Redox processes in glass-forming melts

Redox processes in glass-forming melts

Journal of Non-Crystalline Solids 84 (1986) 129-141 North-Holland, Amsterdam 129 Section HI. Properties / behavior of glass melts REDOX P R O C E S...

636KB Sizes 1 Downloads 31 Views

Journal of Non-Crystalline Solids 84 (1986) 129-141 North-Holland, Amsterdam

129

Section HI. Properties / behavior of glass melts

REDOX P R O C E S S E S IN GLASS-FORMING MELTS H e n r y D. S C H R E I B E R Center for Glass Chemistry, Virginia Military Institute, Lexington, VA 24450, USA

Oxidation-reduction (redox) equilibria of multivalent elements in glass-forming melts can be expressed by several different, yet similar, chemical equations. Each equation attempts to describe the effects that melt temperature, base composition, imposed oxygen fugacity, multivalent element concentration, and the presence of other redox components induce on the established redox equilibrium. Redox kinetics, or the rate at which equilibrium is attained in the melt, demonstrates the interrelationship between the dissolved multivalent species and the gaseous redox components. The diffusion of redox gases, such as oxygen into or out of a static glass-forming melt controls how fast the redox state of the melt can change. Both equilibrium and kinetic redox processes are thus integral to the development of desired glass/melt properties.

1. Introduction Redox processes can be simply defined as those reactions that involve the transfer of electrons from one species to another. Such processes have always played a fundamental role in the production of glass, in particular in the manufacture of c o l o r e d / d e c o l o r e d glass and in the synthesis of h o m o g e n e o u s bubble-free glass. In fact, ancient civilizations realized the importance of controlling redox conditions in their primitive glass recipes in order to achieve their desired end products [1]. The understanding of redox processes in glass-forming melts has wide applicability not only in the glass industry but also in other disciplines. Some of these pertinent areas are compiled in table 1. In order to obtain desirable glass/melt characteristics, the redox processes in the melt have to be carefully controlled [2,3]. One frontier of c o n t e m p o r a r y glass science is the production of "high-tech" glasses with novel electrical, magnetic a n d / o r optical properties [4], but once again the formation of such glasses may depend on the stabilization of certain redox species in the glass-forming melts.

2. Objective The fundamental chemistry of redox processes in glass-forming melts will be examined from both equilibrium and kinetic points of view so that the parameters that control redox properties can be applied to an understanding 0022-3093/86/$03.50 © Elsevier Science Publishers B.V. (North-Holland Physics Publishing Division)

130

H.D. Schreiber / Redox processes in glass-forming melts

Table 1 A sampling of areas in which redox processes in glass-forming melts contribute to the understanding of that property A. Glass production Melt properties gas bubbles, homogeneity refining via redox additives; gas solubility thermal conductivity foaming, oversaturation viscosity and foaming crystallization, evaporation of volatile components corrosion of refractories surface tension, wetting, adhesion of glass-metal junctions Glass properties electrical, semiconduction magnetic stability with respect to corrosion, chemical resistance, devitrification photosensitivity photochromic, solarization, radiation sensitivity color, decolorization light-absorbing properties (optical filters) laser glasses - for stimulated light emission B. Nuclear waste immobilization Favorable melt/glass properties with respect to "exotic" redox elements Solubility, leach resistance Hot isostatic pressing of melts C. Materials science Glass ceramics Electrolytes in melts to isolate metals D. Slag recycling Distribution of species between immiscible melts E. Melt chemistry Probes for melt acidity/basicity Probes for melt structure F. Geochemistry Distribution of elements in planets understanding of magmatic processes comparative planetology Natural glasses tektites, obsidian, lunar glass Space industrialization metals production via melt electrolysis

o f the p r e p a r a t i o n o f glasses w i t h d e s i r e d c h a r a c t e r i s t i c s . T h i s r e v i e w will f o c u s o n o x i d e - c o n t a i n i n g b a s e c o m p o s i t i o n s such as silicate, b o r o s i l i c a t e a n d p h o s p h a t e melts.

