Analytica Chimica Acta 663 (2010) 172–177

Contents lists available at ScienceDirect

Analytica Chimica Acta

journal homepage: www.elsevier.com/locate/aca

Determination of ferrous and ferric iron in aqueous biological solutions

S.E. Pepper 1 , M. Borkowski ∗ , M.K. Richmann, D.T. Reed
Earth and Environmental Sciences Division, Los Alamos National Laboratory, 1400 University Drive, Carlsbad, NM 88220, USA

a r t i c l e i n f o a b s t r a c t

Article history: A solvent extraction method was employed to determine ferrous and ferric iron in aqueous samples.
Received 19 August 2009 Fe3+ is selectively extracted into the organic phase (n-heptane) using HDEHP (bis(2-ethylhexyl) hydro-
Received in revised form 25 January 2010 gen phosphate) and is then stripped using a strong acid. After separation, both oxidation states and the
Accepted 29 January 2010
total iron content were determined directly by ICP-MS analysis. This extraction method was refined to
Available online 6 February 2010
allow determination of both iron oxidation states in the presence of strong complexing ligands, such as
citrate, NTA and EDTA. The accuracy of the method was verified by crosschecking using a refinement
Keywords:
of the ferrozine assay. Presented results demonstrate the ability of the extraction method to work in
Fe2+ determination
Fe3+ determination
a microbiological system in the presence of strong chelating agents following the bioreduction of Fe3+
HDEHP extraction by the Shewanella alga BrY. Based on the results we report, a robust approach was defined to sepa-
Ferrozine method rately analyze Fe3+ and Fe2+ under a wide range of potential scenarios in subsurface environments where
Shewanella alga BrY radionuclide/metal contamination may coexist with strongly complexing organic contaminants.
Published by Elsevier B.V.

1. Introduction The detection of oxidation states of iron in solutions is most

often accomplished using spectrophotometric methods, such as
The importance of iron oxides in controlling the fate and the ferrozine assay [15] and the 1,10-phenanthroline method [16].
transport of many environmentally important radionuclide/metal Both techniques rely on the formation of a colored Fe2+ -complex
species has long been recognized in the literature [1–2]. In the stor- at almost neutral pH and its absorbance is measured in relation
age and disposal of radioactive wastes, which are often envisioned to a set of standard solutions. However, these spectrophotomet-
in steel containers, a variety of iron(II) and iron(III) oxides are gen- ric methods can be limited by the presence of strong complexing
erated as the container corrodes [3–6]. The presence of iron(II) agents and possibly by other interferences (e.g., competition with
generates a reducing environment, and in the case of transuranic other divalent cations or strongly absorbing co-contaminants). A
waste, the actinides are maintained in reduced and therefore less further obstacle in the Fe2+ determination that is often neglected is
soluble and less mobile forms, for example Pu(III) or (IV) vs. Pu(V) effect of residual oxygen in the reagents used, especially when the
or (VI) [7–13]. To establish a mechanistic understanding of these concentration of Fe2+ is lower than 0.5 mM.
coupled subsurface redox processes it is necessary to concurrently A common method for the separation of metal ions having dif-
measure all the pertinent oxidation states of iron and the actinides ferent oxidation states is solvent extraction [17–21]. The solvent
(metals). These coupled redox reactions are also important for bio- extraction technique is based on the distribution of the analyte
logical redox processes when enzymes, iron and other metals are of interest between two immiscible liquids, usually an aqueous
involved [14]. Microorganisms, by controlling the Fe(II)/Fe(III) ratio, solution and an organic solvent containing an extracting species.
can greatly influence the redox environment (Eh ) in a way that may Many organic molecules were developed and utilized in the extrac-
define the oxidation state distribution of the multivalent actinides. tion process. On the basis of our experience in separation science,
An understanding of the key mechanisms of redox control in bio- bis(2-ethylhexyl) hydrogen phosphate (HDEHP) was chosen as the
logical systems may also help differentiate between enzymatic and extractant for this study since it was used to separate various metals
abiotic mechanisms. In this context, a reliable method to quantify including americium, curium, neptunium, plutonium and uranium
the iron oxidation state that is suitable for the complex systems [22–28]. This acidic extractant allows separation of cations in acidic
generally encountered under subsurface conditions is of primary medium. This extraction approach is likely able to deal with inter-
importance. ferences from strong complexants such as NTA (nitrylotriacetate)
and citrate, and to a lesser extent EDTA, which can co-exist with
radionuclide contaminants in environmental biologically active
systems.
∗ Corresponding author.
In this context, the aim of this paper is to define a solvent
E-mail address: marian@lanl.gov (M. Borkowski).
1
Present address: Interfacial Chemistry, Idaho National Laboratory, PO Box 1625, extraction procedure to detect ferrous and ferric iron with a view
Idaho Falls, ID 83415-2208, USA. to coupling this method to the detection of different oxidation

0003-2670/$ – see front matter. Published by Elsevier B.V.

