APTD-1564
OPTIMUM WORKING FLUIDS
FOR AUTOMOTIVE
RANKINE ENGINES
VOLUME II -
TECHNICAL SECTION
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Water Programs
Office of Mobile Source Air Pollution Control
Advanced Automotive Power Systems Development Division
Ann Arbor, Michigan 48105
-------
APTD-1564
Prepared By
D. R. Miller, H. R. Null, Q. E. Thompson
Monsanto Research Corporation
800 North Lindbergh Blvd.
St. Louis, Missouri 63166
Contract No. 68-04-0030
EPA Project Officer;
K. F. Barber
Prepared For
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Water Programs
Office of Mobile Source Air Pollution Control
Advanced Automotive Power Systems Development Division
Ann Arbor, Michigan 48105
June 1973
-------
The APTD (Air Pollution Technical, Data)1; seri'es rof reports is issued by
the Office of Air Quality Planning and Standards'*; 'Office of Air and
Water Programs, Environmental Protection Agency, to report technical
data of interest to a limited number of readers. Copies of APTD reports
are available free of charge to Federal employees, current contractors
and grantees, and non-profit organizations - as supp^lie.s permit - from
the Air Pollution Technical Information Center, Environmental Protection
Agency, Research Triangle Park, North Carolina 27711 or may be obtained,
for a nominal cost, from the National Technical. Information Service,
5285 Port Royals Road, Springfield, Virginia 22151V
This report was furnished to the U.S. Environmental Protection Agency
by Monsanto Research Corporation in fulfillment of Contract No. 68-04-0030
and has been reviewed and approved for publication by the Environmental
Protection Agency.. Approval does not signify that the contents necessarily
reflect the views and policies of the agency. The material presented in
this report may be based on an extrapolation of the "State-of-the-art."
Each assumption must be carefully analyzed by the reader to assure that it
is acceptable for his purpose. Results and conclusions should be viewed
correspondingly. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
Publication No. APTD-1564
ii
-------
ACKNOWLEDGMENT
The authors acknowledge with gratitude the helpful counsel of
Messrs. K. Barber, S. Luchter and F. P. Hutchins, all of the Envi-
ronmental Protection Agency's Advanced Automotive Power Systems
Development Division. Many persons within Monsanto Research Cor-
poration and Monsanto Company contributed substantially to the
research program here reported. These included R. C. Binning,
J. A. Conover, L. L. Fellinger, R. L. Green, R. L. Koch, R. J.
Larson, G. J. Levinskas, R. A. Luebke, N. F. May, J. T. Miller,
L. Parts, A. C. Pauls, P. F. Pellegrin, J. V. Pustinger, W. R.
Richard, A. D. Snyder, W. N. Trump, W. M. Underwood, J. A.
Webster, E. P. Wheeler and F. J. Winslow. Their manifold inputs
are gratefully recognized, as also are those of Professor P. A. S,
Smith of the Department of Chemistry, University of Michigan, who
served as a consultant.
iii
-------
TABLE OF CONTENTS
Page
1. INTRODUCTION 1;
2. SUMMARY 2
3. CONCLUSIONS ; ^5
4. RECOMMENDATIONS J
5. FINAL CANDIDATES '8
5.1 Choice and Composition 8
5.2 Physical Properties fl.3v
5.3 Thermodynamics 26
5.4 Cycle Calculations 26
5.5 Stability/Compatibility Tests 33
5.5.1 Dynamic Loop Testing 33
5.5.2 Capsule Compatibility Tests at 300°F 54
5.6 Toxicity 55
5.6.1 Vapor Hazard Model 55
5.6.2 Rat Inhalation Studies 63
5.6.3 Range Finding Toxicity Tests 64
5.6.4 Products of Combustion 66
5.7 Flammability 71
5.8 Lubrication and Lubricants 71
5.8.1 Lubricant Stability 74
5.8.2 Lubrication Wear Tests 75
5.8.3 Special Lubricant Synthesis 82
5.9 Manufacture and Economics 91
6. THE SEARCH 101
6.1 Selection Criteria 101
6.1.1 Screening Criteria 101
6.1.2 Cycle Criteria 107
6.1.3 Cost Criteria 114
6.1.4 Hazard Criteria 116
6.1.5 Compatibility, Stability, Lubrication
Criteria 117
iv
-------
TABLE OF CONTENTS (Continued)
6.2 Initial Candidate Selection (Criteria 1-3)
6.2.1° Normal Boiling Points
6.2.2 Chemical Structure and Thermal Stability
6.2.3 Candidate Fluids
6.3 Thermal Stability Screening (Criterion 4)
,6.4 Physical Properties, Literature and Estimation
6.5 Thermodynamic Screening, I-Factor (Criterion 5)
6.6 Cost Projections and Screening
6.7 Flammability Testing
6.7-1 • Flash and Fire Point Results
6.7.2 Hot Compartment Spray Ignition Results
6.8 Physical Property Measurements
6.8.1 Vapor . Pressure
6.8.2 Flow Points
6.9 Thermodynamic .Table/Diagram Computations
6.10 Rankine Cycle Computations (Criteria 6-9)
6.10.1 Results of Computations
6.10.2 I-Factor as a Determinant of Cycle
Efficiency
6.11 Special Stability Test Work
6.11.1 Fluorinol-85 Stability Investigations
6.11.2
6.11.3
6.11.4
7 , REFERENCES
Nonvolatiles from Pyridine-Water Type
Fluids
Carboranes
Inorganic Fluids
Page
117
118
118
120
127
137
139
145
147
147
147
.152
154
158
158
167
174
174
176
176
184
186
APPENDICES'
Volume III
-------
CONTENTS OF OTHER VOLUMES , .
Volume I - Executive Summary
Introduction
Summary
Conclusions
Recommendations
Volume III - Technical Section -'Appendices
A. Scope of Work, Contract 68-04-0030
B. Thermodynamic Properties of RC-1
C. Thermodynamic Properties of RC-2
D. Dynamic Loop Test Procedure
E. Acute Vapor Inhalation Report,- RC-1
P. Acute Vapor Inhalation Report - RC-2
G. Acute Vapor Inhalation Report - Hexafluorobenzehe/Penta-
fluorobenzene , : -.- •.. : .-.-
H. Toxicity and Biodegradability of Pyridines - Literature
Abstracts : . .
I. Toxicological Investigation of Advanced Candidates
J. Monsanto Rub-Block Lubrication/Wear Tester
K. Computer Program E1393 - Thermodynamic Properties of
Multicomponent Fluids
L. Computer Program E1375 - Isoteniscope Data Analysis
(Liquid Activities) , . ;
M. Computer Program TSPL0T - Temperature-Entropy Diagram
Plotting
N. Flash and Fire Points, Micro-Cleveland Open Cup Method
0. The Hot Compartment Spray Ignition Test
P. Monsanto Recording Tensimeter
Q. Rankine Cycle Computations - Reference Ideal and
Equivalent Real Cycles
vi
-------
CONTENTS OF OTHER VOLUMES (Continued)
Volume IV - Engine Design Optimization
Introduction
Design Optimization Program Description
Problem Definition
Optimization - Turbine Engine Results
Optimization - Reciprocating Engine Results
Special Subroutine Description
Load Point Analysis for Federal Driving Cycle
Appendices
A. Optimizing Criterion "Pay Off" Function
B. Optimization of ORC Automobile Engine for RC-1
(•Turbine, W.F. = 0.5)
C. Optimization of ORC Automobile Engine for RC-1
(Turbine, W.F. = 0.97)
D. Optimization of ORC Automobile Engine for RC-2
(Reciprocator, W.F. = 0.5)
E. Subroutine SHIFT
F. Subroutine DRIVE
G. Subroutine RAMAR
H. Reciprocator Routine
I. Subroutine WEXPD
J. Federal Driving Cycle Analysis Routine
Vll
-------
. LIST OF TABLES
Table Page
1 RC-1 Composition 9
2 RC-2 Composition 10
3 Summary Properties - Final Candidates 14
4 Vapor Pressures, Final Candidate RC-1 and Related
Compositions . . 15
5 Vapor Pressures, Final Candidate RC-2 and Related-
Compositions ' 16
6 Measured Liquid Densities - Final Candidates 17
7 Measured Liquid Specific Heats - Final Candidates 18
8 Measured Liquid Kinematic Viscosities - Final
Candidates ... . 19
9 Measured Liquid Thermal Conductivities - Final
Candidates 20
10 Calculated Vapor Viscosities at Low Pressure - Final
Candidates ... .. 21
11 Calculated Vapor Thermal Conductivities at Low
Pressure - Final Candidates . 22
12 Pseudocritical Constants for Extension of Transport
Properties - Final Candidates • . 23
13 Physical Property Data Used in Thermodynaraic Table
Generation for Final Candidate RC-1 . . 27
14 Physical Property Data Used in Thermodynamic Table
Generation for Final Candidate RC-2 28
15 Rankine Cycle Summaries, Final Candidates 31
16 Dynamic Loop Particulars 35
17 Dynamic Loop Test Conditions 40
18 Loop Test Summary, RC-1 4l
19 Loop Test Summary, First RC-2 44
20 Loop Test Summary, Second RC-2 51
viii
-------
LIST OF TABLES (Continued)
Table Page
21 Capsule Compatibility Tests - RC-1 56
22 Capsule Compatibility Tests - RC-2 57
23 Vapor Dispersion Calculation - RC-1 6l
2*1 Vapor Dispersion Calculation - RC-2 6l
25 Vapor Dispersion Concentration Histories 62
26 Rat Inhalation Studies: 4-Hour LC50 Concentrations 63
2? Acute Toxicity Test Results 65
28 Combustion Emission Experiments 68
29 Mass Spectrometric Analysis of RC-1 Combustion Products 69
30 Mass Spectrometric Analysis of RC-2 Combustion Products 70
31 Flash and Fire Points and AlT's of Final Candidates
RC-1 and RC-2 and Reference Fluids 72
32 Hot Compartment Ignition Test Results 73
33 Equilibrium Solubility of Methylpyridine Working Fluids
in Various Prospective Lubricants 76
34 Rub Block Wear Tests of Composite Materials 77
35 Rub Block Wear Tests of Lubricant and Mixtures of RC-1
and RC-2 with Lubricant 83
36 Solubilities and Low Temperature Behavior of Aryl
Silane Lubricants 90
37 Manufacturing Cost Estimate - RC-1 96
38 Return on Investment - RC-1 97
39 Propulsion System and Working Fluid Goals 102
40 Regenerator Data Presented at January 1972 Contractors'
Coordination Meeting 112
41 Markups after Plant Sale 114
42 Distribution of Cost 115
ix
-------
LIST OF TABLES (Continued) -,
Table Page
43 Approximate Thermal Decomposition. Temperatures pf
Organic Compounds as a Function of Weakest Bonding Unit 121
44 Pure Fluids Selected for Thermochemical Stability
Screening 122
45 Fluid Mixtures Subjected, tcp The-rmochemical Stability
Tests ' 126
46 Rankine Fluid Candidates - Summary of Thermochemical
Screening Results , 130
4? Stability Testing of Mixed Fluids ... . 133
48 Pure Fluids Surviving 720°F Stability Test for 24 Hours.
or Longer 134
49 Pure Fluids Stable for Longer than 200 Hours at 720°F 136
50 Approximate Order of Stability in Ampoule Tests at
720°F .. 137
51 Pure Component Properties Used in Thermodynamic Prop-
erty Calculations .. - • >.l40
52 Van Laar Constants for Rankine Cycle Mixture Calcula-
tions ...... , . . i/ti
53 I-Factors for Pure Fluids Surviving 720°F Stability
Test for 24 Hours or Longer 142
54 Thermally Stable Components and I-Factprs 144
55 Published and Estimated;Prices of Selected Fluids 146
56 Flash and Fire Points of Mixed Fluids 148
57 Hot Compartment Ignition Test Results 149
58 Vapor Pressures in Psia - Ethanol/Water 153
59 Fluids for Which Thermodynamic Tables and Charts Have
Been Prepared. .... . , ,154
I :'
60 Reference Ideal Cycles - Simple Fluids 159
61 Equivalent Real Cycles - Simple Fluids 160
-------
LIST OF TABLES (Continued)
Table Page
62 Reference Ideal Cycles, MFSN-15, Pyridine-Water l6l
63 Reference Ideal Cycles, 3-Methylpyridine, Water
Mixture, MFSN-21 162
6^J Reference Ideal Cycles, 2-Methylpyridine, Water
Mixture, MFSN-23 163
65 Reference Ideal Cycles, 4-Methylpyridine, Water
Mixture, MFSN-24
66 Special Ideal Cycles, 2-Methylpyridine, Water
Mixture, MFSN-23 165
67 Special Ideal Cycles, 4-Methylpyridine, Water 166
68 Summary of Reference Ideal Cycle Calculations on
Thermally Stable Candidates 168
69 Summary of Equivalent Real Cycle Calculations on
Simple Thermally Stable Fluids 169
70 Formation of Nonvolatile Tars in Methylpyridine-Water
Fluids after Thermal Exposure at 720°F 177
71 Comparison of Fraction of Oxygen Consumed by Three
Carboranes and Benzene Held at 178°C (352°F) in Air
for Various Times 180
72 Ignition Delay Data 182
73 Some Unconventional Inorganic Fluids 185
xi
-------
,. LIST OE FIGURES .-
Figure. Page
1 . Flow Points of Final Candidate Mixtures 11
2 Vapor Pressures - RC-1, and Related 24
3 Vapor Pressures - RC-2 and Related 25
4 Temperature-Entropy Diagram - RC-1 29
5 Temperature.-Entrqpy .Diagram,.- RC-2 . 3Q
6 Dynamic Loop Schematic 3*1
7 Dynamic Loop 36
8 Heater Coil -...Ends .Cut and Sectioned Lengthwise, for
Examination ' ' . - ^7
9 Photomicrographs of'.Heat Exchanger Surfaces before.'Use : 38
10 Aluminum Condenser, Coil 39
11 Photomicrographs of Heat Exchanger Surfaces after
RC-1 Test 43 :
12 Vent Gas Syringe Arrangement 46
13 Photomicrographs of Heat Exchanger Surfaces after.
First RC-2 Test ' ' 47
14 Photomicrograph of Pit in Heater Tube after .Fii-st ,
RC-2 Test ' " 48
15 Photomicrographs of Heat Exchanger Surfaces aft.er
Second RC-2 Test ; . : 53
16 Vapor Inhalation Hazard Model .5,9
17 Rub Block Wear Specimens 78
18 Rub Block Wear Specimens 79
19 Rub Block Wear Specimens 80
20 Scar Widths of Carbon Blocks versus Porosity 8l
21 Rub Block Wear Specimens 84
22 Rub Block Wear Specimens 85
xii
-------
1. INTRODUCTION
This is the second of a four-volume final report on a research
program performed by Monsanto Research Corporation for the U.S.
Environmental Protection Agency, Advanced Automotive Power Systems
Development Division (AAPSDD) under contract 68-04-0030. The
objective of the work was to determine the best working fluids for
Rankine cycle automotive power plants.
The Rankine engine has been identified by EPA as a potential
alternative to the internal combustion automotive engine in the
event the internal combustion engine can not satisfy the 1975/76
Federal Emission Standards. Development work has been proceeding
on the Rankine engine, as well as other alternative power plants,
under the direction of AAPSDD since initiation of the Clean Air
Act Amendments of 1970.
In a Rankine cycle power plant, the working fluid, which is
repeatedly vaporized, expanded and recondensed, plays a key role
in determining the competitiveness of the system. The thermo-
dynamic, chemical and physical properties of the fluid have a
strong influence on the power plant efficiency, materials of con-
struction, component weight and size, the degree of hazard, if
any, and to some extent the power plant cost. Therefore, to make
certain that the Rankine engine could be evaluated at its greatest
potential, this study program was initiated to search out the best
fluids currently available for the automotive Rankine engine
application.
To accomplish this objective, specific guidelines were established
for the fluid screening process. These guidelines are outlined in
the work statement of Appendix A and summarized as follows:
Minimum Carnot Cycle Efficiency L\2%
Minimum Rankine Cycle Efficiency 30%
Maximum Cost to Owner, 5 yr period $100
Melting Point <-^0°F
Health Hazards Nil
Fire and Explosion Hazards Nil
Environmental Hazards Nil
The above restraints were contractual requirements and were intended
to approximate the upper limit of what could be achieved with
presently known materials and current technology. The results
indeed reveal that the criteria were well chosen, as only two
possible fluids from over 100 candidates survived the screening
process, and these did not totally satisfy all of the stated
-------
requirements. In reviewing the results it should be kept in mind
that less stringent or different selection criteria could very
well.have led to the selection of other candidates.
The order o>f presentation of the technical results' in this, report1
has been'.Inverted for the convenience of the reader. Results
specific to the two final candidate fluids fare •cbncentrat'ed':'iri
Section 5, whereas the prior search leading to the choice of the
two candidates is the subject of'Section 6. The chronological'
sequence of the report is through sections 1, 6, 5» 2, 3, ^ •
Companion Volume IV- of this final report covers system' optimization
studies by Sundstrand Aviation (under subcontract to'Mohsarito
Research Corporation) of each of the two final candidate fluids,
one in a turbine, the other in a reciprocating power plant.
2. SUMMARY
Four major tasks were undertaken in response to the basic contract,
namely:
1. Establishing working fluid selection criteria (Section 6.1)'
2. Searching for and screening a large group of fluid candi-
dates (Section"6.2 through 6.11)
3- Developing a comprehensive data base for the most promising
two candidates (Section 5) :
4. Predicting optimum engine designs, one reciprocating and
one turbine, involving the .two preferred working fluids
(Volume IV) ...-••
Selection criteria were first established,to provide as quanti-
tative as possible a basis for seeking and then accepting or
rejecting members of a broad collection of likely fluids. The
criteria covered fluid availability, physical properties,
thermal stability, thermodynamic and cycle efficiency and utility
in principal engine types. It .was., not deemed practical to
quantize all criteria, nor to be rigidly.proscribed by those
that were quantized. The criteria were of greatest utility in
permitting early elimination of entire chemical classes.of . .
materials from further consideration. As one example, .chlorine-
containing organics were eliminated'as a class when no members
were found able to long survive a 720°F stdel ampoule screening
test. Other examples can be found in Sectidn 6.
-------
The search for candidate working fluids centered on pure com-
pounds boiling (at atmospheric pressure) in the range of 150 to
250°F and freezing below 68°F. Oh the bases of availability and
expected thermal stability, about 110 pure fluids were obtained
and subjected to the 720°F steel ampoule test. Those few
surviving 200 hours without serious decomposition were then
subjected to an "I-factor" test, a criterion designed to uncover
fluids capable of high efficiency cycle operation without
requiring prohibitively large regenerative heat exchangers. Com-
puter-aided predictions of vapor heat capacities and heats of
vaporization provided the I-factor values.
The search for working fluids was not confined to pure compounds.
Mixed (or complex) fluids were included to permit more control
over freezing points and thermodynamic characteristics. .Thermo-
dynamic prediction techniques were broadened to include non-ideal
as well as ideal solutions. Graphing of temperature-entropy
diagrams for both simple and complex fluids was mechanized.
As 'a result of the various criterion-guided experimental and com-
putational studies, two final candidate fluids were identified:
\V
RC-1: A 60/40 mole percent mixture of pentafluoro-
benzene/hexafluorobenzene;
RC-^2: A 65/35 mole percent mixture of water/
2-methylpyridine
Both fluids are liquid to -40°F and meet critical performance
criteria for both reciprocating and turbine engines, although'
neither completely satisfies all requirements and criteria.
After identifying the two final candidates, an extensive
experimental program was undertaken to characterize them. Liquid
phy.sicai properties measured included vapor pressure, density,
viscosity, thermal conductivity, flow point and heat capacity.
These data, combined with literature data available for the pure
components of the candidates, were then used to predict other
properties of the mixtures, notably vapor heat capacity, viscosity,
and thermal conductivity. Complete tables of thermodynamic
variables were next computed using a Redlich-Kwong equation of
state specially modified to handle non-ideal solutions. Tempera-
ture-entropy diagrams were also computer-generated.
Predictions of fluid performance in both an ideal and a real
Rankine cycle were next prepared and compared with system
Restrictions. Both qualified in all respects except that RC-2
failed the 30% ideal cycle efficiency criterion.
-------
Both candidates were subjected to 1000-hour dynamic loop testing
under conditions of temperature (720°P max.) and pressure (1000
psig max.) simulating ultimate use. In this test, the RC-1 . .
candidate passed with only a trace of decomposition, whereas!the.
RC-2 candidate failed due to excessive corrosion of the ,SAE (1130
steel tubing utilized in the vapor generator. A .more corrosion-^.
resistant steel-or steel finish .is required before-RC-2 can .be ,;. /
exploited as an automotive working.fluid.
Compatibilities of the candidates with a wide range of both
metallic and non-metallic materials were.determined in a .series
of 300°F ampoule exposure tests. The choice of wetted materials
is wider for RC-1 than for RC-2.
A battery of animal acute exposure tests were performed to gauge
the acute toxicities of the' final candidates.. Included were ...
rat inhalation,, rat oral feeding, rabbit skin absorption, rabbit
skin irritation, and rabbit .eye irritation tests. , Special
attention was given to vapor inhalation as the most likely
avenue of public exposure. Predictions of likely dosages to.be
encountered in a . "worst-case" accident were performed .to compare
with the animal inhalation data. Additionally, experiments were
performed to identify products of combustion of the-working .
fluids exposed to a propane flame.
Flammability parameters were measured, including flash'.and fire
points, autoignition temperature, and ignition-explosion
character. RC-1 was found practically non-flammable. RC-2 is
flammable, although the presence of water does materially reduce;
the energy release below typical fuel values. .
Heat-stable lubricants were sought for each fluid. Only
synthetics were found to withstand the 720°F exposure condi-
tions. A biphenyl-terphenyl eutectic was defined and tested.
with both candidates and .with a wide variety of sliding .element
materials in a special rub-block friction-wear test machine.
This lubricant was soluble in both candidates.
A limited program of lubricant synthesis, undertaken to find an
RC-2 immiscible lubricant, identified the highly silylated ben-
zenes as possessing the required thermal stability and immis-,.
cibility. Two deficiencies of the materials studied .were .high
expected cost and high solidification point.. • •• • .
Future selling prices of the two candidate working fluids were
estimated based on economic and technological information avail-
able in the open literature. The analysis included, in the case
of RC-1, a conceptual plant design employing a fluorination .
process never practiced at larger than bench scale. Using the
projected selling prices, estimates were made of the fluid cost to
-------
the Rankine automobile owner over the first five years of owner-
ship. The RC-1 candidate was projected to cost 12% over the con-
tractual limit of $100; whereas the RC-2 candidate can be
expected to cost less than the limit.
The last task, that of predicting optimum engine designs corre-
sponding to the two final candidate working fluids, is summarized
in Volume IV.
3. CONCLUSIONS
1. No organic or aqueous organic fluids were found to completely
satisfy all the requirements established in the contract for
Rankine cycle automotive power plant working fluids.
2. Fluid RC-1 comes nearest to satisfying the automotive organic
working fluid requirements and goals set by AAPSDD. It meets
all of the established performance criteria for both turbine
and reciprocating engines, including efficiencies, temperature,
pressure and density limitations, and regenerator size.
Additionally it is considered fire-safe and unlikely to cause
serious injury on accidental human exposure (subject to veri-
fication in extended toxicity testing recommended as a follow
on). The fluid is quite stable at temperatures to at least
720°F in low-cost system materials, and it is expected to
last for at least the 5-year span of initial vehicle owner-
ship. It possesses a distinct advantage over water in not
freezing at temperatures down to -40°F. A low cost soluble
lubricant is available if needed. The main contractual
deficiency of the fluid is its . projected cost of no less than
$112/over its expected five-year life, as compared to a $100
contract limit. To arrive at the $112 cost it was necessary
to assume a) successful advancement of the state of fluor-
ination technology beyond that now practiced, and b) the
existence of a 50 million pound per year benzene fluorination
plant. Current prices of the higher fluorinated benzenes are
many times the projected price. Turning to cycle practicality,
the comparatively high flow rate of the RC-1 fluid per net
cycle horsepower (related to its higher molecular weight)
may be considered a disadvantage despite the lack of a con-
tractual limit. All in all, this fluid is worthy of further
development.
3. Fluid RC-2 is a back-up candidate for automotive use, either
In a turbine or reciprocating engine. It satisfies all per-
formance criteria except the 30% minimum ideal cycle efficiency
-------
E;ve,n so, the.fluid is capable of high..ef f iciency .usage, in. a
real cycle, with a compact regenerator. Although -the fluid is
flammable, it is less so than common fuels. In burning it
releases considerably less energy because of its water con-
tent, and its products of combustion are not hazardous to
humans. The fluid itself.-is unlikely to cause serious injury.
It is marginally stable at :720°F in steel. SAE 4130 ste.el is
not a compatible material for boiler tube construction .with
RC-2. The fluid acts chemically to corrode this steel, forming
FesOtt and hydrogen. A more suitable steel alloy needs to be
found before this fluid can be considered a contender with
RC-1. Assuming a satisfactory- s;teel were found, the RC-2
fluid would cost the vehicle owner about $100 in the first
five years of vehicle life. This presumes three system
refills. Traditional lubricants cannot be used with this
fluid above 700°F, but low cost synthetics-offer promise
when mutual solubility: is permissible. Synthetic silarie
lubricants may be considered if immiscibility is'mandatory,
but further development is needed. Fluidity above -40°F is
another feature of Importance. The prime, ingredient of ;:the
fluid is commercially available at low cost now, but in '.._
lesser quantity than would .be .commanded, by a million c.ars- .
per-year market. RC-2 is worthy of further development,, only
in the unlikely event that candidate RC-1 must be. abandoned.
4. Many organic fluids posses the state and thermodynamic
properties needed to attain engine performance goals.
5. Very few fluids possess the requisite thermochemical. stability
in steel at 712°F (and higher). Those that.do fall in these
categories:
Aromatics
Fluoroaromatics
Aromatic tertiary amines
Perfluoroaliphatics
Water ; '
6. Flammability eliminates the aromatics from contention, and
places the aromatic tertiary amines in a contingency
category. Since the aromatic tertiary amines (or pyridines)
are quite flammable and water soluble, their safe usage in
automobiles is contingent upon a suitable reduction in
flammability effected by mixing with water.
7. An excessive demand for regenerator space to maintain
tolerable cycle efficiency rules out the perfluoroaliphatics.
This is clearly shown by the I-factor criterion and the •
supporting cycle, calculations.
-------
Additional restrictions on vapor pressure, melting point
(or "flow" point), cost, and toxicity further narrow the
choice to
mixed fluorobenzens, and
aqueous pyridines,
from which categories the final candidates RC-1 and 2 were
chosen.
The search rationale adopted at the outset was basically
sound, although some adjustments were necessary. Particularly
valuable features were:
a. an efficient thermal stability screening program
b. development and effective use of the I-factor con-
cept for thermodynamic screening - including confirma-
tion via I-factor - efficiency correlations
c. maximal use of computation throughout, but especially
in these areas:
prediction of thermodynamics of non-ideal/
ideal mixtures
I-factor and physical property predictions
ideal and real cycle calculations
d. the concept of progressive advancement of candidates,
exposing each to increasingly rigorous examination.
RECOMMENDATIONS
Relative to fluid RC-1, it is recommended that:
1. It be considered an acceptable organic fluid for a prototype
engine development program.
2. Static thermochemical stability tests be performed to learn
how far above 720°F the fluid might be driven.
3. Providing prototype experience is favorable, the fluid be
subjected to a carefully planned set of exposure tests in
the areas of
-------
chronic inhalation
chronic dermal contact
environmental fate
as necessary to qualify the fluid for widespread use.
4. Encouragement be given to prospective manufacturers to
develop the necessary low-cost processes:for large scale
production. '
Relative to fluid RC-2, and assuming abandonment of RC-1,
it is recommended that:
5. Additional work be undertaken to find a suitable steel
and/or steel passivation technique to avoid the iron-water
reaction at high temperature.
6. If appropriate, additional synthesis work be done to finalize
an immiscible silane lubricant.
7. A fire hazard simulation be performed, preferably involving
a mock-up burner-vapor generator and a simulated high
pressure fluid leak.
8. The first three recommendations under RC-1 be followed.
5. FINAL CANDIDATES
5.1 CHOICE AND COMPOSITION
Details of the extensive search for working fluids conforming to
the demanding goals of this study are given in Section 6. A
brief summary here of the search findings serves to explain the
choice of final working fluid candidates. :
No all-inorganic fluids were uncovered which .were simultaneously
compatible with mild steel at 720°F and potentially inexpensive.
Inorganic compounds qualifying by melting and/or boiling point
are nearly all ionic chlorine-bearing, materials which1 are highly
corrosive to low cost metals, especially at high temperature or
when wet. (Water itself is an obvious exception) ,
Close scrutiny of the 1^0+ organic working fluids entered in the
screening process of Section 6 disclosed none likely to fully
satisfy all search goals simultaneously. Competitive tradeoffs
became painfully evident. Only one way was found to secure high
-------
fire resistance simultaneously with good thermal stability at
720°F. That was to incorporate many fluorine atoms on an organic
nucleus. But fluorination, especially of aromatics, is inherently
expensive. This then threatens the attainment of a 5-year fluid
cost of $100 to the automobile owner. Highly fluorinated chain
compounds (fluoroalkanes and related) can be found which are
stable, firesafe, and potentially less expensive, but all such
compounds cited suffer the defect of demanding impossibly bulky
regenerative heat exchangers to attain reasonable cycle efficiencies
Two candidate working fluids were selected for final evaluation
and verification based upon two distinctly different ration-
alizations of the available fluid data and the competitive
tradeoffs:
1. A fluid qualifying under the screening criteria
most likely to be fire-safe and stable, but with
relaxation of the owner cost requirement ($100 in
5 years).
2. A fluid qualifying under the screening criteria most
likely to be low-cost and stable, but with relaxation
of the goal of non-flammability.
Corresponding to these two rationalizations are the two final
candidate working fluids designated as RC-1 and RC-2.
5.1.1 Final Candidate RC-1
The composition of Fluid RC-1 is given in Table 1.
Table 1. RC-1 COMPOSITION
Concentration
Component Mol. Wt. Mol % Wgt %
pentafluorobenzene 168.1 60.0 57-5
hexafluorobenzene 186.1 ^0.0 42.5
Note: Alternate compositions, including proportions of
other fluorinated benzenes in addition to, or in
lieu of, those listed, are acceptable so long as
they conform to flammability, flow point and other
requirements
Hexafluorobenzene is an attractive working fluid candidate in its
own right because of its outstanding fire resistance and high
thermal stability. Alone, however, it crystallizes at ^1°F (5°C)
-------
and thus fails the -20°P cold start requirement, if not the -*JO°F
no-damage requirement. It is obvio.usly attractive to mix another
stable miscible,fluid with hexafluorobenzene to reduce the
crystallizing point by the requisite amount. Consideration was
given to a number of companion fluoro chemicals, especially
pentafluoropyridine, octafluorotoluene, FC-75, P-1D, and lower
fluorinated benzenes. By far the most attractive of these is
the latter class, as exemplified by pentafluorobenzene, because
chemicals of this class can be produced simultaneously with
hexafluorobenzene in the same chemical processing plant. ;
The 60-40 mol -ratio was selected to place the .flow point just '•
under the -40°P, temperature specified as .the- 'highest permissible
cold-damage fluid .temperature (refer to Figure. 1). If, instead,
-the flow point had been placed at -20°F (-29°C), the specified
highest engine coldstart temperature, the pentafluorobenzene
content would have be.en 52 rather than 60 mol percent. The
presence of additional amounts of lower fluorinated benzenes in
a commercial fluid would tend to lower the crystallizing point
somewhat, permitting a slightly higher hexafluorobenzene con^
tent in either case.
The thermal stabilities of. the ingredients of this candidate have
been established in Section 6.3- Other stable materials con-
sidered as components for this candidate, with the reasons for
their rejection, were:
lower fluorinated benzenes - flammability
fluorinated toluenes - cost
di(trifluoromethyl)benzenes - I-factor (regen. size)
perfluoro alkanes, cycloalkanes - I-factor
(regen. size)
pentafluoropyridine - cost, toxicity
5-1.2 Final Candidate RC-2
The composition of Fluid RC-2 is given in Table 2 on page 12.
Water is a preferred candidate from several viewpoints, especially
cost, hazards, and stability. Two decisive defects for automotive
use are lower cycle efficiency than organics at fixed operating
temperatures and high freezing point. As with hexafluorobenzene,
it is attractive to seek a miscible companion fluid to mix with
water to correct its deficiencies. Of all organic materials
turned up in the search, only pyridine and its three monomethyl
derivatives, 2-, 3- and 4-methylpyridine (also called a-, B- and
Y-picoline, respectively), possessed the requisite 720+°F thermal
stability and water miscibility. While all the pyridines will
1.0
-------
0 -10 -
o
Q_
§ -20 -
1. Hexafluorobenzene
2. Pentafluorobenzene
1. Water
2. 2-Methylpyridine
40 50 60
Mol % Component 2
70 80 90 100
Figure 1. Flow Points of Final Candidate Mixtures
11
-------
Table 2. RC-2 COMPOSITION
Concentration .
Component Mol. Wgt. Mol %Wgt %
water 18.0 65-0 26.4
2-methylpyridine 93.1 35-0 73.6
Note: Alternate compositions, including .proportions
of pyridine, 3- and. 4-methylpyridine in addi-
tion to, or in lieu of, those listed, are.
acceptable so long as flow point requirements
are met.
burn in air, it was expected that a useful diminution of flamma-
bility would occur on being mixed with water, the more so the
lower the. organic content.
Of the various pyridines, the 2-methyl derivative was selected
for these reasons:
1. It is the most effective water freeze point depressant.
2. It has, in water solution, a thermal stability com-
parable to pyridine/water and somewhat higher than
the other methyl pyridines (see.Section 6.3)-
I :•'
3. Its ideal cycle efficiency (Section 6.10) is equal
to or slightly higher than that of its sisters in
the mol percent range of 0 to 40, the preferred low
organic concentrations where the presence of water
tends to reduce flammability and toxicity.
4. Its flash point, while significantly lower than those
of the 3- and 4-methyl derivatives, is appreciably
higher than that of pyridine itself, when compared
at equal molal.concentrations (Section 6.7).
The 65-35 mol ratio was selected to place the flow point just
under the -40°F temperature specified as highest permissible
cold-damage fluid temperature. (Refer to' Figure 1)-. If, instead,
the crystallizing point had been placed at -20°F (-29°C), the
specified highest engine coldstar't temperature, the 2-methylpyridine
content would have been 32 rather than 35 mol percent. The
presence of minor amounts of pyridine and other methyl pyridines
would have only a minor effect on flow point. (Refer to Section
6.8).;
12
-------
Of the various fluid mixtures investigated in an attempt to
radically reduce regeneration requirements (i.e. to control the
shape of the dew line), only the aqueous pyridines were capable
of spanning the gap in thermodynamic behavior between water -- a
condensing fluid -- and such slightly superheating fluids as
thiophene.
5-2
PHYSICAL PROPERTIES
Principal physical properties, measured or calculated, are re-
corded in Tables 3 to 12 and Figures 2 and 3 following. Contents
of these presentations are:
Table
5
6
7
8
9
10
11
12
Content
Summary properties; molecular weights, crystallizing
points, boiling points, condensing pressures, heats
of vaporization, I-factor
Vapor pressures, RC-1 and related
Vapor pressures, RC-2 and related
Liquid densities
Liquid specific heats
Liquid kinematic viscosities
Liquid thermal conductivities
Vapor viscosities
Vapor thermal conductivities
Pseudo criticals - for extension of transport
properties
figure
2
3
Vapor pressures, RC-1 and related
Vapor pressures, RC-2 and related
13
-------
Table 3. SUMMARY PROPERTIES - FINAL CANDIDATES
Final Candidate
Property . RC-1 RG-2
Average molecular weight 175-3 44.3
Flow point, °'F".: -44 -40
Normal boiling point, °F - , 172- 200
Condensing pressure
0 220°F, psia ' 30 21
Heat of vaporization . . . • •
(9'NBFT, Btu/lb.. 79;.i . 378
I-Factor at 220°F 0.72 1.38
Liquid density at
77°F, g/ml 1.549 ... 0.985
Specific heat at 77°F,
Btu/lb°F . 0.290 0.668
14
-------
Table 4. VAPOR PRESSURES, FINAL CANDIDATE RC-1
AND RELATED COMPOSITIONS
Vapor Pressure1, psia
ooiiipus i u ion ,
mo I %
C6H?5
C6F6
Temperature
°C °F
50 122
75 167
100 212
125 257
150 302
175 347
200 392
225 437
250 482
0
100
Lit.2 Meas.
