Beruflich Dokumente
Kultur Dokumente
01-) _ _ _----I
DOLIIffi
ORGANIZING COMMITTEE
PROGRAMME COMMITTEE
ISBN 1-85166-715-6
No responsibility is assumed by the Publisher for any injury and/or damage to persons or property as
a matter of products liability, negligence or otherwise, or from any use or operation of any methods,
products, instructions or ideas contained in the material herein.
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or
transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or
otherwise, without the prior written permission of the publisher.
Foreword
This volume contains the papers that will be presented at 'EMC '91 '-the European
Metals Conference-to be held in Brussels, Belgium, from 15 to 20 September 1991,
and organized by Benelux Metallurgie, GDMB (Gesellschaft Deutscher Metallhutten-
und Bergleute) and IMM (the Institution of Mining and Metallurgy).
'EMC '91' is the first of an intended major series organized at the European level with
the aim of bringing together all those who are involved with the extraction and
processing of non-ferrous metals-European metallurgists and their international
colleagues-to provide them with the opportunity to exchange views on the state and
evolution of their industry.
The programme covers all the different aspects of the metallurgy of non-ferrous metals
from mining to fabricated products. Particular attention is being paid to the European
non -ferrous industry with respect to changes in demand, the technology used, pressures
on the environment and the competitive position of manufacturers.
The contributions of the plenary lecturers (copies of which will appear in the IMM
journal Minerals Industry International in 1991-92) and the many authors are gratefully
acknowledged. Thanks are also due to the referees of the papers, the sponsors, the
companies that have allowed registrants to visit their operations, the chairmen of the
technical sessions and the staffs of the organizing bodies for their efficient administrative
work.
Jean Vereecken
Chairman, Organizing Committee
July 1991
v
Contents
Foreword. . . . . . . . . . . . . . . . . . . . . . . .. v
Mineral Processing
Design of non-cyanide technology for flotation of lead-zinc ores: energy
prerequisites, implementation and results. . . . . . . . . 3
V. Panayotov
Grinding of the 'Jales de Santa Julia de la Compania Real del Monte y Pachuca',
SA de CV, Mexico. . . . . . . . . . . . . . . . . . . . . . 19
F. Patino and J. Ramirez
Hydrometallurgy
Ammonium jarosite formation in ferric chloride leaching processes 65
J. E. Dutrizac
Recovery of copper by dump leaching with use of bacteria and cementation at the
Vlaikov Vrah mine, Bulgaria. . . . . . . . . . . . . . . . . . . 109
S. N. Groudev and V. I. Groudeva
vii
viii
Limitation of lead jarosite formation in the leaching of calcine from high-iron Zn-
Pb-Cu concentrates . . . . . . . . . . . . . . . . . . 127
M. Pedlfk and J. Jandova
Nickel
Intensification of the reductive-roast ammonia leaching process for nickel lateritic
ores . . . . . . . . . 139
J. Jandova and M. Pedlfk
Extraction of Co( II) and N i (II) with Cyanex 272 . .,. . . . . . . . . 175
D. Malikovic, Z. Lenhard and M. Balen
Copper
Technological and technical problems of copper production from chalcosine
concentrates in a flash furnace. . . . . . . . . . . . . . . . . . 185
J. Czernecki, S. Sobierajski and Z. Smieszek
Recycling
Automotive exhaust catalysts: PG M usage and recovery. . . . . . . . . 231
V.Jung
Vibration technique for recovery of non-ferrous, rare and noble metals from
secondary raw materials. . . . . . . . . . . . . . . . . . . . 257
A. Fedotov and G. Denisov
The TBRC as a unit with a promising future for secondary copper plants. . . 287
H. Bussmann
Minor Metals
Purity and long-term stability of 8-hydroxyquinoline-based metal extractants . 301
G. Haesebroek
Improved technology for In, Ge and Ga recovery in an electrolytic zinc plant . 305
Tian Runcang
Trends in the development of processes for recovery of rare metals .... 315
A. A. Titov, I. F. Polelayev, V. A. Krokhin and A. A. Schelkonogov
Environment
Best available technology-a viewpoint on the development and application of
the concept to the European non-ferrous industry. . . . . . . . . . . 351
Tony Connell, Robert Maes and Jose Poncet
Simultaneous microbial removal of sulphate and heavy metals from waste water 391
L. J. Barnes, J. Sherren, F. J. Janssen, P. J. H. Scheeren, J. H. Versteegh and
R.O. Koch
New biological treatment plant for heavy metal contaminated groundwater 403
P. J. H. Scheeren, R. O. Koch, C. J. N. Buisman, L. J. Barnes and
J. H. Versteegh
x
Silver plating from thiosulphate baths . . . . . . . . . . . . . 417
A. Hubin, G. Marissens and J. Vereecken
Light Metals
Status and challenges for modern aluminium reduction technology. . . . . 447
Harald A. flJye and Reidar Huglen
Modelling
Time-dependent simulation of Czochralski growth: application to the production
of germanium crystals. . . . . . . . . . . . . . . . . . . . . 483
N. Van den Bogaert and J. Braet
3
4
,cot
~~...o:;;.IIIIo-.a..""~EC
---E,
rsillll'------Ey
Q
,
Fig. 1 Bending of energy zones on the mineral surface under the influence of a contact with a charged
metal surface
la - enriched zone bending, lb - impoverished zone bending, Ec - bottom of conduction band,
Ev - top of valence band t Ef - Fermi level, Eg - width of forbidden zone
two principal moments: either electrons are in- ging the energy configuration of the mineral sur-
jected in the mineral particle or it is "enri- face and their relation to the flotation results
ched" in these basic carriers (in considering as well as the devices for measuring these chan-
n-type minerals) - Fig. la, or conversely, when ges, the quantitative assessment and prediction
in this contact electrons are lost then the re- of the technological results are given in other
sult is an "impoverished" bending of the energy papers /9,10/.
zones - Fig. lb. We shall discuss briefly the problem raised
The bending of the zones on the mineral sur- above. In order to assess quantitatively and
face is given by the relationships: qualitatively the degree of energy effect on the
mineral surface (in this case reached by elec-
dVI X=O -- --;r;-
dX
dV/ V:V$ -JL
- £Eu
trochemical impact) it is necessary to have in-
formation about the energy state of the mineral
where: ~ - surface barrier which characterizes phase. For this purpose energy diagrams of the
the bending of the energy zones minerals are plotted. In these diagrams all
Qc - density of the surface charge [Clcm2] possible states in nature are reflected. In
If we consider the contact of the mineral par- Fig. Z is presented the joint energy diagram of
ticle with a metal electrode, then it is obvious the most common minerals in dressing lead-zinc
that the depth of penetration of the metal elec- ores: ZnS, FeS Z' PbS, CuFeS Z' in which the con-
tric field in the semiconductor mineral, determi- centrat10n of the free charge carr1ers n [oJ]
em
as -
o 0
ned by the difference in the work functions, mineral surface charge [t fern.!] ;Jlf - num-
should be presented most generally with the ex- ber of possible fixed particles of the collector
pression: ~117 -2] i c;, - the same concentration expressed
in [1179/(1111], which permits the easier deter-
mination of the reagent-collector concentration
and its monitoring. The isolines from 1 to 7
where it is evident that the penetration depth of show, respectively, the degree of bending of the
the field is so much greater as the difference in surface zones. For e 1ff,./11 T= 1 straightened zo-
the work functions is stronger and as the concen- nes are observed and for e 'f}/Ji T = 7 inversion
tration of the charge carriers in the semiconduc- is observed. These are boundary cases and between
tor mineral is lower, i?~ [c m -J 1, for a distance these are all possible states of zone bending
X ~~D in the semiconductor there being no which can be caused by outside impacts, admixtu-
field. The consideration of other effects chan- res, defects, etc., i.e. these are all the
5
i
3.7£-09J 1.2£+10
"6.9[-11
3.9£-12
2.2£+08
L2E+07
1£-"L
2.OE-12 --,--
8
~--T,--~--,''--'--~,---r-
12 14 16
,.-
18
'--,--.-~
20 Lg n
= [CU~ ~ . [Z n 1 + J
:HD
of obtaining particular flotation properties, by
using the so-called coefficient of energy flota-
Ii
tion insufficiency /9/. As far as the optimum [Znb'J .[CU 2t ]
selection of flotation reagents depending on the where: A - activation; D - depression;
energy state of the mineral phase fed for flota- The acceleration of the process of transmitting
tion is concerned, our latest investigations /11/ electrons from the mineral to the liquid phase
give a tentative direction in this respect. speeds up the activation process. 101e start from
When testing the hypotheses for purposeful the basic equation for each semiconductor mineral,
depression and activation of the zinc minerals
2
n.p n.
in the process of flotation without }I/o {'j/f and L
const
- ,lTIl -- .-
2
------1
2 4 6 fO 12 f4 '6 18 ZO
t, m~n
Fig. 3 Results of depression and activation of sphalerite without sodium cyanide and copper sulphate
but with electrochemical impact
Table
Fig. 4 Industrial system for activation of zinc vely, 3,4 - power lines, 5 - studs regulating
minerals located in agitation tanks the distance between electrodes, 6 - rubber li-
miters
8
The relationship between the strength of curr-
ent passed to the working electrode and the qua-
lity of the concentrates obtained was studied.
Fig. 6 and Fig. 7 show respectively the process- fo-n 1%1
ing of these data. The equations derived can be
very useful in controlling these processes. Si-
milar relationships are obtained also when con-
Y= 0,031 X + 45.028
trolling the admixtures of Pb and Zn in the cor-
responding concentrates.
®
71
72
160 1.
•• I (A 1
71
y = - 0, 0756 x + 87,34
Fig. 7 Dependence of the quality of the zinc
concentrate under industrial conditions on the
strength of current of the working electrode in
(A) in the process of activation of zinc mine-
rals without copper sulphate
References
* as PbO.5Fe3(S04)2(OH)6 ** as Fe2(S04)3.9H20
sulphate 3.7%, plumbojarosite 6.9%, Selem 0.05%) cess PbS04' A second synthesis was performed in
is roughly an average, whereas sample 2 (Au 1.3 the same way but with the addition of 5 9 of
gft, Ag 136 gft, ferric sulphate 9.8%, plumboja- lead-jarosite seed from the first synthesis.
rosite 11%, Selem 0.9%) is representative of zo-
nes in the deposits especially rich in ferric The XRD characterization of the resulting product
sulphate and plumbojarosite. indicated only jarosite phase (Fig.5). The chemi-
cal composition (PbO 12.8%, S03 30.5%, Fe203
Tests of alkaline pretreatment were made with 45.2%) was that of a plumbojarosite-hydroniumjaro-
500 9 solid (80% (100 pm) and 1 I of solution in site solid solution whose approximate formula is
a conventional stirred reactor equipped with pH-
PbO.3(H30)0.4Fe3(S04)2(OH)6' similar to the pro-
register. For the cyanidation tests the pulp was ducts obtained by Dutrizac under like conditions.
transferred to a pachuca reactor inmersed in a Most of the material was distributed in aggrega-
thermostatic bath. The process was followed by tes of 5-10 ~m, made up of rhombohedral crystals
solution sampling and analysis of Au and Ag by measuring 0.5-2 pm.
Atomic Absorption Spectrometry (AAS) and CN- by
volumetric analysis. The solid residues were ana- In the experiments for the decomposition of syn-
lyzed by fire assay and selected samples were thetic samples in Ca(OH)2 media dilute pulps were
examined by Scanning Electron Microscopy (SEM) used to prevent contamination of the residues by
in conjunction with X-Ray Energy Dispersive Ana- formation of gypsum. Typically. 2.0 g of sample
lysis (EDS). were treated with 1 1 of Ca(OH)2 solutions of dif-
ferent concentrations at differents temperatures.
Synthetic lead-jarosite The degree of decomposition was determined by
analyzing the sulphates in the solid residue and
Synthetic samples of lead-jarosite were used to by X-ray diffraction. The residues were also cha-
determine certain aspects of its decomposition racterized by SEM and EDS.
in Ca(OH)2 solutions, which are difficult to ob-
serve directly from hematite tailings. The syn-
thesis was carried out by a procedure similar to
RESULTS AND DISCUSSION
that described by Dutrizac~,8 For the firts syn-
thesis, 1 1 0.3M FellI (as Fe2(S04)3.nH20) and
Pretreatment with CaC03
0.03M H2S04 was reacted with 30 9 PbS04 at 100QC
for 24 hours in a stirred reactor. The product
Because of the presence of ferric sulphate, direct
obtained was washed four times with 1 1 10%
conditioning of the tailings with CaD presents
NH4CH3COO at 25QC for one hour to remove the ex-
serious difficulties for stabilizing the pH of
13
the pulp during the conventional cyanidation pro-
cess. A gelatinous precipitate is generated which
coats the CaO particles requiring 50-100 kg/t CaO
; that is an excess of alkali over the theoretical
consumption. Furthermore, the precipitate tends
to clog the pores of the hematite particles and NaCNIGIJ
.0.2.0
part of the ferric sulphate remains unreacted, & 1.0
VG.5
which leads to alterations of the pH in the cyan-
idation process and an increase in the cyanide
consumption (NaCN 2-3 kg/t).
volving tailings.
____
was practically stable. 15r-----------------------____ &IO"C
VIO"C
A 2O"C
Typical results on the effect of the cyanide con- ~ 2.0
SAMPlE 1
~r---------------------------
'"
2.800 keV 7.9 >
B
L
'"
010
40 1M)
Fig.? X-ray spectra of: A) Nucleus of lead-jaro- Fig.8. Temperature effect on the decomposition
site. B) Halo of decomposition (Ca(OH)2 0.8 gil, of plumbojarosite associated to hematite tailings
45 QC, 1 h). (pretreatment time 30 min, cyanidation time 30
min, NaCNi 2 gil)
=8==-==8
_e_e
===8= -----e-
~
E
u 150
___A
,-----A-
I! ,.1::.
~
...CD
~
Vi "1-
-v-V
o eooc A 3C)OC
o 150°C V2O"C
e 40°C
00 30 eo
Time min
Fig.ll. Basic flowsheet
Fig.l0. Cyanidation rates for tailings pretreated
at different temperatures (sample 1, pretreatment
time 15 min, NaCNi 2 gIl)
CONCLUSIONS
References
SYNOPSIS
A study of a scheme about the optil1um yrinding mathematical model for the process alternatives.
process of the "Jales de Santa Julia de la C~mpa With this purpose we use semi-empiric models of
Ria Real del Monte y Pachuca", Mexico, is presen- principal unitary operations involved.
ted. The study consists of an experimental work
of grinding at laboratory scale. The results were Grinding
applied to a digital simulator. The configuration
of two closed grinding circuits in series (direct This operation can be simulated mineral by
primary circuit and inverted secondary circuit) mineral and size by size with a model called of
was defined with the application of the following p-order accumulative, by using concepts of
selection criterion: (a) upper grinding to 48.3% Harris 2 ,3 and Horst, Bassarear4 . The principal
- 400 mesh, (b) optimum utilization water of pulp model characteristics are the following:
dilution, (c) cut correct size of minimum classi-
fication, and (d) to keep with the capacity li- 1.- The mill feed is characterized by a global
mitations recommended by hydrocyclone manufactu- granulometric distribution and the contents
rers. Nominal capacity investigated was of 150 distribution for the constituent mineral
metric ton/hour by grinding line. species.
2.- After a mean residence time in the mill, the
This paper considers the study of a grinding discharge also shows overall granulometric
process of "Jales de Santa Julia", which contain distributions and contents distribution for
residual values of gold and silver. High percent the constituent mineral species.
of the "Jales" were produced by cyanidation of 3.- The volumetric flow of the pulp is needed in
quarziferous ores. At present, the "Jales" the calculation of the mean residence time
production are of two classes: (a) cyanide tails in the mill.
of Pb/Ag flotation concentrates and (b) tails of
this flotation circuit that have not been 4.- The pulp transfer through the mill is defined
cyanided. by the following equation:
Silver recoveries obtained by cyanidation of the Wi,j (T) = Wi,j (0) exp(-FjKi,jTPi,j) (1)
ground ores as well as of the flotation concen- where T is the mean residence time in the mill,
trates were not high in any case. For this in min. Wi j(O) and Wi j(T) are the feed and
reason, the "Jales" re-grinding should be necessa- discharge massic flow in the mill for the j
ry for the recovery of the metallic values con- component of the ore and the size coarser than
tained. the i opening mesh, respectively. Ki j is the
accumulative specific grinding rate,' and Pi j
The practice for many years in the industrial is the grinding kinetics order of the specie'
plant of cyanidation has been to mill the quar- j of a size coarser than the mesh i. Fj is a
ziferous ores only 60-65% -75~m, whereas the scale factor applicable when the grindlng is
optimum size is 75-80% -75~m. simulated at industrial scale, from the grinding
kinetics derived of the tests made at laboratory
The objetive of this paper is to define an scale.
optimum grinding scheme using a mathematical
modelling of unitary operations involved. The Classification
re-grinding is the requisite of these "jales"
beneficiation. Recent metallurgical studies The classification is the more important unitary
suggest this operation 1 . operation, because in a closed grinding circuit
of polymetallic ores, the classifier (hydrocyclo-
ne) operates according to the particles specific
MATHEMATICAL MODELLING gravity as well as the particles size. In the
case of the dense ores, this consideration is
This method consists in a configuration of a very important.
19
20
There are several mathematical models for the of the grinding circuit lines.
hydraulic classification by hydrocyclones. In
this case, we need a hybrid model since it com- 2.- Mass flows of solids in each line.
bines the virtues of three different models in 3.- Pulp volumetric flow pumped to the hydro-
one. These models are the following: cyclones.
-Lynch and oRao classification model 4.- Solids percent by weight and volume in each
line.
This model found a wide success in the ores
industry to predict satisfactorily the clas~i~i 5.- Overall granulometric distribution in each
cation operation under anyone process cond1t10n line.
as the operation variables and the hydrocyclones S.- Granulometric distribution by species in
design. each line.
Two equations of the classificat~on general 7.- Mean residence time in the mill.
model of Lynch and Ra0 5 are app11ed:
S.- Recirculating load by specie.
* Percent of recirculated water, Rf 9.- Corrected overall cut size by specie.
Rf = K3 spigot/WF + K4 / WF + K5 (2)
where K3 and K4 are constant coeffic~entsoof the METHODOLOGY
model and K5 is a plant cons~ant, Sp1g0t 1S the
orifice diameter of coarse d1scharge of the Experimental method
hydrocyclone in cm. and WF is the water flow 0
FPS and Q are the percent solids by weight and Grinding Kinetics Calculation
the volumetric flow of the pulp feed to the
hydrocyclone in It/min., respectively. The grinding kinetics parameters were estimated
-Plitt classification model by lineal regression of the experimental data
applying the following model:
The Plitt modelS relates the classification
index of mineral species to the respective Inln(Ci,j (O)/Ci,j (t» = Pi,jlnt + In Ki,j (S)
corrected cut size, d50c j in accordance to
the following equation: '
jl~i'!m; 1~j~2
Yi,j = 1, - exp(-(ln2) (d/d50c,j)n j ) (4) lm=s; n=2
where Yi j is the classification index, d and
d50c ,j ar~ the g~om~tric midd~e size of parti- where Ci,j(O) and Ci jet) are the accumulative
cles Of the speC1e J between 1 and i+l meshes fractions coarser than the mesh i at time t = 0
a~d corrected cut size of the specie j, respec- and t = t (min) of the batch grinding of specie
t1vely. nj is the slope of the regression line. j.
Figure 1 shows the granulometric distributions 1.- Type and number of grinding circuits.
of "Jales de Santa Julia" recently processed
in the Ex-Hacienda de Loreto, Pachuca, Mexico, The primary circuit was selected of direct type
and subjected to 0, 2 and 8 min. of grinding. considering that the "jales" require a conditio-
Figure 2 includes the silver distribution size ning of their surface previously to the flota-
to size for 0, 2 and 8 min. of grinding. tion or any other separation method. This
Figures 3 and 4 show the curves of accumulative conditioning is referred to the removal of
percent coarser for all material and by specie slimes as well as another products generated
(silver), respectively, for 0,2 and 8 min. of with a prolonged oxidation of the "jales". In
grinding. this primary direct circuit all the material
pass through the mill before the classification
Simulation results step. Otherwise, a fraction of the "jales"
should be classified with the fines and the
Preliminary simulations to define the circuit probability that they pass through the secondary
type to produce the desired grinding grade was mill should be very reduced. The grinding
carried out. The circuit in serie formed by a secondary circuit was selected of the inverse
primary direct closed circuit and a secondary type to optimize the fines production in the
inverted closed circuit resulted from these global scheme of the grinding process. These two
simulation tests. The results are included in grinding circuits are sufficient to produce the
Figure 5. The nominal capacity of grinding was grinding grade required for the silver optimum
defined in 150 t/h of "jales". The appropiate metallurgical recovery.
dimensions of the mill are: 3 m. diameter and 3.5
m. length (outside dimensions). The primary and 2.- Hydrocyclone number.
secondary mills are of the same size and of the
overflow type. Simulations varying the hydrocyclone number in
the grinding primary and secondary circuits were
The number of primary and secondary hydrocyclo- carried out. The criteria used to select the
nes are six. The dimensions are the same in all hydro cyclone optimum number were: (a) using the
of them: (a) hydrocyclone diameter = 38.1 cm.; smaller spigot diameter, that produces the
(b) inlet diameter = 11.43 cm.; (c) spigot diame- smaller corrected cut size, d50c, and (b)
ter = 5.08 cm., and vortex diameter = 10.16 cm. having the pulp flow capacity that is recommend-
The operation parameters to define are: ed by the hydrocyclone manufacturer. Another
restrictions of the hydrocyclones were taken
(a) Tonnage of the fresh feed, in t/h. (b) Granu- into account: c) inlet flow capacity, and d)
lometry of the feed ores. (c) Chemical composition vortex flow capacity. However, the spigot
of the valuable elements. (d) Grinding kinetics. showed higher sensitivity to the limit capacity
(e) Densities of the mineral species of interest of the process. Additionally, the water con-
(A92S and Si02). (f) Solid/water relation in the sumption in the grinding circuit was also
feed to the primary mill. (g) Density of the considered. This way, although low densities of
feed pulp to the primary hydrocyclones. pulp produced corrected cut sizes of finer
classification, the water consumption tends to
Figure 8 shows the percent curves corresponding be high as shown in Figure 8 at pulp densities
to the size -400 meshes vs. hydrocyclone number lower than 1.50 t/m3.
only for the secondary circuit. Figure 7 shows
the curves of water consumption in l/min. used in
grinding of primary and secondary circuits vs.
hydro cyclone number for constant densities of
feed pulp to the hydrocyclone. Figures 8 and 9
illustrate the curves of corrected cut size,
d50c j vs. hydrocyclone number for constant densi-
ties'of feed pulp to the hydro cyclones for silver
and gangue respectively.
The mean residence time in the mill was 1.016
minutes and the scale factor, Fj used was 1.4
(the grinding in the industrial mill was 1.4 ti-
mes more fast than in the laboratory mill). F·
was determined by the experience obtained wit~
the grinding rate of this mineral type at indus-
trial scale. According to Figure 5, the scale
factor used, F' in secondary grinding was of 1.2
because the material was approaching the grinding
limit.
DISCUSSION
The selection of the grinding scheme consisting
of a direct closed circuit (primary circuit)
followed by a inverse closed circuit (secondary
circuit) and six hydrocyclones of the dimensions
above mentioned, was obtained according to the
following considerations:
22
WISH NUMB(R
+ ~ 0 Il10 ••
'0 o. Z .. ,#10
D. • • 111 1 11
100'~---------------------------- ________ ~
:;
. ,r-------~--------------------__A ~t 1~
HI Z10
;0;
.
10
~H"OO
.0
~
...
~
i ..
~
\ Ht .,00
\
Nt 00
.0 .00 ' 00
fIR II CU 51![ IN MIC m 5
CRINBING lIM[ IN Mim ES
Fig. 1.- Granulometric distributions at 0, Fig. 3.- Percent in accumulative weight coar-
2 and 8 grinding minutes. ser vs. grinding time.
MiS~ NUMBER
<00 .,0 '00 <0O <00 00
'0
+ • 0,., ,111 100
0, l ~,,.
..
~ . $; 11'1 1 ",
"
>-
;;-
./'II I 400
+
~
~
10
" "-\ HI 210
- '0
~
~
=
~
.,. /' ~
~H"OO
I I
""
~
0
-- '" /1
/1
'" /, =
""
=
= .. 111 1 ,,~o
/ /
\
'" / "-
~
fill 400
/
'-
/
Fig. 2.- Granulometric distributions silver at Fig. 4.- Percent accumulative coarser of
0, 2 and 8 grinding minutes. silver vs. grinding time.
23
®
-f--
G) IDl @
.6
~----{]r--- -~----~I/L----~~ 0mm~CIDII( • ,,-
In 011't1.E'T .. " ~ ...
rlJIIP£Xt ZN
OVORTEX.F'I""
IIlUlCICII ( ,I~ @
~ . • Xo
OI""-£T.4~J
0AP(X · Z-
OVORTEXI-
1111 ~llI
l ~3o $ milo
uti ~Ill
:) x l~ m It
'oM r-------------------------,
••
,.
., ~r •.jl
,.
i
~ 41 •••
..
u
- - -- _ _ ~_ _ ~. p. 1, 500
0
0
..
14 ••
""I _ - - - - . - - - r.UI
..
~
..... _-_------~ r'UI
_ __-------r.tll
Fig. 6.- Optimum grinding degree % mesh -400 Fig. 7.- Water used in the grinding vs. hydro-
vs. hydrocyclones number at various cyclones number at various feed pulp
feed pulp densities, p. densities, p.
24
References
1.- Espinosa De L. Estudio de ingenieria concep-
tual para la instalaci6n de una planta de benefi
cio de los jales de Santa Julia. Unpublished re?
port, Cia. Real del Monte y Pachuca, Mexico,
'O~--------------------------------I 1982.
2.- Harris C. The Alyavdin Weibull chart in
batch comminution kinetics. Trans.lnst.Min.Mat.
•0
vol. 80, 1971. p. C42-4 .
~P'1:'" 3.- Harris C. Chakravarti A. The effect of time
on batch grinding. Powder Technology. vol. 4,
1970, p. 57-60.
.0 4.- Horst W. and Bassarear J. Use of ore grinda-
p. 1." bility technique to evaluate plant performance.
Trans. AIME. vol. 260, 1976. p. 340-351.
5.- Modelling and scale-up of hydrocyclone cla-
•0 ssifiers. A. Lynch and T. Rao. In: 11th Inst .
~P'160
Min. Process Congress. Cagliari, 1975.
6.- Plitt L. A mathematical model of the hydro-
cyclone classifiers. CIM Bulletin, vol. 69,
•0
1976. p. 114-123 .
7.- Finch J. and Matwijenko O. Individual mine-
ral behaviour in a closed grinding circuit. CIM
Bulletin. vol. 70, 1977. p. 72-164.
mROCYCLONES NUMBER
• oor-------------~~~~==~====~~P~
••~.7~O
'"
""
z~o
=
~
~P.' .••
;;;
~
zoo
~P • .,.60
...
J':\ p. 1.50
27
28
a close eye on the development of new technolo- This improvement is the consequence of nume-
gies with regard both to their technical deve- rous slight modifications in every field. In
lopment and their economic feasibility :c the following chapter we will try to focus on
- The new pyrometallurgical processes (QSL, some of them.
Kivcet, Isasmelt, ••• ).
- The new hydrometallurgical chloride processes
Development and evolution in the last ten years
The present status of these new processes The capacity increase and the coke ratio
and their heavy investment costs lead decrease can be explained by a better operating
Metaleurop to retain its confidence in the ratio of the furnace and a better metallurgical
traditional process and to develop the control, as described hereunder :
potential it still has.
Operating ratio
Ten years ago the four main problems affecting
THE EVOLUTION OF TRADITIONAL METALLURGICAL the operating ratio were :
PROCESSES FACED WITH ECONOMIC PROBLEMS 1. The lack of furnace feed: the sinter plant
was the bottle neck of the lead plant.
Lead bullion production and coke ratio 2. Difficulty of bottom furnacing in slag and
evolution matte tapping.
These two data seem to be a good example of 3. Frequent clean-outs of the top of the furna-
lead blast furnace progresses in the last ten ce to remove accretions.
years. They clearly show us that a great deal 4. Frequent failures, above all in the cooling
of improvement can be achieved with a system.
traditional process (data from one of our lead Today the capacity of the sinter plant has
smelters). been improved due to the investments carried
out on feed preparation and return sinter
treatment.
Lead Bullion output Concerning the bottom furnacing, the new conti-
(M.T/year) nuous slag tapping allows a better working ra-
120000 tio, and has a very positive effect on working
conditions.
The frequency of clean-outs decreased marke-
100000 dly with the better design of the off-gas cir-
cuit, with the better physical and chemical
feed control, and with the improvement of the
80000 feeding.
A new design of the cooling water jackets
Years has been adopted after simulation studies.
Accordingly, their life time has increased
83 85 strongly.
Fig. 1 Lead bullion output versus years
Metallurgical control
In a lead blast furnace, coke combustion is the
Coke ratio main factor.
(kg of coke/M.T. of lead bullion) The combustion of coke in front of the tuye-
240 res supplies most of the heat consumed in the
lead blast furnace. It sets some of the limi-
ting conditions for the heat exchange in the
220 shaft. On one hand, the heat delivered per mole
of blast air is determined by the combustion
ratio Oblast/C. On the other hand, the fraction
200 of this heat supply transferred to the burden
depends on the flame temperature of the
Years combustion.
180 Many fluctuations of the combustion parame-
83 85 87 89 91 ters have been observed at industrial level.
Fig. 2 Coke ratio versus years These variations were investigated in an expe-
rimental furnace, simulating, on a full scale
29
basis the region near one tuyere in the blast Table 1 Influence of the operating parameter
furnace. Cokes with different properties (size upon the combustion ratio
and reactivity) were compared under various
blowing conditions (flow rate and oxygen
content of the blast, tuyere diameter). Coke to Oxygen
Coke Blast burden content
These trials, as a whole, pointed out the size rate ratio of the
general features of the combustion. From the blast
I/ //
cooling box inwards, the coke stack exhibits
two regions corresponding to the two successi- Variation
ve steps of a "layered" combustion : of
- exothermic combustion by the air oxygen parameter
(C +0 2 ---» CO 2 ). This region extends hori-
zontally along the tuyere axis over 40 cm and
\~
vertically along the jacket over a few tens Influence/
of centimeters. This height varies widely upon
according to the permeability of the granular Oblast/C ~
bed,
- endothermic solution loss reaction (Boudouard
C + CO2 ---» 2CO). This reaction occurs in
the furnace center at the tuyere level, but This study led to slight modifications in the
at higher levels, in the front zone, near the industrial way of working, in particular :
jacket. - better utilization of air due to selected
This means that, at any level above the blast impulsion.
tuyeres, the reducing power of the gases is - better utilization of coke due to the quality
higher in the central zone than in the periphe- control and to the optimisation of feeding.
ric zone. Moreover, the furnace center is - control of the lead charge quality and
colder than the periphery due to the predomi- optimisation of feeding.
nance of endothermic reactions.
Sintering
The sintering machine, where the ore concen- With Without
trates are continuously sintered, consists of dilution dilution
two separate areas.
In the first one, known as the reaction
zone, the gases emitted are "rich" (e.g. their
S02 content reaches 4.5 % - 5.5 %). This allows Gas deli very 80 000 Nm3 /h 15 000 Nm3 /h
the sulphur dioxide to be converted directly Gas temperature 135 °c 650 °c
into sulphuric acid by the so-called "contact" Speed of gas in the 12.5 m/s 5.3 m/s
process. pipes
The second zone is where the sintered cake Dust concentration 16 g/Nm3 83 g/Nm3
is cooled by a forced ventilation system. The Composition of gaz
gases collected are "poor" since their content (in %)
is less than 2 %. They are often recycled in - CO in dry gas o 10
the circuit itself using a perfectly tried and - CO2 7 25
tested technology. Yet, due to thermal effect, - °2 16 0
the productivity of the process is reduced by - N2 76 63
approximately 10 % because either the chain - 5°2 0.3 1.5
must be slowed down, or the sinter lead content - Content of H20 2.5 4
must be reduced.
There are other solutions too, such as
31
Apparently, a 1 % S02 bearing gas obtained fic process that does not exist today as a
without diluting could be treated in an acid whole but whose components do exist in other
plant using new processes, but the catalysts industries.
used in this process cannot deal with CO and
dust. Sinter plant
Consequently, the process to be used to In the sinter plant, the treatment of rich
treat the blast furnace gases has to include gases is easy. For the poor gases there are
the three following steps : two alternative solutions :
- sealing the top of the blast furnace, - the first one is to recycle them in the cir-
- removing the CO by a post-combustion process, cuit itself. This has the disadvantage that,
- removing the dust. due to thermal effect, the productivity of
These 3 steps cannot be taken very easily the process is reduced. In this case it is
from a technical and economic point of view. better to invest in a traditional, compact
The main aspects are the following : acid plant instead of a larger sinter
- The sealing of the blast furnace top must not machine.
change the quality of the feeding of coke and - the second one is to mix the poor gases with
sinter. The accretions we had in the past the rich ones. In this case, you do not have
don't have to return with a new sealed top to increase the sinter machine size, but you
furnace. The channelling, with all its have to use a specific process which produces
drawbacks, will also have to be avoided. H2S0 4 from poor gases.
- The post-combustion of offgases could be done
before or after dedusting. In the first case, Blast furnace
post-combustion could be nearly autogenous, Various routes have been considered in the
but we have to avoid any formation of new study of the afterburning and dedusting of the
accretions and moreover the consequences on lead blast furnace offgases (preliminary treat-
decreased operating ratio and unhealthy and ment required before desulfurization in a H2S0 4
arduous tasks due to the clean-out of dust plant). The two following routes are today
accretions. In the second case, the cold considered as the most promising ones :
gases have to be heated in order to avoid any
additional energy consumption.
So, the real progress we obtained in the
past by suppressing unhealthy and arduous tasks
when removing accretions does not have to be
cancelled out by the necessary S02 removal.
This must be studied very carefully.
Therefore we have started off a study in
order to find the best compromise to avoid
these accretion difficulties which seem to be a
problem for all the processes (including the
new ones). We will give details about this
study in the following part.
or
Air quenching
b)
Af terburni ng
Pulverized In the present case, thanks to the preheating
water of the process gases, there is no consumption
Wet of auxiliary gas in the afterburning tower :
quenching - Flow rate of gases after wet
scrubbing ....•..•..............
- Temperature of preheating ••••••
SYNOPSIS
Ais elne offensive Antwort auf dIe zunehmenden Ende August 1990 begann der Probelauf mit Ma-
Forderungen fur den Umweltschutz entschied dIe terial, dIe offlzielle Inbetriebnahme war am
Metallgesellschaft AG im Sommer 1987, In ihrer 29.11. 1990.
Blelhutte In Stolberg dIe besonders emisslons- 1m Februar und Marz 1991 war eln geplanter Re-
trachtlgen Telle Slnter- und Schachtofenanlage paraturstliistand, in dem Mauerwerksschaden,
durch elne QSL-Anlage zu ersetzen. resultlerend aus dem stark schwankenden Inbe-
Das QSL-Projekt beinhaltet Insgesamt neben dem triebnahmebetrieb beseltlgt wurden. Glelchzel-
Reaktor mIt Materialzufuhr, Medlendosierung, tig wurden wahrend des Stliistandes Erkennt-
Abhltzekessel und HelBgasrelnlgung nlsse aus den ersten Betrlebsmonaten konstruk-
1) Anlagen, In denen Arsen und CadmIum auf tlv umgesetzt.
kurzestem Wege selektiv aus dem ProzeBgas Am 30.3. wurde dIe Anlage wIeder In Betrieb ge-
als handelsfahige Produkte ausgeschleust setzt. Ole Betrlebsergebnlsse entsprechen voll
':J8rden konnen, den Erwartungen. Es wurden Pb-Gehalte In der
2) elne neue Schwefelsauregewlnnungsanlage mIt Endschlacke am Reaktorstlch von bis zu 0,4 %
Oampfuberhitzer am Kontaktkessel, erreicht. Damlt wird das Verfahren auch fur
3) elne Energleruckgewlnnungsanlage, In der dIe Betreiber Interessant, dIe zur Verarmung von
Abwarme des QSL-Reaktors und der Schwefel- Schachtofenschlacken elnen Elektroofen betrel-
saureanlage gemeinsam fur den Werkselnsatz ben.
verstromt wlrd und 1m Mai 1991 wurden Versuche mit Erdgas als Re-
4) elne von der Fa. Linde elgens dafur errich- duktionsmittel erfolgreich abgeschlossen.
tete und betrlebene Luftzerlegungsanlage,
dIe aus ca. 100 m Entfernung von der Werks- QSL-TECHNOLOGIE
grenze dIe Belieferung mit O2 und N2 uber Das Verfahren ist benannt nach den Professoren
Rohrleltungen sicherstellt. Queneau und Schuhmann sowie der Lurgi.
Der Anlagenkomplex wird von einer zentralen Es stellt einen kontinulerlichen Direkt-Blei-
MeBwarte uber ein modernes ProzeBleitsystem ge- schmelzprozeB dar, bel dem zwei grundsatzllch
steuert und uberwacht. verschiedene pyrometallurglsche Reaktlonen in
MIt dem QSL-Projekt In Stolberg wird belsplel- elnem ReaktlonsgefaB ablaufen, Abb. 1.
haft demonstriert, Wle durch spezifische Ab- 1. Das autogene Rost-Reaktions-Schmelzen schwe-
stlmmungen der elnzelnen Nebenanlagen In Ihrem felhaltiger bleihaltlger Vorstoffe.
Verbund zum QSL-Reaktor neb en elner emissions- 2. Die karbothermische Reduktlon des In der
armen Verfahrenswelse eln OptImum an Energle- Schlacke eingebundenen Bleloxldes.
ausnutzung und elne abfallfrele Produktlon er- Bel dem QSL-Reaktor handelt es slch 1m Prlnzlp
relcht wud. um eln zyllndrlsches ReaktlonsgefaB, das um dIe
37
38
I_R_e_d_uk_t_io_n_sl_o_n_e_ __ _ __ II ______o_x_id_a_ti_o_ns_l_o_n_e_/
GrOnpellets
Konzentrat + Zuschlage
+ Flugstaub
Abgas
Schlacke
'HuO<:l>&avo 1110 ~g
"Wauer .20 kg
<4 Sou...,''' <00 kg
IlOO kg
-.
.500 kg
<4 K,," 100Jl;g
oS(.
Abga. .595 kg
w_ 615~U
8,,-ulius ,."tal/hUtt."
~ GmbH Vergleich konvMlione-ilto Bltoiverhullung zu QSL
8ltihUtt, 8insf,ldhamm.,-
Stalb.rg.Rhld.
Abb. 3 Lageplanauszug
Les fours tournants dans I'industrie
europeenne de la recuperation des
metaux non-ferreux
M. Rousseau
Stolberg Ingenieurberatung GmbH, Neuss, Allemagne
INTRODUCTION
279
280
TAB. I: PRINCIPAUX TYPES DE FOURS TOURNANTS UTILISES DANS LES INDUSTRIES DE RECUPERATION 1
Four 8 tambour court 1,08 1,5 cit =284 7820t 5000 8 20000 tpa
L =285
Four 8 tambour long 1,583,0 cit = 184 0,58 10 t 500 8 15000 tpa
L = 286
CHARGEMENT
BRULEUR
COULEE
COULEE
CHARGEMENT
FOUR DE RESSUAGE
AVANTAGES:
- Obtention de scories assez pauvres en m6taux et/ou absorbant certaines impuret6s (S,
CI)
- Possibilit6 d'obtenir plusieures phases liquides (p. ex. une matte, un spelss),
- Large 6ventall de granulom6trle des matieres premieres: des fines aux fragments,
INCONVENIENTS:
- Construction plus compliqu6e (que les fours s r6verberes) surtout pour les fours
basculants,
- Capacit6 par charge relativement faible (p. ex. mAme pour les fours Thomas, pas plus de
70 s 100 t)
282
1
Sens de
rotation
AVANTAGES
- Rentable pour oxygene bon marche (p. ex. prix du kg de fuel au moins deux fois celui
du m3 d'oxygene),
- Dans une installation existante: gaz avec plus long temps de sejour (cf rendement
thermique) et vitesse plus faible (cf poussieres),
INCONVENIENTS
les fours tournants possedent pour la plupart les fours rotatifs de dimenSion moyenne (p. ex.
un garnissage basique ou tout au moins peu acide (avec diam. 3 x 4 m) et equipes de brQleurs a air permeHent de
quelques exceptions encore, comme dans I'industrie du traiter environ de 50 1150 tpj de rebuts ou plus. De telles
cuivre: four Thomas, four de perchage). installations (voir Tab. IYJ sont decrites en detail dans la
litterature technique 1, 1 , et deviennent de plus en plus
le type de refractaire depend entre autre des complexes en ce qui concerne les equipements de pro-
metaux et oxydes presents et des temperatures de travail tection de I'environnement. Ceci est iIIustre par la
p. ex. pour les rebuts cuivreux 1100 a 1350 • C, plombi- consommation des fours (diam. 3,6 x 5 m) a I'usine de
feres 1000 I 1100 • C, mais moins de 1000 ·C (jUsqU'1 Villefranche (Ml!taleurop) de 45 kWh/t Pb pour la
750 - 850 • C) pour les rebuts alumineux. reduction contre 105 kWh!t Pb pour la fiHration, et par la
figure 3 (BSB-R, R.F.A.).
les plus basiques sont utilises pour la
recuperation du plomb: principalement magnesie (avec
plus de 80 % MgO) ou magnesie-chrome (de 55 % a 80 %
MgO et jusqu'a p. ex. 16 % Cr20 3). Diverses chamoHes
sont utilisees en plus pour la recuperation du cuivre et les fours a tambour et basculantl (dits fours
surtout de I'aluminium (p. ex. environ 40 % AI20 3). Thomas) sont tres frequemment utilises 1 p. ex. a11a
Norddeutsche Affinerie (N.A.), Hambourg, R.F.A., 6.
leur capacite varie de 8 t jusqu'a 20 a 40 t en moyenne
(c.l.d. env. diam. 3 x 6 m) ce qui peut correspondre a
15000 - 20000 tpa de matieres.
