ANALYST, APRIL 1989, VOL. 114 44s Fourier Transform Infrared Spectrometric Determination of Trace Amounts of Polydimethylsiloxane in Extracts of Plastics Additives Pierre Fux Central Analytical Department, FO 3.32, Ciba-Geig y Ltd., 4002 Bade, Switzerland Fourier transform infrared (FT-IR) spectrometry has been used to determine trace amounts of polydimethylsiloxane in a plastics additive. The efficiency of the proposed extraction method and the utility of the developed FT-IR method were used to determine siloxanes at the vg g-1 level in plastics additives. Various evaluation possibilities of the recorded spectra were compared both in the presence and absence of interfering absorption bands arising from simultaneously extracted components. The method was applied successfully to the determination of siloxanes in a number of other additives.Keywords : Silicones; pol ydime th ylsiloxan e determination; Fourier transform in fra red spectrometry; plastics additives; extraction The entrainment of trace amounts of polysiloxanes in indus- trial products cannot be eliminated completely. By modifying the surface tension, such polysiloxanes can affect adversely the properties of a product and hence its technical applica- tions. It is, therefore, important that analytical methods are available for the determination of trace amounts of siloxanes. Infrared (1R) spectrometry has always been a valuable method for the identification of polysiloxanes. Since the 1960s "classical" IR spectrometry has been used essentially to detect silicones in various substrates,'-' but reliable quantification has been possible only when relatively high polysiloxane concentrations are present in the matrix.6.7 Little work on the determination of trace amounts of silicones has been reported: their determination by inductively coupled plasma atomic emission spectrometry8.9 and atomic absorption spec- trometryl".ll has been described.In the latter instance only the total silicone content was determined and no information was given on the structure of the siloxane present. Few investigations into the determination of trace amounts of silicones have been carried out using infrared spec- trometry. 12.13 The development of Fourier transform infrared (FT-IR) spectrometry and also FT-IR with attenuated total reflectance (ATR FT-IR) has increased the importance of these methods as a tool for the quantification of polysiloxanes, particularly within the last five years.The main areas of investigation have concerned quantitative surface analy- sis,14.15 the determination of silanol in silicones16 and the determination of polydimethylsiloxane on cotton fabrics" and on human skin. 18 Recently, diffuse reflectance infrared Fourier transform spectrometry (DRIFTS) has been used to study the deposition of siloxanes on keratin surfaces. 19 In this paper the quantification of trace amounts of polydimethylsiloxane (PDMS) in a plastics additive using FT-IR spectrometry is described. A liquid - liquid extraction method is used for additives that are insoluble in pentane. The additive is dissolved in N,N-dimethylformamide and the silicones are extracted with pentane using a continuous extractor. Experimental Instrumentation Infrared spectra were recorded in a 4-mm cell on a Perkin- Elmer 1710 FT-IR spectrometer, accumulating 20 scans with a resolution of 4 cm-1.From the recorded spectra the net absorbance values were evaluated using a Perkin-Elmer Model 3600 data station and the QUANT-SINGLE program. Liquid - liquid extractions were performed with a continu- ous liquid - liquid extractor according to Ludwig (Normag, D-6238, Hofheim am Taunus: Code 2071). Reagents Polydimethylsiloxane (Dimethicone 350) was supplied by Wacker. The following solvents, with very low silicone contents, were used: pentane, Burdick and Jackson quality (Fluka); carbon