Anal. Chem. 2001, 73, 332-336
Rotating Electrode Potentiometry: Lowering the Detection Limits of Nonequilibrium Polyion-Sensitive Membrane Electrodes Qingshan Ye and Mark E. Meyerhoff*
Department of Chemistry, The University of Michigan, Ann Arbor, Michigan 48109-1055
A rotating electrode configuration is evaluated as a means to lower the detection limits of newly devised polyionsensitive membrane electrodes (PSEs). Planar potentiometric polycation and polyanion PSEs are prepared by incorporating tridodecylmethylammonium chloride and calcium dinonylnaphthalenesulfonate, respectively, into plasticized PVC or polyurethane membranes and mounting disks of such films on an electrode body housed in a conventional rotating disk electrode apparatus. Rotation of the PSEs at 5000 rpm results in an enhancement in the detection limits toward heparin (polyanion) and protamine (polycation) of at least 1 order of magnitude (to 0.01 unit/mL for heparin; 0.02 µg/mL for protamine) over that observed when the EMF responses of the same electrodes are assessed using a stir-bar to achieve convective mass transport. A linear relationship between ω-1/2, where ω is the rotating angular frequency, and C1/2, the polyion concentration corresponding to half the total maximum ∆EMF response toward the polyion species, is observed. It is further shown that the rotating polycation sensor can be used as an end-point detector to greatly enhance (relative to nonrotated indicator electrode) the analytical resolution and precision for measurement of low concentrations of heparin when such samples are titrated with protamine. The theoretical basis for lowering the detection limits by rotating PSEs is discussed based on the unique nonequilibrium response mechanism of such sensors. Recently, it has been discovered that specially formulated polymer membranes doped with appropriate lipophilic ion exchangers yield large and reproducible potentiometric responses toward various biomedically important polyanions (e.g., heparin, DNA, and polyphosphates) and polycations (e.g., protamine, polyarginine, etc.) at microgram per milliliter levels in the presence of physiological concentrations of common inorganic ions.1-13 The EMF response of these so-called polyion-sensitive * Corresponding author: (phone) (734)-763-5916; (fax) (734)-647-4865; (email)
[email protected]. (1) Baliga, N.; Yang, V. C.; Meyerhoff, M. E. Clin. Chem. 1998, 44, A52-A53. (2) Dror, M.; Baugh, R. F.; Armstrong, A.; Yang, V. C.; Meyerhoff, M. E. Thromb. Haemostasis 1997, (Suppl.), 1156. (3) Esson, J. M.; Meyerhoff, M. E. Electroanalysis 1997, 9, 1325-1330. (4) Fu, B.; Yun, J. H.; Wahr, J.; Meyerhoff, M. E.; Yang, V. C. Adv. Drug. Delivery Rev. 1996, 21, 215-223. (5) Ma, S. C.; Yang, V. C.; Meyerhoff, M. E. Anal. Chem. 1992, 64, 694-697.
332 Analytical Chemistry, Vol. 73, No. 2, January 15, 2001
membrane electrodes (PSEs) has been ascribed to the establishment of a nonequilibrium, steady-state ion-exchange process that occurs at the membrane/sample interface. This exchange occurs because of the very favorable extraction of the analyte polyion into the organic membrane phase by cooperative ion pairing with the lipophilic ion-exchange species.14, 15 In addition to unraveling the fundamental response mechanism of PSEs,16 many useful bioanalytical applications of these devices have already been demonstrated.17 For example, accurate determinations of levels of the anticoagulant heparin in undiluted whole blood can be achieved via a simple potentiometric titration using protamine as the titrant and a polycation PSE as the end point detector.9 Both polycation and polyanion PSEs have also been shown to be useful as detectors for the determination of certain enzyme activities that cleave larger polyionic substrate molecules into smaller fragments of lower charge and molecular weight.3,18 Such enzyme analysis applications of PSEs rely on the fact that these devices exhibit much less EMF response to lower molecular weight polyions, owing to a significant decrease in the strength of cooperative ion pairing between the low molecular weight polyion and the lipophilic ion exchanger within the membrane phase. Very recently, PSEs have also been applied as detectors in the development of a novel, nonseparation, competitive binding (6) Ma, S. C.; Yang, V. C.; Fu, B.; Meyerhoff, M. E. Anal. Chem. 1993, 65, 2078-2084. (7) Meyerhoff, M. E.; Fu, B.; Bakker, E.; Yun, J. H.; Yang, V. C. Anal. Chem. 1996, 68, A168-A175. (8) Meyerhoff, M. E.; Yang, V. C.; Wahr, J. A.; Lee, L. M.; Yun, J. H.; Fu, B.; Bakker, E. Clin. Chem. 1995, 41, 1355. (9) Ramamurthy, N.; Baliga, N.; Wahr, J. A.; Schaller, U.; Yang, V. C.; Meyerhoff, M. E. Clin. Chem. 1998, 44, 606-613. (10) Ramamurthy, N.; Baliga, N.; Wakefield, T. W.; Andrews, P. C.; Yang, V. C.; Meyerhoff, M. E. Anal. Biochem. 