New Selective Method for Quantification of Organosilanol Groups in

Biomacromolecules 2002, 3, 850-856

850

New Selective Method for Quantification of Organosilanol Groups in Silicone Pre-elastomers Anders Karlsson† and Ann-Christine Albertsson*,‡ Medical Products Agency, Box 26, S-751 03 Uppsala, Sweden, and Department of Polymer Technology, Royal Institute of Technology, S-100 44 Stockholm, Sweden Received March 4, 2002; Revised Manuscript Received May 21, 2002

The silicone elastomers used for drug delivery are normally reinforced by fumed silica, which contains a high density of silanol groups. These inorganic silanol groups have to be deactivated in order to avoid stiffening the uncured pre-elastomer, also called creep hardening. One commonly used way of achieving this deactivation is to mix the material with low molecular mass organosilanols at an elevated temperature. It is important to be able to quantify the nonbonded organosilanols remaining in the material after manufacture. Traditional testing does not distinguish between inorganic silanols and organosilanols. A new selective method for the quantification of organosilanol groups in silicone pre-elastomers has therefore been developed. This method is based on derivatization of the silanol groups with a mixture of dimethylphenylchlorosilane and tetramethyldiphenylsilazane, so that the silanol groups are replaced with a dimethylphenyl group. The derivatized organosilanols are then determined by liquid chromatography using a size exclusion column and a UV detector. No interference was found from other groups normally present in medical grade preelastomers, such as vinyls, hydrides, and inorganic bonded silanol on silica or water. The results agreed well with the nonselective Karl Fischer titration for some short chain silanols. Introduction Medical grade silicone elastomers are used in many applications, e.g., in medical devices, as implants, and for controlled drug-release products. Silicone elastomers used for medical applications are normally based on poly(dimethylsiloxane)s (PDMS) and are most commonly cured by hydrosilylation.1 In this process, Si-H adds to vinyl groups in the presence of a small amount of platinum catalyst. The vinyl group is often bonded to a high molecular mass polymer, while the hydride-containing cross-linker is a shortchain polymer. Silicone elastomers are normally reinforced by addition of fumed silica. This type of silica has a lot of silanol groups on the surface. During mixing of the pre-elastomer components, untreated silica tends to agglomerate, which leads to reduction in mechanical strength of the final elastomer. This agglomeration can be overcome by deactivating silanol groups on silica with trimethylchlorosilane or hexamethyldisilazane. However, regarding high-consistency pre-elastomers, it was found early that the inorganic silanol groups must be deactivated in order to avoid stiffening the uncured pre-elastomer, also called creep hardening.2 A commonly used way of achieving this kind of deactivation is to mix the fumed silica with low molecular mass organosilanols at an elevated temperature.2,3 Excess organosilanols remaining in the material after the manufacture of the pre-elastomers may affect the rheological properties of the pre-elastomer, influence drug delivery, and cause leaching from the final † ‡

Medical Products Agency. Royal Institute of Technology.

medical product. It is therefore of importance to quantify these nonbonded organosilanols. Figure 1 shows two different types of silanols present in silicone pre-elastomers, inorganic and organosilanol. In addition, the organosilanol may be free, hydrogen bonded, or covalently bonded to the silica. In the development of new medical silicone-based products, it is important to have well-characterized raw materials. The different components in the pre-elastomers can be quantified using several different techniques. The vinyl content can be measured by proton NMR, hydrides in the cross-linker can be determined by infrared spectroscopy or proton NMR, and the unsaturated olefin inhibitor can be measured by headspace gas chromatography.4 Atomic absorption spectroscopy is a suitable technique for quantification of platinum metal from the catalyst, and the silica filler content can be determined by thermogravimetry. However, the free organosilanol content is more difficult to quantify. The organosilanols, also called hydroxy-terminated PDMS, are produced by hydrolysis of dimethyldichlorosilane.

