ARTICLE pubs.acs.org/ac
Highly Sensitive Elemental Analysis for Cd and Pb by Liquid Electrode Plasma Atomic Emission Spectrometry with Quartz Glass Chip and Sample Flow Atsushi Kitano,† Akiko Iiduka,† Tamotsu Yamamoto,‡ Yoshiaki Ukita,† Eiichi Tamiya,§ and Yuzuru Takamura*,† †
School of Materials Science, Japan Advanced Institute of Science and Technology (JAIST), 1-1 Asahidai, Nomi, Ishikawa 923-1292, Japan ‡ Micro Emission Ltd. Ishikawa Create Lab, 2-13 Asahidai, Nomi, Ishikawa 923-1211, Japan § Department of Applied Physics, Graduate School of Engineering, Osaka University, 2-1 Yamada-Oka, Suita, Osaka 565-0871, Japan ABSTRACT: This paper describes the development of a highly sensitive liquidelectrode plasma atomic emission spectrometry (LEP-AES) by combination of quartz glass chip and sample flow system. LEP-AES is an ultracompact elemental analysis method, in which the electroconductive sample solution is put into a microfluidic channel whose center is made narrower (∼100 μm in width). When high voltage pulses (1500 V) are applied at both ends of the channel, the sample evaporates locally at the narrow part and generates plasma. By the emission from the plasma, elemental concentration is analyzed. In this paper, the limits of detection (LODs) were investigated in various conditions of accumulation time, material of the chip, and the sample flow. It was found that the long accumulation using the quartz chip with sample flow was effective to improve LOD. Authors suggested that this was because bubbles remaining after each plasma pulse were removed from the narrow channel by sample flow, resulting in highly reproducible plasma generation, to enable a high accumulation effect. Finally, LODs were calculated from a calibration curve, to be 0.52 μg/L for Cd and 19.0 μg/L for Pb at optimized condition. Sub-ppb level LOD was achieved for Cd.
T
he technology of elemental analysis is of great significance for commercial and industrial applications worldwide. Elemental analysis methods such as atomic absorption spectrometry (AAS),1 inductively coupled plasma atomic emission spectrometry (ICP-AES),2 and inductively coupled plasma mass spectrometry (ICP-MS)3 have been widely used for the measurement of trace element(s) in environmental fields. These methods are typically highly sensitive and accurate. However, these methods are not suitable for on-site, mobile, and in situ applications due to their size, initial and running costs, and requirement of operators. Recently, the miniaturization of measurement devices is making remarkable progress utilizing microelectro mechanical systems (MEMS),4,5 micro total analysis systems (μ-TAS),6 and lab-on-a-chip (LOC)7 13 technologies. For miniature plasma sources, a large number of microplasma sources have been developed and reported14 43 for the measurement of elements, molecules, and compounds. In 2003, a novel microplasma elemental analysis method was presented by Iiduka et al,46 namely liquid electrode plasma atomic emission spectrometry (LEP-AES). The principle of the LEP-AES is shown in Figure 1. The narrow channel is filled with an electroconductive liquid sample (a). When the voltage is applied from both ends of the channel, the electric field concentrates on the narrow central part, and the solution is locally heated by Joule heating (b). Subsequently, water evaporates and makes bubbles (c). The plasma is generated in the bubble, and r 2011 American Chemical Society
the elements in the sample liquid enter the plasma to emit light with the element-specific spectra (d). From the spectra, elemental analysis can be done. Compared with conventional ICP-AES,44,45 the advantages of LEP-AES are follows: (1) It requires no high power source, no plasma gas, such as Ar, that means no heavy cylinder, and no nebulizer, which needs frequent maintenance. (2) There is high signal intensity because samples are not diluted by carrier and plasma gas. (3) There is low background and low stray light due to no bright emission from Ar from plasma gas, which enables the usage of low specification spectrometer. (4) Samples of only a few microliters can be measured, so that a high ratio preconcentration becomes practical. Therefore, compact battery driven instruments can be easily achieved, with low energy consumption condition. Some studies using LEP-AES are already reported.46 55 Banno et al. reported determination of trace amount of Li and Na in ZrO2 ceramics utilizing no sensitivity of Zr by LEP-AES.50 Lead in soil was measured with pretreatment of solid phase extraction.53 Kumai et al. measured the excitation temperature dependence on channel size.54 Yamamoto et al. determined elements in soil samples Received: August 6, 2011 Accepted: November 4, 2011 Published: November 04, 2011 9424
dx.doi.org/10.1021/ac2020646 | Anal. Chem. 2011, 83, 9424–9430
Analytical Chemistry
ARTICLE
Figure 1. Principle of LEP. The narrow channel is filled with an electroconductive liquid sample (a). When the voltage is applied from both ends of the channel, the electric field concentrates on the narrow central part, and the solution is locally heated by Joule heating (b). Subsequently, water evaporates and makes bubbles (c). The plasma is generated in the bubble, and analyte gets into the plasma and emits light (d). Figure 3. Fabrication process for the quartz chip (a) and for PDMS chip (b).