1999, 266, 116-124. (11) Wahr, J. A.; Yun, J. H.; Yang, V. C.; Lee, L. M.; Meyerhoff, M. E. J. Cardiothoracic Vasc. Anesth. 1996, 10, 447-450. (12) Yun, J. H.; Ma, S. C.; Fu, B.; Yang, V. C.; Meyerhoff, M. E. Electroanalysis 1993, 5, 719-724. (13) Yun, J. H.; Han, I. S.; Chang, L. C.; Ramamurthy, N.; Meyerhoff, M. E.; Yang, V. C. Pharm. Sci. Technol. 1999, 2, 102-110. (14) Esson, J. M.; Ramamurthy, N.; Meyerhoff, M. E. Anal. Chim. Acta 2000, 404, 83-94. (15) Fu, B.; Bakker, E.; Yang, V. C.; Meyerhoff, M. E. Macromolecules 1995, 28, 5834-5840. (16) Fu, B.; Bakker, E.; Yun, J. H.; Yang, V. C.; Meyerhoff, M. E. Anal. Chem. 1994, 66, 2250-2259. (17) Dai, S.; Esson, J. M.; Lutze, O.; Ramamurthy, N.; Yang, V. C.; Meyerhoff, M. E. J. Pharm. Biomed. Anal. 1999, 19, 1-14. (18) Badr, I. H. A.; Ramamurthy, N.; Yang, V. C.; Meyerhoff, M. E. Anal. Biochem. 1997, 250, 74-81. 10.1021/ac000756g CCC: $20.00
© 2001 American Chemical Society Published on Web 12/14/2000
immunoassay scheme in which synthetic polycationic peptides are employed as labels.17 Given the large number of potential applications of PSEs, it is of interest to further enhance the sensitivity of these electroanalytical devices. Previously, it was noted that lowering the plasticizer content in the polymer membrane matrix can improve the sensitivity of PSEs (by reducing the diffusion coefficient of the polyion-exchanger complex).6,16 Further, the shape of the electrode has also been found to affect the sensitivity of PSEs, with a cylindrical membrane electrode design being slightly more sensitive than a planar membrane configuration owing to the enhanced mass transfer of the polyion to the membrane surface by cylindrical diffusion versus planar diffusion.16 It was also noted earlier that stirring the sample solution (with a stir bar) versus nonstirring results in an improvement in analyte sensitivity for PSEs. Indeed, taken together, each of these observations initially helped to determine, definitively, the operative nonequilibrium response mechanism of PSEs.16 Additional improvements in polyion sensitivity could, however, be made by employing a method to further enhance mass transfer of the analyte polyion to the membrane/sample interface in a controllable manner. Herein we introduce an effective methodology to improve the sensitivity of PSEs by designing a rotating potentiometric PSE system. Rotating electrode voltammetry and amperometry are well-established hydrodynamic methods that yield enhanced mass transfer of analyte as a function of the rotation speed of a planar working electrode.19 It will be shown here that the sensitivity of PSEs (both polycation (protamine) and polyanion (heparin)) can also be improved substantially by rotating the potentiometric electrode, with a linear relationship found between the observed sensitivity and ω-1/2 (where ω is the rotating angular frequency). EXPERIMENTAL SECTION Reagents. Tridodecylmethylammonium chloride (TDMAC), dioctyl sebacate (DOS), 2-nitrophenyl octyl ether (NPOE), high molecular weight poly(vinyl chloride) (PVC), and tetrahydrofuran (THF) were purchased from Fluka Chemika-Biochemika (Ronkonkoma, NY). Heparin, sodium salt (from bovine intestinal mucosa, 188 units/mg), protamine sulfate (from herring, grade III), and tris(hydroxymethyl)aminomethane (Tris) were products of Sigma Chemical Co. (St. Louis, MO). Calcium dinonylnaphthalenesulfonate (DNNS) was a kind gift from King Industries (Norwalk, CT). Polyurethane (M48) was supplied by Medtronic Inc. (Minneapolis, MN). All other reagents were analytical grade or better. Deionized water (16 MΩ) was used to prepare all aqueous solutions. Unless otherwise noted, the primary buffer solution used for all experiments was 50 mM Tris-HCl, pH 7.4, containing 0.12 M NaCl. Membrane and PSE Electrode Preparation. Heparinsensitive membranes were prepared by the cocktail solution casting method as described previously.20 The cocktail solution was prepared by dissolving the appropriate amounts of membrane components (polymer, plasticizer, and ion exchanger) into THF. The final thickness of the membrane was ∼150 µm and it contained 1 wt % TDMAC, 33 wt % DOS, and 66 wt % PVC. Protamine-sensitive membranes were made by the same method (19) Bard, A. J.; Faulkner, L. R. Elecrochemical Methods: Fundamentals and Applications; John Wiley & Sons: New York, 1980. (20) Mathison, S.; Bakker, E. Anal. Chem. 1999, 71, 4614-4621.
