Materials Technology

Corrosion
Failure from
Water-Soluble
Contaminants
on Abrasives by William C. Johnson,
KTA-Tator, Inc.

I t is widely recognized that soluble salts can af-

fect the performance and lifetime of protective
coatings. One strong indication of the presence
of soluble salts is flash rusting (or rust-back).1 For
example, in the blast cleaning of a bridge in 1982,
flash rusting that occurred was attributed to salt
contamination of the substrate deposited by the
vised. Flash rusting is merely a symptom of a con-
dition that can lead to accelerated corrosion and
paint breakdown. When soluble salts occur on
the surface, whatever the reasons for their deposi-
tion, efforts should be made to remove the salts
prior to painting. Removal can typically be done by
pressure washing followed by a subsequent light
abrasive used for blast cleaning. Short of actual blast cleaning.
wetting from rain, dew, or condensation, visible The presence of soluble salts on steel also af-
flash rusting will not occur unless salts are pre- fects the relationship between dew point and rec-
sent. This article describes the problem of corro- ommended temperature. Currently, most authori-
sion from soluble salts, including data derived ties recommend that the steel temperature be at
from tests on abrasives. least 3 C (5 F) higher than the dew point prior to
painting.
This helps to ensure that the surface will not
have any excess moisture that can interfere with
the adhesion of the coating. For certain salts, like
The Problem ferrous chloride, this criterion is not sufficient.
Ferrous chloride will absorb water until it becomes
A common misconception is that if a paint is ap- a 20 percent solution at a temperature of 5 C (9 F)
plied before rust-back occurs, the corrosion prob- above the dew point. The presence of this aqueous
lem will be mitigated. If salts are present, however, solution will interfere with the paint’s wetting of
applying paint before flash rusting occurs is ill-ad- the surface.

54 / Journal of Protective Coatings & Linings Copyright ©1990, Technology Publishing Company
 Contaminated Abrasives / Materials Technology

■ Fig. 1
Panels blast-cleaned with abrasives tested
for conductivity

In addition to preventing proper wetting of Courtesy of

the film, salts can also accelerate corrosion. Cer- William Johnson
tain salts are corrosion stimulators.1 The salts
stimulate corrosion by drawing water through the
coating film by osmosis. They also set up a cell po-
tential and lower the pH. The most harmful ac-
tions of salt are the lowering of pH (resulting in a
regeneration cycle that consumes oxygen) and the
absorption of water through osmosis or lowering
of vapor pressure.

Mechanism of Surface Contamination

Ionogenic materials (soluble salts forming ions in
water) can result in a contaminated surface from 3
major sources.
• Contamination can be transferred onto the steel A brief description and some examples follow
surface upon impact from blast cleaning abrasives. for each of the sources described above.
• Anions can accumulate on the anode sites in the
bottoms of corrosion pits from previous exposures Salts on Abrasives
to deicing salts, marine spray, chemical pollution, Ionogenic contamination from abrasives is trans-
bird droppings, or acid rain. ferred uniformly onto steel surfaces. The resulting
• Contamination can also be introduced into the flash rusting (or rust-back) turns the steel surface a
primer from salt-contaminated pigments. homogeneous light gray to black color, depending

