International Journal of Hydrogen Energy 30 (2005) 619 626 www.elsevier.

com/locate/ijhydene

Upper bound and best estimate of the efciency of the iodine sulphur cycle
Stephen Goldstein , Jean-Marc Borgard, Xavier Vitart
CEA/DEN/DPC, CEN, Saclay, 91191 Gif sur Yvette, France Available online 21 August 2004

Abstract An upper bound and a best estimate of the efciency of a thermochemical iodine sulphur water splitting cycle coupled with a High Temperature Reactor operating between 760 K and 1144 K are proposed in this publication. An upper bound of 51% is found assuming ideal reversible chemical reactions and constraints related to the heat exchange between helium and chemical process. It appears that the separations between water and sulphuric acid on one hand and water, HI and iodine on the other hand are the most energy consuming. The best estimate of the efciency is found between 33% and 36% according to the assumptions made for the temperature pinch at the heat exchangers of the hydrogen production section. More realistic values can be obtained after having designed the heat exchangers of this section. 2004 International Association for Hydrogen Energy. Published by Elsevier Ltd. All rights reserved.
Keywords: Reaction; Thermodynamic data; Efciency

1. Introduction Among the large scale, cost effective and environmentally attractive hydrogen production processes, sulphur iodine (I_S) thermochemical cycle [1] seems to be a quite promising one. This cycle was originally studied in the 1980s by General Atomics Co [2] (GA) and can be split into the following reactions: (9I2 )1 + (SO2 )g + (16H2 O)1 (2HI + 10H2 O + 8I2 )1 +(H2 SO4 + 4H2 O)1 (393 K), (1) L2 = (2HI + 10H2 O + 8I2 )1 (2HI)g + (10H2 O + 8I2 )1 (2HI)g H2 + (I2 )1 (600 K), (500 K), (2) (3) L1 = (H2 + SO4 + 4H2 O)1 (H2 SO4 )1 + (4H2 O)1 (H2 SO4 )1 (H2 SO4 )g (570 K), (4) (5) (6) (7)

(630 K), (670 K),

(H2 SO4 )g (SO3 )g + (H2 O)g

1 (SO3 )g (SO2 )g + 2 O2

(1140 K).

Corresponding author. Tel.: +33-1-6908-6410; fax: +33-1-

6908-6802. E-mail address: stephen.goldstein@cea.fr (S. Goldstein).

The temperatures between brackets are approximate and depend upon the pressure which is not necessarily uniform in the different parts of the cycle. The rst reaction, called Bunsen reaction, proceeds exothermically in liquid phase and produces two immiscible aqueous acid phases which compositions are indicated between brackets: L1 phase which is aqueous sulphuric acid and L2 phase which is a mixture of hydrogen iodide, iodine and water named HIx . In the second reaction, HI is separated from L2 . This separation is the most critical phase of the cycle. Reaction (3) is the thermal decomposition of HI. Roth and Knoche [3] proposed to perform reactions (2) and

0360-3199/$30.00 2004 International Association for Hydrogen Energy. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.ijhydene.2004.06.005

620

S. Goldstein et al. / International Journal of Hydrogen Energy 30 (2005) 619 626

Fig. 1. Sketch of the I_S cycle.

