Journal of Chromatography A, 1105 (2006) 208–212

Development of an improved heated interface for coupling solid-phasemicroextraction to high-performance liquid chromatography

Jose Carlos Rodrigues, Alvaro J. Santos Neto, Christian Fernandes,Claudete Alves, Alex S. Contadori, Fernando M. Lanc¸as∗

University of Sao Paulo, Institute of Chemistry at Sao Carlos,13566-590 Sao Carlos, SP, Brazil

Available online 27 December 2005

Abstract

The aim of the study described in this report has been the development and the evaluation of a new improved interface to be operated undercontinuous heating, for on-line coupling solid-phase microextraction (SPME) to high-performance liquid chromatography (HPLC). Heating isdesirable to increase desorption rate and decrease carryover. The results obtained have been compared with that obtained by off-line desorptionand online desorption without heating. The SPME–HPLC interface described here has an inner volume of 60�L, fixation for infinite points and an g the off-linem t performancea©

K

1

ep[ssmcpbpSgtmulc

in theucesg these. A

, thus,rptionn ratesame

ctaseffectrcialt thebingtheiece

tubeheat-tweened byr, in

0d

ovel leak less sealing system. When the heating system was used, the area values were almost 10-fold higher than that obtained usinode. The obtained chromatograms showed an increasing of the area and height of chromatographic peaks and proved the excellennd reproducibility of the interface developed in this work.2005 Published by Elsevier B.V.

eywords: SPME; Heated interface; SPME-HPLC

. Introduction

According to Chen and Pawliszyn[1] many industrially andnvironmentally important compounds, such as, pharmaceuticalroducts[2,3], food products[4,5], drugs[6], some pesticides

7–9] and polycyclic aromatic hydrocarbons (PAHs)[10], areemi or not volatile or have high polarity. So, they are bettereparated by liquid chromatography. Based on it, solid-phaseicroextraction (SPME) coupled with high-performance liquid

hromatography (HPLC) can be an useful tool for online sam-ling and determination of such compounds not amenable toe analyzed by SPME–GC (solid-phase microextraction cou-led to gas chromatography). The main difference betweenPME–GC and SPME–HPLC is the desorption procedure. Inas chromatography (GC) analyses, the fiber is introduced into

he injector of a gas chromatograph where the analytes are ther-ically desorbed from the fiber. The high temperature oftensed in thermal desorption of SPME fibers may cause prob-

ems such as degradation of the polymer. Therefore, in severalases, the desorption with solvent, in HPLC, is the best alterna-

tive. Temperature control enhances the desorption processSPME–HPLC interface. An increase in the temperature redviscosity and increases diffusion rates, thereby enhancinmass-transfer between stationary phase and mobile phahigher mass-transfer rate reduces band broadening andincreases the chromatographic efficiency. Increasing desotemperature is the simplest way to increase the desorptioand decrease the carryover for any analytes, following thebasic effects used in SPME–GC.

Years ago, Daimon and Pawliszyn[11] investigated the effeof heating the injector in SPME–HPLC. In this study it wfound that the desorption temperature have a pronouncedon the efficiency of desorption and separation. A commeSPME–HPLC interface was modified for this study, so thafiber inserted in the injector was extended into a piece of tuoutside the T-joint. This piece of tubing was heated, duringfiber exposure to the mobile phase, with a heating wire (a pof Ni–Cr with 0.1 mm diameter was used) coiled around theusing a dc power supply or a capacitive discharge for quicking. The temperature was measured by a thermocouple bethe heating wire and the stainless steel tubing and controllmanually adjusting the voltage of a power supply. Howeve

∗ Corresponding author. Tel.: +55 16 3373 9983; fax: +55 16 3373 9984.E-mail address: flancas@iqsc.usp.br (F.M. Lanc¸as).

that system conditions had been controlled very carefully forreproducibility.

021-9673/$ – see front matter © 2005 Published by Elsevier B.V.

oi:10.1016/j.chroma.2005.12.002

J.C. Rodrigues et al. / J. Chromatogr. A 1105 (2006) 208–212 209

Fig. 1. Structures of fluoxetine (a) and clomipramine (b).

In other hand, the total inner volume of the commercial des-orption chamber is about 200�L, requiring, at least, that volumeto transfer all the desorbed analytes to the column. This largevolume injection results in broadening of the peaks and loss inanalytical sensitivity and resolution.

