Environ. Sci. Technol.
2010, 44, 4322–4327
Chromium(III) and Chromium(VI)
Surface Treated Galvanized Steel for
Outdoor Constructions:
Environmental Aspects
D A V I D L I N D S T R Ö M , Y O L A N D A H E D B E R G ,
AND INGER ODNEVALL WALLINDER*
Division of Surface and Corrosion Science, School of Chemical
Science and Engineering, Royal Institute of Technology (KTH),
Drottning Kristinas v. 51, 100 44 Stockholm, Sweden
Received February 1, 2010. Revised manuscript received
April 9, 2010. Accepted April 25, 2010.
The long-term degradation of chromium(III) (Zn-Cr(III)) and
chromium(VI)-based (Zn-Cr(VI)) surface treatments on galvanized
steel and their capacities to hinder the release of zinc
induced by atmospheric corrosion at nonsheltered urban and
marine exposure conditions for 2 years are investigated.
Compared to bare zinc sheet, both surface treatments revealed
high corrosion protection abilities and capacities to hinder
the release of zinc, still evident after 2 years of exposure. The
zinc barrier properties of the thinner Zn-Cr(VI) (10 nm)
treatment were during the first 100 days of urban exposure
slightly improved compared with Zn-Cr(III) (35 nm). However,
their long-term protection capacities were inverse. Released
concentrations of total chromium correspond to annual release
rates less than 0.000032 (Zn-Cr(III)) and 0.00014 g Cr m-2
yr-1 (Zn-Cr(VI)) after 1 year of urban exposure. Aging by indoor
storage of the surface treatments prior to outdoor exposure
reduced the released Cr concentrations from the surface
treatments. No Cr(VI) was released from the aged surfaces
but from the freshly exposed Zn-Cr(VI). Marine exposure
conditions resulted in a faster reduction of chromate to
chromium(III)oxide compared with urban conditions, and a
significantly lower amount of both chromium(III) and chromium(VI)
released from Zn-Cr(VI) at the marine site compared with
the urban site.
Introduction
Chromium(VI)-based surface treatments have commonly
been used to prevent white rust staining (corrosion) of zinc
sheet or zinc coatings upon atmospheric storage, transport,
and handling (1-3). Their capacities to temporarily hinder
corrosion are related to the self-healing effect induced by
hexavalent chromium (4). Since hexavalent chromium is a
carcinogen, its use is strongly restricted within the European
Union (5).
The corrosion performance of chromium based surface
treatments is commonly characterized in the literature in,
for example, accelerated corrosion tests or through electro-
chemical investigations in bulk solution (6, 7). Their possibility
to hinder corrosion-induced zinc runoff at different atmo-
spheric conditions is significantly less investigated due to
their intended use as conversion coatings or for temporary
* Corresponding author phone: +46-8-7906621; fax: +46-8-208284;
e-mail: ingero@kth.se.
4322 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 11, 2010
corrosion protection. Recent findings have elucidated their
barrier capacities also at long-term atmospheric conditions
(8, 9).
Triggered by an increased environmental concern related
to the diffuse dispersion of metals in the society, several
investigations have been performed at different atmospheric
conditions to generate short and long-term quantitative data
on corrosion-induced zinc runoff from bare and coated zinc-
based materials in external buildings and roofs in relation
to patina formation and surface properties (9-16), and to
assess the environmental fate of dispersed zinc (9, 17). A
scientifically solid base for sustainable decisions, potential
legislative actions, and accurate risk assessment and man-
agement for zinc in external constructions is hence continu-
ously generated.
Zinc and chromium are both essential and potential toxic
metals for water organisms, depending on, for example, their
concentration, state of oxidation, and formation of complexes
(18, 19). While zinc is of relatively low toxicity to humans, it
can be highly toxic for water organisms (in fresh water), with
a predicted no effect concentration (PNEC) of 7.8 µg L-1 (20).
Active chromium(VI) is highly toxic due to a facilitated
ability to penetrate cell membranes (21) and is therefore
generally considered to be more toxic; however, recent studies
for fresh water algae show that Cr(III) is more toxic than
Cr(VI) to freshwater algae (22). The state of oxidation and
the degree of complexation of chromium are, hence, essential
for ecotoxicological considerations. Within the framework
of environmental risk assessment in the EU, PNEC values of
3.4 µg L-1 (Cr(VI)) and 4.7 µg L-1 (Cr(III)) have been reported
for surface water (23).
This paper aims to: (i) illustrate the capability of chro-
mium(III)- and chromium(VI)-based surface treatments on
galvanized steel to reduce corrosion-induced release of zinc
compared to bare zinc sheet at nonsheltered exposure
conditions during 2 years in an urban and a marine
atmosphere; (ii) correlate time dependent changes in released
zinc with surface treatment degradation, corrosion rates, and
patina formation; and (iii) present quantitative kinetic data
on the amount of chromium and its chemical speciation as
Cr(III) and/or Cr(VI) released from the two surface treatments,
and illustrate differences between freshly exposed and
preaged surface treatments.