(3) in the same reactive distillation column. Reaction (4) is the separation of L1 in H2 SO4 and H2 O. Up to now only distillation has been proposed for this separation. Several distillation owsheets are found, ranging from the simple column [4] to multi effect arrangements [5,7]. This step is energy consuming due to the large number of water moles to evaporate. Reactions (5)(7) proceed in gas phase and produce H2 O, SO2 and oxygen. These gases are cooled down before to bubble in the Bunsen reactor to separate oxygen from SO2 and H2 O. Due to the fact that reaction (7) is incomplete, a residual quantity of SO3 is found in the hot gases at the outlet of reaction (7). This SO3 is recombined to H2 O in a reactor where the reverse of reaction (6) is performed and the produced diluted H2 SO4 is recycled in reaction (4). Membrane separation of oxygen is considered by certain teams to shift to the right reaction (7) and to increase the partial pressure of SO2 in the Bunsen reactor. Besides this separation has the advantage to reduce the amount of H2 SO4 recycled. The whole cycle can be divided into three sections according to the GA nomenclature [2], we shall call section I reaction (1), section II reactions (4)(7) and section III reactions (2) and (3). A sketch of the cycle is shown Fig. 1. In this paper, we would like rst to nd an upper bound of the efciency of this cycle. Therefore, in the next section, we assumed that the energy requirements of reactions (1)(7) are the minimum prescribed by the thermodynamics, and that the decomposition reaction (7) is complete, therefore no H2 SO4 is recycled. In Section 3, we review the published efciency calculations and present our best assessment of section III. A more realistic efciency of the whole cycle is then deduced.

Finally, we conclude with the R&D needs to bring this cycle to an industrial level.

2. Efciency bound We dene the thermal efciency th as the ratio of the enthalpy of the hydrogen and oxygen recombination reaction at 0 ambient temperature and pressure HH O (Ta )=286 kJ/mol 2 to the total heat requirement of the cycle:
th = 0 HH O (Ta ) 2 , Q + W/ r

Q and W are the heat and work requirements and r the efciency of the conversion system. We used for the numerical calculations r = 0.5, which corresponds to the efciency of the current high temperature reactors projects. Let H , S and G be the enthalpy, entropy and Gibbs free enthalpy of a chemical reaction at a given temperature T . Each reaction of the cycle is supposed to be performed at a temperature for which G is as close as possible to zero. The condition G = 0 cannot be fullled systematically due to possible constraints. The reactants are brought from ambient temperature and pressure to the reaction conditions, and the products are brought back from the reaction conditions to the ambient conditions. The heat capacities of reactants and products are supposed to be equal, therefore the heat content of the products can be recovered by the reactants. Phase changes or separations are considered as independent reactions, therefore no phase change occurs during heating of the reactants or cooling of the products.

S. Goldstein et al. / International Journal of Hydrogen Energy 30 (2005) 619 626

621

Thus, the heat Q and work W requirements of each reaction are: Q T S, W H T S = G, W + Q = H.

The signs , are replaced by = when the reactions are reversible. If the turbine, that produces W , is included in the system, then we must consider the heat Qa rejected to the ambient temperature Ta . In this case: W = 0, Q 1 Ta T H Ta S, Q Qa = H.
Fig. 2. Thermodynamic data for reaction (7), p = 1 bar.

All the energy input is heat, and in the reversible case, the turbine is supposed to have a Carnot efciency, which is too optimistic. Consequently we shall use the rst formulation. Hence, the heat requirement is: G
r

+ ( H G) G
r

if

H G>0

and

if

H G < 0.

In the latter case, the reaction is exothermic and the heat of reaction is lost, or can be recovered for internal use if T is high enough. 2.1. Thermodynamic data of the reactions The optimum pressures of the three sections of the cycle are not equal due to thermodynamic reasons. The temperature of section I must be higher than the melting point of iodine but not too high to avoid the occurrence of the reverse reaction. Therefore, we assumed that this reaction (1) is performed at 400 K and 2 bar. The pressure of section II must be low, due to the fact that the number of gaseous moles is increasing, therefore, the low pressure shifts the reactions towards completion. We have chosen 1 bar for this section. Finally, the best pressure for section III is 50 bar. The more the pressure of this section is high, easier is the distillation of HI, because, the position of the azeotrope depends upon the pressure. 50 bar seems a reasonably high value. Of course, pumping power is needed to adjust the pressures. It will be taken into account in the energy balance. The Bunsen reaction (T = 400 K, p = 2 bar) can be split into four elementary reactions: (SO2 )g + (2H2 O)1 (H2 SO4 )1 + H2 , (I2 )1 + H2 (2HI)g , (H2 SO4 )1 + (4H2 O)1 (H2 SO4 4H2 O)1 , (2HI)g + 8I2 + (10H2 O)1 2(HI 5H2 O)1 + 8I2 . (8) (9) (10) (11)

Fig. 3. Thermodynamic data for reaction (6), p = 1 bar.