doi:10.1016/j.aca.2010.01.056
 S.E. Pepper et al. / Analytica Chimica Acta 663 (2010) 172–177 173

states of multivalent radionuclides, such as uranium, neptunium and was caused by an experimental uncertainty of the extraction
or plutonium. The extraction method could permit the simultane- and stripping steps.
ous determination of all metal/actinide species using one analytical Considerable experimental work was carried out to obtain the
approach. Additionally, we modified the ferrozine assay [29] for optimum conditions that would enable both ferric and ferrous iron
the detection of Fe2+ to extend its applicability to these complex to be determined in aqueous solution and in the presence of com-
systems and use this analytical approach to confirm the results plexing ligands such as citrate, NTA and EDTA. The effects of acid
obtained in the HDEHP extraction approach. concentration for the extraction and stripping steps, contact time
for extraction and type of solvent were investigated. Once the pro-
cedure was established, a series of experiments were performed
2. Experimental to assess its performance in mixed oxidation state solutions and in
more complex “real” experimental systems.
The following reagents were used in the present work: fer-
rozine monosodium salt of 97% purity supplied by Aldrich, 2.1. Modified ferrozine assay
bis(2-ethylhexyl)hydrogen phosphate (HDEHP) 97% purity, also
from Aldrich, solvents: cyclohexane, toluene and n-heptane, all A modified version of the ferrozine assay [29] was used to
of HPLC grade from Fisher Scientific, HCl (certified ACS Plus) crosscheck the results of the extraction experiments. Briefly,
from Fisher Scientific, nitrylotriacetic acid (NTA) 99+ % from 0.9 mL of 4.06 × 10−4 M (0.2 g L−1 ) ferrozine in 0.25 M HEPES (N-2-
Aldrich, sodium dihydrogen citrate 99% from Aldrich, EDTA dis- hydroxyethylpiperazine-N’-2-ethanesulfonic acid) buffer (enzyme
odium salt dihydrate 99+ % from Sigma-Aldrich, sodium L-lactate grade) was added to 0.1 mL of acidified sample (0.5 M HCl) and the
approx. 98% from Sigma-Aldrich, lactate reagent kit from Trin- purple color was allowed to develop. The absorbance of the com-
ity Biotech., hydroxylamine hydrochloride 99% Reagent Plus from plex was measured within 15 minutes at 562 nm, using either a
Sigma-Aldrich. CARY 5000 spectrophotometer or a Thermo Spectronic GeneSys 20
All solution preparation and experiments involving Fe2+ were spectrophotometer, and compared to a calibration curve obtained
performed in a nitrogen negative pressure anoxic glovebox from a series of standards prepared in a similar fashion. No
(MBraun Labmaster 130 with re-circulating copper shaving oxygen difference in absorption readings was noted between the two
purification system) equipped with an oxygen sensor. The oxygen instruments. The molar extinction coefficient for the ferrous fer-
levels in the glovebox were maintained below 3 ppm O2 at all times rozine complex was calculated as the average of several calibrations
and were typically < 0.1 ppm O2 . It was found that at or below this and was equal to 27 400 ± 1000, in good agreement with the lit-
concentration of oxygen, significant (∼1%) oxidation of Fe2+ did not erature value [15] of 27 900. A good linear dependence of the
occur in the timeframe of a couple of months. Beer–Lambert law was observed for Fe2+ concentrations ranging
Ferric and ferrous iron solutions were prepared in 0.1 M from 10 to 80 ␮M of Fe2+ with the precision of the Fe2+ determina-
hydrochloric acid to prevent precipitation of the oxide phases. tion being equal to ± 2 ␮M. The determination limit was assessed
With the exception of the ferrous iron solution, which was pre- to be on the level of 6 ␮M Fe2+ .
pared in the glovebox, all solutions were prepared outside the For samples containing up to 8 mM of iron, the total iron con-
glovebox and transferred into the box, where they were purged centration was determined as follows: to 0.1 mL of sample 0.9 mL
to remove oxygen over several weeks by equilibrating with the of 0.28 M hydroxylamine hydrochloride in 0.28 M HCl was added.
nitrogen atmosphere. The ferrous solution was prepared with a After 30 minutes, all the Fe3+ was reduced and 0.1 mL of this
fresh FeCl2 solid (Sigma, analytical grade) and the ferric solution solution was added to 0.9 mL of the ferrozine solution and the
was prepared using the certified iron(III) oxide (Alfa Aesar stock# absorbance measured after 20 minutes. The Fe3+ concentration was
44666, lot# H21R005). Since a standard reference for the iron oxi- determined by the difference between the total iron concentration
dation state mixture is commercially unavailable, these solutions and the Fe2+ concentration directly measured in the sample.
were used as secondary standards for the preparation of iron oxi-
dation state mixtures with different ferrous to ferric ratios. The 2.2. Procedure for the HDEHP solvent extraction method
oxidation states and iron concentrations in these standards were
measured using the ferrozine method, a combination of extrac- The following procedure is applicable to sample aliquots con-
tion with ICP-MS assay and direct ICP-MS measurements. The Fe2+ taining up to 5 mM of ferric iron. The analytical range for Fe2+ was
concentration in the secondary standard solution was equal to assessed and it was determined that up to 8 mM Fe2+ could be
58.9 ± 0.3 mM and the Fe3+ concentration in the secondary stan- detected, although it is believed that this amount could be much
dard solution was equal to 99.7 ± 0.4 mM. These two solutions were higher given that Fe2+ does not partition into the organic phase. The
kept in a nitrogen-controlled atmosphere and were used for further following procedure is also applicable to solutions containing cit-
dilutions. rate and NTA at concentrations up to 5 times that of iron. All steps
Aqueous metal concentrations were determined by ICP-MS were carried out in glass vials with PTFE lids.
(Agilent Model 7500ce), fitted with a hydrogen reaction cell, to Step 1: To 0.5 mL of sample, 1.5 mL of 0.67 M HCl was added to
minimize the interference of the argon–oxygen peak with the iron obtain a final acid concentration of 0.5 M. An aliquot of the sample
peak at mass of 56 and to extend the sensitivity of analysis. The ICP- was removed for analysis to determine the total iron concentration.
MS determination limit of iron was 5 ppb, which corresponds to an Step 2: To the remaining sample, an equal volume of 0.1 M
iron concentration of 10−7 M. The ICP-MS apparatus was calibrated HDEHP in n-heptane was added and shaken for 1 hour. The phases
prior to each analytical run. The certified iron standard solution were allowed to completely separate and an aliquot of the aqueous
(High Purity Standards) contained 1000 ppm of Fe3+ and further phase was removed for determination of Fe2+ content, taking care
dilutions of this sample were used for calibration. Each calibration not to contaminate the extracted portion of the solution with the
used a minimum of six points over the iron concentration range organic phase.
of 0–500 ppb. The R2 of calibration linearity was on the level of Step 3: To a portion of the organic phase, an equal volume of
0.9999 ± 0.0001. Each result reported by ICP-MS was an average of 4 M HCl was added and shaken for 15 minutes. The phases were
five measurements and the precision of single point measurement allowed to completely separate and an aliquot of the aqueous phase
was better than 0.5%. The highest relative difference between the was removed for determination of Fe3+ content, again taking care
synthetic samples having initially the same composition was 9% to minimize contamination by the organic phase.
174 S.E. Pepper et al. / Analytica Chimica Acta 663 (2010) 172–177