Organosilanols are stable but they may condense, especially at high temperatures or in the presence of an alkaline catalyst, with an increase in the average molecular mass. They may also react with silanol groups on silica surfaces to form siloxane bonds or strongly attach to the silanols by hydrogen bonding and thereby deactivate the silica.

10.1021/bm025531p CCC: $22.00 © 2002 American Chemical Society Published on Web 06/19/2002

Organosilanol Groups in Silicone Pre-elastomers

Biomacromolecules, Vol. 3, No. 4, 2002 851

Experimental Section

Figure 1. Two different types of silanol present in silicone preelastomers.

A variety of different methods have been used to estimate the silanol content, but a major weakness of most of these methods is that they do not distinguish between inorganic silanols and organosilanols. The use a Grignard reagent, such as methylmagnesium iodide, leads to the production of molar amounts of methane which have been measured manometrically.5 Another method suitable for manometrical measurement is treatment with the strong reducing agent lithium aluminum hydride, which leads to the formation of hydrogen gas.6 Several titration methods have also been described. In potentiometric titrations, silanols can be titrated as weak acids with a strong base, e.g., lithium aluminum dibutylamide.7 Karl Fischer titration has been widely used for this application. In this method, silanols and methanol (solvent) react and produce silyl ether with water as a byproduct, which simultaneously reacts with the Karl Fischer reagent.8 A major disadvantage of all these methods is, however, that they are not specific for organosilanols. Both water and silanol groups on silica present in pre-elastomers will be determined together. With spectroscopic techniques such as IR and 1H NMR it is difficult to differentiate signals from inorganic silanols, organosilanols, and water. It seems to be possible to determine organosilanol using 29Si NMR,9 but although the technique is highly selective, the sensitivity is fairly low. Several chromatographic methods have also been developed, and they appear to be more or less selective for silanols. The use of reverse-phase liquid chromatography or size exclusion chromatography in combination with an inductively coupled plasma atomic emission spectrometer made the detection highly sensitive and also selective for silicones but, unfortunately, not selective for organosilanols.10 Using an infrared spectrometer as detector for LC makes the determination fairly selective but not very sensitive.11 Gas chromatography (GC) equipped with a mass spectrometric detector can be used for silanol identification and quantification,12 but this technique can only be used for low molecular mass silanols which can be vaporized in the GC injector. Derivatization of the silanol groups makes it to some extent possible to extend the mass range of silanols which can analyzed by GC. On the other hand, for the smaller silanols, GC is an excellent technique with a high separating power, and a flame ionization detector or mass spectrometer makes this technique very sensitive. Another drawback when using the chromatographic methods is the lack of suitable reference standards. The organosilanol standards available are mixtures of polymers. This paper presents a new sensitive and selective method for the determination of the total amount of organosilanols in silicone pre-elastomers.

Materials. Pyridine, with a water content less than 0.01%, nonstabilized tetrahydrofuran (THF) (LiChrosolve), toluene, methanol, and methyl Cellosolve (2- methoxyethanol) were supplied by MERCK Eurolab, Stockholm, Sweden. The THF which was used for derivatization was stored with a molecular sieve (4A) in a closed bottle to reduce the water content. The Karl Fischer reagent, Hydranal Composite 5, was obtained from Riedel-de Haen, Germany. The silylation agents, dimethylphenylchlorosilane (DMPSCl) and 1,3diphenyl-1,1,3,3-tetramethyldisilazane (DPTMDS) were supplied by Sigma-Aldrich, Stockholm, Sweden.

Fumed silica, Aerosil 200, was supplied by Degussa, Frankfurt, Germany. Hydroxy-terminated poly(dimethylsiloxane)s PS340, PS340.5, PS341, and PS342.5 (organosilanol fluids), dimethylvinyl-terminated poly(dimethylsiloxane) (PS437), and methylhydro (30-35%) dimethylsiloxane copolymer (PS123) were all obtained from ABCR GmbH, Karlsruhe, Germany. Three other organosilanol fluids, DCI, DC-II, and DC-III, and several batches of the pre-elastomer Silastic Q7-4735 Part A and Part B were obtained from Dow Corning, Sophia Antipolis, France. The hydroxyl content of the organosilanol fluids was determined by Karl Fischer titration. The principle of this titration is that silanol groups, both inorganic and organic, condense with methanol to form an equivalent molar amount of water.