5-4 5-2
13-0 12.9
27-7 27.1
53-1 49. 53
93-6 91
15^ 152
239 235
355 355
506 500
(RC-1)
60
4o
Meas.
5-9
13.1
26.3
47. 13
85
142
221
340
480
100
0
Lit . 2 Meas .
4.7 8.4
11.2 15-1
23-5 27-6
44.9 47- 63
79-5 80
132 133
207 210
310 315
448 460
Notes :
Actually vapor-liquid equilibrium pressures, 1:1 liquid:
vapor volumes (very close to bubble point pressures)
Reference 1
Extrapolated
15
-------
Table 5- VAPOR PRESSURES, FINAL CANDIDATE -RC-2 .-
AND RELATED COMPOSITIONS
Vapor Pressure1, psia
^~~--4io^nposi.tion • mo] %•
^ ~^^_^ Wa t e r
Temperature^ 2MP?
°C °F \ Note
50 122
75 167
100 212
1?5 257
150 302
175 347
200 392
225 437
250 482
275 527
300 572
325 617
0
100
3
0.8
2.4
6.0
13-0
25.4
45.5
77
124
185
265
380
4406
25
75
4
2.4
5-7
14.1
29-8
59
110
173
266
390
540
710
900
50
50
4
2.5
6.6
16.8
37-9
78
144
240
390
570
810
- '
-
(RC-2)
65
35
4
2.8
7.4
18.4
39-2
75
1-51
252
420
630
900
-
-
100
0
5
1.8
5.6
14.7
33.7
69
129
225
370
577
862
1246
1749
Notes
l
2
3
4
5
B
Actually vapor-liquid equilibrium pressures, 1:1 Liquid:
vapor volumes (very close to bubble point pressures)
2-Methylpyridine
Values below 50 psia from Reference 2, others measured
Measured (by recording tensimeter)
All values from Reference 3
Value by extrapolation
16
-------
Table 6. MEASURED LIQUID' DENSITIES, -- FINAL CANDIDATES
Density'
g/ml (at 1 atm.- abs.; )
Temperature
°C
0
25 .
50
75
82
92
o.p
32
77
..' 122
167
179'.6
197-6
Final Candidate
RC-1
1.607
1.549
1.503
1.448
1.433
RC- 2
1 .007
0.985
0.962
: 0.935
-
0.920
Method: Westphal balance
17
-------
Table . 7• MEASURED LIQUID SPECIFIC HEATS - FINAL CANDIDATES
Specific Heat
cal/g-C° (at 1 atm. abs. or vap. press.)
Temperature
°C °F
25
30
'40
50
60
70
80
90
100
110
120
130
L •'-! 0
Method:
X
77
86
104
122
140
158
176
194
212
230
248
266
284
Perkin
Snmfa 1
.. "-:. , Final Candidate
RC-1 (Al pan)
0.290
0.293
0.296
0.300
0.306
0.310
0.316
0.321
0.330
-
-
-
Elmer DSC-2
nrl 1 pat-.-l on nf*
. RC-2 (Al pan)* RC-2 (Au pan)*
0.655
0.660
0.679
0.690
0.706
0.719
0.735
0.746
0 . 755
-
-
-
differential s
a +: hoY>mu1 Inl-o
0.668
0.684
0.700
0.708
0.720
0.718
0.725
0.741
0.745
0.766
0.789
0.814
0.826
canning calorimeter
•f a r» 1- 1 n n hot-.woon f-ina
candidate RC-2 and the aluminum pan led to a repeat
using a gold (Au) pan, which gave no indication of
such interaction.
18
-------
Table 8. MEASURED LIQUID KINEMATIC VISCOSITIES
FINAL CANDIDATES
Temperat
°C
37.8
51-7
65-6
76.7
ure
op
100
125
150
170
Final
RC-1 -
0.45
0.38
0.33
0.30
Candidate
RC-2
2.00
I. Hi
1.08
0.92
Method: ASTM 04^5; Cannon Manning semi-
micro tubes
19
-------
Table 9- MEASURED LIQUID THERMAL CONDUCTIVITIES - FINAL CANDIDATES
Thermal Conductivity x 105
cal/C°-cm-sec (at 1 atm. abs.)
Temperature
°"C T~~ 5F~~
-40
-20
0
30
60
80
90
-40
-.4
32
86
140
176
194
Final
RC-1
26.4
•24.8
23.8
22.4
21.3
19.8
—
Candidate
RC-2
^
46.8
46.6
46.6
46.7
-
46.8 '"
Method: Monsanto Company developed hot-wire
transient measurement
20
-------
Table 10. CALCULATED VAPOR VISC'OSITIES AT -LOW PRESSURE
•FINAL •GftNrDI;BA:TES
Absolute Viscosity of Final
Candidates (Centipoise. x IP,1*)
Temperature °F RC-1 RC-2
200 112 106
250 120 113
300 127 121
350 135 128 '
400 143 136
450 151 ' 1^4
500 159 153
550 166 160
600 174 168
650 182 177
700 190 184
750 197 192
800 205 200
21
-------
Table 11. CALCULATED VAPOR THERMAL CONDUCTIVITIES AT LOW PRESSURE
- FINAL CANDIDATES
Thermal Conductivity of Final
Candidate Btu/hr-ft-°F x 101*
Temperature °F
200
250
300
350
400
450
500
550
600
650
700
750
800
RC-1
62
71
79
88
97
107
116 .
126
136
146
156
166
176
RC-2
121
131
141
153
165
178
191
,205
219
233
248
263
278
22
-------
Table 12. PSEUDOCRITIQiL CQNSTANTS. FOR EXTEJJS:IG[N OF TRANSPORT
PRbPERTIES - FINAL CAND'tD'ATES-... .
Ps.eudocritlcai RC-1 RC-2
Temperature, °K . 508.9 615.2
Pressure;, atm. abs. 37.17 127.2
Compressibility 0.2^28 0.2567
Volume•, cmVg-mbi 272.8 101.9
23
-------
1000
100,
o
ex
>
10
250°C
200°C
150°C
100°C
j_
20 40 60 80
Mol % Pentafluorobenzene
100
Figure 2. Vapor Pressures - RC-1 and Related
-------
looocr
OJ
1_
3
t/>
IS)
o>
k_
Q.
i_
O
ex.
«J
10
0.1
20
.100°C
40 60 80
Mol % 2-Methylpyridine
100
Figure 3. Vapor Pressures - RC-2 and Related
25
-------
5.3 THERMODYNAMICS
Using the physical property data of Tables 13 and 14 in computer
program E 139 3» thermodynamic data tables were produced for each
of the final candidate fluids. The resulting tables are included
in Appendices B and C. Computation details are covered in Section
6.9. ;
Additionally, computer program TSPLOT was used to generate tem-
perature-entropy diagrams for each candidate. These are included
in Figures *J and 5.' ,
5.1 CYCLE CALCULATIONS
Cycle calculations .for the two final candidates are presented in
Table 15- Two cycles were calculated for each of the two candi-
dates, viz:
1. Reference Ideal Cycle: an ideal Rankine cycle (clockwise
rectangle on a P-S diagram) in which the. isen'tropic expansion
starts at a temperature of 712°F and ends at" the bubble point
pressure of the liquid at 220°F; with the added restriction
that the expansion entropy be the smallest value that (a)
prevents expansion into the two-phase regime, and (b) requires
an expander inlet pressure no greater than 1000 psia;
2. Equivalent Real Cycle: a non-ideal (irreversible) Rankine
cycle having identical working fluid conditions (referred to
the above cycle) at- the two "opposite" cycle corners, i.e. at
the vapor generator exit and the condenser exit, but otherwise
characterized by the following efficiencies and pressure drops:
Efficiencies (fluid indicated)
expander 75%
pump .75%
Pressure losses
vapor side, regenerator 0.09 PI*
condenser 0.14 P^
liquid side, regenerator 0.05 PI
vapor generator 0.10 PI
where PI and PI+ are the absolute pressures at..vapor generator and
condenser exits respectively; with the added restriction that the
regenerator UAk product, (or Q/ATiOg mean) be fixed at 125 Btu/HP-
hr-°F, the "largest" regenerator believed acceptable in a turbine
driven automobile according to Criterion 7, Section 6.1.
Additional explanatory notes keyed to Table 15 are:
26
-------
Table 13. PHYSICAL PROPERTY DATA USED IN THERMODYNAMIC
TABLE .GENERATION FOR FINAL CANDIDATE RC-1
Molecular Wt.
Critical Temp.
168.
(°K) 503.
re (atm) 24.
ssibility 0.
(cm3/g-mol) 356
067
7
7
2181
186.057
516.72
32.61
0.2799
364
Heat Capacity, ideal gas
(Cp = a + bT + cT2 + dT3)
Btu/lb mol-°F or cal/g-mol-°C
T in °K
a
b
c
d
Vapor Pressure
[In P° =. A"+ B/(T + c)]
P° in atm, T in °K
A
B
C
Heat of Vaporization
(cal/g-mol)
at T (°K)
Liquid Density
(g/cm3)
at T (°K)
-1.2077
0.112202
-7.108 x 10~5
1.611 x 10~8
11.72714
-4535.12
28.57
7729
358.15
1.522
298.15
0.2994
0.111666
-7.05 x 10~5
1.575 x 10~8
11.54669
•4369.7
25.287
7651
353.15
1.613
298.15
Van Laar Constants = 0 for this pair; mixture assumed
an ideal solution.
27
-------
Table 14. PHYSICAL PROPERTY DATA USED IN THERMODYNAMIC TABLE
GENERATION FOR FINAL CANDIDATE RC-2
Molecular Wt.
Critical Temp. (°K)
Critical Pressure (atm)
Critical Compressibility
Critical Volume (cmVg mole)
Heat Capacity, ideal gas
(Cp = a + bT + cT2 + dT3)
Btu/lb mole-°F or cal/g mole-
T in °K
a
b
c
d
Vapor Pressure
[In P° = A + B/(T + C)]
P° in atm, T in °K
A
B
C
Heat of Vaporization
(cal/g mole)
at T (°K)
Liquid Density
(g/cm3)
at T (°K)
Van Laar Constants (T in °K);
A12 = 3.9654 + 792.79/T -
2-Methylpyridine
(Component 1)
93.129
621.1
45.4
0.297
333.2
-4.1626
0.116616
-6.68199 x 10-5
1.301498 x 10-8
10.429
-3957
-22.68
8654
402.55
0.9497
288.15
Water
(Component 2)
18.016
647.3
218.2
0.230
56
7'. 136 ' .
0.00264
4.59 x 10-8
0
12.144
4203-9
-26.75
9717
373.15
1.
277.15
0.0089793 T
A21 = 11.984 - 1997.1/T - 0.0146616 T
28
-------
s
i
s
i
3
•
>
2
:
2
s
> i
8
- 5
5 *
i 5
i 9
.;...j._-__Uo--4--+H—:-~4-i-h-4—\-l-4- --.-4- -
I : L .! ! . , i
C6F5H. cere . eo/HO :B»LI renioj 78ososi3>ig
, ^_ _, „' i | i ' .1
Figure 4. Temperature-Entropy Diagram - RC-1
29
-------
_••
iiiwai . ira^Ni
Figure 5- Temperature-Entropy Diagram - RC-2
-------
Table 15- RANKINE CYCLE SUMMARIES, FINAL CANDIDATES
RC-1 RC-2
Cycle
Efficiencies, %
Cycle
% of Carnot
Cycle, 0% Rgn.
100% Rgn.
Carnot
Temperatures, Max, °F
Mln, °F
Pressures, Max, psia
Min, psla
Per 100 Cycle HP:
Fluid Rate, Ib/hr
Eng. Exh., cfm
Pump In, gpm
Engine HP, gross
Pump HP
Heat Flows
Heater, kBtu/hr
Regen, kBtu/hr
Conden, kBtu/hr
Without Regen
Heater, kBtu/hr
Conden, kBtu/hr
Engine: % Effic =
Pressure Ratio
Density Ratio
Isentropic (3)
Exhaust Qual, %
Delta -H, Btu/lb
Isentropic
Nozzles, Coef =
V spout, fps
Mach No. spout
A throat (4)
Regen: % Effect =
Q, kBtu/CHP-hr
UA, kBtu/CHP-hr-°F
Delta TI LM, °F
NTU, OG
I-Factor, 220°F
Pump: % Eff.
Pressure Ratio
Notes covered in text.
Ideal1
30.00
71.45
20.54
30.69
42.00
712
220
1000
27.31
7470
259.1
12.24
107.34
7-34
848.4
390.5
408.5
1238.8
799-0
100
36.62
44.55
44.55
176.5
36.57
36.57
1.00
1353-0
2.48
0.052
95.3
390.5
11.98
32.6
7-47
0.719
100
Real2
20.95
49.88
13.35
22.07
42.00
712
220
1150
27.31
11,521
340.0
18.81
116.84
16.84
1214.7
691.7
949.5
1906.5
1641.2
75.0
29.77
37.91
15.4
189.7
25.81
15.29
0.95
1262. 4
2.40
0.118
91.8
691.7
12.5
55.6
4.92
0.719
75
36.62
42.11
Ideal1
28.8
68.57
26.84
28.8
42.00
712
220
1000
22.94
1756
233.4
3.93
102.43
2.43
883-3
64.8
628.8
948.1
693.6
100
43.59
35.60
35.60
107.5
148.5
148.5
1.00
2726
2.95
0.066
100
64.8
oo
0
00
1.379
100
43.59
Real2
22.15
52.74
18.57
22.23
42.00
712
220
1150
22.94
2521
310.1
' 5-66
105.04
5.04
1148.9
212.8
890.1
1361.7
1102.9
75
35.44
32.94
12.06
118.1
106.0
141.4
0.95
2587
2.74
0.049
98.4
212.8
12.5
23.0
7.36
1.379
75
50.13
31
-------
3- The isentropic density ratio is calculated over a
hypothetical isentropic expansion from engine inlet
condition to an enthalpy equal to real cycle engine
exhaust enthalpy; it is, then, a good approximation
of the minimum permissable expansion ratio [(displacement
+ clearance volume)/(clearance volume)] to be provided
in the design of a -reciprocating expander of matching
performance. Refer to Criterion 9 of Section 6.1.
fl. The combined turbine nozzle throat cross sectional
area required per 100 cycle HP assuming choked flow
(cycle HP is net .fluid-indicated HP, expander less
pump)... Additional computation details are given, in
Section. 6.10..
A comparison of. the data of Table 15 with the fluid selection
criteria of Section 6.1 permits several interesting observations.
Candidate RC-2 fails the 30% ideal cycle efficiency dictum of
Criterion 6, yet it has a higher real cycle efficiency than does
RC-1 by over two percentage points. ,It therefore appears that
RC-2 should not be disqualified solely on the basis of its ideal
cycle efficiency of 28.8%. While RC-1 shows ideal and real cycle
efficiencies, of 30 and 21%, it may be noted that RC-1 was found
to be more thermally stable at 720°F than RC-2 so it actually
could be pushed higher in temperature to where, it would show
even more advantageously in efficiency than RC-2. By Criteria 8
and 9j both candidates qualify for both reciprocating .and turbine
expanders.
The assumption of a "smaller" regenerator than the UAk = 125 Btu/HP-
hr-°F of the Table 15 real cycle calculations, as to account for
the lesser available compartment space with a reciprocating plant,
would lower the real cycle efficiencies on both candidates, and
would therefore make fluid RC-2 look more attractive in a recip-
rocating plant competing with a turbine engine utilizing RC-1.
Table 15 also reveals a significant difference in working fluid
circulation rates compared at equal cycle power levels. About
four times as much RC-1 needs to be circulated as RC-2 (by weight).
One consequence is the need for a three times as powerful pump with
RC-1. Another consequence is the need for larger conduits and com-
ponents with RC-1 to avoid the higher pressure losses normally
accompanying higher mass flow rates. These disadvantages of the
RC-1 fluid are traceable to the lower available energy of expansion
per unit mass associated with higher molecular weight fluids.
Attempts were made to quantize molecular weight as a selection
criterion despite a lack of contractual restrictions. The attempts
failed because of the ill-defined and complex ways molecular weight
enters into system goodness.
32
-------
A fact not apparent in Table 15 is that the RC-1 cycles are super-
critical, whereas the RC-2 cycles are subcritical. Again, no re-
strictions were imposed in this case. It appears that each type
of cycle has its supporters and detractors, and no judgment is
offered here on which may be superior.
5.5 STABILITY/COMPATIBILITY TESTS
5-5-1 Dynamic Loop Testing
Final candidate working fluids were subjected to extended loop
testing under conditions approximating prototype. The purpose of
dynamic loop testing was to demonstrate fluid stability, materials
compatibility and energy exchanges in a real flowing system.
5.5.1.1 Description of Loops - Two identical dynamic loops were
constructed. Figure 6 is the loop schematic showing the arrange-
ment of the various components. Table 16 gives component details.
An overall view of one of the loops appears in Figure 7- Procedure:
followed in conducting the tests are given in Appendix D.
Each heater coil, as noted in Table 16, was made by bending a 100-
inch length of 0.25 inch OD SAE 4130 steel tubing into the trombone
shape shown in Figure 8. The tubing was used as received, with no
internal pretreatment other than the solvent cleaning procedures
of Appendix D. The internal surface of the new tubing was found
to have a smooth dull black oxide finish. A photomicrograph
appears in Figure 9a.
In use, the heater coil was heated by Dowtherm A condensing on
the outer surface of the tubing. This heating mode was provided
to eliminate all question of the presence of hot spots along the
coil. Hot spots are difficult to control, especially when elec-
trical heat is used to boil a fluid to dryness at a pressure below
its critical pressure.
Condenser coils, as typified in Figure 10, were fabricated by
wrapping the 0.25 inch OD 3003 aluminum tubing around a mandrel.
The tubing was used as received. Solvent rinsing (as described
in Appendix D) was the only internal treatment employed. Figure
9b shows the initial surface. To the eye this surface appeared
as a gray oxide finish with very distinct drawing marks.
During testing, the condenser coil was immersed in hot flowing
tap water. In order to prevent electrolytic corrosion of the
outer surface of the aluminum, a sacrificial anode system was
employed. The flowing water (the electrolyte) was confined jn a
plastic beaker. A 1.5-volt dry cell was connected to the conden-
ser coil (the cathode) and to the sacrificial anode (a one inch
33
-------
lOOcc '
Syringe
Needle/ b—
Septum T
XD
Vent Glass
n Let-Down "5
Valve -£
I Aluminum. ^> 3
TP Condensing <; "
Y
1
Sample r
Valve ^
i
^ F
(jHx]—
0-200psig
TC
LOII-
r '^if ^
S.J J|
X
Precooler
•
Water In
0<
Fill
Glass
316 SS i
— vr ^-»
7
L.
i-
-
High Pressure
^-— — Shutdown
/^ Fill-Bleed
i
\ Fluid
Heater
i i
I !!.
rib
0^?
U1 Q.
"5 E,
TC
J
Dowtherm A ;
Heating Loop
Dowtherm
i — ' — i Boiler
2.5 KWe
L
— 1
n
1 Temp. io_
20-30^ ^--> Cont. Z
Filter'
• r+
I k
\
0-3000psig .
A Diaphragm-Type
Circulating Pump
• Alternately located at'X1
Figure 6. Dynamic Loop Schematic
-------
Item
Pump
Filter
Pulse Damper
Heater Coil
Precooler
Table 16. DYNAMIC LOOP PARTICULARS
Description
Lapp CPS-1, variable positive displacement,
SS diaphragm and wetted parts, PTFE seals.
Hoke #6311043, 20-30 micron SS frit element
7 in. length, 1/2 in OD, SS tube, vapor cushion
SAE 4130 steel, 0.25 in. OD x 0.035 in. wall
x 100 in. long, formed into 4-pass hairpin coil
7 in. length* of water jacket around run of
0.25 in. OD x 0.035 in. wall SS tubing
Let-down Valve
Condensing Coil
Sample Valve
Loop Tubing
Loop Fittings
Dowtherm A Loop
Temp. Controller
Temp. Recorder
Pressure Gauges
Republic #644X6-9-4-2, 5/16" D. ball subs.
for poppet
3003 aluminum, 0.25 in. OD x 0.035 in. wall
x 60 in. long, formed into 7-turn helical coil.
Hoke #7H5G4Y with Swagelok #100-R-2-3l6,
fitted with a septum, silver soldered in the
side at ball centerline.
316 SS, 0.25 in OD x 0..035 in. wall (except
heater and condenser coils above)
316 SS Swagelok, 303 & 316 SS Hoke valves,
SS ferrules
Boiler: 2.0 in. OD x 21 in. long, 10 x 250 W
cuff heaters
Exchanger: 2.0 in. OD x 26 in. long
Vapor tubing: 0.5 in. OD, 316 SS
Liquid tubing: 0.375 in. OD, 316 SS
Insulation: 1" silicate pipe cover, 1"
ceramic blanket
Galvanometer type, on/off mode, I/C couple
Brown multi-point type, I/C thermocouples
Steel bourdon tube type, 0-200 and 0-3000
psig, with zero-pressure pointer stops removed.
*See text for exception, second RC-2 test
35
-------
Figure 7. Dynamic Loop
36
-------
Figure 8. Heater Coil - Ends Cut and Sectioned
Lengthwise for Examination
37
-------
9a. Steel (Heater)
9b . Aluminum (Condenser)
Figure 9- Photomicrographs of Heat Exchanger
Surfaces before Use (2?x)
38
-------
Figure 10. Aluminum Condenser Coil
39
-------
square of sheet magnesium) suspended at the center of the coil.
Every 2-3 days the deteriorated anode was replaced. This system
effectively prevented external corrosion, but did not, of course,
prevent calcium/magnesium scaling.
This design was based upon a similar dynamic loop constructed and
operated at Aerojet Liquid Rocket Company, Sacramento, in 1971-2.
Features common to the two designs include the tubing materials
and sif.es and general component arrangement. The major features
of the present loop not borrowed from the Aerojet design are:
- heating of fluid by condensing Dowtherm A vapor
(vs. electric;ML resistance windings)
:.%?•', - • - ..,.,•/.£>:';
- positive displacement.pumping
5-5.1.2 Test Description - Conditions of testing were as listed
in Table 1?.
Table 17- DYNAMIC LOOP TEST CONDITIONS
Controlled..Values ..
pump discharge,pressure
heater exit, 'bulk temperature
pump-' suetio'n ^pressure
f 1 u,id. flow,;lrate"; '
Typical Values1 ' '
pump suction temperature
cooling water in temperature
Dowtherm A condensing pressure
Dowtherm A condensing temp.
system volume (measured)
heater tube volume (calc.)
Elastomer Seais
pump diaphragm, check valves
filter
let-down and sample valve 0-rings
sample valve ball seals
all other valves
1000 psig
720°F
15-30 psig
see text
100-110°F
40-60°F
120-160 psig
735-770°F
190-200 cc
ill.6 cc
PTFE
PTFE
see text
Kel-F or PTFE
Kel-F or PTFE
-------
The detailed test procedure is presented in Appendix D. A test
"pass" is 1000 hours of successful operation with no operational
problems attributable to fluid instability, incompatibility or
corrosion.
5.5.1.3 RC-1 Test Results - The test of RC-1 started 18 October
1972 under the conditions and with the results listed in Table 18.
Early in the test a very slow but persistent leak developed. The
source of leakage was eventually traced to the end fittings on the
aluminum condenser coil (through the use of a sensitive thermal
conductivity vapor detector).
At '198 hours the test was interrupted long enough to remove the
stainless steel ferrules originally installed in the end fittings
and replace them with aluminum ferrules. As this cured the chronic
leakage problem,, it was subsequently adopted as standard. A mis- _
match in elasticity or thermal expansion coefficient between the
alumiinum and stainless steel is believed to have been the cause of
the leakage.
In all other respects the test was nominal. The outcome was,
therefore, rated "pass".
Table 18. LOOP TEST SUMMARY, RC-1
Test & fluid
Outcome
Dates
Duration (at temp.)
Termination Reason
Pumping rate
0-ring elastomer
Zero time exceptions*
Forerun
Condenser Ferrules
Letdown valve
' Filter
Gas measurement
Vent glass tee
Fluid additions
Vented gas
Cumulative RTF volume
Over span of
For average rate of
Filter changes
RC-1
pass
18 Oct. to 6 Dec. 1972
1172 hr
time expired
110 ccprn
Viton (Parker 7^7-7),
16 cc (in first 1000 hr)
0 cc
QUO hr
0.0 ccph
none
*Exceptions to the data of Table 16 or the procedures of
Appendix D.
-------
Microscopic examination of internal heat exchange surfaces,
exposed by slicing sections from each end of each coil, produced
these observations:
1. Heater inlet (cold) end (Figure lla); smooth black
oxide finish as in new tubing, but with a band of
brown deposit at entry end presumed to be ordinary rust,
Fe203.
2. Heater outlet (hot) end (Figure lib); smooth black
oxide finish as in new tube but with off-black
striated patches visible - no loose deposits.
3- Condenser inlet (hot) end (Figure lie); same appearance
as new tubing.
4. Condenser outlet (cold) end (Figure lid); same
appearance as new tubing.
Fluid removed from the loop at 1002 hours appeared water white.
Analyses of the new and used fluid by gas-liquid chromatography
and mass spectroscopy permits the conclusion that only the
subtlest of chemical changes occurred. The major impurities
present in the new and used fluid and their approximate concen-
trations were as follows:
Parts per Million
Contaminant New Used
fluorinated aliphatics 410* 710
chloropentafluorobenzene 4200** 3100
unknown, MW - 258 0 820
*primarily from the hexafluorobenzene sample
**primarily from the pentafluorobenzene sample
These data suggest that the main decomposition mechanism was one
that converted the main initial contaminant, chloropenta-
fluorobenzene, into a higher molecular weight compound and one
or several lower boiling fluoroaliphatics. This is plausible in
light of the finding reported elsewhere that chlorinated compound::;
are simply not as stable as fluorinated compounds. In any event,
the extent of decomposition was well below the levels of the
contaminants alone in the starting fluid.
5-5.1-4 First RC-2 Run Test Results - The first RC-2 test,
summarized in Table 19, revealed several flaws in the original
loop design, one of which was responsible for the early abortion
of the test.
-------
lla. Inlet (Cold) End lib. Outlet (Hot) End
Steel Heater Coll
lie. Inlet (Hot) End lid. Outlet (Cold) End
Aluminum Condenser Coil
Figure 11. Photomicrographs of Heat Exchanger
Surfaces after RC-1 Test (2?x)
-------
Table 19. LOOP TEST SUMMARY, FIRST RC-2
Test & fluid
Outcome
Dates
Duration (at temp.)
Termination Reason
Pumping rate
0-ring elastomer
Zero-time exceptions*
Forerun
Condenser ferrules
Letdown valve
Filter
Gas measurement
Gauge glass tee
Fluid additions
Vented gas
Cumulative RTF volume
in span of
for average rate
Filter changes
RC-2 (first)
Fail - incompatibility
6 to 23 October, 1972
400 hr
fluid loss, flow loss
approx. 56 ccpm (estimated)
EPR rubber (Parker E515-8)
stainless steel
unmodified (as received)
not provided (until 287 hr)
15 cc
450 cc (287 to 400 hr)
113 hr
4.0 ccph
One (305 hr)
*Exceptions to Table 16 or Appendix D.
-------
During the first week of the test, difficulty was experienced in
maintaining a constant 1000 psig pump discharge pressure, and
many adjustments of the let-down valve bonnet were needed. Some
slow leakage was experienced, mainly through the 0-ring seal on
the bonnet. In addition, no diminution in the amount of vent pas
was noted after the first few days operation.
At the 305-hour mark, the test was interrupted to repair leaks in
the sample and let-down valves. The sample valve was replaced on
account of galled and leaky threads and a new EPH 0-ring was
installed in the let-down valve after resurfacing the valve seat
and replacing the original poppet valve with a 5/16 in. D. stee]
ball. Loose black deposits found in the porting areas of the valve
in disassembly were collected and analyzed. Similar material was
found on the filter element (which was replaced) and in fluid
drained from the heater coil.
A 100-cc syringe with 3-way valve was mounted above the vent glasy
to permit measurement of vent gas volumes. The arrangement is
shown in Figure 12. The original fluid drained from the system
was replaced, and 15 additional cc of new fluid were needed to
make up for disassembly losses. The test was then restarted.
Operation continued fairly smoothly for the next 50 hours, but
then erratic high pressure control was again experienced. At 400
hours the test was aborted upon discovering loss of fluid cir-
culation and excessive fluid leakage through the fitting at the
condenser coil inlet.
During the last 113 hours of this test, some 450 cc of vent gas
were collected. The major constituent of this gas was found by
mass spectroscopy to be hydrogen. The black deposits mentioned
above were observed to be fully magnetic. By x-ray diffraction
it was discovered that a major (if not sole) ingredient of this
material was Fe30i+
The test was rated a "fail". Erratic control of pump discharge
pressure is ascribed to black iron oxide deposits in the let-
down valve preventing normal valve seating and free valve movement.
Microscopic examination of internal heat exchange surfaces, exposed
by slicing sections from each end of each coil, produced these
observations.
1. Heater inlet (cold) end (Figure 13a); smooth black
oxide finish as in new tubing, but with an inch wide
band of brown deposit at entry end presumed to be
ordinary rust; one very striking corrosion pit was
noted in this portion of the coil (as shown in Figure
14; no others were spotted, so the one must be con-
sidered atypical.
-------
Figure 12. Vent Gas Syringe Arrangement
-------
13a. Inlet (Cold) End 13b. Outlet (Hot) End
Steel Heater Coil
13c. Inlet (Hot) End 133. Outlet (Cold) End
Aluminum Condenser Coil
Figure 13. Photomicrographs of Heat Exchanger Sur-
faces after First RC-2 Test (2?x)
-------
Figure 14. Photomicrograph of Pit in Heater
Tube after First RC-2 Test (2?x)
-------
2. Heater outlet (hot) end (Figure 13b); smooth black
oxide finish as in new tubing, but overlaid with patchy
clusters of loosely-adherent shiny black crystals of
Fe304
3- Condenser inlet (hot) end (Figure 13c); brown deposit
estimated half mil thick covering surface everywhere
but in isolated silvery patches; scrapings are magnetic
and are presumed to be FesO^ crystals embedded in a
whitish inorganic matrix.
4. Condenser outlet (cold) end (Figure 13d); same as c.on-
denser inlet except deposit is thinner, smoother and
possesses fewer gaps.
Evidence points clearly to a corrosion reaction between water and
iron,
3Fe + 4H20 »• FesO^ + ^H2,
in the steel heater tube; which resulted in the copious release
of gas and .formation of Fe30i+ particles. The former required.
frequent:purging, while the latter ultimately caused malfunctions
in the letdown valve. Both.-of these .constitute' evidence of in-
compatibility. • • .
5.5.1.5 Second RC-2 "Loop Test Results - The first RC-2 loop test
was ended prematurely by .the steady formation of magnetic.iron
oxide and- hydrogen, both of which interfered with nominal operation.
Continuous reaction of water with iron, however, is not the in- '
evitable result of contacting water with steel at elevated tem-
perature. Many years of experience with steel-tubed boilers and
superheaters in power plants proves this. On startup of new steel
tube boilers, it -is common to find significant amounts of hydrogen
in the steam for the first day or so of operation, but the for-
mation rate then declines to practically zero for the rest of the
useful life of the tubing. This self-passivation phenomenon is
thought (ref. ;l) to depend on the formation of a chemically-.
resistant film •of oxide and possibly of hydrogen on the metal
surface. Hydrogen film passivation is called polarization.
Failure of the steel heater tube to self-passivate could be linked
to: .
a. improper surface treatment
b. incorrect choice of steel
c. presence in the water of the 2-methylpyridine
d. combination of above
-------
A second loop test of RC-2 was scheduled to try to resolve this
issue. Attempts to purchase 0.25 in. OD steel tubing ,t.o. ASTM
specifications A210 or A213 Grade Til (typical steels for boiler
and superheater tubing) failed. Stainless steel could .have been
chosen and likely would have solved the corrosion problem, but
stainloss steel appears not to satisfy the criterion.-qf. a "low
cost" metal of system construction. Other low cost .s.fce.e.ls which
mitfht have been chosen were not, since there was no reason to
believo they would be better than the SAE 4130.
The decision was therefore made to retain the SAE 4130 steel for
the heater tube. It was further decided to attempt:to pre-
passivate this tubing in-situ by "running in" on pure water. If
the steel could be prepassivated with water, it was reasoned,
inhere would then be hope that it would remain passive to the
RC-2 fluid. This was the rationale behind the second RC-2 lo.op1
test.
Several changes in the loop hardware were made prior to starting
to ease problems encountered in the first RC-2 test. The most
important of these was the relocation of the filter (to point "X"
in Figure 6) to keep solids from interfering with the let-down.
valve. To accommodate this change, it was also necessary to
enlarge the precooler (to 12 in. length) so as to drop -the filter
body temperature below the 500°F limitation of its PTFE seal
gasket. Another change was the replacement of the vapor-separa-
ting tee at point "Y" of Figure 6 to secure more positive liquid-
vapor separation. This was done because of a suspicion that gas
bubbles had been accumulating in pump suction regions and had been
contributing to flow stoppages.
The change consisted of the substitution of a 1/4 inch IPS stain-
Less steel pipe tee for the original tubing tee. The 1/4-inch OD
inlet tube was allowed to project into the tee body through the
lateral opening. Its end was sectioned laterally for. about one- .
half inch to afford an easy upward escape of bubbles. Although
this change undoubtedly improved the gas-liquid separation'
efficiency of the vent tee, it brought about no noticeable changes
in loop operation or behavior.
On 5 December 1972 the loop was filled with distilled water and
put into operation under standard conditions. After the first
four days of running no bubbles appeared at the sight glass. This
was Interpreted as successful attainment of the passive condition.
This forerun was ended after 285 hours hot operation.
The water was drained and blown from the loop and new RC-2 charged
on 18 December 1972. The second test of this fluid then commenced
50
-------
under the conditions and with the results of Table 20. The run
proceeded smoothly until 167 hours when it was necessary to shut
down to clean scale from the outside of the condenser coil and
clean up a resultant water spill. During this shutdown inspection
of the filter element showed it to be heavily clogged with black
oxido par-tides, so it was replaced.
At 2/;8 hours tne test was again interrupted to check the filter
element. The element was found to be still servicable and only
slightly clogged. The element was nevertheless cleaned and re-
insta.1 led .
Another shutdown occurred at 332 hours. This was occasioned by
leakage through the PTFE seal gasket on the filter body. From
this point to che final shutdown at 1011 hours, operation was
entirely nominal except for gas venting,.
Table 20. LOOP TEST SUMMARY, SECOND RC-2
Test ft fluid
Outcome
Dates
Duration (at temp)
Termination reason
Pumping rate
0-ring elastomer
Zero time exceptions*
Forerun
Condenser ferrules
Letdown valve
Filter
Gas measurement
Vent glass tee
Fluid additions
Vented gas
Cumulative RTF volume
Over span of
For average rate of
Filter changes
RC-2 (second)
fail - incompatibility
18 Dec. 1972 to 31 Jan. 1973
1011 hr
time expired
56 ccpm
EPR rubber (Parker E515-8)
285 hr @ 720°F, dist. H20
relocated to "X" Figure 6
started at zero time
tubing tee replaced by 1/2"
SS pipe tee
38 cc
519 cc
1011 hr
0.5 ccph
one (167 hr)
*exceptions to Table 16 or Appendix D
51
-------
Gas evolution was experienced throughout the run. Cumulative
volumes vented at RTF (room temperature and pressure)-at 'various
times were as follows:
Test hr 93 196 380 456 648 843 1011
cc Gas 112 156 270 290 374 459 '519
When graphed, these data show an initial high gas evolution
followed by a-fairly steady rate of generation.throughout.the
remainder of the test. .
