You are on page 1of 14

SAND2016-10742C

The Influence of Salt Loading Density on the Atmospheric Corrosion of Aluminum

Rebecca F. Schaller Jason M. Taylor


Sandia National Laboratories Sandia National Laboratories
1515 Eubank SE 1515 Eubank SE
Albuquerque, NM, 87123 Albuquerque, NM, 87123
USA USA

Eric J. Schindelholz
Sandia National Laboratories
1515 Eubank SE
Albuquerque, NM, 87123
USA

ABSTRACT

Corrosion of aluminum and aluminum alloys under atmospheric exposure has been well documented
for outdoor conditions. While these studies expose the effects of environmental severity they do not
explicitly establish the dependence of corrosion rate on salt loading. Accelerated laboratory studies
have shown that initial corrosion rates are generally higher with higher salt loadings, but, over time
corrosion appears to effectively stifle for low loadings of NaCl (<100 µg/cm2) under fixed humidity
conditions. This has previously been attributed to the stability or passivation of the surface that is pH
and, in turn, CO2 dependent. Another possible explanation could be the gettering of NaCl by corrosion
product leading to surface drying and depletion of the corrosion aggressor. This paper explores the
effects of selected NaCl loading densities vs. exposure time of Al 1100 at both the macro and micro
scale to illuminate the possible mechanisms leading to corrosion stifling. Through this work, an
understanding of the relationship between corrosion in atmospheric systems versus the variation of a
specific environmental severity factor, NaCl loading density, will be further developed.

Key words: Droplets, Corrosion Stifling, Atmospheric Exposures

INTRODUCTION

Aluminum and aluminum alloys have been extensively studied under field atmospheric exposure
conditions.1-8 These studies characterized effects of environmental severity and exposure factors that
may control or enhance corrosion rates. One severity factor, salt deposition rate, results in higher
corrosion rates with higher salt deposition rates.7, 9 However, as atmospheric field exposures are open
to a wide variety of environmental severity factors, such as salt deposition rates, SO2, UV, Ozone, pH,
temperature, humidity, time of wetness, and other industrial pollutants, understanding the reliance of
corrosion on one single severity factor often becomes convoluted.3, 4, 6-8, 10-13 Simplified laboratory
studies enable establishment of the corrosion rate under more controlled environments and the same
correlative relationship between salt loading and corrosion rate has been reported in the fewer
laboratory studies on this topic.10, 14-17 Figure 1 exemplifies these trends where mass loss of 4N pure
Al, an Al-Mg alloy (AA 6061), and Al 1100 is correlated with salt deposition rate for field and lab
accelerated exposures.10 For all exposures and materials, corrosion rates increased with increasing
deposition rates, except for AA 6061, which showed a stifling in corrosion rate in laboratory exposures
above 100 mg/m2d (10 µg/cm2).

Figure 1. Corrosion rate (mass loss, g/m2) vs chloride deposition rate, mg/m2d, for Al samples
exposed to varying salt deposition rates. Data compiled from Dan et al (1 year field and 5 day
laboratory exposures).10

While these studies expose some environmental severity effects and indicate factors that may be
controlling corrosion rates, they do not provide insight into the dependence of corrosion rate on salt
loading. Other laboratory studies have elucidated a phenomenon for atmospheric exposures of
aluminum, in that corrosion rate diminishes with increased exposure time. Samples of UNS A91070
(99.8% Al) exposed to humid air with 14 or 70 µg/cm2 coverage of NaCl displayed corrosion stifling at
as short as 100 h of exposure.14, 15 When the effect of CO2 was considered in this system, it was
shown that atmospheric concentrations (350 ppm) lowered the corrosion rates. Samples printed with
the same salt loading densities exposed to humid air with 350 ppm CO2 exhibited an order of
magnitude lower corrosion in terms of mass gain, than samples exposed in environments with less than
1 ppm CO2.14 A thorough explanation to the stifling effect of corrosion over time under fixed NaCl
loading conditions has not been fully developed. Possible theories have been put forth, including, the
stability or passivation of the surface that shuts down the cathodic region, which is pH and, in turn, CO2
dependent 14, 15 While results to date demonstrate a strong correlation between corrosion stifling and
CO2 concentration, data is insufficient to pinpoint the underlying mechanisms. Also, the hypothesis that
increased CO2 results in the stifling of corrosion suggests that corrosion would be independent of NaCl
loading density, which current results do not support. Another possible explanation for stifling could be
the gettering of NaCl by corrosion product leading to surface drying and depletion of the corrosion
aggressor.

