6
`Polyalkylene Glycols
`
`Paul L. Matlock, William L. Brown, and Nye A. Clinton
`Union Carbide Corporation
`Tarrytown, New York
`
`INTRODUCTION
`I.
`Polyalkylene glycols are unique among synthetic lubricants because of their high oxygen content.
`As lubricants, they are exceptionally clean, allowing use where petroleum products would build
`tars and sludges. By varying their structure, one can vary their solubilities from water soluble
`to water insoluble. They are the only lubricants available with water solubility. A product of
`World War II, they quickly found uses where petroleum-based lubricants fail. This chapter cov-
`ers lubrication uses only; however, polyalkylene glycols have many applications in addition to
`lubrication.
`
`II. HISTORICAL DEVELOPMENT
`Polyalkylene glycols are one of many important industrial chemicals developed during World
`War II. This work was performed by H. R. Fife, and to a lesser extent by R. F. Holden, as a joint
`development project between Union Carbide Chemicals and Plastics Company Inc. (then known
`as the Union Carbide and Carbon Corporation) and the Mellon Institute of Industrial Research in
`Pittsburgh. Union Carbide Chemicals and Plastics Company Inc. held the original patents for the
`common lubricants [1–3].
`The first use of polyalkylene glycols was in water-based hydraulic fluids [4]. First devel-
`oped for the navy [5] for use in military aircraft, these compounds were being investigated as
`early as 1943. They were formulated from water, ethylene glycol, a polyalkylene glycol that
`acted as a thickener, and an additive package. In military aircraft, it is important that fires not
`break out when bullets or shrapnel sever hydraulic lines. The final test the Navy conducted was
`to fire a 50-caliber incendiary bullet, shredded by first passing through a steel baffle, through 1-
`gallon cans of test fluid. This test was passed by UCON Hydrolube U [6] using a polyalkylene
`glycol thickener.
`
`159
`
`Copyright © 1999 Marcel Dekker, Inc.
`
`Page 1 of 35
`
`Arkema Exhibit 1118
`
`

`
`160
`
`Matlock et al.
`
`More severe flammability requirements were established after the war. Hydraulic fluids to
`be used for missile ground handling equipment were developed that would not burn in a 100%
`gaseous oxygen atmosphere when the fluid was ejected at a pressure of 3000 psi in the presence
`of a continuous electric discharge ignition source [7]. Aqueous solutions of polyalkylene glycols
`could be formulated to pass this test.
`When polyalkylene glycols were first developed, the high viscosity indices and low pour
`points were quickly identified [8], leading to the use of these compounds in all-weather, heavy-
`duty brake fluids. Besides being fluid at temperatures that would cause petroleum products to
`freeze, they were also water tolerant. Small amounts of water contaminants would dissolve, not
`significantly changing the physical properties of the fluid nor crystallizing at low temperatures.
`This is still a major use of polyalkylene glycols today.
`Polyalkylene glycols were extensively used as aircraft engine lubricants in cold climates [9].
`Over 150,000 flying hours were accumulated, mostly in Alaska, using an inhibited polypropylene
`glycol monobutyl ether. The low pour point allowed aircraft engines to start at temperatures as
`low as ⫺30⬚F without diluting the lubricant with fuel, a step that can be used to reduce lubricant
`viscosity. It was possible to hydraulically feather the propellers using the polyalkylene glycol
`based lubricant down to ⫺60⬚F. Clean burn-off, an intrinsic property of polyalkylene glycols,
`resulted in low levels of carbon deposits and sludge, making engine cleanup easier during main-
`tenance. Polyalkylene glycols were finally judged unsuitable for aircraft engine oils because of
`factors: corrosion and deposits. Corrosion, due to the tendency of polyalkylene glycols to absorb
`water, was principally a problem for engine parts exposed to moist air. Corrosion protection
`additives were not available at that time for polyalkylene glycols. The hard deposits consist
`primarily of lead from the fuel. The clean burn-off tendency of the fluid apparently was respon-
`sible for this. The lead deposits formed with petroleum as an engine lubricant are soft and have
`a lower lead content. It is believed that these unusual lead deposits resulted in valve sticking after
`about 300–400 hours of operation [10] although no valve sticking was observed if valve clear-
`ances were adequate.
