H Kataoka et aL Analytica Chimica Acta 655(2009)8-29 (A)Load position(extraction) (B)Inject position(desorption) PEEK tube Column connector Injection loop Capillary njection loop (0) Capillary capillary njection Metering column needle n Au oler Waste d Waste° Mobile phase DA LC column Detector Workstation Detector Workstation Fig. 4. Schematic diagram of automated on-line in-tube solid-phase microextraction(draw/eject extraction system)(A)Load position(extraction).(B)inject position (desorption). DAD is photodiode array detector and MSD is mass detector. with Mili-Q water(1.5 mLmin-1)to remove remaining matrix and The computer controls the drawing and ejection of sample solu norganic residues from the capillary. For desorption( Fig. 3B), the tion, switching of the valves, control of peripheral equipment such x-port valve is switched to the" INJECT"position, and the lc elu- as the HPLC and mSD, and analytical data processing, thus redue acetonitrile-Mili-Q water containing 1% acetic acid(85+15), ing labor and enhancing precision. In addition, a large number of is passed through the column for 4 min. For this step, the flow-rate samples can be autor lly processed by the autosampler with- of the lC pump was reduced to 100 HLmin-I to reduce the back out carryover, because the injection needle and capillary column pressure of the analytical column on the capillary. The desorbed are washed in methanol and the mobile phase before the sample is analytes are transferred to the analytical column for separation and extracted. detected using a UV or tandem mass selective detector(MS-MS). In a similar system, using double column switching valves [78]. the 2.2.2. Optimization of parameters extraction and analysis segments were independent, enabling In-tube SPme depends on the distribution coefficient of each pid, simultaneous performance of several runs, thus shortening analyte as well as its affinity for the fiber SPME, making it important and the time for the whole analysis can be shortened. These sys- the rapidity and efficiency of extraction s y phase to optimize analysis time, with the separation process of the previous sample to raise the distribution factor in the stationa tems, however, may cause some systematic problems, including ciency of the extraction depend on the extraction rate the sample contamination of the switching valve with sample solution because volume, the ph of the sample the type of stationary phase, and the capillary column is directly fixed on the six-port valve. This may the internal diameter, length, and film thickness of the capillary result in inaccurate quantitative information and overestimation of column. Several commercially available capillary columns, which differ In the draw/eject extraction system(Fig. 4), an extraction capil- according to the selectivity of the stationary phase, internal diam ry column is placed between the injection loop and the injection eter, length and film thickness, have been developed. For example, needle of the HPLC autosampler. The capillary connections are facil- a low polarity column with a methyl silicon liquid-phase selec itated by using a 2.5-cm long sleeve of 1/16in polyether ether tively retains hydrophobic compounds, whereas a high polarity ketone(PEEK)tubing at each end of the capillary fixed with 1/16- column with a polyethylene glycol liquid-phase selectively retains in SS unions(0. 25 mm bore stainless steel nuts)and ferrules. An hydrophilic compounds. Since the internal diameter, length and injection loop is installed to prevent sample contamination of the film thickness of the column and other dimensions affect the metering pump and switching valve Building in UV, diode arrays amount of sample that can be loaded and the amount of compound or fluorescence detectors between the HPLC and the MsD, can that can be extracted these parameters should be chosen carefully. enhance the multi-dimensional and simultaneous multidetections, The thin(usually <1 um)coating in such capillaries often results in improving analyte identification. As shown in Fig. 4A, a computer low stationary phase loading reducing sample capacity and extrac controls the injection syringe, which repeatedly draws and ejects tion sensitivity, although extraction equilibrium is quickly attaine sample from the vial, with the analytes partitioning from the sam in the relatively thin layers. Although increasing the film thickness ple matrix into the stationary phase until equilibrium is almost of the stationary phase may solve this problem, it is extremely dif- reached. Subsequently, the