The test was rated a "fail" primarily because of the need to, vent
the hydrogen gas. While the water forerun failed to fully
passivate the steel surface, it did greatly reduce the, rat.e of
reaction, as shown by the eight-fold reduction in the rate of
gas evolution compared to the first RC-2 run. •
Microscopic examination of internal heat exchange surfaces,
exposed by slicing sections from each end of each coil, produced
these observations: •
1. Heater inlet (cold) end (Figure 15a); smooth black
oxide finish as in new tubing but with overlying tan-
colored striations of a thin deposit extending down-
stream from a point 4 inches inside the heater shell;
appearance suggests an inorganic material laid down
when the fluid evaporates. '' ;
2. Heater outlet (hot) end (Figure 15b); smooth black
oxide finish as in new tubing but covered with dense
fine-grained coating of black crystalline material..
presumed to be magnetite, Fe30tt; this coating was much
more dense (i.e. covered larger fraction of the -visible
surface), finer-grained and more tightly adherent t/o
the substrate than that found in the same loca'tion
after the first FC-2 test. '
3. Condenser inlet (hot) end (Figure 15c); tan-'colored
deposit filling in most of the original working marks
(micro-scratches along tube axis) on the aluminum sur-
face; color off-white when topping scraped off; deposit
non-magnetic and tightly adherent.
4. Condenser outlet (cold) end (Figure 15d); 'same ap-
pearance as inlet section above. "
Fluid removed from the loop after 1001 hours of testing was deeply
colored (deep amber) but relatively free of suspended solids.
-------
15a. Inlet (Cold) End 15b. Outlet (Hot) End
Steel Heater Coil
15c. Inlet (Hot) End 15d. Outlet (Cold) End
Aluminum Condenser Coil
Figure 15- Photomicrographs of Heat Exchanger Sur-
faces after Second RC-2 Test (27x)
53
-------
Comparisons of the gas-liquid chromatograms of new and used
(1001 hours) fluid revealed only the subtlest of composition
changes. The new fluid was found to have three minor peaks
(contaminant species) in addition to the two major peaks (water
and 2-methylpyridine). The used fluid showed the same three
peaks but showed additionally a small amount of a higher boiling
material. By area ratios the concentrations of the various
tramp-compounds were as follows:
Species,
Elution Time Concentration (ppm)
(min 6 100°C) New Fluid Used Fluid
3-1 ± .15 1800 1800
4'8 ± '2 1700 960
5-1 ± -25
8.4 ± .3 0 370
By way of comparison, the water peak elutes at 1.4 minutes (with
an exponential tail extending well past 14 minutes) and the 2--
methylpyridine peak at 3-8 ± 0.2 minutes. (The chromatograph
column consisted of a 4 meter length of 1/4-inch OD tubing packed
with SE-52 on high performance Chromosorb W - helium carrier,
thermal conductivity detector.)
The chemical identities of the minor contaminants were not deter-
mined .
5.5-2 Capsule Compatibility Tests at 300°F
In a real Rankine engine, as opposed to the dynamic loops of
Section 5.5.!, a wider variety of materials must be in contact
with the working fluid, and the question of compatibility of the
working fluid with different materials arises. Welded low carbon
and low alloy steels are the preferred materials in the hot, high
pressure portion of the cycle. Many other materials, including
elastomers, plastics, and bearing and sealing elements, are re-
quired in the cooler parts of the cycle to facilitate power trans-
fer, carry moving loads, seal against leakage, and so on.
Capsule tests at 300°F were selected to test the compatibility of
the final candidate working fluids with representative materials.
Capsules constructed of stainless steel and having internal dimen-
sions of 1.5 in. D x 5-5 in. long were used in this work. Into a
capsule were placed the previously weighed solid specimens and
then 30 cc of the liquid sample. The capsule was then closed off
with a PTFE-sealed screw cap and placed in a 300°F thermostated
54
-------
oven for 164 hours. At the end of this time the capsules were
removed, cooled, opened and inspected. Solid specimens were
weighed again and the changes in weight and appearance noted.
Results are summarized in Tables 21 and 22.
5.6 TOXICITY
The problem of assessing toxicological hazard is a complex one.
It may be assumed that accidental and uncontrolled personal con-
tacts with automotive working fluids would inevitably follow any
widespread adoption of Rankine cycle automobiles. The hazard of
such a contact, whether it occur via the skin, eyes, respiratory
system or digestive tract, may be thought of as a product of an
inherent toxicity times a dosage factor times the probability of
reaching that dosage. As an example, the 50% lethal dose (LD50)
for table salt in rats is, by 1972 HEW data, 3 grams salt per
kilogram body weight. A rat force-fed this amount of salt by
stomach tube has a 50 percent probability of dying of acute salt
poisoning. The probability of a. rat accidentally ingesting this
much salt is extremely small. Therefore, despite its inherent
toxicity, table salt ranks far below cats in a listing of every
day hazards faced by the rat population.
Considering how a working fluid might be manufactured, stored,
transported, charged, used, serviced and replaced, it seems clear
that vapor inhalation is the single mode of accidental contact
most difficult to prevent and most likely to occur. This suggests
that the primary concern in consideration of possible toxic
hazards should be with acute intoxication by vapor inhalation. A
model was adopted to assist in estimating accidental vapor in-
halation hazards for candidates RC-1 and RC-2.
5.6.1 Vapor Hazard Model
While a full assessment of toxic hazards is a monumental experi-
mental undertaking beyond present resources, it nevertheless is
desirable to quantify, even if very roughly, the likelihood of
inhaling significant quantities of working fluid vapors. As a
step in this direction, a "worst case" accident is visualized as
follows: a Rankine passenger car suffers a front end collision
which dumps on the ground 0.25 cubic feet of working fluid and
entraps the occupants for 20 minutes within a few feet of the
spilled fluid. A mathematical model describing the evaporation
and dispersion of the working fluid was set up employing these
assumptions:
1. spherical symmetry and directional uniformity
»
2. evaporating fluid puddle represented as a sphere
of constant radius R
o
3. partial pressure equal to vapor pressure p* at
surface of "puddle" sphere
55
-------
Table 21. CAPSULE COMPATIBILITY TESTS - RC-1
164 Hours at 300°F
in Final Candidate RC-1
Specimen Capsule
Brass ; A
Bronze (brg. type) . -A
Silver solder - A
Aluminum 3003 ' • .A :
Teflon (O-.r-ing) ' A
Viton A (0-ring) .. . B
Fiberite Fi'1 4 005 B
Viton coatee Daeron B
Polyimide B
Polyprop;-. ' one B
Gjf ci'p: : '• ' e B
Ryt-..' T3 C. •
Delrin C
DacT on fabric . • C
-ja el ' C
Specimen
Initial
1.3258
1.2895
0.17^5
0.2114
1.9200
1.5932
0.2817
0.1673
0.4438
0.1171
1.2938
0.3222
0.5824
0.0482
0.7615
Wt. (g)
Change
0.0001
0.0187
0.0000
-0.0001
0.1394
1.3
-0.0027
-0.0250
0.0005
-
0.0941
0.0044
0.0256
0.0007
0.1250
Notes
SI. absorption
Mod; absorption
High absorption
Separated
Dissolved
Mod. absorption
Turned yellow
Mod. absorption
Not : or the materials listed, the Viton coated Dac'ron and the
:•••{'•> .i ypropylene are definitely incompatible with RC-1.
.. . l.rin is questionable. Viton A tends to strongly absorb
' he . fluid and. swell, yet it was successfully used to seal
the letdown valve in the -1100+.: hour dynamic loop .test.
All other materials are considered compatible.
56
-------
Table 22. CAPSULE COMPATIBILITY TESTS
164 Hours at 300°F
in Final Candidate RC-2
Specimen
Brass
Bronze (brg. type)
Silver solder
Aluminum 3003
Teflon (0-ring)
Viton A (0-ring)
Fiberite FM-4005
Viton coated Dacron
Polyimide
Polypropylene
Graphite
Ryton PPS
Delrin
Dacron fabric
Specimen Wt. (g)
Capsule Initial Change
A 3-9247 -0.0098
A
A
A
B
B
B
C
C
C
D
D
E
E
4
0
0
1
1
0
0
. 5
0
1
2
3
0
-2332
-2559
• 5913
• 9163
.6076
.4645
• 3829
.7022
.6655
• 3514
.4566
.6003
.1896
0.
0.
-0.
0.
0.
0.
0.
0.
0615
0002
0008
0082
-
0197
-
-
0420
0574
1014
-
-
Notes
Dk. coating,
some attack
SI. absorption
Some attack
SI. absorption
Disintegrated
Mod. absorption
Disintegrated
Black coating
Embrittled
Mod. absorption
Mod. absorption
Dissolved
Dissolved
Note: These data show that RC-2 is more difficult to confine.
Viton, polypropylene, Delrin, Dacron and polyimide are
incompatible under test conditions. Brass (and copper)
should be considered unsuitable. Ethylene-propylene rubber
is a suitable elastomer with RC-2, as proven in the 1000+
hour dynamic loop test.
57
-------
Jr. No vapor in atmosphere at time zero
5. Pick's Law of diffusion applicable, with diffusivity D
(interpreted as an eddy diffusivity in the absence of
an average wind) . '.-...
With these assumptions, the vapor concentration as a function of
time and distance is given by the diffusion equation as:
(1)
where p = partial pressure of working fluid in air
p* = vapor pressure of working.fluid
R = radius of the spilled sphere
r = distance from center of spill
D = diffusivity
t = time
erf = error function
erfc = 1-error function (proportional to the integral
of the normal probability distribution) •
This is illustrated in Figure 16 'where it is seen that at various
times (Tj, T2, T3, etc.) the concentration of vapor with distance
from the spill would vary in the manner shown. Eventually the
puddle would dry up and the concentration profiles then would
decay in the manner of curves T4 and T5.
At infinite time, the asymptotic (maximum) concentration in air
is given by .
_ o_ \j / ^ \
However, the spill could hot last an infinite time,, nor would
trapped occupants remain at the scene an infinite time. Although
there is no analytic solution for the exact lifetime of the puddle,
the rate of evaporative disappearance of spilled liquid, in terms
of the change of radius of a shrinking liquid source sphere, is
given by
-» T-\ HOT-. IE 1 1 -i I
(3)
58
-------
Distance
Concentration Profiles
10'
Trapped
Occupant
Plan View
Figure 16. Vapor Inhalation Hazard Model
59
-------
where M = molecular weight of working fluid,
p = its density,
RG = the ideal gas law constant, and
R = instantaneous puddle radius.
A conservative estimate of the puddle life is obtained by assuming
R in the 1/R term to be constant at R . The puddle life then is
somewhat less than
1 + /I + "j*"^ I (ll)
After time t~, the source of vapor no longer exists. An approxi-
mation of p vs t at t = t , assuming no further vapor generation,
yields the subsequent vapor dispersion by the formula
(5)
Thus, equation (1) applies ..when t < t , equation (4) estimates
t_, and equation (5)'applies when t > t^. The maximum concentra-
tion at any time-position point is given by equation (1) at t =
V
In the calculations reported here, it was assumed that the "occu-
pants" were trapped for 20 minutes, at an effective distance from
the spill of ten feet (actual distance six feet plus four feet
allowed for shielding effect of enclosure). The amount of fluid
spilled was assumed to be 1.9 gallons (0.25 cubic feet) and the
diffusivity was varied over a wide range of values.
Effective atmospheric diffusivity varies considerably, depending
on conditions. However, in a region near the ground and near a
spill, a value of 18 cm2/sec appears to be a reasonable choice for
the hypothetical worst case accident. This was derived from data
given by Turner (ref.. 5). Calculations presented here cover a
substantial range above and below this value. The principal
results are presented in Tables 23 to 25. The first two of these
show average concentration and "dose" (or concentration integrated
over the 20-minute time interval) for each of the candidate fluids
as influenced by diffusivity. It is. interesting to note that there
is a maximum "dose" in each case at some high value of diffusivity.
Table 25 shows the concentration as a time function at a single
diffusivity for both candidates.
60
-------
Ta'n±e 23- VAPOR DISPERSION CALCULATION
Fluid: Final Candidate RC-1 (fluorobenzene mixture)
Molecular Weight 175-3
Molecular Diffusivity* 0.0585 cm2/sec
Ambient Temperature 70°F
Vapor Pressure 0.169 atm. abs.** (169,000 ppm)
Fluid Density 1.58 g/cm3
Assumed Diffusiv., cm2/sec
at 10 ft distance: 1. 4. 16. 64. 256. 1024.
20 min. "dose," ppm-min. 0;2 x I0~k 59.8 6300 34,700 54,900 28,200
-Average cone., ppm 0.9 x 10~6 3-0 315 1,730 - 2,740 1,410
* Method of Fuller, Schettler and Giddings, (ref. 6) assuming (£v)M = x^EvJj + x2Uv)2
** Taken at a temperature midway between normal boiling point and 70°F
Table 24. VAPOR DISPERSION CALCULATION
Fluid: Final Candidate RC-2 (2-methylpyridine/water)
Molecular Weight 44.3
Molecular Diffusivity* 0.144 cm2/sec
Ambient Temperature 70°F
Vapor Pressure** 0.110 atm. abs. (equiv. cone. 110,000 ppm)
Fluid Density 0.950 g/cm3
Assumed Diffusiv., cm2/sec
at 10 ft distance:
20 min. "dose," ppm-min.
Average cone., ppm
* Method of Fuller, et al. (see footnote, Table 23)-
*-* Taken at a temperature midway between normal boiling point and 7r"i°?.
1.
0.00
0.00
4.
38.9
1.95
16.
4100
205.
64.
22,600
1,130
:256.
46,100
2,310.
512.
56,000
2,800
-------
Table 25. VAPOR DISPERSION CONCENTRATION HISTORIES
Assumed Diffusivity 18 cm2/sec
Distance from Spill 10 feet
See Tables 23 and 24 for other
properties
Time Since Concentrations, ppm
Spill, min. RC-1 RC-2
0
2
4
6
8
10
12
14
16
18
20 .
20 min. avg. cone. 390 ppm 254 ppm
. 0.00
0.055
10.8
66.6
171
306
455
609
761
908
1049
0.00
0.036
7.0
43.4
111
199
296
396
495
591
683
62
-------
5.6.2 Rat Inhalation Studies
As already noted, inhalation of working fluid vapors represents
the most likely avenue of accidental exposure. A search of the
literature revealed a paucity of information regarding vapor
toxicity of the methylpyridines. In the case of RC-1 some data,
mainly concerned with the anesthetic and analgesic action of
highly fluorinated benzenes, were available (ref. 7 to 11).
These investigations indicated that vapor toxicity of penta- and
hexafluorobenzenes is quite low. A Russian investigator (ref. 12)
reported the median 2-hr lethal air concentration (LC50) of
hexafluorobenzene with white mice is 95 mg/liter (12,500 ppm) as
compared to 37 mg/1 (13,700 ppm) for benzene itself.
In the present investigation, 4-hr LC50 rat inhalation determin-
ations were made with both final candidate fluids. Complete
details of these experiments are given in Appendixes E and F.
Table 26 summarizes pertinent features of the experiments and the
LC 5 Q results. Here a comparison is also made between the
exposure levels calculated from the vapor dispersion model for a
20-minute exposure at 10 ft from a spill (Table 25) and the 4-
hr rat LC50's. Under conditions set forth for the model, a
substantial margin of safety is indicated. c
Table 26. RAT INHALATION STUDIES: 4-HOUR LC.sn CONCENTRATIONS
Experimental Conditions:
10 Rats per Test
70 Liter Chamber
Air Flow Changes Air Every 3-6 Minutes
Test Fluid Metered into Vaporizer and Thence Carried
into Chamber by Air Stream
'-1-Hour Exposure Time per Test
Concentrations Increased in Subsequent Tests Until
Sufficient Data to Calculate LC50
After Exposure Surviving Rats were Observed for 14 days
4-Hour LC5n Results
RC-1 (Fluorobenzenes) LC50 = 115 mg/1 or 16,000 ppm
RC-2 (2-MP-H20) LC50 = 14.3 mg/1 or 8,000 ppm
Comparison With Calculated Exposure Levels:
RC-1 RC-2
Average 20 min. Dosage @ 10 ft 390 ppm 254 ppm
4-Hour LC50 (For Rats) 16,000 8,000
63
-------
Subsequent to these results, additional mice inhalation work
sponsored by Imperial Smelting Corporation has .been reported
(ref. 13). The full report is included in Appendix G. The
principal results of 30-minute exposure tests were as follows:
30-Minute •
Fluid LCsn, ppm (vol)
Pentafluorobenzene 30,000
Hexafluorobenzene . 65,000
50:50 mixture of above 50,000
Considering that the exposures were 30 minutes rather than 4
hours, these data fall in line with the 4-hour RC-1 data.
Some appreciation for the inhalation hazards of these materials..
can be had by comparing their animal LC50's with concentrations
of more familiar materials which have resulted in human death.
Gasoline is reported (ref. 14) to have killed a man after 5
minutes exposure in a concentration estimated between 5000 and
16,000 ppm. Benzene is reported (ref. 15) to be lethal in 5 to
10 minutes at 19,000 to 20,000 ppm, and hazardous to life in.30
to 60 minutes at 7500 ppm. Both gasoline and benzene are handled
in very large quantities every day with only the rarest reports
of vapor intoxication.
Odor is an intangible but important factor in consideration of
toxic hazards. Both final candidates have characteristic odors.
RC-2 with its content of 2-methylpyridine is immediately detect-
able at very low concentrations by its strong, penetrating, and
extremely disagreeable odor. Its escape into the atmosphere can
be expected to strongly motivate people to move away from the
source. RC-1 has a much milder, sweeter, medicinal odor which
would likely produce a positive but less urgent motivation to move,
5.6.3 Range Finding Toxicity Tests
Considerably more information about oral toxicity of pyridine and
methylpyridines was available in the literature than could be
found relating to the effects of vapor inhalation. Pertinent
data, and literature sources are summarized in some detail in
Appendix H. Here it is sufficient to note that Veselov (ref.
16) reports the following acute oral toxicity levels for 2-
methylpyridine:
LD50 = 674 mg/kg in mice
790 mg/kg in rats
900 mg/kg in guinea pigs
-------
Closely bracketing these results are the toxlcity parameters
for rats given by Polilei (ref. 17) as:
LD
0
LD50
= 550 mg/kg
= 790 mg/kg
LD100 = 950 mg/kg
Using white mice, Lapik and Zimina (ref. 18) report the oral LD50
of pentafluorobenzene as 710 mg/kg and that of pentafluoropyridine
as 280 mg/kg. While pentafluorobenzene had a narcotic effect
with decreased motor activity, pentafluoropyridine caused con-
vulsions in the test animals.
Animal acute toxicity tests carried out in the course of this
investigation included rat feeding and rabbit skin and eye con-
tact experiments, as well as the rat vapor inhalation experiments
covered above. Several fluids of interest other than the final
candidates RC-1 and RC-2 were included in this test series.
In determining rat oral LD50 values, the test material was fed
by stomach tube to mixed sex albino rats. The acute rabbit skin
absorption test involved applying the test material in increasing
doses to the closely clipped skin of white rabbits. The less
severe rabbit skin irritation test was carried out by adding a
single dose of test fluid to a one square inch area of skin and
awarding an irritation score after a 24-hr exposure period. Eye
irritation in rabbits was determine by adding 0.1 milliliter of
sample to the conjunctival sac of the right eye of the test
animals; from this an eye irritation score was determined over a
7-day exposure period. Results of these tests are summarized in
Table 27- Complete data are given in Appendix I.
Table 27- ACUTE TOXICITY TEST RESULTS
Fluid
2-Methylpyridine/H20 (RC-2)
3-Methylpyridine/H20
4-Methylpyridine/H20
4-Methylpyridine
(undiluted)
© CRf-i •>
CgFg/CgHPs^
C6F6/C6F5N
®
Mol %
40/60
40/60
40/60
100
40/60
40/60
LDsn (mg/kg)
Rat Oral
810
710
700
700
-^12,000
310
Rabbit
Dermal
200-316
126-200
200-316
126-200
>8000
•^5000
Motes: ® M.MXimum score = 8
© :^.uximum score = 110
© Hexafluorobenzene/pentafl uorobenrsene
® Kexafluorobenaene/pentafluoropyridine
Irritation Score
Rabbit Rabbit
Skin® Eye©
5-6
5
4.6
Corrosive
3
5-3
71
67
77
72
10
62
65
-------
5.6.4 Products of Combustion
A question often raised in considering working fluid toxicity is
whether exposure of the fluid to fire is likely to generate toxic
materials which then pose a greater threat to safety than the
fire itself. The appearance of .phosgene in the products of
oxidative pyrolysis of certain chlorine-bearing organic fluids
is an oft-cited example.
As in the case of the unburned working fluid, the most likely
route of acute intoxication by combustion products is by in-
halation. Most of the comments of Section 5-6.1 on vapor dis-
persion apply equally to these materials when allowance is made
for the fact that they are formed in the gaseous state. To
establish that a real hazard can exist, it is not sufficient to
show that hazardous substances are formed. It must also be
shown that dangerous concentrations of these substances can
reasonably be expected to be generated and persist for sufficient
time to cause injury to an incapacitated victim.
In interpreting the combustion experiments to follow, it must be
kept in mind that the products of combustion of the working
fluids were confined within a small chamber to simplify iden-
tification and measurement. Concentrations reported are, there-
fore, far in excess of those to be expected in an actual incident.
Hazards are to be assessed only after realistic air dilutions are
established.
5.6.4.1 Experimental Procedure - The final candidate fluids
were burned in an Amihco-NBS smoke density chamber (ref. 19) of
18 cu ft volume. A measured amount of the liquid was transferred
from a syringe into a stainless steel sample holder of special
design. The sample was moved into position so that the surface
of the liquid was directly impinged upon by the flames from a six-
outlet propane/air burner. The flow rates of propane and air were
9*4 cc/min and 6^7 cc/min, respectively.
The combustion gases were confined in the chamber whose internal
wall temperature was maintained at approximately 50°C. At the
completion of burning, the burners were extinguished, and
measured volumes of the contained gases were withdrawn into
evacuated vessels through a sampling port in the ceiling of the
cabinet.
A minimum of two experiments were conducted with each candidate
fluid. From one to two liters of gas were withdrawn for each
analysis.
66
-------
The temperature of the flames produced during combustion of the
fluids were measured with an optical pyrometer.
5.6.^.2 Analytical Methods - Carbon monoxide and total hydro-
carbon concentrations were measured with the Monsanto CO/CH^/C H
Analyzer, which is a gas chromatographic type instrument. x ^
Hydrogen cyanide, which could be generated during the burning
of the water/2-methyIpyridine mixture, was sampled by withdrawing
a known volume of the combustion gases through a tube packed with
Ascarite absorbent (asbestos impregnated with NaOH). The Ascarite
was dissolved in water. An accurate and sensitive fluorometric
method, adapted from procedure outlined by Hanker, J. S., et al.
(ref. 20), was used for cyanide analysis. Parallel fluorometric
measurements were conducted with standard cyanide solutions.
Hydrogen fluoride, produced by the burning of the penta-
and hexafluorobenzene mixture, was sampled by withdrawing measured
volumes of the combustion gases through a Teflon tube, packed with
Ascarite. The fluoride was quantitatively extracted and measured
as fluoride ion, using a specific fluoride ion electrode.
Nitrogen oxides (NOX), from the burning of the water/2-
methylpyridine mixture, were collected by withdrawing a measured
volume of the combustion gases into an evacuated vessel contain-
ing a dilute sulfuric acid/hydrogen peroxide absorbing and
oxidizing solution. The nitrogen oxides (NO + N02), except
nitrous oxide (N20), were measured colorimetrically using the•
phenyIdisulfonic acid (PDS) procedure (ref. 21).
The mass spectra of the combustion gases, other than water,
carbon dioxide, and the gaseous products referred to above, were
recorded with a Consolidated Engineering Co. Model 21-103C mass
spectrometer.
•S . 6 .'I . 3 • Results and Discussion - Experimental and analytical
results are given in Tables 28 through 30. Referring to Table
28, it is seen that combustion gases from the burning of RC-1
with propane-air are relatively high in CO and HF, both recognized
toxic substances. It appears that the presence of the working
fluid significantly inhibited the complete combustion of the
propane, as evidenced by the high CO and unburncd hydrocarbon
(C H ) concentrations.
* y
Combustion of RC-2 was, as might be expected, more complete.
Much lower concentrations of the toxic materials CO, NOX and HCN
were found. Also, the measured flame temperature was somewhat
higher.
Mass spectrometric analyses of additional products of combustion
of the tv.'o fluids, shown in Tables 29 and 30,' fail to reveal any
comparably hazardous concentrations.
67
-------
Table 28. COMBUSTION EMISSION EXPERIMENTS
CO
Fluid
RC-1
RC-2
Average
Sample
Mass
(g)
3.51±0.30
2.45±0.26
Burning
Time
(min)
1.5-2.0
3.0
Flame
Temp.
During
Burning
(°C )
1100-1145
1185-1250
CO
(ppm)
612±66
30±1
cxHy
(ppm)
935±72
49±1?
HF
(ppm)
746*80
—
NOX
(ppm)
-
79±6
HCN
(ppm)
-
<0.5
-------
Table 29.. MASS 3PECTRQMETRIC ANALYSIS OF RC-1 COMBUSTION PRODUCTS*
Concentration
Component (ppm by vol.)
Hcxafluorobenzene 57
Pentafluorobenzene 68
Tetrafluorobenzene 46
Silicon tetrafluoride 18
Perfluoromethane 42
Perfluoroethane . 18
Carbonyl fluoride 1.1
*This table does not include data for H20,
C0,2, .CO, MOX, and HP content of. the chamber
atmosphere.
69
-------
Table 30. MASS SPECTROMETRIC ANALYSIS OF RC-2 COMBUSTION PRODUCTS*
Concentration
Component (ppm by vol.)
Methane 2.4
Ethylene 3
Ethane . 12
Propane 20
Propylene 5
*Thls table does not Include data for H20,
C02, CO and NOX content of the chamber
atmosphere.
70
-------
5 . 7 FLAMMABILITY
Flash and fire points, and autogenous ignition temperatures
measured using the final candidates and related compositions are
listed in Table 31. Comparable literature values for a number of
familiar reference fluids are also listed for comparison. The
data show RC-1 to be non-flammable, and RC-2 to be flammable, in
terms of these specific tests. RC-2 has flash and fire points
comparable to kerosene despite a higher vapor pressure. The
effect of the water content of this fluid is to raise its flash
and fire points about ^5°F above straight 2-methylpyridine.
Flarnmabilities of the ingredients of RC-2 are presented by Pollard
(re.F, 22). He shows that all fluorobenzenes containing four or
more fluorine atoms per molecule are non-flammable in air under
any conditions.
Results of a hot compartment ignition test are listed in Table 32
for the final candidates in comparison with benzene and Fluorinol
85 (a product of Halocarbons Incorporated). In this test, a
pressurized spray of the fluid is admitted into a heated, air-
filled chamber equipped with a continuous high-voltage spark gap.
The intensity of ignition is gaged by the reading of the shock
meter, which is simply an accelerometer attached to a loose sheet
metal lid covering the compartment. A bare thermocouple within
the chamber gives an indication of the amount of heat generated
upon ignition.
The results show RC-1 to be practically inert compared to RC-2
and Fluorinol 85 (which are quite comparable in both energy
release and shock intensity) and far below benzene. RC-2 shows
temperature rises somewhat greater than benzene but shock inten-
sities about half as high. The shock meter readings are measures
of the sharpness of combustion and are indicative of damage
potential.
Test methods are covered in detail in Section 6-7-
5.8 LUBRICATION AND LUBRICANTS
Lubricant requirements of a Rankine cycle power plant employing a
reciprocating expander are substantially different and much more
stringent than the demands of a turbine expander system. In the
latter the main bearings can be mechanically isolated from con-
tact with the working fluid and they operate at relatively low
temperatures. Thus conventional lubricants can be used. Demands
of the reciprocating expander are another matter. Here it is
almost impossible to mechanically isolate the lubricant from the
71
-------
Table 31. FLASH AND FIRE POINTS AND AIT'S OF FINAL
CANDIDATES RC-1 & RC-2 AND REFERENCE FLUIDS
Fluid
Final Candidate RC-1
Final Candidate RC-2
Additional mixtures of
100
75
40
15
Reference Fluids
Gasoline
Benzene
Toluene
Kerosene
Ethyl alcohol
Methyl alcohol
Ethylene glycol
Flash
Point
(°F)
-1 None
-2 130
s of
pyridine
ridine
90
90
120
135
-50
12
40
55
55
1 232
Fire
Point
(°F)
None
145
100
105
145
175
100-160
AIT
(ASTM
D-2155)
(°F)
None*
1060
928
900
685
727
748
*Up to 1200°C., the apparatus limit
72
-------
Table 32. HOT COMPARTMENT IGNITION TEST RESULTS
Test conditions:
Fluid charge
Compartment temp.
Compartment environment
Spray time
30 cc
100°C
Air
sec
Fire Results
'Working Fluid
Benzene
RC-1
RC-2
Fluorinol 85
1000 psi Spray
A Temp Shock
(°C) • Meter
58
-19
72
73
0
8
9*
500 .psi Spray
A Temp Shock
(°C) Meter
80
-13
103
11
1
6*
100 psi Spray
A TempShock
(° C )• Meter
. 138 3
9 1
.159 2
80* • 0*
^Delayed Ignition or 'Fire
73
-------
working fluid. . As a consequence, some minor amount of lubricant
is inevitably carried through the vaporizer. To avoid fouling
the vaporizer and degrading the lubricant, it is necessary that
the latter have thermal stability comparable to .the working fluid.
Vaporizer temperatures in the 710-720°P range required in the
present study exceed the thermal stability of conventional mineral
oil lubricants. This means that a highly stable synthetic lub-
ricant must be employed. Since 2-methylpyridine/water of 0.35
mol fraction 2-methylpyridine, or similar mixtures with the other
methylpyridines, appear to be optimum for a reciprocating Rankine
system, the lubricant must further be both thermally and chemically
stable in such a water-amine environment up to ^720°F. Finally, it
is desirable that the lubricant and working fluid have low mutual
solubility; 'that is, the oil should be insoluble in the working
fluid and vise versa so as to minimize both dilution of the crank-
case oil by working fluid and carryover of lubricant to the
vaporizer. Given these restrictions, lubricant choices are
extremely limited.
Although mutual insolubility of the working fluid and lubricant is
desirable, system design can reduce working fluid contamination by
tight seal design in the pump and expander with lubricant separa-
tion before working fluid condensation. Since the bulk lubricant
operational temperature can be higher than the working fluid boil-
ing point under anticipated crankcase pressures, the working fluid
would not dilute the lubricant significantly during vehicle opera-
tion. Only on shutdown does lubricant dilution seem possible and
this could be corrected by appropriate one-way valving between the
lubrication system and the Rankine cycle system.
5.8.1 Lubricant Stability
A test program was initiated, a) to measure relative solubilities
of methylpyridine-water working fluids in various potential lub-
ricants, and b) to determine thermal stability (compatibility) of
various combinations. The ampoule test described in Section 6.3
was used for the thermocheinical stability work, and a simple pro-
cedure was devised to measure relative solubility of the methyl-
pyridine-water fluids in the lubricant.
Procedure: Four grams of the test lubricant in a 25 x 100
mm test tube were stirred vigorously for 10 min. with 12 g
of the methylpyridine-water working fluid. The mixture was
then allowed to stand undisturbed for 10-15 min. to allow
lubricant and working fluid layers to separate. After
settling, about 1 g of the working fluid layer was withdrawn
and carefully weighed. The weighed sample was then dis-
solved in about 10 cc of glacial acetic acid and titrated
with 0.1 N perchloric acid (in acetic acid solution) using
-------
crystal violet indicator. The amount of amine dissolved
in the lubricant layer could then be calculated from the
amount of perchloric acid consumed.
Results of lubricant-working fluid solubility studies are
summarized in Table 33- Here it can be seen that only the per-
fluorinated oils (experiments 55 6 and 7) have the requisite in-
solubility properties, with hydrocarbon mineral oils (11-15) be-
ing next in solubility. Unfortunately both these oils are un-
stable at 720°F as are all of the other candidates tested except
for the simple aromatic hydrocarbons. This is essentially what
might be predicted on the basis of known thermal stability be-
havior (see section 6.2.1). Only the biphenyl-terphenyl eutectic
(whose composition is roughly 21% biphenyl, ^7% o_-terphenyl, and
32% m-terphenyl) seemed sufficiently available for lubrication
studies .
5.8.2 Lubrication Wear Tests
After the lubricant stability tests indicated that the only ?20°F
stable fluid was the biphenyl-terphenyl eutectic, and considering
the lubricant volumes need for testing, the lubrication tests were
performed in two stages:
1. Screen six composite materials as sealing ring materials
for wear resistance at 600°F on cast iron in the biphenyl-
terphenyl eutectic.
2. .Select the best composite material and run with an alloy
steel, M50, on cast iron at 600°F; and at room tempera-
ture with mixtures of the biphenyl-terphenyl eutectic
and RC-1 and RC-2.
The apparatus used in these test was a Monsanto Company developed
wear test referred to as the Rub Block Test and as described in
detail in Appendix J. Essentially, the apparatus consists of two
flat blocks (ring material) loaded diametrally against a rotating
ring (cylinder block material). This assembly is immersed in the
test lubricant which in turn, is contained in a hermetically
sealed chamber provided with fluid temperature and pressure con-
trols. Examination of the wear scars developed on the rubbing
blocks quantitatively ranks the lubricants and wearing materials.
The texture of the scar can distinguish the types of wear taking
place.
The results from composite material testing are presented in Table
3^ with photographs of the ring and blocks in Figures 17 through
19. A correlation between scar width and carbon porosity is demon-
strated in Figure 20.
75
-------
Table 33. EQUILIBRIUM SOLUBILITY OF METHYLPYRIDINE WORKING FLUIDS
IN VARIOUS PROSPECTIVE LUBRICANTS
Potential Lubricant
1) biphenyl-terphenyl
eutectic
2) OS-124 5-ring polyphenyl
ether •• .
3) F-25.57,, 4-ring';-mixe;d-S- .
and o-polyptieriyl ether
4) tetramethyldi(m-p'henoxy-
phenyl)disiloxane [m-
000Si(CH3)2]20
5) perfluorokerosene •
6) Krytox AB perfluoro ether-
7) Krytox AD perfluoro ether
8) o/m-terphenyl-phenoxy-
biphenyl 1/2 mixture
9) 1-phenylnaphthalene
10) o/m-terphenyl +
blsphenoxybiphenyl
11) SUNISO 3GS hydrocarbon oil
12) TECO Oil #1, hydrocarbon
13) TECO Oil #2, hydrocarbon
11) TECO Oil #3, hydrocarbon
15) TECO Oil #4, hydrocarbon
16) DC 710 Silicone oil
Mole Fraction Amine
in Working Fluid
Used for
Equilibration
0.3 2-methylpyridine
U. 4 2-methylpyridine
0.. 3 2-methylpyridine
0.3' 2-methylpyridine
0.3 2-methylpyridine
0.4 2-methyipyridine
0.3 2-methylpyridine
;0.4 2-methylpyridine
0.4 2-methylpyridine-
0.4 2-methylpyridine
0.4. 2-methylpyridine
0.4 2-methylpyridine
0.4 '2-methylpyridine
0.4 .2-methylpyridine
0.4 3-methylpyridine
0.4 4-methylpyridine
0.4 3-methylpyridine
0.4 3-methylpyridine
0.4 3-methylpyridine
0.4 3-methylpyridine
0.4 3-methylpyridine
0.4 3-methylpyridine
Wt Percent Amine
in Lubricant Phase . . • . ,
after Equilibration Results of• Thermal Stability
at 77:°F (25°C) . . Tests-at 720°F
24
32
17
24
:I ]
:'. 0
18,0 ~
25.2 ._
17
23
0.
0.
0.
32
36
29
32
26
8
7
7
7
5
14
:' ]
24 ,~
1
02
.9
.1
\l
.6 ~
.1
.1
.2
.6
.7
[
\-
\-
-
-
-
-
_
Stable 168 hr- g 720°P, but
o'rily.single, p^hase on .cooling
Extensive decomposition
168 hr @ .720°?
Extensive decomposition
-168 hr @ 72p°F
Extensive decomposition
168 hr-v@ 720°F
Complete decomposition
168 hr § 720°F
Partial decomposition
168 hr g 720°F
Not tested - should be stable
Partial decomposition
168 hr g 720°F
Partial decomposition
90 hr g 72'0°F
Complete decomposition
90 hr g 720°F
-------
Table 34. RUB BLOCK WEAR TESTS OF COMPOSITE MATERIALS
Test Conditions:
Ring diameter = 1.5 in.
Block thickness = 0.25 in.
Ring speed = 3200 rpm
Load between blocks and ring = 44.2 Ibf
Test duration = 50 min.
Lubricant = Biphenyl-terphenyl eutectic
Fluid temp. = 600°F.
