Viscoelastic Properties of Carboxyl-Terminated (Butadiene-co-Acrylonitrile)-Based Composite Rocket Propellant Containing Tris(2,3-Epoxypropyl) Isocyanurate as Bonding Agent

The influence of tris(2,3-epoxypropyl) isocyanurate as a bonding agent on the polymer network characteristics as well as rheological properties of [carboxyl-terminated (butadiene-co-acrylonitrile)/polyglycidyl ether of glycerol and epichlorhydrin]-based composite propellant was investigated. Amplitude sweep tests have been run above the glass transition temperature to establish the linear viscoelastic range (LVR) in a rubbery phase of the composite propellant for the correct physical interpretation of DMTA data. Temperature dependencies of rheological behaviour parameters (storage modulus, loss modulus, loss factor and glass-rubber transition temperature) were analyzed. It was observed that all the tested propellants have shown two relaxation processes in the temperature range –80°C to +80°C. The first one, structural relaxation ( α - transition) related to glass transition and the other one, secondary relaxation process. The activation energy for glass transitions was determined by the Arrhenius plot from dynamic properties measured at three frequencies (1, 5 and 10 rads -1 ). The dynamic mechanical analysis showed that the tris(2,3-epoxypropyl) isocyanurate content did not affect the glass transition temperature of the tested propellants


Introduction
HE essential ingredients of a composite propellant are a crosslinked binder (an organic polymer) and an oxidizer (usually ammonium perchlorate).The binder, as the name implies, holds the composition together and acts as an auxiliary fuel.The use of hydroxyl-terminated (HTPB) and carboxyl-terminated (CTPB) polybutadienes as prepolymers for composite propellants is the most common choice of polymeric materials to achieve the desired mechanical properties of the propellant in rocket motors [1,2,3].The composite propellant grain undergoes high stresses induced by a rapid acceleration, sharp turn, or rapid chamber pressure rise during launch and flying.A proper propellant must possess mechanical properties that prevent these stresses [4].In order to improve mechanical characteristics, an interaction between the ammonium perchlorate (AP) particles and the polymeric binder is necessary, which is achieved by bonding agents added to composite propellant composition.Bonding agents are typically used in HTPB-based composite propellants since these polymers are weakly polar [5].Due to the highly polar carboxyl (-COOH) group, carboxylterminated prepolymers (CTPB) do not require the use of bonding agents.If for a composite propellant production carboxyl-terminated (butadiene-co-acrylonitrile), known as CTBN, is taken, experiments have shown that tris(2,3epoxypropyl) isocyanurate (TEIC) as the bonding agent significantly improves mechanical characteristics [6].
The molecular structure of the polymer based composite propellant is the largest factor in determining the mechanical properties of the cured propellant.The mechanical properties of composite propellants depend strongly on the time, temperature, loading procedure and loading history [7].Dynamic mechanical thermal analysis (DMTA) is one of the most versatile techniques available for the characterization of T time-, frequency-, and temperature-dependent mechanical properties of polymers.No other characterization method provides more information in a single test [8].The technique may be used to characterize various polymers including thermoplastics, thermosets, elastomers and polymer based composites.Moreover, DMTA is often considered as the most accurate test method for determining the glass transition temperature and other structurally related phase transformation temperatures in polymers [9].
Since the bonding agent is an ingredient that is added to modify the mechanical properties of the propellant, precise knowledge of the dynamic mechanical behavior is extremely important.
The main objective of this paper is to broaden our knowledge of applying dynamic mechanical analysis to the study of CTBN-based propellant.As far as we are aware, there are no publications in the extensive literature, except our previous investigation, that have surveyed not only the viscoelastic properties of CTBN-based composite propellant but also the effect of this type of bonding agent on the CTBNbased composite propellant behavior.

