Reaction engineering: The supercritical water hydrothermal synthesis of nano-particles

https://doi.org/10.1016/j.supflu.2005.08.011Get rights and content

Abstract

Supercritical water hydrothermal synthesis (scWHS) is a relatively simple and environmentally friendly process for the production of potentially valuable metal oxide nanoparticles. However, it has never found industrial application to date due to poor process reliability, reproducibility and control. This paper presents the conclusions of collaborative work between chemical engineers and chemists that attempts to optimise the reaction engineering of this process, with the goal of reducing or even eliminating these fundamental process flaws. Initial investigations on the mixing in a T-piece highlighted that the environment within the scWHS reactor is highly unusual in terms of conventional reaction engineering because the fluid properties were significantly different in terms of temperature, viscosity and density. This led to the development of an optimised reactor, termed the Nozzle Reactor, which was designed on the basis of Light Adsorption Imaging (LAI) and Computational Fluid Dynamics (CFD) modeling, both of which show excellent mixing mechanics. Light Adsorption Imaging is an image analysis visualization method using fluids of different densities at ambient conditions. Experimental results using the Nozzle Reactor are presented showing how different metal oxide particles can be produced including titania, ceria, zirconia, copper, zinc and silver. The reactor shows a dramatic improvement in process reproducibility (±5 m2/g for BET surface area) and in reliability such that, given further investigation, will lead to process optimisation. Preliminary evidence suggests that the reactor could eventually lead to the ability to control particle properties, such as size, composition and shape, through the manipulation of process variables.

Introduction

In recent years, both scope of application and demand for nano-scale metal and metal oxides have greatly expanded. Nano-sized materials exhibit many desirable properties and the efficient production of these materials is likely to play a key role in the future of the specialty chemical industry. The range for applications of metal (e.g. Ag) or metal oxide (e.g. CeO2) nanoparticles is greatly expanding; current and potential applications include colloid science, environmental remediation, catalysis and photo-catalysis, electronics, medicinal applications, separations, thin films, inks, and disinfection. However, many of the industrial synthetic routes in use today are generally not easily scalable, involving relatively noxious chemicals and expensive precursors, and a complex and time-consuming sequence of steps. Supercritical water hydrothermal synthesis (scWHS) offers a relatively simple route which is inherently scalable and chemically much more benign than current technology. As a result, this continuous process has been investigated extensively by several research groups [1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], but has yet to be applied industrially due to problems of reliability, reproducibility and process control.

The scWHS process is relatively straightforward; it involves the mixing of an aqueous metal salt stream with a supercritical water stream within a continuous reactor to produce nano-size metal oxide particles. When water is heated towards its critical point (Tc = 374 °C, Pc = 22.1 MPa), it changes from a polar liquid to a fluid with a low dielectric constant and low pH. Kw also increases, giving rise to correspondingly increased concentrations of H+ and OH. These enhanced levels of OH were first exploited for nano-particle synthesis by Adschiri et al. [5], who showed that, under these conditions, hydrolysis of the metal salts was immediately followed by a dehydration step.Hydrolysis: MLx + xOH  M(OH)x + xLDehydration:M(OH)xMOx/2+x2H2OFor the past 6 years [12], [13], [14], [15], [16], [17], the Clean Technology Research Group at the University of Nottingham has been carrying out research into the development and eventual optimisation of the scWHS process with the goal of making the process industrially viable. Recently, our group expanded this approach to the synthesis of single-phase mixed metal oxides [13], [14]. There have also been investigations, by other workers, into the effects of pH [18], metal salt concentration [19], temperature [20], pressure [11], etc. on the morphology and size distribution of fine metal oxide particles.

The flowsheet for the Nottingham scWHS process [12], [13], [14], [15], [20] is given in Fig. 1. Prior to the development of the nozzle reactor, the reactor was a vertically up-right Swagelok® T-piece (0.71 cm internal diameter). Supercritical water entered the reactor via the side arm, whilst the aqueous metal salt was introduced through the top arm. Almost every scWHS investigation conducted in this apparatus was hindered by the unreliability of the process and poor product reproducibility. The root of the problem was particle accumulation and agglomeration within the reactor and its two inlets, which caused a narrowing and eventual blockage of the process pipes. As a result, the apparatus usually required extensive cleaning between runs, and in many cases, the experiments had to be shutdown prematurely because of blockages within the metal salt inlet. Experiments under various process conditions suggested that the key to eliminating these problems lay in the engineering of the reactor. The resulting investigation into reaction engineering and reactor modeling involved the collaboration of chemical engineers and chemists at the University of Nottingham.

Two techniques were used to model the scWHS reactor. The first technique (LAI modeling) involved performing ‘simulations’ of the scWHS mixing environment at ambient conditions in a pseudo-reactor, using the approach described by Blood et al. [15] The results of LAI modeling were then complemented by Computational Fluid Dynamics (CFD) simulations of the reactor under similar conditions. The LAI experiments [15], which examined the efficiency of a T-piece reactor at different flowrates and reactor/inlet orientations, produced some very interesting conclusions. The flow regime within the scWHS reactor was found to be highly unusual in the terms of conventional reactor engineering. Firstly, the Reynolds number (Re) for each of the reactant streams was found to be extremely low (Re = 148 in the supercritical water inlet, Re = 20 in the metal salt inlet), suggesting that both feed streams were highly laminar in nature.

