Mostafa Abbasian-arani,Mohammad Sadegh Hatamipour,Amir Rahimi

Chemical Engineering Department,University of Isfahan,Isfahan,Iran

Keywords: Jet bubbling reactor Liquid-side volumetric mass transfer coefficient Gas holdup Specific interfacial area Experimental analysis

ABSTRACT The hydrodynamics and mass transfer characteristics of a lab-scale jet bubbling reactor(JBR)including the gas holdup,volumetric mass transfer coefficient and specific interfacial area were assessed experimentally investigating the influence of temperature,pH and superficial gas velocity.The reactor diameter and height were 11 and 30 cm,respectively.It was equipped with a single sparger,operating at atmospheric pressure,20 and 40°C,and two pH values of 3 and 6.The height of the liquid was 23 cm,while the superficial gas velocity changed within 0.010–0.040 m·s?1 range.Experiments were conducted with pure oxygen as the gas phase and saturated lime solution as the liquid phase.The liquid-side volumetric mass transfer coefficient was determined under unsteady-state oxygen absorption in a saturated lime solution.The gas holdup was calculated based on the liquid height change,while the specific interfacial area was obtained by a physical method based on the bubble size distribution(BSD)in different superficial gas velocities.The results indicated that at the same temperature but different pH,the gas holdup variation was negligible,while the liquid-side volumetric mass transfer coefficient at the pH value of 6 was higher than that at the pH=3.At a constant pH but different temperatures,the gas holdup and the liquid-side volumetric mass transfer coefficients at 40°C were higher than that of the same at 20°C.A reasonable and appropriate estimation of the liquid-side volumetric mass transfer coefficient(kla)in a pilot-scale JBR was provided which can be applied to the design and scale-up of JBRs.

1.Introduction

Different types of gas–liquid contactors exist such as bubble reactors,plate columns,packed bed reactors,mechanically agitated contactors,etc.which are applied for gas–liquid reactions in chemical industries[1].Bubble reactors are one of the most widely used in oxidation,hydrogenation,chlorination,alkylation,halogenation,Fischer–Tropsch reaction,ozonolysis,carbonylation,carboxylation,fermentation,wastewater treatment,hydrometallurgical operations,steel ladle stirring,column flotation and other chemical processes.Bubble reactors have different applications due to their simple design,having no moving parts and no seals as well as high liquid holdup[2–4].

Jet bubbling reactors (JBRs) belong to bubble reactors and are equipped with a gas sparging device to generate a bubbling layer.In this layer,an extensive gas–liquid interface area together with an intensive liquid phase turbulence provide an efficient mass transfer area for transporting chemical species between the two phases [5].JBRs are prominently applied in flue gas desulfurization plants.Flue gas is sparged intensively into the liquid phase and an extensive gas–liquid interfacial area is developed.The JBRs do not require large liquid recycle pumps and spray heads,but they undergo a high pressure-drop[6–8].

Since the 1970s,many experimental and theoretical investigations have been conducted on bubbling reactors developing many empirical correlations and mathematical models while a few fundamental studies have been reported on JBRs.

Meng et al.[9]proposed a pierced cylinder as a new device for distribution of gas in JBRs.They studied the gas holdup as a hydrodynamic characteristic of JBR and the effect of slurry pH on SO2removal efficiency.Moreover,they compared the performance of this new sparger with the common ones.Zhao et al.[10]and Cheng[5]suggested a mathematical model for SO2removal from flue gas in a JBR.The proposed model included two coupled differential equations that were solved simultaneously consisting of mass transfer with chemical reaction concept in the jet bubbling and rising bubble zones.The effects of the slurry temperature and pH,concentration of inlet SO2and presence of Cl?was studied by Zheng et al.[7].Huang et al.[11]experimentally assessed the hydrodynamics and radial distribution of the liquid velocity along the axial direction of a JBR.They suggested a tanks-in-series model for the JBR and compared the model and their experimental results.Moreover,some authors have reported the effect of JBR slurry pH and pressure drop on its performance and limestone utilization for full-scale JBRs[12,13].

Due to the widespread use of JBRs in the wet flue gas desulfurization(FGD)process and insufficient data for design and development of JBRs,in this study,an attempt was made to assess the hydrodynamics and mass transfer characteristics of a JBR.The obtained results,i.e.gas holdup,liquid-side mass transfer coefficient,and specific interfacial area,with additional experimental data reported by the authors,Abbasian Arani et al.[14],and chemical and physical properties of the system under study can result in the derivation of some relationships and dimensionless numbers with the aid of dimensional analysis,which can be used in design and scale-up of JBRs.

