Interaction between Carbon Black and Polymer in Cured Elastomers R . S. STEARNS'ANDB. L. JOHNSON The Firestons Tire C Rubber Co., Akron, Ohio T h i s research was initiated to determine whether the interaction a t the interface between the surface of finely divided solids, such as carbon black, and cured elastomers was primarily physical or chemical in nature. Further, i t was desired to correlate some physical property of the reinforced stock with the surface properties of the solid pigment. Through an exam ination of the thermodynamic changes accompanying the deformation of loaded stocks i t is shown that physical adsorption of the van der Waal type occurring a t the interface between pigment and polymer is inadequate to account for the experimental observations. How-ever, if chemical bonding occurs a t the interface between polymer and pigment, then the entropy
of deformation of the stock may be correlated with the extent of this bonding. By a calorimetric method i t was demonstrated that the surface of a carbon black particle contains sites that react with bromine to liberate the same amount of heat as low molecular weight olefins, It is therefore proposed t h a t a carbon black particle be considered as a disordered agglomerate of polymeric benzenoid type molecules which contain around their perimeters various functional groups. The existence of olefinic-type unsaturation on the surface of carbon blacks suggests strongly that, in the case of carbon blacks, the polymer and pigment are combined chemically through pigment-sulfur-polymer bonds into a continuous threedimensional cross-linked matrix.
T
t,hen
HE improvement of one or more physical properties of vul-
canized rubber stock8 by the incorporation of a finely divided solid is known as reinforcement. There exist in the literature a number of theories as to the mechanism by i%-hicha solid of high surface area, such as carbon black, improves abrasion resistance and increases the tensile strength and the modulus of a crosslinked rubber matrix ( 1 , b f , 22,38, &, 50). However, no satisfactory conclusions have been drawn as to the nature of the interaction a t the interface betwren the polymer molecules and the surface of the solid inclusion. I n general, it has been considered that van der Waals type adsorption forces are responsible for this interaction (b,26,55, 36). However, there are in the literature a uuinber of references to "chemical binding" between polymer and pigment, usually in reference t o carbon blaclis (54,43). The work reported here was undertaken in order to establish more clearly the nature of the interaction between various carbon blacks and the polymer in the cross-linked rubber matrix. The resulk3 of the study are described in three parts:
f
(g)p* + T(%)
Pl
where .f is the force in grams per square em. required to extend 1 cc;. of stock to a length, 1. The symbols F , E , H , S , P , V , and T represent free energy, internal energy, enthalpy, entropy, pressure, volume, and temperature, respectively. Considerable interest has been shown in the determination of the contribution of the entropy and of the internal energy to the force of retraction of unloaded (gum) stocks (14, 90, 59). Few data exist, however, to show how t>hesequantities are affected by a finely divided solid (61). Equation 3 requires that the tension in a strip of rubber held at constant length should be a linear function of the temperaturc. That part of the force of retraction which is due to a change in the entropy,
(g)PT,
may then be determined from
the slope of the isometric, and the enthalpy contribut,ion,
1. The thermodynamic changes accompanying the extension and retraction of gum and loaded stocks. 2. The chemical nature of the carbon black surface. 3 . The relationship of the physical and chemical nature of the carbon black surfaces to the physical properties of the loaded rubber stocks.
(%)ppJ
may be determined from the intersection of the isometric with the ordinate a t 0" K. If isometrics are determined a t a number of elongations then the int>egralquantities AH, A S , and A F may be evaluated as the area under the appropriate curve obtained by plott,ing the differential quantities as a function of the elongation. Apparatus and Method. The apparatus used to obtain the data necessary to calculate the above quantities was similar to that used by other investigators (18, 39). A beam balance was mounted on top of an air thermostat, and the motion of' the beam was mechanically restricted to *0.05 em. about the rest point. The rubber test strips were in the form of standard 2-inch dumbbells. The samples were held in the thermostat between two clamp^. The upper clamp was suspended from one arm of the balance, and the lower clamp was so arranged that the distance between the two clamps could be varied. The elongation was determined by the distance between bench marks placed on the restricted portion of the dumbbell. The temperature of the air thermostat could be varied from -30" to f75" C. and could be controlled vithin ~ 0 . O 1C. The load required to counterbalance t>heretractive force of the rubber could be read t o h0.5 gram. The distance between bench marks on the rubber test strips was determined to *0.01 em. with calipers and a steel scale.
THERMODYNAMICS O F RETRACTION OF LOADED STOCKS
The properties of a system may he defined by the thermodynamic equation of state. For t h e mechanical deformation of rubberlike materials the follow in^ equations have been derived ( 2 7 , 19):
and since
and
1
=
Present address, University of Chioago, Chicago 37, Ill.
146
January 1951
INDUSTRIAL AND ENGINEERING CHEMISTRY
147
TABLEI. COMPOUNDING INGREDIENTS (Stocks cured a t 300' F. for 60-180 minutes; sample slabs 6 X 6 X 0.075 inches) Parts bv "2 .Weight Sulfur 0 Zinc oxide 3.0 Antioxidanta 0.5 Accelerator b 1.5 GR-So 100.0 Filler or reinforcing pigment 0-45 a Phenyl-B-naphthylamine. b N-cyclohexyl-2-benaothiaaolesulfenamide. C Stearic acid present t o the extent of a t least 3.75%.
z 0 I-
o
-
a
-
a
PROPERTIES OF FILLERS AND REINFORCING TABLE 11. PHYSICAL PIGMENTS
-
"
134%
"
Surface Areaa,
8s.
EPC-1 (easy processing channel black)
EPC-2 Calcined EPC-2 (black heated to 1000° C. in vacuo) RF-1 (reinforcing furnace black)
Meters/ z b Gram Met& 127 23 108 29
..
rc 0.68*0.1 0.80
...