This paper explores the effects of selected NaCl loading densities vs. exposure time of UNS A91100
samples with isohumidity tests at a range of set RH. Pure Al samples were selected for study to
establish a simplified corrosion system to understand atmospheric mechanisms which could then be
expanded to common engineering alloys. Both macro and micro scale studies were undertaken to
elucidate the possible mechanisms leading to corrosion stifling under fixed NaCl loading conditions.
High throughput inkjet salt printing was applied to characterize corrosion rate over time via coupon-level
studies. Gravimetric measurements were utilized to determine mass gain/loss and provide indication of
conditions leading to corrosion stifling. Micro scale explorations with in-situ time lapse imaging of
micro-pipetted droplets exposed to iso-humidity conditions helped to provide basic indication of
electrolyte coverage and stability during exposure. Post-exposure analysis provided characterization of
the evolution of electrolyte chemistry and corrosion product mineralogy. Through this work, an
understanding of the relationship between corrosion in atmospheric systems versus the variation of a
specific environmental severity factor, NaCl loading density, will be further developed for prediction of
Al corrosion including the effects of stifling.

EXPERIMENTAL PROCEDURE

Materials

Test samples were comprised of 3.8 cm2 UNS A91100 coupons. All coupons were ground with 1200
grit silicon carbide paper and then polished using 1 µm diamond paste. After polishing, the coupons
were cleaned with 18.2 MΩ deionized water and isopropyl alcohol and then dried in a compressed
nitrogen stream.

Salt Loading Exposures

Sample sets of Al coupons were loaded on one side with 10, 60, and 150 µg/cm2 of NaCl particles
using an inkjet printing method.18 Salt loading patterns for each of these loadings are exemplified in
Figure 2. After loading, samples were exposed in iso-humidity chambers at 98% RH and 21oC from 1
day to 4 weeks. Post-exposure, coupons were analyzed for mass gain using a calibrated microbalance
with a repeatability of ± 10 µg. Samples were massed, a minimum of three times each, after drying in a
nitrogen desiccator (< 10%RH) for one hour. Drying was carried out to minimize the mass contribution
of free water on the samples present during the 98%RH exposure.

A subset of the exposed samples was additionally analyzed by optical microscopy and grazing
incidence X-ray diffraction (GIXRD) to characterize surface conditions resulting from exposure. GIXRD
was carried out at a grazing angle of 0.5o.

Once mass gain measurements and surface characterization were carried out, mass loss of each
coupon was measured after corrosion product removal following the ASTM G-1 cleaning procedure in
nitric acid.19 An acid cleaning time of two minutes was found to minimize etching of the sample base
metal while still removing all visible corrosion product. Mass loss was determined by weighing the
samples with removed corrosion product on the microbalance compared to initial coupon weight
measurements.
a) b) c)

d)

200
Salt Deposition Analysis:
175 Conductivity
Salt Loading (g/cm )
2
150 Mass Gain

125

100

75

50

25

0
0 1 2 3 4 5 6 7 8 9
NaCl Concentration (M)
Figure 2. Example images of salt printing for a) 10, b) 60, and c) 50 µg/cm2 NaCl loadings and d)
calibration of salt loading with conductivity and mass gain measurements.

