Roberto Moretti

Seconda Università di Napoli Dip. Ingegneria Civile

Via Roma 29 - 81031 Aversa (I)

INGV - Osservatorio Vesuviano Via Diocleziano 328 - 80124 Napoli (I)

Redox in melts and glasses… an introduction ?

Gaillard & Behrens, 2006

Mysen and Richet 2005

!???!

Mysen and Richet 2005

……..

Empirical efforts….

S2- + 2O2 SO42-

44 BaSO

glass

FeSBaSO

FeSglass24

SKSK

SKSKSKSK

nStotnSO

αλΔ

αλΔ=

αλ−αλ

αλ−αλ=

−

[ ][ ]

log . log /S

SOfO T

2

42 2

1 02 25410 10−

−= − − +

Locus of points from Wallace and Carmichael (1994)

Dealing with redox couples: an example from lte literature

Log K = -1.97 logfO2R2 = 0.98

0

5

10

15

20

25

30

35

40

-20 -15 -10 -5 0log fO2 (bars)

Ther

mod

ynam

ic e

quili

briu

m K

12

(log 1

0, ad

imen

sion

al)

Log K = 35078/T - 7.6021

R2 = 0.67

5

15

25

35

0.0005 0.00075 0.001

1/T (K-1)

log

K12

K = [SO42-]/[S2-]fO22

Moretti and Ottonello (2003), J. Non-Cryst. Sol.

X

=> Good compatibility between different techniques!

0

0.2

0.4

0.6

0.8

1

MössbauerXANESRaman

0 0.2 0.4 0.6 0.8 1

Fe3+

/ !Fe

Fe3+/ !Fe (wet chemical analyses)

Magnien V., Neuville D.R., Cormier L., Mysen B.O. and Richet P. (2004) Kinetics of iron oxidation in silicate melts: A preliminary XANES study. Chem. Geol., 213, 253-263.

Galoisy 2006

The questions…

-How do we write chemical reactions for silicate melts

to account for COMPOSITIONAL dependences?

-Which ‘syntax’ do we use?

-Is the chemical syntax for describing chemical

exchanges directly available from structural studies?

Or is just an (useful) approximation of some major

features?

Intact solvation shells Partial disruption of solvation shells

Disruption of solvation shells

Outer sphere ion pair Inner sphere ion pair

(complex)

Mysen and Richet, 2005

Depolymerization of Silicate Melts to accommodate volatiles (e.g. water)

Some observations…

We usually do not perceive the problem of the chemical

syntax as long as:

-we interpret our own data and a few more

Thus we tend to adopt the detected structural

scenario and turn it into the chemical syntax

Some observations…

-  But what happens if we want to find the general chemical mechanism and set models? (….the problem of many data from many compositions…)

-  How “convert“ structural findings into speciation hypotheses for general chemical mechanisms?

Acid-base exchanges The dividing line between the Lewis acids and bases is not sharp a one, and its theoretical interpretation is obscure  Strong role of the molecular structure (hence, bulk composition)  Hard and soft categories…(atomic structures associated with hard acids and bases are rigid and impenetrable, whereas those associated with soft acids and bases are more readily deformable)

...it is not practical if we deal with chemical reactions

For example…

This acid-base and redox exchange synthesizes many (combined) reactions:

SO2(g) + 1/2O2(g) + CaO(m) CaSO4(m) + SO2(g) + 1/2O2(g) + H2O(m) H2SO4(m) +

SO2(g) + 1/2O2(g) + 1/2SiO2(m) 1/2Si(SO4)2(m) +

… … +

… … +

… … …

Marini, Moretti ,Accornero 2011

So we need ions… •  “The formation of ions sensu stricto in liquid water is due to

the high dielectric constant of aqueous medium (efficient shielding of charge) as also manifested by the formation of hydration spheres.

•  In contrast, none of these concepts is applicable to silicate melts. The dielectric properties are different, no volumetric electrostriction is known, electrostatic bond valences are generally greater to allow for such effects.

•  The only and apparent similarity arises from adoption of analogous symbolic notation for melt species that uses negative charge, therefore, the thermodynamic identities look like ions”

•  the connectivity of the silicate structure is such that cations and anions have actual charges lower than formal ones, and the residual charge distribution from bonding of bridging oxygen to silicon allows oxygen bonding with other cations.

