Zeren Shang,Mingchen Li,Baohong Hou,,Junli Zhang,2,Kuo Wang,Weiguo Hu,2,Tong Deng,Junbo Gong,Songgu Wu,

1 School of Chemical Engineering and Technology,State Key Laboratory of Chemical Engineering,Tianjin University,Tianjin 300072,China

2 North China Pharmaceutical Co.,Ltd.,Shijiazhuang 050015,China

Keywords:Cephalexin monohydrate Sonocrystallization Nucleation kinetics Fine-crystal dissolution policy Crystal habit control

ABSTRACT With the high-quality requirements for cephalexin monohydrate,developing a robust and practical crystallization process to produce cephalexin monohydrate with good crystal habit,appropriate aspect ratio and high bulk density as well as suitable flowability is urgently needed.This research has explored the influence of ultrasound on crystallization of cephalexin monohydrate in terms of nucleation mechanism and crystal habit control.The results of metastable zone width and induction time measurement showed the presence of ultrasound irradiation can narrow the metastable zone and shorten induction time.Cavitation phenomena generated by ultrasound were used to qualitatively explain the mechanism of ultrasound promoting nucleation of cephalexin monohydrate.Furthermore,on the basis of classical nucleation theory and induction time data,a series of nucleation-related parameters (such as crystalliquid interfacial tension,radius of the critical nucleus and etc.)were calculated and showed a decreasing trend under ultrasound irradiation.The diffusion coefficient of the studied system was also determined to increase by 72.73%under ultrasound.The changes in these parameters have quantitatively confirmed the mechanism of ultrasound influence on the nucleation process.In further,the calculated surface entropy factor has confirmed that the growth of cephalexin monohydrate follows continuous growth mechanism under the research conditions of this work.Through the exploration of crystallization conditions,it is found that suitable ultrasonic treatment,seeding,supersaturation control and removal of fine crystals are conducive to improving the quality of cephalexin monohydrate product.Optimizing the crystallization process coupled continuous ultrasound irradiation with fine-crystal dissolution policy has achieved the controllable production of monodisperse cephalexin monohydrate crystal with good performance.

1.Introduction

Cephalexin monohydrate (C16H17N3O4S·H2O,CAS Registry Number:23325-78-2,Fig.1) is an important cephalosporin which has been used to treat infections of skin and soft tissue,respiratory tract,urinary tract and etc.[1,2].This drug has a huge market demand and its annual output as an API is greater than 800 tons in China(reported in 2018 China Pharmaceutical Statistics Annual Report)[3].At present,the enzymatic process is mainly used in the industry to produce cephalexin monohydrate [4].As an important unit operation in this process,reactive crystallization determines the quality of cephalexin monohydrate product [5].

However,in the reactive crystallization of cephalexin monohydrate,it is prone to form large local supersaturation which leads to poor crystal habit such as needle or slender rod[5].The poor crystal habit causes many disadvantages including low bulk density,poor flowability and bad post-processing performance [6–10].Therefore,some researchers have reported on the reactive crystallization process of cephalexin monohydrate.Rousseau et al.investigated the influence of cephalexin precursors on the crystallization kinetics of cephalexin monohydrate [11].Hou et al.[5] provided a route of supersaturation control and seeding to obtain plate-like cephalexin monohydrate crystals.Based on its isoelectric point crystallization feature,Gong et al.[12]designed a continuous crystallization process of cephalexin monohydrate with a production capacity of 500 tons per year.Grover et al.[13] developed a 1-D population balance model for reactive crystallization of needle-like cephalexin monohydrate crystals.

Fig.1.The molecular structure of cephalexin monohydrate.

Sonocrystallization has been proven to be an effective way to manipulate the crystal habit,particle size distribution,solid form and purity of the product in many studies [14].Under suitable ultrasound conditions,solution could nucleate at lower supersaturation which is similar to seeding crystallization [15].Moreover,ultrasound could couple with seeding crystallization to promote the consumption of supersaturation for crystal growth synergistically,i.e.,seeds could be fragmented by ultrasound to provide more growth sites which is conducive to improving the crystal habit[16].Recent years,several control strategies (including supersaturation control,direct nucleation control,polymorphic feedback control and etc.) have been successfully developed to control multiple properties of crystals [16,17].Nagy et al.[18] compared four control strategies in crystallization of paracetamol and confirmed that automated direct nucleation control could alternately help crystal growth and fines removal in cooling/heating cycles which resulted in products with better crystal shape.For precise preparation of the metastable form α of L-glutamic acid,Braatz et al.[19]used supersaturation control to implement the corresponding supersaturation profile and obtain large crystal with narrow size distribution successfully.Inhibiting the sudden change of supersaturation in the reaction crystallization process is a key factor to obtaining crystal with well habit,therefore,coupling ultrasonic irradiation with multiple control strategies into the crystallization process of cephalexin monohydrate may be an effective means to improve product morphology.

