ECI Symposium Series, Volume RP5: Proceedings of 7th International Conference on Heat Exchanger Fouling and Cleaning - Challenges and Opportunities,
Editors Hans Müller-Steinhagen, M. Reza Malayeri, and A. Paul Watkinson, Engineering Conferences International, Tomar, Portugal, July 1 - 6, 2007
METALLIC MICRO HEAT EXCHANGERS:
PROPERTIES, APPLICATIONS AND LONG TERM STABILITY
J.J. Brandner1), W. Benzinger, U. Schygulla, S. Zimmermann and K. Schubert
Forschungszentrum Karlsruhe, Institute for Micro Process Engineering (IMVT),
P.O.Box 3640, DE-76021 Karlsruhe, Germany
1)
corresponding author: Phone +49 7247 82 3963, Fax +49 7247 82 3186,
e-mail: juergen.brandner@imvt.fzk.de
ABSTRACT
Micro heat exchangers, which until recently have been
implemented only at laboratory scale, are now being
available for industrial applications. They are well known
for their superior heat transfer properties due to the large
surface-to-volume ratio. But there are little data available on
the long term stability of these devices.
In this paper application several application examples
for micro heat exchangers made of stainless steel are
presented. The devices consist of stainless steel foils
providing numerous micro channels generated by
mechanical micromachining or wet chemical etching. A
number of the foils are arranged in a specific way and
bonded together.
Device property descriptions as well as some possible
application examples show the potential of metallic
microstructure devices.
Results on two crossflow microstructure heat
exchangers running in long term tests are presented. Both
devices have been tested for more than 8000 hours each,
using deionised water as test fluid. Experimental data on the
heat transfer properties and the pressure drop are given and
compared. It was found that the heat transfer capabilities
were significantly decreased within the first few hundred
hours of testing and then run into a saturation state.
Performance degradation may be due to a fouling layer
deposited on the heat exchange surface. Some other
experimental applications in which fouling was expected to
cause problems are described briefly.
INTRODUCTION
Since the early 90s, microstructure heat exchangers
have been widely acknowledged as excellent tools for
laboratory research, but only as a niche product for
industrial applications. More recently, microstructure heat
exchangers have been designed, manufactured and tested
that provide the potential for industrial use.
To design those devices, conventional pre-calculation
techniques like simple fluidmechanical equations based on
the Nusselt theory can be used, as it was shown by
Halbritter et al. (2004). Beside this, Wenka et al. (2000)
showed that CFD calculations may be advantageous to use
for the primary design and to obtain a first idea of the heat
transfer performance of such microstructure devices.
Nevertheless, the manufacturing process is determined by
the possibilities to obtain micro channels or microstructures,
the bonding methods and of course by the needs of the
processes the devices are meant for. Moreover, the process
itself is limited by the long term stability of the
microstructure devices against corrosion, blocking and
fouling.
In this paper, the design and manufacturing technique
of microstructure heat exchangers is described. Examples of
microstructure heat exchangers and their potential are given
by experimental results. Experimental results for the long
term stability of two crossflow type microstructure heat
exchangers are given in detail. Some more experimental
results with regard to fouling and long term stability
obtained with the same or other designs of microstructure
heat exchangers are presented. A summary and short
outlook complete the paper.
MANUFACTURING
Most processes running outside of a laboratory will
need considerably high pressure resistance as well as
resistance against corrosion. This means that the use of
silicon as base material is not longer a suitable choice.
Metallic microstructure heat exchangers are necessary to
provide the demanded resistivity against pressure,
temperature and corrosion. Therefore, it is necessary to
choose a metal which can be microstructured quite easily
and can also be bonded together to form stable and more or
less complex structures. Due to the higher corrosion
resistance, stainless steel is the first choice. Micro channels
or other designs of microstructures in stainless steel foils
may be obtained either by mechanical precision
micromachining (e.g. milling, turning or drilling), as it is
described by Schaller et al. (1999) or by wet chemical
etching as described by Madou (2002). The cross section of
the channels differs by the manufacturing method. While
mechanically cut channels will take on the shape of the
microtool used to cut, wet chemically etched micro channels
have a semi-elliptic cross section due to the isotropic
Produced by The Berkeley Electronic Press, 2016
384
Heat Exchanger Fouling and Cleaning VII
[2007], Vol. RP5, Article 49
etching effect. In Figure 1, a SEM photo of a mechanically
cut microchannel foil with rectangular micro channels is
shown, while Figure 2 shows a SEM picture of an etched
microchannel stainless steel foil.
