When u piymerit i s dispersed, erieryy is needed to merconie llie in/ernalfriclion ($ fhe baich, lo dejZoccnlute /lie pigment, and to create the pigment-rubber interface. These energy factors hate been studied for lhe system rubber-bluck under reproducible conditions. The eflect of lernperulure, shearing forces, and weltin,g on curbon black dispersion has been invesligated. I1 has been shown that lo obtain the masinzurn in black dispersion the butch must be made receptive lo black incorporution, a n d s u f f i c i e n t l y inlense shearingforces must be developed to break dawn the black agglomerates. A method of testing the processing characteristics of channel blacks is presenled.
Dispersibility of Gas Black 11. Evaluation of Gas Blacks' FRaK'
K . SCHOEKFELIJ AND
HAnmhx)
1'. ALLEN,The B. F. Gootlrich Company, Akron, Ohio
NE excellence and rapidity of pigment dispersiurr in rubber batches so greatly influences the ivearirig qualities and tlie economical fabrication of rubber products that the selection of pi-ments on tlie basis of their processing characteristics is a major problem. Carbon black, because it. is the most important of rill rei'nforcement pigments and bccause i t is extremely difficult to incorporate into rubber, deserves special study. This paper presents a method of evaluatiiig the dispersibility of channel blacks. Escli (8) judges tlie quality of a black by determining the quantity required to stiffen the batch so greatly that t,lic mixing mill can IIO longer be operated. T l i e smaller this quantity, t l i e better lie judges t,tie black to be. Wiegand ( l o )einpliasizes that the very property that is renponsilile for cliaii~ielblack's being the "king" of recnforcrnierit piqrents accounts for the difficulty of introducing it into rulrlinr. While this is true if channel black is conipai.ed with otlier pigments, i t is also true that differeiit channel bla.ks, all baviiig the same value as reenforcement pigmeiitfi, vary arriong themselves io their ease of mixing irit,o rubbcr. Tlie t,estiiig procedure here rlcscribed was devcloprii under tlie assuniptiw that any truly witicai test, iuust first have consider'ntiun for the tlieoretical factors ?i-!iicli iriiliiencc dispersibility.
In Figure 1 is sliown the total energy of mixing various piginto rubber by use of an improved Sclriller mixer. All
iiicrits
results reported i n t,liis paper, except in one case specifically not.ed, were obbaincd by use of this mrrcliine. The lowest curve is a control and sliows tiic cnergy input as the volume of the batch is increased by addit.irins of rubber. The other curves arc for cqrial volume additions of barytes, a tliernml decomposition black, a fine particlc size zinc oxide, slid tno channel blacks wliicli differed widi.ly in tlieir factory-pr1.cessing behavior. These curves were driinii from continuous records of kilowat,t, input as succersivc 10-volume portions of pigmelit, ume added at 5minutc intervals to a batch containing 500 volumes of rubbrr. The resiilts are reported as volumes of l & m i t per 100 volumes of rulher.. This 3-niinute interval is iiisufficient to permit, good dispersion, but the c u r ~ c sgrapliically sliow the rlifferoices in the power required to incorporate various pignirnts into rubber. These differcrices are correlated with tlie particle size of the pigments. As Wiegmri (10) has noted from empirical observations, the order of increasing proeessiug difficulty for dhese pibpents is the same as tlie order of increasing rei'iiforcing value. Tlie two chaiinel blacks, tiowever, are riot differentiated suffici~~itly to use this THEOJWPICAL CuxsrutlrAmms method as a critical test. Similar experiments with a nurnber of other elrannel blacks \ V i m , a pigment is dispersed, energy is mec?dcd €or tlir: ingave app,roximately t.lie mnie total energy curves. It was ternal friction of the hatch, to defloceulate the pignietit, and to create the pigmentruhher intcrfacc. If, for convenience, felt, tliereforc, that, in order to develop a satisfactory test ing behavior of blacks, it first would be necesit is assumed tliat these energies are positive, tlien the ener(gy sary to study tlie component energy factors contributing to of mixing can be expressed by the equation: this total energy. E = OR i (P ~ P B en (F~ICTIOX) \There E = tot,nl e n w ~ ysupplied to the lmt~11 r b = energy used liy internal irict,ionnl forcrs \%'hen rubbes is masticated, the mechanical energy of . deformation is transformed by frictinir to lieat energy. This I For Part I, J ~ literaluro D c i L ~ t i u nI.
