Development of a Multi-Stage Liquid Extraction Method for ...

42
Development of a Multi-Stage Liquid Extraction Method for the Quantitative Analysis of Flavors in Tobacco Filler John Mathis, Rayman Stanelle, Jean-Marie Dimandja U.S. Food and Drug Administration Office of Regulatory Affairs/Office of Regulatory Science/ Southeast Tobacco Laboratory 2017_TSRC78_Mathis.pdf TSRC2017(71) - Document not peer-reviewed

Transcript of Development of a Multi-Stage Liquid Extraction Method for ...

Development of a Multi-Stage Liquid Extraction Method for the Quantitative

Analysis of Flavors in Tobacco Filler

John Mathis, Rayman Stanelle, Jean-Marie DimandjaU.S. Food and Drug Administration

Office of Regulatory Affairs/Office of Regulatory Science/Southeast Tobacco Laboratory

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

www.fda.gov

Reference to any commercial materials, equipment, or process does not, in any way, constitute approval, endorsement, or recommendation by the US Food and Drug Administration.

This information is not a formal dissemination of information by FDA/CTP and does not represent Agency position or policy.

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Presentation Outline

• Method Overview

• Product Analysis Review

• Flavored Product Discussion

• Validation Data and Analysis

3

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Acknowledgements

Tobacco Laboratory• Khalil Kerdahi• Ed Phifer• John Casanova• Marcela Mcleod• Javier Vega• Julio Arrecis• Ashley Keys, Alesia

Parker, Danielle Pierre, Freddy Ortiz, Yiming Ye

• Center for Tobacco Products

o Office of Compliance and Enforcement• Racheal Schmidt

o Office of Science• Ken Taylor• Lana Rossiter• Matthew Holman

4

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Method Preparation Details• Liquid extraction of tobacco filler in ethyl

acetate, partition with water, centrifugation, dilution with hexane

• Analysis by GC-MS

• Linear range for 1 g filler

o 10 µg/g − 250 µg/g (ppm)

5

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Highlights: Extraction of Filler• Extraction in ethyl acetate (EtOAc)

o Add EtOAc, vortex• Water Addition

o Vortex, centrifuge; results in two-layers, partition of matrix components

• Remove EtOAc and add Hexane with ISo Vortex, centrifuge; forces water out of EtOAc

• Alternatives investigated

6

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Cigarillo Chromatogram

Reference to any commercial materials, equipment, or process does not, in any way, constitute approval, endorsement, or recommendation by the US FDA. 7

Abundance

Time (minutes)

Targeted Compounds: Listed in Market Sample Results

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Product Analysis• History of tobacco filler analysis

o SPME and GC-MS; 2009-12 • Many flavored products analyzed• Prevalent compounds related to flavors identified

o Focus on finding compounds related to flavorso Liquid Extraction and GC-MS; 2012-2016

• Methods validated for quantificationo Goal to quantify compounds related to flavors

8

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

9www.fda.gov

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Highlights: Extraction of Filler• Extraction in organic solvent(s)

o Weigh sample, Add liquid, vortex

• Water Addition – Liquid-Liquid Extractiono Immiscible solvent; results in two-layers

Vortex, centrifuge, filter, analyze• Partitioning of matrix components

10

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Extraction Optimization• Spiked 3R4F filler analyzed with different treatments

• Solvent Selection• Sample Treatment• Water Addition

o Data units: • Raw Peak area for treatment comparisons • Validation: parts per million (ppm or μg/g) by external

standard calibration using an internal standard 3,4-(Methylenedioxy) acetophenone (MDA, I.S.)

11

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Sample Preparation Summaryo Organic Solvent Extraction

• Investigated different solvents: Ethyl Acetate, Acetonitrile, Methanol, Hexane, methyl t-butyl ether

• Compared sample treatment – grinding, conditioning• Water partitioning- water purity, salt, or buffer (pH)

o Many iterations for optimization

o Additional processing steps tested• Use more or less sample as needed• Dilution with hexane containing IS to remove H2O

12

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

13

0

200000

400000

600000

800000

1000000

1200000

1400000

whole ambient nowater

Methanol extractwhole ambient 3R4F

2 mL waterwholeconditioned

Cryogrind EtOAc3R4F whole

WholeEtOAc extract WholeEtOAc extract

Ethyl Maltol

Hexane Hex Methanol Ethyl Acetate EtOAc/Water

cryogrind

cryogrind

Peak

Are

a /

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Scope of Evaluation• Alternate solvents (no water addition)

o Methanol o Acetonitrile/Ethyl Acetate mixtures

• Add varying amounts of water up to 2 mL to hexane extraction of a 100 ppm spike

• Effect of Conditioning prior to testing• Water Addition – Liquid-Liquid Extraction