The aim of the research described in this report has beenthe development and the evaluation of a new improved inter-face with lower inner volume, to be operated under continuousheating. In this new improved SPME–HPLC heated interface,analytes were completely desorbed with a minimum amount ofsolvent, avoiding, in this way, significant extra column disper-sion. The device described here has a much low inner volumechamber (approximately 60�L). The heating is homogeneouslydistributed for the whole body and a temperature controllerlinked to a solid-state relay rigorously controls the interfacetemperature.

The analyte used in this study was fluoxetine, (d,l-N -methyl-3-phenyl-3- [(�, �, �-trifluoro-p- tolyl) oxyl ] propyl-amine (Fig. 1a); clomipramine (Fig. 1b) was used as internalstandard.

2. Experimental

2.1. Chemicals and reagents

Fluoxetine (FLU) analytical standard was kindly sup-p deQ an-

dard, employed as internal standard (IS) (Fig. 1b) was providedby Sigma–Aldrich (Steinhein, Germany).

Ammonium acetate (Mallinckrodt, Paris, USA), aceticacid (Mallinckrodt), methanol (Mallinckrodt), acetonitrile(Mallinckrodt), sodium tetraborate (Reagen, Rio de Janeiro,Brazil), hydrochloric acid (J.T. Baker, Xalostoc, Mexico), andsodium chloride (Grupo Quımica, Rio de Janeiro, Brazil) werealso used. All solvents and reagents were of HPLC or analyticalgrade. A Milli-Q Ultra-Pure Water System (Millipore, Bedford,MA, USA) purified the water. Drug-free plasma was kindlydonated by Santa Casa de Sao Carlos (Sao Carlos, Brazil) andmaintained frozen at−20◦C.

Spherisorb C18 “pH stable”, a silica-based material forreversed-phase liquid chromatography (RPLC) (spherical parti-cles, average diameter, 3�m) from Phase Separations (Norwalk,USA) was used to pack the analytical column.

2.2. Instruments

A manual fiber holder for SPME, 60�m polydimethylsilox-ane/divinylbenzene (PDMS/DVB) SPME fibers were purchasedfrom Supelco (Bellefonte, PA, USA). A Shimadzu HPLC sys-tem (LC-10A) consisting of two pumps (LC-10ATVP), an oven(CTO-10ASVP), a fixed-wavelength ultraviolet detector (SPD-10AVVP), an auto injector (SIL-10AF), a system controller(SCL-10AVP), a degasser (DGU-14A) and an acquisition datas PLCa

2

vel-o ec ect-i andh mberw iam-e ,

F ace. tiod

lied by Dr. Maria Eugenia Q. Nassur (Departamentouımica-USP/RP). Clomipramine (CLOMI) analytical st

ig. 2. (A) Two-dimensional design (2D) of heating SPME–HPLC interfesorption chamber).

oftware Class-VP (Shimadzu, Japan), was used for all Hnalysis.

.3. Characteristics of the improved heated interface

A homemade interface for coupling SPME–HPLC was deped and built in our laboratory (Fig. 2). This new interfaconsists of a six-port valve Valco (Houston, TX, USA) conn

ng a 60�L (inner volume) homemade desorption chambereating block. Continuous heating of the desorption chaas achieved with a resistance cartridge of 25 W, 9.4 mm dter, 39 mm length purchased from Casa Ferreira (Sao Paulo

(B) Three-dimensional (3D) interface explosion view design (cross-secn of the

210 J.C. Rodrigues et al. / J. Chromatogr. A 1105 (2006) 208–212

Fig. 3. Six-port valve at (A) static desorption and (B) sample injection position.

Brazil). The temperature of desorption chamber and heatingblock were measured with a PT 100 type thermocouple pur-chased from Casa Ferreira (Sao Paulo, Brazil) and controlledby a temperature controller (model MDH002N-220VCA-P017)linked with a solid-state relay (model RSR) purchased fromTholz (Rio Grande do Sul, Brazil).

2.4. SPME–HPLC procedure

The PDMS/DVB SPME fiber was pre-conditioned with 5 mLof mobile phase in a vial for 15 min, under stirring. The con-ditioned fiber was exposed to a stirred FLU/CLOMI plasmasolution for 30 min in order to extract the analytes. Before insert-ing the fiber into the desorption chamber, the six-port valve wasplaced in the “static desorption” position (Fig. 3A) retaining themobile phase in the desorption chamber.

The fiber remained exposed into the static mobile phase inthe desorption chamber, causing a static desorption. Then, thevalve was switched to the “injection position” (Fig. 3B), andthe fiber was exposed to the flowing mobile phase. The desorp-tion chamber was operated at both room temperature and 50◦C.The measured temperature was exactly that of the mobile phase,since the temperature was measured by a thermocouple placedin the heating block of the desorption chamber, after temperatureequilibrium was reached.

