Synthesis, vibrational spectra and thermal stability of Ag3O4 and related Ag7O8X salts (X = NO3-, ClO4-, HSO4-)

Polyhedron 26 (2007) 3310–3322

www.elsevier.com/locate/poly

Synthesis, vibrational spectra and thermal stability of Ag₃O₄and related

ꢀ

Ag₇O₈X salts ðX ¼ NO₃; ClO₄; HSO₄Þ

*

Geoﬀrey I.N. Waterhouse , James B. Metson, Graham A. Bowmaker

Department of Chemistry, University of Auckland, Private Bag 90219, Auckland, New Zealand

Received 12 February 2007; accepted 6 March 2007

Available online 12 March 2007

Abstract

The titled compounds were synthesized in high purity and yield by passing an ozone/oxygen mixture (5 vol% O₃in O₂) through aque-

ous solutions of diﬀerent Ag(I) salts at 273–298 K. Ag₃O₄is obtained by ozonolysis of neutral or mildly acidic solutions of AgF (0.05 M)

or AgBF₄(0.05 M) at 273 K. Ozonolysis of aqueous solutions of AgNO₃, AgClO₄, Ag₂SO₄and CH₃COOAg at 273–298 K aﬀorded

Ag₇O₈NO₃, Ag₇O₈ClO₄, Ag₇O₈HSO₄and Ag₂O₂, respectively. Products were characterized by XRD, SEM, XPS, FT-IR, Raman spec-

troscopy and DSC. The spectroscopic data show that Ag₃O₄and the Ag₇O₈X salts behave as metals at room temperature, due to overlap

between valence and conduction bands in their electronic structure. The correct valence formulations for these compounds are

(Ag^2.67+)₃O₄and (Ag^2.67+)₆O₈Æ Ag⁺X^ꢀ, respectively. The thermal stability in air of the ozonolysis products was determined by DSC

and also complementary XRD, FT-IR and Raman measurements. Ag₃O₄was the least stable product, and thermally decomposed to

1.5Ag₂O_2(s)+ 0.5O_2(g)within 20 h at 298 K. Thermal stabilities in air increased in the order Ag₃O₄< Ag₇O₈ClO₄, Ag₇O₈HSO₄< Ag₇O₈-

NO₃< Ag₂O₂. This work resolves long-standing debate about the valence states of silver in Ag₃O₄and Ag₇O₈X salts, and also provides

valuable new spectroscopic information for these compounds. Attempts to prepare Ag₂O₃by the ozonolysis route proved unsuccessful.

Keywords: Silver oxides; Ozone; Ag₃O₄; Ag₇O₈NO₃; XPS; FT-IR; Raman

1. Introduction

gested [13,14], although supporting evidence for these

phases is lacking.

The oxides of silver are a structurally diverse group of

compounds that attract attention due to the widespread

use of Ag₂O and Ag₂O₂as cathode materials in Zn–silver

oxide ‘‘button cell’’ batteries [1–3]. To date, ﬁve binary oxi-

des have been isolated and structurally characterized in the

silver–oxygen system [4–12]; Ag₆O₂, Ag₂O, Ag₂O₂, Ag₃O₄,

and Ag₂O₃. It should be noted that Ag₂O₂and Ag₃O₄are

both mixed-valence compounds (formally written as

Ag⁺Ag³⁺O₂and Ag²⁺(Ag³⁺)₂O₄, respectively), and that

Ag₂O₂crystallizes in two diﬀerent polymorphic forms

[6–11]. The existence of other binary oxide phases, such

as Ag₄O₃and a hexagonal form of Ag₂O have been sug-

Due to their technological importance, Ag₂O and

Ag₂O₂have been the subjects of numerous structural

and spectroscopic investigations [5–9,14–23]. Accordingly,

literature pertaining to these two compounds is extensive

and generally reliable. By comparison, data for Ag₃O₄

and Ag₂O₃is extremely limited [10–12,22,23], which stems

largely from the thermal instability of these two com-

pounds. Thermal stability decreases rapidly with increas-

ing silver oxidation state (cf. Ag₂O, Ag₂O₂and Ag₂O₃;

D_fG⁰= ꢀ11.2, 27.6 and 121.4 kJ mol^ꢀ1

, respectively)

[24]. Ag₂O₃decomposes completely to Ag₂O₂in less than

1 h at 298 K, whereas Ag₂O is stable indeﬁnitely at this

temperature. Crystalline specimens of Ag₃O₄and Ag₂O₃

have been obtained by electrolytic deposition from aque-

ous solutions of diﬀerent Ag(I) salts at 273 K [10–12], and

their crystal structures solved. Ag₂O₃crystallizes in the

*

Corresponding author. Tel.: +64

9 3737599x85876; fax: +64 9

3737422.

E-mail address: g.waterhouse@auckland.ac.nz (G.I.N. Waterhouse).

doi:10.1016/j.poly.2007.03.006

G.I.N. Waterhouse et al. / Polyhedron 26 (2007) 3310–3322

3311

orthorhombic system Fdd2 with Z = 8, and is isomor-

phous with Au₂O₃[12]. The structure contains approxi-

mately square-planar AgO₄groups linked together via

common oxygen atoms in a three-dimensional network,

nated 8-fold by oxygen at an average Ag–O distance of

2.52 A. Anions occupy sites at the centre of the cubo-octa-

˚

hedra. Fundamental interest in these compounds relates to

their fascinating physical properties which include metallic

conductivity (q = 2.5 · 10^ꢀ3X cm for Ag₇O₈NO₃at 300 K)

and low temperature superconductivity (T_c= 1.04 K for

Ag₇O₈NO₃) [27,28]. The metallic properties of Ag₇O₈X

salts are attributed to promotion of two electrons from

the face-sharing ‘Ag₆O₈’ structural units into a conduction

band that utilizes Ag 4d and O 2p orbitals. The Ag₇O₈X

salts are two electron class III-B mixed-valence compounds

under the classiﬁcation system of Robin and Day [30], and

as such are expected to show an absorption edge (typically

in the infrared) and metallic reﬂectivity in the visible

region. Accordingly, they should be very interesting mate-

rials to study by FT-IR and Raman spectroscopy. To date,

no reliable vibrational spectroscopic data have been

reported for these salts, and minimal experimental data

exists examining their electronic structure (e.g. XPS, UPS

or XANES) [35]. Further spectroscopic studies of the

Ag₇O₈X compounds are needed to ﬁll these gaps.

Traditionally, Ag₃O₄, Ag₂O₃and Ag₇O₈X salts have

been synthesized by electrochemical routes (e.g. by anodic

polarization of a platinum electrode). The disadvantage of

the electrochemical approach is the fact that crystal growth

rates are slow and yields typically low (scale-up procedures

are not practical due to the high cost of Pt electrodes).

Chemical routes, rather than electrochemical, are thus

more attractive for the large-scale synthesis of these com-

pounds. O₃is one of the strongest oxidants known

ðO₂=O₃e^acid¼ 2:075 VÞ and capable of oxidising Ag⁺to

Ag²⁺ð⁰Ag^þ=Ag^2þe₀^acid¼ 1:980 VÞ and Ag⁺to Ag³⁺

ðAg^þ=AgO^þe₀^acid¼ 2:00 VÞ [24]. The application of ozonol-

ysis to synthesize Ag₂O₂from metallic Ag or Ag₂O has

already been demonstrated [9,19,20,37,38]. The synthesis

of Ag₂O₂, Ag₃O₄, Ag₂O₃and Ag₇O₈X salts by ozone-

oxidation of aqueous Ag⁺is also theoretically viable

(Eqs. (1)–(4) below). Minimal reliable work has been

reported in this area to date [39], clearly justifying further

research eﬀort.