In-Situ Droplet Exposure

Droplets of 4 M NaCl were deposited on Al coupons at 80% RH using a capillary microinjection system
Figure 3. A minimum of 10 droplets with average diameters of 200 to 250 µm and spaced a minimum
of 250 µm apart were deposited on each sample. Samples were moved to a smaller chamber, within
10 min of droplet loading, set at 98% RH by saturated K2SO4 and 21o C. The equilibrium concentration
of NaCl solution is expected to be 0.6M, meaning the droplet increased on average 33% in volume as
they equilibrated to 98% RH conditions.20 The droplets were timelapse imaged through windows in the
smaller chamber from both a top down perspective and a side perspective of the droplets (Figure 3)
while temperature and humidity were simultaneously recorded during the 12 day exposure. Post-
exposure sample characterization was carried out using the methods described above for the salt film
experiments as well as scanning electron microscopy (SEM), energy dispersive spectroscopy (EDS),
and fourier transform infrared spectroscopy (FTIR). SEM and EDS were carried out at an accelerating
voltage of 10 keV and a working distance of 10 mm. A Bruker Hyperion IR microscope was used for
total reflection FTIR for corrosion product identification. Samples were again cleaned using the ASTM
G-1 procedure and imaged post-cleaning using optical profilometry to determine the extent of corrosion
damage.
a) b)

Figure 3. a) Schematic of micro-injector system and b) side image of droplet being deposited by
the micro-injector.

RESULTS

Salt Loading Exposures

Results of the salt loading coupon exposures were consistent with previous findings showing higher
mass gain with increased loading and a strong bend over in mass gain at all loadings (Figure 4-a).14, 15
Bend overs were observed at times greater than 150 hours, and are compared to results from Blücher
et. al in Error! Reference source not found.-a.14 One explanation for the discrepancy in initial time to
stifling may be due to the nature of the measurements, as Blücher et. al acquired wet mass gain, and
measurements taken here were of dry samples. Changes to electrolyte volume with exposure time
may affect the mass gain measurements if measured while wet.

Further analysis of the Al coupons through gravimetric measurements of mass loss in Figure 4-b, with
corrosion product removed, confirmed the mass gain measurements. The mass loss results display a
consistent trend with an increase in corrosion with increasing initial salt loading. However, the stifling in
corrosion with time is not as clearly defined, especially at low loading densities as mass loss
measurements are close to the detection limit of the mass loss procedure. This could be due to an
introduction of increased possible error into the mass loss measurements from the nitric acid cleaning
procedure.
a) b)

500
Salt Loading:
1000 Salt Loading:
2 10 g/cm2
10 g/cm
2 400 60 g/cm2
60 g/cm
150 g/cm2
Dry Mass Gain (g/cm2)

800 150 g/cm2

Mass Loss (g/cm )


2
1
Johansson et al. :
2 300
10 g/cm
600 70 g/cm
2

200
400

100
200

Detection Limit
0 0
0 5 10 15 20 25 30 35 0 5 10 15 20 25 30

Time (days) Time (Days)

Figure 4. a) Average dry mass gain vs. exposure time at 98% RH for Al 1100 printed with 10, 60
and 150 µg/cm2 NaCl. Results from Johansson et. al displayed on graph for comparison. b)
Average mass loss vs. exposure time for the same Al 1100 samples. Error bars are included (not
visible at scale shown) for one standard deviation. The detection limit for mass loss is based
on unexposed samples submitted to the same nitric acid cleaning procedure as exposed
samples for corrosion product removal.

Aluminum coupons exposed with initial salt loadings at constant RH displayed an increase in corrosion
product through optical microscopy for both increasing time exposures and increasing salt loading
(Figure 5). GIXRD results for exposed samples only exhibited the presence of Dawsonite
(NaAlCO3(OH)2). The presence of this corrosion product could lead to consumption of the NaCl salt
and, in turn, could cause electrolyte drying, a topic further explored in the next section. It is important to
note that even at longer exposure times, not all NaCl particles show corrosion initiation (Figure 5 –
c&d). Therefore, longer term exposures may be necessary to capture fully the possible corrosion
stifling at these fixed salt loadings. As inconsistencies in initiation of corrosion across the sample
surface can occur within these droplet fields, study of individual droplets provides further insight into the
stifling of corrosion on Al samples.
a) b) c)

d) e) f)

Figure 5. Optical images of macro scale salt printed Al 1100 exposed at 98% RH with 10 µg/cm2
of NaCl for a) 24 h and d) 4 weeks, 60 µg/cm2 of NaCl for b) 24 h and e) 4 weeks, and 150 µg/cm2
of NaCl for c) 24 h and f) 4 weeks. Red outlines on d) and e) indicate undissolved NaCl particles
suggesting corrosion in these areas has not yet initiated.

