behaviour of prussian blue during its interaction with ozone

Hyperfine Interactions 73(1992)285-294 285

B E H A V I O U R OF PRUSSIAN BLUE DURING ITS I N T E R A C T I O N W I T H O Z O N E

E. REGUERA, J. FERNANDEZ-BERTRAN, C. D[AZ and J. MOLERIO Centro Nacional de Investigaciones Cientificas, P.O. Box 6990, Havana, Cuba

Received 27 January 1992; accepted 3 April 1992

The interaction of ozone with Prussian Blue has been studied by M6ssbauer, infrared and XRD techniques. All spectral results reveal a reversible conversion of ferrocyanide to ferricyanide with preservation of the crystal skeleton. Ferric ferricyanide obtained by ozonization of ferric ferrocyanide has M6ssbauer parameters different to those of an ordered Prussian Brown, which has been attributed to secondary reaction products located at the zeolite voids of the former compound.

1. Introduction

In previous papers, we have reported the occurrence of mixed valence states when cobalt ferrocyanide is ozonized [1] or cobalt ferricyanide is reduced tribo- chemically [2]. We have now carried out a systematic study of the interactions of ozone with hexacyanometallates. In this paper, we present our results on the ozonization of ferric ferrocyanide (Prussian Blue) followed by Mc3ssbauer, infrared and XRD techniques. For comparison, a study of ferric ferricyanide (Prussian Brown, PBr) has been included.

As Prussian Blue (PB) are designed four different types of ferric ferro- cyanides [3]: Insoluble Prussian Blue (IPB), Fe4[Fe(CN)6]3. nH20; Soluble Prussian Blue (SPB), KFe[Fe(CN)6]. nH20; Tumbul l ' s Blue (TB), MFe3[Fe(CN)6]z.nHzO, where M is an appropriate anion; and that derived on aging of PBr (APBr), (H30)Fe[Fe(CN)6]. nH20. All these compounds crystallize in the cubic system and are members of a family of compounds named Prussian Blue analogous [4,5]. The structure of SPB and APBr can be described as a cubic network of F e n - c - N - F e 3+ with low-spin Fe H coordinated to six CN ligands at the C end and the high-spin Fe 3+ similarly coordinated to six CN ligands at the N end. The K + and H30 + cations are located in the Fe 3+ environment. In IPB and TB there is an exces of Fe 3+ over Fe H cations due to vacancies created by the omission of [Fe(CN)6] 4- anions in the cubic lattice, in order to maintain charge neutrality. The place of the missing N atoms as ligands of Fe 3+ cations is taken over by water molecules [6]. Additional water molecules have a zeolite character.

Different analytical methods can be used to monitor the ozone interaction with PB. MOssbauer spectroscopy is especially suited, since it can detect valence

�9 J.C. Balt.zer AG, Scientific Publishing Company

286 E. Reguera et al., Prussian Blue interaction with ozone

changes in both low-spin and high-spin iron atoms and gives valuable information on the environmental symmetry of the iron probes [7]. XRD informs us about possible changes introduced in the crystalline structure by the oxidation process. Infrared (IR) spectroscopy is particularly useful since VcN changes nearly 70 cm -1 on the oxidation of the central cation [8,9]. However, with only this last technique it is rather difficult to distinguish between different PB's.

2. Experimental

PB and PBr samples were prepared from dilute solutions of potassium ferrocyanide and ferricyanide, which are added to the appropriate iron salt: ferrous ammonium sulfate or ferric nitrate. The precipitates were filtered, washed several times with distilled water, dried in air up to 60 ~ and kept in a desiccator. Chemical analysis of compounds prepared by this method revealed that their potassium content was one percent (by weight) for IPB and TB and lower than this value for PBr. All samples were aged at room temperature.

Ozone was produced from dried oxygen in a glass ozonizer with a metal high- voltage electrode. The measurements on 03 decomposition in 100 mg PB samples were carried out by ultraviolet (UV) spectrophotometry at 254 nm using a flow reactor.

