Analytica Chimica Acta 689 (2011) 103–109
Contents lists available at ScienceDirect
Analytica Chimica Acta journal homepage: www.elsevier.com/locate/aca
Fabrication and characterization of a ␣,␤,␥,␦-Tetrakis(1-methylpyridinium4-yl)porphine/silica nanocomposite thin-layer membrane for detection of ppb-level heavy metal ions Kyaing Kyaing Latt, Yukiko Takahashi ∗ Top Runner Incubation Center for Academic-Industry Fusion, Nagaoka University of Technology, 1603-1 Kamitomioka, Nagaoka, Niigata 940-2188, Japan
a r t i c l e
i n f o
Article history: Received 8 September 2010 Received in revised form 23 December 2010 Accepted 14 January 2011 Available online 25 January 2011 Keywords: Cadmium ␣,␤,␥,␦-Tetrakis(1-methylpyridinium-4yl)porphine Nanocomposite membrane Dye nanoparticle coated test strip Colloidal silica Colorimetry
a b s t r a c t A new detection membrane for ﬁltration enrichment combined with colorimetric determination of Cd(II), Zn(II), Pb(II) and Cu(II) ions is presented. We have demonstrated the use of a dye nanoparticle coated test strip (DNTS) structured with a reagent layer for on-site analysis of trace metal ions. In this study, a [TMPyP/SA] DNTS coated with a nanocomposite layer (average thickness: 5.39 m) of ␣,␤,␥,␦-Tetrakis(1methylpyridinium-4-yl)porphine (TMPyP) and silica-SA on the top surface of a cellulose ester membrane ﬁlter was fabricated by a simple ﬁltration of an aqueous TMPyP/silica-SA nanocomposite dispersion through a membrane ﬁlter. The nanocomposite formation of cationic TMPyP and negatively charged colloidal SA (9–80 nm) was based on electrostatic interaction and was conﬁrmed in the 120–800 nm diameter range by a dynamic light scattering photometer (DLS). To optimize the DNTS nanocomposite layer, surface uniformity, mechanical strength, the percent retention of TMPyP, and sensitivity to Cd(II) detection for six DNTSs with ﬁve different types of silica were examined. A half[TMPyP/SA] DNTS with an average layer thickness of 2.60 m, which was prepared by controlling the amount of TMPyP and SA, demonstrated the highest sensitivity to Cd(II) ion because it had the lowest background absorbance. In addition, factors that affected the percent retention of TMPyP, such as pH and TMPyP/SA ratio, were determined. More than 99% of the TMPyP was retained on a membrane ﬁlter at pH 7.8 with a TMPyP and SA concentration of 2 × 10−5 M and 4 × 10−5 wt%, respectively. Filtration enrichment of 100 mL of an aqueous solution containing Cd(II), Zn(II), and Pb(II) at ppb levels was achieved by concentrating the metal ions in a nanocomposite layer (the effective TMPyP area was 1.77 cm2 , pH 10.2). The signaling surface changed from a brown color to green when the ions were captured. The percent extraction for metal ions on a half[TMPyP/SA] DNTS were estimated by TLC scanning and ICP-MS. It was observed that, when using the half[TMPyP/SA] DNTS, Cd(II) concentrations as low as 1 ppb were detectable at a ﬁltration rate of 4.0–5.0 mL min−1 . © 2011 Elsevier B.V. All rights reserved.
1. Introduction With industrial progress, pollution of the environment with toxic heavy metals has been increasing throughout the world. Detection of heavy metals is of great importance for environmental reasons. For example, heavy metals such as Cd and Pb, which are naturally occurring elements in the environment, are recognized to have a toxic effect on humans [1,2]. The World Health Organization (WHO) recommended a maximum of 3 ppb (g L−1 ) Cd and 10 ppb Pb in their guidelines for drinking water quality . In the establishment of environmental quality standards by the Ministry of the Environment (Government of Japan), both Cd and Pb concentrations below 10 ppb in ground water and below
∗ Corresponding author. Tel.: +81 258 47 9657; fax: +81 258 47 9610. E-mail address: [email protected]
(Y. Takahashi). 0003-2670/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.aca.2011.01.032
100 ppb in efﬂuent were acceptable for human health . The current standard techniques employed for trace heavy metal analysis include: atomic absorption spectrometry (AAS), inductively coupled plasma, and inductively coupled plasma-mass spectrometry (ICP-MS) [5–7]. These methods offer high sensitivity but require costly analytical instruments that are not capable of being used in the ﬁeld, involve time-consuming sample treatments and troublesome steps. To meet these demands, we have introduced a simpler onsite detection membrane with high sensitivity and selectivity to trace heavy metal ions at parts-per-billion (ppb) levels [8–10]. The membrane, loaded with a thin (several hundred nm) dye nanoparticle/nanoﬁber layer on the top surface, concentrates most target ions and provides a concentrated coloring layer. This layer is designated as the ‘dye nanoparticle coated test strip’ (DNTS). A DNTS is applicable not only to dip tests but also to ﬁltration enrichments in which target metal ions are concentrated by passing sample solu-
K.K. Latt, Y. Takahashi / Analytica Chimica Acta 689 (2011) 103–109
