Polymer 54 (2013) 685e693

Contents lists available at SciVerse ScienceDirect

Polymer
journal homepage: www.elsevier.com/locate/polymer

Confocal Raman imaging, FTIR spectroscopy and kinetic modelling of

the zinc oxide/stearic acid reaction in a vulcanizing rubber
Pellegrino Musto a, *, Domenico Larobina b, Salvatore Cotugno c, Paolo Straffi c, Giuseppe Di Florio a, d,
Giuseppe Mensitieri d
a
Institute of Chemistry and Technology of Polymers, National Research Council of Italy, Via Campi Flegrei 34, Olivetti Buildings, 80078 Pozzuoli (NA), Italy
b
Institute of Composite and Biomedical Materials, National Research Council of Italy, P.le Tecchio 80, 80125 Naples, Italy
c
Bridgestone Technical Center Europe S.p.A., Via del Fosso del Salceto 13/15, Rome, Italy
d
Department of Materials and Production Engineering, University of Naples “Federico II”, P.le Tecchio 80, 80125 Naples, Italy

a r t i c l e i n f o a b s t r a c t

Article history: The reaction of zinc oxide (ZnO) with stearic acid (StA) to form zinc stearate (ZnSt) has been investigated
Received 18 September 2012 experimentally in a model matrix (unvulcanized styreneebutadiene rubber) by using confocal Raman
Received in revised form microscopy and FTIR transmission spectroscopy. The heterogeneous nature of the reacting system has
5 December 2012
been confirmed. The Raman analysis has revealed the coreeshell structure of the product, which is
Accepted 6 December 2012
formed via the gradual shrinkage of the ZnO core and the concurrent formation of a surrounding ZnSt
Available online 16 December 2012
shell of increasing thickness. FTIR spectroscopy has provided information about the molecular state of
aggregation of StA when dissolved in the rubber, as well as quantitative information on the reaction
Keywords:
Zinc oxide
kinetics. The kinetic behaviour of the system has been interpreted using a semi-quantitative heteroge-
SBR rubber neous reaction model grounded on the Raman imaging results, which was able to catch the essential
Vibrational spectroscopy features of the phenomenon and to simulate reliably the experimental conversion vs time data at three
different temperatures.
Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction rubber processing is represented by undesired side reactions,

whose occurrence increases with the number and reactivity of the
Vulcanization (i.e. the cross-linking reaction by sulphur) is one of formulation components. For instance, zinc oxide is still the best
the oldest and best developed processes in the rubber industry. known activator for sulphur vulcanization.
Since its discovery by Goodyear in 1839, it has been continuously Zinc oxide (ZnO) is in many vulcanisation systems a precursor to
improved by the introduction of new ingredients (accelerators, zinc-derived accelerators; it reacts with most accelerators to form
activators, retarders), in order to improve mechanical properties. the highly active zinc salt [1,2] Complex formation of the zinc ion
Nowadays, in the tyre industry, where the process represents the with different accelerators is critical to get efficient curing. In
core technology, a vulcanizing rubber formulation is a very complex addition, there is also evidence that the inclusion of ZnO in
system comprising tens of components. Most of these formulations a compound reduces heat-build-up and improves abrasion resis-
have been developed and optimized by skilful and lengthy tech- tance. Due to its high thermal conductivity, ZnO acts as a ‘heat sink’,
niques based on trial-and-error but, owing to the complexity of the which accepts frictional energy dissipating local heat concentra-
chemistry involved and the multiplicity of ingredients, the mecha- tions without a large increase in internal temperature. It has also
nism of vulcanization has not been fully elucidated yet. It is gener- been found that ZnO improves heat resistance of the vulcanisates
ally recognized that the vulcanizing rubber is a heterogeneous and their resistance to the action of dynamic loading [2e6].
system which, in turn, gives rise to a heterogeneous vulcanizate at Moreover, several fatty acids are employed in rubber formulations
the end of the process. How to precisely measure the degree of as plasticizers, but for a few of them (for instance stearic acid, StA) it
heterogeneity and to control it for a rational product design, has been postulated that, beside acting as an internal lubricant [7],
remains an open issue [1]. Another relevant problem encountered in they play a further role as co-activators [4]. By reacting with stearic
acid, zinc oxide forms water and the hydrocarbon-soluble zinc
stearate (ZnSt), which has been claimed to be the actual activator
* Corresponding author. species [7]. In fact, when the zinc coordinates the carboxylate
E-mail address: musto@ictp.cnr.it (P. Musto). ligands of stearic acid, the solubility and reactivity is strongly

