Effect of the Aramid Pulp on the Physicochemical, Viscoelastic Properties and Rheokinetics of Polyurethanes

doi:10.21203/rs.3.rs-1831592/v1

This publication highlights the effect of aramid pulp (1 wt.%) on equilibrium swelling, enthalpy reaction, solid viscoelastic properties, and rheokinetics of castor oil (CO) and polyether (PE) polyol blend (50/50 CO/PE) with polymeric isocyanate (pMDI). The CO-pMDI system showed a lower solubility parameter difference (Δδ_i) than the PE-pMDI and CO/PE-pMDI. Crosslink density values ν_e are higher for CO-pMDI, and when added to aramid pulp, there is a slight increase due to the possible restriction of swollen volume. Regarding the pseudo-solid behavior, AP restricts the macromolecular movement and confers reinforcing characteristics to PU systems. The Prout–Tompkins autocatalytic model satisfactorily described the reactions, resulting in activation energy values E_a within 45−54 kJ/mol. However, for the CO/PE blend, the difference between the reactivity of the polyols and the solubility parameter with pMDI results in a two-step reaction kinetics. Aramid pulp practically does not change the reaction mechanism, only the viscosity of the medium. Furthermore, the thermodynamic parameters (ΔS* and ΔH*) indicated that the PU with AP resulted in more ordered complexes. The adequate description of the polymerization reaction and the effect of the aramid pulp allowed obtaining PU formulations with tunable characteristics for strict composite processing composites.

Polyurethane

aramid pulp

polymerization kinetics

rheology

viscosity

Polyurethane (PU) based systems are highly versatile materials used in various applications, varying from linear thermoplastics to thermosetting polymers [1]. PUs are commonly produced by the reaction of petroleum-based molecules containing hydroxyl (-OH) groups with isocyanates (-NCO)[2]. However, from abundant biological sources, vegetable oils can also be used as raw material to produce bio-based PUs, with advantages such as low toxicity, inherent biodegradability, and high purity. In addition, polyols from renewable and petroleum-based sources may be blended to tune specific properties such as reactivity or macromolecular characteristics.

The use of fillers, additives, and reinforcing materials can further expand PU applications [1, 3]. For instance, aramid (poly-p-phenylene-terephthalamide) pulp (AP), a common waste from fabric production facilities, has been used to reinforce PUs and increase fire resistance and flame retardancy, among other advantages, tackling significant issues of this polymer [4, 5].

Nonetheless, a filler or reinforcement into a thermoset PU can affect its curing; therefore, the knowledge of reaction kinetics is essential [6]. For instance, when processing PU composites, it is essential to monitor the viscosity increase during polymerization [7]. This increase in viscosity during the pre-gel stage due to the increase in molecular weight indicates the magnitude of chain extension and branching and provides information on reaction rate and profile before and after gelation [8]. In addition, viscosity monitoring and knowledge of the kinetic parameters of a reactive system allow optimization of the composite processing conditions.

Many methods are employed to the PU polymerization kinetics, broadly classified into two groups: (i) indirect methods, which measure a physical property functionally related to the extent of the reaction, such as differential scanning calorimetry (DSC), rheometry and viscometry; (ii) direct methods, which measure the concentration of a reactant or product species, such as Fourier transform infrared spectroscopy (FTIR), titration, nuclear magnetic resonance (NMR) and gel permeation chromatography (GPC)[9–12]. Nevertheless, the very high viscosity and fast gel formation observed when the urethane conversion reaches a significant level brings analytical problems to direct methods, limiting their use ^[15]. In this context, indirect methods are receiving increasing attention [10].

The rheological measurements approach [10, 13, 14] has produced similar results to DSC to determine kinetics [11, 15]. However, the former allows for lower characteristics of the change in free volume from the macromolecular growth [8]. Furthermore, it can help simulate properties better correlate with actual processing conditions [11]. Rheometry is essential since different rheological parameters, such as viscosity, may increase by as much as 6−7 orders of magnitude during polymerization, affecting processability characteristics of great relevance for many PU applications [13, 15, 16]. In addition, this technique has recently shown great versatility in characterizing the different stages during PU formation and liquid-solid transition[7].

Considering that viscosity varies with the length of the polymer chain, it can provide information about its condition[17]. PU polymerization kinetics have been carried out using isoconversional methods based on calorimetry and rheometry [10, 13, 16]. Activation energy values for the polymerization are in the 45 − 55 kJ/mol range[18, 19], considered an autocatalytic reaction mechanism [12], in which the monomer acts as a catalyst for the reaction, increasing the conversion rate. The most used autocatalytic models in these studies were Prout-Tompkins and Kamal-Sourour [14, 16]. Lucio et al. have extensively studied the rheokinetics of functional PUs, HTPB, and (ferrocenylbutyl) dimethylsilane-grafted HTPB and described the reaction kinetics using empirical models (Arrhenius and Kinua) [13, 15]. They have found a significant contribution of the autocatalytic effect during the formation of the urethane group, which is noticeable for reactive systems at elevated curing temperatures.

Santhos et al [12] evaluated the curing kinetics of a composite propellant based on ammonium perchlorate, hydroxyl-terminated polybutadiene (HTPB), and tolylene diisocyanate system using rheometry under isothermal conditions (60–70°C). They reported that the reaction follows first-order kinetics with an activation energy value of ~ 48.3 kJ/mol in a single autocatalytic step [12]. However, chemical species with active hydrogens can alter the PU polymerization mechanism ^[13]. Recently, Echeverria-Altuna and colleagues used the Castro-Macosko model in pre-gel stages in conjunction with DSC data to simulate the processing of PU composites in RTM. They report that using LiCl improves PU processing as the viscosity increases at a slower rate, which is essential for processing with liquid resins and fiber impregnation [20]. In the liquid molding of composites, for example, latency in the first stage of the polymerization reaction becomes necessary to guarantee low viscosity (∼500 Pa.s) values and obtain good impregnation in reinforcements and mold filling [20, 21]. Therefore, optimizing PU formulations becomes fundamental to improving their applications. Using physicochemical and rheokinetic parameters can be helpful tools for understanding and improving PU formulations.

