Gas-liquid
distribution in monoliths: Effect of distributor configurations and scale

Deepali
Chugh and Shantanu Roy^*

Department
of Chemical Engineering, Indian Institute of Technology Delhi, India

(^*
roys@chemical.iitd.ac.in)

Keywords:
Monolith;
Gas liquid distribution; Gas liquid distributor; Distribution index

Monolith reactor is one of the
alternatives that are being actively considered for overcoming the shortcomings
of trickle bed reactors. It has been established in literature that for any chemical
kinetics, it is always possible to optimally design a monolithic reactor for
three-phase applications that will outperform a trickle bed (packed bed)
reactor [1]. However, such a claim is valid only if the liquid distribution and
local irrigation in the monolith channels is uniform. This is a tall order,
particularly since a monolithic reactor that could potentially replace or
retrofit a trickle bed reactor must operate in "trickle-flow" like
flow regime, which in the context of monoliths would correspond to "film
flow" conditions [2]. Liquid distribution is usually poor in these low
flux conditions. Trickle beds having fine particles (of the order of 1 mm) do
suffer from liquid mal-distribution problems as well, but the radial
distribution of liquid in the bed is not a very strong function of the
"distributor-generated" pattern at the top of the co-current trickle
bed, as the liquid finds its path of least resistance through the
inter-particle voids and eventually develops into a "natural"
distribution. In comparison, monoliths, with little or no radial flow, are less
forgiving since a mal-distribution generated at the top in a co-current flow
bed is likely to propagate in the same manner down the length of monolith. A
good distribution generated by the liquid distributor, on the other hand, can
lead to complete utilization of all the channels and hence realize the
locked-in activity of the monolith bed, arguably more than what can be achieved
in a packed trickle bed [1].

In view of the above arguments, it is
clear that an assessment of liquid distribution as a function of operating flow
conditions, liquid (and gas) properties and monolith channel dimensions, is of
crucial importance. Some reports of work in this direction are available in
literature [2-4]. Most of the reports in literature deal basically with the
Taylor flow regime that has been a popular topic amongst researchers owing to
the potential enhanced mass transfer benefits that this flow regime offers. However,
Taylor or slug flow regime is of little significance in the context of a vast
number of applications such as various kinds of hydrotreating, where monoliths
may be retrofit to replace trickle beds that are the reactor of choice today.
Even amongst literature works related with film flow where distribution issues
have been investigated, the emphasis has been on small laboratory scale
monoliths that are a few centimetres in diameter, and little or no results have
been presented on the effect of scaling up distributor and monolith dimensions
to larger sizes.

In
this presentation, results from our experimental investigation of liquid
distribution in three-phase monoliths, considering different distributor
configurations as well as varying the relative sizes of monoliths and
distributors, is considered. Effect of liquid properties on the flow
distribution is also investigated.

Experimental
Setup and Approach

A
co-current monolith reactor is operated in downflow mode using
air and water with monolith of desired cell geometry. For
a comparative study among different distributors various configurations of
single pipe distributors, multi-pipe distributor, showerhead type, packed bed
type and spray nozzle type were used. Details of the flow operating conditions
and distributor configurations are mentioned in Table 1. The range of flow
rates is selected on the basis to cover liquid hour space velocity (LHSV) of
100-2500 hr^-1. Distribution and pressure drop measurements were made
for a wide range of superficial velocities covering both film and Taylor flow
regimes in the interest of retrofit applications of monolith. The laboratory
setup is based on gravimetric liquid collection method using a customized apparatus,
which allows for collection of liquid for a pre-set duration (which is adjusted
as per the flow rate). In particular, the bottom cross-sectional surface of
monolith is divided into different zones and liquid outflow through these
predefined zones is collected and weighed to measure distribution. Schematic
diagram of the setup is shown in Figure 1.

Figure
1. Schematic diagram of laboratory
setup

Table 1. Flow conditions and specifications of the
different distributors used for the study

For each set of the
flow conditions, the extent of deviation from uniformity is measured through a Distribution
index (j) as defined in Equation (1) considering the 
locational value of data (x_i, y_i) in addition to
its variation around the mean.

                                   (1)

where
 (g/s)
the mass flow is rate through zone i,  (g/s)
is the mean of mass flow rates through all collection zones and M is
the number of collection zones.

Results
and Discussion

Within
the flow range under which experiments were conducted, the distribution is
found to vary in a significant manner with different configurations of
distributors. Greater uniformity can be realized by increasing the number of
liquid injecting points over the top of monolith, but with a trade-off of
higher pressure drop of distributor. This was clear from the results of single
pipe, multi-pipe and showerhead distributors.

A
typical set of results for showerhead different showerhead configurations with
same number of holes distributed over the monolith but of different sizes and
therefore different open % area is presented here. Detailed discussion on all
the distributors and across all flow conditions will we part of the final
presentation/submission.

Figure
2(a) shows the raw experimental data as measured for the configuration having
1 mm holes with percent open area of 3.33 at superficial liquid velocity 0.15
m/s. The data represented in each zone is in form of mean ± standard deviation
for five independent experiments. Figure 2(b) shows the suitably
two-dimensional interpolated data. The inset in Figure 2 shows overall mass
balance for the particular case. It has to be noted that all the data reported
in the figures is an average of five experiments.

(a)                                                 
(b)

Figure
2 (a) Experimental data in g/s (mean ± standard deviation) measured in
each zone at superficial liquid velocity 0.15 m/s and (b) two dimensional
interpolated data of the same.

Typical
result of different showerhead configurations with same number of holes
distributed over the monolith but of different sizes and therefore different
open % area for the range of flow velocities examined (Figure 3(a) and 3(b)). It
is clear from the Figure 3(a) that change in open area percentage does not lead
to a significant variation in the distribution index.

(a)

 

(b)

Figure
3. (a) Distribution index and (b) pressure drop,
across distributor as a function of liquid superficial velocities for
showerhead distributors all having 181 holes of diameter 0.5, 1.0 and 1.5 mm respectively
and distributor percent open area of 0.83, 3.33 and 7.55 respectively.

At
low air flow rates the distribution is worse for all the configuration as
obvious from the large deviations in the magnitude. However, beyond a certain
point the distribution is stabilised with negligible deviations. This
limit is lowest for the configuration having smallest open area percentage (i.e.
with 181 holes of 0.5 mm size) but at the cost of high pressure drop
(Figure 3(b)). And if the goal is to identify an optimum distributor for
industrial applicability of monolith, pressure drop of distributor is a key
parameter to be considered while assessing any distributor. Hence stringent
uniformity level relies upon the pressure drop of the distributor.

The
final submission will discuss experimental findings and provide quantitative
and qualitative recommendations regarding the choice of distributors for
three-phase monoliths.

References

1.    Roy S, Heibel AK,
Liu W, Boger T. Design of monolithic catalysts for multiphase reactions. Chem.
Eng. Sci. 2004;59:957-966.

2.    Heibel AK,
Heiszwolf JJ, Kapteijn F, Moulijn JA. Influence of channel geometry on
hydrodynamics and mass transfer in the monolith film flow reactor. Catal.
Today. 2001;69:153-163.

3.    Gladden, LF, Lim
MHM, Mantle MD, Sederman AJ, Stitt H. MRI visualization of two-phase flow in
structured supports and trickle bed reactors. Catal. Today. 2003;79-80:203-210.

4.    Behl M, Roy S. Experimental
investigation of gas-liquid distribution in monolith reactors. Chem. Eng.
Sci. 2007;62:7463-7470.