`Lubrication engineers quickly developed new uses of polyalkylene glycols. The uses devel-
`oped were for petroleum oil replacement in operations where petroleum oil was not entirely sat-
`isfactory and the higher cost of the polyalkylene glycol could be justified. The desirable proper-
`ties of the polyalkylene glycols include a low tendency to form carbon and sludge, clean burn-
`off, solvency, high viscosity indices, tolerance for rubber and other elastomers, low pour points,
`and low flammability.
`Polypropylene glycol monobutyl ethers were tested extensively as lubricants for automobile
`engines [11]. The fluids showed the expected low carbon and low sludge, as well as clean engine
`parts and satisfactory cranking at low temperature. Over 2 million miles of operation using these
`oil were experienced. This market was never developed.
`Because polyalkylene glycols burn off cleanly, they are desirable to use in high temperature
`applications where petroleum lubricants would form sludge. They have been used in glass facto-
`ries to lubricate the turrets of hot cut flare machines or to lubricate the bearings of rollers that
`smooth glass sheets. When mixed with graphite, polyalkylene glycols are very effective at lubri-
`cating bearings of carts being rolled into kilns. After the polyalkylene glycol has burned off, a
`soft, lubricating layer of graphite is left behind.
`Polyalkylene glycols were found to have little or no solvent or swelling effects on most
`synthetic or natural rubbers. This gave rise to many uses calling for the lubrication of rubber
`parts, such as rubber shackles, joints, or O-rings, or in the manufacture of rubber parts, where
`demolding lubricants were needed.
`
`Copyright © 1999 Marcel Dekker, Inc.
`
`Downloaded by [Charles Mitchell] at 20:21 16 January 2016
`
`Page 2 of 35
`
`

`
`Polyalkylene Glycols
`
`161
`
`III. CHEMISTRY
`A. Nomenclature
`Polyalkylene glycol is the common name for the homopolymers of ethylene oxide, propylene
`oxide, or butylene oxide; or the copolymers of ethylene oxide, propylene oxide, and/or butylene
`oxide. Although polyalkylene glycol is the common usage, Chemical Abstracts refers to these
`materials as polyoxyalkylene glycols. The ethylene oxide polymers are generally called poly
`(ethylene glycols) or poly (ethylene oxides). The Chemical Abstracts nomenclature is oxirane
`polymer. The propylene oxide polymers are known as poly (propylene glycols) or poly (propy-
`lene oxides) with a Chemical Abstracts name of oxirane, methyl polymer. The butylene oxide
`polymers are known as poly (butylene oxides) with a Chemical Abstracts name of oxirane, ethyl
`polymer. The copolymers are known as “oxirane, polymer with methyloxirane” or “oxirane,
`methyl polymer with oxirane,” depending on which oxide was used in the greater amount.
`Butylene oxide polymers are treated similarly. The Chemical Abstracts nomenclature does not
`distinguish between random and blocked copolymers (see below). The individual polymers and
`the copolymers all fall into the class of polyalkylene glycols. This latter name leads to the
`acronym PAGs. The acronym PAO has occasionally been used to indicate poly (alkylene oxide),
`but PAO is commonly used to designate poly (␣-olefin).
`
`B. Mechanism of Polymerization
`Polyalkylene glycols are prepared by polymerizing epoxides with a starter that consists of an
`alcohol and a smaller amount of its metal alkoxide, usually the potassium or sodium salt. The
`epoxide reacts with the metal alkoxide form of one of the starter alcohol molecules to give an
`alkoxide derivative of a new alcohol. This new metal alkoxide is in equilibrium with all the alco-
`hols present, so that the next reaction of an epoxide can occur either with the molecule that has
`already reacted or with a different alcohol:
`
`Copyright © 1999 Marcel Dekker, Inc.