•  This makes the relative contribution of nonbridging and bridging oxygens to the oxygen coordination of the other cations poorly known (Mysen and Richet, 2005) and highlights the impossibility to readily distinguish solute and solvent like in aqueous solutions.

•  The anionic framework of silicate melts, in fact, makes solute and solvents so intimately related that one cannot identify a solvation shell and identify directly, from structural studies, the complexes needed to define acid-base reactions.

So, we need ions to write chemical reactions…but…

2O- O0 + O2- Vs. 2Qn Qn-1 + Qn+1

(Polymerization Vs. Connectivity)

…to understand what determines acid-base behavior and to write and inspect chemical reactions we have to assume an understanding of the bonding, structure, and properties of individual molecules also in melts (with its pros and cons…)

Nevertheless…

The “thermochemical knowledge” of a melt system does not seem to require the microstructural

“complexity” that can be revealed by many spectroscopic investigations: the structural

“characterization” exceeding that required for the description of acid-base properties (e.g, in the

Lux-Flood notation) may be not useful.

How deep need we to go with connections between structure and chemical thermodynamics ?

In (essentially aprotic) silicate melts acid-base properties are expressed in terms of Lux-Flood formalism: Acid + O2- Base (1) “Reaction” 1 recalls the Bronsted-Lowry formalism for aqueous solutions: Acid Base + H+ (2)

Which acid-base ‘syntax’ to describe reactivity in melts?

Which redox ‘syntax’ to describe reactivity in melts?

In aqueous solutions the electrode of reference is the “normal hydrogen electrode”, whereas in silicate melts

the reference electrode is the “normal oxygen electrode” ,i.e.:

1/2O2 + 2e- O2- (3)

Even if the main redox couple in oxide melts is given by iron, reaction (3) follows the syntax based on the O2-

exchange

O2- ? Virtual or real ?

Calculated from data in Park and Rhee (2001)

Nesbitt et al. (2011)

Free-oxygen = oxide ion…attached to some metal cation

Polymeric nature of anion matrix: Toop-Samis and Masson models

In polymeric models for silicate melts, it is postulated that, at each composition, for given P-T values, the melt is characterized by an equilibrium distribution of several ionic species of oxygen, metal cations and ionic polymers of monomeric units SiO44-.

−− +⇔ 2m0

mm OO2O

−−−− +⇔+ 267244

44 OOSiSiOSiO

2

2

58 )(O))(O(OK

−

−°

=

polymmixing )/2]RTlnK[(OΔG−=

( ) ( )2ON4

O 2SiO0−−

=

( ) ( ) ( )2ON1O

2SiO2

−− −−=

( ) ( ) ( )( ) ( ) 01NN8N22O1K4O222 SiOSiOSiO2

2=−+++− −−

Kp=(O°)(O2-)/(O-)2

On the basis of simple mass balances we can link the three oxygen species to the melt composition and therefore to the melt compositions

Theory: the polymeric model

⎥⎥⎦

⎤

⎢⎢⎣

⎡−⎟⎟⎠

⎞⎜⎜⎝

⎛γ−γ×= +η+η+ν+ν ∑∑ 1445.1XX662.4expK

jjìi TjTM

iMpolym

Modifiers - Formers KP= exp(4.662Δγ-1.1445)-8

-7

-6

-5

-4

-3

-2

-1

0

0.5 1 1.5 2 2.5γΜν+

ln K

P,M

O-S

iO2

[16]

[17,18]

[24]

[25]

[26]

This study

( )polym.RTlnK2

OΔG mixing−

=

Ottonello et al., Chem. Geol. (2001)

Even simple interaction parameters along limiting binaries cannot be reduced to fitting coefficients of mathematical minimization routines, but must be formally linked to the intrinsic atomistic properties of the interacting ions and molecules

depolymerizing role of water has been overrated with respect to its actual acid-based properties in melts.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 0.2 0.4 0.6 0.8 1

X SiO2

(O- );

(O°)

; (O

2-)

