McMaster University

DigitalCommons@McMaster
Open Access Dissertations and Teses Open Dissertations and Teses
5-1-1984
Kinetics of natural degradation of cyanide from
gold mill efuents
Ljubica Simovic
Follow this and additional works at: htp://digitalcommons.mcmaster.ca/opendissertations
Part of the Chemical Engineering Commons
Tis Tesis is brought to you for free and open access by the Open Dissertations and Teses at DigitalCommons@McMaster. It has been accepted for
inclusion in Open Access Dissertations and Teses by an authorized administrator of DigitalCommons@McMaster. For more information, please
contact [email protected].
Recommended Citation
Simovic, Ljubica, "Kinetics of natural degradation of cyanide from gold mill efuents" (1984). Open Access Dissertations and Teses.
Paper 7009.
KINETICS OF NATURAL DEGRA.DATION OF CYANIDE
FROM GOLD MILL EFFLUENTS
KINETICS OF NATURAL DEGRADATION OF CYAHIDE FROM GOLD KILL EFFLUENTS
by
LJUBICA SIMOVIC, B.Sc. (Eng.)
A Thesis
Submitted to the Faculty of Graduate Studies
in Partial Fulfilment of the Requirements
for the Degree of
Master of Engineering
McMaster University
May, 1984
MASTER OF ERGINEERING (1984)
(Chemical Engineering)
McMaster University
Hamilton, Ontario
TITLE:
AUTHOR:
SUPERVISORS:
KINETICS OF NATURAL DEGRADATION OF CYANIDE FROM GOLD
MILL EFFLUENTS
Ljubica Simovic
B.Sc. (Eng) (University of Novi Sad, Yugoslavia)
Dr. W.J. Snodgrass, Associate Professor, Chemical
Engineering Department
Dr. K.L. Murphy, Professor, Chemical Engineering
Department
NIMBER OF PAGES: 369
ABSTRACT
Most Canadian gold mining facilities utilize the cyanidation
process in which cyanide is added to sequester gold from the ore.
After zinc addition, gold is precipitated from a gold-cyanide complex.
Waste streams from the process generally contain a sufficiently high
concentration of cyanide and heavy metals that treatment is essential.
The oldest treatment method practiced by Canadian gold mines
for cyanide destruction is "natural degradation".
Based on the literature review of the previous studies, the
most important mechanism in the natural degradation of cyanide was
recognized as being volatilization. Among the variables affecting the
volatilization process the three most important were selected for this
study, namely: temperature, UV light and aeration rate. The experimen-
tal pH chosen was 7.0. In order to evaluate the relative importance of
these variables a full 2
3
factorial design was employed and appropriate
experiments conducted over a period of one year.
The synthetic solutions examined were simple cyanide (NaCN),
four single metallo-cyanide complexes (Cu, Zn, Ni and Fe) and two
mixtures - a "low mix" containing a low concentration of metals rela-
tive to the total cyanide concentration and a "high mix", saturated
with metals.
i
Analysis of the results of the experimental design led to the
conclusion that temperature had the largest effect upon reaction
kinetics. The rate of aeration together with temperature had a signi-
ficant effect upon the volatilization rate within the first 48 to
72 hours. Ultraviolet irradiation only had a significant effect upon
the decay rate of the iron cyanide complex and low mix.
A mathematical model for the degradation of a single
metallo-cyanide complex solution was postulated for a batch reactor
taking into consideration the law of conservation of mass, and assuming
that chemical equilibrimn existed between hydrocyanic acid (HCN) and
cyanide ion (CN-). The rate of volatilization of HCN and metal decay
rate were the mechanisms assumed to be controlling the cyanide degrada-
tion process.
The best estimates of the metal decay coefficients from
single metallo-cyanide solutions used to simulate the experimental data
of the mixed metallo-cyanide solutions are as follows:
Na
2
Cu(CN)3
Na2Zn(CN)4
Na
2
Ni(CN)4
K3
Fe
(CN)6
0.00295
0.01783
0.00044
0.00143
0.00753
0.04496
0.00095
0.00473
The model fits the data at the 95% confidence level for five of the
eight test conditions; however, for the other three it is poorer.
The model, calibrated using coefficients from the synthetic
solutions, has been applied to actual gold mill effluents. Results
show that the model can be used for the basic estimate of the cyanide
ii
degradation rate. For a more precise result, a recalibration of the
model is necessary as a function of the particular gold mill effluent's
metal content.
iii
ACKNOWLEDGEMENTS
I would like to express my gratitude to Dr. W.J. Snodgrass
and Dr. K.L. Murphy for their patient guidance, professional assis-
tance and interest during the course of this research.
Many, with whom I have been fortunate to work, have also
contributed to the completion of this project. Mr. J.W. Schmidt, Head,
Physical/Chemical Processes Section, Wastewater Technology Centre,
(WTC) , provided scientific support and encouragement throughout the
program. Mr. K. Conn, Head, Laboratory Services Section, and his
staff, particularly Mr. J. Dupuis, gave excellent analytical support to
the requirements of the project. Thanks are also due to Mr. R. Mercer,
a student, for his assistance in the laboratory, my friend
Mrs. N. Harper for her constant moral support and to Mrs. G. Graham and
Mrs. D. Crabtree for undertaking the typing of the thesis.
Finally and most importantly, my deepest appreciation goes to
my husband, Rade, not only for his support and self-sacrifice but also
for his unfailing trust and optimism during the years required for this
work. Special thanks go to my children, Nancy and Peter for their
patience, love and understanding when it was most needed.
iv
TABLE OF CONTEliTS
ABSTRACT
ACKNOWLEDGEMENTS
TABLE OF CONTENTS
LIST OF FIGURES
LIST OF TABLES
1 INTRODUCTION
2 LITERATURE REVIEW
2.1 Cyanidation
2.2 Forms of Cyanides and Metals in Mining Wastes
2.2.1
2.2.2
2.2.3
2.2.4
2.2.5
Cyanide
Copper
Zinc
Nickel
Iron
2.3 Cyanide Chemistry
2.3.1
2.3.2
2.3.3
Free cyanide
Simple cyanides
Complex cyanides
2.4 Removal of Cyanide by Natural Degradation
2.4.1
2.4.2
2.4.3
2.4.4
2.4.5
2.4.6
2.4.7
Volatilization
Photodegradation
Chemical oxidation
Microbial oxidation
Thiocyanate formation
Cyanide in soil
Conclusion
v
iv
v
vi i i
x
2
2
8
8
9
11
12
13
16
16
18
19
27
27
31
33
33
35
36
37
3
4
ANALYTICAL METHODOLOGY
3.1 Cyanide Determinations
3.2 Cyanate and Thiocyanate Determination
3.3 Metals Determination
SCREENING EXPERIMENTS
39
39
40
41
43
4.1 The Effects of Temperature
t
UV Light and Concen- 43
tration
4.2 Selection of Buffer and pH Level 48
4.3 The Effect of Aeration 53
4.4 The CO
2
Effect 57
5 EXPERIMENTAL PROCEDURES AND ANALYSES OF EXPERIMENTAL RESULTS 59
6
7
8
5.1 Introduction
5.2 Full 2
3
Factorial Design
5.3 Inspection of the Data
5.4 Rates of Reaction
5.4.1
5.4.2
General
Rate calculations
5.5 Examination of the Factors Affecting
Volatilization of Free Cyanide
5.6 Comparison of Volatilization Rates
ANALYSIS OF EXPERIMENTAL DESIGN
6.1 Introduction
6.2 Results
6.3 The Effect of Aeration on NaCN Degradation
MODEL DEVELOPMENT AND CALIBRATION
7.1 Model Development
7.2 Parameter Estimation Method
the
EVALUATION OF SINGLE METALLO-CYANIDE SOLUTIONS RESULTS
AND DISCUSSION
8.1 Four Parameter Estimation Model
8.2 Three Parameter Estimation Model
8.3 Two Parameter Estimation Model
8.4 Parameter Estimation With Fixed MCN/TCN Ratio
vi
59
59
61
611
64
65
67
72
76
76
78
85
90
91
95
98
98
103
115
117
9 VERIFICATION STUDIES
Page
124
9.1 Simulation on the Synthetic Mixes 124
9.1.1 Low mix 128
9.1.2 High mix 136
9.2 Evaluation of Model Calibration with Barren Solutions 143
9.2.1
9.2.2
9.2.3
9.2.4
10 CONCLUSIONS
Introduction
Simulations with the original kl values
Analyses of barren solutions parameters
Simulations with the adjusted kl values
11 RECOMMENDATIONS
REFERENCES
APPENDIX A
APPENDIX B
APPENDIX C
APPENDIX D
APPENDIX E
APPENDIX F
APPENDIX G
APPENDIX H
APPENDIX I
APPENDIX J
vii
143
145
146
151
159
162
164
177
179
192
207
225
229
235
275
341
358
FIGURE
2.1
2.2
4.1
5.1
5.2
LIST OF FIGURES
Simplified Gold Mill Flowsheet
Relative Percentage of Hydrocyanic Acid and Cyanide Ion
in Solution as a Function of pH
The Setup for Bench Scale Experiments
Total Cyanide Degradation in Sodium Cyanide Solution
Typical Results of Total Cyanide Degradation in FeCN,
Cu, Zn and Ni Cyanide Complex Solutions
PAGE
5
17
51
62
5.3 a Volatilization Mass Transfer Coefficient kv versus 68
5.3 b
5.3 c
5.4
6.1
6.2
7.1
8.1
8.2
8.3
Temperature
Volatilization Mass Transfer Coefficient kv versus UV
Light
Volatilization Mass Transfer Coefficient kv versus
Aeration
Typical Examples for 95% CIon Part of the Degradation
Curve Attributed to Volatilization - NaCN versus Cu
(a), Zn (b), Ni (c), Fe (d)
Half Normal Probability Plots (Cu, Zn, Ni and Fe
Cyanide Solutions)
Half Normal Probability Plot for kv for NaCN
Cyanide Degradation Model
Four-Parameter Estimation Model Typical Fit of
Predicted Values to Observed Data (Zn-CN, 20C, AIR,
UV)
Three-Parameter Estimation Model - FCN >MCN
Three-Parameter Estimation Model - MCN >FCN
vii i
69
70
75
82
89
93
99
109
110
LIST OF FIGURES (Continued)
FIGURE PAGE
8.3a Three-Par ameter Es tima tion Model Es tima ted MCN 1 1 4
Values with 95% Confidence Intervals - Cu
9.1 Low Mix Simulation - 20C, NO AIR, UV
9.2 Total Cyanide Degradation - Single Metallo-Cyanide
Complex Solutions and Low Mix Cyanide Solutions (4C,
NO AIR, UV)
9.3a High Mix - First Simulation (4C, NO AIR, NO UV)
9.3b High Mix - Second Simulation (4C, NO AIR, NO UV)
9.4 Barren Solution HI - Original
9.5 Barren Solution HI - Adjusted kCu and kZn
9.6 Barren Solution H2 - a) Original; b) Adjusted kcu and
kZn
9.7 Barren Solution H3 - a) Original; b) Adjusted kcu and
kZn
9.8 Barren Solution H3 - a) Original; b) Adjusted k
Fe
ix
130
133
138
142
148
153
154
155
157
TABLE
2.1
2.2
2.3
2.4
4.1
4.2
4.3
4.4
4.5
4.6
5.1
5.2
6.1
6.2
6.3
6.4
LIST OF TABLES
Summary of Composition of Process Streams and Mill
Effluents for Selected Canadian Gold Milling Operations
Solubility of Minerals in Cyanide Solution
Solubility Products and Formation Constants of Cyanide
Compounds (Ingles, 1981a)
Oxidation States of Metals Forming Cyanide Complexes
Approximate Wavelength Limits for Transmission of
Various Optical Materials Near Room Temperature
The Test Conditions for Bench Scale Experiments
The Effect of pH on the Rate of Cyanide Removal
(Choice of Buffer)
3-
Efeects of Cl- and P0
4
on Cyanide Removal Rate
The Effect of Aeration
The Effect of CO
2
on the pH Readings
Factorial Design Matrix
Factors Affecting the Rate of Volatilization of Free
Cyanide (Results of the Dot Diagram Analyses)
Reaction Rate Constant as the Response Variable
Results of 2
3
Factorial Design in Terms of Relative
Effects
Complete Results of the Full 2
3
Factorial Design
The Response Variables and Effects - NaCN
x
PAGE
7
10
15
20
45
50
52
54
56
58
60
73
80
81
84
87
TABLE
6.5
B.l
8.2
LIST OF TABLES (Continued)
Probability (%) and Effects Arranged in Ascending Order
(NaCN)
Estimated Metal Cyanide Decay Coefficients kl [h-l]
for Single Metallo-Cyanide Solutions (Four-Parameter
Estimation Model)
Estimated Values
Coefficients kv
Model)
of Volatilization Mass Transfer
[em/h] (Four-Parameter Estimation
PAGE
88
100
101
8.3 Metal Cyanide to Total Cyanide Ratio - Four Parameter 104
Estimation Model
B.4 The Estimated Metal Cyanide Decay Coefficient kl [h-l] 105
at kv 8 0.93 [em/h]
8.5 a Three Parameter Estimation Model - Cu and Zn Cyanide at 107
kv Average
8.5b Three Parameter Estimation Model - Ni and Fe Cyanide at 108
kv Average
8.6 Results on Residual Sums of Squares in Two and Three 116
Parameter Estimation Models (PEM)
B.7 Number of Minimum RSS Values at the Particular MCN/TCN 119
Ratio
B.8 Results of the F Test 120
8.9 The Resutls of Three Parameter Estimation Model
MCN/TCN - 0.17
121
9.1 Estimated Metal Decay Coefficient Values kl (h-
1
) at 125
MCN/TCN K 0.17 (Averaged with Respect to Temperature)
9.2 Energy of Activation for kl Between 4C and 20C 127
9.3 Initial Parameters used for the Model Simulation of the 129
Low Mix
xi
LIST OF TABLES (Continued)
TABLE PAGE
9.4 Total Cyanide Values (Observed and Estimated) for low 131
Mix Cyanide Solution
9.4a Verification of the low Mix Model Simulation - Standard 135
F Test (5%) versus Modified F Test
9.5
9.6
9.7
9.8
9.9
The Initial Parameter Values for High Mix Cyanide
Solution (First Simulation)
Total Cyanide Values, Observed and Estimated for High
Mix Cyanide Solution (Two Attempts)
The Initial Parameter Values for High Mix Cyanide
Solution (Second Simulation)
Barren Solution Parameters
The Initial Parameters Required for the Model
Simulation of Barren Solutions
9.10 Original Simulations on Barren Solution
9.11 Additional Barren Solutions' Parameters
9.12 Modified Simulation on Barren Solutions
xii
137
139
141
144
147
149
150
156
1
mente
IRTR.ODUCTION
Cyanide and cyanide compounds are found throughout the environ-
The natural origins of cyanide and its derivatives are many
plants containing cyanogenic glycosides, which when hydrolyzed release
hydrogen cyanide (Ecological Analysts Inc., 1979). Historically,
however, cyanide in the environment has been associated with industrial
point sources. Cyanides are extensively employed in industry and occur
in effluents from many different industrial processes such as the case
hardening of steel, electroplating and metal-finishing, the scrubbing
of coke-oven or blast furnace gases, photography and the extraction of
precious metals such as silver and gold.
A reasonable amount of information on the treatment of
cyanide-bearing wastes from plating industry and iron and steel subpro-
cesses is available. But little information is available on the treat-
ment of gold milling effluents.
Most of the gold mined in Canada is extracted from the ore by
the process of cyanidation. The cyanide-bearing waste streams from the
process are the barren bleed and ~ h washed tailings containing the
waste rock. Typically, these wastes contain simple cyanides of sodium
or calcium, the cyanide complexes of copper, iron (less frequently),
nickel, and zinc, thiocyanates and at some mines, arsenic. In addition
to the toxic effec ts of cyanide these heavy metals can pose toxic
problems. Their discharge may exceed the water quality criteria.
The "methods considered to offer the greatest possibility for
cyanide removal from gold mill effluents are: natural degradation,
oxidation processes (using chlorine gas, hypochlorite, ozone or
hydrogen peroxide), acidification/volatilization/reneutralization,
electrolytic processes, ion exchange and conversion to less toxic forms
(such as ferro cyanide , thiocyanate, etc).
For a long time, natural degradation in tailings ponds has been
the only method practiced by Canadian gold mines for cyanide destruc-
tion. When this method is applied, usually all cyanide-bearing waste
streams are pumped to a common tailings pond. In this pond the solids
settle and a clear solution, the tailings pond overflow, is discharged
to the environment. The liquid residence time in the tailings pond may
be of the order of several days or months during which time the process
of natural degradation of cyanide occurs.
For a long time an engineering goal has been to evaluate the
efficiency of tailings ponds in cyanide removal from a mechanistic
point of view, in order to develop criteria for pond design. In this
study, it was decided, that one way to accomplish this would be to
substitute synthetic mixtures for actual gold mill effluents and
examine the phenomenon of natural degradation of cyanide in more
detail. For this purpose, the study uses a laboratory and modelling
approach. The objectives of this research are as follows:
2
(1) to examine the natural degradation of cyanide in synthetic
solutions;
(2) to delineate maj or factors affecting cyanide removal, and
mechanisms involved; and
(3) to model the kinetics of cyanide removal and compare the
results with observations on actual systems.
3
2 LITERATURE REVIEW
2.1 Cyanidation
Over 90% of the gold mined in Canada is extracted from the ore
by the process of cyanidation. This process consists of addition of
sodium or calcium cyanide solution to the crushed and slurried ore in
agitated and aerated tanks for up to 50 hours under alkali conditions
(Figure 2.1). Elsner's equation has been suggested as being represen-
tative of the reaction that takes place during the dissolution of gold
in dilute cyanide solutions (Hedley and Tabachnick, 1958):
4 Au + 8 NaCN + 02 + 2 H
2
0 - 4 NaAu (CN)2 + 4 NaOH
The solution pH is controlled at a high level (pH: 11) using
lime additions to prevent the volatilization of hydrogen cyanide gas.
After clarification and filtration the pregnant solution
containing gold is separated from the solids, and gold is precipitated
from it by the addition of zinc dust. Zinc precipitation of gold is
represented by the following
2 NaAu(CN)2 + Zn + + 2 Au
ORE FROM MINE
!
CRUSHIIG
GRIIClIIG
AERATION
_ LINE NfJ CYAHUE
FILTER. rASH
IIEPtlP
TAILIJCS
PIIIJ
DISDWIIiE
IlLUTION
CLARIFICATION
lIE-AERATION
PRECIPITATION
Pigure 2.1 Siwp11fied Gold 11111 novsbeet
SOLUrION
I
I
r _ L J ~
DISDWIIiE ,
tll.DIIG
PIIIJ
5
A gold free solution, called "the barren solution" is produced follow-
ing precipitation. The barren solution contains cyanide, heavy metals
(Cu, Zn, Nt and Fe), thiocyanates, cyanates and sometimes arsenic.
Part of the barren solution is returned to the crushing and
grinding circuit to aid in the cyanidation process. The remainder,
called "the barren bleed" is discharged in order to avoid excessive
accumulation of heavy metals in process waters.
As shown in Figure 2.1, the barren bleed constitutes only a
portion of the total cyanide-bearing effluent to be disposed of - the
other main effluent source is the washed filter residue containing the
leached solids (tailings) plus repulping water and other miscellaneous
streams (see Table 2.1). The barren bleed and washed tailings are
usually combined and discharged to the tailings pond. After settling
of solids, the tailings pond supernatant is left in the pond for a
prolonged period of time to enhance natural degradation of cyanide. At
some mines (e.g., Dome Mines Ltd. and Pamour-Schumacher Mines), waste
barren solution (the effluent with the highest cyanide content) is
segregated from the other cyanide-bearing solutions and retained in a
separate pond for the purpose of natural degradation of cyanide.
6
TABLE 2.1. SUMKARY OF COMPOSIfiOR OF PROCESS STREAMS AND MILL EFFLUENTS FOR. SELECTED
CANADIAN GOLD MILLING OPERATIONS
Flow x 10-
3
Total eN eu Zn Ni Fe CNS-
LId pH (mg/L) (mg/L) (mg/L) (mg/L) (mg/L) (mg/L)
Barren 150 to 11.3 to 86 to 6 to 10.2 to 0.28 to 0.1 to 42 to
Solution 560 12.2 650 275 237 10.0 35.7 510
Filtered* 191 to 10.2 to 27 to 3.0 1.02 to 0.1 to 0.1 to 16 to
Tailings 2,300 11.7 150 15.0 22.9 2.36 7.14 120
Tailings Pond o to 7.2 to 0.2 to 0.02 to 0.04 to 0.07 to 0.4 to 2 to
Effluents 13,600 8.3 48 37.3 0.82 0.7 5.2 72
Source: lEe, 1979.
* ~ m p l e s filtered; results indicate soluble components.
.......
2.2
2.2.1
Forms of Cyanides and Metals in Mining Wastes
Cyanide
The forms of cyanide in process waters are a result of reac-
tions during gold cyanidation process. These forms are not well
known. The cyanide may be chemically or physically adsorbed on
tailings mineral surfaces and/or it may have reacted with metallic and
nonmetallic constituents in tailings to produce cyano complexes,
thiocyanates and cyanates at the mineral surfaces or in the
interstitial solutions (Huiatt et al., 1983).
Metal-cyanide complexes are produced by the reaction of cyanide
with a variety of mineral types: oxides, carbonates, silicates and
sulphides of copper, zinc, nickel and iron. In Canadian ores it is the
sulphide minerals that usually predominate in the feed to the cyanida-
tion process (Scott and Ingles, 1981).
It is usually thought that most of cyanide is in solution
rather than adsorbed on suspended particles. The fact that cyanide is
used to leach gold from the ore supports that theory. The large
fraction of metal sulphides in solution oxidize to polysulphides and
thiosulphates. Some cyanide reacting with them becomes converted to
thiocyanate and as insoluble precipitates but much of it remains in
metallo-cyanide complexes and free cyanide forms (Schmidt et al., 1981;
Conn, 1984). Even if adsorbed on suspended particles, some of the
cyanide ion can be slowly re-released into the water assuming that such
adsorption phenomena are reversible (Caruso, 1975).
8
2.2.2 Copper
a n ~ precious metal ores contain copper minerals such as:
azurite, 2CuCO-Cu(OH)2; malachite, CuC0
3
-Cu(OH)2; chalcocite, CU2S;
chalcopyrite, CuFeS2; cuprite, Cu20, etc. In these minerals copper
may have an oxidation state ranging from +1 to +2. Under the condi-
tions normally used in cyanidation these minerals are readily and
completely solubilized in weak cyanide solutions (Table 2.2). Copper
combined with cyanide forms a variety of complexes. The complexation
is particularly rapid at high pH values (Caruso, 1975). The complexes
include simple complexes such as cupric and cuprous forms of cyanide
and thiocyanate (CU(CN)2' CuCN, Cu(CNS)2 and CUCNS) and double copper
1- 2- 3-
cyanides with complex ions such as Cu(CN)2 , Cu(CN)3' and Cu(CN)4 Of
the complex ions
2-
Cu(CN)3 is considered the most probable to occur
(Hedley and Tabachnick, 1958). Copper sulphides are readily oxidized
to sulphates, and copper then released as Cu
2
+. If Cu
2
+ is present in
the ore, some of the cyanide may be oxidized to cyanate (CNO-) in
reducing combined Cu
2
+ to Cu
1
+. For each two cupric ions reduced, one
cyanide ion is converted to cyanate. Cyanide consumed in this way is
not recoverable (Fairbridge, 1972).
Cuprous ion has a complete outer shell of 18 electrons and as
such does not absorb radiation in solution. Cupric ion, however, has
17 electrons in its M-shell and can absorb radiation of suitable wave-
length (3630, 7580 and 11900A), (Noblitt, 1973).
9
10
TABLE 2. 2. S()LUBILITY OF MINERALS Df CYARIDE SOurrIOIf
Main
Element
Gold
Silver
Copper
Zinc
Iron
Calaverite
Argentite
Cerargyrite
Proustite
Pyrargyrite
Azurite
Malachite
Chalcocite
Cuprite
Bornite
Enargite
Tetrahedrite
Chrysocolla
Chalcopyrite
Smithsonite
Zincite
Hydrozincite
Franklinite
Sphalerite
Gelamine
Wi llemite
Pyrrhotite
Pyrite
Hematite
Magnetite
Siderite
Marcasite
Source: R. Tremblay. 1983.
Mineral
AuTe2
Ag2
S
AgCI
Ag3
AsS
3
) Ruby
Ag3
SbS
3
)silver
2CuCO.Cu{OH)2
CUC03Cu{OHh
CU2S
Cu20
FeS 2Cu2S.CuS
3CuSAs
2
Ss
4Cu2S.Sb2Sg
CuSiOg
CuFeS2
ZnC03
ZnO
3ZnCO
g
2H
2
O
{Fe. Mn. Zn)O.{Fe.
Mn)20g
ZnS
H2Zn2SiO ..
Zn2Si04
FeS or FeSS6
FeS2 (isometric)
Fe20g
FegO ..
FeCOg
FeS2
(orthorhombic)
% Dissolved
in 24 hours
Readily soluble
Readily soluble
Readily soluble
Sparingly soluble
Sparingly soluble
94.5
90.2
90.2
85.5
70.0
65.8
21.9
11.8
5.6
40.2
35.2
35.1
29.2
18.4
13.4
13.1
Readily soluble
Sparingly soluble
Sparingly soluble
Practically insoluble
Readily soluble
Readily soluble
11
2.2.3 Zinc
Zinc occurs in a precious metal ore in a number of minerals -
smithsonite, ZnC0
3
; sphalerite, ZnS; willemite, zincite, ZnO
and several others. These zinc minerals are sufficiently soluble under
normal cyanidation conditions. The reaction of cyanide with sphalerite
for example (Ingles, 1981a), proceeds according to the equation:
In the presence of oxygen, sulphites and other intermediate products
such as thiosulphate and thionates may be formed (Hedley and
Tabachnick, 1958). Zinc forms complexes in aqueous solution but these
are not particularly stable or inert (Fairbridge, 1972).
Sodium zinc cyanide is soluble in water (Hedley and Tabachnick,
1958) dissociating at any pH to some extent as follows:
+ + 2- +
+ 2Na + Zn(CN)4 + Zn(CN)2 + 2CN
Another source of Zn in mill effluent is from the gold precipi-
tation process. The amount of Zn dust used for the precipitation of
precious metals depends to a large extent on the amount of precious
metal in solution although other factors such as oxygen content,
cyanide strength and alkalinity play an important part. During the
reactions involved in precipitation, the greater part of the zinc
dissolves in the cyanide solution and forms various complexes such as
sodium or zinc cyanide, zinc cyanide, zinc thiocyanate and zinc
ferrocyanide (Hedley and Tabachnick, 1958).
Zinc's chemical behaviour is dictated by the existence of a
2+
single oxidation state, i.e., Zn In a crystal zinc has the tendency
to form covalent bonds which show a distinct preference for tetrahedral
coordination. The octahedral coordination does not commonly occur.
The Zn
2
+ ion is normally found with ferrous ion and apparently competes
with it during fractionation for vacant octahedral lattice sites
(Fairbridge, 1972).
2.2.4 Nickel
Nickel is most frequently found in certain varieties of the
iron mineral pyrrhotite (3 to 5% Ni) or in combination with sulphur as
millerite NiS. Although nickel can achieve oxidation states of -1,
+1, +3 and +4, the majority of the nickel compounds are of the Ni(II)
oxidation state.
The kinetics of formation and decomposition of the tetracyano-
nickelate (II) complex ion have been studied in some detail (Kolski and
Margerum, 1968). The rate of formation is a fifth-order reaction,
first-order in nickel and fourth-order in total cyanide; HCN is a reac-
tant as well as cr. In the pH range 5.5 to 7.5 the rate is propor-
tional to By the use of iodine as a ion scaven-
2-
ger, the first-order rate constant for dissociation of the Ni(CN)4 ion
12
at pH 3 to 8 in ~ - 0.1 NaCl0
4
is estimated as 4.8 x 10-
4
(s-1). The
2-
yellow Ni(CN)4 ion (a square-planar complex) is both thermodynamically
very stable and kinetically slow to release ~ ion (Crouse and
Margerum, 1974). Nickel is able to displace Zn and Cd from their
cyanide complexes, forming the more stable tetracyanonickelate (II)
complex (Broderius, 1973).
Nickel thiocyanate is also know to form. It is a yellow-brown
hydrated solid that reacts with an excess of SCN- to form complex ions
2- 4_
Ni(SCN)4 and Ni(SCN)6 (Cotton and Wilkinson, 1962). A mixed cyano
thiocyanato complex K2[Ni(CN)2(SCN)2] is obtained from nickel thiocya-
nate and potassium cyanide (nitrogen atom of the thiocyanate ion acts
as ligand) (Sharpe, 1976).
The chemistry of nickel is not simple because of the existence
of complex interconversions of various structures, i.e., octahedral,
tetrahedral and square (Cotton and Wilkinson, 1962).
2.2.5 Iron
The principle sources of iron are the iron sulphide minerals:
pyrite FeS2 (isometric), marcasite FeS2 (orthorhombic) and pyrrhotite
Their reactivities increase in that order (Ingles, 1981b).
Cyanide reacts with iron (II) and iron (III) to form extremely stable
. 4- 3-
octahedral complexes: ferrocyanide Fe(CN)6 and ferricyanide Fe(CN)6 '
respectively. These reactions are slow and incomplete at high pH
values (Caruso, 1975). Together these two complexes form an
oxidation-reduction couple; ferrocyanide, which is the usual form in
13
14
solution at normal environmental oxidation potential levels, being
fairly easily oxidized to ferricyanide (Ingles, 1981b). In addition to
the difference in their stability constants (see Table 2.3),
ferrocyanide forms an insoluble zinc salt while ferricyanide does not.
Perhaps the best-known reaction of aqueous Fe
3
+ is with thio-
cyanate ion to form one or more intensely red coloured thiocyanate
complexes.
The principal iron mineral of concern, pyrrhotite, is reported
to react first via its labile sulphur atom to yield thiocyanate and
ferrous sulphide:
The ferrous sulphide formed oxidizes rapidly to ferrous
sulphate and the ferrous iron then combines with cyanide
2+ 2-
FeS + 20
2
+ Fe + S04
Fe
2
+ + 6 ~ + Fe(CN):-
These reactions will be competing, of course, with the
hydrolysis-oxidation reactions leading to the precipitation of ferrous
and ferric hydroxides (Ingles, 1981b).
Fe
2
+ + 20H- + Fe(OH)2(s)
4 Fe(OH)2(s) + O
2
+ 2H
2
0 + 4 Fe(OH)3(s)
YAILI 2.3. SOLnn.tn PItODOC'l'S AJI) POIMAnow COIISTAlftS OP CYAWIDB CUMPOOilJS (l1ll1_.
1981.)
Metal COlipound
CadalUil It Cd(CNh
Cobalt II
Copper I Cu,CIf
Cold I
trOD II
I rOil III -
M.rcury I fig 2 (CIf>2*
Mercury II
Kickel II 1f12(CIf)4+
+
2_
Ni + Ifi( CN) 4
S11 r t . Ag, eM
Zillc II
Solubility Cumulative Formation Constant8
Product
101 K8P 101 Kl log K2 log K3 101 ~ 101 K6
8.0 5.48
19.49
39.3
8.77
15.92
10.60 15.23
24.0
38.3
2l.1
28.59
21.7
18.78
30.30
41.4
22.0
20.6
16.7
very large
35 (47)**
42 (52)
* IKce81 eyanlde IIve8 metallic .ercury and the Rg2+ complex, fig (CN)!:
.. '.lue8 In parenthese8 fro. Broderlu8, 9 7 3 ~
\11
16
Chemically, there is no evidence that any definite hydroxide, Fe(OH)3
exists, and the red-brown precipitate commonly called ferric hydroxide
is best described as hydrous ferric oxide, Fe203-n H
2
0 (Cotton and
Wilkinson, 1962).
2.3 Cyanide Chemistry
2.3.1 Free cyanide
Cyanide exists in water in several forms. The term "free
cyanide" refers to a summation of the ionic form CN- and the molecular
form, hydrocyanic acid, HCN. The equilibrium between HCN and CNr is pH
dependent.
[HCN]
+
+ K =
a
II: 4.36 x 10-
10
(at 20C)
(2.1)
At the pH value of 9.36 (equal to the pKa) the concentration of
HCN and CNr ion are equal (Broderius, 1973). With an increase in pH of
one pH unit there is a tenfold decrease in the [HCN]/[ClrJ ratio. At
lower pH values (pH 7.0) ;lDOst cyanide exists as molecular HCN (99.5%)
(see Figure 2.2) which is the major toxic form of cyanide. Molecular
HCN has a relatively hlgh vapour pressure (53.1 kPa at 10.2C), and
therefore can readily be volatilized to the atmosphere, even from
stagnant solutions.
The equilibrium constant Ka varies slightly with tempera-
ture. Extrapolating to a temperature more characteristic of northern
100
90
80
70
60
50
40
30
20
10
.........
.
.
~
~ ~
\
\
,
j
i
;
1
,
~ ~
~ .........
C J ~
I
I
J
i
~ I
J
\
\
,
~
"
/
~ -"
06 7 8 9 10 11 12 13
pH
Pigure 2.2 Relative Percentage of Hydrocyanic leid and Cyanide IOD
in Solution as a I'Unction of pH
17
Canada or winter conditions in southern Canada, approximately 94% of
free cyanide would be in the form of HCN at 4C and pH 8.5 (Leduc
et a1., 1982).
Free cyanide in milling solutions may react with various forms
of sulphur (polysulphide or thiosulphate) from pyrites and other
mineral constituents to form thiocyanate (Luthy and Bruce, 1979).
2- +
CN + Sx +
2-
[S(X-l)] + SCN
In lower concentrations in the environment free cyanide will
oxidize to cyanate, and this will degrade chemically to carbon dioxide
and ammonia (Hendrickson and Diagnault, 1973).
This reaction can occur at pH's up to 8.5 but is greatly
accelerated at lower pH's (Hendrickson and Daignault, 1973).
2.3.2 Simple cyanides
Most of the cyanide used in gold milling is in the form of
alkaline metal salts. They are simple cyanide compounds represented by
18
the formula A(CN)x where A can be alkali (sodium, potassium or
ammonium) or metal, and x, the valence of A, represents the number of
cyano groups present in the molecule.
The aqueous hydrolysis of simple cyanides has been examined by
several investigators, e.g., Milne (1950a), Hyatt (1976) etc. Being
the salts of a strong base (NaOH or KOH) and a weak acid (HCN) their
dissolution in water results in the dissociation of the ionic lattice
followed by the hydrolysis of the cyanide ion to form HCN:
NaCN + Na+ + CNr
Na+ + CN- + H
2
0 : Na+ + OH- + HCN
(2.2)
(2.3)
As the very weakly dissociated hydrocyanic acid permits little free
hydrogen ions in solution, hydroxyl ion will predominate to produce an
alkaline solution. The reduction of pH of these cyanide solutions will
always lead to the preferential formation of undissociated HCN
(Ecological Analysts Inc., 1979). Subsequent behaviour would then be
the same as for HCN explained in Section 2.3.1 for free cyanide.
2.3.3 Complex cyanides
There are 28 elements found (Table 2.4) and more than 64
oxidation states of these elements which may form complex cyanides
(Ford-Smith, 1964). The stability of the complex cyanides varies
greatly from one element to another and from one oxidation state to
another (see Table 2.4). The complexes form in stepwise fashion, with
19
20
TABLE 2.4. OXlDATIOlf STATES OF METALS FORKIlfG CYAlfiDE C<lfPLEXES
M M(O) M(l) M(ll) M(lll) M(IV) M(V) M(Vl) M(VII)
Ti(22) Ti(Ill)
V (23) V(ll) V(Ill) V(IV)
Cr(24) Cr(I) Cr(ll) Cr(III) Cr(IV)
Mn(25 Mn(O) Mn(I) Mn(ll) Mn(IlI) Mn(IV)
Fe(26) Fe(I)? Fe(II) Fe(IlI)
Co(27) CoCO) Co(I) Co(ll) Co(III) Co(IV)
Ni(2S) Ni(O) Ni(I) Ni(II)
Cu(29) Cu(I) Cu(ll)
Zn Zn(II)
Ge(32) Ge(IV)
Mo(42) Mo( II) Mo(Ill) Mo(IV) Mo(V)
Tc(43) Tc(I) Tc(IV)
Ru(44) Ru(II) Ru( III)
Rh Rh(O) Rh(I) Rh(III)
Pd(46) Pd(O) Pd(I) Pd(II)
Ag(47) Ag(I)
Cd(4S) Cd(II)
W (74) W(IV) W(V)
Re(75) Re(O) Re(I) Re(II) Re(III) Re(IV) Re(V) Re(VI) Re(VII)
Os(76) Os(II) Os( III)? Os(VI)
Ir(77) Ir(II) Ir(III)
Pt(7S) Pt(O)? Pt(I)? pte II) Pt(III) Pt(IV)
Au(79) Au(!) Au(IlI)
Hg(SO) Hg(II)
T1(Sl) TI(I) TI(Ill)
Pb(S2) Pb(II)
Po(S4) Po(IV)?
U(92) U(VI)
Source: Ford-Smith, 1964.
successively higher cyanide contents, as the cyanide ion concentration
of the solution is increased.
Under certain conditions which are a function of such factors
as pH, temperature, and total cyanide concentration, complex metallo-
cyanides will dissociate in water. This dissociation generates a
complex cyanide ion which may further dissociate to release cyanide
ions. But many complex ions are, in general, more stable than their
original compounds and hence the extent of their subsequent
dissociations may be relatively small (Caruso, 1975). Although a
simple equation may be written to describe the overall dissociation of
a complex cyanide ion, such one-step reactions are rare.
The actual process usually involves a number of intermediate
steps. To appreciate the quantitative effect of altering conditions on
an equilibrium, all intermediate species need be considered from a
kinetic point of view. Although one species could predominate at any
given time, other species could also be present. Unfortunately, very
little work has been done on the elucidation of all intermediate
reactions of the metallo-cyanide species thus far and, hence, one must
rely upon overall or cumulative dissociation or equilibrium constants
for many complex ions (Leduc et al., 1982).
According to Broderius (l973} the extent to which a complex
cyanide dissociates is of little importance in environmental systems;
the important factor is the rate at which cyanide bound in the complex
exchanges with free cyanide in solution as illustrated in Eq. 2.4:
21
+
+ (2.4)
The rate of exchange is independent of total cyanide concentration and
is governed by the pH of the mediwn and perhaps light intensity. A
decrease in pH increases the rate of exchange. Complex ions such as
4- 3-
Fe(CN)6 , and Fe(CN)6 exhibit slow exchange with cyanide, while com-
2- 2-
plex ions such as Ni(CN)4 ,Cu(CN)3 and Ag(CN)2 exhibit immeasurably
rapid exchange with cyanide (MacDiarmid and Hall, 1953). The
thermodynamic stability of a complex should not necessarily be expected
to show a relationship with its rate of exchange (Adamson et a1.,
1950).
Most of the tests carried out on metallo-cyanide complexes
(Asperger, 1952; Doudoroff, 1956, etc.) revealed that irregardless of
how stable the complexes are (e.g., iron or nickel cyanide complex),
they must be regarded as delayed sources of free cyanide which will be
released under specific conditions (UV irradiation or decreased pH).
This free cyanide is then subject to various natural removal
mechanisms.
Since the most common heavy metals present in mill tailing
solutions are copper, zinc, nickel and iron, their chemistry will be
discussed in detail in the following paragraphs.
a) Copper-Cyanide Complexes (Cu-CN)
High levels of copper are often found in cyanide-containing
effluents.
3-
Cu(CN)4
Common aqueous species include
1-
Cu(CN)2 ,
2-
Cu(CN) 3 and
Copper-cyanide complexes in solution undergo a moderate
22
23
degree of dissociation depending upon numerous factors such as: pH,
temperature, dissolved oxygen concen tr a t ion, initial cyanide
concentration, etc. (Broderius, 1973; Caruso, 1975; Amax of Canada
Ltd., 1980).
When a copper cyanide solution was prepared having the mole
ratio CN:Cu = 2.0, equilibrium was established within two to three
days. However, equilibrium was not attained at CN: Cu mole ratios of
2.5 and 3.0 after an extended period (110 days). This finding was
interpreted as a sign of the slow formation and dissociation of the
2- 1-
Cu(CN)3 complex in comparison with Cu(CN)2 (Broderius, 1973).
b) Zinc-Cyanide Complexes (Zn-CN)
Zinc-cyanide is a complex that readily dissociates in an
aqueous solution. In very dilute solutions, virtually complete
dissociation of these complex ions is to be expected at any pH value
(Broderius, 1973). The tetracyano-complex of zinc (II) may exist in
highly concentrated process waters but is expected to dissociate
readily during in-plant effluent treatment. The equilibrium constant
(Ka = 1.3 x 10-
17
) indicates that the complex will almost totally
dissociate upon effluent dilution (total cyanide concentration less
than 0.3 mg/L as CN-) and, therefore, will be a significant source cf
free cyanide and metals (Doudoroff, 1976).
c) Nickel-Cyanide Complexes (Ni-CN)
For all practical purposes, the N i C N ~ - and N i C N ~ - species
are only present at very high eN"" concentrations. In dilute nickelo-
cyanide solutions in which the cyanide to nickel molar ratio is 4 to 1,
the
2-
Ni(CN)4 ion is the only significant stable complex species
(Penneman al., 1962). The dissociation of nickel-cyanide complexes
is very pH dependent. As the pH decreases, the dissociation of the
complex ion increases (Doudoroff, 1976). Broderius (1973)
examined in detail the rates of dissociation and formation of the
tetracyanonickelate (II) complex ion in solutions of varying total
cyanide concentration and pH. He assumed that cyanonickelate (II)
exists in equilibrium with cyanide and nickel ions:
and calculated the
+
+
4 + 4 t 4 HCN
dissociation constant
(2.5)
as 1.004

