The final performance of the stack depends on a variety of parameters:
(1) membrane properties (conductivity, selectivity, osmotic behavior),
(2) cell properties (compartment thickness, spacer type), (3) stack design parameters (way of feed, electrodes), (4) operating conditions (flow rate, electrical load), and (5) water quality (salt content, impurities, temperature, composition). These different parameters often conflict with each other and all together they determine the final power output. Veerman et al.  systematically investigated the performance of a real RED stack with respect to power density and energy efficiency, especially focusing on the effect of the current density, the membrane and spacer resistance, and the feed flow rate. They used a custom-made RED stack with an adaptable number of cells, with a maximum of 50 cells (total effective membrane area of 1 m2). Each cell consisted of an anion and a CEM with an effective membrane area of 100 cm2 per membrane. Commercially available membranes from Fumasep (Germany) were used: FAD as AEM and FKD as CEM. These membranes have a thickness of 0.082 mm. Polyamide woven sheets with a thickness of 200 mm were used as spacer. As electrode system, the authors used a solution of 1M NaCl with 0.05 M K4Fe(CN)6 and 0.05 M K3Fe(CN)6. Sea and river water were represented by NaCl solutions of, respectively, 30 and 1 g/L. The 50-cell stack generated a power output of 0.93 W, which is the highest power output reported for RED using sea and river water. Fig. 13 shows the power output of the stack as a function of the current density for different numbers of cells (N) .
The obtained power increases almost linearly with the number of cells, which indicates that the losses due to limiting currents are limited . The maximum power obtainable in this stack is 0.93 W/m2, which is the highest power reported in literature.
Not only the power output is an important parameter, the energy efficiency also plays a significant role. It represents the fraction of the total available energy available from the mixing of river and seawater that is really used to generate power. In the case of the stack experiments of Veerman et al. , the highest power density reported could be obtained. However, the energy efficiency at that point is no more than 50% . So optimization with respect to obtained power only would result generally in low energy efficiencies and loss of potentially available energy. Post et al.  show that, in principle, no fundamental limitations restrict the energy efficient use of the resources and values as high as 80% can be obtained.
In the real application, the power density obtainable in a RED stack is often reduced due to parasitic currents, or also called current leakage in the
stack. There are two sources of these losses : (1) ion exchange membranes are never 100% selective, which apart from generating the transport of counterions, also generates a transport of co-ions, which reduces the power output. This issue is related to membrane design and optimization. (2) Ionic shortcut currents occur due to the transport ofions in feed and drain channels and this effect is more severe at higher salt concentrations. These ionic shortcut losses are strongly related to stack design.
In principle, three different ionic shortcut currents can be distinguished in the stack :
1. Ionic shortcut currents in the electrode solution (the electrode solution connects the anode and the cathode compartment). These losses can be easily reduced by increasing the length of the tubing that connects the electrodes.
2. Ionic shortcut currents between the river water compartments. Generally, this shortcut current can be neglected because the salt concentration in the river water compartment is too low to cause significant leakages.
3. Shortcut currents between the seawater compartments.
Veerman et al.  investigated the possibilities to reduce the shortcut currents between the seawater compartments. Model calculations show that the effect of these losses can be significantly reduced through proper stack design. Especially important in this respect are the number of cells (N), the channel resistance in relation to the cell resistance (R/r), and the lateral spacer resistance in relation to the cell resistance (r/r), where the latter two are the critical design parameters that need to be optimized . In medium-size stacks, the number of cells and the ratio R/r and p/r need to be as high as possible . Possibilities to do so include (i) increasing the channel resistance (R) by narrowing the channels; (ii) increasing the lateral spacer resistance (p) by using thinner spacers (in the seawater compartment); and (iii) decreasing the cell resistance r by using membranes with low resistances and thin spacers (in the river water compartment). Possibilities to increase R are limited: with narrowing the channels, the hydrodynamical resistance in the channels also increases. An increase in the spacer resistance only induces a very small change in power output and the only way to increase the power output is to decrease the cell resistance, as it increases the efficiency and the power output of the system . In very large stacks, (R/r) (the channel resistance in relation to the cell resistance) should be maximized to obtain the largest power output . A decrease in r can be induced by minimization of the membrane thickness and the thickness of both the sea and the river water compartments, but this results in an equal decrease in the channel resistance, and consequently has no effect. But, at given membrane thickness and river water compartment dimensions, a decrease of the thickness of the seawater compartment induces a lower r and a higher R/r ratio and consequently a higher efficiency and power output .
Was this article helpful?