1 s2.0 S1369703X19302475 Main

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 10

Biochemical Engineering Journal 151 (2019) 107312

Contents lists available at ScienceDirect

Biochemical Engineering Journal


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

Computational fluid dynamic analysis of hydrodynamic shear stress T


generated by different impeller combinations in stirred bioreactor
Ritu Verma, Luv Mehan, Rahul Kumar, Arinjay Kumar, Aradhana Srivastava

University School of Chemical Technology, Guru Gobind Singh Indraprastha University, Sec-16C, Dwarka, New Delhi, 110078, India

HIGHLIGHTS

• Novel combination of impellers in bioreactor for microalgae growth.


• CFD analysis for hydrodynamic shear stress.
• Shear stress varies as 0.99–6.17 Pa with RPM change of 50–225.
• Shear sensitive microalgae cultivation.
• platensis shear stress optima as 2.64Pa at 100RPM producing 3.9g/l cells.

ARTICLE INFO ABSTRACT

Keywords: The production of microalgae has gained attention in food, health and biofuel industries. Optimized growth
Shear sensitive conditions such as agitation, aeration and nutrient concentration govern the hydrodynamic environment and
Microalgae sustainability of photoautotroph microalgae in a bioreactor. A numerical simulation using the k-epsilon model in
Impeller computational fluid dynamics (CFD) was used to investigate the hydrodynamic shear stress generated due to two
Hydrodynamic
combinations of impellers Rushton-Rushton and Rushton-marine in stirred bioreactor at different RPMs.
Bioreactor characteristics were modeled and simulated with different operational impeller speeds to compute the
shear stress generated in the fluid. The four-way flow regime in the fluid was generated by the angular motion of
Rushton-marine impellers. The rotation speeds ranging from 75 to 150 RPM exhibited shear stress of
1.97–3.96 Pa which is tolerable for large microalgae types. Further increase in RPM resulted in formation of
shear zones which was also verified experimentally by observing the growth of A. platensis at different rotational
speeds. The maximum biomass was achieved as 3.9 g/l at 100 RPM generating a shear stress of 2.64 Pa. The
observed growth of microalgae (34.46% higher) was attributed to the improved hydrodynamics in terms of shear
stress and mixing time in the bioreactor through altered impellers combination using the CFD approach.

1. Introduction effective internal mechanical agitation are the conventionally used


stirred-tank bioreactors. They are highly flexible to modifications and
Microalgae find their use in production of single cell proteins, CO2 provides high kLa (volumetric mass transfer coefficient) values for gas-
sequestration, wastewater treatment and biofuel production [1–4]. liquid transfer when pressurized air is supplied through the sparger.
However, only a few out of the 30,000 known strains find their use in The size of the gas bubbles and their dispersion throughout the tank are
industrial applications. It is essential to produce large microalgae bio- critical to the performance of the bioreactor [6,7]. The dispersion of gas
mass for maximizing the products. The cultivation conditions such as in the medium is mainly the function of the impeller instead of the
adequate nutrient concentration, pH and temperature and operating sparger. The selection of impeller and its operation plays a major role in
conditions such as illumination intensities, agitation and aeration di- the generation of flow regime in the reactor fluid which directly in-
rectly influence the growth of microalgae cells in the bioreactors [5]. fluences the growth of microalgae. The impeller must provide sufficient
The selection of cultivation system is based on effectiveness of the rapid agitation to disperse the bubbles throughout the tank. It should
above-mentioned parameters. The prime example of a reactor with also provide increased residence time of the bubbles and nutrients


Corresponding author.
E-mail addresses: ritu15verma2@gmail.com (R. Verma), luvmehan@gmail.com (L. Mehan), rajput060895@gmail.com (R. Kumar),
arinjaykumar@gmail.com (A. Kumar), aradhana.usct@ipu.ac.in (A. Srivastava).

https://doi.org/10.1016/j.bej.2019.107312
Received 12 February 2019; Received in revised form 20 July 2019; Accepted 22 July 2019
Available online 23 July 2019
1369-703X/ © 2019 Elsevier B.V. All rights reserved.
R. Verma, et al. Biochemical Engineering Journal 151 (2019) 107312