Experimental Section
Material. Bare zinc sheet (Zn) (0.08 wt % Cu, 0.06 wt % Ti)
and galvanized steel with two different surface treatments,
Cr(III)-based-35 nm thick, (Zn-Cr(III)) and Cr(VI)-based-10
nm thick (Zn-Cr(VI)), were exposed to determine zinc and
chromium runoff rates (300 cm2), corrosion rates (54 cm2)
and perform surface analytical studies (4 cm2).
Runoff Rate Measurements. Continuous measurements
were performed in an urban (Stockholm, Sweden), and a
marine (Brest, France) environment. The urban site was
characterized by low annual levels (<1 µgm-3) of gaseous
sulfur dioxide (SO2) and annual rainfall quantities of 639 (
11 mm yr-1 (Oct 2007-Oct 2008) and 447 ( 12 mm yr-1 (Oct
2008-Oct 2009). Additional information is given in ref 9.
Measured chloride concentrations in precipitation impinging
Plexiglas surfaces varied annually between 0.1 and 2.5 mg
L-1, implying a small influence from the Baltic Sea. The annual
SO2 concentration of the marine site was less than 2 µg m-3,
the annual rainfall quantities were 684 ( 26 mm yr-1 (Nov
2007-Nov 2008) and 745 ( 25 mm yr-1 (Nov 2008-Nov 2009),
and deposited chlorides varied from 10 to 5000 mg L-1.
10.1021/es1003022  2010 American Chemical Society
Published on Web 05/12/2010
 The exposures were conducted at an inclination of 45°
from the horizontal facing south in agreement with the ISO
9226 standard for corrosion rate measurements. Runoff water
was continuously collected in acid-cleaned HDPE-vessels
with relatively low risk of metal adsorption at the vessel walls
(no risk if the sample is acidified).
To examine effects of preaging of surface treatments (2
years indoor storage), samples were exposed (2 years later)
at the urban site for 40 days (three initial sampling periods)
collecting runoff water in polytetrafluoroethylene (PTFE) or
perfluoroalkoxy (PFA) vessels with even lower risk of metal
adsorption.
All collecting vessels and equipment were acid cleaned
for at least 24 h by immersion in 10% nitric acid, rinsed four
times in ultra pure water (18.2 MΩ cm-1) and dried at ambient
laboratory conditions. Runoff water from a bare Plexiglas
fixture was collected in parallel to determine the background
concentration deposition (blank) of zinc and chromium.
Metal Analysis and Anion Analysis. Total released zinc
concentrations based on five replica readings were deter-
mined with Atomic Absorption Spectroscopy, AAS (Perking
Elmer AAnalyst 800), with a limit of detection of 0.01 mg Zn
L-1 and a relative accuracy of (1%. Total chromium
concentrations based on three replica readings were deter-
mined using Graphite Furnace Atomic Absorption Spec-
troscopy (GF-AAS), with a detection limit of 1 µg L-1.
Calibration curves were consecutively verified during analysis
and quality assurance tests continuously performed by
analyzing standard solution samples every 10th sample. Prior
to analysis, the runoff water was acidified to a pH less than
2 to prevent potential complexation. Blank values due to
matrix effects, if any, have been subtracted for all reported
values and runoff data in this paper.
The content of active chromium(VI) and total chromium
in runoff samples from Zn-Cr(III) and Zn-Cr(VI), and
corresponding blank samples (all below detection limit) was
determined by means of differential pulse adsorptive cathodic
stripping voltammetry, DPAdCSV (Metrohm 797 VA Com-
putrace). Nonacidified (original pH) samples were frozen
after sampling prior to analysis in case of the preaged samples.
All other samples were stored at ambient dark laboratory
conditions (not acidified). Detailed methodological informa-
tion is given in ref 24. Calibration was conducted individually
for each sample by standard addition of chromium(VI). The
determination limit was 0.02 µg L-1 or slightly higher (0.04
µg L-1 at the most) for total Cr and Cr(VI) measurements.
The detection limit (detected but not determined) was
approximately 0.01 µg L-1. In addition, complexation capac-
ity measurements were performed for two different blank
samples adding 0.1 mL of 20 µg L-1 Cr(VI) to the sample
(12.6 mL) seven times, resulting in a final concentration of
1.11 µg L-1 Cr(VI).
Ion chromatography (IC) measurements were performed
to determine concentrations of chloride (Cl-), nitrate (NO3-),
and sulfate (SO42-) in the runoff water from the blank samples
(limits of detection 0.1 mg L-1).
Corrosion Rate Estimations. Corrosion rates were de-
termined after 6 months, 1, and 2 years of exposure on single
sided samples by measuring the mass loss. The groundward
side was covered with a tightly adherent nonpermeable tape
to prevent corrosion and the edges were sealed with a
corrosion resistance paint. This paint was partly removed
during the pickling procedure (8), introducing an uncertainty
in the corrosion rates determined (Zn, 10%; ZnCr(III),
ZnCr(VI), 30%).
