On-line solid-phase extraction coupled with liquid chromatography/electrospray ionization mass spectrometry for the determination of trace tributyltin and triphenyltin in water samples

RAPID COMMUNICATIONS IN MASS SPECTROMETRY

Rapid Commun. Mass Spectrom. 2009; 23: 3795–3802

) DOI: 10.1002/rcm.4321
Published online in Wiley InterScience (www.interscience.wiley.com
On-line solid-phase extraction coupled with liquid

chromatography/electrospray ionization mass

spectrometry for the determination of trace tributyltin

and triphenyltin in water samples

Qian Sun1,2,3, Zuliang Chen2,3*, Dongxing Yuan1, Mallavarapu Megharaj2,3 and

Ravendra Naidu2,3

1State Key laboratory of Marine Environmental Science, Environmental Science Research Center, Xiamen University, Xiamen, China, 3610052Centre for Environmental Risk Assessment and Remediation, University of South Australia, Mawson Lakes, SA 5095, Australia3CRC for Contamination Assessment and Remediation of Environment, Mawson Lakes Boulevard, Mawson Lakes, South Australia 5095,

Australia

Received 16 July 2009; Revised 30 September 2009; Accepted 1 October 2009

*CorrespoAssessmeMawsonE-mail: z

On-line solid-phase extraction (SPE) for pre-concentration and sample cleanup is one strategy to

reduce matrix effects and to simultaneously improve detection sensitivity in liquid chromatography/

mass spectrometry (LC/MS). This paper describes an on-line SPE-LC/MS method for the determi-

nation of tributyltin (TBT) and triphenyltin (TPhT) at trace levels in water samples. The direct

coupling of an on-line C18 pre-column to LC/MS was used to pre-concentrate TBT and TPhT at trace

levels from waters and to remove interfering matrix effects. Pre-concentration was followed by

separation of TBT and TPhT on a C18 column using a mobile phase containing 0.1% (v/v) HCOOH/

5mMHCOONH4 and methanol. While both electrospray ionization (ESI) and atmospheric pressure

chemical ionization (APCI) can be interfaced with MS for the detection of TBT and TPhT, ESI-MS

was preferred for this application. The calibration curve for the targetswas linear in the concentration

range 0.1–30mg L�1. The detection limit (signal-to-noise (S/N) ratio¼ 3) was 0.02mgL�1 when 3.0mL

of sample was enriched on the C18 pre-column. The recoveries of TBT and TPhT in spiked waters

were from 81.0 to 101.9%. The reproducibilities for the analysis of the standardmixture (10mgL�1) for

TBT and TPhT were 13.1 and 5.0%, respectively. The developed method was an easy and fast way to

analyze TBT and TPhT in water samples. Copyright # 2009 John Wiley & Sons, Ltd.

In recent years, analytical methods for the environmental

analysis of organotin (OT) species have become necessary

because of the wide use of tributyltin (TBT) and triphenyltin

(TPhT) in anti-fouling paints and pesticides.1 Since TBT and

TPhT are highly toxic and can promote harmful effects on

aquatic organisms even at trace levels,2,3 analytical methods

for TBT and TPhT with high sensitivity, high selectivity and

simple operation are required.

Gas chromatography (GC) is often used for the determi-

nation of OTs4–7 because of its high resolution and the high

detector sensitivity. However, the GC analysis of OTs

requires derivatization prior to detection due to their poor

volatility. Alkylation by means of Grignard reactions from a

pre-extraction eluate or ethylation with NaBEt4 directly from

aqueous solution are the most common derivatization

procedures, but all are tedious and time-consuming.1 As

an alternative, liquid chromatography (LC) methods provide

simple and rapid techniques for the analysis of OTs in

various matrices since the OTs do not require to be

ndence to: Z. L. Chen, Centre for Environmental Risknt and Remediation, University of South Australia,Lakes, SA 5095, [email protected]

derivatized. Hence, LC coupled with various detection

techniques, such as mass spectrometry (MS) or inductively

coupled plasma mass spectrometry (ICP-MS), can be used

for the analysis of OTs at trace levels. Since ICP-MS provides

excellent sensitivity, selectivity, wide linear dynamic range

and the ability to perform isotopic analysis,8 LC/ICP-MS

coupled with various separation modes (ion-exchange,9,10

ion-pairing,11,12 and reversed-phase chromatography13,14)

has been used for the speciation of OTs. However,

the hyphenation of LC with ICP-MS is limited because the

organic-based mobile phase in LC is often not compatible

with the plasma source. For example, a high concentration of

organic solvents in the mobile phase results in poor plasma

stability and may even cause the plasma to be extinguished.