`
`Downloaded by [Charles Mitchell] at 20:21 16 January 2016
`
`Page 3 of 35
`
`

`
`162
`
`Matlock et al.
`
`If the epoxide is propylene oxide, a propyloxy group results:
`
`If the epoxide is butylene oxide, a butyloxy group results:
`
`The equilibrium between alcohol and alkoxide determines the molecular weight distribution
`of the product. The epoxide monomers react with the metal salts of the alcohol at much faster
`rates than are observed with the alcohols. Whichever alcohol is most acidic will tend to form the
`alkoxide salts and will be the most reactive toward the epoxide. Once each starter alcohol has
`reacted with at least one epoxide, all molecules in the system will have approximately the same
`reactivity. Unless the parent alcohol is extremely unreactive, the fast exchange of metal salt
`between the growing polymer chains then results in what is nearly a Poisson distribution for
`molecular weight. The starter alcohols in commercial polymers use relatively reactive alcohols.
`The Poisson distribution is a much narrower distribution than the most probable or Gaussian dis-
`tribution. In many applications the narrow distribution is critical, since it means that there is no
`significant fraction of low molecular weight, volatile, or low-boiling components. In addition, a
`narrow molecular weight distribution leads to a high viscosity index. Polymerization of ethylene
`oxide produces a structure like the following:
`
`Ethylene oxide has two reactive sites, and the product is the same no matter which one reacts.
`The situation is different with propylene oxide and butylene oxide. In this case the ring opening
`occurs predominantly to produce a secondary hydroxyl group. This result is due to steric factors;
`the methylene ring position is less hindered than the methylene. For propylene oxide, 96% of the
`end groups are secondary hydroxyls and 4% are primary hydroxyls:
`
`This leads to a polymer structure approximated as follows:
`
`Copyright © 1999 Marcel Dekker, Inc.
`
`Downloaded by [Charles Mitchell] at 20:21 16 January 2016
`
`Page 4 of 35
`
`

`
`Polyalkylene Glycols
`
`163
`
`Copolymers of ethylene oxide and propylene oxide have two types of structure, random and
`blocked. In the random polymer, the two epoxides are co-fed to the starter and will both be incor-
`porated throughout the polymer. They react to give a product that is itself reactive and is in the
`acid–base equilibrium with all the other alcohols and metal alkoxylates present. To a first approx-
`imation, the epoxides are incorporated in a random manner dependent on the relative amounts of
`each epoxide present and the molecular weight distribution is still approximated by the Poisson
`model. Polymers with this structure are identified as random copolymers. A portion of the struc-
`ture of a random ethylene oxide–propylene oxide copolymer is shown schematically:
`
`In the block copolymer, an alternative structure is produced by reacting the starter first with one
`of the epoxides to produce a homopolymer. This can then be reacted with a different epoxide to
`produce a block copolymer. This name arises from the presence of a chain of one structure con-
`nected to a chain with a different structure. A block copolymer produced by feeding propylene
`oxide to propylene glycol followed by feeding ethylene oxide is shown schematically as follows:
`
`The polyalkylene glycols that are used commercially as lubricants are of five main types.
`
`1. Homopolymers of propylene oxide (polypropylene glycols), which are the water-insolu-
`ble type. These show limited solubility in oil. These are typically monobutyl ethers:
`
`2. Copolymers of ethylene oxide and propylene oxide, which are the water-soluble type.
`These are typically diols or monobutyl ethers:
`
`Copyright © 1999 Marcel Dekker, Inc.
`
`Downloaded by [Charles Mitchell] at 20:21 16 January 2016
`
`Page 5 of 35
`
`

`
`164
`
`Matlock et al.
