Journal of Environmental Chemical Engineering 4 (2016) 3283–3292

Contents lists available at ScienceDirect

Journal of Environmental Chemical Engineering

journal homepage: www.elsevier.com/locate/jece

Silver recovery from simulated photographic baths by electrochemical

deposition avoiding Ag2S formation
P.F.A. Prado, L.A.M. Ruotolo*
Department of Chemical Engineering, Federal University of São Carlos, Rod. Washington Luiz km 235, 13565-905 São Carlos, SP, Brazil

A R T I C L E I N F O A B S T R A C T

Article history:
Received 13 May 2016 Aiming at recovering silver from a simulated photographic wastewater, a study on electrochemical
Received in revised form 28 June 2016 deposition with modulated current was developed. In cathodic potentials where hydrogen formation
Accepted 30 June 2016 occurs, Ag2S formation was noticed over the electrode due to parasitic reactions. Therefore, in a
Available online 1 July 2016 galvanostatic process, the electrodeposition would only be feasible if conducted under current control to
avoid Ag2S formation. An electrochemical flow reactor was used for silver electrodeposition under
Keywords: current control using a stainless steel plate. The modulated current values for the electrodeposition were
Electrochemical reactor calculated according to the limiting current kinetic control, applying the mass transfer coefficient
Electrodeposition
calculated from experimental values of diffusion coefficient of silver ions and diffusion layer thickness.
Limiting current
Silver electrodeposition was achieved at high efficiency rates, low energy consumption and without the
Modulated current
Silver presence of parasitic reactions. The highly negative potentials were avoided when the limiting current
was applied and, consequently, sulfur sulfide was not formed over silver deposit. The modulated current
electrodeposition of silver revealed to be a feasible process to recover silver from spent photographic
baths.
ã 2016 Elsevier Ltd. All rights reserved.

1. Introduction obtained, the fixer may be recycled, and silver concentration in

Disposal of toxic materials from industrial wastewaters or any
other sort of waste generated by human activities has grown
exponentially in the last decades, drawing attention due to their
impact on the environment and human settlements around these
disposal sites. Severe legislation has been adopted in most
countries regulating, monitoring and determining legal limits on
concentration of these residues, as carbon monoxide emissions by
cars, metallic ions from processing facilities and organic com-
pounds generated by chemical plants.
Photographic, X-ray and cinema processing wastewaters
constitute the main source of effluents containing silver. In these
effluents, silver halide weights from 30% to 40% of the total solution
weight, thus making clear that the removal of these species is
important not only because of environmental concerns, but mainly
due to economic reasons [1]. Although many techniques have been
tested for silver recovery, such as cementation [2], adsorption
using cationic resins or activated carbon [3], and precipitation [4],
electrolysis has been preferred because high-purity silver is
* Corresponding author.
E-mail addresses: pluis@ufscar.br, luis_ruotolo@hotmail.com (L.A.M. Ruotolo).
the wastewater can be considerably reduced [5]. Recently,
Condomitti et al. [6] reported a technique combining magnetically
coated carbon particles and electrodeposition to recover silver
from dilute solutions.
Since the primary reagent is the electron, the electrochemical
technology is even more desirable nowadays due to its environ-
mental compatibility. Furthermore, it is also very economically
competitive, presenting a low cost solution besides both easy
operation and maintenance procedures. However, the application
of electrolysis for silver recovery faces some drawbacks regarding
to the low current densities and its unsatisfactory operation for
solutions with concentrations less than 200–500 ppm silver due to
parasitic electrochemical reactions, leading to precipitation of
silver sulfide [7].
Since 1837, the photographic process using fixing solutions
based on silver halide (AgX) dissolution in ammonium or sodium
thiosulfate solutions has been used by the photograph and X-ray
industry [8]. The AgX dissolution in thiosulfates and thiocyanates
is fast and no catalysts are necessary, being thermodynamically
more stable in slight alkaline pH. The simplified mechanism of
AgCl dissolution in sodium thiosulfate is shown in Eq. (1) [3].


AgClðsÞ þ 2Na2 S2 O3ðaqÞ @Na3 AgðS2 O3 Þ2 ðaqÞ þ NaClðaqÞ ð1Þ

http://dx.doi.org/10.1016/j.jece.2016.06.035
2213-3437/ã 2016 Elsevier Ltd. All rights reserved.
3284 P.F.A. Prado, L.A.M. Ruotolo / Journal of Environmental Chemical Engineering 4 (2016) 3283–3292