The condensation reaction is driven to completion as the water immediately reacts with the Karl Fischer reagent. The titration was performed with a titroprocessor, E 682, and a dosimat, E 665, from Metrohm Ltd, Switzerland. The titration end point, when an excess of iodine remained in solution, was detected amperiometrically with a twin-polarized platinum electrode. Approximately 200 mg of sample was dissolved in 100 mL of a pretitrated water-free methanoltoluene (1:1) mixture. The following instrumental parameters were used: delay time, 10 s; extraction time, 300 s; end point potential (EP), 250 mV. The titration was performed at room temperature. Water present in the standard samples was codetermined with silanol by this procedure, and this caused an error in the hydroxyl determination. A separate specific water determination was therefore made to correct for this

852

Biomacromolecules, Vol. 3, No. 4, 2002

Karlsson and Albertsson

Table 1. Hydroxyl Content Found by Karl Fischer Titration and Calculated Number Average Molecular Mass (〈Mn〉) in the Organosilanol Standards, Triplicate Analyses silanol standard PS340 DC-I DC-II DC-III PS340.5 PS341 PS342.5

% hydroxyl content 6.1-6.3 4.2-4.3 3.3-3.4 2.5-2.6 0.9-1.2 0.8-0.9 0.2-0.2

〈Mn〉 (calcd) 550 800 1000 1300 3200 4000 17000

Table 2. Experimental Design Factors: Investigated Factors and Their Settings factor

low (-)

medium (0)

high (+)

sample mass (mg) reaction time (min) reaction temperature (°C) amount of pyridine (µL) volume of DMPSCla (µL)

10 5 20 0 10

55 17.5 35 50 50

100 30 50 100 90

a Total DMPSCl + DPTMDS ) 100 mL. DMPSCl ) dimethylphenylchlorosilane. DPTMDS ) diphenyltetramethylsilazane.

Table 3. Experimental Conditions and Normalized Peak Areasa

error. In this determination, methyl Cellosolve (2-methoxyethanol) was used as solvent, since negligible condensation of silanol groups occurs with methyl Cellosolve.13,14 Otherwise, the procedure was the same as for the hydroxyl content titration. Data for the organosilanol standards found by this procedure are given in Table 1. Method Development and Optimization. The aim was to develop a method to selectively quantify silicone-bonded hydroxyl groups (organosilanols) in the pre-elastomers. The approach was to combine silanol derivatization with liquid chromatography using a small pore size exclusion column (SEC). All polymeric silicones would then elute within a rather small volume, but only the derivatized polymers would be detected. Derivatization can be achieved with several different agents, to introduce a chemical group to the silanols suitable for detection in a chromatographic system. We chose to react the silanol groups with a mixture of two different agents, similar to a procedure described earlier for trimethylsilylation.15,16 The two agents, dimethylphenylchlorosilane (DMPSCl) and 1,3-diphenyl-1,1,3,3-tetramethyldisilazane (DPTMDS), replaced the hydroxyl group on the silanol with a dimethylphenyl group, suitable for UV detection in a chromatographic system. In the first study, 100 mg of the reference silanol PS340 was dissolved in 5.0 mL of solvent. One hundred microliters of pyridine and 100 µL of each derivatizing agent were then added. After 60 min at room temperature, the reaction was stopped by the addition of 1.0 mL of ethanol to deactivate the reagents. Different solvents were initially evaluated. More or less poor recoveries were found with hexane, methylene chloride, toluene, and ethyl acetate as solvent, although these are all good solvents for PDMS. The highest recovery was found when nonstabilized tetrahydrofuran (THF) was used as derivatization solvent, and THF was thus chosen as the chromatographic eluent. After the initial tests, an experimental design screening study was performed in order to clarify how different factors influence the derivatization yield. The factors and the experimental design are given in Tables 2 and 3. The temperatures of the solutions were controlled by waterbaths. The first eluting part in the chromatogram in each run was integrated (approximately 5-6.5 min), which corresponds to the derivatized organosilanols. The peak areas were then normalized with respect to the sample mass and final volume of the sample solution. The statistical evaluation was carried out with Modde ver 3.0, Umetri AB, Sweden, using multiple linear regression.

