J.

of Supercritical Fluids 88 (2014) 117–125

Contents lists available at ScienceDirect

The Journal of Supercritical Fluids

journal homepage: www.elsevier.com/locate/supflu

Study of transpiring fluid dynamics in supercritical water oxidation

using a transparent reactor
Zhong Chen a,b,1 , Guangwei Wang a,b,1 , Zakaria. A. Mirza a,b , Shu Yang a,b , Yuanjian Xu a,b,∗
a
Chongqing Institute of Green and Intelligent Technology (CIGIT), Chinese Academy of Sciences, Chongqing 400714, PR China
b
Key Laboratory of Reservoir Aquatic Environment, Chinese Academy of Sciences, Chongqing 400714, PR China

a r t i c l e i n f o a b s t r a c t

Article history: The transpiring wall reactor (TWR) is considered to be one of the most promising reactors because it
Received 26 August 2013 minimizes both corrosion and salt precipitation problems that seriously hinder the industrialization of
Received in revised form 29 January 2014 supercritical water oxidation technologies. A transparent reactor is built to study the fluid dynamics of
Accepted 29 January 2014
transpiring flow, which are the foundation of reactor design and optimization. The results showed that
Available online 14 February 2014
the transpiring flow is anisotropic with respect to the surface of the transpiring wall due to both the
static pressure and viscous resistance. Finally, the novel idea of using air as the transpiring fluid instead
Keywords:
of water is presented in an attempt to alleviate current TWR problems such as high energy consumption,
Fluid dynamics
Transpiring fluid
high volume of pure water consumption, and temperature fluctuation in the reaction area. A series of
Transpiring wall reactor experiments and theoretical derivations demonstrate that this novel idea is feasible.
Gas seal © 2014 Elsevier B.V. All rights reserved.
Supercritical water oxidation

1. Introduction
In the 1980s, Modell [1] pioneered the degradation of organic
waste in supercritical water, which marked the birth of supercrit-
ical water oxidation (SCWO) technology. Modell’s process utilizes
the special physical properties of water under supercritical state
(T ≥ 374 ◦ C, P ≥ 22.1 MPa), such as the non-polar character, high
diffusivity, and excellent transport properties [2,3]. The SCWO is
hailed as the greatest potential end-of-pipe technology, but its
commercialization is hindered by corrosion and salt precipitation
problems. Some plants have been interrupted or shut down due to
these two problems [4,5].
In the last three decades, many kinds of reactors have been
designed and studied to solve these problems [5,6]. The transpiring
wall reactor (TWR), which was first reported by McGinness in 1993
[7], is undoubtedly one of the most promising technologies because
it can simultaneously handle both corrosion and salt precipitation
problems.
The typical TWR structure is shown in Fig. 1, and the key design
principles are described as follows. First, a TWR consists of an outer
∗ Corresponding author at: Environmentally-Benign Chemical Process Research
Center, Chongqing Institute of Green and Intelligent Technology, Chinese Academy
of Sciences, No.266 Fangzheng Avenue, Shuitu Hi-tech Industrial Park, Shuitu Town,
Beibei District, Chongqing, P. R. China, 400714. Tel.: +86 23 65935819.
E-mail address: xuyuanjian@cigit.ac.cn (Y. Xu).
1
Zhong Chen and Guangwei Wang contributed equally to this work.
non-porous wall bearing high pressure and an inner porous wall.
All reactions occur inside the porous tube. Second, the annulus
between the two walls is filled with either subcritical or supercrit-
ical pure water; therefore, the fluid prevents the pressure-bearing
wall from contacting the corrosive reaction solution. Third, water
flows through the transpiring wall and forms a protective water
film on the inner wall surface against the corrosive reaction mix-
ture, salt deposition, scaling, and high temperatures [5].
More than 10 organizations have researched TWRs, and three
types of transpiring walls have been designed. A summary of the
different types of transpiring walls is provided in Table 1.
In 1993, Mueggenburg et al. [8] designed the first transpir-
ing wall type, called platelets. Their design was based on Aerojet
platelets technology. In order to provide better wall protection,
Daman [9] and Ahluwalia [10] from Foster Wheeler Development
Corporation patented a better performance platelet tube, which is
a seven-layer structure that can disperse a single inlet stream into
1134 flow paths. This platelet reactor was widely used in SCWO in
the 1990s [11–13].
The second transpiring wall type is more efficient at forming an
isotropic protective film and is well known as the sintered porous
tube. The ITC-CPV [14] group takes the lead in adopting this type of
transpiring wall. Since 2001, various materials have been studied
to improve the corrosion resistance of the transpiring wall, such
as stainless steel [14], ceramic [17,24], Inconel alloy [20,22], and
316L low-carbon stainless steel [26]. Furthermore, both the pore
diameter of the porous tube [15,17,20,23,24,26] and the reactor
design [15,17,20,22,26] have been studied. Wellig et al. [20] and

http://dx.doi.org/10.1016/j.supflu.2014.01.020
0896-8446/© 2014 Elsevier B.V. All rights reserved.
118 Z. Chen et al. / J. of Supercritical Fluids 88 (2014) 117–125

Nomenclature

A test area of transpiring wall (m2 )