Materials
Four CTBN-based composite propellant compositions (Table 1) were selected for this study.They consisted of 80 wt.% of solids loading including AP (oxidizer, with two particle sizes, 200 μm and 10 μm) and 20 wt.% of a polymeric binder, whose mix ratio of ingredients is expressed in phr (parts per hundred of resin based on 100 parts of CTBN).The binder composition consisted of 100 phr of CTBN (1300×15, viscosity at 25ºC: 66.  Tris(2,3-epoxypropyl) isocyanurate (TEIC) was prepared by the reaction of cyanuric acid and epichlorhydrin in accordance with a modified literature method [10].Investigated composite propellant samples differ from each other in the bonding agent content.Composition CP00 is a baseline propellant composition containing no bonding agent.In most of the currently used propellants, an amount from 0.1 to 0.3 is the preferred amount of the bonding agent [11].TEIC content varied from 0.1 to 0.3 wt.% of the total mass of the investigated propellant.
The chemical structures of the propellant ingredients are shown in Fig. 1.The tested propellant formulations were mixed in a 1gallon Baker-Perkins planetary mixer at a temperature of 60ºC.All of the binder ingredients, except the curing agent and the cure catalyst, were blended thoroughly for 15 min at an ambient pressure and then vacuum mixed for 30 min.The AP blend (fine and coarse fractions) was added portionwise and after each portion the mixture was blended for 10 min.The mixture was then vacuum mixed for 60 min.The curing agent was then added to the mixture, mixed for 10 min at an ambient pressure and then vacuum mixed for further 10 min.Finally, the cure catalyst was added to the mixture and the same procedure was repeated.The mixture thus prepared was cured for 5 days at 77±2ºC.

Sol-Gel measurements
The sol content of the propellant binder is the weight of the soluble fraction of the propellant, obtained by the extraction with a solvent, as described in STANAG 4581 [12].The extraction was carried out in three Soxhlet extraction units (with refluxing solvent) for 16 hours using approximately 150 mL of dichloromethane (CH 2 Cl 2 ) in each.The formula for the sol fraction evaluation was modified [13] and, in accordance with Equation (1), the polymeric part S poly of the total soluble part E total of the propellant sample was determined with respect to the polymeric part of the formulation A k-P .

Uniaxial tensile properties of the composite propellant binder samples
The uniaxial tensile properties of the cured propellant samples were evaluated at 20°C using INSTRON 1122 tensile test machine fitted with a 5000 N load cell.«JANNAF C» dog bone samples (120 mm length, 25 mm width, 8 mm thickness) were used.Crosshead speed of the tensile machine was 50 mm/min, while the effective gauge length of the JANNAF C samples was 68.6 mm.

Differential scanning calorimetry (DSC)
DSC analyses were carried out using the DSC Q20 manufactured by TA Instruments with liquid nitrogen cooling.The temperature scale is calibrated using the melting temperature of high purity indium.These measurements were made with the purpose to investigate the thermal based glass transition temperature ( g T ) of the tested samples.The small amount of the samples (~5 mg) was scanned with a heating rate of 10°C min -1 , in a range from -90°C to +50°C.The g T of the samples were detеrmined from the midpoints of the transitions.

Dynamic mechanical thermal analysis (DMTA)
DMTA study of the investigated propellant samples was performed in a torsion deformation mode using the Modular Compact Rheometer MCR-302 (Anton Paar GmbH) equipped with standard fixtures (SRF12) for rectangular bars, temperature chamber (CTD-620) having high temperature stability (±0.1) and liquid nitrogen with automated cooling accessories [14].The DMTA tests, in order to evaluate the influence of TEIC on viscoelastic properties of the tested composite propellant, have included the following specific experiments: -strain amplitude sweep tests, in order to determine the linear viscoelastic range (LVR) for the tested propellant,  The samples for the DMA tests were of a rectangular bar shape (54×10×3-4 mm).The complex shear modulus was determined, G*(ω) = G'(ω) + iG''(ω).The extracted data were the storage shear modulus (G'), the loss shear modulus (G'') and the loss factor tan(δ) = G''/G'.The glass transition temperature (T g ), determined by the dynamic mechanical measurements, was estimated as the temperature at which the loss factor, tan(δ) reached its maximum value.