disulphide , infrared spectroscopic grade (Fluka) ; and N,N-dimethylformamide (DMF) , analytical-reagent grade (Fluka). Preparation of the Test Samples Extraction procedure Approximately 200 g of the additive were dissolved in 200 ml of DMF and extracted for 4 h at 60 "C with 500 ml of pentane in a continuous liquid - liquid extractor according to Ludwig. The pentane emerged from the stirred DMF phase through the perforated, centred tube as a stream of fine droplets; this tube allowed good interfacial contact. The solution was left to stand at room room temperature for 1 h to optimise the phase separation.To remove any remaining additive, the pentane phase was extracted a further 5-7 times with 50 ml of DMF using a separating funnel. A 250-ml portion of the purified pentane solution was rotary evaporated to a volume of approximately 10 ml.This solution was then evaporated to dryness for 2 h at about 50 "C under high vacuum. To the dry residue 0.6-2 ml of carbon disulphide were added depending on the intensity of the IR spectrum. (A blank run, i . ~ . , the extraction of a solvent mixture, was carried out in parallel with the analytical extraction in order to detect whether there was any contamination of the reagents or apparatus with sili- cones.) Preparation of Spiked Test Samples Known amounts of PDMS as a solution in pentane were added to the samples, e.g., 0.05, 0.2, 0.5 and 1 .O mg of PDMS were added to 200 g of the liquid sample. The extraction was performed as described above for the unspiked samples. Preparation of Calibration Solutions Standard solutions (0.02, 0.04, 0.06, 0.08, 0.10 and 0.20 mg ml-1) of PDMS in carbon disulphide were prepared; their IR spectra are shown in Fig.1.446 ANALYST, APRIL 1989, VOL. 114 FT-IR Measurements Polydimethylsiloxane shows a characteristic narrow absorp- tion band at 1260 cm-1, a double band at 1090 and 1020 cm-l and a broad band at 805 cm-1 (Fig. I). The band at 1260 cm-I is generally used for quantificationl.7,’2.lX.t~ because of its sharpness. For reliable quantitative analysis it is important to verify the absorbance ratios of these characteristic bands. In this work, the ratio of the absorbance of the band at 1260 cm-1 (Al26O) to that of the band at 805 cm-1 (Axos) was calculated (Table 1). If the silicone bands are actually single bands, then the ratio of A1260 to Axos will be 1.10 k 0.02.The IR spectra of the sample and calibration solutions were recorded from 700 to 2000 cm-1 in a 1- or 4-mm cell. The absorption of the solvent was compensated for by spectral subtraction. Calibration graphs were constructed at 80.5, 818 (shoulder of the band at 805 cm-I), 1020 and 1260 cm-1 by plotting absorbance versus milligrams of PDMS per millilitre of CS?. The base-line point was taken to be approximately 1900 cm-1. Results and Discussion Evaluation Possibilities Using the IR Measurements If the bands at 1260 and 805 cm-I were “proper” bands, i.e., A1260:AK05 was within the range l . C l . 2 , then the recorded spectrum of the sample could be evaluated from either the first or the second band, using the corresponding calibration graph.If the IR spectrum contained interfering bands (e.g., a narrow band at 805 cm-1 or an intense overlapping band at 1260 cm-1 so that this typical band was not visible) from simultaneously extracted components (impurities, residual traces of additive, etc.) that could affect quantification, then the evaluation was based on the absorbance at 818 cm-1. The PDMS content determined in this way has to be considered as the maximum concentration in the sample. s 85 t 2000 1800 1600 1400 1200 1000 800 Wavenu m bericm - 1 Fig. 1. Infrared spectra of PDMS standard solutions in CS2 in a 4-mm cell. A. 0.02: B. 0.04; C. 0.06: D , 0.08: E. 0.10: and F, 0.20 mg rnl I Table 1. Absorbance ratio, A Il_ho : AH(,5, for standard solutions of PDMS in CS2 in a 4-mm cell Standard w lu t i o ni mg ot PDMS per Abcorbance at Absorbance at ml of CS2 1260 cm 805 cni- 1 A I X ~ : AH05 0.0200 0.0490 0.0453 1 .