Ring material = cast iron*
Test Scar Width
No. Block Material (in.)
1 Pure carbon P-19 0.112
2 Pure carbon P-5-NR .0.183
3 U. S. Graphite 14SC . 0.094
4 U.' S. Graphite 110 0.098
5 Ford I/G (40/60), 16% porosity 0.477
6 Ford I/G (40/60), 12% porosity 0.387
*Ford Cast Iron Material ESE-MIA117-B
Hardness Re 50, case depth = 0.025 in.
Surface finish direction of rotation = 8-12u in RMS
77
-------
17a. P-19 Blocks
Test 1
17b. P-5-NR Blocks 17c. Cast Iron Ring
Test 2
Figure 17. Rub Block Wear Specimens (3.2x)
78
-------
l8a. USC Blocks
I8c. 110 Blocks
Test 3
l8b. Cast Iron Ring
l8d. Cast Iron Ring
Test
Figure 18. Rub Block Wear Specimens (3.2x)
79
-------
19a. I/G (W60) 16? Porosity Blocks
Test 5
19b. Cast Iron Ring
19c. I/G (1*0/60) 12? Porosity Blocks I9d. Cast Iron Ring
Test 6
Figure 19. Rub Block Wear Specimens (3.2x)
80
-------
S
O)
CD
ro
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
Test 1 Pure Carbon P-19
Test 2 Pure Carbon P-5-NR
Test 3 US Graphite 14SC
Test 4 US Graphite 110;3
Test 5 Ford I/G .(40/60) 10% Porosity
Test 6 Ford I/G :(4Q/60) 12% Pprpsity
10 15
Porosity, Vol %
2
9
25
Figure 20. Scar Widths of Carbon Blocks .vers.us .Por.o.si-ty
81
-------
The U.S. Graphite 14SC sintered graphite material was chosen as
the best compp.sit.e ...t.o.~,r..un with the lubricant working fluid mixtures
The others; -evaluated had larger scar widths but should not be con-
sidered inferior to USG 14SC since the 44.2 Ibf normal load is
extremely high for non-conforming contacts. Real piston rings
would operate at much lower ..contact stress levels, but to expedite
a short term wear test the higher load was needed. Also worthy
of note is the low ring wear on the highly worn Ford I/G com-
posites. This could translate 'to'low cylinder wear in a Rankine
reciprocating expander. /•;• . . , .• . . ;
.4 • .-'»::- '-. >..••«• , •
Since the bipheny-lL'terp'henyl* eiitectic has a relatively high pour
point of 50°F, it/;> was;/.Hoped that,, a' beneficial dilution with RC-2 •
would produce a lower flow point. Equal parts of the lubricant
and working fluid were found to have a flow point of -22°F (-30°C),
and the mixture is in two phases at room temperature.
For mixtures of RC-1 and the biphenyl-terphenyl eutectic, a 5%
dilution of the eutectic lubricant was chosen due to the limited
availability of RC-1. The ' test conditions and wear results are
presented in Table 35. Corresponding block and ring photographs
appear in Figures 21 through 25.
The test results indicate the USG 14SC material to be superior to
M50 tool steel in block and ring wear. The wear of the RC-2-
lubricant mixture at room temperature is equivalent to the lubri-
cant wear at 600°F. The lubricant wear is less at ro'om temperature
on the USG 14SC material than at 600°F, while the opposite seems
true with M50. The 5% dilution of the lubricant with RC-1 has
little effect on wear on the USG 14SC and M50 materials at room
temperature. Besides showing high wear, the M50 tool steel wear
characteristics indicate welding and scoring in the wear areas.
It is concluded that!the combination of either candidate working
f luid/eutectic/lubri),cant/cast iron/graphite is a workable one
insofar as lubrication against sliding wear is concerned.
5.8.3 Special Lubricant Synthesis
A limited synthesis effort was undertaken in an attempt to find
new lubricants stable at 720°F in the presence of the methyl-
pyridine-water fluids, yet relatively poor solvents for the work-
ing fluid. Briefly the situation at the onset of this work could
be summarized as follows. Only a few polyaromatic liquid oils are
stable above .71-2°R^-in. the presence of :pyridine-water working
fluids. All of these stable materials are good solvents for
pyridine and the methylpyridi^ne'S,. and -hence are extensively
diluted by the working fluid and, in fact, barely form two phases
at room temperature.
82
-------
oo
OJ
Table 35- RUB BLOCK WEAR TESTS OF LUBRICANT AND MIXTURES
OF RC-1 and RC-2 WITH LUBRICANT
Test
Test
No.
7
8
9
10
11
12
13
14
15
16
conditions :
Ring diameter = 1.5 in.
Block thickness = 0.25
Ring speed = 3200 rpm
Fluid
Biphenyl-terphenyl
Eutectic (lubricant)
Same
Same
Same
50% wt. Lubricant
50% wt. Candidate RC-2
Same
95% wt . Lubricant
5% wt. Candidate RC-1
Same
Suniso 3G Oil
Same
in.
Fluid
Temp
(°F)'
600
600
RT
RT
RT
RT
RT
RT
RT
RT
Test duration = 60 min.
Load between blocks and ri:
Ring material = Cast iron1
Block Material
US Graphite 14SC
M50 Tool Steel2
M50 Tool Steel
US Graphite 14SC
US Graphite 14SC
M50 Tool Steel
US Graphite 14SC
M50 Tool Steel
US Graphite 14SC
M50 Tool Steel
Scar
Width
(in.).
0.083
0.225
0.323
0.058
0.081
C.1123
0.059
0.275
0.067
0.038
1 Ford cast iron - Material ESE-MIA117-B
2 Hardness = 60 RC, surface finish = 8y-in. RMS
3 Friction torque exceeds limit of torque cell, reduce load to 29.5
.2 lb
-------
2la. l^SC Blocks 21b. Cast Iron Ring
Test 7
21c. M50 Blocks 2 Id. Cast Iron Ring
Test 8
Figure 21. Rub Block Wear Specimens (3.2x)
-------
22a. M50 Blocks
22b. Cast Iron Ring
Test 9
22c. USC Blocks 22d. Cast Iron Ring
Test 10
Figure 22. Rub Block Wear Specimens (3.2x)
85
-------
23a. 14SC Blocks 23b. Cast Iron Ring:
Test 11
23c. M50 Blocks 23d. Cast Iron Ring
Test 12
Figure 23. Rub Block Wear Specimens (3-2x)
86
-------
. USC Blocks 24b. Cast Iron Rine
Test 13
24c. M50 Blocks 24d. Cast Iron Ring
Test 14
Figure 24. Rub Block Wear Specimens (3-2x)
87
-------
25a. 14SC Blocks
Test 15
25b. Cast Iron Ring
25c. M50 Blocks
25d. Cast Iron Ring
Test 16
Figure 25. Rub Block Wear Specimens (3.2x)
-------
Oils which are primarily aliphatic in character (i.e. mineral oils
similar to Suniso 3GS) appear to be sufficiently insoluble in the
methylpyridine water fluids, but these materials are not thermally
stable above about 660°F. Similarly, polyfluorinated kerosene or
KRYTOX ethers are quite insoluble but are chemically unstable in
the presence of the working fluids. One possible way to resolve
this dilemma would be in the synthesis of a stable, largely ali-
phatic molecule. This approach was followed.
Advantage was taken of the fact that methyl-silicon bonds are
abnormally stable. A trimethylsilyl group (CH3)3Si- is relatively
large sterically and somewhat more thermally stable than a non-
silicon-containing hydrocarbon group of comparable size. By
attaching several trimethylsilyl groups to the benzene nucleus, it
was hoped that a stable, low-solubility oil might be obtained.
The method of synthesis was that employed by Chaffer and Beck
(ref. 23) for the preparation of sym-tris (trimethylsilyl) benzene
+ 3 Mg + 3 (CH3)3SiCl
+ 3 MgClBr
Si(CH3)3
Two oils, l,3-bis(trimethylsilyl)benzene and 1 , 3, 5-tris (trimethyl-
silyl )benzene, were prepared. Equilibration with methylpyridine-
water fluids showed that the trisilyl-substituted benzene was,
indeed, a relatively poor solvent for 2-methylpyridine-water .
In addition, ampoule tests indicated that both of the silylbenzenes
were about as thermally stable as the methylpyridine-water work-
ing fluids, and were not chemically attacked by the latter. Thus
a poly trimethylsilyl aromatic of suitable physical properties
cou.ld conceivably be an acceptable immiscible lubricant for use
with 2-methylpyridine-water. Difficulties arose, however, upon
investigation of the low temperature properties of the two fluids
when it was found that the trisilylbenzene crystallized at +8l°F.
Data on both 1 , 3-bis (trimethylsilyl )benzene and 1 , 3 , 5-tris ( tri-
methylsilyl )benzene are summarized in Table 36. Here it can be
seen that the melting points of the possible silane candidates
are discouragingly high and even the use of mixed silanes (example
#3) does not provide a -20°F flow point with the 2-methylpyridine
working fluid.
In an effort to improve the situation, two additional trimethyl-
silyl-substituted benzenes, shown in Table 36 as examples ^ and
5, were prepared by procedures similar to the first two. These
89
-------
Table '36. SOLUBILITIES AND LOW TEMPERATURE
- BEHAVIOR OF ARYL SILANE LUBRICANTS
Expt.
No.
1
2
3
4
5
6
7
Silane
Si(CH3)3
^SIWM.'
Si(CH3)3
\3 / 3 ^ J- '^^sx'^ Q -t / f'tJ N
o JL ^ L^ITS / 3
50/50 mix of #1 & #2
S1(CH3)3
Si(CH3)3
Si(CH3)3
(CH3)3Si -^S
Si(CH3)3
50/50 mix of #2 & #4
Si(C2H5)3
/ /-i TT \ o 4 ~^*°°**^^*' o ^ / r> IT \
^02ri5J3ol ^^ oH02Hs^3
f /*i TJ \ o •? ~^+*^*' a 4 f /"* n ^ "™^^~" ij
[U2n5^3ol ol(U2ns)2 T
1 ^
CH3 i
Si(C2H5)3
^^ Si(C2H5)3
mp
•Pure"
+ 12
+ 81
—
+109
+333
+ 5
-- 76
After Equilibration
With RC-2
Flow
Point
--10
-+30
-10
—
--
-13
^-103
Solubility of
2-MP in Lube
(Wt %)
20
14
17
--
•
14.2
11.3
90
-------
compounds unfortunately had still higher melting points and thus
appear to be useless for the intended application, although a
mixture of the isomeric tris trimethylsilylbenzehes 2 and 4
(example 6 of Table 36) offered some encouragement.
In an effort to work around the crystallinity impasse the complex
trietbylsilylbenzene mixture shown in example' 7 of Table 36 was
prepared from 1, 3»5-tribromobenzene and triethylchlorosilane.
It was expected that the use of triethylsilyl groups in place of
trimethylsilyl would drastically lower the melting points of the
polysilylated benzenes and, indeed, this proved to be the case.
Unfortunately the presence of the triethylsilyl groups, 4.. e.
(C2H5)3Si-, also appreciably lowered the thermal stability of the
oil, so that ampoule tests at 720°F revealed extensive decompo-
sition of the lubricant after ^8 hours of exposure. No further
attempts along these lines were made to define ah immiscible
lubricant.
Prospects for finding an immiscible lubricant for RC-2 are not
considered bright. Lubricants possessing the necessary in-
solubility are mostly of chemical types which are not stable to ^
720°F. Best prospects seem to be in extending the silane work
reported here, but ultimate costs of such materials would be high;
5.9 MANUFACTURE AND ECONOMICS
5.9.1 Final Candidate RC-1
Polyfluorobenzenes are currently custom-produced only in small
amounts at high cost by the direct fluorination of benzene with
cobalt trifluoride (ref. 2*O . While the process is well known
and convenient for small scale production, it is too chemically
inefficient to be seriously considered at the 50 million pound
annual production rate needed to support the manufacture of a
million Rankine automobiles a year. Before a reasonable estimate
of thf= ultimate owner's cost for the RC-1 fluid can be made, it
is, therefore, necessary to define a "most probable" process,
based on available technology. The process of choice is -the so-
called halogen exchange.
5-9-1-1 Halogen Exchange Process - This process basically con-
verts hexachlorobenzene into a series of chlorofluorobenzenes by
a fluorine-for-chlorine exchange between hexachlorobenzene and a
fluoride salt, as in the reaction.:
C6C16 + nKF r> C6Cl6_nFn + nKCl
In general, the reaction products are found to contain all
possible combinations of chlorine and fluorine on the fully
halogenated benzene ring. In other words, compounds represented
by all values of n from 1 to 6 are simultaneously formed in the
reaction.
91
-------
The major ingredients.of RC-1 are hexafluorobenzene (n=6) and
pentafluorobenzene." The latter is not a direct.product of the
exchange reaction;,, but' it is readily obtained from the n=5 product
by hydrogenation: ',••.."
•'.. C6C1F5' + H2 -> C6HF5 + HC1
This reaction is quite selective in that the chlorine is
replaced, but not the fluorine atoms.
Other products of.the original exchange reaction, those with
n=l to 4, can be upgraded to a higher degree of fluorination
(higher n)' by,simply being re-exposed to the exchange conditions
of the initial reaction.
These reactions are..the' bases for the RC-1 plant conceived as
depicted in the. flow sheet of Figure..26. Recent patent
literature (ref. 25-28) supplied much of the detail, but many
assumptions were necessary to fill in gaps.
Referring to Figure 26, liquid hexachlorobenzene is pumped into
the tubular substitution'reactor at. a. pressure of about 560
psig. In the; flue-gas-heated reactor, it travels through tubes
packed with solid potassium, fluoride. The halogen exchange
reaction occurs at the solid-liquid interface at a temperature
near 1000°F. On leaving the reactor, the^product stream is
depressurized and cooled to about *J50°F prior to entering a
distillation column.
In the fractional distillation column, the desired products
(n=5,6) are separated from low-boiling by-products and from
the higher boiling chlorofluorobenzenes (n=0 to *O . They are
then lead to a hydrogenation reactor which selectively converts
the chloropentafluorobenzene into pentafluorobenzene, leaving
the hexafluorobenzene unaffected. Hydrogen is admitted to the
reactor at a pressure of about 100-psig. The reaction takes
place at a controlled liquid temperature of 300°F. Hydrogen
chloride resulting from the hydrogenation reaction is recovered
from the hydrogen atmosphere and sold or used elsewhere.
The liquid product from the hydrogenation step is the final RC-1
fluid. Ingredient ratio is controlled by the reaction conditions
maintained in the halogen exchange reactor. Too high a ratio of
C6F6 to C6HF5 can be corrected by reducing the dwell time or by ]
altering other- reaction conditions. . .
The high-boiling bottom fraction from the distillation, which
contains primarily fluorine-lean compounds of n<5, is returned
as recycle to the exchange reactor. These materials are thus
re-exchanged to higher levels of fluorination, eventually to the
92
-------
Cooling
Water
m
© CI2
Furnace f1
•^—
FiueGasOut
Furnace
• B
Flue Gas In
mnr
(T)
UJ
Substitution
Reactor
Regeneration
Cooler
To Vent
H2 from Storage
HCI to Storage
Recycle Reaction'Mass.
Storage Cooler1
Flow'Streams-l-lndicated by.-'Cir'cles')
1. Hexachlorobenzene^CfiGle"- Raw:Material)'
2. Partially FiuorinatefrRetycle. '
3.-4; Fluorinated=Crude;Product •
5. Non-Condensable'Yield'Loss
•6. Distillation:'Column:Reflux..
•7:--H'exafluoro-Pent^fluo'roehloroben'zene;IVlixture
8. Hexafluoro^Pentafiuorbbenzerie'ProductiMixture
9. Fluoririe'Gas'lRaw Material);
10: Nitrogen'Gas'lRawJMaterial)1
I'l- Recycle-F2 - N2 Mixture'
12: By-Product'Liquid;Chlorine
13. By-Product HCI to'-Storage
Tp.Rroducti
Stor(age;-
.Hydrpgenation:
Reaeto'rs.i(-2)
Figure. 26. ,.Manufacture, of Highly Fluorinated, B.e-n-z.enes
(.Cs'Fg/C'gF'sH") 'KE'Regieae-ria-feiSDn in 'Place..
-------
desired n=5,, 6 level. An occasional purge of high molecular weight
polymeric residues (not shown in flow sheet) would be necessary
to prevent their over-accumulation in the recycle loop.
The gradual conversion of:the KF into KC1 accompanying the
exchange reaction slows the reaction to the extent that it must
be periodically interrupted- to allow the salt to be regenerated
to the fluoride form. In -the conceptual process, a make time of
12 hours was assumed. It was also assumed that at the end of the
"make" portion of the cycle, half the original KP is converted
to KC1. At this point, fe:ed to the reactor is stopped and the
reactor is vented to the distillation c.qlumn. After cooling the
reactor contents to the range 500-600°F by circulating nitrogen
(N2), fluorine (F2) is'admitted to the gas stream and a halogen
exchange takes place which converts KC1 back to KF with the
generation of C12. C12 may be condensed from the mixed F2-C12-N2
gas stream for sale as a by-product.
- ' . i
Following regeneration of the KF, the reactor is reheated to
reaction temperature (1000°F), raw material and recycle streams
are restarted, and the reaction proceeds again for another 12
hour period. .
Major assumptions about the process are listed below:
A. Overall Material Balance : "
;'-.-•*
1) Ons'tream, time .would ..be 750,0 hr/yr
2) A 40/60 molar split between C6F6/C6F5H is the optimum
fluid of this type . ' ;
3) Yields: : F2 '. 90% \
, . C6C16 95%
4) Plant design capacity - 50,000,000 Ib/yr
B. Substitution Reactors
1) Retention time - 1/2 hr
NOTE: This sets the requirements of this reaction
step at one reactor with 53 8 in. dia., 20-ft long
tubes.
2) A total downtime for regeneration will be ^ hr
3) The reactivity of the fluoride salt will reach an
unacceptable level after 12 hr running time.
-------
C. Distillation Column
1) Atmospheric pressure
2) 10 theoretical trays are necessary
3) Minimum reflux ratio to achieve the split required
is 1.7/1.
!-\) These parameters set specifications of the column at
3 ft dia. and 60 ft tall with 20 sieve trays.
D. Hydrogenation Reactor
1) Retention time will be 2 hr.
2) Total turnaround time will be 3 hr.
3) The size of the reaction system is then two 1500-gal.
reactors.
5-9-1-2 Halogen Exchange Economics - Capital and operating
costs were estimated for the hypothetical 50 million pound per
year RC-1 plant described above. Because the technology of the
process is very sketchy, the accuracies of the resulting estimates
are not considered high enough to warrant a final decision on
economic suitability of the fluid. The data are considered
sufficiently accurate to support additional process research and
development.
The estimate of capital requirements was made using equipment
pricing factors, installation costs, and overhead factors from
published correlations (ref. 29, 30) or from Monsanto's pro-
prietary experience records, whichever was believed more reliable.
All cost data were referred to 1972 using the Marshall and Stevens
'••iuipment Cost Index. The "battery limits" capital estimate was
:':•;, 000,000. When support capital and working capital are added,
the total capital investment is $37,000,000. This estimate is
•nsidered to have an accuracy of ±25-
in addition, and using a Monsanto-proprietary economic evaluation
program (E0058), the cost of operating such a plant at capacity
•as estimated. Costs normally sensitive to the choice of site
for plant construction were assumed at levels representing an
"economically favorable U.S. location", without specifying where
it might be. Other assumptions were made with the intent of
producing a "moderately optimistic" analysis to define the likely
lower bound on ultimate costs. Both the assumptions and resulting
analysis are given in Tables 37 and 38.
The estimated bulk manufacturing cost of $1.00 per pound is con-
sidered the lowest probable cost after several years operating
experience in a plant operating at full capacity.
95
-------
'Table 37. MANUFACTURING COST ESTIMATE - RC-1
- SO MILLION POUNDS PER YEAR
YIFl.ns: ON ELUOfiTMF = 90*. ON HE X ArH|_rROBFNZENE = 95*
WY-PP-OnuCT CWEniT OF $0.015 PEP LB CL? AND *0.~030 PER LR MCI
M; A N IJ F A C T U R I N G COST ESTIMATE
S600 DESIGN RATE 50000000 LR/YP, 7SOO OPpRATlNG MR/YP
400 PRODUCTION 50000000 L«/Yk lOO.O * nF DESIGN
o 0 0 0
HI.no
UNTTS/YR
«54Q9936
25. 0*
?7500
900
C6CL*
H^ 400000
NP 4796995?
CI.P -545Q9936
MCI. -6249999
TOTAL MATERIALS
EXPFNSE
41600
SyPFiPV/ISlON H320
PAYROLL CHGS
FLECTPICTTY
PWdCFSS WATFP
F-'UFL GAS
RFSAiws M^F fl.oa;
REPAIRS HLOG .4.0*
I AHORATOPY 1700
CLHT-I ^ LAUNO 2.or<
FACT SUPPLIES
TOT4L HIN EXP
T EXPENSE
MR.E
OEPP ^L.HG
I- IF - NON-CONT
FIF-CONTDOLL«RLF
TOTAL 1 vF> EXP
TOTA|. CONV. COST
10,0*
4. 0*
COST
TOTAL FOR COST OF (-OOOS
UNIT OR
* HAS IS
LR
LM
LH
LR
La
LR
«/UNIT
0. 1«0
1 .000
1.000
0.020
0.015
0.0 TO
PRACTICE
IB/LR
1.710
0.650
O.()0«
0.959
-1.09?
-0.1 25
«:M/Yr5
15390
3?500
400
950
-819
-187
4«24?
•B/LB
0.30^
0.650
0.003
0.019
-0.016
-0.004
0.965
MAN MRS 5.50
MAN MRS 6.50
LABOR *SLIC'V
CKWH 1.5QO
MGAL 0.200
0.400
DG
LAHOR+SUPv
o.ooo
RI.PG
5390
?500
400
950
-819
-187
«24?
229
54
71
41
0
24
44fl
16
17
6
30
936
560
16
90
1P3
849
| 7*4
10?7
1027
0.30^
0.650
0.003
0.019
-0.016
-0.004
0.965
0.005
0.001
0.001
0.001
o.ooo
0.000
0.009
0.000
0.000
0.000.
0.001
0.019
0.011
0.000
0.002
0.004
0.017
0.036
1 .001
1 .001
96
-------
Table 38. RETURN ON INVESTMENT - RC-1
HFXA-/DFNTAF|JJOh"^S F OH COST OF GOODS SOi p
690P.1
PROFIT
'; MAT FXPFN^^C, ;Q, m.Oo * OF _
1; CORP IwVFST fHG "f S.OO ^ OF MF T
OF
TNCO'^f. ^FFORE TAX
LFSS JNCOMF TAX a
MFT
TAX
19fl96
699?
1S90
1 131S
5657
5657
NFW FTXFH CAPITAL
UTILITY ANO
O MAT CAPITAL
CAPITAL
TOTAL FT.XFO CAPITAL
NG CAPITAL w 17. On * OF SALFS
T'MVFSTMENT
6000
945
1 1R40
?587?
ROT
ANNUAL ^FTURN ON INVFST^FNT AFTFP TAXF<; 15.oo
97
-------
Of the. various elements contributing to this cost, the raw
materials at $0.97 per pound dominate. Obviously, any appreciable
reduction, in the cost can only be achieved by significant improve-
ment in yields or significant reductions in raw material costs.
Yields on both hexachlorobenzene and fluorine were assumed above
90%, ' so there is no_t- much improvement possible in these respects.
The assumed fluorine cost of $1.00 per pound (ref. 3D was verified
in consultation with Allied Chemical Corporation and is based on
a price-volume extrapolation to quantities involved in an auto-
motive market. This price is considerably below present prices,
by a factor of 'about four. It is unrealistic to expect a more
favorable price.
Fluorine is, however, available in a cheaper form than the
assumed elemental liquid (for example, in combination with other
elements in the minerals fluorspar and cryollite). Perhaps the
best hope for ultimately lower prices lies in the invention of a
way to, exploit a cheap inorganic fluorine compound without going
through the expensive elemental form.
Turning to the return on investment analysis of Table 38, the FOB
plant selling price is next estimated at $1.^0 per pound after
assuming a 15% return on gross investment and applying typical
factors for marketing, administration and technical expenses, and
for income taxes and debt service. The gross investment, in turn,
was calculated from the earlier estimate of new fixed capital by
the use of typical factors for allocated capital and working capital,
As with the bulk manufacturing cost, the bulk FOB price is con-
sidered a minimum after several years of full production. It
could legitimately be argued that a 15% return on investment is
insufficient incentive to interest most potential manufacturers
in a venture as technically and economically speculative as the
one at hand. A substantial and successful technical and com-
mercial development, underwritten by funds not included in this
economic assessment, is considered a likely prerequisite.
The estimated $1.40 per pound FOB bulk sales price may be trans-
lated into a 5-year automobile owner's cost of $112 by entering
Equation 21 (section 6.1.3) with the defendable assumption that
no fluid replacement would be needed in the 5-year period. A
basic assumption in the calculation was that the fluid charge is
40 pounds (equivalent to 3 gallons liquid).
While this projected owner's cost exceeds the goal of $100, it
is close enough to warrant the acceptance of RC-1 as a final
fluid candidate on the strength of its more positive attributes.
98
-------
5-9-2 Final Candidate RC-2
Both the water and the 2-methylpyridine components of this can-
didate are currently available in commercial quantities. Water,
even in the highly purified form demanded in this application,
can be obtained for less then let per pound (ref. 32). The cost
of the water per pound of working fluid would, therefore, be less
than 0.4
-------
market to supply the methylpyridines needed at prices quite near
the current prices. By-product imbalances would not likely be
s.evere or ..long-lived. . . • . . •
5.9.2..2 . RC-.2 ,C.0s.t '-. Based upon current prices and projected
demands, "it is? reasonable to assume that Rankine-grade --RC-2, could
be purchased .in .bulk, FOB producer, for 55
-------
6. THE SEARCH
An ability to recognize one prospect among a host of possible
choices is the key to a wide-ranging search for new automotive
working fluids. Known pure chemical species number in the mil-
lions. Since mixtures of pure species are permissible, the pos-
sible choices are truly infinite. A rationale for accepting or
rejecting specific materials at several levels of familiarity was
clearly required to meet contract goals.
Table 39 lists the various characteristics desired of the working
fluid and those characteristics of the propulsion system impact-
ing on fluid choice. These were gleaned primarily from the con-
tract scope of work, given in Appendix A, and from EPA vehicle
design goals (ref. 35). Using these goals, a set of selection
criteria were established to guide the search for new fluids.
6.1 SELECTION CRITERIA
An idea for a new fluid is normally a mental image of its chem-
ical structure. Given such an-image, it is a simple matter to
check the chemical literature and learn whether the compound is
known. If it is, the chances are good that its melting and boil-
ing points are reported, but not much else of great help. Melting
point, boiling point, and chemical structure thus are core data
upon which a first judgment of suitability may be based. A set
of screening criteria cover this level.
Experiment and computation are normally needed to further expand
knowledge of candidate suitability. These are the main approaches
employed in the later criteria covering cycle utility. Safety,
cost, compatibility and lubricant selection are handled on an ad
hoc basis with the advanced candidates only. Fluid cost criteria
were originally established but were later abandoned when found
untrustworthy.
6.1.1 Screening Criteria
6.1.1.1 Criterion 1. Known and Available - The contract re-
quirement of a search for existing fluids is interpreted to
restrict the search to fluids which are quantified combinations
of known chemical compounds, each of which is recorded in the
open literature and obtainable in multi-gram amounts at reason-
able expense without being synthesized by the contractor (except
as a last resort). This rules out fluids which are ill-defined,
secret, hypothetical, unavailable, or very expensive.
6.1.1.2 Criterion 2. Melting Point Below 20°C (68°F) - The
working fluid must obviously be fully liquid at the lowest re-
quired normal-start ambient temperature of -20°F (-29°C) and
either fully liquid or non-damaging to the engine down to the
101
-------
Table 39. PROPULSION SYSTEM AND WORKING FLUID GOALS
Propulsion System Characteristic
Type.Expander .
Apprqx. Net2 Pull Power
Max,; Weight . , ..
Maxi. Volume
Regenerator Size
Min;. WF6 Condensing Temp;
Max
Min
Max,
Max
Low Ambient Temp.,
High Ambient Temp.
Starting' Time (idle at -20°F.
Starting Time (65% power) @ 60°F.
Average Annual Usage
Life Expectancy '"
Working Fluid; Characteristic
Presently available?
Min,. Carnot Cycle Efficiency
Min. Ideal Cycle Efficiency
Max. Cost5 to Owner, 5 yr. period
Compatible with Low Cost Materials?
Compatible Lubricant Needed?
Health Hazard
Fire and Explosion Hazard
Environmental Hazard
NOTES:
Goal1
f 1-stage recip. and
\1-stage turbine
li»53 HP
1600 Ib
35 cu ft
Nil or small
220°F;
-40°F
125°F
25 sec
^5 sec
350 hr
3500: hr . . ;
Goal1*, .
Yes
30%
$100
Yes
Yes - recip..
Maybe - turbine
Nil
Nil
Nil
2.
3.
/I,
5.
Prime sources: 1971 Vehicle Design Goals (EPA) as amplified in
consultation with EPA .
Gross expander shaft less feed pump shaft
Sundstrand calculation based on performance requirements of
1971 Vehicle Design Goals (EPA)
Prime source: contract work statement
Initial factory fill plus materials only for any subsequent
refills . . '..''"••
"Working Fluid"
102
-------
lowest specified ambient temperature of ^40°F (-40°C). To qualify
as a working fluid, a pure compound mustj therefore, have a melt-
ing point of or below -29°C. This restriction does not apply to
the individual components of a mixed fluid, because the liquidus
or crystallizing point of a solution is often many degrees lower
than the melting point of the highest melting ingredient. The
ability of one compound to depress the crystallizing point of
another varies widely, so the selection of the particular value
of 20°C is somewhat arbitrary. It is understood that compounds
having melting points between -29 and 20°C are only to be.consid-
ered potential components of a mixed fluid of suitably low crys-
tallizing point. The term "flow point" has been adopted to encom-
pass the behavior of either simple or complex fluids. It is the
lowest temperature at which the fluid flows, freely.
6.1.1.3 Criterion 3.
Normal Boiling Point 65 to 120°.C (15CK2500F) - The ''
vapor pressure characteristic of the working fluid determines the
pressure to be encountered wherever liquid and vapor coexist at a
particular temperature. It is particularly important, 'that the
condensing pressure is within reasonable limits. Too high a pres-
sure requires a condenser 'so heavy, bulky, or expensive as- to* be:
impractical. Too low a pressure leads to excessive pressure, .drops
in the vapor passages, and poses difficulties in'providing suffi-
cient net positive suction head to avoid feed pump cavitatio.n;. -\
Inleakage of air or outleakage of working fluid through shaft,. _••
seals is more likely the farther removed the condensing pressure
is from atmospheric. Reasonable condensing pressure limits were
taken to be 5 to 50 psia at the advice.of Thermo Electron"Corpora-
tion system design engineers.
By plotting lines of typical slopes through the two limiting
points (5 psia, 220°F and 50 psia, 220°F) on a log-pressure recip-
rocal-temperature chart, it was found that a normal boiling point
range of 150 to 250°F could be expected to differentiate well
those fluids meeting the 5 to 50 psia pressure limit.
6.1.1.4 Criterion 4. Thermally Stable at 720°F (382°C) - The
requirement of a Carnot efficiency of at least 42%, coupled with
a lowest allowed condensing temperature of 220°F, leads to the.
determination of the lowest high temperature of 712°F:
nc - ^^ > 0.42 .(6)
c TR
TL > 460 + 220 = 680°R (7)
TH > 1172°R = 712°F • X';8)
103
-------
A .working fluid me.eting the Carnot efficiency requirement must,
therefore, be. reasonably stable in an engine environment when
heated-to a maximum bulk temperature of at. least 712°F (preferably
hig'h'er.') v. 'For 'the. purpose of a test definition, this temperature
was'-ro'unded off at 720°F.
v- •. /,' ;_ •• ,' •
The - screening .criterion used to qualify a candidate under the
Carnot efficiency requirement was as follows: the fluid must sur-
vive :200 or more hours of exposure in a steel ampoule held at
720°F without serious distress. Details .of the .test and the in-
terpretations are given in section 6.3'.-,
In arriving at the choice of 720°F for 200 hours as the. test cri-
terion, note was taken of the fact, that in a flame-heated dynamic
system the working fluid is exposed to temperatures considerably
above' and below the peak bulk temperature of the fluid leaving
the vapor generator. The highest such temperature, called the
hot-spot temperature, is a metal skin temperature somewhere near
the exit of the vapor, generator. Depending upon design and oper-
ating factors,.it may be several tens to several hundreds of
degrees Fahrenheit above peak fluid bulk temperature. Hot-spot
temperatures are quite difficult to predict or measure accurately;
they are therefore of severely limited utility in considerations
of fluid life. Even though it is at the hot-spot where fluid de-
composition is most rapid, the fact that only a very thin layer
of fluid sees the temperature suggests that a system-overall de-
composition rate is more -readily related to a temperature near
peak bulk. In any event, the selection of ampoule time-temperature
conditions is.arbitrary but nevertheless useful in distinguishing
between stable and unstable fluids.
6.1.1.5 Regeneration and.I-Factor - Compared to an-internal com-
bustion engine of the same power rating, .the Rankine engine is
inherently bulkier,"primarily because.it uniquely requires the
transfer of large amounts of heat across solid-fluid interfaces.
Heat exchangers peculiar to the Rankine engine are:
vapor generator
regenerator
cond.enser
While the automotive Rankine condenser seems comparable to the
familiar radiator, it is required to reject many times more heat.
In the conventional internal combustion engine, the major portion
of the energy loss is through the high temperature exhaust gases,
whereas in the Rankine engine the primary energy loss is in heat
rejected to the ambient air in the condensing process.
The requirement that the Rankine engine fit in an engine compart-
ment designed to house an internal combustion engine places severe
-------
restrictions on component size, and makes it very attractive to
reduce or eliminate heat exchange components.
The regenerator, the only such component whose size or existence
is sensitive to the choice of working fluid, has the sole function
of recovering superheat from the engine exhaust to keep the cycle
efficiency as high as practical. By judicious selection of work-
ing fluid and cycle operating condition, it is possible to arrange
that the engine exhaust be at or near the saturated vapor curve
and thus in need of li.ttle or no regeneration.
It is not difficult to conclude that an automotive regenerator is
"too big" to fit in the engine compartment with the rest of the
propulsion system if it occupies more than about 1.5 cubic feet
of the approximately 30 cubic feet in the compartment. But this
knowledge is of little help in screening many fluids because of
the impracticality of performing the detailed system design anal-
ysis required to fix the regenerator size for each fluid. There-
fore a parameter characteristic of the fluid and expressing the
tendency of the isentropically expanding vapor to converge or
diverge with the saturated vapor line was chosen as a screening
parameter. This is the so-called I-factor.
6.1.1.6 Definition and Calculation of I-Factor - The enthalpy
form of the thermodynamic equation of state for a closed uniform
system is:
dH = TdS + VdP (9)
where H = enthalpy
T = absolute temperature
S = entropy
V = volume
P = absolute pressure
and the enthalpy change in terms of temperature and pressure dif-
ferentials is:
dH = CpdT + a dP
(10)
where a = (3H/3P)T
C = heat capacity at constant pressure
Combining these two expressions, while restricting them to the
saturated vapor path (or dew line, subscript D), yields:
dT
T/CT
dS
D
1 T
V-a
dP
D
(11)
105
-------
This expresses the' slope of the dew line on., a cpnventional temper-
ature,-;ent'ropy' da':agrairi'in terms of 'the saturated .vapor,-, volume and
specific.'heat and the dew pressure derivative. Direct use of this
S;l;ope is hot .convenient because it acquires an infinite value and
:ehang:e;S' sign right--in1 th'e'Ymiddle of the most interesting region,
!<-. e .-.where:'-the line is nfear'vertical. Also, it. is. desirable to
use dimensionle'ss'quantities when comparing-different fluids. For
these reasons, the' T-factor' is defined as a dimensionless function
of the slope as follows:
'T/C
With this definition,- and. considering a1 point on the dew line near
the proposed condensing; .temperature but- far below the''critical
temperature, the -following conventions are valid":
.Fluid Isentropic
.Expansion - Dew Line-
TS Dome Shape ... • Characteristic Slope "••'' I-Factor
Condensing <0 >1
Saturating »- 1+
Superheating. >0 <1
Other authors have called these expansion characteristics wetting,
isentropic, and drying, respectively.
