TRIPENTYL ESTERS OF TRIMELLITIC ACID

Abstract

Tripentyl esters of trimellitic acid, especially mixtures of triisopentyl esters of trimellitic acid comprising isomeric pentyl radicals in which more than 5 mol % of the isomeric pentyl radicals incorporated in the ester mixture are branched, have good compatibility with PVC and PVC-containing polymers and simultaneously exhibit a lesser tendency to migrate.

Claims

1. A composition, comprising an ester mixture of triisopentyl cyclocyclohexane-1,2,4-tricarboxylates, wherein more than 5 mol % of isomeric pentyl radicals incorporated in the ester mixture are branched.

2. The composition of claim 1, wherein at least 15 mol % of the isomeric pentyl radicals incorporated in the ester mixture are branched.

3. The composition of according to claim 1, wherein at least 50 mol of the branched isomeric pentyl radicals incorporated in the ester mixture are 2-methylbutyl radicals.

4. The composition according to claim 1, wherein more than 10 mol % of the isomeric pentyl radicals incorporated in the ester mixture are linear.

5. A process for preparing the composition of claim 1, the process comprising: reacting trimellitic acid, a trimellitic acid derivative, or both, with a mixture of isomeric pentanols, such that more than 5 mol % of isomeric pentyl radicals incorporated in the mixture of isomeric pentanols are branched, to obtain a mixture of triisopentyl esters of trimellitic acid; and hydrogenating the mixture of triisopentyl esters of trimellitic acid to obtain an ester mixture of triisopentyl cyclohexane-1,2,4-tricarboxylates.

6. The process according to claim 5, wherein: one or more trialkyl trimellitates in which the alkyl radicals of the ester functions each comprise fewer than 4 carbon atoms is/are transesterified with the mixture of isomeric pentanols; and more than 5 mol % of the isomeric pentyl radicals incorporated in the mixture of isomeric pentanols are branched.

7. The process according to claim 5, wherein, in the reaction of the trimellitic acid, the trimellitic acid derivative, or both, with the mixture of isomeric pentanols, the reaction mixture is heated to boiling and at least 0.2 molar equivalent of the amount of mixture of isomeric pentanols needed to introduce all the isomeric pentyl radicals incorporated in the ester mixture is added to the reaction mixture containing the trimellitic acid, the trimellitic acid derivative, or both, only after attainment of the boiling temperature.

8. The process according to claim 5, wherein: at least two isomerically pure tripentyl esters of trimellitic acid which differ in terms of isomerism of the pentyl radicals are incorporated in the mixture of triisopentyl esters of trimellitic acid, at least two mixtures of triisopentyl esters of trimellitic acid comprising isomeric pentyl radicals which differ in isomer distributions of the isomeric pentyl radicals are incorporated in the mixture of triisopentyl esters of trimellitic acid, or at least one isomerically pure tripentyl ester of trimellitic acid and at least one mixture of triisopentyl esters of trimellitic acid comprising isomeric pentyl radicals are mixed with one another, such that more than 5 mol % of the isomeric pentyl radicals incorporated in the commixed ester mixture are branched.

9. A process, comprising: reacting trimellitic acid, a trimellitic acid derivative, or both, with pentanol, the reaction mixture being heated to boiling, such that at least 0.2 molar equivalent of an amount of pentanol needed to introduce all pentyl radicals into an ester mixture formed by the reacting is added to the reaction mixture only after a boiling temperature is attained, to obtain a mixture of triisopentyl esters of trimellitic acid; and hydrogenating the mixture of triisopentyl esters of trimellitic acid to obtain an ester mixture of triisopentyl cyclohexane-1,2,4-tricarboxylates.

10. A plasticizer, comprising the composition of claim 1.

11. A plasticizer composition, comprising the composition of claim 1 and a further polymer-plasticizing compound.

12. A composition, comprising the plasticizer of claim 10, said composition being selected from the group consisting of an adhesive, a sealing compound, a coating material, a paint, an ink and a plastisol.

