EQUIPMENT AND SUPPLIES--Hydrolab #4041 Digital Meter In Reply Refer To: September 30, 1982 EGS-Mail Stop 412 QUALITY OF WATER BRANCH TECHNICAL MEMORANDUM NO. 82.17 Subject: EQUIPMENT AND SUPPLIES--Hydrolab #4041 Digital Meter The Hydrolab #4041 reference electrode problem reported in Quality of Water Branch Technical Memorandum 82.14 has been resolved. Careful study, coupled with extensive testing by both the Hydrolab Corporation and several U.S. Geological Survey offices, confirms that recommendations made by Hydrolab provide an acceptable solution. The problem was most pronounced when pH values below 5 were measured in water with electrical conductivity less than 200 umhos. The #4041 values were lower than values obtained using instruments marketed by other manufacturers because the ionic strength of the filling solution in the Hydrolab reference electrode was too low. The conditions a filling solution should meet to insure a proper liquid junction may be reviewed by reading Chapter 3.2, Reference Electrodes, in pH Measurements by C. Clark Westcott (Academic Press, 1978), particularly pages 61-63. References listed in the attachments prepared by the Hydrolab Corporation are recommended also. Attachments 1 and 2, prepared by Hydrolab, clearly explain the reference electrode problem and the corrective measures needed. Several WRD offices have followed the Hydrolab recommendations and were able to achieve comparable measurements using several different brands of pH meters. Selected results from measurements of standards and stream water are presented in the following table: District Specific conductance A B C D (umhos)/pH 65 202 212 70 29 32 52 Hydrolab 7.5 8.0 8.1 7.2 4.2 6.4 4.0 Sargent-Welch 7.6 7.7 8.0 4.1 Orion 7.0* L & N # 1 6.2 3.9 L & N # 2 6.2 3.9 *drifting The attachments are commended to careful study. Preparation of the filling solution, cleaning and refilling the cavity of the reference electrode, and calibration of the #4041 should be conducted in the District Service Unit. After checking responses to standard solutions, comparative pH measurements using different meters and low conductivity samples with pH less than 5 should be made. As Hydrolab points out, cleanliness of the conductivity electrodes should be assured, and the DO sensor, which is part of the pH measuring circuit, must be immersed during calibration and sample measurement. Attachment number one suggests ways to correct pH measurements that have been made with a Hydrolab #4041. Correction of pH values is not encouraged because (a) repetitive measurements may have been made with different instruments, (b) corrections are empirical, and (c) there are other variables that are difficult to account for with adequate assurance. Please ensure that each person who uses a Hydrolab #4041 has the opportunity to study this memorandum and attachments, and that the information becomes part of the the District "Field Handbook." Questions may be addressed to Herman Feltz, who is now in the Quality of Water Branch, and can be reached by phoning FTS 928-6834. Once each #4041 instrument has been recalibrated following the recommended procedure, pH measurement restrictions outlined in Quality of Water Branch Technical Memorandum No. 82.14 no longer apply. R. J. Pickering, Chief Quality of Water Branch (2) Attachments WRD Distribution: A, B, S, FO, PO Key Words: water quality, instrumentation, quality assurance This memorandum supersedes Quality of Water Branch Technical Memorandum 82.14 Attachment 1 A NEW FILLING SOLUTION FOR MINIMIZING LOWQCONDUCTIVITY ERRORS IN HYDROLAB (TM) pH REFERENCE ELECTRODES HYDROLAB CORPORATION AUSTIN, TEXAS 24 AUGUST 1982 HYDR0LA8 CORPORATION A NEW FILLING SOLUTION FOR MINIMIZING LOW-CONDUCTIVITY ERRORS IN HYDROLAB pH REFERENCE ELECTRODES It has been noticed that Hydrolab pH instruments report pH values that are consistently too low when measurements are made in the low-conductivity, poorly-buffered waters peculiar to certain geologic regions. As can be easily demonstrated, these measuring errors are caused by the particular filling solution that has been used in the pH reference electrode. A new filling solution has been formulated which virtually eliminates the low-conductivity error without affectind the stability of the reference electrode to the large change~ in pressure and temperature to which it is normally subjected. Moreover, the low-conductivity error appears to be sufficiently reproducible to allow errors incurred using the old filling solution to be corrected in most cases. THE REFERENCE ELECTRODE PROBLEM Thc general character of the low-conductivity error can bc seen in Fig. 1. Data for Fig. 1 was collected by measuring the pH of a deionized water sample as its conductivity was increased in steps by adding small amounts of concentrated KCl solution. The reference electrode contained the old pH 7 buffer filling solution. The pH error clearly varies as the logarithm of solution conductivity (straight line data) until the conductivity reaches about 2000 micromho/cm, where the error virtually vanishes. This behavior makes