Din inner diameter of pressure bearing wall (m)
Dout outer diameter of pressure bearing wall (m)
din inner diameter of transpiring wall (m)
dout outer diameter of transpiring wall (m)
dp pore diameter of transpiring wall (m)
Ftw mass flow rate of transpiring water (kg h−1 )
Fcw mass flow rate of cooling water (kg h−1 )
Fww mass flow rate of waste water (kg h−1 )
H height of the transpiring tube (m)
P fluid pressure (Pa)
P1 pressure drop across the transpiring flow inlet and
the reactor outlet with porous tube (Pa)
P2 pressure drop across the transpiring flow inlet and
the reactor outlet without porous tube (Pa)
P pressure drop due to porous tube (Pa),
P = P1 − P2
Qv volume flow rate (m3 s−1 )
Qv,radial volume flow rate of radial flow (m3 s−1 ), radial flow
and transpiring flow are the same in this paper
Qv,axial volume flow rate of axial flow (m3 s−1 )
Rtw ratio of transpiring water to waste water,
Fig. 1. Typical structure of transpiring wall reactor.
Rtw = Ftw /Fww
Rcw ratio of cooling water to waste water, Rcw = Fcw /Fww
R ratio of total water to waste water, [18] identified a pressure drop across the porous wall, which is gov-
R = (Fcw + Ftw )/Fww erned by Darcy’s law under supercritical conditions. However, the
Vannulus volume of annulus (m3 ) local flow condition inside TWRs has not been studied. In recent
Vtube volume of transpiring tube (m3 ) years, computational fluid dynamics (CFD) have been applied to
describe the velocity and temperature profiles inside the reactor
Greek Symbols and around the transpiring wall in SCWO [15,30]. Although CFD can
˛ viscous permeability of transpiring wall (m2 ) be used to describe the flow conditions in TWRs, there are still some
 viscosity (Pa s) inaccuracies in the simulated results due to the limitations of the
 density (kg m−3 ) CFD software and the theoretical models’ inability to predict state of
ı thickness of transpiring wall (m) a real, complex system [30]. Sapphire window installation [31,32]
1 residence time (s) defined by Eq. (4) has been used to visually observe what happened in a SCWO reac-
2 residence time (s) defined by Eq. (5) tor. This method is the only direct way to make visual observations;
3 residence time (s) defined by Eq. (6) unfortunately, the field of vision is too small for a TWR.
In this paper, a simplified and custom designed glass transparent
reactor was developed to understand the fluid dynamics of the tran-
Zhang et al. [26] separately designed a special system for TWRs spiring flow. In particular, a tracing technique (a colored dye) was
that can control both the temperature and flow rate at individual adapted to show the dynamic processes. In order to avoid the high
transpiring flow inlets. Bermejo et al. [22,23] innovatively used a pressure and high temperature associated with subcritical water,
partly porous tube in their reactor that achieved good results with which is commonly used as a transpiring fluid (Table 1), either
a comparatively low transpiring ratio (Rtw ≈ 2). ethanol or acetone was used as a substitute based on their simi-
Xu et al. [28] reported the latest transpiring wall type in 2010. lar physical properties. A series of experiments were performed at
The sintered wire netting transpiring wall was manufactured by 25 ◦ C and 0.1 MPa. The pressure drop across the porous tube was
the sintering metal wire at high temperatures to produce a wire measured. Finally, a novel idea was introduced that uses air as the
mesh sheet. This kind of transpiring wall combines the strengths transpiring fluid instead of water, and its feasibility was assessed.
of the two previous types. In particular, the sintered wire netting
transpiring wall possesses good mechanical strength, which is sim- 2. Experimental sections
ilar to platelets, and the pores have a uniform distribution, which
is similar to the sintered porous tube. 2.1. Reactor design
After careful reflection, there is still a critical problem with TWRs
that must be resolved. In TWRs (Table 1), a large quantity of pure The general design of the transparent reactor is shown in
water is injected into the reactor at a high transpiring ratio (Rtw ) and Fig. 2. The transparent reactor (11) consists of a glass outer tube
low temperature (Ttw ) in order to protect the reactor from corrosion (L = 0.500 m, Din = 0.050 m) and a porous inner tube (Alumina,
and salt precipitation. However, this will cause temperature fluctu- dout = 0.040 m, din = 0.0265 m, dp = 8 × 10−5 m). There is an inlet
ations in the reaction area and requires high energy consumption (0.005 m i.d.) for the transpiring flow in the middle of the glass tube.
[23,27]. The annulus between the two tubes are sealed by rubber O-rings
The reactor design can be improved if the flow behavior of the at both ends. The effective length of the porous tube is 0.460 m.
transpiring fluid can be directly determined and optimized, but the The glass transparent reactor enables visual observation of the
extreme chemical and physical environment in the SCWO reactor dynamic processes of transpiring fluid. In most cases, subcritical
limits the measurements needed for current research. Fauvel et al. water is used as transpiring fluid (Table 1); however, the glass
 Z. Chen et al. / J. of Supercritical Fluids 88 (2014) 117–125 119

Table 1
Detailed structural parameters of TWRs and operational conditions of transpiring water, adapted from Bermejo et al. [5,23] and updated.

Organism Pressure bearing wall Transpiring wall Transpiring water Ref.

