Selective Demethylation of Paraffin Hvdrocarbons J
PREPARATION OF TRIPTANE AND NEOPENTANE
Braricliecl-c*hain paraffin h?drocarbonc redc't w i t h ti, drogen in the presence of a niche1 or cwbalt c8atal;rst I O s i \ e paraffins c-oiitaining one less carbon atom arid methane. The reartion in\ol\e* the renioFal of specific r i i e t h ; r l groups since those at tac.hed to cecoticlar? carlwti atoms are remo\ed much m o r e readil? than those 1,oiititl to the tertiar? carhon a t o m - , \+bile nieth?l proupc a t tarhed t o the quaterriar? c,,irl,on atoni- arc m i ) - t stahlt..
the reac.tion o( vur- within narrow ranges of experimental cwnclitions. Data are presented on the conver-ion of inixtures o f 2,2.~-trinieth?lperitaneatid 2,J,&-trimeth, I pentane to t ripta tie, 2,3-dimeth>1peri t anr, .ind 2 , k l irneth;rlbutane. R e - t i l t s are a150 giieri o n the tlenicth? lation of iieolieuine t o neopeiitane. The effect o f operating \ariables such SI- pres-lire. temperatiire. a n d c p a c ~\ c . l o v i t y i - diwit-ee(1.
.. .
rure plat?, an(I :A. valvr, m d is finslly :ttt:ic~hi~tlt o t h i s iri1i.t i d the reaction t.ul)(.. liydrugc~ri,supplied friirn :L rylinder, gotw t liri~iigh a pressuri' icbgulator, a fine-adjust inelit f1oic.r-tppc v:tlvt., :ttitI H rotanirter before joining with the hytlroc~itrl~riri strwtii :it t I I C i i i 1c.t t o the reaction tube. The reaction rubiii:tiiIy o f 100 cc. The effluent from the reaction rul)c: csnters :t j ~ r ~ ~ - - r i r ' c ~ -cparator, the uncondt*n>edgas going to a p r t w u r c rot11I.oIII.I., \\-bile thtl liquid drains intermittently into a glass ri*i,(%ivi,i,.'1'1112 r,fflucnt,from :he pressure controller g i x s through >L d r y i o ( \ i r : t p anti finally discharges after passing through a wet, t c s t i n i ' t t > t ' . In starting an iispcriment it was founil neccsssrp t,o UBI' rvinlii'rtttures below the c1esirr.d final catalyst temperat,ure. Thi- i,. i l u i , to t,he fact that the reaction is evot,hermic and, as a rule, tho ~ ' : t t * lyst, temperatures are some 10-20" F. higher than the ja4ict temperature. Furthermore, as will be shown hirer, the react'ion is confined to a relat,ively short range of ternpcbrnturc. for any o i i c fixed set of conditions, and it is desirahle to itart, with vcsry l o 5 v
13)
14)
853
854
INDUSTRIAL AND ENGINEERING CHEMISTRY
-5 d
t
-
e2
.---
--
? h
Vol. 39; No. 7
ions and incrc~asi.the t,enilwi,ature in tliic. gritdually. The catalyst eiiiployc~~l inveit igation was niclwl on liiovlyuhr, knon-n as the stanclard UOP 1iyd1,ogi'natioii catalyst ( 2 ) . Prior to t,he usc' oE,tIic. catalyst, it, was found advantage w h j e c t it t o reduction with hydrr approsiniatc~ly1000" F. The liquid product i. b l r n d ( d \vi t 11 t 1 1 t h tlry ice ccindrnaatc and subjected i i ) fmrt ionat ion. The nie~thodof analy-i.; vxricv i t h the charging stock suhjcctcd t o dt~nic~thylation,but a s a rule, infi,nrcd , tlc.te~rniinationsarc niatle of a nunitwr of ruts. In the case of neohesane tlc~nic~thylation, the liquid product is suhjc.ctcd t o a low temperature Podhiclniak distillat ion. The gaseous product.? are ana1yzc.d by :I modified Goeckel niet,liotl. The major amount, of work has lieen tlonr. on the preparation of triptanc using a> charging stock a hydrogc.nat,ed po1ymc.r prntlucetl by the hot-acid (sulfuric I prorti.-> (31.The hot-acid isooctane rontainh t 11~'four trimc~thylpentanesas tlirs major c.on.stituents along with sniallcr amount. of d i n i e t h y l h e s a n e s . S i n c e 2 , 2 , 4 t ririic't hylpcntanc, on deiiietliylation, yic~ltls 2,2-diini~thylpentaneas t h e niaiii p ~ ~ x l u and c t the boiling point of thv 1atti.r is about 1" C ' , loir-er than that of triptane,, thc. preaencf, of 2,2,~-triniethylpc,1it:lnc:in the, charging stock is undcsira1)lc from the standpoint of ultimate purification
PC
of t,hc triptane: fraction.
