Reference s
me tal rec over ies and concent rato r capacit ies fr om the West ern Grou p of
An on., 1987, Mining Ann ual Review, p. 419.
Cayro n, R., 1960, " Cont ribut io n a I'et ude du precede de f lottation de la malac hit e
apres sulf urat io n superf icielle appl iqu e aux minerai s ox ydes dolomi tiq ues du
Katanga ," Doc torate Thesi s, Universite Catho lique de Louvain, Bel gium.
de Cuype r, J., 1981, " Concent rat ion of co balt ores," Proc eedin gs, Vol. 1, Brussels
International Confe rence on Coba lt : Metallurgy and Uses, ATB Meta llu rgi e,
Brussels.
Evrard, L., 1971, " Contributio n a I' etude de la f lo tt abili te des ox yces hydrates du
co balt du gro upe de I' heterogenite. Ameliorat ion de leur flott at io n a I'aid e
d 'alkylhyd roxamates," Doc to rate Thesis, Un iversite Cath ol iq ue de Louva in,
Belgium.
Evrard, L., and de Cuype r, 1975, " Flotat ion of coppe r co balt oxi de ore s with
alk yl hydroxam ates, " Proceed ing s, 11t h Inte rnationa l Mineral Proce ssing ce o-
gress, Cagl iari , Ita ly .
Form anek, V., 1982, " Essais en usine pilote avec les reac tifs Hoec hs t nydrox-
amate s de haute purete," Unpubli sh ed Gec am ines Int ern al Repor t No. 2172 / E·
403.5 /RE, May 3.
Giesekke, E.W., and Harr is, P.J., 1984, " A st udy of the selec ti ve fl otati on from
ca lci te by the use of a si ngle bubbl e st ream mic rot lo tation ce ll ," Mintek 50,
Inte rnat iona l Con fe rence on Recent Ad vances in Mi neral Science and Tech no logy,
April.
Ignat ow, A., Yen, W.T., and Salm an, T., 1970, " The selec t ive f lot ati on of carro il ite
fr om chalco pyri te, born ite and chalc oc ite ," 9th CIM Conf ere nce of Met allurgists,
McM aster Universit y, Hamilton , On tari o, Canada, Augu st.
Kilia n, P.G., 1984, " Improvements in th e f lo tat io n of low grade fluorspar ores,"
Mint ek 50, Internat ional Con ference on Recent Advances in Mineral Scie nc e and
Tech nology, April.
Zinc-dust cementation of silver from
alkaline cyanide solutions analysis of
Merrill-Crowe plant data
J.R. Parga, R.Y. Wan, and J.D. Miller
Abstract - Th e direct recov ery of precious m etals
from alkaline c yanide solution is frequently ac com-
pli shed by th e well-established Merrill-Crowe pro cess
involving zinc -dus t cementation. Th is is particularly
true for sil v er. Plant data from two operations are
presented and analyzed in terms of intrinsic rea ction
f eatures. In such Merrill -Crow e plants a plate-and-
frame filt er pres s acts as a semicontinuous fi x ed-b ed
rea ctor wi th a distribution of particl e ret ention times.
For th e plants studied, 99 % of th e silv er is recov ered
wi th a solution ret ention tim e of less than 2 min.
Differen ces in plant operating conditions and product
cha racte ris tics can be related to th e nature of th e
surfac e deposit of sil v er which forms on th e zinc
part icles. From lab oratory expe ri men ts and plant
samples it has been f ound that th e pres en ce of lead in
th e s yste m activates a chan ge in th e silv er deposit
morphology to th e dendrit ic state which affects th e
rate of si lver depo sition as well as th e f ilt er ca ke
porosity. Finally th e plate -and-frame press is consid -
ere d f rom an en gin eering perspectiv e as a semi-
con tinuous f ix ed-bed rea ctor to describ e th e behav ior
of th e s ystem.
J.R. Parga, R.Y. Wan, and J.D. Miller, member SME, are with
the Department of Metallurgy and Metallurgical Engineering,
University of Utah, Salt Lake City, UT. M&MP paper 87-641.
Manuscript August 1987. Discussion of this paper must be
submitted, in duplicate, prior to Oct. 31, 1988.
170 AUGUST 1988
Lakef ield Research , " Inves t igation for t he imp rovem ent of co nce nt rate grade s,
conc entrators," Projec t No. L. R. 2550, Lakefi eid Research of Canada, Ltd.,
Lakefield, Ontario , Canada.
Mill er, J.D., and Mi sra, M., 1984, " The hyd rophobi c character of semi-solub le
salt minerals ," Min tek 50, Internat iona l Conf eren ce on Recent Advances in
