This project is aimed at the discovery of new methods for the generation of acyl derivatives through the use of carbon monoxide and transition metal catalysts. The work is aimed specifically at replacing phosgene as an acyl source in the synthesis of thiocarbamates, ureas, and isocyanates as intermediates, and at replacing acid chlorides in amide syntheses. These types of processes are central in the synthesis of hundreds of photographic intermediates used at Kodak, and have been identified as an area where environmentally benign chemistry could have a significant impact. The goal of this project will be to develop new transition metal based catalysts for the preparation of acyl derivatives that will avoid noxious chemicals and eliminate waste products. Students and postdoctoral fellows will gain valuable experience working in industry in process development through a collaboration with Dr. Ronald Valente in Kodak's Chemicals Development Division.
The Toxic Release Inventory required by the EPA has made the public acutely aware of the vast quantities of chemicals released into the environment each year as byproducts of industrial chemical manufacturing. The disposal of these byproducts not only are a detriment to the environment but also represent a waste of valuable natural resources. Consequently, there is national pressure to have university and industrial researchers join together in the development of new environmentally benign chemical synthesis and processing methodologies.
To this end, the Eastman Kodak Chemical Process Research Team has joined with the University of Rochester to initiate a cooperative effort investigating the replacement of phosgene with carbon monoxide. This Process Research Team is responsible for the development of chemical processes which are used to manufacture approximately 1,100 metric tons of chemicals annually in Kodak Park, Rochester, NY.
The primary use of these synthetic chemicals are for photographic film and paper product applications. The structural requirements of photographically useful compounds are such that they must contain both a hydrophilic portion for aqueous solution chemistry and significant lipophilic character to prevent migration between adjacent layers when coated in a photographic system. The lipophilic requirement is most often accomplished by acylation of an amino substituted latent chromophore with a long chain acid chloride.
The manufacture of acid chlorides as well as carbamoyl chlorides and the corresponding amides, urethanes, and carbamoyl thioesters represent 50-60% of the annual Kodak Park volume. Thus, novel and efficient methods to prepare these intermediates would have greater impact in reducing process-generated waste being generated at Kodak Park.
Methodology for Acid Chloride and Acid Chloride Equivalent Manufacture
a. Phosgene. Phosgene is an essential building reagent in the chemical industry but the environmental hazards upon accidental release are significant. Eastman Kodak has attempted to limit the use of phosgene and would like to eliminate its use altogether. Currently, Kodak manufactures carbamoyl thioester 1 (~10,000 kg/yr.) as shown below.
Its synthesis employs sequential condensation of secondary amine and a mercaptotetrazole with phosgene. At current levels, 80% of the total Kodak phosgene use is required to produce the annual requirement of 1. Thus, an alternative methodology to synthesize 1 would result in a major reduction of phosgene use at Kodak Park.
b. Thionyl Chloride and Phosphorus Oxychloride. The bulk methodology used at Kodak Park for acid chloride manufacture is thionyl chloride based. Efficient emission air treatment minimizes the byproduct HCl and SO2 off-gases. This procedure involves neutralization with caustic and subsequent bioremediation of the aqueous stream. However, this treatment increases the overall operating costs and raw material usage. Additionally, thionyl chloride based acid chlorides introduce variable levels of sulfur and sulfur-containing impurities to the final compounds. The sensitivity of silver halide based imaging toward sulfur poisoning is well documented and a significant cause for concern for film makers. This sulfur contamination may require further purification of chemicals which add to raw material usage and hinder robust manufacturing efforts.
Alternative use of phosphorus oxychloride is less preferred and not a viable alternative. Typically, its aqueous waste stream after neutralization is limited by the efficiency of the waste water treatment facility in digesting phosphorus to comply with standards set by the EPA for clean water. These limits preclude the possibility of switching from thionyl chloride to POCl3.
c. Potential Hazards and Process Efficiency. The use of these reagents also requires transportation to the plant as well as inventory volumes and capability. There are associated concerns around the potential for accidental spills and releases to the environment during the transport and storage of these reagents. Theoretically, carbon monoxide could be generated on-site and hard piped directly into our reactor systems.
