United States
Environmental Protection
Agency
Environmental Sciences Research
Laboratory
Research Triangle Park NC 27711
EPA-600 3 79 077
August 1979
Research and Development
Vapor Pressure and
Melting Behavior of
Sulfuric Acid-Water
Systems
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development. U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are.
1 Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4 Environmental Monitoring
5 Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7 Interagency Energy-Environment Research and Development
8 "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ECOLOGICAL RESEARCH series. This series
describes research on the effects of pollution on humans, plant and animal spe-
cies, and materials. Problems are assessed for their long- and short-term influ-
ences. Investigations include formation, transport, and pathway studies to deter-
mine the fate of pollutants and their effects. This work provides the technical basis
for setting standards to minimize undesirable changes in living organisms in the
aquatic, terrestrial, and atmospheric environments.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/3-79-077
August 1979
VAPOR PRESSURE AND MELTING BEHAVIOR OF
SULFUR 1C ACID-WATER SYSTEMS
by
G. Raymond Brown and V. Rao Veluri
Clark College
Atlanta, Georgia 3031^
Project Officer
Jack Durham
Chemistry and Physics Division
Environmental Sciences Research Laboratory
Research Triangle Park, N. C. 27711
Environmental Sciences Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, N. C. 27711'
-------
DISCLAIMER
This report has been reviewed by the Chemistry and Physics Division,
Environmental Science Research Laboratory, of the U.S. Environmental Protec-
tion Agency and approved for publication. Approval does not signify that
the contents necessarily reflect the views and policies of the U.S. Environ-
mental Protection Agency, nor does mention of trade names or commercial pro-
ducts constitute endorsement or recommendation for use.
!i
-------
ABSTRACT
An experimental apparatus was designed and constructed to utilize high
vacuum and mass spectrometric techniques to determine total and partial vapor
pressures above bulk liquid samples in the temperature range between -65°C
and 25°C.
Observations on the sulfur5c acid-water system revealed complexities
which interfered with the experimental goal, but which are of some interest
in themselves. These interfering processes included long internal thermal
relaxation times and chemical reactions. An unexpected and most interesting
observation was an apparent pressure effect on the melting curve of the sul-
furic acid-water mixtures. The main result of this study is a description
of this apparent pressure effect.
This report was submitted in fulfillment of Grant Number R8QM70 by the
Department of Physics of Clark College under the partial sponsorship of the
U. S. Environmental Protection Agency. This report covers the period Novem-
ber 1, 197^ to August 31, 1977, and the work was completed August 31, 1977-
111
-------
ACKNOWLEDGEMENTS
The authors would like to acknowledge many helpful conversations with
Dr. Grayson H. Walker. Dr. Homer Utley cheerfully provided help whenever
the enterprise needed an extra hand or two. Mr. Karl Varner assisted in sam-
ple preparation and data collection. The manuscript preparation was ably
handled by Ms. Doris Goodwin. The managerial and administrative aid provided
by Dr. 0. P. Puri was gratefully appreciated.
tv
-------
SECTION 1
INTRODUCTION
The main purpose of this research was to directly measure the very low
vapor pressures of fluid mixtures which, in low concentrations, form aerosols
in the ambient atmosphere. The prime example of such a system, and the one
deemed most important in this study, is the mixture of sulfuric acid and
water. As originally submitted, our proposal for this work foresaw success-
ful measurements on the sulfuric acid-water system (subsequently h^SOjj - ^0)
and sufficient understanding of it to allow for expanded studies of other
systems, including other inorganic acids as mixture components, and even of
organic compounds such as are found in automobile exhausts. Unavoidable
delays in completion of the experimental apparatus, and unexpected complexi-
ties in the h^SOj, - H£0 system, precluded attainment of this program. In the
following, attention is restricted to the single system upon which measure-
ments were actually made in the apparatus, that being H2SOij - H20.
An extensive 1iterature^~9 attests to long scientific interest in the
chemistry and thermodynamics of H^SOjj - 1^0. However, no direct measurements
of vapor pressure of h^SOij - H20 in the interesting region near 100% ^SO^
have been published. This is not very surprising, due to the extremely low
value of this pressure and the difficulty of accurately measuring low pres-
sures. The 1970's brought increasing attention to the role of the formation
of H2SO/J and other sul fates in air pollution. This provided the incentive
for new attempts to directly measure the relevant physical-chemical processes
involved.
Thus followed the present attempt to design an apparatus by which the
expected very low values of the equilibrium vapor pressures could be measured
above the surface of bulk h^SO^ - H20. Because pressure gradients are easily
established along tubes and across apertures at the low pressures being con-
sidered, a vacuum chamber was designed so that the pressure-measuring devices
(nude ion gage, capacitance manometers, and the ion-source region of a mass
spectrometer) would be located physically as close as possible to the sample.
Especially, it was desired to minimize obstructions to vapor motion between
the sample and the measuring devices. This consideration, as well as the
requirement of sample cooling equipment in order to study the temperature
behavior, resulted in a rather large (^ 0.06 m^) vacuum chamber. It was
understood at the outset that this large a vacuum chamber could result in
relatively long equilibrium times. However, this was outweighed by the resul-
ting simplicity of design and minimization of pressure-gradient and mass-
discriminating effects in the apparatus.
-------
It may be noted at this point that a Knudsen-cel1-type design was also
considered. (in a Knudsen cell the sample Is completely enclosed except for
ti pin hole through which vapor may escape. Under low-pressure molecular flow
conditions, the Intensity of the molecular beam exiting the pin hole Is re-
lated to the vapor pressure of the sample). This was rejected as leading to
a design of considerably greater complexity than the present one. Also, the
Knudsen pin hole Is Itself a mass-discriminating device, and thus corrections
would have to be applied In order to estimate the partial pressures of the
mixture systems.