The IR spectra were recorded in an M80 Carl Zeiss spectrometer in Nujol mulls between CaF2 windows and with a maximum error of +2 cm -1. Some IR spectra were also taken in situ during ozonization of the samples. Freshly prepared PBr samples were studied directly as wet precipitates using the fortunate coincidence that the spectral region of interest (1900-2250 cm -1) is practically free of H/O band absorption.

XRD powder patterns were taken in an HZG-4 diffractometer (Carl Zeiss) using monochromatic CuKc~ radiation.

M0ssbauer spectra were recorded at room temperature with a 57Co in Rh source, using a constant acceleration spectrometer in the transmission mode. A special cell was built to obtain Mtissbauer spectra in situ under a controlled flow of an 03/02 mixture. Before recording the in situ MOssbauer spectra, the samples were kept for some hours in an ozone atmosphere. All MOssbauer spectra were fitted with an iterative least-squares minimization algorithm using Lorentzian line shapes. The errors (see table 2) are just the statistical ones resulting from the multi- parameter fit. Physical considerations lead us to estimate errors around three times higher in most cases.

3. Results and discussion

3.1. STUDY OF PRUSSIAN BROWN

Freshly synthesized PBr has three bands in the IR spectrum: 2172 cm -~, strong; 2136 cm -1, weak; and 2078 cm -l, medium (see fig. 1). The highest frequency

287

22OO 2OOO

"/ Ecrn-,]

Fig. 1. 1R spectrum recorded on wet precipitated Prussian Brown.

Table 1

Vc~ frequency values of wet synthesized Prussian Brown and of insoluble Prussian Blue before, in situ, and after the ozonization process.

Starting Spectrum recording Vcs Line Line compound condition *) [cm -1 ] intensity b) assignation c)

PBr After synthesis 2172 S FemFe 3+ (0.10 h) 2136 W (FenFe3+)(FemFe 2+)

2078 M FeIIFe 3.

PBr After synthesis 2172 S FemFe 3§ (100 h) 2078 M FenFe 3.

PBr After synthesis 2172 W FemFe 3§ (1440 h) 2078 S Fe nFe 3§

IPB Before 03 2078 S Fe"Fe 3§

IPB In situ 03 2172 S FemFe 3§ 2078 W FelIFe 3§

IPB After 03 2172 M FemFe 3§ (100 h) 2078 S FeIIFe 3§

IPB After 03 2078 S FelIFe 3§ (1440 h)

*)Aging time (after synthesis or 03 txeatment) in hours (h). b)s: sta'ong, M: medium, W: weak. C)FeIIFe3+: ferric ferrocyanide; FenZFe3+: ferric ferricyanide.


band corresponds to the ferricyanide, while the lowest one belongs to PB. The 2136 cm -1 band could be indicative of the presence of a mixed-valence intermediate, a ferro-ferricyanide [ 1 ]. PBr is not a stable product. After 100 hours, the IR spectrum shows a strong band at 2172 cm -1 and a medium strength band at 2078 cm -1. The colour changes gradually from brown to green and finally to blue. At this latter stage, we have a strong band corresponding to PB at 2078 cm -1 (see table 1).

On aging, wet synthesized PBr samples transform into PB, and at the same time a crystalline ordering process occurs. Diffractograms recorded on freshly synthesized samples are characteristic of an amorphous material, while those aged a few hours show a well-defined pattern. This ordered PBr and the corresponding aging products (APBr) practically have the same XRD powder patterns (see fig. 2).

ID

>:,

c

"E

J '

I0 50 50 De9, [ 2 r

Fig. 2. XRD powder patterns of (a) wet synthesized Prussian Brown, (b) aged Prussian Brown (1 year), (c) insoluble Prussian Blue, and (d) ozonized insoluble Prussian Blue.

