TAN
UNCLASSIFIED
ORNL


.
.
11
.

.
W
,
"
AD
WW
"
.
1.
"
NA" ..

2
.
.
.
.
.
25
.
.
.:-
.
-
-

.
542
okre P-542
16!!!.
...
/
OCT 30 1904
--- EOAT NOTICE ---------------


ond then
there,
the nel
ooowww diam
CALORDETRIC INVESTIGATION OF THE THORIUM OX DDE-WATER INTERFACEY
8. F. Holmes
E. L. Fuller, Jr.
C. H. Secoy
Oak Ridge National Laboratory
Oak Ridge, Tennessee
ABSTRACT
A Calorimeter for the precise determination of heats of immersion
in solid-liquid systems has been developed. The all metal calorimeter
is of the isothermal Jacket type. The temperature sensing element 18
a thermistor whose resistance is continuously measured by a circuit
consisting of a Mueller bridge, a high gain breaker-amplifier, and a
recorder. Energy equivalents of the measured resistance changes were
determined by electrical calibration. For fast reactions the calorime-
ter bas a useful sensitivity of about 0.002 cal. The calorimeter is
capable of bandling five samples for a single assembly of the apparatus.
Performance of the calorimeter and utility of the results are demonstrate
ed with selected data obtained on the thorium oxide-water system.
*Research sponsored by the U.S. Atomic Energy Commission under
contract with the Union Carbide Corporation.
The phenomenon of heat evolution on immeroing a solid in a liquid
has been known since 1822. However, only qualitative conclusions could
be drawn from the carlier measurements as there was no acceptable method
for determining the surface areas of powdered and/or porous materials.
The BET method for the determination of surface areas provided the
foundation for quantitative results on a unit area basis. Following the
introduction of the BET theory the pioneering work or Boyd and Harkins
served to put heat of immersion measurements on a firm thermodynamic
foundation. With the advent of experimental innovations such as breaker-
type amplifiers and thermistors, heat of immersion measurements have been
applied to an increasing number of problems related to the solid-liquid
and solid-gas systems. From a thermodynamic viewpoint the most userul
information is obtained from coupling heat of immersion measurements with
adsorption isotherms for the same solid-liquid system.
The outstanding characteristic of -alorimeters used for heat of
immersion measurements is their sensitivity, rather than extreme accuracy
or precision. The small heat effect associated with immersing a solid in
a liquid demands high calorimetric sensitivity, for in many cases the
total observed heat will be much less than one calorie. Figure 1 is a
section view of the calorimeter used in our work. With the exception of
the stirrer and sample holder, which are brass, the entire calorimeter was
constructed of copper. In order to protect against corrosion and to
reduce radiant heat transfer, the interior and exterior surfaces of the
calorimeter were gold-plated. Inlet and exit connections to the cooling
21NL
coil were made with thin-walled rubber tubing. The "O" ringe shown in
Fig. 1 provided erfective air-tight seals as evidenced by no dirficulty
with ovaporation or condensation. The calorimeter stirrer was driven
directly by « 300 rpta synchronous motor. Total Internal volume of the
calorimeter (including metal parts) 18 300 cms. For immersion export.
ments, the calorimeter was loaded with 230 cm of water which brought
the water level to within about 1.5 cm from the lid. An immersion ex-
perimeat was initiated by manually lifting the sample holder assembly,
which broke a Bampic bulb against the sample breaker. Any one of the
i'ive samples could be selected w rotation of the sample holder Akoembly.
Vertical motion of the sample holder assembly was limited to about 8 mm
by a me cianical stop on the nylon section. This arrangement was
necessary to prevent breaking more than one sample bulb per experiment.
The calorimeter was suspended by four wylon rods inoide a brass
submarine jacket whoso inside surface was gold-plated. The calorimeter
was separated from the submarine jacket by a tour om dead air gap. The
ud of the submarine jacket was fitted with appropriate exit tubes for
electrical and mechanical connection to the calorimeter.
The constant temperature bath consisted of a well-stirred, insulat-
ed, 30 liter water bath operated at 25.0°c. Control was accomplished
by a nickel resistance thermometer used with a proportional controller
with reset action. This combination gave short-term control to better
than + 0.001° with long-term drift le88 than 0.001° per 24 hours.
A simpliried block diagram of the electrical circuito 18 shown in