Other criteria which the design was to meet were the following: (1) the
ability to measure total pressures In the range from atmospheric pressure
(^ 100 kPa) to the lowest estimate as found In the literature for the vapor
pressure of pure H2SOj, (^ 10~^ - 10~5 Pa), (2) temperature measurement and
control from approximately room temperature (25°C) to the low temperatures
found In the stratosphere (^ -65°C), and (3) the ability to detect and measure
the relative concentrations of various mixture components In the vapor. For
this a mass spectrometer was Incorporated Into the design. Construction of
the experimental apparatus began with receipt of the first of the Instruments
ordered In May of 1975.
Reliable operation of the entire apparatus In Its final form was attained
In May of 1977* Most of the experimental data reported here were obtained
during the period from May to August, 1977. Exceptions are certain prelimi-
nary measurements and calibration measurements which were accomplished prior
to this period.
-------
SECTION 2
CONCLUSIONS AND RECOMMENDATIONS
The main purpose for which this apparatus was constructed, namely the
direct measurement of the total vapor pressures and the partial equilibrium
vapor pressures of sulfurlc acid-water mixtures (H2S01* - H20) has proved to
be an elusive goal. Several factors contributed to this lack of success, but
one major reason Is the presence of a very Interesting but completely unex-
pected complication: an apparent pressure effect on the melting curve of
H2S01, - H20.
The strongest evidence for this pressure effect on the melting curve was
seen In the h^SOlj - HgO samples at concentrations near that of the monohy-
drate (84.5 wt$ h^SOlJ. In particular, the experimental runs of July 7 and
July 11, 1977 Indicate a reproducible difference of 15°C between the melting
temperature of this mixture at near-atmospheric pressures (60 to 80 kPa) and
pressures of the order of 1.0 Pa. Reproduclbl1Ity of the melting point reduc-
tion upon temperature and pressure cycling shows that the effect cannot be
the result of distillation In the pumping process, which would lead to changes
In the composition of the sample. In particular the reproduction of the "high
pressure melting point" following an observation of the "low pressure melting
point" In the sample of July 11, 1977 Indicates that the mean concentration
of the sample was not significantly changed during that day's run.
The observations of melting In samples with nominal concentrations of
96.5 and 100 wt* h^SOj, are not Inconsistent with the existence of a possible
pressure effect on the melting curve. However, the conclusions In these cases
are made ambiguous by the effects of persistent supercooling, possible dis-
solved gases, and possible, as yet unidentified, chemical reactions. The
latter two effects are also major obstacles to overcome If the original Inten-
tion of measuring the vapor pressure In this system Is to be attained.
It Is the opinion of the authors that at this point the most pressing
problem Is the further elucidation of the nature of the apparent pressure
effect on the melting curve In h^SOlj - H£0. The existence of the effect
requires confirmation and corroboratIon by further experiment, and a detailed
determination of the extent of melting point lowering with pressure Is Impera-
tive. Assuming for the moment that the apparent pressure effect on the melt-
Ing curve Is corroborated by further experiment, then the thermodynamlc conse-
quences of the effect must be explored. This would Include possible extensive
modification of calculations such as those by Gmltro and Vermeulen.5
-------
It may, of course, prove to be the case that this apparent effect Is due
to impurity gases, supercooling effects, or is in some other way an unfore-
seen experimental artifact. Even so, the course of the effect must be resol-
ved in order that further progress be made towards a direct measurement of
the vapor pressures in the system.
Results of the mass spectrometric observations on this system remain at
this point purely qualitative, and even somewhat tentative. However, this
has not been a major limiting factor in the experimental analysis. Further
work In this direction appears to hold some promise in elucidating the physi-
cal and chemical behavior of the H2S04 - H20 system.
-------
SECTION 3
DESCRIPTION OF THE APPARATUS
GENERAL
A sketch of the experimental apparatus, showing the spatial relationships
between the various measuring instruments, is given in Figure 1. In the text
below, capital letters in parentheses refer to items identified in Figure 1.
A sample dish containing the h^SO^ - H20 liquid is supported inside the
vacuum chamber by a stainless steel, temperature-controlled table (D) . This
table is cooled and its temperature controlled by chilled alcohol from a cir-
culating bath refrigerator (Neslabs Model LT-9) . The chilled alcohol enters
the vacuum chamber through the feedthrough (E) , is routed to the interior of
the cooling table via stainless-steel tubing, circulates through the interior
of the cooling table, is drained by the re-entrant section of the same feed-
through (E) , and is returned to the circulating bath refrigerator- This sys-
tem provided temperature control to within ± 1°C throughout the temperature
range -60°C to -HO°C. During the period in which most of the data reported
here were taken, operation was limited to a lower temperature of ~35°C. This
was due to a malfunction of the circulating bath refrigerator followed by an
improvised repair. The decision was made to operate the system within the
limited temperature range rather than suffering a six- to eight-week hiatus
while the refrigerator was returned to the factory for proper repair. The
cooling table itself was the source of considerable difficulty, in that the
first model of this table proved to be of poor quality. The facilities of the
shop where the table was made were insufficient to insure that the extensive
welding required was vacuum-tight. After a few thermal cycles, persistent
and ever more numerous leaks developed in this table. Eventually it was
necessary to contract with the Georgia Institute of Technology Research shop
to construct a second model of the cooling table. The second model proved to
be quite adequate, and provided trouble-free service.
The temperature of the cooling table was monitored by a chromel-gold
(2% iron) thermocouple, which is the sensing element in a Lakeshore Cryo-
tronics, Inc. Model DRC-5 thermometer. This device incorporates an internal
electronic standard which makes its use extremely convenient and trouble-free
in this application. (I.e., no reference junction maintained at a fixed-point
temperature is needed.) This thermometer could detect temperature changes of
± 0.2°C, and was stable for short periods (^ 30 minutes) to within ± 0.5°C.