Copyright ©1990, Technology Publishing Company September 1990 / 55

Materials Technology / Contaminated Abrasives

on the concentration of salt. The photograph in cleaning. They subsequently absorbed water and
Fig. 3 shows the surface condition of steel panels catalyzed the corrosion of the steel. The chlorides
that were blast-cleaned to white metal and exposed were not uniformly distributed as they were in the
to high humidity over a weekend. The panel on the case of new steel blast-cleaned with contaminated
left was blast-cleaned with uncontaminated abra- abrasive (Fig. 3).
sive and consequently shows no rust-back. The one
on the right was blast-cleaned with contaminated Contamination on Pigments That Are
abrasive; thus, it shows severe rust-back that is ob- Incorporated into Primers
served to be uniform. The uniformity is due to the Salt-contaminated pigments within a primer cause
distribution of contamination. Uniform rust-back blistering5 of overcoats. Contamination on pig-
is a characteristic of corrosion caused by salts ments in topcoats provides a path for moisture
transferred from abrasives. through the coating. This contamination can be
Although the contaminant is uniformly dis- controlled by specifying limits in accordance with
tributed at first, after a spot becomes anodic and ASTM D 2448, “Test Method for Water Soluble
time passes, the anions will begin to segregate Salts in Pigments by Measuring the Specific Resis-
electrochemically and become concentrated at tance of the Leachate of the Pigment.”
anode sites. These sites occur where a material
such as oxygen or a force such as an electropoten- Mitigation of Corrosion by Lead Pigments
tial also stimulates corrosion. Oxygen enters at a Lead pigments incorporated into primers have mit-
coating deficiency such as a pinhole or a thin film igated corrosion by insolubilizing sulfates and
spot. When an anode site becomes dominant, an- chlorides, the predominant contaminating ions,
ions will accumulate there, and, after time, anode thus reducing the osmotic forces and limiting
pitting will become apparent. somewhat the harmful corrosive activity of these
anions. This ability to insolubilize has distin-
Anions in Corrosion Pits guished lead pigments from pigments of other
Anions such as chloride or sulfate ions migrate to- cations. However, lead pigments are essentially im-
ward anodes (hence the name) where metal loss, mobile in the coating; thus, only chloride and sul-
rust bloom, and pitting occur. Here, they are next fate ions that migrate to the particles of lead pig-
to the steel.2,3,4 The pits are filled with black corro- ments will be insolubilized. The chloride and
sion product, which covers the anions and blocks sulfate ions that do not contact the lead particles
their removal. The corrosion pits containing the will be able to continue to corrode the steel. An-
anions are arranged in a speckled pattern. The ions that are isolated in deep anodes are less apt to
most critical areas on contaminated steel surfaces, contact a lead particle than anions that are on the
the bottoms of pits, are difficult to clean. However, surface or are incorporated with pigments.
the objective of each cleaning should be to remove The solubility of a salt is related to its os-
these salts. Emphasis should be on removing the motic pressure, which brings water through the
material in the bottoms of pits rather than on the coating. The solubilities of lead chloride and lead
amount of material allowed to remain on the sur- sulfate are much lower than the solubilities of fer-
face. (In recent years, the industry has begun rous or ferric salts of chloride or sulfate. In terms
much research and development activity to find of osmosis, ferrous chloride will draw 200 times as
means of identifying salts and removing them from much water through a coating as lead chloride
pits and corroded steel surfaces.) will; ferric sulfate will draw 150,000 times as much
The photograph in Fig. 2 shows a speckled water through a coating as lead sulfate will. Thus,
pattern of black rust-back growing out of pits. This lead sulfate will not produce osmotic blistering.
specimen, removed from a bridge, was severely As a consequence, salt contamination has
corroded from deicing salts. It was blast-cleaned to not received adequate attention or emphasis.
white metal with minimally contaminated slag and Moreover, industrial development has brought
allowed to sit at office conditions for 1 month. The greater prevalence of contamination by deicing
pitted areas are spotted with newly formed black salts, chemical pollution, and acid rain, making the
iron oxide corrosion product, giving the area a control of corrosion-stimulating anions essential.
speckled appearance. The corrosion consumed the However, lead compounds, which were in
iron at these spots, forming deep pits. These pits widespread use until their health hazards became
are very small (less than 1/16-inch or 1.6-millime- generally recognized, are, currently, severely re-
ters diameter) compared to cathode areas. Chlo- stricted. Without the incorporation of lead pig-
rides remained at the bottoms of pits during blast ments in primers, chlorides and sulfates must now

56 / Journal of Protective Coatings & Linings Copyright ©1990, Technology Publishing Company
 Contaminated Abrasives / Materials Technology

be minimized by other means, such as with con-

trols on abrasives.