`
`3. Polymers of butylene oxide. These show greater oil solubility than the homopolymers of
`propylene oxide. They have the following structure:
`
`4. Polymers of propylene oxide and higher epoxides designed to give greater oil solubility.
`These have the following structure:
`
`5. Polymers of propylene oxide that are dimethyl ethers:
`
`Polymers consisting of all ethyloxy groups, the polyethylene glycols, are not often used as lubri-
`cants, since they tend to crystallize at room temperature when their molecular weight exceeds
`600. Nevertheless, solid polyethylene glycols are used in specialty lubrication applications where
`the solid formulation is advantageous.
`The tendency of polyethylene glycol chains to crystallize affects the block polyalkylene gly-
`cols. If the blocks of ethyloxy groups are long enough in a block copolymer, pastes or waxes result.
`Block structures also tend to give the polymers surfactant properties in water. As a result, block
`polyalkylene glycols are often used as surfactants. However, surfactant-like properties are of little
`use for most lubrication applications.
`The epoxide polymers formed by base-catalyzed reactions typically have molecular weights
`of less than 20,000. Traces of water in the monomer feed and minor side reactions limit the aver-
`age molecular weight that can be achieved. The major side reaction for the base-catalyzed poly-
`merization of propylene oxide is the rearrangement of propylene oxide to allyl alcohol. This was
`recognized as early as 1956 [12]. The rearrangement involves deprotonation of the methyl group
`on the propylene oxide, followed by intramolecular ring opening:
`
`Copyright © 1999 Marcel Dekker, Inc.
`
`Downloaded by [Charles Mitchell] at 20:21 16 January 2016
`
`Page 6 of 35
`
`

`
`Polyalkylene Glycols
`
`165
`
`This mechanism is supported by kinetic studies with 1, 2-epoxypropane-3,3,3-D, which show a
`large positive isotope effect [13]. The allyl alcohol formed reacts with ethylene oxide and propy-
`lene oxide to yield new polyalkylene glycol molecules, which are monoallyl ethers. This chain
`transfer reaction limits the ultimate molecular weight that can be achieved with base catalysis.
`Although rearrangement of butylene oxide is not reported, this compound undoubtedly undergoes
`a similar base-catalyzed rearrangement. Strong base catalysts that give reduced amounts of unsat-
`uration are reported to be barium and strontium hydroxides [14].
`An alternative technology for polymerization can produce much higher molecular weight
`ethylene oxide polymers. Using coordinate-initiated polymerization, it is possible to produce eth-
`ylene oxide polymers with molecular weights in excess of a million. These compounds are pro-
`duced commercially by Union Carbide Chemicals and Plastics Corporation under the trade name
`of Polyox. A calcium amide alkoxide suspended in a solvent that does not dissolve the polymer
`product is used. Ethylene oxide is added and the polymer, not being soluble in the medium, is
`produced as a granular solid. Polymerization is thought to take place by coordination of the epox-
`ide to an electrophilic site on the catalysts. This coordination activates the epoxide for reaction
`with the growing chain.
`
`C. Synthesis of Polyalkylene Glycols
`Ethylene oxide is a toxic material with a time-weighted average for 8 hours of exposure of 1 ppm
`and a short-term permissible limit of 5 ppm in a 15-minute period, as determined by the federal
`Office of Occupational Safety and Health (OSHA). It is highly flammable and has a wide flam-
`mable range in air of 3.0–100%. It can explosively decompose if exposed to an ignition source.
`The flammability is only heightened by a boiling point of 10.4⬚C, making it a gas at ordinary tem-
`peratures. It can be polymerized with acidic, basic, and coordination catalysts, a polymerization
`that is very exothermic. A very careful study of tile hazards and procedures for safely handling
`ethylene oxide must be undertaken before the use of this substance is attempted. Similar hazards
`exists with propylene oxide and butylene oxide.