H.D. Schreiber / R e d o x processes in glass-forming melts

131

3. Redox equilibrium 3.1. Solvolysis equation

When a single multivalent element is dissolved within a glass-forming oxide melt, it will establish an equilibrium between its two (or more) valence states in this solvent. The equilibrium can be described as a solvolysis reaction, because the melt participates directly in the actual oxidation-reduction process. Several representations of the redox equilibrium equation have been expounded; each possesses certain advantages and disadvantages. For an individual multivalent element " M " in a glass melt, the general redox equilibrium can be easily written as m+

2-

d l~/f ( m - - n ) +

M(melt) + 2 n O ( m e l t ) ~ . . . . (melt)

+ nO2(gas),

(1)

where n is the number of electrons involved in the reduction [2,5]. The participation of the solvent in the establishment of the redox equilibrium is clearly shown via the presence of the oxide ion in this equation. Despite the inherent simplicity of this expression, criticism has often been directed at eq. (1) because it does not display the proper dependence of the redox ratio (concentration ratio of M ( ' - ' ) + to M "+) to the base composition with respect to melt basicity (oxide ion activity) [6]. However, since the redox ions exist in the melt as solvated species typically coordinated to oxygens within the silicate network, the activity redox ratio may not be directly proportional to the concentration redox ratio. Thus, this " i m p r o p e r " dependence of the concentration redox ratio to melt basicity can be rationalized by reference to solvation activity coefficients of the redox ions [7]. In order to obtain directly the compositional effect on the redox ratio without recourse to the ion activities of M m+ and M ( . . . . )+, an alternate form of the redox equilibrium was developed: A jlArf'~(m - 2x) . . . . . ~(melt)

(m-n)+ ~---All/f --r~(melt) +

nO2(gas), (4x -- 2n)O{mel,)+ 2

(2)

where x is the number of oxygens associated with the oxidized redox ion in the melt to form an oxo-anion [8]. Such an equation appears to describe the redox equilibria which involve CrO 2 , VO 3, FeO 2 and other species where discrete oxo-anions are readily identifiable or realistic. However, for some redox systems the postulated oxo-anions involve a certain "phoniness", as if the complexes were created simply to balance the oxygens required by eq. (2). In order to apply eq. (2) to an actual redox equilibrium, it must then be realized that MO~ m 2x) may not be representative of a real entity but instead the value of x implies the relative number of oxygen associations of the oxidized ion in the melt with respect to that of the reduced ion in the melt. In analogy to the solid state reaction, yet another form of the redox equilibrium can be written: 2Fe203~melt) ~ 4FeO(melt) + 02(gas ) ,

(3)

H.D. Schreiber / Redox processes in glass-forming melts

132

as for the case of the Fe(III)-Fe(II) system. Obviously, however, pseudocrystallites of these redox species are not realistically present in the melt but can be used in an operational sense. One advantage of this form of the redox equilibrium is that it does allow ready access to thermodynamic approximations [9]. In order to ascertain which equation is the "best" representation of the redox equilibrium actually established in the glass-forming melt, it must be realized that all equations are by necessity simplifications. It is impossible to show completely in one equation the complexities of solvation of the redox ions of M and of polymerization of the glass-forming network via just the oxide ion. Equations (1-3) become different ways of envisioning the same phenomenon; as long as the subtleties of each equation are recognized, they are essentially equivalent. Thus, the use of a particular redox equilibrium expression becomes one of personal preference. 3.2. Redox components

Redox behavior is exhibited by a wide range of multivalent element in glass-forming melts [3]. Transition metals {Fe(III)-Fe(II)-Fe(0)}, lanthanides {Eu(III)-Eu(II)}, actinides { U ( V I ) - U ( V ) - U ( I V ) - U ( I I I ) } , metalloids {As(V)-As(III)}, and non-metals {S(VI)-S(0)-S(-2)} all can exist in one or more valence states within the melt. A quantitative measure of the degree to which a certain multivalent species is reduced in an aqueous solution is provided by the standard reduction potential of that species. Likewise, in glass-forming melts, each multivalent element possesses its own inherent reducibility. Under a given set of melt

