Polymer formulation

CHAPTER 1

INTRODUCTION

Pigments are additives in a polymer formulation which provide countless possibilities to designers who want to differentiate their product. Legislation and uprising environmental awareness has led to the gradual phasing out of heavy metal inorganic pigments and increased usage of organic pigments. Despite their good heat stability, light fastness, tinctorial strength and low cost, certain organic pigments are widely known to cause significant warpage in polyethylene mouldings (even at pigment concentrations as low as 0.1% wt).[1,2] This phenomenon is especially common in large thin-walled mouldings such as lids, bottle crates and trays.[3]

It is generally accepted that the warpage phenomenon is caused by the nucleating effect these organic pigments have on polyethylene. They act as nucleating agents, increasing crystallisation rate and altering the morphology of mouldings. Morphological changes cause higher internal stress which leads to distortion.[2] Adding on to the problem, different organic pigments nucleate polyethylene to different degrees, making it impossible to produce mouldings with identical dimensions using identical processing conditions when a variety of pigments are used.[4]

Numerous attempts have already been made, with usually moderate success, to solve organic pigment induced warpage. They range from adjusting process parameters, mould design changes, pre-treatment of pigments, to incorporation of additional additives. A review of literature in this research area showed that although some studies have been conducted to investigate the incorporation of nucleating agents to override nucleating effects of organic pigments on polypropylene, limited information of this sort exists for polyethylene. The specific mechanism behind nucleating agents overriding nucleation by organic pigments is also still unclear.

Therefore, it is the aim of this research to study the influence of nucleating agents, based on potassium stearate and carboxylic acid salts, on the crystallisation and warpage behaviour of high density polyethylene containing copper phthalocyanine green pigment. Differential Scanning Calorimetry (DSC) and Optical Microscopy (OM) will be employed to follow the crystallisation behaviour of the formulations and correlations between rate of crystallisation and shrinkage behaviour will also be made.

CHAPTER 2

LITERATURE REVIEW

2.1. Nucleation and Crystallisation of Semi-Crystalline Polymers

2.1.1. Crystallisation Mechanisms

Crystallisation involves the formation of an ordered structure from a disordered phase, such as melt or dilute solution.[5] The crystallization process of polymers is thermodynamically driven. It is governed by change in Gibbs free energy, ΔG.[6]

ΔG = ΔH – TΔS (2-1)

Where ΔH is change in enthalpy, T is absolute temperature and ΔS is change in entropy.

When ΔG is negative, crystallisation is thermodynamically favourable. This occurs when loss of enthalpy upon crystallization exceeds the loss of entropy multiplied by absolute temperature. It can therefore be derived that as the absolute temperature of the system falls, the driving force of crystallisation will increase.[7]

For a polymer to crystallise, it must conform to the following requirements:[8]

• Molecular structure must be regular enough to allow crystalline ordering
• Crystallisation temperature must be below melting point but not close to glass transition temperature
• Nucleation must occur before crystallisation
• Crystallisation rate should be sufficiently high

A hundred percent crystallinity is not possible in polymers due to factors such as chain entanglements, viscous drag and branching. Thus they are termed ‘semi-crystalline’. All semi-crystalline polymers exhibit a unique equilibrium melting temperature above which crystallites melt and below which a molten polymer starts to crystallise. The crystallisation of semi-crystalline polymers is a two-step process consisting crystal nucleation and crystal growth.[6]

2.1.2. Primary Nucleation

Primary nucleation can be defined as the formation of short-range ordered polymer aggregations in melt which act as a focal centre around which crystallization can occur.[9] There are three mechanisms of primary nucleation, namely, homogeneous nucleation, heterogeneous nucleation and orientation induced nucleation.[10]

2.1.2.1. Homogeneous Nucleation

Homogeneous nucleation involves the spontaneous creation of nucleus in a semi-crystalline polymer melt when it is cooled below its equilibrium melting temperature.[7] This process is termed as sporadic as nuclei are formed in timely succession.[11] Creation of nuclei occurs when statistical variation within a polymer melt results in the formation of ordered assemblies of chain segments larger than a critical size[7]; usually between 2-10nm.[11] Below this critical size, the nuclei are unstable and may be destroyed.[11]

Generally, super-cooling to between 50-100°C below equilibrium melting temperature is minimally required to achieve true homogeneous nucleation.[12] The super-cooling is attributed to the energy barrier homogeneous nuclei are required to overcome to reach stability.[7].

When molecular segments pack next to each other to form an embryo, there is a change in free energy, ΔG, caused by two opposing mechanisms. The creation of new crystal surface increases free energy (ΔS is negative) while the reduction in volume of the system decreases free energy (Δ(U+pV) ≈ ΔH is negative). The two opposing mechanisms lead to a size-dependent free energy curve which defines critical nucleus size.[13] A small embryo has high surface to volume ratio and so ΔG is positive; in other words, crystal growth is not thermodynamically favourable.[13,14] However as nuclei grow, the surface to volume ratio decreases up to a point where volume change outweighs the creation of new surface and change in free energy decrease; crystal growth becomes increasingly probable. This point is defined as critical nuclei size and above this point, the energy barrier is overcome.[13,14] Eventually when ΔG becomes negative, nuclei are thermodynamically stable, paving the way for further growth into lamellae or spherulites.[14]

The minimum number of unit cells required to form a stable nuclei decrease when temperature decrease, due to a reduction in energy barrier. In other words, the rate of homogeneous nucleation increases when temperature of the polymer decreases.[7]