˚

with an average Ag–O bond length of 2.02 A. Ag₃O₄

crystallizes in the monoclinic system P2₁/c with Z = 2

[10,11], and has the formal valence electron conﬁguration

Ag²⁺(Ag³⁺)₂O_4.The crystal structure comprises a three-

dimensional network of distorted edge- and corner-

sharing square-planar AgO₄units. Silver cations occupy

two crystallographically independent sites; Ag³⁺cations

are square-coordinated by four oxygens at an average dis-

2+

˚

tance of 2.07 A, while Ag

cations have four nearest

˚

neighbours at 2.03 A. This result is somewhat anomalous,

since shorter Ag–O bond lengths are expected about the

Ag³⁺cations. Recently, van Wullen et al. reported

¨

¹⁰⁹Ag solid state magic angle spinning (MAS) NMR spec-

tra for a number of binary and ternary silver oxides [23].

The ¹⁰⁹Ag MAS NMR spectrum for Ag₃O₄showed only

one broad isotropic signal at ꢀ795 ppm, which was tenta-

tively assigned to Ag³⁺centres (cf. Ag³⁺centres in Ag₂O₂

and Ag₂O₃showed chemical shifts of 1970 and 2015 ppm,

respectively). No signal was observed for the Ag²⁺cations

in Ag₃O₄. Results of the X-ray crystallographic and ¹⁰⁹Ag

MAS NMR investigations can be rationalized if an alter-

native valence conﬁguration of (Ag^2.67+)₃O₄is considered

for the compound. O’Keefe calculated Madelung poten-

tials at Ag and O sites in a number of binary silver oxides

[25]. Madelung potentials are approximately proportional

to valence and may be used to determine valences in

mixed-valence oxides. In the case of Ag₃O₄, no preferred

valence conﬁguration was found between Ag²⁺(Ag³⁺)₂O₄

and (Ag^2.67+)₃O₄, evidence of electron delocalization and

metallic behaviour. Further work is necessary to clarify

the electronic structure of Ag₃O₄. Also, important spec-

troscopic data for Ag₂O₃and Ag₃O₄has yet to be

reported (such as FT-IR and Raman spectra), clearly jus-

tifying further work in this area.

Closely related to the binary silver oxides are the silver

oxide clathrate salts of empir_ꢀical formula Ag₇O₈X

ðX ¼ NO₃^ꢀ; ClO₄^ꢀ; HSO₄^ꢀ; HCO₃or HF₂^ꢀÞ

[26–36].

2Ag^þ_ðaqÞþO_3ðgÞþH₂O_ðlÞ! Ag₂O_2ðsÞþO_{2ð gÞ}þ2H^þ

ð1Þ

ð2Þ

ð3Þ

The formal valence conﬁguration for these compounds is

(Ag²⁺)₂(Ag³⁺)₄O₈Æ Ag⁺X^ꢀ, although their physical prop-

erties and structural data suggest a (Ag^2.67+)₆O₈Æ Ag⁺X^ꢀ

formulation is more accurate [26–32]. At room temperature,

all Ag₇O₈X salts have a fcc crystal structure (space group

ðaqÞ

6Ag^þ_ðaqÞþ5O_3ðgÞþ3H₂O_ðlÞ! 2Ag₃O_4ðsÞþ5O_2ðgÞþ6H^þ

ðaqÞ

2Ag^þ_ðaqÞþ2O_3ðgÞþH₂O_ðlÞ! Ag₂O_3ðsÞþ2O_2ðgÞþ2H^þ

ðaqÞ

7Ag^þ_ðaqÞþX_ð^ꢀ_aqÞþ5O_3ðgÞþ3H₂O_ðlÞ! Ag₇O₈X_ðsÞþ5O_2ðgÞþ6H_ð^þ_aqÞ

ð4Þ

˚

ꢀ

Fm3m, Z = 4, a = 9.82–9.94 A) based on face-sharing, 26-

sided, cubo-octahedra [26–32]. Oxygen anions in the

‘Ag₆O₈’ structural unit are located at the 24 corners of

the cubo-octahedra (each is common to three cubo-octahe-

dral units). The six Ag cations in the ‘Ag₆O₈’ unit occupy

structurally equivalent positions in the square faces com-

mon to two cubo-octahedra (each is square-planar coordi-

For the reasons outlined above, a comprehensive study

of the aqueous Ag⁺/O₃system was undertaken. We

aimed to synthesize Ag₂O₂, Ag₃O₄, Ag₂O₃and Ag₇O₈X

salts by the ozonolysis route (Eqs. (1)–(4)), then compre-

hensively characterize the reaction products using a vari-

ety of spectroscopic and thermo-analytical methods.

Particular emphasis is placed on examination of the

vibrational spectra and electronic structure of these

compounds.

˚

nated by oxygen with Ag–O = 2.05 A). As these cations are

structurally indistinguishable, an intermediate valency of

2.67 is implied. Ag⁺ions in the Ag₇O₈X lattice occupy

cubic sites between the cubo-octahedra, and are coordi-

3312

G.I.N. Waterhouse et al. / Polyhedron 26 (2007) 3310–3322

2. Experimental

2.1. General

X-ray tube and curved graphite ﬁlter monochromator.

XRD data was collected at room temperature, using mono-

˚

chromatized Cu K_aX-rays (k = 1.5418 A). Additional

acquisition parameters were: 2h range, 2–90ꢁ; step size,

0.02ꢁ; and scan rate, 2ꢁ min^ꢀ1. Diﬀraction patterns were

referenced against the JCPDS database for sample identiﬁ-

cation [42–47].

Silver(I) salts were purchased from Aldrich and used

without further puriﬁcation. Ag₂O and Ag₂O₂powders

were prepared from AgNO₃in accordance with established

procedures [40,41], and used as reference materials.

2.3.2. SEM

2.2. Synthesis of the binary silver oxides and Ag₇O₈X salts

Sample morphology was examined using a Philips XL-

30 ﬁeld emission gun scanning electron microscope (FEG-

SEM). All micrographs were collected at an electron gun

accelerating voltage of 5 kV. Specimens were mounted on

black carbon tape for analysis.

During preliminary studies, we determined that the

product obtained during ozonolysis of aqueous solutions

of Ag⁺at 298 K was highly dependent on the anion of the

silver salt used to prepare the silver solution. Ozonolysis

of aqueous solutions of AgNO₃, AgClO₄and Ag₂SO₄aﬀor-

ded Ag₇O₈NO₃, Ag₇O₈ClO₄or Ag₇O₈HSO₄, respectively,

whereas ozonolysis of 0.05 M aqueous solutions of AgF

or AgBF₄yielded Ag₃O₄. To prevent thermal decomposi-

tion of Ag₃O₄during the synthesis, the ozonolysis of aque-

ous solutions of AgF or AgBF₄was performed at 273 K.

Monoclinic Ag₂O₂was the product collected by ozonolysis

of aqueous solutions of CH₃COOAg at 273–298 K.

The general procedure used to prepare these compounds

was as follows. Aqueous solutions of the silver salts were

prepared using milli Q water. Ag⁺concentrations used in

the ozonolysis experiments were determined by the respec-

tive solubilities of the diﬀerent silver salts in H₂O at 273–

298 K. The pH of the Ag⁺solution was measured, after

which it was transferred to a 1 L. glass round-bottomed

ﬂask; ﬁtted with a stirrer, a sintered glass frit bubbler for

introducing gaseous O₃-O₂mixtures and a gas outlet tube.

The ﬂask was then placed in a thermostated water bath and

allowed to equilibrate for 1 h at a speciﬁed temperature in

the range 273–298 K. An O₃–O₂gas mixture (5 vol% O₃in

O₂, total ﬂow rate O₃+ O₂= 200 L h^ꢀ1), generated from

dry O₂using a commercial ozone generator (Fischer model

502), was then bubbled through the silver(I) solution for a

speciﬁed time (0.5–15 h). After the speciﬁed time, ozonoly-

sis was discontinued and the solid product collected by vac-

uum ﬁltration on a pre-weighed sintered glass crucible

(porosity 4). The ﬁltrate was retained and its pH measured.