13. The plasticizer composition of claim 11, wherein the further polymer-plasticizing compound comprises at least one selected from the group consisting of alkyl benzoates, dialkyl adipates, glycerol esters, trialkyl citrates, acylated trialkyl citrates, glycol dibenzoates, trimellitates with radicals other than those described in the present invention, dialkyl terephthalates, dialkyl phthalates, esters of furandicarboxylic acid, dialkanoyl esters, of dianhydrohexitols (e.g. isosorbide), epoxidized fatty acid alkyl esters, polymer plasticizers, for example the polyadipates, and dialkyl esters of cyclohexane-1,2-, -1,3- or -1,4-dicarboxylic acid.

14. A polymer composition, comprising plasticizer composition according to claim 13 and at least one polymer selected from the group consisting of polyvinyl chloride, copolymers of vinyl chloride with vinyl acetate or with butyl acrylate, polyalkyl methacrylate (PAMA), polyvinyl butyral (PVB), chlorosulphonated polyethylene, polyurethane, polysulphides, polylactic acid (PLA), polyhydroxybutyral (PHB) and nitrocellulose.

15. An article, comprising the plasticizer of claim 10, said article being selected from the group consisting of a foam, a synthetic leather, a floorcovering, a wallcovering, a roofing membrane, an underbody protection, a fabric coating, a cable, a wire insulation, a hose, a tube, an extruded article, a film, a toy, a contact sheet, a food packaging and a medical article.

16. An article, comprising the plasticizer of claim 10, said article being selected from the group consisting of a top layer of a floorcovering, a film in the automobile interior sector, a tube of a medical article and a blood bag.

Description

EXAMPLES

Examples 1 to 7

Preparation of Trimellitates by Esterification of Trimellitic Anhydride

[0102] 576.0 g (3 mol) of trimellitic anhydride (Sigma Aldrich, 97% purity) and the amount ml of the alcohol A were introduced into an apparatus comprising a stirred flask with a stirrer, sampling stub, dropping funnel, immersed tube, thermometer and water separator in the case of the preparation of C4 trimellitate. The apparatus was purged with nitrogen (6 l/h) through the immersed tube for one hour, before 0.8 g (0.75 mmol) of tetra-n-butyl titanate (Sigma Aldrich, >97% purity) was added. While sparging with nitrogen (6 l/h), which lasts until the end of the reaction, the mixture was heated to boiling while stirring. Water obtained as a result of the reaction was removed continuously by means of a water separator. From attainment of a reaction temperature of 240 C., the amount m2 of the alcohol A and 0.8 g (0.75 mmol) of tetra-n-butyl titanate were metered in at such a rate that the reaction temperature did not fall below 240 C. As soon as 108 ml (6 mol) of water had been removed by means of the water separator, the acid number was determined on a sample of the reaction solution in accordance with DIN EN ISO 2114. Depending on the particular acid number determined, the reaction mixture was heated further until the acid number of the reaction mixture was below 0.1 mg KOH per g of sample (total reaction time t).

[0103] The cooled reaction mixture from the esterification was transferred into a stirred flask with stirrer, thermometer, immersed tube, Claisen distillation apparatus and receiver flask, and purged with nitrogen through the immersed tube for at least one hour, and then the pressure was reduced to about 1 mbar. The temperature was increased gradually up to 160 C. and the excess alcohol was distilled off. The mixture was cooled down under vacuum with introduction of nitrogen (20 mbar). On attainment of a temperature of 80 C., the pressure was equalized with nitrogen, and the crude produce was admixed with 15 mmol of demineralized water and 0.5% by mass of activated carbon (Cabot Norit Nederland B.V., CAP Super, amount based on the reaction produce assuming a yield of 100%) and stirred with introduction of nitrogen (6 l/h) at 80 C. for 15 minutes. Subsequently, the reaction mixture was heated to 160 C. under reduced pressure with introduction of nitrogen (20 mbar) for 2 hours and the last volatile components were removed. After cooling again down to 80 C., the reaction mixture, while sparging with nitrogen (6 l/h), was mixed with 2% by weight of basic alumina (Sigma Aldrich, Brockman 1 type, amount based on the reaction product assuming a yield of 100%) and stirred at 80 C. for 1 hour.

[0104] The reaction mixture was then filtered at 80 C. through a Bchner funnel with filter paper and precompacted filtercake of filtration aid (D14 Perlite ex Knauf) into a suction bottle by means of reduced pressure. The filtrate was analysed by means of GC analysis with regard to purity and by means of NMR with regard to composition.

