Printer Friendly

Review: Recycling of Nylon From Carpet Waste.


Large amounts of post-consumer carpet are discarded every year. Most of this waste is currently landfilled, while a small percentage is incinerated. The face carpet fibers, consisting primarily of nylon 6 and nylon 6,6, represent the majority component in the carpet waste. Recent financial incentives and environmental constraints have motivated the industrial sector to develop recycling strategies for these fibers. Depolymerization into their constituent monomers is the most complex recycling route, but at the same time it produces the most valuable product. A second alternative involves the use of solvents for the extraction of carpet fiber components in their polymeric form. Finally, a third recycling option yields thermoplastic mixtures by melt blending the carpet waste. The recent literature on the recycling of nylon from carpet waste is reviewed in this paper. The paper also includes a section focusing on the current state of carpet recycling at the industrial level.


In the United States alone, approximately 3.3 billion lbs of carpet fibers, including nylon, polyester, polypropylene, acrylic, wool and cotton, are produced annually [1]. Of this amount, approximately 65%--or 2 billion lbs--is composed of nylon 6 and nylon 6,6. The European annual production of carpet, including tufted, woven and needled products is approximately 1.5 million tons (3.3 billion lbs) [2]. The average life cycle of a carpet is between 8 and 12 years. Carpet production also generates a substantial amount of waste in the form of trimmings and cuts, which typically amounts to 12% of the total production [3, 4]. As a result, an enormous amount of synthetic waste is generated, most of which is disposed in landfills. Landfilling is not an environmentally friendly solution since carpet fibers, like many other synthetic polymers, are not biodegradable. Furthermore, the cost of disposal is increasing continuously owing to limited landfill capacity. Environmental concerns and governmental regulations have spurred efforts In the direction of recycling all non-biodegradable synthetic polymers, of which carpets and carpet fibers constitute a significant percentage.

The literature In the area of carpet recycling is relatively sparse. A review paper on the subject of nylon recycling was published in 1991 by Datye [4], and it provides a detailed insight into the recycling of post-production nylon 6 waste. This review, however, does not focus upon the issue of carpet recycling at both the post-manufacture and post-consumer levels, as this concept has emerged only over the last few years because of the large amounts of non-biodegradable carpet accumulated In landfills. Recycling of the entire carpet poses a significant challenge because of the inhomogeneous nature of the material. Moreover, it is even more difficult to recycle post-consumer carpet waste because of the dirt, cleaning chemicals, and other materials accumulated in the carpets over the years. Studies carried out in Europe reveal that post-consumer carpet is approximately 30% heavier than new carpet owing to dirt accumulated in the piles [5]. It also contains a significant amount of contaminants, mainly the chem icals used for cleaning purposes. It is this inhomogeneous nature of carpets that makes effective recycling a very difficult and costly process.

The recycling of polymers has been broadly classified into four categories [3]:

(a) Primary recycling or depolymerization, which converts the waste into products having a quality equivalent to that of the "virgin" polymer. This category encompasses methods to break down the long polymer chains into their original monomers that can be repolymerized.

(b) Secondary recycling, which involves recovering of the individual components of a polymeric mixture without necessarily breaking them down to the monomeric form. This category includes various extraction and separation methods.

(c) Tertiary recycling, which consists of preparing a thermoplastic mixture by melt-blending the entire carpet waste. This recycling route requires no fiber separation or latex removal, as the carpet components are mixed by means of reactive extrusion and compatibilization. Products of lower quality can be manufactured from this blend, by methods such as injection molding.

(d) Quaternary recycling, which involves only energy recovery during the incineration of the polymer waste.

The intention of this review is to provide an insight into the problem of carpet recycling, along with an analysis of the different methods being proposed or commercially utilized. The reviewed literature includes a limited amount of journal publications, which focus primarily on fundamental aspects, and a large number of patents, which describe the available technologies.


Artificial fibers have been created in an effort to improve the quality and availability of textiles, and to reduce the cost of several products for the consumer. Initial attempts were targeted towards the synthesis of fibers with properties similar to those of the natural materials. During the industrial revolution of the 19th century, the first patent for "artificial silk" was granted to a Swiss chemist in 1855 in England. It was not until the end of the 19th century that the "artificial silk" (rayon) started being produced on a commercial scale by French chemist Count Hilaire de Chardonnet.

Discovered in 1931 by chemists at E. I. DuPont de Nemours and Company, nylon 6,6 quickly became one of the most used fiber materials. Nylon 6,6 is the polyamide formed by the reaction between hexamethylene diamine and adipic acid. Each of the monomers has 6 carbon atoms, thus the designation 6,6. Commercial production of nylon 6,6 by DuPont began in 1939. Its initial applications were sewing thread, parachute fabric, and women's hosiery. When the U.S. entered World War II, its nylon production was allocated for military use. In the post-war industry, nylon became one of the most widely used artificial materials, with a variety of applications. By the end of the 1940s, it was already being used for upholstery and in carpets. During the same period, Paul Schlack of the infamous I. G. Farben Company in Germany obtained a different form of polyamide by using caprolactam as the monomer, and called it "nylon 6." Most carpet fiber yarns in production today are either nylon 6 or 6,6, although small percentage is made of polyester, acrylics, polypropylene and other olefin fibers, wool or cotton [1, 6].

A typical carpet has four main layers (5) as shown in Fig. 1. The top layer, or face yarn, is composed of nylon fibers tufted through a primary backing, which is usually made of polypropylene. Other fibers such as jute, polyethylene, polyesters and rayon may also be used. Latex adhesive is applied under the primary backing in order to secure the face fiber. The adhesive is usually made of a styrene butadiene co-polymer (SBR), which is filled with inorganic materials such as [CaCO.sub.3] or [BaSO.sub.4] [3]. The fillers are added for soundproofing purposes. Finally, a secondary backing (same material as the primary backing) may optionally be added to the primary backing and bonded to it by the same SBR adhesive. The nylon face fibers, containing dyes, soil-repellents (to improve the resistance to stains), and possibly other additives to improve the quality of the carpet, usually account for about half of the total carpet weight.


3.1 General Considerations

As mentioned in the Introduction, there are several recycling approaches, which differ by the type and quality of the product generated, and consequently by the type of process utilized. Depolymerization is the preferred route of carpet recycling, since it breaks down the carpet fibers (nylon 6 and/or nylon 6,6) into the corresponding monomers. This allows the recovery of the monomers that can be re-polymerized into new nylon products of high quality.

Alternatively, the nylon component can be separated from the carpet waste by extraction. In this process, the entire waste carpet material is dissolved in a solvent at elevated temperatures. During extraction, the nylon from the fibers is recovered in its polymeric form and can be reused in injection molding applications. The main problem associated with this approach is the selection of a suitable solvent that selectively dissolves the nylon fibers and does not react with or dissolve any of the other carpet components. Use of a partially selective solvent results in the recovery of nylon containing several impurities, and hence, having limited further use.

Melt blending of the entire carpet scrap generates a thermoplastic mixture that can be used for the manufacturing of a lower quality plastic material. Such a material can be utilized in less "demanding" products. The method consists of melting the entire carpet waste, without a previous separation into its components, to obtain a blend of different polymeric and inorganic materials. The low quality and lack of homogeneity of the resulting mixture are the main drawbacks of this method, restricting the number of applications in which its product can be used. Since carpets consist of several immiscible phases, compatibilizers (i.e. interfacial agents that decrease the surface tension and increase the interfacial adhesion between the different phases in a mixture) often must be added to the system in order to increase the miscibility of the various materials. The composition of the final product depends on the type and composition of the carpet recycled, varying significantly from one batch to another. Despite th ese problems, this recycling approach is still attractive because of its low cost and the utilization of the entire carpet waste, without a requirement for any prior separation.

Disassembling the face fibers off the used carpet material using different mechanical separation methods provides nylon 6 fibers that can be used in several applications. One example that has attracted some interest is their use in the reinforcement of concrete (7). Laboratory studies have shown that adding 1-2 wt% of short fibers separated from waste carpet to the concrete mix improves Its properties, such as its tensile strength and shrinkage.

Used together with landfilling, but to a much smaller extent, incineration is slowly being abandoned, mainly because of the air pollution problems it creates. The main idea behind incinerating this type of waste Is to at least recover some of the energy value that was input during the manufacturing of the carpet. Generally, this approach proves efficient when the energy needed to make the original material is less than twice the energy obtained by incinerating it (3). Otherwise, as is the case with polymeric materials, more advanced recycling techniques are preferred, which make use of the waste in a "smarter" way by transforming it into useful raw material for other applications. Hence, incineration of carpet waste is not regarded as a promising long-term alternative for carpet disposal.

