8
* F. X. Perrin Laboratoire de Chimie Applique ´e-Universite ´ de Toulon et du Var, BP132, 83957 La Garde Cedex (France) H. Girard, J. Pagetti Laboratoire de Corrosion et Traitments de Surfaces, Universite ´ de Franche-Comte ´, 32 rue Me ´gevand, 25030 Besanc ¸on Cedex (France) G. Daufin INRA-Laboratoire de Recherches de Technologie Laitie `re, 65 rue St Brieuc, 35042 Rennes Cedex (France) Aspects of the effects of temperature and electrolyte composition on pitting corrosion of stainless steel in dairy fluids Aspekte des Einflusses von Temperatur und Elektrolytzusammensetzung auf die Lochkorrosion von nichtrostenden Sta ¨ hlen in Molkereiflu ¨ ssigkeiten F. X. Perrin*, H. Girard, G. Daufin and J. Pagetti The pitting corrosion resistance of 304L stainless steel in dairy fluids (milks, wheys, soya juice and peptidic fluids) was studied using electrochemical measurements. The effects of temperature, chloride content and other components of the fluids was particularly investigated. In the range 30 – 70 8C, the pitting potential in whole milk E p is related to the temperature by the relation ln(E p 100) aT 1 b. Above 70 8C, a further phenomenon adds to the common activation effect of temperature. Heat induced conformational changes (denaturation) of the proteins were believed to explain such a behaviour. A typical linear relationship was found between E p and the logarithm of chloride concentration. All fluids are well represented by a single relationship. Therefore, the buffering capa- city of casein micelles in milks do not significantly change the pit- ting resistance of the oxide film. In dairy industry, the corrosion risk is usually estimated from the difference between the pitting poten- tial and the potential of a gold electrode (E g ). It is noteworthy that the pitting risk decreases when temperature increases in the tem- perature range 50 – 90 8C. Such a trend was due to the strong de- crease in dissolved oxygen above 50 8C. Besides, in aggressive peptidic solutions, the resistance of the passive film to localized attack is directly related to the Cr, Mo and N alloy content of stainless steel. Das Lochkorrosionsverhalten des nichtrostenden Stahles 1.4404 in Molkereiflu ¨ssigkeiten (Milch, Molke, Sojasaft und peptischen Flu ¨ssigkeiten) wurde mittels elektrochemischer Messungen unter- sucht. Im wesentlichen wurden die Einflu ¨sse von Temperatur, Chlo- ridgehalt und anderer Komponenten der Flu ¨ssigkeiten untersucht. Im Bereich von 30 – 70 8C ist das Lochkorrosionspotential in Voll- milch E p mit der Temperatur nach der Beziehung ln(E p 100) aT 1 b gekoppelt. U ¨ ber 70 8C kommt zu dem u ¨blichen Aktivie- rungseffekt der Temperatur ein weiteres Pha ¨nomen hinzu. Es wird angenommen, dass wa ¨rmeinduzierte Anpassungsa ¨nderungen (De- naturierung) der Proteine dieses Verhalten erkla ¨ren ko ¨nnen. Zwi- schen E p und dem Logarithmus der Chloridkonzentration wurde eine typische lineare Beziehung gefunden. Alle Flu ¨ssigkeiten wer- den gut durch eine einzige Beziehung wiedergegeben. Das bedeu- tet, dass die Pufferkapazita ¨t der Kaseinmicellen in der Milch den Lochkorrosionswiderstand des Oxidfilms nicht nennenswert vera ¨n- dert. In der Molkereiindustrie wird das Korrosionsrisiko u ¨blicher- weise durch die Differenz zwischen dem Lochkorrosionspotential und dem Potential einer Goldelektrode (E g ) abgescha ¨tzt. Es ist be- merkenswert, dass das Lochkorrosionsrisiko im Temperaturbereich 50 – 90 8C mit steigender Temperatur abnimmt. Dieser Trend ist auf die starke Abnahme des gelo ¨sten Sauerstoffs oberhalb von 50 8C zuru ¨ckzufu ¨hren. Außerdem ist in aggressiven peptischen Lo ¨sungen der Wider- stand des Passivfilms gegenu ¨ber lokalem Angriff direkt gekoppelt mit dem Gehalt des nichtrostenden Stahles an den Legierungsele- menten Cr, Mo und N. 1 Introduction Food processing equipment is subject to strict monitoring regarding the materials and the conditions under which they are used. To guarantee the constant production of good quality food that will not be altered by denaturation or contamination subsequent to contact with the materials constituting the equipment, a regular monitoring of their reactions is neces- sary. In the dairy industry, corrosion generally results from the misuse of equipment (bad maintenance, unsuitable disin- fectant...) or from ignorance of its potential sources. Stainless steel is by far the mostly used material in dairy industry owing to its mechanical properties, easy cleanup and biocompatibil- ity. Corrosion of stainless steel food processing equipment is usually localised owing to the composition of the medium [1]. Pitting corrosion of stainless steel occurs when a critical po- tential called pitting potential, E p , is exceeded. Oxidising strength of the medium, pH and contact time of stainless steel with the solution determine the value of the gold potential E g , which characterises the metal-solution interface in the med- ium. Then, E g is theoretically the maximum potential that must be reached by any metal and, particularly, by stainless Materials and Corrosion 52, 629–636 (2001) Pitting corrosion of stainless steel in dairy fluids 629 Ó WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001 0947-5117/01/0808-0629$17.50.50/0