0032-3861/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.polymer.2012.12.021
686 P. Musto et al. / Polymer 54 (2013) 685e693

enhanced. It has been reported [8aeh] that, in sulphur vulcaniza-

tion of rubbers, the crosslink density increases with increasing zinc
stearate concentration, while the effect of stearic acid consists in
increasing the initial cure and in lowering the level of polysulfidic
cross-links, thus enabling a more efficient use of sulphur [9,10].
It has also been suggested that both these effects are caused by
stearic acid allowing a better dispersion of ZnO. In another study
[11], it was observed that, if the vulcanizing medium contains zinc
stearate but no stearic acid, the end-properties of the final products
worsened.
From the above arguments it follows that a quantitative char-
acterization of the zinc oxide/stearic acid reaction, not only in terms
of overall conversion but also with respect to the heterogeneity of
the system, is of paramount importance towards a deeper under-
Fig. 1. Particle size distribution (in terms of equivalent particle diameter) of ZnO as
standing of the vulcanization mechanism. Previous literature
determined by SLS measurements.
reports have shown that, depending on the mixing conditions, the
two reactants can already form zinc stearate during the com-
pounding of the rubber. The calorimetric (DSC) results of Kruger solution casting on Petri dishes, followed by slow solvent
and McGill [12] indicated the formation of zinc stearate and its evaporation (24 h at ambient temperature). Residual traces of
subsequent reaction with sulphur, but the phase separation of the solvent were eliminated under vacuum at 40
C. Complete
system components was not investigated. elimination of the solvent was assessed by the absence of solvent
In the present contribution we examine the interaction between bands in the FTIR spectrum of the film.
zinc oxide and stearic acid in a medium (uncrosslinked styrenee The sample with only SBR and stearic acid, used to investigate
butadiene rubber, SBR) suitable to simulate the actual vulcanizing the StA monomer-to-dimer equilibrium, and several samples con-
system. The dispersion of the components before and after the taining SBR and zinc stearate, used as standards for analytical
reaction is investigated by confocal Raman microspectroscopy. This purposes, were prepared in the same way as the ternary mixture,
technique offers unique advantages in the characterization of with the appropriate amount of the respective components.
heterogeneous polymer systems: it provides good contrast due to
the specificity of the spectrum, molecular-level information and the 2.3. Techniques
possibility to implement quantitative or semi-quantitative analysis.
The limiting spatial resolution (0.8e1.0 mm in the focal plane) is Raman spectra were collected by a confocal Raman spectrom-
adequate for the systems under scrutiny. Furthermore, the state eter (Horiba-Jobin Yvon Mod. Labspec Aramis) equipped with
of aggregation of stearic acid in the rubber matrix and its a HeeNe laser source emitting at 632 nm. The sample was placed
reaction with zinc oxide are investigated by FTIR spectroscopy, on a piezo-electrically driven microscope stage with a x,y resolution
which provides information on the H-bonding interaction of the of 10  0.5 nm and a z resolution of 15  1 nm and scanned at
stearic acid dimer and kinetic data at different temperatures. a constant stage speed with a 2.0 mm step size. The 180
back-
These are used to test the reliability of a semi-quantitative kinetic scattered radiation was collected by an Olympus metallurgical
model developed on the basis of the spectroscopic evidences and objective (MPlan 50, NA ¼ 0.75) with confocal and slit apertures
based on a simple heterogeneous reaction scheme. both set to 200 mm. A grating with 1200 grooves/mm was used
throughout. The radiation was focused onto a Peltier-cooled CCD
2. Experimental detector (Synapse Mod. 354308) operating in the Raman-shift
range 3200e100 cm1 for single point measurements or in the
2.1. Materials range 1600e200 in the case of the mapping experiments, to
reduce the acquisition time. Exposure times were 30 s for the single
Styreneebutadiene rubber (SBR) had a styrene/butadiene molar point acquisition and 25 s for the mapping experiments. The data
ratio of 20/80 and a trans/vynil/cis C]C ratio of 24/56/20. Zinc oxide gathered by the instrument were converted to ASCII format and
with a purity >99% was supplied by Bridgestone Technical Center transferred to the MATLAB (Mathworks, Natick, MA) computational
Europe (TCE), Pomezia, Italy. The particle size distribution of ZnO, as platform, for further processing. The Raman images were elabo-
determined by static light scattering (SLS), is reported in Fig. 1. The rated by programs written in our laboratory, making use of the
number average equivalent diameter (Dn ) was 1.775 mm. Reagent image processing facilities and the surface interpolating algorithms
grade stearic acid and zinc stearate were supplied by Aldrich of the MATLAB environment.
Co., Milan (Italy) and were used without further purification. FTIR spectra were collected in the transmission mode using
a System 2000 spectrometer from PerkineElmer. The instrument
2.2. Sample preparation was equipped with a Germanium/KBr beam splitter and
a deuterated tryglycine sulphate (DTGS) wide-band detector. The
The adopted composition of the reactive mixture was represen- instrumental parameters adopted for the spectral collection were
tative of a commercial formulation. To prepare the ternary reactive as follows: resolution 4 cm1, scan speed [i.e. Optical Path
mixture, 85 g of SBR were dissolved in cyclohexane (5% wt/vol) at Difference (OPD) velocity] ¼ 0.20 cm s1, spectral range 4000e
ambient temperature for 24 h under vigorous mechanical stirring. 400 cm1. For the kinetic measurements, a single data collection
After complete dissolution, 9.0 g of stearic acid were added to the was performed for each spectrum, which took 6 s to complete. It
highly viscous solution, together with a suspension containing 4.0 g was found that, even with a single acquisition, the signal-to-noise
of Zinc oxide in 50 mL of cyclohexane. The stirring was continued ratio of the spectra (5000:1) was suitable for quantitative
until a homogeneous and stable dispersion of zinc oxide was analysis.
attained, as demonstrated by the appearance of a uniform milky FTIR measurements were performed at different temperatures
colour. Films with a thickness of 20e30 mm were prepared by on SBR containing only StA and on the reactive system comprising
 P. Musto et al. / Polymer 54 (2013) 685e693 687