Sol-gel analysis
Knowing the soluble fraction content (S poly ), the crosslink density (C) of the propellant binder (Table 2) can also be estimated following the modified version of the Charlesby-Pinner [15], see Equation ( 2).(2) The soluble or extractable content is a useful parameter to characterize the extent of cross-linking within a cross-linked polymer [13].The greater the fraction of not cross-linked polymer chains, the higher is the soluble fraction.A soluble fraction equal 1 means that the material is not cured and that all can be extracted.In contrast, a value of S poly tending to 0 means that nothing can be extracted.For a propellant formulation, the soluble fraction S poly should be in the range 0.15 to 0.2 to get an elastomer with a low cross-link density to achieve a suitable rubbery state.
For the specified TEIC content in the tested propellant compositions sol fraction in the polymer network is from 26 up to 37 %.With an introduction bonding agent into propellant composition (CP01), an increasing in the soluble fraction is evident, as well as the decrease in the crosslink density value.The percentage increasing in the soluble fraction was 5.39 % in CP01 (0.1 wt.% TEIC).But with a TEIC content of 0.2 wt.% the cross-link density value reached its maximum.Further increase of TEIC content resulted in a decrease of polymer network cross-link density.Decrease of the crosslink density can be explained by an increase in physical crosslinking due to the high concentration of epoxy curing agent.However, the highest percentage reduction in the soluble fraction was 16.91 % in CP02 (0.2 wt.% TEIC).
We conclude that the cross-link density of a polymer network cannot be determined definitely using this method alone.However, the increase in C value suggests that the degree of cross-linking should increase.

Uniaxial tensile mechanical characteristics
The average and standard deviation [SD] values of uniaxial tensile mechanical properties, i.e. the tensile strength (σ m ), the strain at maximum load (ε m ), the strain at break (ε b ) and the Young's modulus (Е) for the investigated propellant samples are listed in Table 3. SD -standard deviation.
All of the values listed in Table 3 were obtained as an average of five specimens.The uniaxial tensile properties of composite propellant depend on the tensile properties of the propellant binder, which in turn determines the strain capability of the propellant.High strain capability of the binder network is favorable considering long shelf life of the propellant system [16].The uniaxial test data showed that the values of the tensile strength and the Young's modulus increase as the TEIC content increases.Increasing the tensile strength and Young's modulus is caused by a larger amount of hard segments.The most pronounced improvement in the uniaxial tensile strength is shown by the propellant CP02.TEIC content at 0.2 wt.% causes an increase in tensile strength of 24.6% with a simultaneous decrease in the strain at maximum load value of 5.4%.
The incorporation of TEIC into the propellant formulation decreases the value of the strain at maximum load as a consequence of the decreased molecular weight of the polymer chains between the crosslink points.The same tendency can be seen in the strain at break values as a function of the effective polymer chains between the crosslink points [6].

Strain amplitude sweep test
In order to use the DMTA technique to accurately determine rheological properties of materials, a tested polymer must be deformed at the amplitudes that remain within the linear viscoelastic region (LVR).Within LVR, the viscoelastic response of the polymer is independent of the magnitude of deformation [17].As a general rule, this region must be determined for every type of polymer by DMTA amplitude sweep tests, in which a frequency is fixed and the strain amplitude is incrementally increased.Fig. 3. shows the dependence of the shear storage modulus, G', on the strain amplitude for the baseline propellant sample CP00.
G' G'' The logarithm of the storage modulus decreases linearly with the logarithm of increasing the shear strain level, indicating some form of damage function exists which increases with deformation (Figure 4).Similar behavior was exhibited by the loss modulus.This phenomenon of straindependent material properties, termed nonlinear viscoelastic behavior, is often associated with the presence of the filler particles.Data in Fig. 4, representing CTBN-based composite propellant as a material utilizing predominately reinforcing filler, show that the logarithm of modulus is linearly dependent on the strain over approximately 2 decades of strain.This suggests that over the strain range examined in this work, constitutive relationships can be developed of the form: where A and B are the shear storage and loss moduli at 1 % strain, respectively.γ is the strain amplitude and the exponents m and n are the factors representing the degree of nonlinear response [18].Table 4 lists the factors obtained by fitting the data in Fig. 4 using equations ( 3) and (4).Agreement between these equations and data is very high as indicated by the linear regression correlation coefficients, given in Fig. 4. For all tested propellants except the propellant CP00, the absolute value of n is greater than m.Relative values of these nonlinear factors are most likely dependent on the region of the strain investigated.
The composite propellant sample CP02 displays the greatest nonlinear behavior (largest absolute values of m and n).The effect of crosslinking or crosslink density appears to increase the nonlinear behavior.