(I82 0.0400 0.092 1 0,0849 1.085 0.0600 0.1402 0.1271 1,103 0.0800 0.1861 0.1682 1.106 0.1000 0.2339 0.2 1 3 8 1.094 0.2000 0.4521 0.4027 1.123 Recovery of PDMS Extracted from DMF Solutions Solutions of PDMS in DMF in the concentration range 0.1-1.5 pg g-1 were extracted with pentane as described under Extraction procedure; the average recovery was greater than 98%.The various concentration steps necessary to obtain a dry residue were carried out with no loss of PDMS. Reproducibility and Recovery of the Sample Extraction About 200 g of the test substance A1 were extracted alone and also after the addition of 0.05,0.52 and 1.03 mg of PDMS.The results are given in Table 2. For the unspiked sample, only the band at 1260 cm-1 was used in the calculation procedure because the intensity of the band at 805 cm-1 was substantially different from that of the former and from the intensity of the band at 1090 cm-1. For the spiked samples no interfering bands were visible in the recorded spectra, A,26o : Axo5 varying from 1.085 to 1.207. Quantification could be performed at either 805 or 1260 cm- l. The recovery of PDMS from an extracted sample could therefore be determined from these experiments. Extrapol- ation to zero PDMS added, based on a linear regression analysis of the six points (correlation coefficient = 0.9986), gave a PDMS concentration of 1.61 pg g-1 in the sample. Direct extraction of the unspiked sample gave a PDMS content of 1.57 pg g-1.This means that a recovery of about 98% for the silicones extracted from the plastics additive was achieved. To check that the extraction was quantitative, a series of unspiked and spiked samples were extracted. The results are summarised in Table 3. Impurities which changed the shape of the broad absorption band at 805 cm-1 were Table 2. Determination of PDMS content for unspiked and spiked additive sample A1 Amount of sample A 1 extracted/ & 206 222 20 1 205 Maximum PDMS con- PDMS con- PDMS con- Amount of tent found by tent found by tent found by PDMS evaluation at evaluation at evaluation at added1 1260 cm-11 805 crn-11 818 cm-’1 v& g- ’ vgg M g ’ w g-‘ - 1.57 Evaluation not possible because of interfering bands 0.23 1.97 1.82 a2.74 2.57 4.24 4.07 15.22 5.02 6.69 6.89 17.36 Table 3.Rccovery of PDMS from the extraction of plastics additive samples Concentration found for the unspiked sample .‘ i Sample pgg-1 A2 . . S1.03 A3 . . S0.45 A4 . . S0.88 No. ( a ) A5 . . a1.78 A6 . . a1.45 A7 . . a0.98 A8 . . G1.11 Amount of PDMS added1 P& & - I (c) 1.40 1.43 1.27 1.09 1.09 1.09 0.33 Concentration found for the Recovery, spiked %, 2.58 110.7 1.93 103.5 2.23 106.3 2.92 104.6 2.61 106.4 2.09 101.8 1.44 100.0 The PDMS content given is the maximum concentration in the sample.ANALYST, APRIL 1989. VOL. 114 447 100 80 60 40 20 8 a, C m 4z 4- .- 100 t I- 2 80 60 40 20 W 2000 1800 1600 1400 1200 1000 800 Wavenum berlcm-1 Fig. 2. Fourier transform infrared spectra of (uj unspiked and ( h j spiked sample A5 (containing approximately 0.2 mg of PDMS in 200 g of suhtance) recorded from 700 to 2000 cm-I in a 4-mm cell (diluted mith 1 ml of CS?) extracted simultaneously for all the samples studied.This band therefore became abnormally sharp. Hence, quantifica- tion was performed by measuring the absorbance of its shoulder at 818 cm-I. The shoulder was clearly visible in all the recorded spectra (see, for example, the spectra of spiked and unspiked sample A5 shown in Fig. 2). From this series of samples, the average recovery was found to be approximately 105 O/” . Comparison of Quantification at 805 cm-1 and “Shoulder” Evaluation at 818 cm-1 Spiked samples of the additive A1 were extracted and the IR