SBR, ZnO and StA. Tests were carried out in an environmental 1300 and 1000 cm1 is essentially coincident (compare Fig. 2C and
chamber constructed in house by modifying the commercially D), which, in the present context, prevents the possibility to
available SPECAC 20100 cell. This unit was controlled by a Euro- discriminate the two molecular species from their Raman signals.
therm PID temperature regulator mod. 2416, with an accuracy of A preliminary Raman analysis was carried out to test the state of
0.5
C. The chamber was directly mounted in the spectrometer aggregation and the degree of dispersion of stearic acid in SBR
and the spectra were taken at selected time intervals to monitor in- achieved with our solution casting procedure. The mixture was
situ the resulting chemical changes. All measurement was carried initially prepared without ZnO, to prevent the possible reaction
out under a continuous flux of N2 (60 cm3 min1). between the components. The Raman image collected on
a 34  34 mm2 area is reported in Fig. 3. There are regions of the map
3. Results and discussion where the peaks characteristic of the crystalline StA phase are clearly
discernible, while in other areas these peaks are undetectable (see
3.1. Raman spectroscopy inset of Fig. 3). The signals’ intensity was converted to relative
concentration of scattering species making use of a general
In Fig. 2AeD are reported the Raman spectra of the system formalism whereby the sample and laser variables are factorized
components. with respect to the collection variables [16]. Thus, we may define
Zinc oxide displays two sharp peaks at 437 and 98 cm1 a specific scattering intensity L, expressed in photons sr1 cm2 s1
(E2 modes) [13] which occur in an interference-free region and
provide the required spectroscopic contrast. Stearic acid also L ¼ PD bD[ (1)
displays intense inelastic scattering and a series of peaks in a region
where the rubber matrix does not interfere (see inset of Fig. 2C). where
The intense and sharp components at 1296, 1130 and 1062 cm1
have been assigned, respectively, to the CH2 twist, and to the PD ¼ laser power density (photon s1 cm2)
symmetric and anti-symmetric CeC stretching modes [14,15]. The b ¼ ðds=dUÞ=N ¼ Raman cross section
prominent intensity and the sharpness of these peaks are associ- ðcm2 molecules1 sr1 Þ
ated with the ordered crystalline structure, as demonstrated by D ¼ density of scatterers (molecules cm3)
the inset of Fig. 2C which compares the spectrum in the 1600e [ ¼ path length illuminated by the laser beam (cm).
800 cm1 range taken at 30
C (blue trace in web version) and
above the melting temperature, (80
C, red trace). In particular, In the definition of b, s ¼ F=J, where F is the total scattered
the sharp triplet at 1130, 1101, 1062 almost vanishes on melting, power and Jis the irradiance of the incident laser beam. s refers to
thus providing a reliable spectroscopic signature of the crystalline the whole solid angle U around the scattering center, while
phase. b represents the fractional intensity scattered into a given direction
Being the Raman spectrum of stearic acid strongly dominated by within an infinitesimal element of the solid angle dU. N is the
CH and CC modes, no major differences are expected between the number of scattering molecules. Furthermore, an instrumental
latter and the spectrum of Zinc stearate. In fact, the region between collection function C (cm2 sr e photon1) may be defined, which

Fig. 2. Raman spectra of SBR rubber (A); ZnO (B); stearic acid (C); zinc stearate (D).
688 P. Musto et al. / Polymer 54 (2013) 685e693

Fig. 3. Confocal Raman image of a SBR/StA mixture 91/9 wt/wt. Image reconstruction performed by use of the intensity ratio I1126/I620. The inset compares the spectra collected at
the positions indicated.

expresses the fraction of the total scattered light that is collected, rubber matrix, StA segregates in the form of lenticular crystalline
analysed and detected by the instrument: domains having a size from 2 to 5 mm. The observation that, even
when focussing on a StA domain, the SBR spectrum is still dominant
C ¼ AD UD TQ (2) (see inset of Fig. 3, blue trace in web version) is an indication that
the thickness of the crystallites is lower than the spatial resolution
where along the z axis (2.5 mm).
Fig. 4A and B display the Raman images of a sample containing
AD ¼ sample area monitored by the spectrometer (cm2) 4.0 wt % of ZnO and 9.0 wt % of StA. In particular, Fig. 4A was ob-
UD ¼ the collection solid angle of the spectrometer tained by considering the intensity ratio I437/I620 and is therefore
(steradians, st) representative of the concentration ratio CZnO/CSBR, while Fig. 4B
T ¼ instrumental transmission (unitless) was reconstructed in the same way as Fig. 3 and denotes the spatial
Q ¼ quantum efficiency of the detector (e photons1) distribution of StA.
The image analysis reveals a heterogeneous system in which
Finally, the observed intensity I(n), in photoelectrons, is domains of ZnO coexist with a StA crystalline phase broadly
distributed within the sampled area. In the central region several StA
IðnÞ ¼ LCt ¼ ðPD bD[Þ,ðAD UD TQ Þt (3) crystallites coalesced forming a larger aggregate. The ZnO particles
display an elongated shape with main dimensions around 4e5 mm.
with t being the exposure time in seconds. The particle size as measured by Raman imaging compares well
If we consider two fully resolved peaks at frequencies ni and nj, with the particle size distribution of ZnO as revealed by SLS, thus
characteristic of the scattering species i and j, the intensity ratio confirming the efficiency if the dispersion processes. It is noted that
Iðni Þ=Iðnj Þ will be: in several places (i.e. the upper right and the lower left corners) the
domains of the two species are localized at approximately the
Iðni Þ b ,D ,Tðn Þ,Q ðni Þ D n same positions, while in other regions (the central area) their