within the liquid along with breaking larger bubbles into smaller ones. bioreactors which have supreme well mixed zones in the vessel. It
However, inadequate stirring and multiple impeller systems are often provides equitable accuracy in results for an extensive range of turbu-
detrimental owing to the shear sensitivity exhibits by the microalgae lent flows [18,19]. Its stable solution makes it preferable over other
cells to varying degrees [8–10]. Although high hydrodynamic stress and highly refined models including shear stress transport (SST) which is
turbulence generated by the Rushton type impellers and larger bubbles considered in case of low Re number. This software package is freely
released from the ring type sparger in conventional bioreactors is fa- available for students with an educational license (https://www.
vorable for the growth of various micro-organisms such as Tetraselmis autodesk.com/education/free-software/cfd). Fig. 1(a)–(c) depicts the
suecica (which can tolerate shear stress up to 80 Pa) but not suitable for front, top and bottom view of the complete agitation and aeration as-
the shear sensitive microalgae species like Chaetoceros muelleri, Skele- sembly. Rushton impeller has a diameter of 5.5 cm with blades di-
tonema costatum and Isochrysis galbana whose viability has been re- mensions as 1.5 cm x 1.8 cm. Marine impeller has a diameter of 7 cm
ported to decrease gradually from 1.2 to 5.4 Pa [11]. Viability of with 45° angular blades. Both the impellers are attached to the same
Chaetoceros muelleri is found to decrease by 52–66% at the shear stress shaft which is connected to the top driven motor and enters from the
1–1.3 Pa [12]. However, shear rate of 7000 s−1 exhibits the highest upper side of bioreactor. The cylindrical sintered stainless steel micro-
growth rate of Phaeodactylum tricornutum by maintaining a balance porous sparger has 5.8 × 108 pores of the size 0.2 μm each. The total
between the gas-liquid mass transfer limitations and the shear stress length of the cylindrical sparger is 3.78 cm on which the pores are
damage [13]. Shear sensitivity is dependent on hydrodynamics inside distributed along the length of 2.18 cm and placed centrally beneath the
the reactor and microalgae genotype. This is shown clearly in published lower impeller. Continuous sterile air was supplied to the reactor at a
literature where damage to the Protoceratium reticulatum cells is ob- superficial velocity of 0.0016 m/s. The working volume was fixed at
served even at low shear rate of 0.12 s−1 [14]. The same is observed in 2.2 l (i.e. 2/3rd of the bioreactor volume) and filled with water as it has
the case of Phaeodactylum tricornutum and Porphyridium cruentum where negligible viscosity difference with broth at 28+1 °C.
the threshold impeller tip speed is beyond 1.56 m/s and between 2.45 CFD simulations were carried out and verified by experimental
and 2.89 m/s, respectively [15]. A co-axial cylindrical device when work. Different arrangement and configuration of Rushton and Marine
used for the cultivation of Spirulina platensis under appropriate shear impeller were simulated via CFD at different RPM i.e. 50, 75, 100, 125,
stress environment causes cell damage at shear stress beyond 0.3 Pa 150, 175, 200 and 225 whereas 100 RPM was chosen as the control
[16]. The reported data clearly indicates the importance of design RPM further studies to obtain the better arrangement and configuration
parameters in the cultivation of shear sensitive microalgae. A previous of Rushton and marine impeller.
work on a modified conventional bioreactor also describes the same
[17]. Further upgradation of the system is possible by involving im-
peller speed (RPM) and its effect on hydrodynamic shear stress and 2.2. Strain used and cultivation conditions
microalgae growth inside the reactor. This analysis is essential to pre-
dict the optima for high cell density culture of shear sensitive micro- Arthrospira platensis supplied from Algal Research and Supply
algae cells. Therefore, present work was focused on the CFD modeling (Carlsbad, USA) was used for the experimental studies carried out for
and simulation of a stirred tank bioreactor (Bioflo™ 110) to analyze the the verification of simulation results. A. platensis was selected as well-
hydrodynamic shear stress generated at different rotational speeds suited shear sensitive specie due to its high protein content and other
using different impeller combinations i.e. Rushton-Rushton, marine- industrial benefits. It has abridged trichrome chains that contribute to
marine and Rushton-marine. The effects of different rotational speeds its low shear tolerance but high mass transfer and illumination ex-
and corresponding shear stress on the growth of A. platensis were also posure capabilities [16]. It exhibits steady growth rate even at the
experimentally studied. The simulation and experimental studies com- sudden alteration in the shear stress due to regeneration of the tri-
bined provided the desirable rotational speed of combination of im- chromes [20]. It reveals higher growth rate on gentle aeration (0.2–1.2
pellers to be operated in the bioreactor for the generation of tolerable vvm) and doesn’t show a very significant effect on further increments in
shear stress to ensure efficient growth of microalgae. aeration [21]. Increase in shear stress (up to 0.6 Pa) increases photo-
synthetic activity of cells even at low illumination intensity. Turbulent
2. Material and methods motion makes the cells move about to their helix that permits illumi-
nation to scope pigments generation which generally shaded by helical
2.1. Configuration of the conventional bioreactor structure [16].
In present study, the culture was maintained and preserved using
The conventional bioreactor (Bioflo™ 110, M/S Eppendorf, USA) the methodology described elsewhere [17]. The experimental studies
was initially equipped with two Rushton impellers (top) along with one were performed by cultivating A. platensis at various rotational speeds
ring sparger (bottom) housed inside a 1 cm thick vessel of height of RPM 50, 75, 100, 125, 150, 175, 200 and 225 using two combina-
23.5 cm and internal diameter of 13.32 cm. The two impellers are tions of impellers to analyse the effects of respective shear stress on the
placed on the shaft at 13.3 cm and 6.5 cm distance from the base. The growth of microalgae. The entire methodology was same in both the
sparger was placed beneath the impeller at a distance of 3.25 cm from fermentation runs. A. platensis was first cultivated in the conventional
the base of the reactor. After the calculation of optimum spacing be- bioreactor BIOFLO 110 having two Rushton type impellers and then in
tween the impellers, marine impeller replaced the bottom Rushton one with the combination of Rushton-marine impellers. Cultivations
impeller and sintered stainless steel microporous sparger replaced the were carried out to compare the performance of different impeller
ring sparger [17]. Baffles were placed at the centre of the working combinations in terms of the biomass build up. Bioreactor with mod-
volume with the dimension of 12 cm × 1 cm × 0.2 cm (Height x Width ified Zarrouk’s medium [17] was autoclaved and inoculated with 10%
x Thickness). (v/v) inoculum. The illumination intensity of 27, 284 lx was measured
Fig. 1 shows the CAD of the modified bioreactor including the above at the reactor surface using HTC instrument, LX-103 Light meter Lux-
components. These three-dimensional drawings were constructed with meter. Temperature, pH and photoperiod were maintained as 30 °C, 9
the help of Autodesk SimStudio Tools 2016. The software uses Study and 12:12 (Light: Dark) throughout the batch fermentation, respec-
Design Environment as the UI engine with Auto desk CFD as a solver tively. Continuous sterile air was supplied at the rate of 1.36 LPM
component. Autodesk CFD utilizes standard κ-ε model which is the most (0.62vvm) through micro porous sparger. CO2 from air was used as
widely used isotropic two-equation turbulence model for steady in- carbon substrate for microalgae growth under phototrophic conditions.
compressible turbulent flow in other commercial software programs. It
is the most appropriate and robust model for the small stirred tank

2
R. Verma, et al. Biochemical Engineering Journal 151 (2019) 107312

Fig. 1. (a) Front view of the fermenter incorporated with Rushton and marine impeller; and sintered micro-porous sparger (b) top view of the fermenter (c) bottom
view of the fermenter.