-.-
The hot-acid isooctane ' ~x~niovc, first, the 2,2,1-t1 followed by a cut representing a conce~i-
c
_. *
r
trate of 2,2,3-, 2,3,1-, and 2,3,3- trinicthylpentancs. The volume of thi. cut ' wlativc t o t l i c s initial charge dctc~rmiric~i tliil coiicvntration of three isomc,rs i n the tlii. clesircd fraction. Thus, t h e 2,2,3ti,imethylI)entan(, concentration c a n I)? nintle t o vary from about 30 to :rliout (io';, depending upon \r-hether it n:irro\vOI' : t v-ide-boiling cut is u d . PRODUCT DISTRIBL-TIOS
Tiit, denwthylatiori rc.ac.tion caii tic 1ooi;cd upon as a continuous pr(ic[h+ of
tic~tarhingmtxthyl groups; as long 'ti any cxr\ion-carbon bonds riLmain. tlicre is n tcsndvncy for tlenic~thyltition to ploc~cetl. 011riously it is dcxsirahle to stop the rctict ion a t the desired stage of tlc~nic~thylation: otherwise, a 105s of product \vi11 orcur. This means that thr: ojxxrating cwnditions have t o i)e adjusted so th:tt the runversion per p a s dot^ not r c w h esive values. This point ih impe~rtant betcause, although the heat of tlc~nicthylatiori of an octane to a hciptane is Iow(~r than the hcat of hydrogc~nation o l d n , in the cai.c~ of conipletr. con\ of an octane t o nic~tlianc~ there iy an cstrc~nic~lyhigh vvolution of hcat. I t ii i'ortunatc, ho\\cvcr. that iii t h i s cases uE the
INDUSTRIAL AND ENGINEERING CHEMISTRY
July 1947
855
ratio, the logarithm of the operating prcwure is proportional to os PRODUCT DISTRIBUTIOXthe catall-at tcniperat~urc~.This particular reaction appcw3 to be T A ~I.E EFFECTOF CONTERSIOS 1
Run S o . C a t a l y s t t e m p e r a t u r e , O F. Pressure, ib. a q . ill. gage Hydrogen-hydrocarbon ratio Hourlv !iquid space velocity Duratyon of run,hour? 1,iquid r e c o v e r y , vc by vol. Cr,nver.ion t o riisterial boiling beloir95: C , ( 7 by vol. P r o d u c r diarributioii, b y \VI,: Cut 1, initial b . p . t o 54' C. C ('ut '2, 54-70"
cut
4.5 96.0
491 100 4.i 0.57 6.0 94.0
34.1
51.9
66.8
7.2
i .3
12 9
9.0
13 7
16.i
53.2
62 8
61.6
30 6
16 2
8.8
C
C .;C-C!-C-Cs
:3, 7n-850 c'
3
2 505 100 2 5 0.59 5.5 94.0
460 100 2.3 0.60
c c C-c-c-c, c;
pcculiar t,o the denietilylatioii r,c%ction, and differs f r o m the hydrogellation of olefins where a higher pressure alloivs onc to operate at a l o u w temperature. Sonic atlditional ti with a cobalt catalyst iiitlicate a similar relation lwtn.c~eiiprcssure and tenlperaturc~. Figure 4 s h o w the relatii~nhiitivcen conversion ailti wtalybt tcxniperaturt: and Figure 5 correlateo space velocit,y with c;~t:ilyst t.c~nipcratulc. Thi:, series of expcrimcntl; \vas m d e :it a total p r t ~ ~ ~ uofr e100 pounds per square inch gage. .klditioi iorniation on the effect of catalyst temperature upon cox at a nuniher of pressure. is given in Figure 6. At pressure's in the range 300-800 pounds gage, the logarithm of conversion i-: proportional tu the catalyst teniperat,ure.