tions through it. Monitoring the Zn(II), Fe(II) and Hg(II) content by observing the color changes of a DNTS has been reported [8,9]. DNTS fabrication is simple and feasible, but applicable indicator dyes are limited to only those that are hydrophobic, despite the fact that most are water-soluble. In this work, we demonstrate a new type of DNTS that was coated with nanocomposite of a water-soluble metal indicator, ␣,␤,␥,␦-Tetrakis(1-methylpyridinium-4-yl)porphine (TMPyP) and a silica colloid that was used as a concentrated signaling layer. The nanocomposite formation of TMPyP and silica was based on the electrostatic interaction of TMPyP and the silica particles in the aqueous solution and the resulting nanocomposite was ﬁltered using a ﬁnely woven membrane ﬁlter, yielding a nanocomposite layer of [TMPyP/silica] on the top surface of the membrane. Solid oxides in an aqueous suspension are generally electrically charged due to the dissociation of surface MOH groups. Silica colloid SiO2 has an average isoelectric point (pH0 ) of 2.0 ± 0.2 , is negatively charged from neutral to alkaline conditions, and works as a cation exchanger. On the other hand, a positively charged indicator dye, TMPyP has been used to detect metal ions in aqueous media due to the highly sensitive probing molecules for optical analysis [12–14]. Such kinds of TMPyP immobilization have been reported on chemically converted graphite sheets , on glass surfaces as a grafted monolayer or on a sol–gel matrix  and polymeric matrix . In our work, a layer of TMPyP–silica composite was fabricated on a water-permeable mixed cellulose ester membrane ﬁlter and demonstrated for on-site analysis of trace metal ions. A sample solution was ﬁltered through the composite DNTS and a signal produced when the target metal ions present were concentrated in the TMPyP/silica nanocomposite layer. Detection of metal ions at ppb levels can be achieved by ﬁltration enrichment combined with simultaneous colorimetric analysis. The presence of Cd, Zn, Pb and Cu on the DNTS was observed as a color change from brown to green or orange. The fabrication and characterization of [TMPyP/silica] DNTSs for Cd(II), Zn(II), Pb(II) and Cu(II) detection are reported here. The optimum conditions, such as the selection of silica, ratio of TMPyP to silica, solution pH, pore size of membrane ﬁlter, and ﬂow rate in metal detection were examined. Membrane characterization and detection sensitivity to these ions were examined based on the uniformity, strength of the [TMPyP/silica] layer, retention percentage of TMPyP and percent extraction of the metal ions. 2. Experimental 2.1. Reagents and materials ␣,␤,␥,␦-Tetrakis(1-methylpyridinium-4-yl)porphine ptoluenesulfonate (TMPyP) was purchased from Tokyo Chemical Industry Co., Ltd. Colloidal silica particles (4–40 nm diameter) were from the Cataloid-S and Fine Cataloid series of Catalyst & Chemicals Ind. Co., Ltd. All chemicals used in this work, except the
silica, were of analytical grade. Ultrapure water (Simplicity UV, −18.2 M cm, Millipore), on demand from pretreated water by reverse osmosis, was used for all dilutions. The metal ion standard solutions (1000 ppm) for the metal ion detection experiments were purchased from Wako Pure Chemical Industries Ltd. Mixed cellulose ester membrane ﬁlters (47 mm diameter and 110 m thick) with 0.1 and 0.2 m pore sizes were from Advantec Co. and were used as the substrates for the TMPyP/silica nanocomposite DNTSs. 2.2. Buffer solutions Buffer solutions with pH values of 2.0, 3.46, 4.7, 6.15, 6.8, 7.8, 8.4, 9.2, 10.2, 10.8 and 12.1 were prepared to determine the optimum reaction conditions for TMPyP/silica DNTS preparation and metal ion detection. The buffer solutions were prepared from either sodium acetate/acetic acid, 2-(N-Morpholino)ethanesulfonic acid (MES), KH2 PO4 /K2 HPO4 , Tris(hydroxymethyl)aminomethane (Tris), borax, Tris(hydroxymethyl)methylaminopropane sulfonic acid (TAPS), NaHCO3 /Na2 CO3 , or 3-(cyclohexylamino) propanesulfonic acid (CAPS) by adjusting the pH with either 0.1 M HCl or 0.1 M NaOH. 2.3. Preparation and characterization of TMPyP/silica nanocomposite DNTS Five kinds of silica, two irregular shaped silica colloids (USBB120, F-120) and three spherical shaped silica colloids (SI-350, SI-550, SA), each with different particle diameters (4–40 nm, details are shown in Table 1), were examined for TMPyP/silica nanocomposite DNTS preparation. First, 10 mL of nanocomposite aqueous solution was prepared by mixing 100 L of 2 mM TMPyP and 4 × 10−5 wt% of silica at pH 7.8 with 0.01 M Tris buffer solution. Then, by passing the solution through a substrate membrane under 0.1 MPa suction pressure at 20 ◦ C, roughly 100% of the TMPyP/silica nanocomposite was coated as thin layer onto the surface of the membrane ﬁlter through surface ﬁltration (details are described in Section 3.1.1). The resulting DNTS (35 mm in diameter) was dried at room temperature and stored, under vacuum, in a gas barrier aluminum bag with an oxygen absorber. The DNTS was denoted as [TMPyP/silica] and had a thickness of approximately 5.4 m. To compare the thickness of the [TMPyP/SA] DNTS nanocomposite layer with respect to the Cd(II) detection sensitivity, the amounts of TMPyP and silica SA were reduced by half and another DNTS (denoted as half[TMPyP/SA]) was prepared with an approximate 2.60 m thickness. The particle sizes of the TMPyP/silica nanocomposite and silica colloids were checked using a dynamic light scattering photometer (DLS). The mechanical strength of the nanocomposite layers was characterized using a pencil hardness test. The maximum scratch resistance of the layer was determined through visual inspection. The performance of Cd(II) detection with half[TMPyP/SA] DNTS
Table 1 Properties of [TMPyP/silica] nanocomposite DNTSs. Silicaa
Particle diameter (nm)
USBB-120 F-120 SI-350 SI-550 SA SA
20–40 5–7 7–9 4–6 10–14 10–14
Surface ﬂatnessb (S.D.)