Aspects of the effects of temperature and electrolyte composition on pitting corrosion of stainless steel in dairy fluids

Embed Size (px)

Citation preview

Page 1: Aspects of the effects of temperature and electrolyte composition on pitting corrosion of stainless steel in dairy fluids

* F. X. PerrinLaboratoire de Chimie AppliqueÂe-Universite de Toulon et duVar,BP132, 83957 La Garde Cedex (France)

H. Girard, J. PagettiLaboratoire de Corrosion et Traitments de Surfaces,Universite de Franche-ComteÂ,32 rue MeÂgevand, 25030 BesancËon Cedex (France)

G. DaufinINRA-Laboratoire de Recherches de Technologie LaitieÁre,65 rue St Brieuc, 35042 Rennes Cedex (France)

Aspects of the effects of temperature andelectrolyte composition on pitting corrosion ofstainless steel in dairy fluids

Aspekte des Einflusses von Temperatur und Elektrolytzusammensetzung auf dieLochkorrosion von nichtrostenden StaÈ hlen in MolkereifluÈ ssigkeiten

F. X. Perrin*, H. Girard, G. Daufin and J. Pagetti

The pitting corrosion resistance of 304L stainless steel in dairyfluids (milks, wheys, soya juice and peptidic fluids) was studiedusing electrochemical measurements. The effects of temperature,chloride content and other components of the fluids was particularlyinvestigated. In the range 30 ± 70 8C, the pitting potential in wholemilk Ep is related to the temperature by the relation ln(Ep � 100) �aTÿ1� b. Above 70 8C, a further phenomenon adds to the commonactivation effect of temperature. Heat induced conformationalchanges (denaturation) of the proteins were believed to explainsuch a behaviour. A typical linear relationship was found betweenEp and the logarithm of chloride concentration. All fluids are wellrepresented by a single relationship. Therefore, the buffering capa-city of casein micelles in milks do not significantly change the pit-ting resistance of the oxide film. In dairy industry, the corrosion riskis usually estimated from the difference between the pitting poten-tial and the potential of a gold electrode (Eg). It is noteworthy thatthe pitting risk decreases when temperature increases in the tem-perature range 50 ± 90 8C. Such a trend was due to the strong de-crease in dissolved oxygen above 50 8C.

Besides, in aggressive peptidic solutions, the resistance of thepassive film to localized attack is directly related to the Cr, Moand N alloy content of stainless steel.

Das Lochkorrosionsverhalten des nichtrostenden Stahles 1.4404in MolkereifluÈssigkeiten (Milch, Molke, Sojasaft und peptischenFluÈssigkeiten) wurde mittels elektrochemischer Messungen unter-sucht. Im wesentlichen wurden die EinfluÈsse von Temperatur, Chlo-ridgehalt und anderer Komponenten der FluÈssigkeiten untersucht.Im Bereich von 30 ± 70 8C ist das Lochkorrosionspotential in Voll-milch Ep mit der Temperatur nach der Beziehung ln(Ep � 100) �aTÿ1 � b gekoppelt. UÈ ber 70 8C kommt zu dem uÈblichen Aktivie-rungseffekt der Temperatur ein weiteres PhaÈnomen hinzu. Es wirdangenommen, dass waÈrmeinduzierte AnpassungsaÈnderungen (De-naturierung) der Proteine dieses Verhalten erklaÈren koÈnnen. Zwi-schen Ep und dem Logarithmus der Chloridkonzentration wurdeeine typische lineare Beziehung gefunden. Alle FluÈssigkeiten wer-den gut durch eine einzige Beziehung wiedergegeben. Das bedeu-tet, dass die PufferkapazitaÈt der Kaseinmicellen in der Milch denLochkorrosionswiderstand des Oxidfilms nicht nennenswert veraÈn-dert. In der Molkereiindustrie wird das Korrosionsrisiko uÈblicher-weise durch die Differenz zwischen dem Lochkorrosionspotentialund dem Potential einer Goldelektrode (Eg) abgeschaÈtzt. Es ist be-merkenswert, dass das Lochkorrosionsrisiko im Temperaturbereich50 ± 90 8C mit steigender Temperatur abnimmt. Dieser Trend ist aufdie starke Abnahme des geloÈsten Sauerstoffs oberhalb von 50 8CzuruÈckzufuÈhren.

Auûerdem ist in aggressiven peptischen LoÈsungen der Wider-stand des Passivfilms gegenuÈber lokalem Angriff direkt gekoppeltmit dem Gehalt des nichtrostenden Stahles an den Legierungsele-menten Cr, Mo und N.

1 Introduction

Food processing equipment is subject to strict monitoringregarding the materials and the conditions under which theyare used. To guarantee the constant production of good quality

food that will not be altered by denaturation or contaminationsubsequent to contact with the materials constituting theequipment, a regular monitoring of their reactions is neces-sary.