Evaluating the
Effects of Abrasive
Contaminants
In the past 2 decades, blasting has become a pre- ■ Fig. 2
ferred surface preparation method, and the Speckled pattern of
amount of blasting prior to painting steel has in- rust-back growing out of pits
creased markedly. The by-product abrasive indus- caused by deicing salts
try, together with the mineral industry, has grown
to meet the market demand. Unfortunately, both
highly contaminated as well as minimally contami-
nated by-product or mineral abrasives are being
marketed.1 In addition, there is very little informa-
tion available on the effect of various levels of con-
tamination on performance.
As a part of an ASTM testing program, the
author evaluated the degree of flash rusting pro-
duced with abrasives having a variety of conductiv-
ity levels. As discussed in the box on p. 58, conduc-
tivity is considered a relatively accurate indication
of the presence of soluble salts on the abrasives.
Abrasives with a range of contamination lev-
els were obtained, as well as samples of the various
abrasive types such as by-product, sand, and anoth-
er mineral. Four samples of slag abrasive were ob-
tained and analyzed, representing a range of con-
ductivity levels. One minimally contaminated sand
and a proprietary mineral sand were also included. ■ Fig. 3
Surface condition of
Test Data steel panels blast-cleaned
The conductivity of the abrasives was measured in to White Metal and exposed to
high humidity
accordance with the subsequently adopted ASTM
Test Method D 4940. In addition, each of the abra-
sives was used to blast clean new, hot-rolled steel
panels, which were then exposed to high humidity
in an enclosed chamber over the weekend. The
panels are shown in Fig. 1.
Four of the panels (1, 6, 5, and 4) were blast-
ed with slag abrasives. Respectively, the conductiv-
ities of these slag abrasives in micromhos per sq rusting. The absence of rust-back on panel 3, an
cm were nominally 30 (no flash rusting), 150 (in- anomalous result, prompted further investigation.
termediate flash rusting), 450 (substantial flash Discussion with the supplier revealed that an alka-
rusting), and 2100 (very substantial flash rusting). line material is used during the processing of the
Panel 2 was blasted with the sand sample, which product. Such a water-soluble, inhibitive material
had a conductivity of 56 micromhos per sq cm, and would contribute to the total conductivity result,
showed no flash rusting. Panel 3 was blasted with a obscuring the presence of salt. Thus, the result of
proprietary mineral sand that had a conductivity of this abrasive would be invalid as a measure of com-
313 micromhos per sq cm and showed no flash mon harmful ions such as chloride or sulfate.

