Subscriber access provided by United Arab Emirates University | Libraries Deanship
Article
Deconvolution of Soft Ionization Mass Spectra of Chlorinated Paraffins to Resolve Congener Groups Bo Yuan, Tomas Alsberg, Christian Bogdal, Matthew MacLeod, Urs Berger, Wei Gao, Yawei Wang, and Cynthia A. de Wit Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.6b01172 • Publication Date (Web): 17 Aug 2016 Downloaded from http://pubs.acs.org on August 27, 2016
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 22
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
1
Deconvolution of Soft Ionization Mass Spectra of Chlorinated Paraffins to
2
Resolve Congener Groups
3 4 5
Bo Yuan1*, Tomas Alsberg1, Christian Bogdal2,3, Matthew MacLeod1, Urs Berger4, Wei Gao5, Yawei Wang5,
6
Cynthia A. de Wit1.
7 8 9
1
Department of Environmental Science and Analytical Chemistry, Stockholm University, Svante Arrhenius väg
10
8, SE-10691 Stockholm, Sweden
11
2
12
Prelog-Weg 1, CH-8093 Zürich, Switzerland
13
3
Institute for Sustainability Sciences, Agroscope, Reckenholzstrasse 191, CH-8046 Zürich, Switzerland
14
4
Department Analytical Chemistry, Helmholtz Centre for Environmental Research - UFZ, Permoserstraße 15,
15
DE-04318, Leipzig, Germany
16
5
17
Sciences, Chinese Academy of Sciences, Shuangqing Road 18, CN-100085 Beijing, China
Institute for Chemical and Bioengineering, Swiss Federal Institute of Technology, ETH Zürich, Vladimir-
State Key Laboratory of Environmental Chemistry and Ecotoxicology, Research Center for Eco-Environmental
1 ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
18
19
We describe and illustrate a three-step data processing approach that enables individual congener groups of
20
chlorinated paraffins (CPs) to be resolved in mass spectra obtained from either of two soft ionization methods,
21
electron capture negative ionization mass spectrometry (ECNI-MS) or atmospheric pressure chemical ionization
22
mass spectrometry (APCI-MS). In the first step, general fragmentation pathways of CPs are deduced from
23
analysis of mass spectra of individual CP congeners. In the second step, all possible fragment ions in the general
24
fragmentation pathways of CPs with 10 to 20 carbon atoms are enumerated and compared to mass spectra of CP
25
mixture standards and a deconvolution algorithm is applied to identify fragment ions that are actually observed.
26
In the third step, isotope permutations of the observed fragment ions are calculated and used to identify isobaric
27
overlaps, so that mass intensities of individual CP congener groups can be deconvolved from the unresolved
28
isobaric ion signal intensities in mass spectra. For a specific instrument, the three steps only need to be done
29
once to enable deconvolution of CPs in unknown samples. This approach enables congener group-level
30
resolution of CP mixtures in environmental samples, and opens up the possibility for quantification of congener
31
groups.
ABSTRACT
32 33
34
Chlorinated paraffins (CPs) are mixtures of chlorinated n-alkanes. They are widely used product additives, with
35
numerous applications that include use as flame retardants and plasticizers in plastics, paints, sealants, and
36
adhesives, as coolants and lubricants in metal-cutting, and as fatliquors in leather processing.1 Global production
37
of CPs has increased to more than 1 million t/a in 2009.2 CP mixtures can be divided into short-chain (SCCPs,
38
C10 – C13), medium-chain (MCCPs, C14 – C17), and long-chain CPs (LCCPs, C≥18). The Persistent Organic
39
Pollutants (POPs) Review Committee of the UN Stockholm Convention has listed SCCPs as a POP-candidate,
40
citing evidence that SCCPs occur in remote areas such as the Arctic, are toxic to aquatic organisms, and are
41
potentially carcinogenic.3 However, scientific studies on the occurrence and behavior of CPs in humans and the
42
environment are still rare. The major reason for this paucity of data is the challenge posed by CP analysis, even
43
using modern separation and detection techniques.
