An improved microelectrode resistance meter

Journal of Neuroscience Methods, 17 (1986) 335-342

335

Elsevier

N S M 00609

An improved microelectrode resistance meter Robert S. Stephenson and George W. Overbeck Department of Biological Sciences, Wayne State Unioersity, Detroit, MI 48202 (U.S.A.) (Received 15 August 1985) (Revised 20 May 1986) (Accepted 27 May 1986)

Key words: Micropipette - Mlcroelectrode - Resistance - Megohm - Ohmmeter - Megol~mmeter M e t e r - Instrument

A device is described for measuring the resistance of micropipette electrodes. The useful range of electrode resistances that it can measure is 100 kl2 to 1 GI2. It is more convenient to use than previously described or commercially available meters, especially for very high-resistance electrodes. Resistance of even the finest microelectrodes can be measured accurately while their tips are inserted by hand into a test solution. This eliminates the need for special holders, speeding and simplifying the screening of large n u m b e r s of electrodes. Test solutions are stored in interchangeable reservoirs, making it easy to characterize the resistance of an electrode in solutions of different resistivity. Test solutions can also be capped and removed when not in use to prevent evaporation. To protect very high-resistance electrodes from damage during measurement, the measuring current is low (only 300 pA on the highest resistance range) and the test voltage across the electrode is limited to + 1 V.

Introduction

The performance of glass capillary microelectrodes in penetrating and recording from nerve cells and the electrical characteristics of these electrodes depends critically on the geometry of their tips (tip diameter, angle of taper, ratio of inner to outer diameter, etc.). Since the pulling of microelectrodes is more art than science, it is necessary to monitor their geometry routinely to optimize the settings of the puller. Light microscopic observation is of limited value because the tip diameter (generally 0.1/~m or less) is below the limit of resolution, and examination under the electron microscope is too time consuming for routine use. The most rapid way to monitor tip diameter is by measuring the electrical resistance of the microelectrode Correspondence: R.S. Stephenson, Department of Biological Sciences, Wayne State University, Detroit, MI 48202, U.S.A. 0165-0270/86/$03.50 © 1986 Elsevier Science Publishers B.V. (Biomedical Division)

336 placed in a saline test solution. This resistance also depends on the resistivity of the filling and test solutions and the angle of taper, but is critically influenced by the tip diameter (Schanne et al., 1968; Plamondon and Gagn6, 1980). The procedure for measuring electrode resistance is generally to pass a constant direct or alternating current through the electrode and measure the resulting voltage drop, which is proportional to the resistance. For reliable measurements, several criteria must be met: (1) Alternating current should be used to prevent significant errors due to rectification and tip potentials (Rubio and Zubieta, 1961; Schanne et al., 1968), (2) The measuring current must be small enough that the electrode behaves ohmically~ (3) The measuring frequency must be low enough that capacitative current does not introduce errors, (4) The circuit should operate stably and accurately over a wide range of values of electrode resistance and capacitance, (5) The output of the circuit should be proportional to resistance, and cover the range of common microelectrode resistances (1-250 M g ) . A number of circuits have been proposed that satisfy all or most of these criteria (Schanne et al., 1968; Sabah, 1973; Naylor, 1978). The design proposed here, however, has several features in addition to those mentioned above. It is quite insensitive to pick-up from the power lines. This means that microetectrodes of even hundreds of megohms in resistance can be measured accurately while held in the hand, greatly speeding up the task of screening many at a time. The circuit settles quickly and output is displayed directly on an analog meter rather than an oscilloscope, also contributing to speed and convenience of operation. In the meter proposed here, test solutions are readily interchangeable, making it easy to test each electrode in several solutions. When not in use, solutions can be capped to prevent evaporation, or even refrigerated. The variation of microelectrode resistance with the conductivity of the test solution can yield important information about the tip geometry, as well as identify incompletely filled electrodes (Schanne et al., 1968; McCann and Stibitz, 1977; Plamondon and Gagn6, 1980). Thus, it is advantageous to have interchangeable test reservoirs so one can test a microelectrode in a battery of different solutions easily. An additional feature of the circuit is that its open-circuit voltage is limited to + 1 V, slightly more than the value applied across an electrode when the meter reads full scale. This protects the finest-tipped electrodes from possible electrical damage while under test. At the instant the electrode is inserted into the test solution, it is subjected to the open-circuit voltage, which in other resistance meters may be as large as 15 V. Since a significant fraction of this voltage falls across the glass near the tip, where the wall is very thin, the resulting electric field severely stresses the glass. If 10 V falls across a flat wall of 0.01/tm thickness the electric field is on the same order as the dielectric breakdown strength of borosilicate glass (900 V / # m : Holland, 1964). Curvature of the wall and hydration of the glass tend to aggravate this problem. Limiting the open-circuit voltage to 1 V minimizes the chance of damage to very high resistance electrodes.

