Chemical lnstrunrentation S. 2. LEWIN, N e w York University, Washington Square, N e w York 3, N. Y.
T h i s series of articles presents a sumey of the basic principles, characteristics and lGmitations of those instruments which find important applG
..
.
eouinmenl:, anoro&mate mices are ouoted to indicate the order of rnaonitude of " cost of the various types of design. A
.
pH Meters Continuedl In choosing a pH meter from among the many commercially available models, the following factors should be considered: ( I ) numher of pH determinations to be made per working day, ( 2 ) accuracy reqnirod in tho measurements, (3) nature of the ambient conditions under which measurements will be made, (4) intelligence and technicrtl training of operators of the inst,mment, and (5) importance of adaptability to potentiometrie titrations and continuous recording. Laboratory pH meters currently available in the U. S. range in price from $100 to over $1000, reflecting a wide variation in accuracy, speed, freedom from interferences, degree of multifunctionality, and convenience of manipulation. Eleotrometrie pH measurements accurate to &0.2 pH units are relatively easy to make, and the instrument. used may be simple and inexpensive. However, i t should be appreciated that colored indicrttors and indicator papers are also capable of this kind of accuracy, and are much cheaper than an electronic device. A pH meter of 0.2 pH unit accuracy is preferred over indicators in the monitoring of inaccessible or remote solutions, or in making a large number of measurements in a. short time. Most commercial pH meters find their greatest usefulness in providing messurements with an accuracy of &0.l to 0.01 pH units. For work of this order of accnmey, tho limiting factor is often not the electronic instrument, but the electrochemical system; i.e., the characteristics of the electrodes and the solution in which they are immersed. One important source of error is due to temperature, for not only does the proportionality factor h e twocn emf and pH vary with temperature,. but ionization equilibria and junction potentials also have significant temperature coefficients. Consequontly, the most accurate work ahauld be performed on ~olutionswhose temperature is controlled, or, a t least, measured, and the meter should be calibrated and standardized with buffer solutions maintained at the same temperature. For &0.01 pH accuracy, the temperature should be known to 1 2 ° C .
Another important source of error in pH measurement,^ is thc junction potential existing a t the interface between the referonce (calomel) electrode and the solution. If this junction is not made in a reproducible way each time the electrodes are immersed in s. solution, errors ss large as several tenths of a. pH unit can result. In addition, thc surface of the glass electrnde must be kept clean of any film or sediment. Some commercial pH meters are readable to 0.005 to 0.001 pH unit, and are reproducible to almost this degree. Such measurements may have value in providing relative information, such as the rate of variation of acidity, or the location of an equivalence point, but it is unlikely that the pH in most systems can be given absolute significance beyond the 0.01 level. For a basic disoussion of the prohlem of relating meter readings of voltage to hydrogen ion concentration or activity, see Feldman, I., Anal. Chem.., 28, 1859 (1956). The direcereading pH meters are quicker and easier to use than the nullbalance instruments, hut are generally less accurate. If the amplifier circuits of both types of instruments functioned perfectly and introduced no error?, the accuracy of a null-balance instrument would be limited by tho accuracy of its slidewire; the direct-reading meter would be limited by the accuracy of the meter scale. Slideaire aeeumcies are generally 0.1% or bettor: meter scale accuracies are usually of the order of 1%. I n most laboratory situations, the greater part of the time consumed in making a pH determination consists in the manipulation of the glassware and solutions, including t,ho cleaning of the electrodes between samples. The time far finding the null balance is generally under 30 seconds. Hence, if a number oi different solutions is to be measured, with rinsing of the electrodes between samples, s, direct-reading meter will not represent sn important time saving. On the other hand, a direct-reader is indicated if it is desired to follow the change in pH of a single solution with time conveniently, or to use a continuous recorder to obtain a permanent record. Many people find the direct-reading type oi instrument more pleasant to use because it releases them from the psychological burden of
-afeature
conrentrating on the adjustment of a balancing dial. If the pH meter will be used in a laboretory whcrc there are nearby motors, pumps, stirrers, osrillators, fluoresrent lights, u r other nc devires, special attention must be m i d to adeouste shieldine of the solutions, electrodes, and input circuit. An cqually serious problem is presented by high ambient humidity, which can create leakage paths across the input or in the electrometer tube cirewt. If conditions such ns those mentioned above prevail, the user should make certain that the pH meter he has rhosen is specifically designed for adequate accuracy under such conditions. All pH meters require both slandardi2.ation and calibration imm time to time. Same commercial models are designed to make these operations as simple and frre from the necessity of thought on the part of the operator as possible. Tho importance of this feature in inversely proportional to the intelligence and techniral competence of the instrument operator. Of course, even highly intelligent, competent operatom allcome simplicity of manipulation in an instrument.