Table 2 Experimental

r e f e r e n c e r e d u c t i o n p o t e n t i a l s f o r r e d o x c o u p l e s in S R L - 1 3 1

melt composition

2) a t 1 1 5 0 ° C [10] Redox couple

E'

Redox couple

(rel. u n i t s )

E' (rel. u n i t s )

N i = 3,2

1.7

V = 5,4

- 1.9

C o = 3,2

1.4

U = 5,4

- 2.2

A u = 3,0

0.8

C u = 1,0

- 3.3

M n = 3,2

0.8

C r = 3,2

-3.4

A g = 1,0

0.5

M o = 6,5

- 3.8

O = O~,2

0

V = 4,3

-4.0

C e = 4,3

-0.1

E u = 3,2

-4.3

C r = 6,3

- 0.3

T i = 4,3

- 5.0

S b = 5,3

- 0.3

N i = 2,0

- 5.3

C u = 2,1 U = 6,5

-0.8 - 1.5

Sn = 4,2 C o = 2,0

-5.5 - 6.0

F e = 3,2

- 1.7

F e = 2,0

A s = 5,3

- 1.7

M o = 5,0

- 6.3 - 15.8

(fig.

H.D. Schreiber / Redox processes in glass-forming melts

133

conditions, for example, one element may exist predominantly in the reduced state whereas another element may be stabilized in its oxidized state. The value of the equilibrium constant K for eq. 1 or 2 is of course different for different multivalent elements M. The degree to which that reduction equation is shifted to the right can be related to the inherent reducibility of that redox couple or to a relative value of the reduction potential in the melt. Table 2 compiles the most complete series of these reduction potentials in a borosilicate glass-forming melt under reference conditions and with reference to the oxygen-oxide couple, O2(gas ) q- 2e ~ 20(melt)

(4)

assigned a value of zero [10]. Although these potentials are defined for a particular reference glass melt, it has been ascertained that the ordering of the redox couples into an electromotive force series is relatively independent of the base composition and melt temperature. In general, the order mimics that in aqueous solution, indicative of similarities in solvation of the redox ions in the two systems. 3. 3. Base composition

The base composition of a glass-forming melt determines the magnitude of depolymerization of the silicate, borate or phosphate network and, thus, the amount of free oxide ions released into the melt. In order to increase the

10 I

% of R e d o x C o u p l e in Reduct~d S t a t e 1 10 30 50 70 90 99 I t I I ! I ! i

i

9 ~

i/ /

17

.t ,,'/

,'7 ,'/

s.

,'/

I

2

'

0

I

I / - ~+

3-

15000C

t/

//i

/i

Cr3+-Cr2 // -

- -

I -2

F,s

,

I -1

/7

,'/

,7

,/ I

,/I

0

,7

~// i 1

i

I 2

log ~'Reduced~ / EO xidized 3

Fig. 1. The effect of melt composition on redox equilibria. Solid lines indicate equilibria for melt composition FAS (54.2 wt% SiO2, 9.9% A1203, 5.6% CaO, 30.3% MgO) while dashed lines indicate equilibria for the more basic melt composition F A D (51.6 wt% SiO2, 2.3% A1203, 19.2% CaO, 26.9% MgO) at 1500°C [12].