to the irregular current and potential distribution inside the porous

Nomenclature
matrix [15,16].
In this work silver electrodeposition from a simulated sodium
A Membrane area [L2] thiosulfate solution was investigated. In a first attempt, a three-
C Mass concentration [M L3] dimensional cathode (reticulated vitreous carbon) was tested to
Cb Molar concentration in bulk of solution [N L3] carry out the electrodeposition, but the formation of Ag2S and the
Ci Molar concentration of “i” [N L3] difficulty of potential control during the electrolysis stimulated the
Cs Molar concentration on surface [N L3] search for new strategies to deposit silver and avoid Ag2S
D, DAB Diffusion coefficient [L2 t1] formation: (i) an attempt to remove sulfur compounds by reducing
Ecell Cell potential [V] the solution pH was tested and (ii) a potentiodynamic mode to
I Current [I] supply a modulated current (maintaining the limiting current) was
i Current density [I L2] successfully tested.
ilim Limiting current density [I L2]
km Mass transport coefficient [L t1] 2. Experimental
L Membrane length [L]
M Molecular weight [M N1] 2.1. Electrolyte preparation
ND Diffusive flux within boundary layer [N t1 M1]
Vi Volume “i” [L3] To mimic the real conditions found in a photographic
x One-dimensional length [L] wastewater, the solutions used in the voltammetric character-
z Ion charge [I t] izations and for silver electrodeposition were prepared according
b Cell constant [L2] to the silver concentration found in real wastewater. In this case, a
d Boundary layer thickness [L] wastewater sample from a photographic store was analyzed by
e Current efficiency [–] flame atomic absorption spectrophotometry and the average
h Energy consumption [M T2 L1] concentration of 800 mg L1 was achieved, along with a 7.3 pH and
s Conductivity [t3 I2 M1 L3] 35.8 mS cm1 conductivity. This wastewater was a mixture of the
spent photography fixer and the wash water. The average
concentration of sodium thiosulfate in this mixture was estimated
from literature. The value used in this work was 8.75 g L1 [17,18].
The fixing baths also contain a large number of chemicals such The simulated wastewater was prepared by dissolving silver
as acids, preservatives, hardeners, and buffers. Therefore, the chloride in 8.75 g L1 Na2S2O3 solution. Sulfuric acid and sodium
aqueous waste resulting from the photographic process is a hydroxide were used to control the pH. The conductivity was
complex solution, which typical composition was provided by adjusted adding sodium sulfate as supporting electrolyte, when
Adani et al. [3]. The main components of the sodium sulfate necessary. All reagents were of analytical grade and deionized
solutions are Ag+ (2050 mg L1), Na+ (3050 mg L1), and K+ water was used to prepare the solutions.
(3500 mg L1). The silver electrodeposition reaction [9] shown in
Eq. (2) will compete with thiosulfate reduction (Eq. (3)) and 2.2. Voltammetric and chronopotentiometric analyses
hydrogen evolution reaction (Eq. (4)), which constitute the main
parasitic reactions occurring during silver electrodeposition. Cyclic voltammetries were performed in a typical three-
electrode cell using 1.0 cm2 and 1.5 cm2 Pt foils as working and
 þ 
AgðS2 O3 Þ3
2 þ e þ 2H ! 2HS2 O3 þ Ag ð2Þ counter-electrode, respectively. Ag/AgCl 3.0 M KCl was used as
reference electrode. The electrolytes used for voltammetry
contained Ag+ (from AgCl) in concentrations ranging from 10 to
 þ 
S2 O2
3 þ 8e þ 2H ! 2HS þ 3H2 O ð3Þ 1000 mg L1 and 8.75 g L1 Na2S2O3. Potential was scanned at
50 mV s1.
Chronopotentiometric electrodeposition was performed using
2Hþ þ 2e ! H2 ð4Þ a 1.0 cm2 Pt foil as working electrode. In this case, 3.0 mA was

applied, giving an initial electrode potential of 0.7 V and the
At low concentrations, HS reacts according to the mechanism initial Ag+ concentration was reduced to 374 mg L1, since Ag2S
of Eq. (5), precipitating silver as Ag2S. formation occurs preferentially at low concentrations in the
HS þ 2AgðS2 O3 Þ3 þ  conditions studied in this work.
2 þ 3H ! Ag2 S þ 4HS2 O3 ð5Þ

