The Viscosities for Hydrocarbon Systems. The Effect of Composition and Pressure Upon the Viscosity for a Three Component Hydrocarbon System Citation Inman, Byron Noll (1936) The Viscosities for Hydrocarbon Systems. The Effect of Composition and Pressure Upon the Viscosity for a Three Component Hydrocarbon System. Master's thesis, California Institute of Technology. doi:10.7907/mqe2-r454. https://resolver.caltech.edu/CaltechTHESIS:11072024-190004643 Abstract No abstract. Item Type: Thesis (Master's thesis) Subject Keywords: (Chemical Engineering) Degree Grantor: California Institute of Technology Division: Chemistry and Chemical Engineering Major Option: Chemical Engineering Thesis Availability: Public (worldwide access) Research Advisor(s): Sage, Bruce H. (advisor) Lacey, William Noble (advisor) Thesis Committee: Unknown, Unknown Defense Date: 1 June 1936 Funders: Funding Agency Grant Number American Petroleum Institute UNSPECIFIED Record Number: CaltechTHESIS:11072024-190004643 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:11072024-190004643 DOI: 10.7907/mqe2-r454 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 16839 Collection: CaltechTHESIS Deposited By: Benjamin Perez Deposited On: 13 Nov 2024 18:08 Last Modified: 13 Nov 2024 18:26 Thesis Files PDF - Final Version See Usage Policy. 11MB Repository Staff Only: item control page

THE VISCOSI'fIES FOR HYDROCARBON SYST&Ti,[S - - 0 - - TH.E EFFECT OF COMPOSI TION AND PRESSURE UPON THE VISCOSITY FOR A TH.REE COMPONENT HYDROCARBON SY STEM - - 0 - - ':f.'hesis by Byron N. Inman I N PARTIAL FULFILLMENT OF' THE REQUIREMENTS FOR THE DEGREE OF MASTER OF' SCIENCE June , 1936 THE EFFECT OF COMPOSITION AND PRESSURE UPON THE VI SCOSITY FOR A THREE CO!JIPOME.NT HYDROCARBON SYSTEM - - 0 - - I NTRODUCTION As a part of a coordi nated resea.reh program being carried out by the American Petroleum Insti tute ProJeet No. 37 , this s tudy of the effec t of pre ssure and composi t ion upon the viscosity of the me thane ,. propane , crystal oil system was made in an effort to correlate further t he phy sical behavior of hydrocarbons. Since the s tudy of a three component system over the whole range of compositions , temperatures , and pre ssura.s is a large task, this research was limited to a t emperature of 10O° F, and t o t hat region of compositions bounded by a maximum of' 5 mass per cent methane and 20 mass per cent propane . The viscosities were measured. i n the pressl1re range from below bubble point to 3000 pounds per S•':uare i nch , wh ich wa s the maximum _pressure allowed by the design o,f the apparatus . The 'Nork reported i n this thesis consist s of a de scription of the viscometer , of tabulated values of viscosity f or t he experimental compositions inve s tigat ed and of a graphical pres entation of viscosity and phase 2 equilibrium data der ived from the experimental values .. MATERI~LB USED The crystal oil used was a. highly-refined waterwhite paraffin base oil produ.eed from Pennsylvania crud~; its constitu~nts had a narrow range of high molecular weights and consequently the oil was of low volatility , and of moderately high viscosity. Its mole- cular weight was determined by the f reezing-point method in benzene and found to be approximately 342 . density at 100' F was o..8244 grams per cc. Its The sample was obtained through the courtesy of the Union Oil Company of California .• The propane used was obta ined from the Phil.gas Company, of Bartlesville , Oklahoma wh o reported t he spe cial fractionation analysis 100 ..00% propane. Investi- gation showed the bubble-point and dew-point varied by l ess than 0,.5 pound3 per square i nch , thus substantiating the Philgas Compa ny 's a nalysis .• The methane used was prepared from na tural gas in e quipment already developed by tJ1e Project 37 for re- moval of the higher molecular weight hydrocarbons by compression, cooling, and subs equent liquefaction of the undesirable components. Further treatment of the methane with activated charcoal served to remove small amounts of 3 ethane and higher hydrocarbons to the extent that less than 0.02 per cent remained in the methane gas . Partial solidification at liquid air temperatures and 1.05 pound per square inch pressure s.e rved to remove all but a trace of the n1 trogen and similar gases.. DESCRIPTION OF TI-IE APPARATUS The i nstrument u sed wa s the r olling ball viscometer adapted by Sage (1) from the earlier forms ~s ed by Ii~lowers ( 2) and Hersey ( 3) . I n principle it consists of a ball moving down (by a combination of rolling and sliding, whence the nam.e ttrolli ng ball") an i nc lined, liquid- filled tube for a fi xed distance . The annular clearance between t he ball and the walls of the tube is on the order of a . 