The scope of this work was the development of a robust and convenient crystallization process to prepare cephalexin monohydrate with desired crystal habit.To achieve this goal,we firstly measured the solubility of cephalexin monohydrate in waterphosphoric acid system from 288.15 to 318.15 K.Then the influence of concentration,feeding rate,temperature and ultrasound on the metastable zone width of cephalexin monohydrate was studied experimentally.Additionally,the correlation between these parameters and metastable zone width was further confirmed by a statistical analysis.The induction time was also determined at various supersaturation without/with ultrasound.Then on the basis of classical nucleation theory,we used the data of induction time to calculate a series of nucleation-related parameters.Finally,the effect of multiple process parameters and feeding policies on the physical properties of cephalexin monohydrate crystals were assessed.The optimal crystallization process was developed on the basis of improved product quality.

2.Experimental

2.1.Materials

Cephalexin monohydrate (mass purity:98%) was supplied by North China Pharmaceutical Co.,Ltd.(Shijiazhuang,China),phosphoric acid (mass purity:85%) was purchased from Tianjin Jiangtian Chemical Reagent Co.,Ltd.(Tianjin,China),ammonia solution (mass purity:25%–28%) was obtained from Shanghai Aladdin Biochemical Technology Co.,Ltd.(Shanghai,China),deionized water was produced by our NANOPURE system from BARNSTEAD (Thermo Scientific Co.,Ltd.,China).All chemicals were used without further purification.

2.2.Experimental apparatus and characterization of crystals

The weighing operation in the experiment was achieved by an analytical balance (ML204,Mettler-Toledo,Switzerland).Experiments for Sections 2.3–2.5 were performed in a 200 ml jacketed crystallizer equipped with a four-blade propeller stirrer at the stirring rate of 200 r·min-1.A cryo-compact circulator (CF41,Julabo,Germany) was used to control the studied system to maintain at the set temperature.The measurement of pH and temperature was accomplished by using a pH meter (FE20,Mettler-Toledo,Switzerland).A peristaltic pump(BT300-1F,Longer Precision Pump Co.,Ltd.,Baoding,China)was employed for feeding ammonia solution,and the feeding pipe was submerged below the surface of the crystallization solution.The turbidity of the solution was monitored by a laser monitoring equipment (JD-3,Zhengzhou University,Zhengzhou,China) composed of laser generator,photoelectric transformer and laser beam intensity display.An ultrasonic processor (JY92-IIN,Ningbo Scientz Biotechnology Co.,Ltd.,Ningbo,China) was used for the corresponding ultrasoundassisted experiments in this work.This device has a 6 mm tip diameter with a power range of 60–600 W and the specific setting power parameters were described in Sections 2.3–2.5.Furthermore,the horn depth and the ultrasonic irradiation frequency were fixed at 2 cm and 21 kHz,respectively.The overview of experimental setup is depicted in Fig.2,and the experimental setup may be adjusted according to the operations described in Sections 2.3–2.5.

An optical microscope (Eclipse E200,Nikon,Japan)was used to observe and take pictures of the crystal.The aspect ratio(width of the crystal/length of the crystal) was obtained on the basis of statistical calculation of 200 crystals from the microscope images(using NIS-Elements software).The measurement of bulk density(mass of the powder/volume of the powder) and repose angle was carried out using a comprehensive powder characteristics tester (BT-1000,Bettersize Instruments Ltd.,Dandong,China).The solid form of crystals was identified by powder X-ray diffraction(PXRD) and the characterization was conducted on an X-ray powder diffractometer (Rigaku D/max 2500,Rigaku,Japan) using the Cu Kα radiation (λ=0.154 nm) at 40 kV and 100 mA.

Fig.2.The schematic diagram of experimental setup.

2.3.Solubility and metastable zone width measurements

The solubility of cephalexin monohydrate in water-phosphoric acid system was measured by the dynamic method based on a laser monitoring equipment.At a series of constant temperatures(288.15 K,298.15 K,308.15 K and 318.15 K),this measurement began with 140 ml water in the crystallizer.Then open the laser monitoring equipment,and the laser beam intensity became stable gradually (the corresponding value was also recorded as the maximum level).After adding known mass cephalexin monohydrate to the crystallizer,the monitoring equipment showed the decrease of the laser beam intensity due to the suspension of particles in the system.Next,we added phosphoric acid to the suspension liquid dropwise,and cephalexin monohydrate particles dissolved gradually.The interval time of phosphoric acid addition was adjusted to 30 min when the system was about to reach equilibrium.Finally,the system became transparent again(i.e.,the laser beam intensity has become more than 92%of the maximum level),the solid–liquid equilibrium was assumed to be established and the final pH value was recorded.Each measurement was conducted three times,and the arithmetic average of the final pH value was taken as the solubility (pHsolubility).The corresponding concentration of cephalexin monohydrate when the system was equilibrium (c*) could be calculated by Eq.(1).

where m is the mass of added cephalexin monohydrate;V represents the total volume of the final saturated solution.