Depending on the manufacturing method, the surface
quality differs. While mechanically cut microchannels in
stainless steel provide an overall roughness of about 3 µm,
wet chemically etched stainless steel microchannels show
about three times this value – around 9 µm is common.
This, of cause, is to be taken in account when fouling takes
place.
To form the core of a microstructure heat exchanger, a
number of microstructured foils is stacked together in the
desired order and arrangement. The complete stack is then
brought into a furnace and bonded by a diffusion bonding
process. Depending on the material and the connecting face
area, a strictly defined temperature and force program is
applied to run the diffusion bonding process. The forces
used for diffusion bonding are in the range from a few kN
up to around 100 kN, depending on the connection area,
while the applyable temperature range is from 500°C to
more than 1,000°C.
Fig. 2: SEM picture of a wet chemically etched
microstructure foil. The semi-elliptic micro channels
are about 200 µm wide and around 70 µm deep.
Fig. 1: SEM picture of a mechanically cut microstructure
foil. The rectangular micro channels are 200 µm wide
and around 100 µm deep.
Fig. 3: SEM picture of full-elliptically shaped micro
channels of a crossflow micro heat exchanger. The
good alignment is clearly to see.
For microstructure foils, two possibilities for stacking
exist: either the simple stacking (called “face-to-back”), or
to stack them “face-to-face”. Those two possibilities are of
special interest for wet chemically etched micro channels,
were the “face-to-back” version will form semi-elliptically
shaped micro channels with small hydraulic diameters,
while the “face-to-face” option can be used to generate fullelliptically shaped micro channels, as shown in Figure 3.
After the diffusion bonding step, the microstructure
device is similar to a monolithic structure, as shown in
Figure 4, where the micro channel system of a crossflow
heat exchanger is shown. The second passage is laying
across the visible face areas of the channels.
The displacement of the micro channels can be reduced by
aligning the single foils on one or more pins made to fit in
holes in the corners of each foil.
http://dc.engconfintl.org/heatexchanger2007/49
Brandner et al.:
385
the leak tightness, the overpressure resistance and other
parameters. A more extensive description of manufacturing
of microstructures and microstructure devices can be found
in Brandner et al. (2006a).
PRECALCULATION AND MODELING
As mentioned above, for most devices the precalculation using standard Nusselt theory equations is
suitable in many cases. Therefore, the hydraulic diameter of
each microchannel can be calculated using the simple
equation
Fig. 4: Crossflow microstructure heat exchanger made of
stainless steel after diffusion bonding. Only small
hints of the single layer structure are visible.
The core bodies of the microstructure devices are
cleaned and welded into a housing by electron beam
welding. The housing may be equipped either with
conventional tube fittings or any other fitting system
suitable for the desired application. In Figure 5, an example
photo of a crossflow heat exchanger core with an active
volume of 1 cm3 and another device welded to standard
fittings is shown.
dh =
4⋅ F
C
(1)
where dh is the hydraulic diameter, F is the cross section
area of the channel and C is the circumference of the micro
channels.
For further calculations with non-circular channels, a
correction factor has to be applied to approximate a circular
cross section. Those factors may be found, e.g., in VDI
Wärmeatlas (1994). For the pressure drop, equations like
the Darcy equation for laminar flow with a Reynolds
number Re lower than approx. 2300 or an equation
developed by Petukhov and Gnielinski for turbulent flow
with a Reynolds number Re between 3000 and 105 may be
applied (see VDI-Wärmeatlas (1994)).
For a microstructure heat exchanger having a defined
geometry and known mass flows as well as inlet and outlet
temperatures, the heat transfer therefore was calculated by
the power balance equation (2).
m& ⋅ c p ⋅ ∆T = k ⋅ A ⋅ ∆ϑ m
(2)
Here, m& is the mass flow, cp the specific heat capacity,
∆T the temperature difference, k the overall heat transfer
coefficient, A the heat transfer area and ∆ϑm the mean
logarithmic temperature difference.