01 i nch so that viscous flow of the li quid past the ball 1s maintained over the major por ti on or t he viscosity range.. Thu s , . for a given ball , the value of t ( () -E>) 1 is a linear function of absolute viscosity 0 of the liquid- Havrnver for 10·1 ,; viscositi es , as is sh own by a sample cali bration curve., the f low through the annular space become s turbulent , and the relationship between vis cosi ty and the parameter t( e 0 l Notation : -~) i s no t = roll t :i,.me ,. se cond.s p0 = density- of ball grams/cc. ~= densi t y of liquid grams/cc. 4 le>nger linear. Another factor which enters into this deyiation from linea rity is the accele:rat1on during starting,. which is, however , neglig i b l e for all ' except the shortest of roll ti:::ie$ (15 seconds) as may be judged from the fact t hat the compl ete roll time of' the ball down the elean a.ir.-.f iJ.led tube is . <>nly 1.. 3 seconds. A schematic diagram of the instrument devel©p.ed for the measurement of viscosities in t he condensed region a s well a $ for the tw0· phase reg ion is shown in Figure I. The li cruid under te s t was circulated t hrough the satu1·ation cell A, to the variable vol.ume ' ' cell B thence up the roll tube of the viseometer proper~,and :returned to t h e satu:r•ation cell t0 the cain pump.. A vane agi tator was provided in t he upper par t of the saturation cell to assist r apid attainment of phase equilibr ium. The system wa s so arranged that any entrapped gas in t he roll tube was s wept out by the ei.: rculating liquid and into the gas space in t he top of the saturation cell-~ The total volUIUe of the . closed circulating system could be varied about t wenty per cent at the will of the operator by admission to or release of ·mercury f'rom the bottom of the variable volume cell; permitting the measur ement of viscosities in the condensed reg:t.o n. The upper pres sure limit was set at 3000 pounds ·per 5 Figure I VISCOMETER Schematic Diagram of Viscometer 6 square inch by the design of the i nstrument. EXPERI MENTAL METHOD In opera tion, th,e entire system was :filled with !i We i ghed oil sample,. which after attainment of thermal equilibrium, (requiring l to 2 hours) was circulated through the roll tube, forc i ng the ball to the top , above the outlet tube , and removing any gas bubbles from the viscometer tube . Circulati on was th-en: stopped, the ball allowed to roll part of the way down the tube, and then t he upper contact w~.s screwed down a definite small distance- St~rt1 ng the circulati on again pumped. t he ball up the tube and wedged it against the upper ... · t p 1 n. l convac Circuiation was stopped, the vaive at the lower end of the roll tube closed , and after a short wait to allow the turbulent currents to cease the upper contact was lifted, starting the ball down the tube and aetuating an electric chronograplil .. The ball, on striking the l.ower c·onta ct, again actuated the chron@graph and the r,oll time was thus re corded . The valve at the bottom of the roll tube was then opened , circulation started again and the measurement l This procedure gave qui te reproduci.ble results , even though the upper' C(l)ntact may be reset a d o z,en t imes dur ing the course o.f a r un . 7 repeated until a. constant value was o.btained .. The pressure on the system was then i n creased by admission of mercury at a hi gh pressure to the bottom of the variable volume cell, equilibrium reestablished . by circulation, and -a viscosity mea surement then taken at the higher pressure, which was maintained eonstant by the use of fluid pressure scales . When the pre s .s.u re limit had been reached , t he pressure v.,a s. dro pped -t o below bubble-point by relea$ e of mer cury , more gas \Vas then forced in, and the whole serie s- of measurements repeated on the liquid for both the t wo pha se and the con- densed region for the new eomposi t1on . It •.,;a s- the usual experience to obtai n one or two val ue s far pressures below bubble-poi nt , one value at bubblepoint, a..l1.d t hen values for the eondensed region a t gauge pres.sures for even 500 pound i nt-e rval..$ to 3000 Jlound s per s quare i nch. MEASURJ;JMENT S 'I'he pressure on the system. was mea sured by means of a f l u i d pre ssure scale connected to the me.r eury in the bottom of the variable vel.u.m.e eell through a mercury trap , the transm.i,tti.ng .fluid being transformer o-il. 'The fluid pressure seales u sed w~re sensitive to 0.1 pound up to 300 pound s and to, 1 .. pound up t<:) 3000 pound s .. 