+
1102
October, 1933
INDUSTRIAL AND ENGINEERING
CHEMISTRY
I
1103
Failure to recognize this factor results in the persistence of many uneconomical o CHANNELBUCK I factory mixing operations. For example, 0 CURNNfLBUM 2 i n order t o o b t a i n suitable dispersion, 0 ZINC OXIDE 4'0 80 0 T)tE~lLDCCMPaSlTIONBLACK batches are mixed for unduly long periods; B BsRVTEI thus w a s t e f u l a t t e m p t s a r e m a d e to e RUBBER J S I W M O O achieve the desired results by slowly wear3.5. ing down the hard agglomerates. More perfect dispersion can be obtained in a shorter time and with less energy input if o ORIGINALBLFICK ONE P6SS THROUGH MILL a t some early point in the mixing cycle k 0 THREE PRSSESTHROUGNMILL z shearing forces are provided which are 0 F I V E P f i I 5 E S T H R O U ~ HMILL 5 2.5 sufficiently great to break down the harder 3, agglomerates. P t O It has been found helpful to an understanding of the mechanism of gas black dispersion t o postulate a picture of the 56 I. 5 formation of these soft and hard agglomerates. If, when gas black is formed, the free energy of the surface of the particles I. 0 0 IO 20 30 40 50 18 23 28 is large, an appreciable decrease in this M;:IN6 TIMEIN MINUTES VOLVHESPIGMENT 2. EFFECTOF MILL-PACKING free energy would result when the particles FIGURE 1. COMPARISON OF ENERGIES FIGURE TO MIX DIFFERENT PIG- BLACKON ITS EASE OF DISPERSION approach one another closely enough to REQUIRED IN RUBBER MENTS INTO RUBBER eliminate some of this interface. This process of free energy decrease would cause causes a rise in the temperature of the stock with a correspond- flocculation. A sikilar phenomenon has been pointed out by ing increase in plasticity. Until this plasticity reaches a cer- Steele (9) in the case of zinc oxide dispersed in a nonpolar tain point, there is considerable slippage of the rubber past liquid. The gas black agglomerates so formed, because aggrethe metal surfaces of the masticator. When cold rubber is gations of solid particles have few points of contact, would tend introduced into an enclosed mixer, preparatory to mixing to be very soft. In the presence of the various organic prodgas black, the rotors deform the rubber, but because of slip- ucts incidentally formed in the present flame method of preparpage there is little tearing of the rubber and consequently ing carbon blacks, many of these agglomerates might be little new rubber surface is formed. Black added to the mixer cemented more or less tightly together. This may explain why a t this point receives little or no frictional wiping action and gas blacks produced under different conditions have different is exposed to a small new rubber surface. As the rubber ratios of soft and hard agglomerates. In any case, shearing becomes warmer, the plasticity increases, the batch adheres forces that are sufficient to disperse a certain percentage of one to the metal surfaces, the mixer blades pull through the rub- black may disperse a larger or smaller percentage of another ber, tearing it apart, and a fresh rubber surface is continually black. Whenever a pigment is mixed into rubber, a certain amount being formed. I n this condition high shearing forces are available to wipe out the black agglomerates, and a continu- of mill packing of this pigment takes place. Apparently, ally renewed rubber surface is presented to the black. A however, the wiping action of the rubber under usual mill further rise in temperature leads to a still higher plasticity. conditions always is greater than the forces required to break The rubber floms more easily. Little resistance is offered up the mill-packed agglomerates that can be formed under the rotors. Hence, while the black is constantly exposed to these same conditions. A point that has not yet been investia large new rubber surface, the frictional forces are small. If gated is whether different blacks pack to varying degrees of the temperature becomes still higher, the rubber breaks up and tends to become sandy. In this state the black and rubber can only form a coarse mixture. Any test for gas black dispersibility must permit a control of these frictional or shearing forces. 4.5
.
TLMPERRTURC 140'
L
.
0
EP
(DEFLOCCULATION)
Part I of this series (1) has shown that gas black contains various types of agglomerates differing from each other in the strength of their cohesive bonds. I n d i s p e r s i n g such a pigment, relatively small s h e a r i n g forces are required to wipe out the softer agglomerates and r e l a t i v e l y high shearing forces to break down the harder agglomerates. If, during the mixing operations, the maximum shearing forces developed are below the threshold force needed to disrupt the hard agglomerates, these agglomerates cannot be dispersed.