112 81 29 1.02 108 27 1.06 29 218 0.44 19 SRF (semireinforcing furnace black) 80 1 .oo Graphon (EPC black,heated t? 3000' C., 6 , 7 ) 76 30 (SiO*)m(H20)n (precipitated silica) 150 22 0.6 a Surface area by low temperature nitrogen adsorption (6). b Mean particle diameters from electron micrographs. 0 Ratio of area calculated from eleotron microscope data to the nitrogen area (5).
RF-3 R F.3 *I-
...
200
GR-S was used as the medium in which t o disperse the pigments since i t has no tendency t o undergo crystallization on orientation ( 2 3 ) . T h e compounding ingredients added t o the polymer t o effect cross linking are listed in Table I. Table I1 gives the physical properties of the pigments studied. All s t o c k in a series containing any one black or other pigment were treated in a like manner. T h e time of cure was so chosen t h a t the ordinary physical properties, as represented by tensile strength, elongation at break, and modulus a t 3mY0 elongation changed only slightly as the time of cure was increased. A plot of the physical properties of rubber stocks against the time of cure shows a rather broad plateau, and the procedure adopted in this work I placed all stocks on this plateau 5 but not necessarily at the same 13 point on the plateau. No use ,4 is made of the absolute values ff of any property but only of - c l3 6' the manner in which the par3 12 ticular property changes with : addition of pigment. Rubberlike materials exhibit 3? PRE-ELONGATION A S L I L . several phenomena which make Figure 1. Effect of Prethe determination of their physical properties in a state of elongation on Force of equilibrium somewhat difficult. Retraction at 100% Elongation I n particular, the effects of GR-S containing 45.0 parts of creep (44), permanent set RF-1 black (ZO), and previous history (4, 8, 3.3) must be considered. The effect of the initial deformation on the physical properties of a stock is well illustrated by the data plotted in Figure 1. Here the tension a t an elongation of 100% has been plotted as a function of the greatest deformation the test strip has suffered. A period of 1 hour was allowed for relaxation and recovery in these experiments before measuring the force of retraction. T h e value of 20, representing the unstrained length, used to calculate the 100% extension always included the increase in
5
E4 y -
~~~~~~
d
-
0
20
400 a00 TIME I N MINUTES
40 60 90 TIME IN MINUTES
eo0
100
12(
Figure 2. Relaxation and Recovery Curves for GR-S Stock Containing 37.5 Parts of EPC-2 Black
the permanent set introduced as a result of the greater deformation. T h e cause of this behavior is not clear; i t is probably the result of the "untying" of physical entanglements of polymer chains, €he slippage of the adsorbed polymer on the pigment surface, and the possible movement of cross links along the polymer chain through exchange reactions ( 3 1 ) . T h e stocks were given a known history by allowing them to relax a t an elongation of 150y0 for 1 hour. T h e initial relaxation curve at a n elongation of 150y0 and the subsequent recovery curves as the elongation wns decreased in increments of 1 5 % are plotted for one typical stock in Figure 2. The initial relaxation curve shows no tendency to reach a constant value, and the rate of relaxation is a linear function of time in the latter portion of the creep phenomena ( 3 2 ) . T h e authors believe t h a t the linear portion of the relaxation curve is not connected with the physical rearrangement of the polymer chains but is due t o the release of tension as a result of the breaking of cross links and oxidative degradation (4). T h e rate of recovery was found t o be rather rapid; a t the end of 30 minutes the force of retraction was within 5Y0 of the value which would have been obtained had a much longer recovery period been allowed. This is shown in the insert of Figure 2 where the recovery curve for a n extension of 86% has been plotted over a considerable period of time. This method of obtaining the stress-strain curves did not eliminate the hysteresis loop as shown in Figure 3. However, the last point on the ascending curve is close, after a relaxation period of 15 hours, t o the descending curve so t h a t it would appear t h a t the stress-strain curve obtained by retraction is very close t o the true equilibrium curve if such a thing exists for rubber stocks. Results and Discussion. A typical isometric-tension, as a function of temperature at constant length, of a loaded stock is
INDUSTRIAL AND ENGINEERING CHEMISTRY
148
shown in Figure 4. These isometrics indicate that the force of retraction is a linear function of the temperature over the temperature range covered as required by Equation 3. The slopes of the isometrics were obtained graphically and the intercepts were calculated by substitution of the value of the slope into Equation 3. T h e thermodynamic quantities calculated by this procedure for a series of stocks containing the SRF black and a series of stock containing the EPC-2 black are plotted against the elongation in Figurecl.5 and 6.