2.3. Assay methods


Residual nitrate=Absorbance@220nm – 2(Absorbance@275nm) (1)
2.3.1. Growth analysis All sets of experiments were performed in triplicates. And average
Samples were withdrawn after every 24 h for growth analysis. The values of biomass concentration are given in figures.
growth was analysed as absorbance at 680 nm and the dry cell con-
centration was measured using gravimetric method. 5 ml sample was 3. Theory and calculation
poured into an already dried and weighed foil cup. It was then kept at
70 °C for 12 h. Dried cup with biomass was weighed to obtain dry 3.1. CFD model governing equations
biomass concentration.
The specific material properties to specific parts of the bioreactor
2.3.2. Nitrate estimation such as: glass to the bioreactor vessel; steel to all interiors such as pi-
The residual nitrate concentration was measured using Ultraviolet pelines, baffles, impellers, sparger, shaft; water to rest of the working
Spectrophotometric method [22]. 2 ml of collected sample was cen- volume space were applied in this simulation. The pressure at the
trifuged for 20 min at 8000 RPM. Supernatant was diluted to 1000-folds surface of water in the reactor was kept at 1.02 atm. The temperature of
to bring the concentration in the range of standard curve. 1 ml of 1 N the vessel was kept constant at 28ºC. At this temperature, the surface
HCl was added to 5 ml of diluted supernatant to restrict the interference tension and viscosity of water air system was kept at 7.15 × 10 2 N/m
of hydroxides and carbonates. Absorbance was measured at 220 nm and and 0.8324 mPa.s, respectively. All other physical properties of water
275 nm for the analysis of nitrate and organic content respectively. The were selected automatically by the software for the given temperature.
actual nitrate content was calculated using the Eq. 1. Standard curve Water played the part of the primary phase in this study whereas air
was plotted using the sodium nitrate as the nitrate content. played its role as the secondary phase and no other phase was present in

3
R. Verma, et al. Biochemical Engineering Journal 151 (2019) 107312

Fig. 2. Single layered meshed diagram of the fermenter (Layer factor = 0.45 and gradation = 1.1).

the domain that could represent biological cells. The viscosity was ( U)
+ (µeff U ) = . + . (µeff U )T + B
measured for all the sampled broth using LVDV-II + PRO Viscometer t (7)
(Brookfield Engineering Laboratories, USA). The observed viscosity
exhibited the variation with water as 0.01 mPa.s. The velocity of the The Turbulent Kinetic Energy (TKE) equation (Eq. 8) and the
water content was kept at 0 cm/s initially as there were no inflows and Turbulent Energy Dissipation (TED) equation (Eq. 9) [24]:
outflows. The angular motion was specified at 50, 75, 100, 125, 150,
175, 200 and 225 RPM in both the impellers for distinct simulations. ( k) µt
+ . Uk = . µ+ k + Pk
The presence and action of distinct impellers in the domain generated t k (8)
different flow regimes. The governing equations of the model are given
as follows.
( ) µt
The Continuity Equation [23]: + . U = . µ+ + Pk + (C 1 Pk C2 )
t k (9)
+ . ( U) = 0
t (2) Where constants, C 1 = 1.44, C 2 = 1.92, σκ = 1 and σε = 1.3.

Where ρ is the density, t is the time and U is the fluid particle velocity.
For constant density fluid (incompressible fluid), ρ is independent of 3.2. Meshing
space and time.

.U=0 (3) Enhanced mesh with a single layer of layer factor of 0.45 and layer
gradation of 1.1 is given in Fig. 2. The meshing resulted in 25,984
The Navier-Stokes equation [23]: numbers of elements along with 5458 number of nodes. Three different
2U grid sizes of 10,000, 20,000 and 30,000 were selected initially to per-
DUj P
form simulations and compare the results. The purpose was to find if
j
= µ
Dt xi xi xj xj (4) negligible variation in result exists otherwise further increase the grid
size.
Under the assumption that μ and ρ are uniform and . U = 0 , where D
Dt
is the time derivative, μ is the coefficient of viscosity, P is the pressure
and ᴪ is the gravitational potential.
4. Results
The constant eddy viscosity model equations:

K2 Simulations carried out at three different grid sizes i.e. 10,000,


µt = Cµ 20,000 and 30,000, were considered as scenario-1, scenario-2 and
(5)
scenario-3 respectively. The variation in the results of scenario 1 and
Cp scenario 2 was found to be 5.2%, whereas further increase in the grid
Kt = µt size by 10k reflected variation as 0.93%. The steady state was obtained
t (6)
between the iterations 7000–8000 for all angular velocities for all
Where μt is dynamic eddy viscosity, is density, is turbulence dis- scenarios. All the simulations were stopped at 10,000 iterations. CPU
sipation rate, K is turbulent kinetic energy, t is a turbulent Prandtl for the simulation of three different combinations of Rushton-Rushton,
number (usually taken to be 1.0) and Cμ is an empirical constant. Rushton-marine and marine-marine at 100RPM was 94413, 92,839 and
The Momentum Equation: 91984 s respectively.