The main objective was to evaluate the effect of an aramid pulp and the type of polyol (castor oil or polyeter) on the physicochemical, solid viscoelastic, and kinetic characteristics of the reaction. In addition, the monomers' miscibility and reactivity were used to assume that in the system with the polyols blends (50/50), the polymerization reaction occurs practically in parallel. In this way, it is expected to better understand the effect of fibers and particles during the polymerization reaction of PUs to predict future applications and optimize processing conditions.

2.1. MATERIALS

The CO polyol, hydroxyl number 160 mg.KOH/g, molecular weight 933 g/mol, functionality 2.7, viscosity at 25 ºC 640 mPa.s, and specific gravity 0.961 g/cm³, was supplied by Alpha Quimica Ltda, Brazil. The CO polyol was dried at 100 ºC under reduced pressure (700 mm Hg) for 12 h before use. A polyether (PE) polyol was also used, being a polypropylene oxide (PPO) with hydroxyl number 110 mg.KOH/g, molecular weight 1000 g/mol, functionality 2, viscosity at 25 ºC 175 mPa.s and specific gravity 1.000 g/cm³, supplied by Mantoflex Indústria de Plásticos Ltda, Brazil.

The polymeric isocyanate 4.4’ - diphenylmethane-isocyanate (pMDI), with M_n 340 g/mol, CAS number 101-68-8, functionality 2.7, and free isocyanate index 31.40 ± 0.02 wt.% determined using n-dibutylamine titration according to ASTM D2572-10, was supplied by BASF Brazil and used as received. The aramid pulp (AP) was donated by Dupont® and, according to the manufacturer, has fiber length of 0.5–1 mm, bulk density of 48–160 kg/m³, and a specific surface area of 7–11 m²/g.

2.2. PU AND COMPOSITE PREPARATION

In this work, three polyols were used in the formulations, one from castor oil (named CO), one from polyether (named PE), and their 50/50 wt.%/wt.% blend (named CO/PE). In addition, some PU formulations were incorporated with 1 wt.% aramid pulp by weight, as detailed in Table 1.

The content of aramid pulp was defined based on preliminary experiments with 0.5, 1.0, and 1.5 wt.%, evaluating the dispersion using optical microscopy, according to literature recommendations ^[1]. The dispersion procedure employed by Pearson was used, which consisted in adding the AP amount to the polyol and steering it using a Dispermat N1 high shear mixer direct disperser^[1]. The initial speed was 500 rpm, for 3 min., to break the agglomerates, then it was gradually increased, 500 rpm every 3 min., until 2500 rpm, where it remained until complete dispersion of the pulp in the polyol ^[1]. The procedure to prepare the composites is illustrated in Fig. 1. The same method, but without AP, was employed to obtain the PU samples.

2.3. CHARACTERIZATION

The PU reaction enthalpy (ΔH_p) was monitored via DSC (TA instruments Q20) using ∼5 mg sample, N₂ atmosphere (50 ml/min), and heating from 25 to 250 ºC at 10 ºC/min for all samples. The swelling degree was evaluated as a function of time using 5 g of the sample immersed in 200 mL of cyclohexane (based on the solubility parameter) in a container kept at 25 ºC. At pre-determined time intervals, from 1 h until equilibrium (∼480 h), the sample was removed from the solvent, the mass was measured, and the sample returned to the solvent. The swelling degree (Q) was obtained according to Eq. 1 ^[19].

\(Q\left(wt. \%\right)=\frac{{m}_{t}-{m}_{o}}{{m}_{o}}.100\)

(1)

where m_t is the mass of the swollen polymer at a particular time; and m_o is the mass of the polymer before swelling.

The crosslink density \({v}_{e }\)(mol/cm³) was calculated based on the degree of swelling at the equilibrium of the samples in cyclohexane, using Eqs. 2–3 as proposed by Flory-Rehner ^[21] described in ASTM D 6814.

\({v}_{e }=\frac{-[lnln \left(1-{V}_{r}\right) + {V}_{r}+\chi {V}_{r}^{2}]}{\left[{V}_{1}({V}_{r}^{\frac{1}{3}}-{V}_{r})/2\right]}\)	(2)
\({V}_{r}=\frac{{V}_{As}}{{V}_{As}+ {V}_{Sae}}\)		(3)

where \({V}_{As}\) is the volume of the dry sample; and \({V}_{Sae}\) is the volume of absorbed solvent at equilibrium. Kraus' correction was used to calculate the true crosslink density considering the presence of the aramid pulp for the composites [18, 19].

Miscibility of the monomers was evaluated with theoretical solubility parameters based on the Hansen approach using the group contributions method [22]. In addition, Bagley diagrams were built to aid the analysis. In these, the effect of the dispersion (δ_d) and polar (δ_p) components of the solubility parameter was assumed similar, being grouped into δ_v, according to Eq. 4:

\({\delta }_{v}^{2}= {\delta }_{d}^{2}+ {\delta }_{p}^{2}\)

(4)

In contrast, the effects of the hydrogen bonding component (δh) are very different [23]. Therefore, in Bagley's approach, the solubility region can be expressed using a δv × δh plot [23], where the distance between two points is calculated using Eq. 5.

\({D}_{12}=\sqrt{{({\delta }_{1,v}-{\delta }_{2,v})}^{2}+{({\delta }_{1,h}-{\delta }_{2,h})}^{2}}\)

(5)

The solid linear viscoelastic behavior of the materials was investigated by DMA (TA Instruments, DMA 850), using the dual cantilever geometry. The experiments were carried out in the temperature range of -60°C to 100°C, with oscillation amplitude of 10 µm (linear viscoelastic regime), frequency of 1 Hz, and heating rate of 3°C/min. Samples with 60.0 mm × 13.0 mm × 3.0 mm were used.

The rheokinetics behavior was carried out by dynamic oscillation employing an Anton Paar MCR 101. The plate diameter and gap were 25.0 and 0.50 mm, respectively. Measurements were performed in 5 isothermal curves (in the 40–80 ºC range) with 1 Hz frequency and 2 Pa (linear viscoelastic regime). The viscoelastic response during the reaction, in terms of complex viscosity (η*) and shear storage and loss moduli (G’ and G”, respectively), were registered as a function of the reaction time.

