The Evaluation of the Comparative Corrosion Behaviour of Conventional and Low-Nickel Austenitic Stainless Steel: Hercules™ Alloy

Abstract

Austenitic stainless steels require approximately 8% Ni to maintain austenitic microstructure at room temperature for alloys such as 304 stainless steel (304SS). Ni contributes approximately 60% of the total material cost and its price fluctuates, making the cost of austenitic stainless steel unpredictable. The use of low-nickel austenitic stainless steels as a substitute has been considered in order to remedy costs associated with Ni price fluctuations. Alloying elements such as Mn and N have been considered, however they have been found to reduce corrosion resistance. A new alloy namely Hercules™ has been developed with reduced Ni content (1.8–2% Ni). This chapter presents a comparative study of the corrosion behavior of Hercules™ and 304SS in different solutions. The alloys were evaluated using cyclic polarisation technique and immersion tests. The results demonstrated that the corrosion resistance of Hercules™ is comparable to that of 304SS. This presents the alloys as potential industrial substitutes of each other.

1. Introduction

Ni contributes about 60% of austenitic stainless steel manufacturing material price. This means that the price of austenitic stainless steel increases with an increase of Ni. Ni price fluctuation has led to major efforts to reduce its content in austenitic steels. Ni has been replaced with readily available, cheap elements such as Mn and N. Hercules™ is a low-Ni austenitic stainless steel alloy that was developed at Mintek-South Africa in the Advanced Materials Division. The typical content of Hercules™ comprises of 2 wt.% Ni, 9 wt.% Mn and 2.5 wt.% N [1, 2, 3].

When Hercules™ was tested for mechanical properties, it was found that it had higher tensile strength than 304SS in the hot rolled and annealed condition, hence it was termed Hercules™. This indicated that it can be used for structural applications where high strength is required. Possible applications for Hercules™ were targeted at reinforcement bars/rebars, fasteners and hot rolled channels for construction purposes. Construction industry is likely to benefit from Hercules™ because it has higher strength and reduced cost than 304SS. Other industries will soon find benefit also because it could later be available in flat product such as sheet and coil [4].

A minimum yield strength required in structural applications is 400 MPa in hot rolled condition. Typical tensile properties and density for Hercules™ alloy in hot rolled condition are shown in Table 1 [4, 5]. Fastener prototypes of Hercules™ have been manufactured and was divided into two types; the large head bolts (M16 and M24) and roof fasteners.

UTS (MPa)	850
0.2% Proof Strength (MPa)	500
Elongation	50%
Density (g/cm^3)	7.85

Table 1.

Typical mechanical properties of Hercules™ alloy [4].

According to White et al. [6], 240,000 tonnes of fasteners are produced globally per year and South Africa contributes only about 35,000 tonnes per year. Production of corrosion resistant bolts and roof fasteners promises a viable business. It is required that new LNASSs be resistant to corrosion. The latter requirement is in response to specific environment such as swimming pool applications, which requires pitting and crevice corrosion resistance [6].

Furthermore, about 400,000 tonnes of stainless steel rebar is produced per year. South Africa has been using about 95% carbon steel rebars which are cheaper than 304SS. However, there was little success in marketing of carbon steel rebars because of poor corrosion resistance. Corrosion of rebar is detrimental in that it can cause concrete spalling which leads to infrastructure failure. An infrastructure repair is more costly than preventing failure. Hence there is a need to use stainless steel rebar in any concrete because they are more corrosion resistant than carbon steel rebar. The conventional LNASS 201 stainless steel (201SS) is hardly available in South Africa; therefore, Hercules™ could close the gap for applications that require 304SS properties at a lower price. The cost production of Hercules™ bar is around 25% less than 304SS using a similar production route [6].

The corrosion resistance was however compromised by addition of Mn and N in Hercules™. Thus, to counteract this, 0.5 wt. % Mo was added (Hercules™ B) [2]. The focus is on characterisation of the pitting behaviour of Hercules™ B (with 0.5 wt. % Mo) against Hercules™ A (without Mo addition) and 304SS.

Pitting corrosion is the local discontinuity of a passive film which results in small holes through the material. These holes are referred to as pits. Initiation of pits can be caused by mechanical imperfection such as surface damage or inclusions. The composition of stainless steel may cause formation of inclusions which become nucleation sites for pits at the inclusion-austenite matrix interface [7]. Bautista et al. [8] studied the morphology of pits that were formed on the surface of test coupon after polarisation in NaCl. Pits were found to nucleate preferentially at the point of strain and at geometrical irregularities that favoured formation of corrosion cells [8].

The effect of Mo content on pitting and crevice corrosion resistance of stainless steels in chloride environments has been studied by Kaneko et al. [9]. The pitting potential for austenitic stainless steel alloys with 2 wt. % and 5 wt. % Mo contents showed a dramatic increase compared to that of steels without Mo in chloride environment. It is understood that Mo is adsorbed at the dissolving interfaces of a corroding metal and hence inhibiting dissolution kinetics [3, 9].