Fuellourd 110 kg/h 37 kg/h 238 kg/h ou 209 kg/h 114 kg/h
oU'moinsde ou moins de 55 - 65 kg/t ou 40 kg/t
80 kg/t charge 40 kglt charge matieres matieres
Porteur air oxygene pur 2500 Nm3 1900 Nm3 200 Nm3
d'oxygene air/h air/h °ih
Gazde 21000 Nm3/h 9000 Nm 3/h 50008 n.c. n.c.
procede 7000 Nm3/h
I · -·- · - · - · - - - - -.- .- ·l
I r--...:=-===-~-e.--·=-=-~_=r-l t .
I: ~'l L ~.<!
I I i ~ } "1 I ~ ~ I
MAZOUT !
I I ,·.tttf··t;~ I I
~ L .' ..' :.: :: .. : .;:. I ,iiI
!
~ .,'1, I . ,
~ ;~':~ :.:.~ r':':':':':':':;r~: :-:-:-:-:.:-:-~ 'C'
l1$J .~:~
MAZOUT
AIR
Fig, 3 Four rotatif court; arrangement typique du four, refroidisseur et filtre (Approx. 8 l'echelle)2
285
Aluminium TENDANCES FUTURES ET CONCLUSION
Les matieres fines et fragmentlles sont glln€!- De plus en plus de mlltaux non-ferreux doivent
ralement refondues sous couche de sels (KCI, NaCI, evtl. Atre rl!cupilres 8 partir de rebuts complexes et d'alliages.
NaF) dans des fours 8 tambour long (capacitll 5 8 20 t). Les nouvelles mltthodes de preparation permettent
Des fours moyens de 108 12 t (capacHlI de fUS~%n env. 1 jUsqU'8 un certain point une rl!cuperation selective, mais
t/h) peuvent traiter jUsqU'8 12000 tpa de rebuts . les fractions obtenues dOlve.". @tre retrait'es en g'neral
par voie pyromlltallurgique 2 .
Une amlliioration notable (voir Tab. V) peut @tre
aussi obtenue par prllchauffage de I'air. Les rllsuHats Les fours rotatlls gamls de r'fractaires sont
obtenus avec des brOleurs spl!ciaux (pourvus de particullerement bien adaptlls 8 ces travaux complexes
rllcupnteurs cnmiques coupllls), pour les fours de fusion/r8ductlon (voir Tab. II) ou diffllrentes matitres
circulalres de 12 t (capacitll de fusion env. 3,5 t/h) secondalres (assez fragment6es) peuvent @Ire traltees
d'Alumlnium Corp. (Dolgarr09, Royaume-Uni) sont grace a une large gamme de templlratures et
significatlfs 8 cet ltgard. Les con'r§ntrations en NOx sont d'atmospheres de four. Seuls des rebuts agrande valeur
par 18 rilduislbles de 70 8 80 % • L'enrlchissement en etJou volumineux peuvent justifier I'utlllsation d'autres
oxygene permet par contre d'augmenter notablement la fours (rllverberes, plus encore fours 8 induction).
capacltll de traltement des fours toumants: 'W ex.
+ 30 % pour une consomrnation de 20 m3 Oil AI . Pour les matieres mlltalliques 8 bas point de
fusion etJou relativement pures ou seule une refonte est
nllcessaire, les fours rotatifs de ressuage (8 chauffage
direct ou indirect) peuvent constituer une alternative
comme p. ex. pour des plaques d'accumulateurs au
TAB. V: plomb (Ies crasses et oxydes devralent Atre toutefois
FOUR POUR LA FUSION DES RE~~TS DE
retraitlls en fours classiques).
FABRICATION DE L'ALUMINIUM
Parametres BrOleur 8 air BrOleur 8 recu- Pour les matieres tres complexes etJou 8 valeur
perateur importante, des reacteurs fixes de type SIRO-
incorpore SMELT/ISASMELT devralent occuper prW,1lslvlment
des segments bien particuliers du marche ' , 2 .
Capaclte du
brOleur 3,9MW 2,3MW Toutefois les fours toumants sont restlls,
malgrll la concurrence de certains proc8dlls utilisant des
Temperature reacteurs spl!cialislls, un des outils essentiels du
recyclage des mlltaux non-ferreux. La grande flexibillte
* dans Ie four maxi 1100·C maxi 1100·C de ces fours ainsi que I'utilisation d'un air de combustion
enrlchi en oxygene etJou prechauffll ou bien de brOleurs
* des fumees maxi 1200· C maxi 200· C a oxygene, devraient assurer, pas seulement en Europe,
leur perennite pour de nombreuses annees encore.
* de I'airde
combustion 30·C maxi 1000· C
Consommation de
combustible 1500 kWIt (fuel) 920 kWIt (gaz
naturel)
286
Littt§rature
287
288
from the blast furnace is blown together copper anodes are refined electrolyti-
with alloy scrap in scrap metal conver- cally into cathode copper. This common
ters. During this process the slag kind of process shows that there are
arising is returned to the reduction considerable material recycles before
unit. The converter copper is refined the final end product cathode copper is
(Inc, resi-
dues: -- -
du"L~1 ashes,
drosses
WORKING SCHEME -HUTTENWERKE KA YS ER
sl.g •.
~'-'- l ~ !-'-l
shroddor-
... trriols.
[J
Cu-r~-HOLer18h
Olasl rurnnce.
,: I I
70%Cu
,.-._.
t..-------.-------l-----I----
BO%Cu 1 ' -'
slog
slag 9rnnuIoL~'
I I
gun _rL I _ _
£ironcc. elc.
I L _ _ _ _ llnc oxide
converlers
I
I
~-----slog--------~
_ _ _ _ _ _ _ _ tin-lead olloy
Cu-roflnlllg-I'--i
fIInLcrlol I .
bll<vr
cop IHt r
i
____.J'od
~ olocLrolytic
i reflnlng
_ _ _ Cu- ond Nt -
vitriol
slag
I Anode
s1 imQ
I Residues ________
Lo SlIclLcr
~ PreC10US lIIot.ols
plnnt - - ......... Precious IIIIQt.nls
flu~ dust
I
SOHed
production
scrap'"
- - - _ _ _ _ _ r _ _ _ _ _ _ _ _ _ -O-O Wirobar!O
slitg
row milLCTlols
products to ralo
Intrr.rdJnLr nrOdIJCL~
further in the anode furnace, and during reached. Besides energy requirements for
this operation copper refining scrap is melting and reduction these material
added - the slag from this process is recycles cause metal losses by passing
also returned to the blast furnace. The the units several times.
289
HGttenwerke Kayser has carried out For a blast furnace working effecti-
calculations, tests, and planning for vely on reduction work a preparation of
easier processing with material returns the material is necessary by mixing,
as low as possible. crushing, agglomeration, and sieving. If
the reduction work is done in the TBRC
In non-ferrous metallurgy the TBRC this is largely unnecessary. The mixing
(Top Blown Rotary Converter) has been is done by the TBRC's turning. Coarse
used in recent years. The TBRC is a material gets dissolved in the moving
rotating and tilting furnace with burn- melt, and fine material does not matter,
ing and blowing injector lances: it as there is not a feed column with gas
looks like a concrete mixer. flow. By using melted materials, as for
example, liquid slags the energy for
remelting is saved. The charge still can
be influenced while running by addition
/ __{Fi~d
1 Hygiene of flux and reduction agents. By this
the distribution of the elements accom-
panying copper can be influenced favour-
ably for the following process steps.
By working step by step, removing the
FeeiIJ metal under the slag layer, the elements
can be parted according to their reduc-
tion reaction. Moreover, low final
contents in the slags can be achieved
without producing high amounts of metal
with high iron content.
Pouring
Ladle Enclosure
All these processes are helped by the
intensive mixing during rotation, which
IWlIIlIllliF0, allows the fluid-solid and fluid-liquid
diffusion stages to be of minimal length,
producing rapid and complete reactions.
Fig. 2 - TBRC arrangement
For reduction work in the TBRC all
solid reduction agents can be used. For
carbon containing materials no special
requirements have to be met regarding
stability, piece size or ash content.
Therefore, the cheapest reducing mate-
In the blast furnace reactions are rials can be used. Also if using iron as
slow and incomplete due to an incomplete reduction agent one only has to take
mixing. Due to the excellent mixing care that the brick lining does not get
effect of the TBRC the diffusion paths damaged due to the piece size. Moreover,
have been reduced, therefore, the too big pieces unnecessarily prolong the
attainment of equilibrium is accele- reduction times.
rated.
All these facts show that regarding
Also, when using the TBRC for conver- reaction kinetics, requirements for feed
ter work and refining work an accelerat- material, and energy utilization of the
ed and complete reaction is achieved due heat contents of combustible feed mate-
to good mass transfer. rial as metallurgical reactor the TBRC
is superior to the blast furnace.
By using oxygen both for heating and
as reaction gas the required fuel and In oxidation work as necessary upon
the quantities of waste gas are reduced. converting and refining, one also has to
If the TBRe is working in a hood which point out the favourable reaction mecha-
closely surrounds it, the process waste nism. Due to the rotation of the reactor
gases as well as the secondary emissions there is always a good mixture of the
which occur upon feeding and emptying material, which leads to a uniform
can all be captured. This means a consi- concentration and new surfaces for the
derable reduction of environmental gas-melt-reaction. This is especially
pollution compared with the usual pro- important for reactions which do not
cessing. produce gases, as in the metallurgy of
copper secondary materials.
The plant's flexibility is consider-
ably improved by the fact that the TBRC By using oxygen there are high re-
can do the blast furnace's reduction action velocities and minimum losses of
work as well as the converter's and heat recorded, therefore, for processing
anode furnace's refining work, by carry- it is necessary to add only small
ing out the oxidation process in the amounts of energy besides the oxidation
same vessel. energy available. The oxygen necessary
for oxidation can be utilized optimally,
as the melt which wets the vessel's
TBRC -Plant
process diagrem
filter
plant
waste air from material preparation
water
cherging device
water to baUer
I'.)
(0
o
fume for
after burning waste heat gaa mixing tube process gas treatment
slag metal cham bar bailer chamber cooler filter plant
~
G.-J
291
In the early 1980s the stocks of dusts were consumed, and other
materials tested. This included leach residues, lead batteries TREATMENT OF LEAD-BEARING DUSTS
and concentrates. Copper scrap with high organic content was
also successfully smelted. Lead-bearing dusts from the copper converters and the clinker
furnace were in the 1960s sold to other lead producers. Penal-
In 1978 a Copper Kaldo Plant was built for smelting of copper ties for arsenic, antimony, cadmium and halides (table 1) soon
concentrates. It was found that large amounts of scrap could be made the sales unprofitable.
included in the charge, due to good heat efficiency and mixing
in the Kaldo
Table I. Compositions of materials treated in
In 1989 the Copper Kaldo was closed down because of lack of the Lead Kaldo Plant
concentrates and refinery capacity and the Lead Kaldo was
modified for smelting of lead concentrates.
Dust from Dust from Residue from
Significant for the Boliden Kaldo Processes are high produc-
tivity, low energy consumption, low emissions to environment copper zinc clinker zinc clincer
and superior in-plant hygiene. converters plant leaching
% % %
293
294
Plaut description Pelletized dust, granulated slag from the fuming furnace,
soda, lime and dross from the lead kettles were charged
The dusts were stored in a separate building and, via daybin batch wise to the furnace. Heat is supplied by an oxygen/oil
and chargebins, charged to the furnace by drag chains. Fluxes, burner. During the melting the lead sulphates decomposes to
iron and coke are charged from separate bins above the lead oxide and S02' After melting coke was charged conti-
furnace. nuously for the reduction of the lead oxide in the slag. When
a lead content of less than 1 % was reached, the slag was
The furnace is totally enclosed by a ventilated hood. Inside skimmed into ladles and transferred to the fuming plant. The
diameter of the furnace is 3,6 m and the length is 6,5 m. During lead was collected in a kettle. When 60-90 tons of lead were
operation the furnace inclination was first 22 degrees, later on collected in the kettle, the lead was pumped to the Kaldo
changed to 28 degrees to give larger working volume. Maxi- furnace forrefining. In pilot scale it was easy to get a selective
mum rotation speed is now 15 RPM. oxidation of tin to a dry dross that was pulled out of the furnace
before the oxidation was continued to produce a fluid arsenic-
Heat is supplied by an oxygen/oil burner in a watercooled antimony dross.
lance.
Process modification
The offgas hood is water cooled and reinforced by a basic
castable. The time and labour consuming oxidizing refining method
was discontinued quite soon. Instead iron was added to form
For the gas cleaning a Venturi scrubber is used, operating at a iron-arsenic-speiss. The fluid speiss was easy to skim off to
2000 mm pressure drop. Venturi water flows to a thickener for a ladle and could be discarded after solidification. Instead of
settling. The sludge is filtrated, and the filter cake dried in a oil collecting lead in a kettle and pumping it back, it was collected
heated drum dryer and then returned to the process. in the furnace and refined after 2-3 heats, then tapped in a
ladle.
Lead is cooled in ladles before pouring to kettles and cast into
2,5 ton blocks from the kettles. (Figure 1.)
FANS
-----@,
-==::a------'
to CENTRAL EFFWENT
WATER TIlEATIIENT PLANT
~bofden~
295
25 t lead
Smelting
11500 C
Coke
Iron cuttings
OiVoxygen OiVoxygen
Reduction Arsenic
11500 C removal
Slag t---I~ Speiss
'----,----'
oxygen
TIn drossing
6500 C
Tin dross
oxygen
Arsenic/
Antimony
drossing Arsenic/
7500 C Antimony
dross
Crude lead
296
Minor modifications of the slag composition made it possible Later it was found out that the iron could be added together
to operate without the expensive soda. (Table II) with the rest of the charge. Speiss was formed simultaneously
with the lead, slag and speiss were tapped together.
TableU. Products from the Lead ((aldo Plant
In 1982 processing ofleach residues from the leaching of zinc
clinker was started. No major changes in the process were
Slag Crude Lead Speiss needed.
TO CENTRAL EFFLUENT
WATER TIlEATIEHT PLANT
GboIden~
Boliden Metall has traditionally always processed all domestic In 1978 the Copper Kaldo Plant was started up. Ideas to produce
low-grade copper scrap. Organic material has to be removed blister copper were abandoned, because of problems to keep
from cables and electronic scrap before further processing. heat balance and a reasonable converting speed.
In the early 1970s the scrap was burned ina stationary Best concept was to smelt concentrates with oxygen to a rich
furnace. However, the removal of organics was not perfect. In (60-65 % Cu) matte. The matte would then be transferred to the
addition the dust and gas emissions were too large. The autho- Pierce-Smith converters for conversion to blister copper. Cop-
rities demanded new equipment with lower emissions to be per content of the slag was about 1 %, so the slag was discarded.
built. It was found out that the excess of heat created when smelting
with pure oxygen could be used to smelt large amounts (25 %
In this situation it was decided to test burning of scrap in the of the charge) of reverts and scrap. Good mixing by rotation
Lead Kaldo Plant. The results were very encouraging. The ensured that the metals dissolved in the matte. The cold charge
rotation of the furnace makes a perfect combustion possible. could be of any size as long as it would go through the furnace
The organics are burned with oxygen in the furnace and the mouth, and moisture content was not critical. Fine-grained
gases afterburned with secondary air in the water cooled hood concentrates or concentrates with high carbon content, that
at a high temperature. The highly effective Venturi takes care created problems when roasted and smelted in the electric
of the chlorine and dust and thus the emissions are well below furnace, could be used in the Kaldo with no problems.
the legislated values.
The slag can be discharged, and the metal is transferred in Copper scrap srpeJtjpa
molten state to the Pierce-Smith converters of the copper
smelter. Smelting of low-grade copper scrap is also possible in the
Kaldo. Such a process would use an oxygen/fuel-burner for
In 1982 the furnace hood was extended to accommodate a fully smelting and an air/oxygen lance for converting of the "black
enclosed skip hoist for charging of scrap to the furnace. copper". The metallurgy is described in ref.2.
References
1. H.I. Elvander; The Boliden Lead Process, Symposium
on "Pyrometallurgical Processes in Non-Ferrous
Metallurgy", Pittsburgh 1965.
2. D.S Flett et.al "Oxygen refining of black copper in
a top blown rotary furnace". Copper 87 vol
4:Pyrometallurgy of Copper page 425-439.
Minor Metals
Purity and long-term stability of
8-hydroxyquinoline-based metal
extractants
G. Haesebroek
MHO-a Division of ACEC-Union Miniere, Research Department, Olen, Belgium
SUMMARY been shown in the early seventies 1. Their usefulness for the
8-Hydroxyquinoline derivatives such as Kelex ® 100 from recovery of Gallium from Bayer process aluminate solutions
Schering AG are well known commercial extractants for the was demonstrated in the late seventies 2 and industrial
recovery of Gallium from Bayer process aluminate solutions. application followed soon.
Despite their high extraction power and outstanding Despite these encouraging results, the sensitivity of this
selectivities, they suffer from a limited resistance to kind of molecule to oxidation when contacted with strong
oxidation, which in the long run, leads to diminishing physical caustic solutions limited further developments.
and chemical performances. Inert gas blanketing and reduction of the double bond in
In 1985, MHO set up an extensive research program the unsaturated side chain by hydrogenation both help to limit
aimed at improving the long term stability of Kelex ® 100 when the oxidation of 8-Hydroxyquinoline derivatives 3.
contacted with aqueous caustic solutions. Even with both these improvements, the oxidation leads
The study showed that an increase of the active to unfavourable ratios of active compound to oxidation products
component content of the extractant results in a substantial in the solvent phase after a few years of industrial use. This is
improvement of the resistance of Kelex ® 100 against due to the high solubility of the oxidation products in the organic
oxidation, thus avoiding pronounced deleterious effects on the solvent.
composition and performances of the Kelex ® 100 containing Commercial 8-Hydroxyquinoline derivatives are
solvent when used in contact with concentrated aqueous caustic relatively impure products. Kelex ® 100 from Schering contains
solutions in the presence of air. only some 75 vol % of active 8-Hydroxyquinoline derivatives 4.
Eventually, this work led to the development of a new MHO and Schering were both interested in increasing the
Kelex ® 100 type (named Kelex ® 100 S) which is now stability of such compounds. A joint research program started
commercially available. in 1985. The key results of this study are reported in this
paper.
INTRODUCTION
Industrial solvent extraction experience proved that the long EXPERIMENTAL
term stability of the solvent phase is of primary importance to Long term stability tests
control the process economics. A poor stability increases These tests consisted of stirring equal volumes of aqueous and
reagent consumption. In most cases the degradation products organic phases together over a long period of time ranging from
also tends to accumulate in the organic phase which leads to 1000 up to 3000 hours.
depressed physical and chemical performances, decreased At different time intervals, solvent samples were taken
equipment capacity and eventually lower profitability. and analysed for their total and active extractant
This is why process engineers try to avoid using concentrations (Gas chromatography (GC) analysis and Cu
extractants with limited chemical stability. loading capacity measurement according to Schering 's standard
However less stable products are sometimes used because of procedures).
other unique properties. Commercial 8-Hydroxyquinoline Temperature, volumes and gas flow rates above the
derivatives are good examples of such extractants. mixture were monitored and controlled.
Their outstanding metal chelation properties have already
301
302
Organic phase (Solvent) RESULTS
Two different solvents were studied. They contained 10 vol % The stability of the extractants is measured by following the Cu
of Kelex <® 100 (as delivered), 25 vol % of Isodecanol and 65 loading capacity of the extractant-containing solvent as a
vol % of Escaid 110 (from EXXON Chemicals). function of the total contact time between aqueous and organic
These solvents differed only in the purity of the Kelex @ phases.
100 component. Two qualities prepared by Schering were Both solvents were prepared with equal concentrations of
tested. Their properties are given in table I. extractant (Kelex ® 100). Due to the different purities of the
extractants, the initial active component concentration of the
Table I considered solvents slightly differs (7,7 and 9,2 vol %).
Therefore the active component concentration of both the
Kelex ® 100 type Active material Cu loading capacity solvents were normalised to 100 at time 0 in the following
(extractant) content (l)-vol% (g Cu/10g Kelex) graphs (fig. 1 - 3). This must be taken into account in the
further discussions.
(3) Improved purity. In the chosen operating conditions, the loss of activity of
the reagent is almost twice as much with the common 77 %
Aqueous solutions of NaOH, prepared from demineralized water It must be emphasized that these tests have been
and a technical grade 50% NaOH solution, with concentrations developed in order to obtain substantial degradation rates.
varying from 40 to 100 g NaOH per litre. Obviously, a direct extrapolation to industrial practice is not
straight forward. In continuous operations, the solvent phase is
Beside the purity of the Kelex ® 100 extractant, we also looked flow rate).
at the influence of the alkalinity of the caustic aqueous solution Furthermore, in each of its cycles, the solvent is
and of the oxygen content of the gaseous atmosphere above the intimately contacted with the caustic solutions only for limited
aqueous/organic mixture on the stability of Kelex @ 100. periods of time in the mixers of the sections where caustic
% ACTIVE
NaOH : 60 gil
roAPOI'BIT
COOlENT Temperature: 50°C
100 Gaseous atmosphere: air
90
80
60
CONTACTllME
(HOURS)
~+---~--~.---~---.----~---r--~----~--~~~
o 200 400 600 800 1000
solutions are added. The degradation rate in industrial plants limited. Further increase of the alkalinity accelerates the
would therefore be much lower than the rate measured in our degradation process. At 100 gil (2,5 N) NaOH, high oxidation
standard test. rates are measured. These results confirm that the purification
Alkalinity of the caustic aQueous phase agents. They also show that a careful choice of the operating
Figure 2 shows that at concentrations to 60 gIl (1,5 N) sodium conditions is important for maximizing the performances of
% ACTIVE
KELEX 100 S : 90+ %
CQ\PONENT
Temperature: 50°C
CCNTENT
Gaseous atmosphere: air
100
90
80
70
• NaOH 40g11
60 --G-- NaOH 60g/l
NaOH 80g/l
..
--0--
NaOH 100g/l
50
Oxygen content of the gaseous atmosphere the O2 concentration below 5 vol % 02- This is very interesting,
Figure 3 shows that a reduction in the oxygen content of the because it demonstrates that it is not necessary to make the
gaseous atmosphere above the solvent aqueous phase mixture equipment completely gastight to obtain satisfactory gains. This
leads to a pronounced decrease of the oxidation rate. It is technique can therefore easily be used in industrial plants at
important to note that the oxidation rate is quite independent of low costs.
% ACTIVE
C(JvlPGlENT KELEX 100 S : 90 + %
CCNTENT Temperature: 50°C
100~~~~~~~~====__~________~______~N:a:O:H~:~8:0~9:/I====~
90
80
70
Gaseous atmosphere
- - 0 - - 0.5 % O2
60
• 5 % O2
--0-- 20 % O2
50
CONTACT TIME
40+-__~~__~____~____~__~____- r____~__~~__~~~OU~R=S~)
o 200 400 600 800 1000
Figure 3: Effect of the oxygen content of the gaseous atmosphere above the solvent!
aqueous caustic solution mixture on the stability of Kelex ® 100 90+ %.
304
CONClUSIONS
The long term stability of 8-Hydroxyquinoline derivatives
contacted with aqueous caustic solutions can be dramatically
increased by increasing its purity. Schering AG was able to
reduce the content of inactive components in its Kelex @ 100
reagent from ± 23 vol % down to ± 8 vol %. The purer product,
now commercially available as Kelex @ 100 S, proved to have a
better oxidation resistance in caustic solvent extraction
conditions. Kelex @ 100 S has also other advantages 5 as a
metal extractant. We conclude that it is a very valuable
alternative for other commercially available 8-
Hydroxyquinoline derivatives.
ACKNOM.EDGMENT
The author wishes to thank the MHO-Management for having
given the permission to publish this paper and other people of
MHO involved in this study and in the preparation of this paper.
REFEFB\CES
1. W.M. Budde & al - USP 3637711 - 25/1172 - Beta-
alkenyl substituted 8-Hydroxyquinolines.
2. A. Leveque, J. Helgorsky - ISEC '77 - Proceedings -
Vol. 2 - The recovery of Gallium from Bayer Process Aluminate
Solutions by Liquid-Liquid Extraction - p. 439 - 442. •
3. D.L. Gefvert, H.J. Richards - WO 82/01369 - 29/4/82
Stabilization of substituted 8-Hydroxyquinoline
Hydrometallurgical reagents.
4. G.P. Demopoulos, P.A. Distin - Hydrometallurgy 1983, 11
On the structure and composition of Kelex @ 100 - p.389 - 396.
5. A. De Schepper, G. Haesebroek, A. Van Peteghem - USP
4942023 - 1717/90 - MHO a division of ACEC-Union MiniEHe -
Metal extraction process with substituted 8-Hydroxyquinolines.
Improved technology for In, Ge and Ga
recovery in an electrolytic zinc plant
Tian Runcang
Guangzhou Research Institute of Non-ferrous Meta/s, Guangzhou, China
305
306
new technology of recovering In. Ge.Ga in Fig.1. AN64 concentration dependency on
electrolytic zinc plants. special Ge-stripping
research for stripping Ge has been made.
aiming at studying and finding out a new
Ge-stripping agent. This agent should be It can be seen from Fig.l that when AN-64
high in stripping rate. non-toxic. low concentration is in the range of 25-35g/l.
corrosivity. cheap and easy to get as well a peak value will appear in the curve of
as no harmful effects on the next proce- Ge-stripping.When the concentration is
dure. After selection and testing. AN-64 27g/l. the stripping rate of Ge reaches
has been found to meet all the above- its maximum -- 97%; when the concentra-
mentioned requirements. Not only can it tion is less than 18g/I, the stripping
overcome the drawbacks that HF or rate is lower -- less than 75%; when the
NH3• HzO + (NH-'f)..aSO~ possess as stripping concentration is higher than 44g/l. not
agents but it also retains their advan- only does the stripping rate reduce but
tages. emulsification app(lars. Ther·efor·(l. it is
recommendable to take the concentration
of AN-64 as 27g/l.
EXPERIMENTAL (b) Effect of stripping-equilibrium
The experiment was conducted in a 60 ml time on Ge-stripping
glass separatory funnel. The Ge-stripping It is shown in rig.2 that stripping
agent used was AN-64. Ge-rich organic equilibrium can be reached within
phase was prepared by Ge-bearing feed 3 min. In order to ensure enough time for
(provided by smelter) extracting with 20% the reaction to r(lach equilibrium.the
D1EHPA + 1.25% YWIOO-kerosine. After time can be takerl as 5min.
mixing for a given time in an oscillator. Testing conditions:
it was settled to separate into two Gu concn. in Gu-rich organic phase
layers. The aqueous phase was run out to 0.48g/1
analyze its Ge concentration. According ratio O:A=2
to analytical results. stripping ratio AN-64 concn. 27g/1
O/A and Ge concentration in Ge-rich temp. room temp.
organic phase. the stripping rate for Ge
could be calculated. The benzfluorenone
.
~
stripping 2 3 4 5 6 7
Testing conditions: stripping equilibrium time. min.
Ge-concn. in Ge-r ich organ ic phase
0.34g/I, Fig.2 Equilibrium time versus Ge-
ratio O:A=3 stripping rate
mixing time 5 .. in.
temp. room temp.
(c) Effect of phase ratio (O/A) on Ge-
/-
stripping
'0 100 )t ,. "....!
Testing conditions:
..., 80
Q)
Ge concn. in Ge-rich organic phase
10 60
I- 0.47g/1
bD 40
c AN-64 concn. 27g/1
Q. 20 equilibrium time 5min.
Q.
0 temp. room temp.
...,
I-
rJl 10 20 30 40 50 60 70 80
concn. of AN-64. gil
307
Results in Fig.4 indicates that at
M
:5! different phase ratios (the other strip-
,!l100 A 2.5 .~
_ ping conditons are fixed), with the
-
f 80 ~~_::a.I2. 0 ~ enhancement of Ge concentration in Ge-rich
c 60
IIG
1.5 8: organic phase, Curve A falls gradually
a. 40
....
1.0 "- and Curve B drops more rapidly, illus-
..
a.
"- 20 0.5 tl trating that by means of adjusting the
.5 phase ratio or controlling Ge-concen-
CIl 0
Q)
t::I 2 3 4 5 tration in organic phase, a high stripping
phase ratio,O/A rate can be achieved and a given Ge-concn.
can be ensured. For example, in one
Fig.3 Relation between phase ratio and ~
~ Chinese smelter, the Ge concentration in
Ge-stripping the Ge-rich organic phase is usually
A - stripping curve around 0.42-0.52g/1, the phase ratio can
B - Ge-concn. in stripping liquid be chosen as 4; whereas for some other
versus phase ratio smelters, Ge concentration in Ge-rich
organic phase is higher(1.8-2.3g/l). then
the phase ratio can be taken as 1.
It is shown in Fig.4 that when the
stripping ratio Q:A=l to 4, Ge-stripping Test on circulating the organic phase
rates are all as high as 98% or more and After the organic phase is extracted and
Ge-concn. in the stripping liquid is stripped. whether it can be recycled in
increased with the increase of ratio. use or not determines the feasibility of
However, when the phase ratio reaches applying AN-64 as a new stripping agent.
over 4, a downward inflection point For this reason, circulation of the
appears in Curve A, which indicates organic phase was examined in the test
reduction of stripping rates; while Curve and the results are shown in Fig.5.
B tends to be flat, indicating saturation Testing conditions:
of Ge in the stripping liquid. Hence, aqueous phase real plant feed
when the concn. of Ge-rich organic phase Ge-concn. 0.083g/l
is around 0.47g1l. it is suitable to organic phase 20% D2EHPA + 1.25% YVI00 -
choose 4 as phase ratio. kerosine (supplement YW-IOO each tille)
(d) Effect of Ge-concn. in Ge-rich ratio Q:A=I:5
organic phase on Ge-stripping equilibrium time 5min
Test conditions: temp. room temp.
..
M
AN-64 concn. 27g/1 M
equilibrium time 5.in. Q)
100 ....110
Q)
~ 10
..
u
110 40
L.
~ 20 -
40 a.Co
20 .;.>"-
....0 <I> III
..
80 Q) 0 CIl
t::I
Q)
60 2 3 4 5 6 7 8 t::I
110
L. 40 circulation time of organic phase
110
c 20
0 Fig. 5 Circulation time of organic phase
..
a.
a.
0.5 1.0 1.5 2.0 2.5 versus Ge-extraction and Ge-stripping
"-
CIl Ge concn. in organ i c phase, gil A - Ge extraction curve
B - Ge stripping curve
Fig.4 Relation between Ge concn. in Ge-
rich organic phase and Ge-stripping rate
A - ratio O:A=1 Results from Fig.5 show that Ge-extrac-
B - ratio Q:A=4 tion rate for each circulation is in the
range of 94 - 96%(average of 95%) and
Ge-stripping rate is about 96 - 98%
(average of 97%). After 11 times of
308
extraction - stripping circulation. both In consideration of the economic effect.
extraction rate and stripping, rate of Ge it is possible that some smelters may be
fl uctuated within the allow range. not interested in producing GeO~ of high
showing any tendency of declination. purity. So the second technological route
for recovery of Ge has been studied.
namely. the evaporating enrichment --
Technological route for recovery of Ge chlorinaing distillation. In this process.
from stripping solution enriched solution. obtained by evapora-
(a) Evaporation enrichment -- hydrolysis tion. and concentrated hydrochloric acid
to precipitate Ge (process 1) are added separately into the chlorinating
When AN-64 is used as the stripping agent. distillation flask. After mixing. the
the Ge concentration in the stripping total acidity of the solution is at the
solution usually averages I.88g/1. If Ge range of 8 - ION. The solution is put on
is precipitated by hydrolysis directly an electric hot plate to be heated. when
from this stripping solution. the hydro- the temperature reaches 78 - 82°C. Ge
lysis rate will not be high since Ge will be distilled out in the form of
concentration in the solution is not high GeCI+. which is then absorbed by water
enough. Generally. evaporation enrichment and hydrolyzed to GeO~ of higher purity.
is applied to raise the Ge concentration When the temperature is suddenly raised
in the stripping solution. But if evapo- to 100·C. the heating and distillation
ration causes any loss of Ge. this process is stopped. The residual liquid from
should not be used. For this reason. distillaion is analyzed. Ge concentration
special tests have been conducted to in which is less than O.Olg/l. Recovery
confirm that evaporation enrichment will of this method is over 97~ (from stripping
not cause any Ge loss~ solution to Ge0 2 of higher purity).
Normally. Ge concentration in the
enriched solution after evaporation is Analysis of economic and technical indexes
9.3g/1. In order to recover Ge from the Comparison between AN-64 as stripping
solution. 50~ NH3·H~0 is added for neutra- agent (is new process) and HF or NH3'H?0
lization. the end point of which is con- + (NH",,)~SOJf as stripping agent is shown
trolled between pH 3.3 - 9.1. Then the in Table 1.
neutralized solution is filtered and the The data listed in Table 1 show that
precipitate is washed. Finally. the when AN-64 is taken as stripping agent.
product is dried and crude GeO~ containing the average Ge-stripping rate is 97~.
25~ or more Ge is obtained. Recovery of approaching the index of HF -- 98~. but
this process is 96.3~ (from stripping 37 - 4I~ higher than the rate with
solution to crude GeO~). NHjH~O plus (NH~)~SO~. Moreover. when
(b) Evaporating enrichment -- chlori- AN-64 is applied. the unit consumption
nating distillation (process 2) of raw materials is the lowest. 313.7
IlNB/Kg Ge.
Table 1 Comparison of economic and
technical indexes between the new process
and the old one
-;;~~~;;--;t;i;;i~;-;;~~t--G~=;t;i;;i~;------G~-------;~~~;~;;-i----~~it-~~~;;;;;io~-
rate ~ hydrolysis chlorinating of raw materials
distillation RKB/Kg Ge
--~~;-------AH=64-------------97-----------94~8-----------97-----------313~7---------
--~ld--------HF---------------98------------/;------------97-----------415~2---------
10--0-0-:-("0-,-50---56=60----------55-------------55-----------341~O---------
-------------------------------------------------------------------------------------
* Hydrolysis cannot be used here.
309
CONCLUSION
It Is a significant ilProvelent on the
original process to use AN-64 as the
stripping agent, which not only can
guarantee the feasibility of the new
process but also reduce the cost of GeO~
production. Not only are the drawbacks
existing in each of the two stripping
agents used in the original process
over cOle but it retains all the advan-
tages of the other techniques. In addi-
tion, two technological routes for
recovering Ge frol the stripping solution
are offered so that different slelters
can decide, depending on their own condi-
tions.
References
1. Tian Runcang, International Mineral
Processing and Extractive Metallurgy.
Kunlng, China, 1984, 10, P. 615.
2. Li Shuzhen, Tian Runcang, Chen
Xinglong, Rare Metals, 1988, 3, P. 299.
3. Chen Xinglong, Tian Runcang, Li
Shuzhen, Metallurgy of Non-ferrous Metals,
1990, 5, P. 9.
Production of molybdic trioxide by
high-temperature oxidation of
molybdenite in a cyclone reactor
Igor Wilkomirsky
Department of Metallurgical Engineering, University of Concepcion, Concepcion, Chile
11,820
Log PmoO = - --,---- - 7.04 log T + 30.44(mm Hg)
3
By performing the reaction of oxidation above the
boiling point of Mo0 3, the formation of lower
molybdenum oxides, such as Mo0 2 , M?4011 and
Mo0 26 can be minimized, since no ?1ffus10n control
is established through the sol1d Mo0 3 product
laver and interparticles that could d1splace the
thermodynamic equilibri~m from Mo0 1 to other
lower oxides, as occurs 1n mult1ple hearth fur -
naces, and to a lesser extent, in fluidized bed
reactors. Figure 2: Bench scale cyclonic reactor
1. inlet pipe. 2. cyclonic reactor
The reaction rate constant follows approxi- 3. oxygen inlet. 4. condenser. 5. low speed
mately an exponential function with temperature, dust collector. 6. Product cyclone.
as can be observed in Figure 1. At 500 QC it is
about 1.1 x 10-4 cm/seg! iDcreasinq to 6.2 x
10-4 at 600 QC and to 8 x 10-2 at 1100 QC. This The experimental bench scale cyclone used
represents an oxidation rate over 120 times in some of the early versions is shows in Figure
faster for a cyclone reactor operating at 1100 QC 2. The reactor has a long barrel with a single
compared with a fluidized bed reactor at 600 QC(4). tangential entry where a suspension of dry
molybdenite and oxygen enriched air was injected.
At temperatures prevai 1ing in multiple Additional inlets allow the introduction of
hearth furnaces and fluidized bed reactors it oxygen or air along the cyclone barrel to com-
appears that the initial reaction control mecha plete the reaction. Under ideal conditions, a
nism is the chemical reaction at the interface: molten layer of CuO-5i0 2 slag should form,
followed bv a mixed control, which is progressive- discharging through the apex of the cyclone
ly replaced by a diffusional control through the while the gaseous Mo03 together with the gaseous
Mo0 3 layer growing outward, products (502 and 02J exit the reactor via the
vortex to be cooled 1n a water jacketed condenser.
For reactions temperatures above the melting Fine M003 crystals gro.ving along the condenser \\ere continu-
point of Mo0 1 , the dense molten p.hase coats the ously removed by means of a low rotating device.
unreacted core of M05 decreas i ng even further
the diffusion of oxyge~ to the reacting interface. The sol id Mo0 3 was further separated from the
As temperature increases over 800 QC, volatiliza- gases in a conventional cyclone.
tion kinetics of Mo0 3 becomes faster up to 1195 QC,
where the boiling point is reached. At th1s
temperature chemical reaction is extremely fast, Due to the constra i ns imposed by the
and the control of the reaction becomes predomi- construction materials of the reactor, the
nantly the oxygen transfer to the reacting surface upper reaction temperature was about 1150 QC,
of M05 2. just below the boiling point of Mo0 3.
The combined heat, mass and momentum trans- Mol ybden i te concentrates used ranged from
fer in cyclone or flash reactors plays a complex 0.5 up to 2.5% copper (Table I), although the
interaction at temperatures above 1195 QC. Tempe- concentrate u';ed in bench test contains below
rature instability at particle level due to the 0.5% copper on average.
unsteady heat generation and dissipation rates
increases particle temperature up to 600-700 QC Table I. Molybdenite Concentrates Used (Wt.~)
above the nominal reaction temperature, the
condition under which a combined heat and mass Mo 5 Cu Fe Ca Insol.
control is establ ished. In a cyclone reactor an
additional resistance mechanism appears also to 52-56 25-29 0.5-2.5 0.8-1.3 0.1-0.2 5.8
exist due to the molten slag phase that can
entrap reacting molybdenite particles. This
phenomenon it is not present in a flash-type In operating the bench scale units, several
reactor. problems arise, mainly in the cyclone reactor
itsel f due to the formation of accretions and
blockage of the feed inlet. The high viscosity
EXPERIMENTAL RESULTS of the molten phase and its slow movements
along the reactor wall did not allow a continuous
Early attempts in 1975 at the University operation for more than a relatively short
of Concepci6n (3) indicate that a cyclone reactor period of time.
could be an alternative to conventional roasting
technology. These initial studies and later
developments showed that a high grade Mo0 3 could Condensed Mo0 3 as a very fine powder
be produced in this type of reactor. (-400 mesh) could contain variable amounts of
313
-- --
~100·r---------------------------------~
impurities as well as an un reacted MoS 2. Purity ~
varies from 85 up to + 99% of MoO , depending on
several operational variables suc~ as gas/sol id 0'o 90
ratio, gas velocity inside the reactor, reaction ~
:g 70
>-
"0
E
100 a 60
99
~
-
u
::>
"0
98
700 800 900 1000 1100
Reaction temperature I ·C
1200 1300 1400
~ 97
.5 Figure 4:Recovery of molybdenum as Mo0 3 from
.., MoS 2 as a function of temperature .
0
0 96
~
1.0 r!z:~~-",--"""""""'''''''--.,..---lr---,
95
...
0
I-
0.9
u
;S
94 IX 08
0
"
u 0.1
93 IL
0
z
S! 0.&
...u
I-
92
900 1000 1100 1200 1300 IX 0·5
IL
Reaction tl'mpera turl'. ·C
0.4 REACTION TEMPERATURE
d 650 • C
Figure 3: Influence of the reaction temperature
..•
0 700' t
on tne purity of Mo0 3 obtained. (U 790' C
900' C
0 1000 0 C
+ 1050 • C
CVtLO~[
AEACIOA
/ BLOw BACK
A[ACIION
/SHArI
GAS
~(AUBBlA
AIR . OXYG[~
RECOVERY
BIBLIOGRAPHY
1. Wilkomirsky I. ,Petit-Laurent H. and Reghezza
A., "Production of high purity molydic oxide
from molybdenite". Chilean Pat. 37.285, (28
Aug. 1990).
2. Kelly, K.K., U.S. Bureau of Mines Bull.383,
(1935).
3. "Direct production of high purity molybdic
oxide from molybdenite". University of Con-
cepcion, Resea~ch Grant NQ 2.10.14 (1975).
4. Wilkomirsky I., Brimacombe J.K. and Watkinson
P."Kinetics of the oxidation of molybdenite"
Trans.IMM, Vol.86, NQ 3, p C16-C22,(1977).
