(87d) Macromixing Characteristics of Viscous and Shear-Thinning Fluids in Rotor-Stator Spinning Disc Reactors

Macromixing Characteristics
of Viscous and Shear-Thinning Fluids in Rotor-Stator Spinning Disc Reactors

Arnab
Chaudhuri^a ( font-family:" arial>a.chaudhuri@tue.nl), Wyatt
Winkenwerder^b (Wyatt.Winkenwerder@nouryon.com), John van
der Schaaf^a (J.Vanderschaaf@tue.nl)

text-autospace:none">Department
of Chemical Engineering, TU Eindhoven, 5612 AZ, Eindhoven, Netherlands

text-autospace:none">Nouryon-Surface
Chemistry, Croton River Center 281 Fields Lane, Brewster 10509, New York,
United States of America

text-autospace:none"> 

Introduction

In
recent years the chemical industry has been focused on improving both sustainability
and safety of chemical processes. This involves for instance an increase in efficiency
of raw material/energy usage. This drive towards process intensification has
seen the development of many novel, integrated reactors one of which is
the rotor-stator spinning disc reactor (rs-SDR). The rs-SDR consists of a
rotating disc, enclosed by a stationary cylindrical housing. Intensification in
mass transfer and heat transfer rates can be achieved due to the small cavities
which provide a high degree of turbulence and shear. Previous investigations
into the hydrodynamics of rs-SDRs have resulted in an engineering model to
describe the macromixing characteristics for aqueous systems as a function of
flow rate, gap distance and rotation speed[1]. As we look towards
applications for the rs-SDR, it becomes vital to characterize viscous systems,
both Newtonian and non-Newtonian, as these are highly relevant for industrial
systems. In this study, we have characterized the macromixing behavior of
viscous systems using residence time distribution measurements and have looked
at how they differ in comparison with aqueous systems at varying rotation
speeds.

Experimental
Methods

A single stage rs-SDR has been used for this study. The
radius of the rotor is r_d= 143*10^-3 m while the radius of
the cylindrical housing is 145 *10^-3 m. This results in a radial gap
of 2 *10^-3 m. The height of the stage is 9 *10^-3 m with a
disc thickness of 4 *10^-3 m, therefore the resulting axial gap is
2.5 *10^-3 m.

none">

Figure 1 font-style:normal">: Schematic of the experimental setup.

The liquid
was pumped at the top of the reactor with flowrates monitored by a CORIFLOW^TM(Bronkhorst) meter and was withdrawn from the bottom. The temperature in
the system is maintained with the use of a cooling bath (LAUDA ECO RE 630). The
RTD measurements were carried out by measuring the tracer injection through
UV-vis at both the inlet and outlet. By deconvolution of the inlet signal from
the outlet we were able to obtain the final RTD curve[2].
A schematic of the experimental setup is shown in Fig. 1. We have used glycerol
and Xanthan gum (1 w/w%) to study the behavior of viscous and non-Newtonian fluids,
respectively.

Results
& Discussion

" arial>      

font-family:" arial> 

            Figure 2: Experimentally obtained E_exp
curves with the conventional PFR-CSTR in series model fitting (G=0.035, Re_ω=
4.27 * 10⁵, C_ω = 104.7). A) Water B) Glycerol (69%w/w) C)
Xanthan Gum (0. 1% w/w)

Figure 3 font-style:normal">: Effect of rotation speed on the PFR volume fraction for various
fluids in the RS-SDR

The results
indicate that while the model used for aqueous systems was applicable to
viscous, Newtonian systems; for Xanthan gum (non-Newtonian) we found that the
fitting of the E-curve was quite poor (Fig 3C). This is most likely due to the presence
of varying regions of viscosity in the reactor. At the edge of the disc where
the shear is most intense, we can expect that the viscosity to be low, while near
the entrance and exit of the fluid stream we can expect the viscosity to be
much higher. Therefore, an alternative model should be investigated for this
system, mostly likely involving CSTRs with exchange volumes to differentiate
between regions of higher and lower viscosity.  

We have also
observed that the time at which we first observe a signal at the outlet also
varies significantly for the non-Newtonian fluid. For the aqueous model, this
breakthrough time was described as the PFR residence time. Assuming that PFR
volumes are still present in the reactor for non-Newtonian fluids, we can see
that the PFR volume of the reactor decreases significantly for Xanthan Gum in
comparison to Glycerol or Water (Fig 3). This can be expected since the shear
thinning nature of the fluid will mostly likely lead it to transition to CSTR
type behavior at earlier reactor lengths.

 Future Work           

font-family:" arial> 

The results
obtained thus far have illustrated the applicability of RS-SDRs for viscous
flows. We have observed that the macromixing characteristics of the reactor can
still be maintained for Newtonian fluids with kinematic viscosities greater
than 27 times that of aqueous systems. We are currently investigating at what
values the model for the viscous system is unable to mimic the aqueous system
and poor mixing behavior is observed.  We are also investigating a model to
describe the non-Newtonian fluid in a RS-SDR. Furthermore, this work will be
extended to also investigate the effect of flow rate and gap ratio on the
macromixing behavior of viscous flows.

font-family:" arial> 

References

none">[1]       M.M. de Beer,
J.T.F. Keurentjes, J.C. Schouten, J. van der Schaaf, Engineering model for
single-phase flow in a multi-stage rotor–stator spinning disc reactor, Chem.
Eng. J. 242 (2014) 53–61. doi:10.1016/j.cej.2013.12.052.

none">[2]       D.
Bošković, S. Loebbecke, G.A. Gross, J.M. Koehler, Residence Time
Distribution Studies in Microfluidic Mixing Structures, Chem. Eng. Technol. 34
(2011) 361–370. doi:10.1002/ceat.201000352.

font-family:" arial>