2965
Electrophoresis 2014, 35, 2965–2971
Ingrid Hintersteiner Thomas Schmid Markus Himmelsbach Christian W. Klampfl Wolfgang W. Buchberger Institute of Analytical Chemistry, Johannes Kepler University Linz, Linz, Austria
Received May 27, 2014 Revised July 1, 2014 Accepted July 2, 2014
Research Article
Quantitative analysis of hindered amine light stabilizers by CZE with UV detection and quadrupole TOF mass spectrometric detection The current work describes the development of a CZE method with quadrupole QTOFMS detection and UV detection for the quantitation of Cyasorb 3529, a common hindered amine light stabilizer (HALS), in polymer materials. Analysis of real polymer samples revealed that the oligomer composition of Cyasorb 3529 changes during processing, a fact hampering the development of a straightforward method for quantitation based on calibration with a Cyasorb 3529 standard. To overcome this obstacle in-depth investigations of the oligomer composition of this HALS using QTOF-MS and QTOF-MS/MS had to be performed whereby 22 new oligomer structures, in addition to the ten structures already described, were identified. Finally, a CZE method for quantitative analysis of this HALS was developed starting with a comprehensive characterization of a Cyasorb 3529 standard using CZE-QTOF-MS, subsequently allowing the correct assignment of most Cyasorb 3529 oligomers in an electropherogram with UV detection. Employing the latter detection technique and hexamethyl-melamine as internal standard, peak areas obtained for the melamine could be correlated with those from the triazine ring, the UV-absorbing unit present in the HALS. This approach finally allowed proper quantitation of the single oligomers of Cyasorb 3529, an imperative for the quantitative assessment of this HALS in real polymer samples. Keywords: CZE-MS / Hindered amine light stabilizers / Polymer product analysis / QTOF-MS / UV-stabilization DOI 10.1002/elps.201400265
Additional supporting information may be found in the online version of this article at the publisher’s web-site
1 Introduction As plastic materials find their way into more and more areas of application ranging from rather simple “single use” devices to highly sophisticated components integrated in high performance products, design of such materials is confronted with steadily increasing requirements. These advanced specifications are not only met by development of new base materials,
Correspondence: Ingrid Hintersteiner, Institute of Analytical Chemistry, Johannes Kepler University, Linz, Altenberger Strasse 69, A-4040 Linz, Austria E-mail: [email protected] Fax: +43-732-2468-8679
Abbreviations: HALS, hindered amine light stabilizers; HMM, hexamethyl-melamine; PMP, 1,2,2,6,6-pentamethyl1,2,3,6-tetrahydropyridine; TMP, 2,2,6,6-tetramethyl-1,2,3,6tetrahydropyridine
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
i.e. different types of polymers and polymer blends, but also by use of novel types of polymer additives. Additives are highly important constituents of plastics, playing a fundamental role in achieving materials with unique physical and chemical properties. Within the group of additives, stabilizers can be understood as a prominent subgroup, protecting the polymer against thermal, oxidative and UV-light induced stress [1, 2]. Focusing on the latter type of stabilizers, compounds such as 2-hydroxybenzophenones or 2-hydroxyphenylbenzotriazoles, and hindered amine light stabilizers (HALS), which stand for the latest development in this field, are commonly employed when it comes to UV-protection of plastic materials. Typically, HALS include 2,2,6,6-tetramethylpiperidine as structural element being responsible for the photostabilizing effect, as the secondary amine can act as alkyl radical trapping species by forming a stable nitroxide radical [3]. For better understanding of product performance, determination of stabilizers is of substantial concern at all stages of polymer processing such as production of a stabilized base
www.electrophoresis-journal.com
2966
I. Hintersteiner et al.
polymer, compounding, and further processing steps such as extrusion or injection molding [4–6]. The necessity for qualitative and quantitative analysis of polymer stabilizers but also their degradation products has led to the development of a series of analytical methods. Focusing on the analysis of HALS it has to be taken into account that these stabilizers are technical products with often only limited or no information available concerning purity or actual structural composition of the respective product. Particularly for oligomeric HALS, which are not amenable to GC analysis, only a very small number of reports on analytical methods suitable for the characterization and quantitation of these components can be found in the literature. These comprise pyrolysis GC-MS, an approach enabling identification of several HALS additives [7–9] but unable to differentiate between different oligomeric structures of a specific HALS, HPLC [10–12], mass spectrometric techniques [3, 13–18], and CZE with MS detection (CZE-MS) [19]. HPLC analysis of HALS is affected by several drawbacks such as unfavorable adsorption of HALS to silica-based stationary phases and the subsequent need for amine additives in the eluent leading to a substantially reduced column lifetime [10–12]. MS techniques applied for HALS analysis include ESI-MS [13], flow injection-MS [14, 15], desorption ESI-MS [3], liquid extraction surface electrospray-MS [16, 17], and paint spray-MS [18]. Although these techniques can be seen as useful tools for fast screening of plastic samples with respect to the presence of HALS and their degradation products, they do not allow an in-depth investigation particularly of complex high molecular mass HALS and their oligomer composition. Although capillary electroseparation techniques can be regarded as promising complement to HPLC for the separation of polymer stabilizers, only very few papers dealing with this topic exist [19, 20]. In a recent study we presented a CZE-MS method for the characterization of seven monomeric and oligomeric HALS, enabling us to identify a series of new structures in addition to those already displayed in the respective data sheets [19]. The present work is primarily dedicated to the development of a CZE method allowing the quantitative determination of HALS in polymer products. Thereby a special focus is set on the quantitative aspects of potential changes in the oligomer distribution of a specific HALS before and after processing.
2 Materials and methods 2.1 Materials, reagents, and standard solutions Formic acid (98–100%) was provided by Merck (Darmstadt, Germany). TFA (⬎99%) and ammonium formate (⬎97%) were supplied by Sigma-Aldrich (Steinheim, Germany), ACN, acetone, 2-propanol, and toluene, all analytical reagent grade, by VWR (Vienna, Austria). Methanol was obtained from JT Baker (Deventer, The Netherlands). For the internal mass calibration of the employed QTOFMS instrument, a 2.5 mM solution of hexakis(1H,1H,3H C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Electrophoresis 2014, 35, 2965–2971
tetrafluoropropoxy) phosphazine in 90:10 ACN:water provided by Agilent Technologies (Waldbronn, Germany) was employed. Purified water from a Milli-Q system from Millipore (Bedford, MA, USA) was used throughout the study. HALS Cyasorb 3529 (CAS-number 193098-40-7) was purchased from Cytec (West Paterson, NJ, USA) as a technical substance without any specification of purity. Internal standard 2,4,6-Tris(dimethylamino)-1,3,5-triazine (in the following text referred to as hexamethyl-melamine, HMM) was synthesized according to procedures of Pearlman and Banks [21] and Borkovec and DeMilo [22]. Cyasorb 3529 and HMM were dissolved in 0.4 M formic acid containing 10% v/v ACN giving a 2000 mg/L stock solution of the HALS and a 1000 mg/L stock solution of the internal standard. Working standards were prepared daily prior to use in concentrations between 10 and 100 mg/L for Cyasorb 3529 by a dilution with 0.4 M formic acid containing 10% v/v ACN. The internal standard was diluted with acetone resulting in a concentration of 100 mg/L and added to the Cyasorb 3529 standard for the quantitation of the oligomer composition, obtaining a final concentration of 5 mg/L. For the preparation of a lab sample of polypropylene stabilized with Cyasorb 3529, a MiniLab II Haake Rheomex CTW5 lab-scale compounder from Thermo Scientific (Karlsruhe, Germany) was employed. 5 g unstabilized polypropylene and 25 mg Cyasorb 3529 were filled into the extruder and compounded at a temperature of 190°C, a cycle time of 3 min and an extruder screw speed of 60 rpm. 2.2 Sample preparation For the determination of Cyasorb 3529 in a polypropylene sample, about 20 mg of the polymer were dissolved in 750 L toluene in an oven at 130°C (dissolution time 60 min). Subsequently, 750 L of acetone were added to reprecipitate the polymer, whereby the HALS remains in solution. After centrifuging the sample, 300 L of supernatant were brought to dryness and reconstituted in 300 L 0.4 M formic acid and ACN (20:80). 2.3 CE instrumentation and conditions CZE separations were carried out on an Agilent 3D CE system with DAD. Fused silica capillaries (50 m id × 360 m od) were from Polymicro Technologies (Phoenix, AZ, USA). For an optimization of the CZE separation and characterization of Cyasorb 3529 employing UV detection, new capillaries were cut to a length of 100 cm. In order to obtain both UV and MS signals within a single run, for correlation of the MS signals to the corresponding UV peaks, capillaries were cut to a length of 185 cm with a UV detection window at 100 cm. For MS quantitation, the employed capillaries were cut to a length of 70 cm. New capillaries were conditioned by flushing with 1 M NaOH for 20 min, followed by water and BGE for 10 min each. The same steps were applied as a daily routine prior to the first run. The capillary was further flushed with www.electrophoresis-journal.com
CE and CEC
Electrophoresis 2014, 35, 2965–2971
2967
Figure 1. Structures of Cyasorb 3529. The bold series were identified recently by Hintersteiner et al. [19], the elucidation of the other series was done within the present work.