M6ssbauer spectra give greater insight in the observed changes in PBr. On going from PBr to PB by aging, a mixture of ferric ferricyanide and ferric ferrocyanide

E. Reguera et al., Prussian Blue interaction with ozone 289

Table 2

M6ssbauer parameters at room temperature of wet synthesized Prussian Brown and of Prussian Blue before, in situ, and after its ozonization ~).

Starting Spectrum recording 6 b) A F Sub-spectrum compound conditions [mm/s] [mnVs] [mm/s] assignation c)

PBr Freshly 0.018(3) 0.76(1) 0.45(2) FeaULS synthesized 0.629(5) - 0.35(1) Fe3+- HS

PBr Aged 1 year 0.111(1) - 0.364(5) FeN-LS (APBr) 0.645(2) 0.339(5) 0.410(7) Fe3+- HS

IPB Before O 3 0.116(2) - 0.263(5) Fen-LS 0.639(2) 0.476(4) 0.43(1) Fe3+- HS

In situ O 3 0.018(3) 0.649(5) 0.698(9) FenI-LS 0.747(4) 0.404(5) 0.65(1) Fe3+- HS

100 h after 03 0.105(1) - 0.304(4) FetI-LS 0.667(2) 0.366(6) 0.444(7) Fe3+- HS 0.017(3) 0.590(7) 0.57(1) Fem-LS

1440 h after 03 0.118(2) - 0.321(9) FelI-LS 0.657(3) 0.421(9) 0.31(1) Fe3+- HS

TB Before O 3 0.111(2) - 0.280(5) FeU-LS 0.631(2) 0.533(4) 0.44(1) Fe3+- HS

In situ O 3 0.013(3) 0.436(5) 0.51(1) Fem-LS 0.693(4) 0.471(5) 0.54(1) Fe3+- HS

100 h after 03 0.108(2) - 0.312(6) FeU-LS 0.674(2) 0.420(7) 0.489(8) Fe3+- HS 0.027(3) 0.393(7) 0.47(1) Fem-LS

SPB Before O 3 0.129(1) - 0.305(5) Fen-LS 0.679(2) 0.266(8) 0.47(1) Fe3+-HS

In situ O 3 0.065(3) 0.349(7) 0.514(9) Fem-LS 0.756(3) 0.441(7) 0.536(9) FeS+-HS

100 h after O 3 0.122(2) - 0.278(6) Fea-LS 0.681(2) 0.326(5) 0.429(5) Fe3+- HS 0.088(3) 0.342(6) 0.43(1) Fem-LS

APBr In situ O 3 0.086(2) 0.348(5) 0.430(8) Fem-LS 0.628(2) - 0.491(7) Fe3+- HS

a)Fitting error in parentheses after the digit affected by the error. b)Isomer shift values relative to sodium nitroprusside. c)LS: low-spin configuration; HS: high-spin configuration.


is obtained, which has a characteristic deep green colour. This mixture is known as Prussian Green. Finally, when all the ferric ferricyanide has been reduced to ferric ferrocyanide, we have a PB. This spontaneous reduction must be associated with the oxidation of water molecules, allowing the reduction of low-spin Fem to Fe H and compensating the charge deficiency with hydronium ions (H3 O+) according to

1 2 { Fe[Fe(CN)6] } + 3H20 ---) 2 { (H30)Fe[Fe(CN)6] } + ~ 02 �9 (1)

Since low-spin Fe II cations conserve their highly symmetric environment (see table 2 and fig. 3(b)), hydronium ions must be in the neighbourhood of high-spin Fe 3§ explaining the quadrupole splitting observed for these last iron cations. This could be due to the reduction taking place first at the high-spin Fe 3+, which after being reduced to Fe 2+ suffers a fast electron transfer through the CN ligand, reducing the low-spin Fe HI to Fe H.

3.2. OZONIZATION OF PRUSSIAN BLUE

The 03 output as a function of contact time in the reactor shows that the oxidation of the PB sample is the dominant mechanism for ozone decomposition. When the sample has been totally oxidized, its colour has changed from deep blue to brown.