0.375 x 10-
31
He found that formation and dissociation of the complex
are very slow; as much as ten days may be required to achieve
dissociation equilibrium. Hence, Eq. (2.5) is actually the summation
of several reactions, each with its own dissociation constant. He
observed that the exposure of a dilute nickel-cyanide complex solution
to bright sunlight does not result in any photodecomposition of the
complex.
Long (1951) studied the exchange of nickel between tetracyano-
nickelate (II) ion and hydrated nickel ion. He noticed that at pH
values ranging from 4 to 8, the exchange is slow compared to the rate
of precipitation of nickel cyanide. However, the exchange of labelled
24
cyanide
2-
between Ni(CN)4 and aqueous cyanide is fast, being completed
in 30 s at pH 6.5 to 10.5.
d) Iron-Cyanide Complexes (Fe-CN)
Iron forms two complexes with cyanide, hexacyanoferrate (II)
4-
ion, Fe(CN)6' usually referred to as ferrocyanide, and hexacyanofer-
3-
rate (III) ion, Fe(CN)6 , commonly known as ferricyanide.
Potassium ferrocyanide and ferricyanide, the otherwise highly
stable complexes, photodecompose when exposed to bright sunlight
particularly in the presence of oxygen (Broderius, 1973). The actual
decrease in the concentration of hexacyanoferrates in effluents due to
photolysis will depend on the clarity of the effluent, the extent of
contact between the effluent and the air, as well as the rate of air
movement (rate of HCN removal) over the surface of the tailings pond
(Ingles, 1981b).
Specific reports of studies on the natural degradation of pure
hexacyanoferrates solutions, or of effluents containing high concentra-
tion levels of these compounds, are relatively rare. The works of
Broderius (1973), Broderius and Smith (1980) and others, focussing on
their photolytic decomposition, provide valuable and relevant back-
ground information. In their detailed and extensive studies on the
rate and extent of photolysis of (II) and (III)
Broderius and Smith (1980) concluded that the maximum amounts of total
cyanide that can be photochemically released as HCN from these
complexes (each containing 6 mol are 85 and 49%, respectively,
corresponding to 5 moles of for hexacyanoferrate (II), and 3 moles
25
of CN- for hexacyanoferrate (III). They quote Balzani and Carassiti
(1970) in stating that under prolonged irradiation of aerated and
relatively concentrated hexacyanoferrate (II) and (III) solutions,
Fe(OH)3 (in alkaline solutions) and Prussian Blue (in acid solutions)
are formed. At the same time they support these authors' contention
that the overall chemical changes and the reaction mechanisms for the
photolysis of hexacyanoferrate complexes are not well defined.
The removal methods of iron-cyanide complex other than photoly-
sis do not appear to have received much attention. One would antici-
pate, for example, that in an effluent containing a relatively high
proportion of labile metal cyanide complexes (such as those of zinc or
copper), the release of the metal ions resulting from a drop in pH
would be accompanied by precipitation of the corresponding metal ferro-
cyanide (Ingles, 1981b).
Although apparently never investigated, it is possible that
ferro cyanide could be adsorbed on constituents of the tailings solids.
For example, Gaudin (1932) and Taggart (1945) have both indicated that
soluble prussiates (hexacyanoferrates) tend to form relatively stable,
insoluble, polar films of copper ferrocyanide or ferricyanide on copper
sulphide particles.
Hendrickson and Daignault (1973) studied the use of heavy metal
ions to precipitate hexacyanoferrates for removal by precipitation or
centrifugation. The metal ions tested were Fe
2
+ Mn
2
+ ~ . 2 Zn
2
+ and
, '''''''
Cd
2
+. Copper and zinc were the most effective for precipitating hexa-
cyanoferrate (II). Hexacyanoferrate (III) did not precipitate well by
26
itself. Ferrous iron gave 90% removal but the other metals only
removed aboll;t 60%. However, with both hexacyanoferrates present,
precipitation of 99.5% or better in both forms was obtained with Fe
2
+,
Cd
2
+ and eu
2
+.
2.4 Removal of Cyanide by Natural Degradation
For the removal of cyanide from gold mill effluents, the oldest
treatment practiced to date has been natural degradation.
The natural degradation process consists of impounding
cyanide-bearing waste solutions in large shallow ponds, thus allowing
the hydrogen cyanide to escape. Depending upon present cyanide species
and physicochemical conditions, such interactions as hydrolysis, disso-
ciation, volatilization, photodegradation, chemical and bacteriological
oxidation, and precipitation-dissolution may be important.
discussed in the following sections.
2.4.1 Volatilization
They are
The rate of volatilization of HCN from a pond is dependent on
the pH value of the waste and on prevailing atmospheric conditions. In
general, decreased pH, increased temperature, reduced liquid-depth to
surface-area ratios and increased turbulence will accelerate volatili-
zation.
Taking into consideration the fact that atmospheric conditions
vary widely, and that the cyanide-containing gold mill wastewater is
27
strongly alkaline, it would be difficult, but not impossible, to
predict the time necessary for complete removal of cyanide (Dodge and
Reams, 1949).
Since the equilibrium reaction between HCN and ~ is so pH
dependent (Eq. 2.1) the presence of air and CO
2
(air normally contains
300 g/m
3
of CO
2
) will have an effect on the formation of HCN (Clevenger
and Morgan, 1916). '!he pH will decrease as the absorption of carbon
dioxide from the air occurs and carbonic acid is produced by hydroly-
sis:
Carbon dioxide will convert the hydroxide alkalinity as rapidly as it
is formed to bicarbonate alkalinity. The reactions are:
pH = 8.3
Due to the formation of the bicarbonate-carbonic acid buffer
system, further decreases in pH below 8.3 will be relatively slow.
-
However, at equi-molar concentrations of H2C03 and HC0
3
the pH would be
6.5, and small changes in pH In this range (8.3 to 6.5) result in large
changes in the HCN/CNI ratio (Milne, 1950a) (see Figure 2.2).
28
The carbonic acid produced by dissolving CO
2
from the air in
the water will also furnish hydrogen ions necessary to the formation of
hydrogen cyanide.
The pH value can be used as an indicator of the percentage of
total cyanide which has been converted into hydrogen cyanide. The
concentration of cyanide in solution determines the amount of volatile
HCN produced and consequently the partial pressure of HeN at the inter-
face between the liquid and the surrounding gas. The partial pressure
is related to the concentration of HeN in the liquid by an equilibrium
constant (H) called the Henry's Law constant (Dodge and Zabban, 1952).
H may be independent of concentration in a very dilute solution and of
total pressure, provided the latter is nearly atmospheric; however, it
is affected by the temperature of the solution.
Hydrogen cyanide in solution will maintain an equilibrium with
atmospheric HeN. In general the amount of contact required to bring
two phases into equilibrium is dependent on the rate of mass transfer.
The rate at which mass is transferred between phases (in this case gas
and liquid) is controlled by the driving force for mass transfer, the
resistance to mass transfer and the interfacial area between phases
(Perry, 1976) according to:
where = mass transfer rate;
29
~ ~ - liquid and gas phase mass transfer coefficients;
X, Y - bulk liquid and gas mole fractions;
Xi' Y
i
- liquid and gas mole fractions at the interface.
Looking at the problem from the viewpoint of chemical-
engineering unit operations, the objective to be achieved in the
cyanide removal process, based upon release of cyanide as hydrogen
cyanide would be to bring about intimate contact first between cyanide
in the pond and atmospheric pollutants (such as CO
2
) to decrease the pH
and promote the formation of HCN in solution and then between the
solution and air to volatilize HCN.
In batch operations, volatilization of cyanide may occur in two
ways:
i) By keeping the solution stagnant, in which case the solu-
tion is exposed to the surrounding air and the volatile solute is
allowed to escape into the atmosphere. As soon as the surface of the
solution becomes depleted of HCN the rate of removal of cyanide from
the solution by volatilization becomes entirely dependent on the rate
of the molecular diffusion of HeN (Palaty and &rokova-Jakubu, 1959;
Lur'e and Panova, 1962).
When working with stagnant solutions wind effects and
stratification have to be taken into consideration.
ii) By aerating the solution with a gas (usually air) which is
introduced into the solution in fine bubbles. Even gentle aeration of
alkaline cyanide wastes will increase the mass transfer rate and cause
30
31
a decrease of the pH of the solution by natural recarbonation with a
concomitant shift of the HCN/Clf" equilibrium to favour the formation
and volatilization of HCN (Doudoroff et aI, 1966). The difference
between the partial pressure of hydrogen cyanide (which is in
equilibrium with the concentration of HCN in the solution) and that in
the gas used for stripping is one of the factors determining the rate
at which the hydrogen cyanide is stripped from the solution. Stripping
can only occur if the former partial pressure is higher than the
latter, and the rate increases as this difference increases (Dodge and
Zabban, 1952).
In all the experiments to date carried out on aerated solu-
tions, air was used as a stripping agent. This improved the contact
between atmospheric CO
2
, the alkaline compounds in solution and HCN,
decreased the pH, enhanced hydrolysis and brought about loss of HCN
(Lur'e and Panova, 1962).
2.4.2 Photodegradation
Photodegradation occurs naturally in any shallow pond or lagoon
where light can penetrate. Factors such as light penetration, affected
by turbidity, colour, depth and many other parameters, will determine
the impact of photodegradation in a specific body of water. Factors
which enhance photolysis also enhance HCN removal. For example, water
turbulence will enhance mixing and expose more metallo-cyanide
complexes to sunlight. This will enhance volatilization of HCN by
increasing the exposed surface area and the rate of mass transfer
32
(Ecological Analysts Inc. 1979). Numerous metallo-cyanide complex ions
are relatively stable in aqueous solution in the absence of ultraviolet
and visible light. However, under certain conditions, photodecomposi-
tion, with subsequent release of cyanide ions, will occur.
Ultraviolet (UV) and visible radiation cause the decomposition
of complex iron cyanides, especially in the presence of disso1 ved
oxygen (Burdick and Lipschuetz, 1950). After prolonged irradiation,
HeN is released from its iron-cyanide complex. Iron may precipitate in
basic solution as ferrous hydroxide (a white precipitate), or as ferric
ferro-cyanide (Prussian Blue) in acidic solutions (Broderius, 1973).
The pH, temperature and cyanide concentration all have a varying
affect on the photolysis reaction of both iron-cyanide complexes
(ferro- and ferri-). The lower the concentration of the iron complex
ion, the faster the rate of photodegradation (Broderius and Smith,
1980).
The rate of volatilization of HeN from the surface of waters
containing iron-cyanide complexes subjected to photolysis was found to
be approximately first order. Half-lives for hexacyanoferrate (II)
solutions containing 100 IJg/L eN of 50 and 18 min were observed for
fall and summer mid-day conditions, respectively, for a lake surface in
Minnesota. In comparable (III) solutions, the mid-day
half-life ranged from 160 min in late fall to a minimum of about 64 min
in mid-summer (Broderius and Smith, 1980).
2.4.3 Chemical oxidation
Chemical oxidation of cyanide using different oxidizing agents
is a commonly employed treatment method for cyanide-bearing wastewaters
but little data are available on the chemical oxidation of free cyanide
under natural conditions.
Cyanide can be oxidized to cyanate by ozone, which is produced
in water from ultraviolet irradiation by sunlight. Once cyanate is
formed it will not revert to free cyanide, but because of its extreme
instability it will decompose to ammonium ion and carbon dioxide (Amex
of Canada Ltd., 1980; Resnick et al., 1958; Hendrickson and Daignault,
1973).
It is thermodynamically possible to oxidize free cyanide with
molecular oxygen to yield cyanate, but unless catalyzed the reaction is
impracticably slow. For example, Cu, activated carbon, enzymes,
photo-redox reaction or microorganisms can be used as catalysts.
(Pol 'kin and Krylova, 1979; Weber and Corapcioglu, 1981; Schmidt et
a!., 1981).
2.4.4 Microbial oxidation
Many microorganisms are resistant, or can induce resistance, to
cyanide by degrading it to nontoxic products (Knowles, 1976). It has
been shown, primarily through studies of biological wastewater treat-
ment systems, that bacterial, fungal and algal species can convert
cyanide under either aerobic or anaerobic conditions to carbon dioxide
33
34
and ammonia, nitrite or nitrate (Towill et al., 1978; Brunker, 1980;
Ludzack and Schaffer, 1960).
The need for gradual acclimation of the system is usually
stressed as well as an additional source of metabolic energy such as
dextrose, glucose, etc. (Howe, 1965; Raef et al., 1977).
The fact that cyanide can be removed from various effluents by
biological treatment has been confirmed on many occasions. Murphy and
Nesbitt (1964) found that 2/3 of the cyanide carbon used was incor-
porated into bacterial cellular material and 1/3 was used for cell
respiration.Kunz et al., (1978) still had doubts whether the removal is
caused by bio-adsorption or by bio-oxidation.
Cyanate and thiocyanate can be broken down by bacteriological
treatment. The products of decomposition are incoorporated into
cellular constituents within the organism as either a nitrogen or
carbon source or both. It has been proven that CNO- was more difficult
to treat biologically than CN""" am CNS- am that the process is very
sensitive to changes in pH and temperature (Ludzack and Schaffer,
1960).
The extent of microbial degradation of cyanide in mine tailings
ponds is not well documented. Gold mill effluents are not only domi-
nated by high concentrations of cyanide but they usually contain
cyanate, ammonia, thiosalts, thiocyanate, and high concentrations of
metallo-cyanide complexes which may not be amenable to bacterial attack
as well as toxic chemicals such as heavy metals (Cu, Zn, Ni, Fe). In
addition, they are typically initially characterized by a high pH
35
(>10), low nutrient levels, low number of microorganisms and high
concentration of suspended solids.
The initial loss of cyanide from the cyanide-containing wastes
in these ponds has been attributed to volatilization because it was
found to be too rapid for strictly biological activity (Ludzack et al.,
1951; Tarzwe11, 1955).
After a prolonged period of time (some ponds have retention
times of one to two years) the degradation of free cyanide could be
enhanced by some biodegradation because the stagnant solution in the
pond is bound to accumulate some microorganisms from the air, soil or
plants. In order to enhance the process, some researchers have
suggested the enrichment of the effluent organic substances by growing
water plants (e.g., bulrushes) and the introduction of organic carbon
material (such as molasses), and the maximization of oxygen avail-
ability by aerating the pond (I1yalemdinov et al., 1977).
2.4.5 Thiocyanate formation
Cyanide can react with po1ysulphides or thiosu1phate to form
thiocyanate. Both polysu1phide and thiosulphate are oxidation products
of sulphides which are abundant in minerals (Amax of Canada Ltd, 1980;
Luthy and Bruce, 1979).
If both CN""" and CNS- are present in solution, the metallo-
cyanide complexes will exist in solution preferentially to the
respective thiocyanate complexes because the formation constants for
36
metal-thiocyanide complexes tend to be much lower than the
corresponding metal-cyanate complexes (Dean, 1973).
Many of the metal-thiocyanate complexes are insoluble,
therefore, they will precipiate. This mechanism might play a role in
the removal of Zn, Ni and Cu in effluent holding ponds (Schmidt et al.,
1981).
2.4.6 Cyanide in soil
Some metal cyanides in gold mill effluents remain in solution,
while others adsorb onto solids and settle from solution (lEC, 1979;
Ingles, 1981a). Li ttle information is available on the movement of
either free or complex cyanides at low concentrations through soils or
tailings (Renn, 1955).
An EPA laboratory study of cyanide movement in different soils
(Alesii and Fuller, 1976) showed that some cyanides are quite immobile
while others may be more mobile. Ferric ammonium hexacyanoferrate was
apparently very immobile. Cyanide as
3-
Fe(CN)6 and ~ in water were
found to be very mobile in soils. Cyanide as KCN in a natural landfill
leachate was found to be less mobile. Soil properties such as low pH
and percent free-iron oxide, and kaolin, chlorite, and gibbsite type
clays, tend to increase the attenuation of cyanide in the three forms
tested. High pH, presence of free CaC03, low clay content and the
presence of montmorillonite clay tend to increase the mobility of the
cyanide forms (Alesii and Fuller, 1976).
37
The movement of cyanide in soil is usually quite limited
because it is either complexed by trace metals or metabolized by
various microorganisms (Towill et al., 1978).
2.4.7 Conclusion
A review of the literature shows that cyanide in an aquatic
environment is not persistent. Very little information is available in
the literature on the fate of cyanide in gold mill tailings ponds.
In general, the oxidation of cyanide with atmospheric oxygen is
possible but, unless catalyzed, the reaction is impracticably slow. If
it comes in contact with some oxidizing agents (e.g., chlorine, ozone),
cyanide may be oxidized to a harmless level, but these oxidizing agents
are not at hand in any of the mine tailings ponds.
Microbial degradation of cyanide in mine tailings ponds is not
well documented in the literature. Microorganisms can convert cyanide
(up to 200 mg/L) to C02, ammonia, nitrite and nitrate, in biological
wastewater treatment systems, particularly if they are acclimatized to
cyanide and there is an additional source of metabolic energy present.
The possibility of microbial degradation in normally operated tailings
ponds upon release of the mill effluent is remote because of the high
levels of cyanide, heavy metals, pH and suspended solids, and the low
number of microorganisms and low nutrient levels, found in the
effluent. Therefore, microbial degradation cannot be considered as a
removal mechanism during the initial period when natural degradation of
38
cyanide seems to be most rapid. After a prolonged period of time the
number of microorganisms may increase, but their effect on cyanide will
depend upon their ability to oxidize it and break down metal cyanide
complexes.
The only information that is well documented is the ability of
free cyanide to volatilize in an aquatic environment. Basically, most
of the potential cyanide degradation or removal mechanisms apply to
free cyanide. Thus, it is necessary that the metallo-cyanide complexes
dissociate or break down to the metal and free cyanide components
before cyanide removal from an aquatic system can occur. Regardless of
its form, simple or complexed, cyanide has to be treated as a potential
source of HCN. After the initial quantity of free cyanide is released
from a solution, the remaining cyanide forms will be converted into
HCN, the volatile form of cyanide, as a function of the specific condi-
tions occurring, e.g., sunlight, increased temperature, decrease in pH,
presence of air.
Based on the above considerations, the major focus of this
research was concentrated on volatilization and the metallo-cyanide
complex dissociation (metal cyanide decay) as most important mechanisms
for the removal of cyanide from the gold mill tailings ponds, keeping
in mind all the other possible mechanisms.
3 ANALYTICAL METHODOLOGY
Cyanide terminology is defined in Appendix A and used consis-
tently throughout this thesis. All chemical analyses for this project
were carried out by the staff of the Laboratory Services Section at the
Wastewater Technology Centre (WTC).
3.1 Cyanide Determinations
Samples taken during each experimental run were collected in
polyethylene bottles, preserved with NaOH to a pH of 12, stored at 4C
and analyzed at the earliest possible time. The samples from all the
solutions, except the iron cyanide complex solutions, were analyzed for
total cyanide by the picric acid method (Appendix B.l). The picric
acid method recovers all cyanides in the solution except iron cyanides
(Conn, 1981). In this method, reagents are added to the sample
directly and a colour develops which is directly related to cyanide
concentration. Although with standard solutions the method was found
to be both accurate and precise to levels of approximately 0.2 mg/L ~
a concentration of 1 mg/L was adopted as a lower limit in this research
for application of this method.
Below the 1 mg/L level the distillation method was applied
(Appendix B. 2) Iron-cyanide complex solutions were also analyzed by
the distillation method.
39
40
A modified version of the conventional acid distillation
technique, using hydrochloric acid and hydroxylamine hydrochloride as
reagents at pH 2.0 is the more accurate analytical procedure for total
cyanide than the picric acid method used at WTC (Conn, 1981). The
cyanide measurements after distillation are either colorimetric or
titrimetric, depending on the concentration of cyanide. The inter-
laboratory standard deviation was determined to be 0.01 mg/L at the
0.1 mg/L level for synthetic samples and 0.1 mg/L at the 2.6 mg/L level
and 0.003 mg/L at the 0.030 mg/L level for four real samples.
The same distillation procedure is used for the determination
of weak acid dissociable cyanides (all but iron-cyanides) except the
reagents used in this case are an acetic acid - sodium acetate solution
buffered at pH 4.5 and zinc acetate. The purpose of the zinc acetate
is to prevent decomposition of any ferrocyanide present. Cyanide has
been reported to be totally recovered from zinc and nickel complexes
but only about 70% recoverable from copper complexes and 30% recover-
able from cadmium complexes. No recovery is reported from ferro, ferri
and cobalt complexes and thiocyanate is not an interference. The
interlaboratory relative standard deviation on an actual effluent
sample was determined to be 5.5% at a cyanide concentration of 2.5 mg/L
crr.
3.2 Cyanate and Thiocyanate Determination
Cyanate hydrolyses to ammonia when heated at a low pH. This
phenomenon is used in the cyanate determinations. The ammonia content
41
Is measured before and after hydrolysis of cyanate using
electrodes. The cyanate results have, for a 10 mL
aliquot, a detection limit of 0.3 mg/L with a 3% precision at
100 mg/L.
Thiocyanate forms an intense red color with ferric ion at an
acidic pH, which is suitable for colorimetric determination. The thio-
cyanate results have, for a 10 mL aliquot, a detection limit of
1 mg/L. The precision, depending on the sample concentration, is as
follows: 3.7% at 7 mg/L ; 0.6% at 35 mg/L and 0.8% at 70 mg/L
More detailed descriptions of the methods are available in "Standard
Methods" (APHA-AWWA-WPCF, 1980).
3.3 Metals Determination
Samples for the determination of metals were collected in
acid washed polyethylene bottles preserved with nitric acid (pH -2")
and stored at 4C. All metal analyses were performed by atomic
absorption (AA) flame analysis (APHA-AWWA-WPCF, 1980). The detection
limits, working range and precision for the results obtained by AA are
as follows:
Detection Limits (mg/L) Linear Working Range (mg/L)
Cu 0.01 0.005 0.01 - 10.0
Zn 0.01 0.005 0.01 - 3.0
Nt 0.10 0.02 0.10 - 10.0
Fe 0.10 0.02 0.10 - 7.0
42
The precision varies with the sample concentration and its
matrix. Precisions at the top end of the linear working ranges for Cu,
Zn, Ni and Fe are: l%, l%, 3% and 3%, respectively.
4 SCREENING EXPElUKENTS
The review of the literature revealed that the processes of
volatilization of HCN from cyanide solution and the dissociation of
metallo-cyanide complexes were dependent upon many factors, such as pH,
temperature, ultraviolet irradiation, contact with air (aeration),
initial cyanide concentration, metal content, depth of the solution,
etc. In this research, the three factors that were considered to be
the most important were selected: temperature, UV light and aeration.
The effect of depth upon irradiation was taken into consideration later
when UV irradiation rate constant was calculated, and upon other
factors was eliminated by having the solutions well mixed. The assump-
tion was made that the pH effect on metallo-cyanide complexes was
insignificant and only its effect on HCN/Crr ratio was considered late'r
in the modelling process.
The following sections explain how the temperature, UV light
and concentration ranges used in this research were selected and
present the results of the preliminary tests that were carried out in
order to determine suitable pH and air flow levels for the experimental
design.
4.1 The Effect of Temperature, UV Light and Concentration
In order to examine the effect of temperature, all bench scale
experiments were carried out at 4C and 20C. The 20C temperature was
4'3
44
selected to be representative of warmer conditions and was easily
obtained as a set room temperature. The 4C temperature was chosen to
be representative of cold weather conditions at many Canadian gold
mines and was also the lowest temperature that the equipment available
in the laboratory (cold water bath) could handle.
Considering that iron commonly occurs in the mill effluents and
that an iron-cyanide complex solution is photodegradable (see Sec-
tion 2.4.2), the effect of UV irradiation had to be examined to repro-
duce the sunlight effect. Samples tested under UV light were exposed
to an intensity of 2 mW/cm
2
; this intensity of sunlight is typical of
southern Ontario in summer for the 300 to 400 nm range (Shaw, 1983).
Since the experiments were carried out in graduated cylinders made of
2 mm thick Pyrex glass, it filtered out all radiation below 300 nm
(Table 4.1). The range of UV irradiation from sunlight was reproduced
with B-I00A Black Ray Ultraviolet lamp (wavelength 366 nm) mounted ~
cm above the cylinder (Figure 4.1). The intensity was measured with a
"Black Ray" Ultraviolet Meter, Model J-221, with absolute accuracy of
15% and a precision of 5%.
In order to examine the effect of concentration, the fact that
the quality of the actual mill effluent would depend on the type of ore
available, the treatment used on that ore, and chemicals applied in the
process has to be taken into consideration. It would be difficult to
carry out laboratory experiments on one specific mill effluent and
apply the findings to others. Therefore, it was decided to produce a
TABLE 4.1. APPROXIMATE VAV!LDGTB LIMITS POR t'RABSKISSIOR 01' VAlUOUS
OPTICAL MATERIALS JOWt IWOK 'l'EMPEUTUU
Material '1'hickness. 1IIIl Approx. >. for cutoff. A
Window glass (standard) 1 3070
3 3140
10 3300
Optical (white crown) glass 1.8 3090
Pyrex (Corning 774) 1 2800
2 2970
Source: Shaw. 1983.
45
46
"synthetic effluent", i.e., to emulate the general characteristics of
an actual gold mill effluent and use this solution in the experiments
performed under controlled laboratory conditions.
The important metals found in the mill effluent at the selected
gold mine complexed about 17% of the cyanide present. Based on the
large amount of data available on this effluent, the individual metals
most often found in the mill effluent typically tied up: Cu about 4%,
Zn 8%, Ni 2% and Fe 3% of the available total cyanide.
Therefore, the level of 17% was chosen (Le., metal cyanide to
total cyanide ratio of 0.17) for sodium cyanide and the selected single
metallo-cyanide solutions, which were:
i) Sodium Tricyanocuprate
ii) Sodium Tetracyanozincate
iii) Sodium Tetracyanonickelate
iv) Potassium Hexacyanoferrate
Na.2 Cu( CN)3
Na2Zn(CN)4
Na.2 Ni ( CN)4
K
3
Fe(CN)6
The four particular metallo-cyanide complexes were selected for
the following reasons:
i) Upon addition of cyanide in the cyanidation process, copper
sulphides are readily oxidized to sulphates and eu
2
+ is released. When
cyanide is oxidized to cyanate, eu
2
+ is reduced to
1972). Therefore, of all the complex ions formed,
eu
1
+ (Fairbridge,
2-
Cu(CN)3 is con-
sidered the most probable to occur in the mill effluents (Hedley and
Tobachnick, 1958).
ii) Zinc's chemical behaviour is dictated by the existence of
a single oxidation state, i.e., Zn
2
+. Zinc minerals are sufficiently
47
soluble under normal cyanidation conditions. Zinc forms complexes in
aqueous solution but these are never particularly stable or inert
(Fairbridge. 1972). During the reactions involved in gold precipita-
tion. the greater part of the zinc dissolves in the cyanide solution
and forms various complexes. the most common being Na
2
Zn( CN) If (Hedley
and Tobachnick. 1958).
iii) The chemistry of nickel is not simple because of the
existence of the variety of its structures (octahedral. tetrahedral and
square) (Cotton and Wilkinson. 1962). Although nickel can achieve
various oxidation states (see Section 2.2.4) the majority of its
compounds are of the Ni(II) oxidation state (Nriagu. 1980). Since the
formation of tetracyanonickelate (II) complex is a very fast reaction
(a fifth-order reaction) and this complex is considered to be
thermodynamically very stable. this is the nickel-cyanide complex most
likely to occur in the mill effluents (Kolski and Margerum. 1968).
iv) Cyanide reacts with iron (II) and iron (III) to form
extremely stable ferro- and ferricyanide complexes. respectively
(Broderius. 1973). The common form in the cyanidation process would be
ferrocyanide but in an aquatic environment ferrocyanate would fairly
easily oxidize to ferricyanide (Ingles. 1981b).
The simple sodium cyanide solution and these metallo-cyanide
solutions were tested in anticipation that the results would assist in
the interpretation of the data from more complex solutions.
For more complex solutions low and high mixes were tested. The
"low mix" contained 200 mg/L of TeN. of which 17% was in a complex with
48
the four metals mentioned above according to their exact stoichiometric
proportions. The "high mix" was oversaturated with the same number of
metals, hence, it did not contain any free cyanide (calculations were
performed assuming that 117% of available cyanide was complexed with
these four metals).
In order to determine suitable conditions for the experimental
design, several preliminary tests were carried out. The results were
used to delineate the choice of buffer and pH level, and to examine the
effects of aeration and CO
2
on cyanide removal.
4.2 Selection of Buffer and pH Level
Objective - Molecular HCN and the cyanide ion CN'"" (the free
cyanide), exist in water in a definite relationship dependent upon the
pH of the system (Eq. 2.1). Since the pH of cyanide solution changes
when open to the atmosphere, a buffer has to be used in order to
maintain the pH at the certain level. The objective of this experiment
was to determine the type of buffer which could be used for the
experiments with simple and metallo-cyanide complex solutions. The
pH's that were considered were 7.0, 9.0 and 11.0. From Figure 2.2 it
can be seen that at pH 7.0 approximately 99.5% of the free cyanide in
solution is present as HeN, about 64% at pH 9.0 and about 5% at pH of
11.
Conditions - For safety reasons all the experiments were
performed in a fume hood. All the preliminary tests were carried out
at room temperature (20C) and under fluorescent light. The conditions
of the test are presented in Table 4.2, and the setup used is
illustrated ~ Figure 4.1. All cyanide solutions were made on the day
of the experiment. The air flow rate used in this test was
1000 cm
3
/min (arbitrarily chosen).
Discussion - The results of the experiments are presented in
Table 4.3. These show that at pH 7.0 the rate of reaction is the
fastest and as the pH increases the reaction rate slows down substan-
tially. There were some difficulties encountered in maintaining the pH
at 11.0. During the latter part of the tests (after about 17 h) NaOH
had to be added intermittently to maintain the pH constant. Because of
these difficulties the results of the second set of tests are not
available.
The equation used to calculate buffering capacity for the
phosphate buffer (pKa - 7.2) was
pH - pKa + log (Freiser and Fernando, 1963)
The calculated concentration of the phosphate buffer (see Table 4.2)
had a sufficient buffering capacity to maintain the pH at the desired
level of 7.0.
49
50
TABLE 4.2. THE TEST CONDITIONS FOR BENCH SCALE EXPERIMENTS
Apparatus (Fig. 4.1)
Sample Source
Buffers
* Surface area.
** Total cyanide.
Graduate cylinder (2000 mL) (As* - 47.78 cm
2
)
pH meter
Fritted glass diffuser
Air flow meter
Glass impinger (water)
} for aerated samples
Glass impinger (IN KOH) - used for C02 trap
UV lamp - used for conditions with UV light
2000 mL NaCN (100 mg/L TCN**)
2000 mL Na2Cu(CN)3 (100 mg/L TCN)
For pH 7.0
For pH 9.0
For pH 11.0
0.25M KH
2
P0
4
and 0.16M Na2HP04
O.lM H
3
B0
3
and O.lM NaOH
0.05M Na2HP047H20 and O.lM NaOH
mpH
nwter
0
0
0
0
0
0


,.
0
0
0
0
0
0
.0



..
elr flow
!!Ieter
Figure 4.1 'l'be Setup for Bench Seale Experi.ent.
51
eir
distilled
.. ter
52
TABLE 4.3. THE EFFECt OF pH ON THE RATE OF CYANIDE RKMOV AL (CHOICE OF
BUFFER)
pH 7.0 pH 9.0 pH 11.0
0.25M K H
2
P0
4
O.IM H
3
B0
3
0.05M Na2 HP0
4
0.16M Na
2
HP0
4
O.IM NaOH O.IM NaOH
Cyanide
k (h-
1
) k (h-
1
) k (h-
1
)
Solutions
t
l
/
2
(h) t
l
/
2
(h) t
l
/
2
(h)
v v v
NaCN 0.2493 2.8 0.0776 8.9 0.0155 48.8
0.2512 2.8 0.0858 8.1 NA NA
Cu-CN 0.0356 19.5 0.0035 197.8 0.0014 498.1
0.0328 21.1 0.0020 340.7 NA NA
- not available. NA
k
v
- volatilization rate constant calculated assuming first order
reaction.
half life.
53
In order to check if the
3-
P0
4
ligand had any effect on the
removal reaction
3-
of cyanide (i.e., if P0
4
was competing with eN""" for
metals) parallel experiments were carried out using either phosphate
buf fer or 1.2M HCl for pH con trol. The results are presented in
Table 4.4 and they show the same rate of cyanide removal in both
experiments.
Conclusion - Taking into consideration the ease of running the
test at pH 7.0 and the fact that the rate of reaction was the fastest
at that pH (Table 4.3), it was decided to run future tests at pH 7.0.
From the results presented in Table 4.4 the conclusion is that
phosphate buffer had no effect on cyanide removal - same rates of
3-
reaction involving Cl- and P04
4.3 The Effect of Aeration
Objective - To determine if the presence of air had any effect
on cyanide removal rate and, if so, to decide on the air flow rate to
be used in subsequent experiments.
Conditions - For consistency the same setup was used for the
experiments with and without air addition (Table 4.2). In the
experiments with air addition, two air flow rates were used:
Observations - Since the non-aerated samples were stagnant for
the duration of the test, it was necessary to stir the solution before
every sample was taken to avoid stratification effects. Cyanide
concentrations in the layer closest to the surface always were lower
TABLE 4.4. EFFECTS OF a- And PO!- ON CYAHIDE REMOVAL RATE
54
Cyanide Solutions kv (h-
1
) tI/2 (h)
NaCN 1.2M HCl 0.2638 2.6
2-
H
2
P0
4
and HP0
4
0.2512 2.8
Na2Cu(CN)3 1.2M HCl 0.0101 68.6
- 2-
H
2
P0
4
and HP0
4
0.0112 62.0
than the average cyanide concentration. Hence, there was some incon-
sistency in .the results obtained on non-aerated samples.
There was a disadvantage in using an air flow rate of
1000 cm
3
/min under laboratory conditions. The sample splashed around
and the evaporation was 5% over 24 h. Compared to this, the evapora-
tion of 2% per 24 h in case of 100 em
3
/ min air flow was considerably
lower, and in addition there were no problems with splashing during
these tests.
Discussion
i) Air versus no air - From the results shown in Table 4.5 it
is apparent that the presence of air had a definite effect on cyanide
removal. Much longer times were required for cyanide removal from
stagnant solutions.
ii) 1000 cm
3
/min versus 100 cm
3
/min - Cyanide from the solu-
tions receiving the higher air flow rate was removed faster (about four
times faster from NaCN and two times faster from Cu-CN solution) than
from the solution with 100 cm
3
/min of air.
Conclusions - The tests show that the presence and rate of
availability of air have significant effects on the removal of
cyanide. The only dilemma was whether air contributed only to
stripping off cyanide from the solution, whether there was oxidation of
cyanide going on, or a possible reaction with CO
2
from the air. To
determine if cyanide oxidation to cyanate (CNO-) was occurring,
chemical analyses were performed and in all the cases CNO- concentra-
tions were <0.1 mg/L. Therefore, it is assumed that at the applied air
55
56
TABLE 4.5. l'BE EFFECT OF AERATION
AIR.
NO AIR
1000 (cm
3
/min) 100 (cm
3
/min)
Cyanide
Solutions k
t1/2
k
t1/2
k
t1/2
v v v
(111)
(h)
(h-
1
)
(h)
(h-
1
) (h)
NaCN 0.2493 2.8 0.0686 10.1 0.0228 30.4
0.2512 2.8 0.0597 11.6 0.0269 25.8
Cu-CN 0.0356 19.5 0.0211 32.8* 0.0033 210.0*
0.0328 21.1 0.0402 17.2 0.0086 80.6
* Inconsistent result.
Experiments were carried out at the pH of 7.0.
57
flows and for the duration of these experiments (4 to 30 h) there was
no CNO- produced.
4.4 The CO
2
Effect
Objective - To see if there was any effect from CO
2
in the air,
causing a decrease in pH in spite of the phosphate buffer present in
the solution.
Conditions - The conditions of the test are described in
Table 4.2. One solution was examined having a CO
2
trap (IN KOH) on the
air line, and the other solution was without CO
2
trap. The length of
test runs was 103.6 h.
Observations - The results frem this test are presented in
Table 4.6. After four days the pH in both solutions went through
essentially the same change (1.e., decreased from 6.94 (6.92) to 6.86)
indicating that there was a reduction of 1% in the pH under the applied
conditions. Because the permitted change in the experiments was
decided to be 0.5 at pH of 7.0 (7%), this reduction was negligible.
Conclusion - Since the oxidation of cyanide to cyanate was not
detected and there was a negligible effect of CO
2
on the cyanide
solution pH level for the duration of these experiments, the positive
effect of aeration was attributed to agitation of the cyanide solution
and the stripping off of free cyanide.
TABLE 4.6. THE EFFECT OF CO2 ON THE pH READINGS
Time (h)
0.0
3.7
7.1
25.7
30.9
50.7
55.2
95.9
103.6
Experiment 111
(without CO
2
trap)
pH
6.94
6.94
6.95
6.94
6.93
6.91
6.86
6.84
6.86
58
Experiment fI2
(with CO
2
trap)
pH
6.92
6.92
6.91
6.88
6.91
6.86
6.86
6.85
6.86
5 EXPERIMENTAL PROCEDURES AND ANALYSES OF EXPERIMENTAL RESULTS
5.1 Introduction
From the seven variables found in the literature to be affect-
ing the process of natural degradation of cyanide (pH, temperature, UV
light, contact with air, surface area versus depth of pond, initial
cyanide concentration and metal content) only three, considered to be
the most important factors, were selected to be examined under
previously determined pH of 7.0. These are: UV light, aeration and
temperature.
5.2 Full 2
3
Factorial Design
A full 2
3
factorial design was used for the three chosen vari-
ables to determine which variable had the highest effect on the rate of
cyanide degradation. This design required 8 experimental runs to
evaluate all combinations of two levels of each of the three variables
(Box et al., 1978).
Table 5.1 shows the factorial design matrix (k - 3 columns and
N - 2k rows) used for these experiments (a minus sign represents the
low level and a plus sign the high level of each variable). The
criteria used to determine the range for low and high levels was
explained in Section 4.1.
59
60
TABLE 5.1. FAC'fORIAL DESIGN MATRIX
Variables
UV Light Air Temp.
(mW/cm
2
) (cm
3
/min)
(OC)
Run No. A B C
8
6 +
7 +
5 + +
4 +
2 + +
3 + +
1 + + +
Experimental UV light Air Temperature
Settings + + +
0 2 0 100 4 20
61
5.3 Inspection of the Data
p o ~ completion of the experiments, plots of temporal changes
in total cyanide concentrations in the single cyanide solutions were
produced and visually examined. Two examples of these plots are given
in Figures 5.1 and 5.2.
The resulting graphs show that the total cyanide concentration
drastically decreased after 48 to 72 hours under all conditions. The
decrease varied between 20 to 90% for the seven cyanide solutions
examined (see Appendix C, Tables C.1 to C.14).
Sodium cyanide is a typical representative of water soluble
simple cyanides. It will ionize rapidly and completely when dissolved
in water to release cNr and form HCN and, is therefore, considered as
free cyanide. Consequently, if volatile HCN was to be released from
any of the solutions examined, it would be derived initially from NaCN,
then from the metallo-cyanide complex solutions. Therefore, it was
concluded that the initial decrease in cyanide concentration in
metallo-cyanide solutions should be compared to the rate of decrease of
cyanide in NaCN solution since it occurred relatively quickly as the
free cyanide was volatilizing from the solutions.
After the common initial decrease of the first part of the
curve, the slopes of the second part of the curves varied among
different metals, as well as within individual metallo-cyanides under
different experimental conditions. This was attributed to the
different length of time necessary for the dissociation of these
\
-

E
-


a:
;
MO
220
m m
TIME (h)
Pigure 5.1 10tal Cyanide Degradation in Sodiua Cyanide Solution
'"
N
2DO
::;
Fe-CN.20OC
AA. UV
~ o - ~ ~ ~ ~ ~ ~ ~ ~ ~ - - ~ - -
e
150
100
Zn-CN,20t
AIR. NO UV
Cu-CN.20'C
AIR. UV
Hi-CN, 20t
NOAIR,NOUV
63
00 48 12 138 -1B4 i z30 ' 'a22 0 48 12 138 184 230 276 322
Tt. (h)
Figure 5.2 Typical Results of Total Cyanide Degradation 1n Fe. Cu. Zu
and Hi Cyanide Collplex Solutions
metallo-cyanide complexes (Table 2.2) , the release of err ion and
formation of the volatile HCN.
5.4
5.4.1
Rates of Reaction
General
The objective of every study of chemical reaction kinetics is
the determination of the order of a reaction and the provision of one
or more rate equations to adequately describe the reaction at hand.
The rate of reaction and the algebra for describing it is obtained by
observing the change of concentration of reactant as a function of
time.
If the mechanism of a reaction is known it is usually possible
to postulate a rate equation and, hence, the order of reaction. The
mechanism is the sequence of elementary steps that describes how the
final products are formed from the original reactants. The rates of
the individual steps will normally differ from each other, and the rate
of the overall reaction will be determined primarily by the slowest of
the steps.
It has been found experimentally that reaction rates are
proportional to the concentrations of the reactants raised to an
appropriate power. Expressing this for the irreversible
reaction aA + bB + cC + dD the rate is (Smith, 1981):
64
65
dC
A
------
dt
(5.1)
a is the order of the reaction with respect to A and a is the order
with respect to B. The proportionality constant k, called the reaction
rate constant, is dependent on temperature and independent of concen-
tration.
5.4.2 Rate calculations
The following first-order rate reaction was applied to both
parts of the curve - the initial part caused by rapid volatilization of
free cyanide (see Figures 5.1 and 5.2), and the second part
representing the slow decrease of total cyanide concentration due to
metal cyanide decay rate:
(5.2)
66
If the initial concentration is
integration yields
1n (5.3)
This implies a linear relationship between 1n CA/CAo and t.
In order to select the cut-off point between volatilization and
metal cyanide decay, the break point on the curve, and determine the k
value for both parts of the curve the computer program BACKRAT
(Appendix D) was applied. In this program linear regression was
carried out on every set of data from Time 0 to the end of the
experiment and backwards from the last data point to Time (\. The
residual sum of squares (RSS) was calculated and its average value
monitored. When the change in cyanide removal mechanisms occurs RSS
value suddenly changes, and this point is considered to be the cut-off
point. The same check was made on the results of the average residual
sum of squares obtained using backwards regression.
When the cut-off point was determined, the kv value (rate
constant from 0 time to the cut-off point) was available from the same
forward regression analysis. In order to determine the kl value (the
rate of reaction constant for the second part of the curve) the data
set was analyzed by the RATE program (Appendix E). This program
performs linear regression from the cut-off point to the end of the
data set. The values of k's (the rate constants) determined this way
had the units of [h-
1
].
5.5 Examination of the Factors Affecting Volatilization of Free
Cyanide
The values of the volatilization mass transfer coefficients
kv (determined for all 7 solutions examined under 8 different condi-
tions) tabulated in Appendix C (Tables C.1 to C.14), were used to
construct the dot diagrams given in Figures 5.3a,b,c. The diagrams
show the general location of the volatilization mass transfer coeffi-
cient with respect to temperature (Figure 5.3a), UV light (Figure 5.3b)
and air (Figure 5.3c).
The effect of temperature on the rate of volatilization of
cyanide from all examined solutions was very pronounced (higher rates
at 20C), particularly in the reactions with aeration present
(Figure 5.3a). At 4C the reaction was generally slow.
The energy of activation E, the minimum energy the reacting
substances must have for the reaction to occur, was calculated for the
NaCN solution using the semi theoretical relationship of Arrhenius for
the specific rate constant (Perry, 1976):
k - A e-
E
/
RT
(5.4)
67
I
I
I
I
I

NaCN Cu-CN Zn-CN Nl-CN
AIR
x
x
AIR
j r
r
2
X
2
AIR M
x
>
.k:
1
x
I
1 r
1 1

X



)(

X
)(
X
X
X
I

a


4

12 18 2D 24

4

12 11 2D 24 0 4

12 11 2D 24 0 4

12 11 2D 24
TEMPERATURE DEG.C
I I
F.-CN
x
LOW MIX HIGH MIX
:2 2 2
X
2
"-
1 .l
x
l
AIR
x
1 X
)(



x
I
M
x
i
x
0 0

4

12 11 2D 24 0 4

12 18 2D 24 0 4

12 11 2D 24
TEMPERATURE DEG.C
0'
rigure 5.38 Volatilization
co
Maaa Transfer Coefficient It versus Te.,erature
y

NatH Cu-CN ZrrCN NI-CN
x
x
x
x
:2 I l- I I
X
I
l
I I
x
x
>
x
x
s
x
I
s r
1
M
M

x

1 r
I
x

x

x
x
I
D D
G I
tC1UV UV NOUV UV NOUV UV NOUV UV
PRESENCE OF UV LIGHT
I .
I
F.-CN LOW MIX
x
HIGH MIX
:2 I f- I f-
x
I
"-
! 1 l
X
1 l
X
X
X
1
X



X
I
X X
X X
X
>,C
G G
NOUV UV NOUV UV NOUV UV
PRESENCE OF UV LIGHT
0'
Pigure 5.3b Volatilization Hass Transfer Coefficient k versus UV Light
I.D
..

NatH Cu-CN Z,,-CN Ni-CN

x
x
2 I I I
X
I ..
"-
E
I I
M
0
...,
x
>
oX
I
x
I
I r
1 I



x
.1
M

X
X
X
X
I
a a
NO AIR AIR NO AIR AIR NO AIR AIR NO AIR AIR
PRESENCE OF AERATION
I I
F..-CN
X
LOW MIX HIGH MIX
2 I l-
I J..
X
I
"-
i,l
X
3E
x
1 X 1
X
X
X

X
X
X
. I
M
x
x

x
M
a 0
a NO AIR AIR NO AIR AIR NO AIR AIR
PRESENCE OF AERATION
-....J
Pigure 5.3e Volatilization Maaa Tranafer COeffieient k eraU8 Aeration
0
71
where A is the frequency factor, R is the gas constant (R =
8.314 J/oK mol), T is the absolute temperature (OK) and E is the energy
of activation of the reaction (kJ/mo1). Taking the logarithms of Eq.
5.4, gives
log k ... log A -
E
(5.5)
2.303RT
Thus a plot of log k against l/T yields a slope that represents
-E/2.303R. If the rate constants are available for only two tempera-
tures, Eq. 5.5 is applied and when the two resulting expressions are
subtracted from one another, the equation becomes (Perry, 1976):
log
-- -
(5.6)
In this research, the activation energy for NaCN was calculated by
applying Eq. 5.6, and using the rate constant values determined at 4C
(Tl - 277K) and 20C (T2 - 293K), and solved for E the results were:
Conditions
AIR - UV
NO AIR - UV
AIR - NO UV
NO AIR - NO UV
E (kJ/mo1)
21.4
25.3
70.5
65.9
72
The E value is always positive, therefore, the rate constant
always increases with temperature. A rough working guide that the rate
constant is doubled for every increase of 10C has often been applied.
In practice E may range from 4.2 to 418.4 kJ/mol or even wider and a
10C change in temperature may lead to a smaller or greater factor than
two in the rate constant (Perry, 1976). In the above case, for the
temperature increase of 16C (E = 21 to 25 kJ/mol) the factor was about
2 with UV light present, and a factor of about 5 was calculated with no
UV applied (E = 66 to 70 kJ/mol).
The effect of UV irradiation was evident on the kv of iron
cyanide and low mix (Figure 5.3b).
Considering the effect of aeration (Figure 5.3c) the reaction
rate was increased in most cases with the air being present, including
NaCN.
The summarized results of the dot diagram analyses of factors
affecting the volatilization rate of free cyanide from all cyanide
solutions are presented in Table 5.2. The + and - signs in the table
mean that the rate of volatilization was higher or lower, respectively,
when the particular factor was present in the experiment.
5.6 Comparison of Volatilization Rates
When the values for the volatilization rate constant kv were
determined for NaCN and for metallo-cyanide solutions (see Appendix C,
Tables C.l to C.14) it was possible to compare them and at the same
time check if the rate of volatilization of HCN from single
73
TABLE 5.2. FACTORS AFFECTING THE KATE OF VOLATILIZATION OF FREE
CYANIDE (RESULTS OF THE DOT DIAG1WI ANALYSES)
Temp. Air UV
NaCN
+ +
Cu-CN + +
Zn - CN + +
Nt - CN + +
Fe - CN + + +
Low Mix + + +
High Mix + +
74
metallo-cyanide solutions was within the 95% confidence limit of the
volatilization rate of HCN from a solution of NaCN. The computer
program VOLCONF (Appendix F) was used to do linear regression analyses
and calculate the 95% confidence interval on the first part of the
curve of all the data sets under the same condition, e.g., sodium
cyanide is compared to Cu, Zn, Ni and Fe cyanide complex solutions
under 20C, NO AIR, UV (example, Figure 5.4). The general observation
was that the values were within the 95% confidence level. Therefore,
the statement made previously (Section 5.3) that the first part of the
curve was attributed to the volatilization of free cyanide from NaCN as
well as from metallo-cyanide complex solutions is justified by these
results.

...

u
"
In
4ft
1
8
....
.,
...
.,
.'
c

.. ..

8


J
.,
...
cs
.,
..

..

'lIOn 5.4
a.
.. .. .. ..
......

.. .

. . . - - - -
.... ...
II _ _ _ _
.... w
Typlea1 b..,le. for 95% e.l. 011 Part of the Desr .... tloo
Con. Attributed to VolatlU .. tlOli - .CII ftr8ua Co (a). Z.
(h). It (e). ,. (d)
-
"
V'I
6 ANALYSIS OF EXPERIMENTAL DESIGN
6.1 Introduction
The importance of each factor is determined by calculating its
effect on the reaction. The effect of any factor is the change in
response Y produced by a change in the level of the factor. When a
factor is examined at two levels only, the effect is simply the
difference between the average response of all trials carried out at
the first level of the factor and that of all trials at the second
level.
The effects were examined in three ways. They are:
i) Calculated effects.
ii) Half normal probability plots.
iii) Backward elimination procedure.
The theory about three ways is described in the following
sections:
(i) Calculated effects - In these experiments the estimates of
the main effects and the two and three factor interactions -(see
Table 5.1) were calculated by taking the sum of products between the
response elements Y and the corresponding elements of the column (1 to
k) and dividing this product by N/2 e.g.,
76
2 It
Effect (I to It) - - r Y
i
N i-I
77
(6.1)
where {I to It} stands for the elements of the. 1 to k columns and the
summation is taken over all N products (Draper and Smith, 1966).
ii) Half normal probability plots - Each estimated effect has
a variance defined as:
Var (effect) - {x'x)-1 a2
where is the variance of the individual observations and x and
x' represent the design matrix and the transpose design matrix, respec-
tively (Shaw, 1983).
No replicate runs were conducted because of the large number of
experimental runs carried out (7 solutions times 8 conditions = 56) and
the length of time required for each run. Therefore, the a2 value was
determined by analyzing the half-normal probabili ty plots. The value
of the variance was estimated from the plots by considering the effects
which lie close to the straight line joining the response points. The
variables with effects that deviated most from the straight line are
considered the most significant (Daniel, 1959). The confidence
interval at the al2 level of significance for the best estimated
individual effect Si was calculated as:
78
Bi tv' a/2 { Var (effect)
where v represents the degrees of freedom and a./2 = 0.025, i.e., a 5%
level of significance.
iii) Backward elimination procedure - The backward elimination
procedure (Draper and Smith, 1966) was used as another way of verifying
the results by separating significant from non-significant variables.
This method first performs a complete linear regression using all
variables i.e., finds the least squares equation on all independent
....
variables, y - f (zl_ z2_ Z3). In order to be able to eliminate the
non-significant variables the contribution of each of the variables to
the regression sum of squares is determined. The partial F test
values, calculated for every variable treated as though each one were
the last to enter the equation, provide a measure of these contribu-
tions. The lowest partial F test value (F
L
) is always compared with a
preselected significance level F
o
If FL is found to be less than F ,
o
the variable which gave rise to FL is removed from consideration and
the calculation carried out again, until all the variables
that remain exceed the significance level F
o
6.2 Results
i) Calculated effects - For metallo-cyanide single solutions
and two mixes the response variables in the factorial design were the
reaction rate constants (k1's) (see Table 6.1), estimated as explained
1n Section 5.4.2, by using linear regression on the second portion of
. --------
the curve which was attributed to metal cyanide decay.
The response variable for sodium cyanide solution was the
reaction rate constant, k, estimated for the curve,
which was attributed to the removal of free cyanide (Table 6.1). Some
variability was observed at the end of this degradation process (see
the degradation curve in Figure 5.1 after 80 hours).
The main effects and the two and three factor interactions were
determined by using Eq. 6.1 1n the computer program FACT presented in
the Appendix G.1 (example calculation 1n Appendix G.2). The values of
the calculated effects are given in Table 6.2.
From the results in Table 6.2, temperature (C) seems to be the
largest main effect in the majority of cases. The exceptions are Fe
and low mix where UV light ("A") is the main effect.
11) Half normal probability plots - To confirm the findings of
the "calculated results", half normal probability plots were construc-
ted using the values from Table 6.2. The effects were arranged 1n
ascending order and plotted against Pi(%) - 100 (i-1/2) /m where i is
order number and m is the total number of effects. The plots are given
as Figures G.3.1 to G.3.7 1n Appendix G.3. Typical plots are presented
1n Figure 6.1.
When the half normal probability plots were visually inspected
the temperature seemed to be the main effect for NaCN. Zn and Ni
79
TABLE 6.1. REActION RATE CONSTANT AS THE RESPONSE VARIABLE
Response Variables
Test Cyanide
Code Solutions Na Cu Zn Ni Fe
I 0.0050 0.0033 0.0045 0.0027 0.0007
A 0.0139 0.0031 0.0096 0.0041 0.0040
B 0.0074 0.0038 0.0068 0.0030 0.0012
AB 0.0234 0.0036 0.0137 0.0042 0.0051
C 0.0238 0.0071 0.0194 0.0008 0.0026
AC 0.0253 0.0024 0.0181 0.0015 0.0222
BC 0.0394 0.0207 0.0062 0.0027 0.0016
ABC 0.0389 0.0153 0.0448 0.0008 0.0068
Low Mix
0.0046
0.0052
0.0046
0.0087
0.0056
0.0060
0.0047
0.0075
High ~
0.0017
0.0011
0.0013
0.0016
0.0025
0.0022
0.0026
0.0027
ex:>
o
81
TABLE 6.2. RESULTS (F 2
3
FACTORIAL IESIGN IN TERMS (F RELATIVE EFFECTS
Cyanide Solutions
Factors* Na Cu Zn Nt Fe Low Mix High Mix
A 0.0057 -0.0025 0.0123 0.0003 0.0079 0.0019 -0.0001
B
C
AB
0.0088 0.0034 0.0049 0.0004 -0.0037 0.0010
0.0180 0.0078 0.0135 -0.0020 0.0055 0.0002
0.0006 0.0032 0.0104 -0.0007 -0.0034 0.0015
0.0002
0.0011
0.0003
AC -0.0068 -0.0023 0.0063 -0.0009 0.0044 -0.0004 0.0000
BC 0.0029 0.0029 0.0018 0.0002 -0.0045 -0.0007 0.0001
ABC -0.0031 0.0032 0.0095 -0.0006 -0.0038 -0.0003 -0.0001
* A - UV light.
B - Aeration.
C - Temperature.
Single letters represent one factor effects and multiple letters
represent interaction effects.
HII.CulCH1.
..
/
1BoP.

...
.NIt
-
.NR,Tf!NP.
.u
I-

::
F.M
/
.MI,""
-a.o .... -U-GIOCWDOOI.'"
EfFECTIa1D
./
i:./
/'
,.M/
.1BIP.
.lI\( 1BIP.
7.M ,
-4010 DO 1.0 4.D e.o eo
lFFECTSa1D
82
Figure 6.1 Balf Roraal Probability Plots (Cu. ID. 111 aDd Fe Cyanide
So lutioaa)
cyanide solutions and the high mix. Temperature combined with air, had
some effect on Cu cyanide solution, while UV light combined with air
affected the low mix. UV light showed a large effect on the iron
cyanide solution and low mix. By taking the probability transform* as
X and the effects that lie close to the straight line section as Y and
applying a linear regression analyses, the estimates for the variances
and the 95% confidence intervals on random effects were determined
(Appendix G.4). Most factors found by visual inspection to have high
effects, were proven significant at 95% CI. Only factors having an
effect on zinc cyanide solutions were not significant at 95% signifi-
cance level.
iii) Backward elimination procedure - The results of the
backward elimination procedure are presented in Table 6.3. In order to
carry out the backward elimination calculations the highest order
interaction (three factor interaction) was taken as the estimate of
error variance, due to the fact that there were no repeated runs in
these experiments, and used for testing the significance of the other
mean squares. The computer program LJUB7, using the backward elimina-
tion procedure as explained in Section 6.1, is applied and with the
* "Probability transform" term used in this thesis refers to the linear
units (mm) read off the probability graph in order to be able to
perform linear regression calculations.
83
84
TABLE 6.3. cntPLETE RESULTS OF THE FULL 2
3
FACTORIAL IESIGlIf
Half Normal Probability Plots
Type of . The Highest Results of Significant Results of
Cyanide Calculated the Visual Effects Backward
Solution Effect Inspection (95% CI) Elimination
NaCN C C C C
Cu-CN C C,B,BC C,B,BC C,B,BC
Zn-CN C C C ( non- signif )
Ni-CN C C,BC C C
Fe-CN A A,C,AC A,C,AC A ( non- signif )
Low Mix A A,AB,B A,AB,B A,AB,B
High Mix C C C C
A - UV light
B - Aeration
C - Temperature
85
output of the example calculation. presented in the Appendix G.5. The
calculations were carried out at one degree of freedom which affected
the sensitivity of the F test. therefore. the results were considered
slightly inferior to the methods outlined in Sections (i) and (ii).
The summarized results of the full 2
3
factorial design are
presented in Table 6.3.
The temperature (C) is the largest main effect for most
solutions (Na. Cu. Ni and high mix cyanide). while UV light (A) is the
largest main effect for Fe and low mix. There is some effect of
temperature on zinc cyanide removal but it is not significant.
6.3 The Effect of Aeration on NaCN Degradation
To assess the effect of air. UV light and temperature on the
removal of cyanide in NaCN solution it was assumed that the whole NaCN
curve represented the removal of free cyanide. However. neither the
effects of aeration nor UV light are significant for the NaCN solution
at the 5% level. This lack of significance is somewhat surprising.
Therefore. the reaction rate constant for NaCN was recalculated
assuming that a cut-off point existed. The linear regression was
carried out forward and backward on every set of data. The average
residual sum of squares (RSS) values were compared (see computer
program BACKRAT in Appendix D explained in Section 5.4.2). There was
no sudden, but a small gradual change in these values. hence there were
some difficulties in estimating the cut-off point. In the case of
metallo-cyanide complex solutions, when there was a definite change in
the cyanide removal mechanisms (from volatilization to metal cyanide
decay), the sudden change in the average RSS values occurred.
The reaction rate constant ltv established for the first part
of the curve did not differ significantly from estimates for the whole
curve. When applied as the response variable for the estimation of the
largest effects, the results presented in Table 6.4 were obtained.
These effects are arranged in ascending order (see Table 6.5)
and a half normal probability plot prepared (Figure 6.2). The effects
of temperature, air and their interaction were observed.
In order to check the significance of these effects the 95%
confidence interval was calculated and it was found to be between
-6.86 x 10-
3
and 5.46 x 10-
3
Therefore the effects of temperature
(e), air-temperature (Be) and air (B) were determined to be significant
at the 95% confidence level (Figure 6.2).
This confirmed that the air played an important role in removal
of free cyanide together with temperature within the first 48 to
72 hours.
The way the "second part" of the NaeN degradation curve can be
explained is that with the decrease in the concentration of total cya-
nide the volatilization rate of HeN in the aerated solutions decreases
too and the variability in the data at the end of the experiment was
noted. In the non-aerated samples the variability occurred too,
probably due to a change in thickness of the cyanide concentration
gradient with time.
86
87
TABLE 6.4. 'fIlE USPOIISE VAlUA1ILES All) KPFECTS - Baa
Test Code R.esponse ky(h-
1
)
Effects
I 0.0164
A 0.0193 -0.0018
B 0.0176 0.0169
AB 0.0159 -0.0004
C 0.0263 0.0235
AC 0.0205 -0.0024
Be 0.0597 0.0180
ABC 0.0569 0.0019
88
TABLE 6.5. PROBABILITY (X) AND EFFECfS ARRANGED IN ASCENDING ORDER
(MaCN)
Order Number 1 2 3 4 5 6 7
Effect x 10
3
-2.45 -1.S5 -0.4 1.9 16.9 lS.0 23.5
Identity of effect ac a ab abc b be c
P - 100 (i-1/2)/7 7.14 21.4 35.7 50.0 64.3 7S.6 92.9
92.9
78.6
ii 64.3
-
50.
:::::i 35.7

21."
CL
71."
NaCN
-10 -5
TEMP.

AIR TEMP.

AIR

o 5 10 15 20 25
EFFECTSX10
3
Figure 6.2 Balf 1Ibraal. Probability Plot for Irv for RaCK
89
7 K)JEL JEVELOPHEIIIT AIm CALIBRATION
According to the results obtained from the experimental design,
it was concluded that volatilization and metal cyanide decay are the
two predominant mechanisms for cyanide removal from the synthetic
solutions.
The results of the dot diagram analyses on the factors affect-
ing volatilization of free cyanide suggest temperature as the main
factor, followed by air and UV (see Table 5.2).
The experimental design results on the factors affecting the
degradation of metal cyanide shows that the temperature is the largest
main effect for all cyanide solutions except iron cyanide and the low
mix (Table 6.3).
Such a strong influence of temperature on the reaction rates
and the results of the calculations of the energy of activation for
NaCN of 21 to 70 kJ/mol, suggest molecular diffusion and chemical
reaction controlled processes.*
* Stwmn and Morgan (1981). For a very rapid diffusion controlled
reaction, E has a range of 12.6 to 21 kJ/mol and for slow reactions
controlled by chemical steps t E of about 70 kJ/mol (a pseudo-first
order rate constant). Moore (1962) - for Wlimolecular gas phase
decomposition E = 54.4 kJ/mol.
90
91
The reaction rate constants for the two predominant mechanisms
for cyanide removal, and the important factors affecting these
constants, will be taken into consideration for the development of a
model.
7.1 Model Development
It is desirable to develop a mathematical model which is a
fairly general expression of the mechanisms underlying a real process.
The evidence found in the literature and the results obtained in this
study suggest that this model should be based upon the following obser-
vations:
1. Metal cyanide complexes dissociate into metal and cyanide
ions.
2. Cyanide ion, in a hydrolytic reaction with hydrogen from
water, produces HeN and is in equilibrium with it.
3. The dominant form of free cyanide at pH 7.0 is molecular
HeN. Molecular HCN is the species that volatilizes.
4. Temperature influences the rate of the dissociation of all
metallo-cyanides. UV irradiation may affect the dissocia-
tion for a few metals.
5. The only parameter that was measured in the experiments is
total cyanide [TCN]o' which is the summation of all the
different cyanides that exist in aqueous solution (i.e.,
TCN = MCN + CN- + HCN).
92
6. HCN can oxidize chemically to cyanate. But measurements of
cyanate made during the experimental runs showed that there
was no CNO- present, suggesting that chemical oxidation is
not an important mechanism.
In order to develop a model for a single metallo-cyanide
complex solution, each chemical species containing cyanide forms was
considered as a separate component, as given in Figure 7.1.
The first component is the metal cyanide complex denoted as
MCN, the second is the cyanide ion and the third is molecular HCN. If
the second and third are taken together they represent free cyanide.
Each arrow between the components represents a transformation of mass.
The assumption was that one metal complex and free cyanide can explain
the observed characteristics of cyanide loss from solution in the
single metallo-cyanide solutions.
The coefficients kl' k2' k3 are the rate constants (h-
1
); kv
is the volatilization (aeration) mass transfer coefficient (cm/h),
while kuv is the additional effect upon kl due to ultraviolet
irradiation (h-
1
). For a batch reactor, taking into consideration the
law of conservation of mass, the mass balance equations for each
compartment are:
d[MCN]
dt V = - kl [MCN]V - kuv[MCN]V
(7.1)
(7.2)

>
.x:
z
U
I
N (T)
.x: .x:
>
:J
.x:
I
Z
U
.....