Pencil hardness test (maximum grade)
Retained TMPyP on substrate (%)
Integrated area by TLC at 485 nm (detection of Cd(II) at 20 ppb)
[TMPyP/USBB] [TMPyP/F] [TMPyP/SI350] [TMPyP/SI550] [TMPyP/SA] Half[TMPyP/SA]
0.037894 0.085952 0.051110 0.046762 0.039443 0.039743
2H 3H 2H 3H 3H 4H
98.1 100 100 98.5 100 99.9
1914 4242 4715 2763 5412 9151
The data were abstracted from the catalogue of Catalyst & Chemicals Ind. Co., Ltd. The data points were picked up in the range of 15 mm to the left and right from center of the DNTS.
K.K. Latt, Y. Takahashi / Analytica Chimica Acta 689 (2011) 103–109
was compared with those of [TMPyP/silica] by following the below mentioned procedure (see Section 2.4) and evaluated by thin-layer chromatography (TLC) scanning. 2.4. Detection of Cd(II), Pb(II), Zn(II) and Cu(II) ions Detection of ppb levels of the target metal ions was achieved through ﬁltration enrichment combined with colorimetric determination. The DNTS was cut into a circle (25 mm in diameter) and sandwiched between a ﬁlter holder with an attached suction ﬂask. Individual 100-mL aliquots of aqueous solutions containing Cd(II), Pb(II), Zn(II) or Cu(II) ion buffered at pH 10.2 with 0.01 M Na2 CO3 /NaHCO3 were ﬁltered through the DNTS (effective area 1.77 cm2 ) under a constant suction rate of 4.0–5.0 mL min−1 at 20 ◦ C. The brown color of the DNTS changed to green when Cd(II), Pb(II), or Zn(II) ions were concentrated within the [TMPyP/silica] nanocomposite layer. For the detection of Cu(II) ion, the DNTS color changed to orange. The DNTS percent extraction values for Cd(II) were calculated using data collected from both TLC scanning and ICP-MS. 2.5. Instruments The particle sizes of the silica SA colloids and TMPyP/SA composite were observed using a DLS particle size analyzer (ELS Z; Otsuka Electronics Co. Ltd.). The mechanical strengths of the [TMPyP/silica] and half[TMPyP/SA] nanocomposite layers were tested using a Japan Industrial Standards pencil hardness test (JIS K5600-5-4). Reﬂection–absorption spectra were measured using a multi-channel spectrophotometer (MCPD-3700; Otsuka Electronics Co. Ltd.) equipped with a Y-type optical ﬁber as the light projector and receiver. The reﬂection absorbance A is deﬁned by A = −log R/Ro , where R is the reﬂectance of the colored sample and Ro is that of colorless substrate. For collection of a cross sectional image of a DNTS and the distribution proﬁle of the Si and Cd, a ﬁeld emission scanning electron microscope (FE-SEM, S-800; Hitachi Ltd.) equipped with an energy dispersion X-ray spectrometer (EDX, Horiba EMAX) was used. During membrane preparation, the percent of TMPyP retained on a substrate membrane ﬁlter was examined using a UV–vis spectrophotometer (UV-1800; Shimadzu Corp.). A TLC scanner (Dual Wavelength Flying Spot Scanning Densitometer, CS-9300PC; Shimadzu Corp.) was employed to determine the uniformity and color intensity of the initially prepared DNTSs and the DNTSs after metal ion extraction. The signal intensity was measured in reﬂection–absorption mode with linear scanning of a 420 nm light beam (1.0 mm × 1.0 mm) for the prepared DNTSs and a 485 nm light beam (1.0 mm × 16 mm) for the metal ion-detected DNTSs. The measurements were collected as the peak area of a color spot (see Fig. 7(b)). The Cd(II) extracted on a DNTS was calculated from the Cd(II) concentration in the ﬁltrate as determined by ICP-MS (JMS-PLASMAX2; JEOL). 3. Results and discussion 3.1. DNTS fabrication 3.1.1. Preparation of [TMPyP/silica] nanocomposite The particle size distributions of a mixture of 2 × 10−5 M TMPyP and 4 × 10−5 wt% silica SA at pH 7.8 and a solution of 4 × 10−4 wt% silica SA were measured by DLS. The concentration of the colloidal SA was 10 times higher than that of the [TMPyP/SA] mixture because the intensity of the scattered light from the small particles was not sufﬁcient to determine the particle size. Fig. 1 shows that the particle size of the TMPyP/SA mixture was signiﬁcantly
Fig. 1. Particle size distributions of SA silica colloid () and TMPyP/SA nanocomposite ( ). Colloidal SA dispersion contained 4 × 10−4 wt% SA (pH 9.0–10.0). TMPyP/SA dispersion contained 2 × 10−5 M TMPyP and 4 × 10−5 wt% SA (pH 7.8).
