corrosion rates of stainless steels in renewable biofuel ......the materials is critical for...

ISSN 2070-2051, Protection of Metals and Physical Chemistry of Surfaces, 2018, Vol. 54, No. 4, pp. 724–744. © Pleiades Publishing, Ltd., 2018.

PHYSICOCHEMICAL PROBLEMSOF MATERIALS PROTECTION

Corrosion Rates of Stainless Steels in Renewable Biofuel Sourcesof Refined Rapeseed Oil, Waste Cooking Oil

and Animal Waste Lard1, 2

András Gergelya, *, Antal Krójerb, Zoltán Vargac, and Tamás Kristófa

aDepartment of Physical Chemistry, Institute of Chemistry, University of Pannonia, Hungary, 8200 VeszprémbDepartment of Inspection and Maintenance, Mol Nyrt., Hungary, 2443 Százhalombatta

cDepartment of MOL Hydrocarbon and Coal Processing, Institute of Chemical and Process Engineering,University of Pannonia, Hungary, 8200 Veszprém

*e-mail: [email protected], [email protected] March 2, 2017

Abstract—The 1.4541 and the 1.4571 stainless steels and a carbon steel were subjected to immersion corrosiontests in stagnant and stirred biomass fuel sources such as rapeseed oil, waste cooking oil and animal waste lardas well as their emulsions with 5 and 50 wt.% aqueous citric acid solutions at a temperature of 80°C so as tomodel storage, handling and purification conditions. Passivation of carbon steel was facilitated by f low of theless acidic rapeseed and waste cooking oils and increased flux of oxygen. Carbon steel was sensitive for thehigher concentration of proton donor species, acidity of the waste lard. Higher mass loss rates correlated withincreased corrosion currents measured in citric acid solution by electrochemical methods. Flow of the bio-masses and increased acid concentration of the emulsions were beneficial for the passivation of stainlesssteels. Although corrosion related mass loss, dissolution rate of the passive layers increased by f low and highacidity of the f luids, both the formation and compactness of passive layers are facilitated by the biomasseswith higher concentration of oxygen donating species like water, alcohol and acids. Surface transformation ofthe passivating steels was reflected by decreasing electrochemical pseudo-capacity of the interfaces andincreasing resistance of the passive layers derived from the results of Tafel and Stern methods as well asimpedance results. Anti-correlation between mass loss results obtained by immersion in the biomasses andelectrochemical data measured in dilute aqueous citric acid solution is explained by the varied compactness,resistance of the passive layers and exchange currents of the steel electrodes due to the orders of magnitudedifferent activities of the hydrogen ion in the biomass mixtures and citric acid solution.

Keywords: renewable biomass fuel sources, corrosion and f low assisted passivation of carbon steels, biomassand f low assisted passivation of stainless steelsDOI: 10.1134/S2070205118040184

INTRODUCTION

There is a strong incentive to rationalize industryprocesses in terms of energy consumption, minimiseemission of greenhouse gases and develop or substitutetechnologies with an aim to reduce spectrum and theamount of materials to be regularly disposed. Thus,the priority of complete or partial substitution of fossilfuels with renewable biomass fuel sources is an utmostendeavour. Virgin and used vegetable oils besides ani-mal waste lard (AWL) are feasible biomass materials.Natural triacylglycerols are found in vegetable oil sandthey are in majority in waste lard. Chemical stability ofthe materials is critical for long-term reliable utilisa-

tion. Methyl ester of fatty acids is less oxidation-resis-tant and more hygroscopic so their long-term storageis not safe. If vegetable oils like tall oil with high freefatty acid contents [1–5] similar to AWL are not pro-cessed, then they will pose corrosion risks to theequipment of storage, transportation and pretreat-ment during processing. Storage tanks and facilitieswith tube systems can severely impacted by biocorro-sion phenomena [6]. Engine and exhaust parts of thevehicles are particularly subjected to corrosion risks,depending on the fuel characteristics. Corrosion phe-nomena are primarily connected to sulphur com-pounds of the fuel sources, acid and water traces aswell. Acidity and alcohol content of gasoline are thekey factors in corrosivity [7–9]. Compared to ethanol,similarly to water and salt impurities methanol posesmuch greater threat not only to copper, brass, zinc andcarbon steel (CS) but low alloy steels (LASs) too [10].

1 The article is published in the original.2 Supplementary materials are available for this article at

10.1134/S2070205118040184 and are accessible for authorizedusers

724

CORROSION RATES OF STAINLESS STEELS 725

Besides the economic reasons, stainless steels (SSs)with high chromium contents are the most reliablematerials for long-term safe integrity of structures[11]. Stainless steels with overlays of Pb-Sn alloys fea-turing high cathodic overvoltage for hydrogen dis-charge are good candidates for storage tanks in whichacidity and unsaturation of the fuels increases by timedue to oxidation processes. Galvanised steels withchromium-rich top-layers feature firm corrosionresistance just as well as SSs, withstanding the attackof non-compliant fuels. The activity of hydrogen ion(acidity of the system) [12], the presence of criticalanalytes like chloride [13, 14] sulphate and acetate ionsalong water are associated with higher corrosion ratesand severe corrosion damages.

Although limitation of electrochemical characteri-sations was clarified as validation with the weight lossmethod indicated deviation in inhibition efficiencies,impedance spectroscopy was successful to study cor-rosion effects of homogeneous stagnant f luids on mildsteel (AISI 1005) in ethanol in the presence andabsence of ethanolamine type inhibitors [15]. The cor-rosion of SS (SS410) in comparison with a chromiumalloy (Fe–10Cr) and mild steel (SUS430) in aqueousethanol mixed with sulfuric acid indicated decreasingcorrosion rates with increasing relative amount ofalcohol in the reaction mixture and chromium contentof the alloys [16]. Corrosion of the Al-Si alloy in aque-ous ethanol is negligible up to 2% of water [17] but cor-rosion rate increases up to a maximum at 20% of watercontent then it decreases due to passivation. The Fe–Mn–Al–Si alloys with greater aluminium content areparticularly susceptible for passivation. Nonetheless,high resistivity against general corrosion in gas-oilleads to increased susceptibility to pitting in alcoholsolutions. Even stress corrosion cracking can occurwith steels in ethanol in the presence of chloride andoxygen [18] which is aggravated by acetic acid [19].Pure aluminium (A1319) is incompatible with gaso-line-alcohol mixtures but the A16061 alloy with mag-nesium and silicon elements is firmly resistant [20]. Inalcohol-gasoline mixtures, wear rate is proportional tocorrosion rate reaching a maximum at 20% alcoholcontent [21]. Increasing degree of unsaturation, polar-ity results in higher hygroscopic nature of the biofuels.Corrosiveness of the biodiesels is related to the rela-tively high free fatty acid [22], alcohol and water con-tents as well as the contingent microbial activities [23–27]. The study of canola based biodiesel on pure alu-minium established a relationship between fasterintegrity loss and degree of impurities in the biofuels[28]. Stainless steel and aluminium are compatiblewith biodiesel but bronze, brass, copper, lead, tin,zinc, iron and nickel are incompatible. To circumventcontamination of the fuels with biological sources andto obviate subsequent bio-degradation processes arethe key factors in stabilisation of the biofuels and low-ering the risks of integrity loss of structural materials[29]. Interestingly, biofuels with low total acid number

PROTECTION OF METALS AND PHYSICAL CHEMISTR

(TAN) can still cause severe corrosion in comparisonwith the ones of similar characteristic but with higheracid content. Nevertheless, TAN is a sensitive param-eter to forebode quality loss of the fuel. Piston metalsand the liners of engine parts showed heavy corrosionlosses in biodiesels as compared to the results carriedout with neat diesel [30]. Storage stability at varioustemperatures confirmed increasing free fatty acid con-tent of poultry fat and diesel mixtures over time expo-sure [31]. Compatibility assessment of automotivematerials in biodiesel, both static and on-road enginetests led to increased friction and wear in biodiesels incomparison with petrodiesel [32]. Aluminium, copperand bronze experienced rapid and extensive corrosionin biodiesel blends [33]. Although SSs firmly with-stood in biodiesels, the degree of integrity loss washigher than in petrodiesel. In ultra-low sulphur dieselblends with rapeseed methyl ester, aluminium withhigher affinity to passivate under atmospheric condi-tions proved to be less susceptible to corrode than cop-per [34]. Dissolution of copper and aluminium was fargreater in palm biodiesel blends than SS of whichintegrity loss was practically ignorable [35]. UnlikeSSs, copper acts as an oxidiser catalyser contributingto degradation of biofuels which is ref lected in proper-ties of increasing TAN, viscosity and density. Whatinteresting is SS in itself compatible with biodiesels butit also changes fuel properties over time. The mixturesof vegetable oils of crude palm and soybean oil, wastecooking oil (WCO) and AWL were stable in acceler-ated oxidation test at 77°C for 180 days [36]. Free fattyacids in vegetable oils and AWLs did not cause highercorrosion risks to CS (ASTM A 293 Gr C) which isa basic structural material of storage systems.

In this work, corrosion rates of two SSs and a CSimmersed in biofuel feedstock are investigated tomodel storage, transportation and purification. Rape-seed oil (RSO), waste cooking oil and animal wastelard were used as biomass fuel sources for the immer-sion tests. Corrosion resistivity of the EN1.4541 (AISI321) and the EN 1.4571 (AISI 316) SSs was monitoredby electrochemical direct current methods as a func-tion of time. To assess validity of the results obtainedin an aqueous electrolyte, temperature dependence ofthe corrosion rates was analysed and the data com-pared to weight loss results for the sake of recognisingcorrelation between corrosion rates and degradation ofthe fuel properties over time exposure.