In addition, carbon deposits may form on the sampler and

the skimmer cones within the ICP-MS interface region, which

contributes to high noise and reduced signal intensities.1

Furthermore, it is often necessary to confirm the OT species

at trace level in real samples, as ICP-MS is only an element-

specific detector and it cannot provide the inherent

information required to characterize or confirm specific

OT species. However, these issues can be solved by using

LC/MS, which provides molecular information to confirm

Copyright # 2009 John Wiley & Sons, Ltd.

3796 Q. Sun et al.

OT species in addition to a quantitative response. Thus

LC/MS is an attractive technique for the analysis of trace

components in complex matrices.15 Generally, both electro-

spray ionization (ESI) and atmospheric pressure chemical

ionization (APCI) can be interfaced successfully with MS

to detect OTs, where OT compounds follow the same

fragmentation pattern, which is characterized by the

progressive loss of the organic groups bound to the Sn.1

Determination of butyl- and phenyltin in a certified reference

sediment material using LC/APCI-MS has been reported.16

Similarly, LC/APCI-MS was used to determine TBT and

4-hydroxybutyldibutyltin (OHBuDBT) in seawater with

detection limits of 35 and 26mg L�1, respectively.17 Positive

ion LC-/ESI-MS was used to determine TBT and TPhT in

seawater, and off-line SPE was used to improve the detection

limit to the low ng L�1 level and to reduce matrix effects.18

Recently, LC/ESI tandem mass spectrometry (MS/MS) was

used for the identification and quantification of OTs by using

the high selectivity and sensitivity of the multiple reaction

monitoring (MRM) mode, achieving good accuracy and

detection limits of approximately 1 pg for TBT.19 These LC/

MS methods are useful, but LC/MS has rarely been applied

to environmental OT analysis because of its poor detection

limits and because matrix effects lead to mass signal stability.1

Several different sample preparation methods have been

reported for TBT and TPhT. Solid-phase microextraction

(SPME),20 headspace single-drop,21 liquid-phase microex-

traction,22 headspace solid-phase extraction,23,24 and liquid-

liquid extraction25 have all been successfully used for sample

preparation and analysis by GC-based methods. For LC-

based methods, accelerated solvent extraction (ASE),26,27

microwave-assisted extraction (MAE),27 and solid-phase

extraction (SPE)18,28,29 have also been successfully used for

sample preparation for TBT and TPhT. Of these techniques,

SPE has the advantage of not only achieving OT enrichment,

but also cleaning up interferences,18 thus improving detec-

tion sensitivity and reducing any matrix effects in LC/MS.30

Using an off-line SPE approach, cation-exchange SPE29

and C18 SPE18 have been applied to extract OTs from

seawater, resulting in good recoveries and high sensitivities

with low ng L�1 detection limits. However, these methods

can be time-consuming, cumbersome to perform, and only a

small part of the extract is injected.30 In contrast, on-line SPE

has higher sample throughput, allows faster analysis, and

can achieve automation of all parts of the analytical method,

including sample preparation.30 Despite these advantages, to

date only one study has reported the use of on-line SPE for

the pre-concentration of OTs, followed by HPLC with post-

column derivatization and fluorimetric detection.31

In this study, an on-line SPE method was developed for the

pre-concentration of TBT and TPhT at trace level and cleanup

of interferences, followed by LC/MS for the separation and

detection of TBT and TPhT. To achieve method optimization:

(1) the conditions for on-line SPE, such as sample pH and

injection volume, were evaluated; (2) the conditions for the

separation and detection, including the mobile phases, the

ionization mode and the operation parameters, were

investigated; and (3) the proposed method was demon-

strated for the determination of TBT and TPhT in spiked

contaminated lake waters and seawaters.


EXPERIMENTAL

ChemicalsAll chemicals, including tributyltin chloride and triphenyltin

chloride, were analytical grade reagents (Sigma-Aldrich Pty.

Ltd, Castle Hill, NSW, Australia). The mixed standard was

prepared daily from a 2000mg mL�1 stock solution. Milli-Q

water (Milli-Q plus system, Millipore, Bedford, MA, USA),

with a specific resistance of 18.2 MV cm�1, was used to

prepare all solutions. The mobile phases used in LC/MS

were prepared by dissolution of the appropriate amounts of

acetic acid (HAc) and ammonium acetate (NH4Ac) or formic

acid (HFc) and ammonium formate (NH4Fc), and methanol

in Milli-Q water. The mobile phases were filtered through a

disposable 0.45mm cellulose acetate membrane filter (Milli-

pore) and degassed in an ultrasonic bath prior to use.

