Development of sensitive automated pH meter for real-time biosensor applications Sandeep Kumar Jha, Stanislaus F. D’Souza Nuclear Agriculture and Biotechnology Division, Bhabha Atomic Research Centre, Mumbai 400085, India; E-mail: [email protected] A computer controlled pH meter was developed using simple op-amp circuits. Analog to digital converter (ADC) card was used to acquire the potential output from the op-amp. Software was written for acquisition and averaging of the signal. Software based averaging of the signal helped in reduction of electrical noise. The instrument was calibrated against pH and ammonia selective electrodes where it followed Nernstian behavior with slopes of 59.08 and 55.53 mV per decade of pH and ammonia respectively.

Keywords: urea biosensor, potentiometric, ion selective electrode, op-amp.

INTRODUCTION With increasing environmental pollution levels and resulting health concerns demand for precise and rapid measurements of pollutants are growing1. Clinical diagnostics2 and military concerns over security threats from biological warfare agents3 have also produced the necessity of determination of analytes with high specificity. Agriculture, food processing, healthcare and pharmaceutical industries need real-time and in-situ sample analysis4. In these circumstances, biosensors have emerged as the analytical devices that could perform analysis of samples in real-time environments. Being sensitive, reproducible and sometimes reusable, they hold the promise to the market requirements. Since the introduction of glucose sensor in 1962 by Clark and Lyons5 this particular branch of analytical sciences was still in infancy till the growth of cutting age instrumentation techniques, availability of faster computation and data analysis tools in the past two decades. A biosensor can be defined as a device that detects, transmits and records information regarding a physiological or biochemical change4,6. A biosensor comprises two distinct elements: first a biological recognition element e.g. enzymes, antibodies, microbial cells, DNA or RNA, animal or plant tissue or a combination of these that could trigger physicochemical changes upon interaction with the analyte under concern4. The second component is a signal transduction element (e.g. optical, potentiometric, amperometric, acoustic, thermal and electrochemical etc.) connected to a data acquisition and processing system which converts the signal from the biological element into a quantifiable signal, e.g. current, potential or frequency etc. Because of diffusional constraints it is essential that these biological recognition elements should be in close contact with the transducer. This requires some sort of materials that can fix them onto transducer surface. These matrices are mainly classified

into synthetic polymers, metal or semiconductor electrode surfaces and sometimes biopolymers4 and the process is called immobilization. Immobilization not only helps in forming the required close proximity between the biomaterial and the transducer, but also helps in stabilizing it for reuse. Biomaterials can be immobilized either through adsorption, entrapment, covalent binding, cross-linking or a combination of all these techniques7,8. A number of techniques have been developed in our laboratory for the immobilization of viable and non-viable cells as well as cell–enzyme conjugates7-11. Selection of a technique and/or support would depend on the nature of the biomaterial, substrate and configuration of the transducer used. For example, entrapment and adsorption techniques are more useful when viable cells are used4. The physicochemical changes attributed by interaction of analyte with the recognition element can be sensed by converting them preferably into current or potential by electronic circuit. Thereafter, it becomes quite easier to acquire, process and manipulate the signal for measurement of analytes. This may be achieved via two different strategies, either by developing microprocessor or microcontroller based laboratory instruments or alternatively by using microchip based hand-held or computer controlled devices. The later approach supports automated analysis and is convenient for discrete analysis, where it prevents human error, improves accuracy and precision. It can process many samples together and reduces reagent and operator costs. A computercontrolled system also gives users flexibility to design the analysis as per their needs. Automation permits on-line monitoring in industrial and clinical processes, where information is needed in quasi real-time. In this paper we have demonstrated the design, calibration and operation of a simple computer controlled pH meter that may find use in potentiometric biosensor systems especially for urea determination. A urea biosensor works on the principle of enzyme-catalyzed reaction of urease over its substrate urea. The reaction liberates ammonia that can be sensed using potentiometric ammonia selective electrode. These electrodes fall in the category of ion selective electrodes (ISEs)12, where inbuilt pH electrode is separated from surrounding with an ammonia selective Teflon or PVC membrane. These membranes contain ionophore nonactin that allows ammonia to selectively diffuse in the interior of the electrode. This increases the pH of the internal filling solution and producing an alteration in potential, which can be amplified using electronic circuits. We have also devised strategy to make use of software control for better signal processing and noise reduction.

