Preparation of Carbodiimides From Ureas by Dehydration

Synthetic linear polyester filaments, yarns and cords are known to show improved strength under hydrolytic conditions or elevated temperatures when, either by the use of certain additives to the polymer, or chemical coatings on the polyester fiber, the carboxyl end group concentration is lowered.

One of the commonly employed methods of lowering the carboxyl level is by applying to the polyester fiber a surface coating which consists essentially of a carbodiimide or polycarbodiimide. United States patents relating to the use of such carbodiimides or polycarbodiimides include U.S. Pat. No. 3,193,523, U.S. Pat. No. 3,193,524, and U.S. patent application Ser. No. 302,710 filed Nov. 1, 1972.

It is an object of this invention to provide a new and improved method of preparing biscarbodiimides.

Briefly, the object of this invention is attained by dehydration of ureas by triphenylphosphine and carbon tetrachloride reactants according to the reaction sequence shown in the following typical example:

The reactants are rather common and relatively inexpensive and the products are substantially free of residual isocyanates and uretidinedione imines.

Separation and isolation of the carbodiimide from the triphenylphosphine oxide was sometimes difficult, and has been demonstrated by vacuum distillation although fractional crystallization and extraction may well be practical.

In the production of certain biscarbodiimides, it was recognized that the product being isolated was a monocarbodiimide of the terminal groupings (R' and R") of the bisurea rather than the desired biscarbodiimide. The possibility of inter- or intramolecular rearrangement of a reaction intermediate during dehydration is considered a possible explanation. ##STR1## + other products Another possibility is that intra- or intermolecular rearrangement of the desired biscarbodiimide occurred during isolation of the reaction products. ##STR2## + other products

The separation of the biscarbodiimides from other products by distillation, for example, resulted in rearrangement in accordance with the following equation:

where

Notwithstanding problems of rearrangement, biscarbodiimides which are effective as additives for polyethylene terephthalate tire yarn as stabilizers against hydrolytic degradation can be prepared in good yield by this dehydration method.

The following examples should not be construed to limit the scope of this invention.

The amine (0.41 mole) dissolved in 200 ml of tetrahydrofuran was placed in a 1000-ml flask equipped with a mechanical stirrer, dropping funnel, and nitrogen inlet to maintain an atmosphere of nitrogen above the reactants. The flask was cooled by an ice bath. The diisocyanate (0.20 mole) dissolved in 200 ml of tetrahydrofuran was dropped into the amine solution over a 2-hr period. Upon completion of the addition, stirring was continued for 12 hr at room temperature.

The diurea was prepared from aniline (Eastman, MP -6 to -5.degree.) and tolylene-2,4-diisocyanate (Fisher, practical). At conclusion of the 14-hr reaction period, a white precipitate had formed. The precipitate was isolated by filtration and washed with water. The material was then slurried with 2 l of 2 N aqueous hydrochloric acid in a blender, filtered, slurried three times with 2 l portions of water and filtered, and then rinsed with acetone. After drying at 110.degree.C under 20 mm Hg pressure for 12 hr, a white powder was obtained.

The diurea was prepared from phenyl isocyanate (Fisher, BP 55.degree.-56.degree./13mm) and p-phenylene diamine (Fisher, certified). The isocyanate was dropped into the diamine solution. At conclusion of the 14-hr reaction period, a grey precipitate had formed. The precipitate was isolated by filtration and washed with water. The material was then slurried with 2 l of 2 N aqueous hydrochloric acid in a blender, filtered, slurried three times with 2 l portions of water and filtered, and then rinsed with acetone. After drying at 110.degree. C under 20 mm Hg pressure for 12 hr, a grey powder was obtained.

The diurea was prepared from n-hexyl amine (Eastman, BP 128.degree.-130.degree.) and tolylene-2,4diisocyanate (Fisher, practical). At conclusion of 14-hr reaction period, a white precipitate had formed. The precipitate was isolated by filtration and washed with acetone. The material was recrystallized from ethanol/chloroform.

The diurea was prepared from aniline (Eastman, MP -6.degree. to -5.degree.) and hexamethylene diisocyanate (Eastman, MP 40.degree.-42.degree.). At conclusion of 14-hr reaction period, a white precipitate had formed. The precipitate was isolated by filtration and washed with acetone. The material was recrystallized from dimethylsulfoxide/acetone.

The urea (0.050 mole), triethylamine (0.100 mole) were added to 100 ml of dichloromethane in a round bottom flask equipped with a mechanical stirrer and condenser. The dichloromethane was refluxed at approximately 40.degree. C for 5 hr. During the first hour of refluxing, a yellow solution was obtained. Precipitation of triethylamine hydrochloride occurred after several hours at reflux. The reaction mixture was cooled to room temperature, filtered, and the residue washed with dichloromethane. The filtrate was washed twice with 200-ml portions of water. The dichloromethane layer was dried over magnesium sulfate, filtered, and then evaporated to dryness at room temperature under 1 mm Hg pressure using a rotatory vacuum apparatus. This residue was then vacuum distilled. An Ace Glass mini-vacuum distillation head with a 15-cm Vigreux column was used.