Kaneko et al. [9] findings are consistent with the work that was done by Pardo et al. [7]. He studied the effect of both Mn and Mo in the pitting corrosion resistance of 304SS and 316SS. Tests were performed by immersion in 6 wt. % FeCl₃.H₂O and cyclic polarisation technique in 3.5 wt. % NaCl. The scanning electron microscopy was used to examine the morphology of pits on the surface of corroded coupons. Images of corroded samples of 304SS that were electrochemically tested in 3.5 wt.% NaCl are shown in Figure 1 [7].

Figure 1.

Pitting corrosion of 304SS as a function of Mn (adopted from [7]).

Alloys with higher Mn content had larger pits compared to ones with lower Mn content. That is, higher Mn stainless steels experience high pitting because it was thought that since Mn has high affinity for sulphur, it reacts with Mn to form inclusions (MnS), which in turn are precursors for pit nucleation. When Mo was increased to 2.10 wt.%, pit density decreased and the size of pores evidently decreased as shown in Figure 2 [7].

Figure 2.

Pitting corrosion of 304SS as a function of Mo (adopted from [7]).

Pit initiation can also be influenced by surrounding conditions such as gaseous environment, temperature and the nature of the electrolyte. Stainless steels tend to form deep pits at specific areas when exposed to environments that contain solutions with chloride, bromide or hypochlorite [10, 11].

The occurrence of the electrochemical reactions is a result of a potential change created on a conductive metal when exposed to a conductive medium. Electrochemical potential is accompanied by electron movement which leads to electron availability at the metal surface. The electron movement or potential difference can affect the rate of corrosion reactions. Overall, the energy change provides the driving force and control for the spontaneous direction for a chemical reaction. The change in energy can be understood using thermodynamics to show how conditions of the corrosion cell can be adjusted to avoid corrosion. When a metal is immersed in a conductive solution, a charged surface of an alloy forms a complex interface. The interface is formed when the polar H₂O molecules form an oriented solvent layer. The electric field formed around the solvent layer prevents easy charge transfer, thereby limiting electrochemical reactions at the surface of an alloy [10].

The positively charged ions such as Fe²⁺ at the anode are transferred to the conductive solution which acts as an electrolyte for the cell. The electrolyte consists of ions that create electrical connectivity with the metal. Oxygen and water act on the cathodic reaction and accept negatively charged ions to form hydroxyl ions$\left(O{H}^{-}\right)$. Further anodic reactions occur simultaneously with cathodic reactions during corrosion. Typical anodic reactions are shown in Eqs. (1)–(3) and the cathodic reaction in Eq. (4). These type of corrosion reactions occur when the alternative air exposure and water is present, for example in the sea wave condition [12].

$$Fe\to F{e}^{2+}+2{\mathrm{e}}^{-}$$E1

$$F{e}^{2+}+2O{H}^{-}\to Fe{\left(OH\right)}_2$$E2

$$4Fe{\left(OH\right)}_2+2{\mathrm{H}}_2\mathrm{O}+{\mathrm{O}}_2\to 4Fe{\left(OH\right)}_3$$E3

$$4e^{-}+O_2+2H_{2}O\to 4O{H}^{-}$$E4

Stainless steels can corrode by pitting mechanism without a significant loss of weight on a whole structure being recognised. For example, chloride induced corrosion of stainless steel rebars in the concrete occur when there exists a difference of electric potential along the rebar. In the presence of chloride ions, the surface of the rebar is activated to act as the anode and the passivated region becomes the cathode. The reactions involved are shown in Eqs. (5) and (6) [12].

$$F{e}^{2+}+2C{l}^{-}\to {\mathrm{FeCl}}_2$$E5

$$FeC{l}_2+2H_{2}O\to Fe{\left(OH\right)}_2+2HCl$$E6

The chloride ions migrate easily towards the interior of the pit and catalyse the hydrolysis reaction. An acidic environment is created in the pit solution as the reaction continues. During corrosion, more than one anodic reaction takes place because of different elements present in the alloy. The electrons produced by these anodic reactions are consumed by the cathodic reactions which includes hydrogen and metal reduction. Removing the cathodic sites therefore reduces the rate of corrosion. The potential required for corrosion to take place is denoted by potential corrosion E_corr. This is the potential at which the total rate of all anodic reactions is equal to the total rate of all cathodic reactions. The corrosion current density at this point is denoted by i_corr and it is used to measure corrosion rate of the metal as anionic species are released [13].

Electrochemical tests can be used to determine i_corr and can be measured indirectly with the aid of a counter electrode and electronic equipment. This technique uses a potentiostat in conjunction with the reference electrode. Potentiostat is an instrument that applies a potential to a specimen which enables the modification of current flow. The commonly used electrochemical techniques are polarisation resistance technique, electrochemical impedance and Tafel extrapolation. The current can be measured by extrapolation procedure whereby a specimen is initially made to act as a cathode in the electrochemical cell containing the test solution [13]. Some of these techniques can be used to determine the lifetime of a metal by calculating the time required for initiation and propagation of corrosion to cause failure.

Studies have been conducted on commercial stainless steels such as 304SS and 316SS. Pitting or crevice corrosion resistance of these alloys in chloride environments can also be measured by immersion tests in metal-chloride solutions [14].