2.3.3. XPS

XPS spectra were collected using a Kratos XSAM800

instrument equipped with an analytical chamber of base

pressure ꢁ5 · 10^ꢀ10Torr, a dual anode X-ray source

(Mg/Al), and a hemispherical electron energy analyzer

with multi-channel detector. For analysis, samples were

lightly pressed to form a pellet then mounted on stainless

steel stubs using carbon sticky dots. Spectra were excited

using Mg Ka radiation (hm = 1253.6 eV). The hemispher-

ical analyzer was operated in the ﬁxed analyzer transmis-

sion (FAT) mode, and the analyzer axis was normal to

the sample surface. Survey spectra were collected at an

analyzer pass energy of 65 eV. Multiplet spectra were col-

lected over narrow spectral regions at a pass energy of

38 eV. All XPS spectra were referenced against the Fermi

edge. Elemental compositions were determined from the

survey spectra by standard quantiﬁcation procedures

using peak areas and spectrometer-modiﬁed Scoﬁeld

cross-sections. Estimated uncertainty in the percentage

atomic concentration reported for each element is approx-

imately 1–2%.

2.3.4. FT-IR

Far-infrared measurements were performed on a Digi-

lab FTS-60 series spectrometer. Incorporated equipment

included an FTS-60V vacuum optical bench with a

5 lines/mm wire mesh beam splitter, a mercury lamp source

and a pyroelectric triglycine sulfate detector. Spectra were

recorded at 2 cm^ꢀ1resolution over the range 50–

700 cm^ꢀ1using 500 signal-averaged scans. Prior to data

collection, the optical bench was evacuated to a pressure

of ꢁ10^ꢀ2Torr. Mid-infrared studies were carried out using

a Digilab FTS-60 series spectrometer employing a DTGS

detector. Acquisition parameters were: range = 400–

4000 cm^ꢀ1; resolution = 2 cm^ꢀ1; number of signal-averaged

scans, = 500. Samples were analyzed using the polyethyl-

ene disk sampling method (solid state ion exchange reac-

tions between the silver compounds and KBr precluded

the use of the KBr disk technique). Fresh sample (5 mg)

was mixed with polyethylene powder (20 mg) then com-

pressed at a pressure of 5 ton per-square-inch for 5–

10 min to form thin disks of diameter 13 mm. Pellets were

The solid product was then washed repeatedly with ice-cold

ꢀ

deionised water, then acetone. Ag₇O₈X salts ðX ¼ NO₃

;

ClO₄^ꢀ; HSO₄^ꢀÞ were dried at 313 K under vacuum for

2 h. After drying, the sample and sintered glass crucible

were reweighed, and the yield of product determined by

diﬀerence. Ag₃O₄powders were not vacuum dried, and

instead were immediately transferred to a freezer at

255 K to prevent thermal decomposition before spectro-

scopic analysis.

2.3. Product characterization

2.3.1. XRD

Powder X-ray diﬀraction patterns were acquired using a

Philips PW-1130 diﬀractometer, equipped with a Cu anode

G.I.N. Waterhouse et al. / Polyhedron 26 (2007) 3310–3322

3313

mounted in a stainless steel disc holder for the collection of

infrared spectra.

with Fig. 1, which shows XRD patterns for the various

reaction products isolated. As mentioned in Section 2.2,

the anion of the silver(I) salt had a strong inﬂuence on

the ozonolysis product. Ozonolysis of neutral or mildly

acidic solutions of AgNO₃, AgClO₄or Ag₂SO₄at 298 K

aﬀorded Ag₇O₈NO₃, Ag₇O₈ClO₄and Ag₇O₈HSO₄, respec-

tively. XRD patterns for these clathrate compounds were

indistinguishable (Fig. 1a–c), in agreement with the fact

that the crystallographic parameters for all Ag₇O₈X salts

2.3.5. Raman

Raman spectra were taken on a Renishaw RM-2 Raman

microscope system, ﬁtted with an Olympus 20· objective

lens. Raman spectra were excited using either 785 nm radi-

ation from a diode laser or 488 nm radiation from an argon

ion laser, and scattered light was collected at an angle 180ꢁ

to the incident beam direction (180ꢁ backscattering sam-

pling geometry). The laser power at the sample for routine

data collection was ꢁ0.3 mW, although higher powers were

also used to study the photosensitivity of selected samples.

Spectra were collected over the range 200–2000 cm^ꢀ1and

at a resolution of ca. 2 cm^ꢀ1. To improve the signal-

to-noise ratio of the data reported 20 consecutive scans

were added to produce a spectrum.

ꢀ

˚

are similar (space group Fm3m, Z = 4, a = 9.82–9.94 A)

[26–32]. For each synthesis, the molar yield ratio

H⁺/Ag₇O₈X ðX ¼ NO ^ꢀ; ClO₄^ꢀ; HSO₄^ꢀÞ was in the range

3

5.8–5.9, close to the theoretical ratio of 6 predicted by Eq.

(4). The nature of the anion appears to be important for

the stabilization of the Ag₇O₈X clathrate structure. Poly-

oxoanions appear to be ideal in this regard. Ozone treat-

ment of aqueous solutions of AgF or AgBF₄at 273 K

yielded Ag₃O₄in high purity (Fig. 1d and e). This is an

exciting result and the ﬁrst reported synthesis of Ag₃O₄

by non-electrochemical routes. The molar yield ratio

H⁺/Ag₃O₄for both experiments was 2.9, in good accord

with the theoretical value of 3 (Eq. (2)). Monoclinic

Ag₂O₂was the product obtained by ozonolysis of aque-

ous solution of silver acetate at 273–298 K (Fig. 1f). Ace-

tate ions acted to buﬀer the solution, since the pH of the

solution decreased only from 8.7 to 5.9 after 3 h ozonol-

ysis. Ag₂O₂may have formed directly as described by

Eq. (1), or alternatively is a decomposition product of a

higher silver oxide, possibly Ag₂O₃. The latter is very

thermally unstable at 298 K [12]. Our attempts to prepare

Ag₂O₃by ozonolysis of aqueous solutions of AgClO₄or

AgBF₄at low temperature, as has been achieved electro-

chemically [11,12], have proved unsuccessful to date.

To examine the eﬀect of other experimental parameters

on the ozone-mediated syntheses, the precipitation of

Ag₇O₈NO₃was examined in further detail.

2.3.6. DSC

DSC data was obtained on a DSC-2 instrument (Rheo-

metric Scientiﬁc Ltd.). Samples (10–15 mg) were weighed

into open aluminium crucibles, and referenced against an

empty aluminium crucible during calorimetric measure-

ments. DSC traces were recorded at a heating rate of

10 K min^ꢀ1, under a dry air purge.

3. Results and discussion

3.1. Ozone-mediated synthesis of binary silver oxides and

Ag₇O₈X salts

3.1.1. Eﬀect of silver(I) salt

The ﬁrst objective of this study was to identify of the

product(s) formed by sparging ozone through aqueous

solutions of diﬀerent Ag(I) salts. Table 1 summarizes

our key ﬁndings. Results are discussed in conjunction

Table 1

Products obtained by the ozonolysis of aqueous solutions of diﬀerent Ag(I) salts

Silver salt

AgNO₃

Conc. (M)

Vol. (L)

I (M)

Ozonolysis temp. (K)

Ozonolysis time (h)

Initial pH

Final pH

Product

% Yield

0.1

0.75

0.1

298

0.5

2

5

10

15

5.5

2.9

2.2

1.7

1.5

1.4

Ag₇O₈NO₃

1.9

9.1

23.8

42.0

52.7

AgNO₃

0.1

0.75

1.1

298

5

10

0

2

5.5

10

14^*

2.1

1.8

1.7

Ag₇O₈NO₃

Ag₂O₂

11.1

18.9

24.0

24.5

99.8

AgClO₄

Ag₂SO₄

AgF

AgBF₄

CH₃COOAg

*

0.1

0.75

0.30

0.75

0.1

298

273

5

3

5.5

8.7

1.7

1.9

1.6

5.9

Ag₇O₈ClO₄

Ag₇O₈HSO₄

Ag₃O₄

Ag₂O₂

23.6

55.5

46.1

44.7

30.1

0.015

0.05

0.045

0.05

At pH 14, Ag(I) precipitates from solution as Ag₂O, which is oxidized by O₃to monoclinic Ag₂O₂.