Comparison of Equations 11 and 12 gives the basic relation for
calculating I-f actor:
The approximate form, fully valid where the saturated vapor
behaves as a perfect gas, is intended for use only at temperatures
(and pressures) far below critical. An alternate computational
form is obtained when the Clapeyron equation
106
-------
dPn AHRn
£ _ OU ( | c \
dT T'AVBD. ^15;
is substituted in Equation 13, giving:
Vn-a \ AH'Rn
T _ u "u r i ^ ^
1 - V^VB-) C^T- (16)
- CpT
In these equations, the B subscript refers to the saturated liquid
or bubble line. The approximation is good far from the critical
temperature where both a and the saturated liquid specific volume,
Vg, are negligibly small compared to the saturated vapor specific
volume, Vp. Par from the critical temperature the enthalpy change
of evaporation, AHBD, and the speci.fic heat Cp, of the saturated
vapor are relatively constantj so the I-factor is nearly inversely
proportional to absolute temperature over modest intervals.
6.1.1.7 Criterion 5- I-Factor Between 0.65 and 1.5 - The upper
and lower limits on I-factor were chosen somewhat arbitrarily,
but with the knowledge that ideal cycle efficiency must inevitably
deteriorate rapidly as values of the I-factor deviate widely from
unity.
Subsequent comparisons of real cycle efficiencies with I-factors
forj a number of working fluids have tended to confirm the suita-
bility of the stated limits.
6 . .1. 2 Cycle Criteria
Beyond the five screening criteria, it is important that a working
f~'u:icl perform well in the particular Rankine cycle selected for
:! - Cycle efficiency, for example, is dependent upon the cycle
selected as well as the thermodynamics of the fluid. It is thus
necessary to select and define reference cycles before setting
forth cycle criteria. Two such cycles are defined: the "refer-
ence ideal" and the "equivalent real" Rankine cycles.
6.1.2.1 Reference Ideal Cycle - .Like any simple ideal Rankine
cycle, the reference ideal cycle appears as a clockwise rectangle
on a pressure-entropy diagram. On a temperature-entropy diagram,
its trace is typified by cycle 123^51 (solid line) of Figure 27-
Vapor expansion and liquid compression are ideally reversible
adiabatic processes, and, therefore, isentropic. Cooling and con-
densing occur at a constant low pressure P^ over the path 23^.
Heating, which includes boiling and superheating in subcritical
cycles, occurs along path 51 at a constant high pressure Pj. To
107
-------
712
220
Heat
Expand
Condense
Ideal
Real
S —
Figure 27. Reference Ideal and Equivalent Real Rankine Cycles
108
-------
meet the ^2% Carnot efficiency and the 220°F condensing tempera-
ture requirements, the extreme cycle temperatures are 712°F and
220°F at cycle "corners" 1 and 4, respectively. Finally, the two
pressures are fixed by these conventions:
1. Point 4 is on the saturated liquid (or bubble) line;
therefore, Ft, is the bubble point pressure at 220°F.
2. The pressure at point 1, PI, is the lesser of:
a) that pressure just preventing line 12 from pene-
trating the saturated vapor (or dew) line, or
b) 1000 psia.
The rationale behind the second convention is as follows. Given
a particular fluid, the effect of raising PI at fixed Tt and P^
is to increase the ideal cycle efficiency. There are, however,
two limitations. One is that PI not exceed a practical upper
limit where pump design, cost and reliability become problems, or
where the safety of the occupants of the vehicle may be unduly
threatened. This limit was placed at 1000 psia at the expander
inlet to include the design pressures of the several system con-
tractors. The other limitation is on the incursion of the expan-
sion line into the two-phase region of the fluid. While both the
turbine and reciprocating expanders can tolerate some expansion
condensation, normal practice seems to be to design system hard-
ware in the expectation that condensation will rarely, if ever,
occur. For this reason, it was decided that the ideal expansion
line on the T-S diagram should at no point penetrate the saturated
vapor (or dew line) curve.
6.1.2.2 Equivalent Real Cycle - The "equivalent real" cycle is
a hypothetical cycle associated intimately with the "reference
ideal" cycle, while also-being a much closer approximation of the
intended prototype hardware cycle. Because it is closer to real-
ity than the reference ideal cycle, it is of greater utility in
evaluating and comparing working fluids.
The equivalent real cycle is illustrated by cycle 12'3'^5'1 in
Figure 27 and is defined as follows:
1. Cycle corners 1 and 4 are identical to those of the ref-
erence ideal cycle (i.e., expander and pump inlet condi-
tions are identical in the two cycles)
2. Principal irreversibilities, expressed as expander and
pump efficiencies, and as component pressure drops, were
fixed after comparing the values being assumed by system
designers at Thermo Electron Corporation and Aerojet
Liquid Rocket Company in January 1972 as follows:
109
-------
Equiv.
Real
Cycle
75
Jan *72 Design
TECO
75
Aerojet
0
0
0
75
0
.09
.05
.11
.10
P
P
P
P
it
i
4
1
0
0
0
85
0
.09
.01
.11
.17
Pit
PI
Pit
PI
0
0
0
70
. 0
.00
,10
..00.
.025
Pit
PI
Pit
PI
Expander Efficiency, %
Regeh. Press. Loss, psi
Vapor side
Liquid side
Cond. Press..Loss, psi
Pump Efficiency, %
Vap. Gen. Press. Loss, psi
3. Expander and. pump are assumed to be: adiabatic; thus the
enthalpy change through either component is the negative
of the work done on the. surroundings.
6.1.2.3 Cycle Efficiency and Regeneration - .Whether .Ideal or
real, Rankine cycle efficiency is defined as follows:
Net indicated work out
cy° Net indicated heat in
where net indicated work out is the algebraic;sum of the enthalpy
changes across the expander and the pump (with appropriate sign
change), and net indicated heat in .is the enthalpy rise.across
the vapor generator only. Note that whatever heat is .added to
the liquid by the regenerator reduces the amount of heat required
from outside sources and thus increases the cycle efficiency. Any
development of cycle efficiency criteria thus requires a simulta-
neous development of regeneration criteria.
The heat transfer rate in a regenerator is given as:
Q = UAATm
- UAkATlm
where Q = heat flow, energy/tinie
U = overall heat trans-fer coefficient,
energy/time-are a-temp..
A = interfacial area
ATm = mean temperature difference, vapor
. to liquid sides .
ATlm = logarithmic mean temperature differ-
ence, vapor to liquid' side
k = cross flow factor
(18)
(19)
110
-------
As shown in heat transfer texts (e.g. ref. 36) the logarithmic
mean temperature difference is the "correct" mean temperature dif-
ference to use when the exchanger is truly countercurrent . j.:iat
is, k = 1 for simple counterflow. Other flow conditions, includ-
ing multi-pass and cross flows are accommodated empirically by
values of the multiplier k less than one.
Rearrangement of the heat transfer equation gives
UAk = Q/ATlm (20)
either side of which is a measure of regenerator "size". While
UAk is not a direct measure of regenerator bulk, 'It is certainly
well-correlated with bulk through its sensitivity to exchanger
surface area, and is therefore useful as a size criterion. Obvi-
ously UAk follows the regenerator bulk by increasing when a) more
heat must be transferred at constant ATjm, or b) the same amount
of heat must be transported at a lower AT^m. Implicit here is
the notion that comparisons are made between fluids having not
too disparate thermal properties and at nearly equal pressure
drops. In summary, then, the UAk product is selected as a useful,
if imperfect, indicator of regenerator size and as a suitable
basis for a criterion intermediate between the I-factor limits of
Criterion 5 and the ultimate test of reasonableness provided by
the system analysis, wherein actual optimum regenerator dimensions
are generated.
Regenerator duty is conveniently expressed as its "effectiveness",
which is that percent or fraction of the ultimately recoverable
heat which is actually recovered. The ultimately recoverable
heat is that theoretically recoverable in a pure counterflow re-
generator of infinite area. Depending upon the relative heat
capacities of the vapor and liquid streams entering the regenera-
tor, 100$ effective recovery implies stream temperature equality
(zero AT) at one end or the other, usually at the vapor-out end.
6.1.2.4 Criterion 6. Reference Ideal Cycle Effi-
ciency at Least 30% when UAk <; °> - A 30% ideal cycle
efficiency is required of the working fluid irrespective of the
extent of regeneration. In testing a working fluid under this
requirement, it is not instructive to limit the regenerator "size"
(i.e. the UAk product) because regenerator size is strongly sensi-
tive to expander exhaust fluid condition and to fluid flow rate,
which variables are likely to be much different in a real (irre-
versible) cycle. This disparity, in fact, was the stimulus for
the definition of the equivalent real cycle and the limitation of
its regenerator size.
6.1.2.5 Criterion 7. Equivalent Real Cycle Efficiency when
UAk = 125 Btu/HP-hr-F: Report (not limited) - The assign-
ment of the particular UAk product of this criterion drew on proto-
type system designs of Thermo Electron Corporation and Aerojet
111
-------
Liquid Rocket Company as existing in January 1972 and shown in
Table 40.
Table 40 .
REGENERATOR DATA PRESENTED AT JANUARY 1972
CONTRACTORS' COORDINATION MEETING
luid
type
horsepower
or Q, kBtu/hr
TECO
F-85
Recip.
138
414
Aerojet
AEP78
Turbine
154
1300
Temperatures, °P
Vapor in . 376
Vapor out 248
Liquid in 214
Liquid out 283
Difference, Log.-Mean 59.4
Regenerator UAk*, 50.5
Btu/HP-hr-P
Effectiveness, % 79
Regenerator size
Dimensions, inches 27x8x7
Volume, cubic feet 0.85
480
230
170
400
69.5
121.4
81
2(20x10x6)
1.39
*Per net cycle horsepower
The table data demonstrate the type of correlation between UAk
and regenerator volume alluded to earlier. In both cases, but.
especially in the Aerojet design, the regenerator size shown was
thought to approach the maximum tolerable in the space available.
Thus the criterion value.of 125 represents a "tight fit" with a
turbine system, and is likely a bit "oversized" for use with the
somewhat bulkier reciprocating expander.
No specific contractual limit was imposed on permissible values
of this criterion.
6.1.2.6 Criterion 8. Isentropic Expansion Enthalpy Drop < 200
Btu/lb (Turbine Engine Only - Equivalent Real Cycle) -
The only organic-powered turbine expander under consideration by
the Rankine system contractors during this study .was the single-
stage impulse type. In this type the working fluid is.expanded
virtually to expander exhaust pressure in a single converging-
112
-------
diverging nozzle. For the efficient conversion of fluid kinetic
energy into work by the blades of the single turbine wheel, the
velocity of the supersonic vapor stream exiting from the nozzle
mouth (the "spouting" velocity) must bear a definite relation to
the blade tip speed. Blade tip speed, in turn, is limited by the
strengths of the materials used, in wheel fabrication.
Starting with the assumption that automotive turbine wheels must
be mass produced using inexpensive, plentiful metals, Sundstrand
Aviation Company engineers have determined that a nozzle enthalpy
drop of more than 200 Btu per pound of working fluid is excessive,
The argument follows.
'\
The metals of low-cost wheel fabrication would likely have yield
strengths no greater than 155,000 psl. Sundstrand wheel design
practice is to limit the design tensile stress for such a mate-
rial to 45,000 psi. This corresponds to a wheel tip speed of
slightly over 1200 feet per second in a wheel of any diameter
having a uniform stress shape, i.e. thick at the hub and thinning
with increasing radius. This tip speed, when combined with the
design generalizations of Balje (ref. 37) gives, finally, the 200
Btu per pound enthalpy fall limit of this criterion. This is
understood to be the isentropic enthalpy fall to the engine ex-
haust pressure.
6.1.2.7 Criterion 9. Isentropic Expansion Density Ratio < 25
(Reciprocating Engine Only - Equivalent Real Cycle) -
The only reciprocating expander considered economically feasible
in automobiles is the single stage type. In this type the work-
ing fluid is expanded within a cylinder in one step from inlet
cutoff volume to blowdown. If the pressure of the fluid falls
too little, then too little of the available work of the fluid
will be picked up by the piston, and the engine efficiency will
be too low.
The obvious way around this limitation is to increase the expan-
sion ratio provided by the cylinder-valving design. But there
are mechanical limitations which prevent practical attainment of
a cylinder expansion ratio much greater than 25. The limiting
factor is the need to provide a minimum cylinder clearance dis-
tance and volume to permit unobstructed movement of the inlet
valve. Going to a small bore - long stroke design offers no
great relief because piston accelerations increase, along with
problems associated with bearing lubrication and wear.
Because the working fluid expands on flowing through valves and
ports as well as when trapped within the cylinder, and since the
former is a totally irreversible change, the criterion is based
upon a hypothetical two-step expansion as follows:
Step 1 - fluid expands isentroplcally from engine inlet condi-
tion to an enthalpy equal to the expander exhaust enthalpy
113
-------
Step 2 - fluid then expands further at constant enthalpy until
it reaches' the expander exhaust pressure
1 J; r
The cylinder expansion requirement is then simply the, density
ratio occurring in the first step. The process corresponds to a
fully reversible' cylinder expansion followed by a fully irrevers-
ible blowdown of'the vapor to the exhaust.
6.1.3 Cost Criteria .
The contractual restriction on fluid cost . (Appendix A) is to the
effect that the owner-'s cost for working fluid over the first five
years should not exceed $100. This amount includes all direct and
indirect costs directly traceable to the working fluid except (by
later clarification by.EPA) the cost of labor 9niy for any system
refills required. Costs are.to be based upon a production .rate of
one million automobiles;per year.
Basic to a determination of economic: feasibility is..the allowable
bulk selling 'price of the fluid FOB the chemical plant at the ex-'
pected production volume. This can be estimated working backward
from the $100 owner cost by subtracting 'all shipping, packaging,
labor, overhead and profit costs accrued between plant sale and
customer receipt. Inquiries of several transport companies and
internal sources, coupled with operating experience, led to the
assignment of expected cost elements as listed in Table 4l.
Table 41. MARKUPS AFTER PLANT SALE ....
Markup
Agency \ Service . % .$
Original Fill
Auto Builder Ship, store, install, profit 60
Auto Seller Transport, check, profit 25 —
Refill , . ,
(Freight & packaging, per
Distributor S refill (3 gal) . .— 0.60
vWarehouse, profit . 25 —
Service Station Stock, profit . 35
Using these data, the cost distributions of Table 42 were calcu-
lated for cases spanning 0 to 3 refills.in five years.
114
-------
Table 42. DISTRIBUTION OF COST
Bulk price, FOB chemical plant
Auto manufacturer's markup (60%)
Auto dealer's markup (25%)
Packaging cost (60
-------
The amount of working,fluid to be installed in an average automo-
bile has not been fixed with great certainty. By inquiry of Ran-
kine ;system contractors in early 1972 it was learned the expecta-
tion was for about'1 40 pounds per vehicle. As this corresponds
closely to; 3 g'all!oriS'T6f high density (1.6 x water) fluid, the
3-gallon charge .wa.s assumed from this point on. At an annual pro-
duction' rate-of one million automobiles, total fluid demand would
be about '40 million pounds per year. Allowing 25% for refills,
the expected .annual working fluid demand becomes 50 million pounds
per year.
Under the set of assumptions made thus'far, the highest,tolerable
bulk fluid price is $1.25 per pound for a fluid of density 13-33
pounds per gallon or $2.00 per pound for a fluid of 8.33 pounds
per gallon. To sell at these prices, the fluid would have to
function without replacement for five years.
At the outset of this study it was intended to set go/no-go price
criteria for candidate screening. Limits were to be set for each
of the principal price sources:
1. published bulk prices
2. privately quoted or estimated bulk prices
3- quick-rough estimated bulk prices
4. actual prices of research samples
Early experience with . this approach was so poor as to lead to the
abandonment of the idea of screening early candidates using cost
criteria.
6.1.4 Hazard Criteria
Hazard evaluation is without doubt the most difficult aspect of
working fluid selection. Among the reasons for this are the pau-
city of widely accepted standards, guidelines, protocols, test
methods, definitions and the like. Another reason is the economic
impracticality of carrying out more than a handful of the simplest
of animal toxic exposure experiments, employing only a very few of
the most promising materials. While it is recognized that more
exhaustive evaluations must be conducted before an organic fluid
can be fully recommended for widespread use in Rankine powered
vehicles, it was deemed economically impractical to do so before
the Rankine engine truly is a viable alternate to the internal
combustion engine. In addition, at ,this writing no detailed envi-
ronmental and health hazard criteria have:been established for
rating the acceptability of fluids for the application of this
study.
Largely as a consequence, consideration of hazards tends to be
highly subjective, from laying down requirements to recommending
116 .
-------
courses of action. Thus it is a relatively simple matter to
declare that the "working fluid and its derivatives will present
a negligible health hazard in any physical state when exposed to
any ambient or system conditions", but it is quite a different
matter to prove that any particular fluid, including water, meets
this ambitious requirement. What is, for example, a negligible
health hazard? Is a working fluid hazardous simply because it
will kill a mouse when fed via stomach tube in the amount of
1 g/kg body weight? Is the detection of phosgene in the products
of a fuel-air-working fluid fire sufficient grounds for rejecting
the fluid? How can one predict the effect of releasing millions
of pounds of working fluid into the environment over a period of
years?
Because unequivocal answers to such questions are not to be had,
quantitative hazards criteria were not adopted. Instead, a series
of hazard-related tests were selected and performed on only the
more advanced candidate fluids. The test results were presented
in sections 5-6 and 5.7. This does not mean that potential haz-
ards were ignored in screening early fluid candidates. Every
reasonable effort was made to exclude from consideration materials
known or suspected to be fire-hazardous or toxicity-hazardous.
6.1.5 Compatibility, Stability, Lubrication Criteria
The requirement that the working fluid be compatible with common
materials of powerplant construction was partly satisfied in
screening Criterion *J, where demonstration of compatibility with
carbon steel at 720°F for at least 200 hours is demanded. With
this exception, compatibility, stability and lubrication were
explored in depth only with respect to the final fluid candidates,
as discussed in sections 5-5 and 5.8. Quantitative search cri-
teria were, therefore, not set for these requirements, just as in
the case of the hazard requirements.
In choosing candidate fluids, however, these requirements were
,/er.y much in mind. Very many materials, including nearly all
inorganic chemicals having reasonable liquid ranges, were rejected
from the search as being incompatible with steel at the 720°F tem-
perature. In searching for compatible lubricants, petroleum oils
were shown in special testing (section 6.11) to be unstable at
temperatures above 660°F, so only synthetic and non-petroleum oils
were seriously considered as working fluid compatible lubricants.
6.2 INITIAL CANDIDATE SELECTION (CRITERIA 1-3)
While it is apparent from contractual requirements that a success-
ful candidate must meet many exacting demands, choice of initial
candidates was based on the first four selection criteria. First
of all, it was necessary that the prospective fluid be available
from some source in quantities sufficient for preliminary testing,
as no synthesis effort was contemplated in the contract. Beyond
117
-------
this, the initial selection was based primarily on the normal
boiling point: of the;prospective candidates, coupled with their
estimated thermal stability. Melting or flow points were also
taken into account, but this property was given wide latitude in
Criterion. 2 and was not so restrictive.
6 .2 .'-1 'Normal Boiling Points
As will be seen, some candidates chosen for initial screening
boiled somewhat higher than the 250°F (120°C) upper limit of Cri-
terion 3. These were included for various reasons, such as ready
availability, or because of a need to experimentally develop
thermal stability information.
6.2.2 Chemical Structure and Thermal Stability
The ability' of organic materials to withstand high temperature
depends first of all on the chemical structure of the substrate
material. In practice, however, secondary factors, such as the
presence or' absence of impurities, the types of metals and non-
metals present, sometimes strongly affect the ability of the
organic fluid to withstand the thermal stress. '.'In this regard it
is useful to consider the term thermbchemical stability, which is
not synonymous with thermal stability. The latter may be regarded
as a property possessed by, in the present context,-a Rankine
fluid. It is largely independent of the system.in which the prop-
erty is measured. Thermochemical stability, on the other hand, is
used to denote the stability of a fluid in its. operating environ-
ment, i.e., in the presence of metals, lubricants, contaminants,
etc. The distinction is important in a practical sense. Thermal
stability can be estimated, calculated, or. measured with a reason-
able degree of efficiency, whereas thermochemical stability gener-
ally requires experimental determination under conditions.germane
to the actual operating system. It is useful in this connection
to define the terms decomposition temperature (Tjj) and working
temperature (Tw). Tp is the temperature at which a certain arbi-
trarily selected rate of decomposition ensues when a fluid is
heated in an "inert" environment. Tw is the. maximum temperature
at which the fluid has a useful working life1. The. former almost
invariably exceeds the latter by approximately lOO-^OO0!"1. The
TD-TW difference depends on such factors as operational conditions,
chemical structure of the fluid(s) involved, and the required use-
ful life. Because a great deal more information is available in
the literature on thermal stability, initial candidate selection
was based on Tp values.
Numerous methods have been employed in the past to measure thermal
stability of fluid products. Th'at most widely used toclay involves
some adaptation of the isoteniscope vapor pressure method of A.
Smith and A. W. C. Menzies (ref. 38). To measure thermal decompo-
sition temperatures, advantage is taken of the fact that the most
general change in properties attending thermal decomposition of
118
-------
organic molecules is an increase in pressure in a containing sys-
tem. A close relation between isothermal rates of pressure rise
and decomposition rates can be readily shown. A plot of log dP/dt
vs 1/T is usually a straight line indicating that dP/dt follows an
Arrhenius rate law. Hammann (ref. 39) has defined the TD as the
temperature at which dP/dt = 0.014 mm Hg/sec. Details of the
apparatus and experimental procedures for measuring TD are summa-
rized by Hammann (ref. 39), by Blake and co-workers (ref. 40),
and by Johns, et al. (ref. 4l).
Modern definitive works relating thermal stability and structure
of organic compounds stem largely from research on thermally
stable lubricants and hydraulic fluids carried out; in MRC and
Monsanto Company laboratories. Among the best known of these are
the extensive studies of Blake and associates (ref. 42-50), begin-
ning in 1955 and continuing to the present. In addition, some
particularly useful correlations are given by Johns and co-workers
(ref. 41,51) and Sheehan, et al. (ref. 52). Within this large
body of data, and in extensive unpublished internal sources, ther-
mal decomposition points (Tp) of several hundred organic compounds
are to be found. Many diverse structural types are represented.
From these data certain rules of thumb can be formulated which
permit estimation of Tp's with reasonable accuracy. In the esti-
mation process, account must be taken of possible mechanistic
pathways by which decomposition takes place.
The finite strength of chemical bonds limits the vibrational
energy which molecules may possess . Most compounds have pathways
of rearrangement at temperatures far below that necessary for
straightforward bond rupture. Thus, compounds having configura-
tions or electronic features particularly favoring a low energy
pathway will exhibit abnormally low TD points. An important
example illustrates this. Ordinary esters having one or more
hydrogens in the B-position of the alcohol moiety show Tp points
between 500 and 550°F. Those having the B-position blocked de-
coinoo.se from 40-80°F higher because the low energy decomposition
mechanism shown for the unblocked ester is precluded. Many other
0
II
/C-R
> R-C + HO
/^
R CH2
8-hydrogen ester olefin acid
types of organic materials have comparable low energy pathways by
which they may come apart, rearrange, or self-condense. An example
of the latter effect is shown by the stability of 4-benzoyldiphenyl
ether (TD = 745°F) as compared to 4-acetyldiphenyl ether (Tp =
351°F) (ref. 40). In the latter case self-condensation occurs
119
-------
readily with loss of water whereas no such easy reaction is pos
sible in the. first instance.
H20
Internal stresses within molecules promote instability. Thus,
cyclic carbon compounds and most heterocycles with 3- or 4-mem-
bered rings or those with large, 7- to 9-membered rings are appre-
ciably less stable than: their unstrained 5- and 6-membered co-
geners . Perfluorocyclobutanes and some 4-membered silazanes are
exceptions to these generalizations.
A crude but useful comparison can be made between thermal decompo-
sition of a covalent compound and the breaking of a chain. The
strength of the latter depends on its weakest link; thermal decom-
position temperature is a function of the weakest bond or combina-
tion of bonds. Following this notion one may compile a table of
"weak link" bonds along with the approximate temperature at which
.bond breaking (i.e., decomposition) has been observed to occur.
Such a compilation is given in Table 43. With information of this
type and some knowledge of decomposition mechanisms, a reasonably
accurate assessment of Tp can be made merely by inspection. Inas-
much as a maximum working temperature (Tw) of at least 712°F is
required by Criterion 4 in the automotive Rankine application, it
was anticipated that only materials, having Tp ' s over ^800°F would
have sufficient stability. Ultimately this proved to be the case,
but to avoid overlooking any possibilities and as a means of gain-
ing experimental information, many compounds were included 'in the
initial screening list which were judged to have little chance of
survival. Almost without exception these initial estimates were
confirmed.
6.2.3 Candidate Fluids
Based on normal boiling points, estimations of thermal stability
and other considerations, approximately one-hundred pure fluids
were collected. These are listed in the following Table ..44 along
with their normal boiling points and the sample source. For in-
dexing purposes each fluid was assigned a fluid serial number
(FSN). From the original list of pure fluids a number of mixtures
were then selected for additional examination. Water was an
important ingredient in several of the candidate mixtures, as indi-
cated in Table 45.
In practice, each fluid sample was checked to see that its boil-
ing point corresponded to that reported in the literature. The
sample was also checked for purity by gas liquid chromatography
(GLC). If significant impurities were present, the sample was
120
-------
Tab-le: 4'3". APPROXIMATE, THERMAL DECOMPOSITION. TEMPERATURES' OP., ORGANIC
" COMPOUNDS AS: A' FUNCTION. OF WEAKEST BONDING. UNIT.
Functionality; or Partial'. Bond'
Structure-Representing. Weakest: L'into I'm Molecule!
0 S NH;
I H I- I II II
-CHC-CH -CHCR'. -CHCR
I I
Or - 0,
I H II
-CHCN -C-NH2, -C-OH;,, SCN;, NC=0-
-CHN02 -N02, -CHN2,
,
-CHS02R -CHSRi RSH',. RSRv RSSR
;CH-C-halogen.
-CHOH R3COH: R3COOCR', R3COR', etc.
. ROP: R2NP:- (Triva-lent. Phosphorus) and all P-H and ^P-H structures
R3COP=0 Phosphates of tert alcohols
C
C-C and other' 3-membered heterocycles
S.
. I II1
iCHC-O-C-
Appr.oxlmate TD (°F)
off Molecule, Will Be
Low stability,
functions usually
have TD
<350
ROP=0, -C=CCH2-
-C-H 0, 0
-c- c-,; ;CHOCH;:,.-CH2ociC, -c-NR2
350 - 125
500 - 550
R3CCH2OC-
R2N-, R2N-P=0', R-P:
RP=0
-C-C- -C-H a-llphatlc hydrocarbons only
1 ' ' (a) cyclic, with internal strain
(b) acyclic, with internal strain
(•c) acyclic, without Internal strain
I I, i
R-S1-, ROS1-, ROB-
.11.
i i
»ArCH3, Ar.2CH2, ArCH2CH3-, Ar-Sn-, ArGe-
i i
ArR where R ? C3, Ar-P : , ArP=0
0 0
i II I II
Ar-Cl, Ar-Br, ArCF3, ArO-P=0, Ar-S, Ar-S02-, ArC- , (-Si)20, ArOC-
ArF. Ar-Si-. ArOSi-, ArOB-
I '
Ar(F)n polyfluoro but not perfluoro
Perfluoro aliphatlcs Rf-0-, RfS-, Rfl<
Ar-N-, Ar-0-, Ar-H, Ar-Ar
Benzene, C6F6 , certain simple aromatic heterocyclics and many
perf luorocarbons (Rf)
580 - 615
550 - 620
200 - 580
580 - 615
630 - 680
580 - 680
700 - 730
650 - 700
700 - 800
800 - 850
750 - 800
750 - 850
850 - 900
>850
* Ar = Phenyl, biphenyl and certain other relatively simple aromatic systems
121
-------
Table 44. PURE FLUIDS SELECTED FOR THERMOCHEMICAL
STABILITY SCREENING
FSN
1
3
4
6
7
8
9
10
11
12
13
14
15
16
17
18
21
22
23
24
25
26
27
28
30
31
32
33
Boiling
Pt. % Atm.
Compound Press. (°C)
benzotrif luoride
m-bromobenzotri fluoride
p-bromobenzotr if luoride
o-bromof luorobenzene
m-bromofluorobenzene
bromopentaf luorobenzene
p-bromof luorobenzene '
p-bromophenyl trif luoromethyl
ether
o-chlorobenzotri fluoride
p-chlorobenzotri fluoride
m-chlorobenzotri fluoride
o-chlor of luorobenzene
m-chlor of luorobenzene
p-ch lor of luorobenzene
ch lor open taf luorobenzene
1,2-dichlorohexaf luorocyclo-
butane
1,2-dichlorotetraf luorocyclo-
butene-1
o-dif luorobenzene
p-dif luorobenzene
m-dif luorobenzene
2 , 4-dif luorotoluene
2 , 5-dif luorotoluene
2 ,6-dimethylf luorobenzene
2 , 3-dimethy If luorobenzene
1, 3-di (trif luoromethyl )benzene
1 , 4-di (trif luoromethyl ) benzene
o-f luoroanisole
m-f liioroanisole
102
154
154
156
150
137
151
155
152
139
137
138
126
130
117
60
67
92
88
83
115
H7.
140
146
116
116
155
158
, Melting
i Pt.(°C)
-29
<-78
. -4
-4.0
<-l8
-31
-17
-34
-6
-33
-57
-42
<-78
-22
-15
-16
-45
-34
-24
<-78
-49
-35
<-78
-27
-33
3
-35
-34
Sample
Source*
Pierce
Pierce
PCR
Pierce
Pierce
Pierce
Pierce
Pierce
Pierce
Pierce
Pierce
Pierce
Pierce
Pierce
Pierce
PCR
Pierce
Pierce
Pierce
Pierce
PCR
PCR
Pierce
PCR
Pierce
Pierce
Pierce
Pierce
122
-------
Table 44 Continued
FSN
Compound
Boiling
Ft. g Atm.
Press.(°C)
Melting Sample
Pt.(°C) Source*
34
35
37
38
41
42
43
44
45
46
47
48
49
50
51
52
53
55
56
ji '/
58
59
60
62
63
64
65
66
67
p-f luoroanisole
f luorobenzene
m-f luorobenzotri fluoride
o-f luorob en zotri fluoride
p-f luorophenyl trif luoromethyl
ether
2-f luoropyridine
o-f luorotoluene
p-f luorotoluene
m-f luorotoluene
he xaf luorobenzene
me thy Ip en taf luorobenzene
octaf luorotoluene
pen taf luorobenzene
pentafluoropropanol-1
perf luoro(methylcyclohexane)
perf luoroalkane-70
perf luorot rib utylamine
perf luoro-2-butyltetrahydrofuran
perf luoro(dimethylcyclohexanes )
phenyltrif luoromethyl ether
phenyltrif luoromethyl sulfide
1,2,4, 5- tetraf luorobenzene
1,2,3, 5- tetraf luorobenzene
2,3,5 ,6- tetraf luorotoluene
trifluoroacetophenone
1,2, 4-trif luorobenzene
1 , 3 ,5-trif luorobenzene
2 ,2 ,2-trif luoroethanol
tris (trif luoromethyl )-s-triazine
157
85
100
114
105
126
114
116
117
80
117
104
85
81
76
70
170
102
102
106
140
90
83
125
153
90
75
75
95
-44
-42
-81
-46
<-78
<-78
-78
-56
<-78
5
-30
<-78
-48
-49
-37
<-78
-50
-93
-55
-50
-41
4
-48
-35
-40
-32
-5
-44
-24
Pierce
Pierce
Pierce
Pierce
Pierce
Pierce
Pierce
Pierce
Pierce
PCR
Pierce
PCR
Pierce
Pierce
Pierce
Pierce
Pierce
PCR
PCR
PCR
PCR
PCR
PCR
PCR
PCR
PCR
PCR
PCR
PCR
123
-------
Table 44 :Cdntinued
FSN *
69
70
71
73
74
75
76
77
78
79
80
82
84
86
88
89
91
92
93
94
95
96
98
99
100
101
104
105
106
Boiling
Pt. 6 Atm
'Compound . ... Press. .(°C
,3-chloropyri.dlne
3-me thy Ipy r idine
•pyrazine
toluene
me'thanol
1-propanol
benzene
monochlorobenzene
pyridine
tetrachloroethylene
trichloroethylene
2,2,3, 3-tetraf luoropropanol
perf luoroglutaric anhydride
2,4 , 6-trif luoropyrimidine
hexamethyldisilazane
perf luoroether (P-ID)
acetic acid
morpholine
ethanol
neopentanol
ethylene glycol monomethyl ether
anisole
N-methylmorpholine
<;.
thiophene
1,1,1,2,4,4, 4-heptaf luoro-2 ,3,3-
trichlorobutane
perf luoroheptene-1
N-nonaf luorobutyloc tafluoro-
morpholine
2-me thy Ipy r idine
4-met.hylpyridine
149
144
118
••IIO''
65
97
80
132
115
122
87
110
70
98
125
135
118
129
78'
113
124
154
84
98
82
,115
129
143
. Melting
):.. Pt.(°c.)
<-78
-17
52' '
-95
-9'8;
-127.
5
-45
-H.2 ',
-30
-35,
-13
-38 '
-2
-62"
-121
17 ;
-3
-114
52
-85"' '
-37
<-78
-38
4
<-78'
-45
-66
4
Sample
. Source*
' Aldrich
Aldrich
Aldrich
Fisher
Fisher
Fisher
Fisher
Fisher
Fisher
Fisher
Fisher
PCR
PCR
PCR
Pierce
Allied
Fisher
Fisher
Fisher
Aldrich
Aldrich
Fisher
Aldrich
Aldrich
Pierce
Pierce
PCR
.Fisher
Fisher
124
-------
FSN
Table 44 Continued
Compound
Boiling
Pt. @ Atm. Melting Sample
Press.(°C) Pt.(°C) Source*
107
108
109
110
111
112
pentaf luoropyridine
2-chlorothiophene
2 , 6-dimethylpyridine
2-methylpyrazine
hexamethyldisiloxane
trimethyl borate
83
128
143
135
100
68
-44
-72
-7
-25
V
-63
-29
PCR
Eastman
Eastman
Pfaultz
Pierce
Aldrich
* Sources
Pierce Chemical Company, Rockford, Illinois 61105
PCR, Inc., Gainseville, Florida 32601
Aldrich Chemical Company, Inc., Milwaukee, Wisconsin
Fisher Scientific, Pittsburgh, Pennsylvania 15219
Allied Chemical, Buffalo, New York
Eastman Organic Chemicals, Rochester, New York 14650
Pfaultz & Bauer, Flushing, New York 11351
53233
125
-------
Table
FLUID MIXTURES SUBJECTED TO THERMO
". CHEMICAL STABILITY TESTS
Mixed.;/
Fluid No'.
MESN
:;>3
'11
15
. . 16
18
19
20
21
22
23
24
25
26
-, {Pure 'Fluid
-'"Serial No's'.
FSN" - FSN
93
75
78
78
78
96
66
70
71
105
106
109
110
- 90
- 90
- 90
-'23
- 59
~ 23
- 99
-- 90
-:90
- 94. .
-90 '
- 90
- 90
Mixture :- . , .
ethanol-H20
. i '
n-propanol-H20
pyridine-H20 .
pyridine-p-dif luorobenzene
pyridine-1 ,2,3, 4-tetraf luorb
benzene
anisole-p-difluorob.enzene
2,2,2-trif-luoroet-hanol-water
3-methylpyridine-H20
pyrazine-H20.
2-methylpyridine-water
4-methylpyridine-water
2 ,6-dimethylpyridine-water
2-methylpyrazine-water
126
-------
fractionally distilled until material of satisfactory purity was
attained. In a few cases the identity of harcU-to-separate impuri-
ties were determined by combined GLC/mass spectroscopy so as to
make sure that no low-stability materials were present. Usually
such minor impurities proved to be isomers of the compound in
question. Their presence in small amounts (1-5$) was sometimes
tolerated when it was felt that thermal stability tests would not
be compromised .