BGE for 3 min before every run. The separation voltage was +30 kV for CZE separations with formic acid as BGE and +25 kV for CZE separations with TFA. Direct UV detection was performed at 240 nm. Two different BGEs were used throughout the study. A 1 M formic acid with 30% v/v ACN and 100 mM ammonium formate was employed for CZE separations with UV detection, for the structural elucidation of the additional oligomeric compounds using QTOF-MS detection, and for the assignment of the MS signals to the corresponding UV peaks. For the optimization of the MS parameters as well as for the subsequent MS quantitation, 20 mM TFA with 80% v/v ACN was utilized.
2.4 MS and MS/MS experiments MS and MS/MS experiments were performed in the positive ion mode using an Agilent 6510 QTOF mass spectrometer with an ESI source and an Agilent G1607A coaxial sprayer (all from Agilent). For CZE-MS, an optimization of the sheath liquid composition and the MS parameters (see Section 3.1.2) revealed that with a 1 mM ammonium formate solution with 80% v/v methanol and the following MS parameters high sensitivity and a stable spray can be achieved: drying gas flow (nitrogen) 7.6 L/min, drying gas temperature: 300°C, nebulizer pressure 7.3 psi, sheath liquid flow 4 L/min. The sheath liquid was delivered via an HPLC pump (Agilent 1100 Series G1311A) using a 1:100 splitter. To perform an on-line internal mass calibration, hexakis(1H,1H,3H-tetrafluoropropoxy) phosphazine (M + H+ = 922.0097) was added to the sheath liquid as reference mass. The scanning mass range was from m/z 110 to 3000 with an acquisition rate of 3 spectra/s. MS/MS measurements were performed with flow injection using an Agilent Technologies Series 1100 HPLC instrument coupled to the Agilent Technologies 6510 QTOF-MS. As mobile phase, a mixture of methanol and water (3:1) was used at a flow rate of 0.4 mL/min. For the fragmentation, a collision energy of 35 V was applied. C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
2.5 Software for the experimental design R
The Design-Expert software version 9 from Stat-Ease (Minneapolis, MN, USA) was employed to apply an experimental design, in particular a central composite design, to the optimization of the MS related parameters.
3 Results and discussion 3.1 Quantitation with CZE-QTOF-MS 3.1.1 Choice of BGE The main focus of this work was set on the development of a CZE method allowing the quantitative analysis of HALS in real polymer samples. Cyasorb 3529 was chosen as representative of this specific class of UV-stabilizers, whereby information on the basic structures of this HALS can be seen in Fig. 1. Starting point was a paper on the qualitative assessment of several different HALS by CZE-QTOF-MS published recently by our group [19]. At the beginning of the research reported in the present paper only the structures printed in bold letters in Fig. 1 were known. Employing the BGE system based on 1 M formic acid and 10% v/v ACN (the same as used previously for the qualitative analysis [19]), substantial variations of peak areas and signal intensities were observed, preventing us from obtaining useful quantitative data. A similar problem was also reported in the literature where during the analysis of melamine derivativesrepeatability problems were encountered when formic acid in high concentrations was used as BGE [23]. To avoid this problem, employing a stronger acid at lower concentrations was considered as an alternative. Since sulfuric acid and phosphoric acid are both not MS compatible, different concentrations of TFA were investigated for the separation of the oligomeric compounds and their quantitation with QTOF-MS. Too high concentrations of TFA or too low amounts of organic modifier resulted in an unstable CE current. Finally, with a system consisting www.electrophoresis-journal.com
2968
I. Hintersteiner et al.
of 20 mM TFA and 80% v/v ACN, a stable and reliable system was obtained. A drawback of this BGE was that resolution of some peaks was inferior to that achieved with the formic acid BGE and that some compounds were comigrating. This problem could be mitigated using a QTOF-MS for detection and the separation was still sufficient to avoid possible deterring suppression effects caused by coelution of analytes and the polymer matrix.
3.1.2 Optimization of ESI conditions The ESI-MS related parameters were optimized focusing on two principal requirements: a highly sensitive detection of the oligomeric compounds of Cyasorb 3529 and concurrently stable electrospray conditions to obtain reliable quantification results. Starting with the sheath liquid composition, different amounts of organic solvent (2-propanol and methanol), varying concentrations of formic acid as well as different amounts of salt were tested. A 1 mM ammonium formate in 80% v/v methanol turned out to be an optimum composition for the analysis of Cyasorb 3529, so it was chosen as starting point for further investigations. Raising the capillary voltage from 3000 to 4000 V showed an increase in sensitivity, but a decrease of signal stability at the same time, so 3750 V was selected. Concerning the fragmentor voltage, voltages from 150 to 250 V were investigated, obtaining a compromise solution for small and large oligomers at 200 V. For optimization of the other MS parameters an experimental design, more accurately a central composite design model was utilized to account also for possible interdependencies between the parameters of interest. Nebulizer pressure, drying gas flow rate and temperature, and sheath liquid flow rate were selected as design factors (see Table S1 in Supporting Information for optimization parameters). The peak areas for seven oligomeric structures were investigated to find a highly sensitive system, but also the electrospray stability was observed in order to ensure reliable conditions. The results revealed that for all oligomers the sheath liquid flow rate impacts the peak areas significantly, yielding higher values at lower flow rates. A further even though less important influencing factor is the drying gas temperature whereby at elevated temperatures higher peak areas were obtained. Increasing the nebulizer pressure appeared to have a slightly positive impact on the peak areas of some oligomers. The last investigated factor, the drying gas flow rate, did not show a significant impact on the peak areas of the oligomers. Within the design factors, interdependency between the nebulizer pressure and the sheath liquid flow rate was found, showing that higher nebulizer pressure values are required for increased sheath liquid flow rates. Besides these results, the measurements revealed that the electrospray becomes unstable when too extreme values such as very low sheath liquid flow rates (2 L/min), very high drying gas temperatures (350°C) or nebulizer gas pressures (12 psi) are applied. For this reason, an in-range optimization was used to obtain a sensitive and reliable set of MS parameter values. Optimum parameters C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Electrophoresis 2014, 35, 2965–2971
were as follows: nebulizer pressure of 7.3 psi, a drying gas temperature of 300°C, a drying gas flow rate of 7.6 L/min, and a sheath liquid flow rate of 4 L/min.
3.1.3 Measurement of calibration curves The optimized MS-related parameters were employed to measure calibration curves for the oligomers of Cyasorb 3529. Calibration was performed with standards of Cyasorb 3529 at concentrations of 10, 20, 40, 60, 80, and 100 mg/L, whereby each standard was analyzed three times. The concentration range was selected according to the concentrations expected for real polymer sample extracts. Calibration-related parameters such as linear range, the correlation coefficient, and repeatability data are shown in Table S2 (see Supporting Information). Repeatability was measured by injecting a 100 mg/L standard of Cyasorb 3529 five times, the RSDs for peak areas of 21 investigated oligomeric compounds were between 3.0 and 7.2%.