In Eq. 6, viscosity is treated as the combination of two exponential expressions, one for the dependence of viscosity and another for the effect of reaction kinetics [6, 24]:

\(\eta \left(T,\alpha \right)=K{e}^{\frac{{E}_{\eta }}{RT}}.A{e}^{\frac{{E}_{\alpha }}{RT}}f\left(\alpha \right)\)

(6)

where K is a constant; \({E}_{\eta }\) is the flux activation energy; R is the universal gas constant; T is the absolute temperature; E_α is the activation energy; and f(α) is the function that describes the reaction mechanism.

In reactive systems, two effects can be superimposed in rheometric measurements, the dependence of viscosity with temperature and its increase in the degree of reaction [25, 26]. These were determined based on the Pahl-Hesekamp approach [24, 27] in which it is possible to separate the viscosity and kinetics therms. The viscosity of the reactive system is described with a WLF equation, in which the actual temperature is reduced by the difference between the actual glass transition temperature, T_g (x), and the glass transition temperature for the degree of reaction x = 0 (Eq. 7).

\(ln\frac{{\eta }^{*}(T,x)}{{\eta }^{*}({T}_{ref},x=0)}=-\frac{{C}_{1}\left(T-{T}_{ref}\right)}{{C}_{2}+T-{T}_{ref}}-\frac{{B}_{1}[T-{T}_{g}\left(x\right)]}{{B}_{2}+T-{T}_{g}\left(x\right)}+\frac{{B}_{1}[T-{T}_{g}\left(x=0\right)]}{{B}_{2}+T-{T}_{g}\left(x=0\right)} \left(7\right)\)

where C₁ and C₂ are WLF constants of the non-crosslinked polymer, and B₁ and B₂ are constants of the crosslinked polymer. The model considers that the formation of crosslink decreases the intrinsic volume because of the reduced ability of the chains to move.

Opfermann described the temperature behavior using an Arrhenius-type equation using two different activation energies, E_0, and E₁, referring to the beginning and the end of the reaction, that is, the lowest and highest viscosities, respectively[24, 28].

\(ln{ղ}^{*}\left(T, x\right) =ln{ղ}^{*}( {T}_{r} , x=0)+\frac{{E}_{0}}{R}.\left(\frac{1}{T}-\frac{1}{{T}_{r}}\right)\left(1-G\left[T\left(t\right)\right]\right)\)+

\(+ \frac{{E}_{1}}{R}*\left(\frac{1}{T}-\frac{1}{{T}_{r}}\right)*G[T\left(t\right)+ \varDelta l{ղ}^{*}G[T\left(t\right)\)]

(8)

where G[T(t)] is a gain function that measures the change in viscosity, described by Eq. 9:

\(G\left[T\left(t\right)\right]=1-{e}^{-{x}^{({A}_{0}+{A}_{1}*x + {A}_{2}*{x}^{2}+ {A}_{3}*{x}^{3})}}\)

(9)

Dynamic and/or isothermal measurements allow calculation of the A₀, A₁, A_2, and A₃ parameters by fitting the experimental data. Thus, it is possible to predict the viscosity as a function of time at various temperatures based on the viscosity increase along with the polymerization reaction [25, 28]. The data was computed using “Netzsch Thermokinetics: A software modulus for the kinetic analysis of thermal measurements” by non-linear least-squares fitting. The term f(α) was determined by comparing the experimental data with the theoretical mathematical models based on fitting sixteen kinetic models, shown here as supplementary material. The models that best describe the reaction were identified based on statistical tests (f-test)[11, 29].

The kinetic and thermodynamic parameters of the polymerization process were computed using Wynne-Jones–Eyring–Evans theory [30, 31] which relates the temperature-dependent pre-exponential factor with the kinetic constant resulting in:

\(k=\frac{{k}_{B}{T}^{n}}{h}{e}^{\left[N+\frac{\varDelta {S}^{*}}{R}\right]}{e}^{\left[\frac{-{E}_{a}}{RT}\right]}\)

(10)

where k_B and h are the Boltzmann and Planck constants, N is a parameter related to molecularity. \(\varDelta {S}^{*}\) is the activation entropy and , is a function of activation enthalpy. Considering the classical Arrhenius constants have N = 0, whereas when the reaction is in a liquid medium N assumes a value of 1[30] Considering the PU polymerization occurs in the liquid state, N=, the Eq. (10) can be solved using a linear relationship by plotting ln (k/T) vs. 1/T, \({\varDelta H}^{*}\)and \(\varDelta {S}^{*}\) are obtained from the slope and linear coefficients. The values of k can be obtained by using the Arrhenius equation (\(k=A{e}^{-\frac{{E}_{a}}{RT}})\).

Miscibility of the polyol blends was computed using the solubility parameters estimated based on group contribution theories [32]. Table 2 shows the solubility parameters of the monomers, that is, the dispersive (\({\delta }_{d}\)), polar (\({\delta }_{p}\)), hydrogen bonding (\({\delta }_{h}\)), volumetric (\({\delta }_{v}\)) and total \(\left({\delta }_{T}\right)\) components. These results were relatively close to those of similar studies ^[25]. Note that CO has a more significant similarity to pMDI regarding dispersive and total components than PE; therefore, less phase separation is expected for the former.

A simple analysis can be done by building phase diagrams using the Flory-Huggins mixing theory, according to the work of Rusette et al., shown here as supplementary material [33]. The ΔG_m values at 25 ºC for the CO-pMDI pair (0.177 J) are lower than for PE-pMDI (0.870 J) (for mixtures with equal weight fraction). When evaluating the effect of temperature, an increase in miscibility is observed for the CO-pMDI system based on the ΔG_m < 0 values. Increased miscibility between segments reduces phase segregation between rigid and flexible segments, and therefore a single glass transition is observed [2, 34].

The Bagley graph is shown in Fig. 2. The van der Waals forces the miscibility is delimited by a relatively well-defined circular miscibility region, with an “interacting radius” of about 2.5 MPa^1/2, being expanded to the regions 5.5 and 7.0 MPa^1/2 for systems with hydrogen bonding. This figure shows that the polyols used in this study (CO and PE) have a low tendency to phase separation between them, at a distance of 1.87 MPa^1/2. In addition, as observed for ΔG_m, CO is closer to pMDI than PE.