The effective heat transfer area A was calculated by
A = n f ⋅ nCh ⋅ C ⋅ l
Fig. 5: Core of a stainless steel crossflow microstructure
heat exchanger (left) and a similar core welded to
conventional tube fittings. The active heat transfer
volume of both devices is 1 cm3.
Throughout the complete manufacturing process a
quality control is stringent, e.g. measurements are made of
(3)
were nf is the number of microstructured foils, nCh is the
number of microchannels per foil, C is the circumference of
each microchannel and l is the length of the microchannels.
∆ϑm was calculated from the formula given in VDIWärmeatlas (1994).
The heat transfer coefficient was defined for both a
cold and a hot passage providing the same heat transfer area
Produced by The Berkeley Electronic Press, 2016
386
Heat Exchanger Fouling and Cleaning VII
like it is shown in equation (4), in which α is the heat
transfer coefficient, Nu is the Nusselt number and λFluid is
the thermal conductivity of the fluid. Then, the overall heat
transfer coefficient k was calculated using equation (5) with
s being the foil thickness.
α=
Nu ⋅ λ Fluid
dh
(4)
s
1
1
1
=
+
+
k α hot α cold λ solid
(5)
For laminar flow (Re < 2300), the Nusselt number was
calculated using equations (6) to (9), where Pr is the Prandtl
number.
(
Nu lam = 3 Nu13 + 0.7 3 + (Nu 2 − 0.7 ) + Nu 33
3
Nu1 = 3.66
)
(6)
(7)
Nu 2 = 1.615 ⋅ 3 Re ⋅ Pr
dh
l
(8)
1
6
d
2
⋅ Re ⋅ Pr h
Nu3 =
+
⋅
l
1
22
Pr
(9)
For turbulent flow (Re > 2300), the Nusselt number
was defined by equations (10) to (11). Here, ξ is a form
factor (see VDI Wärmeatlas (1994)).
ξ
Nuturb =
⋅ (Re − 1000 )⋅ Pr
d 23
⋅ 1 + h
ξ 2 3 l
⋅ Pr −1
1+12.7 ⋅
8
8
ξ = (1.82 ⋅ log(Re ) − 1.64 )−2
(10)
(11)
When fluids with similar specific heat capacity are
applied to both passages, the heat transfer efficiency ε can
be used easily for comparison of differently designed
(microstructure) devices. A definition is given in equation
(12) or can be taken from VDI Wärmeatlas (1994) or
Wagner (1999).
ε=
Q&
m& ⋅ c mean ⋅ (Thot ,in − Tcold ,in )
[2007], Vol. RP5, Article 49
.
.
In equation (12), Q is the thermal power, m the mass
flow, T the temperature defined by the subscripts and cmean
the mean specific heat capacity.
In the specific case of symmetric fluid flow and using
the same fluid for both passages, the capacity flow of both
fluids is almost the same. Then, equation (12) can be
rearranged to calculate the heat transfer efficiency ε with
equation (13), which is given in VDI Wärmeatlas (1994).
ε=
Tcold ,out − Tcold ,in
Thot ,in − Tcold ,in
(13)
For modeling of the microstructure devices, either the
commercially available CFD software FLUENT or the
simple Nusselt theory equations have been used. With these
equations, a complete microstructure device, which means
the plenum outlet temperature at a certain mass flow, was
precalculated. Using FLUENT, a temperature distribution
for a system consisting of two microchannel foils only was
performed and therefore the temperature distribution across
the complete area of the foils was calculated, as it is
described by Halbritter et al. (2004) and Schubert et al.
(2001).
In Figure 6 an example for the temperature distribution
obtained for an arrangement of two crossflow
microstructure heat exchanger foils is shown. Here, an
asymmetric mass flow of 50 kg ⋅ h-1 for the cold water
passage and 700 kg ⋅ h-1 for the hot water passage was
assumed. The temperature distribution was calculated and
plotted as function of the distance from the inlet and the
channel number.