Using flat plate electrodes, many authors investigated the silver
electrodeposition under galvanostatic and potentiostatic modes
[10–12]. The influence of flow rate, current, and potential on silver
removal rate were investigated and it can be concluded that,
besides high mass transfer rates, the rigorous control of the
potential was fundamental to obtain high current densities and
silver electrodeposition to concentrations lower than 1.0 mg L1
without the formation of Ag2S.
In order to overcome mass transfer constraints and improve
current efficiency, three-dimensional electrodes have been studied
for silver electrodeposition. Packed bed of carbon particles [9] and
reticulated vitreous carbon [13,14] has been used for this purpose.
Although it is recognized that three-dimensional electrodes can
satisfactorily be applied to recover metals from diluted solutions,
the electrode potential control in these electrodes is not easy due
2.3. Silver electrodeposition on RVC
Fig. 1(a) shows a schematic representation of the experimental
system used for silver electrodeposition. It was composed by: 1)
electrolyte tank with maximum capacity of 5L; 2) centrifugal pump
(Bomax, PX-NH40); 3) rotameter; 4) diaphragm valve for reactor
flow rate control; 5) voltmeter for cell potential measurement; 6)
electrochemical reactor; 7) power supply (Minipa, model 3003D)
or potentiostat (Autolab, PGSTAT30); 8) thermostatic bath; 9)
sphere valve for by-pass flow rate control, and 10) sphere valve for
system drainage.
The electrochemical reactor used for silver electrodeposition
(Fig. 1(b)) was built in acrylic. The electrolyte flowed upward in
closed loop between reactor and the electrolyte tank. A polyethyl-
ene mesh covered with a polyamide fabric was placed between the
 P.F.A. Prado, L.A.M. Ruotolo / Journal of Environmental Chemical Engineering 4 (2016) 3283–3292
Fig. 1. (a) Experimental setup. (b) Electrochemical reactor: 1) current feeder (stainless steel); 2) RVC; 3) flow distributor; 4) polyamide fabric; 5) polyethylene mesh, and 6)
counter electrode (Ti/Ti0.7Ru0.3O2).
porous cathode and the anode and worked as a separator in order
to avoid short circuit. A 1.27 cm  4.0 cm  7.0 cm 60 ppi (porous
per inch) RVC (1166 m2 m3 specific surface area) was used as
three-dimensional cathode. In the experiments were the RVC was
not employed, the stainless steel plate (0.1  0.07 m2) was used as
the working electrode. The counter electrode was a Ti/Ti0.7Ru0.3O2
plate. As much for the process utilizing the stainless steel plate as
utilizing the porous electrode, a reference electrode of Ag/AgCl
3.0 M KCl was used to measure the electrode potential during the
electrodeposition and also for cyclic voltammetry measurements.
After the reactor had been assembled and attached to the
experimental system, the electrolyte flow rate was adjusted to
3.0 L min1 and the electrochemical reaction started when the
current source was switched on. In the experiment carried out
using the RVC, the applied current was 120 mA. Electrolyte samples
were taken along the electrodeposition process for analyses of
silver concentration by flame atomic absorption spectroscopy
(Varian, SpecAA200). Cell potential was also measured in order to
calculate the energy consumption.
Current efficiency (e) and the specific energy consumption (h)
of the electrodeposition process were calculated using Eqs. (6) and
3285
(7), respectively. The reaction rates (dC/dt) were determined from
the experimental concentration-time curves.
z  F  V dC
e¼  ð6Þ
M  I dt
I  Ecell
h¼ ð7Þ
V  dC
dt
2.4. Determination of the diffusion coefficient and diffusion layer
thickness
The silver diffusion coefficient was determined in an aqueous
media by the diffusion cell with porous membrane (diaphragm)
schematized in Fig. 2. The diaphragm cell was composed of two
compartments separated by a porous membrane. Two solutions
with different concentrations filled the compartments V1 and V2
(Fig. 2(a)). Diffusion occurred inside the diaphragm with length L
and total area A. Since the volume of the pores is small compared to
the volume of the compartments, the concentration on each one
3286 P.F.A. Prado, L.A.M. Ruotolo / Journal of Environmental Chemical Engineering 4 (2016) 3283–3292
Fig. 2. Diffusion cell: (a) internal view; (b) side view, and (c) general view.
varies slowly, establishing a constant concentration profile inside
the porous diaphragm. To maintain the concentration homogenei-
ty, the compartments were continuously magnetically stirred.
The cell was built in PVC with two cylindrical compartments
with a small orifice on the top to allow bubbles to rise to the
surface. Inside each compartment, a Teflon-covered magnet was
rotated by the movement of an external magnet, which was
connected by an axis to a pulley in order to better control the
angular velocity. The rotation was adjusted in a potentiometer,
which was responsible to transmit movement to the pulleys.
The cell constant (b) was determined measuring the conduc-
tivity (s) of a KCl solution on each compartment of the diaphragm
cell before and after 10 min of contact. The conductivity values
were converted to concentration and, using the diffusion coeffi-
cient for KCl from literature (1.866  109 m2 s1 [19]), the value of
b was calculated using Eq. (B.9).
The same procedure was employed to determine the diffusion
coefficient for silver in the thiosulfate solution. Ag+ concentration
was measured in each compartment of the cell after 10 min using
flame atomic absorption spectroscopy (Varian, SpectrAA200).
Using these concentrations and the aforementioned cell constant
b, D for silver ions was determined by Eq. (B.9).
2.5. Experiment carried out applying modulated current
In this experiment the RVC was substituted by the flat stainless
steel plate as working electrode but the same reactor shown in
Fig. 1(b) was still used.
Modulating the limiting current for the electrodeposition
process was accomplished with calculations of limiting current
for every instant of the process (as shown in Eq. (A.6)) according to
silver concentration expected throughout the electrodeposition
process, determined by a fixed value of e, which was assumed as
100%. Thus, silver concentration as a function of time was
calculated using Eq. (6).
The mass transport coefficient km (D/d) utilized in Eq. (A.6) was
estimated using the value of the boundary layer thickness (d)
obtained from literature for this same reactor [20] and the
diffusion coefficient (D) determined using the diaphragm cell. The
estimated limiting current as a function of silver concentration
were transferred to the NOVA software (Autolab) which automati-
cally controlled the applied current through the potentiostat. The
experimental setup (Fig. S3) used in this experiment is similar to
that schematically shown in Fig. 1(a), but the current source was
replaced by the potentiostat.
3. Results and discussion
3.1. Silver electrodeposition on RVC
As a first attempt to electrodeposit silver from the synthetic
solution, an experiment was performed using the RVC. At 120 mA,
the initial electrode potential was 0.4 V vs. Ag/AgCl. Fig. 3(a)
shows the normalized silver concentration as a function of time.
According to this figure, 70 min were enough to remove silver
completely out of the solution. From this curve and using Eqs. (6)
and (7), e and h, respectively, were calculated. The energy
consumption in this process was 6.4 kWh kg1, which is very
economically attractive comparing the prices of energy and
metallic silver.
Nevertheless all silver had been removed from the solution, the
calculated current efficiency was 156%, revealing that some other
process than silver electrodeposition occurred at the same time. In
relation to that, a black powdery formation was present over the
electrodeposited silver layer (Fig. 3(b)). This powder was
posteriorly identified by X-ray diffraction (XRD) as being silver
sulfide.
These results confirm what was observed and measured by
many authors regarding the irregular potential and current
distribution inside three-dimensional electrodes [15,22]. In the
region of the electrode close to the current feeder, where the
potential are less negative, only metallic silver was electro-
deposited but, on the other hand, in the region close to the counter-
electrode, Ag2S formation (Fig. 3(b)) was favored due to the
existence of very cathodic potentials [16]. Therefore, this experi-
ment indicated that either the electrode potential or the applied
current has to be strictly controlled during electrolysis in order to
avoid Ag2S formation and thus loss of recoverable silver. On these
 P.F.A. Prado, L.A.M. Ruotolo / Journal of Environmental Chemical Engineering 4 (2016) 3283–3292 3287