Mineral Sc ience and Technol ogy, Apr il.
Ngongo, N., 1975, " Sur la si mil itude ent re res gi sements uraniferes (ty pe
Shi nko lo bwe) et le s gi sements cupriferes (typ e Kamolo ) au Shaba, Zaire,"
An nales de te Soci ete Geolog ique de Belg iqu e, Vo l. 98, pp. 449 · 463.
Palac he, C., Ber man, H., and Frod et, C., 1944, Dana 's System 01 Mine ralog y,
John Wiley and Sons , New York .
Rin ell i, G., and Marabi ni, A., 1982, " Flo tatio n of cobalt minerals with chelati ng
agent s as collectors," A TB Metall urgie, Vol. 22, p. 1.
Somasundaran, P.. 1984, " Surface inte ract ions in fin e particle flotat io n," Min tek
50, Inte rnati ona l Con feren c e on Recent Advance s in Mine ral Scie nce and
Technol ogy, Apr il.
Teoh, E.C., Law son, F., an d Han, K.N ., 1982, " Selec tive fl otatio n of co balt -bearing
mine rals with use of specific co lle ct or s," Transac tions, Inst itutio n of Mining &
Metall urgy, Vol. 91, Sec. C, Dec em ber.
Theys, L., 1947, " L'evolut ion de ra tec hnique de f lo tt at io n des minerals ox ydes
du Katanga ," Proceeding s, Congres du Cent ena ire de I'AILg, pp. 439 · 442.
Van Bell e, J., 1975, " Synthese des es sais de separati on chalcosl ne -car rolt ite
dan s res c oncent rateurs de Kam bove, Kolwe zi et Kamoto," Maadini, Bull. Inf.
Gecam ines , No. 5, pp. 23·36 .
Van Belle, J., and Iyol o, M., 1976, " Separati on cnatcos ine-ca rro llite : etu de
fond ament ale de f lottabili te, essais en cell ule Hall imo nd, es sais au banc pilo te
de Kambove," Maadini , Bull . lnt . Geca mines , No. 11, pp. 2- 13.
Whi te, L., 1978, " Cipec's big fou r: Zaire," Engineeri ng and Mining Journal,
November, p. 188.
Introduction
The recovery of precious metals from dilute alkaline
cyanide solution by zin c-dust cementation has been
practiced since the late 19th centur y . The technology
was patented in 1894 by Sulman and Pichard and wa s
first a pplied at the Deloro mine in Ontario, Canada
(Lebla nc, 1942). C.W. Merrill improved the process
a nd used this process strategy at the Homestake mine
in 1897. Later, in 1916, T.B. Crowe's vacuum deaerator
was in corporated to remove oxygen from cyanide
leach soluti on to signific a ntly improve the efficiency
of th e precipitation reaction (Lebla nc , 1942).
Considerable research has been reported on the
cha r a cter isti c features and mechanistic details of
cementation reactions, and this literature has been
r eviewed in terms of electrochemical theory and
transport phenomena on several occasions (Wa ds-
worth, 1986; Miller, 1981; Strickland and Lawson,
1971). In most cases, the cementation reactions follow
first-order reaction kinetics and generally are limited
by diffusion of the noble metal ion through the mass-
transfer boundary layer. The structure and morphology
of the reaction product have a significant effect on
reaction kinetics, either enhancing the rate in some
cases or passivating the surface in other cases.
Only modest attention has been paid to the silver
cementation reaction in alkaline cyanide solution.
Von Hahn and Ingraham (1968) studied the kinetics of
silver cementation on a rotating zinc cylinder in
MINERA LS AND METALLURGI CAL PROCESSING
deoxygenated alkaline silver cyanide solution. They 10 · ,.---r---.,---,-----,---r--.,.--~r-___,
found that the rates were dependent on peripheral
velocity and the cementation rates obeyed a pseudo -
first-order rate at the beginning of the experiments.
The deposits were dense and adherent and had little
or no influence on the reaction velocity constant
except at very low free-cyanide concentration. Ritchie
and Staunton (1984) reported their results on the
electrochemical study and electron-microscopic fea-
tures of the silver/copper cementation reaction in
cyanide solutions. In this case, the silver deposit
u
appeared to be clustered around anodic areas on the
copper surface. z
Most recently research in our laboratory has provid-
ed further information regarding details of the intrin-
.....
~
II:
10 I