DEVICE PROPERTIES AND APPLICATIONS
Microstructure devices manufactured in the described
way provide unique properties. After the diffusion bonding,
a monolithic device core is generated out of single metallic
foils. Thus, the devices are stable against high pressures due
to the plane connection between the single foil layers.
Pressure resistance up to 100 MPa at room temperature is
possible, as it is described by Schubert et al. (2001). The
devices can be scaled to the desired mass flow range,
reaching from some grams up to several tons liquid per hour
and passage of the device. The pressure drop resulting on
the microstructure passages is in the range of up to 1 MPa,
depending on the density and viscosity of the fluid.
(12)
http://dc.engconfintl.org/heatexchanger2007/49
Brandner et al.:
387
Although one of those devices is running continuously
within a production facility for more than one year, we
could not observe any fouling inside the microstructure
system yet.
hot
cold
Fig. 6: Temperature distribution inside a two foil crossflow
arrangement at an asymmetrical water mass flow
(50 kg ⋅ h-1 cold water, 700 kg ⋅ h-1 hot water mass
flow), calculated using FLUENT.
By parallelization of the devices, mass flow ranges of
some tons per hour and per passage can be reached. In
Fig. 7, a stack of five crossflow heat exchanger devices is
shown. With this stack, a maximum mass flow of 35 tons of
water per passage and per hour is possible. Thus, a
maximum thermal power of up to 1 MW can be transferred.
Devices providing an active volume of 1 cm3 can
transfer a thermal power of up to 20 kW, measured with
water of 95°C in one passage and water of 8°C in the other,
at a mass flow of about 700 kg/h each, as it is reported by
Brandner et al. (2006b). The overall heat transfer coefficient
k of the devices with linear microchannels is in the range of
20 kW/(m2 ⋅ K) and can be increased to values about
56 kW/(m2 ⋅ K) and more by changing the design of the
microchannels. A surface-to-volume ratio of up to
30000 m2/m3 can be provided, corresponding to Brandner et
al. (2006b).
With an electrically powered stainless steel heat
exchanger device, a biodiesel flow is continuously heated
from 80°C to 105°C at a mass flow of 1000 kg/h, as it is
presented in Rinke et al. (2006). The manufacturing and
working principle of the device is given by Henning et al.
(2004). In Fig. 8, an electrically powered microchannel
device used for this application is shown, providing a
maximum electrical power of 20 kW.
Fig. 7: Stack of five crossflow microchannel heat
exchangers. 1 MW thermal power can be transferred
at a water mass flow of 35 tons per hour and per
passage.
Fig. 9 shows a specially designed chemical modular
microstructure reactor for the production of fine chemicals.
This device was developed in cooperation with the DSM
company and was used in several campaigns in commercial
production. At a maximum reactant mass flow of 1700 kg/h,
a thermal power of several hundred kW could be transferred
within a number of crossflow microchannel heat exchanger
modules. They have been combined with a number of
specially designed micro mixers to provide fast and
Produced by The Berkeley Electronic Press, 2016
388
Heat Exchanger Fouling and Cleaning VII
sufficient mixing and temperature control for the reaction.
More details can be found in Kraut et al. (2006).
More examples for applications are given in literature,
e.g. in Hessel et al. (2006).
Fig. 8: Electrically powered microchannel heat exchanger
used to heat a production mass flow of 1000 kg/h
biodiesel.
[2007], Vol. RP5, Article 49
EXPERIMENTAL SETUP
For experimental tests of microstructure heat
exchangers a test system with hot and cold water as fluids,
ducted in closed loops, was used. Metal frit filters were used
to clean the water from particles. Results of a water analysis
are given in Table 1.
The system was made from conventional components
as far as it was possible. The pumps were conventional
centrifugal pumps generating a maximum pressure of up to
1.6 MPa and covering a throughput range from 5 kg ⋅ h-1 up
to 1000 kg ⋅ h-1. Throughputs in the range from 0.4 kg ⋅ h-1
to 6 kg ⋅ h-1 can be realized by the use of HPLC pumps.
For the hot water loop, an electrical heater with a
heating power of 30 kW was used. For low throughputs, an
electrically powered microstructure flowthrough heater was
applied directly in front of the microstructure heat
exchanger to be tested. This was necessary to provide a
constant inlet temperature.