3.0

1.5

-2
0.0

i / mA cm
-1.5

-3.0
simulated wastewater
-4.5 real wastewater

-1.2 -0.8 -0.4 0.0 0.4 0.8

(a) E / V vs. Ag/AgCl
4

-2
i / mA cm
-4
supporting electrolyte
simulated wastewater
-8 real wastewater

Fig. 3. (a) Normalized silver concentration against time for electrochemical
reduction at RVC; (b) Pictures of RVC before and after silver electrodeposition. Flow
rate: 3 L min1, I = 120 mA, Initial Ag+ concentration: 800 mg L1.
results, flat plate electrodes were chosen to electrodeposit silver
and continue this study.
3.2. Voltammetric analysis
Cyclic voltammetries revealed the behavior of silver ions
complexed by thiosulfate. Fig. 4(a) shows two contrasting cases.
Analyzing the voltammetry for the synthetic solution, a cathodic
peak can be observed at 0.4 V, denoting the silver electrodeposi-
tion and another peak at 0.2 V referring to oxidation of the
electrodeposited silver. The voltammetric curve for the real
effluent presented the aforementioned two peaks, however a
third one is noticed, an anodic peak at 0.1 V, revealing that other
chemical species are being oxidized, most probably one of the
organic compounds used as additives in photographic processing.
Extending the potential sweep, now ranging from 1.5 V to
1.0 V, as shown in Fig. 4(b), it can be observed three oxidation peaks
in both synthetic and real wastewater. On the voltammetry of the
supporting electrolyte (Na2S2O3), oxygen evolution reaction was
noticed at approximately 0.25 V and a wave at 0.6 V, most probably
referring to thiosulfate oxidation process. When the potential was
swept up to 1.5 V, a new oxidation peak appeared at 0.7 V in the
reverse scan at the second and third cycles. This peak was
attributed to Ag2S oxidation, since the typical black power
-12
-1.5 -1.0 -0.5 0.0 0.5 1.0
(b)
E / V vs. Ag/AgCl
Fig. 4. Voltammograms (third cycle) for the simulated and real wastewater in the
potential range of 1.0 V to 0.5 V (a) and 1.5 V to 1.0 V (b). Scan rate: 50 mV s1.
Working electrode: 1.0 cm2 Pt.
characteristic of this compound could be visually observed on
the electrode surface. This was further confirmed by X-ray
diffraction analysis as shown in Fig. 7. Silver sulfide is not desired
since it does not evoke any economic interest, representing an
impurity in the silver deposit in this case.
During electrodeposition of silver, the pH changed due to
oxygen evolution reaction over the anode, causing an increase of
H+, thus lowering the pH. To verify how the pH could interfere on
oxidation and reduction of the chemical species, cyclic voltam-
metries were performed at different pH values. The results are
shown in Fig. S1. It was verified that the oxidation and reduction
potentials were not influenced by the pH and thus the
electrodeposition process was not influenced by pH variations.
Furthermore it was observed that pHs lower than 4 induced
precipitation of a yellowish solid, indicating that the silver complex
was destabilized.
Silver concentration impact on electrodeposition process was
studied by cyclic voltammetry of solutions with distinct silver
concentrations. Firstly, a voltage sweep ranging from 1.0 V to
1.0 V was performed as shown in Fig. 5(a). A correlation between
the increase of current density due to the increase of concentration
can be observed, however it should be reminded that for this
potential range only the oxidation and reduction peaks for silver
were noticed. Hence, even for the lowest silver concentration
values, the anodic peak at 0.6 V attributed to Ag2S did not occur.
Voltammetries were also performed for the potential sweep
ranging from 1.5 V to 1.0 V to verify if peaks regarding to other
3288 P.F.A. Prado, L.A.M. Ruotolo / Journal of Environmental Chemical Engineering 4 (2016) 3283–3292
Fig. 5. Voltammograms (third cycle) for electrolytes with different silver
concentrations: (a) 1.0 V to 1.0 V and (b) 1.5 V to 1.0 V. Scan rate 50 mV s1.
Na2S2O3 8.75 g L1.
oxidation or reduction reactions would occur. In Fig. 5(b) it could
be observed that the cathodic potential sweep range did not reveal
any other reduction peaks; on the other hand, an extra oxidation
peak which appeared and was attributed to oxidation of Ag2S, as
previously discussed. In conclusion, Ag2S formation occurs at the
same cathodic potentials that hydrogen evolution reaction occurs.
Therefore maintaining a potential lower than 1.0 V would prevent
formation of Ag2S.
3.3. Silver complex destabilization
When the synthetic solution underwent conditions of pH lower
than 4, a yellowish precipitate was formed, causing solution
turbidity and interfering in the voltammetric analyzes. According
to literature review [21], the yellowish precipitate refers to
reaction in Eq. (8).
þ
S2 O2
3 ðaqÞ þ 2H ðaqÞ ! SðsÞ þ SO2ðgÞ þ H2 OðlÞ
Considering that under certain circumstances, thiosulfate in
solution can generate sulfide anions, opening possibility for silver
sulfide formation, it was hypothesized that decreasing the pH
could potentially remove sulfur from solution. Hence, removing
S2O32 from solution according to Eq. (8) would prevent the
undesirable Ag2S during Ag electrodeposition. In order to decrease
the pH and also increase the electrolyte conductivity, addition of
H2SO4 would be recommended from technical and economical
point of view. Therefore, it was necessary to know the impact of the
SO42 presence on the electrochemical process. Moreover, sulfate
anion was chosen due to its higher oxidation state and rather
superior stability. Cyclic voltammetries were performed with
different cathodic vertex potentials using a 0.5 mol L1 sulfuric acid
as supporting electrolyte (Fig. S2). It was observed that the
presence of SO42 did not have influence on the silver electrode-
position, then suggesting that a process to oxidize thiosulfate
anions to sulfate could be an alternative to electrodeposit silver
without Ag2S formation. Moreover, electrodeposition from a
solution containing sulfate would avoid a rigorous control of the
electrode potential or applied current.
In order to investigate whether sulfur removal according to
Eq. (8) would enable silver electrodeposition preventing Ag2S
formation, a series of cyclic voltammetries were performed
varying the cathodic vertex potential on the solution obtained
after decreasing the pH from 6.9 to 2.5. The solids, identified as
elemental sulfur by XRD (not shown), were separated by vacuum
filtration using qualitative filter paper. The conductivity of the final
solution (11.5 mS cm1) and Ag+ concentration (851 mg L1)
remained practically the same of the initial solution (11.3 mS cm1
and 854 mg L1, respectively). The cyclic voltammograms of this
filtered solution revealed the formation of Ag2S at high cathodic
potentials, indicating that possibly not all thiosulfate had been
removed. Based on these results, it was concluded that residual
sulfur as S2O32 was sufficiently enough to still form Ag2S during
cathodic polarization. The experiment was repeated, but in this
case the solution was filtered using a cellulose acetate membrane
with pore size of 0.45 mm was used to separate the sulfur in
suspension. Fig. 6 shows the cyclic voltammetries of this micro
filtered solution for different vertex potentials. According to this
figure, no anodic peak around 0.7 V could be observed, indicating
that no Ag2S was formed.
Considering the results shown in Fig. 6, a chronopotentiometric
electrodeposition was performed and after 2 h the final silver
concentration was 18.5 mg L1 and the pH changed from 3.5 to 4.6.
Conductivity remained practically constant throughout the elec-
trodeposition. Although there was a substantial removal of silver
from solution, a dark gray deposit could be seen over the electrode.
In order to verify its nature, a XRD analysis was performed (Fig. 7),
0
-2
-4
i / mA cm
Vertex potential:
-1.0V
-8 -1.2V
-1.4V
-1.5V
-12 -1.6V
ð8Þ -16
-1.6 -1.2 -0.8 -0.4 0.0 0.4
E / V vs. Ag/AgCl
Fig. 6. Voltammograms of the micro filtered solutions varying the cathodic vertex
potential. [Ag+] = 851 mg L1; pH = 2.5; s = 11.3 mS cm1. Scan rate: 50 mV s1.
Working electrode: Pt (1.0 cm2).
 P.F.A. Prado, L.A.M. Ruotolo / Journal of Environmental Chemical Engineering 4 (2016) 3283–3292 3289

