Ram

Da

Nb

a

ARRAA

KGGPRPS

1

hpeaaalkaflp5

atb

0h

Journal of Chromatography A, 1334 (2014) 118–125

Contents lists available at ScienceDirect

Journal of Chromatography A

jo ur nal ho me pag e: www.elsev ier .com/ locate /chroma

apid and sensitive method for the determination of polycyclicromatic hydrocarbons in soils using pseudo multiple reactiononitoring gas chromatography/tandem mass spectrometry

ayue Shanga,∗, Marcus Kimb, Maxine Haberla

Pacific and Yukon Laboratory for Environmental Testing, Science and Technology Branch, Pacific Environmental Science Centre, Environment Canada,orth Vancouver, British Columbia, CanadaAgilent Technologies Inc., Mississauga, Ontario, Canada

r t i c l e i n f o

rticle history:eceived 30 October 2013eceived in revised form 24 January 2014ccepted 27 January 2014vailable online 3 February 2014

eywords:C/MS/MS

a b s t r a c t

A method for the rapid determination of 18 polycyclic aromatic hydrocarbons (PAHs) in soil has beenestablished based on a simplified solvent extraction and GC/MS/MS operated in pseudo multiple reactionmonitoring mode (PMRM), a technique where the two quadrupoles mass monitor the same m/z. ThePMRM approach proved superior to the classic single quadrupole technique, with enhanced sensitivity,specificity, and significant reduction in time consuming sample clean-up procedures. Trace level PAHscould be readily confirmed by their retention times and characteristic ions. The limit of quantitation insoil was observed to be 20 ng/g for 16 EPA-priority PAHs and 2 additional PAHs specific to Environment

C/MSolycyclic aromatic hydrocarbonsapid extractionseudo MRMoil and sediment

Canada. The developed method was linear over the calibration range 20–4000 ng/g in soil, with observedcoefficients of determination of >0.996. Individual PAH recoveries from fortified soil were in the range 58.1to 110.1%, with a precision between 0.3 and 4.9% RSD. The ruggedness of the method was demonstratedby the success of an inter-lab proficiency test study organized by the Canadian Association for LaboratoryAccreditation. The present method was found to be applicable as a rapid, routine screening for PAHcontamination in soil, with significant savings in terms of preparation time and solvent usage.

. Introduction

Polycyclic aromatic hydrocarbons (PAHs) are ubiquitousydrophobic compounds originating from natural or anthro-ogenic sources. These compounds are widely distributed in thenvironment and detected in soils and sediments, mainly due totmospheric deposition processes [2]. All PAHs in the environmentre an ecological and human-health concern. Of the one hundrednd twenty-six Environment Protection Agency Priority Pollutantsisted by the Clean Water Act, sixteen are PAHs, with seven beingnown carcinogens [1]. It is recognized that an increase in the rel-tive amount of two to four ring compounds, such as naphthalene,uoranthene, and phenanthrene, is usually a good indication of theresence of petrogenic hydrocarbons [1]. Larger PAHs such as the

and 6-ringed compounds are indicative or pyrogenic sources [3].The reserves of oil sands bitumen in Northern Alberta, Canada,

re estimated at 1.7 trillion barrels, with 173 billion estimatedo be economically recoverable. Oil exploration in this region haseen intensified over the past 20 years, with production increasing

∗ Correspondingauthor. Tel.: +1 604 903 4462; fax: +1 604 903 4408.E-mail address: dayue.shang@ec.gc.ca (D. Shang).

021-9673/$ – see front matter. Crown Copyright © 2014 Published by Elsevier B.V. All rittp://dx.doi.org/10.1016/j.chroma.2014.01.074

Crown Copyright © 2014 Published by Elsevier B.V. All rights reserved.

from 100,000 barrels per day to about 1.5 million barrels per daycurrently [2]. Close monitoring of PAH concentrations in soils andsediment has become critical, and large scale surveillance is beingimplemented by government agencies. The characterization andknowledge of PAH concentrations in soil and sediments can beinstrumental in tracing an oil spill source and enabling remedia-tion efforts. A rapid, sensitive, and robust analytical method for thedetermination the PAH concentrations in soil is urgently needed[2].