Finally, the described traditional organic chemical methods result in measurable waste which must be pre-treated before being sent to the waste water treatment facility. A substantial effort has been ongoing at Kodak Park to quantify the hidden costs associated with waste treatment. This effort includes complete waste characterization and a kilogram waste to kilogram product ratio which measures the efficiency of our chemical processes. The waste to product factor for acid chloride and related chemistries is often quite low. However, given the total volume of chemicals being manufactured reduction in byproduct waste would result in significant improvements in overall pollution abatement efforts.
Thus, an alternative synthetic methodology for industrial scale acid chloride and equivalent manufacture without the use of traditional organic reagents would offer Kodak a tremendous advantage in terms of cost, quality, waste reduction at its source, and environmental risk. The joint effort has been initiated by Kodak with the University of Rochester based upon the technical expertise and leadership of its faculty and to encourage close ties with the neighboring intellectual community. The following section provides preliminary studies that form the basis for this proposal.
2. Preliminary Results
The results described in this section were obtained by a postdoctoral fellow whose support for a period of one year was supplied through a seed money grant from Kodak. The work was to have some relation to environmentally benign chemistry that might have an impact on current chemical synthesis. Meetings with several chemists from the company led to the identification of many general processes that could be targeted for the development of new, 'greener' methodology. The area where our expertise in transition metal chemistry appeared to have an obvious applicability was in the area of phosgene replacement chemistry.
As mentioned earlier, current large scale synthetic methods for the synthesis of carbamoyl esters and thioesters makes use of the exothermic reaction of amines with phosgene to generate chloroformamides, which are then coupled with thiols to generate carbamoyl thioesters. This class of product is typical of the types of couplings that are employed in industry, with reactions typically run on the 100-500 kg scale and productions of up to 10,000 kg/year. Several hundred different chemicals are produced for the photographic industry in this fashion. Equation 1 shows an example of the type of reaction that is currently carried out. In addition to the use of large quantities of phosgene, HCl is removed by combination with an amine base that is then either incinerated or 'sewered'.
(1)We chose this type of coupling reaction as our initial target for inventing a new type of reaction that would avoid the problems associated with phosgene use. In approaching the synthesis of new acyl derivatives, the enhanced reactivity of carbon monoxide toward nucleophiles upon coordination to a metal seemed to present many possibilities. Attack of a heteroatom nucleophile on coordinated CO generates an acyl derivative that has Fischer carbene character, enhancing the stability of this type of species (see Scheme I below). This carbene complex could be susceptible to a second nucleophilic attack, and the intermediate formed could then eliminate a difunctionalized acyl derivative. The metal would be reduced by 2 electrons in such a process, and would have to be reoxidized if the process were to be catalytic in metal. Such an oxidation could be accomplished by reduction of protons (from deprotonation of the nucleophiles) to H2, addition of an external oxidant, air oxidation, or the use of an oxidized form of the nucleophile (i.e. Nu-Nu + 2 e- ® 2 Nu-).
With this type of a cycle in mind, we set out to examine the reactions of several metal carbonyl complexes with nitrogen and sulfur nucleophiles. The choice of thiocarbamates served as an excellent test target in these initial studies, as their formation would require that no single nucleophile dominate the chemistry. If nitrogen nucleophiles proved to be much more reactive than sulfur nucleophiles, then ureas would be formed instead of thiocarbamates. An enhanced reactivity of sulfur nucleophiles would have similar deleterious consequences with regard to the production of cross-coupled products. Successful achievement of catalysis would require one nucleophile to be more reactive with the metal carbonyl complex, and the other nucleophile to be more reactive with the Fischer carbene complex.
Scheme I:
Some success was finally obtained when a coordinatively unsaturated non-carbonyl complex was used under an atmosphere of CO. A THF solution of Pd(PPh3)2Cl2 was treated with 3 equivalents of secondary amine followed by 1 equivalent of thiol. (Excess amine was used as a base.) After 24 h, the desired thiocarbamate product was isolated in 60% yield (eq 3). Similar results were obtained using Pd(PPh3)2Br2. Use of the chelating phosphine derivative Pd(diphos)Cl2 resulted in no thiocarbamate product. The nickel complexes Ni(PPh3)2Cl2 and Ni(PPh3)2Br2 also gave good yields of thiocarbamate (~70%). Once again, the diphos derivative Ni(diphos)Cl2 showed no activity.
