Search

The autonomous ordering and assembly of atoms and molecules on atomically well-defined surfaces combines ease of fabrication with exquisite control over the shape, composition and mesoscale organization of the surface structures formed. Once the mechanisms controlling the self-ordering phenomena are fully understood, the self-assembly and growth processes can be steered to create a wide range of nanostructures with exotic and desirable properties, synthesised from the bottom-up, on an industrial scale from metallic, semiconducting and molecular materials. The work of this thesis aims to address questions concerning molecular self-assembly on graphene. Firstly, techniques for fabricating graphene membranes for electron microscopy (EM) are outlined. The complete fabrication process is described, beginning with the growth of CVD graphene, followed by the transfer of graphene from chemical vapour deposition (CVD) foils to a transmission electron microscope (TEM) support, and finishing with the cleaning steps involved to produce pristene regions of graphene. Strategies to chemically functionalise graphene through covalent and non-covalent means are detailed, as well as methods to fabricate more specialised graphene TEM membranes consisting of stacked and sandwiched graphene layers. With the methods used to fabricate and modify graphene EM membranes described, attention is next focused on specific microscopy techniques developed in order to study organic materials that readily damage when exposed to the electron beam in an electron microscope. Strategies to mitigate the damage arising due to beam exposure are investigated for a range of different organic molecules, and the effects of using a range of detection devices are also studied. Next, the growth of two very similar overlayer systems on graphene are studied. Trimesic acid (TMA) and terephthalic acid (TPA) thin films are grown on both freestanding and CVD graphene substrates for a range of thicknesses, and the resulting structures are probed using a range of microscopy techniques. For TMA, van der Waals epitaxy results in two preferred orientations of the assembly structure that grows in a layer-by-layer Frank-van der Merwe fashion, up to a height of ≈ 20 nm. In stark contrast, TPA assembles into a 2D monolayer before rapidly transitioning to its bulk-like structure as further layers are deposited, following a layer-plus-island, or Stranski-Krastanov, growth mode. Continuing the investigations into the structure of self-assembling molecular films on graphene, a pair of porphyrin-based molecules of the octaethyl porphyrin (OEP) class are studied. A monolayer film of OEP molecules is deposited either side of a freestanding graphene membrane, and the resulting assembly structure is driven by a remote interaction across the graphene between the two OEP films. The remote interaction is shown to diminish on the length scale of two graphene layers. Finally, the structure and motion of individual metal nanoclusters (M-NCs) deposited on freestanding graphene is studied using high-resolution TEM. Computational routines involving cross-correlation techniques are developed in order to better study the dynamic behaviour of M-NCs in atomically-resolved time-series image sequences. The strategies developed provide a means for accurately studying more complex systems, structural changes, and chemical reactions at atomic-resolution and in real-time.