expt chlorono. run order mass time temp pyridine silane area (norm) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

11 19 17 10 7 6 5 2 4 12 18 16 8 15 3 9 13 1 14

+ + + + + + + + 0 0 0

+ + + + + + + + 0 0 0

+ + + + + + + + 0 0 0

+ + + + + + + + 0 0 0

+ + + + + + + + 0 0 0

418256 1150 510 723803 336 772863 210555 14425 790149 1543325 1625385 1228927 1759900 686182 1647252 1584078 1596147 1601930 1596541

a The conditions of the settings -, 0, and + are found in Table 1. The areas have been normalized to 10 mg of sample.

Figure 2. Variables and combination of variables which had the greatest influence on the derivatization yield. Only the amount of pyridine and the ratio of reagents were significant.

Results The primary results of the optimization of the derivatization step in the organosilanol determination are shown in Table 3. Low, medium, and high point settings of the five variables, sample mass, reaction time, reaction temperature, amount of pyridine, and ratio of the derivatization agents were combined in a factorial experiment. The peak areas were normalized with respect to sample mass and solution volume. Figure 2 show a graphical presentation of the variables and combination of variables which had the greatest effect on the derivatization yield. It was found that only the amount



Figure 3. The full chromatogram for 100 mg of the reference standard, PS340, 〈Mn〉 ) 550, together with a blank. The first eluting region has been expanded, showing the organosilanol peak.

of pyridine and the ratio of silanisation agents had any significant effect. The medium and high settings gave approximately the same result in both cases. In the final selection of factor settings, the appearance of the total chromatogram was also considered. For the high setting of DMPSCl a large extra peak eluted at approximately 16 min with strong tailing. As a consequence, the chromatographic run time had to be prolonged to 30 min for these samples. This peak was not found for the medium and lower settings. These data where used to develop an optimized derivatization procedure. The chromatographic conditions were also optimized. The small pore, 50 Å, size exclusion column was selected in order to separate the short-chain silanols from the derivatization agents. The optimal detection wavelength was selected after studying repeated injections of the same sample while varying the wavelength between 250 and 270 nm. A wavelength of 264 nm gave the maximum peak area. To gain maximum sensitivity, a large injection volume should be used. However, to maintain the peak resolution between the derivatized silanol and the reagents in the chromatogram, the injection volume should not be too large. Injection volumes up to 200 µL could be applied without losing too much resolution. The resolution can be even better if a more efficient column is used or if several columns are used in series. Final Method Description. Approximately 100 mg of sample or the reference standard silanol oil, PS340, is weighed into 10 mL glass tubes. The materials are dissolved in 5.0 mL of THF, 100 µL of pyridine, 50 µL of DMPSCl, and 50 µL of DMTMDS are then added, and the solution is thoroughly mixed. After 30 min at room temperature, 1.0 mL of ethanol is added to decompose the remaining derivatization agents. After another 30 min, the supernatant solution is filtered through a 0.5 µm filter into autosampler vials and injected into the column.