In the dairy industry, corrosion generally results from themisuse of equipment (bad maintenance, unsuitable disin-fectant. . .) or from ignorance of its potential sources. Stainlesssteel is by far the mostly used material in dairy industry owingto its mechanical properties, easy cleanup and biocompatibil-ity. Corrosion of stainless steel food processing equipment isusually localised owing to the composition of the medium [1].Pitting corrosion of stainless steel occurs when a critical po-tential called pitting potential, Ep, is exceeded. Oxidisingstrength of the medium, pH and contact time of stainless steelwith the solution determine the value of the gold potential Eg,which characterises the metal-solution interface in the med-ium. Then, Eg is theoretically the maximum potential thatmust be reached by any metal and, particularly, by stainless

Materials and Corrosion 52, 629±636 (2001) Pitting corrosion of stainless steel in dairy fluids 629

Ó WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001 0947-5117/01/0808-0629$17.50�.50/0

Page 2: Aspects of the effects of temperature and electrolyte composition on pitting corrosion of stainless steel in dairy fluids

steel in the same electrolyte. If Eg > Ep, pitting corrosion iscertain. Conversely, the larger the difference (Ep ÿ Eg), themore improbable the occurrence of pitting. The stainless steelfree potential Ef in the medium is obviously always inferior tothe gold potential Eg. Thus, the difference (Epÿ Eg) quantifiesthe corrosion risk in a more severe way than the difference (Epÿ Ef) does [2].

The aim of this work was (i) to classify the corrosion resis-tance of a reference 304L stainless steel in several food fluids(milks, wheys and soya juice), (ii) to identify both the role ofthe fluid components (chloride ions, proteins. . .) and other ex-perimental factors (mainly pH, temperature and dissolvedoxygen) on pitting susceptibility.

2 Experimental methods

2.1 Stainless steel

The chemical composition of 304L, 316L and 316Ti stain-less steels are given in Table 1. The plate samples (100 � 50mm2) were passivated in nitric acid (65%) at room tempera-ture, then mechanically polished with a grade 1000 siliconcarbide paper, air dried and finally immersed into the corro-sive fluid. Only the two faces of the samples were polished,

leaving the edge passivated to reduce effects [4]. The stainlesssteel surface was polished with a relatively coarse paper toinduce roughness. Indeed, breakdown of the protective filmand consequently the formation and propagation of pits is ea-sier on a rough surface compared with a smooth one [5]. Then,the selected surface preparation should induce more severeconditions to pitting than cold-rolled or electropolished ma-terials.

2.2 The dairy fluids

The chemical composition of the media studied namelydairy and peptidic fluids is described in Tables 2 and 3, respec-tively. Whole and skim milks and sweet and acid wheys werereconstituted from powder adding milli-Q water to give arange of 100 to 500 g of dry matter per litre. Analyses of dairyfluid components were carried out as follows: total dry matter,proteins, fat, lactose, ash, chloride ions [6]. Owing to theirhigh chloride content and low pH values (between 4.7 and5.1), peptidic fluids are more corrosive than dairy fluids.The more corrosive peptidic fluids have only been used inthe initial study concerning the influence of material typeon pitting potential. Besides, a model whey was preparedby successive addition of chloride ions, inorganic ions, lactose

Table 1. Types and compositions of stainless steels (% w/w)

Tabelle 1. Typen und Zusammensetzungen der nichtrostenden StaÈhle

Element 304L 316L 316Ti

Iron 71.39 64.79 66.65Chromium 17.12 17.14 16.77Nickel 9.08 12.89 11.02Manganese 1.59 1.40 1.56Silicon 0.37 0.53 0.37Cobalt 0.13 0.21 0.35Sulfur 0.01 0.001 0.001Carbon 0.03 0.03 0.04Phosphorus 0.03 0.02 0.02Molybdenum 0.07 2.64 2.11Copper 0.09 0.15 0.62Wolfram 0.04 0.06 0.09Vanadium 0.04 0.11 0.08Titanium 0.03 0.27Aluminium 0.004 0.03Pitting Index (PI)8(a) 17.35 25.85 23.73

(a) PI � Cr � 3.3 Mo � 16 N [6]

Table 2. Composition of raw dairy products (g � kgÿ1 (a) or g � Lÿ1(b))

Tabelle 2. Zusammensetzung der rohen Molkereiprodukte (g � kgÿ1 (a) oder g � Lÿ1(b))

Whole Milk(a) Skim Milk(a) Sweet Whey(a) Acid Whey(a) Soya juice(b)

Total dry matter 950 951 953 941 126Proteins 254 356 126 103 22Fat 287.3 4.9 5 ± 5.2Lactose 353 508 791 768 ±Sucrose ± ± ± ± 81.6Ash 56 55 75 110 2.7Chloride 8 10.1 16.3 21 0.2pH(25 8C)(c) 6.6 (100) 6.6 (100) 6.4 (60) 5.0 (60) 6.3 (126)

(a) powder(b) dry matter (g � kgÿ1) is indicated in brackets

630 Perrin, Girard, Daufin and Pagetti Materials and Corrosion 52, 629±636 (2001)

Page 3: Aspects of the effects of temperature and electrolyte composition on pitting corrosion of stainless steel in dairy fluids

and, finally, whey proteins, respectively. A model skim milkfluid was also prepared following the same procedure andadding native phosphocaseinate in a later stage. The prepara-tion of model solutions is fully described elsewhere [7].