The metastable zone width measurement of cephalexin monohydrate was achieved by laser monitoring method.Based on our measured solubility data,we used cephalexin monohydrate,water and phosphoric acid to prepare saturated aqueous solution of cephalexin monohydrate at certain pH and temperature.Then 150 ml saturated aqueous solution was put into the crystallizer.Under the constant temperature,we used a peristaltic pump to add ammonia solution to the system at different feeding rates(15 μl·min-1,30 μl·min-1and 150 μl·min-1) and the laser beam intensity before the introduction of ammonia solution was recorded as the maximum level.When cephalexin monohydrate started to crystallize out (i.e.,the intensity of laser beam was less than 92% of the maximum level),we recorded the corresponding pH value.Each measurement was repeated three times to reduce the error,and the average pH value was defined as the pHnucleation.The difference between pHsolubilityand pHnucleationis the metastable zone width (ΔpH) of corresponding conditions.The measurement of metastable zone width was implemented in the absence and presence of continuous ultrasound irradiation (50 W per 50 ml solution).In the ultrasound-assisted experiments,the ultrasound irradiation was applied when the system pH was constant at pHsolubilitytill the occurrence of nucleation.

2.4.Induction time measurements

To measure the induction time,150 ml clear saturated solution of cephalexin monohydrate(prepared by cephalexin monohydrate,water and phosphoric acid) was transferred into the crystallizer.Then a certain mass of ammonia solution was quickly added into the system to increase the pH to generate the set supersaturation(S,calculated by Eq.(2)).The time period between generating set supersaturation degree and beginning of nucleation(the judgment standard was the same as that of the metastable zone width measurement)was recorded,and the average value was defined as the induction time (tind) from three measurements.The experiments without/with ultrasound (50 W per 50 ml solution,continuous mode)were carried out to study the effects of ultrasound on induction time,and the ultrasonic processor was opened when the system reached the set supersaturation degree till the detection of nucleation.

where c and c* are the actual concentration and equilibrium concentration of solution,respectively.

2.5.Crystallization process

The physical properties(crystal habit,aspect ratio,bulk density,angle of repose and solid form) of cephalexin monohydrate obtained by different crystallization processes were compared.The pros and cons of these crystallization processes were discussed and the optimal crystallization policy to improve the quality of cephalexin monohydrate product was established.

2.5.1.General process parameters

The studied solution in the crystallization process composed of cephalexin monohydrate,water and phosphoric acid in the pH value of 1.60 and at the concentration of cephalexin monohydrate 120.00 mg·ml-1.The system was maintained at 298.15 K and the volume of studied solution was set as 150 ml.In seeding crystallization,the seed loading was 6% of the total mass of cephalexin monohydrate in the studied solution.When the crystallization was complete,the suspensions were filtered and vacuum dried at 313.15 K for 12 h to obtain the final product.The crystals at different stages from the beginning to the end of the crystallization process were analyzed for physical properties.

2.5.2.Modification of process parameters for ultrasound-assisted spontaneous crystallization and seeding crystallization

The influence of ultrasonic power and ultrasonic time on the spontaneous crystallization was studied at continuous ultrasound irradiation.The ultrasonic power was varied from 20 W per 50 ml solution to 50 W per 50 ml solution.Under the feeding rate of 30 μl·min-1,ammonia solution was added to the system until the spontaneous nucleation of cephalexin monohydrate.Additionally,the ultrasound irradiation lasted throughout the crystallization process.

The seeding crystallization started with the investigation of seeding pH (including 2.1,2.2 and 2.4).Ammonia solution was pumped into the studied solution at the rate of 30 μl·min-1,and the feeding was ceased when the system reached the seeding pH.Then added the seeds and set the breeding time as 100 minutes.

Next,under the seeding pH of 2.1,two different sets of experiments were conducted to study the effect of ultrasonic time and ultrasonic power in the seeding crystallization.It should be noted that the operation before seeding was consistent with the investigation of seeding pH as mentioned above.A set:After adding seeds,the continuous ultrasound irradiation (20 W per 50 ml solution)was introduced to the solution system until the completion of the experiment.Besides the breeding time was extended to 120 minutes.B set:Ultrasound in continuous mode was implemented to the studied solution when the seeding was complete.And the ultrasonic power investigated includes 20 W per 50 ml solution and 50 W per 50 ml solution.After breeding for 30 minutes,ammonia solution continued to be added to the system at the rate of 30 μl·min-1until the pH of system reached 4.94.The ultrasound irradiation lasted for 90 minutes.One control experiment without ultrasound irradiation was performed in the same conditions as experiment of B set for comparison.