8 The crystal oil san:rple was added fr om a glass wei ghing :flask. small steel bomb . The propane was weighed i n f rom a All _weighings were by difference , and i.,ve re a ccurat e t.o 5 mg . j repre,senting an accuracy better than .1% tor the smallest addition made. The methane vtas a-dded lL.'11.d~r pressure from a calibrated . reservoir., with a...11 a ccur acy of .. 3% by weigh t. The temperatur e of the v iscometer was hel d at the desired val ue b-y means of an oil bath and an electric heater controlled by a mercury regulator , which mainta ined the temperature within a .04° :F.. of the fixed temperature. The . maximum obs erved temperature vari.ation from point to p.0 -int i n the ba t h was less than O.. l o F-. T.em.-pera. t ures were measureu~ ·th -1. a mercury- 1·n- _glass them.om.et er vi'i th. gr9'du,at.ionsto O.-1 ° C and having a Bureau of Standards calibration. Exposed. stem: a.or0 rec. tion for the temperature o.f l00°F ( or 37 -~7a c) 0p amounted to only O. 07 .· - ~ .. Ti me was measured with an eleatric chrenograph having an elect.roma.gnet actuated recording :pen, and a drum driven by a. small synchronous motor which was , ef course,, subject to minor variations in the line frequency amounting to not more than 0 .1%. The desi gn was such that one inch of" tracing re:pres ~nted one seeond, thu.s allowi ng the es timatio.n of the e lapsed time to 0 .01 9 ·second. CALIBRATION The calibration of the viscometer was carried out with the liquids of known viscosity by following the general proeedure outlined above,. with the exception of the gas addition and attending manipulations. The· ball was one which had been previously used, but the calibration was cheeked at six points and at t wo temperatures with t wo kinds of erys.tal 011 and a kerosene. The vis- cosity of one oil, (a western oil having a high naphthenic content and hereafter re.f'e.rred to as crystal oil B) was known from previous work (4), while the viscosity of t he previously de scribed oil (an eastern paraffin base oil, and hereafter r eferred to as crystal oil C) was determined at 100° and 1eo9F wi th a previously calibrated Pyrex Ostwald viscometer. The viseo•si ty of the kerosene was similarly determined in a quartz Ostwald i nstrument. These points,. together with the points det.e rmined earlier were used in drawing the calibration curve shown in Figure II .. For the rolling ball viscometer the viscosity is proportional to the roll time and to that component of gravity determined by · the angle of the roll tube and the difference in density between the ball and the liquid. The angle of the roll t ube 1.s constan,t, henee so long as l.O ~ II ~ 500 - 10 0 , - - - - - - - + - - - - - + - - - - - - - - , r - - - - + - - - - - + - ~ - -- , ~ 0 CL :J _J ~ 400 - 801-- - -+-- - - - + ---------j-- - - + -.,,,_-- - + - - -- I ~ JOO - 60>-- - - - + - -- - - - +--------,-,,- - - + - - -- - + - - - - - , vi 0 ~ > W 200 - 40 f- :::, _J 0 U) CD <t JOO - 20 200 1000 2000 400 JOO() - 8 00 4000 T Ce.- t ) Calibrtrtion Curve 1000 ,000 11 the liq,uid flows past the baII is vis·eo-u s., the parameter of roll-time multiplie-d: by the difference in densities of ball and liquid is a li-ne:ar f'unetion of viscosity. Examination of the calibration curve indicated sub~ stantial linearity for the range 3:0 to 500 millipoises:1 I vdth some i ndication of' turbulent flow for viscosities ef less than 30 millipoises . 'I1he s.i:ngle ball used was sufficient to cover th~·-,,\ full range inv.o lved, so no further ea-l ibrations were }\ • neeessary, except as a check upon the accuracy of the .c alibration, for the work of Sage (1) indicated that / / i any change in the surface conditi'on of either ball or / roll tube would ea.use l .a .rge error.s in the results .. / The ball used was precision ground from stainless iron (density 7.675) to within .OQOl i neh of spherical, and extreme care was abs.e.rved in handling the ball to assure maintenance of an aeeurate calibration. The :roll tube was of similar material and machined with equal p~cision. Previous work of Sage Cl) indicated that any effect of pressure on the calibration was negligible. Similarly, the effect of' temperature. on the calibration was found to be negligible •. CALCULATIONS 'l'he calculations involved were simple. .,;_...- s~~r:-l :. The liquid ..f.,L:, 12 density for a given solutien wa.s calculated from the laws of perfect sol utions·:, and the estimated eompres·- sabili ty of the liquid and the value of the parameter t( <:> -e ) found frmn the ei;;,rresponding roll t i.me, t .. 