Water samples were collected in pre-cleaned 250 mL glass

bottles from Mawson Lakes (freshwater) and Glenelg beach

(seawater), South Australia, and stored at 48C immediately

after sampling. Prior to analysis, water samples were filtered

using 0.45mm pore size, mixed cellulose membranes

(Millipore) and acidified to pH 2.7 with HFc solution to

achieve compatibility with the initial mobile phase con-

ditions. Analysis was carried out within 24 h after sampling

to avoid adding chemical preservatives and to minimize

potential degradation and transformation of the analytes. All

samples were tested in triplicate.

InstrumentAn Agilent 1100 sample pre-concentration system (Agilent

Technologies, Palo Alto, CA, USA), consisting of a

quaternary pump, an on-line degasser, an autosampler

and a six-port switch valve was used. Agilent C18 cartridges

(2.1 mm� 20 mm, 3.5mm) were employed for on-line SPE. A

schematic diagram of the on-line SPE sample enrichment

system coupled to LC for the determination of the trace OTs

in water is shown in Fig. 1. A filtered sample (3.0 mL) was

injected by the G2258A dual loop autosampler at 10 mL

min�1. For loading (Fig. 1, position A), pump A (the on-line

SPE pump) delivered the eluent (0.1% HCOOH in Milli-Q

water) through the SPE cartridge for 1.5 min at a flow rate of

2.5 mL min�1 after sample injection. The flow rate was

subsequently decreased to 0.8 mL min�1 over a period of 30 s,

in preparation for the SPE cartridge being switched in-line

with the LC pump (pump B), where the flow rate was 0.8 mL

min�1. The six-port valve was then switched to the inject

position (Fig. 1, position B) at 2.0 min. While in the inject

position, pump B delivered LC buffers to elute the analytes

and transfer them to the separation column, while pump A

pumped to waste. The valve was switched back to the

loading position at 22.0 min to clean the SPE cartridge using

pump A to deliver 100% methanol for 3 min and then to

purge the cartridge with 0.1% HFc in Milli-Q water in

preparation for the next sample.

The separation of TBT and TPhT was performed on

reversed-phase columns (Zorbax, XDB-C18, 150 mm� 3.0 mm

i.d., 5mm, Agilent). The samples were injected using an 1100

autosampler with an injection volume of 20mL. The mobile

phase flow rate was 0.8 mL min�1 and the column

temperature was 308C. The mobile phases used in this study


DOI: 10.1002/rcm

Figure 1. Schematic diagram for on-line SPE-HPLC/MSD:

(A) loading position and (B) injection position

Determination of trace tributyltin and triphenyltin in water 3797

were: mobile phase A, 0.1% HFc, 5 mM NH4Fc in

10:90 MeOH/H2O (v/v); mobile phase B, 0.1% HFc, 5 mM

NH4Fc in 90:10 MeOH/H2O (v/v). ESI and APCI were tested

in both positive and negative ion modes. The operating

parameters for both ESI and APCI are shown in Table 1. Data

were collected using full scans mode (m/z 100–500). The

Agilent Chemstation software package was used to control

the on-line SPE-LC/MS system and to process data.

RESULTS AND DISCUSSION

Separation of TBT and TPhTTwo different water/methanol mobile phases, one contain-

ing 0.1% (v/v) HFc, 5 mM NH4Fc, and another containing

0.1% (v/v) HAc, 5 mM NH4Ac, were compared for the

separation of TBT and TPhT by directly injecting 20mL of a

5 mg L�1 mixed standard. Initial results showed that the TBT

and TPhT could be separated well using both mobile phases

by gradient elution (Fig. 2), where water/methanol (10:90,

v/v) was changed to water/methanol (90:10, v/v) over

5 min, and these concentrations were maintained for 20 min.

The use of 0.1% (v/v) HFc, 5 mM NH4Fc, where the higher

concentration of HFc overcomes the problems of absorption

to the column, leads to better peak shape. Both TBT and TPhT

Table 1. Operating parameters for ESI and APCI

ESI APCI

Drying gas flow rate (L min�1) 12 5.0Nebulizer gas pressure (psi) 35 60Drying gas temperature (8C) 350 350Vaporizer temperature (8C) 325 325Capillary voltage (V) �4000 �4000Corona current (mA) / 15Fragmentor voltage (V) 70 90


are present as cations in acidic solutions,32 and these cationic

species are more easily eluted from a reversed-phase column.