4
R. Verma, et al. Biochemical Engineering Journal 151 (2019) 107312

Fig. 3. Cross sectional view from top across height of (a) single Rushton impeller and (b) single marine impeller; and (c) Flow regime using impeller combinations
(c.1) Rushton-Rushton, (c.2) Rushton-marine and (c.3) marine-marine at 100RPM.

Table 1 4.1. Comparison between agitation systems in the bioreactor


Shear stress obtained from CFD simulations and biomass build-up (Provided
with S.D values) in A. platensis cultivation carried out at different RPMs using Cross sectional view of velocity profile and flow regime in the case
two agitation systems i.e. Rushton-Rushton and Rushton-marine impeller. of single Rushton impeller is shown in Fig. 3(a) where the blue color
Rotational Rushton-Rushton Rushton-marine depicts the non-shear zones, green and yellow color show the enhan-
speed, RPM cing shear zone areas and red colored zones are not suitable for shear
Simulated Biomass Simulated Biomass sensitive cells. The maximum shear for a rotational speed is generated
shear stress, Concentration, shear stress, Concentration,
near the blades of the impeller when in motion. Fig. 3(b) shows the flow
Pa g/l Pa g/l
regime generated due to single marine impeller in the reactor. In case of
50 1.80 1.1 ± 0.02 0.99 2.2 ± 0.06 marine impeller, the maximum shear for a rotational speed is generated
75 3.03 1.3 ± 0.05 1.97 2.9 ± 0.04 in the direction away from the tip of its blades.
100 3.82 2.4 ± 0.07 2.64 3.9 ± 0.03
Fig. 3(c) shows the comparison of three flow regimes generated
125 6.01 1.72 ± 0.03 3.32 3.87 ± 0.05
150 7.5 1.09 ± 0.01 3.96 3.81 ± 0.02 using the three combinations of impellers i.e. Rushton-Rushton,
175 10.17 0.93 ± 0.008 4.64 3.2 ± 0.07 Rushton-marine and marine-marine at 100 RPM. Stagnant zones of
200 13.74 0.89 ± 0.04 5.60 2.8 ± 0.01 varying sizes are visible near the walls and at the bottom for each im-
225 16.65 0.8 ± 0.06 6.17 2.6 ± 0.05 peller combination. The size of stagnant zones depends upon the im-
peller geometry which in turn directly impacts the position and

5
R. Verma, et al. Biochemical Engineering Journal 151 (2019) 107312

Fig. 4. Vertical cross-sectional view of the modified conventional bioreactor equipped with Rushton-marine impellers at (a) 225 RPM, (b) 200 RPM, (c) 175 RPM, (d)
150 RPM, (e) 125 RPM, (f) 100 RPM, (g) 75 RPM, (h) 50 RPM showing velocity magnitude of 0–66.66 cm/s and temperature varying from 27.72 to 28.28 °C.

magnitude of the shear stress. Among all impeller geometric para- time but it provided lower volume coverage for mixing and the larger
meters, the impeller diameter and the blade angle highly influence the stagnant zones as compared to other two combinations. Therefore, this
shear stress magnitude. The mixing time from the simulation results combination was not considered for further studies. Rushton-Rushton
were calculated at the time step number when the turbulent kinetic and Rushton-marine generated the shear stress of 3.82 and 2.64 Pa,
energy and turbulent energy dissipation saturated and achieved steady respectively. Attributing to better mixing by producing least stagnant
state in the convergence plot. The mixing time resulted from the si- zones and acceptable shear stress in the reactor, combination of
mulations for Rushton-Rushton, Rushton-marine and marine-marine Rushton-Rushton and Rushton-marine were used for further studies.
were 11.2, 9.9 and 5.2 s, respectively. Although marine-marine com- The hydrodynamic parameters generated from these three simulations
bination exhibited acceptable shear stress value of 2.67 Pa and mixing (Fig. 3) exhibited the potential of Rushton-Rushton and Rushton-marine

6
R. Verma, et al. Biochemical Engineering Journal 151 (2019) 107312

Fig. 5. Simulation and experimental results obtained at different RPMs varying from 50 to 225 for (a) Rushton-Rushton, (b) Rushton-marine and (c) the overall
analysis at 100 RPM.