A clear liquid boiling at 123.degree.-126.degree.C (0.5 mm Hg) was obtained. Infrared and mass spectra indicated that the compound obtained was N,N'-diphenylcarbodiimide instead of the expected biscarbodiimide. Elemental analysis was consistent with C.sub.13 H.sub.10 N.sub.2 rather than C.sub.21 H.sub.16 N.sub.4 ; calculated 80.38%C, 5.19%H, 14.43%N; found: 78.91%C, 5.23%H, 14.00%N. The nuclear magnetic resonance spectrum of the compound in CDCl.sub.3 exhibited only a multiplet at 7.15 to 7.45.delta. (ppm). The infrared spectrum of the compound was identical to that of C.sub.6 H.sub.5 --N=C=N--C.sub.6 H.sub.5. The mass spectral cracking pattern of the compound also appeared equivalent to that of C.sub.6 H.sub.5 --N=C=N--C.sub.6 H.sub.5. The yield was 45% of theory.

The white precipitate which formed during the dehydration was isolated by filtration, washed with dichloromethane, and dried at 80.degree. C under 20 mm Hg pressure. The precipitate weighed 1.5 g and was identified as triethylamine hydrochloride by comparison of its infrared and nuclear magnetic resonance spectra with those of an authentic sample. The water washes used to remove the amine hydrochloride from the dichloromethane reaction solution were evaporated to dryness to yield 12.3 g of residue with infrared (KBr pellet) and nuclear magnetic resonance (D.sub.2 O solution) spectra identical to those of triethylamine hydrochloride.

A 10-ml portion of the dichloromethane reaction solution was washed with D.sub.2 O and the nuclear magnetic resonance spectrum of the D.sub.2 O wash was then obtained. The spectrum indicated the presence of only triethylamine hydrochloride with no signals in the 7 to 9.delta. region indicative of phenyl protons.

A 10-ml portion of the water-washed dichloromethane reaction solution was evaporated to dryness under a nitrogen sweep at room temperature, and the nuclear magnetic resonance spectrum of the residue was determined in CDCl.sub.3. The spectrum exhibited phenyl proton signals at 7 to 8.delta. and an aryl CH.sub.3 signal at 2.3.delta. ; the spectrum did not exhibit a NH signal in the 8 to 9.delta. region. The nuclear magnetic resonance spectrum thus incicated the presence of the desired biscarbodiimide in the dichloromethane reaction solution. The infrared spectrum of this residue exhibited absorptions attributable to triphenylphosphine oxide and a strong carbodiimide absorption at 2140 cm.sup.-1 but only slight if any amide absorptions at 3200 cm.sup.-1, 1650 cm.sup.-1, or 1550 cm.sup.-1 were observed.

The pot residue after distillation of the reaction products consisted of triphenylphosphine oxide and residual aryl carbodiimides as evidenced by the infrared spectrum of the residue. The spectrum consisted of the characteristic absorptions of triphenylphosphine oxide at 1430, 1180, 1120, 995, 755, 720, and 695 cm.sup.-1; and the carbodiimide absorption at 2150 cm.sup.-1. No absorptions in the 1600 to 1700 cm.sup.-1 region were exhibited, thus indicating no dimers, trimers or polymers derived from the reaction of carbodiimide groups to yield ##STR9## groups.

The dehydration was repeated as described above except that the separation of the products was varied in an attempt to isolate the biscarbodiimide The alternate procedure was employed to verify the rearrangement theory. In this procedure, the filtered dichloromethane reaction solution was evaporated to dryness under 20 mm Hg pressure at room temperature. The residue was then extracted with petroleum ether (38.degree.-51.degree.) for 3 hrs. in a Soxhlet apparatus. The petroleum ether was removed under vacuum to yield 10.1 g of petroleum ether soluble residue, a 60% yield of the crude biscarbodiimide based upon the equation. The nuclear magnetic resonance spectrum of the residue in CDCl.sub.3 exhibited a multiplet from phenyl protons at 7.1 to 7.7.delta. and a singlet from the aryl CH.sub.3 protons at 2.30.delta. (ppm). The relative area of the two signals was (phenyl H) / (CH.sub.3) = 4.5. The mass spectrum exhibited the molecular ion at (m/e) of 324 for the biscarbodiimide ##SPC2##

and the cracking pattern was consistent with this structure.