Standardization and Calibration The distinction betwren the terms standardization and ealibmtion as they are generally employed in the instruction manuals accompanying pH meters may be appreciated hy 8. consideration of the diagrams in Figures 9 and 10.
Figure 9. Standordirotion of the scale of o potentiometric instrument consists in adjusting the riidewire current until the aduoi voltage gradient lor IR-drop) equals the spacing on the inscribed scale.
The process of standardization in a potentiometer-type circuit consists in (Continued on page A536)
Volume 36, Number 7 0, October 1959
/
A595
Chemical Instrumentation adjusting tho voltage gradient in the slidewire to correspond with the markings on the instrument dial or scale. That is, if the spacing between major divisions of the scale (say, from pH = 6.00 to 7.00) is, e.g., 1 inch, then i t is necessary t h a t each inch of the slidewire have n voltage drop of 50.2 millivolts ( = 2.303 R T I F ) a t 25°C. As shown in Figure 9, this is easily arcomplish~dby means of an adjusting resistor in series with the didewire. In direcGdeflection pH meters, standardization consists in adjusting the amplification factor of the eleetromrt,er tube, the cathode or screen grid voltage, or a shunt across the read-out meter, until the change in deflection of the meter needle for a. given pH change coincide8 aibh the spacing of the markings on the meter face. [The first two parameters mentioned above do not apply to instruments employing high gain negative feedback amplifiers.] Calibration involves adjusting the ahsolute d u e of the voltage across the (standardized) slidewire 80 that the markings of the scale may be read directly in trrms of pH units or millivolts. For example, ns shown in Figure 10, a pair of variable resistors may be connected a t apposite ends of the slidewire, so that they function in inverse relation to each other. As one is adjusted to increase its resistance, exactly the same amount of resistance is removed from the ot,her, keeping the total resistance of the pair constant. Hence, the voltage gradient across the slidewire is not affected by these controls, but t h e magnitude of t,hevoltage at any point may be shifted up or down by this adjustor.
VOLTAGE
1
(~IIIIIIII)IIII)IIIIIIIII[IIII[--8
~1111~1111~1111~1111~111!~1111~---~
INSTRUMENT
Fig. 10. Calibration of a potentiometer didewire by means of ganged variable resirtors t o make the markings on the dial direct-reoding i n true voltage.
In direct-reading pH meters, this type of calibration is sceompli~hedby adjusting the bias volrage of the grid of a vacuum tube, i.e., the potential of this grid relative to its cathode vhen there is no input signal. Testing a pH Meter
A very useful devicc to have in any laboratory that depends upon the reliability of pH meter measurements is a p H
A596
/
Journol of Chemicol Education
Chemical Instrumentation met,er tester. Such n test instrument should bc capable of checking the standardization of t,he p H meter, as wdl ar of measuring the grid current in the input circuit. An appropriate device can be constructed in thc laboratory without great difficulty (aee Bates, R. G., "Eleetrometric p H Determinat,ions," p. 276 ff). A handy commercial unit is a v d n l h from Photovolt Corp., New York 16. N. Y. (Model 25 Tester, $68). Thc rircuit. is shown schematically in Figure- 11. The 110 v ac from the line is stepped-up in voltage by n transformer, rectified, and filtercd (by condenser . T ~ Pgas discharge ("glow") tubes, 0A2 and 5651, have the property of maintaining a constant voltage across their elrctroder, independent of fluctuationsin the sopply voltage, as long as these fluctuations are not, too large. Thus, a rather constant reference voltage is available a c r o s the- output terminals. Thia is made to correspond to
Figure 1 1 . Schematic diagram of the circuit of the Photovolt Corp. Model 25 pH Meter Tester.
the voltage diffcrcncc between s glass electrode in buffers of p H 4 and 7 respeetively. I t will, therefore, also correspond to the difference betwecn p H i and 10. This voltage is applied to t h r p H metcr through a large (250 megohm) resistor, thus simulating the high internal imppdance of the glass electrode. If the p H meter is "zero-ed in" a t p H 7, and the tester is applied, it shordd enusc the meter to rehnlanee a t 4.00 or 10.00, depending on the polarity. This procedure permits one to check the electrical circuitry of the p H meter without, mffering the ambiguity of not bring sure that thr electrodee or Ihffrr solutions are functioning p r o p d y . I t should be noted that the calibration of a high quality p H mcter can he checked within the t,oleranco of the instrument only by using a vokage source that is more (Continued on page A598) Volume 36, Number 10, Ocfober
1959
/ A597
Chemical Instrumentation
standardized with a buffersolution several units away from the test p H .
