Metal Release and Corrosion Resistance of Different Stainless Steel Grades in Simulated Food Contact

Stainless steel is an iron (Fe) based alloy with at least 11 wt% chromium (Cr),¹  which may also contain several other alloying elements, such as nickel (Ni), molybdenum (Mo), and manganese (Mn).²  Stainless steel is widely used in food- and beverage-relevant applications because of its high corrosion resistance in combination with good mechanical properties.³  The corrosion resistance of stainless steels is a result of the existence of a very thin self-healing chromium-rich passive surface oxide (sometimes called passive film or passive layer), typically 1 nm to 3 nm thick.^4-5  In this paper, the passive film is denoted the surface oxide and contains usually divalent or trivalent Fe and trivalent Cr oxides, hydroxides, and/or oxyhydroxides.^4,6-7  Nickel is enriched in its metallic form beneath the surface oxide, but usually is not present in the surface oxide.^4,8-11  Oxides of manganese and molybdenum can be present in the surface oxide of some stainless steel grades and at some environmental conditions.^{4-5,7-8,12-13}  The surface oxide composition, thickness, and other properties dynamically change with time and gradually adjust to the environment.^4-6,11,14  For example, chromium is enriched in the surface oxide at acidic conditions.^11,14-18  Passivation of the surface oxide in citric acid has, with significantly reduced released metals as a consequence, been reported.^13-14,19-20 

The extent of released amounts of the main alloying constituents, Cr, Ni, Mn, and Fe, from stainless steel cookware depends on several factors including grade, cooking time, pre-use, and temperature.^19,21  Different metal release mechanisms for the stainless steel surface in solutions of relevance for food applications, including electrochemical (metal corrosion/oxidation), chemical/electrochemical (dissolution of the surface oxide), or physical processes (removal of metal or oxide particles via, e.g., friction), have recently been reviewed.¹⁹ 

In Europe, a new test guideline has recently been published by the Council of Europe (CoE) to ensure safety of metals and alloys in food contact.²²  Main changes compared to the earlier available test, stipulated within the Italian Decree,²³  are the use of citric acid (5 g/L, pH 2.4) instead of acetic acid (31.5 g/L, pH 2.4) to simulate contact with acidic foods. There is also greater freedom in the test setup to enable more application-realistic investigations. Specific release limit (SRL) values, based on available toxicological, daily intake, and/or sensitization data, have been stipulated in the guideline for metals of concern. These values are used in compliance tests for comparison with corresponding levels released from metals and alloys into the test medium at a given surface area to solution volume. Recent findings show that the CoE protocol test conditions provide similar, or even more aggressive, test conditions from a metal release perspective compared with the setup described by the Italian Decree.¹³ 

The main objectives of this study were:

• (i)

  to quantitatively assess the extent of released metals from austenitic (grades AISI 204 [UNS S20431],⁽¹⁾ 304 [UNS S30400], and 316L [UNS S31603]), ferritic (grades AISI 430 [UNS S43000] and EN 1.4003 [UNS S40977]), and lean duplex stainless steels (grade EN 1.4162 [LDX 2101^†, UNS S32101]) exposed in citric acid (pH 2.4) and artificial tap water (pH 7.5), when following the CoE protocol and upon repeated exposure;

• (ii)

  to relate the extent of metal release to differences in surface oxide composition, the open circuit potential (OCP)–time dependence, and corrosion resistance of the same grades; and

• (iii)

  to investigate the combined effect of high chloride concentrations (0.5 M NaCl) and citric acid at pH 2.2 and 5.5 on the extent of metal release with and without prior surface passivation by citric acid (elucidated for AISI 304). The pH 2.2 solution simulates very harsh conditions.

Six different stainless steel grades of 2B surface finish (bright cold-rolled, annealed, pickled, and skin-passed sheet) and sheet thicknesses ranging from 1 mm to 2.5 mm were supplied by the International Stainless Steel Forum. Another stainless steel grade, AISI 201 (UNS S20100), with a different surface finish, 2D (dull cold-rolled, annealed, and pickled), is included for comparison with an earlier study.¹³  The microstructure and nominal bulk composition of the different grades are presented in Table 1.