ACKNOWLEDGEMENTS
The authors wish to thank H.Petit-Laurent,
A.Reghezza,M.E. Alarc6n, Dr.R.Padilla and J.Gaci-
tua for their collaboration. Thanks are also ex-
tended to FONDECYT, Chile, for the Research Grant
NQ 89-0723.
Trends in the development of
processes for recovery of rare metals
A. A. Titov
I. F. Poletayev
v. A. Krokhin
A. A. Schelkonogov
State Institutefor Rare Metals, Moscow, USSR
dioxide gas and carbon monoxide are respect, since this reactant-less waste-
directed in the gaseous state to a free process makes it possible to sepa-
condensation system for fractional rate Nb pentachlorides from Ta penta-
condensation. chlorides and to carry out their
To purify chlorides of desired metal subsequent full purification. 4
values from i~on and aluminium use was Practical implementation of this
made of the salt purification method separation method is limited by the
based on the difference in the thermal highly corrosive working medium, which
stabilities of complex chlorides of leads to considerable corrosion of metal
niobium, tantalum, titanium, iron and values and makes impossible the use of
aluminium. tray-type rectification columns.
As iron and aluminium chlorides react These technical difficulties were
with potassium and sodium chlorides, overcome by using packed-type rectific-
thermally stable compounds are formed - ation columns with a diameter of 300 mm.
for instance, K(Na)Fe(AI)CI 4 and, as a The main working surfaces of such
result, a gaseous mixture of refractory columns are protected with a special
metals chlorides is purified from iron coating.4 The operating parameters of
aluminium. condensation are selected so as to
In the process of chlorination of enable the discharge of distillate (i.e.
titanium niobates of rare-earth metals, Nb and Ta pentachlorides) from the
apart from pentachloride, niobium rectification column in the form of
oxychlorides are also formed. This fact powder.
is demonstrated by thermodynamic calcul- The essential novelty of the above-
ations and tahes place in practice. 2 At described process flow-sheet for separa-
o
a temperature of 1000 C the fraction of tion of niobium and tantalum resides in
niobium oxychloride reaches 50%. the use of rectification, since hitherto
For subseouent separation and deep extraction was used for this purpose.
purification of Nb and Ta chlorides the Effective methods have been developed
need arises for conversion of Nb to enable Nb and Ta pentachlorides to be
oxychloride to Nb pentachloride, the converted to such chemical compounds as
latter being a sufficiently convenient find wide acceptance in engineering,
chemical compound to be amenable to such as niobium and tantalum oxides,
purification by physical and chemical lithium niobates and tantalates, lead
methods. This conversion process is niobates, lead magnesia niobates niobium
based on a chemical reaction between Nb and tantalum carbides. 5
oxychloride and carbon tetrachloride. The majority of the above numerated
Consequently, to produce Nb and Ta in chemical compounds are synthesized via
the form of pentachlorides, use is made intermediate compounds - chloroxoethyla-
of three main chemical reactions, namely tes 4 - in accordance with the following
flowsheet:
2Nb 20 5 + 6Cl 2 + 3C --- 4NbOCl3 + 3C02 Heat _ Oxides,
Chloroxoethylates
Fe(AI)CI 3 + KCI --+ KFe(AI)CI 4 treatment niobates,
tantalates,
2NbOCl 3 + CCl 4 - 2NbCl 5 + CO 2 carbides.
The chlorination approach to proces- The above products contain impurities
sing titanium niobates represents a more in amounts ranging from 10-2 to 10- 5wt%.
rational approach to the problem of Another field of application for
NbiTa separation. The rectification chlorination techniques for processing
method is particularly promising in this titanoniobates of rare-earth metals is
317
envisaged for the first stage of the Processing secondary raw materials
leaching treatment of rare-earth metals Recovery of Nb and Ta from secondary raw
with nitric acid. A solid residue resul- materials constitutes an important sour-
ting from the leaching treatment ce of these metals. Considerable quanti-
contains Ta, Nb and Ti, To recover these ties of Nb are contained in the wastes
elements the residue is subjected to resulting from the production of
chlorination. This treatment offers the electrolytic capaCitors, superconducting
advantages of lower chlorine and reduct- alloys, hard-alloy products, niobates of
ant consumption rates and a lower dis- alkali metals and alkaline-earth metals.
charge of chlorine containing effluents. To process spent niobium capacitors a
Of basic importance is the fact that method has been developed based on a
the chlorination temperature has been reaction between the material of spent
reduced from 1000 0 C to 850oC, the lower capacitors and chlorine, in an aqueous
process temperature giving considerably medium, at a pH 6, in the presence of
milder requirements for the material variable-valency metal chlorides.
used for manufacturing the chlorination Impurities (iron, copper, nickel, manga-
reactor. Its service life, although nese) are solubilized, while niobium and
operating exposed to attack by corrosive tantalum do not react with chlorine and
molten media at 850°C, is prolonged by a remain in a solid residue. The process
factor of three. of chlorinating niobium-containing
Rare-earth metals are concentrated in wastes is a multi stage technique,
nitric acid solutions from which they variable-valency iron ions serving as
are separated and purified by extraction chlorine carriers. It is this fact that
Fig.1 illustrates a promising method explains the possibility to chlorinate
for the recovery of Nb, Ta and rare- copper, nickel and iron at relatively
-earth metals from are concentrates. low temperatures. The above process
Titanoniobates of rare-earth (re)
et Is
~
,
r-------------------------__
Nitric-acid solutions ~ Solid residue
of re metals (Nb, Ta, Ti)
,
Separation by extraction Chlorination
Condensation
~
Precipitation of carbonates
and oxalates
~
Heat treatment Rectification
\
Oxides of re metals Hydrolysis ~ ~assisted
~ ohemical decom-
~ position
proceeds in accordance with the follow- The Nb- and Ti-containing residue, after
ing equations: hydro chlorination and drying, is subjec-
2Fe + 6HCl ~ 2FeC1 3 + 3H2 ted to chlorination with chlorine gas to
Ni(Cu) + 2FeC1 3 --- Ni(Cu)Cl 2 + 2FeC1 2 form NbCl 5 and TiCl 4• When chlorination
is carried out at 700°C for 60 minutes,
2FeC1 2 + C12 ~ 2FeC1 3 the degree of conversion is 97% by
At the final stage, the Nb/Ta con- weight.
centrate thus formed is subjected to Fig.) shows the sequential order in
additional hydrometallurgical treatment which process operations for Nb and Ta
followed by chlorination, in which case recovery from wastes are carried out.
the chlorination degree reaches 90% at Significant possibilities for the reco-
500°C. If chlorination is not preceded very of rare metals from wastes are
by chemical treatment, the chlorination apparent.
degree at the same temperature does not
exceed 50%. The lower chlorination Sintering Zr with CaCO)_
degree under the same process conditions The Soviet zirconium industry is mainly
is explained by the presence of contami- based on processing zircon. Zircon con-
nants forming high-boiling chlorides. centrates containing at least 65% zr0 2
The latter create an obstacle to mass are used for further processing by the
transfer of chlorine and lower the Soviet industry. Latest research into
extent of chlorination. primary decomposition of zircon with
The hydro chlorination method is also lime-salt mixtures made it possible to
suitable for recovery of rare metals develop and partially to master on a
from discarded superconducting alloys commercial scale new processes offering
containing 25% Nb, 25% Ti and 50% Cu. As superior economic parameters and a
such waste material is treated with chlo- higher degree of protection against
rine gas in an aqueous medium at 60°C to market instability.6-8
95°C, copper is solubilized in accordan- As is common knowledge, zircon reacts
ce with a kinetic curve shown in Fig.2. with calcium carbonate at a molar
Zr02 :CaC0 3 ratio of 1:1 in accordance
100 .---------~-=-----,
with the following equation:
ZrSi0 4 + CaO - Zr02 + CaSi03
This process is complicated by a num-
ber of intermediate reactions which
result in the formation of acid-soluble
zircon silicate and calcium zirconate.
In studying the process of primary de-
composition of zircon with a lime/chlo-
ride mixture at a CsC0 3 :Zr02 ratio of
0.1 - 0.6 it was established that with
the increasing amount of the decomposing
agent (CsCO) the degree of decomposit-
J20 200 ion of zircon increases, aDd as this
Time, min figure reaches 90 to 92% the zirconium
Fig.2. Kinetics of extraction of Cu to content in the acid soluble compounds
solution at different temperatures (oC): present in sinter products increases
1 - 95; 2 - 80; 3 - 65. appreciably. A rise in the process tem-
perature, a longer sintering time and
319
Wastes
(Nb, Ta, Fe, Ni, Cu)
Chlori~~
~
I
~drochlorination
Water ,
Filtration
I- - Chloride solution
Drying
separJtion
crushed~
Nb) 70-90%
Processine:
HydrOflUo~ aCi~
,
Dissolution
Extraction
,
Chlorination, 400-500°C
. -----=- ----
Rectification
. 7•,
Precipitation, heat treatment I NbC1 5 1 I TaC1 5 1
I Nb20 51 °
[ Ta2 51
Fig.3. Schematic diagram of process flowsheet for recovery of Nb and Ta from wastes
the use of fluxing agents contribute to a product with minimal zirconium losses
lower the zirconium content in the form (1 to 1.5%) and with a minimal CaO im-
of these compounds in the sinter pro- purity content (0.25 to 0.35%). In the
ducts. latter case, Zr02 contains an admixture
The final product (Zr0 2 ) of this re- of silica (3 to 5%) in the form of zir-
action contains 0.2 - 0.3% CaO an amount con. This product is fairly usable for
that rapidly increases as the zircon the production of refractories and
decomposition degree exceeds 95%. These abrasives.
phenomena associated with the nonequi- Fig.4 shows a typical kinetic picture
librium nature of the solid-phase of the zircon decomposition process
process make it impossible to achieve using CaC0 3 in the presence of CaC12 • To
complete decomposition of zircon result- purify zr0 2 from the bulk of CaSi0 3 ,
ing in the formation of Zr02 alone. If sintered zircon cakes ground to a grain
zircon is fully decomposed, there is in- size of 150 m were subjected to leach-
variably formed a noticeable amount of ing using various acids.
zirconates, which ultimately leads to It was established that the best
zirconium losses, since purification is purification results are attained by
normally conducted with acid solutions. using two-stage acid leaching, namely:
Moreover, when zircon is fully de- the first leaching stage is carried out
composed, the reaction product contains cold with 5 - 6% HC1 (8 to 10% HN0 3 ),
an elevated amount of CaO (up to 3%). while the second leaching stage is
When the zircon decomposition degree is carried out under heating conditions
90 to 94%, it becomes possible to obtain with 10% HC1 or 15 - 18% RN03•
320
CaCl 2
CaC0 3 ZrSi0 4
/
CaC0 3 ~ Sintering
~~~
Sintering Leaching
~- RN0 3
\
2nd leaching stage------------ I
Hydrolysis ~ H2 0
J
Filtration Mother
1
Filtration
.l liquor for ,
Dr y~ng production Washing
"I zr0 2]
Ca ( NO 3) 2 -----===-::::;-;;,,-::::-
~~~--~--~----~~~~
Basic zirconium sulphate
for production of various
zirconium compounds
_S. ; .U; ; ;lLP. . ; h; ; ;a. ;.te. ;.l;. s; ;. o;. .d;.;,;a= =m=i: :x=t:ure~:-._-. ~
-
I
Sintering
~O _ Leaching with water
I
Acid leaching,
• filtration, washing
Filtrate 44_ _ _ _ _ _ _ __
Filtration,lwaShing
CONCLUSION References
The chlorination method for the recovery 1. Morosov I.S. Uses of Chlorine in Me-
of rare metals features high process tallurgy of Rare and Non-Ferrous Metals.
parameters and permits the production of Moscow: Nauka Publishers, 1966. 250pp.
various chemical compounds. The chlori- 2. Nazarov Yu.N., Titov A.A. Tsvetniye
nation method constituted the basis of Metally J., 1987. N 6, p.57-58.
a number of techniques for the product- 3. Krokhin V.A., Solyakov S.P., Mal't-
ion of ultrapure substances for various sev N.A. GIREDMET Transactions, vol.24,
technical purposes. The efficiency of 1969. p.153-160.
this method was confirmed for recovery 4. Titov A.A., Karavainy A.I., Mal't-
of niobium and tantalum from wastes. The Bey N.A. Tsvetniye Metally, no.3, 1988.
metallothermic reduction of chlorides p.56-58.
opens up a possibility for producing 5. Titov A.A. Fundamentals of Processes
metals more ec.onomically. for Producing Special-Purity Niobium
In order to produce zirconium and REM Pentachlorides and Oxides, TSNIITSvetmet
chemicals, effective methods were deve- Research and Development Institute for
loped for their recovery. These methods Economic Studies and Information Servi-
are based on hydrometallurgical proces- ces, 1986. p.48.
ses and on processes of solid-phase 6. Pol,tayev I.F., Krasnenkova L.V.,
interaction of reactants. Farther prog- Smurova T.V. Tsvetniye Metally, vol.12,
ress in this direction in processing REM 1988. p.56-58.
-containing materials makes it possible 7. Poletayev I.F., Krasnenkova L.V., Be-
to lower material losses and considerab- restiuk A.S. Tsvetniye Metally, vol.9,
ly to reduce discharge of harmful sub- 1983. p.67-68.
stances to the atmosphere. 8. Poletayev I.F., Smurova T.V., Kras-
323
nenkova L.V. GIREDMET Transactions. Mos-
cow, 1987. p.97-102.
9. Chuvilina E.L., Poletayev I.F., Z1mi-
na G.V. J.Inorganic Chemistry, vol.34,
no.5. 1989. p.1274-1280.
10. Baryshnikov N.V., Chuvilina E.L.,
Poletayev I.F. J.Inorganic Chemistry,
vol.28, no.5. 1983. p.1303-1308.
La lixiviation chlorurante d'alliages
Fe-Si: une voie d'avenir dans la
production du silicium1
F. Margarido
M. H. Bastos
Centro de Valoriza~iio de Recursos Minerais (CVRM), Departamento de Engenharia de Materials,
Instituto Superior Tecnico, Lisbon, Portugal
RESUME INTRODUCTION
Par c~ntre Ie chlorure ferrique est l'agent Les conclusions se rapportent au procede indu!
lixiviant par excellence du fer. triel a suivre, dans Ie but de la production du
silicium "metallique".
Le rapport etabli entre les rendements de la
lixiviation des impuretes et la composition
structurelle des alliages montre qu'on ne peut PROCEDE5 DE PURIFICATION DU 5ILICIUM PAR
atteindre Ie degre de purification voulu, sauf si LIXIVIATION ACIDE
on parvient a atteindre une structure adequate
des alliages pendant Ie procede industriel de La lixiviation est une des operations utilisees,
fabrication. depuis longtemps, dans la purification du silicium
325
326
"mdtalique". Les premiers travaux remontent aux par l'dtablissement de plans factoriels a deux ni-
annees 30, dpoque ou Becker et Tucker ont prdpard veaux, dont les facteurs (en nombre de k) peuvent
du silicium a 99%1. prendre seulement deux valeurs - correspondantes
aux niveaux infdrieur et supdrieur (ddfinis par
Actu~~ment, on peut envisager deux niveaux rapport a un niveaux "standard", dquidistant de
dans l'utilisation des procddds hydromdtallurgi- ces valeurs extremes).
ques: ceux qui portent soit sur Ie 5i avec une
teneur ~ 98%, soit sur des alliages de Fe-5i L'utilisation de variables codifides de fa90n
(avec une teneur en 5i 260%). Dans Ie premier homogene, au lieu des variables originelles,rend
cas on obtiendra Ie 5i avec une puretd de l'ordre plus facile la construction des plans factoriels
de 99,9% (convenable pour des applications dans d'essais. La formule de codification qui definit
les domaines de l'dnergie solaire et de l'dlectr~ les variables codifides, Xi 8 est la suivante:
nique) tandis que dans Ie deuxieme on ne parvient 2Xi-(XiB+XiA)
pas a depasser 99,2% en 5i (Ie produit dtant uti- i = 1,2, ... ,k, (2)
lisable dans l'industrie chimique).
etant
PURIFICATION D'ALLIAGE5 Fe-5i niveau inferieur de la variable
independante Xi
La lixiviation d'alliages Fe-5i a pour but l'dli-
mination des impuretds prdsentes, par l'action niveau supdrieur de la variable
des acides chlorhydique, fluorhydrique, sulphuri- inddpendante Xi
que ou nitrique, suivant la rdaction gdndrale 2 :
Dans cette formule on utilise la notation + 1
5i,(xFe,YAl,zCa)(s)+KHmA(aq)~5i(S)+xFeAq(aq)+YAl4s(aq)+ pour les niveaux des variables codifiees, ie, Ie
x V niveau supdrieur (+1) et inferieur (-1).
+ zCaA +K H2 (1)
I(aq) ; (g) On peut donc construire la matrice du plan facto-
riel d'essais qui represente l'ensemble des dif-
II n'y a pas beaucoup d'etudes sur la production ferentes possibilites de combinaison des niveaux
du 5i a partir d'alliages de Fe-5i. Dans la re- des variables codifides. Chaque ligne de la matri-
cherche bib~ographique effectude, on trouve seule ce spdcifie la combinaison de niveaux des divers
ment deux brevets norvdgiens qui utilisent, d'ail facteurs qu'il faut adopter, et maintenir fixes,
leurs, la me me technologie sur des all~ages Fe-5i pour chaque essai.
diffdrents (respectivement avec 60% 5i et 90 .~
94% 5i - procedd 5ilgrain 4 . La rdplique d'essais dans Ie niveau "standard"
conduit a l'estimation de la variance de l'erreur
Dans les deux cas on utilise un reacteur con- experimentale (V e ), ce qui correspond a une esti-
tinu, a contre-courant, la lessive constituant Ie mation de la dispersion du procddd d'observation.
flux ascendant. Le rendement de la lixiviation
des impuretds depend du type et concentration des L'ensemble des valeurs de la variable de rd-
phases intermdtalliques, dans les alliages de Fe- ponse V, obtenu experimentalement et concernant
-5i, qui sont solubles. les differentes combinaisons des facteurs, est
ensuite analyse et ddveloppe statistiquement.
~
Valeurs
I_ot 1 I_ot 2
Element Statistiques
En ce qui concerne la phase 6-FeSi , elle est deux niveaux et trois facteurs: Ie rapport liqu!
uniquement decelable dan. Ie lot 1. 2 de/solide, la concentration d'HCl, la concentra-
tion de FeC1 3 . 6 H20,respectivement designes par
Xl' x2 et x3 ·
PROC~O~ EXP~RIMENTAL
Aussi avons-nous dO effectuer les huit essais
Les lots des alliages Fe-Si utilises pour les es- necessaires, dans chaque plan, a l'etude de ces
sais de lixiviation, apres fragmentation et cri- trois facteurs, et encore quatre essais dans Ie
blage prealables,avaient une gamme de granulome- niveau "standard", pour estimer l'erreur experi-
trie fixee entre 10-22,4 mm. mentale .
Ces lots ont ete ensuite soumis a une Ii xi via- La matrice du plan factoriel d'essais, o~ on
tion acide, avec de l'acide chlorhydrique et du represente, les combinaisons des niveaux codifies
chlorure ferrique. des facteurs etudies, est indiquee dans Ie Tableau
II.
les reactions chimiques de solubilisation des
principales impuretes sont les suivantes: Tableau II - Matrice des niveaux codifies du
plan factoriel d'essais - 23
Facteurs
Essais Xl x2 x3
avec:
1 -1 -1 -1
2 +1 -1 -1
n 2 pour Me Fe ou Ca
3 -1 +1 -1
n 3 pour Me Al
4 +1 +1 -1
5 -1 -1 +1
Pour rendre compte de l'action simultanee de
6 +1 -1 +1
l'acide chlorhydrique et du chlorure ferrique sur
7 -1 +1 +1
la solubilisation des impuretes en question on a
8 +1 +1 +1
admis comme reaction globale de la lixiviation ,
pour chaque element, celIe qui resulte de la 9 a 12 0 0 0
La duree des essais, thermostatises a 102 0 c Pour Ie l er plan 23 l'indication des facteurs,
o . et de leurs valeurs dans les niveaux respectifs
(± 2 C), a ete constante (= 4h)· Aucune agltation
mecanique n'a ete utilisee au cours des experien- pour les deux lots, est resumee au Tableau III.
ces, puisque Ie degagement d'hydrogene a ete con-
sidere suffisant pour garantir l'agitation de la
pulpe.
Niveau
Lot Facteurs Inferieur "Standard" Superieur Unites
( -1) (0) (+1)
x1 6: 1 8: 1 10: 1 mllg
1 x2 10 24 38 gil
x3 0 170 340 gil
x1 6: 1 8: 1 10: 1 ml/g
2 x2 8 22 36 gil
er
Les resultats de l'application du 1 plan fa£
toriel mont rent des rendements de solubilisation
insuffisants pour les impuretes.
Pour les rendements d'extraction de l'Al, obte (1) Conclusion obtenue par l'analyse des coeffi-
nus dans l'alliage 1, l'accroissement de LIS est cients qui affectent les facteurs.
nuisible si les concentrations d'HCl et de FeC1 3
330
10.45
7. 9.29
6. 8.13
5 ~ 6.97
OJ
... 5 82
>-
4.Db
2. 3.00
1- 2.34
O.qL' IIII11I1I1I11
0 . 71 1
(a) 0.C4 ~
0.57 1
~
Q; 0.52
20.20 ... 0.43
>-
19.30 0.36
18.30 O.'lS
~
.... 17 .40 0.21
c:[ 16.50
>-
15.67
14.77
13.06
o
o '''>s
o ''''0
'0 'Os Fig. 4 - Surface de reponse pour Ie Fe, modele 5
'0 '<b - 1er plan factoriel 21 alliage 1;
'1 ' 6"s (:>
'q;. a) x3 est au niveau +1; b) x3 est au
nl veau -1.
Niveau
Lot Facteurs Inferieur "Standard"Superieur Unites
(-1) (0) (+1)
xl 6: 1 8: 1 10: 1 ml/g
1 x2 38 79 120 gil
x3
° 170 340 g/1
xl 6: 1 8: 1 10: 1 ml/g
2 x2 36 73 110 g/1
x3
° 150 300 gil
Tableau VII - Rendements d'extraction (%) du Fe, L'ajustement effectue pour Ies modeles elabo-
Ca et AI, pour Ie lot 1 (2~me plan factoriel). res m02tre qu'ils sont tres significatifs et aju~
tes (R = 99%) pour Ie Fe et l'Al, tandis que pour
Rendements d' extraction (y)% Ie Ca il est significatif mais non ajuste(R 2=75%).
Essais Fe Ca Al
Tableau VIII - Rendements d'extraction (%) du Fe,
1 0,43 58,50 22,37 Ca et Al dans Ie lot 2 (2~me plan factoriel)
2 0,78 55,28 24,14
3 4,30 56,92 46,42
Rendements d'extraction(y)%
4 5,05 62,64 46,09
5 4,69 51,00 13,89 Essais Fe Ca Al
6 10,72 61,81 20,43
7 7,81 54,92 20,97 1 0,46 62,72 21,64
8 10,42 49,31 21,09 2 0,69 60,03 23,13
9 4,69 54,56 15,82 3 4,58 63,79 50,41
4 5,34 64,01 51,75
5 4,81 50,36 16,81
L'estimation des nouveaux coefficients des 6 8,77 58,52 17,96
facteurs, et l'analyse respective de significa- 7 5,61 47,64 19,39
tion, montrent que: l'effet du facteur x est 8 9,16 49,18 15,80
tres significatif pour la lixiviation de§ trois 9 3,97 58,24 20,72
impuretes, tandis que Ie facteur x l'est seule-
ment pour l'Alj l'interaction des facteurs x2 et
x3 est significative pour la lixiviation du Ca et I_es modeles de regression du Fe et de I' Al sont
de l'Al, alors que l'interaction des trois fac- les suivants:
teurs l'est pour Ie Ca.
VFe = 4,821 + 1,063X l + 1,245X 2 + 2,160X 3 +
Les tests d'homogeneite des variances de l'er-
reur experimentale et des residus, dans les mode- ( 10)
les elabores pour les trois impuretes, ont revele
qu'ils sont - pour Ie Ca et l'Al - tres significa-
tifs (U=O,Ol) et ajustes (Ie niveau d'ajustement VAl = 26,401 + 7,226X 2 - 9,621x 3 - 7,121x 2x3 (11)
est de 88%). Le modele pour Ie Fe, bien que sig-
nificatif, n'est pas ajuste (R 2= 62%): la varia- L'analyse de l'influence des variables des mo-
tion giobale enregistree n'est expliquee que dans deles mathematiques, dans Ie rendement de solu-
une faible mesure. bilisation des impuretes, met en evidence Ies ef-
fets suivants:
Les modeles de regression qui decrivent Ie ren
dement d'extraction du Ca et de l'Al sont, pour 1 - Pour la variable xl (rapport LIS)
I'alliage 1, les suivants: L'accroissement du niveau de ce facteur est tres
significatif pour les rendements d'extraction du
Fe (alliage 2), quand Ie facteur x ([FeCI ])est
au niveau superieurl n'est pas sigfiificati~ si
(8)
celui-ci se trouve au niveau inferieur (fig. 5).
VAl 25,691 + 6,718x 2 - 7,830X 3 - 4,783x 2 x3 (~ L'augmentation du niveau de ce facteur est encore
favorable a la lixiviation du Ca, quand les ni-
veaux des facteurs x2 et x3 sont opposes (fig.6).
Les resultats des essais de Iixiviation de
I'alliage 2 sont repris dans Ie Tableau VIII. 2 - Pour la variable x2 (concentration d'HCl)
L'analyse de signification des facteurs montre L'accroissement de la concentration d'HCl a un
que, pour Ie Fe, les facteurs xl' x2 ' x3 et l'in- effet tres significatif sur Ie rendement d'ex-
teraction des facteurs x , x3 et x2 ' x3 sont tres traction du Fe (alliage 2) quand Ie facteur x
significatifs tandis que l pour Ie Ca c'est Ie fac- est au niveau inferieur, et peu significatif gi
teur x3 et pour l'Al les facteurs x2 ' x et leur x3 est au niveau superieur (fig. 5). II est favo-
interaction. 3 rable a la lixiviation du Ca et de I'AI si Ie fac-
332
teur x3 est au niveau inferieur (figs. 6 a 8). concentration d'HCl. En effet, la tendance globa-
Ie del'accroissement de la concentration de FeC1 3
est de contrarier la solubilisation du Ca, ce qUI
n'est pas Ie cas dans les essais 4 et 8 (dans Ie
Ca) l er plan factoriel) et 2 et 6(dans Ie 2eme plan).
(1)) 60.99
58.61
~ 56.23
4.
B 53.66
3. >- 51.47
M 3.
Qj'
....
>- 1.
O.
O.
65 . , ·,
yeo yeo
49.36 (-/_) (-1_)
44.69
40.62
E 36.36 Lsl,
----
L!)
......
c:(
30.68 Sf r NJ\UUC
>-
26.01
21.34
• , so; I
!)O 1
., ,,
..
~o ~o
: ,,
JO 30 ,,
YAI YAI
--,
I
,,
(-I..) (°/_ ) 20
-- -- _... ,:'
10
Fig. 7 - SurFace de reponse pour l'Al, modele 11-
0
- 2 eme plan factoriel 23, alliage 2.
HNJU!
11' ,
··
10
II
1", ,,
39.44 YFe ,,
(-I. ) 6 ,
34 . 74
E
.... 34.63
«
>-
26.33
20.62
L 5 '--------'--- ---'
-, Bibliographie
I
I \
\
trutural na Uxivia9 80 de Ugas Fe-Si, These
40 I 40 \, Doctorat. 1ST Ion (1989).
YAI I
I
YAI ,, 3. Christiania Spigerverk, (1972). Brevet Fran-
(%) 30
I
(%) 30 ,,
,, 9ais 7115882.
,,
4. Kol flaath, J.A. (1957). Brevet USA 2923617.
/
/ J
15 eI
20~
,7
--~
, a
5. Cochram, W.J. e Cox, G.M. (1957). Experimen-
tal Designs. 2 n ed. , John Wiley & Sons, N. Y..
:L_
10
6. Kempthorne, O. ( 1983) . The Design and Analysis
I
., of Experiments. Robert Krieger, Florida.
0_, 0 .,
----.I -, 0
7. BOX, G. et Draper, N. (1977) : Empirical Model-
NIVEAUX II:s f ACTELRS NIVEJVJX II:s fACTELRS
-Building and Response Surfaces. John Wiley &
Sons, Nova York.
10 .....:...'---r-----,-' 10 ., -1
8. Khuri, A. et Cornell, J. (1987). Response Sur-
faces. Designs and Analysis. Marcell Dekker ,
a a Nova York.
9. Margarido, F. et Figueiredo, M.O. (1988). Ma-
terials Science and Engineering, A104,249/54.
0_, ., 0_, .,
NIVEAUX II:s fACTELRS
°
fine grinding were subjected to gaseous and the residual Na 2 0 content in the
sulphation at 4S0 0 C with S02 + mix- range 0.7 to 1.3 wt % always remains in
ture to obtain water soluble VOSb 4 . the leached pellets which causes diffi-
Hater leaching at room temperature gave culties in the subsequent iron-making
88.2 % vanadium extraction calculated process.
on the basis of the vanadium content in As it was mentioned above, the smel-
the slag. A slightly lower degree of ting of vanadiferous titanomagnetites or
extraction was obtained by suI phi ding ilmenites under rather reducing condi-
the slag with S2 + C at 1200 0 C and then tions results in the extensive reduction
sulphation with water leaching. of vanadium into a high-carbon pig iron
melt. Bong-Chan and Kruger l have shown
that extensive reduction of titania slags
INTRODUCTION can remove, at the very most, SO % of the
Hore than 80 % of the world's vanadium vanadium content of the slag. They also
production is recovered from titanoma- found that when FeO content of titania
gnetite ores. Titanomagnetites consist slag is reduced to 1 - 2 wt % the most
of a.titanium-rich.magn~tit~ i.e.F~304- of the silicon, manganese and chromium
Fe 2Tl0 4 (ssl' some 11me~lte 1.e.FeTl0 3 - contents are reduced into the metal. At
Fe 20 3 and gangue mlnerals. Vanadlum the same time the Ti+ 4 in the slag is re-
distr±Eution between coexisting magne- duced completly to Ti+3. This extensive
tite and ilmenite phases show some scat- reduction of titania slag requires a
ter, but the magnetite phase 3 is usually large increase in the electric energy
richer in vanadium, where v+ replaces consumption, up to unacceptable level of
Fe+3 as shown by Schuiling et al. l 2000 k\;Jh/t slag.
Vanadium in these ores amounts to be- Smelting of titanomagnetite or ilmen-
tween 0.2 - 0.9 % and can be recovered. ite ores under less reducing conditions,
These ores can not be smelted in an produce a low-carbon iron, and a low
iron blast furnace because of the high melting slag with 8 - 10 wt % FeO, con-
content of Ti02' and the highly reduc- taining nearly all the vanadium, chrom-
ing conditions which exist in such fur- ium and manganese values. This slag may
naces. There is extensive reduction of subsequently be subjected to a refining
Ti+4 which gives a mushy high-melting process Ivhereby the vanadium can be re-
point slag which may contain titanium covered.
carbides and nitrides. These ores can One refining process is by sulphiding
be smelted in electric reduction furn- or hydrosulphiding according to the re-
aces where less reducing conditions can actions;
be achieved. This gives a pig iron (hot (f,1eTi 20 S )a(Ti 30 S ) I-a + (a-l)S2 = arieS +
metal) which contains a certain amount
of vanadium. By blowing the pig iron in (3-a) Ti02 + (a-!) S02 (1)
a converter, a slag rich in vanadium is (J:1eTi 20,,)
obtained. This slag can be used for ~ a (Ti 30 S ) 1 -a + aH')S
~
= aMeS +
Compo-
Series A Series B Series C
nents
wt % Low S High S Low S High S Low S High S
Compo-
nents Series A Series B Series C
wt %
-
experiments is shown by a dashed
line. ~ 20
........
It) (J')
0 ::?!
If) *' Series A
::?! c: 10
30 o Series B
c: ~
0
A Series C
+-
~
0
_ 6_lnduclion ~
+-
~
furnace
> * *
If)
0 20 4 6
2
>N B
........
......
(J')
slag was remelted in a graphite crucible.
::?! 10 o Various degrees of reduction were ob-
c: tained by various holding times at sme-
lting temperature. During smelting fur-
~
0 ther reduction of FeO to metallic iron
+- and Ti +4 to Ti+ 3 took place. Elemental
~
sulphur was added by means of a graphite
> 6
"plunger" and 120 g of MgO as micro-
2 4 pellets was added to the molten slag.
During remelting, due to extensive redu-
ction a certain thickening of the slag
was observed and after the sulphur/MgO
addition the slag once more became
The most interesting feature of these
Figures is the formation of vanadium quite fluid. The slag/sulphide ~ixtures
wi th different degr.ees of reduct10n were
rich sulphides in a wide range of the
poured into cast iron ingot molds and
degree of reduction of slag,fo: s~mp~es were cooled and then analyzed by X-ray
with a high MgO content. Th1S 1S 1nd1-
diffraction and microprobe analysis.
cated by high distribution ratio on both
Similar to the synthetic mixtures
Figures, although this effect i~ more
most of these samples consist of a four
pronounced for samples with a h1gher
phase combination i.e. the pseudobroo~
sulphur contents. Vanadium distribution
ratios between the sulphide and the gla- kite phase M3 0 S ' sulphide p~as~ MS,
ssy silicate were found to be one order glassy silicate, and metall1c 1ron. The
M °
of magnitude higher than between the
phase and the MS phase. These re-
in the M °
overreduced slag for which FeO content
phase was found to be below
0.2 wt%,3i~ addition, to these phases
stll~S suggest that sulphiding of molten
vanadiferous titania slag should be gave also the tagirovite M)03 phase
done under reducing conditions with with the V 03 content in tfie range 1.1-
1.4 wt%. T~e V2 0 3 content of the M3 0 S
MgO added.
References
1. Schuiling R.D. and Fennstra A. Geo-
chemical behaviour of vanadium in Iron-
Titanium Oxides. Chern. Geology, vol. 30,
1980, p. 143-150.
2. Bong-Chan and Kruger J. Upgrading of
ilmenite and tltania slag by DC-ARC
smelting. Extractive Metallurgy and
Materials Science Part 2, Sept. 21-24,
1987, Editors: Li Songren, Jin Zhangpeng
and Z. Yong Jian. Symposium organised
by CSUT and TUC in China.
3. Borowiec K., Rosenqvist T., Tuset J.
and Ulvensoen J.H. Synthetic rutile
from titaniferous slags by a pyrometa-
llurglcal route. In: pyrometallurgy 87,
p. 91-119, 1987, The Institution of
Mining and Metallurgy.
4. Borowiec K. Sulphidization of solid
titania slag. Scand. J. of Metallurgy
in press.
5. Burylew B.P., Mishin P.P. and
Tseitlin M.A. Thermodynamics of the pro-
cess of desulphurizlng liquid and solid
steel by vanadium. Fiz-Khim. Issled.
Metall. Protsessov (Sverdlovsk) vol. 12,
1984, p. 57-62.
6. Pei B. Sulphidization of rare metal
(Cr,V,Ti) oxides. Ph.D. thesis.
Norwegian Institute of Technology,
Trondheim, 1989.
•
Electrolytic reduction of Eu(lll) In
acidic chloride solutions
T. Hirato
H. Majima
Y. Awakura
Department of Metallurgy, Kyoto University, Kyoto, Japan
341
342
As is obvious in this table, the reduction poten- chemicals used in this study were commercially
tial of Eu(III) to Eu(II) is larger than that of supplied and of reagent grade. Deionized water
the other redox systems. Thus Eu(III) can be with a specific resistivity of 5 x 10 5 ohm cm was
selectively reduced more easily than the other used in the preparation of aqueous solutions.
trivalent rare earth ions.
Industrial processing of Eu(III) involves re- Experimental apparatus
duction to Eu(II) by zinc powder or zinc amalgam, The experimental set ups used in this study
and the Eu(II) thus formed is precipitated as are essentially the same as those reported in
sulfate by means of a fractional crystallization previous papers 6,7. Fig.1 shows the apparatus
method 3. However, it is necessary that the zinc used for the determination of cathodic polariza-
introduced during reduction of Eu(III) is removed tion characteristics. A rectangular vessel was
from the resultant solution. Recently, Lu et al. divided into two compartments by an anion ex-
reported the electrolytic reduction of Eu(III) change membrane fixed at the center of the ves-
using a graphite electrode 4. Furthermore, the sel. The size of the titanium cathode was 1.5 cm
electrolytic reduction of U(VI) to U(IV) was x 1.5 cm, while that of the platinum counter
studied in the authors' laboratory using a tita- anode was 5.0 cm x 1.0 cm. The cathode compart-
nium cathode of a high hydrogen overvoltage 5- 8. ment was filled with acidic EuC1 3 solution, and
Since the reduction was achieved with high effi- the anode compartment with 0.1 kmol'm- 3 HCl solu-
ciency using a titanium cathode, the reduction of tion. The catholyte solutions were agitated by
Eu(III) is also expected to be efficient. means of a magnetic stirrer throughout the elec-
In the present study, to obtain better funda- trolysis. A three electrode method was employed
mental understanding of the electrolytic reduc- for polarization measurements during the electro-
tion of Eu(III) using a titanium cathode, cathod- lytic reduction of Eu{III). A AgCl-Ag electrode
ic polarization characteristics of the reduction
of Eu{III) were studied. Batch-type electrolytic
reduction of acidic EuC1 3 solution was also in- Function
g.n.rator
vestigated, using a bipolar electrolytic cell. x-v
R.cord.r
Pot.n!lo
EXPERIMENTAL PROCEDURES galvano
stat
60
0.1 kmol·m-3EuCI 3
III :400min- 1
50
Temp I K
1 : 288
40
2: 298
< 3: 308
_E 30
4: 318
2
20 1
10
Fig.2 Schematic illustration of the experimental
apparatus used for batch-type electrolytic reduc- OL-__ ~~E- __L -_ _L -_ _L -_ _L-__ L-~
tion of Eu(III): (A) Pt anode, (B) Anion selec- -0.2 -0.4 -0.6 -0.8 -1.0 -1.2 -1.4 -1.6 -1.8
tive membrane, (C) Ti cathode, (P) Pump, (G) E 'V vs SH E
Ga1vanostat, (W) Thermostated water bath, (Va)
Vessel for ano1yte, (Vc) Vessel for catho1yte,
(I) C1 2 gas absorber, (N) N2 gas cy1 inder, (T) Fig.3 Effect of temperature on the cathodic po-
Autotitrater. larization curve of aqueous EuC1 3 solution
344
surface product was analysed by ESCA. The energy portional to the square root of agitation
°
spectra of Eu and exhibited peaks at 137.3 and speed l2 . Fig.6 depicts the linear relationship
532.2 eV. respectively. The energy spectra are between the plateau current and EuC1 3 concentra-
very similar to those of Eu 203. These findings tion. Judging from these results. the plateau
suggest that a film. which inhibits reduction.
formed on the cathode surface by hydrolysis of 5O~-------'r-----"""-----"""
-
This value suggests that mass transfer controls E
the electrolytic reduction of Eu(III) in aqueous
solution 9- 11 under the experimental conditions 20
which determine the plateau.
The effect of agitation speed of catholyte on
/
the plateau current is shown in Fig.5. In this 10 ;'
;'
figure. values of the current are plotted against ;'
1.7.----.----.----.-------,.---,
70r---r---.---,---.---.
0.1 kmol·m-3 EuCI 3 Temp: 298 K
III: 400min- 1
60 W: 400 min- 1
1.6
50
<
E « 40
E
1.5
30
Ea = 15.7 kJ ·mol-I 20
1.4
10
Fig.4 Effect of temperature on the plateau cur- Fig.6 Effect of EuCl a concentration on the pla-
rent of the cathodic polarization of aqueous teau current of the cathodic polarization curve
EuCla sol ution of aqueous EuCla solution
345
current of the cathodic polarization curve is a evolution at the titanium cathode, the difference
diffusion-limiting current in the electrolytic between the currents shown by polarization curves
reduction of Eu(III). 1 and 3 should correspond to the reduction cur-
rent of Eu(III) in aqueous 0.1 kmol·m- 3 EuC1 3 -
Cathodic polarization curve for the aqueous 0.1 kmol ·m- 3 HCl solution. This difference in
solution system, EuC1 3-HCl currents is illustrated by Curve 4 in Fig.7,
In industrial solvent extraction processes for which shows that the electrolytic reduction of
the separation and concentration of rare earths, Eu(III) can proceed in an aqueous solution of
aqueous solutions containing rare earth salts EuC1 3-HC1. However, the reduction current de-
acidified with HCl or HN0 3 are generally sub- creases at a cathode potential lower than -1.0 V
jected to treatment. Therefore, cathodic polar- vs SHE, when the evolution of H2 gas becomes
ization curves for aqueous EuC1 3-HCl solutions significant.
are of interest.