.x:
z
U
2:
z
U
LL
z
U
t-
.-4
II
i
g
..
...
i
to
a
II

I
-
...
e
I,

93
d[HCN]
dt V
~ k2 [CNi]V - k3 [HCN]V + k As (C -[HCN])
v s
The total cyanide is the sum of Equations 7.1, 7.2 and 7.3:
d [TCN]
dt V
= -k A [HCN]
v s
94
(7.3)
(7.4)
Here As (cm
2
) is the area of the air-water interface across which
volatilization occurs, C
s
is the saturation concentration of cyanide
(mol/L) corresponding to the atmospheric partial pressure of cyanide
(approximately equal to zero for natural conditions) and V is the
reactor volume (cm
3
).
Rearranging the equations yield:
- for metal cyanide
(7.5)
- for free cyanide (sum of CNr and HCN)
d[FCN]
- - - - ~ . . . k [MCN] + k [MCN]-
dt 1 uv
(7.6)
95
If the UV irradiation effect is insignificant,
kuv becomes zero.
Summing these two equations produces one for total cyanide
d[TCN]
dt
Hence the model is equaions 7.5, 7.6 and 7.7.
(7.7)
Evidently the controlling steps are the rate of volatilization of
the HCN form and the rate of dissociation of the MCN. Since the
kinetics of HCN formation frem CW- and its dissociation to CW- are
fast (Eq. 2.1), it is not rate limiting. This allows one to treat HCN
and CN"" as being constantly in equilibrilDD. relative to other proces-
sese The model thus derived represents a two compartment reaction-in-
series model.
7.2 Parameter Estimation Method
The differential equations (7.5 to 7.7) were solved by nlDD.eri-
cal integration using the Runge-Kutta-Verner fifth and sixth order
method (DVERK) available from the International Mathematical and
Statistical Libraries Inc. for CYBER systems. This method uses given
initial conditions and calculates the change of concentration over time
for total, metal and free cyanides using ten-hour time increments.
96
To obtain the best estimates of the parameters for this model,
an iterative, nonlinear least squares parameter estimation subroutine
was used (GAUSHAUS*). It is based on Marquardt's method which combines
the Gauss method (Taylor series) and the method of steepest descent
(Meeter and Wolfe, 1965).
A starting estimate for the initial concentration of
metal-cyanide [MCN]o was calculated from the initial metal concentra-
tion and stoichiometric relationship of the particular complexes formed
from dissociation of the stock salt namely:
2-
Cu(CN)3 ,
2-
Zn( CN)4 ,
2- 3-
Ni(CN)4 , and Fe(CN)6 The estimate for initial free cyanide concen-
tration [FCN]o was obtained by subtracting [MCN]o from the measured
initial concentration of total cyanide, [TCN]o. An example calcula-
tion is given in Appendix H.1.
In this case there were three parameters that needed to be
estimated:the overall metal decay coefficient (k
1
+ kuv)' the
volatilization coefficient (kv) and [FCN]o. However, the UWHAUS
parameter estimation procedure allows for estimation of the initial
metal cyanide concentration. Treating initial metal and initial free
cyanide concentrations as parameters to be estimated creates a four-
parameter estimation problem which was first evaluated herein.
The results of the four parameter estimation problem are
described in Section 8.1. Three and two parameter estimation problems
* GAUSHAUS now called UWHAUS.
97
are also carried out in this study and the results are presented in
Sections 8.2 and 8.3
t
respectively. The parameters estimated by the
three parameter estimation model specifying the MCN/TCN ratio to be
equal to 0.17 are presented in Section 8.4.
A verification study concentrates on the simulations of low
mixes (Section 9.1.1.) and attempts the simulations of high mixes
(Section 9.1.2). Model calibration was evaluated by carrying out
simulations on actual barren solutions (Section 9.2).
8 EVALUATION OF SINGLE METALLO-CYANIDE SOLUTIONS - RESULTS AND
DISCUSSION
8.1 Four Parameter Estimation Model
A sample graph of the results obtained using four parameter
estimation for a single metal cyanide solution, Zn-CN, is given in
Figure 8.1. It appears that a good fit of the model to the data is
obtained. This implies that, as postulated, volatilization of free
cyanide controls the initial steep slope of the curve and that metal
cyanide decay controls the lower part of the curve (the slower rate of
degradation).
The estimated metal cyanide decay coefficients using four
parameter estimation are presented in Table 8.1 with respect to temper-
ature. There is some variability in these estimates. The calculated
values of R2, the mUltiple correlation coefficient, indicates that 93
to 99% of the total variability in the observed data is explained by
the model (Eq. 7.5 to 7.7).
The estimated volatilization mass transfer coefficients kv
are given in Table 8.2. In reviewing these results, it was noted that
although most were quite close, between some of them there was a
difference of two orders of magnitude. The extent of this range was
difficult to accept. A possible explanation was that the metal
98
5."
4.8
OBSERVED CN
T
- TOTALCN-}
---- METAL CN- ESTIMATED
.------- FREE CN-
3.8
3.1
2.3
W
C
t5

.77
0
0 20 40 60 80 100 120 140 160 180
TIME (h)
Figure 8.1 Four-Para.eter Estiaation MOdel - Typieal Fit of Predicted
Values to Observed Data (Zn-CN 20C, AIR, UV)
200
1..0
1..0
100
TABLE 8.1. ESTIMATED METAL CYANIDE DECAY COEFFICIENTS
kl
[b-
1
] FOR
SIRGLI!' METALI.()-CYANIDE SOLUTIONS (FOUR-PARAMETER ESTIMATION
KODEL)
TEMPERATURE
AIR UV k1 at 4C R2[%] k1 at 20C R
2
[%]
Cu
+ + 4.7 x
10-
9
98 8.9 x
10-
3
99
+ 1.0 x
10-
9
94 5.2 x
10-
3
98
+ 5.8 x
10-
3
98 1.5 x 10-
9
* 98
3.2 x
10-
3
99 1.4 x
10-
1
**
98
Zn
+ + 7.0 x
10-
2
94 1.8 x 10-
2
* 99
+ 7.9 x
10-
3
96 5.7 x
10-
2
99
+ 9.3 x
10-
3
98 7.9 x
10-
2
98
4.4 x
10-
3
98 6.5 x
10-
2
98
Ni
+ + 4.5 x
10-
9
98 1.1 x
10-
3
99
+ 1.1 x
10-
3
94 3.6 x
10-
3
99
+ 3.2 x
10-
3
96 1.1 x
10-
3
93
1.7 x
10-
3
96 4.2 x 10-
5
* 99
Fe
+ + 5.9 x
10-
5
99 4.8 x
10-
3
99.
+ 5.6 x
10-
9
98 4.9 x
10-
3
98
+ 4.9 x
10-
3
99 1.2 x
10-
2
99
8.6 x
10-
4
99 1.6 x
10-
3
99
* Inconsistencies because kl at 20C is less than kl at 4C.
**
Time of observation was too short to allow for adequate estimation
of parameters.
101
TABLE 8.2. ESTIMATED VALUES OF VOLATILIZATION MASS TRANSFER
COEFFICIENTS k" [ca/h]
(FOUR-PARAMETER ESTIMATION HODEL)
TEMPERATURE
AIR UV k
v
at 4C k
v
at 20
0
e
Cu
+ + 0.56 139.12*
+ 0.46 2.02
+ 7.85* 0.55
2.85* 0.53
Zn
+ + 0.74 0.08*
+ 2.14 29.73*
+ 1.66 0.74
0.75 0.70
Ni
+ + 0.55 2.97
+ 0.91 0.77
+
1.78 0.74
1.11 3.33
Fe
+ + 0.29 3.69
+ 0.84 0.56
+ 428.06* 1.43
1.55 0.71
* Inconsistencies.
102
complexes, in a somewhat random fashion acted as surfactants and were
able to change the mass transfer characteristics of the volatilization
coefficient. This was judged not to be probable. Another possible
explanation was that the parameter estimation procedure had difficulty
resolving parameter values for certain experiments, due to large
correlations between parameters.
In all the runs, even when the model fit appeared to be satis-
factory, some parameters were moderately correlated. A very high nega-
tive correlation existed between several parameters (FCN, MCN and k
1
)
for data sets which were termed inconsistent in Table 8.1 (Example in
Appendix H.2; program TST used for four-parameter estimation and the
output shOWing correlation matrix element value of -0.99). This means
that an increase in one parameter is accompanied with a decrease in the
other and that a unique set of estimates is difficult to obtain.
Considering the fact that the temperature had such a strong
effect on the cyanide removal kinetics, the sodium cyanide volatiliza-
tion mass transfer coefficients, used for the ky initial values, were
averaged for 4C and 20C (k4
0
C = 0.52 cm/h and k20
0
C '" 1.33 em/h)
and the same four parameters (FCN, MCN, ky, k
1
) were reestimated.
There was no noticeable improvement, in either the model fit within
correlation matrix elements.
Another way of examining the estimated values is to compare the
estimated ratio MCN/TCN from the four-parameter estimation model to the
initially set ratio of 0.17, i.e., based on how the solutions were
made. In 12 out of 32 cases, the MCN/TCN ratio is greater than 0.17
(Table 8.3). For these experiments the estimated FCN concentration was
much smaller than originally calculated (based upon stoichiometry of
the single metal complex). The estimated MCN concentration was much
greater and metal decay appears to be controlling the cyanide degrada-
tion processes most of the time. Despite some difficulties, the
overall fit of the model to the data for most experiments was good.
8.2 Three-Parameter Estimation Model
Decreasing the number of parameters in a given model usually
decreases the incidence of high correlation but can also decrease the
ability of the model to accurately predict observed phenomena.
Metal decay coefficients (k
1
), and initial concentrations were
reestimated for all four metals by fixing kv at the global average
value* of all the volatilization coefficients, kv of 0.93 (em/h).
This reduced the problem to a three-parameter estimation (FCN
o
'
MCN
o
and k
1
). Contrary to the other metals, the estimates for metal
decay coefficients for Cu were more consistent in the three than in the
four parameter estimation model (See Table 8.4).
* The Arrhenius' equation for the temperature effect on the rate
constant was considered in the individual k values for 4C and 20C
but not in calculating the global average value.
103
104
TABLE 8.3. METAL CYANIDE TO TOTAL CYANIDE IlATIO - FOUR PARAMETER
ESTIMATION MODEL
MCN/TCN Ratio MCN/TCN Ratio
AIR UV 4C 20C 4C 20C
+ + Cu 0.09 0.27 Zn 0.30 0.13
+ 0.17 0.09 0.60* 0.74*
+ 0.58* 0.04 0.47* 0.18
0.69* 0.90* 0.69* 0.40*
+ + Nt 0.12 0.17 Fe 0.08 0.18
+ 0.30 0.21 0.16 0.30
+ 0.34* 0.26 0.97* 0.35*
0.50* 0.13 0.25 0.25
* The MCN/TCN ratio is greater than 0.17.
105
TEMPERATURE
AIR UV 4C 20C
Cu
+ +
0.6 x 10-
2
8.7 x 10-
B

+
0.4 x 10-
2
0.2 x 10+
1

+
0.2 x 10-
2
0.6 x 10-
2
0.2 x 10-
2
1.9 x 10-
2
Inconsistency.
The problem common for the estimation with the overall average
kv was that all the estimates for MCN (except for ten conditions out
of 32) were substantially higher than the theoretical initial
concentrations [MCN]o (Tables 8.5a and b). The reason for this was
that the volatilization coefficients were faster than the
volatilization process in these experiments, forcing the model to treat
the entire curve as one rate limiting process, metal decay, and thus
the MCN concentrations appeared greater.
Typical model fits for these conditions are shown in
Figures 8.2 and 8.3. For total cyanide in Figure 8.2, the model fits
the data well; volatilization of free cyanide controls the initial
portion of the curve while metal cyanide decay controls the plateau
over a longer time period. The total cyanide in Figure 8.3 also shows
a good fit, but the metal cyanide complex is estimated by the model to
control the decay process over 1II.1ch more of the whole curve than is
shown in Figure 8.2.
Additional analyses were performed with a three parameter esti-
mation model on twelve data sets in which the four parameter estimation
model gave a MCN/TCN ratio greater than 0.34 (Table 8.3). In these
analyses a fixed stoichiometry of 0.17 was used and the three
parameters to be estimated were initial free cyanide concentration and
the two decay coefficients. There was apparent improvement. The
estimated values for FCN and the metal decay coefficient were closer to
the initial values and the correlation between all the parameters
106
TMILE 1.!Ie. 11REE MWETER ESTIllATICllIlXEL - C. All) Z. CYANIIE AT ky AVERAGE
20C 20C 4C 4C 20C 20C 4C 4C
ky 0.93 (c_/h) - oyerall AIR NO AIR AIR NO AIR AIR NO AIR AIR NO AIR
average NO UV NO UV NO UV NO UV UV UV UV UV
G [FCN 10 I( 10
2
Cu-CN I (IIIOI/L) 0.632 0.667 0.651 0.629 0.647 0.628 0.620 0.655
V
{:oa:k)
0.13 x 10-
2
E
N kc I( 10
2
u (h -1)
2.07 0.71 0.37 0.33 1.53 0.24 0.36 0.31
FeH I( 10
2
UWHAUS 0.601 0.442 0.473 0.312 0.572 0.597 0.545 0.497
MCN
0.411(10-S. 0.38KIO-
2
0.32)(10-
2
0.39)(10-
2
0.89)(10-4. 0.22)(10-
2
0.25x10-2 0.19)(10-2 UWHAUS
lieu
2.3 1.59)(10-
2
0.3.,0-
2
0.23x10-
2
0.87)(10-
7
0.64)(10-
2
0.6OxI0-
2
0.17)(10-
2
U ~ U S
Zn-cN G [FCN ~ I( 10
2
I IIIOI/L) 0.584 0.518 0.620 0.502 0.624 0.639 0.628 0.601
V
{:oa:k)
0.13 I( 10-
2
E
N k I( 102
Zn (h- 1)
0.62 1.94 0.68 0.45 4.48 1.82
. 1.37 0.95
FCN
0.5&c10-
2
0.45x10-
2
0.511(10-
2
0.19)(10-
2
0.18K10-
2
0.94xI0-
8
0.5&c10-2 UWHAUS NC
Mat
0.,.,0-'+. 0.19)(10-
2
0.241(10-
2
0.45x10-
2
0.59)(10-
2
0.77)(10-
2
0.13x10-2 UWHAUS NC
kZn
2.45x10-
2
0.6OxI0-
2
0.44xI0-
2
4.3Ox 1 0-
2
8.97)(10-
2
0.5OxI0-
2
lMiAUS 2.79 NC
-- 0
NC - tl_tlon routine was not converging.
.....
INCNJ lower than observed.
TaE I.". n. EST_nClt MJDB. - III A. Fe CTMIDE AT Il" IWERIISE
20C 2o-C 4C 4C 20C 20C 4C 4C
"v 0.93 (eM/h) - overal I
AIR NO AIR AIR NO AIR AIR NO AIR AIR NO AIR
average NO UY NO UV NO UV NO UV UV UV UV UV
8 [FCN] x 102
NI-eN I ,...gill)
0.1" 0.635 0.661 0.522 0.636 0.632 0.639 0.541
Y
(::Jt)
0.13 x 10-
2
E
N Ifttl x 1 ~
(h- )
0.21 0.08 0.30 0.21 0.08 0.15 0.42 0.41
FeN x 102
UNHAUS 0.512 0.541 0.541 0.362 0.534 0.590 0.534 0.502
MCN JC 102
tlWHAUS 0.219 0.063* 0.223 0.301 0.011" 0.192 0.283 0.111*
kNI
0.4!K10-2 0.42K10-8 0.59lC10-
3
0.12K10
2
0.34x10-8 0.18K10
2
0.52K10
2
0.32K10-8 lJfHAUS
F.-eN G [FCN] x 102
I
(",gill)
0.644 0.624 0.590 0.409 0.514 0.514 0.613 0.514
,
[ : : J ~
0.13 x 10-
2
E
N ~ . x 1 ~
(h- )
0.16 0.26 0.12 0.01 0.68 2.22 0.51 0.40
FCN x lil
UWHAUS 0.410 0.532 0.615 0.319 0.461 0.623 0.264 0.118
MCN x 102
UWHAUS 0.381 0.242 0.119* 0.100- 0.021* 0.064* 0.413 0.505
k
F

0.6,.10-2 0.2,.10-2 0.3,.10-8 0.1&.10-
9
0.51x10
7
0.54xI0-
2
0.64x'O-2 0.4fM'O-2 tJfHAUS
* [MCN] lower than observed.
0
00
::;
~
~ 8 D
~ .
- C')
~
)(
~
~
~
o
~
~
o
\
\
\
\
\
\
\
\
\
\
"
"
,
OBSERVED CN
r
TOTAL CN-}
. -- METAL C ~ - ESTIMATED
.------- FREE CN
, ..
- .
" .
- - - ~ .. ---- .
- ~ ~ ~ --------
~ - - - - - - - - - - - - - - - - - - -
- - - - - - - - . - . - - - - - - - -----_.
01 , I I I I I ~
o 50 100 150 200 250 3)()
TIME (h)
ftaun 8.2 1bne-'ar_ter .. t1 .. tiOll Model - fCIf > HCiI (c.. 20C, m A.D., Of)
o
\,0
110
-
t
i
..
S

I
I
i
I


I
-
I
I ..
I U
0
I


I
I ...
I
a

I
/
I
::E I
-
-
I
t;;
I
I

.- I
5,
w

I
--
/
I
"
,

I

-/
I

I

I
000
I
I

I
..J..J

I
-

I
I
.r:.
I
-
"""
I II
I
I W
i
I


I
I
I
I
,

F=

I
I
g
I
I
I
I
-
-
I
I
'to'4
I
I ,
..,
I

!
,
I
,
,

, ..,
I
,
,

,
a
/
,
,
..

,
II
/
, ..,
,
I
I
,
,
, ..

:.
/
,
I
t
,"
I
I
2:

/
"
tS
/,/
,.,
, .
,
00
0
2:
l!

Q
., 0
.. t') ,
(1;1ow) x
'to'4
1114
decreased, e.g., the correlation between FCN and kl decreased from 0.99
to 0.43 for .Fe at 4C, NO AIR, UV.
One other problem was that for some experiments the time period
of observations was too short. Accordingly, because the data base was
not sufficiently extensive, the lower portion of the curve did not
adequately define the values of the coefficients. For such curves,
there was a high correlation between estimated parameters in both the
four-parameter problem format and in the three-parameter format.
In a few experiments which ran sufficiently long t the total
cyanide concentration plateaued at a non-zero value whereas the model
predicted that the concentration should decrease to zero. It was hypo-
thesized that an additional metallo-cyanide complex existed in these
solutions whose decay characteristics were much slower than those of
the first complex.
Consequently, it was decided to perform the three parameter
estimation (three parameters being MCN, kv and k
1
) with the volatili-
zation mass transfer coefficient initial values equal to the value
recorded for NaCN under each individual condition. This was possible
since it was proven earlier that
metallo-cyanide complex solutions
interval of kv for NaCN (k
NaCN
).
volatilization coefficients for
were within the 95% confidence
.The computer program was adjusted
so that the first predicted TCN value was equal to the initial observed
TCN concentration and initial MCN was estimated. This resulted in FCN
111
112
being equal to the difference between TCN and MCN. The results of this
estimation. process are presented in Appendix H.3, Tables 1 to 4.
There were three incidents (Cu 20C, NO AIR, NO UV; Zn 4C,
AIR, UV; Zn 20C, NO AIR, UV) where the estimated MCN value was much
greater than the theoretical metal cyanide concentration or was even
greater than the observed TeN concentration. One experiment (Cu 20C,
NO AIR, NO UV) was extremely short in time, 169 hours. Only four data
points were available to use for the calculation of the metal decay
coefficient and this was not a large enough data base to define
adequately the values of the parameters. The correlation between MCN
and kl for Cu (or kcu) was unity.
For Zn at 4C, AIR, UV the values of kl and ley were very
close (0.019 and 0.014 h-
1
, respectively) forcing the model to treat
the entire curve as one process, and consequently giving too high an
estimate for the MCN concentration and at the same time a negative
value for the FCN concentration (since TeN = MCN+FCN). As expected
there was a high negative correlation of -0.94 between ley and for
Zn (or k
Zn
).
In the third case, Zn at 20C, NO AIR, UV the same problem
occurred, i.e., ley and kl values determined for Zn were very close
(0.021 and 0.018 h-
1
respectively). In addition it was calculated that
93% of cyanide was removed during the volatilization process (from
Time 0 to the cut-off point) and only 7% was due to metal decay (from
the cut-off point to the end of the experiment). The observation time
113
for metal decay was only 150 h long, and the change in cyanide concen-
tration was not significant enough to give a good estimate for the
metal decay rate. TIle high correlation of 1.0 found between MCN and
kZn was not surprising.
TIle inability of the model to estimate adequate MCN concentra-
tion for these three cases was observed also when the values were
plotted with their individual 95% confidence intervals, based on the
linear hypothesis calculated by UWHAUS, Figure 8.3a and Appendix H.3,
(Figures 1 to 3). The three conditions analyzed previously were
characterized with very large confidence limits.
In spite of the difference between the original MCN/TCN = 0.17
and the MCN/TCN estimated values, ranging from 0.07 to 0.74, the
calculated multiple correlation coefficient, R2, indicated that 96 to
99% of the total variability in the data was explained by the model
(Appendix H.3, Tables 1 to 4).
Generally, the estimated values for the parameters in the
three-parameter estimation model improved and resulted in a better
model fit compared to the estimates and model fit in the four-parameter
estimation model. TIle correlation between parameters generally stayed
high in the case of copper and zinc and decreased for iron and nickel.
TIle reason for this was that the kv and kl values determined for
copper and zinc cyanide solutions were very close and that created a
difficulty in estimating all the parameters.
I
~
~ 4
-
2
3
-
~
12
oj
-

'\.1
-x
I
a
l
...
~
-
i-I
114
I
I
1 2 3 4 5 II 7
I TEST tu4BER
Figure 8 ~ a Three-Para.eter IBtt.ation MOdel - Eatt.ated HCR values
with 95% Confidence Intervals - Ca.
115
More data which covered the complete range of variation would
be the only way to significantly decouple the correlation among the
parameters.
8.3 TWo Parameter Estimation Model
By specifying the value of the volatilization mass transfer
coefficient as equal to that estimated for NaCN solutions, and total
cyanide concentration as equal to initial observed value, the only two
parameters which need to be estimated are the metal decay coefficient
and metal cyanide concentration. Free cyanide is thus equal to the
difference between total cyanide and metal cyanide. The computer
program used in two parameter estimation modelling is presented in the
Appendix H.4 (see TSTOL2).
In this estimation process, the correlation between these two
parameters (k
1
and MCN) decreased but the model fit was poorer. When
the residual stun of squares (RSS) was compared between three and two
parameter estimation problems the RSS in the two parameter estimation
problem generally increased (Table 8.6).
The conclusion at the end of two parameter estimation modelling
was that by decreasing the number of estimation parameters to two the
ability of a model to predict them accurately also decreased substan-
tially.
'fAILa 8.6. USUL'lS ext USIDU& SUMS or SQUADS D !WO AIm 'I'8aD PAUMITD BSTIIIA.next JD)ILS
(PIlI)
Metallo-
Cyanlde 20C 20C 4C 4C 20C 20C 4C 4C
Co.plex AIR. NO AIR. AIR. NO AIR. AIR. NO AIR. AIR. NO AIR.
Solution NO UV NO UV NO UV NO UV UV UV UV UV
Cu-CN Two PEM
R.SS x 10
6
2.8 1.3 1.1 2.6 16.3 2.3 2.2 10.6
Three !fM
R.SS x 10 0.98 0.97 1.1 0.43 0.65 1.6 1.8 0.72
Zn-CN Two PEM
R.SS x 10
6
13.1 0.73 0.72 0.42 1.1 0.81 1.7 4.8
Three PEM
R.SS x 10
6
0.05 0.72 0.72 0.42 0.16 0.80 1.3 1.2
Rl-cN Two PEM
R.SS x 10
6
0.49 39.5 2.4 1.8 8.6 1.1 . 1.6 10.7
Three PEM
R.SS x 10
6
0.49 0.37 0.32 1.2 0.08 1.0 1.2 1.6
re-CR Two PEM
R.SS x 10
6
2.2 0.36 1.4 6.7 8.6 0.62 0.33 0.64
Three PI"
R.SS x 10 1.3 0.28 0.86 0.21 0.02 0.42 0.26 0.64
13'
117
8.4 Parameter Estimation with Fixed MeN/teN Ratio
As indicated earlier all the metallo-cyanide synthetic
solutions examined were made based on the calculated value of
MeN/TCN .. 0.17, i.e., 17% of the available cyanide was in a complex
with the metal( s) present. After completing four, three and two
parameter estimation using the earlier defined model equations (7.5,
7.6 and 7.7), the three parameter estimation model seemed to give the
best estimated values for all the parameters (MeN, kv and k
1
) but
there were some discrepancies from the original MeN/teN ratio of 0.17.
In order to see whether these estimated values were statisti-
cally different from the estimates obtained by presetting the MeN/teN
ratio to 0.17, an extensive set of calculations were carried out. A
series of different MeN/TeN ratios were stipulated and used in the
three parameter estimation problem for all four metals. MeN/TCN values
used were 0.0425, 0.085, 0.17, 0.34 and 0.68. These values represent
lower, equal and higher values than the theoretical MeN/ TeN value of
0.17.
The complete results from this estimation process are presented
in Appendix H.5 Tables 1 to 4, together with the example program
" TSRATIO" used for the calculations. The residual sums of squares
(RSS) were recorded for all estimation runs (the four metals under all
eight experimental conditions). In 14 of 32 cases the MeN/TCN = 0.17
had the lowest RSS. Only under two conditions at MCN/TeN<0.17, RSS
were the lowest, and all the others with the minimum. RSS (16 of 32)
happened to be with MCN/TCN >0.17 (Table 8.7). The RSS values were
plotted v r ~ u s the five MCN/TCN ratios (see Appendix H.s
t
Figures 1 to
4) to visualize better the minimum RSS conditions.
In order to determine the statistical significance in these
results an F test was performed between two models
t
one using the RSS
at MCN/TCN - 0.17 t and the other for the MCN/TCN ratio having the
minimum RSS of all five ratios examined. The total sum of squares
(TSS) values for both models are identical
t
as is the number of
parameters and degrees of freedom for the RSS being compared.
The results of the F test are given in Table 8.8. These F
tests performed on the individual conditions prove that only five
conditions (5 out of 32) are significantly different from the MCN/TCN
ratio of 0.17. These conditions are mostly experimental data for which
we have had estimation problems before: either too short experimental
runs
t
or experiments with NO AIR where stratification probably
occurred.
The general conclusion from these statistical tests is that the
model fit obtained in 27 out of 32 conditions with the defined three
parameter estimation model (using kv - kNaCN) was not significantly
different (5% level of significance) from the fit made with the three
parameter es timation model under preset MCN/TCN ratio of 0.17. The
values of the three parameters estimated for the different experiments
at fixed MCN/TCN - 0.17 are given in Table 8.9. Whenever there was a
case where volatilization occurred extremely rapidly (e.g.
t
as observed
118
119
TABLE 8.7. RUKBER OF MINIMUM RSS VALUES AT THE PAilTICULAR MCN/TCN
RATIO
Metallo-
Cyanide
Complex
Solution
Cu - CN
Zn - CN
Ni - CN
Fe - CN
Number of
cases with the
lowest RSS
0.0425 0.085
2
2
MCN/TCN
0.17
2
3
4
5
14
0.34
2
1
4
3
10
0.68
2
4
6
UJILIl 1.9. 1!111 IIS1JLTS r. PAUHI'l'IIt IftIlfAt'IOll ., .. L - IICII/tClI - 0.17
Metallo-
Cyanide 20C 20C 4C 4C 20C 20C 4C 4C
Coapl_ Ea tiu ted AIR NO AIR AIR. NO AIR AIR. NO AIR AIR. NO AIR
Solution Par_eten til uv tIlUV til UV tIlUV UV UV UV UV
Cu-CH FeR (aol/L) 0.0064 0.0069 0.0065 0.0062 0.0065 0.0067 0.0065 0.0059
It., (h-
1
) 0.0709 0.0148 0.0151 0.0147 0.1014 0.0196 0.0184 0.0168
kcu (b-
1
) 0.0119 0.0741 0.49z10-
9
.. 0.47xlO-9 .. 0.0066 0.0042 0.0029 0.2&10-
9

ZD-CI FCB (aol/L) 0.0056 0.0053 0.0062 0.0053 0.0063 0.0067 0.0067 0.0058
It., (h-
1
) 0.1062 0.0192 0.0183 0.0069 0.0840 0.0222 0.0212 0.0242
ItZ
n
(b-
1
) 0.0280 0.0671 0.0032 0.0014 0.0219 0.0629 0.0622 0.0045
li-CI lCH (aol/L) 0.0070 0.0063 0.0066 0.0053 0.0064 0.0064 0.0067 0.0049
ky (h-
1
) 0.0177 0.1045 0.0190 0.0153 0.0876 0.0196 0.0171 0.0192
ItNi .(h-
1
) 0.87xlO-
3
0.0017 0.3Oxl0-
9
0.63xlO-
1O
0.0012 0.S9z10""'. 0.0017 0.lb10-
9

Fe-CN PCB (aol/L) 0.0066 0.0063 0.0061 0.0040 0.0058 0.0059 0.0061 0.0056
It., (b-
1
) 0.0140 0.0187 0.0249 0.0183 0.1142* 0.0297 0.0098 0.0069
kPe (h-
1
) 0.0013 0.24x10-
9
* 0.21x10-
3
0.89z10-9* 0.0052 0.0078 0.0027 0.0014
*
High e6timated tv value aa a reault of a faat volatilization proceaa.
** Lo1f eaU.ated Itl value. the reault of either long volatilization proceaa and abort t1ae of obaarvation for
aetal decay. during which not auch change in cyanide concentration occurred. or abort volatilization proceaa
and plateaued value of cyanide decay.
N
in Figures 5.1 and 5.2) the'estimated value for the volatilization mass
transfer coefficient was an order of magnitude higher than the value
experimentally determined. On the other hand some metal decay
coefficients are estimated to have extremely low values (in the range
of 0.6 x 10-
4
to 0.6 x 10-
1
); these low values generally occur when
the metal decay rate plateaued for a substantial period of time (e.g.
Ni) or when the volatilization process dominates most of the curve or
when the experimental data is too short to give an adequate definition
of the metal decay rate.
The estimated values of the metal decay rate coefficient were
plotted on dot diagrams with respect to the three factors: UV light,
air and temperature (Appendix H.6, Figures 1 to 3). Temperature has an
effect on the decay rate of all four metals. There is no observed
effect of either UV or air on nickel decay. There is some relationship
between UV light and iron cyanide decay, and some relationship between
air and copper cyanide decay.
When considering these results in connection with the actual
cyanide-bearing solution in the tailings pond, the implication is that
the worst condition for the degradation of cyanide is the cold
particularly ice covered ponds when the air and UV in addition to
temperature have a minimum effect on degradation.
The results of this modelling excerise have, therefore, proven
to be very realistic and in agreement with the similar information that
can be found stated in the literature (Ingles, 1981a).
122
Contrary to this, it can be concluded that the best conditions
for c y n i d ~ degradation would be warm weather, high temperatures,
sunshine and wind (or some other mechanical source of aeration, Freeman
(1983.
123
9 VEltIFlCAnON STUDIES
9.1 Simulation on the Synthetic Mixes
The parameters estimated by the three parameter estimation
model specifying MCN/TCN ... 0.17 were used to simulate the results from
the low and high synthetic mix solutions.
As observed from the experimental and estimated values only
temperature had a consistent effect on metal cyanide decay coefficients
of all four metallo-cyanide solutions. Therefore the metal cyanide
decay coefficients were averaged at 4C and at 20C and used in the
simulations for the mixes (see Table 9.1). The parameter values that
were judged as inconsistent were rejected before averaging the decay
rates. The three criteria used to specify inconsistency were the
following:
(i) The three values of kl for Cu (kCu) at 4C estimated as
extremely low, would signify almost a zero rate of reaction, while in
reality the reaction was still going on (cyanide removal over that
per.iod averaged about 24%). Therefore, the assumption was that the
more realistic value of 0.0029 (h-
1
) was more representative of kcu
at 4C.
(ii) The value of 7.41 x 10-
2
(h-
1
) for kcu at 20C was
estimated as too
high (kcu > kv) because the estimation routine
124
125
TABLE 9.1. ESTIMATED METAL DECAY COEFFICIENT VALUES kl (h-
1
) AT
KCN/TCN - 0.17 (AVERAGED WITH RESPECT TO TEMPERATURE)
Metallo-
Cyanide
Complex
Solution
Cu-CN
Zn-CN
Ni-CN
Fe-CN
AIR
+
+
uv
+
0.49 x 10-
9
*
0.47 x 10-
9
*
0.00295
+ 0.28 x 10-
9
*
Accepted Average kCu - 0.00295
+
+ +
+
0.0032
0.0014
0.0622
0.0045
Accepted Average kZn - 0.01783
+
+ +
0.30 x 10-
9
0.63 x 10-
10
0.0017
+ 0.12 x 10-
9
Accepted Average k
Ni
- 0.00044
+
+ +
0.0002
0.89 x 10-
9
*
0.0027
+ 0.0014
Accepted Average ~ - 0.00143
* Rejected as inconsistent.
0.0118
0.0741*
0.0066
0.0043
0.00753
0.0280
0.0671
0.0219
0.0629
0.04496
I 0.0009
0.0017
0.0012
0.59 x 10-'+
0.00095
0.0013
0.24 x 10-
9
*
0.0052
0.0077
0.00473
126
did not have enough information due to the limited data base (total
experiment,al time only 169 h). Therefore, the other three k
1
' s for Cu
at 20C (Table 9.1) were used to obtain the average kcu at 20C.
(iii) The kl for Fe (k
Fe
) for two conditions had very low
estimated values (10-
9
h-
1
) and if they were included in the cal cuI a-
tion for the average value they would change kFe drastically for the
other six conditions, therefore they were also rejected as non-
representative.
The energy of activation was calculated using Eq. 5.6 for the
average kl values at 4C and 20C. The values obtained are listed in
Table 9.2. The activation energy permits one to ascertain that values
used for averaging are plausible am consistent with the idea that the
activation energy for aqueous solution reactions are typically within a
range of 10 to 100 kJ/mol*.
These energy of activation values are at the low end of such a
scale. They seem to be on a border where diffusion control ends and
chemical reaction control starts.
* Stumm and Morgan (1981) For very rapid diffusion controlled
reactions they found E to have a range of 12.6 to 21 kJ/mol and a
slow reaction controlled by chemical steps to have E = 70 kJ/mol (for
a psuedo-first-order rate constant). On the same subject, Moore
(1962) referred to a unimolecular gas phase decomposition with
E = 54.4 kJ/mol.
127
tilLE 9.2. DEltGY OF ACTIVATION POll kl BETWEEN 'C aDd 20C
Estimated kl
(h-
1
) at
Metallo-Cyan1de E
Complex Solut1on 4C 20C ltJ/mol
Cu - CN 0.00295 0.00753 39.4
Zn - CN 0.01783 0.04496 39.0
N1 - CN 0.00044 0.00095 33.0
Fe - CN
0.00143 0.00473 50.4
9.1.1 Low mix
The values of all the initial parameters necessary to carry out
the simulation on the low mix cyanide solution are presented in
Table 9.3. An example calculation is given in Appendix 1.1. An
example calculation for the rate constant due to ultraviolet irradia-
tion is presented in Appendix 1.2.
The initial metal cyanide concentration [MCN]o was calculated
taking theoretical stoichiometry of cyanide complexes into considera-
tion. The initial free cyanide concentration for all eight experimen-
tal conditions was calculated by subtracting [MCNlo from the measured
[TCN]o value.
A typical example for the low mix model fit simulation is
presented in Figure 9.1 for the conditions of 20C, NO AIR, UV. The
results of the model simulations on the low mix solutions under the
other seven experimental condi tions (Figures 1.3.1 to 1.3.7) and the
sample program used for simulations are presented in Appendix 1.3.
Predicted cyanide values are compared with observed values near
the end of each low mix experiment in Table 9.4. Generally, the
predicted TCN values are close to the observed values with one excep-
tion which is the 4C, NO AIR, UV condition. The lack of agreement is
obvious when the graph is inspected (Figure 1.3.2, Appendix 1.3). In
order to illustrate the difficulty that the model encountered, the
experimental results of the degradation process of single metallo-
cyanide solutions and the low mix (concentration versus time) under the
128
TAILI 9.3. IlIITIAL PA1WIKTERS USED FOR 1'IIB MODEL SDIULATIOR OF TIIK LOW MIX
20
0
e 20
0
e 4e 4e 20
0
e
Paraaeters AIR NO AIR AIR NO AIR AIR
NO UV NO UV NO UV NO UV UV
PCN TR (1) (mol/L) 0.00638 0.00631 0.00615 0.00692 0.00592
MCN TH (2) (1I01/L) 0.0013
tv TH (3) (h-
1
) 0.0394 0.0238 0.0074 0.0050 0.0389
[eu-CN] TH (4) (1I01/L) 0.00033
[Zn-CN] TR (5) 0.00062
[Nt-CN] TH (6) (mol/L) 0.00014
[Pe-cN] TR (7) (mol/L) 0.00022
keu TH (8) (h-
1
) 0.00753 0.00753 0.00295 0.00295 0.00753
kZ
n
TH (9) (h-
1
) 0.04496 0.04496 0.01783 0.01783 0.04496
kNt TH (10) (h-
1
) 0.00095 0.00095 0.00044 0.00044 0.00095
kPe TH (11) (h-
1
) 0.00473 0.00473 0.00143 0.00143 0.00473
kuv TH (12) (h-
1
) 0.00025
20
0
e 4e
NO AIR AIR
UV UV
0.00635 0.00631
0.0253 0.0234
0.00753 0.00295
0.04496 0.01783
0.00095 0.00044
0.00473 0.00143
0.00025 0.00025
4e
NO AIR
UV
0.00581
0.0139
0.00295
0.01783
0.00044
0.00143
0.00025
N
\.D
N
0
Y""
><
Z
0
-
<
a:
....
z O.
w
()
z
0
()
W
C
Z

()
-OBSERVED CN
r
-TOTALCN-l
---METAL CN- ESTIMATED
..... - .. FREE CN-
'lpn 9.1 Low IUs Sl .. 1atl00 - 20C. IJO AI., UV'
..
\N
o
DILl ;.4. 'l'O'I'AL CYUlDI YALOIS (08SIIVIO AIm ISTDIATID) fOR LOll MIX CYARlDI SCLUTIOII
Observed TCN 20C 20C 4C 4C 20C 20C 4C 4C
verSU8 AIR NO AIR AIR NO AIR AIR NO AIR AIR NO AIR
Predicted TCN NO UV NO UV NO UV NO UV UV UV UV UV
Expert.entall, Exper. Ti.e (h) 312 312 317 312 306 306 288 264
Observed
Values TCN exp. (lIOl/L) 0.0004 0.0004 O.OOlS 0.002l 0.0001 0.0002 0.0004 O.OOOlS*
Model Si.ulated Tiae (h) 310 310 320 310 310 310 290 260
Si.ulaUon
Values TCN predlc. (lIOl/L) 0.0002 0.0002 0.0012 0.0021 0.0002 0.0002 0.0004 0.00068
* Model esti.atiob value considered .ore accurate than experi.entally obtained value because of the discre-
pancies observed in this low .tx experi.ental run.
w
same conditions were plotted on the same graph (Figure 9.2). The
difference in the slope for the initial volatilization process is
obvious. This is an extreme case, i.e., the volatilization process in
the low mix happens so fast that it gives the definite indication that
some of the assumed complexes do not form, instead that the free
cyanide is released as soon as the process starts. This would explain
a sudden decrease in total cyanide concentration for which the model
could not account. Therefore, the low mix model simulation under 4C,
NO AIR, UV condition was assumed correct and the experimental observa-
tion of the low mix under the same condition incorrect.
Under the three conditions: 4C, AIR, UV; 4C, NO AIR, NO UV
and 20C, AIR, NO UV there seems to be a delay in the initial degrada-
tion of cyanide in the low mix. However, when compared to the range of
variations observed in single solutions under these conditions, the
"delay" was within the 95% confidence interval of the single solu:""
tions. Although this slowed down the process in the beginning, after
230 hours the concentration estimated by the model was similar to the
observed values. Generally, it was observed that the conditions with
NO UV would require a longer reaction time than that allowed, in order
to have the model fit the data better. This was not surprising, since
UV light was found earlier to have an effect on the rate of degradation
of the low mix.
In order to check if the model was within the 95% confidence
level of the observed data, a "modified F" test was used. According to
132
1
~ t
~
a:
I
cu
Low mix
Zn
o m ~ ~ ~ m ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
TIME (h)
Pigure 9.2 10tal Cyanide Degradation - Single Metallo-Cyanide o ~ l e x
Solutions and Low Mil[ Cyanide Solution (4C. NO AIR. OV)
w
w
Bard (1974), it is possible to calculate the likelihood ratio test for
two models with nonlinear coefficients by using the following simpli-
fied equation:
Modified F test
n
Mode1
2
RSS 2'
(Modell RSS)
In this case, Modell is based on the mean; Mode1
2
is the
low mix simulation, while n is the number of experiments. For
L(1)/L(2) :::::: B hypothesis H2 is accepted (Mode1
2
is correct), and
for L(1)/L(2) A hypothesis Hl is accepted (Modell is correct),
and if A ::> L(1) / L(2) ::> B the experimentation should continue.
The A and B constants are determined depending upon the confidence we
desire to place on the results. If a = B - 0.05 the values of A and B
1-B B
are A .. -- - 19 and B ,.. -1- - 0.053.
a -a
From the results of these calculations presented in Table 9.4a,
the experiments under 20C, NO AIR, NO UV and 4C, AIR, UV conditions
should be continued (A >4.14 and A ::>1.69) and for the condition of
20C, AIR, NO UV, Modell is correct. Therefore, for five of the
eight test conditions, the model fits the data at the 95% confidence
level (identified with two asterisks in Table 9.4a).
When the calculated values are compared to the standard F test,
only the model fits under two conditions (20C, AIR, NO UV and 20C, NO
AIR, NO UV) are not acceptable.
134
TABLE 9.4a. VERIFICATION OF THE LOW MIX MODEL SIMlJLA.TIOIf - STANDARD F TEST (5%) VERSUS I>DIFIED F
TEST*
df
Modell RSS
Model
2
RSS
F Test
Modified
F Test
20C
AIR
NO UV
13
20C
NO AIR
NO UV
14
4C
AIR
NO UV
18
1.034 0.867 0.053
2.571 1.035 0.006
4C
NO AIR
NO UV
17
0.145
0.0002
20C
AIR
UV
12
2.542
0.735
20C
NO AIR
UV
12
0.707
0.273
4
4C
AIR
UV
4C
NO AIR
UV
8
0.356 5.158
0.432 0.066
2.57
926.5
2.50
4.14
2.23 2.27 2.27 2.69 6.39
1.6x10-
10
** 8.1x10-
29
** 1.7x10-
4
** 1.3x10-
3
** 1.69
3.44
3.4x10-
10
**
* "Modified F Test" is a modified version of an F test for the model with nonlinear coefficients -
See Ref. Y Bard, 1974. Chapter x.
** The model fits the data at the 95% confidence level.
\AI
\.l1
9.1.2 High mix
The metal decay coefficients applied in the high mix model
simulations were the same used for the low mix. These were the average
values of the metal decay coefficients determined by the three-
parameter estimation model carried out on the single metallo-cyanide
complex solutions (preset MeN/TeN ratio of 0.17).
The initial values for cyanide concentrations in the high mix
were obtained as explained in Appendix J.1 (i) and (ii). Because the
high mix cyanide solution contained an excess of metals, some adjust-
ments in cyanide concentrations had to be made before the model
simulation process. The adjustments made for the first model simula-
tion consisted in reducing all MeN concentrations by 15 to 23.1% in
order to reach FeN - 0.0 mol/L. The initial values for all the
parameters required for the simulation are presented in Table 9.5.
A typical example of the first model simulation and its fit are
presented in Figure 9.3a. The model fit of Figure 9.3a is not satis-
factory, throughout the whole domain. But final predictions for TeN
were not too far from those observed (see Table 9.6). In this first
simulation, either the original assumptions made were not warranted or
the estimates of the metal decay coefficients made for the single
metallo-cyanide solutions were not 'suitable for the high mix condi-
tions.
A second set of model simulations on the high mix data were
carried out after some adjustments were made on metallo-cyanide
136
'l'ABLI 9.5. 181 DD'IAL PAIUIIB'l'BI. VALUES POR RIGH MU CDWlDB SOLO'l'IOW (PIRSt' SlMULA'l'IOIl)
(1) 20
0
e 20
0
e 4e 4e 20
0
e 20
0
e 4e
High Mix AIR NO AIR AIR NO AIR AIR NO AIR AIR
Simulation NO UV NO UV NO UV NO UV UV UV UV
reN-rR(l) (mol/L) 0.0000
MCN - TCN - TH(2)
0.0076 0.0075 0.0076 0.0076 0.0074 0.0073 0.0069
k - TH(3) (h-
1
) 0.0394 0.0238 0.0074
v
0.0050 0.0389 0.0253 0.0234
[Cu-CN] - TH(4) (1I01/L) 0.Q019 0.0019 0.0019 0.0019 0.0019 0.0019 0.0018
[Zn-CN] - TH(5) 0.0036 0.0036 0.0036 0.0036 0.0035 0.0035 0.0033
[Hi-CN] - TH(6) 0.0008 0.0008 0.0008 0.0008 0.0007 0.0008 0.0007
[re-CN] - TH(7) . 0.0013 0.0012 0.0012 0.0012 0.0012 0.0012 0.0015
k
ea
- TH(8) (h-
1
) 0.0075 0.0075 0.0029 0.0029 0.0075 0.0075 0.0029
k
Zn
- TH(9) 0.0449 0.0449 0.0178 0.0178 0.0449 0.0449 0.0178
k
Hi
- TH(10) 0.0009 0.0009 0.0004 0.0004 0.0009 0.0009 0.0004
"re- TH(11)
.0.0048 0.0048 0.0014 0.0014 0.0048 0.0048 0.0014
kuv- TH(12) 0.0002 0.0002 0.0002
4e
NO AIR
UV
0.0074
0.0139
0.0019
0.0035
0.0008
0.0012
0.0029
0.0178
0.0004
0.0014
0.0002
w
-...,J
........
..J O.
~
-
N
$2
)(
Z
o
-
~
a::
t-
Z
W O.
U
Z
0
U
W
C
Z
~
o

. '"
\
\ .

' ..
'. .
'.
'.
'.
'. .
'.
'.
' .. ,
.. , ........
OBSERVED CN
r
TOTAL CN- !
---_. METAL CN- ESTIMATED
......... FREE CN- .
..............
.........
,.,.,.-. .
_._._._._._._._._._.--.._._.-


.................................


..... . ......... .
. . ............. .
.. . ....................... _ ........... _ .... __ ...... _ .. .
. -


Q o ~ r ~ ~ ~ ~ ~ r ~ . . ~ r . r . . . . ~ ~ ~ ~ ~
o 35 105 140 175 245 280 70 210
315
TIME (h)
Pigure 9.3. B:l.gh Mis - firat Si .. l.tion (4-C, NO AlR., 80 UV)
w
00
TAILI 9.6. 'IO'rAL C'fAlODI VALOIS, OBSIDID All) 1St'DIA'l'ID fOIl RIGH MIX CDRIDI SOLOTIOlf ('lVO A1IBIIPTS)
RiSh Mix 20C 20C 4C 4C 20C 20C 4C 4C
Obaerved TeN AIR NO AIR AIR NO AIR AIR NO AIR AIR NO AIR
.. Est. TCH NO UV NO UV NO UV NO UV UV UV UV UV
!xper.
Experi_ntall, Tiae (h) 266 266 312 306 306 306 312 317
observed .alues
TCNexp. 0.0015 0.0022 0.0032 0.0031 0.0013 0.0019 0.0026 0.0032
(aol/L)
Model 8t_latioo Siaulated 270 270 310 310 310 310 310 320
tille (h)
TCN predicted 0.0014 0.0015 0.0035 0.0044 0.0012 0.0011 0.0021 0.0026
(1) (.ol/L)
Model 8iaulation Si1lUlated 270 270 310 310 310 310 310 320
tille (h)
TCN predicted 0.0015 0.0016 0.0034 0.0041 0.0010 0.0013 0.0024 0.0025
(11) (1Iol/L)
Vol
\.0
140
concentrations (Appendix J.1(U. The adjustments involved reduction
of eu ani Zn metal concentrations in a complex with cyanide. The
values of all the initial parameters are presented in Table 9.7. The
assumption in this case was that eu ani Zn cyanide complexes were
easily dissociated and cyanide became available for volatilization or
further complexation, while metals could precipitate, adsorb or complex
again.
According to Figure 9.3b cyanide was complexed again as part of
a slow-decaying metallo-cyanide complex, because the model fit subs tan-
tially improved in the metal decaying area. Model predictions agree
much more closely with observations near the end of the experiments in
this simulation (see Table 9.6 ani Figure 9.3b).
The obvious observation concerning the high mix model fit in
both Figures (9.3a and b) was that the model predicted a slower rate of
volatilization and more rapid rate of metal decay than actually
occurred. The problem in the volatilization area could be that the
complexes assumed to be formed had not formed in the time available
therefore, some free cyanide escaped initially, ani once the cyanide
did complex with the metals present, the decay rate slowed down. When
dealing with a solution such as the high mix which has an excess of
metals, the possible competition between metals and ligands (elf", OH-
3-
and P0
4
) has to be suspected, which in itself could slow down the com-
plexation process. According to the program REDEQL2 (McDuff ani Morel,
1974) used for finding the equilibrium compositions of complex
TAILB 9.7. 'l'B IIII1'UL PAIWIB'l'D YALUBS fOIl BIGH MIX CYAlUDB SOLtrrIOIf (SBCOle SDlDLA1'1OW)
(il) 20C 20C 4C 4C 20C 20C 4C 4C
High Mix AIR NO AIR AIR NO AIR AIR NO AIR AIR NO AIR
Siaulatloll NOUV NO UV NO UV NO UV UV UV UV UV
reR-TH(I) (aol/L) 0.0000
MCIf TCIf TH(2)
0.0076 0.0075 0.0076 0.0076 0.0074 0.0073 0.0069 0.0074
k TH(3) (h-
1
) 0.0394 0.0238 0.0074 0.0050 0.0389 0.0253 0.0234 0.0139
[Cu-cR] TR(4) (.ol/L) 0.0015 0.0015 0.0015 0.0015 0.0019 0.0014 0.0013 0.0019
[Zn-cR] TR(5) 0.0037 0.0036 0.0037 0.0037 0.0035 0.0035 0.0032 0.0035
[Ni-cR] TR(6)
0.0009 0.0009 0.0009 0.0009 0.0009 0.0009 0.0009 0.0009
[pe-CN] TR( 7) 0.0015 0.0015 0.0015 0.0015 0.0015 0.0015 0.0015 0.0015
kCu. TR(8) (h-
1
) 0.0075 0.0075 0.0029 0.0029 0.0075 0.0075 0.0029 0.0029
k
Zn
TR(9)
0.0449 0.0449 0.0178 0.0178 0.0449 0.0449 0.0178 0.0178
kwi TR(10)
0.0009 0.0009 0.0004 0.0004 0.0009 0.0009 0.0004 0.0004
k'e TR(ll)
0.0048 0.0048 0.0014 0.0014 0.0048 0.0048 0.0014 0.0014
kuv. TH(l2) 0.0002 0.0002 0.0002 0.0002
....
-l:"
....
0.75
.--..
s.
1
~
o
~
)(
z
Q 0.45
~
a:
t-
z
W
~ 0.
o
o
w
c
-
z
~
o
, .
\

\.
' ..
'.
.OBSERVED CN
r
-TOTALCN-j
'. .
'.
'.
" .
-,
-'-,
._-_. METAL CN- ESTIMATED
FREE CN-
-'-,
---.--.--.... ....,,.. .............. ..................... . = ....
............ .
.-._._._._._._._._._._._._._._.
..... _-...... .
...... .......... .
/ .............. .
" ............... .
..........................................................................
F1gure 9.3b Rlgh Mix - Second S1 ... 18t10n (4C, NO AIR, NO UV) .::-
N
aqueous systems, assuming equilibrium in this system at pH 7.0, zinc
3-
can be found in solid form with a P ~ ligand.
The general conclusion from the simulation attempts on high mix
is that the estimated kl values for the single metallo-cyanide solu-
tions are suitable for the mixes relatively low in metal concentration
but when it comes to the solutions oversaturated with metals, the metal
decay coefficient values may need reevaluation.
9.2 Evaluation of Model Calibration with Barren Solutions
9.2.1 Introduction
Three barren solutions were obtained from the well known gold
mines in Canada. The apparatus used for these experiments was the same
as the one used for synthetic solutions (Figure 4.1). Experiments were
conducted under what was considered to be the best conditions for the
natural degradation of cyanide to occur under a controlled environment,
i.e., 20C, AIR and UV.
The initial cyanide and metal concentrations in all three
barren solutions were compared to the final concentrations recorded at
the end of the experiment (see Table 9.8). It was difficult to compare
values obtained after different times of observation, (968 h vs 1320 h)
but since the concentrations of metals were not available during the
experiments, only at the beginning and the end, these values had to be
sufficient.
143
'fAILI ;.1. ..... IOLftlOW PAIAIII'I'III
TC1f Cu Zn
lanan ( .. IL) ( .. IL) (allL)
8olatlOtl8
I
,
I
,
I
,
'I (t, 132(11) 145 2.3 7.8 7.8 31 21(0.37)
'.
'2 (t, 132(11) 295 7.2 15.0 13.0 131 128(0.24)
13 (t, 968h) 128 6.3 7.7 2.4(2.3) 1.1 .1
Value. In parentha.i. areOaetal concentration. in a filtered ple.
tratlona.
I - Initial tiae.
, - 'inal tlae.
IU
"e
pR
( .. IL) (118/1.)
I
,
I
,
I
,
1.5 1.5 0.1 0.1 11.5 8.2
0.7 0.6 5.7 3.3( 0.1) 12.0 9.0
0.5 0.1 10.0 0.6 (0.5) 11.6 8.0
All other .et.l lue. .ra tot.l concen-


Zinc and copper were the metals with the highest concentra-
tions common for #1 and #2 barren solutions, while the #3 barren had
the highest concentration of iron, followed by copper.
The cyanide removal in all of them was:>95%. The concentration
of metals generally decreased (all except Ni). It is suspected that
with a change of pH from a high of 11.0 to 9.0 or 8.0, the metals
precipitated in the form of hydroxides or were adsorbed on the walls of
the container.
9.2.2 Simulations with the original kl values
At this point it was considered that the only differences
between the barrens and synthetic solutions, were the cyanide and
metal contents, and the pH of the solutions. From that aspect two
adjustments had to be made to the model. The first adjustment (Appen-
dix J.2) involved the assumption that barren solutions only contained
the same metallo-cyanide complexes as the synthetic solutions. The
theoretical stoichiometry for each complex was used in the calculations
for [MCN]o' which was subtracted from the analytically determined
[TCN]O in order to obtain [FCN]o. Since the model equations were
derived for synthetic solutions at pH 7.0, the effects of pH on the
volatilization of HCN had to be taken into consideration when modelling
barren solutions. Hence the appropriate adjustments made on the model
based on the equilibrium equation between HCN and CNr were explained in
145
Appendix J3. The calculated values of all initial parameters required
for the model simulation are presented in Table 9.9.
The results of the model simulation on barren solutions were
very similar to the results of high mix simulation. The model
volatilized cyanide too slowly and decayed metal cyanide too quickly
(Figures 9.4, 9.6a and 9.7a). The reason for this may be the original
assumption that the pH effect on metallo-cyanide complexes is
insignificant and that only pH effect on HCN and CNr ratio is
important. From another viewpoint, if metal cyanide in barren solution
was simply dissociating, as assumed, and releasing CNr, the high pH
would prevent formation of HCN. The released CNr ion may complex with
other metals, or form thiocyanate and precipitate, or oxidize into
CNO-, etc. There are many possibilities but not all of them could be
covered in the model obtained for synthetic solutions.
The final estimated TCN is compared to the experimental TCN
(Table 9.10). The observed and predicted TCN values for #1 barren
solution were at least within the same order of magnitude, but the TCN
predictions for #2 and #3 barrens were an order of magnitude lower than
experimental TCN values.
9.2.3 Analyses of barren solutions' parameters
In order to find some explanation for the discrepancies in the
model fit, other chemical characteristics of the barren solutions were
examined. Pertinent data are given in Table 9.11 .
146
147
TABLE 9.9. THE INITIAL PARAMETERS REQUIRED FOR THE MODEL SIKULATION OF
BARREN SOLUTIONS
#1 #2 #3
FCN=TH(I) (mol/L) 0.0032 0.0022 0.0034
MCN - TH(2) 0.0024 0.0091 0.0015
ky - TH(3) (h-
1
) 0.0389 0.0389 0.0389
[Cu-CN] - TH(4) (mol/L) 0.0004 0.0007 0.0004
[Zn-CN] - TH(5) 0.0019 0.0078 0.0001
[Ni-CN] - TH(6) 0.0001 0.00005 0.00003
[Fe-CN] - TH( 7) 0.00001 0.0006 0.001l
kCu TH(8)
(h-
1
)
0.0075 0.0075 0.0075
k
Zn
- TH(9) 0.0449 0.0449 0.0449
k
Ni
- TH(lO) 0.0009 0.0009 0.0009
kFe"" TH( 11)
0.0048 0.0048 0.0048
k - TH(l2) 0.0002 0.0002 0.0002
uv
1
N
~
)( 0.4
~
~
a:
~
w
~ 0.3:
8
w
c
z
c(
~ o.


,
,

I

I

,
,

\

\

' ..
' ..




