7198 Phys. Chem. Chem. Phys., 2011, 13, 7198–7206 This journal is c the Owner Societies 2011

Cite this: Phys. Chem. Chem. Phys., 2011, 13, 7198–7206

Iron oxyhydroxide colloid formation by gamma-radiolysis

P. A. Yakabuskie,aJ. M. Joseph,

aP. Keech,

aG. A. Botton,

bD. Guzonas

cand

J. C. Wren*a

Received 10th January 2011, Accepted 15th February 2011

DOI: 10.1039/c1cp20084d

Gamma-irradiation of deaerated aqueous solutions containing FeSO4 leads to the formation

of uniform-sized colloidal particles of g-FeOOH. At short irradiation times, or in solutions with

a low initial [Fe2+]0, spherical particles with a size less than 10 nm are formed. These primary

particles grow to form a dendritic structure upon longer irradiation, and the final size of the

large particles is B60 nm with a very narrow size distribution. Further prolonged irradiation

does not change the final particle size. The narrow size distribution is attributed to rapid

homogeneous radiolytic oxidation of soluble Fe2+ to relatively insoluble Fe3+ hydroxides

[Fe(H2O)6�n(OH)n]3�n leading to particle nucleation by spontaneous condensation. These primary

particles then grow into g-FeOOH particles with a dendritic structure. The final size reached at

long times is regulated by the steady-state redox conditions established during long-term

irradiation at the aqueous–solid interface.

1. Introduction

In the heat transport system of a water-cooled nuclear reactor,

corrosion of materials and wear of in-core surfaces can release

metal ions (corrosion products) into the reactor coolant. These

metallic species can take the form of dissolved ions or suspended

particulates (including colloids) that may avoid capture by ion

exchange and filtration systems and can subsequently be

transported into the reactor core where they are subjected to

high intensity radiation fields. Neutron activation of these

metallic species can generate radioactive isotopes that can

recirculate with the coolant (activity transport), creating a potential

source of radiation exposure to plant workers; in addition, the

corrosion products can deposit on piping surfaces, decreasing

heat transfer efficiencies. The intense g-radiation fields in the

reactor core can also affect the water chemistry of the coolant

and induce changes to the redox states of metallic species in

materials that are in contact with the coolant.

Many studies have examined the influence of g-radiation on

the surface oxides and corrosion rates of iron-based materials

such as carbon steel in aqueous systems,1–4 but little research

has focused on the metal speciation in irradiated aqueous

systems containing dissolved metal ions. Radiation-induced

nanoparticle formation in aqueous solutions has been investigated

previously,5–8 but most studies have been limited to noble

metal nanoparticles. These studies primarily examined the

ability of hydrated electrons (�eaq�) and hydrogen atoms (�H)

from the radiolytic decomposition of water to reduce dissolved

metal ions to insoluble metallic particles. The syntheses of

some metal oxides by radiolysis have also been reported.9–11 In

that work, metal-oxide nanoparticles were formed via radiolytic

reduction of a metal oxide from a higher oxidation state to a

lower oxidation state, generally using a short-duration radia-

tion pulse and chemical scavengers to optimize the reducing

radical concentrations in solution. Under continuous steady-

state irradiation conditions (as would be present in reactor

systems), the presence of these scavengers complicates the

water radiolysis chemistry on longer time scales and cannot

be used to elucidate the mechanism of colloid formation in

‘‘pure’’ aqueous systems. We are not aware of any reported

studies on the formation of iron oxide nanoparticles utilizing

the oxidizing radiolytic products of water formed during long-

term g-irradiation in aqueous solutions in the absence of

chemical scavengers.

When exposed to highly ionizing radiation, water decomposes

to form several oxidizing and reducing species (primary

radiolysis products):12–17

ð1Þ

Other reactive species such as O2 and�O2

� are also formed as

secondary species stemming from reactions of these primary

radiolysis products. The radiolysis products react rapidly with

each other and, if present, any additional reactive species in

solution. The probability of ionizing radiation directly inter-

acting with solute species in dilute solutions is very small since

the rate of energy transfer from radiation particles or photons

aDepartment of Chemistry, the University of Western Ontario,London, Ontario, Canada N6A 5B7. E-mail: jcwren@uwo.ca;Fax: +1 519-661-3022; Tel: +1 519-661-2111 ext 86339

bDepartment of Materials Science and Engineering,McMaster University, Hamilton, Ontario, Canada L8S 4M1

cAtomic Energy of Canada Ltd., Chalk River, Ontario,Canada K0J 1J0

PCCP Dynamic Article Links

www.rsc.org/pccp PAPER

Publ

ishe

d on

11

Mar

ch 2

011.

Dow

nloa

ded

by U

nive

rsity

of

Cal

ifor

nia

- Ir

vine

on

29/1

0/20

14 0

6:57

:00.

View Article Online / Journal Homepage / Table of Contents for this issue

This journal is c the Owner Societies 2011 Phys. Chem. Chem. Phys., 2011, 13, 7198–7206 7199

to an interacting medium depends mainly on the density of

electrons in the medium. Thus, radiolysis-driven processes are

predominantly solvent-oriented and the effect of a radiation

field on the chemistry of solute species occurs mainly via reac-

tions of the solute species with the radiolytic decomposition

products of solvent molecules.15

It has been observed that, when noble metal ions are the

solute species, the reduction of the dissolved metal ions by

water radiolysis products leads to nanoparticle formation. For

example, silver nanoclusters with very uniform and narrow

size distributions have been formed by irradiation of aqueous

solutions containing Ag+ in electron accelerators and gamma

irradiation cells.5 The formation of nanoclusters of bi-metallic

species under g-irradiation and the formation of gold nano-

particles in a room temperature ionic liquid (RTIL) have been

also observed.7,8 These studies on radiation-induced nano-

particle formation utilize the strong reducing power of radio-

lytically produced �eaq� (or �H) to reduce the metal ions

present to a zero-valence state. These studies have demon-

strated that the size of the clusters so formed can be ‘tuned’ by

manipulating the radiation dose rate. The simplicity of this

type of radiation-induced particle formation process gives

it an advantage over other particle formation processes that

require chemical additives and stabilizers since those species

can lead to undesired effects on the nanoparticle structure and

chemistry.