INTRODUCTION
2 ACS Paragon Plus Environment
Page 2 of 22
Page 3 of 22
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
44
Technical CP mixtures contain thousands of congeners with the general elemental composition CnH2n+2-mClm
45
(hereafter abbreviated as CnClm). Complete separation of individual CPs has so far not been achieved,4 although
46
two dimensional gas-chromatography (GC×GC) techniques show considerable promise.5 The identification of
47
CP congener groups (i.e., CPs with a given number of carbon and chlorine atoms) relies on detection of pseudo-
48
molecular ions by soft ionization mass spectrometry.6 The large number of isomers, chlorine isotope patterns and
49
multiple fragmentations of different CP congeners all contribute to an extremely complex mass spectrum and
50
result in mass interference for pseudo-molecular as well as possible confirmation ions, both within and between
51
congener groups. High instrument resolving power is required to separate interfering ions of the same nominal
52
mass(i.e. isobaric mass).7 For instance, to separate [C10Cl8 – Cl] – (12C101H1435Cl637Cl.-, m/z 380.8886 amu) from
53
[C10Cl6 + Cl] – (12C101H1635Cl7.-, m/z 380.9072 amu), a resolving power of at least 20 500 (10% valley definition)
54
is required. The required resolving powers are significantly higher than those of commonly used magnetic sector
55
or time-of-flight (TOF) high-resolution instruments in routine resolution mode (typically 10 000). In addition,
56
certain fragment ion species, such as the three species [C10Cl8 – Cl]
57
possess identical chemical sum formulas (in this case, C10H14Cl7), and therefore cannot be resolved by mass
58
spectrometry, even with infinite resolving power.
59
Mathematical deconvolution techniques have been applied to resolve signals of high-mass molecules in the field
60
of protein mass spectrometry and to overcome the limitations of instruments with poor resolution.8-12
61
Deconvolution can allow unique identification of individual ion signals within a cluster of overlapping
62
masses.13,14 Some deconvolution methods determine the relative amounts of an individual ion by reconstructing
63
the unresolved mass spectra from the isolated isotopic patterns of the ion candidates.15 However, these
64
applications depend on knowledge of possible ion candidates, and complete knowledge of overlapping ions of
65
CPs is not available due to their compositional complexity and the lack of information about the content of
66
commercial CP products.
67
Recently, Bogdal and coworkers introduced a method for analysis of CP mixtures in environmental samples
68
based on the identification of chloride adducts by deconvolution of chlorine-enhanced atmospheric pressure
69
chemical ionization (APCI) mass spectra.7 However, possible mass interferences due to overlapping fragment
70
ions were not discussed. The present study is a logical extension of the method presented by Bogdal et al. Here,
– ,
[C10Cl7 – H] – and [C10Cl9 – 2Cl + H] –,
3 ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
71
we describe a data processing approach to identify individual CnClm isobaric mass signals in two commonly
72
applied soft ionization mass spectrometry methods, electron capture negative ionization mass spectrometry
73
(ECNI-MS) and APCI-MS. For either ionization method, all possible fragment ions of C10 – C20 CPs can be
74
deduced progressively by mass spectral analysis of CP chemical standards. Isotopic distributions of all the
75
possible ions are enumerated to identify isobaric overlaps. Thereafter, CnClm (i.e., specific congener groups) can
76
be identified from the isobaric ion signals unresolved by the instrument. Typical applications are described and
77
requirements to extend the method to quantify concentrations of congener groups are identified.
78
79
CP Standards. We purchased eight individual CP congeners with defined molecular structure, including carbon
80
chain lengths 10, 11, 12 and 14, with 4 to 8 chlorine atoms. We also purchased three CP chain length standards
81
with a defined carbon chain length (C12) but varying Cl substitution. Finally, seven CP mixture standards with
82
variable chain length and chlorination degree between 42% and 63% were selected for this study. For a detailed
83
list of standards see the supporting information.
84
Instruments. Three different GC/ECNI-MS and one LC/APCI-MS were used in this study: (1) GC/ECNI-
85
quadrupole time-of-flight (QTOF) HRMS (Agilent 7200, Santa Clara, USA); (2) GC/ECNI-magnetic sector
86
HRMS (MS: Thermo Finnigan MAT95, Bremen, Germany); (3) GC/ECNI-quadrupole LRMS (Thermo DSQ II,
87
Bremen, Germany); (4) APCI-QTOF (Waters QTOF Premier, Manchester, UK). For detailed instrumental
88
methods see the supporting information. The GC was only used to separate solvents from the CP analytes, before
89
introducing the CPs into the ECNI-ion source. For the APCI system, the autoinjector of the LC system was used
90
for direct injection into the ion source without chromatography.