OBSERVED CN
T
TOTAL CN- ~ ESTIMATED
----- METAL CN-J
I .......
0.0 ........... ,,:::;._.r'.-.
, ,. .,
o ;0
150 300 450 600 750 900 1050 1200 1350
TIME (h)
Figure 9.4 Barren Solution '1 - Original
1500
./:""
ex>
149
TABLE 9.10. ORIGINAL SIMULATIONS ON BARREN SOLUTION
Time (h) TeN (mol/L)
Barren Solutions Exper. Simulated Exper. Predicted
#1 1320 1320 0.88 x
10-
4
0.29 x
10-
4
#2 1320 1320 2.8 x 10-
4
1.6
x 10-
5
#3 968 970 2.4
x 10-
4
2.4
x 10-
5
TABLE 9.11. ADDITIORAL BARREN SOLUTIONS' PARAMETERS
Barren CNS- CNO- TCN CNW
Solutions (mg/L) (mg/L) (mg/L) (mg/L)
I F I F I F I F
11 51 57 23 0 145 8.2 145 2.1
112 428 412 125 72 295 7.2 286 4.7
113 1090 1074 56 23 128 6.3 96.5 4.6
~ - Weak acid dissociable cyanide.
I - Initial time.
F - Final time.
Total Phenolphthalein
Alkalinity Alkalinity
(mg/L as CaC0
3
) (mg/L as CaC0
3
)
I F I F
398 42.5 263
1,309 757 890 115
338 292 239
Ca
(mg/L)
I F
116 43
176 105
116 31.5
pH
I
11.5
12.0
11.6
F
8.2
9.0
8.0
V1
o
Initially, most of the total cyanide measured in the first two
barrens was.equal to the weak acid dissociable cyanide (compare TCN and
CN
W
)
This meant that cyanide was tied up in any other complex but
iron cyanide, i.e., susceptible to natural degradation even without UV
light present. The difference between TCN and CNW in the third
barren solution was assumed to be mostly the cyanide complexed with the
iron present in the solution.
The thiocyanate CNS- fluctuated slightly in all three barren
solutions, but generally it did not change significantly. There was no
noticeable production of cyanate (oxidation of CN- to CNO-) but the
initial CNO- gradually decomposed, probably into CO
2
and ammonium ion,
which were lost in the atmos?here.
The total alkalinity was extremely high in the second solu-
tion. This was probably the reason that slowed down the decrease of
its pH value (high buffering capacity). All three solutions had an ~
2-
alkalinity at the beginning, which became C0
3
and HCO; alkalinity at
the end of the experiments. In all barren solutions, some
precipitation of CaC03 also occurred during the experiment (see
Table 9.11).
9.2.4 Simulations with the adjusted.k
1
values
Taking into consideration the high concentrations of Zn and Cu
in two out of three barren solutions and the fact that these are the
metals with the highest decay coefficients,
(kCu
-
0.0075 and
151
kZn - 0.0449 h-
1
) it was postulated that these must control the rapid
metal decay rate produced in the model simulation. Therefore, the kl
values for these two metals were recalibrated using data from the first
barren solution. The values obtained, kCu ... 0.0022 h-
1
and kZn a:
0.003 h-
1
gave a good model to data fit (see Figure 9.5). Then the
recalibrated values were used to simulate the other two barrens.
Simulation of the #2 barren's data was just as good as for #1
(Figure 9.6b) but the model fit in the case of the third barren solu-
tion while improved was still poor (Figure 9.7b). There was a signifi-
cant improvement in the part of the degradation curve attributed to the
metal decay process in the first two barrens. This validates the
theory that Cu and Zn control this part of the curve. There was some
improvement in the model fit for the third barren solution (compare
Tables 9.10 and 9.12) but not significant enough. The estimated TCN
values are compared to the experimentally obtained TCN in Table 9.12.
The final values estimated for TCN are almost equal to the experimental
TCN concentration for #1 and #2 barren solutions.
In a way the minimal change in #3 barren solution simulation
was expected since this solution contained a higher Fe concentration
than Cu and Zn, therefore the changes in kCu and kZn would not have
much effect on the rapid metal de<7ay rate in this particular case.
But, when k
Fe
was recalibrated from 0.0048 (h-
1
) to 0.001 (h-
1
) the
model simulation on #3 barren changed and the fit improved signifi-
cantly (Figure 9.8a and b). All curves still showed discrepancies
152
0.8
j
-
0
....
)(
OBSERVED CN
T
z
o 0.4


TOTAL CN- ! ESTIMATED
----- METAL CN-
a:

w
0
z
0
0
UJ
Q
-
Z

0
O.
."
".
'.
'.
'.
,
. .,
.
. ,
"'.
.
. ' .. "
' .. "



..... .......
- -_.'.AO ......
.---------------



o 150 300 450 600 150 900 1050 1200 1360 1500
TIME (h)
Figure 9.5 Barren Solution '1 - Adjusted k Cu.ad k Zn
IJ1
W
1



.
I
i

Z

OBSERVED (:Ny
- toTAl or ESTIMATED
- METAlorj
\

a.






o a _
TIME (h)
08SERYED (:Ny
- toTAl or I ESTIMATED
- METAlorj
b


-.
. _---
o a _
TIME (h)
Figure 9.6 Barren Solution 12 - a) Original; b) Adjusted keu aDd
154
155
OBSERVED CNr
- tOTAl eN"' I ESTIMATED
1
- METAlCfir
..
52

I
b.
i
I
l




0 1110 IDO IDO III
'101)
IGD 1110 IGD IGD tIICID
T.ME(II)
Figure 9.7 Barren Solution '3 - a) Original; b) AdjuBted keu and kzn
156
TABLE 9.12. ID>IPIED SDIOl.A'l'IOli OR BAJUlD S<LU'I'IORS
Time (b) 'l'CB (1Dol/L)
Barren Solutions Exper. Simulated Exper. Predicted
11 1320 1320 0.88 x 10-4 0.9 x 10-'+
12 1320 1320 2.8
x 10-4 2.4 x 10-
1t
13 968 970 2.4 x 10-4
0.7 x 10-
1t
j 0.4
.....
~
JC .
. 0.
~
I
~ ~
w
e
z 0.1
~
o
i
12


I
~ ~
I
~




0 100 Il1O XI
OBSERVED CNy
- tOTAL eN"" , ESTIMATED
- METALCN""
8.




eoo eoo JOO 100 100 tDOO
TIME (h)
OBSERVED eN
T
- TOTAL eN"" } ESTIMATED
----- METAL eN-
b_
' .....
., .......
................. . ' .
.................... -
-.... _------.... -.... _--

------
100 200 1500 eoo 7DO 800 800
TIME (h)
Figure 9.8 Barren Solution '3 - a) Original; b) Adjusted ~
157
1000
158
during volatilization process (the first part of the curve). This is
suspected ~ be due to a change in pK value, and the increase in ionic
strength of the solution. Ionic strength correction would change the
effect of pK by a small amount; this change would not significantly
influence the curve predicted by the model.
The general conclusion fran the model simulations is that the
volatilization mass transfer coefficient and the values obtained for
metal decay rate for single metallo-cyanide complex solutions, were
very good for these particular solutions and to some extent for a low
mix. However, in order to be used for high mix and barren solutions,
more observation and some recalibration of the parameters is necessary.
10 CONCLUSIONS
The long-term. objective of this work is to model the kinetics
of removal of cyanide from tailings ponds containing gold mill
effluents. The approach in this research was to emulate the general
characteristics of the actual systems and by creating a synthetic
solutions, examine them under controlled conditions and then compare
the results with observations on the actual systems.
The thesis describes the experimental design used in the
laboratory to delineate major factors affecting cyanide removal in the
synthetic solutions and modelling the results of the synthetic and
actual barren solutions.
The general conclusions from the experiments are the following:
1) Volatilization and metal cyanide decay are the controlling
mechanisms for removal of cyanide from solution; volatilization is a
very rapid process, therefore metal cyanide decay is the rate limiting
mechanism.
2) Analysis of the results from the full 2
3
factorial design
show that temperature has the most significant effect on the volatili-
zation coefficients of all cyanide solutions examined, including NaCN
solution. The aeration effect on the volatilization coefficient is
159
160
significant combined with temperature, but only for the first 48 to 72
hours of the volatilization process.
3) Temperature has a large main effect upon the metal cyanide
decay in most of the solutions. UV light has an effect on Fe-CN and
low mix solutions.
4) A model was derived to describe cyanide removal from solu-
tion. It assumes that the volatilization of HCN and the decay rate of
one metallo-cyanide complex control the disappearance of total cyanide
from single metallo-cyanide solutions.
5) The rate of volatilization of HCN (ltv) from the single
metallo-cyanide solutions was determined to be within 95% confidence
interval of the volatilization rate of HCN in sodium cyanide solu-
tions. The means and standard deviations for kv values at 4C and
20C were: 0.0124 0.008 and 0.0318 0.008, respectively. The metal
decay rates varied among different metals and at different test
conditions.
6) The best estimates of the metal decay coefficients from
single metallo-cyanide complex solutions and their energy of activation
using Arrhenius's equation were determined:
Metallo-Cyanide
at 4C (h-
1
) at 20C (h-
1
) Complex Solutions
kl kl
E (kJ/mol)
Na20l( CN) 3 0.00295 0.00753 39.4
Na
2
Zn( CN)I+ 0.01783 0.04496 39.0
Na2Ni( CN) 1+ 0.00044 0.00095 33.0
K
3
Fe(CN)6 0.00143 0.00473 50.4
161
These energy of activation values seem to point out that the reactions
occurring. are on a border where molecular diffusion control ends and
chemical reaction control starts.
7) These metal decay coefficients were used to simulate the
experimental data of the mixed metallo-cyanide solutions. They
produced a gocxi fit for the low mixes. The model fits the data at the
95% confidence level for five of the eight test conditions.
8) The model, calibrated using the rate coefficients from the
single synthetic solutions was applied to the high mix and the actual
gold mill effluents (barren solutions). For the high mixes and barren
solutions the results showed that the model can be used for the basic
estimate of the cyanide degradation rate. For the more precise results
the recalibration of the parameters is necessary as a function of the
solution's metal content.
11 JlEC<ltKKRDA.TIOBS
The following are recommended for future modelling work:
1. Experiments with the solutions containing single metallo-
cyanide complexes with the high concentration of metal, in
order to determine the metal cyanide decay rate that could be
used for modelling high mixes.
2. Recalibration of the metal decay rates for at least 10 to 15
different barren solutions, to determine the best values that
could be used for modelling other similar solutions.
3. Examine the possibility of the formation of more than one
metallo-cyanide complexes in the high mixes and barren
solutions.
4. To develop a model for dynamic systems that could be applied to
cyanide removal from the tailings pond that is continuously
being filled.
The recommendations that could be applied to any future study
concerning cyanide in gold mill effluents, are the following:
1. Detailed characterization of the individual gold mill
effluents is required before treatment.
2. Investigate techniques for obtaining and preserving mine
process samples containing particulates.
162
3. Investigate the possibility of refining existing methods and/or
better methods for quantifying cyanide species in
process effluents.
4. Study microbial degradation of cyanide in mine tailings ponds
over long retention times.
5. Develop a model with which to predict the equilibrium chemical
species of all interactive inorganic and organic chemicals in
mill effluents.
6. Devise a method to accelerate natural degradation of cyanide
and metallo-cyanide compounds in tailings impoundments.
163
REFERENCES
Adamson, A.W., J.P. Welker and G. Volpe. "Exchange studies with complex
ions. I. The exchange of radiocyanide with certain heavy metals complex
cyanides. " Journal of the American Chemical Society 72: 4030-4036, Sept.
1950.
AlesH, B.A. and W.H. Fuller. "The mobility of three cyanide forms in
soil." U.S. Environmental Protection Agency, National Solid and Hazardous
Waste Research Center, EPA-600/9-76-016, 1976.
164
Amax of Canada Ltd., "Background information on cyanide/mining/environ':'"
ment." Vancouver, B.C., Aug. 1980.
APHA-AWWA-WPCF. "Standard methods for the examination of water and waste-
water." 15th ed. Am. Publ. Health Ass., Am. Water Works Ass., Water
Pollute Contr. Fed., Washington, D.C. p. 1134, 1980.
Asperger, S. "Kinetics of the decomposition of potassium ferrocyanide in
ultra violet light." Transactions Faraday Society, 48(7):617-624, 1952.
Balzani, V., and V. Carassiti. Photochemistry of coordination compounds.
Academic Press,. New York, N.Y. 432 p. 1970.
Bard, Y. Nonlinear parameter estimation. Academic Press, New York, N.Y.
1974.
Box, G.E.P., W.G. Hunter, J.S. Hunter. Statistics for experiments. John
Wiley & Sons, Inc., New York, N.Y. 1978.
Broderius, S.J. "Determination of molecular hydrocyanic acid in water and
studies of the chemistry and toxicity of fish of metal cyanide complexes."
Ph.D. dissertation, Oregon State University, Corvallis, Ore. 1973.
Broderius, S.J. and L.L. Smith, Jr. "Direct photolysis of
hexacyanoferrate complexes. Proposed applications to the aquatic
environment." U.S. Environmental Protection Agency EPA-600/3-80-003, 50
pp. 1980.
Brunker, R.L. "The biological degradation of cyanides by autotrophic
organisms. Industrial Wastewater Treatment." Proc. 12th Midatlantic
Industrial Waste Confer., Bucknell Univ. Lewisburg, P.A.
1980.
pp. 146-151,
165
Burdick, G.E. and M. Lipschuetz. "Toxici ty of ferro- and ferricyanide
solutions to fish, and determination of the cause of mortality." Transac-
tion of the Amer. Fish Soc. 78:192-202, 1950.
Caruso, S.C. "The chemistry of cyanide compounds and their behavior in
the aquatic environment." Carnegie-Mellon Institute of Research,
Pittsburg, P.A. 1975.
Clevenger, G.R. and R. Morgan. "Atmospheric decomposition of cyanide
solutions." Mining and Scientific Press. p. 413-423, Sept. 1916.
Cotton, F.A. and G. Wilkinson. Advanced inorganic chemistry. A compre-
hensive text. Interscience Publishers, A division of John Wiley & Sons.
1962.
Conn, K. "Cyanide analysis in mine effluents." Presented at Canadian
Mineral Processors 13th Annual Meeting, Ottawa. Jan. 1981.
Conn, K. Private communication, P S ~ Wastewater Technology Centre,
Burlington, Ont. 1984.
Crouse, W.C. and D.W. Margerum.
p. 1437, 1974.
Inorganic Chemistry, Vol. 13, No.6,
166
Daniel, C. "Use of half-normal plots in interpreting factorial two-level
experiments." Technometrics. 1:311-341, 1959.
Dean, J.A. (Ed.) Lange's handbook of chemistry, 11th ed. McGraw-Hill,
Toronto, Ont. 1973.
Dodge, B.F. and D.C. Reams. "Disposal of plating room wastes. 1. A
critical review of the literature pertaining to the disposal of waste
cyanide solutions." Plating, 36, 463-9, 512, 571-7, 664, 723-5, 728-32.
1949.
Dodge, B.F. and W. Zabban. "Disposal of plating room wastes. IV. Batch
volatilization of hydrogen cyanide from aqueous solutions of cyanides."
Plating 30, 1133-9, 1952.
Doudoroff, P. "Some experiments on the toxicity of complex cyanides to
fish." Sewage and Ind. Wastes 28(8): 1020-1040, 1956.
Doudoroff, P. and G. Leduc and C.R. Schneider. Acute toxicity to fish of
solutions containing complex metal cyanides, in relation to concentrations
of molecular hydrocynic acid." Trans. Am. Fish. Soc. 95:6-22, 1966.
167
Doudoroff, P. "Toxici ty to fish of cyanides and related compounds. A
review." Environmental Research Laboratory. EPA-600/3-76-038, 155 pp.
Duluth, Minnesota, 1976.
Draper, N.R. and R. Smith. Applied Regression Analysis. John Wiley &
Sons, Inc. New York, N.Y., 407 p. 1966.
Ecological Analysts, Inc. "Cyanide: An overview and analysis of the
literature on chemistry, fate, toxicity and detection in surface waters."
Prepared for The Inter-Industry Cyanide Group, Towson, Maryland, U.S.A.
1979.
Fairbridge, R.W. The encyclopedia of geochemistry and environmental
sciences. Volume IVA. Van Nostrand Reinhold Company, New York, N.Y.,
1972.
Ford-Smith, M.R. "The chemistry of complex cyanides: A literature
survey." Her Majesty's Stationary Office, London, 93 p., 1964.
Freeman, P.A. "Wind driven aerators for lagoons." Workshop on Low-Cost
Wastewater Treatment, Clemson, S.C., p ~ i l 19-21, 1983.
Freiser, H. and Q. Fernando. "Ionic equilibria in analytical chemistry."
John Wiley & Sons, Inc., New York, N.Y., 334 p, 1963.
168
Gaudin, A.M. Flotation. McGraw-Hill Book Company, Inc. New York, N.Y.,
165 p. 1932.
Halli burton, D. "Discussion paper on water pollution control in the
primary gold mining and milling industry." Prepared for the Water
Pollution Directorate, Ottawa, Ont., March 1980.
Hedley, N. and H. Tabachnick. "Chemistry of cyanidation."
Cyanamid Company, New York, Number 23, p. 54, 1958.
American
Hendrickson, T .N. and L.G. Daignault.
compounds for reuse and disposal."
Environmental Protection Agency, 1973.
"Treatment of complex cyanide
Report No. EPA-R2-73-269, U.S.
Howe, R.H.L. "Bio-destruction of cyanide wastes - advantages and disad':"
vantages." Int. J. Air Water Pollute 9:463-478, 1965.
Huiatt, J.L., J.E. Kerrigan, F.A. Olson, and G.L. Potter. "Proceedings of
a Workshop. Cyanide from mineral processing." Utah Mining and Mineral
Resources Research Institute, Salt Lae City, Utah. Feb. 1983.
Hyatt, D.E. "The chemical basis of techniques for the decomposition and
removal of cyanides." Society of Mining Engineers, AlME Transactions,
Vol. 260, p. 204-8, Sept. 1976.
169
IEC. "Factors affecting natural degradation of free and metal-complexed
cyanides from gold milling effluents." International Environmental
Consultants Ltd., Toronto, Ont., 45 pp., 1979.
Ilyalemdinov, A.N., P.B. Enker and Z.G. Vlasova. "Role of Microorganisms
in Removing Cyanides from Effluents." Mikrobiologiya 46(2):358-362
(Russ.) Chem. Abs. 87:90174h, March 1977.
Ingles, J. "Chemistry and toxicity of cyanides in gold mill effluents."
A discussion paper prepared for the 10th meeting of the Gold Processors
Working Group, Environment Canada, Hull, Quebec, 1981(a).
Ingles, J. "The treatment of mine mill effluents for the removal of iron
cyanides." Prepared for Mining and Metallurgical Division, Water
Pollution Control Directorate, EPS, Hull, Quebec, 1981b.
Knowles, C.J "Microorganisms and cyanide . Bacteriol. Reviews
Vol. 40(3):652-680, 1976.
Kolski, G.B. and D.W. Margerum. "Kinetics of formation and dissociation
of tetracyanonickelate (II) Ion." Chemistry 7( 11) :2239-2243,
Nov. 1968.
170
Kalthoff, I.M., E.B. Sandell, E.J. Meehan, S. Bruckenstein. Quantitative
Chemical Analysis, Published by The Macmillan Company, Collier-Macmillan
Ltd., London, 1969.
Kunz, R.G., J.P. Casey and J .E. Huff. "Refinery cyanides: A regulatory
dilemma." Hydrocarbon Processing, Oct. 98-106. 1978.
Leduc, G., R.C. Pierce and I.R. McCracken. "The effects of cyanides on
aquatic organisms with emphasis upon freshwater fishes."
Research Council of Canada, Ottawa, 1982.
National
Long, F.A. "Exchange reactions of tetracyanonickelate ion: The structure
of nickel cyanide." Journal of the American Chemical Society 73:537-540,
Feb. 1951.
Ludzack, F.J. and R.B. Schaffer. "Activated Sludge Treatment of Cyanide,
Cyanate and Thiocyanate." Proceedings of the 15th Industrial Waste
Conference, Purdue University, p 439, 1960.
Ludzack, F.J., W.A. Moore, H.L. Krieger, and C.C. Ruchhoft. "Effect of
cyanide on biochemical oxidation in sewage and polluted water." Sewage
Ind. Wastes 23:1, 298-1, 307. 1951.
171
Lur'e, Yu. Yu. and V.A. Panova. "The behaviour of cyanide compounds in
water ponds." Gidrokhim. Materialy 37:133-143, 1962.
Luthy, R.G. and S.G. Bruce. "Kinetics of reaction of cyanide and reduced
sulfur species in aqueous solution." Environ. Sci. Techno!. 13: 1481-1487,
1979.
Luthy, R.G., S.G. Bruce, R.W. Walters, D.V. Nakles. "Identification and
reactions of cyanide and thiocyanate coal gasification wastewaters."
Presented at the 50th Annual Conference Water Pollution Control
Federation, Philadelphia, P.A. Oct. 1977.
MacDiarmid, A.G. and N.F. Hall. "Illumination - pH effects in solutions
of complex cyanides." J. Amer. Chem. Soc., 75, 5204-7, Nov. 1953.
McDuff, R.E., F.M.M. Morrel. "Description and use of the chemical
equilibrium program REDEQL2." Technical Report EQ-73-02, California
Institute of Technology, Pasadena, California. 1974.
Meeter, D.A. and P.J. Wolfe. "Non-linear least squares (GAUSHAUS)."
University of Wisconsin Computing Center, Wisconsin. 1965.
Milne, D. "Equilibria in dilute cyanide waste solutions." Sewage and
Industrial Wastes, 22(7):904-911. 1950a.
172
Milne, D. "Disposal of cyanides by comlexation." Sewage and Ind.
Wastes. 22(9):1192-1199. 1950b.
Moir, J. "Loss of cyanide and value of protective alkali." Journal of
Chern. Met. and Min. Soc. of South Africa, Vol. 16, p. 9. 1910.
Moore, W.J. Physical Chemistry, 3rd Ed., Prentice Hall Inc., Englewood
Cliffs, N.J. 1962.
Murphy, R.S. and J.B. Nesbitt. "Biological treatment of cyanide wastes."
Engineering Research Bulletin B-88. The Pennsylvania State University,
University Part, P.A. 66 pp. 1964.
Noblitt, R.L. "Complex ions of copper and cyanide." Department of
Energy, Mines and Resources. Mines Branch, Ottawa, Sept. 1973.
Nriagu, J .0.
Ont. 1980.
Nickel in the Environment. John Wiley & Sons, Toronto,
Palaty, J. and M. Rorokova-Jakubu. "The course and rate of removal of
cyanides from water under natural conditions." Scientific paper from
Institute of Chemical Technology, Prague, Faculty of Technology of Fuel
and Water V3, Part 1. 1959.
173
174
Penneman, R.A., R. Rain, G. Gilbert, L.H. Jones, R.S. Nyholm, and
G.K.N. Reddy. "An infrared study of pentacyano-,hexacyano-, and
chloropentacyano-nickelate (II) in aqueous solutions."
Chemical Society, p. 2266-2274, 1962.
Journal of the
Perry, R.H. Engineering Manual. McGraw-Hill Book Company. 1976.
Pol' kin, S.1. and V.N. Krylova. "Promising trends in the removal of
cyanide compounds from the waste waters of concentration plants." Moscow
Inst. of Steel & Alloys, USSR, No.2, p. 3-7, 1979.
Raef, S.F., W.G. Characklis, M.A. Kessick, and C.H. Ward. "Fate of
cyanide and related compounds in aerobic microbial systems. I. Chemical
reaction with substrate and physical removal." Water Research,
11:477-483, 1977.
Renn, C.E. "Biological properties and behaviors of cyanogenic wastes."
Sewage and Industrial Wastes. 27(3):297-308, 1955.
Resnick, J.D., W.A. Moore and M.B. Ettinger. "Behavior of cyanates in
polluted water." Ind. Eng. Chem. 50(I)January. 1958.
Scott, J.S.
effluents."
and J.C. Ingles. "Removal of cyanide from gold mill
Presented at the 13th Annual Meeting of Canadian Mineral
Processors, Ottawa, Ont. 1981.
Schmidt, J.W., L. Simovic and E.E. Shannon. "Development studies for
suitable technologies for the removal of cyanide and heavy metals from
gold milling effluents." Proc. 36th Ind. Waste Conf. Purdue University,
Lafayette, Indiana. p. 831-846. 1981.
Sharpe, A.G. "The chemistry of cyano complexes of the transition metals."
Academic Press, New York. 302 pp. 1976.
Shaw, M.P. "A kinetic model for the natural dechlorination of municipal
effluents." M. Eng. Thesis, McMaster University, Hamilton, Ont. 1983.
Smith, J .M.
1981.
Chemical Engineering Kinetics, McGraw-Hill Book Company,
Stumm, W. and J.J. Morgan. Aquatic Chemistry. John Wiley & Sons, New
York, N.Y. 1981.
Taggart, A.F. Handbook of mineral dressing, ores and industrial
minerals. John Wiley and Sons, Inc. New York, Sec. 12, p. 26. 1945.
175
Tarzwell, C.M. "Discussion" of C.E. Renn's paper "Biological properties
and behaviors of cyanogenic wastes." Sewage and Industrial Wastes
27(3):308-310, March 1955.
Towill, L.E., J.S. Drury, B.L. Whitfield, E.B. Lewis, E.L. Galyan, and
A.S. Hammons. "Reviews of environmental effects of pollutants: v.
Cyanide. " Oak Ridge National Lab. ORNL/EIS-81. EPA-600/1-78-027.
190 pp. 1978.
Tremblay, R. "Laboratory testing for flowsheet development and optimiza-
tion. " Presented at the 18th Annual Program, Professional Development
Seminars, Mineral and Metallurgical Engineering, Mineral Processing of
Gold Ores. McGill University, Montreal, Oct. 31-Nov. 2, 1983.
U. S. Envi ronmental Protection Agency. "Cyanides, ambient water qual1 ty
criteria." Criteria and Standards Division, Office of Water Planning and
Standards. u.S. EPA. 1979.
Weber, W.J. Jr., and O. Corapcioglu. ""Catalytic oxidation of cyanides.""
Proc. 36th Annual Industrial Waste Conference, Purdue University,
Lafayette, Indiana. 1981.
176
APPENDIX A
Cyanide Tenainology
177
A CYAliIDE TERMINOLOGY
Cyan,ide ion refers to the free cyanide ion cN"".
Molecular HCN - the cyanide in form of an uncharged intact
molecule in aqueous solution. Therefore the terms hydrogen cyanide,
hydrocyanic acid and HCN all refer to cyanide as molecular HCN.
Free cyanide - the summation of molecular HCN and the cyanide
ion (HCN + CN"") in aqueous solution.
Simple cyanide refers to a cyanide compound which dissociates
directly in water to a cation and the cyanide ion with no soluble
intermediates.
Complex cyanide - cyanide compound which dissociates in water
to a cation and a negative ligand which may be subject to further
dissociation.
Weak-acid dissociable cyanides (CNW) - all forms of cyanide
excluding iron recovered by the chemical analysis.
Total cyanide (TCN) includes free cyanide, and all the simple
and complex cyanides.
178
APPENDIX B
Determination of Cyanide by
Picric Acid and Distillation Methods
179
B.l DETERMINATION OF CYANIDE BY PICRIC ACID METHOD
Principle
Cyanide ion and simple cyanides react with picric acid at a pH
of 10 at elevated temperature to form a stable coloured complex. The
complex obeys Beer's law from 0 to at least 500 pg of CNr (Kolthoff et
al., 1969). With a 5 cm cell, in the spectrophotometer, the detection
limit is 1.3 pg.
Interferences
Thiocyanates and cyanates in concentrations up to 50 mg/L do
not interfere. Sulphide ion causes the most interference, 0.1 mg of
sulphide being equivalent to 0.025 mg of cyanide ion.
interferes at levels in the microgram range.
Precautions
Sulphite ion
Picric acid is also known as trinitrophenol, and is explosive.
It is not readily detonated and is not particularly hazardous to use,
but all spills should be wiped up and solutions should be disposed of
by washing down a sink with copious amounts of water. Aqueous and
organic solutions act as dyes and stain many materials, including the
skin. Most of the stain can be removed from working surfaces and
glassware by means of acetone, but the hands should be protected with
rubber gloves.
Storage
Samples should be preserved by adding enough NaOH (either 50%
w/w or pellets) to raise the pH to 12.
180
181
Apparatus
Regular laboratory glassware, a hotplate capable of heating
samples at 90-95C, and a spectrophotometer for measuring optical
density at 490 nm.
Reagents
1. Stock cyanide solution (1 mL = 1000 ~ CN-)
Dissolve 1.884 g NaCN in distilled water and dilute to 1000
mL. Adjust the pH to 12 with NaOH. Standardize as follows: Dilute 5.0
mL aliquot of stock solution in 100 mL of distilled water at pH 11 or
higher, add 15 drops of rhodanine indicator solution ani titrate with
standard AgN0
3
solution to a salmon-pink end point.
2. Working cyanide solution
3.
Prepare daily by appropriate dilution of stock solution.
Rhodanine indicator solution
Dissolve 1.632 g AgN0
3
in distilled water and dilute to
1000 mL. 1 mL = 500 ~ C ~
4. Standard silver nitrate solution
Dissolve 1.632 g AgN0
3
in distilled water ani dilute to
1000 mL. 1 mL = 500 ~ g CNr.
5. Picric acid solution
- 1% aqueous.
6. O.lM EOTA (37.22 gIL of EOTA - disodium salt)
7. EOTA - Na
2
C0
3
solution
- 15 gIL EOTA (disodium salt)
- 52 gIL Na
2
C0
3
Preparation of Calibration Curve
1. Into a series of 250 mL volumetric flasks, pipet appropriate
amounts of cyanide working solution to cover the working range
up to 500 ~ g C ~
2. Add enough distilled water to bring the volume to approximately
100 mL.
3. Include a flask containing no cyanide as a blank.
4. Continue from the beginning of step 2 under "Analysis of
Samples"
Analysis of Samples
1. Pipet an appropriate aliquot of sample up to 10 mL* into a
250 mL volumetric flask. The aliquot should not contain more
than 500 ~ g CNr. Bulk to about 100 mL.
2. Add 2 mL of 0.1 EDTA solution, 10 mL of the EDTA - Na2C03
solution and 15 mL of the 1% picric acid solution.
3. Heat on the hotplate at 90 to 95C for 20 min.
4. Cool to room temperature with cold water; and dilute to 250 mL
with distilled water.
5. Measure the absorbance in a spectrophotometer at 490 nm with a
5 em cell.
6. Construct a calibration graph by plotting the absorbence vs ~ g
of cyanide.
Note: *Volumes larger than 10 mL may be taken provided that the pH is
adjusted to 10 with pH meter and buffer.
182
7. From the calibration graph, read the number of pg of cyanide in
sample and blank solutions.
8. Calculate the concentration of ~ in the sample:
crr (mg/L) ...
a
b
where a - pg of ~ in sample aliquot
b - volume of sample taken for analysis
183
B.2. CYANIDE DETERKIJlATION BY DISTILLATION METROD
Principles
Most complex cyanides (except gold and cobalt complexes) can be
coverted to molecular HCN gas by boiling with hydrochloric acid and
hydroxylamine hydrochloride. The HCN gas is liberated under reduced
pressure, absorbed in a solution of IN sodium hydroxide and determined
either titrimetrically or colorimetrically, depending on the amount of
cyanide present. Sulphides are removed from the sample by treatment of
the sample with lead carbonate. Sulphides originating from the sample
as ei ther sulphide or thiocyanate are precipitated in the absorbing
solution as cadmium sulphide using cadmium chloride.
Using the titration (for samples above 5 mg/L C ~ , an aliquot
of the sodium hydroxide absorbing solution is titrated with silver
nitrate until an excess of silver ion is detected with a special silver
ion indicator. In the colorimetric test (for samples below 5 mg/L
C ~ , cyanide ion is first converted to cyanogen chloride (CICN) using
chloramine T. This in turn forms a blue complex with a sodium
isonicotinate-sodium barbiturate reagent. Absorbence of this complex,
measured at 600 nm, is proportional to the amount of cyanide present in
the aliquot of sodium hydroxide taken for analysis.
Using the described procedure, this method has been used on
samples with a minimum reportable value of 0.01 mg/L cyanide.
Storage
Samples must be preserved by adding enough sodium hydroxide
(either pellets or 50% w/w) to raise the pH to 12. For samples
184
185
containing high levels of soluble sulphides, one precipitates the
sulphides first using lead carbonate, filter, before raising the pH to
12.
Apparatus
The distillation system (Figure B.l) consists of: (1) 3-neck
one litre distilling flask, (2) Friedricks condenser (315 um), (3)
Thistle tube with ground glass joint to fit flask (the tube must extend
below the level of liquid in the flask), (4) 125 mL gas absorbing
bottle (impinger) with ground glass joint and fritted disc, and (5) low
vaculDD. system. The tube joining the condenser (2) to the gas washing
bottle (4) should be glass as far as possible. The air entering the
thistle tube should be scrubbed with IN NaOH to remove any airborn HCN.
Heating Mantle
Spectrophotometer capable of measuring at 600 nm
Reagents
1. Sodium hydroxide - IN. Dissolve 40 g NaOH in distilled water
and dilute to 1 L.
2. Acetic acid - Glacial.
3. Acetic acid - 1:4. Slowly add 100 mL glacial acetic acid to
400 mL distilled water.
4. Chloramine-T solution - Dissolve 1 g chloramine-T in distilled
water and dilute to 100 mL. Prepare daily.
5. Isonicotinic acid - barbituric acid reagent. In 1000 mL of
distilled water at 60 to 70C dissolve 12 g of NaOH, 20 g of
isonicotinic acid and 10 g of barbituric acid. After cooling,
adj ust t he pH to 8.0 with ace tic acide.
186
(2. )
___ vacuum
(5.)
(4. )
Figure B.l ~ Dtati11atian Brat ..
187
6. p-Nitrophenol Indicator - 0.1% w/v solution in ethanol.
7. Standard Silver Nitrate Solution. Dissolve 1.632 g AgN0
3
in
distilled water and dilute to 1000 mL. 1 mL - 500 ~ CNr.
8. Rhodanine Indicator Solution. Dissolve 0.02 g p - dimethyl-
amino-benzal-rhodanine in 100 mL acetone.
9. Stock Cyanide Solution (1 mL - 1 mg CNr). Dissolve 1.884 g
NaCN in distilled water and dilute to 1000 mL. Adjust pH to at
least 12. Standardize every day as follows: Dilute 1.0 mL
aliquot of stock solution in 100 mL of distilled water at pH 11
or higher, add 0.5 mL of rhodanine indicator solution and
titrate with standard AgN0
3
solution, to a salmon pink end
point.
10. Working Cyanide Solution (l mL - 1 ~ CN-). Prepare daily.
Dilute 1 mL of stock and 10 mL of IN NaOH to 1000 mL with
distilled water.
11. Hydrochloric Acid - hydroxylamine hydrochloride reagent 50%
(V Iv) - 10% (W/v) - Dissolve 100 g NH
2
0H.HCl in 300 mL of
distilled water. Add 500 mL of hydrochloric acid and dilute to
1000 mL.
12. Cadmium Chloride. Powder.
Procedure
1. Place 400 mL of distilled water into the lL boiling flask.
Insert the thistle tube and rinse the diffuser with distilled
water.
2. Add 100 mL IN NaOH, 0.2 to 0.3 g cadmium chloride powder to the
gas absorbing bottle (Note 2). Turn on the vacuum, and insert
the diffuser into the gas absorbing bottle.
3. Close the system by making the connection between the condenser
and the gas absorbing bottle. Into the boiling flask through
the thistle tube add 25 mL of HCI - HH, rinse with distilled
water, add 100 mL of sample (see Note 1).
4. Turn on the heat and cold water and allow the mixture to boil
for 45 min.
5. After boiling, turn off the heat and allow the vacuum to
operate for 15 to 20 min.
6. Disconnect the condenser from the gas absorbing bottle, turn
off pump and water, disconnect and reconnect the vacuum a
couple of times to transfer any cyanide from inside the air
diffuser out to the solution in the bottle.
7. Remove the diffuser from the absorbing solution while the
vacuum is operating. Also rinse the diffuser with 25% HCl and
turn off the vacuum. Filter the absorbing solution through
acid washed #40 Whatman paper (see Note 3) into a polyethylene
bottle.
8. Clean the entire system tho,roughly between samples using 25%
hydrochloric acid (V/V). Rinse the boiling flask with
distilled water.
188
Notes
1. For samples containing high levels of sulphide, ensure that
they have been treated for sulphide removal as soon as possible
after sampling. Add sufficient lead carbonate to remove all
sulphide, filter and raise the pH to 12.
2. Cadmium chloride is added to the gas absorbing bottles when
samples are known or suspected to contain sulphide,
thiocyanate, or thiosulphate.
3. Before filtering the absorbing solution through #40 Whatman
paper, rinse the paper once with 25% HCI and twice with
distilled water.
A. Titrimetric Method - to be used on samples containing 5.0 mg/L
cyanide or greater.
1. Treat sample by distillation procedure.
2. Pipet an aliquot (up to 50 mL) of the IN Sodium hydroxide
absorbing solution into a 300 mL Erlenmeyer flask. Dilute to
about 100 mL. Ensure that the pH is at least 11. Add 0.5 mL
of rhodanine indicator.
3. Titrate with silver nitrate working solution to a salmon pink
end point.
4. Calculations
mg/L Cyanide 8S CN! _ (A-B)500
C
100
x
D
189
where:
A = mL of silver nitrate required to titrate sample aliquot
B - mL of silver nitrate required to titrate blank
C - mL of sodium hydroxide absorbing solution taken for
titration
D = mL of distilled sample
B. Colorimetric Method - to be used when cyanide concentration is
below 5.0 mg/L CN.
1. Treat sample by distillation procedure.
2. Pipet either 1 mL or 10 mL of absorbing solution (depending on
cyanide concentration being measured) into a 25 mL volumetric
flask. For samples in the range 0 to 1.0 mg/L CN:, use 10 mL.
3. Adjust the volume to about 15 mL with distilled water, add 2
drops of p-nitrophenol indicator and neutralize to a colorless
end point using 1:4 acetic acid.
4. Repeat for a blank and standards (1 mL or 10 mL IN NaOH) ,
depending on the level of cyanide being measured.
5. To each flask, add 1.0 mL of chloramine T solution (1%) and
mix. Allow to stand for 2 to 5 min.
6. Add 5.0 mL of the sodium isonicotinate - sodium barbiturate
reagent, dilute to volume n ~ mix.
7. After 30 min read the absorbence at 600 nm (1 cm cells) using
the prepared blank as a reference. If working 1n the 0 to
1.0 mg/L range, use the 10 mL blank; if working 1n the 0 to
10 mg/L range use the 1 mL blank.
190
191
8. Construct a graph of recorded absorbences for standards and
corresponding llg of cyanide present. One graph each for 0 to
1.0 mg/L range and one graph for 0 to 10 mg/L range.
9. Calculation:
A x B
mg/L Cyanide as Crr = ---
C x D
where:
A = llg of cyanide ion from graph
B = volume of absorbing solution (usually 100 mL)
C - original aliquot of sample taken for distillation (usually
100 mL)
D = volume of absorbing solution taken for colorimetric
analyses (lor 10 mL).
192
APPENDIX C
Experimental Results
'rAIL. C.l. IIPIUIaUL DIUL'I'II .... lOa CDIIID& U 4C
4C 4C 4C 4C
All lIO Alit AIR NO All
UV UV NO UV lIO Uf
Synth.tic
Cyaald. % % %
% .
801utloa t(h) CII
T
(q/L) ae.o ... t(h) CRT(q/L) ....... t(h) CII
T
(q/L) ....... t(h) CII
T
(q/L) ........
laC"
0 203.0 0.0 0 200.0 0.0 0 202.0 0.0 0 225.0 0.0
7 176.4 13.1 7 119.0 40.5 19 148.0 26.6 17 158.9 29.4
24 129.9 36.0 24 111.6 44.2 24 140.7 30.4 24 105.7 53.0
48 104.8 48.4 31 88.8 55.6 41 108.2 46.4 41 98.2 56.3
~
76.9 62.1 48 73.3 63.4 48 82.1 59.3 48 91.9 59.1
72 38.4 81.0 72 48.3 75.9 65 (90.0) 55.4 65 (88.3) 60.7
79 29.2 85.6 144 17.2 91.4 72 Sl.O 13.8
1!
65.8 70.7
144 10.6 94.8 168 13.3 93.4 113 ( 58.3) 71.1 137 63.6 71.7
1'51 (10.6) 94.8 175 11.3 94.4 144 41.4 79.5 144 62.5 72.2
168 6.9 96.6 192 10.9 94.5 151 44.1 78.2 161 60.2 13.2
175 5.0 97.5 199 (11.9) 94.0 168 37.1 81.6 168 58.5 74.0
192 3.3 98.4 216 5.3 97.3 175 36.3 82.0 185 56.1 75.1
216 1.2 99.4 223 4.6 97.7 192 36.1 82.1 192 (45.4) 19.8
240 0.5 99.7 240 4.0 98.0 199 (37.6) 81.4 209 46.8 . 19.2
312 (0.5) 99.7 288 2.9 98.5 216 33.1 83.6 216 (39.1) 82.6
312 2.6 98.7 223 24.5 87.9 233 39.3 82.5
319 2.4 98.8 241 23.2 88.5 256 38.3 83.0
247 21.8 89.2 280 36.3 83.9
312 20.4 89.9 296 35.1 84.4
317 28.6 87.3
ky(h-
1
) 0.0159 (0.665 c./h) 0.0193 (0.808 c./h) 0.0176 (0.728 c./h) 0.0164 (0.686 ~ / h
It (h-
1
) 0.0234 (0.979 c./h) 0.0139 (0.588 c./h) 0.0074 (0.318 c./h) 0.0050 (0.209 ~ / h
t 1/2 (h) 30 50 94 139
Value. la par.nth l. war. aot ue.d for .od.l1lna purpo
- Th. cut-off polat.
I.D
w
'rAILI C.2. DPIUIIIft& ..... ,.. ... amnii CDIIIDI A.'l 20C
20C 20C
All NO All
UV UV
S,..theUc
Cyenlde I I
801llUOIl tCh) CII
T
C .. IL) a..o t(h) CII
T
( .. IL) . ae.o
.. ell 0 189.0 0.0 0 190.0 0.0
11 61.0 67.7 11 140.0 26.3
25 33.0 82.5 U (168.0)* 11.6
4Z 17.0 91.0 42 120.0 36.8
49 11.0 94.2 114 20.0 89.5
114 6.8 96.4 121 17.4 90.8
121 1.2 99.4 138 14.9 92.2
138 0.6 99.7 145 4.1 97.5
145 0.15 99.9 162 2.8 98.5
169 2.4 98.7
116 2.0 98.9
193 1.8 99.0
210 1.2 99.4
216 0.7 99.6
306 0.28 99.9
k.,(h-
1
) 0.0569 (2.382 ~ / h 0.0205 (0.858 ca/h)
It (h-
1
) 0.0389 (1.628 ~ / h 0.0253 (1.059 calh)
t 1/2 Ch) 18 27
* 'eloe. 1. par.nth l. were DOt ... d for aod.l11aa purpo
- 1'h. cut-off pol.t.
20C
AIR
10 UV
I
t(h) CII
T
C .. IL) ino
0 200.0 0.0
6 113.7 43.1
23 56.9 71.5
29 29.1 85.1
47 11.2 94.4
53 1.4 99.3
119 1.2 99.4
126 1.0 99.5
143 0.6 99.7
167 0.1 99.9
0.0597 (2.499 calh)
0.0394 (1.649 ca/h)
18
20C
NO All
10 Uf
I
t(h) CII
T
C .. IL) inoftCI
0 196.0 0.0
6 140.3 28.4
23 110.0 43.9
29 74.7 61.9
47 54.7 72.1
53 23.5 88.0
119 9.8 95.0
126 9.1 95.4
143 5.4 97.2
150 4.9 97.5
174 3.7 98.1
197 2.3 98.2
285 1.1 99.4
0.0263 (1.101 ~ / h
0.0231 (1.003 calh)
29
\..0
J:'"
Data C.,. ID'I&:MUI'fAL IUULft .. CDIPLD Itt 4C
4C 4C 4C 4C
AIR NO AIR AIR NO All
UV UV NO UV NO UV
8yathetic
Cyanld. I I I I
Solution t(b) eN ( .. /L) lellO".d t(b) eM (III/L) lellOv.d t(b) CN ( .. /L) lellOv.d t(b)
eN ( .. /L) ".o".d
T T T T
Ba2Cu(eM). 0 195.0 0.0 0 204.0 0.0 0 203.0 0.0 0 197.3 0.0
7 . 185.6 4.8 7 137.0 32.8 19 (62.4) 69.3 17 149.2 24.4
24 172.4 11.6 24 121.9 40.2 24 148.5 26.8 24 136.9 30.6
48 103.4 46.9 31 99.1 51.4 41 135.3 33.3 41 136.0 31.1
54 87.5 55.1 48 80.9 60.3 48 126.7 37.8 48 114.6 41.9
72 68.9 64.7 72 78.2 61.6 65 100.6 50.4 65 112.5 43.0
79 67.6 65.3 144 58.6 11.2 72 11.5 64.8 72 111.3 43.6
144 36.9 81.1 168 30.2 85.2 113 55.6 72.6 137 93.1 52.8
151 37.1 81.0 175 35.8 82.4 144 66.2 67.4 161 90.9 53.9
168 31.0 84.1 192 41.1 79.8 151 59.6 70.6 168 84.7 57.1
175 29.7 84.8 199 37.8 81.5 168 49.2 75.8 185 85.9 56.5
192 (26.5) 86.4 216 33.8 83.4 175 51.8 74.5 192 74.0 62.5
216 27.6 85.9 223 34.4 83.1 192 17.8 61.7 209 70.7 64.2
240 23.7 87.8 240 34.4 83.1 199 37.6 81.5 216 68.9 65.1
312 19.7 89.9 247 31.8 84.4 216 44.3 78.2 233 70.1 64.5
288 23.9 88.3 223 35.4 82.6 256 60.3 69.4
312 22.8 86.3 241 36.8 81.9 280 58.9 70.1
319 28.5 86.0 247 34.7 82.9 296 51.5 73.9
312 34.3 83.1 317 50.2 74.5
ky (h-
I
)
0.0148 (0.62 c./b) 0.0174 (0.73 c./h) 0.0126 (0.53 c./h) 0.0073 (0.30 c-lb)
ILl (h-
1
) 0.0036 0.0031 0.0037 0.0033
t 1/2 (b) 192 220 187 210
Value. In parenthe.l. were not ue.d for IIOd.lllna purpo
- Th. cut-off point.
U)
\J"1
'1'.ltlAt C.4. DPIIIIIIIIUL 1lSUL1'S .. COPPD-<DIlDI COIIfLD AT 20C
20C 20C 20C 20C
AIR NO AIR AIR NO AU
UV UV NO UV NO UV
Synth.tlc
Cyanld. % % % %
Solution t(h) CltT(qIL) 1teaov.d t(h) CltT(qIL) . ItellOved t(h) CltT(qIL) ItellOved t(b) CNT( .. IL) "lIOnel
la2Cu(CNh 0 202.0 0.0 0 197.0 0.0 0 198.0 0.0 0 207.0 0.0
1l 106.5 47.3 18 172.0 12.7 18 94.3 52.4 18 192.5 7.0
14 73.0 6].9 25 156.0 20.8 24 39.8 79.9 24 144.4 30.2
18 46.4 77.0 42 ll5.0 41.6 42 32.3 8].7 42 124.4 39.9
25 45.0 77.7 49 85.9 56.4 48 29.0 85.3 48 99.6 51.9
42 38.0 81.2 114 47.3 76.0 ll4 12.1 93.9 114 42.7 79.4
ll4 24.0 88.1 121 ]2.4 83.5 121 9.7 95.1 121 36.6 82.3
121 15.5 92.] 138 (23.7) 88.1 1]8 9.1 95.4 138 28.8 86.1
145 14.7 92.7
ill
24.2 87.7 169 7.3 95.4 169 28.1 86.4
162 13.7 93.2 162 24.2 87.7 186 7.0 96.3
193 9.3 95.4 169 (27.8) 85.9 192 4.7 96.5
210 9.1 95.5 186 22.7 88.5 210 2.3 97.6
217 8.5 95.8 192 23.6 88.0 265 0.7 98.8
282 2.4 98.8 210 22.4 88.6 282 0.3 99.7
316 0.7 99.6 216 21.3 89.2 99.8
282 20.0 89.9
289 19.0 90.4
]06 15.1 92.]
ky (b-
l
) 0.0]92 (1.64 ca/b) 0.0146 (0.61 ca/b) 0.0402 (1.68 (ca/b) 0.0150 (0.63 ca/b)
Itl (h-
1
) 0.0153 0.0024 0.0207 0.0071
t 1/2 (b) 45 289 34 98
Value. In par.nth l. v.re not u.ed for .ad.1Iin, purpo
\.D
0'
- Th. cut-off point.
'lAILI c.,. IIPIUIIIft& IISUL'IS lOll U.c-cY.&IIID1 ClllPLD &r 4C
.C .C .C .C
AIR NO AIR AIR NO All
UV UV NO UV NO U,
S",thetic
Cyealde % % % %
Solutin t(h) CRr(q/L) IeIIonei t(h) CMr(q/L) Reaonll t(h) CMr(q/L) "110 t(b) CMr(q/L) ...... 11
laZZa(CI) .. 0 197.0 0.0 0 190.0 0.0 0 195.0 0.0 0 164.2 0.0
7 188.3 4.4 7 148.0 22.1 19 140.7 27.8 17 158.9 3.2
24 171.1 13.1 24 131.2 30.9 2. (108.2) 44.5 24 141.6 13.1
48 79.6 59.6 31 17.5 53.9 41 121.1 37.5 41 129.6 21.1
54 54.4 72.4 48 71.7 62.3 41 103.3 47.0 41 125.6 23.5
72 49.1 75.1 72 59.9 68.4 65 79.4 59.3 65 124.6 24.1
79 37.1 11.1 144 (27.6) 85.5 72 58.3 70.1 72 100.0 39.1
144 9.0 95.4 168 27.1 85.3 113 (47.7) 75.5 137 84.1 41.1
151 1.5 95.7 175 20.0 19.5 144 51.3 70.1 144 (61.2) 58.5
.161 1.2 95.1 192 16.9 91.1 151 57.0 70.1 161 72.7 55.7
175 5.6 97.2 199 15.9 91.6 161 (24.6) 17.4 161 71.6 56.4
192 3.9 91.0 216 (16.4) 91.4 175 31.1 84.0 185 69.2 57.9
216 3.8 91.0 223 13.9 92.7 192 29.3 85.0 192 69.2 57.9
240 2.9 91.5 240 11.9 93.7 199 27.0 16.1 209 66.3 59.6
312 2.6 98.7 247 7.3 96.2 216 23.2 88.1 216 50.5 69.2
218 7.3 96.2 223 22.5 18.5 233 50.1 69.5
312 7.3 96.2 241 20.4 89.5 256 44.9 72.7
319 6.4 96.6 247 18.4 90.6 280 43.5 73.5
312 12.6 93.5 296 42.1 74.4
317 33.1 79.4
ky (b-
1
) 0.01873 (0.71 ~ / h 0.01622 (0.68 ~ / b 0.01517 (0.63 ca/h) 0.00602 (0.25 C8/h)
kl (b-
I
) 0.01366 0.0095 0.00682 0.00446
t 1/2 (b) 51 73 102 155
'a1u ia par.ath l. vere aot u.ed for IIOd.lllna purpo
\.D
-....J
- 'lbe cut-off poiat.
TOLB C.6. IXPIUMKJI'UL USOLTS FOIl ZIIIC-CYAJlIDI aJfPLIX AT 20C
20C 20C 20C 20C
AIR NO AIR AIR NO AIR
UV UV NO UV NO UV
Synthetic
Cyanide % % % %
Solution t(h) CNT(mg/L) Removed t(h) CNT(mg/L) Removed t(h) CNT(mg/L) Removed t(h) CNT(mg/L) Removed
Na2Zn(CN>.. 0 196.0 0.0 0 200.0 0.0 0 185.6 0.0 0 168.4 0.0
18 68.3 65.1 18 168.0 16.0 2 139.2 25.0 2 (159.1) 5.5
25 41.0 79.1 25 126.0 37.0 5 117.2 36.8 5 160.4 4.7
42 25.0 87.2 42 98.0 51.0 7 106.1 42.8 7 145.9 13.4
49 15.7 92.0 49 62.3 68.8 24 38.4 79.3 24 115.6 31.3
114 9.7 95.0 114 16.2 91.9 27 34.1 81.6 27 107.4 36.2
121 2.4 98.8 121 15.7 92.1 30.5 29.9 83.8 30.5 (l18.0) 29.9
138 1.2 99.4 138 14.3 92.8 48 (5.3) 97.1 48 79.6 52.8
145 0.6 99.7 145 3.6 98.2 51 7.4 96.0 51 70.3 58.3
162 0.2 99.9 162 2.9 98.5 54.7 5.8 96.9 54.7 47.7 71.6
186 0.1 99.9 169 2.1 98.9 72 .8 99.6 72 45.3 73.1
186 2.0 99.0 75 .7 99.6 75 43.8 74.0
192 1.8 99.1 102.5 .6 99.7 78.5 32.9 80.5
210 1.2 99.4 96 31.6 81.3
216 1.4 99.3 102.5 26.5 84.2
281 0.6 99.7 168 10.3 93.9
289 0.1 99.9 174.5 8.2 95.1
192 5.0 97.0
198.5 3.3 98.0
216 2.1 98.7
240 1.3 99.2
k" (h-
1
) 0.0494 (2.07 em/h) 0.02101 (0.88 em/h) 0.0607 (2.54 em/h) 0.01884 (0.79 em/h)
It} (h-}) 0.0448 0.01815 0.0062 0.01944
t 1/2 (h) 15 38 III 36
Values in parenthesis were not used for modelling purposes.
- The cut-off point.
I..D
(X)
1'AlLI C.7. IIPDDIIIIDL ..... 1'1 .. RCULooCDllJDI QlM.U A1' 4C
4C 4C 4C 4C
All RO All AIR 10 AIl
UV UV IOUV NO UV
l,..th.Uc
C,aDlc1. I I Z I
Bolatloa t(h) ClT( .. IL) leIIO.ecI t(h) ClT(aalL)leIIO cI t(h) ClT( .. IL) le8o .. cI t(h) ClT(aalL) ...... d
211(CI) .. 0 200.0 0.0 0 176.0 0.0 0 207.0 0.0 0 169.5 0.0
7 . 19a.' 1.0 7 125.9 2a.5 19 (196.2) 5.2 17 153.6 '.4
24 169.7 15.1 24 112.7 36.0 24 121.a 41.2 24 140.2 17.'
48 107.4 46.3 31 66.3 62.3 41 112.0
45.'
41 9a.2 42.1
54 104.a 47.6 41 59.4 66.2 4a 100.6 51.4 4a 95.7 43.5
72 (67.6) 66.2 72 50.4 71.4 65 87.9 57.5 65 90.3 46.7
79 6a.9 65.5 144 39.5 77.5 72 a2.2 60.3
1!
88.6 47.7
144 39.2 80.4 168 37.1 78.9
ill
62.2 69.9 137 84.1 50.4
151 40.0 ao.o 175 37.0 78.9 144 62.1 70.0 144 (13.8) 56.5
168 41.6 79.2 192 36.7 79.1 151 (58.3) 71.8 161 7a.l
53.'
175 36.6 81.7 199 33.1 81 . 2 168 57.0 72.5 168 (66.8) 60.6
192 34.5 82.a 216 31.a 81.9 175 59.6 71.2 185 71.6 57.8
216 (36.a) al.6 223 28.5 83.8 192 59.2 71.4 192 69.2 59.2
240 2a.9 a5.5 240 24.5 86.0 199 57.a 72.1 209 (60.9) 64.1
'12
26.3 86.9 216 55.9 13.0 216 64.5 61.'
223 55.2 73.3 233 (64.5) 61.9
241 53.1 74.' 256 61.7 63.6
247 49.0 76.3 2ao 56.1 66.9
312 40.4 aO.5 296 54.7 67.7
317 50.0 70.5
ky (h-
l
) 0.01400 (0.59 ca/h) 0.01633 (0.6a ca/h) 0.01000 (0.42 ca/h) 0.01014 (0.42 ca/h)
Itl (h-
l
) 0.00416 0.00406 0.002978 0.00269
t 1/2 (h) 167 171 233 258
V.1u lD par.nthe.l. vera Dot uaed for aoclel11na purpo
1.0
1.0
- Th. cut-off polat.
rAIL. C.I. DPDDIIIIUL ..... n IOIlIlCIDL-CDIIIDI CCIIPLD U 20C
20C 20C 20C 20C
AIR ItO All All NO All
uv uv NO UV NO UV
Synthetic
C,anida I I I I
hlutioa tCh) CNTC .. IL) a..oncl tCh) CNTC .. IL) .Ieao d tCh) CNTC .. IL) baG .. d tCh) CNTC .. IL) -..0
2
N1
(CI) ..
0 199.0 0.0 0 198.0 0.0 0 229.4 0.0 0 198.9 0.0
18 65.9 66.9 18 158.0 20.2 2 212.2 7.5 2 159.1 20.0
25 55.0 72.4 25 140.0 29.3 5 202.9 11.5 5 144.5 27.3
42 42.0 78.9 42 106.0 46.5 7 197.6 13.9 7 132.6 33.3
!!.
30.9 84.5
.!!.
83.4 57.9 24 (196.2) 14.5 24 41.1 79.3
114 29.7 85.1 114 69.8 64.7 27 151.2 34.1 27 35.8 82.0
121 (28.5) 85.7 121 (49.8) 74.8 30.5 143.2 37.6 30 31.6 84.i
138 28.6 85.6 138 49.8) 74.8 48 116.7 49.1 48 30.0 84.9
145 28.4 85.7 145 39.0 80.3 51 110.6 51.8 51 29.4 15.2
162 28.0 85.9 162 37.5 81.1 55 104.2 54.6 54 29.2 15.3
210 26.6 86.6 169 37.0) 81.3 75 85.4 62.8 72 28.2 15.8
289 25.6 87.1 186 (37.0) ,81.3
1!:1
84.9 63.0 75 27.9 15.9
192 37.2 81.2 102.5 72.4 68.4 78 (27.9) 85.9
210 36.8 81.4 168 43.0 81.3 96 27.8 86.0
216 36.0 81.8 174.5 40.1 82.5 102 27.3 86.3
282 35.0 82.3 192 39.3 82.9 168 27.2 86.3
289 34.2 82.7 198.5 39.2 82.9 174 27.0 86.4
306 32.1 83.8 222.7 35.0 84.7 192 26.0 86.9
240 (35.0) 84.7 198 25.7 87.1
252 33.9 85.2 216 24.9 87.5
It" (h-
l
) 0.03480 (1.46 ca/h) 0.01713 (0.72 ea/h) 0.01264 (0.53 ca/h) 0.06212 (2.6 ca/h)
Itl (h-
l
) 0.00082 0.001487 0.00268 0.000814
t 1/2 (h) 845 466 259 851
'alue. in p.renth i. vera not u." for aod.llinl purpo
N
0
- Th. cut-off point. 0
'fAILI C.t. IIPI&DIft& .... 11 ... H I ClllPLD IJ.'f 4C
4C 4C 4C 4C
All IfO All AIR 10 All
UV UV 10 UV NO U,
Synthetic
Cy.nide % Z Z I
Solutloa t(h) CRT(q/L) lao ... tCh) CRTCq/L) lAIIlo.,ed t(h) CRT( .. /L) Reaoved t(b) CRT( .. /L) ......
K3'e(CI), 0 193.0 0.0 0 183.0 0.0 0 187.0 0.0 0 140.0 0.0
19 160.0 17.1 17 150.0 18.0 24 138.0 26.2 10 108.0 22.9
24 155.0 19.7
!!.
149.0 18.6 41 66.0 64.7 24 62.0 55.7
41 143.0 25.9 41 145.0 21.0 72 57.3 69.4 40 54.0 61.4
41 133.0 31.1 41 144.0 21.3 96 46.0 75.4 4i 50.0 64.3
65 120.0 37.1 72 125.0 31.7 161 34.1. 11.4 60 42.0 70.0
72 116.0 39.9 137 85.2 53.4 192 33.3 12.2 72 34.0 75.7
113 76.6 60.3 144 10.6 55.9 216 33.0 12.3 96 33.0 76.4
144 63.3 67.2 161 72.9 60.2 240 32.8 82.5 168 32.0 77.1
151 62.6 67.6 168 72.6 60.3 264 30.4 83.7 192 29.7 71.8
168 55.0 11.5 185 66.2 63.1 216 31.0 77.9
175 54.0 72.0 192 65.3 64.3 240 30.0 78.6
192 48.6 74.8 209 (58.9) 67.8 264 29.0 79.3
199 46.5 75.9 216 61.1 66.6
216 43.9 77.2 233 56.6 69.1
223 42.1 78.2 256 51.5 11.8
241 31.4 10.1 280 48.0 13.1
247 37.1 10.1 296 43.2 76.4
312 27.8 85.6 317 39.7 78.3
It., (b-
l
) 0.00612 (0.28 ca/h) 0.00915 (0.38 ca/b) 0.01535 (0.64 ca/h) 0.01527 (0.64 ca/b)
Itl (h-
l
) 0.00510 0.00399 0.001189 0.00074
t 1/2 (b) 136 174 583 937
'.lu.. In p.r.nthe.i. v.re Dot u.ed for aod.11ing purpo
N
- Th. cut-off point. 0
'fAILI C.IO. IIPIlIlIIIBIIUL IISIILft 101. ~ l l I I COIIPLU AT 20C
20C 20C 20C 20C
AIR 10 AIIl AIR 10 All
UV UV NOUV NO UV
Synthetic
Cyanide Z Z Z Z
Solution t(h) ClT( .. IL) le.oyed t(h) ClT( .. IL)
.a..o.,ed t(h) ClTC .. IL) Ieaoyed t(b) ClT( .. IL) Ieao .. d
Ilfe(CI), 0 183.0 0.0 0 183.0 0.0 0 201.0 0.0 0 196.0 0.0
18 49.8 72.8 18 130.0 29.0 7 190.0 5.5 7 184.0 6.1
25 37.9 79.3 25 94.1 48.6 24 165.0 17.9 24 150.0 23.5
42 24.8 86.4 42 69.2 62.2 48 134.0 33.3 31 130.0 33.7
49 24.7 86.5 49 62.4 65.9 54 132.0 34.3 48 97.0 50.5
114 (18.3) 90.0 114 26.2 85.7 72 85.8 57.3 72 80.9 58.7
121 18.8 89.7 121 24.5 86.6 79 84.8 57.8 144 53.4 72.8
138 16.4 91.0 138 21.0 88.5 144 45.8 77.2 168 48.5 75.3
145 t6.7 90.9 145 18.7 89.8 151 44.5 77.9 192 45.0 77.0
162 U.O 91.8 162 14.7 92.0 168 44.1 78.1 "199 42.8 78.2
169 14.4 92.1 169 13.0 92.9 175 (41.1) 79.5 216 38.9 80.1
116 12.5 93.2 186 10.2 94.4 192 42.2 79.0 223 (40.3) 79.4
193 11.6 93.7 192 8.7 95.2 216 36.5 81.8 240 38.7 80.3
210 11.1 93.9 210 6.5 96.4 240 (36.5) 81.8 247 38.1 80.6
217 10.3 94.4 216 6.0 96.7 312 35.5 82.3 288 35.4 81.9
282 6.6 96.4 282 0.9 99.5 312 33.8 82.8
289 6.1 96.7 289 0.7 99.6
306 5.5 97.0 306 0.4 99.8
ky (b-t
1
) 0.03976(1.66 c./b) 0.02263 (0.95 C8/h) 0.01042 (0.44 C8/h) 0.01306 (0.55 C8/h)
tl (h-
1
) 0.00679 0.0222 0.00163 0.00264
t 1/2 (h) 102 31 425 262
Valuea In paranthe.la vere not ua.d for .od.l1Ina purpo
- The cut-off point.
N
0
N
r ... C.ll. DPIUIIa'l'& ..... .,. ... LCIf lUX Ar 4C
4C 4C 4C 4C
AIR 10 AU AU 10 AU
UV UV IfO UV 10 Uf
SJRth.tlc
CyaDid. I I I I
lolutloll tCh) C1f
r
C .. IL) ..... tCb) C1f
r
C .. IL) a..o tCh) C1f
r
C .. IL) lao .. tCb) C1f
r
C .. IL) a..o
Low lila 0 198.0 0.0 0 185.0 0.0 0 194.0 0.0 0 214.0 0.0
7 197.' 0.3 7 133.5 27.8 17 145.0 25.2 19 211.0 1.4
24 175.0 11.6 24 35.0 81.1 24 143.0 26.3 24 207.0
3.'
31 1n.0 12.6
'1
27.3 85.2 41 140.0 27.8 41 195.0 8.9
.!!
120.5 39.1 48 12.8 93.1 48 130.0 ".0 48 193.0 9.8
72 63.0 68.2 72 9.2 95.0 65 124.0 ".1 65 (189.0) 11.7
120 35.0 82.3 120 C8.1) 95.6 72 121.0 37.6 72 172.0 19.6
144 34.8 82.4 144 8.3 95.5 137 83.8 56.8 113 148.0 30.8
168 C35.1) 82.' 168 6.3 96.6
ill.
78.1 59.7 144 152.0 29.0
175 34.3 82.7 175 6.2 96.6 161 74.8 61.4 151 144.0 32.7
192 31.0 84.3 192 5.2 97.2 168 72.0 62.9 168 127.0 40.6
199 31.' 84.2 199 5.0 97.3 185 68.6 64.6 175 118.0 44.9
216 26.6 86.6 216 4.9 97.3 192 66.8 65.6 192 92.2 56.9
240 11.1 94.4 240 4.5 97.6 209 65.6 66.2 199 93.3 56.4
247 10.7 94.6 264 3.9 97.9 216 58.5 69.8 216 84.6 60.5
288 10.5 94.7 233 59.4 69.4 223 83.7 60.9
256 48.7 74.9 241 n.2 65.8
280 46.7 75.9 247 72.6 66.1
296 41.9 78.4 312 55.1 74.2
317 38.0 80.4
... (b-
l
) 0.00974 (0.41 e./h) 0.06566 C2.75 c"h) 0.00603 (0.25 c"h) 0.00292 CO.12 e./h)
Itl Ch-
l
) 0.00868 0.00523 0.00460 0.00464
t 1/2 (h) 80 132 151 149 N
0
Value. III paranth l. vere Dot u for .ad.111o, purpo
w
- Th. eut-off poiDt.
YAIU CeiL UPIaDIUDL IISULTB rca LOll NIX AT 2O-C
20C 20C 20C 20C
AIR NO AIR AIR NO AIR
UV UV NO UV NO UV
Synthetic
Cyanide I: I: I: I:
Solution t(b) ClIr(q/L) "110 ..
t(b) CNr(q/L) IellOv.d t(h) CNr(q/L) "IIOYH t(b) CNr(q/L) .... ad
Low Nix 0 188.0 0.0 0 199.0 0.0 0 200.0 0.0 0 198.0 0.0
18 55.3 70.6 18 147.0 26.1 7 197.0 1.5 7 187.0 5.6
25 39.3 79.1 25 129.0 35.2 24 147.0 26.5 24 181.0 8.6
42 18.9 89.0 42 74.1 62.8 48 108.0 46.0 31 122.0 38.4
49 15.5 91.8 49 67.5 66.1 54 93.2 53.4 48 79.4 59.9
114 8.7 95.4 114 19.5 90.2 72 66.8 66.6 72 58.3 70.6
121 7.6 95.9 121 17.5 91.2 79 62.1 68.9 144 28.2 85.8
138 6.4 96.6 138 14.1 92.9 144 23.4 88.3 168 23.0 88.4
145 6.2 96.7 145 13.2 93.4 151 (23.8) 88.1 192 19.0 90.4
162 (5.1) 97.3 162 11.3 94.3 168 19.4 90.3 199 18.2 90.8
169 5.5 97.1 169 10.6 94.7 175 17.4 91.3 216 16.3 91.8
184 4.6 97.5 186 9.5 95.2 192 17.3 91.3 223 15.7 92.1
193 (3.9) 97.9 192 8.7 95.6 216 14.7 92.6 240 15.1 92.4
210 4.5 97.6 210 (9.5) 95.2 240 12.8 93.6 247 14.4 92.7
217 4.3 97.7 216 7.5 96.2 312 10.4 94.8 288 11.8 94.0
282 2.3 98.8 282 6.2 96.9 312 10.5 94.7
289 2.3 98.8 289 6.1 96.9
306 1.8 99.0 306 5.6 97.2
ky (b-I) 0.05039 (2.11 ~ / b 0.02336 (0.98 ca/h) 0.01532 (0.64 ~ / b 0.01788 (0.75 ~ / b
Ita (b-
I
) 0.00749 0.00599 0.00468 0.00562
t 1/2 (h) 92 116 148 123
Value. in parenthe.i. wer. not u.ed for".adellint purpo
N
0
- Th. cut-off point.
-l:'"
uaI.II Cel S. UPDIlIDUL IISOL!S rca 81GB liD at 4C
4C 4C 4C 4C
AIR NO AIR AIR NO AIR
UV UV NO UV NO UV
SJIlthetlc
Cyanide Z Z Z Z
8olutioa t(h) CNT(q/L) Iaoy. t(h) CNT(q/L) . ReIlOY. t(h) CNT(q/L) lallOY. t(b) CNT( .. /L) lallOYad
Bllh IU. 0 180.0 0.0 0 192.0 0.0 0 198.0 0.0 0 198.0 0.0
7 175.0 2.8 17 163.0 15.1 19 185.0 6.6 18 181.0 8.6
24 156.0 1l.3 24 (153.0) 20.3 24 175.0 11.6 25 175.0 11.6
31 148.0 17.8 41 165.0 14.1 41 (158.0) 20.2 42 164.0 17.2
48 135.0 25.0 48 152.0 20.8 48 160.0 19.2 49 155.0 21.7
72 115.0 36.1 65 147.0 23.4 65 150.0 24.2 114 138.0 30.3
144 93.0 48.3 72 140.0) 27.1 72 (150.0) 24.2 121 139.0 29.8
168 84.7 52.9 137 125.0 34.9 113 127.0 35.9 138 132.0 33.3
"192 81.8 54.6 144 123.0 35.9 144 (120.0) 39.4 145 133.0 32.8
199 76.8 57.3 161 119.0 38.0. 151 125.0 36.9 162 (134.0) 32.3
216 74.5 58.6 168 117.0 39.1 168 117.0 40.9 169 126.0 36.4
223 (76.3) 57.6 185 112.0 41.7 175 99.9 49.5 186 (126.0) 36.4
240 (77.7) 56.8 192 (115.0) 40.1 192 98.0 50.5 192 120.0 39.4
247 74.3 58.7 209 109.0 43.2 199 96.7 51.2 210 (120.0) 39.4
288 71.4 60.3 216 92.7 51.7 216 (94.0) 52.5 216 93.0 53.0
312 68.7 61.8 233 91.0 52.6 223 95.4 51.8 282 82.2 58.5
256 89.9 53.2 241 92.4 53.3 289 81.1 59.0
280 88.8 53.7 247 89.4 54.8 306 80.9 59.1
296 83.6 56.5 312 83.8 57.7
317 83.1 56.7
tty (la-I)
0.00628 (0.26 ca/h) 0.00246 (0.10 ca/h) 0.00391 (0.16 ca/h) 0.00231 (0.09 ca/b)
Itl (h-
I
) 0.00159 0.00112 0.00223 0.00167
t 1/2 (h) 435 619 311 415
N
0
\n
Value. in parenth i. wer. not u.ed for lIOd.llinl purpo
- The cut-off point.
TAiLI C.l'" DPIUIdIIDL USUL1'8 rca BIGIl IIIZ A"l 2O"C
20C 20C 20C 20C
AIR NO All AIR NO All
UV UV NOUV NO UV
Synthetic
Cyanide I I I I
Solution t(b) CNT(.,/L) leIIOyed t(b) CNT(.,/L) leIIOyecl t(h) CNT(.,/L) "lIOyeel t(b) CN (.,/L) a..o.
lI1.h Mh 0 193.0 0.0 0 191.0 0.0 0 199.0 0.0 0 196.0 0.0
18 92.4 52.1 18 153.0 19.9 19 92.9 53.3 19 162.0 17.3
25 87.0 54.9 25 146.0 23.6 26 81.0 59.3 26 150.0 23.5
42 70.3 63.6 42 99.5 47.9 43 (67.7) 66.0 43 99.0 49.5
49 66.0 65.8 49 97.2 49.1 SO 68.8 65.4 SO 95.6. 51.2
114 (54.5) 71.8 114 70.1 63.3 67 63.3 . 68.2 67 87.4 55.4.
121 56.0 70.9 121 70.6 63.0 74 (58.4) 70.6 74 (76.9) 60.8
138 53.5 72.3 138 66.7 65.1 91 59.3 70.2 91 77.9 60.3
-U5 53.9 72.1 145 63.2 66.9 98 54.4 72.7 98 75.1 61.7
162 (46.9) 75.7 162 (63.9) 66.5 163 48.8 75.5 163 65.7 66.5
169 50.9 73.6 169 63.0 67.0 170 48.1 75.8 170 64.1 67.3
186 48.8 74.7 186 59.0 69.1 187 45.5 77.1 187 (62.1) 68.3
193 (43.6) 77.4 192 58.2 69.5 194 45.6 77.1 194 62.1 68.3
210 45.2 76.6 210 56.9 70.2 211 43.4 78.2 211 59.5 69.6
217 43.6 77.4 216 55.4 71.0 218 42.2 78.8 218 58.7 70.0
282 36.9 BO.9 282 (52.0) 72.8 235 40.3 79.7 235 (59.5) 69.6
289 35.2 81.8 289 53.0 72.2 242 38.7 80.5 242 58.2 70.3
306 33.2 82.8 306 50.2 73.7 266 38.4 80.7 266 57.4 70.7
ky (b-
I
) 0.03366(1.41 c./h) 0.01106 (0.46 ca/h) 0.03570 (1.49 c./h) 0.01023(0.43 ca/h)
"I (b-
I
) 0.00268 0.00224 0.00264 0.00247
t 1/2 (h) 256 309 262 281
* Valuee In par.ntheele .. re Dot ueeel for . .od.llin. purpo
- The cut-off point.
N
0
(j'\
APPENDIX D
The Cut-off Point Deteraination
by Using Linear Regression
(Ex8llple: Ri-CN at 4C. NO AIR.. RO OV)
207
1
2
3
4
5
&
7
8
9
10
11
12
13
14
15
1&
11
18
lq
20
21
2Z
23
24
25
2& .
27
28
2q
30
31
32
33
34
35
36
37
3ft
39
40
.. 1
42
.. 3
44
45
46
.. 7
1t8
1t9
50
PROGIUH BACKRA r 13/171 OPT= 1 FT N 5.1+538
PROGRAM 8ACKRAT CINPur, OUTPUT t Til P E5 =INPUT tTAPE6=OU TPUT'
ItEAL "lOO.,Cl 100 cCloo"cXUOO'1 CHAT 100)1 RClOO., RSUOOt.
+ PXClOO',CHATHUOO.,CHATP(tOO'.rREV 10D),CREVC 00)
1 CONTINUE
C .................................................................... .
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
THIS PROGRAM CALCULATES RATECS' TOP TO BOTTOH
VIC,= VERSA