In this study we report on the oxidative formation of

uniform-sized colloidal particles of g-FeOOH by gamma-

irradiation of deaerated aqueous solutions of FeSO4 in slightly

acidic (pH 5.5) water. Analyses of dissolved ferric and

ferrous species as well as molecular water radiolysis products

(H2, O2 and H2O2) were performed as a function of irradiation

time. The Fe3+-containing colloidal particles were characterized

using transmission electron microscopy (TEM), selected area

electron diffraction (SAED), Fourier transform infrared (FTIR)

and UV-Vis absorption spectroscopy.

2. Experimental

All solutions were freshly prepared before each experiment

with water purified using a NANOpure Diamond UV ultra-

pure water system; the resulting water had a resistivity of

18.2 MO cm. High-purity iron (II) sulfate was obtained from

Sigma-Aldrich (purity Z 99%) and used without further

purification. Pure water was first purged with ultra high purity

argon (impurity content 0.001%) for at least one hour and

then transferred to an argon-filled glove box for solution

preparation. The oxygen concentration in the glove box was

maintained below 0.1 volume percent. Solutions containing

dissolved Fe2+ in the concentration range 5 � 10�5 to

5� 10�4 mol dm�3 were prepared and the pHs of the solutions

were adjusted to 5.5 using sulfuric acid, as measured with a

Mettler Toledo pH meter inside the glove box. The solutions

were then transferred into 20 ml glass vials (Agilent Technologies)

leaving no headspace and sealed using PTFE silicon septa. The

sample vials were irradiated in a 60Co gamma cell (MDS

Nordion) as discussed in detail in a previous publication.16

The gamma source provided a uniform absorption dose rate of

6.7 kGy h�1 in the water samples.

Following irradiation, the concentration of Fe2+ in a

solution was determined using the ferrozine method,18 in

which ferrozine reacts with Fe2+ to form a coloured complex

that absorbs light at 563 nm with a molar extinction coefficient

of 27 900 dm3 mol�1 cm�1.19 Subsequent reduction of Fe3+

species with hydroxylamine hydrochloride (1.4 mol dm�3 in

concentration) allowed for the measurement of the total iron

concentration in solution with the ferrozine method. The Fe3+

concentration was then calculated from the total iron concentra-

tion by subtracting the previously determined concentration of

Fe2+. The concentration of Fe3+ determined by this method

may include contributions from completely solvated Fe3+ as

well as FeIII oxides or hydroxides in colloidal particles that

were transferred to the sampling syringe and underwent

reduction to Fe2+ by hydroxylamine hydrochloride. Inter-

ference tests were performed to show that the presence of the

molecular water radiolysis product H2O2, which remains

in the post-irradiated solutions, does not interfere with the

Fe2+ and Fe3+ measurements by the ferrozine method.

The concentration of hydrogen peroxide in the irradiated

solution was determined at the end of the irradiation period

by the Ghormley tri-iodide method20 in which I� is oxidized

to I3� by H2O2 in the presence of ammonium molybdate

as a catalyst. The I3� concentration is determined spectro-

photometrically; I3� has a maximum absorption at 350 nm

with a molar extinction coefficient of 25 500 dm3 mol�1 cm�1.21

Dissolved Fe2+ and Fe3+, when present at the concentra-

tions observed at steady state, did not interfere with the

H2O2 analysis. UV-Visible spectrophotometry was also used

to monitor the change in the absorbance properties of the

irradiated solution as a function of irradiation time. All

of the spectrophotometric measurements were carried out

using a diode array detector (BioLogic Science Instruments).

Using this method, the detection limit for [H2O2] was

3 �10�6 mol dm�3 and the uncertainties in the measurement

arising from sampling and instrumental errors were estimated

to be �0.2% at the lower end of the measured concentra-

tion range and �0.05% at the higher end of the measured

range.

For the analysis of dissolved H2 and O2, one half (10 ml) of

the irradiated sample solution was transferred to a new 20 ml

vacuum-sealed vial using a gas-tight syringe. Equilibration of

the gases between the headspace above the solution and in the

aqueous phase was quickly established. A gas sample was

extracted from the headspace using a gas-tight syringe with a

luer lock (Agilent Technologies) and was analyzed using Gas

Chromatography–Mass Spectrometry (GC–MS, 6580, Agilent

Technologies) as discussed elsewhere in detail.16 Using this

method, the detection limit for the aqueous concentration of

H2 was 1.0 � 10�5 mol dm�3 and the uncertainties in the

measurement arising from sampling and instrumental errors

were estimated to be �50% at the low end of the measured

concentration range and �0.005% at the high end of the con-

centration range. Although the gas sampling was performed in

a glove box, where the concentration of oxygen in the gas

phase was maintained below 1000 ppm, oxygen ingress into

the sampling syringe and the GC column contributed to large

background O2 level, limiting the accurate measurements

of the aqueous concentration of radiolytically produced O2.

Publ

ishe

d on

11

Mar

ch 2

011.

