Morphology of colloidal metal pyrophosphate salts

Y.Mikal van Leeuwen,*a Krassimir P. Velikovb and Willem K. Kegel*a

Received 13th July 2011, Accepted 6th January 2012

DOI: 10.1039/c2ra00449f

We report the preparation and characterization of colloidal particles of several pyrophosphate metal

salts, including, for the first time, salts containing multiple metals. These materials are compared in

order to determine the influence of the composition and experimental conditions on particle

morphology. From this we make two observations: first, the metal ion valence determines the degree

of crystallinity of the material. While trivalent metal ions lead to amorphous precipitates, divalent

ones result in crystalline nanoparticles. Furthermore, the nature of the metal ion has a significant

influence on the morphology of the crystalline particles as we find clear shape preference for each

metal ion. Second, methods such as autoclave treatment and surfactant addition can be used to

further change particle morphology. Finally we mixed iron(III) with divalent metals during

preparation and find that this only has a significant effect on the resulting particles if the

FeIII : MII ratio is 1 : 5 or higher in MII content.

I Introduction

The pyrophosphate anion (diphosphate, P2O742 or PPi) is part of

the biological energy cycle and DNA synthesis, released upon

hydrolysis of adenosine triphosphate (ATP) to adenosine

monophosphate (AMP): ATP A AMP + PPi.1 Pyrophosphates

form insoluble complexes in water in combination with most

multivalent cations.2–4 Metal pyrophosphate salts are known for

their chemical and structural complexity5 and wide industrial

applications. They are used in catalysis,6–8 as solid electrolytes,9

as battery electrodes10 and in phosphate glasses for immobilisa-

tion of nuclear waste.11 Biomedical applications include artificial

bone transplant material12 and photoluminescent probes.13

Furthermore, a type of arthritis known as pseudogout is caused

by the deposition of small amounts of calcium pyrophosphate

(CaPPi) in articular tissue.14,15 Studying the formation of

colloidal systems of calcium pyrophosphate might be insightful

for the treatment of this disease. Finally, colloidal iron(III)

pyrophosphate (FePPi) has the intriguing property of containing

iron, yet being transparent/white. Because of this it is commer-

cially available as a food additive and mineral supplement: it is

an easily concealable material and useful for fighting iron

deficiency due to good bioaccessibility.16,17 Combining iron with

other micronutrients such as calcium, zinc and magnesium in a

single colloidal system would make it a multi-purpose, widely

applicable delivery system for micronutrients.18 While these are

many potential applications for colloidal systems not much work

has been published on them. In current literature the materials

are mainly prepared by solid-state synthesis under harsh

conditions8,12,13,19–23 although an intermediate ‘wet chemistry

method’ is sometimes used as well.5,10

Furthermore, control over morphology is important for many of

the biological and food applications since health risks can be induced

by particle size and shape, asbestos being a notorious example.

Here we present for the first time a systematic study into the

morphology and properties of colloidal metal pyrophosphate

salts depending on metal-ion valence and nature. Mixed-metal

systems are studied as well, by which we refer to a system of

pyrophosphates precipitated with a combination of Fe3+ and

either Al3+, Ca2+, Mg2+ or Zn2+. The divalent metal ions were

chosen considering the biocompatibility mentioned above,

aluminum was chosen for valence comparison to iron(III). We

show that the particle morphology is very diverse and depends

on metal ion as well as preparation method.

II Experimental section

Materials

FeCl3?6H2O, CaCl2?2H2O, AlCl3?6H2O, sodium dodecyl sulfate

(SDS), hexadecyltrimethylammonium bromide (CTAB), Igepal

CO-520 (NP-5) and disodium ethylenediaminetetraacetic acid

dihydrate (Na2EDTA) were all obtained from Sigma Aldrich.

Na4P2O7?10H2O and Zn(NO3)2?4H2O were bought from Merck

and MgCl2?6H2O from Fluka. All chemicals were used as

received; aqueous solutions were prepared using water deionized

by a Millipore Synergy water purification system.