MATERIALS AND METHODS

Materials and Immersion Test Procedures

Immersion tests were to model storage, transporta-tion and preparation procedures of the biomasses.Chemical composition of the CS and SSs selected forthe tests are given in Table 1. The biomass sources ofrefined RSO, WCO and AWL provided by the MOLNyrt., Hungary, were used without further processing.

Y OF SURFACES Vol. 54 No. 4 2018

726 ANDRÁS GERGELY et al.

Table 1. Chemical composition of the carbon. low and high alloy steels in wt %

Alloys Fe C Cu Si P S Mn Cr Ni Mo N Ti

St35.8 Balance 0.25 ≤0.04 ≤0.04 0.701.4541 Balance ≤0.06 ≤0.75 ≤0.04 ≤0.015 ≤2.0 18.0 10.0 – 0.701.4571 Balance ≤0.07 ≤1.0 ≤0.045 ≤0.015 ≤2.0 17.5 12.50 2.5 0.70

For the immersion tests in neat biomasses, the quanti-ties of 650 g were used from all sources. In each type ofbiomass, two steel rods of two specific compositionswere tested. Preparation of the steel coupons startedwith stain removal by mechanical grinding and polish-ing with silicon carbide emery papers with the meshsizes of 400, 600 and 800 at a rotation rate of 100 perminute by Buehler Metaserv grinder-polisher (Bue-hler UK, Ltd). After mechanical treatment, water wasimmediately removed with ethanol and high pressurenitrogen stream. Steel coupons were ultrasonicated inisopropanol twice for 5 min, dried and measured on ananalytical balance (OHAUS AS-120 Analytical Stan-dard). Coupons were assembled on Teflon rods(diameter of 10 mm and length of 150 mm) over thelength of 100 mm and the rest (50 mm) occupied byTeflon ring spacers (with length of 5 mm, inner andouter diameters of 10.5 and 13 mm). Immersion depthof the specimens at upper end of the Teflon rods wasminimum 20 mm under level of the f luid phases.At the end of the experiments, solid depositions of thepolymerised biomasses were carefully removed thencoupons were ultrasonicated twice for 5 min in isopro-panol and once in acetone. Corrosion spots wereremoved by careful mechanical polishing then ultra-sonication in isopropanol was repeated in line with theASTM G1 principles [37]. Integral corrosion ratesover the 672 hour periods were defined by the weightloss method based on mass changes of the coupons ofa standard size; thickness of 2 mm, outer and innerdiameters of 32 and 10.5 mm, respectively. Weight lossresults were evaluated and transferred into reactionrates of general corrosion (mm year–1) by the equationgiven underneath:

corrosion rate (mm year–1) =

where Δm: the measured weight loss (kg), ρ: density ofthe alloy (estimated as ~7870 kg m–3), A: the geomet-ric surface (1.702 × 10–3 m2) and t: time of the expo-sure (672 h).

In emulsions of the biomasses, citric acid solutionsof 5 and 50 weight % (prepared at 25°C) were mixedwith the biomasses. At a temperature of 80°C and stir-ring rate of 240 rpm, course and fine metastable emul-sions were obtained from the refined RSO and WCO,respectively. Emulsions had to be vigorously stirred toensure macroscopic homogeneity of the liquids anddissipate the larger size of bubbles with diameter of amillimetre to smaller ones. Proportion of the aqueous

Δ × ×ρ

365 1000,

m

At

PROTECTION OF METALS AND PHYSICAL

and organic phases was 20 to 80% by weight. Stagnantand flow conditions were set to model two extremecases of storage and handling.

Degradation of the biomasses was characterised byacidity, unsaturation and water content during the testprocedures of 672 h. TAN, iodine number (IN) andwater content (WC) are given in the appropriate sec-tion. Samples were taken from the neat organic phasesafter segregation of the emulsions then analysis wasperformed on three parallel samples.

ELECTROCHEMICAL TECHNIQUESCell Configuration and Electrodes

Glass cells with an inner volume of a litre werethermostated for the measurements. Cylindrical plati-num gauze with the length of 40 mm and diameter of35 mm was as the counter electrode. With an immer-sion depth of 57 mm, steel rods with a diameter of5 mm and length of 200 mm were tested in the neatbiomasses and in emulsions of the citric acid solutionsof 5 and 50% by weight. To evaluate corrosion resistiv-ity of the steels by direct current methods without anysignificant potential drop in the f luid phase, citric acidsolution (5% by weight) was used as electrolyte for theelectrochemical measurements unaerated underatmospheric pressure. The silver/silver chloride (1 Msodium chloride solution) reference electrode wasplaced in Haber-Luggin type capillary (with outerdiameter of the capillary of 1 mm) pointing towardsaxial end of the working electrodes(steel rods) with adistance of 3 mm. Open circuit, corrosion potentials arereferred to the reference electrode potential at all tem-perature. Prior to the electrochemical measurements,samples were ultrasonicated twice for 2 min in iso-propa-nol to remove traces of the biomass materials.

Cyclic VoltammetryAs current of the electrochemical interface,

pseudo-capacitance is approximately proportional to thevoltage scanning rates at low amplitudes (±20 mV), dou-ble layer capacitance in the citric acid solution vs. biaspotential was evaluated by series of voltammetry mea-surements performed with scanning rates of 800, 600,500, 450, 300, 200, 100, 50 mV s–1 at direct voltagesbetween –430 and +30 mV set to the reference elec-trode. Electrochemical pseudo-capacitance wasdefined by extrapolating to the scanning rate of 1 V s–1.Data provide basis for comparison with the imped-

CHEMISTRY OF SURFACES Vol. 54 No. 4 2018


ance spectroscopy results to which possible distinctionof time constants arisen from the electrochemicaldouble-layer and passive layer of the steel alloys. Thus,these quantities are expectedly more reliably identifiedin the impedance spectra.

Linear Sweep Voltammetry

Polarization resistance (Rp) measurements anddynamic polarizations (linear sweeps) were performedwith a Solartron SI1287 Pstat/Gstat ElectrochemicalInterface (by Solartron Instruments of SolartronGroup Ltd) at open circuit potentials (OCPs) at a scan-ning rate of 0.1667 mV s–1 starting from the cathodicpotential range between the voltage intervals of ±15 mVand from –55 to 105 mV, respectively. Dynamic polar-izations were also carried out by a Radiometer PGZ-301potentiostat (Voltalab instrument, Radiometer Ana-lytical S.A.) at scanning rates of 0.33, 1.0, 3.33 and10 mV s–1 from 200 to 400 mV vs. the reference elec-trode for Tafel analysis to check validity, reliability andreproducibility of the data in the aspects of state of thesteel samples and the electrolytes. Data evaluation wasperformed with the CorrView software (Scribner,Inc.) and some of the results were transferred into gen-eral corrosion rates (mm year–1) by the equation givenin the following:

where j is the instantaneous corrosion current densityin A m–2, and ρ is the density of the alloy (estimatedas ~7870 kg m–3). The M is molar weight of alloysgiven in the following (kg mol–1): 5.54 × 10–2 for theEN 1.4541, and 5.59 × 10–2 for the EN 1.571, in accor-dance with their chemical compositions. As a compari-son with the SSs, the St35.8 CS was also tested. For thecalculations, the molar weight of 5.58 × 10–2 kg mol–1

was used based on the composition of 0.2% Ni and0.6% MN by weight besides the iron balance. The fac-tor of 31536000 is the proportional rate exchange con-stant to transfer from seconds to a year interval. Withtwo molar electron number change, the Faraday con-stant of 96 485 (C mol L –1) was used for the calcula-tions with geometric surfaces of 9.15 × 10–4 m2 and7.58 × 10–4 m2.

Impedance Spectroscopy

Impedance investigation was to evaluate charge-transfer resistance and the coupled mass transportprocesses in relation with rate limitations. Measure-ments were performed by a Radiometer PGZ-301instrument using sinusoid perturbation of 10 mV overthe frequency range between 10 kHz and 0.05 Hz atdifferent bias voltages vs. the reference electrode.

−

× ×=

× ρ

1corrosion rate (mm year ) 31 536 000 1000

,2 96 486

j M


Impedance spectra were fitted and modelled by theZview software (Scribner Inc.).