3314

G.I.N. Waterhouse et al. / Polyhedron 26 (2007) 3310–3322

3.1.3. Eﬀect of initial pH

To investigate the eﬀect of initial pH on the synthesis of

Ag₇O₈NO₃, we prepared 0.1 M AgNO₃solutions in either

dilute HNO₃(pH 0–4) or NaOH (pH 8–14) and used these

in the ozonolysis experiments. NaNO₃was also added to

achieve a constant ionic strength (1.1 M).

Ag₇O₈NO₃

Ag₇O₈ClO₄

(a)

The data in Table 1 shows that the initial pH had a

strong inﬂuence on product composition and yields. For

initial pH 0–10, ozonolysis yielded Ag₇O₈NO₃exclusively.

Product yields increased with initial pH in this range. For

pH 10–13, the precipitation of Ag₂O and the oxidation

by O₃of Ag₂O to Ag₂O₂compete with Ag₇O₈NO₃forma-

tion. The proportions of Ag₂O and Ag₂O₂in the product

increased with increasing initial pH. At pH 13 and above,

only Ag₂O and Ag₂O₂formed. At pH 14, complete oxida-

tion of the initially formed Ag₂O precipitate to Ag₂O₂was

achieved in 10 h.

(b)

(c)

Ag₇O₈HSO₄

Ag₃O₄

(d)

(e)

(f)

3.2. SEM characterization of the ozonolysis products

The growth morphology of the products was examined

by SEM. Fig. 2a–d shows the morphological evolution of

Ag₇O₈NO₃particles as a function of ozonolysis time. Par-

ticles formed after 30 min were predominantly of cubic

shape (exposing 6 square {100} faces) and submicron size

(Fig. 2a). With increasing ozonoloysis time, the particles

became larger and developed an octahedral habit (exposing

8 triangular {111} faces) or forms derived from octahedra

by truncation of corners and edges (Fig. 2b–d). After 15 h

ozonolysis, the Ag₇O₈NO₃particles were typically of trun-

cated octahedral habit, twinned and of average size ꢁ5 lm

(Fig. 2d). Similar growth morphologies were observed for

the Ag₇O₈ClO₄and Ag₇O₈HSO₄powders (SEM micro-

graphs not shown).

Ag₂O₂

20

10

30

40

50

60

70

2θ (degrees)

Fig. 1. Powder XRD patterns for the products obtained by ozonolysis of

aqueous solutions of (a) AgNO₃(0.1 M, 298 K, 5 h); (b) AgClO₄(0.1 M,

298 K, 5 h); (c) Ag₂SO₄(0.015 M, 298 K, 5 h); (d) AgF (0.05 M, 273 K,

3 h); (e) AgBF₄(0.05 M, 273 K, 3 h) and (f) CH₃COOAg (0.05 M, 273 K,

3 h). Silver salt concentrations, reaction temperatures and ozonolysis times

are given in parentheses. Products identiﬁed are listed next to the XRD

patterns.

3.1.2. Eﬀect of ozonolysis time

Ozonolysis of 0.1 M AgNO₃(initial pH 5.5) at 298 K

for 0.5–15 h yielded Ag₇O₈NO₃as the only solid phase

(Table 1). The yield of Ag₇O₈NO₃increased almost line-

arly with time during the ﬁrst 5 h of reaction, but tended

towards an asymptotic limit of approximately 55% at

longer reaction times. A yield of 52.7% was achieved

after 15 h. The pH of the solution decreased sharply dur-

ing the initial stages of the reaction, but maintained an

almost constant value of ꢁ1.4 after 15 h. The molar yield

ratio H⁺/Ag₇O₈NO₃at all times was between 5.7 and

5.9, again close to the theoretical value of 6 predicted

by Eq. (4).

Results suggest that the precipitation of Ag₇O₈NO₃is

likely aﬀected by competing dissolution processes at low

pH. To test whether competing precipitation-dissolution

phenomena were a factor in our studies, we determined a

dissolution curve for Ag₇O₈NO₃powder in aqueous

HNO₃(not shown). The stability of Ag₇O₈NO₃particles

in HNO₃solution decreased sharply in the pH range 1–2.

This conﬁrmed that competitive precipitation and dissolu-

tion of Ag₇O₈NO₃particles was likely during ozonolysis at

low pH values, and this limits the yield of product

achievable.

Ag₃O₄obtained by ozonolysis of aqueous solutions of

AgF at 273 K displayed a rhombohedral prismatic habit

(Fig. 2e). Monoclinic Ag₂O₂resulting from the ozonolyis

of aqueous silver acetate crystallized in the form of irregu-

lar platelets (Fig. 2f).

3.3. Electronic structure, vibrational spectra and thermal

decomposition of Ag₇O₈X salts

As stated in the introduction, Ag₇O₈X salts are metals

and the only known clathrates to exhibit low temperature

superconductivity [28,29]. Minimal spectroscopic informa-

tion is available for these compounds, motivating a

detailed investigation here.

3.3.1. XPS characterization of Ag₇O₈NO₃and Ag₇O₈HSO₄

Information about the electronic states of silver in these

compounds can be deduced from the high-resolution XPS

spectra shown in Fig. 3. The spectra are discussed in con-

junction with Table 2, which provides XPS data summaries

for the Ag₇O₈X salts and reference data collected for

metallic silver and other relevant silver compounds. It is

useful to note that the spectra presented in Fig. 3 are in

G.I.N. Waterhouse et al. / Polyhedron 26 (2007) 3310–3322

3315

Fig. 2. SEM micrographs of Ag₇O₈NO₃powders obtained by the ozonolysis of 0.1 M AgNO₃at 298 K for (a) 0.5 h; (b) 2 h; (c) 5 h and (d) 15 h. Fig. 2e

shows the Ag₃O₄powder obtained by ozonolysis of 0.05 M AgF at 273 K for 3 h and Fig. 2f monoclinic Ag₂O₂precipitated by ozonolysis of 0.05 M

CH₃COOAg at 273 K for 3 h. Image magniﬁcations are 10000, 10000, 10000, 10000, 20000 and 40000·, respectively.

O 1s

Valence

Ag 3d

(a)

(b)

(a)

(b)

(a)

(b)

0

380 378 376 374 372 370 368 366 364 362

Binding Energy (eV)

10

8

6

4

2

536 534 532 530 528 526

Fig. 3. XPS O 1s, Ag 3d and valence-band spectra for (a) Ag₇O₈NO₃and (b) Ag₇O₈HSO₄. Spectra were referenced relative to the Fermi edge (E_F= 0 eV).

3316

G.I.N. Waterhouse et al. / Polyhedron 26 (2007) 3310–3322

Table 2

Summarized XPS data for metallic silver and various silver compounds

Sample

Atomic composition (%)

O/Ag ratio

Theor.

Binding energy (eV)

Ag

O

N

S

C

Exp.

Ag 3d_5/2

O 1s

N 1s

S 2p_3/2

Ag

Ag₂O

60.9

38.9

15.5

24.6

36.8

21.9

21.4

19.0

22.4

46.8

48.9

39.9

40.3

41.8

20.1

38.7

22.0

14.3

23.3

34.3

33.6

0

0.31

0.58

3.02

1.99

1.08

1.84

1.95

368.2

367.8

368.0

367.9

367.4

531.0

529.6

532.1

531.7

528.6

528.5

0.5

3.0

2.0

1.0

1.57

1.71

AgNO₃

Ag₂SO₄

Ag₂O₂

Ag₇O₈NO₃

Ag₇O₈HSO₄

15.7

3.1

406.1

405.8

12.2

3.2

168.0

167.6

excellent qualitative agreement with data reported by Bao

et al. for electrochemically synthesized Ag₇O₈NO₃[35].