F18. 2. The temperaturo sono ing element was a nominal 100 ohm bead
type thermistor enclosed in a glass probe. The thermistor was aged in
a 150°c oven for 400 hours. Following this, a current or 7.5 mA (rive
times the measuring current) was passed through the thermistor for 70
hours. Stability of the thermistor after this treatment has been
demonstrated by a constant calibration factor with no detectable long-
term drift. The thermistor was uocd as a three-lead thermometer.
A Leeds and Northrup model G-2 Mucller bridge served as the primary
resistance mcasuring instrument. Two volts from a 6-volt lead storage
battery served to supply the bridge current. The unbalance potential
of the bridge was amplified by a Beckman model 14 breaker amplifier.
The output of the amplifier was used to drive a zero-center Brown
recorder having a full scale range of 100 mv. In normal use the gain
on the amplifier was adjusted so that rill scale on the recorder was
5 x 10-3 ohm (approximately 1.6 x 20-3 °c).
The electrical calibration heater was 10 ohms of enameled 36 ga
Manganin wire. An identical heater placed in the constant temperature
bath served as the dummy beater. The heating circuit, which consisted
of a variable resistor, a 10 ohm standard resistor, and the calibration
heater, was powered by four 6-volt lead storage batteries connected in
paralel. Potential drops across the heater and the standard resistor
Tytt X-
were measured with a Leeds and Northrup K-3 potentiometer using a Leeds
and Northrup electronic null-detector.
Un
Duration of the heating period was measured with an electronic
frequency counter. The triggering circuit of the counter was connected
directly across the batteries and the calibration heater so that it was
energized with the same switch contact as the calibration heater.
Periodio checks of the accuracy of the Internal oscillator were made by
comparing it with the standard frequency broadcast by station www. In
normal calorimetric usage the counter was used to measure heating
periods up to 1000 seconds in duration with an accuracy of 0.001 sec.
Prior to an experiment, the temperature of the calorimeter was ad-
justed, f necessary, with the heater or cooling coil to give the
desired drift rate depending on the estimated temperature rise for that
experiment. Drift rates never exceeded 1.5 x 104 °C/min, and were
usually much less as the calorimeter was never operated more than 0.05°C
from bath temperature. Cooling of the calorimeter was accomplished by
blowing dry air through a coil inmersed in liquid nitrogen and then
through the calorimeter cooling coil.
The temperature rise for an experiment was evaluated by a simple
extrapolation of the fore and after drift rates. Electrical noise of
the recorder trace was about 1.8 x 1023 °C peer-to-peak and did not
contribute any uncertainty to the measurements. Total beat evolution
for an immersion experiment varied from about 1 to 50 joules, depending,
on the sample and the outgassing temperature. A total of three
corrections were applied to the observed heat from an immersion experiment.
.6.
These were the beat of bulb breaking, the heat produced by the mechani.
cal motion of the sample holder (~ 0.05 joulo), and the correction for
the vaporization of water into the previously evacuated void space in
the sampla bulb (~ 0.2 joule). The heat of vaporization correction was
calculated from the properties of water and the measured void volume in
the sampin bulb.
The other two corrections were determined experimental.
ly
Some characteristics of the loaded calorimeter and its associated
instrumentation are: a thermal leakage constant of approximately
2.5 x 10°3 min, attainment of an 18othermal condition internally with-
in one part in 20* two minutes after cessation of beating, and a short-
term sensitivity of approximately 0.008 joule. For a reaction period
of one hour, maximum uncertainty in the calorimetric results was no
more than 0.2 joule.
Six to ei.ght electrical calibrations and five heat of immersion
experiments were made for each loading of the calorimeter. In general,
the electrical calibrations always agreed to better than 0.1 per cent
and were independent of the rate of heating, the total heat input, and
the duration of the heating period. Reproducibility of the heat of
immersion results for a given sample was generally about 2 per cent.
Accuracy of the heat measurements was about one per cent for all of the
samples, being limited by the small temperature rise in the case of