0 Th(¥; viscosity was then read .from the ealibration curve and 1·ecorded as rough data. The rough . v-alues were pl&tted on a $<;!ale large enough to give readings of comparable •a.c curaey . with the exp,eri mental. work (0, 5%), the best fitting c;rurves drawn through. the po:i,nt s and the smoothed values of vise.osity :read rram the curve thus drawn. The excellent agreement of the values may be se,e n from Figures I!I, . V , VI , and. VII, in which the a<ltual experi.-- mental. points are sho~m for· the dif:feren_t eompo.si tioris,. The cemp,ositi.ons were calc,u lated both on t~u~ basis .o f weight per cent and mol per cent, using an average value of mol.eeular v1ei ght of i4.-:; for cryst al oil c.. The vis ... cesities f or the experimentally studie°: compositions are t abulated and shown i n Appendix A. Th.e arrange- ment is i n the oraler of i nereas ing mass ratio of propane to crystal. e>il .. At this point i~ . is pefhap s advisable to •e xplain the tt,m.a,.,s ratios» as used i n this thesis. As. i s well known, to repres 1:;m t in full detail a gi van variable a s a function of eomposi ti on and an i nde:pe.u dent variable (pre s sure) for a three component system would require 13 1 !l' Ill I ~ a . •~ !:--- 20 (/) w ~I ~ ~ ~ i,.....--- a: :J ...J ~ 80 >- \Jr!~ I- vi 0 vi > 60 w I- => ...J 0 ~4 0 <i I..-- ---- ...- L - -~ !_ ---- •~!:-- b----- 19o•F. / I 22o•F. 20 500 1000 PRESSURE 1500 2000 POUNDS PER SQ. IN. The Visco it7 of Cry t l 011 J?res.., re 2500 vnct1nn o 14 J'igure I 1.50,i------+--- -+ - - -- - t - -- +---+-------i .. ., ~ 1~1--- -+ - - - + -- C' 500 Th - + - - - - t - , ,,£.- 1000 PRESSURE 1500 2000 POUNDS PER SQ. IN . 2500 in i co ity of Cry t l oil ,ith Pr ssure :xpressed s Ratio of 1sco ity- Und r Fre sure to th ta.tit V por Pre sure. 15 more than three dimensions .. One possible solution is the tabulation of visco sity for eaeh c-ompos.ition and pres sure ,. such as ha.s been done for t he e:x:perimental compositio.ns. The only other .alternative for conven- ient presenta tion is to show one varia.ble as a fµnot .i on of t wo others with the restrietion that the ratie (either molal or wei ght) of two components must remain constant. For this work,, the mass ratio of .propane to \ crystal oil was chosen as the most conveniel'lt ratio to / i hold constant, as it arose directly .from the experi .- ( mental procedure, in which, the propane was added ~irst (. to the crystal oil, thus establishing a fixed weight } ,,/ _.., / ratio o.f the t wo-. All subsequent changes in compesi- tion for this particular ratio were made by addition of me thane to the system . As the experimentally de- termined molecular wei ght of a hydro-carbon such as crystal oil is of doubtf:u l sig:nificance, all subse ... quent. data (unless specifically stated otherwise) ar e pre s ented on a mass basis; thus per c,mt and mass ra tio of propane to crystal oil will be used i.,n re.ferr1-ng to composition. EXPERI MENTAL RESULT.S The effect of temperature and pressure on crystal oil C was first investigated, with the results shovm in Figure III, in which the viscosity is shown .as a function of pres sure for temperature at 30:0 intervals from 100° to 220°F , inclusive . The eff eet of teJllper·a- t .u re on viscosityf. is/as usual, much greater for the lower temperatli/:-re.s. The effe-ct ,of pressure is most marked foJ? 100°F, for which the vis-coslty at 3000 pounds absolute is 144% ef that at ap-pr-oxima tely O pounds per s quare inch absol ute. 1J.1his p~reentage is pr ogressively less for eaeh succ eeding higher temperature , amounting to only 13~"% at 2209F. The ratio for ea.ch t emperature as a function of pres.s ure :Ls shown in Fi gure IV. Entirely similar results te those shown i n Figures III and IV were obtained for crystsa.l o·t l B ( 4),, and for a kerosene ., but no means of correl ation is sugges t ed by the data ., exce.pt the possibility G-f an empiri- cal relationship based on specific gravit y , m-®l.eeular weight,: or some simil~r factor . However , the derivation - of such and empirical formula would require far more data than was obtained in the s tudy ot' thes e substances . METHANE-CRYSTAL OIL SYSTF1\/l: As a result of the work done on crystal oil c, the temperature of l00°F was chosen as having the most convenient range of vi scosity for the subsequent work on the t wo and three component systems. The first of the two component systems studied was that of methane and crystal. oil C;. the experimental. results of whieh are shown :in li'igure V ♦. The pressure ....viscosi ty curve f or pu:re. cryst$.l oil is shown, as well as tho s e for four compositions ranging fr om 0.584 to 4.0o mass per cent methane and ta a maximum bubble point cf 2290 p.o:unds per -square i i-;ieh. The number of' points in the two phase region was suf'..fieieht to permit the d:rawi:ag of t he t wo ... phase curve , . which is of the same nature as the saturatto-n ... p:ressure ; viscosity curves previousl y reported by Sage and eo.. workers (4). It should be :noted t hat t-he curve is not a true bubble point curve.