Elsevier

Chemical Engineering Science

Volume 63, Issue 18, September 2008, Pages 4596-4604
Chemical Engineering Science

Hydrodynamic gas–solid model of cupric chloride particles reacting with superheated steam for thermochemical hydrogen production

https://doi.org/10.1016/j.ces.2008.07.003Get rights and content

Abstract

This paper examines the transport phenomena of a non-catalytic reaction of cupric chloride particles with superheated steam in a fluidized bed, as part of a copper–chlorine (Cu–Cl) thermochemical cycle for nuclear-based hydrogen production. As both cupric chloride and steam participate in the chemical reaction, it is necessary to develop a new model that predicts the conversion of cupric chloride particles, as well as steam. This incorporates features of a uniform reaction model (Volumetric Model; VM) and a Shrinking Core Model (SCM). Due to little or no experimental data available for the hydrodynamics and chemistry of the reaction, the above two models are considered as limiting cases. Separate numerical solution procedures are developed to monitor the effects of various parameters on the conversion of CuCl2 particles and steam. Also, the new solution algorithms are used to predict outputs for a typical bench-scale reactor and operating conditions. From the numerical results, under the assumption of VM or SCM, the conversion of steam decreases with superficial velocity, whereas the conversion of solid particles increases. Also, a higher bed inventory leads to higher conversion of both reactants. SCM predicts higher values for the reactant conversions, compared to VM. The new solution procedures may be utilized for parametric studies that observe the effects of different process parameters on the fluidized bed performance.

Introduction

Hydrogen as a fuel is frequently identified as a major solution to the climate change problem of greenhouse gases emitted by fossil fuels. However, a major portion of the world's hydrogen production is dependent on fossil fuels. The predominant existing process for large-scale hydrogen production is steam methane reforming (SMR). SMR is a carbon-based technology that emits carbon dioxide. In contrast, nuclear energy can be used for large-scale zero-carbon production of hydrogen (Rosen et al., 2006), through electrolysis or thermochemical cycles (Yildiz and Kazimi, 2006). Thermochemical “water splitting” requires an intermediate heat exchanger between the nuclear reactor and hydrogen plant, which transfers heat from the reactor coolant to the thermochemical cycle (Forsberg, 2003). Operating temperatures are key factors for thermochemical hydrogen production; so optimization of heat flows is important for high energy conversion efficiency (Naterer, 2002).

Many types of thermochemical cycles exist for hydrogen production. The sulfur–iodine (S–I) cycle (hydrogen sulfide, iodine–sulfur, sulfuric acid–methanol) and the Br–Ca–Fe cycle are prominent examples (Forsberg, 2003). Thermochemical cycles have received considerable attention because current estimates indicate that hydrogen costs could be as low as 60% of those from room-temperature electrolysis (Forsberg, 2003). Kasahara et al. (2003) reported a maximum thermal efficiency of 56.8% for the S–I process to produce hydrogen, after optimization of various parts of the cycle. Kasahara et al. (2004) investigated the effects of operation parameters of HI synthesis and concentration on the thermal efficiency of the S–I process, based on heat/mass balances. Nomura et al. (2004) successfully employed the Bunsen reaction (SO2+I2+2H2O=H2SO4+2HI) in the thermochemical S–I process to produce hydrogen with an electrochemical membrane reactor. The effects of three typical membrane techniques, i.e., electro-electrodialysis (EED), electrochemical cell (EC) and a hydrogen permselective membrane reactor (HPMR), on the total thermal efficiency were reported by Nomura et al. (2005). A flowsheet evaluation of the S–I cycle using heat from Japan's first high-temperature gas-cooled reactor HTTR, with a hydrogen production rate of 1100Nm3/h, was presented by Sakaba et al. (2006). Kasahara et al. (2007) performed a continuous and stable operation of the bench-scale S–I apparatus, including a preliminary screening of corrosion resistant materials and a ceramic heat exchanger for the H2SO4 vaporizer.

