ch07Acquisition and Restructuring Strategies(战略管理,英文版)

合集下载

Catalytic CO2 hydration by immobilized and free human carbonic anhydrase II

Catalytic CO 2hydration by immobilized and free human carbonic anhydrase II in a laminar ﬂow microreactor –Model and simulationsIon Iliuta,Maria Cornelia Iliuta,Faical Larachi ⇑Department of Chemical Engineering,Laval University,Québec,Canada G1V 0A6a r t i c l e i n f o Article history:Received 2April 2012Received in revised form 9November 2012Accepted 9January 2013Available online 26January 2013Keywords:Laminar ﬂow microreactor CO 2hydrationHuman carbonic anhydrase II Modeling Simulationa b s t r a c tEx vivo applications of human carbonic anhydrase II (HCA II)for its potential in CO 2capture technologies are emerging owing to the formidably large hydration turnover number Nature endowed this enzyme with to catalyze aqueous hydration of CO 2near diffusion limits.In this work,we investigated the CO 2hydration process catalyzed by solution-phase or immobilized HCA II enzyme in a laminar ﬂow microreactor with the purpose to simulate the reaction–transport of HCA II in microchannels.The effects of operating condi-tions as well as the contribution of carbonic anhydrase on the performances of the CO 2hydration process are presented.Numerical simulations indicate that in laminar ﬂow microreactor with HCA II immobilized on the inner surface of the tube,interpreting the data as a one-dimensional plug ﬂow results will lead to signiﬁcant error.Therefore,coupling of transport phenomena and surface enzymic reaction necessitates the use of a two-dimensional analysis.Simulations reveal that hydrodynamic and diffusional constraints do not permit reasonable utilization of the immobilized HCA II enzyme in a laminar ﬂow microreactor,even if HCA II has a very high hydration turnover and the uncatalyzed bulk CO 2hydration is the dominant pro-cess.In the microreactor with solution-phase HCA II enzyme ‘‘plug ﬂow’’is achieved under laminar ﬂow conditions and the contribution of uncatalyzed CO 2hydration process is not considerable.Ó2013Elsevier B.V.All rights reserved.1.IntroductionThe reversible hydration of carbon dioxide catalyzed by human carbonic anhydrase II (HCA II)in aqueous solutions has been exten-sively investigated,mainly from a biochemistry and catalytic standpoint [1–4].HCA II-catalyzed CO 2=HCO À3inter-conversion plays a signiﬁcant role in a multitude of physiological processes such as pH homoeostasis,respiratory gas exchange,photosynthe-sis,ion transport,as well as it is a fairly important reaction for drug design [5]and has been thoroughly investigated and described in a number of reviews [6–9].Emerging ex vivo applications of HCA II for its potential use in CO 2capture and sequestration technologies have recently at-tracted the researchers’attention [10,11].The main incitement to this interest is the very high hydration turnover,k h %106s À1,and 2nd-order rate constant,k h =K CO 2%108M À1s À1,that Nature endowed this enzyme with to effectuate catalytic hydration of CO 2near the limits imposed by diffusion encounters in aqueous media [12].Unfortunately,application of free HCA II enzyme in solution-phase is not always suitable and optimal because of the large volume of enzyme required.Binding of HCA II enzyme on a solid support is an attractive modiﬁcation of its application having several advantages,including easier separation of the reaction products without catalyst contamination,ability to recover and re-use the enzyme,increase of the enzyme stability and operational lifetime,continuous operation of enzymatic processes and ﬂexibil-ity of the reactor design [13,14].However,attaching HCA II to a so-lid macrosurface may lead the enzyme to behave differently [14]because:(i)the immobilization may cause the enzyme molecules to adopt a different conformation;(ii)the immobilized enzyme ex-ists in an environment different from that when it is in solution-phase;(iii)there is a partitioning of substrate between the solution and support,with the result that the substrate concentration in the neighborhood of the enzyme may be signiﬁcantly different from that in the bulk solution;and (iv)diffusional effects play a more important role with immobilized enzymes.The present contribution focuses on the CO 2hydration process catalyzed by solution-phase or immobilized HCA II enzyme in a laminar ﬂow microreactor –which allows strict control of reaction conditions in time.The objective lies on exploring the possibility to use this micro enzyme reactor system as a tool for further under-standing and development of CO 2hydration process with a view to elaborate a comprehensive theoretical framework of these sys-tems and to apply it for experimental data reduction.The behavior of CO 2hydration laminar ﬂow microreactor with human carbonic anhydrase attached to the inner surface of the tube was explored using a detailed kinetic model developed for reversible CO 2hydra-tion catalyzed by solution-phase HCA II (pseudo random Quad1383-5866/$-see front matter Ó2013Elsevier B.V.All rights reserved./10.1016/j.seppur.2013.01.006⇑Corresponding author.E-mail addresses:ion.iliuta@gch.ulaval.ca (I.Iliuta),maria.iliuta@gch.ulaval.ca (M.C.Iliuta),rachi@gch.ulaval.ca (rachi).Quad Iso Ping Pong mechanism with one transitory complex[15]). Particular attention has been given to the following items:(i)util-ity of the numerical simulations for reﬁning the reactor operating conditions when determining the catalyzed CO2hydration kinetics data in a laminarﬂow microreactor with immobilized human car-bonic anhydrase,(ii)numerical identiﬁcation of conditions,if any, to approximate plugﬂow operation,and(iii)evaluation of the ef-fects of uncatalyzed CO2hydration and two-dimensionality of the ﬂow on laminarﬂow microreactor performance.Finally,we reveal the difﬁculties that result in interpreting the data obtained when HCA II is immobilized on the microreactor wall.minarﬂow microreactor modelThe system considered consists of a circular tube with solution-phase HCA II enzyme or with HCA II enzyme uniformly attached on its inner surface.The microreactor is isothermal.The entireﬂow in the tube may be viewed as consisting of three sections[13]:the so-called hydrodynamic inlet section,the concentration inlet section, and the fully developed section.In the hydrodynamic inlet section the initiallyﬂat liquid velocity proﬁle evolves toward a parabolic velocity proﬁle which remains translationally invariant in the downstream direction.The hydrodynamic inlet section is esti-mated to be fairly short(less than1mm)compared to the total length of the tube(0.1m)and we may consider that the laminar ﬂow with a parabolic velocity proﬁle is developed from the en-trance of the tube.Owing to the enzymatic reaction in solution-phase or on the tube wall and the diffusion of substrate towards the wall,the initiallyﬂat concentration proﬁle changes gradually and becomes fully established in the third region.Flat entrance velocity proﬁle would tend to shorten residence time of liquid lay-ers nearby the wall.In the presence of the chemical reaction,this leads to radial reactant concentration gradients which are larger than those with parabolic velocity proﬁles.Hence,‘‘all-through’’parabolic velocity proﬁles are expected to lead to lesser conver-sions thanﬂat entrance velocity proﬁles evolving towards para-bolic.However,it is reasonable to assume that our simulations lead to conservative estimation of CO2conversion in the presence of solution-phase HCA II enzyme or HCA II immobilized enzyme.The pseudo random Quad Quad Iso Ping Pong mechanism with one transitory complex,which implies a possible competitive in-ter-molecular proton transfer step by the CO2=HCOÀ3pair with re-spect to external buffer(B),was used to describe the reversible hydration of carbon dioxide catalyzed by human carbonic anhy-drase II[15]:CO2þZnOHÀðEÞ¢ZnHCOÀ3ðESÞð1ÞH2OþZnHCOÀ3ðESÞ¢HCOÀ3þZnH2OðE WÞð2ÞE W¢H Eð3ÞBþZnH2OðH EÞ¢BHþþZnOHÀðEÞð4ÞHCOÀ3þZnH2OðH EÞ¢CO2þH2OþZnOHÀðEÞð5ÞThe mechanism of uncatalyzed hydration of CO2and dehydra-tion of H2CO3under the conditions used in enzymatic process was described in the following way[16]:H2OþCO2¢k31k13HþþHCOÀ3ð6ÞHþþHCOÀ3¢k12k21H2CO3ð7ÞH2OþCO2¢k32k23H2CO3ð8Þ2.1.Model for CO2hydration laminarﬂow microreactor with immobilized HCAII enzymeThe unsteady-state mass balance equations for a chemical species j in the liquid phase are formulated taking into account that in laminarﬂow regime the transport in the lateral direction occurs as a result of molecular diffusion only and the transport in the longitudinal direction occurs by both advection and diffusion:@C CO2@tþ2m‘1ÀrR2@CCO2@z¼D CO2@2C CO2þD CO21@r@C CO2ÀR ucCO2ðC jÞð9Þ@C HCOÀ3þ2m‘1Àr 2@CHCOÀ3¼D HCOÀ3@2C HCOÀ3þD HCOÀ31@r@C HCOÀ3þR ucCO2ðC jÞð10Þ@C B@tþ2m‘1ÀrR2@CB@z¼D B@2C B@z2þD B1r@@rr@C B@rð11Þ@C BHþ@tþ2m‘1ÀrR2@CBHþ@z¼D BHþ@2C BHþ@z2þD BHþ1r@@rr@C BHþ@rð12ÞTo complete the description of the system,the following initial and boundary conditions are written:t¼0C jð0;z;rÞ¼C injð13Þz¼0C jðt;0;rÞ¼C injð14Þz¼L@C j@zðt;L;rÞ¼0ð15ÞNomenclaturea s speciﬁc surface area,m2=m3reactorC E0enzyme load,kmol=m3reactorC j concentration of species j in liquid phase,kmol/m3D j molecular diffusion coefﬁcient in liquid phase,m2/s L microreactor length,mr radial position within microreactor,mR microreactor radiusR j reaction rate,kmol/m3s t time,sv‘liquid velocity,m/s z axial coordinate,m Subscripts/Superscriptsc catalyzedin microreactor inlet uc uncatalyzed62I.Iliuta et al./Separation and Puriﬁcation Technology107(2013)61–69r¼0@C j@rðt;z;0Þ¼0ð16Þr¼RÀD j @C jðt;z;RÞa s¼R cjðC j jr¼RÞð17ÞThe boundary condition selected for the outlet does not set any restrictions except that convection dominates transport out of the reactor.Thus this condition keeps the outlet boundary open with-out restrictions on the concentration.2.2.Model for CO2hydration laminarﬂow microreactor with solution-phase HCAII enzymeThe unsteady-state mass balance equations for a chemical spe-cies j in the liquid phase are:@C CO2 @t þ2m‘1ÀrR2@CCO2@z¼D CO2@2C CO2@z2þD CO21r@@rr@C CO2@rÀR ucCO2ðC jÞÀR cCO2ðC jÞð18Þ@C HCOÀ3þ2m‘1Àr 2@CHCOÀ3¼D HCOÀ3@2C HCOÀ3þD HCOÀ31@r@C HCOÀ3þR ucCO2ðC jÞþR cCO2ðC jÞð19Þ@C B @t þ2m‘1ÀrR2@CB@z¼D B@2C B@z2þD B1r@@rr@C B@rÀR cCO2ðC jÞð20Þ@C BHþ@t þ2m‘1ÀrR2@CBHþ@z¼D BHþ@2C BHþ@z2þD BHþ1r@@rr@C BHþ@rþR cCO2ðC jÞð21ÞThe initial and boundary conditions are:t¼0C jð0;z;rÞ¼C injð22Þz¼0C jðt;0;rÞ¼C injð23Þz¼L@C j@zðt;L;rÞ¼0ð24Þr¼0@C jðt;z;0Þ¼0ð25Þr¼R@C jðt;z;RÞ¼0ð26ÞThe boundary condition selected for r=R makes sure that nomaterialﬂow through the reactor wall.2.3.Uncatalyzed CO2hydration kineticsThe overall rate of uncatalyzed conversion of dissolved CO2tobicarbonate developed by Ho and Sturtevant(1963)was used[16]:R ucCO2¼k031C CO2Àk013C HþC HCOÀ3where k031¼k31þk32;k013¼k13þk23=K H2CO3ð27ÞThe rate constants at25°C are:k031¼0:037sÀ1andk013¼5:5Â104m3=kmol s[16].Table1The rate constant aggregates,turnover,apparent Michaelis and inhibition constants [15].Rate constant aggregates Turnover,apparent Michaelis and inhibitionconstantsk3k1¼9:5Â10À3M k h¼1:1Â106sÀ1;k d¼9:4Â104sÀ1 K Ea1k1þk5¼8:4Â106MÀ1sÀ1K CO2¼9:5mMk1kÀ1¼1:11Â103MÀ1K B1¼5:2mM;K BHþ1¼0:23mMK HCOÀ3¼22:2mMK i HCOÀ3;1¼14:8mM;K i HCOÀ3;2¼127:9mM;K i HCOÀ3;3¼37:5mM;K i HCOÀ3;4¼61:9mM;K i HCOÀ3;5¼12:2mMTable2The equilibrium constants[15].Equilibrium constant ValueCO2þH2O¢HCOÀ3þHþKa1¼½HCOÀ3 ½Hþ2;mol=l p K a1=5.97Proton-transfer group acid dissociationH E¢EþHþKE¼½E ½Hþ½H E;mol=l p K E=7.1Catalytic group acid dissociationE W¢EþHþKE ¼½E ½Hþ½E W;mol=l p K E%7.1BþHþBHþKa2¼½B ½Hþ½BHþ;mol=lBuffer:Na2HPO4p K a2=7.2 1,2-Dimethylimidazole(1,2-DMI)p K a2=8.2I.Iliuta et al./Separation and Puriﬁcation Technology107(2013)61–69632.4.Catalyzed CO2hydration kineticsThe pseudo random Quad Quad Iso Ping Pong mechanism with one transitory complex,which implies a possible competitive in-ter-molecular proton transfer step by the CO2=HCOÀ3pair with re-spect to external buffer,B,was used to describe the reversible hydration of carbon dioxide catalyzed by HCA II[15]:In addition to enzyme isomerization,the model takes into ac-count the intermolecular CO2=HCOÀ3-subtended proton transfer viaa½CO2 Á½HCOÀ3coupling,the CO2=HCOÀ3-subtended proton transfervia½HCOÀ32and½CO2 Á½HCOÀ32couplings,and the enzyme-substratetransitory complex via½CO2 Á½HCOÀ3Á½B coupling.The hydration and dehydration turnover rate constants,k h and k d,the apparent Michaelis constants,K CO2,K HCOÀ3,K B,KþBH,and the apparent bicarbon-ate inhibition constants,K i HCOÀ3;jare deﬁned as follows:k h%K a1K Ek3k1K EK a1k1þk5;k d%K a1K Ek3k1K EK a1k1þk51þk3À1ð29ÞK CO2%k3k1;K HCOÀ3%21þk3kÀ1K a1K Ek3k1;K B¼K B1K EþK a2K E;K BHþ¼K BHþ1K EþK a2K a2ð30ÞTable3Laminarﬂow microreactor operating conditions.Operating conditions DataChannel diameter 2.0mmMicroreactor length0.1mActive HCA II loading4:32Â10À7kmol=m3reactor Microreactor temperature298KInlet CO2concentration0.017mol/lInlet superﬁcial liquid velocity0.0025–0.0.005m/sR c CO2¼k h C CO2C BÀK a2a1C HCOÀ3C BHþ1þC HCOÀ3i HCOÀ3;3C E0K B C CO2þK CO2C BþK B2K EK a1C HCOÀ3þ2K CO2K a2K EC BHþþC CO2C Bþ2K CO2K HCOÀ3K a2K EC HCOÀ3C BHþþK BK i HCOÀ3;1C CO2C HCOÀ3Â1K CO2K i HCOÀ3;2C B C HCOÀ3þ12K BK i HCOÀ3;3K EK a1C HCOÀ32þK BK i HCOÀ3;1K i HCOÀ3;4C CO2C HCOÀ32þ1K i HCOÀ3;5C CO2C B C HCOÀ3ð28Þ64I.Iliuta et al./Separation and Puriﬁcation Technology107(2013)61–69K i HCOÀ3;1%2K a1K Ek1kÀ1þ2k1k3k5K Ea1k1þk5!À1;K i HCOÀ3;2¼K a1EK CO2;K i HCOÀ3;3¼K a1Ek31K EK a1k1þk55ð31ÞK i HCOÀ3;4¼K2i HCOÀ3;3K i HCOÀ3;3ÀK i HCOÀ3;1;K i HCOÀ3;5¼121K i HCOÀ3;1À1K i HCOÀ3;3!À1ð32ÞThe rate constant aggregates with the inferred turnover and apparent Michaelis constant and inhibition constants are tabulated in Table1.The equilibrium constants are given in Table2.The ki-netic model was developed for reversible hydration of carbon diox-ide in the presence of solution-phase human carbonic anhydrase II. However,the kinetic model is expected to be suitable under immo-bilization enzyme conditions taking into account the comparable CO2removal efﬁciency of the immobilized HCA II and the soluble counterpart for extended periods[17].2.5.Method of solutionIn order to solve the system of partial differential equations,we discretized in space and solved the resulting set of ordinary differential equations.The spatial discretization was performed using the standard cell-centeredﬁnite difference scheme.The GEAR integration method for stiff differential equations was em-ployed to integrate the time derivatives.The relative error toler-ance for the time integration process in the present simulations was set at10À6for each time step.3.Results and discussionThe model was initially used to compare the performance of the laminarﬂow microreactor with solution-phase or immobilized HCA II enzyme under the same HCA II loading.Both catalyzed and uncat-alyzed CO2hydration processes were considered.Figs.1and2show typical CO2and HCOÀ3radial and axial steady-state concentration proﬁles obtained under the same operating conditions(Table3).Dif-fusional effects are more important with immobilized HCA II en-zyme(Fig.1a)and the result is a lower CO2conversion(Fig.2a). On the other side,with solution-phase HCA II enzyme,the species concentration is nearly uniform in the radial direction(Fig.1b) and this is close to the ideal‘‘plugﬂow’’conditions[18].Numerical simulations indicate that in laminarﬂow microreactor with the HCA II immobilized on the inner surface of the tube the coupling of transport phenomena and chemical reaction necessitates the use of two-dimensional analysis in processing the experimental data.I.Iliuta et al./Separation and Puriﬁcation Technology107(2013)61–6965Interpreting the data as one-dimensional plugﬂow results will leadto signiﬁcant error.It is of interest to investigate the behavior of CO2hydration pro-cess by forcing artiﬁcially silencing of the uncatalyzed conversion of dissolved CO2to bicarbonate(Figs.3and4).As expected,with solution-phase HCA II enzyme the contribution of uncatalyzed CO2hydration is reduced and the overall reaction rate is dominated by the enzymatic process(Fig.4b).On the other side,the uncata-lyzed CO2hydration is the prevailing process when HCA II enzyme is immobilized on the inner surface of the tube.Due to consider-able diffusional limitations(Fig.3a),a relatively small amount of HCOÀ3is produced by catalyzed CO2hydration process(Fig.4a). Therefore,the hydrodynamic and diffusional constraints in laminar ﬂow microreactors do not permit a reasonable utilization of HCA II enzyme immobilized on the inner surface of the tube.It is antici-pated that such behavior will generate difﬁculties in interpreting experimental data obtained with immobilized HCA II enzymes.Un-der these conditions,the microreactor conﬁguration must be se-lected accurately to exploit the absorption potential of HCA II enzyme(e.g.,enhanced mixing to disrupt adjacent laminarﬂuid streams by adding internals in the microchannel).In laminarﬂow microreactors,an important parameter which dictates mixing in the radial direction is the molecular diffusion coefﬁcient.The inﬂuence of the diffusion coefﬁcients on the CO2 hydration process under immobilized HCA II enzyme conditions is illustrated in Figs.5and6where both catalyzed and uncatalyzed CO2hydration processes were considered.It is evident that an in-crease in the molecular diffusion coefﬁcient(a10-fold increase with respect to estimated values with Frank et al.[19]and Wil-ke-Chang(Reid et al.[20])correlations)leads to higher mass trans-ferﬂuxes transported between the liquid and the catalytic solid surface(Fig.5)and the result is a higher CO2conversion(Fig.6a). Theoretically,as the molecular diffusion coefﬁcient continues to increase,the mixing in the radial direction will become faster as well as theﬂux of CO2toward the walls will be promoted.The high degree of mixing in the radial direction will lead to a more uniform distribution of mass across the cross section and a higher CO2 conversion.Figs.7and8show CO2and HCOÀ3axial and radial steady-state concentration proﬁles obtained for two values of liquid velocity in the laminarﬂow microreactor with immobilized HCA II enzyme. Two cases were simulated:(i)catalyzed CO2hydration process was considered only,and(ii)both catalyzed and uncatalyzed CO2 hydration processes were considered.As expected,CO2conversion increases with the decrease of liquid velocity due to higher resi-dence time(Fig.7).At lower liquid velocity,diffusional limitation continues to be considerable(Fig.8a)and the uncatalyzed CO2 hydration becomes more important(Figs.7and8b).The perfor-mance of the laminarﬂow microreactor with solution-phase HCA II enzyme increases slowly(not shown)with the decrease of liquid velocity.66I.Iliuta et al./Separation and Puriﬁcation Technology107(2013)61–69Figs.9and10show the effect of the inlet buffer(Na2HPO4)con-centration on the axial and radial steady-state concentration pro-ﬁles without uncatalyzed CO2hydration in the laminarﬂow microreactor with solution-phase or immobilized HCA II,under the same HCA II loading.Buffers in solution participate as pro-ton-transfer agents in the catalyzed CO2hydration process.Low buffer concentration displaces the hydration into a regime where the inter-molecular proton transfer is rate determining and CO2 hydration rate is small.On the contrary,sufﬁciently high buffer concentrations ensures that inter-molecular proton transfer is not rate limiting and CO2hydration rate is large.This behavior ex-plains the raise of CO2conversion with the increase of inlet buffer concentration(Figs.9and10a).The enhancement is less important with immobilized HCA II enzyme because of the lower mass trans-ferﬂuxes of buffer transported between the liquid and the immo-bilized HCA II enzyme due to the diffusional limitation(Fig.9b). Fig.10b shows,once more,that the solution-phase HCA II enzyme system is a non-diffusion limited system and‘‘concentration’’plug ﬂow is achieved in laminarﬂow.Fig.11shows the inﬂuence of the type of buffer(characterized by the equilibrium constant K a2)on the radial steady-state concen-tration proﬁles in the laminarﬂow microreactor with immobilized HCA II enzyme.Both catalyzed and uncatalyzed CO2hydration processes were considered.The buffers used in simulations are: Na2HPO4and1,2-dimethylimidazole.CO2conversion increases with the decrease of equilibrium constant K a2due to the higher CO2hydration driving force.The raise is less important with immo-bilized HCA II enzyme due to diffusional limitation(Fig.11).In the laminarﬂow microreactor with immobilized HCA II en-zyme–a diffusion limited system as mention above–the cata-lyzed CO2hydration process is largely dictated by the diffusive ﬂuxes between the liquid and the immobilized HCA II enzyme. Therefore,the system is not very sensitive to the increase of cata-lyzed CO2hydration kinetic parameters.A10-fold increase in the reactor HCA II loading and hydration turnover does not have a sig-niﬁcant effect on the CO2hydration process.So,because of diffu-sion limitations,the laminarﬂow microreactor with immobilized HCA II enzyme on the internal surface is not suitable for kinetic studies,unless a method to enhanceﬂuid mixing is employed. On the contrary,in laminarﬂow microreactor with solution-phase HCA II enzyme,non-diffusion limited system,the catalyzed CO2 hydration process is largely dictated by the kinetics.A10-fold in-crease in molecular diffusion coefﬁcients values does not have a signiﬁcant effect on the CO2hydration process.4.ConclusionThe behavior of CO2absorption enhanced by the enzyme car-bonic anhydrase in a laminarﬂow microreactor was studied withI.Iliuta et al./Separation and Puriﬁcation Technology107(2013)61–6967the purpose to understand the mechanism of HCA II enzyme reac-tions in a microchannel for further development of this process. The effects of operating conditions as well as the contribution of carbonic anhydrase on the performances of CO2hydration are pre-sented.Numerical simulations indicate that in laminarﬂow mic-roreactor with the HCA II immobilized on the inner surface of the tube the coupling of transport phenomena and chemical reaction necessitates the use of a two-dimensional analysis for data inter-pretation as a one-dimensional plugﬂow description will lead to signiﬁcant error.The laminarﬂow microreactor with immobilized HCA II enzyme is a diffusion limited system and hydrodynamic and diffusional constraints do not permit a reasonable utilization of the immobilized HCA II enzyme even if HCA II has a very high hydra-tion turnover(a better mixing is necessary to disrupt adjacent lam-inarﬂuid streams).The uncatalyzed CO2hydration is the dominant process and this behavior will generate difﬁculties in interpreting the experimental data.Because of diffusion limitations,the laminar ﬂow microreactor with immobilized HCA II enzyme on the internal surface is not suitable for kinetic studies,unless a method to en-hanceﬂuid mixing is employed.In the microreactor with solu-tion-phase HCA II enzyme–a non-diffusion limited system,‘‘concentration plugﬂow’’is achieved under laminarﬂow condi-tions,the enzymatic contribution to the overall reaction rate is large and therefore the contribution of uncatalyzed CO2hydration process is not considerable.Any future modeling studies of mixing in capillaries equipped with internals to enhance mass transfer must benchmark or com-pare the enzymatic performance results with the present empty capillary case.A branch which can take advantage of presentﬁnd-ing is the design of reliable‘‘kinetic study’’devices more appropri-ate for immobilized enzymes.Our present contribution is on the right direction to help elucidating diffusion–reaction couplings for high-turnover enzymes.References[1]S.H.Koenig,R.D.Brown,H2CO3as substrate for carbonic anhydrase in thedehydration of HCOÀ3,Proc.Nat.Acad.Sci.69(1972)2422–2425.[2]D.N.Silverman,S.Lindskog,The catalytic mechanism of carbonic anhydrase:implications of a rate-limiting protolysis of water,Acc.Chem.Res.21(1988) 30–36.[3]S.Lindskog,Structure and mechanism of carbonic anhydrase,Pharmacol.Ther.74(1997)1–20.[4]D.N.Silverman,R.McKenna,Solvent-mediated proton transfer in catalysis bycarbonic anhydrase,Acc.Chem.Res.40(2007)669–675.[5]V.M.Krishnamurthy,G.K.Kaufman,A.R.Urbach,I.Gitlin,K.L.Gudiksen,D.B.Weibel,G.M.Whitesides,Carbonic anhydrase as a model for biophysical and physical–organic studies of proteins and protein–ligand binding,Chem.Rev.108(2008)946–1051.[6]Y.Pocker,S.Sarkanen,Carbonic anhydrase:structure,catalytic versatility,andinhibition,Adv.Enzymol.47(1978)149–274.[7]S.Lindskog,Carbonic anhydrase,Adv.Inorg.Biochem.4(1982)115–170.68I.Iliuta et al./Separation and Puriﬁcation Technology107(2013)61–69。