Traditional sample preparation techniques for the determina-tion of PAHs in soil are time consuming and generally require largevolumes of toxic solvents, together with multi-step extraction andsilica gel or Florisil column clean-up procedures. To address theseissues, and as an alternative to the classic Soxhlet solvent extrac-tion methods, various techniques have been developed and used inthe analysis of PAHs from soil. Alternative processing includes pres-surized liquid extraction or accelerated solvent extraction (PLE orASE), ultrasonic extraction, supercritical fluid extraction (SFE), andmicrowave-assisted extraction (MAE) [4–6]. An automated Soxhlet

method has recently been developed with corresponding reductionin soil extraction time [7,8]. Despite intensive method develop-ment in this area, some of the referenced techniques suffer oneor several shortcomings, including low recovery, expensive initial

ghts reserved.

atogr.

in“bateupc

ddbmhpomrmFetgmMpscaiiaacca

avmawabGdwcA

2

2

A(flfl((a(

D. Shang et al. / J. Chrom

nvestment, frequent equipment malfunction, and lack of robust-ess or ruggedness. Very recently, a new promising approach ofmicroextraction” has emerged using MAE combined with solventar [9]. While this approach is both “green” and effective, widepplication of this method remains to be seen. An elegant approacho the issue would be to take advantage of the modern instrument’snhanced capability of handling less processed sample extracts andse a “dilute and shoot” approach. Perhaps more importantly, sim-lified sample processing improves method ruggedness, which isritical for routine analysis.

Presently the two most frequently employed techniques toetermine PAHs are HPLC with fluorescence, UV, or diode arrayetection [10,11] and GC with MS detection [1,7,8,10]. The HPLCased methods are usually fast in comparison to the GC/MSethods; however, the disadvantages of the HPLC method are

eavy dependence on chromatographic retention time for com-ound identification and the HPLC methods are typically an orderf magnitude lower in sensitivity than GC/MS [12]. In complexatrices, such as soil extract, peak identification based solely on

etention time is subject to interference from other components,aking trace level PAH contamination difficult to characterize.

or this reason, over 15 years the GC/MS technique has becomestablished as the accepted method for PAH determination inhe environment [7,8]. Despite numerous improvements to sin-le quadrupole MS instrumentation however, performance cannotatch the sensitivity and specificity offered by triple quadrupoleS. As a consequence, an increasing number of peer reviewed

ublications have applied GC/MS/MS techniques to PAH analy-is. However, due to the unique structure stability of the PAHompounds, the traditional Multiple Reaction Monitoring (MRM)pproach has been hampered by generally weak fragmentationon responses for this group of compounds [13–15]. Consider-ng the well-established GC/MS single quadrupole method, thepplication of the triple quadrupole presently does not providedequate improvement in sensitivity and specificity to initiate ahange from proven procedures. In this regard we challenged thisonclusion and successfully applied GC/MS/MS techniques to PAHnalysis.

In this paper, we present a rapid analytical method for thenalysis of PAHs in soil and sediments, based on a one step, lowolume solvent extraction followed by GC/MS/MS in pseudo MRMode. Long extraction time, large solvent volume consumption,

nd extensive silica gel column clean-up were eliminated. Thisas made feasible by the increased sensitivity and specificity

chieved by pseudo MRM mode GC/MS/MS. Compelling results wille presented to support the favoring of this pseudo MRM modeC/MS/MS over that of single quadrupole procedures, even forifficult-to-fragment compounds like PAHs. The present methodas validated and applied successfully during an inter-lab profi-

iency study organized by The Canadian Association for Laboratoryccreditation Inc. (CALA).

. Material and method

.1. Reagents and standards

The 18 PAHs analyzed in this study were Acenaphthene (ACE),cenaphthylene (ACY), Anthracene (ANT), Benzo(a)anthracene

BAN), Benzo(a)pyrene (BAP), Benzo(e)pyrene (BEP), Benzo(b)uoranthene (BBF), Benzo(g,h,i)perylene (BGP), Benzo(k)uoranthene (BKF), Chrysene (CRY), Dibenz(a,h)anthracene

DBA), Fluoranthene (FLA), Fluorene (FLU), Indeno(1,2,3-cd)pyreneIND), Naphthalene (NAP), Perylene (PER), Phenanthrene (PHE)nd Pyrene (PYR). A certified standard solution of the 18 PAHs2000 �g/mL each) was provided by SPEX CertiPrep (Metuchen,

A 1334 (2014) 118–125 119

NJ). This solution was stored at −20 ± 10 ◦C in amber glass andhad a shelf life of 12 months. An internal standard solution ofNaphthalene-d8, Acenaphthene-d10, Phenanthrene-d10, andPerylene-d12 was purchased from Supelco (Oakville, Ontario).This internal standard was employed both in the preparationof calibration standards and in fortifying soil samples for spikerecovery.

Calibration standards were prepared in dichloromethane byserial dilution of primary standard to provide final concentra-tions of 10, 20, 40, 100, 500, 1000, 1500 and 2000 ng/mL. Internalstandard at a final concentration of 200 ng/mL was added to allcalibration standards.