134

H.D. Schreiber / Redox processes in glass.forming melts

basicity or the oxide ion activity of a melt, more network modifiers are introduced into a melt at the expense of the network formers. For examples, SiO 2 as a melt component can be systematically replaced by N a 2 0 to enhance the melt basicity. Since the redox equilibria expressed in eq. (1) and (2) include the oxide ions as a part of the solvolysis reactions, the equilibria are affected by the base composition of the melt. As the basicity of the melt is increased, the percentage of that redox couple in the oxidized state also increases [11]. In fact, the redox couples of vanadium and chromium are often employed as internal probes of the acidity/basicity of a particular glass composition [11]. Equation (2) directly relates the redox ratio of the multivalent element to the melt basicity, as explained in the previous section. The redox equilibria of cerium, iron, chromium and titanium in two different melt composition are compared in fig. 1. Of the two melts, FAS contains a higher proportion of alumina and silica at the expense of calcia and magnesia than composition FAD. Since FAS is accordingly more acidic than FAD, that composition stabilizes a greater percentage of the reduced valence states at the same melt conditions [12].

3.4. Oxygen fugacity If the equilibrium constant expression for eq. (1) or (2) is mathematically manipulated, the relationship -log

fO 2 = 4log

[Mm+ ]

+B

(5)

can be derived for dilute solutions of M in the melt [2,3,5]. The term b incorporates all quantities that are functions of temperature a n d / o r base composition, and " n " once again represents the number of electrons transferred in the reaction. Thus, a plot of -log(oxygen fugacity) versus log(redox ratio) should yield a straight line of slope " 4 / n " at constant temperature for a given composition. This relationship is illustrated in fig. 2 for many redox couples, and upon which the results of table 2 are experimentally based [13]. In a qualitative sense, both eqs. (1) and (2) show that as the amount of oxygen in the atmosphere of the melt is increased, the concentration of the oxidized ion of " M " increases at the expense of that of the reduced species. Thus, by control of the oxygen fugacity imposed on the melt by the atmospheric gases, the redox state of the melt can be regulated.

3.5. Melt temperature The reduction of a multivalent species as shown by eq. (1) or (2) is always endothermic; that is, in order for the reduction of M m+ to occur in the melt, heat has to be added to the system'. Thus, as the temperature of the melt is increased, the melt becomes more reduced or the proportion of the multivalent

135

H.D. Schreiber / Redox processes in glass-forming melts

149~ ~

~

PERCENTAGE OF REDOX COUPLE IN REDUCED STATE

1

5

10

20 3040506070 80

15

go

95

,~,o,

99

~,.o~

,,

12 11

o;o,~

' ,~.~ ~

",

2,

.p"

/j// o

6

/

/

~

;3,~



4 3 2 1 0

* -3



. -2

,

.

.



-1

.



i



0

.

.

.

I

I

1

2

3

EOX,D'ZED3 Fig. 2. Relation of imposed oxygen fugacity to analyzed redox ratio of individual multivalent elements dissolved in SRL-131 melt (57.9 wt% SiO2, 1.0% TiO 2, 0.5% ZrO 2, 14.7% B203. 0.5% La203, 2.0% MgO, 17.7% Na20, 5.7% Li 20) at 1150°C. Dopant concentration is nominally about 1 wt% except where noted. Experimental points were fitted to theoretical straight lines predicted by eq. (5). Arabic numerals following element symbols for that relation represent oxidation states for that redox couple [13].

element in the reduced state increases. The enthalpy change for this reduction process (eq. (1) or (2)) can be calculated through an approximation of the Clausius-Clapeyron equation,

log

[M,,, +]

2.303R

+ b,

(6)

where the redox ratio is representative of the equilibrium constant, R is the ideal gas constant, T is the temperature in °K, A H is the enthalpy of reduction, and b is a term that incorporates oxygen fugacity and composition dependencies. This equation is only an approximation because of the replacement of the equilibrium constant by the redox ratio and because of the assumption that the oxide activity is temperature-independent in b. A plot of the log(redox ratio) versus reciprocal absolute temperature for a given melt at a reference oxygen fugacity should be linear with slope ( - A H / 2 . 3 0 3 R ) in accordance with eq. (6). Figure 3 illustrates the temperature dependence of the Fe(III)-Fe(II) equilibria in a reference borosilicate melt. Most enthalpies of reduction in glass-forming melts range from 10 kcal/mol to 70 kcal/mol [5]. On a first approximation, the enthalpy of reduction should also correlate to the ease of reduction of that redox couple as monitored by the reduction potential.