for better growth of microalgae. results from CFD simulations and to compare the two selected combi-
nations of impellers i.e. Rushton-Rushton and Rushton-marine. Eight
batch fermentations were carried out at different RPMs for each im-
4.2. CFD and experimental analysis of shear stress in the bioreactor
peller combination. The effect of shear stress was studied in the terms of
growth maxima of A. platensis and the results are appended in Table 1
The combinations of Rushton-Rushton and Rushton-marine im-
which clearly shows that combination of Rushton-marine impeller is
pellers were further used for CFD simulations at all range of rotational
more efficient than two Rushton impellers in terms of biomass con-
speeds and for experimental studies. The results given in Table 1 were
centration at 100 RPM. Low biomass concentration of 2.2 g/l was
obtained on the completion of 10,000 iterations and provided a com-
achieved in the case of Rushton-marine impeller combination at a 50
parison of averaged hydrodynamic shear stress in the complete fluid
RPM because of settling of cells at the bottom of the reactor. This value
domain generated at different RPMs using both the impeller combina-
was further increased in batches carried out at 75 and 100 RPM. The
tions. The increase in rotational speed (RPM) resulted in the increase in
maximum biomass concentration of 3.9 g/l was achieved at 100 RPM
velocity magnitude and the energy incorporated.
consuming 95% of the nitrates present in the modified Zarrouk’s
Shear stress at different RPM generated in the reactor equipped with
medium.
Rushton-Rushton is approximately twice as that of Rushton-marine
Fig. 5 is highlighting the overall result analysis of the two impeller
(Table 1). The modified bioreactor assembly equipped with the com-
combinations i.e. Rushton-Rushton and Rushton-marine at 100RPM.
bination of Rushton-marine impellers allowed better extent to homo-
Two Rushton impellers required 26.6% higher gassed power than that
genous mixing by considering the mixing time, volume coverage for
of other combination. The maximum biomass concentration achieved
mixing and stagnant zones. Fig. 4 shows the four-way flow regime
with Rushton-Rushton was 38.46% lower than that achieved with
generated inside the bioreactor equipped with Rushton-marine im-
Rushton-marine combination. In brief, the comparison in Fig. 5 high-
pellers at different RPM and temperature range. Comparing C.1 and C.2
lights that Rushton-marine combination exhibited lower mixing time,
of Fig. 3, the major volume coverage for mixing is observed with
power consumption and shear stress but higher biomass build up than
Rushton-marine combination. The velocity in the range 6.66–19.99 cm/
that of Rushton-Rushton.
s at the base of the reactor clearly depicts the effective functioning of
the marine impeller. This prevents the settling of the cells. Under the
assumption of no slip condition and no pressure gradient this simula- 5. Discussion
tion study is an example of pure shear driven flow. The temperature of
the impeller shaft and the wall of the reactor were at 28 °C as a 5.1. Comparison between the two combinations of impellers in the
boundary condition for the simulations. Viscous dissipation acts as a bioreactor
source of thermal energy that causes temperature gradient as no heat
conduction is happening. The viscous dissipation is directly propor- Rushton impellers are known for providing a radial flow of the li-
tional to the square of the velocity gradient. The temperature gradient quid and hence pushing the medium towards the reactor walls. As the
was extremely small in all batches of microalgae cultivation and so the rotational speed increases, the spread towards the wall increases and
viscous dissipation was negligible in the current study. The viscous hence larger shear zones are formed. Whereas, marine impellers pro-
dissipation won’t be important in scale-up of the same bioreactor as the vide axial motion of the liquid in the reactor and circulate the liquid
distance between the impeller blades and wall is large enough for this towards axis and then raise it along the walls. The enhanced rotational
lab scale reactor. However, viscous dissipation makes relevance if more speed reduces settling of biomass and does not contribute in the for-
viscous fluids and a narrow space between the two boundaries exist. As mation of intolerable shear zones. The periphery of fluid flow by axial
shown on the velocity magnitude scale, blue color depicts the shear less impellers forms bigger radius of mixing zones as compared to that of the
zone and up the scale to green, yellow and red show elevating shear Rushton impeller [17]. Shear stress depends on the impeller geometric
zones. The profiles in Fig. 4 also reveal that most of the shear sensitive parameters as well as on the impeller rotational speed [25]. The results
cells are sustainable up to 150 RPM exhibiting colors blue to yellow and obtained from CFD simulations also exhibited that shear stress values
corresponding shear stress up to 3.96 Pa. However, the shear sensitivity increase with increase in the impeller rotational speed and the number
depends on specific strain type; as few strains reported to sustain at of impeller blades. The combination of Rushton-Rushton impeller
higher RPMs. The red color might be unsuitable for numerous shear having 6 blades configured at right angle generated high shear stress
sensitive microalgae as it is visible at 200 RPM and above. values as compared to marine-marine combination having 3 blades
The experimental studies were also performed for verification of the angled at 45°.

7
R. Verma, et al. Biochemical Engineering Journal 151 (2019) 107312

Table 2
Microalgae growth in stirred tank bioreactor with different impeller type and their hydrodynamics.
Ref. Reactor configuration Microalgae type and Hydrodynamic Growth results Inferences and remarks
their properties conditions

[30] Stirred tank


Agitator- Single
Chlorella vulgaris
Green algae
Tip speed-
0-5.89m/s
Biomass concentration-1.48g/l • Effect of hydrodynamic stress on the
photosynthetic activity of cells
spinning plate
Slightly shear tolerant Shear stress-
• Tip velocity and the shear stress on wall affects
cellular activities and may also become cause of
Unicellular 0-3.66 Pa death at high shear stress
Scenedesmus obliquus
Green algae
Tip speed-
0-5.89m/s
Biomass concentration-0.77g/l • Flow regime- Axial motion within the
periphery of spinning plate
Slightly shear tolerant
4-cell colonies
Shear stress-
0-3.66 Pa
• Order of shear tolerance- C. vulgaris > S.
obliquus > C. reinhardtii > A.flos-aquae
C. reinhardtii
Green algae
Tip speed-
0-5.89m/s
Biomass concentration-0.84 g/l
Speicfic growth rate- 0.012 h−1
• A.flos-aquae compeltely damaged at 0.45Pa
Shear sensitive Shear stress-
Unicellular 0-3.66 Pa
A.flos-aquae Tip speed- Biomass concentration- 0.22 g/l
Cyanobacteria 0-1.26m/s
Shear sensitive Shear stress-
Filamentous 0-0.45 Pa
[20] Stirred tank C. reinhardtii Tip speed-0.79-7.85
m/s
Speicfic growth rate- 0.11 h−1 • Tip
growth
speed > 6.28m/s causes decreased

Agitator- Multiple Green algae • Flow regime- Radial motion


vertical bladed turbine
stirrer
• Radial flow regime exibited better growth as
compared to axial in case of C. reinhardtii
Shear sensitive
Unicellular
• Filamentous specie is more sensitive as
compared to unicellular type
S. platensis Tip speed-0.79-
7.85m/s
Speicfic growth rate- 0.07 h−1 • S.lowplatensis being filamentous specie requires
tip speed
Cyanobacteria
Shear sensitive
Filamentous
[31] Stirred tank P. cruentem Tip speed- 0.68-
2.89m/s
Biomass concentration-1.03 g/l • Threshold
2.89m/s
tip speed is between 2.45 and