Dow

nloa

ded

by U

nive

rsity

of

Cal

ifor

nia

- Ir

vine

on

29/1

0/20

14 0

6:57

:00.

View Article Online

7200 Phys. Chem. Chem. Phys., 2011, 13, 7198–7206 This journal is c the Owner Societies 2011

The O2 analysis was used mainly to observe any increase in O2

and not to determine the dissolved O2 concentration.

The diameter and shape of the iron nanoparticle clusters

formed in the irradiated solutions were directly measured

using transmission electron microscopy (TEM) with the electron

microscope operated at 80 keV. The TEM samples were

prepared by dipping a carbon-coated copper grid into the

irradiated solution and then drying the sample grid in flowing

air. Selected area electron diffraction (SAED) patterns for

the iron colloids were collected at various loci on the high

resolution TEM sample grids using an acceleration voltage of

200 kV. Fourier transform infrared (FTIR) spectroscopy

measurements of the iron nanoparticles were performed using

an IR spectrophotometer with Fourier transformation (Bruker)

in the 4000 to 400 cm�1 frequency range. A sample of solution

from an irradiated test vial was centrifuged and the particles

thus collected were transferred to a glass plate and air dried.

The dried sample was incorporated into a KBr pellet for the

FTIR measurement. Spectra of standard samples of different

oxides of iron were also measured using the same method.

3. Results

The optical images of the irradiated solutions and TEM

images of the colloidal particles formed are shown in Fig. 1.

These results were obtained by irradiating deaerated solutions

initially containing 5 � 10�4 mol dm�3 FeSO4 at pH 5.5

at a dose rate of 6.7 kGy h�1. Over the irradiation period,

the appearance of the solution changed from colorless to

increasingly yellow in colour while remaining clear. The

TEM images of the colloidal particles formed at short irradia-

tion times (20 min) show spherical particles with a uniform

size on the order of 10 nm. Larger particles, approximately

60 nm in size, started to appear at longer irradiation times

(Z60 min). These larger particles showed a dendritic structure.

Further prolonged irradiation did not lead to an additional

increase in the final size of the particles formed or the size

distribution. However, longer irradiation did result in an

increase in the number of larger particles formed.

Fig. 2 shows TEM images of particles formed during

irradiation of a lower concentration iron sulfate solution

(1 � 10�4 mol dm�3 FeSO4 at pH 5.5) as a function of irradia-

tion time. For this lower concentration solution, the particles

grew to their final size faster (in less than 20 min) and this final

particle size (on the order of 20 nm) was smaller than that

observed for more concentrated iron solutions. The particles

formed at the lower iron concentration were spherical and did

not display a dendritic structure. Similar results were observed

for the lowest iron concentration studied (5 � 10�5 mol dm�3)

(not shown), although the particles formed were even smaller

than those measured for the 1 � 10�4 mol dm�3 solution.

The ferrous and ferric concentrations of the irradiated

solutions were determined by the ferrozine UV-Vis

method. The results from the solutions initially containing

[Fe2+]0 = 5 � 10�4 mol dm�3 are presented in Fig. 3b. The

conversion of Fe2+ to Fe3+ species was already significant at

20 min, the time of the first measurement. The concentrations

of both redox states then changed more slowly, with the

Fig. 1 (a) Optical images of the bulk solutions, and (b) TEM images showing the size and morphology of iron nanoparticles formed by the

irradiation of [Fe2+]0 = 5 � 10�4 mol dm�3 deaerated solution at pH 5.5 and a dose rate of 6.7 kGy h�1, shown as a function of irradiation time.

Publ

ishe

d on

11

Mar

ch 2

011.

Dow

nloa

ded

by U

nive

rsity

of

Cal

ifor

nia

- Ir

vine

on

29/1

0/20

14 0

6:57

:00.

View Article Online

This journal is c the Owner Societies 2011 Phys. Chem. Chem. Phys., 2011, 13, 7198–7206 7201

concentrations reaching steady state after approximately

3 h of irradiation. At a lower initial Fe2+ concentration of

1 � 10�4 mol dm�3 (Fig. 3a), the same trend was observed,

but the steady state was achieved earlier, after about 2 h. It

must be noted that the mass balance for iron in the irradiated

solutions was poor; only 90% of the initially dissolved iron

was detected in the solution, see Fig. 3. We assume that the

‘missing’ iron collected at the bottom of the sample vial by

gravitational settling and, hence, escaped the aqueous

sampling used for the ferrozine analysis.

The concentrations of stable water radiolysis products, H2,

O2 and H2O2, were also monitored as a function of irradiation

time by gas chromatography and UV-Vis spectrophotometry,

respectively. In the absence of initially dissolved iron, the

concentration of H2 produced from g-radiolysis of water at

the studied pH was below the experimental detection limit at

irradiation times up to 6 h, Fig. 4. The presence of initially

dissolved iron raised the concentration of dissolved H2 from

below the detection limit (1 � 10�5 mol dm�3) to measurable

values within 20 min, and the concentration of H2 measured

over time was nearly independent of the initial dissolved iron

Fig. 2 TEM images showing the size and morphology of iron nanoparticles formed by the irradiation of [Fe2+]0 = 1 � 10�4 mol dm�3 deaerated

solution at pH 5.5 and a dose rate of 6.7 kGy h�1 as a function of irradiation time.

Fig. 3 The steady-state concentrations of Fe2+ (m) and Fe3+ ( )

evolved as a function of irradiation time for a deaerated solution of

(a) [Fe2+]0 = 1� 10�4 mol dm�3; and (b) [Fe2+]0 = 5� 10�4 mol dm�3

at pH 5.5 and a dose rate of 6.7 kGy h�1.