6.3 THERMAL STABILITY SCREENING (CRITERION 4)
All of the pure compounds of Table 4*1 and the mixtures listed in
Table ^5 were subjected to a series of ampoule thcrmostability
tests . Details of the ampoule itself and test conditions are out-
lined in Figure 28.
The SAE 1008 low carbon steel tube used to construct the ampoule
was chosen as representing the type of steel ultimately desired
for vaporizer fabrication.
In thermostability testing an appropriate number of tubes (up to
50 at a time) were charged with about one milliliter each of
fluid. These were then capped with no effort made to exclude
ambient air - this to simulate the unsophisticated handling of
working fluid which might be anticipated in actual service. After
closure, a short wire (about 2 ft in length) was attached to ea.ch
tube . The tubes were then suspended vertically in an electrically
heated fluidized alumina bath. Each alumina bath (two were used
in this test work) contained about 2.5 gal of 80-200 mesh alumina.
The temperature in each bath could be controlled at any tempera-
ture desired up to ^750°F with no more than ±3°F variation. Ex-
cept for some minor thermostability studies (section 6.11) all
test work herein reported was done at 720°F under Criterion *J .
After appropriate exposure times at 720°F, ampoules were removed
for examination. The procedure employed was essentially as fol-
lows. After cooling to room temperature, each ampoule was in-
verted (cap down) and cooled to -75°F in an acetone-dry ice bath.
While cold, the welded end of each tube was cut off and a rubber
septum placed over the open end. A 10 cc disposable syringe was
then inserted into the septum and the tube was allowed to come to
room temperature. If a condensable gaseous decomposition product
was present, its presence was indicated by expansion into the
syringe on warming. In this way, too, a gas sample was made
available for examination. Liquid contents were then removed and
examined in an appropriate manner. Generally this involved quali-
tative observations of color, solids content and, where appropri-
ate, of acidity. The presence of strong acid was judged to be
failing, regardless of other fluid conditions. Other pass-fail
criteria were more subjective. Thus, if an exposed sample was
clear, though colored, it was judged as having passed. If solids
were also present, the sample was usually judged as having failed.
127
-------
O~D
I
Tube Specifications
SAE1008 low carbon steel tubing
6 inches long
0.375 in. OD x a 035 in. wall thickness
Total volume 7.0ml
Top Fitting - Swagelok S 600-C
Bottom - Crimped and Heliarc Welded
Test Conditions
Vol. test fluid: 0.5 - 2ml(calc'dto
afford max. pressure in 800-1500
psi range)
Test temperature - 720°F.
Test time - variable
Freeboard - room air at atmospheric
pressure
Figure 28. Steel Ampoule Thermal Stability Test
128
-------
In general, where doubt existed, the fluid was either retested or
passed to the next time interval - usually 24 hours longer. In
this way the thermochemical stability of each candidate fluid was
screened. These results are summarized in Table 46.
Concurrent with the thermochemical screening of pure fluids a
number of mixed fluids were similarly investigated. These results
are summarized in Table 47.
Not surprisingly, nearly one-third of the fluids failed to survive
even 24 hours at 720°F. These included all of the simple ali-
phatic alcohols and, except for trifluoroethanol, all of the par-
tially fluorinated alcohols. Other simple aliphatic acids, amines
and ethers decomposed extensively. With the exception of chloro-
benzene Itself and bromopentafluorobenzene, bromo- and chlorosub-
stituted aromatics failed. All partially chlorinated or fluorin-
ated alkanes were among the early failures as were perfluorohep-
tene-1, perfluoroglutaric anhydride, anisole and the monofluoro-
anisoles. It is noteworthy also that in mixtures with pyridine
the fluorinated benzenes decomposed - this despite the fact that
separately both components were quite stable at 720°F. Fluids
surviving exposures of 24 hours or longer are listed in Table 48.
At this point the question of what constitutes acceptable thermo-
stability must be considered. No agreement exists as to what
survival time in a hot ampoule means in terms of real fluid life
In a Rankine system. Clearly this is a complicated question. One
might conclude that if 10% of the fluid inventory of a Rankine
engine were exposed to the elevated temperature (712°F) at any
given time, then ampoule life at this temperature might be in-
creased by a factor of ten in a real system. Regardless of the
validity of such reasoning, the ampoule test serves to rank fluids
in their approximate order of stability in a static mild steel
environment. The test serves more to define relative stability of
candidates than to forecast their lifetimes in a working system. ;
Far more sophisticated tests are required to do the latter. In
the present investigation a survival time of at least 200 hours at
720°F has been considered necessary for a fluid candidate to be of
interest. Such fluids are listed In Table 49.
While not entirely evident from the data of Table 46, it became ;•
clear during the screening that benzene, toluene, the perfluoro-
(methylcyclohexanes), P-1D, octafluorotoluene and all of the
fluorinated benzenes were stable at 720°F almost indefinitely.
Generally the samples survived the two-week exposure without any
discernible change. Pyridine, pentafluoropyridine, l,3-di(tri-
fluorome,thyl)benzene and the dimethylfluorobenzenes seemed
slightly less stable but survived reasonably well. Pyridine-
water, the methylpyridines (and their water mixtures) and some
fluorinated toluenes appear to be of lesser stability, with slight
decomposition being detectable after 7-10 days. The various
classes may thus be grouped approximately as shown in Table 50.
129
-------
Table 46. RANKINE FLUID CANDIDATES - SUMMARY OF THERMOCHEMICAL SCREENING RESULTS
Hot Tube Thermal Stability Test (Hours Survival at 720CF)
FSN
1
3
k
6
7
8
9
10
11
12
13
11
15
16
17
18
21
22
23
2H
25
26
27
28'
30
31
32
33
31
35;
.37-
Compound
0 2 i 72 9> 120 TP 168 192 2T5 F5o
2FB 312 33?
benzotrlfluoride
m-bromobenzotrifluoride
p-bromobenzotrlfluoride
o-bromofluorobenzene
m-bromofluorobenzene
bromopentafluorobenzene
p-bromofluorobenzene
p-bromophenyl trifluoromethyl ether
o-chlorobenzotrlfluoride
p-chlorobenzotrifluoride
m-chlorobenzotrlfluoride
o-chlorofluorobenzene
m-chlorofluorobenzene
p-chl'orof luorobenzene
chloropentafluorobenzene
1,2-dichlorohexafluorooyclobutane
1,2-dichlorotetrafluorocyclobutene-l
o-difluorobenzene
p-difluorobenzene
m-difluorobenzene
2,1-dlfluorotoluene
2,5-difluorotoluene
2,6-dimethyIfluorobenzene
2,3-dimethylfluorobenzene
1,3-dl(trlfluoromethyl)benzene
Ijl-dKtrif luoromethyl) benzene
o-fluoroanlsole
m-fluoroanisole
p-fluoroanlsole
fluprobenzene
m-fluorobenzotrifluoride
.Stable iiuDecomposltion
9iResults Uncertain
-------
Table 46 Continued
"Hot Tube Thermal Stability Test (Hours Survival at 720°?)
FSN Compound
38 o-fluorobenzotrifluoride
41 p-fluorophenyl trifluoromethyl ether
42 2-fluoropyridine
43 o-fluorotoluene
44 p'-f luorotoluene
'15 m-f luorotoluene
'16 hexaf luorobenzene
'I? methylpentafluorobenzene
48 octafluorotoluene
49 pentafluorobenzene
50 pentafluoropropanol-1
51 perfluoroCmethylcyclohexane)
52 perfluoroalkane-70
53 perfluorotributylamine
55 perfluoro-2-butyltetrahydrofuran
56 perfluoro(dimethyIcyclohexane)
137 phenyltrifluoromethyl ether
58 phenyltrifluoromethyl sulfide
59 1,2,4,5-tetrafluorobenzene
60 1,2,3,5-tetrafluorobenzene
62 2,3 ,5,6-tetrafluorotoluene
63 •trifluoroacetophenone
64 1,2,4-trifluorobenzene
55 1,3,5-trifluorobenzene
66 2,2,2-triflucroethanol
67 tris(trifluoromethyl)-s-triazine
69 3-chloropyridine
70 3-niethylpyriuine
71 pyrazir.e
73 "toluene
74 methanol
IHt !)B 72 9& 120 IP -168 192 213 275
312 33? 3oO
aiiiiiini'
• iiiiiini
I • Ba • an • I
Stable mil Decomposition
Results Uncertain
-------
Table 46 Continued
Hot Tube Thermal Stability Test (Hours Survival at 720°P)
FSN Compound
75 1-propanol
76 benzene
77 monochlorobenzene
78 pyrldine
79 tetrachloroethylene
80 trichloroethylene
82 2,2,3,3-tetrafluoropropanol
84 perfluoroglutaric anhydride
86 2,4,6-trifluoropyrimldlne
88 hexamethyldisllazane
89 Allied P-ID perfluorpether
91 acetic acid
92 morphollne
!_, 93 ethanol
U) 94 neopentanol
95 ethylene glycol monomethyl ether
96 anlsole
98 N-methylmorpholine
99 thiophene
100 1,1,1,2,4,4,4-heptafluoro-2,3,3-
trichlorobutane
101 perfluoroheptene-1
104 N-nonafluorobutyloctafluoromorpholine
105 2-methylpyridlne
106 4-methylpyridine
107 per.taf luoropyrldine
108 2-chlorothipphene
109 2,6-dlmethylpyridine
110 2-methylpyrazlne
111 hexamethyldisiloxane - .
112 trimethyl borate
72 9? 120
168 192 2l5
IBfllBIBQI
IIIIIIIII
IIIIIIIII
IIIIIIIID
IIIIIIIII
•III!III!
IIOIIIBII
•Illlllll
IIIIIIIII
IBIIIIIII
IIIIIIIII
US' 336 36(
t I. t
i Stable
uDecomposltlon IBBICB Results Uncertain
-------
Table 47. STABILITY TESTING OF MIXED FLUIDS
uo
MFSN
3
11
15
16
18
19
20
21
22
23
24
25
26
Mixture
ethanol-H20
n-propanol-H20
pyrldlne-H20
pyridine-p-dif luorobenzene
pyr Idine- 1,2, 3, 4- tetraf luorobenzene
anisole-p-dif luorobenzene
2 ,2 ,2-trif luoroethanol-H20
3-methylpyridine-H20
pyrazine-H20
2-raethylpyridine-H20
4-methylpyridine-H20
2,6-dlmethylpyridine-H20
2-methylpyrazlne-H20
Mol
Hours Survival at 720°F
Fraction 0 24 48 72 96 120 144 168 192 216 240 264 288 312 336
Tested i i i i i i t t i i • i i i i i
0
0
0
0
0
0
0
0
0
0
0
0
0
• 5 -0
• 5 -0
• 5 -0
.5 -0
-5 -0
• 5 -0
.85-0
• 5 -0
.25-0
• 5 -0
• 5 -0
• 5 -0
.5 -0
.5
. 5 IIIIIIIDI
.5
. 5 aiiiiini
. 5 tllllllll
.75 «..«
r-
Stable
mi Decomposition
Results Uncertain
-------
Table 48. PURE FLUIDS SURVIVING 720°F STABILITY
TEST FOR 24 HOURS OR LONGER
FSN
1
8
14
15
16
17
18
22
23
24
25
26
27
28
30
31
35
37
38
41
43
44
45
46
47
48
49
51
Fluid
benzotri fluoride
bromopentafluorobenzehe
o-chlorof luprbbenzene
m-chlorof luorpbenzene
p-chlorof luorpbenzene
chloropentaf luorobenzene
1,2-dichlorohexafluorocyclobutane
o-dif luorobenzene
p-dif luorobenzene
m-dif luorobenzene
2 , 4-di f luorotoluene
2 ,5-dif luprptoluene
2, 6-dimerthy If luorobenzene
2, 3-dimethy If luorobenzene ;
1,3-di (trif luoromethy Dbenzene
1 , 4-di (trif luoromethy 1 )benzene
f luorobenzene
m-fluorobenzotrifluoride
o-fluorobenzotrif luoride
p-f luorophenyl trif luoromethyl ether
o-f luorotoluene
p-fluorotoluene
m-f luorotoluene
he xaf luorobenzene
me thy Ipentaf luorobenzene
octaf luorotoluene
pent af luorobenzene
perf luoro(methylcyclohexane)
Survival
(hr)
24
48
48
48
24
48
120
>336
>336
>336
>240
>240
>240
>240
>336
24
>336
72
24
>192
168
168
168
>336
96
>336
>336
>336
134
-------
Table 48 Continued
FSN
52
53
55
56
57
59
60
62
64
65
6,6
67
70
71
73
76
77
78
88
89
99
104
105
106
107
109
110
111
Fluid
perfluoroalkane-70
perf luorotributy iamine
perfluoro-2-butyltetrahydrofuran
,perf luoro(dimethylcyclohexane )
phenyltrif luoromethyl ether
1,2 ,4,5-tetraf luorobenzene
1,2 ,3 ,5-tetraf luorobenzene
2 , 3 ,5 ,6-tetraf luorotoluene
1 , 2 , 4- trif luorobenzene
1 a 3, 5- trif luorobenzene
2 ,2 ,2-trif luoroethanol
tris (trif luoromethyl )-s-triazine
3-methylpyridine
pyrazine
toluene
benzene
monochlorobenzene
pyridine
hexamethyldisilazane
Allied P-ID perf luoroether
thiophene
N-nonaf luorobutyloctaf luoromorpholine
2-methylpyridine
4-methylpyridine
pentaf luoropyridine
2 ,6-dimethylpyridine
2-methylpyrazine
hexamethyldisiloxane
Survival
(hr)
96*
48*
144*
>336
48
>336
>336
24
>336
>336
24
^312
• 240
192
>336
>336
48
336
120
>336
72
96
264
216
>336
216
72
^312
* Stability believed to be lowered by the presence
of hydrogen-containing impurities.
135
-------
Table 49. PURE FLUIDS STABLE FOR LONGER
THAN 200 HOURS AT ?20°F
FSN '
22
23
24
25
26
27
28
30
46
48
49
51
56
59
60
64
65
67
70
73
76
78
89
105
106
107
109
111
Fluid
o-dif luorobenzene
p-dif luorpbenzene
m-di f luorobenzene
2 , 4-dif luorotoluene
2 ,5-dif luorotoluene
2 ,6-dimethylf luorobenzene
2 , 3-dimethylf luorobenzene
1 , 3-di ( trif luoromethyDbenzene
he xaf luorobenzene
octaf luorotoluene
pen t-af luorobenzene
perf luoro(methylcyclohexane)
perf luoro(dlmethy Icy c lone xane)
1 >2 ,4 ,5-tetraf luorobenzene
1*2,3, 5-tetraf luorobenzene
1,2, 4-trif luorobenzene
1 y 3 , 5-trif luorobenzene
tris(trif luoromethyl)-s-triazine
3-methylpyridine
toluene
benzene
pyridine
Allied P-ID perf luoroether
2-methylpyridine
4-methylpyridine
pentafluoropyridine
2 ,6-dimethylpyridine
hexamethyldisiloxane
Survival
'(hr )-"•'••
'"336 ..
>336
>336;
>240
>240
>240
%240
>336.
>336
>336
>336.
>336';
>336
>336
>336
>336
>336
^312 '
•^288
>336
>336
^336/ '
>336
•^288
^240 '
>336
•x-288
^312
136
-------
Table 50. APPROXIMATE ORDER OF STABILITY IN
AMPOULE TESTS AT 720°F
Stable Indefinitely
benzene, toluene
fluorobenzenes
octafluorotoluene
P-ID
perfluorinated-
methyIcyclonexanes
Stable to
about 336 hr
pyridine
pentafluoropyridine
l,3-di(trifluoro-
methyl) benzene
methyl-substituted
fluorobenzenes
Stable for
1.2*10-312 hr
pyridine-water
methylpyridines
me thylpyridine-water
some fluorinated
toluenes'1
tris(trifluoro-
methyl)-s-triazine
hexamethyldisiloxane
While thermostability testing was in progress other screening cri-
teria were being applied to further refine the selection process.
Included in this effort were calculations designed to evaluate the
thermodynamic qualities of a candidate fluid or fluid mixture.
These approaches will now be outlined.
6.4 PHYSICAL PROPERTIES, LITERATURE AND ESTIMATION
Since the performance of a Rankine cycle engine is significantly
dependent on the physical properties of the working fluid, some
physical property information had to be retrieved for a large num-
ber of compounds. A standard data sheet for recording properties
was prepared (see Figure 29 for an example). As physical proper-
ties were retrieved from the literature or estimated, they were
recorded on these data sheets and the source was recorded on the
back. Two large loose-leaf binders were filled with these data
sheets and supplementary information. Sheer volume prevents their
inclusion in this report, but the notebooks have been filed with
EPA as a supplement to this project.
The extent of the literature search was not the same for all com-
pounds. Initially, only those data necessary to compute I-factors
(see section 6.5) were retrieved. Properties not readily found
were estimated with Monsanto proprietary computer programs. Gen-
erally, the estimation methods used were based on those reported
in Reid and Sherwood (ref. 53).
As the I-factor and thermal stability tests indicated certain com-
pounds to be more promising as candidates or mixture components of
candidates, a more thorough literature search was made, estimated
properties were replaced by literature values wherever possible,>
and sufficient properties were recorded to allow the computation
of complete tables of thermodynamic properties (see Appendix K for
137
-------
FLUID NAME STRUCTURE FRESH
SOPH.
MOt. FORMULA MOL.WT. JR.
SR.
GRAD.
P . -atm ( ) T ,K ( ) REF-
V-, cu.cm/gmol.^ (
HEAT OF VAPORIZATION
cal/gmol @ deg.C
® ( )
{ )
( )
WATSON EXP.— ( )
I-J FACTORS (calc'd)
T
TU
j
Jv .
SOLUBILITY IN WATER
( )
( )
( )
SOURCE
( )
( )
BULK MFG. COST
S 7cwt ( )
S /cwt ( )
) Z ( )
c ..—,..
r -
LIQUID DENSITY
g/cu.cm-® deg.C
; ®" ' • ' ( )
_ ._ ( )
( )
( )
yO = A + BT + CT3
Rng ( ) ( )
A
B
C
LIQUID HEAT CAPACITY
cal/gmol deg.C - .
Cp= A + BT + CT* + DT3
Rna ( ) ( )
A
B '' •
C
D ' - '
THERMAL STABILITY
( )
( )
T0XICITY
( )
( )
VAPOR PRESSURE
ramHg @ deg.C
® ( )
( )
; •"'•()-'•
( )
logloP = A + B/tT-tC)
Rna ( ) ( )
A
B '• •:..-.--- -• •
C : •. ;.::..:-••-•. .
IDEAL GAS HEAT CAP.
cal/gmol deg.C :.
Cp= A + BT + CTa + DT3
Rna ••{•••)' •"' ( )
A
. B • . .
C
D .- . -
FLAMMABILITY
( )
• ::.- ' (:)-
EXPLOSIVE HAZARD
{ )
( )
:..-.
Parentheses indicate reference source, see over ISSUE___ DATE_
Figure 29. Pure Compound Datasheet
138
-------
input data sheets for the thermodynamic grid calculations). The
physical properties of the pure components of the advanced candi-
dates are given in section 5.3. Some of the other fluids and mix-
ture components having promising physical properties are recorded
in Table 51- Properties of other materials retrieved in this
project are recorded in the physical property notebooks.
The computation of thermoclyriamic properties of liquids requires a
knowledge of liquid phase activity coefficients. These quantities
have powerful influence on the bubble point and dew point vapor
pressures of mixtures. Details of the use of activity coefficients
in saturation pressure calculations are given in the description
of computer programs developed for this project (?ee Appendix L).
The Van Laar equations have been used throughout this project, and
the parameters developed for the more promising fluid pairs are
listed in Table 52.
The Van Laar parameters were obtained by processing vapor-liquid
equilibrium data through proprietary computer programs of Monsanto
Company. Where literature data were not available, vapor pressure
curves were measured by isoteniscope for the pure components and
two or more mixture compositions. A special computer program
(E-1375) was developed to obtain Van Laar constants from the iso-
teniscope data. This program is described in detail in Appendix L.
6.5 THERMODYNAMIC SCREENING, I-FACTOR (CRITERION 5)
The use of the I-factor (sections 6.1.1.5-7) as a screening tool
proved to be extremely useful in the fluid selection process .
Under Criterion 55 the range of acceptable values was given as
0.65
-------
Table 51. PURE COMPONENT PROPERTIES USED IN THERHODYNAMIC PROPERTY CALCULATIONS
Fluid:
Molecular Wt
Tc (°K)/Tb
Tc (°F)/Tb
Pc (atm)
Pc (psla)
*zc
«V^ (ftVlfa)
4Hvan (cal/gmol)
pat °K
PT (g/cm3)
at °K
o»Vapor Pressure
A
B
C
»BOHeat Capacity,
Ideal Gas
A
B
C
D
2-Methylpyrldlne
93.129
621.1/1)02.55
657. 98/261). 59
1)5.14
667-20
0.297
0.0573H
8651
H02.55
0.9197
288.15
11.102
-8986.3
233.73
-t. 1626
0.116616
-6.68199 x 10-5
1.301H98 x 10"8
Water
18.016
617.3/373.15
705.1/212
218.2
3206.2
0.230
0.0503
9717
373.15
1.000
277.15
12.1lll(
-1(203.9
-26.75
7.136
0.002611
1.59 x 10-8
0
it-Hethylpyrldlne
93.13
616/118.51
702.80/293.37
16
676.02
0.297
0.05886
10650
298.15
0.9156
303.15
9.915
-3595.1
-55.27
22.2633
-0.0897015
1.53193 x 10-"
-1.10335 x 10-7
3-Methylpyrldine
93.13
-6
1
615/117.29
701/291.12
51.02
793.88
0.2807
0.0173
8932
117.29
0.91736
303.15
10.810
-1211.
-27.18
-1.1626
0.116616
.68199 x 10-5
.3015 x 10-8
Toluene
92.112
591.72/383-75
605.1/230.75
10.55
595.92
0.261
0.05195
7931
383.75
0.866
293.15
10.5882
-1017.31
-7.762
-8.06559
0.132327
-8.08599 x 10-5
1.8719 x ID-8
Trlfluordethanol
100
199.83/311
110/159.51
U8.652
711.99
8755-5
311
1.3736
295.15
11.8
-3712.6
26.97
5.391
0/071913
-5.309 x 10-5
1.3501 x ID"8
"Redlich-Kwong Eq. of State gives Zc = 0.333 for all fluids.
"LnP (atm) = A + B/[T (°K) + C>; fit between Tb and Tc.
•«Cp (cal/gmol - °K) = A + BT + CT2 + DT3.
-------
Table 52. VAN LAAR CONSTANTS FOR RANKINE CYCLE MIXTURE CALCULATIONS
Component 1
Component 2
2-Methylpyridine 3-Methylpyridine 4-Methylpyridine Trlfluoroethanol
Water Water Water Water
Cl 3.9654
C2 792.79
C3 -0.0089793
Ci» 11.984
C5 -1997.1
C6 -0.0146616
11.574
-255.43
-0.019551
11.704
-2174.7
-0.0134725
68.767
-14325.
-0.07005437
-1.0535
602.58
2.35243 x 10-1*
-3.699628
1861.58
0
-2.565628
1861.58
0
NOTE:
2i
= C
C2/T + C 3T ; T in °K.
C5/T + C6T.
-------
Table 53. I-FACTORS FOR PURE FLUIDS SURVIVING ?20°F
STABILITY TEST FOR 24 HOURS OR LONGER
PSM
1
8
14
15
16
17
18
22
23
24
25
26
27
28
30
31
35
37
38
41
43
44
45
46
47
48
49
51
. Fluid
benzotri fluoride
bromopentafluorobenzene
o-chlorof luorqbenzene
m-chlorof luorobenzene
p-chlorof luorobenzene
chloropentaf luorobenzene
1,2-dichlorohexaf luorocyclobutane
o-dif luorobenzene
p-dif luorobenzene
m-dif luorobenzene
2 , 4-dif luorotoluene
2 3 5-dif luorotoluene
2 ,6-dimethy If luorobenzene
2, 3-dimethy If luorobenzene
l,3-di(trif luoromethyl)benzene
1 , 4-di ( trif luoromethy 1 )benzene
f luorobenzene
m-f luorobenzotrifluoride
o-f luorobenzotrif luoride
p-f luorophenyl trif luoromethyl ether
o-f luorotoluene
p-f luorotoluene ,
m-f luorotoluene
hexaf luorobenzene
me thy Ipentaf luorobenzene
octaf luorotoluene
pent af luorobenzene
perfluoro(methylcyclohexane)
Survival
(hr)
24
48
48
48
24
48
120
>336
>336
>336
>240
>240
>240
>240
>336
24
>336
72
24
>192
168
168
168
>336
96
>336
>336
>336
I-Factor
at NBPt.
0.56
0.43
0.69
0.71
0.70
0.64
0.52
0.78
0.79
0.81
0.60
0.60
0.50
0.50
0.42
0.42
0.77
0.55
0.53
0.52
0.63
0.62
0.63
0.69
0.57
0.55
0.71
0.40
142
-------
Table 53 Continued
FSN
52
53
55
56
57
59
60
62
64
65
66
67
70
71
73
76
77
78
88
89
99
104
105
106
107
109
110
111
Fluid
perf luoroalkane-70
perf luorotributylamine
perf luoro-2-butyltetrahydrof uran
perf luoro( dime thy Icy clohexane)
phenyltrif luoromethyl ether
1,2,4 , 5-tetraf luorobenzene
1,2,3 , 5-tetraf luorobenzene
2 , 3 5 5 ,6-tetraf luorotoluene
1,2 , 4- trif luorobenzene
1, 3, 5- trif luorobenzene
2 ,2 ,2-trif luoroethanol
tris (trif luoromethyl )-s-triazine
3-methylpyridine
pyrazine
toluene
benzene
monochlorobenzene
pyridine
hexamethyldisilazane
Allied P-ID perf luoroether
thiophene
N-nonaf luorobu ty loot afluoromorpho line
2-methylpyridine
4-methylpyridine
pent afluoropyri dine
2 ,6-dimethylpyridine
2-methylpyrazine
hexamethyldisiloxane
Survival
(hr)
96*
48*
144*
>336
48
>336
>336
24
>336
>336
24
^312
240
192
>336
>336
48
336
120
>336
72
96
264
216
>336
216
72
^312
I-Factor
at NBPt.
0.26
0.12
--
0.27
0.54
0.74
0.75
0.59
0.77
0.77
—
0.36
0.82
0.84
0.67
0.89
0.74
0.87
0.33
0.28
1.35
—
0.56
0.82
0.75
0.75
--
0.35
* Stability believed to be lowered by the presence of hydrogen-
containing impurities.
143
-------
Table 5A. THERMALLY STABLE COMPONENTS.AND I-FACTORS
PSN. . . . Name
56 perf luoro(dimethylcyclohexane)
89 perfluorpether P-ID (Allied)
111, hexamethyldisiloxane
6? tris ( trif luoromethyl )-s-triazine
51 perfluoro(methylcyclohexane)
30 1,3-di (trif luoromethyl )benzene
28 2, 3-dimethy if luorobenzene
27 2, 6-dimethylf luorobenzene
48 octaf luorptoluene
105 2-methylpyridine
25 .2,4-difluorotoluene
26 2,5-difluorotoluene
73 toluene
46 hexaf luorobenzene
49 pentaf luorobenzene
59 1, 2, 4 ,5-tetraf luorobenzene
60 1,2, 3, 5-tetraf luorobenzene
107 pentafluoropyridine
109 2,6-dimethylpyridine
64 1, 2, 4-trif luorobenzene
65 1, 3, 5-trif luorobenzene
35 f luorobenzene
22 o-difluorobenzene
23 p-dif luorobenzene
24 m-dif luorobenzene
70 3-methylpyridine
106 4-methylpyridine
78 pyridine
76 benzene
90 water
Hours
Stable
at
720°P
>336
>336
312
>312
>336
>336
>288
>336
>336
>288
>336
>336
>336
>336
>336
>336
>336
>336
>288
>336
>336
>336
>336
>336
>336
288
240
>336
>336
NBPt
I- Result^
Factor Grit. 3
0.27 0
0.28 X
0.35 ~0
0.36 o
0.40 0
0 . 42 0
0.50 X
0.50 X
0.55 '0
0.56 X-
0.60 0
,0.60 , . 0 .
0.66 . 0 ,
0.68 0
0.71 o
0.74 ' 0
0.75 '0: ' •
0:75 0
0.75 X
0.76 : 0 .
0.76 ... o
0-77 ..; 0 . .
0.78 o
0.79 0
0.81 ' 0
0.82 X
0.82 ' ' X
0.87 o
0.89 . 0
2.81 : .0 .
0 = pass, X =. •fail.-
-------
6.6 COST PROJECTIONS AND SCREENING
As described earlier (section 6.1.3), it was originally antici-
pated that a price for each candidate could be easily found or
estimated. Prices were obtained for many of the compounds on the
initial candidate list from the Oil, Paint and Drug Reporter, a
trade paper of the chemical industry, or from private communica-
tions with current suppliers for candidates that were not articles
of commerce. The resulting list, with prices and price sources
noted, is given in Table 55.
Two things became apparent very quickly. "Articles of commerce",
which could be priced quickly from such sources a--; Oil, Paint and
Drug Reporter, were almost as quickly eliminated from the lists
by failing to meet other criteria. Second, those less common can-
didates for which suppliers could be found were usually eliminated
on technical grounds before correspondence concerning their price
could be completed. It was obvious, then, that technical screen-
ing was quicker and more effective than economic screening.
In addition, it became obvious that most common materials, whose
prices would likely meet the contract requirement of $100 maximum
cost over the life of a vehicle, would not meet the technical cri-
teria. The economic question then became, how much will the fluid
cost a vehicle owner (and how can these costs be minimized); not,
which of the fluids that meet the $100 limit are the best techno-
logical choice.
At this point in the work, economic screening was abandoned and
additional work along these lines was confined to predictions of
the cost of only the final candidate fluids, as presented earlier
in section 5-9•
6.7 FLAMMABILITY TESTING
Contractual requirements relating to fire and explosion hazards
stated that the "....working fluid and its derivatives will not
be capable of sustaining combustion under atmospheric conditions
or in a high temperature environment once the auxiliary source of
combustion has been removed. Explosion hazards from working fluid
vapor/air mixtures shall be minimized and a test used to demon-
strate this."
Tests selected to establish flammabilities of candidate fluids in-
cluded traditional flash/fire/autoignition temperatures and a
specialized "hot compartment ignition test". Not all candidates
were subjected to flammability tests. Aside from the two final
candidates, whose test results were given earlier in section 5.7,
most experimental effort was directed to the various organic/water
mixtures proposed as candidates.
-------
Table 55.' PUBLISHED AND ESTIMATED PRICES OF SELECTED FLUIDS
Defined • Organic-Component : Bulk pricet $/lb
FSN
; 22 -.
. 23
.24'
35
46
49
59
60
.70
71
73
76
78
81
82
105 '
106
' . Name
. o-dif-luprobenzehe
p-dif luorobenzene
m-di'f luorobenzene
f luorobenzene
hexaf luorobenzene
pentaf iuorbbenzene ;
1,2,4, 5-tetraf lubrobenzene
1 >2 , 3 ,5-tetraf luorobenzene
3-methylpyridihe
pyrazine
toluene
benzene • •••
pyridine
trichioroethane .
2,y2 , 3 i 3-tetraf luorppropanpl
2-methylpyridine '
4-methylp'yridine
Published1
0-.60
, 0.03
0.03"
0.52
0.13
0.46 .
0.40'
Estimated2
7.20
5-80
10 .00
.1.10
12.00
12.00
12:00
12.00
1.00
12.00
Defined Mixtures
MFSN
1
3
5
6
7
8
9
10'
12
13 .
14
15
17
.23 '.
24
Components
FSN-73 and FSN-78
FSN-90 -and FSN-93
FSN-75, FSN-76 and FSN-90
FSN-76, FSN-90 and FSN-93
FSN-73 and FSN-75
FSN-75 -and FSN-76
FSN-73 "and FSN-74
FSN- 74 and FSN-76
FSN-74, ESN.-78 and FSN-90- .
FSN-73 and FSN-76
FSN-71 and FSN-90
FSN-78 and FSN-90
FSN- 30 and FSN-78
FSN-90 and FSN-105
FSN-90 and FSN-106
0.14
0 . Od
0.03
0.04
0.08
0.05
0.02
,6.02
0.18
0.03
0.62
0.42
1.60
0.30
0.25
1 From Oil, Paint and Drug Reporter, 1972
2 From private estimates
146
-------
6.7.1 Flash and Fire Point Results
A modification of the Cleveland Open Cup method of ASTM procedure
D92 was employed. In this test, a small sample of the working
fluid is heated continuously in a cup. Periodically, a small gas
flame is passed over the cup. The flash point is the lowest fluid
temperature at which ignition is observed, and the fire point is
that temperature where sustained burning occurs. The modifica-
tions, described in detail in Appendix N, were made to reduce the
amount of sample required and to accommodate the above-normal
vapor pressures of the working fluids. Principal results are
listed in Table 56. The data show that increased water content
increased the flash/fire points of the flammable organics, but in
no case at concentrations below 90 mol % did its presence prevent
ignition altogether.
6.7-2 Hot Compartment Spray Ignition Results
This test was used to assess ignition and explosion hazards of
working fluid candidates. The apparatus consists essentially of
a cylindrical chamber with heated walls fitted with a spark ig-
niter, a spray injection assembly, and a loose-fitting lid. Addi-
tional details are given in Appendix 0. In use, the chamber is
brought to a desired temperature (after thoroughly flushing with
air), the arc is started, and the test fluid introduced as a fine
spray through an atomizing nozzle. The occurrence and intensity
of ignition are observed via an accelerometer pickup shock meter
reading, while any resulting energy released by combustion is'
recorded as a compartment temperature increase.
Generally, two types of combustion and variations take place in
the hot compartment spray test. When the fluid is introduced at
high pressure (e.g. 1000 psig) , a fluid-rich atmosphere is created
around the spark ignition source. The ensuing rapid ignition con-
sumes the air present and no sustained fire follows. The second
condition is under low pressure spray where a lower fuel air ratio
exists which favors slow ignition, creating high enough compart-
ment temperatures to induce more air from outside the compartment
to sustain a fire until the fluid is totally consumed.
The results of the hot compartment spray ignition tests of several
fluids are contained in Table 57.
6.8 PHYSICAL PROPERTY MEASUREMENTS
Literature search and estimation techniques were used to produce
physical properties for all simple fluids, as described in section
6.4. When dealing with mixtures, however, predictions are not so
reliable, and literature data is often non-existent. Measurements
are then required to obtain reliable properties. Measurements of
vapor pressures and flow points were taken with several of the
more promising mixture candidates.
-------
Table 56. FLASH AND FIRE POINTS OF MIXED FLUIDS
Fluid
MFSN-15 pyridine/water
mol $ pyridlne =
100
75
50
35
25
15
10
MFSN-20 trlfluoroethanol/water
mol % trifluoroethanol = 85*
MFSN-21 .3-methylpyridine/water
mol % 3-methylpy}ridine = 100
75
50
15
MF.SN-,2.3 2-methylpyrldlne/water
mol % 2-methylpyridine = 100
75
40
15
MFSN-24 4-methylpyridine/water
mol % amine = 75
50
30
15
MFSN-25 2,6-dimethylpyridlne/water
mol % amlne = 35
MFSN-28 pentafluorobenzene/hexafluorobenzene
mol % penta = 60
Flash
Point
(op)
85
90
95
100
115
125
130
105
115
125
130
165
90
90
120
135
125
135
150
175
None
Fire
Point
90
95
105
110
130
150
160
160**
120
145
160
190
100
105
145
175
155
180
205
210
200
None
* Same as "Fluorinol-85"
** Mild, low-intensity, perimeter flame
148
-------
Table 57. HOT COMPARTMENT IGNITION TEST RESULTS
Test Conditions
Initial fluid reservoir charge = 30 cc
Compartment temperature = 100°C (212°P)
Compartment environment = air
Spray time = 1/4 sec
Test
No.
1,2
Fluid
FSN-76
Benzene
13
FSN-76
Benzene
MFSN-15
Py:H20
50 mol %
15
MFSN-15
Py:H20
50 mol %
MFSN-20
TFE:H20
85 mol %
Fluid Reservoir
Pressure Temp.