extracted analytes can be directly des- ficult to reliably immobilize thicker coatings using conventional orbed from the capillary coating by mobile phase flow or by an approaches, and conventionally prepared GC coatings do not bind A Pirated desorption solvent after switching the six-port valve chemically to the fused-silica capillary inner surface. This lack ig. 4B). It is therefore necessary to prevent plugging of the cap- of chemical bonds is mainly responsible for low solvent stabil- illary columns and flow lines during the extraction, as well as to ity, preventing effective hyphenation of in-tube SPME techniques remove particles from samples by filtration before extraction. The that employ nic or organo-aqueous mobile phases. In add desorbed analytes are then transported to the HPLC column for sep- tion, although large amounts of compound can be extracted their aration and detected using a UV or mass selective detector(MSD). quantitative desorption from capillary columns may be difficult by
H. Kataoka et al. / Analytica Chimica Acta 655 (2009) 8–29 13 Fig. 4. Schematic diagram of automated on-line in-tube solid-phase microextraction (draw/eject extraction system). (A) Load position (extraction), (B) inject position (desorption). DAD is photodiode array detector and MSD is mass detector. with Mili-Q water (1.5 mL min−1) to remove remaining matrix and inorganic residues from the capillary. For desorption (Fig. 3B), the six-port valve is switched to the “INJECT” position, and the LC eluant, acetonitrile–Mili-Q water containing 1% acetic acid (85 + 15), is passed through the column for 4 min. For this step, the flow-rate of the LC pump was reduced to 100 L min−1 to reduce the back pressure of the analytical column on the capillary. The desorbed analytes are transferred to the analytical column for separation and detected using a UV or tandem mass selective detector (MS–MS). In a similar system, using double column switching valves [78], the extraction and analysis segments were independent, enabling the rapid, simultaneous performance of several runs, thus shortening analysis time, with the separation process of the previous sample and the time for the whole analysis can be shortened. These systems, however, may cause some systematic problems, including contamination of the switching valve with sample solution because the capillary column is directly fixed on the six-port valve. This may result in inaccurate quantitative information and overestimation of analyte. In the draw/eject extraction system (Fig. 4), an extraction capillary column is placed between the injection loop and the injection needle of the HPLC autosampler. The capillary connections are facilitated by using a 2.5-cm long sleeve of 1/16 in. polyether ether ketone (PEEK) tubing at each end of the capillary, fixed with 1/16- in. SS unions (0.25 mm bore stainless steel nuts) and ferrules. An injection loop is installed to prevent sample contamination of the metering pump and switching valve. Building in UV, diode arrays or fluorescence detectors between the HPLC and the MSD, can enhance the multi-dimensional and simultaneous multidetections, improving analyte identification. As shown in Fig. 4A, a computer controls the injection syringe, which repeatedly draws and ejects sample from the vial, with the analytes partitioning from the sample matrix into the stationary phase until equilibrium is almost reached. Subsequently, the extracted analytes can be directly desorbed from the capillary coating by mobile phase flow or by an aspirated desorption solvent after switching the six-port valve (Fig. 4B). It is therefore necessary to prevent plugging of the capillary columns and flow lines during the extraction, as well as to remove particles from samples by filtration before extraction. The desorbed analytes are then transported to the HPLC column for separation and detected using a UV or mass selective detector (MSD). The computer controls the drawing and ejection of sample solution, switching of the valves, control of peripheral equipment such as the HPLC and MSD, and analytical data processing, thus reducing labor and enhancing precision. In addition, a large number of samples can be automatically processed by the autosampler without carryover, because the injection needle and capillary column are washed in methanol and the mobile phase before the sample is extracted. 