3.2 Analysis of a polymer sample A polypropylene sample containing an unknown amount of Cyasorb 3529 was investigated following a similar sample preparation procedure (see Section 2) as in already published work [12]. Unexpectedly, the ratios of the investigated oligomers were different from those observed for the standard. For the polypropylene sample the highest value for the peak area was detected for oligomer D (n = 1) followed by substantially smaller peaks for C (n = 1) and D (n = 2), whereas for the standard the largest peak area was obtained for C (n = 1) followed by D (n = 1) and C (n = 2). There are two possible explanations for this observation: either the batch of Cyasorb 3529 used for the investigated material was completely different compared to the standard available for our work, or the oligomeric composition changes during processing of the material. These findings led to the conclusion that a straightforward quantitative analysis of real polymer samples using calibration curves based on one Cyasorb 3529 standard is not possible. To solve this issue, an in-depth study of the data recorded for the Cyasorb 3529 standard as well as for the polypropylene sample was performed. Furthermore, a polypropylene sample was prepared in-house containing exactly the same Cyasorb 3529 batch as used for the calibration curves to detect probable structural changes in the distribution of oligomers.
3.3 Structural identification of additional oligomeric components in Cyasorb 3529 In our previously published work we reported six main series of oligomers in Cyasorb 3529 (see bold entries in Fig. 1) [19]. With an in-depth study of the data obtained from measurements of the Cyasorb 3529 standard as well as from the www.electrophoresis-journal.com
Electrophoresis 2014, 35, 2965–2971
polypropylene sample extract, additional structures could be detected as can be seen from Fig. 1. For series D, structures with a higher number of repetitive units (n = 3 and n = 4) could be identified. Furthermore, from the mass spectra of the polypropylene sample extract, additional peaks detected at m/z 14.0157 and 28.0314 less than the already known oligomers were observed. With the QTOF-MS, MS/MS fragmentation experiments were performed to investigate the structural changes leading to these additional components. The oligomer with the highest peak area observed from measurements of the standards, belonging to series C (n = 1) with a calculated exact mass of 685.5731 and the respective compound with m/z 14.0157 less from series C1 with an exact mass of 671.5574 were selected as precursors. From the resulting MS/MS spectrum obtained for the first precursor ion, fragments with m/z of 154.1598 and 140.1432 were detected. The corresponding structures can be assigned to a single charged 1,2,2,6,6-pentamethyl-1,2,3,6-tetrahydropyridine (PMP) and a single charged 2,2,6,6-tetramethyl-1,2,3,6tetrahydropyridine (TMP). As obvious from a comparison of the MS/MS spectra of both precursor ions (see Figs. 1a and 1b in Supporting Information), the ratios of these two fragments are different, whereby the TMP is higher in the case of the precursor with m/z 672.5. So, for the precursor molecule of the C series, both piperidinyl-rings are fivefold methylated and the TMP is only a fragmentation product of the PMP and consequently not present in the precursor, whereas in the case of the C1 series there is also a fourfold methylated piperidinyl-ring causing the higher TMP/PMP ratio in comparison with series C. Finally new structures with one (series B1, C1, D1, and E1), respectively, two methyl groups (series B2, C2, D2, and E2) replaced by hydrogen compared to the oligomeric structures identified in the previous work could be assigned.
3.4 Structural changes caused by polymer processing As described in Section 3.2, different oligomeric ratios were experimental for the standard and the sample containing Cyasorb 3529. To find out whether this could be caused by processing of the polymer, a polypropylene sample containing a defined amount of Cyasorb 3529 (identical batch as used for the calibration curves) was prepared (see “Materials and methods”). Figure 2 illustrates the results of the evaluation of 19 oligomers of the sample extract, whereby the zero line represents the amount of the oligomers in a Cyasorb 3529 standard with identical concentration as expected in the sample extract. For almost all oligomers, a lower amount was detected in the processed polymer than estimated from the amount of stabilizer added before processing. Additionally it was observed that series B, C, and D suffer from a higher degradation than the corresponding series B1, C1, and D1 (with m/z 14.0157 less). A similar behavior can be seen for series B2, C2, and D2 (see Fig. S2 in Supporting Informa C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
CE and CEC
2969
Figure 2. Concentrations of different oligomers in a polypropylene sample after processing, arranged in order of their migration times. The zero line displays the targeted area for each oligomer assuming a full recovery of the added standard. For the determination of these values, three sample extracts were prepared and quantified with CZE-QTOF-MS.
tion). Consequently, during processing the relative amount of the oligomeric series is changing in comparison to the standard, showing a relative increase of series B1, C1, and D1 and B2, C2, and D2 in comparison to B, C, and D. There are even compounds that exhibit higher peak areas than expected, such as C1 (n = 1) or D1 (n = 1), leading to the assumption that the structures with only four methyl groups could be degradation products from the structures initially containing a fivefold methylated piperidine ring. The question whether the difference in the oligomeric pattern for real polymer samples with unknown amounts of Cyasorb 3529 and the Cyasorb 3529 standard (see Section 3.2) can be attributed either to the usage of different batches of this HALS for standardization and for manufacturing of the polymer sample, or can be attributed to changes during polymer processing cannot be answered unambiguously. A major obstacle in quantitative analysis lies in the fact that neither standards for the single oligomers are available nor (under normal conditions) the batch actually employed for production of the sample is on hand for the analyst.
3.5 Quantitation of the composition of a Cyasorb 3529 standard using CZE-QTOF-MS and CZE-UV A practicable approach for solving the quantitation issue is by determining the exact composition of a Cyasorb 3529 standard, so the actual amount of every oligomer identified so far. In the case of Cyasorb 3529 the triazine ring is responsible for its UV absorption. Based on this fact a strategy for the absolute quantitation of the oligomers in a Cyasorb 3529 standard using CZE with UV-detection at 240 nm was developed. First, oligomers have to be characterized by CZE-QTOFMS thereby determining the ratio between UV-absorbing and nonabsorbing moieties within each oligomer. As UV absorption in Cyasorb 3529 is based on a melamine-like structure, www.electrophoresis-journal.com
2970
I. Hintersteiner et al.
Electrophoresis 2014, 35, 2965–2971
Figure 3. Base peak electropherogram of the Cyasorb 3529 standard. MS detection after 185 cm, separation voltage +30 kV, BGE 1 M formic acid, 30% ACN, 100 mM NH4 COOH.
HMM was used as internal standard and UV-absorbance of the HALS was related to this component. Knowing the peak area of the oligomer (AOlig ) and the internal standard (AISTD ) as well as the concentration of the internal standard (cISTD ) and the number of UV-absorbing units x in both, the respective oligomeric structure (xOlig ) and the internal standard (xISTD ) (xISTD = 1 for HMM) a determination of the concentration of the corresponding compound (cOlig ) in mol/L can be accomplished. c Olig =
AOlig · xISTD · c ISTD xOlig · AISTD
(1)
Knowing also the molecular mass (MOlig ) of the oligomer and the concentration of the Cyasorb 3529 solution prepared in gram per liter (wStd ) permits to calculate the amount (m%Olig ) of the respective oligomer in the Cyasorb 3529 standard according to m%Olig
MOlig · cOlig = · 100 wStd
(2)
An indispensable precondition for the success of such an approach is a good separation of the oligomers allowing the accurate determination of the peak areas. Furthermore, it is necessary to have two comparable electropherograms, one with QTOF-MS detection and one with UV-detection, allowing the unambiguous assignment of as many as possible oligomer structures to the corresponding peaks in the UVelectropherogram. To achieve this, the following approach was used: first experiments with dual-detection (simultaneous UV and MS detection) were performed. These experiments allowed the assignment of structures to the major peaks in the CE-UV electropherogram. Primarily due to suction effects (caused by the sheath gas of the ESI source) resolution of minor signals was insufficient to allow a proper assignment of these smaller peaks in the UV-trace. For this reason additional experiments with only UV-detection were performed. Using the previously assigned signals as a clear guide, the minor peaks could then be readily assigned to molecular structures from the CE-MS experiments. The results from these experiments can be seen in Figs. 3 and C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Figure 4. UV electropherogram of the Cyasorb 3529 standard. UV detection after 100 cm, separation voltage +30 kV, BGE 1 M formic acid, 30% ACN, 100 mM NH4 COOH.