EXPERIMETNAL RESULTS AND DISCUSSION

Potential Dynamic Measurements vs Temperature

Corrosion potentials of steel electrodes immersedin 5% citric acid solution are presented in Figs. 1a, 1b.Electrode potentials shifted towards the less noblevoltage range with the increasing temperature duringpolarization measurements at all scan rates. Decreas-ing tendency of the voltage shifts was more sensitivemeasured at decreasing scan rates. The magnitude ofpotential shifts with the 1.4541 SS(~200 mV) wasalmost double than that of the 1.4571 (~110 mV),which means oxidation and dissolution of the 1.4541 isthermally more activated. Regarding the mixed poten-tial theory, the phenomenon of decreasing electrodepotential is connected to increasing proportion of sur-face area for metal dissolution at the expense, shrink-age of the cathodic areas. Instantaneous corrosioncurrent densities of the steels derived from the resultsof Tafel investigation are given in Figs. 2a, 2b. In initialphase of the immersion tests, the magnitude of currentdensities was roughly the same with both steel sam-ples. The currents obtained at low voltage scan ratessometimes remained in the range of standard devia-tion of data. Only the ways of temperature dependenceof the corrosion currents were different. The 1.4541exhibited pure exponential acceleration up to 55°Cand above a somewhat lower thermal dependence.The 1.4571 showed steep acceleration in the tempera-ture range of 25–45°C and above that a very mildreaction rate increase at a nearly constant rate. Thisprobably means altered reaction mechanism of thecorrosion processes in relation with at least the ratelimiting step. Temperature profiles of the extrapolatedcurrents are associated with thermally activatedcharge-transfer resistance of the 1.4541 and increasingdissolution rate of passive layer on the 1.4571 with thetemperature. Current-temperature patterns obtainedat medium and low scan rates are accepted as repre-sentative measure of state of the electrodes for the fol-lowing reasons. The currents varied depending onvoltage scan rates changed around ±1 μA cm–2 withboth steels. The lowest current densities measured at3.33 mV s–1 were lower in comparison with the onesmeasured at 1 mV s–1, substantially higher with the1.4541 and a somewhat higher with the 1.4571 overnearly the entire and a partial voltage range, respec-tively. Considering current contribution of the pseudo-capacitive interface to the anodic Faradaic, instanta-neous corrosion currents, this would equal to anodic cur-rent between 2.5 × 10–1 and 2.5 × 10–2 μA cm–2 at thescan rates of 10 and 1 mV s–1, respectively, calculatingwith a double-layer capacitance (DLC) of 25 μF cm–2

on average. Therefore, other processes should impacton increasing the corrosion currents. Thus, the cur-



Fig. 1. Corrosion potential of steel alloys; (a) the EN1.4541 and (b) the EN 1.4571 as a function of temperaturein 5% citric acid solution.

807060504030–560

20

1

Eco

rr, V

–520–500–480–460–440–420–400

–540

–380

Temperature, °C

Scan rates, mV s–1 at

(a)

(b)

3.33310

80706050403020

1

Eco

rr, V

–500

–450–425–400–375–350

–300

–550–525

–475

–325

–275

Temperature, °C

Scan rates, mV s–1 at

3.33310

Fig. 2. Extrapolated corrosion currents of (a) the EN1.4541 and (b) the EN 1.4571 alloys measured in 5% citricacid solution as a function of temperature.

80706050403020

i corr

, µA

cm

–2

102

101

100

Temperature, °C

(a)

(b)

1Scan rates, mV s–1 at

3.33310

80706050403020

i corr

, µA

cm

–2

102

101

100

Temperature, °C


3.33310

rent difference above 1 μA cm–2 at medium and lowscan rates is connected to passivation, growth of pas-sive layers by electrolysis in the solution during the DCinvestigation, which proven to undergo at this currentmagnitude. In this aspect, the 1.4541 should grow pas-sive layers more readily at higher rates and oxidationcurrents than the 1.4571 as steady currents did notincrease with lower voltage scan rates. DLC decreasesand dissolution rates of the alloys increases with thetemperature, since the currents obtained at varied scan


Table 2. Apparent activation energy of corrosion processes ofnius formalism calculated from logarithm of the corrosion cu

SS alloys Temperature range, °C

1.

1.4541 25–45 11.0 ±55–80 3.28 ±

1.4571 25–45 10.7 ±45–80 3.38 ±

rates should converge to one each other with the tem-perature as it was found with the steel electrodes.Apparent activation energies of the corrosion pro-cesses are low and depend quite moderately on scanrates (given in Table 2).

The slopes of anodic currents measured with theSS alloys at various scan rates are presented in Fig. 3.These slopes at any temperature are far from the theo-


the EN 1.4541 and the 1.4571 SS alloys derived by the Arrhe-rrents (expressed in A m–2 s–1) in the temperature ranges

Scan rates, mV s–1

0 3.33 10.0

0.32 11.1 ± 1.2 8.5 ± 0.57 0.01 3.74 ± 1.55 4.75 ± 0.02 2.52 13.4 ± 3.81 10.2 ± 2.12 0.49 3.94 ± 0.34 2.08 ± 0.65


Fig. 3. Anodic Tafel slopes of (a) the EN 1.4541 and (b) theEN 1.4571 obtained in citric acid solution (5 wt %) as afunction of temperature.

807060504030020

104

103

Temperature, °C

(a)

(b)

1Slopes at scan rates, mV s–1 of

3.33310

807060504030102

20

Ba,

mV

Ba,

mV

103

104

Temperature, °C

1Slopes at scan rates, mV s–1 of

3.33310

retically expected ones of active dissolution sincedesorption and dissolution of oxidised species is ratelimiting. Thus, Tafel slopes at the lowest scan ratesreflect high passivation ability of the electrodes in theelectrolyte. Interestingly, local and absolute maxi-mums of the anodic current slopes were in the highertemperature range (80 and 65°C) with the 1.4541 andin the lower temperature region (35 and 45°C) withthe 1.4571. This means higher passivation susceptibil-ity of the SSs at 60 and 40°C in agreement with tem-perature dependence of the corrosion currents. Thelowest anodic slopes increased less steeply with tem-perature at the highest scan rate with both steels.During the faster voltage ramp, the 1.4571 proved to bemore resistive than the 1.4541 against anodic oxida-tion, electrolysis for passivation. This is connected tolower blockage of the Faradaic processes, higher rateof oxidation coupled with greater contribution of thepseudo-capacitive interfaces. So the currents mea-sured at low voltage scan rates are more representativeof the interfacial and diffusion processes rather thanthe charge-transfer. Cathodic Tafel slopes are pre-


sented in Figs. 4a, 4b. Lower dependence of the slopesthan anodic counter parts on the voltage scan rates areinferred as less disturbed cathodic processes. Theslopes converged to –150 mV/decades with both elec-trodes at the medium and low scan rates. Only the highscan rate resulted in slopes with the 1.4541 decreasingabruptly from 55°C and a decreasing continuously onewith the 1.4541 from 25°C. At the temperatures of 65and 80°C, Tafel slopes from –180 to –230 mV/decadeswere obtained at the highest scanning rates. Thisbehaviour hints on higher exchange current at the1.4541 for hydrogen ion discharge. If that is diffusioncurrent than it should be less temperature dependentor in case of charge-transfer control thermal activationshould compensate loss of the cathodic currents con-sumed by the lower amount of dissolved oxygen.Higher sensitivity of the 1.4541 to the oxygen concen-tration in the electrolyte was noticeable. If cathodiccurrents were under the impact of pseudo-capacitiveinterfaces in the electrolyte with medium activity ofthe hydrogen ion then high voltage scan rate shouldhave resulted in less steep slopes but the opposite hap-pened and it means only the Faradaic processes playgoverning role in magnitude of the cathodic currents.As a consequence of mixed depolarization of the SSsby oxygen and hydrogen ion, cathodic slopes variedfrom –240 to –130 mV/decades with the temperature.Cathodic currents should theoretically translate to theslopes of around –150 mV/decades. Reason for the dif-ference is the less efficient supply of oxygen, mass trans-port of the main depolarisator due to lower solubility inthe electrolyte. This confers with the –130 mV/decadeslope of cathodic reduction of hydrogen ion on iron at25°C [38] as it was experienced at low temperaturealthough this process is vastly affected by solute com-position of the electrolyte, e.g., the presence of citricacids and their salt [39]. At 80°C, the theoreticalanodic Tafel slope of 70 mV/decade was far exceededas a consequence of blocking behaviour of the elec-trodes, whereas the cathodic –140 mV/decade wasclosely approached at many scan rates with bothalloys, corresponding to the theoretically expectedslope of cathodic reactions proceeding via the Vol-mer–Tafel mechanism. That would result in a Stern-Geary coefficient of 20 mV which was close to the onesof ~18 mV derived from experimental data. Accordingto the literature [40], corrosion current of iron gener-ally can exceed negative ionisation current of hydro-gen ion by an order of magnitude in weakly acidicsolutions in which discharge process occurs via elec-tro-sorption and subsequent chemical desorption atlow over-voltages and current densities. Differencebetween the two the SS electrodes with mass transportlimited cathodic reactions instead of charge-transfer ispartly attributed to greater nickel and molybdenumcontents of the 1.4571, contributing to electro-sorp-tion of the hydrogen ion [41].

Polarization resistances of the EN 1.4541 and1.4571 (derived from the corrosion currents) changed



Fig. 4. Cathodic Tafel slopes of (a) the EN 1.4541 and (b)the EN 1.4571 obtained in citric acid solution (5 wt %) as afunction of temperature.