Temperature dependence of rheological parameters
Based on a weakly crosslinked polymeric matrix, composite rocket propellants show very large changes in mechanical properties with temperature [19,20].Figure 5 shows the viscoelastic response of a baseline composite propellant sample CP00 in terms of G', G'' and tan(δ) as a function of temperature.
-100 -80 -60 -40 -20 0 20 40 60 80 100 10  It follows from Fig. 5 that the G' is almost constant at low temperatures and it reaches the maximum value in the region below the glass transition (−90°C to −80°C).The transition from the glassy to the rubbery state (the onset point on the G'-T curve at T= −73°C) has, as a consequence, a drop in the G' value by a factor of about 35 (from 978 MPa at −90°C to 27 MPa at −40°C).With increasing the temperature (between T= −60°C and T=20°C) the G' curve exhibits a two-step character, indicative of two separate relaxation processes.Similar temperature dependence of G' curve was observed in the CTBN-based composite propellant binder composition [21].Thus, the existence of two relaxation process observed in the composite propellant composition could not be attributed only to the influence of AP solid particles on the polymer network mobility.Copolymers often show a broadening of the transition region with a decrease in the slope of the modulus curve.Part of broadening of the transition region could be due to the heterogeneity in the molecular weight between cross-links.Also, the reason for a slow decrease of the storage modulus could be crystallinity within a polymer structure of the tested propellant binder samples [22].
A molecular interpretation of the viscoelastic behavior can be given considering the tan(δ), which describes molecular rearrangement regions, corresponding to the binder fractions with different mobility.There is a question what the glass transition temperature is.According to one group of authors the T g is the maximum of the G''.The other choice for assigning a T g is the maximum of the tan(δ).At the maximum intensity of the tan(δ), the molecular rearrangement process from glassy to rubbery and vice versa is at its maximum.Therefore, in order to characterize the transition between these two states, the maximum of the tan(δ) must be taken as glass (to rubber) transition temperature.Also, tan(δ) is much more sensitive than the G'' in showing the different temperature regions of molecular rearrangements in order to change the basic state from energy to entropy elasticity [23].Polybutadiene-based elastomers contain soft and rigid segments [7].The soft segments consist of large flexible noncrystalline polymer chains, in this case, polybutadiene, while rigid (hard) segments consist of crosslinking units.
The peak of tan(δ) positioned between T= −65°C and T= −55°C (its temperature location is designated as T g ), is attributed to the glass-rubber transition temperature.This relaxation process involves cooperative motions of prepolymer (CTBN) main chains between the crosslinks within the soft segment regions.These polymer network segments have a very low density of ester groups and are the polybutadiene part of binder which reorient at very low temperatures.The T g value for the baseline propellant composition, CP00, is −59.1°C.
Being reinforcing (filler strongly interacts with the binder) or non reinforcing (filler interact only weakly with the binder), viscoelastic properties of composite propellants are strongly affected by the inclusion of solid fillers.Particlepolymer interactions as well as physical confinement effects, have a significant influence on the polymer chain mobility.Although the molecular segments are almost free to move in the rubbery region, the incorporation of highly active fillers such as AP, significantly modifies the molecular mobility.Such systems indicate the appearance of a second, higher T g , reflecting a constrained population of polymer chain segments in the vicinity of the filler particle surface [24].
The second peak (its temperature location is designated as T g restricted ), broader than the first one, is located 60-100 deg higher than the first.The peak is assigned to the glass transition of polymer chains experiencing mobility restrictions due to the interactions with ammonium perchlorate particles.
The Fig. 6 shows temperature dependencies of the loss factor, tan(δ), loss modulus, G'' and storage modulus, G' of the tested composite propellant samples.
CP01 propellant sample (0.1 wt.% TEIC) has the lowest values of G' in the plateau of glassy state.Modulus in the glassy state is determined primarily by the strength of the intermolecular forces and the way the polymer chains is packed [25].The increase in storage modulus values with increasing TEIC content suggests that a tougher, more rigid network is obtained at higher TEIC concentrations.The transition from the glassy to the rubbery state has as a consequence a drop in the G' value, also the G'' maximum value.The loss modulus in the transition region is higher for propellants with higher TEIC content, which may be due to an increase in internal friction, creating energy dissipation.Although a broadening of the loss modulus peak was not observed, this may be attributed to the inhibition of the relaxation process within the composition as a consequence of a higher number of chain segments upon TEIC addition.
Considering the tan(δ) curves of tested propellants, no change in the main glass-rubber transition temperature was observed on varying the content of TEIC.
The T g values determined by G'', G' and tan(δ) are listed in Table 5.According to the temperature dependences of the loss factor, T g restricted , tan(δ) max1 and tan(δ) max2 values were determined, also temperature intervals (W 1 and W 2 ) coresponding to these values.These are parameters used to analyze the influence of applied bonding agent on tested propellant compositions.
For the tested propellant samples only quite small variations in T g values were observed.Regardless the TEIC content, the T g values are located in the narrow temperature range (Т g ~-60°C).The reason is the concentration of crosslink points apart from the polybutadiene parts, which are responsible for the main glass-rubber transition at low temperatures.

