identification of unknown organic compounds based on...

Identification of unknown organic compounds based on

comparison of electron ionization mass spectra

Andrey S. Samokhin

Igor A. Revelsky

Chemistry Department

Lomonosov Moscow State University

Identification of unknown organic compounds

2

Nuclear

magnetic

resonance

(NMR)

spectroscopy

Infrared

(IR)

spectroscopy

Ultraviolet

(UV)

spectroscopy

Mass spectrometry

(MS)

Electrospray ionization

Atmospheric pressure

chemical ionization

Atmospheric pressure

photoionization

Electron ionization (EI)

Chemical ionization

High-performance liquid

chromatography/mass spectrometry

(HPLC/MS)

Gas chromatography/mass spectrometry

(GC/MS)

• Formation of a large number of fragment ions

• High (interlaboratory) reproducibility of mass spectra (in

comparison with other ionization techniques)

Features of electron ionization mass spectrometry (70 eV)

3

Electron ionization mass spectral databases- NIST'11 (213000 compounds)

- Wiley 9th (592000 compounds)

M + e– → M+• + 2e–

Electron ionization ion sourceFilament (cathod)

Anode

Ion lenses

Identification based on search against mass spectral database (Library search)

4

50 80 1100

50

10055

7083

97 112

50 90 1300

50

100

51 64 77 102

128

…

Similarity index

Experimental mass spectrum of

unknown compound

Result of library search

5

• Result of library search is the list of possible candidates (―Hit

table‖). Possible candidates are ranked by Similarity index

• ―Hit table‖ obtained using MS Search 2.0 software:

# Lib. Match R.Match Prob. Name

1 M 855 855 27.4 Nonanoic acid, methylester

2 M 849 849 21.6 Undecanoic acid, methylester

3 M 845 845 18.2 Decanoic acid, methylester

4 M 838 838 14.0 Tridecanoic acid, methylester

5 M 819 819 6.77 Dodecanoic acid, methylester

Correct compound does not occupy the first position in the

“hit table” in approximately 25% of cases[1]

[1] Stein S.E., Scott D.R., J. Am. Soc. Mass Spectrom., 1994, vol. 5, p. 859–866.

Differences between experimental and library mass spectra of testosterone

6

(Spec. Edit) sample19#1289-1293 RT: 23,46-23,51 AV: 5 SB: 29 23,24-23,38 , 23,69-23,84

50 80 110 140 170 200 230 260 2900

50

100

5579

91 109

124

133

147

165 185

203228

246

270

288

(Spec. Edit) testosterone (Thevis' book)

50 80 110 140 170 200 230 260 2900

50

100

55

67

7991

105

124

133

147

159187

203

228

246

273

288

(Text File) testosterone (Sigma)

50 80 110 140 170 200 230 260 2900

50

100

55

79

91

105

124

133

147

159 187

203

228

246

273

288

(Text File) Testosterone (NIST)

50 80 110 140 170 200 230 260 2900

50

100

55677991

105

124

133

147

165185

203228

246

273

288(a)

(b)

(c)

(d)

(a) – NIST mass spectral database (http://webbook.nist.gov/)

(b) – mass spectra registered by us using quadrupole instrument ―DSQ‖ (Thermo)

(c) – Thevis M. Mass Spectrometry in Sports Drug Testing, Wiley, 2010. 360 p.

(d) – Sigma-Aldrich, Product Information, Testosterone (T 5411).

