Mineral oil in vegetable oils : Laboratory and Industrial

Mineral oil in vegetable oils : Laboratory and Industrial Issues Florence LACOSTE, ITERG

Recherche . Innovation . Qualité

DGF Symposium, 10 March 2016

Agenda  Mineral oil composition  EFSA Scientific Opinion on Mineral Oil Hydrocarbons in food  Methods of analysis  Standardization work (ISO & CEN levels)  Origin of the contamination

 Examples of contamination  Effect of refining on the mineral oil contamination  FEDIOL code of practice for the management of mineral oil hydrocarbons in vegetable oils  Conclusion

2


Mineral oil (MO) composition

EFSA Journal 2012;10(6):2704

MOH

Food grade

(C10 –C50)

Paraffins MOSH 65% à 85% Naphtenes

Technical grade MOAH 15 % à 35%

Aromatics (PAHs)

3


Mineral oil (MO) composition Natural hydrocarbons

 Chemical composition of MOH largely unknown  Depending on the crude oil source and the process, MOH products composition can vary batch to batch

Mineral oil : unresolved complex mixture

 Number of isomers increases exponentially with the number of carbon atoms (number of substances > 100 000)  unresolved complex mixture  Quantification of total MOSH and total MOAH is achievable

0

1

2

3

4

5

6

7

8

9

10

11

12

13

Contaminated Sunflower oil (MOSH analysis)

4


14

Mineral oil: EFSA Scientific Opinion

EFSA Journal 2012;10(6):2704

CONTAM Panel conclusions:  Potential concern associated with the current exposure to MOSH in Europe (white oils as release agents for bread and rolls, spraying of grains)  Exposure to MOAH through food considered to be of potential concern because of its specific health risk  Total MOSH and MOAH should be separately determined  2 sub-classes to be distinguished: MOSH up to n-C16 & MOSH from n-C16 to n-C35

5


Mineral oil determination - Critical points Principle of the methods :  First step : isolation of hydrocarbons by liquid chromatography or HPLC  Second step (optional) : extra purification and concentration in order to increase the sensitivity  Third step : gas chromatography analysis with FID detection

Critical points:    

6

Integration of the hump & subtraction of the “natural hydrocarbons” Quantification of hydrocarbons without loosing the volatile ones Removal of interference compounds Limit of quantification as low as possible DGF Symposium, 10 March 2016

Ring test organised by a French producer  6 European laboratories using their own method + 1 reference Labo  8 different samples

 dispersion of the results (between 40 and 130 for sample H) 180

Reference Lab

160

2010

140 120 100 80 60 50 40 20 0 A

B

C

D

E

F

In 2010, it was stated that a reference method was needed 7-


G

H

Standardization : on-going work ISO/TC34/SC11 ISO work :

CEN/TC275/WG13 EC requirement :

 determination of saturated aliphatic  determination of mineral oil hydrocarbons in vegetable oils (MOSH)  method able to distinguish MOSH (Mineral  crude and refined vegetable oils Oil Saturated Hydrocarbons) and MOAH (Mineral Oil Aromatic Hydrocarbons)  pre-trial in 2012  crude and refined vegetable oils  collaborative study in 2013 & 2014  foodstuffs on basis of vegetable oils

ISO 17780: 2015 for MOSH

8


Method for MOSH & MOAH Expected publication : 07/2017

ISO 17780 : MOSH determination « mineral oil »

Fractionation of the sample by LC on silica gel impregnated with AgNO3 Quantification with C18 (IS) Injection of a reference standard blend (C10C40) and a C48 standard GC condition on an apolar column (15 m) Double integration (subtraction of the peaks above the hump)

Silica gel with AgNO3 18,5 g Test portion  1 g Elution with hexane  55 ml

9


« natural hydrocarbons »

ISO 17780 : Validation 2013 Trial : 7 oil samples + 2 foodstuffs  virgin olive O + huile minérale  refined pomace olive O  crude palm O + gasoil  refined grapeseed O  refined sunflower O  refined sunflower O + MOSH  crude soja O + MOSH  margarine & mayonnaise + MOSH

2012 Pre-trial : 2 oil samples  virgin olive O + 50 ppm  crude soja O +450 ppm pomace olive (163 mg/kg)

olive (50 mg/kg)

Laboratories : 39 participants from 12 pays soja (90 mg/kg)

grapeseed (235 mg/kg)

2014 Trial : 6 oil samples  virgin olive O + MOSH  refined pomace olive O  refined rapeseed O  refined sunflower O + MOSH  crude soja O + MOSH

Laboratories : 21 participants from 9 pays

10


ISO 17780 : Validation

160 140 120 100 80 60 40 20 0

Sample 4: 50 mg/kg spiked virgin olive oil Lab mean value (mg/kg)

Lab mean value (mg/kg)