In the U.S., recent advances at the Oak Ridge National Laboratory (Forsberg et al., 2003) used a molten-salt advanced high-temperature reactor (AHTR) as a reactor concept for very high-temperature heat (7501000C) needed by thermochemical production of hydrogen via the S–I cycle. Yildiz et al. (2006) examined hydrogen production using high-temperature steam electrolysis (HTSE) supported by a supercritical CO2(SCO2) recompression Brayton cycle that was directly coupled to an advanced gas-cooled reactor (AGR). The Savannah River National Laboratory (SRNL) has been developing a hybrid sulfur (HyS) cycle since 2004 (Summers and Gorensek, 2006). McLaughlin et al. (2006) estimated that process heat between about 776 and 950 °C was needed by the HyS cycle to perform the decomposition of SO3 to O2 and SO2, using a sulfuric acid separation system as the high temperature heat exchanger. Smith et al. (2006) presented a probabilistic safety assessment to answer the risk-related questions for a combined nuclear and thermochemical hydrogen plant. The Argonne National Laboratory (Lewis et al., 2005) is developing low temperature thermochemical cycles at 550 °C or less, including a hybrid copper-chlorine (Cu–Cl) cycle.

In Canada, Rosen and Scott (1998) performed comparisons, based on energy and exergy analyses, of a wide range of hydrogen production processes, including processes which are hydrocarbon-based (SMR and coal gasification), non-hydrocarbon-based (water electrolysis and thermochemical water decomposition) and integrated (SMR linked to non-hydrocarbon-based processes). Atomic Energy of Canada Ltd. (AECL) (Sadhankar et al., 2006) identified the copper–chlorine (Cu–Cl) cycle as a promising thermochemical cycle, with a heat-to-hydrogen efficiency of about 44% (Lewis, 2007, Serban et al., 2004). Water is decomposed into hydrogen and oxygen through intermediate reactions below 500 °C, involving copper and chlorine compounds. The Cu–Cl cycle is well matched to Canada's nuclear reactors, since its heat requirement for high temperatures is adaptable to the super-critical water reactor (SCWR), Canada's Generation IV nuclear reactor. The primary components of the cycle are four interconnected reaction vessels, with intermediate heat exchangers, and a drying step. The sequence of steps in the Cu–Cl cycle is shown in Table 1. A conceptual schematic of the Cu–Cl cycle is depicted in Fig. 1. This paper focuses on gas–solid transport phenomena of a key hydrolysis reaction within the Cu–Cl cycle, namely the reaction of cupric chloride particles with superheated steam in a fluidized bed reactor (reaction 4 in Table 1). Currently, no experimental data are available in the archival literature on the hydrodynamics and chemistry of the hydrolysis reaction. This paper develops predictive models and performs sensitivity studies to determine the effects of various parameters on operating performance.

In a fluidized bed reactor, a mass balance of gas and solid reactants requires knowledge of the bed properties, through a hydrodynamic analysis (Crowe, 2006, Haseli et al., 2008). Several past studies have modelled the conversion of gaseous species, including those occurring in catalytic gas–solid reactions (CGSR), e.g., gasification in a fluidized bed (Basu, 2006). Kunii and Levenspiel (1991) used three reactor models that depend upon flow regimes to predict the conversion of reacting gas. They developed two limiting models to describe the conversion of solid particles in a non-catalytic gas–solid reaction (NCGSR). These models are based on two limiting behaviors, either a uniform reaction model, or a shrinking core model (SCM). The calculation procedure of Kunii and Levenspiel (1991) uses a combination of different models. It would be beneficial to develop a single model that best describes the conversion of both gaseous and solid reactants. This article addresses this goal by developing a single model that can accommodate both limiting behaviors.

Various two-phase models have been developed to predict the conversion of a gaseous reactant in a fluidized bed reactor under CGSR assumption, where the solid particles do not participate in the reactions. Toomey and Johnstone (1952) introduced a two-phase model of fluidization, which assumes that all gas in excess of the minimum fluidization velocity flows through the bed as bubbles, while the emulsion stays stagnant at the minimum fluidization condition. The model of Davidson and Harrison (1963) follows the two-phase theory of Toomey and Johnstone (1952), where bubbles have a uniform size throughout the bed and the reaction takes place only in the emulsion phase with first-order kinetics. The two-phase model proposed by Partridge and Rowe (1966) also uses the two-phase theory of Toomey and Johnstone (1952), but considers the gas interchange to occur at the cloud–emulsion interface. The bubble and cloud phase are assumed to be well-mixed, with the result of a bubble–cloud phase. Kato and Wen (Yang, 2003) proposed a bubble assemblage model that considers a changing bubble size with varying height in a bed. The bubbles grow continuously, while passing through the bed, until they reach a maximum stable size, or the diameter of the bed column. The bed is assumed to be operating under isothermal conditions, since the effective thermal diffusivity and the heat transfer coefficient are large. In contrast, the bubbling bed model proposed by Kunii and Levenspiel (1991) is a modified version of the two-phase model. In addition to the bubble and emulsion phases, a cloud–wake phase is also considered. A key difference between this model and the other two-phase models is that the inter-phase mass transfer considers two distinct resistances, one from the bubble phase to the cloud–wake phase, and the other from the cloud–wake phase to the emulsion phase. This paper will extend these past models to a new gas–solid formulation that predicts the hydrodynamic behavior of cupric chloride particles in a hydrolysis step of the thermochemical Cu–Cl cycle.

