Comparison of Performance Parameters of ... - Semantic Scholar

3 downloads 0 Views 688KB Size Report
Apr 10, 2011 - conditions.3,4,25–36 A total of 39 procedures (16 for Fe and 23 for. Co) were considered. ... Co with. 2-nitroso-5-dimethylaminophenol,29.
ANALYTICAL SCIENCES APRIL 2011, VOL. 27

381

2011 © The Japan Society for Analytical Chemistry

Comparison of Performance Parameters of Photothermal Procedures in Homogeneous and Heterogeneous Systems Mikhail A. PROSKURNIN,*1 Elena S. RYNDINA,*1 Dmitrii S. TSAR’KOV,*2 Valerii M. SHKINEV,*3 Adelina SMIRNOVA,*4 and Akihide HIBARA*5† *1 M.V. Lomonosov Moscow State University, Vorob’evy gory, d. 1 str. 3, Moscow 119991, Russia *2 Institute for Biomedical Problems of RAS, Khoroshevskoe shosse 76A, Moscow 123007, Russia *3 Vernadsky Institute of Geochemistry and Analytical Chemistry of RAS, ul. Kosygina, d. 19, Moscow 119991, Russia *4 Department of Mechanical Engineering, School of Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo, Tokyo 113–8656, Japan *5 Institute of Industrial Science, The University of Tokyo, 4-6-1 Komaba, Meguro, Tokyo 153–8505, Japan

The main types of analytical procedures used in thermal-lens spectroscopy and microscopy, which are based on photometric reactions in (i) aqueous solutions, (ii) organo-aqueous mixtures, (iii) polymer-containing (nonionic surfactants or polyethylene glycol) aqueous solutions, (iv) water-organic extraction systems, and (v) two-phase aqueous extraction systems, were compared from the viewpoint of both reproducibility and sensitivity. This comparison was made by examples of the determination of cobalt and iron for batch conditions, flow determination, and detection in HPLC, flow-injection analysis (FIA), and μFIA. It was revealed that for all five types, the real analytical efficiency (a decrease in the limit of detection (LOD) as compared to spectrophotometry) is primarily determined by the reaction conditions, provided that excitation of the thermal lens is the same. Aqueous solutions provide more efficient optimization of reaction conditions than do those in organo-aqueous solutions and solvent-extraction water-organic mixtures. The best results are achieved when shifting to polymer-containing aqueous solutions and two-phase aqueous extraction systems, which decreases in the LODs by a factor of 20 – 100%. (Received December 27, 2010; Accepted January 27, 2011; Published April 10, 2011)

Introduction Photothermal-lens or thermal-lens spectrometry (TLS) is a versatile technique of molecular spectroscopy.1,2 Its high sensitivity is achieved not only through instrumentation, but also because of the reaction conditions at trace levels.3 The locality and high spatial resolution of TLS have led to wide-scale applications of thermal-lens microscopy (TLM) as a state-of-the-art method.4–7 TLM perfectly fits micro-totalanalytical systems (μTAS) (reactions and separations occurring at high rates in nanoscale volumes).4,8–10 To exploit the full potential of TLM-μTAS, tuning the reaction conditions of integrated chemistry on the microscale is even more crucial.11–13 Thus, understanding the principles of procedure selection for photothermal measurements seems to be a very important step for success. Previously, we showed that the sensitivity of TLS and TLM is dictated by the same optimization features, and the concentration Moreover, limits of detection (LODs) are the same.14 simultaneous consideration of the accuracy of TLS and reaction conditions at the nanoscale level also enhances the selectivity.15 However, the dependence of the TLS sensitivity on the To whom correspondence should be addressed. E-mail: [email protected]



thermophysical properties of the solution poses a problem: What is more important, an increase in the strength of the photothermal effect (photothermal sensitivity) or an increase in the analytical sensitivity due solvent selection as a chemical component? From the viewpoint of photothermal sensitivity, the best are nonpolar organic solvents,2,16,17 which can be used in extraction procedures and preconcentration. However, they are flammable, toxic, etc. and cannot be used with water-soluble reagents or in biological systems. On the contrary, the most flexible procedures use aqueous solutions. However, they exhibit much lower photothermal sensitivity and cannot fit well when preconcentration is required. A compromise is organo-aqueous solutions combining the thermophysical properties of organic media and sample preparation in aqueous solutions.18,19 Still, the nature of organic components cannot be easily varied in a single reaction. Recently, aqueous solutions modified with surfactants or, more importantly, water-soluble polymers (first of all, polyethylene glycols, PEG) have been introduced as green analytical solvents. Their chemistry has been studied in detail,18,20,21 and they are commercially available, inexpensive, and can be used both as single-phase or extraction two-phase systems. High photothermal sensitivity in these solutions21–23 poses them as an alternative to organic solvents for increasing TLS sensitivity in aqueous media. Therefore, we compared the performance parameters of

382

ANALYTICAL SCIENCES APRIL 2011, VOL. 27

various analytical procedure types in TLS/TLM for new and previously reported procedures. We have chosen Co and Fe determination because (i) the photometric methods for these metals are both sensitive and reliable;24 (ii) they are based on the same reaction type, chelation; (iii) these methods cover all of the selected procedure types; and (iv) optimized TLS/TLM procedures have low practical LODs under batch and flow conditions.3,4,25–36 A total of 39 procedures (16 for Fe and 23 for Co) were considered.

Procedure Comparison Criteria The instrumentation parameters and the main equations of thermal lens theory are given in Supporting Information. The theoretical increase in the calibration slope for TLS was calculated as the ratio of the expected sensitivities of TLS and spectrophotometry, E theor = 2.303Pe E 0ε TL ε phot–1 ,

(1)

where εphot and εTL denote the molar absorptivities at the wavelengths used for spectrophotometry and TLS, respectively; Pe is the excitation power and E0 is the thermo-optical constant of the solvent (see Supporting Information). The corresponding experimental enhancement factor, Eexp, was calculated as the ratio of the calibration slopes (sensitivity coefficients), bTL and bphot, for the TLS and spectrophotometric procedures: Eexp = bTL/bphot.

(2)

LODs obtained for different conditions (excitation power, Pe) and for different substances (various εTL values) were normalized for the same conditions Pe = 100 mW and εTL = 104 L/(mol cm), LOD TL =

LOD ⋅Peε TL LOD ⋅Peε TL = . 10 2⋅10 4 10 6

(3)

The increase in the sensitivity for a given procedure was calculated as an LOD ratio, LR, of the spectrophotometric, LODphot, and normalized LOD for TLS (3), LR =

LOD phot . LOD TL

(4)

The reproducibility and repeatability of measurements were estimating by relative standard deviation (RSD) values calculated according to Ref. 37. The increases in the slope (2) and in the LODs (4) were compared to one another by introducing the total sensitivity enhancement when shifting from spectrophotometry to thermal lensing (SE factor), SE = LR . E exp

(5)

Procedure Development As organic media with very high photothermal enhancement factors, E0,1,2 we selected toluene, chloroform, and m-xylene. As surfactants, Brij-family reagents and Triton X-100 were selected according to previous data.18 PEGs were selected as representatives of water-soluble polymers; molecular weights of 1500, 2000, and 6000 Da provide a threefold enhancement in

E0, as compared to aqueous solutions.38 The absorption-band parameters for all of the reaction products do not change when shifting from aqueous to organic or organic to aqueous solutions with PEGs or surfactants (0.01 – 30% w/v, Table S-3, Supporting Information). The molar absorptivity of water-insoluble Co(III) tris(2-nitroso-1-naphtholate) is much higher in its PEG- or surfactant-containing solutions than its alkaline aqueous solutions (4 × 103 L/(mol cm)) due to micellar solubilization of the chelate in these solutions.9 The following procedures were described previously: Co with diethyldithiocarbaminate;39 Co and Fe with 4-(2-pyridylazo)resorcinol (PAR),40 4-(2-thiazolylazo)resorcinol (TAR) and Xylenol Orange,41 and dithizone;42 Co with 2-nitroso-5-(N-propyl-32-nitroso-5-dimethylaminophenol,29 sulfopropylamino)phenol (Nitroso-PSAP),43 chiral Co complex with triethylenediamine;31 and Fe with 1,10-phenanthroline,32 bathophenanthroline,44 and bathophenanthroline disulfonic acid.45 Below, we briefly describe newly developed procedures. Fe(III) with nitroso-R-salt in water This system is used as a high-strength eluent for The microchromatography of strongly retained anions.46 procedure (Procedure 1, Supporting Information) is simple,47 and does not require optimization. The LOD is 10 nM, which is comparable to other photothermal methods for Fe, but with a very long linear range of 5 orders of magnitude. Thus, a low LOD is achieved for a relatively low εTL of 750 L/(mol cm) (Table S-3, Supporting Information). Co and Fe in surfactant- and PEG-containing aqueous solutions The procedures for the determination of Fe(II) with 1,10-phenanthroline and Co with 2-nitroso-1-naphthol and nitroso-R-salt25,34 were optimized (Procedures 2 and 4 – 6, Supporting Information) so as to exclude the Soret effect,18 the measurement time was decreased to 0.1 s. The optimum pH for reactions of Co with nitroso-naphthols and nitroso-R-salt48–51 in Triton X-100, Brij-30, and Brij-56 and PEG solutions was 7.0  (phosphate buffer solutions), and the reagent-to-metal ratio  was  1:30. The spectrophotometric LOD for Co is threefold  lower than in aqueous solutions, and closer to the extraction–spectrophotometric determination of Co (Table 1). The concentrations of all the surfactants were 0.08 M, which provided well-resolved TLS transient curves, which are in good agreement with the previous data;52 the nature of a surfactant does not affect the procedure. For PEG-modified aqueous solutions, no changes in spectrophotometric determination as compared to surfactant solutions were found (Table 1). The determination of Co (Procedure 4, Supporting Information) and Fe (Procedure 5, Supporting Information) show LODs of 2 and 4 nM, respectively. In the case of Fe as hemoglobin cyanide, we used a specially designed reagent mixture and added PEG at the end of the procedure (Procedure 7, Supporting Information). A good correlation of the LOD ratio for Fe as tris(1,10-phenanthrolinate) and hemoglobin cyanide with the ratio of their εTL confirms that PEGs do not affect the protein molecule. Extraction–photothermal determination of Co(III) with [2,2′]-furildioxime in organic solvents The reagent-to-metal ratio is 1:25. The main factor limiting the photothermal LOD is a high blank signal governed by the excess reagent. We developed a procedure since the Co chelate is extracted to polar organic solvents, while the reagent remains in water (Procedure 3, Supporting Information). The reagent concentration was 2.5 μM, which provided Co determination

Batch, aqueous Triton X-100 2-Nitroso-1-naphthol Batch, aqueous Triton X-100 Batch, aqueous Brij-30 1,10-Phenanthroline Batch, aqueous Brij-56 2-Nitroso-1-naphthol as hemoglobin cyanide Aqueous PEG 2000 Aqueous PEG 1500 Aqueous PEG 6000

Batch extraction Batch extraction Batch Flow extraction, μFIA Flow extraction, μFIA Flow extraction, μFIA Flow extraction, μFIA Flow extraction, μFIA Flow extraction, μFIA Flow extraction, μFIA Extraction

Extraction, PEG 1500 Extraction, PEG 2000 Extraction, PEG 20000

Co Fe Fe Fe Fe Co Fe

Co Co Co Co Co Co Co Co Fe Fe Co

Co Co Co

Use of modified aqueous solutions

Water-organic solvent extraction

Aqueous solvent extraction

Nitroso-R-salt

2-Nitroso-1-naphthol, chloroform 1-Nitroso-2-naphthol, toluene [2,2′]-Furildioxime, chloroform 2-Nitroso-1-naphthol, toluene 2-Nitroso-5-dimethylaminophenol, toluene Dimethylaminophenol, m-xylene 2-Nitroso-1-naphthol, m-xylene Nitroso-PSAP, isoamyl alcohol Bathophenanthroline disulfonic acid, chloroform Bathophenanthroline, isoamyl alcohol Diethyldithiocarbaminate, chloroform (488)

PAR, water–acetonitrile 3:2

Diethyldithiocarbaminate, water–ethanol 1:3 Dithizone, water–acetone 2:1

Nitroso-R-salt, water–ethanol 1:1 Nitroso-R-salt, water–acetone 1:1 TAR, water–ethanol 5:1

Xylenol Orange 2-Nitroso-1-naphthol Xylenol Orange 1,10-Phenanthroline

Batch Batch Batch/FIA, HPLC Batch/FIA, HPLC Batch Batch/FIA Batch/FIA HPLC HPLC

Nitroso-R-salt

Reagent, solvent

Use of organo-aqueous Co solutions Co Co Fe(II) Co Fe Co Co Fe

Condition

Batch Batch Batch/FIA, HPLC Batch Batch/FIA, HPLC Batch/FIA Batch/μFIA

Metal

Fe Co Co Co Fe(III) Fe(II) Fe(II)

Use of aqueous solutions

Procedure type

Table 1 Performance parameters for Co and Fe, achieved using TLS and TLM

1.000 1.000 1.000

0.5 0.5 1.0 0.500 1.000 1.000 1.739 1.000 3.000 1.333 5.000

0.8 0.5 0.5 0.5 0.5 0.8 1.0

1.0 1.0 1.5 5.0 1.0 0.7 0.5 0.4 1.5

3.0 1.0 3.0 3.0 5.0 0.3 3.0

0.03 0.75 0.70 0.09 0.52 0.81 0.81

bTL

40 40 40

300 300 300 75 200 200 200 100 200 100 60

100 100 100 100 100 100 80

0.46 0.55 0.44

61.00 58.00 40.00 12.00 50.00 50.00 33.00 29.00 45.00 22.00 0.77

1.10 1.00 0.90 0.92 0.66 0.80 5.54

100 3.30 100 3.60 120 2.16 120 1.80 60 0.40 120 11.00 120 12.00 40 1.60 40 1.11

100 100 120 45 120 140 100

LODSP/ Pe μM

6 4 5

0.5 0.5 1 2 5 5 2 10 10 10 20

2 3 2 2 1 2 1

5 5 10 30 10 4 4 5 20

10 9 10 30 30 2 10

0.02 0.01 0.03

0.08 0.07 0.05 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.05

0.03 0.04 0.04 0.04 0.02 0.04 0.03

0.04 0.06 0.04 0.06 0.06 0.05 0.05 0.04 0.06

0.05 0.03 0.03 0.05 0.05 0.02 0.02

2.4 1.6 2

1.5 1.5 3 1.5 10 10 4 10 20 10 12

2 3 2 2 1 2 0.8

5 5 12 36 6 4.8 4.8 2 8

10 9 12 13.5 36 2.8 10

LODTL/ LODnor/ RSDmin nM nM

4 2 3

2 2 2.25 2 20 20 5 20 30 15 0.9

2 3 2 2 1 2 3

8 8 14 36 0.5 10 12 3 9

1 14 14 5 32 3 11

417 625 500

333 333 333 333 100 100 435 100 150 133 417

400 167 250 250 500 400 1250

200 200 125 139 167 146 104 200 188

300 111 250 222 139 107 300

95.7 90.9 81.8 83.6 89.2 69.6 162.0

220.0 240.0 149.7 149.7 833.3 458.3 400.0 241.8 241.8

77.4 91.9 72.5

77.4 91.9 72.5

1874.6 1783.6 1577.6 1681.2 1874.6 1777.8 1391.3 1250.0 1250.0 1577.6 1434.8 1450.0 1500.0 1466.7 1874.6 1645.3

91.9 77.4 130.6

92.1

241.8

437.6

806.1

149.7

268.7 243.1

48.4

37.3 50.0 48.4 50.0 48.4 52.6 73.6

1.00 1.00 1.00

0.95 1.07 0.95 0.88 0.79 0.79 0.91 0.92 0.95 0.93 0.88

1.04 0.99 0.89 0.91 0.97 0.90 1.24

0.82 0.99 1.00 1.00 1.03 1.05 0.91 1.00 1.00

0.77 1.03 1.00 1.03 1.00 1.09 1.52

LOD/ LR, Eexp, Eexp/ Etheor, nM, Eq. Eq. (1) Eq. (2) Etheor Eq. (3) (4)

5.385 6.802 6.892

0.187 0.198 0.188 0.240 0.080 0.080 0.303 0.069 0.100 0.091 0.253

4.182 1.833 3.056 2.989 5.606 5.750 7.718

0.9 0.8 0.8 0.9 0.2 0.32 0.260 0.827 0.775

8.0 2.2 5.2 4.4 2.9 2.0 4.1

SE, Eq. (5)

ANALYTICAL SCIENCES APRIL 2011, VOL. 27 383

384 below 0.1 μM. The LOD is 1 nM (Table 1), and the linear range is 4 – 40 nM. Extraction–photothermal determination of Co with nitroso-R-salt using two-phase aqueous PEG systems We used the reaction of Co(III) with nitroso-R-salt,53 with taking advantage of PEG use in spectrophotometry.54 We added 1 M NH4F as an Fe-masking reagent to all of the solutions after heating with nitric acid (to decompose the chelates of interfering metals54) so as to decrease the blank signal. Also, the nitroso-R-salt concentration was decreased by a factor of 20 as compared to the procedure53 (to 0.4 mM, Procedure 8, Supporting Information). According to the existing data,54 the amounts of (NH4)2SO4 and PEG in the upper phase are 25.2 and 5.2%, respectively. The calculation of E0 of PEG extracts from Procedure 5 (Supporting Information) showed no difference compared to the values for manually prepared solutions with the same concentration of PEGs and the existing data.18 The calculations of E0 for (NH4)2SO4 and PEG23 compared to pure water of 1.3 and 1.8, respectively (the total enhancement of 1.3 × 1.8 = 2.34), correlate with the experimental value of 2.5. Thus, the increase in the photothermal sensitivity due to solvent properties is high, but not dominating.

Comparison of Procedures Comparison parameters A direct LOD comparison was hardly possible. A more correct approach is based on photothermal-to-spectrophotometric LOD ratios, but it does not always reveal a true picture because the excitation conditions (power or power density) and the molar absorptivity may differ significantly. Therefore, we introduced the normalized photothermal LOD, Eq. (3): this is the LOD for a given procedure if the excitation power and εTL were 100 mW and 10000 L/(mol cm), respectively. These values were selected as being the most common for TLS, and the real values of these parameters are not too far (40 – 300 mW and 4000 – 20000 L/(mol cm), see Table S-3, Supporting Information, and Table 1). Such scaling is correct in principle, since Beer’s law is obeyed in TLS/TLM,2,15 and the TLS LODs are inversely proportional to the excitation power, as the theory predicts.2 Therefore, the photothermal and spectrophotometric LODs are also compared for the scaled and equal power of 100 mW, Eq. (4). Further, we calculated the theoretical enhancement factors Etheor, Eq. (1), for all of the selected procedures and compared them with the actual experimental enhancement, Eq. (2): a deviation of their ratio from 1 would indicate some complication in photothermal determination. Finally, because the sensitivity coefficient (calibration slope) and LOD are governed by different factors, mainly the sensitivity and reproducibility,14,15,55,56 we compared the photothermal procedures by the minimum value of the reproducibility RSD and by the SE ratio, Eq. (5), which highlights changes in the reproducibility. Reproducibility A comparison of the reproducibility shows that all of the procedures are well-developed, and are characterized by the same RSD level, which is close to the expected instrumental error in TLS/TLM.14 This confirms that the reproducibility can be increased when adapting a spectrophotometric procedure to TLS due to a change in the error curve from transmittance to photothermal measurements.

ANALYTICAL SCIENCES APRIL 2011, VOL. 27 Sensitivity coefficients As expected, the experimental enhancement in the sensitivity coefficient over its value for spectrophotometry correlates well with the theory, giving a ratio of the experimental and theoretical factors close to 1. We believe that slightly lower ratios for TLM flow-extraction procedures (the values of 0.8 – 0.9) result from more complicated flow conditions in a microchip, which degrades Eexp and cannot be simply accounted for by Eq. (1). This problem is discussed elsewhere.5,56 Figure 1(a) shows that aqueous solutions exhibit the lowest increase, about 50-fold, and that the increase in modified aqueous solutions and two-phase aqueous systems are about an order higher. Organo-aqueous mixtures indicate a 300-fold enhancement in the sensitivity coefficient for TLS, as compared to spectrophotometry. Extraction–photothermal measurements show a mean 1500-fold increase. Thus, as expected, the calibration slope in TLS is governed by the solvent thermophysical properties. Normalized limits of detection The first noteworthy point is that normalized LODs, Eq. (3), are at the same level for aqueous, organo-aqueous, and organo-aqueous extraction systems: the average normalized LOD of the latter is only 20% lower than that of the former. This shows that an actual increase in the sensitivity of extraction procedures results from high εTL and high excitation powers in these studies,25–34 and not from the solvent properties. On the contrary, the normalized LODs for organo-aqueous mixtures exhibit the same LOD decrease as do the extraction procedures, but with much less changes in methodology compared to aqueous solutions. The normalized LODs of modified aqueous solutions and PEG extraction indicate the same sensitivity when compared to other procedures based on aqueous solutions. In fact, their normalized LODs are even lower than those for organo-aqueous and extraction systems, but these new procedures are very simple. LR factors In contrast to sensitivity coefficients, a quite different picture is presented by the spectrophotometric-to-photothermal LOD ratio, LR (Fig. 1b). While the LOD decrease in aqueous solutions is still the lowest, the LODs for the procedures based on organo-aqueous mixtures and water-organic extraction systems show the same level despite much higher sensitivity coefficients. For example, an average decrease in LODs for the extraction procedures is only 40%. On the contrary, the procedures using modified aqueous solutions exhibit a 400-fold decrease in the LOD as compared to spectrophotometric procedures, and a more than twofold increase in this sensitivity parameter as compared to the procedures based on aqueous solutions. SE factors The same trend is even more pronounced in comparing the SE factor (Fig. 1c). As mentioned above, this is a ratio of the enhancement in the LOD to that in the sensitivity coefficient, which represents the reproducibility enhancement. However, while RSDmin indicates the precision for the best measurement conditions (the minimum instrumental error), the LOD-based SE factor indicates the reproducibility at the lowest concentrations, the procedure is developed for. Thus, if SE is higher than 1, it implies that TLS has the full advantage over spectrophotometry, providing precise measurements of low concentrations (Fig. 2); if SE is lower than 1, it implies that the reproducibility of measurements at low concentrations is

385

Relative standard deviation

ANALYTICAL SCIENCES APRIL 2011, VOL. 27

0.8

0.6

0.4

0.2

0 1. × 10–4

1. × 10–3

1. × 10–2

1. × 10–1

Absorbance

Fig. 2 Calculated dependences of the instrumental error in measurements (relative standard deviation) on the sample absorbance for TLS (solid line, see14 for the equation) and transmittance measurements with spectrophotometry (dashed line14); the detector noise is 1% for both cases. Calculations done with Maple 13.

Fig. 1 Values of (a) experimental enhancement factor of thermal lensing, Eexp, Eq. (2); (b) limit of detection ratio for spectrophotometry and thermal lensing, LR, Eq. (4); (c) total sensitivity enhancement, SE, Eq. (5) for various procedure types used in TLS and TLM. The solution types, in which the reactions were carried out, are indicated in the figure.

strongly degraded and the sensitivity enhancement originating from the thermophysical properties of the solution is somewhat wasted. The SE factor is less than 1 for procedures based on organo-aqueous solutions and water-organic extraction and above 1, otherwise. This implies that in aqueous solutions, the changes in the RSD are smaller as compared to those expected from the theory (Fig. 2). The general shape of the curves for TLS is sharper than that for spectrophotometry, which accounts for a bigger increase in the LOD as compared to that in the

sensitivity coefficient. This is confirmed by comparing the error curves for presynthesized ferroin and for the photometric reaction of Fe(II) with 1,10-phenanthroline including the sample preparation (Fig. 3a). The sample preparation stage certainly degrades the reproducibility (curve 2 shows higher RSDs for the same iron concentrations); however, the major RSD degradation occurs in the LOD vicinity only. This means that for real TLS procedures, the linear calibration range is long. On the contrary, changes from the procedures for presynthesized solutions to the real procedures for water-organic extraction systems show a different picture in the error (RSD) curves (Fig. 3b): firstly, RSD degrades very significantly, compared to that for aqueous solutions (Fig. 3a); secondly, the shape of the curve for the procedure with sample preparation shows much more similarity to transmittance measurements (see Fig. 2). These two factors result in the observed behavior of SE (Fig. 1c) for these extraction systems: the LOD is much higher than that expected from the sensitivity coefficient, and the calibration range is shortened. Organo-aqueous mixtures are something in between aqueous solutions and organic systems, which results in mutual cancellation of good and bad effects, and SE is close to 1, i.e., the LOD can be predicted from sensitivity coefficients. Thus, comparison of all the factors affecting the sensitivity and reproducibility shows that the LOD of a real procedure in TLS is mainly determined by the reproducibility of photothermal measurements (Fig. 3), which is in turn determined by the thermophysical properties of the solvent and the possible formation of heterogeneity during sample preparation. From this viewpoint, surfactant- and polymer-modified aqueous solutions seem to be promising, since they provide error curves more similar to aqueous solutions as compared to organo-aqueous and organic solvents, and are less affected by the procedure development than these two systems. TLS and TLM Table 1 shows that despite a very different approach to the procedure development in TLM and TLS, all of the parameters (normalized LODs, experimental photothermal enhancement factors, and RL and SE factors) are the same, provided the same procedure type is considered. This is also confirmed for the same reactions in TLS and TLM; the difference is governed by

Relative standard deviation

386

ANALYTICAL SCIENCES APRIL 2011, VOL. 27 In addition, the time for the adaptation of a procedure to TLS is much shorter due to much lower blank signals and quicker emulsion layering. The time for batch analysis is half that of solvent-extraction systems. Very low LODs in solvent-extraction systems are advantageous when dealing with low concentrations and low sample volumes such as in the case of limited amounts of natural or high-purity samples. In addition, they should be used when preconcentration cannot be performed in a simple manner. Their use under batch conditions seems to be disadvantageous, but their implementation in microfluidic applications is a very worthy choice. Still, the very topical problem is high blank signals, which cannot be overcome easily. PEGs in single-phase and extraction systems are rather complementary than competitive systems, and can be changed to one another if practice demands. Thus, their use in TLS is very promising, because they combine the advantages of extraction preconcentration, have the convenience of aqueous solutions, and are free of the drawbacks of organic solutions. It seems promising to implement the developed PEG-based procedures in a flow system so as to fully estimate their potential in μTAS, which would benefit from shorter and simpler sample preparation.

0.4

0.2

2 1 0.0 0

1

2

Iron concentration × 109/M

Relative standard deviation

0.6

0.4

Acknowledgements

2

0.2

1 0.0 0

2

4

This study was supported by the Russian Foundation for Basic Research, Project Nos. 10-03-01018-a and 09-03-92102-YaF_a and by Japan Society for the Promotion of Science under the Japan-Russia Research Cooperative Program.

Cobalt concentration × 10 /M 9

Supporting Information Fig. 3 Dependences of the reproducibility relative standard deviation of TLS on the metal concentration for (a) Fe(II) with 1,10-phenanthroline in aqueous solutions and (b) Co(III) with 2-nitroso-1-naphthol in chloroform; curves 1 are for solutions of pre-synthesized chelates, and curves 2 indicate the results of measurements of photometric reactions occurring during sample preparation. λe = 514.5 nm, P = 300 mW.

the excitation conditions alone, not by the method. These findings were confirmed by a comparison of the same solutions for the spectrometers used (Table S-4, Supporting Information). The comparison shows a linear correlation with the coefficient of 0.99, and the coefficients of correlation for Fe and Co differ insignificantly. From the viewpoint of TLS as a detector in HPLC, FIA and electrophoresis and in microfluidic applications in TLM, Table 1 shows that despite the complications caused by the flow, the procedure parameters within the same group are mainly the same, which is especially interesting for extraction– photothermal procedures.

Conclusions As a whole, the sensitivity of TLS in surfactant and especially PEG solutions is comparable to that in organo-aqueous systems. The comparison of extraction systems shows that PEG-based extraction provides much better analytical parameters as compared to solvent extraction due to much better reproducibility of measurements at low concentrations and simpler sample preparation, despite a lower increase in the sensitivity coefficient.

Supporting information is available free of charge via the Internet at http://www.jsac.or.jp/analsci/.

References 1. R. D. Snook and R. D. Lowe, Analyst, 1995, 120, 2051. 2. S. E. Bialkowski, “Photothermal Spectroscopy Methods for Chemical Analysis”, 1996, Wiley-Interscience Publication, New York. 3. M. A. Proskurnin and V. V. Chernysh, J. Anal. Chem., 2000, 55, 943. 4. Y. Kikutani, H. Hisamoto, M. Tokeshi, and T. Kitamori, Lab Chip, 2004, 4, 328. 5. F. Kitagawa, T. Tsuneka, Y. Akimoto, K. Sueyoshi, K. Uchiyama, A. Hattori, and K. Otsuka, J. Chromatogr., A, 2006, 1106, 36. 6. Y. Tanaka, M. N. Slyadnev, A. Hibara, M. Tokeshi, and T. Kitamori, J. Chromatogr., A, 2000, 894, 45. 7. K. Uchiyama, A. Hibara, H. Kimura, T. Sawada, and T. Kitamori, Jpn. J. Appl. Phys. 1, 2000, 39, 5316. 8. P. A. Auroux, D. Iossifidis, D. R. Reyes, and A. Manz, Anal. Chem., 2002, 74, 2637. 9. D. R. Reyes, D. Iossifidis, P. A. Auroux, and A. Manz, Anal. Chem., 2002, 74, 2623. 10. Y. Tanaka, K. Sato, T. Shimizu, M. Yamato, T. Okano, and T. Kitamori, Biosens. Bioelectron., 2007, 23, 449. 11. T. T. J. Rossteuscher, A. Hibara, K. Mawatari, and T. Kitamori, J. Appl. Phys., 2009, 105, 1020. 12. H. Shimizu, K. Mawatari, and T. Kitamori, Anal. Chem.,

ANALYTICAL SCIENCES APRIL 2011, VOL. 27 2009, 81, 9802. 13. A. Smirnova, K. Shimura, A. Hibara, M. A. Proskurnin, and T. Kitamori, J. Sep. Sci., 2008, 31, 904. 14. A. Smirnova, M. A. Proskurnin, S. N. Bendrysheva, D. A. Nedosekin, A. Hibara, and T. Kitamori, Electrophoresis, 2008, 29, 2741. 15. M. A. Proskurnin, V. V. Chernysh, and V. A. Filichkina, J. Anal. Chem., 2004, 59, 818. 16. V. I. Grishko, V. P. Grishko, and I. G. Yudelevitch, “Lazernaya Analiticheskaya Termolinzovaya Spektroskopiya (Laser Analytical Thermal Lens Spectroscopy, in Russian)”, 1992, Institute of Inorganic Chemistry, Russian Academy of Sciences, Siberian Branch, Novosibirsk. 17. D. S. Kliger, “Ultrasensitive Laser Spectroscopy”, 1983, Academic Press, New York. 18. J. Georges, Spectrochim. Acta, Part A, 2008, 69, 1063. 19. M. A. Proskurnin, S. N. Bendrysheva, V. V. Kuznetsova, A. A. Zhirkov, and B. K. Zuev, J. Anal. Chem., 2008, 63, 1168. 20. N. Arnaud and J. Georges, Spectrochim. Acta, Part A, 2001, 57, 1295. 21. M. Franko and C. D. Tran, J. Phys. Chem., 1991, 95, 6688. 22. C. D. Tran and T. A. Van Fleet, Anal. Chem., 1988, 60, 2478. 23. D. C. Tsar’kov, A. A. Zhirkov, and M. A. Proskurnin, Vestnik MGU Ser. 2, 2007, 48, 257. 24. E. B. Sandell and H. Onishi, “Photometric Determination of Traces of Metals”, 1978, Wiley, New York. 25. V. A. Filichkina, A. G. Abroskin, Y. A. Barbalat, T. V. Belyaeva, M. A. Proskurnin, and V. M. Savostina, J. Anal. Chem., 1992, 47, 368. 26. A. Hibara, M. Tokeshi, K. Uchiyama, H. Hisamoto, and T. Kitamori, Anal. Sci., 2001, 17, 89. 27. T. Minagawa, M. Tokeshi, and T. Kitamori, Lab Chip, 2001, 1, 72. 28. M. Tokeshi and T. Kitamori, Prog. Nucl. Energ., 2005, 47, 434. 29. M. Tokeshi, T. Minagawa, and T. Kitamori, J. Chromatogr., A, 2000, 894, 19. 30. M. Tokeshi, T. Minagawa, K. Uchiyama, A. Hibara, K. Sato, H. Hisamoto, and T. Kitamori, Anal. Chem., 2002, 74, 1565. 31. M. Yamauchi, K. Mawatari, A. Hibara, M. Tokeshi, and T. Kitamori, Anal. Chem., 2006, 78, 2646. 32. K. Sato, M. Tokeshi, T. Kitamori, and T. Sawada, Anal. Sci., 1999, 15, 641. 33. H. M. Sorouraddin, A. Hibara, M. A. Proskurnin, and T. Kitamori, Anal. Sci., 2000, 16, 1033. 34. V. V. Chernysh, M. Y. Kononets, M. A. Proskurnin, S. V.

387

35. 36. 37.

38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56.

Pakhomova, V. V. Komissarov, and A. I. Zatsman, Fresenius J. Anal. Chem., 2001, 369, 535. A. D. Khrycheva, D. A. Nedosekin, M. A. Proskurnin, M. Y. Kononets, S. V. Pakhomova, and W. Faubel, Appl. Spectrosc., 2008, 62, 450. M. Y. Kononets, S. V. Pakhomova, A. G. Rozanov, and M. A. Proskurnin, J. Anal. Chem., 2002, 57, 586. Intenational Union of Pure and Applied Chemistry, “Compendium of Analytical Nomenclature”, ed. J. Inczedy, T. Lengyel, and A. M. Ure, 3rd ed., 1998, Blackwell Science, Oxford. A. A. Zhirkov, D. C. Tsar’kov, V. M. Shkinev, and M. A. Proskurnin, J. Anal. Chem., 2012, 67, in press. A. Smirnova, V. V. Chernysh, and M. A. Proskurnin, J. Anal. Chem., 2004, 59, 424. A. A. Dzyabchenko, M. A. Proskurnin, A. G. Abroskin, and D. V. Chashchikhin, J. Chromatogr., A, 1998, 827, 13. M. A. Proskurnin, V. V. Chernysh, and M. A. Kurzin, J. Anal. Chem., 2001, 56, 31. M. A. Proskurnin, M. A. Kurzin, and V. V. Chernysh, Vestnik MGU Ser. 2, 2001, 42, 116. Y. Kikutani, M. Ueno, H. Hisamoto, M. Tokeshi, and T. Kitamori, QSAR Comb. Sci., 2005, 24, 742. S. Legeai and J. Georges, Anal. Lett., 2000, 33, 3153. T. Imasaka, K. Miyaishi, and N. Ishibashi, Anal. Chim. Acta, 1980, 115, 407. S. Przeszlakowski and R. Kocjan, Chromatographia, 1982, 15, 717. M. Griffing and M. G. Mellon, Anal. Chem., 1947, 19, 1014. Z. Marchenko, “Fotometricheskoe Opredelenie Elementov (Photometrical Determination of Elements, in Russian)”, 1971, Mir, Moscow. V. M. Peshkova, V. M. Savostina, and E. K. Ivanova, “Oksimy (Oximes, in Russian)”, 1977, Nauka, Moscow. E. B. Sendel, “Kolorimetricheskoe Opredelenie Sledov Metallov (Colorimetric Determination of Metal Traces, in Russian)”, 1964, Mir, Moscow. I. V. Pyatnitskii, “Analytical Chemistry of Cobalt”, 1969, Ann Arbor, London. M. A. Proskurnin, V. B. Ivleva, N. Y. Ragozina, and E. K. Ivanova, Anal. Sci., 2000, 16, 397. H. H. Willard and S. Kaufman, Anal. Chem., 1947, 19, 505. M. A. Proskurnin, S. N. Bendrysheva, A. A. Zhirkov, and B. K. Zuev, J. Anal. Chem., 2008, 63, 1279. IUPAC, Stability Constant Database, Royal Society of Chemistry, SCQuery, Version 1.38, 1994. M. A. Proskurnin, M. Tokeshi, M. N. Slyadnev, and T. Kitamori, Anal. Sci., 2001, 17, S454.