XPS spectra collected for non-metallic samples typically

need to be corrected for specimen charging during X-ray

irradiation, unless precautions are taken to dissipate the

build up of positive charge (such as ﬂooding the specimen

surface with low energy electrons during data acquisition).

No such charging correction was necessary for spectra

taken from the Ag₇O₈X salts, evidence of their metallic

character.

the XPS chemical shift [16–18]. The Ag 3d peaks for

Ag₇O₈NO₃and Ag₇O₈HSO₄were also broad compared

to those measured for Ag and Ag₂O (FWHM = 1.25 eV

and 1.47 eV, respectively), and of similar width to those

of measured for Ag₂O₂(FWHM = 1.9 eV), strong evidence

that Ag₇O₈X compounds contain silver in several diﬀerent

oxidation states. The Ag 3d spectra for the Ag₇O₈X salts

also show satellites on the low binding energy side of the

Ag 3d_3/2and Ag 3d_5/2peaks, arising from ﬁnal state relax-

ation eﬀects.

The elemental composition determined by quantitative

XPS analysis for Ag₇O₈NO₃was Ag (21.9%), O (40.3%),

N (3.1%) and C (34.3%) (Table 2). The carbon impurity

results predominantly from the adsorption of residual

hydrocarbons in the UHV chamber of the XPS instrument.

The experimental Ag/N ratio was close to 7, consistent

with the empirical formula of the compound. The O/Ag

ratio of 1.84 exceeded the theoretical value of 1.57.

Independent ICP-MS and thermogravimetric analyses of

the compound revealed a composition very close to

Ag₇O₈NO₃, indicating that the oxygen excess determined

by XPS was likely localized at the specimen surface (XPS

is surface-analytical technique). The cause of this oxygen

excess is examined below. The elemental composition data

for Ag₇O₈HSO₄was similar (Table 2), in that the Ag/S

ratio was approximately 7, but the O/Ag ratio of 1.95

slightly exceeded the theoretical value of 1.71.

The O 1s spectra contain two distinct peak maxima. The

narrow peak at 528.5 eV (FWHM 1.4 eV) is typical for lat-

tice oxygen (formally O^2ꢀ) in silver oxides [15–20], and in

excellent agreement with XPS data reported here for bulk

Ag₂O₂(Table 2). We assign this feature to oxide ions of

the face-sharing Ag₆O₈structural units in Ag₇O₈NO₃and

Ag₇O₈HSO₄. The feature at ꢁ531.4 eV in the O 1s spectra

of both compounds is very broad (FWHM 2.8 eV), sug-

gesting the presence of multiple oxygen species. The poly-

oxoanions of the Ag₇O₈X salts will contribute to this

feature (cf. AgNO₃

O

1s = 532.1 eV; Ag₂SO₄

O

1s = 531.7 eV), as also will surface hydroxyl species

(531.7 eV) and possibly chemisorbed H₂O (533.0 eV) since

the compounds were prepared from aqueous solution

[18–20]. For both compounds, spectroscopic evidence was

found in the C 1s spectrum for a surface carbonate species

(C 1s = 289.0 eV), which typically gives an O1s peak

between 530.5 and 531 eV [19,20]. Carbonate formation

results from the reaction between surface oxygen atoms

and atmospheric CO₂, and is a characteristic common to

all silver oxides (cf. Table 2, carbonate formation is respon-

sible for the high O/Ag ratios observed experimentally for

Ag₂O and Ag₂O₂). The presence of surface hydroxyl and

carbonate species, and possibly chemisorbed H₂O, satisfac-

torily explain the high O/Ag ratios determined by quantita-

tive XPS analysis for the Ag₇O₈X salts.

The core-level Ag 3d spectra for Ag₇O₈NO₃and

Ag₇O₈HSO₄show broad, asymmetric peaks at 373.4 eV

and 367.4 eV (FWHM = 2.0 eV), which are readily

assigned to core-level Ag 3d_3/2and Ag 3d_5/2photoemis-

sions, respectively. The Ag 3d peaks are shifted by

ꢀ0.8 eV relative to a those of Ag metal (Ag 3d_5/2

=

368.2 eV), the magnitude and direction of the shift typical

for highly oxidized states of silver [16–20,35]. Silver is unu-

sual in that its core-levels photoemissions shift to lower

binding energy with increasing oxidation state (cf. Table

2, Ag 3d_5/2binding energies for Ag, Ag₂O and Ag₂O₂are

368.2, 367.8 and 367.4 eV, respectively). For most metals,

positive binding energy shifts are observed when the metal

is oxidized, on account of the extra coulombic attraction of

core-level electrons towards the nucleus in the oxidized

state. The negative binding energy shift observed for silver

with increasing oxidation state is rationalized in terms of

the large contribution of ﬁnal state relaxation eﬀects to

N_ꢀ1s and S 2_ꢀp_3/2binding energies measured for the

NO₃and HSO₄ions in Ag₇O₈NO₃and Ag₇O₈HSO₄,

respectively, were in excellent agreement with those deter-

mined for the same anions in the AgNO₃and Ag₂SO₄

(Table 2). S 2p binding energies are similar for metal salts

containing HSO₄and SO₄(both S⁶⁺) [48].

ꢀ

2ꢀ

Valence-band spectra for Ag₇O₈NO₃and Ag₇O₈HSO₄

(Fig. 3) comprise an intense peak at 4.7 eV and a weaker

feature at 2.0 eV. The 4.7 eV peak exhibits asymmetry on

G.I.N. Waterhouse et al. / Polyhedron 26 (2007) 3310–3322

3317

conditions (DSC) and stepwise isothermal heating condi-

tions (XRD and FT-IR spectroscopy, Section 3.3.3.

below).

The DSC trace for Ag₇O₈NO₃(Fig. 4a) comprises a

large exothermic peak at 430 K, two small endothermic

peaks at 478 K and 698 K, and a large endothermic peak

at 717 K. The trace is qualitatively similar to that reported

for electrochemically synthesized Ag₇O₈NO₃[33], and the

peaks assigned accordingly. The large exothermic peak at

430 K is due predominantly to the thermal decomposition

of Ag₇O₈NO₃(Eq. (5)), but will also contains contributions

from the thermal decomposition of Ag₂O₂(Eq. (6)) and the

orthorhombic ! tetragonal phase transition of AgNO₃

[21,33].

(a)

Ag₇O₈NO_3ðsÞ! 3Ag₂O_2ðsÞþ AgNO_3ðsÞþ O_2ðgÞ

3Ag₂O_2ðsÞ! 3Ag₂O_ðsÞþ 1:5O_2ðgÞ

ð5Þ

ð6Þ

(b)

The endothermic peaks at 478 K, 698 K and 717 K in

Fig. 4a result from the melting of tetragonal AgNO₃, the

thermal decomposition of AgNO₃(Eq. (7)) and the thermal

decomposition of Ag₂O (Eq. (8)), respectively.

300

400

500

600

700

800

Temperature (K)

Fig. 4. DSC curves for (a) Ag₇O₈NO₃and (b) Ag₃O₄. All data was

collected at a heating rate of 10 K min^ꢀ1under a dry air purge.

Ag₇O₈NO₃

(a)

the high binding energy side, consistent with presence of

more than one component. By comparison with valence-

band spectra reported for other silver compounds

[15–20], we assign the 4.7 eV peak to orbitals of Ag 4d

character. The broadness and asymmetry of the peak

reﬂect the mixed-valence nature of the Ag₇O₈X com-

pounds. The feature at about 2.0 eV originates from

valence orbitals of mostly O 2p character. The metallic

conductivity of Ag₇O₈X salts is attributed to promotion

of two electrons from the face-sharing Ag₆O₈structural

units into a conduction band involving Ag 4d–O 2p orbi-

tals (i.e. electron hopping between metal centres of the

‘Ag₆O₈’ units via bridging oxygen atoms). Consistent with

this proposal, the valence-band spectra of the Ag₇O₈NO₃

and Ag₇O₈HSO₄show overlap between valence and con-

duction bands at the Fermi edge (E_F= 0 eV). The XPS

data thus strongly supports the mixed-valence formulation

(Ag^2.67+)₆O₈Æ Ag⁺X^ꢀfor these compounds.