The thermocouple itself was embedded within epoxy in a blind hole 3-2 mm in
diameter and 3.2 mm deep drilled in the stainless steel cooling table. The
epoxy provided a mechanically strong, thermally conducting and electrically
-------
A Vacuum Valve
B To Mass Spectrometer
C Pyrex Sample Dish
D Temperature Controlled Table
E Table Coolant Inlet
F Thermocouple
G Ion Gauge
H Manometer 0 - 133 Pa
J Manometer 0 - 133 kPa
K Bell Jar
L Thermocouple Vacuum Gauge
Figure 1. Experimental Apparatus.
-------
insulating bond between the thermocouple and the cooling table.
The vacuum chamber was pumped via a pumping line 10.16 centimeters in
diameter by a CVC Products, Inc., Model PMC-4B diffusion pump. The pump was
charged with polyphenyl ether oil, and the pumping line contained a liquid-
nitrogen-cooled baffle. This system was capable of producing a bell jar
pressure of 10~6 Pa when the freshly cleaned vacuum chamber contained no sam-
ple and was pumped overnight.
Four independent pressure-sensing devices were used to monitor the total
pressure inside the bell jar. One of these was the thermocouple vacuum gauge
(L), which was used only to monitor the operation of the vacuum pump, and
which played no part in the experimental measurements.
The primary instruments for total pressure measurements in this experi-
ment were two capacitance manometers. These devices sense the changes in
capacitance when the system pressure flexes a thin metal membrane which forms
one half of a flat-plate capacitor. This type of manometer is directly sen-
sitive to force per unit area. Thus it measures pressure directly, and is
not sensitive to the composition of the gas or mixture of gases which produce
the pressure. The manometers used here are the Baratron Model 145A, manufac-
tured by MKS Instruments, Inc. Two manometer sensing heads were used, a high-
pressure one (J) for the range 0-133 kPa and a low-pressure one (H) for the
pressure range 0 - 133 Pa. Pressure differences of 2 x 10"5 of the raage
could be detected with these instruments, and the readings were reproducible
to within 50% at 10"^ of the range. Thus the two sensing heads provided
direct measurement of the pressure throughout the range from atmospheric pres-
sure to 10"3 Pa.
Finally an ionization gauge (G) inside the bell jar allowed measurements
of total pressures below 10~^ Pa to the ultimate vacuum attainable within the
system. This ionization gauge is "nude", i.e., the filament, collector and
heater that constitute the sensing element project from the wall of the bell
jar to the region near the sample, rather than being enclosed in tubulation
projecting out from the bell jar wall. This construction was used in order
to minimize any possible pressure gradient between the sample and the ion
gauge.
The ion gauge has the advantages of excellent sensitivity, precision and
reproducibi1ity. Unfortunately, its response varies with the chemical iden-
tity of the gas. This leads to large uncertainties (factors of two to ten)
in the absolute pressure as measured by this instrument for different gas
mixtures. It was our original intention to calibrate the response of the ion
gauge to each gas mixture used in this experiment by relating the ion gauge
readings to those of the low-pressure capacitance manometer where their
ranges overlapped; i.e., 10~2 - 10~3 Pa. During our study of the f^SOlj - f^O
system the necessity for this procedure never arose, as pressures in the
range below 10"^ Pa were not attained while the sample was in the system.
In order to detect and measure the concentrations of the various chemi-
cal species in the vapor above our samples, a mass spectrometer was made a
part of the apparatus. The sensing element of the mass spectrometer (B)
-------
could communicate with the sample space through a stainless steel tube with
an inner diameter of 'v. k.$ cm and a length of ^ 30 cm. A large valve (A)
.controlled communication between the mass spectrometer and the bell jar.
(This valve, which is referred to as Valve "A" below, has a maximum aperture
of * 5 cm.) The tube connecting the bell jar and the mass spectrometer was
fashioned to be as short and as large in diameter as was physically possible,
again to minimize pressure gradients. This tube could be heated to reduce
the possibility of condensation between the bell jar and the mass spectrom-
eter.
The mass spectrometer itself is a time-of-f1ight instrument, Model MA-
3A, manufactured by CVC Products, Inc. It has a completely independent
vacuum pumping system, lonization is accomplished by electron impact In the
source region of the instrument; the Ions are then accelerated by an electric
field and allowed to drift. More massive ions are accelerated less by the
field, and so the ions are segregated by mass as they drift. High speed
detection and counting of the ions thus provides a mass spectrum of the ori-
ginal gas. This instrument is sensitive to molecules in the mass range from
one to approximately six hundred atomic mass units, and has a resolution of
- 300 at m - 200 AMU.
Two systems were used to monitor the output of the mass spectrometer
analyzer: an oscilloscope in one channel, and an electronic scanner with a
chart recorder in another channel.
The oscilloscope used was a Tektronix Model 5^03. The oscilloscope was
used to provide an easily observed overview of the mass spectrum. This capa-
bility is a requirement for convenient tuning of the mass spectrometer. Also
it allowed one to visually, and therefore quickly, note regions of the spec-
tra of particular interest, occasional transient phenomena in the spectra,
and immediate indication of anything wrong with the mass spectrometer. The
oscilloscope channel is not convenient for providing a permanent record of
data.
The other channel, that containing the scanner, utilizes a gate to se-
lect a small segment of the total mass spectrum from each sweep of the mass
spectrometer. (The instrument produces approximafly 30,000 "sweeps" each
second during operation.) The width of this gated segment was usually set to
be a small fraction of the width of a single mass peak. The signal in the
gated segment could be Integrated over many sweeps of the spectrometer, thus
providing excellent sensitivity. Controls on the scanner allow the gate to
be moved through the spectrum at will, and for the gate width, scan rate and
scan range to be varied. Thus any part of the mass spectrum can be closely
scrutinized using the scanner. This scanner is a CVC Products, Inc., Model
010. The signal from the scanner is amplified within the scanner itself and
displayed on a meter as a direct-current voltage. This voltage is also sup-
plied to a chart recorder, in this case an L S N Model 625, to provide a per-
manent record of the mass spectra as scanned by the instrument.