Additional information on the duplex microstructure of the grade EN 1.4162 is available in the Appendix (Figure A1).

All coupons were prepared with a total geometric surface area of approximately 6 cm² (each defined separately). As-received surfaces (2B or 2D) were investigated following the CoE protocol.²²  All cutting edges of the as-received coupons were abraded using 1200 grit SiC paper. The coupons were then cleaned ultrasonically in ethanol and acetone for 5 min each, subsequently dried with cold nitrogen gas, and aged (stored) for 24±1 h in a desiccator (at room temperature). As the thickness of the coupons varied between 1 mm and 2.5 mm, the edge area to total surface area ratio varied between 10% (grades AISI 201, 204, and 304), 15% (grade EN 1.4003), 19% (grade AISI 430 and EN 1.4162), and 24% (grade 316L). During exposure in solution, the surface area to solution volume ratio was kept constant at 1 cm²/mL. Triplicate coupons and one blank sample (test solution only) were exposed in parallel for each grade, time period, and test solution. All exposures were conducted in an oven at stipulated temperatures (Torrsterilisator^†, Termaks). All vessels were acid-cleaned in 10% HNO₃ for at least 24 h, rinsed four times in ultrapure water (>18 MΩ·cm, Millipore^†), and dried in ambient laboratory air. All chemicals used were of analytical grade (p.a.) or puriss p.a. grade (in the case of nitric acid, used to acidify solution samples to a pH of <2 prior to atomic absorption spectroscopy [AAS] analysis). The pH of all test solutions was measured before and after exposure (with pH changes between 0 and −0.13 for citric acid, and between −0.04 and 0.76 for the non-buffered artificial tap water).

As-received coupons of the different grades of stainless steels were exposed in two different fluids (see Table 2), relevant food simulants for different types of food contact, for time periods up to 10 d (240 h): 2, 26, and 240 h at 70°C (for the first 2 h) followed by 40°C (as stipulated in the CoE protocol²²  to simulate cooling food and for short- and long-term applications). That is, 2 h (70°C), 2 h (70°C) + 24 h (40°C), and 2 h (70°C) + 238 h (40°C).

Abraded and 24-h aged (desiccator, room temperature) AISI 304 and 316L coupons (abraded by commercially available stainless steel wool, “fixa stålboll”^†) were exposed into 5 g/L citric acid (pH 2.4) for three consecutive 30 min periods at 100°C, as stipulated by the Italian Decree.²³  These coupons were then re-abraded, cleaned ultrasonically in ethanol and acetone for 5 min each, subsequently dried with cold nitrogen gas, and aged (stored) for 24 h (room temperature, in a desiccator), followed by another three consecutive 30 min exposure periods in 5 g/L citric acid (pH 2.4) at 100°C. The abrasion by the stainless steel wool and the last three exposures are not stipulated in the Italian Decree, but were chosen to simulate normal household use. The wool material was stainless steel to avoid galvanic effects.

The effect of chlorides, citric acid, and their combination on the extent of metal release was investigated in the solutions shown in Table 3. The effect of pre-passivation in 5 g/L citric acid at 70°C for 2 h was also investigated. This investigation was conducted on as-received AISI 304 coupons. A NaCl concentration of 0.5 M (29.22 g/L) was chosen as representative for a high salt concentration in food.²⁴ 