Figure 3 shows the DSC curves of PU and PU composite samples. PU(CO) has greater reactivity than PE, and the reaction starts at lower temperatures. The PE reactivity is lower due to many secondary hydroxyls due to this polyol's type of synthesis reaction [35]. Also, as the polyether polyol is added to the formulations, the curing peak shifts to higher temperatures, and the reaction takes longer to finish.

The onset, endset, peak temperatures, and reaction enthalpy for PU and PU composites were obtained from the DSC curves and compiled in Table 3. The incorporation of 1 wt.% of aramid pulp shows little effect on all samples' enthalpy of reaction ΔH_p, with slight differences attributed to experimental deviations. It is important to remember that the aramid pulp has terminal groups (−NH₂ and −COOH) that could react with −NCO, but these side reactions (allophanates, urea biuret, etc.) did not contribute to a significant increase in ΔH_p [36, 37].

Concerning reactivity, primary hydroxyls are much more reactive than secondary ones [38]. Ismail et al. studied the reaction kinetics of palm oil PUs and showed that it is reduced by about three times when comparing MDI reactions with secondary and primary hydroxyls ^[41]. Polyester-type polyols react much faster than polyethers due to the observed difference in reactivity. Furthermore, polyesters such as castor oil can stabilize the hydrogen bonds of the produced intermediates and thus react more quickly [39].

Figure 4 shows the swelling curves of the samples in cyclohexane until equilibrium, after approximately 20 days. After this period, the PU(CO), PU(CO/PE), and PU(PE) samples showed 18.1, 20.4%, and 27.9 wt.% swellings, respectively. Samples with PE present more significant swelling than those with CO, attributed to the higher molecular weight between crosslink points of PE compared to CO, affecting the swelling reticulum volume and the polymer-solvent interaction [2].

The swelling curves of PUs and PU composites have a similar shape, but the latter reached lower swelling values than the respective polymers, being 16.1, 17.5, and 25.0 wt.% for PU(CO) + AP, PU(CO/PE) + AP, and PU(PE) + AP samples, respectively. The lower swelling may be related to the reduction in the swollen volume after AP addition. Indeed, the AP swells very little compared to the PU matrix and could reduce the overall swelling. Also, the -NCO group can react with groups containing reactive hydrogen, yielding different products [29].

It is known that more crosslink points, i.e., higher crosslink density, reduce the polymeric network's mobility and hinder solvent permeation [40, 41]. In Fig. 5, the calculated crosslink density is consistent with the previous swelling results. Also, PE tends to produce a structure of higher molecular weight between crosslink points than CO.

The V_e values for the composites, also shown in Fig. 5, considered the Kraus correction to account for the fiber incorporation. A slight increase in V_e values is noticed compared to the respective PU samples, which may be related to the swelling restraint that the pulp exerts on the PU [18].

Figure 6 shows the storage module (E´) and Tan delta vs. temperature (-50 to 100ºC) for PU and composite samples with AP. The moduli values at -50, 100ºC, T_g from Tan delta peak, full width at half maximum (FWHM) in T_g region, and E’’ peak are summarized in Table 4. All samples showed elastic pseudo-solid behavior in the investigated temperature range. The AP increased the rigidity of all samples, mainly above the T_g. The displacement of the modulus values in the glass transition region is related to the restriction to the movement that the fibers result.

For all samples, only one glass transition temperature was observed. For PU (CO) a Tg of ∼28ºC was observed [42, 43], while for the composite with AP a small change to ∼32ºC was noted. However, a broadening in the transition peak (FWHM) was noticed, which is related to the heterogeneity of the T_g of the composites.

PU (PE) showed a T_g at -7.2 ºC, while PU(PE) + AP changed to 12.7 ºC. Two glass transitions were not observed concerning PU(CO/PE). According to the solubility and thermodynamic parameters, there may be immiscibility between the polyols. However, the DMA investigation scale was not evidenced. This shows that phase separation between flexible phases can occur on a much smaller scale. Thus, as for PU (CO), PU (PE), and PU(CO/PE) there is an increase in FWHM values with the addition of AP, which is a reflection of the heterogeneity of the system.

Furthermore, it is noted that the values of Tan delta increase in magnitude in the AP samples, which is a reflection of the higher elastic component (E') compared to the viscous (E''), which reflects lower energy dissipation capacity per sinusoidal deformation cycle. On the other hand, the lower T_g values observed for PU(PE) reflect, the higher molecular weight between crosslink points and the oxygen in the polyol main backbone chain, which results in greater mobility [2].

Note that the modulus values in the elastic plateau region (E`_{r 100ºC}) are higher for all composite systems with AP. It is also noted that E`_r100ºC presents a practically linear behavior concerning pure polyols and the 50/50 blend. Based on the theory of rubber elasticity [42] (\({E}^{\text{'}}=3{\nu }_{e}RT)\) it is noted that a similar trend to the swelling results, that is, as PE is used, there is a greater distance between the crosslink points. Still, movement restriction caused by the AP increases the modulus values as well as ve according to the swelling data.

According to Flory (1953)[44] any cross-linkable polymer system has a sharply defined gel point at a particular critical conversion (extent) of the p_c reaction, which must be independent of temperature, type, and amount of catalyst, among others [45] When polymers gel, a sudden and dramatic increase in viscosity is observed until the molecular weight becomes infinite (degree of polymerization, DP→∞). Thus, the second term of the Carother's equation (\(DP=2/(2-p{f}_{m})\)disappears, and the critical extent of the reaction becomes \({p}_{c}=2/{f}_{m}\). This occurs at a lower reaction degree (0.75) for the castor oil system than for the polyether (0.85). The gel time at each temperature was estimated according to ASTM D7750 as the point at which the storage modulus abruptly increased, shown as Supplementary Material (Table S2). As expected, as the temperature increases, the time for crossover (G’=G’’) to occur decreases, suggesting that molecular mobility increases in the early stages at higher temperatures, allowing the reaction to proceed at a faster rate. This time reduction in relation to temperature can be described with an Arrhenian relation (Fig. S2). Note that there is a reduction in gel time and lower activation energy values (E_gel: 30−52 kJ.mol⁻¹) for PUs with pulp. Because it is highly fibrillated, even with a low degree of reaction in pulp systems, the material formed already reaches a pseudo-solid state long before pure PU. The values found are in agreement with the literature [12].