9.3xl0-4 M
sic reaction kinetics for the Ag " / Zn (CN - ) system, and Z
w 10-4 M
these results are discussed in another publication U
Z 10.5
(P a r ga , et al., 1987). In this contribution, the plant o
u
data on silver cementation from alkaline cyanide III
solution by zinc dust are analyzed in terms of the III
w
intrinsic reaction kinetics and the nature of the surface -'
z
deposit. Further, an engineering analysis of the plate- 2
III
and-frame press as a semicontinuous fixed-bed reactor Z
w
is presented in order to describe the operating charac- :I
teristics of the system . The Escalante plant operations o
in Utah and the Silver Peak plant operations in Nevada Zinc Dust
\\
are considered. mesh 270x400 2oox270 100x150 70x loo

Characteristic features of the

cementation reaction
o 5 10 15 20 25 30 35
Before evaluating the plant data, it is useful to
review recent findings regarding the intrinsic reaction TIME . m in

kinetics . The recovery of metallic silver by cementa-

tion with zinc dust is an electrochemical process Fig. 1 - Flrstorder react ion kinetics plot for si lver cement a-
involving the oxidation of zinc and the reduction of tion from alkaline cyan ide solut ion . Initial silver co nce ntr ation
silver cyanoanions. The overall stoichiometry for the 100 ppm .
reaction is as follows :
2
2Ag(CN) 2 + Zn -+ 2Ag +- Zn(CN) i (1)

The cementation reaction kinetics we re studied by

batch reaction of suspended zinc dust in an oxygen.
free environment «0.5 ppm 02) and generally the
results were found to conform to the first-order rate
expression to over 99% silver removal as shown in
Fig. 1. Further, these rate data have the expected
inverse first-order dependence on particle size for a
mass-transfer, boundary-layer, diffusion- controlled
reaction. The experimental reaction velocity constants
are slightly smaller than the expected mass-transfer
coefficients as shown in Table 1.
The effect of the silver deposition (0.231 mg/cm") on
the rate of cementation was not significant. Subse-
quent examination of the cementation reaction product
showed that the silver formed a uniform layer around
the zinc (see Fig. 2), and it is expected that this layer
has sufficient porosity to account for the observed
boundary-layer, diffusion-controlled, first-order reac-
tion kinetics . Fig . 2 - SEM photom icrographs of sectioned zinc particles
with cemented silver deposited from alkaline cyanide solution .
Table 1 - Comparison of Experimental React ion Velocity
Constants and Calculated Mass-Transfer Coeff icients
for Suspended Particles Further, in this work (Parga, et aI., 1987), th e
structure of the silver deposit was found to be sensitive
Particle Size, Experimental Reaction Calculated Mass ·Transter
Coefficient. em/sec
to small additions of other metal ions. lead in partic-
mesh Veloci ty Constant, em/sec
2 ular. Under these circurnstances . a dendritic silver
100 x 150 1.88 x 10- 2 2.2 x 10-
150 x 200 1.76 x 10- 2 2.4 x 10- 2 deposit can for:n (depending on th e conc entr a tion of
200 x 270 1.63 x 10- 2 2.2 x 10- 2 lead) as shown in Fig. 3, and the r eaction kinetics
2
270 x 400 1.62 x 10- 2 3.2 x 10- exhibit a slightly enhanced rate region but signifi cant-
Source : Parg a, et aI., 1987.
ly less than that observed in other cementation reac-
Ag = 100 ppm (9 x 10- 4 M ), NaCN = 10- 2 M, pH = 10.5. tions studied with rotating-disk geom etry . Thi s rather