spectra evaluated using the band at 805 cm-1.These spectra were also evaluated using the shoulder (at 818 cm-1) of this band. The results are summarised in Table 2. Extrapolation to zero PDMS based on a linear regression analysis (correlation coefficient = 0.9983), gave a PDMS content of 2.58 pg g-1. Quantification using the shoulder at 818 cm-1 resulted in an increase in the PDMS content found of approximately 60% compared with quantification at 805 cm-1. Removal of Interfering Components by Difference Spec- trometry A series of ten samples were extracted and all the recorded IR spectra showed the presence of interfering bands. Fig. 3 shows typical IR spectra of samples A l l and A14 in which strong absorbing bands mask the characteristic silicone bands. To 100 80 60 40 20 $ ai 9 L Y m i I ’ +- ._ $ 100 K F I- 80 60 40 20 2000 1800 1600 00 1200 1000 800 Fig.3. Fourier transform infrared spectra of ( u ) sample A14 and ( h ) sample A1 1 recorded from 700 to 2000 cm I in a 4-mm cell (diluted with 0.8 ml of CS?) remove these interfering bands, difference spectra were generated. A sample from this series, containing less than 0.4 ug g-1 of PDMS but showing qualitatively the same interfering bands, was purified as described under Extraction procedure. This sample was taken as a pure reference sample (“AR”), i.e., silicone-free (containing <0.2 pg g-1 which is the approximate quantification limit of the method). Difference spectra were calculated from the recorded spectra minus this reference spectrum. The absorbance value was defined by measuring the absorbance at 818 cm-1 because this was the wavelength least affected by impurities, and the base-line point was taken to be 1900 cm-1.The results of the quantification are summarised in Table 4. Evaluation of the difference spectra for samples containing 0.2-2.0 pg g-1 of PDMS gave a PDMS content that was approximately 40% lower on average compared with evalua- tion carried out without using difference spectra. Fig. 4(a) shows the difference spectrum of sample A14, which contains approximately 2 pg g-1 of PDMS and the reference substance “AR”. The spectrum is similar to a typical PDMS spectrum with its four characteristic bands. However, for sample A 11, containing only about 0.2 pg g-1 of silicones, the difference spectrum [Fig. 4(b)] shows no evidence of PDMS bands.Comparison of PDMS Quantification by FT-IR Spectrometry and Total Silicone Determination by X-ray Fluorescence Spectrometry Some of the FT-IR results obtained in this work were compared with those obtained for total silicone determination by X-ray fluorescence spectrometry. The total silicone content was determined using the extracts of the plastics additive448 ANALYST, APRIL 1989. VOL. 114 Table 4. Evaluation comparisons for ten additive extracts Amount of sample Sample extract e di No. g A9 . . . . 183.1 A10 . . . . 174.6 A l l . . . . 178.2 A12 . . . . 188.0 A13 . . . . 175.8 A14 . . . . 184.0 A15 . . . . 167.5 A16 . . . . 157.1 A17 . . . . 157.7 A18 . . . . 174.1 Maximum concentration of PDMS found by evaluation at 818 cm-1’ S2.07 S1.20 s1.10 G0.83 60.77 12.32 S1.39 G0.97 61.07 61.32 v g s I PDMS content found by difference spectrometry with a silicone-free reference substance/ v g g- 0.86 0.54 0.24 0.41 0.47 2.06 1 .05 0.68 0.88 1 .oo Table 5.Comparison of X-ray fluorescence spectrovetry and FT-IR spectrometry for the determination of PDMS X-ray fluorescence spectrometry Maximum Amount of PDMS content sample Silicone PDMS content found by FT-TR Sample cxtractcdi found*/ calculatedl spectrometryt/ No. g Ing L% g- ’ I%% I A5 . . 200 56 0.74 d I .78 A6 . . 200 64 0.84 G1.45 A7 . . 200 108 1.43 60.98 A8 . . 165 48 0.79 51.11 A19 . . 