Viscosity vs. time curves for the PU and PU composites at 40, 50, 60, 70, and 80°C isotherms are shown in Fig. 7(A-F). The given viscosity data is required up to the gel time for each temperature since above that time. The red solid reference lines represent the model proposed. The Prout-Tompkins model (\(f\left(\alpha \right)={\alpha }^{m}(1-\alpha {)}^{n}\) Bna mechanism) was found the most adequate to describe the polymerization mechanism, with all determination coefficients higher than 0.970. Initially, 16 mechanisms were tested (supplementary files). However, in this reaction, what he wrote best was based on the Bna model. This model is an expression derived from the model of Šesták–Berggren [46].

The increase in temperature Fig. 7 (A-F) contributes to the increase in reaction rate, decreasing the time to reach the same viscosity value. In addition, according to several studies [47, 48]the aramid pulp is highly fibrillated, with a large surface area and free amine groups, increasing the viscosity of the reaction.

Viscosity at the start and end of the reaction is lower for polyether than for castor oil PU. Still, the increase in the rate of viscosity is less pronounced due to the lower reactivity of the PE polyol [49]. Concerning the polyol blend, an intermediate behavior is noted, with two regions of viscosity increase related to the reaction rate of each polyol, as indicated in the DSC results. Since both polyols have close solubility parameters, there is poor phase separation and apparent homogeneity. The difference in reaction kinetics can induce phase separation, which would yield the two observed reaction stages in Fig. 7(C-D) [2, 22]. Therefore, these observations considered a parallel reaction approach for the CO/PE blend.

The values of the kinetic parameters of the six samples of the best model (B_na mechanism) are presented in Table 5. The single-step (when only one polyol is used) and autocatalytic mechanism have been used in several curing kinetics studies of PU systems, with or without reinforcement/filler [13, 50]. Regarding the pre-exponential factor (log A₁), one can notice a decrease with aramid pulp incorporation. Considering that this parameter reflects the frequency of collisions between groups, decreased reactivity of the system occurs after AP addition. This term also contains information about mobility reduction due to the increase in system viscosity from the pulp addition [51]. Although the fibers have reactive -NH₂ and -COOH terminal groups, their importance is reduced since the fiber has limited mobility. It can also be noted that the pre-exponential factor values are lower for PU(PE) and PU(PE) + AP, justified by the lower PPO reactivity due to the synthesis method used in the production of this polyol type [35].

The values of Eα increase in the systems with AP which results in the increase of the energy barrier for the formation of activated species. The increase in Eα is related to the increase in system viscosity in the pre-gel period. The higher initial viscosity causes the system to pass from the pseudo-liquid state to a pseudo-solid in shorter reaction times, as observed in the gel time values. Although it is evident that the pulp practically does not participate in the polymerization reaction, the effect of increasing viscosity significantly alters the viscous response of the system. Regarding the polyol blend, it is noted that the two activation energy values found are of the same order of magnitude in both steps. However, in PE, the reaction occurs more slowly due to its structural characteristic. The Eα values are consistent with those reported in the literature for PU systems (45 − 60 kJ/mol) [13, 15, 52].

In general, the parameters of the reaction order “m” and “n” are only slightly different, indicating that the same reaction mechanisms prevail. In general, values of m lower than 1 indicate that an autocatalytic reaction greatly contributes to the overall reaction rate. These values also agree with previous studies on PU curing kinetics [13, 20, 41]. Os valores de n encontrados variam de 1.8−1.98 e m de 0.10−0.46. Taking the works of Fernandez d’Arlas et al (2007)[52] and Lucio and de la Fuente (2021)[10] as a comparison, the parameters estimated by the methodology used are not so far from the literature. Both authors used the Kamal–Sourour autocatalytic model and obtained m and n parameters of the same magnitude. In the case of the work by Lucio and de la Fuente, the main difference between the models applied in this work is in the form of obtaining the degree of conversion. Thus, it is evident that the method applied in this work can produce very similar results concerning other methodologies for obtaining kinetic data using viscosity.

From the rheological kinetic model, it is possible to predict the behavior of viscosity as a function of time or temperature. Figures S3 (supplementary files) show the studied systems' time-temperature–transformation diagram curves. These diagrams are essential because they can be used directly to predict the behavior of the polymerization reaction. These diagrams constructed curves to predict viscosity, as shown in Figure S4. In these curves, the viscosity at 25ºC and the conversion curve are estimated as a function of what was obtained from the TTT diagrams generated by the rheokinetic model. As can be seen, in systems with AP, the viscosity starts at higher values but has slight variation concerning the reaction conversion, as also shown in the DSC measurements. However, the increase in viscosity caused by AP can often hinder processes that demand a fast flow of resin inside a mold or cavity.

The thermodynamic reaction parameters, such as enthalpy and entropy change, have been calculated using the Wynne-Jones–Eyring–Evans theory and summarized in Table 6. These parameters help understand the reaction mechanism based on the Prout-Tompkins model of urethane reaction. The activation entropy variation (ΔS*) indicates the degree of randomness of the constituent throughout the reaction, while ΔH* indicates the nature of the reaction heat in activate complex [14, 53].

The negative values for activation entropies (ΔS*) ratify the importance of the association of reagents prior to the polymerization, thus ratifying that the assumed model shows coherence for the PU systems. Also, the reaction process resulted in positive ΔH* values regardless of the temperature, suggesting an endothermic process for forming a complex. Fernandez d’Arlas et al.(2007) determined the thermodynamic parameters for catalyzed and uncatalyzed PU (1,6-hexamethylene diisocyanate/poly(carbonate-co-ester)diol) polymerization reactions from DSC analyses. They used the Wynne-Jones–Eyring–Evans approach equation and reported ΔS* of -173.5 J/K.mol and ΔH* of 40.2 kJ/mol for uncatalyzed reactions [52].