2.2.2. Optimization of parameters In-tube SPME depends on the distribution coefficient of each analyte as well as its affinity for the fiber SPME, making it important to raise the distribution factor in the stationary phase to optimize the rapidity and efficiency of extraction. The selectivity and effi- ciency of the extraction depend on the extraction rate, the sample volume, the pH of the sample, the type of stationary phase, and the internal diameter, length, and film thickness of the capillary column. Several commercially available capillary columns, which differ according to the selectivity of the stationary phase, internal diameter, length and film thickness, have been developed. For example, a low polarity column with a methyl silicon liquid-phase selectively retains hydrophobic compounds, whereas a high polarity column with a polyethylene glycol liquid-phase selectively retains hydrophilic compounds. Since the internal diameter, length and film thickness of the column and other dimensions affect the amount of sample that can be loaded and the amount of compound that can be extracted, these parameters should be chosen carefully. The thin (usually <1 m) coating in such capillaries often results in low stationary phase loading reducing sample capacity and extraction sensitivity, although extraction equilibrium is quickly attained in the relatively thin layers. Although increasing the film thickness of the stationary phase may solve this problem, it is extremely dif- ficult to reliably immobilize thicker coatings using conventional approaches, and conventionally prepared GC coatings do not bind chemically to the fused-silica capillary inner surface. This lack of chemical bonds is mainly responsible for low solvent stability, preventing effective hyphenation of in-tube SPME techniques that employ organic or organo-aqueous mobile phases. In addition, although large amounts of compound can be extracted, their quantitative desorption from capillary columns may be difficult by
H Kataoka et aL/ Analytica Chimica Acta 655 (2009)8-29 石西园店可 extension of the sample bandwidth(peak broadening and tailing) The optimal length of a capillary column is 20-100 cm and the opti mal length of a flow through in draw/eject extraction system is 50-60 cm. The internal diameter of a capillary used in combina tion with HPLC is 0. 25 or 0. 32 mm with a short widebore capillary (0.53 mm)used in combination with capillary electrophoresis(CE), because widebore capillaries are resistant to pressure. Although capillary columns with chemically bonded or cross-linked liquid phases are very stable in water and organic solvents, they readily deteriorate in the presence of strong inorganic acids or alkalis. How- ever, capillary columns are generally stable for the mobile phase usually used in HPLC. In in-tube SPME, complete equilibrium extraction is generally 落二三莴手R的导 not obtained fo y analyte, because these analytes are partially desorbed into the mobile phase during the ejection step For flow through extraction systems, the volume of sample passed through a capillary is usually 0. 2-2 mL, and the optimum extraction flow-rate is 0. 25-4 mLmin-, depending on the capacity of the column. In contrast, an increase in the number and volume of draw/eject cycle can enhance the extraction efficiency of in-tube SPMe, but band- width may widen, and peak broadening has often been observed. A column 20-80 cm in length is usually used for effective extraction. For a capillary column (inner diameter 0. 25 mm, length 60 cm) with a volume of 30uL and an injection needle of 10 HL, optimal condi- tions include a draw/ejection volume of 30-40-FL, a draw/ejection flow-rate of50-100-HLmin-I and 10-15 draw/ejection cycles [15] Generally, it is possible to increase the extraction efficiency of ana- lyte into a stationary phase in Spme by changing the ph and salt level of the sample solution. Acidic and basic compounds can be effectively extracted from acidic and alkaline sample solutions respectively. However, the stability of each compound at the ph 目创|导N,,=,,,=,, of the sample solution must be determined beforehand. Although salting out increases extraction efficiency in fiber SPMe, the salt deposits can clog the column during in-tube SPME. Furthermore, the presence of a hydrophilic solvent such as methanol in the sam- ple decreases the extraction efficiency by increasing the solubility of the compound in the sample, but methanol concentrations ,, <5% have little effect on extraction efficiency. The amount of com- pound extracted into the stationary phase is dependent on the 2. 