2.3 Electrochemical study

2.3.1 Electrodes

The working electrode was stainless steel as previously de-scribed; the free area in contact with the solution was 60 cm2.

The reference electrode was a silver/ silver chloride (SSC)electrode (Ingold), the potential of which is � 0.197 V/SHE(Standard Hydrogen Electrode). In this work, all potentialswere referred to the SSC.

The counter-electrode was a titanium grid.The gold electrode used to assess the oxidising strength of

the medium had a 4 mm2 working area. For all experimentsthe electrode was mechanically polished (grade 1000), rinsedin deionised water and air dried.

2.3.2 Apparatus and procedure

Electrochemical measurements were performed with stir-ring in a standard three-electrode cell that allows temperaturecontrol (polystat type I Bioblock Scientific). The experimen-tal setup included a Tacussel potentiostat (PRT 10-0.5L) as-sisted by a Tacussel signal generator (Servovit 13). E �f(t) and I � f(E) curves were recorded on a Sefram recorder(X-Y-t).

The general procedure was to monitor the stainless steelcorrosion potential Ef over a 4 h period (Ef � f(t) curves).The potentiodynamic polarisation was subsequently appliedat a rate of 250 mV � hÿ1 starting from ÿ 600 mV (SSC).The pitting potential, Ep, was reported as the potential atwhich an anodic current of 100 lA � cmÿ2 was exceeded.The slow scan rate favours the natural and controlled forma-tion of a protective oxide film. Consequently, breakdown po-tentials should be reproducible (of course within the limits ofthe hazardous character of the pitting event) and not far fromthe breakdown potentials in real service conditions.

Besides, the gold potential Eg was also recorded over a 4 hperiod (Eg � f(t) curves).

Each experiment was repeated six times to take into ac-count the statistical character of the pitting phenomenonthat involve dispersion of pitting potential values [8].

3 Results and discussion

3.1 Material type: an initial survey

Preliminary potentiodynamic runs carried out in peptidicfluids with different material types (304L, 316L and 316Tistainless steels) reveal that the pitting potential significantlyincreases with the pitting index (PI) [3] of the material (Fig. 1).Such a behaviour is a common trend of stainless steels ex-posed to chloride solutions.

In stainless steels, the areas immediately adjacent to man-ganese sulfide inclusions make ideal pit initiation sites. Tita-nium is believed to act as a sulfur trapping element that pre-vents the formation of manganese sulfide inclusions [9]. Then,the lower susceptibility to pitting corrosion of 316L comparedto 316Ti shows that the low sulfur content (0.001% w/w)likely makes the titanium presence in 316Ti useless as faras pitting corrosion behaviour is concerned. Therefore, the pit-ting corrosion of stainless steel exposed to food fluids in con-ditions of our study is mostly directed by the pitting index ofthe material equipment. In the next experiments, only the304L stainless steel material, one amongst the mostly usedmaterial in dairy industry, has been investigated.

Fig. 1. Variation in pitting potential as function of pitting index(PI) in peptidic fluids at 65 8C; fluid codes are specified in brackets

Abb. 1. Variation des Lochfraûpotentials als Funktion des Pitting-index (PI) in peptischen FluÈssigkeiten bei 65 8C; die FluÈssigkeits-codes sind in Klammern angegeben

Table 3. Composition of peptidic fluids (g � Lÿ1)

Tabelle 3. Zusammensetzung der peptischen FluÈssigkeiten (g � Lÿ1)

PEP1(a) PEP2(a) PEP3(a)

Total dry matter 420 300 400Total nitrogen content 30 22 13Sodium chloride 170 ± 180 150 ± 165 170 ± 185Acetic acid ± ± 6Citric acid ± 14 14pH (at 65 8C) 5.1 4.8 4.7

(a) Fluid codes

Materials and Corrosion 52, 629±636 (2001) Pitting corrosion of stainless steel in dairy fluids 631

Page 4: Aspects of the effects of temperature and electrolyte composition on pitting corrosion of stainless steel in dairy fluids

3.2 Gold potential (Eg) and stainless steel free potential(Ef)

Whatever the food fluid, the potential-time characteristicsshow the same pattern (Fig. 2). After a fast decrease, Eg and Efincrease in a relatively short time (15 ± 20 min) and finallyslightly decrease to reach a relatively constant value beyondtwo hours of immersion. The same pattern in potential-timecurve of gold and stainless steel might suggest that the passiveoxide film formation on stainless steel contributes poorly to Efevolution. Then, on both electrodes, the change in metal-elec-trolyte interface would be controlled by the components indairy fluids. The adsorption of proteins on different metalshas previously been reported by several authors [10± 12].The latter phenomenon likely explains the evolution of Efand Eg towards more anodic values that follows the initialbreakdown of the native passive film.