Bergstrom et al. [8] followed guidelines outlined on ASTM G48 method A and B [15] to test susceptibility of 201SS and 304SS to pitting corrosion. ASTM G48 method A is the practice for measuring pitting resistance and method B for crevice corrosion resistance. These two methods included immersion of coupons in 6 wt.% FeCl₃.6H₂O and measuring mass loss due to either pitting or crevice corrosion after 72 hours as shown in Table 2. The results showed that there was no significant difference in the mass loss of 201SS and 304SS. Therefore, this means that according to Bergstrom et al. [8], there was no difference in the corrosion behaviour of 201SS and 304SS in 6 wt.% FeCl₃.6H₂O even in the presence of an artificial crevice.

		201SS	304SS
ASTM-G48 A	Mass loss	0.0228 g/cm^2	0.0280 g/cm^2
	Max. Pit depth	0.0762 mm	0.0762 mm
ASTM-G48 B	Mass loss	0.0211 g/cm^2	0.0205 g/cm^2

Table 2.

Results of ASTM-G48 A and B tests conducted at 22°C by Bergstrom et al. [8].

Similarly, Garcia-Alonso et al. [16] also performed corrosion tests of 304SS and 316SS rebars embedded in concrete with different chloride concentrations. Tests were also done for carbon steel and LNASSs for comparison. LNASSs had Ni composition of 0.2 wt. % and 1.5 wt. % with addition of Mn. The concrete was manufactured with additions of 2 wt. % and 4 wt. % chloride content of cement. Other rebars were also embedded in a portion of concrete without chlorides and immersed into 3.5 wt.% NaCl solution to evaluate the effect of diffusion of chloride ions through the non-chlorinated concrete in the corrosion behaviour of stainless steel rebars [16].

The i_corr values of carbon steel, 304SS, 316SS and LNASSs were measured using electrochemical methods. In the absence of chlorides, i_corr values for all test alloys were measured around 0.1 μA/cm². The i_corr for carbon steel was observed to moderately increase after 30 days of immersion, which is attributed to diffusion of chlorine at the surface of the rebar. The i_corr values for alloys that were embedded in concrete with 2 wt. % chlorides were measured to be 3–5 times higher for carbon steel compared to that of LNASSs and 316SS. In the slab with 4 wt. % chlorides, carbon steel was measured to have i_corr value 10 times higher than of stainless steels. 304SS was measured to have an i_corr value that is of at least one magnitude lower than other alloys in 4 wt.% chlorides concrete slab [16].

The response of these stainless steel and carbon steel alloys towards increased chlorides concentration can be attributed to local breakdown of the passive layer depicting the occurrence of pitting corrosion [1, 16].

Furthermore, Bautista et al. [14] also performed corrosion experiments for LNASS Type 204Cu stainless steel (204SS) in a solution simulating “pore solution” of the concrete. The composition of 204SS consisted of 1.89 wt. % Ni and 8.25 wt. % Mn. Cyclic polarisation technique was used to test the susceptibility of 204SS rebars towards pitting corrosion against conventional 304SS and 316SS.

A number of mixtures of saturated calcium hydroxide concrete solutions were used with different NaCl additions. The pitting potential values were measured from the cyclic polarisation curves at the potentials where the current sharply increases when the working electrode is anodically polarised. No pitting was detected on the media without NaCl addition. However, pitting was detected for tests done with additions of NaCl. The greater the amount of NaCl added for each test, the lower the pitting potential obtained. The presence of chloride ions causes the passive layer to break down at potentials below the transpassive region and results in pitting corrosion. An example (Adapted from [14]) of pitting scans showing the effect of addition of NaCl in the corrosion behaviour of 204SS is shown in Figure 3 [14].

Figure 3.

Influence of chloride concentration on the pitting of 204SS (adopted from [14]).

The cyclic polarisation curve labelled 0% NaCl illustrates the typical behaviour of stainless steels in the absence of chloride ions. The curve shows passivity until it reaches a potential above 650 mV and evolution of oxygen is observed and referred to as transpassive region. The reverse scan for 0% NaCl curve shows current density values that are less than that of the forward scan. This means that in this medium, 204SS is not susceptible to pitting corrosion. Increasing NaCl concentration reduced the pitting potential value with a certain significant order and also increases the current density. This therefore means that the presence of chloride ions speeds up the initiation and propagation of pits. Moreover, it was noted that no repassivation or protection potential (E_pro) was determined for tests with NaCl additions. Comparison of corrosion behaviour of 204SS with 316SS and 304SS in the solution with 0.5 wt.% NaCl is shown in Figure 4 [14].

Figure 4.

Pitting of alloys in concrete with 0.5 wt.% NaCl (adopted from [14]).

The pitting potential for 204SS was measured closer to that of 304SS and 316SS around 700 mV vs. SCE. Non-carbonated solution on its own is alkaline with the pH of 12.6 and without NaCl alloys did not experience pitting corrosion. When NaCl was increased to 5 wt. %, it was observed that 204SS had significantly lower pitting potential. The difference in E_pit obtained at different NaCl concentrations is shown in Figure 5 [14].

Figure 5.

Variation of pitting potential of stainless steels with variation of NaCl [14].

The general interpretation of pitting scans is: If the E_pit value is more noble than E_corr of a tested alloy, then pitting will not occur. Thus if the test solution constituents raises the pitting potential above what is estimated by the polarisation curve, that solution will protect the steel from pitting corrosion. Moreover, if the potential of the test solution is below the measured value of E_pro, pitting corrosion will not take place. [Ca(OH)₂] with 0.5% NaCl as an additive proved to be the safer one to use for concrete solution for all alloys that were tested because higher E_pit values were obtained compared to solutions with 1 wt.% and 5 wt.% NaCl additive [14].