2.Materials and Methods

2.1.Experimental setup

The experiments were conducted in a glass lab-scale JBR of 11 cm diameter and 30 cm height.Pure oxygen was consumed as the gas phase and saturated lime solution was used as the liquid phase with a height of 23 cm in all experiments.A schematic diagram of the experimental setup is shown in Fig.1.The JBR was embedded in a glass jacket,and the temperature was controlled by water circulating therein.A gas sparger tube,similar to a full-scale JBR,was mounted at the top of the reactor to sparge the gas into the liquid phase through a horizontal 0.5 mm×5 mm slot.The inlet oxygen flow rate was regulated by a gas rotameter and pure nitrogen was consumed for stripping oxygen from the liquid phase.The nitrogen flow was manually controlled by a ball valve.A finite liquid stream was continuously circulated through a small diaphragm pump and sampled in a sampling container using a dissolved oxygen(DO)probe(YSI 5739).

2.2.Materials

In this study,both the oxygen gas of 99.999%purity(Ardestan Medical&Industrial Gas Complex,Iran)and nitrogen gas of 99.999%purity(Tarkib Gas Pars Complex,Iran)were consumed as the gas phase.The liquid phase was an aqueous saturated lime solution consisting of Ca(OH)2with a purity of about 95%(Sadad Sepahan Co.,Iran).In addition,sulfuric acid with a purity of about 98%(Kimia Tehran Acid,Iran)was applied for the liquid phase pH adjustment.

Fig.1.Schematic diagram of the experimental setup.1-Feed tank,2-Peristaltic pump,3-Gas cylinders,4-Gas regulators,5-Valves,6-Gas rotameter,7-Static mixer,8-Sampling bottle,9-Gas analyzer,10-JBR,11-Gas sparger,12-Temperature sensor,13-Water pump,14-Water bath,15-Temperature controller,16-Diaphragm pump,17-liquid waste tank.

2.3.Experimental procedure

The experiments for measuring liquid-side volumetric mass transfer coefficient and gas holdup were conducted as the following steps:

1- The liquid height in the reactor was adjusted at the determined value of 23 cm.

2- The temperature and slurry pH were adjusted for experiments at 20°C for both the pH values of 3 and 6,and the same was done for the temperature of 40°C.

3- Prior to the oxygenation,the pure nitrogen gas was consumed for stripping the liquid phase from dissolved oxygen to achieve zero concentration of dissolved oxygen.

4- After removing the dissolved oxygen,the oxygen flow rate was regulated to the proper value and then was transported to the JBR.

5- The liquid phase was sampled continuously,and the concentration of the dissolved oxygen was recorded at specified intervals until the liquid phase was saturated with oxygen.

6- At the end of each experiment,the height of the slurry was measured and recorded.

In this study,all experiments were repeated three times,and the average value of each run was reported.

2.4.Gas holdup

The gas holdup is one of the most significant operating parameters in bubble reactors,which affects the gas phase residence time and mass transfer between the two phases.The gas holdup is influenced mainly by superficial gas velocity and gas–liquid system properties.The gas holdup is defined as the proportion of the gas phase volume to the total of the two phases.Gas holdup(εg)is calculated using:

where Vgis the gas phase volume,Vlis the liquid phase volume,H0is the initial liquid height measured before the oxygenation and Hlis the final liquid height determined at the end of oxygenation.

2.5.Mass transfer coefficient

According to the two-film theory,the overall mass transfer coefficient is expressed as[15]:

where Kl,kland kgare the overall,liquid-side and gas-side mass transfer coefficients,respectively,and H is the Henry's constant.

In this study,the oxygen absorption process was assessed with a Henry's constant higher than 1.0 with considerably great kgdue to the purity of the gas phase [16].Therefore,the second term of the right side in Eq.(2)is negligible resulting in:

The mass transfer resistance in the gas phase was unimportant,and the main resistance for mass transfer was located on the liquid side.

In general,the mass transfer flux (N) is the product of the mass transfer coefficient and the driving force concentration difference:

where Cl*is the saturation concentration and Clis the instantaneous concentration of dissolved oxygen in the liquid phase.Both the mass transfer rate and mass transfer flux are from:

where a is the specific interfacial area.Thus,the mass transfer rate can be written as:

The product of mass transfer coefficient with specific interfacial area(kla)is known as the volumetric mass transfer coefficient,which is an essential parameter to design and scale-up bubbling reactors[17].