`In the laboratory, it is possible to use glass equipment at atmospheric pressure to prepare
`ethylene oxide, propylene oxide, and mixed ethylene oxide–propylene oxide polymers. A
`nitrogen-flushed flask is charged with the starter solution and fitted with a dry-ice condenser.
`A small amount of the epoxide is fed to the heated flask (typically 100⬚C or more) and
`allowed to reflux from the dry-ice condenser. The epoxide charge will be slowly consumed by
`the polymerization reaction, and the reflux rate will decrease. More epoxide is added at a rate
`sufficient to keep the system at reflux [15]. The rate can be increased by keeping the appara-
`tus under a slight pressure from a dip tube immersed in an inert liquid. The higher pressure
`increases the concentration of monomer in the reaction solution. To make a random copoly-
`mer, the two oxides are co-fed; a block copolymer requires sequential feeds of the two differ-
`ent epoxides. A similar system can be designed for coordinate-initiated polymerization.
`The use of an autoclave for the polymerization will result in much faster rates, since opera-
`tion at higher pressures is possible, resulting in much higher liquid phase concentrations of
`the monomers. The epoxide can be fed either by forcing it into the autoclave from a pressurized
`feed vessel with nitrogen pressure or by pumping it into the reactor. The reactor needs to be
`equipped with a cooling system and a control scheme to follow and regulate both pressure and
`temperature. The reactor is heated to the desired operating temperature and the epoxide fed until
`the pressure has reached the desired level. As the reaction progresses, the pressure will fall and
`more epoxide can be fed. Pure ethylene oxide vapor can explosively decompose upon exposure
`to an ignition source. A sufficient amount of nitrogen present before the initiation of the epoxide
`feed will ensure that the vapor phase does not reach the flammable limit at any time during the
`
`Copyright © 1999 Marcel Dekker, Inc.
`
`Downloaded by [Charles Mitchell] at 20:21 16 January 2016
`
`Page 7 of 35
`
`

`
`166
`
`Matlock et al.
`
`run. It is critical to keep the inventory of unreacted oxide in the reactor at a level such that the
`heat of polymerization (20 kcal/mol) can be removed by the cooling system. A critical factor in
`keeping the oxide concentration low is the reactor temperature. If pressure is the control mecha-
`nism, a low temperature in the reactor will allow the oxide to build to a potentially unsafe con-
`centration. Concentration of unreacted epoxides is the cause of the greatest number of reactor
`failures. The problem becomes larger with propylene oxide and especially butylene oxide, where
`the vapor pressure of the oxide may not be a reliable indication of liquid phase concentration. The
`reactor should have a safety relief device sized to handle a runaway reaction due to loss of cool-
`ing. One of the authors has seen an autoclave and its high pressure cell catastrophically destroyed,
`with the autoclave top thrown many hundreds of feet. The cause was the inadvertent feeding of
`ethylene oxide at a low temperature, an error that allowed the accumulation of a large inventory
`of ethylene oxide. The uncontrolled polymerization that followed proved to be uncontainable.
`To avoid exposure of personnel to unreacted ethylene oxide or propylene oxide, it is neces-
`sary to hold the reactor contents at temperature after the end of the feed until the concentration
`of unreacted epoxides has dropped to an acceptable level. This procedure is called a cook-out or
`digestion. A cook-out may be necessary during synthesis because the vessel will fill with liquid
`as the reaction proceeds and the polymer is produced. This will compress the nitrogen in the ves-
`sel, and the partial pressure of the monomer will therefore decrease (the system is run by keep-
`ing pressure constant). The reaction rates will fall to unacceptably low levels. Venting of the
`excess nitrogen after a cook-out will allow feed to be resumed at faster rates. It may even be
`necessary to remove some of the reactor’s liquid contents to allow room for further reaction.
`This is most likely to occur during the synthesis of higher molecular weight products.