807060504030–240

20

–200

–180

–160

–140

–120

–220

–100

Temperature, °C

(a)

(b)

Bc,

mV


3.33310

807060504030–240

20

–200

–180

–160

–140

–220

–120

Temperature, °C

Bc,

mV


3.33310

from ~40 to 7 kΩ cm2 and ~8 to 0.5 kΩ cm2, respec-tively (Figs. 5a, 5b). In comparison, orders of magni-tude lower resistances were measured by the Sternmethod of short voltage scanning around the OCPs,decreasing from 1.6 to 0.7 kΩ cm2 and from 2.6 to 1.3kΩ cm2 with the temperature (Fig. 5c). In this case,the 1.4571 indicated higher resistances than the 1.4541over the entire temperature range, with lower tempera-ture dependence between 55 and 65°C. This is partlydue to the high specific exchange current densitiesaround the corrosion potentials, surpassing the overallresistance affected currents extrapolated from a moredistant Tafel range. The resistance was relatively high(~1 kΩ cm2) at 80°C and this range is related to thecharge-transfer resistance which was mainly affectedby the cathodic half reactions. Charge-transfer resis-tances derived from the impedance spectra are pre-sented in Figs. 6a, 6b. The same order of magnituderesistances were obtained with both SS electrodes asobtained by the Tafel method. The 1.4571 showeddecreasing resistance with the temperature far lessdependent on the impressed DC voltage than that wasexperienced with the 1.4541. The former confers withthe Tafel results and the latter is attributed to a morestable passive state of the 1.4571 (less temperature andvoltage dependent) compared to the 1.4541. In addi-tion, such a difference between the SSs agrees qualita-tively with the resistance differences characterised bythe Stern method. Below 80°C, only the charge-trans-fer resistance locating in the medium-high frequencyrange could be estimated but not the Faradaic nor thepolarization resistances could be assessed due to thelow time constants of the mass transport-coupled pro-cesses inaccessible by the lowest frequency thresholdof the impedance measurement. When an electrodepotential of 0.3 V was impressed at 80°C, Faradaicresistance composed of charge-transfer resistance,resistance of the passive layer, solution resistance anddiffusion impedance was allowed to estimate. None-theless, polarization resistance was not allowed toevaluate by the AC method even at 80°C because ofinstability of the electrodes and the lowest frequencyof 0.01 Hz was not low enough to assess diffusionregime (reaching only around quarter perimeter of thesemi-circles or beginning range of diffusion imped-ance). Only at the temperature of 80°C and DC volt-age of 0.3 V some mass-transport limitation could berecognised besides the low Faradaic impedancesbetween 0.1 and 0.01 Hz. Difference between thepolarization resistances derived from the Tafel analysisand the resistances obtained by the impedance resultsarises by resistance of the passive layer. In case ofeither the Volmer-Heyrovsky or the Volmer-Tafelmechanism there is no simple relationship betweenFaradaic and polarization resistance considering onecathodic half reaction, i.e., discharge of the hydrogenion at the electrode.


Pseudo-capacitive nature of the SS electrodes wasassessed by cyclic voltammetry series and the resultsare given in supplementary data (Fig. A.1). Double-layer capacitance of the 1.4541 was ~65 μF cm–2 and asomewhat lower 60 μF cm–2 (qualitatively the same)was related to the 1.5671 with moderate and increasedtemperature dependence, respectively. These valuesare mainly affected by geometrical inhomogeneousnature of the surface but the surface roughness factorwas unknown. Most minima of the differential capac-itances shifted towards the cathodic range in accor-dance with decreasing corrosion potential of the elec-trodes as a function of temperature. These data were inno agreement with the outcome obtained by potentialstair step series (with DC voltages of 5, 10 and 20 mV)exhibiting inception slopes of logarithm of thecharging currents of around –0.1 V over time. Capac-itances multiplied with the solution resistance asa function of temperature would expectedly result in



Fig. 5. Derived polarization resistances of (a) the EN1.4541 and (b) the EN 1.4571 alloys obtained from the cor-rosion currents by applying the Stern method based on theTafel slopes, (c) measured polarization resistances of steelalloys based on the Stern method, in citric acid solution(5 wt %) as a function of temperature.

80706050403020

1

2

3

80706050403010–1

20Pola

riza

tion

resi

stan

ce/k

ohm

cm

2Po

lari

zatio

n re

sist

ance

/koh

m c

m2

101

100

Temperature, °C

Temperature, °C


3.33310

80706050403010–1

20Pola

riza

tion

resi

stan

ce/k

ohm

cm

2 102

101

100

Temperature, °C


3.33310

Steel alloys:

(a)

(b)

(c)

1.45411.4571

time constants multiplied by a factor of ~2 matchingthe decreasing currents vs. time. So, lower solutionresistance should mean less effectively blocking inter-


face for Faradaic reactions in vicinity of the corrosionpotentials. Thus, the electrodes were probably stableenough for high rate voltage scanning and voltage stepsof small amplitudes up to 55°C with limitations up to80°C. Interface of the SSs is also characterised by theconstant phase elements (coupled parallel with thecharge-transfer resistance). Time constants and expo-nents of the SS electrodes are given in Figure A.2.Greater variation of pseudo-capacity of the 1.4541interface was ascertained with DC voltages and lowertemperature dependence compared to that of the1.4571 in agreement with the results of CV series.

IMPEDANCE SPECTROSCOPY MEASUREMENT WITH TEMPERATUREAdmittances and time constants of the interfaces

were translated to real capacitances over frequenciesby a normalisation program (provided by ZAHNER-Elektrik GmbH and CoKG). The DLCof ~130 μF cm–2

was similar to estimation by the potential stair-stepresults, about doubled of the ones obtained by CVmeasurements. Exponents indicated increasinghomogeneity of the 1.4541 interface with the tempera-ture at all DC potentials, while exponents obtainedwith the 1.4571 were less affected by the temperaturerather the electrode potential. These data were appar-ently not conferring with DLCs results based on tran-sient DC methods, which is attributed to resistiveinhomogeneity of the passive layers, different charge-transfer resistance in lateral distribution, yieldinghigher inhomogeneity factor to the 1.4541 and moder-ate to the 1.4571. This means a more coherent, uni-form and less transforming passive layer, beyond thepossible shunting effect of the charge-transfer resis-tance. Time constants of the Faraday impedance andthe passive layers may not clearly be separated in thespectra as charge-transfer resistance of 1 and 10 kΩ cm2

with an average DLC of 15 uF cm–2 gives 67 and 6.7 Hzare quite similar to time constants of the passive layersof 50 and 5 Hz resulted by the resistances of 104 and105 kΩ cm2 and the capacitance of ~2 μF cm–2. Theexperimentally determined diffusion impedance wasbetween 20 and 100 kΩ cm2. The theoretical charac-teristic time constants for diffusions of oxygen andhydrogen ion in organic acidic aqueous solutions(δ2/D) were around 500 and 34 s over distances of 5and 0.05 cm, respectively. Diffusion coefficient (D =5.4 × 10–9 m2 s–1) [42] and solubility (c = 0.02 mol m–3)[43] of oxygen, whereas the concentration of 30.1 mol m–3

and diffusion coefficient of 7.4 × 10–9 m2 s–1 as Sternestimated for hydrogen ion were used for the calcula-tions. But theoretical Warburg coefficients werearound 100 and 1 Ω cm2 s–0.5 which seemed to be toolow to cause measurable contribution in the frequen-cies of 10 kHz and 10 mHz. Calculated diffusionimpedance of oxygen was 400 Ω cm2 at 10 mHz at adistance of 5 cm, whereas 2 Ω cm2 at 0.5 cm was



Fig. 6. Resistance of (a) the EN 1.4541 and (b) the EN 1.4571 steels as Faradaic impedance derived from impedance spectra mea-sured at various impressed electrode potentials vs the reference electrode in citric acid solution (5 wt %) as a function of tempera-ture.

70 806050403020

–0.1 V

Cha

rge-

tran

sfer

resi

stan

ce, O

hm c

m2

101

100

Temperature, °C

(a)

(b)At bias potentials

vs corrosion potential

OCP0.10.20.3

70 806050403020

–0.1 V

Cha

rge-

tran

sfer

resi

stan

ce, O

hm c

m2

102

101

100

Temperature, °C

At bias potentials vs corrosion potential

OCP0.10.20.3

roughly equal to the diffusion impedance of hydrogenion over distance of 0.05 cm. So, appreciable contri-bution of diffusion impedance of the depolarisators tothe impedance spectra besides charge-transfer resis-tance of the steel rods is not expected over the fre-quency range. The difference between time constantsof mass transport processes in the f luid phase and solidstate seems also not to be clearly separated. The War-burg impedances relating to diffusion of charge-carriersin the solution and the passive layer could be resolved asestimated time constants of depolarisator species in thesolution should appear at 0.05 and 0.005 Hz by diffu-sion lengths of 0.01 and 0.1 cm calculated with a diffu-sion coefficient of 5 × 10–6 cm2 s. Diffusion processes


in the solid state should appear at 0.003 and 0.03 Hzwith diffusion coefficients of 10–16 and 10–15 cm2 s–1

(with high alloy steels) over a distance of 2 × 10–7 cmexpected average thickness of the passive layers,respectively. The former would not be detectable as afeature of a stable electrode being not in equilibriumbut at least in a steady state by the applied investigationfrequency range in the experiment. As time constantsof the processes are closer than a magnitude [44], res-olution of these impedance contributions in the spec-tra would rather be questionable.



Table 3. Weight loss derived corrosion rates of the EN 1.4541 SS immersed in stagnating and stirred phases of neat biomasssources tested at 80oC for 672 h (modelling the circumstances of storage and handling)

Typeof the biomass fuel source Flow condition

Average corrosion rate,μm year–1

STD of corrosion rate, μm year–1

rRSO Stagnant 0.21 0.11WCO 0.39 0.25AWL 0.29 0.07rRSO Stirred 1.28 0.51WCO 1.77 0.29AWL 0.40 0.21