4. Peaks series A and A1 are not visible in the UV electropherogram as these compounds do not include structural elements showing UV absorption. For the peaks that could not be clearly separated, an average molecular mass per triazine of 597.62 g/mol was used for the calculation, as this value is quite similar for the oligomers in this part of the electropherogram. The values used for the calculations and the results obtained for the Cyasorb 3529 standard can be found in Table 1. Adding up the calculated weight fractions of the investigated oligomers, results in 90.40% of the total amount
Table 1. Characterization of the Cyasorb 3529 standard a)
Oligomer
M
xOlig
B B1
1035.70 1021.67
1 1
C (n = 4) C1 (n = 4) C (n = 3) C1 (n = 3) C (n = 2) C1 (n = 2) D (n = 4) D1 (n = 4) D (n = 3) D1 (n = 3) C (n = 1) C1 (n = 1) D (n = 2) D1 (n = 2) E D (n = 1) D1 (n = 1) Sum
2440.74 2426.71 1855.84 1841.82 1270.94 1256.92 2675.99 2661.97 2091.09 2077.07 686.05 672.02 1506.19 1492.17 584.90 921.29 907.27
-
d)
4 4 1 1 3 3 1 2 2
b)
cOlig
m%Olig
RSD%
3.745 × 10−6 2.710 × 10−6
1.77 1.26
0.87 5.42
1.644 × 10−4
44.84
2.34
3.689 × 10−6 2.502 × 10−6 3.008 × 10−5 5.257 × 10−6 1.026 × 10−5 4.894 × 10−6
3.52 2.37 9.42 1.61 7.05 3.33
3.22 4.90 2.16 2.73 3.45 4.43
2.854 × 10−5 7.762 × 10−6
12.00 3.21 90.40
2.34 4.54
c)
a) Molecular mass of the oligomer. b) x = number of absorbing units. c) RSD calculated for three replicates. d) Average molecular mass of per 597.62 triazine used for calculation.
www.electrophoresis-journal.com
Electrophoresis 2014, 35, 2965–2971
of Cyasorb 3529. This value is reasonable, as there are structures that do not absorb UV light (A and A1), and furthermore, some other signals detected in the CZE-UV run could not be clearly assigned and peaks for some minor components such as those from the B2, C2, D2, and E2 series were not evaluated, as these structures exhibit only very small peaks in CZE-MS and CZE-UV measurements. With this basic characterization of a Cyasorb 3529 standard a novel approach for analyzing real samples prepared from an unknown batch of Cyasorb 3529 is available. It might be used for product characterization or quality control purposes whenever evaluation of the actual amount of this HALS present in the final polymer product is needed.
4 Concluding remarks In this work we present a strategy for the quantitation of a widely used HALS, Cyasorb 3529, in polymers. Thereby, in addition to the ten previously known oligomeric structures present in this technical product, 22 additional oligomeric structures, some of them potential degradation products, could be identified. The additionally elucidated series exhibit one or two methyl groups less than the oligomer series they are originating from, whereby MS/MS experiments revealed that the methyl is missing at the piperidine ring. Furthermore we were able to detect changes in the oligomer distribution of Cyasorb 3529 occurring during a polymer processing (in particular compounding) step. Combining data from CZEUV and CZE-QTOF-MS runs together with the selection of an appropriate internal standard (HMM) allowed the quantitative analysis of HALS on the basis of quantitation of the individual oligomers. This research work was performed in WP-01 “Performance and Test Methods” of the cooperative research project SolPol3 on novel polymeric encapsulation materials for PV modules (www.solpol.at). This project is funded by the Austrian Climate and Energy Fund (KLI:EN) within the program “Neue Energien 2020.” The authors have declared no conflict of interest.
5 References [1] Zweifel, H., Maier, R. D., Schiller, M., Palstics Additives ¨ Handbook, Hanser, Munchen 2008. [2] Jenke, D. R., Encyclopedia of Chromatography, 3rd edition, CRC Press, Boca Raton 2010.
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
CE and CEC
2971
[3] Paine, M. R. L., Gryn’ova, G., Coote, M. L., Barker, P. J., Blanksby, S. J., Polym. Degrad. Stab. 2014, 99, 223–232. [4] Bart, J. C. J., Polym. Degrad. Stab. 2003, 82, 197–205. [5] Klampfl, C. W., Trends Anal. Chem. 2013, 50, 53–64. [6] Paine, M. R. L., Barker, P. J., Blanksby, S. J., Anal. Chim. Acta. 2014, 808, 70–82. [7] Di Pasquale, G., Giambelli, L., Soffientini, A., Paiella, R., HRC & CC J. High. Resolut. Chromatogr. Chromatogr. Commun. 1985, 8, 618–622. [8] Coulier, L., Kaal, E., Tienstra, M., Hankemeier, T., J. Chromatogr. A. 2005, 1062, 227–238. [9] Jansson, K. D., Zawodny, C. P., Wampler, T. P., J. Anal. Appl. Pyrolysis 2007, 79, 353–361. ´ ˜ ´ [10] Noguerol-Cal, R., Lopez-Vilari no, J. M., Fernandez´ Mart´ınez, G., Gonzalez-Rodr´ ıguez, M. V., Barral-Losada, L. F., J. Sep. Science 2010, 33, 2698–2706. [11] Andersen, T., Skuland, I. L., Holm, A., Trones, R., Greibrokk, T., J. Chromatogr. A. 2004, 1029, 49–56. [12] Reisinger, M., Beissmann, S., Buchberger, W., Anal. Chim. Acta. 2013, 803, 181–187. [13] Lowe, T. A., Paine, M. R. L., Marshall, D. L., Hick, L. A., Boge, J. A., Barker, P. J., Blanksby, S. J., J. Mass. Spectrom. 2010, 45, 486–495. [14] Beißmann, S., Reisinger, M., Toelgyesi, L., Klampfl, C. W., Buchberger, W., Anal. Bioanal. Chem. 2013, 405, 6879–6884. [15] Beißmann, S., Reisinger, M., Reimann, A., Klampfl, C. W., Buchberger, W., Rapid Commun. Mass Spectrom. 2014, 28, 939–947. [16] Paine, M. R. L., Barker, P. J., Blanksby, S. J., Anal. Chim. Acta. 2014, 808, 190–198. [17] Paine, M. R. L., Barker, P. J., MacLauglin, S. A., Mitchell, T. W., Blanksby, S. J., Rapid Comm. Mass Spectrom. 2012, 26, 412–418. [18] Paine, M. R. L., Barker, P. J., Blanksby, S. J., Mass Spectrom. Lett. 2012, 3, 25–28. [19] Hintersteiner, I., Himmelsbach, M., Klampfl, C. W., Buchberger, W., Electrophoresis 2014, 35, 1368–1374. [20] Hilder, E. F., Klampfl, C. W., Buchberger, W., Haddad, P. R., J. Chromatogr. A. 2001, 922, 293–302. [21] Pearlman, W. M., Banks, C. K., J. Am. Chem. Soc. 1948, 70, 3726–3728. [22] Borkovec, A. B., DeMilo, A. B., J. Med. Chem. 1967, 10, 457–461. [23] Himmelsbach, M., Schmid, T., Buchberger, W., Schwarzinger, C., List, M., Electrophoresis 2010, 7, 1194–1200.