design was also microfabricated in order to compare the accuracy, the limits of detection, to the quartz chip. The effects of chip material and accumulation with and without sample flow were investigated.
’ EXPERIMENTAL SECTION Figure 2. Chip design for the mask pattern (a) and detailed geometry around narrow channel (b).
in the mountain district of Shikoku using a handy-type LEP atomic emission spectrometer.55 The limits of detection (LODs) in those reports were, however, not sufficiently low, for example, around 1 ppm for Cd and Pb. In those reports, the used plasma chips were made of resins, which were considered to be damaged easily by heat of plasma. This should cause a variation of the intensity. From another viewpoint, in those reports, the voltage was applied as pulse repeatedly and the emission intensities from each pulse were accumulated to obtain a measured value. After each pulse, small remaining bubbles were frequently observed locating randomly at narrow channels. Due to no flow system in those reports, those bubbles stayed to next plasma generation by pulse, and should affect the nucleation condition of bubbles for plasma to change the intensity also. Those intensity changes areconsidered to degrade LOD and accuracy. In this report, quartz glass was used for the chip material due to its heat resistance. In order to confirm the effect of removing the air bubbles from the channel, emission intensity and its variation were measured with and without sample flow. The quartz glass was more permeable in the deep ultraviolet light region, so that the emission in the region can be detected more strongly than resin chip. A polydimethylsiloxane (PDMS) chip with the same
Chip Design. Figure 2 shows the design of the microfluidic channel used in this paper. The chip (20 mm 20 mm) has a microfluidic channel that is 100 μm at the narrow part and approximately 90 μm deep. For the chip material, we used quartz and PDMS. Chip Fabrication. The quartz chip fabrication process is shown in Figure 3a. The fabrication of the chips was carried out in a clean room environment for semiconductor processes (at JAIST, Japan). A piece of quartz glass substrate (20 mm 20 mm 0.5 mm) was washed using a solution of H2SO4/H2O2 (4:1, each EL grade, Kanto Chemical Co., Inc., Japan) and then rinsed using ultrapure water (ultrapure water system, PURIC-MX II and pure water equipment, PRO-0100-002, Organo Corporation, Japan). A 6 μm thick Cr thin film was sputtered on the top side of the quartz glass substrate as a metal mask using sputtering equipment (MNS-2000, ULVAC, Japan). The photoresist (OFPR-800 30 cP, Tokyo Ohka Kogyo, Co., Ltd., Japan) was spin-coated onto the Cr-sputtered substrate using a spin coater (1H-DXII, Mikasa, Japan). The coated photoresist thickness is about 1.2 μm. The mask pattern of the flow channel was used for the photolithography process. The mask pattern of the flow channel was transferred onto the photoresist by UV light exposure using a mask aligner (PEM-800, Union Optical Co., Ltd., Japan). After exposure, the pattern was developed using a developer (NMD-3, Tokyo Ohka Kogyo, Co., Ltd., Japan) to remove the exposed part. Then, the quartz chip was dipped into the diluted Cr etchant (Cr etchant MPM-E30, DNP Fine Chemicals 9425
dx.doi.org/10.1021/ac2020646 |Anal. Chem. 2011, 83, 9424–9430
Analytical Chemistry
ARTICLE
Figure 5. Typical emission spectra using PDMS chip for 0.1 mg/L Cd in a 0.1 mol/L HNO3 aqueous solution at 0 and 250 μL/min flow rate. Figure 4. Experimental setup used as the LEP-AES with the quartz and PDMS chips (a). The photograph of quartz chip (b) and PDMS chip (c) settled in chip holders.