(frac

tions

)

fO-

fO2-

fO°

T=1600°C ; P=1 barKP=0.0003 CaO-SiO2

Ottonello et al., 2001

Back to basics: amphoteric behavior of trivalent iron

melt2,2melt

-2melt

3melt

melt2,-2melt

2melt

-3melt

O41FeO

21Fe

_______________________________

O41O

21

FeFe

+⇔+

=+⇔

+⇔+

++

−

++

e

e

melt2,-2

melt2meltmelt2,

melt2,melt2,

-2melt

2meltmelt2,

O41 O

23 FeFeO

_________________________________

O41 O

21

2O FeFeO

++⇔

=+⇔

++⇔+

+−

−−

+−−

e

e

That’s the way electron transfer operates. It is not evident from the equilibrium involving macroscopic components: to get that you must adopt the ionic notation.

(From the Fraser’s approach to Europium)

FeO + 1/4O2 FeO1.5

In the above reaction the redox potential is expressed by the oxygen fugacity . But how does the electron transfer take place?

5.12 FeOO41FeO ⇔+ 1)

−− ⇔+)melt()melt()melt( 2

232 FeO2OOFe 2)

Fe O Fe Omelt melt melt

2 33 22 3

( ) ( ) ( )⇔ ++ − 3)

FeO Fe Omelt melt melt( ) ( ) ( )

⇔ ++ −2 2 4)

Adopting The ionic model of Fused Salts after Temkin (1945):

[ ][ ] [ ]∑ ∑

∑+−

+

+==⎟⎟

⎠

⎞⎜⎜⎝

⎛

cationsFeanionsFeOcationsFe

nn

FeFe

32

2

Fe

FeIII

II

III

I1

∑ ∑∑+

×=⎟⎟⎠

⎞⎜⎜⎝

⎛

−

−

cationsKanionsaKcationsKa

KFeFe

O

O

OIII

II

213

2212

421

411 2

2

2f1

Ratio of activities (aFeO/aFeO1.5)

Ratio of activity coefficients (γFeO1.5/γFeO)

Iron redox model

The message:

1.   Silicate melts are polymerized liquids 2.   Polymeric units are highly reactive

Polymerization and redox state are intimately interrelated. This melt affects properties, including oxidation state, volatile solubility et cetera.

Fig. 4.12: Computed versus measured activity of ferrous oxide components in melts equilibrated with metallic iron at various T anf fO2 conditions and 1 bar pressure.

Fig. 4.10: Thermodybamic activity of stoichiometric ferrous oxide FeO in equilibrium with pure iron metal at T=1600K and various aO2.

Ottonello et al., 2001

ο This work eqn 4.61

Fig. 4.14: fO2 estimates based on thermochemical model (eqn. 4.61) (4.13) and those based on Sack equation 4.86 (b). The dataset is the same used for model calibration, in both cases.

Ottonello et al., 2001

Fig. 4.15: Comparison between fO2 estimates based on thermochemical model and estimates of the Kilinc equation. Dataset is the same adopted in constraining iron equilibria in molten systems.

Ottonello et al., 2001

Fig. 4.13: Iron redox ratio in quenched melts and glasses equilibrated with a fO2 buffered atmosphere at various T conditions.

Ottonello et al., 2001

Fig. 4.20: Quantile-quantile representation of the observed Vs calculated FeO and Fe2O3 wt% amounts in the two subsets.

calc

ulat

ed

measured

Fig. 4.19: Statistics of residuals for the Mossbauer and wet chemistry data subsets. The distribution of residuals for wet chemistry subset (444 samples) is normal, centered on x= 0.0003. The distribution of residuals for the Mossbauer dataset (100 samples) is non-normal, centered on x = -0.2254.

Ottonello et al., 2001

Let us introduce the amphoteric behavior of water, i.e. an acidic dissociation: H2O + O2- 2OH-

Which can be coupled to the basic dissociation (see Fraser): 2H+ + O2- H2O

By subtracting: H+ + O2- OH-

∑−+

−

− −⋅=

cationsnn

nn

K OHTOTH

OH

O2OHOH

INH 2n2nn ⋅=+ −+

Moretti (2005, AoG)

At parity of water content, high free hydroxyl concentration is expected in more basic melts, in

agreement with Xue and Kanzaki (2004), Behrens et al. (2004) for water diffusivities etc…

Iron redox ratio in hydrous silicate melts is also origin of controversies. Current hypotheses about the role of water on the FeIII/FeII value are:

-Water does not affect ferric to ferrous ratio. -Water causes a decrease of the ferric to ferrous ratio. -Water causes an increase of the ferric to ferrous ratio.