[0105] GC Analyses:

[0106] The GC analysis took place with the following parameters: [0107] Capillary column: 30 m DB5; 0.25 mm ID; 0.25 m film [0108] Carrier gas: Helium [0109] Column pressure: 150 kPa [0110] Cool on-column injection [0111] Oven temperature programme (duration: 51 min): 50 C. (for 1 min), heating at 15 C./min to 350 C. (hold temperature for 14 min) [0112] Injector: 50 C. [0113] Detector (FID): 425 C. [0114] Injection volume: 0.3 l

[0115] Components in the sample chromatogram were identified using a comparative solution of the relevant esters. This was followed by standardization of the signals in the sample chromatogram to 100 area %. The molar ratios were determined in sufficient approximation from the area ratios of the individual signals.

[0116] The purity was determined via the fraction of the product signals as a proportion of the total areas in the chromatogram.

[0117] As an alternative to the NMR method described below, the nature and number of the isomeric pentyl radicals present in the triisopentyl esters can be determined by hydrolysis of the ester mixture in basic solution and subsequent GC analysis of the resulting alcohol mixture. It should be noted here that the GC conditions (especially column material and column dimensions, and temperature programme) permit separation of the alcohols into the individual isomers. Components in the sample chromatogram are then identified using a comparative solution of the relevant esters. This is followed by standardization of the signals in the sample chromatogram to 100 area %, such that it is possible to determine the molar ratios in sufficient approximation from the area ratios of the individual signals.

[0118] NMR Analyses:

[0119] The composition of the triisopentyl esters mixtures, i.e. the respective proportion of the different isomeric pentyl radicals in the totality of all pentyl radicals, can be ascertained, for example, by 1H NMR and 13C NMR spectroscopy. The determination of the composition was conducted here with the aid of 1H NMR spectroscopy on the solution of the triisopentyl esters mixture in deuterochloroform (CDCl3). For the recording of the spectra, 20 mg of substance are dissolved in 0.6 ml of CDCl3 (containing 1% by mass of TMS) and transferred to an NMR tube having a diameter of 5 mm. Both the substance to be analysed and the CDCl3 used were first dried by means of molecular sieve in order to rule out distortion of the measurements by any water present. The NMR spectroscopy studies can in principle be conducted with any commercial NMR instrument. For the present NMR spectroscopy studies, an instrument of the Bruker Avance 500 type was used. The spectra were recorded at a temperature of 303 K with a delay of d1=5 seconds, 32 scans, a pulse length of about 9.5 s (90 C. excitation pulse) and a sweep width of 10 000 Hz with a 5 mm BBO (broad band observer) sample head. The resonance signals were recorded relative to the chemical shift of tetramethylsilane (TMS=0 ppm) as internal standard. Other commercial NMR instruments give comparable results with the same operating parameters.

[0120] FIG. 1 below is intended to serve for comprehension of the determination method elucidated. Of relevance here is the assigned numbering of the carbon atoms of the methyl groups and the methylene groups adjacent to the oxygen atom in an n-pentyl radical (C14H3 and C10H2), in a 2-methylbutyl radical (C22H3, C23H3 and C19H2) and in a 3-methylbutyl radical (C18H3, C18H3 and C15H2).

[0121] FIG. 1: Numbering of the Carbon Atoms in the Different Radicals

##STR00001##

[0122] The resultant 1H NMR spectra of the mixtures of triisopentyl esters of trimellitic acid have, in the range from 0.5 ppm up to the minimum of the lowest valley in the range from 0.7 to 1.2 ppm, resonance signals which are formed by the signals of the hydrogen atoms of the methyl group(s) of the isomeric pentyl substituents (C14H3, C18H3, C18H3, C22H3, C23H3). The signals in the range of chemical shifts from 3.60 to 4.55 ppm can be assigned essentially to the hydrogen atoms of the methylene group adjacent to the oxygen in the alcohol radical (C10H2, C15H2, C19H2). In this context, the protons on C19 are subject to a high-field shift because of the adjacent tertiary carbon atom and appear between 3.95 and 4.29 ppm, while the protons on C10 and C15 give signals at lower shifts of 4.29 to 4.55 ppm.