In-Situ Droplet Exposure

Time lapse imaging of exposed droplets at high RH (> 90%) in Figure 6Error! Reference source
not found. displays a decrease in droplet volume combined with a replacement of electrolyte with
apparent corrosion product and can be correlated to the time to bend over in the previously discussed
salt loading exposures. Two main droplet morphologies were observed at the end of exposure, a “wet”
droplet with a gel-like structure (drop 3, Figure 6-a) remaining and a “dry” droplet (drop 4 Figure 6-a)
with no visible electrolyte and significant amounts of corrosion product. These two morphologies were
further analyzed post-mortem to determine corrosion damage, morphology, and identification.
a) b)

30 100

Relative Humitdity (%)


Temperature (oC)
28 90

26 80

24 70

22 60

20 50
0 5000 10000 15000
Time (Minutes)

Figure 6. Exposure of Al 1100 coupon with 0.6 M NaCl droplets at constant humidity, a) side
view images of droplets with increasing time, with drops 3 and 4 noted and b) temperature and
RH vs. exposure time.

SEM and EDS results in Figure 7 and Figure 8 display the differences in the two primary droplet
morphologies, showing that the distribution of NaCl varies between the “wet” and “dry” droplet. The
“wet” droplet analysis is seen in Figure 7, where NaCl particles are still clearly present in the post-
mortem sample (b, e, and h) giving indication as to why the droplet is still wet at the end of exposure.
However, in the “dry” droplet shown in Figure 8, these similar particles do not exist, and instead the
majority of the chloride is not wholly associated with the sodium, but is present in a large portion of
corrosion product (b, e, and h). This gives indication of the possible consumption of NaCl in the
corrosion process leading to the drying out of the droplet over time.
a) b)

c) d) e)

f) g) h)

Figure 7. a) Side view of initial and final droplet exposures, with droplet 3, a “wet” droplet
highlighted, b) SEM SE image of droplet 3 post mortem and EDS maps of droplet 3 post mortem
for c) C, d) O, e) Na, f) Al, g) Si, h) Cl.
a) b)

c) d) e)

f) g) h)

Figure 8. a) Side view of initial and final droplet exposures, with droplet 4, a “dry” droplet
highlighted, b) SEM SE image of droplet 3 post mortem and EDS maps of droplet 3 post mortem
for c) C, d) O, e) Na, f) Al, g) Si, h) Cl.

Optical profilometry of corrosion damage underneath the two droplet morphologies revealed much
more severe corrosion attack, in terms of maximum corrosion depth, under the “dry” droplet as
compared to the “wet” droplet, 24 and 66 µm respectively (Figure 9). Also, under the “dry” droplet, a
large pit can be associated with the area where large amounts of corrosion product formed (Error!
Reference source not found.-c). This suggests possible indication of the droplets acting as a
combined corrosion system, where the “wet” droplet is the cathodic region, and the “dry” droplet, where
NaCl is consumed and corrosion product formed, is the active anode.
a)

b) c)

Figure 9. a) Side view of initial and final droplet exposures, with droplets 3 & 4 highlighted,
optical profilometry post mortem after corrosion product removal of b) droplet 3 and c) droplet
4. Depth of corrosion in b is less than sample tilt, and therefore may not be significant.

FTIR analysis of these corrosion products revealed only the presence of dawsonite (NaAlCO3(OH)2),
Figure 10. FTIR results were taken for various point analyses from positions in both “wet” and “dry”
droplets post-mortem and are shown in Figure 10. The only corrosion product detectable was
dawsonite and is shown compared to a reference dawsonite spectra. The shifting or broadening in
some of the bands could likely be due to water coordination and/or the presence of other counter ions
such as chloride. This again supports the idea that consumption of NaCl in the corrosion products may
lead to drying of the electrolyte across the surface. However, as all measurements were taken post-
mortem and ex-situ, possible formation of dawsonite could have occurred after the exposure, and may
not be wholly representative of the corrosion products and species present in the droplets.