Finally,the other two ultrasonic modes were tested.One is the continuous ultrasound irradiation was only introduced in the breeding period(set as 90 minutes),the other is pulsed ultrasound irradiation(10 s on with 10 s pause).These two modes’ ultrasonic power was constant at 20 W per 50 ml solution,and other operating conditions were same as experiment of B set.

2.5.3.Policy of variable-rate feeding and fine-crystal dissolution

The variable-rate feeding policy has different feeding rate of ammonia solution at four pH stages,the specific parameters are shown in Table 1.Except for changing feeding rate and omittingthe breeding stage,the rest of the operation was the same as the control experiment as described above.

Table 1 Feeding rate of ammonia solution in the variable-rate feeding policy

Moreover,we also attempted to eliminate the fine crystals produced by system fluctuations.The fine-crystal dissolution policy,i.e.,add phosphoric acid at certain pH (2.15 and 2.95) to adjust the pH back to about 0.05 and maintain for 20 minutes under this condition.The feeding rate of ammonia solution was 0 μl·min-1during the stage of fine-crystal dissolution.In addition,other parameters of this policy were the same as the variable-rate feeding policy.

2.5.4.The optimal crystallization policy

The optimal crystallization policy combined continuous ultrasound irradiation with fine-crystal dissolution policy.The ultrasound irradiation (20 W per 50 ml solution,continuous mode)was applied after seeding at pH=2.1 and lasted for 90 minutes.The control of process pH was achieved by following the finecrystal dissolution policy.

3.Results and Discussion

3.1.Solubility

Cephalexin monohydrate is prone to degradation under alkaline conditions,therefore,its reactive crystallization process is usually carried out under acidic conditions in practice [5].In this work,cephalexin monohydrate solubility data in water-phosphoric acid system from 288.15 to 318.15 K are shown in Table 2.It was found that the pHsolubilityof cephalexin monohydrate decreased with the increasing c*,which met with the characteristics of amphoteric substance.Moreover,under constant c*,pHsolubilitygradually increased with the rising temperature.The solubility of cephalexin monohydrate in water system has also been reported by other researchers,i.e.,Haghtalab et al.measured the solubility of cephalexin monohydrate in pure water from 283.1 to 323.1 K[20],Rousseau et al.[11]used the process analytical technology to obtain the cephalexin monohydrate solubility in aqueous solution from pH=4–8 at 278.15 K and 298.15 K,Hou et al.[5]displayed the solubility of cephalexin monohydrate at various pH values from 278.15 to 308.15 K.Due to the difference in the measurement range between this work and Refs.[11,20],we only converted the solubility data of Ref.[5] into the form of this work and listed them in Table S1.Then the graphical comparison of solubility data between Ref.[5] and this work is shown in Fig.S1,we could find the solubility values are very close and the slight deviation may be caused by the different acid used (Ref.[5]:sulfuric acid with mass purity ＞98%,this work:phosphoric acid with mass purity of85%).Therefore,the solubility determined by this work has high credibility.

Table 2 Solubility (pHsolubility) of cephalexin monohydrate from 288.15 to 318.15 K

3.2.Metastable zone width

The results of metastable zone width measurement are listed in Table S2 and depicted in Fig.3,it was found that c*,feeding rate,temperature and ultrasound irradiation had impact on the metastable zone width.In the absence of ultrasound irradiation,narrower metastable zone width could be found at the larger c*value,slower feeding rate and higher temperature.The reasons for this trend are as follows:(1) c* value is positively related to the probability of solute molecules colliding to form nuclei,therefore,larger c*value is conducive to nucleation of cephalexin monohydrate which is shown by narrower metastable zone width;(2)Due to the existence of the nucleation induction period,slower feeding rate gives the solution system more time for mixing,mass transfer and cluster formation,which narrow the metastable zone width;(3) Higher temperature intensifies the thermal movement of solute molecules and promotes nucleation.When the ultrasound irradiation was introduced,the metastable zone width was further reduced and this impact was more obvious at slower feeding rate.For example,at 298.15 K and c*=53.33 mg·ml-1,the metastable zone width was reduced by 44.62%(from 0.65 to 0.36)at the feeding rate of 30 μl·min-1and 21.52%(from 0.79 to 0.62)at the feeding rate of 150 μl·min-1.The mechanism behind the impact of ultrasound on metastable zone width could be explained by cavitation phenomena.Under the irradiation of ultrasound,the process of vibration,expansion,compression and collapse of the cavities or small bubbles contained in the liquid is defined as cavitation phenomena.It is well agreed that cavitation phenomena created by ultrasound waves could promote intense mixing,enhance mass transfer and increase energy transfer in crystallization systems,which are beneficial to nucleation.Thus,narrower metastable zone width of cephalexin monohydrate could be obtained by process intensification of ultrasound.