The kla is a measurable parameter,determined by experiments.Steady-state,unsteady-state,dynamic and sulfite test methods are the four common methods for determining the liquid-side volumetric mass transfer coefficient.Some differences between these methods are evident;however,they all follow the same general principles.In this study,the unsteady-state method was applied to measure the kla.To this end,first,the dissolved oxygen was eliminated from the liquid phase by distributing the pure nitrogen.Thereafter,the oxygen or air was sparged in the liquid phase,and then the variation of dissolved oxygen was measured until the liquid phase reached complete saturation[18].

Assuming a completely mixed liquid phase and a short time constant of DO meter electrode,the volumetric oxygen transfer coefficient kla was obtained by integration of Eq.(7)yielding:

where kl,SO2and kl,O2are the liquid-side mass transfer coefficients of SO2and O2,respectively,while DSO2?slurryand DO2?slurryare the diffusion coefficients of SO2and O2in the liquid phase,respectively.

2.6.Specific interfacial area

The specific interfacial area is of particular importance in the gas–liquid reactors design.Chemical and physical methods can be used for experimental determination of the specific interfacial area in bubbling reactors.In the present research,the specific interfacial area was obtained by a physical method based on the bubble size distribution.The photographic method was applied,and bubble size distribution was measured using the images captured along the height of the reactor.

2.6.1.Photographic method

The photographic method maybe the most commonly applied technique for bubble size measurement.This method consists of a camera with suitable lighting for photography.The images of bubbles are usually captured by a digital camera.Then bubble size measurement is performed from the captured images.The main advantage of this method is that it can measure both the bubble size and the projected shape of the bubble on the view plane.In other methods the shape of bubbles are not always available.Further,in some methods spherical shape assumption is applied for bubble size estimation,especially for very small bubbles.While in the industrial systems the bubbles have large sizes and different shapes.But,the most important disadvantage of this method is the tedious analysis and very time consuming of images which usually processed manually.However,this problem has been solved by using advanced image processing software.

2.6.2.Bubble size measurement

An appropriate front-light was used to illuminate a minor sector(i.e.π/4 rad)of the bubble reactor and remaining of the reactor(i.e.major sector)covered by a thick dark paper.A ruler was used as a millimeter reference scale placed in the bubble reactor.Then,the photos were taken using the Canon EOS 760D camera.A typical image recorded by photographic method is presented in Fig.2.Undoubtedly,some refraction problems may still arise and due to cylindrical surface of the reactor,the effect of the curvature on the images is unavoidable,but these effects have been proved to be negligible[20].

Image analysis has been applied to get bubble size distributions and bubble shape data.The 2D pictures of the bubbles were estimated by ellipsoid shape,then maximum and minimum axes of the ellipsoid were measured.A typical image analysis is shown in Fig.3.The bubbles,which were partially overlapped,were segmented manually;then,an ellipsoidal fit was used to obtain the bubble shape[21].Assuming the bubbles were symmetric around the minimum axes,an equivalent spherical bubble diameter was computed using:

The equivalent bubble diameter was then used to calculate the Sauter mean diameter(d32)[22]:

where niis the number of bubbles with the equivalent diameter of di.Then,the specific interfacial area was calculated using the gas holdup and Sauter mean diameter[22]:

Fig.2.A typical image recorded by photographic method.

Fig.3.A typical image analysis for bubble size measurement.

Although the measurement method of this work is relatively simple,some errors can always occur.In this study,the shape of the bubbles was analyzed and evaluated by hand which comprises some errors that are difficult to assess.However,for assaying some results three persons evaluated the same data,almost the same averaged results with insignificant differences were obtained.Therefore,it was assumed that human error is negligible in this work.

3.Results and Discussion

Fig.4 shows the calculated gas holdup values as a function of superficial gas velocity under different conditions.The superficial gas velocity changed from 0.010 m·s?1to 0.040 m·s?1.

A parabolic shape is observed in Fig.4 indicating an increase in the superficial gas velocity leading to a growth in the height of the liquid phase and an increase in the gas holdup.This increase in gas holdup gets slower since bubble coalescence and bubble velocity are increased.Moreover,the observations indicate that increasing the gas flowrate(which increases superficial gas velocity in thereafter)intensifies the liquid back-mixing and enhances the turbulency of phases as a result of increasing the jet-flow speed in the sparger.

Fig.4.Gas holdup vs.superficial gas velocity under different temperature and pH values.