我的翻译原文

ORIGINAL PAPEREggshell crack detection based on acoustic response and support vector data description algorithmHao Lin ÆJie-wen Zhao ÆQuan-sheng Chen ÆJian-rong Cai ÆPing ZhouReceived:21May 2009/Revised:27August 2009/Accepted:28August 2009/Published online:22September 2009ÓSpringer-Verlag 2009Abstract A system based on acoustic resonance and combined with pattern recognition was attempted to dis-criminate cracks in eggshell.Support vector data descrip-tion (SVDD)was employed to solve the classiﬁcation problem due to the imbalanced number of training samples.The frequency band was between 1,000and 8,000Hz.Recursive least squares adaptive ﬁlter was used to process the response signal.Signal-to-noise ratio of acoustic impulse response was remarkably enhanced.Five charac-teristics descriptors were extracted from response fre-quency signals,and some parameters were optimized in building model.Experiment results showed that in the same condition SVDD got better performance than con-ventional classiﬁcation methods.The performance of SVDD model was achieved with crack detection level of 90%and a false rejection level of 10%in the prediction set.Based on the results,it can be concluded that the acoustic resonance system combined with SVDD has signiﬁcant potential in the detection of cracked eggs.Keywords Eggshell ÁCrack ÁDetection ÁAcoustic resonance ÁSupport vector data descriptionIntroductionIn the egg industry,the presence of cracks in eggshells is one of the main defects of physical quality.Cracked eggsare very vulnerable to bacterial infections leading to health hazards [1].It mostly results in signiﬁcant economic loss in the egg industry.Recent research shows that it is possible to detect cracks in eggshells using acoustic response analysis [2–5].Supervised pattern recognition models were also employed to discriminate intact and cracked eggs [6].In these previous researches,training of discrimination models needs a considerable amount of intact egg samples and also corresponding defective ones.However,it is more difﬁcult to acquire sufﬁcient naturally cracked eggs samples than intact ones.Artiﬁcial inﬂiction of cracking in eggs is time-consuming and a waste.Moreover,the artiﬁcially cracked eggs may not provide completely authentic information on naturally cracked ones.So,the traditional discrimination model shows poor performance when the numbers of sam-ples from the two classes are seriously unbalanced,because the samples of minority group cannot provide sufﬁcient information to support the ultimate decision function.Support vector data description (SVDD),which is inspired by the theory of two-class support vector machine (SVM),is custom-tailored for one-class classiﬁcation [7].One-class classiﬁcation is always used to deal with a two-class classiﬁcation problem,where each of the two classes has a special meaning [8].The two classes in SVDD are target class and outlier class,respectively.Target class is assumed to be sampled well,and many (training)example objects are available.The outlier class can be sampled very sparsely,or can be totally absent.The basic idea of SVDD is to deﬁne a boundary around samples of target with a volume as small as possible [9].SVDD has been used to solve the problem of unbalanced samples in the ﬁeld of machine faults diagnosis,intrusion detection in the network,recog-nition of handwritten digits,face recognition,etc.[10–13].In this work,the algorithm of SVDD was employed to solve the classiﬁcation problem of eggs due to imbalancedH.Lin ÁJ.Zhao (&)ÁQ.Chen ÁJ.Cai ÁP.ZhouSchool of Food and Biological Engineering,Jiangsu University,212013Zhenjiang,People’s Republic of Chinae-mail:zjw-205@;zhao_jiewen@ H.Line-mail:linhaolt794@Eur Food Res Technol (2009)230:95–100DOI 10.1007/s00217-009-1145-6number of samples.In addition,recursive least squares (RLS)adaptive ﬁlter was used to enhance the signal-to-noise ratio.Some excitation resonant frequency charac-teristics of signals were used as input vectors of SVDD model to discriminate intact and cracked eggs.Materials and methods Samples preparationAll barn egg samples were collected naturally from a poultry farm and they were intensively reared.These eggs were on maximum 3days old when they were measured.As much as 130eggs with intact shells and 30eggs with cracks were measured.The sizes of eggs ranged from peewee to jumbo.Irregular eggs were not incorporated into the data analysis.The cracks,which were 10–40mm long and less than 15-l m wide,were measured by a micrometer.Both,intact and cracked samples,were divided into two subsets.One of them called calibration set was used to build a model,and the other one called prediction set was used to test the robustness of the model.The calibration set contained 120samples;the number of intact and cracked samples were 110and 10,respectively.The remaining 40samples constituted the prediction set,with 20intact eggs and 20cracked ones.Experimental systemA system based on acoustic resonance was developed for the detection of crack in eggshell.The system consists of a product support,a light exciting mechanism,a microphone,signal ampliﬁers,a personal computer (PC)and software to acquire and analyze the results.A schematic diagram of the system is presented in Fig.1.A pair of rolls made of hard rubber was used to support the eggs,and the shape of the support was focused to normal eggshell surfaces.The excitation set included an electromagnetic driver,an adjustable volt DC power and a light metallic stick.The total mass of the stick was 6g,and its length 6cm.The excitation force is an important factor that affects the magnitude and width of the pulse.The adjustable volt DC power was used to control the excitation force.Based on previous test,the voltage of excitation was set at 30V.In this case,optimal signals were achieved without instrumentation overload.The impacting position was close to the crack in the cracked eggshells,which was placed randomly among intact eggshells.Data acquisition and analysisResponse signals obtained from the microphone were ampliﬁed,ﬁltered and captured by a 16-bit data acquisition card.The program of data acquisition was compiled based on LabVIEW8.2software(National Instruments,USA)that allows a fast acquisition and processing of the response signal.The sampling rate was 22.05kHz.The time signal was transformed to a frequency signal by using a 512-point fast Fourier (FFT)transformation.The linear frequency spectrum accepted was transformed to a power spectrum.A band-pass ﬁlter was used to preserve the information of the frequency band between 1,000and 8,000Hz,because the features of response signals were legible in this frequency band and the signal-to-noise here was also favorable.Brief introduction of support vector data description (SVDD)SVDD is inspired by the idea of SVM [14,15].It is a method of data domain description also calledone-classFig.1Eggshell crackmeasurement system based on acoustic resonance analysisclassiﬁcation.The basic idea of SVDD is to envelop samples or objects within a high-dimensional space with the volume as small as possible byﬁtting a hypersphere around the samples.The sketch map in two dimensions of SVDD is shown in Fig.2.By introducing kernels,this inﬂexible model becomes much more powerful and can give reliable results when a suitable kernel is used[16]. The problem of SVDD is toﬁnd center a and radius R, which have the minimum volume of hypersphere contain-ing all samples X i.For a data set containing i normal data objects,when one or a few very remote objects are in it,a very large sphere is obtained,which will not represent the data very well.Therefore,we allow for some data points outside the sphere and introduce slack variable n i.As a result,the minimization problem can be denoted in thefollowing form:min LðRÞ¼R2þCX Ni¼1n i;s:t x iÀak k2R2þn i;n i!0ði¼1;2;...;NÞ;9>>>>>=>>>>>;ð1Þwhere the variable C gives the trade-off between simplicity (volume of the sphere)and the number of errors(number of target objects rejected).The above problem is usually solved by introducing Lagrange multipliers and can be transformed into maximizing the following function L with respect to the Lagrange multipliers.For an object x,we deﬁnef2ðxÞ¼xÀak k2¼ðxÁxÞÀ2X Ni¼1a iðzÁx iÞþX Ni¼1X Nj¼1a i a jðx iÁx jÞ:ð2ÞThe test objects x is accepted when the distance is smaller than the radius.These objects are called the support objects of the description or the SVs.Objects lying outside the sphere are also called bounded support vectors(BSVs). When a sphere is not always a goodﬁt for the boundary of data distribution,the inner product(x,y)is generalized by a kernel function k x;yðÞ¼/xðÞ;/yðÞf g;where a mapping/ of the data to a new feature space is applied.With such mapping,Eq.(2)will then becomeL¼P Ni¼1a i kðx i;x iÞÀP Ni¼1P Nj¼1a i a j kðx i;x jÞ;s:t0a i C;P Ni¼1a i¼1and a¼Pia i/ðx iÞ:9>>>=>>>;ð3ÞIn brief,SVDDﬁrst maps the data which are not linearly separable into a high-dimensional feature space and then describe the data by the maximal margin hypersphere.SoftwareAll data-processing algorithms were implemented with the statistical software Matlab7.1(Mathworks,USA)under Windows XP.SVDD Matlab codes were downloaded from http://www-ict.ewi.tudelft.nl/*davidt/dd_tools.html free of charge.Result and discussionResponse signalsSince the acoustic response was an instantaneous impulse, it was difﬁcult to discriminate between the different response signals of cracked and intact eggs in the time domain.The time domain signals were transformed by FFT to frequency domain signals for the next analysis.Typical power spectra of intact egg and cracked egg are shown in Fig.3,and the areas under the spectral envelope for the intact eggs were smaller than that of the cracked eggs.For the intact eggs,the peak frequencies were prominent, generally found in the middle place(3,500–5,000Hz).In contrast,the peak frequencies of cracked eggs were dis-perse and not prominent.Adaptive RLSﬁlteringSince the detection of cracked eggshells is based on acoustic response measurement,it is vulnerably interfered by the surrounding noise.This fact is reinforced by the much damped behaviors of agro-products[17].Therefore, response signal should be processed to remove noise in further analysis.Adaptive interference canceling is a standard approach to remove environmental noise[18,19].The RLS is a popular algorithm in theﬁeld of adaptive signal processing. In adaptive RLSﬁltering,the coefﬁcients are adjusted from sample to sample to minimize the mean square error(MSE) between a measured noisy scalar signal and itsmodeledvalue from the ﬁlter [20,21].A scalar,real output signal,y k ,is measured at the discrete time k ,in response to a set of scalar input signals X k ði Þ;i ¼1;2;...;n ;where n is an arbitrary number of ﬁlter taps.For this research,n is set to the number of degrees of freedom to ensure conformity of the resulting ﬁlter matrices.The input and the output sig-nals are related by the simple regression model:y k ¼X n À1i ¼0w ði ÞÁx k ði Þþe k :ð4Þwhere e k represents measurement error and w (i )represents the proportion that is contained in the primary scalar signal y k .The implementation of the RLS algorithm is optimized by exploiting the inversion matrix lemma and provides fast convergence and small error rates [22].System identiﬁcation of a 32-coefﬁcient FIR ﬁlter combined with adaptive RLS ﬁltering was used to process the signals.The forgetting factor was 1,and the vector of initial ﬁlter coefﬁcients was 0.Figure 4shows the fre-quency signals before and after adaptive RLS ﬁltering.Variable selectionBased on the differences of frequency domain response signals from intact and cracked eggs,ﬁve characteristic descriptors were extracted from the response frequency signals as the inputs of the discrimination model.These are shown in Table 1.Parameter optimization in SVDD modelThe basic concept of SVDD is to map nonlinearly the original data X into a higher-dimensional feature space.The transformation into a higher-dimensional space is implemented by a kernel function [23].So,selection of kernel function has a high inﬂuence on the performance of the SVDD model.Several kernel functions have been proposed for the SVDD classiﬁer.Not all kernel functions are equally useful for the SVDD.It has been demonstrated that Gaussian kernel results in tighter description and gives a good performance under general smoothness assumptions [24].Thus,Gaussian kernel was adopted in this study.To obtain a good performance,the regularization parameter C and the kernel function r have to be opti-mized.Parameter C determines the trade-off between minimizing the training error and minimizing model complexity.By using Gaussian kernel,the data description transforms from a solid hyper-sphere to a Parzen density estimator.An appropriate selection with width parameter r of Gaussian kernel is important to the density estimation of target objects.There is no systematic methodology for the optimization of these parameters.In this study,the procedure of opti-mization was carried out in two search steps.First,a comparatively large step length was attempted to search optimal value of parameters.The favorable results of the model were found with values of C between 0.005and 0.1,and values of r between 10and 500.Therefore,a much smaller step length was employed for further searching these parameters.In the second search step,50parameter r values with the step of 10(r =10,20–500)and 20parameter C values with the step of 0.005(C =0.005,0.01–1)were tested simultaneously in the building model.Identiﬁcation results of SVDD model inﬂuenced by values of r and C are shown in Fig.5.The optimal model was achieved when r was equal to 420and C was equal to 0.085or 0.09.Here,the identiﬁcation rates of intactandFig.3Typical response frequency signal ofeggsFig.4Frequency signals before and after adaptive RLS ﬁlteringcracked eggs were both 90%in the prediction set.Fur-thermore,it was found that the performance of the SVDD model could not be improved by smaller search parison of discrimination modelsConventional two-class linear discrimination analysis (LDA)model and SVM model were used comparatively to classify intact and cracked eggs.Gaussian kernel was recommended as the kernel function of the SVM model.Parameters of SVM model were also optimized as in SVDD.Table 2shows the optimal results from three dis-crimination models in the prediction set.Identiﬁcation rates of intact eggs were both 100%in the LDA and SVM models,but 50and 35%for cracked eggs,respectively.In other words,at least 50%of cracked eggs could not be identiﬁed in conventional discrimination model.However,detection of cracked eggs is the task we focus on.The identiﬁcation rates of intact and cracked eggs were both 90%in the SVDD pared with conventional two-class discrimination models,SVDD model showed its superior performance in the discrimination of cracked eggs.LDA is a linear and parametric method with discrimi-nating character.In terms of a set of discriminant functions,the classiﬁer is said to assign an unknown example X to thecorresponding class [25].In the case of conventional LDA classiﬁcation,the ultimate decision function is based on sufﬁcient information support from two-class training samples.In general,such classiﬁcation does not pay enough attention to the samples in minority class in building model.It is possible to obtain an inaccurate estimation of the centroid between the two classes.Conventional LDA clas-siﬁcation always poorly describes the speciﬁc class with scarce training samples.Therefore,it is often unpractical to solve the classiﬁcation problem using tradition LDA clas-siﬁer,in case of imbalanced number in training samples.The basic concept of SVM is to map the original data X into a higher-dimensional feature space and ﬁnd the ‘optimal’hyperplane boundary to separate the two classes [26].In SVM classiﬁcation,the ‘optimal’boundary is deﬁned as the most distant hyperplane from both sets,which is also called the ‘middle point’between the clas-siﬁcation sets.This boundary is expected to be the optimal classiﬁcation of the sets,since it is the best isolated from the two sets [27].The margin is the minimal distance from the separating hyperplane to the closest data points [28].In general,when the information support from both positive and negative training sets are sufﬁcient and equal,an appropriate separating hyperplane can be obtained.How-ever,when the samples from one class are insufﬁcient to support the separating hyperplane,it will result in the hyperplane being excessively close to this class.As a result,most of the unknown sets may be recognized as the other class.Therefore,compared with other discrimination models,SVM showed poorest performance in discrimi-nating cracked eggs.Differing from conventional classiﬁcation-based app-roach,SVDD is an approach for one-class classiﬁcation.ItTable 1Frequencycharacteristics selection and expressionSome Low frequency band:1,000–3,720Hz,Middlefrequency band:3,720–7,440HzVariables Resonance frequency characteristics Expression X1Value of the area of amplitudeX 1¼P512i ¼0PiX2Value of the standard deviation of amplitude X 2¼ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃðPi ÀP Þq=nX3Value of the frequency band of maximum amplitude X 3¼Index max ðPi ÞX4Mean of top three frequency amplitude values X 4¼Max 1:3ðPi Þ=3X5Ratio of amplitude values of middle frequency bands to low frequency bandX 5¼P 200i ¼1Pi P 400i ¼201Pi200Fig.5Identiﬁcation rates of SVDD models with different values ofparameter r and CTable 2Comparison of results from three discrimination models ModelIdentiﬁcation rates in the prediction set (%)Intact eggsCracked eggs LDA 10050SVM 10035SVDD9090focuses mainly on normal or target objects.SVDD can handle cases with only a few outlier objects.The advantage of SVDD is that the target class can be any one of two training classes.The selection of the target class depends on the reliability of the information provided from training samples.In general,the class containing more samples may provide sufﬁcient information,and it can be selected as target class[29].Furthermore,SVDD can adapt to the real shape of samples andﬁndﬂexible boundary with a mini-mum volume by introducing kernel function.The boundary is described by a few training objects,the support vectors. It is possible to replace normal inner products with kernel functions and obtain moreﬂexible data descriptions[30]. Width parameter r can be set to give the desired number of support vectors.In addition,extra data on the form of outlier objects can be helpful to improve the performance of the SVDD model.ConclusionsDetection of crack in eggshell based on acoustic impulse resonance was attempted in this work.The SVDD method was employed for solving classiﬁcation problem where the samples of cracked eggs were not sufﬁcient.The results indicated that detection of crack in eggshell based on the acoustic impulse resonance was feasible,and the SVDD model showed its superior performance in contrast to conventional two-class discrimination models.It can be concluded that SVDD is an excellent method of classiﬁ-cation problem with imbalanced numbers.It is a promising method that uses acoustic resonance technique combined with SVDD to detect cracked eggs.Some relative ideas would be attempted for further improvement of the per-formance of SVDD model in our future work,such as follows:(1)introduce new kernel functions,which can help to obtain a moreﬂexible boundary;(2)try more methods for selection of parameters to obtain the optimal ones,since parameters of kernel functions are closely related to the tightness of the constructed boundary and the target rejection rate,and appropriate parameters are important to improve the performance of SVDD models;(3)investigate the contribution of abnormal targets to the calibration model and develop a robust model,which has an excellent ability to deal with abnormal targets.Acknowledgments This work is a part of the National Key Tech-nology R&D Program of China(Grant No.2006BAD11A12).We are grateful to the Web site http://www-ict.ewi.tudelft.nl/*davidt/ dd_tools.html,where we downloaded SVDD Matlab codes free of charge.References1.Lin J,Puri VM,Anantheswaran RC(1995)Trans ASAE38(6):1769–17762.Cho HK,Choi WK,Paek JK(2000)Trans ASAE43(6):1921–19263.De Ketelaere B,Coucke P,De Baerdemaeker J(2000)J Agr EngRes76:157–1634.Coucke P,De Ketelaere B,De Baerdemaeker J(2003)J SoundVib266:711–7215.Wang J,Jiang RS(2005)Eur Food Res Technol221:214–2206.Jindal VK,Sritham E(2003)ASAE Annual International Meet-ing,USA7.Tax DMJ,Duin RPW(1999)Pattern Recognit Lett20:1191–11998.Pan Y,Chen J,Guo L(2009)Mech Syst Signal Process23:669–6819.Lee SW,Park JY,Lee SW(2006)Patten Recognit39:1809–181210.Podsiadlo P,Stachowiak GW(2006)Tribol Int39:1624–163311.Sanchez-Hernandeza C,Boyd DS,Foody GM(2007)Ecol Inf2:83–8812.Liu YH,Lin SH,Hsueh YL,Lee MJ(2009)Expert Syst Appl36:1978–199813.Cho HW(2009)Expert Syst Appl36:434–44114.Tax DMJ,Duin RPW(2001)J Mach Learn Res2:155–17315.Tax DMJ,Duin RPW(2004)Mach Learn54:45–6616.Guo SM,Chen LC,Tsai JHS(2009)Pattern Recognit42:77–8317.De Ketelaere B,Maertens K,De Baerdemaeker J(2004)MathComput Simul65:59–6718.Adall T,Ardalan SH(1999)Comput Elect Eng25:1–1619.Madsen AH(2000)Signal Process80:1489–150020.Chase JG,Begoc V,Barroso LR(2005)Comput Struct83:639–64721.Wang X,Feng GZ(2009)Signal Process89:181–18622.Djigan VI(2006)Signal Process86:776–79123.Bu HG,Wang J,Huang XB(2009)Eng Appl Artif Intell22:224–23524.Tao Q,Wu GW,Wang J(2005)Pattern Recognit38:1071–107725.Xie JS,Qiu ZD(2007)Pattern Recognit40:557–56226.Devos O,Ruckebusch C,Durand A,Duponchel L,Huvenne JP(2009)Chemom Intell Lab Syst96:27–3327.Liu X,Lu WC,Jin SL,Li YW,Chen NY(2006)Chemom IntellLab Syst82:8–1428.Chen QS,Zhao JW,Fang CH,Wang DM(2007)SpectrochimActa Pt A Mol Biomol Spectrosc66:568–57429.Huang WL,Jiao LC(2008)Prog Nat Sci18:455–46130.Foody GM,Mathur A,Sanchez-Hernandez C,Boyd DS(2006)Remote Sens Environ104:1–14。

Rapid, highly efficient extraction and purification of membrane proteins using a microfluidic

