Chemical Engineering Journal

Expressing variable mass transfer coefficients for gas fermentation in trickle bed
reactor
--Manuscript Draft--

Manuscript Number: CEJ-D-23-14259

Article Type: Research Paper
Section/Category: Chemical Reaction Engineering
Keywords: volumetric mass transfer coefficient; trickle bed reactor; dimensionless equation; syngas
biomethanation; mass transfer dynamic model
Abstract: Gasification of lignocellulosic biomass and syngas fermentation are promising technologies for
producing biofuels like biomethane. Trickle bed reactors (TBR) are efficient bioreactors for
syngas fermentation allowing for higher concentration of microbial cells and very efficient for
mass transfer due to high surface area. Optimized scaling up and dimensioning of the reactor is
crucial when it comes to industrial applications. The first step towards this goal is to develop a
model able to simulate the mass transfer of gases to the liquid phase as a function of the
operating conditions and reactor geometry; this is exactly the scope of the present study.
Experiments were performed in a lab scale (220 ml) TBR with co-current flow of syngas (45%
H2, 20% CO, 25% CO2 and 10% N2) and water, representing the growth medium, to determine
the volumetric mass transfer coefficient (kLa) of each gas compound and its dependence on the
flowrate of syngas and water and reactor structural characteristics. The kLa achieved were in
the range of 6.18 – 13.6 hr-1 for H2, 2.6 – 5.79 hr-1 for CO and 1.53 – 3.43 hr-1 for CO2. A
dimensionless equation was fitted to the experimental data and a model was developed to
predict kLa of a specific gas at various gas and liquid flowrates, and reactor configurations. A
kinetic model was also developed to simulate the concentrations of syngas in gas phase and
liquid phase by applying the kLa values predicted by the dimensionless correlation. Finally, the
model was validated using literature data where the predicted kLa had a coefficient of variance
of ±25% within a realistic range of liquid and gas flowrates for syngas biomethanation.

Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation
Graphical Abstract
Highlights (for review)

Highlights


Mass transfer coefficient, kLa, of syngas was determined in a trickle bed reactor.

kLa was expressed as a function of reactor geometry, gas, and liquid flowrates.

Mass transfer dynamic model was developed and validated.

Model can be used for liquid Reynolds number less than 15.

Model can be used for gas Reynolds number less than 1.
Manuscript Click here to view linked References

1 Expressing variable mass transfer coefficients for gas fermentation in trickle

2 bed reactor
3 Sambit Dutta, Hariklia N. Gavala, Ioannis V. Skiadas*

4 Department of Chemical and Biochemical Engineering, Technical University of Denmark, Søltofts Plads 228A,
5 2800 Kongens Lyngby

6 *Corresponding author.

7 E-mail address: ivsk@kt.dtu.dk, ioannis_sk@yahoo.co.uk (I. V. Skiadas)

8 Abstract
9 Gasification of lignocellulosic biomass and syngas fermentation are promising technologies
10 for producing biofuels like biomethane. Trickle bed reactors (TBR) are efficient bioreactors
11 for syngas fermentation allowing for higher concentration of microbial cells and very efficient
12 for mass transfer due to high surface area. Optimized scaling up and dimensioning of the reactor
13 is crucial when it comes to industrial applications. The first step towards this goal is to develop
14 a model able to simulate the mass transfer of gases to the liquid phase as a function of the
15 operating conditions and reactor geometry; this is exactly the scope of the present study.
16 Experiments were performed in a lab scale (220 ml) TBR with co-current flow of syngas (45%
17 H2, 20% CO, 25% CO2 and 10% N2) and water, representing the growth medium, to determine
18 the volumetric mass transfer coefficient (kLa) of each gas compound and its dependence on the
19 flowrate of syngas and water and reactor structural characteristics. The kLa achieved were in
20 the range of 6.18 – 13.6 hr-1 for H2, 2.6 – 5.79 hr-1 for CO and 1.53 – 3.43 hr-1 for CO2. A
21 dimensionless equation was fitted to the experimental data and a model was developed to
22 predict kLa of a specific gas at various gas and liquid flowrates, and reactor configurations. A
23 kinetic model was also developed to simulate the concentrations of syngas in gas phase and
24 liquid phase by applying the kLa values predicted by the dimensionless correlation. Finally, the
25 model was validated using literature data where the predicted kLa had a coefficient of variance
26 of ±25% within a realistic range of liquid and gas flowrates for syngas biomethanation.
27 Keywords: volumetric mass transfer coefficient, trickle bed reactor, dimensionless equation,
28 syngas biomethanation, mass transfer dynamic model
29 1.0 Introduction
30 Gasification of lignocellulosic and other residual biomass is an emerging technology that has
31 been gaining attention due to its ability to convert biomass carbon into a wide range of products.
32 One of the promising products is biomethane, which has the potential to replace fossil derived
33 natural gas in the global gas grid [1]. This is particularly important given the current need to
34 reduce greenhouse gas emissions and mitigate the impact of climate change. The process of
35 producing biomethane involves the conversion of syngas, a mixture of H2, CO and CO2
36 produced from biomass gasification, into methane through a process known as syngas
37 fermentation. This process has already reached a Technology Readiness Level (TRL) of 4-5
38 using an efficient trickle bed bioreactor (TBR) based on mixed microbial cultures [2]. The
39 technology has shown promise in efficiently transforming syngas into methane and other
40 valuable liquid products such as alcohols and organic acids [3]. TBR offers several advantages
41 over other reactor types, making them a more efficient option for syngas biomethanation.
42 Primarily, TBR has a high surface area-to-volume ratio, which facilitates efficient mass transfer