REAo.N
WRITEC. ,200' N
200
300
.. 00
500
IFCN,GT1IOD1GO TO 717
FORHArl
CHECK IF IT IS OK TO PROCEED
IF ((N.GT.OI .ANO.CN.LT.I00HTHEN
300.
FORHAT( TO PROCEED#'
ENO IF
REAO tCO
WIUTE ... 001
FORI1ATclx.nIHE T IN HRS CONC IN "GIL'.)
DO 10 1=1, N
CHECK IF N IS WITHIN SPFCIFIEO DIMENSION RANGE
IF (N. E 0.01 SlOP
IF C(N.LT!01.0R.CN.GT.100.)
W RITE C. 5 u 0 )
FORI1AT,Ix,.N MUST BE GREATER THAN ZERO ANt LESS THAN 100.)
ENDIF
READ IN GIVEN SETS OF DATA
REAO,TCIl,CtCll

T HOLDS INDEPENDENT VARIA9LES
C HOLDS I)EPENCENT VARIABLES
I HOLDS NUI1QER OF ElEHENIS IN AltRAY
ST IS SUH OF INOEPfNOENT V IHARES
SC IS SUM OF DEPENOENT VARIABLES
STC IS SUM OF Of T r.
'iTT IS SUM OF T
sec IS SUM of C
A IS COFFFICIENT REPRESENTING SLOPE OF THE STRAIGHT LINE
8 IS COEFFICIENT REPR[SENTING tNrERCEPT OF THE STRAIGHT LINE
N
o
00
51
5Z
53
54
55
5&
57
58
59
60
61
6Z
63
6 ..
65
66
67
6"
69
70
71
7Z
13
7ft
75
7&
77
78
79
flO
1'1
liZ
113
8 ..
85
fir,
87
86
89
90
ef1
9Z
133
9ft
95
qr,
q7
138
139
100
101
10Z
103
10"
105
106
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
600
10
650
C
ZO
C
C
C
C
C
c
C
CORR IS COQRElATlON COEFfICIENT
CHAT RfPRESfNTS PREDICTED DFPFNDENT VARlAClE
SCHAT HOlOS THE SUH Of PREOICTEO DEP. VARIABLES
R HOlOS RESIDUALS
SR HOLDS SUH Of RESIDUALS
RS HolDS SQUARE 0 RES IOUAlS
RSS HOLDS SU'1 Of SQUARED RESIDUAlS
SEE STANDARD ERPOR OF [STIHATICN

CX(U=Clnt/CO
C Ht::A l OG(Cl( n) )
600) TIU,ClCll
FORHAT(lX,F10.5,5l(,Fl0.5.)
CONXINUE
RFl G= 0
NOMOA T-= N
CONTINtJf
NTOTAl=N
If( NTOT Al.GT .15) NTOTAl=15
00 5000 N=3,NTOTAl
COMPUTE SUMMATIONS fOR lEAST SQUARS
ST=O.O
SC=O.O
STC=O.O
STT =D. 0
SCC=O.O
00 20 I =l,N
ADD THE VALUES OF T ANC C TO THE ACCUMULATING VALUES
IN SUMMATIONS
ST=STH (t)
SC=SCtC (J)
S TC=STC. U ( [)C I II )
STT=STT.(T([)Z)
sCC =SCC, (C ( lin 2.
CONTINUE
COMPUTE COEFFICIENTS ,A, AND ,a,
A l=STC- (IST.SC) IN)
AZ=STX- ((STSTt/N)
A=A 11 Z
IH=sC- I A.sTJ
B=BlIN
COMPUTE CORRELATION COEFFICIENT '::ORR'
Al=SCC- ((SC.SC) IN)
CORR=Al/(SQRTfA2tSQRTfA3)
USE THESE A q TO CALCULATE PREDICTED
N
o
U>
107
108
109
110
111
ttl
113
114
115
11&
117
118
l1q
Il0
III
ll2
123
124
125
Il&
127
Il8
llq
130
131
132
133
134
135
13&
137
138
139
140
1 .. 1
14l
1"3
1 ....
145
146
147
14"
149
150
151
152
153
15 ..
155
15&
157
10:;11
159
160
161
c
c
c
C
C
C
C
C
c
C
C
C
C
C
30
1100
HOD
ItO
IItOO
1500
5000
C
C
C
50
60


DEPENDENT
SR=O.O
SCHAT=O.O
RSS=IJ.O
DO 30 I =1.N
CHATUJ=ATClhA
II
COHPUTE THE PREDICTED ERRORS I.E .RESIDUALS
RlIJ=C C IJ-CHATC 1)
SR::SR+RCU
RS ( IJ :: R C IJ .. 2
RSS:: RSS+R5UJ
CONTINUf
COHPUTE RESIDUAL VARIANCE
RV=RSSI CN-Z,
SIGMAS= CA3-U.U)J /(N-Z)
COMPUTE STANDARD OF
SEE =SQRT fRV I
WRITE ALL CALCULATED VAlUES
WRITEC,1100J
FORHATC 11.'1 ,,)
00 itO 1=I.N
WRITEC,1200JTfIJ,CICIJ
FORt1ATI3X,FI0.5,20X,FI0.5.1
CONTINUE
WRITE C .,1400t A.fJ,CORR.RSS
FORMA T I 1,5X, t SLOPE A.: t,F 12 .q. 5X. 'IN TERC. B=,. FIZ .9. 5X,
t,FlO.5.!'iX} tRSS= "FIO.5.11
R<.iSAVlI=RSS !If
MRITE( 15 ODIN. RSS AVG
FeRHAT( 1,3X, tAVERAGE RSS FORt ,1l.tPOtNTS=t .FI!. 5.11.' .....
....................................................... ,)
CONUllfUF

DO REGRESSION ON OUR DATA BACKWARDS fROM LAST TO FIRST

N::NU"4DA T
00 50 1=1. N
TREVn) =TCN-hU
CREvnt=C(N-l+IJ
00 &0 I=I.N
TCIJ=TREVCIJ
C C lJ::CR[VC It
RFlAG= RFLAG+l
IF(PFLAG
1
En!1ITHEN
WRITE( 70uJ
N
o
162
163
161t
165
t 6fi
t,.,7
168
1700 FOPH4TC 11I,3)(,,9ACKWAPDS PEGRESSION"II/)
F'NDIF
tF(RFLAG.EQ.ltGO TO 650
GO TO t
N: 16
01( TO
T I"E T Itf
0.0000)
11.00000
Z".OOOOO
Itt .000!l1l
1t".00000
65.00001
72.00000
137.000011
161.000o:J
185.000111
lQZ.OOOOD
Z16.00000
256.00000
180.00000
Z9r,.0001l1l
317.80010
777 CONTINUE
STOP
END
CN CoNt IN HG/L
169.50000
153.60000
1,.0.20000
98.Z0000
95.70000
90.30000
88.60000
81t. tOOOO
78.10000
11.60000
69. ZOOOO
&,..50000
61.70000
56.10000
5,..70000
50.00000
.............................
0.0110011 169.50000
153.6000n
lItO.ZOOOO
17.00000
2,..01000
SLOP:: _= -.00751,.532 INTERC .Il: .006603923
AVERAC:: ItSS FOR 3pOINTS: .000 Z7



1.0011011
17.00000
1&9.50001)
153.60001)
-.OJ 7716 RSS: .00081
N
21t.00000
Itl.OOIOO
SLOPE -= -.013305600
140.20000
911.20000
INTERC.B= .064232109
RSS FOR 4POINTS= .00415
COQR=

.............................
0.00000
17 .00000
24.00000
Itt.OODOO
ItS.OOOOO
SLOPE _= -.013261664
1&9.50000
153.&000(J
14 O. 200 (l0
H.20000
95.70000
.06370219"
AVERAGE FOR 5POINTS: .00332
CORR=

.............................
0.001)00
17.00000
24.00000
41.00000
1t".00000
&5.0011 0 0
SLOPE h -.0112&0590
1&9.50000
153.&0000
140.20000
91\.20000
95.70000
90.30000
INTERC.8:: .026122353
.VERAGE RSS FOR &POINTS= .004119
CQRR=

............................................................
-.94988
-.9&9.19 RSS=
-. 'H) 0 2& RSS=
.01&59
.01&59
o2<J 33
N
N

..

..


..
..
..
..

..
..
..

..

e ~ c e c c c
c=occcc
Cc=cccc:t
Cc==OCc
II\\lINN"'-""\lI
.......
O"""COOI/'IOIl)
\lI1I\oI'O"crcroo
.........
cece=ae
.OcoOOOCOO
aoeCCCCD
.. ceClcccc
.====000

a "'- .... 0011\ N
.. .c', 01' 01' '" "'-

..

..
"
VI
VI
~
II
at
~
o
(.)
til
til
...
til
'"
...
I':
o
C
II
CD

U
~
W
~
Z
...
N
11\
N
'"
""
...
a
...
co

t
It
..
Ie
o
....
VI
"
VI
~
Z
...
o
A-
~
~
o
~

..
..
..
..
..
..
..
..

..
..
..



..


..
..
..

..

..
..


..
..


..

..

..

..


..
..
..
..
..
..
..
..
..
..
..


..
..
..
..


..
..
..
".
N
N
"'-
...

II
VI
VI
~
C
\D
"'-
N
IC

t
II
&'
IY
o
(.)
""
""
...
...
N
...
IC
".
COCOcCCCcoO ...
ceocooee
COCOOOC C
CCCCCCet C
"'\lINN",-""",... "
CD
cr""COOLl\CtII.8
..oLl\oI'O"crCf'CIC U
......... at
W
~
Z
...
...
.8
"'-
.. 11\
.8
01'
11\
co .. co
.. .
.. t

.CDCCCCDD II
.CCCCCCCC C
.OetOOCtooet
.. a c 0 aCtO C = hi
: ~ ~ ~ ~ ~ ~ ~ ~ ~
.c",-.8"'ICII\N .......
..... NoI''''..o ... ", VI
...
..
..
'" 11\
...
N
o
"

Z
...
o
A-
IC
VI
VI
III!
'"
L!I
C
D:
W
>
C
..
..

..


..


..
..
..
..

..
..
..
..
..
..


..
..
..

..


..
..
..
..
..


:
..
..




..
..





..
..





..


..
:


..







..
..
..
ceoe
00==
eeteto
-==Cet
",..oC\lC\l
. ...
Cf'""Coo
..oLl\"'Cf'
... ......

.ecce
.COCC
.=cec
.... CDC
.... acc

..
..
:



a ........
... N".
213
,
48.00001)
65.00000
12.00000
131.000011
161.00000
Cl5.70000
ClO.30000
88.60000
IIIt.l0000
78.10000
SLOPE h INTERC.8= -.186001461
aVERAGE RSS FOR .02246
CORR:


O.OOOQIJ
11.00000
21t.00000
"1.01l1J00
"8.00000
65.00000
12.00000
137.00000
161.001100
185.00000
SLOPE 11= -.0039880 78
1&9.50000
153.60000
litO .200 DO
CI".20o 00
Q5.70000
90.30000
"".60000
84.10000
78.10000
11.60000
INTERC.8:
aVERAGE FORIOPOINTS: .02131
CORR=

.......................................................
.............................
0.00000
11.0000]
21t.oOOOO
Itl.OOO DO
"8.00000
65.00000
16CJ.5 00 00
153.60000
IltO.20000
98.20000
95.7000n
90.30001]
-.1131 CJ8 RSS=
-.85661 RSS=
.20212
.21306
N
J;:-
72.00000
131.00000
161.00000
1115.00000
192.00000
SLOPE - = -.0 O.J 83 oct 71t
111.60000
81t.l0000
18.1000!!
11.&00011
&Q.20000
INTERC.g=
AVERAGE IlSS FORI1POIHTS= .01970
CORR=

. ..................................................... .
.............................
0.00000
t1.00000
21t.00000
1t1.00000
"'.00000
6S.00000
12.0001)0
137.00000
161.00000
185.01)000
192.00000
216.00000
_= -.003699990
I&Q.50000
153.60001'1
IltO.20000
98.20000
95.700011
90.30000
88.600011
8".10000
711.10000
71.60000
&Q .200 00
61t.50000
INTERC.B= -.2165ct3397
AVERAGE ttss FORIZPOINTS= .01835


.............................
0.00000
17.001100
21t.00000
Itl.00000
"8.00000
1&9.50000
153.60000
IltO.20000
98.20000
95.70000
-.87138 RSS=
-.8952& RSS=
.21667
.22020
N
\J1
65.000011
72.00000
137.00000
161.000011
181.;.001100
192.00000
216.00000
256.00000
SLOPE h -.003 ...... 885
90.30000
111'.60000
"".10000
78.10001)
71.60000
6 9.Z00 00
6".50000
61.70000
INTERC.8: -.23251.;,.613
lVERAGE R5S FOR13POINTS= .01816
caRR:

. ............... _ .................................... .
.............................
0.00000
17.011000
2".00000
"1.000n
..... 00000
61.;.00000
72.001100
137.00000
161.00000
18r;.00000
192.0011011
216.00000
21.;6.00000
280.001100
SLOPE A. -.00331922 ..
169.1.;0000
153.60000
lIto.zonoo
98.200 Of)
95.70000
90.30000
811.600011
81t.l0000
711.10000
71.(0001)
&9.20001)
6".50000
61.70000
1.;6.10000
INTERC.B: -.2 .. 1230645
lVERAGE ~ FOR1"POINTS= .01729
CORR:

.......................................................
-.90351 RSS,.
-.<J1671 RSS=
.23612
.2 .. 205
N
'"

1.00000
17 .00000
2".noono
\1.00000
1t&.OOOOD
65.00000
7Z.00000
137.00000
161.00000
185.00000
19Z.00000
216.00000
256.00000
280.00000
296.00000
SLOPE A= -.0032H217
16Q.500 00
153.60000
140.20000
98.Z0001)
QIj.700o0
90.30000
811.(01)00
84. too 00
711.10000
71.60000
69.Z0noo
64.500no
61.700011
56.10000
5".70000
-.2\&97Q504
AVERAGE IfSS FOR15POINTS= .01657


REGRESSION

317.00000
291).000011
ZID.OOOOO
SLOPE 4= -.003170"66
169.50000
153.60000
1 .. 0.Z0000
[NTERC.IJ:: -.20&170105
AVEfUGE R55 FOR 3POINTS= .00013
CORR=
-.92579 RSS=
-.'HZ .... RSS=
.Z"& .. 9
DO 0 itO
/
N
-....J
317.01D00
lCJ6.00ll00
280.00000
256.00000
SLOPE "= -.0033Z4CJ36
I&CJ.500 00
153.&0000
140.Z0000
98.20000
-.161CJ38551
AVERAGE RSS FOR 'POINTS= .00011
CORR=


317.00000
296.00000
280.00000
256.00000
216.0DGOO
SLOP! _= -.00Z514362
169.50000
153.60000
140.20000
98.20000
95.70000
INTERC.B= -.400437877
AVE'UGE FOR 5POINTS= .000 .. 7



317.00000
296.00000
280.00000
256.00000
216.00000
192.00000
SLOPE -= -.00Z .... 0198
169.5000'1
153.60000
1"0.200'10
98.20000
9'i.70000
90.30000
INTERC. B= -. 4Z 1792819
aVERAGE RSS FOR 6PoINTS= .00041
COQR=
-.'lCJO'i2 RSS=
-.91037 RSS=
-.98266 RSS=
.00042
.00237
.OOZ ....
N
00

117.00000
ZQ6.00000
ZSO.OOOOO
256.0010'
216.00000
In.ooooo
185.00000
SLOPE A:
16Q.500 00
153.60000
IltO.ZOOOO
98.2001)1)
'l5.10001)
90.300111)
88.60000
INTERC.O: - ... 0 895"Z17
aVERAGE Rss FOR 7PoINTS= .00036
CORR=

. ..................................................... .

117 .00000
Z'l6.01J1J01J
Z80.00000
256.00000
216.00000
lQ2.0001J0
1115.00000
161.011000
A= -.0025Q'l389
16'l.50000
l'il .&0000
140.20000
H.20000
'lC;.10000
QO.30000
88.60001)
8".10000
INTERC.O= -.377515'J55
AVERAGE Rss FOR 8PoINTS= .00041
CQRR=


117.00100
296.00000
280.00000
256.0000fl
216.00000
16'J.50001)
153.60000
1"11.20000
'l8.20000
'lS.10000
-.'J8769 RSS=
-. 'lll913 RSS=
.00252
.00325
N
I.D
ecoc
eeoc
c:tcce
eeee

. . . .

!f'Cc",
ct=c:te
=e==
ccce
=0=_
000=

...
cr-c"'..,
.. ..
..,
a-
..,
o
o
"
III
III
at
"
at
at
o
f..l
a
c
...
e
a-
a
..,
'"
..,

..

a

cr-
:e
a
a


..
..
L&J
Q..
o
-
III
"
III
...
Z
...
o
Q..
cr-
at
o

III
III
at






..

..












..
..
..




..
..





..

:






..



..



..




..
..



:












..
t=cccccecCtc::t
ccecceoccc
cc:tOOCtCccc:tet
=ocoeceaeea
"'''''''I\j ... ..,''' ..
. . . . . . . ...




.ClC3cecac:tc:teo
.=oeeaeoaao
.eecc:tClc::tc:tcCte:t
.Cccc=OCClCta

... "'a"'''''''''' ..... 1\j
.. cr-.,"'.., ...
........


..
I\j
'" cr-
a
o
"
III
III
at
'"
I\j
II'
.,
cr-

I
"
at
at
o
f..l
...
..
'"
'"
I\j
..
'"
o
a

I
"
C

o
-
III
'" a-
=
a
a
"
III
...
Z
...
o
Q..
a
..


III
III
at
L&J
I.lI
..
at
\I,j
>
c


..



..

..

..

..

..



..
..

..
..
..








..

..
:




:
















..



..
..


..
..






ecceec:C"o
cccOCtOCO
Ctoo=o=oo
CCCCOQClCJ
"''''I\jl\j ...
. ...... .




.ccCtccetcc:t
.COOoeOClc
.COCeocce
.cceeoetc:tc
.=0:0=00=00

.... "'0"'''''''''' ..

....



220
131.000OD
72.00000
65.00008
SLOP; A= -.002330363
78.10001)
71.&0000
&9.20000
INTERC. FJ= -.442986127
AVERAGE qSS FORIIPOINTS= .00103
C O R ~

. ............... _ .................................... .

317.00000
29&.00000
280.00000
256.00001
216.001100
lQ2.00000
185.1)0000
161.00000
131.000oa
12.00000
65.00800
1t8.00000
SLOPE A= -.1102305122
1&9.50000
153.&0000
IltO.200 00
98.2001)0
95.10000
90.30000
88.&0000
84.10000
78.10000
71.60000
&9.20001)
&4.50000
INTERC.8= -.41t"954360
aVERAGE ~ S S FORI2POINTS= .00097
CORR=


311.00000
296.00000
280.00000
256.00000
21&.00000
192.00000
185.00000
16Q.500 00
153.60000
IltO.20000
9ft.20000
Q<j.70000
QO.3001)1)
"".60001)
-.98595 RSS=
-.98856 RSS=
.01138
.01158
N
N
161.00800
137.00001
72.0010]
65.0000a
Its.ooIDa
"1.00000
SLOPE A= -.002302938
84.10000
7".10000
71.60000
69.20001)
6".50000
61.700 00
INTERC.A: - ..... "621037
RSS fORI3POINTS= .00089



317.00000

280.00000
256.00000
216.00000
192.00000
155.00iJOIJ
161.00000
137.0DIJOO
12.00000
65.00000
1t8.00000
"1.00000
21t.OOIDO
SLOPE A= -.002631 .. 73
16Q.50000
153.60000
140.20000
98.20000
95.70000
90.30000
88.60000
8".10000
78.10000
71.60000
6).20000
61t.50000
61.70000
56.10000
tNrERC.9= -.31337lt311
AVE RAG:: RSS fORlltPOINrs= .00638


.............................
311.00000
Z96.000 011
16 'J.500 01)
I1j3.60000
-. ))046 RSS=
-. 951j Olt RSS=
.01159
.08)3"
N
N
N
2110.00000
256.00000
216.00000
tlJ2.00000
185.0113011
161.00000
131.00000
12.00000
(1).00001
1t1l.00000
Itl.00000
ZIt.OOOOO
11.00000
lltO.20000
98.20001)
90;.100 DB
90.30000
88.60000
8".1001)0
18.100110
11.60000
69.200011
6,..50000
61.10000
'i6.10000
5".70000
SLCPE _= -.002915535 INTERC.9: -.3083183Z6
AVERAGE RSS FORI5POINTS= .011 .. 8
CORR=

N=500

-.9 .. 003 RSS= .17226
/
N
N
W
224
Comparison of the Average Residual Sum of Squares
Forward Backward
(from 0 to 317 h) (from 317 to 0 h)
For first 3 points 0.00027 0.01148
0.00415 0.00638
0.00332 0.00089
0.00489 0.00097
The cut-off point 0.00570 0.00103
0.02153 0.00095
0.02246 0.00044
0.02131 0.00041
0.01970 0.00036
0.01835 0.00041
0.01816 0.00042
0.01729 0.00011
0.01657 0.00013 For first 3 points
APPENDIX E
Calculation of the Reaction Rate
Constant by Using Linear Regression
225
PROGRA/'I RATE(INPUT.OUTPUT.TAPE5:INPUT.TAPE6:OUTPUT)
REAL T(1ee).C1(lee).C(lee),CK(1ee). CHAT(1ee). R(1Be). RSC1ee).
C .*.**.*.**.**** **********.*.**************** *.* **.*******
C
C
C
THIS PROGRA/'I CALCULATES RATEcS)
C
*.** ************.********************************.*****.**********
READ N
C
C'
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
1.e
2ee
3ee
.. ee
see
6ee
1.
WRITE(*.1.e)N
IF(N.GT.lee)GO TO 777
FOR/'IAT(1X.N=.13)
CHECK IF IT IS OK TO PROCEED
IF (CN.GT.e).AND.(N.LT.l.eTHEN
WRITE(*.2ee)
FOR/'IAT (lX. "OK TO PROCEED')
ENDIF
READ*.C.
WRITE(*,3e.)TESTT
FOR/'IATelX."TESTT:".Fl 5.)
WRITE(* ... ee)
FORMAH1X,"CN CONC IN "G .... L".SX."CX:Cl .... Ce".5)(."NAT.LOG.OF CK".)
DO 1. l=l.N
CHECK Ir N IS WITHIN SPECIrlED DI"ENSION RANGE
IF eN.EQ ) STOP
IF eeN.LT ).OR.(N.GT.l THEN
WRITEe*,se.)
FORMAT(1x.N /'lUST BE GREATER THAN ZERO AND LESS THAN 1 ')
ENDIF
READ IN GIVEN SETS or DATA
READ Tel).Cl(lt
**.************************************************************.
T HOLDS INDEPENDENT VARIABLES
C HOLDS DEPENDENT VARIABLES
I HOLDS NUMBER OF ELEMENTS IN ARRAY
ST IS SUM or INDEPENDENT VARIABLES
SC IS SUM or DEPENDENT VARIABLES
STC IS SUM or PRODUCT OF T AND C
STT IS SUM or SQUARED T
SCC IS SUM or SQUARED C
A IS COErrICIENT REPRESENTING SLOPE or THE STRAIGHT LINE
B IS COErrICIENT REPRESENTING INTERCEPT or THE STRAIGHT LINE
CORR IS CORRELATION COErrlCIENT
CHAT REPRESENTS PREDICTED DEPENDENT VARIABLE
SCHAT HOLDS THE SUM or PREDICTED DEP. VARIABLES
R HOLDS RESIDUALS
SR HOLDS SUM or RESIDUALS
RS HOLDS SQUARED RESIDUALS
RSS HOLDS SUM or SQUARED RESIDUALS
SEE REPRESENTS STANDARD ERROR or ESTIMATION
******************** ****************************************
CX(I):Cl(I) .... C.
C ( I) : ALOG ( CX ( I ) )
WRITE(*.688) Clel).cxel).C(I)
FORMATC1x.r18.5.sx.r18.5.5x.rl 5.)
CONTINUE L
226
C
C
C
C
C 7BB
C +
C
C
C
C
ST:8.8
SC:8.8
STC:8.8
STT:8.8
SCC:8.8
WRITE (*, 7BB)
f"ORI'IAT(" SUI'I T",9X,"SUI'I C",9X,"SUI'I TC",9X,"SUI'I TT",9X,
SUI'I CC",)
DO 28 1:1. N
ADD THE UALUES Of" T ANC C TO THE ACCUI'IULATING UALUES
IN SUI'II'1ATIONS
ST:ST+TC I)
SC:SC+C(I)
STC:STC+(T(I)*C(I
STT:STT+(T(I)**2)
SCC:SCC+(C(I)**2)
2B CONTINUE
C
C COI'IPUTE COEf"F'ICIENTS "A" AND "B"
C
C
C
C
C
C
C
C
9BB
18BB
C
C
C
C
C
aB
Al:STC-CCST*SC)/N)
A2:STT-CCST*ST)/N)
A:Al/A2 "
Bl:SC-CA*ST)
B:Bl/N
COI'IPUTE CORRELATION COEf"F"ICIENT "CORR"
A3:SCC-CCSC*SC)/N)"
CORR:Al/CSQRT(A2)*SQRT(A3
USE THESE A AND B COEf"f"ICIENTS TO CALCULATE PREDICTED
DEPENDENT UARIABLE "CHAT"
WRITE (*, 9BB)
f"ORI'IAT(" CHAT",12X,"SCHAT",)
SR:8.B
SCHAT:B.B
RSS:B.B
DO 38 I:l,N
CHAT C I) :A*T (I)+B
SCHAT:SCHAT+CHATCI)
WRITE(*,lBBB)CHAT(I),SCHAT

COI'IPUTE THE PREDICTED ERRORS I.E.RESIDUALS
R(I):C(I)-CHATCI)
SR:SR+RCI)
RS( I ):R( 1)**2
RSS: RSS+RSCI)
CONTINUE
COI'IPUTE RESIDUAL VARIANCE
227
1I!l5B
C
C
C
C
C
C
UBB
+
12BB
eB
13BB
+
leBB
+
RV=RSS/(N-2)

WRITE(.,lB5B)RV,SIGMAS

SEE=SCilRTCRV)
WRITE ALL CALCULATED VALUES
WRITE(.,118B)
TIME T IN HOURS .. ,12)(, "CONCENTRATION C .. ,
3)(,"RESIDUALS R=C-(;HAT",9X,"CHATM",B)(,"CHATP",)
DO e8 I:1-N
WRITE(.,128B)TCI),CCI),RCI),CHAT"CI),CHATPCI)

CALL PLOTPT CT(I).C(I),2)
CALL PLOTPT(T(I),CHAT(I).28)
CONTINUE
WRITEC.,13BB)ST,SC,SR
rORMATCC)(,"SUM OF T:",r18.5.C)(,"SUM OF C:".FIB.5.C)(,"SUM OF
R:',r18.S/)
WRITE (.,lCBB) A,B,CORR,RSS
A=",FIZ.9,S)(,"INTERC.B:",FIZ.9,5X,"CORR=
",F18.5,S)(, "RSS=
CALL OUTPLT
STOP
END
228
APPENDIX F
Linear Regression and 95% Confidence Interval
Calculations Perfot1led on the Part of the
Degradation CUrve Attributed to Volatilization
(BaCN and CU-CN. Zn-CN. Nt-CN. Fe-CN Collplexes)
229
PROGRA" VOLCO., 73;'171 OPT=1 F"TN 5. 1+538
1
C
C
C
C
C
C
C
C
C
PROGRA" VOLCONF"(INPUT.OUTPUT.TAPE5=INPUT.TAPE6=OUTPUT)
REAL T(I ).Cl(I ).C(I ).CX(I ). CHAT(I ). R(I ). RSel ).
r PX(I ).CHAT"Cl ).CHATPC1 )
CONTINUE

THIS PROGRA" CALCULATES RATECS) PLOTS THE
GRAPH AND GIVES 95. CONF". INTERVAL

WRITEC 1 )
1.. F"OR"AT("1 INPUT NU"8ER OF" DATA POINTS")
READ N
WRITEC.,2 .. )N
IF"CN.GT.l JGO TO 777
2..
3 ..
CHECK IF" IT IS OK TO PROCEED
IF" N.GT ).AND.(N.LT.I88THEN
WRITEC.,3 )
F"OR"ATUX. "OK TO PROCEED")
ENDIF"
READ.,C.
READ.,TESTT
35.
WRITEC 35.JTESTT
F"OR"AT(IX."TESTT=".F"I'.5.)
C
C
C
..
WRITE( ..... )
... CN CONC IN "G;'L ... 5X."CX=Cl;'C .... SX ... NAT.LOG.0F" CX".)
DO 1. 1:1.N
CHECK IF" N II WITHIN SPECIF"IED RANGE
IF" (N.EG.') STOP
IF" N.LT.'J.OR.(N.GT.l" THEN
WRITE( S"J
N
W
o
c
c
c
c
c
c
c
C
c
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
s.
a.
I.
KUST IE GREATER THAN ZERO AND LESS THAN 188')
END IF" '
READ IN GIVEN SETS DATA
READ.,TCI),CICI)

T HOLDS INDEPENDENT VARIABLES
C HOLDS DEPENDENT VARIABLES
I HOLDS NU"BER OF" ELE"ENTS IN ARRAY
ST IS SU" OF" INDEPENDENT VARIABLES
SC IS SUM OF" DEPENDENT VARIABLES
STC IS SUM OF" PRODUCT OF" T AND C
9TT IS SlI" OF" SQUARED T
SCC IS SUM OF" SQUARED C
A IS COEF"F"ICIENT REPRESENTING 9LOPE OF" THE STRAIGHT LINE
B IS COEF"F"ICIENT REPRESENTING INTERCEPT OF" THE STRAIGHT LINE
CORR IS CORRELATION corF"F"ICIENT
CHAT REPRESENTS PREDICTED DEPENDENT VARIABLE
SCHAT HOLDS THE SU" OF" PREDICTED DEP. VARIABLES
R HOLDS RESIDUALS
SR HOLDS SUM OF" RESIDUALS
RS HOLDS SQUARED RESIDUALS
SRS HOLDS SU" OF" SQUARED RESIDUALS
SEE REPRESENTS STANDARD ERROR OF" ESTI"ATION

hCl (I);'C.
cc
WRITEc.,a )
5,)
CONTINUE
CO"PUTE F"OR LEAST SQUARS
ST:
SC:
8TC:
N
W
C
C
C
C
C
C
C
C
STT:
SCC:
WRITE(.,?.)
'1" FOR"AT(" SU" T",9X,"SUH C",9X, "SU" TC",9X, "SU" TT",9X,
+
Z.
"SU" CC",)
DO Z. I=I,N
ADD THE VALUES ()f" T ANC C TO THE ACCUHULATINO VALUES
IN SU""ATIONS
ST:ST+T( I)
SC:SC+C(I)
STC:STC+(T(I).C(I))
STT:STT+(T(I) Z)
SCC:SCC+(C(I) Z)
CONTINUE
CO"PUTE COEF'F'ICIENTS "A" AND .....



.1:SC-(A.Sn
.=.UN
C CO"PUTE CORRELATION COEF'nCIENT "CORR"
C
C
C
C
C


USE THESE A AND COEFFICIENTS TO CALCULATE PREDICTED
DEPENDENTVARIA.LE "CHAT"
NRITE(.,9 .. )
9.. F'ORMT(" CHAT' ,12X, "SCHAT",)
SR=B
SCHAT:
SRS= II
DO 311 I:l,N
CHAT(I):A*T(I)+8
SCHAT:SCHAT+CHAT(I)
NRITE(.,lI1B.)CHAT(I),SCHAT
N
W
N
C
C
C
C
C
C
ta.
3a
ta5a
C
C
C
C

CO"PUTE THE PREDICTED ERRORS I'.E.RESIDUALS
RCI):CCI)-CHATCI)
SR:SR+RCI)
RSC U :RC 1).*2
SRS: SRS+RSCI)
COHTlHUE
CO"PUTE RESIDUAL VARIANCE
RV:SRS'CH-2)
SIG"AS:(A3-CA*At'CH-2)
NRITEC*.laSa)RV.SIG"AS

CO"PUTE STAHDARD ERROR or
SEE:SQRTCRU)
C CO"PUTE 95. COHr.LI"ITS rOR A
C
C
AA2:SQRTCA2)
ALPHA":A-CCTESTT*SEE)'AA2)
ALPHAP:A+CCTESTT*SEE)'AA2)
C CO"PUTE 95. COHr.LI"ITS rOR 8
C
C
AH2:H*A2

8ETAP:8+CCTESTT*SEE*SQRTCSTT'SQRTCAH2
C CO"PUTE 95. COHr.LI"ITS rOR CHAT
C
REVH:"H
T"EAH:ST'H
C"EAH:SC'H
DO 35 l:l.H
P)(CI):(TCI)-T"EAH)**2

N
\N
\N
3S
C
C
C
U_
+
12 ..
...
1311
+
14 ..
+
ISII

+

CONTINUE
WRITE ALL
WRITEC UII)
F'OR"ATC" TI"E T IN HOURS ".SX."CONCENTRATION C ".
S)(. "RESIDUALS R=C-CHAT". S)(. UI)C. "CHATP". )
DO 41 I=I,N
WRITEC., 1211)TC I). CC I). RU). .CHATPU)
F'OR"ATCS)(,F'II.S,21)(,F'11.S,lS)(,F'II.S.S)(,F'11.S.SX,F'II.S.)
CALL PLOTPT CTCI),CHATCI).2)
CALL PLOTPTCTCI).CHAT"CI).21)
CALL PLOTPTCTCI).CHATPCI).21)
CONTINUE
WRITEC.,I38I)ST.SC.SR
F'ORI'IATC"X,"SU" OF' T:".F'tl.S ... X,"SU" OF' C:",F'11.S.4)(,"SU" OF'
R:",F"ll.S/)
WRITE C*.I"' )
F'OR"ATC/.S)(."SLOPE =".F'12.',SX."CORR:
",F'II.S.S)(. "SRS: ",F'II.S,/)
WRITEC 1SII)ALPHA".ALPHAP,BETAI'I.8ETAP,T"EAH.C"EAN
F'ORMTC3X. "ALPHA":",F'12.9,S)(, "ALPHAP=".
F't2.9,SX."8ETA":".F'12.9,S)(."BETAP:",F'12.9,S)(
"T"EAN:". F'S. 2. 5)(. "C"EAH=",F"S.2,/)
GO TO I
.,.,., CONTINUE
CALL OUTPLT
STOP
END
N
W
.s:-
APPENDIX G.l
The Coaputer Prograa for the
Analysis of 2
3
Factorial Design Data -
Calculation of the Effects by USing
Matrix of Independent Variables and Equation 6.1
235
PROGRAM FACT
.***********************************************************************
* ANALYSIS ()F fACTORIAL DESIGN DATA

NOMENCLATURE:
NF = NUMBER OF FACTORS
NT = NUMBER OF TESTS
TYPE = "FULL" FOR A FULL FACTORIAL DESIGN
: "FRAC" FOR A FRACTIONAL DESIGN
(TYPE IS A CHARACTER VARIABLE)
TEST : ARRAY OF TEST CODES
* (TEST IS A CHARACTER VARIABLE)
R : ARRAY OF TEST RESULTS






*
*

*
*
* *
INPUT: *
* INPUT ALL DATA USING A FILE CALLED "FDAT" *
IN THE fOLLOWING ORDER: *
NF. NT. TYPE. TEST(I). R(I), TEST(2). R(2) TEST(NT), R(NT) *
* E.G. 3 a 'FULL' 'I' .832 'A' .854 '8' .71 ETC *
OUTPUT: *
* INPUT DATA "TYPE", "NF", "TESTCNT>" AND "R(NT>" ARE REPEATED *
* FOR CONFIRMATION. EFFECTS ARE LISTED IN THE FOLLOWING ORDER: *
* MAIN. 2-FACTOR. 3-FACTOR. THE 3-FACTOR ARE *
CALCULATED ONLY FULL FACTORIALS (TEMPORARY LIMITATION). *
* LIMITATIONS: THIS PROGRAM WILL ACCEPT TWO-LEVEL FULL OR FRACTIONAL
FACTORIAL DESIGN DATA IN ANY ORDER. THE "AXI"UM
SIZE IS 2**& (a FACTORS). *
* NOTE: THIS IS A FORTRAN-5 PROGRAM *
* *
************************************************************************
DIMENSION R(256)




CHARACTER TYPE*4.TEST(2S6).&.EC3)*I.EFF1(a)*1
READ (le.*) NF.NT,TYPE. (TEST(I).RCI).I:l.NT)
WRITE (3e. S8e)
IF (TYPE.EQ.FULL') GO TO 168
WRITE C3e.68.) NF
GO TO 17.
160 WRITE (30.7.0) NF
170 WRITE (30.&00) (TESTCI).R(I).I=I.NT)
WRITE Ci0.9.0)
DEFINE EFFECT LABELS
EFF1(1) : 'A'
EFF1(2) : 'B'
EFFU3) : 'C'
EFFt (4) : 'D"
EF"F"US) : 'E'
EFFt (6) : 'F'
EFFt (7) : 'G'
EFF1<a) : 'H'
DETERMINE "AIN EFFECTS

DO 58 K=I,NF
SUM:e.
DO .8 1=I,NT
L=t
236

Ie IF"(TEST(I)CL:L).EQ.EF"F"I(K GO TO 2e
IF"CL.EQ.B) GO TO 38
L:L+l
GO TO Ie
2e SU" : SU" + R(I)
GO TO 48
38 SU" = SU" - R(I)
48 CONTINUE
END = SU"/CNT/2)
58 WRITE (38,1888) EF"F"ICK).END
DETER"INE TWO-F"ACTOR INTERACTIONS


NF"l = NF"-1
DO B8 N=I,NF"1
Nl = N+l
DO B8 " : Nl,NF"
E (1) : EF"F" 1 C N )
E(2) = EF"F"IC")
SU" : 8.
DO 78 I:l.NT
KT = 8
DO 63 KL:l.2
L = 1
68 IF"(TEST( I)CL:L).EQ.ECKL GO TO 61
IF"(L.EQ.B) GO TO 62
L:L+l
GO TO 68
61 KT : KT+l
GO TO 63
62 KT : KT-l
63 CONTINUE
IF"CKT.EQ.8) SU" : SU" - R(I)
IF"CKT.NE.8) SU" : SU" + R(I)
78 CONTINUE
END:SU"/CNT/2)
Be WRITE (38,2888) E(1).EC2).END
DETER"INE THREE-F"ACTOR INTERACTIONS

IF" CTYPE.EQ.'F"RAC) GO TO 9B
-NF"l : NF"-1
NF"2 : NF"-2
DO 96 N:l,NF"Z
Nl:N+l
DO 96 ":Nl.NrI
"1:"+1
DO 96 "N:"I.NF"
E(1) = ErrICN)
E(2) : EF"F"IC")
E(3) : EF"rU"N)
SU" : 8.
DO 95 I=I,NT
KT : 8
DO 94 I.1K:l.3
L = 1
89 IF"CTESTCI)CL:L).EQ.ECI.1K GO TO 91
JF"CL.EQ.5) GO TO 93
L :L+l
GO TO 89
91 KT:KT+l
GO TO 94
93 KT : KT-l
94 CONTINUE
IrCKT.E$.3) SU"
237

I;CKT.EQ.-l) SU" : S ~ + R c I
IFCKT.EQ.-3) SU" : SUM-ReI)
I;eKT.EQ.l) SU" : SUM-RCI)
9S CONTINUE
END=SUM/CNT/2)
96 WRITE(38.3888) E(1).E(2).E(3).END
98 CONTINUE
FOR"AT STATE"ENTS





S8e FOR"ATU//dex. "ANALYSIS OF FACTORIAL DESIGN DATA.///
+.SX."PRO.7ECT NUMBER 81 .... 18X.
+"EXPERII'IENTAL SOLUTION NACN ...... //.SX.
+"RESPONSE PARA"ETER IS REACTION RATE CONSTANT K(1.I'H)")
6ee FORI'IATU/.2X."THIS EXPERIMENT IS A f"RACTIONAL FACTORIAL OF".
+IX. I2.1X. "'ACTORS". //. SX. "TEST CODE". SX. "RESPONSE")
7ee FOR"ATCO'.2X. "THIS EXPERI"ENT IS A FULL f"ACTORIAL OF".
+IX. 12.1X ... F'ACTORS" ..... ". Sx. "TEST CODE". s)(. "RESPONSE")
see f"ORMAT(/.7X.AS.3x.Fle.6)
98e F'ORMATU/ .... SX. "EF'f"ECTS")





1888 F'OR"ATC/.SX.Al.1X.":".IX.F'18.6)
28ee F'ORI'IAT(/.SX.2AI.1X.":".IX.F'18.6)
3eee F'ORMATc/.SX.3Al.1X.":".lx.F'le.6)
STOP
END
238
239
APPENDIX G.2
Examples for the Calculation of the Effects
240
G.2 CALCULATION OF THE EFFEC'rS
~ following is an example calculation of the effects on
nickel cyanide complex solution solved by using given matrix of
independent variables and Equation 6.1 (applied in program FACT).
I A B C AB AC BC ABC Yx10-
3
+ + + + 2.69
+ + + + 4.06
+ + + + 2.98
+ + + + 4.16
+ + + + 0.81
+ + + + 1.49
+ + + + 2.68
+ + + + + + + + 0.83
Effect of A _ (-2.69 +4.06 -2.98 +4.16 -0.81 +1.49 -2.68 +O.83)x1Q-3
4
- 0.34 x 10-
3
Effect of B (-2.69 -4.06 +2.98 +4.16 -0.81 -1.49 +2.68 +0.83)x10-
3
- 4
- 0.40 x 10-
3
Ef
(-2.69 -4.06 -2.98 -4.16 +0.81 +1.49 +2.68 +O.83)x10-
3
feet of C - 4
- -2.02 x 10-
3
Effect of AB _ (+2.69 -4.06 -2.98 +4.16 +0.81 -1.49 -2.68 +O.83)x10-
3
4
. - -0.68 x 10-
3
Effect of AC .(+2.69 -4.06 +2.98 -4.16 -0.81 +1.49 -2.68 +0.83)x10-
3
4
- 0.21 x 10-
3
Effect of ABC= (-2.69 +4.06 +2.98 -4.16 +0.81 -1.49 -2.68 +0.83)x10-
3
4
= -0.59 x 10-
3
241
242
APPENDIX G.3
Half Nbraal Probability Plots
G.3 PROBABILITY (%) AND EFFECTS ORDERED IN ASCENDING MAGNITUDE
- HALF NORMAL PROBABILITY PLOTS
HaCH
Order Humber i 1 2 3 4 5 6
Effect x10
3
-6.78 -3.10 +0.58 +2.90 +5.68 +8.80
Identity of
effect ac abc ab be a b
P-100 (i-1/2)/7 7.14 21.4 35.7 50.0 64.3 78.6
NaCN
92.9

TEMP.
.-.
78.6

......
>


./
:J
CD
<
m
0
Q:
21.4
-6.0 -3.0 0.0 3.0 6.0 9.0 12.0 15.0 1&0
EFFECTx10
3
Figure G.1. Half normal probability plot for HaCH
7
+18.04
c
92.9
243
244
Na
2
Cu(CN)3
Order Number i 1 2 3 4 5 6 7
Effect xlq3 -2.60 -2.44 -0.185 -0.175 +6.39 +6.85 +7.92
Identity of
effect a ac ab abc bc b c
P(%) 7.14 21.4 35.7 50.0 64.3 78.6 92.9
TEMP.
92.9

AIR
.....
78.6

lit
"-'
AIR ,TEMP.
> 64.3

5
SO.O
m

35.7
L
2\4
7.14

/
-4.0 -2.0 0.0 2.0 4.0 6.0 8.0
EFFECTS X 10
3
Figure G.2. Half normal probability plot for copper cyanate complex
245
Na2Zn(CN)4
Order number i 1 2 3 4 5 6 7
Effect xl0
3
+1.78 +4.97 +6.33 +9.55 +10.39 +12.33 +13.52
Identity of
effect bc b ac abc ab a c
P(%)
7.14 21.4 35.7 50.0 64.3 78.6 92.9
Na
2
Zn [ N ] ~
OD 2D .D 6D ao 10.0
EFFECTS x 10 3
Figure G.3. Half normal probability plot for zinc cyanide complex
246
Na2 Ni(CN)1t
Order number i 1 2 3 4 5 6 7
Effect xlO
3
-2.02 -0.93 -0.68 -0.59 +0.21 +0.34 +0.40
Identity of
effect c ac ab abc bc a b
P(%) 7.14 21.4 35.7 50.0 64.3 78.6 92.9
Na
2
Ni[CNJ.
/
92.9

..... 78.6
.uv
-
.....
AIR, TEMP.