H.D. Schreiber / Redox processes in glass-forming melts

136

T (~) m

1150

1050

950

850

I

I

I

I

0

I

0

SRL-131

"~ - o . 8 \

o

*L

-1.o

lwt% Fe~ 0 ~ 4 5

Q

~

Kcal/mole

-1.5 @

-2.0 t /NH-73Kcal/mole \lOwt% Fe \

0

-2.5

[

I

,

0.0007

I

0.0008

1/T I"1(-11

,

Q-I

0.0009

Fig. 3. The temperature dependence of the Fe(III)-Fe(ll) equilibrium in melt composition SRL-131. The AH for the reduction is computed from the slope of the straight-line relationship.

3.6. Multivalent element concentration

As long as the concentration of the multivalent element dissolved in the glass-forming melt remains sufficiently dilute, the redox equilibrium established by that element should remain invariant. "Sufficiently dilute" is defined as the activity coefficients (due to solute-solvent interactions) of the redox ions being constant with concentration. Typically, this statement is valid up to about 2 - 4 wt% of that multivalent element. Perturbations in the redox equilibrium at very low redox concentrations were reported, but later attributed to other factors [14]. When the multivalent element concentration is increased so that it becomes a minor or major component Of the melt, the redox element contributes to changes in the melt basicity and thus deviations from the redox equilibrium established for dilute solutions. Figure 2 illustrates the dependence of the F e ( I I I ) - F e ( I I ) redox equilibrium on the total iron concentration in a reference melt. As the concentration of iron, usually a network modifier, is increased in the melt, the melt becomes more basic and thus stabilizes more of the iron as Fe(III). 3. 7. Mutual interaction

If two or more multivalent elements are simultaneously introduced in a glass-forming melt, the redox couples may interact by internal electron ex-

H.D. Schreiber / Redox processes in glass-forming melts

137

change reactions m+ --1- vii?( r s)+ ~ ~l.f(m n)+ sM(melt) . . . . . (melt) ~ &"(mert) q'-

r+ nR(me],l

(7)

where M and R are the two interacting redox elements. In this melt equilibrium, M " + is reduced to the M (m-n)+ ion while R ( r - ' ) + is oxidized to the R r+ species with a transfer of ns electrons. The interaction between different redox ions in melt systems can be viewed as analogous, at least qualitatively, to oxidation-reduction reactions in aqueous solution [7]. However, this point of view is not universally accepted, as another hypothesis is that each individual redox equilibrium ( M and R) is buffered by the melt composition (eq. (1)) so that the electron exchange cannot occur at melt temperature but instead occurs during the quench [15]. Recent developments [16,17] in the interpretation of mutual interactions in silicate melts indicate the identification of the mechanism involving the solvent in the electron exchange, the concept of redox buffering to minimize the internal rearrangement of two redox couples in the presence of the third, and the correlation of the magnitude of interaction to net electrode potential as in aqueous systems - all of which substantiate the viewpoint that the electron exchange occurs at melt temperature. Such internal mutual interactions, such as the M n - F e coupling, have been employed for the decolorization of glass melts. 3.8. O t h e r r e d o x i n t e r a c t i o n s

The ideal container for studying redox chemistry in melts is totally inert to the melt. Unfortunately, the container introduces unwanted disturbances on the measured redox processes in many cases so that care must be taken to minimize such melt-container interactions. Pt, P t / R h or other metals are typically the preferred container materials under most oxidizing conditions. But such metals are known to alloy with iron and copper, although such redox interferences can be minimized by use of Pt rings instead of capsules. Alumina, silica, and iron containers will all slowly corrode into the melt, changes the melt composition and accordingly upset the established redox equilibria. Graphite and molybdenum containers can typically be used under reducing conditions. The presence of other gases in the melt atmosphere and their subsequent dissolution in the melt may also affect the redox equilibrium. For example, carbon dioxide, water and sulfur dioxide dissolve in melts as 22C02(gas ) q- O(melt) ~ C03(melt),