Total iron, ferrous and ferric, concentrations were determined

directly from the analysis of the acidified sample by ICP-MS mea-
surements.

2.3. Coupled effects of both oxidation states and complexing

ligands

A series of solutions containing a total iron concentration of

0.5 mM were prepared using the secondary standard solutions
described above and were treated as a reference material since con-
centrations of Fe2+ and Fe3+ were precisely known. In the absence of
complexing ligands, the Fe2+ content was varied from 0 to 0.5 mM.
In the presence of citrate, NTA or EDTA, the ratio of Fe2+ to Fe3+ was
kept constant at 1:1 and the concentration of the ligands was var-
ied from 0.05 to 2.5 mM, giving a total iron to ligand ratio of 1:0.1
to 1:5.

2.4. Reduction of Fe3+ by Shewanella alga under anoxic conditions

Fig. 1. Effect of time on the extraction of 0.4 mM Fe3+ from 0.5 M HCl by 0.1 M HDEHP
in n-heptane. Squares represent total Fe3+ (sum in aqueous and organic phases) and
Full details of the experimental procedure can be found in Reed
circles represent Fe3+ extracted into the organic phase.
et al. [14]. Under anaerobic conditions, aqueous Fe3+ , as a stabilized
Fe3+ -NTA complex was added, at an approximate concentration of
[36–37]. Extraction rate can be increased by increasing the temper-
6 mM, to a solution containing the metal-reducing bacteria S. alga
ature [33]; however the effect of temperature was not investigated
(strain BrY), which had been grown anarobically [30]. Lactate, as
in the present study.
sodium L-lactate, was used as the electron donor. Total iron, ferric
The effects of three solvents, cyclohexane, toluene, and n-
and ferrous iron concentrations were measured as the experiment
heptane, on the extraction of 0.4 mM Fe3+ from a 0.5 M HCl solution
progressed using both methods as described above. Lactate was
by 0.1 M HDEHP were also established (Table 1). The relative polar-
analyzed using a lactate reagent and standards kit (Trinity Biotech)
ities of these solvents are: cyclohexane = 0.006, n-heptane = 0.012,
based on the colorimetric technique as recommended by the man-
and toluene = 0.099 [38]. Altering the solvents has a small but sig-
ufacturer.
nificant effect on the rate of partitioning of Fe3+ into the organic
phase. When cyclohexane is the solvent, the partitioning occurs on
3. Results and Discussion
the same timescale as n-heptane, i.e., one hour. Toluene appears to
increase the time for Fe3+ partitioning; after one hour only 75% of
The experimental results we report were obtained for solutions
the Fe3+ has transferred into the organic phase. This trend is corre-
that were carefully degassed to remove all trace levels of oxygen.
lated with the polarity of the solvents. In solvents of low polarity,
We note that the presence of even trace-levels of oxygen can lead
HDEHP via the formation of dimers [39] extracts trivalent cations
to significantly higher errors when the total Fe concentrations are
according to the following mechanism [40]:
0.1 mM or below due to the oxidation of Fe2+ by this residual oxygen
content leading to spurious and inconsistent analytical results. M3+ (aq) + 3(HDEHP)2(org) ↔ M[H(DEHP)2 ]3(org) + 3H+ (aq) (1)