2.1.2.2. Heterogeneous Nucleation

In practice, one usually observes heterogeneous nucleation and not homogeneous nucleation.[15]

Heterogeneous nucleation involves the formation of nuclei on the surface of foreign bodies present in the molten phase of a semi-crystalline polymer. The foreign bodies can take the form of adventitious impurities such as dust particles or catalyst remnants, nucleating agents added on purpose or crystals of the same material already present in the molten phase (self-seeding).[7,8]

The presence of foreign bodies greatly reduces the energy barrier for the formation of stable nuclei. This reason for this is, polymer molecules which solidify against pre-existing surfaces of foreign bodies create less new liquid/solid interface than the same volume of polymer molecules forming a homogeneous nucleus.[6] In turn, critical size of nuclei is smaller in heterogeneous nucleation as compared to homogenous nucleation so that heterogeneous nucleation always occurs at lower supercooling.[16]

Foreign bodies with crystallographic spacings matching the semi-crystalline polymer are especially effective heterogeneous nucleating agents. Favorable nucleation sites include crystal grain boundaries, cracks, discontinuities and cavities.[7]

2.1.2.3. Orientation-Induced Nucleation

Orientation-induced nucleation is caused by some degree of molecular alignment in the molten phase of a semi-crystalline polymer. Molecular alignment reduces the entropy difference between the molten and crystalline state of the polymer. This kind of nucleation is important in various processes such as fibre melt-spinning, film-forming and injection moulding. In these processes, polymer melt is sheared before and during crystallisation.[8,17]

2.1.3. Crystal Growth

2.1.3.1. Primary Crystallisation

Primary crystallisation occurs when melt of a semi-crystalline polymer is cooled below its equilibrium melting temperature. It involves molecular segments depositing onto the growing face of crystallites or nuclei. The resultant crystal growth occurs along the a and b axes, relative to the polymer’s unit cell.

These additions of molecular segments can occur through two mechanisms: tight fold adjacent re-entry or independent deposition (illustrated in Figure 2.3).[6] Tight fold adjacent re-entry requires that chain stems be laid down continuously from a single polymer molecule in a series of hairpin bends until its length is exhausted. This single molecule is thought to be ‘reeled in’ from surrounding molten material.[7] This mechanism requires that molecular motions along the polymer molecule’s contour length to be several times faster than the rate of crystal growth. On the other hand, the independent deposition mechanism only requires localized motion of molecular segments. Molecular segments only need to re-organise sufficiently to align with molecular segments at the crystallite face.[6]

1. tight fold adjacent re-entry
2. independent deposition[6]

2.1.3.2. Secondary Crystallisation

After a semi-crystalline polymer is cooled to room temperature, crystallisation is still thermodynamically favourable but restricted by the low mobility of molecular segments in its amorphous regions. Over an extended period of time, which can span from hours to weeks, re-arrangement of molecular segments within amorphous regions can lead to further crystal growth. This process is defined as secondary crystallisation. Secondary crystallisation can take two forms; either thickening of pre-existing crystallites by re-organisation of amorphous chain segments adjacent to crystallite surface or creation of new crystallites by re-organisation of amorphous chain segments in interstitial regions between pre-existing crystallites. [6]

2.1.4. Rate of Crystallisation

The crystallisation of semi-crystalline polymers is a two-step process and therefore overall crystallisation rate is governed by both nucleation rate and crystal growth rate. Both factors are highly temperature dependent, as illustrated in Figure 2.4. When temperature is just below equilibrium melting point, there exists a meta-stable region where rate of nucleation is low as nuclei that are formed dissolve easily due to high thermal motions.[8] As super-cooling increases, thermodynamic conditions become more favourable and rate of nucleation increases and reaches a maximum near the glass transition temperature. On the other hand, kinetic conditions are less favourable as super-cooling causes viscosity to increase. This results in a shift in maximum rate of crystal growth to higher temperatures where viscosity decrease is balanced by formation of nuclei.[8,18]

Overall crystallisation rate at a given temperature is usually expressed as the inverse of time needed for half of the crystals to grow in the polymer (1/ t1/2).[8]

When crystallisation occurs under isothermal conditions, its progress can be expressed by the Avrami equation:[8]

Xc(t) = 1 – exp (-K.tn) (2-2)

Where Xc(t) is the fraction of material transformed at time t, n is the Avrami exponent and K is the Avrami rate constant.

Equation (2-2) may also be written as:[19]

ln ( -ln |1-Xc(t)| ) = n ln (t) + ln K (2-3)

So that n and K may be obtained by plotting ln ( -ln |1-Xc(t)| ) against ln (t); n is the slope while ln K is the y-intercept.[19]

The value of the Avrami exponent, n, is dependent on mechanism of nucleation and geometry of crystal growth. Theoretical values of n corresponding to different nucleation modes and crystal growth shape are tabulated in Table 2.1.[19]

Table 2.1: Relation between n and nucleation mode / crystal growth shape[19]

When crystallisation occurs under constant-cooling-rate conditions, its progress can be expressed by the Ozawa equation:[8]

Xc(t) = 1 – exp (-ĸ(t) / ϕm) (2-4)

Where ĸ(t) is the Ozawa rate constant, ϕ is the constant cooling rate (- δT/δt) and m is the Ozawa exponent.