==> Parameterization of the ferric to ferrous ratio should consider the “impact” of water on melt acid-base properties and then polymerization together with the effect of pressure on the previous reactions.

New experimental evidences

Botcharnikov et al. (GCA, 2005)

QFM + const redox buffer holding on each line

Model comparison

Botcharnikov et al. (2005)

Fe/ F

eII

I II

H O (wt

%)

2

0.2

0.6

1

1.41.8

2.22.63

About the effect of water on the iron oxidation state of melts…

Redox models simply relating Fe2+/Fe3+ to fO2 are not applicable !

Moretti (in prep.)

Iron redox model: unexpected features

We can explain, for example, the increase of oxidation with decreasing T. This depends on oxygen fugacity, that is, on how fO2 varies with T. We propose that under constant fO2 , or more precisely for

a decrease of T can cause iron oxidation. 22000

12 −≈>

⎟⎠

⎞⎜⎝

⎛∂

∂

T

fO(depending on composition…)

-1.5

-1

-0.5

0

0.5

1

1.5

2

2.5

3

3.5

400 600 800 1000 1200 1400 1600 1800

T (K)

log fO2= -8

P = 2000 barH2O = 2.5 wt%

log fO2= -17000/T + 2

NNO

QFM

Log F

eIII/F

eII

Métrich et al., 2006

S as sulphide S as sulphate

Sulfide equilibrium: (L)S(G)O

21(G)S

21(L)O 222

-2 −+⇔+

Sulfate equilibrium:

(L)SO(G)O23(G)S

21(L)O 2422

-2 −⇔++

S as sulphide S as sulphate

S solubility: the Conjugated-Toop-Samis (CTSFG) model (Moretti and Ottonello, 2005)

0

0.5

1

1.5

2

2.5

3

-4 -3 -2 -1 0 1 2 3 4ΔQFM

log

S(pp

m)

Model calculation

Katsura andNagashima (1974)

Katsura and Nagashima (1974), Tholeiite, 1250°C

SO2 (i) = 0.21%

Slope = -1.5

Slope = 0.5

for QFM > -1

Moretti and Ottonello (2005 GCA)

Sulfide equilibrium: (L)S(G)O

21(G)S

21(L)O 222

-2 −+⇔+

Sulfate equilibrium:

(L)SO(G)O23(G)S

21(L)O 2422

-2 −⇔++

CaO-SiO2

-7

-6

-5

-4

-3

-2

-1

0 0.2 0.4 0.6 0.8 1CaO (molar fraction)

log

Cs

(wt%

)

1650 °C

1500 °C

MnO - SiO2

-6

-5

-4

-3

-2

-1

0

0 0.2 0.4 0.6 0.8 1MnO (molar fraction)

log

Cs

(wt.

%)

1500 °C

1650°C °C

MnO - Al2O3

-2.5

-2

-1.5

-1

-0.5

0

0 0.2 0.4 0.6 0.8 1MnO (molar fraction)

log

Cs

(wt.%

)

1650 °C

Moretti and Ottonello, 2003

(L)MSO(G)O23)(

21MO(L) 422 ⇔++ GS

⎟⎟⎠

⎞⎜⎜⎝

⎛−

Δ−+==

−− )1(exp''

''21

23

22

4

4P

RTv

Tbaff

aa

k LiiSOOM

SOMSOO-MM

i

i

ii

Sulfate-metal oxide reactions:

(G)O21MS(L))(

21MO(L) 22 +⇔+ GS

⎟⎟⎠

⎞⎜⎜⎝

⎛−

Δ−+=⎟

⎟⎠

⎞⎜⎜⎝

⎛= )1(exp

''

21

2

2 PRTv

Tba

ff

aa

k LiiS

O

OM

SMSO-MM

i

i

ii

Sulfide-metal oxide reactions:

iSOSMOanneal

N

iiSOS kCNk

OX

),,(1

, 2422

42 lnln −−

+

−− −=∑=

ν

Gas-liquid reactions: Flood and Grjotheim (1952) thermochemical cycle:

∑=

+

++ =

OXN

iii

iii

n

nN

1ν

νν

where Niν+ represent electrically equivalent fractions, and ni are mole fractions computed over the appropriate matrix, either cationic or anionic (Temkin model of fused salts)

Theory of the S solubility model (CTSFG)

−− +⇔ 20 OO2OFincham and Richardson (1954):

O- are singly bonded (or non-bridging) oxygens

•  O0 are doubly bonded (or bridging) oxygens

•  O2- are free oxygens The Toop-Samis (extended) model is applied

[ ]21

%2

22 ⎟

⎟⎠

⎞⎜⎜⎝

⎛=−

S

OwtS f

fSC

[ ] 21S23

O%wtSO 2224ffSC

−−=−

The model computes Canneal. (entropies of annealing). kMO-(S2-,SO42-) constants are from independent thermodynamic compilations when available.

SiO2 Al2O3

CaO

25

50

75

25 50 75

25

50

75

-5.18-5.15

-3.52-3.84

-4.81-4.98

-4.21

-5-5.5-4.5-4-3.5-3-2.5

-2.5-3.83

-4.1-4.33-4.58

-4.60

-3.7

-3.71

-4.04

-4.17-4.38-4.64 -4.12

-3.98-4.2

-2

b)

SiO2-Al2O3-CaO, T = 1500°C

SiO2 Al2O3

CaO

25

50

75

25 50 75

25

50

75

-4.43

-5.5-4.5-4-3.5-3-2.5

-2.5

a)

-2

-5

-2.28

-4.30-4.22

-4.10

-3.62-3.60

■

Moretti and Ottonello (2003), Metall. Mat. Trans. B

The S solubility (CTSFG) model: verification and check of the simplex and its features

Moretti and Ottonello (2005)

Simple metallurgical slags (log Cs contouring)

Natural-like melts (sulfur content)

0500

1000150020002500300035004000

-10 -9 -8 -7 -6 -5 -4logfO2 - logfS2

O'Neill and Mavrogenes(2002) - measured

Cape Vogel Andesite

d)present model

0

5001000

150020002500

30003500

4000

-10 -9 -8 -7 -6 -5 -4logfO2 - logfS2

O'Neill andMavrogenes (2002)- measured

MORB

b)present model

0

2000

4000

6000

8000

10000

12000

14000

-12 -11 -10 -9 -8 -7 -6 -5 -4logfO2 - logfS2

S (p

pm)

Buchanan and Nolan (1983)

present model

a)

0500

1000150020002500300035004000

-10 -9 -8 -7 -6 -5 -4logfO2 - logfS2

S (p

pm)

O'Neill andMavrogenes(2002) - measured

Olivin Tholeiite

c)present model

S solubility at P: considering volumes for systems in which they have never been determined !

( )∫ +−+=ʹ′ −−−−P

1gm

'MS,O

'MS,O

TPMS,O dPΔVΔVRT

1TBAKln

1-bar lnK

Our value for the (partial) molar volume of VSO3 is 57.71 cm3/mol at 1673K… must be checked

experimentally (no data at present) 3

υ24

υ2 SOOMSOM VVV +=

++

/mol)(cm41VVΔV 3SMSOMS-SOυ24

υ224

=−=++

Moretti and Ottonello (2005)

The CTSFG solubility model: features of the sulfide and sulfate capacity surfaces

Kilauean Tholeite

-6.5-6.25

-6-5.75-5.5

-5.25-5

-4.75-4.5

-4.25-4

-3.75

0 1 2 3 4 5 6 7 8 9 10H2O (wt%)

log

CS

1050°C

1100°C

1150°C1200°C

5 kbar

4 kbar

3 kbar

2 kbar

1 kbar

0.5 kbar

1 bar

0.1kbar

a)

Arc Basalt

-7

-6.5

-6

-5.5

-5

-4.5

-4

0 1 2 3 4 5 6 7 8 9 10 11H2O (wt%)

log

CS

1000°C

1050°C

1150°C1100°C

1 bar

0.5 kbar2 kbar

1 kbar

3 kbar

4 kbar

5 kbar0.1 kbar

b)