Journal of Chromatography A,1218 (2011) 171–177Contents lists available at ScienceDirectJournal of ChromatographyAj o u r n a l h o m e p a g e :w w w.e l s e v i e r.c o m /l o c a t e /c h r o maRapid,highly efﬁcient extraction and puriﬁcation of membrane proteins using a microﬂuidic continuous-ﬂow based aqueous two-phase systemRui Hu a ,Xiaojun Feng a ,Pu Chen a ,Meng Fu b ,Hong Chen c ,Lin Guo b ,Bi-Feng Liu a ,∗aBritton Chance Center for Biomedical Photonics at Wuhan National Laboratory for Optoelectronics –Hubei Bioinformatics &Molecular Imaging Key Laboratory,Department of Systems Biology,College of Life Science and Technology,Huazhong University of Science and Technology,Wuhan 430074,China bCollege of Life Science and Technology,Wuhan University,Wuhan 430072,China cKey Laboratory of Oil Crops Biology of the Ministry of Agriculture,Oil Crops Research Institute,Chinese Academy of Agricultural Sciences,Wuhan 430062,Chinaa r t i c l e i n f o Article history:Received 28July 2010Received in revised form 22October 2010Accepted 25October 2010Available online 30 October 2010Keywords:Microﬂuidic chipAqueous two-phase system Membrane proteins Puriﬁcationa b s t r a c tMembrane proteins play essential roles in regulating various fundamental cellular functions.To investigate membrane proteins,extraction and puriﬁcation are usually prerequisite steps.Here,we demonstrated a microﬂuidic aqueous PEG/detergent two-phase system for the puriﬁcation of mem-brane proteins from crude cell extract,which replaced the conventional discontinuous agitation method with continuous extraction in laminar ﬂows,resulting in signiﬁcantly increased extraction speed and efﬁciency.To evaluate this system,different separation and detection methods were used to identify the puriﬁed proteins,such as capillary electrophoresis,SDS-PAGE and nano-HPLC–MS/MS.Swiss-Prot database with Mascot search engine was used to search for membrane proteins from random selected bands of SDS-PAGE.Results indicated that efﬁcient puriﬁcation of membrane proteins can be achieved within 5–7s and approximately 90%of the puriﬁed proteins were membrane proteins (the highest extraction efﬁciency reported up to date),including membrane-associated proteins and integral mem-brane proteins with multiple transmembrane pared to conventional approaches,this new method had advantages of greater speciﬁc surface area,minimal emulsiﬁcation,reduced sample con-sumption and analysis time.We expect the developed method to be potentially useful in membrane protein puriﬁcations,facilitating the investigation of membrane proteomics.© 2010 Elsevier B.V. All rights reserved.1.IntroductionMembrane proteins constitute approximately 30%of the pro-teome [1],playing essential roles in regulating various fundamental cellular functions,such as cell recognition,selective transportation of metabolites and receptor-mediated signal transduction [2,3].In addition,more than half of the known membrane proteins are pre-dicted to be pharmacological targets [4].However,researches in membrane proteins are relatively hampered since most membrane proteins are of natural low abundance.Thus,extraction and puriﬁ-cation of membrane proteins is usually a prerequisite step in such investigations.Puriﬁcation of membrane proteins has proven to be a chal-lenging task due to their hydrophobic nature as complexes of proteins and lipids [5].Consequently,solubilization of membrane proteins by detergents is necessary to separate them from crude cell extract.Currently,detergent/polymer aqueous two-phase system (ATPS)is a common approach for membrane protein enrichment∗Corresponding author.Tel.:+862787792203;fax:+862787792170.E-mail addresses:bﬂiu@ ,bifeng liu@ (B.-F.Liu).without denaturation [6–8].ATPS involves the use of two aque-ous phases to extract target molecules by vigorous agitation.Although it has been widely adopted in laboratories,the sepa-ration efﬁciency of ATPS still requires improvement.In addition,emulsiﬁcation during agitation can also elongate the separation time [9].The emerging microﬂuidic technology has provided an oppor-tunity for the integration and miniaturization of existing biological tools to address issues like speed,throughput and sample cost [10–12].Previously,Kitamori and co-workers reported a microﬂu-idic liquid–liquid extraction system,which was applied to the isolation of metal ions based on multi-phase laminar ﬂows.Extrac-tion of different metal ions was successfully realized,including Fe 2+,Co 2+,Ni 2+,K +,Na +and Al 3+[13–15].The extraction effect of laminar ﬂows in microchannels was equivalent to that of vigorous agita-tion.However,the microﬂuidic-based method had advantages of enhanced extraction speed,simpliﬁed operation and potential for miniaturization.Recently,Meagher et al.developed a microﬂuidic aqueous two-phase system (␮ATPS)for isolating speciﬁc proteins from sub-microliter volumes of Escherichia coli cell lysate [16].In this method,PEG-salt two-phase system was realized in a Y-shaped microﬂuidic channel for continuous extraction of target proteins0021-9673/$–see front matter © 2010 Elsevier B.V. All rights reserved.doi:10.1016/j.chroma.2010.10.090172R.Hu et al./J.Chromatogr.A1218 (2011) 171–177Fig.1.Microchip fabrication and system setup.(A)Schematic of the microﬂuidic chip design.(B)Size comparison of the fabricated PDMS microchip with a U.S.one cent coin.(C)Schematic of the system setup for ␮ATPS and image acquisition.into the PEG phase with reduced separation time,enhanced speed and throughput.Compared to the traditional batch techniques with agitation,␮ATPS has distinctive advantages of fast extraction rate,high separation efﬁciency and sample enrichment [17].In traditional methods,two liquid phases are highly scattered by vigorous agi-tation to maximize the speciﬁc interface area between the two phases and to improve the extraction rate and efﬁciency.With microﬂuidic based method,the laminar ﬂow in microchannels usu-ally results in greater speciﬁc interface area,avoiding the use of agitation and preventing the occurrence of emulsiﬁcation.In addi-tion,liquid–liquid extraction is typically a low-throughput batch technique in laboratory,but it is well-suited for continuous oper-ation on microﬂuidic chips [18,19].Therefore,we believe that the ␮ATPS method is potentially useful in the puriﬁcation of membrane proteins [20,21].Here,we demonstrated the use of a PEG/detergent ␮ATPS sys-tem for the puriﬁcation of membrane proteins from crude cell extract.Our ␮ATPS system combined the use of the zwitterionic detergent Zwittergent 3-10,sodium dodecyl sulfate (SDS)and the nonionic detergent Triton X-114,resulting in a complementary solubilization of proteins [22,23].The PEG/detergent two-phase system partitioning allowed successful removal of soluble pro-teins.Integral and peripheral membrane proteins remained in the detergent phase,while soluble proteins were found in the PEG-rich phase.Extraction of FITC-labeled IgG from detergent to PEG phase was ﬁrst conducted to evaluate the developed ␮ATPS method.Capillary electrophoresis of the puriﬁed samples suggested efﬁ-cient puriﬁcation of IgG within 5–7s.We further applied the ␮ATPS method to the puriﬁcation of membrane proteins from HeLa cell extracts.Results indicated that 90%of the extracted proteins are membrane proteins,including membrane-associated proteins and integral membrane proteins with multiple transmembrane domains,which represented one of the highest extraction efﬁ-ciency among existing approaches.2.Experimental2.1.Chemicals and reagentsTris (hydroxymethyl)aminomethane (Tris),HCl,NaOH,KCl,NH 4HCO 3,ACN,NaCl,NaHCO 3,KH 2PO 4,Na 2HPO 4·12H 2O,formic acid,ethylene diamine tetra acetic acid (EDTA)were purchased from Tianjing Chemical Co.Ltd.(Tianjing,China).N,N -methylene Bisacrylamide,Coomassie Brilliant Blue G250,zwitterionic deter-gent Zwittergent 3-10were purchased from Fluka (MO,USA).Dithiothreitol (DTT),iodoacetamide (IAA),acrylamide,glycerol,bromophenol blue,␤-mercaptoethanol,polyacrylamide,glycine,polyethylene glycol #6000(PEG 6000),Trypsin (proteomics sequencing grade)were purchased from Amresco (OH,USA).N,N,N ,N -tetramethylethylenediamine (TEMED),ammonium per-sulfate (AP),sodium dodecyl sulfate (SDS),Triton X-114were purchased from Sigma–Aldrich (MO,USA).DMEM were pur-chased from GIBCO (Invitrogen corporation,USA).Membrane Protein Extraction Kit was purchased from XinHan (Shanghai,China).All reagents were of analytical grade unless speciﬁed otherwise.Water was puriﬁed by the Millipore-Q system (Mil-lipore,USA)before use for the preparation of all solutions.Samples and all buffer solutions were autoclaved (121◦C;0.12MPa)and ﬁltered (0.45␮m microporous membrane ﬁltration)before experiments.For chip experiments,the PEG-rich inlet stream was 35wt%PEG,and the detergent-rich inlet stream was pre-pared with a volume ratio of 9:5:1(20%(w/w)Zwittergent 3-10:100%(w/w)Triton X-114:100mM SDS),resulting in a pH of approximately 7.4.2.2.Chip design and fabricationWe designed the PDMS microchip with serpentine microchan-nels as shown in Fig.1A.The widths of the inlet channel a–o,b–o,c–o and the outlet channel p–e,p–f are 80␮m.The width of the outlet collection channel p–d is 40␮m.The separation channel o–p has a width of 180␮m and a total length of approximately 140mm.All channel depths are 50␮m (Fig.2C).We fabricated the microchip using previously reported protocols [24,25].Fabri-cated PDMS structures are then irreversibly bonded to a planar glass substrate (76mm ×26mm ×1mm)to form the ﬁnal device.A comparison of the microchip with a US one-cent coin is given in Fig.1B.Micro-syringe pumps are used to control the ﬂuid ﬂow in the microchannels.Typical injection speed was 0.8–1.2␮L min −1of the PEG-rich inlet,and 3.5–5.0␮L min −1of the detergent-rich inlet.2.3.Image acquisitionExperiments were conducted on an inverted ﬂuorescence microscope (IX 71,Olympus,Japan).A mercury lamp was used as the excitation source.For FITC and FQ,the light emitted from the mercury lamp was ﬁltered by a 460–490nm band-pass ﬁlter,reﬂected by a 505nm dichroic mirror,and then focused on the microchannel by a 10×objective (NA 0.7)as illustrated in Fig.1C.During experiments,ﬂuorescence images of the each local channel were collected through the same objective with a 510nm high-passR.Hu et al./J.Chromatogr.A 1218 (2011) 171–177173Fig.2.Extraction of membrane proteins by ␮ATPS.(A)Schematic of the ␮ATPS extraction mechanism.Side streams,PEG-rich phase;middle stream,crude membrane protein extract dissolved in detergent.Black arrows indicate the direction of ﬂow.Hollow arrows indicate the direction of membrane protein migration.(B)Channel width dependence of speciﬁc interface area and diffusion time in microchip.S ,speciﬁc interface area;V ,volume;W ,diffusion distance;t ,diffusion time;D ,diffusion coefﬁcient.Dotted line indicates the microchannel width of our device (180␮m).(C)Proﬁle of extraction channel.L ,length;W ,width;H ,height.ﬁlter and monitored by a CCD camera (CoolSNAP cf2,Photometrics)with 200ms exposure.2.4.Cell cultureHeLa cells were grown in DMEM (Invitrogen Corporation,GIBCO,CA)supplemented with 10%NCS (Invitrogen Corporation,GIBCO,CA)and maintained in standard culture conditions (37◦C,95%humidiﬁed air,and 5%CO 2).Cells were allowed to grow toa density of 80%and then were harvested using sterile PBS/EDTA (pH 7.4)before experiments.As they spread out across the cell cul-ture dish,when two adjacent cells touch,this signals them to stop growing,loss of contact inhibition is a classic sign of oncogenic cells.2.5.Crude membrane protein preparationCrude membrane protein was prepared by using extraction kit.Brieﬂy,HeLa cells ((5–10)×107)were collected bycentrifugationFig.3.Evaluation of the ␮ATPS method.(A)Fluorescence images of the extraction of FITC labeled IgG from the detergent phase to the PEG-rich phase.Microchannel outline is indicated by dotted lines.(B)Schematic of the microﬂuidic chip design.a–c,inlets;e–f,outlets.The red rectangles 1–4indicate the locations where the four ﬂuorescence images in (A)were captured.(C)Capillary electrophoresis of standard FITC-IgG with concentrations of 0.01mg mL −1,0.05mg mL −1and 0.1mg mL −1.(D)Capillary electrophoresis of the FITC-IgG in the solutions before and after ␮ATPS puriﬁcation.(For interpretation of the references to color in this ﬁgure legend,the reader is referred to the web version of the article.)174R.Hu et al./J.Chromatogr.A1218 (2011) 171–177at1000×g for5min at4◦C,then were resuspended in1mL of homogenate buffer with protease inhibitor cocktail in an ice-cold homogenizer and were homogenized on ice for30–50times after been washed once with1mL of ice cold PBS.The homogenate was centrifuged in700×g for10min at4◦C.The supernatant was trans-ferred to a new vial and centrifuged at10,000×g for30min at4◦C, cytosol fraction in the supernatant was collected.The pellet con-tained proteins from both plasma membrane and cellular organelle membrane.It was resuspended in200␮L lysis buffer with protease inhibitor cocktail for30min at4◦C,and centrifuged at10,000×g for 30min at4◦C,membrane protein in the supernatant was collected.2.6.Capillary electrophoresisCapillary electrophoresis–laser-inducedﬂuorescence detec-tion platform,including capillary,excitation and emission light path,high voltage power supply,optical signal conversion and collection,ampliﬁcation,ﬁltering devices,signal acquisition pro-grams and hardware,and other related components.The main components include Olympus IX71research-oriented inverted ﬂuorescence microscope excitation,emission-axis optical system; high-voltage power supply(Institute of Nuclear Research,Shang-hai):the maximum voltage of30,000V,maximum current of 0.3mA;electrode diameter of0.6mm of P99platinum wire;argon ion laser(maximum power42mW)as aﬂuorescence excitation light source;PMT photomultiplier tube and associated circuitry for ﬂuorescence signal collection and ampliﬁcation;hardware signal ﬁlter(Stanford Research System Model SR570);NI6035data acqui-sition card and LabView acquisition program.Uncoated fused-silica capillaries(i.d.75␮m,o.d.375␮m)were purchased from Yongnian Optical Fiber Factory(Hebei Province,China)and treated following the protocol introduced by Hjérten[26].2.7.SDS-PAGE and nano-HPLC–MS/MSA200-␮g sample of PM protein was separated by SDS-PAGE on a5%stacking gel and a12%separation gel run according to standard laboratory procedures.After electrophoresis,the gels were stained with Coomassie Brilliant Blue G250.In-gel digestion was performed as described previously with slight modiﬁcation[27].In brief,random selected bands of SDS-PAGE gel was cut into many1mm3gel slices,each of which contained a stained protein(s).The resulting slices were washed with100mM NH4HCO3containing50%ACN(pH8.0)for3times,till the dye(Coomassie Brilliant Blue)was completely removed.After being dried in a SpeedVac concentrator,the gel-bound proteins were reduced in10mM DTT/50mM NH4HCO3(pH8.0)and then alkylated in55mM iodoacetamide/50mM NH4HCO3(pH8.0)for 30min in a darkroom.The gel pieces were then washed with10mM NH4HCO3,and again dehydrated with ACN and dried in a SpeedVac. The dry gel pieces were reswollen with35␮L of10mM NH4HCO3 containing0.5␮g of promega trypsin.Digestion was carried out at37◦C overnight.The peptides were extracted two times with 40␮L of60%ACN/5%formic acid by sonication/ultrasonic oscilla-tion/sonic oscillation for10min,then centrifuge for2min to collect the supernatant.The combined extracts were evaporated to about 2␮L in a SpeedVac and stored at−80◦C.The digested peptides were injected into a nanoLC system (Eksigent)andﬁrst desalted and preconcentrated on a CapTrap (0.5mm i.d.,2mm long;MICHROM)precolumn.The peptides were then eluted onto a C18column(100␮m i.d.,15cm long; MICHROM)coupled to a quadrupole time-of-ﬂight(Q-TOF)hybrid mass spectrometer(QSTAR ELITE,Applied Biosystems)equipped with a MicroIonSpray ESI source.The gradient proﬁle consisted of a linear gradient from5%to40%B(0.1%formic acid/1.9% H2O/98%acetonitrile,v/v)over40min into A(0.1%formic acid/1.9%acetonitrile/98%H2O,v/v),followed by10min of85%B and then 15min reconditioning with95%A.Theﬂow rate was300nL min−1. The peptides were detected in the positive ion MS mode or the data-dependent MS/MS mode.The data-dependent mode was used for survey scans(m/z400–1800)in order to choose up toﬁve most intense precursor ions.For collision-induced dissociation(CID) mass spectrometric(MS/MS)analysis,collision energies were cho-sen automatically as a function of m/z and charge.The collision gas was nitrogen.The temperature of a heated interface was150◦C and the electrospray voltage was2000V.3.Results and discussion3.1.Puriﬁcation by ATPSWe demonstrate a continuousﬂow PEG/detergent␮ATPS.FITC labeled hydrophilic proteins migrated into the PEG phase,while hydrophobic plasma membrane protein from intact membrane protein complexes remained in the detergent phase.This opera-tion is illustrated schematically in Fig.2A.Sample stream of crude plasma membrane protein with unwanted(tagged)proteins are hydrodynamically focused between twoﬂowing streams contain-ing PEG.Theﬂow-rates of the liquid samples and extracting reagent were controlled by microsyringe pumps(model210,KD Scientiﬁc, Boston,MA)when precise control of the velocity of the sample stream in the mainﬂow channel was required.Each syringe nee-dle was connected to aﬁlter through a fused-silica capillary tube (GL Sciences,0.25mm i.d.0.5mm o.d.)using epoxy-based glue.TheFig. 4.One-dimensional SDS-PAGE of the prepared crude membrane protein extract.CM,crude membrane proteins;PM,puriﬁed membrane proteins.Molecular mass markers are shown on the right.The gel was stained with colloidal Coomassie brilliant blue and12bands were randomly selected for further LC–MS/MS analysis.R.Hu et al./J.Chromatogr.A1218 (2011) 171–177175 outlets were also connected to a fused-silica capillary tube with EPcentrifuge tube for collection in the same way.For experiments thatdemonstrate the operations,the velocity of plugs in the mainﬂowchannel was∼5␮L min−1.Typically,we found the lower theﬂowrate,the better the extraction efﬁciency.However,the laminarﬂowbecomes unstable at too low aﬂow rate due to the limitation of thesyringe pumps.Thus,we chose5␮L min−1as the optimalﬂow ratefor the extraction of hydrophobic membrane proteins.3.2.Theory for puriﬁcation by ATPSMicroﬂuidics has several characteristic features different frombulk scaleﬂuidﬂow,such as short diffusion distance,high spe-ciﬁc interface area,and small heat capacity.These characteristicsof microﬂuidic systems are essential keys to control chemical unitoperations,such as mixing,reaction,extraction and separation.Especially,to control molecular transport in microﬂuidic channels,the molecular transportation time and the speciﬁc interface areamust be considered[28].The molecular transportation time is givenby:t=W2D(1)where t,W,and D are the molecular transportation time,diffusion distance and coefﬁcient,respectively.The speciﬁc interface area, , can be expressed as:=SV∝1W(2)where S and V are the interface area and the volume,respectively.In our method,samples were injected into the separating chan-nel byﬂow focusing.For an ideal sandwich-type laminarﬂow with each stream occupying one third of the channel,the relationship between the microchannel width and the molecular transporta-tion time and the speciﬁc interface area is summarized in Fig.2B. Given the dimensions of our device(Fig.3C),the speciﬁc inter-face area is approximately167cm−1,which represents a dramatic increase compared to the conventional mechanical shaking method (1–10cm−1)and previously reported Y-shaped microﬂuidic sys-tems(80cm−1)[29,30].In consequence,a signiﬁcant decrease in the transportation time can be expected.For molecules with a diffu-sion coefﬁcient of10−9m2s−1,the transportation time is less than 5s.3.3.System evaluationExperiments were conducted withﬂuorescent tracer molecules to visualize the performance of the␮ATPS system.Fig.3A illus-trates theﬂuorescence images captured during experiments with locations indicated by the four rectangles shown in Fig.3B.As shown in Fig.3A,detergent phase containing water-soluble FITC-IgG was injected into the middle stream by hydrodynamic focusing. The FITC-IgG was continuously extracted from the detergent-rich stream into the two PEG-rich side plete extraction of FITC-IgG was observed at the end of the microchannel.Capillary electrophoresis was used to investigate the extraction efﬁciency of the developed␮ATPS system.Fig.3C illustrates a comparison of electropherogram of increasing concentrations of standard FITC-IgG solutions.To determine the recovery of extracted proteins, samples containing0.05mg mL−1FITC-IgG were quantitatively analyzed before and after␮ATPS extraction.As shown in Fig.3D, FITC-IgG only existed in the solutions collected from the two side streams(outlet e and f),consistent with the optical observations shown in Fig.3A.Quantitative analysis indicated a recovery of 90.8%.The loss of proteins could have resulted from the nonspeciﬁc adsorption of proteins on PDMS surfaces.Table1Categories of puriﬁed membrane proteins by␮ATPS.Identiﬁed membrane proteins Categories4F2cell-surface antigen heavy chain Plasma membrane Alkaline phosphatase,placental type Plasma membrane Alkaline phosphatase,tissue-nonspeciﬁc isozyme Plasma membrane Annexin A6Plasma membrane Antithrombin-III Plasma membrane Calcium-binding mitochondrial carrier proteinAralar2Plasma membrane Complement decay-accelerating factor Plasma membraneEzrin Plasma membrane Heterogeneous nuclear ribonucleoprotein M Plasma membrane Intestinal alkaline phosphatase Plasma membrane Junction plakoglobin Plasma membrane Lamin-A/C Plasma membrane Moesin Plasma membrane Olfactory receptor5AC2Plasma membrane Prostaglandin G/H synthase1Plasma membrane Scavenger receptor class B member1Plasma membrane Steryl-sulfatase Plasma membrane Transferrin receptor protein1Plasma membraneWD repeat and FYVE domain-containing protein3Plasma membrane78kDa glucose-regulated protein MembraneCarbamoyl-phosphate synthase[ammonia]MembraneCarnitine O-palmitoyltransferase2Membrane Cytoskeleton-associated protein4MembraneGlycerol-3-phosphate dehydrogenase MembraneGPI transamidase component PIG-S MembraneGPI transamidase component PIG-T MembraneLamin-B1MembraneNADH-ubiquinone oxidoreductase75kDa subunit MembraneNitric oxide synthase,brain MembraneSuccinate dehydrogenase[ubiquinone]ﬂavoprotein subunitMembraneTrifunctional enzyme subunit alpha MembraneAFG3-like protein2Integral to membrane Calnexin Integral to membrane Dolichyl-diphosphooligosaccharide–proteinglycosyltransferase subunit1Integral to membrane Heterogeneous nuclear ribonucleoprotein R Integral to membrane Mitochondrial import receptor subunit TOM70Integral to membrane Protein disulﬁde-isomerase A4endoplasmic reticulum RNA-binding protein FUS NucleusATPase family AAA domain-containing protein3A CytoplasmElongation factor1-alpha1Cytoplasm3.4.Extraction of membrane proteins and SDS-PAGEFor membrane proteins,it usually involves a crude mem-brane protein extraction procedure before further puriﬁcation and enrichment.Previously,Cao et al.reported the use of conventional aqueous two-phase agitation method for the puriﬁcation of crude membrane protein extracts,yielding the highest extraction efﬁ-ciency of67%[31].In this work,the developed␮ATPS system was used in combination with detergents for the puriﬁcation of crude membrane proteins.After extraction,puriﬁed membrane proteins were separated by one-dimensional SDS-PAGE(Fig.4).Twelve bands were selected for further MS analysis to verify the devel-oped method,similar to approaches reported previously[22,23]. Proteins from the selected bands of SDS-PAGE were digested by trypsin;the tryptic peptides were then extracted from each gel band and further separated by reversed-phase nanoLC,and then detected and sequenced with waters Q-TOF micro mass spec-trometer.All MS/MS samples were analyzed using Mascot(Matrix Science,London,UK;version Mascot).Mascot was set up to search the SwissProt57.7database(selected for Homo sapiens)assuming the digestion enzyme trypsin.Mascot was searched with a frag-ment ion mass tolerance of0.40Da and a parent ion tolerance of 200ppm.176R.Hu et al./J.Chromatogr.A1218 (2011) 171–177Fig.5.Classiﬁcation of the functional categories of the identiﬁed plasma membrane proteins in HeLa cells.(A)Subcellular localization of the identiﬁed proteins accord-ing to the GO annotation terms.(B)The functional categories of the characterized proteins.3.5.Identiﬁcation of membrane proteinsAfter being processed with analytical software Analyst QS2.0, samples were utilized to search the Swiss-Prot database with Mas-cot search engine for protein identiﬁcation.To assess the efﬁcacy of the developed protocol for the enrichment of integral mem-brane proteins and to estimate contamination by other cellular organelles,including mitochondria and endoplasmic reticulum (ER),we classiﬁed the40identiﬁed proteins according to the gene ontology(GO)annotation and other currently available data (Table1).Of the annotated proteins,36(90%)were previously assigned as integral membrane or membrane-associated proteins. The90%purity of membrane proteins represented one of the highest extraction efﬁciency among existing approaches.Of the reminder proteins with a subcellular annotation,10%were anno-tated as cytoplasmics,nucleus and ER,this group may include proteins that exist at more than one site in the cell.These data indicate that the contamination by mitochondria and ER in the membrane fraction and soluble non-target protein in cytoplasmics was greatly reduced by use of the microﬂuidic aqueous two-phase extraction process.In Fig.5A,of the36PM proteins,19(47.5%)were plasma membrane,12(30%)were membrane proteins,5(12.5%) were integral membrane proteins.We also categorized the iden-tiﬁed proteins according to their functions,except for5.56%of the protein function is not clear,based on universal GO annota-tion terms Fig.5B:2.94%have signal activity,5.88%of proteins have electric carrier activity,17.65%have catalytic activity,41.18% are involved in cellular binding,and8.82%are structural pro-teins.In addition,17.65%of proteins are transport proteins which allow the passage of inorganic ions and other small,water-soluble molecules into the cells,5.88%proteins were not easily categorized and labeled“others”.Since the types of all membrane proteins are still unknown in HeLa cells,we did not compare the number of collected membrane proteins to the total number of membrane proteins.In addition,the loss of proteins during␮ATPS could not be determined due to unknown number of total proteins and possible loss of proteins during crude membrane protein extraction.4.ConclusionsIn this paper,we demonstrate a microﬂuidic aqueous PEG/detergent two-phase system for the puriﬁcation of membrane proteins from crude cell extract.The method was applicable to hydrophobic proteins such as membrane proteins extracted from eukaryotic cells.Our␮ATPS combined the use of the zwitteri-onic detergent Zwittergent3-10,sodium dodecyl sulfate(SDS) and the nonionic detergent Triton X-114,resulting in a comple-mentary solubilization of proteins.The PEG/detergent two-phase system partitioning allowed successful removal of soluble pro-teins.Integral and peripheral membrane proteins remained in the detergent phase,while soluble proteins were found in the PEG-rich phase.Results indicated that approximately90%of the puriﬁed pro-teins were membrane proteins,including membrane-associated proteins and integral membrane proteins with multiple transmem-brane pared to conventional approaches,this new method had advantages of greater speciﬁc surface area,minimal emulsiﬁcation,reduced sample consumption and analysis time.We expect the developed method to be potentially useful in membrane protein puriﬁcations,facilitating the investigation of membrane proteomics.AcknowledgementsThe authors gratefully acknowledgeﬁnancial support from the National Basic Research Program of China(2007CB914203and 2007CB714507)and the National Natural Science Foundation of China(30970692,20875035and30800286).References[1]C.C.Wu,M.J.MacCoss,K.E.Howell,J.R.Yates,Nat.Biotechnol.21(2003)532.[2]A.Abbott,Nature426(2003)755.[3]M.C.King,C.P.Lusk,G.Blobel,Nature442(2006)1003.[4]C.C.Wu,J.R.Yates,Nat.Biotechnol.21(2003)262.[5]C.Smith,Nat.Methods2(2005)71.[6]M.W.Qoronﬂeh,B.Benton,R.Ignacio,B.Kaboord,J.Biomed.Biotechnol.2003(2003)249.[7]G.Munchow,F.Schonfeld,S.Hardt,K.Graf,Langmuir24(2008)8547.[8]H.Everberg,R.Peterson,S.Rak,F.Tjerneld,C.Emanuelsson,J.Proteome Res.5(2006)1168.[9]T.Chapman,Nature434(2005)795.[10]A.Hibara,M.Tokeshi,K.Uchiyama,H.Hisamoto,T.Kitamori,Anal.Sci.17(2001)89.[11]T.Minagawa,M.Tokeshi,T.Kitamori,Lab Chip1(2001)72.[12]K.Sato,M.Tokeshi,T.Sawada,T.Kitamori,Anal.Sci.16(2000)455.[13]H.Hisamoto,T.Horiuchi,K.Uchiyama,M.Tokeshi,A.Hibara,T.Kitamori,Anal.Chem.73(2001)5551.[14]M.Surmeian,A.Hibara,M.Slyadnev,K.Uchiyama,H.Hisamoto,T.Kitamori,Anal.Lett.34(2001)1421.[15]H.B.Kim,K.Ueno,M.Chiba,O.Kogi,N.Kitamura,Anal.Sci.16(2000)871.[16]R.J.Meagher,Y.K.Light,A.K.Singh,Lab Chip8(2008)527.[17]G.Munchow,S.Hardt,J.P.Kutter,K.S.Drese,Lab Chip7(2007)98.[18]J.Atencia,D.J.Beebe,Nature437(2005)648.[19]P.J.A.Kenis,R.F.Ismagilov,G.M.Whitesides,Science285(1999)83.[20]J.Blonder,M.B.Goshe,R.J.Moore,L.Pasa-Tolic,C.D.Masselon,M.S.Lipton,R.D.Smith,J.Proteome Res.1(2002)351.[21]K.K.Hixson,N.Rodriguez,D.G.Camp,E.F.Strittmatter,M.S.Lipton,R.D.Smith,Electrophoresis23(2002)3224.[22]H.Everberg,T.Leiding,A.Schioth,F.Tjerneld,N.Gustavsson,J.Chromatogr.A1122(2006)35.[23]H.Everberg,U.Sivars,C.Emanuelsson,C.Persson,A.K.Englund,L.Haneskog,P.Lipniunas,M.Jornten-Karlsson,F.Tjerneld,J.Chromatogr.A1029(2004)113.。