www.electrophoresis-journal.com
Electrophoresis 2014, 35, 2965–2971
Ingrid Hintersteiner Thomas Schmid Markus Himmelsbach Christian W. Klampfl Wolfgang W. Buchberger Institute of Analytical Chemistry, Johannes Kepler University Linz, Linz, Austria
Received May 27, 2014 Revised July 1, 2014 Accepted July 2, 2014
Research Article
Quantitative analysis of hindered amine light stabilizers by CZE with UV detection and quadrupole TOF mass spectrometric detection The current work describes the development of a CZE method with quadrupole QTOFMS detection and UV detection for the quantitation of Cyasorb 3529, a common hindered amine light stabilizer (HALS), in polymer materials. Analysis of real polymer samples revealed that the oligomer composition of Cyasorb 3529 changes during processing, a fact hampering the development of a straightforward method for quantitation based on calibration with a Cyasorb 3529 standard. To overcome this obstacle in-depth investigations of the oligomer composition of this HALS using QTOF-MS and QTOF-MS/MS had to be performed whereby 22 new oligomer structures, in addition to the ten structures already described, were identified. Finally, a CZE method for quantitative analysis of this HALS was developed starting with a comprehensive characterization of a Cyasorb 3529 standard using CZE-QTOF-MS, subsequently allowing the correct assignment of most Cyasorb 3529 oligomers in an electropherogram with UV detection. Employing the latter detection technique and hexamethyl-melamine as internal standard, peak areas obtained for the melamine could be correlated with those from the triazine ring, the UV-absorbing unit present in the HALS. This approach finally allowed proper quantitation of the single oligomers of Cyasorb 3529, an imperative for the quantitative assessment of this HALS in real polymer samples. Keywords: CZE-MS / Hindered amine light stabilizers / Polymer product analysis / QTOF-MS / UV-stabilization DOI 10.1002/elps.201400265
Additional supporting information may be found in the online version of this article at the publisher’s web-site
1 Introduction As plastic materials find their way into more and more areas of application ranging from rather simple “single use” devices to highly sophisticated components integrated in high performance products, design of such materials is confronted with steadily increasing requirements. These advanced specifications are not only met by development of new base materials,
Correspondence: Ingrid Hintersteiner, Institute of Analytical Chemistry, Johannes Kepler University, Linz, Altenberger Strasse 69, A-4040 Linz, Austria E-mail: [email protected] Fax: +43-732-2468-8679
Abbreviations: HALS, hindered amine light stabilizers; HMM, hexamethyl-melamine; PMP, 1,2,2,6,6-pentamethyl1,2,3,6-tetrahydropyridine; TMP, 2,2,6,6-tetramethyl-1,2,3,6tetrahydropyridine
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
i.e. different types of polymers and polymer blends, but also by use of novel types of polymer additives. Additives are highly important constituents of plastics, playing a fundamental role in achieving materials with unique physical and chemical properties. Within the group of additives, stabilizers can be understood as a prominent subgroup, protecting the polymer against thermal, oxidative and UV-light induced stress [1, 2]. Focusing on the latter type of stabilizers, compounds such as 2-hydroxybenzophenones or 2-hydroxyphenylbenzotriazoles, and hindered amine light stabilizers (HALS), which stand for the latest development in this field, are commonly employed when it comes to UV-protection of plastic materials. Typically, HALS include 2,2,6,6-tetramethylpiperidine as structural element being responsible for the photostabilizing effect, as the secondary amine can act as alkyl radical trapping species by forming a stable nitroxide radical [3]. For better understanding of product performance, determination of stabilizers is of substantial concern at all stages of polymer processing such as production of a stabilized base
www.electrophoresis-journal.com
2966
I. Hintersteiner et al.
polymer, compounding, and further processing steps such as extrusion or injection molding [4–6]. The necessity for qualitative and quantitative analysis of polymer stabilizers but also their degradation products has led to the development of a series of analytical methods. Focusing on the analysis of HALS it has to be taken into account that these stabilizers are technical products with often only limited or no information available concerning purity or actual structural composition of the respective product. Particularly for oligomeric HALS, which are not amenable to GC analysis, only a very small number of reports on analytical methods suitable for the characterization and quantitation of these components can be found in the literature. These comprise pyrolysis GC-MS, an approach enabling identification of several HALS additives [7–9] but unable to differentiate between different oligomeric structures of a specific HALS, HPLC [10–12], mass spectrometric techniques [3, 13–18], and CZE with MS detection (CZE-MS) [19]. HPLC analysis of HALS is affected by several drawbacks such as unfavorable adsorption of HALS to silica-based stationary phases and the subsequent need for amine additives in the eluent leading to a substantially reduced column lifetime [10–12]. MS techniques applied for HALS analysis include ESI-MS [13], flow injection-MS [14, 15], desorption ESI-MS [3], liquid extraction surface electrospray-MS [16, 17], and paint spray-MS [18]. Although these techniques can be seen as useful tools for fast screening of plastic samples with respect to the presence of HALS and their degradation products, they do not allow an in-depth investigation particularly of complex high molecular mass HALS and their oligomer composition. Although capillary electroseparation techniques can be regarded as promising complement to HPLC for the separation of polymer stabilizers, only very few papers dealing with this topic exist [19, 20]. In a recent study we presented a CZE-MS method for the characterization of seven monomeric and oligomeric HALS, enabling us to identify a series of new structures in addition to those already displayed in the respective data sheets [19]. The present work is primarily dedicated to the development of a CZE method allowing the quantitative determination of HALS in polymer products. Thereby a special focus is set on the quantitative aspects of potential changes in the oligomer distribution of a specific HALS before and after processing.
2 Materials and methods 2.1 Materials, reagents, and standard solutions Formic acid (98–100%) was provided by Merck (Darmstadt, Germany). TFA (⬎99%) and ammonium formate (⬎97%) were supplied by Sigma-Aldrich (Steinheim, Germany), ACN, acetone, 2-propanol, and toluene, all analytical reagent grade, by VWR (Vienna, Austria). Methanol was obtained from JT Baker (Deventer, The Netherlands). For the internal mass calibration of the employed QTOFMS instrument, a 2.5 mM solution of hexakis(1H,1H,3H C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Electrophoresis 2014, 35, 2965–2971
tetrafluoropropoxy) phosphazine in 90:10 ACN:water provided by Agilent Technologies (Waldbronn, Germany) was employed. Purified water from a Milli-Q system from Millipore (Bedford, MA, USA) was used throughout the study. HALS Cyasorb 3529 (CAS-number 193098-40-7) was purchased from Cytec (West Paterson, NJ, USA) as a technical substance without any specification of purity. Internal standard 2,4,6-Tris(dimethylamino)-1,3,5-triazine (in the following text referred to as hexamethyl-melamine, HMM) was synthesized according to procedures of Pearlman and Banks [21] and Borkovec and DeMilo [22]. Cyasorb 3529 and HMM were dissolved in 0.4 M formic acid containing 10% v/v ACN giving a 2000 mg/L stock solution of the HALS and a 1000 mg/L stock solution of the internal standard. Working standards were prepared daily prior to use in concentrations between 10 and 100 mg/L for Cyasorb 3529 by a dilution with 0.4 M formic acid containing 10% v/v ACN. The internal standard was diluted with acetone resulting in a concentration of 100 mg/L and added to the Cyasorb 3529 standard for the quantitation of the oligomer composition, obtaining a final concentration of 5 mg/L. For the preparation of a lab sample of polypropylene stabilized with Cyasorb 3529, a MiniLab II Haake Rheomex CTW5 lab-scale compounder from Thermo Scientific (Karlsruhe, Germany) was employed. 5 g unstabilized polypropylene and 25 mg Cyasorb 3529 were filled into the extruder and compounded at a temperature of 190°C, a cycle time of 3 min and an extruder screw speed of 60 rpm. 2.2 Sample preparation For the determination of Cyasorb 3529 in a polypropylene sample, about 20 mg of the polymer were dissolved in 750 L toluene in an oven at 130°C (dissolution time 60 min). Subsequently, 750 L of acetone were added to reprecipitate the polymer, whereby the HALS remains in solution. After centrifuging the sample, 300 L of supernatant were brought to dryness and reconstituted in 300 L 0.4 M formic acid and ACN (20:80). 2.3 CE instrumentation and conditions CZE separations were carried out on an Agilent 3D CE system with DAD. Fused silica capillaries (50 m id × 360 m od) were from Polymicro Technologies (Phoenix, AZ, USA). For an optimization of the CZE separation and characterization of Cyasorb 3529 employing UV detection, new capillaries were cut to a length of 100 cm. In order to obtain both UV and MS signals within a single run, for correlation of the MS signals to the corresponding UV peaks, capillaries were cut to a length of 185 cm with a UV detection window at 100 cm. For MS quantitation, the employed capillaries were cut to a length of 70 cm. New capillaries were conditioned by flushing with 1 M NaOH for 20 min, followed by water and BGE for 10 min each. The same steps were applied as a daily routine prior to the first run. The capillary was further flushed with www.electrophoresis-journal.com
CE and CEC
Electrophoresis 2014, 35, 2965–2971
2967
Figure 1. Structures of Cyasorb 3529. The bold series were identified recently by Hintersteiner et al. [19], the elucidation of the other series was done within the present work.