Fig. 4 The dissolved H2 concentration measured as a function of

irradiation time for deaerated solutions initially containing Fe2+

concentrations of 0 mol dm�3 (i.e., below experimental detection

limit); 1 � 10�4 mol dm�3 (’); and 5 � 10�4 mol dm�3 ( ) at

pH 5.5 and a dose rate of 6.7 kGy h�1.

Publ

ishe

d on

11

Mar

ch 2

011.

Dow

nloa

ded

by U

nive

rsity

of

Cal

ifor

nia

- Ir

vine

on

29/1

0/20

14 0

6:57

:00.

View Article Online

7202 Phys. Chem. Chem. Phys., 2011, 13, 7198–7206 This journal is c the Owner Societies 2011

concentration. This initial rapid rise in concentration was

followed by a slower increase in [H2] with irradiation time.

Conversely, the concentration of radiolytically produced H2O2

remained below the detection limit of 3 � 10�6 mol dm�3; the

O2 concentration did not increase as a function of irradiation

time for all iron concentrations studied (data not shown).

Since the net mass balance of radiolytic decomposition of

water dictates that the total number of hydrogen atoms

produced must be twice that of oxygen, these results suggest

that the oxygen from water decomposition by radiolysis was

consumed to form solid iron oxide/oxy-hydroxide species.

To determine the chemical composition and phase of the

nanoparticles formed, the FTIR spectra of the colloidal particles

were compared with the FTIR spectra of standard powder

samples of various mineral iron oxides and oxy-hydroxides. In

Fig. 5 only two standard spectra are shown, for simplicity, with a

colloidal particle spectrum. The colloidal particle spectrum

matches that of g-FeOOH well, except for three additional peaks

located between 1000–1300 cm�1. These peaks are characteristic

of S–O stretches and are attributed to terminal SO42� groups

that adsorbed onto the g-FeOOH nanoparticle surface during

the spectroscopy sample preparatory drying stage.22

The SAED patterns of the colloidal particles following 5 h

irradiation were also collected and compared to ED reference

spectra of various iron oxide/oxyhydroxides. Fig. 6 shows a

characteristic ring pattern from the colloidal samples along

with a corresponding reference intensities profile. The small

size of the crystallites caused broadening of the observed

reflections in the diffraction pattern. However, intense reflec-

tions at 2.53 A and 1.53 A, and a weak reflection at 2.1 A are

consistent with the g-FeOOH reference structure (ICSD file

108876). The colloid sample diffraction patterns were found to

match the referenced line pattern for g-FeOOH at all locations

analysed, confirming that the composition of the nanoparticles

is homogeneous and single phase g-FeOOH. In addition, the

TEM images and SAED patterns obtained for colloidal

particles formed in the same initial ferrous solution but

irradiated for much longer time (20 h) were the same.

Fig. 5 FTIR spectrum of iron nanoparticles formed by the radiolysis

of a [Fe2+]0 = 5 � 10�4 mol dm�3 deaerated solution at pH 5.5 and a

dose rate of 6.7 kGy h�1.

Fig. 6 The SAED ring pattern and the associated diffraction

spectrum (inset) of iron nanoparticles formed by the radiolysis of a

[Fe2+]0 = 5 � 10�4 mol dm�3 deaerated solution at pH 5.5 and a dose

rate of 6.7 kGy h�1 for 5 hours.

Fig. 7 UV-Vis spectra of irradiated [Fe2+]0 = 5 � 10�4 mol dm�3

deaerated solutions at pH 5.5 and a dose rate of 6.7 kGy h�1.

(a) Spectra of the initial unirradiated and g-irradiated ferrous sulfate

solutions (solid lines; irradiation times from bottom to top are 0, 20,

40, 60, 120, 240 and 300 minutes), along with the spectrum for

unirradiated ferric nitrate solution (dashed line). (b) Spectra of the

irradiated ferrous sulfate solutions with the absorption contribution of

the hydrated ferric ion removed (irradiation times from bottom to top

are 20, 40, 60, 120, 240 and 300 minutes).

Publ

ishe

d on

11

Mar

ch 2

011.

Dow

nloa

ded

by U

nive

rsity

of

Cal

ifor

nia

- Ir

vine

on

29/1

0/20

14 0

6:57

:00.

View Article Online

This journal is c the Owner Societies 2011 Phys. Chem. Chem. Phys., 2011, 13, 7198–7206 7203

The UV-Vis absorption spectra of the irradiated solutions

were also compared with those from standard solutions con-

taining Fe2+ (prepared with FeSO4) and Fe3+ (prepared with

Fe(NO3)3), Fig. 7a. The spectra of irradiated solutions show a

UV-Vis absorption band centered atB304 nm that is associated

with the presence of hydrated Fe3+ species (from comparison

with the Fe(NO3)3 solution spectrum also shown in Fig. 7a)

and an additional band at B370 nm. The contribution of

absorption by hydrated Fe3+ species was subtracted from

the UV-Vis absorption spectra of the irradiated samples

(by matching the peak intensities at 304 nm), and the resulting

‘hydrated-Fe3+ subtracted’ spectra show a well-defined absorp-

tion band at B370 nm, Fig. 7b. This peak is tentatively

attributed to absorption (and/or scattering) by the colloidal

particles formed by gamma-irradiation. At a given [Fe2+]0, the

peak position and width of the absorption band at 370 nm

did not change significantly, but the absorbance intensity

increased with irradiation time. This is consistent with an

increase in concentration, but not size, of the g-FeOOH

nanoparticles as a function of irradiation time, as was also

observed in the TEM images.