-------
INSTRUMENT CALIBRATIONS
Since the available facilities for calibration of pressure and tempera-
ture sensors was extremely limited, care was taken to procure only those
devices which could be delivered with a factory-certified calibration. This
was one consideration in buying the thermocouple thermometer and the capaci-
tance manometers. (The ionization gauge could not be so certified, by the
nature of the instrument. Our intended procedure for calibrating this
instrument with the low-pressure capacitance manometer is briefly described
above. As noted there, the necessity for this calibration did not arise in
the experiment).
The thermocouple thermometer contained an internal electronic standard
which related the potential difference produced by the thermocouple to the
Kelvin temperature, which was digitally displayed by the instrument. This
standard could however be "shifted" by a bias voltage. Before installation
into the apparatus this voltage was set to provide a reading of 273-12 ±
0.50K when the thermocouple was submerged in an ice-point bath. This bias
setting was then checked by dipping the thermocouple into liquid nitrogen,
and noting that the thermometer displayed the correct value of this tempera-
ture as 77 K.
Installation of the thermocouple sensor into the vacuum chamber involved
two solder joints in each of the thermocouple leads. After the thermocouple
was installed It was noted that readings of room temperature by the thermo-
meter appeared to be too high by some three or four degrees, as compared to
a mercury-in-glass thermometer. This shift was confirmed by data taken in
the system on the vapor pressure of pure water- The pure-water runs are
described in greater detail below. It was found that the initial readings
of the vapor pressure of pure water as a function of temperature would agree
with the accepted values only after a downward shift of the recorded tempera-
tures by 3-8 degrees. The thermometer bias voltage was adjusted to provide
this shift. Subsequent measurements of the pure-water vapor pressure were
always found to agree with the accepted values.
The pure water vapor pressure runs were also confirmation of the cali-
bration of the capacitance manometers. These manometers did show a small
day-to-day drift in the reading for "zero" pressure. This drift was typi-
cally 3 x 10"^ of the instrument range. Small daily adjustments of a bias
voltage to reset the zero for these instruments was part of the routine
during data collection. So long as care was taken to reset the zero read-
ings, no discrepancy was found between the two capacitance manometers when
they were each exposed to the same system pressure. This could, of course,
only be noted within the range of the low-pressure capacitance manometer.
-------
SECTION k
EXPERIMENTAL MEASUREMENTS
PURE H20 RUNS; AND GENERAL SAMPLE HANDLING PROCEDURES
As a final check out of instrument calibrations and general system per-
formance several data runs were made on the pure water system. The pure
water system is also, of course, an important limiting case of the H2SO^ -
H20 system, and thus it was important to measure the ability of the apparatus
to reproduce the well-known vapor pressure curve of water.
Most of the sample handling techniques were similar for all sample runs
and that will be described here.
Pure distilled water (Fisher Reagent W-2) was poured into a clean sample
dish. The sample dish and the sample were then immediately lowered to the
surface of liquid nitrogen in a dewar, thereby becoming solidly frozen at a
temperature of 77 K. At this temperature the vapor pressure of water is
negligibly small, and that of HjSOj, - h^O is even considerably less than that
of pure water- Thus a frozen sample at this temperature could be safely
pumped from atmospheric pressure to high vacuum with no danger of distilla-
tion causing changes in the sample composition.
Meanwhile the cooling table inside the bell jar was refrigerated to its
lowest temperature (-65°C in early runs and -35°C after difficulties devel-
oped with the refrigerator). The pressure inside the bell jar was increased
to atmospheric pressure by venting with high purity dry nitrogen gas. The
dry nitrogen gas was allowed to continue flowing into the bell jar as the
vacuum space was opened to permit the admission of the experimental sample.
Thus a slight dynamic over pressure of dry nitrogen gas was maintained in the
vacuum chamber volume as the sample was introduced. This procedure minimized
the introduction of ambient air into the system along with the sample. There
was particular concern not to introduce atmospheric water vapor with samples
containing high concentrations of K^SOi,, since these samples are extremely
hygroscopic.
Thus a solidly frozen sample was introduced under an inert, dry nitrogen
atmosphere and placed on the cooling table at its lowest temperature. The
bell jar was then closed and the vacuum chamber evacuated as quickly as pos-
sible.
The system was pumped down until the pressure was well below that expec-
ted as the equilibrium value for the sample at the cooling table temperature.
10
-------
When this point was reached, the pumping aperture was closed, and the system
was allowed to come to pressure equilibrium. Here "pressure equilibrium" was
operationally defined as having been reached when no change was noted in the
pressure reading during a reasonable observation time - generally five to ten
minutes. Typically then the reading of the pressure and temperature would
be recorded, and the temperature of the cooling table increased by whatever
increment was desired to reach the next intended temperature.
By this process we were able to reproduce the accepted vapor-pressure
curve of pure water throughout the temperature range of -30°C to +15°C,
corresponding to vapor pressure between J»0 Pa and 1500 Pa. Data from one
such pure water run are shown in Fig. 2.
Long equilibrium times were of considerable concern in the experiment.
In the measurements on the pure water system it was found that for the higher
pressures (above 1 kPa) , equilibrium times as observed in the apparatus were
always negligibly short (< 100 sec). A limiting factor in determining the
high pressure equilibrium times was the fact that the system temperature
would respond to small changes !n the room temperature. At the lowest pres-
sures in the pure water runs, the equilibrium times became quite long, on the
order of one hour. As a result of the experience with the pure water system
in the apparatus, we were prepared to find that long equilibrium times would
be a limiting factor in our ability to accurately measure total vapor pres-
sure in the h^SOjj - 1^0 system, particularly at high concentrations of ^SO^.