Disposable centrifuge filter tubes (15 and 50 mL, Polypropy-lene/Polyethersulfone) were supplied by Pall Corporation (PortWashington, NY). Disposable 50 mL polypropylene centrifuge tubeswere purchased from Sarstedt (Numbrecht, Germany). Florisil®

adsorbent (60–100 mesh) was from Fisher Scientific (Fairlawn, NJ.USA). OmniSolv solvents dichloromethane (DCM), acetone (ACE),hexane, isopropanol (IPA), acetonitrile (ACN), pesticide grade, werepurchased from EM Science (Gibbstown, NJ. USA).

2.2. Sample extraction and clean up

Aliquots of 10 ± 0.1 g of air dried free flow homogeneoussoil sample were weighed and placed into a 50 mL polypropy-lene centrifuge tube with screw caps. To the sample, 200 �Lof 20 ppm internal standard mixture were added, followed by5 g of sodium sulphate (pre-dried at 350 ◦C). The mixture wasthen hand-shaken to mix sodium sulphate with the soil sam-ple, with occasional spatula use to break any soil lumps toensure homogeneity. After mixing, 15 mL of dichloromethanewas added and the mixture was vortexed briefly. The slurrywas further shaken for 10 min at room temperature using amechanical wrist action shaker. The sample was centrifuged at5000 rpm (4696 g) for 5 min. The supernatant was decanted andretained in a clean 50 mL polypropylene centrifuge tube. Theremaining pellet was subjected to a second extraction in 5 mLdichloromethane (breaking up the pellet “cake” with a spatula ifnecessary), employing only a 5 min shaking time. Supernatant fromboth extractions were pooled and the volume adjusted to 20 mLwith dichloromethane.

An aliquot of the 20 mL extract was transferred to a 15 mLcentrifuge filter tube with 0.2 �m filter device. Following cen-trifugation for 5 min at 5000 rpm (4696 g), the filtrate extractwas ready for GC/MS/MS analysis. Refer to Fig. 1 for a flowchartof sample extraction steps. For soils contaminated with lubeoil, vegetable oil, or animal oil and grease, the filter insertof the centrifuge tube may be pre-packed with approximately3 g of Florisil to improve clean up. These materials may loweranalyte recovery and the inclusion of an isotope dilutiontechnique may be required to compensate (Supplementary mate-rials).

2.3. GC-MS analysis

A gas chromatograph (GC) HP 7890A from Agilent Technologies(Palo Alto, CA., USA) equipped with an Agilent 7693B automatic liq-uid sampler with 10 �L syringe was used for the separation of PAHs.Analysis employed a 1 �L sample injection in pulsed splitless mode(pulsed pressure at 50 psi with the split valve closed for 1 min). Allanalytes were separated on a Restek Rtx-5MS with Integra-guardcolumn (30 m x 0.25 mm id, 0.25 �m). A 4 mm i.d. single tapered,

deactivated inlet liner with glass wool at the bottom (Agilent Tech-nologies) was installed into the injector. The oven temperatureprogram was as follows: initial temperature at 50 ◦C (hold 2 min),then 6 ◦C/min to 310 ◦C, hold for 20 min. The total run time was

120 D. Shang et al. / J. Chromatogr. A 1334 (2014) 118–125

Add 10 g b ank soil to a 50 mL PP

centr ifug e tub e. Spike in 0.2 mL of

surrog ate. Add 5 g sodium

sulph ate.

Vortex mix. Centrifuge at 5000 rpm for 5 minutes. Decant into a clean 50 mL

centrifuge tube.

Vortex mix. Ce ntr ifug e at 5000 rpm for 5 minu tes and d ecant in to the same tub e fr om

the first extraction.

Add 5 mL o f sampl e to a 15 mL di sposabl e ce ntr ifug e fil ter tub e and ce ntr ifug e at

5000 rpm for 5 minutes.

Make volume up to 20 mL with DCM. Vortex mix.

Wrist-action shake the sampl e and solve nt for 10 minu tes.

Add 15 mL of DCM. Vortex mix.

Add 10 g o f sampl e to a 50 mL PP

centr ifug e tub e. Spike in 0.2 mL of

surrog ate. Add 5 g o f sodium

sulph ate.

Add 5 mL o f DCM to solid ‘ cake’. Break up with manu al agi tation (with spatul a if

necessary). Shake for 5 minutes.

GC/MS/MS Analysis

Discard filter insert and transfer 1.5 mL to a GC vial.

QC Real

Sample

ple p

6gamiMmaaooqb

tsdttbt

Fig. 1. Schematic of sam

5.33 min. High purity helium gas (>99.999%) was used as carrieras with the constant flow rate of 1.0 mL/min. Detection of thenalytes employed an Agilent Technologies 7000 triple quadrupoleass spectrometer (MS) operated in electron impact positive mode

onization at 70 eV. Analyte ions were monitored in either “Classic”ultiple Reaction Monitoring (CMRM) or “Pseudo” MRM (PMRM)ode. The GC/MS transfer line and inlet temperatures were set

t 300 and 320 ◦C respectively. Ion source temperature was sett 325 ◦C and quadrupole temperature at 150 ◦C. A solvent delayf 4.5 min was employed. Table 1 lists the PAHs along with theirbserved retention times and their characteristic quantitation andualifier ions. The N2 collision cell and He Quench Gas flows wereoth set to 1 mL/min.