4. Discussion

Under g-irradiation the concentrations of water radiolysis

products (reaction (1)) reach pseudo-equilibrium (or steady

state) very quickly at pH values o9.15–17 In the presence of

dissolved species at low concentrations (o10�2 mol dm�3), the

steady-state concentrations of radiolysis products dictate the

bulk aqueous chemistry of the dissolved species. For dissolved

metal ion redox reactions, the key radiolytic products of water

are the �OH and �eaq� radicals due to their high oxidizing and

reducing powers. The elementary reactions of ferric and

ferrous ions with water radiolysis products have been studied

and their rate constants are well established:23

Fe2+ + �OH - Fe3+ + OH�

kR2= 3.2 (�1.0) � 108 dm3 mol�1 s�1 (2)

Fe3+ + �eaq� - Fe2+

kR3= 6 � 1010 dm3 mol�1 s�1 (3)

The ferric and ferrous ions in these reactions are also in fast

hydrolysis equilibria:

Fe3+ + nH2O # Fe(H2O)6�n(OH)n(2�n)+ + nH+ (4)

Fe2+ + mH2O # Fe(OH)m(2�m)+ + mH+ (5)

where n and m are integers from 1 to 4 and from 1 to 3 for

reaction (4) and (5), respectively. The rate constants listed

above, kR2and kR3

, are the net rate constants including the

contributions of all the iron species involved in the hydrolysis

equilibria (4) and (5). Due to the fast hydrolysis equilibria,

all the hydrated ferric or ferrous species will be, hereafter,

referred to as Fe3+(aq) or Fe2+(aq) respectively.

Reactions (2) and (3) are expected to be fast due to the

relatively high concentrations of �OH and �eaq� generated

during radiolysis, and the dynamic system reaches initial

steady state very quickly. The conversion from ferrous to

ferric ion was, indeed, observed to be initially very fast, more

than 30 to 50% conversion after 20 min. However, this change

was followed by a period of progressively slower increase in

the ferric iron concentration which is not predicted by a

homogeneous aqueous reaction kinetic model.15 At longer

times (>120 min) the relative concentrations of ferrous and

ferric ions did not change, but only 90% of the initially

dissolved iron was detected in the samples irradiated. The

ferrozine method used to detect iron in solution can, in

principle, reduce suspended FeIII colloidal particles, but the

water sampling method may not collect any gravitationally-

settled particles. This may account for the apparent loss of

iron in the sampled solutions.

To explain the observed three stage ferrous to ferric con-

version behaviour we propose a series of processes, schematically

shown in Fig. 8. The first stage (t o 20 min) is the nucleation

and primary particle formation stage. Since the reactions of�eaq

� and �OH with other radiolytic products of water as well

as ferrous and ferric ions are very fast, the concentrations of

the radicals reach pseudo-steady state very quickly (within ms)

under continuous irradiation. Thus, [Fe2+(aq)] and [Fe3+(aq)]

would be in pseudo-equilibrium, with the equilibrium con-

centrations determined by:

½Fe3þðaqÞ�t½Fe2þðaqÞ�t

¼ ½Fe3þðaqÞ�t½Fe2þðaqÞ�0 � ½Fe3þðaqÞ�t

� kR2� ½�OH��t

kR3� ½�e�aq�t

ð6Þ

Under the conditions studied, this ratio is greater than 1 since

the steady-state concentration of �OH is typically 2–3 orders

of magnitude higher than that of �eaq� due to the higher

reactivity of �eaq�.16 This would lead to supersaturation of

Fe3+(aq) in the solution.

The solubility of ferric hydroxide species is several orders of

magnitude lower than that of ferrous hydroxide species at

pH o 8.5; for example, at pH 5.5, the solubilities of

Fe(OH)3 and Fe(OH)2 are on the order of 10�8 mol dm�3

and 10�1 mol dm�3, respectively.24 Thus, although the initial

solution contained Fe2+ at a concentration far below its

solubility limit, the net radiolytic oxidation of Fe2+(aq)

produces Fe3+(aq) at a concentration that is well above the

ferric ion solubility limit. (Note that since the aqueous chemistry

in the presence of ionizing radiation is not in thermal equilibrium,

the solubilities of iron species under irradiation conditions

could be somewhat different than their solubilities at thermal

equilibrium.) This supersaturation is then followed by condensa-

tion/dehydration of hydrated ferric species to form primary

particles composed of Fe(OH)3(s):

Fe3+(aq) + 3OH�# Fe(H2O)3(OH)3 # Fe(OH)3(s)+3H2O

(7)

The condensed ferric hydroxide could be partially hydrated

(at least on the surface), but hereafter, Fe(OH)3(s) will be used

to represent the condensed ferric hydroxide particles.

This primary particle formation occurs rapidly under gamma-

irradiation, compared to thermal processes, because of the

rapid homogeneous generation of supersaturated Fe3+(aq)

throughout the solution. Compared to this process, chemical

syntheses of iron oxy-hydroxide nanoparticles by other processes

are slower. For example, in forced hydrolysis methods, in

Publ

ishe

d on

11

Mar

ch 2

011.

Dow

nloa

ded

by U

nive

rsity

of

Cal

ifor

nia

- Ir

vine

on

29/1

0/20

14 0

6:57

:00.