401
START
w
a
1.0
1.5 2.0 R E L A T I V E LENGTH
Vol. 43, No. 1
Assuming that Gee's analysis holds a t least qualitatively for GR-S loaded stocks then
TVR = AF
= -2'ASd
=
fjdl
(5)
where Ci'E is the equilibriuni work of retraction and A S d is the configurational entropy change accompanying deformation of the matrix due to orientation of the polynier segments. -4large number of isometrics niust be obtained in order to construct the curves presented in Figures 5 and 6. This is a rather laborious procedure, but little error will be made if the area under the equilibrium stress-strain curve is considered equivalent to TASd. ii series of GR-S stoclis, containing different amounts of the nine pigments listed in Table 11, were compounded and the equilibrium stress-strain curves were obtained as previously described. T h e area under the stress-strain curves from 0 to 125% elongation was evaluated by means of a planimeter and the work of retraction, [ W R ] ~was ; ~ , computed. The work of retraction as a function of concentration is plotted in Figure 7 for several representative pigments. This may be expressed by the equation
where IT:' is the work of rcitraction ot the gum stoch. V I is the volume fraction of pigment, and rp is a constant. This equation represents the data accurately for the nine pigment8 considered. According to Equation 6,+ is an inherent property of the pigment, and it may be considered that $J is the additional amount of work required t o deform the polymer matrix per unit volume of pigment as the concentration of pigment approaches zero, that is,
2.5
LIL.,
Figure 3. Stress-Strain Curve Illustrating Hysteresis Effect Stock contained 45 parts RF-1 black in GR-S
The SRF black has a relatively low area and consists of particles having a mean diameter of approximately 800 A. T h e EPC-2 black has a n area five times larger and is composed of particles having r~ mean diameter of 290 A. The EPC-2 black will then contain on the order of 1016 particles per gram whereas the SRF black will have on the order of IOl4 particles per gram. T h e two blacks are therefore entirely different in their physical properties. However, Figures 5 and 6 indicate that these two blacks exhibit similar behavior when incorporated in the cross-linked polymer matrix. The origin of the enthalpy changes which were found to accompany the deformation of the stocks can probably be ascribed to volume changes taking place during deformation, as discussed by Gee (19) for Hevea gum stocks. Since for rubber
the volume change may be calculated from the equation
(%)
=rpasV2+0
Figure 4. Isometric at Elongation of 300Yo for GR-S
Stock Containing 37.5 Parts EPC-2 Black
(4)
Cross-seotional area test piece, 0.0556 s q . cm.; slope, 0.179 kg.fsq. cm. X C.; intercept, 1.78 kg./sq. om.
k
where K is the isothermal compressibility of the elastomer and p is the linear coefficient of expansion of the elastomer. An increase in the volume should be observed therefore on deforination at low extensions; at greater extensions a decrease in volume should be anticipated. T h e volume change is probably the result of a n increase or decrease in the intermolecular spacing (in the absence of crystallization), and Gee has shown that a t small extensions the change in the internal energy associated with the volume change is approximately balanced by an equivalent change in entropy so t h a t the force of retraction in the rubber matrix remains the same as if a n external hydrostatic pressure had been applied to prevent the volume change.
R a i l ( 4 8 ) has shown that for cross-linked polymers the entropy change taking place during deformation of the matrix is given by A& = k In P
Po where Po is the number of possible configurations of the polymer chains before deformation and P is the reduced number of possible configurations in the strained state. The following equation may therefore be written
INDUSTRIAL AND ENGINEERING CHEMISTRY
January 1951
149
I
,,/
ko0
37.5 b0.0 FREE E N E R G Y
37 5
I I I I I I I I I I I I I I I I I I I I I I I IO
1.5
2.0 2.5 RELATIVE LENGTH L / L o
3.0 I
Figure 5. Curves Representing Thermodynamic Changes Accompanying Deformation of GR-S Stocks Containing EPC-2 Black Numbers on curves refer t o parts black/100 parts polymer; only two of six enthalpy curves are shown; A G on ordinate represents thermodsnamic quantities A F , AH, and TAS
where the second term on the right may be considered as representing the further decrcase in the entropy of deformation due to the presence of the pigment as a result of an additional decrease in the number of possible configurations of the polymer segments 111 the strained state. ];quation 6 may be eypanded to give
l l ' ~= 1.1'8 4- +Vi 11hich
c
.,
(10)
has the same form as the equations developed by Rehner (.%'e),and Smallwood ( 4 1 ) to represent the increase in the tensile properties of a polymer when a finely divided pigment is present. T h e higher terms of Equation 10 represent the effect of particle-particle interaction, and the squared term represents the interactjon between pairs of particles. Table I11 gives the experimental values of [+] observed for the nine pigments used in this work. It has been considered ( 1 1 , 1 2 ) t h a t the variation in the values of 4 is primarily due to changes in particle shape and degree of aggregation of the partirles since, according t o Smallwood ( 4 1 ) and Rehner (58), cp should be independent of particle size. This type reasoning might conceivably explain the variation in c#.among the carbon hlacks, but it does not explain the value of zero observed for Graphon and the silica. Also there is no correlation between the surface available for the adRorption of polymer and the value of +-that is, Graphon and silica have a large surfare area compared t o the SRF black which in turn has a value of cp greater than several of the EPC blacks. It might be argued that the differences in adhesion between the polymer and the surface of the pigment account for the large values of 4 observed for the seven blacks on the one hand and the zero value observed for the t u o inorganic pigments on the other hand. However, recent surface work (6, 9,24, 26) has indicated t h a t there is not a sufficient difference either in the work of adhesion or in the heat of adsorption between the pigments and low molecular weight (38), Guth
~
+ 24Vi -t 347.';.
1.0
l
l
/
I
1.5
I
1
I
I
2.0
l
l
I
I
25
RELATIVE LENGTH
l
l
/
L/L.
,
1
1
1
30
Figure 6. Curves Representing Thermodynamic Changes Accompanying Deformation of GR-S Stocks Containing SRF Black Numbers on curves refer to parts black/100 parts polymer; A G on ordinate represents thermodynamic quantities A F , AH, and TAS
hydrocarbons to account for the observation of a value of 4 equal to zero.
