Document GknndOR1VxK93685GYDbE9Zv

!- , V " DISTRIBUTION REACTION PRODUCTS IN : BENZENE CHLORINATION . i' BATCH VS. CONTINUOUS PROCESS PROCEDURES . ' .. ROBERT B. MacMULLIN . ' R. B. MacMullin Associates, Niagara Falls, New York /CHLORINATION of hydrocar- ination products which result in the he is aware of the penalties involved, Ldwns is a unit process * in chem substitutive chlorination of hydrocar proven capacity and performance ical engineering which is of consider bons. A later paper will deal with may be in practice very different able importance. Monochlorobenzene addition reactions and the compli from what he designs on the basis of is produced commercially on a large scale as the essential intermediate in the production of phenol, aniline, DDT, and a host of other organic intermediates. Dichlorobenzene is manufactured on a large scale, the para-isomer being used as a larvicide and the ortho-compound, as a solvent ; ji'id heat-exchange medium. Chlor';;,,lition of toluene.is a primary unit process in the preparation of benzyl alcohol, benzaldehyde and benzoic acid. Chlorination of ethyl benzene is a primary unit process in the prep aration of dichlorostyrene whose jKilymers and copolymers are of con siderable interest. Chlorination of phenol on the one hand and the chlor ination of acetic acid on the other, are primary unit processes in the prep cated problem of parallel chemical batch-chlorination data. reactions which compete for chlorine The purpose here is to show how during the chlorination process.f The the whole family of curves, which mathematical treatment which fol represent compositions of the various lows has been deduced in general components as functions of the terms so that it is of general applica amount of chlorine introduced, de tion within the limitations of the pends solely upon the relative reac assumptions which are clearly stated tion rates of the compounds being in this paper. By way of illustration chlorinated. These relative reaction the method has been applied to the rates may be obtained by a single chlorination of benzene up to and in experiment preferably carried out at cluding trich lore benzene. the point which represents the max- ' Distribution of. the various substi imum concentration of any particular tution products is of considerable im- , component, one experiment to a max portance to the chemical engineer. He imum. Moreover, the degree of I? it : is usually interested in producing chlorination at which the maximum certain of these products to the ex will occur can be predicted within elusion of others, or to produce them very close limits. The theoretical ip: M in definite proportions, depending composition curves can then be de- r aration of the important weed-con upon economic consideral ions. He duced and extrapolated with confi trol compound, dichlorophenoxyace- 'usually approaches the problem of dence into regions where experimen- .. tic acid (2-4-D). Chlorination of plant design by demanding from the tal technique would be difficult." aliphatic hydrocarbons is carried out research department experimental There-will be a separate family of industrially on a large scale to pro data showing the distribution of the curves for batch chlorination, another \ ,`!Vduce a variety of chlorinated solvents successive chlorination products as a for single-stage continuous chlorina and intermediates. The present paper is a contribution, function of the amounts of chlorine introduced. This will usually be fur tion, and another for two-stage con-, tinuous chlorination. Each family of ai) tr. to the technology of chlorination as a nished by the research chemist on the curves is determined by the relative !? j ; unit process and is concerned with basis of experimental chlorination of reaction rates which, as pointed out, jt I , the distribution of successive chlor- the hydrocarbon in question as a can be determined by a surprisingly '>!