In addition, higher signal intensity was observed using 0.1%

(v/v) HFc, 5 mM NH4Fc, mainly because of the formation of

formic adducts, such as [Mþ2Fc]�. Hence, 0.1% (v/v) HFc,

5 mM NH4Fc was chosen as the mobile phase for the

remainder of the experiments in this study.

Optimization of API-MS parametersAPCI and ESI were compared as the ionization source in both

positive and negative ion modes. The degree of fragmenta-

tion and formation of adducts was greatly affected by the

choice of APCI or ESI interface and the polarity (Table 2). It

should be noted that the assignment of the weaker ions in all

these spectra can only be tentative because they all contain

the characteristic tin clusters, based the ten tin stable isotopes,

of which 120Sn (32.58%), 118Sn (24.22%), 116Sn (14.54%), 119Sn

(8.59%) and 117Sn (7.68%) are the most prominent. The ions

cited below are all based on the 120Sn isotope.

TBT and TPhT yield similar ESI and APCI spectra in the

negative ionization (NI) mode, characterized by strong

signals for [TBTþ2Fc]� at m/z 381.1 and [TPhTþ2Fc]� at m/z

441.0. In addition, weak signals for [TPhT]� at m/z 351.0 and

[TPhTþClþFc]� at m/z 431.0 were observed, while another

weak signal at m/z 371.0 which we tentatively assign to

[TBTþClþFc]� was observed for TBT using the APCI

interface.

In comparison, as shown in Table 2, more ions were

observed from these OTs in the positive ionization (PI) mode

than in the NI mode. For example, in both ESI and APCI, TBT

shows the most prominent ion [TBT]þ at m/z 291.1, accom-

panied by fragment ions corresponding to the loss of one or

two butene groups, [TBT–C4H8]þ at m/z 235.0 and [TBT–

2�C4H8]þ at m/z 179.0, as well as ions that we assign as the

formate, methanol and ammonia adducts, [TBTþHFcþFc]þ

atm/z 382.2, [TBTþMeOH]þ atm/z 323.1, and [TBTþNH3]þ at

m/z 308.1. In the TPhT spectrum, the most abundant ion was

[TPhT]þ at m/z 351.0. Methanol and ammonia adducts,

[TPhTþMeOH]þ at m/z 383.1 and [TPhTþNH3]þ at m/z 368.1,

were also observed. ESI also showed ions at m/z 442.1, 428.1

and 419.0, corresponding to [TPhTþHFcþFc]þ, [TPhTþMeOHþFc]þ and possibly [TPhTþNaþFc]þ, respectively. Similar

ions for TBT and TPhT were observed using LC/APCI-MS

for the analysis of organotin compounds.13

Higher signal intensities for TBT and TPhT were obtained

in the NI mode when either APCI or ESI was interfaced with

the mass spectrometer (Fig. 3). For example, with ESI the

signal intensity for TBT and TPhT in the NI mode was 12.7

and 4.2 times, respectively, more than that obtained in the PI

mode. Thus, the NI mode was chosen for the remaining

studies. ESI was also more sensitive than APCI. The signal

intensities for TBT and TPhT were 8.4 and 5.9 times higher in

ESI than in APCI.

On-line SPE-LC coupled with ESI and APCI-MS was

evaluated for the quantification of tin species by determining

standard calibrations in the ranges 0.1–30mg L�1. The slope

of the calibration curve from ESI was 14.5 and 6.5 times

greater than from APCI for TBT and TPhT, respectively. ESI-

MS in the NI mode gave the highest signal intensity and


DOI: 10.1002/rcm

min2.5 5 7.5 10 12.5 15 17.5 20 22.5

0

10000

20000

30000

40000

50000

60000

70000

80000

MSD1 469, EIC=468.7:469.7 (C:\HPCHEM\1\DATA\080819SN\H3.D) APCI, Neg, Scan, Frag: 90

TPhT

min2.5 5 7.5 10 12.5 15 17.5 20 22.5

0

5000

10000

15000

20000

25000

30000

35000

40000

45000


TBT

min2.5 5 7.5 10 12.5 15 17.5 20 22.5

0

20000

40000

60000

80000

100000


TPhT

min2.5 5 7.5 10 12.5 15 17.5 20 22.5

0

10000

20000

30000

40000

50000

60000


TBT

(A)

(B)

Figure 2. Separation of TBTand TPhTusing different mobile phases by LC/APCI-MS. The separation was

achieved on a XDBC18 reversed-phase columnwith a gradient from 10%B to 90%B in 5min, maintained at

90% B for 20min. The mobile phase for (A) is mobile phase A, 0.1% HAc, 5mM NH4Ac in 10:90MeOH/