`The commercial preparation of poly(alkylene oxides) is carried out in a manner analo-
`gous to that described for the laboratory autoclave. A semibatch stainless steel system with
`a recirculation loop and an agitator has been described. The reactions are carried out at
`100–120⬚C at pressures of 60 psig. The oxide feed rate is controlled by pressure, and feed
`times are on the order of 15 hours or more [16].
`Pressindustria Company has reported a novel method for synthesizing polyalkylene gly-
`cols [17]. The solution of growing polymer is sprayed through the headspace of a horizontal
`reactor. The reaction with oxide monomer is reported to take place at the gas–liquid interface.
`Rapid reactions without large increases in pressure or temperature are reported. Cooling takes
`place with an external heat exchanger.
`The first commercial-scale syntheses were performed at the Union Carbide production facil-
`ity in Charleston, West Virginia. The first preparations were similar to that used today. Sodium
`salts of alcohols were used as starters at reaction temperatures slightly above those used currently.
`Butanol was the starter alcohol of choice for monoethers. The products developed at this time
`form the backbone of the UCON® Fluids product line of Union Carbide and has been widely
`duplicated by other manufacturers. The method first used at Union Carbide to neutralize these
`fluids was unusual. The crude fluids were diluted with water, acidified with carbon dioxide,
`extracted with hot water, and then stripped of water at high temperature. Decolorizing with acti-
`vated charcoal was the last step.
`
`D. Preparation of Capped Polyalkylene Glycol
`Polypropylene glycol dimethyl ether has become available as a commercial lubricant. This
`requires that the hydroxyl end groups be converted to the methyl ether:
`
`Copyright © 1999 Marcel Dekker, Inc.
`
`Downloaded by [Charles Mitchell] at 20:21 16 January 2016
`
`Page 8 of 35
`
`

`
`Polyalkylene Glycols
`
`167
`
`Although the conditions used commercially to effect this transformation are not reported, it is
`almost certainly done by a Williamson ether synthesis. This involves converting the alcohol to its
`alkoxide form, followed by reaction with methyl chloride:
`
`The art of this transformation is the conversion of the alcohol into its alkoxide. Sodium hydrox-
`ide is the most convenient base, and it works well for low molecular weight ethoxylates [18,19].
`Higher molecular weight ethoxylates can be capped by means of sodium hydroxide and a phase
`transfer catalyst [20]. Propoxylates, which are secondary alcohols, hence more difficult to con-
`vert to their alkoxide form, are normally capped by adding sodium or potassium methoxide and
`driving the equilibrium reaction between the different alkoxide forms to the polyether alkoxide
`by removing the methanol [21]. Capping efficiency is limited by the difficulty of converting all
`the end groups to the higher energy alkoxide species. Improved yields can be obtained by using
`sodium hydride after the methoxide reaction has been driven as far as is practical [21] or by using
`sodium hydride alone [22].
`
`E. Oxidative and Thermal Stability
`The bond strength of the carbon–carbon bond is 84 kcal/mol (ethane), which is slightly stronger
`than the 76 kcal/mol carbon–oxygen bond of an ether (dimethyl ether) [22]. Other authors have
`reported that carbon–oxygen ether bonds are comparable to, or slightly stronger than, the usual
`carbon–carbon bonds [23]. However, from a thermochemical standpoint, polyalkylene glycols
`are usually considered less stable than typical hydrocarbons. In the absence of air, they can be
`used up to about 250⬚C.
`The poly(alkylene oxides) are all polyethers with an oxygen atom in every third position of the
`polymer backbone. As with all ethers, a secondary or tertiary carbon adjacent to the ether oxygen
`is susceptible to oxidative attack. The mechanism involves a free radical abstraction of the hydro-
`gen on the ␣ carbon, resulting in a carbon-based radical stabilized by the adjacent oxygen atom.