Equation (2-4) may also be written as:

ln ( -ln |1-Xc(t)| ) = – m ln (t) + ln ĸ(t) (2-5)

So that m and ĸ(t) may be obtained by plotting ln ( -ln |1-Xc(t)| ) against ln (t); m is the slope while ln ĸ(t) is the y-intercept.

Qiu et al. combined the Avrami and Ozawa equations to make a connection between the Avrami and Ozawa exponents:[20]

log ϕ = log F(T) – a log t (2-6)

Where a = n/m and the kinetic function F(T) = (ĸ(t) / K)1/m. At a given degree of crystallinity, a plot of log ϕ against log t will yield a and log F(T) as the slope and y-intercept respectively.[20]

2.2. High Density Polyethylene (HDPE)

2.2.1. Chemical Structure, Crystallisation Rate and Morphology

High density polyethylene, HDPE, is a semi-crystalline polymer made up of repeat units (C2H4)n and has a general form as illustrated in Figure 2.5. It consists mainly of unbranched molecules with very few defects to disrupt its linearity or hinder crystalline packing. As such, HDPE has a high rate of crystallisation, degree of crystallinity and density (0.94-0.97 g/cm3).[7]

Being a semi-crystalline polymer, HDPE exhibits a three-phase morphology consisting of submicroscopic crystals surrounded by a non-crystalline phase comprising a partially ordered layer adjacent to the crystals and disordered material in the intervening spaces. This is illustrated in Figure 2.6.[7]

The unit cell of HDPE, defined as the smallest arrangement of its chain segments that can repeat in three dimensions to form a crystalline matrix, is orthorhombic; a cuboid with each of its axes having different lengths while the angles of adjoining faces are all 90°. Each unit cell is made up of two ethylene repeat units; a complete unit from one chain segment and parts of four others from surrounding chain segments.[7] Bank and Krim[21] reported that the a, b and c axes of a polyethylene unit cell are of dimensions 7.417, 4.945 and 2.547Å respectively. This is illustrated in Figure 2.7.

1. orthogonal view,
2. view along c-axis[7]

HDPE unit cells pack together in a three dimensional array to form small crystals known as crystallites. Most commonly, crystallites of HDPE take the form of ‘lamellae’; crystallites with a and b dimensions that are much greater than their c dimensions. Lamellae thicknesses are usually between 50 to 200Å while lateral dimensions can range from a few hundred angstroms to several millimetres. Figure 2.8 illustrates a HDPE lamella.[7]

Various models have been proposed to explain the arrangement of molecular chains in lamellae. They include adjacent re-entry with tight folds, switchboard, loose loops and a model with combined features (illustrated in Figure 2.9). As molecular length of HDPE is known to be many times greater than lamellae thickness, all models indicate some form of chain folding. However, they differ in their specific nature of folding.[7]

d) composite model[7]

In HDPE, the most common large scale-structures composed of crystalline and non-crystalline regions are known as ‘spherulites’. A spherulite consists of lamellae growing outward radially from a common nucleation site. As this growth advance into amorphous molten polymer, local inhomogeneities in concentrations of crystallisable segments will be encountered. This causes the folded chain fibrils to inevitable twist and branch. As illustrated in Figure 2.10a, a spherulite will resemble a sheaf in its early stage of development. Fanning out of the growing lamellae will subsequently produce a spherical structure but true spherical symmetry is never achieved due to impingement of neighbouring spherulites. This growth of spherulites also involves the segregation of non-crystalline materials into regions between lamellar ribbons. Thus the overall structure of a spherulite consists of twisted and branched lamellae with polymer chains mostly perpendicular to their long axis and amorphous regions (illustrated in Figure 2.10b).[22]

2.3. Organic Pigments

2.3.1. Copper Phthalocyanine Pigments: Copper Phthalocyanine Green

Copper phthalocyanines are a class of organic pigments which dominate the sectors of blue and green coloration of polymers. This dominance can be attributed to desirable properties such as high tinctorial strength, bright hues, excellent light and weather fastness excellent heat stability and good chemical resistance.[23] In addition, in spite of its structural complexity, this class of pigments is inexpensive as they are manufactured in high yield from low cost starting materials.[24]

The parent compound of copper phthalocyanine pigments is extremely easy to prepare; a phthalic acid derivative is condensed with a source of nitrogen such as urea and a copper salt such as cuprous chloride in the presence of a metal catalyst such as vanadium or molybdenum. This is usually done in organic solvents, at elevated temperatures (approximately 200°C) and sometimes under increased pressure. The resultant crude copper phthalocyanine (yields of over 90%) is purified commercially by one of several processes; salt attrition, solvent-free salt attrition, acid pasting and acid swelling.[3,25]

Figure 2.11 illustrates the chemical structure of the copper phthalocyanine parent compound. It consists of a tetrabenzoporphyrazine nucleus containing a central copper atom. The planar molecule is in the form of a quadratic shape with length and thickness of 1.3nm and 0.34nm respectively.[27] This parent copper phthalocyanine compound, which is characterised by unsubstituted benzene rings, is used as blue pigment. Copper phthalocyanine blue is polymorphous and exists in five crystal forms. Out of the five, the two of commercial importance are the alpha and beta forms while the other three are distorted α forms.[27] Different crystal forms bring about a variation in the blue shade. Alpha crystals exhibit a bright-red-shade blue while beta crystals exhibit a green-shade blue.[26]