Sample 2: 50 mg/kg spiked olive oil

26 3 8 23 21 34 15 31 1 19 4 33 36 9 28 17 30 27 29 24 11 10 35 5 32 12 6 22 2 18 20 14

160 140 120 100 80 60 40 20 0 20 12 14 10

9

13 16 15

1

11 19

4

8

Laboratories

Laboratories

Mineral oil – 50 mg/kg spiked olive oil

7

sample 2-2013

sample 4-2014

Spiking level (mg/kg)

50

50

Mean value (m)

52

52

Repeatability limit (r)

19,3

15,3

Reproducibility limit (R)

51,7

28,1

Horrat value

4,1

2,2

 Improvement of the precision data by selecting experienced laboratories 11


5

2

6

17

ISO 17780 : 2014 collaborative trial sample 1-2014 sample 2-2014 sample 3-2014

pomace olive oil refined rapeseed oil spiked refined sunflower oil

sample 4-2014 sample 5-2014 sample 6-2014

spiked virgin olive oil = sample 2-2013 spiked refined sunflower oil crude soybean oil

Mineral oil - MOSH

1

2

3

4

5

6

Number of participating laboratories

20

20

20

20

20

20

Number of laboratories retained after elim. outliers

18

18

18

18

17

18

-

-

100

50

25

-

Mean (m)

119

11

107

52

38

90

Repeatability standard deviation (sr)

4,4

3,1

3,7

5,5

2,7

4,9

Repeatability relative standard deviation (RSDr)

3,7%

28,1%

3,4%

10,5%

7,2%

5,4%

Repeatability limit (r)

12,4

8,7

10,3

15,3

7,6

13,7

Reproducibility standard deviation (sR)

14,8

8,2

11,2

10,0

9,6

9,2

12,4%

73,9%

10,4%

19,3%

25,7%

10,2%

41,4

22,9

31,2

28,1

27,0

25,6

Spiking level (mg/kg)

Reproducibility relative standard deviation (RSDR) Reproducibility limit (R)

Scope  Method applicable from 50 mg/kg 12


MOSH and MOAH Determination – 2014 trials Determination of aliphatic and aromatic hydrocarbons in animal and vegetable fats and oils with on-line-HPLC-GC-FID analysis” Principle of the method  Fractionation of MOSH & MOAH with HPLC/UV  On-line injection with large volume of each fraction followed by GC/FID analysis  Optional purification by LC on a SilicaALOX column (to get rid of C23-C33 nalkanes)  Optional clean-up by an epoxidation procedure (to avoid interferences with olefinic substances ) Biedermann, Grob. Journal of Chromatography A, 1255 (2012) 56– 75 13


MOSH & MOAH - 2015 Collaborative study - Samples sample

14

matrix

Spiking

Mineral oil used

Level of spiking

1 & 13

refined sunflower oil 1

none

2 & 16

crude soybean oil

none

3 & 20

virgin olive oil

none

4 & 17


yes

MOSH & MOAH (75:25)

110 mg/kg

5 & 19


yes

MOSH

20 mg/kg

6 & 11


yes

MOSH & MOAH (70:30)

150 mg/kg

7 & 15

refined olive pomace oil

8 & 12

mayonnaise

yes

MOSH & MOAH (75:25)

150 mg/kg

9 & 18

palm oil

yes

MOSH & MOAH (75:25)

70 mg/kg

10 & 14

margarine


Epoxidation purification asked

ALOX purification asked

Epoxidation purification asked

none

none

Remarks

Epoxidation purification needed Epoxidation purification needed

MOAH - 2015 Collaborative study – impact of the epoxidation step For pomace olive oil samples with very long chain MOSH, part of the hump may be lost by incomplete elution

refined olive pomace oil

palm oil

Palm oil and olive oil samples need an epoxidation step for the MOAH determination 15


MOSH & MOAH - 2015 Collaborative study - Results - Some of the participants reported data as “< LOQ” for MOAH in non-spiked oil - It was decided to use the value of the LOQ as the result for the calculation of the precision data

Laboratories

Laboratories

Sunflower 3 is Sunflower 1 spiked only with MOSH  no difference in MOAH content expected Sunflower 1 : variations between 0 and 10 mg/kg for both MOSH or MOAH

Laboratories

16


Laboratories

MOSH & MOAH - 2015 Collaborative study - Results

Laboratories

Laboratories

When samples are spiked (> 10 mg/kg), the dispersion of the laboratories is less important

Laboratories

17


Laboratories

2015 Collaborative study – MOSH – Precision data sample 1 sample 2 sample 3

refined sunflower oil crude soybean oil virgin olive oil

samples 4-5-6 sample 7 sample 8

refined sunflower oil refined olive pomace oil mayonnaise

sample 9 sample 10

palm oil margaribe

Sample

1& 13

2& 16

3& 20

4& 17

5& 19

6& 11

7& 15

8& 12

9& 18

10 & 14

Number of laboratories

12

12

12

12

12

12

12

12

12

12

Number of laboratories retained after eliminating outliers

12

12

12

11

11

10

10

12

11

11

Number of outliers (laboratories)