Surface Analysis. Morphological investigations were
conducted after 1, 7, 15, 30, 90, and 180 days and after 1 and
2 years of exposure by means of scanning electron micros-
copy, SEM, and energy dispersive spectroscopy, EDS, using
a table-top SEM, Hitachi TM-1000 and a stereomicroscope
Leica M205C (×10) with video-camera Leica DFC 290 at
standard conditions.
Phase identification was conducted for the bare zinc
samples after 1 and 2 years of exposure to the urban and
marine environment by using FTIR, Fourier transform
infrared reflection spectroscopy, measurements using a Bio-
Rad FTS175C spectrometer and a TDGS detector at 8 cm-1
resolution.
The presence of crystalline phases in the patina was
determined using X-ray powder diffraction, XRD, using a
Panalytical Expert-Pro powder diffractometry instrument
with monochromatized Cu KR radiation (λ ) 1.5405981 Å).
The diffractograms were recorded from 15 to 89° (2θ) and
reflections were identified using reference data from the
ICDD-JCPDS database.
Results and Discussion
Zinc Runoff Rates, Corrosion Rates, and Patina Formation
at an Urban Site. The corrosion (oxidation) process is
primarily electrochemically induced, whereas the release
process is governed by a combination of electrochemical,
chemical, and physical (e.g., wear) processes. Both reactions
depend on many interacting environmental factors such as
presence of atmospheric pollutants, rain pH, rain quantity,
and dry-wet cycles. The metal release/runoff rate can be
higher compared with the corrosion rate, when chemical
and physical processes are dominant, or lower when the
patina layer is growing faster than dissolved. As expected,
the corrosion rate of bare zinc sheet decreased with time as
a consequence of a gradual patina formation process. In
agreement with literature findings (8, 9, 11, 16), the annual
runoff rate of zinc (Zn, 2.1; Zn-Cr(III), 1.4; Zn-Cr(VI), 1.4 g
m-2 yr-1, after 2 years of exposure) was throughout the
exposure period significantly lower compared to the corrosion
rate. The runoff rate was 47% (Zn), 61% (Zn-Cr(III)), and
33% (Zn-Cr(VI)) lower than the corrosion rate after 2 years
of exposure (Figure S1). The only exception was Zn-Cr(VI)
after 1 year of exposure revealing a twice as high runoff rate
than the corrosion rate. The high initial corrosion protection
(low initial corrosion rate) which rapidly decreased with time
after 1 and 2 years of exposure for the Cr(VI)-based surface
treatment (increasing corrosion rates and zinc runoff rates
with time) was primarily related to changes in chromium
speciation (as Cr(VI) (chromates, high corrosion protec-
tion) converts to Cr(III) (stable chromium(III)oxide) with
time), as will be discussed later.
Any assessment or extrapolation of the extent of runoff
from traditional corrosion data (mass loss or corrosion
resistance) is highly erroneous and a too conservative
measure of released metals (9, 11, 12, 16).
Both surface treatments aiming to protect the surface
during storage/transport acted despite their original purpose
(short-term (a few months) protection) as barriers reducing
the runoff and the corrosion rate throughout the exposure
period. Despite a thinner layer (≈10 nm) of the Cr(VI)-based
surface treatment, its corrosion protective ability was still
higher (lower corrosion rates) compared with the Cr(III)-
based surface treatment (35 nm) after 2 years of exposure,
showing corrosion rates only slightly lower than observed
for bare zinc sheet (after 2 years of exposure). This is believed
to be a result of the self-healing function of the Cr(VI) surface
treatment (4, 7), not occurring for Cr(III)-based surface
treatments. However, the capacity of both surface treatments
to hinder the release of zinc was still evident also after 2
years of nonsheltered exposure.
Since the impinging rainfall quantity is one of the main
factors influencing the zinc runoff quantity, normalization
to this parameter enables comparative measures between
different sampling periods and test sites (8, 10, 17), Figure
1 (top). Small variations in released zinc at given rainfall
VOL. 44, NO. 11, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 4323

FIGURE 1. Released zinc quantity per given surface area and
rainfall unit during five consecutive urban exposure periods
(top); and corresponding released amounts of zinc per surface
area and corroded surface area (bottom).
quantities from bare zinc sheet reflect differences in prevailing
environmental conditions as previously reported (8, 10, 15, 17).
The chromium-based surface treatments display long-term
capacities (compared to the original purpose of the treatment)
to reduce the released amount of zinc, capacities that were
correlated to the gradual local degradation of the surface
treatments and the concomitant formation of corrosion
products, illustrated by changes of the estimated corroded
area fraction with time, Figure S3. This time-dependent local
degradation of the surface treatments and the formation of
corrosion products in these areas are illustrated in Figure S2.
The degradation process was initially fast (up to 30 days) for
Zn-Cr(III) releasing slightly more zinc compared with
Zn-Cr(VI) in the first 30 days, followed by an enhanced
degradation process of the Cr(VI) surface treatment with time.