`This can then react with oxygen to produce a peroxy radical. The chain process is continued, with
`the peroxy radical abstracting a hydrogen atom to give another oxygen stabilized radical:
`
`Copyright © 1999 Marcel Dekker, Inc.
`
`Downloaded by [Charles Mitchell] at 20:21 16 January 2016
`
`Page 9 of 35
`
`

`
`168
`
`Matlock et al.
`
`The process does not continue to build peroxide levels, since a number of mechanisms lead to
`peroxide destruction. In the early stages of oxidation the peroxide will increase, but as the reac-
`tion proceeds and carbonyl levels build, the peroxide level reaches a steady state. Further reac-
`tion will lead to the formation of acidic material. As the acidic oxidation products build, the vis-
`cosity begins to drop. Studies of the oxidation of polypropylene glycols [25] and polyethylene
`glycols [26] show that the polymers degrade into lower molecular weight products as they oxi-
`dize. This is in contrast to the behavior of petroleum products, which build higher molecular
`weight products as they oxidize [27]. It is these higher molecular weight materials that precipi-
`tate from the oil and produce sludge.
`It is the fate of the hydroperoxides that determines the tendency of a lubricant to form sludge.
`In the case of petroleum, the hydroperoxides form hydrocarbon-based aldehydes, which can
`undergo aldol condensations to form high molecular weight ␣,␤-unsaturated species that them-
`selves are reactive [28]. This leads to the formation of high molecular weight polar species that
`are not soluble in the parent nonpolar base oil and therefore precipitate out of solution to form
`sludge.
`
`The fate of hydroperoxides is different in polyalkylene glycols and in petroleum. Because
`every third atom is an oxygen, a high proportion of the chain cleavage products in the former case
`are esters. While esters can condense, the reaction is difficult and is unlikely to occur. The type
`of end groups can be determined by proton magnetic resonance. For polyethylene glycols, a
`prominent formate peak is formed at 8.1 ppm, indicating the following reaction:
`
`Copyright © 1999 Marcel Dekker, Inc.
`
`Downloaded by [Charles Mitchell] at 20:21 16 January 2016
`
`Page 10 of 35
`
`

`
`Polyalkylene Glycols
`
`169
`
`In the case of polypropylene glycols, the reaction products are more complex owing to the greater
`complexity of the molecules. The end groups produced are shown in Table 1 [25]. These end
`groups are not simple aldehydes and will undergo condensation with difficulty.
`When sufficient degradation has taken place, volatile products will be produced. These have
`been quantified in the case of polypropylene glycol degradation [25], as shown in Table 2.
`The oxidation of polyalkylene glycols yields polar, oxygenated products. Polyalkylene gly-
`cols themselves are polar and will dissolve these oxidation products. In contrast, oils are nonpo-
`lar and their oxidation products are polar, consisting of peroxides and carbonyl species as well.
`Oils will not dissolve these polar species, and this property contributes directly to their tendency
`to form sludge and varnish.
`The tendency of polyalkylene glycols to solubilize their own degradation products, and their
`cleavage to form volatile species, can be cited to explain their low Conradson carbon and
`Ramsbotton carbon (ASTM D189 and D 524), typically less that 0.01%. When sludges do form
`from polyalkylene glycols, it is usually in oxygen-starved systems as a result of aldehyde conden-
`sation. Under conditions of exhaustive oxidation, the chain cleavage will have occurred to such an
`extent that the oxidation products evaporate. The volatilization of the fluid, together with the ten-
`dency not to form carbon or sludge, means that the polyalkylene glycol will be removed under
`high temperature applications in a property known as clean burn-off. Clean burn-off is important
`in a number of the applications for these products—for example, as a carrier for graphite on chains
`being used in ovens or kilns. As with the pyrolysis products of any organic material, good venti-
`lation, should be installed to ensure the removal of the vapors from the workplace.