Curve 1 in Fig.7 shows the cathodic polariza- Batch-type electrolytic reduction of acidic EuC1 3
tion curve obtained with aqueous 0.1 kmol·m- 3 solution
EuC1 3 - 0.1 kmol·m -3 HCl solution. For the sake Electrolytic reduction of Eu(III) was per-
of comparison purpose, cathodic polarization formed in solutions of different pHs using the
curves obtained in aqueous 0.1 kmol.m- 3 EuC1 3 bipolar cell shown in Fig.2. The time variation
solution and in aqueous 0.1 kmol·m- 3 HCl solu- of the concentration of Eu(II), which was pro-
tion, are depicted by Curves 2 and 3, respective- duced by the electrolytic reduction, is shown in
ly, in the same figure. Comparison of Curve 1 Fig.8. The broken line indicates the theoretical
and Curve 2 shows that the addition of HCl to the time variation of the Eu(II) concentration, which
EuC1 3 solution slightly decreased the cathode was derived assuming that a current efficiency of
potential for initiation of Eu(III) reduction. unity. As is clear in this figure, there is no
Also, there is no longer a distinct plateau cur- significant difference between the reduction
rent, or diffusion-limiting current in the rates at pH 2, and at pH 1. The reduction of
Eu(III) reduction. Eu(III) was almost complete in about 4 h. The
Since the cathodic polarization curve, which current consumed by the reduction of H+, which
was determined in 0.1 kmol·m- 3 HCl aqueous solu- was calculated from the amount of Hel added by
tion, is believed to correspond to hydrogen gas the autotitrater during the experiment, showed
80r---~---r---.--~r---~---r---.-, good agreement with the current calculated from
Temp: 298 K
III : 400 min- 1
the difference between the reduction rate curve
70
1 : 0.1 kmol'm-3EuCI3
obtained experimentally and that predicted theo-
.0.1 kmol·m- 3 HCI retically. This suggests that a possible side
so 2 : 0.1 kmol·m-3 EuCl3
3 : 0.1 kmol'm-3 HCI O.I.----.-----r----r----=--=----r-----,
4 : Difference between
I
"' S O l a n d 3 I
E '",
E I
- 40 ..:. I EuCl3 : 0.1 kmol·m- 3
o I
E I Temp: 298 K
"" I
30 - 0.05 I : 0.5 A
I
I
I pH t:. 1
20
I
I o 2
~ o 3
10
V 4
O~--~--~--~=-~--~--~--~~
-0.2 -0.4 -O.S -0.8 -1.0 -1.2 -1.4 -1.S 2 3 4 5 6
E I V vs SHE h
Fig.7 Cathodic polarization curves in various Fig.8 Effect of pH of catholyte on the electro-
aqueous EuC1 3 solutions lytic reduction of Eu(III)
346
reaction is the reduction of H+. On the other evolution of H2 gas. Fig.IO shows the mean cur-
hand, the reduction of Eu(III) was limited at pH rent efficiency of the electrolysis depicted in
3 and pH 4. At these higher pHs, it was observed Fig.9. Although the mean current efficiency de-
that the cathode surface was coated with a white creased as the reduction progressed, it remained
product. It is believed this product formed by above 0.6 at the end of the electrolysis at 0.25
the hydrolysis of Eu(III), because the pH of the A. Thus, in the batch-type electrolytic reduc-
solution in the vicinity of the cathode surface tion, excess current was not always desirable,
was raised by H+ reduction. Therefore, in the from the viewpoint of current efficiency.
experiments which followed, the pH of the catho- The data shown in Fig.9 are replotted in Fig.
lyte was maintained at a level of 2 throughout 11, according to the first order rate law. In
the electrolysis. Fig.lI, the logarithm of the ratio of Eu(III)
Fig.9 shows the effect of electrolytic current concentration at time, t, to the initial concen-
on Eu(III) reduction. The times required for tration is plotted against the reduction time.
complete reduction of Eu(III) were 5 h at 0.25 A In the electrolysis at 0.5 A, a linear relation-
and 4 h at 0.5 A. The reduction rate of Eu(III) ship remains throughout the reduction. This re-
at 0.75 A decreased after 1.5 h due to a violent sult indicates that the reduction rate is pro-
portional to the concentration of Eu(III), and
suggests that the reduction of Eu(III) proceeded
0.75A 0.5A 0.25A under the condition of a diffusion-limiting cur-
I ... 1'I.•.. . - ....•.•••..../............................
0.1 -.......... r' I
..... rent. On the other hand, for the electrolysis at
I I
I I 0.25 A, the initial reduction rate was less than
I I
I I the limiting current. After 3 h of reduction,
I I
I I Temp: 298K when the Eu(III) concentration became lower, the
I I
I
I
I
I
EuCI3: 0.1 kmol·m- 3 reduction proceeded at the limiting current and
I I exhibited the same slope as that obtained at 0.5
:0.05 I I pH 2
1:1 I I
I I I A
A. At 0.75 A, the reduction of Eu(III) proceeded
:J
W I I initially at the limiting current; however, after
I I. 6 0.25
1.5 h of reduction, the reduction rate became
I
I "
o 0.5
I smaller. This may be attributed to a decrease in
I o 0.75
OL-__ ~ __- J_ _ _ _ ~ _ _- L_ _ _ _L -_ _ ~~
cathode surface area available for Eu(III) reduc-
o 234 5 6 tion. The cathode surface is covered with H2 gas
h
u•
u
-1
CI
0
I I A
IFO.
I I A 6. 0.25
Temp: 298 K A 0.25 -1.5 0 0.5
pH : 2 0 0.5 0 0.75
EuCI3 : 0.1 kmol'm-3 C 0.75
0
0 2 3 4 5 6 7
t I h 2 3 4 5 6
I h
Fig.l0 Mean current efficiency of the electrolyt-
ic reduction of Eu(III) using different reduction Fig.ll First order rate law plot of the data
currents shown in Fig.9
347
more significantly at a higher current density. order rate law plots of the data for 0.05 and
Electrolytic reduction experiments were per- 0.025 kmol'm- 3 EuC1 3 are almost identical to that
formed at a constant current of 0.5 A using aque- of 0.1 kmol.m- 3 EuC1 3 shown in Fig.11. These
ous 0.1 kmol'm- 3 EuC1 3 solution as catholyte, findings suggest that the electrolytic reduction
under a flow rate of catholyte ranging from 3.3 proceeds under the condition of the limiting cur-
to 1.5 1·min- 1. The reduction rate of Eu(III) rent. The low current efficiency observed at low
decreased dramatically with a decrease in flow EuC1 3 concentration was caused by current con-
rate. This result can be understood from the de- sumption by the evolution of H2 gas.
crease in the diffusion limiting current density
due to the increase in thickness of the diffusion CONCLUSIONS
layer.
The effect of temperature on the electrolytic To more fully understand fundamental aspects
reduction of Eu(III) was examined at 298, 308 and of the electrolytic reduction of EuC1 3, the po-
318 K, using aqueous 0.1 kmol'm-3 EuC1 3 solution larization characteristics of electrolytic reduc-
as catholyte and maintaining the reduction cur- tion of Eu(III) using a titanium cathode were
rent at 0.5 A. Although the reduction rate in- examined. Batch-type electrolytic reduction of
creased with increasing temperature, the effect Eu(III) was also investigated using a bipolar
of temperature was not significant under these electrolytic cell, in which a titanium plate was
experimental conditions. This is reasonable, used as the cathode. The principal results ob-
since an increase in temperature causes an in- tained are as follows:
crease in the reduction rate of H+, as well as 1. The cathodic polarization curve for aqueous
that of Eu(III). 0.1 kmol'm- 3 EuC1 3 solution exhibited an increase
The results of the electrolytic reduction of in electric current, at the start of the electro-
EuC1 3 solution containing different concentra- lytic reduction of Eu(III) to Eu(II), at around
tions of EuC1 3 are shown in Fig.12. Although the -0.6 V vs SHE. The plateau current for the re-
reduction rate of Eu(III) decreased gradually duction of Eu(III) was observed at around -1.2 V
with the progress of reduction, almost complete vs SHE, but further lowering of the cathode po-
reduction was achieved, in each case, after about tential resulted in a decrease in the reduction
4 h. The mean current efficiencies at the ini- current. The electrolytic reduction of Eu(III)
tial stage of electrolysis were around 0.4 and could proceed in aqueous 0.1 kmol.m- 3 EuC1 3 - 0.1
0.2 for 0.05 and 0.025 kmol'm -3 EuC 13 so lutlons,
' kmol'm- 3 HCl solution, although the reduction
respectively, and these efficiencies decreased current decreased at a cathode potential lower
with the progress of the electrolysis. The first than -1.0 V vs SHE, when the evolution of H2 gas
AC KNOWLEDGMENT
Financial assistance from The Kawakami
Memorial Foundation to T. Hirato is greatly
acknowledged.
REFERENCES
1. W.M. Latimer; "Oxidation States of the
Elements and their Potentials in Aqueous
Solutions", 2nd Ed., Prentice Hall, New York
(1952), p.294.
2. L.B. Aspery, B.B. Cunningham; "Progress in
Inorganic Chemistry", Vol.11, Edited by F.A.
Cotton, Interscience, New York (1960), p.267.
3. C.G. Byown and L.J. Sherrington; Journal of
Chemical Technology and Biotechnology, 1979,
Vol.29, p.193.
4. D. Lu, J.S. Horng and C.P. Tung; Journal of
Metals, 1988, Vol.40, p.32.
5. H. Majima, Y. Awakura and S. Hirono; Metal-
lurgical Transactions B, 1986, Vol.17B, p.41.
6. H. Majima, Y. Awakura, K. Sato and S. Hirono;
Metallurgical Transactions B, 1986, Vol.17B,
p.69.
7. Y. Awakura, K. Sato, H. Majima and S. Hirono;
Metallurgical Transactions B, 1987, Vol.l8B,
p.19.
8. Y. Awakura, H. Hiai, H. Majima and S. Hirono;
Metallurgical Transactions B, 1989, Vol.20B,
p.337.
9. M. Eigen, W. Kruse, G. Maass and L. de Maeyer;
Progress in Reaction Kinetics, 1964, Vol.2,
p.287.
10. M. Anbar, Z.B. Alfass and H. Bregman-Reisler;
Journal of the American Chemical Society,
1967, Vol.89, p.1263.
11. B. Cercek and M. Ebert; Journal of Physical
Chemistry, 1968, Vol.72, p.766.
12. V.G. Levich; "Physicochemical Hydrodynamics",
Prentice Hall, New York (1962), p.70.
Environment
Best available technology-
a viewpoint on the development
and application of the concept to
the European non-ferrous industry
Tony Connell
Britannia Refined Metals, Northjleet, England
Robert Maes
MHO-a Division of ACEC-Union Miniere, Hoboken, Belgium
Jose Poncet
Asturiana de Zinc, Aviles, Spain
PREFACE
PART A: GENERAL ASPECTS
AI Introduction
All Legal Provisions of Member States
Relevant for plant Authorisation
AlII Processes and Installations Included in
this Technical Note
AIV Heavy Metals And Their Compounds Emitted
by Non-Ferrous Industrial Plants
AV References to Part A
PART D: ANNEXES
DI Examples of the Performance Characte-
ristics of Dust Collectors
DII Example of Emissions Figures of
Primary Lead Production
DIll Plant Authorisation Procedures in
EC Member States
DIV Recommended emission limits for
lead-, copper-, and zinc-works
354
resulting costs and the way in which these would REMARKS MADE BY EUROMETAUX ON
be assessed in different countries. BEHALF OF THE EUROPEAN NON-FERROUS
Similar concerns were also expressed by the INDUSTRY
authors of this paper during the Technical Wor- Following a request of the Chairman at the meet-
king Group meetings and these were notified to ing of February 1990, Eurometaux presented some
the Environment Department (DGXI) of the Commis- serious concerns to the Commission that it con-
sion. For example, competition between processes sidered should be addressed in the follow-up
for a lowering of emissions by a few per cent meeting for the heavy metals sector. They summa-
was considered futile. Furthermore, there were rized the main difficulties that had been en-
differences of opinion regarding processes countered in the implementation of the BAT con-
applying for BAT that were not unanimously con- cept for the sector.
sidered as "proven" technology. 1. It was clear from the descriptions brought
together in the technical note that the pro-
"PLENARY" MEETINGS FOR EXCHANGE OF duction processes of non-ferrous metals were
INFORMATION BETWEEN COMMISSION AND very varied and had a very large number of
NATIONAL EXPERTS unit operations that were dependent upon
Following the establishment of "final versions" speciflc economic and social conditions and
of the technical notes for the different produc- possible feed materials. The aim was to
tion sectors meetings took place in the frame- designate a process for the production of a
work of the BATNEEC Exchange of Information Com- given metal as the best available technology
mit tee , bringing experts from the Commission and and this therefore appeared irrealistic; even
from national Governments together, under the if one concentrated on the unit operations,
chairmanship of a representative of DGXI respon- attention had to be paid to the entire pro-
sible for BAT activities; industrial experts cessing circuit in which they were linked in
were invited as observers. The following pro- order to treat given feed materials .
gress was achieved : 2. Primary feed materials were, in essence, min-
. December, 1989 : Most of the established tech- eral and thus variable in both form and com-
nical notes were agreed, but reservations were position. Recycled materials were also ex-
expressed for the heavy metals sector, where tremely varied in shape and combination.
the large number of processes mentioned and Metallurgical processes, particularly in
problems underlined by industry were recogni- Europe, often combined both types of feed
zed to need further consideration. It was also materials (for the sake of environment pro-
concluded that a firm legal ground for BAT ac- tection), with corresponding implications on
tivities should be established by appropriate the choice and performance of unit opera-
Council legislation . tions.
- February, 1990 : Requests to the Council to 3. With regard to environmental protection,
formally set up the BAT Committee and to defi- emissions at the stack for a given process
ne the legal status of the technical notes were important, but not the only polluting
were proposed. Owing to the complexity and factor to be taken into consideration. Fugi-
variety of processes listed in the technical tive emissions were also important, but not
note on heavy metal emissions, some reserva- directly related to the selected process. The
tions were expressed on making the total BAT overall environmental impact, including all
technical notes legally binding and it was effluents, residues and waste products, as
suggested that they should remain of an infor- well as recycling capabilities, should there-
mative nature and be completed by a separate fore be considered in the assessment of BAT.
section defining the standards that would be 4. Futile competition between processes on the
binding and enforceable. basis of their emissions should be avoided by
355
coupling a given reduction of the particular tion with the implementation of the BAT concept.
emission (air, water, waste, etc. ) with a The industry convinced that it was not merely
fair appreciation of the corresponding reduc- manoeuvring to delay enforcements but that fun-
tion of harmfulness and with the price that damental obstacles were encountered that would
has to be paid for it, as referred to in have to be cleared in order to enable the Direc-
article 8 of the Directive. tive to be properly and effectively enacted. The
5. One had to be aware of the tendency for some principle of "pilot exercise" being practised
companies or groups to outbid their competi- with the establishment of these first technical
tors by pushing their own process in order to notes was fully realized and it is not surpri-
favour their engineering divisions. There was sing that some major interventions were made by
consequently a risk of distorting the infor- the non-ferrous sector, to which environmental
mation released, particularly about new pro- questions were particularly sensitive.
cesses, on how they achieve the required per- The workload required to defme all the rele-
formances and how far they may be considered vant technologies was recognized and time would
as proven, reliable and flexible technology . be given to perfect what had already been achie-
It was therefore important to define the ved. Practical difficulties to define emissions
parameters of proven technologies for each associated with given processes were realized.
specific plant or equipment relative to the For the delicate point of "proven" technology,
process. Eurometaux was requested to. contribute to the
6. The assessment of costs associated with the development of a definition.A climate of mutual
application of a process or unit operation appreciation was established, inspiring confi-
would always be complicated by local circum- dence that a way would be found to apply the
stances, such as the cost and availability of Directive with the necessary flexibility, taking
energy, specificity of feed materials or im- into account the complexity of involved proces-
plications of the overall processing circuit ses, the variety of feed materials and the di-
in which the introduction of the process or versity of local situations. The hope was ex-
unit operation was considered. pressed that a completed version of the techni-
7. Notwithstanding the above remarks, it was cal note on heavy metal emissions would be fina-
considered that the BAT action was fruitful, lized in the course of 1991.
at least by the information that it would
bring together and the inventory of processes THE CONCEPT OF PROVEN TECHNOLOGY
that it would establish. The belief was ex- The feeling of the authors of this paper is that
pressed that the action should be extended in no authority can be conceived that could settle
a positive direction by giving the produced out of hand whether a given technology is proven
documents a certain enforcement in order to or not. The decision can only be established by
progressively reach harmonization of the some form of jurisprudence coupled with past ex-
environmental standards for the non-ferrous perience. Therefore, it was decided that a sta-
industry all over the EC. tement should be developed to define the minimum
criteria that a technology should meet to be
"RESTRICI'ED" MEETING AT EUROMETAUX'S considered as best available technology. After
OFFICE consultation of all Eurometaux' members, this
An informal discussion to exchange views on statement became the following :
Eurometaux's reservations took place with the "In order to be taken into consideration as best
reponsible representative of DGXI and the coor- available technology, a technology must be
dinator. This meeting was very constructive and proven i. e. the technology must have been suffi-
probably played an essential role in the future ciently operated during a reasonable period of
decisions which were going to be taken in rela- time, according to the complexity of the pro-
356
cess, at a stated capacity, processing the decisions. A suggestion had been made before-
various feedstock materials intended, so that hand by the coordinator, based on a proposal
appropriate assessments can be made of recog- from the German Umweltbundesamt, to define
nised performance criteria and results recorded eDllSSlon limits associated with metallurgical
in an appropriate form, in terms of reliability , processes on a general TA Luft-type base6 and to
consistency of environmental performance and seek agreement on a draft document established
economic factors, which can be proved and con- along this line. After consultation of Eurome-
trolled, if need be." taux' members, this procedure was accepted by
It was also felt that minimum criteria for the industrial experts. It presented the advan-
technology evaluation should include details on tage of overcoming the difficulty of collecting
essential operating data, such as : information on emissions for each individual
- a description of the process and the equipment process. However, it did not exactly follow the
and/or plant in which it is practiced, with approach recommended by the Directive, but did
mention of capacities appear as an acceptable initial solution.
- physical and chemical characteristics of pro- Discussions took place in an excellent atmos-
cess materials phere. Most proposed amendments to the sugges-
- the period under evaluation and the degree of ted emission limits were unanimously approved.
availability of equipment and/or plant Furthermore, the comments of industry on proven
- process cross media aspects technology were not only accepted for inclusion
- a technical description of environmental con- within the technical note but would also be used
trols for emissions to air, water and waste, as a reference to all industrial sectors techni-
and their operational performance, including cal notes.
ambient air quality
- economic factors. FUTURE TRENDS
This list should be considered as a recommen- At the end of the meetings of February, 1991,
dation of minimum information to be released on the responsible representative of DGXI summari-
processes applying for BAT. The list was also zed the future orientations of the BAT activi-
established with the intention of providing sup- ties as follows7 :
port to companies being constrained by authori- - The technical notes that had been approved
ties requiring them to apply technologies that unanimously by the experts of the Technical
in their view, were not proven for the condi- Working Groups would be presented to the Com-
tions of the intended application. It may be mission and to national government experts for
stressed, in this connection, that the interpre- final agreement.
tation of these principles remains quite intrac- - The Commission would organize at least two
table, as exemplified by the case of a well es- meetings per year of the Exchange of Informa-
tablished process whose transposition to slight- tion Committee in order to approve technical
ly different conditions has recently resulted in notes, improve their quality, define further
significant start-up difficulties. 5 work programmes and decide upon updating of
technical notes whenever necessary.
FINALIZATION OF THE TECHNICAL NOTE - Approved technical notes would be published by
ON HEAVY METAL EMISSIONS the Commission as EUR documents.
A last round of consultations was organized in New technical notes would not contain emission
February, 1991, with the aim of fmalizing the limits but only a description of the BAT,
technical note. The Technical Working Group was including relevant operating and investment
split into sub-groups, dealing separately with costs and associated emission~.
lead, zinc and copper, which, in fact, agreed - The technical notes would not be binding docu-
more or less unanimously with their respective ments for the Commission, neither for the Mem-
357
ber States nor for the industry. They would
constitute a technical basis for the Commis-
sion to prepare directives on specific indus-
trial sectors and annexes defining emissions
limits.
- The annexes would be inserted in a more gene-
ral directive prepared in the framework of an
integrated approach for industrial pollution.
It was intimated that this integrated ap-
proach would consider the overall environmental
impact in the assessment of BAT.
These new DG XI guidelines were received very
positively by all the attending experts .
References
1. Directive du Conseil du 28 juin 1984 relative
a la lutte contre la pollution atmospherique en
provenance des installations industrielles
(84/360/CEE), Journal Officiel des Communautes
EurQP6ennes,16.7.84.
2. Bartaire J. -G., La politique de la Commission
des Communaut6s Europ6ennes en matiere de pollu-
tion atmospherique, Pollution Atmospherique,
juillet-septembre 1990, pp. 267-278.
3. Technical Note on Best Available Technology
Not Entailing Excessive Costs for Heavy Metal
Emissions from Non-Ferrous Industrial Plants,
Final Report, April 1991, to be published by the
Commission as EUR document.
4. Barbour A.K., Europe on the 1992 horizon -
Views of the European non-ferrous industry - Its
hopes and concerns, Politiques "Environnement.
Hygiene et S6curite" des Communautes Europeen-
nes. Eurometaux, 28-29 septembre 1989, pp. 89-
91.
5. Anon., World's largest flash smelter - Under
strict environmental demands, Magma meets rules
and decreases costs as well, Enaineering and
Mining JOurnal, January 1990, pp. C37-39.
6. Technische Anleitung zur Reinhaltung der
Luft, TA Luft. Deutscher Wirtschaftsdienst.
7. Bartaire J.-G., New orientations of the BAT
activity, Document DGXI/A/3/JGB/kl, 18.1.91
Environmental legislation and
advances in tailings disposal
technology in North America and
Europe
1. P. Haile
Knight and Piesold Ltd, Vancouver, British Columbia, Canada
M. Cambridge
Knight Piesold & Partners, Ashford, Kent, England
359
360
INTRODUCTION concentrations, to an allowable change up to prescribed
Knight Piesold have been involved in the disposal of a receiving water quality limits. In many cases this requires
wide range of mine waste products throughout the world a zero discharge of any liquids associated with the waste
for in excess of forty years, and have been r~sponsible for being stored.
planning, design and construction supervision, and The basis for any waste disposal facility design
decommissioning of tailings dams during this period. The generally requires a comprehensive knowledge of existing
resulting mine waste disposal experience has been or background conditions at the site, the physical and
developed against a background of widely varying chemical characteristics of the waste to be stored, and
legislation and often in its absence. The consistent site specific containment features or construction
design requirement has been the safe, efficient and materials.
environmentally acceptable disposal of waste products. Establishing background conditions requires
During the last ten years, the mining industry has found identification of resources potentially at risk, whether they
itself operating under increasingly stringent environmental be water quality, fisheries, soils, air quality, vegetation or
legislation, much of which has been developed in the habitat of a rare and endangered plant or animal
response to public concern, inadequate environmental species. Particular emphasis is required to establish site-
standards and the effects of industrial pollution. The specific hydrology, hydrogeology and water quality data
leading role in the increasing tightening of legislation has over at least one year of monitoring. This information is
been North America with the European Community required to assess potential impacts of the proposed
following their environmental lead. As a result mining design and to provide actual monitoring data during
companies have found themselves operating under strict operations. For existing operations undergoing
guidelines with little flexibility during planning and expansion or modification, background water quality data
development of new mines due to the often restrictive may be significantly modified by past practices, and the
legislative framework within which they are constrained to assessment of new designs may be based on
work. improvements to background conditions.
The experience of the firm based on worldwide Site specific geotechnical information and
project experience and legislative regimes is that a more containment features, such as natural clays, glacial tills or
pragmatiC and flexible approach to waste disposal is impervious bedrock, and the extent and possible use of
feasible without detriment to the environment. This paper these materials as part of the waste facility deSign, also
compares the current legislative environment on either form an essential part of the site data relevant to any
side of the Atlantic and indicates where such a flexible design. Geotechnical data can often be obtained in
approach has been successful and provides case conjunction with baseline hydrogeological investigations.
histories which show the benefits of this attitude in mine
planning and environmental protection. Waste Characterisation
Characterisation of the waste to be stored forms an
WASTE DISPOSAL LEGISLATION IN NORTH AMERICA essential starting point in the design of any containment
General Approach facility. For mining wastes, some standard tests are
Legislation governing the design of mining and industrial usually required to determine the potential for acid
waste disposal facilities throughout North America is generation and whether the material should be classified
generally based on the prevention of any significant as a special waste.
impact on beneficial uses of groundwater and surface Acid generation is the atmospheric oxidation of
water downstream of the site. The focus is on the control sulphides, with the ensuing chemical and biological
of liquid wastes, it being assumed that solids wastes can processes resulting in a lowering of pH and leaching of
be fairly easily contained. Significant impact on heavy metals into water passing through the material.
downstream beneficial water usage is variously interpreted The prediction of acid generating potential is based on
to mean no detectable change in any background static acid/base accounting methods or more extensive
361
kinetic tests. A static test includes the following: • Leach tests with carbonic acid (to simulate long-
• Sulphur content determination (%S) term leaching with rainfall).
• Calculation of the theoretical acid generating • A full analysis of the tailings liquid phase. and any
potential (AGP) other liquid waste streams from the process.
• Neutralisation potential determination (NP) Once the waste characterisation data has been
• Calculation of the net neutralisation potential developed. it forms the basis. together with baseline
(NNP) environmental conditions. of the prescriptive design or the
The results of these tests determine if more development of a site specific design to prevent or
extensive kinetic tests are required. which involve mitigate environmental impact. How the above data is
weathering under laboratory or site conditions in order to applied in different legislative arenas is discussed in the
establish the potential for and rate of acid generation. following sections.
One such laboratory test is referred to as the Humidity TABLE 1
Cell test. Since the onset of acid generating conditions is WATER QUALITY CRITERIA FOR CAUFORNIA (1)
rate controlled. the results of kinetic tests can be very Water Quality Criteria (3)
Maximum
important in determining the need for short-term control Contaminant Drinking Freshwater
technology. which may differ from the long-term Parameter Levels (1) Water Aquatic LHe
for characterisation of the waste materials from the actual (3) Source: Quality Criteria for Water (USEPA, 1976).
process flow sheet(s) is as follows: California
• Physical characterisation of the solids to determine The design of any waste management unit in California is
geotechnical properties. regulated by the California Administrative Code. Title 23.
• Acid/base accounting on the solids (to test for Chapter 3. Sub-chapter 15 "Discharge of Waste to Land".
acid generating potential) and kinetic tests if which is administered by Regional Water Quality Control
appropriate. Boards. Sub-chapter 15 is an all embracing piece of
• Leach tests with acetic acid (required for special legislation written primarily for municipal and industrial
waste claSSification). wastes. It includes some special provisions for the
mining industry in Article 7 "Mining Waste Management".
362
The requisite waste characterisation testwork in Based on the classification of the waste, Table 7.3
California differs from all other States, in that the EP of Sub-chapter 15 provides prescribed liner requirements
toxicity test is replaced by the California WET (Waste as reproduced in Table 2. Most non-hazardous waste
Extraction Test) test, which uses a citric acid leach in products would be classified as Group B, and would
contrast to acetic acid. Relevant maximum contaminant require a double liner under the entire facility, consisting
levels for the WET test and water quality standards for of an outer liner of 2 feet of clay with a permeability of
California are shown on Table 1. 1 x 10.6 cm/s or less, an inner liner of clay or synthetic
On the basis of the results of acid/base membrane, and a blanket drain between the two to
accounting, the WET test, water quality standards and monitor leakage through the inner liner and reduce the
antiCipated process reagents, waste materials are head acting on the outer liner. This drain is termed the
classified as follows: Leach Collection and Removal System (LCRS).
Group A: Hazardous wastes that pose a significant While attempting to provide complete protection
threat to water quality. by the provision for double liner systems, these
Group B: Waste containing non-hazardous prescriptive standards have some in-built deficiencies
constituents with a low risk, but with some which can result in increasing seepage and the creation
potential for water quality degradation. of a long-term liability problem for the mine operator.
Group C: Inert material causing no degradation of Implicit in the regulations is the notion that the
water quality except for turbidity. prescribed liner systems will reduce impacts on
(1) Synthetic liner may be used for short-term containment (see Sub-section 2572 (f)(1) of this article).
(2) Liner and leachate collection an removal system for tailings pond must be able to withstand the ultimate weight of wastes.
(3) Permeability of ~1 x 10-6 em/sec or natural geologic materials may replace outer liner of double liner system.
(4) Single clay liner ~1 x 10-6 em/sec) for surface impoundment, to be removed before last 25 percent (minimum 1 foot thickness) of liner is
penetrated by fluid, including waste and leachate.
363
groundwater to acceptable levels. However, a simple conditions to assess net changes in baseline
calculation using Darcy's Law to calculate seepage concentrations in receiving waters (groundwater or
through a 10.6 cmjs outer liner under unit hydraulic surface water). British Columbia, however, is prolific in
gradient (for vertical flow under zero head, i = 1.0), shows fish, which are protected under both Provincial and
that for any reasonably sized facility the seepage quantity Federal legislation by very strict requirements. Examples
can be quite large (8.6 m3jdjha) and could in fact have a of special waste criteria and maximum allowable
significant impact. Furthermore, by containing all liquid concentrations for selected constituents in receiving
wastes under fully saturated conditions within an inner waters are shown on Table 3.
liner, seepage through the inner liner can be very This approach differs from California in that the
significant and will persist for a very long time. design of any waste containment facility can be varied to
California regulations do include, however, an meet the specific receiving water criteria, depending on
allowance for engineered alternatives to prescriptive actual waste characteristics and site specific conditions.
standards providing that equivalent or better water quality
protection is demonstrated. Current Trends
which, despite its shortcomings, was responsible The Mines and Quarries (Tips) Act and the subsequent
for the reduction of air pollution in the UK 1971 Regulations do not impose pollution controls other
than those relating to safety i.e. control is imposed on the
• The 1951 Prevention of Pollution Act in Rivers,
instigated to prevent uncontrolled river discharges disposal of solids alone. The pollution element in the UK
and, for the first time, imposing licenses on has historically been controlled by the Town and Country
discharges into rivers. Planning Act and air and water pollution legislation
Each of the above were non specific with general previously mentioned. These were considered to be
highly developed planning controls designed to ensure
365
that the environmental and other consequences of Early directives from Brussels were aimed at specific
individual developments were fully considered before pollutants the so called "fire fighting" directives which set
being allowed to proceed. County Planners thus limits and standards for the control of specific
undertook a controlling role to the benefit of the public substances. These led to the 1976 EC directive on
and of the environment but generally in a non-specific control of discharges to water limiting such elements as
guise. Where appropriate as for mine dumps, advice was mercury, cadmium and arsenic. At this stage legislation
taken at the planning stage from relevant statutory was reactive but by 1977 a preventive or precautionary
consultees such as water authorities who imposed approach was evident. In 1985 environmental
regulations on discharge consent limits. However, the assessments were enshrined in legislation and became a
lack of guide-lines for specific environmental elements centre piece for all future development plans under EC
together with absence of local mining experience often regulations and in 1987, care of the environment, was
led to a lack of uniformity in discharge controls for new enshrined in the Treaty of Rome. An increasing number
mining ventures. of environmental controls have been imposed on member
Since 1970, with increasing public awareness states since this time.
further legislation has been prepared, initially in advance
of European Community legislation, but more recently in The aim of the European legislation has been to provide
direct response to Brussels. The 1989 Water Act and the specific limits for air and water discharges and for land
establishment of the independent water regulating contamination with the aim of establishing common
authority, the NRA, imposed real controls on industrial standards across Europe. This approach though laudable
discharges. Similarly the Town and Country Planning is fraught with anomalies, particularly where geographical
Regulations 1988 enshrined formal environmental impact features conflict with EC standards, natural radon
emissions in Cornwall are such an example. The
assessment in the planning process for all new
proposed controls therefore make little allowance for
developments. More recently the 1990 Environmental
individual site conditions and the following example
Protection Act of the United Kingdom has been enacted
indicates the potential effect on new mine projects.
which with the 1989 Water Act, has established
A new mining venture in the UK required consent
centralised regulatory authorities ie. the Inspectorate of
to discharge effluent from the tailings dam which would
Pollution, the Health and Safety Executive and the
be based under EC guidelines on a consent standard for
National Rivers Authority. The Act is a move towards
the receiving water. In the specific locality the receiving
stricter regulation introduced integrated pollution control
water standard to be achieved was the directive for
rather than EC element specific standards. The
salmonides 78/659/EEC. Quality analysis indicated that
introduction of integrated pollution control will enable the
the natural river quality upstream of the site was, for both
UK to move towards a more site specific approach, and
hydrogeological and geomorphological reasons, acidic
alone amongst its European competitors enable
with a natural heavy metal content resulting from contact
consideration, at any site, of the effect of individual
with mineral outcrops within the river valley. The natural
elements, from any process on air, water and soil. This
level of copper in the river was significantly above the
system, once fully enacted, will enable rational site
requirements of the EC directive. This standard therefore
specific consent and discharge controls to be imposed on
could not logically be applied at this site on tailings dam
all future mining projects and provide greater emphasis
discharges. For this project a more flexible approach
on good engineering practice to the benefit of all.
based on site conditions was clearly required and could
be employed under new UK integrated pollution control
Present European Controls
criteria.
The European Community has had an increasing
A strict but more flexible approach has been taken
influence on environmental legislation throughout Europe
in Ireland which has a shorter history of mining but has
and will continue to do so particularly as a result of the
had some notable environmental problems from old tips.
move towards the creation of a genuine single market.
366
In particular full protection against dust and of Control of seepage must involve best appropriate
groundwater is require (80/68/EEC) to prevent discharges technology given the site conditions. However, the
of significant quantities of "certain dangerous" substances, temptation is to impose full containment by an impervious
Table 4. Full containment by an impervious liner is not liner regardless of site conditions. New mines thus face
mandatory but it must be demonstrated that the quantity significant costs at the planning stage in order to avoid
and quality of seepage to groundwater is acceptable. imposition of inappropriate quality standards by the strict
interpretation of EC directives.
Removal of all excess water during operations will The initial stage of tailings management facility
eliminate the potential for long-term seepage. Air-drying was constructed in 1986 with on-going expansions of the
and consolidation of the tailings will facilitate surface liner system and embankment raises. Performance of the
capping and revegetation measures. TMS has been carefully monitored by piezometers and
CASE HISTORIES flowmeters within the facility and water quality testing on
Jamestown Mine. California process streams and groundwater monitoring wells. The
Sonora Mining Corporation constructed the Jamestown monitoring programs confirm that the system is
mine, a 7,000 ton per day open pit gold mine with a froth performing as designed and that the water quality within
flotation ore concentrator in 1986. The flotation tailings, the facility is very good, and little different to the natural
368
groundwater. Detection monitoring for process reagents The tailings were to be separated by gravitational and
has confirmed that there is zero impact by the TMS on flotation methods and deposited behind a dam
local groundwater. constructed in an adjacent valley. The tailings dam was
designed to incorporate the coarse fraction of the waste
Gold Processing Plant. Greece within the confining wall. Thus fully drained tailings
Preliminary designs have been prepared for the tailings comprised the dam wall whilst the fine tailings were
storage facility for a proposed gold processing plant in depOSited into the reservoir by controlled beaching to
Greece. The process involves pressure oxidation of an achieve a semi-drained state (Fig. 2). The zoning of the
arseno-pyrite concentrate, and carbon-in-Ieach gold
) t
N
recovery, with a tailings product consisting of gypsum,
calcium carbonate, ferric hydroxide and ferric arsenate.
Waste characterisation testwork was carried out on
pilot test samples of the tailings as outlined in
Section 2.0, resulting in a classification as non-hazardous STREAM DIVERSION / \
CHANNEL ~./ ~--"'.......-____~
based on the EP toxicity test criteria. The carbonic acid
leach test indicated leachable constituent concentrations
slightly above drinking water standards for a few /
parameters. The leachate generated in the carbonic acid /
leach was then passed through the proposed clay liner
/
material to investigate the effects of attenuation, which
/
was indicated to be significant. I
The resulting proposed design of the tailings I
storage facility consisted of a clay liner with overlying I
drainage blanket over the base of the facility, an
\ I DECANT
impervious confining embankment, and a separate
recycle pond for the underdrainage and decant water.
\ RETURN WATER
,\UMP STATION I
r
II EFFLUENT
CHANNEL
N
trials are a positive indication that with a more flexible
approach to deposition, appropriate technology can be
applied to prevent environmental effects, improve
geotechnical characteristics, and therefore long-term
restoration potential.
CONCLUSIONS
The experience of mine waste disposal schemes
worldwide against a background of increaSingly restrictive
legislation indicates that economic development is not
necessarily served by rigid global standards. A more
FIG 4 CERRO DO LOBO TAILINGS DAM - SCHEMATIC LAYOUT flexible approach will promote good engineering practice
and environmental protection. Future legislation should
take account of the need for site specific requirements for
Knight Piesold were commissioned to review
new developments and promote technologically
tailings disposal processes in 1988 and to consider
appropriate solutions to the improvement of industrial
methods of improving the deposition characteristics of the
activity and environmental protection.
tailings. Initial laboratory studies had indicated that
significant density enhancement could be achieved
References
without acid generation in the tailings solids. Sub-aerial
1. California Administrative Code, Title 23, Chapter 3,
deposition trials using spray bars were therefore
Sub-chapter 15 "Discharge of Waste to Land".
undertaken within the tailings depository. The spray bar
2. Pommen L.W. Approved and Working Criteria for
trials were undertaken during the hottest and driest period
Water Quality. Government of British Columbia, Ministry
of 1989, and measurements of pH and moisture content
of Environment and Parks, Water Management Branch,
were carried out on the deposited tailings on a frequent
Resource Quality Section. April, 1987.
basis. Accurate measurements of the dry density of the
3. Guidelines for Freshwater Aquatic Life. Canadian
semi-solid tailings was not undertaken but an indication of
Council of Resources and Environment Ministers
density was obtained from the moisture content
(CCREM).
relationship and showed a significant improvement, by a
4. Special Waste Regulation. Waste Management
factor of between 25 and 30%. At the same time, where
Act, Government of British Columbia, April, 1988.
tailings deposition was controlled, a pH value of above 7
5. Knight A.B. and Haile J.P., Sub-Aerial Tailings
was maintained throughout the trials showing little
Deposition with Underdrainage. Seventh Pan American
evidence of acid generation. Where uncontrolled
Conference on Soil Mechanics and Foundation
deposition occurred, rapid oxidation and depression of
Engineering. June 19-24, 1983, Vancouver, B.C.
pH occurred, resulting in low pH values of between 2 and
6. Haile, J.P. and East, D.R., "Recent Developments
3 locally.
in the Design of Drained Tailings Impoundments".
371
The conventional fluidized bed furnace con- In some cases, special combustion systems are
sists of a vertical cylindrical combustion cham- needed to achieve the required clean up effect. A
ber provided with a refractory lining. The typical example of such a case is the DAG site in
fine-grained soil is decomposed, dried and its Stadtallendorf/Hesse (Fig. 4). In this reme-
organic constituents are burnt in a suspended diation projekt, concentrated TNT sludge ob-
bed of quartz sand. Uniform air distribution, tained form an upstream I urgi DECONTERRA soil
which is crucial to complete destruction, is a- washing plant is injected into a vertical combus-
chieved through a special configuration of the tion chamber. Using natural gas as a fuel, the
nozzle grate (Fig. 2). The residue is entrained sludge has to be burnt at a high air-to-fuel
with the flue gas and collected in downstream ratio to achieve complete destruction of the
dust collectors. trinitrotoluene.
~
g
-
11!
Classical -I--- CirculaUng
Auldlzed Bed.
I
,
Auld Bed
- I
~ Gas
'1e\oe\\'I
y,eal\
Mean Solids Velocity
.:
: CNWlH ~fL:. SAICI
.... .
: -- • iJIl
~
.... II.aA.U".AI
.1 1.IIAt,H!JAllltt
.A.
III AVY MAli I....,
oICl LN"II
:• • •. .
COOIt....-,,,,,,
2) Paticle sizes of less than 1 mm have an ad- impossible. Therefore the fines fraction has to
verse effect on gravimetric sorting. be separated and discharged as residue.
In the DECONTERRA process, fines separation,
3) Particles coarser than 1 mm interfere with the i.e. desliming, is carried out in hydro-cyclones
subsequent selective flotation process as they applying a d 80 mesh of separation. The cyclone
tend to be sedimented. overflows with particle sizes of 0.015 to 0.02 mm
are descharged as slurry. This fraction is routed
The size fraction minus 1 mm makes the highest to the treatment residue without any further
demands on the sorting and classification pro- treatment.
cess and hence requires sophisticated process The specific surface area of the particles in
technology. Aprat from the particle size distri- the cyclone underflow is abt. 1000 cm2/g.
bution, Lurgi also determines the specific sur-
face area of the solids using the air permeabi-
lity method to obtain a precise picture on the TREATMENT OF COARSE FRACTION
type of bulk material involved. These measur-
ments yielded specific surface areas of less The coarse fraction is further processed in a
than 4000 cm 2/g for the fines fraction of the jig where the material is lifted up and down by
soil. In individual cases, values up to 16,000 a stream of pulsating water while being slowly
cm 2/g were analyzed. It is a known fact that the moved in horizontal direction. In the process,
adsorption capacity rises with decreasing parti- the heavy material is separated from the light
cel size and increasing porosity. However, as material, each fraction being drawn off sepa-
the desorption capability decreases to the same rately. The two fractions can be sorted into one
extent, this means that desorption of the conta- light and one heavy fraction each in another
minants by mechanical methods is practically downstream sorter. The heavy fraction which is
378
free from wood, coal and porous components con- bour of Hamburg alone, 2 million cubic meters of
stitutes the coarse soil yield. dredging mud are obtained per year. Here again,
it is the fine silt fraction that presents the
greatest environmental hazard. Lurgi has solved
FINES TREATMENT this problem by agglomerating the highly conta-
minated silt fraction and immobilizing it in a
If required, the cyclone underflows are treated ceramic matrix (Fig. 6), i. e. vitrification.
in a second attrition scrubbing stage. Here, A brief description of the principle of the
they are washed in the absence of the highly process is given below:
contaminated fines fraction. This not only fore-
stalles re-contamination but is also beneficial The dredging mud is feed from the sand fraction
to subsequent sorting as the fraction 0.6 mm + and the highly contaminated silt fraction de-
is reduced through the scouring effect. watered to a residual moisture content below
50 % in a separation and dewatering circuit. The
The principal contaminant sources such as
dewatered silt is then mixed with ground return
coal, wood and roots have to be removed from the
fines (so-called soft-fired pellets) from the
fines fraction as completely as possible. This
travelling grate process.
is best achieved by selective flotation. After
Next, the silt/return fines mix is formed
comprehensive testwork, it has been possible to
into green pellets on continuously operating
find a suitable and ecologically neutral col-
Lurgi pelletizing disks before being indurated in
lector. The latter attaches to the contaminants
the Lurgi travelling grate.
and, after addition of a foaming agent, causes
Return fines are added as a leaning agent to
them to be concentrated in a layer of froth on
lower the silt's organic carbon content to a
the liquid surface. The highly contaminated froth
level which rules out uncontrolled temperature
is then united with the cyclone overflow and
peaks in the pellets during the subsequent fi-
dewatered in a thickener.
ring process.