The gas holdup at different pH and temperature values are compared in Fig.4.As observed in Fig.4 at the same temperature but different pH values,the gas holdup versus superficial gas velocity follows almost the same trend in this system.These results are consistent with that of[22].The patterns of the gas holdup variations are considerably different at the same pH but different temperatures which are observed in Fig.4.As shown in this figure,an increase of 20°C in the liquid phase temperature increases the gas holdup by 23%due to the changes in the gas and liquid physicochemical properties.In other words,an increase in the liquid phase temperature decreases both its viscosity and surface tension;therefore,small and stable bubbles are formed,increasing the probability of bubble coalescence[23,24].There exists a duality on which some authors[25,26]refer to the existence of an indirect relation between temperature and gas holdup,while some refer to a direct relation[27,28].As observed in Fig.4 an increase in temperature increases gas holdup,thus there is a direct relation.

In this study,the obtained gas holdup values are in good agreement with the results of Akita and Yoshida [29],Cho and Wakao [30],Hughmark[31],Lau et al.[32]who reported on gas holdup(εg=0–0.1 in the range of us=0–0.040 m·s?1)for a single-nozzle sparger in bubble columns.

The DO concentration was measured at a determined interval whose typical profile is shown in Fig.5.It is assumed that due to well-mixed condition of the liquid phase,no change occurred in the DO concentration throughout this reactor.This assumption is relatively reliable based on the observed complete back-mixing of the liquid phase and intensive turbulence caused by the jet stream of the gas sparger.As observed in Fig.5,the DO concentration measurement continued until the oxygen concentration of the liquid phase is saturated.It should be note that according to YSI 58 instruction manual[33],the probe response time(τ90)was 10 s.Therefore,since the condition kla ?1/τ90was satisfied,the electrode dynamics was disregarded in analysis of the experiments.

Fig.5.A typical dimensionless dissolved oxygen concentration vs.time for calculation of liquid-side volumetric mass transfer coefficient.

The variation of the volumetric liquid mass transfer coefficient vs.superficial gas velocity is shown in Fig.6 under different conditions.It can be seen that an increase in the superficial gas velocity leads to an increase in the liquid-side volumetric mass transfer coefficient.As mentioned,the liquid-side volumetric mass transfer coefficient (kla) is a product of the mass transfer coefficient (kl) and specific interfacial area(a).According to Eq.(10)klin the bubbling reactor is calculated through Higbie's equation[15]:where klis the mass transfer coefficient,DABis the gas diffusivity and tcis the contact time of liquid around the bubble,defined as the time required for a bubble to travel a length equal to its diameter in the liquid phase[15].The contact time is inversely proportional to the bubble rising velocity,thus larger bubbles have higher rising velocity,shorter contact time and higher mass transfer coefficient.Accordingly,at higher superficial gas velocities the probability of bubble coalescence and formation of larger bubbles is high,where the mass transfer coefficient is higher than its value at lower superficial gas velocities.Moreover,the gas–liquid interfacial area is directly proportional to the gas holdup.In other words,the gas holdup increases by an increase in the superficial gas velocity,making the gas–liquid interfacial area larger at higher superficial gas velocities.Therefore,the liquid-side volumetric mass transfer coefficient increases by an increase in the superficial gas velocity.

In Fig.6,the variation of the liquid-side volumetric mass transfer coefficient vs.superficial gas velocity is compared for different pH values and temperature.It is observed that the liquid-side volumetric mass transfer coefficient in pH of 6 is higher than that in pH of 3.Similar results were reported by[22,34]who confirmed the influence of pH on the liquid-side volumetric mass transfer coefficient.According to[22],variations in liquid-side mass transfer coefficient could be related to bubble surface contamination,while no decisive reasons exist regarding the liquid-side volumetric mass transfer coefficient variation versus pH.Fig.6 reveals that the difference between volumetric liquid mass transfer coefficients versus superficial gas velocity at 40°C is higher than the same at 20°C.

Fig.6 compares the variation of the liquid-side volumetric mass transfer coefficient versus superficial gas velocity under different temperatures at the same pH.It can be observed that at higher temperature the liquid-side volumetric mass transfer coefficient is higher than its value at a lower temperature.The obtained results are consistent with findings by[35,36].According to Higbie theory,at higher temperatures the molecular diffusivity rises,increasing the liquid mass transfer coefficient.On the other hand,the liquid viscosity and surface tension decrease,leading to the formation of small and stable bubbles,increasing the gas–liquid interfacial area.The contents of Fig.6 illustrate that a 20°C increase at both pH values of 3 and 6 increases the liquid-side volumetric mass transfer coefficient up to 39.9%and 45.1%respectively,an almost equal increase.