EXPERIMENTAL Electronic assembly An electronic circuit was developed using a high input impedance; ultra low offset voltage, low bias current op-amp LF 35613. Due to the high impedance characteristics of the pH and ammonia selective electrodes a voltage follower was used at the first stage. However, the signal was quite feeble to achieve a high sensitivity. Hence the potential was further amplified ten fold by an op-amp inverter at the second stage so that the voltage can be measured by a simple 4 ½ digit digital multimeter. Automated

data acquisition was carried out by connecting the amplified output to a 12 – bit analog to digital conversion card with digital I/O capabilities (Dynalog India Ltd.). This card is placed in the ISA slot of a computer and is responsible for the conversion of the analog signal produced by the electrode. The complete electronic system with electrode connectors, except the digital I/O and A/D converter card, was placed in a grounded aluminium box as a further noise-suppression device. The ADC card has a resolution of 2.44 mV for a full-scale voltage of 10 V (± 5 V). Software Since the circuit did not have any inbuilt noise filters it was mandatory to average the signal using the software. The software that controls the analyzer and the signal processing modules was developed in our laboratory by using C++ language. The machine language routines supplied with the A/D conversion card were embedded in the program. These subroutines manage the communications between the A/D card and the computer. The software acquires the data after receiving manual keyboard start command, reads the potential from the sensor and the time of the measurement. It also identifies between positive, negative or differential signal peak and its size and calculates the analyte concentration after a predefined sampling period from the stored calibration data. When the system measures concentration of samples, the program automatically repeats the cycle, until an abort order is placed through the keyboard. The program is capable of reading 30000 voltage measurements per second. It averages and translates one averaged potential readings per second. The data points could be recorded and stored in files for further analysis. Chemicals and reagents All the chemicals used in the experiments were of high purity and analytical reagent grade. Standard buffer tablets were purchased from Qualigens fine chemicals, Mumbai. All the reagents were prepared in double distilled water. pH of the solutions were adjusted by addition of dilute acid or alkali while keeping the ionic strength constant with addition of KCl. For calibration of the instrument with ammonia, 20 ml solution of ammonium chloride to different concentrations were placed in a beaker. Ammonia selective electrode (Orion Research Inc.) was immersed in this mixture, while avoiding trapping of air bubbles on the tip. 0.5 ml of 5 M NaOH was added to it and data acquisition was initiated using keyboard shortcut.

RESULTS AND DISCUSSION Signal conditioning An electrical noise in the circuit reduces the performance of the sensor not only by sacrificing the lower detection limit but also the sensitivity of the instrument. To overcome this problem a number of data points were averaged per unit time using software. It was observed that the number of sample points averaged and the averaging time had significant contributions to the signal quality (Fig.1 A-E). For justification, a pH electrode was placed in the standard buffer of pH 4. For uniformity, all the potential

measurements were carried in the same solution. Averaged data points were recorded with varying sampling time (100 m sec to 3 sec). Non-averaged signal had wide fluctuation from the mean. The standard deviation (SD) of data points from their mean position was as high as 17.1 mV for 100 m sec averaging period, whereas 1.09 mV for 2 3 sec of averaging. For a rapid measurement6, biosensor sampling should be done within 10 to 30 sec and for this reason, we have chosen the averaging time of 1 sec. Interestingly in another observation it was noted that by stirring the solution also causes a drift in the potential (Fig. 1 F). For that reason we have chosen not to stir the solution during sampling. Calibrating the instrument with pH and ammonia selective electrodes The pH meter along with the software designed for proper signal conditioning was calibrated with respect to a commercially available analog pH meter (Model LI 127, Elico Ltd., Mumbai). Potential of the solutions having different pH values were first recorded with the analog meter and then with the instrument under development. The slope of the straight lines drawn between the observed potentials of the pH solutions, where nearly equal for the first op-amp (follower circuit) and the analog meter (Fig. 2). The op-amp at the second stage inverted the sign of the potential for ease in calculations at basic pH range. The straight-line equations for the first and second op-amps were V = 0.40839 - 0.05832 × pH and V = - 4.1544 + 0.59082 × pH respectively, following Nernstian behavior12. Since the instrument was to be used alongwith an ammonia selective electrode for development of urea biosensor, the instrument was calibrated with ammonia solutions. Upon addition of alkali to the ammonium chloride solution, the sudden pH change liberated most of the ammonia, giving rise to an increase in potential response from the ammonia selective electrode. The plot of potential change versus ammonia concentration expressed as logarithmic scale obeyed Nernstian behavior with the slope of 555.3 mV per decade of ammonia concentration with respect to second op-amp (Fig. 3).