H2O (v/v); mobile phase B, 0.1% HAc, 5mM NH4Ac in 90:10MeOH/H2O (v/v); The mobile phase for (B) is

mobile phase A, 0.1% HFc, 5mM NH4Fc in 10:90MeOH/H2O (v/v); mobile phase B, 0.1% HFc, 5mM

NH4Fc in 90:10MeOH/H2O (v/v). The operating parameters for APCI were capillary voltage: �4000V,

corona current: 15mA, fragmentor voltage: 90V, dryin gas flow rate: 5.0 L min�1, nebulizer pressure: 60 psi,

drying gas temperature: 3508C, and vaporizer temperature: 3258C.

3798 Q. Sun et al.

strong signals for [TBTþ2Fc]� at m/z 381.1 and [TPhTþ2Fc]�

at m/z 441.0 and was subsequently used in the following

study.

The applied cone voltage and capillary voltage are usually

the most significant parameters affecting the enhancement or

inhibition of fragmentation, so these parameters were

initially adjusted to give optimized signal intensities.13 The


effect of applied cone voltage on the signal intensity of TBT

and TPhT was investigated in the range 50–130 V. The signals

for both TBT and TPhT reached their highest intensity at

70 V. As the applied cone voltage increased, the analytes

fragmentation increased, and characteristic fragment ions

were observed. The effect of the capillary voltage on the

signal intensity was examined in the range 2500–4500 kV.


DOI: 10.1002/rcm

Table 2. Assignment� of ions detected for TBT and TPhT from ESI and APCI in negative and positive ionization modes

Positive Negative

ESI APCI ESI APCI

ion m/z % ion m/z % ion m/z % ion m/z %

TBT [TBT]þ 291.1 100 [TBT]þ 291.1 100 [TBTþ2Fc]� 381.1 100 [TBTþ2Fc]� 381.1 100[TBTþHFcþFc]þ 382.2 29 [TBTþNH3]þ 308.1 42 [TBTþClþFc]� 371.0 6[TBT-Butene]þ 235.0 18 [TBTþMeOH]þ 323.1 35[TBTþMeOH]þ 323.2 11 [TBT-Butene]þ 235.0 18

[TBT-2xButene]þ 179.0 14 [TBT–2�butene]þ 179.0 14TPhT [TPhT]þ 351.0 100 [TPhT]þ 351.0 100 [TPhTþ2Fc]� 441.0 100 [TPhTþ2Fc]� 441.0 100

[TPhTþHFcþFc]þ 442.1 59 [TPhTþMeOH]þ 383.0 88 [TPhT]� 351.0 6 [TPhTþClþFc]� 431.0 4[TPhTþMeOH]þ 383.1 41 [TPhTþNH3]þ 368.0 30 [TPhTþClþFc]� 431.0 4 [TPhT]� 351.0 3[TPhTþNH3]þ 368.1 14

[TPhTþMeOHþFc]þ 428.1 14[TPhTþNaþFc]þ 419.0 11

� Assignment of some of the weaker ions is only tentative.

Figure 3. Relative abundances of the most abundant ions of

TBTand TPhT in different ionizationmodes at a concentration

of 5.0mg L�1. The mobile phase contained 0.1% HFc, 5mM

NH4Fc in methanol and Milli-Q water solution. The operating

parameters for ESI and APCI were the same as in Table 1.


However, the signal intensity was not significantly affected

by the capillary voltage, and 4000 V was chosen as the

optimal value.

One of the major issues in ESI-MS for quantitative analysis

is matrix effects since the ESI source is highly susceptible to

other components present in the matrix, which may result in

signal suppression or enhancement, leading to erroneous

results.33,34 On-line SPE sample cleanup and chromato-

graphic separation is one of the preferred strategies to reduce

matrix effects.34 In this work, matrix effects were evaluated

because of the widely reported issues of MS signal

suppression or enhancement effects when complex matrices

are analyzed.35 The MS response obtained from standard

Table 3. Characteristics for TBT and TPhT determined using the

Analytes Calibration range (mg L�1) Regression equation

TBT 0.1–30 y¼ 92912xþ 8031.2TPhT 0.1–30 y¼ 72575xþ 55856

a Injection of a 10mg L�1 mixed standard (n¼ 4).