200 340 4.49 G1.66 * Total silicone found in the extract. f Values determined by reference to a calibration graph at 818 cm l . samples. The PDMS content was determined by multiplying the total silicone concentration found by a factor of 2.64, taking into account the following structure for PDMS: The results are summarised in Table 5 .It is worth comparing the results given by the two evaluation methods for the samples studied. The values determined spectrometrically for samples A5, A6 and A8 are the maximum concentrations, the actual PDMS concentrations are approximately 40-60% lower. These results are in good agreement with those obtained by X-ray spectrometry. Samples A7 and A19 contain about 3-4 times less PDMS according to the FT-IK measurements than indicated by the X-ray spectrometry results (Table 5 ) . Therefore, the determi- nation of total silicone cannot be used as a means of quantifying PDMS in additives, nor even as a means of measuring their PDMS content. Conclusion Trace amounts of PDMS can be quantified reliably by FT-IR spectrometry, after their complete extraction from industrial products, e.g., plastics additives. Interference resulting from the remaining extracting solvents, additive residues or simul- taneously extracted impurities are eliminated by spectral subtraction. The detection limit of the proposed method is approximately 0.1-0.2 pg g-1 of PDMS in the additive. These resuits concern plastics additives that are insoluble in pentane. However, a solid - liquid extraction of silicones with butan-1-01 has been carried out for solid additives that are 100 80 60 40 $? 2o a i C rn =: .- 5 100 C 2 I- 80 60 40 * O l 1 I I, I I 2000 1800 1600 1400 1200 1000 800 Wavenum bericrn - Fig.4. Difference spectra of ( a ) sample A14 and ( h ) sample A l l with reference substance “AR” (silicone-free) generated on a Perkin-Elmer 1710 spectrometer soluble in pentane; the quantitative determination of PDMS is carried out using the same FT-IR subtraction method. In this instance the detection limit of PDMS is approximately 10 pg g-1 for the additives examined, because the less selective extraction step leads to higher spectral interferences. The author thanks P. Acker for recording and evaluating the IR spectra. 1. 2. 3. 4. 5. 6. 7 . 8. 9. 10. 11. 12. 13. References Pashenkova, L. F., and Yablochkin, V. D., Gig. Sanit., 1968, 33, 55. Danielak, R . , Ludwicki, H., and Ostrowska, E., Acta Pol. Pharrn., 1978, 35, 467.Wieczorek, H . , Seijen, Oele, Fette, Wachse, 1985, 111, 115. Wilkowa, T . , Chern. Anal. (Warsaw), 1976, 21, 399. Mayhan, K. G., Thompson, L. F., and Magdalin, C. F., J . Paint Technol., 1972, 44, 85. Hauptmann, G., Keil, G., and Eberhardt, E., Schrnierst. Schrnierungstech . , 1968, 29, 38. Kojima, S . , and Oba, T., Eisei Shikenjo Hokoku, 1970,88,26. Watanabe, N., Nagase, H., Ose, Y., and Sato, E., Eisei Kugaku, 1985, 31, 391. Watanabe, N., Yasuda, Y., Kato, K., Nakamura, T., Funa- saka, R . , Shimokawa, K., Sato, E . , and Ose, Y., Sci. Total Environ., 1984, 34, 169. Nishijima, M., Kanmuri, M., Takahashi, S . , Kamimura, H., Nakazato, M., and Kimura, Y., Shokuhin Eiseigaku Zasshi. 1975, 16, 110. McCamey, D. A . , Tannelli, D. P., Bryson, L. J . , andThorpe, T. M., Anal. Chim. Acta, 1986, 188, 119. Rotzche, H., Clauss, H., and Hahnewald, H . , Pluste Kautsch., 1979, 26, 630. Sinclair, A., and Hallam, T. R., Analyst, 1971, 96, 149.ANALYST, APRIL 1989, VOL. 114 449 14. Reikichi, I.. and Koji. O., A p p f . Specfrosc., 1984. 38, 359. 15. Fuhrmann, J . , and Glanzer, K.. Pharm. Znd., 1985, 47, 652. 16. Griffith, G . W . , Znd. Eng. Chem. Prod. Res. Dev., 1984, 23, 590. 17. Shreedhara Murthy, K. S., Leyden. D. E., and D'Alonzo, R . P.. Appl. Spectrosc.. 1985, 39, 856. 18. Klimisch. H. M., and Chandra, G., J . SOC. Cosmet. Chem., 1986. 37, 73. 19. Klimisch, H. M., Kohl, G. S . , and Sabourin, J . M., J . SOC. Cosmet. Chem., 1987, 38, 247. Paper 8104076A Received October 13th, 1988 Accepted December 5 t h , 1988