When normalizing the released amount of zinc to the
corroded surface area, and not the geometrical sample area,
Figure 1 (bottom), it was evident that the presence of few
and small local areas of degradation results in an initially
enhanced release of zinc from Zn-Cr(VI) compared with
Zn-Cr(III), typical for degrading coated metals due to a higher
anodic (degraded coating area) to cathodic (coated area)
area ratio which determines the corrosion rate. As the
corroded surface area fraction gradually increased and
became relatively similar between the materials, only small
differences between materials were observed after longer
exposure periods.
On the basis of FTIR, XRD, and SEM/EDS findings,
hydrozincite (Zn5(CO3)2(OH)6) was determined to be the
predominant patina constituent on bare zinc sheet, Zn-Cr(III),
and Zn-Cr(VI), both at the urban and marine site, with the
local presence of sulfur-rich and/or chloride-rich corrosion
products, especially at local degraded areas of the chromium
film for the chromium treated samples, see also Figure S4.
Detailed patina evolution on bare zinc sheet and galvanized
steel at the urban site is reported elsewhere (9, 17, 25).
Zn Runoff Rates and Corrosion Rates at a Marine Site.
Bare Zn and Zn-Cr(VI) were exposed for 5 years at the marine
4324 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 11, 2010
FIGURE 2. Average annual zinc runoff quantity per given
rainfall unit from bare zinc sheet and ZnCr(VI) exposed for 5
years at the marine site and ZnCr(III) exposed during the last 2
years of this period.
site, whereas Zn-Cr(III) only was exposed during the last 2
years. As seen for the urban site, Zn-Cr(VI) revealed a
relatively high initial capacity to reduce the release of zinc
compared to Zn, a capacity almost completely lost after 5
years of exposure, Figure 2 (note different exposure years for
Zn-Cr(III)!). An obvious barrier effect was also observed for
the Cr(III)-based surface treatment. Contrary to the findings
at the urban site, Zn-Cr(III) exhibited improved barrier
properties compared with Zn-Cr(VI) after 1 year of exposure,
when normalized to the annual rainfall quantity comparing
two different exposure periods. This effect was partially
attributed to changes in chromium speciation, as discussed
later, and partially to a thicker surface treatment (35 nm
compared with 10 nm for Zn-Cr(VI)). The capacity to reduce
the release of zinc per given rainfall unit was throughout the
first 2 years of exposure higher for Zn-Cr(III) compared with
Zn-Cr(VI), with an exception of the initial 42 days of exposure
(data not shown) when still soluble Cr(VI) was present (see
section Effect of the Exposure Site and Chlorides on
Chromium Speciation).
The local degradation process of the Cr(III)-based surface
treatment was significantly faster at the marine site compared
to the urban site probably as a result of the chloride rich
environment and the lack of self-healing capacity of the
surface treatment. Already after 2 months, the barrier
properties were significantly reduced, changing the runoff
water pH from pH 4.6 (similar to the blank value) to 6.7 as
a result of increased zinc release and the formation of zinc-
rich corrosion products in locally degraded surface areas.
Patina evolution at the marine site has previously been
reported (8). The fast degradation of the surface treatment
with time was also evident from corrosion rate measurements
showing only slightly lower rates for Zn-Cr(III) (6.9 g m-2
yr-1) after 1 year compared to bare zinc sheet (9.5 g m-2 yr-1)
and rates decreasing with time, indicative of barrier effects
due to a fast patina formation.
The total amount of released zinc from bare zinc sheet
was higher at the marine site after a given time period (5.2
g m-2 after 2 years) compared to the urban site (4.0 g m-2
after 2 years) primarily due to higher annual rainfall quantities
at the marine site. However, in agreement with previous
findings (8), a lower amount of zinc was released per given
rainfall unit from bare zinc sheet at the marine site. This is
a consequence of significantly longer wet periods at the
marine site induced by higher rainfall quantities, higher
humidity levels, and deliquescence of sea-salt aerosols
reducing the number of consecutive dissolution and repre-
cipitation cycles. Similar observations were observed for
Zn-Cr(III) and Zn-Cr(VI) after 2 years of exposure. After 1
year of exposure, Zn-Cr(III) and Zn-Cr(VI) displayed either
similar or higher released amounts of zinc per given rainfall
unit at the marine site. This implies also that the degradation
of the chromium layer may be accelerated at the marine site.
 From a corrosion perspective, the marine site is more
corrosive compared with the urban site primarily as a result
of significantly higher deposition rates of chlorides. This effect
was evident from the corrosion rate measurements showing
an average annual rate after 2 years almost twice as high (8.0
g Zn m-2 yr-1) as measured at the urban site for bare zinc.
As a result of a relatively fast degradation of the chromium
based surface treatment, similar observations were seen for
Zn-Cr(III) with decreasing corrosion rates with time due to
a rapid patina development.