Co., Ltd., Japan), and the Cr film of the flow channel pattern was removed. Here, the photoresist on the Cr material functions as a mask. After removing the Cr thin film, the photoresist was removed using an acetone (EL grade, Kanto Chemical Co., Inc., Japan) and ethanol (EL grade, Kanto Chemical Co., Inc., Japan) mixture. Next, the flow channel pattern on the quartz chip was dry-etched using this Cr film as a hard mask. For this dry etching, deep reactive ion etching (D-RIE) with high-density plasma (RIE-200iPB, SAMCO, Japan) was used. The flow channel pattern was vertically etched approximately 90 μm in depth on the quartz substrate in 5 h. After etching, the bottom of the channel was kept smooth enough to be optically transparent for detecting the plasma emission. The Cr mask film was removed using a sulfuric acid (H2SO4) and hydrogen peroxide (H2O2) mixture. Finally, the quartz glass chip with a flow pattern was bonded to quartz glass with the same dimensions and 2 holes (inlet and outlet) in a 1% hydrofluoric acid (aqueous HF, wet etching agent for semiconductors, Daikin Industries, Ltd., Japan) solution for 12 h. Figure 3b shows the fabrication process of a PDMS chip. Thick photoresist (SU-8 3050, permanent epoxy negative photoresist, Nippon Kayaku Co., Ltd., Japan) was spin-coated onto a Si wafer by a spin-coater (1H-DXII, Mikasa Co., Ltd., Japan). The pattern of the flow channel was transferred to the resist by photolithography using a mask aligner (PEM-800, Union Optical Co., Ltd., Japan). The photoresist on the Si wafer was developed using a developer (SU-8 developer, MicroChem Corp.) to be a mold. PDMS (Dow Corning Toray Silicone Co., Ltd., Japan) was cast into the mold and was cured in the oven at 75 °C for 90 min. The cured PDMS was peeled from the Si wafer, and two holes of 2 mm in diameter for the inlet and outlet were punched. The punched PDMS and another quartz glass substrate (20 mm 20 mm 0.5 mm) were bonded at room temperature. Here, as illustrated in Figure 3b, the upper part of the PDMS chip was made of PDMS, and the bottom part was made of quartz glass. The emission from plasma was designed to be measured from the bottom through the quartz glass to minimize the absorption of UV light. Sample Preparation. For sample preparation, solutions of Cd and Pb (concentrations from 0.001 to 1 mg/L) were prepared by dilution of 1000 mg/L standard solutions (Cd 1000, and Pb 1000, Kanto Chemical Co., Inc., Japan.), with 0.1 mol/L nitric acid (Kanto Chemical Co., Inc., Japan). The 0.1 mol/L nitric acid was also used as a blank sample.
Experimental Setup. Figure 4a shows the experimental setup used for the LEP-AES with the quartz and PDMS chips. The quartz and PDMS chip are set in chip holders, as shown in Figure 4b,c, respectively. The PFA tubes were connected to the chips as the inlet and outlet using a PDMS sheet as a sealing material. Pt wires 0.3 mm in diameter were inserted into the tube as electrodes. The liquid sample solution was introduced from the tube through the chip to waste using a syringe pump. For an electrical power source, a specifically designed pulsed dc voltage power supply (Nissei-Giken Co., Ltd., Japan) was used. The voltage conditions such as pulse width, height, and pulse sequence were controlled by laptop computer. A spectrometer (Shamrock SR-303i-A, Andor Technology plc., U.K.) and detector (Newton DU970N-UVB, Andor Technology plc., U.K.) were used to acquire the emission spectra of the LEP. The spectral data were transferred to the laptop computer for display and analysis. Before measurements, the inner tubes and flow channel of the chips were carefully washed using ultrapure water and 0.1 mol/L HNO3 aqueous solution. The emission lines that were used for measurement were 228.8022 nm for Cd and 405.7807 nm for Pb, and both of them are atomic line.56,57 The standard deviation (SD) and relative standard deviation (RSD) values of the averaged measurement data for each concentration of samples were determined . The limits of detection (LODs) for Cd and Pb were calculated using Currie’s definition method,58,59 3.29 σ/m, where σ is the SD for the blank solution measurement data and m is the slope of the calibration curve.
