1286
INDUSTRIAL AND ENGINEERING CHEMISTRY
with rubber alone. This situation is modified on milling in air by t h e radical-acceptor activity of oxygen. The rapid attaining of a gel maximum not much less than the value under nitrogen is due t o t h e rapid reaction with carbon black. However, once t h e network reaches a rubber concentration where its rate of rupture compensates for its rate of formation, the superimposed action of oxygen produces degradation. T h e rubber chains sheared from the network may be the links binding some of the carbon black into the network, and so a proportion of carbon black passes into the sol fraction. T h e low intrinsic viscosity of the soluble rubber, the decrease in rubber content and swelling of t h e gel (Table I ) , and resistance of the carbon black t o centrifuging are consistent with this interpretation. Different grades of rubber and carbon black give different gel contents under identical milling conditions (6). T h e normal nonrubber constituents are of small influence, b u t certain rubbers may contain efficient radical acceptors, as is the case with chemically bleached pale crepe. T h e influence of the rubber itself is primarily t h a t of its molecular weight, gel content increasing with molecular weight (Figure 4). T h e radical-acceptor efficiency of carbon black should increase with fineness of particle size, structural factors remaining constant, owing t o the higher surface area for termination of rubber radicals. This trend is noted with different grades of carbon black. However, structural factors are also important, as revealed by the differences between lampblack and fine thermal black, although of comparable particle size. T h e relatively greater effect of carbon blacks on physical properties of GR-S and neoprene vulcanizates than of natural rubber, and the lesser effect on Butyl rubber under normal processing are coincident with gel formation. GR-S and neoprene on cold milling form networks with carbon black incorporating most of the elastomer, whereas Butyl rubber gives no network with carbon black (Figure 6). T h e resistance of these networks t o shear degradation is explicable by t h e low radical-acceptor ac-
Vol. 47, No. 6
tivity of oxygen with these elastomers compared with natural rubber. The demonstration of chemical combination of rubber and carbon black during cold milling raises immediately its relevance t o t h e reinforcement of t h e subsequent vulcanizate. This work has not reached the stage when a n answer to this question can be given. Preliminary results indicate t h a t a correlation exists between the reinforcing action of carbon blacks and other fillers (silicates, activated carbonates, whiting, etc.) and the occurrence and amount of gel on cold milling. Also, the properties of t h e vulcanixate are affected by the extent of milling with carbon black (6). There appears t o be sufficient evidence from gel formation t o modify the view currently held that milling is merely a convenient physical method of dispersing t h e filler. ACKNOWLEDGMENT
T h e author gratefully acknowledges the cooperation of R. I. Wood of the technology section, and t h e experimental assistance of John Davey and Norman Williams. This work is part of t h e program of research of the British Rubber Producers’ Research Association.
z
-
LITERATURE CITED
(1) Kolthoff, I. M., Gutmacher, R. G . , and Kahn, A., J . P h y s . &
Colloid Chem., 55, 1240 (1951). (2) Parkinson, D., I n d i a Rubber J . , 105,976 (1953). (3) Pike, M., and Watson, W. F., J . Polymer Sci., 9,229 (1952). (4) Schweitser, C. W., Goodrich, W. C., and Burgess, K. A., Rubber A g e , 65, 651 (1949). (5) Van Den Waarden, NI., J . Colloid Sci.,5 , 317 (1950). (6) Watson, W. F., Proc. 3rd Rubber Technology Conference, London, 1954,in press. (7) Watson, W. F.,and Wilson, D., J . Sci. Instr., 31, 98 (1954). RECEIYED for review June 4, 1954. -4CCEPTED November 8, 1954. Presented in part before the Third Rubber Technology Conference, London, England, June 22 to 25, 1954.
Effect of Heat Treatment on Reinforcing Properties o f Carbon Black W. D. SCHAEFFER
AND
W. R. SMITH
Godfrey L. Cubot, Inc., Cambridge 42, Muss.