[0123] Quantification is effected by comparative determination of the area beneath the respective resonance signals, i.e. the area enclosed by the signal from the baseline. Commercial NMR software has program functions for integration of the signal area. In the present NMR spectroscopy study, the integration was conducted with the aid of the TopSpin software, Version 1.

[0124] To determine the mean degree of branching of the isomeric pentyl radicals in the mixture according to the invention, the integral value of the signals in the range from 0.68 to 1.18 ppm (I(CH3)) is first divided by the integral value of the signals in the range from 3.95 to 4.55 ppm (I(OCH2)). In this way, an intensity ratio which states the ratio of the number of hydrogen atoms present in a methyl group to the number of hydrogen atoms present in a methylene group adjacent to an oxygen is obtained. Since three hydrogen atoms per methyl group and two hydrogen atoms in every methylene group adjacent to an oxygen are present, the intensities have to be divided by 3 and 2 respectively, in order to obtain the ratio of the number of methyl groups to the number of methylene groups adjacent to an oxygen in the pentyl radical. Since a linear n-pentyl radical having only one methyl group and one methylene group adjacent to an oxygen does not contain any branch and accordingly has to have a degree of branching of 0, it is necessary to subtract the quantity of 1 from the ratio.

[0125] The mean degree of branching DB can thus be calculated by the formula

DB=*I(CH3)/I(OCH2)1

from the intensity ratio measured. In this formula, DB means mean degree of branching, I(CH3) means area integral assigned to the methyl hydrogen atoms, and I(OCH2) means area integral of the methylene hydrogen atoms adjacent to the oxygen.

[0126] The product may contain 2-methylbutyl radicals and 3-methylbutyl radicals each having a degree of branching of 1, and also n-pentyl radicals having a degree of branching of 0, which means that the maximum means degree of branching of any triisopentyl ester is always 1. From the deviation of the mean degree of branching from the value of 1, it is therefore possible to determine the molar proportion of n-pentyl radicals (xpentyl) in the molecule.

xpentyl=1DB

[0127] The proportion of 2-methylbutyl radicals can be calculated with the aid of the integration of the baseline-separated signals in the range from 3.95 to 4.55 ppm. Here too, the signals of the protons on C19 (C19H2; multiplet between 3.95 and 4.28 ppm) are separated from the signals of the protons on C10 and C15 (C10H2 and C15H2; multiplet between 4.29 and 4.55 ppm) in the minimum of the valley cut between the signal groups.

[0128] The molar proportion of 2-methylbutyl radicals (x2-methylbutyl) can be calculated by the formula

x2-methylbutyl=I(OC19H2)/(I(OCH2)

by forming the ratio of the intensity of the signals for the OC19H2 protons (I(OC19H2)) to the intensity of all OCH2 protons (I(OCH2)).

[0129] The molar proportion of 3-methylbutyl radicals (x3-methylbutyl) is thus calculated from the difference between the two previous molar proportions and 1.

x3-methylbutyl=1x2-methylbutylxpentyl

TABLE-US-00001 TABLE 1 Preparation of trimellitates by esterification of trimellitic anhydride Composition: Purity proportion of Reaction by radicals in the Alcohol A 1(A) 2(A) time GC ester mixture x. [molor ratio] gl gl t [h] [%] [molor ratio] n-butanol 16.9 52.1 5 97.8 n-butyl n-pentanol 97.4 97.4 9 99.2 n-pentyl n-hexanol 74.8 74.8 2.5 98.0 n-hexyl * 2-methylbutanol/ 95.8 95.8 8.5 99.3 2-methylbutyl/ n-pentanol n-pentyl 1:4 0.20:0.80 * 2-methylbutanol/ 96.7 71.0 8 98.8 2-methylbutyl/ n-pentanol n-pentyl 1:1 0.50:0.50 * 2-methylbutanol/ 95.8 95.8 7 98.9 2-methylbutyl/ 3-methylbutanol/ 3-methylbutyl/ n-pentanol n-pentyl 1:1:1 0.33:0.33:0.34 * 2-methylbutanol/ 95.8 95.8 8 99.0 2-methylbutyl/ n-pentanol n-pentyl 4:1 0.79:0.21 * inventive