(b)

Ag₂O₂+

AgNO₃

x 2

(c)

Ag₂O + AgNO₃

(d)

(e)

Ag₂O + AgNO₃

Ag₂O

(f)

Ag₂O + Ag

(g)

3.3.2. Thermal decomposition of Ag₇O₈NO₃in air under

dynamic heating conditions

Ag

(h)

Detailed knowledge about the thermal stability of the

Ag₇O₈X salts synthesized in this work is important as these

compounds show commercial application potential in sev-

eral areas; most notably in antibacterial coatings on

account of the enormous oxidising power of Ag²⁺and

Ag³⁺centres [36]. Here, we examined the thermal decom-

position of Ag₇O₈NO₃in air under both dynamic heating

20

30

40

50

60

2θ (degrees)

Fig. 5. XRD patterns for the thermal decomposition of Ag₇O₈NO₃in air.

Fresh samples were calcined 1 h at (a) 363 K; (b) 388 K; (c) 398 K; (d)

423 K; (e) 573 K; (f) 623 K; (g) 648 K and (h) 673 K. All data was

collected at room temperature. The main phases identiﬁed are listed on the

left side of the ﬁgure.

3318

G.I.N. Waterhouse et al. / Polyhedron 26 (2007) 3310–3322

The XRD pattern obtained following calcination of

Ag₇O₈NO₃at 363 K (Fig. 5a) was identical to that col-

lected for the freshly prepared salt (cf. Fig. 1a), indicating

no decomposition occurred during this treatment. The cor-

responding mid-IR and far-IR spectra for the compound

dispersed in a low-density polyethylene disk are shown in

Fig. 6a The spectra contain no discernable peaks, and were

identical to that of a pure polyethylene disk of the same

thickness (not shown), except for the fact they showed sig-

niﬁcantly increased absorbance (i.e. relative to the spec-

trum of pure polyethylene, spectra for Ag₇O₈NO₃in

polyethylene were shifted up the absorbance axis by an

amount proportional to the Ag₇O₈NO₃loading). The fact

that bands due to the component ions of Ag₇O₈NO₃are

not evident in Fig. 6a is indicative of metallic behaviour

AgNO_3ðlÞ! 0:5Ag₂O_ðsÞþ NO_2ðgÞþ 0:25O_2ðgÞ

ð7Þ

ð8Þ

3:5Ag O ! 7Ag_ðsÞþ 1:75O_2ðgÞ

ðsÞ

2

The overall decomposition reaction for Ag₇O₈NO₃in air is

thus

Ag₇O₈NO_3ðsÞ! 7Ag_ðsÞþ NO_2ðgÞþ 4:5O_2ðgÞ

ð9Þ

DSC curves for Ag₇O₈ClO₄and Ag₇O₈HSO₄heated in air

were also collected (not shown). The curves were similar in

most respects to that presented in Fig. 4a for Ag₇O₈NO₃,

although the onset temperature for the thermal decomposi-

tion of these compounds was ca. 5 K lower than that ob-

served for Ag₇O₈NO₃, suggesting they have slightly lower

thermal stability.

3.3.3. Vibrational spectra and thermal decomposition of

Ag₇O₈NO₃in air under isothermal heating conditions

Structural changes occurring during the thermal decom-

position of Ag₇O₈NO₃in air under isothermal conditions

were followed by powder XRD (Fig. 5) and FT-IR spec-

troscopy (Fig. 6). Freshly prepared Ag₇O₈NO₃powders

were calcined for 1 h at selected temperatures in the range

323–773 K, then quickly cooled to room temperature and

analyzed by the two techniques. Due to the nature of the

heating program, onset temperatures for reactions 5–8 reg-

istered by XRD and FT-IR spectroscopy were consider-

ably lower than those determined by DSC.

and lends further support for the valence formulation

ðAg^2:67þÞ O₈ꢂ Ag^þNO

.

ꢀ

6

3

The thermal decomposition of Ag₇O₈NO₃commenced

around 373 K, in agreement with literature reports that

Ag₇O₈NO₃is transformed to 3Ag₂O₂+ AgNO₃+ O₂with

prolonged immersion in boiling water [7]. Spectroscopic

evidence to support this is found in Fig. 5a–c, where

XRD peaks characteristic for Ag₇O₈NO₃disappeared on

increasing the calcination from 363 K to 398 K. Simulta-

neously, new peaks appeared and were indexed to mono-

clinic Ag₂O₂[43], orthorhombic AgNO₃[44] and Ag₂O

[45]. The same changes manifest themselves in the FT-IR

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

0

200

400

600

800

1000

1200

1400

Wavenumber (cm^-1)

Fig. 6. FT-IR absorbance spectra for the thermal decomposition of Ag₇O₈NO₃in air. Fresh samples were calcined 1 h at (a) 363 K; (b) 388 K; (c) 398 K;

(d) 423 K; (e) 573 K; (f) 623 K; (g) 648 K and (h) 673 K. All data was collected at room temperature using the pressed polyethylene disk sampling method.

The spectrum of the polyethylene disk was removed by subtraction.

G.I.N. Waterhouse et al. / Polyhedron 26 (2007) 3310–3322

3319

spectra of Fig. 6a–c, where bands characteristic for mono-

clinic Ag₂O₂(60, 69, 97, 155, 192, 254, 380, 520, 556 cm^ꢀ1

to a spectrum collected at the same laser power using

488 nm excitation (not shown). Results conﬁrm the metal-

lic reﬂectivity of Ag₇O₈NO₃powders over the visible and

near-infrared regions. Excitation at 785 nm and higher

laser power (ca. 10 mW at sample) produced the spectrum

shown in Fig. 7b, which contains bands at 219, 300, 376,

428, 467, 485, 525 (broad), 972 and 1043 cm^ꢀ1. By compar-

ison with Raman spectra collected for monoclinic Ag₂O₂

(Fig. 7c), and orthorhombic AgNO₃(Fig. 7d), it is clear

that the use of a high laser power caused sample heating

and the thermal decomposition of Ag₇O₈NO₃to

3Ag₂O₂+ AgNO₃+ O₂(Eq. (5)). The broad band at

525 cm^ꢀ1is characteristic for Ag₂O [19–21], indicating par-

tial decomposition of Ag₂O₂also occurred (Eq. (6)).

Fig. 7b is qualitatively similar to a spectrum reported by

Pettinger et al. for electrochemically synthesized

Ag₇O₈NO₃[34], suggesting laser damage to the sample also

occurred in that study (a laser power of 20 mW was used

for data acquisition).

)

and orthorhombic AgNO₃(80, 103, 125, 170, 801, 1043,

1275, 1350 cm^ꢀ1), then Ag₂O (86, ꢁ540 cm^ꢀ1), develop

with increasing calcination temperature up to 398 K.

Assignments for these bands may be found in Refs

[21,49,50]. The XRD and FT-IR data obtained following

sample heating at 423 K (Figs. 5d and 6d, respectively)

were those of Ag₂O and AgNO₃, indicating that complete

thermal decomposition of the initially formed Ag₂O₂was

eﬀected by this treatment.

Diﬀraction peaks and IR bands characteristic for

AgNO₃lost intensity with temperature between 573 K

and 623 K (Figs. 5e–f and 6e–f), which is explained by

the thermal decomposition of AgNO₃(Eq. (7)) [51]. With

increasing temperature above 623 K, spectral data typical

of Ag₂O was progressively replaced by that of Ag metal

(Figs. 5f–h and 6f–h), reﬂecting the thermal decomposition

of Ag₂O to its component elements (Eq. (8)). Silver was the

only phase identiﬁed by XRD after sample calcination at

673 K (Fig. 5h). Note the similarity between the FT-IR

spectra of Ag₇O₈NO₃(Fig. 6a) and silver metal (Fig. 6h).