Dissimilarity of registration conditions

7

• Technique used for introducing sample into a mass

spectrometero direct inlet

o combination with gas chromatograph

• Temperature of ion source

• Type of mass analyzero quadrupole

o time-of-fligth

o ion trap

o magnetic sector

• Method of instrument calibration

Prediction of electron ionizationmass spectra

8

Electron ionization mass spectra cannot be fully predicted

by experts and simulated using any software:

• Theory of molecular ion fragmentation gives only general

information about set of ions

• Software (e.g. Mass Frontier, ACD Labs, MOLGEN) can be

used to predict only the set of m/z values corresponding to

possible fragment ions (intensities of mass spectral peaks

cannot be predicted)

Therefore identification of unknown compounds (by means

of electron ionization mass spectrometry) should be based

on comparison with experimental mass spectra

Reliable identification by means of electron ionization mass spectrometry

9

For the most reliable identification by means of electron

ionization mass spectrometry the full mass spectra of

unknown compound and possible candidate should be

registered under identical experimental conditions and

compared

The main problem is how to compare full mass spectra,

which may contain hundreds of peaks

Possible approaches for mass spectra comparison

10

• Visual comparison

• Application of mathematical algorithms used in library search

programs

• Comparison of intensities of individual mass spectral peaks

(Text File) sub N7 Ion Trap (average) (min_I=5) v.0.2.

60 80 100 120 1400

50

10057

71

8597 111 126


60 80 100 120 1400

50

10057

71

8597 110 126

Visual comparison

11

• Visual comparison is very subjective

• Mass spectra similarity cannot be estimated quantitatively

n-Dodecane(MW = 170)

n-Tridecane(MW = 184)

Do these mass spectra correspond to the same

compound?

960 980 10000,00

0,15

0,30

0,45

Similarity index

Application of mathematical algorithms used in library search programs

12

Samokhin A.S., Revelsky I.A., Revelsky A.I. Book of Abstracts. Fourth All-RussianConference "Fundamental issues of mass spectrometry and its analyticalapplication―. October 10–14, 2010, Zvenigorod, Russia. P. 58–59.

Test set of model compounds

GC/MSanalysis

Creation ofin-house mass spectral library(by averaging n

replicate spectra)

n replicate

experiments

Search againstin-house

mass spectral library

k replicate

experiments

Distribution of values of the

similarity indexes(corresponding to

correct identification)

Calculating the parameters of the

distributionMean = 977StD = 9.2


13


960 980 10000,00

0,15

0,30

0,45

Similarity index

Normaldistribution

960 980 10000,00

0,15

0,30

0,45

Similarity index

95%

Normaldistribution Critical value of

Similarity index = quantile95%

(corresponding to reliable

identification)

Mass spectra correspond to the same compound, if

similarity index between these spectra is more than critical

value (equaled to quantile95%)


14

• Magnitudes of critical values calculated for different

instruments are not equal

The main limitation is that the magnitude of critical value

depends on registration conditions

• The experiments (including registration of hundreds of

spectra) should be repeated every time, when conditions are

changedo ion source contamination

o method of instrument calibration

o etc.


Comparison of intensities of individual mass spectral peaks

15

[1] Commission Decision 2002/657/EC, Off. J. Eur. Commun., 12 August. 2002[2] Guidance for industry: mass spectrometry for confirmation of the identity of animaldrug residues: final guidance. FDA, Washington, DC; 2003.[3] WADA Technical Document—TD2010IDCR, 2011.

• only several (usually 3) mass spectral peaks are considered

• width of tolerance window is quite large

Tolerance windows for relative intensities of mass spectral

peaks[1–3]

Intensity[3]

(normalized to the base peak)Width of window[3]

>50% ±10%relative

20–50% ±15%relative

10–20% ±20%relative

≤10% ±50%relative

Distinguishing electron ionization mass spectra of isomers by PCA

16

• Hejazi et. al have used PCA to distinguish geometrical

isomers of α-linolenic acid methyl ester[1]

• We have shown possibility of distinguishing between

o-xylene, m-xylene and p-xylene[2]

-5 -3 -1 1 3

-2

2

PC1

PC2

o-Xylene

m-Xylenep-Xylene

[1] Hejazi L., Ebrahimi D., Guilhaus M., Hibbert D.B., J. Am. Soc. Mass Spectrom.,2009, vol. 20, p. 1272–1280.[2] Samokhin A., Revelsky I., Eur. J. Mass Spectrom. 2011, vol. 17, p. 477–480.