’ RESULTS AND DISCUSSION Figure 5 shows typical emission spectra for 0.1 mg/L Cd in a 0.1 mol/L HNO3 aqueous solution measured in a PDMS chip with and without sample flow (solid line and dotted line, respectively). The voltage was applied as a series of pulses; the pulse height, ontime, and off-time were 1500 V, 3 ms, and 2 ms, respectively. The emissions from the plasma for 100 pulses were accumulated to obtain one spectrum. The flow rate was 250 μL/min for the case with sample flow. At this flow rate, the average flow velocity at narrow channel and the distance of the sample motion in 2 ms of offtime are estimated as 463 mm/s and 926 μm, respectively. Therefore, this flow rate is high enough to push out the bubble from the narrow channel. The peaks for Cd are clearly observed at 228.8 nm on both spectra. The peak intensity for the case without sample flow is around 8 au and is less than one-tenth of the peak intensity (80 au) for the case with sample flow. This is considered to be caused by bubbles remaining in the narrow channel after each pulse. As mentioned before, the 9426
dx.doi.org/10.1021/ac2020646 |Anal. Chem. 2011, 83, 9424–9430
Analytical Chemistry
ARTICLE
Figure 7. Photographs of narrow channel in PDMS chip before (a) and after (b) 143 times measurements (totally 14 300 pulses of 3 ms on, 2 ms off, 1500 V) with flow. This condition is same as that for the case of PDMS chip 250 μL/min in Figure 6.
Figure 6. Stability of the Cd peak intensity during repeated measurements as a function of number of times (N) using PDMS or quartz chip. The flow rate of the sample solution (Cd 0.1 mg/L in 0.1 mol/L HNO3) was maintained at 0 or 250 μL/min. Other conditions follow: applied voltage, dc 1500 V; pulse width, 3 ms-on, 2 ms-off; accumulation, 100 pulses for each measurement. One plot represents one measurement. The measurements were repeated 143 times for the quartz chip with flow, 139 times for PDMS chip with flow, 143 times for quartz chip without flow, and 100 times for the PDMS chip without flow. The thick solid lines: 10 points moving average for each case.
plasma is generated in a bubble formed at the narrow channel by evaporation of water when the voltage pulse is applied. After the voltage was turned off between pulses, the evaporated water condenses to a liquid as the temperature decreased. However, it was frequently observed that some part of the bubble remained. This is considered to be caused by dissolved or decomposed nonwater gas in the sample solution. For the plasma generation on the next pulse, the remaining bubbles would either change the bubble/plasma-generation conditions and, therefore, the intensity, or disturb the plasma generation by breaking the electrical path in the solution. During the 100 pulse measurement for the spectra without sample flow in Figure 5, pulses without emission were frequently observed. In contrast, in the case with sample flow, emission occurred stably for every pulse, and the total intensity significantly improved. Under these conditions, the remaining bubbles were removed from the narrow channel before the next pulse by the sample flow. These facts suggest that the decrease of intensity in the case without sample flow is caused by the remaining bubbles and is significantly improved by sample flow. The broad peaks around 225.5 and 229.9 nm were observed in Figure 5. These emission spectra are not observed in the quartz chip, so that they are conducted to be molecules including C or Si. Figure 6 shows the stability of the Cd peak intensity during repeated measurements as a function of the number of measurement (N). Each plot represents an intensity derived from a spectrum with 100 pulses accumulated in the same manner as for the results shown in Figure 5. The pulse height, on/off time, flow rate, and sample solution conditions are also the same. One series of repeated measurements started using a fresh chip and was repeated over 100 times without chip exchange. Thus, over 10 000 pulses were applied on one chip. The plots indicated by crosses, open triangles, open squares, and open circles represent the intensity in the case of the PDMS chip without flow, the PDMS chip with flow, the quartz chip without flow, and the quartz chip with flow, respectively. The thick solid lines show the 10 point moving average for each case. In the case of the PDMS chip and quartz chip without flow, only a few plots have detectable but weak intensity at this concentration, as