acquisition 英语解释

acquisition 英语解释Acquisition is the process of getting something new, whether it's a physical object, a skill, or knowledge. It's like adding a new piece to your puzzle or filling a gap in your understanding. When you buy a new book, that's an acquisition. Learning a new language? Yeah, that's an acquisition too. It's the act of taking in and making something your own.Acquiring things can be exciting. It's like finding a hidden treasure or unlocking a new door in a mystery game. It might take time and effort, but the satisfaction you get from it is often worth it. Whether you're acquiring a new hobby, a piece of art, or a valuable skill, it's always a growth experience.Sometimes, acquisition is part of a strategy. In business, for example, companies acquire other companies to expand their reach or gain access to new technologies. It's a way of growing and evolving, and it takes carefulplanning and execution. But the result can be powerful and transformative.Acquisition isn't just about material things. It's about growth, discovery, and the joy of learning. It's about taking risks, trying new things, and embracing what comes next. So, go ahead and acquire something new today. You never know what amazing adventures it might lead to!。

英文翻译

强化氨氮去除序批式反应器工艺中使用一致的生物改性和氨沸石再生交换摘要改进后的沸石--法可以推荐用于新的氮去除工艺。

因为其对于一贯的氨再生交换和沸石的生物再生具有很好的效果。

序批式反应器，可控制的反应器，沸石--序批式反应器，改进的沸石——序批式反应器，三个工艺都可用于评估氮的去除效率可控反应器包括缺氧填充，曝气混合，沉淀转移三个阶段。

这就意味着氮的去除效率取决于一个循环周期的移除体积。

沸石——序批式反应器以同样的方式作为一种控制反应器运行。

除了在序批式反应器中的沸石粉末。

在改进的沸石——序批式反应器工艺中运行次序是不断变化的。

缺氧填充阶段通常在曝气混合阶段之后，而在改进后的沸石——序批式反应器曝气混合阶段却在缺氧填充阶段之后并且携带氨氮到下一个运行周期，同时减少出水中总氮的浓度，在改进后沸石——序批式反应器工艺中，硝化作用和生物再生作用产生于起初的曝气混合阶段，而反硝化作用和氨吸附作用产生于接下来的厌氧填充阶段。

在改进后的沸石——序批式反应器工艺中，不断变化的运行次序以促进氨的吸附和沸石的生物再生，从而提高氨的去除效率。

随着工艺的不断运行在可控和沸石——序批式反应器内氨的去除效率能达到68.5--70.9%，这个效率主要取决于一个循环周期内33%的移除体积。

沸石——序批式反应器工艺表明一致的氨吸附交换和生物再生在缺氧填充阶段和曝气混合阶段是两个独立的过程。

与此同时在改进的沸石序批式反应器工艺中，通过沸石的氨吸附作用出水的总氮浓度能达到50--60mgN/L。

当出水中的氨浓度达到315mgN/L,这就表明和相同的条件下的可控反应器和沸石——序批式反应器相比总氮去除效率被提高了10%，氨的吸附产量能够达到6--7mg/gFSS，也就是相当于40mg/L氨氮被去除关键词：氨氮沸石生物再生氨离子交换略写词：SBR代表序批式反应器；可控反应器代表在缺氧填充和曝气混合阶段连续运行的反硝化作用和硝化作用的序批式反应器；沸石—序批式反应器代表除了日常投加沸石粉末的方式不同外，其他的运行方式都与可控反应器一样；改进的沸石—序批式反应器代表除了运行方式的连续不同外，其余的运行方式都与沸石—序批式反应器一样；HRT代表水力停留时间；SRT代表固体停留时间；SS 代表悬浮固体；FSS代表混合液悬浮固体；COD代表化学需氧量：TOC代表总有机碳；TIC代表总无机碳。

Restructuring (换序译法1)

(8). Many scientists hold that man’s social practice alone is the criterion of the truth of his knowledge of the external world. 许多科学家认为，只有人们的社会实践，才是人们对外界认识的真理性的标准。 (9). This is no class war, but a war in which the whole British Empire and Commonwealth of Nations are engaged, without distinction of race, creed or party.
由于英语和汉语的语法结构和表达习惯有很大的不同在翻译过程中根据译文语言的表达习惯对原文的词序进行调整如原文中在句子中处于前面位置的词或短语在译文中调整到句子后面的位置使译文做到最大程度上的通顺自然
Restructuring 换序译法
1.Definition Restructuring—the necessary or inevitable change of word-order or the logical rearrangement of expression in a sentence in conformity with the standard or good usage of the target language on account of syntactic differences between the source language and the target language. 由于英语和汉语的语法结构和表达习惯有很大的不同，在翻译过程中，根据译文语言的表达习惯对原文的词序进行调整，如原文中在句子中处于前面位置的词或短语，在译文中调整到句子后面的位置，使译文做到最大程度上的通顺自然。