Eqs. (8) and (9) are endothermic and their heats of reaction can be computed with standard thermodynamic tables [8]. One nds for (8): H = 53 kJ/mol, G = 95 kJ/mol

and for (9) H = 34 kJ/mol, G = 11 kJ/mol. Reactions (10) and (11) stand for the exothermic dissolutions of H2 SO4 and HI in water. Their heats of solution, computed using Engels strong acid model [9], are: H = 58 kJ/mol, G = 40 kJ/mol for (10), and H = 122 kJ/mol, G = 116 kJ/mol. In (11) iodine is supposed to have no inuence on the heat of solution. This assumption is probably wrong because evidence of ion solvation with iodine molecules has been demonstrated in several papers. Calabrese [10] for ex ample, shows experimentally that I3 ions exist in dilute solutions and theoretically that I2x H+ should exist as well. Unfortunately, there is no direct measurement of the heat of the Bunsen reaction. We shall then assume that we can add the heats of (4)(7), which gives 93 kJ/mol. The data for reactions (7), (6) and (3) are found for example in [8] and are represented Figs. 24. Given the fact that the sulphuric acid boils at 612 K at 1 bar and that the heat of vaporisation is 58 kJ/mol, we can draw the Q/T diagram of section II (Fig. 5): We assumed that the maximum helium temperature was 1144 K and the return to the reactor temperature was 760 K. The pinch point was taken equal to 17 K. We have now all the inputs to calculate the total heat request.

622

S. Goldstein et al. / International Journal of Hydrogen Energy 30 (2005) 619 626

Fig. 4. Thermodynamic data for reaction (3), p = 50 bar.

Reaction (3): According to Fig. 4, G = 12 kJ/mol, H = 24 kJ/mol, therefore the heat request is 12/0.5 = 24 kJ/mol. The two reactions occurring in the same reactor, we added the G and H , which results in G = 89 kJ/mol and H = 98 kJ/mol. The heat requirement is then: 89/0.5 + 98 89 = 187 kJ/mol. 2.1.3. Pumping power The pumping power (mechanical efciency of 0.75) to raise the pressure of SO2 from 1 to 2 bar is RT Ln (2)/0.75= 3 kW and the pumping power of section III (see next section) is 5.3 kW. The additional heat due to the pumping power is then (3+5.3)/0.5 = 17 kW. According to these gures, an upper bound for the efciency is: th = 286/(352 + 187 + 17) = 0.51. It can be noticed that the separation steps (Eqs. (2) and (4)), are the most energy consuming (98 and 199 kJ/mol).

Fig. 5. QT diagram of section II.

2.1.1. Section II Reaction (4): The temperature has a slight inuence on the heat requirement of this reaction. Therefore, we supposed that this reaction is the reverse of reaction (10). Hence, H = 58 kJ/mol and G = 40 kJ/mol. The reaction can be performed with 40 kJ of mechanical work and 18 kJ of heat. The heat demand is then: 18+40/0.5 = 98 kJ/mol. Reaction (5): Heat demand: 58 kJ/mol. Reaction (6): According to Figs. 3 and 5, the reaction is performed at G = 0, therefore, the heat demand is T S = 94.3 kJ/mol. Reaction (7): Due to the pinch of 17 K, the reaction occurs with a slightly positive G i.e: G = 4.5 kJ/mol and H = 97.6. The heat request is then: (97.6 4.5) + 4.5/0.5 = 102 kJ/mol. Hence, the total amount of heat request for this section is: 98 + 58 + 94.3 + 102 = 352.3 kJ/mol. 2.1.2. Section III Reaction (2) is the reverse of reaction (11), but occurs at higher temperature and pressure. At 500 K and 50 bar, we have H = 66 kJ/mol and G = 77 kJ/mol. Due to the fact that the mixing enthalpy depends upon the temperature, we cannot assume that the specic heats of the reactants and products are the same. Therefore we shall take H = 122 kJ/mol (opposite of reaction (11)) and G = 77 kJ/mol. Hence the heat requirement is: (122 77) + 77/0.5 = 199 kJ/mol.