 ¼ i i
i
 ¼ K i ¼ K i (4)
I nj bj Dj ,T nj ,Q nj Dj nj spatial distribution is largely different. Recalling that the analytical
signals corresponding to StA and ZnSt are coincident, this
where K ¼ ðbi ,Tðni Þ,Q ðni ÞÞ=ðbj Tðnj Þ,Q ðnj ÞÞ depends only on the observation may be interpreted assuming that, during the sample
choice of the analytical peaks and is invariant with the very many preparation procedure, although the temperature was purposely
instrumental parameters. In Equation (4), ni, nj represent moles of kept as low as possible, a partial reaction occurred between ZnO
the respective species. and StA, with subsequent precipitation of solid ZnSt crystallites
The image reconstruction was performed by use of the intensity around the ZnO particles. In other words, during the solution
ratio I1126/I620 which, according to Equation (4), is proportional to casting and subsequent drying, StA may undergo two competing
the concentration ratio CStA/CSBR. This analysis reveals that, in the processes, crystallization and reaction with ZnO: the former is

Fig. 4. Confocal Raman images of a SBR/StA/ZnO mixture 87/9/4 by wt. Image reconstruction performed by use of the intensity ratio I437/I620 (A) and of the intensity ratio
I1126/I620 (B).
 P. Musto et al. / Polymer 54 (2013) 685e693 689

strongly predominant but the latter plays a non-negligible role. This strength of the H-bonding interaction [17]. The component at
interpretation is confirmed by the FTIR spectra of the as-prepared 1755 cm1 is the n(C]O) mode of the free carbonyls (i.e. the
solution cast films which display low intensity peaks characteristic monomeric form), whereby the considerable frequency enhance-
of ZnSt (vide infra). ment is due to the stiffening of the C]O force constant as
In a further experiment, the reactive mixture was allowed to a consequence of the H-bonding dissociation [17,18]. It is evident
react at 100
C for 1 h and then cooled at ambient temperature prior that most of the StA is present in the rubber compound in the
to be analysed by confocal Raman spectroscopy. FTIR analysis dimeric form, albeit the concentration of free monomers is non-
demonstrated that, in these conditions, the StA conversion reached negligible.
its maximum value (80%). The Raman images of the above sample As expected, increasing the temperature causes a decrease of
are reported in Fig. 5A, B. the main carbonyl component and a concurrent increase of the
Firstly, it is explicitly noted that, although the thermal treatment secondary component, that is, the concentration of self-associated
has been carried out in-situ, the region sampled in Fig. 5 is different StA decreases with temperature. Fig. 7 shows a plot of the overall
from that sampled in Fig. 4. This is because the sample undergoes n(C]O) band area as a function of temperature: from 60 to 120
C
a volume expansion when heated, which prevents to localize the the integrated absorbance remains essentially constant, while
specific area imaged before rising the temperature. When the dropping afterwards. A significant decrease is also observed for the
reaction is complete the position of most residual ZnO domains remaining StA peaks, while, in the same temperature range, the
coincides with that of the ZnSt crystallites. It is inferred that at rubber bands remain essentially constant. This behaviour indicates
100
C the StA melts (Tm ¼ 60
C, as evaluated by FTIR) and migrates that above 120
C a significant amount of StA leaves the mixture by
towards the ZnO domains where it reacts forming ZnSt which evaporation; thus, only the 60e120
C range can be reliably used to
immediately crystallizes around the oxide particles being its investigate the monomer-to-dimer equilibrium.
melting point (120e130
C) well above the reaction temperature. The absorbance vs temperature curves relative to the peaks at
Some of the ZnSt domains (i.e. those denoted 1 and 2) appear 1712 and 1755 cm1 are reported in Fig. 8A.
considerably larger than the corresponding ZnO particles, while From the above data the equilibrium value for ln ðA1712 =A21755 Þ
in one case (domain denoted as 3) no Zinc oxide is detectable at and, hence, the enthalpy of the dimer formation (see Scheme 1) can
the corresponding position. It is likely that, in these cases, the be evaluated.
ZnO particles have been largely depleted by the reaction, leaving As usual, the DH value results from the slope of the
an inner core much smaller than its original size. ln ðA1712 =A21755 Þ vs 1/T diagram (see Fig. 8B) on account of the Van’t
DM ¼ eðDG=RTÞ which, in terms of absorbance
Hoff equationKEQ
3.2. FTIR spectroscopy
becomes ln ðA1712 =A21755 Þ ¼ ðDH=RTÞ þ ðDS=RÞ þ ln ðε1712 =ε1755 Þ.
As for the Raman analysis, we started with the molecular The resulting value (7.6 kcal mol1) is in reasonable agreement
characterization of StA when dispersed in the rubber matrix with with literature data on carboxylic acids (DH ¼ 10.4 kcal mol1 for
no ZnO added. For a 91/9 wt % SBR/StA mixture the peaks due to StA acetic acid in CCl4 and 13.4  0.5 kcal mol1 for stearic acid in the
appear in the form of a doublet centred at 1755 and 1712 cm1. This liquid state) [17].
is shown in Fig. 6, where the 1820e1650 cm1 range is reported for To implement the kinetic model (see Section 3.3), it is advisable
spectra taken at different temperatures in the 60e120
C interval. to obtain an estimate of the equilibrium constant at the different
It is noted that, at ambient temperature (data not reported) both investigated temperatures. In order to transform the spectroscopic
observables into KEQ DM values (whereK DM ¼ C =C 2 ), we made use
carbonyl peaks are more complex, exhibiting an unresolved fine EQ D M
structure with strong asymmetry. This is a consequence of crystal of a εD =εM ratio equal to 1.6  0.1 [19], which is representative of
field effects which vanish upon melting. The spectral data in Fig. 6 a typical carboxylic acid forming a cyclic H-bonding structure as
confirm the complete melting of StA at 60
C. The component at that represented in Scheme 1. On this basis and using a mass
1755 cm1 displays a complex profile which turns out to be due to balance on StA, one can evaluate KEQ DM according to:
the presence of underlying rubber peaks. This is demonstrated by
difference spectroscopy (see inset of Fig. 6), whereby removal of the Rubber;INI
CStA ¼ CM þ 2CD
SBR spectrum reveals the symmetric, fully resolved shape of the
1755 cm1 peak. The prominent absorption at 1712 cm1 is due to Rubber
It is explicitly noted that, here and in the following, CStA
the n(C]O) vibration of the carboxyl groups self-associated to form represents the equivalent monomeric StA concentration, i.e. the sum
dimers. The relatively low frequency of this mode reflects the of the concentration of the monomeric form and of twice the