The stainless steel free potential Ef was checked to be sys-tematically inferior to Eg in each experiment (Eg ÿ Ef � 200mV). Eg values after 4 h immersion has been systematicallyreported. Firstly, the low discrepancy between Eg values mea-sured in the same conditions must be noted (DEg < 20 mV).This is likely due to the fact that, contrary to pitting, the goldpotential is not subjected to a statistical process. As expected,the change in chloride concentration in the range 0 ± 11 g � Lÿ1

does not influence the gold potential of a given fluid (Fig. 3).Consequently, differences (Ep ÿEg) for a given fluid will bemainly controlled by the decrease in Ep when chloride concen-tration increases (see 3.3). Nevertheless, the dairy fluids givesignificant differences in gold potential. For example, caseinsand the higher pH of milks contribute to a lower Eg thanwheys. The lower pH of acid whey compared to sweetwhey (DpH � 1.4 unit) might explain the higher Eg valuein the former fluid (DEg � 50 mV) in satisfactory agreementwith the Nernst law. Besides, the strongly higher concentra-tion of fat in whole milk compared to skim milk might con-tribute to the lower Eg values (DEg� 50 mV). Variations of Egas a function of temperature for whole milk (Fig. 4) showsmall changes up to 50 8C. Above 50 8C and more dramati-cally above 70 8C, Eg values decrease when temperature in-

creases (DEg � 180 mV between 70 8C and 90 8C). Contraryto the pitting potential change above 70 8C, heat induced de-naturation of proteins can not explain such a strong decrease inEg. A more likely explanation would be the change in the oxy-gen content at higher temperature. Indeed, as the temperatureis increased, the Henry's law constant decreases meanwhilethe water pressure increases. That cause a decrease in dis-solved oxygen. At 90 8C the level of available oxygen atthe cathode sites is low (< 2.5 � 10ÿ5 mol � Lÿ1). The presenceof salts in dairy fluids adds to the latter phenomenon sinceoxygen is usually less soluble in salts solutions comparedto pure water solutions. Then, the strong decrease of Eg be-tween 70 8C and 90 8C is ascribed to negligible content of dis-solved oxygen at 90 8C.

Fig. 2. Variation of the free potential of stainless steel (Ef) and gold(Eg) with exposure time in whole milk (10% w/v total solids) at40 8CAbb. 2. Variation des Freien Korrosionspotentials von nichtrosten-dem Stahl (Ef) und Gold (Eg) in AbhaÈngigkeit von der Auslage-rungszeit in Vollmilch (10% w/v Gesamtfeststoff) bei 40 8C

Fig. 3. Variation of the gold potential Eg with chloride concentra-tion for food fluids at 40 8CAbb. 3. Variation des Goldpotentials Eg mit der Chloridkonzentra-tion fuÈr LebensmittelfluÈssigkeiten bei 40 8C

Fig. 4. Variation of gold potential Eg and pH with temperature forwhole milk (10% w/v total solids) ; error bars for pH denote max-imum and minimum values

Abb. 4. Variation des Goldpotenzials Eg und des pH-Wertes mitder Temperatur fuÈr Vollmilch (10% w/v Gesamtfeststoff); die Feh-lerbalken fuÈr den pH-Wert kennzeichnen die maximalen und mini-malen Werte

632 Perrin, Girard, Daufin and Pagetti Materials and Corrosion 52, 629±636 (2001)

Page 5: Aspects of the effects of temperature and electrolyte composition on pitting corrosion of stainless steel in dairy fluids

3.3 Pitting potential of 304L stainless steel in dairy fluids

3.3.1 Effect of chloride ions

Fig. 5 shows that the pitting potential fits the law estab-lished by Leckie et al. in 1966 [13] and confirmed later byseveral authors [14 ±16]: Ep � c log[Clÿ] � d

Thus, chloride ions in dairy fluids play an active role in thelocalised breakdown of the passive film. Correlation coeffi-cients r2 is satisfactory (r2 � 0.92). Values of Ep calculatedfrom model equations did not differ by more than 60 mVfrom experimental values while experimental deviationmay be as large as 70 mV. Interestingly, only one group ofexperimental results, hence including all the studied fluids,is found. Therefore, the decrease in Ep is mainly due to theincrease in chloride concentration, whatever the fluid. Conse-quently, the buffering capacities of casein micelles in milks donot result in a slower acidification rate in pitting nuclei at anygiven chloride concentration as compared to wheys or soyajuice. Values of slopes c are usually reported to be nearfrom ÿ120 mV at room temperature. In this study, the calcu-lated slope stands for the same range of values. However, thefact slope c is somewhat inferior to ÿ 120 mV might suggestthat chloride ions in dairy fluids have even more dramatic ef-fects on pitting potentials than in more standard electrolytes.

Besides, the pitting potential remains unchanged (maxi-mum variation below 30 mV) when inorganic ions, lactoseand proteins are successively added to the chloride ions solu-tion during the preparation of model fluids. This result con-firms that components other than chloride ions in dairy fluidshave low effect on the pitting process. For example, the com-parison between skim milk and whole milk results (Fig. 5)shows that fat has little influence on Ep.

3.3.2 Effect of temperature

The 304L pitting potential Ep of whole milk (10% w/v so-lids) decreased by about 10 mV in the temperature range 30 to70 8C and by about 50 mV in the range 70 to 90 8C. The sametrend was observed with the other fluids. The decrease of pit-ting potential for austenitic stainless steels with increasingtemperature has been reported to be a common trend [17 ±18]. Indeed, an increase in temperature alters the propertiesof passive layers [18]. Both the rate of chemisorption and dif-fusion process in passive layers of chloride and metal ions in-crease with temperature, hence promoting the initiation andpropagation of pits.