For the cold water loop, a plate heat exchanger with a
maximum power of 30 kW was used. The temperatures of
the loops are set to 10°C and 95°C, measured at the inlet of
the micro heat exchanger, at a pressure of 0.9 MPa.
The mass flow through the microstructure heat
exchangers are coarsely set by a bypass valve adjustment. A
fine trimming is done by mass flow controllers.
Table 1. Chemical analysis results of water used in the
experimental setup.
total
electrical
conductivity
Ca
Mg
Al
Mn
Cu
Zn
Fig. 9: Microstructure reactor for chemical production. In
cooperation with the company of DSM, this device
was developed and manufactured. It is about 650 mm
long and weighs 290 kg. Several thousand
microstructures are integrated to provide excellent
mixing and heat transfer.
30.55
µS/cm
2.400
1.610
0.040
0.031
0.029
0.085
mg/l
mg/l
mg/l
mg/l
mg/l
mg/l
The sensors used in the test rig are membrane pressure
transducers with an accuracy of 0.25% of the end value
(range: 0 – 1.6 MPa), PT100 resistance sensors with a
resolution of 0.7 K, Type K thermocouples with a resolution
of 1 K and mass flow meters based on the Coriolis
principle. The accuracy of the mass flow meters is 1% o.r.
for the flow range up to 10 kg ⋅ h-1 and decreasing to about
0.03% o.r. for 100 kg ⋅ h-1 and above. A schematic view of
the test system is shown in Figure 10.
http://dc.engconfintl.org/heatexchanger2007/49
Brandner et al.:
389
pressure drop over the hot and cold water passage at
different mass flow values have been taken as function of
the time. From this, the transferred thermal power and the
overall heat transfer coefficient at the same mass flow
values have been calculated as function of time.
Table 2: Technical data of the crossflow microstructure heat
exchangers compared in this publication.
Fig. 10: Schematic view of the test system for
microstructure heat exchangers.
With this setup, microstructure heat exchangers are
tested in a symmetrical way, which means the mass flow
through the hot and cold water passage is kept at an almost
constant value for a certain time. Plenum temperatures and
pressures at the inlet and the outlet of each passage are
measured. The data are saved in a computerized data
archiving system using LABView software. For each
microstructure heat exchanger, the mass flow is increased as
high as possible, e.g. up to a level where the pressure drop
measured over one passage has reached a value around
0.8 MPa. The experiments have been done at least twice
(one time with increasing mass flow, one time with
decreasing mass flow) to obtain consistent data and to
exclude hysteresis effects.
EXPERIMENTAL RESULTS
Results for different experiments with crossflow
microchannel devices as well as electrically heated devices
are given in the following section.
Comparison of Crossflow Microstructure Heat
Exchangers
In this section the focus is set to the comparison of two
crossflow microstructure devices with an active volume of
1 cm3. The design data for those devices are given in Table
2. Both devices have been used for several thousand hours
in long term run within the experimental setup shown
above.
For both devices, integral measurements have been
taken. the maximum mass flow per passage as well as the
66082
1246
channel
rectangular
fully elliptic
cross section
channel
200 µm
160 µm
dimensions
100 µm
100 µm
(width,
14 mm
14 mm
height,
length)
hydraulic
133 µm
121 µm
diameter
number of
850
1118
channels per
passage
effective
0.0051 m2
0.00868 m2
heat transfer
area
material: W316L, heat conductivity: 15 W / m⋅ K
Unfortunately, it was not possible to look inside the
microstructure heat exchangers without destroying them.
Thus, although it can be seen from the measurement values
that some performance degradation occurred probably due
to fouling inside the microchannel system, it was not
possible to confirm this.
Experimental Results for the Comparison of
Crossflow Heat Exchanger Devices. Fig. 11 presents the
maximum mass flow plotted over the time for both the hot
and the cold water passage of the rectangular channel
device. It is clearly to see from Fig. 11 that within the first
1000 hours a significant decrease of the maximum
applyable mass flow is shown, which takes place in both
passages simultaneously. After about 1600 hours of
operation, the maximum mass flow is almost constant.