Fig. 5. Confocal Raman images of a SBR/StA/ZnO mixture 87/9/4 by wt reacted at 100
C for 1 h. Spectra taken at ambient temperature. Image reconstruction performed by use of
the intensity ratio I437/I620 (A) and of the intensity ratio I1126/I620 (B).
690 P. Musto et al. / Polymer 54 (2013) 685e693

110
C) are reported in Table 1 along with the values of the equi-
DM ¼ C =C 2 .
librium constant, KEQ D M
When ZnO is present in the mixture the reaction is evidenced by
a gradual lowering of the n(C]O) peaks of the monomer and the
dimer and the concurrent increase of a peak at 1538 cm1 due to
the asymmetric stretching vibration of the carboxylate anion. The
corresponding symmetric stretch occurs at 1395 cm1 and is
almost completely overlapped with intense rubber peaks. It has
a very limited analytical usefulness and was not considered further.
In Fig. 9 is shown the time evolution of the FTIR spectrum of
a reactive mixture comprising 9.0 wt % of StA and 4.0 wt % of ZnO,
treated isothermally at 80
C.
To monitor the extent of reaction, the 1538 cm1 peak is to be
preferred over the 1755 or 1712 cm1 peaks because it is fully
resolved, interference-free and avoids any ambiguity resulting from
monomer-to-dimer equilibrium.
In the following, we will describe the evolution of the system by
means of the relative conversion of equivalent monomeric StA, a,
defined as:

Fig. 6. FTIR spectra in the wavenumber range 1820e1650 cm1 of a SBR/StA mixture Rubber;INI
CStA Rubber ðtÞ
 CStA
91/9 wt/wt. Spectra taken at different temperatures as indicated. The inset displays the aðtÞ ¼ Rubber;INI
(5a)
difference spectra obtained by subtracting the plain SBR spectrum collected at the CStA
same temperature.

It is assumed here, in defining a, that stearic acid is mostly

concentration of the dimeric form. In the interconversion equilib- present within the rubber and that the amount dissolved in the
rium experiment, this value is constant since no reaction of the shell of the reacting ZnO particle is negligible for the purposes of
Rubber;INI
stearic acid occurs and it is equal to its initial value, CStA . evaluating the total conversion. Accounting for the stoichiometry of
Using the values of AM (i.e. A1712) and AD (i.e. A1755), which are the reaction, which dictates that each ZnO mole reacts with 2 mol
available from the spectroscopic analysis, one can write: of monomeric StA, the relative conversion of StA can also be
9 expressed in terms of concentration of ZnSt, i.e.:
AM >>
>
CM ¼ >
εM [ = A ε
0CD ¼ CM D M CZnSt ðtÞ
AD > > AM εD aðtÞ ¼ 2, (5b)
CD ¼ >
>
ε [ ;
Rubber;INI
D
CStA

which can be written as:

where [ represents the sample thickness. It follows that:
C Rubber;INI
CM ¼  StA  A1538 ðtÞ
A ε aðtÞ ¼ 2,
1þ2 D M Rubber;INI
ε1538 ,[,CStA
AM εD
and, in terms of absorbance:
Rubber;INI
where, in the case at hand, CStA ¼ 0:291 mol=l. The values of
CM and CD at the three temperature of interest (i.e. 80, 100 and
A1538 ðtÞ
aðtÞ ¼ að0Þ (5c)
A1538 ð0Þ