TABLE111. OBSERVED Pigment EPC-1 EPC-2 Calcined EPC-2 RF-1 RF-2 RF-3 SRF Graphon Si02
VALUES FOR CONSTANT
4
L S I ' ~Cal./Cc. ~, 0.31 0.37 0.49 0.50 0.65
0.57 0.46 0.00 0.00
Let it be assumed t h a t the difference in behavior between the carbon blacks and the two inorganic pigments is primarily due t o the type bonding between polymer and pigment. If a chemical bond exists between carbon blacks and the polymer in the cured rubber stocks, then the entire carbon black- olymer system could be considered as being cross-linked into a tEree-dimensional network; if i t is assumed that only van der Waals type adsorption forces exist between the silica and the Graphon, then the low values of 4 obtained from these latter pigments can be understood. The greater the amount of polymer which is bound or immobilized a t the surface of the pigment the larger will be the change in Sd on deformation. Beebe (6) has pointed out that the heat of adsorption of hydrocarbons on solids is of the same order of magnitude as the potential barrier t o free rotation around single bonds. If these results can be carried over t o the polymer system then migration of polymer segments over the surface of pigments would be expected in the cases of physical adsorption of the polymer. This migration of polymer segments would tend t o reduce the value of A&. If on the other hand it can be shown t h a t a covalent chemical bond is posRible between the polymer and the
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
150
surface of some pigments, then migration of polymer segments could not take place. I n this case the polymer could be considered as immobilized, and this would tend to increase the value of ASd. Furthermore, in the derivation of Equation 10, Smallwood ( 4 1 ) and Rehner (38)considered that the polymer segments are immobilized a t the surface of the solid inclusion. This condition cannot be fulfilled if only physical adsorption exists a t the polymer pigment interface. CHEMICAL NATURE O F CARBOY BL4CK SURF4CE
The crystal structure of carbon blacks has been examined in considerable detail ( 7 , IO); the adsorption of nitrogen and hydrocarbons on the surface of blacks has yielded much valuable information (3, 6, 6); and the morphology of carbon blacks as obtained from electron micrographs is of considerable interest (88. 49); but i t is only recently that any research has been inaugurated on the chemical nature of the carbon black surface.
I
/
r_ E P C .2
Figure 7. Work of Retraction as Function of Volume Fraction o f Black Present in Stocks
It ha5 been known for some time that carbon blacks contain considerable oxygen, hydrogen, and other so-called volatile constituents ( 2 5 ) . T h a t the pH of carbon blacks is dependent on the extent of surface oxidation was recognized by Zapp (52, 5 3 ) and further studied by T'illers (46). The study of the reaction of the Grignard reagent with carbon blacks undertaken b y TXers ( 4 7 ) has suggested t h a t there exist on the surface of carbon blacks a number of organic functional groups containing oxygen 0
// such as -OH,
-OOH,
-C
, and
\
Vol. 43, No. 1
eter. The entire assembly v a s immersed in a constant tempcrature bath. iibout 10.0 grams of carbon black, which had been dried in vacuum over calcium sulfate, was dispersed in 200 nil. of dry carbon tetrachloride and placed in the Deviar. After thcrmal equilibrium had been obtained, a known amount of bromine dissolved in carbon tetrachloride was introduced by breaking a glass bulb containing the bromine immersed in the black slurry. The type of temperature rise versus time curve obtained is illustrated in Figure 8. After the bromine TWS introduced the temperature change was followed for a period of about 1 hour, and t,he temperature rise was computed by making a linetw extrapolation of the temperature-time curve to the instant a t which the bromine was introduced as indicated. The heat capacity of the calorimeter was found by passing a known current from a storage battery t,hrough the heater of known resistance. The temperature was read on the Reckniann thcrmomctcr to *0.001' C. with the aid of a cathetometer. At the completion of t,he experiments the black was filtcred off and the concentration of any excess bromine was measured o n a colorimeter. Thc accuracy of the individual experiments was of the older of 5 : ; and the over-all accuracy is of t.he ordcr of loyo.
Results and Discussion. The intt:gral lieat of reaction as M function of the amount of adsorbed bromine is plotted in Figure 0 for the EPC-2 black; Figure 10 is a siniilar plot for the RF-2 black. It is evident t,hat the integral heat of reaction of bromine with carbon black may be represented by two intersecting straight lines. This is in distinct cont,rast t o the integral heat of reaction curve obtained for Graphon, Figure 11, which is typical 01 that expected for physical adsorption. The linear and intersecting heat of reaction curves obtained in the case of carbon blarks are characteristic of those t o be expected if the bromine is r c w t ing chemically rather than being adsorbed physically on the surface of the black. Figure 12 is a plot of the values of the integral heats of reaction for the seven carbon blacks studied, a t bromine concentrations which lie to the left of the discontinuity. T h e values for six of the seven blacks fall on the same straight line which passes through the origin. These six blacks have t.hvrcfore an average differential heat of reaction in this region of 18.2 kg.-cal. per mole of bromine. The SRF black has a differc.ritial 4.30
I I-
1 I
\
C=O, and this is further /
OH substantiated b: the pirxliminary work of Smith and bchaeffer (48). If, in addition to the ahove functional groups, evidence could
(-A=&-" " >
be found for the existence of eth) lenic type bonds on the surface of carbon black particles, which may be considered as part of the carbon black structure rather than being present in physically adsorbed organic units, then it may be postulated t h a t during vulcanization of the carbon black-polymer-sulfur mixture, sulfur cross links between the polymer and the black will occur, and the black and polymer will be combined chemically into a continuous three-dimensional matrix. The determination of unsaturation is rather difficult and uncertain under the best of conditions. When a large surface, capable of physically adsorbing considerable amounts of reagent, is present the number of difficulties is not lebsened. The problem n-as solved with some degree of success by the following caiorimrtric method:
Ex erimental. The calorimeter consisted of a silvered Dewar whicl! contained a stirrer, a heater, and a Beckinann thexmom-
TIME IN MINUTES
Figure 8. Temperature Rise on Addition of 2 X Moles of Bromine to 10 Grams o f EPC-1 Black in 200 A l l . of cc14 Heat capacity of calorimeter wag 90.4 G a l . / " C.; A H i b 2.51 cal. f gram of black
heat of reaction which amounts to 27.5 1rg.-cal. per mole of l r o mine. T h e equilibrium n-hicli exists t o the right of the disroiitinuity, between the free bromine and the adsorbed bromine, follows a Freundlich-tl-pe isotherm, as shown in Figure 13. T h e average slope of the integral heat of reaction curves in this reyioii is 11.9 kg.-cal. per mole of bromine. Table IV gives the va.lues of the differential heats of reaction in the regions to the right, and left of the discontinuity and the amount of adsorbed bromine at the point of discontinuity.