CONCLUSION A potentiometric system follows Nernstian equation with usual slope in the range of 55 to 58 mV per decade of analyte12. The instrument developed obeys this equation perfectly and will find use in biosensor studies where the sensitivity of the instrument would be as low as 2.5 mV. With online data acquisition capability and permissible modifications in the software the instrument will be very useful in the development of potentiometric biosensors.

ACKNOWLEDGEMENT Mr. Sandeep Kumar Jha is a recipient of a fellowship from Council of scientific and Industrial Research, New Delhi, India. References 1. I. Karube, Y. Nomura, J. Mol. Catal. B: Enzymatic, 10, 177–181(2000).

2. J. Wang, J. Pharma. & Biomed. Anal., 19, 47–53 (1999). 3. S. S. Iqbal, M. W. Mayo, J. G. Bruno, B. V. Bronk, C. A. Batt, J. P. Chambers, Biosens. & Bioelectron., 15, 549–578 (2000). 4. S.F. D’Souza, Biosens. & Bioelectron., 16, 337–353 (2001). 5. L.C. Clark Jr., C. Lyons, Ann. NY Acad. Sci., 102, 29-45 (1962). 6. A. P. F. Turner, I. Karube, G. S. Wilson, Biosensors, Fundamental and applications, Second Edn., Oxford university press, NY. 7. S. F. D'Souza, Ind. J. Microbiol., 29 (2), 83-117 (1989). 8. S. F. D'Souza, Current Sci., 77, 69-79 (1999). 9. S. F. D'Souza, Appl. Biochem. Biotechnol., 96 (1/3), 225 – 238 (2001). 10. S. F. D'Souza, Ind. J. Biotechnol., 1, 321-338 (2002). 11. S. K. Jha, S. F. D’Souza, J. Biochem. Biophys. Method, (In print) (2005). 12. P.R. Unwin, A. J. Bard, M. Stratmann, (Eds.), Encyclopedia of Electrochemistry, Vol. 3: Instrumentation & Electroanalytical Chemistry, Wiley – VCH GmbH & Co., KGaA, Weinheim, Germany, pp 7 -13 (2003). 13. P. Horowitz, W. Hill, The art of electronics. Second Edn., Cambridge University press, Cambridge, UK, pp 202 (1995).

Potential (V)

-1.66

A

B

-1.66

-1.67

-1.67

-1.68 -1.66

-1.68 -1.66

C

D

-1.67

-1.67

-1.68

-1.68

Time (sec) Fig. 1: Standardization of signal averaging period for the ADC card in standard buffer of pH 4. [A] Averaging time of 100 m sec (SD 17.1 mV); [B] 500 m sec (SD 4.45 mV); [C] 1000 m sec (SD 2.19 mV); [D] 2000 m sec (SD 1.09 mV); [E] 3000 m sec (SD 1.05 mV from mean position); [F] Drift in signal by stirring of buffer (at 1000 m sec averaging time); the two sudden drops in potential represents stirring points.

3

4

Potential (V)

Potential (V)

2 1 0 -1 -2

3

2

1

2

4

6

8

10

12

14

pH

Fig. 2: Calibration of the computer controlled pH meter with respect to commercial analog pH meter. Straight-line equations for [○] commercial pH meter, (V = 0.40913-0.05845 × pH); [+] 1st op-amp output, (V = 0.40839-0.05832 × pH) and [■] 2nd op-amp output, (V = 4.1544+0.59082× pH) of the fabricated instrument.

-10

-8

-6

-4

-2

0

Log10 [NH3 Concentration, M]

Fig. 3: Nernstian behavior of pH meter with respect to different concentrations of NH3 (dotted line). Straight-line equation at central portion of curve represents Y = 0.55533 × X + 4.57511, Chi2=0.00121, R2=0.99902.