In order to quantify the correlation between metastable zone width (ΔpH) of cephalexin monohydrate and the four operating parameters,we used the HIPLOT platform to perform the correlation heatmap analysis.Fig.4 shows the ΔpH is positive correlation with the feeding rate,whereas c*,temperature and ultrasound irradiation present negative correlation.These results are consistent with the above analysis.From the correlation coefficients displayed,c*has the strongest negative correlation with the ΔpH,followed by temperature and ultrasound irradiation.In brief,it can be confirmed that the metastable zone width is subject to the combined influence of the four operating parameters.

3.3.Induction time

At 298.15 K,the induction time (tind) of cephalexin monohydrate within 1.700–4.017 supersaturation scope determined in the absence and presence of ultrasound irradiation was shown in Table S3 and Fig.5.Under the measurement conditions,the induction time decreased with the increasing supersaturation,which is similar to results for other systems.What’s more,the ultrasound experimental group showed shorter induction time with the decrease rate from 40.31%to 48.18%.Thus,same as the conclusion drawn in Section 3.2,it can be identified that the ultrasound irradiation has the effect of nucleation promotion due to the acoustic cavitation.

To quantify the dependence of crystallization kinetics on supersaturation and ultrasound,we used the classical nucleation theory to estimate a series of kinetic parameters.This theory states that the relationship between induction time(tind)and supersaturation(S) is written as Eq.(3) [21],

Fig.3.The metastable zone width(ΔpH)of cephalexin monohydrate:(a)288.15 K,(b)298.15 K,(c)308.15 K and(d)318.15 K.ΔpH1,ΔpH2,ΔpH3 are obtained in the absence of ultrasound irradiation with the different feeding rates of 15 μl·min-1,30 μl·min-1 and 150 μl·min-1 respectively,ΔpH4 and ΔpH5 are obtained in the presence of ultrasound irradiation with the different feeding rates of 30 μl·min-1 and 150 μl·min-1 respectively.

Fig.4.The heatmap of metastable zone width (ΔpH) related parameters.

where γCLis the crystal-liquid interfacial tension,Vmrepresents the molecular volume of solute molecule,k refers to Boltzmann constant,T is the absolute temperature,α and β indicate the slope and intercept of the linear equation of lntindagainst ln-2S.Therefore,the crystal-liquid interfacial tension(γCL)which is an index of the ability of solute to crystallize out of solution,could be calculated by Eq.(4) [22]:

Fig.5.Dependence of induction time (tind) of cephalexin monohydrate on supersaturation and ultrasound.

Growth mechanism of solute is estimated by surface entropy factor (f) which is obtained by Eq.(5) [23]:

The change of Gibbs free energy drives the nucleation and growth of solute,ΔGVis the Gibbs free energy change per unit volume and its calculation equation is as Eq.(6) [24],

The radius of the critical nucleus (r*) is calculated by Eq.(7)[25],

The extreme value of Gibbs free energy change that forms the critical nucleus is defined as the critical Gibbs free energy change(ΔG*),which is given by Eq.(8) [26]:

Finally,the number of molecules required to form the critical nucleus (i*) can be calculated by Eq.(9) [27]:

Fig.6 shows the relationship between lntindand ln-2S at 298.15 K without/with ultrasound.By linear fitting,it is observed that both of the fitting results are composed of two straight lines with different slopes(corresponding values are shown in Table 3).Based on literatures reporting similar phenomena [19,20],the following conclusion can be drawn that different slopes represent different nucleation mechanisms at corresponding supersaturations,i.e.,the large slopes (3.117 and 3.032) in the high supersaturation range indicate the homogeneous nucleation,and the small slopes(0.433 and 0.430) in the low supersaturation range indicate the heterogeneous nucleation.The reason for this phenomenon is large homogeneous nucleation rate in the high supersaturation range leads to the dominance of homogeneous nucleation,however,the inducing effect of existing impurities (dust or bubbles) and surfaces (crystallizer wall or propeller stirrer) in the studied systems on nucleation of cephalexin monohydrate is obvious in the low supersaturation range,which makes heterogeneous nucleation dominate.According to the slope corresponding to homogeneous nucleation,the crystal-liquid interfacial tension(γCL)in the studied uninsonated/insonated solutions could be calculated by Eq.(4),and the results are listed in Table 3.Accordingly,the smaller the value of γCLis,the easier to crystallize.Thence,the slight decrease in γCLunder the irradiation of ultrasound is further confirmed that ultrasound can promote nucleation of cephalexin monohydrate.To explore the growth mechanism of cephalexin monohydrate,we used Eq.(5) to obtain the surface entropy factor (f) values.Due to the results of f shown in Table 3 are all less than 3,the continuous growth mechanism is considered to be the mechanism of cephalexin monohydrate growth.

Fig.6.Plot of lntind versus ln-2S for cephalexin monohydrate at 298.15 K without/with ultrasound.