Bautista et al. [14] studies were in agreement with work done by Berke et al. [17], where carbon steel rebars were tested in solutions containing different chloride concentrations. It was observed that pit nucleation became more active with increasing chloride content [17]. It is therefore important to measure the chloride content of the test solution in order to determine if the test alloy will experience corrosion in a certain chloride environment or not.

In order to determine whether certain steel will corrode in a chloride environment, a critical chloride threshold level (CCTL) is calculated [18]. The CCTL is the measure of the chloride level that is enough to cause pitting. According to Bautista et al. [14], the CCTL for 204SS was measured to be 1 wt.% chloride and for 304SS was 5 wt.% [14].

Stainless steel producers have also ran in-solution tests to determine the CCTL of stainless steels and carbon steel in the concrete solution. The CCTL for carbon steel was measured to be less than 0.35 wt. % chlorides and 2.51 wt. % chlorides for 304SS. Garcia-Alonso et al. measured the CCTL of 304SS to be 2 wt.% chlorides in similar test conditions [16, 19].

Fajardo et al. [20, 21] also tested LNASS (4.32 wt.% Ni) against 304SS in a carbonated concrete solution with different chloride concentrations. The cyclic polarisation curves that were obtained showed that LNASS had almost similar pitting behaviour to that of 304SS, with 304SS having a slightly higher pitting potential at all chloride concentrations.

The results that were obtained by Fajardo et al. [20] were in agreement with the results that have been obtained by other researchers [14, 15, 16, 17]. LNASS was expected to have pitting potentials significantly lower than those of 304SS, but further analysis of corroded samples showed that both LNASSs and 304SS had similar behaviour.

Thus, for the current work, we study the general corrosion behaviour of Hercules™ -a LNASS alloy using standard testing methods in comparison to 304SS. Cyclic polarisation technique and immersion tests are used. The objective was to evaluate whether or not the newly developed alloy has corrosion behaviour comparable to that of 304SS and therefore can be a candidate for applications such as reinforcement bars, fasteners and hot rolled stainless steel sheets.

2. Test materials

The chemical composition of test alloys was analysed using spark emission spectroscopy. The typical composition (by weight) of both Hercules™ alloys is: 2% Ni, 16% Cr, and 10% Mn and 0.25% N, with a difference on Mo content where Hercules™ A (no Mo addition) and Hercules™ B (0.5% Mo). The test samples were in the hot rolled and annealed condition, while 304SS (8%Ni, 18%Cr, 1.8%Mn, 0.03%N) was commercially received.

Figure 6 shows the general microstructures of Hercules™ alloys. Grinding, polishing and etching of the sample surface was performed in order to reveal the microstructural features. The Beraha tint etchant was used to colour and highlight the twin boundaries white and the prior austenitic grains brown. The microstructural evaluation showed that Hercules™ alloys have fully austenitic microstructure with an insignificant amount of ferrite (etched dark spots) dispersed on the austenite matrix at room temperature.

Figure 6.

The general microstructure (scale bar 20 μm) of Hercules™ a (left) and Hercules™ b (right).

3. Corrosion tests

Table 3 represent the summary of corrosion tests that were conducted.

3.1 Cyclic polarisation

Cyclic polarisation technique was used to investigate the susceptibility of Hercules™ and 304SS alloys to pitting. This technique is generally used to measure the pitting tendencies of alloys in a given metal-solution system. The experiment starts by applying the potential scan beginning at E_corr and continuing in the anodic direction until there is a large increase in current (e.g. a sustained anodic current density ≥ 10 μA/cm⁻²). The resulting graph is a plot of applied potential vs. the logarithm of current density. When the scan reaches the programmed current density limit value, it reverses and begins scanning in the negative direction. An example of a typical cyclic polarisation plot is shown in Figure 7 [22, 23].

Figure 7.

A typical cyclic polarisation plot [22].

The E_pit is the potential at which stable pits initiate and propagate as applied potential increases. E_pro is the potential below which no initiation of pits will occur. Pits that form above E_pit will eventually repassivate below E_pro hence the potential is also referred to as the repassivation potential. Both E_pit and E_pro are used to explain the kinetics of pitting and repassivation [23].

The size of the hysteresis loop can give a rough indication of the extent of propagation of initiated pits. The longer time it takes for pits to repassivate, the bigger the hysteresis loop. This imply that formed pits are severe and stable. In some cases, pits show no tendency to repassivate by the hysteresis loop closing at a potential less than E_corr [23].

In our work cyclic polarisation technique has been used to evaluate the corrosion behaviour of Hercules™ alloys and conventional 304SS. The procedure outlined in the ASTM G61 standard [24] was used to conduct pitting corrosion tests. The 12 mm diameter disc shaped samples were prepared from each alloy. Three test solutions were used, 3.56 wt. % NaCl, reduced concentration to 1 wt. % NaCl and 5 wt. % H₂SO₄.