2.3. Capillary coatings GC capillary column coatings were initially compared with 至三R,,,, uncoated fused-silica(a retention gap capillary) as extraction devices for in-tube SPMe. These comparisons included silica mod ified with PDMs (e.g, SPB-1, PTE-5 and SPB-5) or polyethylene glycols(PEG), such as Omegawax 250 and Supelcowax, porous DVB and Supel-Q-Plot 3, 79-81. The relatively polar poly(ethylene glycol)(PEG) coating(Omegawax 250)is sufficiently bonded and cross-linked to prevent loss of phase when solvent is passe through the capillary, showing the highest yield of analyte. Si ica modified columns have been found more suitable for the analysis of non-polar compounds. Use of the adsorptive coated cap- illary Supel-Q-PLOT(DVB polymeric material) has recently been found to be more efficient for the analysis of estrogens, because 9=,, of its large surface area, enhancing mass-transfer kinetics, and because estrogens are of intermediate polarity [82]. Table 1 shows a comparison of some commercially available GC capillary columns used in the extraction of environmental pollutants, food contami nants and biological compounds For most organic comp porous polymer-type capillary column(Supel-Q PLOT) showed bet- ter extraction efficiency than the other liquid-phase type capillary columns(CP-Sil 5CB, CP-Sil 19CB, and CP-Wax 52CB) As the PLot column has a large adsorption surface area and a thick-film layer, the amount extracted was greater than that with liquid-phase type columns. However, patulin was effectively extracted with another
14 H. Kataoka et al. / Analytica Chimica Acta 655 (2009) 8–29 Table 1 Ability of commercially available capillary coatings to extract various compounds by in-tube SPME. Compound DB-1 (0.25 m) CP-Sil 5CB (5 m) DB-17 (0.25 m) CP-Sil 19CB (1.2 m) Omega wax (0.25 m) CP-Wax 52CB (1.2 m) Pora PLOT amine (10 m) Supel-Q PLOT (17 m) Carboxen 1006 (15 m) Ref. Nonyl phenol 9.7 – 10.4 – 3.5 – – 21.6 – [161] Bisphenol A 1.6 – 1.8 – 1.2 – – 22.8 – [161] Dibutyl phthalate 8.6 – 11.0 – 2.7 – – 24.1 – [161] Di-2-ethylhexyl phthalate 3.6 – 5.1 – 4.0 – – 5.8 – [161] Genistein 1.1 – 2.3 – 7.2 – – 11.9 – [166] Patuline – 0.4 – 0.4 – 0.2 11.8 0.8 30.3 [163] Aflatoxin B1 – 3.1 – 2.5 – 2.8 23.7 56.2 2.8 [164] -Estradiol 0.8 – 2.4 – 1.8 – – 30.0 – [82] Testosterone – 13.3 – 9.3 – 4.2 20.5 44.9 5.8 Cortisol – 8.8 – 8.4 – 9.5 – 15.9 5.9 [122] Nicotine – 3.6 – 3.0 – 2.3 50.5 21.4 26.7 [123] Atrazine 1.5 – 2.3 – 1.1 – – 58.8 12.0 [149] Perfluorooctane sulfonate – 0.4 – 1.9 – 1.2 57.5 12.8 5.9 Phenanthrene – 50.8 – 63.7 – 17.7 – 54.8 1.1 Benzo(a)pyrene – 28.7 – 61.9 – 44.7 – 49.2 2.6 Each data represents relative ratio of peak height of compound by in-tube SPME against that by direct injection. extension of the sample bandwidth (peak broadening and tailing). The optimal length of a capillary column is 20–100 cm and the optimal length of a flow through in draw/eject extraction system is 50–60 cm. The internal diameter of a capillary used in combination with HPLC is 0.25 or 0.32 mm with a short widebore capillary (0.53 mm) used in combination with capillary electrophoresis (CE), because widebore capillaries are resistant to pressure. Although capillary columns with chemically bonded or cross-linked liquid phases are very stable in water and organic solvents, they readily deteriorate in the presence of strong inorganic acids or alkalis. However, capillary columns are generally stable for the mobile phase usually used in HPLC. In in-tube SPME, complete equilibrium extraction is generally not obtained for any analyte, because these analytes are partially desorbed into the mobile phase during the ejection step. For flow through extraction systems, the volume of sample passed through a capillary is usually 0.2–2 mL, and the optimum extraction flow-rate is 0.25–4 mL min − 1, depending on the capacity of the column. In contrast, an increase in the number and volume of draw/eject cycles can enhance the extraction efficiency of in-tube SPME, but bandwidth may widen, and peak broadening has often been observed. A column 20–80 cm in length is usually used for effective extraction. For a capillary column (inner diameter 0.25 mm, length 60 cm) with a volume of 30 L and an injection needle of 10 L, optimal conditions include a draw/ejection volume of 30–40-L, a draw/ejection flow-rate of 50–100- L min − 1 and 10–15 draw/ejection cycles [15] . Generally, it is possible to increase the extraction efficiency of analyte into a stationary phase in SPME by changing the pH and salt level of the sample solution. Acidic and basic compounds can be effectively extracted from acidic and alkaline sample solutions, respectively. However, the stability of each compound at the pH of the sample solution must be determined beforehand. Although salting out increases extraction efficiency in fiber SPME, the salt deposits can clog the column during in-tube SPME. Furthermore, the presence of a hydrophilic solvent such as methanol in the sample decreases the extraction efficiency by increasing the solubility of the compound in the sample, but methanol concentrations of ≤5% have little effect on extraction efficiency. The amount of compound extracted into the stationary phase is dependent on the concentration of the compound in the sample. 