FWDC (USA) Material: inconel 625 Material: inconel 600 Type: two transpiring inlets for transpiring water [12,13]
L(upper) = 0.406 m Type: platelets (seven floors) and one transpiring inlet for quench water
L(lower) = 1.219 m din = 0.152 m
ITC-CPV (Germany) Material: 1.498 steel Material: 1.4404 steel Type: two transpiring inlets for transpiring water [14–16]
L = 0.950 m Type: porous and two transpiring inlets for quench water
Dout = 0.120 m dout = 0.066 m
Din = 0.080 m din = 0.060 m Qtw = 6.5 kg h−1
dp = 3.5 × 10−5 m Ttw = 550 ◦ C
Rtw = 1.18
Qqw = 13.0 kg h−1
Tqw = 30 ◦ C
Rqw = 2.36
R = 3.54
CEA (France) Material: 316 Material: ␣-alumina Type: two transpiring inlets for transpiring water [17–19]
L(reaction) = 0.500 m Type: porous and an external jacket for cooling
L(cooling) = 0.300 m dout = 0.019 m
Din = 0.024 m din = 0.015 m Qtw = 0.6–1.8 kg h−1
dp = 8 × 10−5 m and 5 × 10−6 m Ttw = 380 ◦ C
Rtw = 1–3
ETH (Switzerland) Material: inconel 625 Material: inconel 625 Type: three independent transpiring inlets for [20,21]
transpiring water and one transpiring inlet for
quench water
L(reaction) = 0.177 m Type: porous Qtw = 3.6 kg h−1
L(cooling) = 0.104 m dout = 0.0295 m Ttw = 35–250 ◦ C
Dout = 0.095 m din = 0.022 m Rtw = 2.26
Din = 0.034 m dp = 3 × 10−6 m and 5 × 10−6 m Qqw = 216–252 kg h−1
Tqw = 28–35 ◦ C
Rqw = 60–70
R = 62.26–72.26
Uva (Spain) Material: steel Material: inconel 600 Type: one top inlet for transpiring water [5,22,23]
L = 1.500 m Type: only the central part of
the wall is porous
Dout = 0.145 m dout = 0.080 m Qtw = 10–72 kg h−1
Din = 0.100 m din = 0.074 m Ttw = 25–250 ◦ C
dp = 2.4 × 10−5 m Rtw ≈ 2
DHU (China) Material: C-276 Material: ceramic Type: one axial inlet for transpiring water [24,25]
L = 0.500 m Type: porous
Din = 0.015 m din = 0.009 m
dp = 2 × 10−6 m
SDU (China) Material: 321 steel Material: 316 L Type: three independent transpiring inlets for [26,27]
L = 0.750 m Type: porous transpiring water
Dout = 0.114 m dout = 0.060 m Qtw = 20–85 kg h−1
Din = 0.080 m din = 0.055 m Ttw = 200–350 ◦ C(upper) 100–250 ◦ C(middle)
Room temperature (lower)
dp = 2.08 × 10−5 m Rtw = 2.5–5 (approximate)
XJTU (China) Material: 316 Material: C-276 Type: six transpiring inlets for transpiring water [28,29]
L(upper) = 1.180 m Type: wire netting Ttw < 374 ◦ C
L(lower) = 0.245 m din = 0.300 m R = 10
dp = 3 × 10−5 m

transparent reactor cannot sustain the high temperature and high

pressure associated with supercritical conditions. Therefore, it is
essential to find suitable fluids whose physical properties at room
temperature and atmospheric pressure are close to those of tran-
spiring water. In our study, ethanol and acetone were selected as
substitute fluids because their physical properties are similar to
subcritical water. The primary physical properties of ethanol and
acetone are compared with water under a pressure of 25 MPa at
different temperatures (Table 2). Because both ethanol and acetone

Table 2
Physical properties of typical fluids, data collected from [33,34].

Fluids Viscosity Density Surface tension

(10−6 Pa s) (kg m−3 ) (10−3 N m−1 )

Ethanol at 0.1 MPa and 25 ◦ C 1074.0 787.2 22.0

Acetone at 0.1 MPa and 25 ◦ C 306.0 786.0 23.5
Water at 0.1 MPa and 25 ◦ C 902.9 997.0 73.4
Fig. 2. Schematic of the simulation experiment system. (1) air compressor, (2) pure
Water at 25 MPa and 95 ◦ C 304.0 973.0 59.9
feed tank, (3) feed with tracer tank, (4) pressure reducing valve, (5) valve, (6) air-
Water at 25 MPa and 270 ◦ C 103.1 792.4 21.3
mass flow meter, (7) globe valve, (8) pump A, (9) pump B, (10) pressure meter, (11)
Water at 25 MPa and 400 ◦ C 29.2 116.5 Inexistence
transparent reactor, (12)collection tank.
120 Z. Chen et al. / J. of Supercritical Fluids 88 (2014) 117–125