>' . batchwise process. The data may be limited number of experiments. * Defined by R. Norris Shreve. t Since writing this manuscript the au thor has learned that an exhaustive investiration of this subject has been made by of questionable accuracy, and the chemical engineer may wish to extend the experimental curves into a region In the treatment of the subject, the plan will be first to set up.the funda mental equations; second, to state the if.: it. Dr. John W. Churchill (Mathieson Alkali beyond the range of the data. The final equations in usable form; and Works, Inc.), and that a paper is in whole design of the plant including third, to illustrate their use by apply (reparation. t Batch chlorination of benzeile has also _ mathematically treated by M. F. reaction vessels, fractionating col umns and auxiliary equipment, will ing them to the chlorination of ben zene. Mathematical derivations of ,i ourion, Annalct de Chcmie (9), 14, 215 depend upon just such considerations. the final formulas have been omitted. 1 (1920), but the results are presented The chemical engineer may get the So far as known, these equations in less usable form. His. experimental re- ' mlts agree well with those presented in this paper. - - . . . . > "bee in his bonnet" to des gn a con tinuous chlorination plant, and unless have not appeared elsewhere in the literature.^ .. .. Vol. 44, No. 3 CHEMICAL ENGINEERING PROGRESS DSW 297235 Page' 183 STLCOPCB4067145 -V-, Table 1.--Batch Chlorination of Benzene , Calculated Compositions . V -. II 00 6* II S3 Dj : j;. . ' r. a . , It 1 1. l;,-: 1 0.000 . 0.414 1.000 ' ` . .600 0 .386 0.760. .300 .640 ', 1 1.057 . .100 .743 ' .*{ ,i 1.172 1.349 1.595 '.050 . : ' .010 . .728 .631 .422 : : 2.045 2204 hi' 2.768 ' .038 .024 . .005 . C 0 .014 . .060 .157 332 .359 .561 .879 .748 .222 D ( 0 0 0 0 0 1 .... .017 .083 .228 .773 ` > * Method , ' . Equation 20 , .' '. . Equation 21 v = .70 . v = .96 v = .97 v = '.98 ' ,l:i i'1 :!. - :r: : I. J >f. r 0.740 0.890 0.970 .316 .207 .155 1.140 .063 1.477 . ' :;! :! :i 1.715 1.910 1.918 V HV-. 2.090 ii::r ' Observed Compositions .627 .697 .725 .735 .564 .320 .140 .130 .025 .057 .096 .123 .203 .437 .646 .820 .825 .855 .013 .034 .044 .046 .125 sp. gr. 30/20 , 1.050 1.089 1.105 1.146 1.211 1.250 1.273 1.275 Assumptions 1. Conditions are such that chlor ine enters the benzene ring by substi tution of hydrogen only. 2. Reactions are irreversible. 3. Relative rates of reaction are proportional to the relative molal concentrations of the components reacting. 4. Relative rates of reaction are referred to a specific set of reacting conditions, such as temperature, catalyst. 5. Distribution of components is independent of the speed with which chlorine is introduced. 6. Concern is with the distribution of mono-, di-, tri- and higher substi tution products without regard to the proportion of isomers in any such product. 7. The course of the reaction is A-* B~* C^+D . .. where A -- benzene B -- monochlorobenzene C = diehlorobenzene D =? trichlorobenzene \:M\. 