1. C.I. pigment green 7, b) C.I. pigment green 36 (3y),
2. C.I. pigment green 36 (6y)[28]

Copper phthalocyanine green, the pigment of interest in this project, is produced from the copper phthalocyanine blue by replacing the hydrogens on the four benzene rings with halogens. Unlike its blue counterpart, where variation of shade is achieved by modification of crystal form, variation in the green shade is controlled by degree of chlorination or bromination. Copper phthalocyanine green only has one known crystal form.[26] The two types of copper phthalocyanine green pigments are colour index (C.I.) pigment green 7 and colour index (C.I.) pigment green 36. C.I. pigment green 7 is a blue-shade green made by introducing thirteen to fifteen chlorine atoms to replace hydrogens in the benzene ring of the copper phthalocyanine blue molecule (illustrated in Figure 2.12(a)). C.I. pigment green 36 is a yellow-shade green made by gradual replacement of chlorine atoms in C.I. pigment green 7 with bromine atoms. The most brominated C.I. pigment green 36, known as 3y, has an extreme yellow shade (illustrated in Figure 2.12(c)) while the least brominated C.I. pigment green 36, 6y, has a much more bluish shade (illustrated in Figure 2.12(b)).[28]

The outstanding tinctorial and fastness properties of both copper phthalocyanine green pigments allow their application under the harshest conditions. They can be used effectively in masstone tints and shades down to the very palest depth. Both green pigments can be processed at temperatures in excess of 260°C with little colour change. They have even better chemical and colour stability than copper phthalocyanine blues. On comparison, C.I. pigment green 7 is preferred over C.I. pigment green 36. The latter is weaker and more opaque and accounts for less than 5% of copper phthalocyanine greens used in the polymer industry.[3]

2.3.2. Effect of Copper Phthalocyanine Green and Other Organic Pigments on Properties and Crystallisation Behaviour of Moulded Polyolefins

Although the combination of spectacular performance and low cost make copper phthalocyanine green ideal pigments, its use is not without challenges. It is widely known that copper phthalocyanine green and a few other pigments can cause unacceptable levels of shrinkage and warpage in moulded parts of polyolefins.[2,29] The problem persists even at pigment concentrations as low as 0.1% wt.[2] Shrinkage can be described as reduction in moulded part dimensions in reference to mould cavity dimensions.[30] Warpage is a measure of out-of-plane distortion and commonly arises from the relaxation of unbalanced residual stress in a moulded part or unbalanced shrinkage in flow and transverse direction.[30]

The early work of Turturro et al.[2] demonstrated that this shrinkage and warpage phenomenon is only limited to organic pigments. It was reported that no distortion occurred in HDPE mouldings containing inorganic pigments such as BBS red (cadmium selenide), 21 M yellow (blend of PbCrO4, PbSO4 and PbMoO4) and 500 L yellow (complex of Ni and Ti). Findings from later studies by Bugnon et al.[31] and Suzuki & Mizuguchi[29] are in good agreement. Suzuki & Mizuguchi[29] reported similar observations when they incorporated inorganic pigments, TiO2, Fe2O3 and Cd Y into HDPE and PP. Using scanning electron microscopy, Bugnon et al.[31] were able to show that when inorganic pigments such as CdS or CrTiO4 are incorporated into HDPE, there is no interaction between pigment surface and polymer. The polymer essentially builds a cavity around the pigment. On the other hand, an organic pigment of diketo-pyrrolo-pyrrole chemistry was found to blend into the HDPE matrix. This led them to propose that inorganic pigments do not induce shrinkage and warpage as their chemical constitutions and polar hydrophilic surfaces have no interactions with polymers and do not influence their crystallisation behaviour.

It is generally agreed that the shrinkage and warpage of polyolefins induced by copper phthalocyanine green and other organic pigments is associated with the nucleating effect these compounds have on the polymers.[2,29,31] These compounds provide a foreign surface that reduces the free energy of formation of a new polymer nucleus.[27] Vonk[32] was one of the first few individuals who pointed out that organic pigments can act as nucleating agents for polyethylene. The nucleating effect of organic pigments on polyolefins has since been the focus of intensive studies over the years. The key literature identified from this research area is that produced by Koh[33] for Clariant (Singapore) Pte Ltd. Koh[33] studied the influence of C.I. pigment green 7 and C.I. pigment green 36 on the crystallisation and properties of HDPE. It was reported that the high level of differential shrinkage in HDPE mouldings incorporated with copper phthalocyanine greens was accompanied by increased crystallisation rate, increased peak / onset crystallisation temperature and reduced spherulite size. These findings clearly indicate that copper phthalocyanine green can act as a nucleating agent for HDPE. It was also reported that increasing pigment concentration will cause an increase in crystallisation rate and level of differential shrinkage.

Koh’s[33] findings are in line with those from similar studies carried out by Turturro et al.[2], Suzuki & Mizuguchi[29] and Silberman et al.[34] Turturro et al.[2] observed a similar nucleating effect of copper phthalocyanine green on HDPE with the aid of depolarisation and dilatometry techniques. In addition, they found that the Avrami exponent value of HDPE decreases with increasing concentration of copper phthalocyanine green; which indicates a shift in morphology, away from the spherulitic one characteristic of pure polyethylene. They proposed that the strong nucleating effect of copper phthalocyanine green causes only the development of fibrils in HDPE, which subsequently do not organise into spherulites. Interestingly, they also found that pigments do not affect the absolute level of crystallinity in HDPE; implying that these compounds affect only the kinetics and not the thermodynamics of the crystallisation process.[2] Suzuki & Mizuguchi[29] and Silberman et al.[34] showed that, apart from HDPE, copper phthalocyanine green can also act as a nucleating agent for PP. Moreover, Silberman et al.[34] found that the addition of copper phthalocyanine green into PP would increase its lamellar size and decrease the activation energy (Uact) of its crystallisation process. The explanation they put forward for these observations was based on the specific chemical structure of the pigment. The symmetry of nitrogen in the copper phthalocyanine green molecule, with an absence of complex structures was thought to promote the dynamic adsorption of PP molecules on the pigment surface and the subsequent crystallisation process. This will lead to the formation of a perfect crystal structure of large lamellar size. Together, the works from all three authors demonstrated that, besides copper phthalocyanine green, organic pigments of anthraquinone, perylene, quinacridone, copper phthalocyanine blue and condense azo chemistries can also act as nucleating agents for polyolefins.[2,29,34]