3.1. Experimental refinement of the HDEHP solvent extraction Cyclohexane and n-heptane have low and comparable polarities
method that enable HDEHP to form dimers more rapidly, which leads to a
more rapid extraction of Fe3+ . Toluene, however, with its somewhat
HDEHP is an acidic extractant, thus the distribution ratio of higher polarity, causes a decreased rate of Fe3+ extraction.
the metal, defined as the ratio of metal in the organic phase to The Fe3+ extraction yield is a function of free HDEHP concen-
metal in the aqueous phase, depends on the acidity in the aque- tration and decreases with increasing Fe3+ concentration in the
ous phase. Therefore, the effect of the acid concentration on the organic phase. Fig. 2 shows the isotherm for Fe3+ extraction. The
extraction of 0.4 mM solutions of ferric and ferrous iron using 0.1 M extraction efficiency of Fe3+ is significantly influenced by its ini-
HDEHP was measured. Between 0.05 M and 1.0 M hydrochloric acid tial concentration in the aqueous phase. As mentioned before, the
concentration, the extraction of Fe3+ into the organic phase was extraction is better than 98.5% for Fe3+ concentrations below 1 mM.
essentially quantitative (>98.5%). Above 1.0 M HCl, the amount of A 5 mM initial concentration leads to the extraction of 92% Fe3+ .
Fe3+ extracted decreased rapidly and was almost zero at 4 M. In The amount of Fe3+ extracted decreases as the initial concentra-
the case of Fe2+ , less than 8 ± 1% was extracted into the organic tion increases; 77% of Fe3+ is extracted at an initial concentration
phase at all acid concentrations investigated. Consequently, it was of 10 mM, and 48% of Fe3+ is extracted at 35 mM. From the data
decided that acidifying the solutions to 0.5 M HCl would provide
the optimum conditions for separation of Fe3+ from Fe2+ . Under Table 1
these conditions, the Fe3+ /Fe2+ separation factor is 1000 ± 150. Effect of different solvents on the Fe3+ extraction by 0.1 M HDEHP measured for
various extraction times. Errors represent one standard deviation from the mean
These results are in agreement with previous studies that have value.
demonstrated that HCl concentrations below 1.0 M give the best
separation between Fe3+ and either Fe2+ [31] or divalent cobalt and Time (minutes) % Fe3+ extracted into organic phase

nickel, which behave similarly to Fe2+ [32]. Cyclohexane n-Heptane Toluene

The extraction of Fe3+ into the organic phase (Fig. 1) is rela- 10 35.92 ± 0.45 47.94 ± 0.88 32.38 ± 0.64
tively slow compared to other metals; for example, the trivalent 20 56.78 ± 2.67 67.86 ± 2.12 46.84 ± 1.22
lanthanides, trivalent actinides, and U(VI) were extracted in min- 30 71.84 ± 1.97 86.11 ± 0.20 51.78 ± 0.45
utes [33–35]. Almost quantitative partitioning (greater than 98.5%) 45 84.53 ± 0.03 91.70 ± 0.19 62.46 ± 0.14
60 93.68 ± 0.03 98.68 ± 0.30 75.04 ± 0.26
of Fe3+ occurs after 1 hour and is in agreement with previous studies
 S.E. Pepper et al. / Analytica Chimica Acta 663 (2010) 172–177 175

Fig. 2. Extraction capacity of 0.1 M HDEHP in n-heptane. The amount of extracted

Fe3+ (circles) was measured after the back-extraction from the organic phase. The
dotted line represents the ideal case (100% of Fe3+ extraction) where the capacity of
the HDEHP is not limiting. The theoretical maximum capacity of 0.1 M HDEHP for
Fe3+ is shown by the solid line.