Quantification employed the integrated peak area ratio of thearget ion to internal standard. A weighted (1/x) linear regres-ion of the calibration standard responses was employed toefine the calibration curve from which measured concentra-

ions were calculated. The PAH analytes were identified by theirarget ions and retention time order. Retention times had toe within ± 0.1 min of the expected time for positive confirma-ion.

reparation procedure.

2.4. Method validation

The linearity of the analytical GC/MS/MS PMRM method wasassessed by analyzing duplicate calibration standards prepared at10, 20, 50, 100, 500, 1000, 1500 and 2000 ng/mL (equivalent tosoil samples spiked at PAH concentrations from 20 to 4000 ng/mL).Linearity using weighted (1/x) least-squares regression was con-sidered acceptable when the correlation coefficient (r) was >0.995.

The limit of quantitation (LOQ) and limit of detection (LOD)of the method were assessed based on the signal to noise (S/N)response of the relevant analyte peak response at the loweststandard concentration of 10 ng/mL. A S/N of >10:1 and >3:1 for LOQand LOD respectively were considered acceptable for each analyte.

Method accuracy (expressed as percent recovery) and precision(expressed as percent relative standard deviation (%RSD)) weredetermined by recovery studies in PAH-free soil samples spikedat low, mid, and high PAH concentrations. Eight replicate soil sam-

ples spiked with PAH standard at 20, 200 and 2000 ng/mL wereprocessed and analyzed. Results showing an accuracy of 60% to120% recovery from nominal concentration and a precision of <20%RSD were considered to be acceptable. The percent recovery was

D. Shang et al. / J. Chromatogr. A 1334 (2014) 118–125 121

Table 1List of PAHs and pseudo MRM acquisition parameters.

Compound name ISTD Indicative RT (min) Precursor> product ion Qual ion 1 Qual ion 2 Dwell (ms) quant/qual CE (V)

Time Segment 1: 14.50 mind8-Naphthalene X 15.19 136 > 136 137 > 137 15 5Naphthalene 15.26 128 > 128 129 > 129 10 5

Time Segment 2: 21.00 minAcenaphthylene 21.32 152 > 152 153 > 153 76 > 76 10 10d10-Acenaphthene X 21.94 163 > 163 162 > 162 15 5Acenaphthene 22.06 153 > 153 154 > 154 76 > 76 10 5

Time Segment 3: 23.00 minFluorene 24.02 166 > 166 165 > 165 83 > 83 10 5

Time Segment 4: 27.00 mind10-Phenanthrene X 27.61 188 > 188 189 > 189 15 10Phenanthrene 27.69 178 > 178 179 > 179 89 > 89 10 5Anthracene 27.86 178 > 178 179 > 179 89 > 89 10 10

Time Segment 5: 31.00 minFluoranthene 32.35 202 > 202 203 > 203 101 > 101 10 5Pyrene 33.17 202 > 202 203 > 203 101 > 101 10 10

Time Segment 6: 37.00 minBenz(a)anthracene 37.92 228 > 228 114 > 114 229 > 229 10 10d12-Chrysene X 37.98 240 > 240 121 > 121 15 10Chrysene 38.08 228 > 228 229 > 229 114 > 114 10 10

Time Segment 7: 41.00 minBenzo(b)fluoranthene 41.89 252 > 252 253 > 253 250 > 250 10 10Benzo(k)fluoranthene 41.97 252 > 252 253 > 253 250 > 250 10 15Benzo(e)pyrene 42.80 252 > 252 253 > 253 250 > 250 10 5Benzo(a)pyrene 42.95 252 > 252 253 > 253 250 > 250 10 10d12-Perylene X 43.16 264 > 264 132 > 132 15 15Perylene 43.24 252 > 252 253 > 253 250 > 250 10 5

Time Segment 8: 45.00 minIndeno(1,2,3-cd)pyrene 46.40 276 > 276 277 > 277 138 > 138 10 10Dibenz(a,h)anthracene 46.50 276 > 276 277 > 277 138 > 138 10 10

I

c×

2

(CpbbpP

3

3

tppcoa(iu

mstca

Benzo(g,h,i)perylene 47.10 278 > 278

STD: Internal Standard; RT: Retention Time; CE: Collision Energy.

alculated from the equation: Mean (calculated)/Mean (spiked)) 100.