At this point, with the aid of various papers, it is established that shrinkage and warpage of polyolefins induced by copper phthalocyanine green and other organic pigments are associated with these pigments serving as nucleating agents for the polymer. However the specific mechanism correlating nucleating effect and shrinkage or warpage has yet to be discussed. Both Turturro et al.[2] and Suzuki & Mizuguchi[29] proposed the same explanation for this phenomenon. In a moulding process such as injection moulding, the quench rate is not the same at different parts of the polymer. Polymer melt in contact with mould walls crystallise and ‘freeze’ very quickly, which results in crystals of low perfection with polymer chains oriented in the direction of flow. This layer of imperfect crystals in turn impedes heat exchange between polymer melt in the core regions and the mould walls. As a result, polymer melt in the core regions cool slowly and give rise to regular crystals. As the surface ‘freezes’ very quickly, contraction in the core regions due to crystallisation will produce stress in the ‘frozen’ outer layer and cause distortion. In addition, relaxation of oriented regions after removal of polymer from the mould will also cause internal stress and lead to distortion. The presence of a strong nucleating agent such as copper phthalocyanine green will limit the time available for oriented chains to recover during cooling and also increase the thickness of the skin layer. Both factors will lead to more pronounced distortion.[2,29]

Apart from altering the shrinkage and warpage behaviour of polyolefins, the nucleating effect of copper phthalocyanine green and other organic pigments is thought to also have a marked influence on the mechanical properties of polyolefins. An investigation of how certain organic pigments affect the mechanical properties of HDPE was undertaken by Lodeiro et al.[1] They found that tested pigments, copper phthalocyanine blue and irgalite yellow do affect the principal mechanical properties of HDPE. In particular, it was observed that the presence of small amounts of phthalocyanine blue in HDPE is sufficient to cause an increase in ductility, reduction in Young’s modulus (up to 10%), reduction in yield stress and increase in failure strain. They attributed these consequences to smaller and more numerous spherulites induced by the pigment; smaller spherulites in larger numbers, each surrounded by amorphous material, results in a polymer that will deform more readily and have lower yield stress and higher failure strain.

2.4. Nucleating Agents

2.4.1. Heterogeneous Nucleation of Polyethylene: Nucleating Agents Based on Potassium Stearate and Carboxylic Acid Salts

Nucleating agents have traditionally been added to semi-crystalline polymers to enhance processing and end product characteristics. The incorporation of these compounds results in shorter cycle time as they increase the crystallization rate of semi-crystalline polymers, ensuring faster solidification from the melt upon cooling. Their addition also results in the formation of smaller spherulites in semi-crystalline polymers. This change in spherulite size improves mechanical properties (such as tensile strength, hardness and modulus) and optical properties (such as haze and transparency).[8,35]

Polyethylene, and in particular high density polyethylene, has an extremely fast rate of crystallization, which makes it very hard to nucleate.[8,35] This is probably the reason why little has been published on its nucleating agents. That being said, a handful of nucleating agents have been identified to date. Together, the works of Solti et al. and Ge et al. showed that benzoic acid, talc and Na2CO3 can effectively nucleate polyethylene.[8] Besides the use of particulate or low molecular weight nucleating agents, polyethylene can also be nucleated by epitaxial crystallization on another polymer substrate. Loos et al. was able to demonstrate the melt crystallisation of LLDPE on oriented HDPE.[8]

Potassium stearate is another nucleating agent that has been shown to be effective for polyethylene. This compound is one of the two nucleating agent chemistries that are of interest in this project. Potassium stearate is an organic acid metal salt with molecular formula, C18H35KO2, and a chemical structure as illustrated in Figure 2.13.[36] It is prepared from starting materials potassium hydroxide (KOH) and stearic acid (CH2(CH2)16COOH).[37] In their research, Narh et al.[37] generated compelling data that clearly suggests that potassium stearate can effectively nucleate HDPE. They reported observing an increase in crystallisation temperature from 113.9°C to 116.9°C and a decrease in final spherulite radius from 47µm to 13.3µm when potassium stearate was incorporated into HDPE. In addition, they also noticed an increase in nucleation density from 1.94×104/cm3 to 8.56×105/cm3.