Agitator- 2 four bladed Red algae


turbines
Shear sensitive
Unicellular
[32] Stirred tank S. platensis Tip speed- 1.1m/s Biomass concentration- 4.44g/l • Flow regime- radial motion
Agitator- Gas inducing
4-bladed impeller
Cyanobacteria • Low biomass productivity at high tip speed
Shear sensitive
Filamentous
[14] Stirred tank P. reticulatum Tip speed- 0.288m/s Biomass concentration- 2.5×104
cells/ml
• Maximum permissible RPM suggested as 50
to avoid shear stress effects.
Agitator- 3 bladed
impeller
Dinoflagellate • P.requires
reticulatum being highly shear sensitive
low tip speed
Non baffled Shear sensitive
[33] *
Shaken flask P. reticulatum Shear stress- 0.00016
Pa
– • Cell shear sensitivity is higher in dark phase
Agitation- orbital shaker Dinoflagellate
Shear sensitive
[34] *
Couette flow and shaken
flask
I.polyedrum Shear stress-0.004 Pa Biomass concentration-3.25×103
cells/ml
• Turbulent
growth
conditions reduces dinoflgellate

Highly shear sensitive


Red tide dinoflagellate
• Response to turbulence depend on
environemntal and physiological factors
Present work Stirred tank A.platensis
Cyanobacteria
Tip speed- 0.33m/s
Shear stress- 2.64 Pa
Biomass concentration-3.9g/l • Combination of radial alongwith axial
motion exhibited high growth even at low tip
Shear sensitive speeds
Filamentous

* Note: Highly shear sensitive microalgae such as P. reticulatum, L. polyedrum are not capable to sustain stress generated in stirred tank configuration so grown in
shaken flask [33,34].

It was important to have an agitation assembly that provides ef- observed during fermentation. There was no peripheral movement of
fective mixing because the cultivation of microalgae requires the gas to cells in the reactor. Brown dead cells were also observed at the bottom
remain in the bioreactor periphery for longer duration. Higher mixing after 3–5 days which further increased with time.
time of 9.9 s for Rushton-marine and 11.2 s for Rushton-Rushton im-
pellers were obtained which resulted in the movement of gas in the
medium for longer duration. These results from CFD analysis motivated 5.2. CFD and experimental analysis of shear stress in the bioreactor
the authors to study hydrodynamic shear stress effects on A. platensis
with the combination of Rushton-Rushton and Rushton-marine im- The major problems in the cultivation of shear sensitive microalgae
pellers in the bioreactor. In case of two Rushton impellers, the cells in bioreactor were settling of the cells to the bottom, stagnant zone
settled down at the bottom of the reactor as larger stagnant zones were formation and cell damage. As Rushton impeller has been proven as the
main contributor to the high shear stress and settling of cells, one

