• No results found

Analytical Method Development of Fluorinated Silanes using Mass Spectrometry

N/A
N/A
Protected

Academic year: 2021

Share "Analytical Method Development of Fluorinated Silanes using Mass Spectrometry"

Copied!
32
0
0

Loading.... (view fulltext now)

Full text

(1)

Analytical Method Development of Fluorinated Silanes

using Mass Spectrometry

Author: Tea Eklundh Odler

Supervisor: Ulrika Eriksson

Examinor: Anna Kärrman

Date: 2018.05.30

Project in Chemistry 15hp

School of Science and Technology

Örebro University, Sweden

(2)

Abstract

The aim of this study was to develop an analytical method for fluorinated silanes. Furthermore, as a secondary aim, to explore if there would be possible to detect

1H,1H,2H,2H-perfluorooctyl triethoxysilane (6:2 PTrEtSi) and 1H,1H,2H,2H-perfluorodecyl triethoxysilane (8:2 PTrEtSi) in two different matrices, sludge and cosmetic extract. The method development included experiments using LC-MS, LC-MS/MS, UPC2, GC-MS and APGC-MS/MS and was carried out using standards containing 6:2 PTrEtSi and 8:2 PTrEtSi. The analytical method that worked best for the compounds was GC-MS/MS and an analytical method using APGC-MS/MS was developed for fluorinated silanes. The IDL for 6:2 PTrEtSi was 0.0012 µg/mL and 1.32 µg/mL for 8:2 PTrEtSi. This makes the developed method suitable for high contaminated samples, such as extracts from cosmetic products. It was concluded that a method using LC as the analytical instrument would not work for the two target compounds since they were too reactive with the mobile phase. However, LC could be a good choice for siloxanes, compounds that are formed from hydrolysis and condensation of fluorinated silanes. The samples analyzed in this study were three sludge extracts and one extract from a cosmetic product. 6:2 PTrEtSi was expected to be detected in the cosmetic sample since the compound was stated on the table of contents of the cosmetic product. No detection of 6:2 TrEtSi or 8:2 TrEtSi could be made in either of the samples. The reason for this was suspected to be transformation or degradation of the compounds into other

compounds. Therefore, a full scan of the cosmetic sample using LC-MS/MS was included in the experiment as an addition to verify the suspicions that compounds such as siloxanes could have been formed. An interesting peak was discovered with m/z 947 which could be a

disiloxane of 6:2 PTrEtSi.

Keywords: PFAS, Fluorinated silanes, Fluorinated siloxanes, 1H,1H,2H,2H-Perfluorooctyl

triethoxysilane, 1H,1H,2H,2H-Perfluorodeyl triethoxysilane, FTOH, PFCA, APGC-MS/MS, LC-MS/MS

(3)

TABLE OF CONTENTS

1.INTRODUCTION ... 1

1.1BACKGROUND ... 1

1.2FLUORINATED SILANES AND SILOXANES ... 2

1.2.1HEALTH AND ENVIRONMENTAL CONCERNS OF FLUORINATED SILANES ... 4

1.3INSTRUMENTS ... 5

1.4AIM ... 6

2.MATERIALS AND METHODS... 6

2.1CHEMICALS ... 6

2.3METHOD DEVELOPMENT... 7

3.RESULTS ... 7

3.1 METHOD DEVELOPMENT – LC-MS ... 7

3.2METHOD DEVELOPMENT –UPC2 ... 8

3.3METHOD DEVELOPMENT –GC-MS ... 8 3.4APGC-MS/MS ... 9 3.5ANALYSIS OF SAMPLES ... 11 4.DISCUSSION ... 12 4.1LC/MSMETHOD DEVELOPMENT ... 12 4.2GC-MSMETHOD DEVELOPMENT ... 13 4.2.1LOW-RESOLUTION GC-MS ... 13 4.2.2APGC-MS/MS ... 13 4.3SAMPLE ANALYSIS ... 16 5.CONCLUSION ... 16 6.REFERENCES ... 18 7.APPENDIX ... 20

(4)

1. Introduction

1.1 Background

Per- and polyfluoroalkyl substances (PFAS) consists of a family with many different classes and homologues, more than 3000 anthropogenic chemicals (KEMI, 2015. Some common PFAS are shown in table 1. PFAS have been used in industrial production since the 1950s and have contributed to pollution of the environment ever since (Kissa, 2001). They have been used by industries, as surfactants, as well as in household products such as cleaning products and paint. About 20% of the compounds have functions as surfactants and has a broad application field (KEMI, 2015). One third of the compounds are used in chemical synthesis, electronics, cosmetics, different types of impregnation, pharmaceuticals and many more. One common way PFAS is used is in aqueous-film-forming foams (AFFF) (Moody, et.al. 2003). PFAS properties help with reducing the surface tension of AFFFs and is therefore suitable for use in fire-fighting foam, thus, creating a problem with pollution around fire-fighting training facilities. Studies have been made from airport sites with this type of activity in America, Sweden, Norway, Canada, The Netherlands and Germany and levels of PFAS could be reported in the groundwater, surfacewater, the biota and in the soil around the facilities (Hale et.al. 2017).

Table 1. Examples of common PFAS, their chemical structures, names and abbreviations (Kjølholt, J. et.al. 2015).