Graphene — a two-dimensional sheet of carbon atoms — has surged into recent interest with its host of remarkable properties and its ultimate thinness. However, graphene combined with other materials is starting to attract more attention. These heterostructures can be important for production routes, incorporating graphene into existing technologies, or for modifying its intrinsic properties. This thesis aims to examine the role of van der Waals epitaxy within these heterostructures. First, the graphene-copper interaction during chemical vapour deposition of graphene is investigated. Graphene is found to grow with a mismatch epitaxy of 8 relative to the [001] direction of the Cu(100) surface, despite a mismatch in symmetry and lattice parameter between two. Further, the electronic structure of both graphene and copper is unchanged by the interaction. This highlights the weak interaction between the two, owing to its van der Waals nature. Functionalised graphene is another important heterostructure, and is intensively studied for both graphene production routes and for altering graphene’s properties. Here, it is the change to the homogeneous graphene surface that makes it interesting for van der Waals epitaxy. The effect of functionalisation of graphene with atomic oxygen and nitrogen is presented next. In both cases, only small amounts of functionalisation ( 5 at%) is sufficient to significantly deteriorate the -band structure of the graphene through localisation. For small amounts of nitrogen functionalisation, and greater amounts of oxygen functionalisation, extended topological defects are formed in the graphene lattice. Unlike epoxide oxygen groups, these disruptions to the pristine graphene are found to be irreversible by annealing. Next, the interaction between graphene and the organic semiconducting molecule vanadyl-phthalocyanine (VOPc) is presented. As a result of the van der Waals nature of the graphene surface, VOPc molecules can form crystals microns in size when deposited onto a substrate with an elevated temperature of 155 C; at ambient temperatures, the crystals are only tens of nanometres across. In contrast, the functionalised graphene oxide surface prevents large crystal growth, even at elevated temperatures, because surface functionalities inhibit molecule diffusion. This highlights the importance of graphene as a substrate for molecular crystal growth, even when the growth is not epitaxial. Finally, the supramolecular assembly of trimesic acid (TMA) and terephthalic acid (TPA) is presented. Despite their chemical similarity they display different behaviour as they transition from monolayers to three-dimensional structures: for TMA, the epitaxial chicken wire structure seen at a monolayer templates up through the layers as molecules stack, until a thickness of 20 nm, when random in-plane orientations appear; on the other hand, TPA forms a brickwork structure at the monolayer, which quickly transitions to fibre-like crystals with a bulk structure for the thin films. However, the TPA orientation is still determined by the epitaxy with the graphene substrate, although this is significantly weaker than for TMA.

Many of the highlights in the surface science field in the last years have arisen from the study of organic molecules adsorbed upon, but physically decoupled from, metallic substrates. When adsorbed on bare surfaces like Cu, the strong interaction between the two can have a number of different effects on the chemistry, functionality and assembly of the adsorbates. These effects can be reduced or even prevented by incorporating an ultra-thin insulating film between the molecule and the metal, enabling the retention of the desired molecular properties. Previous studies have almost exclusively used films of inorganic material, such as alkali halides and metal oxides. In this work, an all-organic alternative approach has been developed, which should allow a greater degree of control over the nature of the decoupling film and its corresponding interactions. To demonstrate the potential impact of the molecule-substrate interaction on the formation of molecular nanostructures, the adsorption and assembly of a prototypical molecular tecton, terephthalic acid (TPA), has been explored on Cu(110) and Cu(111). On Cu(110), the TPA is deprotonated and goes on to form a range of metal-organic or all-organic monolayer films. The exact structure obtained is determined by the total molecular coverage, which is rationalised by consideration of the TPA adsorption energies and the molecular density in each phase. On Cu(111), TPA adsorbs intact, and assembles into a brickwork-like structure characterised by both uncharacteristically short and elongated hydrogen bonds. This unusual assembly is the result of the oft-cited, but in general poorly understood, ‘interplay of intermolecular and molecule-substrate interactions’. In a bid to understand the observed supramolecular assembly of TPA on Cu(111), a simple algorithm has been developed that is capable of predicting both its quantitative and qualitative aspects. Moreover, this analytical model can be readily expanded to more complex assemblies, such as those using more complex molecules and surfaces. In addition, the formation of Cu- and Fe-terephthalate complexes has been explored. Oxalic acid (OA) monolayers have been investigated as prototype organic decoupling layers for the Cu(110) and Cu(111) surfaces. This molecule was chosen as, in previous work, it had been reported to adsorb in an upright orientation and assemble into a densely packed monolayer on Cu(110). In this work, a flat-lying, low density film has been obtained on Cu(110), which later proved to be incapable of supporting molecular ‘overlayers’unless it was cooled to low temperature prior to the deposition of molecular overlayers. In contrast, a densely-packed, upright monolayer was obtained on Cu(111). This decoupling layers was effective even at room temperature, and moreover appears to plays a significant role in determining the assembly of the overlayer species. These studies indicate the potential of vertically standing organic films as both decoupling layers and also as a tool with which to directly modify and control the assembly of organic nanostructures.