8
R. Verma, et al. Biochemical Engineering Journal 151 (2019) 107312

Rushton impeller at bottom was replaced with the marine impeller. A combination also resulted lesser power input as compared to that of
combination of Rushton-marine impeller provided lesser shear en- Rushton-Rushton placed in the conventional bioreactor. The angular
vironment in the bioreactor and favorable hydrodynamic conditions (radial and axial) motion of both the impellers fixed at 75–150 RPM
when compared to that with the combination of two Rushton impellers. exhibited shear stress of 1.97–3.96 Pa which promotes numerous mi-
This modified conventional bioreactor would also be suitable for cul- croalgae growth [30–32]. Operating the reactor at 100 RPM not only
tivating C. muelleri, S. costatum and I. galbana at an RPM of 150 and 175 make the hydrodynamic environment more suitable for A. platensis but
which are reported to withstand the shear stress of 1.2–5.4 Pa [11]. A also keeps its production feasible even at 2.64 Pa by building up 3.9 g/l
study on Cyanobacteria shows that a shear rate value of 2.2 s−1 can of cells. Enhanced cell production was due to the combination of
decrease its ability of carbon fixation and nitrate metabolism [26]. On homogeneity and appropriate shear stress generated during cultivation.
the basis of published literatures on tolerable shear rate for various The simulation can be further extended for bioreactors with high aspect
microbial cells, the authors believe that the designed assembly and ratio by providing multiple sets of Rushton-marine combination. The
optimized operational conditions would be appropriate for not only present studies are also useful for growing all type of shear sensitive
microalgae cells but for other shear sensitive cells as well [27–29]. unicellular or multicellular microbes.
Batch fermentations were carried out to analyze the effects of shear
stress generated due to the combination of Rushton-Rushton and Declaration of Competing Interest
Rushton-marine impellers at various speeds on the growth of A. pla-
tensis. The Rushton-Rushton impellers generated higher shear stress and None.
lower biomass concentration at all range of RPM than that generated by
the combination of Rushton-marine impeller (Table 1). Although, al- Acknowledgements
most similar results of biomass were attained in the range of 100–150
RPM with Rushton-marine impeller combination but further increase in We are thankful to University School of Chemical Technology, Guru
impeller speed resulted in a decrease in biomass concentration due to Gobind Singh Indraprastha University, New Delhi, for all the lab fa-
high hydrodynamic shear stress. In the case of Rushton-marine im- cilities and infrastructure provided. We acknowledge the university
pellers, dead cells were also observed in the samples collected after 3–4 support for the same.
days at 200 RPM and above due to vigorous mixing. These experimental
studies clearly supported the results obtained from CFD simulations. References
The impeller speed range of 100–150 RPM causing shear stress as
2.64–3.96 Pa exhibited nearly comparable maximum biomass con- [1] B. Capelli, G.R. Cysewski, Potential health benefits of spirulina microalgae,
centration but 100 RPM exhibiting the highest biomass among all re- Nutrafoods 9 (2010) 19–26, https://doi.org/10.1007/bf03223332.
[2] N. Abdel-Raouf, A.A. Al-Homaidan, I.B.M. Ibraheem, Microalgae and istewater
sulted as the most appropriate impeller speed for the growth of A. treatment, Saudi J. Biol. Sci. 19 (2012) 257–275, https://doi.org/10.1016/j.sjbs.
platensis. 2012.04.005.
[3] R. Verma, R. Kumar, L. Mehan, A. Srivastava, Carbon dioxide sequestration/utili-
zation for microalgal growth in photobioreactor, Int. J. Environ. Eng. 3 (2016)
5.3. Validation of simulations 162–165, https://doi.org/10.15224/978-1-63248-084-2-42.
[4] R. Verma, A. Srivastava, Carbon dioxide sequestration and its enhanced utilization
Simulation results in terms of hydrodynamic shear stress and tip by photoautotroph microalgae, Environ. Dev. 27 (2018) 95–106, https://doi.org/
10.1016/j.envdev.2018.07.004.
speed and experimental results in terms of biomass concentration were [5] B. Wang, C.Q. Lan, M. Horsman, Closed photobioreactors for production of mi-
qualitatively and quantitatively compared for stirred tank bioreactor croalgal biomasses, Biotechnol. Adv. 30 (2012) 904–912, https://doi.org/10.1016/
used in the present work with the previously published studies on si- j.biotechadv.2012.01.019.
[6] G. Besagni, G.R. Guedon, F. Inzoli, Computational fluid-dynamic modeling of the
milar bioreactor (Table 2). Studies on stirred tank bioreactors with
mono-dispersed homogenous flow regime in bubble columns, Nuclear Eng. Des. 331
impeller types such as spinning plate, gas induced 4-bladed impeller; 4- (2018) 222–237, https://doi.org/10.1016/j.nucengdes.2018.03.003.
bladed turbines and multiple vertical bladed turbine stirrers reported [7] G. Besagni, F. Inzoli, T. Ziegenhein, Two-phase bubble column: a comprehensive
hydrodynamic shear stress of the order of 0–3.66 Pa. Similarly, present review, Chem. Eng. 2 (13) (2018) 1–79.
[8] W. Freedman, J.F. Davidson, Hold-up and liquid circulation in bubble columns,
work on stirred tank bioreactor with combination of Rushton-marine Tans. Inst. Chem. Eng. 47 (1969) T251–T262.
impellers produced hydrodynamic stress range up to 0.99–2.64 Pa for [9] H. Qiang, A. Richmond, Productivity and photosynthetic efficiency of Spirulina
optimum growth which validates the present study. The tip speed in the platensis as affected by light intensity, algal density and rate of mixing in a flat plate
photobioreactor, J. Appl. Phycol. 8 (1996) 139–145, https://doi.org/10.1007/
published literature ranges from 0 to 7.85 m/s which is a quite high bf02186317.
value as compared to 0.33 m/s used in present work. The flow regime [10] I.S. Suh, C.G. Lee, Photobioreactor engineering: design and performance,
generated due to combined radial and axial flow even at low tip speed Biotechnol. Bioprocess Eng. 8 (2003) 313–321, https://doi.org/10.1007/
bf02949274.
has efficiently improved the mixing and provided a highest biomass of [11] M.H.A. Michels, A.J. van der Goot, M.H. Vermue, R.H. Wijffels, Cultivation of shear
3.9 g/l in 5 days at 100RPM. stress sensitive and tolerant microalgal species in a tubular photobioreactor
equipped with a centrifugal pump, J. Appl. Phycol. 28 (2016) 53–62, https://doi.
org/10.1007/s10811-015-0559-8.
5.4. Conclusion [12] M.H.A. Michels, A.J. van der Goot, N. Norsker, R.H. Wijffels, Effects of shear stress
on the microalgae Chaetoceros muelleri, Bioprocess. Biosyst. Eng. 33 (2010)
In conclusion, the impacts of combination of impellers on the hy- 921–927, https://doi.org/10.1007/s00449-010-0415-9.
[13] A. Contreras, F. García, E. Molina, J.C. Merchuk, Interaction between CO2-mass
drodynamic shear stress in the bioreactor were quantitatively assessed
transfer, light availability, and hydrodynamic stress in the growth of Phaeodactylum
using the CFD simulation and its validation was performed through the tricornutum in a concentric tube airlift photobioreactor, Biotechnol. Bioeng. 60
experiments carried out at different RPMs. The combined radial and (1998) 317–325, https://doi.org/10.1002/(sici)1097-0290(19981105)
axial motion produced unique flow regime for generating lesser shear 60:3<317::aid-bit7>3.0.co;2-k.
[14] F.G. Camacho, J.J.G. Rodríguez, A.S. Miron, E.H. Belarbi, Y. Chisti, E.M. Grima,
stress and complete mixing. The selected combination of impellers in Photobioreactor scale-up for a shear-sensitive dinoflagellate microalga, Process
CFD analysis has successfully provided the clear view of stagnant zones Biochem. 46 (2011) 936–944, https://doi.org/10.1016/j.procbio.2011.01.005.
and flow regimes in the bioreactor fluid. The bioreactor was modified [15] T.M. Sobczuk, F.G. Camacho, E.M. Grima, Y. Chisti, Effects of agitation on the
microalgae Phaeodactylum tricornutum and Porphyridium cruentum, Bioprocess
and the optimum rotational speed of 100RPM was achieved for the Biosyst. Eng. 28 (2005) 243–250, https://doi.org/10.1007/s00449-005-0030-3.
cultivation of shear sensitive microalgae. It was not possible to main- [16] S. Mitsuhashi, K. Hosaka, E. Tomonaga, H. Muramatsu, K. Tanishita, Effects of shear
tain good amount of mass transfer without any mechanical agitation. flow on photosynthesis in a dilute suspension of microalgae, Appl. Microbiol.
Biotechnol. 42 (1995) 744–749, https://doi.org/10.1007/bf00171956.
The combination of Rushton-marine impeller in Bioflo™ 110 promises [17] R. Verma, R. Kumar, L. Mehan, A. Srivastava, Modified conventional bioreactor for
homogeneity throughout the working volume of the bioreactor. This