samples with small specific surface areas and by the extended reaction
period in other cases.
The sample bulbs used for these measurements were blown from 4 mm
Pyrex tubing and had a diameter of 2 cm. 8ize of the bulbs was control.
lod by shaping them in a hemispherical graphite mold during the blowing
oporation. In this mannor the wall thickness of the bulbs was control.
Led between the limita of 3 to 6 mils. Those bulbs had an internal
volume of approximately 3.2 cm, were able to withstand vacuum outgas-
sing to 500°c, and were easily broken in the calorimeter.
The heat released on breaking an evacuated glass sample bulb has
been a source of difficulty in precise measurements of small quantities
of beat. To determine the magnitude and reproducibility of the cor-
rection for the heat of bulb breaking, a series of evacuated sample bulbs
were broken in the calorimeter. The average value, after correcting for
the heat of vaporization of water and mechanical motion of the sample
holder assembly, was 0.513 + 0.016 joule. The significance of this
correction varied Inversely with the specific surface area of the sample,
amounting to about 15 per cent for the sample with the smallest specific
surface area. However, the uncertainty in the correction was only twice
us large as the ultimate calorimetric sensitivity.
The thorium oxide samples used in tbis study were prepared by the
thermal decomposition of thorium oxalate. Four samples were prepared
from the same lot of thorium oxalate. Physical properties and impurities
for the thorium oxide samples are given in Fig. 3. It is apparent from
the geometric standard deviation that each of the samples bas a wide
.
1.
--
distribution of particle sizes. Specific surface areas of the samples,
as determined by nitrogen adsorption at 77°K, are much too large to be
related to the size of the particles. This has been attributed to the
oxide particles retaining the relic structure of the oxalate crystal
with the oxide particles consisting of much smaller crystallites or
thorium oxide. This 18 in agreement with the crystallite sizes quoted
10 F18. 3.
Sample pretreatment consisted of outgassing for 24 hours at tem-
peratures ranging from 100 to 500°c to a pressure of approximately
1 x 10 mm of mercury. Ultimate pressure was attained within one to
two hours after the sample reached outgassing temperature. Outgassing
temperatures were controlled to + 3°C and the samples were sealed off
under vacuum at the end of the outgassing period.
The beats of immersion of these samples of thorium oxide are shown
in Fig. 4 as a function of the outgassing temperature. The experimental
heat data were reduced to a unit area basis by means of the measured
nitrogen surface areas. Two important points are obvious from Fig. 4.
The first of these is the general increase of the heat of immersion
with increasing outgassing temperature. The second obvious point is the
fact that these four samples have remarkably different energetics with
respect to the solid-water interface. This is especially true with
respect to sample D, whose heat of lumersion 18 lower and much less
dependent on outgassing temperature than that of the remaining three
samples.
If the only function of the outgassing procedure was to remove
physically adsorbed water then the heats of immersion should be practi.
cally independent of the outgassing temperature over this entire range,
for there should be no physically adsorbed water remaining under these
outgassing conditions. A much more plausible explanation is that we
are, progressively removing more strongly bound chemisorbed water as we
increase the outgassing temperature. On subsequent exposure to liquid
water during the imersion process this chemisorbed water is replaced
with a large net heat of adsorption. Indeed heata of adsorption as large
as 45 kcal/mole have been observed for water on thorium oxide which was
dehydrated at 593°c. These high heats of adsorption are in line with
the generally beld belief that the surface of polar oxides 18 populated
with surface bydroxyl groups. It has been shown that temperatures as
high as 1300°c, under atmospheric conditions, are required to reduce the
surface hydroxyl content of thorium oxide to the point where they are no
longer detectable by infrared spectroscopy. This mechaniøm involving
surface hydroxyl groups may not be as operative in the case of sample D
because of the fact that the heat of immersion of this sample is relative-
ly Independent of outgassing temperature.
. There is no obvious correlation of the observed heats of immersion