Andesite

-9

-8.5

-8

-7.5

-7

-6.5

-6

-5.5

-5

0 1 2 3 4 5 6 7 8 9 10H2O (wt%)

log

CS

850°C900°C

1000°C

950°C

1 bar

0.5 kbar

1 kbar

2 kbar

3 kbar

4 kbar

5 kbar

0.1 kbar

c)

Rhyolite

-14

-13

-12

-11

-10

-9

-8

-7

-6

0 1 2 3 4 5 6 7 8 9 10H2O (wt%)

log

CS

800°C850°C

900°C

1 bar

0.5 kbar 1 kbar

2 kbar 3 kbar

4 kbar

5 kbar

0.1 kbar

750°C

700°Cd)

Kilauean Tholeite

88.5

99.510

10.511

11.512

12.513

13.514

14.5

0 1 2 3 4 5 6 7 8 9 10H2O (wt%)

log

CSO

4

1050°C

1100°C

1150°C

1200°C

3 kbar2 kbar

4 kbar

1 kbar

0.5 kbar1 bar

5 kbar

a)

0.1 kbarArc Basalt

8

9

10

11

12

13

14

15

16

0 1 2 3 4 5 6 7 8 9 10 11H2O (wt%)

log

CSO

4

1150°C

1100°C

1050°C

1000°C

1 bar 0.5 kbar

1 kbar

2 kbar3 kbar

4 kbar5 kbarb)

0.1 kbar

Andesite

11

12

13

14

15

16

17

18

19

0 1 2 3 4 5 6 7 8 9 10H2O (wt%)

log

CSO

4

850°C

900°C

950°C

1000°C

1 bar

0.5 kbar

1 kbar

2 kbar

3 kbar

4 kbar

5 kbar

0.1 kbar

c)

Rhyolite

11

13

15

17

19

21

23

0 1 2 3 4 5 6 7 8 9 10H2O (wt%)

log

CSO

4

800°C

850°C

900°C

1 bar

0.5 kbar1 kbar

2 kbar3 kbar

4 kbar

5 kbar

d)

0.1 kbar

750°C

700°C

[ ] [ ] [ ]21

2

21

2S

21

223

2Ssulfatewt%,sulfidewt%,totwt%, fOfSCfSfOCSSS 26 −+ +=+=

Lesne, 2008

The CTSFG model for sulfur: bulk performance and suggestions for experimentalists

Limits - Most data are at 1 bar ! => more data at pressure are needed

- Few data for sulfate solubility (high fO2) even at 1 bar !

- Lack of experimental data on partial molar volumes of sulfide and sulfate liquid species - Need of fS2 probes for high-P pressure vessel experiments

-12

-10

-8

-6

-4

-2

0

-12 -10 -8 -6 -4 -2 0

log Csulfide (experimental)

log

Csu

lfide

(cal

cula

ted)

Anhydrous dataset (P = 1 bar)1081 compositions

Hydrous dataset (via Papalemodel) 60 compositions

-20%

+20%

a)

R2 = 0.946 (1141 data)σy,x= 0.205

Annealing volumetric terms added2.5

5

7.5

10

12.5

15

17.5

20

2.5 5 7.5 10 12.5 15 17.5 20

log Csulfate (experimental)

log

Csu

lfate

(cal

cula

ted)

Anhydrous dataset (P = 1 bar)172 data

Hydrous dataset (via Papalemodel) 49 compositions

e)

-10%

+10%

R2 = 0.990 (221 data)σy,x= 0.278

Moretti and Ottonello (2005)

Lehmann and Nadif, 2011

Are complex approaches really needed for data interpretation?

Scaillet and Pichavant (2003) Geol. Soc. Spec. Publ. 213 Moretti et al. (2003) Geol. Soc. Spec. Publ. 213

Mass partitioning in Nature is non-linear. The expected linearity required by theory (stoichiometry in this case) is embodied, not

cancelled. “Complex” models may reveal it.

Tholeite, 1400K . H2Otot=3%wt; CO2tot=1%wt; Stot=0.5%wt

Moretti et al. (2003)

Baker & Moretti (2011)

Baker & Moretti (2008)

STRONG ROLE OF COMPOSITION (e.g. H2O content) ON SULFUR SPECIATION !