Perfluorooctane sulfonic acid (PFOS as acid)

Perfluorooctanoic acid (PFOA)

8:2 Fluorotelomer alcohol (8:2 FTOH)

6:2 Fluorotelomer phosphate/mono[2- (perfluorohexyl)ethyl] phosphate

Although the properties of the compounds can be useful, they can also be harmful. Many PFAS are highly persistent and stay in the environment for a long time, others can degrade or transform into persistent compounds (KEMI, 2015). Some of the compounds are toxic and bio-accumulative which means that they accumulate in living organisms and can cause harm. Due to the many applications of PFAS, it is unavoidable that they end up in wastewater treatment plants (Clara et.al. 2008). The conventional wastewater treatment has been found to not be sufficient to remove PFAS from the water. This contributes to pollution of sewage

(5)

sludge which is one of the biggest emission sources into the environment. One reason why the knowledge of the occurrence of PFAS in Sweden and EU is limited can be due to the fact that many PFAS are potent and the desired effect can be achieved even with very low

concentrations (KEMI, 2015). If the amounts used by manufacturers or suppliers are below 100 tons/year, it is not required to report any information about the usage. It is important to learn more about the compounds as well as establish the usage of PFAS to prevent further health problems and environmental pollution.

The reason PFAS are so commonly used is because of their oil and water repellency properties (KEMI, 2015). The typical structure consists of a hydrophilic, active head and a lipophilic fluorinated alkyl chain (Kissa, 2001). The compounds have one or more of the hydrogen atoms have been replaced with fluorine atoms. The bond between a carbon and fluorine is known to be one of the strongest chemical bond. The fluorine atoms also protect the carbons from being chemical attack since the small size of the fluorine contribute to a “shielding effect” without steric stress, which contributes to their persistence. PFAS can be divided into groups and a large group is the different types of PFAS which are polymers, these can have fluorinated carbons incorporated in the chain or have fluorinated alkyl-groups attached to the polymer (KEMI, 2015). Another large group is the non-polymer which can be divided into two classes; perfluoroalkyl substances and polyfluoroalkyl substances (Buck et.al 2011). Perfluoroalkyl substances consists of compounds where all hydrogens on the carbon-chain have been substituted by fluorine atoms. Polyfluoroalkyl substances are compounds where all hydrogen atoms are exchanged to fluorine atoms on at least one but not all carbons in the molecule.

1.2 Fluorinated Silanes and Siloxanes

The general structure of a fluorinated alkyl silane are like other PFAS, in which they have a polyfluorinated carbon chain. They have a similar structure as FTOHs, which is shown in table 1, however, fluorinated alkyl silanes consists of a silicon head group with alkyl-groups of varying chain lengths attached to silicon through ester linkage. The structures of 6:2

PTrEtSi and 8:2 PTrEtSi are shown in table 2. Siloxanes are molecules that are polymeric and is composed of a silicon bound to oxygen, which can form cyclic or linear backbone

structures (Lee et.al. 2018). The carbon chain can have hydrogen atoms partially or fully substituted by fluorine. The base atom for the organosilanes is silicon which is a metalloid (Nørgaard et. al 2010). Silicon has unique abilities since it can both form partly ionic bonds as well as stable covalent bonds with carbon. This result in bridges between inorganic structures with organic functionality.

(6)

Table 2. The chemical structures, names, CAS-number and synonyms that occur for the compounds of the fluorinated silanes used for the method development (Sigma-Aldrich). The abbreviations 6:2 PTrEtSi and 8:2 PTrEtSi will be used in this study.

Compound Name

CAS-number Synonyms 1H,1H,2H,2H-Perfluorooctyl triethoxysilane 51851-37-7 6:2 polyfluoroalkyl triethoxysilane, 6:2 PTrEtSi, Triethoxy(1H,1H,2H,2H-perfluoro-1-octyl)silane, Triethoxy(3,3,4,4,5,5,6,6,7,7,8,8,8-tridecafluoro-1-octyl)silane 1H,1H,2H,2H-Perfluorodeyl triethoxysilane 101947-16-4 8:2 polyfluoroalkyl triethoxysilane, 8:2 PTrEtSi Triethoxy-1H,1H,2H,2H-perfluorodecylsilane, Triethoxy(3,3,4,4,5,5,6,6,7,7,8,8,8-tridecafluoro-1-decyl)silane The manufacture volume of 6:2 PTrEtSi in EU is between 10-100 tons/year. The further use of about 10% of the manufactured amount is unknown (ECHA, 2017). The import is

estimated to be small in comparison to the manufactured volume within EU. Fluorinated organosilicon derivatives are of interest due to their broad application in the production of modern materials (Dopierala et. al. 2013). They are used as surfactants, for surface

modification of lenses and optical fibers, components of many cosmetic preparations and in the production of oil-, dirt- and water-repellant surfaces.

When fluorinated silanes are used in example film forming spray products, the compounds need to be “activated” in order to enable the formation of a self-assembling film during evaporation of the solvent, which induces the desired properties to the surface that is treated (Nørgaard et.al. 2010). This is done through controlled hydrolysis and condensation reactions shown in figure 1. These reactions will result in the formation of a network of hydrolyzed silanes and functionalized siloxanes and makes it likely that other compounds such as siloxanes can be detected in matrices where silanes are present.

(7)

Figure 1. Hydrolysis (1), homofunctional condensation (2) and heterofunctional condensation (3) of an organosilane (Nørgaard et.al. 2010).

1.2.1 Health and Environmental Concerns of fluorinated silanes

The toxic effects of fluorinated silanes have not been particularly studied yet. However, in a study done by Nørgaard et. al (2010) they investigated different commercial spray products containing fluorinated silanes and organic solvents (ECHA, 2017). They found that

polyfluorooctyl trialkoxysilane most likely were the parent compound in the products. The risk of exposure is not directly related to the compounds listed in the table of contents on the products, but to the substances that forms from the parent compounds during hydrolysis in combination with organic solvents. To investigate the health effects, Nørgaard et. al (2010) let mice inhale the sprays which showed that they have toxic effects on the deep lung tissue, even with short term exposure.

