CORROSION

Section 2: Basic Corrosion Measurements

By

D.H. Lister & W.G. Cook

Department of Chemical Engineering
University of New Brunswick
1
If a piece of metal:

corrodes to this:

How would we know how much corrosion had occurred?

Clearly, we need before and after measurements ………

but of what?
2
Size?
Length, breadth and thickness before …
• What about after?
Volume before, volume after ….. by liquid displacement ….
• Sensitive enough?

Mass?
Weigh before, weigh after (if piece small enough for reasonable
balance … note that measuring weight losses to a fraction of
a mg in a few grams is the norm in corrosion testing).
To be of any use, a weight loss measurement has to be scaled to
the size of the specimen … corrosion is a surface related
phenomenon.
3
Because corrosion is a surface phenomenon, weight loss is
related to surface area (SA)

For example:
a metal “coupon” having a total surface area of 10 cm2 loses 1
mg when exposed to a particular environment for a month;

1 mg
total corrosion 
10 cm 2

which should be a characteristic amount for that metal in that

environment in a month.
mg
rate  0.1 2
cm month 4
• Corrosion engineers traditionally work in units of
dm2, thus:
mg cm 2 mg
0.1 2 100 2
 10 2
 10 mdm
cm month dm dm month

• BUT … even though mdms give reasonable

numbers for most corrosion situations, the month
is a bad unit … Why?

mg
• So … are often used …. mdd
dm 2day 5
If we know the metal density, we can convert the weight
loss to “penetration”:
for ρ in g/cm3:

x mg cm3 1g dm 2 x 5
     10 cm / day
 dm day g 1000 mg 100 cm
2 2


Which isn’t a very practical unit.

6
• However, we note that 1 cm = 104 mm, thus:

x 5 cm x mm day x mm
10   0.1  365  36.5
 day  day yr  yr

• Another common unit for corrosion penetraion (in

the US) is the mil … or milli-inch per year
(=0.001 in/year).
mm z mil
z 
yr 25.4 year 7
NOTE: these rates are averaged over the exposure time, and are
quoted as if corrosion were constant with time … it often is not …

M1 M2
measure rate 1 = measure rate 2 =
T1 T2
8
Often, our bit of metal ..

after exposure, looks like this …

It is covered with scale.

How would we assess corrosion under these conditions?

9
Scaled (e.g., oxidized) specimens often experience a
weight gain … thus, the high-temperature oxidation of
iron proceeds …

4Fe + 3O2 → 2Fe2O3

(456)g (456 + 616)g

If we know scale composition (e.g., type of oxide) and

weight gain, we can calculate the Fe in the oxide and,
therefore, the Fe corroded.
10
BUT … we can rarely be sure that the scale
accounts exactly for the corrosion … some
might fall off (spall) … some might have
been deposited from the environment …

NOW WHAT?

11
Consider the scaling process…

Weight before = Wo g
Surface area = A dm2

Exposure time = θ days

Find condition; weight after = W1 g

12
We can only be sure that the remaining metal provides an
unequivocal reference, so …

Descale

Descaled weight = W2 g
13
 ( W0  W2 ) 1000 mg  mg 
Total corrosion    2
A 1 g  dm 
(divided by θ for rate … mdd)

We also have a measure of the weight of scale …

( W1  W2 ) g
scale 
A dm 2
If we assume the scale is composed of Fe2O3, we can
calculate the iron in scale as …
( W1  W2 )
Fein scale   fraction of Fe in scale
A
( W1  W2 ) (4  56) 1000 mg
Fein scale   
A (4  56)  (6 16) 1g 14
 ( W1  W2 ) mg
Fein scale  700
A dm 2

The difference between iron in scale and total

amount corroded is either …
• Iron released to environment … OR …
• Iron deposited from environment

15
How do we descale?

Chemically or electrochemically … usually in weak acids.

BUT … such descaling can also dissolve some of the remaining

metal, so … how do we account for this?

• employ an “inhibitor”
• use a blank …

16
A blank is a “bare” specimen put into the descaling solution
along with the specimen;

during the descaling, if the blank corrodes by:

Wb  mg 
weight loss of blank  1000  2
A b  dm 
we apply this blank correction to the specimen measurement

DRAWBACKS? DISCUSS

EXAMPLE.
17
Note: such corrosion measurements and units are useful for general
or uniform corrosion but they don’t work for localized corrosion
(e.g. pitting or cracking). A component can fail by stress corrosion
cracking (SCC) with no detectable weight loss …

18
• For these situations, we use concepts such as crack
propagation rate or pit propagation rate.

• For example:
• “percent through-wall” for tubes, pipes, vessels, etc.;

• “crack depth” or “pit depth” – usually in mil or mm;

mm mm
• “propagation rate” – , etc.
year year

19
 Examples of Corrosion
General corrosion ….

General corrosion of the reactor vessel head of the Davis-Besse PWR.

(note: this exterior general corrosion was indicative of a much more
20
serious local corrosion problem due to a leak in the pressure boundary)
• Classic crevice corrosion.
In this case caused by a
carbon rubber exhaust
flapper clamped onto a
stainless exhaust flange
plate. Notice that the
worst corrosion is where
the clamp pressed the
rubber the tightest.

21
Stress Corrosion …

Stress corrosion of stainless steel type 304 autoclave. (Mallinckrodt 22

Chemical Works)
• How would you “measure” (estimate) the
corrosion of plant components?
• For example:
• pressure vessels or pressure tubes;
• feeder pipes
• steam generator tubes;
• feedwater pipes and heaters;
• etc.

• DISCUSS …
23
• Methods involve …
• Non-Destructive Evaluation (NDE) or
• Non-Destructive Testing (NDT)

• and include …
• Ultrasonic Testing (UT);
• Eddy Current;
• Acoustic Emission;
• etc.

24
Some useful conversion factors:

25