0

0

0

1

1

2

2

0

1

1

MOSH mean value, mg/kg

4,0

190,3

6,0

80,9

21,4 101,3 196,6 101,9 53,9

5,0

Repeatability standard deviation, mg/kg

1,1

6,1

1,7

4,1

2,0

1,8

7,5

2,5

3,4

0,2

Repeatability relative standard deviation (RSDr), %

28%

3%

29%

5%

9%

2%

4%

2%

6%

4%

Repeatability limit r, mg/kg

3,1

17,2

4,9

11,6

5,5

5,2

21,1

6,9

9,6

0,5

Reproducibility standard deviation, mg/kg

3,0

26,0

3,0

9,7

3,9

13,7

30,9

8,8

12,2

3,5

Reproducibility relative standard deviation (RSDR), %

75%

14%

50%

12%

18%

14%

16%

9%

23%

70%

Reproducibility limit R, mg/kg

8,4

72,9

8,4

27,1

11,0

38,3

86,6

24,7

34,2

9,8

Application limit fixed at 10 mg/kg for MOSH with the method as described 18


2015 Collaborative study – MOAH – Precision data sample 1 sample 2 sample 3

refined sunflower oil crude soybean oil virgin olive oil

samples 4-5-6 sample 7 sample 8

refined sunflower oil refined olive pomace oil mayonnaise

sample 9 sample 10

palm oil margaribe

Sample

1& 13

2& 16

3& 20

4& 17

5& 19

6& 11

7& 15

8& 12

9& 18

10 & 14

Number of laboratories

12

12

12

12

12

12

12

12

12

10

Number of laboratories retained after eliminating outliers

11

12

11

10

11

12

11

12

11

9

Number of outliers (laboratories)

1

0

1

2

1

0

1

0

1

1

MOAH mean value, mg/kg

2,0

1,6

1,7

29,3

4,2

50,7

44,7

36,1

11,4

2,4

Repeatability standard deviation, mg/kg

0,7

0,5

0,5

1,6

0,3

5,6

2,5

3,2

1,0

1,3

Repeatability relative standard deviation (RSDr), %

36%

29%

33%

6%

7%

11%

6%

9%

9%

53,4%

Repeatability limit r, mg/kg

2,0

1,3

1,5

4,6

0,8

15,7

7,0

9,0

2,9

3,5

Reproducibility standard deviation, mg/kg

1,9

1,3

1,3

6,6

2,7

15,6

23,0

7,4

1,9

3,6

Reproducibility relative standard deviation (RSDR), %

94%

81%

79%

23%

64%

31%

51%

20%

17%

152%

Reproducibility limit R, mg/kg

5,2

3,7

3,7

18,5

7,5

43,6

64,3

20,7

5,4

10,0

Application limit fixed at 10 mg/kg for MOAH with the method as described 19


Origin of the contamination Plant protection

Seeds

Crop Storage Transport Crushing Meal


Diesel or lubricant leakage

Treatment of the seeds (anti-dusting)

Transport in contaminated containers Diesel or lubricant leakage Technical oils used in the oil mill

Crude oil Mineral oil

20

Environmental pollution (air, soil)

Multiple sources of mineral oil

EFSA Journal 2012;10(6):2704

The quality of the mineral oils, in terms of molecular mass range and MOAH content, differs for the sources.

21


Anti-dusting agent White mineral oil may be safely used in food in USA (21 CFR 172.878)  As a dust control agent for wheat, corn, soybean, barley, rice, rye, oats, and sorghum  Applied at a level of no more than 200 mg/kg of grain.

22


Plant protection Paraffinic white oils are used for plant protection for: - winter and spring insecticide treatment (pome fruit, stone fruit), - adjuvant for herbicide spray (beetroot, rapeseed) TRACE GC-Channel 1 Vegefix

TRACE GC-Channel 1 Oviphyt

450

450

400

400

350

350

300

300

Millivolts

Millivolts

Retention T ime Name

250

250

200

200

150

150

100

100

50

50

0

0 0

1

2

3

4

5

6

7

8

9 Minutes

23


10

11

12

13

14

15

16

17

18

Contamination during transport 2008: mineral oil in crude sunflower oil

1999: diesel in crude palm oil 17 000 µV C5 C7

16 000 15 000

ech ITERG1114.DATA N-Paraffines-1067.DATA

IFP analysis simulated distillation

14 000

Sample prepared by ITERG (saponification of 15 g + purification on silica gel)