337

Basic circuit operation The basic circuit is shown in Fig. 1. The unknown microelectrode resistance R e forms a negative feedback loop with operational amplifier A], so that the amplifier's inverting ( - ) input is maintained at ground potential. The voltage generator and large, calibrated resistor R c cause a square wave of current to flow to the inverting input. Since the op-amp A~ itself draws negligible current, an equal but opposite

Fig. 1. Simplified circuit diagram of resistance meter. R e represents the resistance of the microelectrode immersed in the test bath.

square wave of current must flow through R e. The voltage at the output of amplifier A 1 is equal to the voltage drop across R e. This square wave voltage is AC-coupled by C] and R~, and then rectified by amplifier A 2, which has a gain of + 1 or - 1 depending on the phase of the square wave generator. The output of A 2, a DC voltage proportional to electrode resistance R e, is displayed on the meter. The amplitude of the square wave generator and the value of resistor R c determine the

15+

i3M 1M

v/4 U~

I/4 U 1

~/4U1

500

K5K

F!

:,..~L~o. K

Fig. 2. Circuit diagram of resistance meter. Integrated circuits are identified in Table I.

A3

338

amount of test current to flow through electrode R e. By changing Re, the circuit can accommodate electrodes of different resistances. The test bath and most of the micropipette being measured lie within a grounded A ~

0

0

q

0

0

0

t

I

Fig. 3. Pattern for printed circuit, actual size; A: foil side. B: component side, showing parts placement. For component values, see Table I. Leads terminating with large, sofid dots are soldered directly to the ground plane on the component side of the board, All other holes are countersunk so that the lead passes through without touching ground plane. The d o t t ~ circle around pin 2 of A 1 (the input terminal) indicates that the IMa ~ ~ a 3 mm h¢~ ht t h ¢ ~ board b¢fore ~ sctdCt~ to a contact

339

shield. The part of the shank of electrode R e which protrudes from the shield is connected to a very low impedance (viz. the output of A 1) and is thus insensitive to hum pickup. Since the inverting input of A 1 is already a virtual ground, the B

Z u

insulated from the foil side of the circuit board by a Teflon stand off. Components R24-27 and C9 run from this contact directly to the range switch, and are not shown in Fig. 3B. The pad labelled S on the circuit board is connected to the wiper of the range switch, and a lead from pad E connects to a platinum wire inserted in the' shank end of the electrode under test.

340 capacitance between shield and test bath has little effect on the circuit's performance.

Circuit construction details

The detailed circuit is shown in Fig. 2. The input amplifier referred to as A 1 in Fig. 1 is actually composed of two op-amps: A 1 and A 4. A 1, the input amplifier, is an AD515 FET-input electrometer with very low input bias current. When the input of A1 is open-circuited, its output saturates at close to the positive or negative supply voltage ( + 1 5 V). The voltage divider composed of the 1 MI2 and 68 KS2 resistors limits the open-loop voltage at the output of m 4 to about + 1 V. The square wave generator is a comparator (actually part of A 2, a GAP-01 integrated circuit) set up to oscillate at about 30 Hz. Its output is AC-coupled through the 1 # F condenser and reduced by the voltage divider composed of the 1 MI2 resistor and the 50 kI2 potentiometer. The resulting square wave (about + 0.3 V) is applied to the selector switch which chooses the value of R c. The 4 resistors shown give full-scale ranges of 1 MI2, 10 M~2, 100 MI2 and 1 GO. The corresponding peak measuring currents are 300 nA, 30 hA, 3 nA and 300 pA, respectively. The output of A1 is AC coupled, inverted and amplified by m 3 before being applied to the inputs of A 2. This GAP-01 integrated circuit is composed of two

TABLE I COMPONENT VALUES Resistors RI,2,7,s,I3,14

Capacitors

Cl C2

g16 R17 R18,22

1MI2 wirejumpers 330 kI2 470 k[2 68 kI2 50 kI2 trimpot 500 k9 trimpot 5 k9 trimpot 39 kl2 680 k~ 10 k12

RI9

6.8 k9

A2

R2o,2!

20 k9 1%

A3,4

R23

2.2 M9

U1

R24 R25 R26 R27

1 MQ 1% 10 M~ 1% 100 Mt2 1% 1000 MQ 1%

R 3.4 R5 R6 R9,12

Rio Rll R15

C3, 7

C4 C5 C6 C8 C9

22 pF 500 pF 0.0039 pF 0.047 pF 33 pF 1.0 #F 0.47 #F 10 pF

Integrated circuits

A1

ADS15JH (AnalogDevices, Nor~vood,MA, U.S.A.) GAP-01 (Precision Monolithics, Clara, CA, U.S.A.) LM308H (National Semiconductor, Santa Clara, CA, U,S.A.) CD 4093BC(National Semiconductor, Santa Clara, CA, U.S.A.)

341

op-amps whose outputs are connected by CMOS analog switches to the input of a third, which functions as a voltage follower. One of the input operational amplifiers in Fig. 2 is configured for a gain of + 1, the other for - 1 . The overall gain of A 2 may thus be either + 1 or - 1 depending on which analog switch is closed. If both switches are open, the 500 pF capacitor holds the last input voltage. The circuit in Fig. 2 takes advantage of this sample and hold capability to eliminate capacitative transients occurring when the square wave changes sign. The Schmitt-trigger N A N D gates (U1) form two circuits that delay the onset of the positive and negative phases of the square wave, respectively, by about 4 ms. During these delays the last output of A 2 is held. After the delay has passed (and the capacitative transient at the input has settled) the corresponding analog switch closes and the appropriate input t o A 2 is connected to its output, and thus to the meter. The circuit of Fig. 2 was built on a 10 × 16.5 cm, double-sided printed circuit board whose pattern is reproduced full-size in Fig. 3A. Fig. 3B shows the component placement, and Table I gives the component values. Care was taken to position

Fig. 4. Details of the completed instrument. A: bottom view, showing component layout. Power supply and meter are at the top. Note the two shielding partitions, and the copper ground plane on the component side of the printed circuit board. Input amplifier A~ is at bottom. The test solution reservoir fits into a metal shielding tube located behind the circuit board, immediately adjacent to A~. B: top view of resistance meter, showing range selector switch and test reservoir. The wire emerging below the meter is normally inserted into the shank of the electrode under test. An extra test solution reservoir is shown at left. Note the contact on its bottom.

342

the input op-amp A I , the four R c resistors and the range selector switch away from the rest of the components. The copper foil on the component side of the circuit board forms a ground plane which further shields the sensitive portions of the circuit. All components are mounted, together with an AC power supply, in a 9 × 15 × 25 cm box, as shown in Fig. 4A. An internal partition shields the power supply and the AC power lines from the rest of the circuit. Fig. 4B shows the front of the instrument. A short length of 26-gauge platinum wire, soldered to the end of a flexible lead coming out of the box, serves to make contact with the shank of the microelectrode. The electrode's tip is inserted into the test solution in the reservoir shown at the bottom of Fig. 4B. This reservoir was made from a 30 ml polyallomer centrifuge tube (Nalgene number 3119-0030), shown on the left of the figure. A 30-gauge platinum wire passes through a small hole in its bottom and serves to make contact with the solution inside. On the outside it is soldered to a small washer, and the whole assembly made waterproof and held together with epoxy cement. The centrifuge tube fits into a metal sleeve, 26 mm i.d. and 80 mm long, set vertically into the face of the box and grounded to it. The bottom of the sleeve stops 1 cm above the foil side of the circuit board, directly above the AD515 input amplifier. Pin 2 of this operational amplifier (the inverting input) passes up through holes in the circuit board and in a small teflon stand off, and is soldered to a washer. When the test solution reservoir is inserted in the metal sleeve, the two washers make contact, effectively connecting the test solution to the input operational amplifier. The circuit requires a -F 15 V power supply and draws about 15 mA. The power supply we used was Model P33, Polytron Devices (Patterson, NJ, U.S.A.). The analog voltmeter has a 0-1.5 V movement (Simpson Electric, Elgin, IL, U.S.A.). A milliammeter could be used instead by substituting an 8.2 kf2 resistor for the 6.8 kf2 in series with the meter. Parts cost, exclusive of power supply, meter and chassis box, was about $70 U.S. Total parts cost was about $160 U.S.

Acknowledgements This work was supported by N.I.H. Grant EY-04186 to R,S.S.

References Holland, W. (1964) The Prope,rtics of Glass Surfaces, Wiley, Now York, 546 pp. Lavallte, M., Schanne, O.F. and H4bort, N.C. (1969) Glass MAcroelectrodos, W ~ , Now York, 446 pp. McCann, F.V, and Stibitz, G.R. (1977) Indirect dotcrmination of microb-qcotrod¢ geomot.,% IEEE Trans: Biomed. Eng., BME-24: 297-300. Naylor, G.R.S. (1978) A simple circuit for automatic continuous recording of microelectrode resistance, Pfliigers Arch., 378: 107-110. Plamondon, R. and Gagnt, S. (1980) On the influence of diffusion, double layer and glass conduction on the electrical resistance of open tip glass microelectrodes, IEEE Trans. Biomed, Eng., BME-27: 260-270. Ruhio, R. and Zubieta, G. (1961) The variation of the electric resistance of mciroelectrodes during the flow of current, Aeta Physiol. Lat. Am., 11: 91-94. Sabah, N.H. (1973) A microelectrode resistance meter, J. Appl. Physiol., 34: 722-723. Schanne, O.F., Lavall6e, M., Laprade, R. and Gagn6, S. (1968) Electrical properties of glass microelectrodes, Proc. IEEE, 56: 1072-1082.