(8)

2~ 2OH(melt), H 2 0 ( g a s ) + O(melt)

(9)

1 2~ S04~melt ) SO2(gas ) q- ~O2(gas ) "It- O(melt)

(10)

using oxide ions from the melt [8]. Consequently, the presence of these gases should result in the uptake of oxide ions from the melt and the formation of

138

H.D. Schreiber / R e d o x processes in glass.forming melts

more reduced species at the expense of the oxidized ion in accordance with eq. (2). This should lead to a synergistic effect whereby the presence of the redox component should enhance the solubility of the gas. In particular, the reverse reaction of eq. (1) or (2) could simply be regarded as the chemical solubility of oxygen in the melt.

4. Redox kinetics

Because of time constraints a n d / o r changing melt conditions in most glass-making operations, true redox equilibrium of the melt may be practically unattainable. Correspondingly the rate at which redox equilibrium is established - redox kinetics - becomes a crucial propertY of the melt. 4.1. Static melts

If a melt is initially at redox equilibrium and some parameter is changed so as the perturb the equilibrium (for example, melt temperature, imposed oxygen fugacity), the equilibrium will attempt to respond by minimizing the effect of the perturbation, in accordance with Le Chatelier's principle. How rapidly the equilibrium adjusts in the establishment of the new redox equilibrium will be regulated by the rate-determining step of this kinetic process. The actual electron transfer step as well as the atmosphere to melt surface re-equilibrium are both expected to be very rapid. Current theory [19,20] postulates that the rate-determining step in redox kinetics is the diffusion of oxygen into (oxidation) or out of (reduction) the melt. In order for the gaseous oxygen to be utilized in the redox equilibrium of eq. (1) or (2), the oxygen must first be transferred to or away from the site of the multivalent element. Equation (1) becomes in effect m+

2--

~

AliA(m-n)+

4m(melt) + 2 n O(melt) . . . . .

(melt)

-+- F/O2(gas ~ melt)"

(11)

In particular this is illustrated by a change in the imposed oxygen fugacity on the melt, since Henry's law requires the dissolved oxygen content in the melt to be proportional to the gaseous concentration. Figure 4 illustrates this diffusion semi-quantitatively by the color boundary monitoring the following reaction •4+ + 20(melt) 2- ~~ 4Tl(melt) .3+ + 02(melt ~ gas), 4Tl(meh)

(12)

whereby the colorless glass containing only Ti(IV) is systematically turned purple due to the formation of Ti(III) as oxygen diffuses out of the melt when a more reducing atmosphere is imposed on the melt. Redox kinetics can also be quantitatively related to the diffusion of oxygen via Fick's laws of diffusion for a gas into a liquid of known geometry [19,20]. Figure 5 shows the correlation between oxygen concentration, redox ratios,

H.D. Schreiber / Redox processes in glass-forming melts

139

t=Osec t=6OOsec t=2700sect=3600sect=9OOOse~. ~clear, ~A,u r /0.1mm p l e ~ ' ~ u rn pie ~'l~lPUrple~'~ ] " ~PtlmMq~ ~,b~ ~~purple Vcolorless ~/~°l°rless~//~°'O~%e~//~ "°~e ~olorless t=14400sec t=O0 ~purple ~purple iOI scale(mm) Fig. 4. Schematic diagram of the cross sections of a set of glass samples, SRL-131 + 5 wt% Ti, that were initially synthesized at l l 5 0 ° C in air, quenched, then remelted at 1150°C in an atmosphere of f O 2 = 10- 14 arm as a function of time. The sample for t = o¢ was synthesized at 1150°C and f O 2 = 10-14 arm for 24 h. The purple color is characteristic of the presence of Ti(IlI) in the melt, while Ti(IV) is colorless [19].