In order to account for the temperature effect, a linear re-lationship ln(Ep� 100)� aTÿ1� b can be found. For example,with whole milk (10% w/v solids) in the range 30 to 90 8C, theequation was ln(Ep � 100) � 411 Tÿ1 � 4.4 (Ep in mV(SSC)and T in K) with a correlation coefficient r2 of 0.83 (Fig. 6).The low r2 value clearly shows that the variations in Ep are notonly due to the direct activation effect of temperature. Con-versely, the linear fitting is much better when only the tem-perature range 30 ± 70 8C is considered (r2 � 0.98). In thatcase, the equation was ln(Ep � 100) � 197 Tÿ1 � 5.1 (Epin mV(SCC) and T in K). Then, temperatures above 70 8Clikely involve modifications of the electrolyte that overall re-duce the breakdown potential of the passive film, hence add-ing to the activation effect of temperature.

As a general rule, the higher a, the higher the influence oftemperature on Ep and the higher b, the higher the influence ofparameters other than temperature on Ep. Figs. 7 and 8 show

that the influence of temperature on Ep increases with the drymatter content (and obviously with the chloride content) whilethe influence of parameters other than temperature decreases.With whole and skim milks, the values of a and b are similarthroughout the dry matter range. Then, fat does not signifi-cantly contribute to the variations in the pitting potential Epwith temperature. With acid and sweet wheys, the values

Fig. 5. Variation of pitting potential values Ep with chloride con-centration logarithm for food fluids at 40 8C; Ep�ÿ 150 log[Clÿ]�197 (r2 � 0.92)

Abb. 5. Variation des Lochfraûpotentials Ep mit dem Logarithmusder Chloridkonzentration fuÈr LebensmittelfluÈssigkeiten bei 40 8C,Ep � ÿ 150 log[Clÿ]� 197 (r2 � 0.92)

Fig. 6. Effect of temperature on the pitting potential Ep for wholemilk (10% total solids); the continuous line stands for the linearfitting in the temperature range 30 ± 90 8C while the dotted linestands for the linear fitting in the temperature range 30 ± 70 8C;pH values are indicated in brackets (mean values)

Abb. 6. Einfluss der Temperatur auf das Lochfraûpotential Ep fuÈrVollmilch (10% Gesamtfeststoff); die durchgezogene Linie stehtfuÈr die lineare Anpassung im Temperaturbereich 30 ± 90 8C, waÈh-rend die gestrichelte Linie die lineare Anpassung im Temperatur-bereich 30 ± 70 8C wiedergibt; die pH-Werte sind in Klammernangegeben (Mittelwerte)

Materials and Corrosion 52, 629±636 (2001) Pitting corrosion of stainless steel in dairy fluids 633

Page 6: Aspects of the effects of temperature and electrolyte composition on pitting corrosion of stainless steel in dairy fluids

of a and b are different from the values for milks and variedmore with dry matter content, particularly for acid whey.

Coming back to the strong decrease in pitting potentialabove 70 8C, it is unlikely that heat release of chloride ionsfrom micelles or other components of the fluids would explainsuch a behaviour. Indeed, in milk, chloride ions are known tobe totally soluble [19]. Then, the available content of freechloride ions should not vary with temperature.

Heat denaturation of proteins is the main change sufferedby the dairy fluids at temperatures above 50 ±60 8C. Heat de-naturation may release SH groups and other sulfur compounds[20]. However, most denaturation processes consist ofchanges in secondary bonds (ion-dipole, hydrogen and Vander Waals bonds) and in the rotational positions about singlebonds which are controlled by the secondary bond structure[19]. The reaction between whey proteins (mostly b-lactoglo-buline) and casein micelles also occur upon heating, hencecausing their irreversible denaturation [21]. The main proteinsin dairy products are caseins, b-lactoglobuline and a-lactalbu-mine. Although their isoelectric point (IP) notably depends onionic strength and content in other components of the fluids, inmilk, it is about 4.65, 5.23± 5.30 and 4.3, respectively [19]. Inthis study, pH values of solutions are superior to the latter IPvalues. Therefore, electrostatic attractions between the anodi-cally polarized metal and proteins likely promote the forma-tion of a strongly adsorbed layer of proteins at the metal-elec-trolyte interface. When temperature increases, the bulk pH(see explanation below) and, consequently, the excess in ne-gative charges of proteins decreases. Then, the increase intemperature should result in a less strongly adsorbed organiclayer at the metal. However, the latter process must not ex-plain the strong decrease of Ep above 70 8C since, if signifi-cant, it would preferably induce a continuous decrease of Epthrough the whole temperature range. The soluble proteins(whey proteins) of milk mostly consist of b-lactoglobulineand a-lactalbumine. They are much more sensitive to heating

as compared to caseins [19]. From ellipsometric measure-ments, Hegg et al. [22] studied the change in thickness ofthe adsorbed layer of proteins with temperature. For a-lactal-bumine, the thickness remains unchanged in the temperaturerange 25 ± 70 8C (few ten Angstroms) but dramatically in-creases up to 200 ± 300 Angstroms when temperature reaches90 8C. Other proteins of milk (caseins and b-lactoglobuline)showed the same trend as above. Therefore, we suggest thatthe passive film undergoes heat-induced conformationalchanges that gives thicker film but more porous and, thereforeless protective. This would explain the peculiar behaviour ob-served above 70 8C.

Besides, the precipitation of calcium phosphates and therupture of POH-Ca bridged structures in micellar caseinhave been reported by several authors [23]. In wheys, calciumions have a greater tendency to precipitate as calcium phos-phates when temperature is increased than in milks [24]. Then,their contribution to adsorption on cathodic sites of the metalprobably decreases when temperature increases in wheys.That might at least partly explain the stronger influence oftemperature on Ep values in wheys compared to milks (Fig. 7).

So far, pH effects have not been totally considered. Actu-ally, pH decreases both with temperature (Fig. 4) and with drymatter increase (Fig. 7). As the temperature is raised, the fol-lowing acid-base equilibria shift to the right, hence decreasingpH of the solution [19]:

HCit2ÿ $ Cit3ÿ � H�

H2POÿ4 $ HPO2ÿ4 � H�

3Ca2� � 2HPO2ÿ4 $ Ca3�PO4�2 � 2H�

Moreover, lactose decomposes at high temperatures to giveformic acid and, consequently, a decrease in pH [19]. The lowpH values of acid whey solutions at a 50% dry matter content(pH is inferior to 4 at 90 8C), must be responsible for its pe-culiar behaviour respective to the other fluids. Actually, theprecipitation of calcium phosphate in wheys contribute tothe decrease in pH while the casein micelles in milks enhancethe pH buffering capacity and, consequently, stabilise the pas-sive state. Moreover, as it was already noted, owing to theiramphiprotic nature, the adsorption of proteins on metals isstrongly affected by the pH of the bulk electrolyte [10].The low pH values of acid whey solutions at a 50% dry mattercontent might be responsible for a peculiar structural organi-sation of soluble proteins at the metal-electrolyte interface(namely nature of the interactions, thickness of the adsorbedlayer, fraction of uncovered metal. . .). Anyway, the results ofFig. 7 suggest that an increase in temperature would havemuch more detrimental effects on such a layer as comparedto layers formed in other conditions (either lower dry mattercontent or other fluids).

Finally, further experiments showed that pH effects are notresponsible for the stronger decrease of pitting potential withtemperature above 70 8C. Indeed, the pitting potential of awhole milk model solution at 40 8C slightly increases whenpH value is artificially decreased from 6.60 to 5.05 by additionof a 1N citric acid solution. It is noteworthy that pH value inwhole milk does not decrease below 5.2 even at 90 8C. Thatwould confirm that the peculiar behaviour observed withwhole milk above 70 8C is mostly due to an heat-induced con-formational change of the adsorbed protein layer.

Fig. 7. Values of a coefficients (in K) from equations ln (Ep� 100)� aTÿ1 � b (Ep in mV/SSC and T in K) obtained after linear fittingin the temperature range 30 ± 90 8C as a function of dry matter con-tent and food fluids; extreme pH values are indicated in brackets;dotted lines are only a guide for the eye

Abb. 7. Werte fuÈr den Koeffizienten a (in K) aus den Gleichungenln (Ep � 100) � aTÿ1 � b (Ep in mV/SSC und T in K), die nachlinearer Anpassung im Temperaturbereich 30 ± 90 8C als Funktiondes Trockenmasseanteils und der LebensmittelfluÈssigkeiten ermit-telt wurden; extreme pH-Werte sind in Klammern angegeben; diegepunkteten Linien sind nur ein visuelles Hilfsmittel

634 Perrin, Girard, Daufin and Pagetti Materials and Corrosion 52, 629±636 (2001)

Page 7: Aspects of the effects of temperature and electrolyte composition on pitting corrosion of stainless steel in dairy fluids

3.4 Assessment of corrosion risks

3.4.1 Effect of chloride content and other components

As previously noted, the differences (Epÿ Eg) and Ep showthe same variations respective to the chloride ions concentra-tion (Fig. 9). However, the aggressive chloride anions are notthe only compounds that govern the pitting corrosion of 304Lstainless steel in dairy fluids. Actually, Fig. 9 gives the follow-ing classification according to increasing risk to pitting corro-sion at a given chloride ion content:

Whole milk < sweet whey � skim milk < acid wheySoya juice is not included in the above classification since

only one chloride concentration was available for this fluid.The same classification is found when considering the dairy

fluids compositions in real service conditions (chloride ionscontent somewhat different). The lower aggressiveness to-wards pitting corrosion of whole milk compared to skimmilk and sweet whey can be ascribed to the presence of fatin the former fluid that decreases the stainless steel free po-tential towards more negative values.

When the concentration of chloride ions exceeds 6 g � Lÿ1,(Epÿ Eg) in acid whey is negative, which can be considered asa clue to potential pitting risk. Nevertheless, (Ep ÿ Ef) stillremain positive even in these worst conditions. That wouldexplain why no pitting was observed on unpolarized 304L la-boratory samples.