The froth flotation unit can be operated with
either aerators or agitators.
In a final prrcess step, the green pellets
are subjected to controlled heat treatment on the
RESIDUE TREATMENT Lurgi travelling grate. After cooling in the
discharge-side section of the travelling grate,
The two fines fractions - cyclone overflows and the pellets are discharged followed by separa-
flotation froth - are flocculated and thickened. tion into product pellets and soft-fired pel-
Subsequently, the thickener underflow is dewa- lets.
tered on a filter press. This normally yields a
filter cake with a residual moisture of < 30 % The flue gases generated by the firing process
wt. As this filter cake contains more than 95 to are reacted out in a post-combustion chamber
99 % of the contaminants originally present in before being routed to a flue gas cleaning sys-
the raw soil, it has to be disposed of by ther- tem for desulfurization, dust collection and
mal methods. mercury removal. The flue gas cleaning system is
equipped with the latest control technology and
ensures consistently low emission levels.
AGGLOMERATION AND VITRIFICATION OF HARBOUR The strengths of this process can be summa-
SEDIMENTS rized as follows:
The highly contaminated harbour sediments ob- Exact control of the temperature profile and a
tained from dredging operations pose a special sufficiently long retention time in the high-tem-
challenge to remediation technology. In the har- perature range, i.e. at temperatures just below
379
the softening point of the dry silt, while ensu-
ring ample oxygen supply in the reaction zone.
The end produkt takes the form of ceramic pel-
lets, abt. 10 to 15 mm diameter, featuring good
compressive, shear and temperature cycling
strengths.
Treatment at temperatures above 1070 C re-
0
381
382
Table II shows in a summary a more de- T able II. Yield and Use of Wastes
tailed picture of the specific amount of wastes in Iron and Steel Production
obtained in the different steps of iron- and
steelmaking processes; the figures are average
values of the German works.
Process step Specific amount Main current use
The present use of these residues is also Ki nd of Wastes in kg/t
marked in this table. It can be seen that an
amount of about 20 - 25 kg wastes per ton of
steel produced today is dumped although it
contains a high amount of iron. I. Si nter Plant
With respect to the enviromental situation, EAF dust Grade "A" 10 - 15 kg/t St >20\ Zn content transfer
dumping of these materials cannot be consi- to B.U.S Waelz kiln
dered as a long-term solution. It also repre-
sents a loss of substantial quantities of iron as Grade "B" 10 - 15 kg/t St 10 - <20\ Zn recycled
well as zinc and lead for the national economy. to increase Zn content
Since dumping will become more and more Grade "e" 10 - 15 kg/t St dumped or disposed of
restricted, it is only a matter of time until all by outside companies
companies will be forced to look for other pos-
sibilities to get rid of these materials.
As the requirements to keep the air and IV. Rolled Steel Production
water clean have become more and more Mill scale, coarse ca. 40 kg/t RSt <0.5\ 0 il content:
stringent in recent years, the quantities of recycling to sinter feed
filter dust and sludge have been increasing
consistently . oi l-contami nated 3 - 4 kg/t RSt Mainly in-plant dumping,
mill scale sludge rest disposed of by
It was therefore mandatory to develop a outside companies
process to recycle these residues right back
into the process and thus reduce the quanti-
ties to be landfilled.
383
PROCESS SELECTION
contained in BOF dust are used for covering a
An essential condition for the selection was part of the process heat.
that the process should be based on proven
elements wherever possible so that the time Any feasible concept for the treatment of
and money to be spent on the development steel mill residues and the recirculation of the
could be kept to a minimum; the process products into the industrial cycle has to follow
should also be tolerant to different residue these lines.
mixes and reliably remove all impurities to
ensure that the iron materials can be safely This was the starting point for the
added to the production feed. In addition, the MET ALLGESELLSCHAFT Group to initiate the
Zn and Pb enrichment of the secondary development of an overall concept for the
product should be high enough to make it treatment of steel mill residues by ~erzelius
suitable as non-ferrous smelter feed. Umwelt Service (B.U.S) and LURGI in contact
with niYssEN STAHL AG [2]. This concept
All the steel mill residues to be treated provides for a firculating fluidized ~ed (CFB)
occur in the form of sludge or dust, i.e. of to recover materials suitable as feed for the
fine particulates. In order to reduce capital ferrous and non-ferrous metallurgy sector.
expenditure these particulates should be treat- The principle of Circulating Fluidized Beds was
able without agglomeration. This at the same developed by LURGI and today is proven tech-
time would ensure short heating and reaction nology. CFB reactors are used in numerous
periods, and hence high specific plant industrial processes of the chemical and metal-
throughputs, owing to the large reaction lurgical industry. Their applications range
surfaces of the particulates. However, since from simple physical processes to complex gas/
the Fe contents of the residues are not solids reactions. As evidenced by numerous
comparable with those of rich ores and since applications such as calcining of alumina or
the residues may contain certain impurities, combustion of coal in power stations, LURGI
such as sulfur, the Fe product is unsuitable has for more than 20 years now employed and
for direct use in the steel mill. No attempt is fully controlled the circulating fluidized bed
therefore made to achieve complete reduction to technology also in large industrial complexes.
metallic iron, and this in turn leads to savings
of energy and reducing agent. A special feature of the circulating flui-
dized bed is that it can handle fine particu-
This approach is supported by the possibi- late material of some particle size distribution
lity of using the fuel content in the mix for without any previous conditioning. This ensu-
process purposes; 20 % carbon in the sludge res that the residues can be treated as a mix
from the blast furnace gas scrubber, 4 - 10 % even if the particle sizes of the individual
oil in the mill sludge, and the metallic iron components exhibit considerable variations.
Q Gas
"C" 0
"S" {1
+--- Solids
"A" {1
.. '
Increasing Expansion
Filter 2
• 1 1J J
100
..,.
!"'
1 II'-'IIII"-
%
90
I (\.
:n ( ~~ '''
80 Crp
-:=
!5
VI
VI
oQ.
o
Z
<
o
<
LU
-'
..
-
u
z
N
~
o
-'
:;
i
LU
a::
0~-4-~~4-4-~~+-+-+-~~~~~~~~
o i 8 12 l' 20 2i i 8 12 l' 20 2i i 8 12 l' 20
13.12.1989 li.I2.1989 15.12.1989
Fig. 6. Removal of ln, Pb and K from steel mill residues in CFB-pilot plant tests
388
FURTHER DEVELOPMENT
After the preliminary tests on a bench Another part of the solids is collected in a
scale and in the pilot plant at LURGlls R&D secondary cyclone and led to the primary
Centre have proved successfully, the prepara- product discharge system. If necessary this
tions are now under way for tests on an ex- part of solids can be recirculated to the
tended pilot scale in order to develop the CFB-reactor as well.
technical systems, optimize the process and
derive the basic data required for a commercial The use of hot cyclones ensures that most
plant. of the volatilized elements which are unwanted
in the steel mill are separated from the solid
For this purpose, a pilot plant with a residue.
throughput of approx. 4 t/h currently is un-
der construction on the premises of THYSSEN As a consequence of evaporation and re-
STAHL AG. oxidation Zinc, lead and alkaline compounds
are concentrated in the waste gas dust load in
A process flow diagram of this pilot plant the form of oxides.
is shown in B.9..:....l.
The hot gases leaving the secondary cy-
The granulate is fluidized by means of clone are cooled directly by injecting water in-
preheated coke oven gas which is blown into to an evaporation cooler. This cooler is fol-
the reactor through the nozzle grate. This lowed by a multi-stage fine dust collection sys-
primary gas is heated to approx. 600°C in a tem including a cyclone and a cartridge filter.
gas heater. A second gas heater raises the The secondary dust recovered in this system
temperature of the combustion air to approx. contains Zn/Pb oxides. It is further cooled,
750°C before it is blown into the CFB reactor stored in a bin, wetted for dust-free trans-
as secondary gas. Oxygen may be added to port, and prepared for shipment to downstream
the combustion air to increase the reactor processing units. Some of the coarser fractions
throughput. of this secondary dust can be recycled to the
process, upstream of the granulation stage, in
The injection of primary and secondary gas order to obtain higher percentages of zinc and
leads to the formation of an expanded fluidized lead in the fine dust as secondary product.
bed in the CFB reactor in which the particles
from the gas flow are reacted without reducing The clean gas is led into the existing blast
the iron oxide to metallic iron. A certain per- furnace gas mains so that the steel mill can
centage of the solids which are entrained with benefit from its remaining calorific value. This
the waste gas at the reactor top are removed process thus does not produce any residues
in a recycle cyclone and fed back to the reac- that have to be landfilled nor does waste water
tor bottom via a pressure seal. need further treatment.
to existing
~-'/Gas System
Zn+Pb
Air Evapora'-l-io-n-C-o-o-le-r--JL..Fi-ll-er.l.-----contalning Dust
Air Preheater
Gas Preheater
The selective reduction process was veri- 1. J.A. Philipp, R. Goren et aI., "Umwelt-
fied by tests with residues from THYSSEN schutz in der Stahlindustrie, Entwick-
STAHL AG in a modified 0.2 m CFB plant at lungsstand - Anforderungen - Grenzen",
LURGl's R&D facility. An extended pilot plant Stahl und Eisen 107 (1987), No. 11,
to be used for long-term tests and to derive pp. 507-514.
upscaling data is currently being constructed
on the premises of THYSSEN STAHL AG. To 2. M. Hirsch, A. Kaune and H. Maczek,
ensure that all aspects of CFB technology can "Recovery of Zinc and Lead from Steel-
be covered, this pilot plant will also be Making Dusts, in particular by the Circu-
equipped to reduce the iron content of the lating Fluid Bed", 28th Annual Conference
residues to metallic iron. of Metallurgists, Halifax, 1989.
After some 20,000 t of THYSSEN residues 3. L. Reh, "The Circulating Fluid Bed
have been treated, the data collected by then Reactor", First International Conference on
will allow us to say whether the concept is Circulating Fluidized Beds, Halifax, 1985.
technically and economically feasible on an
industrial scale. 4. M. Hirsch, K. Janssen and H. Serbent,
"The Circulating Fluid Bed as Reactor for
Further tests will then follow to find out Chemical and Metallurgical Processes" ,
whether the know-how derived in this way can First International Conference on Circu-
also be applied to steel mill residues from lating Fluidized Beds, Halifax, 1985.
other producers.
5. F. Johannsen, "Die Technik des Wiilzver-
3 to 4 regional CFB-plants - integrated in fahren", Metall und Erz (1984),
the gas power system of the respective smelter pp. 235239.
- will be able to treat the whole amount of the
said residues of the German steel industry. We 6. H. Maczek and R. Kola, "Recovery of Zinc
believe, that the presented CFB-process will and Lead from Electric-Furnace Steelmaking
be a contribution to the solution of an increa- Dust at Berzelius", UNEP Industry and
singly burdensome environmental problem. Environment, July/August/September 1986,
pp. 11.
391
392
for removing heavy metals from the effluent to (2) Scale up:
levels acceptable to the authorities. The three
- using a 9m 3 sludge blanket reactor at
most suitable processes selected for research
Budelco's manufacturing site to establish the
were, ion exchange, liquid membrane extraction and
feasibility of the process with the site's
anaerobic microbial sulphate reduction. Of these
groundwater.
three processes only the last will remove sulphate
economically.
THE SRB PROCESS
This paper concentrates on the experimental work
Background
carried out in establishing the feasibility of the
Sulphate-Reducing Bacteria (SRB) process. A Many natural anaerobic aqueous environments, as
second paper l describes further process found in oceans, lakes and sediments 2 - 14 , contain
development work at Budelco b.v., Budel-Dorplein, micro-organisms which use organic compounds for
The Netherlands, a company owned jointly by growth and as the energy source for reducing
Billiton (Shell) and Pasminco; this will culminate sulphate to sulphides. When heavy metals are
in commercialisation of the process in 1992. present, they will precipitate as extremely
insoluble sulphides. These organisms playa major
The key features of this process are in situ
role in the formation of certain sedimentary metal
generation of H2S by microbes in a reactor and
sulphide deposits as well as producing the iron
concomitant precipitation of heavy metals as
sulphides found in coal IS -17. The overall
highly insoluble sulphides (Table 2).
reaction in such an anaerobic environment can be
Table 2 Divalent metal ion solubilities for metal represented by:
sulphides
Metal sulphate + Carbon substrate )
- identification of a simple, readily available, (1) Neutral pH. In nature this is achieved
economically attractive carbon substrate; through chemical equilibria involving either
precipitation or dissolution of carbonates and
- process studies in continuously operated 1.6 1
sulphides, which are products of organism
reactors using a synthetic waste water to
growth;
ascertain the operating parameters;
393
(2) Low redox potential. This is maintained in Synthetic solutions, based on Bude1co groundwater
nature by the presence of sulphide ions in the (Table 1), were used as feed media during the
aqueous phase and the absence of oxygen. studies in order to investigate the effect of
various carbon substrates, nutrients (nitrogen and
Table 3 Some sulphate-reducing organisms
phosphorus), potential microbial inhibitors and
heavy metals. General purpose grade chemicals
Organisms Growth Substrate
were used to prepare these solutions.
(Max. Temp. °c) (Product)
Screening for suitable SRB organisms
Desulfovibrio vulgaris (44) La (Ac)
Batch screening of the environmental samples was
Desulfomonas pigra (45) yeast (Ac + H2)
carried out by inoculating 100 m1 of non-sterile
Desulfobulbus propioniC'US (43) La,Pr,Et (Ac)
synthetic medium, in a septum sealed 150 ml
Desulfococcus multivora.ns (40) La,Ac,Et,Me (C02 )
bottle, with 5 m1 of each sample. The bottles
Desulfoba.cter postga.tei (40) AC,Et (C0 2)
were made anaerobic by addition of sodium sulphide
Desulfosa.rcina. variabilis (38) La,Ac,Et,Me (C02 )
(10 mg) and purging with nitrogen. Resazurin was
Desulfonema magnu.m (35) Be (C02)
added as a redox indicator to show when an excess
Desulfotomaculu.m orientis (38) H2 + CO 2 (-)
of hydrogen sulphide was produced by active SRB.
The synthetic media used for these experiments
La Lactate, Ac = Acetate, Pr = Propionate
were doped with a variety of carbonaceous
Be Benzoate, Me = Methanol, Et = Ethanol
substrates. Lactate was used since it is
Since these conditions can be attained in an known 3 ,4,18-21 to be an excellent growth substrate
anaerobic bioreactor, the expected concentrations for most SRB. Both ethanol and methanol were used
of heavy metals in the aqueous phase of such a as they have been reported 2 ,6-8,19,21-23 to be
reactor will be insignificant. The bioreactor may suitable carbon substrates for some SRB and are
operate under non-sterile conditions because cheaper than lactate. Since acetate is
sulphide is known to be a strong inhibitor for known 2 , 3,21,23,24 to be a metabolite of SRB when
other micro-organisms, therefore preventing grown on lactate or ethanol, acetate analysis was
microbial contamination. used to monitor complete degradation. The bottles
were inspected daily for solution colour changes
It should be noted that the quantity of sulphate
and for sediment colony growth, which initially
reduced is proportional to the growth of the SRB,
showed as small black specks in the grey sediment.
which in turn is related to the energy available
Small samples (5 m1) of the aqueous phase were
in the carbonaceous growth substrate. The
taken weekly, using a hypodermic syringe, and
sulphate conversion can therefore be controlled by
analysed for substrate, acetate, sulphate and
varying the amount of carbon substrate supplied to
various metals.
the growing culture. Acetate, which is often
produced by growing SRB, is only slowly degraded Environmental samples were obtained from a variety
by these organisms. However, methanogens, which of locations, including an oil refinery, sewage
normally exist in the naturally-occurring organism works, effluent treatment plant, river estuary,
consortium, rapidly degrade this compound. stagnant pools, water channels, the Bude1co zinc
refinery and an anaerobic biotreater. The
EXPERIMENTAL majority of these cultures showed SRB activity
when grown on lactate and a large number of the
Laboratory studies
cultures were also able to utilise ethanol.
Test feed compositions Although methanol was consumed by some cultures
394
practically no sulphate was reduced. perfect steady state operation was never achieved
in this reactor system some valuable information
Continuous experiments
on residence time and sludge bed buffer capacity
In total, eight continuous reactor runs were was obtained. Complete degradation of ethanol and
carried out in the laboratory using three precipitation of heavy metals were obtained at a
different types of reactors. residence time of only 4 hours. However, during
operation we noticed channelling (possibly owing
1. Stirred tank
to compaction).
Stirred tank reactors are very suitable for
studying the effects of various operating 3. Raked sludge blanket
parameters on the chemistry and microbiology of
Biomass can be retained in a properly designed
the process since they offer consistent, reliable
sludge blanket reactor where the sludge blanket is
and rapid equilibrium conditions.
mechanically raked or suspended by liquid recycle
Basic data required for the process were obtained to prevent compacting or channelling or both of
from this reactor type. Three runs were carried the high density sludge. Successful operation of
out in 1.6 litre stirred (200 r.p.m.) reactors laboratory scale raked sludge blanket reactors was
with pH and temperature control. Figure 1 is a demonstrated.
schematic diagram of the equipment used.
The last four runs were carried out in a 0.75
Figure 1 Schematic of laboratory reactor system litre reactor where the reactor temperature was
controlled at 35 0 C. This reactor configuration
overcame the sludge compaction and channelling
that occurred in the fixed bed reactor.
V&I!!nl
Process description
SI.UOOE
TANK SlUllGE 10
RESULTS AND DISCUSSION
-
bIl -fr- Sulphate
The minimum residence time required in a sludge 8 --<>- Zinc xl00
blanket reactor to achieve the desired sulphate 300 -0- Ethanol
I::
0
reduction and complete degradation of carbon ~
-a
<> initial Pbo.phate - '1Omal-1
potential below -lOOmV, then heavy metal j' C initial Pbo.phate - 140mgl-1
t il)
concentrations of only parts per billion will
remain in solution in the reactor I s aqueous ...
II)
lIS
effluent. Whenever our reactors were operated .r:j;I.,
!Il
0
under conditions of good microbial growth, heavy .r:
Il.
metal removal was practically complete even for 0.1
-
II)
Fate of calcium and magnesium The inhibitory effect of potential feed components
on microbial growth and hence sulphate reduction
Because the sludge produced in the SRB reactor
must be considered. As expected, alkali and
will be added to the zinc refinery feed a build up
alkaline earth cations were found to have no
of calcium and magnesium in the sludge should be
deleterious effect since the organisms used in
minimised to prevent calcium precipitation in the
this process can be isolated from marine
zinc refining process and magnesium build-up in
environments 4 . Heavy metal cations appear to be
the electrowinning circuit. Experiments show that
non-toxic, presumably because the presence of
399
excess soluble sulphide maintains a very low metal These depend on the level of sulphate reduced, the
concentration in solution (see Table 2). Certain heavy metal concentration in the feed, and reactor
anions are, in general, known to be powerful pH providing optimal microbial conditions exist
microbial inhibitors. However, the following within the reactor.
anions were found not to inhibit microbial growth
or sulphate reduction at the concentrations tested CONCLUSIONS
(mgl- l ): molybdate (10), selenate (3.5), arsenate
Tighter environmental control within some Western
(3.3), fluoride (48) and sulphide (500).
European countries requires that sulphate and
Mass balance over the SRB reactor
heavy metal contaminated groundwaters should be
The build up of a SRB sludge blanket in the cleaned up to levels demanded by the authorities.
reactor is very slow, and it can only be This has resulted in the development of a sulphur
considered to be in perfect equilibrium with the and heavy metal removal/recovery technology based
feed and effluents after a long period of on the capability of some types of bacteria to
operation (months), providing the feed composition reduce sulphate in an anaerobic environment.
and reactor conditions have remained constant
From data obtained during batch and continuous
during the whole period. However, a sensible and
experiments in the laboratory and during the
reasonably accurate overall mass balance can be
demonstration unit operations, we conclude that:
attained from analysis of the feed, outflow liquid
and gas streams assuming the remaining elemental 1. Removal of a wide range of heavy metals and
balance is associated with the sludge. Figure 5 sulphate from aqueous streams can be achieved
shows such a mass balance over a SRB reactor which with the anaerobic Sulphate-Reducing Bacteria
has been operated under pseudo-steady state (SRB) process.
conditions.
2. SRB utilise ethanol as carbon/energy substrate
Figure 5 Mass balance over the SRB reactor (g) for reducing sulphate and subsequently form
GASEOUS EFFlUENT
acetate as a major product. The acetate can
Hydrogen Sulphide 94
be successfully degraded to carbon dioxide,
Carbon Dioxide 4
Methane (max) 132
methane and biomass by introducing methanogens
Ethanol 1132
BIOREACTOR 1739
Carbonate 3. The quantity of sulphate that can be reduced
Sulphate
Zinc
3050
250
pH = 7.5 Sulphate
Sulphide
500
639
Cadmium 1.5 Zinc <0.01
by the SRB process is controlled by the amount
Cobalt 0.25
30a C Cadmium <0.01
Capper 0.75 Cobalt <0.01
of ethanol fed to the reactor. One mol of
Calcium 900 Capper <0.01
Calcium 750
ethanol is capable of reducing one mol of
SLUDGE sulphate.
Biomass (CH,O) 245
Carbonat. 225 4. Dissolved heavy metal concentrations in the
Sulphide 124
Zinc 250 aqueous effluent from the SRB reactor are
Cadmium 1.5
Cobalt 0.25 extremely low (ppb) provided sufficient
Copper 0.75
Calcium 150 sulphate is reduced and the redox potential in
solution is maintained below -100 mV.
It should be noted that although there is a
5. Nitrogen and phosphorus concentrations in the
multitude of equilibria within the reactor, only
reactor feed must be greater than 5 %mol and
the elemental ratios of sulphur and carbon
0.2 %mol, respectively, of the ethanol
compounds in each effluent change drastically.
consumed to sustain microbial growth. Excess
400
phosphate should be minimised to avoid calcium p. 330-335.
precipitation.
4. Bergey's Manual of Systematic Bacteriology.
6. Temperature effect on the performance of the Volume 1. Krieg, N.R. and Holt, J.G. Published
SRB reactor is negligible in the range l5 0 C to by Williams & Wilkins. 1984, p. 663-679.
40 o C. For the organisms investigated,
5. Oremland, R.S. and Polcin, S. Methanogenesis
temperatures higher than 42°C must be avoided
and sulphate reduction: competitive and
to prevent killing the organisms.
noncompetitive substrates in estuarine sediments.
7. pH control in the reactor may not be required Applied and Environmental Microbiolology, vol. 44,
for treatment of groundwater with a pH greater no 6, Dec 1982, p. 1270-1276.
than 4 owing to the buffer capacity of the
6. Oremland, R.S., Marsh, L.M. and Polcin, S.
system.
Methane production and simultaneous sulphate
8. Addition of flocculant to the SRB reactor feed reduction in anoxic, salt marsh sediments. Nature.
is essential i f short liquid residence times vol. 296, 11 Mar. 1982, p. 143-145.
and adequate solid retention are to be
7. Widdel, F. and Pfennig, N. Studies on
obtained within a single reactor.
dissimilatory sulphate-reducing bacteria that
9. Unwanted sulphide present in the three product decompose fatty acids II. Incomplete oxidation of
streams (i.e. gas, liquid and sludge) from the propionate by Desulfobulbus propionicus gen. nov.,
SRB reactor must be removed before discharge. sp. nov. Archives of Microbiology. vol. 131,
1982, p. 360-365.
10. The anaerobic mixed culture employed is
robust, can handle many potentially inhibitory 8. Battersby, N.S., Stewart, D.J. and Sharma,
cations and anions, and recovers readily from A.P. Microbiological problems in the offshore oil
(substantial) process upsets. and gas industries. Journal of Applied
Bacteriology Symposium Supplement. 1985,
11. Energy consumption in the process is low since
p. 227S-235S.
it operates near ambient temperature.
9. Widdel, F. and Pfennig, N. A new anaerobic,
REFERENCES sporing, acetate - oxidis ing, sulphate - reduc ing
bacterium, Desulfotomaculum (emend.) acetoxidans.
1. Scheeren,-P.J.M., Koch, R.O., Buisman, C.J.N.,
Archives of Microbiology. vol. 112, 1977,
Barnes, L.J., and Versteegh, J.H. A new
p. 119-122.
biological treatment plant for heavy metal
contaminated groundwater. This conference. 10. Klemps, R., Cypionka, H., Widde1, F. and
Pfennig, N. Growth with hydrogen, and further
2. Laanbroek, H.J., Geerligs, H.J., Sijtsma, L.
physiological characteristics of Desulfotomaculum
and Veldkamp, H. Competi tion for sulphate and
species. Archives of Microbiology. vol. 143,
ethanol among Desulfobacter, Desulfobulbus and
1985, p. 203-208.
Desulfovibrio species isolated from intertidal
sediments. Applied and Environmental Microbiology, 11. Sorensen, J., Christensen, D. and Jorgensen,
vol. 47, no. 2, Feb. 1984, p. 329-334. B. B. Volatile fatty acids and hydrogen as
substrate for sulphate-reducing bacteria in
3. Laanbroek, H.J. and Pfennig, N. Oxidation of
anaerobic marine sediment. Applied and
short-chain fatty acids by sulphate-reducing
Environmental Microbiology. vol. 42, Jul. 1981,
bacteria in freshwater and in marine sediments.
p. 5-11.
Archives of Microbiology, vol. 128, 1981,
401
SUMMARY INTRODUCTION
Soil and ground water underneath the zinc Budelco, a company jointly owned by Billiton
production plant of Budelco (The Netherlands) (The Netherlands) and Pasminco (Australia),
are contaminated with heavy metals and is a zinc refinery in the Netherlands
sulphate. To avoid contamination of nearby producing over 200,000 tpa of zinc. The soil
drinking water aquifers in the distant and ground water underneath the plant have
future, a Geohydrological control System been contaminated with heavy metals and
(GCS) and a treatment plant for the extracted sulphate due to 100 years of industrial
contaminated ground water have been activity. Ground water analysis data, linked
developed. with computer based aquifer simulation models
For the design of the GCS system ground water indicate that the contamination will
analysis data and geological profiles of the eventually surpass the site perimeter. In
site were used to model the existing order to avoid contamination of nearby
geohydrological situation. From the model drinking water aquifers in the distant
several combinations of water extraction future, a Geohydrological Control System
wells were deduced. The model was then (GCS) has been developed. This GCS is
improved by introducing additional field work designed to contain the contaminated ground
data. Finally the calculated results of the water within the site borders by pumping up
GCS options were compared to the objectives ground water from strategically located
of the system, leading to the choice of a wells. The GCS will produce water
design base. Some 6,000 m3 /day ground water contaminated with sulphate and zinc as well
will have to be extracted from a combination as traces of other heavy metals.
of 17 shallow and deep wells.
From the extracted water heavy metals and Before using the water in the zinc production
sulphate must be removed before discharging process or discharging the water to the
the water to the public domain. Several water public domain, heavy metals and sulphate have
treatment options have been studied and to be removed.
piloted. A process based on activity of Different water treatment processes have been
Sulphate Reducing Bacteria (SRB) combines studied. Three options were selected for
sulphate removal and heavy metal removal in further investigations :
one process step. The anaerobic bacteria
reduce the sulphate to sulphide and - Ion Exchange (IX)
consequently metal sulphides will - Liquid Membrane Permeation (LMP)
precipitate. A demonstration plant showed - Sulphate Reducing Bacteria (SRB).
promising process results. A Dutch company
specialized in anaerobic water treatment was
contracted for the design of a commercial
scale plant. This design should include
effluent treatment for excess sulphide and
solids in order to meet the discharge The IX process was studied by Billiton
criteria. The proposed design was tested in a Research Arnhem (1). The process is based on
pilot plant including all main process units. an exchange of heavy metal ions with hydrogen
pilot plant test results for all process ions. In practice the water is led through a
units are presented. The collected column packed with a cation exchange resin.
information was used to adjust the plant Heavy metals are extracted from the water to
design. Final design for the commercial scale the resin. The resin has to be regenerated by
plant is described. a stripping liquor, in this case an acid
solution. Thus the process is semi-
continuous : one or more columns are in use
for heavy metal extraction producing clean
water while another column is stripped
producing an acid metal solution.
In summary it was concluded that :
- the process would remove heavy metals to
acceptable levels
- the process would not remove sulphate
- the process could be considered to be
standard technology and thus upscaling to a
commercial plant would be relatively easy
403
404
- for the regeneration of the resins a Based on the positive results achieved (4),
relatively large volume of stripping liquor it was decided to run a process demonstration
would be needed, resulting in a low metal test at Budelco's site. A 9 m' demonstration
concentration bleed stream to be handled by plant was developed with support from SRC.
Budelco's production process This plant was operated for 10 months to
- depending on the cation exchange resin the prove the process on a large scale and to
process would only be effective for a establish operational ranges for process
certain (group of) metal(s) parameters.
- operation costs of the process were Results of lab tests and demonstration runs
estimated to be relatively high. are presented in a separate paper (5).
Because of the proven capability of the
process to remove sulphate and heavy metals
simultaneously at acceptable operation costs,
Bude1co decided to start the development of a
The LMP process is a development of the full scale water treatment plant based on the
University of Graz (Austria) (2). The process SRB process. Paques, a Dutch company with
is a solvent extraction: metal ions in the experience in anaerobic water treatment
water are exchanged with hydrogen ions of the systems, was contracted for Basic Design.
extractant. Intensive surface contact is Although the SRB process was proven, the
achieved by mixing the extractant with the subsequent treatment of the process effluent
water in a column in counter-current mode. to remove excess sulphide and remaining
Again metals have to be stripped from the solids still had to be developed. In addition
extractant. In the LMP option the stripping Paques needed more information on design
liquor is emulsified in the extractant before parameters. An integrated pilot plant was
entering the extraction column. By breaking therefore installed, containing all basic
up the emulsion after the extraction, the components needed for a complete water
strip liquor, containing the metals, is treatment. The plant was operated for 8
separated from the extractant. The extractant months to prove the efficiency of all
is then ready for reuse. Thus LMP can be components and to fix design parameter
designed as a continuously metal removing values.
process.
The process was piloted at Budelco's site in This paper describes the design of the GCS,
1980, first in a one step and later on in a results of pilot plant tests and the design
two step pilot plant. (3) of a commercial scale SRB water treatment
In short, conclusions of the test work were plant.
- LMP would remove heavy metals to levels
just acceptable by the authorities
- the process would not remove sulphate GEOHYDROLOGICAL CONTROL SYSTEM
- extractant efficiency could be affected by
iron hydroxide in the feed water
- removal was selective for some metals Background
- the stripping liquor was an acid solution
with high metal concentration and thus easy
to handle by Bude1co's zinc production In 1892 the Kempensche Zinkmaatschappy was
process. founded in Brabant (The Netherlands). The
process used was a thermal reduction of
A 200 m3/h scale process design was made at oxidic zinc ores in horizontal retorts using
the beginning of 1989. From the design it coke as the reductant. Later the oxidic ores
appeared that capital costs would be were replaced by roasted sulphidic ores.
relatively high. Annual production was 50,000 tpa zinc. The
residues of the process, the so-called zinc
ashes, were used as landfill to level the
swampy site areas and make them suitable for
plant extensions. The zinc ashes, containing
up to 15 , of heavy metals were at that time
The role of Sulphate Reducing Bacteria in the considered to be inert material. As we now
environment is well known. Possible know, heavy metals are leached by percolating
application of SRB's in a process for waste ground and rain water. Zinc and cadmium
water treatment has been studied by Shell levels up to 200 mg/l and 20 mg/l
Sittingbourne Research Centre (SRC) since respectively are found at a depth of ten to
1987. The bacteria reduce sulphate to twenty meters underneath the zinc ashes.
sulphide. Consequently heavy metals in the In 1973 an electrolytic zinc plant was
water will precipitate as metal sulphides erected to replace the thermal process.
because of the extremely low solubility
products of these components. In principle Zinc is recovered with 98 , efficiency from
the process is capable of removing metals as sulphidic ores by roasting and subsequent hot
well as sulphate. acid leaching. Annual design capacity was
150,000 tonnes of special high grade zinc,
Several electron donors for the process have capacity today is 210,000 tpa. The residue of
been tested. First indications suggested that the electrolytic zinc process is an iron-
operation costs for the process, based on a ammonium- sulphate compound called jarosite.
lactate carbon source, would be too high. Jarosite, containing traces of heavy metals,
However, by the end of 1988 SRC reported that is considered to be chemical waste and is
ethanol was very efficient as a carbon source stored in HOPE lined ponds. Monitoring of the
for the SRB's. Budelco supported a labscale zinc concentration in the drainage systems
study by SRC to apply the SRB process for indicated that one pond is leaking. The
treatment of Budelco's GCS water. leakage contributes to ground water
contamination.
405
In the seventies a geophysical electrical Phase 3
resistivity investigation of soil and ground
water underneath the zinc plant was
conducted. The resistivity measurements The additional field work, mainly consisting
proved to be unsuited to quantify the extent of drilling new holes, was executed. Because
of the contamination. of new uncertainties arising from
After 1983 sampling probes were installed to interpreting the data, the scope of the field
investigate the soil and ground water work was extended and the number of wells
contamination caused by the leaking jarosite increased to 100.
pond. Results indicated a relatively small
contribution to the overall contamination
although locally high concentrations of heavy Phase 4 and 5
metals were found.
Two further studies on contamination by
leaching zinc ashes showed severe Reliability of the model was improved by
contamination of ground water to a depth of introducing the additional field information
30 meters. Contamination is however confined in the model. The adjusted model was then
within the industrial site. applied to calculate the effect of proposed
It was concluded to design a geohydrologica1 control options. Combining the results with
system to contain the contamination within the objective to contain the contamination
the site perimeter (6). within site perimeters led to the selection
of the control system.
For the design of the GCS two companies were The Basic Design of the selected GCS was made
contracted : Delft Geotechnics for model by Tauw. The design includes pumps, control
development and calculations, Tauw Infra valves and pipeline manifolds. A PLC based
Consult for field work, design and flow control system has been developed.
engineering.
The design program of the GCS involved 6
phases : Results
BUDELCO
MALPIE STRAMPROY
--------~oc~~xx~~~~~~~QQQP~~~ 40
25
NAP
$1
1 aquifer
25
50
o
75 :c::
z
~~~J~~~~~~~' 100 »
-0
125
~~;or150
(zanden van Pay)
175
200
ground waler
monllorlng
wells
Surlaca
- 25/30 m
ground water
flow ~
-100 m
0 ..
o,~
a.
0,
• biogas
ffluent
UASB
V-2 ~ tan<
R-l U/I.93
S-5 Striwer
Ibly : Ibl),,!lectrolytel
rIocx:ulint
E%l1 . : ~ m t
to
"-J
The Tilted Plate Settler is a well known During the tests the following process flows
settling unit for solids. Settling is were analyzed
improved by shortening the settling distance
through parallel tilted plates. The settler - feed water heavy metals
removes solids consisting of sulphur produced and sulphate,
in the SFF, metal sulphides and biomass redox
flushed with SRB effluent. The pilot plant - UASB effluent heavy metals,
settler had a volume of 12 m3 • sulphate,
sulphide,
acetate,
DynaSand filter phosphate,
redox
- UASB off-gas methane, carbon
The filter is a continuously refreshed sand dioxide,
bed. The water flows up through the bed hydrogen
leaving the solids at the bottom of the bed. sulphide
Air is pumped through a small diameter pipe (partly)
at the centre of the filter. This airlift - SFF effluent heavy metals,
transports the sand particles with attached sulphate,
solids to the top of the filter. During sulphide,
transport the sand is separated from the phosphate,
solids by sheer forces. At the top of the redox,
filter the sand is separated from the wash - TPS effluent heavy metals,
liquor by gravity and the solids pass out as sulphate
a slurry. The pilot plant contained a - DynaSand effluent heavy metals,
DynaSand filter with a bed height of sulphate,
approximately 1. 2 m and a surface of 0.6 m2 • sulphide,
phosphate,
A schematic view of the plant is shown in redox.
Figure 6.
Heavy metals are measured on clear solutions
to check the conversion efficiency. Analysis
Operation conditions on heavy metals of the total solution
(including solids) show the efficiency of
solid removal steps. Under normal operating
As a result of the demonstration tests of the conditions solids content is too low to be
SRB process the operational ranges of process measured directly.
parameters were known
The hydraulic operation range is 1.5 to 3 m3 /h The oxygen measurement was troublesome
(residence time of 4 - 8 hrs). Higher flow because of calibration problems. The redox
velocities cause solids flush-out. At lower value shows a sudden increase at low sulphide
flow rates the sludge bed mixing is levels which can be expected from theory. The
insufficient and the sludge solidifies sulphide concentration is the best control
causing channelling and inactivity. parameter.
During the tests a blockage of the reactor
The gas production was 7 m3/day at 2 m3/h feed occurred caused by a flocculant overdose. For
flow rate. The gas contained 70 - 80 % final reactor design cleaning facilities are
methane, 5 - 10 % carbon dioxide and 5 - 10 % important.
hydrogen sulphide. Gas production is strongly
dependent on activity of the methanogens for
acetate conversion. TPS and DynaSand results
Table 2 shows a mass balance of sulphur The TPS is very effective for sulphur
components over the UASB and SFF reactors. removal. In different tests at which
different flocculants were applied the
Sulphide removed by precipitation of metal efficiency was always better than 89 %.
sulphides was 50 mgtl. Sulphide removal via
off-gas was 140 g/m which can be converted to The DynaSand filter is an effective polishing
22 mg/l S. Sulphide stripped in SFF was 3 g/m3 step in the process. It is a rugged filter
to be converted to 40 mg/l S. that needs only a minimum maintenance.
However, an overdose of flocculant can cause
Efficiency of the SFF reactor is high : up to poor separation of sand and solids. The
300 mg/l sulphide in the SFF influent are result is a that sand is flushed out with the
decreased to less than 1 mg/l in the effluent washing liquor.
under aerobic conditions. Incidentally the
sulphate concentration in the effluent At stable operation solids concentration in
increases because of excess oxidation the UASB effluent is less than 100 mg/l. This
occurring during underload conditions but will result in less than 1 mg/l solids in the
increase never exceeds 50 mg/l SO•• DynaSand filter effluent and the
corresponding total zinc discharge is less
The conversion is strongly dependent on the than O. 3 mg /1.
air supply. (Table 3)
The redox and dissolved oxygen concentration
in the effluent were measured to check if
both parameters can serve as control
parameters for the conversion.
412
Table 2 Sulphur mass balance UASB and
SFF.
Influent flow: 1.5 m]/h, UASB off-
gas flow 0.24 ~/h, air flow 20
m]/h.
All components expressed as mq/l
S.
o 1/
m
g 0 .6
/
0.1
. - - - .d -- --\i O. Jl
0.2
0 . 07
o . .-'- ~ __L--,-...--...----L...-_ .~. _ __'__ _'___'--~ _ _ L _ --'-_~---'
o 3 6 9 12 15
10 10.1
u
9
/ ;l
7.6
"
1 -..l.
o 3 6 jj 12 15
Zn504·solulion
flare
CH4
INFLUE EFFLUENT
Final Design
417
418
conditions (25.0 ± 0.1 0c) by making use of a double walled For the calculation of the concentrations, activities and
electrolytic cell. Before each experiment, the electrode was activity coefficients (hereinafter called concentration para-
polished with diamond spray (Buehler Metadi, compound 5 meters) of all components in the solution, a set of 12 equa-
and 1 ~) and rinsed ultrasonically successively in twice tions with 16 unknowns is formulated 10,12. This means
demineralized water and chloroform. that 4 of the unknowns must be measured or calculated in-
Linear sweep voltammetry experiments were performed dependently.
at a scan rate of 1 mY s-l using a potentiostatlgalvanostat The activity of the uncomplexed silver ions (aAg+)f is de-
with built-in generator (Princeton Applied Research, PAR duced from the measurement of the equilibrium potential Eo
model 273) controlled by an mM/PS2 personal computer. of the solution versus a silver electrode14 :
Characterization of the silver deposits was done with
three surface analytical techniques: Auger electron spectro-
E /NHE = 0.799 + 0.059 log (a +)f (1)
scopy, AES, (PHI 545 with cylindrical mirror analyser), 0AglAg+ Ag
scanning electron microscopy, SEM, (Jeol JSM 50-A) and
electron diffraction (Seifert). Auger spectra were recorded The activity coefficients of uncomplexed S2032- and Na+,
in the E[dN(E)/dE] derivative mode under constant condi- (YS2032Jf and (YNa+)f and of the complex NaS203-,
tions 10•
YNaS203-' are calculated with the ion interaction model,
using tabulated values of specific parameters lO,15. This
I
RESULTS AND DISCUSSION model is chosen because of the high ionic strength of the
solutions (I > 0.1 mol kg-I).