Four other nucleation parameters including Gibbs free energy change per unit volume (ΔGV),radius of the critical nucleus (r*),critical Gibbs free energy change (ΔG*) and number of molecules required to form the critical nucleus (i*) were calculated by Eqs.(6)–(9)respectively.These results are listed in Table 4 and depicted in Fig.7.It could be observed that these nucleation parameters are negative correlation with supersaturation in both the uninsonated and insonated systems,which is evidenced that nucleation is more likely to occur in the case of high supersaturation.Moreover,compared to the system without ultrasound,the application of ultrasound at same supersaturation also shows a certain reduction in r*,ΔG*and i*,which is consistent with the results that ultrasound could shorten the induction time of cephalexin monohydrate.

However,the reduction degree of the above parameters does not seem to match the reduction degree of the induction time of cephalexin monohydrate in the presence of ultrasound.The reason behind this phenomenon may be due to this work is carried out under constant temperature conditions which leads to the rapid elimination of the effects of heat release caused by ultrasonic cavitation.Therefore,the crystal-liquid interfacial tension which is sensitive to temperature change shows slight change and other parameters calculated based on it also show the same trend.Hence,there may be other factors that cause the effect of ultrasound on nucleation.Guo et al.[28] proposed the intercept (β) in Eq.(3) could be expressed as Eq.(10):

where K is the proportion constant,DABindicates the diffusion coefficient,C is the molar concentration,NArepresents the Avogadro’s number,CCis the molecular density of solid.

As shown in Table 3,the intercept corresponding to homogeneous nucleation is 2.225 without ultrasound,while the corresponding value under ultrasound is 1.683.Based on Eq.(10),it could be found that the diffusion coefficient in the studied system was increased by 72.73% under ultrasound.With the larger diffusion coefficient,the solutes could move faster from solution to make more clusters grow into crystal nuclei,which exhibits as shorter induction time in the presence of ultrasound.Therefore,we have determined that the effect of ultrasound irradiation on nucleation of cephalexin monohydrate is also reflected in the enhancement of the mass transfer and diffusion of the crystalliza-tion system.Similar conclusion is also reported in some literatures[28–30].

Table 3 The fitting parameters of lntind versus ln-2S,the crystal-liquid interfacial tension (γCL) and the surface entropy factor (F)

Table 4 The results of Gibbs free energy change per unit volume(ΔGV),radius of the critical nucleus(r*),critical Gibbs free energy change(ΔG*)and number of molecules required to form the critical nucleus (i*)

Fig.7.Primary nucleation parameters of cephalexin monohydrate at 298.15 K without/with ultrasound:(a)Gibbs free energy change per unit volume(ΔGV),(b)radius of the critical nucleus (r*),(c) critical Gibbs free energy change (ΔG*) and (d) number of molecules required to form the critical nucleus (i*).

3.4.Effect of ultrasonic power and time on spontaneous crystallization

Previous work reported unseeded crystallization of cephalexin monohydrate could only obtain undesired fine needle-like crystal habits [5].This work confirms that ultrasound irradiation could promote the nucleation of cephalexin monohydrate under lower supersaturation,which could theoretically suppress the burst nucleation phenomenon in the spontaneous crystallization process and improve the crystal habit.Therefore,we have introduced ultrasound into spontaneous crystallization of cephalexin monohydrate and hoped to explore its effect on crystal habit.And the microscopic images of cephalexin monohydrate at different stages under different ultrasonic power are represented in Figs.8 and 9 respectively.From Fig.8,it could be seen that the cephalexin monohydrate crystals are long needle-like and even have some deflection at the beginning of crystallization.After ultrasonic treatment for 73 minutes,the crystals still maintain the long needle-like morphology.However,the result of high power ultrasonic treatment(50 W per 50 ml solution)shown in Fig.9 indicates that some long needle-like crystals are broken along the radial direction and become shorter after ultrasonic treatment for 50 minutes.This behavior is attributable to higher-power ultrasound can produce stronger shock wave to break the long needle-like crystals.Although we have determined that the crystal length can be reduced by applying high-power ultrasound during the spontaneous crystallization process,the crystal habit is still needle-like which is deviated from the original intention of crystal habit improvement.So we continued to implement the ultrasoundassisted seeding crystallization of cephalexin monohydrate.

Fig.8.Microscope images of cephalexin monohydrate obtained by ultrasoundassisted(20 W per 50 ml solution)spontaneous crystallization:(a)at the beginning of crystallization,(b) ultrasonic treatment for 73 minutes.

Fig.9.Microscope images of cephalexin monohydrate obtained by ultrasoundassisted(50 W per 50 ml solution)spontaneous crystallization:(a)at the beginning of crystallization,(b) ultrasonic treatment for 50 minutes.

3.5.Effect of seeding pH

In the seeding crystallization,seeding pH is the key parameter to obtain product with good crystal habit.The morphology of the crystal after breeding is shown in Fig.10 and the dependency of aspect ratio on seeding pH is depicted in Fig.S2.It is evident that the aspect ratio of crystal decreases with increasing seeding pH.Therefore,we used pH=2.1 as the seeding pH in the next experiments.