3. Best estimate of the efciency Many estimations are found in the literature, either for the whole cycle or for sections II or III taken separately. The rst efciency estimation was published by GA [2] in 1981. H2 SO4 was separated from water by a constant pressure multi effect distillation and HI was separated from HIx by using phosphoric acid that had to be regenerated. The total efciency was 0.47. Other papers deal with the heat demand of section II. In a recent work GA [7] using a more accurate thermodynamic model, estimated the heat demand at 420 kJ/mol. The pressure of the distillation stages ranged from 35 to 0.07 bar. The reason of this variation is to recover the maximum heat for internal use in the rst stages of the distillation. Ozturk [5] proposed a owsheet where the produced oxygen is used to boil by direct contact the sulphuric acid. The heat demand was 441.5 kJ/mol. In an alternative owsheet, Schepers [6] proposed an increasing pressure multi effect distillation process. The heat demand could be decreased to 389 kJ/mol.

S. Goldstein et al. / International Journal of Hydrogen Energy 30 (2005) 619 626

623

Finally the author [4] found that with a simple distillation column a heat demand was 520 kJ/mol. The conclusion of these results is that there should be a compromise between efciency and complexity and therefore cost of this section and that the optimum design is not necessarily the one with the minimum heat request. In the following we shall take the GA value of 420 kJ/mol as a best estimate value. Much less results are available for section III, due to the lack of thermodynamic data of the HIx mixture. Roth and Knoche [3] calculated a reactive distillation column and external circuits and found an overall heat request of 237 kJ/mol. The research teams of JAERI use electrodialysis to concentrate HI in HIx before to distillate the enriched HIx , but no best estimate efciency calculation is yet published. The calculation of the reactive distillation of section III is very difcult because of the complex behaviour of the vapourliquid equilibrium of HIx . In the next section we describe briey the model used and we give preliminary results concerning the heat requirement of this section. 3.1. Available experimental data Experimental data are available for the binary systems I2 H2 O, HIH2 O and for the ternary HII2 H2 O. The I2 H2 O system has been studied by Kracek [11] in 1931. He measured the solubility of iodine in water and put in evidence a miscibility gap lying between the solidliquid equilibrium point at 385.3 K (very close to the melting point of iodine: 386.7 K) and an upper temperature of approximately 553 K. At 385.3 K, the light aqueous liquid contains a very small amount of iodine (0.05 mol %) and the heavy liquid 98 mol % of iodine. Total pressures of HIH2 O mixtures have been measured by Wster [12] and the mixing enthalpies by Vanderzee and Gier [13]. This system exhibits an azeotrope whose precise location depends on temperature and pressure. At ambient temperature, the azeotrope corresponds to a molar fraction of HI of almost 15%. For HI concentrations higher than the azeotrope, the vapour phase is very rich in HI, and for high temperatures (> 500 K), HI might be dissociated in the vapour phase in H2 and I2 . The kinetics of this reaction is in principle very slow but possible catalytic effects or hydrogen production in the liquid phase might appear as discussed by Berndhauser [14]. We shall assume in this paper that the decomposition reaction of HI takes place only in the vapour phase. These experimental results have been used by Engels [9] to t the binary interaction parameters of his model, using the Wilson model for the calculation of the activity coefcients. Detailed total pressure measurements for the HIH2 OI2 ternary system for a [HI]/[H2 O] ratio up to 0.19 and various iodine concentrations are found in Neumann [15]. A synthesis of these results has been published by Engels and Knoche [16]. Experimental location of borderlines of the ternary phase diagram are found in [2]. From these results, it appears that the liquid phase exhibits two miscibility gaps