c c Cur 4 , S5-95"
-drocarbon. , it is possible t o at'tain high conversions pi'r pass nithout appreciable loss of desired product. 'r:ltJli' 1 illustrates the effect of extent of conversion upon product di.jtrii)ution. The charging stock for this set of esperinients i.onsist,td of 59?> 2,2,3-trimetliylpentane and 4lC; 2,3,4-tring at 107-113" C. These results indicate on is increased, t'he triptane cut '(70-85' C., triptane concentration in excess of 83%) is not affected, while the 2,3-dinietliylpt,iitaii(! i y gradually converted into 2,3-dimethylbutane and linvi~i~ ~iiiilingdemethylated products. Figure 2 s h o w a high t o n i i x ~ l ~ a t uPotltiiclniak w distillation of the prqtiuct.
lize the exit gas for further dernetliylatiun, it is nc~essary to remove the methane from the effluent gas. Since it is impractical t o concentrate the hydrogen a t low niethane content.;, a bcttCr procrdure involves the ri~?-clingof a mixture t,hat contains approximately equal amounts of methane and hydrogen, fol1 u n . d by partial removal of niet,hane equivalent to t,he amount ing demethylation. For that purpose it is neethat, dpmethylation occurs equally well in the presence c)r a hydrogen-iiiethant? mixture as in t,lie presence of
T . ~ H L11. E DEAICTHTL.ITIO?~ O F TECHSICAL ~-\'EOHEXISE 4 532 ,505
EFFECTS O F I'KESSCRE. TEJII'ERATURE, A S D SPACE \ ELO( ITY
0 55
The partial hydrogcn prcmure has a pronoullwd csffect upon tile i~eartioii. I t was found that a highcr pressure i,iquircsi a higher operating tc:mperatuw; and since the reaction is confined i o a I i q u i d recovery. b y vol. 1.iquid product aii:iIy>w, inole rc~lativi~ly narrow tcmpcrature range, the reaction i. stopped when T o t 1 b u t a ne\ the operation is shifted froin a lon. pressure, such as 100 pnuii~ls pcr q u a r e inch gag-st Temperature Catalyst Cobalt Zjickel I SlL.hel I1
Ratio, H2 to Hydrocarbon 4.8
3.9-4.2 3.1-3.9
Space Yeloci ty 9
2 1.5
Conversion
70 21-30 28-33 48-52
856
INDUSTRTAL AND E N G I N E E R I N G CHEMISTRY
Vol. 39, No. 7
\ubjecicti t o fractionstion to remove a loiver boiling cut that CIJIItained about 74TGt.riptane. Table I11 gives the results obrhiiied on subjecting this materid to demethylation in two successive Ytages xhere the residue from stage 1 is demethylated in atage 2. These data show that the impurities prescnt8in t,he triptane c u t demethylate to a much greater estent than triptane itself; t h ( w fore, triptane can be concentrated by this method. S i i w ttiptane and 2,2-dimethylpentane (the major impurity! give I I ~ I P hesarie on tiemethylation, it may be expected that the ion.1.r boiling material produced in t,he purification shoultl c.iint;ritr >I large aniount of this material. The neoht~s:ttic~i w n t ( 1 t i t I ~i i l w lowrr Iioiling matci'ial was found to bo 67tW0,
OL
-
-1
_I
-I--
510
1
520 530 540 55C C A T A L Ys T T E M P E R A T u R E ,OF.