Two bio-based comonomers, isosorbide (ISB) and 2,5-Furan dicarboxylic acid (FDCA) was used to partially or completely replace the diol, Ethylene glycol (EG) or diacid, Terephthalic acid (TPA) of petroleum-based polyester, Polyethylene terephthalate (PET). The effect of the co-monomers on the alkaline hydrolysis reaction of bio-based co-polyesters, PEIxT and PETFy or polyester, Polyethylene furanoate (PEF) was investigated to compare the rate of depolymerization for their reaction kinetics and the factors contributing to the reaction mechanism were studied.Different compositions of ISB (3%, 8% and 12%) was used to replace the EG in PET to produce the PEIxT co-polyesters and different composition of FDCA (10% and 100%) was used to replace the TPA to produce the PETFy co-polyesters. The 100% replacement of TPA with FDCA, produced the furan analogous of PET, PEF. The alkaline hydrolysis reaction kinetics were studied at temperatures between 130oC-150oC in 1.1 M Sodium Hydroxide (NaOH) solution in a glass reactor system and a Parr reactor system. The comonomers exhibited increase in the reaction rates under alkaline hydrolysis relative to the pure PET for all systems.Since alkaline hydrolysis occurs in the amorphous region near the surface of the polymers, the crystallinity, surface wetting and water uptake properties of the polymers have an impact on the hydrolysis reaction kinetics. To understand the cause for the improved reaction kinetics with inclusion of the co-monomers, the impact of comonomers on the physical properties such as crystallinity, surface wetting and water uptake were studied. The reactive bonds of the co-monomers can have a significant impact on the ester linkage in the co-polyesters that make them susceptible to hydrolytic degradation. Literature on the hydrolytic degradation in the environment suggests that the ester-linkage of isosorbide with TPA might be more susceptible to hydrolysis than the linkage of EG with TPA, which could be a factor in facilitating the hydrolysis reaction at lower temperatures. Similarly, the FDCA-EG linkage could be more reactive under hydrolysis than the EG-TPA bond. To compare the reactivity of the ester linkage of the oxygen-based furan rings of FDCA to the aromatic benzene ring of TPA, the reaction kinetics of small molecule analogues, Bis (2-hydroxyethyl) terephthalate (BHET) for PET and Bis-hydroxyethyl 2,5-Furandicarboxylate (BHEF) for PEF were compared. To determine the reaction rate constants of the PEIxT co-polyesters, the conversion rates at 130°C were fit into an established first order kinetic model. Because of the higher reaction rates of the PETF10 and PEF at the experimental temperatures studied (130°C -150°C), the kinetics of the samples could not be captured. Fourier Transform Infrared Spectroscopy (FTIR) was used to analyze the precipitate following filtration of the co-polymers from product of the alkaline hydrolysis reactions. FTIR data confirmed the presence of functional groups consistent with monomers present in the co-polyesters qualitatively. Proton (1H) and Carbon (13C) Nuclear Magnetic Resonance (NMR) was also performed on the co-polyester samples to quantify the monomers present in the samples. A recovery method for the co-monomers, TPA and FDCA based on selective solubility of TPA and FDCA in water and dimethyl sulfoxide (DMSO) co-solvent system was investigated. The solubility of TPA and FDCA in water is limited, but in DMSO, TPA has a19 wt% solubility and FDCA has a 55 wt% solubility at room temperature. Preliminary work demonstrated that using the difference in their solubility in the DMSO solvent, TPA can be selectively precipitated from the TPA/FDCA mixture in DMSO using water as an anti-solvent, however, further studied need to be done for the recovery of FDCA from the FDCA/DMSO solution. Since, isosorbide is readily soluble in water, the isosorbide following the alkaline hydrolysis of PEIxT co-polyesters were expected to be washed off with water in the EG/water residue, during washing of the recovered product. Literature suggests a distillation and cold-crystallization method can be used as an energy optimized isosorbide recovery method from a EG/water solution.Bio-based co-polyesters exhibited higher reaction rates compared to PET, while PEF exhibited the highest reaction rate at all temperatures. The impact of the co-polyesters on crystallinity, surface wetting, water uptake and the bond reactivity could be facilitating the higher reaction rates. TPA following the alkaline hydrolysis of PETFy can be recovered using the selective solubility of TPA and FDCA in a co-solvent of DMSO and water. ISB can be recovered from the residue solution of EG and water using a distillation and cold crystallization method.