(3)In an effort to increase product yields, the use of amine as solvent was examined. The reaction now showed catalytic formation of thiocarbamate product (4-6 turnovers based on metal). As these reactions were conducted in septum capped flasks in a laboratory hood under a slow purge of CO, it seemed possible that air might have been responsible for turning over the metal. Intentional addition of air as an oxidant was examined and found to result in a substantial increase in catalytic activity. In one experiment, an approximately 50% conversion of the thiol limiting reagent was converted to product (disulfide accounted for the remainder), as shown in eq 4.
(4)These results are very promising, demonstrating the feasibility of this approach for the synthesis of thiocarbamates. The reaction has the advantages of not requiring a solvent that must be reclaimed and not using potentially harmful phosgene as a reagent. Consequently special production facilities would not be required with any process that derives from this type of catalysis, although further development would rely on the use of a different amine than the substrate amine as the base. This project has many of the hallmarks of environmentally benign chemical processing: the elimination of hazardous waste byproducts, the elimination of toxic chemicals in the manufacturing process, and eliminating the use of organic solvents.
3. Proposed Studies.
a. Thiocarbamate Synthesis using ML2X2 Catalysts.
As outlined in the previous section, preliminary work accomplished using seed money from the Eastman Kodak Company has allowed us to establish a new approach for the synthesis of acyl derivatives. While the 1 year funding period for this project ends in June 1996, we intend to continue to work in the area of discovering new environmentally benign chemistry using transition metal catalysts. Specifically, we will continue to develop the system we have discovered and to investigate new catalysts.
With regards to the present system, based on nickel and palladium M(PPh3)2Cl2 catalysts, many questions arise about the mechanism of the reaction. We anticipate that study of the mechanism of this reaction will allow us to extend the acyl formation chemistry to a wider range of compounds, including ureas, amides, and esters. In order to accomplish this goal, we will need not only to know the nature of the intermediates involved, but also their relevance to the kinetics of the catalytic reaction.
The first studies will be to examine the reactions in deuterated solvents using NMR spectroscopy. Obvious studies will be to look at the interaction of Ni(PPh3)2Cl2 with CO in the presence of a secondary amine, such as piperidine. 31P NMR spectroscopy has the advantage of providing an easy method for determining the number of (phosphine containing) species, whereas 1H NMR spectroscopy will provide information about the formation of adducts. Once the nature of the interaction of the amine, CO, and Ni complex has been determined, thiol can be introduced and further intermediates characterized. Should any of the reactions be too rapid for observation of intermediates, the temperature can be lowered to slow rates of reaction. The amine and the thiol can be varied to provide a sample that enhances the ease of analysis. 13C NMR spectroscopy provides a useful way of identifying carbene complexes, as the resonances for these ligands appear far downfield from other resonances.
The reaction will also be studied using infrared spectroscopy. Samples can be characterized using IR to verify the presence of carbon-monoxide and oxycarbene ligands. The extensive literature on these types of compounds will be useful in assigning structures to intermediates. Our collaborators at Eastman Kodak also have the capability of performing in situ IR spectroscopy using a small probe that can be inserted into the reaction chamber, thereby allowing characterization of intermediates during catalysis.
Another approach to characterization of the reaction intermediates will be to synthesize likely intermediates. For example, it should prove possible to make species such as Ni(PPh3)2(NMe2)Cl or Ni(PPh3)2(SAr)Cl, and then examine their reactions with carbon monoxide. Alternatively, reactions of cationic carbonyl compounds such as [Ni(PPh3)2(CO)Cl]+ with thiolate and amide nucleophiles can be examined. A comparison of the spectral data from these reactions with those obtained during the catalytic reaction might help identify the intermediates involved.
The kinetics of the reaction will also be examined. By varying the concentration of amine, the concentration of thiol, and the concentration of catalyst we can determine the reaction order in each component. CO pressure can also be varied. Once the order of each reactant is known, we can construct a mechanistic scheme that is consistent with the kinetic equation. We will begin with our original model for how this reaction might work, i.e., formation of a metal carbonyl is followed by nucleophilic attack by the amine, followed by attack by the thiol. It is possible that nucleophilic attack by an amine is slow, and that the thiol is the first to attack in a rate determining step, as indicated in Scheme III below. The excess of amine present could then rapidly lead to thiocarbamate formation as opposed to dithiocarbonate product. It is also possible that one or more of these steps is reversible, which would lead to the reaction being first order in all reagents in or prior to the rate determining step. Alternatives also include the possible formation of free isocyanate.