Substances based on a fluorotelomer are suspected to be able to transform into perfluorinated carboxylic acids (PFCAs) (Fasano, et. al. 2006). They may degrade to form PFCAs as well as contain PFCAs as an unintended residual from the reaction. Since some fluorinated silanes and siloxanes have fluorotelomer structures, for example 6:2 PTrEtSi, it should not be excluded that these compounds can degrade into compounds such as PFCAs as well, with FTOH as a possible intermediate. The molecular structure of PFOA and FTOH can be seen in table 1. Amongst the PFCAs is perfluorooctanic acid (PFOA), an eight-carbon molecule, which is considered to have toxic effects to humans and environmental persistent properties (Yu et.al. 2018). The pathways from FTOH to PFOA are shown in figure 2. The

biotransformation of compounds that have similar structures to fluorinated silanes, such as PAPs, have been studied (D’eon and Mabury 2011). Furthermore, it has been observed that PAPs degrade to form FTOH through hydrolysis which in turn oxidizes and creates PFCAs.

(8)

Figure 2. The pathways 8:2 FTOH biotransformation under aerobic conditions (Buck. et.al 2011).

1.3 Instruments

The most common method when analyzing PFAS is LC coupled to MS/MS with ESI- as ionization method (Lorenzo, M. et. al. 2018). Sometimes MS/MS can show insufficient selectivity which makes other methods worth to consider. These are high-resolution MS (HRMS), time-of-flight (TOF) or Orbitrap-MS when doing trace analyses in complex matrices. In a study by Zacs and Bartkevics (2015) they evaluated different ionization techniques to determine which would provide best results for PFOS and PFOA. The ionization techniques were electrospray ionization (ESI), atmospheric pressure chemical ionization (APCI) and atmospheric pressure photoionization (APPI) (Zacs and Bartkevics 2016). Although ESI have a higher risk for matrix interference, they concluded that it would work best since it provided highest instrument sensitivity compared to the two other

techniques evaluated. The most common ionization technique to use is ESI since it provides high sensitivity. However, in some studies, APCI have been preferred as the ionization technique instead because it yields less fragmentation of the molecule (Lorenzo et. al. 2018). The most common column is a non-polar bonded silica (C18) and reversed phase is usually used. The mobile phases usually consist of water and methanol, with addition of ammonium acetate.To improve peak shape some of the most common modifiers is 1-methyl-piperidine for higher pH and formic acid for lower pH. They are added to the mobile phase to increase the chromatographic resolution and detection sensitivity for PFAS that are anionic.

In some cases, GC is selected instead of LC and the choice depends on the properties of the analytes (Lorenzo et. al. 2018). Compounds that are thermally stable and volatile are suitable for GC analysis, examples of such compounds are organophosphorus flame retardants (PFRs) and novel brominated flame retardants (NBFRs). However, some PFAS can be tricky to quantify due to their ionic nature which can provide tailing and broad peaks. This can be

(9)

avoided if the compounds undergo derivatization. For ionization technique, electron

ionization (EI) is most popular, however, for the more labile compounds, a softer ionization technique is a better option. A new combination of GC and APCI have shown to work well for analysis of these compounds (Lorenzo et. al. 2018). When analyzing PFAS with GC, the most common mass analyzer is the single quadrupole but QqQ and TOF have recently gained popularity.

In a previous study, GC-MS with EI as ionization method has been used for detecting fluorinated silanes (Nørgaard et al. 2009). For fluorinated siloxanes LC-MS with ESI was used. The hydrolyzed compounds that were detected with LC could not be detected using GC in this study.

1.4 Aim

The objective for this study was to develop an analytical method for fluorinated silanes. The secondary objective was to use the developed method to analyze extracts from cosmetics and sludge, which are suspected to contain the target compounds.

2. Materials and methods

The materials used were washed with methanol before use. All standards and samples were prepared in a fume hood and were stored in a freezer.

2.1 Chemicals

The standards, Perfluorooctyltriethoxysilane (98%) and 1H,1H,2H,2H-Perfluorodeyltriethoxysilane (97%), formic acid (95%) and N-methylpiperidine (99%) came from Sigma-Aldrich. The methanol used was HPLC grade from Fisher Scientific (Ottawa, Canada). Milli-Q water (18.2 MΩ) and NH4Ac was used.

2.2 Method

The method developed for the target compounds 6:2 PTrEtSi and 8:2 PTrEtSi are shown in table 3 and was carried out on an APGC-MS/MS instrument.

Table 3. The developed GC analytical method for fluorinated silanes.

Column 30m*0,25mm*0,25µm DB-5MS column

(Agilent Technologies)

Temperature program 45°C for 1 min

10°C/min to 180°C

30°C/min to 260°C hold for 2.2 min

Injection volume 1 µL

Solvent Methanol

Solvent delay 2 min

Transitions MRM method (m/z) for 6:2 PTrEtSi

511 195 (Qualifier), 511 219,

(10)

511 239 (Quantifier), 511309 (Qualifier)

Transitions MRM method (m/z) for 8:2 PTrEtSi

611 295 (Qualifier), 611339 (Qualifier), 611 389,

611409 (Quantifier)

Ionization method APCI

Carrier gas Nitrogen

Gas flow 1,5ml/min

Temperature at ion source 200°C

Temperature transfer line 200°C

Temperature MS quad 200°C

2.3 Method Development

The method development was done using a stock solution containing two compounds. The compounds used were 6:2 PTrEtSi with a molecular weight of 510 and 8:2 PTrEtSi with a molecular weight of 610, their molecular structure can be seen in table 2.