Release of Chromium, Chemical Speciation, and Com-
plexing Capacities. Time-dependent measurements of the
total concentrations of chromium released from Zn-Cr(III)
and Zn-Cr(VI) were conducted throughout the 2-year
exposure by means of GF-AAS. However, despite a relatively
low limit of detection for chromium (1 µg L-1), released
concentrations were only possible to quantify for the first
three sampling periods from Zn-Cr(VI) at the urban site.
The total released concentration of chromium for all
subsequent sampling periods from Zn-Cr(VI) and Zn-Cr(III)
at the urban site and from Zn-Cr(III) at the marine site were
below the limit of detection. On the basis of a highly
conservative assumption considering the limit of detection
of GF-AAS as the released chromium concentration, the
annual release rate of chromium were estimated to less than
0.0013 g-2 yr-1 for Zn-Cr(III) and Zn-Cr(VI). When using
data based on the more sensitive DPAdCSV method, the
estimated annual release rates of chromium after 1 year of
exposure were 0.000032 and 0.00014 g-2 yr-1 for Zn-Cr(III)
and Zn-Cr(VI), that is, about 10 times lower than the
estimated rates based on GF-AAS measurements. This
illustrates that the use of an analytical technique of too low
sensitivity (concentrations below the limit of detection) can
result in highly overestimated released quantities of metals,
in this case chromium.
To more accurately determine released chromium con-
centrations and to investigate the oxidation state of released
chromium from the different surface treatments, measure-
ments were conducted with differential pulse adsorptive
cathodic stripping voltammetry (DPAdCSV) with a signifi-
cantly lower limit of detection (<0.01 µg L-1). The technique
was also applied to investigate the complexation capacity of
chromium in the runoff water. The DPAdCSV technique has
a large advantage compared with GF-AAS which showed large
deviations between different blank samples and, hence, a
strong matrix effect on the total chromium concentration
measured (limit of detection 1 µg L-1), effects not observed
with DPAdCSV. Furthermore, measurements of total chro-
mium concentrations in samples containing chromium(VI)
are significantly more reliable compared with GF-AAS
measurements since the sensitivity of GF-AAS is different for
chromium(III) and chromium(VI). Measured total concen-
trations of chromium will, hence, be overestimated when
chromium(VI) is present in the solution samples (26).
Chromium(VI) should be reduced to chromium(III) at acidic
conditions as used in the case of GF-AAS (pH < 2). However,
when comparing findings for the three first sampling periods
at the urban site (containing both Cr(VI) and Cr(III)), with
measurable total Cr concentrations for both techniques, GF-
AAS measured chromium concentrations twice as high
compared with DPAdCSV (14.0 ( 0.1 compared with 6.7 (
0.6 µg L-1), while only chromium(III) containing samples
(e.g., the three initial samples periods in the preaging study,
see Figure 3, bottom) revealed concentrations in the same
range (0.2-2 µg L-1) for both GF-AAS and DPAdCSV. This
indicates that GF-AAS cannot be used for an accurate
measurement of Cr(VI) containing samples at these low
concentration levels, even if the samples are acidified.
Measurements were conducted on collected runoff water
from Zn-Cr(III), Zn-Cr(VI) and blank samples after selected
FIGURE 3. Released amounts of chromium(III) and chromium(VI)
per surface area and given rainfall unit from Zn-Cr(III) and
Zn-Cr(VI) in runoff water collected during the initial three
sampling periods, after 1 year and after 2 years of nonsheltered
urban exposure (top); and during the initial three sampling
periods after preaging (indoor storage for 2 years) (bottom).
sampling periods (the initial three sampling periods and one
sampling period after 1 year and one after 2 years of exposure
at the urban site) stored at indoor conditions for periods up
to 2 years in total.
Neither chromium(VI) nor any total chromium was
detected (<0.01 µg L-1) in any of the blank samples, that is,
significantly lower concentrations than possible to determine
with GF-AAS were present. DPAdCSV analysis of total
chromium revealed Cr concentrations below or about blank
levels for all sampling periods from Zn-Cr(III) and from all
sampling periods with the exception of the initial three
sampling periods for Zn-Cr(VI). These findings imply that
chromium(VI) was not released after the fourth sampling
period (after 30 days).
In parallel, DPAdCSV measurements were performed for
two blank samples, investigating if runoff water has a capacity
to complex released chromium, and if this capacity is different
for different rain occasions. Both complexation capacity
measurements adding active Cr(VI) into a blank runoff sample
showed no, or a very weak capacity of runoff water at the
urban site to complex chromium. This means that released
active chromium(III) in these samples can be considered to
be mainly active (not complexed). The difference between
the measured chromium(VI) and the total chromium con-
centration is in the following nevertheless denoted active
and inactive Cr(III).
Released concentrations of total chromium were very low
ranging from 7 µg L-1 for the first sampling period of
Zn-Cr(VI) rapidly decreasing with time to concentrations
below or close to 0.02 µg L-1, the determination limit.
Measured concentrations of total chromium released from
Zn-Cr(III) were below this limit for all sampling periods. All
blank runoff sample values were below detection limit of
0.01 µg L-1. Except for the first sampling period of runoff
water collected from Zn-Cr(VI), released concentrations of
total chromium were similar or lower compared with the
release of chromium from stainless steels AISI 304 and AISI
316 and from concrete exposed at similar atmospheric
exposure conditions (27, 28).