Reversed-phase liquid chromatography (RPLC) is by far the most important separation technique in analytical chemistry. It is an essential separation technique that has been extensively used in science and industry fields. However, this technique has some drawbacks such as the harmful environmental impact due to the thousands of gallons of toxic organic solvents used as a mobile phase, which usually end up as waste. An RPLC developed method used methanol and acetonitrile as organic solvents and compared the efficiency and selectivity of these solvents using the same C18 column to separate terephthalic acid (TPA) and eight of its contaminants. Methanol showed a different selectivity and faster separation than acetonitrile. The developed method was successfully used to analyze two different industrial samples.Using surfactants in the RPLC mobile phase (MP) has many advantages such as decreasing the environmental impact, lower cost, higher efficiency in separating structural isomers, and, most importantly, the ability to inject samples directly with minimal treatment. Tween 20 was used for what we believe is the first time to dynamically coat a C18 column, which was used to separate TPA impurities. The developed method is simple, fast, green and showed excellent coated column stability. The analysis took only 19 minutes to achieve baseline separation. This method was successfully applied to measure TPA contaminants in industrial samples.There is little published literature reporting the use of ultra-high performance micellar liquid chromatography (MLC); it seems the surfactants Tween 20 and Tween 40 have never been used in this mode as mobile phase components. Cinnamic acid derivatives are well known for their antioxidant properties, and they are present in food, beverages, pharmaceuticals, and supplements. A comparison study included separation of nine cinnamic acids (CA) derivatives using Tween 20 and Tween 40 in the ultra-high-performance MLC mode. In general, there are no significant chromatographic behavior differences between the two surfactants; however, Tween 20 shows faster and better peak shape than Tween 40. MLC using Tween 20 was used successfully to analyze red wine, spiked and unspiked, for seven of the cinnamic acids; the results showed excellent recoveries for the spiked sample.To the best of our knowledge, amino acid-based surfactants have not been used for as mobile phase or stationary phase modifier for HPLC. Changing the separation mode of a C18 column from pure RPLC to ion exclusion liquid chromatography (IELC) was achieved using an amino acid surfactant lauroyl-L-glutamate (LLG). The LLG is a negatively charged surfactant, and this property was exploited to separate ten short-chain aliphatic carboxylic acids successfully. Our novel approach showed an excellent peak shape and fast analysis. The IELC modified column was used to analyze ten different juices for their aliphatic acid contents.