480
4 9 0 500
E ' i q r c 1. C2oti\er*ion 1 5 . ( d t a l > z t Temperature 100 Pound> per Sqiiare Inrli and H>drogeii1Ijdrocarl)on R a t i o of i t o 1
dt
Iiydrogeri Alone. Table I1 gives t h e results ohtailid on IirlJIY'3ing technical neohesane (Phillips Petroleum Company, 90f; t iiure) in the preeence and ;thsi~ticc~ of niethanca in t h c hyilmgi>n sti'('ani. Three data indicatc that tlt~iuethylatioiiof Iicuhes:tlics t o rirvilit>ritsne occurs readily in 1he presence of a m e t l ~ a n c - l i y ~ l ~ ~ ~ ~ ~ i ~ ,-t YIWYI. Additional dittit are given on procedsing neiiiii various pressures in the presence of hydrogen aloilc.. Iitc*livit,yof reaction is aho~vnt o tie relatively high. Previously it was indicated that, in the c [li some of the highly branched conipnuntl$. the primary 1)rodiii.I .[If demethylation is more r iant, to iurthcr tic~nic~thy1:ttion: m I , !herefore, relatively high ron\/ibr4on p r p ivithout apprcciithle loss of priinnry product -ing applies to possible use of t1enicrhyl:il iiin ing highly branched ronstiruents of :L hytiroc,arhon niist i ~ r t ' which cannot tie wonomically purified if the Iioiliug 11 ~ndi~-idual coinpimaits lie close togchthi*r. Iri t lic c ruethylation of a hot-acid isooctane ft,:rrticm, thc: ~ r i p t : i ~ l ifrai,Tiiirl , ,,ithe product contains 2,2-dimethyIpi~rit~rii. :diitig ivitli FIII,LIIL,I riiiounts of 2,1-dimethylpentnne as impuritiw. TSi dctcriiiinc, t i t t t I 50icvwibility of concentrating the triptanv frar: ion !,il.1, ~~l(~inrthy1ation. :I triptane cut oht:iiiicd i n ;L I . I . ~ I I I ~ Io. p e ~ ii to t i \ \ : I 40-
/,oo
PSI 2 S.V.
P
si3Ic.i.i
/
.
I
Z
I
0
I
2sv
,
I
, I I
1 470 480
490 500 510 520 530 540 C A T A L Y S T f E & i P E R A T U R €,OF
550
560
Figure 5. Spare Velocity os. Catalyst Temperature iit 100 Pounds per Square Inch and H~dropen-H?droc*:rrI,or~ Ratio of 1 to 1
I
500 PSI
/ IOL
4
sv
--I_--
530 540
550
__
1
-__I
5 7 0 5 0 0 5 9 0 600 A V E R A G E c A T A L Y s T T E M P E R A T u R E,%. I'igure 6.
560
Cmnversion 1's. Catalyst Temperature
I N D U S T R I A L A N D E N G I N E E R I N G CHEMISTRY
July 1947
lation t o rimove the 2,2,i-trimt.thylpentane first, then the charging stock for demet,hylation. The reaction product was fractionated. anti the unconvertcd material was recycled to the reaction zonr. T h r following arc the over-all yield? starting x i t h 1005 of hot-acid i.sooctane: "c b y Volume
Compound
c
Total
These data indicate that, approsimately 36 volunic~r;of ti,iptatic,, having a purity in excess of 83y0,can be ohtairicd from 100 v ~ d unies of hot acid isooctane, along with about 19 volunios oi' 2,2,~-trinic'thylperitaneand smaller vo1umc.s of 2,3-dirri~~tIi~l~i1~iitanc and 2,3-dimeth~ll~ut~aiie. LITERATURE CITED
Haensel, Y.,and Ipatieff, V. N., J . Am. Chem. Soc., 68, :N.j (1946) and Corson, B. B., IBD.ENG.CHEJI.,30, lO:{!l ( 2 ) Ipatieff, V. N., (1938). (3) Mc.Illiater, 8 . H., .Vat!. Petmlettm .'.-eu.,s, 29, S o . 4(j, l{-:{:i? (1937). (lj
3.5
2,3-Dimethylbutane Intermediate fraction T r i p t a n e , 83% p u r e Triptane,, approx. 90% pure Intermediate fraction 2 3-Dimethylpentane 2:2,4-Trirnethylpentane R ii t t o ms
a57
11.1
I
2.9 3;: 3 3 5 . 6 0.6 15.6 1s 9
2 5 __ 90.6
PAESBVTED before t h e Division of Petroleum Cheiniitry a t tlie l l l t i i i n g of t h e A V K R I C A X CHEWCAL SOCIETY, Atlantic C i t y , S . .J.