1
43 of gaseous substrates from the gas phase to the liquid phase [4,5]. This is because the TBR is
44 packed with small-sized particles, which create a large surface area for contact between the gas
45 and liquid phases. This high surface area promotes effective contact between the microbial
46 cultures responsible for biomethanation and the gaseous substrates, leading to efficient and fast
47 conversion of syngas to biomethane [6–8]. Finally, co-current TBR is more resistant to liquid
48 flooding and has improved liquid distribution [9]. The high surface area of the packing material
49 in the TBR promotes effective mixing of the liquid and gas phases, which enhances the
50 availability of gases to the immobilized methanogenic microbial consortia on the packing
51 material [10] and higher productivity of biomethane.
52 The potential of biomethane as a sustainable and renewable source of energy is enormous, and
53 syngas fermentation offers a promising avenue for its production from lignocellulosic biomass
54 [11]. While the technology has already demonstrated significant progress in lab- and pre-pilot
55 scale experiments, there is still a need to optimize the process and scale up to pilot and
56 commercial scales for continuous operations [12,13]. However, the challenge in optimizing,
57 modeling, and scaling up TBR lies in adequately simulating the mixing of liquid and gas inside
58 the reactor and the mass transfer of gaseous substrates from gas to liquid phase. Mixing and
59 mass transfer is crucial in ensuring efficient conversion of syngas to biomethane, as it facilitates
60 the distribution of gaseous substrates and nutrients to the microbial cultures responsible for the
61 conversion process [14]. Research efforts are underway to address this challenge through
62 advanced modeling techniques and experimentation. For instance, computational fluid
63 dynamics (CFD) simulations have been used to model the mass transfer and mixing processes
64 in TBR, while experiments have been conducted to validate the model predictions [15].
65 Additionally, alternative reactor designs, such as fluidized bed reactors and membrane
66 bioreactors, have been proposed as potential solutions to improve the efficiency of syngas
67 fermentation for biomethane production [16]. Another hindrance towards scaling-up of syngas
68 biomethanation is the economic viability of the process. Studies have revealed that operational
69 cost can be reduced almost 15 times by reducing the recirculation rate of liquid from 1500
70 ml/min to 500 ml/min during syngas fermentation without affecting the conversion efficiency
71 of syngas and productivity of methane [17]. Another approach to reduce the production cost of
72 methane could be intermittent liquid recirculation [18,19]. Studies have shown that periodic
73 liquid recirculation improved the content of methane in outflow gas by 15% and productivity
74 of methane by 5 – 20 % [20,21].
75 From the literature survey it is evident that syngas biomethanation provides a promising
76 alternative for biofuel production and TBR has several operational advantages to support
77 syngas biomethanation on the lab scale. Nevertheless, significant challenges still exist
78 regarding scaling up of TBR for continuous operation. The present work focuses on the
79 determination of the volumetric mass transfer coefficient of gases in a trickle bed reactor. The
80 volumetric mass transfer coefficient from gaseous phase to liquid phase and its dependence on
81 the operating and structural characteristics of the reactor is an important aspect of design and
82 operation of a TBR, as it affects the growth pattern of microbes, which constitute the core of
83 the process [22]. Therefore, it is essential to determine the volumetric mass transfer coefficient
84 of a TBR and simulate its variability with the operating conditions to optimize the reactor
85 efficiency for syngas fermentation and biomethane production. The volumetric mass transfer
86 coefficient is a measure of how efficiently gaseous substrates are transferred from the gas phase
87 to the liquid phase and it is also affected by the gas and liquid flowrates and reactor geometry

2
 88 [23,24]. Therefore, determining the volumetric mass transfer coefficient is crucial in
89 understanding the rate of conversion of syngas to biomethane, as it provides insight into the
90 effectiveness of mass transfer and how to optimize it for efficient biomethane production. Also,
91 by determining the mass transfer coefficient of a TBR, it is possible to optimize the reactor
92 design for efficient conversion of syngas to biomethane. This includes optimum selection of
93 the packing material, liquid and gas flow rates, reactor size and height to diameter ratio, and
94 other parameters that affect mass transfer and microbial growth. This optimization can lead to
95 higher conversion rates, lower production costs, and more sustainable production of
96 biomethane. In addition, mass transfer coefficient determination is an important step for scaling
97 up the TBR from lab-scale to pilot-scale or commercial-scale production. This ensures that the
98 optimum parameters are applied consistently across all scales, leading to efficient and
99 sustainable production of biomethane. In this study, experiments were carried out to determine
100 the volumetric mass transfer coefficient in a trickle bed reactor with co-current gas and liquid
101 flow and further develop a model to predict mass transfer coefficient of a gas in a liquid at
102 different gas and liquid flow rates in a trickle bed reactor.
103 2.0 Materials and Methods
104 A lab scale TBR was used in this study with a bed volume of 220 ml. More information about
105 the set-up can be found in the work of Asimakopoulos et al. [25].
106 2.1 Determination of bed porosity
107 The porosity of lab scale TBRs was determined by the water displacement method. The
108 polypropylene/polyethylene packing material (BioFLO 9-Smoky Mountain Biomedia, USA)
109 was packed inside a measuring cylinder of a volume equal to the volume of the TBR (220 ml
110 for the lab-scale) and water was poured into it. The porosity of the bed was calculated by
111 dividing the volume of water required to fill the packed measuring cylinder by the total volume
112 of the cylinder.
113 2.2 Volumetric mass transfer experiments
114 Volumetric mass transfer experiments were conducted in a lab-scale TBR (Fig. 1) using a
115 syngas mixture (H2: 45%, CO: 20%, CO2: 25% and N2: 10%) and distilled water as liquid phase
116 to determine the transfer rate of gas molecules from gas phase to liquid phase by following the
117 two-film theory [26]. The reactor operating temperature was maintained at 60 ℃ throughout
118 the experiment. Prior to the introduction of syngas and water, the reactor was initially flushed
119 with nitrogen gas for 10 to 15 minutes, with no flow of syngas and water. After the initial
120 flushing, syngas and water were continuously introduced at the top of the reactor at the
121 volumetric flow rates indicated in Table 1. The gas and liquid flowrates used for this
122 experiment was taken from previous work [5] and are the flowrates the system have
123 demonstrated maximum conversion of syngas to methane and maximum methane productivity.
124 For simplicity, there was no liquid recirculation but average recirculation rates by previous
125 study were used as resulting equivalent continuous liquid flowrate (See supplementary material
126 for detailed calculation). Gas and liquid samples were collected from the bottom of the reactor
127 every 60 minutes for a duration of 4 hours. Duplicate experiments were performed for each set
128 of gas and liquid flow rate to determine the volumetric mass transfer coefficient (kLa) of H2,
129 CO, and CO2. The liquid working volume in the TBR was experimentally determined after
130 each experiment by stopping the liquid flow in the top of the reactor and then collecting the

