SciELO - Scientific Electronic Library Online

vol.20 issue3A study of the process of desorption of hexavalent chromiumA statistical approach to the experimental design of the sulfuric acid leaching of gold-copper ore author indexsubject indexarticles search
Home Pagealphabetic serial listing  

Services on Demand




Related links


Brazilian Journal of Chemical Engineering

Print version ISSN 0104-6632On-line version ISSN 1678-4383

Braz. J. Chem. Eng. vol.20 no.3 São Paulo July/Sept. 2003 

Cr (VI) electromechimal reduction using RVG 4OOO graphite felt as the electrode



E.O.Vilar; E.B.Cavalcanti; H.R.Carvalho; F.B.Sousa

UFCG/CCT/DEQ - Laboratório de Engenharia Eletroquímica , Av. Aprígio Veloso 882, CEP 58109-970, Phone 55 (83) 310-1314, Fax 55 (83) 310-1114, Campina Grande PB, Brazil, E-mail:




Even in at very low concentrations, heavy metals in industrial waste constitute environmental and health risks. The U.S. Department of Health and Human Services has recognized as chromium compounds and defined carcinogens the level acceptable in drinking water as being only 0.05 ppm. The objective of this work was the electrochemical reduction of hexavalent chromium Cr (VI) to Cr (III) ions in a dilute synthetic solution of K2Cr2O7 and Na2SO4 (0.05N). A plug-flow reactor with an RVG 4000 graphite felt (Le Carbone Lorraine, France) electrode was used for this work. Its morphological characteristics such as specific variables surface, porosity, average fibre diameter and permeability were determined. The influencing process selectivity such as initial concentration of Cr (VI), solution pH, current intensity and conversion yield are considered. The fractional conversion achieved in the plug-flow reactor in the present work, was about 90%.

Key words: graphite felt, hexavalent chromium, electrochemical reduction, plug-flow reactor.




Wastewater polluted with hexavalent chromium is produced by industries such as textile, tanning, metal finishing, pigment, drug and organic chemical and by cooling systems where hexavalent chromium is added as a corrosion inhibitor.

Porous electrodes can be used to effectively remove low concentrations of metal or toxic substances from solutions. Of the porous electrodes, carbon or graphite felts have aroused the most interest owing to their higher area per unit volume and mass transfer characteristics.

Numerous studies on the removal of metal ions from diluted water solutions using graphite felt, carbon or RVC electrodes have been reported in the literature. Walsh et al. (1992) studied electrolytic removal of cupric ions from dilute liquors using reticulated vitreous carbon cathodes. This work reported on the deposition of copper from synthetic, acid sulphate solutions in order to characterise the performance of RVC electrodes on a laboratory scale.

Abda et al. (1991) were the first to use RVG-2000 graphite felt (Le Carbone Lorraine, France) to reduce hexavalent chromium Cr (VI) to Cr (III). Oren and Soffer (1983) have also measured removal of mercury ions using graphite felt. They found the variation between limiting current and flow rate to be linear, owing to the near-completeness of conversion under the conditions of their experiments. This type of electrode has received considerable research attention during the past decade (Tricoli et al., 1993; Podlaha and Fenton, 1994; Winder et al., 1997). They also used a plug-flow reactor (PFR) showed that graphite felt cathodes were better at removing metal ions. Generally a distinction is made between the flow-through and the plug-flow concepts, based on whether the fluid flow and the electrical current are parallel or perpendicular to each other. A comparison between these two alternative reactor configurations shows that at a given maximum ohmic drop and reactant conversion the maximum permissible flow rate in a plug-flow electrode is higher than in a flow-through electrode, provided that the length-to-depth ratio is greater than 5 (Simonsson, 1984). Iglia et al. (1996) studied the reduction of Cr (VI) using an RVG (reticulated vitreous graphite) in an attempt to process the deposition of this metal. Use of the GBC reactor for reduction of chromate in dilute solution with a carbon-particle packed-bed electrode is reported by Wijnbelt and Janssen (1994). Recently the same GBC reactor (gas-diffusion electrode packed- bed electrode cell), consisting of a hydrogen gas-diffusion electrode in direct contact with a graphite-felt packed-bed electrode was studied by Njau and Janssen (1999). A number of studies have shown that the electrochemical kinetics of reduction of hexavalent Cr depend heavily on initial pH solution (Abda et al., 1991).

The main reactions involved in the electrochemical reduction are:

anode reaction

cathode reaction

cathode reaction

Golub and Oren (1989) studied the mechanisms and electrochemical principles related to these reactions. By using a cyclic voltammeter, they were able to observe that the reaction (II) occurred at – 600mV in comparison to the mercury-mercurous sulphate reference electrode. They used an RVG 2000 graphite felt with a work electrode and a solution made up of dichromate potassium with sodium sulphate as a supporting electrolyte.