with the physical properties given in Fig. 3. At outgassing tempera-
tures greater than about 250°C there 18, with the exception of sample B,
a decrease in the beat of immersion with increasing crystallite size or
decreasing surface area. Experimentally there 18 not enough difference
-10.
in the particle sizes to draw any correlations. Neither could impurities
account for the observed dirferences in the beats of immersion since all
of the samples were prepared from the same lot of material. As a matter
of fact, the impurity concentration 18 not large enough to account for
the observed differences even ir they were localized entirely in the
surface. The gross particles of the present samples are agglomerates of
smaller crystallites formed during the calcination process. It is quite
probable that in the present thorium oxide-water system the relative
heats of immersion are dictated by the relative proportions of the more
energetic crystalline edges and corners included in the available surface.
In view of the fact that the heat of immersion of these samples is
dependent on the outgassing temperature over the entire range of tempera-
tures studied it is interesting to look at weight loss measurements over
this same range of temperatures. Figure 5 18 the weight lose for sample
A as obtained with a vacuum microbalance apparatus. Some 16 to 20 hours
were required to reach a constant weight at each temperature with no
further weight change occurring after about 20 hours. This 18 excellent
evidence that our 24 hour outgassing period 18 sufficient to obtain an
equilibrium state. The most significant feature of Fig. 5 18 the fact
that there 18 no indication of the weight loss leveling out at 500°c. In
other words, a high vacuum at 500°C 18 not sufficient to remove all of
the water from the surface of thorim oxide. This fact supports the
postulated chemisorption mechanism.
-11.
The vast majority of reported beat of immersion measurements are
rapid proceuses, i.e., they are considered to be instantaneous within the
measurement apabilities of the calorimeters used. This is not the case
with samples A, B, and C of the present study. Pigure 6 shows the typi-
cal behavior observed as tweersing these samples in water. This is
essentially a seni-log plot of the unreleased heat as a function of time
after initiating the imersion reaction. Detectable quantities of heat
continued to be liberated for periods of time as long as 90 minutes in
the case of sample B. Linearity of the semi-log plots extended over
practically the entire period in all cases. Extrapolation of the semi-
log plots back to the time of Lumersing the sample gives a value which we
refer to as the slow heat of immersion. Half-lives for tube slow beats of
Imersian were 7.2, 10.7, and 5.3 minutes for samples A, B, and C, respec-
tively. Within experimental uncertainty, these ball-lives were independent
of the outgassing temperature. In contrast to this the extrapolated
values for the slow beats of immersion were dependent on the outgassing
temperature.
Figure 7 shows how the slow heats of immersion for samples A, B, and
C vary with the outgassing temperature. Within the experimental capabili.
ties, no slow heat was ever observed with sample D. The fraction of the
total heat of Lumersion which 18 released slowly varies from about 4 to
18 per cent and increases with outgassing temperature. This fraction 16
generally larger for sample B which again 18 out of line with respect to
surface area and crystallite size.
Slow heats of emersion have previously been reported for the
immersion of Alyo, and 510, la water. For the Algog-water maten the
quantity of slow heat and its half-life lacreased with outgassing ter
perature. Correspooding data were not even for the 810,-water syota.
For both systems the slow heat phenomena were attributed to a slow mo-
bydration of surface oxide groupe resulting from the outgassing procedure.
If the slow step in the lumersion process was rebydration of surface ord.de
groups one would expect the ball-life of the alow heat to vary with out-
sussing temperature. The fact that the ball-life of the slow beat in the
present system is independent of outgassing teaperature is evidence that
the rate controlling step is also independent of the outgassing tempen.
ture. In view of this we are lead to postulate that the rate controlling
step for the slow beat must be slow arrusion of water into the porous
structure of the thorium oxide particles. However, the dependence of the
N
magnitude of the slow beat on the outpussing temperature must be due to