T
HE effect of progressive heat treatment from 1000° to 3000” C. on t h e structure and surface composition of a series of commercial carbon blacks has been described (9). Treatment in this temperature range first removes chemically combined oxygen and hydrogen, and i t may safely be assumed t h a t above 1500” C. t h e carbon surface is essentially bare. The area of the carbon blacks as determined by nitrogen adsorption, and the particle size as observed under t h e electron microscope, are not significantly altered b y this treatment. However, x-ray diffraction measurements reveal t h a t t h e size and eventually t h e ordering of the quasi-graphitic parallel layer groups ( 7 ) composing t h e carbon black particles increase with increasing temperature (9). T h e extent of “graphitization” or internal growth and ordering depends on t h e particle size of t h e black. T h e degree of graphitization increases with particle size. These d a t a are summarized in Table I. Beebe and Young ( 2 ) have measured heats of adsorption of argon on t h e series of heattreated M P C blacks used in this study and report a progressive
change toward a more uniform and less active surface with increasing degree of graphitization. Argon adsorption isotherms on t h e series of heat-treated SRF and FT samples suggest a similar transition (8). The present paper reports t h e change in properties which semireinforcing and fully reinforcing blacks graphitized t o varying degrees display when compounded in natural rubber. EXPERIMENTAL
.
T h e induction furnace and procedures for heat treating t h e carbon blacks have been described (9). T h e carbon blacks (except t h e FT grade) were all in the pelleted form. Their properties are reported in Table I. All stocks were mixed in a Type B Banbury mixer and sheeted on a 6-inch roll mill. I n general, 1000-gram batches were prepared. All batches were mixed from a single lot of No. 1 natural rubber smoked sheet in the recipe given in Table 11. I n the case of the slower curing untreated channel blacks, 0.8 p a r t of accel-
.
INDUSTRIAL AND ENGINEERING CHEMISTRY
June 1955 Table I. Treatment
MPC
Carbon Black Spheron 6
HAF
Vulcan 3
SRF
Sterling S
FT
Sterling FT
Type
a
~
~
~
C. 1000 1500 2000 2700 1000 1500 2000 2700 1000 1500 2000 2700 1000 1500 2000 2700
Structural and Analytical Properties of Heat-Treated Carbon Blacks Structure , La Lo c 22.8 13.4 36.2 1 5 . 5 7126 44.8 23.3 7.03 75.4 41.9 6 . 9 4 78.86 48.2 6.90 23.5 13.9 32.5 1 5 . 6 7:26 46.5 24.7 7 . 0 3 72.9b 58.3 6.88 76.7b 5 8 . 3 6.88 25.8 15.2 36.2 1 8 . 3 7:23 57.2 36.4 6.95 llOb 9 8 . 8 6.87 132 121 6.86 27.6 16 8 17:5 7:26 37.5 62.0 38.6 6 . 9 8 116b 133 6.87 271h 276 6.82
Time 2 hours.
Electron microscope surface area z Z nidi 0 eda = d d n = __
Surface Area, Sq. Meters/G. Nitrogenb E.M.O 114 94 91.9 88.0 85.4 84.1
~ X-Ray . a
b B.E.T. area from nitrogen adsorption a t 78' 6
1287
71.5 68.1 65.4 62.8 61.9 27.9 26.0 26.5 25.2 24.9 15.5 13.1 12.9 12.6 12.5
Av. Particle Size, A.
volatile Matter,/
Ash,
dnd
da4
%
277
344
5.6
0.07 0.03 0.05 0.00 0.00
74
235
436
0.7
26
800
1220
0.4
1650
2250
0.4
0.20 0.22 0.12 0.03 0.00 0.75 0.55 0.20 0.06 0.00 0.00 0.04 0.03 0.03 0.05
14.3
70
pH 4.5 9.4 9.7 10.2 9.0
Electrical Resistivity0 Ohm-Cm. 17.8 0.44 0.41 0.58 0.70
9.4 8.2 9.5 7.9 7.3 9.8 10.3 10.6 10.1 9.0
0.56 0.44 0.39 0.70 0.77 0.58 0.25 0.26 0.51 0.58 100-110 0.24 0.19 0.34 0.38
8.? 8.0 9.9 7.6 6.5
Density, G./Cc. 0.45 0.30 0.36 0.39 0.42 0.43 0.40 0.45 0.47 0.48 0.65 0.65 0.69 0.74 0.74 0.74 0.66 0.76 0.84 0.91
K.