3.6.Effect of ultrasonic time,power and mode on seeding crystallization

Firstly,for the experiment related to exploring the effect of ultrasonic time on the crystal habit of cephalexin monohydrate,continuous ultrasound of 20 W per 50 ml solution was used.Under different ultrasonic time,the change of crystal morphology is shown in Fig.11 and the results of corresponding aspect ratio are depicted in Fig.12.It has been observed that the particles agglomerate when the ultrasonic time is less than 60 minutes,but the crystals are broken and the resulting fine crystals adhere to the crystal surfaces when the ultrasonic time is up to 120 minutes.Moreover,Fig.12 displays the results of aspect ratio are around 0.35 which prove the crystal habit has been improved compared with the breeding experiment without ultrasound.Thus it is better to choose 90 minutes of ultrasonic time to improve the physical properties of the cephalexin monohydrate crystal.

Fig.10.The morphology of the crystal after breeding:(a) pH=2.1,(b) pH=2.2,(c) pH=2.4.

Fig.11.Comparison of crystal morphology under different ultrasonic time:(a)30 min,(b) 60 min,(c) 90 min,(d) 120 min.

Fig.12.The dependency of aspect ratio on ultrasonic time.

Next,we continued to explore the suitable ultrasonic power.The final product morphology under different ultrasonic power(0 W per 50 ml solution,20 W per 50 ml solution and 50 W per 50 ml solution)is presented in Fig.13 and the corresponding physical property information is listed in entry(1),(2)and(3)of Table 5 respectively.Furthermore,we also sealed the suspension with a parafilm after crystallization under different conditions,and observed the sedimentation status after standing for 30 minutes.The results of sedimentation are shown in Fig.S3.From Fig.13 and Fig.S3,it could be seen that product with poor crystal morphology and aspect ratio is formed without ultrasound.In addition,there are many fine crystals in this product,and the agglomeration phenomenon is serious.The suspension of this product shows obvious stratification,but the supernatant is turbid which is caused by suspension of fine crystals.Therefore,the product obtained by the control experiment has poor physical properties which could reduce yield and filtration efficiency.In the condition of 20 W per 50 ml solution,the morphology of the final product is good and the agglomeration phenomenon is not obvious.Meanwhile,its sedimentation supernatant is clear without obvious fine crystals,and its filtration performance is also good.Because the suitable ultrasonic power promotes nucleation at low supersaturation and makes crystal dispersion which could act as growth site for supersaturation consumption.However,the ultrasonic power of 50 W per 50 ml solution results in the product with poor morphology,which is attributed to excessive power causes excessive fragmentation of crystals.The phenomenon of crystal interweaving and agglomeration is also serious,which leads to the difficulty in stratified sedimentation and the extremely poor filtration performance.The results shown in Table 5 also prove ultrasonic power of 20 W per 50 ml solution could effectively improve the physical properties of cephalexin monohydrate product.Therefore,we selected 20 W per 50 ml solution as the suitable ultrasonic power.

Fig.13.Final product morphology with ultrasonic treatment of different power:(a) 0 W per 50 ml solution,(b) 20 W per 50 ml solution,(c) 50 W per 50 ml solution.

Finally,the effect of two other ultrasonic modes on seeding crystallization was explored.Fig.S4 represents the crystal morphology of final products under different ultrasonic modes.It can be found that the final product under the ultrasonic mode that continuous ultrasound irradiation was only introduced in the breeding period has long rod crystal habit with inevitable agglomeration phenomenon.The specific physical property information of product in this mode listed in entry(4)of Table 5 also proves product quality has not improved compared to the control experiment.Additionally,the final product of pulsed ultrasound mode also doesn’t significantly improve the quality,and its corresponding information is listed in entry (5) of Table 5.Possible cause of the above phenomena is the two ultrasonic modes have longer period without ultrasound relative to continuous ultrasound mode,and the agglomeration phenomenon could reoccur during the period without ultrasound which is not conducive to provide enough crystal surfaces for crystal growth.Therefore,continuous ultrasound mode could be a good selection for ultrasound-assisted seeding crystallization of cephalexin monohydrate.

3.7.Effect of different feeding policies

In order to achieve supersaturation control during the crystallization of cephalexin monohydrate,we divided the feeding rate of ammonia solution into four stages to avoid sudden changes in supersaturation.From the crystal morphology at different stages of variable-rate feeding operation shown in Fig.S5,it can be found that there is no burst nucleation in the whole process and the final product quality has been improved.The good result is due to the gradually slowing down of the feeding rate makes the crystallization system have more time to reach the crystallization equilibrium and proceed to Ostwald ripening when the crystallization is near the end point.Compared with the control experiment,the bulk density of variable-rate feeding operation product (shown in entry (6) of Table 5) has increased by 47.37%,and its morphology and aspect ratio have also been improved.