Thermodynamic Aspects of Solution Composition Details concerning the combined measurement/calcula-
tion routine are found in 10,13. The concentration parame-
In AgN03 - Na2S203 - NaN03 solutions, all silver is, ters of the AgN03 - Na2S203 - NaN03 solutions are given
in good approximation, present as Ag(S203>23- and in Table I as a function of the total silver and thiosulphate
Ag(S203>35- 10-12. The amount of uncomplexed Ag+ is concentrations.
negligible, but its activity is nevertheless an important ther-
modynamic quantity. Between S2032- and Na+ the com- Determination of the Mechanism and the Kinetic Parameters
plex NaS203- is formed 13. Beside the complexes, uncom-
In order to determine which of the elementary steps, re-
plexed S203 2- and Na+ are present.
presented in Fig.l, are actually occurring and at which rate,
If (cAg+)t (CUt (aAg+)r ('i)r (YL)r (cNaL-)r (YNaL-)r cAgL23- YAgL23- cAgL3 5- YAgL35-
1 1.2 10-3 2.710-2 1.0 10- 13 2.310-2 0.65 1.6 10-3 0.87 1.2 10-3 0.51 - -
2 4.9 10-3 6.010- 2 1.010- 13 4.310-2 0.58 7.210- 3 0.85 4.9 10-3 0.39 - -
3 1.110-2 9.710-2 1.210- 13 6.210- 2 0.53 1.3 10-2 0.85 1.1 10-2 0.35 - -
4 2.010- 2 1.410- 1 1.3 10- 13 8.210- 2 0.49 1.9 10-2 0.91 2.010-2 0.31 - -
5 4.010- 2 5.8 10- 1 1.610- 14 2.910- 1 0.39 2.010- 1 0.98 2.410-2 0.24 1.4 10-2 0.09
6 1.010- 1 7.2 10- 1 4.210- 14 3.010- 1 0.32 1.810- 1 0.89 6.3 10-2 0.18 3.8 10-2 0.06
7 7.010- 2 9.210- 1 1.010- 14 4.010- 1 0.35 3.710- 1 1.04 3.410-2 0.18 3.010-2 0.06
8 1.110- 1 1.0 1.210- 14 4.110- 1 0.31 3.4 10- 1 0.96 5.210-2 0.11 5.710-2 0.Q3
9 1.4 10- 1 1.1 2.110- 14 4.210- 1 0.30 3.310- 1 0.95 7.3 10-2 0.13 7.010- 2 0.03
10 1.110- 1 1.3 0.910- 14 4.610- 1 0.32 5.110- 1 1.04 4.810- 2 0.11 6.1 10-2 0.03
11 2.010- 1 1.4 1.5 10- 14 4.610- 1 0.27 4.710- 1 0.97 9.010- 2 0.08 1.210- 1 0.01
12 1.3 10-3 2.5 10- 1 1.710- 15 1.710- 1 0.50 8.010- 2 0.98 9.9 10-4 0.36 3.3 10-4 0.18
13 5.5 10-3 5.110- 1 1.9 10- 15 2.810- 1 0.46 2.210- 1 1.08 3.410-3 0.27 2.010-3 0.11
14 1.3 10-2 7.8 10- 1 1.9 10- 15 3.6 10- 1 0.43 3.810- 1 1.15 6.8 10- 3 0.20 6.1 10-3 0.07
15 2.510-2 1.1 1.5 10- 15 4.110- 1 0.41 5.910- 1 1.21 1.110-2 0.12 1.5 10-2 0.03
(2a) Composition
-1/6 1/2 2/3 2!3 The Auger analysis gives a wide variety of informations
j = - F 0.62 v ro (D 1 2 c 1 2 + D1 3 c 1 3) (2b)
m " " about the surface layer of specimen of about 10 Athickness:
not only component identification 17-19, but also chemical
with: kO,O = 1.1 105 state identification20-26 and structural information27 . Com-
kl,O = 2.9 10-2 ponent identification is straightforward and goes by the de-
termination of the energy of the Auger peaks. Chemical and
u = 0.77
structural state identification are based on the line shape
1311 = 6.6 105 mol- 1 m3
analysis of the Auger peaks, and this requires the compari-
Dl ,2 = 6.8 10- 10 - 6.4 10- 11 I m2 s-1 (I moll-I) son of the spectra of the unknown sample with those of
Dl,3 =7.210- 10 _1.310- 10 1 m2 s-1 (I moll-I) standard samples of known composition.
T =298 K The complete analysis is described in detail in lO,16 (sil-
ver deposit and standard sample preparation, line shape
In the equations are : °
analysis of Ag, S and peaks on deposits and standards,
j, jc and jm respectively the total current density, and the identification by comparison), but only the results are re-
contributions of charge and mass transfer to it; ported here.
u and ki,O the kinetic parameters of charge transfer, res- A typical example of an Auger spectrum of the surface
of an electrodeposited silver layer is given in Fig.5, sho-
pectively the transfer coefficient and the potential inde-
pendent rate constant of Ag(S203)i 1-2i, with i = 0 or 1; °
wing the presence of S, Ag, and Na. ° and Na are not
present all over the surface, but always together. S is de-
Dl,i the diffusion coefficient of Ag(S203)il-2i, with i =
tected everywhere, but the intensity of its peak, relative to
2 or 3 as a function of the ionic strength I of the solution, the intensity of the Ag peak, is varying.
characterizing mass transfer;
420
From the line shape analysis in 10,16 is known that Ag is Also for the morphological aspects, the results are good.
not bounded to an oxygeneous anion, such as S2032-. S is In a wide range of plating conditions, smooth deposits are
obtained. A compromise must be reached between level-
thought to be present as an SO adsorption layer, in accor-
ness of the deposit and plating rate.
dance with the results 0(28. Remnants of the supporting
Based on the conclusions of this investigation, it is
electrolyte NaN03 on the surface are responsible for the
worth while to examine the engineering aspects of silver
presence of O. plating from thiosulphate baths.
Depth proftles indicate S, 0 and Na only to be present in
the top layers as contaminants. The bulk of the deposits is
composed of silver, and is polycrystalline, since the diffrac-
tograms completely correspond to those of pure polycrys- REFERENCES
talline silver lO,16,29.
1. LUCE B.M., FOULKE D.G.
"Modern Electroplating", ed. F.A. Lowenheim, 3rd
Morphology Ed., J. Wiley and Sons.
9. LEVICH V.G.
CONCLUSION "Physicochemical Hydrodynamics"
Prentice-Hall, Inc. Englewood Cliffs, New York, 1962.
Using a combination of electrochemical and surface ana-
lytical techniques, the possibility to perfonn silver plating 1O.HUBIN A.
from thiosulphate instead of cyanide containing baths is "Bijdrage tot de Studie van de Elektrochemische Reduc-
examined. tie van Zilverthiosulfaat- en Zilverthiocyanaatcom-
plexen", Ph.D. Thesis, Vrije Universiteit Brussel,
The kinetic investigation reveals two important facts.
Brussels, 1989.
The kinetic parameters are such that current densities of the
same order of magnitude as in the case of cyanide baths 11. "Gmelins Handbuch der anorganischen Chemie"
(100 A m- 2 1) can be achieved. So, from a kinetic point of Springer Verlag, Berlin-Heidelberg-New York, achte
view, the behaviour of the thiosulphate complexes is very Auflage, Ag[B3], 1975.
promising.
12.HUBIN A., VEREECKEN J.
The mechanism predicts the formation of a deposit compo-
1. Electroanal. Interfacial Electrochem., paper submit-
sed of Ag and [Ag(S2~)"]-' entailing that S2~2- is incor- ted.
porated in it. This is undesirable, but it is shown that poly-
crystalline silver is deposited, independent of the thiosul- 13.GIMBLETT F.G.R., MONK C.B.
phate concentration. Trans. Farad. Soc., 51 (1955) 793.
421
ted.
t+ 1
17.BRIGGS D., SEAH M.P.
"Practical Surface Analysis by Auger and X-ray Photo-
electron Spectroscopy", J. Wiley & Sons, New York,
1983.
18.THOMPSON M., BAKER M.D., CHRISTIE A., Fig.1: Schematic representation of the elementary steps
TYSON J.F. for the electrochemical formation of a metal deposit:
"Auger Electron Spectroscopy", J. Wiley & Sons, New E = electrode
York,1985. Ox = initial state of the oxidant
Ox* = electroactive state of the oxidant
19.CHUNG M.F., JENKINS L.H. Red = reductant, deposited on the electrode
Surf. Sci., 22 (1970) 479. S 1,S2 = electroinactive components
step 1 = charge transfer
20. MADDEN H.H.
step 2 = mass transfer
J. Vac. Sci. Technol., 18-3 (1981) 677.
step 3 = homogeneous chemical reaction
step 4 = heterogeous chemical reaction (4a and
21. CARRIERE B., DEVILLE J.P., HUMBERT P.
4b are adsorptions of respectively Ox*
J. Microsc. Spectrosc. Electron., 10 (1985) 29.
and Sl)
22.QUINTO D.T., ROBERTSON W.D. step 5 = electrocrystallization
Surf. Sci., 27 (1971) 645.
23.CHATTARJID.
"loa -I oa .) oa '1 oa
"The Theory of Auger Transitions", Academic Press,
J lIlEH/a2
1976.
25.WEISSMANNR.
Solid State Communications, 31 (1979) 347.
28.PERDEREAU M.M.
C.R.Acad. Sc. Paris, 274 (1972) 448.
·I~ ·1
kl
Ag(S203)- + e ~ [Ag(S203)-]-
·1 !O
Fig. 4: Mechanism of the reduction of silver thiosulphate
complexes. k and ~ are respectively the rate and
·1 III the equilibrium constant.
a)
.J 00
( 1·1'
!tJ
......
w
-Bw
5
0
1 c)
Ag
~--l
0 500 1000
E (eV)
Fig. 5 : Auger spectrum of the silver deposit. Fig. 6: SEM pictures of the silver deposit:
(a) at low current density,
(b) for a thick deposit,
(c) at limiting current density.
Magnification: 3000 x.
Use of peroxygens in treating cyanide
effluents from gold processing
E. N. Wilton
P. J. Wyborn
J. A. Reeve
Interox Research and Development, Widnes, England
LABORATORY STUDIES
Caro's Acid
Experimental Procedures
i) Generation of Caro's Acid
In the H202 detoxification reaction, a
catalyst such as copper is often required to Caro's acid was prepared in the laboratory
allow a faster reactions. This catalyst by the controlled addition of 70% H202
may have to be added, if it is not already into a cooled, stirred flask containing 98%
present in the effluent. The cyanate H2S04. 9 These concentrations were
produced from the detoxification reaction is chosen as the most convenient to handle
then hydrolysed by alkali to give carbonate safely in the laboratory. The reaction is
and ammonia, exothermic and its temperature was kept
below 10oC, to prevent decomposition.
CNO- +H2 O+OH- ----. COa 2-+NHa
The reaction can be represented as:-
or by acid to give carbon dioxide and
ammonium,
CONCLUSION 0
0.
0 lIII <40 80 80 100 Il111 1<40 180
All the laboratory work carried out to date TIME (rmuas)
has shown that Caro's acid can be used with
great effectiveness, in terms of its speed FlG1d
of reaction and its ability to detoxify many
cyanide complexes, for the treatment of COMPARISON OF H202 WITH H2S05 AT SAME STOICH
cyanide containing effluents.
120 r--;:::=============:;---'
J Zn(CN)4~+H202 Zn(CN)4~~+H2soa .1
100
80
FlG1a CIi
E 80
COMPARISON OF H202 WITH H2S05 AT SAME STOICH ~
40
I
1mr-------~================~------~
FREE~ FREE eN- + H2SD6
100 --&:- 20
FIG 1.
Reactions of Cyanide and Caro's Acid
stVsrloue pH
r----------;:======::;----------,
I~-~~ . ~~1:1
120
10
!z ao
0
40
20
......
0
0 3
Time (minutes)
4
•
427
Run No. 2 3 4 5 6 7 B
Initial CN- conc 130 129 94 115 160 115 145 102
mg/1
CN- conc after B.4 7.9 0.4 10 3.B 2.0 0.2 0.5
treatment mg/1
Caro's Acid Mole Ratio 2.3:1 1.2:1 1.7:1 1.B:1 1. B: 1 1.4 1.1 1.9
H2S04:H202
H2S05:CN Mole ratio 5.B:1 5.2:1 5.3:1 3.5:1 4.0:1 4.0:1 5.0:1 6.5:1
428
....:
..". . . : ,..)' . ~4.
~
Alkali ....... ~
................. \.
......... ....••• EFflUENT IH~
Cara's acid .............. 1 - '-------~
OrI,lnal 01...... polilan
Tl 2 - AlI.....at... 01...... .,.elllon
Feed tank 3 - OrI,lnal Caro'. acid doln, .,.Int ••'
'. 4 - AII.....atIv. Caro'. acid do.ln, .,.oIlon ....
Effluent ..................... ......
............
.::::::::::u,,"······
.... T2
........~.~~.~~ ..~~~ ....................
FIG 2
T3
Final tank
an organic foulant Unless these effects are height of 150 mm. The packed column contained
properly accounted for in models used for the 12,32 g of dry carbon and had a void fraction of
design of adsorption columns, such models will not 0,38, while in the case of resin it contained a
be reliable to predict performance over the volume of 0.075 L of wet-settled resin with a void
complete range of operation. fraction of 0.86. A sintered glass disk at the
bottom of the column facilitated an even
EXPERIMENTAL
distribution of solution. A metering pump was
used to provide a constant flow rate of 1.4 x 10-
Materials
6 m3 Is from a continuously stirred 600 liter
stainless steel tank The flow rate had to be
Potassium aurocyanide, potassium
monitored continuously to compensate for the
tetracyanocuprate and potassium argentocyanide
decrease in the level of the solution in the feed
were used as adsorbates. The latter was dissolved
tank A pH of 9 was maintained throughout the
in a potassium cyanide solution at a pH of 8 and
experiment
temperature of 90 C. The reacting solutions used
0
ultraviolet absorbance on a LKB Biochrom 4050 The mass balance for surface diffusion inside the
ultraviolet spectrophotometer. The pH of aU pores is:
solutions was monitored by using a Beckman
combination pH-electrode coupled to a Beckman
.71 pH-meter. (2)
MATHEMATICAL MODEL
If no accumulation occurs at the external surface
of the carbon, the boundary condition becomes:
A dual resistance model involving both external
ftlm diffusion and intraparticle surface diffusion Oq
Ou - (3)
was applied to the proftles for the uptake of gold, Or r=R
copper and silver cyanide by activated carbon and
ion-exchange resins in a packed column in order to For the resin (f = 1/.
estimate kinetic parameters. This model is similar
to the model used by Van Deventer 6 for the Local eqUilibrium at the particle surface:
adsorption of gold cyanide.
[for carbon] (4)
The following main assumptions have been made
in the development of the model: [for resin] (5)
qs = - - - - - - -
k2 + Cs
(a) The carbon and resin particles can be
treated as equivalent spheres for modelling Figures 1 and 2 show the isotherm results for gold,
purposes, silver and copper cyanide and the fitted
equilibrium expressions for activated carbon and
(b) The radial transport of silver or gold ·esin respectively.
cyanide, and the organics into the pores can be
described by a surface diffusion mechanism. Pore When equations (3) and (4) or (5) are applied to a
diffusion is assumed to be negligible, packed column, the value of q refers to a specific
height in the column. The value of Cs is guessed at
(c) Accumulation of metal cyanide or organics each step and the calculation of equations (2) and
in the liquid phase within the pores of the carbon (3) repeated with a fourth order Runge-Kutta
or resin is negligible, routine until equation (4) or (5) is satisfied.
Equation (1) is discretized using a backwards
(e) It is assumed that the adsorption reaction difference method and solved together with the
on the carbon, and the ion-exchange reaction on other equations.
the resin occur instantaneously, so that eqUilibrium
exists at the solid-liquid interface, This study also involved the simultaneous
adsorption of two components on activated carbon.
(t) In the packed column the liquid is in plug Although simultaneous asdorption was also
flow, and the carbon or resin is assumed to consist performed with ion-exchange resins as adsorbents,
of a number of completely mixed reactors. the latter did not reveal any equilibrium influence,
and will therefore only be considered to changes in
The liquid phase material balance for a packed
the kinetic parameters. The equilibrium model
column is:
equations are therefore developed on the basis of
3c oC6kf(1-e)A two solutes as described by Sheindorf et al. 3:
eA-- + v-- + ---------(C-Cs)=o (1)
at ax
434
,...., 80
z
0
m •
a 64
<r:
u
(.!)
~ 48
"-
...J
<r: u
I- 32
w
L
(.!) • COPPER
'-' 16 a GOLD
",. SILUER
(\)
CI
0
0 9 18 27 36 45
Ce [MG/U
10
,....,
z
..... •
(f)
w 8 • • • •
a
...J
"- 6
...J
<I:
I-
W
L 4
(.!)
'-' • SILUER
a GOLD
2 ",. COPPER
(\)
CI
0
0 9 18 27 36 45
Ce [MG/U
FIG URE 2: Equilibrium loading of silver, gold and copper cyanide on DU A161.
Equations ( 6) and (7) can be written in the Equations (6) and (7) are based on the assumption
form: that species 1 and 2 both follow the Freundlich
[a s.J
isotherm in a single solute solution. By replacing
1
equation (4) or (5) with equations (6) and (7) and
f(C s ,1) 1 c 1 1-n1 Cs ,1 by duplicating the remaining equations for each
Qs,1 adsorbate, the above mentioned model can be
applied successfully to a two component system
B12CS ,2 (8)
using packed column.
[a s' J
1
A computer programme was developed for the
f(C s ,2) 2c 2 1-n2 numerical solution of these simulation models.
Cs ,2
Qs,2 This programme, written in Turbo Pascal 4, can be
used on an IBM compatible personal computer
B21 Cs ,1 (9)
equipped with a 8087 mathematical coprocessor.
The value of the film transfer coefficient kf can be
With f ( C s , 1 ) and f (C s , 2) plotted against estimated from short-column experiments 7, while
Cs ,2 and Cs , 1 respectively, the competition the surface diffusivity 0 can be estimated by using
coefficients B12 and B21 can be obtained from the the model in a least squares routine.
slopes of the curves.
435
RESULTS AND DISCUSSION volumetric flowrate: 1.4 x 10-6 m3 Is
0.6
0
u ,'I
u"- 'I
0.4 ,'/
,'/ a = 6.4
, '/ a = 12
a = 30
~ '/'
0.2
. ~'h
~..-.:::
0.0 0
15 30 45 60 75
Time [hrJ
0,8 .~
.' ~
, 'h-
, ' /-.
0.6
, ../..
0 /,
u
/,
u
0.4 ,,'
-12 ~
, I o = 2.5 x 10 l1.m /5
" / oo =
= 3.5
'1.5
X to-'I. m"/5
x 10- m2 /5
0.2 .'/
,'/.
./
.."
./:.
0.0
0 15 30 45 60 75
Time [hrJ
FIG URE 4: Sensitivity of the packed column to variations in the surface diffusivity D.
436
0.8 , '"",,,
'%
, , /:-
" /,
0
0.6 .j
u
'\.
u ,'1
0.4 :1 -5
:1 Kf 2.0xIOsm/s
0.2 ,./,. I Kf
Kf
3.2 x 10- 5 m/s
4.0 x 10- m/s
,~
,/:.
, ..~
0.0
0 15 30 45 60 75
Time [hr]
FIGURE 5: Sensitivity of the packed column to variations in the film transfer
coefficient kf
1.0
/.
/.
.
..
/"
0
0.6 /
/, ,
..
u
'\. 7,'
u + HUMIC ACID
0.4 o GOLD
B 0.20
0.2 B 0.72
B 0.95
0.0
0 15 30 45 60 75
TIME [HOURS]
1.0
0.8
0
0.6
u
'\.
u + HUMIC ACID
0.4 0
GOLD
a2 20
0.2 --- a2 50
- a2 70
+
0.0
0 15 30 45 60 75
TIME [HOURS]
FIGURE 7: Sensitivity of the effluent concentration of component 1 (gold) from
the packed column to variations in the isotherm parameter a2 (for humic acid) for
the gold /humic acid system.
437
obtained here) will be used in the next section to Equations 6 and 7 show clearly that, if 8 12 is
interpret the experimental data. relatively low, the effect of a2 on the adsorption
behaviour of component 1 will be less significant
Variations in the size as well as the external The nature of the interaction between different
roughness of the carbon particles can influence the adsorbates during the adsorption process is poorly
effective thickness of the liquid film surrounding understood in the literature, and deserves much
the carbon or resin particle. Variations in the more attention.
resistance to bulk phase transport are depicted in
Figure 5 by varying the film transfer coefficient kf Pre-treatment prior to adsorption
It is evident that the model is more sensitive to
changes in kf during the early adsorption period The resin and carbon in Figures 8 to II were
than during the subsequent slower period of pretreated in solutions of organics, rinsed with
adsorption when intra particle diffusion controls. water, and then contacted with a clear solution of
silver, gold or copper cyanide. The organic
Variations in the competition factor 8 12 and the loadings adsorbed onto the resin or carbon during
equilibrium parameter a2 in the two-component the pre-treatment period varied from 46 mg
adsorption system are shown in Figures 6 and 7 organiclg adsorbent to 520 mg organiclg
respectively. It appears as if variations in a2 have adsorbent UV -spectrophotometry confirmed a
a more pronounced effect on the adsorption of loss of only 5 to 7 % of the organic loading from the
component 1 than do variations in 8 12. Also, it is resin or the carbon during the adsorption of metal
significant that the adsorptivity of component 2 cyanide.
affects the adsorption behaviour of
component 1 (gold) already in the early stages of a Figures 8 to 10 clearly show that low loadings of
column run. Variations in 8 12 in Figure 6 did not organics on either activated carbon or ion-
affect the adsorption of component 2 (humic acid), exchange resins, inhibited the rate of uptake of
and seemed to be the same for the three different gold, silver or copper cyanide, but did not affect
values of 8 12. However, variations in a2 as the equilibrium attainment, and confirms earlier
depicted in Figure 7, revealed an influence in the results by Petersen and Van Deventer 10. It has to
adsorptivity of component 2, which shown a be stressed that the organic loadings recorded on
downward shift as the value of a2 increases (The the carbon or resin used in these experiments
curve for component 2 as indicated in Figure 7, varied between 46 and 51 mg organiclg carbon or
presents only the case where a 2 = 50).
1.0
0.8
0
0.6
u
u"-
0.4
_ model
x untreated carbon
0.2 <>p,-etreated with methanol for 3 hours
pretreated with phenol for 3 hours
o p,-etreated with heptane for 3 hours
• pretreated I~ith PEX for 3 hours
0.00
15 30 45 60 75
Time [hrJ
1.0
0.8
0
0.6
u
'\.
u
0.4
- model
.... untreated carbon
0.2 a pretreated uith ghenol for 3 hours
• pretreated uith pEX for 3 hours
0.00 45 60 75
15 30
Time [hr]
FIGURE 9: Inhibited mass transfer (kinetic influence) of copper cyanide to
activated carbon owing to fouling of organics in a packed column. Co = 20 mg
CulL; 1= 15 cm; v = 1.4 x 10-6 m3 Is; M = 12.32 g.
1.0 A
'"
0.8
0
0.6
u
'\.
u
0.4
0.0
0 15 30 45 60 75
Time [hr]
FIGURE 10: Inhibited mass transfer (kinetic influence) of silver cyanide to DU
A161 owing to fouling of organics in a packed column. Co = 20 mg Ag/L; 1= 15 cm;
v = 1.4 x 10-6 m3 Is; V = 0.075 L.
0.8 A +
0
0.6
u
"-
u
0.4 _ model
<> untreated carbon
v 150 mg PEX/g carbon
0.2 o 290 mg PEX/g carbon
+ 462 mg PEX/g carbon
0.0
0 15 30 45 60 75
Time [hrJ
FIGURE 11: Inhibited mass transfer (kinetic + equilibrium influence) of gold
cyanide to activated carbon owing to fouling of organics in a packed column. Co =
20 mg Au/L; 1= 15 cm; v = 1.4 x 10-6 m3 Is; M = 12.32 g.
22 and 36 mg organic IL resin. The results loadings of organics were recorded on the resin as
obtained from these Figures (8 to 10) indicated a well, the latter did not show any changes in the
constant equilibrium value, so that a mechanism of equilibrium values. At these high loadings of
competitive adsorption between the metal cyanide organics on the carbon, it was impossible to obtain
and organic species, is irrelevant. Furthermore, it acceptable fits for the model by changing only the
appeared that fouling of the resin or carbon had no values of D. It was therefore necessary to use
effect on the value of the film transfer coefficient different values of A and n in the Freundlich
kf' but had a marked effect on the surface isotherm as summarized in Table II. The values of
diffusivity D. This implies a mechanism of pore A and n were determined from equilibrium
blocking, whereby the organic foulants that had experiments on the pre-treated carbon. The
diffused into the pores of the adsorbent, partially mechanism of adsorption onto activated carbon is
blocked the pores and thereby hindered the extremely complex and largely unknown. It is
passage of metal cyanides to such an extent that possible that the competitive adsorption between
diffusion remained possible, but at a reduced rate. organics and metal cyanides is due to competition
This is supported by the kinetic parameters in for active sites, but it is also possible that the
Table II estimated for each of the sets of data A to organics merely block certain regions in the carbon
K in Figures 8 to 10. When a mechanism such as and thereby rendering them inaccessible to metal
pore blocking by organic foulants is operative cyanides. It was impossible to obtain acceptable
during the adsorption of metal cyanides, the fits of the data in Figure 11 by using only different
relationship between the size and shape of the A and n values. The surface diffusivity D also
foulant and the pore distribution of the adsorbent decreased, which means that a further inhibition of
intraparticle transport occurred when the loading
becomes important. Basically, the pore wall
of organic increased.
reduces the freedom of movement of larger
organic molecules, resulting in retarded diffusion.
This is confirmed by the results depicted in Table I Simultaneous adsorption
which indicates that the value of the surface
diffusivity decreased with an increase in the In most studies on the simultaneous adsorption of
molecular mass of the organic foul ant. organics onto activated carbon, the Ideal Adsorbed
Solution (lAS) theory was used to correlate
At higher loadings of organics (more than ]20 mg results 4. In the work presented here, the
organiclg adsorbent) both the rate of adsorption as Freundlich-type multicomponent isotherm was
well as the equilibrium loadings are affected, as used to describe the simultaneous adsorption of
depicted in Figure 1l. Although relatively high gold or silver cyanide and the organic specie onto
activated carbon, simply because the lAS theory
440
80
•
64
•
'":J 48 • • slope = 0.7214
<I:
'V
La.. 32 •
16
9 18 27 36 45
Ce (HUM. ACID) [MG/LJ
FIGURE 12: Estimation of the competition factor B12 for the competitive
adsorption of gold cyanide and humic acid on activated carbon.
80
64
'"
0
......
•
u 48 •
<I: •
i • slope 1. 934
::J
I 32
La..
16
9 18 27 36 45
Ce (Au) [MG/l]
FIGURE 13: Estimation of the competition factor B21 for the competitive
adsorption of gold cyanide and humic acid on activated carbon.
1.0
0.8
0
0.6
u - MOOEL
u"- • HUM[C ACro
0.4 a GOLD
0.2
0.0
a 15 30 45 60 75
TT MF r HnllRSl
FIGURE 14: Rate curves for the simultaneous adsorption of gold cyanide and
humic acid on activated carbon in a packed column. Co(Au) = 20 mg/L; Co(humic
acid) = 70 mg/L; I = 15 cm; v = 1.4 x 10-6 m3 Is; M = 12.32 g.
441
does not hold for partially dissociated solutes. this paper are useful to demonstrate the sensitivity
Figures 12 and 13 present the results of the bi- of column operation to well-defmed variables, but
solute isotherm of gold cyanide and humic acid in cannot be applied conveniently when the available
linear form. From the linear relationship in Figure knowledge on adsorbates is incomplete.
12, the slope of the line, which represents the Moreover, it is frequently necessary to include
competition that humic acid offers to the qualitative information and order-of-magnitude
adsorption of gold cyanide, was determined by estimations in the design of columns.
least squares regression to be 0.7214. The This type of ill-defined problem is not unique to
equilibrium gold loading is then given by: the minerals industry, and has received attention in
the literature. Recently, knowledge-based systems
(KBS) have found increasing application for
treating ill-defined problems, especially in the
chemical industry. Most KBS in the chemical
By using the slope of 1.934 from Figure 13, the
industry have been formulated for fault
equilibrium humic acid loading may be given as:
diagnostics, steady-state circuit design and
simulation, process selection, operator training,
qe ha =5OC e ha(C e ha + 1.93 Ce Au)O.II-1
, " , data management and scenario analysis by means
of "what if?" questions. However, the dynamic
As explained in the theory by Sheindorf et.aI 3, the
simulation of ill-defined processes by use of KBS is
one competition factor should be the reciprocal of
a new concept, and as such has received little
the other in a bi-solute system. The fact that this is
attention in the literature. Reuter and Van
not the case here, shows that these equations are
Deventer 11,13 have explained the methodology
merely convenient empirical expressions.
involved in detail in previous papers.
An expen system is a computer program that
Figure 14 shows the simultaneous adsorption of
behaves like an expert in a usually narrow field of
gold cyanide and humic acid on activated carbon in
application. One of the features of an expert
a packed column. If the results obtained in Figure
system is the separation of knowledge from the
14 are compared to that obtained in Figure 8,
techniques that are used to think about this
which involved both gold cyanide, the surface
knowledge. In view of the fact that expert systems
diffusivity D is the only parameter that did not
require expert knowledge in some form, they are
remain constant The decrease in the intrinsic
also called KB~ An expert system should be able
diffusivity can be explained through the fact that
to explain its behaviour and decisions to the user.
two mechanisms were operative at the same time
In addition, a KBS should be capable of dealing
namely, the competitive adsorption for active sites
with uncertainty and incompleteness of
on the carbon, as well as that of pore blocking.
information. An expert system consists usually of
a knowledge base, an inference engine and a user
APPLICA TION OF KNOWLEDGE-BASED interface. Although different architectures are
SYSTEMS used to define a KBS, most KBS use both
qualitative (experiential and heuristic) as well as
Nature ofKBS systems
quantitative (equation orientated) knowledge, also
In most industrial columns used for the removal of
termed shallow and deep knowledge respectively.
metals and organics from plant effluent, complex
multicomponent solutions are encountered. Treatment of adsorption data
Existing phenomenological models are usually
incapable of considering the effects of pH, ionic The basis of the adsorption model is batch
strength, carbon deactivation, and all the complex concentration-time data, and the corresponding
interactions between the many species in solution. loading-time curve that covers the adsorption
The more conventional fundamental models profile up to the equilibrium loading. From such a
discussed and applied in the previous sections of batch adsorption curve, the "rate variable"
442
k [ q (t) ], which is defined here to be a function entering / editing of all relevant experimental and
of only the loading q ( t ), may be determined. heuristic information; (2) An inference procedure
This "rate variable" is subsequently used to predict for system diagnostics and the simulation of
the change in concentration and carbon loading at different reactor and column systems; (3) A
any position in the column at any moment in time working memory which contains the current input
at the prevailing chemical process conditions. and status of the specific adsorption column or
It must be noted here that such KBS models do other reactor being investigated. The different
not attempt to suggest a mechanism for the modules of the system communicate via files in
reaction under consideration. What they do imply, order to increase the working memory; (4) A user-
however, is that a kinetic curve which describes a friendly menu-driven interface used to
process up to equilibrium is a finger-print of that communicate with the above three components.
reaction and therefore of the reaction mechanism. As expected, the KBS is heavily dependent on the
Consequently, the KBS model can be used as a range, accuracy and population density of the
basis to describe the reaction under all possible available data.
conditions, given that the mechanism does not
change. The database will obviously cover all CONCLUSIONS
possible mechanisms for a particular system. This
basic finger-print curve and its associated shallow Both activated carbon and ion-exchange resins
and deep level knowledge, defined as objects, are used in adsorption columns for the removal of
termed the pivot-data, or the standard condition. metals from plant effluent are fouled by organics
These pivot-data serve as a reference with which usually present in industrial effluent streams. At
other curves are compared, and which can low loadings of organics on either the resin or
subsequently be used for fault-diagnosis or process carbon the adsorption of metal cyanides is
identification. influenced in a kinetic way, and can be ascribed to
Any condition that differs in whatever way from pore blocking. At higher loadings of organics, both
this standard condition or the pivot-data, is a kinetic as well as an equilibrium influence were
considered to be non-standard. Consequently, any observed. This can be attributed to two effects,
change in the empirical parameters or the profiles viz.: (1) permanent pore blocking, where the pores
of loading should be considered as the combined
of the adsorbent are totally sealled off by the
effect of different deviations from the standard organic species, and (2) competitive adsorption for
condition, and therefore as taking all interactions active sites between the metal cyanide and organic
into account Such changes in the adsorption foulant It has also been illustrated that KBS
behaviour may be caused by changes in the modelling is a valuable tool in effective design of
chemistry of the feed solution, deactivation of the multi-component adsorption columns.
adsorbent, or changes in the nature of competition
between the known and unknown adsorbates. It is NOMENCLATURE
convenient to express deviations from the pivot-
data or standard parameters as percentages, which a parameter in Freundlich isotherm
are estimated through experience, directly from A flow area ofthe column [m2]
the plant or from experimental data. This KBS
B equilibrium competition factor
approach of modelling developed at the University
of Stellenbosch has been applied successfully to C solution phase concentration [g/m 3]
the simulation of small-scale columns under D surface diffusion coefficient [m 2 Is]
various modes of operation, including that of
continuous countercurrent flow.
dp particle diameter [m]
2. Fritz, w. and Schliinder, E. U. "Competitive 10. Petersen, F.W. and Van Deventer,1.S.1.,
adsorption of two dissolved organics onto activated "The inhibition of mass transfer to porous
carbon-I. Adsorption Equilibria., Chemical adsorbents by fIne particles and organics",
Engineering Science, vol. 36, 1981, pp. 721-730. Chemical Engineering Communications, in press,
1991.
3. Sheindorf, C.H., Rebhum, M. and
Sheintuch, M., "A Freundlich-type multicomponent 11. Reuter, M.A. and Van Deventer, 1.S.1., "A
isotherm", Journal of Colloid and Interface knowledge-based system for the simulation of
Science, vol. 79,no. 1, 1981,pp.136-142. batch and continuous carbon-in-pulp systems",
Extraction Metallurgy '89, Institution of Mining &
4. Van Deventer, 1. S. 1. , "Competitive Metallurgy, London, 1989, pp. 419-442.
equilibrium adsorption of metal cyanides on
activated carbon", Separation Science and
12. Van der Merwe, I.W., Van Deventer, 1.S.1.
Technology, vol. 21, no. 10,1986, pp. 1025-1037.
and Reuter, M.A., "Knowledge-based computer
simulation of gold leaching in batch and
5. Petersen, F.w. and Van Deventer,1.S.1., continuous systems", 14th Congress of the Council
"The influence of organics on the determination of of Mining and Metallurgical Institutions,
gold, silver and copper by atomic absorption Institution of Mining & Metallurgy, London, 1990,
spectrophotometry", Minerals Engineering, vol. 3, pp. 147-160.
no. 5, 1990, pp.415-420.
13. Reuter, M.A. and Van Deventer, 1.S.1., "A
6. Van Deventer,1.S.1., "Kinetic models for knowledge-based system for the simulation of
the adsorption of gold onto activated carbon", batch and continuous carbon-in-Ieach systems",
MINTEK-50 Int. Con! Mineral Science & Proceedings of APCOM '90, TUB-
Technology, MINTEK, Randburg, South Dokumentation, Technical University of Berlin,
Africa, L.F. Haughton (ed), vol. 2, 1985, pp. 487- vol. 51, no. 1, 1990, pp. 343-356.
494.
0
LJ
447
448
CHARACTERISTICS OF MODERN CELL smelters for AIF 3 , has made it possible to add
TECHNOLOGY alumina at known amounts and thereby
Old potlines with fully paid-back investment costs and monitor alumina and bath composition more
possibly also linked to a hydro-electric power station closely. Point feeders also require specific
built long ago may be cost effective. Nevertheless properties of the alumina which is different from
such plants are steadily phased out due to increased previous experience with side- and
direct production cost, but especially due to stricter centerbreak cells.
requirements to the aluminum industry's effect on the
inner and outer environment. It is also found that 6. Process control. The point feeder and
some plants with relatively simple technology show close monitoring of alumina concentration is
good technical results close to the more advanced the basis for the modern cell control. Presently,
designs, due to workers' dedication and good the only continuously measured control
management. But a steady modernization is occurring parameters are the cell voltage and potline
in the aluminum industry either in form of more or less current which also enable calculation of the cell
deep-going retrofit programs or by building new resistance. This resistance depends on the
plants. alumina content in the cell, and is the basis of
the control system which contains rather
1. Size. Small cells are usually more labour sophisticated logics with additional inputs and
intensive and less energy efficient. Cells with safeguards. It is possible to control the alumina
amperage less than 150 kA are hardly concentration, keep a nearly constant anode-
considered modern to-day. cathode distance and also to spot irregular cell
behaviour at an early stage which in turn
2. Magnetic Compensation. A direct stabilizes bath composition and operational
consequence of the higher amperage is the temperature. The improved alumina
need for magnetic compensation to reduce concentration control will also reduce the
metal and bath velocity and counteract the anode effect frequency, usually to about 0.1 a
tendency to a non-horizontal metal surface. day.
New cells are usually arranged side by side
with distributed quarter rizers for more efficient
7. Mathematical Models. Mathematical
compensation of the magnetic fields.
models based on knowledge of physico-
chemical properties of the materials and the
3. Off-gas Cleaning. Dry-scrubbing systems
processes coupled with accurate experimental
are applied thus reducing the emission to less
data on cell behaviour are the corner-stones of
than 0.5 kg fluorides per ton aluminum
modern cell technology. The models will be the
produced. The off-gas fluorides and
basis for designs, choice of materials, effect of
hydrocarbon vapors are absorbed on the
mechanical forces, process control, and will
alumina constituting the ore feedstock to the
describe voltage drops, magnetics, fluid
cells.
dynamics, heat balances, time dependent
4. Inner Environment. Modern plants changes, and disturbances, evaporation and
emphasize the protection of workers from PAH transport of impurities.
and fluoride vapors and dust. Sometimes it is
accomplished by personal protection 8. Bath Chemistry. The bath composition
equipment, but the long-time solution is an has been optimized with respect to current
improved pot-room atmosphere determined by efficiency and energy consumption and to-
cell-emissions, hooding efficiency and suction day's electrolytes contain 7 - 13 % excess AIF3
capacity. added to cryolite. In some cells LiF and MgF2
are added. With point feeding of alumina a
5. Point feeders. Introduction of point smaller difference between the liquidus and
feeders not only for AI 20 3 but also by some solidus of the electrolyte can be allowed.
449
9. Cathode Materials. There is a tendency creates new options for improved fume collection
to use more graphitic cathode block materials through reduced crust-breaking and design
to reduce expansion due to sodium modifications for the gas skirts. To overcome the main
intercalation and reduce cathode voltage drop. S0derberg drawbacks it is, however, also necessary
Redesign of collector bars contributes also to to have a major reduction of the anode fumes and get
reduced cathode drop. The thickness of a stable and improved anode quality with less dusting
refractory and insulation materials are tailored and anode problems. The ability to do so is likely to
to each design, and it is often used three kinds be the only way to a viable S0derberg technology for
of materials to obtain the optimum positions of the future.
the isotherms.
CURRENT PLANT OPERATION
10. Steel Shell. It is found that a strong and Competitive advantages are also created by the ways
elastic steel shell prolongs life for the high the potlines are operated. During the last decade, the
amperage cells. dominating aluminum reduction technology licensee
has been Aluminium Pechiney with its 175/180 kA
11. Handling of Materials. An aluminum cells. Even though the companies that have acquired
plant may have the appearance of a busy AP technology have received the same specifications
storage building with trucks, cars, and conveyor and instructions, there are Significant differences in
belts going in all directions. Shortened flow operational results. 5 For the period 1982 - 1988
lines of raw-materials, products and production differences in obtained current efficiencies of more
equipment are emphasized in modern plants. than two percent have been observed. The
Alumina is transported automatically directly to corresponding differences in energy consumptions
the different cells. are about 0,3 kWh/kg AI. Generally, potlines with the
highest current efficiency also have obtained the
12. Pot-Tending Machines. Pot-tending lowest energy consumption. Notably, one potline has
machines are mUltipurpose and substitute the both achieved the highest current efficiency (>95 %)
work that previously was carried out by and the lowest energy consumption (12,7 kWh/kg AI).
specialized trucks or by operators. For a potline of 120.000 tons/year, two percent
current efficiency amounts to 2.400 tons of aluminum,
The self-baking continuous S0derberg anode the related production cost being that of alumina,
is currently not considered to be modern technology. solely. Conversely 0,3 kWh/kg AI amounts to 37 GWh
The inherent advantages of the continuous anode of electrical energy saved annually. The operational
have been more than offset by: performance data illustrates a general trend of
- Less efficient fume collection learning by experience in that practically all potlines
- Higher anode voltage drops have improved their current efficiency by one percent
- Stud pulling is a difficult and often an during the time of operation. 5
unpleasant work operation As pOinted out in two previous papers 6 ,7 when
- Point feeders cannot be placed in the the amperage of the reduction cells increases, an
middle of the cell increase of productivity is obtained by a reduction in
- A lower practical limit to the cell size man hours/ton AI. Likewise, as the total production of a
- Inferior anode quality compared to prebake smelter complex increases, an increase in overall
due to lower baking temperature productivity should be expected.