To be able to visualise the morphological structure of the fibres and their spatial distribution, a Scanning Electron Microscope (SEM) LEICA model S4401 was used.

A sketch of the plug-flow reactor is show in Figure 1. Each reactor compartment has a capacity of 4x 3x1 cm3, and was machined from two identical blocks of Plexiglas. Two RVG 4000 graphite felt electrodes were clamped together with two acrylic frames, separated by a cation exchange membrane (Nafion 115) that divided the cell into two compartments. The electric connection to the electrodes was made with the graphite felt and a sheet of Papyex (Carbone-Lorraine). Cathode potential (- 600 mV) was controlled and measured with a [Hg|Hg2SO4|Na2SO4 (1M)] as reference electrode using an AMEL potentiostatic-galvanostatic model 555B. Synthetic solutions were used to up 7, 14 and 25 ppm of Cr (VI) in the form of K2Cr2O7, and Na2SO4 (0.05N) as an electrolyte support at a working temperature of 30oC.



The pH of the feed solution was acidified to 3.0 with H2SO4. The treated solute ion was passed through the porous electrode at flow rates (measured using a flow meter) between 3.0 x 10-6 and 6.5 x 10-6m3.s-1.The solution was pumped and recycled. The electrolyte leaving the cell was recycled in a reservoir with a volume of 0.05 dm3.Catholyte samples were removed at regular intervals and analysed for Cr (VI) concentration with the aid of an SBC atomic absorption spectrophotometer, model AA7000. Interference by Cr (III) was avoided by prior precipitation (pH » 10), followed by filtration of its hydroxide Cr (OH)3.



Structural Aspect of the RVG 4000

The microstructure and fibre size of a porous RVG 4000graphite and its surface morphology are illustrated in Figures 2a, 2b, 3aand 3b.





Figures 2a and 2b show an anisotropic medium composed by randomly dispersed fibre. In Figures 3a and 3b, it can be seen that each graphite fibre is the result of a fusion of others melted together lengthways and shaped like an ellipse. The entangled cross-sections of the fibre depict the characteristic of fibrous structure that produces a larger reaction area. A maximum fibre diameter of 18 mm (Figure 3a) and minimum of 9.57 mm (Figure 3b) were obtained from the dimensions taken from micrographs. Because of these characteristics, some properties, such as pore diameter, have no meaning from a geometric point of view.

The morphologic characteristics of the RVG 4000 (1.0 cm thick) were estimated from measurements of static pressure drop (DP.z-1) as a function of linear flow velocity under laminar conditions, using the well-known Kozeny-Carman equation, which was corrected by Kyan (Dullien, 1975) for fibrous beds. Kyan proposed a modification of the equation constant, as expressed in the following form:

where z corresponds to the height of the porous bed; m the dynamic viscosity of the electrolyte; ae the specific surface per unit volume of the bed; the average percolating velocity; the mean porosity of the bed, calculated from the apparent and true densities giving a value of 0.96; and k' the modified constant value equal to 7, estimated by graphical interpolation from Figure 4.



By determining the value of ae, we can estimate fibre diameter, df, as defined by the following equation:

where as corresponds to the specific surface per unit volume of solid. To be able to determine permeability B, the classic Darcy equation shown below was employed:

Using the Kyan model the values found for the principal morphologic parameters were df = 8.7 mm, ae = 18,400 m-1 and B = 3.73 x 10-10 m2. Table 1 shows these results in comparison to those for other materials (González-Garcia et al., 1999).



Variation in Current Intensity, I

Current intensity profiles as a function of time at different flow rates are shown in Figure 5. This same behaviour was observed by Njau and Janssen (1999), when using a GBC reactor. It can be observed that the maximum value was achieved after approximately 10 minutes of process followed by an exponential reduction. The same behaviour can also be seen for different initial concentrations used in this work. This can be explained by the close relationship between current intensity I, and the initial concentration Ci of the electrochemically active ionic species, which is present in the boundary layer concentration on the graphite fibres in the first moments in the operation. This relationship can be expressed by the following equation:



where Rf is the Faraday yield, V the electrolyte volume, C the ionic concentration on the electrode/electrolyte interface, Á the Faraday constant , and t time.

The gradual increase in intensity current during the first minutes of the experiment was probably due to the increase of activity of electrode surface by reduction of oxides present initially at the surface. At the same time, a parallel reaction associated with the production of hydrogen gas was observed (except at a 7 ppm concentration). The quantity of gas bubbles generated is believed to be associated with the initial concentration of Cr ions. The high concentration of chromium ions also produced a high concentration of H+ ions, which are adsorbed on the graphite surface by the generic mechanism of adsorption, production of gas and desorption (Navarro et al., 1999).