are transparent, the tracing technique was adapted with a colored

dye to show the dynamic processes. The color change in the annulus
shows the flow behavior of transpiring fluid.
As shown in Fig. 2, pure liquids with and without the tracer
are stored in tanks (2) and (3), respectively, which are pumped
into the reactor (11) through valve (5). Valve (5) is used to switch
between different liquids. Two plunger-metering pumps (2J-X,
Hangzhou Zhejiang Petrochemical Equipment CO., LTD.) are used
to deliver the fluids with a flow rate ranging from 1.3 × 10−7 to
1.3 × 10−6 m3 s−1 . The axial flow delivered by pump A (8) induces
an upward stream, which ensures that the reactor is fully filled with
liquid. The transpiring flow delivered by pump B (9) flows through
the porous wall, mixes with the axial flow, and exits the top of
reactor into the collection tank (12).
Air, as an alternative transpiring fluid, is compressed and de-
watered by the air compressor (1) from Atlas (GX4FF-10), and the
air is pressure regulated to about 0.1 MPa. The flow rate is con-
trolled by an air-mass flow meter (6) from SEVENSTAR (Beijing,
China), which is accurate up to the second decimal place and has
a maximum flow rate of 1.67 × 10−4 m3 s−1 under standard condi-
Fig. 3. Measure results of the viscous permeability using air as the test fluid. (Exp.
tions. 1 in Table 3).
Because the outlet of the transparent reactor is open to the atmo-
sphere, the pressure measured at the inlet of the transpiring flow
The viscous permeability coefficient (˛) can be determined
is the same as the pressure drop across the transpiring flow inlet
using Eq. (1) by measuring P and Qv . Previously, water [18] and
to the reactor outlet (P1 in Fig. 2). Two precision pressure meters
water–glucose syrup [35] were used as the test fluid to measure ˛.
(10) from SYCIF (Shanghai, China) were used to measure the pres-
The particulates in liquid have the potential to cause pore block-
sure. A pressure meter with a measuring range of 0.0 to 16.0 kPa is
age in the transpiring wall. Some non-condensable gases are likely
utilized when the transpiring fluid is liquid, while another pressure
sealed in the micro pores of the porous medium, which can lead to
meter with a measuring range of 0 to 160 kPa is utilized when the
gas locks caused by the surface tension between liquids and gases.
transpiring fluid is air. All photos are taken by a Nikon D7000.
When liquid is used as the test fluid, these two phenomena make
the accurate measurement of ˛ difficult. Consequently, according
2.2. Reagents
to BS EN ISO 4022-2006 [36], gases are more suitable than liquids
for the determination of ˛.
Ethanol (99.7%) and acetone (99.7%) are supplied by Chongqing
In this paper, air is used as the test fluid at 27 ◦ C and 0.1 MPa
Chuandong Chemical (Group) Co., Ltd. De-ionized water with
(Exp. 1 in Table 3). The average pore diameter and length of the
a resistivity approximately 18.25 M cm is obtained from a
test tube are 8 × 10−5 m and 0.107 m, respectively. The pressure
MOLATOM 1850D ultra-pure water machine.
drop due to the porous tube (P) [18] is defined as the difference
Methyl orange (Ind) and methyl red (AR) are purchased from
between the measured pressure with (P1 ) and without the tube
Sinopharm; the former is used as a tracer in water, while the latter
(P2 ). The plot of P as a function of Qv is linear and can be fit with a
is used as a tracer in both ethanol and acetone. The tracer concen-
straight line (Fig. 3, square of related coefficient is 0.9995). This lin-
tration is 0.01 g L−1 . All liquids are filtered through a membrane
earity indicates that the flow is in the laminar flow regime and that
filter with a pore size of 2 × 10−5 m.
Darcy’s law is applicable to our measurement system. The viscous
permeability coefficient (˛) of our porous tube was determined to
3. Results and discussion be 1.15 × 10−13 m2 according to Eq. (1).

3.1. Viscous permeability coefficient of transpiring wall

3.2. Dynamic process of liquids flowing through the transpiring
wall
As a porous medium, the viscous permeability of the transpiring
wall plays an important role in the dynamic process of the transpir-
In the Exp. 2–4 (see Table 3), the dynamic processes of the three
ing fluid. For a laminar flow regime, the incompressible fluid flow
fluids flowing through the porous tube were photographed. The
in a porous medium is governed by the Darcy’s law [18], in which
images at several characteristic times (in s) are chosen to show the
˛ is the viscous permeability coefficient.
dynamic processes (Fig. 4).
ı The theoretical residence time  1 is the time of transpiring flow
P = Qv (1)
˛A (radial flow) in annulus space and can be calculated according to
Eq. (4). The theoretical residence time  2 is the time of both the
The viscosity of fluid () is available from Table 2. Both the
transpiring flow (radial flow) and axial flow in the overall reactor
thickness (ı) and the test area (A) of the porous cylindrical tube
space (including the annulus and tube space, Fig. 1). The theoretical
can be calculated from the following relationships [34], using the
residence time  3 is the time of transpiring flow (radial flow) in the
measured H, din , and dout :
overall reactor space. Both  2 and  3 were defined by Fauvel et al.
 2 [17] and are expressed by Eqs. (5) and (6), respectively.
dout lndout /din
ı=   (2)
Vannulus
2 dout /din − 1 1 = (4)
  Qv,radial
dout H lndout /din
A=   (3) Vannulus + Vtube
2 = (5)
dout /din − 1 Qv,radial + Qv,axial
 Z. Chen et al. / J. of Supercritical Fluids 88 (2014) 117–125 121

Table 3
Experimental conditions.

No. Axial flow Transpiring flow

Fluid Flow rate (m3 s−1 ) Fluid Flow rate (m3 s−1 )

1 Air 0.0 Air 1.67 × 10−7 –1.67 × 10−4

2 Water with tracer 1.9 × 10−7 *
Switching from water with tracer to water without tracer 8.1 × 10−7
3 Ethanol with tracer 1.9 × 10−7 *
Switching from ethanol with tracer to ethanol without tracer 8.1 × 10−7
4 Acetone with tracer 1.9 × 10−7 *
Switching from acetone with tracer to acetone without tracer 8.1 × 10−7
5 Water with tracer 1.9 × 10−7 Water without tracer 1.6 × 10−7 –1.3 × 10−6
6 Ethanol with tracer 1.9 × 10−7 Ethanol without tracer 1.6 × 10−7 –1.3 × 10−6
7 Acetone with tracer 1.9 × 10−7 Acetone without tracer 1.6 × 10−7 –1.3 × 10−6
8 Water with tracer 1.9 × 10−7 Air 1.67 × 10−7 –1.67 × 10−4
9 Ethanol with tracer 1.9 × 10−7 Air 1.67 × 10−7 –1.67 × 10−4
10 Acetone with tracer 1.9 × 10−7 Air 1.67 × 10−7 –1.67 × 10−4
*
Switch when the reactor is full filled with colored fluid.