2.2.3. Capillary coatings GC capillary column coatings were initially compared with uncoated fused-silica (a retention gap capillary) as extraction devices for in-tube SPME. These comparisons included silica modified with PDMS (e.g., SPB-1, PTE-5 and SPB-5) or polyethylene glycols (PEG), such as Omegawax 250 and Supelcowax, porous DVB and Supel-Q-Plot[3,79–81]. The relatively polar poly(ethylene glycol) (PEG) coating (Omegawax 250) is sufficiently bonded and cross-linked to prevent loss of phase when solvent is passed through the capillary, showing the highest yield of analyte. Silica modified columns have been found more suitable for the analysis of non-polar compounds. Use of the adsorptive coated capillary Supel-Q-PLOT (DVB polymeric material) has recently been found to be more efficient for the analysis of estrogens, because of its large surface area, enhancing mass-transfer kinetics, and because estrogens are of intermediate polarity [82] . Table 1 shows a comparison of some commercially available GC capillary columns used in the extraction of environmental pollutants, food contaminants and biological compounds. For most organic compounds, the porous polymer-type capillary column (Supel-Q PLOT) showed better extraction efficiency than the other liquid-phase type capillary columns (CP-Sil 5CB, CP-Sil 19CB, and CP-Wax 52CB). As the PLOT column has a large adsorption surface area and a thick-film layer, the amount extracted was greater than that with liquid-phase type columns. However, patulin was effectively extracted with another
H Kataoka et aL/ Analytica Chimica Acta 655 (2009 )8-29 porous polymer-type capillary column(Carboxen-1006 PLOT)but is convenient for coupling of miniaturized samples to micro-or not Supel-Q PLOT For the extraction of nicotine and perfluorooc scale separation technologies such as micro-LC and-CE The ne sulfonate, a CP-Pora PLOT amine gave superior extraction configuration of the on-line preconcentration is the same as that efficiency because of its affinity to relatively polar compounds In of the coated capillary microextraction system described in Section contrast, CP-Sil 19CB (liquid-phase type capillary) was superior for 2.2.1, with construction of these systems involving two Microfeeder aromatic hydrocarbons, although the film was thin. MF-2 pumps equipped with MS-GAN microsyringes. Several research groups have attempted to synthesize new A comparison of PDMs columns packed with Zylon(a materials to improve extraction efficiency and selectivity. These fibrous rigid-rod heterocyclic polymer; poly (p-phenylene -2, 6 nclude the preparation of a series of electrochemical coating benzobisoxazole) fibers[89 and stainless steel wire 88 for the based on polypyrrol(PPY) by an oxidative polymerization method determination of antidepressant drugs showed that up to 246 fila- [183]. The extraction efficiencies of PPY coatings were better than ments could be packed into a capillary, with an optimum packing those of commercial GC columns, due to the numerous types of density of 52% of the lumen volume. Using this system, analytes interactions between these multifunctional (i.e. T-T, polar, hydro- could be optimally desorbed in 2 FL of acetonitrile, with about gen bonding and ionic interactions)coatings and the analytes. 4 nL being injected for CE. Compared with wire-in-tube techniques Another advantage of electrochemical polymer-coated over cor in combination with micro-LC, The fiber-in-tube SPE technique in mercial capillaries for in-tube SPME is the ability to manipulate combination with CE resulted in a 3-76-fold increase in precon- extraction efficiency and selectivity by regulating the thickness of centration, depending on whether analytes interacted significantly the coating (ie the number of electrochemical polymer cycles). with the sorbent fiber in the lumen of the capillary. Zylon has Chemically or electrochemically deposited PPY coatings have been also used in the same way for determination