values obtained from S.A. = 6 X 104/1.86 do. z Z ntdhs
i
0 2
0"
0
Z
f Weight loss on heating for 7 min. a t 950' C. in covered platinum crucible.
nidi2
0
h
D.c. resistivity of black bed a t indicated density under 150 Ib./sq. inch compression. Average values from (10)and (If) maxima.
erator was employed. This way decreased to 0.5 part for the furnace- and heat-treated black compounds. The T-50 data indicated t h a t about identical cure rates were attained for all Spheron 6 compounds and comparable Vulcan 3 compounds. Compounds containing coarser particle size Sterling S and Sterling FT blacks had consistently lower T-50 values. All stocks were cured at 15, 30, 60, and 90 minutes at 280' F. All stress-strain data reported are for optimum cures. With the exception of the 60-minute cures for the untreated channel blacks, this optimum generally occurred a t 30 minutes.
4000
--
3000
In d
$
2000
W
a I-
(0
1000
Table 11.
Compounding Recipe
Ingredients
No. 1 smoked sheet
Carbon black Stearic acid Zinc oxide Sulfur Agerite Hipar Santocure
a
*
Parts 100 50 2.5 5.0 2.5 1.0 0.5"
__
161.5 0.8 part Santocure used in batches containing original Spheron 6.
0
100 PO0 x x ) 400 500 600
100 200 300 400 500 600
ELONGATION(%)
ELONGATION (%I
Figure 1. Stress-strain curves of rubber compounds containing heat-treated MPC and HAF blaclts
RESULTS AND DISCUSSION
Stress-strain curves 'for the vulcanizates containing 50 parts of the reinforcing blacks Spheron 6 (MPC) and Vulcan 3 (HAF) and for similar vulcanixates containing the same blacks after heat treatment are reproduced in Figure 1. The most striking effect produced by heat treatment of the carbon black is a drastic reduction in modulus. With the fully reinforcing carbons this reduction is evident in the 1000" to 1500" C. range. Heating these blacks higher-Le., to 2700' C.-produces no further decrease in modulus. I n the case of the semireinforcing carbon Sterling S (SRF), however, a further slight decrease can be obtained by heating the black to 2000" C., as illustrated in Figure 2. With the coarser thermal black, Sterling FT, there is little change in modulus on heating to 1500' C. The marked change occurs a t the 1500' to 2000" level. With the exception of the H A F black (Vulcan 3) there appears to be no significant loss in tensile strength with heat treatment. This is in line with generally accepted opinion t h a t tensile strength is primarily a matter of particle size of the black, which is not
significantly altered by heat treatment. The loss in tensile strength displayed by the treated HAF black may possibly be associated with some degree of sintering of the particles on heat treatment, although why it should be so pronounced in this case and not with the channel black is not clear. Turning t o the drop in modulus occasioned by heat treatment of carbon black, the question of paramount interest concerns the change in property of the carbon responsible for this decrease. Heating the blacks over the temperature range 1000" to 2700" C. has progressively driven the combined oxygen and hydrogen from the surface of the blacks, and, in addition, has induced some internal ordering or "graphitization" of the blacks. These effects are shown in Figures 4, 5 , and 6 of the earlier publication ( 9 ) . As a measure of the degree of graphitization the average dimensions, L, and La may be employed. These dimensions, derived from x-ray diffraction measurements, indicate the average thickness or number of platelets per parallel layer group,
INDUSTRIAL AND ENGINEERING CHEMISTRY
1288
+
4000
3000
-
-0
Vol. 47, No. 6
I
v)
n
200c
Y
+ 1000
0 IO0 200 300 400 500 600
ELONGAT I ON ( %)
ELONGATION(%)
Figure 2. Stress-strain curves of rubber compounds containing heat-treated SRF and FT blacks
Smallwood has derived the expression (10) M = Mo (1 2.5 V )to describe the modulus of a system consisting of an elastomer with gum modulus M o and containing V volume % of filler. He observed t h a t with reinforcing carbon blacks the observed values were much higher than those predicted by this equation. Guth (6) introduced higher terms and a shape factor to account for the discrepancy. It is interesting to note in Figure 3 that the values of all stocks containing carbon blacks heated to 1500' C. and higher are, with the exception of the H A F stock, of the order predicted by Smallwood's expression. The gum stock modulus, Mo, is about 500 pounds a t 400% elongation. All stocks had an equal black loading of 50 parts per 100 of rubber; thus V is of the order of 207& Accordingly, the classical stiffening computed from Smallwood's expression is 750 pounds per square inch. With the exception of the H A F stocks, all other blacks heated above 1500" C. have 400% moduli, falling in the range 900 to 500 pounds per square inch. I n view of the fact