Test samples were ground to 600 grit SiC paper finish. The corrosion test conditions were set as shown in Table 4. An ACM potentiostat was used to apply the potential on the working electrode and to measure the current flow between the counter electrode and the working electrode. A graphite rod was used as a counter electrode and a saturated calomel electrode (SCE) was used as a reference electrode. The scan was set to increase the potential stepwise starting from the corrosion potential to 1200 mV. Duplicate scans were performed.

Technique	Standard testing procedure	Solution
Cyclic polarisation	ASTM G61	3.56 wt.% NaCl and 1 wt.% NaCl
Cyclic polarisation	ASTM G61	5 wt.% H2SO4
Immersion	ASTM G31	5 wt.% H2SO4
Immersion	ASTM G48 A	6 wt.% FeCl3.6H2O

Table 3.

The summary of corrosion tests.

Parameter	Conditions
Temperature	Room temperature
Scan rate	10 mV/min
Reverse current density	5 mA/cm^2

Table 4.

Cyclic polarisation test parameters.

The corrosion potential (E_corr), pitting potential (E_pit) and protection potential (E_pro) were measured from the cyclic polarisation curves and they were analysed using the Origin program. The corroded coupons were further taken for analysis under stereomicroscope at 50X magnification. Information about the extent of passivation region, stability of passive state and the ability of tested alloys to spontaneously passivate in a given environmental system was obtained. The corrosion rate (CR) was calculated using Eq. (7) [25]. The critical current density (i_corr) was obtained by extrapolating the Tafel region of the cathodic and the anodic regions of the polarisation curve.

$$CR=0.011\times i_{\mathit{corr}}\times 1000\left(\mathit{mm}/y\right)$$E7

3.2 Immersion tests

Immersion tests were performed in 5 wt. % H₂SO₄ for 10 days. Tests were done following the guidelines outlined in ASTM G31 [26]. 25 by 50 mm² coupons of Hercules™ A, Hercules™ B and 304SS were ground to 600 grit SiC paper finish. Test coupons were left for a minimum of 24 hours to allow them to passivate in air, in order to spontaneously form the passive layer possibly disrupted by sample preparation [27]. Test coupons were then immersed in a test solution at room temperature. The amount of solution in the beaker was calculated to a ratio: 0.20 ml/mm² [26].

ASTM G48-Method A was used for immersion tests in FeCl₃.6H₂O. The temperature was maintained at 26°C ± 2°C. The 25 by 50 mm² coupons were also prepared. Test coupons were immersed in 6 wt. % FeCl₃.6H₂O for 72 hours. Then coupons were removed and rinsed with water and ethanol. Corroded coupons were then reweighed and examined for pitting [15]. Pit depth and density were measured by visual examination of images taken at 10X magnification using the Light microscope.

Mass loss due to corrosion was measured and corrosion rate calculated using Eq. (8) [26].

$$\mathrm{Corrosion}\mathrm{rate}=\frac{\mathrm{K}\times \mathrm{W}}{\mathrm{A}\times \mathrm{T}\times \mathrm{D}}$$E8

Where:

$\mathrm{K}$= $8.76\times {10}^4$ for corrosion rate in millimetres per year (mm/y), $\mathrm{W}$ = mass loss in grams, A = area in cm², T = time of exposure in hours, D = density in g/cm³.

4. Corrosion performance

4.1 Localised corrosion mechanisms in sodium chloride

Polarisation scans are shown in Figure 8. All test alloys performed poorly in 3.56 wt. % NaCl, with just a slightly higher E_pit around 72 ± 19.4 mV. SCE for 304SS. However similar corrosion behaviour was observed for all test alloys. The E_pit values for Hercules™ A and Hercules™ B were measured to be almost similar at more electronegative potentials. E_pro was measured at potentials below the pitting potential E_pit for all test alloys, which is an indication of severe pitting corrosion as observed by higher anodic currents at the E_pit, that is, there was a sudden increase of current at lower potentials. The Tafel extrapolation showed that E_corr values were almost similar for all test alloys and presumptuously this could be expected because of similar corrosion constants K.

Figure 8.

Polarisation scans of austenitic SSs in 3.56 wt. % NaCl.

In 1 wt. % NaCl the E_pit was measured at 219 ± 17.5 mV. SCE for Hercules™ A which is more positive than the one obtained from 3.56 wt. % NaCl test, as shown in Figure 9. The E_pit for Hercules™ B was measured at 579 ± 0.8 mV. SCE. There was a significant difference in E_pit values for each alloy due to reduction of NaCl concentration. All test alloys scans showed incapability to repassivate as the E_pro was measured below E_corr.

Figure 9.

Polarisation scans of austenitic SSs in 1 wt. % NaCl solutions.

The behaviour of 304SS and Hercules™ in 1 wt. % NaCl can be explained by considering the effects of Cr since LNASSs have lower Cr content than 304SS. Ujiro et al. [28] has studied the effect of Cr additions in the corrosion behaviour of austenitic SSs in NaCl solution. It was noted that increasing Cr reduced the rate of increase of current density above E_pit, that is, the size of hysteresis loop was reduced. In the alloy with less Cr content, the hysteresis loop will show higher anodic currents than one with higher Cr contents as observed with 304SS and LNASSs comparisons. This demonstrates the inhibitory effect of Cr on initiation of localised corrosion.