Although the overall supersaturation control is good under variable-rate feeding policy,however,the system concentration fluctuations caused by factors such as uneven mixing and local supersaturation will cause the generation of local fine crystals,which will reduce the final product quality.To overcome this problem,we used the fine-crystal dissolution policy and realized the effective removal of fine crystals in the crystallization process(Fig.S6).The information in entry (7) of Table 5 also shows the final product of fine-crystal dissolution policy has increased bulk density (with increase rate of 68.42%) and reduced repose angle(with decrease rate of 24%) compared to the control experiment.

Table 5 Physical properties and yields of final cephalexin monohydrate products

3.8.The establishment of the optimal crystallization policy

Comprehensive consideration of the pros and cons of the above crystallization methods,we added 20 W per 50 ml solution continuous ultrasound irradiation to the fine-crystal dissolution policy.Fig.14 shows less fine crystals appear during this crystallization process and the removal of fine crystals is also efficient.The final product has short rod crystal habit with good post-processing performance.The values shown in entry (8) of Table 5 also confirm this policy product has the best aspect ratio,bulk density and repose angle in this research.Compared with the control experiment,this policy product has achieved a 153.23% increase in the aspect ratio,a 68.42% increase in the bulk density and a 26.67%reduction in the repose angle.The reason is the mass transfer diffusion is enhanced by ultrasound,which reduces the appearance of local supersaturation and fine crystals.Meanwhile,the strengthening of diffusion also leads to faster removal of the fine crystals in the fine-crystal dissolution stage.And these factors are conducive to the improvement of product morphology and performance.The solid form of the crystals with different crystal habits obtained in this article are all proved to be cephalexin monohydrate(Fig.S7).Furthermore,this combined policy increases the yield to 92%,which further proves that this crystallization policy is robust and optimal.

Fig.14.Crystal morphology at different stages of the optimal crystallization policy:(a)seeding,(b)unexpected occurrence of fine crystals,(c)dissolution of unexpected fine crystals,(d) final product.

4.Conclusions

This research mainly explored the effect of ultrasound on nucleation behavior and crystal habit of cephalexin monohydrate.Ultrasound is proved to be negatively correlated with metastable zone width.And the induction time could also be reduced in the presence of ultrasound.The mechanism of promoting nucleation with ultrasound is cavitation phenomena which could enhance mass transfer and lower the energy barrier for nucleation.The nucleation-related parameters including γCL,r*,ΔG* and i* which were calculated by induction time,have a decreasing trend under ultrasound irradiation.Moreover,the diffusion coefficient in insonated system was determined to increase by 72.73% compared to uninsonated system.Regardless of whether there is ultrasound irradiation,cephalexin monohydrate is dominated by homogeneous nucleation in the high supersaturation range and heterogeneous nucleation in the low supersaturation range respectively.The mechanism of cephalexin monohydrate growth without/with ultrasound is also determined to be the continuous growth mechanism.

It is found that suitable ultrasonic power,ultrasonic time,ultrasonic mode,seeding,supersaturation control and removal of fine crystals are conducive to improving the quality of cephalexin monohydrate product.The continuous ultrasound with ultrasonic power(20 W per 50 ml solution)and ultrasonic time(90 minutes)was coupled with the fine-crystal dissolution policy to establish the optimal crystallization policy,which could effectively control the production of monodisperse cephalexin monohydrate crystal with short rod crystal habit,nice aspect ratio (0.471),high bulk density(0.32 g·ml-1)and improved flowability(repose angle=55°).In short,the influence of ultrasound on crystallization of cephalexin monohydrate in terms of nucleation promotion and crystal habit improvement has been identified for the first time.

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

The authors are grateful to the financial support of National Natural Science Foundation of China (22078238) and Special Project for the Transformation of Major Scientific and Technology Achievements of Hebei Province (19042822Z).

Supplementary Material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.cjche.2021.07.034.

Nomenclature

C molar concentration,kmol·m-3

CCmolecular density of solid,kmol·m-3

c actual concentration,mg·ml-1

c* equilibrium concentration,mg·ml-1

DABdiffusion coefficient,m2·s-1

f surface entropy factor

ΔG* critical Gibbs free energy change,J

ΔGVGibbs free energy change per unit volume,J·m-3

i* number of molecules required to form the critical nucleus

K proportion constant

k Boltzmann constant,J·K-1

m mass of added cephalexin monohydrate,mg

NAAvogadro’s number,mol-1

pHnucleationnucleation point expressed in pH form

pHsolubilitysolubility expressed in pH form

ΔpH metastable zone width

r* radius of the critical nucleus,nm

S supersaturation

T absolute temperature,K

tindinduction time,s

V total volume of the final saturated solution,ml

Vmmolecular volume of solute molecule,m3

α slope of the linear equation of lntindagainst ln-2S

β intercept of the linear equation of lntindagainst ln-2S

γCLcrystal-liquid interfacial tension,mJ·m-2