classic example uf the null-balance electrometer circuit. A simolitied diaernm of
Beckman Instruments, Inc. accurate than that calibration. The accuracy of the reference voltage in the Photovolt Tester is probably not better than &0.02 pH unit. T o measure the grid current, the tester is used as follows. The pH meter is "zero-ed in" with the electrometer tube grid connected directly to it8 cathode. I n most pH meters, this corresponds to the setting of the pH didewire a t 7.00. Then, wibhout changing m y of the dial settings, the input of the pH meter is connected to the tester, which is equivalent to placing a 250-megohm resivtor between grid and cathode of the electrometer tube. Any change in the deflection or balance pomt of the pH meter is due to the flow of grid or leakage current through the 2.50-megohm resistor, and the magnitude of this current can be readily computed from this measurement. With the foregoing as background, we are now ready t,o consider the design of the variorw common commercial lahoratory pH meters. I n the following. most of the data given relative t,o accuracy and reprod~mibilityare taken from manulnctoren' literature, and in some cases may be optimistic. For example, if the met,cr is standardized with a reliable hufier solution having a pH value close to thnt of t,he test solution, accurate results are possible even with an inferior instrument. A high quality, precision meter should be capable of 0.01 pH accuracy even when
A598
/
Journol o f Chemical Education
Reckmen pH meters are more widely distributed in lohorntory use than those
L STD.F q & CELL
Figure 12. Simplified schematic of the circuit of the Beckman Model G pH Meter. The actual circuit contoins tetrodes instead of the triodes shown here, and hor oddition01 reri$torr, switches, ek.
of any other mnkc. The Model G ($475, complete with electrodes) is a.
The current models of this pH meter contain two vacuum tubes: a No. 932, which is the electrometer tuhe, and a No. 031, which amplifies the output of the former. The current flowing through the 032 tube is controlled by the differenae of potential between it,s grid and cathode. The magnitude of the current reaching the plate of this tube determines the potential a t the lower end of R , (because of thc IR-drop in that resistor), and since this point is connected to the grid of the 831 tube through the grid bias battery, E,, its potential is added to the total bias on that grid. This is an example of s single-stage, direct-couplrd amplifier. Thus, the deflection of the meter needlo is a measure of the potential on the grid of the clcetromeber tube. The potential on the grid of the 932 tube is composed of (1) the potentisl difference between the sensing and reference deetrodes in the ~olutionunder study, plus (2) the potential ditiermce between the sliding contsct on the " p H slidesire and that on the "zero ndjust" slidemire, plus ( 3 )the potential of t h r diding contact on the "No. 1" variable resist,or. The role of each of the variable adjustments can best he appreciated in terms of the sequence of operations thnt are carried out in the corme of a pH measurement. A . With t,he snitches S, S' in the "I"
(Continued on page 9600)
Chemical lnstrumentation position, the metcr nefdlc is brought to "0" (which is not zero rurrent, hut rathcr an a r h ~ t r a r y rdprence current) hg adjustment of the No. 1 rontrol. Thia varies t h r potential of the cnthorles of t,he vacuum tubes and permits the instrument to be adjusted to compensate for aging of the vacuum tuhrs and hattcrics. B. The temperature adjust is then set in necortlnnee with n sralc mnrkcd in trrms of the temperature of t h r rleetrodrs. This inserts a known resistanre in series with the pH slidewire which is of thp propcr magnitude to limit the current to a. value that will givc the correct v o h g e gradient lor thc tcmperatnrr involved, as governed by tht! Nernst eqoation. This is, thrreforr, a form of standardization control. C. With the switches S and S' in the "2" position, the S o . 2 control is adjusted until the meter needle comes to its "zero," or refermce point. This placer the st,.zndnrd cell in series with the slidewire, and attaches the electrometer cathode to one side of the standard cell, and the grid to one sidc oi the slidesire. This permits adjustment of the slidewirr current to a constant reference value, by wing the electrometer tube grid-to-cathode patmtial ns the indicator. Actually, this adjustment regulates t,he current through the "scro adjust?' slidewire, lrut since the "pH" slide~rireis in parallel, its current, is also aut,ometirxlly regulated in the DIOPPSS.