3
131 liquid trickling down in a measuring cylinder placed at the bottom of the reactor. The liquid
132 samples, with a volume of 8 ml, were transferred to anaerobic serum vials with a total volume
133 capacity of 11 ml and subjected to heating at 100℃ for a duration of 30 minutes. The gas
134 components (H2, CO and CO2) were thus transferred out of the liquid to the gas phase. A 100
135 µL gas sample was then extracted using a gas-tight syringe to determine the composition of the
136 liquid phase in gaseous components according to the methodology described in Grimalt-
137 Alemany et al. [27]. Gas samples (100 µl) were also taken from the bottom of the reactor using
138 a gas-tight syringe to determine the composition of gas flowing out of the reactor. These
139 samples were analyzed to determine the concentration of H2, CO, and CO2 in the gas phase.
140 Overall, this experimental procedure aimed to investigate the mass transfer coefficient of H2,
141 CO, and CO2 in a gas-liquid system using a syngas mixture as the gas. The obtained mass
142 transfer coefficient values were used to calculate the rate at which the gas components are
143 transferred from the gas phase to the liquid phase, which can be a critical parameter of syngas
144 biological conversions in TBR. As mass transfer coefficient depends on several factors, a
145 dimensionless correlation (Eq. 4) consisting of terms representing physical properties of both
146 liquid and gas along with physical properties of TBR i.e., porosity of bed, size of packing and
147 dimension of reactor was used to express the dependence of the kLa on these properties [24].
148 The dimensionless equation was fitted to experimental data to estimate the parameters or
149 exponents of the dimensionless terms using the function fitnlm in MATLAB. A kinetic model
150 of transfer of gas components from gas phase to liquid phase was executed in Aquasim 2.1g,
151 as a gas compartment and a mixed liquid compartment with a constant volume linked by a
152 diffusive link [28]. The dynamics of acid dissociation i.e., acid-base equilibrium equation was
153 included in the kinetic model as described in Eq. 5. It was assumed that the rate of acid
154 dissociation was very fast [29].
155155

4
 Figure 1: Illustration of volumetric mass transfer in lab scale trickle bed reactor for co-current flow of gas and
liquid from the top of reactor

156 Table 1: Experimental conditions for mass transfer experiments

Experiment no. Gas flow rate Liquid flow rate
(ml/min) (ml/min)
1 6.11 24.5
2 6.11 8.25
3 6.11 3.4
4 3.7 24.5
5 3.7 8.25
6 3.7 3.4
157

158 2.3 Theoretical Calculations

159 The volumetric mass transfer coefficient of a specific gas ((kLa)g) is given by Eq. (1) where
160 Qin,g and Qout,g are the molar inflow and outflow of the gas, respectively, VL is the liquid
161 working volume inside the reactor and Cg* and Caq,g are the saturation concentration of gas in
162 liquid phase and the actual concentration of gas in liquid phase, respectively.

163 (𝑘𝐿𝑎)𝑔 = (𝑄𝑖𝑛,𝑔 − 𝑄𝑜𝑢𝑡,𝑔)⁄𝑉𝐿(𝐶∗ − 𝐶𝑎𝑞,𝑔) (1)

𝑔
164 The saturation concentration of the gas (C *) is calculated by applying Henry’s law (Eq. 2)
g
165 where Cg is the concentration of gas in the gas phase.
166 𝐶∗ = 𝐾 ∗ 𝐶 (2)
𝑔 𝐻 𝑔

167 Temperature correction for henry’s constant (KH) was done using Van’t Hoff equation (Eq. 3)
168 where –ΔHo is the standard enthalpy of solution of a specific gas as found in Alberty [30] and
169 Haynes [31].
𝑜 −∆𝐻𝑜 1 1
170 𝐾𝐻 = 𝐾𝐻 ∗ exp ( ( − )) (3)
𝑅 𝑇 𝑇𝑜

171 A dimensionless correlation was used to express the mass transfer coefficient for a gas
172 compound ((kLa)g) in a gas mixture, and it is given in Eq. 4. It includes dimensionless terms
173 for gas properties (XG and ScL), and liquid properties (XG, ReL, WeL and ScL) and a
174 dimensionless term for reactor configuration (θ). The exponents of the dimensionless terms
175 i.e., a, b, c, d, and e were estimated using MATLAB. In the present work, water was used as
176 the liquid and syngas comprising 45% H2, 20% CO, 25% CO2 and 10% N2 in molar fraction
177 was used as the gas mixture.

178 𝑆ℎ𝐿, = Ω ∗ 𝑋𝐺𝑎 ∗ 𝑅𝑒𝐿𝑏 ∗ 𝑊𝑒𝐿𝑐 ∗ 𝑆𝑐𝐿𝑑 ∗ 𝜃𝑒 (4)

179 For the kinetic model in Aquasim, the acid-base equilibrium equation (Eq. 5) was included to
180 describe the kinetics of acid dissociation, where Ka is the acid dissociation constant, and Ca and
181 Cb are the concentration of compound a and b respectively.
𝐶𝑏∗𝐶𝐻+
182 𝐾𝑎⁄ = (𝐶𝑎 − ) (5)
𝐾𝑎
𝑏

5
183 Temperature correction for Ka was done using Eq. 6, where –ΔHo is the standard enthalpy of
184 reaction as found in Alberty [30].
𝑜 −∆𝐻𝑜 1 1
185 𝐾𝑎 = 𝐾𝑎 ∗ exp ( ( − )) (6)
𝑅 𝑇 𝑇𝑜

186 Table 2: Terms, variables, and values for Eq. 4

Abbreviation Name Value Description/Source

It is the ratio of
(𝑘𝐿𝑎)𝑔𝑑ℎ 2 convective mass
Liquid phase Sherwood
ShL,g transfer to the rate of
number for a specific 𝐷𝑒𝑓𝑓,𝐿,𝑔 diffusive mass
gas
transfer.