In the following sections the technical advantages of each one of the most promising recycling techniques (depolymerization, extraction, melt blending, and mechanical separation) are reviewed in detail. An additional section is devoted to the current state of carpet recycling at the industrial level.

3.2 Depolymerization

During depolymerization, polymer chains are broken down into their monomeric constituents. A detailed description of the chemistry of the depolymerization of nylon 6 to its caprolactam monomer was presented by Agrawal in 1975 [8]. More recently, Datye [4] reviewed the depolymerization of waste nylon generated during the manufacturing of nylon 6 and filament yarn. Another review on the topic of e-caprolactam regeneration from wastes in the manufacture of nylon fibers and yarns has also been published by Dmitrieva et al. (9). In the following paragraphs we provide a summary of the most important issues related to the depolymerization process as applied to the recovery of nylon from carpet waste and review the more recent developments in this area.

The depolymerization of nylon 6 is a first-order, endothermic reaction, which takes place in initiation and de-propagation steps. Water is an initiator for the depolymerization process. This process is endothermic and requires high temperatures, achieved by the use of superheated steam. Steam not only acts as a heat source and water supply, but also provides a better agitation for the reaction system. Temperatures above the boiling point of caprolactam (around 267[degrees]C) are typically used. Under such conditions a heterogeneous system is formed (a liquid polymer melt and a gas-phase caprolactam product). The caprolactam monomer is removed from the reactor along with the steam, resulting in a shift of the equilibrium towards further monomer formation.

The depolymerization reaction is thermodynamically favored at low pressures, since the number of moles in the system increases as the polymeric chain breaks into monomeric units. Furthermore, at low pressures it becomes easier to separate the monomers (which usually have relatively high boiling points) from the melt. On the other hand, water solubility in the melt is favored at high pressures. Consequently, in most cases, there is an optimum process pressure (usually above atmospheric value), which maximizes the reaction rate for a given temperature.

Catalysts may also be added to the reaction mixture to accelerate the process. Both Lewis and Brnsted acidic and basic catalysts have been used for this application. Each class has certain advantages and disadvantages that will be discussed later in this section. The presence of a catalyst also allows for easier processing of the polymeric melt by decreasing its viscosity.

Several types of depolymerization processes have been proposed and patented. All of them result in the full recycling of the nylon component of a carpet into recycled nylon that has a quality comparable to the quality of the original nylon used. Nylon-producing companies have undertaken most of these efforts in the past, although carpet manufacturers have recently entered this area as well. The majority of the available depolymerization processes focus on the recovery of caprolactam, the basic monomer of nylon 6. Companies such as BASF, AlliedSignal (which in 1999 merged with Honeywell under the Honeywell name), and DSM. which produce nylon 6, have led the efforts in this area. DuPont, on the other hand, which manufactures nylon 6,6, has focused its depolymerization efforts on the recovery of hexamethylenediamine (HDM) and adipic acid (the corresponding monomers of nylon 6,6).

3.2.1 Depolymerization With Ammonia (Ammonolysis)

Nylon 6 and/or nylon 6,6 can be converted into a mixture of the respective monomers through an ammonolysis process [10]. This process involves heating of a polyamide mixture (such as a mixture of nylon 6 and nylon 6,6) in the presence of ammonia at high temperatures and pressures. At least one mole of ammonia (more is preferable) is needed per mole of amide group in the polymer. The reaction takes place at temperatures between 300[degrees]C and 350[degrees]C. In addition to temperature, the reaction rate also depends on the reactor pressure, with pressures in the range of 500-2500 psig being preferred. The products generated during this process include hexamethylene diamine (HMD), 5-cyanovaleramide (CVAM), adiponitrile (ADN), caprolactam (CL), 6-aminocaproamide (6-ACAM) and 6-aminocapronitrile (6-ACN). Water is also formed during the reaction, and is known to suppress the complete conversion of the intermediately formed amides to nitriles by shifting the equilibrium of this process. This can be avoided by con tinuous removal of the water, achieved by flowing excess ammonia through the reaction zone. The monomers produced are then fractionally distilled into separate streams containing HMD/6-ACN, CL and ADN [11]. HMD, 6-ACN and ADN are then hydrogenated to form pure HMD, the monomer required for the production of nylon 6,6. Caprolactam can be either purified, or recycled into the ammonolysis process. Although the ammonolysis reaction does not require a catalyst, the addition of a phosphate catalyst (such as phosphoric acid, ammonium phosphate or boron phosphate) has been shown to increase its rate.

McKinney, who invented the ammonolysis process, further explored the effect of Lewis-acid catalyst precursors on the rate of this process. He reported that the presence of such precursors makes the ammonolysis process more efficient [12]. The term "catalyst precursors" was used in these studies instead of "catalyst" since the catalyst may or may not maintain its original structure during the ammonolysis process. These precursors were selected from a group consisting of [ScX.sub.3], [TiX.sub.4], [MnX.sub.2], [ReX.sub.5], [FeX.sub.3], [CuX.sub.2], CuX, [ZnX.sub.2], [MoX.sub.6], [WX.sub.6], and [AIX.sub.3] (where X can be Cl, Br, or I).

A process patented by Hendrix et al. for DSM [13] also involves a method of depolymerizing a polyamide to its monomer in the presence of "at least one nitrogen-containing compound." According to its inventors' claims, this process can produce monomers with a higher selectivity as compared to DuPont's ammonolysis process. The nitrogen-containing compounds that can be used include ammonia, or a primary, secondary or tertiary amine. Additionally, the nitrogen-containing compound chosen has to have a boiling point lower than the boiling point of caprolactam (267[degrees]C). A concentration of two moles of the nitrogen-containing compound per mole of amide group in the polymer is recommended, while at least one mole of nitrogen-containing compound per amide group is required. Additionally, the authors claim that "accelerators" (i.e. catalysts, which increase the reaction rate and/or improve selectivity towards monomers, can be also used. Examples of such "accelerators" include Lewis (such as [Al.sub.2][O.sub.3]) or Bronsted acids (such as [H.sub.3][PO.sub.4] and [H.sub.3][BO.sub.3]). Reaction conditions include temperatures between 250[degrees]C and 350[degrees]C and pressures between 0.9 and 3 atmospheres.

The DSM process can be carried out in a batch, semi-batch, or continuous reactor; the continuous steady state process is preferred, since it improves the monomer selectivity. The nitrogen-containing compound reacts with the polyamide mixture in a molten form, and the products are removed from the reactor through the gas phase. The products are then separated by steam and vacuum distillation. The products include caprolactam and aminocaproic acid (a caprolactam precursor) together with small amounts of aminocaproamide and aminocapronitrile. The desired monomer in the DSM process is caprolactam. Consequently, the DSM process can be operated in the presence of water, to minimize nitrile formation. This is not the case in the DuPont ammonolysis process, which targets the formation of nitriles, and subsequently, HMD. Th preferred feed in the DSM process is nylon that has been separated from the other non-polyamide components of the post consumer carpet. However, the DSM patent states that the process can also be o perated without prior separation, but does not present any performance data in this case. Finally, mixtures of nylon 6 and nylon 6,6 can be used, although all examples shown in the patent used pure nylon 6 as the reactor feed.

3.2.2 Acid-Catalyzed Depolymerization

Acid catalysts ere utilized in early efforts to depolymerize nylon 6 to caprolactam, and several processes utilizing such catalysts have emerged from these efforts. When such processes are applied to the depolymerizatlon of carpet fibers, however, a separation of the fibers from the other carpet components (I.e., "beneficiation") is required, since the [CaCO.sub.3] usually present In the fillers can consume an equivalent amount of the acid catalyst used.