3. Results

3.1 Method Development – LC-MS

The system used for the LC-MS/MS experiments was a Acquity Ultra Performance (Waters) coupled to a Quattro Premier XE (Waters). Initially, standard solutions of 6:2 PTrEtSi and 8:2 PTrEtSi in methanol was infused directly into the mass spectrometer. The initial LC method was based on a previous method used for PFAS and the original settings are shown in table 4. Firstly, MS experiments were performed to explore which conditions could obtain signals for the analytes. Secondly, MS/MS experiments were carried out to see if any fragments of interests could be detected. The concentration of the extracts analyzed were 2 ng/µL and the solvent was 100% methanol.

Table 4. The initial settings during the LC experiment.

Conditions Initial settings

Column ACQUITY UPLC BEH

C18 1.7 µm 2.1*50mm

Ionization mode ESI

-Capillary voltage (kV) 0.8 Cone voltage (V) 55 Desolvation temperature (C°) 410 Source temperature (C°) 100 Desolvation (L/h) 950 Cone (L/h) 50

(11)

With ESI- no signal was detected which lead to the change to positive mode. When increasing the capillary voltage to 1.8 and decrease the cone voltage to 25, some peaks could be

obtained, however, they were still quite unstable. If the capillary or cone voltage was too high or too low, it was difficult to detect any peaks at all. The peaks that could be seen when injecting 6:2 PTrEtSi had the m/z 534, 523 and 508 and when injecting 8:2 PTrEtSi the peaks with m/z 634, 623 and 608 could be detected.

To investigate if the signal could become more stable, infusions of standard solutions were combined with mobile phases using a valve prior to the ESI. Three different mobile phases were tested, one with neutral pH, one with lower and one with higher pH. Both ESI- and ESI+ were tested, however, no peaks could be seen during experiments using ESI-. The ions obtained from ESI+ can be seen in table 5. The flow of both the mobile phase and of the sample from the syringe were kept at 0.025 µL/min during this part of the experiment. Table 5. The m/z obtained while tuning with the three different mobile phases in ESI+ mode for both 6:2 PTrEtSi and 8:2 PTrEtSi.

100% 2mM NH4Ac in H2O (neutral pH) 0.1% formic acid in H2O (lower pH) 2mM NH4Ac + N-methylpiperidine (higher pH)

6:2 PTrEtSi 8:2 PTrEtSi 6:26 PTrEtSi 8:2 PTrEtSi 6:2 PTrEtSi 8:2 PTrEtSi

m/z 512 m/z 600 m/z 99 m/z 135 m/z 101 m/z 102

m/z 534* m/z 612* m/z 135 m/z 166 m/z 634 m/z 150

* Shows the peak with highest intensity.

Further experiments were made in positive mode using neutral pH with 100% 2mM NH4Ac in H2O as the mobile phase. An MRM method was tested using the developed method with m/z 534 and 612 as targets, but no peaks could be detected. An auto-tuning was carried out which only detected noise.

3.2 Method Development – UPC2

A SIR method was carried out with the system Acquity UPC2 (Waters) coupled to a XEVO-TQS (Waters), both ESI- and ESI+ were tested. The m/z scanned was 510, 511, 512, 610, 611 and 612. The column used was Torus DIOL 130Å 1.7µm 3.0mm*150mm. The critical gas used was CO2 and the mobile phase was 2mM NH4Ac in H2O.The capillary voltage was 3.5 kV and the cone voltage was 25 V. In negative mode, there was only noise detected and in positive mode the compounds eluted together at the beginning of the run and the peak shapes were tailed and broad. A reason for this can be that the compounds were not polar enough for the column that was used.

3.3 Method Development – GC-MS

The GC-MS experiments were carried out on a low-resolution GC-MS with a GC system from Hewlett Packard coupled to a Mass Selective Detector (Hewlett Packard). The initial settings were derived from two previous methods, one that Nørgaard et.al. (2009) used for

(12)

analyzing commercial spray products containing fluorinated silanes, and one method found in a paper written by Hjortsberg (2017) where organic fluorinated compounds were analyzed. Both methods are shown in table 6. To conclude what temperature program to develop further, the two different methods were tested using a low-resolution GC-MS.

Table 6. The GC-methods tested on the low-resolution GC-MS used by Nørgaard et.al (2009) and by Hjortsberg (2017). Method by Nørgaard et.al (2009) Method by Hjortsberg (2017) Column 60m*0,25mm*0,25µm

(Varian Chrompack Sil-19),

(30m*0,25mm*0,25µm DB-5MS column (Agilent Technologies) was used instead)

DB-5MS column (Agilent Technologies

30m*0,25mm*0,25µm

Temperature program 45°C for 1 min

10°C/min to 230°C hold 5 min 25°C/min to 300°C hold for 2,2 min 80°C hold 2 min 10°C/min to 205°C hold 5 min 15°C/min to 260°C hold 5 min Injection volume 1 µL 1 µL

Solvent Methanol Methanol

Solvent delay 2 min 2 min

Scan range 50-700 50-700

Ionization method EI EI

Carrier gas Helium Helium

Gas flow 1,5ml/min 1,5ml/min

Temperature at ion source 200°C 200°C

Temperature transfer line 200°C 200°C

Temperature MS quad 200°C 200°C

The method that was chosen to continue developing was the one used by Nørgaard et.al (2009). The column used was a 30m*0,25mm*0,25µm DB-5MS column (Agilent

Technologies). The temperature program of the method was changed, initial temperature was 45°C for 1 min then a ramp of 10°C/min to 180°C, lastly a ramp of 30°C/min to 260°C which was held for 2.2 min. When doing a fullscan the m/z observed for 6:2 PTrEtSi was 59, 149 and 163 and for 8:2 PTrEtSi the m/z detected was 119, 149 and 163.