64.3
-' 50.0
m
I
35.7
/.
21.4

7.14
-2D -1.6 -t2 -0.8 -0.4 0.0 0.4 0.8 1.2
EFFECTS x 10 3
Figure G.4. Half normal probability plot for nickel cyanide complex
247
K
3
Fe(CN)6
Order Number i 1 2 3 4 5 6 7
Effect xlO
3
-4.49 -3.76 -3.71 -3.43 +4.39 +5.56 + 7.97
Identity of
effect bc abc b ab ac c a
p (%) 7.14 21.4 35.7 50.0 64.3 78.6 92.9
K
3
Fe[CNJ
e
92.9
.UV
,....,
78.6
TEtvP.

.....
>
l-
64.3 UV. TEMP.
i
50.0

<
J

35.7
I
Q. 21.4
7.14

I
-4.0 -2.0 0.0 2.0 4.0 6.0 6.0
EFFECTS x 10 3
Figure G.5. Half normal probability plot for iron cyanide complex
Low Mix
Order number i 1 2 3 4 5
Effect xlQ3 -0.71 -0.37 -0.26 +0.16 +0.99
Identity of
effect
bc ac abc c b
p(%) 7.14 21.4 35.7 50.0 64.3
lDN MIX
92.9
.UV
-
7a6
.UV. AIR


> 64.3
AIR
to-

SO.O
c(

35.7
0.. 21.4
/
7.14
-08 -Q4 QO QA 08 12 1.6 2.0
EFFECTS x 10 3
248
6 7
+1.48 +1.96
ab a
78.6 92.9
Figure G.6. Half normal probability plot for low mix cyanide solution
High Mix
Order number i 1 2 3 4 S
Effect x1o.
3
-0.135 -0.12 +0.025 +0.12 +0.185
Identity of
effect abc a ac bc b
P(%) 7.14 21.4 35.7 50.0 64.3
HIGH MIX
92.9
.TEMP.
~
7&6
lit
~
>
64.3
l-
i
SO.O
c(
35.7
~
21.4
7.14
~ 00 ~ ~ ~ ~ W u
EFFECTS x 'K}.
249
6 7
+0.27 +1.08
ab c
78.6 92.9
Figure G.7. Half normal probability plot for high mix cyanide solution
APPENDIX G.4
Calculation of 95% Confidence Intervais
from the Half Morael Probability Plots
250
251
G.4. THE DETERMINATION OF SIGNIFICANT EFFECTS BY CAI.CULATING 95%
CONFIDENCE INTERVAL (CI)
NaCN (See
Mean -
a=
0.025
ts
-
Fig. G.l)
Probability
Transform*
X
7.14
9.40
10.80
11.95
13.15
14.50
2.35 x
10-
3
5.92 x
10-
3
2.571
Effects
Y x 10
3
-5.97
-2.27
1.27
4.32
6.47
10.27
0.025
Therefore 95% CIon random effects = Mean t a
u
- 2.35 x 10-
3
2.571 (5.92 x 10-
3
)
- (-12.9 x 10-
3
, 17.5 x 10-
3
)
The temperature effect is significant at 95% confidence level.
* "Probability transform" term used in this thesis refers to the lower
units (mm) read off the probability graph 1n order to be able to
perform linear regression calculations.
Cu-CN (See Fig. G.2)
Mean -
a =
0.025
Probability
Transform
X
7.14
9.40
10.80
11.95
-1.35 x 10-
3
1.353 x 10-
3
t3 3.182
Effects
-2.600
-2.440
-0.185
-0.175
The 95% CI is - -1.35 x 10-
3
3.182 (1.353 x 10-
3
)
- (-5.65 x 10-
3
, 2.95 x 10-
3
)
252
Therefore temperature, air and their interaction are significant at the
95% confidence level.
Zn - CN (See Fig. G.3)
Probability
Transform
Effects
Mean -
0=
0.025
ts
-
X
7.14
9.40
10.80
11.95
13.15
14.50
7.56 x 10-
3
3.91 x 10-
3
2.571
95% CI - 7.56 x 10-
3
2.571 (3.91 x 10-
3
)
- (-2.49 x 10-
3
, 17.6 x 10-
3
)
1.78
4.97
6.33
9.55
10.39
12.33
253
Therefore, none of the effects are significant at the 95% confidence
level.
Ni - CN (See Fig. G.4)
Probability Effects
Transform
X Y x 10
3
9.40 -0.93
10.80 -0.68
11.95 -0.59
13.15 0.21
14.50 0.34
16.60 0.40
Mean -
-2.08 x 10-
3
a- 5.89 x
10-
4
0.025
ts - 2.571
95% CI is - -2.08 x 10-
3
2.571 (5.89 x 10-
4
)
- (1.31 x 10-
3
, -1.72 x 10-
3
)
The effect of temperature is significant at 95% CI.
254
Fe - CN (See Fig. G.5)
Probability Effects
Transform
X Y x 10
3
7.14
9.40
10.80
11.95
Mean - -3.85 x 10-
3
a -
4.5 x 10-
4
0.025
t3 =
3.182
95% CI - -3.85 x 10-
3
3.182 (4.5 x 10-
4
)
- (-5.29 x 10-
3
, -2.41 x 10-
3
)
-4.49
-3.76
-3.71
3.43
Therefore, UV, temperature and their interaction are significant at 95%
confidence level.
255
Low Mix (See Fig. G.6)
Probability Effects
Transform
X Y x 10
3
7.14 -0.71
9.40 -0.37
10.80 -0.26
11.95 0.16
Mean ... -0.295 x 10-
3
(J ..
0.36 x
10-
3
to.0025
3 -
3.182
95% CI = -0.295 x 10-
3
3.182 (0.36 x 10-
3
)
= (-1.43 x 10-
3
, 0.84 x 10-
3
)
256
Therefore UV, air and their interaction are significant at 95%
confidence level.
High Mix (See Fig. G.7)
Probability
Transform
X
9.40
10.80
11.95
13.15
14.50
Mean - 9.6 x 10-
5
a =
1.50 x 10-
4
0.025
t4 - 2.776
95% CI - 9.6 x 10-
5
2.776 (1.5 x 10-
4
)
- (0.51 x 10-
3
, -0.32 x 10-
3
)
Effects
-0.12
0.02
0.12
0.18
0.27
Therefore, temperature is significant at 95% confidence level.
257
APPENDIX G.5
Backward Eli.Ination Procedure for the
Deter.Ination of Significant Effects -
Linear Regression with Sequential F-tests
258
tit",. '"'''''' itt ""'" hi if , .......... , .. ""'"
PROGRAM LJUB7
*, , "., .. t, "'i'.iiii"i ... ' .. "ii .. iiiii ..... ' ...... ti ........ ' ...... 'i'i' .
TO FAtILlllllE Bl!I;KWlli<D ELItilNIiTlON
. 1.) tLIMINATI0N rROCtDURE: DRAPER SMITH P. 167-169.
2.) StQUENTIAL F-TlS1S:- SMITH, P. 7,-72,
3.) XTRA SUH Of SwUARES:- RAPER SHITH, p, 67-hS.
4.) LlNEAk USING KIITklClS:- SMllH, P.
S.) SUM (jf fOR lACK OF fIT:- DRAPEf. SMITH, P, 3v-31.
6,) COMPUTATlONAL FUK COkklLA110N hAiKIX:-
DRPAi:.R StH1H.
INrURMATION TO SUPPlltb IN fILl: AS tOLLOWING:
(------- X MATR1X ----------) RtsrONStS:
1 X1i Xi2 Xil Xi4 ... XiK Yi
1 >:21 X22 X23 , X2K Y2
1 Xli X32 ............. X3K '(3
1 XNi XN2 XNJ XN4 .. XNK YN
iNSTRUCTIONS FGR RUNNING PROGRAH:
RU,BAKLK,LU.N.Nl,NC.HF. WMERE:
LU=READ LuGICAl UNIT HO.
li:-LlS1 LOGICAl. UN11 NO.=6 IN PkOGRAM)
N=NO. OF RUHS=NO. OF IN X MATRIX
Nl=Nu, OF CuLS. IN INhlT X-MATRIX
NC=NO. OF COlS, IN FOR:
NOD: RlAIlS IN X-MAnlX AS II (NI,H) HAH:IX
THEN KASSAGES Ift?UT KATRIX TO OBTAIN A (HC.HI
MATRIX; sa: PkOb.:Ati CODE. LINE 139 flELQlo!.
NC{O : INPUT KATR!X IS USED AS IS.
ti=DETERKINlS TYrt OF OUTPUl AS FOLLOWil:
ftR)O : RlPICATES:EXPANDlD ANQVA TABLE'
EXTRA SUn Of SQUARES CALtUi.Al10NS,
ftR=O : NO RPS: AHOVA TABLE
lXTRA sun Of SgUAr.ES
NR{O : NO REPS: BASIC ANOVA TABLE ONLY.
(SEt DRAPER SMITH P. 62 COMMENT ON
USlk KUST IN ADDlTION:
1.) HAKE OF FILL X MATRlX & Y RESPOHSE VECTOK.
2.) fROM fABLE: t(n-np,951). FOR b-VALUE CONFIDENCE LEvELS.
l.) FOR RlII-!S "HlCH INCLUDE. REPICilIES: SUM OF SQUAP.ES FOt: fUKE
ERROR & DEGREES OF FREEDOH ASSOCIATED WITH
PUkE EkROk (DtPl).
l(n-np-ne ,ne,95,) , FOR LACK Of TES1,
s.) TABLES: f(1.n-nD.95Z). FDR TESTING PARTIAL F-VALUES.
--------------------------------------------------------------------
259
TO LOAD BAKU!:
LOADR: Ll.
LfiADR: LI.lnLiB2
lOADP.:
LOADR: SL
LOII!'R: EN
IF SILt OF INi'Ul MTRIX IS f:iRi:.AH.k THAN 22 USt.:
LOADR: OP.LB
----------------------------------------------_. ---.... _----_ ... __ ..
HHTrALllE
ltifLlCIT DlIUkE rKEtlS10N HdLP-Z)
DiiURf rRECISION
VELlOkS FUR COL. KEANll" STD. IILV.: CtiEANtNC).CSID{Kc)
IHHNS1I}N CHEAN(J) .CSTli(l)
ttAH:ICES FUR OUT LlNi:.AF; RlGklSS10Io!:
X-HArKIX ON: X(N ,XfX(NC ,NC) .xn (NC.1!
DlnENS10N X(8.7),Xl(7.B) ,XIX(7.7).XTY(7.1)
Y-vCTtiR HANIPULAT!ON: Y(N.i),YT(t,M)
DltitHSION
CORRELATION tiAIRIX: XWCNU ,XCHNl; ,N) ,)'.t1XUNL .NU JuHNC .NU
D!HENSION XC(8.7),XCT(7.8) ,Y.CTXC(7,7' ,COR(7.7)
B-VE.CTOR HANIPUi..H1!ON: B(Ni.: ,1> ,lIH1 ,Nc)
DIHNSION B(7,i),FT(i,7)
FOR SUit Of
DfnNSIOH SSYti(1.1),SSH(1.1)
lWK) ,lC(NU ,[(N,!)
DrHNSIUN IR(7l.IC(7),(B,1)
DHii.N5ION NAH.<S) .lDBl1(144)
KATkl[;ES NElES5?lK'i fOR CALCULAlING lXH:1l SUn Of
260
XD MAlkIX I'iANlPlILAHON: XD!N .t-!::-i) ,XDHNC-1,N) .Y.DXDTlN\:-1.NL-1) .XD1Y (Nl- i . U
DIMNSION XD(B.6) .XDT<6.Sl .XDTXD(6.6} ,XDTY<6.1>
Bu VECTOk HAN.IPUi..Al1UN:
DIHfNSION BD(6.1).BDT(1,6)
FOY. SUH OF
DIMENSION SSKD(1.1)
MISLELLANEOU5: IDR(Nt-1),IDC(NC-1)
DLHfNsrON IDR(6).lDC(6)
CALL <IP)
LU=lP(1)


NC=IP(4)
Nk=lP(S)
LL=6
WRITE(LL.311)LU.LL.H,NI,NC,Hk
11

IN AND WRITE OUT NAME OF FILE:
WRlll(LU,300i
00 fUkHAHi!X,ttHl:.v. fiLE IU,Hi:. CI.!NIA1N1Hi: X HAH:!X t y vtCTOK
C
)
Rt\D(LU,J01}HAHE
Ii f OKriAT (;sK2)

82 fOknAT(II.2X.rlLE C0H1A1NING CALIBr.AT10N
--.- -- ---_._--_ .... --------- ------... _----- -------_ ... _ .. _----- ---- .. _------ .----
OfLN FILE DATA READ IN OUT DATA:
CALL urN(IDFC1 ,IERR,MAKE)
If(lERR.GE.O)GO TO 303
WR ITt:: (LU ,504)
14 fURnAT(2X , "rlLE TO OPEN")
\iu TO 1060
13 CONIHiUE
II:!
IS CON11NUl:.
LL=SO
CALL READ!' (IDl:ll ,I!:.Rk I IlL ILl.Hi
If(LN.Q.-l)GO Til 30b
CALL CODt.

IHILGLNIGO TO 306
11=11+1
Gu TO 30;
16 CALL CLOSUHtBli,II:.RU
CtiNvRT FLOWS IN L/HIN TO FLOWS IN LISEe.
DO 370 Il"'l.N
CLARIfER
X(11.3)=X(II,3)/6.0Dl

X(11.4)=X(11,4)/6,OUl
370 CON flftU
LF(NC.LT.O)Gu TO 391
iiANUULATION OF LHIlUT Kt\TRIX IN ORDER TO EXAMINE ItHERAI'IONS
DO 390 11=1.N
DO in U=3.S
K2=-Kl+1
X(II.K1)=X(11.K2)
717 CONT lNUE
X(11.6)=X(lI.3)'A(Il,S)
10 117 j(1=4,6
K2q(1+1
X(Il.K1)=X(11,K2)
717 CONTINUE
90 CONllNUE
tf=NC
TO 724
261
91 ftP=NI
24 WRIll(LL,90i)
01 f'OktillHI! ,1DX,"it",-,-,-Hi LlNlAR SIJIJIlr.lS fhH".H'=)
WIHWLL.310)
10 HJRI'iIlT (/ 1,10;':, I-------------X HATHX Y
DO 11 1=1,N
lti=lO
IFlNr ,l T .10) IH=rit'
WRIlULL ,99ti )(X(l ,J) ,J=l, 110, Y(l)
1f(NP.LE.10)GO TO 11

1 CONTiNUE
90 FORhlll(/,lX,11(Dl0.J,lX
23 fORHAT(5X,30(DiG.3.1X
DU EUilNE COLUMN IiEANS ST ItNDAkD U!:. V1II TI ONS:
DO i81 J=2,NP
JJ=J-1
C$iJ!\=O.ODO
DO 702 l=l,N
02 CSUn=CSUM+X{I,J!
CHlAN(JJ)=CSUM/N
11 CllHTlHUE
YSU;;=O.OllO
DO 103 1=1,1
13 YSun=YSU;;+Y(l)
CKtiltHHP
YM!:.hN=CM!:.Ato!(NP)
WP.if our TH COLUMN MEANS:
VB TE(LL. SO)
o HlANS:
C
)
ftX=i'iP-i

1 fORnAf{11X.20{Dl0.J,lX
DE1H:lilNE COLUnN STANDARD
0 i06 J=2.HP
13=J-1
CSUt\=O.ODO
DO 7ti7 I=l.N
87 CSUM=CSUM+X(1.J)-CMEAN(JJ,t2)
CSfu(J)=DSQRT(CSUH/(N-1.8DO
16 CuNTlNUE
i'Sutl=8.8DI
DO 70S l=l.N
IS

WKITE OUT TH COLUKN STANUltkD DlV1ATIOftS:

2 fOknKl(II.11X.
1
CULunN SlAHDltkD
WRIT(LL,Sl)(CSTD(J),J=2,HP),YSTD
Of X KATR1X
;,
262
DO 13 !=1.N
DO 12 1=1.Nr
12 XT(J,I)=X(I.l!
13 CONHhuE
WRITE(ll.991)
'91 ,SX. I X
DO 14 1=1.NP
14 WklTUll, )(XHl.J) .J=1.N)
CAKKY GUT MATKIX "ULT1PLICATION: XTX=XT*X
CALL NUT{XT .Ni'.NF)
wRITE(ll.992)
t92 X PKltiE X HIIlR1X L)
00 15 !=1.HP
15 WKi lULL. 990)(XTX(l,J) .1=1.Ni')
WRITE OUT MAHIX:
DC! 811 I=i.N
ftXX=NP-i
IiO BH 1=1.NXX
JJ:J+1
10 Xt(l,J)=X(I.IJ)
XC<I,ftP)=Y(!)
11 CONTHiijE
WRITt::(lL.812)
Bi2 fOKnAl(//,5X."XL hHlIX:")
DO 611 I=Ui
B17
TRAN!)flO$ Of XC MATRIX:
DO 813 1=1,N
DO 814 J=l,N'
14 X[T(J,l)=Xt(l,J)
13 COIHINUE
WP-lTUll
815 I [)RnA"! (f / SX. Xi:l tiIITHX: ")
DO ti16 !=i,NP
Bi6
OUT MATRIX nUll.:
tHLl MUT(XCT,XC,XCTXL.Nf.N,N,Nr.NP.Nf)

804 I UknAl(//.lX,'XLTXC KIIlklX:')
DO HOD I=1,NP
CSUtH =Nil.nt.iIN ( 1 )
DO tid1 J=l,NP
CSUHJ=NiCfttllft(l) .
Ii

t8 lOKi IHUl
wr rE(LL .805)
os KHTklX:
C
)
DO a02 l=l,NP
DO J=1.NP
263
IITll\=XCl XC (1 I ) .AllXL< J J)
03 LuR(l.J)=XCTXC(I.J)/DSQP.i(ZITERI
lti=iIi
IFHk .Ll.l0HII=Ni'
HCOR <I .J) .J=1. 1M)
Tu 802
W?lf(LL.881)(COR(I.J).J=11.NPI
80
rut<i'lAHSX.30(FS.2.SX
02 CONllNUl
CAkn OUl lNvE.k::iillN Ur XIX tiAlklX:
CALL GBSIN(XTX.NP .MP .Iel
WUTE(LL.994i
994 Hlkhll'[ (f / SY..X X' Hiiii:.kSi:. L)
j)Q 16
16 H:.(LL. 9S'U )(XTXn .J).J=1,NP)
ct\;{r.( OUR MATRiX iiiJUIPLICATIGN: X1Y=XltY
CI:LL MUHXi Y . XlV .Nf IN.N .1.NP .1)
WRHt:(LL.220)
20 MATRIX:")
DO 221 1=1.NP
221 WKHULL.99u)XTH1.1>
CARI<Y OUT MAW1X tlULTlPLlCAHuH: B=XTX-Ui:n
CALL KUT(XrX.XTY .B.Nr.N?NF .1.NP.i)
DLIERiHNi:. Ot EI KA1R1X:
00 i8 I=1 ,NP
i8 l:Hl,I>=[i(l.1l
Y KATR1X:
DO l=l.N
19 Yl(1.I)=Y(1.1)
CARkY OUT MATH X KULllPLlCAllON:
MUT(YT.Y.SSYII.1.N.N.1.1.1)
1,1)
BSS=ri*(YMEAN**2i
lSSC=TSSti-[lSS
CAkRY OUT KAlklX tiULllPLlCAHON: SSIi:.BHi:1Y
CALL KUT(ET.XTY.SSH,1.NP.NP.l.1.1)

CALCULATE SUh Ot

DllU=N
DfrC=H-i.onc
1)f K"'-NP-i. Oili!
fR=DFTC-DfK
DfB=1.DDO
ArlSH=XHSS/DFM
Xh)i(=RSS/DfR
F:Xi'lSHiXi'lSR
WRllE<LU.
83 TO DtTEkH1NL CUNrlDEWCE LIH1TS")
READ(LU,')STUDE
264
CALCULATE CONr lOCNCI:. LllmS FOj( 1:
1
5:
tIlUTE(LL,7BO)
BO fOkiiAHl1 .iX, W1TH 9SI. ClJNi"IDl:.NCE
DO 7&1 !=1.NP

ID=:lTUDHDSQRT(iriSK4IAl
BPLUS=B \1.1)+21:
,i)-IS
WU1E(LL.7B'S)STUDE,2A.2B
7BS rlJktiH! (iOX, Lt- VALUi:.=" ,F1S. 3. "ZA=" .ns. 3. "U:' .t-1'J. 3)
WRITt(LL.7B2)B(I,l),BMIN3.BPlUS
82 IUKnHl(lx.r,u.s.10X,f'O.5."-,F,e.S)
81 CONTINUE
DHEki'ilNl:. C!.lEtI- lClENl or MULTIPLE DETWilNATlON
CliDfT:(XtiSS/TSS;:)*i.OD2
WRlll(Ll.7B4)CnDl:.l
84 Of" MULTlI'Ll:.
(1X,f4.1. :X")
DlltknlNE PklD1CTED YIS & ReSIDUALS:
UF HATRIX MULTIPLICATION: E=XtB
WK HI::( LL. 'lBS)
88 fOF:nAT (/ I ,ax, "i'Kclll[;llJi VilLUES:' L)
CALL MUT(X,B.E.N.NP.NP,l.N,l)
DO 7i:1e I=1.N
RES: Y (l) -E ( 1.1>
WRIIE(LL.787)[(1.1),RlS
FUKMAT(1X,F20.S.SX.FlO.S)
86 CuNTINUE.
IF(NK.GT.O)CO 10 lSO
WRlTE(LL,230)
30 .SX." ... , ..... ", .. BASIC IINliiiil TABi..L
WRH(LL .231>
H I Oi<i'iAl(1 I .3X. I .iSX. SS: .1SX. al'i : a .17>:,


32 rOknHT(f I .3X. "nOlli:.L: .12>: J 10.5 .7XJB. 0 .10X .FlO. S .UX .riD. S)
WRHE(LL.233)RSS.nFR,XltSR
33
WRIT{LL,S4)BSS,DfB
4 I Oki'illHI I ,3X. 'nt.AN (1:0)-, S'X.F10. S .7X JI:L.)
.DfTC
(COP.k).&X
WR!TE(LL,234)TSSU.DfTU
34 fOknAT(II.3X.-lOTAL (UftLuKj()",3X.f1D.S.1X,FB.8)
If(ftR.LT.O)GO TO 3000
GO 10 264
SO CONfIftuE
RE.AII iN SUn OF FOR PUki. lkRIJI.: (kEPEA1S) , DE.!;Rt.!:.S OF n[lDOK:
WRH{LU.2S1>
51 fOkMAT(1X."ENT[k SS{g.e.) l n+(P.t), ie. SSrE.Df')
265
)SSPE.DfPE
SSLr=RSS-SSPI:..
DfLf=D;:I1-DFPE
Xtt5LF =SSU lIlt U
Xtl5ft:=SSPE!DFPE
FU=Xtt:iLf
WRITt::<Ll.2S3i
53 FOkiiAHII.10X."""''''''. EXhiNDEIi Mnuvil TAW:.

S4 I ORiiAl (f 1.3X. aSUiii<:LE: I "55: t .1SX. 'ilF :. ,
(1SX,;nS:",1SX,F:)

70 Fln:hlil (/ / ,3X, liuuEL: ,12X ,FiO. 5, 7X.tB. 0 ,lOX ,t H. 5 ,tox ,FiO. 5i
WRIT(LL,771)RSS,DFR,XKSR
71
WR!TE(LL,772}SSLF,DiLF,XH3LF,FLf
72 I,SX !i.Ai..f: 01- FlP ,SX ,HO. 5 .7X.F8. 0,
(10X.f10.S.10X,FI0.5)
WUTE(LL,773)SSPl.IiFPE,XtiSPl:.
73 FOknAl (/ / ,SX. 'i'Ut:L L ,6X .?d S. 0 ,iI/X
IIRHE(LL.774H{SS .DFB
74 F Oti'iAHII .3X. tit.RN .9X.FiIl. .'ll: JB. 0)
WRITt:(LL.17S)TSSC.DFTC
75 I OkiiAT(II.3X,"T01AL
WR IT!:: (LL. 776) TSSU .DFTU
16 I OU;liHII .3X. TOTAL AX J H .5 .7X.FB. 0)
64 CONHNUE
IIl:.TlRlilNE lX1RA sut! m tOR lALE B:
WRITt:(LL.492)
12 [XTkA sun or SwuARES
iifD.::HP-1
110 lO@O lK=2.Ni'
DUiiKi XD. USl:.D iN LXTRA SUn OF SQiJARES
JJ=l
DD 4iii J=1.NP

DO 4iiil ll:.l.N
00 XD(II.J)=X(II.JJ)
J3=3J+1
'1 COtHLNUE
W;;lTt<LL,403)
403 rORItAHII,lOX. a _________ Xl) HIITUX YD vtCTOR ------.---.)
DO 404 I=l.H
484 WklTl(LL,9YO)(XJ)(l,J).J=1.HPIII,Y(1)
TRANSPOSE OF XD KATRIX
DO 405 1=1.N
DO 406 J=l,NPD
16 XDT(J.I)=XD(l.J)
IS COtHINUE
WRIlE<LL,4i17i
266
87 fORMAT(//.5X," XD
00 l=I,Hrv
4DB wr.llt(ll,99u)(XDT(l,J),J=1.N)
CAKKY OUT MATRIX HUlTIPLICATION: XXTXD=XDT,XD
CALL ItUT (XiiT ,Xli ,XllTXli .Ni'D.N.k .Nf'li ,Ni'ii)
WRH(Ll.409)
409 tOl(MAT(f/ .5X, XD I'W;[ XIi hHH:l).: L)
DO 410 I=1.HPD
410 Wio:i.TULL, <XDIXDn.J) ,J=l,NPD)
CAiRY OUT INVERSION OF XDTXD MATRIX:
C/iLL GllSHHXDTXII ,NFlI,Hrli DEl l1 .Ill( .IDR .IDC>
IiiR!T(Ll,41il
11 Hlki'iA 1 (/ / .5X, eXI, XD- Hii.it.P.SE L)
DO 412 1=1.NPD
412 Wj.;UULL,990)(XDTXI,(l .J) ,J=l,NPD)
OUT MIlTklX HULTH'Utl1TWN: XD1Y=XDlfi
CALL KUT(XDT.Y.XDTY.HPD.H.N,I.NPD,l)
W!U lULL .4 i3)
13 t'uKriATU! .5X. "XDTY iiATRIX: "j
DO 4i4 l=i,NPD
WRITE(ll,990)XDTY(l,1)
CAHY om KATR1X nULllt'LICA1l0N: J.Ill=XDTXD- UXi,TY
CALL HUf(XDrXD.XDTY.BD,NPD.N?v,NPD.l,HPD.l)
Wt:llUll.4iS)
415 F llknl11 (/ I,SX, tLLAS'1 LLTHIRiLS- Bil-VALIJES")
DO 416 I=l,NPD
416 Willl(ll,990) Eli(l.l)
DTtKiilNE Of BD HATRIX:
111.1 417 I=i.HPD
17 BDT(l,I)=BD(I,ii
Cilir.'i OUT KRTklX ItULTn'l!lATIQN:
CALL MUT(BDT.XDTY.SSnD.i,NPD,NrL.i.l.1)

WRllf(LL,718)XniiSS
7iO IllknAT(II.7X,-KOD[L SUn Or

Df XS=1. allb
fX$=XSS/XHSR
Ill=lk-l
IF(lll.GT,l)GO Tll 419
WU T<Ll ,420)
20 LAST:,11X."S:",1SX,uf:.17X,
("KS: ,l1X, F:)
19
21 tuRKAT(II,3X,"B(",!Z,")".13X,F18.S.7X.f8.0,
(18X,Fl0.S,18X.flb.S)
IDO t:ONTlNUE
800 CONTINUE
267
080 STOP
END
SUllh!UTlNi. ,DETA lEI-: ,n: ,lei
DOUBLE fRECI$ION (A-H,P-Z)
DlhENSIOk HiIM,lAl,lk(nAl,lC(hH)
DCIW KA1klX
IF (lA-ttA) 1.2,2
1 WklTt.<6 ,ij)
1R=1
Rl:.TURN
2 it::R=O
DD 3 I=1.tlM
IR (1)=0
3 lC(l)=O
DdA=1.
DU 8 IJKL=1.tlA
CALL
PIV=A(LJ)
bUA=PIViilETA
If(PIV.EQ.O.O} GO TO 12
lknl=J
IC(J)=1
PIV=1./PIV
DO 4 K=l,l\li
4 Mn ,K)=A(l,KltDv
A(!.J)=PIV
DO 6 K=1.KA
If(K.t:W.I) GO TO 6
P1V1=A(K,J)
DO S L=1,ttA
5
A(K,J):PIVl
b CONTI NUl:.
r lv1=AU ,])
DO 7 K=1,KA
7 A(K,J)=-PIV*A(K.J)
A(l,1>=PlV1
a CUtHlHUC:
DO 11 1=1,KA
K=!Cm
K:.l!(l)
IF(K.EQ.I) GO TO 11
DTA=-TA
DO 9 L
e
1,tIIt
ThP=A(K.U
A(F;,L)=A<I,U
9 A(I,U=TEtlP
J,I(J U L=1.tIIt
268
TI::KP=A(L,I1)
A(Uil=A(L,I)
10 fHL,Il=TEtiP
lC(I1)=K
IR(K)=K
. 11 COHHHUl:.
RIURN
12 Illo::1
I!:JURN
i3 rOkriATUH - DIKlNSlmi or Wui-i IS ( OF 1'i
*ATRiX .)
END
SUllki.iUTlNl:. GBSS'j(A, 111 ,lA ,tiA .NI'I. n .It .1.J)
!hrL!C!T DGUBLE PRECISION (A-H,P-Z)
PkLt:lSWN
A(lH,JA),IR(HA),IC(NAI
HilS SUBK 15 tili..LEii Ifj THE MAl ltivt.bltlN SUBF
1=0
J'"iJ
TST.:O.O
liv 2 K=1,tIA
IHIR(K) .iiE.O) GO TO 2
1>[1 1 l=l,HIi
IF(!C(Ll.NE.O) GO fO 1
X=DHhS(A(i( ,Ll)
If(X.LT. TEST) GO TO 1
l=K
J=L
TlSl=X
1 CONfINUE
;: CONTlNUi::
RTiiRN
END

DOUBLE PRCISION (A-H,P-II
DIME.NSION AUI,N) ,t:<Il,Ni ,C(ti.N)
00 1 1=1.K
DO 1 J=1,'"
1 C(1.J)=A(1.J)+X*P(I.!)

END
SURkOUTINE.
DOUBLE PRlClSIO'" (A-H,P-Z)
DIHlNSION
00 1 1=1,HB
DO 2 1=1."11
SUK=D.
DO 3 \(z1.NA
269

3 COlflltilii.:
!:(LJ)=SUK
2 CUtHi.Niii:.
1 COHilNUE
RE1URN
END
SUI:i<uUlIN
IHPLllll DUUBLE PRt.CiS10N (Il-H,P-2)
DIMNSION A(H,NI,C(H,N)
DO 1 1=1,1'!
0 1 J=l,N
C<I,J>=X'k(l,J)
RfTliRN
ENi!
270
The Output fro. the BaCkward Elimination Program
(UUB])
271
272
ltl,. 1 U= i. H::. e HI:: 7 "'[:-1 Nr.= 0
FIlf COWTAIHINf, CAlIBRHTIOH SD07
FILE CONTAINING CALIBRATION DATA:
U**UUu. LINEAR LEAST UUUUU
----. --------- X KATRIX " Y VECTOP ----------.. ---
. 1000101 - .10GD+Oi -.110])+01 .10er'+01 .180D+01 .10DD+Ol .so:m-02
.118Dt01 .110])+01 - .119])+01 -.180D+Ol -.100D+Ol -. U@Dl01 .100D+Oi 139D-Oi
.118])+01 - .110])+81 .110])+81 -.180])+01 180D+Ol .186D+01 -.100D+Oi .740D-02
.1001HI1 .1IID+01 .110])+01 -.110])+01 .1I0D+01 -.119])+01 -.1D01i+01 Oi
.110+01 -.108])+01 -.1I8D+01 188D+81 .1I0D+61 -.116])+81 -.110])+01 .238D-Oi
.180]){81 iDOD+01 -.110])+01 .100D+01 - .1001)+01 .100])+01 -.1I0D+01
.tlID+1i -.180])+01 180D+01 .180D+01 -.1&0])+01 -.180D+01 .nOD+01 .3941i-6i
.108D+81 .180D+01 180D+01 196D+81 .19OD+81 .lIBD+01 .1881+01 .389t'-@i
COLUMN MEANS:
.101])+00 .180Dt80 .088])+00 .IOOD+OO .8DO])+00 .100D+08 22H'-9i
COLUMN STANDARD DEVIATIONS
.117])+01 .1I7D+01 .117+01 .107D+01 .1071)+01 .1I7D+01 129D-D1
CORRELATION MATRIX:
1. ..
'.00
'.11
.... . ... .... .27
'.10 1."
....
t."
1.10
'.ID
.42
....
'.10
1."
....
I.'.
'.10 .80
.... . ... 1.11 1.ID 1.88 1.18 .IS
....
'.18
.... . ... 1.11 1.11 -.25
'.10
.... .... ....
.27 .42 .so .8S
WITH 9SZ CONfIDENCE LIMITS
.82214
.11323
.USi3
.10971
.IOOtA
-.10296
.10217
....
-.25

.10044-
.10234-
.80692-
. 00215-
-.eOS7e-
-.80063-
COEFrlCIEWT OF MULTIPLE 99.1%
PREDICTED VAlUES:
.10389
.81504
.10854
.12226
.12494
.12416
.13826
.14004
RESIDUALS:
.00114
-.10114
-.e01H
.19114
-.OHH
.09114

-.10114
*tt'.**i** BASIC ANOVA TABLE ,._,tt*t**
SOURCE: st: Df:
KODEl: .18116 6.
RESIDUAL .11881 1.
MEAN (Bt) .11392 1.
TOTAL (COlR) .18111 7.
TOTIIL <UHCDRR) .11519 8.
1.10
.1B
.1S
1.00
.82493
.10603
.80793
.I12S4
.00343
-.DOOiP
.00496
KS
.01019
180 Ii
273
r:
le.5b48b
274
uunun rxnK SUfi or SQUARES *tuu __ u
ADDED LAST: SS: 1If': liS: r:
B( 1)
.9880B 1. OOOS 8.02081
B( 2) .08021 1. .80021 20.23522
i( 3) .80075 1. .DD075 72.3ii68
B( 4) .80000 1. .80000 .31571
J( S) .18007 1. .DOOO? 6.83539
B( 6) .80104 1. .16004 3.66293
APPENDIX B.l
Example Calculation of the Initial Theoretical
Metal Cyanide Concentration
275
276
H.I CALCULATION OF THEORETICAL METAL CYANIDE CONCENTRATION AT THE
BEGINNING OF EXPERIMENT (EXAMPLE CALCULATION FOR. Fe-CN at 4C,
AIR.. UV)
Measured total cyanide [TCNJo - 193 mg/L - 0.00742 mol/L.
Analytically determined metal concentration [Fe
3
+J = 12.04 mg/L. The
stoichiometric relationship for ferricyanide is
--- - 2.795
where molecular weight for cyanide is M W ~ = 26.02 g/mol and
molecular weight for iron is MW
Fe
3+ = 55.85 g/mol. Therefore, the
initial cyanide concentration tied up with iron present will be:
[MCN Jo - 2.795 x 12.04 - 33.65 mg/L ., 0.00129 mol/L and the initial
free-cyanide value will be:
[FCNJo - [TCNJo - [MCNJo - 0.00742-0.00129 = 0.00613 mol/L.
APPENDIX B.2
Computer Prograa Used for Four-Para.eter Estlaatlon Model
(FCN, MCN, ltv, 1t
1
)
(Example: Synthenlc Iron-Cyanlde at 4C, AIR, UV)
277
PROGRA"
+DE8UG=OUTPUT,TAPE11)
DI"ENSIOH Y(2S),TH(4),DIF'F'C4),SIGNSC4),SCRATC60BB)
C ********************.*.*.*****************.*****.***
C F'OUR PARA"ETER "ODEL
C*.**.******************.****.****************************
CO""ON /8/ X(25),YPC6,61),DELT,NTST[P,TT(61),NT1,YINIT
EXTERNAL "ODEL
DATA NP,"IT/4,15/
DATA HOB/1S/
DATA NPR08/!/
DATA [PSI, EPS2, F'LM.F'NU/l.E-6.1. E-6. B. Bl. lB. /
READCS ) CTHCI),I:l.NP)
WRITEC6.1B8) (TH(I).I:l.NP)
IBB F'OR"AT(4F'12.9)
READCS.*) C)(I).V(I).I:l,NOB)
WRITE(6.*) ()(I).V(I).I:l,NOB)
VAU=B.
DO 38 l:l,NOB
38 VAU:VAU+V(I)
YAU:YAU/F'LOAT(NOB)
TSS:8.
DO 35 I:l,NOB
35 TSS=TSS+CVCI)-VAU)**2
TSOS:TSS/(26888 *2)
WRITE(6.288)TSOS
2B8 F'OR"ATC1X,"TOTAL Sutt OF' SQUARES IS"E14.1)
DO 48 J:l,NOB
VI :Y(J)/268B8.
Y(J):Yl
CALL PLOTPT()(CJ).Yl.5)
48 CONTINUE
YINIT:VU)
DELT=2.
NTSTEP:)(CNOB)/DELT
DO 4S I:l,NP
DIF'f"(I):B.Bl
45 SIGHS(I):1.8
CALL UWHAUSCNPROB,"ODEL.NOB,V,NP,TH,DIf"f".SIGHS,EPS1,EPS2,"IT,
+f"LA",f"NU.SCRAT.6)
C CALL "ODEL CNPROB.TH.F'.HC>B,NP)
DO 58 J:l.NTl
N
......,
ex>
T:T1(J)
CALL PLOTPT
CALL PLOTPT (T,VP(2,J),2)
CALL PLOTPT (T,VP(3,J),2.)
5. CONTINUE
CALL OUTPLT
STOP
END
SUBROUTINE
DI"ENSION
COMMON X(ZS),VP(6,6l),DELT,NTSTEP,TT(6l),HT1,VINIT
INTEGER H,IND,NW,IER,K
REAL IX,IV
REAL
COMMON RKC,VK,UVKN,XK.VK,IX.IV
EXTERNAL
NW=3
N:3
XX=
IT"AX:l
)+1
VV(l):TH(l)
VV(ZhTH(Z)
n l):VINIT
C CONSTANT DElT "UST BE LESS THAN OR EQUAL TO "INI"U"
C TIME BETWEEN OBSERVATIONS
T=B
.10BS:2

XK:TH(3)

IX:5


TTU ):T
VP U, 1 ) : VV ( 3 )
VP(Z,1):IX*VV(1)+IV*VV(2)
VP(3,1):VP(1,1)+VP(Z,1)
TOl: l
IND:l
DO 55 K:l,NTSTEP
N
-....,J
'-"

CALL DVERK
(IND.LT.8.0R.IER.GT.8) GOTO 488
T:T+DELT
JPRED:(X(JOBS)/DELT)
YVV:IX*VV(1)+IV*VV(2)+VV(3)


TO 388
GO TO 55
388 KK:(K/IT"AX)
TT<KK+U:T
VP(1,KK+l):YV(3)
YP(2,KK+l):IX*VY(1)+IY*YY(2)
VP(3.KK+l):YY(3)+VP(2,KK+l)
55 CONTINUE
48. CONTINUE
NT1:KK+l
RETURN
END
SUBROUTINE
COMMON /A/ RKC,VK,UVKN,XK,YK,IX,IY
REAL VV(N),VPRI"E(N),XX,IX,IV
VPRI"t( 1) :-XK* (VV( 1
VPRIME(2):(XK*(YY(1)-(VK*(YY(2)
YPRIME(3):(IX*XK*YY(1-(IV*XK*YV(1+(IV*YK*(YV(2)-
+(VK*. 99*YY( 3
RETURN
END
SUBROUTINE
1 MIT, rLA". rNU. SCRAT,LL)
DI"ENSION SCRAT(l)
1
IA:l
IB=IA+NP
IC:IB+NP
ID=IC+NP
If"=IE+NP
IG:Ir+NOB
IH:IG+NOB
II : IH + NP * NOB
IJ : IH
CALLHAUS9(NPROB,"ODEL.NOB.V,NP.TH.DIrr,SIGNS.EPS1.EPS2."IT
,rLA".rNU,SCRAT(IA), SCRAT(IB). SCRAT(IC), SCRAT(ID),
N
CO
o
2 SCRAT(IE). SCRATCIG). SCRATCIH). SCRAT(II).
3 SCRATCIJ).LL)
RETURN
END
SUBROUTINE HAUS9(NPRIO. "ODEL. NIO.

C.
C
C DI"ENSION THCNQ). SIGNSCNQ). YCNIO)
C DI"ENSION Q(NQ). P(NQ). E(NQ). PHI(NQ). TBCNQ)
C DI"ENSIOH RCNIO) .
C DI"ENSION ACHQ.NQ). DCNQ.NQ). DELZ(NBO.NQ)
C
DI"ENSION TH(l). SIGNS(1). Y(l). Q(l). P(l), E(l).
1 PHI(l), TB(l). R(l). A(1). O(1). DELZ(l)
ACOS(X) : ATANCSQRT(I.a/X**2 - 1.a
HP = HQ
HPROB = NPRIO
NOB : NIO
EPSI = EPIS
EPSZ : EPZS
NPSQ : NP * NP
NSCRAC = 5*HP+NPSQ +Z*NOB+NP*NOI
WRITECLL.IBBB) NPROI. NOI. NP. NSCRAe
WRITECLL.laBl)
CALL GASSaCl. NP, TH, TE"P, T"EP.LL)
WRITE(LL.laa2)
CALL GASSaCl. NP, TE"P. TE"P.LL)

15 99. 16
16 CONTI HUE
DO 19 I=I.NP
:
AISCTH(I99. 99. 19
19 CONTINUE
GA : rLA"
NIT: 1
LAOS = a
IrCEPSI) 5.7a.7a
5 EPS1: a
7a SSQ : a
CALL "ODELCNPROI. TH. ,., NOI. NP)
DO 9a I = I. NOI
N
00
C
C
C
c
ReI) : vel) -
98 SSQ:SSQ+Rel).Rel)
WRIT[eLL.1883)SSQ
188
181
18
OA : GA /
INTCNT : 8
WRIT[eLL.1884)NIT
JS : 1 - HOB
DO 138 J:1> HP
TEMP: TH(J)

TH(J): TH(J)+peJ)
Q(J):8
JS : JS + NOB
WRITEC6. HI) J
FORMATC1H 15)
C *
CALL MODELeNPROB. TH. DELZeJS). NOB. NP)
IJ : JS-l
DO 128 I : 1. NOB
IJ : U + 1
DELZeIJ) : DELZ(IJ) -
128 Q(J): Q(J) + DELZeIJ) ReI)
Q( J): Q( n/pu)
BEGIN ITERATION
C Q:XT.R eSTEEPEST DESCENT)
138 THeJ) : TEMP
IFCLAOS) 131.131.414
131 DO 158 I : 1. NP
DO 151 J:l.I
SUM : 8
KJ : NOB*U-1)
KI : H08*(I-1)
DO 168 K : 1. NOI
KI : KI + 1
KJ : KJ + 1
168 SUM: SUM + DELZeKI) DELZeKJ)
SUM/(peI)*PCJ
JI : J + HP.CI-l)
DUJ) : TEI1P
IJ : I + HP*CJ-l)
151 D(IJ): TEMP
N
00
N
158 Eel) : SQRTCDeJI
666 CONTINUE
DO 153 I: 1. NP
IJ : I-NP
.8.T8CI0 153 J:l.1
221
221
231
663
665
664
662
669
229
225

248
265
266
CONTINUE
SUI'tI:8
CALL 1't0DEL(NPROI. TB. F. NOI. NP)
DO 238 l:l.NOI
R ( I) : Y ( 1 ) -f"( I )
SU"I:5U"I+R(I).R(I)
WRITE(LL.1843)SU"1
- (1.8+EPS1).SSQ) 662. 662. 663
A"INleTE"p-31.8. GA 665. 665. 664
STEP:STEP"2.8
INTCNT : INTCNT + 1
- 36) 178. 2788. 2788

INTCNT : INTCNT + 1
- 36) 666. 2788, 2718
WRITECLL.1817)
DO 669 l:l.NP
THU):TBCI)
CALL GASS8(1, NP, TH, TE"P, TE"P,LL)
GA, SU"B
229.229.225
278.278,265
DO I : 1, NP
248, 241
278,278.265
CONTINUE
WRITE(LL,lSe9)EPS2
GO TO 2B8
- SSQ) - [PSI.SSQ) 266, 266, 278
WRITE(LL.1818) EPSI
GO TO 288
278 SSQ:SUMB
NIT:NIT+l
- "IT) 181, 188, 288
2788 WRITE(LL.2718)
2718 FOR"AT(""11SH8* THE SU" OF SQUARES CANNOT BE REDUCED TO THE SU"
10F SQUARES AT THE END THE LAST ITERATION - ITERATING STOPS ,,)
N
00
w
C
C
EHD IT[RATIOH
" 288 WRITE (LL.181l)
WRITE(LL.2B81) (F(I). I : 1. HOB)
WRITE(LL.1B12)
WRITECLL.2BB1) (RCI). I : 1. HOB)
WRITE(LL.IB17)
1811 FOR"AT("".H X PRI"E-X "ATRIXH)
CALL GASS8C4.HP.TE"P.TE"P.D.LL)
SSQ:SU"B
IDF:HOB-HP
WRITE(LL.181S)
1=8
CALL "ATIN(D. HP. P. I. DET)
DO 7692 l:l.HP
II : 1 + HPe(I-l)
1692 E(I) : SQRT'D(II
DO 348 I=loHP
348
JI : 1 + HP*CI-l) - 1
IJ : 1 + HPeeI-2)
DO 348 J : I. HP
JI : JI + 1
A(JI) : DeJ1) ([el).[(J
IJ =" IS + HP
A(U) : AU!)
CALL GASS8C3, HP, TE"P, TE"P, A,LL)
WRITE(LL. Un6)
CALL GASS8(1. HP, E. TE"P. TE"P.LL)
IFeIDF) 341. 418. 341
341 DEU: SSQ IDr
WU) SDEU. IDF
DEU)
DO 391 I:t.HP
P(I):THeI)+2.8.E(I).SD391 TBel):THel)-2.8
CALL GASS8(Z. HP. TB. P. TE"P.LL)
LAOS : 1
GO TO 181
414 DO 415 K : 1. HOB
TE"P : 8
DO 428 l:l.HP
DO 428 J:l.NP
WRITE(LL.1839)
ISUB : K+H DEBUG1: DELZ(ISUB : DELZ(K + HOB.(I-l
ISUB : K+HOB.(J-l)
N
00
.J::-
DEBUGZ : DELZ(ISUB)
C DEBUGZ : DELZ(K + HOB.(J-l
11 : I + HP*(J-I)
DEBUG3 : D(IJ)/(DIFZ(I).THDIFZ(J)*TH(J
+ DEBUGI * DEBUGZ TEMP: Z.8.SQRT(TEMP(K):F(K)+TEMP
415 F(K):F(K)-TEMP
WRITE(LL,188S)
IE:8
DO 425 I:I,NOB,5
IE:IE+5
Ir(HOB-IE) 438,435,435
438 IE:HOB
435 WRITE(LL,2BBI) (R(J), J : I, IE)
425 WRITE(LL,2BB6) (F(J), 1 : I. IE)
418 WRITE(LL,IB33) NPROB
RETURN
99 WRITE(LL,1834)
GO TO 41B
laaa FORMAT(3BHINON-LIHEAR ESTIMATION, PROBLEM NUMBER 13,// IS,
1 14H OBSERVATIONS, IS, IIH PARAMETERS 17, 17H SCRATCH REQUIRED)
IB81 FORMAT(/25HBIHITIAL PARAMETER VALUES )
18B2 FORMAT(/S4HBPROPORTIOHS USED IN CALCULATIHG DIFFERENCE QUOTIENTS )
18B3 FORMAT(/25HBINITIAL SUM OF SQUARES: EI2.4)
IBB4 FORMAT(/////45X,13HITERATIOH NO. 14)
1887 FORMAT(/32H8PARAMETER VALUES VIA REGRESSION )
188S FORMAT(////S4H8APPROXIMATE CONFIDENCE LIMITS rOR EACH FUHCTION VAL
IUE )
18a9 FORMAT(/62H8ITERATION STOPS - RELATIVE CHANGE IH EACH PARAMETER LE
ISS THAH EI2.4)
lala FORMAT(/62H8ITERATION STOPS - RELATIUE CHAHGE IH SUM OF SQUARES LE
ISS THAN EI2.4)
1811 FORMAT(22H1FINAL FUHCTION VALUES
lal2 FORMAT(////laH8RESIDUALS )
1814 FORMAT(//24H8VARIANCE OF RESIDUALS: ,EI2.4,IH,I4,
128H DEGREES OF FREEDOM )
1815 FORMAT(////19HBCORRELATIOH MATRIX )
IBI6 FORMAT(////21HBNORMALIZIHG ELEMEHTS
1833 FORMAT(//19HBEND OF PROBLEM HO. 13)
lB34 FORMAT(/16HBPARA"ETER ERROR )
1839 FOR"AT(/7IHBINDIUIDUAL CONFIDENCE LIMITS FOR EACH PARAMETER (ON LI
IHEAR HVPOTHESIS) )
1848 rORMAT(/9H8LAMBDA :Ela.3, 4X,33HSU" or SQUARES ArTER REGRESSIOH :
IEI5.7)
N
00
\J1
e
p
1.41 FOR"ATCI4H DETER"IHAHT: EI2.4. 6X. 25H AHGlE IH SCALED COORD. :
i F5.2. BHDEGREES )
la43 FOR"ATC2BHaTEST POIHT SU" OF SQUARES = EI2.4)
2aal FOR"ATC/5EI2.4)
2a86 FOR"ATC5EI2.4)
EHD
SUBROUTIHE "ATIHCA. MUAR. B. H8. DET)
DI"EHSIOH ACHVAR. I). BCHVAR. I)
PIVOT" = ACI.I)
DET = 1
DO 558 leOl = I. HVAR
PIVOT = ACleOL. leOl)
PIVOT" = A"IH1(PIVOT. PIVOT")
DET = PIVOT DET
e DIVIDE PIVOT ROW 81 PIVOT ElE"EHT
e
ACleOl. leOl) = I.a
PIUOT = A"AX1CPIUOT. 1.[-2a)
PIVOT = ACICOL. leOl)/PIVOT
DO 3S8 l=l.HVAR
35. ACICOl. l) = ACleOl. l).PIVOT
IF"CHB. EQ. II> GO TO 311
DO 37. l=l.HB
31.
e
BCleOl. l) = ICICOl, l).PIUOT
C
e
371
45a
REDUCE HOH-PIVOT ROWS
DO 55. ll=I.HVAR
IFCll .EQ. leOl) GO TO 55.
T = ACll. leOL)
ACll. leOl) =
DO 45. L=I.HVAR
ACl1. l) : ACll. L) - ACleOl. l).T
IFCHB .EQ. a) GO TO 55.
DO 588 l=I.HB
5.. BCll. l) : BCll. L)-BCleOL.L).T
55. eOHTlHUE
RETURH
EHD
SUBROUTIHE GASSaCITVPE. HQ, A. I, e,LL)
DI"EHSIOH ACHQ),ICHQ).eCHQ.HQ)
N
co
(T\
NP :NQ
NR : NPI'18
LOW: 1
LUP : 18
18 NR )15,28,38
15 RETURN
28 LUP:NP
.GT. LUP) RETURN
38 WRITE(LL,588J (J,J:LOW,LUPJ
GO TO
WRIT(LL,6aaJ(A(J),J:LOW,LUP)
GO TO taa
68 WRIT(LL,6aa) (8U),J=LOW,LUP)
GO TO
8a 00 9a I:LOW,LUP
9a WRITE(LL,72BJI,(C(J,I),J=LOW,I)
LOWZ=LUP+1
IF(LOW2 .GT. NP) GO TO laa
00 95 I=LOWZ,NP
95 WRITE(LL,72BJI,(C(J,I),J=LOW,LUPJ
lBa LOW = LOW + la
LUP = LUP + 18
NR = NR - 1
GO TO la
S8B FORMAT(1'5X,17,9(SX,17
688
72a
END
N
ex>
-.....J
SUBROUTINE r .. sso
1
0;
18
15
21
Z5
31
.1061217

10
15
21
30

60
110
go
95
to.
5DD
60D
,20
'112' __
71/171 OPT=1
GA"SOUTYP(l NO,. A, 8, e,lll
AINOl,BINOl,e NO,NOI
NP = NO
NP = ,,"Pit n
lOll = 1
lUP = 10
I FI 115,20,30
RETURN
lUP= NP
IfllCII .GT lUPI RfTUqN
IIRITlll
I
5001 'JtJ=lOIl,lUPI
GO TO '4U,60 80