CORROSION RATES BY TIME EXPOSURE

Electrochemical measurement results performedin 5 wt % citric acid solution with the SS alloys testedin stagnating and stirred neat biomasses as a functionof time are presented in Fig. 7. The results of weightloss and electrochemical methods are summarised inTables 4 and 5, respectively. Corrosion potentials andcurrents of both SS alloys, polarization resistanceschanged in the same manner over time and data didnot show any remarkable variation by the biomass fuelsources. OCPs were cathodic by ~150 and 200 mV tothe 1.4541 and the 1.4571, respectively, ~350 mVanodic to the zero-charge potential (–350 mV vs stan-dard hydrogen electrode, SHE). Corrosion currentdensities were lower with both steels in comparison ofthe initial and final data almost regardless of the bio-mass sources. This hints on increased surface area tooxidation and decreasing tendency of dissolution ofthe oxidised metal, resulting in an overall lower corro-sion rates, increasing passivity. The 1.4541 experi-enced higher corrosion rates in the moderate acidicWCO and the more acidic AWL than in refined RSOregardless if the media was stagnating or stirred (Table 3).Appreciable amount of hydrogen ion with variousactivities was available for cathodic half reactions inthe biomasses. Weight loss results unveiled almost anorder of magnitude higher corrosion rates in stirredbiomasses than in stagnating ones except for the resultsobtained in AWL. This is attributed to a better supplyof oxygen as the main depolarisator in the media cou-pled with more intense dissolution of the corrodedmetals in the media through shrunk diffusion layers.The exception of a somewhat higher corrosion rate inthe WL as a result of restrained stream of the f luidphase is explained by the fact the highly viscous mediadid not allow boundary layer to thin sensibly aroundmetal coupons (by better dissipation) compared to theless viscous RSO and WCO streaming at much highrates leaving thinner stagnant boundary layers nearsurface of the coupons and for the depolarisators todiffuse. This seems to be contradicted first by corro-sion test results with plain CS (St35.8) which experi-enced somewhat low rates in RSO and WCO but muchhigher rate in AWL in f lowing f luids. This is explained


by difference between SSs and the plain CS fromwhich the latter is not able to passivate in moderateacidic environments like AWL only in the less acidicRSO and WCO despite the more intense oxygen sup-ply to build up compact passive layers. An order ofmagnitude lower corrosion current densities wasobtained with the 1.4541 in stirred media (Table 4).The highest corrosion rate was obtained in WCO,medium in RSO and the lowest in AWL. The electro-chemically assessed corrosion rate of the 1.4571 wasthe highest in RSO, medium in WCO and the lowestin AWL. Similarly to the 1.4541, the 1.4571 tested instirred rather than stagnant f luids indicated lower cor-rosion rates during the electrochemical characterisa-tion. The remarkably changing patterns of the OCP,current and polarization resistance obtained with theSS electrodes immersed in stagnating biomasses (Fig. 7)became much more stabilised in time when steel rodswere tested in stirred f luids (Fig. 8). OCPs were moreanodic, nobler by ~100 mV and corrosion currentsshifted quite moderately around 10–7 A cm–2 an orderof magnitude lower than measured with samples testedin stagnating media over the entire experiment. BothSS alloys immersed in AWL indicated lower corrosionrates in the characterising citric acid electrolyte andshowed similar time dependent current variation withthe samples tested in RSO and WCO. Regarding thecharacterising method, DC polarization technique inan aqueous electrolyte was able to differentiate state ofthe SSs tested in the biomasses, giving instantaneousreaction rates much closer to the results provided bythe validating type weight loss method when couponstested in stirred media. Corrosion resistance and elec-trochemical behaviour of the SSs tested in stagnatingand stirred f luids under scores the relevance of greatersupply of depolarisator species with the resultantincreased susceptibility of passivation of the SSs in theneat biomasses. Nonetheless, a reasonable explana-tion to the highly different corrosion rates of the 1.4541obtained by the weight loss (Table 3) and electrochem-ical methods performed in different media under stag-nant condition is given in later section.

The results of weight loss and electrochemicaltechniques with the SS coupons immersed in emul-



Fig. 7. Open circuit potential (Eocp), polarization resistance (Rp) and corrosion current (Io) of steel alloys measured in citric acidsolution (5 wt %) as a function of time; (a) the EN 1.4541 and (b) the EN 1.4571 tested in stagnating biomass fuel sources such asrRSO, WCO and AWL.

600500400300200100–0.35

0

Eoc

p, V

vs A

g/A

gCl

700

–0.25

–0.20

–0.15

–0.10

–0.05

–0.30

0 10–4

10–5

10–6

10–7

10–8

105

104

103

Time, h

1.4571 in rRSO1.4571 in WCO1.4571 in AWL

I0I0I0

RpRpRp

I 0, A

cm

–2

Rp,

Ohm

cm

2

600500400300200100–0.35

0

Eoc

p, V

vs A

g/A

gCl

700

–0.25

–0.20

–0.15

–0.10

–0.05

–0.30

0 10–4

10–5

10–6

10–7

10–8

105

104

103

Time, h


I0I0I0

RpRpRp

I 0, A

cm

–2

Rp,

Ohm

cm

2

(a)

(b)

sions of the biomasses with aqueous citric acid solu-tions (5 and 50 wt %) are summarised in Tables 5 and 6,respectively. The weight loss method indicated accept-able low corrosion rates in both types of emulsions(Table 5). The same sort of variation was obtained overtime exposure with both SSs immersed in emulsion ofcitric acid solution (5 wt % ) with RSO and WCO.Only the emulsion of WL led to measurable highercorrosion rates. In emulsions of the essential citricacid solution (50 wt %), both alloys experienced thesame corrosion rates in RSO but they behaved in


opposite ways in the emulsions of WCO and AWL. Incomparison with the mass changes, electrochemicalresults indicated roughly an order of magnitude highercorrosion rates of the SSs (Tables 5 and 6) regardless ofconcentration of the citric acid solutions. Both alloystested in biomass emulsions of the 5% citric acid andthe 1.4541 in emulsion of the 50% solution exhibitedgreater variation of corrosion rates depending on typeof the biomasses, which suggest altered passivationefficiencies of the alloys in the presence of 5% citricacid solution. When essential citric acid solution was



Table 4. Corrosion currents and rates of steel alloys derived from electrochemical measurements performed in 5 w/w% cit-ric acid solution during the immersion test carried out at 80oC for 672 h in neat biomass sources of stagnating and stirredphases (modelling circumstances of storage and handling)

Steelalloys

Type of the biomass source

Flowcondition

Average corrosion rate, μA cm–2

Average corrosion rate, μm year–1

1.4541 rRSO Stagnant 1.12 13.0WCO 1.62 18.8AWL 1.06 12.3

1.4571 rRSO 1.88 21.8WCO 1.40 16.2AWL 0.76 8.78

1.4541 rRSO Stirred 0.14 1.59WCO 0.32 3.75AWL 0.06 0.74

1.4571 rRSO 0.46 5.28WCO 0.38 4.43AWL 0.09 1.06

Table 5. Weight loss derived corrosion rate of steel alloys immersed in f luids of the biomass sources emulsifying aqueousphases (20 w/w%) of citric acid solutions of 5 and 50 w/w%, tested at 80°C for 672 h (modelling the circumstances of puri-fication)

Steel alloys Type of the biomass source

Conc. of citric acid (w/w%) emulsified

in the biomasses


STD of corrosion rate, μm year–1

1.4541 rRSO 5 1.81 0.63WCO 2.69 0.77AWL 3.97 1.88

1.4571 rRSO 1.46 0.21WCO 1.48 0.80AWL 3.00 1.28

1.4541 rRSO 50 4.09 0.91WCO 1.36 0.77AWL 5.64 0.13

1.4571 rRSO 4.14 2.61WCO 2.94 0.62AWL 1.13 0.48

used for the experiment, only the 1.4571 showednearly the same corrosion rates regardless of the bio-masses, which means better passivation susceptibilitythan the 1.4541 regardless of the fuel sources. Electro-chemical results of the SSs tested in emulsions of the 5and 50 wt % citric acid solutions as a function of timeare given in Figs. 9 and 10, respectively. During the testin emulsion of the 5 wt % citric acid solution, OCPsbecame stabilised and corrosion currents decreasedover time (Fig. 9). At the end of the experiment, cor-rosion currents of the SSs were in the same magnitudeat ~10–7 A cm–2. Corrosion potentials of both alloys


shifted around –160 mV which was quite similar to theones obtained with the samples immersed in neatmedia under f lowing condition, around 400 mVnobler than zero-charge potential of the Fe(II) con-taining passive layer. As mass losses became greater inemulsions of the essential citric acid solution, bothalloys indicated the same patterns over time as OCPsdecreased at 260 h which stabilised at –0.2 V after 420 hthen remained stable over the entire period (Fig. 10).Stable polarization resistance of the steels around105 Ω cm2 hints on the formation of compact passivelayers in the emulsions of essential citric acid solu-



Fig. 8. Open circuit potential (Eocp), polarization resistance (Rp) and corrosion current (Io) of steel alloys measured in citric acidsolution (5 wt %) as a function of time; (a) the EN 1.4541 and (b) the EN 1.4571 tested in stirred biomass fuel sources such asrRSO, WCO and AWL.

600500400300200100–0.5

Eoc

p, V

vs A

g/A

gCl

700

–0.3

–0.2

–0.1

–0.4

0

10–6

10–7

10–8

105

104

Time, h


E0E0E0

RpRpRp

I 0, A

cm

–2

Rp,

Ohm

cm

2

600500400300200100–0.5

Eoc

p, V

vs A

g/A

gCl

700

–0.3

–0.2

–0.1

–0.4

0

10–6

10–7

10–8

105

104

Time, h


E0E0E0

RpRpRp

I 0, A

cm

–2

Rp,

Ohm

cm

2

(a)

(b)

tion as compared to the experiments performed withdilute aqueous solution (polarization resistances of~6 × 104 Ω cm2), in spite of the dissolution relatedhigher mass losses. This is similar to the differenceobtained by immersion tests with the biomasses underchanged flow condition. The excessive corrosion ratesassessed by the electrochemical methods are con-nected to the larger real electroactive surfaces than thegeometrical ones and much greater activity of depolar-isators in the characterising aqueous solution than inemulsion of the biomasses. Dissolution rate of themetal oxides and hydroxides are more intense under


f lowing condition and better supply of the electro-chemically active organic species besides increasingacidity and unsaturation of the biomasses.

