Tianrong Cao ,Weipeng Zhang *,Jingcai Cheng ,Chao Yang ,*

1 Key Laboratory of Green Process and Engineering,Institute of Process Engineering,Chinese Academy of Sciences,Beijing 100190,China

2 Nanjing Jiuzhang Chemical Technology Co.Ltd.,Nanjing 21180,China

3 University of Chinese Academy of Sciences,Beijing 100049,China

1.Introduction

As an important unit operation,crystallization is widely used in the production of basic chemicals,pharmaceuticals,pigments and ceramic powders[1–4].By now,diversified stirred tanks prevail as crystallizers owing to their good mixing performance and technical maturity[5].However,secondary nucleation and crystal attrition are unavoidable due to the existence of mechanical agitation by impellers[6].As for an airlift-loop reactor(ALR),there are no moving mechanical parts in the reactor,and the shear rate is moderate and relatively uniform throughout the reactor volume.Therefore,it is widely used in fermentation,biofuel production and wastewater bio-treatment,to take the advantage of low extent of damage to microbes[7,8].In view of these excellent properties,ALR has been considered as a new prospective crystallizer[6,9–13].

To date,there are only a few papers on the crystallization in ALRs.Soare et al.[13]studied the secondary nucleation in an airlift crystallizer and found it was flexible in controlling the size of crystalline products by manipulating seed load and air flow rate.Rigopoulos and Jones[10]proposed a model for simulating the semi-batch agglomerative gas–liquid precipitation of CaCO3by CO2absorption into lime in an airlift reactor.By comparison with the crystallization of L-ascorbic acid in a stirred crystallizer,Lakerveld et al.[6]found that an air-lift crystallizer can suppress secondary nucleation at a higher super-saturation.As for the study on liquid–liquid reactive crystallization in an ALR,fewre ports have been found in the literature.Therefore,the objective of this work is to study the process of reactive crystallization in an airlift-loop reactor.

Asan important crystalline material for advanced energy conversion and storage,nickel hydroxide(Ni(OH)2)is extensively used as cathode material in rechargeable alkaline batteries[14–22].There have been many reports exploring the preparation and crystallizing mechanism of Ni(OH)2in stirred tanks.E et al.[22]explored the precipitation of nickel hydroxide in astirred tank by online measure ment and got useful data on preparing Ni(OH)2with small size and small sizespan.Pengand Shen[23]came up with the template growth mechanism of spherical Ni(OH)2in a continuous stirred-tank reactor(CSTR).Borho[24]explored the crystallite aggregation and simulated the process of aggregation in a stirred column.Shen et al.[25]investigated the microstructures of spherical Ni(OH)2particles and found that spherical particles were more stable during the charge/discharge cycles compared with the conventional non-spherical ones.Cai[26]studied the production of spherical Ni(OH)2using an oscillatory baffled crystallizer and indicated that tap density of Ni(OH)2had great Influence on electronic properties.Bigger tap density means bigger capacity per volume of Ni(OH)2particles,which makes the electronic properties better.Watanabe and Kikuoka[27]proposed that the nickel hydroxide of small size showed better charge–discharge behavior due to its faster proton diffusion.On the whole,the experimental data and theoretical analyses on preparation of Ni(OH)2in stirred tanks are abundant and the electro-chemical properties of Ni(OH)2particles with spherical shape,small size and big tap density are preferred.In light of this knowledge basis,the precipitation of Ni(OH)2is chosen as the model reaction to study the reactive crystallization in an ALR.

In this paper,the conditions for preparing sphere-like Ni(OH)2in an ALR are explored first by the orthogonal experiment.Then,the Ni(OH)2particles are analyzed by the scanning electron microscope(SEM),X-ray diffraction(XRD),laser particle analyzer,tap densitometer and optical microscope.The temporal evolution of size and tap density of Ni(OH)2particles is characterized by comparison with those prepared in astir red tank.On this basis,the evolution process of Ni(OH)2particles in the ALR is elaborated.

2.Experimental

2.1.Setup and procedure

The experiment was conducted in two reactors,i.e.,an ALR and a stirred tank.A schematic diagram of the ALR is shown in Fig.1.The reactor is composed of an outer cylinder of diameter D1=200 mm and height of H1=400 mm,a draft tube of diameter D2=120 mm and height of H2=300 mm,and an air distributor.The clearance between the reactor bottom and the draft tube is 50 mm.As for the stirred tank,it is a cylinder with the same diameter of the ALR and an up-pumping 45°six-pitched-blade turbine(PBTU).