shown in Figure 5, and for most of the numbers of times, the intensities are almost zero. This was due to the bubbles that remained in the narrow channel, as mentioned above. In the case of the PDMS chip with flow, the intensity maintained a relatively high level of 80 100 au for the number of times from 1 to about 35. Therefore, the advantage of sample flow is also confirmed by these plots. However, the intensity still fluctuated. The average intensity, SD, and RSD values calculated from the data are 82 au, 20 au, and 0.24, respectively. After about 35 times, the intensity decreases suddenly and becomes more unstable. The average intensity, SD, and RSD values calculated using all the data for the PDMS with flow are 41 au, 32 au, and 0.79, respectively. This degradation is considered due to the damage to the chip by the plasma because PDMS is a resin and is easily melted and decomposed by the heat of the plasma. Actually, widening and deformation of the narrow channel was observed in the used PDMS chip. Herein, Figure 7 shows photographs of the narrow channel in PDMS chip before and after 143 measurements (totally 14 300 pulses) in the same condition for Figure 6. The width of narrow channel after the measurements is apparently larger than that before the measurement by approximately 1.5 times. This is due to plasma damage. The widening of the narrow channel is considered to cause two opposite effects on the intensity. One is the increase of the plasma volume, and the other is the decrease of excitation temperature. The increase of plasma volume results in increase of intensity. The excitation temperature is one of the important factors for atomic emission intensity, and can be measured by Boltzmann plots.60 The excitation temperature of LEP was reported by Kumai et al.54 in the PDMS chip to be 6000 8600 K and decreased with increase of channel cross section. In Figure 6, the intensity by PDMS chip 250 μL/min seems to increase at N = 1 30 and to decrease at N > 30 with large fluctuation. The emission intensity changes approximately linearly to plasma volume and exponentially to excitation temperature. Therefore, as the widening of narrow channel, it is considered that the initial increase of intensity is caused by the increase of plasma volume and subsequent decrease is caused by decrease of the excitation temperature. In Figure 6, the same measurements were performed using the chip made of quartz, which is more heat resistant than PDMS. As expected, in the measurements using the quartz chip with flow, a constant Cd emission intensity around 32 au was clearly obtained at N = 1 143. The pulse, flow rate, and solution conditions were the same as in the experiment using the PDMS chip. The average intensity, SD, and RSD values were calculated to be 32 au, 6 au, and 0.18, respectively. The intensity stability improves much by using quartz chip compared to PDMS chip. This is considered due to higher heat resistance of quartz than that of PDMS. Actually, no severe damage was observed in the flow channel of 9427
dx.doi.org/10.1021/ac2020646 |Anal. Chem. 2011, 83, 9424–9430
Analytical Chemistry
Figure 8. Cd emission spectra by the LEP-AES using the quartz chip. Sample flow rate: 250 μL/min. Cd concentration region 0.001 to 1 mg/L (0.001, 0.005, 0.01, 0.05, 0.1, 0.5, and 1 mg/L in 0.1 mol/L HNO3). Blank solution is 0.1 mol/L HNO3 (with Cd 0 mg/L). Applied voltage: dc 1500 V. Pulse width: 3 ms-on, 2 ms-off. Accumulation for each concentration: 3000 pulses.
Figure 9. Cd calibration curve using the quartz chip. Each point is the average from three measurements (n = 3). Each measurement was done with a 3000 pulse accumulation. Error bars are defined as (SD.
the used quartz chip. Therefore, the quartz chip allows longer time accumulation than the PDMS chip. The RSD value of 0.18 for the measurements using the quartz chip with flow is considered to be too large for most applications. This value is derived from the accumulation of 100 pulses. However, by increasing the accumulation, the RSD value improves. For example, the RSD value for the accumulation of 1000 pulses was calculated to be 0.07 au From these results, accumulation was determined to be effective for improving the accuracy and limits of detection when using the quartz chip with flow because the intensities of each pulse fluctuate significantly but the average stayed constant. Interestingly, the emission intensity using the PDMS chip with flow at N = 1 35 times is approximately 2 times greater than the emission intensity of the quartz chip. This seems due to the PDMS chip being a soft material while the quartz chip is a hard material. During plasma emission, the actual width at narrow channel of the PDMS chip should become wider by high pressure due to superheating of water. This small widening can be a reason for high intensity of PDMS chip, by above discussion. Because the quartz chip with flow results in high stability of intensity, we tried to increase accumulation using this condition. Figure 8 shows the Cd emission spectra with 3000 pulse
ARTICLE
Figure 10. Cd calibration curve using the PDMS chip. Each point is the average from three measurements (n = 3). Each measurement was done with a 5000 pulse accumulation. Error bars are defined as (SD.
