International Journal of Hydrogen Energy 32 (2007) 531 536 www.elsevier.

com/locate/ijhydene

Experimental study of two phase separation in the Bunsen section of the sulfuriodine thermochemical cycle
A. Giaconia, G. Caputo, A. Ceroli, M. Diamanti, V. Barbarossa, P. Tarquini, S. Sau
ENEA, Casaccia Research Center, via Anguillarese, 301, 00060 S. M. di Galeria, Rome, Italy Available online 25 September 2006

Abstract The Bunsen reaction for the production of hydriodic and sulfuric acids from water, iodine and sulfur dioxide, has been studied with the evaluation of the effect of some operative parameters on product phase behavior. Results show that operative temperature has a minor effect on the phase behavior. In contrast, both iodine and water loads can be adjusted to enhance the downstream operations of the sulfuriodine thermochemical water-splitting cycle: the effect of iodine and water excess on resulting phases purity, side reaction occurrence and acid concentration was studied and, then, the most favorable operative conditions dened. 2006 International Association for Hydrogen Energy. Published by Elsevier Ltd. All rights reserved.
Keywords: Hydrogen production; Sulfuriodine thermochemical cycle; Bunsen reaction

1. Introduction In the prospective of a future energy scenario based on the intensive use of hydrogen as energy vector, sustainable production routes essentially based on the utilization of renewable raw materials and alternative (non-fossil) primary energy sources must be considered for the long term. Hence, the evaluation of all the viable processes for hydrogen production from water is a must and, in principle, thermochemical cycles one of the most interesting options. To date, the most widely studied thermochemical watersplitting cycle is the sulfuriodine (SI) cycle, consisting of the following three reactions involving sulfuric and hydriodic acids as intermediates: 2H2 O2 + I2 + SO2 H2 SO4 + 2HI, 2HI I2 + H2 ,
1 H2 SO4 H2 O + SO2 + 2 O2 .

(1) (2) (3)

Corresponding author. Tel.: +39 06 3048 6334; fax: +39 06 3048 6779.

E-mail address: salvatore.sau@casaccia.enea.it (S. Sau).

Reaction (1), called Bunsen reaction, provides water-splitting into two acids (hydriodic and sulfuric) via reaction with iodine and sulfur dioxide. Afterward, produced HI and H2 SO4 are separated and sent to the concentration/decomposition sections for conversion to H2 and O2 by reactions (2) and (3), respectively (Fig. 1). Intense experimental and theoretical research activity has been carried out worldwide during the last decade to investigate the potentialities of this process [13] and, as a matter of fact, the scientic feasibility demonstration of the SI chemical cycle has been recently achieved with a continuous bench scale loop [4,5]. Nevertheless, further effort is still required for process advancement and optimization, in order to evaluate the technological feasibility and the competitiveness of the process (energetic efciency, costs, etc.). From the engineering point of view, the development of the three sections in which the whole cycle can be divided (Fig. 1) imposes the resolution of different problems. Hence, the sulfuric acid concentration/decomposition section (referred to as sulfuric section) is critical for the highest temperatures involved (8001000 C), with consequent issues dealing with material corrosion and the coupling with the primary high temperature energy source, as well as the study of catalyst activity and stability [610]. Material resistance issues are also important for the development of the hydriodic acid concentration/decomposition section

0360-3199/$ - see front matter 2006 International Association for Hydrogen Energy. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.ijhydene.2006.08.015

532

A. Giaconia et al. / International Journal of Hydrogen Energy 32 (2007) 531 536

heat (600-950C) Sulfuric section H2SO4 SO3 + H2O H2O + SO2 + O2 SO2/H2O H2O H2SO4/H2O Bunsen section 2H2O + I2 + SO2 H2SO4 + 2HI I2/H2O 2HI I2 + H2 HIx section heat (300-450C)
Fig. 1. Schematic of the sulfuriodine thermochemical water-splitting cycle.

O2

HI/H2O H2

(for reaction 3) [10]; furthermore, in this case, reliable thermodynamic data for HI/I2 /H2 O mixtures are required for plant and process design, while the identication (and subsequent development) of the most convenient route has not been realized: yet, process schemes based on reactive distillation, on HI concentration by extractive distillation, electro-electrodialysis (EED) or more conventional distillation operations, and on the utilization of permselective membrane reactors, have been proposed [1114]. Bunsen reaction (1), on the other hand, can be carried out at relatively mild temperatures (< 120 C) to obtain high yields with a very small energy input (the reaction itself is exothermic in water media). However, since this section is interfaced to both the sulfuric and hydriodic concentration/decomposition sections (Fig. 1), it can be considered as the key step of the SI thermochemical cycle. Therefore, the operative conditions for this reaction must be optimized in order to reduce the possibility of side reactions (side products must be avoided especially for such a cyclic chemical process) and to obtain two concentrated acid streams to ease the operations in the other sections. In practice, water is reacted with I2 and SO2 to obtain hydriodic and sulfuric acids that can split into two liquid phases: a sulfuric phase (containing sulfuric acid and water) and a HI phase (containing HI with residual iodine and water). Some experimental and theoretical studies reveal that a large iodine excess (x) is needed to enhance phase (product) separation and to avoid side reactions [1517]. Hence, the hydriodic acid phase will be referred to as HIx phase and this chemical process step can be described as follows: (x + 1)I2 + SO2 + (n + 2)H2 O [H2 SO4 + (n m)H2 O]Sulfuric phase + [2HI + (x)I2 + mH2 O]HIx phase . The larger the water excess (n), the higher are the resulting capital and operative expenses, and the lower is the energetic efciency of the entire chemical loop because of the higher heat duty for the concentration of both acids prior decomposition. A reduction in the iodine excess (x), on the other hand, deter-

mines benets on the management of the HIx section due to the lower iodine recirculation ow rates. For this reason, one of the major objective of the work presented here is to select the most suitable process conditions for the Bunsen reaction to obtain two separated liquid streams as pure and concentrated as possible in the two acids (sulfuric and hydriodic) to facilitate the processes in the downstream sections: rst, I2 and HI traces into the sulfuric acid must be minimized because they can lead to poisoning of the catalyst used for sulfuric acid decomposition [7]; on the other hand, sulfates in the HIx phase can lead to unwanted reactions with HI with possible sulfur or H2 S formation. 2. Materials and methods Phase behavior of the four-component H2 SO4 /HI/H2 O/I2 mixtures was investigated using a 50 mL jacketed quartz test tube whose temperature was controlled by means of a water/ ethylene glycol recirculation system and a thermostatic bath (Julabo). A schematic of this simple apparatus is shown in Fig. 2. First, a selected amount of iodine (>99.9%, Aldrich) was added to a 57 wt% HI/H2 O liquid solution (Aldrich) past heating. After complete iodine dissolution, sulfuric acid

thermostat

sulfuric acid phase

hydroidic acid phase (HIx)

Fig. 2. Apparatus used for liquidliquid phase behavior tests.

A. Giaconia et al. / International Journal of Hydrogen Energy 32 (2007) 531 536

533

Table 1 Effect of temperature and Iodine concentration: water excess n = 11.2 mol/molH2 SO4 (constant), iodine excess (x ) variable from 3.9 to 16.0 mol/molH2 SO4
Entry T C Feed compositiona HI 1 2 3 4 5 6 7 80 95 120 0.113 0.106 0.113 0.106 0.113 0.093 0.069 H2 O 0.614 0.572 0.614 0.572 0.614 0.503 0.371 I2 0.217 0.271 0.217 0.271 0.217 0.359 0.527 H2 SO4 0.055 0.051 0.055 0.051 0.055 0.045 0.033 Sulfuric phase compositiona HI 0.004 0.003 0.005 0.004 0.004 0.004 0.001 H2 O 0.84 0.85 0.86 0.85 0.86 0.84 0.79 I2 0.001 <0.001 0.001 <0.001 0.001 0.001 <0.001 H2 SO4 0.15 0.15 0.14 0.15 0.14 0.15 0.21 density (g/cm3 ) 1.3 1.3 1.3 1.3 1.3 1.3 1.4 HIx compositiona HI 0.15 0.14 0.16 0.14 0.16 0.12 0.08 H2 O 0.53 0.48 0.50 0.48 0.51 0.40 0.30 I2 0.30 0.37 0.32 0.37 0.32 0.47 0.62 H2 SO4 0.02 0.02 0.02 0.02 0.02 0.01 0.01 density (g/cm3 ) 3.2 3.3 3.0 3.2 3.2 3.4 3.6

a Composition expressed in molar fractions. (maximum experimental error on molar fractions rounded to the 2nd and 3rd decimal place is 0.01 and 0.002, respectively).

(98 wt%, Aldrich) and the residual amount of water were slowly added, always at 80 C, and afterwards the mixture was heated to the desired operative temperature. The system was then energetically mixed and, afterwards, kept at rest for at least 1 h to allow complete phase separation and equilibrium achievement. The absence of solid iodine in the bottom was checked by immersion of a glass bar. Although both liquid phases appeared dark-red colored in the presence of iodine, it was always quite easy to locate the phase boundary. Thus metered volumes of the supernatant sulfuric phase were sampled, and then weighted and analyzed. Iodine concentration was determined by titration with a sodium tiosulfate (Na2 S2 O3 , Aldrich) standardized solution, while HI and sulfate content determined by ionic chromatography, using a Metrohm 761 Compact Ionic Chromatograph, with an anion column (Metrosep A SUP 4). The eluent solution was 1.8 mM in Na2 CO3 and 1.7 mM in NaHCO3 . Several samplings and analyses were made for each test to assess the repeatability of measurement and the effective physical equilibrium attainment, as well as to estimate the experimental accuracy. The composition of the lower HIx phase was at last determined by mass balance. Finally, the upper sulfuric phase was quantitatively removed and the residual HIx phase weighted for density measurement and, then, washed with a KI solution and ltered in order to attest and quantify the possible solid sulfur formation. 3. Results and discussion In Table 1 the composition and densities of both phases (HIx and Sulfuric) at different temperatures (80, 95, 120 C) and overall iodine concentration are reported. Results are expressed in term of molar fractions. The corresponding plots are represented in Fig. 3 as molar ratios (Z) of HI and H2 SO4 in the HIx and sulfuric phases, relative to the total minus water. As a result of the much higher iodine solubility in the HIx phase, the density of the lower phase was rather sensitive to the iodine excess (x), ranging from 3.0 to 3.6 g/mL, while the density of the sulfuric phase is always in the 1.31.4 g/cm3 range regardless the iodine overall concentration (Table 1); the

0.40 0.30

HIx phase

feed Z(HI) 0.20 0.10 0.00 0.0 0.2 0.4 0.6 0.8 Sulfuric phase

Sulfuric phase 80C Sulfuric phase 95C Sulfuric phase 120C HIx phase 80C HIx phase 95C HIx phase 120C feed 80C feed 95C feed 120C

1.0

Z(H2SO4)
Fig. 3. Liquidliquid phase behavior: effect of iodine excess. Compositions expressed as molar ratios (Z) of the two acids relative to the total minus water.

sulfuric/HIx phases volume ratio decreased from 3 up to ca. 10 as the iodine excess (x) was increased. Owing to solubility considerations, the higher is the temperature the larger is the amount of iodine that can be loaded to the system; therefore, it was possible to operate at the highest temperature level (120 C) with a large amount of iodine (I2 overall molar fraction of about 0.53). On the other hand, temperature at equivalent iodine content was found to have an insignicant effect on the composition of the two phases (Table 1 and Fig. 3). As shown in Fig. 3, as the iodine load was increased both phases became purer in the two acids, regardless the temperature, in agreement with what elsewhere reported [16]. In this case also the phase behavior at temperatures as high as 120 C was investigated. Particularly, less than a few unit percent of I2 and HI actually introduced in the vessel were transferred into the sulfuric phase: this amounts to < 4 mol% of HI and <1 mol% of I2 relative to the H2 SO4 in the sulfuric phase (Fig. 4 ). Hence, a good segregation of the iodine-containing compounds was achieved under these conditions. A quite different situation regards the amount of H2 SO4 transferred into the HIx phase: in all experiments the H2 SO4 initially loaded and lastly transferred into the HIx phase was

534

A. Giaconia et al. / International Journal of Hydrogen Energy 32 (2007) 531 536

0.04 (HI/H2SO4)Sulfuric phase 0.03 mol/mol 0.02 0.01 0.00

(I2/H2SO4)Sulfuric phase

80C 95C 120C

0.010 0.008 mol/mol 0.006 0.004 0.002 0.000

80C 95C 120C

4 8 16 iodine excess, x (mol/molH SO ) 2 4

16

iodine excess, x (mol/molH2SO4)

Fig. 4. Effect of iodine excess on HI and I2 concentration in the sulfuric phase. Water excess n = 11.2 mol/molH2 SO4 (constant).

0.15

80C 95C

0.12 (H2SO4/HI)HIx phase

120C

mol/mol

0.09

0.06

0.03

0.00 4 8 iodine excess, x (mol/molH2SO4)

Fig. 5. Effect of iodine excess on H2 SO4 concentration in the HIx phase. Water excess n = 11.2 mol/molH2 SO4 (constant).

16

the order of 1525% (usually over 20%), with only a slight decrease obtained with the increase in I2 excess (x). In other words, only at the largest I2 excess (molar fraction of 0.53 at 120 C, with x = 16) the amount of H2 SO4 relative to HI in the HIx phase was the order of 7 mol%, being in the 915 mol% (relative to HI) in the other cases (Fig. 5). This behavior is of major concern when we deal with purication of the HIx phase from H2 SO4 . Moreover, unless operating with a huge I2 load that would nally lead to high xed and operative costs, such an unfavorable partitioning of the H2 SO4 practically determines a lower effective SO2 conversion than what predictable on the basis of reaction thermodynamics. In fact it is reasonable that the H2 SO4 will be recovered from the HIx phase by reversing the Bunsen reaction and, therefore, reduced to SO2 and recycled back to the Bunsen reactor. Maximum experimental error of 0.01 and 0.002 on molar fractions reported in Table 1 was estimated on values rounded to the 2nd and 3rd decimal place, respectively: more precise data can be obtained only with more suitable but even much

more complex sampling and analytical tools. Nevertheless, the results of this semi-quantitative analysis have been conrmed by repeating each experiment more than a few times. The effect of temperature on the H2 SO4 partitioning is unclear. As far as the concentration of both acids in the corresponding phases is concerned, we observed that under the adopted conditions H2 SO4 concentration in the sulfuric phase was in the 4750 wt%, except for the experiment at the largest I2 content (molar fraction of 0.53 at 120 C), where measured H2 SO4 concentration in the sulfuric phase was about 59 wt%. The HI/H2 O ratio in the lower phase did not seem to vary sensibly with the iodine content, being around 0.3 mol/mol. According to the results recently reported [17], the concentration of the HI in the HIx phase here determined should be higher than the corresponding predicted value for the azeotrope of HI/I2 /H2 O ternary mixtures at about 22 bar: hence, after quantitative removal of H2 SO4 from the HIx phase, the recovery of pure or highly concentrated HI for the subsequent decomposition step could be performed by high pressure distillation operations, signicantly reducing the management costs of the downstream HIx section. The effect of water excess on liquidliquid phase behavior was also studied by carrying out a set of experiments at different overall water content in the feed, and results are represented in: composition (molar fractions) and densities of both phases (HIx and sulfuric) at two different temperature levels (80 and 95 C) are reported. The increased water content did not affect the extent of contamination of both phases except for a slight increase of I2 and HI impurities in the sulfuric phase, as shown in Fig. 6. Similarly to what earlier determined, sulfuric phase composition comprised <5 mol% of HI and <1.5 mol% of I2 relative to the H2 SO4 . Again, the amount of H2 SO4 initially loaded and lastly transferred into the HIx phase was the order of 1525%, with a concentration in the HIx phase in the 814 mol% range (relative to HI), and almost not affected by the water addition (Fig. 7). As expected, a certain dilution of both phases was observed when water overall concentration was increased: when the water excess (n) was increased from around 11 to 16 mol/mol,

A. Giaconia et al. / International Journal of Hydrogen Energy 32 (2007) 531 536

Table 2 Effect of temperature and water concentration: I2 excess x = 3.9 mol/molH2 SO4 (constant), water excess (n) variable from 11.2 to 15.8 mol/molH2 SO4
Entry T C Feed compositiona HI 1 8 3 9 10 80 95 0.113 0.090 0.113 0.100 0.090 H2 O 0.614 0.695 0.614 0.659 0.695 I2 0.217 0.172 0.217 0.192 0.172 H2 SO4 0.055 0.044 0.055 0.049 0.044 Sulfuric phase compositiona HI 0.004 0.006 0.005 0.005 0.006 H2 O 0.84 0.86 0.86 0.87 0.88 I2 0.001 0.001 0.001 0.001 0.002 H2 SO4 0.15 0.13 0.14 0.12 0.12 density (g/cm3 ) 1.3 1.3 1.3 1.3 1.2 HIx compositiona HI 0.15 0.12 0.16 0.15 0.12 H2 O 0.53 0.63 0.50 0.55 0.62 I2 0.30 0.23 0.32 0.29 0.24 H2 SO4 0.02 0.01 0.02 0.01 0.02