The formation of Ni(OH)2proceeds through the following reactions[22]:

Fig.1.Schematic diagram of the ALR under investigation.1—inlet A;2—inlet B;3—over flow;4—thermometer;5—sample tap;6—gas distributor.

With the addition of ammonium hydroxide,most of Ni2+ions are complexed with NH3to form[Ni(NH3)n]2+,which decreases the super-saturation of Ni(OH)2.

The main reactant stream A contains NiSO4,and the reactant stream B contains NH3and NaOH(0.2 mol·L?1).All chemicals are of analytical grade.Two streams are fed separately through two inlets.Distilled water is used throughout the experiment.The experimental procedure is described as below.

Before the start of an experimental run,distilled water is heated to the required temperature in an electric kettle and then added into the reactor,and air begins to bubble in through the distributor.After the temperature is stabilized,the NH3and NaOH(0.2 mol·L?1)solution is added into the reactor to adjust p H value.Then,NiSO4is added into the reactor with the feed rate of 55 ml·h?1and reaction begins to happen.Besides,the NH3and NaOH(0.2 mol·L?1)solution is added continuously throughout the whole process to maintain the p H value.Samples are taken every 4 h for tests including XRD,PSD(particle size distribution),SEM and tap density.

2.2.Experimental design

Preparation of spherical Ni(OH)2particles in an ALR is relevant to many factors and has not been explored so far.For this purpose,an orthogonal experiment is performed firstly following the design of L25(56),which is shown in Table 1,and the experimental factors and levels are listed in Table 2.

Table 1 The orthogonal table of L25(56)

Table 2 The table of factors and levels

Fig.2.SEM images of Ni(OH)2 in T6.

Fig.3.Tap density of Ni(OH)2 particles in two reactors(1—ALR and reaction time of 8 h,2—ALR and 20 h,3—stirred tank and 20 h).

Through the orthogonal experiment,the conditions for preparing spherical Ni(OH)2particles can be obtained.Then,the contrast experiment in a stirred tank under the same conditions is conducted.

3.Results and Discussion

3.1.Conditions for preparing spherical particles in ALR

Ni(OH)2particles with sphere-like morphology have been proved to present better electronic properties,such as higher storage[22].Compared with other samples,the ones prepared in T6(abbreviation of Test 6)have the highest sphericity,as shown in Fig.2.Therefore,the conditions of T6 are selected for studying the process of reactive crystallization.

Fig.5.The PSD in ALR and stirred tank.

3.2.Factors in fluencing tap density

Though the morphology of Ni(OH)2prepared in T6 is spherical,the tap density(0.2233 g·cm?3)is too small,while there ported tap density of Ni(OH)2prepared in conventional stirred tanks reaches up to 1 g·cm?3[16,18].To find out the reason,an experimental run in a stirred tank with the same conditions of T6 has been done.

Fig.4.SEM images of Ni(OH)2 in two reactors.

Fig.6.Micro structures of Ni(OH)2 prepared in two reactors when time is 20 h.

Fig.7.Variation of tap density of Ni(OH)2 with time in an ALR and a stirred tank.

For particles,its tap density can be expressed by[26]

where ρ represents the tap density,ρ0represents the real density,f represents the voidage in crystallites which depends on the phase of Ni(OH)2,g represents the void fraction in particles,and ε represents the void content between particles which depends on the shape,size and size distribution[28].

From Fig.3,it can be seen that the tap density of Ni(OH)2prepared in a stirred tank is much higher than that prepared in an ALR.According to Eq.(3),there are 5 factors in fluencing the tap density,i.e.,the shape,particle size,size distribution,the phase of Ni(OH)2and void content g,which are examined and analyzed one by one.

Fig.8.Variation of XRD patterns of Ni(OH)2 prepared in an ALR and a stirred tank with time.

Fig.9.The size of crystallite varying with time in an ALR and a stirred tank.

Fig.10.The micros tructure of Ni(OH)2 in 4 tests.

Peng[28]showed that the spherical Ni(OH)2had a higher tap density.Fig.4 shows the SEM pictures wherein particles prepared in an ALR are closer to spheres compared with the ones prepared in a stirred tank.That means the tap density should be higher in an ALR.However,the higher tap density in a stirred tank suggests that shape isn't the main reason for the lower tap density in an ALR.