which are the extensions in the ternary diagram of those found for the binaries I2 H2 O and HIH2 O. The I2 H2 O miscibility gap tends to disappear rapidly with increasing HI concentration probably because of the formation of polyio dides ions like I3 or I2 H+ . 3.2. Neumanns thermodynamic model The thermodynamic model for the H2 OHII2 H2 reactive-liquidliquidvapour system proposed by Neumann is based on the following assumptions: hydrogen is only present in the gas phase; the vapour phase is ideal, despite high pressures; the following solvation reaction is taken into account in the liquid phase [9]: mH2 O + HI {(mH2 O, H+ ) + I } with m = 5,

HI decomposition 2 HI H2 + I2 takes place in the gas phase only. The NRTL activity coefcient model is used to take into account the non-ideality of the liquid phase and the binary interaction parameters (including solvent-complex) have been estimated by Neumann [15] from the experimental data of Engels and Knoche [16] for the H2 OHII2 H2 system. The entire thermodynamical model includes solvation equilibrium, mass balances, HI decomposition equilibrium in vapour phase and vapourliquid-phase equilibrium. With such model, the phase diagrams can be calculated with condence on the left-hand side of the binary azeotropic H2 OHI or for low iodine content. Uncertainties remain for high iodine contents (> 20%) and high temperatures (> 540 K). The main features of this model have been published in [4]. We recall here the behaviour of the ternary mixture at 22 bar, which is the pressure chosen by Roth and Knoche [3] for their reactive distillation calculation. 3.3. Ternary system HII2 H2 O A fundamental aspect for the reactive distillation design is the inuence of I2 concentration on the maximum bubble temperature of the ternary system H2 OHII2 . The calculated results are presented on Fig. 6. Each curve corresponds to the bubble curve of a ternary system H2 OHII2 with constant I2 mole fraction varying in the interval: 1 (pure I2 ), 0 (binary H2 OHI). Thus the last curve on the right is the bubble curve of the binary H2 OHI with a maximum azeotrope at xHI = 0.133. The crest line, locus of the maximum bubble points, joins the pure I2 point (T = 359.17 K) to the binary azeotrope H2 OHI (T = 245.13 C). Fig. 7 shows the ternary diagram of the system H2 OHII2 at 22 bar, limited to the region of interest for the reactive distillation design. We have notably reported the

624

S. Goldstein et al. / International Journal of Hydrogen Energy 30 (2005) 619 626

Fig. 6. Inuence of I2 on the maximum bubble temperatures.

Fig. 7. Ternary diagram of the system H2 OHII2 .

singular points: pure I2 (origin) and pure H2 O; the binary azeotrope H2 OHI and the binary heteroazeotrope H2 OI2 . The distillation frontier is the projection of the crest shown in Fig. 6. Note that a linear approximation of this frontier is not correct. Finally, Fig. 7 shows the liquidliquidvapour domain for low HI content.

This diagram gives information on feasible distillation paths at 22 bar. Thus, as mentioned by Roth and Knoche [3], from the feed composition (HI/H2 O/I2 = 0.1/0.51/0.39), we could obtain pure iodine as residue and, by dissociation of HI, a distillate with H2 , the feed point being just on the right of the distillation frontier. The straight line in Fig. 7

S. Goldstein et al. / International Journal of Hydrogen Energy 30 (2005) 619 626

625

Fig. 8. Flowsheet of section III.

Table 1 Energy requirements and temperatures Component E1 E2 E3 Column boiler Column condenser Pump Separator Load (kW) 482 69.1 369.6 310 221 5.3 61.6 Tin (K) 393 584 548.4 584 548.4 548.4 Tout (K) 575.4 548.4 393 584 548.4 548.4

Table 2 Inlet/outlet streams compositions Stream Entry Exit 1 Exit 2 H2 O 10.2 0.987 9.21 HI 2 0.004 1.33 I2 7.8 0.117 8.02 H2 0 0.333 0 Total mol/s 20 1.441 18.56

represents the mass balance and gives the composition of the distillate we could obtain without chemical reaction (HI/H2 O = 0.17/0.83). As a matter of fact we see that the number of water moles to evaporate is signicant. Finally, we present Fig. 7 the liquid prole obtained by simulation of a reactive distillation column, with HI dissociation in the vapour phase. This prole follows very closely the distillation frontier. These preliminary simulation and thermodynamics results give the basis for the optimal design of the reactive distillation process [17]. 3.4. Flowsheet of section III The above Neumanns model has been implemented in the chemical engineering code PROSIM and we proceeded by trial and error to converge on the owsheet of Fig. 8 . The pump P raises the pressure to 50 bar and the valve V brings it back to 2 bar. The column has 25 stages, feed is at stage 18 (stage No 1 is the upper stage). Heat exchanger E2 cools the residue at the same temperature as the condenser: 548.4 K. The liquid vapour separator S condenses part of the steam and HI of the distillate. This additional condensation occurs because the composition of the distillate and the residue are different. The energy requirements of the main components are found in Table 1 and the stream compositions in Table 2.

The heat released by E2 and E3 can be partly recovered by E1. The amount of heat recovered depends on the temperature difference assumed between the primary and the secondary circuit, which depends on the size of the heat exchangers. To size the heat exchangers, the physical properties of the uids are needed, but are not known. A priori these physical properties are not very favourable. For example, the Prandtl number for liquid iodine is P r = 1500 at 400 K and P r = 1600 for pure HI at 230 K. Anyway, in this paper, the heat exchangers are not dimensioned and we assumed minimum temperature differences between primary and secondary sides of the heat exchangers of 10 and 20 K. The heat recovered can then be deduced from the QT diagrams of E1 and E2+E3+E4. We nd that, if the temperature difference is 10 K, then the net heat request is 37 kW, and if the temperature difference is 20 K, it is 63.5 kW. Due to the fact that the temperature difference between the condenser and the boiler is small (35.6 K), a water heat pump can be used to raise the temperature of the heat released at the condenser and to use it for the boiler. The total heat released at the condenser of the column and the separator is 221+61.6 = 282.6 kW at 548.4 K. The total heat request at the boiler of the column is 310 kW at 584 K. If we assume 5 K of temperature drop between the condenser of the column and the boiler of the heat pump, and 5 K between the condenser of the heat pump and the boiler of the column, then the heat pump has to transfer 282.6 kW from 543.4 K to a temperature of 589 K. Now, a water heat pump operating at these temperatures and with