Owing to the overdependence on fossil fuels, global effort has been initiated to identify alternate renewable sources for PET Polyethylene terephthalate (PET), a polyester of terephthalic acid (TPA) and ethylene glycol (EG). As these efforts to make PET from renewably sourced TPA and EG make inroads, alternate renewably sourced comonomers need to be identified to modulate its properties depending on the needs of the application. The goal of this work was to develop polyesters of PET with two commercially viable, renewably sourced comonomers, namely, 2,5-furandicarboxylic acid (FDCA) and isosorbide (ISB), and establish a fundamental understanding of the structure-property relationship for these polyesters.In the first part of this work, effect of copolymerization of PET with FDCA (described as PETF copolyesters) on the esterification kinetics was evaluated in detail. The presence of FDCA and esterified FDCA products during copolymerization enhanced the solubility of TPA in the reaction media and improved the esterification kinetics. The kinetics study revealed that the esterification step of PETF copolyesters can be performed at shorter times or lower temperatures to gain energy savings. Following this study, high molecular weight PETF copolyesters were synthesized and processed into continuous films. In the unoriented state, incorporation of furan units in the PET backbone reduced the crystallization rate and enhanced mechanical and barrier performance of PET. Improvement in the barrier properties was primarily attributed to the suppressed chain segment mobility due to the presence of rigid, non collinear furan ring. The PETF copolyester films were biaxially stretched to evaluate the effect of furan units on the orientation and strain induced crystallization of PET. Differential scanning calorimetric studies revealed that the furan units resulted in decreased strain induced crystallization upon orientation. However, even with lower crystallinity values, PETF copolyesters showed similar enhancement in properties as seen in the unoriented state. Analysis of amorphous phases suggested that the rigid amorphous phase trapped between the lamellae and chain folds increased by incorporation of furan units. It was proposed that this rigid amorphous fraction compensated for the loss of crystallinity upon orientation and retained the property enhancement for biaxially stretched PETF samples. Lastly, the framework developed for structure-property evaluation of PETF copolyesters was extended to the copolyesters of PET with isosorbide (PEIT copolyesters) and terpolyesters of PET with ISB and FDCA (PEITF terpolyesters). Incorporation of bulky isosorbide units improved the glass transition temperature significantly. However, mechanical and barrier performance of PET was negatively affected due the increased brittleness because of the isosorbide units. The PEITF terpolyesters exhibited intermediate properties between PETF and PEIT copolyesters. This work, which focused on understanding the fundamentals of reaction chemistry and structure property relationship for the copolyesters and terpolyesters of PET, can therefore be envisioned to provide a framework to evaluate the effect of potential comonomers on the synthesis and properties of PET relevant for packaging.

The large increase in the generation of post-consumer plastic in past few decades has led to an increased interest in eco-friendly recycling technologies. Polyethylene terephthalate (PET) is a highly valued packaging material with broad applications because it is strong, lightweight, non-reactive, non-toxic and shatterproof. To extend its applications, the packaging industry adds co-monomers, additives, multilayered structures and forms polymer blends to improve the mechanical and barrier properties of the base polyester. These additives can pose challenges to the mechanical recycling methods that are commonly used in the industry. While mechanical recycling is economical and broadly commercially used, the recycled PET (RPET) tends to have reduced molecular weight and can degrade in the presence of impurities (i.e. polyvinyl chloride (PVC)). Chemical recycling is an attractive alternative approach that results in recovery of monomers and other chemical constituents that can be used as precursors for new polymers. Several chemical recycling methods were reported in literature to address the end-of-life PET waste, but little work was done on co-polyesters that are of interest to the packaging industry. The focus of this thesis is to investigate alkaline hydrolysis of traditional PET and a co-polyester (will be referred to as PETF20) containing ethylene glycol, 80% terephthalic acid (TPA) and 20% 2,5-furan dicarboxylic acid (FDCA). Studies on chemical/mechanical recycling of PETF20 were not reported in the literature. Alkaline hydrolysis of PET and PETF20 was investigated at atmospheric pressure and a range of temperatures (= 150¿) using sodium hydroxide solution (1.1 M) to recover TPA and FDCA. The impact of time, temperature, co-solvent (i.e. ¿- Valero lactone) and impurity (i.e. PVC) on conversion of PET was investigated at = 150¿, rate of depolymerization and impact of co-solvents (¿-Valero lactone, ¿- butyral lactone, ethylene glycol diacetate, propylene glycol diacetate and triglycerol) on PET and PETF20 were studied at 90¿. The chemical structure of the products was confirmed via FTIR and NMR. The conversion of PET obtained at 150¿ and 180 min with and without impurity (PVC) is approximately 81%. PETF20 flakes exhibited high conversions of 88% compared to PET i.e. 42% at 90¿. Addition of triglycerol to PET flakes resulted in high TPA yields of 63% while the other co-solvents resulted in either lower or same yields as that of base NaOH solution. PETF20 with and without co-solvents resulted in the same yields. Research was extended to separate TPA from FDCA using precipitation, 20 wt.% water in DMSO solution exhibited promising results with 68.3% recovery of diacid from 20:80 molar fraction of TPA and FDCA.