The a(0) term in Equation (5c) arises from the fact that both
during the film preparation procedure and the heating step to reach
the isothermal temperature, the reaction already starts. As
a consequence, at time t ¼ 0, (when the acquisition of isothermal
Rubber ðt ¼ 0Þ < C Rubber;INI , i.e. a(0) > 0. The
kinetic data initiates) CStA StA
a(0) value has been estimated from the normalized value of
A1538(0) by means of a calibration curve constructed with SBR/ZnSt
standards (see experimental).
At the investigated temperatures (i.e. 80, 100 and 110
C) the
stearic acid is assumed to be homogeneously dissolved within the
rubber.
The a vs time curves are collectively reported in Fig. 10. As ex-
pected, the reaction rate increases considerably with temperature
but the final conversion is never complete: As shown in Fig. 10 and
confirmed by the presence of residual StA carbonyl bands at the end
Fig. 7. Overall area of the carbonyl peaks (i.e. A1755 þ A1712) as a function of temper- of the reaction, the process seems to approach, in all cases,
ature for the difference spectra of the SBR/StA mixture 91/9 wt/wt. a limiting value amax.
 P. Musto et al. / Polymer 54 (2013) 685e693 691

Fig. 8. (A): Integrated absorbance area of the 1712 cm1 peak (B) and the 1755 cm1 peak (,); (B): Van’t Hoff diagram obtained with the data of Fig. 8A.

O H O
O
CH3 CH2 C C CH2 CH3
2 CH3 CH2 C
16 16
16 O H O
OH

M D

Scheme 1. Scheme of equilibrium between monomeric and dimeric forms of stearic acid.

3.3. Modelling the reaction kinetics demonstrates that stearic acid is prevalently present in its dimeric
form, in equilibrium with the monomer in a much lower concen-
In the light of the results of Raman imaging and time-resolved tration. In the model view, both species are assumed to diffuse
FTIR spectroscopy, a theoretical model has been formulated to towards the external surface of the reacting particle (whose initial
interpret the physicalechemical process of reaction of ZnO parti- radius is indicated as R(t ¼ 0)), where, however, only the smaller
cles with stearic acid dispersed in the unvulcanized rubber matrix monomeric form is likely to dissolve, as ruled by a partition coef-
making some relevant assumptions which afford a considerable ficient, K0 . Then, StA molecules diffuse through the ZnSt shell to
simplification of the problem without missing the main features of reach the surface of the unreacted ZnO core. Here the reaction
the occurring phenomena. occurs, which is assumed to attain an instantaneous equilibrium,
Indeed, the Raman analysis shows that the reaction of ZnO and the product ZnSt precipitates, instantly increasing the thick-
particles occurs in the form of an unreacted core surrounded by ness of the shell layer.
a shell of zinc stearate. This experimental observation suggests that The evolution of molar concentration of monomeric StA within
a reaction front, within each particle, sets in at the beginning of the Shell ,is ruled by the following mass balance:
the shell, i.e. CStA
reaction process and then moves towards the center of the particle
itself, gradually converting ZnO into ZnSt. Hence, in the course of !
Shell
vCStA Shell
1 v 2 Shell vCStA
the reaction, the particles can be envisaged as a shrinking inner ¼ 2 r DStA (6)
core made of ZnO surrounded by an evolving shell of ZnSt.
vt r vr vr
A pictorial representation of the model is given in Fig. 11.
Rext(t)R(t) represents the growing thickness of the ZnSt layer, with boundary conditions:
while R(t) represents the time dependent radius of inner ZnO
core. at r ¼ RðtÞ : Shell ¼ C EQ
CStA
In formulating the model it has been assumed that the ZnO
StA  
at r ¼ Rext ðtÞ : Shell ¼ K 0 ,f C Rubber
CStA
reacting particles i) are evenly distributed throughout the rubber StA
domain, ii) have a spherical shape with mono-dispersed diameter,
and iii) do not interfere with each other, being considered as where
independent reacting systems. Moreover, the stearic acid molecules
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
are taken as uniformly dispersed within the rubber matrix at the   1 þ 1 þ 8KEQ DM C Rubber
StA
reaction temperatures. As already discussed, the FTIR analysis f Rubber
CStA ¼ DM
4KEQ

Table 1 DShell
StA is the stearic acid mutual diffusivity in the shell layer,
DM as determined at the temperatures of interest for SBR/StA
Values for CD, CM and KEQ EQ
system.
t stands for time and r is the radial coordinate. CStA represents the
stearic acid molar concentration at chemical equilibrium at the
Temperature (
C) CD [mol/l] CM [mol/l] DM [l/mol]
KEQ
shellecore interface and determines an upper limit for the overall
80 0.131672 0.027655 172.1613 conversion of stearic acid (i.e. amax < 1, see definition of amax in
100 0.127078 0.036845 93.60877
110 0.123934 0.043132 66.61962
Equation (14)). The boundary condition at r ¼ R(t) is dictated by the
assumed instantaneous reaction equilibrium, while the one at
692 P. Musto et al. / Polymer 54 (2013) 685e693

