Corrosion depth

Fig. 4-16 presents the result of corrosion depth after 120 days after applying Eq.

(4.2). Note that these data are derived from the full length of the steel bars. When recovered from mortars, the bars actually showed significant rusting only at their endpoints, portions covered with epoxy resin. This means that the data in Fig. 4-16 is based primarily on the corrosion depth outside of the mortar, which is not correct because the target area is within the mortar.

Figure 4-16. Uncorrected corrosion data considering the whole length (200 mm) of steel bars.

Rust stain coverage (%) E corr (mV vs. CSE)

Corrosion rate (μm/day)

Average corrosion depth based on Eq. 4.2(μm)

Exposed time (days)

0 30 60 90 120 150

Cracked

CT NT

PC

0 10 20 30 40 50 60

0 30 60 90 120 150

No crack

CT NT

PC

E_corr= 166.5r - 364.0 R² = 0.09 -600

-500 -400 -300 -200 -100 0

0.0 0.2 0.4 0.6 0.8 1.0

E_corr= -21.1c - 319.2 R² = 0.08

0 1 2 3 4 5 6

10% corrosion Uncertain 90% corrosion Severe

Steel Uncracked Cracked

PC ■

CT ▲

NT ●

100

Figure 4-17. Details of cutting the bars to lengths within the mortar.

To consider only the area embedded in mortar, Eq. (4.2) was modified by cutting the ends of the bars, retaining only a length of about 140 mm, as described in Fig. 4-17.

Then, the actual length and mass of newly cut bars were measured, and Eq. (4.3) is applied as the true estimate of corrosion depth. The whole expression similarly assumes that thickness loss is uniform over an exposed area, which in this case is the lateral surface area of the newly cut bar. Such assumption is reasonable for comparison purposes.

The result of analysis is presented in Fig. 4-18.

where = average thickness loss due to corrosion of the steel bars inside the mortar (μm); = initial mass of original length (g); = actual length after cutting (mm);

= original length before cutting; = final mass of the cut length (g); = average metal density (g/cm³); = diameter (mm).

𝜋 (4.3)

After cutting Before cutting

L_c 140 mm, m_fc

Target area

Rust happened mainly outside the mortar

L_m 150 mm L_o 200 mm, m_io, m_fo

101

Figure 4-18. Change of corrosion depth with time. Error bars represent one standard deviation from the mean.

First, it can be observed that the behavior of steels in cracked and uncracked mortars are almost similar, consistent with the result of corrosion potential. This was again due to limited amount of oxygen and chloride that reached the rebar surface, causing small variation in corrosion depth at the end of the test.

Additionally, even after correction NT remains slightly higher than other types in terms of corrosion depth. Fig. 4-19(a) was prepared to compare these data with those obtained in IM-B case (permanent immersion in alkaline solution) of the liquid test because the steel bars appear to have similar surface conditions throughout the entire test period (see Fig. 4-19(b)).

Average depth of corrosion based on Eq. 4.3(μm) 0 10 20 30 40 50 60

0 30 60 90 120 150

No crack

CT NT

PC

0 30 60 90 120 150

Cracked

CT NT PC

Exposure time (days)

102

Figure 4-19. (a) Comparison of average corrosion depths in IM-B case (solution containing 3.5% NaCl and NaOH with pH = 13.1) and mortars; (b) surface conditions of steels in mortar and liquid phase; (c) schematic illustration of ionic distributions near the rebar level in liquid and mortar phase.

It becomes apparent that the sequence of steel type is not consistent from solution to mortar. That is, corrosion depth in mortar is arranged from highest to lowest as NT >

PC ≥ CT, while the difference is not as apparent in liquid phase. This difference results from different ionic distribution present in the each phase, as schematically illustrated in Fig. 4-19. In liquid phase, high amount of chloride ions are uniformly distributed near

(a)

(b)

Average corrosion depth(μm)

Steel type

(c)

0 10 20 30 40 50 60

PC CT NT

120 days 90 days 60 days

Solution (IM-B)

PC CT NT

120

60 90

Cracked mortar

PC CT NT

120

60 90

Uncracked mortar

120 days in mortar 120 days in simulated solution

Uncracked Cracked IM-B IM-A

PC

CT

NT

steel bulk

mortar

pores, cracks

dissolved oxygen

Ni Fe

e^–

steel bulk solution

e^–

Fe H₂O

103

the rebar level. As a result, corrosion gets deeper and larger, reducing the effect of added elements. Meanwhile, chloride, moisture and oxygen exist only in pores and discontinuities present in steel-mortar interface. Because of these local sites, corrosion becomes more localized, and thus, the effect of added elements becomes more pronounced in mortar. From the above discussions, it appears, therefore, that Ni-type WS is relatively susceptible to penetration when rust is still developing and only covers small areas. Localized corrosion occurs when anodic areas, where metal dissolution takes place, is small relative to cathodic areas, which in this case are those regions below the mortar or in the vicinity of dissolved oxygen. Ni, as an element, is known to be more cathodic, i.e. has greater tendency to gain electron, than Fe or Cr. So, the inclusion of Ni is thought to increase the cathodic areas and, therefore, enhance the rate of cathodic reaction initially provided by dissolved oxygen. Therefore, an improvement of localized penetration in NT is probably caused by slightly higher rate of iron release in active (anodic) areas as it balances the increase of said cathodic reaction (Fig. 4-19(c)).

Complementary studies derived from atmospheric settings have also shown that localized depths on WS appear in greater number at early stage ^15,16.

However, it should be noted that rusts that formed were only few and localized.

The computed values of corrosion depths, as depicted in Fig. 4-19(a), were merely less than 30 μm, which is deemed significantly small relative to the values when rust is fully grown and uniform. It follows that the behavior of high-Ni type WS bars in mortar may still be considered practically the same as those of conventional steels.

Furthermore, since corrosion are still localized, the result presented in this study does not take into account the fact that in real corroding structures, rust covers much larger areas. It should also be noted that Ni-type WS obtains its resistance from the appearance of a dense and uniform rust layer. It is possible that that the initially high penetration in NT may stabilize as the protective rust appears after a sufficient period has passed. Since a uniform layer has not yet formed within the period and scope of conditions considered in this study, the decrease in corrosion curve by Ni addition could not be established as of this time. In the future, the possibility of forming the protective rust by using much longer time of exposure should be further explored.

104