However this did not prove to be the case. As described below, complexities
within the H2SO^ - H20 system itself proved to be the limiting factors.
PRESSURE -TEMPERATURE RUNS ON SAMPLES NEAR 100$ H2SOj,
Two commercially available formulations of H20 - H^SOjj were used to pre-
pare the samples. These are:
Fisher A-300 : 95 - 98 wt* of H2S04 and
Fisher A-30*» : Sulfuric Acid Fuming, 15-18 wtfc excess SOj
In the case of H2SOjj - H£0 it is reasonable to talk of mixtures contain-
ing a weight percentage of H2SO^ greater than one hundred. This can come
about due to the fact that the acid molecule can undergo the dissociation:
*• H20 + S03
Removal of f^O molecules increases the concentration of the acid, and removal
to the point of an excess of SQy corresponds (by convention) to a concentra-
tion of H2SO/, in excess of 100%. The weight percentage of H2SO/j is then
always simply related to the weight percentage of SO^ by the ratio of the
molecular weights:
Wt* HjSOj, = (Wt* SO.) (98.082/80.066)
The A-304 commercial preparation can be seen to have a mean composition of
84.7 wtlfc SOj, from which we determine the H2SOJf concentration:
11
-------
ro
1.4
1.2
1.0
.8
0)
(0
0)
0)
CL .4
.2
— = Handbook of Chemistry
and Physics 58 th Ed.
• = present work
250
260
270
280
Fioure 2. Vaoor Pressure of Water,
-------
Fisher A-304 : 103-104 wt% H2SO|,
Thus, samples with concentrations in the range 95 < wt% H2S04 < 104
could be produced by mixing A-300 and A-304 together. Since the variation of
specific gravity between these two concentrations is negligible10, a
straightforward volumetric mixing can produce any desired concentration in-
termediate between A-300 and A-304.
Samples were prepared in this way such that 99 < Wt% h^SO^ < 101. These
are referred to below as "100% H2SOj,"- The actual mixing was accomplished
utilizing calibrated pipets to prepare approximately 2-ml samples of the
desired mixture. As these high concentration samples are extremely hygro-
scopic, care was taken to minimize the time during which they were exposed to
air. Typically the entire process of mixing the sample and getting it under
a dry nitrogen atmosphere at 77 K required less than one minute.
Pressure-temperature runs were performed on A-300 samples and on 100%
H2SOjt samples and these are described in this section. Observations using
the mass spectrometer are given at the end of this section.
The initial runs on the A-300 and 100% H2S04 samples produced very
similar results, and these will be described together. A solidly frozen sam-
ple of 100% H2SOj, at 77 K was transfered under dry nitrogen atmosphere to
the vacuum chamber cooling table, which was maintained at a temperature of
240 K. The bell jar was then quickly lowered to close the vacuum chamber
and the vacuum space was evacuated as quickly as possible. Typically it was
found that the vacuum system could not be pumped down below pressures less
than about 0.4 Pa. If this figure were to represent the equilibrium vapor
pressure It is surprisingly high, approximately two orders of magnitude lar-
ger than the estimates of Gmitro and Vermeulen for H2SO/j - H20 in this con-
centration range. However the other observations described below strongly
indicate that this can not be interpreted as an equilibrium vapor pressure.
As the pumping line was shut off, the pressure in the bell jar was noted
to quickly rise to the neighborhood of 1.0 - 3-0 Pa, and then to more slowly
increase at a rate not perceptively different from the leak-up rate observed
when the vacuum chamber was empty of the sample. The solid sample, mean-
while, showed signs of local melting. In fact, bubbles were always noted
which would rise and burst at the surface. This was not a totally unexpected
observation, as it was suspected that some amounts of dissolved gases would
be present in the sample, and that these could possibly cause local heating
effects as they came out of solution. (At higher temperatures, the bubbling
behavior could be noted to occur at higher pressures.)
An observation that was quite unexpected was that the areas of local
melting did not refreeze, even though the temperature of the cooling table
was far below the melting temperature of the sample. The relevant part of
the melting curve In H2SOjj - H20 is shown in Fig. 3. From this we see that
the 100% H2SOj} sample has a melting point at 282 K and a 95 wt% H2SOj, sample
would melt at approximately 250 K.
13
-------
300
280
o>
2
(5
a
1^260
240
80
90
100
110
Figure 3. Melting Point of H2S04 from Gable, Betz, and Maron (Ref. 3)
-------
The sample would continue to melt, often forming porous solid structures
that would coexist with the liquid for periods up to two hours in length.
Eventually the sample would become completely melted and the liquid would be
noted to take on a greenish color, which would darken as the sample aged.
Two conclusions could immediately be drawn from these observations:
(i) The very long life times of porous and f i lamentatious solid
coexisting with the liquid to which it was melting suggest that
thermal equilibrium times within the sample itself were quite long,
regardless of how rapidly external thermal equilibrium or quasi-
equilibrium was obtained.
(ii) The greenish discoloration indicated that quite possibly some
chemical process was proceeding.
Either of these conclusions would hamper the ability of the present
technique to provide measurements of the equilibrium vapor pressure of
- H20 system.
The total melting of these 100% I^SO^ and A-300 samples at temperatures
well below the melting curve was quite unexpected. From Fig. 3, only samples
with concentrations between 93.5 ~ 9^.0 wt% h^SO/j should be liquid at a tem-
perature of 240 K. Distillation of the samples was discarded as a possible
cause, as this would result in an increase in the concentration of ^SO^,
rather than the needed decrease. Supercooling effects were considered as
well, since it Is known that these mixtures strongly supercool. However,
thermodynamics does not permit the melting of an equilibrium solid to a
supercooled liquid.