Preparation

Colloidal particles of iron pyrophosphate were prepared by

dissolving 0.857 mmol FeCl3 in 50 ml water and adding this

aUtrecht University, Van’t Hoff Laboratory for Physical and ColloidalChemistry, Padualaan 8, 3584, CH Utrecht, the Netherlands.E-mail: Y.M.vanleeuwen@uu.nl; W.K.Kegel@uu.nl; Fax: +31-302533870;Tel: +31-302532873bUnilever Research and Development Vlaardingen, Olivier van Noortlaan120, 3133, AT Vlaardingen, the Netherlands

RSC Advances Dynamic Article Links

Cite this: RSC Advances, 2012, 2, 2534–2540

www.rsc.org/advances PAPER

2534 | RSC Adv., 2012, 2, 2534–2540 This journal is � The Royal Society of Chemistry 2012

Dow

nloa

ded

on 2

8 Fe

brua

ry 2

012

Publ

ishe

d on

07

Febr

uary

201

2 on

http

://pu

bs.r

sc.o

rg |

doi:1

0.10

39/C

2RA

0044

9FView Online / Journal Homepage / Table of Contents for this issue

solution dropwise in about 15 min to 0.643 mmol Na4P2O7 in

100 ml while stirring. A turbid white dispersion forms during the

addition of the final y5 ml. Mixed systems were prepared by

substituting part of the iron content by calcium, magnesium or

zinc (together referred to as MII) in the following Fe : MII ratios:

Fe10 MIIPPi8 (referred to as ‘10 : 1 ratio’), Fe8MII2PPi7 (‘4 : 1’),

Fe4MII4PPi5 (‘1 : 1’), Fe2 MII

11PPi7 (‘1 : 5’) or Fe2MII19PPi11

(‘1 : 10’). Here we assume complete precipitation without inclusion

of the Na+ and Cl2 from the reactants. Full substitution of iron

results in the pure MII pyrophosphate, MIIPPi. Substitution of

Fe3+ with Al3+ allows for direct stoichiometric exchange,

pyrophosphates precipitated with a trivalent metal ion will in

general be denoted MIIIPPi. For some experiments, the addition

was performed in an ultrasonic bath. These dispersions were

stirred for another 30 min while ultrasonicating after complete

addition. For autoclave experiments, after complete addition the

reaction mixture was transferred to a Teflon holder in a steel

autoclave cylinder and stored at 100 uC for four days without

stirring. All reaction mixtures were cleaned after preparation by

centrifugation and washing with millipore water twice.

The influence of surfactants was also investigated. For these

experiments surfactant (anionic SDS, cationic CTAB or non-

ionic NP-5) was used in a four times excess to the pyrophosphate

concentration; 2.6 mmol of the surfactant was dissolved in the

pyrophosphate solution before adding the metal ion solution.

This resulted in a concentration of 17 mM in the final volume:

beyond the critical micelle concentration (cmc) for all surfac-

tants. These samples were centrifuged five times instead of two to

remove excess surfactant.

Analysis

Dynamic light scattering (DLS) measurements were performed on a

Malvern Instruments Zetasizer Nano series machine in backscatter

mode at 25 uC with 5 min equilibration time. Samples were dried on a

carbon-coated copper grid prior to transmission electron microscopy

(TEM) and energy-dispersive X-ray spectroscopy (EDX) performed

on a Tecnai 12 and Tecnai 20 from FEI Company, respectively.

Samples for material analysis using X-ray powder diffraction (XRD)

and Fourier transform Infrared spectroscopy (FTIR) were prepared

by centrifuging the dispersion again, washing the sediment three

times with acetone, and drying in an oven at 40 uC for 24 h. FTIR

spectra were measured on a Perkin-Elmer IR spectrometer using the

KBr pellet technique, XRD measurements were carried out on a

Bruker-AXS D8 Advance powder X-ray diffractometer in Bragg–

Brentano mode equipped with automatic divergence slit (0.6 mm

0.3 u) and a PSD Vantec-1 detector. The radiation used was Co-

Ka1,2, l = 1.79026 A, operated at 30 kV, 45 mA. Elemental analysis

using inductively coupled plasma atomic emission spectroscopy

(ICP-AES) was performed on a solution prepared by dissolving the

dried samples using a four times excess of Na2EDTA compared to

the (estimated) metal ion content. Measurements performed using a

SPECTRO CIROSCCD ICP (8.13) with a radial plasma ion source

and Paschen–Runge spectrometer.

III Results

Metal pyrophosphate morphology

In a comparison between the influence of divalent and trivalent

metal ions on the resulting precipitates, we find that the

precipitation method yields similar particles for aluminum, iron

and calcium pyrophosphates. TEM analysis shows clusters

consisting of primary particles of roughly 20 nm, see Fig. 1a–c.