Lockhart-Martinelli number for

X
aG 𝑣𝐺 𝜌𝐿 It represents the flow
gas √ of gas in a gas-liquid
𝑣𝐿 𝜌𝐺 fluid mixture.
𝑣𝜌𝑑 It represents the flow
𝐿 𝐿 𝑃
ReL Liquid phase Reynolds number conditions of liquid
𝜇𝐿 i.e., laminar, transient,
or turbulent.
It represents the effect
𝑣𝐿2𝜌𝐿𝑑𝑃 of inertial force and
WeL Liquid phase Weber surface tension on the
number 𝜎𝐿 flowing liquid in a
system.
It represents the

ScL 𝜇𝐿 transport of
Liquid phase Schmidt number momentum and mass
𝜌𝐿𝐷𝑒𝑓𝑓,,𝑔
of liquid in a system.
Dimensionless term for 𝑎𝑣𝑑ℎ
θ reactor properties
1−𝜀
Krischer and Kast
d
hydraulic 3 16𝜀3
h
diameter, m 𝑑𝑃 √
9𝜋(1 − 𝜀)2
Cross sectional area of the
Ω TBR, 2 1.257E-03
m
ε Porosity of the TBR 0.79
dP Diameter of packing material, m 0.01
𝐹
𝐺
vG Superficial velocity of gas, m/s
𝐹Ω
𝐿
vL Superficial velocity of liquid, m/s
Ω
6
 Volumetric flowrate of syngas to
FG the reactor, m3/s
Volumetric flowrate of liquid to
FL
the reactor, m3/s
ρL Density of liquid at 60℃, kg/m3 983.19 [32]
ρG Density of syngas at 60℃, kg/m3 0.6917
Viscosity of liquid at
µL 60℃, kg/m.s 4.658E-04 [33]
Surface tension of liquid at
σ
60℃,
L 2 6.62E-02 [34]
kg/s

av Specific surface area of 800 [25]

packing material, m2/m3
Effective diffusivity of a gas in 𝐷𝑒𝑓𝑓,𝐿,𝑔𝐴

Deff,L,g 1 − 𝑦𝐴
liquid with the gas being part of = 𝑦 [35]
a gas mixture, m2/s 𝐷 𝐵 𝑦𝐶 𝑦𝐶
+𝐷 + 𝐷 +⋯
𝐿,𝑔𝐵 𝐿,𝑔𝐶 𝐿,𝑔𝐶
Diffusivity of H2 in water at
DW,H2 60℃, m 2/s 1.31E-08 [36]

Diffusivity of CO in water at
DW,CO 60 , m 2/s 5.68E-09 [37]
℃
Diffusivity of CO2 in water at
DW,CO2 60 , m 2/s 3.80E-09 [38]
℃
187

188

189 3.0 Analytical Methods

190 For gas mixture samples, the content (mol %) of H2, CO and CO2 were determined using gas
191 chromatography. An SRI 8610C chromatograph was equipped with a Molsieve 13x column
192 followed by a silica gel column. The flow of gas towards any of the two columns was controlled
193 by a rotating valve. Gas samples of 100 µl (SGE 1MR-VLL-GT gas tight syringe by Mirkolab
194 Aarhus A/S) were injected and Helium was used as the carrier gas at a flowrate of 50 ml/min.
195 The column temperature was kept constant for 3 min at 65 ℃ and then it was increased by 10
196 ℃/min up to 95 ℃. Then the rotating valve changed the gas flow towards silica gel column
197 followed by a temperature ramp of 24 ℃/min up to 140 ℃ [25].
198 4.0 Results and Discussions
199 4.1 Bed Porosity of TBR
200 Using the water displacement method, the porosity of the bed was found to be 0.79 in the lab
201 scale TBR.
202 4.2 Mass Transfer Experiments

7
203 Estimating the volumetric mass transfer coefficients (kLa) of gases in a liquid medium and
204 combining it with microbial growth kinetics is important for determining product formation
205 rates. However, it can be challenging to accurately determine kLa coefficients due to potential
206 biological activity that could interfere with the measurement process. To overcome this
207 challenge, abiotic experiments were conducted at a temperature of 60℃ to determine the kLa of
208 all gases. This approach leads to a more accurate estimation of the volumetric mass transfer
209 coefficient [27].
210 The kLa of each gas was determined using Eq. 1; and the data used for the determination and
211 calculated kLa values are presented in Table 3. The results of the abiotic experiments revealed
212 that the kLa coefficients for all gases increased as the liquid flow rate decreased at a specific
213 gas flow rate. This is because a decrease in liquid flow rate increased the residence time of the
214 liquid in the TBR, promoting longer contact between gas and liquid phases, and thus an
215 increase in kLa. The findings may also be seen in Fig. 2, which depicts the relationship between
216 liquid flow rate as expressed by the Reynolds number, also shown in Table 3, and kLa
217 coefficients for all gases. The liquid phase Reynolds number (ReL) was calculated according
218 to the formula mentioned in Table 2 and using the value of the superficial velocity of the liquid
219 (𝑣𝐿) from Table 3. Additionally, the results showed that reducing the gas flow rate resulted in
220 a decrement of the concentration gradient between the gas and liquid phases, thereby
221 decreasing the driving force of mass transfer and lowering the kLa [39]. This underscores the
222 importance of carefully controlling the liquid and gas flow rates during bioprocess engineering
223 to optimize microbial growth and product formation. It is worth noting that the accurate
224 determination of kLa coefficients is crucial in enhancing bioprocess performance. The values
225 of kLa coefficients are used in mathematical models to simulate the behavior of bioprocesses,
226 which enables researchers to understand and optimize the processes. Furthermore, the
227 determination of kLa coefficients allows researchers to predict the impact of changing operating
228 conditions, which can be used to optimize bioprocess performance. Ultimately, accurate
229 determination of kLa coefficients is essential for enhancing bioprocess performance and
230 improving our understanding of bioprocess behavior.