90 ratch___ V ' jltf V 60 vtjl `=13:1.1 riil i Iv i! ' 1 )k!* Ii m m "I !| .? I':-H h 70 Z IaJ o ce to CL 60 111 -o1 z 50 ' - - . r r.oNf^ too5 ----- \ / - . --; /^ ' ,' - . '' ' : ; '. MAXIMUM * CONCENTRATION J '' . -4 . ; OF FIRST ' REACTION PRODUCT 'Ml-:. 40 ;l -i. ; ii :-!k i ; 0*5 : ; f / /' 30 11 20 t `V 4 / 12 ' . - MOLE ^PERCENT vs P . V. . . * :^ ' - ' ' - - Figi 1 - . 1 1 C . '> i ' . ' . ` . ' - . ' '. ,. / ' f ! .\ 16 20 K, r= Kt 24 28 DSW 297236 STLCOPCB4067146 g The accompanying mathemat- ,We will now consider these cases in Differentiating Equation (12) ical treatment-is limited to the chlor- order. ' ; with respect to x and setting. ination up to, but not beyond tri- . ; Batch Chlorination ^ V dB n . : . V- 9. In continuous single or multi- ' stage reaction systems, the liquid .. composition within each stage is assumed to be uniform and constant. The linear homogeneous equation ax - . . , ' dB ' B -- rA' we obtain the maximum value of B* . - = . - , (7)' ;^ follows, , : . . . : r/l .. -.;V - . .. -., . iv _ , "''X .. r'integrates readily to . :. Bm -- r */(r X) (1^);'. .. ` , e g . Fundamental Equations ' .** 'AB --=mmool,lee f,rfaracct.i.toionn.bemnzoenneochloro: v-.{:' p -- r--,,r 1 'r ai/t _ a)y , (/o)\ ... > . .;.ia<ntgwvhailcu.he mo_rfajx.r.imisum, , the . -; corresrpond- V. i'; <7= mde^fSction dichloroben-' ^ >?- for ^uzene, D is neg-^; . zene . hgible as long as A is appreciable, - 2 - Bm( 1+^-Y\ \ ' V . <y.. . , \ rj ; r- D mole fraction trichloroben-;;. y ' zene ; ' // ; ' ; x -- total atoms of chlorine per , ; . moje of benzene /constit- <. 2A + 2B'+ 2C -- 2 , ' ! ' \ B + 2C -- x ' ' V 2A + B ~ 2--x . . . . /iay . ' 2= -- (14) '. uent. . . . f..; - ' .. : . Equations (13) and (14) are , A+B + C + D = 1(1) . Combining with Equation (9), .. ;* evaluated in Table 4 and are plotted. ' . ' -B + 2C + 3D ='x\:V;(2)'".l I, / r-2\ / r. '-'v. Hin Figures 1 and 2. Inspection re- aa' V ' . ----=.-- kiA* . : ' V ~-- 1 }^1/r veals that for all values of r between ' /. ' 4 and 32, the maximum occurs within the narrow range of x = 1.05 to 1.07. 4: Ia ' ' dB , , ri - ' ' r r/2-- B -- *-W* . dt = M " kzB^r'JA). ^ :B = A.. ^ , A single experimental analysis for B at thif critical value for x is suffi- dC = k2B -- k&C /. v ;'V C:-'(5) r,L V` . . *-.> ;; . ' ( 2-^- B -- x\~\ dt ,. /: ' V 2 JJ ' . cient to determine r by means of ,''/10V. Equation (13) or the plot. Figure 1. , [ ' From this value of' r, A', B, and' C" dD dt =,kzC / -(6). r = JL- . Let k\/k2 ~ r and k2/k3 -- s ... . . HO 10 15 20 25 30 35 t, . From Equations (3) and (4) we obtain .. . ' :; '. .I ' 3]n - . r. ; J- ! &i .I, ; it V; s-.1 . l,\ !M dB _ B -- rA dA rA .. (7) 1.05 ^ BATCH _ - : (' . ii*r:. i From Equations (5) and (6) we ' obtain . V dC sB -- C Tu c~ sq\ (8);, There are five unknowns--A, B, 7- C, D, and x, and four Equations (1), ' (2), (7), and (8), from which it is A 'desired to express A, B, C, and D as functions of x. - . , ' Equations (7) and (8) apply, in differential form, to batch chloritia-^O, LOO .95 90 1 ' 1 'v ' * -e r.OMTII uous sv . ST- : < ' r' . ' * r' ' ' *1. . ' _ ` -. . I.'; 1 lit! tions, in which the components vary , ;, :`x* i V* with respect to time, and to \ continuous chlorination, after: a' -My05 1 i nr ATirIN nc ` ' " i -.'i uaVima '' i steady state is reached, ,r /' iV . ^ ddIABi -**B1BAa findA. dC_C dD D ' ... ,80 It is clear therefore that batch ' ** . ' `" " / ...Values OF. X1 FOR WHICH COMPONENT/ i i . ' .; 1;. - .. , . A MAXIMUM " FOR VARIOUS ' VALUES 1' , chlorination will result in a different distribution of products than cOn- ' : "of r i r'',-: CO tinuous chlorination, when compared ' at corresponding values of 'x. Like- wise, continuous chlorination will be-|.;-., 75 ; ,1. . Fig. 2 . V " a> effected by the number of stages into ' which the reaction system is divided. .70 o *ln Equations (3)-(6), chlorine con-.