It proved to be exceedingly difficult to refreeze any of the high wt%
H2SO/J samples after they had become completely melted — this only occurred
with one sample, a 100% h^SOj, sample. In this case, the pressure in the bell
jar was very rapidly increased to atmospheric pressure with dry nitrogen gas,
while the temperature of the cooling table and the melted 100% h^SOj, sample
remained below 250 K. As the pressure reached approximately 93 kPa the sam-
ple suddenly refroze to its expected equilibrium solid. Several attempts to
reproduce this behavior met with no success. In this isolated refreezing
Incident it appeared that the only parameter to have changed significantly
was the external pressure, and thus there appeared to be some possible pres-
sure effect on the melting curve.
For this reason, It was decided to make a series of measurements on sam-
ples of H2SO^ - H£0 with concentrations near that of the monohydrate, f^SO/^ •
H-0. The monohydrate mixture has a concentration of 84.5 wt% h^SOl}.
Reference to Fig. 3 indicates this concentration corresponds to a rather
broad maximum in the melting curve, with the conveniently high melting tem-
perature of 281 K. Thus the behavior of the system in this region is rela-
tively insensitive to small changes in the concentration. Therefore any
effect of pressure variation can be isolated with relative ease.
15
-------
PRESSURE - MELTING POINT RUNS ON MONOHYDRATE SAMPLES
General
Samples with concentrations near that of the monohydrate, h^SOjj • H20,
were prepared by mixing distilled water with the A-300 sample acid, in the
same manner as described previously. In this case a volume correction was
necessary because of the large difference in specific gravity between pure
water and the concentrated sulfuric acid. Two samples were prepared and runs
were made on separate days. These runs will be described in detail.
Run of July 7, 1977
A sample with nominal concentration of 84.6 wt% sulfuric acid and a
volume of 1.26 ml was prepared and immediately frozen at liquid nitrogen
temperature. This was transferred under a dry N2 atmosphere to the cooling
table, which was maintained at a temperature of 256 K. The bell jar was only
slightly evacuated, to a pressure of 81>6 kPa, simply to hold the bell jar
down and isolate the sample from the ambient atmosphere.
The temperature of the cooling table was then raised in steps of roughly
five degrees. At a temperature between 281 K and 282 K it was noted that the
sample was in the process of melting. This agrees with the expected melting
point, from Fig. 3 of .281.6 K for this sample. The sample completely melted
before the temperature could be reduced again.
The cooling table temperature was reduced to 238 K but the sample super-
cooled and did not refreeze into solid. The bell jar was pumped down using
the forepump. The forepump could not reduce the pressure over the super-
cooled sample to below 0.9 Pa. In an attempt to freeze the supercooled
liquid, the pressure was increased to 38.7 kPa by introducing dry nitrogen
gas, while maintaining a temperature of 238 K. A small crystal was then
noted to appear in the sample, and this grew rapidly, causing the entire sam-
ple to transform to the crystalline solid within approximately ten minutes.
After obtaining a solidly-frozen sample, the pumping line was again opened to
the forepump. The system pressure was reduced to 0.48 Pa and could not be
pumped below this value. The pumping line was again closed and the tempera-
ture increased in five degree steps.
The sample underwent the solid-to-1iquid transition during one of the
steps. The sample was noted to be roughly 50% melted at a temperature of
266 K and a pressure of 1.3 Pa. The observations of this run indicate the
following:
(i) The system exhibits substantial supercooling. This is not unexpec-
ted behavior and is indicated in the literature.
(ii) Two distinct melting points were observed, the expected one at a
temperature of 281 - 282 K and a pressure of 81.6 kPa, and an unex-
pected one at a temperature of approximately 266 K and a pressure
of about 1.3 Pa.
16
-------
One might assume that the pumping had distilled the water from the mix-
ture, thus reducing the melting temperature. However, from Fig. 3, this
would require an increase in concentration from 84 wtifc to 90.5 wt% ^SOj,.
Distillation to this extent would appear to be highly unlikely under the
experimental conditions. Also the observed melting pressure of about 1.3 Pa
is difficult to reconcile with the pressure of ^ 2.824 x 10~4 Pa obtained
from the tables of Gmitro and Vermeulen for h^SO^ - H£0 at 90.5 wt$ H2SO/, and
266 K.
A second independent run on another sample near the monohydrate concen-
tration appeared to be necessary in order to see if the behavior shown by
the run of July 7, 1977 could be reproduced, to measure more carefully the
transition temperatures and pressures, and to attempt to estimate the extent
of possible distillation.
Run of July 11. 1977
A sample with a nominal concentration of 84 wtfc t^SO/^ and a volume of
approximately 2.5 ml was prepared. The standard sample handling technique
was observed, with the mixed sample quickly frozen at 77 K with liquid nitro-
gen and under a dry N£ atmosphere transferred to the cooling table, which in
this case was maintained at a temperature of 255 K. As before, the atmos-
phere within the bell jar was dry nitrogen gas.
The vacuum chamber was slightly pumped down to a pressure of 80.9 kPa.
The pumping lines were closed off and the temperature of the cooling table
slowly raised while the pressure and the sample appearance were monitored.
The sample was observed to melt over a range of approximately eight degrees.
The temperature, pressure and visual estimates of the extent of melting of
the sample are given in Table 1.
TABLE 1. HIGH PRESSURE MELTING. MONOHYDRATE SAMPLE
Extent of Sample Melting
T (K) Pressure (kPa) (Visual Estimation)
275.7
279.1
281.4
282.4
283.5
283-9
81.281
81.349
81.407
81.424
81.444
81.460
Barely discernable
20%
40 - 50*
50%
80 - 85*
90%
It was decided to term the "melting point" that temperature at which
50% of the sample was estimated to be in each of the solid and liquid states.
By this criterion, the melting point of this sample was determined to be
282.0 ± 0.5 K. This is quite consistent with the value of 281.7 K taken
from the melting curve (Fig. 3). This constitutes good assurance that this
17
-------
sample is indeed within approximately 2% of the monohydrate concentration.