BGE for 3 min before every run. The separation voltage was +30 kV for CZE separations with formic acid as BGE and +25 kV for CZE separations with TFA. Direct UV detection was performed at 240 nm. Two different BGEs were used throughout the study. A 1 M formic acid with 30% v/v ACN and 100 mM ammonium formate was employed for CZE separations with UV detection, for the structural elucidation of the additional oligomeric compounds using QTOF-MS detection, and for the assignment of the MS signals to the corresponding UV peaks. For the optimization of the MS parameters as well as for the subsequent MS quantitation, 20 mM TFA with 80% v/v ACN was utilized.
2.4 MS and MS/MS experiments MS and MS/MS experiments were performed in the positive ion mode using an Agilent 6510 QTOF mass spectrometer with an ESI source and an Agilent G1607A coaxial sprayer (all from Agilent). For CZE-MS, an optimization of the sheath liquid composition and the MS parameters (see Section 3.1.2) revealed that with a 1 mM ammonium formate solution with 80% v/v methanol and the following MS parameters high sensitivity and a stable spray can be achieved: drying gas flow (nitrogen) 7.6 L/min, drying gas temperature: 300°C, nebulizer pressure 7.3 psi, sheath liquid flow 4 L/min. The sheath liquid was delivered via an HPLC pump (Agilent 1100 Series G1311A) using a 1:100 splitter. To perform an on-line internal mass calibration, hexakis(1H,1H,3H-tetrafluoropropoxy) phosphazine (M + H+ = 922.0097) was added to the sheath liquid as reference mass. The scanning mass range was from m/z 110 to 3000 with an acquisition rate of 3 spectra/s. MS/MS measurements were performed with flow injection using an Agilent Technologies Series 1100 HPLC instrument coupled to the Agilent Technologies 6510 QTOF-MS. As mobile phase, a mixture of methanol and water (3:1) was used at a flow rate of 0.4 mL/min. For the fragmentation, a collision energy of 35 V was applied. C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
2.5 Software for the experimental design R
The Design-Expert software version 9 from Stat-Ease (Minneapolis, MN, USA) was employed to apply an experimental design, in particular a central composite design, to the optimization of the MS related parameters.
3 Results and discussion 3.1 Quantitation with CZE-QTOF-MS 3.1.1 Choice of BGE The main focus of this work was set on the development of a CZE method allowing the quantitative analysis of HALS in real polymer samples. Cyasorb 3529 was chosen as representative of this specific class of UV-stabilizers, whereby information on the basic structures of this HALS can be seen in Fig. 1. Starting point was a paper on the qualitative assessment of several different HALS by CZE-QTOF-MS published recently by our group [19]. At the beginning of the research reported in the present paper only the structures printed in bold letters in Fig. 1 were known. Employing the BGE system based on 1 M formic acid and 10% v/v ACN (the same as used previously for the qualitative analysis [19]), substantial variations of peak areas and signal intensities were observed, preventing us from obtaining useful quantitative data. A similar problem was also reported in the literature where during the analysis of melamine derivativesrepeatability problems were encountered when formic acid in high concentrations was used as BGE [23]. To avoid this problem, employing a stronger acid at lower concentrations was considered as an alternative. Since sulfuric acid and phosphoric acid are both not MS compatible, different concentrations of TFA were investigated for the separation of the oligomeric compounds and their quantitation with QTOF-MS. Too high concentrations of TFA or too low amounts of organic modifier resulted in an unstable CE current. Finally, with a system consisting www.electrophoresis-journal.com
2968
I. Hintersteiner et al.
of 20 mM TFA and 80% v/v ACN, a stable and reliable system was obtained. A drawback of this BGE was that resolution of some peaks was inferior to that achieved with the formic acid BGE and that some compounds were comigrating. This problem could be mitigated using a QTOF-MS for detection and the separation was still sufficient to avoid possible deterring suppression effects caused by coelution of analytes and the polymer matrix.
3.1.2 Optimization of ESI conditions The ESI-MS related parameters were optimized focusing on two principal requirements: a highly sensitive detection of the oligomeric compounds of Cyasorb 3529 and concurrently stable electrospray conditions to obtain reliable quantification results. Starting with the sheath liquid composition, different amounts of organic solvent (2-propanol and methanol), varying concentrations of formic acid as well as different amounts of salt were tested. A 1 mM ammonium formate in 80% v/v methanol turned out to be an optimum composition for the analysis of Cyasorb 3529, so it was chosen as starting point for further investigations. Raising the capillary voltage from 3000 to 4000 V showed an increase in sensitivity, but a decrease of signal stability at the same time, so 3750 V was selected. Concerning the fragmentor voltage, voltages from 150 to 250 V were investigated, obtaining a compromise solution for small and large oligomers at 200 V. For optimization of the other MS parameters an experimental design, more accurately a central composite design model was utilized to account also for possible interdependencies between the parameters of interest. Nebulizer pressure, drying gas flow rate and temperature, and sheath liquid flow rate were selected as design factors (see Table S1 in Supporting Information for optimization parameters). The peak areas for seven oligomeric structures were investigated to find a highly sensitive system, but also the electrospray stability was observed in order to ensure reliable conditions. The results revealed that for all oligomers the sheath liquid flow rate impacts the peak areas significantly, yielding higher values at lower flow rates. A further even though less important influencing factor is the drying gas temperature whereby at elevated temperatures higher peak areas were obtained. Increasing the nebulizer pressure appeared to have a slightly positive impact on the peak areas of some oligomers. The last investigated factor, the drying gas flow rate, did not show a significant impact on the peak areas of the oligomers. Within the design factors, interdependency between the nebulizer pressure and the sheath liquid flow rate was found, showing that higher nebulizer pressure values are required for increased sheath liquid flow rates. Besides these results, the measurements revealed that the electrospray becomes unstable when too extreme values such as very low sheath liquid flow rates (2 L/min), very high drying gas temperatures (350°C) or nebulizer gas pressures (12 psi) are applied. For this reason, an in-range optimization was used to obtain a sensitive and reliable set of MS parameter values. Optimum parameters C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Electrophoresis 2014, 35, 2965–2971
were as follows: nebulizer pressure of 7.3 psi, a drying gas temperature of 300°C, a drying gas flow rate of 7.6 L/min, and a sheath liquid flow rate of 4 L/min.
3.1.3 Measurement of calibration curves The optimized MS-related parameters were employed to measure calibration curves for the oligomers of Cyasorb 3529. Calibration was performed with standards of Cyasorb 3529 at concentrations of 10, 20, 40, 60, 80, and 100 mg/L, whereby each standard was analyzed three times. The concentration range was selected according to the concentrations expected for real polymer sample extracts. Calibration-related parameters such as linear range, the correlation coefficient, and repeatability data are shown in Table S2 (see Supporting Information). Repeatability was measured by injecting a 100 mg/L standard of Cyasorb 3529 five times, the RSDs for peak areas of 21 investigated oligomeric compounds were between 3.0 and 7.2%.
