreprChemSocDiv406

PetChem2000;45:595–8.407

[8] ShahA,FishwickRP,LeekeGA,WoodJ,RigbySP,GreavesM.Experimental408

optimizationofcatalyticprocessinsituforheavy-oilandbitumenupgrading.JCan409

PetTechnol2011;50:33–47.doi:10.2118/136870-PA.410

[9] HartA,ShahA,LeekeG,GreavesM,WoodJ.OptimizationoftheCAPRIprocessfor411

heavyoilupgrading:Effectofhydrogenandguardbed.IndEngChemRes412

2013;52:15394–406.doi:10.1021/ie400661x.413

[10] HartA,GreavesM,WoodJ.Acomparativestudyoffixed-bedanddispersedcatalytic414

upgradingofheavycrudeoilusing-CAPRI.ChemEngJ2015;44:1–11.415

doi:10.1016/j.cej.2015.01.101.416

[11] Al-MarshedA,HartA,LeekeG,GreavesM,WoodJ.Optimizationofheavyoil417

upgradingusingdispersednanoparticulateironoxideasacatalyst.Energy&Fuels418

2015;29:6306–16.419

[12] HashemiR,NassarNN,PereiraAlmaoP.Nanoparticletechnologyforheavyoilin-situ420

upgradingandrecoveryenhancement:Opportunitiesandchallenges.ApplEnergy421

2014;133:374–87.doi:10.1016/j.apenergy.2014.07.069.422

[13] Al-MarshedA,HartA,LeekeG,GreavesM,WoodJ.Effectivenessofdifferent423

transitionmetaldispersedcatalystsforinsituheavyoilupgrading.IndEngChemRes424

2015;54:10645–55.425

[14] WoodJ,BodenesL,BennettJa,DeplancheK,MacaskieLE.Hydrogenationof2-426

Butyne-1,4-diolUsingNovelBio-PalladiumCatalysts.IndEngChemRes2010;49:980–427

8.doi:10.1021/ie900663k.428

[15] DelBiancoA,PanaritiN,DiCarloS,ElmouchninoJ,FixariB,LePerchecP.429

Thermocatalytichydroconversionofheavypetroleumcutswithdispersedcatalyst.430

ApplCatalAGen1993;94:1–16.doi:http://dx.doi.org/10.1016/0926-860X(93)80041-431

N.432

[16] SchüthC,ReinhardM.Hydrodechlorinationandhydrogenationofaromatic433

compoundsoverpalladiumonaluminainhydrogen-saturatedwater.ApplCatalB434

Environ1998;18:215–21.doi:10.1016/S0926-3373(98)00037-X.435

[17] ChaplinBP,ReinhardM,SchneiderWF,SchüthC,ShapleyJR,StrathmannTJ,etal.436

CriticalReviewofPd-BasedCatalyticTreatmentofPriorityContaminantsinWater.437

EnvironSciTechnol2012;46:3655–70.doi:10.1021/es204087q.438

[18] CokerVS,TellingND,vanderLaanG,PattrickRAD,PearceCI,ArenholzE,etal.439

HarnessingtheExtracellularBacterialProductionofNanoscaleCobaltFerritewith440

ExploitableMagneticProperties.ACSNano2009;3:1922–8.doi:10.1021/nn900293d.441

[19] CokerVS,BennettJA,TellingND,HenkelT,CharnockJM,vanderLaanG,etal.442

Microbialengineeringofnanoheterostructures:biologicalsynthesisofamagnetically443

recoverablepalladiumnanocatalyst.ACSNano2010;4:2577–84.444

doi:10.1021/nn9017944.445

[20] CuttingRS,CokerVS,TellingND,KimberRL,PearceCI,EllisBL,etal.OptimizingCr(VI)446

andTc(VII)remediationthroughnanoscalebiomineralengineering.EnvironSci447

Technol2010;44:2577–84.448

[21] WattsMP.Novelbionanotechnologicalapproachesfortheremediationof449

contaminatedland.UniversityofManchester,2014.450

[22] ByrneJM,TellingND,CokerVS,PattrickRAD,VanDerLaanG,ArenholzE,etal.451

Controlofnanoparticlesize,reactivityandmagneticpropertiesduringthe452

bioproductionofmagnetitebyGeobactersulfurreducens.Nanotechnology453

2011;22:455709.454

[23] WattsMP,CokerVS,ParrySA,ThomasRAP,KalinR,LloydJR.Effectivetreatmentof455

alkalineCr(VI)contaminatedleachateusinganovelPd-bionanocatalyst:Impactof456

electrondonorandaqueousgeochemistry.ApplCatalBEnviron2015;170-171:162–457

72.doi:10.1016/j.apcatb.2015.01.017.458

[24] CreanDE,CokerVS,vanderLaanG,LloydJR.Engineeringbiogenicmagnetitefor459

sustainedCr(VI)remediationinflow-throughsystems.EnvironSciTechnol460

2012;46:3352–9.doi:10.1021/es2037146.461

[25] MurrayAJ.RecoveryofPlatinumGroupMetalsfromSpentFurnaceLiningsandUsed462

AutomotiveCatalysts.UniversityofBirmingham,2012.463

[26] MacaskieLE,MikheenkoIP,YongP,DeplancheK,MurrayAJ,Paterson-BeedleM,et464

al.Today’swastes,tomorrow'smaterialsforenvironmentalprotection.465

Hydrometallurgy2010;104:483–7.doi:10.1016/j.hydromet.2010.01.018.466

[27] LloydJR,ByrneJM,CokerVS.Biotechnologicalsynthesisoffunctionalnanomaterials.467