presented in Fig. 2, the organic phase is saturated at an initial Fe3+

concentration of 35 mM, with only ∼17 mM extracted. In these
experiments, an HDEHP concentration of 100 mM was used. Since
six HDEHP molecules are required to extract one Fe3+ cation (see Eq.
(1)), this concentration of HDEHP allows a maximum of 16.7 mM
Fe3+ to be extracted. The net effect of this Fe3+ “capacity” is that
there will be a limiting Fe3+ concentration that depends on the con-
centration of HDEHP used in the organic phase, and this limit will,
in part, define the error in the analytical approach. Therefore, under
the conditions of the procedure presented here, a maximum con-
centration of 5 mM Fe3+ is recommended, which results in an error
of less than 10%. To measure higher concentrations of Fe3+ , the sam-
ple could be diluted or a higher concentration of HDEHP could be
used up to a maximum of 0.3 M.
For the range of concentrations investigated, the Fe2+ remains
in the aqueous phase and does not interfere with the extraction Fig. 3. Comparison of the results obtained from the ferrozine assay (a) and solvent
of Fe3+ , which is discussed in Section 3.2. For this reason, the Fe2+ extraction method (b) at Fetotal concentration = 0.5 mM. Fe total data are represented
by squares, Fe2+ by triangles, Fe3+ by circles and dashed lines represent the true
concentration limit in this extraction procedure is defined simply
values.
by its solubility in the aqueous phase.

3.2. Mixed oxidation state system and the effect of highly

complexing ligands

In all the experiments reported herein, the ferrous and fer-

Table 2
ric secondary standard solutions were used as the reference. The
Effect of citrate, NTA and EDTA on the detection of iron by the solvent extraction
separation of iron oxidation states in the synthetic mixtures and method (SE) and the ferrozine assay (FA). Total iron concentration = 0.5 mM, with a
analytical sample preparations were conducted in accordance with 1:1 ratio of Fe2+ to Fe3+ . The ratio of total iron to ligand varied from 1:0.1 to 1:5.
the modified ferrozine method and the extraction method using
System Fe:ligand Total iron (mM) [Fe2+ ] (mM) [Fe3+ ] (mM)
our anoxic nitrogen glovebox. The results of the experiments per- ratio
formed on mixed Fe2+ /Fe3+ oxidation state systems using these two SE FA SE FA SE FA
methods are shown in Fig. 3. The total iron concentrations mea- Fe/citrate 1:0.1 0.518 0.576 0.209 0.265 0.232 0.311
sured by ICP-MS were always greater than those determined by 1:1 0.518 0.597 0.222 0.248 0.263 0.349
the ferrozine method. The explanation for this observation is that 1:2 0.528 0.600 0.242 0.269 0.278 0.331
1:5 0.567 0.567 0.259 0.321 0.290 0.246
the reduction of Fe3+ was not complete. However, agreement to
within our target experimental error could be achieved. There is Fe/NTA 1:0.1 0.529 0.498 0.250 0.270 0.276 0.228
also a slight difference between the two methods in the determi- 1:1 0.512 0.499 0.242 0.247 0.244 0.252
1:2 0.516 0.484 0.228 0.254 0.235 0.230
nation of Fe3+ concentration, since the solvent extraction method
1:5 0.524 0.508 0.274 0.258 0.256 0.249
determines the concentration directly, whereas the ferrozine assay
determines the concentration by difference. The results obtained Fe/EDTA 1:0.1 0.517 0.455 0.244 0.270 0.252 0.185
1:1 0.519 0.258 0.320 0.242 0.158 0.015
by both methods, however, agree to within 7 ± 4%. 1:2 0.515 0.235 0.368 0.230 0.141 0.005
In the presence of citrate and NTA (Table 2), the detection of both 1:5 0.500 0.224 0.411 0.217 0.109 0.007
oxidation states is not significantly affected. However, the presence
176 S.E. Pepper et al. / Analytica Chimica Acta 663 (2010) 172–177

Table 3
Log K values for the complexes formed between citrate, NTA or EDTA with Fe2+ or
Fe3+ . Values are for I = 0.1 M at 25 ◦ C for [ML]/[M][L] [41].