.5. Application to real samples

The validated method was applied to a set of four soil sampleslabeled C-18-01, C-18-02, C-18-03 and C-18-04) supplied by theanadian Association for Laboratory Accreditation Inc. (CALA) asart of a proficiency testing program. The samples were analyzed byoth a well-established GC/MS single ion monitoring method andy the new validated GC/MS/MS PMRM method. Results were com-ared as a demonstration of the performance of the new GC/MS/MSMRM procedure.

. Results and discussion

.1. Sample extraction and clean-up

For PAHs analysis, the ideal extraction solvent(s) should havehe characteristics of high extraction efficiency of the targeted com-ounds from soil, capability of application directly to wet soils (torevent loss of low volatile PAHs during the drying process), andompatibility with GC/MS. To meet these requirements, and basedn practical experience and literature review [1,8], several solventsnd their combinations were examined during the present studyhexane, acetone, acetonitrile (ACN), dichloromethane (DCM) andsopropanol). A detailed discussion of solvent selection and clean-p method development can be found in Supplementary materials.

The most practical approach was achieved by first reducing soiloisture and extraction in neat DCM solvent. The incorporation of

odium sulfate allowed for soil moisture reduction and DCM extrac-ion yielded consistently high recoveries at over 60% for all 18 PAHompounds. Furthermore, all 18 PAH compounds exhibited accept-ble chromatographic peak shape and were satisfactorily separated

279 > 279 139 > 139 10 5

with adequate sensitivity. A representative GC/MS/MS PMRM chro-matogram of a blank soil sample spiked at 0.1 �g/g of PAHs standardand extracted under these conditions is shown in Fig. 2.

For relatively clean soil samples, the main issue is the simulta-neous extraction of interference components, particularly isobariccompounds. These interferences are significant in GC/MS analy-sis but could be overcome efficiently by the proposed GC/MS/MSPMRM approach. For this proposed procedure, the current elab-orate routine sample clean-up and concentration steps, such asroto-vap and nitrogen gas blow down, were rejected in lieu of a“dilute and shoot” approach, with only a filter 0.2 �m centrifugetube filtration step included to remove particulates.

3.2. GC/MS/MS determination

In environmental analysis, co-eluting isobaric matrix interfer-ences derived from soil and sediments usually make MRM thetechnique of choice to achieve high signal-to-noise ratios andmethod specificity. Polyaromatic hydrocarbons possess an excep-tionally stable and rigid macrocyclic ring structure and yield verylittle fragmentation during collision induced dissociation. There-fore, in this work PAHs were analyzed by GC/MS/MS with pseudoMRM monitoring (GC/MS/MS PMRM), a technique where bothresolving MS quadrupole mass filters monitor for the same molec-ular ion m/z that is employed for quantitation. In PMRM, the targetions are selectively transferred to the second resolving quadrupolemass filter without collision induced dissociation.

In the present study, a systematic optimization of targeted PAHcompound ionization and daughter ion fragmentation was con-ducted. Results supported observations by other groups [13–15] in

that the most abundant peak in the electron impact ionization (EI+)fragmentation spectrum was consistently the molecular ion (M+).Other observed peaks were present at lower relative intensitiesdespite efforts to optimize conditions for fragmentation and even

122 D. Shang et al. / J. Chromatogr. A 1334 (2014) 118–125

DCM

ui[trntfpG

atGttwiIeftabi[tqr

swlmos

Fig. 2. Combine extracted ion chromatogram of 18 PAHs spiked in

nder extreme collision energy conditions. Other weaker intensityons that commonly formed were the multiply charge ions such asM − 2H]+ or double-charged analytes M2+. The results suggest thathe highly condensed and stable PAHs are not easily amenable tooutine MRM analysis without loss of sensitivity due to the limitedumber and low response of daughter ions; especially with respecto the low intensity qualifier ions. The lack of increased sensitivityor a triple quad over a single quad GC/MS analysis explains theaucity of peer reviewed papers published using triple quadrupoleC/MS/MS for PAHs analysis [13–15].

The high structural stability of PAH compounds was exploited asn advantage under the present PMRM approach. The characteris-ic PAH analyte parameters used in established single quadrupoleC/MS methods for PAHs were employed for monitoring in both

he first and third quadrupole of the GC/MS/MS. No fragmen-ation of the target ions was attempted, while collision energyas fine tuned to achieve best signal to noise ratio by decreas-

ng or eliminating co-eluting isobaric interference compounds.n effect, the collision energy was tuned to reduce interferingffects by fragmentation or creation of unfavorable energy transferor interfering compounds, while targeted ions remained rela-ively intact. As a result, the PMRM technique provided two maindvantages: (1) potential destruction and mass filtering of iso-aric interferences; (2) the collisional focusing of the ion beam

n a high pressure RF device as described by Douglas and French16] to focus the ions toward the centre axis where they are bet-er able to enter the acceptance aperture of the 2nd resolvinguadrupole mass filter resulting in better transmission at a givenesolution.

