Original papers

STRUCTURAL AND MECHNICAL RESPONSE

TO A THERMO-RHEOLOGIC HISTORY OF SPINEL SLUDGE
IN HIGH-LEVEL WASTE GLASS
MILOŠ JIŘIČKA, PAVEL HRMA*

Laboratory of Inorganic Materials, Institute of Inorganic Chemistry ASCR and Institute of Chemical Technology, Prague,
Technická 5, 166 28 Prague, Czech Republic
E-mail: [email protected]

*Pacific Northwest National Laboratory, Richland, Washington 99352, U.S.A.

Submitted April 4, 2001; accepted October 18, 2001.

The composition and structure of a sludge sample from a high-level waste glass melter were studied using optical and
scanning electron microscopy, and x-ray diffraction. At isothermal heat treatments between 1050°C and 1350°C, spinel
crystals partly dissolved to form on cooling tiny (~ 10 µm) star-like crystals or dendrites. The shear stress in sludge was
measured at a constant shear rate (from 0.005 s-1 to 1.0 s-1) and temperature (from 1050°C to 1350°C) during repeated
deformation and after idling. The initial thixotropic character of the loose structure of the settled sludge turned on subsequent
deformation (and idling) to rheopectic behavior. As the spinel concentration in the sludge decreased from 28 mass% (sludge
as received) to 15 mass% at 1300°C, the sludge turned into a Newtonian suspension.

Keywords: Glass, Spinel, Rheology

INTRODUCTION Brookfield digital rotating viscometer with a disk

spindle was used to study the response of the sample
Little attention has been paid to the mechanical (which was large enough to fill the viscometer crucible)
behavior of sludges of solid particles in molten glass. to repeated shear deformation at different temperatures.
Yet solid phases, mainly spinel, can precipitate and Thus, all measurements were performed consecutively
accumulate in high-level waste (HLW) melters, po- on a single sample during 3 days as shown in figure 1.
tentially shortening the melter's lifetime and hindering, For simplicity, each experiment run is represented in
even jeopardizing, process operation. Plodinec [1] and figure 1 as the end values of temperature, shear stress,
Mika et al. [2] reported that spinel sludge is a rheopectic and spindle speed, not as a time interval. As seen in
pseudoplastic liquid. The viscosity of such a liquid figure 1, the temperature ranged from 1050°C to
decreases with increasing deformation rate and 1350°C, and the spindle speed varied from 0.005 to
increases with the time of deformation. This type of 1 RPS. The sludge in the platinum crucible was
behavior is encountered in suspensions in which solid allowed to cool in air between measurements. With
particles form loosely bounded aggregates or alignment a few exceptions, measurements ran for 15 min.
groupings that break down with continuing deformation Before each measurement, the sample was held for
and increasing shear rate [3]. This study investigates the 90 min at measuring temperature. The idle time for
response of spinel sludge to a complex temperature and spindle-speed change was 5 min and for temperature
deformation history as a preliminary attempt to assess change 30 min.
the potential for removing the sludge from the melter. The shear stress (τ) and the shear rate (∇v) were
calculated using the formulas*
EXPERIMENTAL
M
A small (~300-g) sample of spinel sludge from a τ= (1)
pilot-scale HLW glass melter was characterized using 2πRdL
optical microscopy, scanning electron microscopy with
energy-dispersive spectroscopy (SEM-EDS), x-ray 2ϖ Rc2 Rd2
diffraction (XRD), and an inductively coupled argon ∇v = (2)
plasma spectrometer (ICP). Pure spinel crystals were x2 (Rc2 - Rd2)
obtained from a portion the sludge sample by dissolving
the amorphous and metallic phases in 18 % HNO3 for
24 h, adding HF for 48 h, and then separating spinel * Brookfield Viscometer Guide, Brookfield Engineering
from amorphous gel with a magnet. Laboratories, Brookfield, Massachusetts, USA