Total concentrations of released, non-precipitated Fe, Cr, Ni, and Mn (and Mo only for grade AISI 316L) in the different test solutions were determined for acidified samples (pH < 2) by means of graphite furnace atomic absorption spectroscopy (Perkin Elmer AA800^† analyst). The atomization temperature was 2,400°C (for Fe, Ni, and Mo), 2,300°C (for Cr), and 1,900°C (for Mn). All analyses were based on three replicate readings for each solution sample, and a quality control sample of known concentration was analyzed every 8th sample. The limits of detection, as determined from the average blank value + 3 times the maximum standard deviation of all blanks (solution samples without any stainless steel coupon) were 0.01 μg Fe/cm² (10 μg/L), 0.0004 μg Cr/cm² (0.4 μg/L), 0.004 μg Ni/cm² (4 μg/L), 0.0009 μg Mn/cm² (0.9 μg/L), and 0.0005 μg Mo/cm² (0.5 μg/L). Calibration was conducted using four calibration standards: 0 (ultrapure water), 10, 30, and 100 μg/L for Cr, Ni, Mn, and Mo, respectively, and 0, 50, 100, and 200 μg/L for Fe. The exposed solutions were diluted up to 20 times to ensure metal concentrations within the calibration range. Twenty times diluted control samples of Fe, containing 200 μg Fe/L, showed a maximum dilution-induced error of 2%. However, control solution samples of known concentration in the NaCl-containing solutions revealed an overestimation of up to 30%. This is accounted for in the error bars of the corresponding figure. The blank values in all solutions were <9.5 μg Fe/L, <2.3 μg Cr/L, <3.9 μg Ni/L, <4.0 μg Mn/L, and <0.1 μg Mo/L. All observed metal concentrations were statistically significantly higher compared with the blank values and exceeded the limits of detection, with the exception of released metals in artificial tap water, as indicated in the figures.

Released amounts of metals (in the unit μg/cm²) correspond to the concentration of released metals (μg/L) normalized to the exposed geometrical surface area (cm²) and solution volume (L). For the standard surface area to solution volume ratio (loading) of 1 cm²/mL, 1 μg/cm² is equivalent to 1,000 μg/L or approximately 1 mg/kg, where kg is the mass of the food simulant (in this case the test solution). All release data is presented in the unit μg/cm² as the mean value of triplicate coupons exposed in parallel, with the blank exposure concentration exposed in parallel subtracted (if > 0). Error bars show the standard deviation between these triplicate individual coupons.

X-ray photoelectron spectroscopy (XPS, UltraDLD^† spectrometer, Kratos Analytical) measurements using a monochromatic Al K_α x-ray source (150 W) were performed on two separate surface areas approximately sized 700 × 300 μm² for compositional analysis of the outermost surface oxide (with information depth of 5 nm to 10 nm). Elements of the outermost surface oxide were distinguished by running a wide spectrum and high-resolution spectra (pass energy of 20 eV) for the main alloying elements: Fe2p, Cr2p, Ni2p, Mn2p, Mo2p, O1s, and C1s (as energy reference). The results are presented as the relative mass content of oxidized iron, chromium, nickel, manganese, and molybdenum (only for grade 316L) in the outermost surface oxide, e.g., [Cr_ox/(Cr_ox + Fe_ox + Ni_ox + Mn_ox)]. Sensitivity factors provided with the XPS software were used to determine relative metal concentrations. Peak overlap between nickel and manganese was accounted for. Relative changes in the surface oxide thickness (before and after exposure) were assessed roughly based on the absence or presence of non-oxidized metal peaks.

The OCP of as-received surfaces (at least two replicates for each grade and solution) was monitored for a time period of 2 h at 70°C followed by 24 h at 40°C—the same as in the metal release investigations—in citric acid (5 g/L, pH 2.4) for grades AISI 204, 304, 316L, EN 1.4003, and EN 1.4162, and in artificial tap water (pH 7.5) for grade AISI 204. Furthermore, the behavior of grade AISI 304 was also investigated in 0.5 M NaCl (pH 5.5) and 5 g/L citric acid + 0.5 M NaCl solution (pH 2.2). Sample preparation was performed as described earlier. OCP was monitored using a Metrohm μ-Autolab Type II Potentiostat^† equipped with the Nova 1.5^† software and an Ag/AgCl saturated KCl electrode as reference electrode. All measurements (at least duplicate measurements) were performed at aerated conditions.