- More prone to dusting. Figures 2 and 3 illustrate the productivity of a
random sample of smelters measured over a wide
There are, however, efforts going on to range against cell amperage (cell size) and overall
modernize the existing S0derberg technology. tonnage (smelter size). At first glimpse, the
Several companies have installed pOint feeders on a illustrations show no correlation between productivity
limited number of cells. Point feeders are a major step and size, neither cell nor smelter size. The extremes
forward in order to achieve improved feed control and in productivity are found for small and large size,
higher current efficiencies. This way of feeding also respectively, but in-between a large variety is
450
> 0 0
0 0
I-
LJ Degree of subcontracted services.
=> 00 0 0
= - Differences due to plant management
D
0:::
100 0 0 <6 Q:>
CL attitudes.
Figure 2: Productivity (tons per worker-year) vs. cell several of these factors. The need for high productivity
amperage (kA) for a random sample of is especially important in high cost countries.
2: 200 1-0 0
00 smelter complexity and size, including cell amperage,
I- Q)
L.)
=> 0 is the fundamental effect of management attitudes. In
o 0
o
fr:: o 0 0 this respect tendencies towards relative positions of
u.. 100 l- -
I I I I I I high and low productivity is strongly influenced by
50 100 150 200 250 300 management capability.
SMELTER CAPACITY (klon)
DIFFERENCES IN OPERATIONAL PHILOSOPHY
Figure 3: Productivity (tons per worker-year) vs. The major mission of all businesses is to comply with
smelter size (annual aluminum production) the demands of customers and society, justifying the
for a random sample of smelters. use of resources in the long term which also means
being profitable.
For larger smelters (>200,000 t/yr.), productivity The location of aluminum smelters during the
ranges from 170 to 300 tons per worker-year, while for last century has been very strongly connected to the
small smelters (50,000-100,000 tlyr.), the productivity location of readily available (abundant) energy
span is from 120 to 220 tons per worker-year. For sources, and where the energy production potential
newly built smelters with a capacity of more than and capacity has been greater than the local
200,000 t/yr., a productivity of 250-300 tons per communities could absorb themselves. Also, the
worker-year is typical. For very large, single- availability and cost of capital and labour have
technology smelters a productivity in the range 400 influenced both location and operational philosophy.
tons per worker-year may be reached. It appears that The aluminum smelting industry is both energy and
potlines with cell amperages more than 175 kA capital intensive.
combined with large scale are required to achieve These factors have been determinant for the
productivity of more than 250 tons per worker-year. historical difference between the operational
There may be some apparent reasons that philosophy of the North-American and the European
explain the observed differences: aluminum smelting industries. Capital and labour
451
have been traditionally expensive in the USA, while and energy consumptions are significantly better for
energy so far has been readily available. This has technologies I and II due to this being the latest
favoured a tendency towards high current densities released technologies where the impact of computer
and high current efficiencies in the USA at the modelling and up-to-date process control have
expense of high energy consumption. In continental brought the state of the art forward.
Europe, with high population density, electrical For normally well operated end-to-end prebake
energy have been an in-demand resource as well as cells with modified magnetics and thermal balance
not very abundant, which in turn has generated a including up-dated process control, current
tendency towards lower energy consumption in efficiencies of more than 93 % and energy
European aluminum smelters. Typically, the consumption down below 14 kWh/kg AI can be
difference in energy consumption between the two achieved.
continents has amounted to 1 - 2 kWh/kg AI. Carbon consumption numbers are clearly
Table 1 summarizes operational data (from different between point-fed and centerbreak cells as
1989 and 1990) from eight different types of prebake seen between the groups including Types I and II,
reduction technologies. Types II and IV data are an Type III, and Type IV, respectively. The difference is
average for several potlines with similar technology. most clearly displayed in the frequency of anode
Types I and II are pointfed cells, Type IV have effects.
centerbreakers, while data for Type III both represent The reduction technology represented by the
pointfed and centerbreak cells. Types III and IV are operational data in column I, illustrates an extreme in
more than fifteen years old. several aspects of aluminum reduction technology. It
Bath chemistry and average operational is an old low- amperage technology that has been
temperatures seem to be fairly similar for the boosted to a current density of 1.33 A/cm 2 by
technologies I - III. Also the current densities for these employing low melting lithium-magnesium containing
technologies are not too different. Current efficiencies bath and SiC as cathode side lining to dissipate
excess heat. Current efficiency and energy area, uses less materials and may have a more
consumption are in line with the very best operated stable side ledge. On the other hand the
S"derberg potlines, but the technology has to pay a refractory layer is thinner with greater chances
high carbon consumption due to centerbreakers and for complete penetration of the bath. This
high current density. This technology is probably an design may also prevent the use of graphitic
example of an old potline where for a limited period of cathode blocks due to their higher thermal
time, there will be available inexpensive electrical conductivity.
energy locally, and the metal output is maximized.
The technology also illustrates the robustness and Insulation. The insulation design may show
flexibility of old cells when using up-dated knowledge large variations illustrated by the following
to improve cell characteristics. It would be unthinkable ranges in heat loss:
with modern technologies to do such drastic changes Top: 35 -70 %
to the original cell design. Side: 20 - 45 %
The cells represented by column II - VIII are all Bottom: 5 - 10 %
high amperage and magnetically compensated, some This is a consequence of the initial design and
built a few years ago, some up to 28 years old. Some the need to keep a proper heat balance.
common features are apparent. The current density Changes may also be related to the wish to
varies within a narrow range being between 0.77 and boost the amperage when the aluminum price
0.83 Alcm 2. Current efficiency is above 92 %, the is high. The heat loss from the top is the only
highest being 94.6 %. The content of Si + Fe is easily adjustable parameter by varying the ore
between 0.15 and 0.20 wt %. The AIF3 consumption is cover.
within the narrow range 18.8 - 22 kg/t AI. The bath
chemistry is similar, CaF2 being between 4.9 and 6.6 Automation and Mechanization. Although
wt % and excess AIF3 between 10.2 and 12.0 wt %. all smelters see the benefit of point-feeders, the
This is reflected in a narrow operational temperature complexity of pot-tending machines varies
range between 951 and 963 °C. greatly. The cost aspect, skill and fine-tuning of
The energy consumption varies between 13.2 present operation are important factors in
and 15.1, a 14 % difference. It is typical that the oldest deciding the degree of automation. It was for
designs II, III and VII have the highest energy instance at some time built pots with
consumption. Again it is demonstrated that high individually adjustable anodes. Presently it is
current efficiency often means high energy found that satisfactory results are obtained
consumption. The difference between centerbreak without this extra option, and individual anode
and pOint feeders is illustrated by cell-type II. regulation may even disturb the interpretation
Centerbreak gives a considerable higher carbon of cell resistance measurements.
consumption and a higher frequency of anode effects.
Preheating. The most common preheating
For the other cells the carbon consumption is not
methods are resistive heating with a coke bed
clearly correlated to cell parameters, and the 5 %
and gas burner heating. Coke bed preheating
difference in net consumption probably reflects
has the advantage that no special equipment is
differences in raw materials, efficiency of the carbon
needed as the line current provides the power
plant and operational procedures.
source. The success of the method is, however,
In many areas there is general consensus how
dependent on an even current distribution to
to design and operate a successful modern aluminum
avoid local overheating. This problem is
cell. There are, however, some notable differences in
reduced by some smelters with the use of
opinion:
shunting avoiding full line current at the start,
Shell Design. It i~ generally agreed that but at the expense of a longer heat-up period,
high-amperage cells' need strong pot-shell, but which in turn is favorable for minimum thermal
opinions vary whether the shell should be deep stresses. Gas heating gives a more uniform
or shallow. A shallow pot-shell with a shallow heat-up rate and improved temperature
distribution in the cathode introducing minimum
metal pad (10 cm) has a smaller heat-loss
453
developed, i.e. an anode that may have a life of "History of Electrical Energy Consumption by Hall-
months instead of weeks. This will then eliminate the Heroult Cells", The Metallurgical Society, Warrendale,
release of C02. On the other hand the reversible cell PA, (1986), pp. 106 -113.
voltage will increase one volt due to the lack of 2. Richards, N.E. in Hall-Herault Centennial. 1886-
carbons depolarizing effect. This voltage may, ~. "Evolution of Electrolytes for Hall-Herault Cells",
however, be partly or fully offset by reduced ohmic The Metallurgical Society, Warrendale, PA, (1986),
resistance and anode overpotential. pp. 114-119.
3. Welch, B.J. in Hall-Herault Centennial. 1886-1986.
"Gaining That Extra 2 Percent Current Efficiency", The
paper: Harald A. 0ye and Reidar Huglen: "Managing 4. Belitskus, D. in Hall-Herault Centennial, 1886-
Aluminium Reduction Technology. Extracting the Most .1.9.aa.. "Carbon Electrodes in Hall-Heroult Cell: A
from Hall-Heroult", which appeared in JOM 42, Nov. Century of Progress", The Metallurgical Society,
Warrendale, PA, (1986), pp. 130-143.
1990, p. 23-28.
5. Langon, D. and Peyneau, J.M. "Current Efficiency
in Modern Point Feeding Industrial Potlines",.l..ig,b..1
Metals, (1990), pp. 267-274.
6. Welch, B.J. "Aluminium Reduction Technology
Entering the Second Century", J. Metals, 40, (1988)
No. 11, pp. 19-25.
7. Grjotheim, K. and Welch, B.J. "Aluminium Smelting
as the Industry Enters the 21 st Century, J. Metals, 41,
(1989) No. 11, pp. 12-16.
8. S"rlie, M. and 0ye, H.A. Cathodes in Aluminium
ElectrolysiS, Chap.II, Aluminium-Verlag, Dusseldorf
(1989).
New approach to acid methods for
special alumina production
Rumen D. Kanev
Kurillo Metal SPE, Polymet Co., Sofia, Bulgaria
phenomena;
- high degree of ammonium aluminium sulfate salt (4)
formation;
- high degree of ammonium aluminium sulfate salt - Basic ammDnium aluminium carbDnate salt thermal
purification by means of counter-current precrystalization decomposition and calcinatiDn:
with complex-forming agents; 400 DC t
- high degree of basic ammonium aluminium carbonate NH4AI2C03(OHh ----------- am-Al20 3 + NH3t+ C02 +
formation by means of neutralization of ammonium 11500 C
aluminium sulfate into the ammonium bicarbonate + H 20r -------------- alpha-ALUMINA (5)
solution;
- low temperature of ammourphous alumina formation by The leaching Df aluminium wastes with sulfuric acid is a
means of basic ammonium aluminium carbonate thermal very well knDwn prDcess fDr producing aluminium
decomposition with full utilizatiDn of the separated sulfate.Usually, the wastes are treated with sulfuric acid
gaseous phase, and diluted to. SDme chDsen extend and metallic aluminium
- direct ammDurphDuS alumina calcinatiDn to. high-purity reacts according to. the abDve shown reactiDn (1). This
alpha-alumina at IDwer temperature. reactiDn is eXDthermic and proceeds with cDmparatively
In the present paper some theDretical aspects Df this high reactiDn speed and heat and hydrogene eVDlvatiDn at
method, concerning all productiDnal steps and mainly the the beginning. HDwever, as the reaction proceeds and the
new approach to. the basic ammDnium aluminium amount of free acid in the system decreases, the reaction
carbDnate salt formatiDn and its decDmpDsitiDn to a high speed and heat eVDlvatiDn decrease tDO. TherefDre , to.
purity alumina are cDncidered, and the results Df the pilDt achieve a mDderate final reactiDn speed additional external
plant investigatiDn fDr determinatiDn Df Dptimal process heating is used. FrDm the Dther hand, into. the aluminium
parameters with mDnitDring the impurities CDntent into. the wastes coexist as impurities some metallic and nDn-metallic
semi-prDducts and final product are given. inclusiDns. Due to. their cDmparatively high chemical
affinity towards sulfuric acid they also. react and pass
quantitatively into. solutiDn and cDexist in the final prDduct
EXPERIMENTAL as corresponding metal sulfates.
These basic problems Df that well knDwn process are
Some theoretical aspects overcomed to. a great extend in this productiDnal step by
The main chemical interactiDns which take place in this percDlatiDn leaching Df the wastes using a specially
methDd can be summarized as fDllDWS: designed percolatiDn cDlumn and creating cDnditions fDr
impurities separatiDn by attaining maximum pDssible
- Aluminium wastes acid leaching: effect frDm cementatiDn and reductiDn phenDmena [9].
The direct mixing Df liquid aluminium sulfate and
crystaline ammonium sulfate is also a well knDwn process
fDr producing ammDnium aluminium sulfate. Usually this
- Aluminium and AmmDnium sulfates mixing: process takes place within the temperature range from 65
to 950 C and after a cDmplete dissDlving Df crystaline phase
the solution is cDoled and Dbtained according to reaction
(2) crystals of ammonium aluminium sulfate are separated
(2) by filtratiDn, leaving into. the liquid phase the surplus Df
ammDnium sulfate. This process has the fDllDwing basic
- Ammonium aluminium sulfate purification: problems: a complete aluminium transfDrmatiDn from
aluminium sulfate to the ammDnium aluminium sulfate,
and impurities separatiDn frDm the crystal semi-product. In
the newly develDped methDd the abDve mantioned
problems of this prDductiDnal step are DverCDmed by
using into the stage Df mixing around 120 - 130 pct
- AmmDnium aluminium sulfate neutralization: stoichiometric surplus of ammDnium sulfate and so.,
creating conditiDns fDr cDmplete aluminium utilizatiDn,
and fDr impurities separatiDn, a step-by-step
(NH4)2A12(S04)4.24H20 + 8NH4HC03 + nH20 = precrystalization of the crystals is used. For attaining a
biger purificatiDn effect however, in any precrystalization
= NH4AI2C03(OHh + 4(NH4)2S04 + (n+24)H20 +
step cDmplex-fDrming agents are added. (see reactiDn 3 ).
459
The amount of these reagents is calculated upon the base ions in the resulting solution has to be at least 125 pct of
of the impurities content as 120 - 150 pct stoichiometric that nesessary for basic ammonium aluminium carbonate
surplus. In this manner for reaching the requisite level of salt formation. Second, into the prepared solution, as it
impurities content under 10 ppm in the finally purified was described above, the crystaline ammonium aluminium
crystals only three stages of counter-current sulfate is added with preliminary determined rate, but not
precrystalization are nessesary ( against six or seven more than 0,3 kg/dm3 .min-I. After finishing the addition
without usage of complex-forming agents ). In the first two of the crystals the resulting suspension is leaving with
stages the salicilic acid prove to be the best complex- stirring for agening 15-30 min. Third, after finishing the
formiming agent, and for the last one, EDTA II agening the crystals of basic ammonium aluminium
( ethylene-diamine tetraacetic acid ) can be used carbonate are separated from the suspention by means of
successfully. In many methods developed for producing simple settling and suction of the overflow, which actually
high-purity alumina the purified ammonium aluminium contains very pure ammonium sulfate and can be used for
sulfate is the final semi-product from which the high-purity washing of the obtained in the previous productional step
alumina can be produced. From the technological and ammonium aluminium sulfate crystals. The separated
technical point of view however, its thermal decomposition crystals of basic ammonium aluminium carbonate
is a very difficult process connected with handling and undergoes three steps of counter-current washing and the
utilization of the separated within the temperature range outstream is used for dissolving of ammonium bicarbonate
from 100 to nearly 750 0C highly corrosive gaseous phase at the beginning of this productional step. The final
with low sulphur oxides and ammonia content. crystals has 36-37 pct alumina content and can be used
In view of the above, in the newly developed method for directly into the following productional step.
production of high-purity alumina an additional The final productional step of this method is the thermal
productional step is involved. That is the neutralization of decomposition and calcination of the obtained basic
the purified ammonium aluminium sulfate with ammonium alluminium carbonate. Usually, it can be
ammonium bicarbonate in the presence of a small amount accomplished in the static conditions using a corundum
ammonium sulfate, according to reaction (4). This crusibles within two temperature ranges from 100 to
reaction however, is highly endothermic and has to be 5500C and 550 to 11500C. Within the first temperature
carried out very carefully into the frameworks of the range the decomposition of that salt with separation of
previously determined conditions, because, there exist ammonia, carbon dioxide and water vapour takes place.
some possibilities as a final product, instead of basic The very important fact here is, that the time for
ammonium aluminium carbonate salt, the aluminium decomosition has to be sufficient for complete separation
hydroxide to be obtained. Because of its colloidal nature of the volative components from the salt, otherwise there
and pure sedimentation and filtration behaviours this exists a possibility for the secondary permutation among
semi-product is highly undesirable for further processing. the reaction product ( see reaction 5 ) with ammonium
By contrast with aluminium hydroxide, the basic aluminate formation - NH4 AI02 • This semi-product has a
ammonium aluminium carbonate salts has a very well- higher point of thermal decomposition and actually can
formed small crystals which settle down in the suspension affect the intermediate products of calcination. Within the
very quickly and possesses highly desirable sedimentation second temperature range only polymorphous
and filtration behaviours. From the other hand, its transformation of alumina takes place from
thermal decomposition takes place within the temperature ammourphous phase through gamma-modification to the
range from 90 to 265 0C (according to DTA-curves) with final alpha-modification. The product obtained possesses
separation of a less corrosive gaseous pha.,e containing all chemical and physical properties of high-purity
ammonia, carbon dioxide and water vapour. The full alumina.
utilization of that off gases can be accomplished in a very In fig. 1 the flowsheet of the method is given.
sophisticated way by means of absorbtion and the solution
obtained can be used back into the process for dissolving Pilot plant investigation
new amounts of ammonium bicarbonate. The investigation on development of this method in a
The technological manner developed for successful close-circuit productional steps have been carried out in a
neutralization of ammonium aluminium sulfate with specially designed and constructed pilot plant unit with
ammonium bicarbonate has a few stages [10]. First, into capacity of 4 tpy. The main equipment and apparatus
the solution, made by mixing of wash waters from the assemblied in the different productional sections of this
washing of final semi-product and deionizide water, and plant were as follows:
absorbtion water, containing certain amounts of In the section for percolation leaching two percolation
ammonium ions, the ammonium bicarbonate is dissolved columns ( dia 300 mm and height 8 m ), assemblied with
by means of permanent stirring within the temperature glass cylindrical sections, heaters with internal serpantine
range from 35 to 600C. The total amount of ammonium worm-pipes, lids and valves and two PVOF percolation
460
~[
....;..AL.;;...w...;ut;.;.;.•.;..;•....;..... ..._ _ _ _ _ _ _ _....._ _ _ _ _ _ _ _ ~,) .. Sulfuric Acid
Salldflc Add
Mother Uqulor
Washings
Moth... Uqulor
Waahlnge EDTA II
Mother Uqulor
Ammon.Blcarb. Ammon.Blcarb.
Crystal
Delon.Water
Off ~. . .
AbaorbUoin
pumps ( MDR 105V2 JOHNSON PUMN,Sweden ) were In the pilot plant investigations the following raw material
fitted. There were also a horizontal vacuum-filter and four and chemical reagents were used:
glass made vessels with total capacity of 0.8 mS for filtration - technical grade sulfuric acid ( 92 pct );
and transitional storaging of aluminium sulfate obtained. - deionized water;
The total productional rate of one run was 0.4 mS per - technical grade ammonium sulfate;
column. - chemicaly pure ammonium bicarbonate;
In the section for ammonium aluminum sulfate formation - complex-forming agent Salicilic Acid and EDTA II;
a cylindrical glass vessel with capacity of 0.15 mS for - flocculants TIOFLOC B31; MAGNAFWC 342 and
aluminium sulfate dosing, internaly epoxinated stainless DRIMAX ( produced by ALLIED COLLOIDS
steel reactor with steam-jacket for mixing of aluminium LTD,Bradford,England)
and ammonium sulfate ( 0.2 m3 ) , two heated with steam- As a main raw material the aluminium containing cuttings
jacket horizontal vacuum-filters, two internaly epoxinated (Si - 10.11 pct; Fe - 0.18 pct; Mg - 0.31 pct, Cu - 0.01 pct; Pb
stainless stell reactors for crystalization of the filtered hot - 0.01 pct, Zn - 0.05 pct; Mn - 0.01 pct and Ti - 0.001 pct )
solution ( 0.4 mS each) and finaly one glass made horizon- were used.
tal vacuum-filter for filtration of the suspension containing All chemical analysis for the impurities content into the
ammonium aluminium sulfate crystals and mother liquior semi-products, mother liquiors, washing waters and final
were fitted. product were made using Perkin Elmer AAS PC 5100.
In the section for purification of ammonium aluminium The pilot plant investigation have had as a main object
sulfate salt three sets of apparatus for the first, second and determination of the impurities separation and their actual
third precrystalization were fitted.In any set there were destribution within the different productional steps at the
glass made reactor with internal serpantine worm-pipe preliminary determined process parameters. Therefore, the
heating ( 0.1 mS ) for dissolution of the crystals; two glass results are commented only for the particular impurity
made reactors with external cooling-jacket ( 0.05 mS level and its degree of separation in any productional step.
each) for crystalization of the hot solution and one glass
made horizontal vacuum-filter for filtration of the
suspension containing purified crystals and mother liquior. RESULTS AND DISCUSSIONS
In the section for neutralization of the purified ammonium
aluminium crystals there were four glass made reactors Aluminium wastes acid leaching
with external steam-jacket (0.05 mS ) for neutralization In this productional step the following optimal process
and six glass made reactors with external steam-jacket parameters were used:
( 0.05 mS ) for washing of basic ammonium aluminium salt - leaching temperature ( max ) - 85 0C ;
obtained. The suspention was filtered on three glass made - leacing time ( total) - 260 min;
vacuum-filters. - acid concentration - inlet 220 g/dm3 ; outlet 7 g/dms;
All equipment and apparatus of the above mantioned - alumina concentration in the semi-product -73.9 g/dm 3.
sections were connected in accordance with technological In table 1 the results for impurities content at the inlet and
flowsheet by glass and PVC made pipes, tubes and pumps. outlet of the process and degree of their separation in
The main equipment in the last technological section for connection with created conditions for cementation and
thermal decomposition and calcination of the basic reduction phenomena are given.
ammonium aluminium carbonate salt was an electrical The results shown in table I prove the statement that the
furnace NABERTHERM WlOOO with moving loading-car percolation leaching of aluminium wastes with sulfuric acid
(1360 0C -1 m3 ). This equipment was fitted with externaly within the limits of the optimal process parametes is a very
water-cooled pipe-line and glass made ( dia 300 mm and effective way for impurities separation by means of
height 5 m ) absorbtion column for off gases utilization. cementation and reduction phenomena, especially for most
For storaging and dosing of different mother liquiors and of the heavy metals cations. The degree of K and Na ions
washing waters twenty four glass made vassels ( 0.15 m3 separation however is extremely low in this process
and 0,2 m3 ) were used, and for deionized water conditions and that have to explained with their higher
production there were fitted five glass made apparatus acidic actifity than that of aluminium.
with total capacity of 0.24 mS/h ( PCL-60 and PCL-
30,KAVALIER,Chehoslovakia).In the unit there were also Aluminium and ammonium sulfates mixing
a sulfuric acid storage with dosing system, vassels for In this productional step the following process parameters
deionized water with dosing system and different were used:
ventilation systems and aspirations. - mixing temperature - max - 850C ;
All glass made vessels, apparatus, reactors, pumps and - mixing time - max - 75 min;
pipes were disigned as a standart equipment by - complex-forming agent - salicilis acid 125 pct to the
KAVALIER and GLASSEXPORT from Chehoslovakia. impurity content;
462
Table I. Impurities content and degree of separation Table II. Impurities content and degree of separation
- Si - 4808 96.16 98
- Fe - 87 26.10 70 - Si - 173,16 32,90 81
- Ca - 0.0381 0.008 79 - Fe- 47,05 9,41 80
- Mg- 148.88 13.40 91 - Ca- 0,1351 0,05 63
- K - 0.0053 0.004 24 - Mg- 24,15 4,83 80
- Na - 0.0052 0.004 26 -K - 0.0051 0,004 21
- Cu - 0.0125 0.001 92 - Na- 0,0051 0,004 20
- Pb - 0.4667 0.014 97 - Cu- 0,16 0,04 75
- Zn - 24.75 0.99 96 - Pb- 0,163 0,05 69
- Cd - 0.02 0.001 96 - Zn- 4,5 0,99 78
- Mn - 0.02 0.001 95 - Cd- 0,1 0,01 90
- Ni - 0.0238 0.005 79 - Mn- 0,417 0,1 76
- Co- 0.0318 0.007 78 - Ni - 0,192 0,05 74
- Cr - 0.0028 0.001 65 - Co- 0,0921 0,07 14
Total: 5069.2562 136.35 96 - Cr - 0,3334 0,07 79
------------------------------------------------------------------------_ .. _---- - Ti - 0,2857 0,1 65
Total 250,7485 48,678 80,6
- rate of crystalization - 1,45 to 1,55 oC/min; -------------------------------------------------------------------------------
components separation and for that purpose EDTA II has - basic ammonium aluminium carbonate salt obtained:
to be used. avarage moisture - 56,3 pct;
The reached after third step of precrystalization impurities avarage alumina content - 37,6 pct;
level in the purified ammonium aluminium sulfate salt of avarage ammonium ions content -14,9 pct.
less than 5 ppm is more than it is required to further avarage particle size ( D50 ) - 1,0 micron.
processing of this semi-product in the following In table 4 the results for impurities content at the inlet and
productional steps. outlet of this productional step is given.
- temperature and time for decomposition: 1. This newly developed acid method for producing high-
purity alumina by a new approach for impurities separation
100 C ---------- 120 min' and semi-products obtaining has a remarcable priorities
300 C ---------- 120 min; over the known untill now acid methods used for that
500 C ---------- 120 min; sub-total time - 360 min. purpose.The impurities separation starts already in the first
- temperature and time for calcination: productional step - percolation leaching of aluminium
900 C ---------- 15 min; wastes with sulfuric acid, using the created process
conditions for implementation of cementation and
1000 C ---------- 30 min;
reduction phenomena and proceeds in the following step
1150 C ---------- 120 min; sub-total time - 165 min. for ammonium aluminium sulfate salt formation and
- total time for thermal treatment - 525 min; purification by means of precrystalization with complex-
The chemcal composition of the separated during the step forming agents.
2. The results obtained during the pilot plant investigation
for thermal decomposition gaseous phase, directed toward
show that:
absorbtion column, indicate that the separation of the - by percolation leaching more than 80 pct of the impurities
physicaly and chemicaly combined into a basic ammonium presented into the inlet acid solution and dissoluable
aluminium carbonate salt water starts at 95 C and finish at aluminum can be separated; and, in the followed three
around 150 C, and that of carbon dioxide starts at 90 C and stages counter-current precrystalization of the ammonium
aluminium sulfate salt, using as a complex-forming agents
finish at around 170 C. The separation of ammonia
Salicilic acid ( for 1st and 2nd stage ) and EDTA II (for
however, starts at around 110 C and finish at the 3rd stage ), more than 85 pct of the inlet impurities can be
temperatures over 450 C. Having in mind the chemical separated too, and so, to obtain semi-product with final
composition of this salt it can be suggested that during total impurities content less than 5 ppm;
thermal decomposition some secondary phase permutation - the neutralization of the purified ammonium aluminium
sulfate salt with ammonium bicarbonate for obtaining a
are implemented and a new metastable intermediate phase
basic ammonium aluminium carbonate salt has a notable
are formed. Most probably that may be the case of the effect on the method as whole, because its higher alumina
secondary ammonium aluminate formation [ NH4AJ02 ]. content (aroun 37 pct ) than that of ammonium
The chemical composition and some physical properties of aluminium sulfate salt ( around 10,9 pct ) and lower
465
temperature for thermal decomposition and calcination, is
a fact leading to a great energy savings and sophisticated
off gases utilization.
3.The physical properties and impurities content of the
final product prove that this newly developed method can
be used successfuly for producing a high-purity alumina.
ACKNOWLEDGEMENT
References
l.Bes de Berg 0., Rev.LAluminium, 9, 1, 1975,397.
2. Obzornaya informacia dlya polucenia spec glinozemov.,
Obz.Inf.MINF.Moskva.. 1986.
3.Tanev P., Diis.Doc.Deg. Inst.Chem.Techn .. Burgas,1984
4. Moriwake T., Jap.Light Metals Assos.,14,2,1984,67
5. Kirbi RM., US Pat.Off.. 4,216,010, 1979.
6. Kanev RD. et all., Proc.Inst.Non-Ferr.Met.,Plovdiv,1987.
7. Kanev RD., Diss.Doc.Deg., Techn.Univ.,Sofia,1986,66.
8. Kanev RD.,et all., Metallurgia.Sofia,S,1990,6.
9. Kanev RD., Proc.IMM Conf..Birmingham,S, 1990, Ill.
10.Kanev RD. et.all., Bulg.Pat.Off.,88224,1989.
Hazelett twin-belt aluminium
strip-casting process: caster design
and current product programme of
aluminium alloy sheet
Peter C. Regan
Wojtek Szczypiorski
Hazelett Strip-Casting Corporation, Colchester, Vermont, USA
SYNOPSIS
----.......... ;;-==-
techniques including a single roll, tv~ rolls, maintenance and belt changing.
and a roll and a ring. He eventually concluded
that parallel, endless, moving mold surfaces .,..~
CASTER DESCRIPTION
467
468
Thickness of the cast strip Is establ ished by METAL FEEDING
interchangeable carriage spacers and can vary
between 12mm and 7~. Most aluminum strip is In operation, molten meta 1 .IS " .InJec
. t e d" .I.nto
cast at a thickness of 19mm. the mold cavity by a tundish and nozzle (Fig.
4). Inert gas shrouding is applied in the
Casting width is governed by the lateral slight gap between the nozzle and the belt.
spacing of the side dam blocks and can vary up This acts both to preclude oxygen and control
to the maximum deSign capacity of any parti- heat transfer at the interface between the
cular machine. Current designs allow a maximum sol idifying metal and the belt.
casting width of up to 2 metres.
A starter bar is placed across the mold, and as
For aluminum, the casting angle is set at 6°; the pool reaches the desired height, the caster
a mold length of 1900mm is employed. drive is started so that mold speeds match
metal entry rate. Casting runs of 24 hours are
BELT COOLING SYSTEM common; maximum runs exceed 70 hours.
Mold heat extraction is by fast film water During operation, the operator uses the PLI
cooling on the back side of the belts, repeated (Pool Level Indication System) to monitor the
at intervals (Figs. 2 and 3). The unique, high level of molten metal inside the mold. Belt
efficiency cooling maintains belt temperatures temperatures are recorded (Fig.4), showing the
below 110°C on the casting side and 80°C on the operator the metal pool location on the
water side. INTERACQ display screen (Fig.6). Normally, the
level is kept up close to the tip of the
Belts are supported by grooved back-up rolls nozzle.
which allow full coverage by the cooling water
over the full mold area. These can be adjusted Shutdown of the caster is simple.and changeovers
to shape the mold. of cast width can be accomplished by changing
the tundish.
Exiting water is entrapped and returned to the
recirculation pit while an exhaust system
controls the humidity in the mold area prior to
the casting zone, to prevent condensation on
the belts. POOL lEVEL
INDICATOR (PLI)
Casting belts are low carbon, cold rolled steel USED WITH
CLOSED POOL
strip 1.2mm thick. Side dams are low carbon
steel. Belts and side dams are easily replace- FEEDING 't
able and can be changed in 20 minutes, using
specially designed fixtures.
t
Lower Belt
_ Bottom
Nip Pulley
End Products
Cast Ccmrerclally
HAZELETT ALLMINLI'I CASTING AND ROLLING INSTALLATIONS
.
powered software. W"1'ER EXIT CASTF.R
Real time data display and storage is accanpll"'..- I
n: II', TE IP. SI'Em
I
ed with the INTERACQ system using approximately 0 111
20 to 30 channels of data. Screen displays are
custom generated, with their number in the •
range of hundreds. (Figs. 6 and 7) Functions .. :JII'j; 3e " I.,rnout,....
rn Itn II ~R I
.... ';5 '111 - IICNIII. 3 5
such as belt temperature and flatness information , S 1
.I-.
equipment. The quality of the product will
depend on the quality of operation in each of
these areas .
~
• Tn 1[" _
'-....
•
... .. ,.,
lU'.~ :
e> .....LPI: : _
The Hazelett process can be included in the
" rLl'. , ,.,.
PIIJI , . . . t2
:
I group of high solidification rate processes.
.... , . , . 0 .es From the solidification point of view, it is
'''''''14 ...
" :n
I..
=J
'"
similar to that of the Alusuisse Caster II
(block caster), faster than the conventional DC
.-........
uS) I
In 51
I
(direct chill) process and slower than the twin
lCI1't2TC:t
950 roll TYPE caster. The specific heat removal
-,
I Inc:ao.r 1M.
'MSI"f nut flCl\ft. 1119
COtto on ' " rate determines metallurgical and mechanical
characteristics of alloys cast on the Hazelett
caster. Solidification rates can be increased
or decreased to a certAin extent by using a
combination of Matrix coatings and inert gas
Fig. 6 INTERACQ Display shrouding.
Caster Parameters
When comparing Hazelett-cast and DC-cast
devices capable of detecting belt movements or aluminum alloys, the differences will become
displacement of other machine components. quite obvious.
Sensitivity is 0.2~ with a range of 4mm. The
probes do not contact the belt and do not A. Grain Structure
seriously interfere with cooling water.
Higher solidification rates achieved on the
Belt temperature is sensed by thermocouple-based Hazelett caster will result in smaller grain
units which sense the temperature of the water size, finer dendritic cells, smaller size of
side of the casting belts. They are arranged secondary phase particles and their different
at various locations, in particular close to morphology as well as higher retention of
the point where the molten metal fi~st contacts certain elements in solid solution. This will
the belt, as shown in Fig. 4. also ensure the necessary surface quality for
very demanding appl ications such as can stock.
The caster/pinch roll relationship is monitored However, in canparison with twin roll cast
by both the actual rotational speed of both material, the Hazelett-cast strip has bigger
components, as well as with industrial quality grains, coarser dendritic cells, larger size
rotating transformer-coupled strain gauge secondary particles and less retention of
transducers installed between the caster and elements in solid solution.
pinch roll drive motors and their respective
gear systems to provide information on both The high solidification rates ensure fine,
relative speed and torque. equiaxed and randomly textured grains throughout
471
II .66 .25 .15 .59 .23 .042 .015 .021 .006 .007
Table III represents some typical grain size
measurements for corrmon aluminum alloys cast on
the Hazelett caster. The work hardening curves for the samples of
both chemistries are shown on Fig. 8.
TABLE II I
As can be seen, Mn content has a significant
effect on mechanical properties of Hazelett
cast AA3105.
ALLOY GRAIN SIZE (Pm) WOAK-tWlDENINa Cl.llVES FOR
Surface Center HAZELETT CAST AI. 3105 SlEET
(CI£IIISTRY II
E
\
AA 3004 80 150 \
\
16
\ 14 !
;!
AA 3105 130 250 \6
\ 12 ..
-'....x:'
\
~
~
improve with increasing cooling rates which 10
result in finer and more homogeneous structure, ~
especially for higher alloys.
\
,, 8
8
iii
21,
B. Solid Solution . . . . _-"'a-- 7o------_A
A
__ -A.._·r [
4
2
SUtvlt'IARY
SYNOPSIS
monium tartrate, boric acid, ammonium citrate, etc., have
little or no capacity to dissolve the oxide. The films are
The increase of the applications of aluminium and its
relatively thin (less than 1 JlIl1) and essentially non-porous l .
alloys is accompanied by the possibilities of using surface
Anodizing is usually combined with pre and post
treatments to create specific surface properties. These sur-
treatments in order to obtain the required AI-surface. The
face treatments are used to increase the corrosion or wear
type of pre and post treatments is dependent on the ap-
resistance, or to create special electrical, optical, hydrophilic
plication of the anodized AI-surface. It is very important to
or adhesion properties
understand the interaction between the various treatments
The specific surface characteristics are related to the
that are imposed on the AI-surface.
morphology, structure, composition and even charge of the
The variety of pre and post treatments that can be
modified Al surface.
imposed is rather large due to the high number of appli-
The success of a specific treatment is strongly dependent
on the precise understanding of the surface behaviour of the cations. The pretreatments can be classified as mechanical,
chemical and electrochemical, commonly used to grain,
used Al alloy during the treatment.
In this article three types of treatment are considered: etch, polish, ... , the surface prior to anodizingl.
AC electrolytic graining, porous anodizing and sealing. After anodic oxidation, the porous oxide layer is usually
sealed to improve the corrosion resistance. When the po-
rous films are heated in water, steam or a salt solution, the
INTRODUCTION pores are closed due to the reaction between the oxide and
the sealing medium.
Anodizing and allied processes have been studied The success of the treatment is related to the precise
widely because of the commercial importance of the formed understanding of the surface behaviour of the selected Al al-
anodic oxide film, which can give the AI-surface specific loy during the treatment, which requires the understanding
properties, for example : a high corrosion resistance, a de- of interaction between the material input (type of AI-alloy,
corative aspect in architectural applications, adhesive, mechanical and thermal pretreatment) and the parameters of
electrical or optical properties. the treatment.
Anodizing is defined as an electrochemical treatment in In this article three important surface treatments are
which stable oxide films are formed on the surface of me- discussed, namely AC electrolytic graining of Al which is
tals I. Anodic coatings can be formed on aluminium used as a pretreatments to roughen the AI-substrate, the
employing a variety of electrolytes using either AC or DC porous anodizing process itself and its sealing.
current or a combination of both. The complementary information deduced from electro-
Anodic coatings are classified according to the chemical and surface analytical techniques was used to
dissolving action of the particular electrolyte on the anodic characterize these treatments in detail I.
oxide produced in the reaction. Anodic oxide films Table I gives an overview of the used methods with
produced in sulphuric, phosphoric of chromic acid their respective characteristics. The obtained results are dis-
electrolytes are of the porous typel. cussed more in details in a number of pUblications indicated
Anodic oxide fllms produced in electrolytes such as am- in table I.
473
474
Table I
AC electrolytic graining of AI involves the application of Through variation of the electro graining frequency in a
an alternating voltage or current between aluminium elec- region from 0.1 to 1000 Hz and by variation of the anodic
trodes in a suitable electrolyte, e.g. hydrochloric acid or and cathodic charges of the AC cycle, it was evident that the
nitric acid. The object of this industrial treatment is to hemispherical pits propagate through the development of a
develop a uniformly pitted and convoluted surface topo- high population of cubic pits, formed by the (100) faces of
graphy, which improves the water retentive properties of the FCC lattice. These pits are fonned by the anodic charge.
aluminium offset plates for lithography. It is used as a An example of these pits is shown in the TEM micrograph
pretreatment prior to anodizing. of Fig.3. The size of the cubic pits can be related to four
The depth and distribution of the pits are dependent important grain parameters :
upon particular electrical conditions (applied current den- 1) the number of pits, initiated each time the positive cycle
sity, graining frequency, fonn of the AC signal) and the is passed;
electrolyte parameters (composition of the electrolyte, pH, 2) the amplitude of the sinoidal current:
temperature, ... ). 3) its frequency;
In order to understand how the morphology of the 4) the anodic current efficiency in the process.
grained surface is fonned and how the treatment parameters It is especially on the first point that the electrode pre-
act, it was necessary to determine the AC electrolytic treatment (selected AI-alloy, thennal and mechanical pre-
graining mechanism. treatment) acts.
The obtained morphology after AC electrolytic graining It was also proved that during the negative half-cycle the
at 50 Hz in a hydrochloric acid solution is shown in the non-unifonn etch film is developed over the aluminium sur-
SEM micrograph of Fig.l. The surface roughness is face (Fig.2). By examination of the etch film development
created by the presence of a large density of hemispherical as a function of the AC wave fonn and frequency, it was
pits, having no characteristic dimension. Beside the for- obvious that the mm develops during the negative half cycle
mation of these pits, a thick friable etch film is present, re- only when sufficient alkanity occurs. This can be explained
presented in Fig.2, which has to be removed prior to suc- by the hydrogen evolution at negative polarization, causing
cessive treatments. a significant increase of the pH.
475
From the determination of the mechanism of the AC oxide film is shown in the TEM view of Fig.6. The oxide
electrolytic graining process it is now possible to layer formed on flat electropoli~hed surfaces, has a very
understand the impact of following parameters : regular structure. Cell, pore and barrier layer dimensions
current density of the applied signal; depend on the anodizing conditions and have different
frequency of the applied signal; anodizing ratios (nm/V).
form of the applied signal; The shape of the voltage-time transient described is
electrolyte condition; retained over a wide range of anodizing conditions. The
pH of the electrolyte; rate of the voltage rise and the steady-state voltage are
temperature of the electrolyte; influenced by the applied current density and decrease
influence of the Al preconditioning. according to the relative aggressiveness of the acid used.
The composition and roughness of the aluminium sub-
Study of the Porous Anodic Oxidation of AI. strate may influence the anodic oxide film formation. A
significant part of the literature deals with growth of porous
(TEM) Transmission electrode microscopy studies of layers upon electropolished surfaces 1, and little information
the growth of porous layers in sulphuric and phosphoric is available about the role of the composition and
acid permitted description of the morphology. A more topography of the aluminium surface. However this is of
detailed reviewing can be found in 1,14,19. In these elec- industrial importance where the oxide film is often formed
trolytes porous oxide films are developed, having a barrier on extruded profiles and on rolled or even grained sur-
layer in contact with the metal and a porous film above it, faces 1.
usually thicker than the barrier layer. The pores appear as The growth of porous oxide layers on relatively rough
cappilaries through the anodic film. A schematic represen- surfaces, such as rolled and AC electrograined aluminium
tation of this porous anodic film on rolled Al is given in has been studied.