To demonstrate the utility of PMRM approach, a systematictudy of the relationship of collision energy to peak height and S/Nas conducted. All 18 targeted PAHs produced a “sweet spot” col-

ision energy where the peak area and S/N were optimized (Fig. 3,ore to be found in Supplementary Fig. S1). In other words, at this

ptimized collision energy, one can expect strong peak and reducedignal to noise ratio for a particular compound. Occasionally, lower

at 100 ng/mL (GC/MS/MS, EI ionization, pseudo-MRM acquisition).

or higher collision energy may be selected to increase eitherpeak height (in the case of poorly ionized compounds and cleanmatrices) or S/N ratio (in the case of strong ions and dirty matri-ces). The present method for PAHs employed collision energies atthe S/N “sweet spot” with a few exceptions. Under the determinedoptimized collision energy conditions for PMRM mode ion opera-tion, it was theorized likely that the GC/MS/MS method would showimproved limit of detection (LOD) and limit of quantitation (LOQ)for “dirty” samples when compared to routine GC/MS.

To illustrate this point, a GC/MS single ion monitoring methodwas created based on a well-established in-house SOP for the anal-ysis of PAHs in soil and sediments and using a single quadrupoleinstrument (7890A/5975C, Agilent Technologies). An experimentwas carried out to compare the GC/MS/MS in PMRM mode to GC/MSin SIM mode. For this experiment, a series of samples were preparedin solvent laden soil extract spiked at various low PAH concentra-tions from 2 to 30 ng/g. The duplicated soil extract samples wereanalyzed by both GC/MS/MS in PMRM mode and GC/MS in a well-established SIM mode. A summary of the results is provided inTable 2. The advantages of PMRM over GC/MS SIM were clearlydemonstrated. A number of PAHs at 10 ppb were not detected bythe GC/MS SIM method due to severe matrix effects. The samesamples produced strong peaks for all 18 PAHs in GC/MS/MSPMRM analysis, with excellent S/N at this trace level. Further-more, at a 2 ng/g spiking level, minor peaks did not achieve anadequate S/N for quantitation of any of the PAHs by GC/MS SIMwhile, remarkably, GC/MS/MS PMRM still exhibited strong defini-tive peaks adequate for quantitation for the majority of the targetedcompounds.

In justification for the PMRM approach, one must answer thefundamental question: how does PMRM analysis compare withClassic MRM (CMRM) analysis in terms of specificity and sensi-

tivity? With respect to specificity, the PMRM offers equivalentconfirmatory identification of analytes by a combination of reten-tion times, quantitative and qualitative ions, and ion ratio betweenquantitative and qualitative ions in the mass spectrum. In addition,

D. Shang et al. / J. Chromatogr. A 1334 (2014) 118–125 123

F dardsa noise

p

ttpeaC

TI

*

ig. 3. Signal to noise ratio in relation to collision energy for 3 typical PAHs. All stannd selected collision energy used in the method. The symbol (�) indicates signal toeak area of the targeted compound at this collision energy (v).

he PMRM method was shown to be superior to CMRM with respecto sensitivity, as a result of the reduction of interfering isobaric com-

ounds. A comparison of S/N for the PAH analyte peaks from soilxtracts fortified 1 ppm PAH standard mixture is shown in Fig. 4nd clearly demonstrates the improved sensitivity of PMRM overMRM for the majority of the analytes (12 out of 18 PAHs). Although

able 2nstrument comparison: area counts at PAH concentrations from 2–20 p p b.

Agilent 7000 triple quadrupole

Compound 2 p p b 10 p p b

Napthalene 999 3909

Acenaphthylene 537 2619

Acenaphthene 343 1190

Fluorene ND 1551

Phenanthrene 1372 4810

Anthracene 930 2046

Fluoranthene 1325 3659

Pyrene 1641 4056

Benzo(a)anthracene ND 955

Chrysene 1029 2791

Benzo(b)fluoranthene ND 1859

Benzo(k)fluoranthene ND 652

Benzo(e) pyrene 405 1841

Benzo(a) pyrene 355 1159

Perylene 813 1976

Indeno(1,2,3-cd) pyrene 202 587

Benzo(g,h,i)perylene 455 1531

Dibenz(a,h)anthracene ND 804

ND = Not detected.

were prepared in DCM and at 1000 ng/mL. The dotted line indicates the optimizedratio of the targeted compound at this collision energy (v). The symbol (�) indicates

detection of some individual PAH compounds could be significantlyimproved by the PMRM method, the elimination of isobaric com-

pounds for all PAH analytes as a group was not complete. In certaincases, notably Acenaphthene, Benzo(e)pyrene and Perylene, inter-fering compounds were observed that possessed similar chemicalstability to the target ion and the original CMRM remained the