according to the definitions in Eqs. (5) and (6), the transpiring

fluid will flow through the whole reactor space after 1043 s ( 2 ,
considering the influence of the axial flow) or 1287 s ( 3 , ignoring
the influence of the axial flow). However, the experimental results
do not match the theoretical predictions (see Fig. 4). In particular,
colored dye is still present in the annulus after 402 s ( 1 ), and the
colored dye does not fade away in the lower section of annulus
after 1287 s ( 3 ). The colored fluid initially moves away from the
inlet, dissipates in top section, and gradually fades away in the
lower section. In this study, the color change is used to track
the streamline of the transpiring fluid. In general, a high color
dissipation rate indicates a high flow speed. The phenomena
mentioned above suggests that the majority of the transpiring
fluid flows into the porous tube through the middle and upper
section of the annulus, while the rest of the transpiring fluid flows
down to the lower section of the annulus before entering the
porous tube. The visual images (Fig. 4) confirm that the transpiring
fluid is anisotropic with respect to the surface of the porous tube.
Based on Fig. 4, a 2-D scheme (Fig. 5) was developed to depict
the transpiring fluid dynamics from the side and top cross-sectional
views of the reactor. After injection into the annulus, the transpiring
fluid simultaneously flows around the porous tube (around flow
path [37]) and flows through the porous wall into the tube (seepage
flow path [37]). The anisotropy of the transpiring fluid is partially
caused by the inlet position. However, the primary cause of the
transpiring fluid anisotropy is the flow resistance, which can be

Fig. 4. Dynamic processes of water, ethanol and acetone flowing through the tran-
spiring wall. Images at several characteristics moments (s) were captured to show
the dynamic processes. (Exp. 2–4 in Table 3).