14

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Extraction Solvents

15

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Scope of Evaluation• Alternate solvents (no water addition)

o Methanol o Acetonitrile/Ethyl Acetate mixtures

• Add varying amounts of water up to 2 mL to hexane extraction of a 100 ppm spike

• Water Addition – Liquid-Liquid Extractiono Comparison of treatments

16

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Water Addition

17

0

0.2

0.4

0.6

0.8

1

1.2

wholeambient no

water

0.1 mLwaterwhole

ambient

0.2 mLwaterwhole

ambient

0.5 mLwaterwhole

ambient

2 mLwaterwhole

ambient

Peak

Are

a/M

illio

ns

Ethyl Maltol

Ethyl Maltol

Organic 0.1 mL 0.2 mL 0.5 mL 2mLno water | water in 5 mL Hexane |

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Water Addition

18

0

0.05

0.1

0.15

0.2

0.25

0.3

wholeambient no

water

0.1 mLwaterwhole

ambient

0.2 mLwaterwhole

ambient

0.5 mLwaterwhole

ambient

2 mLwaterwhole

ambient

Peak

Are

a/M

illio

ns

Vanillin

Vanillin

Organic 0.1 mL 0.2 mL 0.5 mL 2mLno water | water in 5 mL Hexane |

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Preparation Findings• Add water to hexane – Some differences

for highly water soluble compounds but not as effective as using other solvents

• Effect of conditioning – No difference• Alternate solvents – Methanol and EtOAc• Water Addition – Highest recoveries using

EtOAc + water, 5mL each

19

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Scope of Evaluation• Alternate solvents (no water addition)

o Methanol o Acetonitrile/Ethyl Acetate mixtures

• Add varying amounts of water up to 2 mL to hexane extraction of a 100 ppm spike

• Water Addition – Liquid-Liquid Extractiono Comparison of treatments

20

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Liquid-Liquid Extraction

21Drying salt: Sodium Sulfate, Sodium Chloride Buffer salt: Ammonium Acetate

Organic: Ethyl Acetate (EtOAc) and/or Hexane (Hex)

0

1

2

3

4

5

6

Benzaldehyde Benzyl Alcohol Benzyl Acetate Ethyl Maltol Eugenol Vanillin g-decalactone MDA Nicotine

Peak

Are

a / M

illio

ns Organic/H2O Organic Salt Buffer EtOAc/H2O+Hex30%

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Ethyl Acetate and Water

• Suitable Solvents: Methanol and Ethyl Acetate

• Methanol extracts polar components, including humectants and triacetin

• Ethyl Acetate provides consistent recovery and enhanced with water

22

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Extraction Method Summary• Evaluated different solvents

o Ethyl Acetate (EtOAc) with addition of water

• Additional optimization: o Water solubility in EtOAc:

• Effective for extraction but not GC-MSo Internal Standard (IS) additiono Centrifugation to provide layer separation

23

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Ethyl Acetate and Water• Water solubitility is 3.3% in EtOAc

o Known degradation of GC injection and separation

• Internal standard, 3,4-MDA water solubleo Add at Hexane addition

• Use practical autosampler wash solvents: o Ethyl Acetate and EtOAc/Hexane mixture

24

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Sample Preparation Summary• Using ethyl acetate and water preparation

• Results from mixed filler comparable to reduced particle size – no grinding required

• Additional processing steps tested• Use more or less sample as needed

• Deionized water-lab supplied, no salt or buffer (pH)

• Dilution with hexane (I.S.) to remove H2O

• Centrifugation sufficient to push filler at water level

25

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Method Preparation Details• Liquid extraction in ethyl acetate,

partition with water, centrifugation, dilution with hexane (containing I.S.)

• Analysis by GC-MS

26

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

27www.fda.gov

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Method Preparation Details• Filler weighed or broken into pieces

• Liquid extraction in ethyl acetate, partition with water, centrifugation, dilution with hexane (containing I.S.)

• Analysis by GC-MS using built-in calibration, QA verification

• Linear range: 10 µg/g − 250 µg/g (ppm)

28

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Flavored Sample Analysis• Sample Description

o Samples for validation was 3R4F filler • Spiked at low, mid, and high levels for recovery and precision• Blank matrix used to prepare standards

• Market Sampleso Commercially available samples of filler from cigarettes,

filtered little cigars, cigarillos, and large cigars• Single sample of each matrix spiked at mid-level for recovery

evaluation

• Samples stored in freezer prior to analysis

29

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Experimental Details• Matrices: Tobacco filler and others• Applied method to various brands of

cigarettes, filtered little cigars, and cigarillos• Remaining outline

o Cherry flavored productso Method validation and capabilities o Spiked sample results