larger compared to the colloidal SA. It was noted that aggregation occurred when the positively charged TMPyP was added to the silica SA dispersion and a nanocomposite was formed based on the electrostatic interaction. The diameter range of the TMPyP/SA nanocomposite was between 120 and 800 nm while the colloidal SA diameters were between 9 and 80 nm. Measurement of the potential was attempted; however, the measurement was unsuccessful due to weak scattering signals. Only a simple ﬁltration of the [TMPyP/silica] nanodispersion through the membrane ﬁlter was required to form a [TMPyP/silica] nanocomposite layer on the surface of the substrate. A crosssectional SEM image of a half[TMPyP/SA] nanocomposite layer on a membrane ﬁlter is shown in Fig. 2. The average thickness of the [TMPyP/SA] nanocomposite layer was 5.39 m (n = 7, 6.3% RSD). Similarly, by reducing the concentration of the TMPyP and SA by half, the average thickness of the half[TMPyP/SA] was observed to be 2.60 m (n = 8, 15% RSD). The surface densities of TMPyP were 2.08 × 10−8 mol cm−2 and 1.04 × 10−8 mol cm−2 for [TMPyP/SA] and half[TMPyP/SA], respectively. In addition, the elemental distribution of the whole cross-sectional area of the [TMPyP/SA] nanocomposite layered membrane was determined by EDX. The distribution of silica was dominant in the nanocomposite layer and no Si signal was observed within the 110 m thickness of the substrate membrane ﬁlter. Therefore, it can be conﬁrmed that the TMPyP/SA nanoparticle layer was located only on the top surface of the membrane ﬁlter. 3.1.2. Selection of colloidal silica The properties of the [TMPyP/silica] DNTSs prepared are summarized in Table 1. According to the pencil hardness test, the nanocomposite layers of [TMPyP/silica] DNTSs can resist the 2H–3H grade at maximum and the half[TMPyP/SA] DNTS can resist the 4H grade. These results indicate that the nanocomposite layers are ﬁrmly coated on the substrate and will not easily peel off. Furthermore, the ﬂatness of the composite layers was estimated as the standard deviation (S.D.) of the observed values from TLC scanning with a 1 × 1 mm light beam at 420 nm. The data points for the surface ﬂatness calculations were collected between 15 mm to left and 15 mm to right of center on the DNTS. According to the S.D. values, almost all of the top surfaces of the [TMPyP/silica] DNTSs are fairly uniform. Particularly, lower S.D. values were observed for the
K.K. Latt, Y. Takahashi / Analytica Chimica Acta 689 (2011) 103–109
Fig. 2. SEM images of (a) whole cross section of half[TMPyP/SA] DNTS and (b) the magniﬁed view of half[TMPyP/SA] nanocomposite layer.
[TMPyP/USBB], [TMPyP/SA] and half[TMPyP/SA] than the others, indicating better uniformity. The retention of TMPyP on the substrates was checked for all of the [TMPyP/silica] DNTSs and it was found that 98–100% of the TMPyP was maintained on all DNTSs. For Cd(II) sensing at 20 ppb, [TMPyP/SA] demonstrated a higher Cd(II) color intensity compared to the other DNTSs. For the half[TMPyP/SA] DNTS, where the amounts of TMPyP and silica were decreased by half, the yellowish-brown background color was lower and a more distinct green color of the Cd(II)–TMPyP complex was observed. Based on all considerations listed in Table 1, the [TMPyP/SA] and half[TMPyP/SA] DNTSs were chosen to evaluate the other factors in this study. 3.1.3. Inﬂuence of pH The pH dependence of the [TMPyP/SA] DNTS fabrication was evaluated for pH values between 2 and 10. The percent of retained TMPyP on the membrane ﬁlters is shown in Fig. 3. Notably, for pH values between 6.8 and 7.8, 99.8% of the TMPyP was retained. In principle, the amount of negative surface charge on the silica particles increases with a rise in pH, based on the isoelectric point (pH0 1.8–2.2) . The water-soluble, positively-charged metal indicator TMPyP binds onto the surface of the negatively-charged silica due to electrostatic interactions, yielding a TMPyP/SA nanocomposite. When the pH range was between 6 and 8, the repulsive forces between the silica particles were drastically reduced due to the surface adsorption of the positively-charged TMPyP, after which aggregation readily occurred. In the alkali region (pH > 10), silica particles will dissolve and become unstable.
Fig. 3. Inﬂuence of pH on retention percentage of TMPyP in [TMPyP/SA] DNTS fabrication. TMPyP/SA nanocomposite dispersions were prepared in 10 mL of 2 × 10−5 M TMPyP and 4 × 10−5 wt% SA with relevant buffer solutions given in Section 2.2.