Fig. 12 presents the same data for the fully elliptic
channel device. Here, the maximum mass flow in the cold
water passage is almost constant, while the maximum mass
flow through the hot water passage decreases significantly
within the first 600 hours of operation, and is then slowly
increasing again. This gives a clear hint that there has to be
some fouling inside the hot water passage, since the
Produced by The Berkeley Electronic Press, 2016
390
Heat Exchanger Fouling and Cleaning VII
experimental setup was not changed and worked perfectly
within the normal uncertainty ranges.
[2007], Vol. RP5, Article 49
8
pressure drop 100kg/h (hot)
7
pressure drop 200kg/h (hot)
6
rectangular_warm
rectangular_cold
800
mass flow [kg/h]
700
600
∆p [bar]
pressure drop 100kg/h (cold)
900
5
pressure drop 200kg/h (cold)
4
3
2
500
1
400
0
300
0
2000
4000
6000
8000
10000
time [h]
200
100
0
0
2000
4000
6000
8000
10000
time [h]
Fig. 11: Maximum mass flow in both passages of the
rectangular channels device, plotted versus the total
operation time. The plots are laying above eachother.
300
The processes happening inside the hot water passage
seems to be reversible, since the maximum water mass flow
as well as the pressure drop starts to recover slowly with
increasing operation time. This can be seen in Fig. 12 and
Fig. 13. It is not clear yet whether the original values will be
reached anymore, but the experiment is still running, and
measurements are taken every 1000 hours of operation time.
Nevertheless, it is very unlikely that the values will recover
to the starting point values.
In Fig. 14, the overall heat transfer coefficients for the
rectangular and the elliptic device are shown, plotted versus
the total operation time. The values for two mass flows are
presented, 100 kg/h and 200 kg/h of the warm water
passage. It is clearly to see that the overall heat transfer
coefficient is reduced drastically within the first few
hundred hours of operation and then remains almost
constant for the complete operation time. Thus, the heat
transfer capabilities of the devices are first reduced by about
5% to 10%, but then kept at this level. No further decrease
is shown.
200
Crossflow Heat Exchangers for Bio-Technology
100
Within a first study, a stainless steel crossflow heat
exchanger was tested for its suitability to instantaneously
cool down a sample flow of a mammalian cell culture
suspension from 37°C to 0°C by the use of a mixture of
ethylene glycol and water as cooling fluid. This was
necessary to almost stop the cell metabolism and to analyse
the metabolites inside the cells. In Fig. 15, the temperature
dependency of molar densities of four metabolites plotted
against the temperature are given as examples. Here it is
clearly to see that with increasing sample temperature the
molar mass of metabolites is decreasing in the most cases.
In Fig. 13, the pressure drop measured over the hot
water passage and the cold water passage of the elliptic
channel device is given for a mass flow of 200 kg/h and
100 kg/h, plotted versus the total operation time. The steep
increase of the hot waters pressure drop indicates some kind
of fouling taking place inside the hot water passage, while
only a small increase along the time is seen for the cold
water passage.
600
elliptic_warm
elliptic_cold
500
mass flow [kg/h]
Fig. 13: Pressure drop plot versus the total operation time
for the hot and cold water passage of the elliptic
channel device.
400
0
0
2000
4000
6000
8000
10000
time [h]
Fig. 12: Maximum mass flow in both passages of the elliptic
channels device, plotted versus the total operation
time.
http://dc.engconfintl.org/heatexchanger2007/49
Brandner et al.:
391
Thus, to observe the intracellular density, a snapshot-like
sampling and analysis procedure is necessary.
environment after cleaning was ducted through a metal
mesh filter and then through the micro heat exchanger. No
significant fouling was observed.
14
k_rectangular, 100kg/h
k_rectangular, 200kg/h
k_elliptic, 100kg/h
k_elliptic, 200kg/h
12
G 6PF6P
DHAP
NAD
NADP
350
intracellular metabolite / µM
k [kW/(m2 K)]
10
8
6
4
2
0
0
2000
4000
6000
8000
10000
time [h]
300
250
200
150
100
50
0
-5
Fig. 14: Overall heat transfer coefficient k for the
rectangular and the elliptical channels device, plotted
versus the operation time, for two different mass flow
values.