Intact orange quality prediction with two portable NIR spectrometers

Postharvest Biology and Technology 58(2010)113–120Contents lists available at ScienceDirectPostharvest Biology andTechnologyj o u r n a l h o m e p a g e :w w w.e l s e v i e r.c o m /l o c a t e /p o s t h a r v b ioIntact orange quality prediction with two portable NIR spectrometersJoséA.Cayuela a ,∗,Carlos Weiland ba Instituto de la Grasa,CSIC,Avda.Padre García Tejero 4,41012Sevilla,SpainbDepartamento de Ciencias Agroforestales,Universidad de Huelva,21819La Rábida,Palos de la Frontera,Huelva,Spaina r t i c l e i n f o Article history:Received 10February 2010Accepted 6June 2010Keywords:AcidityFruit weight Firmness JuicinessMaturity index NIR OrangeSoluble solid contenta b s t r a c tTwo commercial portable spectrometers were compared for orange quality non-destructive predictions by developing partial least squares calibration models,reﬂectance mode spectra acquisition being used in both.One of them was a Vis/NIR spectrometer in which the radiation reﬂected by the fruit is collected and conducted by optic ﬁber to the three detectors (350–2500nm)of the instrument.The other is an AOTF-NIR with a reﬂectance post-dispersive optical conﬁguration and InGaAs (1100–2300nm)detector.Four orange varieties were included in calibrations.The parameters studied were soluble solids content,acidity,titratable acidity,maturity index,ﬂesh ﬁrmness,juice volume,fruit weight,rind weight,juice volume to fruit weight ratio,fruit colour index and juice colour index.The results indicate good perfor-mance of the predictive models,particularly for the direct NIR prediction of soluble solids content,and maturity index,the prediction of this last parameter being notable for its relevance and novelty.The RPD ratios for these parameters were in the range from 1.67to 2.21with the Labspec spectrometer,which showed better predictive performance,and from 1.03to 2.33with the Luminar instrument.©2010Elsevier B.V.All rights reserved.1.IntroductionConsumers purchase citrus fruit on the basis of quality,this being a combination of characteristics and attributes signiﬁcant for acceptability.Citrus are non-climateric fruit,hence the ripening process stops once separated from the tree and,consequently,fruit can only be harvested and marketed once adequate maturity has been reached (Watkins,2008).Moreover,the content of sugars and acids in citrus fruit is fairly stable before and after harvest,sugars-to-acid balance being the key to acceptability in these fruit.The content of sugars is generally measured by refractometry as solu-ble solids content (SSC),sugars representing the main component,and acids content is commonly measured as titratable acidity (TA).The ratio of soluble solids content to titratable acidity (SSC/TA)is widely used as a maturity criterion for non-climateric fruit (Fellars,1991),for the reasons indicated above,and particularly used as a maturity index in citrus.The contribution of organic acids to the SSC in citrus fruit is in the 10%range.Total acidity prediction by near infrared spec-troscopy (NIR)has been considered difﬁcult to achieve,due to the relatively low levels of organic acids in fruit (McGlone et al.,2003;Guthrie et al.,2005).Several authors have reported various levels of success in predicting titratable acidity (TA)of pineapple (Shiina et al.,1993),plum (Onda et al.,1994),apple (Sohn et al.,2000),mango∗Corresponding author.Tel.:+34954611550;fax:+34954616790.E-mail address:jacayuela@ig.csic.es (J.A.Cayuela).(Schmilovitch et al.,2000),Imperial mandarin (Guthrie et al.,2005)and Satsuma mandarin (Hernández et al.,2006).Citrus fruit are anisotropic objects,showing different physical and chemical prop-erties when measured in different directions.Hence,equatorial measurements are reasonable since at least in citrus,SSC is great-est in the distal apex of fruit decreasing towards the proximal,the opposite happening with acids and TA as reported by Peiris et al.(1999).Colour is considered as one of the most important external fac-tors of fruit quality (Francis,1995),since the fruit’s appearance greatly inﬂuences the consumer.The change of colour in citrus is a consequence of the maturation process,although it is also highly dependent on cool temperatures at night,not always present under tropical and subtropical growth conditions,which is the reason why a green citrus fruit may or may not be physiologically mature (Olmo et al.,2000).Fruit softening is often used as a criterion for selecting the most suitable harvest date for several commodities (Lehman-Salada,1996).The most common method to determine the ﬁrmness of a fruit is destructive and measures its resistance to penetration (Lehman-Salada,1996;Ahumada and Cantwell,1996;Mercado-Silva et al.,1998).Other methods based on fruit resistance to compression do not necessarily destroy the fruit,but they do require its harvest (Polderdijk et al.,1993;Brovelli et al.,1998).Other methods,in addition to being non-destructive,can be used directly on the tree,such as the use of the hand densitometer (García et al.,1998),or those based on the transmission of acoustic waves through the fruit (Muramatsu et al.,1996).In citrus fruit,the0925-5214/$–see front matter ©2010Elsevier B.V.All rights reserved.doi:10.1016/j.postharvbio.2010.06.001114J.A.Cayuela,C.Weiland/Postharvest Biology and Technology58(2010)113–120Table1Statistical data of the sets of samples and orange varieties.Sanguinelli Valencia Salustiana Navelate TotalN ¯X N ¯X N ¯X N ¯X N ¯XSSC2500.848.5350 1.048.16440.8410.94520.7211.86396 1.569.19 A2500.15 3.14500.13 3.17440.12 3.87520.13 3.523960.28 3.28 TA2500.190.78500.21 1.48440.090.70520.16 1.06396 3.129.38 MI250 2.4511.35500.63 5.5644 2.3015.8552 1.7911.44396 3.3611.14 F250 1.96 5.7950 1.397.6744 1.06 5.7652 1.437.44396 1.91 6.24 JV2507.2531.88509.4048.5844 6.2976.80527.8774.2739619.8044.55 FW25020.7181.525023.98124.05449.41214.605219.22290.1939677.88129.08 JV/FW2500.040.39500.020.39440.020.36520.030.263960.060.37 RW25014.1643.625014.7072.0944 6.62128.155217.59209.7239659.2978.42 FCI2508.6221.05508.4015.0744 6.0826.4852 3.5418.513968.3520.56 JCI25088.3562.975024.52−71.054414.80−35.625230.18−72.4639693.9017.31SSC,soluble solids content(%);A,acidity(pH);TA,titratable acidity(g/L citric acid);MI,maturity index;F,ﬂeshﬁrmness(N);JV,juice volume(mL);FW,fruit weight(g); RW,rind weight(g);FCI,fruit colour index;JCI,juice colour index; ,standard deviation;and¯X,mean.relationship between the degradation of the cellular wall and the loss ofﬁrmness that accompanies fruit maturation has also been observed(Goto and Araki,1983).Other important attributes of the internal quality of fruit,along with those mentioned above,are texture and rind thickness.In the same way,juiciness is another important fruit attribute,which can eventually be reduced in citrus by factors affecting the content of the juice sacs,such as freezing or excess of nitrogen fertilization during summer and early autumn (Flint,1991).Excepting fruit colour,none of them are able to be assessed by the appearance of the fruit to the consumer,whose decision to choose fruit of a desired quality,therefore,is not sup-ported by sufﬁcient objective information(Poole et al.,2006).There is a need for techniques for swift,non-destructive deter-minations of fruit internal quality,to ensure that all fruit meet a minimum level of acceptance.Simpliﬁcation of the analysis and the possibility to monitor practically all the fruit in real time are important reasons for this objective.Conversely,improving the environmental sustainability of human activities is a current challenge that should be emphasized.Non-destructive analyti-cal techniques can contribute,since they do not require chemical reagents or solvents and no waste is generated.The most suitable technology depends on the main quality parameter to be measured.Among several techniques,NIR has great potential for non-destructive determination of internal and maturity attributes(Abbot,1999).The measurement modes most frequent for intact fruit SSC and TA prediction are reﬂectance, transmittance and interactance.Although slightly higher predic-tive outcomes using transmittance have been reported,in contrast with reﬂectance and interactance with intact mandarins(McGlone et al.,2003),good results using the reﬂectance mode have also been reported with mandarins(Guthrie et al.,2005;Hernández et al., 2006)and oranges(Cayuela,2008).Reﬂectance is the easiest mode to obtain measurements,since no contact with the fruit is required and light levels are relatively high(Mowat and Poole,1997).In the transmission mode,the measurements are expected to be more inﬂuenced by fruit size,the amount of light penetrating the fruit often being very small,thus making it difﬁcult to obtain accurate transmission measurements at grading line speeds(Kawano et al., 1993).NIR spectra are the result of the interaction of radiation with the sample,and their physical and chemical properties are reﬂected in it.Fruit juiciness and fruit weight are fruit physical properties. Successful results of NIR calibrations for citrus juiciness prediction have been reported(Guthrie et al.,2005).The possibility of estimat-ing fruit weight by NIR has rarely been reported,the exploration of this possibility being of great interest,since fruit weight could be added to other fruit quality parameters such as SSC,TA or fruitﬁrm-ness as different outputs from a single NIR measure.In fact,some good outcomes have been found recently regarding NIR measur-ing nectarine fruit weight(Pérez et al.,2009).Previous research (Cayuela,unpublished data)has also indicated that nectarine fruit weight can be predicted by NIR.One of the advantages of NIR spectrometry is its portability when the parameter must be measured in situ.A few models of portable NIR spectrometers of several brands are available,but the applications to fruit monitoring are few.Furthermore,the techni-cal and constructive characteristics of NIR spectrometers are very diverse,and research into their suitability for use in new applica-tions is needed.Riquelme(2008)has carried out a full revision of the commercial models of on-line NIR instruments and the portable NIR spectrometers applicable to fruit.In the present work,the feasibility of non-destructive NIR pre-diction of quality parameters on orange fruit,comparing two commercially representative portable spectrometers using predic-tive models constructed by partial least squares(PLS),has been evaluated.The successful prediction of some parameters analyzed in this work is reported for theﬁrst time.2.Materials and methods2.1.FruitSanguinelli,Valencia,Salustiana and Navelate oranges were hand-picked,at random,during the commercial harvest period, from a local experimental grove belonging to the College of Agri-cultural Engineering,University of Huelva,transported to the laboratory and used immediately or after storage at4◦C for up to one week.The orange varieties included in this study are taxonom-ically all Citrus sinensis(L)Osb.The number of samples from each variety contributing to the calibration set is indicated in Table1. The oranges were harvested atﬁve different dates from January to April2009,and therefore,wide diversity in the fruit quality param-eters was assured.Before testing,samples were taken out from cool storage and maintained at room temperature(23–25◦C)for18h in order to allow for acclimatization to the experimental conditions. Fruit were cleaned with a cloth moistened in sterile water,then dried in the laboratory environment prior to measurement.Each fruit unit constituted the sample,and was numbered in the fruit peduncle area.2.2.Spectral acquisitionThe spectral acquisition of every sample was performed using two portable spectrophotometers with different optical and constructive features:Labspec(Analytical Spectral Devices Inc., Boulder)and Luminar5030(Brimrose Corp.,Maryland).Labspec isJ.A.Cayuela,C.Weiland/Postharvest Biology and Technology58(2010)113–120115 a Vis/NIR spectrometer equipped with three detectors.The detectorfor the visible range(350–1000nm)is aﬁxed reﬂective holographicdiode array with a sensitivity of512pixels.Wavelengths in the vis-ible spectrum can carry information relating to some of the qualityparameters analyzed,such as fruit colour and colour of the juice,and perhaps additional ones such as fruit size.The wavelengthrange of1000–1800nm is covered by a holographic fast scan-ner InGaAs detector cooled at−25◦C.The same aforementioneddevice coupled with a high order blockingﬁlter operates for the1800–2500nm interval.The instrument is equipped with internalshutters and automatic offset correction,the scanning speed being100ms.The acquisition of spectra was performed using the highintensity contact probe accessory of the spectrometer,with lightsource diameter20mm,and standard SMA905ﬁber optic con-nectors.The whole spectrum was acquired,each spectral variablecorresponding to2nm interval.The repeatability of the instru-ment,expressed as standard deviation on the average absorbanceof350–2500nm ofﬁve measures of a white tile,is6.00×10−4.Theorange spectra acquisition was carried out using Indico Pro soft-ware(Analytical Spectral Devices Inc.,Boulder).The portability ofthe equipment is assured by the weight of the spectrometer8.5kg.Luminar5030is an AOTF(acousto-optic tunableﬁlter)NIR spec-trophotometer,equipped with a reﬂectance post-dispersive opticalconﬁguration and InGaAs(1100–2300nm)detector.The referencespectrum is automatically taken by the instrument,just as anUV–Vis spectrophotometer dual beam;the beam is divided beforeleaving and a small portion is sent to a second detector that makesthe reference.The scanning speed in Luminar5030is60ms.Thespectrometer is formed by the hand-held unit,shaped and usedsimilarly to a‘gun’,the diameter of the exit cone of the light sourcebeing8mm,and the computer unit;the spectrometer set,with atotal weight of5.26kg,offers good portability,with4h of autonomyusing a set of batteries that allows in situ measuring on a crop;thisis an important advantage to note.The hand-held unit is equippedwith a base for facultative use in the laboratory.The whole spec-trum was acquired,each spectral variable corresponding to a2nminterval.The repeatability of the instrument,expressed as standarddeviation on the average absorbance of1100–2300nm ofﬁve mea-sures of a white tile,is6.76×10−4.The signals were acquired withsoftware Acquire(Brimrose Corp.,MD).The sample unit was the fruit,an averaged spectrum beingobtained for each sample,resulting from a total100spectracorresponding to two measures of50spectra each,it,for bothspectrometers,was taken at opposite equatorial locations.2.3.Reference analysisThe quality attributes of each fruit were evaluated by analysis oftheir physical and chemical parameters.Additionally,parameterscalculated arithmetically were assessed.2.3.1.Physical parametersFruit weight(FW,g)and rind weight(RW,g)were measuredusing an electronic precision balance(0.001g).RW was deter-mined once the fruit was peeled by separating the rind from theﬂesh.For the measurement of intact fruit colour a spectral colouranalyser(colorimeter PCE-RGB1002)was used,with triplicatemeasurements for each fruit.This instrument has a RGB(red,green,blue)colour scale0-1023.The Easy-RGB software(Logicol ColourTechnology Co.)was used for the conversion into the L,a and bparameters of the Hunter scale.The results expressed by the colourindex(FCI)were obtained from the mathematical formula(1).CI=1000a(1)Fleshﬁrmness(F,N)was quantiﬁed using a hand penetrom-eter(TR FT-327Turoni S.r.l.,Forcy,Italy)with a7mm diametercylindrical plunger,twice on the peeled fruit at the equatorial cir-cumference.The fruit were halved through the equatorial planeand juice extracted with a commercial juice extractor.Juice vol-ume(JV,mL)was measured with a test tube.Juice volume to fruitweight ratio was calculated and expressed as a percentage(JV/FW,%).Juice colour was determined with the same colorimeter andunits indicated above for fruit colour and expressed as juice colourindex(JCI).The measurement was performed on a juice sample ofeach individual orange fruit in a Petri dish,this measurement beingmade through the glass at the bottom of the dish.This procedurewas used to avoid the risk of any introduction in the juice of thecolorimeter light source.2.3.2.Chemical parametersSSC was measured on juice of each fruit by a hand-held digitalrefractometer(Atago Co,PAC-1Brix-Meter,Tokyo)and obtainedfrom two replicates,expressed as percentage.Acidity(A,pH)was measured on the juice of individual fruit fromtwo replicates,using a digital pH-meter.The titratable acidity(TA)was analogously measured from two replicates by direct titration ofa10mL juice sample added with10mL distilled water,neutralizedwith NaOH0.1N until pH8.2and expressed as citric acid(g/L).2.3.3.Arithmetically calculated parametersThe maturity index reference(MI)was arithmetically obtainedfrom the ratio between SSC and TA reference analysis values.Thearithmetical computation from SSC and TA obtained using the NIRpredictive models developed in this work for both parameters(SSC P/TA P)was compared with the prediction outcomes from themodel developed for directly predicting MI.2.4.Chemometrics and calibration proceduresPartial least squares models were obtained with Unscrum-bler9.7(CAMO Software AS,Norway).For the Labspec spectra,noise intervals350–499nm and2301–2500nm were removed.Inturn,tests were carried out for the same spectrometer excluding600–750nm,a range strongly affected by the skin pigment chloro-phyll that absorbs red light,whose absorbance band correspondsto680nm.This band is not included in the Luminar’s wavelengthrange.Exclusion of the initial andﬁnal areas of spectrum to elimi-nate noise,in the case of Luminar was also considered unnecessary.Before calibrations,the reﬂectance data were transformed toabsorbance,mean normalized,and optionally treated by mul-tiplicative scatter correction(MSC)using Unscrumbler9.7.Theinﬂuence of the pre-processing on the prediction of the calibrationmodels was tested by different gap and smooth combinations forﬁrst and second gap-segment derivatives.Standard normal variatetransformation(SNV)was also tested.Full-cross internal validation(FCV)was used for building the models.Calibration tests were alsoconducted with different numbers of principal components in orderto determine the number of PCs optimum and the results assessedin terms of standard error of cross validation(SECV).Exception-ally,where indicated in the tables,points clearly separated fromthe calibration sets in the scatter plots were identiﬁed as outliersand removed with Unscrumbler9.7using its speciﬁc applicationfor this purpose.2.4.1.Validation procedureTwo external validation exercises were carried out using the cor-responding models for predicting the parameters.One validationexercise was conducted using1/5of the total number of samplesfor each parameter,the set for validation being formed by Unscrum-bler’s speciﬁc application,theﬁrst from every5samples taken116J.A.Cayuela,C.Weiland /Postharvest Biology and Technology 58(2010)113–120Fig.1.(A)Labspec and (B)Luminar 5030.Examples of absorbance spectra from the same ﬁve oranges.for this purpose (V1).Another exercise was conducted for the set of samples 51–100,corresponding to Sanguinelli (V2),which was excluded from the corresponding calibrations that were developed previously for this purpose.In assessing the soundness of the calibrations performance,the main considerations were the root mean square error of prediction (RMSEP)and the residual predictive deviation (RPD),described by Williams and Sobering (1996)as the ratio of the standard deviation of the reference data for the validation set to the SEP.Paired samples T -tests for dependent samples were also con-ducted to verify the results from RPD and RMSEP analysis.Not established as a statistic speciﬁcally for assessing PLS model per-formance,the T -test was applied here exclusively referred to the data pairs included in the external validation exercise.For this pur-pose,data pairs of the reference value and the resulting prediction corresponding to the external validation sets were compared.The paired T -test is a parametric procedure,useful for testing whether the means of two groups are different,where the samples are drawn in pairs.The T -test was carried out using SPSS Statistics software (SPSS Inc.,Chicago).The compliance with the null hypothesis in this test (P >0.05)indicates that the measure NIR provides at least the same accuracy as the reference method.The prediction output from the calibration model for direct NIR MI measuring was compared both with the reference values and with SSC P /TA P above described,also using the paired samples T -test.3.Results3.1.Fruit NIR spectraTypical diffuse absorbance spectra for intact oranges acquired by both instruments (Labspec and Luminar 5030)are shown in Fig.1.The spectra from the two instruments showed characteristics simi-lar to those also previously described in oranges (Cayuela,2008)or in Satsuma mandarin (Hernández et al.,2006)beyond differences regarding the wavelengths acquired.High intensity peaks are noted,related to the strong water absorbance bands present from their ﬁrst overtone at 1400–1500nm and combination band at 1880–2100nm with an interval of relative intensities of high absorption,approximately between 1400and 2200nm.Absorbance falls around the 1500nm and rises again from approximately 1850nm where oscillation exists probably due to carbohydrates.Overall,the pattern of the absorption curves is similar to that for other fruit such as peach (Peiris et al.,1997),kiwifruit (McGlone and Kawano,1998),mango (Budiastra et al.,1998),apple (Lu et al.,2000)and cherry (Lu,2001).3.2.Population characterizationThe mean and standard deviation values of the populations of orange varieties used in the calibrations and the external valida-tions for the parameters analyzed are shown in Table 1.The MI in the different varieties shows that Valencia oranges corresponded to a less advanced stage of maturation,Sanguinelli and Navelate being at approximately the same maturity stage and Salustiana show-ing,according to its MI ratio,the most advanced maturity of the four varieties.As can be seen,wide ranges of variation of all the parameters analyzed were included in the populations.The characteristics of the sample sets used for the validation exercises,conducted using models obtained for predicting each parameter in samples independent from the calibrations sets,are shown in Table 2.The contribution of each variety to the validation sets were 50Sanguinelli,10Valencia,9Salustiana and 10Nave-late samples,according to the proportion 1/5regarding calibration sets.3.3.Calibration developmentThe statistical coefﬁcients of best calibration models with Lab-spec and Luminar 5030for predicting orange quality parameters according to the treatment used for the two instruments tested,are given in Tables 3and 4,respectively.Table 2Statistical data of the sets of samples used in V1.Parameter Samples Range¯X SSC 76 5.5–12.4 1.599.08A 76 1.1–3.90.28 3.27TA 750.6–2.60.290.91MI 75 4.0–16.7 3.2210.85F 76 1.5–11.7 1.93 5.96JV 7418.0–88.020.5845.00FW 7547.4–311.479.19127.08JV/FW 760.2–0.50.060.37RW 7618.5–228.560.076.5FCI 728.7–31.0 4.6420.54JCI76−141.9–223.487.4016.40SSC,soluble solids content (%);A,acidity (pH);TA,titratable acidity (g/L citric acid);MI,maturity index;F,ﬂesh ﬁrmness (N);JV,juice volume (mL);FW,fruit weight (g);RW,rind weight (g);FCI,fruit colour index;JCI,juice colour index; ,standard Deviation;and ¯X,mean.J.A.Cayuela,C.Weiland/Postharvest Biology and Technology58(2010)113–120117Table3Labspec:statistics of calibrations and validations(wavelength500–2300nm).Parameter Treatment Calibration ValidationPCs Outl.RMSECV R R CV RMSEP RPD TSSC MN1000.600.920.910.74 2.130.719A MN,MSC1040.120.900.880.15 1.850.714 TA MN1000.160.860.830.17 1.690.978 MI MN80 1.810.850.81 1.92 1.670.765 FW MN,MSC10219.890.970.9616.52 4.760.716F MN73 1.180.760.72 1.05 1.850.356 JV MN937.390.920.917.05 2.940.511 JV/FW MN,MSC930.030.890.870.04 1.610.327 RW MN10214.610.970.9612.98 4.540.349 FCI MN,MSC80 3.450.900.87 2.65 1.750.055 JCI MN,MSC100 1.590.860.8366.78 1.310.286SSC,soluble solids content(%);A,acidity(pH);TA,titratable acidity(g/L citric acid);MI,maturity index;F,ﬂeshﬁrmness(N);JV,juice volume(mL);FW,fruit weight(g);RW, rind weight(g);MN,mean normalization;MSC,multiplicative scatter correction;Outl.,outliers;RMSEC,root mean square error of calibration;R,coefﬁcient of calibration; R CV,coefﬁcient of cross validation;RMSEP,root mean square error of prediction;RPD,residual predictive deviation;and T,P-value from paired samples test.Table4Luminar5030:statistics of calibrations and validations(wavelengths1100–2300nm).Parameter Treatment Calibration ValidationPCs Outl.RMSECV R R CV RMSEP RPD TSSC MN,MSC1000.670.910.890.68 2.330.273A MN,MSC1010.150.840.810.16 1.750.931 TA MN1020.180.800.770.19 1.540.321 MI MN,MSC91 2.060.790.74 1.96 1.640.313 FW None9322.690.960.9526.26 3.030.414F MSC100 1.270.710.66 1.39 1.390.280 JV MN,MSC927.970.910.918.00 2.560.977 JV/FW MSC910.030.830.800.04 1.670.148 RW None9217.070.960.9518.86 3.230.249SSC,soluble solids content(%);A,acidity(pH);TA,titratable acidity(g/L citric acid);MI,maturity index;F,ﬂeshﬁrmness(N);JV,juice volume(mL);FW,fruit weight(g);RW, rind weight(g);MN,mean normalization;MSC,multiplicative scatter correction;Outl.,outliers;RMSEC,root mean square error of calibration;R,coefﬁcient of calibration; R CV,coefﬁcient of cross validation;RMSEP,root mean square error of prediction;RPD,residual predictive deviation;and T,P-value from paired samples test.The range from600to750nm could be affected by the skin pigment chlorophyll,as has been previously indicated.Calibra-tion tests were conducted with the Labspec excluding this range, without any improvement when including it.This result probably reﬂects the fact that no green areas were present in the skin of any of the orange samples used in the calibrations.This test was unnec-essary with the Luminar,since its spectral range does not include 600–750nm.The calibration statistics for V1were relatively close with both instruments for SSC,A,MI,JV and JV/FW.Hence,with the Lab-spec the RMSECV were0.60,0.12,1.81,7.39and0.03,whereas for the Luminar they were0.67,0.15,2.06,7.97and0.03,respec-tively for these parameters(Tables3and4).Some differences were found between both spectrometers for FW,RW and F,the Labspec showing RMSECV19.89,14.61and1.18,whereas for the Luminar it was22.69,17.07and1.27.In all the cases,the lower RMSECV values from the Labspec imply a better predictive per-formance for these parameters.Calibration for FCI and JCI was attempted only with the Labspec,since it integrates visible and NIR.Generally,mean normalized data provided the bestﬁts for most parameters analyzed.In some cases,shown in Tables3and4,in both or either spectrometers,MSC treatment alone or after mean normalization facilitated the best calibration coefﬁcients.3.4.Model external validationThe outcomes from the external validation exercises V1for each orange quality parameter with both spectrometers are shown in Tables3and4.The predictions versus the analyzed value of each parameter in this external validation exercise are shown in Fig.2.As can be seen from the validation plots,and according to the validation coefﬁcients RMSEP and RPD shown in Tables3and4, prediction accuracy was similar for most fruit quality parameters with both spectrometers and slightly better with the Labspec for FW,RW and F.The validation exercise V2,corresponding to Sanguinelli,pro-vided better performance for all the parameters analyzed with the Labspec.The statistical values are shown in Table5.The best RPD with the same spectrometer in this exercise were reached for SSC and MI predictions.Table5Validation V2(N=50;wavelength500–2300nm).Labspec Luminar5030 Parameter RMSEP RPD RMSEP RPD SSC0.87 2.21 1.12 1.03 A0.13 1.050.400.80 TA 2.47 1.26 2.07 1.07 MI 1.54 1.75 2.57 1.26 FW43.51 1.1132.630.75 F 1.82 1.20 1.53 1.03 JV8.38 1.1012.130.84 JV/FW0.04 1.180.05 1.10 RW16.07 1.1114.710.73 FCI 6.48 1.64JCI55.690.67SSC,soluble solids content(%);A,acidity(pH);TA,titratable acidity(g/L citric acid); MI,maturity index;F,ﬂeshﬁrmness(N);JV,juice volume(mL);FW,fruit weight (g);RW,rind weight(g);RMSEP,root mean square error of prediction;RPD,residual predictive deviation;and N,number of samples of the validation set.。