On ozonization, the IR spectra of PB with one intense absorption at 2078 cm -1 changes to that of PBr with a strong sharp band at 2172 cm -~, and a weak band at 2078 cm -~ of remnant PB. One hundred hours after ozonization, the intensities of the bands change, the 2172 cm -1 being of medium intensity and the 2078 cm -1 band being intense. After 1440 hours, the reaction has reverted to the original PB (see table 1) but without the formation of mixed valence states, as was observed in ozonized cobaltous ferrocyanide [1].

XRD powder patterns of ozonized PB recorded immediately after the ozonization process (fig. 2(d)) show diffraction peaks which match those corresponding to the PB starting sample (fig. 2(c)) and also to PBr (figs. 2(a), 2(b)). The ozonization does not affect the skeletal structure of the compound. The fact that the oxidation can take place readily in the solid phase indicates that the small ozone molecule can travel through channels in PB. In this sense, IPB and TB should be more easily ozonized than SPB and APBr, as we have observed experimentally.

M6ssbauer parameters of PB samples before, in situ, and after ozonization are reported in table 2. From this table and figs. 3, 4 and 5, one can observe that due to the ozone action, the typical singlet for low-spin Fe II changes to a quadrupole doublet with a smaller isomer shift ~5, showing that ferric ferrocyanide has transformed into ferric ferricyanide. The quadrupole splitting of this low-spin Fem ranges from 0.649 mm/s for IPB to 0.348 mm/s in APBr, indicating a lattice contribution to the observed A value. Since PB and PBr have the same crystalline structure, as shown by XRD powder patterns (see fig. 2), this additional contribution to the quadrupole


0"

o

o

2~

O"

4

;,-.:"""

�9 /7,,t f f',

:.,: .,. . . "-, ~ . . ...~

- 2 J i i

0 2 Velocity, Emm/s]

Fig. 3. M 6 s s b a u e r spectra of (a) freshly syn thes i zed wet Prussian Brown, (b) Prussian B r o w n aged for one year, and (c) in situ ozon iza t ion of Prussian B r o w n aged for one year.

splitting value must arise from secondary reaction products located in the neighbourhood of low-spin Fe IH cations.

In the oxidation of PB by ozonization, with the features discussed above, a third molecule must partake in the reaction. The likely candidates are CN- and H20; however, since the complex is not destroyed, this role could correspond to water molecules. In IPB, for instance, water molecules should participate in the oxidative process according to the following mechanism:

2[FeII(CN)6] 4- + 03 + H20 --e 2[Fem(CN)6] 3- + 2 O H - + 02. (2)

Excess 03 will react quickly with OH- to produce a series of complex reactions leading to charged species (O~, O~, HO~) and free radicals (OH', HO'3) [ 1 0 - 1 2 ] .


0 f

4

rm

E E -4

Q_

M) i 3 <f. �9

/ /

/ : " b

- 2 0 2 Velocity, [ m m / s ]

Fig. 4, M6ssbauer spectra of insoluble Prussian Blue (a) before, (b) in situ, and (c) aged-100 hours after its ozonization.

The large quadrupole splitting observed for low-spin F e III in ozonized IPB samples could be explained by the presence of these species close to iron cations. A similar mechanism might be present in the oxidation of TB.

In SPB there is a potassium cation per molecule, KFe[Fe(CN)6].nH20. This cation can participate in the total reaction through the following mechanism:

2 {KFe[Fe(CN)6]} + 03 + H20 -") Fe[Fe(CN)6] + 2K + + 2OH- + 02. (3)

As in the case of IPB, the OH- will be destroyed by ozone leading to charged species, which will increase A of the Fe 3+ cations. The low-spin Fe ul has an isotropic neighbourhood, with a A value similar to that measured in ozonized APBr (discussed below). On aging, this ferric ferricyanide returns to a ferric ferrocyanide but with M6ssbauer parameters similar to those observed in APBr (see table 2 and fig. 3).