.4t this point, the instrument ha8 bcrn standardized ( ~ fFig. . 9). I t n o v remains to perform the ealibrat,ion. D. 1 5 t h the a d c h e s in tho "1" position, and the electrode pair immersed in ;t buffer solution of known pH, the pH dial is set to the pH of the bufier, and the push-hntton switEh is depresa~d. The metcr needle is brought t,o its refercnee point by means oi thc "zrm adjust" control. Dcpresaing thp push-button switch disconnects the grid of t,hc rlectrompter tuhe from its cathode, and ronnccts it instead to t,he sensing electrode. Vnrring t,he "zero adjust" changes the potential of the eleetromet,er tuhe cathode. Bringing the meter needle to "zero" with the electrodes in tho buRcr is equivalent t o displacing t,he voltagc scale up or d a r n , ns in Figure 10, until the voltage of each point on the pH slidenire mineides with the dial markings. The instrument is now both rt,and:~rdiacd and cnlibmted, and may b~ used t o determine pH'a of unknown solntions. Thc Model G pH meter is rqxoducihle to +0.01 pH unit over the range 0 to 13, and can he used ss a. millivoltm~berwith a. range of 1300 mv. The grid current is lrsa than 2 X 10-12amperes. A modification of this device is available RS the Model GS (X625, complete v-ith electrodes). This meter is designed t o provide ~ensit,ivityto +0.00?5 pH units. I t contains the same cirrrdt as the Model G, v i t h two exceptions. (1) The sero adjust resistor is a ten-turn "Hclipot," which e m be set with 10 times greater precision than a single-turn variable resis-
(Cmlinued on page AAOS)
A600
/
Journal of Chemical
Education
Chemical instrumentation tor. ( 8 ) The ad-out meter is approsimately 20 times more sen~itivethan the one used in the Model G. I n order to permit the instrument to have the wide range of the Model G, coupled with a much higher sensitivity over a narrower range, t,he read-out meter is connected in the manner shown in Figure 13. When the selector switch is in position A, the meter is shunted by the resistor RL,which reduces its sensitivity to that of the Model G's meter. When the seleotor snitch is in position B, the meter is being used a t full sensitivity, and is "bucked" by the battcry E,,,, whieh has the effect of esneelling out the major psrt of the current flowing through it. This arrangement makes it possible to use a highly sensitive meter to show up small
fluctuations in a large current without running the risk of driving the meter needle off scale or burning out the meter coil. The bucking current essentially displaces the origin of coordinates of the meter to a point to the left of the scale zero. When the instrument is being used a t maximum aensitivity, the meter is brought to it8 "zero" position by using the pH slidewire as a. coarse control, and the zero adjust as a fine eontrol. The latter settings have no absolute significance, but are used to pmvido information about relative pH's. All the other Beckman pH mctcrs are direct-deflection-type instruments. The Model H-2 is a. line-operated electrometeramplifier with feedback stahilieation. This model ia no longer in production, having been replaced by the Zeromatic; however, since man." of these instruments are still in oso, thc d e ~ i g nfeatures will be
discussed here. A simpliSed version of the circuitry is shown in Figure 14. The sensing electrode is connected to the grid of the 032 electrometer, and controls the current through the tube. This affects the potential of tho plate (point A ) , whieh is direct-coupled to the grid of the next stage of amplification, the 1U5 tube. This in turn is direct-coupled to the LiAKG, RO that the current flowing t o its plate (point C) is a function of the original input voltage, and the ZR-drop through the 27,000-ohm plate load resistor is a measure of that input. The voltage diffcrenre between C m d D causes the met,er deflection.
Figure 13. Read-out meter arrangement of the Beckman Model GS p H Meter. In position A, the meter is shunted b y R, and har low renritiv. ity. In position 8, Em bucks out the mtlior portion of the current, and meter rhowr up rmoll Ruduotion%with high sensitivity.
The feedback stabilization is pmdueed by the psrt of the circuit involving the resistance from E to 0. To understand its role, consider that a fluctuation has
Figure 14. Basic circuit of the Beckmon Model H-2 line-operated p H meter. The regulated power supply section is not shown; tubes repre. rented ar triodes in the diogrom are tetmder or pentodes in the inrtrument.