Ceramics − Silikáty 46 (1) 1-7 (2002) 1

 M. Jiřička, P. Hrma

where M is the torque, ϖ is the angular velocity, x is the RESULTS AND DISCUSSION
radial distance from spindle axis, Rd is the spindle
radius, Rc is the crucible radius, and L is the Sludge characterization
spindle height (in this study, Rd = 7.3 mm, L = 2.0 mm,
and Rc = 33 mm). The lower surface of the spindle Figure 2 shows a micrograph of the sludge as
was 13 mm above the crucible bottom. The sludge received. The major crystalline phase in the sludge was
height in the crucible was approximately 30 mm. identified by XRD as spinel. Spinel content in the
Equations (1) and (2) are valid for a uniform material. as-received sample was 27.55 mass%. Spinel crystals,
As discussed below, this may not be the case in sludge 20 to 100 µm in size, formed a loose structure with large
deformation. (several hundred µm) areas of crystal-free melt. The
agglomerates of fine needles in the middle of the
micrograph are crystals of silver. Figure 3 shows a
micrograph of the sludge quenched from 1250°C at the
end of the study. The crystals appear somewhat smaller
than in the original sludge, their edges are rounded, and
crystals are surrounded with layers of clear glass that
are approximately 20 mm thin. Beyond these
concentration-boundary layers, the bulk of the
amorphous phase is dark brown in color and filled with
small (10 µm) star-shaped crystals of spinel as seen in
figure 4. The star-shaped crystals probably formed
during cooling.

Figure 2. Sludge as-received (27.55 mass% spinel).

Figure 1. The thermo-mechanical history.

The numbers of the time intervals denote experimental runs;
the figures associated with these runs are listed above.
1 - figure 8, 2 - figure 10, 3 - figure 12, 4 - figure 13,
5 -figure 8, 6 - figure 14, 7 - figure 9, 8 - figure 11, 9 - figure 9, Figure 3. Sludge heat-treated at 1250°C (16.36 mass% spinel).
10 - figure 15

2 Ceramics − Silikáty 46 (1) 1-7 (2002)

 Structural and mechnical response to a thermo-rheologic history of spinel sludge in high-level waste glass

At the end of the study, the sample was transferred Table 1 lists the sludge composition, determined
to a quartz crucible and stirred with the spindle at with ICP, and the compositions of the amorphous and
1100°C. A sample of the sludge that adhered on the crystalline phases, determined with SEM-EDS. The
spindle surface contained 15.69 mass% spinel, while a amorphous-phase composition was calculated using the
bulk sample contained 21.47 mass% spinel. Two small mass balance equation
samples of the bulk sludge were subsequently heat
treated in platinum crucibles for 24 h, one at 1035°C, k

and the other at 1303°C. The corresponding micro- gi - ∑ cijCj

j=1
graphs are shown in figure 5 and 6. Figure 5 resembles ai = (3)
figure 2 except for a small number of tiny crystals that k

probably precipitated during the heat treatment, 1- ∑ Cj

j=1
indicating that the structure of sludge was similar to that
as received. At 1303°C (figure 6), the sludge re-boiled
during heat treatment (as it did previously at T > 1300°C where ai and gi are the i-th component mass fraction in
in the viscometer). The crystals are half of their original amorphous phase and sludge, respectively, cij is the i-th
size (≤ 50 µm), are rounded, and are surrounded with component mass fraction in j-th crystalline phase, and
layers of clear glass. Large volumes of crystal-free bulk Cj is the j-th crystalline phase mass fraction in sludge.
glass are dark brown in color. Figure 7 shows an in- We used spinel composition as measured (in sludge) and
teresting dendritic growth of spinel in a sample of the assumed that the metallic phase was pure silver, and all
original sludge that was heat treated for about 1 h at the Ag2O was reduced to a metallic phase. The
1400°C and then quenched (small sporadic dendrites composition of spinel isolated by dissolving the
were also seen in the sample heat treated at 1303°C). amorphous phase in acid shows a smaller content of

Figure 4. Sludge heat-treated at 1250°C (Detail). Figure 6. Sludge heat-treated at 1303°C (15.30 mass% spinel).

Figure 5. Sludge heat-treated at 1035°C (25.78 mass% spinel). Figure 7. Sludge heat treated at 1400°C and quenched
(28.58 mass% spinel).

Ceramics − Silikáty 46 (1) 1-7 (2002) 3

 M. Jiřička, P. Hrma

Table 1. Composition of sludge, spinel, metallic phase, and amorphous phase (mass fractions).