The second nucleating agent chemistry of interest in the project is that of carboxylic acid salts. A carboxylic acid salt is formed when a carboxylic acid reacts with a base.[38] The most common and well known nucleating agent of this chemistry is sodium benzoate, an aromatic carboxylic acid salt. Apart from sodium benzoate, salts of mono- or polycarboxylic acids such as sodium succinate, sodium glurate, sodium cinnamate, sodium capronate, sodium-4-methyl valerate, aluminium phenyl acetate are also known nucleating agents. The above-mentioned carboxylic acid salts are moderately effective in nucleating polyethylene.[39,40]

Horrocks and Kerscher[10] showed that nucleating agents from Milliken and Company based on dicarboxylic acid salts (Hyperform® HPN-68L and HPN-20E) are highly effective in nucleating polyethylene. They demonstrated that HPN-20E in particular exhibits superior nucleation performance. It was reported that the inclusion of HPN-20E reduced peak crystallisation temperature of LLDPE, mLLDPE and HDPE by 5.4°C, 4.4°C and 3°C respectively. This was accompanied by significant decrease in spherulite size and crystallisation half-time. For example, HPN-20E inclusion reduced crystallisation half-time of LLDPE from 57 minutes to 5.5 minutes. In addition, they also showed that HPN-20E incorporation has a positive impact on barrier properties, clarity and physical properties of LLDPE blown films. Hyperform® HPN-68L is of disodium bicycle [2.2.1] heptane-2,3-dicarboxylate chemistry as disclosed within U.S. Patent Number 6,465,551[42] while Hyperform® HPN-20E is of calcium hexahydrophthalic acid chemistry as disclosed within U.S. Patent Application Number 11/952,372.[43] The chemical structures of both compounds as disclosed within U.S. Patent Number 7,659,336 B2[41] are illustrated in Figure 2.14.

2.5. Solving Organic Pigment Induced Shrinkage and Warpage in Polyolefins

2.5.1. A Review of Possible Approaches

To date, various approaches have been proposed to overcome the problem of organic pigment induced shrinkage and warpage in polyolefins. They range from adjusting process parameters, mould design changes, pre-treatment of pigments, to incorporation of additional additives.

Bugnon et al.[31] outlined that unlike organic pigments, which have lipophilic and non-polar surfaces, inorganic pigments do not induce shrinkage and warpage in polyolefins because their hydrophilic and polar surfaces present no interactions with the polymers. Based on this argument, they investigated the feasibly of simulating an inorganic pigment surface on organic pigments by surface treatments in order to negate their negative nucleation effect. They found that precipitating a layer of metal oxide (e.g. zirconia, silica or alumina) or a layer of polar polymer (e.g. PVOH or cullulose derivatives) onto the surface of organic pigments will reduce their negative nucleating effect and the resultant warpage of pigmented HDPE. The improvement is especially evident for the latter method.

Tomlins et al.[44] experimented with the adjustments of injection moulding processing parameters in attempt to improve the dimensional stability of HDPE mouldings incorporated with organic pigments. Out of the five process parameters investigated (holding time, holding pressure, injection speed, melt temperature and mould temperature), they found that high holding pressure had the largest influence on shrinkage and warpage. High holding pressure improves dimensional stability. It was also observed that increasing the speed at which the mould cavity fills also reduces out-of-plane distortion. This can be achieved by reducing injection time, increasing melt temperature or increasing mould temperature. In addition, they noted that holding time does not have any influence on out-of-plane distortion. This is interesting as it is common practice to extend holding time to anneal out internal stresses in mouldings that may cause warpage.

Other additives in a pigmented polyolefin formulation may also override the negative nucleating effect of the organic pigments. In their study of the influence of organic pigments on HDPE mechanical properties, Lodeiro et al.[1] observed that the carrier (LLDPE) and/or the wetting agent in the organic pigment masterbatches they used serves to reduce the the negative nucleating effect of the pigments. The introduction of nucleating agents into pigmented polyolefin formulations may also overcome the unfavorable nucleating effect of the organic pigments. This is the approach adopted in this project and will be discussed in detail following this section.

Tomlins[30] suggested that small changes in mould design features may serve to provide more dimensional stability to moulded parts. He mentioned that the addition of stiffening ribs, correct placement of cooling channels, correct selection of gate with reference to part geometry and the use of rounded corners instead of square edge corners all assist in improving dimensional stability.

Although not a feasible solution, it is interesting to note that the work of Broda[27] indicated that organic pigments do not participate in the nucleation process of PP when a high degree of orientation is induced. He reasoned that under high levels of orientation, very effective row nuclei are formed in PP, and in the presence of such nuclei, heterogeneous nuclei formed from pigment crystals become insignificant and the crystallisation process only occur on row nuclei. In this case, crystallisation is also no longer spherulitic.

Comparing the different approaches, it is evident that some are more practical than others. Surface treatment of organic pigments is effective but there is an inherent problem associated with this approach; It is highly possible that new naked organic pigment surfaces will be formed (e.g. by means of shearing) when pigmented polyolefins are processed into final articles.[31] Changing of processing parameters is the simplest and most cost effective method, but it is by the authors own admission that because organic pigments promote very pronounced anisotropic shrinkage, totally eliminating distortions by just adjusting process parameters is very difficult.[44] Furthermore,, when a variety of pigments are used, it is to practical to change a new set of processing parameters each time the pigment is varied. Changing of mould design to compensate for shrinkage and warpage is not always possible while inducing high levels of orientation is not at all feasible. It would be ideal if the negative nucleating effect of organic pigments can be negated just by adding nucleating agents.

2.5.2. Eliminating Shrinkage and Warpage by Incorporation of Nucleating Agents

To date, some work has been done to show that the incorporation of strong nucleating agents can solve or at least reduce pigment induced shrinkage and warpage. These investigations have primarily been conducted on polypropylene.