View Article Online

7204 Phys. Chem. Chem. Phys., 2011, 13, 7198–7206 This journal is c the Owner Societies 2011

which the generation of primary particles is also based on the

condensation and dehydration of ferric ions, the starting iron

species are soluble ferric species and the condensation/dehy-

dration of the ferric species is achieved by changing tempera-

ture to shift the hydrolysis equilibria of ferric species.25 The

rapid and high production of powerful redox radicals by

ionizing radiation shortens the time required for production

of the primary particles (and their growth, discussed below),

contributing to the uniform size distribution of the resultant

particles.

The concentration of radiolytically-produced H2 is inversely

proportional to the concentration of �OH.15–17 The significant

increase in the net radiolytic production of H2 in the presence

of ferrous iron can then be attributed to the competitive

reaction kinetics of �OH with Fe2+(aq), H2 and H2O2. The

near independence of the net production of H2 on the initial

ferrous concentration is attributed to the initial steady-state

balance, reaction (7), and the consumption of oxygen during

the formation of Fe2+(aq) to Fe3+(aq), reactions (2) and (4) to (6).

This will also affect the concentrations of hydrated electron

and H2O2 via complicated kinetics, since the hydrated electron

undergoes competitive reactions with H+, Fe3+(aq), H2O2,

O2, etc.15–17 The net result of these coupled reactions is that

the ratio of Fe2+(aq) to Fe3+(aq) in the first stage (o20 min) is

not sensitive to the initial ferrous concentration. In summary,

the first stage (o20 min) involves fast net radiolytic oxidation

of soluble Fe2+(aq) to Fe3+(aq), followed by poly-condensation/

dehydration to form primary amorphous ferric-hydroxide

(Fe(OH)3) particles.

In the second stage, the conversion from ferrous to ferric

species still occurs, but at a slower rate compared to that in

stage 1, Fig. 3. The duration of this stage depends on the initial

ferrous ion concentration at the studied pH; it lasts approxi-

mately 20 min for the lower concentration case and up to

200 min at the higher concentrations tested. During this stage

the particles grow and establish a specific oxide or oxyhydroxide

phase. The chemical composition, phase and size of the

nanoparticles eventually formed depend on pH, temperature,

ionic strength, aqueous redox conditions and initial ferrous

ion concentration since these parameters will affect the free

energy of formation of the solid phase, particle surface reactivity

and charge, and, hence, the solid-water interfacial partitioning

of ferric species.26–28 Studies are currently underway to quantify

the effects of these parameters on iron oxide nanoparticle

formation. In the second stage, the redox reactions (2) and

(3) can continue on the surface of the particles as well as in the

aqueous phase. The increasing importance of heterogeneous

reactions on the particle surface explains the slowing in the

rate of conversion of ferrous to ferric ion. Molecular products

from water radiolysis, such as H2O2, and the secondary

Fig. 8 The proposed 3 stage growth mechanism of g-FeOOH particles induced by the g-irradiation of deaerated FeSO4 solutions at pH 5.5.

Publ

ishe

d on

11

Mar

ch 2

011.

Dow

nloa

ded

by U

nive

rsity

of

Cal

ifor

nia

- Ir

vine

on

29/1

0/20

14 0

6:57

:00.

View Article Online

This journal is c the Owner Societies 2011 Phys. Chem. Chem. Phys., 2011, 13, 7198–7206 7205

radiolysis products such as O2 and �O2�, can accumulate to

substantial concentrations and, hence, can participate in the

redox reactions; see further discussion below.

Under the conditions of this study, the final nanoparticles

formed are composed of g-FeOOH, as identified by FTIR,

UV-Vis and SAED, Fig. 5 to 7. The formation of g-FeOOH

from the primary Fe(OH)3(s) particles requires either dehydra-

tion of Fe(OH)3(s):

Fe(OH)3(s) - g-FeOOH(s) + H2O (8)

or oxidation of the ferrous ion by �OH or H2O2 on the surface

of the primary particles:

Fe2+(aq) + �OH + H2O - g-FeOOH(s) + 2H+

on primary particles

(9a)

2Fe2+(aq) + H2O2 + 2H2O - 2g-FeOOH(s) + 4H+

on primary particles

(9b)

At room temperature with radiation present, the phase con-

version from solid ferric hydroxide to lepidocrocite, reaction

(8), is very slow relative to the formation of g-FeOOH by

reaction of dissolved ferrous ion with water radiolysis products�OH and H2O2, and, hence, reaction (9) is the preferable path

for g-FeOOH growth. Reaction (9) is also an electrochemical

redox reaction and should occur at a faster rate on a more

conducting surface. Compared with insulating Fe(OH)3,

g-FeOOH is a better electrical conductor, as it exhibits semi-

conductive properties,29 and has been known to support the

reduction or oxidation of adsorbed species such as H2O2.30

The growth of g-FeOOH is accelerated as the more conducting

g-FeOOH layer thickens and this could explain the observed

particle size growth behaviour.

The TEM images (Fig. 1) show that the particles formed

following 60 min of irradiation have only two sizes, B20 nm,

the size of the primary particles, and B60 nm, the final size of

the g-FeOOH nanoparticles, and that longer irradiation does

not affect the final particle size, but only increases the number

of the final-sized particles. These results suggest that once a

distinct layer of g-FeOOH is formed on the surface of a

primary particle, further iron oxidation by water radiolysis

products occurs at a faster rate on the g-FeOOH surface layer

than on the Fe(OH)3 surface layer. This preferential oxidation

of ferrous to ferric species on the g-FeOOH layer over the

primary Fe(OH)3 particles results in the g-FeOOH nano-

particles growing to the final size one at a time. This explains

the absence of intermediate-sized particles in the TEM images.