As shown in Fig.6,the obtained kO2,la is comparable to the values reported by Akita and Yoshida [29]for a single-nozzle sparger and absorption of pure oxygen into water in a bubble column (kO2,la=0.002–0.050s–1for us=0.004–0.117 m·s–1) as well as Cho and Wakao [30]for a single-nozzle sparger and oxygen desorption from oxygen-saturated aqueous solutions of five organic solutes in a bubble column (kO2,la=0.004–0.011 s–1for us=0.009–0.070 m·s–1).Moreover,kO2,la varies ason average and the trend of its increase versus the superficial gas velocity is reasonably comparable to the results of Cho and Wakao [30],Shah et al.[3],and Deckwer et al.[37]who reported exponents of 0.81,0.82 and 0.88,respectively.

Fig.6.Liquid-side volumetric mass transfer coefficient of oxygen vs.superficial gas velocity under different temperature and pH values.

Fig.7.Liquid-side volumetric mass transfer coefficient for sulfur dioxide vs.superficial gas velocity under different temperature and pH values.

The volumetric liquid-side mass transfer coefficient for the physical absorption of SO2into the solution of Ca(OH)2was calculated using Eq.(9)where the diffusion coefficients of O2and SO2were estimated from the diffusivity of O2and SO2in water[38].Fig.7 shows the variation of the liquid-side volumetric mass transfer coefficient for SO2versus superficial gas velocity at different conditions.

Using the images,the size of the bubbles in each gas flowrate was obtained.For each gas flowrate,several photographs were taken,then among the photographs,the best three images were chosen to measure the size of bubbles.These data were applied to make a mean bubble size distribution.The obtained mean bubble size distribution is shown in Fig.8.

As mentioned,the Sauter mean diameter and specific interfacial area were experimentally determined.Fig.9 shows the variation of Sauter mean diameter against the superficial gas velocity.It shows that the Sauter mean diameter decreases with an increase in superficial gas velocity.Since at higher gas velocities the interaction of jet bubbles becomes stronger,the separation and breakup of bubbles are increased,forming smaller bubbles.

Fig.8.Bubble size distribution.

Fig.9.Variation of Sauter mean diameter of bubbles vs.superficial gas velocity.

Fig.10.The variation of the specific interfacial area vs.superficial gas velocity.

The specific interfacial area was calculated using the results of Fig.4 and the photographic method at different superficial gas velocities.As illustrated in Fig.10,the specific interfacial area increases with the gas flowrate.As mentioned,higher gas flowrate results in higher gas holdup.Thus,the superficial gas velocity enhancement increases the specific interfacial area.Moreover,as can be seen in Fig.10,the specific surface area at 40°C is greater than that at 20°C,which is due to the higher gas holdup at 40°C.

4.Conclusions

The gas holdup,liquid-side volumetric mass transfer coefficient and specific interfacial area as the crucial parameters of design and scale-up of bubble reactors were experimentally assessed in a lab-scale JBR.The experiments were conducted at atmospheric pressure,at temperatures of 20°C and 40°C,and pH values of 3 and 6 under various superficial gas velocities.The main findings include:

? Regardless of temperature and pH values,an increase in the superficial gas velocity increased the gas holdup,liquid-side volumetric mass transfer coefficient and the specific interfacial area.

? At the same temperature,changing the pH did not affect the gas holdup,while a change in temperature from 20°C to 40°C,increased the gas holdup by 23.0%and 23.3%at pH values of3 and 6,respectively,where this increase was almost similar in both pH values.

? At the same temperature,the liquid-side volumetric mass transfer coefficient was of a higher value at pH of 6 compared to 3.Differences between the liquid-side volumetric mass transfer coefficients at different pH values were 14.0%and 18.2%at temperatures of 20°C and 40°C,respectively.

? The liquid-side volumetric mass transfer coefficient increased by an increase in the temperature at a similar pH,that is 39.9%and 45.1%at pH values of 3 and 6,respectively.

? The specific interfacial area was enhanced by an increase in the temperature and superficial gas velocity.

Finally,the estimated liquid-side volumetric mass transfer coefficients of SO2in different conditions could be used to design and scale-up of JBRs.

Declaration of Competing Interest

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.

Acknowledgements

We are thankful to Dr.M.Momeni Shahraki and Dr.M.Gholami for their precious assistance during the experimental works.Also,the authors appreciate the vice-chancellor of research and technology of the University of Isfahan for supporting this work under Grant No.911401707.