Figure 1. Schematic of the rotating polyion-sensitive potentiometric membrane electrode.
but with the following composition: 1 wt % DNNS, 49 wt % NPOE, and 49 wt % M48. Disks of polyion-sensitive membranes were cut with a cork borer (o.d. 7.0 mm) and were glued at one end of 1-cm-long Tygon tubes (i.d. 4.2 mm, o.d. 7.35 mm; Fisher Scientific, Pittsburgh, PA). Rotating PSE System. A Pine Instrument Co. (Grove City, PA) analytical rotator (model ASR) and an ASR motor control box (1000 rpm/V, 200-10 000 rpm range) were used for all experiments. A 2-cm-long connecting tube made of black Delrin plastic (McMaster-Carr, Cleveland, OH) was used to connect the rotator and the 1-cm-long Tygon tube with the polyion-sensitive membrane disk glued at the distal end. Both the Delrin tube and the Tygon tube were filled with internal filling solution (Tris buffer with 0.12 M NaCl). The internal reference electrode was then inserted through the central void space of the rotator and down to near the surface of the polyion-sensitive membrane surface, as shown in Figure 1. The internal reference electrode was made with a thin silver wire (o.d., 0.076 mm, Medwire, Mt. Vernon, NY) inserted through a PEEK tube (i.d. 0.13 mm and o.d. 1.6 mm; Supelco, Bellefonte, PA) with ∼0.5-cm-long piece of Ag exposed. This exposed region was chloridized with a 1 M HCl solution containing 0.1 M FeCl3. The PEEK tubing sheath around the inner Ag/AgCl electrode must be used to minimize the electrical noise during high-speed rotation. The tubing prevents any electrical contact between the inner Ag wire and the rotator. In addition, the internal reference electrode needs to be mechanically isolated from the rotator to avoid any vibration coupling from the rotator. The external reference electrode employed was a 1-mm-diameter silver wire chloridized with solution of 1 M HCl containing 0.1M FeCl3. Potentiometric Measurements. The EMF responses were measured at ambient temperature (∼23 °C) via a Macintosh IIcx computer equipped with a LAB-MIO-16XL-42 16 bit A/D I/O board (National Instruments, Austin, TX) and VF-4 electrode interface module (World Precision Instruments, Sarasota, FL), controlled by LabView 2 software (National Instruments, Austin, TX). Heparin Concentration Measurements via Continuous Protamine Titration. Titrations of 0.05 unit/mL heparin in 3 mL of buffer (50 mM Tris-HCl, pH 7.4, containing 0.12 M NaCl) were Analytical Chemistry, Vol. 73, No. 2, January 15, 2001
333
carried out by continuous infusion of 0.1 mg/mL protamine aqueous solution with a syringe pump (model MD-1001, BAS Inc., West Lafayette, IN), at a infusion rate of 5 µL/min. The titrations were monitored with protamine-sensitive electrodes either with a rotating (at 3000 rpm) or static mode configuration (using a stir bar) to achieve solution-phase convection. Blank titration curves for protamine-sensitive electrodes (static and rotating ones) were also recorded with continuous protamine infusion into a buffer-only solution (0 units/mL heparin) under the same experimental conditions. Results were averaged and the confidence intervals were calculated using the Student’s t-test at the 95% confidence level. RESULTS AND DISCUSSION Principle. In contrast to conventional ion-selective membrane electrodes that operate under equilibrium conditions, the EMF responses of PSEs are generated by a nonequilibrium quasi-steadystate ion-exchange process that occurs at the membrane/sample interface.16 It has been shown that this steady state occurs when the flux of polyions diffusing from the sample phase to the membrane surface equals the flux of the polyion-ion-exchanger ion pair that diffuses away from the membrane surface into the bulk of the polymer membrane. At low polyion concentrations, where a significant fraction of the original ion-exchanger counterions (inorganic cations for DNNS-based polycation sensors and inorganic anions for TDMA-based polyanion sensors) are still present at the outer surface of the organic membrane, the ∆EMF observed under such conditions can be described as follows:16
∆EMF ) (
(
RT z Daδm c ln 1 F RT Dmδa poly,bulk
)
(1)
where, RT is the total concentration of ion-exchanger sites within the membrane phase; z is the charge number on the analyte polyion; Da and Dm are the diffusion coefficients of polyion in the aqueous and membrane phases, respectively; δa and δm are the diffusion layer thicknesses for the polyion in the aqueous phase and the membrane phase, respectively; Cpoly,bulk is the bulk concentration of polyions in the sample solution; + is for polyanion response and - is for polycation response; T is the temperature in kelvin; and R and F are the gas and Faraday constants, respectively. From eq 1, it is clear that to achieve the same ∆EMF response, a simple way to lower detection limits toward given polyions (i.e., smaller Cpoly,bulk) is to reduce the diffusion layer thickness in the aqueous phase (δa) while keeping all the other parameters constant. An effective approach to reproducibly control and further reduce the diffusion layer thickness is to rotate the membrane electrode. For a disk electrode, δa is related to the angular rotating frequency ω as follows:21
δa ) 1.61/Da1/3υ1/6ω-1/2
(2)
where ν is the kinematic viscosity (defined as the ratio of the normal viscosity η to the solution density F). In accordance with (21) Hamann, C. H.; Hamnett, A.; Vielstich, W. Electrochemistry; Wiley-VCH: Weinheim, 1998.
334 Analytical Chemistry, Vol. 73, No. 2, January 15, 2001
Figure 2. (a) Heparin calibration curves obtained with rotating heparin-sensitive electrode in Tris buffer, containing 50 mM Tris and 0.12 M NaCl, at different rotation speeds. Calibration data for 0 rpm are obtained by convection of sample with magnetic stir bar. (b) Relationship between measured C1/2, the polyanion concentration corresponding to half the total maximum ∆EMF response toward heparin, and ω-1/2, where ω is the rotation angular frequency.
eq 1, a decrease of 10-fold in δa should require 10-fold lower polyion concentration to yield the same EMF response. Rotating Polyanion- and Polycation-Sensitive Electrodes. The above concept can be demonstrated for both polyanion- and polycation-sensitive membrane electrodes. The calibration curves for polyanionic heparin obtained by using several rotation speeds in Tris buffer are shown in Figure 2a. Also shown is the response observed when the electrode is not rotated, but the sample is mixed using a conventional stir-bar. Obviously, the potentiometric response curves are shifted toward much lower concentrations by rotating the membrane electrode. Specifically, without rotation, the lower limit of detection (LLOD), defined as the polyion concentration that yields an average ∆EMF value from background buffer signal of (3 mV (+ in the case of polycation measurements; - in the case of polyanion measurements), is ∼0.1 unit/mL; however, with rotation at 5000 rpm, the LLOD was lowered to 0.01 unit/mL, a 10-fold improvement. A similar effect was observed for the polycation PSEs response toward protamine (see Figure 3a), where rotation at 6000 rpm yields a detection limit of 0.02 µg/mL. In addition, by defining the polyion concentration that corresponds to the ∆EMF that is half of the total maximum ∆EMF (maximum occurs when sample concentration of polyion is high enough to achieve full equilibrium at the membrane/sample interface15) as C1/2, which is proportional to the LLOD, a linear
Figure 3. (a) Protamine calibration curves obtained with rotating protamine-sensitive electrode in Tris buffer, containing 50 mM Tris and 0.12 M NaCl, at different rotation speeds. Calibration data for 0 rpm are obtained by convection of sample with a magnetic stir bar. (b) Relationship between measured C1/2, the polycation concentration corresponding to half the total maximum ∆EMF response toward protamine, and ω-1/2, where ω is the rotation angular frequency.