The reduction in the concentration of Cr2O7= ions in the boundary layer concentration gradually reduces this process. This results in a decrease of H+ ions at the electrode surface.

Variation in initial concentration of Cr (VI)

Figures 6 and 7 show the variation in normalized concentration of Cr (VI) with time of operation. The reduction rate decreased quickly and was established after a period of approximately 120 minutes of operation elapsed. This behaviour was independent of initial concentration (Figure 6) and flow rate (Figure 7). In Figure 6, for example, the experimental data average (Qv = 4.4x10-6m³/s) can be represented approximately by an exponential decay curve:





Equation (5) can be associated with the parameters of Equation (a.7) (see Appendix) by:

In this case, the correlation coefficient is 0.90. On the other hand, it can be observed in Figure 7 that reducing the residence time in R = VR .Qv-1 in the reactor (VR is the reactor volume (24x10-6 m³) and Qv the volumetric flow rate; improvement in the mass transfer by convection and diffusion), the yield of the process is not favoured by a long time of operation, which shows evidence of kinetic/diffusion control.

Variation in pH

Figures 8 and 9 show the influence of initial pH of the solution which was withdraw at the outlet of cell on the rate of conversion of Cr (VI) to Cr (III) ions. As was seen above, Figure 8 shows that the high concentration of Cr (VI) ions favours a rapid consumption of H+ ions with a rapid increase in pH. Figure 9 shows that a low initial concentration of H+ produces a linear decrease concentration of Cr ions, which is associated with a gradual increase in pH. It can be verified that, with the increase in initial pH from 3.0 to 4.0 and with a solution of 14 ppm in Cr (VI), a reduction of 40% in the conversion rate occurs and the process yield is considerably reduced by the low H+ concentration, which is not favourable for adequately carrying out the reaction (II).





Conversion rate, X

The change in conversion was determined from the ratio:

Where C 0 corresponds to the initial concentration of Cr (VI), C(t) the variation in concentration at time t and X(t) the conversion rate. Figures 10 and 11 show the change in conversion with time for flow rates of 3.0 x 10-6 and 4.4 x 10-6 m3.s-1 , respectively. It can be seen that the conversion rate is practically independent of initial concentration of Cr ions, which has an asymptotic profile after 40 minutes of operation. The conversion expression for the PFR in batch recycle process is represented by Equation (8) (Walsh et al., 1992), which is valid for t >> R. Figure 12 shows the influence of the dimensionless factor, d.Ae/QV, for a PFR operating in the batch recirculation mode as a plot of fractional conversion as a function of the number of recycles through the reactor, t/R . The constant value 0.003 found for thedAe/QV ratio in the present work, agrees with theoretical Equation (8). The experimental results presented in Figures 10 and 11 show that in both cases the number of recycles in 120 minute of operation corresponds to 900 and 1323 recycles to achieve an approximate conversion of 0.9.









Under the experimental conditions used in is study, the results confirmed that a rate of conversion Cr (VI) to Cr (III) is highly dependent on the initial pH of the solution. Values on the order of 90% were obtained after two hours of operation, independent of the initial concentration of Cr ions used. From the flow rates used, it can be seen that the process is controlled by mixed control. On the other hand, the production of hydrogen gas was mainly associated with the initial concentration of Cr (VI). This reduced the energy yield of the process, without however negatively affecting the conversion rate. The results found for the conversion rate, agree with theoretical data from the literature. In this case volume of the mixer tank is large in comparison to that of the reactor, and the hydrodynamics within the reactors is not significant in terms of the residence time distribution of flow. The process seems to be favourable. Other studies are also necessary to evaluate the kinetics of electrochemical reduction, space velocity and whether are really advantages to using graphite felt as work electrodes.





The authors thank the LCCQS laboratory UFPB/DQ for the chromium analysis and ALCOA-Aluminio S.A.-Itapissuma-PE for technical support in SEM photomicrographs. They also thank CNPq for financial support.