January 1951
INDUSTRIAL AND ENGINEERING CHEMISTRY
It is assumed that AH,,
HEATOF BROMINATION DATA
TABLE IV.
WZd.
Moles/ la A H A b, AHBC SP. Mol& Kg.-Cal./ Kg.4hi.j Meter G. x 104 Mole Mole x 107 Black 12.2 EPC-1 0.24 18 2 1.89 EPC-2 0.27 18.2 2.52 11.9 1 .oo 18 2 13.3 Calcined EPC-2 9.25 12.4 RF-1 18 2 0.44 5.44 RF-2 17 0 9.2 0.80 7.39 1.31 18 2 10.1 6.02 RF-3 0.36 SRF 27 5 13.3 19.0 Graphon 0.00 ... ... a 6 = Moles Brz adsorbed per gram of black a t point of discontinuity assumed to be equivalent to the double bonds existing on the surface of the black. 6 Differential heat of reaction below point of discontinuity. C Differential heat of reaction above point of discontinuity. d d expressed as moles per square meter of black surface. . . I
MOLES
Bf2
A D S O R B E D per
X104
GRAM
Figure 9. Integral Heat of Reaction of Bromine with EPC-2 Black The heat of reaction of bromine with a double bond in a simple organic molecule is 28 to 31 kg.-cal. per mole (29, SI). This heat of reaction refers to the reactants and products in the gas phase. In order to compare the differential heat of reaction obtained between bromine and carbon black in a carbon tetrachloride slurry with the above value the following thermochemical cycle must be written:
'
1
Kg.-Cal. /Mole AH1 = -7.5 AH2 = - 0 . 5
Br4g) +Br,(l) Brdl) CCI4 +(Br2-CCI4) CB CC14 +(CB-CCla) (CB-CCI,) (BrrCCl4)+(CB~Brp-CCl4) (CB,Br2-CC14)+(CB,Br2) CCll
AH3 . AHc = - 18.2( A H A ) AH5 = . . .
Br? ( g )
AH8 = -26
++
+
+ CB --+
+
CB.Br2
. .
151
the heat of immersion, is about equal to AH,, the heat of emersion. Therefore, to the observed heat of reaction, AH,, must be added the heat of condensation oi bromine and the heat of solution of liquid broCARBON B L A C K mine in t h e s o l v e n t . The heat of reaction of 0 gaseous bromine with I carbon black in vacuo, AHR, is then about -26 M O L E S B r ~ A D S O R B E D ' ' 9 G R I I h lX kp.-cal. per mole of broFigure 11. Integral Heat of mine. This figure comReaction for Graphon pares favorably with the Upper curve = integral heat of reacexpected value and must tionof bromine with Graphon in CClr be due to the addition slurry; lower curves = chord plot of differential heat of reaction of broof bromine to double mine and Graphon as a function of amount of bromine adsorbed; two b o n d s . T h e h e a t of straight lines with zero slope = the substitution of bromine type differential heat of reaction curve obtained from bromine and is c o n s i d e r a b l y less carbon blaek than the observed heat of reaction-approximately 10 to 12 kg.-cal. per mole ( I S ) . Resonance interaction of the ethylenic-type double bonds with the benzenoid rings, which constitute most of the carbon black particle, should be accounted for as well as stearic effects resulting from the strained state of the final configuration. Yo explanation can be offered for the fact that the SRF black has a greater value of AHA than the RF and EPC blacks. The heat of reaction for the reversible addition of bromine to phenanthrene is 7.4 kg.-cal. per mole ( 3 7 ) . The somewhat greater values which were observed as the differential heat of reaction to the right of the discontinuity are possibly due t o the equilibrium of bromine with structures resembling phenanthrene which exist on the surface of the carbon blacks. This portion of the curve does not remain linear when the equilibrium concentramoles per liter. tion of bromine reaches a value greater than
I 4 K
W \
2.0
z 0 t-
a 1.5
v)
0
a
IA.
0
'i
2 I A
a
a
s 0.5 c
Y 2.0
5
I .o
o
1.0
w
r
a
0.5 MOLES
1.0
Sf,
1.5
A D S O R B E D per
2.0
2.5
GRAM x104
Figure 10. Integral Heat of Reaction of Bromine with RF-2 Black
0.5 1.0 1..5 . MOLES Br, ADSORBED per GRAM a IO4
0
Figure 12. Integral Heat of Reaction for All Blacks Studied in Region below Point of Discontinuity-
152
INDUSTRIAL AND E N G I N E E R I N G C H E M I S T R Y
The point of discontinuity, in the plots of the integral heat of reaction against the amount of adsorbed bromine, has been taken as a measure of the number of double bonds existing on the surface of the carbon black>.