The size and size distribution of Ni(OH)2are shown in Fig.5.The particle size in the stirred tank is smaller than that in an ALR for the same preparation time of 20 h,but larger than that in an ALR for a shorter preparation time of 8 h.No matter which samples from the ALR,the tap density is lower than that prepared in a stirred tank.This indicates that the size isn't the main reason for the lower tap density prepared in an ALR.The PSDs of all samples are similar(close to the normal distribution)except the PSD of samples in an ALR(t=20 h)which have 2 peaks.Research[26]shows that the tap density of particles which are made up with small particles and big particles is bigger than ones which just haveonepeak.Even though,the tap density of particles prepared in an ALR when time is 20 h is still lower than the ones prepared in a stirred tank.Therefore,we can infer that the size distribution is also not the main reason.

Fig.11.The sizes of crystallitesin 4 tests(1—Test 5,2—Test 9,3—in astirred tank,4—T6).

Fig.12.The tap densities of 4 tests(1—Test 5,2—Test 9,3—in a stirred tank,4—T6).

Microstructures of Ni(OH)2particles prepared in an ALR and a stirred tank are shown in Fig.6.It can be seen that the voids(g in Eq.(3))in particles prepared in an ALR are bigger than those prepared in a stirred tank.Thus,the key to the tap density of Ni(OH)2seems to be the growth of crystallites in particles.And the process of crystallization can be described as below.The crystallites aggregate together randomly to form many voidsin particles.Then the crystallites in particles can go on growing up to fill up the voids so that the tap density increases.

To verify the guess proposed above,the tap density of Ni(OH)2in an ALR and a stirred tank against time has been measured firstly.The data in Fig.7 show that the tap density in a stirred tank grows up steadily with time while the one in an ALR keeps constant or even decreases slightly.

If the guess proposed above is right,the sizes of crystallites will vary with time following the same trend with tap density.Then,all of these samples are measured by XRD and the sizes of crystallites are also obtained by Jade(a software used to analyze XRD data)based on the data of XRD.The XRD patterns and the sizes of crystallites have been shown in Figs.8 and 9.From Fig.8,it can be found that all samples prepared in both reactors are β-Ni(OH)2,while peaks of the sample from a stirred tank are sharper.According to Fig.9,the sizes of crystallites vary with time following the same trend with tap density regardless of whether in an ALR or a stirred tank,which indicates that the guess above is right.

To further verify that the claim made above is suitable to the particles in an ALR,two more samples prepared in an ALRby orthogonal experiment,which have similar micros tructure with the ones prepared in a stirred tank,are chosen to be compared with the samples of T6 and the ones prepared in astir red tank.The micro structure of them is shown in Fig.10.

Then,all of theses amples are measured by XRD and the sizes of crystallites are shown in Fig.11.According to Fig.11,the sizes of crystallites from the highest to the lowest are the samples prepared in a stirred tank,Test 5,Test 9 and Test 6.The tap densities of these 4 samples are shown in Fig.12 and the results confirm again our claim that the bigger the sizes of the cryst allites are,the higher the tap density is.In summary,it can beconcluded that the growth of tap density depends on the size of crystallites.

Fig.13.Microstructure of Ni(OH)2 varying with time in an ALR.

As for the big difference of tap density of Ni(OH)2particles in an ALR and a stirred tank,it can be attributed to the different driving forces to fluid movement in two crystallizers.Under the same operation conditions in two crystallizers,volatilization of ammonium hydroxide in an ALR is quicker in contrast with stirred tank owing to the gas flow.The volatilization of ammonia makes the concentration of NH4+decrease and Ni2+increase in the solution,which promotes greatly the nucleation but less the growth of crystallites.Therefore,the size of crystallites in Test 6 doesn't change much along with time,while crystallites growup quickly in a stirred tank,which finally causes the difference of tap density.

3.3.Ripening of Ni(OH)2 particles

The shape and morphology of Ni(OH)2with the variation of time in an ALR and a stirred tank have been shown in Figs.13 to 16.In an ALR,the shape of Ni(OH)2particles varies from the irregular shape to the spherical one,the surface gets smoother and the boundary between small particles disappears gradually.As for the particles in astirred tank,the shape of them doesn't change much,the surface maintains the structure with much inter-space,and the boundary is obvious with the whole process though the small particles grow up gradually.In contrast,the conditions in an ALR are more advantageous for preparing spherical Ni(OH)2particles.