The final size is reached when the rate of reduction of

g-FeOOH to Fe2+(aq) by reducing water radiolysis products,

such as �eaq� or �O2

�, reaction (10), becomes equal to the rate

of the g-FeOOH formation, reaction (9):

g-FeOOH + �eaq� + H2O - Fe2+(aq) + 3 OH� (10a)

g-FeOOH+ �O2�+H2O- Fe2+(aq) + O2 + 3 OH�

(10b)

At the early stage of g-FeOOH growth in stage 2, the ferrous

oxidation reaction (9) on the surface of the particles is faster than

the ferric reduction reaction (10). However, as the g-FeOOH

particles grow in a dendritic shape (see further discussion below),

their surface areas increase relative to the primary particles

that have near spherical shapes and the rate of ferric reduction

(reaction 10) increases. The g-FeOOH particles reach their

final size when the ferric reduction rate equals the ferrous

oxidation rate. Note that the concentrations of molecular

water radiolysis products such as H2O2 and H2 and the

secondary radiolysis products such as O2 and �O2� can also

reach relatively high concentrations at longer times.15–17 Thus,

not only the redox reactions of radical products but also those

of the molecular products should be taken into account in the

longer term redox kinetics occurring on surfaces.

In stage 2, the concentrations of water radiolysis products,�OH, H2O2,

�eaq� and �O2

�, are at (pseudo-) steady state, and

reactions (2) and (3) continually occur in the aqueous phase.

In addition, Fe3+(aq) is in aqueous-solid phase equilibrium

with ferric hydroxide via reaction (7). These (pseudo-) equilibrium

reactions maintain the ratio of the aqueous ferric and ferrous

concentrations with time, see eqn (6). Thus, although most of

the initially added Fe2+(aq) has been oxidized to Fe3+(aq)

and then quickly converted to the primary Fe(OH)3 particles

in stage 1, most of these primary particles dissolve back

into solution to continuously provide Fe2+(aq) as iron is

incorporated into the solid phase g-FeOOH on a few of the

particles by reaction (9) in stage 2. Once they reach a critical

thickness, some of these g-FeOOH particles grow at a rate that

accelerates as the growth progresses. The net result is the slow

conversion of the primary particles to g-FeOOH nanoparticles.

Thus, under irradiation, the conversion of the primary

particles to the final particles occurs via radiolytically-induced

oxidation and reduction reactions, and does not appear to be

by aggregation of the primary particles.

The dendritic structure observed for the nanoparticles produced

after long irradiation times is consistent with solid-state growth of

the g-FeOOH crystals rather than homogeneous aggregation and

is consistent with the particle growth mechanism proposed

for stage 2. A similar formation of dendrites has been observed

in g-FeOOH deposited chemically as a thin film on glass25 and

g-FeOOH formed electrochemically on carbon steel.31

The net surface redox reaction rates eventually reach steady

state (via reactions 9 and 10), regulating the further growth of

the particles and this steady state explains the lack of growth

of particles beyond a certain size. Stage 3 is reached when

the aqueous ferric and ferrous ion concentrations, and the

g-FeOOH nanoparticle size, reach steady state.

5. Conclusions

Uniform-sized colloidal particles of g-FeOOH were formed by

gamma-irradiation of deaerated aqueous solutions containing

soluble Fe2+ species (as FeSO4) at pH 5.5. At short irradiation

times or in solutions containing initially low [Fe2+]0, spherical

particulates less than 10 nm in size with a narrow size

distribution were formed. These particles grew to form larger

particles with a dendritic structure upon longer irradiation,

but the growth terminated with particles having a very

Publ

ishe

d on

11

Mar

ch 2

011.

Dow

nloa

ded

by U

nive

rsity

of

Cal

ifor

nia

- Ir

vine

on

29/1

0/20

14 0

6:57

:00.

View Article Online

7206 Phys. Chem. Chem. Phys., 2011, 13, 7198–7206 This journal is c the Owner Societies 2011

uniform and narrow size distribution, approximately 60 nm in

diameter. Further prolonged irradiation did not change the

final particle size, but increased the number of the particles

having the final size. The final colloidal particles were identi-

fied as single-phase g-FeOOH.

The uniform size distribution obtained under g-irradiationis attributed to a multi-stage process initiated by rapid homo-

geneous radiolytic oxidation of soluble Fe2+ to relatively

insoluble Fe3+ hydroxides [Fe(H2O)6�n(OH)n]3�n which then

dehydrate to form amorphous ferric-hydroxide primary

particles. Upon longer irradiation, ferrous ion is oxidized by

water radiolysis products on the primary particle surface to

form g-FeOOH particles. The growth of g-FeOOH is acceler-

ated as the g-FeOOH layer thickens since g-FeOOH is more

conducting than ferric hydroxide. The ferrous ion required for

the growth of g-FeOOH is supplied by dissolution of other

primary particles via pseudo-equilibria between Fe2+(aq),

Fe3+(aq) and Fe(OH)3(s). The reduction of g-FeOOH

by water radiolysis products also occurs on the g-FeOOH

surfaces, and the rate increases with time (and particle surface

area). The g-FeOOH particles reach the final size when the

ferric reduction rate equals the ferrous oxidation rate on the

surface. Thus, under irradiation the conversion of primary

particles to final particles occurs via radiolytically-induced

oxidation and reduction reactions, and not by aggregation

of the primary particles. The radiolytic oxidation and reduc-

tion on the g-FeOOH particle surface is a function of the

steady-state concentrations of oxidizing and reducing water

radiolysis products. The final size reached at long times

is regulated by the steady-state charge transfer processes

at the aqueous-solid interface established during long-term

irradiation.