Magnesium and zinc show different morphology: magnesium

pyrophosphate forms large micrometer-sized needles and spheres

(Fig. 1d), while zinc results in thin platelets with either sharp or

rounded corners, see Fig. 1e. Since it is generally known from

colloidal synthesis that the preparation method can have a

drastic influence on the resulting particles, we tested three

Fig. 1 TEM image matrix summarizing the influence of preparation method and metal ion on colloidal MPPis. Rows indicate the preparation

method, columns the metal ion used. The red circles indicate what we refer to as primary particles in the amorphous aggregates (i is an SEM image).

This journal is � The Royal Society of Chemistry 2012 RSC Adv., 2012, 2, 2534–2540 | 2535

Dow

nloa

ded

on 2

8 Fe

brua

ry 2

012

Publ

ishe

d on

07

Febr

uary

201

2 on

http

://pu

bs.r

sc.o

rg |

doi:1

0.10

39/C

2RA

0044

9F

View Online

frequently used methods for their influence on metal pyropho-

sphates.

The first method is autoclave treatment, used in literature for

changing gel-like materials into crystalline particles.24–26 This

treatment has a strong influence on the metal-pyrophosphate

morphology. For the trivalent materials the primary particles

have become hollow, see Fig. 1f and g. We were unable to take

high resolution images of these hollow particles during TEM

analysis since these seem to be much more beam sensitive than

the untreated particles shown in Fig. 1a and b. Even at low-dose

illumination the particles shrink within seconds and the hollow

centre is no longer visible. CaPPi grows into rods resembling

those found from similar treatment of the chemically related

hydroxyapatite,27 Fig. 1h. The size of these rods can be slightly

influenced by the autoclave temperature and treatment time

(results not shown). MgPPi forms micrometer sized, ellipsoidal

platelets and autoclave treatment of the ZnPPi particles results in

thicker, rounded platelets, see Fig. 1i and j.

In the second method, ultrasonication of the dispersions

during and/or after preparation does not have a significant

influence on particle shape on any material except for CaPPi.

This system changes into micrometer sized, crystalline platelets,

see Fig. 1k.

The third method makes use of surfactant molecules during

preparation, which can have strong effects on particle morphol-

ogy.28,29 Three types of surfactant were used: anionic (SDS),

cationic (CTAB) and nonionic (NP-5). We find that while

surfactants have no discernable influence on FePPi and AlPPi,

they do change the morphology of the MIIPPis. The results are

summarized in the image-matrix in Fig. 2. Two samples are

missing from the matrix: Ca2+ forms a precipitate with SDS and

therefore CaPPi could not be prepared with SDS. This effect is

not observed for any of the other metal ions used in this work.

Using CTAB in the reaction mixture of MgPPi results in a clear

solution: no particles are formed.

For SDS, magnesium forms very thin, elongated platelets

(Fig. 2a) while zinc only forms the geometric shapes found in

Fig. 1e (Fig. 2b). CaPPi in the presence of CTAB forms large

spherical clusters of platelets and ZnPPi results in y3 mm

ellipsoidal platelets consisting of multiple layers as shown in

Fig. 2c and d. Finally, CaPPi grown in the presence of NP-5

shows morphology similar to CTAB treatment but consisting of

Fig. 2 TEM image matrix summarizing the influence of surfactants on MIIPPis. Rows indicate the surfactant, columns the metal ion used (images c, e

and the inset of f are SEM images).

2536 | RSC Adv., 2012, 2, 2534–2540 This journal is � The Royal Society of Chemistry 2012

Dow

nloa

ded

on 2

8 Fe

brua

ry 2

012

Publ

ishe

d on

07

Febr

uary

201

2 on

http

://pu

bs.r

sc.o

rg |

doi:1

0.10

39/C

2RA

0044

9F

View Online

thinner platelets with frayed edges as shown in Fig. 2e. MgPPi

shows particles smaller but similar to those formed with regular

synthesis, ZnPPi yields platelets with truncated corners when

prepared with NP-5 (see Fig. 2f and g).

Comparing these results we find that MIIIPPis form amor-

phous precipitates while MIIPPis form different, possibly crystal-

line, shapes. MgPPi preferably forms elongated, needle-like

particles. The spheres shown in Fig. 1d seem to be an exception

to this, but they actually consist of platelets and needles clustered

together, just like the particles shown in the inset of Fig. 2f.