8
231 Table 3: Volumetric mass transfer coefficient of H2, CO and CO2 under different operating conditions

232232

Experiments
1 2 3 4 5 6
Operating Syngas flow rate (ml/min) 6.11 3.7
conditions Water flow rate (ml/min) 24.5 8.25 3.4 24.5 8.25 3.4
Superficial velocity of water (m/s) 0.00032 0.00011 4.5E-05 0.00032 0.00011 4.5E-05
Saturation CH2* (mmol/l) 0.291 0.291 0.291 0.291 0.291 0.291
concentrations CCO* (mmol/l) 0.116 0.116 0.116 0.116 0.116 0.116
CCO2* (mmol/l) 3.46 3.46 3.46 3.48 3.47 3.46
Measured Caq,H2 (mmol/l) 0.027 0.069 0.13 0.018 0.05 0.0996
Caq,CO (mmol/l) 0.005 0.014 0.03 0.0034 0.0098 0.022
Caq,CO2 (mmol/l) 0.077 0.233 0.537 0.05 0.156 0.3726
VL (l) 0.0163 0.0135 0.0117 0.0163 0.0135 0.0117
Calculated Qout,H2 − Qin,H2 (mmol/min) 0.0006723 0.0005854 0.0004269 0.0004584 0.0004456 0.0003474
Qin,CO − Qout,CO (mmol/min) 0.0001164 0.0001135 9.722E-05 7.97E-05 7.741E-05 7.146E-05
Qin,CO2 – Qout,CO2 (mmol/min) 0.0021439 0.0020917 0.0019582 0.0014291 0.0014255 0.0013592
ReL 6.85 2.3 0.95 6.85 2.3 0.95
kLaH2 (hr-1) 9.37 11.72 13.6 6.18 8.21 9.3
kLaCO (hr-1) 3.85 4.94 5.79 2.6 3.23 3.89
kLaCO2 (hr-1) 2.33 2.88 3.43 1.53 1.91 2.25
233

9
 Figure 2: Variation of mass transfer coefficient (kLa) of H2, CO and CO2 with Reynolds number for liquid
(ReL) at gas flowrates FG: 6.11 ml/min and 3.7 ml/min
234

235 Empirical correlations are often utilized to predict the volumetric mass transfer coefficient in
236 various types of reactors. One such reactor is the trickle bed reactor, for which literature
237 provides the correlation of kLa with several dimensionless numbers as explained in section 2.3.
238 In the present study, Eq. 4 was used to fit the experimentally obtained kLa and estimate the
239 parameters a, b, d, and e [23,24,40,41]. Due to the crucial influence of surface tension effects
240 on the rate of mass transfer from gas to liquid, the exponent of WeL, denoted as c, was assumed
241 to be one. This is because the specific surface area of packing material directly influences the
242 surface tension of liquid. In the case of TBR, the liquid flows through the channels between
243 the packing and passage of liquid through the channels is governed by capillary phenomena
244 where the surface tension of liquid plays a substantial role on how strongly the liquid molecules
245 are attracted to each other and to the packing material [42,43]. Thus, in high surface area
246 packing materials, capillary phenomena become prominent as the number of channels for liquid
247 flow decrease, and surface tension becomes fundamental in determining the effective surface
248 area available for mass transfer [44]. The result of parameter estimation from the nonlinear
249 fitting of Eq. 4 to the experimentally determined kLa values is presented in Table 4. The table
250 shows that the standard error (SE) for all parameters is less than 0.05, indicating minimal
251 deviation of estimated parameter from its true value. The large t-statistic (tStat) suggests that
252 all four estimated parameters are statistically significant and far from the null hypothesis value.
253 The null hypothesis can be rejected as the p-value (pValue) of the estimated parameters is less
254 than 0.05, indicating their statistical significance. The model fitting of Eq. 4 resulted in an
255 adjusted R2 value of 0.99 which suggests that the model fits the experimental data accurately,
256 and F-statistic vs zero model value of 9330 with a pValue of 8.32E-24 indicates the model fits
257 the data compared to the null model, with a low pValue, providing strong evidence in favor of
258 the fitted model. Fig. 3 illustrates the variation of experimental values and model-predicted
259 values of kLa with ReL at syngas flow rates of 6.11 ml/min and 3.7 ml/min. It may be seen that
260 the model accurately fits all three gases, with an R2 value greater than 0.99 for each of them.
261 In summary, empirical correlations are widely used to predict the volumetric mass transfer
262 coefficient in different types of reactors, and Eq. 4 is commonly used to estimate kLa in trickle
263 bed reactors. The estimated parameters (a, b, d, and e) from the nonlinear fitting of Eq. 4 were
264 statistically significant and accurately fit the experimental data. The results provide useful
265 insights into the mass transfer phenomena in a trickle bed reactor, which can be beneficial in
266 reactor design and optimization. The equation obtained to predict the kLa at different gas and
267 liquid flowrates and different reactor geometry is given below:

10
 𝐷𝑒𝑓𝑓,𝐿,𝑔 0.7436 −1.444 1 3.2919
268 (𝑘 𝑎) = ∗Ω∗𝑋 ∗ 𝑅𝑒 ∗ 𝑊𝑒 ∗ 𝑆𝑐 ∗ 𝜃2.0977 (7)
𝐿 𝑔 𝑑ℎ2 𝐺 𝐿 𝐿 𝐿

269

270 Table 4: Estimated parameter of from non-linear fitting of Eq. 4

Parameter Estimated value Standard Error tStat p-value
a 0.7436 0.022088 33.665 8.4825e-15
b -1.444 0.023098 -62.535 1.5423e-18
d 3.2919 0.042025 78.332 6.6429e-20
e 2.0977 0.049723 42.187 3.7081e-16
271

Figure 3: Comparison of experimental (Exp) and Model (Mdl) estimated values of kLa vs ReL at syngas
flowrate (FG) of 6.11 and 3.7 ml/min
272 4.3 Mass transfer simulations
273 Following the experimental determination of kLa and developing a dimensionless equation (Eq.
274 7), a kinetic model was developed to simulate the data obtained from the mass transfer
275 experiments in the TBR. The developed model was used to predict the concentration of gas
276 components in the gas and liquid phase of the TBR. There was an excellent quantitative
277 agreement between the model predictions and the measured concentrations of H2, CO and CO2
278 in the gas and liquid phase of TBR. The kinetic model also exhibited high conformity with all
279 six sets of experimental data, as demonstrated in Fig. 4. Based on the model predictions, the
280 steady state was achieved within one hour for a gas flow rate of 6.11 ml/min and 1.5 hours for
281 a gas flow rate of 3.7 ml/min. This implies that higher gas flow rates will result in a faster
282 attainment of steady state in both gas and liquid phases. The agreement between the model
283 predictions and experimental measurements implies that the model can be reliably used to
284 predict concentrations of H2, CO and CO2 in the gas phase and liquid phase of TBR. It also
285 validated the dimensionless model for kLa determination which can be used for estimating kLa
286 at different gas and liquid flowrates in a TBR. This provides a more versatile tool for the
287 prediction of mass transfer rates in TBRs and can be used to optimize reactor design and
288 operating conditions for a wide range of applications. Overall, the validation of the model used
289 in this study is crucial for ensuring that the results obtained are reliable and can be used to
290 improve the performance of TBRs. The use of a dimensionless model for kLa is also significant,
291 as it provides a more flexible approach for predicting mass transfer rates in TBRs, which can
292 be useful for process optimization and design.