9
R. Verma, et al. Biochemical Engineering Journal 151 (2019) 107312

microalgae cultivation, J. Biosci. Bioengg. 125 (2018) 224–230, https://doi.org/10. org/10.4319/lo.2002.47.1.0108.


1016/j.jbiosc.2017.09.003. [27] L. Pol, J. Tramper, Shear sensitivity of animal cells from a culture medium per-
[18] Handbook of industrial mixing, in: E. Paul, V. Atiemo- Obeng, S. Kresta (Eds.), spective, Feature 16 (1998) 323–328, https://doi.org/10.1016/S0167-7799(98)
ANSYS Fluent Release 13. Manual. Theory Guide, Canonsburg, Penn.: SAS IP, 2010, 01209-8.
Wiley, New Jersey, 2004. [28] P. Jonczyk, M. Takenberg, S. Hartwig, S. Beutel, R.G. Berger, T. Scheper,
[19] N. Jia, M. Gourma, C.P. Thompson, Non- Newtonian multiphase flows: on drag Cultivation of shear sensitive microorganisms in disposable bag reactor systems, J.
reduction, pressure drop and liquid wall friction factor, Chem. Eng. Sci. 66 (2011) Biotechnol. 167 (2013) 370–376, https://doi.org/10.1016/j.jbiotec.2013.07.018.
4742, https://doi.org/10.1016/j.ces.2011.06.067. [29] J. Chalmers, Shear sensitivity of insect cells, Cytotechnol. 20 (1996) 163–171,
[20] R. Bronnnenmeier, H. Markl, Hydrodynamic stress capacity of micro-organisms, https://doi.org/10.1007/BF00350397.
Biotechnol. Bioeng. 24 (1982) 553–578, https://doi.org/10.1002/bit.260240304. [30] M. Leupold, S. Hindersin, G. Gust, M. Kerner, D. Hanelt, Influence of mixing and
[21] S.R. Ronda, C.S. Bokka, C. Ketineni, B. Rijal, P.R. Allu, Aeration effect on Spirulina shear stress on Chlorella vulgaris, Scenedesmus obliquus, and Chlamydomonas re-
platensis growth and ϒ-linolenic acid production, Braz. J. Microbiol. 43 (2012) inhardtii, J. Appl. Phycol. 25 (2013) 485–495, https://doi.org/10.1007/s10811-
12–20, https://doi.org/10.1590/S1517-838220112000100002. 012-9882-5.
[22] L. Mehan, R. Verma, R. Kumar, A. Srivastava, Illumination wavelengths effect on [31] T.M. Sobczuk, F.G. Camacho, E.M. Grima, Y. Chisti, Effects of agitation on the
Arthrospira platensis production and its process applications in River Yamuna water microalgae Phaeodactylum tricornutum and Porphyridium cruentum, Bioprocess
treatment, J. Water Process. Eng. 23 (2018) 91–96, https://doi.org/10.1016/j. Biosyst. Eng. 28 (2006) 243–250, https://doi.org/10.1007/s00449-005-0030-3.
jwpe.2018.03.010. [32] H.H. Kshirsagar, M.S. Revankar, M.Y. Kamat, S.S. Lele, Cultivation of Spirulina in
[23] S. Pope, Turbulent Flows, Cambridge University Press, Cambridge, 2000 Chapter 2 gas induced photobioreactor and isolation of phycobiliproteins, Ind. J. Biotechnol. 1
and 10. (2002) 255–262.
[24] B.E. Launder, B.I. Sharma, Application of the energy-dissipation model of turbu- [33] F.G. Camacho, J.J.G. Rodriguez, A.S. Miron, M.C.C. Garcia, E.H. Belarbi,
lence to the calculation of flow near a spinning disc, Lett. Heat Mass Transf. 1 E.M. Grima, Determination of shear stress thresholds in toxic dinoflagellates cul-
(1974) 131–138, https://doi.org/10.1016/0094-4548(74)90150-7. tured in shaken flasks implications in bioprocessing engineering, Process. Biochem.
[25] J.Y. Xia, S.J. Wang, S.L. Zhang, J.J. Zhong, Computational investigation of fluid 42 (2007) 1506–1515.
dynamics in a recently developed centrifugal impeller bioreactor, Biochem. Eng. J. [34] A.R. Juhl, V. Velazquez, M.I. Latz, Effect of growth conditions on flow-induced
38 (2008) 406–413, https://doi.org/10.1016/j.bej.2007.08.006. inhibition of population growth of a red-tide dinoflagellate, Limnol. Oceanogr. 45
[26] P.H. Moisander, J.L. Hench, K. Kononen, H.W. Paerl, Small scale shear effects on (2000) 905–915.
heterocystous cyanobacteria, Limnol. Oceanogr. 47 (2002) 108–119, https://doi.

10

You might also like