Among the several general acid depolymerization processes, one patented by Crescentini et al. (14) for Allied Corporation has several similarities to depolymerization processes used for carpet materials. The Crescentini process was developed for the depolymerization of cyclic oligomers of nylon 6 that are present in nylon 6 chip wash water, and it involves an initial separation of unreacted caprolactam from the cyclic oligomers present In the wash water. This can be achieved by feeding the aqueous mixture to a wiped-film evaporator at a temperature of 200[degrees]C to 300[degrees]C and a pressure of 10 to 250 Torr (0.013-0.329 atm). Caprolactam is carried off by steam in the overhead stream, and the cyclic oligomers are recovered from the bottoms. The oligomers are then fed to the depolymerization kettle. The operating temperature is between 230[degrees]C and 290[degrees]C, and the caprolactam formed is stripped off with superheated steam. A phosphoric or orthophosphoric catalyst is used at a level of 0.5 to 5 wt% of the cyclic oligomers fed. More than 75% of the oligomers can be recovered in the form of caprolactam. It should be pointed out, however, that when acid catalyzed processes are used, carpet separation Is required, since the [CaCO.sub.3] usually present in the fillers can consume an equivalent amount of the acid catalyst used

Corbin et a!. [15] invented a process for BASF Corporation for the continuous recovery of caprolactam from nylon 6 fibers, as well as carpets made from such fibers. Corbin et al. claim that beneficiation is not required in their process and offer an example to support this claim in the patent. Consequently, the non-nylon carpet components can either be separated (via shredding) before the reaction, or separated from the caprolactam product after depolymerization. In the latter case the separation costs are minimized, but the other carpet components, such as the polypropylene backing, cannot be recovered and recycled. In the BASE process, pure nylon 6 or a carpet containing nylon 6 fibers is fed to a continuous depolymerization reactor unit. Nylon 6 production wastes such as yams, chips or extruder slag, along with nylon 6 wash water, can also be included in the reactor feed. A thin film evaporator design may be used for the reactor and in this case the feed Is introduced in a molten form. The process can also be carried out in a batch reactor, although a continuous process is preferred. Phosphoric acid at a concentration of 5 to 7 wt% is also added as a catalyst. Other active catalysts for this application include boric acid and phosphate salts. Superheated steam is also introduced to the reactor, preferably at temperatures between 250[degrees]C and 280[degrees]C. The amount of steam added can vary depending upon the auxiliary heating requirements of the reactor (such as electrical heat to the reactor wall). The gas phase products are carried out of the reactor by the steam, and are distilled in order to recover the caprolactam monomer. The non-volatile wastes and by-products are directed to a powerhouse for incineration and heat recovery. It Is also possible to recover the used phosphoric acid catalyst during heat recovery. Examples are presented in the Corbin et al. patent, where both nylon 6 fibers and carpet waste containing nylon 6 fibers were used as the feed. In both cases, nearly pure caprolactam suitable for nylon 6 production was recovered. A caprolactam yield of 56% was reported for the case when only nylon 6 fibers were fed to the reactor.

BASF Corporation has also developed a semi-continuous process for the depolymerization of nylon 6 to caprolactam [16]. According to the patent's claims, this process successfully addresses some of the problems encountered with the continuous process, such as low efficiency in the presence of contaminants and large amounts of catalyst waste. The semi-continuous process uses acid catalysts such as ortho-phosphoric or para-toluenesolfonic acid at a loading between 5 and 35 wt% of the polyamide reactant. Polyamides that have been produced from a single monomer (nylon 6) are preferred, although mixtures of polymers can also be depolymerized As claimed by the patent, this invention is suitable for polymers In various forms such as molded articles, chips, fibers, films, as well as polymer wastes containing solid contaminants (pigments, [TiO.sub.2]. polyethylene, etc.). The primary reaction product Is once again caprolactam. The semi-continuous reactor could be any conventional acid-resistant electrically heated reac tor equipped with an inlet for the addition of the polyamide feed and a condenser. Steam, supplied through nozzles, is used as a carrier to remove the caprolactam product from the reactive mixture. Usually the reactor is also purged with an inert gas, such as nitrogen, in order to avoid the presence of oxygen in the system, which can adversely affect the reaction rate.

A typical run involves adding the polymer and the acid catalyst in ground or molten form to the semi-continuous reactor. The reactor is maintained at a temperature between 260[degrees]C and 280[degrees]C and superheated steam between 100[degrees]C and 450[degrees]C is added below the level of the polyamide/catalyst melt. The steam carries off the caprolactam product to a condenser, where the caprolactam-containing distillate is collected for further processing. When a sufficient conversion of the polyamide to the monomer (40% to 90%) is reached, a new batch of polymer is added and the process is repeated. No additional catalyst is added with the new polymer batch. When orthophosphoric acid is used as the catalyst, a loading of 5 to 25 wt% is required. For para-toluenesulfonic acid, the required loading is 10 to 35 wt%. According to Kotek, who invented this process, it is these high catalyst loadings that lead to the effective breakdown of the polymer into its constituent monomers, and to the minimization of u ndesirable by-products. We should point out, however, that the patent in this case does not provide quantitative examples to allow us to assess its effectiveness when compared with other available depolymerization processes.

A process targeting the depolymerization of nylon 6 and/or nylon 6,6 in the form of sheared nylon carpet face fibers, using aliphatic carboxylic acids, has been invented by Moran for DuPont [17]. Unlike the previous depolymerization processes, the acid used in this case is also one of the reactants. Suitable carboxylic acids include acetic and propionic acid. In the depolymerization of nylon 6, one mole of acid is required for every two moles of repeat polymer units, and the resulting monomer is caprolactam. In the depolymerization of nylon 6,6, two moles of acid are required for every mole of repeat polymer units, and the monomer produced is adipic acid. A particular advantage of this process, as claimed by the patent, is that mixtures of nylon 6 and nylon 6,6 can be simultaneously depolymerized to caprolactam and adipic acid, respectively, without requiring any prior separation. Generally, the carboxylic acid is added in excess, and the by-products of the reaction are 6-alkylamidohexanoic acid (from nylon 6 ) and N,N'-hexamethylene bisalkylamide (from nylon 6,6). The reaction is carried out in an autoclave at temperatures preferably above 250[degrees]C, and autogenous pressure. Once substantial conversion has been achieved, the reactor is cooled down, and the monomers are separated by steam distillation or other methods. In the examples given in the patent, significant amounts of the by-products were formed, with approximately 3% caprolactam produced, and 10O%-20% adipic acid.

An improvement of the aliphatic carboxylic depolymerization method was proposed by Moran and McKinney also for DuPont [18]. In this case, the 6-alkylamidohexanoic acid and N,N'-hexamethylene bisalkylamide byproducts are further oxidized to adipic acid. The oxidation (by air, oxygen or hydrogen peroxide) can take place either simultaneously with the depolymerization or in a separate step. Alternatively, electrochemical oxidation techniques can also be applied. According to the patent, this method also applies to carpet materials. In this case, waste carpets can be entirely dissolved in the aliphatic acids used, at temperatures of about 110[degrees]C and atmospheric pressures. Carpet components, such as the primary and secondary backings and adhesive binder, which are insoluble in the acid solution, can be separated by hot filtration. The acidic filtrate, containing the nylon fibers, can then be subjected to the depolymerization process.

3.2.3 Base-Catalyzed Depolymnerization

A process based on the thermal depolymerization of nylon 6 to caprolactam in the presence of potassium carbonate was invented by Nielinger et al. for Bayer AG [19]. In the context of this patent, polyamide 6 is understood to be pure polyamide 6 as well as mixtures of polyamides with over 50% (and preferably over 80%) polyamide 6 content. The polyamides and the potassium carbonate catalyst (0.5 to 2.5 wt%) are heated in an inert nitrogen atmosphere to temperatures up to 270[degrees]C-300[degrees]C. The caprolactam formed is distilled off under reduced pressures (below 100 mbar). The distilled caprolactam can then be fractionated to obtain a pure monomeric product. Unlike other processes, the Bayer process does not utilize liquid water or steam. As a result, there is no need to remove the water from the caprolactam before or during fractionation. Additionally, the caprolactam can be recovered from this process in very high yields (greater than 90%) and with high purity.

Mixtures of nylon 6 and/or nylon 6,6 can also be depolymerized by alkali or alkaline earth hydroxide catalysts. In a process developed by Moran for DuPont [20], such catalysts are used for the depolymerization of polyamide mixtures, and efficiently produce caprolactam from nylon 6 and hexamethylene diamine from nylon 6,6. This process claims to accommodate various feeds, which is expected when mixed post-consumer and industrial waste are being processed. Also, this invention can be used to process either pure nylon 6, and nylon 6,6 as well as any combination of the two. The same group has also been involved in the development of acid depolymerization routes (see previous section) and notes that acid catalysts are very effective in converting nylon 6 to caprolactam. but also form a variety of undesirable by-products in the case of nylon 6,6, such as pentyl amine, pentyl nitrile, aminocapronitrile and butyl amine. On the contrary, basic catalysts produce negligible amounts of the undesirable by-products. In the DuPont process, a mixture of nylon 6 and nylon 6,6 is added to the reactor along with a sufficient amount of a basic catalyst, such as sodium hydroxide. The reactor is maintained at a temperature of at least 275[degrees]C, and high-temperature steam is continuously added to carry the monomer products out of the non-volatile polymer melt. In the examples shown in the patent, analysis of the steam distillate from the reactor revealed only the presence of caprolactam and hexamethylene diamine.