accumulation using the quartz chip with flow, for Cd concentrations of 0, 0.001, 0.005, 0.01, 0.05, 0.1, 0.5, and 1 mg/L in a 0.1 mol/L HNO3 aqueous solution. The spectra for the concentration range from 0.001 to 0.01 mg/L are shown in the figure inset. The voltage height, on/off time, and flow rate conditions were the same as those for the measurements shown in Figure 5. Background subtraction was performed before plotting using a background spectrum obtained from a blank 0.1 mol/L HNO3 aqueous solution. The emission line for Cd at 228.8 nm can be clearly observed even in the spectra for the 0.001 mg/L Cd solution. Figure 9 shows the calibration curves for the concentration range from 0.001 to 1 mg/L, derived from the spectra obtained using the same conditions as for Figure 8 using a quartz chip with sample flow. The calibration curve for the concentration range from 0.001 to 0.01 mg/L is shown in the figure inset. One measurement was done with 3000 pulse accumulation and background subtraction. The measurement was repeated 3 times (n = 3) at each concentration, and the average intensity, the SD, denoted by the error bars, and RSD values were calculated. The lines in the figure show the linear fittings of the data. The coefficients of correlation (R2) of the linear fitting for the range from 0 to 0.01 mg/L and from 0 to 1 mg/L were calculated to be 0.9992 and 0.9998, respectively. The SD for the blank solution (0.1 mol/L HNO3) was calculated to be 0.56 au. The slope of the linear fitting for the range from 0.001 to 0.01 mg/L was 3538 au. From these values, the LOD for Cd using quartz chip was determined to be 0.52 μg/L. Figure 10 shows the calibration curves for Cd obtained using a PDMS chip with flow. One measurement was done with 3000 pulse accumulation and background subtraction. The measurement was repeated 3 times (n = 3) at each concentration, and the average intensity, the SD, denoted by the error bars, and RSD values were calculated. The pulse height, on/off time, and flow rate were the same as those for Figure 5. Here, the coefficients of correlation (R2) for the concentration range from 0 to 0.01 mg/L and from 0 to 1 mg/L were 0.9664 and 0.9997, respectively. The SD for the blank solution (0.1 mol/L HNO3) was calculated to be 3.62 au. The slope of the calibration curve of the Cd for the range from 0.001 to 0.01 mg/L is 6975.85 au. From these values, the LOD for Cd using PDMS chip was determined to be 17 μg/L. From these results, it was confirmed that the increased accumulation is very effective for improving the limits of detection in LEP-AES and that the quartz chip is more suitable than the PDMS chip because of its increased resistance to damage by the plasma. 9428
dx.doi.org/10.1021/ac2020646 |Anal. Chem. 2011, 83, 9424–9430
Analytical Chemistry
Figure 11. Pb calibration curve using the quartz chip. Each point is the average from three measurements (n = 3). Each measurement was done with a 3000 pulse accumulation. Error bars are defined as (SD.
Finally, we performed the Pb measurements using the same setup. Figure 11 shows the Pb calibration curve region from 0 to 1 mg/L in a 0.1 mol/L HNO3 aqueous solution. This measurement was carried out using a quartz chip with flow and 3000 pulse accumulation, which is the same as the conditions used for Figure 8. The coefficients of correlation (R2) for the range from 0 to 0.05 mg/L and from 0 to 1 mg/L were 0.9274 and 0.9992, respectively. The SD of the blank aqueous solution (0.1 mol/L HNO3) was 0.52 au. The slope of the calibration curve for Pb is 90.03 au, and the LOD for Pb was determined to be 19.02 μg/L.
’ CONCLUSIONS We developed quartz-made chip for LEP-AES with the sample flow system to obtain high sensitivity and accuracy. The effect of sample flow and chip material (quartz or PDMS) was investigated. It was found that the long accumulation using quartz chip with sample flow was very effective to improve LOD. This was because bubbles remaining after each plasma pulse were removed from the narrow channel by sample flow, resulting in highly reproducible plasma generation, to enable high accumulation effect. Limit of detection for Cd and Pb were estimated as 0.52 μg/L and 19.02 μg/L, respectively. Those values are 2 3 orders of magnitude improved from the conventional case using resin chip without sample flow. ’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected].