Scheme III:
Another aspect of the kinetics that will be important to control is the pH of the solution. In the best preliminary experiments, the amine was employed as the solvent, and the solution was rather basic. This may be an important feature of the reaction in that anionic amines and thiols are likely to be more nucleophilic than their neutral counterparts. Certainly, under the current reaction conditions, the aryl thiol is fully deprotonated by the amine solvent, which means that the mixture of amine/aminium ion is serving to buffer the pH of the reaction. We can control this buffer pH by choice of the amine and the ratio of amine/aminium ion, and will examine the effect of pH on the reaction. It will also be important to establish that other amines, such as triethylamine, can serve as the base in this reaction. It should prove easier to study pH effects with this amine as the base since it will not be consumed in the acyl product forming reaction. Commercial implications are better for triethylamine as the base rather than the amine substrate.
We will also need to investigate the role of air in the reaction. Is it acting to directly reoxidize the metal back to the +2 oxidation state? Also, we need to know if it is possible to use co-catalysts for the oxidation, such as CuCl/CuCl2 as is done in the Wacker process. Stoichiometric oxidants, such as ferrocinium or benzoquinone could also be employed to control the oxidation step of the reaction in these studies, although air is certainly a more desirable oxidant for a large scale process.
Many other variables need to be examined in this reaction. Does raising the temperature result in enhanced reaction rates? Can disulfides be used as the oxidant in place of air? What, if any, are the effects of solvent polarity on the reaction? Are phosphine or chloride lost from the initial NiL2Cl2 complex, and does added ligand enhance or retard the rate of reaction? Does a phosphine or halide 'sponge' enhance the rate of reaction? What effect does changing the phosphine attached to the metal have on the reaction? What is the scope of reactants that can be coupled using this catalyst? All of these questions fit together with the mechanistic studies that are planned for this series of catalysts to allow us to better understand and design new catalysts for enhanced reactivity and general performance.
b. Scope of Reaction
We will also examine the thiocarbamate forming reaction to determine the scope of amines and thiols that can be employed. These studies will begin by varying the thiol RSH, varying its basicity and nucleophilicity. We will look at substituted aromatic thiols to determine Hammett substituent effects on the reaction. Aliphatic thiols will also be examined for their reactivity.
Similarly, the amine will also be examined for the types of amines that can be used. Our preliminary studies used only a secondary benzylic amine, since this is used in one of the large scale process reactions at Kodak. For aliphatic amines, primary and secondary reactivities will be compared. The nucleophilicity of the amine can be varied. The use of aniline and heterocyclic amines will also be tested to expand the scope of the reaction.
We suspect that some of these substrates will be unreactive towards the CO/catalyst system. In this case, we will examine the effects of added Lewis acids, which are known to promote attack on coordinated carbon monoxide. Catalysts that have been shown to promote CO insertion reactions include AlCl3, AlBr3, and AlPh3.
c. Other Catalysts for Thiocarbamate Synthesis.
The above studies focus on our earlier work with nickel and palladium M(PPh3)2Cl2 complexes as the catalyst for acyl group formation. We will also explore other catalysts for the thiocarbamate coupling reaction. Alper recently reported the use of palladium on clay with ruthenium carbonyl for the catalytic reductive carbonylation of nitroarenes to produce carbamates in good yield (eq 14).
(14)The role of the clay was stated to reduce the nitro group to an amine, with the ruthenium acting to then couple the amine with the alcohol. One can imagine a similar process using a thiol rather than an alcohol, and using an amine instead of the nitroarene. Such a process would produce thiocarbamates similar to our nickel and palladium based catalysts.
Indeed, the concept of using bipyridine rather than phosphine catalysts is a novel one. We could examine the reactivity of complexes such as Ni(bipy)Cl2 and Pd(bipy)Cl2 as catalysts for comparison with the phosphine derivatives.