626

S. Goldstein et al. / International Journal of Hydrogen Energy 30 (2005) 619 626

one mole of water circulating, can transfer 272.3 kW of heat with a mechanical work of 37.1 kW (assuming 0.75 mechanical efciency for the pump). In our case we need then 282.6/272.3 = 1.04 mol of water. The heat recovered is then 1.04(272.3 + 37.1 0.75) = 312 kW, which is almost the heat required at the boiler of the column. The total heat demand for one mol of H2 produced is nally: (37 + (37.1 1.04 + 5.3)/0.5)/0.333 = 375 kW if the temperature difference is 10 K and (63.5+(37.11.04+ 5.3)/0.5)/0.3331 = 454 kW if the temperature difference is 20 K. It can be noticed that in this owsheet, the number of mol/s at the entry is exactly that at the exit of the Bunsen reaction. The H2 production being only 0.333 mol/s, it means that an important quantity of materials must be recirculated between sections III and I. Typically, 20/0.333 = 60 mol/s must enter section III. The head losses of such a mass ow rate have not been taken into account. The best estimate of the efciency with todays knowledge and a temperature difference of 10 K for the heat recovery between E1 and E2, E3 is then 286/(375+420) = 0.36. If the temperature difference is 20 K, the efciency becomes 286/(454+420) = 0.33. 4. ConclusionR&D needs With the current reactions (1)(7), widely adopted for the iodine sulphur cycle, we found an upper bound of the efciency of 0.51. This value has been found by calculating the heat and work requirements of the reactions, assuming that they were reversible and a maximum temperature of the helium of 1144 K. A best estimate could be calculated as well by taking the latest GA value for the heat request of section II together with a reactive distillation calculation of section III. We found a value comprised between 0.33 and 0.36, depending on the temperature difference assumed for the heat recovery between E2+E3 and E1. These gures will be more accurately specied in the future when the heat exchangers will be sized. They do not take into account the head losses due to the large material circulation. Many improvements can be imagined to rise both these gures: Reduction of the amounts of iodine and water in the Bunsen reaction. We have seen that these excesses of iodine and water burden heavily the heat balances of both sections II and III. Use of membranes at various steps of the cycle: To separate SO2 from O2 , to separate liquid H2 O from H2 SO4 , to separate HI from HIx (electrodialysis or distillation membranes?) and nally to separate H2 from (HI+ H2 O)g . Many teams, in particular in Japan, currently work actively on

these topics. Membranes can be foreseen if their permeability and selectivity are high enough. Finally the HIx section still represents a real chemical engineering challenge. Additional thermodynamic measurements are necessary and the best distillation design has to be found. Further experiments are planned, to be performed by the Commissariat lEnergie Atomique (CEA), in order to improve the database of the ternary mixture, more specifically at high iodine contents and to measure the vapour partial pressures as well.

References
[1] Funk JE. Thermochemical hydrogen production: past and present. Int J Hydrogen Energy 2001;26:18590. [2] Norman JH, Besenbruch GE, Brown LC, OKeefe DR, Allen CL. Thermochemical water splitting cycle. Bench scale investigations and process engineering. DOE/ET/26225, 1981. [3] Roth M, Knoche KF. Thermochemical water splitting through direct HI decomposition from H2 OHII2 solutions. Int J Hydrogen Energy 1989;14(8):5459. [4] Goldstein S, Borgard JM, Joulia X, Guittard P, Thermodynamic aspects of the iodine_sulphur thermochemical cycle, AIChE, Spring national meeting, New Orleans, April 2003. [5] ztrk IT, Hammache A, Bilgen E. Trans IChemE, Vol 72, Part A, March 1994. [6] Knoche KF, Schepers H, Hesselmann K, WHEC 1984. [7] Brown LC, Funk JF, Showalter SK. High efciency generation of hydrogen fuels using nuclear power. GA, A23373, June 2003. [8] Barin I, Knache O. Thermodynamic properties of inorganic substances. Berlin: Springer. [9] H. Engels, Phase equilibria and phase diagrams of electrolytes. Chemistry data series. Dechema. [10] Calabrese V. Polyiodine and polyiodide species in an aqueous solution of iodine +KI. Theoretical and experimental studies. J Phys Chem A 2000;104:128792. [11] Kracek FC. Solubilities in the system wateriodine to 200 C. J Phys Chem 1931;35:417. [12] Wster G. Dissertation, RWTH, Aachen 1979. [13] Vanderzee C, Gier L. The enthalpy of solution of gazeous hydrogen iodide in water and the relative apparent molar enthalpies of hydriodic acid. J Chem Thermodyn 1974;6:441. [14] Berndhauser C, Knoche KF. Experimental investigations of thermal HI decomposition from H2 OHII2 solutions. Int J Hydrogen Energy 1994;19(3):23944. [15] Neumann D, Diplomaufgabe RWTH Aachen, 28-01-1987. [16] Engels H, Knoche KF. Vapour pressures of the system HI/H2 O/I2 and H2 . Int J Hydrogen Energy 1996;11(11): 7037. [17] Thery R. New methodology for the feasibility analysis and the design of reactive distillation processes. INPT PhD thesis, 2002.