To explain this difference,the mechanism of ripening should beintroduced firstly.The Gibbs–Thomson effect[23,29],i.e.,the equilibrium concentration at the surface of a cluster,is used in a truncated Taylor's series form as

where the index GTstands for Gibbs–Thomson,r is the cluster radius,c∞is the equilibrium solubility at the in finite cluster radius,γ is the surface tension of the cluster,vMis the atomic volume of the material,k is the Boltzmann constant and T is the temperature.

According to Eq.(4),every particle or bulge on surface has its own equilibrium concentration.The smaller the radius is,the bigger the equilibrium concentration is.As the super-saturation decreases with time,the solution concentration is easy to get smaller than the equilibrium concentration.This induces the small particles or bulge in solution to dissolve[23].In other words,small particles will disappear and the surface will get smoother,which is the phenomenon of ripening.

Based on the analysis on ripening above,it can be known that the degree of ripening depends on the value of super-saturation in solution.The super-saturation decreases quicker in an ALR owing to the volatilization of ammonium hydroxide under the effect of gas.Therefore,the solution concentration is easier to get smaller than the equilibrium concentration in an ALR and the ripening of Ni(OH)2particles is quicker.

3.4.The effect of mixing on Ni(OH)2 particles in an ALR

Mixing in there action crystallization plays akey role in the quality of the products.Mixing effect in ALR mainly depends on the gas flow,which Influences liquid circulation velocity much.This section discusses the Influence of gas flow on growth of Ni(OH)2crystallites by a single factor experiment.Fig.17 shows the size of Ni(OH)2crystallite varying with gas flow.With the increasing of gas flow,the crystallite son 3 lattice planes get bigger.

Zhang[30]reports that macro-mixing time decreases with the increasing of super ficial gas velocity in an ALR.One important step of crystallization from an aqueous solution[31]is the bulk diffusion of hydrated ions through the diffusion boundary and adsorption layers:the mechanism is linearly dependant on concentration gradient and inversely on diffusion layer thickness(or crystal size).With the increasing of gas flow,turbulence in the bulk liquid phase is enhanced,the boundary layer becomes thinner,thus mass transfer efficiency gets higher,which decreases the thickness of diffusion boundary and promotes the growth of crystallites.Therefore,the size of crystallite gets bigger with the increasing of gas flow.

Fig.14.Microstructure of Ni(OH)2 varying with time in a stirred tank.

Fig.15.Morphology of Ni(OH)2 varying with time in an ALR.

Fig.16.Morphology of Ni(OH)2 varying with time in a stirred tank.

Fig.17.The size of crystallite varying with gas flow.

3.5.Growth process of Ni(OH)2 particles

There are a few reports describing the growth process of Ni(OH)2.However,there still lack some convincing explanations.Shangguan et al.[20]thought the formation process of spherical Ni(OH)2mainly included three steps:nucleation,aggregation and crystal growth.Peng and Shen[23]proposed the template growth mechanism of spherical Ni(OH)2,namely the growth on the crystallite templates stretching in the radius direction is free and quick,while the growth rate for crystallites in other directions is confined due to lower monomer concentration and tends to dissolve.But,no one shows the connection between the growth of tap density and the growth of Ni(OH)2particles.

Based on the analysis above and the conclusion in Section 3.2,we add two supplements to the growth process of Ni(OH)2as outlined in Fig.18.

The improved understanding of the growth process of Ni(OH)2mainly includes 2 points as shown in Figs.19 and 20.The first one is about the growth of particle size.It can be seen that the sizes of particles are at the level of μm from Fig.5,while the sizes of crystallites are at the scale of tens of nm as shown in Fig.9.This difference between particles and crystallites indicates that the growth of particles doesn't depend on the growth of crystallites.Then,the aggregation of crystallites can be visualized from Fig.6.Based on these phenomena,it can be suggested which indicates the transmission of light color is good.With the tap density becoming bigger,the color of samples becomes darker,which indicates that the transmission of light is reduced.The phenomenon evidences the decrease of voids visually and confirms the deduction in Fig.20.

Fig.18.The formation process of Ni(OH)2 particles.

Fig.19.The growth process of particle size.

Besides,ripening occurs in the whole process and the shape of Ni(OH)2particles will vary from irregular to spherical slowly.