A similar process for the depolymerization of polyamides in the presence of water and a basic catalyst was developed by Thomissen for DSM [21]. This process operates at pressures between 0.5 and 1.8 MPa and suitable basic catalysts include alkaline-earth oxides, hydroxides, and carbonates, sodium and potassium amino caproates, the sodium and potassium salts of the linear dimer or trimer of amino caproic acid or mixtures thereof. Preferably, sodium hydroxide and/or sodium amino caproate are used. The reaction is carried out in the absence of oxygen at a temperature between 250[degrees]C and 300[degrees]C, with the feed being in its molten form. The amount of the basic catalyst required is 0.2% of the weight of the polyamide. Superheated steam is passed through the polymer melt, and the products are withdrawn through the gas phase, which is then condensed in order to recover the monomer. The patent states that this process can be used for a mixture of different polyamides; however, only examples using nylon 6 and its cyclic oligomers were presented. A very high conversion (over 98%) to caprolactam was obtained in these examples. According to its inventor, the advantages of this process over commonly used acid-catalyzed depolymerization processes are the shorter reaction times required, and the elimination of phosphate wastes (from the phosphoric acid catalyst used in acid catalyzed processes).

3.2.4 Depolymerization With Water

In addition to acid-catalyzed or base-catalyzed depolymerization processes described in the previous sections, the patent literature also reveals the development of processes for the depolymerization of nylon 6 to caprolactam that require only the presence of superheated steam. One such process was invented by Sifniades et al. [22] and Braun et al. [23] for AlliedSignal. The use of carpet made from nylon 6 is emphasized in this case, since the desired monomer to be recovered is caprolactam. Nevertheless, the feed to the reactor may contain up to a total of 10% of non-nylon 6 carpet fibers, such as nylon 6,6 or PET, since these low levels of "contamination" do not interfere with this depolymerization process. Unlike most other depolymerization processes that require separation of the nylon fibers from the other carpet components, this process works efficiently even when the feed consists of the entire carpet waste material, not subjected to any prior separation. It is preferred, however, that the feed be in th e form of a melt. This is achieved by the use of an extruder, gear pump, or other similar devices. It is also recommended that the melt be briefly contacted with water in the extruder, in order to initiate the depolymerization reaction at this early stage and facilitate the overall process [24]. The unseparated polymer melt is then sent to a continuous reactor where it is heated to a temperature between 290[degrees]C and 340[degrees]C, and pressurized to a pressure between 3 and 15 atm. The melt is then contacted with a stream of superheated steam, which acts as the primary heat source. Owing to contact between the polymer and steam at elevated temperatures and pressures, the caprolactam formed is collected via the overheads as a vapor stream, and the nylon 6-depleted melt is collected at the bottoms. The overhead can then be condensed in order to recover caprolactam, which can be further purified as desired. The role of the steam, in addition to heating the reactor, is to promote the cleavage of the amide bo nds of the polymer, which initiates the depolymerization process. Pressure is also found to have an effect on the reaction. At low pressures (e.g. 1 atm), relatively large amounts of caprolactam cyclic dimer are formed, (approximately 4% of the caprolactam in the overheads). Increasing the pressure while maintaining the temperature and amount of steam constant also results in an increase in the amount of water dissolved into the nylon 6 melt, and hence accelerates the depolymerization reaction. The examples used in the patent show conversions of nylon 6 to caprolactam of approximately 90 mol%, with side products such as the dimer and ammonia being less than 1% and 3%, respectively. The patent, however, does not incorporate any recovery, recycling or further processing of the nylon 6-depleted polymer melt that exits from the reactor bottoms.

In a second AlliedSignal patent, Jenczewski et al. [25] describe a process in which the carpet is first shredded and then introduced into the reactor via a pump so that a high pressure can be maintained. Water is then introduced into the reactor, and the mixture is heated to a temperature between 250[degrees]C and 350[degrees]C at pressures between 465 and 2700 psi. Significantly higher pressures are required in this process, since it is desirable to maintain water in its liquid state. Consequently, the operating pressure has to be higher than the vapor pressure of water at the elevated temperatures used. Under these conditions, the nylon 6 fiber is dissolved in the water, and the depolymerization is initiated. At the completion of the depolymerization process, a two-phase mixture remains in the reactor, consisting of the nylon 6 depolymerization products dissolved in the aqueous phase, and an insoluble component consisting of the non-nylon 6 components of the initial unseparated carpet material. At that poin t, the reactor is cooled down in order to facilitate the separation of the soluble monomers from the insoluble components. The cooling tends to solidify the insolubles, thus making their separation much easier. These materials, which consist of polypropylene, fillers and rubber adhesives, can be sent off for heat recovery. The aqueous portion of the reaction is distilled and condensed in order to recover the dissolved caprolactam monomer, which can then be purified as necessary. High purity caprolactam yields in the order of 80% are reported in the patent. Although the inventor claims that this process is also suitable for the depolymerization of nylon 6,6 carpet, yielding adipic acid and hexamethylene diamine as the monomers, only examples using nylon 6-containing carpet were presented.

3.2.5 Pyrolysis

The depolymerization of nylon 6 in the absence of oxygen at high temperatures (pyrolysis) has been investigated as another possible route of nylon recycling. A study involving the heterogeneous catalytic pyrolysis of nylon 6 has been recently published by Czernik et al. [26]. Using a fluidized bed reactor, with nitrogen as the fluidizing gas, this group investigated the depolymerization of nylon 6 to caprolactam in the presence of an alumina-supported KOH catalyst at temperatures between 300 and 700[degrees]C. The product contained caprolactam, a dimer of caprolactam, and a series of low-molecular-weight fragments. In the absence of a catalyst, high caprolactam yields were obtained at temperatures above 450[degrees]C. However, such high process temperatures can be detrimental when applied to polymeric mixtures, when the decomposition of other plastic materials present in the carpet mixture can contaminate the caprolactam. The use of a catalyst allows for higher reaction rates at lower operating temperatures, and thus improved selectivity and caprolactam yield. Czernik et al. reported that a 5 wt% KOH/[Al.sub.2][O.sub.3] catalyst best satisfies the optimization criteria at temperatures between 330[degrees]C, and 360[degrees]C, yielding 85% of the possible caprolactam with a purity of 86%-89%. Although this study focused on the pyrolysis of pure nylon 6, it can be potentially applied for the recovery of caprolactam from carpet waste material.

Using TG/MS and closed-loop reactor analysis, Bockhorn et al. [27] have investigated the kinetics of the catalyzed and non-catalyzed pyrolysis of nylon 6 under isothermal as well as transient conditions. Kinetic parameters (apparent activation energies and reaction orders) were calculated in all cases. At 450[degrees]C the non-catalyzed thermal degradation was found to generate primarily caprolactam (92% yield). Among the by-products the cyclic dimer was the main one (approximately 4% yield), and its yield increased at temperatures above 420[degrees]C. The use of a 10 wt% ortho-phosphoric acid catalyst increases the caprolactam yield to 97%, while the degradation temperature is decreased by approximately 100[degrees]C. Once again, the main by-product was the cyclic dimer (2%) and its yield increased above 320[degrees]C. A lower apparent activation energy and a higher apparent kinetic order (163.9 kJ/mol vs. 211 kJ/mol, and 1.9 vs. 1.3, respectively) were observed as compared to the non-catalyzed case. When a base catalyst was used (60:40 molar ratio eutectic mixture of NaOH and KOH) the apparent activation energy was further reduced and the caprolactam yield was increased to 98% at 260[degrees]C, with all the by-products being below 1%.

3.2.6 Mathematical Modeling of the Depolymerization Process

Although a significant number of patents focus on depolymerization of carpet fibers, a very limited amount of fundamental work has been published in this area. Most of these fundamental studies have focused on the kinetics of the depolymerization process. Kinetic models that capture the basic chemistry taking place during depolymerization are critical for the design of process equipment and the selection of optimal operating conditions. In this section, we review some of the modeling efforts for the depolymerization of nylon 6 to caprolactam [28] as well as the ammonolysis of nylon 6,6 and nylon 6 mixtures [29].