determined measuring the final concentrations of Ag+ in each

compartment of the diaphragm cell after 10 min. Applying
Ag Eq. (B.9), the silver diffusion coefficient was calculated as being
6.1 1010 m2 s1. This value was pretty close to that reported by
Hubin and Vereecken [17] for a thiosulfate solution with similar
concentration used in this work. In their case, the value obtained
using the rotating disk electrode technique was 5.7  1010 m2 s1.
Intensity

Using the values of d as function of flow rate [20] and silver

diffusion coefficient, the values of km against flow rate were plotted
Ag in Fig. 8(a). Considering the energy consumption with regards to
pumping operation yet maintaining a relatively high km, the flow
rate of 5.0 L min1 (corresponding to 0.40 m s1) was chosen to be
Ag2S
Ag2S
Ag2S
Ag2S

Pt Pt used in the experiment applying modulated current. Therefore

Ag2S
Ag2S

Ag2S

using km = 1.7  105 m s1 the limiting current against time was
Ag2S
Ag2S

Ag2S

Ag2S
estimated combining Eqs. (6) and (A.6). Finally, an electrolysis
involving the synthetic effluent was performed with modulated
20 30 40 50 60 current according to the values of Ilim shown in Table 1, which also
2θ / degress includes the Ag+ concentration calculated by Eq. (6) (considering
e = 100%) and the experimental results of Ecell, Ag+ concentration
Fig. 7. X-ray diffraction pattern for the silver deposit obtained after chronoampero- and e.
metric electrodeposition (2 h) at 3.0 mA using the micro filtered electrolyte.
According to Table 1 and Fig. 8(b), silver was almost entirely
removed from solution after 210 min. Considering that the process
revealing an intense peak indicating silver, however, many peaks was operated at 100% e, it can be concluded that the procedure
did indicate Ag2S as well. used to calculate the limiting current was very reliable and,
In order to verity if pure silver could be obtained from an consequently, the concentration-time curve showed the typical
electrolyte only containing SO42, a new chronopotentiometry was exponential trend regarding to a mass transport controlled
carried out using an electrolyte prepared with Ag2SO4 as Ag+ process, once again confirming that it is under limiting current
source and Na2SO4 as supporting electrolyte. After 2 h of condition. Current efficiency was practically kept at 100%
electrodeposition, a grey spongy deposit could be seen over the throughout the process; exceptions occurred at the beginning of
electrode, indicating that silver was electrodeposited on the region
of hydrogen formation, but noAg2S black precipitate was found.
It can be concluded that even though S2O32 can be removed
from solution, there is still residual amounts which lead to Ag2S
formation, mainly at low silver concentrations in which high
cathodic potentials create the appropriate conditions for Ag2S
formation. Indeed, although the initial electrode potential (0.7 V)
is low enough to ensure that Ag2S would not be formed at the
beginning of the electrodeposition, over the galvanostatic process
this potential plunged to very negative values, thus favoring
hydrogen evolution and Ag2S reaction.
After these results, a new strategy to efficiently remove silver
from solution preventing Ag2S reaction was investigated. It was
based on the modulated current method [23,24], in which the
applied current is reduced over the time in order to avoid highly
negative potentials. In this case, the simulated wastewater was not
submitted to any pretreatment.