MINERAL S AND METALLURGICAL PROCESSING AUGUST 1988 111

 small, almost insignificant, increase in rate with Plant operations
dendritic formation undoubtedly reflects the relatively
low slip velocity for suspended particles when com- The extraordinary efficiency of the zinc-dust precip-
pared to rotating-disk systems. The change in surface itation reaction for silver recovery requires that the
character is vividly illustrated by comparison of the pregnant leach solution first be clarified and deaerated
SEM photomicrographs of sectioned particles shown before reaction initiation. Deaeration decreases the
in Fig. 3. tendency of the silver to redissolve by chemical reac-
In addition to these characteristic intrinsic reaction tion with oxygen from the atmosphere . A typical
features, it has also been found (Parga , et al ., 1987) flowsheet for silver recovery by ce m e nta ti on is shown
that solution chemistry variables , particularly cyanide in Fig. 4. The pregnant leach solution is clarified by
concentration, have a significant effect on the reaction filtration . The clarified solution is deoxygenated in a
kinetics. vacuum tower. In the plant, the zinc dust (99.9% zinc,
spherical in shape, and 97% - 45 Jim) is added to a
small mixing cone, through which an auxiliary stream
1.0 ~------------------_--, of barren cyanide solution passes, carrying the zinc
dust into the feed pipe. The cementation reaction
starts at this point and is completed as the zinc
particles are collected and held in the filter press. For
continuous operation, a process will have three filter
presses in parallel so that one filter can be cleaned
while the other two are in operation, typically on a
seven-day cycle. When a cycle is completed, the cake
is blown dry with air, the press is opened, and most of
=
u the cake readily falls into a wheeled tray placed
z 0 ,2
beneath the press. The remainder of the cementation
o cake is removed by scrapers. Normally, the filter cake
....C( is soft but sometimes becomes hard as a result of
<C
.... oxidation and the presence of calcium ca r b ona te.
Z
UJ It is evident that the cementation reaction occurs
U
Z
0 .1 both in the pipe and in th e plate-and-frame filter press.
o The filter press acts as a semicontinuous fixed-bed
o 0
Ul
reactor with accumulation of zinc dust and cemented
Ul
UJ 0 2
silver during the plant cy cle of approximately one
..J
week. Of course, most of the time the zinc particles are
~ 0 .05 6 10 stationary in the press as they react. This situation
0 50
Ul
Z makes detailed analysis of the system somewhat
UJ
:I
difficult because each zinc particle has a different
o retention time in the reactor and mass-transfer coeffi-
AgICN I; 8010 '· M
5 010 '2 M
cients for flow through such a fixed bed (the ca ke ) are
NaCN
pH 10 .5
not well-established.
0.02
A typical plate-and-frame press is shown in Fig. 5.
The pregnant solution with the zinc dust is pumped
2 4 6 under pressure into the filter cavities . Typically
r rus , min diatomaceous earth is used as a pre coat for the
chambers and the zinc dust particles covered with
Fig. 3 - First-order reaction kinetics plot for s ilver cementa- silver are captured at this surface or by the cake while
tion with different concentrations of dissolved lead. Initial permitting the clarified barren solution to pass through.
silver concentration 85 ppm. Comparison of the nature of the Filtration generally proceeds until the chambers are
surface deposit for sectioned particles in the presence and filled with the cementation product. By releasing and
absence of lead. shifti ng the plates, the cementation reaction product
is removed as cake, and the filter is ready for reuse .

LEACH
SOLUTION
VACUUM
PUMP RECYCLE

r-----{l
TO THE
LEACHING
CIRCUIT
CYANIDE
SOLUTION
ZINC OUST

BARREN
SOLUTION
TANK

PRECIPITATE
L..- TO REVERBERATORY
FURNACE
Fig. 4 - Typical Merrill -Crowe flowsheet for silver cementation.