solutions prepared in Milli-Q water was compared with

those measured in spiked (10mg L�1) lake surface waters,

where both standard solutions and samples were introduced

to the on-line SPE system for sample pre-concentration and

cleanup, LC separation and ESI-MS detection. The results

showed that the abundances of TBT and TPhT were

7.30� 105 and 8.95� 105, respectively, using spiked lake

water (10mg L�1), while abundances of 9.02� 105 and

9.17� 105 were obtained using Milli-Q standards. The

response from lake water for TBT and TPhT was reduced

to 80.95% and 97.70%, respectively, compared with that

using Milli-Q water. This clearly showed that signal

suppression occurred when lake waters were used as the

sample matrix. This effect was mainly attributed to the

relative high total organic carbon (TOC) in lake water

(37.77 mg L�1), since organic compounds can be transferred

to the ESI-MS system together with the analytes, and cause

matrix effects and signal suppression. However, use of on-

line SPE reduced the matrix effect and the suppressions to

below 20% and therefore on-line SPE was essential and

effective for the detection of TBT and TPhT. Further

avoidance of any matrix effects in the quantification of these

species was achieved by using standard addition calibration

in the following study.36

Optimization of the on-line SPE conditionsIn the development of any on-line SPE procedure,

parameters such as pH and injection volume can affect the

extraction efficiency and need to be optimized. In this

study, C18 pre-columns were used for the on-line enrichment

of trace levels of TBT and TPhT (3.0 mL of standard solution

containing 30mg L�1) because they are non-selective

materials.18 The results indicated that C18 pre-columns can

proposed method

s R2 Detection limits (mg L�1) RSD (%)a

0.9918 0.02 13.080.9873 0.02 5.02


DOI: 10.1002/rcm

Figure 4. (A) Mass spectra of TBT in ESI NI mode from spiked Mawson Lakes water. (B) Mass spectra of TPhT in ESI NI

mode. (3.0mL filtered lake water was injected to on-line SPE and separated on a XDB-C18 reversed-phase column). The

operating parameters for APCI were capillary voltage: �4000V, corona current: 15mA, fragmentor voltage: 90V, drying

gas flow rate: 5.0 L min�1, nebulizer pressure: 60 psi, drying gas temperature: 3508C, and vaporizer temperature: 3258C.

3800 Q. Sun et al.

be used as the sorbent for on-line enrichment. However,

since sample pH relates directly to the ionization of analytes

and affects both water-solubility and extractability,37 it is an

important parameter for the effective extraction of analytes

during SPE.38 The effect of sample pH on extraction

efficiency was evaluated using 30mg L�1 standard solutions

in the pH range from 2.73 to 6.80. The extraction efficiency

decreased with increasing sample solution pH, and lower

sample solution pH yielded significantly higher SPE


extraction efficiency. This was due to TBT and TPhT bonding

with the C18 cartridge during SPE by hydrophobic inter-

actions as the pKa values of TBT and TPhT are 6.25 and 5.2,

respectively.32 A pH of 2.73, which is the pH of a 0.1% (v/v)

HFc solution, was chosen as optimal.

The injection volume, which is the volume which can be

reliably passed through the on-line SPE cartridge with

maximum extraction efficiency, to obtain sensitivity, but

without breakthrough occurring, is a key parameter. The


DOI: 10.1002/rcm

Table 4. Characteristics for TBT and TPhT determined in lake water and seawater

SamplesVolume

(mL)Initial concentration

(mg L�1)Spiked concentration

(mg L�1) nTBT recovery

(%)TPhT recovery

(%)

lake water 3000 n.d. 10 3 81.0� 6.8 97.7� 3.8seawater 3000 n.d. 10 3 91.5� 2.1 101.9� 2.0

n.d.: not detected.


effect of injection volume on sensitivity was investigated by

injecting a range of volumes (2.0–5.0 mL) of 30mg L�1

standard solutions (the highest concentration of the cali-

bration) at an injection speed of 10 mL min�1. The results

showed that the injection volume significantly affected the

sensitivity of the method. The signals for both TBT and TPhT

greatly increased as the injection volume increased from 2.0

to 3.0 mL, and slightly increased as the injection volume

increased from 3.0 mL to 4.0 mL, decreasing when 5.0 mL

was injected. As TBT and TPhT easily form cations in water,

the OTs that initially bonded to the SPE sorbent can be eluted

by later sample solution. In addition, an increased injection

volume can also increase matrix effects. The optimal injection

volume was therefore 3.0 mL.

The injection speed and the delivery speed can potentially

also affect sensitivity. The injection speed is the rate at which

the plunger ejects sample from the metering device, while the

delivery speed is the rate at which the on-line SPE pump

delivers the mobile phase. The effects of both injection speed

and on-line SPE deliver speed on sensitivity were tested by

injecting 3.0 mL of 30mg L�1 standard solutions at a range

of 1.0–10.0 mL min�1 and 0.8–2.5 mL min�1, respectively.