A1er.i-
Heat Transfer to Gases through
Packed Tubes
U
GENERAL CORRELATION FOR SMOOTH SPHERICAL PARTICLES 3L4X LEVA Central E x p e r i m e n t Station, 1.. 3'. Btrrearr of'.Mine..*,P i t t a b u r g h , P a . dinieriaiotiallj homogeneous equation ha5 been de\eloped which permits the prediction of heat transfer coefficients to gases flowing through tubes pached with low thermal conductiiit: material. The equation giies satisfactor) results for ratios of D p D t TarJiiigfrom about 0.05 to 0.3. 1 h e equation predicts a maximum coefficient of heat transfer for a ratio of D, D t = 0.1,5, and this finding w a y \ erified by experimental midence. The equation ma) be u*ed to calriilate coefficients for gases other than air as long a5 the Prancltl group of those gases doe* not differ materially from that of air and carbon dioxide. The iticlusioti of the Prandtl group in general i.i thorouphl? disrus-ed. E\periinetital eiiidence is preietited to show that to 3 the equation holds for tube 4 7 e s \ar>ing from inches i n dianieter. ' Oter this range the agreement is good, and the application to larger tube sizes is discussed. The effect of loids in the parhinp was studied irith great care, and no coordinated relation qeein- to e\i*t between ioids arid heat tran-fer c-oefficients.
T
HE iiiveqtigation of steady state heat traiisfer to gases fiowing in turbulent motion through packed tubes was untlert:iken in connection with a broad development program on .synthetic liquid fuel proce.sies. The prohleni \vas treated in a general manner, so that fundamental result. were ohtained rather than sprcific information. For thi; reason the jireient study zliould be of interest to those engaged primarily in the dehign and operation of equipment which utilizes, in one form or another, the principles brousht forth by this researc?~. ed of deterniining t h e heat Specifically the problem con transfer film coefficients and developing from these data a general n-orking equation, which could be used to predict heat transfer coefficient- for similar systems. Rome early pioneer work in this field \ m a done by Colburn'. T h e present research is an extension in which new data are presented. rl some\yhat different method of correlation was used. ' C'olhurn, h. l',, I N D E N G .C H E X , 23, 910-13 (1931).
EQUIPME.1T A \ 1) OI'EHATIOS
Figure 1 i h :i .ketch of the apparatu,- uaed for, this invc,-tig;itimi. Air \vas supplied by a hloner, and the Ho\v \vit. rn~~:i.-uI~('~i t J j - t w o rotameter6 \Titi1 overlapping ranges. The rot:unr~tc~r. \!-ere calibrated dii,ectly by a n-et teat meter and the calibt~atio~~,. corrected to stantlard conditions. -4 mercury ni:inunieter a11il tlierniometer inscrtrd into tlie upstream path u i the air iridic-ate11 the pressure and temperature of the air pas.-ing through tlrc, ters. All flon- re:idirigi \\-ere rorrrctrd for I 'iituie drriation. from the st:indard coiidi n-a> fed to either one of the te&.unit,. 9 by-p;~-h:irrangtxtiieI1t licrinitted the u\e of ritliei, t l i p 2-inch 0 1 the i,'2-iiicll pachrd tuljc,. 'The air entered the 2-inch unit througli a tec'. A 3-ir1cal1htalldarti pipe provitird a 36-inch high steam cheat around thr 2-i1ic]l ht:intlsrd pilie. The steam chert i v ~ i > properly veiited anti trapped t o prevent accumulation of ail, or ronden>atr. -4lirehi u r e gage \!-as provided at the center of the tuhe. In the annulus lietween the 3-inch :inti 2-inch standard pipe.< :t 3--in(,h long, ,j,'le-inch dialnetel. steel tube vas imtalied. Tlik tube, rveidctl *hiit at the lower end, touched the inner pipe over it. e~itirta length and served as a thermonell. At thc lover end of the 2inch pipe was a perforated plate covered hy a wire sc'reen i11:~c.ed tliirctly a t ttir lo\\-er end of the >team cheht. The 2-inrl1 p i p , extending out of the lower end of the ,team chect \vas imnwtli:ltelJ. riduced t o 1 inrli, anti a 1-inch gate valve n-as used to c o n t r o l tlic. air iioiv. Tile preisure drop acrws tlie apparatus vas measureti nit11 ;, \\-atc)i'inariometer (Figure 1). The height of t h e packing wa,~i l l all instances 36 inche3, arid the heat transfer coefficients W C ~ I ' I ~ c,:Llr*ulatedon tlie ha& of t h e inside pipe area. Thc entire apparatus W ~ Plagged froin the upper end of the steam chest to tliv throttle valre. h tliermonicter \va.s inserted into tlie t o p of til\ver enti tigations indicated that the differences in tempmitures, 3 5 determined by the thermocouple and obtained from thr, saturation hteam pressure, n-ere negligihle. For thi. wa.-oii tiit