3

itionb se.o minep aturi tingi ase

lyted tinctw anal parto eated nsfe

ence from the fiber coating to the extraction solvent, improvinganalyte desorption.

The interface developed in this study presents a low volumechamber (60�L) for desorption of the analytes. This volumewas chosen since Chen and Pawliszyn[1] determined that thevolume needed to desorb PAHs from 7�m polydimethylsilox-ane (PDMS) fiber coating, without carryover phenomenon, wasfound to be less than 0.2�L (when 90 + 10 (v/v) of CH3CN–H2Owas used as solvent). In order to improve the desorption pro-cess, it is required not only to increase the solubility but alsoto increase the diffusion coefficient of the analytes between thestationary phase and the solvent.

The chromatograms obtained through the analyses ofFLU/CLOMI mixtures using SPME extraction/injection via off-line and online modes, at both room and higher temperature(50◦C) are shown inFig. 4. In this Figure, the heating effecton desorption and chromatographic performance can be seen.The dashed line illustrates a chromatogram obtained in off-linemode at ambient temperature; the dotted line illustrates a chro-matogram obtained in online mode at room temperature; and

Fe .D houth

. Results and discussion

SPME involves a diffusion process in which analytes partetween a sample phase and a polymeric stationary phaccurs when the fiber is exposed to a sample for a predetereriod of time. As a consequence of the desorption temper

ncrease, the diffusion coefficient of the analytes in the coancreases and the coating-liquid distribution constant decre

SPME interface, built in our laboratory, improved anaesorption, when compared to off-line process, in two disays. Firstly, because the whole amount of the desorbed

yte is injected, different from off-line process, where only af the solution is injected. Secondly, the interface has a hesorption chamber, which increases the analyte mass tra

Itde

s.

-

dr-

ig. 4. Chromatograms of FLU (500 ng mL−1) and CLOMI (1000 ng mL−1)xtracted under the following conditions: 30 min, 50◦C, without salt, pH 9.0esorption was carried out in off-line mode and in on-line with and witeating mode.

J.C. Rodrigues et al. / J. Chromatogr. A 1105 (2006) 208–212 211

the straight line represents a chromatogram in online mode at50◦C. Excellent retention time reproducibility was obtainedwith this new interface (Fig. 4). The chamber volume wasreduced to 60�L, eliminating the potential problem of signifi-cant extra-column dispersion caused by a large injected volume(ca. 200�L) has founded in commercially available chamber.The sample injection band was also narrowed resulting in higherdetection sensitivity and reproducibility. In addition, the devicepresented here allows using simultaneously two mobile phasesor one mobile phase and one cleaning solvent, because of itsexclusive second inlet.

Blanks were run between all samples and no peaks weredetected indicating no carryover phenomenon when using thisnew SPME–HPLC interface. By using the heated SPME–HPLCinterface, retention times of FLU/CLOMI are the same than thatobtained using SPME–HPLC without heating.

3.1. Heated desorption

Since the stainless steel chamber has a low inertia and alow specific heat, the temperature of desorption chamber canbe changed rapidly and effectively with adequate reproducibil-ity.

The SPME methods using off-line and on-line desorptionwithout and with heating were compared.Fig. 4 shows theresults obtained using 30 min desorption at room temperaturea vedt eres . Thec opti-m rmald -o

orp-t e tot iszyn[ om-p atingw linem s wea

omeb rac-t uide ceda logic MEa aredw

3

3el-

o angw -c e w

Table 1Concentration and RSD values obtained in the evaluation of within- andbetween-day precision (n = 5)

Concentration(ng mL−1)

Within-day precision(%RSD)

Between-daysprecision (%RSD)

25 3.5 16.850 4.5 n.e.a

100 2.3 11.0200 3.6 n.e300 2.8 n.e.500 3.3 6.9

a n.e., not evaluated.

higher than 0.997. The linear regression equation obtained was:Y = 0.00398x − 0.02126.

3.2.2. PrecisionThe RSD obtained in the evaluation of within day was smaller

than 5% in all evaluated concentrations. In the evaluation ofbetween days RSD was smaller than 17% in all concentrationsevaluated. These results demonstrate that the developed methodhas an adequate precision (Table 1).