-, sin.ee fot Jrress:ures n1ueh be-low bubbl.e point the volume of the gas phase was an appx~e- eiahle fraction of the whole .. The effect of the relative volume of ga s space on the liquld vi$eosi ty is diffi cult to estimate ., for though some tra.11..s,f e:re:n<le• o.f the lighteP components to the gas phase .takes pla'.e e, particularly at the higher pressures, the, necessary information is la.ek'-' ing which would permit a quantitative ealcmlation. It is the belie_f of the au·thor that such a transf'ere.nee or the lighter compo.nents to the gas pha.;se would tend to increase the 11-quid viscosity, Criti.cal exa.min.atio:n of the experimental result.s permits the conclusion that for bubble point pressures up to 2500 pounds per s quare inch and a ga.s volume of not m.o.re than J.5% .o f the whole ., such LS Fi ,.re V 12 0 CJ) w CJ) 0 100 Q. ~ _J ~ 80 >- f-- vi 0 u "'> 60 w f-- :::, __J ti 40 ro <( 20 500 Vi s codti e 1000 PRESSURE 1500 2000 POUNDS PER SQ. IN. 2500 ·t B1:1'b. le- ~oint and n the c n - densed Regi on for the Methane Cr y~tal Oil :,st em 19 sat uration- pres sure , vi s cosi ty curves a s obtai ned with this vi scometer i n the t wo phase region represent substantially the bubbl e- point viscosi ty curves~ I n the condensed region,. it will be noteel i n Fi gure V,. the pressure visc-osity eurve for the lowest mas s per cent methane is of similar shape as that for pure crys t al oil , showing a slight upward eur~ature and indicating a s for crystal oil.,, an i ncreasing pressure coeffi.cient of viscosi -ty 1 with pres.s ure.. F'or the/ ) .~ J · • higher rn.,e thane concentrations, the effeilt of pressure on the absolute viscosity Ls less., However , the per cent ine-rease is still large as may be seen from Figur e VIII , in w'hich the ratio of viscosity f or the eondensed region to t hat for bubble p(')int is :plotted as a function of the pres~s.ure increment above bubble :point . Examination of' the rigure indicates, also, the much greater effect of :methane a.t the lower concentrationi. F.'ur t her di s cussi on of the effect of concentratl on on vi scosity is given i n a later section, i n which an attempt i s l:llad.e to correl ate the results of the experiment a l studies . THE PROPANE•- CRY STAL OIL SYSTEM The pr.o pane- crystal a·i l syste-m was studied over l Pres sure coeff i cient of vi scosity is defined as (~ d p / x_ the bubble point pressure .range fr om 20 to 130 pou.nds per square inch absolute,, with a m.axt mum concentration of approxima te.1 y 17 per cent propane. The bubble point curve for this sylStem, as shown im Figure VI, indicates· a large change in viscosity with change in pressure, as might be expe.eteo. from the large solubility 0f propane in the oil at 100°F ... The experimental points for the t wo· phase r egion all lle on a single curve which represents, within the limits of experimeutal accuracy,. the bubble point viseosity. The effect of pressure on the vi~rcosity in the single phase region is best shown in Fig-u:re IX_,. in which the ratio o:f the viscosity for a mixture i :a. the condensed region to that at it;; bubble point is plotted against the corresponding pre ~sure inerem.ent above bubble point for the eoncentra tions lt 5, and 15% prepa;ne . The eurves a re in many respec t .s similar to th0se for methane , ex .... cept that the effect of pr opane is not~_$ large as that Qf methane , when compared on a mass :per cent basi,s . Fur ther discussion i.s glve}i 1,.n a late:r section. THE METHANE, PROPANE, CRYSTAL OIL SYSTEM This system was studied for compositions. up to 17% propane and 4% methane ,, with propane to crystal oil ratios ranging from 0,.018 to .0 .. 213. A graphical pre-sen.ta- 21 igo.re VI 120 (J) w V>IOO 0 a. :J ....J ~ 80 >- I- vi 0 u ~ 60 w I:::, ..J 0 fil 40 9, 19 <( 20 17. 5 500 L 1000 PRESSURE 1500 2000 2500 POUNDS PER S Q. IN. Viecositi sat Bubble-~o1nt and in the Condensed Region for the Propane , Crystc.l on Syetem 22 tion of the actual experim,e ntal re sults for the t wo ratios mentioned is given in Fi gure VII. As may be seen, the f i gure is in effect a combination of the t wo types represented in Fi gure s V and VI for methane and propane respectively . For the family of curve s having a propane to crystal oil r atio of' 0 . 