41.4
The invariance in Т g values indicates that distances between the crosslink points within the polymer network of tested propellant samples are such that the restrictions affected by the presence of the bonding agent during the crosslink process do not contribute to differences in Т g value.In other words, cooperative motions of polymer network segments, related to main relaxation transition are not influenced.The corresponding relaxation time is only changed minor by the structural changes in propellant binder cross-link structure due to the TEIC addition.
It is observed that, as the TEIC content increases, the width of the second tan(δ) peak (designated as W 2 ), which corresponds to tan(δ) max2 value, gets smaller.This can be explained by the glass transition of polymer chains experiencing more mobility restrictions due to the interactions with the amonium perchlorate particles.Formation of hydrogen bonds between the TEIC as a part of a polymer network on one side, and ammonium perclorate particles on the other side can be a reason for the mobility restrictions.This is the most pronounced in the CP03 propellant composition.In comparison with the CP00 composition, reduction in W 2 value is 53.1%.

Effect of TEIC on activation energy of relaxation events
Dynamic mechanical analysis is one of the most appropriate methods to investigate relaxation events.Bihari et al. in their article have demonstrated that the interpretation of viscoelastic relaxation strength in polymers depends on the tangent of the phase lag angle (tan δ) or real (in-phase) or imaginary (out-of-phase) components of the storage and loss modulus, respectively [26].The glass transition temperature (T g ) of a material also referred as α-transition is normally observed as a peak of tan(δ) considering the variation of tan(δ) with temperature (Figure 7).At this point the material absorbs energy and the modulus of the material decreases as it passes through the glass transition phase.Normally, at a lower temperature, sometimes other relaxation processes are observed for polymeric materials.The secondary transition, known as β-relaxation, is normally attributed to the polymer backbone configuration reorganization.The activation energy (E a ) represents the energy barrier of the material of which has to overcome in order to undergo the structural reorganization of the relaxation process.The E a activation energy corresponding to α-transition was determined by the Arrhenius equation, see Equation (5).
where F is frequency (Hz), A is the preexponential factor, E a is the activation energy (kJmol -1 ), R is the gas constant (kJ(molK) -1 ) and T is the temperature (K).The activation energies were calculated from tan(δ) peaks and onset values of dynamic properties measured at frequencies 1, 5 and 10 rads -1 using the Arrhenius plots for each tested propellant.
The E a values corresponding to α-transition are given in Table 6.It is found that the E a for α-transition of tested propellants is from 160 up to 199 kJmol -1 .
The E a associated with the T g process, which is attributed to the cooperative motions of long soft segments, is higher for a baseline propellant composition (CP00) compared to the propellant compositions CP01 and CP02.This result suggested that the relaxation of CTBN was made easier by the addition of 0.1 and 0.2 wt.% of TEIC.In other words, lower contents of TEIC affect the concentration of short linking hard segments, whereas 0.3 wt.% of TEIC can readily interfere with soft segment motions in CTBN polymer network.