Goal

17

The goal of this work was to develop simple approach

(based on using principal component analysis) for reliable

comparison of mass spectra registered under identical

experimental conditions

Scheme of analysis

18

n1 replicate mass spectra of unknown

compound

n2 replicate mass spectra of possible

candidate

Intensities of mass spectral peaks were normalized to the total ion current:

%100absolut

/

absolut/relative

/

i

izm

jzm

jzmI

II

m/z values

mass s

pectr

a

Principal component

analysis

What can we expect?

19

Unknown compound

ispossible

candidate

Unknown compound

is notpossible

candidate

Principal component

analysis

1. There is only one groupof objects on the scoreplot

2. All PCs contain onlynoise

1. There are two groups ofobjects on the score plot

2. PC1 should separateobjects corresponding todifferent groups

PC1

PC1

Score plot

PC1

PC1

Score plot

Therefore we can useonly PC1 to comparemass spectra

The first principal component

20

PC1 = p1∙I1 + p2∙I2 + … + pn∙In

It is believed, thatintensities of massspectral peaks havenormal distribution

PC1 has normaldistribution as well

Student's t-test canbe used tocompare groups ofobjects

Therefore we transform multidimensional analytical signal

to the one-dimensional one

Scheme of analysis

21

n1 replicate mass spectra of unknown

compound

n2 replicate mass spectra of possible

candidate

m/z values

mass s

pectr

a

Principal component

analysis

PC1

Mean1

StD1

Mean2

StD2

Student’s t-test:

21

2121

StD

MeanMeanξ

nn

nn

If ξ ≤ t (P, f = n1 + n2 – 2),unknown compound is possible candidate

If ξ > t (P, f = n1 + n2 – 2),unknown compound is not possible candidate


60 80 100 120 1400

50

10057

71

8597 111 126

n-Dodecane


60 80 100 120 1400

50

10057

71

8597 110 126

n-Tridecane

Example #1(discrimination of hydrocarbons)

22

Mass spectra corresponded to 1 ng of respective analyte23 mass spectral peaks were taken into account

-2 -1 0 1 2

-2

-1

0

1

2

PC1

PC2

n-Dodecanen-Tridecane

Mean = –1.73StD = 0.17

Mean = 1.73StD = 0.44

ξ = 12.7

P t (P, 4)

0.95 2.8

0.99 4.6

0.999 8.6

Mass spectra correspond to the different compounds

Example #2(discrimination of β-HCH and δ-HCH)

23

(Text File) b-HCH

60 90 120 150 180 210 240 2700

50

10050

61

77

85

96

111

146

181

219

254

β-hexachlorocyclohexane

(Text File) d-HCH

60 90 120 150 180 210 240 2700

50

10050

61

77

85

96

111

146

181

219

254

δ-hexachlorocyclohexane

-1 0 1

-1

0

1

PC1

PC2

-HCH-HCH

Mean = –1.15StD = 0.08

Mean = 1.15StD = 0.18

Mass spectra corresponded to 0.5 ng of respective analyte45 mass spectral peaks were taken into account

ξ = 20.5

P t (P, 4)

0.95 2.8

0.99 4.6

0.999 8.6


tR

I

Example #3(identification of “unknown” compound)

GC/MS analysis

Extraction of pure mass spectrum

(Text File) 2,4-X, 0.1ng (Av.: 1,3,6)

50 60 70 80 90 100 110 120 1300

50

100

5165

77

91

94

107

122

m/z

24

Library search against mass

spectral database

# Name Similarity index

1 2,3-xylenol 8962 2,5-xylenol 8883 2,6-xylenol 8754 2,4-xylenol 8705 3,4-xylenol 859

RI ~ 1155 iu

―Unknown‖ sample

Name RI [1], iu

2,6-xylenol 11152,4-xylenol 11502,5-xylenol 11523,5-xylenol 11702,3-xylenol 11803,4-xylenol 1196

[1] Mjos S., Meier S., Boitsov S, J. Chromatogr. A 2003, vol. 1123, p. 98–105.