018, it will be -seen tha t the effect of methane on t he bubble p.o int Viscosity is still qui te large , and t he condensed region curve s appear similar to those for methane-crystal oil alone, with the obvious exception that the 'bubble point viscosity for zero per cent met hane is now the viscosity of' the propane , cr ysta l oil mixture. I ncrea sing the weight ratio of propane to crystal oil has t he eff eet of .s hifting the ini tial viscosi ty down the propane bubble - point curve. The e.ffe ct on the vis- ,cos1 ty o.f adding methane is much less at the higher pro- pane eoncentrations , as will be noted from the smal l ehange i n viscosity (see Fi gure VII) upon addition of 4% methane to a solution having t he high propane to crystal oil ratio of 0 . 213 . The transference of some of the less volat ile components to the gas phas·e undoubtedly t.akes place i n the t wo phase region for the hi gher pressures , but as is explained i n the se ction cover i ng the methane , crystal oil system, the net effect on the viscosity is negligible . In an effort to corr elate t he effect of 23 iJI! "' "' w ~ 8 0 11.o-==--- - - - + - - - - + - - - + - - - - - - - 1-;:--c,..----------j;;~---""'~ 7 _J aa! ~ >- f- "'u 0 "' :; w :, f- _J 0 "'~ 20 PERCENT >-+- PROPANE /CRYSTAL OIL .. 0.213 o. METHANE ENT 2.1 3 3 .93 500 1000 1500 2000 PRESSURE POU NDS PE R SQ. IN . 'fl co 1t1 densed e 2500 t ble- oint and in th• Co on f. the e t an P~o ane ~ cry tal on s7 t cornposi tion on bubble- :p.oint viscosi ty ., graphiea.1 interpola.ti on was used to obta_in the family of curves presented in Figure X, in which the effect on the bubble poi nt viscosity o.f the mass per eent methane. i n the solution is shown for even values of the propane to crystal oil ratio. An examination o.f the figure sh ows that a simple correlation of bubble point viscosity and mass composition is impro-b able . In the condensed region, a similar correlation is out of the ques tion, as is evident f rom a comparison of Figures VIII and IX, which give the viscosity ratio t\ / I\ bp for fixed compositions as a functi on of the pressure increment above bubble point. A correlation would not be expected, s i nce the total pre ssures for these t wo eas es are widely d i f f'erent for the same value of pressure 1,.n ... crement . As sho'V\"D: tn the upper curve of Fi gur-e IV, the vi s cosity c oeffi cient of p:r essure, (d rt./dp) , i n creases with pressure for the crystal oil. Secondl y , t he effect o-f methane and propane and the tt l tt bp rat io cannot be com... pared satisfactorily by a diagram similar t o Fi gure VIII, si nce the pres sure limitations of t he viscometer precluded the necessary measurements on the methane, crystal oil system. 25 1001\----i------,1-----t---,----i Cf) w Cf) 0 (l_ ~oo ~ - + ~ 1 - - - - - - , - - - - - t - - - , - - --i _J ~ Cf) 0 u Cf) > 1.0 MASS •e 2.0 PERCENT 3.0 METHANE ubbl - oi.nt v , cod t rops.n~. cryst l oil 40 the 1 te~ a~ :r-el- ted t 26 e • II 151--- - + - - - -+ - - - - - + - - - - + - - - - + - -- ----t ~ .; ~ 1.3 t - - -- +--- - + - - - - - + - - ~ . ~ 500 1000 1500 2000 2500 CP-P6 , ) LBS. PER SQ IN. 'l'he I e f or s of ·eth e end vr.i t .ti xt Condensed Region . t o th th The ordin te r pr bubble p int of th the pres ur 'fhe hc1 ~ae r ent bove the 'bubbl e no nt of th I nth l nts mb:tur . i ncre- 27 IX 1.5 1--- - + - - - - - + - - - - + - - - - - - + - -- -----t------1 . c! ' 1.31----- - + - - - - - + - - - - + - - - ~ c' 500 1000 1500 200 0 2500 CP - Po, ) LBS PER SQ IN e Iner .~ ln 1:r.tu.re . op e ,i n . ·~c ,s ty t'h Cry tt1l ress,,,r r i l in the c or . n te re~r sent e r ti~ of the vi co i ty- und.e r pre e to ht at the oub le point of th c1 th pre sure 1ncr ov t e b bbl point of the mixture. The nt A POSS!BLE MEANS OF CORRELATION Previous discussion and the f"igures pres ented therewith have show-.£1 the difficulties of correlation of viscosity as a function of mass composition. Therefore a comparison on a molal composition basis was. attempted. The results are shown in Figures XI, XII, and XIII. In Figure XI, the bubble point visco.sity is plotted as a funct ion O•f the mol per cent crystal o;il for the methane, crystal oil system. Viscosities at constant pressure in the condensed region -:ire shov-in for 1000 t;1.nd 2500 pounds per square inch. It is noted that the constant pressure lines terminate on· the bubble point line at their respective bubble point composit ions. In Fi gure VII a similar plot is sho-w n for the pro- pane , erystal oil system.. At once the similarity of the bubble-point viscosity cu1--ves for the t wo systems is evident , although the constant pressure lines fen· the condensed re gion in the case of' propane do not i ntersect the bubble point curve. It is seen, however , that the bubble point viscosity of each t wo component system is practically the -same function of' the m.ol per cent of crystal oil c. 1 This might be expected, since each 1 This is not true for the systems methane, erystal oil B, and propane , crystal oil B, as is shown by the work of Sage (4), in wn ich marked dif.ference of molal effect on viscosity is noted .. However, cr ystal oil Bis from a western stock and is of high naphthenic content, and t herefore is not a member of the paraffin s eries . 