ZnPPi has a clear preference for thin platelets and CaPPi

morphology differs per preparation method. In general, the

MIIPPis have a preference for thin platelets or needles.

DLS measurements of FeIIIPPi yield volume-average sizes on

the order of 200 nm, which is larger than the particles found

from TEM (y20 nm). This indicates that the precipitates formed

are present as clusters in dispersion, possibly resembling what is

found from TEM analysis (Fig. 1a and b). It should be noted

here that aggregates observed from TEM are not necessarily

representative for the situation in dispersion, as TEM samples

are dried prior to analysis. DLS analysis of AlPPi and the

MIIPPis was not successful due to colloidal instability. The fact

that DLS analysis on MgPPi and ZnPPi would not be possible

could have been expected from TEM analysis: micrometer sized

(aggregated) particles of these materials sediment too rapidly.

But while AlPPi and CaPPi look morphologically similar to

FePPi in TEM images (Fig. 1a–c), sedimentation was too rapid

for these samples as well. From this it might be concluded that

much larger aggregates are formed in AlPPi and CaPPi

compared to FePPi. This is confirmed by the macroscopic

observation of the AlPPi and CaPPi dispersions sedimenting

significantly faster than the FePPi dispersion (see Fig. 3). This

could not be influenced by ultrasonication during or after

preparation.

XRD analysis confirms the difference between MIIPPi and

MIIIPPi found from TEM analysis in Figs. 1 and 2. While the

XRD spectra for both MIIIPPis show only noise indicating

amorphous materials, the spectra for MIIPPis show clear signals,

see Fig. 4a. However, the high signal-to-noise ratio indicates that

part of the MIIPPi particles is amorphous as well. The influence

of the preparation method is also evident in XRD analysis.

For example, autoclave treatment of calcium pyrophosphate

increases crystallinity and changes its crystal structure as can be

seen from the reduction in background noise and shift in peak

location in Fig. 4b, comparing the spectra of CaPPi before

(yellow, bottom) and after autoclave treatment (orange, middle).

Similar effects have been shown in literature for solid-state

preparation methods at much higher temperatures.20,22,30

Ultrasonication also changes the XRD spectrum of CaPPi, see

the red top spectrum in Fig. 4b. Magnesium shows similar

changes and the presence of SDS slightly influences the obtained

pattern but not as drastically as the change in preparation

Fig. 4 XRD analysis of the metal pyrophosphates. Comparison

between MIIPPi and MIIIPPi in (a) shows the MIIIPPis to be amorphous

while the MIIPPis are (partly) crystalline. Using different preparation

methods such as addition of surfactant, ultrasonication or autoclave

treatment results in different crystal structures for CaPPi (b) and MgPPi

(c).

Fig. 3 Sedimentation in three metal pyrophosphate dispersions over

time: freshly prepared (a), after 24 h (b) and after three days (c). From

left to right: FePPi, CaPPi and AlPPi. After three days the gel-like state

formed in (b) has collapsed completely into dense aggregates for CaPPi.

This journal is � The Royal Society of Chemistry 2012 RSC Adv., 2012, 2, 2534–2540 | 2537

Dow

nloa

ded

on 2

8 Fe

brua

ry 2

012

Publ

ishe

d on

07

Febr

uary

201

2 on

http

://pu

bs.r

sc.o

rg |

doi:1

0.10

39/C

2RA

0044

9F

View Online

method, see Fig. 4c. Comparison of these XRD spectra with

literature yields little to no agreement. This could be explained

by the fact that, as mentioned before, most of those character-

izations are performed on macroscopic crystals obtained from

solid-state preparation often at high temperatures, likely to

induce different crystal structures.5,21–23,31–33

FTIR analysis shows that peak positions coincide between

materials due to the fact that the main vibrationally active

species is the same in all materials (PPi), see Fig. 5. The one

exception to this is AlPPi, for which the spectrum is slightly

shifted: e.g. the maximum at 1120 cm21 shifts to 1170 cm21. This

could indicate a stronger influence of Al3+ on the PLO vibration

compared to the other metal ions, but why a similar effect is not

observed for Fe3+ is not clear.