11
12
 Figure 4: Comparison of experimental (Exp) and model (Mdl) predicted values of concentration of H 2, CO and CO2
in gas phase (Gas volume) and liquid phase (Liquid Volume) for TBR for 6 set of experimental data
293293

294 The dimensionless model for kLa estimation i.e., Eq. 7 with the estimated parameters (Table 4)
295 was used to predict volumetric mass transfer coefficient of oxygen in a TBR as determined and
296 reported by other researchers in the field. Therefore, the kLa values predicted by the model
297 developed in the present study were compared with the kLa values of oxygen experimentally
298 obtained by Orgill et al. [39]. The experiments Orgill et al. performed covered Reynolds
299 number of gas from 0.01 to 0.56 and Reynolds number of liquid from 1.7 to 24.18. The results
300 are presented in a parity plot shown in Fig. 5, where the x-axis represents the experimentally
301 obtained kLa values [39], and the y-axis represents the model-predicted kLa values. The parity
302 plot shows that the model predicted kLa values are in good agreement with the experimentally
303 obtained data for Reynolds number of liquid (ReL) less than 15 and Reynolds number of gas
304 (ReG) less than 1. The coefficient of variance (CV) for this range is ±25%, which indicates a
305 good fit between the two variables. The data points are clustered around the diagonal line,
306 indicating that the model is accurately predicting the kLa values. However, for higher ReL, the
307 CV was greater than 25%, indicating a poor prediction of kLa with respect to experimentally
308 determined values. This may be due to the limitations of the model for predicting kLa at high
309 gas and liquid flowrates. Thus, the dimensionless equation can be used to predict mass transfer
310 coefficient for ReG less than 1 and ReL less than 15. Overall, the dimensionless equation for
311 kLa estimation showed promising results in predicting kLa values in a TBR, within certain
312 limitations. Future research can focus on improving the model's accuracy for higher FG and FL
313 values, as well as applying the model to other types of reactors and processes.

13
 Figure 5: Comparison of model predicted and experimental values of kLa (hr-
1)
314314

315 5.0 Conclusion

316 The volumetric mass transfer coefficient, kLa, of H2, CO and CO2 in water in trickle bed reactor,
317 TBR with co-current flow of gas and liquid was experimentally determined. The highest mass
318 transfer coefficient of H2, CO and CO2 achieved were 13.6 hr-1, 5.79 hr-1, and 3.43 hr-1
319 respectively at a syngas flowrate of 6.11 ml/min and water flowrate of 3.4 ml/min. At a constant
320 gas flowrate, kLa increases with decrease in liquid flowrate as a decrease in liquid flow rate
321 increases the residence time of the liquid in the TBR, promoting longer contact between gas
322 and liquid phases, and thus an increase in kLa. At constant liquid flowrate, kLa decreases with
323 decrease in gas flowrate as it results in decrement of the concentration gradient between the
324 gas and liquid phases, thereby decreasing the driving force of mass transfer and lowering the
325 kLa. A dimensionless equation was applied to correlate kLa with the liquid and gas flowrates,
326 reactor configuration and physical properties of liquid and gas. This equation had an excellent
327 fit with the experimental data with R2 of 0.99. A kinetic model developed to simulate the data
328 obtained from the mass transfer experiments in the TBR exhibited high conformity with all
329 experimental data for both gas and liquid phase of the TBR. Further, the dimensionless
330 equation for kLa determination was used to predict kLa data reported in literature. All the
331 predicted kLa values were found to lie within ±25% of literature data. The correlation was thus
332 validated and hence can be utilized for liquid Reynolds number, ReL less than 15 and gas
333 Reynolds number, ReG less than one. Thus, the correlation can be used to predict volumetric
334 mass transfer coefficient for trickle bed reactors of different geometry and operating conditions.
335 Declaration of Competing Interest
336 The authors declare that they have no known competing financial interests or personal
337 relationships that could have appeared to influence the work reported in this paper.
338 Acknowledgements
339 The work was financially supported by the Novo Nordisk Foundation within the framework of
340 the Fermentation-based Biomanufacturing Initiative (FBM), grant number:
341 NNF17SA0031362.
14
342 References
343 [1] A. Grimalt‐Alemany, I. V. Skiadas, H.N. Gavala, Syngas biomethanation: state‐of‐the‐art
344 review and perspectives, Biofuels, Bioproducts and Biorefining. 12 (2018).
345 https://doi.org/10.1002/bbb.1826.

346 [2] K. Asimakopoulos, H.N. Gavala, I. V. Skiadas, Reactor systems for syngas fermentation
347 processes: A review, Chemical Engineering Journal. 348 (2018).
348 https://doi.org/10.1016/j.cej.2018.05.003.

349 [3] H.N. Gavala, A. Grimalt-Alemany, K. Asimakopoulos, I. V. Skiadas, Gas Biological Conversions:
350 The Potential of Syngas and Carbon Dioxide as Production Platforms, Waste Biomass
351 Valorization. 12 (2021). https://doi.org/10.1007/s12649-020-01332-7.

352 [4] F. Ebrahimian, N. De Bernardini, P. Tsapekos, L. Treu, X. Zhu, S. Campanaro, K. Karimi, I.