To investigate the corrosion resistance of the different grades in solution, dynamic polarization measurements were conducted in 5 g/L citric acid (pH 2.4) at 40°C (for approximately 2 h) for abraded (1200 SiC), not pre-passivated, coupons of the ferritic grade EN 1.4003, the austenitic grades AISI 304 and 316L, and the lean duplex grade EN 1.4162. Furthermore, they were conducted in 0.5 M NaCl (pH 2.2 and 5.5) and 5 g/L citric acid + 0.5 M NaCl solution (pH 2.2 and 5.5) at 40°C for abraded (1200 SiC), not pre-passivated, coupons of AISI 304. All measurements were performed at aerated conditions. At least 2 replicates of each grade and solution were examined. The total exposed geometric surface area of the coupons was kept constant (2.25 cm²) in 100 mL of the test solution. A three-electrode cell, with the coupon as working electrode, an Ag/AgCl saturated KCl electrode as reference electrode, and a Pt wire as the counter electrode, was used. During polarization, the potential was swept anodically from the OCP with a scan rate of 0.0005 V/s. The backward scan (in chloride-free solutions) started when the potential reached 1.3 V, going back to the starting potential (i.e., the measured OCP for the specific grade and solution). All data are shown until a current density of 160 μA/cm² (0.0004 A) was reached.

Released amounts of Fe, Cr, Mn, Ni, and Mo from as-received stainless steel grades into citric acid (5 g/L, pH 2.4) and artificial tap water (pH 7.5) are presented in Figures 1 and 2. Parts of the data shown here have been presented earlier by the authors in a short industry report²⁵  and data for the grade AISI 201 (from an earlier study¹³ ) is included for comparison. All released amounts of metals from the investigated grades were well below their corresponding specific release limits in both test solutions (Table 4).

Released amounts of metals, including Mn and Ni, were higher in the acidic solution of citric acid (pH 2.4) compared with the tap water solution (pH 7.5), which was expected.^13,19  The total amount of released metals from each grade into citric acid (Figures 1 and 2) was in all cases dominated by released Fe (80% to 98%), which is in agreement with literature findings in similar solutions.¹⁹  It is expected that the corrosion resistance of stainless steels increases with increasing bulk content of chromium.⁶  The total amount of released metal (Fe + Cr + Ni + Mn + Mo), dominated by Fe, decreased with increasing chromium bulk content (Figure 1[a]). However, this trend (decreasing release with increasing chromium bulk content) is not true if the individual alloying elements are looked at separately (Figures 1[b] and 2), for which no influence of the chromium bulk content could be observed for any of the released alloying elements. This finding is typical for stainless steel that is not actively corroding.¹⁹  A higher extent of Mn or Ni release was seen from alloys of higher manganese or nickel bulk content. It should be noted that the release of Mn or Ni is not necessarily governed by their corresponding bulk content, as judged from literature findings.^12,19,26  Figure A2 (Appendix) shows the metal release, normalized to the bulk content. Mn was the preferentially released element, as compared to its bulk content, in both test solutions for most grades. Preferential release of Mn in citric acid-containing solutions and water has previously been shown for AISI 201.¹³  In addition, Fe was released preferentially in citric acid (pH 2.4) from all grades. Preferential release of Fe from stainless steels at passive conditions is in agreement with literature findings.¹⁹  Released Fe might precipitate from solution at a pH of 7.5 in artificial tap water.^27-28  Preferential release of Fe and Mn from the surface oxide on stainless steel can be explained partly by their higher oxide solubility compared with chromium oxides^29-31  and partly by the absence of nickel in the outermost surface oxide.^5,19,32 

The largest proportion of the total amount of released metals into citric acid was released during the first 2 h, followed by diminished released amounts (Figures 1 and 2), which indicates surface passivation (increased barrier properties of the passive surface oxide). To investigate whether enhanced surface passivation also takes place during repeated usage of pots and pans at boiling conditions, AISI 304 and 316L were exposed to repeated immersions at 100°C for six times for 30 min. The surface was abraded using a household stainless steel wool prior to the first and fourth repetition. The results are illustrated in Figure 3 for grade AISI 304 and in Figure A3 (Appendix) for grade AISI 316L. The amounts of alloy constituents released upon repeated exposure were also reduced (about 10-fold) for both grades, which indicates surface passivation in citric acid at these conditions. This would explain why the extent of metal release of subsequent exposures following the exposure after the abrasion is comparable, e.g., the 2nd and 5th exposure in Figure 3. Any effect of the deposition of particles from the stainless steel wool cannot be excluded; however, the surfaces were thoroughly cleaned after abrasion. The results are consistent with literature findings for the different stainless steel grades that show a reduction of metal release for repeated exposures to different food simulants,¹⁹  and a strong enrichment of chromium in the surface oxide upon exposure in citric acid.^13,19-20 

The influence of exposure of stainless steel in citric acid or artificial tap water on the outermost surface oxide composition is shown in Figure 4 by means of XPS findings. Oxidized chromium was enriched with time in the outermost surface oxide of all investigated grades after exposure in citric acid (Figure 4[a]), in agreement with other studies.^13,19-20  No statistically significant effects were observed for the grades when exposed in artificial tap water, except for grade AISI 201 (Figure 4[b]). These findings are in agreement with OCP results (Figure 5) that show passivation of different grades (increased OCP with time) upon exposure in citric acid (pH 2.4), effects not observed in artificial tap water (investigated for grade AISI 204, Figure 5[c]).

Surface passivation in citric acid, expressed as the change in relative amount of oxidized chromium compared with other oxidized metals in the outermost surface oxide [Cr_ox/(Cr_ox + Fe_ox + Ni_ox + Mn_ox)] with time (Figure 4[a]) was most pronounced for the lowest corrosion resistant grade EN 1.4003 (from approximately 18.5% to 82% after 10 d). The lowest change, though still substantial, was observed for the austenitic grades AISI 304 and 316L (from approximately 40% to 60%), findings in agreement with their corresponding change in OCP with time (Figures 5[a] and [b]). The enrichment of chromium in the surface oxide of stainless steel with time is generally expected in acidic solutions.^13,19-20  A more rapid surface passivation in citric acid was evident for grade AISI 316L compared with AISI 304, which showed a more gradual enrichment with time (Figure 4[a]). In agreement with other studies,¹²  oxidized manganese was only observed in the outermost surface oxide for grades with high manganese bulk content. In the case of the high manganese duplex grade EN 1.4162 (in both solutions) and AISI 201 (in tap water), manganese was depleted from the outermost surface oxide (Figure 4). That manganese may be depleted from the surface oxide for some grades and conditions has also been observed in an earlier study on gas-atomized AISI 316L particles in different solutions of pH between 4.5 and 6.5 (including citric acid solutions).³³  It may be related to the low chemical stability of manganese oxide in citric acid containing solutions.³⁴  In contrast, molybdenum (present in the surface oxide of 316L) was not depleted, but enriched in citric acid (Figure 4[a]). Molybdenum was previously found to be enriched in the outer layer of the surface oxide of stainless steel AISI 312L (UNS S31254).³⁵  Oxidized molybdenum is rather insoluble at pH 2.4,³⁶  which may explain its enrichment.

The electrochemical investigations (OCP measurements), in Figure 5, clearly show that no active corrosion or stable pitting took place for the investigated stainless steel grades either in citric acid or in artificial tap water. The ferritic grade of EN 1.4003 (11 wt% Cr), which released the highest amount of metals among the grades (Figure 1), showed a few occasions during which the OCP suddenly was reduced followed by its subsequent increase back to the same level (Figure 5[a]). This is possibly related to the dissolution of inclusions and hence repassivation of metastable pits. The OCP level is not only influenced by the extent of passivation, but also by the oxidation state and presence of manganese in the surface oxide.³⁷  This might contribute to a higher OCP for the grades that contain manganese in the surface oxide (AISI 204 and EN 1.4162 in Figures 5[a] and [b]).