91
Data Acquisition. Mass spectra of the CPs using ECNI-QTOF MS and APCI-QTOF MS (full scan acquisition,
92
centroid mode, background subtracted) were obtained by the MassHunter B.06.00 and the MassLynx V4.1 data
93
system, respectively. The magnetic sector HRMS was operated in selected ion monitoring (SIM) mode using the
94
Xcalibur V2.2 data system. The flow injection peak of an individual m/z in APCI-MS was integrated by an
95
automatic algorithm to avoid inconsistencies of manual integration. For integration of “humps” resulting from
96
the gas chromatographic introduction see the supporting information including Figure S-1. In QTOF MS data
CHEMICALS AND INSTRUMENTS
4 ACS Paragon Plus Environment
Page 4 of 22
Page 5 of 22
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
97
acquisition, signals were recorded as centroid m/z including all ions that were unresolved within a given mass
98
range. The mass range can be calculated from the resolving power; for example, an observed centroid m/z
99
444.8706 amu using a QTOF instrument with a resolving power of 10 000 represents the ions within m/z
100
{444.8706 – 444.8706/10 000, 444.8706 + 444.8706/10 000} = {444.8261, 444.9151} amu.
101
102
We processed the data acquired for the three types of CP standards (individual CP congeners, CP chain length
103
standards, CP mixture standards) in three sequential steps with repeated application of deconvolution to find
104
isobaric overlaps that occur at each centroid m/z. The approach is summarized in the flowchart in Figure 1. For a
105
specific instrument, these three steps only need to be done once, then the output of step 3 can be applied for
106
resolving CnClm in unknown samples. The principles of the deconvolution and the data processing steps are
107
described below.
DATA PROCESSING APPROACH
5 ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
STEP 1 INPUT Mass spectra of individual CP congeners
STEP 2 PROCESS INPUT Deconvolution (Equation 1) Identified fragment ions Fragment ion types
Mass spectra of reference standards: • CP chain length standards • SCCP mixture standards • MCCP mixture standards • LCCP mixture standards
PROCESS OUTPUT General fragmentation pathways
General fragmentation pathways Possible fragment ions of all CnClm Deconvolution (Equation 1)
STEP 3 INPUT
OUTPUT
Identified fragment ions
Identified fragment ions of all CnClm
PROCESS Isotope permutation Isotopic distributions of all ions
INPUT Resolving power of instrument
Isobaric overlaps checking
PROCESS
OUTPUT CnClm that have isobaric overlaps at each resolvable centroid mass
Isobaric overlaps to the confirmation ions of each CnClm
OUTPUT
108 109
Instrument-specific deconvolution matrix of each CnClm (Equation 1)
Figure 1. Flowchart summary of the data processing steps.
6 ACS Paragon Plus Environment
Page 6 of 22
Page 7 of 22
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
110
Principle of Deconvolution
111
We apply mathematical deconvolution to resolve individual abundances of overlapping ions. Its principle has
112
been described elsewhere,15,16 and is briefly presented here. The intensity of the m/z observed in mass spectra is
113
equal to the sum of the intensities of the individual overlapping ions at the same centroid m/z, which in matrix
114
notation is
115 116
… = …
…
… … ∙ → = ∙ … … … …
(1)
where {I} is a vector containing the intensities of measured m/z in the unresolved spectra (input), i is the number
117
of measured m/z, and j is the number of possible ions (i ≥ j). {A} is an i × j matrix containing the isotope pattern
118
abundances of all ions (calculated). Solution by least square optimization generates the relative amounts of
119
individual ions in overlapping mass spectra in the vector {X} (output, unknown of the equation system). For the
120
isotopic calculation (matrix A) and the least-square algorithm, see the supporting information.
121
Quality control of the deconvolution consists of mass spectrum reconstruction and centroid mass shift analysis.
122
The former calculates the percent difference value (DV) between the reconstructed and the measured mass
123
spectra in order to evaluate the performance of deconvolution. The equation used to determine DV is16
124
= 100 ∙
125
Centroid mass shift analysis compares the observed centroid mass and the calculated one (based on the
126
knowledge of instrumental mass resolution) to verify the deconvolution result {X}. The calculated centroid mass
127
is approximated from the weighted average of the overlapping m/z.15
128
G/I =
129
where m/zj is the exact mass-to-charge ratio of the fragment ion j.
130
Data Processing Steps
131
Step 1: Deduction of General Fragmentation Pathways. The general fragmentation pathway for a molecule
132
specifies the fragment ions that could be observed using mass spectrometry. The first step in our approach is to
! "#$%&'($) *+,-./012,0+3 456789:6;