The first attempt to describe the kinetics of the nylon 6-to-caprolactam reaction was made by Ogale in 1984 (28). Kinetic models were developed in this case for batch and semi-batch reactors, and the model predictions were compared with experimental data. These models are based on the mechanism of the hydrolytic polymerization of caprolactam and on the assumption of reversibility. The three main reaction steps are as follows (28):

Ring Opening: C + W [S.sub.1] (1)

Polycondensation: [S.sub.n] + [S.sub.m] [S.sub.n+m]W (2)

Polyaddition: C + [S.sub.n] [S.sub.n+1] (3)

where C stands for caprolactam, W for water, and [S.sub.n] for a polymer molecule containing n monomer units.

Rate expressions were written for the reverse reactions (depolymerization), and can be used in the development of reactor models (e.g. batch or semibatch). Good agreement between model predictions and experimental data was obtained. Simulation results were compared to experimental data reported in the literature. The acid-catalyzed depolymerization can also be simulated by this method, since acid endgroups are present in the reaction scheme. In this case, however, a good fit between the model predictions and the experimental data was obtained only at high acid catalyst concentrations.

A kinetic model of the depolymerization of nylon 6 and nylon 6,6 mixtures via DuPont's ammonolysis route was more recently proposed by Kalfas (29). In this case, the polyamide mixture reacts with excess ammonia at elevated temperatures and pressures, and yields hexamethylene diamine (nylon 6,6 monomer) and caprolactam (nylon 6 monomer). The reactions involved in this process are shown below (29):


This mechanism Involves amide link breakage and amide end dehydration (nitrilation) reactions, as well as ring addition and ring opening reactions for the conversion of the cyclic lactams present In nylon 6. The model proposed by Kalfas is based upon this mechanism, and was developed for an isothermal batch reactor. Simulation studies conducted with this model examined the effect of different parameters such as the ammonia concentration and the values of the equilibrium constant for the amide end dehydration reaction. Simulations were conducted for homogeneous polyamide systems (nylon 6 or nylon 6,6), as well as mixtures of the two components. However, no comparison was shown between the model predictions and any experimental results, presumably because of the proprietary nature of such data. Nevertheless, this study is valuable, since it demonstrates the effect of the different parameters on critical process characteristics such as polymer conversion and monomer selectivity.

3.3 Extraction of Polyamides

Extraction methods attempt to separate and recover the polyamides from the other carpet components without converting them back to the original monomers. Usually, the separated nylon obtained from such processes is injection molded Into other products. Several extraction processes have been proposed in the patent literature. Most of them utilize organic solvents, which, at elevated temperatures, can separate different carpet components in sequential steps. The main challenge in choosing a solvent for such extraction processes Is the required selectivity in dissolving only the nylon fibers. The more selective such a solvent is, the better the purity of the separated nylon obtained from the process. Furthermore, the potential chemical Interaction of the solvent with other carpet components, and especially the inorganic fillers and adhesives, is an issue that needs to be considered. If the solvent indeed interacts with these components, the separated nylon obtained contains impurities that alter Its properties a nd reduce Its value.

A DuPont process developed by Subramanian [30] recovers polyamides from mixtures containing "foreign" materials (such as carpet backing, glass fiber, other polymers, etc.) with minimal loss of molecular weight (i.e. reclamation without depolymerization). The procedure consists of dissolving the material containing the polyamides at elevated temperatures in a polar solvent such as a polyol or an aliphatic carboxylic acid. Ethylene glycol and propylene glycol are typical examples of the polyols used. Aliphatic carboxylic acids having 2 to 6 carbon atoms (such as acetic acid or proplonic acid) are also effective solvents. The solvents used must be anhydrous, since water reduces the solubility of the polyamide. Cooling the solution causes the dissolved polyamide components to precipitate out, and these can be further isolated by filtration.

The main problem associated with this method is the reactivity of the solvent with the polymeric material, which causes degradation of the polyamide chain. Examples used in the DuPont patent show that prolonged exposure of nylon to a glycol solution at the boiling point of the glycol significantly reduces Its molecular weight. In fact, the decrease in molecular weight, and thus the degradation of the polyamide chain, is proportional to the exposure time. To prevent this reaction, the process includes a quenching step following the dissolution of the polyamide components. During quenching, a low enough temperature is reached so that any potential reaction between the polyamide and the solvent is inhibited. The two main advantages of the quenching technique, as described in the DuPont patent, are the recovery of the polyamides with minimal degradation (since the exposure time at high temperatures is minimized), and potential reuse of the solvent. Since the solubilities of different polyamides vary with temperat ure, controlling the quenching temperature enables one to precipitate the polyamides sequentially, thus separating them from mixtures. However, In the examples shown in the patent, the only results presented are from the recovery of nylon 6,6 from carpet and automotive radiators.

A process based on extraction with pure and mixtures of aliphatic alcohols, as well as mixtures of water and aliphatic alcohols, has been developed by Booij et al. for DSM [31]. As in the DuPont process, the face fibers are separated from the rest of the carpet components by selectively dissolving the fibers in an appropriate solvent at high temperatures. The solvent containing the polyamides is then physically separated from the insolubles, and the nylon is precipitated out of the solution by lowering the temperature. Once again, separation of different types of nylon fibers can be achieved by controlled decrease of the temperature. Different solvents, such as [C.sub.1]-[C.sub.12] aliphatic alcohols, and monohydric and dihydric alcohols can be used, with methanol and ethanol being designated as the preferred extraction agents. The advantage of this process over the previous DuPont process is that there is minimum interaction between the polyamides and the solvent and it also allows separation of the dyes fro m the fibers. The DSM process applies to carpets containing a variety of components (such as polypropylene, jute, fillers, rubbers, and adhesives) and it generates a directly reusable nylon, since the polyamide chains are not decomposed during the process.

Along the same lines, extraction of polyamides with a glycolic compound was also proposed by Stefandl [32]. The Stefandl process applies well to mixed nylon carpet scrap, which can be separated based on the relative solubility of the nylon fractions as a function of temperature. Nylon 6, for example, is soluble in glycerol at 155[degrees]C, whereas nylon 6,6 is soluble at 195[degrees]C. Carpet scrap containing both nylon 6 and nylon 6,6 is first dissolved in glycerol at 165[degrees]C to solubilize the nylon 6 component. The solution obtained is withdrawn In a separate vessel, where, by rapid cooling to 40[degrees]C, nylon 6 precipitates. The residue containing nylon 6,6 along with the rest of the carpet components is dissolved in a second vessel at 200[degrees]C. The precipitation of nylon 6,6 from the hot solution is also achieved by rapid cooling to 40[degrees]C. Both polymeric precipitates are washed with water to remove traces of glycerol, and then dried under vacuum. Since, at high temperatures, nylon ca n undergo oxidation with oxygen and/or hydrolysis with water present in the solvent, an inert atmosphere is required. Relative viscosity and differential scanning calorimetry were used to determine the purity of the obtained polyamides in the patent examples. According to the results, both recovered nylon samples exhibited single melting points corresponding to those of the pure polymers.

An extraction method that uses supercritical [CO.sub.2] as the extraction agent to separate individual polymeric materials from carpet scraps has been developed by Sikorski for the Georgia Tech Research Corp. [33]. Although other extraction agents such as ethane and n-butane are also capable of performing the supercritical stage-wise separation, [CO.sub.2] is preferred because of its low cost, low chemical reactivity, and low toxicity. The carpet components in this case are gradually dissolved in supercritical [CO.sub.2] in the increasing order of their melting points. Carpet scrap is introduced in an extraction vessel along with supercritical [CO.sub.2]. Temperature and pressure are adjusted so that the carpet component with the lowest melting point is dissolved first In the extraction fluid. The solution obtained is withdrawn In a separation vessel, where, by decreasing the temperature and pressure, the carpet component is separated from the extraction fluid, In subsequent steps, the temperature and pressu re are systematically increased so that each individual component of the carpet scrap mixture can be eventually recovered. The recovered extraction fluid can then be recirculated. The carpet components are separated in the following order: grease and lubricants, latex (styrene-butadiene rubber), backing (polypropylene), and, finally, face yarns (nylon 6 or 6,6).