By considering the total amounts of released metals (Figures 1 and 2) and assuming that metals are only released from the outermost surface oxide (c.f., Appendix), the total released amount of metals would correspond to a reduction in surface oxide thickness between 3.4 nm and 14 nm in citric acid (pH 2.4) and less than 0.1 nm in artificial tap water (pH 7.5) for all grades. These calculations indicate that the total amount of released metals in citric acid cannot solely originate from the few nanometer-thick surface oxide that is already present on the surface prior to exposure,^4-5  and that the surface oxide continuously is reformed/changed. The OCP results further suggest that chemical processes, such as protonation,^19,27  are the dominant metal release mechanisms in citric acid at pH 2.4. Literature on the time-dependent release of Fe from AISI 304 indicates a slower passivation and higher release of Fe in citric acid (5 g/L, 37°C) after exposure periods up to 1 week¹⁴  compared with findings from exposures up to 10 d according to the CoE protocol (70°C/40°C) of this study. The reason is most probably related to a more rapid passivation at a higher temperature.

In agreement with the OCP results, accelerated corrosion tests by means of dynamic polarization measurements for abraded grades EN 1.4003, AISI 316L, and EN 1.4162 in citric acid of pH 2.4 did not indicate any active corrosion or pitting corrosion (Figure A5 in the Appendix).

For materials with passive surface oxides, such as stainless steel, chlorides in food may cause problems as a result of their potential capacity to locally destroy the passive surface oxide and induce pitting corrosion if the incorrect grade has been chosen for a specific application.²¹  The metal release and electrochemical behavior (OCP and critical pitting potential [CPP]) of the commonly used grade AISI 304 were therefore investigated in aqueous solutions with a high content of chlorides (0.5 M NaCl) both with and without citric acid (5 g/L). The effect of pH was also investigated, as this affects both pitting corrosion and different chemical dissolution mechanisms.¹⁹  As citric acid has a surface passivating effect, the effect of pre-passivation in citric acid (5 g/L, 2 h, 70°C) was also investigated. The results are shown in Figures 6 and 7 and Table 5. In Figure 6, earlier results on the effect of nitric acid passivation are included for comparison.¹⁴ 

The pre-passivation in 5 g/L citric acid for 2 h at 70°C resulted in a reduction (up to 27-fold) of the extent of released metal (Figures 6[c] through [f]). This is comparable with a surface passivation in 6 M HNO₃ at 60°C for 2 h for which the Fe release was reduced to slightly lower levels (Figure 6[a]). Both citric acid and nitric acid exposure resulted in strong chromium enrichment of the surface oxide (Figure 6[b]). The citric acid pre-passivation also resulted in a higher reproducibility among triplicate coupons, and a reduced dependence of solution composition and pH compared with coupons that were not pre-passivated (Figures 6[c] through [f]).

The presence of citric acid was more important for the release of Fe and Cr (present in the surface oxide) compared with the release of Mn and Ni, which were more dependent on the presence of chlorides for coupons that were not pre-passivated (Figures 6[c] through [f]). In contrast, the release of Fe and Mn were not at all affected by the solution composition or pH when the coupons had been pre-passivated in citric acid. Consistent with the coupons that were not pre-passivated, the release of Cr was slightly higher in solutions containing citric acid, whereas the release of Ni was slightly higher in solutions containing NaCl for pre-passivated coupons (Figures 6[d] and [e]). When citric acid was added to 0.5 M NaCl, or when the pH was reduced, the released metal levels were higher and the OCP values lower for coupons that were not pre-passivated (Figure 7 and Table 5). The effect of pH on the extent of metal release and corrosion resistance for surfaces of stainless steel that were not pre-passivated is expected.^19,38-39  The presence of citric acid was, as discussed previously, shown to increase the amount of released metals for stainless steel that was not pre-passivated both at pH 2.2 and 5.5 as a result of a surface passivation effect.^13-14,19-20  In 0.5 M NaCl at pH 5.5, the released amounts of metals were below the detection limit of Cr, and very low for the other elements investigated (Figures 6[c] through [f]). Consequently, this solution was the only solution for which the pre-passivation did not result in an overall reduction of the extent of metal release.