Elemental analysis for the petroleum ether soluble material was 77.19% C, 5.02% H, and 14.37% N. The calculated values for C.sub.21 H.sub.16 N.sub.4 are 77.75%C, 4.97%H, and 17.28 %N.

Apparently the dehydration reaction yields the expected biscarbodiimide, but this compound rearranges during distillation to yield N,N'-diphenylcarbodiimide as a product.

The petroleum ether soluble residue was vacuum distilled to yield 3.5 g of clear liquid boiling at 160.degree.-170.degree.C (0.35 mm Hg). The yield of N,N'-diphenylcarbodiimide was 36% of theory. The infrared spectrum appears identical with that of N,N'-diphenylcarbodiimide. The mass spectrum cracking pattern appears equivalent to that of diphenylcarbodiimide. The nuclear magnetic resonance spectrum in CDCl.sub.3 exhibited only a multiplet at 7.1 to 7.4.delta.(ppm). Elemental analysis was satisfactory for C.sub.13 H.sub.10 N.sub.2 ; calculated: 80.38%C, 5.19%H, 14.43%N; found: 79.75%C, 5.10%H, 13.98%N.

The dehydration was repeated; and after filtration, the dichloromethane reaction solution was extracted with water and then dried over magnesium sulfate. The dichloromethane was removed, and the residue was extracted with petroleum ether. The petroleum ether solubles were isolated and distilled. Results of this separation were equivalent to those described above for the separation using an unwashed dichloromethane solution.

From distillation of the dichloromethane soluble products, the urea was dehydrated as described in Example 6A. A clear liquid boiling at 112.degree. C (0.6 mm Hg) was obtained. Infrared and mass spectra indicated that the compound obtained was N,N'-di-n-hexylcarabodiimide instead of the expected biscarbodiimide. Elemental analysis was consistent with C.sub.13 H.sub.26 N.sub.2 rather than C.sub.21 H.sub.32 N.sub.4 ; calculated:74.22%C, 12.46%H, 13.32%N; found: 73.2%C, 10.95%H, 13.53%N. The nuclear magnetic resonance spectrum of the compound in CDCl.sub.3 exhibited the following signals: ##EQU1## The infrared spectrum of the compound was identical with that of CH.sub.3 (CH.sub.2).sub.5 --N=C=N (CH.sub.2).sub.5 CH.sub.3. The mass spectral cracking pattern of the compound also appeared equivalent to that of CH.sub.3 (CH.sub.2).sub.5--N=C=N--(CH.sub.2).sub.5 CH.sub.3. The yield of CH.sub.3 (CH.sub.2).sub.5--N=C=N--(CH.sub.2).sub.5 CH.sub.3 was 51% of theory.

The urea was dehydrated as described in Example 6A. From distillation of the dichloromethane soluble products, a clear liquid boiling at 139.degree. -143.degree. C (0.2 mm Hg) was obtained. Infrared and mass spectra indicated that the compound obtained was N,N'-diphenylcarbodiimide instead of the expected biscarbodiimide. Elemental analysis was consistent with C.sub.13 H.sub.10 N.sub.2 rather than C.sub.20 H.sub.22 N.sub.4 ; calculated: 80.38%C, 5.19%H, 14.43%N found: 80.05%C, 5.55%H, 13.80%N. The nuclear magnetic resonance spectrum of the compound in CDCl.sub.3 exhibited only a multiplet at 7.0 - 7.4.delta. (ppm). The infrared spectrum of the compound appeared identical to that of C.sub.6 H.sub.5 --N=C=N--C.sub.6 H.sub.5. The mass spectral cracking pattern of the compound also appeared equivalent to that of C.sub.6 H.sub.5 --N=C=N--C.sub.6 H.sub.5. The yield was 56% of theory. EXAMPLE 9 ##SPC5##

The urea was dehydrated as described in Example 6A. From distillation of the dichloromethane soluble products, a clear liquid boiling at 149.degree. -151.degree. C (1.1 mm Hg) was obtained. Infrared and mass spectra indicated that the compound obtained was N,N'-diphenylcarbodiimide instead of the expected biscarbodiimide. Elemental analysis was consistent with C.sub.13 H.sub.10 N.sub.2 rather than C.sub.20 H.sub.14 N.sub.4 ; calculated: 80.38%C, 5.19%H, 14.43%N; found 79.63%C, 5.47%H, 13.97%N. The nuclear magnetic resonance spectrum of the compound in CDCl.sub.3 exhibited only a multiplet at 7.0 to 7.4.delta. (ppm.) The infrared spectrum of the compound appeared identical to that of C.sub.6 H.sub.5 --N=C=N--C.sub.6 H.sub.5. The mass spectral cracking pattern of the compound also appeared equivalent to that of C.sub.6 H.sub.5 --N=C=N--C.sub.6 H.sub.5. The yield was 58% of theory.