172 AUGUST 1988 MINERALS AND METALLURGICAL PROCESSING

 No. 1 in the reactor, while those added just prior to the
breaking of the press spend only minutes in the
FRAME No. 26 reactor. This situation leads to some interesting
characteristics of the filter cake and provides for a
No. 52 challenging engineering analysis.
Another important observation in comparing the
data for the two plants is the large difference in
zinc-dust consumption and cyanide-solution concen-
tration. The zinc consumption at Silver Peak is about
1.85 Zn / Ag weight ratio, 3.7 times the consumption at
Escalante, which is only 0.5 (the stoichiometric weight
ratio is 0.3). Also , note that the cyanide concentration
at Silver Peak is 1.5 times higher than that at the
Escalante plant. These results suggest that the zinc
consumption may be related to the cyanide and oxygen
level in the system.
Fig. 5 - Typical plate-and -frame press, a semicontinuous
fixed-bed reactor for cementation .
Characterization of cementation product
Data were collected from two Merrill-Crowe plants
to evaluate the plant performance and to determine Examination of samples of the filter cake from the
the characteristic features of the cementation reaction Escalante plant shows another interesting feature of
product. The two Merrill-Crowe plants examined were particles from the cementation cake with respect to
Escalante Silver mine, Redco Silver Inc., Enterprise, frame number and location in the frame. The composi-
UT, and Silver Peak operations, Sunshine Mining and tion of samples for the silver cake from frames 1, 26,
Minerals, Silver Peak, NV. Both plants have a Merrill- and 52 are listed in Table 3 and the ca ke profile for
Crowe process similar to that described previously. each frame /sample location are displayed in Fig. 6.
The silver cementation reaction begins as the particles As shown in Fig. 6, the first frame was found to be
are transported through a pipe leading to the filter about 90% full, and the filling decreased with frame
press. The plate-and-frame filter press collects the number away from the front of the press with the last
particles and acts as a semicontinuous fixed-bed frame having about 20% volume filling by the silver
reactor. The operating conditions for zinc-dust cemen- cake. From the data in Table 3 and Fig. 6, it seems that
tation at the two plants are presented in Table 2. there is only a small variance in average composition
from frame to frame . Specifically the silver analysis
systematically decreases from the first frame (No. 1 )
Table 2 - Plant Data for Zinc-Dust Cementation at Different
to the last frame (No . 52); further, it is evident that the
Merrill -Crowe Plants
copper content in the last frame is significantly
greater than that found in the other two frames , In
Escalante Silver Peak each frame the middle part of the cake contains more
Solution impurities with less silver compared to the upper and
Flow rate , gpm 330 265 bottom portions of the cake. This is because the fresh
Ag, pregnant, oz per ton 2.87 2.4
Ag , barren , oz per ton 0.02 0.02 pregnant leach solution mixed with zinc dust enters
NaCN, Ib per ton 6.0 10.0 from the top on one side of the frame . The reacting
Solution Retention Time zinc particles become sediment and simultaneously
Pipe , sec 7.0 12
Filter press , sec 112 168
collect on the plates, creating the profiles shown in
Zinc Dust
Fig. 6. Of course , the particles in the middle are those
Zn/Ag by w eight 0.5 1.85 that have been in the press the shortest period of time
Zn /Ag by stoichiometry 0.82 2.86 and hence have not reacted to the same extent as those
Filter Press at the top and bottom.
Number of units 3 3
Units per cy c le 2 2 In addition to determining the composition of the
Number of frame s per unit 52 45 silver cake at various locations for each frame, the
product was examined with respect to surface -deposit
characteristics as shown in Fig. 7, which presents
It is evident from Table 2 that the retention time of SEM photomicrographs of the silver deposit on zinc
the solution in the press is approximately 14 times dust particles for the Escalante plant, frames 1, 26,
greater than the retention time in the pipe. As can be and 52. The structure of the silver deposit in each
imagined from the flowsheet presented in Fig. 4, some frame was found to have a distinct dendritic character
reaction may occur passing through the pipe . Analysis that is expected to increase the cathodic surface area
of samples taken from the entry to the filter press at and enhance the cementation reaction rate . This
the Escalante plant indicates that 10% of the silver deposit morphology is similar to that observed in the
reacted in the pipe. The remaining 90% Of the silver is laboratory when lead was added to the cementation
removed in the filter press as the solution passes system . This lead effect may, in fact, account for the
through the cake, giving an overall recovery in excess dendritic growth. Note from Table 3 that a substantial
of 99% for a total solution retention time of approx- amount of lead, typically 3%, is found in the Escalante
imately 2 min. Based on laboratory data (Parga, et al ., cake . This natural lead content in Escalante's preg-
1987), it is expected that such a retention time is nant solution ( ± 0.8 ppm Pb) is probably responsible
sufficient to remove over 99% of the silver as was for the dendrltlc morphology .of the silver deposit . For
observed in both plants. Of course, for this semi- such a particle structure, a faster reaction rate should
continuous fixed-bed reactor, the retention time of the be realized, and the pressure drop across the press
reacting zinc dust varies significantly. Zinc -dust should be minimized because of increased cake
particles that enter early in the cycle spend seven days porosity.