Observed OCP values for the surfaces that were not pre-passivated were lower compared with corresponding CPP in each solution (Figure A6 in the Appendix and Table 5). The CPP decreased with reduced pH and was unaffected by the presence of citric acid at pH 2.2 (no pitting occurred in citric acid when chlorides were absent). The CPP was slightly lower in chloride-containing citric acid of pH 2.2 compared with pH 5.5 (Figure A6 in the Appendix and Table 5). No active corrosion was observed in any solution as judged from the observation that the OCP increased/stabilized with time and was lower compared with the CPP. The OCP results for coupons that were not pre-passivated showed some metastable pitting events that mainly took place initially (Figure 7). Taking all of these observations together, it is likely that manganese-rich inclusions are dissolved during the pre-passivation step or during the initial exposure to citric acid containing solutions. This results in less metastable pitting events, higher reproducibility, less overall metal release, and a lower influence of solution composition and pH.

This study cannot distinguish between the effects of stainless steel composition, the microstructure, the impurity content, and the inclusion density and size on the metal release in citric acid containing solutions. To address this aspect, further studies should investigate the same grade of varying impurity contents. This study was performed at temperatures stipulated by the CoE protocol guideline²²  and/or the Italian Decree.²³  The effect of temperature should further be investigated separately. Solutions containing both chlorides and citric acid have been addressed in this study, but only for two pH conditions and at a high concentration of chlorides. To assess prevailing metal release mechanisms in the presence of relevant concentrations of citric acid and chlorides at pH values relevant for food, further studies are needed. It is speculated that the different mechanisms involved, such as pitting corrosion, inclusion dissolution, protonation, and surface complexation,¹⁹  may have different pH dependences, and have antagonistic or synergistic effects on the metal release process from stainless steels.

A new European food application test protocol (CoE protocol) using citric acid as food simulant was published as a technical guide in September 2013 to ensure the safety of using metals and alloys as food-contact articles. The objectives of this study were to quantitatively assess the extent of metal release from stainless steel grades of different microstructure (austenitic grades: AISI 201, 204, 304, and 316L; ferritic grades: AISI 430 and EN 1.4003; and a lean duplex grade: EN 1.4162) in citric acid (5 g/L, pH 2.4) and in artificial tap water (pH 7.5) following the CoE protocol test procedure, and to assess how these exposures influence the corrosion resistance, passive properties, and composition of the surface oxide. The objectives were further to investigate the effect of repeated exposures of stainless steel surfaces in citric acid at boiling temperature and to elucidate the combined effect of relatively high chloride concentrations (0.5 M NaCl) and citric acid on the metal release process in dependence of pre-passivation. The following main conclusions were drawn:

• Exposures of stainless steel grades of different microstructure (EN 1.4003, AISI 430, 204, 201, 304, 316L, and EN 1.4162) in citric acid or artificial tap water up to 10 d (at 70°C for 2 h followed by 40°C for the remaining time period) resulted in lower metal release levels for all grades than the specific release limits stipulated in the CoE protocol. They all passed the compliance test.

• Released amounts of metals were close to the detection limits for all grades in artificial tap water.

• Exposures in citric acid initially increase the release of metals from stainless steel, but result in increased surface passivation and reduced metal release with time resulting from the enrichment of chromium in the passive surface oxide.

• The total amount of metal release (dominated by Fe, >80%) in citric acid (pH 2.4) decreased with increasing bulk alloy content of chromium. No such correlation to the chromium bulk content was evident for released amounts of Cr, Ni, Mn, or Mo.

• Fe (in citric acid) and Mn (in all solutions) were preferentially released into solution from the stainless steel grades. Ni was released to the lowest extent, as compared to its bulk alloy content.