Fig. 9. FTIR spectra in the 1800e1480 wavenumber range of a SBR/StA/ZnO mixture

87/9/4 by wt. Spectra taken at increasing reaction times under isothermal conditions
(80
C).
Fig. 11. Schematic representation of a reacting ZnO particle (assumed to be of spherical
shape) at a generic instant of time t. The original size of the ZnO particle (R), the
r ¼ Rext(t) is dictated by partition of StA between rubber and ZnSt current size of the ZnO particle (R(t)) and the current size of the ZnO particle and of the
shell. surrounding ZnSt shell (Rext(t)) are evidenced.
Rubber , although variable with time, is considered
The value of CStA
to be uniform across the rubber matrix surrounding the particles. In Based on the stoichiometry of the reaction, it is possible to relate
fact, in developing the model, it has been assumed that the whole the consumption of StA to the dimension of the internal, R(t), and
process is controlled by StA diffusion through the shell, since, the external, Rext(t), radii of the shell, as follows:
characteristic time of diffusion of monomeric and dimeric StA
  4  
through the rubber and the characteristic time of the surface Rubber;INI
CStA Rubber
 CStA ðtÞ ¼ brZnO Np p R3p  R3 ðtÞ (8)
chemical reaction are considered to be much lower than the 3
characteristic time of StA diffusion within the shell layer. It is worth
   
to note that, this approach is obviously less accurate in the initial R3ext ðtÞ  R3 ðtÞ rZnSt ¼ R3p  R3 ðtÞ rZnO (9)
stages of the process, when the shell is not sufficiently thick to
guarantee the assumed diffusion control. In view of these consid-
Rubber can be evaluated using the where Rp represents the initial radius of a ZnO particle, before the
erations, the time evolution of CStA
reaction starts (i.e. Rp ¼ R(t ¼ 0)), b is the stoichiometric coefficient
following mass balance on equivalent monomeric stearic acid in the of the reaction, rZnO is the molar density of the ZnO particle and
bulk of the rubber matrix: rZnSt is the molar density of zinc stearate. After some algebra
! Equations (8) and (9) give:
Rubber
dCStA Shell
vCStA

¼ Np 4pr 2 DShell (7)
dt StA
vr RðtÞ 1
r¼Rext ðtÞ ¼ ð1  K1 aÞ3 (10)
Rp
where Np represents the number of ZnO particles per unit volume
of the system.
!1
Rext ðtÞ
R3 ðtÞ 3
1
¼ rþ 1r ¼ ðr þ ð1  rÞð1  K1 aÞÞ3 (11)
Rp R3p

where:

Rubber;INI
CStA
K1 ¼
4
brZnO Np pR3p
3

rZnO
r¼
rZnSt2
In summary, solution of the system of Equations (6), (7), (10) and
(11) provides the time evolution of CStARubber which can be used to

evaluate a, from Equation (5a), to be compared with experimentally

determined values.
Fig. 10. Relative conversion of stearic acid in zinc stearate in the SBR/StA/ZnO mixture
However, to simplify the mathematics of the model, we further
87/9/4 by wt, monitored isothermally at the temperatures indicated. Symbols refer to assumed a quasi-steady state regime for the diffusion process of StA
the measured values; continuous lines represent the kinetic model predictions. within the particle’s shell. This assumption consists in imposing
 P. Musto et al. / Polymer 54 (2013) 685e693 693

Table 2 interpreted theoretically by means of a kinetic model developed on

Values of best fitting parameters of the kinetic model. the basis of a simple heterogeneous reaction scheme.
Temperature (
C) amax k (s1) Confocal Raman microspectroscopy revealed that, at ambient
80.000 0.51 0.129 temperature, both components are phase-separated in the form of
100.000 0.74 0.138 microcrystals. When the reaction temperature is reached (80
C
110.000 0.75 0.525 and above) only zinc oxide is present in the form of particles while
stearic acid melts and gets molecularly dispersed within the rubber
that the concentration profile inside the shell instantaneously matrix. The analysis also points to a coreeshell structure of the
accommodates to the change of boundary conditions determined reacting system: stearic acid diffuses to the surface of the ZnO
by the evolution of the system. In other words, the system of domains causing the shrinkage of the ZnO core and the formation
Equations (6) and (7), is replaced by: of a shell of increasing thickness made of zinc stearate.
h   i FTIR spectroscopy confirmed that, when molecularly dispersed
Rubber
dCStA K 0 ,f CStA
Rubber  C EQ
StA
in the rubber matrix, StA is predominantly self-associated in
StA 4p
¼ Np DShell   (12) a dimeric form. Time-resolved measurements also yielded quanti-
dt 1 1 tative information on the kinetic behaviour of the reacting system.