Polyethylene Terephthalate (PET) is an industrially important polyester due to its properties. Because of PET’s resistance to degradation in nature, mechanical recycling have been in practice. The Recycled PET(R-PET) has a lower Molecular Weight (MW) and Intrinsic Viscosity (IV) than virgin PET because of degradation of PET during the recycling process. With increasing demand for high performance PET in packaging industry several oxygen barrier additives have been introduced to PET. Polyamide MXD6 is one such additive, which is commercially available, and, the effect of it on the recycling process and recycled product properties are not well established. The purpose of this study was to determine the impact of MXD6 on the structure and properties of mechanically recycled PET blends.A series of pure PET, pure MXD6 and PET/MXD6 (95/5 wt.%) blend samples were prepared and recycled up to 3 times using a twin screw extruder to simulate multiple recycling process in the lab scale. Physical and structural properties of the recycled products were characterized using different methods such as DSC, FTIR, TGA, Solution and Melt Viscosity. The results showed that adding MXD6 results in lower IV of the R-Blend compared to the R-PET (37% comparing to 23% after 3 recycling step respectively). Moreover isothermal and dynamic crystallization kinetic studies showed that MXD6 increases the crystallization kinetics of the blend. In addition, the properties of MXD6 were studied following the mechanical recycling process and they changed differently than PET and Blend properties due to branching/crosslinking of MXD6 due to degradation during processing.Owing to the importance of color generation problem in all the recycled products as well as PET/MXD6 blend in particular; as a second part of the project, a novel method to reduce the color generation in the PET/MXD6 blend during recycling process has been developed and applied. Based on this method, the color generation can be decreased by an interchange reaction between PET and MXD6 chains that leads to elimination of amine end groups. Different amounts of Terephthalic Acid (TPA) and Sodium para toluene sulfonate were added to enhance the interchange reaction. Colorimetry and NMR results showed that the addition of 0.5% TPA was more successful in reducing the yellow color as the result of lower conjugated imine generation. NMR results indicated that free TPA scavenged MXD6 amine end groups during the recycling process. The method reduced the Yellowness Index of the blend by 20% in the blending step and 40% after 2 times recycling while other properties including IV and thermal transitions remained unchanged.

X-ray diffraction methods have been used to investigate the solid-state structure of wholly aromatic main chain liquid crystalline copoly(ester carbonates) and copolyesters incorporating kinked units. The second chapter deals with the copolymers of p-hydroxy benzoic acid (HBA), hydroquinone (HQ) and carbonate units (CO), of HBA, HQ and isophthalic acid (IPA) and of HBA, resorcinol (RC) and terephthalic acid (TPA). The x-ray fiber diagrams for the three systems show that the copoly(ester carbonates) have the highest degree of order. Chapters 3 through 5 deal primarily with the copoly(ester carbonate) of 75/25/25 HBA/HQ/CO, which has been studied in more detail. As a first step the one-dimensional structure was determined for the chain direction. The x-ray fiber diagram show the occurrence of non-periodic meridional peaks with d-spacings of 6.47, 3.22, 2.15 and 2.05A, which can be reproduced by calculating the scattering of chains with a random sequence distribution of the monomers along the chain axis. In a second step the three-dimensional structure was determined. It is found that a pseudo-dimer must be assumed in order to generate all the observed layer lines. In the as-spun fibers, the chains are packed in an orthorhombic unit cell with the base plane dimensions a = 5.0A and b = 8.9A. T he discrete equatorial and off-equatorial Bragg maxima arise as a consequence of packing of the chains with a very limited off-register between adjacent chains. Register in the arrays of chains is achieved when the linkage units, ester or carbonate, lie in a specific plane perpendicular to the chain axis. The last chapter deals with the modelling of discrete systems using the SYBYL°ler and CERIUS software packages. The carbonate units were found to have all possible values of distortion between cis-trans, trans-trans and trans-cis, with a 60° maximum deviation from chain direction. (Abstract shortened with permission of author.)