• Repeated usage of stainless steel (elucidated for AISI 304 and AISI 316L) in citric acid solutions at 100°C resulted in decreased amounts of released metals with time.

• No active corrosion was observed in any solution. Metastable pitting events occurred initially for coupons that were not pre-passivated in solutions containing 0.5 M NaCl.

• 0.5 M NaCl induced a very low extent of metal release (close to detection limits) from grade AISI 304 at pH 5.5. When combined with citric acid (5 g/L) and at lower pH (2.2), 0.5 M NaCl induced a slightly higher extent of metal release compared with citric acid (pH 2.4) without NaCl for coupons that were not pre-passivated. Pre-passivation largely reduced this solution dependence.

• Pre-passivation reduces the metal release up to 27-fold from stainless steel AISI 304 in solutions containing both chlorides and/or citric acid at pH 2.2 to 5.5. It results further in a high reproducibility of observed release data. Relevant pre-passivation steps, e.g., for 2 h in the test solution as in this study, are therefore recommended to accurately mimic the metal release behavior of stainless steel in simulated food contact.

The International Stainless Steel Forum, Belgium, and Jernkontoret, Sweden, are highly acknowledged for financial support and for providing stainless steel coupons. The members of Team Stainless, especially Dr. Jacques Charles, are highly acknowledged for financial support and valuable discussions.

Assumptions:

• (i)

  Metal release (Meⁿ⁺) only from the outermost surface oxide (Figure A4)

• (ii)

  No corrosion

Based on these assumptions and by considering Equation (A1), it is possible to calculate the reduction in surface oxide thickness (h) resulting from the metal release process.

In Equation (A1), ρ is the density of surface oxide (g/cm³), m is total mass of released metals (g), and V is the volume of the surface oxide (cm³), which corresponds to a surface area (A, 1 cm²) × surface oxide thickness reduction (h).

Thus, the surface oxide thickness reduction (h, in nm) for a 1 cm² surface area corresponds to:

XPS analysis shows the presence of oxidized chromium, iron, and manganese in the outermost surface oxide of the investigated stainless steels. As the density ρ of the mentioned oxides is very similar (ρ_Cr2O3 = 5.22 g/cm³, ρ_Fe2O3 = 5.24 g/cm³, and ρ_MnO2 = 5.03 g/cm³), the density of the predominating oxide (Cr₂O₃) was considered as ρ_{surface oxide}.

Taking EN 1.4003 stainless steel grade as an example, which showed the highest amount of released metals in citric acid (5 g/L, pH 2.4), the surface oxide thickness reduction can be calculated using a surface oxide density, ρ_{surface oxide} of 5.22 g/cm³ and the total amount of released metals of 7.1 μg/cm². Thus for 1 cm² surface area, the released mass corresponds to m = 7.1 × 10⁻⁶ g. The calculated reduction in surface oxide thickness, h, then equals approximately 14 nm.

The current peak at 1 V_Ag/AgCl derives most probably from the formation of Fe(III)-citrate complexes (0.9 V_Ag/AgCl to 1.2 V_Ag/AgCl) or from the anodic decomposition of citric acid (1 V_Ag/AgCl to 1.1 V_Ag/AgCl).⁴²  It might also be explained by a thickening of the passive surface oxide (at approximately 0.9 V_Ag/AgCl),^42-43  the transpassive oxidation of Cr(III) (above 0.8 V_Ag/AgCl),⁴⁴  and the formation of a soluble citric complex resulting from the transpassive dissolution of metals (at approximately 1.1 V_Ag/AgCl to 1.2 V_Ag/AgCl).⁴²  At potentials exceeding 1.2 V, the current increased as a result of oxygen evolution. No hysteresis was seen during the reverse sweep for any of the grades, indicative of lack of pitting corrosion. The ferritic grade EN 1.4003 showed a cathodic current peak at approximately 0.12 V_Ag/AgCl, which can be caused by the reduction of the iron oxide layer formed during the previous anodic polarization of the surface.⁴²