Rext ðtÞ RðtÞ In the light of the Raman imaging analysis a simple, yet physi-
By substituting the relationships for R(t) and Rext(t) into Equa- cally sound, model was developed to interpret the FTIR spectros-
tion (12) and expressing a according to Equation (5a), we obtain the copy results. The model is based on the assumption that the
following differential equation: controlling step of the reaction kinetics is the diffusion of the

4pNp DShell 0n    o
StA Rp K
Rubber;INI
Rubber;INI
f CStA ð1  amax Þ  f CStA
Rubber;INI
ð1  aÞ
da CStA
¼ (13)
ð1  ð1  rÞK1 aÞ3 ð1  K1 aÞ3
1 1
dt

with stearic acid within the zinc stearate shell surrounding the reacting
EQ
! particles. The model is able to catch the main experimental features
Rubber;INI CStA
CStA  f 1 and displays a reasonable fitting capability of the conversion vs
K0
amax h (14) time data at three different temperatures (80, 100 and 110
C).
Rubber;INI
CStA
References
The constant K1 corresponds to the molar defect of stearic acid
[1] Ikeda Y, Higashitani, Hijikata K, Kokubo Y, Morita Y, Shibayama M, et al.
with respect to the stoichiometric ratio between stearic acid and
Macromolecules 2009;42:2741e8.
zinc oxide and, in the case at hand, takes a value of 0.32. The [2] Susamma AP, Mini VTE, Kuriakos AP. J Appl Polym Sci 2001;79:1e8.
Rubber;INI
nominal initial concentration of stearic acid, CStA , is [3] Moore CG, Saville B, Watson AA. Rubber Chem Technol 1961;34:795.
0.310 mol/l. Moreover, based on the molar densities of zinc oxide [4] Kresja MR, Koenig JL. The nature of sulfur vulcanization. In: Elastomer tech-
nology handbook. New Jersey: CRC Press; 1993. p. 475.
and zinc stearate, a value of 40 for ð1  rÞ was estimated. Inserting [5] Coran AY. Vulcanization in science and technology of rubber. In: Eirich ER,
these values into Equation (13) one obtains: editor. San Diego: Academic Press; 1978. p. 291.
[6] Domka L, Krysztafkiewicz A. Int Pol Sci Technol 1980;7. T/18.
da kff ½0:31ð1  amax Þ  f ½0:31ð1  aÞg [7] Sullivan AB, Hann CJ, Kuhls GH. Rubber division meeting. Toronto, Canada:
¼ (15) American Chemical Society; 1991.
ð1 þ 12:8aÞ3 ð1  0:32aÞ3
1 1
dt
[8] (a) Adams HE, Johnson BL. Ind Eng Chem 1953;45:1539;
(b) Duchacek V, Kuta A, Pribyl P. J Appl Polym Sci 1993;34:743e6;
(c) Bravar M, Jelencic J, Dabetic M. Kautschuk Gummi Kunststoffe 1986;39(4):
where k stands for 4pNp DShell Rubber;INI
StA Rp K’=CStA . Equation (15), in its 314e8;
numerically integrated form, has been used to fit the experimental (d) Coran AY. Rubber Chem Technol 1964;37(3):679e88;
data available from the FTIR analysis, with, amaxand k as adjustable (e) Xie HQ, Ao ZP, Guo JS. J Macromol Sci Phys 1995;B34(3):249;
(f) Campbell RH, Wise RW. Rubber Chem Technol 1964;37(3):650e67;
parameters. The initial condition for a has been taken as the (g) Zaborski M, Slusarski, Donnet JB, Papirer E. Kautschuk Gummi Kunststoffe
experimental value measured spectroscopically at t ¼ 0 (respec- 1994;47(10):730e8;
tively equal to 0.15, 0.28 and 0.44 at 80, 100 and 110
C). (h) Powell CE, Preparation of rubber compositions with organoclays, US Patent
Appl. 2005, n. 2005/0090584 A1.
The results of the model fitting are compared with the experi- [9] Baldwin FP. Rubber Chem Technol 1972;45:1348.
mental FTIR data in Fig. 10, while the values determined for best [10] Saville B, Watson AA. Rubber Chem Technol 1967;40:100.
fitting parameters, amaxand k, are reported in Table 2. The fitting [11] Bridgestone TCE, Private communication.
[12] Kruger FWH, McGill WJ. J Appl Polym Sci 1991;42:2651.
quality is satisfactory, despite the adopted simplifying assumptions. [13] Damen TC, Porto SPS, Tell B. Phys Rev 1966;142:570e4.
The dependence of the fitting parameters on temperature is [14] Ishioka T, Shibata Y, Takahashi M, Kanesaka I. Spectrochim Acta Part A 1998;
reasonable in view of the expected increase of DShell StA with T.
54:1811e8.
[15] Kobayashi M, Tadokoro H, Porter RS. J Chem Phys 1980;73:3635.
[16] Chalmers JM, Griffiths PR, editors. Handbook of vibrational spectroscopy, vol.
4. Conclusions 1. Chichester, U.K.: Wiley; 2002. p. 921e32.
[17] Pimentel GC, McClellan L. The hydrogen bond. San Francisco, CA: Freeman and
The interaction between zinc oxide and stearic acid in a medium Co.; 1960.
[18] Chalmers JM, Griffiths PR, editors. Handbook of vibrational spectroscopy, vol.
suitable to simulate a vulcanizing system has been investigated 3. Chichester, U.K.: Wiley; 2002. p. 1919e34.
experimentally using vibrational spectroscopy techniques and [19] Lee JY, Painter PC, Coleman MM. Macromolecules 1988;21:346e54.