Atmospheric Exposure
1.0
Dawsonite Reference

0.8
log (1/R)

0.6

0.4

0.2

0.0
4000 3500 3000 2500 2000 1500 1000 500

Wavenumber (cm-1)

Figure 10. FTIR spectra for various point analyses taken across the exposed Al 1100 coupon
with 0.6 M NaCl droplet. Spectra are compared to a dawsonite reference spectra shown in blue.

Preliminary modelling enabled exploration of the effects of the consumption of NaCl by corrosion
products in these droplets in Figure 11 showing electrolyte volume decrease with NaCl consumption.
OLl Analyzer1 strong electrolyte modeling software (OLI Systems), was applied to model droplet
chemistries and possible volume changes that might occur through the corrosion process.21
Stoichiometric replacement of NaCl with the anodic corrosion product; Al(OH)2Cl, Al2(OH)5Cl, and
NaAlCO3(OH)2, was assumed. For all cases, the water volume shrank as NaCl was converted to
corrosion product. This further supports the idea of gettering or consumption of NaCl by corrosion
product leading to surface drying and depletion of the corrosion aggressor, thus possibly initiating
corrosion stifling.

Figure 11. Calculation of change in NaCl droplet volume if NaCl present in the volume is
stoichiometrically replaced (1:1 mole basis) with corrosion product at 98% RH.

.
CONCLUSIONS

 Salt loading coupon exposures displayed increased mass gain with increasing initial salt loading
density, and a stifling in mass gain with increasing time. Gravimetric mass loss measurements
were in agreement with the mass gain measurements. These exposures confirm dependence
of Al corrosion on salt loading and the occurrence of stifling, but further study is needed to
understand the mechanisms for stifling for corrosion prediction.

 An increase in corrosion product was shown through optical microscopy for both increasing time
exposures and increasing salt loading, and was identified as dawsonite through GIXRD. This
increase in corrosion product over time supports possible electrolyte drying as NaCl salt is
consumed.

 A decrease in droplet volume combined with a replacement of electrolyte with apparent


corrosion product was seen for constant humidity exposures and can be correlated to stifling in
the salt loading exposures.

 Two primary droplet morphologies formed, with different final distributions of NaCl and corrosion
damage morphologies, which could be linked to anodic and cathodic active sites. These sites
can form the driving force for corrosion and thus NaCl redistribution and consumption.

 FTIR results displayed dawsonite as the only corrosion product with possible Cl- association,
supporting theories of possible consumption of NaCl leading to droplet drying. This was further
explored through preliminary modeling efforts, which confirmed droplet volume shrinking
through corrosion.

 Future work to better support these preliminary findings will include in-situ monitoring of

1
OLI Systems, Inc., 108 American Road, Morris Plains, NJ 07950.
corrosion product formation as well as in-situ corrosion rate monitoring. Findings from this study
can help to predict corrosion rates for Al with fixed salt loading conditions including the effects of
stifling as well as provide enhanced empirical data for modelling prediction.

ACKNOWLEDGEMENTS

The Authors gratefully acknowledge Bonnie McKenzie for her aid with SEM and EDS measurements
and analysis, Alice Kilgo for her aid with metallographic sample preparation, Mark Rodriguez for his aid
in GIXRD measurements, and Laura Martin and Kathleen Alam for their aid with FTIR spectroscopy.
Sandia National Laboratories is a multi-program laboratory managed and operated by Sandia
Corporation, a wholly owned subsidiary of Lockheed Martin Corporation, for the U.S. Department of
Energy’s National Nuclear Security Administration under contract DE-AC04-94AL85000.