3.1.4. Inﬂuence of the TMPyP to silica ratio The effect of the ratio of TMPyP to silica on the percent retention of TMPyP was examined (Fig. 4). Initially, experiments using [TMPyP/SA] and half[TMPyP/SA] DNTSs were conducted by maintaining the silica concentration at 4 × 10−5 wt% and 2 × 10−5 wt%,
Fig. 4. (a) Inﬂuence of TMPyP concentration on retention percentage of TMPyP in [TMPyP/SA] (–䊉–) and half[TMPyP/SA] (- -- -) DNTS fabrication; [SA]constant = 4 × 10−5 wt% for [TMPyP/SA] and 2 × 10−5 wt% for half[TMPyP/SA], (b) inﬂuence of silica concentration in [TMPyP/SA] DNTS fabrication; [TMPyP]constant = 2 × 10−5 M.
K.K. Latt, Y. Takahashi / Analytica Chimica Acta 689 (2011) 103–109
respectively, and varying only the TMPyP concentration. For each of the above experiments, the percentages of TMPyP retained were plotted (Fig. 4(a)) as a function of the initial TMPyP concentration in a 10 mL solution. At initial concentrations of 3 × 10−5 M for the [TMPyP/SA] DNTS and 1 × 10−5 M for the half[TMPyP/SA] DNTS, respectively, the observed growth was consistent with that of the nanocomposite when the TMPyP concentrations were increased. About 99.9% of the TMPyP was retained on both DNTSs at these points. Higher concentrations of TMPyP provided an overabundant amount of cations, which overloaded the silica surface with TMPyP and caused it to take on a positive charge; therefore, the nanocomposite particles repelled each other. Because no further aggregation occurred, any excess cations dissolved in solution passed through the membrane. The effect of varying the silica concentration during [TMPyP/SA] DNTS fabrication on the percent TMPyP retention was also studied and the results are shown in Fig. 4(b). The concentration of TMPyP was held constant at 2 × 10−5 M. The results showed that over 99.3% TMPyP retention was achieved for silica SA within a range of 4–6 × 10−5 wt%. 3.1.5. Pore size of the substrate membrane The ﬁltration rate is as important as the quantitative retention of TMPyP for on-site analysis of trace metal ions. [TMPyP/SA] DNTSs with 0.1, 0.2, 0.45, 0.65, 0.8 and 1.0 m pore-size mixed cellulose ester membrane ﬁlters were prepared as described in Section 2. Highly quantitative values of TMPyP retention were observed as follows: 100% for 0.1 m, 99.8% for 0.2, 0.45 and 0.65 m, 99.7% for 0.8 m and 99.6% for 1.0 m pore sizes of the membrane ﬁlters, respectively. 3.2. Performance of the [TMPyP/SA] DNTS 3.2.1. Cd(II) ion determination as a function of reaction pH To determine the optimum reaction conditions, ﬁltration experiments using 100 mL aliquots of Cd(II) solutions at a ﬁltration rate of 4.0–5.0 mL min−1 , with pHs between 2 and 12, were analyzed using [TMPyP/SA] DNTSs. Fig. 5 shows both the extraction percentage of 20 ppb Cd(II), as conﬁrmed by color intensity measurements using TLC scanning at 485 nm, and the stability of the [TMPyP/SA] nanocomposite layers versus the solution pH. The amount of TMPyP remaining on the membrane ﬁlter was calculated from the ﬁltrates using UV–vis spectrophotometry after the ﬁltration experiments. The pH dependence of the remaining TMPyP
Fig. 5. Inﬂuence of pH on TLC integrated intensity of Cd(II) detected DNTS at 485 nm (- -- -) and remaining TMPyP% (–䊉–) in 20 ppb Cd(II) detection with [TMPyP/SA] DNTS. Sample solution: 100 mL, pH 10.2.
