fast, automatic, and accurate determination and

Fast, Automatic, and Accurate Determination and Identification of Targeted Analytes in High-Throughput

Analysis by Chromatography – Tandem Mass Spectrometry

Steven J. Lehotay and Yelena Sapozhnikova

Agricultural Research ServiceEastern Regional Research Center

Wyndmoor, Pennsylvania; USA

DisclaimerMention of brand or firm name does not constitute an endorsement

by the USDA above others of a similar nature not mentioned.

Contact: Steven. [email protected]

Acknowledgments

CTC Analytics

ITSP Solutions

Gerstel

Restek (J & J)

Jessie Matarrita Alan Lightfield

Robyn Moten Limei Yun

Tawana Simons Lijun Han

Agilent Hill Labs

Conclusions

• Aminoglycosides can be analyzed at the same time as other veterinary drugs by adding the ion-pairing reagent to combined final extracts from two sample preparation methods.

• Smaller test portions (2- 5 g) in QuEChERS are possible using the Blixer for many (but not all) commodities.

• High quality, rugged results can be achieved for hundreds of ultratrace analytes in diverse foods using automated high-throughput analysis by QuEChERS + ITSP+LPGC-MS/MS and UHPLC-MS/MS in parallel without matrix-matched calibration followed by summation function chromatographic peak integrations and post-run processing to yield accurate determinations and trustworthy identifications with little need for human review.

High Throughput Efficiency Start to Finish

1) Sample Processing (Blixer and/or Cryomill?)

2) Batch extraction by QuEChERS

3A) UHPLC-MS/MS 3B1) Automated clean-up (ITSP)

3B2) Fast, low-pressure (LP) GC-MS/MS

4) Trustworthy automatic peak integrations and identifications without human review

0 5 10 15 20 25 30

DataProcessing

Analysis

SamplePreparation

SampleProcessing

minutes/person/sample

Recent Past

Current

Near Future

Sample Throughput to Analyze Chemical Residues mini-QuEChERS + UHPLC- & ITSP+LPGC- MS(/MS)

UHPLC- and Fast GC-MS/MS

Automated in Parallel

Summation Function Chromatographic Peak Integration

This future is now!

• Lehotay et al. (2016) "Automated mini-column solid-phase extraction cleanup for high-throughput analysis of chemical contaminants in foods by low-pressure gas chromatography – tandem mass spectrometry" Chromatographia, 79, 1113-1130

•Han et al. (2016) “Method validation for 243 pesticides and environmental contaminants in meats and poultry by tandem mass spectrometry coupled to low-pressure gas chromatography and ultrahigh performance liquid chromatography” Food Control 66, 270-282

• Sapozhnikova and Lehotay (2015) “Review of recent developments and applications in low-pressure (vacuum outlet) gas chromatography” Anal. Chim. Acta 899, 13-22

• Lehotay et al. (2015) “Current issues involving screening and identification of chemical contaminants in foods by mass spectrometry” Trends Anal. Chem. 69, 62-75

•Lehotay and Cook (2015) “Sampling and sample processing in pesticide residue analysis” J. Agric. Food Chem. 63, 4395-4404

• Sapozhnikova and Lehotay (2015) “Evaluation of different parameters in the extraction of incurred pesticides and environmental contaminants in fish” J. Agric. Food Chem. 63, 5163-5168

Some Recent Publications of Note

Major Classes of Antibiotics

Aminoglycosides

Sulfonamides Tetracyclines-Lactams

Macrolides Quinolones

Penicillin G Sulfadimethoxine Tetracycline

Gentamicin C1 EnrofloxacinErythromycin

Currently, 219 vet. drugs (including >100 antibiotics) are on our list,but have targeted and evaluated ≈180 so far in (UHP)LC-MS/MS.

Dihydrostreptomycin

Streptomycin

Hygromycin

Spectinomycin hydrate

Spectinomycin

Neomycin

Gentamicin (C2+C2a)

Apramycin

Amikacin

Kanamycin

UHPLC w/ agent added

50 mM sodium 1-heptanesulfate in final extract

Aminoglycosides Multiclass, Multiresidues

2 g tissue + 20 mL of 10 mM NH4OAc, 0.4 mM EDTA,2% trichloroacetic acid, and 0.5% NaCl in water + IS

2 g tissue + 10 mL 4/1 (v/v) acetonitrile/water + IS

Shake 5 min on pulsed vortex platform shaker (80% setting, max pulsation)

Centrifuge 3 min at 3700 rcf

Centrifuge 3 min at 3700 rcfTransfer 10.75 mL (1 g equiv. sample) to 15 mL tube

Adjust pH to 6.5 ± 0.1 using a pH meter

Load extract in 3 portions onto 50 mg WCX DPX tips

Wash DPX tips with 5 mL water

Elute DPX tips with 1 mL 10% formic acid in water

Condition 50 mg WCX* DPX† tips with 3 mL eachof methanol and water

Tissue equivalence 0.174 g/mL

(no cleanup)

407 µL extract(71 mg sample equiv.)

71 µL extract(71 mg sample equiv.)

+ 272 µL 138 mM sodium 1-heptanesulfate ion-pairing (IP)reagent in water/acetonitrile

Yields 95 mg/mL final extract for each method in 34/66 (v/v) acetonitrile/watercontaining 50 mM IP reagent and 0.85% HO2CH 4 µL injection = 0.38 mg equiv. sample on column

*WCX = weak cation exchange sorbent†DPX = dispersive pipette extraction

Updated Veterinary Drug Residue Method

Table 1: Results for the veterinary drugs spiked at 0.5X, 1X, and 2X levels, n=10 each, in the bovine tissues; (tR = retention time); aminoglycosides in blue text.

Drug AnalytetR

(min)1X Level(ng/g)

Kidney Liver Muscle

13C6-Sulfamethazine 3.75 2002-Mercaptobenzimidazole 3.66 252-Mercapto-1-methylimidazole 1.95 200Quinoxyaline-2-caboxylic acid 3.82 1002-Thiouracil 0.96 400Abamectin (Avermectin B1a) 8.80 50Albendazole-2-amino sulfone 3.81 50Albendazole sulfoxide 4.13 50Albendazole 5.45 50Albendazole sulfone 4.57 50Amikacin 3.71 100Amoxacillin 3.50 50Ampicillin 3.89 20Apramycin 3.78 100Acetopromazine 5.09 10Azaperone 4.21 10Bacitracin 4.68 1000Beclomethasone 6.07 100Betamethasone 5.96 100Bithionol 8.09 10Bromchlorobuterol 4.29 10Brombuterol 4.35 10Cambendazole 4.55 10Chloramphenicol 4.72 50Carazolol 4.43 10Carbadox 3.74 30Carprofen 6.97 50Cefazolin 3.81 100Cephapirin 3.48 100Cimaterol 3.57 10Ciprofloxacin 3.96 50Clencyclohexerol 3.88 10Clenbuterol 4.22 10Clenbuterol-d9 4.20 200Clenpenterol 4.43 10Clindamycin 4.58 100Clorsulon 4.54 100Closantel 8.82 50Cloxacillin 6.20 10Chlorpromazine 5.58 10Cortisone 5.48 100Chlortetracycline 4.39 1000Danofloxacin 3.99 200Dapson 3.86 100DCCD 3.40 400Desacetyl-cephapirin 2.65 100Desethylene ciprofloxacin 3.86 100Diclofenac 7.10 200Dicloxacillin 6.53 100Difloxacin 4.17 50Dipyrone (metabolite) 3.64 200Dimetridazole 3.19 50Dimetridazole-hydroxy 2.73 50Doramectin 8.99 100Doxycycline 4.56 100Dihydrostreptomycin 3.66 500Emamectin B1a 7.14 50Enrofloxacin 4.03 100Eprinomectin 8.64 100Erythromycin A 5.20 100Fenbufen 6.46 50Fenbendazole 6.18 400Fenbendazole sulfone 5.17 400Fenoterol 3.67 50Florfenicol 4.31 300Florfenicol Amine 3.09 300Flubendazole 5.68 10Flubendazole-2-amino 4.43 10Flumethasone 5.85 100Flumequine 5.62 300Flunixin 6.69 25Flunixin-d3 6.69 200Gamithromycin 4.56 100Gentamicin C1 3.80 300Gentamicin C1a 3.81 300Gentamicin C2+C2a 3.81 300Haloperidol 4.96 10Haloxon 6.65 100Hygromycin 3.64 100Indoprofen 5.94 50Ipronidazole 4.58 10Ipronidazole-hydroxy 3.95 10Ivermectin 9.25 50Josamycin 5.82 100Kanamycin 3.72 100Ketoprofen 6.28 50Lasalosid A 9.65 100Levamisole 3.83 100

Drug AnalytetR

(min)1X Level(ng/g)

Kidney Liver Muscle

Lincomycin 3.78 100Mabuterol 4.42 10Marbofloxacin 3.85 100Mebendazole 5.47 10Mebendazole-2-amino 4.32 10Meclofenamic acid 7.53 200Meloxicam 6.42 1006-Methyl-2-thiouracil 1.36 400Melengesterol acetate 7.57 25Morantel 4.22 100Moxidectin 8.93 100Metronidazole 2.83 10Metronidazole-hydroxy 2.47 10Nafcillin 6.39 100Nalidixic acid 5.48 200Naproxen 6.35 100Neomycin 3.84 1000Niclosamide 7.76 10Niflumic acid 7.15 200Nitroxynil 5.75 50Norfloxacin 3.91 50Novobiocin 7.78 1000Oxyphenylbutazone 6.18 100Orbifloxacin 4.10 50Oxytetracycline 3.96 1000Oxacillin 5.98 100Oxbendazole 4.63 10Oxyclozanide 7.46 10Oxfendazole 4.70 800Phenylbutazone 7.05 100Phenylbutazone-d10 7.02 200Penicillin G 5.47 50Penicillin G d7 5.43 2006-Phenyl-2-thiouracil 4.23 400Pirlimycin 4.48 300Piroxicam 5.77 100Propionylpromazine 5.48 10Prednisone 5.38 100Prednisolone 5.51 100Promazine 5.06 10Procaterol 3.58 100Propyphenazone 5.80 1006-Propyl-2-thiouracil 3.53 50Pyrantel 3.97 100Ractopamine 3.98 30Ractopamine-d3 3.96 200Rafoxanide 9.11 10Ritodrine 3.76 10Ronidazole 2.96 10Salbutamol 3.51 10Sarafloxacin 4.18 50Sulfabromomethazine 5.54 100Sulfachloropyridazine 4.09 100Sulfadiazine 3.02 100Sulfadimethoxine 4.79 100Sulfadoxine 4.26 100Selamectin 9.20 200Sulfaethoxypyridazine 4.42 100Sulfisoxazole 4.35 100Sulfamethizole 3.72 100Sulfamethoxypyridazine 3.79 100Sulfamerazine 3.42 100Sulfamethoxazole 4.19 100Sulfamethazine 3.76 100Sulfanilamide 1.42 100Sulfanitran 5.49 100Spectinomycin 3.52 100Sulfapyridine 3.34 100Sulfaquinoxaline 4.85 100Streptomycin 3.65 500Sulfathiazole 3.20 100Thiabendazole 3.87 1005-Hydroxythiabendazole 3.71 100Tetracycline 4.03 1000Triclabendazole 7.51 50Triclabendazole sulfoxide 7.15 50Triflupromazine 5.79 10Tildipirosin 3.90 500Tilmicosin 4.64 100Tiamulin 5.31 600Tobramycin 3.78 500Tolfenamic acid 7.73 200Tulathromycin 4.11 1000Tylosin 5.34 200Virginiamycin M1 6.28 100Xylazine 4.22 10Zeranol 5.99 100Zilpaterol 3.51 12

Gold = 80-110% Recovery, ≤15% RSD Silver = 70-120% Recovery, ≤25% RSD Bronze = 50-150% Recovery, ≤40% RSD

Red = <50 or >150% Recovery or >40% RSD

-100%

-80%

-60%

-40%

-20%

0%

20%

40%

60%

80%

100%

120%

140%

0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0

Mat

rix

Effe

ct

Retention Time (min)

Matrix Effects in Different Bovine Tissue Extracts

Kidney

Liver

Muscle

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

110%

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 110% 120% 130% 140% 150% 160%

RSD

Recovery

Recoveries and RSDs in Bovine Kidney

2-mercapto-1-

methylimidazole

abamectincephapirin

closantel

dipyronemetabolite

dora-mectin

ivermectin off scale %Recovery (%RSD)desacetyl cephapirin 300 (8)moxidectin 9 (206)6-methyl-2-thiouracil 31 (354)clorsulon NDdimetridazole-hydroxy NDmetridazole-hydroxy ND

eprinomectin

rafoxanide

ronidazole

tylosin

tobramycin

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

110%

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 110% 120% 130% 140% 150% 160%

RSD

Recovery

Recoveries and RSDs in Bovine Liver

abamectin

cephapirin

closantel

doramectin

off-scale %Recovery (%RSD)desacetyl cephapirin 238 (11)fenbendazole 170 (46)moxidectin 28 (167)6-methyl-2-thiouracil NDivermectin NDclencyclohexerol NDclorsulon ND

eprinomectin

rafoxanide

ractopamine

haloxon

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

110%

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 110% 120% 130% 140% 150% 160%

RSD

Recovery

Recoveries and RSDs in Bovine Muscle

streptomycin

phenylbutazone

ivermectin

off scale %Recovery (%RSD)6-methyl-2-thiouracil NDclorsulon ND2-mercapto-1-methylimidazoleapramycin NDflorfenicol amine NDspectinomyin NDzilpaterol ND

sulfamerazine

chloramphenicoltobramycin

selamectin

oxyphenylbutazone

neomycin

kanamycin

hygromycin

gentamicins

dihydrostreptomycin

amikacin

84% of analyteswithin the box



-100%

-80%

-60%

-40%

-20%

0%

20%

40%

60%

80%

100%

120%

140%

0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0

Mat

rix

Effe

ct


Matrix Effects in Different Bovine Tissue Extracts

Kidney

Liver

Muscle

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

110%

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 110% 120% 130% 140% 150% 160%

RSD

Recovery

Recoveries and RSDs in Bovine Kidney

2-mercapto-1-

methylimidazole

abamectincephapirin

closantel

dipyronemetabolite

dora-mectin

ivermectin off scale %Recovery (%RSD)desacetyl cephapirin 300 (8)moxidectin 9 (206)6-methyl-2-thiouracil 31 (354)clorsulon NDdimetridazole-hydroxy NDmetridazole-hydroxy ND

eprinomectin

rafoxanide

ronidazole

tylosin

tobramycin

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

110%

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 110% 120% 130% 140% 150% 160%

RSD

Recovery

Recoveries and RSDs in Bovine Liver

abamectin

cephapirin

closantel

doramectin

off-scale %Recovery (%RSD)desacetyl cephapirin 238 (11)fenbendazole 170 (46)moxidectin 28 (167)6-methyl-2-thiouracil NDivermectin NDclencyclohexerol NDclorsulon ND

eprinomectin

rafoxanide

ractopamine

haloxon

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

110%

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 110% 120% 130% 140% 150% 160%

RSD

Recovery

Recoveries and RSDs in Bovine Muscle

streptomycin

phenylbutazone

ivermectin

off scale %Recovery (%RSD)6-methyl-2-thiouracil NDclorsulon ND2-mercapto-1-methylimidazoleapramycin NDflorfenicol amine NDspectinomyin NDzilpaterol ND

sulfamerazine

chloramphenicoltobramycin

selamectin

oxyphenylbutazone

neomycin

kanamycin

hygromycin

gentamicins

dihydrostreptomycin

amikacin




Validation Results

works for eggs and fish, too (not milk)

Multi-Application, Multiresidue Analysis

Goal: Develop a multi-class, multi-residue method for analysis of LC- and GC- amenable pesticides as well as legacy and emerging environmental contaminants in food:

Pesticides

Polychlorinated biphenyls (PCBs), including dioxin-like PCB congeners

Polycyclic aromatic hydrocarbons (PAHs)

Polybrominated diphenyl ethers (PBDEs)

Novel alternate flame retardants (FRs)

150 pesticides + 69 environmental &

other contaminants

99 pesticides & internal/ QC standards

LPGC-MS/MS

UHPLC-MS/MS

55 overlapping pesticides

>240 Analytes in Parallel using 10 min Analyses

See Han et al. (2016) Food Control 66, 270-282

Fast Low-Pressure (LP)GC-MS/MS

MS(/MS)

GC Oven

Injector

Restriction Capillary

Mega-BoreColumn

5 m

0.18mm

15 m 0.53 mm 1 m xx-5ms

No special adaptations needed; can be implemented in any GC-MS(/MS).

Review of dozens of publications using LPGC-MS(/MS):Sapozhnikova and Lehotay, Anal. Chim. Acta 899, (2015) 13-22

LPGC-MS is Much Faster

600 800 1000 1200 1400 1600 1800

2e+006

4e+006

6e+006

8e+006

1e+007

Time (s)

100 150 200 250 300 350 400 450 500

2.5e+006

5e+006

7.5e+006

1e+007

1.25e+007

1.5e+007

1.75e+007

2e+007

2.25e+007

Time (s)

25 min

LPGC-MS

Traditional GC-MS

and more sensitiveIs fast micro-bore with back-flush better? Watch out for carry-over.

1. New GC-QQQ provided high-speed, sensitive, selective, and

precise detection of nearly all targeted analytes at <10 ng/g

LOQ/LOI using 9:1 SPLIT INJECTION. Merely 100 µg equivalent

sample introduced! BEST PERFORMANCE, LONG-TERM!

2. LPGC-MS(/MS) is a faster and more rugged technique shown

to give taller peaks with less tailing (and lower LOQs) than

conventional GC-MS(/MS). MEGA-BORE = MEGA-RUGGED!

3. The micro-bore restrictor is inserted ≈1 cm into the mega-bore

analytical column, achieving zero dead volume, and metal

ferrules are used for a leak-free system. COMMERCIALLY

AVAILABLE AS CUSTOM ORDER AT NORMAL COLUMN COST.

4. Sandwich injection (air gap) using 5 µL syringe to ensure full

sample introduction and avoid analyte discrimination.

Aspects in the LPGC-MS/MS Analysis

5. A 1 m integrated guard column goes into the hot transfer line,

which avoids damage to the stationary phase from H2O/O2 in

every extract, even when the oven is cool. This improves long-

term chromatographic performance and reduces column

bleed.

6. An insert reduced oven volume, and 220 V rapid heating

option installed. Oven program of 50°C/min ensured that

the temperature was within consistent control.

7. Ion source and final oven temperature were 320°C (4.6 min)

to reduce ghost peaks and keep system clean. Post-sequence

oven at 250°C help, too.

8. MS/MS dwell times set to >7 points across analyte peaks

yielded greater analytical accuracy than using fewer points.


9. Final extracts contained 0.1% HCO2H to improve stability of

base-sensitive pesticides, such as chlorothalonil.

10. Sample preparation used automated mini-cartridge SPE to

provide better cleanup than dispersive-SPE.

11. Analyte protectants were added to all extracts to coat active

sites in the liner, columns, transfer line, and ion source, even

if matrix material has built up from previous injections.


HO O

OH

O O

OH

ethylglycerol

1 mg/ mL

ethylglycerol

1 mg/ mL

O

HO

HO

O

OHHO

O

HO

HO

O

OHHO

gulonolactone

0.1 mg/ mL

gulonolactone

0.1 mg/ mL

HO OH

OH OH

OH

HO OH

OH OH

OH

sorbitol

0.1 mg/ mL

sorbitol

0.1 mg/ mL

shikimic acid

0.05 mg/ mL

shikimic acid

0.05 mg/ mL

HO

HO

HO

OH

O

2.5 0.25 0.25 0.125

Effect of Analyte Protectants

w/ analyte protectants

w/o analyte protectants

NOTE: BE SURE TO WASH SYRINGE WELL WITH H2O IN SOLVENT MIX

See: Mastovska et al., Anal. Chem. 77 (2005) 8129-8137

Slope = 5.3% ME/tR

Slope = -0.5% ME/tR

-75%

-60%

-45%

-30%

-15%

0%

15%

30%

45%

60%

75%

90%

105%

2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0

Avg

Mat

rix

Effe

ct (

ME)


Avg MEs among the 4 Matrices for the Analytes

ME vs. QC Std

ME vs. Int. Std

Oh, and we may have eliminated matrix effects in GC-MS …

… via use of appropriate int. stds + analyte protectants in split inj’n

-0.04

-0.03

-0.02

-0.01

0.00

0.01

0.02

0.03

0.04

2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00

Dif

f. in

Ret

. Tim

e f

rom

Rgt

Day

1 (

min

)


Cattle Day 1 Chicken Day 2 Pork Day 3

Rgt Day 2 Rgt Day 3

All of which leads to very consistent peaks and tR!3-Day Validation Experiment of 202 Pesticides analyzed by LPGC-MS/MS

40 matrix (muscle) spks and blks + QC = 70 injections per dayAvg tR (min) of reagent stds and matrices throughout the run (SD <0.040)

Analyte Protectants Added to all Final Extracts

phenanthrene

azinphos

dibenz(ah)-anthracene

Don’t Trust the “Advanced” Software

And don’t trust the analyst, either. This mistake was caught after preparing the previous slide for this presentation.

× ×

Int. Std.

Quant.Ion

Qual.Ion

RESOLVED: Human review takes too long!

High-throughput (or even low-throughput) multi-analyte monitoring applications:

G.F. Pang et al. (Beijing, China) include 1,138 pesticides in their GC- and LC- MS/MS monitoring approach. Large team of chemists conduct analyses and review results.

USDA: 240 analytes × 2-4 ion transitions × 50 samples/batch = 36,000 peaks!Analyst review and re-integration at 1 s per peak = 10 hours w/o breaks

on each instrument!

Summation Integration in Chromatography

SIMPLIFY, don’t COMPLIFY!

• Draw a straight line at the baseline just before the start of the expected peak to just after its expected end EASY PEASY!

• e.g. Elkin et al. “Computer-controlled mass fragmentographywith digital signal processing” J. Chromatogr. 81 (1973) 47-55

• Advanced ≠ Better

• Function ≠ Beauty

• Time = Money

2 ng/g Pyriproxyfen in Orange

LOQ/LOI Qualitative(ng/g) Result

Height 0.9/0.9 IdentifiedArea 1.4/1.8 False Negative

Qual. Ionm/z 198 102

Quant. Ionm/z 198 129

tR = 5.6 min

stopstart

Summation integration is consistent and reliable

Ion 2Ion 1 Ion 3

Traditional Integration

Rep A

Rep B

Pain to set many integration parameters that still don’t work!

Summation Integration

Rep A

Rep B

Overall LPGC-MS/MS ResultsOut of 195 analytes and 73 injections in 6 matrices = 14,235 decisions

“Advanced” SummationFalse Pos. 0.19% 0.11%False Neg. 11.2% 9.5%

True 91.6% 92.9%

192 times net overall times that summation did not yield a false result vs. “advanced” = 1.3% improvement

5 ng/mL endosulfan sulfate in reagent-onlyand matrix-matched calibration standards

LOQ ≈2 ng/mL in all matrices; even after 325 injections, including 230 food extracts

p,p’-DDD and o,p’-DDT partially co-elute but can be consistently integrated individually

Pear Cilantro

10ng/g

Spikes

100ng/g

Spikes

Original QuEChERS Acetate-Buffered

1ng/g

Spikes


after ≈90injections



p,p’-DDD

o,p’-DDT

Orange Tilapia

10ng/g

Spikes

100ng/g

Spikes


1ng/g

Spikes





Continued:

Chrysene 1 µL 9:1 Split Injection

1 ng/gStds

MeCN Orange Tilapia

partial co-elution with benz(a)anthracene – summation integration at mid-point

● MeCN stds w/APs (○ outlier)

▪ Chrysene-d12 (IS)

Orange MM stds w/APsTilapia MM stds w/APs

second sequence of 107 injections(214 total with same liner)

184 of matrix extracts

R2 = 0.9996

Issue of S/N and Peak Height vs. Area

Sample (signal height = 0.25) S/N = 1.25 (< LOD)

Blank Sample (rms noise = 0.2) S/N = 0

However, if 16 injections of blanks yield peak area of 0 ± 1 cps

Then, sample peak area of 10 cps S/N = 10 ( LOQ)

Methamidophos (m/z 141 95 at 2.29 min)

Blk 2 ng/mL Blk 2 ng/g

340 apple

180 green bean

290 squash

320 peach

350 banana

360 broccoli

240 potato

480 celery

grape175

Calibration Stds in MeCN Calibration Stds in Matrix

scale

430 orange

day 1

day 2

day 4

day 5

day 6

day 7

day 8

day 9

day 10

day 3

Issue of S/N and Peak Height vs. Area

• Made 490 injections for 70 pesticides in 10 matrices over 10 days.

• Compared LOQ calculations using summation integrated peak heights vs. peak areas.

• Note: instrument software used lowest signal point in integration window as the baseline to ensure all results were positive.

Peak Height Results Led to Lower LOQs and LOIs

2 ng/g Pyriproxyfen in Orange (tR = 5.6 min)Quant. Ion = m/z 198 129Qual. Ion = m/z 198 102

Peak Height Peak AreaQuant. Qual. Quant. Qual.

Signal 722 483 924 698 Noise 50 18 129 35 S/N 14 27 7 20 Slope 553 357 716 461 LOQ/I 0.9 0.9 1.8 1.4

Identified False Negative

• LOQ = limit of quantification (for quant. ion) = 10(SD)/slope• LOI = limit of identification (for qual. Ion) = 10(SD)/slope,

where SD = noise = std dev of 32 blank matrix injections

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

9.0

10.0

Ave

rage

LO

Q (

ng

/g)

MeCN (height) MeCN (area) Matrix (height) Matrix (area)

Avg LOQs in the 10 Commodities over 10 Days

0%

5%10%

15%20%

25%30%

35%40%

45%50%

55%60%

65%70%

75%80%

< 1 1 -10 1 - 25 25 - 100 > 100

Pe

rce

nt

of

Tim

es

Re

sult

in C

on

c. R

ange

Concentration (ng/g)

LOQ MeCN

LOQ Matrix

LOI MeCN

LOI Matrix

Height Area

LOQs and LOIs among ≈700 Calibrations

Rules in Automatic Post-Run Identification(e.g. in Excel or Instrument Software)

1) Ret. time (tR) for each ion (Quant. and Qual.) must be ≤|0.1| min from the contemporaneous tR(ref.)

2) Ion Ratio (IR) = (peak area ion 2)/(peak area ion 1), 3/1, 4/1, etc. (in %); IR(ref.) and tR(ref.) = avg. of contemporaneous high conc. calibration stds in solvent [note: IR(ref.) ≤ 110%]

IR must be |±10| for ≥1 ion or |±20| for ≥2 ions vs. IR(ref.)

3) Conc. must be > reporting level

Conc.(ng/g)

Peak Heights Peak AreasMeCN Matrix MeCN Matrix

2 53% 53% 58% 57%8 16% 18% 20% 24%

32 2% 4% 4% 6%128 1% 10% 10% 17%512 0% 0% 0% 1%

False Negatives Among the ≈700 Calibrations

The true positive identification rates for the calibration stdsat each level are 100% minus the false negatives shown.

Instrument Top Sample Preparation (ITSP)Morris and Schriner (2015) “Development of an automated column solid-phase extraction cleanup of QuEChERS extracts, using a zirconia-based sorbent, for pesticide residue analyses by LC-MS/MS” J. Agric. Food Chem. 63, 5107-5119

www.nacrw.org/2014/presentations/O21-Morris.pdf

Determined performance results in the use of automated mini-SPE cleanup in the LPGC-MS/MS analysis of pesticides and other contaminants in QuEChERS extracts of 10 different matrices.

Used mini-cartridgesshowing removal of chlorophyll and othermatrix components

Robotic liquid handler:3 min cleanup step at 2 µL/s+ 5 min for addition of APs and switching/washing syringes

Final extract volumes = 278 ± 5 µL (n = 255) after 50 µL addition of APs (and/MeCN) solution

See Lehotay et al. (2016) Chromatographia, 79, 1113-1130

ITSP+LPGC-MS/MS takes 13 min per injection cycle

FDA Sampling for Pesticides

• < 25 g units (berries) 1 kg (2.2 lbs)

• 25 – 250 g (apples) 1 kg (≥ 10 units)

• > 250 g (cabbage) 2 kg (≥ 5 units)

• Grains, Tree Nuts 1 kg

• Herbs 0.5 kg

• Spices 0.1 kg

Slide adapted from Jo Marie Cook

CODEX: 1 kg (2.2 lbs)

Pesticide Data Program: 3–5 lbs fresh, 2 lbs processed

USDA-FSIS: 1 lb meat, poultry, fish

Cryogenic Sample Processing

fried bacon

Spex FreezerMill(Cryomill)

Comminuted BroccoliBlixer Blixer + Cryomill

250X magnification

Robot Coupe Blixer has a spatula in the lid to ease and improve comminution

Conclusion: Cryomill = Overkill

250X magnification

Incurred Residues in Peach

0.0

2.0

4.0

6.0

8.0

10.0

1 g 2 g 5 g 10 g 15 g

Co

nce

ntr

atio

n (

ng

/g)

tetrahydrophthalimideBlixer

Cryomill

0

20

40

60

80

100

1 g 2 g 5 g 10 g 15 g

Co

nce

ntr

atio

n (

ng/

g)

fludioxynil Blixer

Cryomill

0

2

4

6

8

10

12

14

16

18

20

1 g 2 g 5 g 10 g 15 g

Co

nce

ntr

ati

on

(n

g/g

)

propiconazoleBlixer

Cryomill

0

2

4

6

8

10

12

14

16

18

20

1 g 2 g 5 g 10 g 15 g

Co

nce

ntr

atio

n (

ng/

g)

azinphos-methylBlixer

Cryomill

Error Contribution due to Sample Processingvs. Test Portion (Subsample Size)

0%

5%

10%

15%

20%

0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0

RSD

fro

m S

amp

le P

roce

ssin

g

Test Portion Size (g)

Calculated from all residues in 10 commodities processed with Blixer and Cryomill

Conclusions

• Aminoglycosides can be analyzed at the same time as other veterinary drugs by adding the ion-pairing reagent to combined final extracts from two sample preparation methods.

• Smaller test portions (2- 5 g) in QuEChERS are possible using the Blixer for many (but not all) commodities.

• High quality, rugged results can be achieved for hundreds of ultratrace analytes in diverse foods using automated high-throughput analysis by QuEChERS + ITSP+LPGC-MS/MS and UHPLC-MS/MS in parallel without matrix-matched calibration followed by summation function chromatographic peak integrations and post-run processing to yield accurate determinations and trustworthy identifications with little need for human review.

+

+ !

Contact: Steven. [email protected]

Sampling and Sample Processing

• For particulate materials

• Finite Elements

• Infinite Elements & Increments

• Compositional Heterogeneity and Fundamental Error

• Distributional Heterogeneity

• Sample Correctness and Tools

Slide adapted from Jo Marie Cook

Sample Processing w/ Dry Ice

Experimental

Sample preparation (final method): 5-8

10 g homogenized fish + internal standards

Add 10 mL MeCN and shake 10 min on vortex shaker

at 80% setting with max. pulsing

Add 5 g HCO2NH4, shake 1 min,centrifuge 2 min at 3700 rcf

Filter-vial dispersive-SPE:•Add 0.5 mL extract to the PVDF (0.2 µm) filter-vial shell containing 75 mg each anh. MgSO4 + 1/1/1 PSA/C18/Z-Sep• Partially depress the filter-vial plunger and shake for 30 s in an autosampler vial tray• Fully depress the plunger into the filter-vial shell

Sample Preparation

We have validated the approach for analysis of beef, pork, and chicken muscle for implementation by USDA-FSIS

Extraction of Incurred Fish Samples

RESOLVED: Good analysts are able to conduct peak

integrations better than current advanced software tools (but good analysts are hard to find, earn wages, get bored reviewing data, and still make misteaks).

RESOLVED: Despite technology and software advancements,

no set of peak integration parameters works consistently for all analytes, concentrations, and matrices in the real-world (at least not yet in my experience).

RESOLVED: Garbage In = Garbage Out• Correct and consistent chromatographic peak integration is

essential to achieving high quality final results.

What the Heck?

× ?Quant.

IonQual.Ion

100 ng/gAzinphos

Analytical chemists are hard-working lazy people who look for short-cuts rather than make manual re-integrations.

Real samples can’t be dropped as outliers.

Integrations of replicate injections were correct 2 out of 3 times

In analytical chemistry, 2 out of 3 ain’t good

In chromatography, the primary parameters are ret. time (tR) and peak shape (width, height/area)

If tR and peak widths are so important and consistent in good methods, why do most (all?) sophisticated (and expensive)

chromatographic peak integration software programs so often choose peaks at the wrong tR with quite variable peak shapes?

Retention Times and Peak Widthsare Rock Solid in UHPLC-MS/MS

# AnalyteDay 1 = 7/17/15 Day 2 = 7/22/15 Day 3 = 7/28/15

Rgt Cattle Rgt Chicken Rgt Pork

1 Methamidophos 0.965 0.963 0.950 0.955 0.962 0.963

8 Oxamyl 1.977 1.970 1.950 1.950 1.962 1.963

18 Dimethoate 3.055 3.055 3.030 3.030 3.045 3.045

28 Oxadixyl 4.030 4.030 4.012 4.010 4.020 4.022

42 Metalaxyl 5.023 5.017 5.000 5.000 5.008 5.010

67 Azinphos 6.067 6.065 6.048 6.045 6.052 6.057

90 Profenophos 7.023 7.018 7.003 7.007 7.017 7.018

100 Methoprene 8.013 8.013 8.000 8.005 8.010 8.010

3-Day Validation Experiment of 101 Pesticides analyzed by UHPLC-MS/MS40 matrix (muscle) spks and blks + QC = 65 injections per dayAvg tR (min) of reagent stds and matrices throughout the run (SD <0.020)

New mobile phase added for each sequence

Summation Integration Function• ≈1 min to integrate a batch of >60 samples of

≈660 MRMs per sample WITHOUT REVIEW!

• This is a >40 year-old integration function, but LACKING IN SOME DATA PROCESSING SYSTEMS!

% of times results in range

Peak Height Results Peak Area Results

LOQ (ng/g) LOI (ng/g) LOQ (ng/g) LOI (ng/g)

MeCN Matrix MeCN Matrix MeCN Matrix MeCN Matrix

< 1 ng/g 38% 20% 30% 8% 27% 11% 21% 3%

1 - 10 ng/g 54% 70% 56% 59% 62% 78% 60% 55%

10 - 25 ng/g 5% 6% 9% 17% 7% 7% 12% 21%

25 - 100 ng/g 3% 4% 3% 15% 4% 4% 6% 17%

> 100 ng/g 0% 0% 2% 1% 0% 0% 1% 3%

Table III. Average limits of quantification and identification (LOQs and LOIs) ± standard deviations for the pesticide analytes in the 10 commodities analyzed by LPGC-MS/MS on 10 different days using peak heights or areas in summation integration. Bold text indicates when LOI < LOQ and italics signify the lower average between the paired result based on peak height or area.

IntegrationFunction

RgtR2

PearR2

TomatoR2

CucumberR2

EggplantR2

LemonR2

“Advanced”0.981 ±0.005

0.995 ±0.008

0.994 ±0.007

0.994 ±0.012

0.996 ±0.009

0.989 ±0.018

Summation0.980 ±0.005

0.997 ±0.002

0.995 ±0.004

0.997 ±0.002

0.998 ±0.002

0.995 ±0.005

IntegrationFunction

Rgt%True

Pear%True

Tomato%True

Cucumber%True

Eggplant%True

Lemon%True

Overall%True

“Advanced” 100 ± 0 100 ± 0 99 ± 2 96 ± 8 97 ± 8 97 ± 8 98 ± 3

Summation 100 ± 0 100 ± 2 100 ± 2 99 ± 4 98 ± 6 97 ± 8 99 ± 3

Fast UHPLC-MS/MS Results for 21 Pesticides

Qualitative identifications (1,638 “yes/no” decisions overall)

No diff. in all but 10 pest/matrix pairs; summation better in 8 cases

ITSP of QuEChERS Salmon ExtractsGC-MS Full Scan m/z 100-500, 1 µL injection including APs

Conclusion: Ok cleanup, and 200-300 µL is better

No Cleanup

MeCN

Salmon

200 µL

300 µL

500 µL

600 µL

400 µL

Extract VolumeCartridge = 45 mg of

20/12/12/1 mg each of MgSO4/PSA/C18/CarbonX

ITSP of QuEChERS Pork ExtractsGC-MS Full Scan m/z 100-500, 1 µL injection including APs

Conclusion: Good cleanup, and 200-600 µL is similar

MeCN

200 µL

300 µL

500 µL

600 µL

400 µL

Extract Volume

Pork No Cleanup

ITSP of QuEChERS Kale ExtractsGC-MS Full Scan m/z 100-500, 1 µL injection including APs

Conclusion: Ok cleanup, and 200-600 µL is similar

No Cleanup

MeCN

200 µL

300 µL

500 µL

600 µL

400 µL

Extract Volume

Kale

45 mg MgSO4+C18+PSA+CarbonX ITSP of QuEChERS Kale ExtractsUV/Vis Absorbance Results

Spectrum of Extract before Cleanup

Ab

sorb

ance

Ab

sorb

ance

Spectrum of 300 µL Cleaned Up Extract

Wavelength (nm)

Chlorophyllsand Xanthophylls

600 µL 500 µL 400 µL 200 µL None300 µL

45 mg MgSO4+C18+PSA+CarbonX ITSP of QuEChERS Kale ExtractsChlorophyll Removal and HCB Results vs. Extract Vol. Added

Conclusion: 300 µL extract needed for 80% HCB elution

50%

60%

70%

80%

90%

100%

50%

60%

70%

80%

90%

100%

150 200 250 300 350 400 450 500 550 600 650

%R

em

ove

d b

ased

on

Ab

sorb

ance

at

68

0 n

m

HC

B P

eak

Are

a vs

. Max

Added Extract Volume (µL)

ITSP Cleanup vs. Recovery of HCB in Kale

% vs. Max HCB

Chlorophyll removal

ITSP of QuEChERS ExtractsRecovery vs. Extract Vol. Added

Conclusion: most analytes were not retained by the sorbents

50%

60%

70%

80%

90%

100%

110%

120%

150 200 250 300 350 400 450 500 550 600 650

Pea

k ar

ea v

s. t

hat

fo

r 6

00

µL

Added Extract Volume (µL)

Rel. Responses of bifenthrin in ITSP

Kale Salmon

Pork Avocado

Injection liner and septum after 325 injections in 5 daysincluding 230 matrix extracts (1 mg equiv.) of 10 diverse food commodities

A little “dirt” here and there, but the analyte protectantsdid their job and results still looked great from start to finish.

fast, automatic, and accurate determination and

Documents