redox kinetics and oxygen diffusion coefficients for a particular melt [19]. The rate of oxidation and reduction in pure molten iron oxide, likewise, was found to be controlled by oxygen diffusion [18]. Thus, redox kinetics are controlled by oxygen diffusion through the melt [19-21], and should not be correlated to any particular rate law [22]. Other gaseous diffusions could control redox kinetics as well. For example, sulfur gases, water and carbon dioxide by their incorporation in the glass-forming melt could upset the established redox equilibria as they diffuse into or out of the melt through 1) changes in the oxide ion activity, 2) concurrent changes in oxygen fugacity, or 3) mutual interactions such as the F e - S amber chromophore. Redox kinetics in amber glass formation are probably related not only to oxygen diffusion but also "sulfur" gas diffusion through the melt

[211. 4.2. Dynamic melts Other variables such as convection and mechanical stirring are encountered in the production of large quantities of glass melts. These factors should enhance the rate of redox equilibration. However, from an understanding of the redox kinetics in the static melts, it becomes known that an increased exposure to the atmospheric gases will speed up the rate of redox equilibration. This will minimize the effect of the rate-determining step, that of gaseous diffusion through the melt. The influence of the melt atmosphere and dissolved gases on melt processing present not only technical difficulties in glass production but also a mechanism for redox control [23]. If a redox additive is added to the melt, the mutual interaction of eq. (7)

140

H.D. Schreiber / Redox processes in glass-forming melts D=9 x lO-6cm2/sec t = 2 0 0 0 seconds 1.0

~

t--lO,O00 seconds

I

'

I

'

I

"

I

"

I

"

I

"

I

"

I

I

'

!

!

'

!

"

I

"

f

'

I

"

I

I, I

"

I . I

'

I, I

"

I, I

"

I, I

"

I, I

"

I. I

"

J I

I, !

'

I, I

'

I, I

'

I, I

'

I , |

"

! . I

'

I I

, "

I !

I I

. "

I I

, "

I i

, '

I I

, '

I I

, "

I I

, '

I I

, '

I I

,

I

.

I

,

I

,

I

,

I

,

I

,

C

0.8

r~ 0.6

% 0,4 0.2

%o '~'8 ~4 II

2

~e

2

o

I, I '

I , I '

I,

I,

I.

I.

I"

I '

I"

I

"

I,

I

,

I

I

'

I

I

,

I

8o

g so x 4o 0

>0 2o

°~ m

¢J

I

,

a

0 .1 bottom

,

I

.2

,

I

.3

,

I

.4

,

.5

.6

z, melt depth (cm)

.7 top

I

0 .1 bottom

.2

.3

.4

.5

.6

I

.7 top

z, melt depth (cm)

Fig. 5. Depth profiles for a cerium-containing melt (of total depth 0.7 cm), SRL-131, at 1150°C. For oxidation (solid line) the melt was presumed to possess initially a ICe 3+ ]/[Ce 4+ ] of 7.9 for 1150°C in f O 2 = 1 0 4 atm, after which it was remelted at 1150°C in f O 2 = 1 0 -°7 arm to produce a ICe 3+ ]/[Ce 4+ ] = 1.3. For reduction (dashed line) the conditions were reversed. Depth profiles are given for (A) relative oxygen concentration using 90 as the concentration in the oxidized glass and 10 as that in the reduced glass, (B) redox ratio of the Ce(IV)-Ce(III) couple, and (C) the concentration of Ce 3+, with 1.01 wt% total cerium. An assumed diffusion coefficient of 9×10 -6 cmZ/s was employed in the calculations for two different "equilibration" times according to redox kinetics [19].

becomes operational. Likewise the rate of this reaction should then be cont r o l l e d b y t h e i n t e r d i f f u s i o n o f t h e m u l t i v a l e n t species.