Significantly more chromium was released from Zn-Cr(VI)
compared with Zn-Cr(III) at the urban site, Figure 3. This
was attributed to more dynamic properties of chromate on
Zn-Cr(VI) compared with a more stable chromium(III) oxide
layer on Zn-Cr(III). Measurements conducted on runoff
VOL. 44, NO. 11, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 4325
water collected from Zn-Cr(VI) clearly showed decreasing
chromium(VI) and chromium(III) concentrations with time
(Figure 3). Chromium(VI) was determined only in samples
collected during the first month of exposure (66 mm rain).
These findings are in agreement with a significantly reduced
capacity of the Cr(VI)-based surface treatment to hinder zinc
release from the galvanized steel surface occurring for
Zn-Cr(VI) between 30 and 90 days of exposure, an effect not
seen for Zn-Cr(III), cf. Figure 1. This indicates that the
complete reduction of chromium(VI) after the initial sampling
periods may be responsible for the deterioration process of
the chromium(VI)-based surface treatment.
Since all analyses were conducted on stored samples, not
frozen upon collection, changes in speciation may have
occurred during storage. However, at the pH-interval between
5.8 and 6.8 of the collected runoff water from Zn-Cr(III) and
Zn-Cr(VI), Cr(VI) is normally considered to be unstable and
would with time be reduced to chromium(III). It is hence
possible that the determined released fraction of chromi-
um(VI) in reality was higher if measured directly after runoff
water sampling. However, no significant differences were
observed when repeating the measurements of total chro-
mium in the runoff water samples with significant chromium
content as both chromium(VI) and chromium(III), collected
during the first three periods in the urban environment, stored
for 2 years by means of GF-AAS, as discussed earlier. Since
GF-AAS measurements are sensitive to the oxidation state of
chromium (26), any significant effects as a result of storage
were excluded. However, changes in speciation meaning that
the chromium(VI) part of the total chromium was higher
originally, could also occur due to complexation with, for
example, the zinc ions. It is however believed by the authors
that those complexes are weak and the DTPA-Cr complex
used during determination in the DPAdCSV method is
significantly stronger. Other complexing agents can be
excluded due to the complexation capacity measurements
of runoff blank samples (without zinc). However, certain
uncertainties (underestimation of chromium(VI)) remain
regarding the chromium speciation, except the preaged
samples stored frozen; see next section.
Effects of Preaging. Prior to a separate 40-day urban field
exposure including 3 sampling periods, the surface treated
galvanized steel materials (Zn-Cr(III) and Zn-Cr(VI)) were
preaged for 2 years at ambient indoor laboratory conditions.
Similar to the findings for nonaged samples, the released
amount of total chromium rapidly decreased with time,
Figure 3 (bottom). The results clearly illustrate that aging of
the Cr(VI)-based surface treatment on galvanized steel
influences the total amount of released chromium per given
surface area and rainfall unit and its speciation. Compared
to nonaged samples, aging reduces the release of total
chromium. In contrast to findings for the nonaged surface
treatment, no chromium(VI) was released (<0.01 µg L-1) from
any of the aged Zn-Cr(VI) samples during the initial sampling
periods.
These effects were not observed for aged Zn-Cr(III).
Slightly more chromium was released from the aged Cr(III)-
based surface treatment compared with nonaged surfaces.
However, since all measured concentrations were close to,
or lower than, the determination limit, this observation
cannot be stated with evidence. The released amount of total
chromium as well as the speciation (only chromium(III)
released) can be explained by surface aging. This aging process
results in a (time-dependent) reduction of chromium(VI)
(chromate) to thermodynamically stable chromium(III) oxide
during electrochemical processes (corrosion and self-healing
effect of the chromate layer by migration to the anodic area
(lower pH) and subsequent reduction to chromium(III)
oxide), also observed during exposure of freshly treated
samples.
4326 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 11, 2010
FIGURE 4. Momentary release rates of chromium(III) and
chromium(VI) per surface area and rainfall unit from Zn-Cr(VI)
in runoff water collected after the initial three sampling
periods, after 1 year and after 2 years of exposure at the urban
site, and during the two initial exposure periods at the marine
site.
Effect of the Exposure Site and Chlorides on Chromium
Speciation. When comparing concentrations of chromi-
um(VI) released from Zn-Cr(VI) at marine exposure condi-
tions with the parallel urban exposure, differences between
the sites were evident, Figure 4. Significantly lower amounts
of both chromium(III) and chromium(VI) were released per
surface area and given rainfall unit from Zn-Cr(VI) exposed
at the marine site compared with the urban site. This effect
was partially caused by longer sampling periods at the marine
site. Significantly higher concentrations of chlorides from
sea-salt aerosols and longer wet periods at the marine site
seem to rapidly reduce the soluble concentration of chro-
mium(VI) in the surface treatment and to accelerate its
reduction to chromium(III) oxide, with a lower released
amount of chromium as a result. These findings are in
agreement with the literature (7), showing a significant
reduction of Cr(VI) to Cr(III) with time, and slower corrosion
rates of zinc at atmospheric weathering conditions with
increasing chloride to chromate ratio.