(psig) (°C)
1000
1000
1000
1000
500
500
100
70
1000
1000
1000
1000
500
500
100
100
1000
1000
1000
1000
500
500
100
100
1000
1000
1000
1000
500
500
100
100
1000
1000
1000
1000
500
500
100
100
100
5^0
565
595
620
500
530
330
500
500
550
580
610
500
190
360
380
520
540
550
560
470
480
320
320
520
540
550
560
460
500
335
410
410
440
460
970
390
400
270
280
285
Fire Results
Compart-
ment Temp .
Rise (°C)
87
60
63
58
107
75
143
134
45
30
60
58
74
62
140
135
70
102
80
74
138
142
140
128
64
100
74
62
80
118
205
163
98
88
47
78
85
87
0
160
80
Shock
Meter
Reading
15
25
12
8
10
7
4
5
8
15
15
10
20
5
2
2
10
15
10
12
20
7
1
1
12
8
14
18
18
3
1
2
3
10
15
16
4
3
0
0
0
Flash
Ignition
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Delayed
Delayed
Delayed
Delayed
Delayed
No
Delayed
Delayed
Post
Ignition
Fire
Long
None
None
None
Long
None
Long
Long
None
None
None
None
Long
None
Long
None
None
Long
None
None
Long
Long
Long
Long
None
Long
None
None
None ••
Long
Short
Short
None
None
None
None
None
None
None
Long
Long
1*19
-------
Table : 57 Continued.
Fluid'Reservoir
Test;
No.
14
Fluid
MFSN-20
TFE:H20
85 mol %
MFSN-21 •-
3CH3Py:H20
50 mol %
11
MFSN-21
3CH3Py:H2P
50 mol %
MFSN-23
2CH3Py:H20
35 mol %
MESN-23
.2CH3Py:H20
35 mol %
Pressure.
(psig)
1000
1000
. 1000-
" 1000
500
560:
500"
100
100
100.
1000
iooo
1000
1000
500
500
100
100
1000.
1000-
1000
1000
500
500
100" :
100
1000 .
1000-
1000
1000
500
500 :-
100
100 :
iooo;
1000
1000
1000.
500
500
100 . :
100
100
.Temp."
390
425
450
465,'
405
. 400 '
400
265
270
270
560
590
610
620
530
545
360
320
560
580
605
630
550
590
440
470
520
530
535
535
450
460
320
350
460
510
530
540
450
460
330
320
310
. . Fire Results
Compart-
ment Temp .
Rise (°C)
93
53
62
62 .
76 .
84
88
142
__
96
102
67 •
70
67
108
127
112 '. ".
119
80 .
52
66
80
124
119
125
140
90
64
68 ,
77
92
100
180
158 / .
74 -
60
72 .
72
88
132
159
132
166
. Shock
. Meter
Reading
1
4
15
.-. 5 .
13
2
"8
1
__ '
1.
1'
7
7
4
4
3
1,' , '
2'. .
20
7
4
2 .
2
2
2
1 .
8
6
8
6 .
3
7 "
l - :
2 ,.
5
8
10
10
3
r
i, .
2
2'
Flash
Ignition
Delayed
Delayed
Yes
Delayed
Delayed
Delayed;
Delayed
Yes
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes .
Yes
Yes
Yes
Yes
Yes
Yes
Yes .
Yes .
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Post
Ignition
Fire
None
None
None
None
None
None
None
Short
None
Short
None
None
None
None
Long
Long
None
None
None
None
None
None
Long
Long
None
None
None
None
None
None
None
Long
Short
Short
None
None
None
None
None
None
Long
Long
Long
150
-------
Table 57 Continued
Fluid Reservoir
Test
No.
3 .
12 '
4
10
16
Fluid
MFSN.-24
4CH3Py:H20
35 mol %
MFSN-24
4CH3Py:H20
35 mol %
MFSN-25
2,6-Luti-
dine:H20
35 mol %
MFSN-25
2,6-Luti-
dine:H20
35 mol %
MFSN-28
C6F5H
60 mol %
40 mol %
Pressure
(psig).
1000
1000
1000
1000
500
500
100
60
1000
1000
.1000
. 1000
500
500
100
100
1000
1000
1000
1000
10 00
500
100
100
1000
1000
1000
lobo
500
506
100
100
1000
1000
500
500
500
100
100
100
Temp.
(°C)
540
550
550
550
480
490
380
500
535
5.45
560
565
475
485
390
420
550
565
570
590
625
540
480
550
520
550
565
570
480
520
320
320
500
545
480
495
550
305
380
500
Fire Results
Compart-
ment Temp.''
Rise (°C) -
35
0
75
72
132
?3
155
150
67
65
60
. 65
79
88
122
125
65
107
70
80
77
105
116
108
__
80
84
80
95
91
134
152
-18
-19
-15
-14
-11
17
14
6
Shock
Meter
Reading
- 1-3
0
15
-12 . .
2 •
6
1
.1
15
8
5
25
5
1
1
1
0
10
7
13
7
30
2
2
— -—
10
8
20
4
4
2
2
0
0
0
1
1
1
2
1
Flash
Ignition
Yes
No
• Yes
Yes
Yes
Yes
Yes
Yes'
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
Slight*
Slight
Slight
Post
Ignition
Fire
-None
None
None
None
•Long
None
Short
Short
None
None
None
None
None
None
Short
Short
None
None
None
None
Short
Long
None
None
None
None
None
None
None
None
Long
Short
None
None
None
None .
None
None
None
None
* During sparking, the pyrolysls products from the fluid caught in
the spark do burn but do not sustain burning after spark stops.
151
-------
6 .8.1 Vap'or Pressure
Vapor pressures .were measured using a Monsanto-developed recording
tensimeter described in.detail in Appendix P. The pressure of a
confined .liquid-vapor 'sample at a particular temperature generally
depends upon both;the.Coverall sample composition.and the fraction
of the sample which is in the vapor state. If the sample is all
liquid but vaporization,is just starting, the pressure,is referred
to as the bubble point pressure and the fluid Is said to be at its
bubble point. When the:sample is all vapor with condensation Just
starting, the equivalent condition is the dew point. Between
these extremes, which may differ considerably at either fixed
pressure or temperature, the pressure is meaningful only when re-
lated to the degree of vaporization.
A simple fluid shows no change in pressure during isothermal va-
porization. Its dew and bubble points are identical, and its
single pressure of vaporization is properly called its vapor pres-
sure. Most complex fluids have .dew points several to many degrees
higher than bubble points at a fixed pressure. An azeotrope,
which is a highly non-ideal complex fluid, is an exception to
this, but only at one particular temperature (or pressure) where
its' dew and bubble points coincide.
"Vapor pressures" measured with the recording tensimeter are the
pressures obtained with the sample bulb half filled (by volume)
with liquid at the beginning of the measurement. This corresponds
closely to the bubble point condition at temperatures and pres-
sures well below critical, where vapor densities are negligibly
small compared to liquid densities.
In Table 58 appear vapor pressures measured with the water-ethanol
system. This mixture was used as a well-known standard exhibiting
non-ideal .behavior for the development of the mathematical proce-
dures for calculating mixture thermodynamics. Vapor pressures of
the final candidate fluids were given earlier in section 5.2.
6.8.2 Flow Points
When a liquid solution is slowly cooled a temperature is reached
where first crystals appear. The only visual' evidence may be a
cloudiness of the fluid. If cooling is continued, the crystals
grow in size and number to the point where the liquid's mobility
is noticeably diminished. This point, peculiar to complex fluids,
is referred to in this study as the flow point.
A simple fluid or a eutectic solidifies completely at a single
temperature normally called the freezing or solidification point.
Complex fluids otherwise solidify over a range of temperatures.
The flow point then represents the lowest temperature at which the
fluid, simple or complex, can be expected to function in an engine.
152
-------
Table 58. VAPOR PRESSURES IN PSIA - ETHANOL/WATER
(°C) 0.0 0.25
50
75
100
125
150
175
200
225
250
1.8;
5.6;
14. 7
33.. 7
69
129
225
370
577
- 4 ..3 .
11.6
28 . 2
(59)
122
224
370
580
850
• 0.50
..-5.0
.13-3
31.. 3
(6;6)
132
246
410
660
950
'•• '0.75 1.00
5.. 0 " '
13/9 '•
32.8
(70)
142
264
430
690
(1020)
•4.; .3
13.0
3:2. 8<
72 :
143
254
429
676
— —
NOTE: Figures in brackets are literature data.
Figures in parentheses are extrapolated.
153
-------
The flow point's^ measured with several of the advanced candidates
are presented in Figures 30 to 32.
6.9 THERMODYNAMIC TABLE/DIAGRAM.COMPUTATIONS
Two special computer programs were developed to compute thermo-
dynamic. tables and diagrams for this project. Program E-1393 pre-
pares thermodynami-c property tables. It is described in detail in
Appendix K. Program TSPLOT interfaces with E-1393 to produce tem-
perature-entropy diagrams. It is described in detail in Appendix M,
Thermodynamic; tables, and TS diagrams were prepared > for a number of
fluids. Tables and charts for the advanced candidates are covered
in section 5.3. The other tables and charts are not included in
this report because of excessive volume. They have been filed
with EPA. 'A; list of fluids for which charts have been filed is
given in Table 59.
Table 59. FLUIDS FOR WHICH THERMODYNAMIC TABLES
AND CHARTS HAVE BEEN PREPARED
Component 1
Trifluoroethanol
Trifluroethanol
Pentafluorobenzene
Pentafluorobenzene
Pentafluorobenzene
4-Methylpyridine
4-Methylpyridine
Toluene
Ethanol
Pyridine
2-Methylpyridine
3-Methylpyridine
Component 2
Water
Water
Hexafluorobenzene
Hexafluorobenzene
Hexafluorobenzene
Water
Water
Water
Water
Water
Water
Mol percents (1/2)
100/0
85/15
100/0 . :
60/40
0/100
40/60
30/70
100/~
60/40
10/90;
40/60;
100/0
10/90;
40/60;
100/0
10/90;
40/60;
100/0
20/80; 30/70;
50/50; 75/25;
20/80; 30/70;
50/50; 75/25;
20/80; 30/70;
50/50; 75/25;
154
-------
40
30
20
10
0
-10
-20
S. -30
-40
-50
-60
-70
-80
-90
<» 4 MP in Water
o 3 MP in Water
„ 2 MP in Water
D Pyridine in Water
2 MP = 2-Methylpyridine
3 MP = 3-Methylpyridine
4 MP = 4-Methyipyridine
10 20 30 40 50 60 70 80
Mol % Methylpyridine (or Pyridine)
90 100
Figure 30. Flow Points - Binary Aqueous Pyridines
155
-------
o Equimolar 2MP/4MP jn Water
• Equimolar 2MP/3MP in Water
Q Equimolar 3MP/4MP in Water
2MP = 2-Metnylpyridine
3MP • 3-Methyipyridine
4MP • 4-iyiethylpyridine
20
30
40 : 50 60:; 70
AAol % Methylpyridines
80 90
100
Figure 31.. Flow Points - Ternary Aqueous Pyridines
156
-------
0 10 20 30 40 50 60 70
Mol % Pentafluoro Component
90 100
Figure 32. Plow Points - Binary Fluoroaromatics
157
-------
6.10 RANKINE CYCLE COMPUTATIONS (CRITERIA 6-9)
In order to investigate the utility of the various candidate
fluids as Rankine cycle working fluids, and to test them against
Criteria 6 through 9, standard cycle analyses were ..provided and
performed. As presented in section 6.1.2, two particular.cycles
were defined which had in common the same fluid,state conditions
at expander and pump inlets. Beyond this, the "'reference ideal
cycle" assumed 100% mechanical efficiencies for expander .and pump,
zero pressure losses between expander and pump, and no limitation
on the extent of heat exchange surface available in the regenerator,
In the "equivalent real cycle", expander and pump efficiencies
were assumed to be 75%, representative pressure losses were assumed
(section 6.1.2.2) in the various components between expander and
pump, and a realistic limit was imposed (section 6.1.2.5) on the
extent of regenerator heat exchange surface. In both cycles it
was assumed that the expander and pump' operated adiabatically.
The particular definition of cycle efficiency used throughout is:
Cvcle efficiency = Net indicated work out (W) i
oycie eniciency Net indicated heat in (Q)
where W = the algebraic sum of the enthalpy changes
across expander and pump (with appropriate
sign change)
Q = the enthalpy rise across the vapor generator
only
The limitation imposed on the regenerator size in the equivalent
real cycle was actually a .ceiling on the value of its UAk product
(overall heat transfer coefficient times exchange surface area
times cross-flow'factor) of 125 Btu/HP-hr-°F.
A detailed description of the computational methods is given in
Appendix Q.
6.10.1 Results of Computations
Mixed fluid "cyc.le calculations were carried out by the desk calcu-
lator-program E-1393 interaction described in Appendix Q. Simple
fluid (single component) cycle calculations were made using a
Monsanto proprietary "Cycle Trace Program". The latter employs
the Hirschfelder-Buehler-McGee-Sutton (HBMS) equation of state
(ref. 5*0 j but the results are comparable to the Redlich-Kwong
equation used in the. mixture calculations.
Cycle computations for advanced candidates are given in section
5 • ^ • Cycle calculations for other fluids are summarized in Tables
60 through 67- Unless otherwise designated, the Ideal Reference
Cycle (R2) was used. Fluids identified by serial number in the
tables are:
158
-------
Table 60. REFERENCE IDEAL CYCLES - SIMPLE FLUIDS
EFFICIENCIES
CYCLF
* OF
CYCLE.
CAHNOT
U10 NO.
MCE
ES. *
APMOT
C X RbN
no x R b n
£ , MAX F
f» 1 11 F
MAX PS1A
I-'IN PS1A
23
?080
3C.OO
71 .44
^1.77
31.68
41.99
712.0
220.0
nio.4
23.64
24
2080
30.no
71.44
21.05
31.63
41.99
712.0
220.0
1000.0
28.47
35
2080
30.00
71.44
21.10
31.90
41.99
712.0
220.0
1000. 0
25.94
46
11211
30.00
71.44
18.18
31.39
41.99
712.0
220.0
1000.2
30.51
fiO
?080
30.00
71.44
21.05
30.64
41.99
712.0
2?0.n
1 000.4
20.18
70
2080
30.00
71.44
26.97
32.42
41.99
712.0
220.0
823.4
4.18
73
2081
30.00
71.44
23.85
31 .67
41.99
712.0
22P.O
1000. 0
12.28
76
2081
30.00
71.44
23.33
31.81
41.99
712.0
2PO.O
icon. 7
29.45
78
2080
30. PP
71.44
27. ?7
31.45
41 .99
712. n
220. n
i o a o . 6
10.54
90
12051
18. Jb
43. ?3
10.36
20.51
41.99
712. P
220. n
170. fa
17. 2t
107
20B1
30. uO
71.44
20.78
30.^3
41.99
712.0
220.0
1000.5
28.16
Of K 1UO CYCLE HP
KLUIO KATE .LB/HR
LNC. EXH.CFM
PUt'P IN. GPM
fNGIME HP, GROSS
PUMP HP
HEAT FLOWS
HEATEK KQTU/HR
REGCM KBTLI/HR
COMJEI'J KBTU/HR
WITHOUT REGEN
HEATER KBTU/HR
CONOEW K8TU/HR
ENGINE. * EFFIC =
PRESSURE RATIO
DENSITY RATIO
ISENTROPIC (1)
EXHAUST QUALi *
DELTA -H. BTU/LB
ISEWTRCPJC
KAPCA (2)
MOZZLES. COEFF =
V SPUUTi FPS
fACH NO. SPOUT
A THROAT (3)
i
-------
Table'" 61- EQUIVALENT REAL CYCLES - SIMPLE FLUIDS
FLUID NO.
KEFERENCE .
EFFICIENCIES, X
. CYCLE
* OF CARNOT
CYCLF, 0 * RGW
100 x RGN
C.ARliOT •
TF.MPERATURE, MAX F.
- ' : MIN F
PRESSURE i MAX PSIA
WIN PSIA
PER 100. CYCLE HP
FLUID PATE«LB/HR .
ENG EXH.CFM . :
PUPTP IN, -GPH
ENGINE HP. GROSS
PUMP HP -.•••••
HEAT FLOWS • -.
HEATER KBTU/HR
RE-GEN KBTU/HR
CONDEN- KBTU/HR
WITHOUT REGEN .
HEATER KBTU/HR
CONDrw KBTU/HR
TNGIME, X EFFIC =' .
PRESSURE RATIO
DENSITY RATIO . '
ISENTROPIC (1)
EXHAUST DUAL. * .
DELTA -H, BTU/LB
ISEMTROPIC
NOZZLES, COEFF .=
V SPOUT, FPS
PACH NO. SPOUT
A THROAT (3)
RCGEN, X EFFECT =
0, K9TU/CHP-HR
UA.KBTU/CHP-HR-F
TEL^TA T (L'H)i F
MTU (OGI
I-FACTOR
AT T, F, =
PUMP, X EFFIC =
PRESSURE RATIO
i-3
2080
22.25
53.00
11.11-
23.21
11.99
712.0"
220.0.
1150.2
23.61
6961. •-.
362. '6
16..91
111.3
1.1.31'
1111.2
622.8
889.5
1767.0
:1512.1
' 75..00
. 31.39
16.68
19.50
181.0
11 1 82
55.76
0.931
0.950
1608.
?.17
0.068
92.71
622.8
12.199
19.8
5.11
0.671
218 iO '
71.99
48.65
21
2080
21.87
52.08
13.81
23.05
11.99
. 712.0
220.0.
1119.7
28.17
7322.
319'. 9
18.67
115.5
15.19
1161. 11
679.8
. 909.7
1811.2
1589.5
75.00
28.56
37.22
16.21
1B8.'2
10.16
53.55
0.937
0.950
1575.
2.11
0.072
91.92
679.8
12.501
51.1
5.11
0.661
218.0
75.01
10.39
35
2080
22.15
52.71
13.88
23.26
11.99
712VO
220.0
1150.0
25.91
5919.
311.0-
17.81
115.2
15.17
1119.7
681.6
895.1
1831.3
1579.7
. 75.00
31.31
39.69
16.97
190.5
19.51
66.06
0.916
0.950
1718.
2.11
0.063
92.53
681.6
12.199
51.8
5.21
0.608
?18.6
71.97
11.33
16
11211
20.67
19.21
11.95
22.91
11.99
712.0
220.0
1150.2
30.51
11711.
302.1
16.53
111.3
11.27
1232.1
898.0
977.5
2130.1
.1875.5
75.00.
26.65
33.99
15.15
221.6
21.77
33.03
0.939
0.950
1231.
2.38
0.091
88.11
898.0
12.500
71.8
1.09
0.522
218.0
75.00
37.70
60
5080
21.23
50.56
13.72
22.10
11.99
712.0
220.0.
1150.1
28.18
9792.
325.7
20.18
117.0
17.02
1199.2
656.7
911.5
1855.9
1601.3
75.00
28.85
36.20
15.67
180.3
30.13
10.57
0.918
0.950
1371.
2.11
0.083
93.31
656.7
12.500
52. 5
5.25
0.721
218.0
75.01
10.83
70
2080
21.77
58.98
18.89
25.02
11.99
712.0
220.0
915.9
1.18
3860.
1303.0
8.81
106.1
6.12
1028.0
319.6
773.1.
1317.6
1093.0
75.00
159.69
296.11
83.57
112.9
70.19
93.58
0.902
0.950
2093.'
3.01
0.051
96.90
319.6
12.502
25.6
8.17
0.820
218.0
71.97
226.01
73
2081
22.72
51 . 12
15i95
23.39
11.99
712.0
220.0
1119.5
12.28
5129..
608.6
16.09
113.9
13.86
1120.5
175.7
865.9
1596.1
1311.5
75.00
66.23
125.52
16.11
161.6
56.52
75.36
0.876
0.950
1S71.
2.61
0.056
93.69
175.7
12.199
38.1
6.02
0.693
218.0
71.99
93.61
76
208]
?2.7b
51.26
15.17
?3.1t
11.99
712. U
220. C
1119.5
29.15
5032.
303.1
15.89
113.1
13.12
1117.5.
528.4
R62.9
1615.9
1391.3
75.00
27.61
31.72
15.. 06-
167.7
57.39
76.52
0.919
0.950
1886.
2.11
0.060
91.70
528.1
12.50]
12.3
6.3?
0.796
218.1}
75.02
39.01
78
2080
23.72
56.18
18.83
23.89
11.99
712.0
220.0
1119.0
10.51
1089.
623.1
9.11
108.0
8.05
1073.6
278.6
818.9
1352.2
1097.6
75.00
77.11
128.23
13.16
135.1
67.27
89.70
0.909
0 . 950
2051.
2.76
0.018
97.21
278.6
12.502
22.3
8.58
0.9P1
21P.O
75.05
109.02
90
12051
13.11
31.91
13.11
11.89
11.99
712.0
220.0
228.5
17.26
1518.
582.2
3.18
100.5
0.52
1898.6
0.0
1611.0
1898.6
1611.0
75.00
9.21
b.16
3.29
105.5
165.39
220.52
1.306
0.950
3251.
2.03
0.188
0.00
0.0
0.000
0.0
0.00
3.157
21P.O
75.27
13.21
107
2081
20.86
••19.67
13.17
21.78
11.99
712.0
220.0
1150;3
28.18
11163.
323.9
23.10
118.3
18.33
1220.7
670.0
966.1
1890.7
1636.0
75.00
28.85
35.56
15.37
179.9
26.99
35 . 9.*
0.953
0.950
1291.
2.12
0.089
92.36
670.0
12.501
53.6
5.11
0.711
218.0
75. T6
10.82
NOTES- (1) FROM ENGINE INLET TO EXHAUST ENTHALPY
C2) AVG ISENTROPIC EXPONENT, ENGINE IN TO EXH PRESS
(3) SO IN (COMBINED) PER 100 CYCLE HP
160
-------
Table 62. REFERENCE IDEAL CYCLES. MFSN-15, PYRIDINE-WATER
Mol ••'raction 0. 1 0.5 0.6 0.7 0.8 0.9
Pyridine
Efficiencies, %
Cycle, Regen
f of Carnot
Cycle, Nonregen
% of Carnot
Carnot
Temperature, Max F
Min F
Pressure, Max psla
Min psia
Per 100 Cycle HP
Fluid Rate, Ib/hr
Eng Exh. CFM
Pump In, GPM
Engine HP
Pump HP
Regen Cycle*
Regen KSTU/hr
Heater KBTU/hr
Conden KBTU/hr
Nonregen Cycle
Heater KBTU/hr
Conden KBTU/hr
Engine, Effic %
Pressure Ratio
Density Ratio
Exhaust Qual %
Intake Qual %
Delta H, BTU/lb -1511.92 -138.25 -125.90 -116.10 -106.HI -102.67
Pump, Effic % 100 100 100 100 100 100
Pressure Ratio 16.97 18.38 51.03 55-10 62.39 73.78
Density Ratio 1.0031 1.0020 1.0025 1.0030 1.0035 1.001
I-Factor at T min 1.663 1.1780 1.329 1.205 1.071 1.020
Regen 30* Cyc Eff
Effectiveness, % t>3 • 3b
Q, KBTU/CHP-Hr 98.50
UA, BTU/HP-Hr-F 206.3^
Del T UK), F I'll*
NTU(OG) 2.115
27
65
26
63
12
712
220
1,000
21
1,685
218
3
102
2
26
929
671
955
700
100
16
36
103
.38
.21
.61
.12
.00
.29
.5
.03
.768
.60
.60
.18
.25
.71
.13
.92
.97
.27
.88
27
66
26
62
12
712
220
1,000
20
1,891
261
i)
102
2
51
915
661
967
712
100
18
11
107
.80
.18
.31
.63
.00
.67
-36
.23
.261
.90
.90
.87
.62
.11
.19
.98
.38
.00
.93
28
67
26
62
12
712
220
1,000
19
2,085
281
1)
103
3
73
898
613
972
717
100
51
17
111
-33
.16
.18
.31
.00
.60
.11
.79
.722
-15
.15
.71
.26
.75
.00
.50
.03
.05
.16
28
68
26
62
12
712
220
1,000
18
2,265
305
5
103
3
93
879
621
972
718
100
55
55
111
.93
.90
.17
.30
.00
.05
.38
.81
.163
.31
.31
.17
.16
.95
.63
.12
.10
.72
.91
29
69
25
61
12
712
220
1,000
16
2,175
318
5
103
3
115
868
611
981
730
100
62
69
118
.30
.76
.85
.55
.00
.03
-53
.61
.673
.53
.53
.91
.61
.13
.55
.01
.39
.62
.87
30
7 3
26
61
12
712
220
1,000
13
2,566
393
5
103
3
118
828
571
916
692
100
73
93
120
.71
.12
.87
.00
.00
.55
.77
.31
.909
• 53
.53
.16
.70
.19
.86
.35
.78
.15
.58
"100$ Effective Regenerator
161
-------
Mol Fraction
3-Methylpyridlne
Efficiencies, %
Cycle, Regen
% of Carnot
Cycle, Nonregen
% of Carnot
Carnot
Temperature, Max F
Min F
Pressure, Max psia
Min psia
Per 100 Cycle HP
Fluid Rate, Ib/hr
Eng Exh. CFM
Pump In, GPM
Engine HP
Pump HP
Regen Cycle*
Regen KBTU/hr
Heater KBTU/hr
Conden KBTU/hr
Nonregen Cycle
Heater KBTU/hr
Conden KBTU/hr
Engine, Effic %
Pressure Ratio
Density Ratio
Exhaust Qual %
Intake Qual %
Delta H, BTU/lb
Pump, Effic %
Pressure Ratio
Density Ratio
I-Factor at T min
Table 63.
0.3
27.27
61.93
26.97
61.22
12.00
712
220
1,000
19.29
1560.0
263.65
3.101
102.10
2.10
10.32
933.25
678.70
913.57
689.02
100
51.81
39.07
101.52
-167.06
100
51.81
1.0006
1.719
0.3"
28. Kg
67.83
26.97
61.22
1)2.00
712
220
1,000
19.29
1560.0
263-65
3. 101
102. tO
2.10
50.26
893.31
617.76
913.57'
689.02
100
51.81
39.07
101.52
-167.06
100
51.81
1.0006
O.'l
27.35
65.12
26.21
62.10
12.00
712
220
1,000
18.71
1803.21
283.88
3.962
102.76 .
2.76 '
10,57
930.19
675-98
971.07
716.56
100
53-37
15.12
106.01
-115.01
100
53.37
1.0007
0.1«»
.-,29.17
6 9. '16
'28.21
62.10
12.00
712
220
1,000 •
18.71
1803.21
283.88
3.962-
102.76
2.76
98.69
872.37
617.86
971.07
716.56
100
53-37
15.12
106.01
-H5-01
100
53-37
1.0007
0.5
, -27.15
65.37 .
25.55
60.83
12.00
712
220 .
1,000
17.60 '
2010.7
309-71
1.530
103.11
3.11.
69.13
927.03
672.51
996.16 :
711.65
100
56.81
53.22
110.28
-121.72
-100
56.81
1.0009
0.5"
:29.12
69.33
r25.55
60.83
12.00
' 712'.
•220
1,000
17.60
2010.7
309.71
1.530
103.11
3.11
122.09
871.08
619.56
• 996.16 :
711.65
' 100
56.81
53.22
110.28
-121.72
100 .
56.31 '
1.0009
0.6
27.68
- 65V92
25.09
• 59.71
12.00
712
220
1,000
15.89
2266.15
315.63
5.082
103.13
3.13
95.18
919.18
661.68
1011.37
759.86
100
62.93
65.36
111.32
-112.31
100
62.93
1.0012
1.719
1.531
1.531
1.350
1.350
1.205
0.6»»
•' 30.09
• •71.65
25.09
59.71
12.00
712
220
1,000
15.89
2266.15
315.63
5.082
103.12
3.13
168.65
817.11
591.21
1011.37
759.86
100
62.93
65-36
111.32
-112.31
IOC
6:. 93
1.0012
1.205
•100? Effective Regenerator
•"Condensation in Regenerator
-------
Table 64. REFERENCE IDEAL CYCLES. 2-METHYLPYRIDINE. WATER MIXTURE. MFSN-23
Mol Fraction
2-Methylpyridine
Efficiencies, %
Cy^ie, Reger.
t of Carnot
Cyco.e, Nonregen
% of Carnot
Carnot
Temperature, Max F
Min F
Pressure, Max psla
Mln psia
Per 100 Cycle HP
Fluid Rate, Ib/hr
Eng Exh. CFM
Pump In, GPM
Engine HP
Pump HP
Regen Cycle*
Regen KBTU/hr
Heater KBTU/hr
Conden KBTU/hr
Nonregen Cycle
Heater KBTU/hr
Conden KBTU/hr
Engine, Effic %
Pressure Ratio
Density Ratio
Exhaust Qual %
Intake Qual %
Delta H, BTU/lb
Pump, Effic %
Pressure Ratio
Density Ratio
I-Factor at T mln
Regen 30? Cyc Eff
Effectiveness, %
Q, KDTU/CH?-Hr
UA, BTU/HP-Hr-F
Del T CLM), F
:;TU(OG) -
0.?
NA
26.97
64.21
42.00
712
220
893.37
22.76
1375-51
220.51
3-04
101.79
1.79
NA
943.7
650.94
100
39.255
26.327
100
-373.23
100
39.255
1.0002 .
0.3
27.93
66.50
26.85
63.92
42.00
712
220
1000.
22.76
1619.85
228.44
3.609
102.31
2.31
35-356
911.26
657.47
946.62
692.11
100
43.93
33.78
105.53
-160.75
100
43-93
1.0011
0.4
28.06
66.81
25.88
61.61
42.00
712
220
1000.
22.70
1969.71
242.50
4.291
102.75
2.75
76.53
907-05
6?2.54
982.58
729.77
100
44.05
38.00
111.72
-132.76
100
44.05
1.0013
0.5
28.07
66.83
25-01
59.54
42.00
712
220
1000.
22.38
2170.3
254.72
4.982
103-17
3.17
111.09
906.72
652.21
1017.81
763.30
100
44.69
42.82
117.03
^120.99
100
44.69
1.0017
0.6
27.91
66.45
24 .09
57.36
42.00
712
220
1000.
21.59
2456.2
271.42
5.721
103-59
3.59
144.58
911.86
657.35
1056.44
801.93
100
46.32
49-39
121.99
-107.34
100
46.32
1.0021
1.985
NA
NA
NA
NA
NA
NA
1.480
;>100
NA
NA
NA
NA
1.190
>100
NA
NA
NA
NA
1.0046
>100
NA
NA
NA
NA
0.8854
>100
NA
NA
NA
NA
27.81
66.22
23.36
55.61
42.00
712
220
1000.
20.17
2742.6
295-21
6.468
103.95
3-95
174.56
915.12
660.01
1089.67
835.17
100
49.59
59-79
126.42
-96.47
100
49.59
1.0027
0.7857
>100
•NA
NA
NA
NA
•naoj Effective Regenerator
163
-------
OY
-Cr
Table 65.
Mol Fraction
4-Methylpyridine
Efficiencies, %
Cycle, Regen
% of Carnot
Cycle, Nonregen
% of Carnot
Carnot
Temperature, Max F
Min F
Pressure, Max psia
Min psis
Per 100 Cycle HP
Fluid Rate, Ib/hr
Eng Exh. CFM
Pump In, GPM
Engine HP
Pump HP
Regen Cycle*
Regen KBTU/hr
Heater KBTU/hr
Conden KBTU/hr
Nonregen Cycle
Heater KBTU/hr
Conden KBTU/hr
Engine, Effic %
Pressure Ratio
Density Ratio
Exhaust Qual %
Intake Qual %
Delta H, BTU/lb
Pump, Effic %
Pressure Ratio
Density Ratio
I-Factor at T mln
REFERENCE IDEAL CYCLES. 4-METHYLPYRIDINE, WATER MIXTURE. MFSN-24
0.3 0.4 0.-5 ''0.6 ' :0.7
27.55
65.60
27.^6
65.38
42.00
712
220
1,000
18.86
28.06
66.81 .
27.13
64.60
. •' 42 .00
712
220
1,000
17.81
28.37
67.54
. 26.68
. ." 63.53
42.00 .
712
220
•-•1,000
16.26-
28.62
68.14
26. .23 .
62.45
42.00 •
712
220
1,000
" 14.33 •
28.61
68.13
25.85
61.55
•• - ;42.00;
712
220
1,000
12.16
-165:45
100 '
53.03
1.0012
1.728
-143.83
'100 ''.
•'5.6.15
- 1.0014.
1.510
-127.53
100
61.5'-- '
•'1.-.0017''
1.342
-114.79
100
:-69'.!79.
.;' .. 1,028
-104.09
100
"' 82.3
'. 1.0027
1.091
0.8
28.52
67-91
- -25.55
60.83
42.00
712
220
967.73
. 9-653
1,577.81
269.16
3.535
102.57
2.57
3.11
923.71
669.20
926.82
672.31
100
53.03
40.;09
100. -46
1,818.77
295.84
4.116
102.78
2.78
31.04
907-00
652.49
.938.04
683.54
100
56.15
.48.01
• 104.76
2,055.59
330.09
4. 7P 3
103.01
3.01 :
56.67
897.23
642.73;v
953-91
699.40
100
61.50 '
59.09
••' 108.82-
2,289.28
376.60
5.289
1 103.26
3.26
79.06
889.35
636.73
968.41
715:79
100
,- 69.79
- 76.36
.112:42
2,531.5
444.04
5.897
103.53
'.' , 3-53
95.00
889.48
634.97
- 984.48
729.97
100
- 82 .-30
-108.13
. 114.96 .
2,799.2
552.58
6.563
103.74
3.74
103.92
892.32
63-7.81
996.23
741.73
100
100.24
170.38
116.29
-94.32
100
100.24
1.0031
1.001
*100? Effective Regenerator
-------
Table 66. SPECIAL IDEAL CYCLES. 2-METHYLPYRIDINE. WATER MIXTURE. MFSN-23
Mol Fraction
2-Methylpyrldine
Efficiencies, %
Cycle, Hegen
% of Carnot
Cycle, Nonregen
% of Carnot
Carnot
Temperature, Max F
Min F
Pressure, Max psia
Min psia
Per 100 Cycle HP
Fluid Rate, Ib/hr
Eng Exh. CFM
Pump In, GPM
Engine HP
Pump HP
Regen Cycle*
Regen KBTU/hr
Heater KBTU/hr
Conden KBTU/hr
Nonregen Cycle
Heater XBTU/hr
Conden KBTU/hr
Engine, Effic %
Pressure Ratio
Density Ratio
Exhaust Qual %
Intake Qual %
Delta H, BTU/lb
Fump, Effic %
Pressure Ratio
Density Ratio
.[. • :-i,iu--.:or at T mln
Ke«on 30!i Cyc Eff
. •"•• f '.•/.-.ness ; %
i A, BTU/HP-Hr-F
Del T (LM), P
NTU(OG)
0.1
(a)
0.1
(b)
0.1
(c)
1.0
(d)
0.0
(e)
NA
NA
26.99
61.28
12.00
712
220
1,781.9
22.70
2,001.97
220.65
1.525
105.23
5.23
NA
NA
NA
912.70
688.19
100
78.62
82.06
100
-133.78
100
78.62
1.0023
1.190
NA
NA
NA
NA
NA
30.18
66.20
26.19
57-53
16.01
800
220
1,000
22.70
1,688.12
239.62
3.816
102.11
2.11
126.01
831.83
580.32
960.81
706.33
100
11.05
36.63
121.71
-151.12
100
11.05
1.0013
1.190
89.26
112.18
1,926.1
58.38
2.851
30.05
66.36
26.11
58.33
15.28
782.5
220
1,000
22.70
1,721.08
239.91
3.897
102.50
2.50
116.59
817.08
592.58
963.68
709.17
100
11.05
36.81
119-67
-151.31
100
11.05
1.0013
1.190
100
116.59
03
0
00
29.33
69-83
23.33
55.56
12.00
712
220
1,000
10.92
3,197.0
537.3
8.156
101.50
1.50
221.71
867.80
611.18
1,090.69
836.18
100
91.61
191.21
137.20
-76.05
100
91.61
1.0011
0.588
>100
NA
NA
19.02
15.29
12.00
712
220
3J2.7
17-28
1,123.1
119 .9
2.616
100.21
0.239
NA
NA
NA
1,338.0
1,083-5
100
9.991
5-857
100
-227.2
100
5.857
1.0003
3.192
NA
NA
NA
NA
NA
* 100? effective regenerator
(a,1 1000 psi pressure maximum restraint omitted
(b) 800°F temperature allowed In gas generator
(c) Upper temperature adjusted for 30? cycle efficiency
(d) Pure component, 2-methylpyrldlne
(e) Pure component, water
165
-------
Table 67. SPECIAL IDEAL CYCLES, 4-METHYLPYRIDINE. WATER
Cycle Description 0.3**
and m.f. Organic'
Efficiencies, %
Cycle, Regen 2.8.89
% of Car not. 68.78
Cycle, Nonregen 27.46
% of Carnot 65.38
Carnot 42.00
Temperature, Max'? 712
Min F 220
Pressure, Max psia 1,000
Min psia 18.86
0.3+
0.6**
26.95
64.17
26.03
•61.97
42.00
712
220
900
18.86
32.06
76.32
26.23
62.45
42.00
712
220
1,000
14.33
Per 100 Cycle HP, •
Fluid Rate, ,lb/hr
Eng Exh. CFM .