3.2 Analysis of a polymer sample A polypropylene sample containing an unknown amount of Cyasorb 3529 was investigated following a similar sample preparation procedure (see Section 2) as in already published work [12]. Unexpectedly, the ratios of the investigated oligomers were different from those observed for the standard. For the polypropylene sample the highest value for the peak area was detected for oligomer D (n = 1) followed by substantially smaller peaks for C (n = 1) and D (n = 2), whereas for the standard the largest peak area was obtained for C (n = 1) followed by D (n = 1) and C (n = 2). There are two possible explanations for this observation: either the batch of Cyasorb 3529 used for the investigated material was completely different compared to the standard available for our work, or the oligomeric composition changes during processing of the material. These findings led to the conclusion that a straightforward quantitative analysis of real polymer samples using calibration curves based on one Cyasorb 3529 standard is not possible. To solve this issue, an in-depth study of the data recorded for the Cyasorb 3529 standard as well as for the polypropylene sample was performed. Furthermore, a polypropylene sample was prepared in-house containing exactly the same Cyasorb 3529 batch as used for the calibration curves to detect probable structural changes in the distribution of oligomers.
3.3 Structural identification of additional oligomeric components in Cyasorb 3529 In our previously published work we reported six main series of oligomers in Cyasorb 3529 (see bold entries in Fig. 1) [19]. With an in-depth study of the data obtained from measurements of the Cyasorb 3529 standard as well as from the www.electrophoresis-journal.com
Electrophoresis 2014, 35, 2965–2971
polypropylene sample extract, additional structures could be detected as can be seen from Fig. 1. For series D, structures with a higher number of repetitive units (n = 3 and n = 4) could be identified. Furthermore, from the mass spectra of the polypropylene sample extract, additional peaks detected at m/z 14.0157 and 28.0314 less than the already known oligomers were observed. With the QTOF-MS, MS/MS fragmentation experiments were performed to investigate the structural changes leading to these additional components. The oligomer with the highest peak area observed from measurements of the standards, belonging to series C (n = 1) with a calculated exact mass of 685.5731 and the respective compound with m/z 14.0157 less from series C1 with an exact mass of 671.5574 were selected as precursors. From the resulting MS/MS spectrum obtained for the first precursor ion, fragments with m/z of 154.1598 and 140.1432 were detected. The corresponding structures can be assigned to a single charged 1,2,2,6,6-pentamethyl-1,2,3,6-tetrahydropyridine (PMP) and a single charged 2,2,6,6-tetramethyl-1,2,3,6tetrahydropyridine (TMP). As obvious from a comparison of the MS/MS spectra of both precursor ions (see Figs. 1a and 1b in Supporting Information), the ratios of these two fragments are different, whereby the TMP is higher in the case of the precursor with m/z 672.5. So, for the precursor molecule of the C series, both piperidinyl-rings are fivefold methylated and the TMP is only a fragmentation product of the PMP and consequently not present in the precursor, whereas in the case of the C1 series there is also a fourfold methylated piperidinyl-ring causing the higher TMP/PMP ratio in comparison with series C. Finally new structures with one (series B1, C1, D1, and E1), respectively, two methyl groups (series B2, C2, D2, and E2) replaced by hydrogen compared to the oligomeric structures identified in the previous work could be assigned.
3.4 Structural changes caused by polymer processing As described in Section 3.2, different oligomeric ratios were experimental for the standard and the sample containing Cyasorb 3529. To find out whether this could be caused by processing of the polymer, a polypropylene sample containing a defined amount of Cyasorb 3529 (identical batch as used for the calibration curves) was prepared (see “Materials and methods”). Figure 2 illustrates the results of the evaluation of 19 oligomers of the sample extract, whereby the zero line represents the amount of the oligomers in a Cyasorb 3529 standard with identical concentration as expected in the sample extract. For almost all oligomers, a lower amount was detected in the processed polymer than estimated from the amount of stabilizer added before processing. Additionally it was observed that series B, C, and D suffer from a higher degradation than the corresponding series B1, C1, and D1 (with m/z 14.0157 less). A similar behavior can be seen for series B2, C2, and D2 (see Fig. S2 in Supporting Informa C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
CE and CEC
2969
Figure 2. Concentrations of different oligomers in a polypropylene sample after processing, arranged in order of their migration times. The zero line displays the targeted area for each oligomer assuming a full recovery of the added standard. For the determination of these values, three sample extracts were prepared and quantified with CZE-QTOF-MS.
tion). Consequently, during processing the relative amount of the oligomeric series is changing in comparison to the standard, showing a relative increase of series B1, C1, and D1 and B2, C2, and D2 in comparison to B, C, and D. There are even compounds that exhibit higher peak areas than expected, such as C1 (n = 1) or D1 (n = 1), leading to the assumption that the structures with only four methyl groups could be degradation products from the structures initially containing a fivefold methylated piperidine ring. The question whether the difference in the oligomeric pattern for real polymer samples with unknown amounts of Cyasorb 3529 and the Cyasorb 3529 standard (see Section 3.2) can be attributed either to the usage of different batches of this HALS for standardization and for manufacturing of the polymer sample, or can be attributed to changes during polymer processing cannot be answered unambiguously. A major obstacle in quantitative analysis lies in the fact that neither standards for the single oligomers are available nor (under normal conditions) the batch actually employed for production of the sample is on hand for the analyst.
3.5 Quantitation of the composition of a Cyasorb 3529 standard using CZE-QTOF-MS and CZE-UV A practicable approach for solving the quantitation issue is by determining the exact composition of a Cyasorb 3529 standard, so the actual amount of every oligomer identified so far. In the case of Cyasorb 3529 the triazine ring is responsible for its UV absorption. Based on this fact a strategy for the absolute quantitation of the oligomers in a Cyasorb 3529 standard using CZE with UV-detection at 240 nm was developed. First, oligomers have to be characterized by CZE-QTOFMS thereby determining the ratio between UV-absorbing and nonabsorbing moieties within each oligomer. As UV absorption in Cyasorb 3529 is based on a melamine-like structure, www.electrophoresis-journal.com
2970
I. Hintersteiner et al.
Electrophoresis 2014, 35, 2965–2971
Figure 3. Base peak electropherogram of the Cyasorb 3529 standard. MS detection after 185 cm, separation voltage +30 kV, BGE 1 M formic acid, 30% ACN, 100 mM NH4 COOH.
HMM was used as internal standard and UV-absorbance of the HALS was related to this component. Knowing the peak area of the oligomer (AOlig ) and the internal standard (AISTD ) as well as the concentration of the internal standard (cISTD ) and the number of UV-absorbing units x in both, the respective oligomeric structure (xOlig ) and the internal standard (xISTD ) (xISTD = 1 for HMM) a determination of the concentration of the corresponding compound (cOlig ) in mol/L can be accomplished. c Olig =
AOlig · xISTD · c ISTD xOlig · AISTD
(1)
Knowing also the molecular mass (MOlig ) of the oligomer and the concentration of the Cyasorb 3529 solution prepared in gram per liter (wStd ) permits to calculate the amount (m%Olig ) of the respective oligomer in the Cyasorb 3529 standard according to m%Olig
MOlig · cOlig = · 100 wStd
(2)
An indispensable precondition for the success of such an approach is a good separation of the oligomers allowing the accurate determination of the peak areas. Furthermore, it is necessary to have two comparable electropherograms, one with QTOF-MS detection and one with UV-detection, allowing the unambiguous assignment of as many as possible oligomer structures to the corresponding peaks in the UVelectropherogram. To achieve this, the following approach was used: first experiments with dual-detection (simultaneous UV and MS detection) were performed. These experiments allowed the assignment of structures to the major peaks in the CE-UV electropherogram. Primarily due to suction effects (caused by the sheath gas of the ESI source) resolution of minor signals was insufficient to allow a proper assignment of these smaller peaks in the UV-trace. For this reason additional experiments with only UV-detection were performed. Using the previously assigned signals as a clear guide, the minor peaks could then be readily assigned to molecular structures from the CE-MS experiments. The results from these experiments can be seen in Figs. 3 and C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Figure 4. UV electropherogram of the Cyasorb 3529 standard. UV detection after 100 cm, separation voltage +30 kV, BGE 1 M formic acid, 30% ACN, 100 mM NH4 COOH.