Example 8

Production of Dryblends, Rolled Sheets and Pressed Plaques

[0130] The test specimens required for the examples which follow are produced by dry mixing (dryblend production), calendering (rolling) and pressing of the following formulations:

TABLE-US-00002 TABLE 2 Drybend formulation phr PVC (Solvin S271 PC; from Solvay) 100 Ester or ester mixture from Example 1, 2, 3, 4*, 5*, 6*, 7*, 67 Tri-2-ethylhexyl trimellitate or DINP Co-stabilizer-epoxidized soya bean oil (Drapex 39 ex Galata) 3 Thermal stabilizer based on Ba/Zn (Mark BZ 965 ex Galata) 2 Processing aid-fatty acid salts (Mark CD 41-0137 ex Galata) 04 phr: parts per hundred parts resin Tri-2-ethylhexyl trimellitate: trioctyl trimellitate ex Sigma Aldrich, purity 99% DINP: Vestinol 9 ex Evonik Industries, purity > 99%

[0131] With dry mixtures, which are referred to as dryblends, it is possible, for example, after thermoplastic processing (e.g. calendering or extrusion) to produce cable and wire insulation, hoses or floors and roofing membranes.

[0132] The dryblends were produced in a Brabender planetary mixer. A thermostat (Lauda RC6) ensured temperature control of the mixing vessel in the planetary mixer. A PC recorded the data sent by the mixer.

[0133] The Winmix software was used to set the following parameters in the Brabender planetary mixer: [0134] Speed program: active [0135] Profile: speed 50 rpm; hold time: 9 min; rise time (of speed): 1 min; [0136] speed 100 rpm; hold time: 20 min [0137] Temperature: 88 C. [0138] Measurement range: 2 Nm [0139] Damping: 3

[0140] The temperature in the mixing vessel was 88 C. after the one-hoar equilibration period. Once the planetary mixer had conducted an internal calibration, the solid constituents (PVC, stabilizer), which had been weighed out beforehand in four times the amount (four times the amount in g based on Table 2 in phr) into a PE cup on a balance (Mettler XS6002S), were fed to the mixing vessel via a solids funnel and the filling stub present in the Brabender mixing vessel. The program was started and the powder mixture was stirred and equilibrated in the mixing vessel for 10 minutes, before the liquid constituents, which had likewise been weighed out in four times the amount in a PE cup on the balance, were fed in via a liquid funnel and the filling stub present in the Brabender mixing vessel. The mixture was stirred in a planetary mixer for a further 20 minutes. After the program had ended, the finished dry mixture (dryblend) was removed.

[0141] These dry blends were used to produce rolled sheets. The rolled sheets were produced on a Collin W150 AP calender. The Collin calender has an automatic sample turner and its temperature is controlled by means of an additional oil thermostat (Single STO 1-6-12-DM). Control was effected by means of Collin software.

[0142] A five-stage program was used to produce the rolled sheet

TABLE-US-00003 Gap Temp. Duration width Speed Stage Designation [ C.] [s] [mm] [rpm] 1 Plastification of the dryblend 165 60 0.2 5 2 Increasing the gap size 165 30 0.5 20 3 Activation of the sample turner 165 170 0.5 20 4 Rolled sheet optimization 165 30 0.5 25 5 Rolled sheet removal 165 60 0.5 7

[0143] On attainment of the roll temperature, the roll gap was calibrated. To start the measurement, the roll gap was adjusted to 0.2 mm. 160 g of each dryblend were weighed in and introduced into the roll gap with the rollers stationary. The program was started. The rollers started with a circumferential speed of 5 rpm and a friction of 20%. After about 1 mi, the plastification was complete for the most part, and the roll gap was increased to 0.5 mm. Homogenization was effected 6 times by means of the automatic turning unit in the calender. After about 6 min, the rolled sheet was removed from the roller and cooled.