- ; centration has been omitted' for the sake: of simplicity. - It cancels out in Equation , (7) and (8). ; 65 ' 'P. . 1 .. ' Vol. 44, No. 3 \ . CHEMICAL ENGINEERING PROGRESS . 1' Pago 185 STLCOPCB4067147 '/bz readily calculated for all Table 2.--Single-Stage Continuous Chlorination of Benzene Values of x up to about 1.5, or the ' ' place where D becomes appreciable, .vf Noting that when D becomes ap- ; V redable, A becomes insignificant. , x 0201 v-V A .713 Calculated Contpositions t:= 30 BC D Method 274 .0136 ..0000 Equation 25-30 inch , ' B +' C .+ 1 . v/;1' x : ,-. ', x-\ * 0.485 :'. ...556 '.:.404 1 .0404 .0001 : 0.720 " *' -.3845 . - .5125 :: .1025 - . .0007 1X74 ' . -.1990 . .5301 . 2654 .0045 1.358 v.-.; . -.1095 . .4375.'* .4375 ' .0151 1.615 s - .0558 : .2974 .595 V. .0425 Equation 25-30 inch Equation 25-30 inch Equation 25-30 inch . Equation 25-30 inch Equation 25-30 jnch, 2.'- ' 'V.: 1.960 > 2232 '. : X . ,' 2.635 . . .0210 .0085 .0021 .. .1400 .700 E .1400 .0620 .620 ;/' .3100 .. .0164 2275 ' .6545 Equation 25-30 inch Equation 25-30 inch Equation 25-30 inch B -D = 2- x imi whence Table 3.---Two-Stage Continuous Chlorination of Benzene . -c i : ; C= r-\-2D + x c (15) Calculated -Compositions, Second Stage B = ' 2+ D -- x (16) , : X :A r=8 BC d ;; ' ' s = 30 Method From . (15), ' .* 4 ' .; j/C, , _ dx 1 dD - dD . (17)! Substituting these values- in the previous Equation (8) ' dC sB-C 0.500 : .750 1.000 1250 1.522* 1.730t 2.000 2250 2200 : .535 237 .187 .094 , .039 .010 4 ` . . t .430 : .035 .577 .086 .627 .186 .563 243 .415 231 288 -- .661 .124 ; 752 .039 .672 .007 . .486 . .. .V '. . *, v ` .' .015 .041 J .124 289 , .507 Equation 38 B, = .235 Equation 38 B, = 234 Equation 38 Bi = .416 Equation 38 Bi = .479 Equation 38 Bi = .520 Equation 42 Ci = .164 Equation 42 C, = .228 Equation 42 Ci = 296. Equation 42 Ci = 270 C, = .008 C, = .020 Ci -- .041 Cl = .071 Cl = .110 D1 = 0 V 23.= .002 V 23. = .004 . 23, = .010 dD. '* Originally calculated for x = 1.5, assuming no D, then adjusted. ` t Originally calculated for.X -- 1.75, assuming no A, then adjusted. we obtain dD X-2D-1 dx (1 -s)x+ (s-2)D + (2s -- 1) v: ; Data are recorded in Table l and1. (18) . plotted as large circles in Figure 3., . From these data7 the maximum con- ; : The above equation cart.be integrated by letting z -- -D--- -- 1 . , . centratiori of monochlorobenzene -: .' .. was judged to be Bm -- 0.745 at x \X; = 1.070, from which r = 8.0 by;-. . v~ 1 JI-.S [',+ -.*,1 '* + !-. ................. ` a( -1) = 3+K s~^ I [1+ (s-3)v-(j-2)^]~% Equation (13) or Figure 1. ; The maximum concentration of di- . chlorobenzene was judged to be " Cm = .885 at x -- 2.(40, from which . (19) s = 30(1) by Equation (13) or In the above equation, K is the in-. tcgration constant. We note that D is readily calculated, and at these when 'X = 0, D = 0,. and v = values of x, the values of B and C from which K can be readily calcu \ follow from Equations (15) and Figure 1. '. .. . Theoretical Chlorination , Curves Using the above values, qf >: and s, Equation (9) becomes . >''yv-v- lated. If experimental data are avail ; (i6).