80
and the average size of the platelets, respectively. Both increase with increasing treatment temperature, as shown in Table
r
1
I
I
iI
iI 9,I -I 'I 1 I
I
I
I
I. I n Figure 3 the modulus at 400% elongation of all stocks studied is plotted against the corresponding Laand Lo dimensions of the carbons. The striking feature is that the major loss in modulua occurs a t low temperatures, where graphitization has not advanced significantly. The slope changes sharply a t La and L, values of around 50 and 30 A., respectively, and from here on loss of modulus is not pronounced. From this it may be concluded t h a t the degree of graphitization is not a primary factor in determining the stiffening or modulus properties of carbon black compounds. Referring to Table I, it may be noted t h a t
56 700 _ . 500
300
'
r 100 88
78
I
W t - r T o I
68
58
Lo ( A l CRYSTALLITE DIMENSIONS
Figure 3.
-
0.45
f3
Effect of crystallite dimensions on modulus at 400y0 elongation
L , values of 50 A. are attained a t 1500' C. At this temperature all carbon blacks are' essentially free of combined oxygen and hydrogen (9). Ultimate analyses have confirmed this ( I , 11). 1.0~s of volatile, particularly combined hydrogen, is evidently a prerequisite for graphitization ( 5 ) . Accordingly, i t must be concluded t h a t the decrease in modulus properties noted on heating to 1000" to 1500' is associated principally with removal of surface hydrogen. The implication is that the outstanding modulus,properties which carbon black displays over other filler pigments is due to some type of undefined association between the rubber molecule and a thermally unstable constituent on the surface of the black.
x Y
0.05
IO00
2000
3000
TEMPERATURE, OC. Figure 4.
Effect of carbon black heat treatment on rubber properties
INDUSTRIAL AND ENGINEERING CHEMISTRY
lune 1955
t h a t a 50-part black loading and 400y0 elongation exceed t h e restrictions of Smallwood's expression, these rough calculations cannot be accepted in a n y quantitative sense. They are, however, adequate t o indicate t h a t the anomalous stiffening effect displayed by the carbon black originally studied by Smallwood is due to a n enhanced carbon-to-rubber association. When the source of this association, the chemisorbed hydrogen, is removed by heat treatment, t h e predicted classical stiffening effect is approached.
1289
5. The "scorch" properties or premature vulcanization which furnace blacks display in rubber stocks are usually associated with their low volatile content. Thus, as may be observed in Figure 5, the unheated HAF black with a volatile content of only 1% had a scorch time of 20 minutes, while the MPC black with a volatile content of 5% had a scorch time of 27 minutes. T h e scorch time of a furnace black can be improved b y increasing its volatile content b y air aftertreatment (4,6).B y the same token, it would be anticipated t h a t removal of volatile matter from a black should decrease t h e time required t o scorch. Surprisingly enough, the authors' data indicate quite the reverse; when both the chemisorbed oxygen and hydrogen are completely removed by heating to 1500" C. the time required t o scorch actually increases. T h e d a t a shown in Figure 5 are for only t h e H A F and EPC stocks; however, similar results were found for the other blacks. Tentatively, the results can be explained b y assuming t h a t the combined hydrogen on the surface of the black accelerates vulcanization and t h a t effect of combined oxygen is one of removing or neutralizing active surface hydrogen. When a n y black is heated t o 2000" C. a n essentially bare carbon surface is obtained, and the scorch and cure time of all blacks should be essentially the same. D a t a in Figure 5 indicate this to be the case.