Vol. 43, No. I
ference in the adsorption 01 simplc olefins and dihydroinyrceiie on carbon blacks, observed by Schaeffer, Polley, and Smith (40), is due to chemisorption, which ifi suggested as a possible explanntion of the experimental data by these authors, then their Jvorli would seem to sub8t)antiatcthe proposal that bound rubber is due to covalent bonding between the black and polymer in solution. The existence of both double bonds and reactive hydrogt,n atoms on the surface of the black alloivs the sulfur to react n-it11 the black during vulcanization in the same manner t,hat suifui reacts t,o cross-link polymer chains. On this basis, the black a n d the polymer are chemically combined during vulcanization into a continuous three-dimensional cross-linked matrix. The strwsstrain properties of such a inatris should depend on the cstent to which the polymer has been imnioi)iliztd at the surfrtcr ol tl,c tilack through carbon ~)laclr-aulfur-I,ol!,mei. Iionds. DEPESDENCE O F STRESS-STRAIN PROPERTIES OF CRO LIKKED MATRIX Oh- CHE3IlCAL PROPERTIES O F CARBOK B1.A SURFACE
A carbon black particale is, accordiiig to s - w y c,vitlence, coniposed of disordered groups of microcr\.stwiIilie graphite units known as "parallel layer groups." However, clements other than carbon are constit,ueiits of carbon black. and it would not seem unreasonable to assunic that thcse elements are situated predominantly a t the edges of the cartlon plnnes. On this basis the carbon black particle may be considcrcd as :i disordcred agglonierate of layers of large polynuclear lienzerioid h>-drocarbons. A scheinatic diagraiii of one such plane is shown in Figure 14. The representat,ion of a carbon black surfacc as containing various organic functional groups as part of the molecular structure of the particle is useful in correlating man>- of thc propcrties associated with the material. The pH characteristic of channcl-t~.pcI~laclcsarises from thc existence of large numbers of hydroxyl and cahosyl groups on the surface of these blacks. The origin of thesc groups on channc.1 blacks and their lack of abundance o n the surfacc of lurnace blacks is probably the result of the method of prqtaratioii of t'hc two types of carbon blacks. Ch:irinol blacks are pr(q1ared i n such a manner that there will be frce acress of the nr~n.lytlcposited carbon to the oxygen of t.he atiiiosphere a t a relativcly high temperature. The oxygen can react ait.h the doublc bonds or active hydrogen atoms to give, as the end product in a series of reactions, the various oxygen-containing functional groups found by Villers (41). Furnace blacks apparently formed in an atmosphere deficient in oxygen and are not exposed to the oxygen of the atmosphere until they have been cooled t o iL temperature where reaction with oxygen will be relatively slow. Furnace blacks would be expect,ed, on this basis, to have a Iargcr number of double bonds or active hydrogen atoms per unit surface area than channel blacks. Column 5 of Table IV indicates that the above conclusion is in accord with the esperiinent'al evidence. When channel blacks are heated at, temperatures between 500" and 1000" C. the o. en-containing groups are drcomposed and removed as carbon dioxide, carbon monoxide, and water. The remaining carbon and hydrogen a t o m o n the surface undergo rearrangement a t these temperatures, and it is therefore not surprising to find that the calcined EPC-2 black has a much greater number of double honds t,han the parent black. The existence of hydroperoxides arid a-methylene carbon atoms on the surface of carbon blacks may account for the familiar phenomena of bound rubber-that portion of polymer found to be insoluble after milling polymer arid blac~ktogether. It may be assumed that eit>herthe hydroperoxide present in the polymer chains will react to link t'he polymer to the black through a carbon-carbon bond or that a hydroperoxide group on the surface of the black will react in a similar manner with the p o l p i e r segments. There is also the possibility of carbon-oxygen-carbon bonds being formed between black and polvmcr. If t,he dif-
In the opinion of thc authors the w-orli of rerract,ion 01' lo:ic!cd stocks is dependent o n the extent to which the polymer is Louiitl or immobilized at the surface of the solid inclusion. If this poiiit of view is correct it should be possible to express the work of' rctraction of a polymer-pigment syst'em, as characterized by til(: constant as a function of the number of possiblc sites 011 the surface of the pigment, where the polymer may be cheniic*ally bound. This latter quantit'y has been derived from the heat oi' reaction of bromine with carbon black as described. Figure 15 shows the values of [+]'E5 plotted as a function of the number o/' available sites for chemical reaction per gram of pigment. 111 calculating the number of available double bonds it was assuiii('(1 that some fract,ion of the total surface area of t.he pigment Tvill not be avtiilable for adsorption of polynicr. The fraction oi' t,lie total area which appears as internal area has been taken as 1 - I , where r is the ratio of the area calculat,ed from electron niicrographs to the area obtained from nitrogen adsorption data. Ilaiineiiberg and Coliyer ( 1 6 ) in their work with bound rubher also came to the conclusion that the surl:rcc area as mtwsured l)y
+-
Figure 14. Schematic Diagram of Single Plane i n Parallel Layer Group of Carbon Black Particles Various organic functional groups have been placed around edges of polynuclear benzenoid hydrocarbon; figure at upper right = proposed reaction between polymer, sulfur, and carbon black during vulcanization
O4
153
INDUSTRIAL AND ENGINEERING CHEMISTRY
January 1951
I-
OF SULFUR COMBINED WITH PIGMENT TABLE V. AMOUNT RF-3
Pigment
EPC-2 RF-1 RF-2 Graphon Silica
0.4
Moles Sulfur/Gram Pigment X 104
0.39 0.55
1.17
0.00 0.00
The concept that carbon black is chemically bound to the polymer mainly through sulfur bonds whereas the inorganic pigments are bound principally through van der Waals type adsorption forces explains many of the properties of carbon blacks and also the failure of inorganic pigments of equivalent size and area to behave in a manner similar to carbon blacks.
0.2
( 6 r ) ' 5 i d (AVAILABLE REACTIVE SITES per GRAM x IO*)"^
Figure 15. $I Plotted against Square Root of Number of Available Reactive Sites per Gram of Black
I
nitrogen adsorption was not all available for adsorption of polymer. Except for the calcined EPC-2 black all the nine pigments examined fall on a smooth curve so that to a first approximation it may be written that (11)
$I = b(6r)l'Z
where b is a constant and 6 is the number of reactive sites per gram of pigment. Thus it is possible to express the effect, of a pigment on the elastic properties of a cross-linked matrix as a function of the ability of the pigment to immobilize the polymer at the eurface through chemical bonding.