Taking into account that ΔS* reflects the variation between a final and an initial state and that PU(CO), PU(CO/PE), and PU(PE) are in similar initial states (i.e. when molecules are very separated one to each other ) a lower value for the entropy of the activation state would suggest a final state of the PU(CO/PE) formation reaction is more ordered and thermodynamically unfavorable. The addition of AP increased the values of ΔS* in all compositions, while ΔH* was noted for a reduction when comparing the systems with AP with castor oil and polyether polyol. The greater heterogeneity of the reaction medium caused by the presence of fiber seems to cause the ΔS* values in relation to the respective pure resins.

Due to the nature of the complexes formed, the polymerization reaction of PUs with AP becomes more complex and tends to be less favorable in thermodynamic terms. Therefore, an optimal balance of viscosity increase and reactivity must be devised when PU resins are used in different processes, such as liquid composite molding, where viscosity and gel time are crucial parameters.

The effect of aramid pulp addition on the physicochemical properties and rheokinetics of polyurethanes based on CO, PE, and a CO/PE blend as a polyol and isocyanate (pMDI) was investigated. Using solubility parameters (δ) together with ΔG_m contributed to elucidating the miscibility behavior of polyols. Furthermore, the number of steps in the PU polymerization reaction can be determined with the help of physicochemical parameters and the reactivity of the polyols. AP restricts the macromolecular movement of the solid material and confers reinforcing characteristics to PU systems. The Prout-Tompkins autocatalytic model, with one reaction stage for PU(CO) or PU(PE) and two stages for the polyol blend, best described their polymerization reaction. The aramid pulp did not affect the reaction's enthalpy, but the consequent increase in viscosity reduced the reaction rate and increased the activation energy. Besides, the reaction complexes formed in a blend CO/PE tend to be more ordered than those in PU(CO) and PU (PE). Furthermore, the greater heterogeneity of the reaction medium caused by AP makes the reaction thermodynamically unfavorable. Finally, the approach described in this work allows direct parameters for the liquid molding of composites in general since viscosity and gel time are crucial processing parameters.

Competing Interests: The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

ACKNOWLEDGMENTS

The authors acknowledge Brazilian Agency Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES, Brazil) - (Finance Code 001) and Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq, Brazil) (grants 305814/2021-4).

1. Pearson A, Naguib HE (2017) Novel polyurethane elastomeric composites reinforced with alumina, aramid, and poly(p-phenylene-2,6-benzobisoxazole) short fibers, development and characterization of the thermal and dynamic mechanical properties. Compos Part B Eng 122:192–201. https://doi.org/10.1016/j.compositesb.2017.04.017

2. Mihail L (2005) Chemistry and Technology of Polyols for Polyurethanes, 1st ed. Rapra Technology, Shropshire.

3. Zhang Y, Seveyrat L, Lebrun L (2021) Correlation between dielectric, mechanical properties and electromechanical performance of functionalized graphene / polyurethane nanocomposites. Compos Sci Technol 211.: https://doi.org/https://doi.org/10.1016/j.compscitech.2021.108843

4. Kerche EF, da Silva VD, da Silveira Jankee G, et al (2021) Aramid pulp treated with imidazolium ionic liquids as a filler in rigid polyurethane bio-foams. J Appl Polym Sci 138.: https://doi.org/10.1002/app.50492

5. Ertekin M (2017) Aramid fibers. Elsevier, Amsterdam.

6. Lipatov YS, Kosyanchuk L V., Nesterov AE, Antonenko OI (2003) Filler effect on polymerization kinetics and phase separation in polymer blends formed in situ. Polym Int 52:664–669. https://doi.org/10.1002/pi.937

7. Macosko CW (1985) Rheological Changes During Crosslinking. Br Polym J 17:239–245. https://doi.org/10.1002/pi.4980170228

8. Lipshitz SD, Macosko CW (1976) Rheological Changes During a Urethane Network Polymerization. Polym Eng Sci 16.: https://doi.org/https://doi.org/10.1002/pen.760161205

9. Juan Li, Shengling Jiang, Liang Ding, Lingfang Wang (2021) Reaction kinetics and properties of MDI base poly (urethane-isocyanurate) network polymers. Designed Monomers and Polymers 24:265–273 https://doi.org/DOI:10.180/15685551.2021.1971858.

10. Lucio B, de la Fuente JL (2021) Chemorheology and Kinetics of High-Performance Polyurethane Binders Based on HMDI. Macromol Mater Eng 306.: https://doi.org/10.1002/mame.202000617

11. Agnol LD, Dias FTG, Nicoletti NF, et al (2019) Polyurethane tissue adhesives for annulus fibrosus repair: Mechanical restoration and cytotoxicity. J Biomater Appl 34:673–686. https://doi.org/10.1177/0885328219864901

12. Santhosh G, Reshmi S, Reghunadhan Nair CP (2020) Rheokinetic characterization of polyurethane formation in a highly filled composite solid propellant. J Therm Anal Calorim 140:213–223. https://doi.org/10.1007/s10973-019-08793-6

13. Lucio B, De La Fuente JL (2016) Kinetic and thermodynamic analysis of the polymerization of polyurethanes by a rheological method. Thermochim Acta 625:28–35. https://doi.org/10.1016/j.tca.2015.12.012

14. Dall Agnol L, Ornaghi HL, Monticeli F, et al (2021) Polyurethanes synthetized with polyols of distinct molar masses: Use of the artificial neural network for prediction of degree of polymerization. Polym Eng Sci 61:1810–1818. https://doi.org/10.1002/pen.25702

15. Lucio B, de la Fuente JL (2018) Kinetic and chemorheological modelling of the polymerization of 2,4- Toluenediisocyanate and ferrocene-functionalized hydroxyl-terminated polybutadiene. Polymer (Guildf) 140:290–303. https://doi.org/10.1016/j.polymer.2018.02.058

16. Lucio B, De La Fuente JL (2014) Rheological cure characterization of an advanced functional polyurethane. Thermochim Acta 596:6–13. https://doi.org/10.1016/j.tca.2014.09.012

17. Pacheco MFM, Bianchi O, Fiorio R, et al (2009) Thermal, chemical, and morphological characterization of microcellular polyurethane elastomers. J Elastomers Plast 41:323–338. https://doi.org/10.1177/0095244309095016