As a consequence of f lowing of the biomasses, the1.4541 experienced changes of corrosion rates by 7, 6and 0.8 times in RSO, WCO and AWL, respectively.The less passivating type St35.8 CS showed oppositetendency of changes of the weight losses 0.3, 1 and8 times in RSO, WCO and AWL, respectively as aresult of f lowing f luids. Comparison of weight lossresults of the CS and SSs in stagnant neat biomasses,the St35.8 indicated 3, 2 and 100 times of corrosion



Table 6. Corrosion currents and rates of steel alloys derived from electrochemical measurements performed in 5 w/w% cit-ric acid solution during the immersion test carried out at 80°C for 672 h in neat biomass sources of stagnating and stirredphases (modelling circumstances of storage and handling)

Steel alloys Type of the biomass source

Conc. of citric acid (w/w%) emulsified

in the biomasses

Average corrosion current, μA cm–2


1.4541 rRSO 5 1.14 13.2WCO 0.62 7.23AWL 1.02 11.8

1.4571 rRSO 1.19 13.8WCO 0.79 9.24AWL 0.55 6.32

1.4541 rRSO 50 1.10 12.8WCO 0.56 6.45AWL 0.72 8.30

1.4571 rRSO 0.66 7.70WCO 0.61 7.12AWL 0.65 7.51

rates of the 1.4541 in RSO, WCO and AWL, respec-tively. In the f lowing biomasses, corrosion rates of theCS changed to 0.1, 0.5 and 200 times of the 1.4541. CSis more sensitive to acidic species of the biomassesthan the SSs and its corrosion rate was under masstransport limitation, diffusion flux of oxygen. For oxygentransport, the corrosion current of 4.25 × 10–14 A cm–2 s–1

is maintained when diffusion coefficient of 1.1 ×10‒13 cm2 s–1 over diffusion length of 1 cm apply. If dif-fusion length is calculated in accordance with the cor-rosion rates than the lengths would decrease to unreal-istic 2.5 × 10–5 or even 1.1 × 10–7 mm. So, majority ofthe anodic currents behind weight losses are related tocathodic discharge of hydrogen ion. If surplus corro-sion currents correspond to hydrogen evolution andcalculated with diffusion coefficient of 7.4 × 10–5 cm2

s–1 and length of 0.05 cm as originally proposed byStern then medium ion activity, the pH of ~4 equiva-lent in aqueous solutions should be enough for the1.4541 and it would change to the pH of ~1.5 for theSt35.8 CS in a highly acidic environment. The corro-sion currents measured in citric acid solution were oneand two orders of magnitude higher than hypotheticalcurrent densities equivalent of the mass losses. Ifcathodic limiting currents of oxygen and hydrogen arecalculated over the lengths of 3 and 0.05 cm, diffu-sion coefficient of 2 × 10–5 cm s–1 [45], solubility ofoxygen at 5.56 × 10–5 mol dm–3 and activity of hydro-gen ion at pH 2.75, then all electrochemically charac-terised corrosion currents correspond to weight lossresults in neat and emulsions of the biomasses, exceptfor the results obtained with 1.4541 immersed in stag-nant RSO and WCO. Corrosion rate of the 1.4541 isclearly limited by dissolution of the passive layer which


was more intense in RSO and WCO but decreasing inAWL. It can only be interpreted as development of amore compact, dense passivating oxide layer whichwas allowed by higher f lux of oxygen and the oxygencontaining potential nucleophiles besides unchangedthickness of the Prandtl type boundary layer. The lat-ter must have been unaffected in vicinity of the cou-pons in AWL as a result of its high dynamic viscositywhile supporting better oxygen supply and other depo-larisators constituting partly the passive layer enoughto counterbalance f low condition during the experi-ment. Passivation of the St35.8 CS by the higher oxy-gen flux was also experienced in the less acidic RSOand WCO but f low of the AWL caused increasedweight loss reflecting lower passivation ability. Whenthe 5% citric acid solution was emulsified with the bio-masses, corrosion rates of the 1.4541 increased by 1.7,1.6 and 30 times in RSO, WCO and AWL, respectively.In emulsions of the 50% citric acid solution, weightlosses increased by 3.4, 0.8 and 3.3 times in RSO,WCO and AWL, respectively. Interestingly, 10 times ofcorrosion rate increases of the 1.4541 were obtainedwith the biomasses from RSO, WCO to the AWLemulsions with 5% citric acid solution. In contrary, 10times decrease of corrosion was found when the 50%citric acid solution was used for the immersion test. Asa comparison, the 1.4571 SS indicated the same ten-dency of weight losses with the biomasses andincreased corrosion rates than the 1.4541 by 2.7 and2.1 times in WCO and AWL, respectively.



Fig. 9. Open circuit potential (Eocp), polarization resistance (Rp) and corrosion current (Io) of steel alloys measured in citric acidsolution (5 wt %) as a function of time; (a) the EN 1.4541 and (b) the EN 1.4571 tested in stirred citric acid (5 wt %) emulsions(20 wt.%) of the biomass fuel sources such as rRSO, WCO and AWL.

600500400300200100–0.25

010–7

Eoc

p, V

vs A

g/A

gCl

700

–0.15

–0.10

–0.05

–0.20

0 10–5

10–6

105

104

Time, h


I0I0I0

RpRpRp

I 0, A

cm

–2

Rp,

Ohm

cm

2

600500400300200100–0.25

0

Eoc

p, V

vs A

g/A

gCl

700

–0.15

–0.10

–0.05

–0.20

0 10–5

10–6

10–7

105

104

Time, h


I0I0I0

RpRpRp

I 0, A

cm

–2

Rp,

Ohm

cm

2

(a)

(b)

DEGRADATION OF BIOMASSESAND CHANGES OF ELECTROCHEMICAL

INTERFACES BY TIME EXPOSURE

Regarding complexity of integrity loss and passiva-tion of steels in the biomasses, besides acidity of thematerial overall characteristics and aging of the bio-masses catalysed by the citric acid solutions shouldplay a dominant role in causing corrosion of SSs. Inaddition, the lowest weight losses in emulsions of theAWL can only be inferred by the assessment of passiv-ation susceptibility and state of passive layers of the


SSs. In the meantime of immersion tests of CS and theSSs, biomasses indicated decreasing water contents inall experiments and increasing TAN in most of theexperiments as a function of time. Except for the stag-nating biomasses, the IN characterised unsaturationof the biomasses decreased over time. The results aresummarised in the supplementary data (Fig. A. 3–6).The higher relative amount of water is assumed toassist passivation of the St35.8 CS in stirred f luids,whereas increasing TAN played no decisive role inchanging the corrosion rates in RSO and WCO unlikein the AWL which caused considerable integrity losses.



Fig. 10. Open circuit potential (Eocp), polarization resistance (Rp) and corrosion current (Io) of steel alloys measured in citric acidsolution (5 wt %) as a function of time; (a) the EN 1.4541 and (b) the EN 1.4571 tested in stirred citric acid (50 wt %) emulsions(20 wt %) of the biomass fuel sources such as rRSO, WCO and AWL.

600500400300200100

Eoc

p, V

vs A

g/A

gCl

700

–0.25

–0.20

–0.15

–0.30

–0.10 10–5

10–6

10–7

10–8

105

103

104

106

Time, h


I0I0I0

RpRpRp

I 0, A

cm

–2

Rp,

Ohm

cm

2

600500400300200100

Eoc

p, V

vs A

g/A

gCl

700

–0.25

–0.20

–0.15

–0.30

–0.10 10–5

10–6

10–7

10–8

105

103

104

106

Time, h


I0I0I0

RpRpRp

I 0, A

cm

–2

Rp,

Ohm

cm

2

(a)

(b)

So, higher ability of passivation of the 35.8 CS is partlyrelated to the increased water contents of the f luidsunder more of an unchanged acidity of the environment.Besides steady TAN in neat biomasses, increased corro-sion rate of the 1.4541 SS occurred along with increas-ing water contents of the RSO and WCO, whereassteady water content and the nearly constant TAN ofthe AWL underwent with almost invariant corrosionrates over time. Increasing acidity of biomasses of theemulsions with dilute and essential citric acid solu-


tions seems first correlates with increased corrosionrates of the SSs.