In a study that closely relates to the present research, Tomlins et al.[45] investigated the influence of nucleating agents on the dimensional stability of pigmented PP mouldings. Although the nucleating agents and pigments examined in their study were not revealed due to commercial sensitivity, several important conclusions can be drawn from their findings. Their work showed that addition of nucleating agents can substantially reduce in-plane warpage in pigmented PP and the effect becomes more significant with increasing nucleating agent concentration. Reduction in out-of-plane warpage is not as significant. It was also demonstrated that difference in anisotropic shrinkage ratio (ratio between flow direction and transverse direction shrinkage) that exists between PP mouldings containing different pigments can be reduced by nucleating agent incorporation. This reduction in difference is however, insensitive to nucleating agent concentration but depends more on nucleating agent type.

Many commercial nucleating agents available in the market have been alleged to solve the shrinkage and warpage problem caused by organic pigments. Halstead and Jones[4] showed that the carboxylic acid salt based nucleating agent from Milliken and Company, Hyperform® HPN-68L, promotes relatively isotropic shrinkage in PP mouldings as compared to other nucleating agents such as sodium benzoate.[4] They attributed this to the plate-like particle shape of HPN-68L having no preferred flow orientation whereas in the case of sodium benzoate, ruler-shaped particles orientate in the direction of flow (illustrated in Figure 2.15). Their work demonstrated that when HPN-68L is incorporated into pigmented PP mouldings, the isotropic shrinkage behaviour associated with this nucleating agent will still persist while the warpage behaviour associated with the incorporated organic pigment will essentially be cancelled out. It was also indicated that incorporation of HPN-68L would serve to ‘level out’ the difference in anisotropic shrinkage ratio and crystallisation temperature that exists between PP mouldings containing different organic pigments. These findings clearly suggest that HPN-68L has nucleation power that overrides the nucleating ability of organic pigments. This overriding power allows for the production of mouldings with identical dimensions using identical processing conditions even though they contain different organic pigments.

Apart from Milliken and Company, BASF[46], Borealis[47] and Ampacet[48] also claim that their respective nucleating agents have the ability to override the negative nucleating effect of organic pigments. The nucleating agent from BASF is of zinc monoglycerolate chemistry (tradenamed Irgastab® NA 287)[49] while that from Borealis is made from a special reactor technique where catalyst is pre-polymerised with monomers (tradenamed BNT)[50]. The nucleating agent from Ampacet can be found in materbatches Ampacet 103003 and Ampacet 103004 but its chemistry is undisclosed. Similar to Milliken and Company, Ampacet also mentioned that the plate-like particle shape of their nucleating agent may be the reason why it can provide better dimensional stability to pigmented mouldings.[48]

In 2006 and 2010, Milliken and Company filed for two patents on nucleating agent blends; U.S. Patent Application Number 11/078,003[51] and U.S. Patent Number 7,659,336 B2[41] respectively. The former is a blend of Hyperform® HPN-68L with a phosphate ester salt based nucleating agent, NA-11 (from Asahi Denka Kogyo K.K.), while the latter involves a blend of Hyperform® HPN-68L and HPN-20E.[41,51] It was reported in both cases that a blend of two nucleating agents does not result in one compound dominating or overriding the nucleating effect of the other. Instead, beyond the expectations of the authors, the nucleating effect of co-nucleants both contributed to resultant physical properties of the polymer and even yielded synergistic effects in some cases.[41,51] Two important conclusions can be interpreted from the patents. Firstly, it has been discussed that for nucleating agents to eliminate shrinkage or warpage in pigmented polyolefins, they are required to override the nucleating effects of the pigments. However, it is shown in these two patents that even a strong nucleator such as Hyperform® HPN-68L does not necessarily always override other nucleating agents. In other words, there is a possibility that Hyperform® HPN-68L may not be able to negate the negative nucleating effects of all organic pigments. Secondly, it is also shown in these two patents that the use of two nucleating agents may yield synergistic properties and this could be applied to the present research. A blend of potassium stearate and a carboxylic acid salt could be tested to investigate if it imparts synergistic improvements to dimensional stability.

2.6. Characterisation Techniques

2.6.1. Differential Scanning Calorimetry (DSC)

Differential scanning calorimetry, DSC, measures the heat flow into or out of a polymer specimen as a function of either time or temperature. Intergration of peaks in a DSC trace (heat flow vs. time or temperature) gives the enthalpy change in a specimen. When there is heat flow into the specimen, the enthalpy change is endothermic and when the specimen releases heat, the enthalpy change is exothermic.[52]

The two main types of commercial DSC instruments are the ‘heat-flux’ and ‘power compensated’ types. A ‘power compensated’ DSC instrument measures the difference in power supplied to a polymer sample and a reference, in order to keep their temperatures the same (illustrated in Figure 2.16a).[53] In a ‘heat flux’ DSC instrument, one single heater is used to increase the temperature of both the sample and reference cell (illustrated in Figure 2.16b). Temperature difference between sample and reference pan occurring due to exothermic and endothermic effects in the polymer are recorded.[53]

1. ‘heat flux’ DSC instrument[54],
2. ‘power compensated’ DSC instrument[55]

A review of literature involving the use of differential scanning calorimetry in polymer crystallisation studies revealed that the value of this technique lies in its ability to quantify various aspects of the process.