TEMPERRTURL 75' o TEMPERATURE30. TEPlPERnTuRc 110 a TEMPERATURE l2.0' TEMPERATUXE 130' e TEMPERATURE 145' B TEMPERATURE 15.5' W
D o
M I X I N G T I M E I N MINUTES
FIGURE 3. COMPARISON OF GASBLACK DISPERSION IN RUBBERAS THE INTENSITY OF SHEARING FORCES VARY
64
-
INDUSTRIAL AND ENGINEERING CHEMISTRY
1104
TLMPCRRTURL
FIGURE 5 . EFFECTOF ZINC STEARATE ON GAS BLACKDISPERSION IN RUBBER
hardness. If they do, this would be one explanation for the difference in the rapidity of dispersion. By forming hard agglomerates in a black, and then dispersing the black, a difference in the dispersion before and after compression is noticed. For example, in Figure 2 are shown dispersion results for batches of rubber and blacks run in an improved Schiller internal mixer. The dispersion was measured by the method described in the first paper (1) of this series. The blacks tested were obtained by compressing a channel black to vnrying degrees of hardness by passage through a roll mill. Every succeeding pass through the mill compresses the black more and gives a product that is increasingly difficult to disperse. Shepard (8) has observed that having rubber too stiff when the black is added results in serious aggregation of the pigment. Undoubtedly this is the same as the phenomenon of mill-packing. In the discussion of friction it has been shown that the easiest experimental method of varying the shearing forces is t o vary the temperature a t which the batches are run. This, of course, may also vary other factors, such as the wetso 88.
Vol. 25, No. 10
the other temperature effects. Figure 3 shows the same black mixed for increasing time intervals a t different temperatures. As the shearing forces are decreased by increasing the temperature, the rate of dispersion and the total amount of black dispersed are decreased. At 150" C. only 80 per cent of the black is well dispersed, a t 110" 95 per cent of the black is well dispersed. At 75" the shearing forces are a t first low because of slippage effects, but, after the black is once in, these slippage forces are to a large extent eliminated and dispersion takes place a t a rapid rate. Figure 4 shows a striking example of the effect of high shearing forces. If the batch is removed from the improved Schiller mixer after 28 minutes of mixing and milled on open rolls for 3 minutes, the dispersion improves from 85 to 90 per cent. If the batch is placed on the roll mill after 8 minutes in the internal mixer and milled 3 minutes, the dispersion increases from 65 to 95 per cent. Because the rubber is tougher when
42 1
2
5
4
5
6
7
8
4
PERCENT MOllTURC ONBLnCK
FIGURE 7 . EFFECTOF MOISTURE ON GAS BLACKDISPERSION IN RUBBER
removed earlier in mix, higher shearing forces can be developed on the roll mill. Consequently, the rate of dispersion is increased to the point where a mixing time of 11 minutes produces a better dispersed stock than a mixing time of 31 minutes.
I-
$ 84-
EPR
5
atz eo-
n.