Typically, in stainless steels a Cr-rich passive oxide film will form when exposed to an oxidising environment. It has been previously reported that the film consists of layers of different compositions formed due to dissimilar conditions at the metal-oxide interface and oxide-environment interface of the passive film. Depending on the composition of an alloy, it has been observed that the metal adjacent to the passive film is enriched with Ni, while the passive film itself consists of Cr-rich oxide inner layer and an Fe-rich hydroxide outer layer. The Ni-rich layer is formed because of diffusion of Fe and Cr out the bulk metal into the oxide layer. Figure 10 shows that Cr and Mo are more significant in the formation of oxide layer because of their high affinity for oxygen. On the other hand, Ni does not participate in the formation or stabilisation of a passive layer [29].

Figure 10.

Contents of optimal passive layer formed in NaCl (adopted from [29]).

Moreover, Ujiro et al. [28] studied the effect of Mo in the corrosion behaviour of austenitic SSs containing 26 wt. % Cr and varying Mo contents from 0 to 4 wt. %. It was observed that addition of Mo decreased the anodic current density; that is, the current measured before E_pit or before the onset of pitting. Extended passive region was evident in Hercules™ B compared to Hercules™ A and 304SS. However, both Hercules™ alloys show similar hysteresis loop behaviour. The hysteresis loop closes at potentials lower than E_corr.

To further substantiate polarisation results micrographs of corroded samples surface were evaluated. The micrographs of corroded samples surface are shown in Figure 11. It is evident from the optical micrographs that all samples experienced crevice corrosion underneath the crevice washer and as a result E_pro was measured at potentials below E_pit.

Figure 11.

Micrographs of corroded coupons (scale bar 1 mm) showing crevice corrosion in 3.56% NaCl solution for: Hercules™ a (left), Hercules™ B (middle) and 304SS (right).

In a crevice metal-solution system, there lies a crevice critical solution, of which a minor shift in potential gradient changes the corrosion behaviour of an alloy from passive to active. The longer it takes for an alloy to reach that crevice critical solution defines the resistance of an alloy to corrosion [30]. Amongst other factors, crevice critical solution is mainly affected by alloy composition. The cationic metal species react with water to generate acidity in the crevice region. In stainless steels, the typical chemical reaction to take place under a crevice is:$F{e}^{2+}+2H_{2}O\to Fe{\left(OH\right)}_2+2H$⁺.

The reaction involves other alloying elements such as Cr, Mn and Mo. Mo was added in Hercules™ B with an expectation that it will inhibit the localised corrosion reactions by lowering rate of generation of acidic hydrogen and consequently lower the corrosion rate. However, the reaction rate is controlled by two different kinetic phenomena. The first is charge transfer or activation control. In this case, the reaction rate is controlled by the size of the driving force, which is either hydrogen evolution or water reduction reaction. As the driving force increases, so does the reaction rate [30].

Other mechanism controlling the rate of reaction is mass transfer through the electrolyte to the electrode surface, that is, oxygen reduction reaction. Since the reaction rate is controlled by diffusion, it cannot increase indefinitely as the driving force increases. Instead, the current reaches a maximum current density which is itself a function of the concentration of the species of interest in the solution as well as its diffusivity. Once the rate for a particular reaction has reached its limiting value, further increases in driving force will not result in any additional increase of the reaction rate. For this current work, addition of Mo was expected to reduce rate of propagation of corrosion by slowing down diffusion at the crevice. However, the concentration of added Mo (0.5%) was not sufficient to lower the corrosion rate in Hercules™ B and thus, similar corrosion behaviour was observed for all test alloys [29, 30]. Perhaps, higher amount of Mo should be added, but bearing in mind cost-related issues. Micrographs are in agreement with the polarisation scan measurements, which showed similar corrosion behaviour.

The difference was observed with micrographs obtained from 1 wt. % NaCl test, as shown in Figure 12, which showed pitting corrosion within the crevice area. Ujiro et al. [28] explained that crevice corrosion can occur either by depassivation or pitting. Depassivation type occurs by corrosion of surface underneath the crevice, due to pH drop and extensive destruction of the passive film under the crevice washer (observed with 3.56 wt. % NaCl). Pitting type occurs by pitting inside the crevice area as a result of chloride concentration increase in the inner solution.

Figure 12.

Micrographs of corroded coupons (scale bar 1 mm) showing pitting corrosion in 1 wt. % NaCl solution for: Hercules™ A (left), Hercules™ B (middle) and 304SS (right).

Ujiro et al. [9] also investigated the relationship between the type of corrosion and the E_corr. It was observed that ferritic alloys which corroded by depassivation had lower E_corr than the ones that corroded by pitting. Crevice corrosion by depassivation is related to E_corr because it involves an intensive metal dissolution at lower potentials and not just a single pit. Similar to the current work, E_corr values obtained from 1 wt. % NaCl were measured to be higher than those obtained from 3.56 wt. % NaCl. This means that 1 wt. % NaCl did not contain a critical chloride concentration required to reach crevice critical solution for the complete dissolution of a metal underneath the crevice. Hence, instead of depassivation, pitting was observed. The solution in the pit is more aggressive once pitting has started as shown by an increase of current density. In some alloys, once pitting initiates, propagation is faster and it becomes difficult for a formed pit to repassivate as observed in the CP scans. However, the onset of pitting was delayed for Hercules™ B in 1% wt. % NaCl, as indicated by an extended passive region than in other test alloys, owing to the inhibiting effect of Mo.