`The oxidation of polyalkylene glycols could result in shorter than desired service life for
`some applications. Oxidation can effectively be controlled through the addition of antioxidants
`to interrupt the chain transfer oxidation mechanism [29]. Typical antioxidants that have been
`used include butylated hydroxyanisole, phenothiazine, hydroquinone monomethyl ether, butyl-
`
`Copyright © 1999 Marcel Dekker, Inc.
`
`Downloaded by [Charles Mitchell] at 20:21 16 January 2016
`
`Page 11 of 35
`
`

`
`170
`
`Matlock et al.
`
`ated hydroxytoluene [30], and phenyl-␣-naphthylamine [31, p. 109]. The poly(alkylene oxides)
`are dramatically stabilized toward oxidation by the addition of antioxidants. In many cases antiox-
`idants at levels of a few hundred parts per million are sufficient to stabilize against oxidative degra-
`dation under mild conditions and higher levels will stabilize these systems under much more
`severe conditions. The uses of inhibited polyalkylene glycols as heat transfer fluids and as gear and
`calender lubricants are all examples of successful high temperature applications. These applica-
`tions show that if the system is protected against oxidative attack, either by the addition of an
`antioxidant or by removal of oxygen, polyalkylene glycols will have very good high temperature
`stability.
`
`F. End Group Chemistry
`The polyalkylene glycols all have at least one hydroxyl group on the end of the molecule. If they
`have been produced from water or a multifunctional starter, they will have more than one hydroxyl
`group. The polyols used for urethane applications perform by virtue of the reaction of the
`hydroxyl groups reacting with isocyanate groups to give the urethane linkage. In this application,
`stearic factors cause the primary hydroxyl group to be more reactive than a secondary hydroxyl.
`The urethane polyols are formed from propylene oxide. Since this results in a less reactive
`secondary hydroxyl end group, it is necessary to end-cap the polyol with a small amount of
`ethylene oxide to increase the number of primary alcohol terminated molecules.
`Other end group reactions are used to functionalize polyalkylene glycols. These reactions use
`the known alcohol derivatization reactions. Esters are formed by reaction with either organic or
`inorganic acids. In addition to the etherification reactions already discussed, it is possible to react
`the alcohol with a strong acid and an olefin to give an alkyl ether cap.
`Other derivatization reactions of the alcohol end group are possible. It has been reported that
`poly (dichlorophosphazene) can be reacted with polyalkylene glycols to give poly(phosphazene)
`derivatized with grafted polyalkylene glycol chains [32]. These materials are reported to have
`utility as metalworking lubricants.
`
`G. Coordination Chemistry
`The presence of an ether oxygen atom at every third position of the polymer backbone leads
`to the rich coordination chemistry of these compounds. The use of these polymers as phase
`transfer agents has been reviewed [33]. Complexation with phenols, phenolic resins, bromine,
`
`Copyright © 1999 Marcel Dekker, Inc.
`
`Downloaded by [Charles Mitchell] at 20:21 16 January 2016
`
`Page 12 of 35
`
`

`
`Polyalkylene Glycols
`
`171
`
`iodine, gelatin, sulfonic acids, mercuric slats, tannic acid, poly(acrylic acid), and urea all have been
`reported [34]. The use of poly(ethylene oxide) polymers as flocculation agents is related to their absorp-
`tion on colloidal silica, clay, and minerals.
`The facile wetting of metal parts in lubrication applications is related to the ability of the
`polymer to associate with me metal surface. Like the complexation of other chemical species, the
`ability of the polyalkylene glycol to wet a metal surface is due to the presence of an ether oxy-
`gen atom at every third position of the polymer chain. This results in good extreme pressure and
`metalworking performance. The solution properties of these polymers in water are also directly
`related to the association of the water with the ether oxygen atoms.
`
`IV. PHYSICAL PROPERTIES
`A. Base Fluids
`The physical properties of poly(alkylene oxides) are best understood by considering them as a
`series of homologous derivatives. Thus the polymers derived exclusively f