Copyright ©1990, Technology Publishing Company September 1990 / 57

Materials Technology / Contaminated Abrasives

The basic precision of the The data are presented in Table

Development of method (i.e., reproducibility and re-
peatability) was established through an
1. These data confirm that commercially
available abrasives have a wide range of
ASTM Test interlaboratory evaluation program to
satisfy the requirements for an ASTM
conductivity. The data also show that
the method is appropriate for abrasive
Method method. An additional purpose of the in-
terlaboratory study was to reveal any
materials with both very low and very
high conductivity. A full report of the
Potential salt problems were reported at vague, ambiguous, or incomplete lan- study is available from ASTM.6
a symposium, “New Concepts for Coat- guage under conditions of actual use of The statistical interpretation of
ing Protection,” which was held in Lake the method by operators inexperienced these data is as follows.
Buena Vista, FL, on January 26, 19831 with the exact requirements. Therefore, • Repeatability: Two results, each the
under the sponsorship of ASTM Com- an interlaboratory test program was de- mean of 2 runs obtained by the same op-
mittee D-1 on Paint and Related Coat- signed in accordance with the require- erator, should be considered suspect if
ings and Materials in conjunction with ments described in ASTM D 3980, “Prac- they differ by more than 5 percent rela-
the SSPC. In this symposium, the au- tice for Inter-Laboratory Testing of tive.
thor presented the results of an analysis Paint and Related Materials.” • Reproducibility: Two results, each the
of salt contamination on abrasives using mean of 2 runs obtained by operators in
procedures developed by KTA-Tator, Inc. Design of Analytical Test Program different laboratories, should be consid-
(Pittsburgh, PA). Prior to designing the program, the task ered suspect if they differ by more than
The data showed a wide range of group directed that the method be user- 22 percent relative.
contamination levels. Severe rust-back friendly. Commercial laboratories have A brief summary of the test
was observed after the use of abrasives had experience with analytical methods method is given below.
with high concentrations of salt on com- that are based on electrical conductivity The ASTM D 4940 procedure re-
mercially available abrasives. Persons (or resistance). Equipment costs are quires adding an equal volume of abra-
attending the conference expressed in- modest, and, with brief training, an op- sive to deionized water and stirring
terest in the analytical test method erator can become skilled. However, the mixture. The water and dissolved
(which was based on conductivity mea- among abrasive producers, painting salts are filtered, and the filtrate is mea-
surements) to quantify the contamina- and lining contractors, paint manufac- sured for conductivity with a conductivi-
tion, and the need for standardization turers (except for larger or more diversi- ty meter.
and consensus of a test method for salts fied ones), and owners (again, except It is important to note that be-
was recognized. larger ones), such experience or neces- cause the specific conductance increases
As the result of this interest as sary equipment is not readily available. as the temperature increases, an error
well as concern shown by the SSPC Sur- The task group sought a method that can be introduced by a temperature de-
face Preparation Committee regarding could be readily conducted and mas- viation of approximately 2 percent per
contamination on abrasives, work on a tered by operators and technicians with degree. This error for the temperature
standard test method was initiated by little sophisticated analytical back- deviation can be avoided when a setting
ASTM Committee D 01.46.02 in its ground or training, and, furthermore, to compensate is adjusted on the instru-
meeting held following the symposium. one that could be conducted under field ment, or when the filtrate sample is at
After 5 drafts of the method and an in- conditions. 25 C (77 F). The change in specific con-
terlaboratory study for a precision state- Seven volunteer laboratories ductance in the range of 18 C (64 F) is
ment, Draft 6 of the specification, were enlisted from ASTM and SSPC reported to be 0.0216 percent per degree
“Method for Conductimetric Analysis of memberships. The 6 abrasives, including for chloride ions and 0.0227 percent per
Water Soluble Ionic Contamination of 4 slags, 1 silica sand, and 1 heavy miner- degree for sulfate ions.
Blasting Abrasives,” was approved by al sand, were the same ones evaluated
ASTM Main Committee D-1 and has for flash rusting.
been designated ASTM D 4940. The overall experimental design
consisted of the following:
• 6 abrasive lots,
• 7 testing laboratories,
• 1 test method, and
• 2 tests (replicates) per sample.

where slags are quenched with sea water. Abrasives

Discussion with medium and low levels are found inland.
Based on experience with commercial abrasives, an
Range of Contamination Levels estimated 40 percent of the abrasives has a low
A wide range of contamination levels was found in level of contamination; 40 percent has an interme-
abrasives at various geographical locations in the diate level; and 10 percent to 20 percent has a high
US. Abrasives with high and medium level. Because of the cost of painting a bridge
concentration levels are found in coastal regions or other structures these days, which can be

58 / Journal of Protective Coatings & Linings Copyright ©1990, Technology Publishing Company
 Contaminated Abrasives / Materials Technology