收购与重组

9
Reasons for Making Acquisitions:
Reshaping the Firms’ Competitive Scope

Firms may use acquisitions to reduce their dependence on one or more products or markets Reducing a company’s dependence on specific markets alters the firm’s competitive scope
4
Reasons for Making Acquisitions:
Increased Market Power

Factors increasing market power
– when a firm is able to sell its goods or services above competitive levels or – when the costs of its primary or support activities are below those of its competitors – usually is derived from the size of the firm and its resources and capabilities to compete

Acquisition of a competitor may result in
– more predictable returns – faster market entry – rapid access to new capabilities
7
Reasons for Making Acquisitions:

could retain secondary structure

could retain secondary structureSecondary structure refers to the local arrangement of amino acid residues in a protein. It is determined by the hydrogen bonding patterns between the backbone atoms of the polypeptide chain. The two most common types of secondary structures are alpha helices and beta sheets. In addition, turns and loops are also considered as secondary structures.Alpha helices are right-handed helical structures formed by the hydrogen bonds between the carbonyl oxygen of one amino acid residue and the amide hydrogen of a residue located four residues away. This regular repeating pattern stabilizes the helix structure. The R groups of the amino acids protrude outward from the helix, avoiding steric clashes. Alpha helices are often found in the transmembrane regions of membrane proteins, where their cylindrical shape allows them to span the lipid bilayer.Beta sheets are formed by the alignment of beta strands, which are extended polypeptide chains connected by hydrogen bonds. Beta strands can be either parallel or antiparallel, depending on the direction of the hydrogen bonds between the strands. The R groups are arranged above and below the plane of the sheet. Beta sheets are commonly found in the core of globular proteins and can also form the basis of protein-protein interactions.Turns and loops are less regular structures compared to alpha helices and beta sheets. Turns are sharp bends in the polypeptide chain, often occurring when there is a change in the direction of the protein fold. Loops, on the other hand, are more extended regions connecting secondary structure elements. Turns and loopscan have various conformations and lengths, often contributing to the overall shape and flexibility of the protein. They play important roles in protein folding, stability, and function.The secondary structure of a protein is influenced by several factors, including the amino acid sequence and the local environment. Some amino acids have a high propensity to form alpha helices or beta sheets, while others have a preference for turns or loops. For example, alanine and leucine are often found in alpha helices, while glycine and proline are commonly present in turns and loops. The local environment, such as the presence of charged or aromatic amino acids nearby, can also affect the stability of secondary structures.Several methods can be used to predict secondary structure from the amino acid sequence. These methods are based on statistical analysis of known protein structures and the observation of specific sequence motifs associated with secondary structure elements. Machine learning algorithms, such as neural networks and support vector machines, have been developed to predict secondary structure with high accuracy. These predictions can provide valuable insights into protein function and folding, as secondary structure is closely related to the overall three-dimensional structure and stability of the protein.In summary, secondary structure refers to the local arrangement of amino acid residues in a protein. Alpha helices, beta sheets, turns, and loops are the most common types of secondary structures. The secondary structure is determined by hydrogen bonding patterns and is influenced by the amino acid sequence and the localenvironment. Predicting secondary structure from the amino acid sequence is a valuable tool in protein structure and function prediction.。

Modeling and simulation of the injection of urea-water-solution

Modeling and simulation of the injection of urea-water-solutionfor automotive SCR DeNO x -systemsFelix Birkhold a ,*,Ulrich Meingast a ,Peter Wassermann a ,Olaf Deutschmann baRobert Bosch GmbH,Robert-Bosch-Platz 1,70049Stuttgart,GermanybInstitute for Chemical Technology and Polymer Chemistry,University of Karlsruhe,76131Karlsruhe,GermanyAvailable online 30June 2006AbstractThe evaporation of water from a single droplet of urea water solution is investigated theoretically by a Rapid Mixing model and a Diffusion Limit model,which also considers droplet motion and variable properties of the solution.The Rapid Mixing model is then implemented into the commercial CFD code Fire 8.3from A VL Corp.Therein,the urea water droplets are treated with Lagrangian particle tracking.The evaporation model is extended for droplet boiling and thermal decomposition of urea.CFD simulations of a SCR DeNO x -system are compared to experimental data to determine the kinetic parameters of the urea decomposition.The numerical model allows to simulate SCR exhaust system conﬁgurations to predict conversion and local distribution of the reducing agent.#2006Elsevier B.V .All rights reserved.Keywords:Evaporation;Thermal decomposition;Urea-water-solution;Simulation;Injection;CFD;SCR;NO x1.IntroductionSelective catalytic reduction (SCR)of NO x is an effective technique for the reduction of nitrogen oxides emitted from various sources.In automotive applications,the urea-water-solution based SCR is a promising method for control of NO x emissions.Urea-water-solution (UWS,contains 32.5wt.%urea;brand name:AdBlue)is sprayed into the hot exhaust stream [1].The subsequent generation of NH 3in the hot exhaust gas proceeds in three steps [2,3]:(i)evaporation of water from a ﬁne spray of UWS droplets,ðNH 2Þ2CO ðaq Þ!ðNH 2Þ2CO ðs or l Þþ6:9H 2O ðg Þ;(1)(ii)thermolysis of urea into ammonia and iso-cyanic acid,ðNH 2Þ2CO ðs or l Þ!ðNH 3Þðg ÞþHNCO ðg Þ;(2)(iii)and hydrolysis of isocyanic acid,HNCO ðg ÞþH 2O ðg Þ!NH 3ðg ÞþCO 2ðg Þ:(3)As the evaporation and spatial distribution of the reducing agent upstream the catalyst are crucial factors for the conversion of NO x ,the dosing system has to ensure the proper preparation of the reducing agent at all operating conditions.Table 1gives an overview of the different exhaust gas and spray characteristics occurring in passenger cars and trucks.Appropriate spray properties of the urea solution will also avoid deposition of urea on walls,which could lead to melamine complexes [4].To our knowledge,there are no studies published on the inﬂuence of urea on the evaporation of water from a UWS droplet.Van Helden et al.[5]used water instead of UWS in a CFD study and estimated the concentration of the reducing agent from the water vapor concentration.Wurzenberger and Wanker [6]modeled the thermal decomposition as a homo-geneous gas phase reaction,which followed the evaporation of UWS to water vapor and gaseous urea.Chen and Williams [7]and Deur et al.[8]assumed the thermal decomposition to occur/locate/apcatbApplied Catalysis B:Environmental 70(2007)119–127*Corresponding author.Tel.:+4981141659;fax:+49811262203.E-mail address:felix.birkhold@ (F.Birkhold).0926-3373/$–see front matter #2006Elsevier B.V .All rights reserved.doi:10.1016/j.apcatb.2005.12.035instantaneously relative to the evaporation rate.Kim et al.[9] used a single kinetic rate de-volatilization model neglecting the appearance of isocyanic acid.Cremer et al.[10]assumed a fast decomposition of urea after the evaporation of pure water, because their focus was the selective non-catalytic reaction with temperatures above1100K.Analyzing the literature it does not seem to be clear in which state of aggregation urea appears after the evaporation of water and during the thermal decomposition.Yim et al.[2]indicated liquid or gaseous urea,Schaber et al.[11,12]reported that molten urea evaporates to gaseous urea at temperatures above 413K,but mainly decompose directly to NH3and HNCO above425K.Schmidt[13]excluded the existence of molten urea when urea particles are exposed to temperatures above 553K in aﬂuidized bed reactor.Koebel et al.[3]revealed that atomization of UWS in hot exhaust stream yields to solid or molten urea.The relevance of the physical condition of urea depends on the different inertia and tracking of gaseous and solid/molten urea in the gas stream and the differences in enthalpy.It is essential to know about spatial enthalpy variations due to evaporation,thermolysis,and hydrolysis, which will be discussed in this article.As the behavior of an UWS droplet in a heated environment is presently not well understood,a theoretical investigation of the evaporation of UWS droplets is presented.Differences in the evaporation between UWS and water are shown and a model to describe the thermal decompostion of urea is developed.The model containing both evaporation of water and the following thermal decomposition of urea is imple-mented into the CFD code Fire v8.3[14].The kinetic parameters for the decomposition model are determined comparing numerical simulation with experimental data from Kim et al.[9].The model derived in the present study allows to judge different SCR exhaust system conﬁgurations with respect to conversion and local distribution of reducing agent. Furthermore the behavior of droplets which impinge on the walls or catalyst can be estimated,because the physical conditions such as temperature and urea concentration of the droplets are determined from the simulation.2.Evaporation of urea-water-solution dropletThe inﬂuence of urea on the evaporation of water from a UWS droplet is investigated theoretically by different evaporation models considering droplet motion and variable properties of UWS and the ambient gas phase.2.1.Liquid phaseTo evaluate the inﬂuence of solved urea on the evaporation of water,three different evaporation models are used:-Rapid Mixing model(RM model):Within the RM model inﬁnite high transport coefﬁcients are assumed for the liquid phase,resulting in spatial uniform temperature,concentration andﬂuid properties in the droplet,but the quantities willF.Birkhold et al./Applied Catalysis B:Environmental70(2007)119–127 120Nomenclaturea thermal conductivity,m2/sA frequency factor,kg/s mB M,T Spalding numbersc molar concentration,mol/m3 c p heat capacity,J/kg KD diameter,mE a activation energy,J/molh speciﬁc enthalpy,J/kgH molar enthalpy,J/molLe Lewis numberm mass,kg˙m massﬂow,kg/sNu Nusselt numberP pressure,Par radius,mr rate of reaction,mol/m3sR unversial gas constant,J/mol K Sh Sherwood numbert time,sT temperature,Ku velocity,m/sv volume fractionw dimensionless radius,r/r dY mass fractionGreek symbolsl heat conductivity,W/m Kr density,kg/m3G diffusion coefﬁcient,m2/s Subscripts0initial*characteristic1ambientd,g droplet,gashy hydrolysisl,vap liquid,vaporu urearef referencerel relatives droplet surfaceth thermolysis Table1Exhaust gas properties and spray parameters for urea dosing system using UWS in automotive applicationsExhaustExhaust velocity5–100m/s Exhaust temperature400–1000K Wall temperature350–900K SpraySauter mean diameter20–150m m Injection velocity5–25m/s Injection temperature300–350Kchange in time[15,16].The variation of urea concentration of the droplet can be evaluated byd Y u d t ¼À˙m vapm dY u:(4)Massﬂow from liquid to gaseous phase is deﬁned to be negative.-Diffusion Limit model(DL model):Neglecting internal convection only diffusive transport of energy and mass are assumed.The diffusion equation for species and energy in the droplet is solved,considering variableﬂuid properties[17]:@Y u @t ¼G urd@2Yu@wþ2wþ1r d@r d@wþ1G u@G u@wþr dG uwd r dd tÀu r@Yu@w(5)@T d @t ¼a drd@2Td@wþ2wþ1l d@l d@wþr da dwd r dd tÀu r@Td@w(6)witha d¼l dr d c p;d:(7)u r denotes a radial convective velocity which accounts for the variable densityﬁeld within the droplet.The derivation of this approach has been described in detail by Schramm [18].-Effective Diffusion model(ED model):The ED model accounts for internal circulation due to forced convection.It is based on the DL model and additionally considers internal circulation by an empirical correction of the transport coefﬁcients G u and l d[19].The RM and DL models describe the physical limits of inﬁnite high and only diffusive transport.It can be assumed that the real droplet behavior is within the range of their solutions.As the temperature and the urea concentration change strongly during evaporation it is obvious to use variableﬂuid properties.In Table2the sources for the used correlations applied for the liquidﬂuid are summarized.For the calculations it is additionaly assumed that no crystallization of urea occurs. The droplets are assumed to be spherical throughout the evaporation and decomposition processes.2.2.Phase changeThe water vapor concentration on the droplet surface is inﬂuenced by the surface urea concentration.This vapor concentration has a decisive impact on the evaporation.For the calculations the correlation from Perman and Lovett[20]is used.The evaporation enthalpy is taken for pure water,because our measurements with UWS show no signiﬁcant deviaton from the values for water given in literature[23].2.3.Gas phaseFor the gas phase the quasi-steady model[15]is used.This approach is suitable to describe the evaporation process in the range of the present conditions even in a free convection situation[25]using the1/3-rule[26]for the reference values for ﬂuid properties.Integration of the transport equations for mass and enthalpy outside the droplet yields analytical expressions for the diffusive transportﬂuxes.The differential equations for droplet mass and temperature can be derived from mass and energy balances[19,15]d m dd t¼Àp D d r g;ref G g;ref ShÃlnð1þB MÞ(8)d T dd t¼À˙m vapm d c p;dcp;vap;refðT gÀT dÞB TÀh vap(9)The Spalding heat and mass transfer numbers B M and B T are calculated asB M¼Y vap;sÀY vap;g1ÀY vap;s(10) andB T¼ð1þB MÞxÀ1;x¼c p;vap;refc p;g;refShÃNu1Le(11)If boiling temperature is reached during the evaporation,it is assumed that the droplet remains at this temperature.Thus the evaporating mass can be determined fromd m dd t¼Àp D dl g;refc p;vap;refNuÃlnð1þB TÞ(12) withB T¼c p;vap;refðT gÀT sÞh vap(13)The approach accounts for non-unit Lewis number and the effect of Stefanﬂow on heat and mass transfer.Convective transport is considered by a modiﬁed Sherwood and Nusselt number using the well-established Frossling correlations[27].3.Thermal decomposition of urea particleUrea melts at406K[28]and the thermal decomposition of urea into ammonia and isocyanic acid[Eq.(2)]starts. Thermolysis becomes fully evident above425K[29,12].F.Birkhold et al./Applied Catalysis B:Environmental70(2007)119–127121 Table2Thermophysical properties of liquid phaseProperty CorrelationDensity Perman and Lovett[20]Heat capacity Gmelin[21]Dynamic viscosity Jaeger et al.