* _ able in the range of x = 1.5 to 2.5, it Experimental Benzene was chlor B = ~- (A*- A) C ,(20y is preferable to .determine K by sub inated in a 2 1. flask provided with stituting known values of D and x. - excellent agitation, and a reflux con- rand Equation (19) becomes In Equation (19) s is the relative re denser. A temperature of 55 C. was ' action velocity of mono and di,. and may be estimated by noting that where C is a maximum! ; . s= C,,_ B maintained. Anhydrous ferric chlor ide was used as a catalyst. The amount of chicline introduced was judged by specific gravity. A series of runs was made, each tc a definite X' = 3 + v+ .0357- [i+27v-28v*]-* (21) V, When x -- 2.0, D observed = .065, specific gravity. The product from l ean also be estimated with only: each run was settled, then fraction slight error from the maximum value ated in a 10-plate laboratory column. itself (see Figure 1). ' Product was divided into cuts so Having thus determined K and s, the value of x at arbitrary values of that in each cut not more than two components were, present. Composi whence v = 0.935. From this, we evaluate K = --5.642. From the above, A, B, C, and D were calculated for values of x rang ing from 0 to 2.5, and they are tabu lated in Table 1 and plotted in Figure o CcOv CD C\| *In the foregoing analysis, Equations. tion of each cut was estimated by (7), (9), and (13) are strictly valid for all specific gravity from previously de nines of r. On the other hand, the simpli- _ tying assumptions made in deriving Equa-'. twn (19) are valid only for values of r termined density-composition plots for the s.ystem in question. In this *nd t greater than about 4., When higher- ' way an accurate material balance was nction.products appear in appreciable per- ^obtained. High boilers were driven WiU.ges in the early stages of chlorination, .owr bv the llsp nf a <Vbn^r " i,e.(j f *nd i obviously have lower values, and a ' ^rc` "7, usc. * ^cr' pre rigorous mathematical* treatment'is-..- diphenyl. Results are judged- ac-.' dicated ' . . . curate to 0.5%. . 3. It is noted that the observed ex perimental points in all cases lie on the theoretical curves, within limits of experimental error. The curves are not exact beyond x = 2.5 because of the appearance of tetrachlorobfen- zene, which has not been taken into account in the derivation*of the for mulas. :: .! ->- (0 Q ' ' '* ' Vol. 4A. No. a 1 ' V f'HPilir'AI CMrilMCCDikir? tnnrnc^r. STLCOPCB4067149 V Table 4.--Location of Maximum for Component B Batch Single Stage Continuous is therefore less favorable than batch chlorination at any given value of x, but can be made as favorable by chlorinating to some lower value of x and recycling the underchlorinated Combining Equations (36) and (37) with (33) and solving for By. i (2 --r)Z?22+ (2*2 + 2r-4C, . -2Cir-4Bx)B2+[rx22 \ r . - , P- 0.5 , .250 1.0? .368 Xm . '.750 - .896 ' Xm .172 .656 .250 .750 materii The ben::en -- (2r + 2Cir)* + 4Cir] =0 (38) which can be written 2 .500 1.000 .343 .829 : P`B2s + Q-B2 + R = 0 (38a) .! .1i fci- i! : -J 4 9 16 25 36 .630 .760 . .832 .875 .903 1.055 1.070 1.065 1.055 1.047 .444 .563 .640 ,695, .735 ' .889 .937 .960 .972, .980 at ^Yrt Bm =: (>% + !)