Agilent 5975 single quadrupole

20 p p b 2 p p b 10 p p b 20 p p b

8121 ND 29 575758 ND 25 452877 ND 33 332790 ND 22 426462 ND ND ND4213 ND ND ND7925 ND 46 869323 ND 46 942521 ND ND ND6306 ND ND ND4778 ND 35 701411 ND 35 853922 ND 42 853230 ND 33 744363 ND 49 1022237 ND 17 473963 ND 29 632293 ND 0 45

124 D. Shang et al. / J. Chromatogr. A 1334 (2014) 118–125

F ollisio1 cates

i

oGPc

3

oi(2tcwtwm

mlt

ig. 4. Signal to noise comparison for pseudo-MRM and classic MRM in relation to c000 ng/g of 18 PAHs, extracted with DCM and run duplicated. The symbol (�) indi

ndicates signal to noise ratio of the targeted compound using classic MRM.

ptimal choice for analysis. Nevertheless, many well establishedC/MS single quad methods may benefit from this straightforwardMRM approach without extensive new method development pro-ess.

.3. Method validation

The linearity of calibration curves was determined by analysisf duplicate preparations of calibration standards over the nom-nal concentration range of 10–2000 ng/mL for all PAH analytesequivalent to soil samples spiked at PAH concentrations from0 to 4000 ng/g). The MS response for all analytes was observedo be linear in this concentration range, with correlation coeffi-ients of >0.996. The limit of quantitation (LOQ) of the methodas determined to be 20 ng/g for all the targeted compounds. Fur-

hermore, prepared stock standard solutions and working solutionsere found to be stable when stored at –20 ± 10 ◦C for up to 6onths.

The precision and accuracy of the analytical GC/MS/MS-PMRM

ethod was confirmed employing eight soil samples spiked atow (20 ng/g), mid (200 ng/g), and high (2000 ng/g) concentra-ions of PAHs (Supplementary Tables S1, S2 and S3). Accuracy for

n energy for 3 typical PAHs. The results were obtained from a soil sample spiked atsignal to noise ratio of the targeted compound using pseudo MRM. The symbol (�)

determination of the PAH compounds was demonstrated by a per-cent recovery in the range 58.1–110.1%, with an observed precisionof <5%RSD for each of the individual compounds. Results for recov-ery of the PAH compounds at low, mid and high levels were in linewith those reported by other authors using sonication [6,17], ASE[7,8], SFE [4,8] or an automated Soxhlet procedures [7]. The methodshowed a limit of detection limit (LOD) of between 5–10 ng/g forPAH compounds extracted from the soils.

3.4. Analysis of proficiency testing samples

The validated method was applied to a set of four soil samples(labeled C-18-01, C-18-02, C-18-03, and C-18-04) supplied by theCanadian Association for Laboratory Accreditation Inc. (CALA) aspart of a proficiency testing program. Due to the high concentra-tion of spiked of PAHs in the samples, only 1 g of each materialwas processed and analyzed. Table 3 presents the typical results for

one of the samples. The results were compared with the assignedvalue and found to be satisfactory (Supplementary Table S4). Theaccuracy for this test with the current method ranges from 63 to139%.

D. Shang et al. / J. Chromatogr. A 1334 (2014) 118–125 125

Table 3CALA Proficiency testing sample C-18-04: measured vs. assigned concentrations.

Compound Sample Actual concentration (p p b) Calculated concentration (p p b) Accuracy (%)

Acenaphthene C-18-04 1119 1161 104Acenaphthylene C-18-04 1343 1501 112Anthracene C-18-04 1224 1543 126Benzo (a) anthracene C-18-04 6076 5363 88Benzo (a) pyrene C-18-04 4007 3222 80Benzo (b) fluoranthene C-18-04 7869 5198 66Benzo (g,h,i) perylene C-18-04 4452 3843 86Benzo (k) fluoranthene C-18-04 4119 4127 100Chrysene C-18-04 6784 5947 88Dibenzo (a,h) anthracene C-18-04 1029 1202 117Fluoranthene C-18-04 19517 14658 75Fluorene C-18-04 1322 1095 83Indeno (1,2,3 - cd) pyrene C-18-04 5045 4400 87Naphthalene C-18-04 36372 33325 92Phenanthrene C-18-04 16835 13059 78

M

4

orosmc

PpaGpwr

dtsiattap

A

tvS

[

[

[

[

[

62.[15] V. Varlet, T. Serot, F. Monteau, B. Le Bizec, C. Prost, Food Addit. Contam. 24

Pyrene C-18-04 13368

easured concentrations vs. assigned concentrations in CALA PT-C-18-04.