[22]Thermal conductivity VDI-Wa¨rmeatlas[23]Enthalpy of evaporation VDI-Wa¨rmeatlas[23]Diffusion coefﬁcient Longsworth[24],Gmelin[21]The RM model,describing the evaporation,is extended to calculate the following thermal decomposition of urea.From literature two different ways for the thermal decomposition can be derived,as outlined in Fig.1:(a)evaporation of molten/solid urea to gaseous urea,which decomposes in the gas phase into NH 3and HNCO or (b)the direct way from molten/solid urea to gaseous NH 3and HNCO.These two ways differ mainly in the way how the heat of the reaction is provided.Furthermore,the inertia and thus the particle tracking of solid or molten urea is quite different to the behavior of gaseous urea in the exhaust stream.In the ﬁrst case,the sublimation enthalpy of +87.4kJ/mol (at standard condi-tions,298K and 1bar)yields a cooling of the particle (reaction rate r 1),whereas the following decomposition (r 2)with an enthalpy consumption of +98.1kJ/mol results in a direct decrease of the gas phase temperature.The results shown below reveal,that the thermal decom-position of urea is limited by kinetics.Thus,urea must be present for a certain time during decomposition in the solid/liquid or gaseous state.But gaseous urea does not seem to be stable at elevated temperatures,because there is no report on the presence of gaseous urea.Therefore,we assume a very fast reaction from gaseous urea to NH 3and HNCO for the ﬁrst case (r 2>>r 1).Thus,if this reaction occurs,it takes place in the boundary layer of the particle only.The enthalpy change due to evaporation of solid urea and reaction of gaseous urea is +185.5kJ/mol in total and has to be transported to the particle surface via the boundary layer.Hence,we have the same situation concerning heat transfer as in the second way where the complete reaction enthalpy of thermolysis of +185.5kJ/mol affects the particle (r 3).From energetic aspects it is not important if urea melts ﬁrst or not,since molten urea likely remains spherical,resulting in the same heat transfer condition.We modeled the direct decom-position from solid/molten urea to NH 3and HNCO and simply included the melting process with the modest urea melting enthalpy of +14.5kJ/mol in the thermolysis.As there is no reasonable condition for the phase change of urea,an alternative way as used for the evaporation of water must be taken to calculate the decomposition rate:For the description of the decomposition an extended Arrhenius expression is used,d m u¼Àp D d Ae ðÀE a =RT d Þ:(14)The heat transfer to the particle is calculated using q d ¼p D d l g ;ref Nu ÃðT g ÀT d Þþd m ud th th (15)as a result of heat balance.4.Modeling of of urea-water-solution sprayFor the calculation of the injection of UWS,the models for evaporation and thermal decomposition are implemented into the CFD code Fire v8.3from A VL [14].In Fire the UWS droplets are treated with Lagrangian particle tracking,which solves the equation of motion for parcels of droplets with identical properties using the Discrete Droplet Method of Dukowicz [30].Turbulence dispersion is deﬁned by the Eddy-Lifetime model [31].Between droplets and gas phase two-way coupling is considered for momentum,mass and heat.For turbulence kinetic energy and dissipation one-way coupling is applied.Hydrolysis of HNCO [Eq.(3)]is considered as a homogeneous gas phase reaction using the coupling interface of Fire to the CHEMKIN chemistry solver [32].5.Results and discussion 5.1.Evaporation modelingWhen UWS is atomized into a hot gas stream,the droplets are heated up and,due to the low vapor pressure of urea compared to the vapor pressure of water [21],water evaporates ﬁrst from the droplets.The evaporation of water leads to a spatial urea concentration gradient with a maximum at the droplet surface.Convection and diffusion smooth out the concentration gradients inside the droplet.Thus solved urea at the droplet surface causes a decrease of the vapor pressure of the water.This effect is heightened during the evaporation.TheF .Birkhold et al./Applied Catalysis B:Environmental 70(2007)119–127122Fig.1.Two possible ways for thermal decomposition of an urea particle with reaction rates ri and corresponding enthalpies D H (denoted for solid urea,at standard conditions,298K and 1bar):with gaseous urea (a)and without gaseous urea(b).Fig. 2.Decrease of droplet mass.Conditions:D d 0=70m m,T d 0=300K,T 1=673K,u rel =0m/s,p =0.11MPa.Calculations stopped at T s =373K.decrease of vapor pressure results in a lower evaporation rate [33].Fig.2.shows the slower decrease of droplet mass during the evaporation for UWS compared to pure water.For UWS both models DL and RM predict a similar mass change,with a little smaller rate for the DL model.The results for the droplet surface temperature are depicted in Fig.3.While the temperature of a water droplet remains constant after the heating period the UWS droplet surface temperature increases continuously.The decrease of vapor pressure at the UWS droplet yields a smaller evaporation rate.Identical heat transfer to the droplet and less cooling due to evaporation enthalpy at UWS results in an increase of droplet temperature and therefore vapor pressure.But as the droplet temperature rises,the heat transfer decreases,because the temperature difference,T g ÀT s ,decreases.Slower evaporation as shown in Fig.2is the result.For UWS a difference between DL and RM model in the heating behavior of the droplet is observed.However,there is no signiﬁcant difference in the evolution of droplet mass,although the DL model predicts concentration gradients.Fig.4shows the time evolution of the urea concentration inside the droplet,from the center (w ¼0)to the surface of the droplet (w ¼1).Since thecharacteristic heating time t $r 2d =a d %7Â10À3s is much shorter than the evaporation time no signiﬁcant gradients in droplet temperature appear.For a droplet in a free convection situation the RM and DL models predict almost the same evaporation rates.The strongest inﬂuence of diffusion resistance inside the droplet is expected at elevated ambient temperature and forced convection,because both yield in high evaporation rates.So the different evaporation models,including the ED model,are tested for UWS at these conditions.Fig.5shows the decrease in droplet mass for the three evaporation models for an ambient temperature of 900K and a relative velocity of 100m/s.As expected the result for the ED model is between the prediction of RM and DL model;as in the free convection situation,there is no signiﬁcant difference between the models.F .Birkhold et al./Applied Catalysis B:Environmental 70(2007)119–127123Fig.3.Droplet surface temperature during evaporation for H 2O and UWS droplet:effect of vapor pressure change due to increase of urea surface concentration,predicted by different models.Conditions:D d 0=70m m,T d 0=300K,T 1=673K,u rel =0m/s,p =0.11MPa.Calculations stopped at T s =373K.Fig.4.Urea concentration during evaporation,predicted with DL model.Con-ditions:D d 0=70m m,T d 0=300K,T 1=673K,u rel =0m/s,p =0.11MPa.parison of evolution of UWS droplet mass with DL,ED and RM model with forced convection at elevated temperature.Conditions:D d 0=70m m,T d 0=300K,T 1=900K,u rel =100m/s,p =0.11MPa.Calcu-lations stopped at T s =373K.Fig.6.Mass ﬂow for water vapor and urea during evaporation and decom-position.Conditions:D d0=70m m,T d 0=300K,T 1=673K,u rel =0m/s,p =0.11MPa,A =0.42kg/sm,E a =6.9Â104J/mol.Without the need of discretization of the droplet interior and without a time consuming solution of transport equation,the RM model is a numerically effective method to predict the evaporation of water from UWS droplets.Since it is based on algebraic equations,about 10times less computational time is required compared to the DL and ED model.The results reveal that the RM model for UWS is suitable for the use in a multi-phase CFD code and a good compromise between accuracy and numerical effort.In technical applica-tions,the calculation of a few thousand droplet packages is usually needed to simulate realistic spray propagation.5.2.Thermal decompositionAfter the evaporation of water urea melts at 406K and thermal decomposition starts.In Fig.6the calculated mass ﬂuxes of water vapor and urea due to evaporation and decomposition are shown.The evaporation mass ﬂow increases as the droplet heats up and decreases as the droplet diameter decreases.The small peak at the end of the evaporation occurs when boiling temperature is reached since the droplet temperature remains constant.The decomposition of urea occurs at lower rates than the evaporation of water.This is a combination of the higher particle temperature and thus a lower heat transfer,the higher reaction enthalpy (h th %3088kJ/kg >2300kJ/kg %h vap )and the limiting kinetic.The model for the decomposition rate [Eq.(14)]results in a constant particle temperature during decomposition since both convective heat transfer to the droplet and decomposition rate are proportional to the droplet diameter.As the decomposition temperature and thus the rate increases,the heat transfer to the particle decreases,until transferred heat and enthalpy of decomposition are balanced.Fig.7shows the evolution of the droplet/particle temperature during the evaporation and decomposition process.We can assume that the urea is molten since the particle stays for a certain time at a temperature above the melting temperature of 406K.Fig.8depicts the associated D 2-ratio with the different slopes for evaporation and decomposition.5.3.Simulation of urea-water-solution-injectionTo evaluate the Arrhenius parameters of the decomposition [Eq.(14)]the simulation is compared to an experimental investigation of Kim et al.[9]studying the conversion from injected UWS to ammonia.The UWS (here a solution with 40wt%urea is used)was directly injected at the axis of a tube (Fig.9)at gas temperatures of 573K,623K and 673K and average velocities varying from 6.0to 10.8m/s.The average conversion rates were measured at distances of 3m,4.5m and 6m downstream the injection,yielding to residence times between 0.3s and 1.0s.The droplets are initialized with a Rosin-Rammler distribution v ðD Þ¼1Àe ðÀD3:27=44Þ(16)an injection velocity of 10.6m/s and an UWS ﬂow rate of 3.3Â10À4kg/s.The rate of hydrolysis of HNCO [Eq.(3)]is given by Yim et al.[2]byr hy ¼c HNCO 2:5Â105e ðÀ62220=RT Þ:(17)Evaporation and thermolysis of the droplets is calculated using the RM model.The kinetic paramters of the thermal decom-position,frequency factor A and activation energy E a [Eq.(14)]are ﬁtted to match the experimental data using least squares method.The activation energy of 7.3Â104J/mol proposed byF .Birkhold et al./Applied Catalysis B:Environmental 70(2007)119–127124Fig.7.Droplet/particle temperature during evaporation and decomposition.Conditions:D d 0=70m m,T d 0=300K,T 1=673K,u rel =0m/s,p =0.11MPa,A =0.42kg/sm,E a =6.9Â104J/mol.Fig.8.Evolution of squared droplet diameter:change of slope after complete evaporation of water due to higher reaction enthalpy compared to evaporation enthalpy of water.Conditions:D d 0=70m m,T d 0=300K,T 1=673K,u rel =0m/s,p =0.11MPa,A =0.42kg/sm,E a =6.9Â104J/mol.Fig.9.Sketch of the experimental setup of Kim et al.[9](not to scale).F .Birkhold et al./Applied Catalysis B:Environmental 70(2007)119–127125Fig.10.Numerical results of spray propagation (the droplets are colored with urea fraction),gas temperature,water vapor concentration and NH 3concentration in gas phase (T g =623K,u g =9.1m/s).Fig.11.Calculated conversion to NH 3for different gas velocities and gas temperatures compared to experimental data of Kim et al.[9].Fig.12.Predicted conversion to NH 3and NH 3equivalent at different gas temperatures.Buchholz[34]is used as initial guess.Theﬁtting procedure yieldsA¼0:42kg=smE a¼6:9Â104J=mol:(18)Fig.10exemplary depicts the results of the numerical calculation for T g=623K and u g=9.1m/s.Small droplets show a minor radial penetration and evaporate and decompose faster than the bigger droplets.Therefore,the main drop in temperature and the maximum concentrations of water vapor and NH3appear in the middle part of the tube.While the highest water vapor concentrations occur near the nozzle the production of NH3occurs further downstream due to the kinetic of the thermal decomposition.Fig.11shows the calculated urea conversion to NH3 compared to the experimental data for T g=573K,T g=623K, T g=673K and varying average gas velocity.Conversion is deﬁned as the ratio of the amount of NH3measured or calculated to the maximum concentration when urea is transformed completely into NH3.The simulations agree well with the experimental data at 623K.While at573K the calculation underestimates the experiment the simulation predicts slightly higher conversion rates at673K.One should consider that the conversion to NH3 is a result from both thermal decomposition of urea and hydrolysis of HNCO in the gas phase.Uncertainties could occur in the description of both reactions.The hydrolysis reaction does not occur signiﬁcantly at temperatures below573K[9]. So conversion at this temperature is only due to thermal decomposition resulting in approximately the same amount of both gaseous products at all positions.Fig.12shows the calculated conversion to NH3and HNCO at varying gas temperature as a function of the residence time.At673K most of the HNCO is already hydrolyzed at the highest residence times.The decrease of conversion rate with increasing residence time is due to slow evaporation and thermolysis of large droplets.This effect is pronounced at high temperatures.6.ConclusionsThe inﬂuence of urea on the evaporation of urea water solution has been studied.The decrease in vapor pressure due to increasing concentration of urea in the droplet results in a continuous increase of the droplet temperature and a slower evaporation compared to pure water.Describing the evapora-tion process with different physical models,the Diffusion Limit model predicts a higher urea concentration at the surface than the Rapid Mixing model due to gradients in the droplet. Nevertheless the Diffusion Limit and Rapid Mixing models predict a similar variation in droplet diameter during evaporation at exhaust conditions.The Rapid Mixing model has been extended to describe the thermal decomposition of urea using an Arrhenius formulation after the evaporation of water is completed.The kinetic parameter for the decomposition has been determined by comparing numerical CFD simulations with experimental data from Kim et al.[9].The results agree sufﬁciently.The CFD model predicts the urea concentration and the temperature of the urea water solution droplets and urea particles,which is important for the understanding of their impingement on catalyst and wall.Furthermore,the conversion into gaseous reducing agents,NH3and HNCO,serves as boundary condition for dimensioning of the catalyitc converter, in which the reduction of the nitrogen oxides by ammonia will be conducted.The results reveal that in real exhaust conﬁgurations the urea water solution does not evaporate and decompose completely. The catalyst must have a sufﬁcient capability for the hydrolisis reaction,especially at temperatures below573K,at which no signiﬁcant hydrolysis in the gas phase occurs.Catalyst or surfaces of the exhaust gas system will be important parts for spray mixing and reducing agent preparation. AcknowledgementsWe gratefully acknowledge the discussion with Hee Je Seong and Seung Hyup Ryu from Hyundai Heavy Industries Co.,Ltd.References[1]M.Koebel,M.Elsener,T.Marti,NO x-reduction in diesel exhaust gas withurea and Selective Catalitic Reduction,Comb.Sci.and Techn.121(1996) 85–102.[2]D.S.Yim,S.J.Kim,J.H.Baik,I.Nam,Y.S.Mok,J.W.Lee,B.K.Cho,S.H.Oh,Decomposition of Urea into NH3for the SCR Process,Ind.Eng.-Chem.Res.43(1)(2004)4856–4863.[3]M.Koebel,M.Elsener,M.Kleemann,Urea-SCR:a promising techniqueto reduce NO x emissions from automotive diesel engines,Catalysis Today.59(2000)335–345.[4]H.L.Fang,H.F.M.DaCosta,Urea thermolysis and NO x reduction withand without SCR catalysts,Applied Catalysis B:Environmental46(2003) 17–34.[5]R.van Helden,R.Verbeek,F.Willems,Optimization of urea SCR deNO xSystems for HD Diesel Engines.SAE,2004-01-0154,2004.[6]J.C.Wurzenberger,R.Wanker,Multi-Scale SCR Modeling,1D KineticAnalysis and3D System Simulation.SAE,2005-01-0948,2005.[7]M.Chen,S.Williams,Modelling and Optimization of SCR-ExhaustAftertreatment Systems.SAE,2005-01-0969,2005.[8]J.M.Deur,S.Jonnavithula,S.Dhanapalan,K.Schulz,B.Raghunathan,H.Nakla,E.Meeks,C.P.Chou,Simulation of Engine Exhaust Aftertreatment with CFD using Detailed Chemistry,in:Proc.12th International Multi-dimensional Engine Modeling User’s Group,Engine Research Center, Detroit,MI,USA,2002.[9]J.Y.Kim,S.H.Ryu,J.S.Ha,Numerical prediction on the characteristicsof spray-induced mixing and thermal decomposition of urea solution in SCR system,in:Proc.2004Fall Technical Conference of the ASME Internal Combustion Engine Division,Long Beach,California USA, 2004.[10]M.A.Cremer,E.Eddings,T.Martz,L.J.Muzio,Q.Quartucy,R.Hardman,J.Cox,J.Stallings,Assessment of SNCR Performance on Large Cole-Fired Utility Boilers.1998U.S.DOE Conference on SCR an SNCR for NO x Control,Pittsburg,PA,1998.[11]P.M.Schaber,J.Colson,S.Higgins,E.Dietz,D.Thielen,B.Anspach,J.Brauer,Study of the urea thermal decomposition(pyrolysis)reaction and importance to cyanuric acid production,American laboratory(1999).[12]P.M.Schaber,J.Colson,S.Higgins,D.Thielen,B.Anspach,J.Brauer,Thermal decomposition(pyrolysis)of urea in an open reaction vessel, Thermochimica Acta424(2004)131–142.F.Birkhold et al./Applied Catalysis B:Environmental70(2007)119–127 126。