Neither of these parameters significantly affected the

sensitivity of the method. Thus, the default injection speed

of 10 mL min�1 and the highest on-line SPE delivery speed of

2.5 mL min�1 were chosen.

Using the optimal conditions, the targets were eluted

directly by an LC mobile phase using a 20 min gradient

elution program. At the same time, the SPE cartridge was

cleaned by methanol and conditioned for the next sample.

The proposed method can achieve pre-concentration,

separation and simultaneous detection of TBT and TPhT

in 25 min without sample pretreatment.

Analytical performance and analysis ofreal samplesThe analytical performance characteristics of the proposed

method are summarized in Table 3. For quantification,

standard addition was used by adding five known standard

solutions in the range 0.1 to 30mg L�1 to the unknown

samples. Calibration curves for quantification were obtained

by plotting peak area versus the concentration of the

corresponding target species. All calibrations were linear

over a concentration range 0.1–30mg L�1 with correlation

coefficients greater than 0.987. The detection limits of both

TBT and TPhT were 0.02mg L�1. The reproducibility (n¼ 4),

from injection of a 10mg L�1 standard solution, showed that

the relative standard deviations (RSDs) were 13.08 and 5.02

for TBT and TPhT, respectively. These results demonstrated

good accuracy and precision.

The proposed method was applied to the detection of the

TBT and TPhT in the water samples, including lake water


(LW) and seawater (SW), which were analyzed in triplicate.

However, no TBT and TPhT were detected in these samples,

indicating there was either no TBT or TPhT in theses samples,

or that the concentrations of TBT or TPhT were below the

detection limits of the method. To validate the method, both

lake water and seawater were spiked with 10mg L�1 TBT and

TPhT, and shaken in an end-over-end shaker for 24 h. The

recoveries of TBT and TPhT for seawater were 91.5% and

101.9%, and for lake water were 81.0% and 97.7%,

respectively (Table 4). The ESI-MS spectra (Figs. 4(A)

and 4(B)) confirmed the presence of TBT and TPhT in the

samples. Characteristic clusters that included the ten tin

stable isotopes were obtained.

Compared with other reported techniques, the developed

method is simpler to perform since no sample pre-treatment

is needed, and has higher sample throughput which reduces

the required sample volume. The detection limits are in the

range of reported LODs for triorganotin analysis by LC/MS,

which were 35mg L�1 for TBT by pre-concentration from

5 mL seawater using liquid-liquid extraction,17 0.016mg L�1

and 0.008mg L�1 for TBT and TPhT, respectively, by pre-

concentration from 250 mL seawater using off-line SPE18

compared with 0.02mg L�1 for TBT and TPhT obtained using

the proposed method when a 3.0 mL sample was enriched on

a C18 pre-column.

CONCLUSIONS

In this study, on-line SPE-LC/ESI-MS provided a rapid and

sensitive method for the analysis of TBT and TPhT in water

samples. On-line SPE was successfully used for sample

cleanup and pre-concentration of TBT and TPhT at trace

levels to reduce matrix effects and improve detection

sensitivity, and ESI-MS was used for the identification of

TBT and TPhT at trace levels in complex samples. A mobile

phase containing 0.1% HFc and 5 mM NH4Fc was suitable for

both LC separation and MS ionization. ESI and APCI, in both

PI and NI modes, were compared to improve the sensitivity

and specificity, with ESI in the NI mode showing the best

response. Under optimized conditions, detection limits for

both TBT and TPhT were 0.02mg L�1. The proposed method

was demonstrated by speciation of TBT and TPhT at trace

levels in spiked freshwater and seawater samples.

AcknowledgementsQian Sun gratefully acknowledges the State Scholarship

Fund of China for sponsoring the study in Australia as a

joint PhD candidate and the Centre for Environmental Risk

Assessment and Remediation for proving access to labora-

tory facilities. We thank Dr Catherine Dandie and Dr Gary

Owens for editing the manuscript.


DOI: 10.1002/rcm

3802 Q. Sun et al.

REFERENCES

1. Gonzalez-Toledo E, Compano R, Granados M, Prat MD.Trac-Trend. Anal. Chem. 2003; 22: 26.

2. Fent K. Crit. Rev. Toxicol. 1996; 26: 1.3. Antizar-Ladislao B. Environ. Int. 2008; 34: 292.4. Carlier-Pinasseau C, Astruc A, Lespes G, Astruc M.