followed the same trend as that of the TMPyP retained during DNTS fabrication described in Fig. 3. However, the TMPyP/SA nanocomposite layer showed more resistance at acidic pHs compared to that of the aggregation process during DNTS fabrication. To discuss the pH dependence of Cd extraction described as TLC signal intensities at 485 nm (Fig. 5), the relationship between the pH effect on the green coloration of [TMPyP/SA] DNTS and the cadmium hydroxide complex formation was considered. The stability constants of the cadmium species, CdOH+ and Cd(OH)2 , are log K1 = 3.9 and log ˇ2 = 7.2 from analytical chemistry handbook , respectively, and molar fractions of divalent cadmium, CdOH+ , and Cd(OH)2 versus pH were plotted. At a pH below 7, cadmium of 10 ppb exists predominately as the free divalent ion, and at pH 7–12 as CdOH+ and no Cd(OH)2 as well. No Cd(II)–TMPyP complex signal observed at pH < 7 shows that free Cd(II) ion cannot react with TMPyP contained in [TMPyP/SA] nanocomposite layer under a ﬂow condition at 4–5 mL min−1 . In an aqueous phase, to reach reaction of a molecular TMPyP and Cd(II) equilibrium at pH 7–8 more than 20 h was reported to be necessary after mixing 1 M TMPyP and 3 M Cd(II) . On the other hand, our experiment in an aqueous solution at pH 10 revealed the complex formation of 2 M TMPyP and 0.18 M (20 ppb) Cd(II) reaches equilibrium within 40 s. CdOH+ species, which exists at alkaline pH predominantly, is expected to accelerate the complexation with TMPyP drastically in comparison with Cd(II) ion. A ﬂow rate of 4–5 mL min−1 is supposed to be not enough for the complex formation with Cd(II) ion and to be enough for that with CdOH+ , resulting in characteristic pH effect on Cd extraction. The Cd(II)–TMPyP complex was quantitatively formed at pH values between 9 and 10.5 (almost 80%), where the CdOH+ is overwhelmingly present, as described above. The decrease of TLC integrated intensity at pH > 10.5 is assumed due to dissolution of colloidal silica. The optimum pH for Cd(II) extraction was found to be 10.2 and further extraction experiments were performed at this pH value. 3.2.2. Inﬂuence of ﬂow rate on Cd(II) detection The function of ﬂow rate on Cd(II) ﬁltration enrichment (20 ppb) was investigated using half[TMPyP/SA] DNTSs with different substrate pore sizes as well as by controlling the suction pressure. To achieve lower ﬂow rates (0.7–4.2 mL min−1 ), a 0.1 m pore size half[TMPyP/SA] DNTS was used under suction pressures of 0.025, 0.04, 0.06, 0.08 and 0.1 MPa. For higher ﬂow rates (5.6–31.3 mL min−1 ), a 0.2 m pore size half[TMPyP/SA] DNTS was used under pressures of 0.02, 0.025, 0.03, 0.04, 0.06, 0.08 and 0.1 MPa. The integrated color intensity of Cd(II) on the DNTS, detected at 485 nm, versus ﬂow rate is shown in Fig. 6. Simultaneously, the percent of Cd(II) extracted was conﬁrmed using ICP-MS and is presented in the ﬁgure. These two results demonstrate that Cd(II) extraction gradually decreased with increasing ﬂow rate (i.e., the ﬁltration time decreased). In Fig. 6, it was noted that the results of these two observations were slightly different and shifted at 15 mL min−1 . The ICP-MS results were closer to the actual extraction amounts of Cd(II) into the TMPyP/SA nanocomposite layer and the extraction percent at ﬂow rates above 15 mL min−1 stayed constant. On the other hand, the curve of TLC scanning sequentially declined along ﬂow rate, which is because TLC scanner only measures the light reﬂected from the top surface of the sample and cannot detect the signals inside of the nanocomposite layer with 2.60 m in thickness. Since light is signiﬁcantly attenuated when it travels into the solid substrate, the reﬂection–absorption intensity drastically decays along with the depth direction. To understand the TLC scanner data more clearly, three cross sectional graphics of Cd(II) detected using a half[TMPyP/SA] DNTS at low (Fig. 6(a)), medium (Fig. 6(b)), and high ﬁltration speed (Fig. 6(c)) are illustrated. At a low ﬁltration speed (Fig. 6(a)), Cd(II) ions are adsorbed and enriched mainly in
K.K. Latt, Y. Takahashi / Analytica Chimica Acta 689 (2011) 103–109
Fig. 7. (a) A photograph of Cd(II) ion detected half[TMPyP/SA] DNTSs. Sample volume: 100 mL, suction pressure: 0.1 MPa, pH 10.2, (b) photographs and TLC densitograms for (1) the prepared DNTS and (2) the metal ion-detected DNTS. Scanned condition for (1): linear scan in reﬂection–absorption mode at 420 nm, light beam size: 1.0 mm × 1.0 mm. Scanned condition for (2): linear scan in reﬂection–absorption mode at 485 nm, light beam size: 1.0 mm × 16.0 mm. Fig. 6. Extraction percentage of 20 ppb of Cd(II) in the half[TMPyP/SA] DNTS by ICPMS (- -- -) and TLC integrated intensity at 485 nm (–䊉–) as a function of ﬂow rate. Sample volume: 100 mL, [Cd(II)] = 20 ppb, pH 10.2, suction pressure: 0.02–0.1 MPa. DNTS with 0.1 m pore was used for the ﬂow rate range of 0.7–4.24 mL min−1 and DNTS with 0.2 m pore size was for the range of 5.56–31.25 mL min−1 . See explanations of illustration (a), (b) and (c) in text.