Sludge Amorphous phase Amorphous phase Metallic phase Spinel Isolated spinel
(ICP) (Calculated) (SEM-EDS) (SEM-EDS) (SEM-EDS) (SEM-EDS)

Ag2O 0.0018 -- -- 0.9134 -- --

Al2O3 0.0888 0.1933 0.2242 0.0217 0.0162 0.0158
B2O3 0.0541 0.0748 -- -- -- --
BaO 0.0004 0.0006 -- -- -- --
CaO 0.0071 0.0067 0.0078 0.0002 -- --
CdO 0.0003 0.0004 -- -- -- --
CoO 0.0001 0.0002 -- -- -- --
Cr2O3 0.0081 0.0021 -- -- 0.0241 0.0920
Fe2O3 0.2533 0.0743 0.0862 0.0179 0.7236 0.6567
K2O 0.0039 0.0034 0.0039 -- -- --
Li2O 0.0342 0.0473 -- -- -- --
MgO 0.0040 0.0009 0.0010 0.0024 0.0034 0.0086
MnO2 0.1016 0.0213 0.0247 0.0035 0.1633 0.1578
Na2O 0.0691 0.1621 0.1879 0.0168 -- --
NiO 0.0118 0.0064 -- -- 0.0259 0.0229
P2O5 0.0034 0.0047 -- -- -- --
SiO2 0.2987 0.3404 0.3947 0.0187 -- --
SrO 0.0388 0.0495 0.0574 0.0037 -- --
TiO2 0.0008 0.0010 -- -- -- --

Fe2O3 (possibly caused by leaching), and a higher

content of Cr2O3 as compared to spinel in sludge.

Mechanical behavior

Figure 8 shows the shear stress as a function of

time at a constant spindle speed (0.01 RPS and
0.025 RPS) and temperature 1100°C. The measurement
was repeated several times with increasing idle time
between subsequent runs. During the starting run at
0.01 RPS, shear stress of the as-received sludge gradual-
ly decreased. This thixotropic behavior can be attributed
to the breakdown of the loose structure that formed as
spinel crystals settled in the crucible and subsequent
motion of the crystals caused by velocity gradient
resulting in the formation of a low-concentration zone Figure 8. Shear stress versus time at 1100°C.
near the spindle surface. The existence of such a zone
was experimentally verified (see previous subsection).
By equation (2), the Newtonian shear rate at the
perimeter of the spindle is 20.4 times larger than the
shear rate at the crucible wall. For non-Newtonian mate-
rial, it seems likely that only a limited volume of sludge
around the spindle (the deformation zone) moved with
the rotating spindle whereas the sludge farther away
from the spindle was stagnant. Assuming that the shear
was restricted to a sludge layer between Rd and Rm < Rc.
For Rm - Rd << Rd, equation (2) simplifies to
ϖ Rd
∇v ≈ (4)
R m- R d

Hence, the shear rate decreases as the deformation zone

increases. In this paper, we report the value of shear rate
(velocity gradient) at the spindle surface, i.e., x = Rd. Figure 9. Shear stress versus time at 0.05 RPS.

4 Ceramics − Silikáty 46 (1) 1-7 (2002)

 Structural and mechnical response to a thermo-rheologic history of spinel sludge in high-level waste glass