J. Chromatogr. A 1996; 750: 317.5. De Smaele T, Moens L, Dams R, Samdra P, Van der Eycken J,

Vandyck J. J. Chromatogr. A 1998; 793: 99.6. Tao H, Rajendran RB, Quetel CR, Nakazeto T, Tominaga M,

Miyazaki A. Anal. Chem. 1999; 71: 4208.7. Uveges M, Abranko L, Fodor P. Talanta 2007; 73: 490.8. Montes-Bayon M, DeNicola K, Causo JA. J. Chromatogr. A

2003; 1000: 457.9. Yang L, Lam JWH. J. Anal. At. Spectrom. 2001; 16: 724.

10. Garcia-Alonso JI, Sanz-Medel A, Ebdon L. Anal. Chim. Acta1993; 283: 261.

11. Vela NP, Caruso JA. J. Anal. At. Spectrom. 1996; 11: 1129.12. Yang HJ, Jiang SJ, Yang YJ, Hwang C. Anal. Chim. Acta 1995;

312: 141.13. White S, Catterick T, Fairman B, Webb K. J. Chromatogr. A

1998; 794: 211.14. Chiron S, Roy S, Cottier R, Jeanot R. J. Chromatogr. A 2000;

879: 137.15. Rosen AL, Hieftje GM. Spectrochim. Acta B 2004; 59: 135.16. Rosenberg E, Kmetov V, Grasserbauer M. Fresenius J. Anal.

Chem. 2000; 366: 400.17. Bekri K, Saint-Louis R, Pelletier E.Anal. Chim. Acta 2006; 578: 203.18. Gonzalez-Toledo E, Compano R, Prat MD, Granados M.

J. Chromatogr. A 2002; 946: 1.19. Banoub JH, Miller-Banoub J, Sheppard GV, Hodder HJ.

Spectroscopy 2004; 18: 95.20. Bancon-Montigny C, Maxwell P, Yang L, Mester Z, Sturgeon

RE. Anal. Chem. 2002; 74: 5606.


21. Colombini V, Bancon-Montigny C, Yang L, Maxwell P,Sturgeon RE, Mester Z. Talanta 2004; 63: 555.

22. Shioji H, Tsunoi S, Harino H, Tanaka M. J. Chromatogr. A2004; 1048: 81.

23. Liu JY, Jiang GB. J. Agric. Food Chem. 2002; 50: 6683.24. Vidal JL, Vega AB, Arrebola FJ, Gonzalez-Rodriguez MJ, Sanchez

MC, Frenich AG. Rapid Commun. Mass Spectrom. 2003; 17:2099.

25. Zachariadis GA, Rosenberg E. J. Chromatogr. B Analyt. Tech-nol. Biomed. Life Sci. 2009; 877: 1140.

26. Wahlen R, Catterick T. Rapid Commun. Mass Spectrom. 2004;18: 211.

27. Wahlen R, Catterick T. J. Chromatogr. B Analyt. Technol.Biomed. Life Sci. 2003; 783: 221.

28. Jones-Lepp TL, Varner KE, Heggem D. Arch. Environ. Con-tam. Toxicol. 2004; 46: 90.

29. Ide K, Kohri M, Sato K, Inoue Y, Okochi H. Bunseki Kagaku1999; 48: 245.

30. Rodriguez-Mozaz S, Lopez de Alda MJ, Barcelo D.J. Chromatogr. A 2007; 1152: 97.

31. Gonzalez-Toledo E, Ortuno A, Compano R, Granados M,Prat M. Chromatographia 2002; 55: 19.

32. Arnold CG, Weidenhaupt A, David MM, Muller SR, Hader-lein SB, Schwarzenbach RP. Environ. Sci. Technol. 1997; 31:2596.

33. Zwiener C, Frimmel FH. Anal. Bioanal. Chem. 2004; 378:862.

34. Reemtsa T. J. Chromatogr. A 2003; 1000: 477.35. Antignac JP, de Wasch K, Monteau F, de Brabander H,

Andre F, le Bizec B. Anal. Chim. Acta 2005; 529: 129.36. Reemtsa T. Trac-Trend. Anal. Chem. 2001; 20: 533.37. Monteil-Rivera F, Beaulieu C, Deshamps S, Paquet L,

Hawari J. J. Chromatogr. A 2004; 1048: 213.38. Wang S, Huang W, Fang GZ, He JX, Zhang Y. Anal. Chim.

Acta 2008; 606: 194.


DOI: 10.1002/rcm

Documents

On-line solid-phase extraction coupled with liquid chromatography/electrospray ionization mass spectrometry for the determination of trace tributyltin and triphenyltin in water samples