the removal of chenisorded water from the laternal surface of the particles
during the outgassing process.
Much usefu laformation can be obtained by measuring the beat of
immersion for a series of samples containing komora amounts of preadsord-
ed water. Such data for sample A are shown in Fig. 8. Zero on the pre-
adsorbed water eis represents the 500°C sample weight in high vacuum
which, as we know trou the weight loss data, does not represent a water
free surface. Tvo features are immediately epident from this slide. There
1o a sharp decrease in the heat of Laersion with increasing coverage in
-13
.
the region of small coverages. This is indicative of a rather large
hoat of adsorption in this region. Then there is the levelling out of
the best of immersion at high coverages which indicates a rather small
boat of adsorption in this region. The horizontal line labeled by in this
slide corresponds to the surface energy of water at this temperature.
These beat of America data are on a unit area basis calculated by means
of the masured nitrogen surface area of 14.7 78. An important point
18, that if this thorium oxide sample, covered with a fila of water, actu-
ally had a surface area of 14.7 '/8, the heats of immersion could not
full below the surface energy of water. The obvious conclusion 18 that as
the porous structure of the sample is filled with water the effective
surface area decreases. The point at which the heat of immersion 18 equal
to the surface energy of water corresponds to an equilibrium water vapor
pressure of about 60 per cent saturation, so many of the smaller pores
should be filled.
This data also supply additional information concerning the mechanism
responsible for the slow heat of immersion. All of the samples corresponde
ing to equilibrium water vapor pressures less than 60 per cent saturation
bad a slow beat of immersion. Those samples corresponding to equilibrium
pressures above 60 per cent did not have a slow beat. Any chemisorption
process should be complete at much lower equilibrium pressures and much
lover surface coverages. So a physical process, such as dirrusion of
water into the small pores of the thorium oxide particles, must be respons-
Ible for the rate-controlling step for the slow heat.
-14.
By a simple differentiation of the data sbrua la Flg. 8 one can
obtain the net differential beat of adsorption of water on this sample
of thorium oxide. These quantities are shown in Fig. 9 as a function of
the amount of water on the surface. Once again, I should 11kce to empha-
size that zero on the x/m axis le relative to the 500°C high vacuum
sample weight and does not represent zero water on the surface. Data
such as shown in this slide are usually plotted as a function of the
fraction of the surface covered so that inflections in the heat data can
be correlated with the completion of a monolayer, etc. We would also
like to do this but, as yet, we do not know where to place our zero. The
large heat values obtained with small quantities of water on the surface
are much too large to be attributed to any physical proce88, that 18, the
large net heat of adsorption must be due to chemisorption.
Also given in this figure are the 1808teric heats of adsorption cal-
culated from the adsorption branch of isotherme we have determined with a
vacuun microbalance. These 18otherms contained bystersis loops and thermo-
dynamic data derived from such isotherms are highly questionable. Never-
theless, if for comparison purposes only, we have calculated the 1Bosteric
beats of adsorption. These values have been corrected for the heat of
vaporization of water to make them comparable to the calorisetric data.
Completion of the first physically adsorbed layer occurs at about 12 mg of
water per gram of thorium oxide. In the region where there 18 physical ad-
sorption only, there 18 quite good agreement between the calorimetric and
.....
....
......
-15-
180steric heats. The points which deviate from the curve occur at equi.
librium pressures of 20 microns or less. This, plus the fact that the
sample weight in high vacuum 18 dependent on temperature in the vicinity
of 25°C introduces a large amount of uncertainty in the 180steric heats
in this region. Heats of adsorption derived from the calorimetric data
are not subject to the limitations and uncertainties associated with
the 18osteric beats.
These studies are currently being extended to other types of thorium
oxide material in order to explore the effect of substrate structure in
the thorium oxide-water system.