occurred in the power supply so that the voltage across the circuit has decreased. This lowers the potential at A (since it is (Continued on page A608)
A602
/
Journal o f Chemical Education
Chemical lnstrumentution a point on a voltage divider het,woen the low and high voltage parts of the poa-rr supply), ~ ~ h i makes ch the grid of the 1U5 more negative than it, had hpen, decreasing that tuhe's current.. This raises the potent,inl at point H (because of the smaller IR-drop in tho plate resistor), making the grid of the 6.41'3 more positive, increasing the tube current, and thereby decreasing the potential a t C. This last cffcct lowers the potential a t point E. Since the r r f c r ~ n e relectrode is connected to E, and the grid potcntial of t,he 932 tube differs from this by s fixed amount (determined by the rhemical system), this chain of events has the ~ f f e cof t making t,he gl.id of the 932 become more negative than it had bcen relative to its cathode. This would tend to decrease the tube current, raising the potent,ial a t point A (brrar~seof the dccreamd 112-drop in the 44mcgohm resistor), and thereby opposinf the o+nd &el of the assumed voltage flurtuation. Hence, the inverse icedhack present here stabilizes the output against vnrintions in circuit characteristics. The fredback can account effectively for variations occurring in the circuit hctween the input grid and the read-uut meter, hot not for drifts in the grid tias itself. The latter (as well as variations in the asymmetry potcntial of the electrode pnir) arc adinsted far by variahle rcsis tors, in a manner similar to that, described in rom~ertion with the Model G. T o cherk the grid bias, the meter is set to give x deflertion corresponding to the pH of a known hnffer with the electrodes immersed in the solution. Then, the elwtrodes nrc ~ a i t c h e dout of the circuit and, with n fixed voltage on thc electrometer tohe grid, the deflertion of the meter needle is noted. This is called the chwk-point. For convenience, an auxiliary pointer is mounted in the meter case, so that it ran he movod by hand to this point as a visual reminder of its value. Subsequently, during use of the instrument, the electrometer screen grid potential is adjusted from time to time to keep the meter needle a t the check-point value a-henever the ~lectrodes3R switched out. This procedure obviates the need for checking the instrument frerluent,ly with I~uffersolutions. The blodel H-2 has a reproducibility of +0.03 pH units, and may he used as n millivoltn~eterwith a. range of 1400 mv. A specinl circuit is provided so that n polarizing cmrent of about 10 mieroamperes can be applied t o the electrod~s for use in "dead-stop" titrxtions. The grid rnrrcnt is less than 2 X 10-12 amperes. An ingenious modification of the Model H-2 that rliminat,es the neremity ior t,hc msnnnl eompnrisan of the inst,rument with it,s check-point by carrying out this cnlihrntion automatically every second is the Zeromatic pH Meter ($295, complete with electrodes). The input circuit, of this instrument is shown schematically in Figure 15. When the relay is off, the voltage difference existing between the two electrodes is impressed on the electrometer tube through a condenser, which (Continued on page A814)
A608
/
Journal o f Chemical Education
Chemical Instrumentation charges up to a voltage determined by the output of the amplifier. When the relay pulls in, the electrodes are disconnected, and the pH-measuring circuit is connected to a so-oslled "Corrector Amplifier," through the same condenser. Now, the oondenser charges up to s voltage characteristic of the reference state of the pH amplifier. This is, therefore, x measure of the "null" output of the amplifier, just as the check-point was in the Modd H-2. When the relay drops out again, the input from the eloct,rodes is now impressed on the electrometer tuhe through the eondenser which had been charged to the output voltage of the corrector amplifier. By causing these voltages to oppose each other, the net input is made equal to the voltage of the electrode pair, and the oubput of the instrument has been correeled for the amplifier null. Since the relay dperates once each accond, the amplifier characteristics need bc stable only for one second a t a time, in order for this correction scheme to '~.ork.
I
TOCORRECTOR AMPLIFIER
MEMORY
L -
-
.
I
I
7
Figure 15. Input circuit of the Beckmon Zeromatic pH meter, rimplifled.
The Zeromatie is reproducible to +0.02 pH units, accurate t,o 0.1 unit, has a, range in pH of 0 to 14 and in volts of 1400 mv. The input current is less than 1 X lo-'$ amperes. It is designed for application also to polarized electrode measurements, and has terminals for ilttachement to a ~otentiometricrecorder. Tho front panel controls are push-button operated. The Beckman Model N-1 ($310, oomplete with electrodes) is a battery-operated, direct-reading, electromet,er-amplifier instrument designed for compnctnem and portability. Reproducibility is +0.02 pH units, accuracy is 0.05, and range is 0 to 14 in pH and 0 to +420 millivolts. A cheek-pointer is used in maintaining the reference condition of the amplilinr circuit. Modcl N 3 ($360) is the same electrically, but has a convenient carrying case dcsigned for field use. The Pocket pH Meter (590.50, complete with electrodes) is a miniaturized version of the battery-operat,cd, direct-reading pH meter, weighing only 2 pounds. I t is readable to 0.1 wH unit:, renroducibilitv . and accuracy are somewhat poorer than this. Thc design is bnsicxlly a simple elcetrometer circuit.
.Ve.rt: Conclusion of the survey of conrmereid 1at.oratory pH meters.
A614 / Journal o f Chemical Educofion