Ligand log K

Fe2+ Fe3+

Citrate 4.62 11.2

NTA 8.90 16.0
EDTA 14.3 25.1

of EDTA affects both of these methods with the influence increas-

ing with increasing EDTA concentration. In the HDEHP solvent
extraction procedure, the determination of total iron concentra-
tion in the presence of EDTA is unaffected but the measurement
of the relative concentration of each oxidation state is affected.
More specifically, the measured concentration of Fe2+ is some-
what elevated whereas the concentration of Fe3+ is lower than the
value expected. In the extractions carried out on systems contain-
ing EDTA and only one iron oxidation state, the presence of EDTA
did not affect Fe2+ determination. However, EDTA caused Fe3+ to
remain in the aqueous phase during the initial extraction step. In
the solvent extraction method, EDTA interferes with the formation
of Fe3+ –HDEHP dimers, thus prevents the complete extraction of
Fe3+ into the organic phase, resulting in a lower Fe3+ concentration
than expected. Altering the experimental conditions, for example
increasing the acid concentration and increasing the time for par-
titioning to occur in the initial extraction step, did not resolve this
problem. Therefore, in the mixed oxidation state system, the appar-
ent increase in Fe2+ concentration is due to Fe3+ that remained
in the aqueous phase, which led to a correspondingly lower Fe3+
concentration in the organic phase.
In the ferrozine assay, the measurement of Fe2+ was unaf-
fected by the presence of EDTA but the total iron measurement
and therefore the determination of Fe3+ content were impacted.
Experiments performed on systems containing EDTA and only one
oxidation state of iron (results not shown) revealed that hydrox-
ylamine hydrochloride was unable to quantitatively reduce Fe3+
to Fe2+ because of EDTA complexation and as a result the Fe3+ con-
Fig. 4. Bioreduction of Fe3+ (circles), initially present as Fe3+ -NTA, to soluble Fe2+
centration determined was lower than expected. The effect of EDTA
(triangles) by S. alga. Lactate was utilized as the electron donor. Iron concentra-
on both these methods can be explained in terms of differences in tions were determined using (a) the HDEHP solvent extraction method and (b) the
stability constants. The strength of the complexes formed between ferrozine method. Lactate concentration (squares) was determined using a lactate
Fe2+ or Fe3+ and the three ligands investigated in this study are com- reagent kit.
pared in Table 3. The Fe3+ -EDTA complex is much stronger than
the other Fe3+ organic complexes as well as all the Fe2+ organic shown). Both the solvent extraction procedure and the ferrozine
complexes. method, Fig. 4a and 4b respectively, were able to track the reduc-
In summary, for the range of experimental conditions inves- tion of Fe3+ to Fe2+ over the course of the experiments and gave very
tigated, both the solvent extraction procedure and the ferrozine similar results. This agreement in a relatively complex media appli-
method could not overcome the effects of strong EDTA complexa- cation confirms the applicability of both these analytical methods
tion to permit reliable detection of Fe3+ . Determination of ferrous to the oxidation state specific analysis of iron in environmental
and ferric iron in the presence of EDTA could, however, be done media.
by using ICP-MS to measure the total iron concentration and the
ferrozine method to establish the Fe2+ content. 4. Conclusions

3.3. Reduction of Fe3+ by S. alga under anoxic conditions The simultaneous measurement of Fe2+ and Fe3+ is needed to
establish and understand key redox processes in complex environ-
The analytical procedures described herein to determine dif- mental systems and must be done very carefully to get the correct
ferent oxidation states of iron were applied to biologically active results. All reagents used in the analytical procedure must be oxy-
environmental samples in the presence of a moderately strong gen free, since residual oxygen will easily oxidize trace-levels of
complexant, NTA. Fig. 4 shows the reduction of Fe3+ (as an NTA com- Fe2+ and introduce significant error into the analysis. When the Fe2+
plex) to Fe2+ by S. alga that was performed in separate experiments. concentration is lower than 0.1 mM, this error can reach 100% in the
It was impractical, due to the time-intensive nature of the analyses ferrozine and extraction methods indicating that the ferrous ions
in the glovebox, to perform the ferrozine method and extraction have disappeared from the system. Alternatively, the presence of a
simultaneously. There is an excellent correlation between the uti- strong complexing agent in the system can mask the Fe3+ concen-
lization of lactate, as an electron donor, and the reduction of Fe3+ . tration leading to an overestimation of the Fe2+ content. For these
There is a strong relationship between the production of Fe2+ and reasons, an incorrect result in the Fe2+ and Fe3+ determination is
the growth of the cells over the course of the experiment (data not obtained if analytical conditions are not carefully controlled.
 S.E. Pepper et al. / Analytica Chimica Acta 663 (2010) 172–177 177