GO TO lOu
IIAITElLl,flOOI IBtJI,J"'lOW,lUPI
GO TO .. 0
DO 9D I=lOIl lUP
IIRITEtll,72&II.tetJ,II,J=lOIl,II
lOW7.=lUP+l
IfllOll2 .GT
1
NPI CO TO 100
DO 90; I=lOIl<:,NP
IIRITftll t720II,tCtJ,II,J=lOM,lUPI
lOM 2 luM + 1
lUP .. lUP 1
NR=NR-l
GO TO 10
FOR"'ATI/5X
1
I'r,t5x,I7JI
FORI1ATl5E1z.1o
FOR"'AT UHO ,13,' ttx ,F12._, I
END
.0215182
FTN 4.II+5H "4/04/29. 22.40.43 PAGE
I. 19l. 19. 160. 2_. lS5. _1. 1_3. _I. 13l. 65. 120. 72. 116. lt3. 76.6 63.3 151. 62.6 168. 55. 175. 192.
19'. _6.5 216. _3.' 223. 37.1 312. 27.8
TOTAL SU" OF IS .6955119E-0-
1
N
00
00
NO"-LINfAR ESTIMATION, PROBLEM NUI1BER 1
19 OaSE RVA TI ONS, It PARAMETERS
INITIAL VALUES
1
.61Z9F-DZ
3
.ZI5Ze:-01
150 SCRATCH REQUIP.ED
It
.6lt8!1E-02
PROPORTIONS USED IN CALCULATING QUOTIENTS
1
.1000E-01
Z
.1000E-Ol
INtT IAL SUM OF SQUARES =
3
.1000E-0 1
It
.10 COE-O 1
ITERATION NO. 1
DETER"'INA NT = .1678E-02 ANGtE IN SCALED COORD. =ltO.Z9nEGREES
TEST POINT VALUES
."852E-02 .2lt09f.-02 .1189E-01 -.18 a E-02
DETER"INANT = .767ltE-02 ANGLE IN SCALED COORD. =33.68DEGREES
TEST POINT VALUES
.5291E-02 .1983E-02 .1015E-01 -.31t51E-02
DETERMINANT = .1023E+00 ANGLE IN SCALED COORD.
TEST POINT PARAMETER VALUES
.5631E-02 .186 .. E-02 .1125F-Ol -.3783E-02
'EST POINT VALUES
.5880E-0? .1579E-02 .1638E-01 l31t9E-02
TEST POINT SUM OF SQUARES = .5lt98E-D5
PARAMETER VALUES VIA REGRESSION
1
.588OE-02
2
.1579E-02
3
.1638E-Ol
It
.13"9E-02
SUM OF SQUARES AFTER = 51t98
ITERATION NO. 2
DETER "INA NT E .8736E-02 ANGLE IN SCALED COORD. =51.66DECREES
TEST POINT PARAMETER VALUES
.5597E-02 .1701E-02 .9553E-02 .2303-D2
TEST POINT SUM OF SQU.RES = .1870E-05
289
PARAMETER VALUES VIA REGRESSION
2
.1701E-02
3
.9553(-OZ
It
.Z3C3E-02
SUN OF SQUARE S AFTER SSION = .1870 273E -05
ITERATION NO. :5
DETERP'INilNT = ANGLE IN SCALED COORD. =73.1"DHREES
TEST POINT VALUES
.6271E-02 .1092E-02 .9496-02 .ZZ56E-OZ
TEST POINT SUM OF SQUARES = .Z .. 79E-06
VALUES VIA REGRESSION
1
.6271E-02
2
.109ZE-OZ
3
.9496E-02
,.
.2256E-02
LA"BDA = .100E-OZ SUM Of SQUARES AFTER REGRESSION =
ITERATION NO. 4
.79Z,.E-04 ANGLE IN SCALE') COORD. =63.11DEGREES
TEST POINT PARANETER VALUES
.7951E-IZ .6345E-02 -.31.0[-02
DETERI'INANT = .3009E-03 ANGLE IN SCALED COORD. =62.97J)HREES
TEST POINT PARAMETER VALUES
.6719E-02 .63ZZf-03 .8665E-OZ .7854E-03
TEST POINT SUM OF SQUARES = .Z465E-06
VALUES VIA REGRESSION
1
.6119E-02
2
.6322E-03
3
.8665E-OZ
It
.785"E-03
LAMBDA = .100E-OZ SUN OF SQUARES AFTER REGRESSION = .Z46,. ;99E-OE
ITERATION NO. 5
DETERMINANT .1181- 03 ANGLE IN SCALED COORD.
TEST POINT PARAMETER VALUES
.1116[-12 -.39ItDE-tIl .7023-02 -.5751tE-OZ
DETERMINANT a .3511-03 ANGLE IN SCALED COORD. =&6.090ErREES
290
TEST POINT VALUES
.7053-02 .Z562E-03 .8154E-OZ
DETERMINANT = .2921-OZ ANGLE IN SCALED COORD. =71.11DEGREES
TEST POINT VALUES
.6761[-02 .5597E-03 .5651E-OZ -.Q406E-04
DETERMIN,ANT a .5690-01 ANGLE IN SCALED COORD. =47.Z10EGREES
TEST POINT VALUES
.6116E-02 .6306E-03 .8682E-02
TEST POINT SUM OF SQUARES = .2360E-06
PARAMETER VALUES VIA REGRESSION
1
.6716E-OZ
2
.6306E-03
3
.868ZE-OZ
4
.Z925E-03
SU" OF SQUARES AFTER = .2360 175E-06
TION NO. 6
OETER"INANT = .3133E-02 ANGLE IN SCALED COORD. =64.380EGREES
TEST POINT PAPAMETER VALUES
.6769E-02 .5817E-03 .5640E-02 -.1362E-g3
TEST PCINT SUM OF SQUARES = .2342 E-06
PARAMETER VALUES VIA REGRESSION
2
.5817-03
3 '+
.5640E-OZ -.1362E-03
LAMBDA = .100E-Ol SUM OF 5QUARES AFTER REGRESSION = .2 3lt1 S5ltE-O 6
ITERATION NO. 7
OETERI'INANT = .4533- 03 ANGLE IN SCALED CCORD. =67.590EGREES
TEST POINT VALUES
.7a77E-02 .2572E-03 .8075E-02 -.2275E-02
TEST POINT SUH OF SQUARES =
DETERMINANT = .3293E-[2
TEST PCINT PARAMETER VALUES
.2628 E-06
ANGLE IN SCALEO CeDRO. =66.720EGREES
.&516-02 .533 3E-0 3 .6560E-D2 -.48"5 E-03
TEST POINT SUH OF SQUARES = .2327E-06
VALUES VIA REGRESSION
291
1
.6816E-02
2
.S333E-03
3 ..
.8560E-02 -."8'5E-03
SUM OF SQUARES AFTE" REGRESSION = 2321136E-0
ITERATICN NO. 1
DETERf'INANT = ANGLE IN SCALED COORD. =63.541)E(REES
TEST POINT PARAHETER VALUES
.7095E-02 .2362E-03 .1038E-02 -.2630E-02
.2613E-06 TEST POINT SUM OF =
DETER"INANT = .3"02E-02 ANGLE IN SCALED COORD.
TEST POINT VALUES
.6860E-02 .4871E-03 .8482E-02 -.1414E-03
TEST POINT SUM OF SQUAPES = .Z313E-06
PARAHETER VlLUES VIA REGRESS[ON
1
.6860E- 02
2
.4811E-03
3 4
.8482E-02 -.8"14E-03
LAMBDA = .100E-01 SUM OF SQUARES AFTER = .2313454E-0
ITERATION NO. 9
OETERf'INA NT = .5299E-03 ANGLE IN SCALED COORD. =65.260ECREES
TEST POINT PARAMETER VALUES
.7110E- 02 .2191E-03 .8001E-02 -.2910 E-02
.2519E-06 TEST POINT SUM OF SOUARES =
DETER"INANT = .3519E-02 ANGLE IN SCALED COORD. =64.51DEGREES
TEST POINT VALUES
.6901E-02 .41t37E-03 .8 .. 01E-02 -.1210E-02
TEST PCINT SUM OF SQUARES = .2301-06
PARAMETER VALUES VIA REGRESSION
3 ..
8 .. 07E-02 -.1210E-02
LAMBDA = .100E-Ol SUM OF SQUARES AFTER REGRESSION = .2300E51E-OE
292
NO. 10
= ANGLE IN SCALED COORD. =67.090E(REES
TEST POINT PARAMETER VALUES
.71Z0E-02 .2075E-03 .7983-02 -.3289E-02
.Z528-06 TEST POINT SUt04 Of SQUARES =
DETERMINANT = ANGLE IN SCALED COORD.
TEST POINT PARAMETER VALUES
.&938E-OZ .,035E-03 .8337E-02 -.1587E-02
TEST POINT SUM OF SQUlRES = .Z289E-06
PARAMETER VALUES VIA REGRESSION
2
.'035E-03
3
.8337E-02 -.1587E-02
SUM Of SQUARES AfTER REGRESSION =
ITERATION NO. 11
NT = ANGLE IN SCALED COORD.
TEST POINT PARAHETER VALUES
.1990E-03 -.3576E-02
.Z'+6'+E-06 TEST POINT SUM Of SQUARES =
DETERMINANT = .3791E-02 ANGLE IN SCALED COORD. =67.8Z0EGREES
TEST POINT VALUES
.6972E-02 .3669-03 .8272E-02 -.1965E-02
TEST PCINT Of SQUARES =
PARAMETER VALUES VIA REGRESSION
1
.6972-02
2
.3669E-03
3 It
.8272E-02 -.1965E-02
SU" OF SQUARES AFTER REGRESSION = .2278730E-0
ITERATION NO. 12
DETERMINANT = ANGLE IN SCALED COORD. =69.550EGREES
TEST POINT PARA"ETEP. VALUES
.7131F-OZ .1935E-03 .7950E-OZ -.3827E-02
TEST POINT SUM OF SQUARES =
DETERMINANT =
.Z .. 00E-06
ANGLE IN SCALED COORD. =68.83DEGREES
293
TEST POINT PARAMFTER VALUES
.7DD2E-02 .3344E-03 .8213E-02
TEST POINT SUM OF SQUARES = .2270E-0&
PARAMETER VALUES VIA REGRESSION
1
.7."2E-02
2
.3344E-03
3 t.
.S213E-02 -.2335E-02
294
LAMBDA = .100E-01 SU,", OF SQUAR E S AFT ER REGRE SSION = .2270D04E-06
ITERATION NO. 13
OETER"INANT = .7550E-03 ANGLE IN SCALED COORD. =7D.OSOEGREES
TEST POINT VALUES
.7133E-02 .190t.E-03 .7940E-02 -.,03S-02
.23t.4E-06 TEST POINT SUM OF SQUARES =
= ANGLE IN SCALED COORD. :69.36DEGREES
TEST POINT PARA .. ETER VALUES
.702SE-02 .3061E-03 .e160E-02 -.2689E-02
TEST POINT SUM OF SQUARES = .2263E-06
VALUES VIA REGRESSION
1
.7028E-02
2
.3061E-03
3 t.
.&160E-02 -.2689E-02
SUM OF SQUARES AFTER REGRESSION = .2262!!93E-OE
ITERATION NO. 1 t.
.8229E-03 ANGLE IN SCALED COORD. =70.09DEGREES
TEST POINT PARAMETER VALUES
.18e7E-03 .7933E-02 -.42[8E-02
TEST POINT SUM OF SQUARES =
= .4267E-02
TEST PCINT PARAMETER VALUES
.2302-06
ANG LE IN SCALED COORD. =69.39DEtREES
.7050E-02 .2819E-03 .S11'E-02 -.3018E-02
TEST POINT SUM OF SQUARES = .2257-06
PARAMETER VAlUES VIA REGRESSION
1
.7058E-12
2
.2819E-03
3 It
.811ItE-82 -.30.E-D2
SUH OF SQUARES AFTER REGRESSION = 2257:!52E-OE
ITERATION NO. 15
DETERf'INA NT = .8915E-03 ANGLE IN COORD.
TEST PCINT PARAMETER VALUES
.1879E-03 .7928E-02
2275E-06 TEST POINT SU" OF SQUARES =
DETERf'INA NT = 02 ANGLE IN SCALED tOOPD. =68.860EGREES
TEST POINT PARAMETER VALUES
.7068E-02 .2617E-03 -.3315E-02
TEST POINT SUM OF SQUARES = .2253 E-06
PARAMETER vaLUES VIA REGRESSION
1
.7068-02
2
.2617E-03
3 It
.807"E-02 -.3315E-02
SUH OF SQUARES AFTER REGRESSION = .2253223[-06
295
FINAL FUNCTION VALUES
.7330-02
.""91E-02
-.216 .. E-02
.1577-12
RESIDUALS
.6375-02
.4230E-02
.2083E-02
.1"0-02
.60&7E-02
.3161E-02
.1867E-02
.1It00-02
.538"-02
.2533E-02
.1802E-02
.1097E-02
.506 7E- 02
.2"33-02
.1629E-02
.931 .. E-0.. -.2211-03 -.1252E-03 .1156-03 .4888E-04
.1241E-03 .2312E-03 -.2150-03 -.98C9E-0" -.2539-0"
.3689-04 .2129-04
X PRIHE-X MATRIX
1
1 1t.7290
2 8.8111
3 -1.5559
.. -.1.1 07
CORRELATION "ATRIX
1
1 1.0000
2
3
..
-.9808
-.9197
-.9811
NORMALIlING ELEMENTS
2 3
31.3 DOS
-5.5953
-.8885
2
1.0000
.9712
.9890
1
.3 .... 8E.01
2
.3808E+01
3
776".01
1.1320
.1It22
It
1.0001]
..
.5338E+02
.0367-
1.0000
296
VARIANCE OF RESIDUALS = .1502E-07. 15 DEGREES OF FREEDOH
INDIVICUAl CONFIDENCE LIMITS FOR EACH PARAMETER (ON LINEAl< HYPOTHESIS)
1
.79111E-82
.6223(-02
2
1195E-02
-.6716E-03
3
.9977E-02
.6171E-02
..
.9769E-02
-.16"OE-Ol
APPROXIMATE CONFIDENCE LIMITS FOR EACH FUNCTION
.7525f-02 .6"89-02 .6189[-02 .5 .. 81-02
.7135E-02 .6261[-OZ .5985E-02 .52e8E-02
.,,51)'>E-02 ."336E-02 .3253E-02 .Z623E-02
.,,38EE-02 ."IZ5E-02 .3069[-02 .2""2E-02
.2Z68E- 02 .2192f-OZ .1988E-02 .19281';-02
20e 0[- OZ .1975E-OZ .1, .. 5E-02 .1676E- 0 Z
.1716E-02 .1511"E-02 .15 .... -OZ .128'>-02
.1439E-02 .1296-02 .1255-D2 .9093E-03
END OF PROBLEM NO. 1
VALUE
.5166E-02
... 967E-02
.25Z6E-02
.2340E-OZ
.17&5-02
.14Q3E-OZ
297
X"IN" O. XHAl(= 11ZOOOEtQl. YHIN= .2&1705E-03 YHAX= POINTS PLOTTED.
x x x x 'C x x x x X )I X X X X lC X J(
.76SE-IZ=Y Y=
.
S
S
.6"IE-Il:Y
S

HSE-OZ"Y
SItIE-Il",
.1t6SE-llaY
.!':IOE-OZa'
.11SE-IZ",
.lltIE-OZaY
16SE-OZ.,
.9aOE-OJ",
S
41,S
,
S
S,
s

S
I
I ,

,
I
S
I
I

,
S

,



t ... t
.t'5IE-13aY
,
S
, S
S
IS

S

, , S
t
S II
. ..
S S
UIS

, S S

, s s ,
S US
.. ..
S SUS S
.. .. s ,
.. .. ..
.. ..
t
t. t ....
t t t tt ..

x x x X x x x X X X )I X X X X X X X X X
Y=
Y=
Ya
Ya
Y=
Yz
Ya
ya
Y=
Yz
0.000 .333 .66& .qqq 1.H2 1.&61j 1.qq" 2.331 2.1)64 Z.qQ7 3.330
7&IjE - 02
.1)90E-OZ
615E-OZ
.St.OE-OZ
1t6SE-OZ
3<J8E- az
.315E-OZ
.ZltOE-OZ
.tr.SE-az
.900E-13
.tSOE-a!
TIMES .tEt03 N
U)
00
APPENDIX B.3
Results of Three-Parameter Estimation Model (Tables 1 to 4) and the
Plots of Estimated MeN Values w t ~ their 95% Confidence Intervals
(UWHADS Estimation Based on the Linear Hypothesis)
- Parameters estimated: MeN, k", k
1

- [TCN]o Equal Initial Measured Value that Changes in Time.
299
DILl L3.1. 'IBII PAIMITIIl ISTDIAftOR IIOIJIL (MCW. t.. It.) - COPPER ctAIfIDI CQlPUX
20C 20C 4C 4C 20C 20C 4C 4C
It -It AlR NO AIR AIR NO AIR AIR NO AIR AIR NO AIR

RaeN NO UV NO UV NO UV NO UV UV UV UV UV
G It., (c_{b) 1.65 1.00 0.31 0.21 1.63 1.06 0.98 0.58
I (b- ) (.0394) (.0238) (.0074) (.0050) (.0389) ( .0253) (.0234) (.0138)
CIrCII
,

! .,1.'L) 0.13 11: 10-
2
I 1tcu (b-
1
) .0207 .0071 .0037 .0033 .0153 .0024 .0036 .0031
t.
UlfHAUS .0635 .0154 .0187 .0829 .1049 .0153 .0148 .2259
MeN 11: 102
UlfHAUS .092 5.67 .237 .523 .146 .052 .066 .451
1tcu (b-
1
)
UWHAUS 88dO-
2
.94 .2611:10-
2
.32dO-
2
.7211:10-
2
.2911:10-
7
53d0-2
TeN 11: 102
(1IOIe/L) .761 .796 .781 .759 .777 .758 .750 .785
MeN/TeN .12 NA .30 .69 .19 .07 .09 .57
TSS 11: 10
6
55.9 59.3 . 58.1
38.0 58.9 81.5 74.1 50.2
ISS 11: 10
6
1.0 1.2 1.1 .43 .65 1.6 1.8 .72
R2 (%)
98 98 98 99 99 98 97 98
NA - not acceptable
w
0
0
UJtIl L 3. 2 !IIUI PAUIIIt'a IISTDIATIOIf IIOIJKL (Nell. t.. Ita) - lIRe CtAIIIDI COIPUX
20C 20C 4C 4C 20C 20C 4C 4C
It -It AIR NO AIR AIR NO AIR AIR NO AIR AIR NO AIR
.,
NaClf NO UV NO UV NO UV NO UV UV UV UV UV
G ltv (c_(h) 1.65 1.00 0.31 0.21 1.63 1.06 0.98 0.58
I (h- ) (.0394) (.0238) (.0074) (.0050) (.0389) (.0253) (.0234) (.0138)

y

E 1101 L) 0.13 x 10-
2
N ItZ
n
(h-
I
) .0062 .0194 .0068 .0045 .0448 .0181 .0137 .0095
.8323 .0195 .0194 .0172 .0770 .0257 .0314 .0613
MeR
tNRAUS 53x10-
2
11 xl 0-
2
.15x10-
2
.4OxI0-
2
95x10-
3
87x10-
1
.IOxIO-
1
.38x10-2
ItZ
n
UWHAUS .0559 .0612 .0039 .0043 .0181 1.10 .0698 .0096
TCN x 102
(lIOle/L) .714 .648 .750 .631 .754 .769 .758 .731
MeR/TeN .74 .17 .20 .63 .12 NA NA .52
TSSx 10
6
67.3 97.8 62.9 43.7 50.2 100.9 106.5 69.5
ISS x 106 .03 .72 .72 .42 .16 .80 1.3 1.2
12 (%) 99 99 98 99 99 99 98 98
KA - not acceptable
w
0
'fAlUl II. 3. 3. ftUI PAItAIII!U BS'fDIA'fIOII JIOIII!L (MCR. kyo It,) - IfICDL C'fAIfIDI!: COIPLU
20C 20C 4C 4C 20C 20C 4C 4C
It -It AIR NO AlIt AIR NO AIR AIR NO AIR AIR NO AIR
"
MaCR NO UV NO UV NO UV NO UV UV UV UV UV
G It" (c_(h) 1.65 1.00 0.31 0.21 1.63 1.06 . 0.98 0.58
I (h- ) (.0394) (.0238) (.0074) (.0050) (.0389) (.0253) (.0234) (.0138)
.-at
y
[MCNl
o
I
0.13 x 10-
2
"
ItN! (h- ) .0027 .0008 .0030 .0027 .0008 .0015 .0042 .0041
.0382 .0972 .0395 .0304 .0873 .0204 .0139 .0581
MCN It 102
UWHAUS .419 .103 .322 .345 .129 .153 .083 .239
It,,!
UWHAUS .615xlO-
2
.232xI0-
2
.179xlO-
2
.116x10-
2
.8OxI0-
3
.87xI0-9 .36x10-
2
TCN It 102
(1I01/L) .882 .765 .796 .652 .765 .761 .769 .677
MCNITeN .47 .13 .40 .53 .17 .20 .11 .35
TSS It 10
6
106.9 82.4 38.2 28.6 38.2 59.3 69.7 39.1
RSS x 106 .49 .37 .32 1.2 .07 1.0 1.2 1.6
R2 (%) 99 99 99 96 99 98 98 96
NA - not acceptable
w
0
N
uatB 8.3.4- 'IBD PAUIIE'l'D ISTDIAnOlf IIOIJIL (Mew. kyo It.) - DOIf C!AJ([DI PIlOCKSS
20C 20C .oC 4C 20C 20C 4C .C
It -It AIR NO AIR AIR NO AIR AIR NO AIR AIR NO AIR

ReClt NO UV NO VV NO UV NO UV VV VV VV VV
C 1ty (c_(h) 1.65 1.00 0.31 0.21 1.63 1.06 0.98 0.58
I (h- ) (.0394) (.0238) (.0074) (.0050) (.0389) (.0253) (.0234) (.0138)
J'e-GI
,

I 1101 L) 0.13 x 10-
2
I kPe (h-
1
) .0016 .0026 .0012 .0007 .0068 .0222 .0051 .0039
.0123 .0194 .0235 .0438 .1098 .0391 .0099 .0111
MCIt x 102
tMlAUS .094 .185 .118 .129 .113 .233 .123 .289
kPe
48dO-'J 13dO-
2
63d0-8 .51xl0-
3
.4911:10-
2
.11xl0-
1
.24xl0-2 .3211:10-2 UWHAUS
TCN x 102
(.al/L) .773 .754 .71.9 .539 .704 .704 .742 .704
MClt/TeI .12 .24 .16 .24 .16 .33 .17 .41
TSS x 10' 67.4 67.3 38.2 20.1 41.0 65.4 69.5 51.3
RSS x 10' 1.3 .28 .86 .21 .02 .42 .26 .64
12 (%) 98 99 98 .99 99 99 99 99
IA - not acceptable
w
0
w
304
IS
5
I

S4

-
i
3

-

"2
S
i
..

.,.1
-
x

a
l
1
....
2 3 4 5 IS 7
8 TEST NJMBER

-
....
13-1
-2

Figure 8.3.1 three-Paraaeter FBtiaation Iblel - Eatiaated HCR values
with 95% Confidence Intervals - Za
305
e
!4
~
I
z
-
2
3
-
~
M2
m
oj
I
..

I
'\a 1
-
x

~
1 2 3 5 e 7
8 TEST NJMBER
~
-

13 -1
Figure 8.3.2 'lhree-Paraaeter FBtUation 1bde1 - Estlaatecl Hell Value.
with 95% Confidence Interval. - RI.
e
5
I-
-
~
-
~
... 2
m
~
..

.,.1
-x
~
fill
...
i
-
...
1:1-1
-2
306
I
I
I 2 3 5 7
8 TEST tu49ER
Figure 8.3.3 1.'hree-Paraaeter Irstilution tbcIel - Estiluted JEll Value.
with 95% Confidence Intervals - Fe
307
APPENDIX 8.4
Exa.ple Prograa for Two-Parameter Estimation Model
- HCN, k1-parameters to be estimated.
- (TCNl
o
Equal Initial Measured Value which Changes in Ti.e.
- [FCN] - [TCN] - [HCN].
TS TOlZ 71'171opr:1 4."
r.
C

c
100
30
35
ZOO
.. 0
45
50
TSTOlt' (INPUT, OUTPUT, TAPE!j:I NPUT,TAPE6:0UTPUT.
+or.,/Jr,=OUTPUr TAPO it
-oIH""NSIoN V ,oIFFfJ' fl) !SCRI\Tff;OIJO)
COttMON I!II )((t'5 -,VP(7,&U.0LT,N'STP,fT(61) ,NJ1,YINrT
'WO PAQAItE" fSrI"A'ION HOO:-l
PH" IfNAl HcnEl
OAT4 NP,HIT/2,15/
OATA NO'l/1101
oArA NPROB/lI
OATA EPSl,fPS?FlA",FNU/l.E-&,1.E-6,0.01,10.1
RAO (5,.) fTH(rJ ,r-=I,NP)
WRlTF(f>
1
100) (fH(I),I:1,NP)
FOP MAT (J F 1 Z 7J
REAO(5,) (l(UrY(O,I=l,NOR.
"'RUEff".' 'XU ,YfI.,I=I,NOfl)
YAV: R.
00 3IJ 1=1, N03
yAV=UVtY( It
YAV=VAV/FlaATeNORJ
TS<;= I).
00 '=1, t..0!J
TSS= TSS+CY' It -YAV)"Z
T'iOS= TSSIf <'flO 00." Z)

FORttATU)("TOTAl SUit OF SQUARES IStEl".7)
00 "0 J=1 J tWII
Vl=YeJlnbOOO.
Y(J) =Yl
CALL PlOTPTe" (JI ,YI,')
CONTINUE
YIN IT-= YClI
DEL T=Z.
NTSTEP=X(N09)/OElT
0045I:l,NP
llII'F (I) =0
1
01
sIGNS(U= .0
CALL UWHAUS(NPROA,HODfl,N08,Y,NP,TH,oIFF,SICNS,EPS1,EPSZ,ttIT,
+FlAIt,FNU,SCRAT
CALL "OOEl
00 50 J=1, Nfl
T:TT'J)
CAll PlOTPT (T,YP(I,J),4)
CAll PLorp, e"Ypez,J',Z)
CAll PLorPT
CONTINUE
CAll OUTPl,
STOP
ENO
w
o
00
l
;
U
H
20
25
30
l5
.. 0
.. 5
50
iJARQUTINF HO[lfL 7311 71 OPT= 1 rTN ".8+538
SUOROU TINE HOOfL TH,F, Non, NPt
nrHn4sIor, TH.? rFf ?5t
COMMON
P[AL YY(J"CIZIt',W 3,q"XX,ToL,XENO
COMMON IAI RI(C, VK, UVKN .
FCN1

C
C
C
C
300
55
ItO 0
)( x= 0 .0
ITHA X= 1
IFINTSTFP.GT.60IITHAX=INTSTEP/60.'+1
yy ( ?t = TH ,
YY'lJ =YIN IT
YV. l' ='tY( J t -YVe 21
Fet '=YlNIT
CONSTANT OfLr MUST RE LESS THAN OR EQUAL TO
TIM OlTWEEN oDSERVATIONS
T= o.
JOIlS=?
VK=.0138080
IlKC= TH (11
AS=.Olt7711
fTC U=T
v-"
ypH ,1 '=YV H'
VpCZ,1t=yy(zt
YPCl 11t=YV (l'
I
Ol:O.(H!01
NO:1
00 K=l,NTrTEP
xum=FLoATiK"I)ELI
CAll CN,FCU
IF GOTO ltoO
T=T+nrlT
JPHE 11= (XC JOflst InEl TJ
IFCJPRfO.EU.K'F(J01st=YYIJI

IFe CK/ITHAXt 'IrMAlO .EQ.KIGo TO JOO
GO 10 55
KK= K/ITHAlO
HC I(I(.lt = T

Y"(Z,I(I(.lt =YV(?t
YPCltKK+1t=YYlll
C ONT NIJ[
CONT INtJE
NT1=KK+1

END
w
o
I..D
1
,
1
i
11
11
Fr.N1 ('49 xx, YVl'PRIMEI
COMMON IAl RKC. VI(, UVI(N' S,V
P.rA L " IN) Y FR I HE (N" XlI
V.,.' Ut = e "I(C.yveltl-1 vIC.ne 11 I
'1P"'IHEIZt .nl?t .CRKCI
YfIR I HE UI VI(.YY (11

END
SUAQOUTtNE UMHaUSI .NOe,Y ,NP ,TH, oIff ,SIGHS,EPS1,EPSZ,
1 "If, FLA"! FNU. SCRATtLLt
DIMENSION seRlTU'
U=l
Ill: U.NP
. IC: III.tlP
n=lC.NP
IFF=InHIP
I :lHW'
IG= IF. NOt)
IH= IG.NO]
It IH NP NOB
I... H
C-Ll "ODfLtNOB. 'I. N"l tH.Dlff.S IGNStEPS!, EPs2 ,"IT
1 .FUHrFNu.sr.p'l'fatr scpal fbl
s
"ICt, $C'UTe Dt,
2 sCRa, lEI. T IF , SCR r Iii , seRAr IH , SCRA" II ,
3 SC HIT IIJ I ,L .. I
PrrUHN
END
"'"
o
1
5
11
15
21
ts
31
35
...
Ie'
51
55
CI
c
r
C

C
15
16
19
5,.
SltlIPOUl tNE HAUS9C NfIIt'lO, "OUEl, H:'O, "N!b TH, olF l ,S IG NS ,EP IS ,EPtS,
I"U
t
Fl ".FNU, 0IIl,f,PHI,Tn, ,R,',O,Ofllt.llJ
STOH:ESCG-, o,n "P, ,p, SDEV,FHU,O,R,F.: PS,[)ELzJ
OIHrNSION THINQI, nIFlCNOl, YCHOot
OhIOI, P CHilI, fiNO .. PHIl NO , TB INa)
DIHENSION FeNnol, RIN'101
o HENSION AIHO,Ne), Oh,Q'HOI, OELZCNBO,HUI
DIMeNSION THltl, OIHIlI
I
')IGN')llI, Yeti, QUI, rlU,
1 'p .. nu, TIJ U, FlU! Hftl, ACUI" DIU, onllt
ACOSIXI a 2 - 1.0 I
NP = NO
N""'On = NPQIJQ
NOR .. NIIO
[PS1' '" P1S
p<-;2 '" fPZS
Hf>!;r) = NP HP
= !J'NP. HPSI) .Z.HOe.Np.HOB
WPITEILl,lOOOI HPROO, NOH, HP, HSeRAe
MtUrElLl,lOlllt
CALL GAS'iOU, HP, TH, TEMP, '"EF.LLI
MRTTf="Ill, 100Z1
CAll GASSOfl
J
HP, OtFl, IFHP, TFHP,LLI

IfCFNU-l.O)'N, 9'), 16
Cmn IHur
no lq 1=1, NP
TEHP = ADSIOIFlll))
IFfA"INU1.D-TEItP, ."SITHIU.U", 9'), 19
C I'INTI NilE
tiA '" FUH
NIf = 1
lA05 = iI
IFIEPSU 5, ro, n
fPSl = a
SSQ .. 0
CAlL "OlELfHP"OIt, TH, F, NO!), HPI
DO qO I = I, NOll
Rill", Yell - FII)
Ell I,
.. S'iQ=ssa.tU [J'IH It


C
101
111
liA = GA I f-NU
I N\r: Nl = 0
MR IE LL,100"INII
JS = 1 - NOll
DO 130 J=l,NfI
= flit Ja
P(Jl=ol lIJI'THIJ)
THIJt= THIJI.pCJJ
Q CJ t=o
JS a JS Nell
REGIN ITERATION
\oN
60
,.,c;
70
"
'"
115
911
95
100
10;
110
c
111
c
c
120
130
131
160
151
lU
153
C
155
C
C
zu
WRlTFI6
I
tll'J
FOHtliTI H ,"1'1',15'
CAll ,.C:-IF LI NI'PO:l, TH, DEl ZI JS I, NO", NPI
[J =
no 170 I = 1,
rJ::IJ.l
OFllflJ' = IJCl1UJ' - FlU
I)(JI = iHJ' OlZIIJ' RI II
Ill ... " 'lIJ"PIJ'
THe J' :: T r
HILAO';' t
()1 1",,) I = 1, NP
110 l'il J=I,1
:iUH ;:. 'i
1(..1 =
1(1:: ,wl-ft-l'
nn I ;"1 I( = I, rfO:1
",r = 1( 1
I(J = I(J 1
SUH : SU .. + 'l[llfl(n DElze KJI
rrMP: SUH/(p(r).peJ
.II = J NI"( 1-0
OIJ() :: TF" P'P
IJ :: I NP.(J-l'
o ( r J I :! r F.: HP
EIU =

no 153 t:: 1, NP
I J = I -NP
-00 IIjJ J=I.t
IJ = IJ NP
AfIJ) :: OfIJ' I (E'U.E(J))
..II = J NP-II-1)
AIJI) :: I"IJI
II = - NP
00 155 1= 1 ,NP
p(lJ=O( I/flIt
PHIl IJ =PI n
II :: NP + 1 It
AUII = AI II' GA
1-::1
CAll MATINU, NP, p, I, DET)
STfP=1.0
SUH1=0.
= O.
SIJHJ=O.
DO Z II I:: 1 ,NP


su,d= PHI( T) PHIIII + SUM]
PHI(U : PII.
TEMP =
'27XT'" ISTErPE'iT J('iCt=:NrJ
-= SCALED HOMENT MATRIX
PIE : CoRRECTION VECTOR
W
N
IH
1211
12;
1311
t.15
litO
Ilt5
1'5!)
155
1&0
I&S
170
TEMP = 1.0)
TEMP = 57 ?CJ5'ACoSfTE14PJ
WRITfCLL,iOItU OCT, TEHP
170 00 220 I = 1, NP
p n J = PH I fIt. S TF. P , E e IJ
nUl) = THCU p(1I
221 COt! TI NUE
wRIrr fLL, 7000)
7000 POINT PARAMETER
WRITF. LL,200f)) (TAU, 1= 1, NPJ
no 27.1 I = 1, NP
IFCSIGNSCI JJ 2211 2211 222 .
222 IFfSIGNI1.0,TH(I 'SIGNC1.o,Tocun E63, 221, 2Z1
221 CONTI NUE
SUl1fJ=O
CALL HO:)ELlNPPOB, TO, F, NOR, NPJ
00 ZJO I=l,NOR
RCIJ=YfIJ-F(I)
230 SUMII=SU'IhRCIJ"UIJ

IF'SlJl1tl - ft.Oi'EPSU'SSa) 66?, 662, 661
663 If( A,HNl CTCMP-H.O, GAf) 6&5, 665, 6& ..
665
= INTCNT .. 1
IF. - 36J 170, 2700
66 .. GA::I;A.FNU
INTCNT:: INTCNT 1
IF' INTeNT - Jf,J 660, 2700, 2700
662
DO 6f,Q 1=1,IP
669 TH. JJ=rrlf It
CALL GASSO(t, NP, TH, TEMP. TE'fP,LLJ
. GA, C;Ul1fJ
IFlf'Ps:?J Z;><J.('Z'),22C;
2Z1 2/0,(,70,265
22 5 0 0 7 41) I :: 1, N P
IF.A'>S',JCIJI1f1
l
E-7fhAilSITH([lJJ-EPS;>1 21t0, 21t0, 21t1
Zlt1 270,2711,2&5
2ltO rOtn hw'
W'Urt CLlo100'Hf P ')l
GO TO ;>I\!J
265 - sse) - EPS1'SSQJ 26&, 26f, 270
2&6 WRITFlLL
1
1fJ1DJ [PSl
GO TO l/tl)
2 70 SUM'I
Ntr=NI H1
IFPHT - HIT) 1011, 100, 280 .
2700 WPITrCLL,Z7101
2710 !'"OR",Ar (II llt;HO .... THE SUM OF CANNOT OF Rfntr.EO TO THf SUM
10F SQUARES AT THE END OF THE LAST ITERATION - iTERATING STOPS IJ
r:
g END ITERATION
Z80 WRJTEfLL,ln11J
NRITEfLL,ZOOll IFeU, I = I. NOll.
WRIr((ll, 1 012.
WRITFfLL.ZOOU IRln, I = I. NO,)I
w
w
17;
1" 0
185
190
11)5
ZOJ
ZUi
110
?H
Ul
Z?5
WRlr[ Ill; 1 017.
1017 FOQ .. 4r (/1I1.t )( MATRIX"
CAll GASSOf4.NP.TE"P,TEHP,D,LL'
SSQ= SUMIJ
IOF=NO'l-NP

1=0
CALL HAlItHIl. HP, p, t. DEn
00 76'l2 =l,NP
II = I + NP.(I-U
7692 ElIt = SQRTfo'II"
00 l40 1=I,NP
JI = I + NP.fl-l' - 1
IJ = 1 + NP.(J-2'
00 3\0 J = I. HP
JI = .II + 1
AfJIt = O'JrJ I 'En"rCJU
IJ = IJ + NP
3\0 Aft JI = A ( JU
CALL GASSOU, HP, TEHP, TEHP, A,LU
101&.
CALL GASsB(l, HP, E. TE"P, TEMP,lLI
tFUOFI 341, 410, 141
lit 1 = SSQ I IOF
WRITEfU
t
tOl ... SOEV, toF
SUEY = s'lPJ(SOEVI
391
1= I .NP
P'IJ=THfIJ+? O.E,r'.SOFV

Cll.l OJ')
r.
c
CALL GASSllC2, HP, TB, p, TE"P,Ll)
_ LAOS = 1
GO TO 101
Itl. 00 : 1, Nol
TEMP = oJ
00 4?O 1=I.NP
00 It?O J=I,NP
Isuf1 :: K+NO:I'fI-lI
Of our.[ = ntLl( IS'J'1'
o(ruGl : O[llfK +
IS(III =
O[I.UftZ = orL7fIsuO,
OflllJ It? = 0, II (I( ,.01.' J-1I ,
IJ " t +
Of-',ur.3 = nfTJIIfIHfZff'.TH(IhOIF7'J'9TH(J"
"ZO Tn"" = rrHfI + OC'lur.t orHuG2 or:'WGJ
Tr"''''' = ? ,,9'; 'lP r IT t-"'" , 9 SlIfli
PCI( J;f '1(1+ 1'"'11"

WqUf Cll.I 011"1
fr: ,
JI)

fret/on-I(' 4.1t.l.4'J'i.103r;
It.f 11 f r:. :,1)'1
101'0 (oICJI, J : f, H'
It:> w f II " I) I)., , (r e J', J = (, (tJ
lei 3 (ll,IOB' "41
1P
I)'1
VJ
J;:-
Zll
1li
,-.
z .. s
251
255
261
lIe Pl"N
'" WIU T d II tl OJ ...
IiO TO ItlJ
1100 FOfI"A, n"HtNON-LlN!:n ESft"ArION. PRORLF. .. NUf'P.t,q 11 .. 11 IS,
t tltH I'i, 11H (1, 11" sCH , .. H REQUiRED.
tOOt FORMATCn'iHOINlfUl V_llJF:S
11) Oz F()IoIHA fe/'>4/1 OpoOPORTIOtfS USF D It! CAl CuL AlING OIFF[ RENer QUOTIENTS
1001 FOthtATflZ'>HOINIrUL 'iIlH Of SO [12."J
too" FOPHATflll"ltr;x,tJHtTERATJON NO, HI
t007 VALulS VI RrGQrSSION
1001 FoPHATCIII'5ItHOAPPROXIHAJE CONFIOENCEoLIHIrs FOR EACH FUNCTION VAL
lur I
uoq FORHAT C/!iZ HOlTEAA TlON STOPS - RELAT IVE CHANGE IN EACH pARAMETER L[
ISS THAll Ell .. 1
lOU FOPHAl":'2HOIhRATlON STOPS - PELUIVE CHANGE IN SU" OF SQUAlfES LE
ISS TH N r12.41
1011 FQI'HATC2lHtFlNAl VALurs
10tZ I U l. I
101ft FOR HAT liNBr. 0 R{S IDUALS. ,E12.'.lH, I .. ,
ll0H OF
ton HATPIX
10 16 ,II 11Z1HO HAL 11 INI;
103' OF PROBLEH NO.3.
tOl .. F()/of .. J
1039 LIMITS FOR EACH PARAHETER CON LI
IN[AQ HYPOTHEStS'
11 .. 0 SlO.!, 1tIC,.)3HSUH of SQUARES AFTER REGRESSION.
lEl';.1t
11'1 FQPHATUItH HZ.'" U, 2SH ANGLE IN SCALED COORO
1 fCj 21 IIHr.ECP,-ES
10 .. 3 f()fo!H" ?IIHOtE'S POIN' SUH OF SOU_RES. E12 ....
2001 FOl)ttATCl5[lZ."
2016 fO
Q
"ATCSEl Z."
END
\AI
V1
t
5'
10
15
20
0'5
30
35
r.
c
c
350
370
C
r.
C
.111
.. 50
500
550
MATHHAl NVAR, r:.I NA,
DIMENSION ), I1(NVAI<, 1
pIVOT" = Ali. 11
ntT=10
00 550 = 1.
PIVOT = AfICOl, reOl)
PIvor" = AHIN1IpTVOT, plVorH'
= PIvor nET
oIvrOF.: PIVIJT ROW BY PIVOT ELfH[NT
AI[COL. ICOl) : t.D
PIVOT = A"AX1IPtVOT. 1.E-20)
pIVOT = A(ICOl, lCOl"PIVOT
DO JSO l=1,NVAP.
l' = AIICOl, U-PIVOT
IFINO 0) GO TO 371
DO 370 l= Nil
SIICOl, II = SCICOl, L)'PIVOT
REDUCE HON-PIVOT ROWS
DO 550 ll=l.NVAQ
IFIll .Ea. ICOl) GO TO 550
T :: ACll. ICOLt
Alll. IcoLl = O.
OF'TJ
('10 450 l=l.NVAR
Alll. LJ = A (L 1. U - A He Ol, l ,. T
rFf NA .EQ. 0. GO TO 550
DO 500 l= 1 ,Nil
8elt, L) = 8fll, Ll-fHICOl,U-T
CONT {NUE
RETURN
ENU
W
0'\
1
.;
til
15
20
10
30
"0
15 60.
80
90
20
95
100
25
500

20
III
NQ, A, B, C,Llt
o HENSION NQ,NOJ
til' = NQ
= 0
lOl4 = 1
= 10
IF NR )15,20,311
RFTflftN

I (lOW .GT. lUI" RETIJQN
WRIrFfll,500' IJ,J=lOIf,lIJPJ'
GO ro 80,e" ,ITYP
WRIrEIll, oof IA J ,J=LOlftLUPJ
GO TO 100
WRUE Ill, 6001 IBI.II ,J=lON,lUPJ
GO 41)
00 II l::lOW lUP
WRITfILl,726'I,fCCJ,tt,J=LON,IJ
L =l UP .. 1
r CLOW? .GT? GO TO 100
'l5 [=lOH .,NP
W IrEIll,7l0)I,ICIJ,It,J=lOW,lUpJ
lOW = lOW .. 10
lUP = lUP .. 10
1
Go 10 10
FORM TI/5X,J7,9(5K,17)t
4'
FO"'MAJ (1 0 ,h,9IlX ,Ftl."U
FNO
VJ
-...s
APPENDIX H.5
1.'bree-Paraaeter FBtimation Ibde1 (FClf, ltv, ~ with
the Flxed. HCN/l'CN Ratio
- Exaaple Prograa.
- Tables of Rstimated. Results.
- Plots of RSS vs KCN/l'CN Ratios.
318
1
:;
1!)
H
20
?;
30
H
itO
1t5
I)!)
r.

OIM[NSION 'SC
P
ATC60001
COMMON 111 X(251,YP(7,611,OELT,NTSTEP,TT(61I,Nfl
HcorL
C ................................................................ ..
r.
r.
E
THIS HAS PRESET RATIo
IT IS ESrIl'4 ArTNGtFQEE CN,VOL.COEFF. liN!: I1[TAL DECAY r:OEFF.
C. ..... ...... ................................................................ .
c
20&
113
114
115
12
1&
c
13
OATA NP,I1IT/3.151
nATA
OATA
DATA
RrtO(St 1 fTHI It rI::1fNP) .
WRITEI"',20t:-1 (TH II, =l,NPI
fOf'/iATnf12 n
READ(5,.1
(I),. I (X f I' ,y (n ,I=l ,Non)
YAV=O.
00 113 [=1,NOP
YIIV=YIIV+Y( II
YAV=YAy/FLOAT (N0I31
TSS= o.
no 111t t=I,NO'-'
TSS=TSS+(Y II-YAY) ".2

WRITF(6,11 1;1 TSOS
FOPHAT(lX,tTOTAL SUM OF SQUAPE5 IStEllt.71
00 12 J::l, NO!"!
Yl=Y( J1I26000.
Y(JJ =Y1
CALL PLOTPT(X(J).Y1,51
CONT tNtlE
OEL T=2.
NT:; TFP='<C NOq) IOEl T
00 it, I=I,NP
oIFF(JI=O 01
<-IGNS(I1=l.O
,NOB ,Y ,NP, TH ,'I IFF ,SIGNS,EPS 1,(PS2.l1lT,
.Fil
CALL 110D""L ('IPR1B,TH,F,NOR,NPI
00 J=i,NTl
T::TT (JI
CALL PLOTPT (T,Yp(1,JI,41
CALL PLOTPT (T,YP(2,JI,21
CALL PlOTPT (T,YP(3,JI,201
cmlT
CALL T
STOP
EN;1
w
U>
1
5
10
15
r.
C
lO
25
30
35
?1
itO
1t5

TINF H()r.[ U N TH ,F ,NOlh NP)
OtM[NSitlN THI3J JF(?5)
COMPiON IfJl )(f2<; ,YPf7,61t,OELT
1
NT<;fF.:P,TT(61JtNTl
R[AL YYIJ),CIllt.,W(3,Q),KX,lOL,KENO
C OM"IIJN 141 QI(C, V II, UVKN
FeNt
NW=3
N=3
X)(= 0.0
r THA K=l
If'( NTSTr:P
1
r.r.60) tTHAK=( NTSTEP/GO.) + 1
yy( 1. =Tlil )
Y Y I 2 J :: IJ 2 0 r; "T H ( 1
YYIl)=YY(l)+YY!?J
FU l::YY(IJ +VYIl'J
CON'iT4NT f)f LT liU';T BE Lr:3S TH4N f lU"l TO
TIME Ii::TwnN OIlSrQl/4T rON5
T=1.
J()i' '):7.
VI(= Tlil?'
RKC= TH (,I)
TTfl'::T
YPU,1)=YY ttl
YPI2,1'=YYfl'
YP(1t ll =YYIJ'
TaL = o. OOf) 1
I NO= 1
DO 10 1(=1,
XI' NO=FlOATiKJ "DEL T
C4Ll UVEWK (N,FCN1,l(l(,YY,XENO,TOl,(ND,C,NW,W,IERJ
IF GOTa 20
T=T+f)ELT
JPRE lJ = IX (J 01S" !J[l TJ
IF( JPP.f-fl. E G.K' F IJ01SJ =YY r:u

IF((CKlIrHIIX'.ITMAX).EQ.I(JGO To 21
GO TO to
KK= C K' TTMA X,
TTfI(I(.U=T
Y pI 1 ,K K.l ) :: YY I 1 J


CONT Nur
r. atl T TNU[
NTl =KK+1

END
/
\.f.J
N
o
l
5
5
10
15
SURQr:JUTINF FCNt (N.XX,VY,YPRIHEI
COMMON IAI
REAL
y pP I HI" (1) = (qI(C. YY f ?I 1 - ( VI( YY (lit
YpP I "E ( 2 1 YY ( ? 1 f - RI(C 1
IME (31 =- VKYV( t I

END
s U" I II) TT'I C UI44 I) ') ( ,.co E: L. NO,], Y ."1 P TH , D IFF. ') I G "S "c PC;i E P')?
t FLAil, HoHlt .
oIWNSTaN
111= f
IH= IA+NP
IC= IfhN'>
In=IC+NP
1=1 f'JHID

IG=IF+NO]
TH=
II = I H + NP NO"
I.J = IH
r:ALL HA US'H HDnfl! NO!). Y, NIl. TH. n IFF.$ IG N'i, [PSI, EPS2 ,HIT
1 ,FlArI,FNU,<;CQATfUIJ SCRATlldl
J

2 SCPArfIO. SeQAr fIF , SCRAn IG , SCRAT(JH" f( II,.
J SCRATfIJI,lll
RETURN
END
/
W
N
1
5
10
15
21
Z'i
30
35
ItO
.. 5
50
55
CS
C
r
c
8
C
r.
C
C
15
16
19
5
Sif,l"'OUTlNE HOIlEl.! N:'lO, 'f ,N(hTH,DIFI 'S (GNS ,fP IS ,EP2S,
IHIT tFUII1,FIW, (ltP.E,PHI.TI1.f,R,' ,D. Dfllt.llI
STOt<ES Iii'" O. T HP ,A ,p, SOE V,fNU.O, R,F PS, CEll I
OIl1rNSION TH(NQI, nIFZINOI, 5Ir.N'lCNQI, 'flNOOI
oII1FN'lION ()frlOI, PINQI, F(NQIt PHIlNQIt TelNQI
DIMENSION f(Nnol,
DIHENSION A(NO,N!:I, D(NQ,NOI, OElIINBO,NQI
TH(ll, OIFlUI, SIGN<;(U, ,((U,
1 . P Ii If it, T'1 f1 I, f U I, R (l I, A ( It ,. 0 f1 I ,
ACOS(XI = ATANISQRT(I.0/X.'2 - 1.0 I
NP = NO
: NPQIJO
NOr = NIlO
[PSI = fltts
[p<;2 = EP2s
NP$'J = NP NP
= .Z.NOe.NP.Noe
WPITF.lll.1000I NpROn, NOH, HP, NSCRAC
WRITElll,101l11
CAll GAS<;O(1, NP, TH, TEMP, THEP,LU
WRlTflll, 100?1
OUI. PItt,
OHIC 11
CAll GASSOU, NP, oIFl, THiP, TEHP,lLJ

IFIFNU-t.019q, 9'1, 16
COtIT I NUl'"
no 19 I=l,NP
TEMP = ABS(OIFZltll
IFfAMINlf1.0-TEHP, AQSITHCUI .. 99, 9'),19
CONTINUE
- GA = FLAH
NIT = 1
lA05 = it
IFIEPS11 5,10,70
FPS 1 = 0
70 ssa = 0
CAll HO'lE If NpPOIJ, TH, F, NO!), HP I
DO 9/1 [ = 1, NOr)
RUI = Yet) - F(1I
90 S'i(1=SSQ.RI U.R( It
WRITEClL,100lISS1
Eft I,
REG!N ITERATION
too GA = GA I fNU
101
INlr. NT = 0
WRJrE(Ll,1004)NIT
JC; = 1 - Non
DO 130 J=I,NP
Tf"HP = T!f(JI
prj) =OIfz I JI'THIJ)
TH(JI= THIJI.pfJ'
Q(JI=O
JS = JS HOO
W
N
N
60
fllj
70
75
In
fI5
CltI
95
100
10;
110
c
in
c
C
120
130
131
160
151
AU
153
C
155
C
C
231
W'UTrCfi
1
1Q'J
H ,.'U.,15'
CALL TH, O[LZIJS', NP'
IJ .:
no 170 I .: 1, NOI)
rJ=IJ.l
OrllCIJ' ::: I)fL7ftJ' - Flft
IJCJt = '/IJt .. OLZIIJ' HIt
0(",'.:: 'l J"P(J)
THU' ::: rr"p
IFfL.\O" IH,tJ1,"I"
01 1'>1.1 I .: 1, NP
1J011j1J=I,I
;: Ii
KJ .:
I( r = 140 1. ( I-I ,
110 1"" I( = 1, !H)1I
.... 1 .: I( I .. t
I(J .: I(J .. 1
SUM = ... 'J[LlfI(U DElli kJ'
TfHP::
..II = J .. NII( 1-1)
0( .11' = TI'""P
It = I NPfJ-ll
o I J' .! rE
E (f I = TID (J n I
CONTINU".:
no IS3 I:: 1, NP
, I J = I -NP
-00 153 J.:l. I
I J = IJ NP
AUJ' = DCIJ' I fEcn.ECJIt
JI :: J .. N''''U- U
AIJIt = HIJ'
II = - NP
00 155 [= I, NP
P(tt=ofU/rn,
pHIl It =p( It
II = NP .. 1 .. I I
HIl' = AI III .. (iA
r,,1
CALL MATINIA, NP, p, I, DEfI
srFP=I.0
SUM1=0.
=0.
SIJ"3.::0.
DO l31 I=I,NP


SUM.I: PHI f n PHIl It .. SUM]
PHI f II :: P II'
TE HP = SU",lIS0QTI SUH? SUM3,
l'7xr.R (STEEPE';T JE,>CiNT)
A= SCALED HOMENT MATRIX
PIE = CORRECTioN VECTOR
W
N
W
tn
1Zu
12;
130
t.1S
tI,ll
tlt5
t;!J
155
t60
165
170
TEHP = ""IN1(TEHP, too.
= 57
WRITflLL,l04u DEf, TEHP
178 00 220 I = 1( NP
p n. = PH I ( [ S TF.( piE II
T1HIJ = THfIt. P II
221 CONTINUE
WRIrr(LL.7000.
7800 POINT pARAHETER VALUES
NIUTEILL,200G' (fRU', I = 1, NPJ
00 221 I = I, NP
IFC SIGNSII)) ZZ 1, 2211 222 .
222 E63, 221,221
221 CONT INUE
SUHfl=O
CALL HOllEL INPPOB, TB, F, NOR, NP'
00 230 hl,NOR
RU,=Y(()-FU'
230 SUHn=SU'flh R( [ 'H I'
WRITE(LL,l04J'SUMB
IffSUHA - UoO.(psU.ssa' 667., 662, 66'
663 Iff AHIN1fT[MP-l0.0, GAr) 6&5, 6&5, 6& ..
665
[NTCII/T = INTeNT' 1
[F(IlI/rrNT - 36 170, 2700
66 .. C,,=f;A.FNU
INTCNT = INTCNT 1
IF'INTeNT - 66&, 2700, 2700
662 WRlrr(LL,1007)
DO &6q [=1,,.,.
TH( U=Tflf It
CALL GIISSOfl, NP, TH, TEMP, TE"fP,LLJ
GAt
22Q,Z2'),l2';
669
Z2Q 2/0,Z711,265
ZZ5 01) ;>411 I:: I, NP
IFUr,S,,,(ltJl(l.lE-?O.4!lSfTHCUn-EPS?' 240,2,,0,241
Z .. 1 IFft.:PSlI 270,27",265
Z .. O rntnTtllJ'
W'HH fLL,100'H[P')t!
GO ro 211(J
265 - SSQI - EPSl.SSQ' 2eG, Z6E, 270
Z66 WRITfflL
J
lf)lOJ (PSl
Go TO ?ltu
Z 70 ';Uri'l
NIT= NIT.l .
[FINIT - HIf' 101), 100, 280
2700 WR
2710 THE SUM OF SQUARES' CANNOT Bf Rrnlr.ED
r: tor SQUARES AT THE END OF THE LAST ITERATION - lTERATING
C
C
2110 NRTTEfLl,1011'
WRITEflL,20U' Iff II, I =
WRlT['fll,1 012'
WRITFfLLoZOOlJ IRI IJ, r =
1, NOll'
1, NO'lt
END I T[RATION
TO THE SUM
STOPS It
W
N
.::-
In
1"0
185
no
11)5
lOl
l.H
110
?u
Ul
Z?5
1017
769Z
3 .. 0
31t!
391
r
c
,,1ft
.. zo
"l'i
Co HI
.. ,.,
4:-'.
413
,,'HrE Ill; 1017.
FOPM4r CIIII,t l(
C_ll G
ft
SSOC4,NP,TEMP,rEMP,0,lLJ
SSQ:::SUMIi
10F::NO'l-NP
Np.lTrILL,1015J
1=0
CALL HATtrlCr', HP, p, I, DETJ
00 7f,CfZ I::l.NP
It :: I + NP. cr- U
fCIJ ::: SQRTCnfIIJJ
00 3"0 1:::1.MP
JI ::: I + NP.CIIJ 1
IJ ::: I + NP-cr.2J
00 340 J ::: I. NP
JI ::: .II + 1
ACJI' ::: oCJr. I I[IIJ-CIJJJ
1.1::: IJ + HI'
AfIJI = AfJJJ
CAll NP, TEMP, TEMP, A,LLI

CALL G_ssnCI, NP, E. TEHP, TE"P.LL)
IFCIOFI l41. 410. 1"1
S'lI'V ::: SS() I 10F
SDEV, IOF
= S'lPTCSDEVJ
00 1 Cfl 1= 1 .NP
PCI)=THCI)tZ O-ECI)-SOFV
TqCIJ:::THCII.',O'CIJ.SDEV
W r r f C LL 1 03'J
CALL NP, TO, p, TEMP,LLJ
LAOS ::: 1
.GO T I) 101
00 41G K ::: 1,
TEMP ::: ol
00 [=I.NP
00 4 ZO J= 1 NP
Isu" = K+NO'I' CI-ll
ornur.l =
UffUGI = D[l7'K +
ISIIf1 = KHICrJ.IJ-tJ
O[I.ur.Z :: 0 r L7 , I suI) J
or[lu G2 :: O. II C I( NO!)- r J-ll )
I J :: I C J- It
nrfiur.3 :: 0.rJIII'UFZCf)'THCIJ'nrF7'J"THCJ))
TF"'" = rr:ttf' OC"lur.l orAtJGl
Tr1P = ?
P(lO:f 11('. Tn .. "
r CKI :FlI(' - T['W
rt: CLL, 1 00" I
r r= 1
JI)
It"= .e,
lfe'HIIl-I[) 4H. "1'1,43<;
lr=-af)"
Wl(lrrCll.?OOU J : r. IrJ
crcJ" J:: I. hI
.. qu. C II dill H ,.."I'rH
W
N
\J1
ZJ]
?Ji
?"D
2 .. S
250
255
2&D

qq wR r H ll, 1 i) lit'
(i0 Tn C,lJ
1000 ESTIMArloN. PROAlF-H 11.1/ 15.
. 1 1ftH 11H [7. 17H SCRATI,;H REQUIRED'
1001 FORMATI/?:>HOtNITIAl VHIIF:S )
1002 fO"'MAT(lSftHOpPOPOIHIONS USFO Itl CAlCuL"TING OIFF[RENC[ QUOTIENTS'
1001 fO"'HAT IIZSHOINITUl 'iIJH Of =. E12."J
100 .. FOPHATfllllll.5 X,lJHtTERATlON NO tit.
1007 VALUES VIA RrGRrSSION
1008 fOPHAT(11115C,HOAPPRQXIMATE CONFtUENCElIHITS FOR rACH FUNCTION VAL
lur
1009 FORMATI/o2HOITERATION STOPS - RELATIVE CHANGE IN EACH pARAHETER lE
ISS THAtl 12 ....
1010 fOPHAl,,:.zHorTERATION STOPS - PElATIVE CHANGE IH SUM OF SQUARES LE
ISS TH N rlZ.4'
1011 for'HATIZZH1FINAl VAlurs
101Z FO"'HATflllllOHORf r UAl I
101ft 0 RESIDUALS = .E12.".lH, I",
12014 OF
1015 fORMATflllll<)HOr:r)R'''F.:LATION "ATPIX
101& .. INr,
1031 FORHAT(lllqHO[NO OF PRnBlEM NO.3.
10llt FO"''1AT IIlf>HOPARA'1ETER J
1039 fORHATC171HOJNOIVIDUAL CONFIOENCe LIMITS FOR EACH pARAMETER (ON LI
IN(AR HYPOTHESIS'
10,.0 =EI0.3. 1t1(.3lHSU" OF SQUAReS AFTER REGRESSION =
IE 1',.71
10 .. 1 fORMATUftH DFH:RHINANT = F1Z .... &X. 25H ANGLE IN SCALED COORD. =
1 F S 2, "H 1)[(; PE E'i
10ft3 FO"'Hlf(?IIHOTrST poINT SUM OF SQUARES = E12.4'
2001 FODMAT(/SEIZ.It)
200& fo
Q
HAr(5E12.'"
END
W
N