the upper part of the TMPyP/SA nanocomposite layer through complex formation. As the ﬁltration speed increased (Fig. 6(b) and (c)), adsorption of Cd(II) is occurred mainly in the middle and lower parts to the TMPyP/SA nanocomposite layer where low-intensity light is present. Basically, acquisition of Cd(II) into the nanocomposite layer is based on the complexation of Cd(II) in an aqueous solution and TMPyP immobilized in the nanocomposite layer. Typically, the complexation reaction of metal ions with porphorins is incomparably slower than other noncyclic ligands because of low ﬂexibility of the cyclic structure. The formation of Cd–TMPyP complex is also slow to such a extent that accelerators are necessary to incorporate Cd(II) into TMPyP ring effectively . In case of dithiozone nanoﬁber DNTS for the detection of Hg(II) we previously reported, because both rates of 1:1 and 1:2 complex formations of Hg(II) and dithizone are quite high and the thickness of dithizone nanoﬁber layer (440 nm) is thick enough, quantitative and constant extraction by TLC scanning was observed irrespective of ﬂow rate, in the range of 1.3–9.3 mL min−1 . Accordingly, the thickness of the TMPyP/SA nanocomposite layer (2.60 m) and the slow rate of the complexation reaction of Cd(II) and TMPyP make the difference of two curves in Fig. 6. In this case, lower speed ﬁltration has an analytical advantage for both visual and photometric detection, 0.1 MPa pressure (average ﬂow rate: 4.48 mL min−1 ) using a 0.1 m pore size half[TMPyP/SA] DNTS was selected as the standard conditions for ﬁltration experiments for Pb(II), Zn(II) and Cu(II) detection. 3.2.3. Determination of Cd(II) with half[TMPyP/SA] DNTS Fig. 7(a) demonstrates the color series with half[TMPyP/SA] DNTSs obtained by ﬁltration of various concentrations of Cd(II) in a 100 mL of sample solution, and depicts a detection limit of 1 ppb. On the other hand, a similar color proﬁle as the function of Cd(II) concentration using [TMPyP/SA] DNTS resulted in a detection limit of 10 ppb by eye observation. Additionally, integrated intensities of relative peak area of a yellowish-brown initial half[TMPyP/SA] and a brown initial [TMPyP/SA] were estimated 79,359 and 86,170, respectively, by TLC scanning at 420 nm. The facts suggested that
the lower background absorbance of half[TMPyP/SA] emphasizes a green color signal of Cd(II)–TMPyP complex and enables the lower detection limit. Fig. 7(b) shows the TLC scanning and the illustrations of the integrated intensity of peak area. A calibration curve for Cd(II) detection using a half[TMPyP/SA] DNTS was calculated using integrated signal intensities, as measured by TLC, for samples between 1 and 200 ppb. The equation of the non-linear calibration curve, at 485 nm detection, is expressed as I = 16506.44 (1 − e−0.05c ), where I is the integrated intensity observed by TLC and c is the Cd(II) concentration. The correlation coefﬁcient (R) was 0.9929. The 3 detection limit based on eight blank responses was 0.102 ppb. Precision at 10 ppb was calculated by analyzing eight samples and the standard deviation was found to be 0.628 ppb. There were a few interfering species in the determination of Cd(II) at 10 ppb when using a half[TMPyP/SA] DNTS. Using a 1:1 ratio of Fe(III), Mn(II) and Al(III) to Cd(II), no interference was observed. However, the presence of 100-fold greater Fe(III) and Mn(II) concentration (1 ppm), compared to Cd(II) (10 ppb), seriously interfered with the Cd(II) detection by disturbing the green color of Cd(II) as well as by clogging the DNTS. The reason for this is because, at pH 10.2, hydroxide precipitates of iron and manganese are generated that stain yellow and brown, respectively. These precipitates adsorbed Cd(II) easily, and they also aggregated into particles and clogged the pore of the supporting membrane. Pre-ﬁltration with a mixed cellulose ester membrane ﬁlter (0.1 m pore) before Cd(II) detection was effective to curb the Mn(II) inﬂuence. For detection of Cd(II) in the presence of Mn(II), the percent recovery was drastically improved from 134.4% at 1.02 mL min−1 without pre-ﬁltration to 106.41% at 5 mL min−1 with pre-ﬁltration. No coloration was observed when 100-fold higher concentration of Al(III) was present. The addition of other heavy metal ions such as Pb(II), Zn(II), and Cu(II) at more than 100 ppb also seriously interfere to the detection of 10 ppb of Cd(II). In an attempt to remove the interferences, masking reagents such as 2-aminoethanethiol, 2,3-dimercapto-1-propanesulfonic acid and mercaptoacetic acid were tested, although it resulted in failure due to the usage of these reagents under alkaline conditions. The addition of a 1000fold increase in concentration for Ca(II), Mg(II), Na+ and K+ versus Cd(II) showed no interference to the determination of 10 ppb Cd(II). Additionally, an excess amount (290,000–3,500,000-fold) of several anions (Cl− , Br− , SO4 2− , CO3 2− , NO3 − , PO4 3− ) was tested and no inhibition was observed for the detection of 20 ppb Cd(II).
K.K. Latt, Y. Takahashi / Analytica Chimica Acta 689 (2011) 103–109
nanocomposite formation was conﬁrmed by DLS analysis as the aggregation of negatively charged SA colloid by addition of cationic TMPyP. Good uniformity and a ﬁrm coating with high mechanical strength for both the [TMPyP/SA] and half[TMPyP/SA] nanocomposite layers were observed. The DNTSs provided more than 99% TMPyP retention in the nanocomposite layer. The preparative conditions for a [TMPyP/SA] DNTS were optimal at pH 6.8–7.8 with a TMPyP concentration and SA ratio of 2 × 10−5 M and 4 × 10−5 wt%, respectively. The [TMPyP/SA] and half[TMPyP/SA] nanocomposite layers were localized on the top surface of the membrane ﬁlter with 5.39 and 2.60 m thickness, respectively. Filtration enrichment combined with colorimetric determination of Cd(II) in an aqueous solution was achieved at a ﬁltration rate of 4.0–5.0 mL min−1 . The higher sensitivity of the half[TMPyP/SA] DNTS was conﬁrmed by detection of Cd(II) at 1 ppb while the detection limit of the [TMPyP/SA] DNTS was 10 ppb. Furthermore, the TMPyP/SA nanocomposite DNTS is applicable for the detection of Pb(II), Zn(II) and Cu(II) ions in aqueous solution. Acknowledgements
Fig. 8. Reﬂectance–absorption spectra of 100 ppb Cd(II) (—), 100 ppb Pb(II) (– –· ·– –), and 1 ppm Cu(II) (- - -) detected half[TMPyP/SA] DNTS comparing with half[TMPyP/SA] (—) DNTS used: half[TMPyP/SA] DNTS (0.1 m pore), Sample volume: 100 mL, suction pressure: 0.1 MPa, pH 10.2.