ACKNOWLEDGMENT
The authors wish to thank F. W. Stavely and members of the Firestone Research staff for their interest and assistance in this work and gratefully acknowledge the helpful discussions with Walter Nudenberg of the University of Chicago and F. T. lYa11 of the University of Illinois. LITERATURE CITED
(1) Aleksondrov, A. P., and Lararkin, J. S., Rubber Chenr. and Technol.. 19. 42 (1946). (2) Amon, F. H., Smit'h, W: R., and Thornhill, F. S., IND. ENG. CHEM.,ANAL.ED., 15, 256 (1943). (3) Anderson, R. B., and Emmett, P. H., J . Applied Phys., 19, 367 (1948). (4) Andrews, R. D., Tobolsky, A. V., and Hanson, E. E., J . Applied Phys., 17, 352 (1946). (5) Beebe, R. A., Biscoe, J., Smith, W. R., and Wendell, C. B., J. Am. Chem. Sac.. 69. 95 (1947). (6) Beebe, R. A., Poll'ey, 'M. H., Smith, W. R., and Wendell, C. B., Zbid., 69, 2294 (1947). (7) Biscoe, J., and Warren, G. E., J . Applied Phys., 13, 364 (1942). (8) Blanchard. A. F.. and Parkinson. D., in Dawson. T. R.. "Pro-
ceedings of the Second Rubber Technology Conference," p, 414, London, Heffer, 1948. (9) Boyd, G. E., and Harkins, W.D., J . Am. Chcm. Sac., 64, 1190
i 3 u)
1.7
m
A
0 -
I.&
A
m -
A
Y
I
w -
SILICAY
' A
cb
GRAPHON
In * J 0 - l 0,osJ lI ~ ' 0. "10 '
V&V,
I , ' I l 0,20 " ' ' ' I
' I 0,15 VOL.PIGMENT/ VOL. POLYMER
Figure 16. Combined Sulfur in Cured Stocks as Function of Volume Fraction of Black Present
f
The determination of the amount of combined sulfur ( 2 7 ) in vulcanizates containing these blacks has yielded further evidence in support of the concept that carbon black and polymer are bound through carbon-sulfur-carbon bonds into a continuous matrix. Figure 16 shows the percentage of combined sulfur based on the weight of polymer present plotted against the amount of pigment present in the stock per unit weight of polymer. For the carbon black the curve has a positive slope whereas for Graphon and silica the amount of combined sulfur remains constant as the aniount of pigment is increased. The slope of the curve gives the amount of sulfur either combined with the black or combined because of the presence of the black; Table V shows that this value is close to the number of reactive sites found to be present on the black surface. The fact that the amount of combined sulfur is always greater than the number of reactive sites may be due to the fact that some of the cross links are in the form of disulfide rather than thio ether linkages. It may seem somewhat surprising that all the available sites on the surface of the carbon blacks have been utilized. However, this is not unreasonable when it is remembered that accelerator and sulfur are probably in excess in the region of the carbon black particle because of the adsorptive properties of the surface.
(1942). (10) Clark, G. L., Eckert, A . C., and Burton, R. L., IND.ENG.CHEM., 41, 201 (1949). (11) Cohan, L. H., Zndia Rubber World, 117, 343 (1947). (12) Cohan, L. H., and Spielman, R. S., IND. ENG.CHEM.,40, 2204 11948). (13) Conn, J. B., Kistiakowsky, G. B., and Smith, E. A., J . Am. Chem. Soc., 60,2764 (1938). (14) Copeland, L. E., and Mooney, M., J . Applied Phys., 19, 450 (1948). EXG.CHEM.,40, 2199 (1948). (15) Dannenberg, E. M., IND. (16) Dannenberg, E. M., and Collyer, H. J., Zbid., 41, 1607 (1949). (17) Elliott, D. R., and Lippmann, S. A , , J . Applied Phus. 16, 50 (1945). (18) Flory, P. J., Rabjohn, N., and Schaffer, M. C., J . Polymer. Sci., 4, 225 (1949). (19) Gee, G., Trans. Faradav Soc., 42, 585 (1946). (20) Gehman, S. D., J . Applied Phys., 19, 456 (1948). (21) Goldfinger, G., J . Polymer Sci., 1, 58 (1946). (22) Guth, E., J . Applied Phys., 16, 20 (1945). (23) Hanson, E. E., and Halverson, G., J . Am. Chem. Soc., 70, 7 7 9 (1948). (24) Harkins, W. D., Jura, G., and Loeser, E. H., Ibid., 68, 554 (1946). (25) Johnson, C. R., IND. ENG.CHEM.,20, 904 (1928). 66, 1356 (26) Jura, G., and Harkins, W. D., J . Am. Chem. SOC., (1944). (27) Kelly, W. J., Zndia Rubber World, 66, 491 (1922). (28) Ladd, W. A., and Wiegand, W. B., Rubber Age ( N . Y . ) ,57, 299 (June 1945). (29) Lister, M. W., J . Am. Chem. Soc., 63, 143 (1941). 130) Meser. K. H.. and Van der Wvk. A. J. A,. Helu. Chim. Acta, 29, 1842 (1946). (31) Mochulsky, M., and Tobolsky, A. V., IND. ENG.CHEM.,40, 2155 11948). (32) Mooney, M., Wolstenholme, W. E., and Villers, D. E., J. Applied Phys., 15, 324 (1944). (33) Mullens, L., J. Rubber Research, 16, 275 (1947). (34) Nauton, W. J. S., and Waring, J. R. S., Trans. Insl. Rubber Znd., 14,340 (1939).
INDUSTRIAL AND ENGINEERING CHEMISTRY Parkinson, D., Ibid., 16, 87 (1940). Ibid., 25, 267 (1949). Price, C. C., J . Am. Chem. SOC.,58, 1834 (1936). Rehner, J., Jr., J . Applied Phys., 14,638 (1943). Roth, F. L., andwood, L. A., Ibid.,15, 7 4 9 , 7 8 1 (1944). Schaeffer, Polley, a n d Smith, J . P h y s . & Colloid Chern., 54, 227 (1950).