4. Conclusion

The modified interface has been successfully applied tothe analysis of a drug in plasma samples. The analysis ofFLU/CLOMI with heating was obtained with improvementin the detection sensitivity, and in qualitative and quantita-tive results when compared to the SPME–HPLC off-line andinterface without heating. Results showed that the total amountdesorbed from the fiber increased with the increasing of the des-orption temperature, especially for compounds highly retainedin ambient temperature, as basic compound. When the heatingsystem was used the area values were almost 10-fold higherwhen compared to off-line mode without heating.

The chromatograms obtained show an increase of the chro-matographic peak areas and peak heights and proved the excel-lent performance and reproducibility of the interface developeda acep anip-u stepi ion,w reals ainedw (SPEa temd drug( , sub-s ts org opet sitiv-i

A

portf and

nd at 50◦C with continuous heating. Temperature improhe method in distinct ways. The desorption conditions welected to decrease carryover to a minimum possible valuearryover decreased with the increase of temperature. Theal experimental conditions found in this assay, for the theesorption step were: interface temperature of 50◦C and desrption time of 30 min.

Extraction conditions were exactly the same for both desion methods. So, the difference in precision values is duhe desorption methods. As reported by Daimon and Pawl11], we found that heating improved the precision for all counds. Area values obtained with the interface without heere more than two-fold higher than that obtained in off-ode. When the heating system was used, the area valuelmost 10-fold higher when compared to off-line mode.

SPME techniques coupled on-line with HPLC has senefits over conventional techniques [liquid–liquid ext

ion (LLE), solid-phase extraction (SPE) and supercritical flxtraction (SFE)], such as, simplicity, low cost, use of redumounts of organic solvents, and reduced contact with bioal fluids[12]. On the other hand, the use of heating in SPllows a gain in sensitivity and a lower carryover when compith conventional SPME.

.2. Performance of the developed interface

.2.1. Linearity and rangeThe experiments employing FLU/CLOMI to test the dev

ped extraction system have shown adequate results. Ras evaluated from 25 to 500 ng mL−1. Correlation coeffiient between the concentration and the obtained respons

re

-

e

as

nd described in this work. In addition, this new interfresents a comprehensive heat system control of easy mlation and automation and low constructing cost. The next

n this investigation, after system optimization and validatill be the application of this method to the analyses ofamples, as well as perform a comparison of the results obtith those obtained using conventional extraction methodsnd LLE). A different approach will be the use of the syseveloped in this study for application in trace analyses offorensic analyses), as well as to detect anabolic steroidstances sometimes used illegally by people playing sporiven to an animal to affect their performance in a race (d

est), in biological fluids, due to the noticed increase in senty obtained when using this system.

cknowledgements

The authors gratefully acknowledge the financial suprom FAPESP (project no. 02/08409-0/PD-Br, 02/07075-1

212 J.C. Rodrigues et al. / J. Chromatogr. A 1105 (2006) 208–212

02/03039-0) and CAPES. We are grateful toElvio Caetano, fromthe electronic shop of the Institute of Chemistry at Sao Carlosfor his assistance in accomplishing the electronic project.

References

[1] J. Chen, J. Pawliszyn, Anal. Chem. 67 (1995) 2530.[2] S. Strano-Rossi, F. Molaioni, F. Botre, J. Anal. Toxicol. 29 (2005) 217.[3] Y. Wen, Y. Fan, M. Zhang, Y.Q. Feng, Anal. Bioanal. Chem. 382 (2005)

204.[4] C.G. Zamboni, L. Balest, G.E. De Benedetto, F. Palmisano, Talanta 66

(2005) 261.

[5] M.J. Jordan, K.L. Goodner, M. Castillo, J. Laencina, J. Sci. Food Agric.85 (2005) 1065.

[6] K.J. Chia, S.D. Huang, Anal. Chim. Acta 539 (2005) 49.[7] J.S. Aulakh, A.K. Malik, R.K. Mahajan, Talanta 66 (2005) 266.[8] P. Canosa, I. Rodriguez, E. Rubi, R. Cela, J. Chromatogr. A 1072 (2005)

107.[9] M.J. Gonzalez-Rodriguez, F.J.A. Liebanas, A.G. Frenich, J.L.M. Vidal,

F.J.S. Lopez, Anal. Bioanal. Chem. 382 (2005) 164.[10] R. Barro, S. Ares, C. Garcia-Jares, M. Llompart, R. Cela, J. Chromatogr.

A 1072 (2005) 99.[11] H. Daimon, J. Pawliszyn, Anal. Commun. 34 (1997) 365.[12] S.A. Poole, T.A. Dean, J.W. Oudsema, C.F. Poole, Anal. Chim. Acta

236 (1990) 3.