13 000 C8

12 000 11 000 10 000

C 17

C 18

C 15

C 16

C 13

C 14

5 000

C 12

C 10

6 000

C 11

7 000

C9

8 000

C6

9 000

4

6

8

10

White MO included in the EU list of previous cargoes DGF Symposium, 10 March 2016

C 64

C 68

RT [min] 2

24

C 66

C 58

0

C 62

C 60

C 54

C 56

C 52

C 46

C 50

C 48

C 42

C 44

C 40

C 38

C 36

C 34

C 32

C 30

C 22

1 000

C 24

2 000

C 26

C 20

3 000

C 28

4 000

12

14

16

18

20

22

24

26

28

30

32

34

36

Contamination with mineral oil during crushing Hexane recovery plant: absorption tower where mineral oil absorbs residual hexane in air before being exhausted to the air

25


Contamination with thermal heating fluid 2009 : contamination of walnut oil with a food grade thermal heating fluid during refining process MOSH < 50 mg/kg 400

MOSH = 372 mg/kg 800

TRACE GC-Channel 1 6114

800

400

TRACE GC-Channel 1 613 B

Retention Time

Name Retention Time

Natural hydrocarbons

350

350

700

300

700

300

600

600

250

250

500

500

Millivolts

Millivolts

200 Millivolts

200

400

150

400

150

300

300

100

100

200

200

50

50

100

100

0

13.208

0

-50 1

2

3

4

5

6

7

8

9

10

11

12 Minutes

Pressed Walnut oil

26


13

14

15

16

0 17 0

0

-50 18

2

19

420

21 6

22

8

23

1024

12

14

16

18

Minutes

Contaminated Refined walnut oil

20

22

Migration from packaging 2014 : contaminated sample with a high boiling fraction of paraffin oil (POSH) coming from the low-density-polyethylene-bottles (food grade bottle) Extraction with hexane of a new bottle

Mineral oil used to spike the oil

27


Bottled oil

Effect of deodorization on mineral oil Wagner, Neukom, Grob, Moret, Populin, Conte, Mitt. Lebensm. Hyg. 92, 499–514 (2001)

Crude oil with 2 types of mineral paraffins: - one centered at n-C22 - one from C28 to beyond C40 Deodorization completely removed: - MOSH up to C23 - about 50% of C25 alkanes.

Step of the process

28

MOSH content (mg/kg)

Crude oil

50

Neutralized soap

40

After bleaching

55

After deodorization

14


Effect of deodorization on mineral oil Wagner, Neukom, Grob, Moret, Populin, Conte, Mitt. Lebensm. Hyg. 92, 499–514 (2001)

4 samples of deodistillates: - A : mineral paraffins (C16 - C33) centered at C23 - B : diesel or heating oil (C9-12) + mineral paraffins (C20 - C40) centered at C29 - C & D : mineral paraffins (C16 - C33) centered at C23 + PAO (C34;C41) Number of analysed samples of deodistillates

28

MOSH min (mg/kg)

120

MOSH max (mg/kg)

6 800

MOSH mean value (mg/kg)

1200 mg/kg

230 mg/kg

350 mg/kg

250 mg/kg

650

poly-alpha-olefins (PAO): synthetic base oils used for high grade motor oils or as food-compatible technical oils.

29


FEDIOL Recommendations FEDIOL code of practice for the management of mineral oil hydrocarbons presence in vegetable oils and fats intended for food uses (24 February 2016)  Prevention in vegetable oil and fat crushing and refining plants  Only lubricants suitable for incidental contact with food to be used  Mineral oil used as absorber in the hexane recovery system free from MOAH  Steam used for heating in processing installations

 Control of possible sources of MOH contamination in the supply chain  Transport in bulk of oils: thermal heating fluids not used in direct heating systems  Spraying of white mineral oil as anti-dusting: monitoring of MOH levels in soybean oils and verification of the absence of MOAH

 Monitoring scientific developments and next steps  More precise identification of the possible entry sources of MOH  Exploration of the MOH mitigation strategies through refining

30


Conclusion  Mineral oil hydrocarbons are present at different levels in nearly all foods, including in vegetable oils (EFSA)  MOAH fraction may be both mutagenic and carcinogenic (EFSA)  There is a potential concern associated to the background exposure to MOSH (EFSA)  no EU legislation regulating the limits of MOH in vegetable oils  Standardized method for MOSH published in 2015  Standardized method MOSH & MOAH soon available with application limits at 10 mg/kg  Mineral oil may be introduced at different stages of the vegetable oil and fat production

 Total removal of mineral oil during refining seems very difficult 31


Pour nous contacter :

ITERG 11 rue Monge Parc Industriel Bersol 2 33600 PESSAC Tél : 05.56.36.00.44 [email protected] www.iterg.com

Recherche . Innovation . Qualité

32

Mineral oil in vegetable oils : Laboratory and Industrial

Recommend Documents