Within the deformation zone, crystals settle and As shown above (see figures 3 and 6), spinel
move in the direction toward lower velocity, creating an crystals partly dissolved with increasing temperature.
area of low concentration in the vicinity of the spindle. The effect of temperature on sludge rheology was tested
The existence of this area may explain the remarkable at two spindle speeds, 0.01 and 0.025 RPS (figures 12
decrease in the shear stress at 0.01 RPS, seen in and 13). At 0.01 RPS, the temperatures of the
figure 8, when the measurement was repeated after a subsequent runs were changed to 1095, 1046, 1095,
period of idling. This effect was much less pronounced 1144, and 1095°C. As expected, the shear stress
at 0.025 RPS. After the first run at 0.01 RPS, during increased when the temperature decreased from 1095°C
which the shear stress decreased, the shear stress was to 1046°C and returned to the previous value when the
nearly constant or increased with time during
subsequent runs. This rheopectic behavior, which was
more pronounced at a higher spindle speed, could result
from stirring settled crystals and their agglomerates into
the deformation zone. At both spindle speeds, the first
data point (at 60 s) after 1800 s of idling shows a higher
shear stress than the corresponding point after 600 s of
idling. This indicates that spontaneous strengthening of
the structure can occur with extended idling.
As figure 9 shows, the sludge was rheopectic from
the start when deformed at a higher temperature
(1250°C) and a higher speed (0.5 RPS). Moreover, the
shear stress increased during idling. Except for the
sharp drop in the shear stress at 1300°C during
deformation after 600 s idling, which was probably
caused by a breakdown of large agglomerates, the shear Figure 11. Shear stress versus shear rate at spindle perimeter at
stress increased with increasing idle time while the 1300°C.
shear stress was constant with time.
Apart from the trends displayed in figure 8, the
shear stress exhibited small-scale oscillations of a
frequency ~0.025 s-1, and the amplitude ≥ 0.2 Pa. The
amplitude decreased with increasing temperature and
time. No oscillations occurred at T ≥ 1300°C.
Figure 10 displays the results of changing spindle
speed at a constant temperature after 9 min of constant
shear rate (with 5-min intervals between runs). The
sludge possessed a yield stress of 16 Pa and approached
Newtonian behavior at higher deformation rates. At
1300°C, the pseudoplastic sludge turned to virtually
Newtonian (figure 11), but as figure 9 shows, its
viscosity increased with idle time. Crystal
agglomeration probably occurred while the sludge was
at rest.
Figure 12. Shear stress versus time at 0.01 RPS.

Figure 10. Shear stress versus shear rate at spindle perimeter at Figure 13. Shear stress versus time at 0.025 RPS.
1100°C.

Ceramics − Silikáty 46 (1) 1-7 (2002) 5

 M. Jiřička, P. Hrma

a) a)

b) b)

Figure 14. a) Shear stress versus time at 0.025 RPS, b) Shear Figure 15. a) Shear stress versus time at 0.5 RPS, b) Shear
stress versus temperature at 0.025 RPS. stress versus temperature at 0.5 RPS.

temperature was increased back to 1095°C. Contrary to displayed in figure 14b. Figure 14b shows that the
expectation, the stress sharply increased when the response of the shear stress to a temperature decrease
temperature increased to 1144°C and then gradually was reversible, but a temperature increase caused an
decreased with time. After the return to 1095°C, the irreversible decrease in the shear stress.
shear stress was at a higher level than during the initial At higher temperatures (1250°C to 1349°C), the
run. The unexpected increase in shear stress at 1144°C shear stress increased with decreasing temperature
was probably associated with an increase in the volume (figure 15). As in the previous experiment (figure 14),
of the deformation zone, which resulted in a higher the shear stress increased with time at the minimum
stress resistance by the dense portion of the sludge. The temperature of this interval. At 1349°C, discontinuous
subsequent decrease of the shear stress was similar to decreases in the shear stress occurred, probably because
that of the initial run in figure 8. A similar sequence of of dissolving crystal agglomerates. The irreversible
events was observed at the higher spindle speed except response of the shear stress to a temperature increase
that the stress decrease was much sharper. was less pronounced as compared to the experiment
Figure 14 shows the outcome of a similar conducted at temperatures below 1200°C (figure 14).
experiment as in figure 13, performed at a later date (in
between the two experiments, the sludge cooled down Potential for sludge removal
to room temperature). The quiescent sludge was heat
treated at 1099°C for approximately 90 min, and then Sludge can accumulate during a melter's lifetime to
the shear stress was repeatedly measured at 1100°C a certain level without unacceptable risk. Sludge
with 0.025 RPS as shown in figure 10. The last of the removal may be considered if an unacceptable level of
runs is represented by the starting line (1099°C) in sludge accumulates in the melter before the end of the
figure 14a. When the temperature was subsequently campaign. Advanced melters may allow sludge to be
increased to 1148°C, the shear stress did not increase as regularly removed. Spinel sludge can be removed either
much as in the previous measurement (figure 13). The chemically or mechanically. Chemical removal can be
final shear-stress values at each temperature are accomplished by dissolving sludge in a borosilicate