5. D i s c u s s i o n T a b l e 1 e m p h a s i z e s t h e i m p o r t a n c e o f r e d o x p r o c e s s e s to a n u n d e r s t a n d i n g o f a n d t h e p r o d u c t i o n o f m a n y t y p e s o f glasses. T h e r e d o x p r o c e s s e s , b o t h

H.D. Schreiber / Redox processes in glass-forming melts

141

equilibrium and kinetic in conjunction with each other, have been used to explain certain aspects of glass refining [19]. "High-tech" glasses require materials with unique magnetic, optical or electrical properties. The redox processes necessary for the production of these substances and certain redox states stabilized therein depend on both equilibration and kinetic aspects of redox chemistry. For example, interesting glasses containing a known redox ion in the top part of the glass and another redox ion in the bottom part of the glass could be manufactured by the procedure illustrated in fig. 4. Differences in redox chemistry may also be attributed to differences in sol-gel versus melt preparation of glasses [24]. The research reported in this paper was supported by NASA (NAG-9-102), NSF(CPE-8406089), and the Research Corporation. This paper constitutes contribution No. 1 of the Center for Glass Chemistry operating in part under the auspices of the Virginia Center for Innovative Technology and VMI Research Laboratories. The author thanks Anita Fuller for typing the manuscript and Kim McManus for drafting the figures. References [1] R.W. Douglas, in: Recent Advances in Science and Technology of Materials, vol. 2, ed. A. Bishay (Plenum, New York, 1973) p. 47. [2] H.D. Schreiber, J. Non-Cryst. Solids 42 (1980) 175. [3] H.D. Schreiber, in: Advances in Materials Characterization, eds. D.R. Rossington, R.A. Condrate and R.L. Snyder (Plenum, New York, 1983) p. 647. [4] N. Kreidl, Ceramic Bull. 63 (1984) 1394. [5] W.D. Johnston, J. Amer. Ceram. Soc. 48 (1964) 185. [6] S.M. Budd, Phys. Chem. Glasses 7 (1966) 210. [7] H.D. Schreiber, T. Thanyasiri, J.J. Lach and R.A. Legere, Phys. Chem. Glasses 19 (1978) 126. [8] S. Holmquist, J. Amer. Ceram. Soc. 49 (1966) 228. [9] H.J. Tress, Phys. Chem. Glasses 1 (1960) 196. [10] H.D. Schreiber, Lunar Planet. Sci. XVI (1985) 738. [11] A. Paul, Chemistry of Glasses (Chapman Hall, London, 1982)p. 148. [12] H.D. Schreiber and G.B. Balazs, Lunar Planet. Sci. XIII (1982) 692. [13] H.D. Schreiber, G.B. Balazs, B.E. Carpenter, J.E. Kirkley, L.M. Minnix and P.L. Jamison, Comm. Amer. Ceram. Soc. 67 (1984) C106. [14] D.S. Goldman, J. Amer. Ceram. Soc. 66 (1983) 205. [15] A. Paul and R.W. Douglas, Phys. Chem. Glasses 7 (1966) 1. [16] H.D. Schreiber and G.B. Balazs, J. Non-Cryst. Solids 71 (1985) 59. [17] H.D. Schreiber, B.E. Carpenter, J.P. Eckenrode and G.B. Balazs, Phys. Chem. Glasses 26 (1985) 24. [18] E.T. Turkdogan, in: Physicochemical Properties of Molten Slags and Glasses (Metals Soc., London, 1983) pp. 248, 326. [19] H.D. Schreiber, S.J. Kozak, A.L. Fritchman, D.S. Goldman and H.A. Schaeffer, Phys. Chem. Glasses 27 (1986) in press. [20] D.S. Goldman and P.K. Gupta, J. Amer. Ceram. Soc. 66 (1983) 188. [21] H.A. Schaeffer, H. Lachenmayr and L.D. Pye, Glastechn. Ber. 56K (1983) 602. [22] R. Pyare, S.P. Singh, A. Singh and P. Nath, Phys. Chem. Glasses 23 (1982) 158. [23] R. Bruckner, J. Non-Cryst. Solids 71 (1985) 49. [24] S. Dave and R.K. MacCrone, J. Non-Cryst. Solids 71 (1985) 303.