Finally, a method that combines the use of a solvent and a supercritical fluid antisolvent (non-solvent) to separate nylon 6,6 from other carpet materials has been recently reported by Griffith et al. [34]. This process takes place in three steps. First, the carpet material is treated with a solvent (e.g. formic acid solution) to separate the nylon 6,6 from the other carpet components. In the second step, nylon is precipitated out of this solution by the use of supercritical [CO.sub.2]. The operating conditions are chosen such that nylon 6,6 is soluble only in formic acid and [CO.sub.2] and formic acid are miscible. Contact of the two solvents causes precipitation of nylon 6,6, due to the lowering of the solution's solvent strength. In the third and final step, the pressure is lowered, formic acid and [CO.sub.2] separate and are recycled back into the process, and nylon is recovered in powder form (particles smaller that 20 [micro]m). The operating conditions, such as solvent and nylon concentration, and ups tream and downstream pressures have been found to have little to no effect on the mean nylon particle size and particle size distribution.

3.4 Melt Blending

Tertiary recycling of carpet materials consists of reusing the carpet as a whole, by melting or extruding It to form a blended mixture that is subsequently used in the production of injection molded polymers and thermoplastics. An advantage of this process is low cost, since no expensive separation and depolymerization procedures are required. However, depending on the chemical compatibility of the waste carpet components melt blending usually results to low quality melt-blended plastics with limited uses. Kotlair and Fountain [6] developed a relatively simple process that converts carpet waste (industrial and post consumer) to fibrous composites that can be used as synthetic wood." The carpet waste is first converted to individual "shreds" approximately one-half inch in length. These shreds (still containing the polypropylene backing and latex adhesive) are then coated with low viscosity prepolymers used as structural adhesives. Examples of such prepolymers include phenol-formaldehyde and epoxy resins. The r esin-coated shreds are then cured at approximately 150[degrees]C-190[degrees]C and pressed into the fibrous composites. Methods of curing include high temperature pressing, or the use of a vented extruder. A large variety of composite materials can be produced following this approach. Their properties and quality depend on the composition of the original carpet (carpet fiber to backing ratio) and the type and amount of resin used. These parameters also affect the process conditions and the potential uses of the final recycled product.

Another method of carpet scrap recycling, which converts the heterogeneous carpet into a homogeneous thermoplastic blend has been developed by Young et al. for Lear Corporation [3]. This process focuses primarily on automotive carpet scrap. The typical automotive carpet consists of nylon 6 or nylon 6,6 face fibers, a polypropylene or polyester primary backing, a poly(ethylene-co-vinyl acetate) (or EVA) precoat, and a [CaCO.sub.3] or [BaSO.sub.4] filled poly(EVA) back coat. This composition differs from the composition of typical residential carpet in the type of backing and the amount of inorganic particulate fillers (i.e. [CaCO.sub.3] and [BaSO.sub.4]) used. Increased amounts of fillers are added in this case as sound insulators. In the Lear Corporation process the carpet scraps are first shredded and granulated before being fed to a twin-screw extruder. The extruder is operated at a temperature between 210[degerees]C and 230[degrees]C. The melt produced is then pelletized and used in injection molding appl ications. Such applications include the production of other automotive parts, such as flexible floor mats to replace the currently used vinyl or rubber mats, or automotive carpet backing with superior sound-insulating properties. Additionally, rigid materials such as door panels and truckliners can be manufactured from the recycled material. Although the melt-blended material already behaves as a thermoplastic, the addition of a compatlbllizing agent greatly Improves Its quality. Generally, thermoplastics resulting from a mixture of polymers produce melts with poor mechanical properties because of incomplete mixing of the different components. Compatibilizing agents promote such mixing, and hence, improve the overall characteristics of the blend and of the resulting injection molded products. EVA, which is already present in automobile carpets, is an example of such a compatibillzer.

The compatibilizatlon between nylon 6 and polypropylene was investigated by Datta et al. (35). This group studied the effect of the addition of a compatibilizer (maleic anhydride grafted polypropylene, MA-g-PP) on the final properties of a blend of nylon 6 and polypropylene. They used two samples of shredded carpet waste containing nylon 6 and polypropylene face fibers, with and without calcium carbonate-filled SBR The properties of the final blend were optimized by varying the extrusion and compression molding parameters, and the percentage of the compatibilizer added. The results show that this approach can be used to produce resins with good mechanical properties.

The compatibilization of nylon 6 and polypropylene was also investigated by Dagli et al. (36). An acrylic acid grafted modified polypropylene (PP-g-AA) was used as the compatibilizer in this study. The results indicate the formation of a grafted copolymer, which suggests that PP-g-AA is another effective compatibilizer that can be used for this type of application.

An alternative method for the melt blending of waste carpet has been developed by David et al. (37, 38) for Monsanto. This process also uses unseparated, postconsumer and post-manufacture waste carpet and melts It at high temperatures and pressures to produce a thermoplastic material. The carpet waste is first cleaned to remove excess dirt, which can have a negative effect on the quality of the final product, and is then introduced into a conventional twin-screw extruder at a shear rate of 200 to 400 [sec.sup.-1], a temperature between 250[degrees]C and 350[degrees]C, and a pressure of 350 to 450 psi. The intense mixing taking place at these temperatures and pressures converts the carpet into a flowing homogeneous mixture containing all of its components (face fiber, backing and SBR adhesive). The process conditions result In effective blending of all the carpet materials, and thus, a compatibilizer is not required (although It may be optionally used). The blend Is then cooled, and cut into easily transportable pellets or ch ips. According to the inventors of this process, the resulting material exhibits thermoplastic characteristics. Consequently, It can be used to produce high quality injection-molded products, either directly or mixed with other thermoplastics.

3.5 Separation of Carpet Waste Components

As was pointed out in section 3.2, one of the most attractive methods for the recycling of carpet waste is the depolymerization of the tufted nylon fibers to their constituent monomers. Depolymerization processes usually require the nylon fibers to be free of any other carpet material, and hence, a separation step is necessary prior to the depolymerization. Such a separation step can increase the cost of depolymerlzationbased recycling processes significantly, making most of them prohibitively expensive. The different available separation methods are reviewed in this section. It should also be noted that current development efforts in the carpet recycling field focus on recycling routes that utilize the carpet scrap as a whole, or require minimum separation.

In 1993 Hagguist and Hume (39) invented the "Carpet Reclaimer," an apparatus that disintegrates and separates post-consumer carpet into Its principal components. The carpet is first moved along a conveyor belt to a rotating cutter, which shears the loops of the continuous carpet yarn (tuft fiber). The carpet is then sent to another station where it is treated with air, water, steam and several chemical solutions under conditions of high temperature and pressure (the choice of which depends on the type of the carpet processed). The purpose of this step is to "debond" the latex binder from the secondary backing of the carpet. The debonded secondary backing is then removed using a series of mechanical impingement devices that strip the backing off the rest of the carpet matrix. Subsequently, the carpet is treated with cup brushes and rotating high-pressure nozzle heaters that loosen and remove the SBR adhesive from the rest of the carpet, leaving the cut tufted fibers and primary backing intact. The final step of this process is the removal of the pile fibers from the primary backing. This is achieved with the help of a number of rotating mechanical combs suspended below the primary backing, which separate the nylon fibers and deposit them into a collection chamber for further processing.

A significantly different and much simpler carpet separation process has been developed by Sharer for JPS Automotive Products (40). During the first step the carpet is cut into even-sized square pieces between two and four inches in length. These pieces are then fed to a granulating apparatus that further grinds them with the help of high-speed cutting rolls. This granulating process also initiates the separation of the fibers from the backing materials. The fine pieces exiting the granulating apparatus are then transferred to an elutriator that separates the backing from the fibers by employing an air stream, which carries off the fibrous material vertically through a separate conduit, At the same time the heavier backing material gravitates onto a conveyor located below. The fibers and the backing material are then transported to storage containers. We should point out that this process does not separate the latex binder from the backing material. However, this is not significant if the recycled backing Is remelted with a virgin feed stream of backing material and remolded onto new carpets as secondary backing. Unfortunately. no experimental results regarding the effectiveness of the separation were presented in the patent, which makes evaluation and comparison with other available methods very difficult.