3.4. Silver electrodeposition applying modulated current

Considering the complex and uneven current and potential

distribution inside a porous electrode [15,22], the replacement of
RVC by a flat electrode was necessary in order to ensure an
equipotential surface and consequently, a better control of the
current to be applied during the modulated procedure. Moreover,
considering the price of silver compared to the process operation
costs (6–20 kWh kg1), the stainless steel plate electrode still
remains competitive compared to the RVC, even when process is
carried out under mass transfer condition and lower current
efficiency. In this case, energy consumption regarding the
electrochemical reaction has a marginal impact on the total
process cost, especially considering the profits coming from the
price of silver. In summary, the main challenge regarding the
electrochemical technology is not the energy consumption but
preventing Ag2S formation in order to avoid silver loss.
Fig. 8. (a) Mass transfer coefficient against flow rate; (b) Silver concentration
Firstly, the 1.05  105 m2 diffusion cell constant b was against time ([Ag+]0 = 770 mg L1, 8.75 g L1 Na2S2O3,V = 0.5 L, km = 1.76 106 m s1);
determined using KCl. Silver diffusion coefficient was then (c) photograph of the cathode after the electrolysis.
3290 P.F.A. Prado, L.A.M. Ruotolo / Journal of Environmental Chemical Engineering 4 (2016) 3283–3292

Table 1 support. Moreover, P. F. A. Prado acknowledges CNPq by the PIBIC

Electrodeposition process under modulated current control.
fellowship.
t/min t/s Ilim/mA [Ag+]/mg L1 Ecell/V [Ag+]/mg L1 e/%
(calculated) (measured) Appendix A.
0 0 82 770 2.24 770 –
10 600 71 664 2.13 752 16
20 1200 61 573 2.04 648 100
A. Mass transfer control: limiting current
30 1800 53 494 1.88 556 100
40 2400 45 427 1.91 479 100
50 3000 39 368 1.84 413 100 In electrochemical processes controlled by mass transport, in
60 3600 34 317 1.78 340 100 general there are three different mechanisms: 1) Diffusion; 2)
80 4800 25 236 1.84 241 100 Migration; and 3) Convection [25] as the equation below shows:
100 6000 19 176 1.66 166 100
130 7800 12 113 1.65 88 100 zk  F
160 9600 8 72 1.62 43 90 Nk ¼ Dk  rC k   C k  mk  rfs þ C k  v ðA:1Þ
Mk
180 10,800 6 54 1.62 22 98
210 12,600 6 – 1.66 13 38 Eq. (A.1) considers three different mechanisms of transport.
While diffusion contributing to mass transport occurs as a result of
a concentration gradient, driving the transport of species in the
solution, migration occurs due to an electric potential gradient
responsible by transporting charged species (thus an ionic
the experiment, probably due to transient caused by modifications current). In most cases an inert/support electrolyte in excess is
on the electrode surface, and at the end of the electrodeposition, used to make sure that migration does not have influence neither
after 180 min, when the modulated current was interrupted and contributes to overall mass transport mechanism. The third mass
the process started to work galvanostaticaly at 6.0 mA, thus the transport mechanism occurs due to mechanical forces in solution,
applied current surpassed Ilim and then e sharply decreased. In induced by agitation of electrolyte and, in some cases, with a
summary, the average current efficiency for the electrodeposition moving electrode.
process was very expressive, 93%. Regarding energy consumption, For systems consisting of turbulent flow regimes, a straightfor-
two factors contributed to its reduction: high values of e and Ecell ward model denominated film model is generally applied. The first
reduction over the electrodeposition as a consequence of the hypothesis assumed is that mass transport occurring due to a
attenuation of the applied current during the modulated proce- convection mechanism in the solution bulk is faster than molecular
dure. diffusion, hence molecular diffusion can be ignored and a uniform
The average energy consumption was calculated taking into concentration profile in the bulk can be assumed. The second
account the average values of Ecell and e; as expected, the value was hypothesis involves attributing all resistance to mass transport to
very low, 0.49 kWh kg1, confirming that the electrochemical molecular diffusion in the laminar layer, which can be simplified by
process is affordable from the economic point of view. However, a linear concentration gradient. The point where the linear
the most interesting contribution of the process operating at concentration gradient reaches the same concentration as in the
modulated current is that silver can be efficiently recovered while bulk is defined as the diffusive boundary layer thickness (d) [26].
preventing Ag2S formation, since practically no silver sulfide was Therefore, the mass flux can be calculated taking into account the
observed on the electrode surface at the end of the electrolysis, as molecular diffusion of the electroactive species through a stagnant
shown in Fig. 8(c). Likely, the few black spots observed on Fig. 8(c) boundary layer of thickness d, as shown in Eq. (A.2) [27].
refer to Ag2S formation after 180 min, when the process was not !
under modulated mode. On contrary, carrying out the electrode- @C @2 C
¼D ðA:2Þ
position under galvanostatic conditions, after few minutes of @t @x2
electrolysis the electrode surface would appear like that shown in
Fig. S4, i.e., almost completely covered by a thick layer of Ag2S. Considering steady state and the boundary conditions: x = 0,
C = Cs and x = d, C = Cb, Eq. (A.3) for diffusive flow in the boundary
4. Conclusion layer is obtained:
D
Destabilization of the thiosulfate complex and removal of ND ¼ ðC b  C s Þ ðA:3Þ
d
colloidal sulfur by filtration in order to avoid sulfide generation
during silver electrodeposition was not successful. On the other D/d is known as mass transport coefficient (km).
hand, silver was satisfactory recovered from a synthetic solution, Considering a process controlled by mass transfer, the electric
simulating a photographic wastewater, by electrodeposition using current can be expressed by Eq. (A.4):
a modulated current procedure. The recovery was achieved at high zF
efficiency rates, low energy consumption and without the presence I¼  ND  A ðA:4Þ
M
of parasitic reactions. The highly negative potentials were avoided
Combining both Eqs. (A.3) and (A.4):
when the limiting current was applied and, consequently, sulfur
sulfide was not formed over silver deposit. Furthermore the zF
I¼  km  AðC b  C s Þ ðA:5Þ
technique of diaphragm cell was an easy and suitable way to obtain M
the diffusion coefficient in order to determine the mass transfer
On Eq. (A.5), when the concentration over the electrode surface
coefficient and allow the calculation of the limiting current.
is close to zero, due to either mass transfer limitations or fast
surface kinetics, the current density reaches a maximum known as
Acknowledgments
limiting current or limiting current density (Ilim/A) [28], as shown
in Eq. (A.6):
The authors are grateful to FAPESP (Fundação de Amparo à
Pesquisa do Estado de São Paulo) and CNPq (Conselho Nacional de zF
ilim ¼  km  C b ðA:6Þ
Desenvolvimento Científico e Tecnológico) for the financial M
 P.F.A. Prado, L.A.M. Ruotolo / Journal of Environmental Chemical Engineering 4 (2016) 3283–3292 3291