4.2 Severe pitting in ferric chloride

The mass loss of test alloys was measured and the corrosion rate due to pitting was calculated. No significant difference in mass loss of all test alloys was observed, with corrosion rates measured to be 11.8 ± 0.2 mm/y for Hercules™ A, 12.6 ± 0 mm/y for Hercules™ B and 14.1 ± 0.1 mm/y for 304SS.

This behaviour is similar to results that were obtained by Bergstrom et al. [8] when 201SS and 304SS were tested in 6 wt.% FeCl₃ for 72 hours. Both alloys showed a corrosion rate of 0.0228 g/cm² and similar pit depth of 0.0762 mm [8].

The 6 wt. % FeCl₃.6H₂O solution is generally used as a test for localised corrosion for accelerated tests. Therefore, test alloys were immersed in the solution for 72 hours as recommended in the ASTM G84 standard. However, this test solution is used to simulate a very rough composition environment within a localised corrosion site in a stainless steel. It can be very aggressive for low-alloyed steels such as 304SS and 201SS. This has also been confirmed by tests that were conducted by Ujiro et al. [28]. Alloys with higher Mo and Cr contents (above 26 wt.% Cr and 4 wt. % Mo) showed more corrosion resistance than the ones with the compositions approximately similar to that of 304SS [28].

Furthermore, 6 wt. % FeCl₃.6H₂O serves as a chemical potentiostat by forming the Fe³⁺/Fe²⁺ redox couple which has an approximate potential of 450 mV with high chloride concentration. The solution is highly acidic with a pH of 1.44, which is enough to create a large current without a need to polarise the specimen as with the electrochemical tests [27]. From the electrochemical tests, it has been established that chloride concentration from neutral 1 wt. % NaCl was enough to cause a E_pit that is less than 450 mV. The 6 wt. % FeCl₃.6H₂O solution has higher chloride concentration than NaCl, hence Hercules™ alloys and 304SS corroded aggressively. The high potential of the test solution almost guarantees that the pitting potential of each alloy was exceeded. The acidic nature of FeCl₃.6H₂O also inhibits repassivation and lowers passive film strength by cathodic reactions that occur on the surface of the sample via Fe³⁺/Fe²⁺ ions [27].

The difference in the individual pit morphology was observed. Hercules™ A had irregular shaped pits and with high depth, whilst Hercules™ B had wide and round shallow pits. This means that pitting propagated quicker in Hercules™ A than in Hercules™ B. Figure 13 shows the one of the deepest representative pit that was observed in Hercules™ A after 72 hours of immersion in FeCl₃.6H₂O, along with the pits measurements. Although pit density of all test alloys was almost similar, Hercules™ A showed severe pitting because of larger pit opening.

Figure 13.

Micrographs (scale bar = 1 mm) and representative pit of Hercules™ A in FeCl₃.H₂O.

Average	1557 ± 206	291 ± 77	1587 ± 199
Max.	1862	374	1881
Min.	1259	165	1289
Range	603	208	592
No.	Width[μm]	Height[μm]	Length[μm]
1	1424	372	1472
2	1676	165	1684
3	1563	374	1607
4	1862	266	1881
5	1259	275	1289

Figure 14 shows that the size of pit opening observed in Hercules™ B is smaller than that of Hercules™ A. Figure 15 shows the extent of pitting that was observed for 304SS. Overall, the pit evaluation proves that FeCl₃.H₂O is an aggressive solution for testing LNASSs and 304SS. Even the addition of 0.5 wt. % Mo for Hercules™ B is not enough for it to be used for applications in aggressive environments.

Figure 14.

Micrographs (scale bar = 1 mm) and representative pit of Hercules™ B in FeCl₃.H₂O.

Average	997 ± 332	114 ± 7	1005 ± 328
Max.	1388	129	1394
Min.	379	109	395
Range	1008	20	999
No.	Width[μm]	Height[μm]	Length[μm]
1	379	109	395
2	1073	110	1079
3	1073	110	1079
4	1073	110	1079
5	1388	129	1394

Figure 15.

Micrographs (scale bar = 1 mm) and representative pit of 304SS in FeCl₃.H₂O.

Average	1552 ± 191	274 ± 76	1592 ± 150
Max.	1863	372	1772
Min.	1324	145	1325
Range	539	227	447
No.	Width[μm]	Height[μm]	Length[μm]
1	1324	324	1576
2	1678	145	1683
3	1556	372	1604
4	1836	255	1772
5	1368	275	1325

4.3 Passivity behaviour in sulphuric acid

A stainless steel can be considered resistant to uniform corrosion in a particular environment if the corrosion rate does not exceed 0.1 mm/y [19] . In the current work all test alloys demonstrated resistance in 5 wt. % H₂SO₄ as shown by polarisation curves in Figure 16.

Figure 16.

Cyclic polarisation scans of austenitic SSs in H₂SO₄.