relation was found between C1/2 and ω-1/2 for both the polyanion and polycation sensors (see Figures 2b and 3b). This linear relation can be theoretically predicted by inserting eq 2 into eq 1. Thus, the improvement in sensitivity with increasing rotation speed is the result of a decrease in diffusion layer thickness, not an artifact resulting from a possible change in the threedimensional structure of protamine or heparin (i.e., uncoiling or unfolding) caused by the vigorous hydrodynamic convection. A direct comparison of the mass transfer of polyion to the surface of the membrane for the two hydrodynamic cases (stir bar convection vs rotating the electrode) can also be made by determining the rotation speed required to achieve the same LLOD value for the two configurations. Indeed, in the case of protamine measurements with the polycation-sensitive membrane electrode, it has been found that that the equivalent LLOD is obtained when the rotation speed is 250 rpm (data not shown) (i.e., same LLOD as when stir bar convection is used). Assuming a kinemetic viscosity of 10-6 m2/s for the aqueous test solution and an aqueous phase diffusion coefficient for protamine of 5 × 10-7 cm2/s, this equivalent diffusion layer thickness corresponds to 11.6 µm. In principle, since the sensitivity (as indicated by C1/2) is controlled by the rotation speed, additional lowering of the LLOD
should be possible by further increasing the rotation speed. However, because C1/2 is proportional to ω-1/2 rather than to ω-1 itself, further increasing the rotation speed above 6000 rpm will not significantly decrease the value of ω-1/2. Additionally, the mechanical noise becomes much more substantial at rotation speeds above 6000 rpm. Hence, there is no analytical advantage gained by attempting to operate PSEs at rotation speeds above this value. The improved sensitivity achieved by rotating the PSEs can also be applied to detect lower concentrations of one polyionic species when titrated with another. Indeed, for real sample measurements with PSEs, titrations are advantageous, since the steady-state EMF responses shown in Figures 2 and 3 are also dependent on the background cation or anion activities in the sample solution (e.g., for biological samples, sodium and potassium in the case of polycation sensors, chloride in the case of anion sensors).3,16,22 Indeed, more reliable analytical results for measurement of polyion levels in complex samples, including whole blood, have been achieved by carrying out such potentiometric titrations.9,10 Figure 4a shows the average (( confidence interval at 95% level with respect to time axis) results for at least four separate titrations (via continuous infusion of 0.1 mg/mL protamine solution) of 0 and 0.05 unit/mL heparin using static and rotating (3000 rpm) protamine-sensitive PSEs. Although 0.05 unit/mL heparin can be distinguished statistically from the blank using the nonrotating PSE as the end point indicator, much greater precision in the titration data is obtained using the more sensitive rotating protamine PSE. Indeed, Figure 4b shows an expanded time scale plot for the same rotating PSE-based titrations shown in Figure 4a, further illustrating the dramatic enhancement in precision that is achieved. Given this improved precision, it appears that levels of heparin down to 0.01-0.02 unit/mL could easily be resolved from the blank using the rotating PSE as the indicator electrode in such titrations. In addition, as shown in Figure 4a, much more rapid titrations can be completed owing to the improved detection limits of the rotating PSE design. In contrast, to increase the titration speed when using a static protamine PSE as the detector, one must increase the infused protamine concentration to 1 mg/mL (using the same flow rate); however, at such high concentrations of titrant even 0.1 unit/mL heparin could not be distinguished statistically from the blank (results not shown). It should be noted that, in previous electrochemical experiments,9 it was found that 1 unit of heparin requires ∼10 µg of protamine for neutralization when conducting manual titrations of heparin. In the experiments associated with Figure 4, a total amount of 0.15 unit of heparin (0.05 unit/mL heparin in 3 mL of buffer) was titrated with 0.1 mg/mL protamine at a speed of 5 µL/min. For the rotating electrode experiment, the time difference between C1/2 of the blank and that containing 0.15 unit of heparin is 104.0 ( 5.3 s; i.e., the infused protamine at this end point is 0.1 mg/mL × 5 µL/min × 104 ( 5.3 s ) 0.867 ( 0.044 µg of protamine or 5.78 ( 0.29 µg of protamine/unit of heparin. This is less than the previous reported stoichiometry between protamine and heparin (∼10 µg of protamine/unit of heparin). This difference is likely due to the combination of using continuous infusion of (22) Fu, B.; Bakker, E.; Yun, J. H.; Wang, E. J.; Yang, V. C.; Meyerhoff, M. E. Electroanalysis 1995, 7, 823-829.
Analytical Chemistry, Vol. 73, No. 2, January 15, 2001
335
can thus be calibrated for quantitative determinations of heparin using the rotating electrode configuration. Beyond improvements in detection limits and precision as demonstrated above in the 3-mL samples, use of rotating PSEs may also be advantageous from the standpoint of achieving more reproducible results in small sample volumes (