4.Conclusions

The growth process of Ni(OH)2particles in an ALR is studied in this work by analyzing micros tructure,shape,size and tap density of particles.Comparing with the experimental results in a stirred tank,the following conclusions can be drawn from this work.

(1)The growth of tap density mainly depends on the size of crystallites.The bigger the size of crystallites is,the bigger the tap density is.

(2)Mixing in the reaction crystallization plays a key role in the quality of the products.With the increasing of gas flow,the crystallites on 3 lattice planes get bigger.

(3)The ALR has more advantageous ripening conditions for preparing spherical Ni(OH)2particles compared with the stirred tank.

(4)The phenomenological growth process of Ni(OH)2particles has been supplemented with more details.Based on this work,the growth of particle size mainly depends on the aggregation of numerous crystallites,while the growth of tap density mainly relies on the size of individual crystallites.that the growth of particles depends on the aggregation of crystallites,as expressed by Fig.19.

The second one is about the growth of tap density,which is shown in Fig.20.The sheet crystallites aggregate together randomly to form many voids in particles.Then the crystallites in a particle go on growing up to take up the voids between crystallites so that the particle density increases.

This process can also be proved by Fig.21,which shows the microscope image of 2 samples with different tap densities.It can be seen that the samples are white when the tap density is 0.2233 g·cm?3,

Nomenclature

CNiSO4concentration of NiSO4,mol·L?1

D1diameter of outer cylinder in an ALR,mm

D2diameter of draft tube in an ALR,mm

H1height of outer cylinder in an ALR,mm

H2height of draft tube in an ALR,mm

qggas flow rate,L·min?1

T temperature,°C

t reaction time,h

VNH3volume of ammonium hydroxide per liter alkaline liquor,ml

VNaOHvolume of NaOH solution,L

Fig.20.The growth process of tap density.

Fig.21.Microscope images of two samples with different tap densities.

[1]B.Yu.Shekunov,P.York,Crystallization processes in pharmaceutical technology and drug delivery design,J.Cryst.Growth 211(2000)122–136.

[2]E.L.Paul,H.H.Tung,M.Midler,Organic crystallization processes,Powder Technol.150(2)(2005)133–143.

[3]Z.Q.Yu,J.W.Chew,P.S.Chow,R.B.H.Tan,Recent advances in crystallization control,Chem.Eng.Res.Des.85(7)(2007)893–905.

[4]D.A.Berry,K.M.Ng,Synthesis of reactive crystallization processes,AIChE J.43(7)(1997)1737–1750.

[5]D.Logashenko,T.Fischer,S.Motz,E.D.Gilles,G.Wittum,Simulation of crystal growth and attrition in a stirred tank,Comput.Vis.Sci.9(3)(2006)175–183.

[6]R.Lakerveld,J.J.H.Krochten,H.J.M.Kramer,An air-lift crystallizer can suppress secondary nucleation at a higher supersaturation compared to a stirred crystallizer,Cryst.Growth Des.14(7)(2014)3264–3275.

[7]Q.S.Huang,C.Yang,G.Z.Yu,Z.-S.Mao,3-D simulations of an internal airlift loop reactor using a steady two- fluid model,Chem.Eng.Technol.30(7)(2007)870–879.

[8]M.?im?ík,A.Mota,M.C.Ruzicka,A.Vicente,J.Teixeira,CFD simulation and experimental measurement of gas holdup and liquid interstitial velocity in internal loop airlift reactor,Chem.Eng.Sci.66(14)(2011)3268–3279.

[9]J.Bao,K.Koumatsu,K.Furumoto,M.Yoshimoto,K.Fukunaga,K.Nakao,Optimal operation of an integrated bioreaction–crystallization process for continuous production of calcium gluconate using external loop airlift columns,Chem.Eng.Sci.56(2001)6165–6170.

[10]S.Rigopoulos,A.Jones,Modeling of semibatch agglomerative gas–liquid precipitation of CaCO3in a bubble column reactor,Ind.Eng.Chem.Res.42(2003)6567–6575.

[11]D.H.Jenkins,Salting-out crystallisation using NH3in a laboratory-scale gas lift reactor,Can.J.Chem.Eng.87(6)(2009)869–878.

[12]P.Gonzalez-Contreras,J.Weijma,C.J.N.Buisman,Bioscorodite crystallization in an airlift reactor for arsenic removal,Cryst.Growth Des.12(5)(2012)2699–2706.