With the melting point determined, the cooling table temperature was reduced
to 253 K, whereby the system pressure dropped to 81.06 kPa. The sample did
not immediately refreeze, but appeared to be a viscous supercooled liquid.
The pumping line to the forepump was opened and system pressure reduced,
while at the same time the temperature was reduced further- At a pressure of
2.93 kPa a nuclear crystal of solid was observed in the sample, and this sub-
sequently grew, transforming the entire sample into the crystalline solid
within a period of approximately six minutes. Pumping was continued until a
pressure of 0.07 Pa was obtained. At this point the pumping lines were
closed, and the system pressure immediately rose to 0.17 Pa and then contin-
ued to rise more slowly.
At a temperature of 2^1.1 K and pressure of 0.639 Pa, a partial melting
of the sample was noted. Bubbles formed in the liquid parts and rose to the
surface. There they burst, causing an increase in the rate of pressure rise.
This behavior continued for approximately ten minutes during which the pres-
sure rose to 1.5 Pa. The bubbling then ceased, and the sample again appeared
to be solidly frozen in the crystalline state. It was tentatively assumed
that this bubbling behavior was due to out-gassing of dissolved gasses from
the sample.
The temperature of the sample was then slowly increased while close
visual observation of the sample was maintained. At a temperature of 257-9 K
and a pressure of 1.612 Pa definite signs of melting were observed in the
sample. The data obtained during the melting of this sample are given in
Table 2.
TABLE 2. LOW PRESSURE MELTING, MONOHYDRATE SAMPLE
Extent of Sample Melted
T (K) Pressure (Pa) (Visual Estimation)
257.
262.
266.
266.
9
5
5
8
1.
1 .
1.
1.
612
657
709
761
Detectable melting
25*
40*
80*
266
mel
Thus a "low
.5 K, some 15.
ting point".
pressure mel
5 K reduced
ting
from
point" was observed at a temperature of
the previously measured "high pressure
While a small amount of unmelted solid yet remained in the sample, the
pressure in the bell jar was increased to 65-3 kPa by the injection of dry
18
-------
gas. It was noted that the sample immediately began to refreeze, the solid
phase growing at the expense of the liquid phase. After five minutes from
the time of pressure increase, the system was judged to be approximately
solid at a temperature of 267.3 K and a pressure of 65.62 kPa. However, the
sample did not solidly refreeze, but was extensively slushy, making a reason-
able estimate of the extent of freezing difficult. After forty-five minutes
from the pressure increase, the sample was seen to be composed of approxi-
mately bQ% crystalline solid with the remainder appearing to be a glassy
solid. It was finally decided to presume that the sample, now at temperature
of 263.7 K and a pressure of 65.5A kPa, was "solidly frozen". Observations
of the.subsequent melting behavior described below appeared to support this
presumption: the melted liquid could be visually distinguished from the
glassy solid, although this distinction was admittedly more difficult than
was the case when only the crystalline solid was present.
Once again the sample temperature was slowly increased, while close
visual surveillance of the sample was maintained. The third observation of
a melting point in this sample is detailed in Table 3.
TABLE 3. SECOND HIGH PRESSURE MELTING. MONOHYDRATE SAMPLE
Extent of Sample Melted
T (K) Pressure (kPa) (Visual Estimation)
263-7
279.6
281.5
283.8
65.539
65.773
65.796
65.805
Sol idly frozen
Definite signs of
melting noted
60 - 80%
95%
From these data the melting temperature of this "second high pressure
melting point" is estimated to be 281.0 ± 0.5 K. This agrees, within the
combined estimated errors, with the previous melting temperature, obtained
from the data of Table 1. Thus it is concluded that no substantial distilla-
tion had occurred during the course of the run, and that the final concentra-
tion of the sample was not significantly different from the initial concen-
tration. There is definite evidence of an unexpected pressure effect on the
melting point of the sample, with a reduction in pressure causing a reduction
in the melting temperature.
This experimental observation is graphically presented in Figure k. The
broken line connecting the high-pressure and low-pressure apparent melting
points is not intended to represent any functional relationship, but only to
indicate the extent of the temperature and pressure changes between the ob-
servations.
19
-------
(SI
O
10s
10*
co
0)
tl
260
A -July7 1977
• -July 11 1977
270
280
Figure 4. Apparent Pressure Effect of H2SC>4 • H20.
-------
PRESSURE - MELTING POINT RUNS ON H2SO/j - H20 NEAR 100% H2SOi,
General
An effort was made to observe the possible pressure effect on the melt-
ing curve in samples with higher concentrations than the monohydrate. Runs
using the commercial sample A-300 (95 - 98 wt% h^SO^) and the 100% H2SO^
were made on July 12, 1977 and July 13, 1977, respectively.
Run of July 12. 1977 (A-300 Sample)
A sample of H2SOi, - H20 was taken directly from the A-300 commercial
preparation and handled in the usual fashion: it was frozen at liquid nitro-
gen temperature, and transferred to the cold cooling table within the bell
jar under a dry N2 atmosphere.
The pressure within the bell jar was reduced to 79.3 kPa and the melting
point was measured as described above for the monohydrate sample. The melt-
ing temperature was determined to be 255.3 ± 0.5 K. This corresponds, by
Fig. 3, to a stilforic acid concentration by weight of 95-5%, which is within
the range stipulated by the manufacturer.
The temperature of the sample was then reduced to 245 K in an attempt to
refreeze it.- The sample did not refreeze easily. After approximately forty-
five minutes the sample had only partially refrozen to a slushy consistency,
and growth of the solid phase was proceeding exceedingly slowly. Since it
did not appear that this sample would become solidly frozen within a reason-
able time, the bell jar was evacuated while part of the sample remained
liquid. As the chamber was evacuated, the solid phase in the sample at first
appeared to increase in size. Then, numerous small bubbles appeared, fol-
lowed by a short period of vigorous bubbling, which suddenly culminated in
the complete melting of the sample.