MINERALS AND METALLURGICAL PROCESSING AUGUST 1988 173

 CHEMICAL COMPOSITION ANALYSIS ACROSS THE CAKE

-_
%- % %
Ag Zn Ag Zn Ag Zn
TOP __ 84 0 .6 _ 82 .0 0 .9 _ _ 80.1 1.1 _ _

MIDDLE x 81 .2 0 .9
73 .7 1.5

69.5 1.8
BOTTOM __ --J( 87.5 0 .5_ _ ___ 80 .8 0 .8 _ _

FRAME 1 FRAME 26 FRAME 52

Fig. 6 - Cake composition for various locations in frames 1, 26, and 52 for the plate-and -
frame press at the Escalante operation.

The mass -transfer coefficient for reaction in a packed-

Table 3 - Chemical Composition of Silver Cake Samples bed reactor can be calculated from the following
from the Escalante Plant (see Fig. 6 for Samp le Location)
equation (Hines and Maddix, 1985):

Frame Position Ag Cu
%
Zn Fe Pb
J
d
= k*(N sc ) 2/ 3 y - l (2)
To p 84.0 2.2 0.6 0.1 3.0
Mi ddl e 81.2 3.0 0.9 0.1 3.4 where k* is the packed-bed mass-transfer coefficient ,
Bottom 89.0 0.3 0.4 0.1 3.0
N Sc is the Schmidt number, Y is the velocity of the
26 Top
Middle
82.0
73.7
2.2
6.8
0.9
1.5
0.1
0.2
3.7
42
flow, and J d is the mass-transfer flux in the packed
Bottom 87.5 0.5 0.5 0.2 2.7 bed.
52 Top 80.1 3.1 1.1 0.1 2.4 The flux for a packed bed can be estimated for very
Middle 69.5 12.4 1.8 0.1 3.5 low Reynolds ' numbers according to the equation
Bottom 80.8 3.3 0.8 0.1 2.3
(Hines and Maddix, 1985),

EJd = 1.09 N Re- 2/ 3 (3)

The particle surface morphologies and the chemical where E is the porosity of the packed bed, and N Re is
compositions of silver-cake samples from the Escalante the Reynolds' number.
and Silver Peak operations are compared in Fig. 8. In this regard, from Eqs. 2 and 3, the packed-bed
From the SEM photomicrographs, it is evident that mass-transfer coefficient for the Escalante operation
the surface deposit at Silver Peak is smooth and can be estimated to be 3.5 x 10- 2 ern /sec. or 2.07
produces a dense layer of silver on the zinc dust cm/rntn (Parga, 1987).
particles, which is different from the dendritic product Now , consider a particular cycle of the cementation
formed in the Escalante plant. The retention time for operation. The mass balance for silver at any partic-
the leach solution in the filter press at Silver Peak is ular time in this semicontinuous fixed-bed reactor
1.5 times greater than the retention time for Escalante (filter press) can be expressed by the following
(see Table 2) , and the surface deposit morphology general expression for a first-order reaction,
may account for the longer retention time due both to
its effect on the electrochemical reaction kinetics and
its effect on the cake porosity. _ Q dC - A k ' C =A dC (4)
dx dt

where Q is the flow rate, cmv /rntn: A is the cross-

Engineering analysis sectional area of the reactor, em": C is the concentra-
tion of Ag(CN)i , mol/ern": and k / is the apparent
In Merrill -Crowe plants, the plate-and-frame filter rate constant, m in - 1 .
press acts as a semicontinuous fixed-bed reactor with If the concentration gradient of the Ag(CN) i is
accumulation of zinc dust and cemented silver. It is considered to be a linear relationship with the trans-
expected that, during the initial stage of the cycle, the port distance x into the filter cake , then
reacting zinc particles are suspended in the silver
cyanide solution, but most of the time the particles are
stationary in the filter press as the packed bed grows. dC
= (5)
Study of the intrinsic reaction kinetics for the dx
Ag/Zn(CN -) cementation system confirms that the
cementation reaction follows diffusion-controlled first- where O, is the inlet concentration of silver, Co is the
order reaction kinetics. The experimental mass- outlet concentration of silver, and d is the filter-cake
transfer coefficient for this cementation reaction with thickness at time t.
suspended zinc-dust particles is estimated to be 1.6 x Thus, the concentration of Ag(CN)i ' can be solved
10- 2 ern/sec (or 0.9 ern/min) as indicated in Table 1. as a function of time :