’ REFERENCES (1) Walsh, A.; Sullivan, J. V. Atomic Absorption Spectrophotometer. U.S. Patent 3,503,686, Mar 31, 1970. (2) Greenfield, S.; Jones, L.; Berry, C. T. Analyst 1964, 89, 713–720. (3) Houk, R. S.; Fassel, V. A.; Flesch, G. D.; Svec, H. J. Anal. Chem. 1980, 52, 2283–2289. (4) Nathanson, H. C.; Newell, W. E.; Wickstrom, R. A.; Davis, J. R., Jr. IEEE Trans. Electron Devices 1967, 4 (ED-14), 117–133. (5) Terry, S. C.; Angell, J. B.; Angell, J. B. IEEE Trans. Electron Devices 1979, 3 (ED-26), 1880–1886. (6) Manz, A.; Graber, N.; Widmer, H. M. Sens. Actuators 1990, B1, 244–248. (7) Whitesides, G. M. Nature 2006, 442, 368–373. (8) Janasek, D.; Franzke, J; Manz, A. Nature 2006, 442, 374–380.
ARTICLE
(9) Psaltis, D.; Quake, S. R.; Yang, C. Nature 2006, 442, 381–368. (10) Craighead, H. Nature 2006, 442, 368–393. (11) deMello, A. J. Nature 2006, 442, 394–402. (12) El-Ali, J.; Sorger, P. K.; Jensen, K. F. Nature 2006, 442, 403–411. (13) Yager, P; Edwards, T.; Fu, E.; Helton, K.; Nelson, K.; Tam, M. R.; Weig, B. H. Nature 2006, 442, 412–418. (14) Eijkel, J. C. T.; Stoeri, H.; Manz, A. Anal. Chem. 1999, 71, 2600–2606. (15) Eijkel, J. C. T.; Stoeri, H.; Manz, A. J. Anal. At. Spectrom. 2000, 15, 297–300. (16) Rahman, M. M.; Blades, M. W. J. Anal. At. Spectrom. 2000, 15, 1313–1319. (17) Marcus, R. K.; Davis, W. C. Anal. Chem. 2001, 73, 2903– 2910. (18) Davis, W. C.; Marcus, R. K. J. Anal. At. Spectrom. 2001, 16, 931–937. (19) Wilson, C. G.; Gianchandani, Y. B. IEEE Trans. Electron Devices 2002, 49, 2317–2322. (20) Davis, W. C.; Marcus, R. K. Spectrochim. Acta, Part B 2002, 57, 1473–1486. (21) Miclea, M.; Kunze, K.; Franzke, J.; Niemax, K. Spectrochim. Acta, Part B 2002, 57, 1585–1592. (22) Franzke, J.; Kunze, K.; Miclea, M.; Niemax, K. J. Anal. At. Spectrom. 2003, 18, 802–807. (23) Broekaert, J. A. C.; Siemens, A. V. Anal. Bioanal. Chem. 2004, 380, 185–189. (24) Yagov, V. V.; Getsina, M. L.; Zuev, B. K. J. Anal. Chem. 2004, 59, 1037–1041. (25) Jenkins, G.; Franzke, J.; Manz, A. Lab Chip 2005, 5, 711–718. (26) Hsu, D. D.; Graves, D. B. Plasma Chem Plasma Process. 2005, 25, 1–17. (27) Franzke, J.; Miclea, M. Appl. Spectrosc. 2006, 60, 80A–90A. (28) Venzie, J. L.; Marcus, R. K. Spectrochim. Acta, Part B 2006, 61, 715–721. (29) Karanassios, V.; Johnson, K.; Smith, A. T. Anal. Bioanal. Chem. 2007, 388, 1595–1604. (30) Miclea, M.; Franzke, J. Plasma Chem. Plasma Process. 2007, 27, 205–224. (31) Wu, J.; Yu, J; Li, J.; Wang, J.; Ying, Y. Spectrochim. Acta, Part B 2007, 62, 1269–1272. (32) Webb, M. R.; Andrade, F. J.; Hieftje, G. M. Anal. Chem. 2007, 79, 7807–7812. (33) Agiral, A.; Groenland, A. W.; Chinthaginjala, J. K.; Seshan, K.; Lefferts, L.; Gardeniers, J. G. E. H. J. Phys. D: Appl. Phys. 2008, 41, 194009. (34) Mitra, B.; Levey, B.; Fung, T.-C.; Gianchandani, Y. B. In Proceedings of Micro Electro Mechanical Systems (MEMS 2006); 19th International Conference, Istanbul, Turkey, January 22 26, 2006; pp 554 557. (35) Jo, K.-W.; Kim, M.-G.; Shin, S.