29 component is a member of t he paraffi·n. serie s;- and methane and propane i n solution are rea sonabl y s i milar in properties. For the t hree component system, methane , propane , erystal oil c., the bubble.point vis cosity curve was drawn as shown i n Figure XIII, based on t he data for al the systems studied. Each system is differentiated by a particular notation and it may be seen t hat a ll the poi nts lie close to the single curve. Some random / sca ttering e>f :poi nts may be obs erved, attributable to ~ errors i n computatio-n of molal composition en the basis \ j of an average mol ecular weight f"cJr t he crystal oil. 'l'herefore , the eonc lus1on is drawn t hat t he. bubble- point viscosity f er t his system is a definite function of t he ;:nol fraet1on of ery stal oil , irrespectl ve of the proporti.0ns of the other t wo components. However , the uniq_u e 1•elationship of the components of this syste-m _must be kept i n mind i n the extension of the foregoing conclusion to systems other than t ho s e pres1::inted in thi s ~hesi s . That is, each of the compo.n ents i n t he systems studied i s a member of the homologous paraffin series ; two are o f low molecu lar weight , of high volatility ,_ and i n solution have reasonably similar properties; the t hird l,rhe cry sta l oil C i s fr om Pennsylvani a crude , h i ghly refined, and therefore largely a pure paraffin oil. 30 ~ o o -- -~ -- + - - - - + -- - - + - - - - iii 0 bl > w 40t-- - - - +- -----'lc'<---t----+-----t f-- :::, ~ <D <{ 201---- - + - -- -- - l - - ----+",,----+-----i 80 60 40 20 MOL PERCENT CRYSTAL OIL n Cone for th Oil . th thane . Cry t l tion of th fl Cryst i 1 ~!ol .ter Cent 3 XII i >~ 00~- - ~~ ~~ - -~ - - - ~ - - - - - - - i 8 V> > · ao oo 40 20 MOL PERCENT CRYSTAL OIL fr tie~ • u. b e- o n n t or 11 rystal th . l er nt 1 120 KEY : x • Clil -C RYSTAL OIL ; : S.t.:~1.~•;T~L~IL 100 80 \\ ~ 60 ~ \ 1ii 0 uen 5 w 40 t::::, 5en m <t 20 ~ '~ eo eo 40 MOL PERCENT CRV&TAL 01.. Vlscos ti 20 the Con- tr1c.t1on of th r .et 17st ~ Oil . ol is of h igh molecular weight and low volatility'! Therefore , for complex hydrocarbon systems of the same nature a.s those here report e d it should be possible ·to pre.diet t h e bubble .... :point viscosity of the system with fair aceuTaey· with data obtained .from the study of a representative t wo component. system, i .. e .. , a vise-o si ty s t udy for a system eons.tituted of' t he heavy component and one li ght component .. For example , it is possible to predict with re·a scmable •accru.racy t he bubble• point viseosity o.f t he methane,. p:r opane, crystal oil system from data obtained in t he .study ef t lae propane, crystal oil system.. Furthermore t~e bubble-poi.n t vis.cosi ty curv e :for such a t wo cmuponent system ma,y be determined without great experime ntal difficulty , in an i nstrument of the type herein des-eribeda- Unfortunataly, for the condensed region t here is no sati.ifa etory means of correlation of viscosity, pressure , and composition . Although the experimental work indicated a mode r at e percentage i nc r ea se in vis- co.si ty lNi th pressure 1 11. the condensed region for the experim.entally covered pre s-sure range,. the widely varying voTa tiliti es of the li ghter component s greatly complicate any attempt at correlation .. ( See Fi gures XI and XII) ~ 1~11e conclusion here is that more extended experi mental studies may develop some s a ti sfactory empirical correla - tion. PHASE EQUILI BRIUM DATA Since the experimental procedure has already provided a certain amount of data on bubble- point pressures for a limited composition range , i t is of inter est to present this additional information i n the t entative dia.g:ram sho•m in Fi gure XIV. The curves re present t he bubble-point pressure as a .function of mass per cen t in.ethane for f ixed mass ratios of propane to cry stal oil. Dat a f or the me t hane, propane curve wer e takeri from previous work by Sage and co- wo-rker s ( 5). I t should be noted that i n t he region 0% to 5% me thane, f or a fixed propane to crystal oil r atio t he bubblepoi nt pre s sure is proportional to t he per cent methane, indicating that Henryts l aw for me thane i s sub stantially true f'or the com.positions in the range studied . How- ever , the addition of propane to the methane erystal oil system m.ru:•kedly de c reas e .: ;- the bubbl e point pressure. The change in Henryl s law constant liJn addition of propane apparently is not proport iona l. to either mass or mol p.er cent . I f i n the future a. more extended study of this system is made , these tentative conclusions can be cheeked .. l 2500 I u ;;::; d (f) er 2000 w a. 