4. Peaks series A and A1 are not visible in the UV electropherogram as these compounds do not include structural elements showing UV absorption. For the peaks that could not be clearly separated, an average molecular mass per triazine of 597.62 g/mol was used for the calculation, as this value is quite similar for the oligomers in this part of the electropherogram. The values used for the calculations and the results obtained for the Cyasorb 3529 standard can be found in Table 1. Adding up the calculated weight fractions of the investigated oligomers, results in 90.40% of the total amount
Table 1. Characterization of the Cyasorb 3529 standard a)
Oligomer
M
xOlig
B B1
1035.70 1021.67
1 1
C (n = 4) C1 (n = 4) C (n = 3) C1 (n = 3) C (n = 2) C1 (n = 2) D (n = 4) D1 (n = 4) D (n = 3) D1 (n = 3) C (n = 1) C1 (n = 1) D (n = 2) D1 (n = 2) E D (n = 1) D1 (n = 1) Sum
2440.74 2426.71 1855.84 1841.82 1270.94 1256.92 2675.99 2661.97 2091.09 2077.07 686.05 672.02 1506.19 1492.17 584.90 921.29 907.27
-
d)
4 4 1 1 3 3 1 2 2
b)
cOlig
m%Olig
RSD%
3.745 × 10−6 2.710 × 10−6
1.77 1.26
0.87 5.42
1.644 × 10−4
44.84
2.34
3.689 × 10−6 2.502 × 10−6 3.008 × 10−5 5.257 × 10−6 1.026 × 10−5 4.894 × 10−6
3.52 2.37 9.42 1.61 7.05 3.33
3.22 4.90 2.16 2.73 3.45 4.43
2.854 × 10−5 7.762 × 10−6
12.00 3.21 90.40
2.34 4.54
c)
a) Molecular mass of the oligomer. b) x = number of absorbing units. c) RSD calculated for three replicates. d) Average molecular mass of per 597.62 triazine used for calculation.
www.electrophoresis-journal.com
Electrophoresis 2014, 35, 2965–2971
of Cyasorb 3529. This value is reasonable, as there are structures that do not absorb UV light (A and A1), and furthermore, some other signals detected in the CZE-UV run could not be clearly assigned and peaks for some minor components such as those from the B2, C2, D2, and E2 series were not evaluated, as these structures exhibit only very small peaks in CZE-MS and CZE-UV measurements. With this basic characterization of a Cyasorb 3529 standard a novel approach for analyzing real samples prepared from an unknown batch of Cyasorb 3529 is available. It might be used for product characterization or quality control purposes whenever evaluation of the actual amount of this HALS present in the final polymer product is needed.
4 Concluding remarks In this work we present a strategy for the quantitation of a widely used HALS, Cyasorb 3529, in polymers. Thereby, in addition to the ten previously known oligomeric structures present in this technical product, 22 additional oligomeric structures, some of them potential degradation products, could be identified. The additionally elucidated series exhibit one or two methyl groups less than the oligomer series they are originating from, whereby MS/MS experiments revealed that the methyl is missing at the piperidine ring. Furthermore we were able to detect changes in the oligomer distribution of Cyasorb 3529 occurring during a polymer processing (in particular compounding) step. Combining data from CZEUV and CZE-QTOF-MS runs together with the selection of an appropriate internal standard (HMM) allowed the quantitative analysis of HALS on the basis of quantitation of the individual oligomers. This research work was performed in WP-01 “Performance and Test Methods” of the cooperative research project SolPol3 on novel polymeric encapsulation materials for PV modules (www.solpol.at). This project is funded by the Austrian Climate and Energy Fund (KLI:EN) within the program “Neue Energien 2020.” The authors have declared no conflict of interest.
5 References [1] Zweifel, H., Maier, R. D., Schiller, M., Palstics Additives ¨ Handbook, Hanser, Munchen 2008. [2] Jenke, D. R., Encyclopedia of Chromatography, 3rd edition, CRC Press, Boca Raton 2010.
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
CE and CEC
2971
[3] Paine, M. R. L., Gryn’ova, G., Coote, M. L., Barker, P. J., Blanksby, S. J., Polym. Degrad. Stab. 2014, 99, 223–232. [4] Bart, J. C. J., Polym. Degrad. Stab. 2003, 82, 197–205. [5] Klampfl, C. W., Trends Anal. Chem. 2013, 50, 53–64. [6] Paine, M. R. L., Barker, P. J., Blanksby, S. J., Anal. Chim. Acta. 2014, 808, 70–82. [7] Di Pasquale, G., Giambelli, L., Soffientini, A., Paiella, R., HRC & CC J. High. Resolut. Chromatogr. Chromatogr. Commun. 1985, 8, 618–622. [8] Coulier, L., Kaal, E., Tienstra, M., Hankemeier, T., J. Chromatogr. A. 2005, 1062, 227–238. [9] Jansson, K. D., Zawodny, C. P., Wampler, T. P., J. Anal. Appl. Pyrolysis 2007, 79, 353–361. ´ ˜ ´ [10] Noguerol-Cal, R., Lopez-Vilari no, J. M., Fernandez´ Mart´ınez, G., Gonzalez-Rodr´ ıguez, M. V., Barral-Losada, L. F., J. Sep. Science 2010, 33, 2698–2706. [11] Andersen, T., Skuland, I. L., Holm, A., Trones, R., Greibrokk, T., J. Chromatogr. A. 2004, 1029, 49–56. [12] Reisinger, M., Beissmann, S., Buchberger, W., Anal. Chim. Acta. 2013, 803, 181–187. [13] Lowe, T. A., Paine, M. R. L., Marshall, D. L., Hick, L. A., Boge, J. A., Barker, P. J., Blanksby, S. J., J. Mass. Spectrom. 2010, 45, 486–495. [14] Beißmann, S., Reisinger, M., Toelgyesi, L., Klampfl, C. W., Buchberger, W., Anal. Bioanal. Chem. 2013, 405, 6879–6884. [15] Beißmann, S., Reisinger, M., Reimann, A., Klampfl, C. W., Buchberger, W., Rapid Commun. Mass Spectrom. 2014, 28, 939–947. [16] Paine, M. R. L., Barker, P. J., Blanksby, S. J., Anal. Chim. Acta. 2014, 808, 190–198. [17] Paine, M. R. L., Barker, P. J., MacLauglin, S. A., Mitchell, T. W., Blanksby, S. J., Rapid Comm. Mass Spectrom. 2012, 26, 412–418. [18] Paine, M. R. L., Barker, P. J., Blanksby, S. J., Mass Spectrom. Lett. 2012, 3, 25–28. [19] Hintersteiner, I., Himmelsbach, M., Klampfl, C. W., Buchberger, W., Electrophoresis 2014, 35, 1368–1374. [20] Hilder, E. F., Klampfl, C. W., Buchberger, W., Haddad, P. R., J. Chromatogr. A. 2001, 922, 293–302. [21] Pearlman, W. M., Banks, C. K., J. Am. Chem. Soc. 1948, 70, 3726–3728. [22] Borkovec, A. B., DeMilo, A. B., J. Med. Chem. 1967, 10, 457–461. [23] Himmelsbach, M., Schmid, T., Buchberger, W., Schwarzinger, C., List, M., Electrophoresis 2010, 7, 1194–1200.
www.electrophoresis-journal.com