-M.; Lee, J.-H. Appl. Phys. Lett. 2008, 92, 011503. (36) Broekaert, J. A. C. Nature 2008, 455, 1185–1186. (37) Lui, S. L.; Godwal, Y.; Taschuk, M. T.; Tsui, Y. Y.; Fedosejevs, R. Anal. Chem. 2008, 80, 1995–2000. (38) Staack, D; Fridman, A.; Gutsol, A.; Gogotsi, Y.; Friedman, Gary. Angew. Chem., Int. Ed. 2008, 47, 8020–8024. (39) Weagant, S; Karanassios, V. Anal. Bioanal. Chem. 2009, 395, 577–589. (40) Tombrink, S.; M€uller, S.; Heming, Ri.; Michels, A.; Lampen, P.; Franzke, J. Anal. Bioanal. Chem. 2010, 397, 2917–2922. (41) Olabanji, O. T.; Bradley, J. W. Surf. Coat. Technol. 2011, 250, S516–S519. (42) Micle, M.; Kunze, K.; Musa, G.; Franzke, J.; Niemax, U. K. Spectrochim. Acta, Part B 2001, 56, 37–43. (43) Webb, M. R.; Andrade, F. J.; Hieftje, G. M. Anal. Chem. 2007, 79, 7899–7905. (44) Harris, D. C. Quantitative Chemical Analysis, 7th ed.; W. H. Freeman & Co Ltd: New York, 2007. (45) Greenfield, S. J. Chem. Educ. 2000, 77, 584–591. 9429
dx.doi.org/10.1021/ac2020646 |Anal. Chem. 2011, 83, 9424–9430
Analytical Chemistry
ARTICLE
(46) Iiduka, A.; Morita, Y.; Tamiya, E.; Takamura, Y. In Proceedings of Micro Total Analysis Systems 2004; The Royal Society of Chemistry: Malm€o, Sweden, September 26 30, 2004; pp 423 425. (47) Matsumoto, H.; Iiduka, A.; Yamamoto, T.; Tamiya, E.; Takamura, Y. In Proceedings of Micro Total Analysis Systems 2005; Transducer Research Foundation, Inc: Boston, October 9 13, 2005; pp 427 429. (48) Kumagai, I.; Matsumoto, H.; Yamamoto, T.; Tamiya, E.; Takamura, Y. In Proceedings of Micro Total Analysis Systems 2006; Society for Chemistry and Micro-Nano Systems (CHEMNAS): Tokyo, Japan, November 5 9, 2006; pp 497 499. (49) Yamamoto, T.; Takamura, Y. Bunseki (in Japanese) 2009, 1, 32–36. (50) Banno, M.; Tamiya, E.; Takamura, Y. Anal. Chim. Acta 2009, 634, 153–157. (51) Kumai, M.; Nakayama, K.; Furusho, Y.; Yamamoto, T.; Takamura, Y. Bunseki Kagaku (in Japanese) 2009, 58, 561–567. (52) Kagaya, S.; Nakada, S.; Inoue, Y.; Kamichatani, W.; Yanai, H.; Saito, M.; Yamamoto, T.; Takamura, Y.; Tohda, K. Anal. Sci. 2010, 26, 515–518. (53) Nakayama, K.; Yamamoto, T.; Hata, N.; Taguchi, S.; Takamura, Y. Bunseki Kagaku (in Japanese) 2011, 60, 515–520. (54) Kumai, M.; Takamura, Y. Jpn. J. Appl. Phys. 2011, 50, 096001 1–096001 4. (55) Yamamoto, T.; Kurotani, I.; Yamashita, A.; Kawai, J.; Imai, S. Bunseki Kagaku (in Japanese) 2010, 59, 1125–1131. (56) National Institute of Standards and Technology (NIST) Atomic Spectra Database Home Page. http://physics.nist.gov/PhysRefData/ ASD/lines_form.html (accessed Oct. 25, 2011). (57) Phelps, F. M., III. M.I.T. Wavelength Tables Vol. 2: Wavelengths by Element; The MIT Press: Cambridge, MA, 1982. (58) Currie, L. A.; Svehla, G. Pure Appl. Chem. 1994, 66, 595–608. (59) Currie, L. A. Pure Appl. Chem. 1995, 67, 1699–1723. (60) Wang, J.; Li, B.; Zhang, B. G.; Huang, X.; Dong, J.; Li, H. Spectrosc. Lett. 1998, 31, 243–252.
9430
dx.doi.org/10.1021/ac2020646 |Anal. Chem. 2011, 83, 9424–9430