REFERENCES

1. T.E. Graedel, "Corrosion Mechanisms for Aluminum Exposed to the Atmosphere," J


Electrochem Soc 136, 4 (1989): p. C204-C212.
2. A.S. Elola, T.F. Otero, A. Porro, "Evolution of the Pitting of Aluminum Exposed to the
Atmosphere," Corrosion 48, 10 (1992): p. 854-863.
3. J.J. Friel, "Atmospheric Corrosion Products on Al, Zn, and AlZn Metallic Coatings," Corrosion
42, 7 (1986): p. 422-426.
4. S.Q. Sun, Q.F. Zheng, D.F. Li, J.G. Wen, "Long-term Atmospheric Corrosion Behavior of
Aluminum Alloys 2024 and 7075 in Urban, Coastal, and Industrial Environments," Corros Sci
51, 4 (2009): p. 719-727.
5. H.P. Godard, The Corrosion of Light Metals, (New York,: Wiley, 1967).
6. A.R. Mendoza, F. Corvo, "Outdoor and Indoor Corrosion of Non-ferrous Metals," Corros Sci 42,
7 (2000): p. 1123-1147.
7. J.A. Gonzalez, M. Morcillo, E. Escudero, V. Lopez, E. Otero, "Atmospheric Corrosion of Bare
and Anodized Aluminum in a Wide Range of Environmental Conditions. Part I: Visual
Observations and Gravimetric Results," Surf Coat Tech 153, 2-3 (2002): p. 225-234.
8. R. Vera, D. Delgado, B.M. Rosales, "Effect of Atmsopheric Pollutants on the Corrosion of High
Power Electrical Conductors: Part 1. Aluminum and AA6201 Alloy," Corros Sci 48, 10 (2006):
p. 2882-2900.
9. Pearlste.F, L. Teitell, "Corrosion of Aluminum and Magnesium in Tropics," Mater Performance
13, 3 (1974): p. 22-27.
10. Z.H. Dan, I. Muto, N. Hara, "Effects of Environmental Factors on Atmospheric Corrosion of
Aluminum and its Alloys Under Constant Dew Point Conditions," Corros Sci 57, (2012): p.
22-29.
11. M. Natesan, G. Venkatachari, N. Palaniswamy, "Kinetics of Atmsopheric Corrosion of Mild
Steel, Zinc, Galvanized Iron, and Aluminum at 10 Exposure Stations in India," Corros Sci 48,
11 (2006): p. 3584-3608.
12. J. Kobus, "Long-term Atmospheric Corrosion Monitoring," Mater Corros 51, 2 (2000): p. 104-
108.
13. D. de la Fuente, E. Otero-Huerta, M. Morcillo, "Studies of Long-term Weathering of Aluminum
in the Atmosphere," Corros Sci 49, 7 (2007): p. 3134-3148.
14. D.B. Blucher, R. Lindstrom, J.E. Svensson, L.G. Johansson, "The Effect of CO2 on the NaCl-
Induced Atmospheric Corrosion of Aluminum," J Electrochem Soc 148, 4 (2001): p. B127-
B131.
15. D.B. Blucher, J.E. Svensson, L.G. Johansson, "The NaCl-Induced Atmospheric corrosion of
Aluminum; the Influence of Carbon Dioxide and Temperature," Elec Soc S 2001, 22 (2001): p.
741-748.
16. S.B. Lyon, G.E. Thompson, J.B. Johnson, G.C. Wood, J.M. Ferguson, "Accelerated Atmospheric
Corrosion Testing Using a Cyclic Wet Dry Exposure Test - Aluminum, Galvanized Steel, and
Steel," Corrosion 43, 12 (1987): p. 719-726.
17. G. El-Mahdy, K.B. Kim, "AC Impedance Study on the Atmospheric Corrosion of Aluminum
Under Periodic Wet-Dry Conditions," Electrochim Acta 49, 12 (2004): p. 1937-1948.
18. E. Schindelholz, R.G. Kelly, "Application of Inkjet Printing for Depositing Salt Prior to
Atmospheric Corrosion Testing," Electrochem Solid St 13, 10 (2010): p. C29-C31.
19. ASTM2 G-1-03, "Standard Practice for Preparing, Cleaning, and Evaluating Corrosion Test
Specimens" (West Conshohocken, PA).
20. ASTM1 E104-02, "Standard Practice for Maintianing Constant Relative Humidity by Means of
Aqueous Solutions" (West Conshohocken, PA).
21. OLI Systems Inc1 , OLI Systems Analyzer Version 3.0, M. Plains, (New Jersey: OLI Systems Inc,
2002).

2
ASTM International (ASTM), 100 Barr Harbor Dr, West Conshohocken, PA, 19428-2959.

You might also like