Decreasing electrochemical pseudo-capacity char-acterised as DLC of the 1.4541 SS electrodes mea-sured in 5 wt % citric acid solution during the immer-sion test reflects increasing uniformity of the interfacewhich depended on the f luids. In stagnant neat bio-masses (contributing to a less effective passivation ofthe SSs), well discernible decrease of the capacitancereflecting transformation of the surface occurredaround 350 h of immersion experiment (Fig. 11a). In



Table 7. Polarization resistance of the EN 1.4541 and the 1.4571 SS electrodes (tested in stagnant and flowing neat bio-masses) calculated from the average corrosion currents over time by the Stern-Geary coefficients, B = 0.016 and 0.021 Vused for the SS alloys respectively based on average of the measurement results obtained at 80°C. These data are comparedto resistance of the passive layer developed in a reversible process appraised by the Stern formula (1), allowing the estima-tion of charge-transfer resistance

Stainless steels Biomass sources ConditionPolarization resistance,kΩ cm–2

Resistanceof the passive layer,

kΩ cm–2

Charge-transfer resistance,kΩ cm–2

1.4541 rRSO Stagnant 14.29 13.58 0.701.4541 WCO 9.88 9.39 0.481.4541 AWL 15.09 14.35 0.741.4571 rRSO 11.17 8.09 3.081.4571 WCO 15.00 10.87 4.131.4571 AWL 27.74 20.10 7.64

1.4541 rRSO Stirred 116.8 111.1 5.731.4541 WCO 49.54 47.11 2.431.4541 AWL 251.9 239.6 12.361.4571 rRSO 46.15 33.44 12.711.4571 WCO 54.97 39.83 15.141.4571 AWL 230.5 167.0 63.50

relation with passivation, this process did not undergowith either stabilisation of the OCPs (Fig. 9) orincreased polarization resistance. So, geometricaluniformity of the interface should have occurred viadissolution of the more active sites without develop-ment of compact passive layer as a consequence oflimited transport of oxygen and other potential nucle-ophiles. Except for minor differences at the beginningof the immersion test, surface of the 1.4571 electrodesbehaved in similar ways with the 1.4541. Faster andmore pronounced transformation of the surfaceshould have taken place in stirred neat biomasses asDLCs decreased markedly within the first 250 h(Fig. 11b) to the extent over 648 hours by testing instagnant f luids. The geometric type transition wentwith stabilisation of the OCPs (Fig. 10) and significantincrease of the polarization resistance (Table 7) whichunderline the formation of compact passive layer.When dilute citric acid solution was emulsified in thebiomasses, then decreasing capacity of the interfacewas only measured after 350 hours (Fig. 12a) parallelwith the slightly or moderately increasing polarizationresistance (Table 8) reflecting slow and less effectivepassive layer formation. Almost the same behaviourand impact of processes were manifested as in thestagnant biomasses. Testing with biomass emulsions ofthe essential citric acid solution lead to well decreasingcapacity of the SS interfaces in the first 250 hours andthe final value of ~14 μF cm–2in the end (Fig. 12b).This is similar to that was obtained in stirred bio-masses. The polarization resistance of the SS elec-trodes was a bit higher and less scattering (Table 8)denoting moderate assist to the development of com-


pact passive layers. Compactness and thickness ofthese layers are assessed by overall sum of resistance ofthe passive layers and charge-transfer resistance inseries. The latter was appraised by subtraction of thepassive layer resistance (estimated by the Stern for-mula (1)) from the entire resistance assessed by theaverage corrosion currents and the Stern-Geary coef-ficient. The results are summarised in Tables 7 and 8.Estimated resistance of the passive layers formed instirred biomasses especially in AWL were considerablyhigher than the ones developed in stagnant f luids. Thisis the reason that the corrosion rates derived from thecurrents obtained by electrochemical methods indilute citric acid solution were closer to the validatingweight loss data. The same kind of difference wasobserved with immersion test results performed withbiomass emulsions. The high cathodic current densitybased on major contribution of hydrogen ion dis-charge owes to the extrapolated excessive assessmentof corrosion rates (by the Tafel method) in compari-son with the weight loss data. The electrochemicallymeasured decreased corrosion currents correlated wellwith an order of magnitude increase of the charge-transfer resistance for both SSs and the multipliedresistance of the passive layers.

(1)

In other aspects, the resistance assessed by theStern method of low voltage amplitude scanning ismore related to charge-transfer resistance rather thanresistance of the passive layers, whereas Tafel methodproved to be better suited forgiving information on

=corr

.

RTRnF j



Fig. 11. Double-layer capacitance derived from low amplitude cyclic voltammetry series (anodic and cathodic branches) mea-sured at 80°C with the EN 1.4541 steel rods in citric acid solution (5 wt %) in the course of immersion tests performed in (a) stag-nating and (b) stirred neat biomass fuel sources.

600 70050040030020010010

0Time, h

20

25

30

35

15

40

(a)

(b)

Pseu

do c

apac

itanc

e, µ

F c

m–

2

Anodic and

Anodic and

Anodic and

Cathodic in rRSO

Cathodic in AWL

Cathodic in WCO

600 7005004003002001000Time, h

30

40

50

60

20

70

Pseu

do c

apac

itanc

e, µ

F c

m–

2

Anodic and

Anodic and

Anodic and

Cathodic in rRSO

Cathodic in AWL

Cathodic in WCO

state, compactness of the passive oxide layers. Basedon the corrosion currents, anodic and cathodic Tafelslopes, net anodic currents were estimated by theMcLaurin expression (2), quadratic formula of theButler-Volmer equation. If an order of magnitude dif-ference between the corrosion and net anodic currentis assumed so as the changes between the electro-chemical and the weight loss based data, then esti-mated cathodic potential shift, an overvoltage of‒0.125 V from the steady electrode potential should beenough. That means electrode potential of the SSs stillremains in the ennobled, anodic potential range incomparison with the SHE.


(2)

When the equation (3) proposed by Stern was usedfor evaluation, nearly the same cathodic polarization(–0.124 V) would result in an order of magnitudeexcess of cathodic currents in relation with the hydro-gen ion accelerated corrosion of the electrodes as ananalogy with discharging hydrogen ion which causeddifferences between the hypothetical corrosion cur-rents of immersion tests and the electrochemicallyderived data.

( )= × +

⎛ ⎞+ +⎜ ⎟⎝ ⎠

net corr

22 2

1 12.303 η

1 11.1 2 .5 η

j jba bc

ba bc



Fig. 12. Double-layer capacitance derived from low amplitude cyclic voltammetry series (anodic and cathodic branches) mea-sured at 80°C with the EN 1.4541 steel rods in citric acid solution (5 wt %) in the course of immersion tests performed with stirredemulsions of the biomass fuel sources mixed at 20 wt % with citric acid solutions of (a) 5 and (b) 50 wt %.

(a)

(b)

600 7005004003002001000Time, h

16

20

22

26

12

14

18

24

28

Pseu

do c

apac

itanc

e, µ

F c

m–

2

Anodic and

Anodic and

Anodic and

Cathodic in rRSO

Cathodic in AWL

Cathodic in WCO

600 7005004003002001000Time, h

20

25

30

35

40

45

55

15

55

Pseu

do c

apac

itanc

e, µ

F c

m–

2

Anodic and

Anodic and

Anodic and

Cathodic in rRSO

Cathodic in AWL

Cathodic in WCO

(3)

When the equation (4) is used for the estimation,then cathodic polarization of –0.152 V is obtainedwhich is commensurable with the experimentallyobtained cathodic polarization of iron by –0.172 V (vsSHE) corroding in deaerated 0.1 M citric acid solution[46].

(4)

By estimating f luxes of the main depolarisator, i.e.,the oxygen flux to the electrodes in both in the stag-nant and flowing biomassess should well exceed (cor-

⎛ ⎞Δ = − × + ⎜ ⎟⎝ ⎠

corrcorr 10

0

log .

jRTE pH BnF j

( ) corrcorr

0

.

jRTEnF j

⎛ ⎞Δ = ⎜ ⎟⎝ ⎠


responding to 9 × 10–3 and 9 × 10–2 and A cm2 over thedistances of 5 and 0.5 cm) the corrosion rates, oranodic currents in relation with mass losses of the SSs.So, corrosion of the SSs did reach the maximumattainable cathodic currents of the main depolarisatorflux. It is not the redox reaction but partly solubility ofthe corroded species limited further corrosion pro-cesses. Only the plain CS showed far greater rate ofcorrosion (corresponding to 1.9 × 102 A cm–2) in theacidic AWL (with the highest content of active hydro-gen ions and redox active species) than theoreticaloxygen flux would suggest. Such a difference becamemuch greater when neat biomasses were stirred (massloss rate equalling to an instantaneous current of 2.5 ×103 A cm–2).



Table 8. Polarization resistance of the EN 1.4541 and the 1.4571 SS electrodes (tested in f lowing biomass emulsions withdilute and essential citric acid solutions) calculated from the average corrosion currents over time by the Stern-Geary coef-ficients, B = 0.016 and 0.021 V used for the SS alloys respectively based on average of the measurement results obtained at80°C. These data are compared to resistance of the passive layer developed in a reversible process appraised by the Sternformula (1), allowing the estimation of charge-transfer resistance

Stainless steels Biomass sourcesConcentration

of the citric acid solutions

Polarization resistance,kΩ cm–2

Resistanceof the passive layer,

kΩ cm–2

Charge-transfer resistance,kΩ cm–2

1.4541 rRSO 5 14.04 13.35 0.691.4541 WCO 25.68 24.42 1.261.4541 AWL 15.69 14.92 0.771.4571 rRSO 17.65 12.79 4.861.4571 WCO 26.35 19.09 7.261.4571 AWL 38.53 27.92 10.61

1.4541 rRSO 50 14.55 13.83 0.711.4541 WCO 28.78 27.37 1.411.4541 AWL 22.35 21.25 1.101.4571 rRSO 31.63 22.91 8.711.4571 WCO 34.20 24.78 9.421.4571 AWL 32.41 23.48 8.93

CONCLUSIONS

The St35.8 CS passivates better in the less acidicRSO and WCO due to increased f lux of oxygen andmoderate f low rate of the f luids. Transition from stag-nant to mediocre f low of the acidic AWL results in upto two orders of magnitude higher corrosion ratewhich is inferred by unable passivation of the CS andincreased rate of cathodic discharge of hydrogen iondonated by active species. In all biomasses, corrosionrate of CS is limited by reduction rate of the hydrogenion (mass transport of the corrosive organic speciesfeaturing high oxidising redox potential) which con-sumes majority of the anodic corrosion currentbesides reduction of oxygen.

Integrity loss of the EN 1.4541 and 1.4571 SSs isprimarily limited by mass transport of oxygen and dis-solution rate of the passive layers in the less acidicRSO and WCO. Corrosion of the SS sensibly intensi-fies by f low of the biomasses by thinning of the bound-ary layer. The f lowing AWL with higher contents ofoxygen donating species and potential nucleophilesprovides increased supply of oxygen for the formationof denser, more compact passive layers. Thus, lowercorrosion rates of the EN 1.4541 and 1.4571 SSs wereobtained under stagnating condition, whereas closermatches of the integral and instantaneous corrosionrates evaluated by the weight loss and assessed by elec-trochemical methods, respectively.