[13]A.Soare,R.Lakerveld,J.Royen,G.Zocchi,A.I.Stankiewicz,H.J.M.Kramer,Minimization of attrition and breakage in an airlift crystallizer,Ind.Eng.Chem.Res.51(33)(2012)10895–10909.

[14]Y.Wang,Q.S.Zhu,H.G.Zhang,Fabrication of β-Ni(OH)2and NiO hollowspheres by a facile template-free process,Chem.Commun.(2005)5231–5233.

[15]D.B.Kuang,B.X.Lei,Y.P.Pan,X.Y.Yu,C.Y.Su,Fabrication of novel hierarchical β-Ni(OH)2and NiO microspheres via an easy hydrothermal process,J.Phys.Chem.C 113(2009)5508–5513.

[16]X.J.Han,P.Xu,C.Q.Xu,L.Zhao,Z.B.Mo,T.Liu,Study of the effects of nanometer β-Ni(OH)2in nickel hydroxide electrodes,Electrochim.Acta 50(14)(2005)2763–2769.

[17]J.X.Ren,Z.Zhou,X.P.Gao,J.Yan,Preparation of porous spherical α-Ni(OH)2and enhancement of high-temperature electrochemical performances through yttrium addition,Electrochim.Acta 52(3)(2006)1120–1126.

[18]E.Shangguan,Z.R.Chang,H.W.Tang,X.-Z.Yuan,H.J.Wang,Synthesis and characterization of high-density non-spherical Ni(OH)2cathode material for Ni–MH batteries,Int.J.Hydrog.Energy 35(18)(2010)9716–9724.

[19]M.Aghazadeh,A.N.Golikand,M.Ghaemi,Synthesis,characterization,and electrochemical properties of ultra fine β-Ni(OH)2nanoparticles,Int.J.Hydrog.Energy 36(14)(2011)8674–8679.

[20]E.Shangguan,Z.R.Chang,H.W.Tang,X.-Z.Yuan,H.J.Wang,Comparative structural and electrochemical study of high density spherical and non-spherical Ni(OH)2as cathode materials for Ni–metal hydride batteries,J.Power Sources 196(18)(2011)7797–7805.

[21]S.Presto,D.Giraud,A.Testino,C.Bottino,M.Viviani,V.Buscaglia,Growth of polycrystalline nickel hydroxidefilms from aqueous solution,solution chemistry,deposition methods, film morphology and texture,Thin Solid Films 552(2014)1–9.

[22]W.W.E,J.C.Cheng,C.Yang,Z.-S.Mao,Experimental study by online measurement of the precipitation of nickel hydroxide:effects of operating conditions,Chin.J.Chem.Eng.23(5)(2015)860–867.

[23]M.X.Peng,X.Q.Shen,Template growth mechanism of spherical Ni(OH)2,J.Cent.S.Univ.Technol.14(2007)310–314.

[24]K.Borho,The importance of population dynamics from the perspective of the chemical process industry,Chem.Eng.Sci.57(2002)4257–4266.

[25]X.Q.Shen,M.X.Peng,M.X.Jing,Y.H.Wei,Study on structural characteristics of spherical Ni(OH)2electrode active material,J.Funct.Mater.36(11)(2005)1798–1801(1805).

[26]Y.Q.Cai,Research on Spherical Ni(OH)2Production Using a Oscillatory Baf fled Crystallizer Master Thesis Zhejiang Univ.,Hangzhou,China,2012.

[27]T.Kikuoka,K.Watanabe,Physical and electrochemical characteristics of nickel hydroxide as a positive material for rechargeable alkaline batteries,J.Appl.Electrochem.25(1995)219–226.

[28]M.X.Peng,Formation Mechanism of the Microstructures and the Electrochemical Performance for Spherical Nickel Hydroxide Ph.D.Thesis Central South Univ.,Changsha,China,2004.

[29]H.Xia,M.Zinke-Allmang,Rate equation approach to the late stages of cluster ripening,Physica A 261(1998)176–187.

[30]W.P.Zhang,Investigation of Macro-and Micro-mixing in Multiphase Loop Reactors Ph.D.Thesis,The University of Chinese Academy of Sciences,Beijing,China,2013.

[31]C.Sist,Precipitation of Nickel Hydroxide from Sulphate Solutions Using Supersaturation Control Ph.D.Thesis McGill University,Montreal,Canada,2004.