Table 1 Abrasive Analysis

Volunteers
Conductivity Micromho/cm

Abrasive Sample 1 2 3 4 5 6 7 Average

1. Powerhouse slag 30 44 16 32 32 23 18 27
2. Sand 57 70 52 57 57 60 52 56
3. Mineral sand 345 357 280 309 315 337 254 313
4. Slag 2287 1800 2000 2250 2163 2035 1763 2100
5. Powerhouse slag 445 470 405 418 479 518 405 453
6. Powerhouse slag 141 148 105 155 159 153 112 144

millions of dollars, methods are needed for obtain- become common. Improved practices may be William C. Johnson
ing uncontaminated abrasives and limiting the oc- needed for recycled abrasives. These abrasives, in- was the task group chairman
currence of salts on painting surfaces, whatever advertently used to clean contaminated steel, could under ASTM D 1.46 for the
preparation of a standard for
their source. accumulate contamination and pass it along. analyzing salts on abrasives.
He is a member of the staff at
Editor’s Notes: In 1987, the US Navy issued a KTA-Tator, Inc., a consulting
Contamination Limits specification with the maximum conductivity engineering firm specializing in
Limits of contamination for various uses have not level of 4,166 microsiemens (when converted to industrial protective coatings.
Prior to joining
been set. The Steel Structures Painting Council the method used in this study). The SSPC Surface KTA, Johnson earned a degree
(SSPC), in its committee on abrasives, is working Preparation Steering Committee has recom- in Chemical Engineering from
on establishing such limits. The Federal Highway mended a maximum level of 1000 microsiemens the University of North Dakota
Administration (FHWA) has contracted for a study for incorporation into the SSPC Abrasives Specifi- in 1941.
After college, he conducted
that includes identifying the influence of the level cation. research on paints for the
of soluble contaminants on coating lifetime and For further discussion about the effects of DuPont Company.
corrosion. Considerations for such limits will be salt-contaminated abrasive, see Peart and Fultz He became interested in
based on the severity of the end use (atmospheric (pp. 21-27). primers for steel and accepted
or immersion), the confidence level of the life ex- a position with the
Bethlehem Steel Corporation.
pectancy, and the supply or cost of abrasives at He later retired from there.
each limit. Based on data reported in 19831 and ex- He is the author of
perience since then, a typical value of conductivity
for an abrasive containing a high level of contami-
References patents on paint primer
compositions and of
publications on principles of
nation is 500 micromhos per sq cm. A typical value 1. W.C. Johnson, “Detrimental Materials at the Steel Paint In- primer compositions;
for an abrasive with a low level of contamination is terface,” New Concepts for Coatings Protection of Steel on contamination of
50 microsiemens per sq cm. Based on the rust- Structures, ASTM STP 841, D.M. Derger and R.F. Wint, eds., steel surfaces by salts, oxides,
back results of the present study, abrasives with American Society for Testing and Materials, 1984, pp. 28-43. grease, and moisture; and
(SSPC Report 83-01) on procedures for
conductivities in the 30 microsiemens per sq cm 2. C. Calabrese and J.R. Allen, “Surface Contamination of At- painting steel bridges.
range would be preferred for immersion exposure; mospherically Corroded and Blasted Steel,” Corrosion, Vol. Readers can con-
abrasives up to 150 micromhos per sq cm might be 34, 1978, pp. 331-338. tact him at KTA-Tator, Inc.,
acceptable for mild atmospheric end uses. 3. H.J. Cleary, “Technical Note: Some Observations on Auto- 115 Technology Drive,
body Rust Layers,” Corrosion-NACE, November 1984, pp. Pittsburgh, PA 15275.
606-608.
Impact of Adopting a Standard Criterion 4. W.P. Cathcart, “On the Cleanliness of Steel Before Coating
Adoption of a criterion for acceptable levels of salts Application,” Journal of Protective Coatings & Linings,
on abrasives may cause major changes in the in- November 1986, pp. 8-10.
dustry. But sufficient uncontaminated abrasives 5. W.C. Johnson, US Patent No. 2,858,285, Oct. 28, 1958.
6. W.C. Johnson, Research Report, RR: D 01-1061, “Interlabo-
will now be or are expected to become available to ratory Analysis of Abrasives for Ionogenic Contamination.”
supply the market. For example, the use of less May be obtained on loan from ASTM Headquarters, 1916
contaminated water in the quenching of slag may Race Street, Philadelphia, PA 19103.

Copyright ©1990, Technology Publishing Company September 1990 / 59