The low release of chromium(VI) from Zn-Cr(VI) at the
marine site is in agreement with the increased amount of
released zinc per surface area and given rainfall unit, cf.
Figures 1 and 2, where the difference in zinc runoff for
Zn-Cr(VI) and Zn-Cr(III) was significantly smaller at the
marine site compared with the urban site. For example, the
zinc runoff is significantly higher from Zn-Cr(III) initially at
the urban site (approximately 100 days) but shorter at the
marine site (first sampling period of 42 days) (data not shown).
This indicates that chromium(VI) in the surface treatment
is more rapidly reduced to Cr(III) at the marine site compared
with the urban site. However, literature findings (7) have
shown that the reduction of chromium(VI) to chromium(III)
on the surface of galvanized steel at atmospheric weathering
conditions and varying chloride concentration was signifi-
cant, but proceeded with a logarithmic rate during 2 months
of exposure, hence, not as rapidly as observed in this study.
These findings imply that both the reduction of chromium(VI)
and an additional effect, for example, the formation of
corrosion products, could reduce the solubility of chromi-
um(VI) in the surface treatment causing a significantly less
release of chromium(VI).
Acknowledgments
The authors are grateful for the financial support provided
by Nordic Galvanizers Association, Sweden; Rheinzink,
Germany; Saferoad, Norway and SSAB, Sweden. Cusanuswerk
is gratefully acknowledged for the financial support of
Yolanda Hedberg. The French Corrosion Institute, Brest,
France, is highly acknowledged for their invaluable help in
collecting runoff water at the marine site. The Swedish Steel
Producers’ Association and the Carl Trygger Foundation are
acknowledged for the financial support of the polarography
instrument.
Supporting Information Available
Four figures comparing zinc runoff and corrosion rates (S1),
SEM images of the bare zinc and surface treated galvanized
steel prior and after exposure periods (S2), corroded surface
area fraction of the surface treated samples (S3) and SEM
images of local corrosion products after exposure of the
different materials (S4). This material is available free of
charge via the Internet at http://pubs.acs.org.
Literature Cited
(1) Hazan, J.; Coddet, C.; Keddam, M. Study of chromate coatings
on zinc by means of D.C, A.C and gravimetric methods in alkaline
electrolyte - Correlation to humid-storage test and to Cr VI
content of the conversion film. Corros. Sci. 1990, 31, 313–318.
(2) Hörnström, S. E.; Hedlund, E. G.; Klang, H.; Nilsson, J. O.;
Backlund, M. Surface study of a chromate pretreatment before
coil coating of Al-43.4Zn-1.6Si coated steel. Surf. Interface Anal.
1993, 29 (5), 427–433.
(3) Williams, L. The formation and performance of chromate
conversion coatings on zinc. Surf. Technol. 1977, 5, 105–117.
(4) Sarmaitis, R.; Rozovskii, V. The mechanism of protection of
zinc against atmospheric corrosion by chromate coatings. Proc.
Int. Congr on Met. Corros. 1984, 1, 390–395.
(5) [EC] European Council. RoHS, Directive 2002/95/EC, restriction
of the use of certain hazardous substances in electrical and
electronic equipment, 2003.
(6) Tomachuk, C. R.; Elsner, C. I.; DiSarli, A. R.; Ferraz, O. B.
Morphology and corrosion resistance of Cr(III)-based conver-
sion treatments for electrogalvanized steel. J. Coat. Technol.
Res. [Online eraly access] DOI: 10.1007/s11998-009-9213-1.
Published online Sept 17, 2009.
(7) Prosek, T.; Thierry, D.; Olsson, M.; Bexell, U. The effect of chloride
to chromate ratio on the protective action of zinc surface films
under atmospheric weathering conditions. Corrosion 2007, 63
(3), 258–267.
(8) Sandberg, J.; Odnevall Wallinder, I.; Leygraf, C.; Le Bozec, N.
Corrosion-induced zinc runoff from construction materials in
a marine environment. J. Electrochem. Soc. 2007, 154 (2), C120–
C131.
(9) Bertling, S.; Odnevall Wallinder, I.; Leygraf, C.; Berggren Kleja,
D. Occurrence and fate of corrosion-induced zinc in runoff
water from external structures. Sci. Total Environ. 2006, 367,
908–923.
(10) He, W.; Odnevall Wallinder, I.; Leygraf, C. A laboratory study of
copper and zinc runoff during first flush and steady state
conditions. Corros. Sci. 2001, 43 (1), 127–146.
(11) He, W.; Odnevall Wallinder, I.; Leygraf, C. A comparison between
corrosion rates and runoff rates from new and aged copper and
zinc as roofing material. Water, Air, Soil Pollut: Focus 2001, 1,
67–82.