353 Angelidaki, Effect of pressure on biomethanation process and spatial stratification of
354 microbial communities in trickle bed reactors under decreasing gas retention time, Bioresour
355 Technol. 361 (2022). https://doi.org/10.1016/j.biortech.2022.127701.

356 [5] K. Asimakopoulos, M. Łężyk, A. Grimalt-Alemany, A. Melas, Z. Wen, H.N. Gavala, I. V. Skiadas,
357 Temperature effects on syngas biomethanation performed in a trickle bed reactor, Chemical
358 Engineering Journal. 393 (2020). https://doi.org/10.1016/j.cej.2020.124739.

359 [6] B. Dahl Jønson, M. Ujarak Sieborg, M. Tahir Ashraf, L. Yde, J. Shin, S.G. Shin, J. Mi Triolo,
360 Direct inoculation of a biotrickling filter for hydrogenotrophic methanogenesis, Bioresour
361 Technol. 318 (2020). https://doi.org/10.1016/j.biortech.2020.124098.

362 [7] P.G. Kougias, P. Tsapekos, L. Treu, M. Kostoula, S. Campanaro, G. Lyberatos, I. Angelidaki,
363 Biological CO2 fixation in up-flow reactors via exogenous H2 addition, J Biotechnol. 319
364 (2020). https://doi.org/10.1016/j.jbiotec.2020.05.012.

365 [8] T.L. Dupnock, M.A. Deshusses, High-Performance Biogas Upgrading Using a Biotrickling Filter
366 and Hydrogenotrophic Methanogens, Appl Biochem Biotechnol. 183 (2017).
367 https://doi.org/10.1007/s12010-017-2569-2.

368 [9] A. Azarpour, N. Rezaei, S. Zendehboudi, Performance analysis and modeling of catalytic
369 trickle-bed reactors: a comprehensive review, Journal of Industrial and Engineering
370 Chemistry. 103 (2021). https://doi.org/10.1016/j.jiec.2021.04.020.

371 [10] C. Feickert Fenske, F. Kirzeder, D. Strübing, K. Koch, Biogas upgrading in a pilot-scale trickle
372 bed reactor – Long-term biological methanation under real application conditions, Bioresour
373 Technol. 376 (2023). https://doi.org/10.1016/j.biortech.2023.128868.

374 [11] S. Paniagua, R. Lebrero, R. Muñoz, Syngas biomethanation: Current state and future
375 perspectives, Bioresour Technol. 358 (2022).
376 https://doi.org/10.1016/j.biortech.2022.127436.

377 [12] K. Asimakopoulos, M. Kaufmann-Elfang, C. Lundholm-Høffner, N.B.K. Rasmussen, A. Grimalt-

378 Alemany, H.N. Gavala, I. V. Skiadas, Scale up study of a thermophilic trickle bed reactor
379 performing syngas biomethanation, Appl Energy. 290 (2021).
380 https://doi.org/10.1016/j.apenergy.2021.116771.

381 [13] B.D. Jønson, P. Tsapekos, M. Tahir Ashraf, M. Jeppesen, J. Ejbye Schmidt, J.-R. Bastidas-
382 Oyanedel, Pilot-scale study of biomethanation in biological trickle bed reactors converting

15
383 impure CO2 from a Full-scale biogas plant, Bioresour Technol. 365 (2022).
384 https://doi.org/10.1016/j.biortech.2022.128160.

385 [14] F. Garcia-Ochoa, E. Gomez, Bioreactor scale-up and oxygen transfer rate in microbial
386 processes: An overview, Biotechnol Adv. 27 (2009).
387 https://doi.org/10.1016/j.biotechadv.2008.10.006.

388 [15] P.R. Gunjal, V. V. Ranade, R. V. Chaudhari, Liquid Distribution and RTD in Trickle Bed Reactors:
389 Experiments and CFD Simulations, Can J Chem Eng. 81 (2008).
390 https://doi.org/10.1002/cjce.5450810365.

391 [16] B. Gunes, A critical review on biofilm-based reactor systems for enhanced syngas
392 fermentation processes, Renewable and Sustainable Energy Reviews. 143 (2021).
393 https://doi.org/10.1016/j.rser.2021.110950.

394 [17] K. Asimakopoulos, Biomethanation of synthesis gas in trickle bed reactors (Doctoral Thesis),
395 Chapter 6 (189-190), Technical University of Denmark, 2019.

396 [18] H. Porté, P.G. Kougias, N. Alfaro, L. Treu, S. Campanaro, I. Angelidaki, Process performance
397 and microbial community structure in thermophilic trickling biofilter reactors for biogas
398 upgrading, Science of The Total Environment. 655 (2019).
399 https://doi.org/10.1016/j.scitotenv.2018.11.289.

400 [19] M. Burkhardt, I. Jordan, S. Heinrich, J. Behrens, A. Ziesche, G. Busch, Long term and demand-
401 oriented biocatalytic synthesis of highly concentrated methane in a trickle bed reactor, Appl
402 Energy. 240 (2019). https://doi.org/10.1016/j.apenergy.2019.02.076.

403 [20] T. Ullrich, A. Lemmer, Performance enhancement of biological methanation with trickle bed
404 reactors by liquid flow modulation, GCB Bioenergy. 11 (2019).
405 https://doi.org/10.1111/gcbb.12547.

406 [21] T.L. Dupnock, M.A. Deshusses, Detailed investigations of dissolved hydrogen and hydrogen
407 mass transfer in a biotrickling filter for upgrading biogas, Bioresour Technol. 290 (2019).
408 https://doi.org/10.1016/j.biortech.2019.121780.

409 [22] B.J. Azzopardi, R.F. Mudde, S. Lo, H. Morvan, Y. Yan, D. Zhao, Hydrodynamics of Gas‐Liquid
410 Reactors, Wiley, 2011. https://doi.org/10.1002/9781119970712.

411 [23] M.J. Ellman, N. Midoux, G. Wild, A. Laurent, J.C. Charpentier, A new, improved liquid hold-up
412 correlation for trickle-bed reactors, Chem Eng Sci. 45 (1990). https://doi.org/10.1016/0009-
413 2509(90)87046-U.

414 [24] A. Gianetto, V. Specchia, Trickle-bed reactors: state of art and perspectives, Chem Eng Sci. 47
415 (1992). https://doi.org/10.1016/0009-2509(92)85029-B.