535

density (g/cm3 ) 3.2 2.8 3.0 3.0 2.8

expressed in molar fractions. (maximum experimental error on molar fractions rounded to the 2nd and 3rd decimal place is 0.01 and 0.002, respectively).
a Composition

0.014 0.05 (HI/H2SO4)Sulfuric phase mol/mol 0.04 0.03 0.02 0.01 0.00 11 13 16 ) 2SO4 water excess, n (mol/molH 80C 95C (I/H2SO4)Sulfuric phase mol/mol 0.012 0.010 0.008 0.006 0.004 0.002 0.000 11 13 16 water excess, n (mol/molH SO )
2 4

80C 95C

Fig. 6. Effect of water excess on HI and I2 concentration in the sulfuric phase. I2 excess x = 3.9 mol/molH2 SO4 (constant).

0.15 0.12 (H2SO2/HI)HIx phase 0.09 0.06 0.03 0.00

80C 95C

11

13

16

water excess, n (mol/molH2SO4)

Fig. 7. Effect of water excess on H2 SO4 concentration in the HIx phase. I2 excess x = 3.9 mol/molH2 SO4 (constant).

higher dilution of both acid streams lowers the energy efciency of the cycle and increase the xed and operative costs of the process, results obtained suggest to minimize the water excess in the Bunsen reaction, with respect of the reaction thermodynamics and kinetics constraints. Finally, from the results described above (Tables 1, 2) we can recognize that temperature has not a signicant effect on liquidliquid phase behavior in the 80120 C range and under the adopted composition. Moreover, no sulfur formation was observed after all experiments provided that the sulfuric acid was slowly added when a large enough amount of I2 was dissolved. This last result is in agreement with some earlier studies about side reaction occurrence in the Bunsen reaction [15]. 4. Concluding remarks Liquidliquid phase behavior of products from the Bunsen reaction in the SI thermochemical cycle has been investigated. Our results show that: (1) Temperature and I2 content have little effect on the sulfuric acid and HI concentrations in the two phases. Increasing water excess (n) leads to a dilution of the acids (as expected), whereas HI concentration in the lower phase is always well over the azeotrope point [17], provided that distillation is carried out at pressures of at least 22 bar.

H2 SO4 concentration in the sulfuric phase decreased from ca. 50 to 41 wt%, and the HI/H2 O molar ratio in the HIx phase from 0.33 to 0.19. The HIx /sulfuric phase volume ratios were rather unaffected by temperature and water excess, being in all experiments in the 34 range (Entries 1,3,8,9,10). Hence, since the addition of water did not lead to benets on the degree of pureness of both phases, whilst the subsequent

mol/mol

536

A. Giaconia et al. / International Journal of Hydrogen Energy 32 (2007) 531 536

(2) Purity of both phases is almost unaffected by iodine content and temperature, unless we use a very huge amount of iodine at 120 C; also in this latter case we found small practical improvements. Iodide and HI in sulfuric phase are always 12% of the total, while sulfates in hydriodic phase are always more than 15% of the total. (3) We never observed secondary reactions in our experiments, provided that the components are initially mixed at a temperature not over 80 C. According to the work of Sakurai et al. [15], in order not to have secondary reactions iodine excess (x) should be larger than 3.4 molI2 /molH2 SO4 . In our experiments with the lowest iodine content we had x = 3.9 molI2 /molH2 SO4 , and so, taking into account a security gap, we are close to the limit previously indicated. For these reasons we can assess that this is the smallest iodine initial molar fraction suitable for the Bunsen reaction. The only alternative is to increase the water excess (n) but, as mentioned above, this drastically decreases the acids concentration. (4) The reaction temperature of 80 C seems a good compromise between the exothermicity of the reaction and iodine dissolution. Furthermore, the condition x > 3.4 molI2 /molH2 SO4 on the I2 excess must be respected [15]. However, if one lowers the mixture temperature the iodine content approaches the saturation point [16], and uctuations of the mixture composition can produce formation of solid iodine. Hence, a temperature of at least 80 C seems a good choice to avoid this problem. It follows that a temperature of 80 C and an initial molar fractions of HI/H2 O/I2 /H2 SO4 = 0.11/0.61/0.22/0.06 (Entry 3) is the best choice for us. This way we can carry out this exothermic reaction at a relatively mild temperature and using the lowest iodine excess (x). Other parameters do not produce signicant improvements either in the phase purity or in the acids concentration. A preliminary experiment concerning the Bunsen reaction under these conditions has already been carried out (this work will be completed within the next months) and the results indicate fast formation of the two phases with the expected compositions. In order to implement a bench scale reactor for the Bunsen part of the SI cycle, further studies must be performed, especially on the viscosity of the two phases, the iodine rate of dissolution at several temperatures, and materials resistance. Finally, our future work will be mainly dedicated to phase purication, especially with regard to the hydriodic phase.