Abda, M., Gavra, Z. and Oren, Y., Removal of Chromium from Aqueous Solutions by Treatment with Carbon Electrodes: Column Effects, J. Appl. Electrochem., 21, pp. 734-739 (1991).         [ Links ]

Iglia, R.A.Di, Sousa, M.F.B. and Bertazzoli, R., Redução do Cr(VI) sobre eletrodo de Carbono Vítreo, X Simpósio Brasileiro de Eletroquímica e Eletroanalítica, Anais, pp. 272-274, October (1996).        [ Links ]

Dullien, F.A.L., Single-phase Flow through Porous Media and Pore Structure, The Chem. Engineering Journal, 10, pp.1-34 (1975).        [ Links ]

Golub, D. and Oren, Y., Removal of Chromium from Aqueous Solutions by Treatment with Porous Carbon Electrodes: Electrochemical Principles, Journal of Applied Electrochemistry, 19, pp. 311-316 (1989).        [ Links ]

Gonzaléz-Garcia, J., Bonete, P., Expósito, E., Montiel, V., Aldaz, A. and Torregrosa-Maciá, R., Characterization of a Carbon Felt Electrode: Structural and Physical Properties, J. Mater. Chem., 9, pp. 419-426 (1999).        [ Links ]

Njau, K.N. and Janssen, L.J.J., Electrochemical Reduction of Chromate Ions from Dilute Artificial Solutions in a GBC-reactor, J. Appl. Electrochemical, 29, pp.411-419 (1999).        [ Links ]

Oren,Y. and Soffer, A., Graphite Felt as an Efficient Porous Electrode for Impurity Removal and Recovery of Metals, Electrochim. Acta, 28, 11, pp. 1649-1654 (1983).        [ Links ]

Podlaha, E.J. and Fenton, J.M., Characterization of a Plug-Flow RVC Electrode Reactor for the Removal of Heavy Metals from Dilute Solutions, Journal of Applied Electrochemistry, 25, pp. 299-306 (1994).         [ Links ]

Silva Júnior, J.G., Goulart, M.O.F. and Navarro, M., Electrocatalytic Hydrogenation of Diethyl Fumarate: a Simple System Development, Tetraedron, 55, pp. 7405-7410 (1999).        [ Links ]

Simonsson, D.," A Flow by Packed Bed Electrode for Removal of Metal Ions from Wastewater, Journal of Applied Electrochemistry, 14, pp. 595-604 (1984).        [ Links ]

Thomas, Z.F., Principles of Electrochemical Reactor Analysis, Elsevier Science Publishers, New York (1985).        [ Links ]

Tricoli, V., Vatistas, N. and Marconi, P.F., Removal of Silver Using Graphite-Felt Electrodes, J. Appl. Electrochem. 23, pp. 390-392 (1993).         [ Links ]

Walsh, F.C., Pletcher ,D., White, I. and Millington, J.P., Electrolytic Removal of Cupric Ions from Dilute Liquors Using Reticulated Vitreous Carbon Cathodes, J. Chem. Tech. Biotechno. 55, pp. 147-155 (1992).        [ Links ]

Wijnbelt, E.C.W. and Janssen, L.J.J., Reduction of Chromate in Dilute Solution Using Hydrogen in a GBC-cell, J. Appl. Electrochem. 24, pp. 1028-1028 (1994).        [ Links ]

Winder, R.C., Lanza, M.R.V., Souza, M.F.B. and Bertazzoli ,R., Electrolytic Removal of Metals from Industrial Wastewater, Plating and Surface Finishing, pp.59-62 (1997).        [ Links ]



Received: March 12, 2002
Accepted: April 15, 2003




Batch Recycles Operation

We know from the literature (Walsh et al., 1992; Thomas, 1985) that batch operation combined with recycling can be efficient if mass transfer can be kept at reasonably higher rates due to the relatively low active ion concentration of the recirculating electrolyte. The system can be represented by Figure A.



The mass balance in the reactor is

where VR is of the volume reactor; Qv, the flow rate; C1 and C2, the feed and out concentrations of the reactor, respectively; ne, the number of electrons involved in the electrode reaction; Á , the Faraday constant; and I, the current intensity. The last term in Equation (a.1) is

where Ae is the electrode area and d, the average mass transfer coefficient. The mass balance in the mixing-recirculation tank is

Taking the residence time; and , in the reactor and tank, respectively, Equations (a.1) and (a.3) may be rewritten as and ,


Eliminating C1, we obtain the single second-order expression

for which the general solution is

l1 and l2 are the eigenvalues of Equation (a.4)

The A and B constants can be determined for the pre-electrolysis conditions, i.e., let t = 0 when C2 = C 02 = C 01, since there is no current flow in the reactor. We finally obtain

Equation (a.5) predicts an exponential decay of the limiting current, which corresponds to a time-varying electrolyte concentration. When the volume of the mixing tank is much larger than the volume of the reactor (the case in this present work), T >> R . Then, in Equation (a.5) l1 = 0 and l2 = -a /R and Equation ( a.6) is simplified to

Creative Commons License All the contents of this journal, except where otherwise noted, is licensed under a Creative Commons Attribution License