One practical problem posed by homogeneous catalysts is that they must be separated from the product at the end of the reaction. There are several solutions to this problem. One is to have the catalyst be so efficient that there is only a trace present in the final product that does not need to be recovered, as in Zieglar-Natta polymerization. Another is to immobilize the catalyst on a heterogeneous support. Polymer bound triphenylphosphine could be used to achieve catalyst separation, and could be used with our nickel system. It should also be possible to extract the catalyst into an aqueous layer while separating the product in an organic layer by choosing the appropriate ligands on the catalyst.
Another alternative is to attempt the same type of coupling chemistry using a heterogeneous catalyst, such as palladium on charcoal or nickel chloride on alumina. In general, however, heterogeneous catalysts of this type generally provide reactivity only at elevated but at the expense of selectivity. In the thiocarbamate synthesis, one might expect to see significant homo-coupling to generate ureas at the temperatures required for reaction with CO, making this a less likely route to the desired products.
d. Synthesis of Other Acyl Derivatives
Our primary interest in focusing on the synthesis of thiocarbamates in the above studies was that this type of product offers us the chance to study systems that will have to demonstrate high selectivity in the competitive reactivity of the nucleophiles with metal carbonyls and carbenes. We will also explore the use of metal carbonyls for the synthesis of other acyl derivatives.
A recent report by Murai and coworkers suggests an interesting approach. They reported that N-methylimidazole will react with CO and olefins in the presence of Ru3(CO)12 catalyst to generate ketones arising from C-H activation and acylation of the position a to the imidazole nitrogen (eq 15).
(15)We will examine similar reactions using amines with acidic N-H bonds, thereby generating amides rather than ketones as products. The idea will be to use a transition metal to N-H activate an amine through oxidative addition, and then insert an olefin into the metal-hydride bond. At this point, CO insertion followed by reductive elimination would not only generate the amide product but also produce the metal in its reduced oxidation state, ready for another cycle. Scheme IV shows this process. Several metal complexes will be probed for this type of reaction, including species such as Ni(bipy)(COD), Ni(PPh3)4, and RhCl(PPh3)2(CO). The general properties required of the metal are (1) an open or readily available coordination site, (2) an accessible oxidation state higher by 2 units, (3) a site for olefin/CO insertion. The square planar 16 electron complexes of the platinum metals nicely fit these requirements, and are known to activate N-H bonds in amines.
Scheme IV:
e. Extension to Other Nucleophiles
The preliminary results establish that amine and thiol nucleophiles can be combined with CO using a nickel metal catalyst, and that this class of products is of interest to Kodak Chemicals. The generalization of this reaction to include other nucleophiles would make the process of even greater interest. These nucleophiles could include not only heteroatom nucleophiles but also carbon nucleophiles.
One example of a carbon nucleophile that would be of interest would be interest would be the use of enamines. These masked nucleophiles could attack coordinated CO along with a second nucleophile to allow the generation of a,b-unsaturated products. Once again, a base will be required to remove protons and an oxidation of the metal will be required to make the process catalytic. Scheme V shows a possible sequence.
f. Industrial Opportunities
While the research projects outlined above will address long-term, generic problems of the photographic intermediates business, the students and postdoctoral fellows involved in this project will benefit from 'front-line' interactions with the industrial partner. The fundamental research performed in the academic laboratory will be meshed with a more applied approach in the industrial laboratory. Each project participant will spend 6-12 months in the Chemicals Development Division of Eastman Kodak Company, under the supervision of Dr. Ronald Valente. Dr. Valente is an expert in chemicals synthesis and processing, and will provide valuable opportunities for learning to the project participants.
Scheme V:
The period of time spent in the industrial lab will begin with training to bring the student/postdoc up to speed on modern process development. After this initial training, biweekly meetings will keep the student up to date on current processing problems, and will expose the student to a variety of chemical transformations and problem solving approaches.
In addition to the basic chemistry involved, the student/postdoc will be exposed to process development topics not usually encountered in the academic laboratory. Examples of these topics will include equipment capability concerns and isolation concerns. What does one do if a product doesn't filter well on the 100 kg scale? What if the product is air-sensitive? What if the product is dangerous to the worker?
The student/postdoc will be exposed to new chemicals and process
development as well as existing chemicals processes. They will
learn that the concerns for these two distinct areas are different.
Perfectly good existing processes might be under pressure to
change due to environmental or business reasons. They will learn
about capability studies and see examples of statistically based
approaches for problem solving in existing processes.