Section snippets

Formulation of gas–solid conversion

The extension of two-phase models to NCGSRs is difficult, because solid particles take part in the reaction as well. The method of Kunii and Levenspiel (1991) may be used to describe the conversion of solid particles in an NCGSR, through two extremes of behavior: uniform reaction model and an SCM. It will be assumed that the reacting solids are covered by gas of the same mean composition. If a reaction is slow and the concentration of gaseous reactant does not change significantly when passing

Fluidized bed model for cupric chloride particles

Since both cupric chloride particles and steam take part in the reaction, it is necessary to develop a NCGSR model to analyze the bed performance. The following assumptions will be made when analyzing the bed behavior.

  • The bed consists of two regions according to two-phase theory: a bubble phase and an emulsion phase.

  • The temperature gradient within the bed reactor is negligible; so the bed experiences an isothermal process.

  • There exists only one hydrolysis reaction, namely the reaction of cupric

Results and discussion

This section applies the previous algorithms to predict the effects of various process parameters on the reaction of cupric chloride particles with superheated steam in a typical bench-scale fluidized bed reactor. The reactor has a diameter of 2.66 cm and height of 16 cm. The hydrodynamic behavior of the fluidized bed, where the cupric chloride particles react with steam, has been examined previously by Haseli et al. (2008). The section will present new results for the overall performance of the

Conclusions

This paper has focused on transport phenomena involving a reaction of cupric chloride particles with superheated steam in a fluidized bed reactor, as part of a Cu–Cl thermochemical cycle. Since both cupric chloride particle and steam take part in the chemical reaction, the method of Gómez-Barea and co-workers has been extended to analyze the non-catalytic gas–solid reaction (NCGSR), for the purpose of the current study. Due to a lack of experimental data regarding hydrodynamics and chemistry of

Notation

Ac,bedbed cross-sectional area, m2
bstoichiometric factor of the reaction
CAbgaseous reactant concentration in the bubble phase, mole/m3
CAegaseous reactant concentration in the emulsion phase, mole/m3
CAiinlet gaseous reactant concentration, mole/m3
CAooutlet gaseous reactant concentration, mole/m3
DasDamkohler number of the reactor scale, Eq. (7)
fgeneral function
f1(xc0,λ)function defined in Eq. (15)
f2(xc0,λ)function defined in Eq. (24)
Fi(xc)kinetic function, Table 2
F0inlet flow rate of solids, kg/s
F

Acknowledgments

The authors gratefully acknowledge the support provided by Atomic Energy of Canada Limited (particularly Dr. S. Suppiah and Dr. A. Miller) and the Ontario Research Excellence Fund. The first author acknowledges personal communication with Dr. Alberto Gómez-Barea, who provided valuable assistance through his past articles.

References (30)

  • T.C. Crowe

    Multiphase Flow Handbook

    (2006)
  • J.F. Davidson et al.

    Fluidized Particles

    (1963)
  • C.W. Forsberg et al.

    Molten-salt-cooled advanced high-temperature reactor for production of hydrogen and electricity

    Nuclear Technology

    (2003)
  • Haseli, Y., Dincer, I., Naterer, G.F., 2008. Reactive transport phenomena of copper oxychloride particles in a...
  • S. Kasahara et al.

    Effects of process parameters of the IS process on total thermal efficiency to produce hydrogen from water

    Journal of Chemical Engineering of Japan

    (2003)
  • Cited by (0)

    View full text