MIIIPPis show broad, smooth peaks in FTIR analysis, while

MIIPPis result in sharper peaks. This is probably due to the

partly amorphous nature of the MIIIPPis found from XRD

analysis in Fig. 4. This lower degree of crystallinity also explains

the poor agreement in detail with FTIR analysis of crystalline

(bulk) materials reported in literature, such as LiFePPi10 or

CaPPi.21,34,35 More amorphous materials such as CoFePPi give

better agreement in the smoothness of the curves.19

Mixed-metal pyrophosphates

For systems containing mixed metals, charge neutral ratios of

FeIII : MII : PPi are prepared, see the Methods section for details.

Most resulting precipitates are similar to the amorphous clusters

observed in Fig. 1a–c, see Fig. 6a and b. The samples deviating

from this trend are Fe : Mg and Fe : Zn in ratios of 1 : 5 and

higher in MII content. For FeMgPPi needle-like particles are

formed resembling those seen before in Figs. 1 and 2 although

some demixing seems to occur: smaller y20 nm particles can be

seen on the surface of the needles, see Fig. 6c. In the case of

FeZnPPi rough, irregular platelets together with elliptical shapes

are found, see Fig. 6d. XRD and FTIR analysis of the Fe : Ca

10 : 1 material shows it to be similar to pure FePPi: amorphous

and with little detail in FTIR (results not shown).

From ICP-AES elemental analysis we find that the particle

composition is very similar to what is added during reaction,

except for the Na+ and Cl2 which are predominantly left in solution.

For pure FePPi particles this results in Na0.2Cl0.3Fe4(P2O7)2.9,

compared to the Na12Cl12Fe4(P2O7)3 added. Most likely the Na and

Cl indicated in the formula of the precipitate are largely present as

counterions in solution not washed away during cleaning of the

reaction mixture, not so much incorporated into the material.

EDX line scans indicate homogeneous composition of the

nanoparticles (on the resolution of the analysis method); as can

be seen the normalized intensities follow each other closely, see

Fig. 7.

Discussion

We have shown that the degree of crystallinity of the colloidal

particles is determined by the valence of the metal ion used as we

find the MIIIPPis to be amorphous and the MIIPPis to be mostly

crystalline. It can be argued that this is caused by the ‘valence

mismatch’ in MIIIPPis. While MIIPPi can easily form a charge-

neutral MII2PPi unit for (bulk) crystalline material, the combina-

tion of Me3+ and PPi42 needs a more complicated stoichiometry

to reach charge neutrality. During fast precipitation this could

result in an amorphous structure. It is therefore slightly

surprising to find so little Na+ or Cl2 in the chemical analysis

since either could facilitate smaller charge-neutral subunits.

Similar complicating effects could cause CaPPi to initially form

amorphous material if the Ca–PPi bond is so strong that

amorphous precipitation is faster than crystallization. It might

also explain why the mixed-metal pyrophosphates remain

amorphous up to a FIII : MII ratio of 1 : 5, simply because the

presence of FeIII disrupts the crystal formation.

It is interesting to note that the chemically related metal

phosphate salts seem to follow the opposite trend in valence

dependent crystallinity, as colloidal aluminum phosphate is

crystalline (vascerite)36 while phosphates containing divalent

metals can be prepared as amorphous particles even after

autoclave treatment.37,38 This might be caused by the same effect

Fig. 5 FTIR analysis of the metal pyrophosphates prepared by

coprecipitation. At wavenumbers beyond 2000 cm21 only a broad –OH

peak is found (not shown).Fig. 6 Morphology of mixed metal pyrophosphates resembles that of

MIIIPPi found in Fig. 1, as shown here for Fe : Al 10 : 1 (a) and Fe : Ca

1 : 10 (b). The only mixed materials that show different morphology are

Fe : Mg 1 : 10 (c) and Fe : Zn 1 : 5 (d). Ratios with higher MIIcontent

show similar morphology (not shown).

2538 | RSC Adv., 2012, 2, 2534–2540 This journal is � The Royal Society of Chemistry 2012

Dow

nloa

ded

on 2

8 Fe

brua

ry 2

012

Publ

ishe

d on

07

Febr

uary

201

2 on

http

://pu

bs.r

sc.o

rg |

doi:1

0.10

39/C

2RA

0044

9F

View Online

mentioned above: since the phosphate ion is trivalent, the

divalent metals would have the ‘valence mismatch’ and the

trivalent metals can immediately form a charge-neutral complex.