3.4 APGC-MS/MS

To obtain a softer ionization technique for the experiments a APGC-MS/MS (Waters) system was used, coupled to a XEVO-TQS (Waters). A MRM scan was started using the same settings as before, shown in table 3, in splitless mode. This resulted in peaks that could be of

(13)

interests with the corresponding m/z to the mass of the compounds, however, the peak shapes were split. The 6:2 PTrEtSi eluted approximately around 9.7 min and 8:2 PTrEtSi at 11.05 min which was the reason the temperature program was changed to a total of 17 min. The inlet temperature was lowered to 260°C which did not change anything and it was changed back to the initial temperature (280°C). To continue the cone voltage was optimized, to do this, several scans was done with the cone voltages 5, 10, 20, 30, 40 and 50 V. To attempt to better the peak shapes the injection mode was switched from splitless to split mode, the results from this can be seen in figure 6-7.

The daughters of the precursor compounds were determined by doing a daughter scan with different collision energies. The energies tested were 8, 10, 15, 20, 15, 30, 40 and 50 V. The collision energy chosen for the method was the one where the fragments showed the highest intensity respectively.

A calibration curve was prepared with 9 points between 0.02ng/mL – 200ng/mL, this was to determine the IDL of the method. The calibration curves obtained from this is showed in the figures 3-4, they range from 2ng/mL-200ng/mL for 6:2 PTrEtSi and from 10 ng/mL-200 ng/mL for 8:2 PTrEtSi.

Figure 3. A 5-point calibration point for the m/z 239 that had the highest intensity of the fragments analyzed for 6:2 PTrEtSi, this was obtained using an MRM-method.

(14)

Figure 4. A 4-point calibration point for the m/z 409 that had the highest intensity of the fragments analyzed for 8:2 PTrEtSi, this was obtained using an MRM-method.

The IDL was calculated using the lowest concentration that could be detected which was 2 ng/mL for 6:2 PTrEtSi and 10 ng/mL for 8:2 PTrEtSi the results can be seen in table 8.

3.5 Analysis of samples

Four samples were analyzed with the method developed from this study (table 3). Three sludge extracts which previously had been analyzed for PFASs and one extract that were from a cosmetic product. The cosmetic extract was left over from a study by Ashhami, A. (2017) where she analyzed the fluorine content in different cosmetic products. In the table of contents of the product that is in the extract it stated that 6:2 PTrEtSi is in the product. The three other extracts were sludge samples from Denmark, Sweden and Norway. The four extracts were diluted 10 times with methanol and injected on the APGC-MS/MS, neither 6:2 PTrEtSi or 8:2 PTrEtSi could be detected in the samples.

In the study by Nørgaard et. Al. 2009 siloxanes and other hydrolyzed silanes were detected using LC-MS/MS ESI. Since no fluorinated silanes could be detected in the cosmetic sample, a full scan was carried out to investigate if other fluorinated compounds could be detected such as siloxanes and hydrolyzed silanes. The analytical instrument used was LC-MS/MS ESI+ with the system of Acquity UPLC (Waters) coupled to a XEVO-TQS (waters) and the method used was one already developed for analyzing PFAS.

A detection of m/z 947 was made, when analyzing with LC-MS/MS, in the cosmetic sample which could be a disiloxane of 6:2 TrEtSi. The pattern showed a possible loss of m/z 44 for each of the ethoxy-groups of the molecule, shown in figure 5. A theoretical isotope model of the m/z 947 showed a similar pattern to the pattern detected in the sample. However, the sample contains too many interferences to confirm this theory. Other m/z was investigated but nothing of interest for this study was detected, which reinforces that this is a theory and future studies must be made to be able to draw conclusions.

(15)

Figure 5. The fragmentation pattern of the suspected disiloxane, which shows the loss of m/z 44 from m/z 947, and the theoretical isotope model of m/z 947.

4. Discussion

4.1 LC/MS Method Development

The reason this analytical method was tested was because it is the most common method to use while analyzing other PFAS, for example PFOS (Trojanowicz 2013). One advantage is that multiple compounds can be analyzed with the same method. However, when injecting the compounds 6:2 PTrEtSi and 8:2 PTrEtSi directly on the column no stable peaks could be detected, the settings are shown in table 3. During tuning by direct infusion of the analytes some peaks was detected in ESI+ mode, though they were unstable. For the compound 6:2 PTrEtSi the peaks detected had the m/z 534, 523 and 508, which indicated that some adducts occurs. The m/z 534 probably have an adduct of sodium, [M+Na]+, which also was detected and identified by Nørgaard et al. (2010). When analyzing the compound, 8:2 PTrEtSi, the corresponding m/z (634, 623 and 608) could be detected. This could indicated that the 6:2 PTrEtSi and 8:2 PTrEtSi was detected, however, containing adducts.

The experiment continued with an attempt to stabilize the compounds by adjusting the mobile phase to low and high pH. There were still no peaks detected in ESI- mode and the peaks detected in ESI+ mode, with both lower and higher pH, showed very low m/z, such as 91. This indicated strong fragmentation and unstable molecular ions. The conclusion from this was that neutral pH would work best for these particular compounds since the m/z detected with 2mM NH4Ac in H2O as mobile phase showed peaks at m/z 512 and 534 for 6:2 PTrEtSi and 600, 612 and 634 for 8:2 PTrEtSi. There was no or little fragmentation, however, there were still some adducts which could be explained by impurities in the standards.