Vannulus + Vtube
3 = (6)
Qv,radial

Based on the size of the reactor, Vannulus and Vtube are

3.25 × 10−4 m3 and 7.15 × 10−4 m3 , respectively, and the  1 ,  2 , and
 3 are 402, 1043, and 1287 s, respectively.
According to the definition in Eq. (4), the transpiring fluid will
flow through the whole annulus space after 402 s ( 1 ). Additionally,
Fig. 5. Schematic diagram of transpiring flow was visualized from Fig. 4. The tran-
spiring flow was dispersed to around-flow and seepage flow and the process is
anisotropy.
122 Z. Chen et al. / J. of Supercritical Fluids 88 (2014) 117–125
Fig. 6. Pressure drop due to the porous tube as a function of transpiring fluids flow
rate. (Exp. 5–7 in Table 3).
attributed to the static pressure (Eq. (7)) and viscous resistance (Eq.
(1)).
P = gH
The resistance of the transpiring fluid increases as the height
(H) of the transpiring tube increases because the static pressure
increases from the top (H = 0.000 m) to the bottom (H = 0.500 m)
of the transpiring tube. This explains why the color disappears in
upper section before it disappears in the lower section.
The pressure drop due to the porous tube (P) was measured
for water, ethanol, and acetone (Exp. 5–7 in Table 3). Before each
measurement, the air in both the reactor and the porous tube is
purged by liquid in order to avoid a gas lock [36]. Fig. 6 shows the
plot of P and Qv for both the experimental values and the theo-
retical values calculated from Eq. (1). All of the experimental values
are higher than the theoretical values calculated by Eq. (1). The dif-
ference between experimental and theoretical values decreases as
the liquid changes from water to ethanol to acetone. On the other
hand, for all liquids, the difference between the experimental and
theoretical values increases as the transpiring flow rate increases.
According to Darcy’s law, P is proportional to Qv and is
inversely proportional to the surface area (A) of a fluid. As shown
in Figs. 4 and 5, the transpiring fluid is anisotropic, which occurs
because the majority of the transpiring fluid (Qv ) flows through
the small surface area (A) near the transpiring fluid inlet. This
anisotropy may explain why all of the experimental values are
higher than the theoretical values (Fig. 6). In addition, it seems that
a higher Qv magnifies this effect because the increase in the tran-
spiring flow results in bigger difference between the experimental
and theoretical values for P (Fig. 6). In other words, the process
is more anisotropic at a high transpiring flow rate, which suggests
that increasing the transpiring flow may not be an effective way to
form an isotropic protective water film.
According to Eq. (1), P is also proportional to the viscos-
ity () of the fluid. The viscosity of acetone and ethanol are
3.06 × 10−4 Pa s and 1.074 × 10−3 Pa s, respectively (Table 2). The
viscosity of ethanol is three times greater than that of acetone,
and the fluid dynamics of ethanol is more anisotropic than that
of acetone (Fig. 4). The difference in viscosity can explain why the
difference between the experimental results and the theoretical
calculation of P is larger for ethanol than it is for acetone (Fig. 6).
However, this does not apply for water. The viscosity of water lies
between that of acetone and ethanol at 9.029 × 10−4 Pa s (Table 2),
but the P difference between the experimental results and the
theoretical calculation is greater than that of ethanol. Due to the
high surface tension between water and air (7.34 × 10−2 N m−1 ),
some of the porous tube pores are sealed by air, leading to a reduc-
tion of the active area (A). Moreover, these results explain why air
is chosen as the test fluid to measure the permeability of the porous
tube instead of an aqueous liquid (Section 3.1).
For acetone, the experimental P results are close to the theo-
retical calculations. Under near critical and supercritical conditions,
the viscosity of water is less than 1.000 × 10−4 Pa s, which is sig-
nificantly smaller than that of acetone (Table 2). Therefore, under
near critical and supercritical conditions, we can assume that the
differences between the experimental results and the theoretical
calculations are negligible. As a result, the pressure drop across the
porous tube follows Darcy’s law, and these findings are in agree-
ment with a previously published work [18].
As shown in Fig. 4, the colored dye completely disappears in less
than 2400 s when acetone is used as the transpiring fluid; however,
some colored solution remains in the bottom right corner of the
tube after 3600 s when both water and ethanol are used as the tran-
spiring fluid. These behaviors can also be explained by the viscous
resistance (Eq. (1)) and the static pressure (Eq. (7)), which essen-
tially correlate with the fluid density and viscosity (see Table 2).
The viscosity of ethanol is greater than water by a difference of
0.1711 Pa s, while the density of ethanol is less than water by a
difference of 209.8 kg m−3 . The counter-effect of viscosity (Eq. (1))
(7)
and density (Eq. (7)) on the fluid flow resistance causes the move-
ment of the dye in the tube to be similar for ethanol and water.
Acetone is different from both ethanol and water because its vis-
cosity is approximately one-third that of ethanol, but the densities
are comparable. Fig. 4 shows that the fluid dynamics of acetone
is more isotropic than those of water and ethanol, which suggests
that fluids with low viscosity tend to form an isotropic protective
film.
During the experiments (Exp. 5–7 in Table 3), the axial fluid
was colored, and the transpiring fluid was colorless. The annulus
remained colorless regardless of the transpiring fluid type. This
indicates that the transpiring fluid flows through the entire sur-
face of the porous tube, and there is no fluid in the porous tube
flowing back to the annulus. All of the colored fluid in the annu-
lus gradually fades away (Exp. 2–4, see Table 3 and Fig. 4), which
further verifies this conclusion.
3.3. A novel design and the feasibility analysis
Because the transpiring flow is anisotropic, most TWRs have
more than one inlet for the transpiring flow. Those designs maybe
helpful for the formation of an isotropic protective water film;
however, these systems have high costs, and a high volume of
transpiring water is required. Currently, all TWRs use water as
the transpiring fluid, which has some disadvantages such as high
energy consumption due to the high heat capacity of water and the
corrosive nature [38] of subcritical water. Therefore, a new idea is
needed.
A new reactor is designed at the Chongqing Institute of Green
and Intelligent Technology (CIGIT), Chinese Academy of Sciences,
and the transpiring fluid is air instead of water. The feasibility of
the proposed design depends on the transpiring air confinement
of the corrosive reaction mixtures in the inner space of the porous
tube, which we call the dynamic gas seal.
A series of experiments (Exp. 8–10 in Table 3) were carried out
to test the feasibility of the proposed design using the transparent