Never forget this!

... the S2- + 2O2 SO42- equilibrium computed by the CTSFG model is then applied to an independent set of data:

-8

-6

-4

-2

0

2

4

6

8

−8 −6 −4 −2 0 2 4 6 8

ΔQFM (experimental or estimated)

ΔQ

FM (c

alcu

late

d)

Model calculation

Wallace andCarmichael(1994) equation

Validity range of ΔλSKα technique

in melts.

a)

-8

-6

-4

-2

0

2

4

6

8

−8 −6 −4 −2 0 2 4 6 8

ΔQFM (experimental or estimated)

ΔQ

FM (c

alcu

late

d)

Model calculation

Wallace andCarmichael(1994) equation

Validity range of ΔλSKα technique

in melts.

Volumetric annealing terms

included

b)

Application to S speciation and fO2 determination

Moretti and Ottonello 2005

The H2O-CO2-H2S-SO2 saturation model

Morett i and Papale (2004) Chemical Geology

Gas phase H 2 O, CO 2 , SO 2 , H 2 S

homogeneous equilibrium ( chemical reactions )

(SUPERFLUID – Belonoshko et al., 1992 )

Liquid phase H 2 O + CO 2 saturation

( Papale, 1999 )

Liquid phase mass balance

reactive and/or dissolution effects

Redox conditions

Gas phase H 2 O, CO 2 , SO 2 , H 2 S

homogeneous equilibrium ( chemical reactions )

(SUPERFLUID – Belonoshko et al., 1992 )

Liquid phase H 2 O + CO 2 saturation

( Papale, 1999 )

Liquid phase mass balance

reactive and/or dissolution effects

Redox conditions

Gas phase H 2 O, CO 2 , SO 2 , H 2 S

homogeneous equilibrium ( chemical reactions )

(SUPERFLUID – Belonoshko et al., 1992 )

Liquid phase H 2 O + CO 2 saturation

( Papale, 1999 )

Liquid phase mass balance

reactive and/or dissolution effects

Redox conditions

Gas phase H 2 O, CO 2 , SO 2 , H 2 S

homogeneous equilibrium ( chemical reactions )

(SUPERFLUID – Belonoshko et al., 1992 )

Liquid phase H 2 O + CO 2 saturation

( Papale, 1999 )

Liquid phase S saturation S saturation S saturation S

mass balance

reactive and/or dissolution effects

Redox conditions

(Moretti et al., 2003) al., 2003)

FeII/FeIII model (Moretti, 2004)

Theory of the revised and extended H2O-CO2 saturation model (Papale et al., 2006)

• Fully non-ideal • Fluid phase of any composition in the system H2O+CO2

• Liquid phase of any composition from two/three components to natural (12 components)

Equilibrium equations

oLCOCOCOCOCO

LCO

GCO

oLOHOHOHOHOH

LOH

GOH

LG

LG

fxPyff

fxPyffTTTPPP

2222222

2222222

γφ

γφ

=⇒=

=⇒=

==

==

22

22

22

22

221

COCO

COTCO

OHOH

OHTOH

COOH

xyxx

xyxx

yy

−

−=

−

−

=+

Mass balance equations

The revised H2O-CO2 saturation model (Papale et al., 2006)

Excess Gibbs free energy of the liquid: ijjn

i

n

iji

E wxxNG ∑∑−

= +=

=1

1 1

Water:

Carbon dioxide:

)0(

1

'2

2

22222222)1)(1()1(ln OiH

n

COiiCOOHOHOCOHCOOHOH wxxxxwxxRT ∑

=≠

−−−+−=γ

⎥⎦

⎤⎢⎣

⎡+−−− ∑∑

=≠=≠

n

OHiiCOio

n

OHiiCOiCOOHCO wxP

Pwxxxx1

)1('

1

)0('

2

2

2

2222ln)1(

)22()1(1

1, 1,

''2

22 22

22 ∑ ∑−

=≠ +=≠

−−−n

COOHi

n

iCOOHjijjiCOOH wxxxx

)0(

1

'2

2

22222222)1()1(ln OiH

n

COiiCOOHOHOCOHOHCOCO wxxxxwxxRT ∑

=≠

−−−−=γ

⎥⎦

⎤⎢⎣

⎡+−−−+ ∑∑

=≠=≠

n

OHiiCOio

n

OHiiCOiCOOHCO wxP

Pwxxxx1

)1('

