Effects of Effects of Oxygen, Temperature and Salinity Oxygen, Temperature and Salinity on Carbon Steel Corrosion on Carbon Steel Corrosion in Aqueous Solutions in Aqueous Solutions --Model Predictions versus Laboratory Model Predictions versus Laboratory

Steven L. GriseSteven L. GriseBrian J. SaldanhaBrian J. SaldanhaOct 23, 2007Oct 23, 2007

November 6, 2007

2

Brine Corrosion of Carbon SteelBrine Corrosion of Carbon Steel

Water and saline solutions (brine) with oxygen can be found in essentially any industrial site– Cooling water ~0-300 ppm Cl(-)

Could include oxidizing biocides– Seawater ~ 3.4% NaCl– We all have this problem !Much of the pipe which contains these solutions is carbon steel– Waste headers, WWT operations, brine refridge systems, etc– Cost is a significant issueFor example, a carbon steel waste header which needs to accept a new brine stream– Near neutral pH– Low concentration of NaClSeems innocent enough, right ?

November 6, 2007

3

Corrosivity of BrineCorrosivity of Brine

NaCl solutions are corrosive because– Good electrolyte– Presence of chloride – forms metal chlorides– Affected by

pHTemperatureOxygenVelocity

Metals that display lowest corrosion ratesCopper alloys in the absence of oxygen (low strength is issue)High Ni-Cr-Mo alloys to minimize SCC and crevice corrosion, Hastelloy C-276® ($$) Titanium ($$)Stainless steels suffer from pitting attack

Decision becomes financial

November 6, 2007

4

Corrosion ModelingCorrosion Modeling

Multi-disciplinary. Requires knowledge of– Solution chemistry– Surface electrochemistry– Mass transport / Fluid flow– Metallurgy

Still an “infant” technology– 1st comprehensive approach was Pourbaix (1950’s)

Purely thermodynamic

Some commercial products available– Tend to be industry specific (i.e., Oil & Gas)– OLI CorrosionAnalyzer® (CA or CSP)

Most comprehensiveSolution based, broad chemistry coverage

November 6, 2007

5

Three Rules of Corrosion ModelingThree Rules of Corrosion Modeling

1. “All models are wrong, but some are useful”– Prof George E.P. Box, Univ. of Wisconsin, 1979

2. Models are not completely predictive – Data are required to develop rate parameters– Need to incorporate BOTH thermodynamics and kinetics

3. Since data are always required, models will never replace laboratory and field corrosion testing !

November 6, 2007

6

Effect of pH on CS Corrosion in BrineEffect of pH on CS Corrosion in Brine

Extremes of pH have profound effect on corrosion rate– Brines concerned primarily in pH 5-9 range

Reference:Kirby, G.N., Chemical Engineering,March 12, 1979Pg 72-84

November 6, 2007

7

Effect of pH on CS Corrosion in BrineEffect of pH on CS Corrosion in BrineSlow moving H2O or 3.5 wt% Brine;

3.9 ppmw O2

0

5

10

15

20

25

30

35

4 5 6 7 8 9 10

pH

Cor

rosi

on R

ate

(mpy

) Kirby data @ 22 C

Kirby data @ 40 C

CA - Water @ 22 C

CA - Water @ 40 C

CA - Water @ 55 C

CA - 3.5% Brine @ 22 C

CA - 3.5% Brine @ 40 C

CA - 3.5% Brine @ 55 C

November 6, 2007

8

Corrosion Rate as a Function of TemperatureCorrosion Rate as a Function of Temperature

Temperature will affect corrosion rateEffect is multi-faceted– Kinetics

Corrosion rate (kinetics) increase with temperatureUsually follows Arrhenius kineticsTemperature can affect equilibrium composition of solution

– O2 solubilityO2 solubility decreases with temperatureIncreases rate of diffusion through layers of Fe(OH)x

– pH Not significant through reasonable temperature ranges

– Mass transfer will also increase with temperature

Difficult to experimentally keep all other variables constant

November 6, 2007

9

OO22 Solubility Solubility vsvs TemperatureTemperature

0

2

4

6

8

10

12

14

16

0 5 10 15 20 25 30 35 40 45

Temperature ( C )

O2 S

olub

ility

(mg/

kg s

oln)

Carpenter dataOLI PredictionLab Data - Water

p(O2) = 0.2094 atm; NaCl conc ~ 0.34 g/kg solnO2 Solub Reference: 1) Carpenter, J.H., Limnology and Oceanography , Vol. 11, No. 2 (1966), pg. 264

November 6, 2007

10

pH pH vsvs TemperatureTemperature

6

6.2

6.4

6.6

6.8

7

7.2

7.4

7.6

7.8

8

0 5 10 15 20 25 30 35 40 45

Temperature ( C )

pH OLI Calc'd pH

p(O2) = 0.2094 atmNaCl conc ~ 0.34 g/kg soln

November 6, 2007

11

Equilibrium Concentration of OEquilibrium Concentration of O22 in Brinein Brine

Brines exposed to air will absorb O2

Amount of O2 absorbed will depend on – p(O2)– Salt concentration– Temperature– Brine / gas surface area

For example, splash zones in seawater exposures up to 10x higher corrosion rates than subsurface exposure– High surface area for O2 transfer– Concentration of NaCl due to evaporation of splash

November 6, 2007

12

Comparison of OLI Predictions to DataComparison of OLI Predictions to Data

0

1

2

3

4

5

6

7

8

0 5 10 15 20 25

wt% NaCl

O2 S

olub

ility

(mg/

l)

LangIwaiOLI Aq

p(O2) = 0.2094 atm; 37oCReferences: 1) Lang, W. and Zander, R., IEC Fundamentals , V 25, No. 4, pg 775 (1986)2) Iwai, Y., et al, Fluid Phase Equilibria , V 83, pg 271 (1993)

NOTE: MSE framework does not yet include Setchenov-type correlations.

November 6, 2007

13

Corrosion Rate as a Function of OCorrosion Rate as a Function of O22 ConcentrationConcentration

O2 provides the cathodic reaction for corrosion

Metal + ½ O2 + H2O Metal(2+) + 2 OH(-)

Higher O2 conc higher corrosion rates

…but as we have discussed, the variables are interdependent. So what happens in real experiments ?

November 6, 2007

14

Lab Data Lab Data vsvs OLI PredictionsOLI Predictions

0.00

1.00

2.00

3.00

4.00

5.00

6.00

7.00

8.00

15 20 25 30 35 40 45 50 55 60

Temperature (deg C)

Cor

rosi

on R

ate

(mpy

)

Tap water in air

OLI

Tap + 3.5% NaCl in air

OLI

Tap water w/ N2 purge

OLI

Tap + 3.5% NaCl w/ N2 purge

OLI

CorrosionAnalyzer assumptions: - Static flow conditions - Liquid saturated with O2

- p(O2) = 0.21 atm - p(O2) = 0 for N2 purge cases

November 6, 2007

15

Corrosion Rate as a Function of VelocityCorrosion Rate as a Function of Velocity

Limiting current density is a function of species movement to/from the surface

Role of velocity and flow conditions is to define mass transfer coefficient, km

Velocity also affects adherency of passive filmsHigher velocity can remove films, leaving base metalIt can also cause erosion or mechanical removal of metal

)( bulk,,lim, iimii akFni ⋅⋅⋅=

November 6, 2007

16

Different Flow ConditionsDifferent Flow Conditions

Flow conditions available in Corrosion Analyzer– Static

km ~ 0– Rotating Disk

km = 0.6211 D2/3 ω1/2 / ν1/6

– Rotating Cylinderkm d/D = 0.0791 Re0.70 Sc0.356

– Pipe Flowkm d/D = 0.0165 Re0.86 Sc0.33

– Complete Agitationkm = very large number ( infinity)

November 6, 2007

17

Compare Predictions, Similar ConditionsCompare Predictions, Similar Conditions

Flow Conditions Water Water1018 CS 1018 CS

40oC 40oCpH ~ 6.8 pH ~ 6.8

p(O2) = 0.21 atm p(O2) = 0 atm

Static 4 0.44Pipe flow (d = 2 m; v = 0.0127 m/s) 10.3 0.44Rotating Disk (d = 10 cm; 2.5 cycle/min) 8.7 0.44Rotating Cyclinder (d = 10 cm; 2.5 cycle/min) 9.5 0.44Pipe flow (d = 2 m; v = 6.1 m/s) 229 0.81Rotating Disk (d = 10 cm; 1165 cycle/min) 195 0.52Rotating Cyclinder (d = 10 cm; 1165 cycle/min) 515 0.77Completely Agitated 211 3.9

~ 1/2 in/sec

~ 20 ft/sec

November 6, 2007

18

CorrosionAnalyzer Velocity CorrosionAnalyzer Velocity vsvs Data at 23°CData at 23°C

1

10

100

1000

0 1 2 3 4 5 6 7 8

Velocity (m/s)

Cor

rosi

on R

ate

(mpy

)

Melchers/LaQue

OLI Pipe flow

OLI Rotating Disk

OLI Static

OLI CompleteAgitation

T = 23oC, Mild SteelAssume p(O2) = 0.2094 atm and 3.5% NaCl brineReferences: 1) Melchers, R.E. and Jeffrey, R., Corrosion , V 60, No. 1, pg 84 (2004) who references LaQue, F.L., in Uhlig's Corrosion Handbook

November 6, 2007

19

SummarySummary

Velocity and oxygen concentration appear to have most significant impact on corrosion in near neutral brines– Both in experimental and OLI prediction

Temperature is secondary effectpH is not a significant factor between 5 and 9OLI CorrosionAnalyzer® predictions– Very good for static conditions– Velocity effect needs further development

November 6, 2007

20

SummarySummary

Corrosion modeling is multi-faceted and interdisciplinary– Use all your resources– Study all the variables

Modeling and experimental programs are interactive– Model suggests experiment conditions– Experimental data validate and/or feed model– Improve the model – Suggest new experiments– Etc…

Did not discuss MIC and microbiological, including impact of oxidizing biocides– Very important considerations for water corrosivity!

November 6, 2007

21