At a temperature of 241.1 K and pressure of 0.12 Pa the sample was com-
pletely liquid, with a calm surface, and was taking on the now-familiar
greenish tint. Subsequent attempts to refreeze this sample by increasing
the pressure failed.
Run of July 13. 1977 (100% »2^L Sample)
A sample of approximately 2 ml in volume of the 100% H2SO^ sample was
prepared and placed frozen on the cold cooling table inside the bell jar.
It was necessary to vent the bell jar In this instance with ambient air,
since the supply of dry N2 was exhausted. At a pressure of 77.2 kPa this
sample exhibited a melting point of 270 K. By Fig. 3, this corresponds to a
concentration of 97.5 wt% of H2SOj,. (It was noted that the difference from
100% H2SOj| could possibly have been absorbed water vapor from the air which
was used to vent the bell jar.)
Subsequent to the melting of the sample, all attempts to refreeze It
inside' the apparatus proved futile.
21
-------
OBSERVATIONS OF VAPOR BY MASS SPECTROMETER
Several spectra of the vapor of ^50^ - 1^0 system were obtained using
the mass spectrometer. At this point, only qualitative conclusions could be
drawn from the mass spectrometer data. In addition to a high water back-
ground (Mass 18), peaks at mass numbers 16, 28, 29, kB, 64, 81 were dominant.
These peaks were observed with sulfuric acid concentrations near 100% by
weight and also near the monohydrate concentrations. Qualitative indications
from initial calibrations, point toward a definite possibility of a detailed
and accurate measurement of the molecular fragments SO, S02, 50^, HS03 etc.,
of the sulfuric acid - water system at varying concentrations. However, as
mentioned earlier, the apparent pressure effect on the melting point of sul-
furic acid has taken precedence over the quantitative study of the molecular
fragments of sulfuric acid - water system with a mass spectrometer.
22
-------
REFERENCES
1. Abel, E. On the Experimental Bases for the Calculation of the Sulfurfc
Acid Vapor Pressure Above the Sulfuric Acid - Water System. J. Phys.
and Colloid Chem., 52 (1948), 903-914.
2. Abel, E. The Vapor Phase Above the System Sulfuric Acid - Water- J.
Phys. Chem., 50 (1946), 260-283.
3. Gable, C. M., Betz, H. F., and Maron., S. H. Phase Equilibria of the
System Sulfur Trioxide - Water. J. Am. Chem. Soc., 72 (1950), 1445-
4. Giauque, W. F. , Hornung, E. W. , Kunzler, J. E., and Rubin, T. R. The
Thermodynamic Properties of Aqueous Sulfuric Acid Solutions and Hydrates
from 15 to 300°K. J. Am. Chem. Soc., 82 (1960). 62-70.
5. Gmitro, J. I. and Vermeulen, T. Vapor - Liquid Equilibria for Aqueous
Sulfuric Acid. A. I. Ch. E. Journal, 10 (1964), 740-746.
6. Greenwalt, Crawford H. Partial Pressure of Water Out of Aqueous Solu-
tions of Sulfuric Acid. Ind. Eng. Chem., 17 (1925), 522-523.
7. Kunzler, J. E. and Giauque, W. F. The Freezing Point Curves of Concen-
trated Aqueous Sulfuric Acid. J. Am. Chem. Soc., 74 (1952), 5271-5274.
8. Rubin, T. R. and Giauque, W. F. The Heat Capacities and Entropies of
Sulfuric Acid and Its Mono- and Di hydrates from 15 to 300°K. J. Am.
Chem. Soc., 74 (1952), 800 -804.
9. Verhoff, Fi H. and Banchero, J. T. A Note on the Equilibrium Partial
Pressures of Vapors Above Sulfuric Acid Solutions. A. I. Ch. E. Journal
18 (1972), 1265-1268.
10. Handbook of Chemistry and Physics, 58th Edition, The Chemical Rubber
Publishing Company , 1 976 .
23
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing}
REPORT NO.
EPA-600/3-79-077
3. RECIPIENT'S ACCESSION-NO.
TITLE AND SUBTITLE
VAPOR PRESSURE AND MELTING BEHAVIOR OF SULFTJRIC
ACID-WATER SYSTEMS
5. REPORT DATE,._.
August 1979
6. PERFORMING ORGANIZATION CODE
AUTHOR(S)
G. Raymond Brown and V. Rao Veluri
8. PERFORMING ORGANIZATION REPORT NO.
PERFORMING ORGANIZATION NAME AND ADDRESS
Clark College
Department of Physics, Box 167
Atlanta, Georgia 30314
10. PROGRAM ELEMENT NO.
1AA603 AH-06 (FY-77)
11. CONTRACT/GRANT NO.
804470
2. SPONSORING AGENCY NAME AND ADDRESS
Environmental Sciences Research Laboratory - RTP, NC
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final 11/74 - 8/77
14. SPONSORING AGENCY CODE
EPA/600/09
15. SUPPLEMENTARY NOTES
16. ABSTRACT
An experimental apparatus was designed and constructed to utilize high vacuum
and mass spectrometric techniques to determine total and partial vapor pressures
above bulk liquid samples in the temperature range between -65°C and 25°C.
Observations on the sulfuric acid-water system revealed complexities which
interfered with the experimental goal, but which are of some interest in themselves,
These interfering processes included long internal thermal relaxation times and
chemical reactions. An unexpected and most interesting observation was an apparent
pressure effect on the melting curve of the sulfuric acid-water mixtures.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Air pollution
*Sulfuric acid
*Vapor pressure
*Melting points
13B
07B
07D
20M
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report)
UNCLASSIFIED
21. NO. OF PAGES
28
20. SECURITY CLASS (Thispage)
UNCLASSIFTF.n
22. PRICE
EPA Form 2220-1 (9-73)
24
------- |