MRM experiments were performed on the m/z of 534 and 612 since these had the highest intensity. Furthermore, auto-tuning of the same m/z was performed. The MRM-method and the auto-tuning results did only detect noise which could be explained by that the compounds

(16)

are too non-polar to be able to elute them with the C18 column used with the LC. In UPC2 the compounds eluted together at the beginning of the method which is an indication that the compounds are not polar enough for the method or the diol column. Therefore, it would be interesting to investigate if it would be possible to analyze them using normal phase instead of reversed phase. However, this was excluded from this study due to time restrictions.

4.2 GC-MS Method Development 4.2.1 Low-resolution GC-MS

After the two GC-methods were tested with low-resolution GC-MS, the method from Nørgaard et.al (2010) was used for continued development, this was since there were no peaks detected for 6:2 PTrEtSi with the method by Hjortsberg (2017). However, the peaks that could be detected with the low-resolution GC and EI as ionization method were highly fragmented with m/z of 59, 149 and 163 for the peak suspected to be 6:2 PTrEtSi and 119, 149 and 163 for the suspected 8:2 PTrEtSi. The m/z 163 could be the molecule with a loss of the fluorinated tail (C3H9SiO3) and the m/z 119 is suspected to be the end of the fluorinated tail (C2F5). It was concluded that a softer ionization method would be needed to be able to determine if the compounds detected were the molecular ions.

4.2.2 APGC-MS/MS

The method development was continued with an APGC-MS/MS to obtain a softer ionization technique. APCI ionization is preferred for compounds that are labile or when the molecular ion is needed as the technique does not fragment the compounds as hard as EI does (Lorenzo et. al. 2018). In splitless mode the peak shapes were split which could be due to that the instrument was overloaded with sample. Thus, split mode was tested, the results for both splitless mode and split mode are shown below in figures 6-7. It can be stated that the split mode did improve the peak shapes significantly.

(17)

Figure 6. The peak shapes of 8:2 PTrEtSi 6:2 PTrEtSi with split-less mode.

Figure 7. The peak shapes of 8:2 PTrEtSi and 6:2 PTrEtSi with split mode.

The optimization of the cone voltage showed that the highest intensity of the precursor ions was around 20 V. The results for 6:2 PTrEtSi can be seen in figure A4 in appendix.

The m/z detected from the daughter scan with different collision energies showed some interesting fragments which can be seen in table 7. They followed a typical fragmentation for fluorinated compounds with loss of CF3 and additions of CF2 in the pattern, which was confirmed by comparing the fragments detected by Nørgaard et al. (2010) which is shown in figure 9. The collision energies that are not shown in table 7 did not show any fragments of interest. When the collision energy was too low the precursor compound could not be fragmented enough. If the collision energy was too high the precursor compound was too fragmented instead. These are the reasons why no fragments of interest could be detected when the collision energy was below 10 V and over 25 V.

Table 7. The different collisions energies tested and the ion fragments that showed the highest intensity at the different collision energies. Both fragments of 6:2 PTrEtSi and 8:2 PTrEtSi are shown. Collision Energy (V) Fragments of 6:2 PTrEtSi Fragments of 8:2 PTrEtSi 10 195, 309 15 195, 239 409, 489 20 169, 195, 219, 239 295, 389, 409, 489 25 169, 219, 239 389

(18)

The fragment at m/z 309 is most likely the fluorinated tail (C8H4F11) which are shown in figure 8 (a). The figure shows the fragmentation of a perfluorinated disiloxane, however the fragmentation pattern is similar for 6:2 PTrEtSi. The figure is therefore used to explain the fragmentation of 6:2 PTrEtSi. This fragment corresponds to the fragment detected at m/z 409 for 8:2 PTrEtSi, which would be the same fragment with two extra fluorinated carbon atoms (C10H4F15). Other fragments that corresponds for the both compounds are m/z 195 for 6:2 PTrEtSi and 295 for 8:2 PTrEtSi. These fragments are probably further fragmentation of the fluorinated carbon chain where three more carbon atoms have been lost, one carbon atom with hydrogen and two fluorinated carbons. This will result in fragments with the structures C5H2F9 for 6:2 PTrEtSi and C7H2F13 for 8:2 PTrEtSi. The m/z 169, 219 and 239 are typical fragmentation of a fluorinated carbon chain, where m/z 169 is the fragment C3F7 and m/z 219 C4F9 and so on. This indicates that the compounds detected are the target compounds. Since the two target compounds does not differ in structure except for length of the fluorinated alkyl chain, they should have a fragmentation pattern similar to each other.

Figure 8. Fragmentation pathways of (a) perfluorinated disiloxane, modified from Nørgaard et.al (2010).

The R2-value of the calibration curves were 0.959 for 6:2 PTrEtSi and 0.944 for 8:2 PTrEtSi (figure 3-4). Good values are supposed to be above 0.98, however, values above 0.95 are considered to be acceptable. This concludes that the calibration curve for 6:2 PTrEtSi is usable. The R2-value for 8:2 PTrEtSi is below the acceptable value. Furthermore, the

(19)

calibration curve only has four point instead of five, which also makes it less reliable. One improvement of the calibration curves that was intended to be done to was to prepare a

narrower range, for example, 9 points between 0.02-10 ng/mL. This would probably make the R2-values better and thus make them more reliable. Another step to the study would have been to do a repeatability test and calculate the RSD which would have been a much-needed validation of the method. The reason that these two things was excluded from the study was due to time restriction of the study and the high pressure on the instruments in the laboratory where the study took place, time simply ran out.