[0144] The pressed plaques were produced with a Collin laboratory press. The prefabricated rolled sheets (see above) were used to produce the pressed plaques. The lateral edges of the rolled sheets were removed with the air of a cutting machine, then the rolled sheet was cut into pieces of about 14.514.5 cm in size. For pressed plaques of thickness 1 mm, 2 rolled sheet pieces in each case were placed one on top of the other into the stainless steel pressing frame of size 1515 cm.

[0145] A three-stage program was used to produce the pressed plaques;

TABLE-US-00004 Temp, Pressure Stage Designation [C.] [bar] Duration[s] 1 Initial pressing 175 5 60 2 Pressing 175 200 120 3 Cooling 40 200 270

[0146] The excess compression lip was removed after the press plaques had been produced.

[0147] Example 9

Plasticizer Absorption in the Dryblend

[0148] During the production of each dryblend, the rate of plasticizer absorption was determined, which serves as a measure of processability. A short plasticizer absorption time results in practice in low processing temperatures of elevated product throughput in an existing system.

[0149] The degree of plasticizer absorption was determined via the torque of the mixer as a function of time in the course of dryblend production.

[0150] For this purpose, by means of the hard- and software described in Example 8 for the dryblend production, the torque of the mixer was plotted against time. The torque rose to a maximum value from the addition of plasticizer or the plasticizer composition and dropped again from this, in order finally to attain a constant value or to drop further only minimally from what is called the dry point.

[0151] The period of time between the addition of the plasticizer or the plasticizer composition and the dry point is referred to as time t for plasticizer absorption.

TABLE-US-00005 TABLE 3 Drybend plasticizer absorption Radical(s) in the Time t [min] for trialkyl trimellitate plasticizer (mixture) absorption Ester from Example 3 n-hexyl 4.9 Trioctyl trimellitate ex 2-ethylhexyl 9.0 Sigma Aldrich, purity 99% Ester mixture from 2-methylbutyl/n-pentyl 4.1 Example 4* Molar ratio 0.20:0.80 Ester mixture from 2-methylbutyl/n-pentyl 4.4 Example 5* Molar ratio 0.50:0.50 Ester mixture from 2-methylbutyl/3- 4.5 Example 6* methylbutyl/ n-pentyl Molar ratio 0.33:0.33:0.34 Ester mixture from 2-methylbutyl/n-pentyl 4.6 Example 7* Molar ratio 0.79:0.21 Vestinol 9 ex Evonik Comparative substance: 5.4 Industries, purity > 99% diisononyl phthalate (DINP) *inventive

[0152] In the case of the inventive mixtures of triisopentyl esters of trimellitic acid from Examples 4*, 5*, 6* and 7*, homogenous dryblends were formed within a shorter time than in the case of diisononly phthalate, the tri-n-hexyl ester of trimellitic acid (from Example 3) and the tri-2-ethylhexyl ester of trimellitic acid. It can be concluded directly from this that mixtures according to the invention have better processability with PVC than DINP, than the tri-n-hexyl ester of trimellitic acid and than the tri-2-ethylhexyl ester of trimellitic acid. Moreover, it an be inferred from the values in Table 3 that processability with PVC in mixtures according to the invention rises with the proportion of linear pentyl radicals in the pentyl radicals incorporated in the ester mixture.

[0153] The tri-2-ethylhexyl ester of trimellitic acid has a particularly high plasticizer absorption time and is therefore unsuitable as a plasticizer for numerous PVC applications, which is the reason why the migration properties thereof were not determined hereinafter.

Example 10

Determination of Migration Properties

[0154] The test for determination of the migration properties of a test specimen comprising a plasticizer composition allows conclusions as to how plasticizer-containing formulations behave in application on direct contact with other materialsfor example on contact between a plasticized PVC layer and a rigid PVC layer (called multilayer systems). If, for example, plasticizer migrates within an article from a plasticized polymer-containing component to a non-plasticized polymer-containing component, this can lead to unwanted embrittlement of the plasticized component and softening of the non-plasticized component, which can lead to leaks, declining stability and reduced service life of the article. Accordingly, a low migration tendency is an important property for a plasticizer usable in numerous application or a corresponding component of a plasticizer composition.