18. Kraus G (1963) Swelling of Filler-Reinforced. J Appl Polym Sci 7:861–871. https://doi.org/doi.org/10.1002/app.1963.070070306

19. Valentini L, Bittolo Bon S, Hernández M, et al (2018) Nitrile butadiene rubber composites reinforced with reduced graphene oxide and carbon nanotubes show superior mechanical, electrical and icephobic properties. Compos Sci Technol 166:109–114. https://doi.org/10.1016/j.compscitech.2018.01.050

20. Echeverria-Altuna O, Ollo O, Calvo-Correas T, et al (2022) Effect of the catalyst system on the reactivity of a polyurethane resin system for RTM manufacturing of structural composites. Express Polym Lett 16:234–247. https://doi.org/10.3144/expresspolymlett.2022.19

21. Younes, Usama E (2018) Polyurethane Composites Produced by a Vacuum Infusion process U.S. Patent No. 9,580,598. 28 Feb. 2017.

22. Hansen CM (2007) Hansen solubility parameters: A user’s handbook: Second edition. London.

23. Meaurio E, Sanchez-Rexach E, Zuza E, et al (2017) Predicting miscibility in polymer blends using the Bagley plot: Blends with poly(ethylene oxide). Polymer (Guildf) 113:295–309. https://doi.org/10.1016/j.polymer.2017.01.041

24. Hesekamp D, Pahl MH (1996) Curing effects on viscosity of reactive epoxy resin adhesives. Rheol Acta 35:321–328. https://doi.org/10.1007/BF00403532

25. Opfermann J (2000) Kinetic analysis using multivariate non-linear regression. I. Basic concepts. J Therm Anal Calorim 60:641–658. https://doi.org/10.1023/A:1010167626551

26. Flammersheim HJ, Opfermann J (2001) Investigation of Epoxide Curing Reactions by Differential Scanning Calorimetry – Formal Kinetic Evaluation Macromolecular materials and engineering 286: 143-150. https://doi.org/10.1002/1439-2054(20010301)286:3<143::AID-MAME143>3.0.CO;2-Z

27. Hesekamp D, Broecker HC, Pahl MH (1998) Chemo-Rheology of Cross-Linking Polymers. Chem Eng Technol 21:149–153. https://doi.org/10.1002/(SICI)1521-4125(199802)21:2<149::AID-CEAT149>3.0.CO;2-P

28. Flammersheim HJ, Opfermann J (1999) Formal kinetic evaluation of reactions with partial diffusion control. Thermochim Acta 337:141–148. https://doi.org/10.1016/S0040-6031(99)00162-8

29. Ourique PA, Ornaghi FG, Ornaghi HL, et al (2019) Thermo-oxidative degradation kinetics of renewable hybrid polyurethane–urea obtained from air-oxidized soybean oil. J Therm Anal Calorim 137:1969–1979. https://doi.org/10.1007/s10973-019-08089-9

30. Eyring H (1935) The activated complex in chemical reactions. J Chem Phys 3:63–71. https://doi.org/10.1063/1.1749604

31. Avery HE (1974) Basic reaction kinetics and mechanism. Macmillan Education, London.

32. Boudouris D, Constantinou L, Panayiotou C (1997) A Group Contribution Estimation of the Thermodynamic Properties of Polymers† Ind. Eng. Chem. Res. 1997, 36, 9, 3968–3973 https://doi.org/10.1021/ie970242g

33. Ruzette A-VG, Mayes AM (2001) A Simple Free Energy Model for Weakly Interacting Polymer Blends. Macromolecules, 34, 6, 1894–1907 https://doi.org/10.1021/ma000712

34. Van Krevelen DW, Te Nijenhuis K (2009) Properties of Polymers - Their correlation with chemical structure; their numerical estimation and prediction from additive group contributions. 4th Edition, Elsevier, Amsterdam.

35. Miyajima T, Nishiyama K, Satake M, Tsuji T (2015) Synthesis and process development of polyether polyol with high primary hydroxyl content using a new propoxylation catalyst. Polym. J. 47:771–778. https://doi.org/10.1038/pj.2015.64

36. Denchev Z, Dencheva N (2012) Manufacturing and Properties of Aramid Reinforced Composites, First Edit. Carl Hanser Verlag GmbH & Co. KG, München.

37. Chattopadhyay DK, Raju KVSN (2007) Structural engineering of polyurethane coatings for high performance applications. Prog Polym Sci 32:352–418. https://doi.org/10.1016/j.progpolymsci.2006.05.003

38. Ismail TNMT, Palam KDP, Bakar ZBA, et al (2016) Urethane-forming reaction kinetics and catalysis of model palm olein polyols: Quantified impact of primary and secondary hydroxyls. J Appl Polym Sci 133.: https://doi.org/10.1002/app.42955

39. Caraculacu AA, Coseri S (2001) Isocyanates in polyaddition processes. Structure and reaction mechanisms. Prog Polym Sci 26:799–851. https://doi.org/10.1016/S0079-6700(00)00033-2

40. Formela K, Hejna A, Piszczyk Ł, et al (2016) Processing and structure–property relationships of natural rubber/wheat bran biocomposites. Cellulose 23:3157–3175. https://doi.org/10.1007/s10570-016-1020-0

41. Olejnik A, Gosz K, Piszczyk Ł (2020) Kinetics of cross-linking processes of fast-curing polyurethane system. Thermochim Acta 683:178435. https://doi.org/10.1016/j.tca.2019.178435

42. Wang C, Dai L, Yang Z, et al (2018) Reinforcement of castor oil-based polyurethane with surface modification of attapulgite. Polymers (Basel) 10:1–11. https://doi.org/10.3390/polym10111236

43. Ionescu M, Radojčić D, Wan X, et al (2016) Highly functional polyols from castor oil for rigid polyurethanes. Eur Polym J 84:736–749. https://doi.org/10.1016/j.eurpolymj.2016.06.006

44. Flory PJ (1953) Principles of Polymer Chemistry, 1^o ed. Cornell University press, New York

45. Flory PJ, Rehner J (2004) Statistical Mechanics of Cross-Linked Polymer Networks I . Rubberlike Elasticity. J Chem Phys 11:512–520