DSC measurements
As an essential ingredient, the binder largely determines the mechanical properties of the composite solid propellants.Low glass transition temperature is preferred for the composite propellant binder, so that it can withstand a large spectrum of stress transients in operation.Therefore, the glass transition temperature (T g ) of elastomer bonded composite propellant is one of the most important properties determining their in-service application [27].T g values of the tested propellant binder samples determined through DSC are shown in Table 7.The differences between T g values of tested propellant samples determined through DSC are within experimental error.Obtained T g values show that it cannot be observed any regularity in its change with increasing TEIC content in the tested propellant formulations.
Glass transition temperatures, determined by DSC, are significantly lower than the ones determined by DMA.The reason is the ''static'' determination, which means no mechanical deformation causing strain hardening is applied, neglecting the volume change by thermal expansion [28].The invariance in Т g values, regardless of the method of determining, indicates that distances between the crosslink points within the polymer network of the tested binder samples are such that the restrictions affected by the presence of the bonding agent during the crosslink process do not contribute to differences in Т g value.In other words, cooperative motions of polymer network segments, related to main relaxation transition are not influenced.The corresponding relaxation time is only changed minor by the structural changes in binder cross-link structure due to TEIC addition.

Conclusion
Samples of the composite propellant based on carboxylterminated (butadiene-co-acrylonitrilе) cured with polyglycidyl ether of glycerol and epichlorhydrin (LX-112) containing the bonding agent tris(2,3-epoxypropyl) isocyanurate (TEIC) were investigated with sol-gel analysis, dynamic mechanical analysis and uniaxial testing.The temperature dependences of rheological parameters disclose that the glass-rubber transition temperature of the investigated propellants do not change by the addition of tris(2,3epoxypropyl) isocyanurate.The glass transition temperature, T g , of all propellant samples was about -60°C.TEIC as the bonding agent was found to alter uniaxial tensile properties of the tested composite propellants.Both the tensile strength and Young's modulus increased, while the strain at maximum load and the strain at break decreased.Mots clés: propergol composite, butadiène acrylonitrile, iso cyanure, température de vitrification.

Figure 1 .
Figure 1.The chemical structures of the ingredients of the investigated propellants o T=20°C, o shear strain, γ from 0.001 up to 10 %. -temperature sweep tests at different frequencies, with the aim to evaluate viscoelastic relaxation processes, o temperature range from -80 to +40°C, o the heating rate was 2°C/min, o shear strain, γ=0.01%, o angular frequencies 1, 5 and 10 rads-1.-temperature sweep tests in heating mode, o temperature range from -80 to +80°C, o the heating rate was 5°C/min, o shear strain, γ=0.01%, o single frequency point of 1 Hz.The equipment and details of a fixed specimen are shown in Fig.2.

Figure 2 .
Figure 2. The Modular Compact Rheometer MCR-302 equipped with the temperature chamber

Figure 3 .
Figure 3. Amplitude sweep test: variation of the storage and loss shear moduli with the strain amplitude for the baseline propellant sample CP00

Figure 4 .
Figure 4.The effect of the strain amplitude on the shear storage modulus for tested propellant sample for the baseline propellant sample CP00

Table 2 .
S prop , S poly and C values for the tested propellant samples

Table 3 .
Values of uniaxial tensile characteristics of tested CTBN-based propellant samples at T=20°C.

Table 4 .
Nonlinear factors and reference moduli

Table 5 .
tan(δ) max values and T g values determined by G'', G' and tan(δ)

Table 6 .
Activation energy of main relaxation process for tested propellants

Table 7 .
Glass transition temperature values of the tested propellant binder samples