The liquid chromatography system consists of a pump M616, an autoinjector, WISP 712, a detector, M486, and a Millennium 2020 chromatography data system from Waters, USA. The size exclusion column, Plgel, Polymer Labs, England, is 7.5 mm × 300 mm with 5 µm particles of crosslinked polystyrene with a pore size of 50 Å. Single-use 25 mm PTFE filters with a pore size of 0.5 µm Millipore, USA, are used. The chromatography conditions are as follows: The flow rate of THF is 1.0 mL/min, the injection volume is 25 µL, and the detection wavelength is 264 nm. The peak area was obtained by integration after first setting a straight baseline from the start to the end of the run. The area was measured between approximately 5.0 and 6.5 min. The area corresponding to a blank is also measured using the same integration time limits. Figure 3 shows the whole chromatogram for approximately 100 mg of the reference standard together with a blank. The major peaks are components from the derivatization reagents and pyridine. An expansion of the first part of this chromatogram shows the region were the organosilanol peak elutes. Validation. The method for silanol determination was validated with regard to selectivity, linearity, detection limit (D.L.), precision, and accuracy. The selectivity of the method was investigated with regard to the different compounds present in a silicone pre-elastomer for hydrosilylation. Silicone oils containing vinyls (PS437), hydrides (PS123), and fumed silica (Aerosil 200) were analyzed according to the described method. None of these compounds gave any response. In addition, these compounds were also included in mixtures with the standard silanol oil PS340 to see whether they had any influence on the response, i.e., whether they inhibited the derivatization reaction. The same response was achieved with and without the added compounds. We therefore conclude that the method is selective for organosilanol groups in pre-elastomers.

854


Karlsson and Albertsson

Figure 4. The peak area for the reference standard, PS340, vs amount of silanol hydroxyl groups.

Figure 5. Comparison between results obtained by the new method and the Karl Fischer titration for six organosilanols standards.

The linearity of the response was studied for the standard oil, PS340 (Figure 4). The chromatographic peak height for the largest amount of standard was approximately 1.0 absorbance unit. This is probably the upper limit for linearity, as larger peaks will be outside the linear response of the UV detector. Dilute solutions of PS340 and blank solutions (THF) were used to estimate the detection limit. The injection volume was increased to 200 µL. Three times the standard deviation of the areas representing six blank injections was defined as the detection limit. This area corresponds to 10 µg of PS340 or 0.5 µg of hydroxyl groups. Triplicate silanol determinations were made on all seven silanol oils (Table 1).

The precision (pooled relative standard deviation) of the peak areas was estimated to be 3%. The accuracy of the method was investigated by comparison with data obtained from the Karl Fischer titration. Figure 5 shows that the LC method gives results comparable to those obtained by conventional Karl Fischer titration. Figures 6-8 show chromatograms of the three silanols PS340.5, PS341, and PS342.5. The different sizes of the organosilanol molecules (Table 1) lead to different chromatogram patterns. The reference standard, PS340, shown in Figure 3 has the lowest average molecular mass and is fully separated from the column void. In contrast, the largest organosilanol, PS342.5, elutes at the void of the column


Figure 6. Chromatogram of 94 mg of the organosilanol fluid PS340.5 together with a blank, 〈Mn〉 is 3 200.

Figure 7. Chromatogram of 110 mg of the organosilanol fluid PS341 together with a blank, 〈Mn〉 is 4000.

Figure 8. Chromatogram of 100 mg of the organosilanol fluid PS342.5 together with a blank, 〈Mn〉 is 17 000.

(Figure 8). Furthermore, the higher the average molecular mass, the lower is the area per mass, as can be seen by comparing the scales of the chromatograms. Organosilanol Content in Pre-elastomers. The organosilanol content in six batches of the pre-elastomer Silastic Q7-4735 Part A and Part B was determined by the new LC method. The results for the different samples were quite consistent, and the values obtained corresponded to 0.050.10% OH. Discussion Short-chain organosilanols are normally used for deactivation of silanol groups on the silica surface in silicone preelastomers. Excess of these organosilanols may be a potential migration problem, and they may also influence the rheological properties of the material. A new method for the