Fig.4. Therefore electrochemical measurements 1,4,5,6,10,14,18,
Porous anodic films may be formed under constant 21 have been recorded during the growth of the duplex
current density (or constant voltage) and the voltage (cur- oxide layer and combined with high voltage electron
rent) time behaviour is characteristic of the way the oxide microscopy. The study leads to following conclusions. The
film grows. Such a voltage time transient is given in Fig.5. growth of the porous layer on rolled and AC electrolytic
Oxidation under constant current density shows an grained aluminium is similar to that of the porous oxide
approximately linear increase in voltage from the beginning. films formed on electropolished flat surfaces. Steady state
The growth of a porous oxide layer on aluminium in- porous cell dimensions, barrier layer thickness, pore and
volves a competition between oxide formation and disso- cell diameter are proportional to he anodizing voltage.
lution. Initially a compact barrier layer is formed by ionic Fig.8 shows a cross-sectional view of the porous layer
migration. Fig.6 shows a cross-sectional TEM view of formed on the AC grained surface. Although the steady-
such a barrier layer. It is believed that A13+ and 02-IOH- state porous cells were observed near the metal-film
ions are both mobile in the field. The extend of solid film interface, complete distribution of irregularly shaped pores
formation at the electrolyte interface depends critically upon was observed on the outer surface of the oxide film This is
the anodizing current density, pH and electrolyte composi- due to the fact that each pore tends to orientate perpen-
tion. A non porous oxide continues to grow for as long as dicularly to the surface. Crossing of the pores and irre-
ionic current continues to flow. The thickness of the barrier gularities will particularly occur as the curvature of the alu-
layer depends upon the voltage drop across the oxide and is minium surface increases.
usually characterized by the anodizing ration, expressed in From the study of the porous oxide film formation it is
nmN. now possible to understand the influence of the following
Mter a time the linear relation ceases and the value parameters :
increases slowly to a maximum. This behaviour reflects the current density;
onset of non-uniform film growth and pore development. potential of the anodic oxidations process;
After reaching a maximum the voltage declines to a steady- composition of the electrolyte;
state accompanying the formation of a regular porous temperature of the electrolyte;
anodic film which thickens according to Faraday's law. influence of the Al roughness.
Non uniform growth appears in aggressive electrolytes,
due to dissolution effects, caused by field assisted dis- Study of the Sealin& Process of Porous Oxide Films
solution and Joule heating, which results in the formation of
pores. Thereafter regular porous anodic film growth occurs Sealing of the porous anodic films is usually carried out
and the oxide film thickness in proportion to anodic to improve the corrosion resistance of the layer. During the
oxidation time while the barrier layer is maintained at a sealing process the pores are filled due to the reaction of
constant thickness beneath each pore. This regular porous Al203 in boiling water, which results in the formation of
476
boehmite4.6.7.11.18.19. The structure and composition of applicable in situ.
the porous layer are significantly changed, especially in the It also presents the advantage of being extremely sensi-
upper layer of the Al203 film. Beside the closing of the tive to minute surface effects, like thin film formation or
pores, hydrated aluminium products are also deposited on morphological changes (e.g. roughening). From the mea-
top of the oxide surface. Fig.9 shows a cross section of a sured reflective properties, a characterization of the physical
thoroughly sealed porous film formed in phosphoric acid. and chemical state of the surface can be deduced by means
The closing of the pores clearly improves the corrosion of simulation and regression procedure. In this procedure,
resistance. Yet, the presence of the sealing products on the the ellipsometric response of a theoretical multilayer model
oxide surface hinder the adhesive properties of the Al20 3 of the surface structure is fitted to the measured response by
layer and makes it more unstable in time, again due to the a least square regression analysis. The model which allows
formed boehmite (Fig.lO). an optimal accordance is retained as the characterization of
the surface structure, provided it is physically realistic. As
Characterization of surface treatments by impedance spec- an example Fig. 12 shows the ellipsometric data recorded
troscopy and scanning ellipsomeny on a porous oxide layer, from which the thickness can be
easily determined.
The characterization of surface treatment is important to
develop surface analytical techniques, which are able to
determine the surface characteristics in a fast, non- CONQ..USION
destructive way.
Indeed,techniques as TEM, Auger, SIMS, XPS, require Using the complementary information, issued from
high vacuum and often a difficult sample preparation. electrochemical and surface analytical techniques it is pos-
Impedance and ellipsometry measurements are used to sible to characterize some important AI-surface treatments
analyse the surface where respectively from the electrical and to understand the influence of most important para-
and optical characteristics of the surface information is meters. The results allow to tailor the AI-surface properties
gained. The attention is focussed on the characterization of in a known way and possible undesirable side-effect can be
oxide thickness (barrier and porous layer thickness) hy- avoided. The knowledge and experience obtained this
dration of the oxide layer, roughness, and detection of the study can be extrapolated to other surface treatments.
corrosion resistance.
5. TERRYN H., VEREECKEN I., VANHELLEMONT 13.HUBRECHT I., TERRYN H., VAN DER LINDEN B.
1., VAN LANDUYT I. VEREECKEN I.
"HVEM Investigation of AVAl203 Structures for Offset "Characterization of Protective Coatings with Electro-
Purposes", Proceedings Symposium on Aluminum chemical Impedance Spectroscopy", Abstracts Interna-
Surface Treatment Technology, Ed. Alwitt, Electro- tional Symposium Corrosion Science and Engineering,
chemical Society, Vo1.86, 11, 1986, pp.291-301. in honour of M. Pourbaix's 85th birthday, Brussels,
12-15/3/89, pp. 72-73.
6. VAN HELLE MONT I., TERRYN H., VAN LAN-
DUYT 1., VEREECKEN 1. 14.TERRYN H., VAN HELLEMONT I., VAN LAN-
"HVEM Study of Porous Al203 Layers Formed by DUYT I., VEREECKEN I.
Anodic Oxidation in Sulphuric and Phosphoric Acid", "Influence of the Aluminium Pretreatment on the
Proceedings XIth Int. Congres on Electron Microscopy, Growth of Porous Oxide Films", The Insitution of
9/9/86, Kyoto, lapan, 1986, p.1069. Metal Finishing "Surface Finishing", 11-14/04/89,
Brighton, England, pp. 135-153.
7.VAN DER LINDEN B., TERRYN H., VEREECKEN 1.
"The Use of Impedance Measurements to Characterize 15. TERRYN H., VEREECKEN I.
Various Aluminium Pretreatments", Preprints, Euro- "Impedance Analysis of Porous Oxide Films Formed on
corr'87, 6-10/4/ 87, Karlsruhe, Germany, 1987, Different Pretreated Al Surfaces", Extended Abstracts of
pp.579-761 the 1st International Symposium on Electrochemical
Impedance Spectroscopy, 22-26.05.89, Bombannes,
8. HUB RECHT I., TERRYN H., VAN DER LINDEN France, p. C4.8.
B., VEREECKEN I.
"Caracterisation des traitements de surface it l'aide des 16.VAN DER LINDEN B., TERRYN H., VEREECKEN
mesures d'impedance", 2eme Forum sur les Impedances I.
Electrochimiques, 28-29/10/87, Montrouge, France, "Fractal Approach of the Impedance of A1203/ Al
1987, pp.43-52. Layers", Extended Abstracts of the 1st International
Symposium on Electrochemical Impedance. Spectrosco-
9. TERRYN, H., VEREECKEN, 1., rnOMPSON G. py, 22-26.05.1989, Bombannes, France, p. C2.15.
"Investigation of the AC Graining Process of AI",
Proceedings of the Anual Technical Conference and 17.TERRYN H., VEREECKEN, I., rnOMPSON G.
Exhibition, 27-29/4/88, Stratford upon Avon, United "Electrograining of Aluminium in Hydrochloric Acid"
Kingdom, pp.113-134. part 1 and 2, Accepted in Corrosion Science
12.TERRYN H., VEREECKEN 1., THOMPSON G. 21.VAN DER LINDEN B., TERRYN H., VEREECKEN
"Investigation of the AC Graining Process of AI", I.
Iournal of the Institute of Metal Finishing, Vo1.66, "Investigation of anodic aluminium oxide layers by
1988, p.116. electrochemical impedance spectroscopy", Iournal of
Applied Electrochemistry, 20,1990, pp. 798-803.
478
o - - - + '......
Fig.6 : TEM cross sectional view of a barrier layer. Fig.8b : at a higher magnification in one hemispherical pit.
I
!.
.
.:.'"
' 3
.
2
"
::' . ' . 1[; • <::::11,--
... . ..
.. '
: " ! II! I! II'
.1 1
O+-~--r-~~--__'-~--r-__~--__'-~~
200 300 .00 500 600 100 800 100
W..... t.nglh (nm)
'.
., +-----...L..,.....!!-.-~-.,r--~""T"~--r__~""T"~____:_l
200 300 .00 500 600 700 800 DOO
100
-
.00
~~D---------~------------------D
:i;~~~~=·-=F:=m~====F===~=='
• '00
SEAUNG TIME
000
(5)
483
484
similar way as Derby et aI.2. Flexible meshes are used in quasi-steady state contribution proportional to the pulling
order to represent the time dependence of the geometry. rate and a time-dependent contribution dependent upon the
motion of the interface with respect to the melt level, due
Three types of constituents corresponding to different heat to the inertia of the interface. The latter contribution may
transfer modes are considered. Each type is characterised reach high values in rapidly changing geometries.
by a typical equation which is discretised in space and in
time if necessary and leads to a local system of equations. When allowing the tri-junction radius to be free, we
These local systems are then coupled in order to obtain a determine the shape of the crystal on the basis of a
global solution. In what follows, we present each type of coupling between the Laplace-Young equation governing
constituent. the meniscus shape with the thermodynamic condition of
contact angle (.0) constancy which relates the meniscus
Radiatiye enclosures angle. to the growth angle 'I' (see figure 1):
Solid bodies
aT
pc at + pc X<!' t) . YT - y.[ k(~ YT] = r(!, t)
where Tfl and Tn are the temperature and the time derivative
of the temperature at instant tn and ~tn+l=tn+l_tn. A
Global solution
condensation technique allows one to eliminate internal
temperatures and to obtain a linear system relating nodal Global computation is performed at every time step by
temperatures and nodal heat fluxes on the boundary of the assembling the local equations relating nodal temperatures
element. and heat fluxes on the boundary of each element. After
elimination of the heat fluxes, one obtains a non-linear
system which is solved by means of a Newton method. A
Ctystal-melt element complete description of the method is given in
This element may be considered as a solid body. Dupret et aJI.
However, as the crystal solidifies, heat is released along
The overall numerical procedure may be summarized as
the solidification front. This solidification heat includes a
follows:
485
- computation of an initial quasi-steady state solution: this /
,--------------------------------- ........
.
solution is obtained by solving the equations with all
time-derivatives set to zero; ,/////,/, ...............
- for each discrete instant:
~
I
,
I
I
I
I
i) automatic generation of a new geometry: this includes I
I
I
mesh deformations and generation of the I
I
:=J
I
Lr 1
A detailed description of our time-dependent numerical
method may be found in Van den Bogaert et al.3.
e
the isotherms in the melt during conical and during Table I. Thermal properties of the different constituents of
cylindrical growth is quite different. The heater power is the furnace.
486
~
.....
~
-y
~
t-'{" ~
..loor"C
~~
i.- ~
~ y,,~
~ _'\ "\.,,'~
:I \ '\. '''~
\ \. \..~
'\. ""~
'"'~
~ '" '""'......
~
1211 1211
1231 1231
1171
1171
1211
1231
1231
1131
1131
1171
1231
1231
Figure 4. Isothenns in the melt and in the crystal, separated by 10 K for various stages of growth.
488
higher during cylindrical growt~. As a result, !he The heater power computed with our ti~e:dependent
temperature in the melt is also hIgher and the meltmg model is compared to the heater power ansmg from a
isotherm rises: the interface becomes thus less conyex. In quasi-steady state calculation in figure 6. The steady and
cylindrical growth, the vertical crystal wall re~~cts the time-dependent curves differ considerably during conical
amount of heat leaving the melt surface by radiation and growth and during the shouldering proce~s. A~ a n,tatter
isotherms become more perpendicular to the surface of the of fact, transient effects are important dunng thIS stage of
melt. growth, as geometrical changes are fast. The two c,urv~s
The resulting heater power is represented in fi~ 5. The are very close to each other once a constant radIUS IS
heater power is highly discontinuous at two precIse stages: achieved as geometrical changes are much slower at that
first when starting the simulation, and second, when the stage of growth. This means that transient effects m:e
gro~th angle starts decreasing. pus d!scontinui~ of the negligible during constant radius growth and that quasl-
result with respect to the data IS tYPICal of an mverse steady state simulations are valid.
problem: a slight m?~ifica.tion ?f the situation in ~he During conical growth, the time-dependent heater power ~s
environment of the tn-JunctIon - mduced by the varymg
radius - may yield an important response of the heater much lower than the quasi-steady state hea~r .pow~r. ThIS
power. This problem is addessed in N. Van den Bogaert may be attributed in part to the release of solidification heat
due to the motion of the solidlliquid interface which is
et al. 3 . In order to obtain a legible result we have taken into account in the time-dependent model. The
smoothed the heater power time variation curve. To assess increase of power during the shouldering proces occurs
the validity of the smoothing, we have solved the problem once the growth angle starts decreasing. The power
in a direct way: the smoothed power variation was impo~ increases more rapidly in time-dependent calculations as
and the corresponding radius was calculated. The resulting some delay is necessary for the changes arising in the
crystal shape was extremely close to the \nitial imposed heater to reach the tri-junction line. In quasi-steady state
shape: the maximum difference was about 0.3 mm only. calculations one assumes that the whole power variation
effect is felt instantaneously at the tri-junction. In time-
We may observe that the heater power is abo~t 20.572 W dependent calculations the major effect is felt about one
during conical growth and about 21.19.5 W dunng cons~t hour later due to the thermal inertia of the crucible and
radius growth, thus 3 % higher. Dun~g the shou~de~ng, melt. In ~er to produce the same radius variation in the
i.e., during the transition between comcal and cyhndrical early instants following the perturbation, the required
growth the power rises rapidly at first in order to stop power variation ~ust b~ much higher in ti~e-dependent
radial ~wth and decreases subsequently in ~r to ~v?id calculations than m'quasl-steady state calculations.
that the radius starts decreasing. The shgbt time vanation
of power during constant radius growth is due to the
changes of geometry.
KW
q
!!
II h
II
II
'I
n II
21.4 )1 II
ft
II 11f1 !I!1
II II
:1 'I II
': • 'I :11
;1 il i I r, PI r\
II
Ii n
II I, " II I \ I I
I
II II
II
II
II
II
:J i
!!1;
11
o 2 4 6 8 10 em
Figure 5. Heater power as a function of crystal
length. Time-dependent computation
(dashed line). Same curve after
smoothing (solid line).
489
KW em
21.4
III
21.0
,
,,
o
,,
I
/ {-
I
"
I
,
"
20.6
o 2 4 6 8 10 em
4
mm
tri-junction
axis of symmetry
2 ,"~ " ...... --~----
l
/~
I
I '.
I
I
o I
melt level
·2
·4
·6
III
Figure 7 shows the evolution of the position of the Figure 8 represents the heat released at the interface. The
solidlliquid interface with respect to the melt level. Two quasi-steady state contribution, due to the pulling rate, is
points of the interface are represented: the point on the axis proportional to the crystal section. As a result, such
of symmetry and the tri-junction. The interface thermal contribution varies as the square of the diameter during
inertia considered in the time-dependent simulation causes conical growth and remains constant during cylindrical
a delay to appear in the interface position with respect to growth. The contribution due to the motion of the
the steady simulation. The position of the interface on the interface is high during the shouldering as the interface
axis of symmetry decreases during conical growth, rises rapidly. It reaches up to 40% of the quasi-steady
increases once the growth angle starts decreasing, reaches state contribution and is of opposite sign. The quasi-
a peak value during the shouldering, and decreases steady state contribution corresponds to heat release and
towards a constant value corresponding to constant radius the time-dependent contribution corresponds to heat
growth. The interface remains convex with respect to the absorption. This latter contribution induces the delay of
crystal during the whole process. The difference between the interface position which has been observed in figure 7.
the vertical position of the tri-junction line and the interface Once cylindrical growth is reached, contribution due to
axial position, also called deflection, is about 5 mm during interface velocity almost vanishes as the solidification front
conical growth, decreases down to 2mm during the remains close to the same position.
shouldering and reaches 4 mm for constant radius growth.
The tri-junction height (or meniscus height) is an Figure 9 shows the history of the successive interface
immediate result of the imposed radius. The initial positions. An interface is represented for every
meniscus height is 0.9 mm corresponding to a growth 600 seconds period. One may observe that the interfaces
angle of 60°. It rises rapidly when the growth angle starts are close to each other during the shouldering.
decreasing and reaches an almost constant value of 4 mm
during cylindrical growth.
491
II ,-, III
40 I
,, ,,,
:! \\ '\",
I ,
,, '
:
I "
I '
................
o , ... -~~----
::..................
,....... I
/
"......... I
" ......... /
......... _,,'
- 40
-80
- - - - - _._..
- 120
o 4000 8000 12000
s
Figure 8. Solidification heat at the interface as a
function of time: quasi-steady state
contribution, due to the pulling rate (dash-
dotted line); time-dependent contribution,
due to the change of interface position
(dashed line) and total amount (solid line).
CONCLUSION
SYNOPSIS
An equation-oriented simulation model is of design in order to provide guidance as to
developed and presented for a new the interplay of the operating variables and
hydrometallurgical process for recovery of as a basis for quantitative evaluation. In
non-ferrous metals from Iberian complex sulphide pract i ce the development and complexity of
ore concentrates, by means of concentrated a model evolves iteratively by means of on-going
ammonium chloride aqueous solution. validation and integration of the information
gathered experimentally with the practical
The modelling approach employed for the and theoretical background knowledge available.
elementary unit operations is analyzed and
discussed together with the strategy adopted for While for process design it is required of
representing the diagram structure, with the a model that it should simulate adequately,
objective of conferring on the model a high on the basis of mass and heat balances and
degree of responsiveness in terms which are within an acceptable numerical approximation,
described as self-regulating steady-state the process steady-state conditions, this same
characteristics. objective can be met with a variable degree
of elaboration. In particular in the case of
The reference conditions for processing 20000 highly concentrated electrolyte solutions and
kg/h of Portuguese ore concentrate are presented complex patterns of interconnecting streams
and simulation runs are applied to reassess the to be found in many hydrometallurgical processes,
val idity of some design parameters namely, the a simulation model should ideally al so enable
extent of the copper reduction reaction, the pulp the estimation of operating parameters and
density and the by-pass stream flow which feeds variables which need to be preset in order
the silver cementator. The effects are measured to meet process specifications.
in terms of changes in the rate of consumption
of by-feeds, process water and oxygen as well The development of models which meet these
as on the solubility and recovery yields of the goals while presenting the necessary flexibility
metal values. On the basis of these results and ease of use required for repeated prospective
design alternatives are analysed and discussed. changes and examination of alternative diagram
configurations, without excessive programming
effort, can be satisfactorily achi eved on the
basis of the so-called equation-oriented
Process simulation is a well established simulation packages of which SPEEDUP is a well
technique in process design which can in addition known example.
be highly effective in effort reviewing and
p1anni ng when experimenta 1 development work In this study a simulation model with the above
is also involved. characteristics is presented which was set
up in the course of the experimental development,
This is particularly relevant for under an EEC cofinanced project, of a new
hydrometallurgical processes whose complexity hydrometallurgical process for recovery of
is usually great both as a result of the non-ferrous metals from sulphide concentrates
composition and constitution of the raw materials of Iberian ores by direct attack with a highly
which contain a vast number of components and concentrated ammonium chloride leaching solution.
reactional mechanisms not fully understood
and diagram configurations based on intricate The process involves two major leaching
networks of operations and streams required operati ons and a number of metal recovery
for satisfactory process performance. operations which include the solvent extraction
of copper and zinc.
Greater advantage would be made of the simulation
model if it were available at the early stages The simulation results obtained for the assumed
493
494
reference conditions are presented for the Figure 1 shows the information diagram for
Portuguese ore concentrate and the discussion the global process whose detailed description
is centred on the handling of some by-pass can be found elsewhere 2• The former is a
streams and the recycle of the zinc solvent simplification of the latter used as a basis
extraction raffinate which controls the dilution for the computer representation.
required to prevent the precipitation of .zinc
and/or copper(II) diamines in the crystallization It shoul d be noted that batteri es of reactors
of lead chloride. are depicted by single models (Rl to R8) and
the solvent extraction operations (El and
E2) by models of the black-box type. The reasons
SPEEDUP SYSTEM for this approach are three-fold:
Speedup is an "equation-oriented flowsheet (i) Most reactor models are empi ri ca 1 and
simulation and optimization system for process based on preset chemical reaction sequences
engi neeri ng systems" I, typi ca lly those with and maximum reaction extents, with the
a modular structure. exception of the lead chloride crystallizer
(R5) and the sulphates removal unit (R6)
It is beyond the scope of thi s paper to di scuss which are based on multiple reaction
the many features of the package but reference equilibria and the silver and the lead
is next made to those which influenced directly cementators (R4 and R7) which use solubility
the modelling approach. curves fit to experimental data over a
wide range of temperatures and compositions.
(i) While the calculation of the units within Thus, whi 1e these models respond correctly
the flowsheet is not order-dependent, the to changes in flow composition and also
modular structure of the process diagram to some extent to temperature variations
is maintained both for problem input and (through the equil i bri um constants and
data output - physical units are represented solubility curves) they remain insensitive
by models which can be developed to residence times.
independently; the output can be inspected Consequently there are no gains with regard
on an unit basis. to the calculation of the steady-state
operating condi ti ons to extend the number
(ii) Provided that the problem is correctly of reactors of each type unless there
specified in terms of equations and set were intermediate feeds or changes in
variables and/or parameters, the unknowns the operating temperatures, which is not
to be calculated may be input or output the case.
streams and/or parameters, or combinations
of these - by-feeds, make-up streams, (ii) The solvent extraction operations have
process water requirements can thus be in common a fairly large number of
calculated together with output streams. counter-current stages and closed-loop
streams which, if incorporated into the
(iii) The models consist essentially of equations global process model, would add
which need not to be variable assignments, considerably to its complexity. This
whose order is arbitrary and which are was considered undesirable and the detailed
solved simultaneously within blocks, treatment of these two operations was
(matrices of lower order), and of carried out separately and falls outside
procedures which gather the calculation the scope of this paper.
routines whi ch are structured sequenti ally To illustrate this point in table. I a
or which are trivial and would increase comparison is drawn between these three
unnecessarily the total number of equations simulation systems and it can be noted
the non-linear algebraic equations the considerable magnitude of the solvent
associated with the chemical equilibria extraction systems, in particular the
are an example of the former while the zinc extraction which is the most complex
empirical models employed for leaching because of its upstream position in the
and the convers i on of phys i ca 1 un its process diagram.
(e.g., kmol/h to kg/h) are examples of
the latter. (iii) The main focus of the process is on
the interplay of the leaching operations
(iv) The same problem input file is used for (including copper reduction) and its
simulation, optimization or parameter effect on metal values solubil ization
estimation while the steady-state and recovery, on the process
operating conditions are calculated by self-sufficiency in leaching agent,
simulation, some of the equilibrium the distribution of the pair NH3/NH4+,
constants were estimated from experimental the inhibition of copper and/or zinc
data pri or to s imul ati on and one of the precipitates, the consumption of by-feeds
recycles requires an optimization procedure. and so on.
The switching between running modes requires These global effects require that a
only minor input file modifications. considerable emphasis should be placed
on the correct representation of the
diagram structure as expressed by the
PROCESS DIAGRAM pattern of interconnecting, by-pass
and recycle streams and positioning
Modelling approach of the make-up units and inlet points
Legend:
BYl - Sil ver By-pass FL i - Flash Tanks R6 -
Sulphates Removal
BY2 - Zinc Extraction By-pass MI - NH4Cl Make-Up R7 -
Lead Cementation
Cl - Concentrate Feeder M2 - NH3 or H2S04 Make-Up R8 -
Copper(I) Oxidation
COi - Condensers Mxi - Mi xers PI -
Heat Exchanger
CTi - Centrifuges RI - Neutral Leaching Tl -
Pulp Tank
Oi - Decanters R2 - Copper(II) Reduction T2 -
Gaseous Tank
DIVi - Stream Dividers R3 - Acid Leaching Gaseous/Aqueous Streams
El - Zinc Extraction R4 - Silver Cementation ..... - Pulp/Solid Streams
E2 - Copper Extraction RS - Lead Crystallization - Information streams
Fi - Fi lters
H20 H20
I
I
I
I
I
IL ___________ • .j:>.
(0
r---------------J
.-------~----~~ H20 I C11
I
I
I
I
I
I
I
I
I
FINAL :
PE;IDUE I
I
I
I
I
I
~------------------- I
r----------------------J I
I
I
I
r-------------------------------------------------- -------------~1 t
IIH3
H2:0~
Average Average
Total Number No.Equat.= Speedup Validation time Run time
Designation of variables = No.Unknowns Input File
(Kbytes) (CPU s) (CPU s)
Tl/FL3 Water flow required in Tl to prepare To estimate the water flow to be removed as
concentrate feed pulp low pressure steam in FL3.
Tl/DIV2 Water flow required in Tl to prepare To estimate aqueous phase to be fed to Tl
concentrate feed pulp
RI/PI Estimated temperature of inlet To control the extent of heat exchange in PI
aqueous phase in RI for adiabatic and indirectly the head load in R3
operation
R3/M2 H+ requirements in R3 for leaching To estimate make-up of NH3 or H2S04 in M2
the hydroxi-chlorides
R5/BYI (Cl-) free in solution To estimate the (Ag+) at saturation in R5 and
either act on the by-pass to allow a main stream
flow with the equivalent amount of Ag+ or apply
a 10% split ratio. whichever in less restrictive
R5/DIV R5 inlet and outlet concentration To estimate the minimum recycle from DIV to
of Zn(NH3)22+. Cu(NH3)22+ and free MX2 to avoid diamines precipitation (optimization
(Cl-) procedure)
El/R7 ZnC12 flow required in zinc To estimate the process water feed in R7
extraction for the scrubbing of
copper
El/BY2 ZnC12 aqueous solution composition To estimate the by-pass flow to R8. assuming
total Zinc extraction in El and (Zn 2+)=5 g/1000g
H20
To Sliver
Ce-n'\enl&bon
--- 01
Prom Coppe;:--'-
Olodellon bnd
Soh'tonl
ExtractIOn
.j::>.
(0
00
R3 -------
H.O
FL2
OIV2
..L--
Pm8l
L--------::;"---=---.:---11 0 1· t residue
'-
H2 S0 4 0 0 8.
Q.
.0.1
u:l 0.10
Zn 3.297 2.155 E+2 "0
Cu 4.130 E-1 2.626 E+1 E
Q.
Pb 1.240 E-1 2.589 E+1 c:
0
0.08
Ag 3.830 E-1
~
c ._-
Evaporated Water 9.888 E+2 1. 780 E+4 Q)
()
c:
0,06 -' --,-
0
Final Residue 1. 259 E+4 ()
Q)
c: 0.04
Gypsum 5.897 E+1 1.015 E+4 'E
IV
'6
()
c: 0.02
N Oparation
Table XI - Non-ferrous metals solubilization 0.00
and recovery yields 0.0 0.2 0,4 0.6 0,8 1.0
R2.Cr
""
60 percent, while the oxygen consumption
increases as a result of the unreacted sulphides
"
N
o
~
100
'/
~-
~
/
99
SoIubilizalion
-
:::I
() 98 L ./
.......
----
o
0.0 0.2 0.4
R2.Cr
0.6 0.8 1.0
V
......
'# Recovery
97
/ Figure 7 - Copper consumption versus copper(II)
reduction conversion
3.90
96
U.O 0.2 0.4 0.6 0.8 1.0
/
R2.Cr
V
3.89
yields versus copper(II) reduction conversion
€
~
/
M
0
5.2...---...----r----....---....------, .....
~
() ..-.
3.88
3.87 1.0
0.0 0.2 0.4 0.6 0.8
R2.Cr
V
5.4
of lead. Figures 15 and 16 show that less
make-up feeds in both NH4Cl and NH3 are required
to adjust the composition of the aqueous phase
which acts as the leaching agent. The observed
decrease in NH4Cl is directly related to the
lower volumetric flow of aqueous phase while
for the decrease in NH3 a number of factors
contribute to explain the greater requirements
-
o 5.3 .--.-..... ---- -.---t-----i----j,----j
~ of acid in R3 namely, the lesser input of
J:
M
NH4+ from DIV3 due to a reduced recovery of
Z copper and an increase in the NH3 input from
MX3 due to impregnation.
Finally figure 17 shows the effect of the
5.2 L---'-_..l--""_-L._"----L_-'---'_-'----'
U.O 0.2 0.4 0.6 0.8 1.0 0.08
R2.Cr
U Saturation
:::l
'C
Figure 9 - NH3 consumption versus copper(II) ec. 0.06
reduction conversion c:
0
~
'E
Q)
2.644 0
c:
0
0.04
_ _ w_
-
0
Q)
1.
2.643 c:
'E
.__..._.. co
15 0.02 r-
L
2.642
Qj ~
E
c.
c. ~
0 ~
.-- ~ Operation
.L
~ 2.641 ()
...0 I
0.00
... _. __ ., .. I
is 2.640 U.04 0.06 0.08 0.10 0.12 0.14 0.16
J: '"
.........---
V Tl.dens...,pulp I kglkg H P
2.639
2.638 .. ~ ./
V Figure 11- Copper diamine concentration product
versus pulp density
2.637
U.O 0.2 0.4 0.6 0.8 1.0 100
R2.CR Sot<blizalion
'"
observed, while for lead a reverse trend is Recovery
found (figure 13). A denser pulp implies higher
concentrations of metals in the aqueous phase 80 ....-.-...-- r---....
"--r--
and hence greater losses by impregnation.
The separation of lead by precipitation in
the form of lead chloride crystals explains
the trend reversal observed for this metal
since phase saturation can be attained with 70
U.04 0.06 0.08 0.10 0.12 0.14 0.16
less residual lead in solution. In the case
of silver a similar effect could be expected Tl.dens...,pulp I kglkg H P
but, as already mentioned, by design the by-pass
for silver cementation is set at 10 percent Figure12-Silver solubilization and recovery
of the main stream, which implies a considerable yields versus pulp density
inventory of silver in the process which offsets
any gains similar to those found for lead, variation over the range 0 to 0.10 of the
due to greater losses by impregnation. fraction of the main stream at BY1, which
forms the by-pass stream for silver cementation.
Other significant effects are found which are It is found that even under conditions of
illustrated in figures 14 to 16. In figure non recovery the operating line is well below
505
1.38
"
102.5
"~ ~
1.36
-E
100.0 f----. ~
SoIlbilization M
0 1.34 --.-..
:::.
..
~
()
J:
Z
97.5
~
Recovlry ~
"
1.32
1.30
U.04 0.06 0.08 0.10 0.12 0.14 0.16
95.0 0.16
U.04 0.06 0.08 0.10 0.12 0.14
n.dens'pulp I kglkg H2 0
T1.densJlUlp I kglkg H P
Figure 15 - NH4Cl consumption versus pulp density
Figure 13 - Lead solubilization and recovery
yields versus pulp density
5.6
4.20
,
5.5
V
--
~
""
4.19 ;;-
/' .s::.
~
/ - 0
~
5.4 1--"---'
"" '"
-E
~
4.18 ...- ..- .
/ J :M
Z
- i/
/
N
0
~
5.3
c:
N
4.17
/ 5.2
0.04 0.06 0.08 0.10 0.12 0.14 0.16
T1.dens.pulp I kglkg H P
4.16 0.16
U.04 0.06 0.08 0.10 0.12 0.14
T1.dens'pu1p I kglkg H P Figure 16 - NH3 consumption versus pulp density
0.00
concentrates, under an EEC research contract. U.OO 0.02 0.04 0.08 0.08 0.10 0.12
By1.pdiv
The model is based on and takes advantage of the
features of the Speedup equation-oriented Figure 17 - Silver concentration in lead
simulation package and thus ensures the crystallization versus silver cementation
flexibil ity required for the gradual building-up by-pass stream expressed as a fraction of the
and exploration of the model. This is an main stream
506
essential aspect in the successful pursuit of ammonium chloride aqueous solution", Final
a common design objective by an interdisciplinary Report, CEC Contract MA1M-0045-C, June,
team, in particular when the complexity of the 1990.
process under consideration is great and requires
on-going experimental data validation and
integration, in mixed design and performance
modes.
The steady-state reference conditions for 20000
kg/h of Portuguese ore concentrate are presented
and the use of the model in investigating
prospect i ve changes in some of the des i gn
parameters is illustrated. These include the
extent of the copper reduction reaction, the
density of the pulp and the silver cementation
by-pass stream.
It is conc 1uded tha t wi th rega rd the forma t ion
of precipitates only copper diamine requires
consi derati on but for operati ng condi ti ons well
below design. In the lead crystallization reactor
which is considered the most critical point of
the process, because of the prevailing high metal
concentrations, following evaporation and
cooling, none of the three remaining metals can
be expected to form preci pitates. It is found,
in fact, that the extent of the copper reduction
reaction could be reviewed and the economics
associated with copper reduction conversion
closer to 0.60 assessed, because the need for
less leaching agent make-up is apparent together
with a reduction in CaO and process water
requirements.
The same applies to the assumed 0.10 split
fraction for the silver by-pass stream where an
increase in revenue associated with an increase
in this parameter should be balanced out against
the additional investment required by larger
equipment capacity for silver cementation.
The possibility of employing denser pulps should
be examined but only in conjunction with the
performance of the sil ver cementation by-pass,
because the decrease in copper and zinc recovery
yields can be expected to offset any gains in
make-up feeds, unless the trend in silver
recovery can be significantly reversed and
approach the one observed for lead.
ACKNOWLEDGMENT
This work was carried out under an EEC research
contract, CEC Contract MA1M-0045-C and i nvol ved
CENIM (Spain) and LNETI (Portugal) under the
coordination of Dr. J.L. Limpo (CENIM).
The authors wi sh to thank Drs. J. L. Limpo and
S. Amer (CENIM) and Eng. J.M. Figueiredo
(LNETI/DTM) for the advice and information
received which were essential for the mode11 ing
and simulation work.
REFERENCES
1. The Speedup User Manual, Prosys Technology
Ltd, Cambridge, England, 1988.
2. "Study of hydrometallurgical treatment of
complex sulphide ores by leaching with ferrous
chloride and oxygen in highly concentrated
Modelling in a Pachuca tank-flow
and mixing phenomena
J. A. Trilleros
M. Martinez
Department of Materials Science and Metallurgical Engineering, Complutense University,
Madrid, Spain
TABLE III
A) One Jet
TABLE II A. One Jet
L = 1100 L = 550
L = 1100 L = 550
a F'
1 F1 F'
g
Q1
d F' t F' t
g m p, m
28 0.86 0.72 0,86 1 ,51
I 0.47
0,18
0.56
0,41
0,50
0,19
0,92
0,60 28 0.86 1388 0.86 688
61 0,86 2,26 0,86 3.27
I 0.47
0.18
1760
2635
0.50
0.19
940
1448
I 0,50
0,19
L35
L32
0,50
0.19
2,59
1.29 61 0.86
0.50
443
551
0.86
0.50
262
354
84 0.90 3.37 0.90 4.02
I 0.19 752 0.19 692
I
0.54 2.89 0.54 2.83 84 0.90 303
0.18 1.79 0.21 1.52 0.90 255
0.54 370
I
125 0.86 5.20 0.86 3.77 0.54 383
0.18 585 0.21 613
I 0.54
0.21
4.30
2.35
0.54
0.21
2.77
1.39
125 0.86
0.54
235
255
0.86
0.54
246
354
250 2.78
0.86
8.43
3.41 I 0.21 440 0.21 638
L d FI L 1100 550
g tm
I
125
0.40
0.14
2.74
340
580
110 28 -0.41 -0.48
61 -0.35 -0.67
I
1.05 200 84
0.68 205 -0.40 -0.57
0.17 330 125 -0.49 -0.44
250 2.74 116 250 -0.59
500
I 0.97 210 -0.58
0.41 270 28 -0.53 -0.41 (1)
550 2.78 130 28 -0.61 -0.42 (2)
28 -0.57 ---- (3)
550
I
125
1.08
0.18
1.00
205
430
177 (1) Bundle 12 Pipes
I I 0.54
0.17
186
367 (2) Btmdle 6 Pipes
(3) Air Free Airlift
TABLE III
C) BLmdles
D = 28 I11III
n t FI tm
g
6
I
1100
0.50
0.19
0.90
200
370
410
I
550
0.43
0.15
0.94
568
780
320
I 0.43
0.18
440
645
TABLE III
D) Air Free Airlift
Jet FI tm
g
1 0.86 194
12
I 0.50
0.21
2.95
315
540
120
I 0.86 160
513
HOT ZONE
U.N~
ORIFICE H£T[R
DIFFERENTIAL ~NOMlTER
, ~NO/I£TER
6 S ROTAI'lETER
, JET
7 HOT LIQUID DISTRIBUTOR
• TANKS
FIG. 1.
,
~
'"
"- 11
en ""'" Ii
'",0
"-
""'" ..... IS ~.
'"0
I X
!
......
x & u. '" 21
u. '" .&
:,
.1 .2 .J .4 .S .1 •Z .J .4 .S
FL (KG/S)
FL (KG/s)
FIG. 2.
FIG. 3.
514
00
'"
o 28 MM. o = 28 MM.
~~
e
"
$~
\
.~
u
0
;:: '0
u
0 '0
J~
\~
I-
~
I
J4
I ~~
J.J
Jo ~~-
2.
,J /' - ~'r-=:'\,"'r
~ I
~
..
20
I
20 to
0 O. 01 12 24 '00 ' 00
"
T<S) X 10 3 T(s)
FIG. 4.
FIG. 5.
oiO
•• G ,
l-
'J"I, 0 I
'\
,
I- .;.
'0 .\ \
J.
).,.
'"'-
1\ '''', 2~
Jo I\t~
If N
I
l!i -...,/
r- IV
,I\!
,o,
I
lO I~
"
~ ~ 0 200 ' 00
T(S) r(5)
FIG. 6. FIG. 7 .
.. r - - - - - - - - - - - - - - - .
o = 125 Mr..
o = 84
"
N II MM.
550 MM .
"
II
I' L = 550 MM. u
L =
Fe, = 0,90 KG/H
• I~ Fe, = 0,90 KG/H 0
;::
~If \ .0
'.'I
u
'0 i~
J' I
o JO
I-
Z,
~.)
I I
IY ,I
" I
I ~.-,
,~+-------r_----~------_r------r_----~------~
"'" "'"
~oo
T<S) r(5)
FIG, 8.
FIG. 9.
515
.
~J
~
D = 125 i'IM.
L = llOO MM.
w
c
;::: ..
So
.......-1
o = 250 MM •
L = 1100 MM.
-0
Fci = 2,74 KG/H F'; = 2,74 KG/H
l'
12 JETS 12 JETS
""\
.,..,
1~
w
0 J"
..... •
le
I' \
y~
!-j
I'
-,
~i\;(~
\ " n
01
I \.•:...
'.
III • " 'I t '\
i .. ~ r ~~OU_I!:>'IO-----------
JV)'6'3VV ••
,t ' / \ 1,-
I I I '.r.' '.~",,-- ~
;'\1':' ~--------------
I I'
,. r
i I ; I ::'
f
I ;
I ,"
L
I I I~---,
00 "50 ,oe 200 leo
00 120 160 loe
T(S) T(S)
~. ,-------------------------------------------~
-, ~~----------------------------------~
-1
BUNDLE 12 PIPES I
I BUNDLE 12 PIPES
D = 28 M
-. L =
Fci
500
= 0,19
i'IM.
KG/H
I,
n
I
I
o = 28 i'IM.
L = 1100 MM.
w -0
I FG = 2 .63
, KG/H
o
1\
:h'I ,n
..... 1•
w
o
..... ~
&:_----..
II
III
,IV
I ,
~--..,
,
1 "
I I -4i'""ll"4.'l:.-_'!!"'_..
.,......
I
t AI • I
• I • ..L-J'I
I
I
o ~--~~--~==~----------------~
.• +---~----._--~----.---~----r_--_r----r_--~
o ",n
T(S)
TCs)
FIG. 12.
FIG, 13.
,.,-------r-----~------~~--~--~~~------~
.. ., BUNDLE 12 PI PES I
44 i D ::: 28~. L=llOO MM. BUNDLE 12 F IPES I
a J Fci= 0,86 KG/H V = II M' D = ; 8 MM . L=55C MM. i
:•• ~'
F';= C,86 KG/H,V= Q M3
..
:;' U I
... .,
w
"' -I o
.... ~
...
" ... I
:O J I
I.
... ...
"
II ~±==-±==L~---l---J ... ••• ...
( S. ) t ( s. )
FIG. 14.
FIG. 15.
516
"~------------------------------------,
.,
AIR FREE AIRLIFT ~ AIR FREE AIRLIFT
ONE JET, FG - 0,21 KG/H MJ 12 JETS, FG ~ 0,86 KG/H
•• I
00
I\
u
0
;::
<oj
o
"
I- JQ
~~
I
:'J V
~+----r---r--~--~----r---~--~--,---~
..,., '. n lOG -no
T(S) T(S)
FIG. 17.
FIG. 16
,
~,i------------------------------------~
~-'1 AIR FREE AIRLIFT 60-28 MM., L.1100 MM. , D-61 ~~ .• L-1100 MM.
'!oO I 12 JETS, FG - 2.95 KG/H o D-84 HM .• L-1100 HM • • D-125 HM .• L.1100 HM.
Ci 8.12 PIPES. L-llOO MI'I _ 8.6 PIPES. L.llOO MI'I.
<oj
o <0 \
• AIR FREE AIRLIFT
I-
:J IL
" OVUFCO--- ....:c
I
T( s)
FIG. 18.
FIG. 19.