10s
18
'
105 I
I
1000
2000
3000
I ' 0 I
TEMPERATURE, OC. Figure 5. Mooney viscosity and scorch point of rubber compounds containing heattreated MPC and HAF blaclis
104
*c
z
% T h e fully reinforcing furnace black (HAF) is unusual in this regard. T h e modulus properties of this grade of black are not readily destroyed b y calcination; as shown in Figure 3, even after heat treatment t o 2700" C. they are still capable of producing a 1600-pound modulus a t 400% elongation. This modulus cannot be ascribed t o volatile matter, and a possible interpretation may involve the fact t h a t H A F blacks poesess a grea,ter degree of particle-to-particle association or "structure" than other blacks. This structure is not destroyed by heat treatment. I n this instance the high residual modulus may well be accounted for on the basis of a shape factor as suggested b y G u t h (6). This interpretation, however, must be accepted with caution in view of the fact t h a t t h e resistivity d a t a of Figure 6 show the heat-treated H A F black stocks to be higher in resistivit,y than the M P C stocks. If the former possessed a higher degree of structure, a lower resistivity might be anticipated. Standard laboratory test d a t a were collected on all vulcanisates. (Copies of the detailed test d a t a on all stocks will be supplied by the authors a t the request of interested readers.) I n general, the results were much less striking than those observed on modulus properties. Shore hardness showed no significant change with increasing heat treatment of the blacks. Tear resistance and resilience properties decreased slightly, while permanent set showed a slight increase. Torsional hysteresis went through a slight maximum for stocks loaded with blacks treated a t 1500" C. These results are summarized in Figure 4. Mooney viscosity d a t a reproduced in Figure 5 show only a slight increase. An unexpected result was obtained when the Mooney scorch time of the stocks containing the heat-treated blacks was measured. These d a t a are also presented in Figure
103
5 w
a
IO' IO I
0
~
2000' C. TREAT M ENT T E M PER ATU R E
1000'
Figure 6. Direct current resistivity of rubber compounds containing heattreated carbon blacks
I n an earlier publication (9) describing t h e preparation and properties of heat-treated carbon blacks, the direct current resistivity of all heat-treated carbon blacks passed through a minimum a t 1500" C. T h e major decrease in resistance coincided with cleaning all chemisorbed oxygen and hydrogen from the surface-i.e., heating at 1000" t o 1500" C. T h e increase in resistivity noted on heating t o 1500" to 2000" was associated with internal crystallite growth and a decrease in the number of contacts within the particle. I n these measurements, made on the dry black, the coarsest blacks developed the lowest resistivity on heat treatment. I n the present work the resistivity of natural rubber vulcanizates containing 50 parts of these same blacks has been measured. T h e results are presented in Figure 6. Here the observations with regard t o conductance and particle size are reversed. T h e coarsest black, which in t h e d r y state showed the minimum resistarice, produces vulcanizates of t h e
1290
INDUSTRIAL AND ENGINEERING CHEMISTRY
highest resistance. This result, of course, is not unexpected. It does, however, emphasize t h e fact t h a t in a carbon-loaded vulcanizate t h e number of carbon-to-carbon contacts, or conducting paths, is of primary significance. Presumably if i t were possible t o load a rubber stock high enough t o approach closest packing of carbon particles, the effect of particle size would disappear and the conductance of the carbon itself would determine resistivity.