6 Ceramics − Silikáty 46 (1) 1-7 (2002)

 Structural and mechnical response to a thermo-rheologic history of spinel sludge in high-level waste glass

melt, but this is a slow process that may take several References
weeks and generate additional waste. The effectiveness
of mechanical removal by gravity-driven flow through a 1. Plodinec M. J.: Advances in Ceramics 20, 117 (1986).
bottom drain depends on the inclination of the wall, the 2. Mika M., Hrma P., Schweiger M. J.: Ceramics - Silikáty
flow rate, the yield stress of the sludge, and sludge 44, 86 (2000).
viscosity. The yield stress of the sludge is low (16 Pa at 3. Govier G.W.: The Flow of Complex Mixtures in Pipes,
1100°C), but may increase with densification and aging. R. E. Krieger Publishing Company, INC., Florida 1972.
Because of the pseudoplastic nature of spinel sludge, a
faster flow-rate would facilitate its removal. Sludge,
once subjected to deformation, is likely to change its
structure from loosely bounded to consolidated. Denser STRUKTURNÍ A MECHANICKÁ ODEZVA
sludge will have different mechanical properties. NA TERMO-REOLOGICKOU HISTORII
SPINELOVÉ USAZENINY VE SKLE
CONCLUSIONS S VITRIFIKOVANÝMI VYSOCE AKTIVNÍMI ODPADY

Spinel sludge is a material with memory, and thus MILOŠ JIŘIČKA, PAVEL HRMA*
its behavior depends on its thermo-mechanical history.
Undisturbed sludge has a loose structure formed by Laboratoř anorganických materiálů,
slowly settling crystals. Mechanical disturbance by Společné pracoviště Ústavu anorganické chemie AV ČR
sheer deformation breaks down the original structure. a Vysoké školy chemicko-technologické v Praze
Structural changes continue after deformation when the Technická 5, 166 28 Praha 6
sludge is at rest. The response of the sludge to shear
may vary from thixotropic to rheopectic dependent on *Pacific Northwest National Laboratory, Richland,
the degree of structural alteration. Below 1300°C, the Washington 99352, U.S.A.
sludge is pseudoplastic. At T ≥ 1300°C, the spinel
content in the sludge drops to 15 mass% from the as- Pro studium struktury a složení spinelové usazeniny
received value of 28 mass%, and the sludge becomes získané po výhasu experimentální tavící pece pro sklo se simu-
Newtonian. lovanými vysoce aktivními odpady bylo použito optického
a skenovacího elektronového mikroskopu a rentgenové di-
Acknowledgments frakční analýzy. Při tepelných expozicích za isotermních
podmínek v teplotním intervalu od 1050°C do 1350°C se
The authors thank Gert Patello and Gary Smith for krystaly v matečné tavenině částečně rozpouštěly a při
providing the sludge sample and Jesse Alton for isolating spinel ochlazení se v tavenině tvořily malé (~ 10 µm) krystaly hvězdi-
from a sludge sample and helping with XRD calibration. They covitého nebo dendritického uskupení. Usazenina byla
are also grateful to Pavel Izak for preparing samples for opakovaně podrobena působení konstantní smykové rychlosti
analysis and Hong Li, Bruce Arey, Ben Wilson, Michael od 0,005 s-1 do 1,0 s-1 při konstantní teplotě od 1050°C do
Schweiger, and Jarrod Crum for laboratory support and 1350°C. Průběh smykového napětí usazeniny byl sledován
insightful advice. Miloš Jiřička is grateful to Associated v závislosti na čase při různých teplotách a rychlostech smyku
Western Universities for the fellowship appointment at Pacific a po různých časových prodlevách. Počáteční thixotropní
Northwest National Laboratory, which is operated for the U.S. charakter volné struktury usazeniny se vlivem opakované
Department of Energy (DOE) by Battelle under Contract DE- deformace a časové prodlevy změnil na rheopektický. Snížením
AC06-76RLO 1830. This work was funded by the DOE through koncentrace spinelu v usazenině z původních 28 hmot.% na 15
Environmental Management Science and Tanks Focus Area hmot.% za teploty 1300°C, se chovaní usazeniny přiblížilo
Programs. k chování Newtonské suspenze.

Ceramics − Silikáty 46 (1) 1-7 (2002) 7