That iron phosphate forms micrometer-sized, amorphous

particles could be caused by a similar effect as mentioned for

calcium pyrophosphate.39

IV Conclusions

In this work we have studied the morphology of (mixed) metal

pyrophosphates. First, we have shown the influence of the metal

ion valence on the morphology of metal pyrophosphates; while

the trivalent metal ions yield amorphous precipitates, the

divalent ones result in crystalline particles. Furthermore, the

type of divalent metal ion determines the shape of the particles

formed as we find different particles for the different metal ions.

Second, we have shown that the morphology of these materials

can be further modified by changes in experimental conditions

by using methods such as autoclave treatment or addition of

surfactants. In general, MIIPPis have a preference for thin, flat

platelets or needles differing in size or organization per

preparation method, while the MIIIPPis form the same amor-

phous nanoparticles for every preparation method. The only

exception is autoclave treatment: amorphous metal pyropho-

sphates aged in an autoclave slightly increase in size and become

hollow. They do not form crystalline particles as one might

expect from other work.24–26 Analysis using XRD shows that

these changes in shape coincide with different crystal structures,

but no exact match is found with materials in literature. Finally,

we have investigated the effects of mixing divalent and trivalent

metal ions in a single precipitate. From this, we find the resulting

materials to mainly resemble the amorphous MIIIPPis unless the

FeIII : MII ratio is larger than 1 : 5 at which point larger particles

tend to form. From EDX analysis we find the components to be

homogeneously distributed over the precipitates while ICP-AES

analysis indicates overall particle composition to be consistent

with what was added during preparation. The difference in

crystallinity between MIIPPi and MIIIPPi could be caused by the

greater complexity needed for forming an uncharged unit of

MIIIPPi compared to MIIPPi, thereby preferring an amorphous

structure upon fast precipitation.

Acknowledgements

We thank the following individuals for characterization and

analysis: F. Soulimani and M. Versluijs-Helder of the Inorganic

Chemistry and Catalysis group at the University of Utrecht for

FTIR and XRD measurements, respectively. J. D. Meeldijk of

the Electron Microscopy group in Utrecht is thanked for SEM

and EDX analysis and H. C. de Waard of the Geolab Analytical

services in Utrecht for ICP-AES measurements. We thank Senter

Novem for financial support; this work is financially supported

by Food Nutrition Delta program grant FND07002.

References

1 B. Alberts, A. Johnson, J. Lewis, M. Raff, K. Roberts and P. Walter,Molecular Biology of the Cell, Garland Science, 5th edn, 2007.

2 E. Hogfeldt, Stability Constants of Metal-Ion Complexes (SpecialPublication), 1964.

3 L. G. Sillen and A. E. Martell, Stability Constants of Metal-IonComplexes (Special Publication), 1970.

4 L. G. Sillen and A. E. Martell, Stability Constants of Metal-IonComplexes (Special Publication), 1964.

5 J. Belkouch, L. Monceaux, E. Bordes and P. Courtine, Mater. Res.Bull., 1995, 30, 149–160.

6 J. Belkouch, B. Taouk, L. Monceaux, E. Bordes, P. Courtine and G.Hecquet, Stud. Surf. Sci. Catal., 1994, 82, 819–828.

7 J.-M. Millet and J. C. Vedrine, Appl. Catal., 1991, 76, 209–219.8 G. J. Hutchings, Appl. Catal., 1991, 72, 1–32.9 N. Y. Morozova and N. Y. Selivanova, Russ. J. Inorg. Chem., 1976,

21, 878–879.10 A. Ait Salah, P. Jozwiak, K. Zaghib, J. Garbarczyk, F. Gendron, A.

Mauger and C. M. Julien, Spectrochim. Acta, Part A, 2006, 65,1007–1013.

11 C. S. Ray, X. Fang, M. Karabulut, G. K. Marasinghe and D. E. Day,J. Non-Cryst. Solids, 1999, 249, 1–16.

12 U. Gbureck, T. Holzel, I. Biermann, J. E. Barralet and L. M. Grover,J. Mater. Sci.: Mater. Med., 2008, 19, 1559–1563.

13 A. Doat, F. Pellee and A. Lebugle, J. Solid State Chem., 2005, 178,2354–2362.

14 P. J. Groves, R. M. Wilson, P. A. Dieppe and R. P. Shellis, J. Mater.Sci.: Mater. Med., 2007, 18, 1355–1360.

15 G. S. Mandel, P. B. Halverson and N. S. Mandel, Scanning Microsc.,1988, 2, 1177–1188.

16 R. Wegmuller, M. B. Zimmermann, D. Moretti, M. Arnold, W.Langhans and R. F. Hurrell, J. Nutrition, 2004, 134, 3301–3304.