25

GC/MS analysis

GC/MS analysis

“Unknown” compound

Possible candidate

(2,5-xylenol;C = 0.1 ng uL–1)

Retention times

tR(―unknown‖) = 261.70 sec

tR(2,5-xylenol) = 261.85 sec

∆tR = 0.15 sec∆tR = 0.06%

(Text File) 2,4-X, 0.1ng (Av.: 1,3,6)

60 80 100 1200

50

100

5165

77

91

107

122

―Unknown‖

(Text File) 2,5-X, 0.2ng

60 80 100 1200

50

100

5165

77

91

107

122

2,5-xylenol

Mass spectra


26

(Text File) 2,4-X, 0.1ng (Av.: 1,3,6)

60 80 100 1200

50

100

5165

77

91

107

122

―Unknown‖


60 80 100 1200

50

100

5165

77

91

107

122

2,5-xylenol

Intensities of chromatographic peaks (of ―unknown‖ compound and2,5-xylenol) were approximately equal26 mass spectral peaks were taken into account

-2 -1 0 1 2

-2

-1

0

1

2

PC1

PC2

"Unknown"2,5-xylenol

Mean = –1.58StD = 0.04

Mean = 1.58StD = 0.18

ξ = 29.8

P t (P, 4)

0.95 2.8

0.99 4.6

0.999 8.6


tR

I


GC/MS analysis

27

RI ~ 1155 iu

―Unknown‖ sample

Name RI [1], iu

2,6-xylenol 11152,4-xylenol 11502,5-xylenol 11523,5-xylenol 11702,3-xylenol 11803,4-xylenol 1196

[1] Mjos S., Meier S., Boitsov S, J. Chromatogr. A 2003, vol. 1123, p. 98–105.


28

(Text File) 2,4-X, 0.1ng (Av.: 1,3,6)

60 80 100 1200

50

100

5165

77

91

107

122

―Unknown‖

(Text File) 2,4-X, 0.1ng (Av.: 2,4,5)

60 80 100 1200

50

100

5165

77

91

107

122

2,4-xylenol

Intensities of chromatographic peaks (of ―unknown‖ compound and2,4-xylenol) were approximately equal26 mass spectral peaks were taken into account

-1 0 1

-1

0

1

PC1

PC2

2,4-xylenol"Unknown"Mean = –0.08

StD = 0.30Mean = 0.08StD = 0.34

ξ = 0.62

P t (P, 4)

0.95 2.8

0.99 4.6

0.999 8.6

Mass spectra correspond to the same compound


29

(Text File) 2,4-X, 0.1ng (Av.: 1,3,6)

60 80 100 1200

50

100

5165

77

91

107

122

―Unknown‖


60 80 100 1200

50

100

5165

7791

107

122

2,4-xylenol

Intensity of chromatographic peak corresponding to 2,4-xylenol was at 5times more than intensity of peak corresponding to ―unknown‖ compound26 mass spectral peaks were taken into account

-1,0 -0,5 0,0 0,5 1,0

-1,0

-0,5

0,0

0,5

1,0

PC1

PC2

"Unknown"2,4-xylenolMean = –0.47

StD = 0.07Mean = 0.47StD = 0.32

ξ = 5.0

P t (P, 4)

0.95 2.8

0.99 4.6

0.999 8.6

Conclusions

30

• We have developed simple approach (based on usingprincipal component analysis) for reliable comparison of fullelectron ionization mass spectra

o Mass spectra should be registered under identicalexperimental conditions

o Mass spectra should correspond approximately the sameamounts of analytes

o Both unknown compound and possible candidate should beanalyzed at least three times

• Applicability of developed approach was shown in a numberof examples

• Dependence of electron ionization mass spectra on amountof substance was shown

Acknowledgement

31

• We are grateful to Leco company for providing us with a

time-of-flight mass spectrometer coupled with a gas

chromatograph for our research

• We are grateful to Dr. A.V. Garmash for discussion of

presented results

identification of unknown organic compounds based on...

Documents