Biomass fuel sources indicated decreasing watercontents and increasing acid number as a function oftime during most of the experiments. Except for thestagnating neat biomasses, the iodine number charac-


terised unsaturation decreased over time. The higherrelative amount of water was found to assist passiva-tion of CS under f lowing condition, whereas theincreasing acidity played probably no decisive role inchanging corrosion rates in RSO and WCO unlike inAWL which contributed to considerable integrity loss.Better passivation ability of the St35.8 CS is partlyrelated to increased water contents along with more ofan unchanged acidity of the f luids. Decreasing watercontent and steady acidity of neat biomasses resultedin less affective passivation of the SSs in stagnant f lu-ids. In the f lowing biomasses with better oxygen sup-ply, increasing water contents occur with better passiv-ation of the SSs even though it goes with higher corro-sion losses as a consequence of faster dissolution of thepassive layer. In f lowing emulsion of the biomassesfeaturing decreasing water content and increasingacidity over time, increased concentration of the citricacid solutions leading to altered corrosion rates isexplained by higher mass transport of oxygen andpotential nucleophiles, oxygen donating species.

ACKNOWLEDGMENTS

This work was supported by the GINOP-2.3.2-15-2016-00053 project with title of “Development ofengine fuels with high hydrogen content in theirmolecular structures (contribution to sustainablemobility)”. Authors are thankful to Tóth Csaba, whoworked at the Department of MOL Hydrocarbon andCoal Processing, Institute of Chemical and ProcessEngineering, the University of Pannonia, for his vari-ety of contributions to the experimental work. Authors



are grateful to Balázs Margit (Bay Zoltán NonprofitLtd. for Applied Research) for lending the analyticalresults of the biomass fuel sources.

REFERENCES1. Craig, W.K. and Soveran, D.W., US Patent 4992605,

1991.2. Stumborg, M., Wong, A., and Hogan, E.N., Bioresour.

Technol., 1996, vol. 56, no. 1, pp. 13–18.3. Monnier, J., Tourigny, G., Soveran, D.W., Wong, A.,

Hogan, E.N., and Stumborg, M., US Patent 5705722,1998.

4. Klass, D.L., Biomass for Renewable Energy, Fuels andChemicals, Encyclopaedia of Energy, Oxford: ElsevierScience, 2004, pp. 1–20.

5. Knothe, G., van Gerpen, J., Krahl, J., The BiodieselHandbook, Champaign, IL: AOCS Press, 2005.

6. Franklin, M.J. and White, D.C., Curr. Opin. Biotech-nol., 1991, vol. 2, no. 3, pp. 450–456.

7. Klokova, I.V., Klimova, T.A., Emelyanov, L.V., andKrylov, I.F., Chem. Technol. Fuels Oils, 2005, vol. 41,no. 4, pp. 319–322.

8. Deshmukh, G.M., Randive, S.N., and Shekh, M.M.,Chem. Eng. World, 2003, vol. 38, pp. 75–76.

9. Knoblauch, U.L. and Heitz, E., Electrochim. Acta,1987, vol. 32, no. 6, pp. 901–907.

10. Foulkes, F.R., Kalia, R.K., and Kirk, D.W., Can. J.Chem. Eng., 1980, vol. 58, pp. 654–659.

11. Paramonov, V.A. and Filatova, N.G., Prot. Met., 2004,vol. 40, no. 1, pp. 82–84.

12. Cavalcanti, E., Wanderley, V.G., Miranda, T.R.V., andUller, L., Electrochim. Acta, 1987, vol. 32, no. 6,pp. 935–937.

13. Weidig, C.F., US Patent 4448586, 1984.14. Perilstein, W.L., US Patent 4521219, 1985.15. de Souza, J.P., Mattos, O.R., Sathler, L., and Take-

nout, H., Corros. Sci., 1987, vol. 27, no. 12, pp. 1351–1364.

16. Sekine, I., Yuasa, M., and Kohara, K., Corros. Sci.,1990, vol. 31, pp. 579–584.

17. Trasatti, S.P., Sivieri, E., and Mazza, F., in Trends inElectrochemistry and Corrosion at the Beginning of the21st Century, Brillas, E. and Cabot P., Eds., Barcelona:Universitat de Barcelona, 2004, pp. 1083–1095.

18. Sridhar, N., Price, K., Buckingham, J., and Dante, J.,Corrosion, 2006, vol. 62, no. 8, pp. 687–702.

19. Jiang, Y., Wu, Y., and Wang, K., Mater. Corros., 2006,vol. 57, no. 5, pp. 418–421.

20. Nie, X., Li, X., and Northwood, D., Mater. Sci. Forum,2007, vols. 546–549, pp. 1093–1100.

21. Yahagi, Y. and Mizutani, Y., Wear, 1984, vol. 97, no. 1,pp. 17–25.

22. Fazal, M.A., Haseeb, A.S.M.A., and Masjuki, H.H.,Renewable Sustainable Energy Rev., 2011, vol. 15, no. 2,pp. 1314–1324.

23. Kaul, S., Saxena, R.C., Kumar, A., Negi, M.S., Bhat-nagar, A.K., Goyal, A.B., and Gupta, A.K., Fuel Pro-cess. Technol., 2007, vol. 88, pp. 303–307.

24. Knothe, G. and Dunn, R.O., J. Am. Oil Chem. Soc.,2003, vol. 80, no. 10, pp. 1021–1026.

25. Knothe, G., J. Am. Oil Chem. Soc., 2006, vol. 83,no. 10, pp. 823–833.

26. Grabowski, M.T. and McCornick, R.L., Prog. EnergyCombust. Sci., 1998, vol. 24, no. 2, pp. 125–164.

27. Cvengros, J., Fett/Lipid, 1998, vol. 100, no. 2, pp. 41–44.

28. Ballote-Díaz, L., López-Sansores, J.F., Maldonado-Lopez, L., and Garfias-Mesias, L.F., Electrochem.Commun., 2009, vol. 11, no. 1, pp. 41–44.

29. Passman, F.J., Int. Biodeterior. Biodegrad., 2013,vol. 81, pp. 88–104.

30. Kaul, S., Saxena, R.C., Kumar, A., Negi, M.S., Bhat-nagar, A.K., Goyal, H.B., and Gupta, A.K., Fuel Pro-cess. Technol., 2007, vol. 88, pp. 303–307.

31. Geller, D.P., Adams, T.T., Goodrum, J.W., and Pend-ergrass, J., Fuel, 2010, vol. 89, no. 3, pp. 792–796.

32. Haseeb, A.S.M.A., Fazal, M.A., Jahirul, M.I., andMasjuki, H.H., Fuel, 2011, vol. 90, no. 3, pp. 922–931.

33. Singh, B., Korstad, J., and Sharma, Y., Renewable Sus-tainable Energy Rev., 2012, vol. 16, no. 5, pp. 3401–3408.

34. Norouzi, S., Eslami, F., Wyszynski, M.L., and Tso-lakis, A., Fuel Process. Technol., 2012, vol. 104,pp. 204–210.

35. Fazal, M.A., Haseeb, A.S.M.A., and Masjuki, H.H.,Fuel Process. Technol., 2010, vol. 91, pp. 1308–1315.

36. Melero, J.A., Calleja, G., Garcia, A., Clavero, M.,Hernandez, E.A., Miravalles, R., and Galindo, T.,Fuel, 2010, vol. 89, no. 3, pp. 554–562.

37. ASTM G1-03: Standard Practice for Preparing, Clean-ing, and Evaluating Corrosion Test Specimens, WestConshohocken, PA: ASTM Int., 2011.

38. Miyata, Y., Kuwahara, Y., Asakura, S., Shinohara, T.,Yakou, T., and Shiimoto, K., Int. J. Corros. ProcessesCorros. Control, 2014, vol. 49, no. 4, pp. 269–274.

39. Beck, W. and Fischer, P., Corros. Sci., 1975, vol. 15,nos. 6–12, pp. 757–766.

40. Dafft, E.G., Bohnenkamp, K., and Engell, H.J., Cor-ros. Sci., 1979, vol. 19, no. 7, pp. 591–612.

41. Huang, Y., Nielsen, R.J., Goddard, W.A., andSoriaga, M.P., J. Am. Chem. Soc., 2015, vol. 137,no. 20, pp. 6692–6698.

42. Han, P. and Bartels, D.M., J. Phys. Chem., 1996,vol. 100, no. 13, pp. 5597–5602.

43. Mulazzani, Q.G., Sun, H., Hoffman, M.Z., Ford, W.E.,and Rodgers, M.A.J., J. Phys. Chem., 1983, vol. 12,pp. 163–178.

44. Walter, G.W., J. Electroanal. Chem., 1981, vol. 118,pp. 259–273.

45. Jamnongwonga, M., Loubiered, K., Dietrich, N., andHébrard, G., Chem. Eng. J., 2010, vol. 165, no. 2,pp. 758–768.

46. Revie, R.W. and Uhlig, H.H., Corrosion and CorrosionControl: An Introduction to Corrosion Science and Engi-neering, Hoboken, NJ, John Wiley and Sons, 2008.

PROTECTION OF METALS AND PHYSICAL CHEMISTRY OF SURFACES Vol. 54 No. 4 2018

corrosion rates of stainless steels in renewable biofuel ......the materials is critical for...

Documents