The results presented herein show that a solvent extraction pro- [11] S.R. Qui, C. Amrhein, M.L. Hunt, R. Pfeffer, B. Yakshinskiy, L. Zhang, T.E. Madey,
cedure that utilizes HDEHP for the separation of ferrous and ferric J.A. Yarmoff, Characterization of uranium oxide thin films grown from solution
onto Fe surfaces, Appl. Surf. Sci. 181 (2001) 211–224.
iron in aqueous solutions, when combined with ICP-MS analysis, [12] D.T. Reed, J.F. Lucchini, S.B. Aase, A.J. Kropf, Reduction of plutonium(VI) in brine
can provide at least the same degree of accuracy and sensitivity under subsurface conditions, Radiochim. Acta 94 (2006) 591–597.
as the ferrozine method. This approach has the added benefit of [13] K. Nakata, S. Nagasaki, S. Tanaka, Y. Sakamoto, T. Tanaka, H. Ogawa, Reduction
rate of neptunium(V) in heterogeneous solution with magnetite, Radiochim.
also separating oxidation states of multivalent actinides so that it Acta 92 (2004) 145–149.
can be combined with radioanalytical methods (e.g., liquid scintil- [14] D.T. Reed, S.E. Pepper, M.K. Richmann, G. Smith, R. Deo, B.E. Rittmann, Subsur-
lation counting) to analyze these oxidation states at the same time. face bio-mediated reduction of higher-valent uranium and plutonium, J. Alloys
Compounds 444–445 (2007) 376–382.
Both the ferrozine and HDEHP extraction methods were capable of [15] L.L. Stookey, Ferrozine – a new spectrophotometric reagent for iron, Anal. Chem.
detecting both oxidation states of iron in systems containing com- 42 (1970) 779–781.
plexing agents with log K values less than 17, in this case citrate [16] H. Fadrus, J. Maly, Suppression of iron(III) interference in determination of
iron(II) in water by 1,10-phenanthroline method, Analyst 100 (1975) 549–554.
and NTA, and in a more complex system containing microbes.
[17] G.M. Ritcey, A.W. Ashbrook, Solvent Extraction: Principles and Applications to
However, the presence of EDTA affected the oxidation-specific Process Metallurgy, Elsevier, Amsterdam, 1984.
analysis of iron in both methods. The extent of this effect increased [18] A.K. De, S.M. Khopkar, R.A. Chalmers, Solvent Extraction of Metals, Van Nos-
with increasing EDTA concentration. In the solvent extraction pro- trand Reinhold, London, 1970.
[19] K.L. Nash, G.R. Choppin, Separations of f Elements, Plenum Press, New York,
cedure, total iron concentration was unaffected, whereas the Fe2+ 1995.
concentration was higher than expected and the Fe3+ lower than [20] S. Yu, J. Chen, C. Chen, Stripping of Fe(III) extracted by di-2-ethylhexyl phospho-
expected. In the ferrozine method, Fe2+ concentration was unaf- ric acid from sulfate solutions with sulfuric acid, Hydrometallurgy 22 (1989)
267–272.
fected but the total iron and thus Fe3+ concentrations were lower [21] G. Cote, Hydrometallurgy of strategic metals, Solvent Extr. Ion Exch. 18 (2000)
than expected. In both cases, this is attributed to the strength of 703–727.
the Fe3+ -EDTA complex (log K = 25.1). On the basis of these obser- [22] M. Borkowski, J.R. Ferraro, R. Chiarizia, D.R. McAlister, FT-IR study of third phase
formation in the U(VI) or Th(IV)/HNO3 , TBP/alkane systems, Solvent Extr. Ion
vations, we proposed to use the following method in the presence Exch. 20 (2002) 313–330.
of EDTA: determine the total iron concentration directly from the [23] M. Borkowski, J. Krejzler, S. Siekierski, The effect of the pseudohalide ligand
sample using ICP-MS, equipped with a hydrogen reaction cell, on solvent-extraction of trivalent lanthanides, yttrium and americium by tri-
n-butyl phosphate, Radiochim. Acta 65 (1994) 99–103.
determine the Fe2+ concentration, using the modified ferrozine
[24] M. Borkowski, S. Siekierski, Factors affecting the position of Y and actinides(III)
method we described, and determine the Fe3+ concentrations by with respect to lanthanides in the NH4 SCN-ADOGEN-464SCN extraction sys-
difference. This procedure has broad applicability and is indepen- tem, Radiochim. Acta 56 (1992) 31–35.
[25] K.L. Nash, R. Chiarizia, M. Borkowski, P.G. Rickert, E. Out, Radioanalytical tech-
dent of the presence of strong chelating ligands in the aqueous
niques and the characterization of new separations reagents, in: C.A. Laue,
solution. K.L. Nash (Eds.), ACS Symposium Series: Radioanalytical Methods in Interdisci-
Future work is focused on extending the applicability of these plinary Research, Vol. 868, American Chemical Society, Washington, D.C., 2004,
procedures to complex environmental systems and developing pp. 140–160.
[26] M. Borkowski, L. Fuks, Preparation of Cm-242 tracer from irradiated Am-241
a co-extraction approach when multivalent actinides are also by extraction chromatography, Nukleonika 33 (1988) 275–281.