174 AUGUST 1988 MINERALS AND METALLURGICAL PROCESSING

 QC j C jk 'Ad Q By combining Eqs. 6 and 7. the concentration of
C (t) = + exp ( - [ - + k '] t) (6) silver during the reaction period ca n be expressed as
o Q + k ' Ad Q + k:' Ad Ad
a function of time,
The distance d can be calculated fr om the packed-bed
volume per unit cross-sectional area, +
1 + --:--- - ' - -
mt
d = (7)
Cj
+ kt I) (8 )
where m is the zinc -dust m a ss flow rate, g /mm : QQZn (l - E)
Q Zn is the dens it y of zinc dust, g/ cm 3; A is the reactor 1 -
k 'mt
cross-sectional area, em>: and E is the porosity of the
packed bed.

FRAME 1

ESCALANTE
SILVER PEAK PLANT
0/0
PLANT Zn Cu Pb
All
%
All Zn Cu Pb 86.3 0 .6 0 .8 2 .2
82. 1 0 .8 1. 1 2 .2
95.2 0 .6 0.1 0 .1
60.5 1.7 12.3 4 .9
94.9 0 .7 0 .1 0 .4
94.5 0 .9 0 .1 0 .1

Fig . 8 - Comparison of the surface morphologies and the

chemical compos ition of the silver cementation product for
t he Silver Peak and Escalante plants.

Using this approach, data from the plant operations

Fig . 7 - SEM photomicrographs of t he si lver deposit on zinc
dust for the Escalante filter press.
was exa m ined to determine if the engineering analysis
could be confirmed. The parameters such as filter-
cake porosity (E = 0.60) and zinc-dust density (Q ZII =
5 g/ cm 3) for the Escalante plant were measured in our
laboratory. The average zinc-dust particle size of th e
Escalante plant is 8 x 10- 4 ern , and the specific surface
area of the zinc dust as measured by air permeametry
was found to be 867 cm-zg. Generally, the second term
of the right side of Eq. 8 can be neglected. since
QQZn(l - E) /m »1. A plot of silver concentration vs .
reaction time for Escalante plant data is compared in
Fig. 9 with the predicted response according to Eq . 8.
The flow rate of solution to this semicontinuous fixed -
bed reactor is an important variable which has the
greatest influence on the performance of the system.
The data analysis suggests that there are three
periods during the operation of th e pJate -and -frarne
filter press :