A similar method of carpet separation has been also developed by Sferrazza et al. for BASF Corporation and Shred-Tech Limited [41]. Once again, the different carpet components are separated by initially shredding and hammering the waste carpet in order to loosen the nylon fibers from the backing materials of polypropylene and SBR adhesive. The shredded carpet is then sent to a primary reduction unit. In this unit, the carpet fibers are removed from the backing materials. This is achieved by subjecting the carpet to a hammer mill-like process, where the more brittle backing materials get dismantled from the fibers. The heterogeneous mixture is then sent to an air elutriator where the lighter nylon fibers (overhead stream) are separated from the heavier backing components (bottom stream). The nylon fibers are then screened to separate any impurities of backing materials present in the overhead stream, and transferred to a metal particulate removal unit, which magnetically removes any ferrous metal particles pre sent in the fibers. The primary bottom stream (backing materials) is re-sent to a grinding/reduction unit, where a second attempt is made to separate any remaining fibers. The BASF/Shred-Tech process represents an extensive yet effective way of separating the nylon fibers from the backing materials. However, no separation of the polyethylene backing material from the SBR adhesive is achieved in this process.

A process that has been developed by Dilly-Louis et al. for Zimmer AG also starts with the pre-shredding of the waste carpet material (42). In this case, however, an aqueous suspension is prepared by grinding the shredded carpet in a commercial cutting mill. During grinding, the carpet components separate, and the particles in suspension are uniform (i.e. each particle consists mainly of either the nylon fibers, the primary backing, or the secondary backing, but not combinations of them). The presence of water controls heat generation during the grinding process and cleans the carpet materials of any impurities. The density of the suspension is adjusted to a level between the densities of the two main carpet components by adding an aqueous salt solution. Separation of the main components of the carpet material is achieved during this step. The concentration of solids (i.e. carpet materials) is approximately 10-30 wt%. The suspension is passed through a 5 to 8 mm filter and fed to a receiving tank where the co ncentration of the solids is lowered to between 3 and 10 wt%. A salt solution of calcium chloride or potassium carbonate is added to the receiving tank in order to adjust the density of the suspension to a level between the highest and second highest density of the main carpet components. The suspension is then sent to a double-cone full-jacketed screw centrifuge, which can separate phases with different densities. As a result, the heaviest of the three carpet components (usually the fillers) are separated from the suspension. The suspension containing the other two solid components (i.e. tufted fibers and polypropylene primary backing) is then, once again, passed through a receiving tank, where its density is adjusted for a second time, and re-sent to the centrifuge for separation. Carpet components recovered in this process usually have a purity of over 90%.

A method similar to the one described above has been developed by Costello and Keller for AlliedSignal and DSM (43). This process also utilizes pre-shredding and/or impact size reduction of the waste carpet materials. The shredding process reduces the size of the carpet and initiates the separation of the latex adhesive from the rest of the carpet materials. The shredded particles are then subjected to screen separation to remove the adhesive and post-consumer dirt from the nylon fibers and primary backing. Next, the separated material is washed to remove additional amounts of the adhesive. Finally, the residual carpet components are mixed with an aqueous solution and introduced into a hydrocyclone, where they are separated. The components that are heavier than water (nylon 6 and nylon 6,6) flow out through the bottom of the hydrocyclone, whereas the lighter components (such as the polypropylene backing) flow through the overflow of the hydrocyclone along with the water.

An outline of a proprietary carpet separation process used by DuPont was presented by Smith and Gracon at Recycle '95 (11). In this case, the post-consumer carpet is first put through a dry separation process where it is cut, shredded, hammer milled, separated by size and further ground to 1.5 mm pieces. The carpet component mixture formed is then mixed with water, washed and wet screened prior to introduction into a density separation process. This process produces a stream of 98% pure polypropylene. The separated nylon stream is further dried, extruded and melt filtered to produce a 98.5% pure nylon resin that can be sent to depolymerization units or other recycling processes.

The Zimmer, AlliedSignal and DuPont separation processes are similar in that they all use an initial dry separation process (which involves shredding and grinding the carpet) followed by a specific gravity based wet separation. High purity nylon fibers and polypropylene backing are the separated components. It should be pointed out that in addition to the nylon fibers, the high purity polypropylene co-produced during the separation could also be easily recycled or reused.


In this final section we examine some of the recycling processes that have been put into practice at either the pilot or commercial scale, and attempt to offer a synopsis of the current state of the carpet recycling industry and the challenges that lie ahead. Information in this section has been obtained from special reports In recent issues of High Performance Plastics (44) and the International Fiber Journal (5, 45), and Carpet and Rug Indus try [46].

Reconditioning of carpets is a procedure in which an old discarded carpet is cleaned, trimmed and re-colored. The "reconditioned" carpet is then sold for reuse, making the reconditioning process one of the most cost-effective approaches for recycling. Independent operators and carpet manufacturers (such as Milliken) are involved in these efforts. The extent of such operations is currently limited.

Several companies such as Monsanto. United Recycling Inc., and Collins & Alkrnan are Involved in the melt blending of carpet waste. United Recycling Inc. produces two grades of thermoplastic material from nylon 6 carpets with added polypropylene. A similar process Is carried out by Monsanto, which produces molding compounds from nylon 6,6. Collins & Aikman Is limited to recycling only Its own carpet tiles. In its carpet construction process, Collins & Aikman has replaced the primary and secondary backing, and the filler with vinyl, while the face yam is made of nylon. During the extrusion process, post-consumer low-density polyethylene (LDPE) is added for improved finish of the recycled product. The resulting material has properties similar to wood, and can be used as industrial block flooring, parking blocks and plastic wood.

The depolymerization of waste carpet nylon fibers is currently the most active route of carpet recycling. The processes described in section 3.2 involve the separation of the nylon face fibers from the other carpet components and the depolymerlzation of these fibers to the corresponding monomer. AlliedSignal, DSM, BASF and DuPont (all fiber producing companies) are actively involved in carpet recycling via depolymerization routes.

BASF recycles nylon 6 fibers that are depolymerized to caprolactam in the presence of steam and an acid catalyst. The other carpet materials (i.e. backing and latex adhesive) are incinerated. The ash produced Is used as filler for other plastic products. BASF accepts for recycling all carpets made under its "Six Again" program, and has established for this purpose four used carpet collection centers in the U.S. and two in Canada. DuPont also recycles used carpets from seventy-six collection centers, which are first sent to a separation unit in Chattanooga, Tennessee. The nylon is then sent to a pilot ammonolysis plant in Kingston, Ontario, for depolymerization. Some of this DuPont recycled nylon is incorporated in mineral-reinforced nylon 6,6, reused by Ford Motor Co. in engine air-cleaning housings for approximately three million vehicles each year [47].

AlliedSignal and DSM are currently operating the largest carpet recycling program as a 50/50 joint venture. A new $80 million plant called "Evergreen Nylon Recycling" was built for this purpose in Augusta, Georgia, and began operation in November 1999. The plant has an annual deploymerization capacity of 200 million lbs of nylon 6-containing post-consumer carpet waste, and generates 100 million lbs of caprolactam (48). The Evergreen project also recycles the non-nylon carpet materials that are separated during the process. Evergreen claims that the quality of its recycled caprolactam is equivalent to that of the virgin material, with the possibility of repeated recycling of the same material without any detrimental effects on its properties [48]. The AlliedSignal share of the recovered caprolactam is included in polymeric materials and fibers trademarked under the name Infinity Renewable Nylon, whereas DSM sells its share of caprolactam into caprolactam markets.

Major carpet recycling efforts are also under way in Europe. A three-year joint European project called RECAM (recycling of carpet materials) was funded by the European Commission to explore the economical feasibility of developing a closed-loop system for recycling of post-consumer and industrial carpet waste. The whole process consists of collection, identification, size reduction, mechanical recycling, chemical recycling, and energy recovery steps. The collection step was studied in Germany at a pilot plant facility. The identification and sorting step was done using a portable apparatus developed by DSM and AlliedSignal that uses near-infrared technology to identify, within seconds, the fiber type. The size reduction and pelletizing step was performed by Recotex at a pilot plant also located in Germany. The mechanical recycling step utilized equipment developed by LaRoche and included the recovery of wool for reuse in non-wovens. Finally, the chemical recycling step used the processes developed by DSM and Enichem to produce caprolactam from nylon 6 face fibers. It is expected that on a long-term basis, the RECAM project will prove the cost effectiveness of such an approach.


The authors acknowledge useful discussions on the carpet recycling topic with Dr. George Kalfas, and express their gratitude to the South Carolina Hazardous Waste Management Research Fund for financial support of the preparation of this review.

(*)Corresponding author. Electronic mail:


(1.) M. Reisch, Chemical and Engineering News. 20 (1998).