of phthalates in oupled to either HPLC [84 for determination of aromatics and wastewater[91-93 Exhaustive extraction was achieved through a anions, respectively. The sensitivity and selectivity of these coatings combination of the increased sorbent capacity and reduced sample for in-tube extraction can be adjusted by altering film thickness In volume. Desorption into the mobile phase was very efficient for the contrast, a simple Spme device has been fabricated for use in on- fiber-in-tube method, eliminating the need for a separate desorp- line immunoaffinity capillaries [85 Immunoaffinity-SPME, which tion solvent, although in this case the mobile phase itself contained combines the inherent selectivity of antibodies and the advantages a high proportion of organic solvent(90% methanol). of SPME, is prepared by immobilization of an antibody in in-tube SPME, using a sensitive, selective, and reproducible method Impor- 2.3.2. Sorbent-packed capillary in-tube SPMe tant aspects of the optimization of in-tube SPMe conditions and An alternative approach using a small section of capillary packed evaluation of capacity of immunoaffinity capillaries have been with microsphere beads is similar to SPE. Although this technique described [85]. Furthermore, sol-gel titania-PDMS-coated capillar- is easy to implement in existing autosampler systems, sorbent s have been used for on-line in-tube SPme and the analysis of packed capillaries can easily break under high pressure. When PAH, ketones, and alkylbenzenes [ 86]. and this method can be easily liquid samples are analyzed by direct immersion, the main dis- automated by using standard HPLC equipment. More recently, new advantage of this technique is that even very tiny particles are ic liquid-mediated sol-gel coatings were developed for capillary able to block the capillaries, making it necessary to use very microextraction of PAHs[87] clean samples. Phases better suited to the extract relatively polar compounds from aqueous samples have been developed to ed capillary in-tube SPME enhance the sensitivity and overall utility of capillary microex traction methods. These include a molecularly imprinted polymer 2.3.1. Wire-in-tube SPME and fiber-in-tube SPE (MIP), consisting of cross-linked synthetic polymers produced by Several methods have been developed to increase extraction copolymerizing a monomer with a cross-linker in the presence of a efficiencies and extend this method to microscale applications, template molecule as an in-tube SPMe adsorbent [ 90. A capillary including"wire-in-tube SPME, using modified capillary columns packed with MiP particles in an 8-cm PEEK tube (inner diameter with inserted stainless steel wires [88]; and"fiber-in-tube SPE", 0.76 mm) has been used for the selective analysis of B-blockers using capillary tubes packed with fibrous rigid-rod heterocyclic biological fluids. In addition, a highly biocompatible SPME-capillary polymers(Fig. 1B)89, 90. These techniques, which require fixed packed with alkyl-diol-silica particles(ADS) particles was deve sample volumes, have also been called miniaturized SPE, rather oped as a restricted access material(RAM)(94]. The bifunctionality than SPME. These distinctions are important, as SPME is an equi- of the aDs extraction phase prevented fouling of the capillary by librium extraction technique where sample volume is significantly adsorbed protein while simultaneously trapping the analytes in the arger than sorbent capacity, with calibration based on the par- hydrophobic porous interior. this oach required a simplified titioning or affinity of each analyte for sorbent For wire-in-tube apparatus compared with existing RAM column switching proce- SPME, internal capacity can be significantly reduced by insertion of dures, as well as overcoming the need for ultrafiltration a narrow stainless steel wire into the extraction capillary while the deproteinization step prior to handling biological samples, thus fur- For fiber-in-tube SPE, several hundred fine filaments of polymeric pre-concentrate resulted in low-ng/mL detection limit ability to surface area of the polymeric coating material remains the same. ther minimizing sample preparation requirements. the ability to materials packed longitudinally into a short polyether ether ketone (PEEK) capillary tube serve as the extraction medium. This tech- 2.4. Monolithic capillary in-tube SPME ique not only can reduce the internal void volume of the extraction capillary but the fine polymer filaments can be employed as the An alternative approach consists of in-tube