glass reactor (Fig. 2). The images of the process were captured, and
the pressure drop of the porous tube (P1 ) was measured. The
results show that none of the test liquids leave the annulus over
the whole flow range (Fig. 7). As shown in Fig. 8, the P1 of water
 Z. Chen et al. / J. of Supercritical Fluids 88 (2014) 117–125
Fig. 7. Images of the reactor. (Exp. 8–10 in Table 3).
is much higher than that of the other two liquids, which correlates
with their surface tension values (Table 2). This indicates that the
surface tension of a liquid plays a leading role in the wall transpiring
process.
Because the surface tension of ethanol is close to that of water at
25 MPa and 270 ◦ C, the results suggest that the new design should
work at temperatures below 270 ◦ C. The surface tension effects
should gradually disappear as the reactor temperature increases.
Under supercritical conditions, water and air form a single fluid
phase [39], and the gas-liquid interface no longer exists. At this
point, the axial fluid and the transpiring fluid can be considered to
be one phase. As we showed in Section 3.2, the axial fluid does not
enter the annulus when the transpiring fluid and axial fluid are the
same. Under supercritical conditions, the resistance of the water
flow through a porous tube with pore diameters of 8 × 10−5 m is
negligible [18], and the viscosity of air is approximately the same
as water (Fig. 9B). Thus, the resistance caused by viscosity may be
Fig. 8. Pressure drop with porous tube as a function of flow rate using air as the
transpiring flow. (Exp. 8–10 in Table 3).
123
Fig. 9. Physical properties of air and water at a pressure of 25 MPa versus temper-
ature, data collected from [33,40,41].
neglected under supercritical conditions. Because the densities of
air and water are different under supercritical conditions, density
plays the major role in the fluid flow resistance. In order to achieve
the dynamic gas seal effect, the static pressure in annulus must be
higher than that in reaction area (Fig. 1) at every point along the
porous tube. In other words, the density of the transpiring fluid (air)
should be greater than that of axial fluid (water) according to Eq.
(7). When the temperature is greater than 422 ◦ C (Fig. 9C), the den-
sity of compressive air is greater than that of supercritical water,
and the dynamic gas seal effect should be achieved.
Hence, the new design is feasible. The interfacial tension plays
the leading role for the fluid flow resistance under subcritical con-
ditions, while the density difference between compressive air and
water plays the leading role under supercritical conditions.
Based on the above analysis, a novel supercritical water oxi-
dation reactor was developed at CIGIT [42], named the dynamic
gas seal wall reactor (DGSWR). Sewage sludge with a solid con-
centration of 5 g L−1 was degraded at 400 ◦ C and 25 MPa with the
flow rate of 2.0 × 10−7 m3 s−1 . The chemical oxygen demand (COD)
removal efficiency reached 99.7%. After one month of intermittent
operation, no obvious corrosion in the pressure bearing wall or salt
precipitation were observed (Fig. 10), which shows that the DGSWR
effectively avoids the corrosion and salt precipitation problems
seen with TWRs.
In order to achieve a better dynamic gas seal effect, the inlet
temperature should be greater than 422 ◦ C for waste water (axial
fluid) and room temperature for air (transpiring fluid). On one hand,
the density of compressive air is greater than that of supercritical
water (Fig. 9C). On the other hand, the gas–liquid interface will
appear due to the cooling effect of the cold air. Both the density
124 Z. Chen et al. / J. of Supercritical Fluids 88 (2014) 117–125
Fig. 10. Inner surface of pressure bearing wall after one month intermittent opera-
tions under 25 MPa and 400 ◦ C.
difference and surface tension work in synergy to ensure an effec-
tive gas seal under this condition. Because of the special physical
properties of compressive air (Fig. 9), the DGSWR possesses some
advantages compared with current TWRs.
First, the consumption of high quality water is reduced. Current
TWRs use pure water as transpiring fluid. Table 1 shows that the
consumption of transpiring water is about 2 to 10 times that of
treated wastewater. However, in DGSWR, air functions as transpir-
ing fluid, and no pure water is required. At the same time, air serves
as the oxidant; therefore, no additional oxidant delivery system is
required.
Second, because the isobaric heat capacity of compressive air is
only one-fifth of the average value of water (Fig. 9A), the energy
consumption for DGSWR should be much less than current TWRs.
Therefore, less energy is required to heat the transpiring fluid. This
will also reduce the temperature fluctuation in the reaction area
and improve the reactor stability.
Third, the DGSWR has enhanced reactor protection. The analysis
in Section 3.2 suggests that fluids with low viscosity tend to form
an isotropic protective film, Fig. 9B shows that the viscosity of air
is much lower than that of water.
4. Conclusions and future work
The development of TWRs was reviewed and their problems
were discussed. The permeability coefficient of the porous tube
was determined using air to be 1.15 × 10−13 m2 . A transparent glass
reactor was used to visually study the dynamics of the transpir-
ing fluid in TWRs, which are anisotropic due to the static pressure
and viscous resistance. The DGSWR is designed to solve both the
corrosion and salt precipitation problems found in TWRs, while
minimizing the energy and high-quality water consumption asso-
ciated with current TWRs. Its feasibility was demonstrated by the
fact that there were no obvious corrosion in the pressure bearing
wall or salt precipitation observed after one month of intermittent
operation.
The methodology used in this paper provides an easy and effec-
tive way for performing fundamental fluid dynamics research in
very harsh chemical or physical environments. However, the effects
of the temperature profile on the fluid dynamics are not explored
in this paper and should be investigated in the future.
Acknowledgments
This work was financially supported by two Key Tech-
nology R&D Programs of Chongqing (grant numbers:
cstc2012gg-sfgc20001 and cstc2011ggC20014) and the 100-talent
program of the Chinese Academia of Sciences.
Appendix A. Supplementary data
Supplementary material related to this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.supflu.2014.
01.020.
References
[1] M. Modell, Processing methods for the oxidation of organics in supercritical
water, U.S. Patent #4338199 (1982).
[2] H. Schmieder, J. Abeln, Supercritical water oxidation: state of the art, Chemical
Engineering and Technology 22 (1999) 903–908.
[3] P. Kritzer, E. Dinjus, An assessment of supercritical water oxidation
(SCWO)—existing problems, possible solutions and new reactor concepts,
Chemical Engineering J. 83 (2001) 207–214.
[4] P.A. Marrone, Supercritical water oxidation-current status of full-scale com-
mercial activity for waste destruction, J. Supercritical Fluids 79 (2013) 283–288.