However, with the initial LAI and CFD modeling experiments with the T-piece reactor, highly turbulent macro-mixing could be clearly observed as the two streams converged in the reactor. This was unexpected when considering the highly laminar nature of the reactor feeds; further investigation highlighted that the driving force behind this turbulent macro-mixing phenomenon was not the inertial forces produced by the process pumps, but the strong buoyancy forces present in the reactor. The two feed streams have significantly different fluid densities (supercritical water density = 371 kg m−3, metal salt density ≈998 kg m−3) such that, when these fluids interact, relatively large natural convection or buoyancy forces are induced. The extent to which these buoyancy forces dominate in these reactor systems was calculated using dimensional analysis [15], i.e. Grashof number (Gr) and Reynolds number. The results of analysis showed that the natural convection forces within the scWHS system were relatively large enough to induce turbulent flow/mixing (Gr = 9.5 × 109), which dominate over the much smaller inertial forces produced by the pumps (Gr/Re2 > 1). Thus, the main conclusion was that a T-piece scWHS reactor is very inefficient at handling such an atypical mixing system. This paper presents a new reactor design that has been specifically developed to: (a) handle this unusual mixing environment efficiently (b) exhibits flow properties that have an increased ability to handle the particulate product (c) be capable of producing a greatly improved heat transfer profile.

Section snippets

Experimental

This optimal scWHS reactor was assessed using modelling techniques (LAI and CFD) identical to those presented in a previous publication [15]. Direct observation of fluids under supercritical water conditions is highly impractical due to the engineering restrictions at the high pressures and with the small pipe diameters used in this process. Therefore, to assess the nature of the mixing mechanism within each reactor design, work was carried out to physically recreate the mixing scenario in a

Design criteria

The LAI and CFD modeling illustrated the importance, of the density difference between the two reactant fluids [15]. This suggested that the key to optimising the reactor was to exploit the density difference between the two fluids. An ‘ideal’ scWHS reactor should be able to satisfy the following criteria:

  • Instantaneous strong and uniform mixing of two reactant streams, to aid in the formation of many small metal oxide nuclei which is desirable for small particle formation;

  • Short average

LAI and CFD modelling

The steady state concentration map and its equivalent CFD simulation for the Nozzle Reactor are shown in Fig. 3a and b, respectively. The reactor takes advantage of the density difference between the two reactants so that the resultant mixing pattern satisfies all four of the criteria above. As the two reactant fluids are introduced, the mixing is instantaneous and strong. The resultant turbulent macro-mixing eddies are streamlined downstream to the outlet of the reactor. No contamination and,

Testing at supercritical conditions

The Nozzle Reactor was constructed used Swaglok® high pressure fittings; the outer tube consisted of a 3/8″ tube (316 Stainless Steel, 0.065″ wall thickness) and the inner tube was constructed from a 1/8″ tube (316 Stainless Steel, 0.035″ wall thickness). The construction of this reactor is shown schematically in Fig. 4. The ‘nozzle’ was omitted for these initial tests as a precaution due to the large reduction in available flow area it would have produced. Table 1 summarises the size of

Process reliability and reproducibility

Experiments were performed to generate CeO2 to test the reliability and the reproducibility of the new optimised reactor design. The initial test was to run the process for prolonged periods (up to 9 h) to assess the process reliability.

The following process conditions were used:

  • Cerium ammonium nitrate (0.2 M) set at 5 mL min−1 flow rate.

  • Supercritical water set at 10 mL min−1 flow rate.

  • Supercritical water temperature 410 °C.

  • System pressure was set at 24.8 MPa.

  • Water cooling coils were added the reactor

Product tunability

The reliability and reproducibility of the experiments with the Nozzle Reactor both mean that better and more accurate experimental control can be achieved in future experiments. For example, Fig. 6 shows the change in surface area, determined by BET analysis, of CeO2 as the flow rate of the aqueous metal salt feed flow rate was changed. A peak surface area of 100 m2/g was obtained at 8 ml/min, giving the optimal ratio of supercritical water to aqueous metal salt for small particle formation.

Conclusions

The dominant driving force behind the mixing of supercritical water and aqueous metal salts streams in a scWHS reactor appears to be the bouyancy forces induced by the large density difference between the two reactants. This realisation means that the orientation of the reactor, as well as the relative positions of the inlets and outlets, become key properties in the design of the reactor. Subsequent investigation in alternative reactors suggested that the key to generating an optimal reactor

Acknowledgements

We thank the EPSRC and ICI Strategic Technologies for financial support. We Thank Dr P.A. Hamley, Mr M. Guyler, Mr P. Fields and Mr R. Wilson for their help.

References (27)

  • F. Cansell et al.

    Supercritical fluid processing: a new route for materials synthesis

    J. Mater. Chem.

    (1998)
  • L. Cote et al.

    Continuous hydrothermal synthesis and crystallization of magnetic oxides

    J. Mater. Res.

    (2002)
  • L.J. Holm et al.

    Synthesis and crystallization of magnetic nanoparticles

  • Cited by (231)

    View all citing articles on Scopus
    View full text