-
;7a-
2
P
76
-
o UNTREATEDBL~CK DEGaSJEDBLaCK
74-
The energy needed to create the pigment-rubber interface is equal to the energy needed to destroy the interface pigmentair and rubber-air, minus the energy gained by any specific attractive forces operative between pigment and rubber:
I- 7 2 -
~ P R=
70-
Q
where
68-
" i i 66-
eo
MIXING TIMEIN MINUTE$
FIGURE6. EFFECTOF HEATTREATMENT ON GAS BLACKDISPERSIOKS IN RUBBER
ting of the black by the rubber, the rate at which equilibrium wetting is obtained, and flocculation of the black. Goodwin and Park (4) believe flocculation of the pigment is made possible by the increased mobility of rubber a t high temperatures. Grenquist ( 6 ) found that flocculation of dispersed pigment takes place in rubber-black mixes on heating. Repeated observations on rubber-black mixes in this laboratory indicate that in the temperature range of these experiments the variation in shearing stresses with temperature greatly outweighs
(WETTING)
IFPA
+ I F R A - IFPR
energy of adhesion IF = interfacial tension energy Subscripts P A = pigment-air; R A = rubber-air; ment-rubber 6
=
.PR
=
pig-
I F P R is, therefore, the energy factor which determines the wetting of the pigment by the rubber mix. The greater this value, the smaller will be E P R and the easier the pigment can be dispersed. Theoretically this value could be so great that dispersion would take place on mere contact. The value of I F P R can often be changed by the addition of rubber-soluble materials; by the adsorption of substances on the pigment surface; by changes in the rubber on milling, such as oxidation, depolymerization, or molecular rearrangements; by temperature changes which not only may affect the value of the interfacial tensions-rubber-air, pigment-air, and rubber-pigmentbut also may affect the rate a t which equilibrium conditions are obtained. Endres ( 2 ) states that a filler cannot be completely dispersed in rubber unless the rubber wets it. Green ( 5 ) believes the more effective the wetting, the more complete is the dis-
I N D LIST 1%1 A 1~. A R 1) P; N G I I Y 1:, E 1%I N G C 11 E hl I S T 11 Y
oc1o1,er. 1YJ3
1103
persion. Steele (9)metitioris tirat f i surface-aotivc mat,crial culitaiiicd in clrnimel black assists dispcrsiori. Eridres (2) conewitrates at tlie interface, ~ C T interfacial S ion, aud also dates that finely divided carboi eeonies gritty vlien ~ O ~ U WUTS eliminates the tciidericy of pigments mulate. st,roitgly lii:atcil. Tlie treated black 11 1 Ilcre was licated to IKgure 5 s h o w s how a 1050' C. at a pressure less c h a n g e i n x e t t i n g can tlian 2 mm. or n i c r c u r y change the energy needed for 30 lioilrs. If t'rrcrc ivas perse a given black to any grit iurrried, it was innitc degree at various sufficient to cliange the retemperatures. Kith a mixenforcing cliaraeteristics of ture of rubber and black tlie the black. TIE increased srnallest amourit r r i en1:rgy dispersibility of this black niay be diio to the remoral is n r e d r r l between 100" and 120" C', l3elow this ten,of a surface film. It also perature range, ivlrere tlie could be accounted for if, during tlie 1,igIi-teinperaslippage effectsoreliigli arid the rate of ivctting is p o b tnre treatment, the bindably slow, inore energy is ing agents wliich cemcnt the hard agglamerates tor e q u i r e d . Blrove t h i s range, as the shearing forces gether were dest.roycd. fall off rapidly, the batch An intcresting s u r f a c e must be mixed for a long effect is t h a t c a u s e d by time, arid the t,ota.l energy variations in the moisture input is large. If a softener content of the black. Water is added to tlic batch, an is n o t c o m p a t i b l e with entirely differentpict,urererubber, yet, strangely s u l t s . In this case zinc enough, dry black disperses s t e a r a t e was used. Tire less readily than black containing a small amount of wetting forces increase witli temperature s u f f icieii t l y water. 'Under the condirapidly to offset the detions of the experiment an crease in alrca.rine forces o n t i m u r n disnersion rate arid fiiier:olating tendency of the pigment. These curves crossi was ubtained r& blacks containing between 1 and 2 per a t 125°C. An operator mixing his factory stocks hot would cent of water. At higher moisture contents the water acts say that zinc stearate helped disperse tlie black, aiiotlrer as expected and inliibits dispersion (Figure 7). The moisoperutw Illixing his batches a t alower temperature wmld con- ture content was deterniiiied hy the heat loss a t 105" C . at elude that zinc stearate had no dispersing effect 011 gas black. tlic end of 2 liours. Figure 6 shows t l i c result of altering the surface of a black. TEST IN^ GASI ~ L A C K S Tire upper dispersion curve is for the same black as t:he lower curve except that the black has been degassed for several Tile recugnitioir and application of these fundamentals of liours a t a high temperature and under low prcssure. This pigment ilispcrsion hare permitted the develognient of a treated black disjrer more rapidly than tlie saixie black not t,reated. Endres (e) postulates that t,lie difficulty in milling
e
8*
IS
iD
el
28
MlXlnqTlnL IN M ~ U Y T C ~
F i w n e 8. COMPARISON OF EASE OF DI~PERSION AT 115" C. OF SEVEN CHANNEL BLACKS enrbun black in rubber is due to the adsorbed gases forming surface film which prevents contact between rubber and black. .Jolinson (7) takes the opposite view. He feels that the gas
9. COaPARISON O F EASI3 OF DISPERSIONAT 140' C. OF SEVEN CHANNEL BLACKS
FlGiUHE
laboratory test for the dispersibility of gas blacks. Tliis testing metlrod has been in use for over a year and bas always given an excellent correlation wit11 factory results.