(2.) R. Wolff, International Carpet Yearbook, 20 (1996).

(3.) D. C. Young. S. J. Chlystek, R. Malloy, and I. Rios, Lear Corporation. Recycling of Carpet Scrap, U.S. Patent 5,859,071 (1999).

(4.) K. V. Datye. Indian Journal of Fiber & Textile Research, 16, 46 (1991).

(5.) J. Hagewood, International Fiber Journal, 48 (1999).

(6.) A. M Katlair and D. P. Fountain, Georgia Tech., "Synthetic wood from waste" fibrous materials, U.S. Patent 5,629,939 (1997)

(7.) Y. Wang, J. Materials in Civil Engineering, 103 (1997).

(8.) S. R. Agrawal. Popular Plastics, 11 (1975).

(9.) L. A. Dmitrieva, A. A. Speranskii, S. A. Krasavin, and Yu. N. Bychkov. Fibre Chemistry, 17 (4), 229 (1986).

(10.) R. J. McKinney, E. I. DuPont de Nemours, "Ammonolysis of Nylon," U.S. Patent 5,302,756 (1994).

(11.) R. A. Smith and B. E. Gracon, "Polyamide 66 and 6 Chemical Recycling," Recycle '95. Davos, Switzerland (1995).

(12.) R. J. McKinney, E. I. DuPont de Nemours, "Lewis Acid Catalyzed Ammonolysis of Nylon," U.S. Patent 5,395,974 (1995).

(13.) J. A. Hendrix, M. Booij, and Y. H. Frentzen, DSM N.V., "Depolymerization of Polyamides." U.S. Patent 5,668,277 (1997).

(14.) L. Crescentini, W. B. Blackman, Jr., J. D. DeCaprio, W. B. Fisher. R. J. Lilley, Jr., and J. W. Wagner, "Recovery of Caprolactam from Nylon 6 Oligomers," U.S. Patent 4,311.642 (1982).

(15.) T. F. Corbin, E. A. Davis, and J. A. Dellinger. BASF Corporation. "Reclaiming Epsilon-Caprolactam from Nylon 6 Carpet," U.S. Patent 5,169.870 (1992).

(16.) R. Kotek, BASF Corporation, "Semi-continuous Depolymerization of Nylon 6 Polymer," U.S. Patent 5,294,707 (1994).

(17.) E. F. Moran, E. I. DuPont de Nemours, "Conversion of Nylon 6 and/or Nylon 6,6 to Monomers." U.S. Patent 5,310,905 (1994).

(18.) E. F. Moran and R. J. McKinney, E. I. DuPont de Nemours, "Conversion of Nylon 6 and/or Nylon 6,6 to Adipic Acid," U.S. Patent 5,468,900 (1995).

(19.) W. Nielinger, E. Ostlinning, K-J Idel, D. Freitag, and H-J. Buysch, Bayer Aktiengesellschaft. "Production of Caprolactam by Thermal Depolymerization of Polyamide 6," U.S. Patent 5,233,037 (1993).

(20.) E. F. Moran, E. I. DuPont de Nemours, "Nylon Component Reclamation," U.S. Patent 5,266.694 (1993).

(21.) P. J. H. Thomissen, DSM N.y., "Depolymerization of Polyamides," European Patent 0,875,504, Al (1998).

(22.) S. Sifniades, A. B. Levy, and J. A. J. Hendrix, AlliedSignal Inc., and DSM N.y., "Process for Depolymerizing Nylon-containing Waste to Form Caprolactam." U.S. Patent 5,681,952 (1997).

(23.) M. Braun, A. B. Levy, and S. Sifniades, Polym.-Plast. Technol. Eng., 38(3), 471 (1999).

(24.) S. Sifniades and A. B. Levy, AlliedSignal Inc., "Monomer Recovery From Multi-Component Materials," U.S. Patent 5.457.197 (1995).

(25.) T. J. Jenczewski. L. Crescentini, and R. E. Mayer, AlliedSignal Inc., "Monomer Recovery From Multi-Component Materials." U.S. Patent 5,656,757 (1997).

(26.) S. Czernik, C. C. Elam, R. J. Evans, R. R. Meglen, L. Moens, and K. Tatsumoto, J. Analytical and Applied Pyrolysis, 46, 51 (1998).

(27.) H. Bockhorn, A. Hornung, U. Hornung, and J. Wichmann, Termochimica Acta, 337, 97 (1999).

(28.) A. A. Ogale, J. Applied Polymer Science, 29, 3947 (1984).

(29.) G. A. Kalfas, Polymer Reaction Engineering, 6(1), 41 (1998).

(30.) P. M. Subramanian, E. I. DuPont de Nemours and Company, "Recovery of Polyamide Using a Solution Process," U.S. Patent 5,430,068 (1995).

(31.) M. Booij, J. A. J. Hendrix, Y. H. Frentzen, and N. M. H. Beckers, DSM N.V., "Process for Recovering and Recycling Polyamide From Carpet Waste," U.S. Patent 5,840,773 (1998).

(32.) R. E. Stefandi, "Carpet Recycling Process for Nylon Containing Carpeting," W.O. Patent 98/35998 (1998).

(33.) M. E. Sikorski, Georgia Tech Research Corporation, "Recycling of Polymeric Materials from Carpets and Other Multi-component Structures by Means of Supercritical Fluid Extraction," U.S. Patent 5,233,021 (1993).

(34.) A. T. Griffith, Y. Park, and C. B. Roberts, Polym.-Plast. Technol. Eng., 38(3), 411 (1999).

(35.) R. K. Datta, M. B. Polk, and S. Kumar, Polym.-Plast. Technol. Eng., 34(4), 551 (1995).

(36.) S. S. Dagli, M. Xanthos, and J. A. Biesenberger, Emerging Technologies in Plastics Recycling, 240 (1992).

(37.) D. J. David, J. L. Dickerson, and T. F. Sincock, "Thermoplastic Composition and Method for Producing Thermoplastic Composition by Melt Blending Carpet," U.S. Patent 5,591,802 (1997).

(38.) D. J. David, J. L. Dickerson, and T. F. Sincock, Monsanto Company, "Thermoplastic Composition and Method for Producing Thermoplastic Composition by Melt Blending Carpet," U.S. Patent 5,294,384 (1994).

(39.) J. A. E. Hagguist and R. M. Hume, "Carpet Reclaimer," U.S. Patent 5,230,473 (1993).

(40.) P. C. Sharer, JPS Automotive Products Corp., "Method for Recycling Carpet Forming Components From Waste Carpet," U.S. Patent 5,518,188 (1996).

(41.) R. A. Sferrazza, A. C. Handermann, C. H. Atwell, and D. K. Yamamoto, BASF Corporation; Shred-Tech Limited, "Carpet Recycling Process and System." U.S. Patent 5,535,945 (1996).

(42.) W. Dilly-Louis, J. Seelig, and R. Wolf, Zimmer Aktiengesellschaft, "Process for the Separation of Carpet Materials," U.S. Patent 5,598,980 (1997).

(43.) M. R. Costello and B. R Keller, AlliedSignal Inc./DSM N.V., "Process for Separation and Recovery of Waste Carpet Components." U.S. Patent 5,722,603(1998).

(44.) High Performance Plastics, 6 (1999).

(45.) H. Miller, International Fiber Journal, 54 (1999).

(46.) J. Herlihy, Carpet and Rug Industry, 17 (1997).

(47.) Modern Plastics, 74(3), 14 (1997).

(48.) R. Westwervelt, Chemical Week, 161(44), 18 (1999).
COPYRIGHT 2001 Society of Plastics Engineers, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2001 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Publication:Polymer Engineering and Science
Geographic Code:1USA
Date:Sep 1, 2001
Previous Article:Modeling Heat Transfer in Screw Extrusion With Special Application to Modular Self-Wiping Co-Rotating Twin-Screw Extrusion.
Next Article:A Review on Interface Modification and Characterization of Natural Fiber Reinforced Plastic Composites.

Related Articles
A recycling first: carpets!
New recycled content standards proposed.
Separation or compatibilization? Alternative recycling solutions emerge.
C&A's carpet recycling program tops 100,000,000 pounds.
Creative carpeting cuts costs.
Recyclers now targeting higher-value engineered resins.
Textile recycler saved by carpets: this is a tale of globalization with a happy ending. One of the world's two largest carpet recyclers got into the...
Circle of life.

Terms of use | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters