Anal. Chem. 2004, 76, 2568-2574
Quantitative Mass Spectrometric Determination of Methylphenidate Concentration in Urine Using an Electrospray Ionization Source Integrated with a Polymer Microchip Yanou Yang,† Jun Kameoka,† Timothy Wachs,‡ Jack D. Henion,‡ and H. G. Craighead*,†
Applied and Engineering Physics, Cornell University, Ithaca, New York 14853, and Analytical Toxicology, Cornell University, 927 Warren Road, Ithaca, New York 14850
We have demonstrated the use of a simple microfabricated electrospray ionization source for coupling microfluidic chips to mass spectrometry (MS). A polymer-based microchip, coupled to a triple quadrupole mass spectrometer, has been employed for direct infusion quantitative bioanalysis of methylphenidate (Ritalin) extracted from human urine samples. The approach used a microfabricated polymer electrospray emitter to couple a microfluidic channel to a stable electrospray ionization source. The microchip was fabricated from cycloolefin plastic plate by hot embossing and thermal bonding. This microfluidic chip contained two independent microfluidic channels, integrated with two corresponding electrospray emitters and an internal gold electrode. Liquid-liquid extraction was used to prepare urine samples, spiked with methylphenidate. A trideuterated analogue of methylphenidate (methylphenidate-d3) was used as the internal standard for the analysis. The system showed good electrospray stability and reproducibility with different spray tips. Four different electrospray tips were used to analyze the same sample, and the results showed very small variation with a relative standard deviation of 1.4%. A standard curve prepared for methylphenidate in urine (R2 ) 0.999) was linear over the range of 0.4-800 ng/mL. The precision of the quality control samples for three different concentrations ranged from 19.1% at 20 ng/mL, 3.2% at 200 ng/mL, to 3.5% at 667 ng/mL while the accuracy was 96.3% at 20 ng/mL, 101.2% at 200 ng/mL, and 101.6% at 667 ng/mL. No system carryover was detected even when the same device was used for sequential analysis. These results suggest the potential of this microdevice for MS-based quantitative analysis in drug discovery and development. Recently developed microfabrication technology has created opportunities for micro total analysis systems1,2 Miniaturization * Corresponding author: (e-mail)
[email protected]; (phone) 607-255-8707; (fax) 607-255-7658. † Applied and Engineering Physics. ‡ Analytical Toxicology. (1) Reyes, D. R.; Iossifidis, D.; Auroux, P. A.; Manz, A. Anal. Chem. 2002, 74, 2623-2636.
2568 Analytical Chemistry, Vol. 76, No. 9, May 1, 2004
and the use of microfluidic systems are motivated by possibilities for improved performance, the use of small sample volumes, lower cost, and higher throughput.3 The integration of chip-based microfluidic devices to mass spectrometry (MS) promises to be a powerful addition to the micro total analysis systems. MS can provide both molecular weight and structural information, which cannot be provided by optical spectroscopic or laser-induced fluorescence detectors. The flow rates used in the microchips are similar to those required for electrospray ionization-mass spectrometry (ESI-MS).4 There is a need, therefore, for simple and effective methods of coupling microfluidic devices to mass spectrometry. Early examples of coupling microfabricated devices to electrospray mass spectrometry have been demonstrated by Karger’s and Ramsey’s groups in 1997.5,6 Efforts directed toward developing interfaces to microfluidic sample pretreatment continues, with work on microfluidic systems made of a range of substrate materials.7-20 One of the most important issues in microchip-ESI-MS is the creation and maintenance of a stable electrospray. The first examples of integrating a microchip to mass spectrometry were (2) Auroux, P. A.; Iossifidis, D.; Reyes, D. R.; Manz, A. Anal. Chem. 2002, 74, 2637-2652. (3) Voldman, J.; Gray, M. L.; Schmidt, M. A. Annu. Rev. Biomed. Eng. 1999, 1, 401-425. (4) Wilm, M.; Mann, M. Anal. Chem. 1996, 68, 1-8. (5) Xue, Q.; Foret, F.; Dunayevskiy, Y. M.; Zavracky, P. M.; McGruer, N. E.; Karger, B. L. Anal. Chem. 1997, 69, 426-430. (6) Ramsey, R. S.; Ramsey, J. M. Anal. Chem. 1997, 69, 1174-1178. (7) Figeys, D.; Ning, Y.; Aebersold, R. Anal. Chem. 1997, 69, 3153-3160. (8) Xu, N.; Lin, Y.; Hofstadler, S. A.; Matson, D.; Call, C. J.; Smith, R. D. Anal. Chem. 1998, 70, 3553-3556. (9) Moore, R. E.; Licklider, L.; Schumann, D.; Lee. T. D. Anal. Chem. 1998, 70, 4879-4884. (10) Zhang, B.; Lui, H.; Karger, B. L.; Foret, F. Anal. Chem. 1999, 71, 32583264. (11) Li, J.; Thibault, P.; Bings, N. H.; Skinner, C. D.; Wang, C.; Colyer, C.; Harrison J. Anal. Chem. 1999, 71, 3036-3045. (12) Licklider, L.; Wang, X.-Q.; Deasi, A.; Tai, Y.-C.; Lee, T. D. Anal. Chem. 2000, 72, 367-375. (13) Schultz, G. A.; Corso, T. N.; Prosser S. J.; Zhang, S. Anal. Chem. 2000, 72, 4058-4063. (14) Kim, J.-S.; Knapp, D. R. J. Am. Soc. Mass. Spectrom. 2001, 12, 463-469. (15) Yuan C.-H.; Shiea, J. Anal. Chem. 2001, 73, 1080-1083. (16) Meng, Z.; Qi, S.; Soper, S. A.; Limbach. P. A. Anal. Chem. 2001, 73, 12861291. (17) Wachs, T.; Henion, J. Anal. Chem. 2001, 73, 632-638 (18) Rohner, T. C.; Rossier, J. S.; Girault, H. H. Anal. Chem. 2001, 73, 53535357. 10.1021/ac0303618 CCC: $27.50
© 2004 American Chemical Society Published on Web 04/03/2004
achieved by spraying directly from the exposed channel at the edge of the glass microchip.5,6 Although attractive because of its simplicity, the liquid tends to spread out and form large liquid droplets at the chip surface. This is not compatible with on-chip separation due to excessive band broadening and sample dilution.11 Attempts to solve this problem included changing the surface property of the outlet edge by coating5 and derivatizing6 with hydrophobic reagents or assisting the droplet formation with a pneumatical nebulizer.10 An alternative approach was to spray from a short transfer capillary attached to the microchip.7 Some of these approaches led to increased dead volume at the outlet, which is an issue that continues to be addressed in ongoing research.11,16,21 Recent exploratory devices used fabricated electrospray emitters made of silicon,13 polycarbonate,22 PDMS,14 and parylene.12 One of the trends in the microchip-ESI-MS field is to utilize polymeric materials for microchip fabrication. The advantages of polymer-based microfabricated devices include reduced cost, simplified manufacturing procedures, availability of a wide range of plastics with different properties, and disposability. Solvent resistance of the polymeric substrate is one of the critical issues for microchip-ESI-MS application. If the polymer is not stable or soluble in the organic ESI solvents, the device can swell, distorting the microchannels, or leached materials can cause chemical noise. Solvent-resistant cycloolefin polymer substrates, which are widely used in the manufacture of compact disks and digital video disks, were recently used to make microchips for coupling with ESIMS.23 One of the fields that could benefit from microchip-ESI-MS developments is pharmaceutical drug discovery and development, which requires high-throughput qualitative and quantitative determination of small-molecule drugs with high sensitivity and selectivity.20,24 Due to the combination of high sensitivity and selectivity, liquid chromatography, coupled to tandem mass spectrometry (LC/MS/MS), can achieve significant improvement in quantitative detection limits, compared to traditional methods using UV or fluorescence detection.25 Thus, LC/MS/MS has become the preferred technique for quantitative determination of drugs and metabolites in biomatrixes. Although direct bioanalysis is likely to suffer from ion suppression26 or interference27 from the matrix components or drug metabolites, direct MS analysis eliminates the time required to run the LC separation, and there are examples of successful bioanalyses, done without the LC separation.28-30 A microfluidic format for direct bioanalysis by infusion through an ESI chip was recently described for quanti(19) Chen, S.-H.; Sung, W.-C.; Lee, G.-B.; Lin, Z.-Y.; Chen, P.-W.; Liao, P.-C. Electrophoresis 2001, 22, 3972-3977. (20) Jiang, Y.; Wang, P.-C.; Locascio, L. E.; Lee, C. S. Anal. Chem. 2001, 73, 2048-2053. (21) Pinto, D. M.; Ning, Y.; Figeys, D. Electrophoresis 2000, 21, 181-190. (22) Tang, K.; Lin, Y.; Matson, D. W.; Kim, T.; Smith, R. D. Anal. Chem. 2001, 73, 1658-1663. (23) Kameoka, J.; Craighead, H. G.; Zhang, H.; Henion, J. Anal. Chem. 2001, 73, 1935-1941. (24) Lee, M. S.; Kerns, E. H. Mass Spectrom. Rev. 1999, 18, 287-279. (25) Allanson, J. P.; Biddlecombe, R. A.; Jones, A. E.; Pleasance, S. Rapid Commun. Mass Spectrom. 1996, 10, 811-816. (26) King, R.; Bonfiglio, R.; Fernandez-Metzler, C.; Miller-Stein, C.; Olah, T. J. Am. Soc. Mass Spectrom. 2000, 11, 942-950. (27) Jemal, M.; Xia, Y.-Q. Rapid Commun. Mass Spectrom. 1999, 13, 97-106. (28) Chen, S.; Carvey, P. M. Rapid Commun. Mass Spectrom. 2001, 15, 159163.
tation of verapamil and norverapamil with the limit of quantitation in the low-nanogram per milliliter range.31 Although several different formats of an on-chip electrospray ionization source for interfacing microchips to mass spectrometry have been introduced, few have been investigated for application in quantitative determination of analytes in complex samples. Recently, a polymer-based microchip integrated with a triangular parylene emitter at the terminus of the channel has been coupled to a time-of-flight mass spectrometer.32 The planar processing of this device and the ease of coupling to polymer microfluidics makes this an attractive candidate for microchip-ESI-MS. In this report, we further investigate the utilization and performance of this type of microfluidic-coupled electrospray device for quantitative analysis. We have used a slightly modified version of the previously reported microchip, now including an integrated gold electrode for applying the electrospray voltage. The electrode was fabricated with standard photolithography and liftoff techniques, before thermal bonding of the polymer plates. This microchip was coupled to an API-III plus triple quadrupole mass spectrometer, and stable electrospray was achieved. Fortified urine extracts, containing a representative small-molecule drug and a stable isotope-labeled analogue, were infused through the electrospray tip via the microfluidic channel for subsequent selected reaction monitoring (SRM) quantitative analysis without prior liquid chromatography separation. The quantitative results obtained with this strategy for methylphenidate (Ritalin) extracted from fortified human urine sample are presented. EXPERIMENTAL SECTION Chemicals and Materials. Methylphenidate hydrochloride and sodium tetraborate were purchased from Sigma (St. Louis, MO). A 1 mg/mL solution of methylphenidate-d3 in acetonitrile was purchased from Isotec (Miamisburg, NJ). Methanol was from VWR Scientific (West Chester, PA). Cyclohexane and formic acid were obtained from Aldrich Chemical Co. (Milwaukee, WI). Cycloolefin plastic plates (ZEONOR1020R, 2 × 100 × 150 mm) were obtained from Zeon Chemicals (Louisville KY). Parylene C dimer was purchased from Specialty Coating Systems (Indianapolis, IN). Water was purified in-house with a Nanopure water system (Barnstead Thermolyne, Dubuque, IA). A stock solution of methylphenidate (1 mg/mL) was prepared in methanol and stored in a refrigerator. Working solutions for all samples were prepared in 75% methanol, 25% water, and 0.1% formic acid. Device Fabrication. The microchip, containing two independent microfluidic channels and parylene electrospray tips, was fabricated from a cycloolefin polymer. The microchannel was fabricated by a hot embossing technique as described previously.23 As can be seen in the schematic drawing of the microchip in Figure 2, the two channels on the microchip are closer at the sprayer end with a 1-mm separation, while they are separated by 2 cm at the inlet end. The purpose of this design is to make alignment of sprayers to the channel exit easier while the nanoport (Upchurch Scientific, Oak Harbor, WA) connectors can still be (29) Zheng, J. J.; Lynch, E. D.; Unger, S. E. J. Pharm. Biomed. Anal. 2002, 28, 279-285. (30) Wachs, T.; Henion, J. D. Anal. Chem. 2003, 75, 1769-1775. (31) Dethy, J.-M.; Ackermann, B. L.; Delatour, C.; Henion, J. D.; Schultz, G. A. Anal. Chem. 2003, 75, 805-811. (32) Kameoka, J.; Orth, R.; Ilic, B.; Czaplewski, D.; Wachs, T.; Craighead, H. G. Anal. Chem. 2002, 74, 5897-5901.
Analytical Chemistry, Vol. 76, No. 9, May 1, 2004
2569
Figure 1. Device fabrication. (A) Gold electrode fabrication: (1) photoresist spinning; (2) photolithography; (3) gold evaporation; (4) lift-off. a, photoresist; b, cycloolefin polymer plate; c, gold film. (B) Fabrication process of the device with two electrospray tips by thermal bonding: a, top plastic chip with embossed fluidic channels; b, reservoirs; c, microfluidic channel; d, parylene film with two electrospray tips; e, bottom plastic chip with fabricated electrode; f, gold electrode; g, hole to fit a metal screw.
Figure 2. Configuration of microchip interfacing to API-III plus triple quadrupole mass spectrometer (not drawn to scale): a, microchip; b, syringe pump; c, power supply; d, X, Y, Z stage.
accommodated at the input. The total length of the channel is 2.5 cm and the depth is 20-25 µm. The microfabricated approach makes the design flexible and the creation of arrays of devices possible with no additional processing steps. Hand drilling was used to make the reservoir holes after embossing, and a wider channel was formed at the inlet end. However, a smaller channel is desired at the tip end to provide a stable spray. The final design was established with the width of the channel at the sprayer end 20 µm while it was 60 µm wide at the reservoir end. A gold electrode with a width of 500 µm was fabricated onto one plastic chip by standard photolithography and liftoff techniques (Figure 1A). Gold film with a thickness of 80 nm was deposited onto the surface by thermal evaporation. A hole was drilled through the other chip surface, which contained the microfluidic channels. The parylene electrospray tips were fabricated by photolithography and reactive ion etching as described before.32 The parylene film thickness was 3-4 µm, and the width of the source was ∼100 µm at the base of the triangle with a 90° angle at the tip. The plate with the gold electrode and the plate with microfluidic channels were bound together with a parylene film containing two electrospray emitters sandwiched between them (Figure 1B). A copper screw was glued into the hole to make contact with the electrode with silver paint. Nanoports were attached to the 2570 Analytical Chemistry, Vol. 76, No. 9, May 1, 2004
reservoir holes with TorrSeal adhesive and cured at 60 °C for 12 h. Figure 3A shows an optical micrograph of the electrosprayer tips. Figure 3B shows the entire fabricated device. The finished microchip dimension, containing two independent channels, was 2 cm × 3 cm × 4 mm. Sample Preparation Procedure. Methylphenidate and methylphenidate-d3 internal standard were fortified in human urine samples and extracted by liquid-liquid extraction.33 A 0.5-mL sample of saturated sodium tetraborate (pH ) 9.3) and 800 or 300 ng of internal standard solution were added to 2-mL urine aliquots spiked with different concentrations of methylphenidate. The methylphenidate was extracted into 2.5 mL of cyclohexane by mixing on a Rotorack at the maximum speed of 40 rpm for 10 min followed by centrifugation at 300g for 10 min. Two milliliters of the cyclohexane layer was transferred to a small glass tube, evaporated to dryness under N2 at room temperature on a Pierce Reacti-Therm III heating/stirring module (Rockford IL), and reconstituted in 1 mL or 200 µL of working solution containing 75% methanol, 25% water, and 0.1% formic acid. For studies of interchip reproducibility, urine samples were spiked with 200ng/ mL methylphenidate, 20 µL of methylphenidate-d3 solution con(33) Bach, G. A.; Henion, J. J. Chromatogr., B 1998, 707, 175-285.
Figure 3. Fabricated devices: (A) array of two parylene electrospray tips with a distance of 1 mm; (B) picture of the microchip for interfacing to mass spectrometer (3 cm × 2 cm, 4 mm thick).
taining 40 µg/mL methylphenidate-d3 was used, and the residue extract material was reconstituted in 1 mL of working solution after blowing to dryness. For the calibration curve preparation, the calibration standards were prepared in duplicate at six different concentrations (0.4, 1.6, 40, 200, 400, 800 ng/mL) plus a blank, which contained only methylphenidate-d3. Two different sets of QC samples were prepared at three different concentrations, respectively. The first set has three different concentrations (10, 300, and 500 ng/mL) with two duplicates. The concentrations of the second set of QC samples were 20, 200, and 667 ng/mL with five or six duplicates, which were used to calculate precision and accuracy. Separate methylphenidate stock solutions were used in preparing calibration standards and QC samples. Thirty microliters of a 10 µg/mL internal standard solution was used in all calibration standards, and QC samples and the extracted materials were reconstituted in 200 µL of working solution. Interfacing the Microchip to Mass Spectrometry. The experimental configuration for interfacing the microchip device to an API-III plus triple quadrupole mass spectrometer (PE Sciex, Concord, ON, Canada) is shown in Figure 2. The microchip was mounted on an X, Y, Z translation stage (Newport, Irvine, CA) to adjust the position of the individual sprayers relative to the ion sampling orifice of the mass spectrometer for maximum MS signal. The microchip was positioned ∼8 mm away from the orifice. Each chip contained two channels integrated with their respective sprayers, but only one sprayer was used at a time. A syringe infusion pump (Harvard Apparatus, Holliston, MA) was used to generate a desired flow rate of the samples. The pump was connected to the nanoports through silica capillary tubing (360-µm o.d. and 50-µm i.d.). A voltage of ∼2 kV was applied between the MS orifice and gold microelectrode, using an external power supply (Bertan, Franklin Park, IL). The liquid on the tip was observed by a mirror and video camera. Once electrospray was established by applying the voltage, the flow rate was finetuned to balance the rate at which the sprayer removes the solution. A flow rate of 0.1-0.2 µL/min and a voltage of ∼2 kV were satisfactory for most sprayers with the used working solution. Mass Spectrometry. Mass spectrometer tuning was accomplished by infusing a solution of the analytes through an ion spray originally coupled to the API-III plus mass spectrometer. Both mass-analyzing quadrupoles (Q1 and Q3) were adjusted to unit mass resolution and operated in the positive ion mode. The orifice was maintained at a potential of 53V for methylphenidate
and methylphenidate-d3 to provide optimal ion transmission. The interface plate was maintained at 450 V. The flow of the N2 curtain gas was set at 120 mL/min. For the SRM mode, the dwell time was 200 ms. The collision energy used in the SRM mode was 24.5 eV. The SRM transitions monitored in the quantitative studies corresponded to the protonated molecules for the methylphenidate (m/z 234.2) and methylphenidate-d3 (m/z 237.2), and the major product ion observed in the collision-induced dissociation mass spectra for both was m/z 84.1. Q1 was set to monitor m/z 234.2 and 237.2 for methylphenidate-d0 and methylphenidate-d3, respectively, while Q3 monitored m/z 84.1. A dummy ion (m/z 250 > 84.1) was used to eliminate cross-talk in the collision cell since one common fragment ion was chosen here for both analyte and internal standard. Data were collected in individual files for each sample. The files were exported into test files and analyzed with Microsoft Excel software. Ion current was averaged over the appropriate region of the total ion current trace for ∼15-30 s, and the ratio of analyte to internal standard was calculated. The internal standard is necessary for quantitation here since there is no discrete chromatographic peak as in LC/MS. Safety Considerations. Because of high voltage used in the experiments, shielding and proper insulation should be used along with normal precautions to ensure that the high-voltage supply is sufficiently current limited so as not to be lethal in case of accidental contact. Because potentially toxic or flammable materials were used or sprayed in this study, proper ventilation and vapor removal should be provided. Normal safe practices for proper handling of laboratory chemicals and biological samples should always be followed. RESULTS AND DISCUSSION Electrospray Stability. The electrospray stability of this device was studied with an extracted urine sample containing only methylphenidate-d3, which corresponds to the blank calibration standards. An API-III plus triple quadrupole mass spectrometer was used for the MS analysis as described above. Figure 4 shows the total ion current of the SRM data (m/z 237.2 > 84.1, 234.2 > 84.1) from the blank sample, which was stable with a relative standard deviation of 3.05% over a period of 5 min. A data acquisition time of 15-30 s with stable electrospray is more than enough for quantitation. Interchip Study. We tested the reproducibility of the device performance by comparing the MS data of several devices with Analytical Chemistry, Vol. 76, No. 9, May 1, 2004
2571
Figure 4. Selected reaction monitoring (m/z 237.2 > 84.1) ion current for extracted urine sample containing 150 ng/mL methylphenidate-d3.
Figure 5. SRM ion current of the same extracted urine sample containing methylphenidate and methylphenidate-d3 using four different electrospray tips: (A) tip 1 from chip 1, I/I0 ) 0.470, and acquisition time 30 s; (B) tip 1 from chip 2, I/I0 ) 0.463, and acquisition time 30 s; (C) tip 2 from chip 2, I/I0 ) 0.460, and acquisition time of 25 s; (D) tip 1 from chip 3, I/I0 ) 0.457, and acquisition time 20 s.
Table 1. Intersprayer Study for Methylphenidate Fortified in Urine Sample tip 1-1 I/I0 I/I0 average RSD (%)
0.470
0.470 X h ) 0.470 0
tip 2-1 0.470
tip 2-2 I/I0 I/I0 average RSD (%) stats of 4 electrosprayer tips
0.460 0.460 X h ) 0.460 0
0.468
0.466 X h ) 0.466 0.3
0.463
tip 3-1 0.460 0.457 0.455 0.454 X h ) 0.455 0.3
X h ) 0.463 RSD ) 1.4%
the same sample. Four different electrosprayers from three different microchips were used to obtain the SRM data of the same extracted urine sample containing methylphenidate and methylphenidate-d3. Three measurements were made on each of the four electrospray tips. Between each measurement, the chip was removed from the X, Y, Z stage and washed with methanol for 5 min at a flow rate of 5 µL/min generated via a second syringe pump. The results are summarized in Table 1. For the same 2572
Analytical Chemistry, Vol. 76, No. 9, May 1, 2004
electrospray tip, different measurements gave the same intensity ratio of averaged SRM ion current of analyte and internal standards over a period of 15-30 s (RSD e 0.3%). The variation between different electrospray tips was also very small (RSD ) 1.4%). The SRM data of one measurement from each tip are shown in Figure 5. These results indicate that the quantitative data from one sample is device-independent if an internal standard is used, which is promising for future development of disposable devices in quantitative bioanalysis. Array-based devices are desirable for high sample throughput applications. While we have demonstrated the use of a two-channel system with independent control, a similar microfabrication approach could be used for larger numbers of channels of other microfluidic designs. Linearity Study. The relationship between the SRM intensity ratio of the analyte to the internal standard for methylphenidate concentrations in a range of 0.4-800 ng/mL was studied. To determine the dynamic range of methylphenidate using the microchip with on-chip electrosprayers, two duplicates of each calibration standard were prepared at six different concentrations (0.4, 1.6, 40, 200, 400, and 800 ng/mL) of methylphenidate spiked in urine plus two blank samples containing only methylphenidated3. Two duplicates of each QC sample at three concentrations (10, 300, and 500 ng/mL) were also prepared from a different stock
Figure 6. SRM ion current for calibration standards and QC sample in urine: (A) 1.6 ng/mL methylphenidate, I/I0 ) 0.0447, acquisition time 30 s; (B) 10 ng/mL methylphenidate, I/I0 ) 0.11, and acquisition time 30 s; (C) 200 ng/mL methylphenidate, I/I0 ) 1.51, and acquisition time 34 s; (D) 800 ng/mL methylphenidate, I/I0 ) 5.50, and acquisition time 18 s.
Figure 7. SRM ion current of the blank sample for methylphenidate analyzed with one electrospray tip indicating the signal obtained before and after running 800 ng/mL calibration standards: (A) before running the highest level of calibration standard, I/I0 ) 0.0312, and acquisition time 30 s; (B) after running the highest level of calibration standard, I/I0 ) 0.0314, and acquisition time 30 s.
solution to check the calibration curve. The prepared calibration standards and QCs were infused to the microfluidic channel and electrosprayed sequentially into the mass spectrometer. The device was washed between each sample to mitigate crosscontamination. The intensity ratio of the averaged SRM ion current of analyte and internal standards over a period of 15-30 s for each calibration standard and QC was used to obtain the calibration curve. The calibration curve was linear over the range tested (0.4-800 ng/mL). The equation and correlation coefficient for methylphenidate were y ) 0.0069x + 0.0439 and r2 ) 0.9993, respectively. Figure 6 shows the representative SRM ion current traces for three calibration standards with different concentrations (1.6, 200, and 800 ng/mL) and the low QC (10 ng/mL). The measured intensity ratio was linear over a methylphenidate concentration dynamic range of 2000. The good reproducibility of the quantitative data from the two duplicates of the calibration standards with the lowest concentration (0.4 ng/mL) indicates a subnanogram per milliliter lower limit of quantitation (LLOQ). If one considers the age of the API-III plus used here, a lower LLOQ may be expected with a newer mass spectrometer with a higher sensitivity.
Table 2. Precision and Accuracy Summary Table for Quantitative Determination of Methylphenidate in Urine Sample QC concn (ng/mL)
calc concn (ng/mL)
20
25.4 18.3 17.2 15.9 19.5 206.8 208.5 207.6 201.7 193.7 195.6 665.9 699.3 699.8 685.6 651.6 663.0
200
667
statistics 19.3 X h 19.1% RSD 96.3% accuracy 202.3 X h 3.2% RSD 101.2% accuracy
677.5 X h 3.5% RSD 101.6% accuracy
A total of five electrospray tips were used for preparing the calibration curves. Both the fact that different electrospray tips Analytical Chemistry, Vol. 76, No. 9, May 1, 2004
2573
can be used for preparing one calibration curve and the results from the interchip assay indicate the potential for reliable use of microfabricated systems for quantitative assay when a suitable internal standard is used. Precision, Accuracy, and Carryover. The precision and accuracy of the method were determined with another three sets of quality control samples with five to six duplicates, prepared and analyzed in the same manner as the calibration standards. The intensity ratio values are calculated from the standard curve described above. Table 2 summarizes the accuracy and precision data of this method for methylphenidate. The precision ranged from 19.1 to 3.2%, and the accuracy ranged from 96.3 to 101.6% for the three concentrations, 20, 200, and 667 ng/mL. To test the system carryover, a blank sample was analyzed with the same tip after the highest concentration of calibration standard was run followed by washing the microchannel and electrospray tip with methanol for 5 min off-line at a flow rate of 5 µL/min. As shown in Figure 7, the carryover was negligible. CONCLUSIONS Quantitative MS analysis of small drug molecules fortified in human urine, with a lower limit of quantitation at the subnanogram
2574
Analytical Chemistry, Vol. 76, No. 9, May 1, 2004
per milliliter level, has been demonstrated using a multichannel polymer-based microchip with integrated electrospray sources. The microfabricated system has a stable electrospray performance, and different channel/sprayer systems performed reproducibly. The linear dynamic range of methylphenidate (0.4-800 ng/mL) fortified in urine sample indicated a lower limit of quantitation employing infusion bioanalysis, which precludes the need for liquid chromatography. The approach we have shown for the electrospray source coupling could be used with other on-chip processing. We envision these plastic systems, when mass produced, to be disposable. ACKNOWLEDGMENT This work was supported by the Nanobiotechnology Center (NBTC), an STC Program of the National Science Foundation under Agreement ECS-9876771. The authors appreciate access and use of the Cornell Nanofabrication Facility.
Received for review October 14, 2003. Accepted January 30, 2004. AC0303618
