The overall reaction rate of salt with steam was measured during synthesis and decomposition reactions, and plotted on figure 6. The synthesis reaction rate depends on the thermodynamical constraints. For the synthesis, the evaporator temperature was fixed at 18 °C and the reactor temperature at 30 °C. Several syntheses have been achieved using the same reactive bed and with the same thermodynamical constraints (figure 6a).

The first reaction was difficult to complete: after nearly 20 hours, the reaction advancement was lower than 0,5 and a plateau appeared. The reaction advancement of the second one was about 0,8. We observed that the chemical reaction started up again after pumping the gas in the reactor during a few seconds. Every rough change in slope of the kinetic evolution is caused by pumping. Therefore it seems that the plateau does not result from a mass transfer limitation due to the swelling of the reactive block and the closing of pores. The most probable explanation is the infiltration of an inert gas inside the reactor. The Fick diffusion of steam through air is low in this range of working pressure. Thus the flow of reactive steam could have transported the inert gas into the pores of the composite block.

Figure 6: Overall reaction advancement versus time. — a — synthesis Tevap = 18 °C, Treact = 30°C — b — Effect of

pumping on overall reaction advancement evolution.

Despite this problem, the experimental kinetic evolutions were compared to simulated evolutions using a model of transformation of reactive bed. This model was described in a previous paper [2]. It involves two reactive fronts f1 and f2, controlled by mass and heat transfers respectively. It assumes thermodynamical equilibrium on each front, as displayed on figure 7. Between the mass diffuser and the front f1, the diffusion of steam is controlled by Darcy’s law and temperature is assumed uniform. Between the heat exchanger and the front f2, thermal diffusion is controlled by Fourier’s law and the pressure is uniform. Between the two fronts, Darcy’s and Fourier’s laws control respectively the mass and the heat transfer.

Figure 7: Schematic representation of the model of transformation of reactive bed

The mass transfer parameters, identified as described above, have been used in this model to simulate the transformation of the experimental blocks. Figure 8 shows a comparison between experimental and simulated evolutions of the reaction advancement. Up to now we are able to identify only k0 and b0 (i. e. mass transfer coefficients at X=0). So we use the correlations developed previously by our team [3-4] to estimate the missing parameters k1, b1, and the thermal parameters. These parameters were calculated according to the thickness, zs, the mass ratio of ENG, wl and the apparent density of the block p, Table 2. The simulated kinetics (figure 8, plot n°3) is significantly faster than the experimental one (plot n°1). So, this correlation cannot suit to the present composites whose mass ratio of ENG and apparent density are very low. In fact, the permeability value of the reactive bed (k0) resulting from the correlation, is an order of magnitude higher than the experimental one.

The other coefficients determined by the correlation are summarized in ‘Table 3’. As reference, the plot 2 on figure 8 uses this whole set of parameters.

Table 2: Transfer coefficients calculated by correlations used in the model

Dec k0 ko ki bo bi

(kWh. m-3) (W. m-1.K-1) (W. m-1.K-1) (m*) (m2) (Pa) (Pa)

226 0,9803 0,4568 1,899.10-12 7,281.10-13 185 459

As our current correlation cannot suit, we have to determine k1 and b1 parameters. The plots 4, 5, 6 correspond to decreasing values of k1. The lower is the value of k1, the smaller is the difference between simulated and experimental kinetics. Anyway, even for a very small value of k (10-16 m2 as shown in plot 8) the two kinetics are not in good agreement. Therefore, we examined some other operating parameters which can also influence this simulation, as the evaporator temperature, plot 7. If the temperature of the evaporator is lower of 2°C the simulation fit to the experimental evolution for the first 5 hours.

The first results obtained by the experimental apparatus allow to identify the permeability of the reactive bed after decomposition (X=0). Its value is close to the result obtained from previous correlations developed by our team. In order to identify these parameters in a wider range of advancement X we still have to solve a slight problem of inert gas.

References

[1] G. Rambaud, M. Mauran, N. Mazet, Modeling of steam transfer at low pressure through reactive porous media, Eurotherm seminar n°81 (2007), Albi, France.

[2] H. Lahmidi, S. Mauran, V. Goetz, Solar Energy, 80 (2006) 883-893.

[3] R. Olives, S. Mauran, Transport in porous media, 43 (2001) 377-394.

[4] S. Biloe, S. Mauran, Carbon, 41 (2002) 525-537.