MINERAL S AND M ETALLURGI CAL PROCESSING AUGUS T I YH8 17',

 • Induc ti on Period - Suspe ns ion of zinc-dust parti - were colle cted from two Merrill-Crowe pl ants: the
cles flows into and fill s up th e filter-press chambers . E scalante Silver m ine , Redco Silver Inc., Enterprise ,
• Sus pe nded-P a r ticle R eaction P eriod - Reaction UT, a nd the Silve r P eak operations , Sunshine Mining
continues wit h suspended zin c-du st particles. and Minerals, NV, and in clude operational details,
• Packed-Bed P eriod - R eaction with stationary plant performance , retention-time calculations, and
zin c-dust particles in th e pack ed bed of the filter press. product filter -cake characterization. The presence of
It is ev ide nt that the eng ineering analysis adequately lead in th e sy stem has a marked effec t on the silver
ex plains the plant data. Though the filter pre ss a cts a s de posit m orphology as evidenced by com pa r is on of th e
a semicontinuous fixed-bed r eactor with accum ul ation dendritic product from th e E scalante Silver mine with
of zin c du st and cem ented silver during th e plant cycle the product from the Silver P ea k operation.
of a ppr ox im a tely se ven days , the periods of indu ction In addition, engineering ana lysis of the plate -and-
time a nd sus pended zinc particle r eaction account for frame filter press operation as a sem icontinuous fixed-
onl y 20 min of the total cycle time. The cem enta ti on bed reactor adequately expla ins the plant data ob-
r eaction for silver r ecovery r eaches 95% after onl y tained during start-up of a particular cycle. •
1 hr of operation. It is evident th at th e plate-and-frame
filter press is a very effecti ve r eactor for the zin c-dust Acknowledgment
cementation process.
This research activity has been supported in part by
funds from National Science Foundation Grant No .
CBT 8512522 and by funds from the Mineral Leasing
Pack ed Bed Fund, University of Utah. Appreciation is extended to
100 Y. Ye for his help in the engineering analysis of plant
E
0.
I I data. Our grateful acknowledgment is made to th e US
q I Bureau of Mines Salt Lake Research Center and
u 80 ~81
"\6
I
I A. King for help in the SEM studies .
....- I 0 I The cooperation of personnel at the Escalante and
W 6
......J::::> I I Silver Peak plants is appreciated, particularly a ssis-
0 60 I I Escal ante Plant Data, Jan 26 ,1987
tance and efforts of D . Va ll Dyk, B. Chase, S. Hartman,
I I 6 6 Fil ter Press No.2
~
z I o Filler Press No. 3
a nd B. Huff.
0 o
~
a: 40 - Pr edi ct ed References
....Z
w
o Hines, A.L. , and Madd ix, R.N ., 1985 , Ma ss Transfer Fundamenta ls and App lic a·
Z
0 tions, Prentice -Ha ll, Inc., Englewood Cliffs, NJ, p. 195.
o 20
a: Leblanc, R., 1942 , " Prec ipitat ion of Gold f rom Cyanide Soluti on by Zin c Dust ,"
w Canad ian Mi ning Jo urnal, April, May, Jun e, pp . 213 ·219, 297-306 , 371 ·379.
:'j
iii Mill er, J.D ., 1981 , " So lut ion Con centration and Purifi cation," Mefallurgical
0 Treet i ses, J.R. Tien and J. F. Ell iott , eds ., AIME, New Yor k, pp. 95 -116 .
0 10 70
Parga, J.R., 1987, " Analysis of the Zinc Cementat io n Reac t ion fo r Rec overy of
TIME , minut es
Precious Metal s fro m Cyani de Solutions ," Ph.D. Dissert at ion, Dep t. of Met allurg y
and Met all urgical Engin eering, Un iversity of Utah, Sail Lake City, April.
Fig . 9 - Plot of silver concentration VS. react ion time fo r Parga, J.R., Wan, R.Y., and Miller, J.D., 1987, " Intrinsic React ion Ki net ic Features
Escalante plate -and -frame press data dur ing cycle start-up in Zinc -Dust Cementa tion of Preciou s Metals from Alkal ine Cyanide Solu ti on,"
SME Annual Meetin g, Denver, CO, March .
compared with the predicted behavior.
Rit ch ie, I.M ., and Sta unton, W. P., 1984, " Elec t roc hemic al Invest igations and
Electro n Microsc op y of Sil ver (1)·Copper Cement ati on Reac ti on ," Pro c eed ing s,
Inte rnationa l Symp osium on Elect roch em ist ry in Min eral and Met al Processing,
Summary P.E. Rich ardson, S. Srin ivasan, and R. Wood s, eds ., Th e Elec troche mical Societ y,
Penn ingt on , NJ , pp. 486·500.
The zinc-dust cementation r eaction of silver from St rick land, P.H ., and Laws on , F., 1971 , " The Cementat ion of Metal s f rom Dilute
Aqueo us Solu ti on s," Pro ceeding s, Austr alasi an Inst itute of Min ing & Metallurgy,
alkaline cya nide solution is controlled by mass-transfer Melbourn e, No. 237, pp. 71 · 78.
processes as has been established in laboratory inve s- Von Hahn, E.A., and Ingra ham, T.R., 1968, "Ki net ics of Sil ver Cementat io n on
tigations. Plant data for silver recovery from cyanide Zi nc in Alkal in e Cyan id e Solutio ns and Perc hlor ic Acid Sol utions," Canadia n
Me fall urgic al Quar terl y, Vol. 7, pp. 15 · 26.
soluti on by the zinc-dust cementation process using a
Wadsworth , M.E., 1986 , " Prec ipitat io n Reac t ions in Hydrom etallu rg ical System s,"
plate-and-frame filter press a s a se micontinuous fixed- Ad van ces in M ineral Processing, P. Somasundaran, ed., Society of Min ing
bed r eactor have been pres ented a nd a na lyze d. Data Engineers, Inc., Littleton , CO, pp . 487·492.

176 AUGUST 1988 MINERALS AND METALLUR GICAL PROCESSING