(12) Cramer, S. D.; McDonald, L. G. Atmospheric Factors Affecting
the Corrosion of Zinc, Galvanized Steel, and Copper. In Corrosion
Testing and Evaluation: Silver Anniversary Vol., ASTM STP 1000;
Baboian, R.; Dean, S. W., Eds; American Society for Testing and
Materials.: Philadelphia, PA, 1990; pp241-259.
(13) Jouen, S.; Hannoyer, B.; Barbier, A.; Kasperek, J.; Jean, M. A
comparison of runoff rates between Cu, Ni, Sn and Zn in the
first steps of exposition in a French industrial atmosphere. Mater.
Chem. Phys. 2004, 85, 73–80.
(14) Odnevall Wallinder, I.; Verbiest, P.; He, W.; Leygraf, C. Effects
of exposure direction and inclination on the runoff rates of zinc
and copper roofs. Corros. Sci. 2000, 42, 1471–1487.
(15) Robert-Sainte, P.; Gromaire, M. C.; De Gouvello, B.; Saad, M.;
Chebbo, G. Annual metallic flows in roof runoff from different
materials: test-bed scale in paris conurbation. Environ. Sci.
Technol. 2009, 43, 5612–5618.
(16) Veleva, L.; Acosta, M.; Meraz, E. Atmospheric corrosion of zinc
induced by runoff. Corros. Sci. 2009, 51, 2055–2062.
(17) Lindström, D.; Odnevall Wallinder, I. Long-term use of galva-
nized steel in external applications. Aspects of patina formation,
zinc runoff, barrier properties of surface treatments and coatings
and environmental fate. Environ. Monit. Assess. [Online early
access] DOI: 10.1007/s10661-010-1377-8. Published online Mar
7, 2010.
(18) Ohnesorge, F.; Wilhelm, M. Chapter II.36 Zinc. In Metals and
Their Compounds in the EnvironmentsOccurrence, Analysis and
Biological Relevance; Ernest, M., Ed; VCH: Weinheim, 1991; pp
1309-1342.
(19) Gauglhofer, J.; Bianchi, V. Chapter II.7 Chromium. In Metals
and Their Compounds in the EnvironmentsOccurrence, Analysis
and Biological Relevance; Ernest, M. , Ed; VCH: Weinheim, 1991;
pp 853-878.
(20) Bodar, C. W. M.; Pronk, M. E. J.; Sijm, D. T. H. M. The European
Union risk assessment on zinc and zinc compounds: the process
and the facts. Integr. Environ. Assess. Manage. 2005, 1 (4), 301–
319.
(21) Cohen, M.; Kargazin, B.; Klein, C.; Costa, M. Mechanisms of
chromium carcinogenicity and toxicity. Crit. Rev. Toxicol. 1993,
23 (3), 255–281.
(22) Vignati, D. A. L.; Dominik, J.; Beye, M. L.; Pettine, M.; Ferrari,
B. J. D. Chromium(VI) is more toxic than chromium(III) to
freshwater algae: A paradigm to revise? Ecotoxicol. Environ. Saf.
[Online eraly access] DOI: 10.1016/j.ecoenv.2010.01.011. Pub-
lished online Feb 7, 2010.
(23) [CSTEE] Scientific Committee on Toxicity, Ecotoxicity and the
Environment. Opinion on the results of the Risk Assessment of:
Chromium trioxide (CAS No. 1333-82-0), Sodium chromate (CAS
No. 7775-11-3), Sodium dichromate (CAS No. 10588-01-9),
Ammonium dichromate (CAS No. 7789-09-5), Potassium dichro-
mate (CAS No. 7778-50-9), 2003.
(24) Midander, K.; De Frutos, A.; Hedberg, Y.; Darrie, G.; Odnevall
Wallinder, I. Bioaccessibility studies of ferro-chromium alloy
particles for a simulated inhalation scenario. A comparative
study with the pure metals and stainless steel. Integr. Environ.
Assess. Manage [Online early access] DOI: 10.1002/ieam.32.
Published online Dec 29, 2009.
(25) Odnevall, I.; Leygraf, C. The formation of Zn4C12(OH)4SO4 · 5H2O
in an urban and an industrial atmospheric. Corros. Sci. 1994,
36 (9), 1551–67.
(26) Perkin Elmer Instruments LLC Analytical Methods for Atomic
Absorption Spectrometry; Perkin Elmer, Inc.: Norwalk, CT, 2000;
p 67.
(27) Odnevall Wallinder, I.; Bertling, S.; Berggren, D.; Leygraf, C.
Corrosion-induced release and environmental interaction of
chromium, nickel and iron from stainless steel. Water, Air, Soil
Pollut. 2006, 170 (1-4), 17–35.
(28) Persson, D.; Kucera, V. Release of metals from buildings,
constructions and products during atmospheric exposure in
Stockholm. Water, Air, Soil Pollut.: Focus 2001, 1, 133–150.
ES1003022
VOL. 44, NO. 11, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 4327