_2 g

o

~ 4

/ II / ~

4

- 2 0 +2 Velocity, Emm/s]

Fig. 5. M6sbauer spectra of soluble Prussian Blue (a) before, (b) in situ, and (c) aged 100 hours after its ozonization.

When an APBr is ozonized again PBr is obtained. In this case, hydronium ions will participate in the conversion of ferric ferrocyanide into ferric ferricyanide according to

2[FetI(CN)6] 4- + 03 + 2H3 O+ --~ 2[Fcm(CN)6] 3- + 3H20 + 0 2. (4)

The ozonization process restores the symmetry of the environment, eliminating the H3 O§ charges without creating OH- anions. This explains the lower value of A for low-spin Fem and A = 0 for high-spin Fe 3+ in the ozonized product, in comparison with ozonized IPB (see fig. 3(c) and table 2). In table 2, one can also note that freshly synthesized PBr and ozonized APBr have different M6ssbauer parameters, which is attributed to the poor crystallinity of the former. According to eq. (4), the parameters of an ozonized APBr must be close to those expected for an ordered PBr.


4. Conclusions

The ozonization of ferric ferrocyanide produces the oxidation of this compound to a ferric ferricyanide, which returns to its initial state when the ozone action has ended. In this oxidation process the crystalline structure is conserved, indicating that the small ozone molecule can arrive at the reaction centers without breaking the skeletal structure of the compound.

M6ssbauer parameters of ozonized PB samples are closely related to the composition of the specific compound. In IPB, TB and SPB, water molecules should participate in the oxidation process producing OH- anions. Excess 03 will react quickly with OH- and produce a complex chain of reactions which include O~, HO~ and free radicals. From the presence of these species could arise the large quadrupole splitting values for low-spin Fem observed in IPB and TB in comparison with the ones corresponding to ozonized APBr. In APBr, the role of intermediate is played by hydronium ions.

Acknowledgement

The authors would like to thank C. Portilla, M. Echag~ie, C. Baluja and J. Duque for assistance in many of the measurements reported in the present paper. We also appreciate the help of L. Nufiez in the preparation of the manuscript.

References

[1] E. Reguera, J.F. Bertr~in, C. Dfaz, J. Blanco and S. Rond6n, Hyp, Int. 53(1990)391. [2] J.F. Bertr~n, J. Blanco, M. Hern~ndez and R. Rodn'guez, React. Solids 5(1988)95. [3] E. Reguera, J.F. Bertr~n, A. Dago and C. Diaz, Hyp. Int. 73(1992)295. [4] A. Ludi and H.U. Gildel, Structure and Bonding 14(1973)1. [5] P.G. Rasmussen and E.A. Meyers, Polyhedron 3(1984)183. [6] H.J. Buser, D. Schwarzenbach, W. Petter and A. Ludi, Inorg. Chem. 16(1977)2704. [7] P. Gt~tlich, M6ssbauer spectroscopy in chemistry, in: M6ssbauer Spectroscopy, ed. U. Gonser (Springer,

Berlin/Heidelberg, 1975), p. 53. [8] J.F. Bertr~n, J. Blanco and E. Reguera, Spectrochim. Acta 46A(1990)685. [9] J.F. Bertr~n, E. Reguera and J. Blanco, Spectrochim. Acta 46A(1990)1679. [10] J. Hoigne, in: Handbook of Ozone Technology and Applications, ed. R.G. Rice and A. Netzer (Arm

Arbor Science, 1982), p. 100. [11] J.L. Sotelo, F.J. Beltrfin, F.J. Benftez and J.B. Heredia, Ind. Eng. Chem. Res. 26(1987)39. [12] K. Chelkowska and D. Grasso, in: Proc. IX Ozone Worm Congress, Vol. 1, New York 1989, ed.

L.J. Bollyky (Port City Press Inc., New York, 1989), p. 106.

behaviour of prussian blue during its interaction with ozone

Documents