SPme using mono- extraction medium. Because the filaments are arranged parallel lithic capillary columns comprised of one piece of organic polymer to the outer tubing, narrow coaxial channels can form inside the or silica with a unique flow through double-pore structure. Capil capillary. therefore, fiber-in-tube spe device involves a reduced laries with monolithic sorbents can be easily synthesized in situ pressure drop during the extraction and desorption compared with initiated thermally or by radiation, using a mixture of monomer. a conventional particle-packed SPE cartridge. Furthermore, the cross-linker and proper porogenic solvent. It generally results effective interaction of the sample solution with a number of fine in monolithic structures with different functional groups that fibrous extraction capillaries suggests further miniaturization as are biocompatible and pH-stable. A C18-bonded monolithic sil- a microscale sample preconcentration device, as this technology ica column, prepared by in situ hydrolysis and polyc
H. Kataoka et al. / Analytica Chimica Acta 655 (2009) 8–29 15 porous polymer-type capillary column (Carboxen-1006 PLOT) but not Supel-Q PLOT. For the extraction of nicotine and perfluorooctane sulfonate, a CP-Pora PLOT amine gave superior extraction efficiency because of its affinity to relatively polar compounds. In contrast, CP-Sil 19CB (liquid-phase type capillary) was superior for aromatic hydrocarbons, although the film was thin. Several research groups have attempted to synthesize new materials to improve extraction efficiency and selectivity. These include the preparation of a series of electrochemical coatings based on polypyrrol (PPY) by an oxidative polymerization method [83]. The extraction efficiencies of PPY coatings were better than those of commercial GC columns, due to the numerous types of interactions between these multifunctional (i.e. –, polar, hydrogen bonding and ionic interactions) coatings and the analytes. Another advantage of electrochemical polymer-coated over commercial capillaries for in-tube SPME is the ability to manipulate extraction efficiency and selectivity by regulating the thickness of the coating (i.e. the number of electrochemical polymer cycles). Chemically or electrochemically deposited PPY coatings have been coupled to either HPLC [84] for determination of aromatics and anions, respectively. The sensitivity and selectivity of these coatings for in-tube extraction can be adjusted by altering film thickness. In contrast, a simple SPME device has been fabricated for use in online immunoaffinity capillaries [85]. Immunoaffinity-SPME, which combines the inherent selectivity of antibodies and the advantages of SPME, is prepared by immobilization of an antibody in in-tube SPME, using a sensitive, selective, and reproducible method. Important aspects of the optimization of in-tube SPME conditions and evaluation of capacity of immunoaffinity capillaries have been described [85]. Furthermore, sol–gel titania-PDMS-coated capillaries have been used for on-line in-tube SPME and the analysis of PAH, ketones, and alkylbenzenes [86], and this method can be easily automated by using standard HPLC equipment. More recently, new ionic liquid-mediated sol–gel coatings were developed for capillary microextraction of PAHs [87]. 2.3. Packed capillary in-tube SPME 2.3.1. Wire-in-tube SPME and fiber-in-tube SPE Several methods have been developed to increase extraction efficiencies and extend this method to microscale applications, including “wire-in-tube SPME”, using modified capillary columns with inserted stainless steel wires [88]; and “fiber-in-tube SPE”, using capillary tubes packed with fibrous rigid-rod heterocyclic polymers (Fig. 1B) [89,90]. These techniques, which require fixed sample volumes, have also been called miniaturized SPE, rather than SPME. These distinctions are important, as SPME is an equilibrium extraction technique where sample volume is significantly larger than sorbent capacity, with calibration based on the partitioning or affinity of each analyte for sorbent. For wire-in-tube SPME, internal capacity can be significantly reduced by insertion of a narrow stainless steel wire into the extraction capillary while the surface area of the polymeric coating material remains the same. For fiber-in-tube SPE, several hundred fine filaments of polymeric materials packed longitudinally into a short polyether ether ketone (PEEK) capillary tube serve as the extraction medium. This technique not only can reduce the internal void volume of the