1、下载文档前请自行甄别文档内容的完整性，平台不提供额外的编辑、内容补充、找答案等附加服务。
2、"仅部分预览"的文档,不可在线预览部分如存在完整性等问题,可反馈申请退款(可完整预览的文档不适用该条件!)。
3、如文档侵犯您的权益，请联系客服反馈,我们会尽快为您处理(人工客服工作时间：9:00-18:30)。

Chapter 7
Acquisition and Restructuring Strategies
Michael A. Hitt R. Duane Ireland Robert E. Hoskisson
©2000 South-Western College Publishing
Ch7-1
Strategic Inputs
Firms may use acquisitions to restrict its dependence on a single or a few products or markets Example: General Electric’s acquisition of NBC
Ch7-6
Problems with Acquisitions
Strategic Actions
Chapter 8 International Strategy
& Innovation
Outcomes
Strategic
Feedback
Strategic Competitiveness Above Average Returns
Ch7-2
Mergers and Acquisitions
Integration difficulties
Overcome entry barriers
Cost of new product development Increased speed to market Lower risk compared to developing new products Increased diversification Avoid excessive competition
Acquisition
A transaction where one firm buys another firm with the intent of more effectively using a core competence by making the acquired firm a subsidiary within its portfolio of businesses
+ Maintain Financial Slack
Provide enough additional financial resources so that profitable projects would not be foregone
Ch7-9
Attributes of Effective Acquisitions
Ch7-5
Reasons for Acquisitions
Increased Speed to Market
Closely related to Barriers to Entry, allows market entry in a more timely fashion Example: Kraft Food’s acquisition of Boca Burger
Inadequate Evaluation of Target
“Winners Curse” bid causes acquirer to overpay for firm Example: Marks and Spencer’s acquisition of Brooks Brothers
Large or Extraordinary Debt
Merger
A transaction where two firms agree to integrate their operations on a relatively coequal basis because they have resources and capabilities that together may create a stronger competitive advantage
Entrepreneurship
Strategy Formulation
Chapter 4 Business-Level Strategy Chapter 7 Acquisitions & Restructuring Chapter 5 Competitive Dynamics Chapter 6 Corporate-Level Strategy Chapter 9 Cooperative Strategies
Chapter 2 External Environment
Strategic Intent Strategic Mission
Chapter 3 Internal Environment
The Strategic Management Process
Strategy Implementation
Chapter 10 Corporate Governance Chapter 12 Strategic Leadership Chapter 11 Structure & Control Chapter 13
Inability to Achieve Synergy
Justifying acquisitions can increase estimate of expected benefits Example: Quaker Oats and Snapple Acquirer doesn’t have expertise required to manage unrelated businesses Example: GE--prior to selling businesses and refocusing
Ch7-11
Restructuring Activities
Leveraged Buyout (LBO)
A party buys a firm’s entire assets in order to take the firm private. Example: Forsmann Little’s buyout of Dr. Pepper
Diversification
Quick way to move into businesses when firm currently lacks experience and depth in industry Example: CNET’s acquisition of mySimon
Reshaping Competitive Scope
Continue to invest in R&D as part of the firm’s overall strategy
Ch7-10
Restructuring Activities
Downsizing
Wholesale reduction of employees Example: Procter & Gamble’s cutting of its worldwide workforce by 15,000 jobs
+
Low-to-Moderate Debt
Merged firm maintains financial flexibility
+
Flence at managing change and is flexible and adaptable
+
Emphasize Innovation
Ch7-4
Reasons for Acquisitions
Increased Market Power
Acquisition intended to reduce the competitive balance of the industry Example: British Petroleum’s acquisition of U.S. Amoco
+ Friendly Acquisitions
Friendly deals make integration go more smoothly
+ Careful Selection Process
Deliberate evaluation and negotiations is more likely to lead to easy integration and building synergies
Takeover
An acquisition where the target firm did not solicit the bid of the acquiring firm
Ch7-3
Reasons for Acquisitions
Increased market power
Problems in Achieving Success
Overcome Barriers to Entry
Acquisitions overcome costly barriers to entry which may make “start-ups” economically unattractive Example: Belgian-Dutch Fortis’ acquisition of American Banker’s Insurance Group
Inadequate evaluation of target
Large or extraordinary debt
Acquisitions
Inability to achieve synergy Too much diversification Managers overly focused on acquisitions Too large
Integration Difficulties
Differing financial and control systems can make integration of firms difficult Example: Intel’s acquisition of DEC’s semiconductor division