In many electrochemical processes, especially those targeting and substituting Eq. (B.4) into Eqs. (B.2) and (B.3):
effluent treatment, the concentration of the toxic compounds in
A I 
the solution is low and need to be further reduced. These processes V 1 dC 1A þ DAB C A  C IIA dt ¼ 0 ðB:5Þ
L
are mainly controlled by mass transport since the toxic chemical
species are present in rather low concentrations and, in this way,
one of the main objectives of the Electrochemical Engineering for A I 
these issues is to use mathematical correlations based on V 2 dC 2A  DAB C A  C IIA dt ¼ 0 ðB:6Þ
L
adimensional numbers such as Sherwood, Reynolds and Schmidt
as a practical tool to design electrochemical reactors and to Subtracting Eq. (B.5) from Eq. (B.6):
establish operational parameters. Therefore, simple and functional  
d C IA  C IIA  
techniques to determine the mass transport coefficient aim at A 1 1
  ¼ þ DAB dt ðB:7Þ
facilitating the limiting current calculation for an effluent C IA  C IIA L V1 V2
treatment process and to allow a correct design of the electro-
chemical reactor. A constant labeled as “cell constant” (b) can be obtained grouping
the geometric terms of Eq. (B.7):
B. Diaphragm cell for diffusion coefficient measurement  
A 1 1
b¼ þ ðB:8Þ
L V1 V2
The technique of diaphragm cell was introduced by Northrop
and Anson in 1928 and has the main objective to obtain in a simple Integrating Eq. (B.8) between time t = 0 to t = t, labeling initial
way to determine the diffusion coefficient of chemical species. The concentrations C1A0 and C2A0, and final concentrations C1Af and C2Af,
fundamental idea consists in confining diffusion in a porous it can be obtained Eq. (B.9):
diaphragm, minimizing vibration [29]. 2 1 3
The diffusion cell consists of two compartments, each one filled C Af  C 2Af
ln4 5 ¼ bDAB t ðB:9Þ
with the same solution at different concentrations, separated by a C 1A0  C 2A0
porous membrane (diaphragm). The diffusion occurs inside the
diaphragm with length L and total area A. If the compartments can The constant b can be experimentally determined considering a
be considered sizable enough compared to the number of pores system in which the diffusion coefficient is known. Once
inside the diaphragm, the concentration variation is slow, hence a determined b, the diffusion coefficient of the component for an
constant concentration profile inside the diaphragm can be unknown system can be then determined from the initial
assumed during the experiment. Both compartments are agitated concentrations in each compartment and measuring after time t
continuously to avoid any concentration variation. Additionally, the final concentrations in each compartment of the cell.
membrane pore diameter must be sufficiently small compared to
molecular dimensions in order to avoid influence of agitation on Appendix B. Supplementary data
diffusion and assuring that mass transport occurs under molecular
diffusion [29]. Supplementary data associated with this article can be found, in
The agitation in the compartments, besides homogenizing the the online version, at http://dx.doi.org/10.1016/j.jece.2016.06.035.
concentration, creates turbulence which minimizes the hydrody-
namic boundary-layer over the diaphragm's surface. Therefore References
diffusion in the liquid film can be neglected. Fig. 2(a) shows a
schematic view of the diffusion cell. [1] Kodak Alaris Publication, Recovering silver from photographic processing
Once the stirring is carefully selected to neglect the effect of the solutions, Technical Data/Environmental Information J-215, 2015.
[2] V.I. Galván, A.G.M. Ruelas, A.L. Valdivieso, Radiological waste processing for
hydrodynamic boundary layer, the mass transfer transport inside the recovery of silver through cementation with zinc powder, Sep. Sci. Technol.
the pores can be considered one-dimensional, with diffusion being 44 (2009) 2695–2706.
the only mass transfer occurring, assuming that there is no [3] K.G. Adani, R.W. Barley, R.D. Pascoe, Silver recovery from synthetic
photographic and medical X-ray process effluents using activated carbon, Min.
convective movement. Considering binary diffusion with a Eng. 18 (2005) 1269–1276.
constant diffusion coefficient (DAB) and assuming that the process [4] I. Rivera, A. Roca, M. Cruells, F. Patiño, E. Salinas, Study of silver precipitation in
occurs in a quasi-stationary regime, the mass flux of species A can thiosulfate solutions using sodium dithionite. Application to an industrial
effluent, Hydrometallurgy 89 (2007) 89–98.