[0155] The migration tendency of the inventive mixtures of triisopentyl esters of trimellitic acid and the comparative substances DINP, tri-n-butyl trimellitate, tri-n-pentyl trimellitate and tri-n-hexyl trimellitate was determined as follows:

[0156] The press plaques of thickness 1 mm produced in Example 8 were cut into blanks of 100 mm.Math.100 mm and stored at 23 C. for 24 hours before their mass was determined.

[0157] In accordance with the method DIN EN ISO 177 (published 1999), the blanks were positioned between two contact sheets (Example 10.1: high-impact polystyrene (HIPS) (manufacturer: VINK Kunststoffe; 1001002 mm plaques); Example 10.2; non-plasticized PVC (rPVC) (manufacturer: VINK Kunststoffe; 1001002 mm plaques)) which had the same dimensions as the blanks apart from a thickness of 2 mm, and two of the layer constructions that arise in this way were stacked one on top of the other. These stacks each comprising blanks of the same plasticizer, weighed down with a weight of 2 kg, were stored in an oven (70(1) C. for 28 days.

[0158] After 14 days and after 28 days, the loss of weight of each blank was determined (per cent weight loss based on the original weight). After the first measurement, the stacks were reassembled again as before.

[0159] Table 4 shows the averaged Values of the weight loss of the test specimens.

TABLE-US-00006 TABLE 4 Migration into high-impact polystyrene (HIPS) after 14 and 28 days Radical(s) in the Difference in Difference in trialkyl weight after weight after Blank contains trimellitate 14 days 28 days . . . (mixture) in % in % Ester from n-butyl 5.8 8.0 Example 1 Ester from n-pentyl 5.1 6.9 Example 2 Ester from n-hexyl 3.7 5.6 Example 3 Ester mixture from 2-methylbutyl/n-pentyl 3.6 5.5 Example 4* Molar ratio 0.20:0.80 Ester mixture from 2-methylbutyl/n-pentyl 2.0 3.2 Example 5* Molar ratio 0.50:0.50 Ester mixture from 2-methylbutyl/3- 2.5 3.7 Example 6* methylbutyl/ n-pentyl Molar ratio 0.33:0.33:0.34 Ester mixture from 2-methylbutyl/n-pentyl 1.6 2.7 Example 7* Molar ratio 0.79:0.21 Vestinol 9 ex Evonik Comparative substance: 11.2 14.6 Industries, purity > 99% diisononyl phthalate (DINP) *inventive

TABLE-US-00007 TABLE 5 Migration into rigid PVC (rPVC) after 14 and 28 days Radical(s) in the Difference in Difference in trialkyl weight after weight after Blank contains trimellitate 14 days 28 days . . . (mixture) in % in % Ester from n-butyl 9.3 11.7 Example 1 Ester from n-pentyl 9.0 11.3 Example 2 Ester from n-hexyl 8.0 10.1 Example 3 Ester mixture from 2-methylbutyl/n-pentyl 6.2 8.3 Example 4* Molar ratio 0.20:0.80 Ester mixture from 2-methylbutyl/n-pentyl 5.9 7.9 Example 5* Molar ratio 0.50:0.50 Ester mixture from 2-methylbutyl/3- 7.4 9.4 Example 6* methylbutyl/ n-pentyl Molar ratio 0.33:0.33:0.34 Ester mixture from 2-methylbutyl/n-pentyl 5.6 7.5 Example 7* Molar ratio 0.79:0.21 Vestinol 9 ex Evonik Comparative substance: 7.7 9.7 Industries, purity > 99% diisononyl phthalate (DINP) *inventive

[0160] It is apparent from Tables 4 and 5 that the inventive mixtures of triisopentyl esters of trimellitic acid consistently have a lower tendency to migrate than DINP and n-butyl, n-pentyl and n-hexyl triesters of trimellitic acid. In addition, it can be inferred from a comparison of the measurements for the blanks comprising the inventive ester mixtures from Examples 5, 6 and 8 that the migration tendency of the mixtures according to the invention decreases with rising proportion of the branched pentyl radicals in the pentyl radicals incorporated in the ester mixture. As the comparison of the measurements for the blanks comprising inventive ester mixtures from Examples 5 to 7 shows, a high proportion of 2-methylbutyl radicals is the branched isomeric pentyl radicals incorporated in the ester mixture has a positive effect on a low migration tendency.