3.4. Electronic structure, vibrational spectra and thermal

decomposition of Ag₃O₄

3.3.4. Raman spectra for Ag₇O₈NO₃

The synthesis of Ag₃O₄in high purity is a key ﬁnding of

the present study. In view of fact that the electronic struc-

ture of Ag₃O₄is unclear and minimal spectroscopic infor-

mation is available for the compound, we used DSC,

Fig. 7a shows a Raman spectrum for Ag₇O₈NO₃excited

at 785 nm and low laser power (ꢁ0.3 mW at sample). The

spectrum shows no discernable features, and was identical

(a)

Ag₃O₄

(a)

x 0.5

(b)

(c)

(d)

(e)

x 0.05

(b)

(c)

x 0.03

(d)

Ag₂O₂

25

(f)

200

400

600

800 1000 1200

20

30

35

40

45

50

Raman Shift (cm^-1)

2θ (degrees)

Fig. 7. Raman spectra excited at 785 nm for (a) Ag₇O₈NO₃, 0.3 mW; (b)

Ag₇O₈NO₃, 10 mW; (c) monoclinic Ag₂O₂, 5 mW and (d) orthorhombic

AgNO₃, 5 mW. Laser powers at the sample are also given.

Fig. 8. XRD patterns for the thermal decomposition of Ag₃O₄in air at

298 K. Time after preparation was (a) 0.5 h; (b) 4 h; (c) 8 h; (d) 12 h; (e)

16 h and (f) 20 h.

3320

G.I.N. Waterhouse et al. / Polyhedron 26 (2007) 3310–3322

XRD, FT-IR and Raman spectroscopy to examine the

thermal decomposition of Ag₃O₄in air. From the vibra-

tional spectra, we hoped to establish the correct valence

structure for the compound. Attempts to collect XPS data

for Ag₃O₄proved unsuccessful due to the thermal instabil-

ity of the compound.

The overall decomposition reaction for AgO₄in air is thus

Ag₃O_4ðsÞ! 3Ag_ðsÞþ 2O_2ðgÞð13Þ

3.5.1. Vibrational spectra and thermal decomposition of

Ag₃O₄in air at 298 K

XRD, FT-IR and Raman spectroscopy were used to

document the thermal decomposition of Ag₃O₄to

1.5Ag₂O₂+ 0.5O₂in air at 298 K. Freshly prepared

Ag₃O₄powders were left in air at 298 K for 24 h. At regu-

lar time intervals, samples were taken and quickly analyzed

using the three techniques. The data obtained is shown in

Figs. 8–10, respectively.

Fig. 8a shows the XRD pattern collected for freshly pre-

pared Ag₃O₄, and is identical to that presented in Fig. 1d.

On standing in air at 298 K, the XRD peaks characteristic

for Ag₃O₄were progressively replaced by those typical for

monoclinic Ag₂O₂(Fig. 8a–f). Complete decomposition of

Ag₃O₄was achieved after about 20 h.

3.5. Thermal decomposition of Ag₃O₄in air under dynamic

heating conditions

The DSC trace for Ag₃O₄(Fig. 4b) shows a large exo-

therm at 375 K, several weak exotherms at 398 K and

458 K, and a large endotherm at 704 K. These peaks are

assigned as follows. The endotherm at 375 K results from

thermal decomposition of Ag₃O₄(Eq. (10)); the weak

endotherm at 398 K is due to crystallization of monoclinic

Ag₂O₂; the endotherm at 458 K is due to thermal decompo-

sition of Ag₂O₂(Eq. (11)); and the strong exotherm at

704 K reﬂects thermal decomposition of Ag₂O to its ele-

ments (Eq. (12)).

Fig. 9a shows the far-IR spectrum of Ag₃O₄powder dis-

persed in a low-density polyethylene disk. The spectrum

contains no discernable bands and was similar to that

reported in Fig. 6a for metallic Ag₇O₈NO₃. Similarly, the

Raman spectrum collected at low laser power for Ag₃O₄

Ag₃O_4ðsÞ! 1:5Ag₂O_2ðsÞþ 0:5O_2ðgÞ

ð10Þ

ð11Þ

ð12Þ

1:5Ag₂O_2ðsÞ! 1:5Ag₂O_ðsÞþ 0:75O_2ðgÞ

1:5Ag O ! 3Ag_ðsÞþ 0:75O_2ðgÞ

ðsÞ

2

Ag₃O₄

(a)

(b)

(a)

(b)

(c)

(d)

(c)

(d)

(e)

(f)

Ag₂O₂

(f)

Ag₂O₂

x 0.05

(g)

0

100 200 300 400 500 600 700

Wavenumbers (cm^-1)

200

300

400

500

600

700

Raman Shift (cm^-1)

Fig. 9. FT-IR absorbance spectra for the thermal decomposition of

Ag₃O₄in air at 298 K. Time after preparation was (a) 0.5 h; (b) 4 h; (c) 8 h;

(d) 12 h; (e) 16 h and (f) 20 h. The data was collected using the pressed

polyethylene disk sampling method. The spectrum of the polyethylene disk

was removed by subtraction.

Fig. 10. Raman spectra excited at 785 nm for the thermal decomposition

of Ag₃O₄in air at 298 K. Time after preparation was (a) 0.5 h; (b) 4 h; (c)

8 h; (d) 12 h; (e) 16 h and (f) 20 h. The laser power at the sample for all

measurements was ꢁ0.3 mW. Fig. 10g was taken from the same sample as

Fig. 10f, at a laser power of 5 mW.

G.I.N. Waterhouse et al. / Polyhedron 26 (2007) 3310–3322

3321

(Fig. 10a) contained no bands. The fact that no discrete

vibrational bands are observed in the FT-IR or Raman

spectra for Ag₃O₄, even though some are expected based

on a factor group analysis of the oxide’s crystallographic

structure, is clear proof that Ag₃O₄behaves as a metal at

298 K. From the metallic properties it follows that one

electron per Ag₃O₄formula unit is delocalized over the

Ag₃O₄lattice, and the correct valence formulation for this

compound is (Ag^2.67+)₃O₄rather than Ag²⁺(Ag³⁺)₂O₄. The

observation that Ag₃O₄behaves as a metal is not un-

expected, given that the unit cell formula for the oxide

Ag₆O₈is identical to that of the face-sharing Ag₆O₈poly-

hedral units responsible for the metallic conductivity of

Ag₇O₈X salts. The grey colour and metallic lustre of both

Ag₃O₄and the Ag₇O₈X salts is further evidence of similar-

ities in the electronic structure of these compounds.

With increasing time at 298 K, the FT-IR and Raman

spectra characteristic for Ag₃O₄are slowly replaced by

those of monoclinic Ag₂O₂(Figs. 9a–f and 10a–f, respec-

tively). Fig. 10f and g was taken sequentially from the same

sample, the latter at a higher laser power and longer acqui-

sition time to achieve a better signal-to-noise ratio. The

value of using XRD and vibrational spectroscopic methods

in combination to characterize silver compounds and fol-

low their thermal decomposition is highlighted by this

study.

and Nanotechnology for ﬁnancial support given to this

project.

References

[1] A.P. Karpinski, S.J. Russell, J.R. Serenyi, J.P. Murphy, J. Power

Sources 91 (2000) 77.

[2] D.F. Smith, J.A. Gucinski, J. Power Sources 80 (1999) 66.

[3] D.F. Smith, C. Brown, J. Power Sources 96 (2001) 121.

[4] W. Beesk, P.G. Jones, H. Rumpel, E. Schwarzmann, G.M. Sheldrick,

J. Chem. Soc., Chem. Commun. (1981) 664.

[5] H.E. Swanson, Standard X-ray Diﬀraction Powder Patterns, Natl.