Smallwood, H. M., J . Applied Phiis., 15, 758 (1944). Smith, W. R., and Schaeffer,W. D., in Dawson, T. R., "Proceeding of the Second Rubber Technology Conference," p. 403, London, Heffer, 1948. Thornhill, F. S.,and Smith, Vi-.R., IXD.ENG.CHEM.,34, 218 (1942).
Vol. 43, No. I
(44)
Tobolsky, A. V., Prettyman, I. B., a n d Dillon, J. H., J . Applied
(45) (46) (47) (48) (49) (50) (51) (52) (53)
Uberreiter, K., Angew. Chem., 54, 508 (1941). Villers, D. S..J . Am. Chem. Soc.. 69, 214 (1947). Ibid.,70, 3655 (1948). ITall, F. T., J . Chem. Phys., 10, 132, 485 (1942). Watson, J. H. L., J . Applied Phys., 20, 747 (1949). ITeiss, J., Trans. Inst. Rubber Ind., 18, 32 (1942). Wiegand, W. B., IND.ENG.CHEar., 17, 939 (1925). Ibid., 29, 963 (1937). Z a p p , R. L., IND.ETG.CHEM.,36, 128 (1944).
Phys., 15, 380 (1944).
RECEIVED Akpril24, 1950. Presented before the Division of Rubber Chemistry, 117th Meeting .IXERICAX CHEUICAL SOCIETY, Detroit, Mich.
Structure-Property Relationships for Neoprene Type W W. E. MOCHEL AND J. B. NICHOLS Experimental Station, E . I . drc Pont de Nemours & Co., Inc., Wilmington, Del.
T h e molecular structure of Neoprene Type W, a new chloroprene polymer having improved stability and processing characteristics, was investigated t o elucidate t h e structural basis for some of the improved properties shown by the polymer. Neoprene Type W was shown by fractional precipitation to have a molecular weight distribution more uniform t h a n t h a t of other neoprenes or GR-S and approaching t h a t of natural rubber. Greater uniformity of molecular structure than in other neoprenes or GR-S was indicated b y viscometric and osmotic molecular weight measurements. Ozonolysis and examination of fragments gave no evidence of lateral double bonds in Neoprene Type W.
The more uniform molecular weight distribution of Neoprene Type W appears a t least partially responsible for the improved processing characteristics observed. The improved compression set and greater ease of crystallization of Neoprene Type W are attributed to its uniformit) of molecular weight distribution and structure. These results illustrate the important effects relatively small changes in molecular structure ma? have on the physical properties of high polymers. Furthermore, they indicate the value of structure studies in the understanding of requirements for t h e preparation of improved polymers. Knowledge of the structure of Xeoprene Type W should aid in making the most effective use of this new neoprene.
NE
moleculitr weight distribution hsvc iwrn shown to be pletlominant factors jn determination of the procesahility of GIL-S :tnd natural rubber ( 7 ) . Fractions of GR-S ranging in osmotic molecular weight from 23,600 to 1,650,000 exhibited re1ativc:ly constant compression set values in tread-type vulcanizates, but the values were all 1ov;er than those for the whole polymer (21). TVhile the effect of molecular weight distribution on stress-strain characteristics is not clear, it has been shown that GR-S vulcanimtcs increase in tensile strength with increasing molecular weight to a limiting value a t a number average molecular w i g h t of about 400,000. Material of molecular weight below 24,000 appears to act cssentiall? as an inert diluent ( 7 ) .
OPREXE Type VV is a general-purpose polychloroprene elastomer which is superior to other commercial neoprenes in many properties (3,4). For example, it is outstanding in stability, showing essentially no change in plasticity during prolonged storage, whereas Seoprene Type GK exhibits a marked plasticity increase and thcn becomes progressively tougher. Similarly, dilute solutions of Xeoprene Type FV in benzene show relatively little viscosity change in 3 months, illustrating the stability obtainable in cements. The processing characteristics of Neoprene Type 15' are much like those of natural rubber in conventional operat,ions, such its milling, calendering, and cstrusion. Properly compounded vulcanizates of Type W rcsemble those of the other neoprenes in stress-strain properties and resistance to oils, sunlight, ozone, and flame, but, in addition, the Type 'iV vulcanixates have much lower coinprcssion ,set. However, the vulcanixates have slightly lower resilience and generally poorer resistance to crystallization a t low temperatures. These differences in properties have been attained by an improved emulsion polymerization of chloroprene. Since Sooprene Type W is different from other neoprenes in some of its properties, an investigation of the molecular structure of Type m' was undertaken. Some properties of elastomers are influenced markedly by average molecular weight and molecular weight distribution (8, 16). For example, optimum processing properties for a series of sodiumcatalyzed polybutadienes were reported for a polymer having a moderately high average molecular weight and relatively narrow molecular weight distribution ( 2 ) . Molecular weight and 1
EXPERIMENTAL
~IATERIALS. Standard commercial samples of Neoprene Type
\Y (lek 111 and 112) having a Mooney viscosity of 48 were
used for some of the molecular weight measurements. For the fractionation and some other experiments similar polychloroprenes prepared in laboratory equipment were employed. They corresponded t o t,he commercial plant samples in Mooney viscosit.y and intrinsic viscosity as well as vulcanizate properties. T h e molecular weight measurements for GR-S reported here were: obtained by exatninat'ion of Glt-9 X-478, a 41' F. rubber believed to be one of t.he most uniform commercial types of GR-S. ( A sample from lot FA 91,003 was supplied by the Mansfield Tire and Rubber Co.) T h e polymers were dissolved in benzene, the solutions filtered through sintered glass plates to remove talc and other inorganic insoluble materials, and the polymer precipitated with methanol. This treatment was repeated and the products were finally dried