3.3.2 Effect of temperature

The difference (Ep ÿ Eg) slightly decreases in the range30 ±50 8C when temperature increases (Fig. 10). Therefore,in that temperature range, the pitting risk increases with thetemperature. The activation effect of temperature controlsthe pitting risk. Conversely, the difference (Epÿ Eg) increasesat temperatures above 50 8C with temperature and mostly atthe maximum tested temperature (90 8C). Then, Ep is not theonly criterion that determines pitting resistance and, above50 8C, Eg variations that are mainly due to decrease in dis-solved oxygen weigh more than Ep variations.

4 Conclusion

From electrochemical results obtained in this study, it isclear that chloride ions is the major component that controlsthe localized attack of stainless steel in dairy fluids. Neverthe-less, other factors among which the presence of fat or caseinsand the pH values of the fluids were found to contribute to asignificant extent in the pitting corrosion risk. Besides, thechange in the corrosion risk with temperature reveals the ex-istence of a critical temperature near from 50 8C. Indeed, be-

Fig. 8. Values of b coefficients from equations ln (Ep � 100) �aTÿ1 � b (Ep in mV/SSC and T in K) obtained after linear fittingin the temperature range 30 ± 90 8C as a function of dry matter con-tent and food fluids; extreme pH values are indicated in brackets;dotted lines are only a guide for the eye

Abb. 8. Werte fuÈr den Koeffizienten b (in K) aus den Gleichungenln (Ep � 100) � aTÿ1 � b (Ep in mV/SSC und T in K), die nachlinearer Anpassung im Temperaturbereich 30 ± 90 8C als Funktiondes Trockenmasseanteils und der LebensmittelfluÈssigkeiten ermit-telt wurden; extreme pH-Werte sind in Klammern angegeben; diegepunkteten Linien sind nur ein visuelles Hilfsmittel

Fig. 9. Variation of (Ep ÿ Eg) with chloride concentration for foodfluids at 40 8C; dotted lines are only a guide for the eye

Abb. 9. Variation von (Ep ÿ Eg) mit der Chloridkonzentration fuÈrLebensmittelfluÈssigkeiten bei 40 8C; die gepunkteten Linien sindnur ein visuelles Hilfsmittel

Fig. 10. Variation of (Ep ÿ Eg) with temperature for whole milk(10% w/v total solids)

Abb. 10. Variation von (Epÿ Eg) mit der Temperatur fuÈr Vollmilch(10% w/v Gesamtfeststoff)

Materials and Corrosion 52, 629±636 (2001) Pitting corrosion of stainless steel in dairy fluids 635

Page 8: Aspects of the effects of temperature and electrolyte composition on pitting corrosion of stainless steel in dairy fluids

low 50 8C, the activation effect of temperature controls thecorrosion risk while, above 50 8C, the decrease in dissolvedoxygen when temperature increases controls the corrosionrisk.

5 References

[1] International Dairy Federation, Bulletin 288 (1993) pp 2 ± 16.[2] International Dairy Federation, Bulletin 236 (1988) pp 10 ±

19.[3] Sandvik in Sandvik Sanicro Und SAF (1990) 2507.[4] J. L. Crolet, L. Seraphin, R. Tricot: Revue de MeÂtallurgie 11

(1977) 647.[5] T. Hong, M. Nagumo: Corros. Sci. 39 (1997) 1665.[6] P. W. Wild: Modern analysis for electroplating, Finishing pub-

lication Ltd., England 1974.[7] H. Girard: Thesis, BesancËon 1996.[8] B. Baroux: Revue de MeÂtallurgie 12 (1988) 683.[9] A. J. Sedrik: Internat. Metals Reviews 28 (1983) 295.

[10] N. K. Adam: The physics and chemistry of surfaces, DoverPublications Inc., New York 1978.

[11] T. Arai , W. Norde: Colloids and Surface 51 (1990) 1.

[12] R. J. Colton, J. S. Murday, J. R. Wyatt, J. J. de Corpo: Surf.Sci. 84 (1979) 235.

[13] H. P. Leckie, H. H. Uhlig: J. Electrochem. Soc. 113 (1966)1262.

[14] N. D. Stolia: Corros. Sci. 9 (1969) 455.[15] C. T. Liu, W. T. Lindsay: J. Sol. Chem. 1 (1972) 45.[16] T. Hong, M. Nagumo: Corros. Sci. 39 (1997) 285.[17] M. A. Streicher: J. Electrochem. Soc. 103 (1956) 375.[18] Z. Szklarska-Smialowska, J. H. Wang, C. C. Su: Corros. Sci.

44 (1988) 732.[19] P. Walstra, R. Jennes: Dairy chemistry and physic, J. Wiley

Ansons, New York (1984).[20] A. R. Hill: Milchwissenschaft 43 (1988) 565.[21] M. Lalande, J. P. Tissier, G. Corrieu: Biotechnology Progress

1 (1985) 131.[22] P. O. Hegg, K. Larson: Proceedings TyloÈsand (1981) 250.[23] J. C. Heughebaert, G. Montel: C.R. Acad. Sci. Paris 270 (C)

(1970) 104.[24] C. Holt: International Dairy Federation, Special issue n89501

(1995) 105.

(Received: September 14, 2000) W 3527

636 Perrin, Girard, Daufin and Pagetti Materials and Corrosion 52, 629±636 (2001)