Since there are no previous studies there is nothing to compare the IDL of the method to. However, the IDL of the develop APGC-MS/MS method was calculated and are shown in table 8. In matrices were the concentration is suspected to be quite high, such as cosmetic extracts, the developed method is applicable. The large difference in the IDL between 6:2 PTrEtSi and 8:2 PTrEtSi provides information about the two compounds properties and that they do not behave similarly to each other.

Table 8. The IDL of the APGC-MS/MS method for both compounds calculated from the lowest point in the corresponding calibration curves.

Compound IDL

6:2 PTrEtSi 0.0012 µg/mL

8:2 PTrEtSi 1.32 µg/mL

4.3 Sample Analysis

There was no detection of 6:2 PTrEtSi and 8:2 PTrEtSi in the four samples. This is probably due to that the compounds have degraded or transformed into other compounds. One way it could have happened is through hydrolysis or formation of a fluorinated siloxane. It should be possible to detect 6:2 PTrEtSi in the cosmetic sample, since it is stated on the table of contents that it contains the compound. However, since the cosmetic product also contains water, hydrolysis of the compound is possible or formation of siloxanes. The LC-MS/MS analysis of the cosmetic sample detected an interesting peak. The peak had the m/z 947 which

corresponds to a disiloxane of the 6:2 PTrEtSi. It showed a fragmentation pattern with a loss of 44 which could be the ethoxy-group (C2H5O) of the compound. The scan range was m/z 500-1000 and m/z 1000-1500 which did not make it possible to see if the compound would have the typical fragmentation of the fluorinated tail. Moreover, the molecular m/z of the possible hydrolyzed compounds fell outside the scan range.

5. Conclusion

The method developed on APGC-MS/MS are applicable for fluorinated silanes. However, if they are in matrices or used in applications where hydrolysis or condensation is possible, other analytes might need to be targeted as well. Even though the four samples analyzed in this study is not enough to be able to draw any conclusion regarding occurrence and

concentrations of fluorinated silanes in products and environmental samples, other methods can be better suitable, such as LC-MS/MS. Furthermore, using both methods when analyzing fluorinated silanes could be complementary to be able to detect the total amount in the

(20)

matrices. Future recommendations would be to perform a method development for fluorinated siloxanes. Since no standards for fluorinated siloxanes could be found it will be a challenge. A suggestion is to synthesize hydrolysates of fluorinated silanes with hydrolysis through addition of acid to a standard containing fluorinated silanes. Another interesting approach would be to perform a degradation experiment to confirm if the compounds could degrade to FTOHs and PFCAs. An investigation of the toxicity other than the lung damage already confirmed would also be interesting.

(21)

6. References

Ashhami, A. (2017) Assessment of Extractable Organic Fluorine (EOF) Content and Contribution of Per- and Polyfluoroalkyl Substances (PFASs) in Cosmetic Products Buck, R., et.al (2011) Perfluoroalkyl and Polyfluoroalkyl Substances in the Environment: Terminology, Classification, and Origins. Integrated Environmental Assessment and Management. 7 513–541


Clara, M., Scheffknecht, C., Scharf, S., Weiss, S., and Gans, O. (2008) Emissions of perfluorinated alkylated substances (PFAS) from point sources—identification of relevant branches. Water Science & Technology. 58.1: 59-66

D’eon, J. and Mabury, S. (2011) Exploring Indirect Sources of Human Exposure to Perfluoroalkyl Carboxylates (PFCAs): Evaluating Uptake, Elimination, and

Biotransformation of Polyfluoroalkyl Phosphate Esters (PAPs) in the Rat. Environmental Health Perspectives. 119(3): 344–350

Dopierala, K. et. al. (2013) Alkyl- and fluoroalkyltrialkoxysilanes for wettability modification. Applied surface science. 283 453-459

European Chemicals Agency (ECHA), Committee for Risk Assessment (RAC), Committee for Socio-economic Analysis (SEAC) (2017) to the Opinion on the Annex XV dossier proposing restrictions on (3,3,4,4,5,5,6,6,7,7,8,8,8-tridecafluorooctyl)silanetriol and any of its mono-, di- or tri-O- (alkyl) derivatives

Fasano, W., Carpenter, S., Gannon, S., Snow, T., Stadler, J., Kennedy, G., Jr, et al. (2006) Absorption, distribution, metabolism, and elimination of 8–2 fluorotelomer alcohol in the rat. Toxicol Sci. 91:341–355

Hale, S., Arp H., Slinde G., Wade E., Bjørseth K., Breedveld G., Straith B., Grotthing Moe K., Jartun M., Høisæter Å. (2017) Sorbent amendment as a remediation strategy to reduce PFAS mobility and leaching in a contaminated sandy soil from a Norwegian firefighting training facility. Chemosphere. 171 9-18.

Hill, P., Taylor, M., Goswami, P., Blackburn, R. (2017) Substitution of PFAS chemistry in outdoor apparel and the impact on repellency performance. Chemosphere. 181 500-507 Hjortsberg, T. (2017) Analysis of New Emerging Organic Fluorinated Substances with Gas Chromatography-Mass Spectrometry: Development of a GC-MS method for air samples Kemikalieinspektionen (KEMI) (2015) Förekomst och användning av högfluorerade ämnen och alternativ. Rapport från ett regeringsuppdrag. Rapport 6/15

(22)

Kjølholt, J., Astrup Jensen A., Warming, M. (2015) Short-chain Polyfluoroalkyl Substances (PFAS). The Danish Environmental Protection Agency

Lee, S., Lee, S., Choi, M., Kannan, K., Moon, H. (2018) An optimized method for the analysis of cyclic and linear siloxanes and their distribution in surface and core sediments from

industrialized bays in Korea. 236. 111-118

Lorenzo, M., Campo, J., Picó, Y. (2018) Analytical challenges to determine emerging persistent organic pollutants in aquatic ecosystems, Trends in Analytical Chemistry Moody C., Hebert G., Strauss S., Field J. (2003) Occurrence and persistence of

perfluorooctanesulfonate and other perfluorinated surfactants in groundwater at a fire-training area at Wurtsmith Air Force Base, Michigan, USA. Journal of Environmental Monitoring. 341-345

Nørgaard, A. (2010) Mass spectrometric study of nanofilm products Chemistry, exposure and health effects – Ph.D. thesis.