It was not clear whether a large number of the cells will
be damaged by the heat exchanger or fouling of the
microchannels will take place due to the agglomeration of
the cells within the suspension. Moreover, it was not clear
whether the heat exchanger could be cleaned well enough to
run the device for more than only one cell sample or
different cell culture types.
Experimental Results with a Crossflow Heat
Exchanger for Bio-Technology. Within the study, Wiendahl
et al. (2007) showed that the number of disrupted cells is
negligible small and, thus, the heat exchanger is suitable for
cell culture sampling. No relevant fouling or blocking was
observed while the samples have been taken, and no cell
agglomeration was found. Moreover, cleaning and
fumigation was easy to perform with conventional methods
like, for example, an autoclave. It was shown that the
sampling method is suitable for gathering relevant
intracellular concentration data with a precision not known
yet. Thus, micro heat exchangers provide great advantages
compared to the standardized sampling methods.
Other Experiments
Other experiments with microstructure devices have
been performed. A crossflow heat exchanger was used in a
side-arm of the wastewater cleaning facilities of the
Karlsruhe Research Center. For a couple of months, a small
part of the waste water running from the facilities to the
0
5
10
15
20
25
30
35
sam ple tem perature / °C
Fig. 15: Dependency of the intracellular density of four cell
metabolites against the sampling temperature.
A special type of electrically powered device was tested
with a defined solution of CaCO3. The device was running
in a test loop, heating it up to a temperature of 80°C for two
minutes and cooling it down again then. This was repeated
several tens of thousand times. After about 60 kg of the
CaCO3 solution was pumped through the device, the
pressure drop increased significantly and the heat transfer
opportunities was decreased significantly by a deposited
calcination fouling layer. Thus, a cleaning step was
arranged, using conventional citric acid to clean the device.
After this step, the experiment could be continued and the
device was regenerated completely.
Some more experiments are still running. More results
about fouling in microstructure devices, prevention of
fouling and possible cleaning methods are expected in near
future.
CONCLUSIONS
Fouling and blocking in microstructure devices is
definitely a point to keep in mind. The microstructures may
be in the same size range as blocking particles are.
Generation of fouling films inside of microstructure systems
may lead not only to less performance with respect to heat
transfer but also to complete blocking and, thus, the
possibility of a non-working device as can happen in
conventionally sized heat exchangers.
Produced by The Berkeley Electronic Press, 2016
392
Heat Exchanger Fouling and Cleaning VII
Unlike those, most microstructure devices can not be
disassembled and cleaned as single parts. Thus, on-stream
cleaning procedures have to be developed.
There are certain countermeasures for conventional
devices dealing with fouling and blocking, such as filters,
protective layers, etc. Some of those may also be useful for
microstructure devices, some are not applicable. Due to the
huge variety of different sizes, types, manufacturing
methods and applications of microstructure systems, there
are lots of different more or less important points to deal
with when talking about long term stability.
Like with the corrosion resistance, it has to be taken in
account that the restraints and limitations are different from
those existing for conventional sized devices, and, thus, new
regulatories have to be generated. Moreover, special designs
have to be developed and applied to prevent fouling inside
microstructure devices.
OUTLOOK
For fouling and blocking in microstructure devices as
well as for their long term stability, new design concepts
and manufacturing methods as well as integrated
countermeasures could lead to a significant better
acknowledgement of this technology in the industry. In
many cases, not only a lack of experimental data but also
the fear to run into difficulties generated by fouling in
microstructures often prevents the use of these highperformance devices. Overcoming this restraints will be one
of the most important tasks in the development of
microstructure devices for the next few years.
NOMENCLATURE
Latin symbols
A
C
F
Nu
Pr
effective heat transfer area m2
microchannel circumference m
microchannel face area
m2
Nusselt number
Prandtl number
.