All alloys displayed the ability to passivate spontaneously in 5 wt. % H₂SO₄. Lower E_corr for Hercules™ A can be attributed to the kinetics of passivity for stainless steels, whereby if the cathodic reaction becomes more dominant at lower potentials and thus remaining in the active region will result in favourable conditions for anodic reactions to overtake the redox reactions at lower potentials. Polarisation scans of some stainless steel will show these cathodic reactions by the presence of anodic current peak, which is an indication of non-uniform passivity due to less OCP test times. Thus, it can be observed that with all test alloys, if given enough time to form a passive layer during OCP tests, the anodic peak current can be avoided and alloy are fully passivated with no disruption of the protective film when exposed to H₂SO₄. Therefore, H₂SO₄ is considered a safe environment for LNASSs and so is 304SS.

Furthermore, the absence of a hysteresis loop is an indication that tested alloys did not undergo any type of localised corrosion even though an artificial crevice was introduced in each sample. The presence of an artificial crevice creates passivation current (i_pass) (Current density at the passive region) that is higher than i_corr. However, the conditions were not sufficient to activate sample surface for formation of pits or cause crevice corrosion since the test solution did not contain chlorides [31].

The corrosion rates calculated from polarisation curves for Hercules™ B and 304SS were comparable at 0.001 ± 0.008 mm/y and 0.002 ± 0.004 mm/y, respectively. The corrosion rate of Hercules™ A was measured to be 0.016 ± 0.029 mm/y, which is a magnitude higher than 304SS and Hercules™ B.

However, the corrosion rate of test alloys in 5 wt. % H₂SO₄ was calculated to be higher in the immersion tests. The corrosion rate of Hercules™ A was calculated from the mass loss incurred to be 1.863 ± 0.028 mm/y. Hercules™ B and 304SS had the corrosion rate less than 0.100 ± 0.020 mm/y, with 304SS having the lowest. It is often assumed that corrosion rate of stainless steels is linear with the function of time during immersion tests, but this is not always true for some stainless steels immersed for a longer time [32].

Hercules™ A was observed to react aggressively for the first 24 hours at a presumably higher corrosion rate but the vigorous reaction decreased as days progressed. Hercules™ B did not react aggressively in the beginning and throughout the entire exposure time [32].

Based on observations made in the current project, the solution in which Hercules™ A was immersed had a dark bluish precipitates after a few hours of immersion and the reaction was more aggressive than Hercules™ B. The solution with Hercules™ B and 304SS did not show any change of colour and the reaction was less aggressive. Therefore, it can be established that passive films formed by stainless steels may be broken down in a prolonged exposure period and hence higher corrosion rates are obtained in immersion tests than polarisation tests. Although mass loss is negligible after a certain period it can still add up to the final mass loss measurement. Electrochemical tests took less than 4 hours to obtain a complete scan. Therefore, the corrosion rate measured tend to be less than that obtained from immersion tests [27].

5. Conclusion

Hercules™ alloy was developed with an aim to reduce cost of austenitic stainless steel and for use as an alternative to 304SS with target applications being reinforcement bars and fasteners, taking advantage of higher strength of Hercules™. In order to fast forward acceptability for this alloy, additional applications were proposed where corrosion resistance is of importance such as rolled sheets for manufacturing of water tanks and other corrosion resistant products.

The general corrosion behaviour of Hercules™ alloy and 304SS have been evaluated using cyclic polarisation technique and immersion tests. Polarisation curves of NaCl tests showed that in higher chloride content of 3.56 wt. %, all test alloys corroded severely. When the concentration was reduced to 1 wt. %, the passive region of Hercules™ B was significantly extended, making this alloy more resistant to initiation of localised corrosion. However, once pitting initiated, it propagated faster, which was observed by the size of hysteresis loop and confirmed by visual examination of corroded coupons. This nullifies the anticipated effects of Mo and this has been explained using thermodynamics of corrosion. Immersion tests in FeCl₃ also proved to be extremely aggressive making these alloys not suitable for use in aggressive chloride environments such as swimming pools and sea water. Additional polarisation and immersion tests in H₂SO₄ showed that all test alloys had an ability to spontaneously passivate in this environment, thus making Hercules™ useful in such reducing environments at room temperatures.

It is however, arguable whether or not Hercules™ alloy (with 0.5% Mo) can be used as a substitute for 304SS for when Ni prices are high because the tests conducted here did not provide sufficient rigour to validate the corrosion resistance of Hercules™ against 304SS. Thus, this leaves a gap for further work to be conducted. The work presented here simply outlines the comparative behaviour of Hercules™ to 304SS and more corrosion tests via in-situ environments should be performed in order to qualify Hercules™ alloy. Temperature, composition and test parameters are other factors that can be investigated further in order to perform application-based tests. Evidently, the solutions outlined by the corrosion test standards are too aggressive for these alloys. Building from this work, a more suitable test procedure can be developed and thus, providing newly developed LNASSs a chance to demonstrate their corrosion resistant strengths, as already observed with 1 wt. % NaCl and H₂SO₄ tests.

Acknowledgments

Authors would love to thank:

University of Cape Town (Centre of Materials Engineering).

Professor R. Knutsen for the overall supervision provided throughout the entire project.

Mintek-Advanced Materials Division (Physical metallurgy group).

Ms. M. Smit for her assistance with executing electrochemical corrosion tests in the AMD corrosion laboratory.

Special thanks to Mintek for funding the project.