1
5'
10
15
20
('5
30
35
IiAThlfAr NVAR, Nnr OFf)
DIHfNSION ), ACNVA , 1
pIVOT" :: A Cl,U
IlQ=10
00 SO = 1, yVAQ
PIVOT:: f ICOl, COLI
PIVOT" = AMINIIPIVOT, pIVOT")
r.
Olf = PIVOT OET
C DIVIOI'.: PIVUT ROW BY PIvor
C
AI [COL, ICOU = 1.0
PIVOT = 1.E-20t
pIVOT = AfICOl, oll'PIVOT
DO l=1tNVAR
350 AI[r.Ol, l) :: ACICOl, l).PtVOT
tFCNfl .EQI OJ GO TO 371
og 370 l= NO
370 CICOl, II = anCOl, U.PIVOT
C
r. REDUCE NON-PIVOT ROWS
C
311 oq 550 ll=1,NVAq
Fill .fa. [COl) GO TO 550
T = A I l , Ie oLl
AClI
0
ICOl) :: O.
00 It') l=1,NVAR
I,SO AlLt, U = Afll, LJ - A([COl, l ).T
IFe Nfl .Elll ot Go TO 550
DO 500 l= ,Nfl
. 500 Bfll, l) = OCl1, lJ-ACICOl,lJ.T
550 CONrtNUF.:
RErURN
ENO
W
N
'-J
1
.;
16
15
20
10
30
ItO
15 60.
liD
90
20
95
100
?5
500
6011
77.0
111
GA,),)O (ITYPE t Na, A, B, C, LU
OMENS 0,. CNQ) ,rJCNQ) ,c NQ,NO)
tIP = NQ
NP :: 0
lOl4 = 1
lVP :: 10
IF NI1 11,,20, 31)


I flOW .GT. LUf" RETUT1,.
WRlrf(U,500' IJ,J=lot.f,LIJP)
GO TO
WRIff.CLL, 00 eA J ,J::LOW,LUP)
GO TO 100
WRITEfLL,600' CBe.Il,J=LON,LUP'
GO 41)
00 I) I:lOW LyP
WRITfCLL,72&' ,eceJ,I)'J=lOlhIJ

r eLOW? .GT
l
GO TO 100
02 <)5 [=lOW.,tlP
W ITEeLL,7Z0.r,er.eJ.I),J=low,lUP)
LOW = LOW 10
lUP = lUP + 10
1
Go 10 10
FORM Te/SX,J7,9'SX,I7
Z It'
O,13,9C1X,F12.1t')
f tiD
W
N
00
TABU ~ 5.1. ESTIMATED PARAMETER VALUES POR COPPER-CYANIDE COMPLEX AT PIV! FIXED MCN/TeN RATIOS
20C 20C 4C 4C 20C 20C 4C 4C
AIR NO AIR AIR NO AIR AIR NO AIR AIR NO AIR
NO UV NOUV NOUV NO UV UV UV UV UV
QI-CIf G PCN
O
6.hl0-
3
6.7xlO-
3
6.5xlO-
3
6.3xlO-
3
6.5x10-
3
6.3xlO-
3
6.2xlO-
3
6.6xlO-
3
1
,
Ity 3.9x10-
2
2.4xlO-
2
7.4xlO-
3
5.OxI0-
3
3.4xI0-
2
2.5x10-
2
2. hlO-2 1.4xlO-2
E
"
kl
2.lxlO-2 7.1xl0-
3
3.8xl0-
3
3.hlO-
3
1.5xl0-
2
2.4xlO-
3
3.611:10-
3
3.1xlO-3
PCN
o
6.8xl0-
3
8. Oxl0-
3
6.6xlO-
3
6. hl0-
3
6.6xl0-
3
6.611:10-
3
6. 4x 1 0-
3
6.5xlO-
3
"CIt
4.611:10-
2
1.411:10-
2
1.8xlO-
2
1.611:10-
2
5. h10-
2
2.2xlO-
2
2.1x10-2 1.4xlO-2
TCii 0.0425 Ity
kl
5.2xlO-
10
6.611:10-
2
9.911:10-
10
1.8xlO-
9
I. 1x I 0-10 4.7xlO-
10
1.2xlO-
10
1.911:10-
9
PCN
O
7. Ox 10-
3
7.611:10-
3
6.7xlO-
3
6.3x10-
3
7.2x10-
3
7.4xlO-
3
6.911:10-
3
6.4xlO-
3
"CIt
It.,
5. 9x I 0-2 1.5xl0-
2
1.7xl0-
2
1. 5x 1 0-
2
8.2xlO-
2
1.7xl0-
2
1.7xl0-2 1.4xlO-2
TcN 0.085
kl
4.611:10-
3
6.9x10-
2
2.911:10-
10
1.8xl0-
10
2.8xl0-
3
6. 5x 1 0"" 1.911:10-
10
1.5xlO-
9
PCN
o
6.hl0-
3
6.911:10-
3
6.5xlO-
3
6.2x10-
3
6.5x10-
3
6.7xlO-
3
6.5xlO-
3
5.911:10-
3
"CIt
Ity 7. 1x I 0-
2
1.5xlO-
2
1.5xl0-
2
1.5xlO-
2
I. Ox 1 0-
2
2. Oxl0-
2
1.811:10-2 1.711:10-2
TcN 0.17
kl 1.2xlO-2 7.4xl0-
2
4.911:10-
10
4.7xlO-
10
6.611:10-
3
4.3xlO-
3
2.911:10-
3
2.8x10-
1O
PCN
o
5. Ox 10-
3
5.4x10-
3
5.2xlO-
3
4.5x10-
3
5. Ox 10-
3
5.3xlO-
3
5.2x10-
3
4.8xlO-
3
"CIt
Ity 9. 8x 1 0-2 1.5xl0-
2
2. Oxl0-
2
1.1x10-
2
1.611:10-
1
2.5x10-
2
2.4xlO-2 3.7xlO-2
m 0.34
kl
2.3xlO-
2
8.911:10-
2
3.1xl0-
3
4. Ox 10"" 1.2xlO-
2
8.2xlO-
3
6.7xI0-
3
2.7xI0-3
FCN
o
2. 5x I 0-
3
2.6x10-
3
2.5xlO-
3
2.4xlO-
3
2.511:10-
3
2.5x10-
3
2.511:10-
3
2.4xlO-
3
"CIt
Ity
5.2x10+
1
3.6x10-
2
5.2xlO-
2
7.911:10-
2
5.6x10+
1
4.6xlO-
2
4.7xl0-2 3.911:10-
1
m 0.68
kl
3.4xlO-
2
1.711:10-
2
. 6.211:10-
3
3.1x10-
3
4.1x10-
2
1.2xlO-
2
l.lxlO-
2
6.lxl0-3
w
N
U)
'l'AlUK L5.L II'l'DIAftD PAUMI'I'R YALtIIS POll ZDC-cYMI .. COfPLD A'I' ,1ft '1110 MCR/'l'CII JA'l'lOS
20C 20C 4C 4C 20C 20C 4C 4C
AIR NO AIR AIR NO AIR AIR NO AIR AIR NO AIR
NOUV NO UV NOUV NO UV UV UV UV UV
G
I
'CNo
S.8xlO-
3
S.2xlO-
3
6.2xlO-
3
S.OxlO-
3
6.2xlO-
3
6.4xlO-
3
6.3xlO-
3
6. OxlO-
3
V
ky
3.9xlO-
2
2.4xlO-
2
7.4xlO-
3
5. OxlO-
3
3.4xlO-
2
2. SxlO-2 2. lxlO-2 1.4xlO-2
II:
"
kl
6.2xlO-
3
1.9x10-
2
6.8xlO-
3
4. SxlO-
3
4.SxlO-
3
1.8xlO-
2
1.4xlO-
2
9.hI0-3
'CRo
6.3xlO-
3
6.2xI0-
3
6.4xlO-
3
S.lxlO-
3
7.2xI0-
3
7. 8x I 0-
3
7.8xlO-
3
6.lxI0-
3
MCR
TciJ -.0.0425 ky
7.hlO-
2
1.8xlO-
2
1.8xlO-
2
I. Sx10-
2
S.9x10-
2
2.1x10-
2
2. Ox 10-2 I. SxlO-2
kl
I. SxIO-2 6. 1x 1 0-2 1.hI0-
9
S.lxIO-
IO
21.7xlO
O
6.1x10-
2
9. Ox 10-2 I. lxlO-
1O
'CWo
6.lxI0-
3
S.9xlO-
3
6.8xlO-
3
5.lxlO-
3
6.9x10-
3
7.4xlO-
3
7.4xlO-
3
6. lx10-
3
MCR
'lCir - 0.08S ky
8. 4x I 0-
2
1. 9x I 0-
2
I. h10-
2
1.4xlO-
2
7.1x10-
2
2. 1x 1 0-
2
2. Ox I 0-2 2.lxlO-2
kl
2.1xl0-
2
6.2xlO-
2
7.hI0-
IO
1.lxl0-9 1.4xlO-
2
6.lxlO-
2
6.3xl0-
2
1.4xlO-3
'C"o
S.6xlO-
3
S.3xlO-
3
6.2dO-
3
S.3xlO-
3
6.3xlO-
3
6.7xlO-
3
6.7xlO-
3
S.8xlO-3
MCR
l.lxl0-
1
1.9xlO-
2
1.8xlO-
2
6.9xlO-
3
8.4xlO-
2
2.2xlO-
2
2. 1x 1 0-2 2.4xlO-2
TCii - 0.17 ky
kl
2.8xlO-2 6.7xlO-2 3.2xlO-
a
1.4dO-
3
2.2xlO-
2
6.3xlO-
2
6.2xlO-
2
4.SdO-3
. 'CNo
4.hl0-
3
4.2xlO-
3
4.9xl0-
3
4.2xlO-
3
4.9xl0-
3
S.3xlO-
3
S.3xlO-
3
4.6xlO-
3
MCR
ky
1.7xlO-
1
2.6dO-
2
2. Sxl0-
2
9.lxI0-
3
1.1xl0-
1
2.4xlO-
2
2. 2x 1 0-2 3.4xlO-2
TCii - 0.34
kl
3.8xlO-
2
2.3xlO-
2
6.4xlO-
3
3.OxI0-
3
3. Sx10-
2
6.6xlO-
2
6.lxlO-2 7.7xlO-3
'CWo
2. lx10-3
2. Oxl0-
3
2.4xlO-
3
2.1xl0-
3
2.4xI0-
3
2. SdO-
3
2. SxlO-
3
2. lx10-
3
MCR
- 0.68 ky
S.6dO-
1
4.9x10-
2
6.6dO-
2
2.2dO-2 8.2xlO
O
2.6xI0-
2
2. 6x 1 0-2 I. lxl0-
1
TC" kl
S.4dO-
2
1.9xl0-
2
9.2dO-
3
4.3xlO-
3
4.3xlO-
2
8.3x10-
2
S.7xlO-
2
1.1x10-
2
w
w
0
TOLl "'.3.
IS'I'DIAftD PAIMK1R YALUU fOR IIIc:an.-cTAIf1DI com.IX AT rIft rDlD IICW/TCII IATIOI
20e 20C .C 4C 20C 20C 4C 4C
AIR NO AIR AIR NO AIR AIR NO AIR AIR NO AIR
NOUV NO uv NOUV NO UV UV UV UV UV
.-eII G rcw
o
7. Sx I 0-3 6. h10-
3
6. 5.2xlO-
3
6. h10-
3
6.3xlO-
3
6.4xlO-
3
5. Sx I 0-3
I
V

3.9xlo-
2
2.4xlO-
2
7.4xlO-
3
5. Ox I 0-
3
3.4xlO-
2
2. Sx I 0-
2
2. h10-
2
1.4xlO-2
It
"
1t1
2.7xI0-
3
8. lx I 0"" 3.OxI0-
3
2.7xlO-
3
8.3xlO ..... I. Sx I 0-
3
4.2xlO-
3
4.lxlo-3
rc"o
7. Sx10-3 6.hI0-
3
6.7xlO-
3
S.3xlO-
3
6.4xlO-
3
6.4xI0-
3
6.6x10-
3
S. Sx10-3
MCW
0.0425 ky
2.4xlO-
2
1.8xlO-
2
I. 6x I 0-
2
3. Sx I 0-
2
2.4xlO-
2
2. lx10-
2
1.4xlO-2
fCW
1t1
7.OxI0-
9
2.7xlO-
9
6. lx I 0-
9
3.hI0-
9
3. SxIO-
IO
3. Sx10-
9
2.4xlO-
1O
2.2xlO-
9
rcw
o
7. Sx10-3 6.3xlO-
3
6.7dO-
3
5.3xlO-
3
6.4xlO-
3
6.4xlO-
3
6.8xlO-
3
S.4dO-
3
!!! 0.085
ky
3.lxI0-
2
2. SxIO-2 1.8xI0-
2
1.6x10-2 3.7xI0-
2
2.3xlO-
2
I. 9x I 0-
2
1.5xlo-2
'fCIf Itl
2.lxI0-
IO
7.7xI0-
lO
4.2xI0-
lO
6.3xI0-9 4. Ox 10-10 7.2xlO-
1O
I. Ox 10-
9
7.txI0-9
rcw
o
7. Oxlo-
3
6.3xlO-
3
6. Sx10-
3
S.3xlO-
3
6. 3x I 0-
3
6.4xlO-
3
6.7xlO-
3
4.txI0-
3
MCW
ky
1.8xlo-
2
I. Ox 10-
1
l.txI0-
2
I. Sxlo-2 8. 8x I 0-
2
I. tx 1 0-
2
1.7xI0-
2
I. txlO-2
- 0.17
TCW
Itl
8.7xlO ..... 1.7xlO-
3
3.OxI0-
10
1.2xI0-
3
5.txI0-
5
1.7xlO-
3
1.2xlO-
lO
rcw
o
5.6x10-
3
4.8xlO-
3
S.2dO-
3
4. Sx10-
3
4.9x10-
3
S.lxlO-
3
S.3xI0-
3
4.3xlO-
3
!!! 0.34
Ity
2. Sxl0-
2
1.4xlO-
1
3.lxI0-
2
I. 9x I 0-2 1.lxI0+
1
2.8xlO-
2
2.2xlO-
2
S.OxIO-2
'fCIf
Itl
4. Sx10-
3
9.txI0-
3
1.4xlO-
3
1.7xI0-
lO
S.6xI0-
3
3. Sx10-
3
S.2xlO-
3
3.2xlO-3
rCNo
2.8xlo-
3
2.4xlO-
3
2.4xlO-
3
2. Ox10-
3
2.3xlO-
3
2.4xlO-
3
2. Sx10-
3
2. Ox10-
3
!!E!. 0.68
It.,
7. txlO-2 3.4xI0-
l
3.3dO
O
4. SdO-
2
6.4xlO+
1
6. tx10-
2
4.8xlO-
2
2.2dO-
1
fCW
7.2xlO-'3
1t1
8.lxI0-
3
3.8xlO-
2
4.4xlO-
3
2.8xlO-
3
2.3xlO-
2
8. tx10-
3
7.lxI0-3
w
w
'fAl'" L 5. 4- IftDIAt'ID pAltAllftft YALUU POR com.n AT PlVI PlDD 1ICR/'I'CIf RATIOS
zoe zoc 4C 4C zOC ZOC 4C "4C
AIR NO AIR AIR NO AIR AIR NO AIR AIR NO AIR
NOUV NOUV NO UV NO UV UV UV UV UV
I'.-eII G PCR
o
6.4xlO-
3
6.ZxlO-
3
5.911:10-
3
4. lx 1 0-
3
5.7xlO-
3
5.7xlO-
3
6. hdO-
3
5.7xlO-
3
I
,
Ity 3.911:10-
2
Z.4xlO-
2
7.4xlO-
3
5. Ox 1 0-
3
3.4xlO-
2
Z.5x10-
2
Z.3xlO-2 1.4xlO-2
Z
"
kl
1.6x10-
3
Z.6x10-
3
I.ZxlO-
3
7.4xlO-it 6.8xlO-
3
Z.ZxlO-
2
5.lxlO-
3
4. Ox 10-'
PCR
o
6. 5x I 0-
3
6.hI0-
3
5.911:10-
3
4. b:10-
3
6. Ox10-
3
6.6xlO-
3
6. ZxlO-
3
5.&dO-
3
NCR
TCN 0.04Z5 Ity
3.5x10-
2
Z. hl0-
2
1.911:10-
2
I. 6x I 0-
2
4.4xlO-
2
Z. hl0-
2
Z.lxl0-2 I. hl0-2
kl
3.ZxlO-
9
8. Ox 10-
9
7.3xlO-
9
Z.91I:I0-
10
1.4xl0-
10
1.5xl0-
3
3.911:10-9 3.hI0-
9
PCR
o
6.5x10-
3
6.4xlO-
3
6. Ox 10-
3
4.lxlO-
3
6.5x10-
3
6.4xlO-
3
6.ZxlO-
3
5.8xlO-
3
NCR
TCN 0.085
Ity
3. Ox 10-
2
Z.2xlO-
2
1.911:10-
2
1.7xlO-
2
8. h10-
2
Z.5x10-
2
I. 1tx10-2 1.lxlO-Z
kl
Z.ltxIO-
IO
1.lxI0-
9
Z.2xI0-
10
4.911:10-
9
1.7xlO-
3
4.4xlO-
3
5.hI0-
11
4.6xl0-
10
PCR
o
6.6x10-
3
6.3xlO-
3
6.lxlO-3 4. Ox I 0-
3
5.8xlO-
3
5.911:10-
3
6. lx I 0-
3
5. 6x I 0-
3
NCR
Tcif 0.17 Ity
I. 4x I 0-
2
1.911:10-
2
Z.5x10-
2
I. Itx I 0-
2
1.lxlO-
1
Z.91I:10-
2
9.ltxlO-3 6.911:10-
3
kl
l.hI0-
3
Z.4xlO-
1O
Z.lxlO"" 8.911:10-
10
5.ZxlO-
3
7.7xI0-
3
Z.7xlO-
3
1.4xlO-'
PCR
o
5. ZxlO-
3
5.1x10-
3
4.8xlO-
3
3. 5x I 0-
3
4.5x10-
3
4. 4.911:10-
3
4. 5x I 0-
3
NCR
Tcif 0.34 It.,
I. 1tx10-2 Z.5xlO-
2
3.5x10-
2
5.911:10-
2
O. ItxlO+
1
4.1xlO-
2
I.ZxlO-2 &.ltxI0-
3
kl
4.911:10-
3
2.911:10-
3
4. Ox 10-
3
2.5x10-
3
9.7xlO-
3
1.1xlO-
2
4.911:10-
3
3.2xlO-'
PCR
o
Z.5x10-
3
2.4xlO-
3
2.2xlO-
3
1.5xlO-
3
Z.2xlO-
3
2. Zx10-
3
Z.3xlO-
3
2. ZxlO-
3
NCR
1fCff 0.6& Ity 3.5xlO-
2
6.3xlO-
2
8.7xI0-
2
1.5xlO-
1
7.6x10+1 I. Ox10-
1
Z.7xlO-2 1.911:10-2
kl
&.2xlO-
3
6.bI0-
3
8.4xlO-
3
7.911:10-
3
4.2xlO-
2
1.5xlO-
2
6.5x10-
3
4.5x10-'
w
w
N
&.
o
20 C. AIR. NO UV
&a
...
CD
sa L.
....
m 1.1
...
.. a
... .. II .. It ........ .
&.
o .
20 C. AIR. UV
&a
...
CD
sa L.
....
X
~ 1.1
...
.. a
... .. II ILIt IL. IL. 1L8I
an
o
20 C.NO AIR. NO UV
I-
I-
I - ~ M ~
""

an
51
....
m
0::
o
4 C.AIR.NO UV
.. III ILII 1L24 IL. IL. IL. ILIII 1.12 0.Z4 a.. a.. a..
MCN/TCN RATIO
o o
28 C.NO AIR.UV 4 C. AIR. UV
"" ""
m I-
....
an
m
m ~
x
m ,
~
.1. ~
~ ~ : __J .-.l_
I
*--
" ----::-L .. __ I _ _ -----.L
o.ao 1.12 1.24 IL. IL. IL. ILIII 1L12 1.24 IL. IL. a..
MCN/TCN RATIO
Figure R.S.l Residual Sa. of Squares versus MeN/TeN Ratios - Co
~
<lilt
m
....
~
~
an
m
....
x
m
0::
o
4 C. NO AIR. NO UV
ILIII 1L12 1LZ4 a.. IL. a.
o
4 C.NO AIR.UV
ILIII 0.12 1L24 0.. 0.. IL
w
w
w
LI ,
o
28 C. AIR_ NO UV
La
...
CD
sa 1.1
....
X
~ 1.1
.. I
.. a
..... 12 .. M ........ .
LI
o
28 C. AIR. UV
La '"
...
CD
sa 1.1 '" ....
X
~ 1.1
.. I '"
... ~
...
~ . L
... .. 12 .. :U ........ ..
o
28 C.NO AIR. NO UV
'"
'"
an
sa
....
x
(J)
'"
~ ..
o
4 C. AIR. NO UV
1",""","---"",- ~ ~ ~ ..
I-
r\

..III .. 12 .. M ... ... ... ..III .. 12 .. :U ........ .
MCN/TCN RATIO
o o
28 C.NO AIR.UV 4 C. AIR. UV
..III .. 12 .. :U ... ... .... .. ... 12 .. :U ........ ..
MCN/TCN RATIO
Pigure B.5.2 Residual Sua of Squares Yersus MeN/TeN Ratios - Zn
...
~
sa
....
x
fa
cr:
o
4 C. NO AIR. NO UV
.. ... 12 .. at ........ .
o
.. C. NO AIR. UV
.. ... 12 .. :U ........ .
w
W
.t-
I.,
o
20 C .. AIR .. NO UV
1.0 ..
L. I-
...
~ I..
J
en
0: 1.2 ..
~
I..
1.0
...
an
51 I..
....
X
en
~ 1.2
a.,
0.0
o.DO 0.12 0.2. a.. a... o.eo
o
20 c.. AIR .. UV
o.DO 0.12 o.U a.. a. .. o.eo
-'"
an
51
....
~
0:
~
o o
20 C.NO AIR. NO UV <4 C. AIR.. NO UV
o.DO 0.12 o.U a.. a.. o.eo o.DO 0.12 o.U a.. a. .. o.eo
MCN/TCN RATIO
o o
20 c.. NO AIR. UV <4 C. AIR. UV
I-
"'
I-
.

o.DO 0.12 a. 24 a. a. .. a. eo o.DO a. 12 a. U a. a. 48 a. eo
MCN/TCN RATIO
Pigure R.S.l Residual Sua of Squares versus MeN/TeN Ratios - Hi
~
...
SI
....
X
(I)
en
0:
o
4 C. NO AIR.. NO UV
o.DO a. 12 a. 2. a.. a. .. 0.111
o
4 c.. NO AIR. UV
o.DO 0.12 o.U a.. a.. o.eo
w
w
V1
o
20 C- NO AIR .. NO UV
o
4 C .. AIR .. NO UV
11.111 11.12 II.:U G.. G." 11.10 11.111 11.12 II.:U G.. G... 11.10
MCN/TCN RATIO
o o
20 C- NO AIR .. UV 4 C- AIR .. UV

m
9"1
m
11.111 11.12 II.:U G.. G... 11.10 II. GO II. 12 II.:U G.. II. .. 11.10
MCN/TCN RATIO
Pigure 8.5.4 Residual Sua of Squares versu8 MON/TeN Rati08 - Pe
o
4 C- NO AIR.. NO UV
II.GO 11.12 1I.Z4 G.. G. .. 11.10
o
4 C- NO AIR .. UV
II.GO 11.12 1I.Z4 G.. G. .. G..
w
W
0'\
APPENDIX H.6
Dot Diagrams
Estimated Itl with Respect to Temperature
UV Light and Aeration at MCN/TCN - 0.17
337
338
1.0 1.0
Cu-CN Zn-CN
-
115:1 IL' X
I.,
-
X

x
I ILl
'"
I..
-

1.2 !" I..
X
X
X
1.0


-'
1.0
W

0 ..

12 II ID Ie 0 ..

12 II ID 1M
TEMPERA TURE DEG. C
1.0 1.0
Ni-CN F.-CN
N
X
115:1 IL' i"
I.,
-
x
,..
l ILl i"
I.,
X
-

I.. 1.2
X
X X
X X

ILO
w
-'-
. . ...
1.0


-
r
0 ..

12 II

1M

..

12 II

..
TEMPERATURE DEG. C
Pigure 8.6.1 Eatiaated kl at MCR/TCR - 0.17 versus Te.peratare
339
2.0 2.0
-
Cu-CN Zn-CN
s 0.7
-
l-
X
0.7
)(
,..
X

j o.S
o.S
-
X
0.2 I-
0.2
X
X
x
...

I

X
0.0 o.D
NOUV UV NOUV uv
PRESENCE OF UV
2.D 2.1
Ni-CN F.-CN
N
x
S 0.7
0.7
-
)(
,..

)o.s I- o.S
X

-
X
0.2 I- 0.2 l-
X
X X
X
X
X
X
0.0
w

w
o.D
... .
y
NOUV UV NOUV UV
PRESENCE OF UV
Figure 8.6.2 Estiaated 1t1 at MCN/TCN - 0.17 versus UV Light
340
LO LO
-
Cu-CN Zn-CN
m 0.7
-
1.7
X
~
-
X
.r.
.......
~ 0.5 o.s I-
'""
-

0.2 ..
X
0.2
X
)(
0.0
~ I I 1
0.0
~ I .
NO AIR AIR . NO AIR AIR
PRESENCE OF AIR
.LO ~ G
.Nl-CN F .. -CN
N
X
m Jl,?
47
-
X
-
.r.
) o.s o ~
X
0
'""
-

0.2 0.2
X
X X
X
X
~
o.G
-
.
-
0.0
.... . ~
NO AIR AIR NO AIR AIR
PRESENCE OF AIR
Pigure B.6.3 Estimated kl at KCN/TCH - 0.17 versus Aeration
APPENDIX I
Low Mix Model Simulations
342
1.1 EXAMPLE CALCULATION FOR THE INITIAL VALUES OF METAL AND FREE
CYANIDE IN THE LOW MIX
The single metallo-cyanide solutions and low-mix solutions
theoretically have 17% of total cyanide tied up with the metal (17% of
200 mg/L). The difference is that in the low mix, the total
concentration of four metals complexed 17% of the cyanide present. The
four metals involved were Cu, Zn, Ni and Fe. This synthetic mix
emulates the composition of one real gold mill effluent which contained
about 7, 10, 2 and 2 mg/L of Cu, Zn, Ni and Fe respectively.
Therefore, the low mix had the same concentrations of these metals.
The value used for [MCNlo in the low mix was based upon the
theoretical stOichiometry. It is calculated as follows:
Concentration Theoretical Stoichiometry
Metals (mg/L) for the assumed complex [MCNlo (mol/L)
Cu 7
(CN)3/
Cu

1.228 0.00033
Zn 10 (CN)4/Zn = 1.592 0.00062
Ni 2
(CN)4/
Ni
- 1.773 0.00014
Fe 2 (CN)6/Fe - 2.795 0.00022
0.00131
Hence the total initial MCN (equal to 17% of TCN) is 0.00131 mol/L.
The total cyanide and metal concentrations are the only
measurable initial concentrations. The total and free initial cyanide
concentrations for the prepared low mix solutions are tabulated in
Table 1.1.1:
TABLE 1.1.1. INITIAL VALUES FOR TOTAL AND FREE CYANIDE CONCENTRATIONS
Conditions [TCN]o (mol/L) [FCN]o
(mol/L)
Air UV 4C 20C 4C 20C
+ + 0.00761 0.00722 0.00631 0.00592
+ 0.00746 0.00769 0.00615 0.00638
+ 0.00711 0.00765 0.00581 0.00635
0.00822 0.00761 0.00692 0.00631
The initial free cyanide concentration was calculated by subtracting
calculated [MCN]o from measured [TCN]o concentrations.
e.g., 0.00746 - 0.00131 - 0.00615 mol/L of [FCN]o.
344
1.2 THE EXAMPLE CALCULATION OF THE ULTRAVIOLET IRRADIATION RATE
CONSTANT
The intensity of the light obtained from the UV lamp was
calculated to be 10 - 7.68 x 10-
4
einstein*/h.
Out of four single metallo-cyanide solutions only iron cyanide
was affected by UV light (see conclusions from the experimental
design). Therefore, the UV light rate constant (kuv) was calculated
using the results obtained on the iron cyanide solution.
If the metal decay constants ki were plotted against 10=0 and
I 7 48 10
-4 /
o x einstein h, the slope (m) would be, for the
conditions with and without UV light calculated as follows (assuming
the full depth of the vessel is operative):
20, AIR, UV & NO UV m
a
(0.0222 - 0.0026)/7.68 x
10-
4
.. 25.5
20, NO AIR, UV & NO UV m ..
(0.0068 - 0.0016)/7.68 x
10-
4
., 6.9
4, AIR, UV & NO UV
m -
(0.00399 - 0.00074)/7.68 x 10-
4
- 4.3
4, NO AIR, UV & NO UV
m -
(0.0051 - 0.0012)/7.68 x 10-
4
- 5.2
Since the first slope is an order of magnitude different from
the other three it was rejected in taking an average. Therefore, the
average m value for the other three conditions is ; - 5.5 (einstein-I).
* Einstein - a unit of light energy used in photochemistry equal to
Avogadro's number times the energy of one photon of light of the
frequency in question.
345
The UV light rate constant is proportional to the average light
intensity, i.e., kuv - m1a
According to Beer's law (Moore, 1963)
the average light intensity is equal
where is the absorption coefficient or extinction coefficient (e:
assumed to be equal to 10 em
2
/mol) (Kolthoff et a1., 1969) and Z is the
depth of the medium that lights go through (determined from the
absorbance spectra by UV spectrophotometry to be 33 em).
If the term in parenthesis is assumed negligible the kuv
becomes
I 5.5 [ 1 ]
o einst.
kuv = mTz=- 2
10 [em ] x 33
mol
If [einst.]
x 7.68 x 10-
h
55.4 mol 1L
em x x
1L 1000 cm
3
(
APPENDIX 1.3
The Results of the Low Mix Hodel Pi t and
Ex8llple of l'be Progr8118 Used
...... 0 +

,a

..

,

,

a

,

,
>- >- ... ... ... ... >- . .. ...
.. II .. to .. II If II ..
I I I I I I I I I
I I I I I I I. I
,
t
I
at
&(
t
I
*I'
III

.. III 1\


1\
"
III ..
. . . . . . .
Cl/LOW)
,
,
,
,
,
,
a
,
a
a
a
a

a

a








...
II
=
I
*I
1
,
I:
x
111
a .
x
: G
.G
N
x
111
,a .
x
a. G
G
.m
x .
N
a .
x
a G
, .G
..
a x
.N
a
a x
a.

G
G
a. x
.111

a x
a

.G

a x .
...


x
:1:
.N

X
...


x

: .

G

x .


x

.G
..
x
x
:1

x
.>-
..
:
I
*I

III
.
347
-
oJ:
-
N
0
-
)(
GJ
E
-....

Ii

-=
=
Ii

u

0
N
M
it
!
""'"

fI"I

...

fc

75K- '"
.'
.175- .,.

.... -.. .,.
I OBSERVED TCN


$ TCN}

+ HCN ESTIMATED
525- Y
* FCN
-
,.
-.I
.......
-
a

-
5K- .,.

e
0
op-
+.J
IU .I?!IE- Y
s..
+.J

C
QJ

U


C
.1eeE- Y

8


III

"t:J


0p-
e
2251:- Y

IU





15K- .,.


.+
,


+


+ +



+ +


."K-n.y + + + +

+ + + + + +


,
+ + + + +

+ + + + + + +
, , , ,

+ + +
, , ,
"
,

.Y
K i.K K K K K K K . K K K K K K K K K K . K K K
..... .333 .556 .999 1. 332 1. 865 1. 998 2.331 2.56" 2.997 3.33.
Figure 1.3.2 Low Mix - 4C. NO AIR. UV
Ti me x 10
2
(h ) W
J:-
00
.,..-R .,
.'
,
.... -R .,
,

, ,
I OBSERVED TCN
? .. -R .,


, $ TCN}

+ HCN ESTIMATED
-
* FeN
....
-.2 .,


"-
,
.-



,

-
c:
.' .. -R .,


0
,
....

,
..,

ItS
s..


.., ..... -R .,

c


QI

. ,
u
c:
8


........
.I'.-Rs"


QI
... ... .
"t:I

. ,..


c:

ItS
.2 -R .,
;?J







+
.1 -R ., + + +
+ + + +
+ + + + + + + + +
+ + + + + + + + + + + + + + + + + +
.Y
Ie Ie Ie Ie Ie Ie Ie Ie Ie Ie Ie Ie Ie Ie )( )( )( )( )(
..... ..... 1.2" 1.&" 2.'" 2..... 2 . " 3.Z'. 3.&"
T1 me x 10
2
(h )
Figure I.l.] Low Mix - 4C. ItO AIR. NO UV
'-"
.::-
\.D
-
...,
.......
.-
a
-
c
0
....
..,
IU
'- ..,
C
ell
U
c
8
ell
'0
....
C
IU

.?SK-lZaV
.,
?SE-IZ:V
.... -Nt:Y


.SZSE-lZaV
SK-lZaV


.'?SE-lZaV


.'''-HaV


.ZZSE-H:V
,


.
ISK-NaV
,
.+

+
+
.?SK-.'aV +
,
+
,






+

+ + + +


+
." +

I OBSERVED TeN
$ TeN J
+ HeN ESTIMATED
'" FeN
+ ,,+ ,,+ + + + + + + +
." " "" + + + +". + +
:V ".
Ie Ie , Ie Ie Ie Ie Ie Ie Ie Ie Ie Ie Ie Ie Ie Ie Ie Ie Ie Ie Ie
... .313131 .666 .999 1. 31312 1. 665 1. 998 2. 331 2.664 2.997 31.3131.
Ti me x 10
2
(h )
Pigure 1.3.4 Low Mix - 20C, AIR., OV
W
\.f1
o
-...I
.......
-
a
-
c
o
....

"'
So.

c:
CII
u
c:
8
CII
"0
-C

.,ee-R."
.,
,
.,.n-_."
.'2K-_."
. ,
.IIK-_,."


. ,
. "---="
,

. ,11:- "

.2IK-_="


11K-g."


+
.,,.-0=,, + +

+ +
+
.
.
,
,



+ +
,

,

+ +

, , ,
I OBSERVED TCN
$ TCN }
+ MCN ESTIMATED
* FCN
,
+ + + + + + ,
+ + +. + + + + + + + +

1( JC JC JC JC JC 1( 1( 1( 1( JC IC IC IC IC IC 1( IC IC
..... .4.. . 1.2.. 1.&.. 2.... 2.4.. 2.88. 3.2.. 3.&
11 me x 10
2
(h )
Pigure 1.3.5 IDw Mil[ - 20C. AIR. NO UV
w
V1
-.....
.....
-
S
-
c
o
.,...
+J
"'
s..
+J
c::
QI
V
c
8
QI
"0
.,...
c
5
.ffn:- .,
., ,
.f .. -N_.,
, ,
.-
IK-ft .,
., .. sc-.! .,
-

.
..... -.. .,
,

-
. .,.-.. .,

-
.111E-12_.,

-

.IIK-'!:"
,
-


- . "K-H."

-



,


-

IIE-.I:"

-


-
,




,



I OBSERVED TCN
$ TeN }
+ MCN ESTIMATED
* FCN
,
,



. ".

. . . - .


"+

1'.-.4:'1
M M M M M M M M M M M M M M M M M M M M M
... III &&& !I!I!I 1. 332 1. &&5 1. '!III 2 33 t 2 664 2. !I!I? 3 II.
Plgure 1.3.6 1Dw Mix - 4C, AIR, UV
Time x 10
2
(h )
W
\.n
N
--'
.......
.-
i
-
e:
o
.....
+I
,.,
s..
+I
e:
cu
u
e:
8
cu
"'0
....
e:

.7 -U .. 'I'
.,
-.7.51:-.2:'1'
, .
6:1.-82:"
,

I OBSERVED TeN


$ TeN }

+ HeN ESTIMATED
,
* FeN

.5551:-U",.

,
4"-82:'1'
, .

,



4851:-82:'1'






13K-82 a 'I'


, .

, .
,. ,
.2551:-82:'1'


, ,



,




I"-82e'l'


, ,

,

,


.1851:-82:'1'


+ + + +

+ + + + + + +

+ + + + + + + + + + + + + + +
.... -.:1:'1' + + + +
x .. x X x x .. X x . X x .. x . x x . x x .. x .. x . x .. x X
..... . . 1.2" 1.&" 2.... 2.... 2 . " 3.2" 3.&"
. 11 me x 10
2
(h )
Plpre 1.3.7 Low 1U.z - 4C, AIR, 110 UV
w
V1
W
PROGRAM MIX (INPUT.OUTPUT.TAPES:INPUT.TAPE6:0UTPUT.TAPE11)
DIMENSION TH(lZ)
REAL V(7).C(24).W(7.9).X.TOl.XEND
COMMON RKC.RKZ.RKN.RKF.RkV.VK.V.AS
FCN!
READ(11 ) eTHCI).I:l.12)
WRITEC6 ) (TH(I).I:l.12)
NW:7
N:7
x:e.e
VC 1) :TH( 1)
VCZ):THC2)
V(3):V(I)+VeZ)
V(4):THe4)
VeS):THeS)
V(6):THe6)
V(7):THC7)
RkC:TH(8)
RkZ:TH(9)
RKN:THCle)
RKf":THC 11)
RKU:TH(1Z)
VK:TH(3)
v:zeee.
AS:47.78
TOl:e.eeal
IND:l
CAll PlOTPTeX.vel).4)
CALL PLOTPTCx.v(2).2)
CALL PLOTPT(X.Y(3).2a)
WRITE(6.1aae)
laae FORMAT( lex. "X" .12X. "ve 1)". lZX. "yez) .12X. "V(3)")
WRITC6.2eea) X.V(1).Y(Z).YC3)
2eaa FORMAT(6X.F6.2.6X.f"18.7.6X.f"18.7.6x.f"1e.7)
DO la K:l.32
XEND:f"LOATeK).18.
CALL DUERK CN.FCN1.X.V.XEND.TOl.IND.C.HW.W.IER)
IF eIND.LT.e.OR.IER.GT.8) GOTO 2e
WRITEe6.3eea) X.V(1).ve2).Y(3)
3aaa FORMATe6X.f"6.2.6X.f"18.7.6X.FI8.7.6X.f"18.7)
CALL PLOTPTeX.Y(1).4)
CALL PLOTPTCX.Y(2).2)
CALL PLOTPTCX.Y(3).Z8)
18 CONTINUE
READ ( 11 ) NOBS
DO 15 3':I.NOBS
READe u . )XO. YO

WRITE(6 )XO.Vl
CALL PLOTPTexo.Vl.S)
15 CONTINUE
CALL OUTPLT
STOP
2a CONTINUE
STOP
END
SUBROUTINE
DIMENSION TH(lZ)
COMMON RKC.RKZ.RKN.RKF.RKV.UK.U.AS
REAL VCN).VPRIMEeN).x
YPRIME(4):-RKC.Y(4)
YPRIME(S):-RKz.YeS)
VPRIME(6):-RKN.YC6)
VPRIME(7):e-RKr.V(7-RKv.ve7)
VPRIME(Z):VPRIMEC4)+YPRIME(S)+VPRIMEe6)+VPRIM(7)

,
RETURN
END
354
1
2
3

S
6
7
a
9
18
II
12
13
14
lS
16
17
lB
19
28
21
22
23

2S
26
27
2B
29
3B
31
32
33
34
3S
36
37
3B
39



c
PROORA" DATA3CIHPUT.OUTPUT.TAPES:IHPUT.TAPE6:0UTPUT.TAPEll)
DI"EHSIOH TH(11)
INTEGER N.IND.NW.IER.K
REAL
CO""OH /A/ RKC.RKZ.RKH.RKF.UK
c
C
C
C
C
THIS PROGRA" DOES "ODELLIHO OF A "IX WITH HO UU EFFECT. IT REQUIRES
'lHCll).H08S AHD DATA SET
C
C
EXTERNAL FCNl
READCll ) CTHCI).I:l.ll)
WRITEC6 ) CTHCI),I:l,ll)
NW:7
H:7
X:8.8
YUhTHCU
Y(2):THC2)
Y(3):YUHYC2)
Y(4):THC4)
YCS):THCS)
Y(6):THC6)
Y(7):THC7)
RI<C:THCB)
RKZ:TH(9)
RI<N:THUB)
RI<F:THUU
UK:TH(3)
TOL:B.8BBl
IND:l
WR ITE C 6. 2BBB)
2888 FOR"ATC1BX."X",12X."Y(1)",12X,"YC2)",12X,"YC3)")
WRITEC6.7BBB) X,Y(1),Y(2),YC3)
7888 FOR"ATC6X,F6.2,6X,F1B.7,6X,F1B.7,6X,F18.7)
DO lB 1<: t. 35
XEND:FLOATCK).18.
CALL DUERK CH,FCN1,X,Y,XEHD,TOL,IND,C,NW,W,IER)
IF CIND.LT.8.0R.IER.OT.B) OOTO 28
WRITEC6,3BB8) X,Y(1),YC2),YC3)
w
V1
V1
S
--EOR--
43
....
.. 5
.. 6
.. 1
.. 8
.. ,
5.
51
52
53
5 ..
55
56
57
58
59
I
2
3
..
5
&
1
a
9
I.
11
12
13
14
IS
69231 .1.129
)(

I
2 .
3 8.
......
5 8
3188 FOR"AT(6)(.F6.2.6)(.FI ?6)(.FI ?&)(.FI 1)
CALL PLOTPT()(.V(I)."J
CALL PLOTPT()(.V(2J.2J
CALL PLOTPTC)(.V(3J.28)
I. CONTINUE
READC 11 )NOBS
DO 12 J:l.NOBS
READUI. III )(0. VO
VI:VO'2&888
WRITE(6 (0. VI
CALL PLOTPT()(O.VI.S)
. 12 CONTINUE
CALL OUTPLT
stop
21 CONTINUE
STOP
END
SUBROUTINE FCNI(N.X.V.VPRI"E)
DI"ENSION TH(II)
COMMON 'A' RKC.RKZ.RKN.RKF.UK
INTEGER N
REAL V(N).VPRI"E(N).)(
VPRIMECI):(RKC.V("JJ+(RKZ.V(5+(RKN.Y(&+(RKF.Y(7-
+(UK.V(l) )
VPRIME(2):-(RKC.YC .. -CRKZ.VCS-CRKN.VC6-CRKF.VC7
VPRIME(3):-UK.V(I)
VPRIME("):-RKCIIIV(")
VPRIME(5):-RKZ*V(5)
VPRIME(&):-RKN*V(6)
VPRIME(7):-RKF.V(7)
RETURN
END
5.26 .e 33 .. 2
V( 1)
..869231
.8066952
.086 .. 621
.8862266
.0859912
.8857576
88618 .88.1377
V(2)
.BB129B8
.0011756
.B818781
.00099 .. 7
.00B9231
BeeS615
8217 . 295
V(3)
.8882131
.8B787e8
.8075"02.
.B072213
.80691"3.
.8B6619B
.81783 .108 .. 373 1 .... 5
IJ.J
'J1
0-.
68.88
78.88
88.88
99.89
188.98
119.98
128.98
138.88
148.88
158.88
169.89
178.89
188.89
198.88
298.89
218.88
228.89
239.88
248.98
'258.88
269.88
279.88
298.88
298.88
388.88
'318.88
e . 23.76923.76'
I' 88115384615385
24 887961538461538
41. ..875
48. ..81423.76,23.77
72. .886615384615385
144 885946153846154
151 885538461539462
168. .884884615384615
115 884538461538462
199 883589461538462
216 883253846153846
223 883219238769231
241 882815394615385
247 882792381692388
112 88211'238169231
.8955213
.8853814
.8858987
.8848661
.8846518
.8844563
.8842617
.8848741
.8838936
.8831282
.8835531
.8833948
.8832418
.8838946
.0829545
.0828285
.8826925
.8025181
.0024533
.fi!823417
.0022353
.8821336
.8828367
.8819442
.8819568
.8817719,
.8889882
.8881628
.8887217
.8886865
.8806554
.8886288
.8886835
.0885816
.8885619
.8885441
.0005218
.0085129
.8004991
.0004864
.0004745
.0004634
.0004529
.0004431
.0084337
.0004248
.0004162
.0004081
.0004082
.0003927
.0003954
.0003784
.8863355:
.8868634
.8858825
.0855526
.8853133
.8858842
.8848652
.8846551.
.0844555
.8842642
.B040815
.8839069
.8037402
.8835818
.B834298
.B832839
.B831454
.8838132
.B828878
.8827665
.8826515
.8025417
.0824369
.8023369
.0822414
.8821503
\AI
V1
-.....J
APPENDIX J
Exa.ple Calculationa for:
- High Mix Model Staulationa.
- Barren Solution Model Sigulationa.
- Effect of pH.
358
359
J.l EXAMPLE CALCULATION FOR mE INITIAL VALUES OF METAL AND FREE
CYANIDE IN 'l'IIE HIG[ MIX
The high mix cyanide solution was designed to have metals
complexing with 117% of total cyanide (200 mg/L). Hence, when
complexes with the metals Cu, Zn, Ni and Fe are formed, there is an
excess of free metal. The characteristics of the real gold mill
effluent was emulated as with the low mix, but there was proportional
increase in the metal concentration. Therefore, the values were:
Metals Theoretical
[MCN]o [MCN] 0
(mg/L) StOichiometry (mg/L) (mol/L)
Cu
=
48.7 (CN) 3/Cu =
1.228 59.79 0.002298
Zn
= 69.5 (CN)4/ Zn = 1.592 110.68 0.004254
Ni
=
13.9 (CN) 4/Ni =
1.773 24.66 0.000948
Fe
= 13.9 (CN)6/Fe
= 2.795 38.87 0.001494
0.008994
Total initial metal cyanide concentration 234.00 mg/L
0.008994 (mol/L).
The total cyanide concentrations recorded for the prepared high
mix solutions were as tabulated:
Conditions
[TCN]o(mol/L)
AIR UV 4C 20C
+ + 0.00692 0.00742
+ 0.00761 0.00765
+ 0.00738 0.00734
0.00761 0.00753
360
In order to determine metal cyanide concentrations and avoid
having negative initial values for free cyanide (because our metal
cyanide is higher than the total cyanide concentration) the
calculations were carried out in two ways:
i) For the first simulation assuming that we were dealing with
MCN
o
= TCN
o
' 15 to 23.1% of each [MCN] was calculated and subtrac-
ted from the metal cyanide concentration, in which case the sum of
individual concentrations became equal TeN
o
'
For example, for the
lowest TeN = 0.00692 mol/L, 23.1% of each MCN was substracted in order
to obtain MCN
o
= TCN
o
'
Adjusted
Cu - CN = 0.001767 mol/L
Zn - CN = 0.003271
Ni - CN = 0.000729
Fe - CN = 0.001149
... 0.00692 mol/L
Other conditions are as tabulated:
Conditions % MCN Substracted
AIR UV 4C 20C
+ + 23.1 12.4
+ 16.0 15.0
+ IS.0 lS.3
16.0 16.2
361
ii) When high mix cyanide solution was made the Cu, Zn and Ni
3-
salts were added to the cyanide solution first and Fe(CN)6 the last.
The assumption for the second simulation was that during this process
the complexes of Cu-CN and Zn-CN break up easily releasing C ~ Excess
Ni and Fe maybe present to react with freed C ~ before it can be lost
by volatilization. Hence, net effect, kinetically, is that Fe and Ni
have their full equivalent of C ~ and the remainder is assigned to Cu
and Zn. Therefore, 56 to 66.5% of total Cu and 75 to 86.5% of total Zn
concentration form a complex with present cyanide, as tabulated:
Conditions % Cu - CN % Zn - CN
AIR UV 4C 20C 4C 20C
+ + 56.0 63.0 75.0 83.0
+ 65.0 66.5 86.5 86.0
+ 63.0 62.0 82.0 82.5
65.0 64.0 86.5 85.0
In the case of the lowest measured total cyanide, for example
the individual metal cyanide concentrations became:
Adjusted
Cu - CN (56%)
Zn - CN (75%)
Ni - CN (100%)
Fe - CN (100%)
- 0.001287 mol/L
- 0.003190
- 0.000948
- 0.001494
[MCN]o - 0.00692 mol/L
362
J.2 EXAMPLE CALCULATION OF mE I.NITIAL CYANIDE CONCENTRATIONS FOR.
THE K>IEL APPLICATION ON BARREN SOLUrIONS
When the experiments commenced the main characteristics for the
examined barren solutions were:
Barren TCN CUT ZnT NiT FeT
Solution (mol/L) (mg/L) (mg/L) (mg/L) (mg/L) pH
f1l 0.00558 7.8 31.0 1.5 0.1 11.5
112 0.01134 15.0 127.0 0.7 5.7 12.0
113 0.00490 7.7 1.1 0.5 10.0 11.6
The possible presence of any other metal or ligand in these
barren solutions was at this point omitted from consideration.
Assuming only the presence of the same complexes as in
synthetic solutions and taking into consideration their theoretical
stoichiometry, the initial metal and free cyanide concentrations were
calculated as in the following example (#1 barren solution):
Metal
Cu = 7.8 mg/L
Zn - 31.0
Ni ... 1.5
Fe ... 0.1
Theoretical Stoichiometry
(CNh/Cu = 1.228
(CN)4/Zn "" 1.592
(CNh/Ni == 1.773
(CN)6/Fe == 2.795
[MCN]o (mol/L)
0.000368
0.001898
0.000102
0.000011
0.002379
Total theoretical [MCN]o concentration = 0.002379 (mol/L)
The total cyanide determined analytically in this barren
solution was [TCN]o'" 0.0055769 (mol/L). The initial free cyanide
concentration was obtained by subtracting calculated [MCN]o from the
measured [TCN]o
The same stoichiometry was applied for the other two barren
solutions according to their metal and total cyanide content.
364
J.3 MODELLING EFFECl'S OF pH UPON VOLATILIZATION OF II:N
The model equations used for the synthetic mixes were all made
with the fixed pH of 7.0 in mind. When the model was to be used on the
barren solutions certain adjustments had to be made on the equations to
compensate for the change in pH during the degradation process.
It has been assumed (Chapter 7) t that HCN and ~ are in
equilibrium according to the reaction:
[HCN]
+
+
The dissociation constant for this reaction is
Therefore
[H+] [CN-]
[HCN]
and the [CN""'] is defined as
= 4.93 x 10-
10
(pKa = 9.3)
(1)
365
The term "free cyanide" is defined as the sum of the molecular
HCN and cyanide ion CN-j i.e.,
[FCN] - [HCN] + [ ~ ]
If we replace the [CNr] with the value from Eq. (1) total free
cyanide will be equal
[FCN] - [HCN] +
From here the concentration of molecular HCN is equal to
[HCN] _ ---'o.r_FC_N .... l _
1+ Ka
[H+)
From the definitions for pH and pKa values
[1'"] - 10-
pH
and Ka - 10 -pK
When replaced in the Eq. (2) the HCN concentration becomes equal
[HCN] -
1 +
[FCN]
10 -pK
10 -pH
(2)
(3)
366
At pH of 7 [HeN] == 0.995 [FeN], justifying the approximation used in
modelling the synthetic solutions that [HeN] equals FeN.. Therefore,
the value for molecular HeN in the model was replaced by the equation
(3)
The computer program used for the model simulation on barren
solutions called BARRMIX is attached.
Exaaple COmputer Prograa USed for Modelling
on Barren Solutions
367
I
2
3
4
S
6
7

9
I.
11
12
13
14
IS
16
17
I.
19
2.
21
22
23
24
2S
26
27
28
29
3.
31
32
33
34
35
36
37
38
39
4.
IS
c
PROGRA" 8ARR"IXCINPUT.OUTPUT.TAPES:INPUT.TAPE6:0UTPUT.TAP(11)
DI"ENSION TH(12).PHC4.).T08SC4.)
V(7).CC24).WC7.9).X.TOL.XENO
:O""ON 'A'
E)(TERNAL
REAOCll ) CTH(I).I:l.1Z)
WRITEC6 ) (TNCI).I:I.1Z)
REAO(I1 )N089
DO 15 .1:I.N089
REAOCI1 )(0.VO.PHl
.. 1:YO .... 26
PH(1) :PHI
T08S<1 h)(O
WRITEC6 )XO.Vl.PHl
CALL PLOTPTC)(O.Vl.S)
CONTINUE
PHCN08S+I):PH(N08S)
T08SCN08S+1):T08SCN08S)+S
"':7
N:7
X:
VClhTHCl)
V(2):THC2)
V(3):VCl )+V(2)
V(4):THC4)
V(ShTH(S)
VC6hTH(6)
V(7):THC7)
RKC:TH(8)
RKZ:TH(9)
RKN:THU.)
RKF':THUI )
UUK:THCl2)
UK:TH(3)
TOL: ..... 1
IND:l
CALL PLOTPT(X.VC1).4)
CALL PLOTPTCX.V(2).2)
CALL PLOTPTCX.V(3).2.)
NRITEC6.1 )
W
0'\
00
" .. 2
.. 3
....
.. 5
.. 6
.. 7
...
"9
5.
51
52
- 53
5 ..
55
56
57
58
59
68
61
62
63
6 ..
65
66
I
2
3
..
5
6
7

9
I.
11
12
13
I ..
I F'OR"ATCI.)(,")( .. ,12)( ...... (I)",I2X ...... C2) .. ,12)(, ..... C3) .. )
WRITEC6,2 ) )(,"'(I), ... (Z), ... C3)
2 ... F'OR"AT(3)(,F'I 5,6)(,F'I 7,6)(,F'18.7,6X,F'18.7)
DO 1. 1<=1,132
)(END=F'LOATCI<).18
12:XEND
DO 3. I:l,NOBS
I1=TOBSC I)
13=TOB9CI+l)
IF'Cll.LE.12.AND.I2.LE.I3)GO TO 31
38 CONTINUE
31 PH1=PHU)
WRITEC6,.)PHI
CALL DUERI< (N,F'CN1,X ..... XEND,TOL.IND.C.NN.N.IER)
IF' CIND.LT OR.IER.GT.8) Q()TO 2.
) )(.VC1) .... (2),VC3)
-3 8 F'OR"AT(3)(.F'I 5,6X,F'I 7.6X.F'18.7.6X.F'18.7)
C CALL PLOTPTC)(.YC1), .. )
CALL PLOTPTC)(,VC2),2)
CALL PLOTPTC)(, ... C3),2.)
18 CONTINUE
CALL OUTPLT
STOP
Z8 CONTINUE
STOP
END
SUBROUTINE F'CN1(N.X .... ,VPRI"E)
DI"ENSION THCI2).PHC ... ).TOBSC .. 8)
CO ""ON 'A' RI<C.RI<Z.RI<N.RI<F',UUI<,UI<,PHI
REAL
PKI=9.8
VPRI"EC .. ):-RI<C .... C .. )
VPRI"E(5):-RI<Z.VC5)
"'PRI"E(6):-CRKN.VC6
... PRI"E(7):(-RKF'.YC7-UUI<.VC7)

VPRI"[CI):-VPRI"EC2)-CUI<.VC1)'CI+I C-PI<I)'I C-PH1)
VPRI"E(3):-UI< .... Cl)'CI+I C-PI<I)'I C-PHI
RETURN
END
w
C1'