3.2.4. Sensing of Pb(II), Zn(II) and Cu(II) ions According to Section 3.2.3, ﬁltration enrichment and colorimetric detection of Pb(II), Zn(II), and Cu(II) using [TMPyP/SA] DNTS were also available at ppb level. Similar to Cd(II), a green coloration was observed with Pb(II) and Zn(II) detection. Cu(II) ions were detected by a change in the DNTS color to orange. Fig. 8 shows the relative reﬂectance–absorption spectra of these metal ions detected using a half[TMPyP/SA] DNTS. The absorption spectra of a DNTS used for detection of Cd(II), Zn(II), Cu(II), and Pb(II) are similar to the spectra collected for the ions in aqueous solution. The Soret band of TMPyP, around 400–450 nm, was slightly red-shifted due to complex formation with Cd(II), Cu(II), and Pb(II). The Q bands were also red-shifted (550–700 nm) due to complexation with Cd(II), Cu(II), and Pb(II) and have characteristic wavelengths of reﬂection–absorption maximum (max ) and relative reﬂection–absorption intensity to standard white plate strongly dependent on the individual metal ions. The values of max are 590 and 645 nm for Cd(II), 595 and 665 nm for Pb(II), 570 nm for Zn(II) and 555 nm for Cu(II). Fig. 8 suggests the potential of simultaneous multi-wavelength spectrophotometry of these ions by comparing a few absorption intensities at inherent max . 4. Conclusion We have described the fabrication of nanocomposite DNTSs with water-soluble porphyrin TMPyP and silica SA coated on a ﬁnely woven cellulose type membrane ﬁlter. The TMPyP/SA
Y. T. is grateful to the special coordination fund from MEXT of Japan, which supports activities for the ‘Promotion of Independent Research Environment for Young Researchers’. This work was ﬁnancially supported by JFE 21st Century Foundation and the Environment Research and Technology Development Fund (B-1005) of the Ministry of the Environment, Japan. The authors thank Dr. Yoshito Wakui at the National Institute of Science and Technology for his support with ICP-MS measurements. References  http://www.cadmium.org/environment.html (International Cadmium Association).  L. Järup, Br. Med. Bull. 68 (2003) 167–182.  Guidelines for Drinking-water Quality, vol. 1, 3rd ed., World Health Organization, Geneva, 2008 (incorporating 1st and 2nd addenda).  http://www.env.go.jp/en/water (Ministry of the Environment, Government of Japan).  J. Feng Peng, R. Liu, J. Fu Liu, B. He, X. Lin Hu, G. Bin Jiang, Spectrochim. Acta Part B 62 (2007) 499–503.  M. Tuzen, M. Soylak, Anal. Chim. Acta 504 (2004) 325–334.  M. Lerchi, E. Bakker, B. Rusterholz, W. Simon, Anal. Chem. 64 (1992) 1534–1540.  Y. Takahashi, H. Kasai, H. Nakanishi, T.M. Suzuki, Angew. Chem., Int. Ed. 45 (2006) 913–916.  Y. Takahashi, S. Danwittayakul, T.M. Suzuki, The royal society of chemistry, Analyst 134 (2009) 1380–1385.  Y. Takahashi, T.M. Suzuki, Nanotechnology Applications for Clean Water, William Andrew, Norwich, New York, 2009, pp. 417–425 (Part 4).  G.A. Parks, Chem. Rev. 65 (1965) 177–198.  M. Biesaga, K. Pyrzynska, M. Trojanowicz, Talanta 51 (2000) 209–224.  K. Kawamura, S. Igarashi, T. Yotsuyanagi, Anal. Sci. 4 (1988) 175–179.  Y. Xu, L. Zhao, H. Bai, W. Hong, C. Li, G. Shi, J. Am. Chem. Soc. 131 (2009) 13490–13497.  D. Delmarre, R. Meallet, C. Bied-Charreton, R.B. Pansu, J. Photochem. Photobiol. A 124 (1999) 23–28.  A. Morales-Bahnik, R. Czolk, J. Reichert, H.J. Ache, Sens. Actuators B 13 (1993) 424.  Bunseki Kagaku Binran, 5th ed., Marzen, Tokyo, 2003, p. 648.  K. Kawamura, S. Igarashi, T. Yotsuyanagi, Inorg. Chim. Acta 360 (2007) 3287–3295.