Q
Re
T
cmean
cp
dh
k
l
.
m
nCh
thermal power
Reynolds number
temperature
mean specific heat capacity
specific heat capacity
hydraulic diameter
overall heat transfer coeff.
microchannel length
W
K, °C
J · kg-1 · K-1
J · kg-1 · K-1
m, µm
W · m-2 · K-1
m
mass flow
kg · h-1
number of microchannels per foil
nf
s
[2007], Vol. RP5, Article 49
number of microstructured foils per passage
path length, foil thickness m
Greek symbols
α
λ
ε
ξ
heat transfer coefficient
thermal conductivity
heat transfer efficiency
mean logarithmic
temperature difference
form factor
Subscripts
cold
fluid
hot
in
lam
out
solid
turb
cold water passage
specified fluid
hot water passage
inlet position
laminar
outlet position
solid material
turbulent
∆ϑm
W · m-2 · K-1
W · m-1 · K-1
%
K
REFERENCES
Brandner, J.J., Microfabrication in Metals and
Polymers, in: Advanced Micro- and Nanosystems Vol.5, eds.
Korvink et al., Wiley-VCH, Weinheim, Germany, 2006, pp.
267-320
Brandner, J.J., Anurjew, E., Bohn, L., Hansjosten, E.,
Henning, T., Schygulla, U., Wenka, A., Schubert, K., 2006,
Concepts and Realization of microstructure heat exchangers
for enhanced heat transfer, Experimental Thermal and Fluid
Science 30, 801-809 (2006)
Halbritter, A., Schygulla, U., Wenka, A. and Schubert,
K., 2004, Comparison of Numerical Simulation and
Experimental Results for Crossflow and Counterflow
Microchannel Heat Exchangers, Proc., 2nd Int. Conf. On
Microchannels and Minichannels, S.G. Kandlikar et al.,
eds., ASME, New York, N.Y. 10016, USA, pp.967-975, 2004
Henning, T., Brandner, J.J., Schubert, K., 2004,
Characterisation of electrically powered micro heat
exchangers, Chem. Eng. J. 101/1-3, 339-345 (2004)
Hessel, V., Hardt, S., Löwe, H., 2005, Chemical Micro
Process Engineering 1+2, Fundamentals, Modelling and
Reactions / Processes and Plants, Wiley-VCH, Weinheim,
Gemany, 2005
Kraut, M., Bohn, L., Wenka, A., Schubert, K.,
Successful Upscale of laboratory micro reactor into
industrial scale, 2006, Proc. 6th ANQUE Int. Congress of
Chemistry “Chemistry and Sustainable development” 2006,
Gráficas DEVA, Spain, T2-L-9, 315-316
Madou, M, 2002, Fundamentals of Microfabrication,
2nd ed, CRC Press LLC, Boca Raton, FL, USA, 2002
http://dc.engconfintl.org/heatexchanger2007/49
Brandner et al.:
Rinke, G., Kerschbaum, S., Wenka, A., Holpe, H.,
Micro process engineering for industrial production of
biodiesel, 2006, Proc. 9th Int. Conf. on Micro Reaction Eng.
(IMRET 9) 2006, 257-258
Schaller, Th., Bohn, L., Mayer, J., Schubert, K., 1999,
Microstructure grooves with a width of less than 50µm cut
with ground hard metal micro end mills, Precision Eng. 23,
1999, 229-235
Schubert, K., Brandner, J.J., Fichtner, M., Linder, G.,
Schygulla, U., Wenka, A., 2001, Microstructure devices for
applications in thermal and chemical process engineering,
Microscale Thermophysical Engineering, 5, 2001, 17-39
393
VDI Wärmeatlas, 7.ed., VDI-Verlag, Düsseldorf,
Germany, 1994
Wagner, W., 1999, Wärmeaustauscher, 2nd ed., VogelVerlag, Würzburg, Germany, 1999
Wenka, A., Fichtner, M. and Schubert, K., 2000,
Investigations of the Thermal Properties of a Micro Heat
Exchanger by 3D Fluid Dynamics Simulation, Proc., 4th Int.
Conf. on Microreaction Technology, I. Rinard et al., eds.,
AIChE, 2000, 256-263
Wiendahl, C., Brandner, J.J., Küppers, C., Luo, B.,
Schygulla, U., Noll, T., Oldiges, M., 2007, A Microstructure
Heat Exchanger for Quenching the Metabolism of
Mammalian Cells, Chem. Eng. Technol. 2007, 30, No. 3, 18
Produced by The Berkeley Electronic Press, 2016