t
*
A
TLY WHY
w
sij
i
nerie
"
.
UNCLASSIFIED
ORNL-LR-DWG 74139
"
..

NYLON SECTIONS-
.
.
.
. .
.
. .
"O" RINGS
.
.
.
STIRRER-
-COOLING COIL
SAMPLE BREAKER-
-THERMISTOR IN
WOOD'S METAL
HEATER IN OIL-
SAMPLE HOLDER
ASSEMBLY
SAMPLE BULBS
(5 TOTAL)
5 cm.
;
.
Color
Calorimeter for Heat of Immersion Studies.

R
UNCLASSIFIED
ORNL-LR-DWG 74140
$-tu

10
ERMISTOR
!
_
__
G-2
MUELLER BRIDGE
BREAKER
AMPLIFIER
RECORDER
.
.
.
-
-
TEMPERATURE MEASURING CIRCUIT
. VA
-
*+1
.

7.
T
-n
Hilitha
"
FOUR 6-VOLT
STORAGE BATTERIES
AVMX
Holila
..:-*-*-os
.
..
..
...
,
.
: 31012
HEATER
1012
DUMMY
.
ins
.
ELECTRONIC
COUNTER
K-3
POTENTIOMETER
NULL
DECTECTOR
eta-sanmitor..
1013
STOS
. M
.
i'w.
0-12 *52
ELECTRICAL CALIBRATION CIRCUIT
Instrumentation for Heat of Immersion Studies.
10
-
-
-
*..'.
UNCL ASSIFIED
ORNL DWG. 64-4704
PROPERTIES OF THO2 SAMPLES
SAMPLE
CALCINING TEMPERATURE (°C)
CALCINING TIME (hr)
NITROGEN SURFACE AREA (m2/g)
GEOMETRIC MEAN PARTICLE DIAMETER (microns)
GEOMETRIC STANDARD DEVIATION
CRYSTALLITE SIZE (angstroms)
A
650
4
14.7
2.63
1.37
194
B
800
4
11.5
2.07
1.37
220
C
1000
4
5.64
2.72
1.38
682
1200
4
2.20
2.97
1.39
1700
IMPURITIES (in ppm): F <10; NO3 <10; SO4, 140; PO4, 15; Si <10; CI < 10; Fe < 10; Ni <10;
Cr <10; Pb <10; Na, 20; K <10; Li <10; Ca, 51; Ba, 0; AI, 2
-
*
-
;
..
*
***
*
*
*
**..'.
i
A.,
w
UNCLASSIFIED
ORNL-DWG 64-4705

HEAT OF IMMERSION (ergs/cm2)
500
100
200
300 400
OUTGASSING TEMPERATURE (°C)
Heat of Immersion of Tho, in Water at 25.0°C.
UNCLASSIFIED
ORNL-DWG 64-2021

WEIGHT LOSS (%)
O EQUILIBRATED AT LEAST 16 hr
•3 TO 6 hr EQUILIBRATION
100
400
500
200 300
TEMPERATURE (°C)
-
-
V
- 1.
-
-
Vio
.
YA
UNCLASSIFIED
ORNL-DWG 64-4706

HEAT RELEASED AFTER TIME (ergs/cm2)
0 4 8 12 16 20 24 28
1, TIME AFTER IMMERSING SAMPLE (min)
Time Dependence of the Slow Heat of Immersion of Tho2
in Water at 25.0°C (Samples Outgassed at 450°C).
E
.
.
--amaren med henvenuto ****
ORNL-DWG 64-4707
SLOW HEAT OF IMMERSION (ergs/cm2)

.
. .
.
-
-
-
-
-
200
100
300
400
500
OUTGASSING TEMPERATURE (°C)
Slow Heat of Immersion of Tho, in Water at 25.0°C.
WU
Wim
W
'
MW
-
.
.
W
.
.
.
YUNI
UNCLASSIFIED
ORNL - DWG 64-3726
1200

1900
1000
O BY DIRECT DOSING
• BY EQUILIBRATION WITH
KNOWN H2O PRESSURE
HEAT OF IMMERSION (ergs/cm?)
300
TO 5 10 15 20 25 30 35
PREADSORBED H20 (mg of H20/9 of Thoz!
Heat of Immersion of DT-37-100 (650) Tho, at 25°C.
( With Preadsorbed H2O)
**
*
samen menys
in
the risert for
at
ku
.
.)
Immunsystem
UNCLASSIFIED
ORNL - DWG 64-3730
x/m (mg of H20% of Thoz!

NET AH (kcal/mole)
--
--
-
-
--
-
-
-
-
T9-egte
p
olo
2 4 6 8 10 12 14 16 18
x/m (mg of H20/9 of Thoz!
Net Differential Heat of Adsorption of H2O on DT-37-100
(650) Tho, at 25°C.
.
DATE FILMED
12/ 30 164

I.
-
. -
P
1. UNTU
..'
+
- LEGAL NOTICE -
This report was proporod as an account of Government sponsored work. Neither the United
States, nor the Commission, nor any person acting on behall of the Commission:
A. Makos any warranty or reproson La tion, expronsod or implied, with respect to the accu-
racy, complotonors, or vuofulness of the information contained in this report, or that the uso
of any information, apparatus, method, or procon disclosed in this report may not Infringe
privately owned rights; or
use of any information, apparatus, method, or process disclosed in this report.
As used in the above, "person acting on behalf of the Commission" includes any em-
ployee or contractor of the Commission, or omployee of such contractor, to the extent that
such omployee or contractor of the Commission, or omployee of such contractor prepares,
disseminates, or provides accos. to, any information pursuant to his omployınent or contract
with the Commission, or his omploymont with such contractor.
Marcadores
END