* (31) (r% + l)2 1 (r + l)2 whence _ -g + VQa-4PP P2 = 2P (39) . . Continuous Chlorination Single Stage: The fundamental equations are .. A + B+ C+ D = 1 (1) . \ ' B + 2C + 3D -- x (2) (32) The above maxima are evaluated in Table 4 and plotted in the lower curve of Figure 1. No experimental data are offered in confirmation of these curves.. However, in commercial operation of continuous lk-uid-phase chlorination In solving this equation for a de- f finite value of x2, one has only to read j off the values of B\ and Ci at Xi = j \A.x2, from the single-stage contin- : uous curves which are plotted in Figure 1. . i Case 2--A negligible, for x over 1.75 iii c. -\ B rA-B . . Xr.i . -c ' .C _ sB-C - \D C (22) (23) plants making monochlorobenzene, the usual operating point is such that the chlorinator product contains ..about 40 Wt. % monochlorobenzene,. 5 Wt. % dichlorobenzene, balance I?2 + C2 + D2 -- 1 (40) - B2 + 2C2 + 3D2 = x2 (41) Combining Equations (40) and (41) with (34) and solving for Ps: /ii'i,!!-'iI The above set of simultaneous equations are most easily solved in terms of m, the allowable ratio of monochlorobenzene to dichloroben zene in the reactor product. That is. benzene. Such an operating point lies close to the predicted curve. (x = .387, A -- .645, B = .325, ' C = .031) (s-2)D22+[(3-s)x2 - (s + 2)Dt + (2s - 3 - C^D) + (--x2 -f- (2 + Ci + Djs + D{)x Tzvo-Stagc: The simplifying as - (l + Cx + 2DiS + DX)] =0 Let m -- ~ (24) sumption is now made that an equal _ : (42) `'i!. 1 S: i i.f'J " then a -- 'm(m + 1) (sm -- 1) Q (25) B _ rm(sm -- 1) _Q (26) amount of chlorine is introduced in each stage of the reaction system, while the liquid flows continuously through the first and second stage in series. It is further assumed that mixing is thorough, and the composi tion uniform and steady in each ves sel. Denoting the stage by subscripts, which can be written , P'-Z>22 + Q'-D2 + R' = 0 whence (42') -Q' + yfQ'2-4P'R' D2 2P' c__r(sm_-- 1) , (27) Equations (22) and (23) become, for the second stage, As before, read off C\ and Dy at X\ t.i-':;! .. . I ; . d = 1T (28) B2 -- B1 C2 -- Ci rA2 -- B2 B2 -(33) ~ Vzx2 from the single-stage con tinuous curves already plotted in i : I'l ! k: ! 'i :p;i' `>w: '`i -^ii.l'. * -Id. /Li ;i _ r[ (sm -- 1) (m + 2) *f 3] -A ` where (29)' C2 -- Ci D2 -- Di sB2 -- C2 (34) C2 x2 = 2xy (35) This set of equations is not so Figure 1. ' In the overlapping region between Case 1 and Case 2, where x = l.S to 1.75, the above equations are not strictly correct, but may be adjusted tej take account of the small amount ' :i.- j : ! ; l\ Q = m(m + 1) (sm -- 1) easily solved as those for single-stage of D or A present respectively chlorination. The biquadratic equa Calculated values of A, B, C, and + rm(sm -1)4- r(sni -- 1) + r tions that result are moreover incon D are tabulated in Table 3, and are (30) venient to use. We therefore divide plotted in Figure 3. the solution into- - tw.o cases: , 1 It is seen that two-stage chlorina i Putting in tn at values ranging from 20 to .05 compositions are ob Case 1--where D is negligible tion gives, a distribution of products intermediate between single-stage V* -:;T I. J? :ilf ) : I tained as listed in Table 2, and plotted in Figure 1. It is readily seen that continuous chlorination results in a greater pro portion of higher chlorinated prod ucts than batch chlorination, at all - Case 2--where A is negligible . Case 1--D negligible, for x up to about 1.5, - , A2 + B2 + C3 = 1 (36) chlorination and batch chlorination. The method can be extended if de sired, to calculate the distribution for a greater number of stages than two The greater the number of stages, the closer the approach to batch distribu J; values of x. Continuous chlorination ' B2 + 2Ca = *2 (37) tion. T DSW 297240 STLCOPCB4067150