'z IQOQ, t - - - - - f H ' H - - - t - -- - - - t - - ~ w _J ffi ffi so e 10.0 150 20.0 MASS PERCENT METHANE bble- • l to 36 L J TERATTJRE CI'fED (1) Sa.ge ., B. Ii,., I ndustrial and Engin.s.ering Chemistry:, Anal ytical Edition, 5; 26l, (1933) . ( 2) E'l ower s , Proceedings of the June-r i ean Societ y for Testing Materi~l.s., l •t 1 5€5 ,- (J.al 4) . (3) Her sey , ~ -• D. , Journal Wa,shi ngton AC:ademy Gf Sti enee s , 6 , 525 , (1 916). ( 4) - Sage , B. It., She:rborne , J . E.,, Lacey,,. ·,w . N. , Production Bulletin No .. 215~ American Pe~roleum I nstitute. i. 11 . V.i.s~e.$1.tg cf!Ji: F»".~ Ml C.. . . ...... .-- . . , . . . . . . . . . . .. page 38 Viseoe:i t y fO.;t" the me%b$ne , , ·()pane, cry$t.a l oi l s.7tJ•t •em:. . ••~ .. ._ e • ~ f; . . • • :• ff e ,. • ·9 •• • • · - • e, • • 'D • •,• • e • • e> ,. e •. • * • e page 39 Ackn..ewl,ed.gtttents ...... , .. :., ,. •.. ... . .... ~ ., . ~ ... -. .,........... page 43 38 AlHh Pressure, tb.,/Sq. In. 1oo~IF. Abs-ol\:lt.e Vhe0s:ity,. Millil}oises 130~F. 160GtF . 190":F .. 0 103, 0 62. 5 40 . 9 500 109 .,l 66. l 4.3 . 2 1000 115-.B 69 ..,8 45 . 7 3-1 •.5 1500 123. 5 ?3.., 8 48-•.2 3:3 . l 25. 2 0000 13.l.6 77. 7 50.6 34. 7 26 .}5 2600 13~ . 9 81.8 53 . 2 36. 4 27. 5 3000 148 .4 85. 9- 55 .6 . 38. 0 28 . 6 3500 156. l 90. 2 58 . 0 39. 6 29 . 8 a 28. 2 21..6 29. 9 39 Oomposi ti-onci::a as Follows o.o o.o o.o o.o 0 . 018 _, 0 . 58 1. 22 2 •.42 4 .00 o.o o.o o.o o.o o.o 1 . 7~ 99.4 98 . 8 .97 . 6 96 .0 98 . 2 27G. · 775,. 1374 . 2290_. 34 . 103 .0 85 •.3 62. 3 46 . 1 32.5 '79 . 5 500 109 . l 87. 7 - - · 83::5 1000 115. 8 ~3 . 2 .., -88: . 3 1500 123.5 . '8 . 9 46_. ? - 93 . 5 2000 131-5 104;9 49 . 3 • ~ -'9 2i00 139 . 9 110 .. 9 62. l 33 . 5 1-04. 5 300!J 14-S:. ~ · 1;[7. 1 55. l 36 .l 1:09 . 9 3500 157,:0 123 . 4 58 . l 38 . 6 ~15:'5 C3H9 .% o.o o.o o.o cob t/; 100 ..0 Blibble- Pt. Pr ~ssure Abs .· f r e ssu:te.,, Lb ./Sq. Il'l. a 'b 63. 9 7.5.. 5 83 . 9 ; -. ¼ Compositions are e a.leulated on a mass p$rcent oasis:. ,<Jc is the abbre11i a ti on for er;srs.tal on C . 40 1 00 , . J}001p0s i tiont a s Follow e~_;e,/ co CH4 -Oe:01133 0 . 018~ L58 1i: e;1a % 0 ~0183 L?S l.76 l . 74 0 . 0430 0 . 0430 0 . 0430 o.o 0. 45 1..23 4 ...1 2 4 . 10 4 ~07 co0 Bubble..,. -t". Presstire 414 .. 601 . 54.. .Abs . Pre~sure. tb. / S4. In. 00 . 6 :Bubble ...J?t . 63 .0 1000 66 . ? 51. 8 lf500 70 . 6 54 .9 .rooo 74 . 5 3000 3500 82 . 6 58 . 2 51.l 41.9 52. 4 500 2£00 36 . 8 64 .9 55. S 36 . 9 68...r 38 . 9· 72 .. l 61. 6 61. 4 4v.9 "16 . 0 64.8 51"';4 $4. 5 42 . 9 80 .1 68 .1 54 . 0 71. 5 VISCOSITY FOR fflE ME'l'IANE PiOP'.Al.'l'fil CRYSTAi',. OIL $YSffli l .0 0°F. Cempo.S:i tt ons as !fol lows. c1}1,8/eo a..o4ZO o •.0495 o.-0697 o . 069? o.OG9? 0 ~1012 % 2 . 56 o.o • o..o o..-40 1 .. oa o,#.o esns 1 4-. 02 4 . 72 6 ; 52 6 .. 49 6..,45 g:,_.:19 00 93.-4 95 ..3 93. 5 93 . 1 s~..5 to.a 1164.. 61. <m:4 :h:ltble-ft . Pressure Abs. Pre$wre,. Lb,s ./Sq. In.. 3'.l. 2 000 5~ . 7 1000 ~ .'7 35.. 2 41.2 -4 0 . 5 UiCO 46,. 8 43,;Ji> 2000 ~9 ., 46.. 5 3500 o/.3.• ·S . 61.1 76 . 8 $4... 0 56.0 42 Co.mposi ti<>as as- Follows OR ~$./00 .. 0 . 1073 0~1073 0,. 2lil?? 0:.2i27 , 0 .. 2:1217 - 0 . 2127 CK4 0 .0 '/; C/fs 1:, eo % :81..8 lhrbble.,Pt. Pres~~ 132. .. . A:hs .. fres·s we. Lb$ ./S.q,., ln. . 29 •.9 .tro'Q ,.. . . l.4 .. 4 12.,,.1 3:1 ,.l 1000 .l.500 34,·f 32:.,? $&,.,!$ us.. s. l'?.9 , .. !/ 2(?'GG a6.'.0 34. S . 19.l l&..,4 13. 6 10. 2 iaoo 31.$ $6,,.,(l l~. , li?,Q 13..:9: Ul.? .300& S9p3 37/1 20. 6 t ,7 .$ 14.~ 11.l ~ 41,,0 3f ~,4 21 . 3 11. 2 l,5. 8 ll,.& 43 ACKNOV/LEDG:MENTS It is with deep appreci a tion the financial support of the American Petroleum Institute is a cknowledged, wh.ich made . possibl~ this researeh . Much gratitude . is felt for the- helpful guidance so cheerf-ally given by Dr . Bruce H.- Sage . The assistanee of Dr . Wm. N. Lacey in the preparation o.f t his thesis is gr eatly appreciated . The ass istan ce of Mr . J . E. Sherborne in the t aking · of measurements in t he l aboratory, and of Mrs . B. N. Inman in t he preparati on of the figures deserv,es many thanks .