Bur. Stand. Monograph 25, section 1, 45 (1962).

[6] J.A. McMillan, J. Inorg. Nucl. Chem. 13 (1960) 28.

[7] J.A. McMillan, Chem. Rev. 62 (1962) 65.

[8] V. Scatturin, P.L. Bellon, A.J. Salkind, J. Electrochem. Soc. 108

(1961) 819.

[9] K. Yvon, A. Bezinge, P. Tissot, P. Fischer, J. Solid State Chem. 65

(1986) 225.

[10] B. Standke, M. Jansen, Angew. Chem., Int. Ed. Engl. 25 (1986)

77.

[11] B. Standke, M. Jansen, J. Solid State Chem. 67 (1987) 278.

[12] B. Standke, M. Jansen, Angew. Chem., Int. Ed. Engl. 24 (1985) 118.

[13] A.N. Mansour, J. Phys. Chem. 94 (1990) 1006.

[14] S.S. Kabalkina, S.V. Popova, R.N. Serebyanaya, L.F. Verschchagin,

Dokl. Akad. Nauk. SSSR 152 (1963) 853.

[15] L.H. Tjeng, M.B.J. Meinders, J. van Elp, J. Ghijsen, G.A. Sawatzky,

Phys. Rev. B 41 (2001) 3190.

[16] J.S. Hammond, S.W. Gaarenstroom, N. Winograd, Anal. Chem. 47

(1975) 2193.

[17] S.W. Gaarenstroom, N. Winograd, J. Chem. Phys. 67 (1977) 3500.

[18] J.F. Weaver, G.B. Hoﬂund, J. Phys. Chem. 98 (1994) 8519.

[19] G.I.N. Waterhouse, G.A. Bowmaker, J.B. Metson, Appl. Surf. Sci.

183 (2001) 191.

[20] G.I.N. Waterhouse, G.A. Bowmaker, J.B. Metson, Surf. Interface

Anal. 33 (2002) 401.

[21] G.I.N. Waterhouse, G.A. Bowmaker, J.B. Metson, Phys. Chem.

Chem. Phys. 3 (2001) 3838.

[22] P. Behrens, S. Aßmann, U. Bilow, C. Linke, M. Jansen, Z. Anorg.

Allg. Chem. 625 (1999) 111.

4. Conclusions

(1) Ozonolysis of neutral or mildly acidic solutions of

Ag(I) represents a simple and eﬀective route for the

sythesis of Ag₂O₂, Ag₃O₄and Ag₇O₈X salts

ðX ¼ NO₃^ꢀ; ClO₄^ꢀ; HSO₄^ꢀÞ. Previously, Ag₃O₄and

Ag₇O₈X compounds have only been accessible by

electrochemical routes. Attempts to prepare Ag₂O₃

were unsuccessful.

[23] L. van Wu¨llen, S. Vensky, W. Hoﬀbauer, M. Jansen, Solid State Sci. 7

(2005) 920.

[24] D.R. Lide (Ed.), CRC Handbook of Chemistry and Physics, 84th ed.,

CRC press, London, 2003.

(2) FT-IR and Raman spectra were reported for the ﬁrst

time for Ag₃O₄and Ag₇O₈NO₃. The vibrational spec-

tra indicate these compounds are metallic, a result

also conﬁrmed by XPS for the Ag₇O₈NO₃. Based

on structural and spectroscopic considerations, the

electron conﬁgurations of these compounds can be

[25] M. O’Keefe, J. Solid State Chem. 85 (1990) 108.

´

[26] I. Naray-Szabo, K. Popp, Z. Anorg. Allg. Chem. 322 (1963) 286.

´

[27] I. Naray-Szabo, G. Argay, P. Szabo, Acta Crystallogr. 19 (1965) 180.

[28] M.B. Robin, K. Andres, T.H. Geballe, N.A. Kuebler, D.B. McWhan,

Phys. Rev. Lett. 17 (1966) 917.

formulated as (Ag^2.67+)₃O₄and ðAg^2:67þÞ O₈ꢂ

6

Ag^þNO₃

.

[29] A.C. Gossard, D.K. Hindermann, M.B. Robin, N.A. Kuebler, T.H.

Geballe, J. Am. Chem. Soc. 89 (1967) 7121.

ꢀ

(3) The thermal decomposition of Ag₃O₄and Ag₇O₈NO₃

powders in air was examined by DSC, XRD, FT-IR

and Raman spectroscopy. The thermal stabilities of

compounds prepared by ozonolysis were comparable

to samples prepared by traditional electrochemical

routes.

[30] M.B. Robin, P. Day, Adv. Inorg. Chem. Radiochem. 10 (1967) 247.

[31] C.H. Wong, T.H. Lu, C.N. Chen, T.J. Lee, J. Inorg. Nucl. Chem. 34

(1972) 3253.

[32] M. Jansen, S. Vensky, Z. Naturforsch. 55b (2000) 882.

[33] J.C. Jack, T. Kennedy, J. Therm. Anal. 3 (1971) 25.

[34] B. Pettinger, X. Bao, I. Wilcock, M. Muhler, R. Schlo¨gl, G. Ertl,

Angew. Chem., Int. Ed. Engl. 33 (1994) 85.

[35] X. Bao, M. Muhler, Th. Schedel-Niedrig, R. Schlo¨gl, Phys. Rev. B 54

(1996) 2249.

Acknowledgement

´

[36] S. Djokic, J. Electrochem. Soc. 151 (2004) C359.

[37] R. Dallenbach, J. Painot, P. Tissot, Polyhedron 1 (1982) 183.

[38] R.O. Suzuki, T. Ogawa, K. Ono, J. Am. Ceram. Soc. 82 (1999) 2033.

[39] J. Selbin, M. Usategui, J. Inorg. Nucl. Chem. 20 (1961) 91.

[40] J.A. Allen, P.H. Scaife, Aust. J. Chem. 19 (1966) 715.

[41] R.N. Hammer, J. Kleinberg, Inorg. Synth. 4 (1953) 12.

GINW thanks the Foundation for Research, Science

and Technology for the award of a NZ Science and Tech-

nology Postdoctoral Fellowship (Contract # UOAX0412)

and the McDiarmid Institute for Advanced Materials

3322

G.I.N. Waterhouse et al. / Polyhedron 26 (2007) 3310–3322

[42] Jcpds-International Centre for Diﬀraction Data, Ag₇O₈No₃, card 73-

[47] Jcpds-International Centre for Diﬀraction Data, Ag₃O₄, card 84-

1261.

1327.

[43] Jcpds-International Centre for Diﬀraction Data, Ag₂O₂, card 84-

1547.

[44] Jcpds-International Centre for Diﬀraction Data, AgNO₃, card 84-

0713.

[45] Jcpds-International Centre for Diﬀraction Data, Ag₂O, card 75-1532.

[46] Jcpds-International Centre for Diﬀraction Data, Ag, card 04-0783.

[48] A.A. Audi, P.M.A. Sherwood, Surf. Interface Anal. 29 (2000) 265.

[49] Z.X. Shen, M.H. Kuok, S.H. Tang, W.F. Sherman, Spectrochim.

Acta, Part A 48 (1992) 1317.

[50] J.R. Ferraro, A. Walker, J. Chem. Phys. 42 (1965) 1273.

[51] F. Paulik, J. Paulik, M. Arnold, Thermochim. Acta 92 (1985)

787.

Article Doi

Synthesis, vibrational spectra and thermal stability of Ag3O4 and related Ag7O8X salts (X = NO3-, ClO4-, HSO4-)

DOI: 10.1016/j.poly.2007.03.006

Source and publish data:

Authors:

Article abstract of DOI:10.1016/j.poly.2007.03.006

Full text of DOI:10.1016/j.poly.2007.03.006

Products guided by the article

R&D Labs maybe for 7440-22-4

Relevant to this article

Hot Product