Pump In, GPM
Engine HP
Pump HP
Regen Cycle* '. •
Regen KBTU/hr
Heater KBTU/hr
Conden KBTU/hr
Nonregen Cycle
Heater KBTU/hr
Conden KBTU/hr
Engine, Effic %
Pressure Ratio
Density Ratio
Exhaust Qual %
Intake Qual %
Delta H, BTU/lb
1,577.81
269 . 16
3.535
102.57
2.57
82.53
844,29
589.78
926.82
672.31
100
53.03
40.09
100.46
— —
-165.45
1,626.54
300.19
3.644
101.84
: 1.84
33.51
944.38
689 . 87
977.90
• 723.39
100
37.12
28.04
104.85
__
-159.35
2,289.3
376.6
5.289
103.26
3.26
176.35
792.05
539.43
968.41
715.79
10.0
69.79
76.36
112.42
_—
-114.79
Pump, .Effic %
Pressure Ratio
Density Ratio
I-Factor at T min
100
53.03
1.0012
1.728
100
37.12
1.0008
1.728
100
69.79
1.028
1.203
* 100$ effective regenerator
** Condensation in regenerator
=}= Lower pressure in boiler
-------
23 p-difluorobenzene
24 m-difluorobenzene
35 fluorobenzene
46 hexafluorobenzene
60 1,2 ,3 ,5-tetrafluorobenzene
70 3-meth'ylpyridine
73 toluene
76 benzene
78 pyridine
90 water
107 pentafluoropyridlne
Reference ideal cycles calculated for many of the Table 60-67 can-
didates are extracted in Table 68. The first column is a test of
Criterion 6, which requires an ideal cycle efficiency of at least
30$ at any level of regenerator effectiveness. All straight
organic fluids pass, whereas the pyridine-water mixtures fail.
While water itself appears disqualified, it is capable of higher
efficiencies by raising its upper operating temperature.
Other entries in the table demonstrate compliance with certain .
interim (non-criterion) goals; namely,
1. condenser pressure from 5 to 50 psia
2. ideal expansion density ratio less than about 40
(to serve in a single expansion reciprocator)
3. ideal expansion enthalpy drop less than about 200
Btu/pound (to serve in a single-stage impulse
turbine)
4. I-factor in the approximate range of 0.65 to °°
(components) or 0.75 to 1.5 (final fluids)
Equivalent Real Cycle calculations employing many of the same
simple fluids recorded in Table 68 are extracted in Table 69.
Comparison of the entries with the reciprocating/turbine suitabil-
ity Criteria 8 and 9 reveals all fluoroaromatics and benzene suit-
able for either engine type and all aromatics but benzene suited
only for turbine use.
6.10.2 I-Factor as a Determinant of Cycle Efficiency
The data of Tables 60 and 6l permit an analysis of the efficacy
of I-factor as a screening criterion. Referring first to the ref-
erence ideal cycle data of Table 60, in Figure 33 is plotted the
regenerator UAk product (linked to regenerator size in section
167
-------
Table 68. SUMMARY OP REFERENCE IDEAL CYCLE CALCULATIONS
ON THERMALLY STABLE CANDIDATES
J
Ideal
Cycle
Effic.C*)*
Aromatics
Benzene
Toluene
Pyridine
3-Methylpyridine
Fluoro Aromatics
Fluorobenzene
m-Dif luorobenzene . '
p-DIf luorobenzene
1.2,3, 5-Tetraf luoro-
benzene
1.2,1, 5-Tetraf luoro-
benzene
Pentaf luorobenzenei
Octafluorotoluene
Pentaf luoropyridine
Hexaf luorobenzene
Water
Pyridine/Water Solutions
lOSK \ (
cna i Pyridine >
DU?b " *
30% \ I
fio- f 2-Methylpyridine \
30% > I
6Q% f 3-Methylpyridine \.
30% » /
60% * v
31.8
31.7
31.5
32.1
31.9
31.6 .
31.7
30.6
30.8
30.3
30.2
30.3
31.1
20.5
.27.1
28.3
27.9.
27.9
27-3
27.7
•27-. 6
28.6
Condenser
Pressure
(psia)
29.1
12.3
10.5
1.2
25-9
28.5
23.6
28.2
22.8
26.6
11.9
28.2
30.5
17-3
21.3
19.6
22.8
21.6
19-3
15.9
. 13-9
11.3
Expansion
Density
Ratio
10.3
115-9
115.1
338.1
16.6
13-7
51.6
12.1
51.7-
15.2
86.9.
11.6
10.3
5.8
36.3
17.1
.33.8
19.1
39.1
65.1
10.1
76.1
Expansion
Enthalpy
Drop
(Btu/lb)
. 81.3
79-3
93-9
97.3
70.1
56.9
59.1
13-0
11.8
35.3
29.7
38.2
35.2
226.0
151.9
125-9
160.8
107.3
167-1
112.3
165-5
111.8
I-Factor
at
220°F
0.85
0.75
0.98
0.89
0.65
0.70
0.71
0.76
0.77
0.68
0.55
0.75
0.51
3-31
1.66
1.33
1.18
0.89
1.75
1.21
1-73
1.20
* Full regeneration on all simple fluids; full regeneration or regen-
eration only'to" the dew line, whichever occurs first, on all solutions.
** Concentrations shown are, mol -percent.s.
168
-------
Table 69. SUMMARY OF EQUIVALENT REAL CYCLE CALCULATIONS
ON SIMPLE THERMALLY STABLE FLUIDS1
Aromatics
benzene
toluene
pyridlne
3-methyIpyridine
Fluoro Aromatics
fluorobenzene
m-dlfluorobenzene
p-dlfluorobenzene
1>2,3,5-tetrafluorobenzene
pentafluoropyridine
hexafluorobenzene
Real
Cycle
Effic. (2)2
22.8
22.7
23.7
ine 24.8
Expansion
Density
Ratio3
15.1
46.4
43.5
83.6
Enthalpy
Drop (Btu/lb)
57.4
56.5
67-3
70.2
22.1
21.9
22.3
21.2
20.9
20.7
17.0
16.2
19.5
15.7
15.4
15.2
49.5
40.2
41.8
30.4
27.0
24.8
1 Condenser pressures and 220°F. I-factors same as in Table 68.
2 Regenerator Q/AT(log mean) = UAk = 125 Btu/HP-hr-°F.
3 Isomtroplc, from engine inlet to exhaust enthalpy.
169
-------
16,000
14,000
12,000
7 10,000
i_
f 8,000
Q-
X
6,000
3 4,000
CO
Z 2,000
0
46
0.5
107
o
o60
Cycle Efficiency = 30%
Numerals are candidate
FSN numbers
* Suitable only for
turbine expander
70*o
0.6
0.7
I - at 220 F
0.8
0.9
78"
1.0
Figure 33- Reference Ideal Cycle Regenerator "Size" (UAk)
as Affected by I-Factor
-------
6.1.2.3) versus I-factor for nine simple fluids. This graph
clearly reveals the expected trend to larger regenerator sizes
with diminishing I-f actor.. It also reveals a fairly wide scatter,
which argues against .setting too narrow limits on an I-factor cri-
terion .
As discussed at some length in sections 6.1.2.3-5, a. regenerator
sized to give a particular ide.al cycle efficiency is likely to be
far too small for the real cycle because of the much higher heat
loads in the real system associated with the need to circulate
considerably more fluid per net cycle horsepower. This argument
lead to the concept of the -equivalent real cycle. Using this
cycle it is possible to show, with reasonable relation to reality,
the relationship between cycle efficiency and I-factor when the
regenerator "size" is held constant at its ceiling value.
The relationship is demonstrated in Figure 3^ for the same nine
fluids. Data were from Table 61. The general downward trend of
cycle efficiency with decreasing I-factor is apparent despite a
fairly wide scatter. This graph clearly presents the argument for
a high I-factor fluid in the automotive application, where regen-
erator space is at a premium. It also buttresses the initial
selection of I-rfactor criterion limits of 0.65 to 1.5.
Another presentation tending to support the validity of the I-
factor screen is given in Figure 35. Here are plotted ideal cycle
efficiencies versus I-factor for a large number of candidates,
both simple and complex. Three different families of points
resulted from the three different degrees of regeneration assumed,
namely :
100$ effective (curve 1)
regeneration of vapor only to saturation (dewpoint)
0% effective
The gap between the 0 and 100% effectiveness curves is a qualita-
tive indication of the size regenerator needed to achieve reason-
able cycle efficiency. Data were from Tables 60 through 67.
An additional point of interest in Figure 35 is the fall-off in
cycle efficiency associated with I-factors above 1.0. All the
fluids in this range contain water. The reason for this fall-off
is traceable to the fact that water, compared to the organics,
has an unusually high critical pressure. At the 1000 psia maximum
allowable fluid pressure, water boils at the unusually low temper-
ature of 5^5°F. A large amount of the heat added to the water to
bring it to the 712°F (167 degrees of superheat) peak cycle tem-
perature is therefore added at temperatures far below 712°F.
Besides leading to a lower cycle efficiency, this explains one of
the reasons for the current interest in organic working fluids.
Dilution of water with soluble organic fluids tends to improve
171
-------
25
24
23
—
'o
Q>
o 22
21
20
,70*
UAk = 12,500 BTU/ lOOHP-Hr-F
Numerals are candidate FSM numbers
•Suitable only for turbine expander
0.5
• i
0.6
0.7 0.8
I - Factor at 220°F
0.9
78*c
1.0
Figure 3'4. Equivalent Real Cycle Efficiency as Affected
by I-Factor at-Fixed :Regeher'ator-Size : '
172
-------
36
o Pure Fluids, Std.
Pure Fluids, Non-Reqen
2-MePyr., H20, Std."
2-MePyr., H20, NR
3-MePyr., H20, Std.
H20, NR
H20. Cond.
a 4-Me Pyr., H20. Std.
4-Me Pyr. , H20, NR
4-Me Pyr., H20, Cond.
Pyridine, H30, Std.
Pyridine, H20, NR
Fluorinol
Fluorinol
Pure Fluids with Regenerator and
^ Mixtures with Condensing Regenerators
Aqueous Pyridine and Methyl Pyridines
^ with Single Phase Regenerators
3) All Fluids without Regenerators
'220
Figure 35. Correlation of Cycle Efficiencies of Candidate
and Reference Fluids versus I-Factor
-------
this situation by reducing the apparent critical pressure and the
amount of superheat required.
6.11 SPECIAL STABILITY TEST WORK
Several limited thermochemical stability studies were made in
areas slightly aside from the central objectives of the contract.
6.11.1 Fluorinol-85 Stability Investigations
Fluorinol-85 is an 85 mol percent mixture of 2 ,2 ,2-trifluoro-
ethanol with 15 mol percent water manufactured by Halocarbons
Incorporated. It is currently the design fluid for the automotive
Rankine engine under development by Thermo Electron Corporation.
Their maximum bulk fluid temperature is presently set at 550°F.
The question inevitably arose as to the possibility of operating
Fluorinol, along with its lubricant (a refrigeration oil) at a
higher temperature. Accordingly, a series of ampoule tests were'
carried out on Fluorinol-85 at 600°F. As in the Criterion 5 am-
poule tests, no attempt was made to remove freeboard air from the
ampoules . Samples were run both in stainless steel and in the
standard SAE 1008 low carbon steel ampoules. Both survived 336
hours at 600°F with no evidence of degradation aside from devel-
oping a straw yellow color. No strong acidity (HP) formed nor
were significant changes in composition found, as evidenced by a
gas-liquid chromatogram. These tests were repeated with various
amounts (2%, 12$ and 25%} of lubricating oil present. Again the
appearance after 336 hours was quite good and it was concluded
that Fluorinol-85 systems could probably be successfully operated
at a 600°F boiler temperature provided hot-spot problems were min-
imal. In one experiment, a sample of Fluorinol-85 in a stainless
steel ampoule was held at 600°F for 740 hours. At this point the
fluid was still almost colorless but a test with moisture indicator
paper suggested the presence of strong acid (presumably HP in small
amounts).
Inasmuch as results at 600°F were encouraging, a series of
Fluorinol-85 samples both with and without refrigeration oil were
run at 660°F. The results indicated that Fluorinol-85/oil combi-
nations were every bit as stable as Fluorinol-85 without oil.
Samples containing no oil developed an unpleasant odor after 24
hours exposure and after 48 hours showed traces of weak acid as
evidenced by the decolorization of aqueous bromophenol blue indi-
cator solution. The quantity of weakly acidic material did not
appreciably increase with time, and no strong acid could be
detected up to 336 hours exposure. The use of stainless rather
than mild steel tubes did not significantly change these results.
An identical series of samples containing 15% refrigeration oil
did not develop odor nor give evidence of weak acid until about
144 hours. Again the weak acid did not increase significantly
with time and no strong acid (HF) was detected at 336 hours. At
174
-------
500 hours, however, some gas pressure remained in the cooled
sample tubes and traces of strong acid were in evidence. The oil
layer appeared light-colored and in reasonably good condition.
Several modified versions of the basic refrigeration oil were also
tested along with Fluorinol-85. Results indicated that all were
about the same. This was not unexpected since all were hydrocar-
bon products of similar composition.
Some additional experiments were carried out to determine the
effect of inerting the sample with nitrogen before thermal expo-
sure. As discussed earlier, standard procedure had previously
been to seal the Fluorinol-85 in the pressure tube without exclud-
ing ambient air. In connection with this investiga i;ion a simple
procedure for detecting the weak acid formed was adopted. First,
an indicator solution was prepared'which consisted of 6 drops
0.04% bromophenol blue solution (Fisher Scientific Cat. No. 5-
985-F) in 20 ml of distilled water. Under neutral conditions its
color is light blue. Weak acid in exposed Fluorinol-85 samples
is then measured by dropwlse addition of test sample from a 0.25
ml graduated syringe to 2.0 ml of the indicator solution. The
volume of sample required to discharge the blue color of the indi-
cator provides a relative measure of the amount of weak acid pres-
ent. Fresh Fluorinol-85 does not cause the blue color to change,
thus the less sample required to cause the indicator to turn from
blue to pale yellow, the more weak acid is present. The method
is, however, somewhat imprecise since the acid being detected is
quite weak. This causes the indicator color to fade over a range
rather than change abruptly as it would under the influence of
strong acid.
Results obtained on exposing Fluorinol-85 samples under nitrogen
and with ambient air are summarized as follows:
.'vnrrole Suniso Exposure Exposed Fluorinol-85 Required
SiaAeo 30S Oil Time at to Discharge Color in 2.0 ml
.•: ..- - Present 660°F (hr) Indicator Solution (ml)
Air No 72 ^0.1.
N2 Mo 72 ^0.2
Air Yes 168 ^0.05
N2 Yes 168 ^0.05
Air Yes 336 ^0.05
N2 Yes . 336 ^0.05
From these experiments it was concluded that at short exposure
times nitrogen-inerting may be somewhat beneficial, but after
long exposure in the presence of lubricant no significant advan-
tage accrues from nitrogen-blanketing. It was also observed that
samples at 168 hours or longer were under slight pressure when
brought back to room temperature. This indicates the presence of
175
-------
some gaseous decomposition products, possibly of hydrogen formed
by reaction of water with the iron ampoule surface. The identity
of this gas was not established.
Because of the unexpectedly good stability of Fluorinol-85 and
oil at 660°P additional experiments were carried out at 720°F.
Results conformed closely with earlier observations at this tem-
perature (see MFSN-20, Table 47). After 24 hours at ?20°F a
Fluorinol-85 + 15% refrigerant oil sample showed evidence of a
small amount of gas but no acids, either weak or strong. Both
the Fluorinol-85 and oil layers were relatively clean and showed
little color change. After 48 hours, however, extensive decompo-
sition was apparent. A heavy, black, dusty deposit had formed,
and large amounts of strong acid were detectable. One may con-
clude from these results that Fluorinol-85 and hydrocarbon oils
might function satisfactorily at ^660°F but have a fairly sharp
break point around 700°F.
6.11.2 Nonvolatiles from Pyridine-Water Type Fluids
As thermal stability appeared to be the principal weakness of
pyridine and methylpyridine-water fluids an effort was made to
better judge the extent and seriousness of decomposition as ob-
served in ampoule tests. This presents an experimental problem,
since these fluids do not give an easily measured entity, such as
formation of acid, by means of which the decomposition can be
followed. Ultimately, the formation of nonvolatile tars was
found to be a convenient means of following the decomposition
process. Results of a series of experiments are summarized in
Table 70.
The table shows that the results are not entirely consistent.
In any event, formation of nonvolatiles remains fairly low in each
of the fluids and shows no dramatic increase over the time inter-
val studied. The appearance of the residues differed substan-
tially. Thus the residues from pyridine and 4-methylpyridine-
water solutions appeared as dry, finely divided brown solid
whereas residue from the other methylpyridine solutions appeared
as a semi-liquid tar.
6.11.3 Carboranes
Interest in the recently discovered class of chemical compounds
called closo-carboranes (ref. 55) stems from the recognition that
polyhedral carbon-boron "cage" molecules are relatively stable
thermodynarnically, some extremely so. While these compounds are
currently laboratory curiosities, and, therefore, "unavailable"
in a practical sense, there is some reason to believe they could
be manufactured in large quantities at reasonable prices if a
demand developed. Through the generosity of Professor M. F. Haw-
thorne, University of California, Los Angeles, small samples of
176
-------
Table 70. FORMATION OP NONVOLATILE TARS IN METHYLPYRIDINE-
WATER FLUIDS AFTER THERMAL EXPOSURE AT 720°F
Mixed
Fluid
No.
M-15
M-21
M-23
M-24
Composition
0.5-0.5 mol
Fraction
Pyrl dine /water
3-Methylpyridine/
water
2-Me thy Ipyri dine/
water
4-Me thy Ipyri dine/
water
Nonvolatile Materials* Produced (wt$)
at 720°F after Various Exposure Times
4
0
0
0
0
8 hr
.013
.0
.0
.125
96 hr
0
0
0
0
.013
.080
.0
.033
168 hr
0
0
0
0
.044
.018
.016
.051
240 hr
0
0
0
0
.054
.148
.049
• 545
264 hr
0
0
0
0
.00
.475
.121
.286
*Procedure:
A carefully weighed sample (approximately 1 g) of
exposed fluid evaporated for 16 hours in an oven
at 149°F. The residue remaining is reported as
nonvolatile.
three substituted closo-carboranes were obtained for evaluation
The three were (ref. 55):
A. C,C'-dimethyl-2,4-C2B5H5, or (CH3)2B5C2H5
melting point
boiling point
cage geometry
references
sample volume
B. C,C'-dimethyl-l,7-C2B6H6, or (CH3)2B6C2H6
melting point -40°C
boiling point 63°C % 134 torr
cage geometry dodecahedron
references 58, 59
sample volume 2 ml
-29°C
9
pentagonal bipyramid
56, 57
1 ml
177
-------
• C. C,C.I-.dimethyl-l,7-C2B7H7, or
melting point -22°C
. boiling point ?
cage geometry tricapped trigonal prism
references 58, 59
sample volume 3 ml
While all three of these are known to be very thermally stable,
their ability to resis't ignition and burning is no.t. well estab-
lished. Flammability is thought to be the key issue in consider-
ing the carboranes as candidates, but the available samples fell
far short of the amounts required for even flash and fire point
measurement .
For these reasons, it was decided to commit the limited samples
to microliter-scale tests of oxidation stability in side-by-side
comparison with reference fluids of known flammability.
6.11.3.1 Oxygen Consumption Measurements - Method (ref. 60) -
Thin wall borosilicate glass melting point tubes (Size D, 1 mm ID)
were cut to a length of ^7 mm. After the tubes were quickly
cooled in dry ice, 1 pi of sample was injected into each tube
with a 5 yl syringe. The tubes were sealed by fusing the open
end with a natural gas-oxygen flame; the same amount of air was
trapped in each tube. Control tubes . containing air, but no sam-
ple, were prepared in the same manner.
All but two of the sealed tubes were placed in an electric fur-
nace at 178°C (352°F). Duplicate tubes were removed from the fur-
nace at intervals of 10, 30, and 60 minutes- for gas chromato-
graphic analyses of the residual 02/N2 ratio. Two sets of con-
trols - (a) air with no sample and (b) air with sample, but not
heated - were also analyzed.
The residual gases in the tubes were released into the injection
port of a gas chromatograph by crushing the capsules with a solid
sample injector. Analysis conditions were:
Instrument - F&M Model 700
Column - 9 ft x 1/4 in. stainless steel tubing packed
with 100-120 mesh Molecular Sieve 5A
Column Oven Temperature - 110-120°C
Injection Port Temperature - 50°C
TC Detector Temperature - 180°C
TC Detector Current - 250 mA
178
-------
Carrier Gas and Flow -He, 50 psig, 120 ml/min
Attenuation - X2 and X8
Ratios of the peak heights for 02 and N2 peaks in the gas chromat-
ogram of the residual gases from the test samples were compared to
those for ambient air. The calculations are as follows:
% 02 in Re- N2 peak height
=
sidual Gas 02 peak
02 peak height
.
~ x 21'°* (02 in
N2 peak height
% 02 consumed = 21.0% - % 02 in residual gas
% 02 consumed
Fraction of 02 consumed = - ^ — pr^ -
c. J. • U fo
6.11.3-2 Oxygen Consumption Measurements - Results - Data for
the fraction of oxygen consumed as a function of time at 178°C
(352°F) are reported in Table 71 and Figure 36 for the three car-
boranes and benzene. All analyses were done in duplicate.
As judged by the consumption' of available oxygen, the oxidative
stability in air at 178°C (352°F) is progressively poorer in the
sequence
(CH3)2B7C2H7 > (CH3)2B6C2H6 > . (CH3 ) 2B5C2H5
All three of the carboranes are, however, far less resistant to
oxidation than is benzene, which consumed no oxygen.
6.11.3-3 Ignition Delay Time - Method - The shock tube used to
measure ignition delay (or induction) Fimes is 3-9 cm in diameter
and 3.6 meters long, equally divided lengthwise into reaction and
driver sections. Incident shock speeds are measured by timing
the passage of the initial shock wave between a pair of SLM model
603 pressure transducers mounted 37.5 cm apart near the downstream
end of the reaction section. Gas temperatures and flow velocities
are calculated from incident shock speeds .
Pressure is detected and recorded using a third SLM transducer 5
cm from the downstream end of the reaction section. Light emitted
upon ignition is detected by a photomultiplier tube "looking"
laterally into the shock tube through a slitted quartz window at
the same axial position. Ignition delay time is taken as the
interval between the detection of the pressure rise due to the
reflected shock wave and the first appearance (10% of peak) of
emitted light of combustion.
179
-------
Table 71. COMPARISON OP FRACTION OF OXYGEN CONSUMED
BY THREE CARBORANES AND BENZENE HELD AT
. 178°C (352°F) IN AIR FOR VARIOUS TIMES
Time Fraction of
Sample (min.) Oxygen Consumed
Room air 0 0
10 0
30 0
60 0
(CH3)2B5C2H5 0 0
10 0.28
30 0.19
70 0.59
111 0.62
(CH3)2B6C2H6 0 0
10 0.20
30 0.44
60 . 0.55 '
(CH3)2B7C2H7 0 0
10 0.17
30 0.28
60 0.41
Benzene 0 0
10 0
30 0
60 0 .
180
-------
o (CH3)2B5C2H5
(CH3)2B6C2H6
-------
The "air" originally confined in the reaction section was actually
a mixture of 21% oxygen and 79% argon. Into this mixture was
introduced 10 microliters of the test fluid, which was then
allowed to evaporate and diffuse through the mixture. The driver
gas was a mixture containing 90% helium and 10$ argon. By varying
the initial pressures in the two halves of the shock tube, the
temperature of the highly compressed "air" after shock reflection
is controlled. The shock compression lasted about 5 milliseconds
at a pressure near 60 psia under the conditions of the tests.
Skinner and Ruehrwein (ref. 6l) give the general techniques in
greater detail.
6.11.3-4 Ignition Delay Time - Results - One carborane,
(CH3)2B7C2H7(theone found most resistant to oxidation in section
6.11.3.2), was chosen for.comparison against two reference fluids;
benzene and 1,3,5-trifluorobenzene. These were selected as repre-
senting materials "too flammable" and "marginally flammable" for
the application.
Using the methods of the previous section, the results of Table 72
were obtained.
Table 72. IGNITION DELAY DATA
Sample
Benzene,
1,3,5-Trifluoro
benzene, C6H3F
Shock Temp
830
848
890
898
1165
1165
1288
1367
1392
1367
1397
1402
1559
Ignition Delay
Time (msec)
0.90
1.78
0.60
0.22
No
ignition
The data of this table are plotted in traditional Arrhenius form
in Figure 37 where the least-mean-square fitted curves correspond
to the relation
T = A exp(B/T).
(22)
182
-------
10.0 r
1.0
0.1
0.01
• C6H6 (benzene)
* (CH3)2B7C2H7 (carborane)
0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4
1000
Inverse Absolute Temperature
Figure 37. Ignition Delay Times - Carborane and Reference Fluid
183
-------
This figure clearly shows that the ignition delay time of the car-
bor.ane is about 1/300 the delay time of benzene at any particular
temp;era;ture. At a particular delay time, the corresponding igni-
tion temperature of the carborane is about 350°K lower than ben-
zene. Either way, it is clear that the carborane is much more
readily ignited in "air" than is benzene. Benzene in turn is
much more readily ignited than trifluorobenzene. Under the test
conditions, the trifluorobenzene never did'ignite within the 5
millisecond span of the compression.
6.11.4 Inorganic Fluids
Many covalent inorganic fluids exist which have boiling points in
the range desired for the present application. Examples of these
are listed in Table 73. It is tempting to seek useful fluids
within this class of materials. Closer investigation, however,
reveals that without exception fluids of the types shown are so
extremely reactive (corrosive) and/or noxious that they cannot be
considered in any application involving contact(s) with materials
ordinarily used in automotive construction. To prove this point,
24-hour ampoule tests (at 720PP) were run on phosphorus oxychlo-
ride (POCls) and germanium tetrachloride (.Ge.Cljt). As expected,
the fluids survived the thermal stress easily but the SAE 1008
steel ampoules were very extensively corroded inside. Despite
extensive literature search and consultation with outside author-
ities, no inorganic fluids of any promise could be identified.
184
-------
Table 73. SOME UNCONVENTIONAL INORGANIC FLUIDS
Melting Point Boiling Point
Material (°C) (°C)
AsCl3
CSe2
CSSe
Cr02Cl2
0=P(OH)F2
GeH2Br2
GeH2Cl2
GeCl^
Fe(CO)5
PC13
POBr.Cl2
POC13
PSC13
Re03Gl
SeOF2
SiH2Br2
SiHBr3
SiBr2Cl2
SiBr3Cl
S1C13F
SiBri,
SiCli,
S12C16
Si.+ Hio
Cl3SiOSiCl3
S 2 C 1 2
S02C12
SOC12
SnClit
TiCli,
UF5
VOC13
- 8.5
-45.5
-85
-96.5
-75
-15
-68
-49
21
-112
. 13
2
-35
4.5
4.6
-70.1
-73
-45
-28
-60
5.4
-70
- 1
-108
28
-80
-54
-105
-33
-25
—
-77
185
63
126
8/1
117
116
89
69.5
84
103
76
138
105
125
131
124
66
109
104
128
113
154
56
145
84
137
136
69
79
114
136
111
127
-------
7. REFERENCES :
I. Evans,' P. i>. j and' P. F. Tiley , J.: Chem. ' Soc . (B) 1966, 134-6.
2,. - Scott, D;. W.-, et al., J. Phys. Cheni. 6?., 680-5 (19.63).
3>.- , "Handbook of Chemistry Physics, 40th Ed., p 2325,
Chemical Rubber Co., Cleveland (1959).
4. _ , Metals Handbook, Vol. I, p 257.
5. Turner, D. B.', Workbook of Atmospheric Dispersion. Estimates,
U.S. P.H.S. No. 999 AP-26, 1969.. •
6. Fuller, E. N., P. D. Schettler and J. C. Giddings, Ind.
Eng. Chem. 58. (5), 18-27 (1966)..,
7- Neal, M. J., and J. M. Robson, Brit. J. Pharmacol. 24_ (2),
519 (1965); Chem. Abstr. 6_3_, 1101 (1965).
8. Neal, M. J. and J. M. Robson, Brit. J. Pharmacol. Chemotherapy,
.261 (2), 482. (1966); Chem. Abstr., 6^, 14769 (1966).
9. Garmer, N. L. , and J. M. Leigh, Brit. J. Pharmacol, Chemo-
therapy, 31 (2), 345 (1967); Chem., Abstr., 68., 1853 (1968).
10. Gage, J. C., Brit. J. Ind. Med..l970, 27; Chem. -Abstr. 73, .
12650 (1970) • .'-. .
11. Burns, T. H. S., et al. Anaesthesia, 1£ (2), 167 (1964);
Chem. Abstr., 61, 2351 (1964).
12. Lapik, A. S., Izv. Sibirsk . Otd. ; Akad. Nauk, SSSR, Ser. Biol.-
Med. Nauk 1965 (3), 91; Chem. Abstr. &4, 20497 (1966).
13- Riley, R. A., Personal Communication, Jan. 1973-
14. _ , API Toxicological Review - Gasoline, 1st Edition,
API, N.Y., 1967.
15. _ , API Toxicological Review - Benzene, 2nd Edition,
API, N.Y., I960.
16. Veselov. V. G., Mater. Nauch.-Prqct . Konf. Mdlodykh Gig.
Sanit. Urachei, llth 1967, 121; Chem. Abstr. 72, 41134 (1970).
17. Polilei, S. A. et al . , Nauch. Tr. , Omsk. Med. Inst . 1969,
No. 88, 319; Chem. Abstr. JJ>, 107,789 (1971).
186
-------
18. Lapik, A. S. and L. N. Zlmina, Izv. Sib. Otd. Akad, Nauk
SSSR, Ser. Blol.-Med. Nauk 1967. 3, 81-83; Chem. Abstr. 6£,
17721 (1968).
19. Gross, D., J. J. Loftus, T. G. Lee, and V. E. Gray, "Smoke
and Gases Produced by Burning Aircraft Interior Materials,"
Final Report on Contract No. FA66NF-AP-7, National Bureau
of Standards, 1968.
20. Hanker, J. S., R. M. Gamson, and H. Klapper, Anal. Chem. 29,
879 (1957).
21. , "Standards of Performance for New Stationary
Sources," Federal Register 3j>, No. 247, 24891-24893 (1971).
22. Pollard, R. T. , Combustion and Flame, 17., 337-42 (1971).
23- Chaffee, R. G., and H. N. Beck, J. Chem. Eng. Data, 8, 453
(1963).
24. Kirk, H. F., and D. F. Othmer, "Encyclopedia of Chemical
Technology," 2nd Edition, Vol. 9, p 698, John Wiley, New
York (1966).
25. Bennett, R. H., and D. W. Cottrell, U. S. Patent 3,453,337,
July 1, 1969-
26. Bennett, R. H., and D. W. Cottrell, U. S. Patent 3,453,338,
July 1, 1969.
27. Cottrell, D. W., U. S. Patent 3,472,789, October 14, 1969.
28. Fuller, G., U. S. Patent 3,574,775, April 13, 1971.
29. Mills, H. E., "Costs of Process Equipment", Chem. Eng. March
16, 1964,
30. Guthrie, K. M., "Capital Cost Estimating", Chem. Eng., March
24, 1969.
31. Kirk, H. F., and D. F. Othmer, Encyclopedia of Chemical
Technology, 2nd Edition, Vol. 9, p 524, John Wiley, New
York (1966).
32. Kirk, H. P., and D. F. Othmer, Encyclopedia of Chemical
Technology, 2nd Edition, Vol. 22, p 77, John Wiley, New
York (1966).
187
-------
33. , Chemical Marketing Reporter, Jan. 8, 1973-
3^. Aries., R. S., U. S. Patent 2,698,849, Jan. 4, 1955-
35i , "Vehicle Design Goals - Six Passenger Automobile,"
Revisions B and C, EPA, Ann Arbor, 11 February and May 28,
1971.
36. -Jakob,. M., "Heat Transfer," Vol. II, Wiley, New York (1957).
37. Balje, 0. E., J. Eng. for Power, 84, 83-102 (1962).
38. Smith, A., and A. W. C. Menzies, J. Am. Chem. Soc. , 32,
897, 907, 1412, 1448 (1910).
39- Hammann, W. C., WADC Technical Report 57-657, ASTIA Docu-
ment No. AD-142285, Contract No. AF33(6l6)-38l9, PP 158-71
(1958).
40. Blake, E. S., W. C. Hammann, J. W. Edwards, T. E. Reichard
and M. Ort, Chem. and Eng. Data, 6_, 87 (1961).
41. Johns, I. B., A. E. McElhill and J. 0. Smith, Ind. and Eng.
Chem., Prod. Res. and Dev., !_: 2. .(1962).
42. Blake, E. S., J. W. Edwards, W. C. Hammann, WADC Technical
Report 54-532, Part 1, (1955), ASTIA Document No. AD65414.
43. Blake, E. S., J. W. Edwards, W. C. Hammann and T. Reichard,
WADC Technical Report 54532 Part 2 (1956), ASTIA Document
No. AD 94172; part 3 (1957), ASTIA Document No. AD 118179.
44. Blake, E. S., J. W. Edwards, W. C. Hammann and T. Reichard,
WADC Technical Report 57-437 (1956) ASTIA Document No.
AD 142188.
45. Blake, E. S., and G. A. Richardson, "Proceedings of the
USAF Aerospace Fluids and Lubricants Conference," pp 130-9
(1963).
46. Blake, E. S., G. A. -Richardson, J. A. Webster and R. E.
Debrunner, ASLE Transactions 9, 47-58 (1966).
47- Blake, E. S., G. A. ^Richardson and J. A. Webster, "High-
Temperature Fire Resistant Fluids," RTD-TDR-63-4186 (1963).
48. Blake, E. S., R. E. DeBrunner and J. A. Webster, "Fire
Resistant Hydraulic Fluid for Submarine Use," (1964).
188
-------
49- Richardson, G. A., and E. S. Blake, Ind. and Eng. Chem. Prod.
Res. and Dev., 7_, 22 (1968).
50. Blake, E. S., Ind. Eng. Chem., Prod. Res. and Dev., 7_, 101
(1968).
51. Johns, I. B., E. A. McElhill and J. 0. Smith, J. Chem. Eng.
Data, 7, 277 (1962).
52. Sheehan, D., A. P. Bentz and J. Petropoulos, J. Appl. Polymer
Sci., 6, 47 (1962).
53. Reid, R. C., and T. K. Sherwood, "The Properties of Gases
and Liquids", Second Edition, McGraw-Hill (1968).
54. Kirschfelder, J. 0., R. J. Buehler, H. A. McGee, Jr., and
J. R. Sutton, Ind. Eng. Chem. 50., No. 3, 375-90 (1958).
55. Grimes, R. N., "Carboranes", Academic, New York (1970).
56. Olsen, R. R. and Grimes, R. N., J. Amer. Chem. Soc. 92,
5072 (1970).
57. Onak, T. P., Gerhart, F. J., and Williams, R. E., J. Amer.
Chem. Soc. 8J5, 3378 (1965).
58. Tebbe, F. N., Garrett, P. M. and Hawthorne, M. F., J. Amer.
Chem. Soc. £0, 869 (1968).
59- Tebbe, F. N., Garrett, P. M., Young, D. C. and Hawthorne,
M. P., J. Amer. Chem. Soc. 88, 609 (1966).
60. Butler, J. M., Wheeler, G. and Ross, W. D., Anal. Chem. 40,
466 (1968).
61. oio.nner, G. B. and Ruehrwein, R. A., J. Phys. Chem. 63, 1736
(1959).
189
-------
TECHNICAL RE. ORT DATA
(/'lease rc
------- |