Non-isothermal differential scanning calorimetry (constant cooling rate) can yield the following information:

• Onset crystallisation temperature (Tc, onset)[66]
• Peak crystallisation temperature (Tc, peak)[66]
• Percentage crystallinity[56], which is given by,

% Crystallinity = (ΔHf° / ΔHf) x 100%

Where ΔHf is measured heat of fusion and ΔHf° is heat of fusion of the polymer when 100% crystalline

• Ozawa exponent (m) and rate constant (ĸ(t)) by using equation (2-5)

Isothermal differential scanning calorimetry (constant crystallisation temperature) can yield the following information:

• Crystallisation half-time (t1/2)[57], which is the time taken for 50% of total crystallisation to occur
• Rate of crystallisation[8], which is given by,

Rate of crystallisation = 1/t1/2

• Avrami exponent (n) and rate constant (K) by using equation (2-3)
• Tomlins and Richardson[57] showed that kinetic parameters from the Avrami equation, n and K, can be further used for calculation of activation energy for crystallisation and spherulitic growth rate.

2.6.2. Polarised Light Optical Microscopy (PLOM)

Polarised light optical microscopy, is the most widely used method to characterise the supermolecular structure (e.g. spherulites) of semi-crystalline polymers.[58] The popularity of this characterisation technique can be attributed to ease of use, simple sample preparation and low cost.[59]

As illustrated in Figure 2.17, a polarised light optical microscope adds two polarising filters to an ordinary optical microscope.[60] These polarising filters cause light that passes through it to vibrate in only one plane.[61] The first is located below the microscope stage to polarise light supplied by a light source (e.g. halogen or arc lamp) and the second serves to analyse the polarisation of light after it passes through a specimen (therefore it is also known as an ‘analyser’). Generally, the polarisers are oriented such that their polarisation directions are at right angles and no light passes through the system. However, when a birefringent or optically anisotropic specimen is placed on the sample stage and rotated through 360 degrees, they will be illuminated at some angles during the rotation. For an optically isotropic specimen, the field of view will remain dark at all angles of rotation.[60]

Spherulites can be view using a polarising light optical microscope because they are ‘spherically birefringent’ objects with two unique refractive indices; radial (nr) and tangential (nθ). The difference in these two indices gives rise to the ‘Maltese cross’ pattern as illustrated in Figure 2.18. Polyethylene spherulites have larger refractive index in the tangential direction nr< nθ; negative spherulites.[62]

polarised light optical microscope[63]

A survey of literature involving the use of polarised light microscopy in polymer crystallisation studies showed that this technique can yield the following valuable information[2,64,65]:

• Spherulite size, nucleation density and overall spherulitic texture
• The entire crystallisation process can be followed and onset crystallisation temperature can be determined with the aid of a hot stage and a video camera mounted on the microscope.
• Also with the aid of a hot stage and mounted video camera, micrographs can be taken at fixed time periods to measure radius of spherulites (R). Measured radius can be used to determine spherulitic growth rate (dR/dt).[64]
• Krumme[65] showed that by using binary conversion and pixel counting, it is possible to calculate percent crystallinity, nucleation density, Avrami exponent (n) and Avrami rate constant (K) from a micrograph taken at a specific time (t). Binary conversion essentially involves converting amorphous regions in a micrograph into black colour and crystalline regions into white colour so that their areas can be used for further calculations.

1. normal light,
2. polarised light,
3. after binary conversion[65]

To investigate spherulitic texture using polarised light microscopy, polymer specimens must be in the form of thin films. If thermal history of the polymer is of importance, thin film specimens can be prepared by microtomy.[59] If thermal history is not important, thin films can be obtained either by solvent casting[66] (dissolve polymer in a solvent and place a drop of the solution on a glass slide and allow solvent to evaporate) or by melt pressing[2] (between microscope glass slide and cover-slip at elevated temperatures).

Although there are many advantages in using polarised light microscopy, this technique does have its disadvantages:

• In general, only spherulites with diameters above 5µm can be resolved and use for calculations using this technique. Smaller spherulites, such as that produced by the addition of nucleating agents, may not be clearly resolved and required the use of other techniques such as electron microscopy or small angle light scattering.[62]
• Cooling rate of commercial hot-stages (max 20°C/min) cannot prevent crystallisation from starting before isothermal crystallisation temperature is reached.[65]

2.6.3. Shrinkage Isotropy Measurements

Commonly used standards to measure shrinkage from mould cavity to moulded dimensions of thermoplastics include ASTM D955 and ISO 294.[30] In this research, the former standard will be adopted.

ASTM D955[67] describes three different geometries that are applicable for shrinkage measurements; Type D2 – 60x60x2mm square plaque, Type A – 12.7x127x3.2 rectangular bars and Type B – disc-shaped specimens with 100mm diameter and 3.2mm thickness. It is mentioned that Type A specimens are more applicable when shrinkage in machine direction is expected while Type D2 specimens are more applicable when shrinkage is expected in both machine and transverse directions.

According to the standard, mould shrinkage is calculated and reported for both the machine direction, MD, and the transverse direction, TD:

Percent mould shrinkage in MD

= (mould dimension in MD – specimen dimension in MD) x 100%

mould dimension in MD

Percent mould shrinkage in TD

= (mould dimension in TD – specimen dimension in TD) x 100%

mould dimension in TD

In their work, Halstead & Jones[4] and Koh[33] both used a shrinkage isotropy ratio to characterise the differential shrinkage in moulded specimens:

Shrinkage isotropy = percent mould shrinkage in MD

percent mould shrinkage in TD

A shrinkage isotropy ratio of 1 would indicate uniform shrinkage. As the value moves further away from 1, it becomes more likely that the moulded specimen will warp.[4]