[5] M.D. Bermejo, M.J. Cocero, Supercritical water oxidation: a technical review,
AIChE J. 52 (2006) 3933–3951.
[6] A. Martin, M.D. Bermejo, M.J. Cocero, Recent developments of supercritical
water oxidation: a patents review, Recent Patents on Chemical Engineering
(4) (2011) 219–230.
[7] T.G. McGinness, Supercritical oxidation reactor, U.S. Patent #5384051 (1995).
[8] H.H. Mueggenburg, D.C. Rousar, M.F. Young, Supercritical water oxidation reac-
tor with wall onduits for boundary flow control, U.S. Patent #5387398 (1995).
[9] E.L. Daman, Process and apparatus for supercritical water oxidation, U.S. Patent
#5571423 (1996).
[10] K.S. Ahluwalia, Internal platelet heat source and method of use in a supercritical
water oxidation reactor, U.S. Patent #5571424 (1996).
[11] C.A. LaJeunesse, B.L. Haroldsen, S.F. Rice, B.G. Brown, Hydrothermal oxidation
of navy shipboard excess hazardous materials, in: Report No. SAND 97-8212,
Sandia National Laboratories, 1997.
[12] P.J. Crooker, K.S. Ahluwalia, Z. Fan, J. Prince, Operating results from supercriti-
cal water oxidation plants, Industrial and Engineering Chemistry Research 39
(2000) 4865–4870.
[13] P.A. Marrone, S.D. Cantwell, D.W. Dalton, SCWO system designs for waste treat-
ment: application to chemical weapons destruction, Industrial and Engineering
Chemistry Research 44 (2005) 9030–9039.
[14] J. Abeln, M. Kluth, G. Petrich, H. Schmieder, Supercritical water oxidation
(SCWO): a process for the treatment of industrial waste effluents, High Pressure
Research 20 (2001) 537–547.
[15] J. Abeln, M. Kluth, M. Bottcher, W. Sengpiel, Supercritical water oxidation
(SCWO) using a transpiring wall reactor: CFD simulations and experimental
results of ethanol oxidation, Environmental Engineering Science 21 (2004)
93–99.
[16] J. Abeln, M. Kluth, M. Pagel, Results and rough cost estimation for SCWO of
painting effluents using a transpiring wall and a pipe reactor, J. Advanced
Oxidation Technologies 10 (2007) 169–176.
[17] E. Fauvel, C. Joussot-Dubien, E. Pomier, P. Guichardon, G. Charbit, F. Charbit, S.
Sarrade, Modeling of a porous reactor for supercritical water oxidation by a res-
idence time distribution study, Industrial and Engineering Chemistry Research
42 (2003) 2122–2130.
[18] E. Fauvel, C. Joussot-Dubien, P. Guichardon, G. Charbit, F. Charbit, S. Sarrade, A
double-wall reactor for hydrothermal oxidation with supercritical water flow
across the inner porous tube, J. Supercritical Fluids 28 (2004) 47–56.
[19] E. Fauvel, C. Joussot-Dubien, V. Tanneur, S. Moussiere, P. Guichardon, G. Charbit,
F. Charbit, A porous reactor for supercritical water oxidation: experimental
results on salty compounds and corrosive solvents oxidation, Industrial and
Engineering Chemistry Research 44 (2005) 8968–8971.
[20] B. Wellig, K. Lieball, P.R. von Rohr, Operating characteristics of a transpiring-
wall SCWO reactor with a hydrothermal flame as internal heat source, J.
Supercritical Fluids 34 (2005) 35–50.
[21] B. Wellig, M. Weber, K. Lieball, K. Prikopsky, P.R. von Rohr, Hydrothermal
methanol diffusion flame as internal heat source in a SCWO reactor, J. Super-
critical Fluids 49 (2009) 59–70.
[22] M.D. Bermejo, F. Fdez-Polanco, M.J. Cocero, Effect of the transpiring wall on the
behavior of a supercritical water oxidation reactor: modeling and experimental
results, Industrial and Engineering Chemistry Research 45 (2006) 3438–3446.
[23] M.D. Bermejo, E. Fdez-Polanco, M.J. Cocero, Experimental study of the opera-
tional parameters of a transpiring wall reactor for supercritical water oxidation,
J. Supercritical Fluids 39 (2006) 70–79.
[24] W.J. Gong, F. Li, D.L. Xi, Oxidation of industrial dyeing wastewater by supercrit-
ical water oxidation in transpiring-wall reactor, Water Environment Research
80 (2008) 186–192.
[25] W.J. Gong, X.J. Duan, Degradation of landfill leachate using transpiring-wall
supercritical water oxidation (SCWO) reactor, Waste Management 30 (2010)
2103–2107.
[26] F.M. Zhang, S.Y. Chen, C.Y. Xu, G.F. Chen, C.Y. Ma, Experimental studies on
supercritical water oxidation of glucose with a transpiring wall reactor, in:
 Z. Chen et al. / J. of Supercritical Fluids 88 (2014) 117–125
Proceedings of 2011 Asia–Pacific Power and Energy Engineering Conference,
IEEE, Piscataway, NJ, 2011.
[27] F.M. Zhang, S.Y. Chen, C.Y. Xu, G.F. Chen, J.M. Zhang, C.Y. Ma, Experimental study
on the effects of operating parameters on the performance of a transpiring-wall
supercritical water oxidation reactor, Desalination 294 (2012) 60–66.
[28] D.H. Xu, S.Z. Wang, Y.M. Gong, Y. Guo, X.Y. Tang, H.H. Ma, A novel concept
reactor design for preventing salt deposition in supercritical water, Chemical
Engineering Research and Design 88 (2010) 1515–1522.
[29] D.H. Xu, S.Z. Wang, X.Y. Tang, Y.M. Gong, Y. Guo, Y.Z. Wang, J. Zhang, Design of
the first pilot scale plant of China for supercritical water oxidation of sewage
sludge, Chemical Engineering Research and Design 90 (2012) 288–297.
[30] M.D. Bermejo, A. Martin, J.P.S. Queiroz, I. Bielsa, V. Rios, M.J. Cocero, Computa-
tional fluid dynamics simulation of a transpiring wall reactor for supercritical
water oxidation, Chemical Engineering J. 158 (2010) 431–440.
[31] R.M. Serikawa, T. Usui, T. Nishimura, H. Sato, S. Hamada, H. Sekino, Hydro-
thermal flames in supercritical water oxidation: investigation in a pilot scale
continuous reactor, Fuel 81 (2002) 1147–1159.
[32] K. Prikopsky, Characterization of continuous diffusion flames in supercritical
water, in: Doctoral Thesis No. 17,374, ETH, Zurich, Switzerland, 2007.
[33] W. Wagner, A. Kruse, International Steam Tables—Properties of Water and
Steam Based on the Industrial Formulation IAPWS-IF97, second ed., Springer,
Berlin, 2007, pp. 258–259.
125
[34] H.Z. Lu, et al., Handbook of Petrochemical Basic Data, first ed., Chemical Industry
Press, Beijing, 1982, pp. 28–92.
[35] B. Wellig, Transpiring wall reactor for supercritical water oxidation, in: Doctoral
Thesis No. 15,038, ETH, Zurich, Switzerland, 2003.
[36] BS EN ISO 4022:2006, Permeable Sintered Metal Materials—Determination of
Fluid Permeability, BSI, London, 2006.
[37] A.V. Gavrilov, I.V. Shirko, Fluid flow in an inhomogeneous granular medium, J.
Applied Mathematics and Mechanics 74 (2010) 267–277.
[38] P. Kritzer, Corrosion in high-temperature and supercritical water and aqueous
solutions: a review, J. Supercritical Fluids 29 (2004) 1–29.
[39] E.U. Franck, Fluids at high-pressures and temperatures, J. Chemical Thermody-
namics 19 (1987) 225–234.
[40] E.W. Lemmon, R.T. Jacobsen, S.G. Penoncello, D.G. Friend, Thermodynamic
properties of air and mixtures of nitrogen, argon, and oxygen from 60 to 2000 K
at pressures to 2000 MPa, J. Physical and Chemical Reference Data 29 (2000)
331–385.
[41] E.W. Lemmon, R.T. Jacobsen, Viscosity and thermal conductivity equations
for nitrogen, oxygen, argon, and air, International J. Thermophysics 25 (2004)
21–69.
[42] S. Yang, G.W. Wang, Y.J. Xu, New design of supercritical water oxidation reactor
for sewage sludge treatment, Advances in Materials Research 774–776 (2013)
212–215.