17 M. C. Fidler, T. Walczyk, L. Davidsson, C. Zeder, N. Sakaguchi,L. R. Juneja and R. F. Hurrell, Br. J. Nutr., 2007, 91, 107–112.

18 F. M. Hilty, M. Arnold, M. Hilbe, A. Teleki, J. T. N. Knijnenburg,F. Ehrensperger, R. F. Hurrell, S. E. Pratsinis, W. Langhans andM. B. Zimmermann, Nat. Nanotechnol., 2010, 5, 374–380.

19 B. Boonchom and N. Phuvongpha, Mater. Lett., 2009, 63,1709–1711.

20 J. Bian, D. Kim and K. Hong, Mater. Lett., 2004, 58, 347–351.21 A. M. El Kady, K. R. Mohamed and G. T. El-Bassyouni, Ceram.

Int., 2009, 35, 2933–2942.22 Z. W. Xiao, G. R. Hu, Z. D. Peng, K. Du and X. G. Gao, Chin.

Chem. Lett., 2007, 18, 1525–1527.23 R. Pang, C. Li, S. Zhang and Q. Su, Mater. Chem. Phys., 2009, 113,

215–218.

Fig. 7 EDX analysis of Fe : Ca 1 : 10 ratio. (a) dark-field TEM image

showing the position of the scan line (red line, scan direction from 1 to 2).

The normalized intensities (b) show a uniform material distribution

throughout the material. Other materials give similar results (not shown).

This journal is � The Royal Society of Chemistry 2012 RSC Adv., 2012, 2, 2534–2540 | 2539

Dow

nloa

ded

on 2

8 Fe

brua

ry 2

012

Publ

ishe

d on

07

Febr

uary

201

2 on

http

://pu

bs.r

sc.o

rg |

doi:1

0.10

39/C

2RA

0044

9F

View Online

24 S. Shen, P. S. Chow, F. Chen, S. Feng and R. B. H. Tan, J. Cryst.Growth, 2006, 292, 136–142.

25 L. Hickey, J. T. Kloprogge and R. L. Frost, J. Mater. Sci., 2000, 35,4347–4355.

26 G. Lagaly, O. Mecking and D. Penner, Colloid Polym. Sci., 2001, 279,1097–1103.

27 R. E. Riman, W. L. Suchanek, K. Byrappa, C-W. Chen, P. Shuk andC. S. Oakes, Solid State Ionics, 2002, 151, 393–402.

28 B. Nikoobakht and M. A. El-Sayed, Chem. Mater., 2003, 15,1957–1962.

29 L. Manna, E. C. Scher and A. P. Alivisatos, J. Am. Chem. Soc., 2000,122, 12700–12706.

30 M. Ciopec, C. Muntean, A. Negrea, L. Lupa, P. Negrea and P.Barvinschi, Thermochim. Acta, 2009, 488, 10–16.

31 L. K. Elbouaanani, B. Malaman, R. Gerardin and M. Ijjaali, J. SolidState Chem., 2002, 163, 412–420.

32 J. Bian, D. Kim and K. Hong, Mater. Lett., 2005, 59, 257–260.33 B. Boonchom and R. Baitahe, Mater. Lett., 2009, 63, 2218–2220.34 P. J. Linstrom and W. G. Mallard, NIST Chemistry WebBook, NIST

Standard Reference Database Number, p. 69.35 B. B. Parekh and M. J. Joshi, Cryst. Res. Technol., 2007, 42, 127–132.36 J. Gomez Morales, R. Rodrıguez Clemente and E. Matijevic, J.

Colloid Interface Sci., 1992, 151, 555–562.37 L. L. Springsteen and E. Matijevic, Colloid Polym. Sci., 1989, 267,

1007–1015.38 T. Ishikawa and E. Matijevic, J. Colloid Interface Sci., 1988, 123,

122–128.39 R. B. Wilhelmy and E. Matijevic, Colloids Surf., 1987, 22, 111–131.

2540 | RSC Adv., 2012, 2, 2534–2540 This journal is � The Royal Society of Chemistry 2012

Dow

nloa

ded

on 2

8 Fe

brua

ry 2

012

Publ

ishe

d on

07

Febr

uary

201

2 on

http

://pu

bs.r

sc.o

rg |

doi:1

0.10

39/C

2RA

0044

9F

View Online