Acknowledgements

The work described herein was funded under the NSERC

(Natural Science and Engineering Research Council of Canada)

Discovery grant. The support from the Canada Foundation

for Innovation New Opportunity grant and the Ontario

Research Fund Excellence in Research: Nuclear Ontario grant

are greatly acknowledged for the purchase of the UV-Vis

absorption and FTIR spectrometers, respectively. The authors

would also like to recognize Dr Richard Gardiner at the

University of Western Ontario Biotron Facility and Dr Glynis

de Silveira and Mr Fred Pearson at the Canadian Centre for

Electron Microscopy, Brockhouse Institute for Materials

Research at McMaster University for their assistance with

TEM imaging.

References

1 K. Daub, X. Zhang, J. J. Noel and J. C. Wren, Electrochim. Acta,2010, 55, 2767–2776.

2 T. Yamamoto, S. Tsukui, S. Okamoto, T. Nagai, M. Takeuchi,S. Takeda and Y. Tanaka, J. Nucl. Mater., 1996, 228, 162–167.

3 N. Fujita, C. Matsuura and K. Saigo, Radiat. Phys. Chem., 2000,58, 139–147.

4 K. Daub, X. Zhang, J. J. Noel and J. C. Wren, Corros. Sci., 2011,53, 11–16.

5 A. Henglein and M. Geirsig, J. Phys. Chem. B, 1999, 103, 9533–9539.6 Y. Ni, X. Ge, Z. Zhang andQ. Ye,Chem.Mater., 2002, 14, 1048–1052.7 J. Belloni, M. Mostafavi, H. Remita, J. L. Marignier andM. O. Delcourt, New J. Chem., 1998, 22, 1239–1255.

8 S. Chen, Y. Liu and G. Wu, Nanotechnology, 2005, 16, 2360–2364.9 G. R. Dey, H. Remita and M. Mostafavi, Chem. Phys. Lett., 2006,431, 83–87.

10 P. Yadav, R. T. Olsson and M. Jonsson, Radiat. Phys. Chem.,2009, 78, 939–944.

11 E. B. Gracien, Z. Ruimin, X. LiHui, L. Kanza Kanza andI. Lopaka, J. Radioanal. Nucl. Chem., 2006, 270, 473–478.

12 A. O. Allen, The radiation chemistry of water and aqueous solutions,Van Nostrand, New York, 1961.

13 J. W. T. Spinks and R. J. Woods, An introduction to radiationchemistry, Wiley, New York, 1990.

14 Farhataziz and M. A. J. Rodgers, Radiation chemistry: principlesand applications, VCH Publishers, New York, 1987.

15 J. C. Wren, inNuclear energy and the environment, ed. M. W. Chienand B. J. Mincher, ACS Symposium Series, American ChemicalSociety, Washington, DC, 2010, pp. 271–295.

16 J. M. Joseph, B.-S. Choi, P. Yakabusie and J. C. Wren, Radiat.Phys. Chem., 2008, 77, 1009–1020.

17 P. A. Yakabuskie, J. Joseph and J. C. Wren, Radiat. Phys. Chem.,2010, 79, 777–785.

18 L. L. Stookey, Anal. Chem., 1970, 42, 779–781.19 E. Viollier, P. W. Inglett, K. Hunter, A. N. Roychoudhury and

P. Van Cappellen, Appl. Geochem., 2000, 15, 785–790.20 C. J. Hochanadel, J. Phys. Chem., 1952, 56, 587–594.21 I. Stefanic and J. A. LaVerne, J. Phys. Chem. A, 2002, 106, 447–452.22 J.-F. Boily, P. L. Gassman, T. Peretyazhko, J. Szanyi and

J. M. Zachara, Environ. Sci. Technol., 2010, 44, 1185–1190.23 G. V. Buxton, C. L. Greenstock, W. P. Helman and A. B. Ross,

J. Phys. Chem. Ref. Data, 1988, 17, 513.24 C. F. Baes and R. E. Mesmer, The hydrolysis of cations, Robert

E. Krieger Publishing Co., Malabar, Florida, 1986.25 D. Fu and J. C. Wren, J. Nucl. Mater., 2008, 374, 116–122.26 D. Fu, P. Keech, X. Sun and J. C. Wren, Iron oxyhydroxide nano-

particlesformed by forced hydrolysis: dependence of phase compositionon solution concentration, Phys. Chem. Chem. Phys., 2011, submitted.

27 L. S. Macary, M. L. Kahn, C. Estourne’s, P. Fau, D. Tremouilles,M. Bafleur, P. Renaud and B. Chaudret, Adv. Funct. Mater., 2009,19, 1775–1783.

28 J. Park, E. Lee, N. M. Hwang, M. Kang, S. C. Kim, Y. Hwang,J. G. Park, H. J. Noh, J. Y. Kim, J. H. Park and T. Hyeon, Angew.Chem., Int. Ed., 2005, 44, 2872–2877.

29 The Iron Oxides, Structure, Properties, Reactions, Occurrences andUses, ed. R. M. Cornell and U. Schwertmann, 2nd edn,Wiley-VCH, Weinheim, 2003.

30 D. Fu, X. Zhang, P. G. Keech, D. W. Shoesmith and J. C. Wren,Electrochim. Acta, 2010, 55, 3787–3796.

31 W. Xu, K. Daub, X. Zhang, J. J. Noel, D. W. Shoesmith andJ. C. Wren, Electrochim. Acta, 2009, 54, 5727–5738.

Publ

ishe

d on

11

Mar

ch 2

011.

Dow

nloa

ded

by U

nive

rsity

of

Cal

ifor

nia

- Ir

vine

on

29/1

0/20

14 0

6:57

:00.

View Article Online