46. Šesták J (2017) Šesták–Berggren equation: now questioned but formerly celebrated—what is right: Commentary on the Burnham paper on logistic equations in kinetics. J Therm Anal Calorim 127:1117–1123. https://doi.org/10.1007/s10973-015-4998-x

47. Okhawilai M, Parnklang T, Mora P, et al (2019) The energy absorption enhancement in aramid fiber-reinforced poly(benzoxazine-co-urethane) composite armors under ballistic impacts. J Reinf Plast Compos 38:133–146. https://doi.org/10.1177/0731684418808894

48. Elaheh K, Vahid H-A, Farshad BA, Jafar H (2019) Aramid fibers composites to innovative sustainable materials for biomedical applications. In: Materials for Biomedical Engineering. Tehran - Iran, pp 173–204

49. Szycher M (1999) Handbook of Polyurethane. CRC press, New York

50. Shen CH, Springer GS (1976) Moisture Absorption and Desorption of Composite Materials. J Compos Mater 10:2–20. https://doi.org/10.1177/002199837601000101

51. Quintanilha RC, Orth ES, Grein-Iankovski A, et al (2014) The use of gum Arabic as “Green” stabilizer of poly(aniline) nanocomposites: A comprehensive study of spectroscopic, morphological and electrochemical properties. J Colloid Interface Sci 434:18–27. https://doi.org/10.1016/j.jcis.2014.08.006

52. Fernandez d’Arlas B, Rueda L, Stefani PM, et al (2007) Kinetic and thermodynamic studies of the formation of a polyurethane based on 1,6-hexamethylene diisocyanate and poly(carbonate-co-ester)diol. Thermochim Acta 459:94–103. https://doi.org/10.1016/j.tca.2007.03.021

53. Bazzo A, Dias SLP, Vaghetti JCP, et al (2021) Caesalpinia ferrea: a potential feedstock for biochar production. Biomass Convers Biorefinery. https://doi.org/10.1007/s13399-021-02068-7

Table 1: Formulation of PU and PU composite samples.

Nomenclature	Isocyanate (NCO)^a	Polyol (CO/PE)^a	AP
PU(CO)	1.05	1.00 / 0.00	---
PU(PE)		0.00 / 1.00
PU(CO/PE)		0.50 / 0.50
PU(CO)+AP	1.05	1.00 / 0.00	1.00 wt.%
PU(PE)+AP		0.00 / 1.00
PU(CO/PE)+AP		0.50 / 0.50

^aMolar ratio.

Table 2: Solubility parameters (MPa^½) of monomers used to synthesize PU formulations calculated according to the group contribution method.

Table 3: Onset, endset, peak temperatures (T_max) and reaction enthalpy (ΔH_p) for PU and PU composites.

Sample	*T_onset* (^oC)	*T_endset* (^oC)	*T_max* (^oC)	Δ*H_p* *(J/g)*
PU(CO)	45.2	198.5	87.1	92.9
PU(CO/PE)	58.2	248.5	115.4	96.5
PU(PE)	77.1	248.3	145.8	94.8
PU(CO)+AP	44.2	198.5	87.5	94.2
PU(CO/PE)+AP	56.3	248.7	113.1	99.7
PU(PE)+AP	79.1	248.4	140.4	96.5

Table 4: Summarized DMA values for Polyurethane and reinforced AP composites.

Sample		E`_{g -50ºC} MPa	E`_{r 100ºC} MPa	T_g (ºC)	FWHM (ºC)	PH	Peak E" (ºC)
PU(CO)	2612		38.5	27.9	26.6	0.78	14.6
PU(CO) +AP	2922		71.6	32.1	32.5	0.56	27.3
PU(CO/PE)	1861		16.9	8.8	29.8	0.91	1.5
PU(CO/PE)+AP	2441		22.6	21.9	39.4	0.66	9.2
PU(PE)	1434		4.7	-7.2	24.1	1.28	-12.7
PU(PE)+AP	1790		8.5	12.7	33.8	0.76	4.9

Abbreviations: FWHM: full width at half-maximum values; PH, peak height

Table 5: Calculated parameters from the fitting using an autocatalytic model (B_na).

Kinetics parameters	PU(CO)		PU(CO) +AP		PU(CO/PE)		PU(CO/PE)+AP		PU(PE)		PU(PE)+AP
log A₁[s^-1]	6.23	3.54		5.95		4.75		7.39		5.76
E_α₁ [kJ/mol]	50.00	40.02		45.15		41.64		58.14		51.86
n₁	1.85	1.99		1.98		1.12		1.75		1.76
m₁	0.26	0.34		0.42		0.10		0.47		0.29
log A₂[s^-1]	-	-		1.25		6.31		-		-
E_α₂ [kJ/mol]	-	-		42.69		57.86		-		-
n₂	-	-		1.54		1.35		-		-
m₂	-	-		0.28		0.25		-		-
E_visc0[kJ/mol]	6.91	9.24		1.85		5.89		9.81		19.64
E_visc1[kJ/mol]	44.93	41.29		47.96		40.55		51.86		53.21
R²	0.998	0.991		0.992		0.982		0.992		0.990

Table 6: Activation parameters calculated according to Wynne-Jones–Eyring–Evans equations for PUs materials with and without aramid pulp.

Thermodynamic parameters		PU(CO)	PU(CO) +AP	PU(CO/PE)		PU(CO/PE) +AP	PU(PE)	PU(PE) +AP
	ΔS* [J/K.mol]	-134.8	-186.3	-140.2	-163.2		-112.6	-143.8
				-230.2	-133.3
	ΔH* [kJ/mol]	-47.2	-37.4	-42.3	-38.8		-55.3	-49.1
				-39.9	-55.1

SupplementalMaterialPaper0127062022.docx

Effect of the Aramid Pulp on the Physicochemical, Viscoelastic Properties and Rheokinetics of Polyurethanes

Status:

Version 1

Abstract

Figures

Introduction

Experimental

2.1. MATERIALS

Results And Discussion

Conclusions

Declarations

References

Tables

Supplementary Files

Status:

Version 1