416 [25] K. Asimakopoulos, H.N. Gavala, I. V. Skiadas, Biomethanation of Syngas by Enriched Mixed
417 Anaerobic Consortia in Trickle Bed Reactors, Waste Biomass Valorization. 11 (2020).
418 https://doi.org/10.1007/s12649-019-00649-2.

419 [26] W.G. Whitman, The two film theory of gas absorption, Int J Heat Mass Transf. 5 (1962).
420 https://doi.org/10.1016/0017-9310(62)90032-7.

16
421 [27] A. Grimalt-Alemany, K. Asimakopoulos, I. V. Skiadas, H.N. Gavala, Modeling of syngas
422 biomethanation and catabolic route control in mesophilic and thermophilic mixed microbial
423 consortia, Appl Energy. 262 (2020). https://doi.org/10.1016/j.apenergy.2020.114502.

424 [28] P. Reichert, AQUASIM – A TOOL FOR SIMULATION AND DATA ANALYSIS OF AQUATIC
425 SYSTEMS, Water Science and Technology. 30 (1994). https://doi.org/10.2166/wst.1994.0025.

426 [29] D.J. Batstone, J. Keller, I. Angelidaki, S.V. Kalyuzhnyi, S.G. Pavlostathis, A. Rozzi, W.T.M.
427 Sanders, H. Siegrist, V.A. Vavilin, The IWA Anaerobic Digestion Model No 1 (ADM1), Water
428 Science and Technology. 45 (2002) 65–73. https://doi.org/10.2166/wst.2002.0292.

429 [30] R.A. Alberty, Thermodynamics of biochemical reactions, John Wiley & Sons, 2005.

430 [31] W.M. Haynes, CRC handbook of chemistry and physics, CRC press, 2016.

431 [32] G.S. Kell, Reanalysis of the density of liquid water in the range 0–150 °C and 0–1 kbar, J Chem
432 Phys. 62 (1975). https://doi.org/10.1063/1.430986.

433 [33] C.J. James, D.E. Mulcahy, B.J. Steel, Viscometer calibration standards: viscosities of water
434 between 0 and 60 degrees C and of selected aqueous sucrose solutions at 25 degrees C from
435 measurements with a flared capillary viscometer, J Phys D Appl Phys. 17 (1984).
436 https://doi.org/10.1088/0022-3727/17/2/006.

437 [34] N.B. Vargaftik, B.N. Volkov, L.D. Voljak, International Tables of the Surface Tension of Water, J
438 Phys Chem Ref Data. 12 (1983). https://doi.org/10.1063/1.555688.

439 [35] D.F. Fairbanks, C.R. Wilke, Diffusion Coefficients in Multicomponent Gas Mixtures, Ind Eng
440 Chem. 42 (1950). https://doi.org/10.1021/ie50483a022.

441 [36] D.L. Wise, G. Houghton, The diffusion coefficients of ten slightly soluble gases in water at 10–
442 60°C, Chem Eng Sci. 21 (1966). https://doi.org/10.1016/0009-2509(66)85096-0.

443 [37] D.L. Wise, G. Houghton, Diffusion coefficients of neon, krypton, xenon, carbon monoxide and
444 nitric oxide in water at 10–60°C, Chem Eng Sci. 23 (1968). https://doi.org/10.1016/0009-
445 2509(68)89029-3.

446 [38] W. Lu, H. Guo, I.M. Chou, R.C. Burruss, L. Li, Determination of diffusion coefficients of carbon
447 dioxide in water between 268 and 473 K in a high-pressure capillary optical cell with in situ
448 Raman spectroscopic measurements, Geochim Cosmochim Acta. 115 (2013).
449 https://doi.org/10.1016/j.gca.2013.04.010.

450 [39] J.J. Orgill, H.K. Atiyeh, M. Devarapalli, J.R. Phillips, R.S. Lewis, R.L. Huhnke, A comparison of
451 mass transfer coefficients between trickle-bed, hollow fiber membrane and stirred tank
452 reactors, Bioresour Technol. 133 (2013). https://doi.org/10.1016/j.biortech.2013.01.124.

453 [40] I. Iliuta, F. Larachi, B.P.A. Grandjean, G. Wild, Gas–liquid interfacial mass transfer in trickle-
454 bed reactors: state-of-the-art correlations, Chem Eng Sci. 54 (1999).
455 https://doi.org/10.1016/S0009-2509(99)00129-3.

456 [41] C. Wang, D. Song, F.A. Seibert, G.T. Rochelle, Dimensionless Models for Predicting the
457 Effective Area, Liquid-Film, and Gas-Film Mass-Transfer Coefficients of Packing, Ind Eng Chem
458 Res. 55 (2016). https://doi.org/10.1021/acs.iecr.5b04635.

17
459 [42] C.W. Green, J. Farone, J.K. Briley, R.B. Eldridge, R.A. Ketcham, B. Nightingale, Novel
460 Application of X-ray Computed Tomography: Determination of Gas/Liquid Contact Area and
461 Liquid Holdup in Structured Packing, Ind Eng Chem Res. 46 (2007).
462 https://doi.org/10.1021/ie0701827.

463 [43] R.E. Tsai, P. Schultheiss, A. Kettner, J.C. Lewis, A.F. Seibert, R.B. Eldridge, G.T. Rochelle,
464 Influence of Surface Tension on Effective Packing Area, Ind Eng Chem Res. 47 (2008).
465 https://doi.org/10.1021/ie070780l.

466 [44] R.E. Tsai, A.F. Seibert, R.B. Eldridge, G.T. Rochelle, Influence of viscosity and surface tension
467 on the effective mass transfer area of structured packing, Energy Procedia. 1 (2009).
468 https://doi.org/10.1016/j.egypro.2009.01.157.

469

18
Supplementary Material

Click here to access/download

Supplementary Material
Supplimentary File_Final.docx
Declaration of Interest Statement

Declaration of interests

☒The authors declare that they have no known competing financial interests or personal relationships
that could have appeared to influence the work reported in this paper.

☐The authors declare the following financial interests/personal relationships which may be considered
as potential competing interests: