Detection and quantification of trace airborne transfluthrin concentrations via air sampling and thermal desorption gas chromatography-mass spectrometry

This efficient method allows for the remote collection of samples and rapid analysis of airborne transfluthrin from industrial applications, optimization studies of commercial products as well as domestic/household monitoring.

j ou rn a l h om ep a ge :w w w e l s e v i e r c o m / l o c a t e / c h r o m a

Short communication

Michael W.C Kwana, Jason P Weisenseelb, Nicholas Giela, Alexander Bosaka,

Christopher D Batichc,d, Bradley J Willenberga,∗

a r t i c l e i n f o

Keywords:

GC–MS

Transfluthrin

Pyrethroids

a b s t r a c t

Arapidthermaldesorption-gaschromatography-electronionization-massspectrometry(TD-GC-EI-MS) methodforairbornetransfluthrindetectionisstudied.Activeairsamplingof9Lover1hat23◦Cthrougha Tenax®-loadedtuberesultedinefficientcaptureofairbornetransfluthrin.Subsequentthermaldesorption wasemployedtoachieveanLODof2.6ppqv(partsperquadrillionbyvolume).Aminimumprimary desorptiontemperatureof300◦CisnecessaryforoptimalrecoveryofsamplefromtheTenax®adsorbent Thematrixeffectsofindoorairleadtoanerrorof10.9%and10.5%recoveryofsample(10pgand100pg loadedtubes,respectively).Thelinearrangewas74–74,000ppqvwithacorrelationcoefficientof0.9981 Activeairsamplingofanovelpassivereleasedevicerevealeda∼150pg/Lairborneconcentrationgradient over1m,providingspatialcharacterizationofthedevice’sperformance.Thisefficientmethodallowsfor theremotecollectionofsamplesandrapidanalysisofairbornetransfluthrinfromindustrialapplications, optimizationstudiesofcommercialproductsaswellasdomestic/householdmonitoring

©2018TheAuthor(s).PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBYlicense

(http://creativecommons.org/licenses/by/4.0/)

1 Introduction

Transfluthrin

((1R,3S)-3-(2,2-Dichlorovinyl)-2,2-dimethyl-1-cyclopropanecarboxylic acid (2,3,5,6-tetrafluorophenyl)methyl

ester)isasemi-volatileorganiccompoundofthepyrethroidclass

of insecticides that is routinely used as an indoor insecticide

Transfluthrin (TF) works as a potent dipteran sodium channel

agonist and can elicit effects such as repellency, restlessness,

knockdownanddeathin mosquitoes[1 Commercial useof TF

along with the inevitable exposure to humans has generated

interestinquantifyinglowlevelsofairborneTFwiththeultimate

goal of understanding what constitutes an effective airborne

concentrationagainst mosquitoes withoutplacing undue harm

onpeople[2–4] Previously,ourgrouphasdevelopeda passive

release devicethat releases airborneTF at a constant rateinto

the air over several hundred hours [5, unpublisheddata] The

precisequantificationofairborneTFemanatingfromthedevice

willprovideinsightintodeviceperformanceandmayhelpguide optimization Furthermore,the lower limit of efficacyfor each

of TF’s effects towards mosquitoes (i.e confusion, excitation, knockdownand death) arecurrently unknownwitha previous studysuggestingthatTFcaneffectivelyknockdownmosquitosat onepartspertrillionbyvolume(1pptv)[6 Properdissemination

ofTFisparamountasoverandunderexposureofinsecticideshave beenlinkedtospurringresistancedevelopment[7–10]

Gas chromatography-electron ionization-mass spectrometry (GC-EI-MS)isapowerfultoolinseparating,analyzingand quan-tifyingtheconstituentsofacomplexmixtureofvolatiles.Previous methods for monitoring TF airborne concentrations via

GC-EI-MSincludeairsamplingfollowedbyultrasound-assistedsolvent extraction[2 solid-phasemicroextraction[11]anddirectair mea-surementsusingProton-Transfer-ReactionMassSpectrometry[4

Thesemethodshavereportedairborneconcentrationsintherange

of␮g/m3,pg/gramandsingle-digitng/m3,respectively

Theshifttowardsthereductioninextractionmass(aka microex-tractions),beitliquidorsolidphaseextraction,isattractiveasthe analytecanbeconcentratedfromtheadsorbentintomicrolitersof solventthatcanbeinjectedintothecolumn.However,thehandling

https://doi.org/10.1016/j.chroma.2018.08.066

occur.Inthisregard,thermaldesorption(TD)isadvantageousover

microextractionsasthelackofextractionmatrixresultsinhigher

recoveryofthesample,reducedpotentialcontamination,andmore

rapidanalysis[12].Additionally,thisstudyusedanautomatedTD

systemwhichhelpedtoreducehumanerror.Althoughprevious

workshaveutilizedthismethod,nodetailedmethodologyforthe

quantificationofTFviaTD-GC-EI-MShasbeen,toourknowledge,

reported[6,13] Furthermore,theInternationalOrganizationfor

Standardization(ISO)standardonindoorairsamplingofvolatile

organiccompounds(VOCs)withTD-GC-EI-MS(ISO16000-6:2011)

isbroadinscopeasthechemicalandphysicalcharacteristicsof

VOCsencompassawiderange.Tothisend,wereportthe

quan-tification ofindoor airborneTF viaTD-GC-EI-MS and detailthe

instrumentalandexperimentalconditionsneededtoobtainsimilar

results.Thismethodwasthenusedtocharacterizetheemanations

ofanovelpassivereleasedevice

2 Materials and methods

2.1 Instrumentationandmaterials

TD-GC-EI-MSwasaccomplishedusingaPerkinElmerClarus®

SQ 8Cmass spectrometer,Clarus® 580gas chromatographand

TurboMatrixTM 650automatedthermal desorber(Waltham,MA,

USA) A PerkinElmer TurboMatrixTM TC 220conditioningoven

(Waltham,MA,USA)wasused toconditionthermal desorption

tubes.AnSKC, Inc.AirChek air samplerpumpModel 224-44XR

(Eighty Four, PA, USA) was used to collect air samples onto

MarkesInternational,Inc.thermaldesorptiontubes(Sacramento,

CA,USA)filledwiththeadsorbentTenax®35/60.AMesaLabsBios

Defender510airflowcalibratorwasusedtocalibratetheSKC,Inc

quadadjustablelowflowholder(SKC,Inc.,EightyFour,PA,USA)

ChromasolvTMmethanol(HPLCGrade)waspurchasedfromFisher

Scientific,Inc.(Hampton,NH,USA).Ultra-highpurityheliumwas

providedbyAirgas,Inc.(Radnor,PA,USA)andnitrogenwas

gener-atedin-housebyaParkerBalstonModelN2-35nitrogengenerator

(ParkerHannifinCorporation,Lancaster,NY)

2.2 Standardpreparation

StandardsofTFwerepreparedusingBayerTFobtainedfrom

United States Department of Agriculture-Agricultural Research

Service-Center for Medical, Agricultural and Veterinary

Ento-mology (USDA-ARS-CMAVE, Gainesville, FLDaniel L Kline, see

acknowledgements).Astocksolutionof10 mg/mLofTF in iso-propanolwasprepared volumetricallyand serial dilutionswith isopropanolwereusedtoprepareasetof4–8workingstandards rangingfrom10pg/␮Lto10ng/␮L.Standardtubeswerespikedat theinletwith0.5␮Lofliquid workingstandardandtransferred ontotheTenax® bedwithaflowofN2at100mL/minfor15min AllthermaldesorptiontubeswereconditionedwithN2 atarate

of100mL/minpriortospikingwithworkingstandards.The condi-tioningcyclewas250◦Cfor20min,followedby300◦Cfor20min andthen335◦Cfor30min.Theheatingrateforallrampperiods was10◦C/min

2.3 Parametersforquantitativeanalysis

To recovertheTF from thesample tubes, a two-stage ther-maldesorptionmethodwasused.Theprimarydesorptionfrom thesampletubewasperformedatvarioustemperaturesranging from200to335◦Cforthreeminutestoassesstheoptimum des-orptiontemperature.Priortosecondarydesorption,thesamples werecollectedonaTenax®concentratingtrapheldat-20◦Cviaa flowofhighpurityheliumatarateof75mL/min.Theheatedvalve

inthethermaldesorptioninstrumentwasmaintainedat300◦Cand transferlinetotheGCwasmaintainedat280◦Catalltimesto pre-ventcondensationoftheanalytesineitherthevalveortransferline

TodeliverthesampletotheGCcolumnforanalysis,secondary des-orptionofthesamplefromtheconcentratingtrapwasperformed

byheatingthetrapto335◦Catarateof40◦C/secwithahelium flowrateof2mL/min

ThetemperatureprogramfortheGCwastwominutesat55◦C followedbya20◦C/minrampto290◦Candthenaholdat290◦Cfor 4.25min.Heliumwasusedasthecarriergasandtheflowratewas

2mL/min.ThecolumnutilizedwasaPerkinElmerElite624(Cat

No.N9315068)midpolarcolumn(6%cyanopropylphenyl–94% dimethylpolysiloxane)withdimensions30mlength,0.25mmID and1.4␮mfilmthickness.ThetransferlinefromtheGCcolumnto theMSsourcewasmaintainedat250◦Candtheelectronionization (EI)sourcewasmaintainedat280◦C

The mass spectrum of TF wasidentified (positive ion mode – 70eV) forTF byidentifyingthetargetion atm/z 163andits fragmentation qualifyingpeak atm/z 91 (Fig.1 which is con-sistentwithpreviouslyobservedmassspectraofTF[14–19].For quantitativeanalysis,themassspectrometerwassettoselected ionrecording(SIR)modeat163m/zwiththeresultingSIR chro-matographintegratedandthepeakareaoftheTFpeakusedfor quantification

2.4 Statisticaldeterminationoflimitofdetection(LOD)andlimit

ofquantification(LOQ)

TheLODandLOQweredeterminedexperimentallybysignalto

noiseratio(SNR)aswellasthestandarderrorofestimates(SEE)

approach.FortheSNRmeasurement,theLODmusthaveanSNRof

atleastthreewhiletheLOQwasdeterminedattheSNRof10[20]

ThecalculationfortheLODfollowingtheSEEapproachisshown

inEq.(1)wheresy/xisthestandarderroroftheestimate,misthe

slopeofthefittedregressionlineandk-factoris3.3and10forthe

LODandLOQasnotedbypreviousauthors[21,22]

LOD=

k·sy/x

Eq.(2) shows thecalculationfor the standard error of

esti-mateswhereyiisthesampledataandyFisthevalueofthefitted

regressionlinefrommatrixspikedsamplesthatcorrespondstothe

concentrationofTFusedtoachievevalueyi.Concentrationsnear

thelowendofthelinearrangewasusedforSEEcalculations

sy/x= 

Duetothelargebackgrounddrift(columnbleed),the

Euro-peanPharmacopoieachromatographicmethodofselectingalarge

backgroundregiontosamplethenoiseisnotachievablewithout

potentiallymisleadingbaselinecorrectiontechniques[20].Thus,

thebackgroundnoisewastaken0.1minawayfromtheanalyte

peakin eachdirection.Data wascompiled fromthreedifferent

experimentsoverdifferentdayswithsamplescollectedin

tripli-cate

2.5 Airsampling

AirsamplingofairborneTFwasaccomplishedusinganovel

pas-sivereleasedevicecreatedwiththesameTFsourceusedtocreate

thestandardsinSection2.2.Thisdevicewascomposedofa

cot-tonwickwithaninteriorreservoirofisopropanolandTFusingthe

sameBayerTFfromSection2.2[5 Thedevicewasactivatedand

subsequentlyallowedtoreleasevolatilesinthecenterofaroomof

5.2×3×2.7m.Theairsamplingoccurredwithin3cmofthedevice,

referredtohereasthepointofgeneration(POG),aswellasone

meterfromthedeviceatthesameheight(1m)abovethefloor

Boththeairsupply (∼0.17m3/s)andexhaust(∼0.21m3/s)were

continuouslyrunningandlocatedontheceilingoftheroom

Thermaldesorptiontubeswereplacedontoadjustablelowflow

holdersuspendedonemeteroffthegroundatbothPOGandone

meter.Flowrateforeachflowportwasadjustedto150mL/min

andvalidatedusingtheairflowcalibrator.Airsamplingwithouta

passivereleasedevicepresentwasaccomplishedtoestablishthe

backgroundsignal.Afterwards,adevicewasactivatedandplacedin

theroomandallowedtoemanatefor60minafterwhichairsamples

werecollectedonTenax®35/60providinganaverageof9Lof

sam-plevolumepassingthrougheachtube.Brasscapsequippedwith

polytetrafluoroethyleneferrulesweresecuredontotheseair

sam-pletubestoprotectthesamplefromcontaminationandpotential

samplelossbetweencollectionandanalysis

Airborne concentration of TF was recorded in parts per

quadrillionbyvolume(ppqv)usingthefollowingequation:

ppqv=

1015·molTF·24.6

WhereppqvisthepartsperquadrillionbyvolumeofTF,molTFisthe

molesofTFquantifiedbyTD-GC-EI-MS,24.6istheratiooflitersto

molesaccordingtotheidealgaslawat23◦Cand101,325PaandV

istheaveragevolume(9L)ofairsampledinliters

ThematrixeffectofindoorairontherecoveryofTFwasstudied

byexposingTF-spikedtubestoindoorairat150mL/minforone(1)

hourorthenitrogengeneratorforone(1)hourat100mL/min

3 Results and discussion

3.1 Optimizationofparametersandmethods

ToassesstheretentionofTFonthecolumn,blanktubeswere analyzedbetweeneachspikedtube.Themeasuredinstrumental transfluthrincarry overwastypicallyless than1%of thesignal intensityofthespikedsampleTF Thesmallmass ofTF onthe column(<10ng)allowedtheuseofasplitlesstransferfromthe secondarydesorption.TheretentiontimeofTFwas12.94min Primarydesorptiontemperatureswerevariedtoelucidatethe optimaltemperatureforrecoveryofTFfromthesampletube.The resultsindicatethata200◦Cdesorptiontemperatureonly recov-ers70.5%ofpotentialTFandthatadesorptiontemperatureabove

300◦Cisneededformaximumrecovery(Fig.2).Thesignal inten-sitywasnormalizedtothelargestsignalobtained,typicallyfrom the325◦treatedtubes.Therecoveryapproachesamaximumat

300◦Cwithanaveragerecoveryof92.9%.Sampletubeswere ana-lyzedtwicetocheckforanyTFcarryoverandnosignalwasdetected

onthesecondrunwhenhigher(+300◦C)desorptiontemperatures wereusedonthefirstrun.Thisexperimentwasrepeatedthree timeswithsamplescollectedintriplicate

3.2 Performanceofmethod/figuresofmeritlowerdetectionlimit, lowerquantificationlimit,accuracyandprecision

Thetotalionchromatograms(TIC) ofunspikedTenax® tubes didnotyieldanymeaningfulnoiseattheretentiontimeofTFdue

tothebackgroundfromthecolumnbleedconvolutinganynoise measurements.Therefore,noaccuratestandarddeviationofblank tubescouldbeascertainedforstatisticalLODorLOQdetermination withoutapplyinganypotentiallymisleadingbaselinecorrections Forexample,abaselinecorrectionmethodutilizingGram-Schmidt (GS)orthogonalizationwasadaptedfromaGC-IRstudy[23]butthe resultingchromatogramwasriddledwithartificialpeaks(datanot shown).Toavoidusingbaselinecorrection,amodifiedstatistical approachbasedontheSEEwasemployedasdescribedinSection

2.4[22]

Analytepeakareawasusedforquantitationsinceusingpeak heightgenerallyhadalargervariance.Thus,rootmeansquared (RMS)SNRwasusedtodeterminetheLODandLOQanddefinedas theTFquantitythatyieldsapeakareaof3and10timeslargerthan thenoise[24].Bothempirical(RMSSNR)andstatistical(SEE)values wereobtainedandreportedhereastheyprovidecomplimentary informationonthedetectionlimitofthismethod.Theempirical methodgivesinsightintothedetectionofaverylowsignalthatis distinctfromablankbackgroundandhelpstakeintoaccountthe instrumentalnoise;itismoreakintoaninstrumentaldetection limit.TheSEEprovidesinsightintothestatisticalvariationofthe

Table 1

Compound SIM (Target, Qualifier) LOD via SEE

(pg, ppq v )

LOQ via SEE (pg, ppq v )

LOD via SNR (pg, ppq v )

LOQ via SNR (pg, ppq v )

Linear Range (pg, ppq v )

R 2, a

methodology(i.e.humanerror,samplepreparation,matrixeffects,

etc.)andmaybecomparedtothemethoddetectionlimit

TheLODandLOQobtainedfromtheSNRmethodwere2.16pg

(16ppqv)and10.3pg(76ppqv)respectively.TheSEEatthelowest

linearconcentrationyieldedanLODandLOQof7.60pg(56ppqv)

and23.0pg(170ppqv)respectively.SincethespikedTFstandards

werenotgeneratedusingactiveairsampling,thecorresponding

ppqvvaluesabovewerecalculatedusingEq.3basedona9Lvolume

tofacilitatecomparisonwiththeairsamplingexperiment.These

figuresofmeritwereobtainedfromaveragingfourexperimental

repeatswithsamplescollectedintriplicateandaresummarized

inTable1.PreviousworkonTD-GC-EI-MSdeterminationof

air-borneTFreportedadetectionlimitof2000pg/tubewhileherewe

reportthat10.28pg/tubecanbequantified[13].Whencompared

toultrasound-assistedsolventextractionofactivelysampledair,

theLODsreportedherewereasmuchasfivetimeslessthanthat

ofthepreviousauthors’LODwhenusingsimilarLODcalculations,

demonstratingasignificantimprovementinsensitivity[2

Theinstrumentusedherehadalinearrangeoffourordersof

magnitudewhenmeasuringTF.Athighconcentrations(20ngper

tube)theresponsebecamenonlinearduetodetectorsaturation

Therefore,thelinearrangeusedinthisstudywasfrom10pg–10ng

TFontube.Regressionwasperformedoverthisconcentrationrange

usingatleast3spikedstandardswithanormalizedslopeof0.9661,

intercept of 0.0005,standard errorof 0.01996 and R2 value of

0.9981.They-interceptwasnotforcedtozerobecausetheputative

criteriafordoingsoinvolvesastandarddeviationmeasurementof

theblankwhichwasnotcompatiblewiththismethod

The use of negative chemical ionization (NCI) may further

improvethedetectionofTFinGC-EI-MS[25] Apreviousstudy

ofTFusingliquid-liquidextractionGC-EI-MSreportedanLOQof

1ng/mL[26].Thisimprovementisattributedtothelarge

molec-ularionformedinNCI;thislargerionhaslessinterferencewhen

comparedtothesmallerfragmentsobtainedfromEI[25]

Elec-troncapturedetection(ECD)hasalsobeenshownpreviouslyto

improvedetectionofpyrethroidswhencomparedtoGC-EI-MS[2]

andtandemmassspectrometrysuchascollision-induced

dissoci-ation(CID-MS/MS)mayprovideyetanotherapproachtofurther

increasethesensitivity/detectionperformance

3.3 AirsamplingofTF

Totestthemethodwitharelevantmatrix,airwassampledin

thevicinityofapassivereleasedevice[5 Thisdevicepassively

releasesTFdissolvedinisopropanolviaawickandreservoirdesign

[5 Samplingnearthedeviceandonemeterawayrevealeda

gra-dientinreleasewith∼200pg/L(13.1ppb)beingadsorbednearthe

deviceand∼50pg/L(3.44ppb)atonemeter(Fig.3).Airsampling

withoutadevicepresentrevealedlowbackgroundsignal∼15pg/L

(∼1ppb).Airsamplingdatawererecordedoverthreeexperiments

andsampleswerereplicated4–8times.Oneoutlierwasrejected

fromthedatasetusingaDixon’sQ-testwithaconfidenceinterval

of99%.ThesourceofanomalousbehaviorissuspectedtobeaTF

contaminatedcap.Duringtheanalysis,nosignalwasobservedin

thesecondarybreakthroughtubessuggestingthatTenax® 35/60

effectivelycapturesalltheTF.Finally,theeffectofindoorairhad

a10%effectonrecoverywhencomparingspikedsamples

Table 2

Spike (pg) RSD with Standards (%) Matrix Effect on Recovery (%)

10 13.0 ± 10.1 10.90 ± 9.41

100 6.30 ± 1.25 10.5 ± 5.83

4 Conclusion

Herewedescribeamethodtoquantifyairborneconcentrations

of TFusing TD-GC-EI-MSwithactive airsampling.A two-stage desorption via a cryotrapwas employed toeliminate a typical microextractionstepthusimprovingtherecoveryofcollected sam-ple.Theuseofasplitlessinjectionensuresthatallthesamplewill reachthedetectortherebyincreasingtheaccuracyofthemethod Finally,activeairsamplingcanbeusedtocollectandanalyzelarge volumesofairforexceedinglytraceamountsofcompound– effec-tivelyincreasingthedetectabilityofthemethod

ThecolumnbleedfromthehightemperatureGCmethodwas notcircumventedwithaGram-Schmidtbaselinecorrection.Future studiesexploringhightemperaturecolumnsorasingularvalue decomposition(SVD)method[27]toeliminatethecolumnbleed backgroundmayfurtherimprovethismethod

TheISOforactiveairsamplingofVOCsusingTD-GC–MS sug-gestsarangeoftemperatures,flowratesandtimestosuccessfully desorptheanalyte(ISO16000-6:2011(E)).Herewereporthigh per-formanceTDwithvaluesforseveralparametersthatlieoutsidethe typicalrangeintheISO.Theuseofamoreaggressivedesorption methodwasneededtoachievea>90%recoveryofspikedsamples

dis-playsuniqueeffects;atextremelylowairborneconcentrationsTF

hasbeenshowntobeanattractanttomosquitoswhileincreasing

concentrationseventuallyleadstodeath[1 Currently,the

concen-trationsatwhichtheseuniquebehaviorsarisearenotknownbut

apreviousstudyhassuggestedthataconcentrationof1pptvmay

beenoughtorepelmosquitos[6 WiththeLODandLOQreported

inthismethod,theairborneconcentrationsontheorderofsingle

digitpptvcanbequantified.Thislevelofdetectionmakesit

pos-sibletostudythevariousbehavioraleffectsoflowconcentrations

ofTFonmosquitobehavior.Themethodwastestedbymeasuring

theairborneconcentrationofTFemanatingfromanovelpassive

releasedevice,demonstratingthatlowlevelsofairborneTFcan

reliablybequantified

Disclaimer

Anyopinions,findings,andconclusions orrecommendations

expressedinthismaterialarethoseoftheauthor(s)anddonot

necessarilyreflectthepositionorpolicyoftheGovernmentandno

officialendorsementshouldbeinferred

Conflict of interest statement

BJWhasminorownershipinterestsinSustainedRelease

Tech-nologies,Inc andPestNatural,Inc.(«5%);theseentitiesdidnot

contributetothesupportofthisstudy

Acknowledgements

This project was sponsored in part by the Department of

the Army,U.S Army Contracting Command, Aberdeen Proving

Ground,Natick ContractingDivision, FtDetrick MDviaa grant

toCDBfromtheArmedForcesPestManagementBoard(AFPMB)

DeployedWarfighterProtectionResearchProgram(DWFP-Grant

No.:W911QY-15-1-0003) and by UCF PreeminentPostdoctoral

(P3) Program awardto MWCK Theauthors would also liketo

thankJedidiahKline(UniversityofFlorida)forassistancein

creat-ingthenovelpassivereleasedevices.Transfluthrin(TF)utilizedin

thisprojectwassuppliedtoUSDA-ARS-CMAVE(Gainesville,FL)by

BayerthroughaMaterialTransferandResearchAgreement(MTRa)

TheMTRaallows transferofTFtoathird partyparticipatingin

researchassociatedwithDWFP

References

http://dx.doi.org/10.1080/15216540701352042

01.093

1016/j.envint.2013.07.011

[5] B.J Willenberg, P.G Koehler, C Batich, G Georgiades, Methods and Devices

for Sustained Release of Substances USA 9,258,988, 2015, January 22.

[6] L.M Meyers, Mosquito Mortality and the Characterization of Airborne

Transfluthrin Concentrations in a Semi-Field Setting, Uniformed Services

University, School of Medicine, Bethesda, MD, USA, 2015.

dx.doi.org/10.1038/nbt1056

1098/rstb.1998.0329

2008.02.080

http://dx.doi.org/10.1371/journal.pntd.0005455

dx.doi.org/10.1021/jf1019592

doi.org/10.1365/s10337-006-0092-7

02772248.2012.729838

doi.org/10.1365/s10337-006-0092-70009-5893/06/12

1007/s00216-012-6003-x

[21] D.A Armbruster, M.D Tillman, L.M Hubbs, Limit of Detection (LOD)/Limit of Quantification (LOQ): comparison of the empirical and the statistical methods exemplified with GC–MS assays of abused drugs, Clin Chem 40 (1994) 1233–1238, PMID: 8013092.

2153-2435.1000355

http://dx.doi.org/10.1021/ac981179n

doi.org/10.1016/j.trac.2010.03.011