Nørgaard, A. et.al. (2010) Characterization of nanofilm spray products by mass spectrometry. Chemosphere. 80 1377-1386

Nørgaard et.al (2009) Release of VOCs and Particles During Use of Nanofilm Spray Products. Environment Science Technology. 43 7824–7830

Trojanowicz, M. and Koc, M. (2013) Recent developments in methods for analysis of perfluorinated persistent pollutants. Mikrochimica Acta. 180: 957–971.

U.S. Environmental Protection Agency (EPA). (2009) Long-Chain Perfluorinated Chemicals (PFCs) Action Plan

Yu, X., Nishimura, F., Hidaka, T. (2018) Enhanced generation of perfluoroalkyl carboxylic acids (PFCAs) from fluorotelomer alcohols (FTOHs) via ammonia-oxidation process. Chemosphere. 198: 311-319

Zacs, D., Bartkevics, V. (2016) Trace determination of perfluorooctane sulfonate and perfluorooctanoic acid in environmental samples (surface water, wastewater, biota,

sediments, and sewage sludge) using liquid chromatography – Orbitrap mass spectrometry. Journal of Chromatography A. 1473. 109-121.

(23)

7. Appendix

Figure A1. Shows the chromatogram obtained with the low-resolution GC for the compound 6:2, using the method from the paper by Nørgaard et.al. (2010).

(24)

Figure A3. Shown the fragmentation from the peak at 5.127 in the chromatogram in figure A1.

(25)

Figure A5. The chromatogram from an injection of 6:2 PTrEtSi with the m/z 309, 239, 219 and 195.

Figure A6. The chromatogram from an injection of 8:2 PTrEtSi with the m/z 409, 389, 339 and 295.

(26)

Figure A7. The calibration curve obtained from injection of standard solutions in the APGC-MS/MS. Fragment ion m/z 195 from the compound 6:2 PTrEtSi.

Figure A8. The calibration curve obtained from injection of standard solutions in the APGC-MS/MS. Fragment ion m/z 219 from the compound 6:2 PTrEtSi.

(27)

Figure A9. The calibration curve obtained from injection of standard solutions in the APGC-MS/MS. Fragment ion m/z 309 from the compound 6:2 PTrEtSi.

Figure A10. The calibration curve obtained from injection of standard solutions in the APGC-MS/MS. Fragment ion m/z 295 from the compound 8:2 PTrEtSi.

(28)

Figure A11. The calibration curve obtained from injection of standard solutions in the APGC-MS/MS. Fragment ion m/z 339 from the compound 8:2 PTrEtSi.

Figure A12. The calibration curve obtained from injection of standard solutions in the APGC-MS/MS. Fragment ion m/z 389 from the compound 8:2 PTrEtSi.

(29)

Figure A13. The chromatogram obtained from analysis with APGC-MS/MS of cosmetic sample when scanned for 6:2 PTrEtSi.

Figure A14. The chromatogram obtained from analysis with APGC-MS/MS of cosmetic sample when scanned for 8:2 PTrEtSi.

(30)

Figure A15. The chromatogram obtained from analysis with APGC-MS/MS of sludge sample 1 when scanned for 6:2 PTrEtSi.

Figure A16. The chromatogram obtained from analysis with APGC-MS/MS of sludge sample 1 when scanned for 8:2 PTrEtSi.

(31)

Figure A17. The chromatogram obtained from analysis with APGC-MS/MS of sludge sample 2 when scanned for 6:2 PTrEtSi.

Figure A18. The chromatogram obtained from analysis with APGC-MS/MS of sludge sample 2 when scanned for 8:2 PTrEtSi.

(32)

Figure A19. The chromatogram obtained from analysis with APGC-MS/MS of sludge sample 3 when scanned for 6:2 PTrEtSi.

Figure A20. The chromatogram obtained from analysis with APGC-MS/MS of sludge sample 3 when scanned for 8:2 PTrEtSi.

References

Related documents

Data analysis and metabolite annotation is one of the most time consuming and challenging steps in any metabolomics study and the reduced cluster and adduct formation, achieved

Environmentally-relevant contaminants discharged from OSSFs were identified using untargeted analysis with two-dimensional gas chromatography mass spectrometry

From observations of the establishment of tourism in small coastal villages in Zanzibar, local people’s ability to bargain for compensation and shares in revenue was identified to

It was possible to separate and identify each strain of bacterium using multivariate data analysis models with the fatty acid profile of the cell wall.. The fatty acid profile for

Therefore, in order to improve our understanding on MeHg chemical speciation, we developed a novel analytical method for direct determination of MeHg complexes

A first analysis of eight human plasma sample pools and the 20 chosen highly reactive antigens (PrESTs) coupled to beads in a suspension bead array (SBA) was performed to

Här finns exempel på tillfällen som individen pekar på som betydelsefulla för upplevelsen, till exempel att läraren fick ett samtal eller vissa ord som sagts i relation

Mass spectrometry in combination with Hydrogen/deuterium exchange, HDX, has been used for a decade to obtain information about protein folding,10, 97 protein-ligand