720
8
640
-z 560 I
3 480
0
> I 400 v)
rn
2
320
2
2
t
I
I
1000
2000
240
v)
4
5U
160
3000
abrasion properties on heat treatment. This was not unexpected, as these blacks are not considered t o be fully reinforcing. However, the laboratory d a t a on the fully reinforcing M P C and H A F grades were somewhat unexpected. T h e M P C black did not show significant change in abrasion loss on h e a t treatment. The maximum spread in hardness of t h e M P C stocks was only one Shore point and in the case of the HAF stocks the spread was b u t two points. However, results obtained on a laboratory angle abrader often bear little relation t o t h e results obtained in actual tire performance. T h e Lambourn abrader is held in much higher regard and the authors are indebted to Donald Parkinson for providing data from the F o r t Dunlop laboratory. Samples of M P C and M P C heated t o 2700' C. wereevaluatedbyparkinson, who reports a wear rating of 70% for t h e vulcanizates containing the 2700' sample, based on 100% for the original MPC. A. E. Juve, Brecksville Research Center, B. F. Goodrich Co., kindly consented to build and road test a natural rubber tread loaded with 50 parts of the 2700" heat-treated M P C black. H e reports a wear rating of 71 yofor t h e heat-treated black based on 100% for the original MPC. Electron microscope examination of these treads showed about equivalent dispersion of the blacks. The 70% ratings must, of course, be accepted over the laboratory abrasion data presented in Figure 7 . While the present data, particularly with regard to abrasion resistance, are inadequate, they demonstrate t h a t the chemical nature of the carbon black surface is the major factor in determining modulus, curing properties, and electrical ,conductivity of carbon-reinforced vulcanizates. On the other hand, when one considers reinforcement in terms of resistance to abrasive wear, i t appears t h a t the major factor is particle size of the carbon black.
TEMPERATURE, PC. Figure 7. Abrasion loss of rubber c o m pounds
containing heat-treated blacks
carbon
The question of paramount interest is the effect of heat treatment of carbon black on its ability t o impart abrasion resistance to rubber. I n view of. the drastic reduction in modulus which seems t o follow from the removal of surface chemisorbed hydrogen, one might anticipate an equally drastic reduction in resistance t o abrasive wear if modulus does indeed play a dominant role in rubber reinforcement. Since the heat treatment of these blacks was carried out in a laboratory induction furnace, the quantity of black available for abrasion testing was necessarily limited. Consequently, most of the study was confined to laboratory evaluation on the Akron angle abrader. Some of the results are presented in Figure 7. Neither the FT nor S R F grades of carbon black showed any very significant change in
Vol. 47, No. 6
LITERATURE CITED (1) Anderson, R. E., and Emmett, P. H., J . Phys. Chem., 56, 753
(1952).
(2) Beebe, R.A., and Young, D. M . , Ibid.,58, 93 (1954). (3) Biscoe, J., and Warren, B. E., J . A p p l . Phys., 13, 364 (1942). (4) Cines. R9. R.. Rubber Aoe ( N . Y.). 69. 183 (1951). ( 8 ) Dannenberg, E. hl., and Collyer', H. ',J., IND. ENG. CHEM.,41,
1607 (1949). (6) Guth, E . , J . A p p l . Phys., 16, 20 (1945). (7) Houska. C. R.. and Warren, B. E., Ibid.. 25, 1503 (1954). (8) Polley, >I. H., Schaeffer, W. D., and Smith, W. R., J . Phys. Chem., 57, 469 (1953). (9) Schaeffer, W. D., Smith, W. R., and Polley, 11. H., IND. ENG. CHEM.,45, 1721 (1953). (10) Smallwood, H. M., J . A p p l . Phys., 15, 758 (1944). (11) Studebaker, &I., Division of Rubber Chemistry, 122nd Meeting, ACS, Los Angeles, Calif., March 1953. ACCEPTED December 27, 1954. RECEIVED for review September 14, 19.54. Presented in part before the Division of Rubber Chemistry a t the 124th Meeting of the AMERICAN CHEMICAL SOCIETY, Chicago, Ill., 1953.
b
*