present. [27] W. Smulek, M. Borkowski, Separation of some burn-up monitors from UO2 -Al
nuclear fuel, Nukleonika 31 (1976) 31–44.
[28] H. Ramebäck, M. Skĺlberg, Separation of neptunium, plutonium, americium
Acknowledgements and curium from uranium with di-(2-ethylhexyl)-phosphoric acid (HDEHP) for
radiometric and ICP-MS analysis, J. Radioanal. Nucl. Chem. 235 (1998) 229–233.
This work was performed under the Los Alamos Actinide [29] Y. Sung, Isolation and ecology of bacterial populations involved in reductive
dechlorination of chlorinated solvents, PhD Dissertation, Georgia Institute of
Chemistry and Repository Science Program at the Carlsbad Environ-
Technology, Atlanta, 2005.
mental Monitoring and Research Center operated by New Mexico [30] F. Caccavo, R.P. Blakemore, D.R. Lovley, A hydrogen-oxidizing, Fe(III)-reducing
State University. Funding was provided in part by DOE-Carlsbad microorganism from the Great Bay Estuary, New Hampshire, Appl. Environ.
Microbiol. 58 (1992) 3211–3216.
Field Office and by the Environmental Remediation Science Pro-
[31] A. Haggag, W. Sanad, A. Alian, N. Radroes, Extraction studies on iron, J.
gram (DOE-OBER/OS). Radioanal. Chem. 35 (1977) 253–267.
[32] E. Cerrai, G. Ghersini, Reversed-phase chromatography of Al, Ga, In, Tl and
transition metals of iron group on paper treated with di-(2-ethylhexyl)
References
orthophosphoric acid in chloride medium, J. Chromatogr. 18 (1965) 124–133.
[33] Z. Kolarik, G. Koch, W. Kuhn, Acidic organophosphorous extractants. 18. Rate of
[1] W. Stumm, J.J. Morgan, Aquatic Chemistry, Wiley, New York, 1981, 252-348. lanthanide(III) extraction by di(2-ethylhexyl) phosphoric acid from complex-
[2] D. Langmuir, Aqueous Environmental Geochemistry, Chapter 12, Iron and Sul- ing media, J. Inorg. Nucl. Chem. 36 (1974) 905–909.
fur Geochemistry, Prentice-Hall, Englewood Cliffs, NJ, 1997. [34] P.N. Pathak, G.R. Choppin, Thermodynamic study of metal silicate complexation
[3] R.M. Cornell, U. Schwertmann, The Iron Oxides, Wiley, New York, 2003, pp. in perchlorate media, Radiochim. Acta 94 (2006) 81–86.
491–503. [35] S. Girgin, N. Acarkan, A. Ali Sirkeci, The uranium(VI) extraction mechanism of
[4] Ph. Refait, J.B. Memet, C. Bon, R. Sabot, J.-M.R. Genin, Formation of the Fe(II)- D2EHPA-TOPO from a wet process phosphoric acid, J. Inorg. Nucl. Chem. 251
Fe(III) hydroxysulphate green rust during marine corrosion of steel, Corros. Sci. (2002) 263–271.
45 (2003) 833–845. [36] J.W. Roddy, C.F. Coleman, S. Arai, Mechanism of slow extraction of iron(III)
[5] S.J. Oh, D.C. Cook, H.E. Townsend, Characterization of iron oxides commonly from acid perchlorate solutions by di(2-ethylhexyl)phosphoric acid in octane,
formed as corrosion products on steel, Hyperfine Interact. 112 (1998) 59–65. J. Inorg. Nucl. Chem. 33 (1971) 1099–1118.
[6] A.V.R. Kumar, R.K. Nigam, S.S. Monga, G.N. Mathur, Effect of inhibitors on the [37] H. Matsuyama, Y. Miyake, Y. Izumo, M. Teramoto, Kinetics and mechanism
nature of corrosion products of mild steel by Mossbauer and FTIR spectroscopy, of metal extraction with acidic organophosphorous extractants. 2. Extraction
Anti-Corros. Methods Mater. 49 (2002) 111–117. mechanism of Fe(III) with di(2-ethylhexyl) phosphoric acid, Hydrometallurgy
[7] T. Missana, C. Maffiotte, M. Garcia-Gutierrez, Surface reactions kinetics 24 (1990) 37–51.
between nanocrystalline magnetite and uranyl, J. Colloid Interface Sci. 261 [38] http://www.speckanalytical.com.
(2003) 154–160. [39] J. Szymanowski, G. Cote, I. Blondet, C. Bouvier, D. Bauer, J.L. Sabot, Interfacial
[8] B. Grambow, E. Smailos, R. Geckeis, R. Muller, H. Hentschel, Sorption and activity of bis(2-ethylhexyl) phosphoric acid in model liquid-liquid extraction
reduction of uranium(VI) on iron corrosion products under reducing saline systems, Hydrometallurgy 44 (1997) 163–178.
conditions, Radiochim. Acta 74 (1996) 149–154. [40] D.F. Peppard, G.W. Mason, W.J. Driscoll, R.J. Sironen, Acidic esters of orthophos-
[9] B.H. Jeon, B.A. Dempsey, W.D. Burgos, M.O. Barnett, E.E. Roden, Chemical reduc- phoric acid as selective extractants for metallic cations–tracer studies, J. Inorg.
tion of U(VI) by Fe(II) at the solid-water interface using natural and synthetic Nucl. Chem. 7 (1958) 276–285.
Fe(III) oxides, Environ. Sci. Technol. 39 (2005) 5642–5649. [41] A.E. Martell (Ed.), NIST Standard Reference Database 46, Version 8.0, NIST Crit-
[10] J. Farrell, W.D. Bostick, R.J. Jarabek, J.N. Fiedor, Uranium removal from ground ically Selected Stability Constants of Metal Complexes, 2004.
water using zero valent iron media, Ground Water 37 (1999) 618–624.