30

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Method Application

• Quantitated compounds shown, plus ethyl vanillin and vanillin

• Additional unknowns observed

31

Cherry Cigarillo Sample

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Cherry Sample Results

32

Concentration Profiles (ppm) of Filtered Little Cigars and Cigarillos

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Market Sample Results

33

Compound Little Cigar A %RPD Little

Cigar B %RPD Cigarillo A %RPD Cigarillo B %RPD

Benzaldehyde 450 5% 15 -20% 40 3%

Benzyl Alcohol 12 9% 1290 1% 3600 0%

Ethyl Maltol 65 8% 910 3% 210 5%

Benzaldehyde Propylene Glycol 34 7% 20 2%

Piperonal 430 -1% 875 1% 735 1%

Eugenol 65 2%

Vanillin 675 -1% 450 4% 4300 7% 3365 -6%

Ethyl Vanillin 365 -1% 8000 -9% 4590 -7%

RPD: Relative Percent Difference between duplicate preparations

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Method Validation Results

Reference to any commercial materials, equipment, or process does not, in any way, constitute approval, endorsement, or recommendation by the US Food and Drug Administration. 34

Quantitation Ion

Qualifier Ion - 1

Qualifier Ion - 2

Next Compound, different ion

Method Specificity illustrated by resolution between common ions, no interfering peaks, and consistent ion ratio

Abundance

Time (minutes)

Quant ion chromatogram

Apex spectrumPeak Area, concentration and ion ratios

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Method Validation Results

35www.fda.gov

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Validation: Recovery ResultsCompounds (%) LOQ Low Level Mid Level Upper Level

(ppm) 10 50 100 150% Recovery % RSD % Recovery % RSD % Recovery % RSD % Recovery % RSD

Benzaldehyde 98% 5% 111% 4% 106% 4% 106% 4%Benzyl Alcohol 102% 4% 103% 5% 101% 5% 101% 5%

Linalool 98% 4% 117% 3% 111% 4% 110% 3%

Menthol 90% 5% 112% 5% 109% 5% 109% 4%Benzyl Acetate 97% 4% 113% 4% 108% 4% 108% 4%Ethyl Maltol 116% 6% 97% 7% 94% 6% 95% 7%Pulegone 91% 4% 111% 4% 107% 4% 107% 4%Benzaldehyde Propylene Glycol 93% 4% 110% 4% 106% 4% 107% 4%

Piperonal 98% 6% 109% 4% 105% 5% 106% 4%Methyl Anthranilate 95% 5% 111% 4% 107% 5% 107% 4%gamma-Decalactone 86% 5% 116% 3% 111% 4% 111% 3%Eugenol 97% 5% 112% 4% 108% 5% 108% 4%Vanillin 108% 8% 107% 5% 102% 6% 103% 5%Coumarin 101% 5% 109% 4% 104% 5% 105% 5%Ethyl Vanillin 106% 6% 110% 4% 105% 6% 106% 5%Raspberry Ketone 109% 6% 113% 5% 104% 7% 106% 6%

36

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

37

Deviation from Predicted Value

Comparison of Spiking before/after Extraction

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Method Validation Results

38Overlay of TIC of cigarillo sample (black) and LOQ spike (blue)

Abundance

Time (minutes)

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Detection of linalool oxide at LOQ

• Extracted Ion Chromatograms, 10 ppm standard• cis-/trans- isomers: ≥90% Recovery, 5% RSD

39

S/N: 169

S/N: 36

S/N: 11

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Method Validation Results

40

Compounds (%) CCV, % %RSD Spiked Filler %RSDBenzaldehyde 102 2% 104% 5%Benzyl Alcohol 102 2% 101% 6%Linalool 104 2% 100% 4%Menthol 101 2% 94% 5%Benzyl Acetate 102 2% 104% 4%Ethyl Maltol 102 2% 108% 7%Benzaldehyde Propylene Glycol 101 1% 102% 4%Piperonal 101 1% 104% 4%Methyl Anthranilate 101 1% 99% 5%gamma-Decalactone 104 1% 100% 4%Eugenol 101 1% 106% 5%Vanillin 101 1% 98% 5%Coumarin 101 1% 106% 5%Ethyl Vanillin 101 1% 100% 5%Raspberry Ketone 102 3% 101% 6%

CCV: Continuing Calibration Verification; Spiked after extraction vs Spiked Filler, before(Bold: compounds found in market samples)

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

Flavor Method Summary• Optimization of Sample Treatment

• Practical approach to solvent selection and liquid-liquid extraction procedure

• Validation of preparation method • Samples and matrix matched standards used to

demonstrate applicability

• Evaluated various brands o Illustrated profiles for cherry samples

41

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed

2017

_TS

RC

78_M

athi

s.pd

fT

SR

C20

17(7

1) -

Doc

umen

t not

pee

r-re

view

ed