. Conclusions

A sensitive and rapid GC/MS/MS method for the determinationf PAHs in soil samples has been developed. Sample extraction waseduced to a rapid mechanical shaking procedure using just 20 mLf solvent, and further sample clean-up was eliminated for mostamples as a result of the application of GC/MS/MS in pseudo MRMode. An additional Florisil clean-up was all that required in the

ase of more heavy contaminated soils.A series of experiments were conducted to compare GC/MS/MS

MRM and GC/MS/MS CMRM detection modes for soil extract sam-le analysis. Results for the recovery of PAHs demonstrated thedvantage of PMRM triple quad analysis over single quad SIM modeC/MS. In the present study, the highly stable nature of PAH com-ounds was taken advantage of and the collision cell fragmentationas optimized for fragmentation of potential interference ions,

ather than for fragmentation of the highly abundant parent ions.In summary, the PMRM provided additional mass filtering due to

ual quadrupole ion focusing and potential reduction or destruc-ion of isobaric interferences in the collision cell, thus improvingensitivity for most of the targeted PAH compounds. Consider-ng the significant saving in time and solvent volume by PMRMpproach, we suggest that labs currently using GC/MS SIM moveo GC/MS/MS PMRM to improve method productivity and sensi-ivity, especially in the case of structurally stable compounds suchs alkylated PAHs, oil biomarkers, dioxins and furans, and someharmaceutical drugs.

cknowledgements

The authors gratefully acknowledge the support and input ofheir colleagues, notably Randy Englar, Liane Chow, Oxana Blajke-itch, Lauretta Liem and Norman Berke of the Pacific Environmentalcience Centre of Environment Canada, North Vancouver, BC.

[[

14165 106

Appendix A. Supplementary data

Supplementary data associated with this article can befound, in the online version, at http://dx.doi.org/10.1016/j.chroma.2014.01.074.

References

[1] M. Olson, J. Iverson, E. Furlong, M. Schroeder, USGS Water Resources Inves-tigations Report 03-4318, 2004, http://nwql.usgs.gov/rpt.shtml?WRIR-03-4318

[2] J. Kurek, J.L. Kirk, D.C. Muir, X. Wang, M.S. Evens, J.P. Smol, Proc. Nat. Acad. Sci.110 (2013) 1761.

[3] Z. Wang, K. Li, P. Lambert, C. Yang, J. Chromatogr. A 1139 (2007) 14.[4] S. Hawthorne, C. Grabanski, E. Martin, D. Miller, J. Chromatogr. A 892 (2000)

421.[5] Y.Y. Shu, R. Lao, C. Chiu, R. Turle, Chemosphere 41 (2000) 1709.[6] D.R. Banjoo, P. Nelson, J. Chromatogr. A 1066 (2005) 9.[7] Z. Wang, S.A. Stout (Eds.), Oil Spill Environmental Forensics: Fingerprinting and

Source Identification, 1st ed., Academic Press, New York, 2007.[8] Z. Khan, J. Troquet, C. Vachelard, Int. J. Environ. Sci. Tech. 2 (2005) 272.[9] L. Guo, H.K. Lee, J. Chromatogr. A 1286 (2013) 9.10] U.S. Environmental Protection Agency, Method 1654, in: Revision A. PAH Con-

tent of Oil by HPLC/UV, U.S. Environmental Protection Agency, 1992 (EPAPublication EPA-821-R-92-008) http://water.epa.gov/scitech/methods/cwa/oil/upload/2007 07 10 methods method oil 1654.pdf.

11] U.S. Environmental Protection Agency, Polynuclear aromatic hydrocarbons,in: Method 8310, U.S. Environmental Protection Agency, 1986 (EPA SW-846)http://www.epa.gov/osw/hazard/testmethods/sw846/pdfs/8310.pdf.

12] L.D. Gratz, S.T. Bagley, D.G. Leddy, J.H. Johnson, C. Chiu, P. Stommel, J. Hazard.Mater. 74 (2000) 37.

13] B. Veyrand, A. Brosseaud, L. Sarcher, V. Varlet, F. Monteau, P. Marchand, F. Andre,B. Le Bizec, J. Chromatogr. A 1149 (2007) 333.

14] N. Barco-Bonilla, R. Romero-Gonzáleza, P. Plaza-Bolanos, J.L. Fernández-Moreno, A. Garrido Frenich, J.L. Martínez Vidal, Anal. Chim. Acta. 693 (2011)

(2007) 744.16] D.J. Douglas, J.B. French, J. Am. Soc. Mass. Spectrom. 3 (1992) 398.17] C. Sánchez-Brunete, E. Miguel, J.L. Tadeo, J. Sep. Sci. 29 (2006) 2166.