I N D U S T 11 I A L A N D E N G 1 Y E E R I U G C I1 E M I S 'I' R Y
I LO6
Val. 25, No. 10
METHODOF TESTING. Five liundred volumes of rubber are masticated for 4 minutes in an improved Scliiller mixer using 13anbury type blades. During this time the temperature is adjusted. Then 150 volurncs of tlie black to be tested are added to tire rubber in four increments at %minute intervals. A sample for dispersion examination is taken a t the end of the last 2-minute period. Mixing is continued, sample' being taken a t convenient intervals. DESCRWPIOK OF M.ACHINE.The Sclriller type of internal mixer is a two-blade nrachine with a capacity of 1500 cc. The bowl is jacketed for the purpose of cooling with water or heating with steam; the steam is supplied through a regulator which permits as high as 140 pounds per square inch (9.8 kg. per sq. em.) pressure.
-1 'L
h
8
lclOUHE OF
o ;
A
1s 7.) MlXINtTWC IN MlNYTEJ
11.
ze
COMPAHISON OF R A T E
DISPERSION OF Two CIXANNEL HL**CKS
I,,
I
56
8
io
,
,
13 iB 29 M I X I N O TinF IN MiNUiLJ
,
ea
Fl0"HE 10. Co?dennrson OF 13.4m OF DISPERSION AT 140' C. OF SEVEN CiiANNEL BLACKS IN THE PRESENCE OF PrNE
run on Seven channel blacks at 115" C. The difference in the ease of dispersion of these blacks is clearly shown. In Figure 9 are the same blacks run at 140". At this temperature the shearing forces arc appreciably smaller than at 115". AU the blacks show a slower rate of dispersion and the difference between the good-dispersing and poor-dispersing Llacks has widened. In this particnlar test these same blacks were factory-tested. The two lower blacks would not tube at all. The two intermediate blacks tubed rouelr. The t.hrec h k.,h blacks tubed excellently. Krrurc 10 shows acrain these same blacks run at 140" C. but with t.lie additionto the rubber of 5 per cent of pine tar. The same order is still preserved, but all blacks disperse at a faster rate. Figure 11 shows two blacks that arrive at approximately the same dc,gree of dispersion but disperse a t quite diffcront rates. I n this particnlar case the black whioh dispersed a t the slower rate contained 5 large number of soft agglomerates.
T*n
Tlie inaciiine is fitted with a compression cover and a movalile ram operated by gravity. The blades used in this investigation are known as Banbury type. The drive is posilive tiirough a reducing gear attached to a 3-horsepowor motor; the front rotor has a speed of 44 and the back rotor 30 r. p. m. PowEn Mmsurtmiem. The energy of mixing is followed by means of a recording wattmeter. The power curves duplicate in forln those obtained from factory I3anburys. The mixing temperature or time of running has no effect on the idle load of the machine in t.liese expcrirnents. TEmErtATunE C O ~ T R OTlie L . temperature is measured by means of a therniocouplc wliich projects into the batch through the lid of the macliine. This reads low, but the difference between the recorded and actual temperatures is fairly constant. Several times during the mixing of the batch this temperature difference is checked by means of a needle thermocouple pushed into tlic stock. For example, in one experinleiit tile recorded temperature was 130" C. while the observed temperature was 138". TIE temperature of the machine is controlled by manual operation of t.he stearn and water lines. DUPLICATIONOF RESULTS. h p e r s i o n rcsults can be duplicated closely. The s a w black is used RS a blank in all tcstti. Samples of these test blanks niixed 1.5 years ago accurately check blanks mixed today. TYPICAL TESTRESULTS.Figure 8 shows results of a test
I,ITEliATURE
ClTED
( I ) Allen and Sairoenieid, IXD. EYO.C ~ P M .25, , 9Ll4 (lil33). 0) Erxires. Ihid., 16, 1148 (1924). (3) Escb. G ~ m m i - Z l g .40, , 1917 (lW26). (4) Goodwin and Park, Ino. EXG.Cam