determination of organosilanols silicone pre-elastomers has been developed. The new method is sensitive and selective for organosilicone-bonded hydroxyl groups. Other compounds normally present in these pre-elastomers, i.e., vinyls, hydrides, and silanol groups in silica or water, do not interfere. In many other published methods, the interference of especially water and silica-bonded silanol groups is common. This new method is based on derivatization of the organosilanol groups followed by a chromatographic analysis using a size exclusion column for separation and a UV detector. The derivatization is achieved with a mixture of dimethylphenylchlorosilane and a tetramethyldiphenylsilazane in the presence of pyridine as HCl acceptor. Tetramethyldiphenylsilazane alone gave a poor derivatization yield. Dimethylphenylchlorosilane alone gave a good derivatization yield, but there was a problem with a large additional peak in the chromatogram. Furthermore, when dimethylphenylchlorosilane is used alone, hydrochloric acid is formed as a byproduct, and this not compatible with the stainless steel in the chromatographic system. An advantage of using equal amounts of the two derivatization agents is that the byproduct is ammonium chloride, which precipitates and can be filtered off before the chromatographic analysis is performed. In the derivatization of silanols, both organosilanols and inorganic silanols groups, the hydroxyl groups are replaced by a dimethylphenyl group. No other group normally present in a medical grade silicone pre-elastomer reacted with the derivatization agents. The final selectivity for organosilanols was obtained by a chromatographic analysis. By setting the UV detector at 264 nm, only the phenyl group attached to the silanols is monitored. In addition, a column with 50 Å pores of cross-linked polystyrene is able to separate the short-chain organosilanols from the reagents, which gives a high selectivity for this method. The validation of this method showed that the results obtained for six short-chain silanols are in good agreement with the results found by the conventional Karl Fischer titration method. The precision was estimated at 3% and the limit of detection to 0.005% OH. The method is linear up to a least 10% OH. The new method was used to measure the organosilanol content in some pre-elastomers. It was found that the organosilanol content was in the range between 0.05 and 0.10% OH. References and Notes (1) Szycher, M. Biocompatible Polymers, metals and Composites; Technomic Publishing Co., Inc.: Lancaster, PA, 1983. (2) Warrick, E. L. Rubber Chem. Technol. 1976, 49, 909. (3) Polmanteer, K. E. Rubber Chem. Technol. 1981, 54, 1051. (4) Karlsson, A.; Singh, S.; Ablertsson, A.-C. J. Appl. Polym. Sci. 2001, 79, 2349. (5) Guenther, F. O. Anal. Chem. 1958, 30, 1118. (6) Barnes, G. H.; Daughenbaugh, N. E. Anal. Chem. 1963, 35, 1308. (7) Kellum, G. E.; Uglum, K. L. Anal. Chem. 1967, 39, 1623. (8) Kellum, G. E.; Smith, R. C. Anal. Chem. 1967, 39, 1877. (9) Kendrick, T. C.; Pharbhoo, B.; White, J. W. Siloxane polymers and copymers. In The Chemistry of Organic Silicon Compounds; Wiley: New York, 1989; p 1289.

856


(10) Dorn, S. B.; Frame, E. M. S. Analyst 1994, 119, 1687. (11) Kohn, E. M. Energy Res. Abstr. 1985, 10, No. 13664. (12) Smith, A. L. The Analytical Chemistry of Silicones; John Wiley & Sons, Inc.: New York, 1991. (13) Smith, R. C.; Kellum, G. E. Anal. Chem. 1966, 38, 647. (14) Mika, V.; Cadersky, I. Z. Anal. Chem. 1972, 258, 25.

Karlsson and Albertsson (15) Pierce, Alan E. Silylation of Organic Compounds; a technique for gasphase analysis; Pierce Chemical Company: Rockford IL, 1968. (16) Sweely, C. C.; Bently, R.; Makita, M.; Wells, W. W. J. Am. Chem. Soc. 1963, 85, 2497.

BM025531P

New Selective Method for Quantification of Organosilanol Groups in

Recommend Documents