be described as: [5] J.M. Grau, J.M. Bisang, Silver electrodeposition from photographic processing
solutions, J. Chem. Technol. Biotechnol. 53 (1992) 105–110.
1dmA dC
NAz ¼ ¼ DAB A ðB:1Þ [6] U. Condomitti, A.T. Silveira, G.W. Condomitti, S.H. Toma, K. Araki, H.E. Toma,
A dt dx Silver recovery using electrochemically active magnetite coated carbon
particles, Hydrometallurgy 147–148 (2014) 241–245.
Similarly the same expression can be written for component B. [7] J.P. Guerlet, A. Storck, G. Valentin, M. Giron, Recovering silver from photo-
With a global mass balance in the membrane thickness (L), the graphic fixing solutions using perpendicular electrolyser with volume cathode
mass variation of A inside each compartment as function of time separator from anode by ion exchange membrane. European Patent, No.
183602-A1 (1986).
can be written as: [8] A.A. El-Sattar, M. El-Shayeb, H.A. Dief, Silver recovery from photographic and
X-ray films and fixer solution waste using a packed bed reactor, Alex. Eng. J. 43
dm1A dC 1 dC 1 (2004) 393–399.
¼ V 1 A ¼ DAB A A jx¼0 ðB:2Þ
dt dt dx [9] J.V. Zee, J. Newman, Electrochemical removal of silver ions from photographic
fixing solutions using a porous flow-through electrode, J. Electrochem. Soc. 124
(1977) 706–708.
[10] M. Spitzer, R. Bertazzoli, Selective electrochemical recovery of gold and silver
dm2A dC 2 dC 2 from cyanide aqueous effluents using titanium and vitreous carbon cathodes,
¼ V 2 A ¼ DAB A A jx¼L ðB:3Þ Hydrometallurgy 74 (2004) 233–242.
dt dt dx [11] G.N.K.R. Bapu, C. Eagammai, S. Jayakrishnan, Electrolytic recovery of silver
Assuming steady state, from low concentrated silver cyanide spent plating solutions, Trans. Inst. Metal
Finish. 86 (2008) 66–72.
[12] P.A. Ramirez, V.E. Reyes, M.A. Veloz, Silver recovery from radiographic films
dC A dC A C I  C IIA
j ¼ j ¼ A ðB:4Þ using an electrochemical reactor, Int. J. Electrochem. Sci. 6 (2011) 6151–6164.
dt x¼0 dt x¼L 0L
3292 P.F.A. Prado, L.A.M. Ruotolo / Journal of Environmental Chemical Engineering 4 (2016) 3283–3292
[13] Q. Radiguès, R. Santoro, J. Proost, Kinetic transitions during Ag and Cu
electrorecovery on reticulated vitreous carbon electrodes in flow-by mode,
Chem. Eng. J. 162 (2010) 273–277.
[14] V. Reyes-Cruz, I. González, M.T. Oropeza, Electro-recovery of gold and silver
from a cyanide leaching solution using a three-dimensional reactor,
Electrochim. Acta 49 (2004) 4417–4423.
[15] L.A.M. Ruotolo, J.C. Gubulin, A mathematical model to predict the electrode
potential profile inside a polyaniline-modified reticulated vitreous carbono
electrode operating in the potentiostatic reduction of Cr(VI), Chem. Eng. J. 171
(2011) 1170–1177.
[16] L.A.M. Ruotolo, J.C. Gubulin, A factorial-design study of the variables affecting
the electrochemical reduction of Cr(VI) at polyaniline-modified electrodes,
Chem. Eng. J. 110 (2005) 113–121.
[17] A. Hubin, J. Vereecken, Electrochemical reduction of silver thiosulphate
complexes. Part I: Thermodynamic aspects of solution composition, J. Appl.
Electrochem. 24 (1994) 239–244.
[18] P.R. Anderson, B. Kim, C. O’Connor, Photocatalytic Process for Silver Recovery
and Wash Water Reuse, Report, Illinois Institute Of Technology, 2000.
[19] L.A. Woof, J.F. Tyllei, Revised values of integral diffusion coefficients of
potassium chloride solutions for calibration of diaphragm cells, J. Phys. Chem
71 (1967) 1962–1963.
[20] R.B.A. Souza, Electrochemical degradation of phenolic compounds present in
oil refinery effluents, M.Sc. Thesis, Federal University of Sao Carlos, 2011.
[21] A.F. Holleman, E. Wilberg, Inorganic Chemistry, Academic Press, New York,
2001.
[22] J.S. Newman, Electrochemical Systems, Prentice-Hall, Englewood Cliffs, 1973.
[23] L.A.M. Ruotolo, J.C. Gubulin, Optimization of Cr(VI) electroreduction from
synthetic industrial wastewater using reticulated vitreous carbon electrodes
modified with conducting polymers, Chem. Eng. J. 149 (2009) 334–339.
[24] P.H. Britto-Costa, L.A.M. Ruotolo, Electrochemical removal of copper ions from
aqueous solutions using a modulated current method, Sep. Purif. Technol. 46
(2011) 1205–1211.
[25] A.J. Bard, L.R. Faulkner, Electrochemical Methods: Fundamentals and
Applications, 2nd ed., McGraw-Hill, Chichester, New York, 1980.
[26] R.E. Treybal, Mass-Transfer Operations, McGraw-Hill, New York, 1981.
[27] F. Goodridge, K. Scott, Electrochemical Process Engineering, Plenum Press,
New York, 1995.
[28] D.K. Pletcher, F.C. Walsh, Industrial Electrochemistry, Chapman and Hall,
London, 1990.
[29] J.C. Gubulin, J.T. Freire, Didactic Laboratory: Special Topics on Transport
Phenomena, UFSCar, São Carlos, 1990.