The solid state structures of some wholly aromatic thermotropic copolyesters have been investigated by x-ray diffraction and molecular modeling techniques. The copolymers were prepared from p-hydroxybenzoic acid (HBA), biphenol (BP), hydroquinone (HQ) and terephthalic acid (TPA): in one copolymer the TPA was unsubstituted and in others, it was substituted with thiophenyl or oxyphenyl side groups (STPA or OTPA). The fiber diagrams of the unsubstituted copoly(TPA/BP/HBA) shows a series of sharp intensity maxima along the fiber axis direction and on layer lines. The meridional d-spacing are aperiodic suggesting that the copolymer chains have random monomer sequence and highly extended conformation. In chapter 2 we have analyzed the packing of non-identical sequences on the X-ray fiber diagrams of the BP series of substituted copolymers show a number of aperiodic intensity maxima along the fiber axis direction. The positions of these intensity maxima are similar to those for the unsubstituted system and have been reproduced by a method developed in this laboratory and applied to various copolymer systems. This method calculates the scattered intensity along the chain axis direction for a single aperiodic chain and averages over all possible sequences. A very good agreement between the calculated and observed data has been obtained for a structure consisting of chains with random monomer sequence and highly extended conformations. The effect of side groups on chain conformation has been analyzed by introducing different levels of non-linearity in the copolymer chains. The chains with substituents are about 1-2% more sinuous than those without X-ray data for the HQ series of substituted copolyesters contain periodic d-spacings along the fiber axis direction. This periodicity arises because the copolymer backbone has alternating phenylene and ester groups where the later may be COO or OCO experimentally observed lattice. Many arrays have been modeled to study different packing strategies. Using the molecular mechanics calculations, the potential energy of the starting arrays has been minimized by different methods. X-ray diffraction patterns have been simulated for the initial and optimized arrays. Almost all the observed Bragg reflections have been reproduced for the models in which chains were staggered along the chain axis to introduce registration between neighboring segments a nd then packed in herringbone fashion. Hence, an effective monomer repeat is observed. The calculated intensity maxima considering the random monomer sequence along the chain for these copolyesters is in excellent agreement with the observed data. Based on the results for the analogous BP series of substituted copolyesters. we believe that the HQ copolyesters also have random monomer sequence

An algorithm for modeling the continuous melt-phase production of linear polyesters is presented with results. The purpose is to determine optimal plant designs taking into account plant cost, product quality, and flexible operation. Using a sensitivity analysis, designs can be developed with a low estimated variance in product quality. The novel methodology used allows optimal designs to incorporate a specified degree of flexibility. Designs are developed to incorporate the desired degree of process flexibility at the lowest cost. Multiple reactors are supported based on two reactor archetypes, a continuous-stirred-tank reactor and a disk-ring reactor. Variables include equipment sizes, monomer feed rates, and reactor operating conditions. Areas investigated include the effects of production rate, degree of polymerization, degradation rates, catalyst rates, rates of mass-transfer, product variability, and others. The algorithm was based on information available from open literature. Development was based on coupling a smooth objective function with a gradient-based optimization routine. Process flexibility was incorporated into the model using two classes of variables, operating and design, to enable multiple operating regimes for a plant design. Data was gathered for polyethylene terephthalate (PET) and polybutylene terephthalate (PBT). Optimization results are primarily reported for PET with a terephthalic acid (TPA) feedstock. Additional results have been obtained for PBT with TPA and PET with dimethyl terephthalate (DMT). The algorithm successfully finds optimal designs for a variety of production levels, product specifications, and product-quality- variability levels. The most pertinent results include the relationship between plant cost and product variability, the optimal number of reactors for various production rates, the trend of operating parameters under varying kinetic parameters, and the cost of process flexibility. The additional cost of a flexible-process design is typically low, but this cost increases as process demands vary to greater extents.