References
[1] Norman JH, Besenbruch GE, Okeefe DR. Thermochemical water splitting for hydrogen production. GRI-80/0105, 1981. [2] Brown LC, Besenbruch GE, Lentsch RD, Schultz KR, Funk JF, Pickard PS, et al. High efciency generation of hydrogen fuels using nuclear power. General Atomic Project 30047, June 2003. [3] Kasahara S, Hwang G-J, Nakajima H, Choi H-S, Onuki K, Nomura M. Effects of process parameters of the IS process on total thermal efciency to produce hydrogen from water. J Chem Eng Jpn 2003;36:88799. [4] Kubo S, Nakajima H, Kasahara S, Higashi S, Masaki T, Abe H. et al. A demonstration study on a closed-cycle hydrogen production by the thermochemical water-splitting iodinesulfur process. Nucl Eng Des 2004;233:34754. [5] Kubo S, Nakajima H, Shimizu S, Onuki K, Hino R. A bench scale hydrogen production test by the thermochemical water-splitting iodinesulfur process. Proceedings of GLOBAL 2005, Tsukuba, Japan. 913 October 2005. [6] Dokiya M, Kameyama T, Fukuda K, Kotera Y. The study of thermochemical hydrogen preparation. III. An oxygen-evolving step through the thermal splitting of sulfuric acid. Bull Chem Soc Jpn 1977;50:265760. [7] OKeefe DR, Norman JH, Williamson DG. Catalysis research in thermochemical water-splitting processes. Catal RevSci Eng 1980;22(3):32569. [8] Barbarossa V, Brutti S, Diamanti M, Sau S, De Maria G. Catalytic decomposition of sulfuric acid in sulfuriodine cycle for hydrogen production. Int J Hydrogen Energ 2006;31:88390. [9] Ginosar DM, Glenn AW, Petkovic LM. Stability of sulfuric acid decomposition catalysts for thermochemical water splitting cycles. Proceedings of the 2005 AIChE spring national meeting. Atlanta (Georgia, USA), 1014 April 2005. [10] Onuki K, Ioka I, Futakawa M, Nakajima H, Shimizu S, Tayama I. Screening tests on materials of construction for the thermochemical IS process. Corr Eng 1997;46:1419. [11] Kasahara S, Kubo S, Onuki K. Nomura thermal efciency evaluation of HI synthesis/concentration procedures in the thermochemical water splitting IS process. Int J Hydrogen Energ 2004;29:57987. [12] Hwang G-J, Onuki K, Shimizu S, Ohya H. Hydrogen separation in H2 H2 OHI gaseous mixture using the silica membrane prepared by chemical vapor deposition. J Membrane Sci 1999;162:8390. [13] Hwang G-J, Onuki K. Simulation study on the catalytic decomposition of hydrogen iodide in a membrane reactor with a silica membrane for the thermochemical water-splitting IS process. J Membrane Sci 2001;194:20715. [14] Caputo G, Felici C. Membrane distillation of H2 O/HI mixtures for the sulfuriodine thermochemical process. Proceedings of the second European hydrogen energy conference. Zaragoza (Spain). 2225 November 2005. [15] Sakurai M, Nakajima H, Amir R, Onuki K, Shimizu S. Experimental studies on side-reaction occurrence condition in the iodinesulfur thermochemical hydrogen production process. Int J Hydrogen Energ 2000;25:6139. [16] Sakurai M, Nakajima H, Onuki K, Shimizu S. Investigation of 2 liquid phase separation characteristics on the iodinesulfur thermochemical hydrogen production process. Int J Hydrogen Energ 2000;25:60511. [17] Goldstein S, Borgard J, Vitart X. Upper bound and best estimate of the efciency of the iodine sulphur cycle. Int J Hydrogen Energ 2005;30: 61926.