1

)0('

2

2

2

2222ln)1)(1(

)23()1(1

1, 1,

''2

22 22

22 ∑ ∑−

=≠ +=≠

−−−n

COOHi

n

iCOOHjijjiCOOH wxxxx

Activity coefficients: inTPi

E

RTnG

ij

γln,,

=⎟⎟⎠

⎞⎜⎜⎝

⎛

∂

∂

≠

The model computes H2O(m)-oxides interaction parameters and CO2(m)-oxides interaction parameters. All other oxide-oxide interactions are from Ghiorso et al. (1983)

(Papale et al., 2006)

Tamic et al., 2001

Moretti et al. (2003) Geol. Soc. Spec. Publ. 213

Volatile mixed “solubility”

Papale and Moretti, work in progress

Single-step volatile separation

(applicable to glass inclusions) (applicable to the geochemical sensing of volcanoes)

40030020010000

1000

2000

3000

4000

5000

6000

pressure (MPa)

diss

olve

d S

(ppm

)

total S in the system (wTS) = 0.5 wt%

wTH2O = 2 wt%

wTH2O = 4 wt%

wTH2O = 6 wt%

wTH2O = 8 wt%

Tholeite 1400K

Moretti & Baker (2008)

STRONG ROLE OF COMPOSITION (e.g. H2O content) ON SULFUR SPECIATION !

Never forget this!

FeO(m) + ½ S2 FeS(m) + ½ O2 ; FeS(po/liq) FeS(m)

aFeO and aFeS from “built-in” polymeric modeling and Flood and Grjotheim thermochemical cycle

0

500

1000

1500

2000

2500

0 5000 10000 15000 20000P (bar)

SCSS

(ppm

)

Tholeiite, H2O = 0.5 wt% - Liquid FeS

1200°CNNO - 0.5

NNO -1

c)

NNO

Falcone et al. 2011

-2melt4,

2melt

-2melt

-2melt

3melt SOFe8O4S8Fe +⇔++

++

SO Fe O S FeOmelt melt melt melt melt42 2 2 2

28 12 8, ,− + − − −+ + ⇔ +

meltmeltmelt OFeSFeOSO ,3222

,4 48 +⇔+−−

If you write:

We can also study the mutual redox exchanges between more redox couples (when existing…) such as Fe, Ni, S and so on… But even when studying mutual interactions, O2- and therefore oxygen species via Toop-Samis equilibrium, intervene through the normal oxygen electrode (see Moretti and Ottonello, 2003 JNCS). For example, if you consider Fe and S redox equilibria you should write the following ionic equilibria:

so that O2- cancel out, you make a very huge mistake, since you are mixing different notations, confusing species with components and therefore mixing up standard states! Remind that the basis of redox reactions in melts is the “normal oxygen electrode”

Fe-S mutual interactions

-6-4-202468

1012

-2 -1.5 -1 -0.5 0 0.5

dry

2.5 H2O wt%

AndesiteP = 2 kbarT = 1050°C

NNO-3-2

-1

NNO

+1

+2

+3

Log FeIII/FeII

Log

SO42

- /S2-

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

-3 -2 -1 0 1 2 3ΔNNO

FeIII/Fetot DRY

FeIII/Fetot 2.5 wt%

H2O

S6+/Stot 2.5 wt%

S6+/Stot DRY

AndesiteP = 2 kbarT = 1050°C

0

0.2

0.4

0.6

0.8

1

1.2

600 800 1000 1200 1400 1600

T (K)

S6+/Stot dry

S6+/Stot 2.5 H2O wt%

AndesiteQFM bufferP = 2 kbar

FeIII/Fetot dry

FeIII/Fetot 2.5 H2O wt%

Fe-S mutual interactions

Remember, it is the following connection:

acid-base properties => polymerization

polymerization => redox state

redox state => solubilities

solubilities => acid-base properties, and so on…

That promotes our full understanding of the role of bulk composition on the partioning of

volatiles in magmatic systems