CurrOpinBiotechnol2011;22:509–15.doi:10.1016/j.copbio.2011.06.008.468

[28] CuttingRS,CokerVS,FellowesJW,LloydJR,VaughanDJ.Mineralogicaland469

morphologicalconstraintsonthereductionofFe(III)mineralsbyGeobacter470

sulfurreducens.GeochimCosmochimActa2009;73:4004–22.doi:DOI471

10.1016/j.gca.2009.04.009.472

[29] WattsMP,CokerVS,ParrySA,PattrickRAD,ThomasRAP,KalinR,etal.Biogenic473

nano-magnetiteandnano-zerovalentirontreatmentofalkalineCr(VI)leachateand474

chromiteoreprocessingresidue.ApplGeochemistry2015;54:27–42.475

doi:10.1016/j.apgeochem.2014.12.001.476

[30] LovleyDR,PhillipsEJP.Organicmattermineralizationwithreductionofferricironin477

anaerobicsediments.ApplEnvironMicrobiol1986;51:683–9.478

[31] LloydJR,LeangC,HodgesMyersonAL,CoppiMV,CuifoS,MetheB,etal.479

BiochemicalandgeneticcharacterizationofPpcA,aperiplasmicc-typecytochromein480

Geobactersulfurreducens.BiochemJ2003;369:153–61.doi:10.1042/BJ20020597481

BJ20020597[pii].482

[32] HartA,LeekeG,GreavesM,WoodJ.Down-holeheavycrudeoilupgradingbyCAPRI:483

Effectofhydrogenandmethanegasesuponupgradingandcokeformation.Fuel484

2014;119:226–35.doi:10.1016/j.fuel.2013.11.048.485

[33] HartA,LewisC,WhiteT,GreavesM,WoodJ.Effectofcyclohexaneashydrogen-486

donorinultradispersedcatalyticupgradingofheavyoil.FuelProcessTechnol487

2015;138:724–33.488

[34] HartA,OmajaliJB,MurrayAJ,MacaskieLE,GreavesM,WoodJ.Comparisonofthe489

effectsofdispersednoblemetal(Pd)biomasssupportedcatalystswithtypical490

hydrogenation(Pd/C,Pd/Al2O3)andhydrotreatmentcatalysts(CoMo/Al2O3)forin-491

situheavyoilupgradingwithToe-to-HeelAirInjection(THAI).Fuel2016;180:367–76.492

[35] CornellRM,SchwertmannU.Theironoxides.Second.Weinheim:WILEY-VCHVerlag;493

2003.494

[36] OmajaliJ.Novelbionanocatalystsforgreenchemistryapplications.Universityof495

Birmingham,2015.496

[37] HartA.Advancedstudiesofcatalyticupgradingofheavyoils.Universityof497

Birmingham,UK,2014.498

[38] HabibFK,DinerC,StrykerJM,SemaginaN,GrayMR.Suppressionofaddition499

reactionsduringthermalcrackingusinghydrogenandsulfidedironcatalyst.Energy&500

Fuels2013;27:6637–45.501

[39] DunleavyJK.Sulfurasacatalystpoison.PlatinMetRev2006;50:110.502

[40] KorteNE,ZutmanJL,SchlosserRM,LiangL,GuB,FernandoQ.Fieldapplicationof503

palladizedironforthedechlorinationoftrichloroethene.WasteManag2000;20:687–504

94.505

[41] TongYJ.Unconventionalpromotersofcatalyticactivityinelectrocatalysis.ChemSoc506

Rev2012;41:8195–209.doi:10.1039/C2CS35381D.507

[42] McCueAJ,AndersonJA.Sulfurasacatalystpromoterorselectivitymodifierin508

heterogeneouscatalysis.CatalSciTechnol2014;4:272–94.doi:10.1039/C3CY00754E.509

[43] AngelesMJ,LeyvaC,AncheytaJ,RamírezS.Areviewofexperimentalproceduresfor510

heavyoilhydrocrackingwithdispersedcatalyst.CatalToday2014;220:274–94.511

doi:10.1016/j.cattod.2013.08.016.512

[44] ForzattiP,LiettiL.Catalystdeactivation.CatalToday1999;52:165–81.513

doi:http://dx.doi.org/10.1016/S0920-5861(99)00074-7.514

[45] ByrneJM,MuhamadaliH,CokerVS,CooperJ,LloydJR.Scale-upoftheproductionof515

highlyreactivebiogenicmagnetitenanoparticlesusingGeobactersulfurreducens.JR516

SocInterface2015;12.517

518

519

520

521

Fig.1.BrightfieldTEMandpowderXRDdiffractogramofbiogenicmagnetite.522

523

524

525

Fig.2.XRDdiffractogramof9.5wt.%Pd-coatedbiogenicmagnetite.526

527

Fig.3.Trueboilingpointdistributioncurvesoffeedandupgradedoilsafterthermalcracking528

andafterreactionwithPd-coatedbiomagnetite(BnM).529

530