extraction capillary, but the fine polymer filaments can be employed as the extraction medium. Because the filaments are arranged parallel to the outer tubing, narrow coaxial channels can form inside the capillary. Therefore, fiber-in-tube SPE device involves a reduced pressure drop during the extraction and desorption compared with a conventional particle-packed SPE cartridge. Furthermore, the effective interaction of the sample solution with a number of fine fibrous extraction capillaries suggests further miniaturization as a microscale sample preconcentration device, as this technology is convenient for coupling of miniaturized samples to micro- or nano-scale separation technologies such as micro-LC and -CE. The configuration of the on-line preconcentration is the same as that of the coated capillary microextraction system described in Section 2.2.1, with construction of these systems involving twoMicrofeeder MF-2 pumps equipped with MS-GAN microsyringes. A comparison of PDMS columns packed with Zylon® (a fibrous rigid-rod heterocyclic polymer; poly(p-phenylene-2,6- benzobisoxazole) fibers [89] and stainless steel wire [88] for the determination of antidepressant drugs showed that up to 246 filaments could be packed into a capillary, with an optimum packing density of 52% of the lumen volume. Using this system, analytes could be optimally desorbed in 2 L of acetonitrile, with about 4 nL being injected for CE. Compared with wire-in-tube techniques in combination with micro-LC, The fiber-in-tube SPE technique in combination with CE resulted in a 3–76-fold increase in preconcentration, depending on whether analytes interacted significantly with the sorbent fiber in the lumen of the capillary. Zylon® has also used in the same way for determination of phthalates in wastewater [91–93]. Exhaustive extraction was achieved through a combination of the increased sorbent capacity and reduced sample volume. Desorption into the mobile phase was very efficient for the fiber-in-tube method, eliminating the need for a separate desorption solvent, although in this case the mobile phase itself contained a high proportion of organic solvent (90% methanol). 2.3.2. Sorbent-packed capillary in-tube SPME An alternative approach using a small section of capillary packed with microsphere beads is similar to SPE. Although this technique is easy to implement in existing autosampler systems, sorbentpacked capillaries can easily break under high pressure. When liquid samples are analyzed by direct immersion, the main disadvantage of this technique is that even very tiny particles are able to block the capillaries, making it necessary to use very clean samples. Phases better suited to the extraction of relatively polar compounds from aqueous samples have been developed to enhance the sensitivity and overall utility of capillary microextraction methods. These include a molecularly imprinted polymer (MIP), consisting of cross-linked synthetic polymers produced by copolymerizing a monomer with a cross-linker in the presence of a template molecule, as an in-tube SPME adsorbent [90]. A capillary packed with MIP particles in an 8-cm PEEK tube (inner diameter 0.76 mm) has been used for the selective analysis of -blockers in biological fluids. In addition, a highly biocompatible SPME-capillary packed with alkyl-diol-silica particles (ADS) particles was developed as a restricted access material (RAM) [94]. The bifunctionality of the ADS extraction phase prevented fouling of the capillary by adsorbed protein while simultaneously trapping the analytes in the hydrophobic porous interior. This approach required a simplified apparatus compared with existing RAM column switching procedures, as well as overcoming the need for ultrafiltration or another deproteinization step prior to handling biological samples, thus further minimizing sample preparation requirements. The ability to pre-concentrate resulted in low-ng/mL detection limits. 2.4. Monolithic capillary in-tube SPME An alternative approach consists of in-tube SPME using monolithic capillary columns comprised of one piece of organic polymer or silica with a unique flow through double-pore structure. Capillaries with monolithic sorbents can be easily synthesized in situ, initiated thermally or by radiation, using a mixture of monomer, cross-linker and proper porogenic solvent. It generally results in monolithic structures with different functional groups that are biocompatible and pH-stable. A C18-bonded monolithic silica column, prepared by in situ hydrolysis and polycondensation