the case for liquid junction potential variation as the error mechanism. Rough calculation of junction potential as a function of conductivity (using Henderson's equation) for the case described confirms that the equivalent pH error should be expected to be negative, should vary with the logarithm of conductivity, and should decrease rapidly above some threshhold conductivity. The liquid junction potential referred to here is the voltage that appears across the porous plug by which the reference electrode makes contact with the sample. The voltage arises from the outward diffusion of the filling solution and the inward diffusion of the sample. Polarity and magnitude of the voltage depend upon the compositions of the two solutions. pH measuring circuits can not distinguish between changes in the potential of the pH electrode and changes in the potential of the reference electrode. Hence, changes in the junction potential of the reference electrode are reported, erroneously, as pH changes. This source of error can be minimized by using the correct filling solution. Further calculations using Henderson's equation indicated that substituting a buffered, concentrated KCl solution for the old buffer-only filling solution should eliminate the low-conductivity error by making the junction potential, for all practical purposes, independent of the properties of the sample. As will be demonstrated, that is indeed the case. BUFFERED POTASSIUM CHLORIDE FILLING SOLUTION The new filling solution should be prepared as follows: 1. Dissolve 225 grams plus or minus 5% of KCl in a liter of deionized water to obtain an approximately 3 molar solution. 2. Using a portion of this solution instead of deionized water, prepare a 7-buffer solution in the (otherwise) usual way. The dramatic improvement in the low-conductivity performance of the Hydrolab reference electrode using the new filling solution is shown in Fig. 2. Data for Fig. 2 was gathered by measuring the pH of a carbon dioxide buffer solution (see An Easily Prepared Low-Conductivity Acid pH Buffer, Hydrolab Corporation, August 1982) as its conductivity was increased in steps by addition of KCl. A further test of the new filling solution was made by taking simultaneous pH readings with a Hydrolab instrument and an Orion system comprising a Ross electrode pair and a Model 601A Orion Instruments pH meter. The test solution, in that case, was a deionized water sample which was gradually acidified with bubbling carbon dioxide. Comparative results are given in Table 1. Finally, Fig. 3 illustrates the behavior of the two filling solutions in non-Hydrolab equipment. These data were taken using an Orion Instruments 601A pH meter, a standard pH electrode, and an Orion sleeve-type reference electrode containing first the old and then the new filling solution. CORRECTING OLD pH DATA pH readings taken using the old filling solution can be corrected by either of two methods. The quickest is to read from Fig. 1 the pH error corresponding to the conductivity at which the original reading was made. Alternatively, a correction table can be constructed by measuring the conductivity and pH of water samples typical of the region, using first the old filling solution and then the new.(Note: re-standardize when filling solutions are changed.) Corrections by either method should generally agree. PERFORMANCE TESTING To avoid the "which one is right?" problem when comparing pH instrument readings at very low conductivities, it is strongly recommended that the carbon dioxide buffer be used instead of makeshift test solutions. The C02 buffer is reproducible, has a pH of 3.91, and a conductivity (when freshly prepared) of less than 50 micromho per centimeter. TABLE 1 TIME CONDUCTIVITY ORION pR HYDROLAB pH (min.) (micromho/cm 0 1 5.66 5.56 0.25 10 4.88 4.84 0.5 14 4.60 4.57 0.75 17 4.51 4.50 1 19 4.43 4.41 1.5 23 4.36 4.33 2 25 4.30 4.27 3 30 4.24 4.22 4 33 4.15 4.16 5 38 4.10 4.10 10 44 4.05 4.04 20 52 3.99 3.98 30 56 3.97 3.96 40 56 3.98 3.96 50 61 3.96 3.94 60 61 3.94 3.93 120 66 3.93 3.93 REFERENCES 1. R. G. Bates, Determination of pH, Wiley, 1973. 2. J. A. Dean, Lange's Handbook of Chemistry, Twelfth Ed., McGraw Hill, 1979. 3. E. V. Condon and H. Odishaw, Handbook of Physics, Second Ed., McGraw Hill, 1967. 4. A. Lerman, Gecchemical Processes Water and Sediment Environments, Wiley, 1979. 5. W. B. Guenther, Chemical Eouilibrium, Plenum, 1975. 6. D. T. Sawyer and J. L. Roberts, Jr., Experimental Electrochemistry for Chemists, Wiley, 19~4. 7. R. A. Horne, Marine Chemistry, Wiley, 1969. AN EASILY PREPARED LOW CONDUCTIVITY (LOW IONIC STRENGTH) ACID pH BUFFER HYDROLAB CORPORATION AUSTIN, TEXAS 20 AUGUST 1982 As an aid to predicting the performance of pH instruments in acid rain studies and the like, the availability of a reliable pH benchmark solution of very low ionic strength is essential. Ordinary pH buffers are of little help because their high conductivities (over 5000 micromhos/cm) make them too unlike the highly dilute solutions to be studied. Millimolar solutions of strong acids have been tried but the initial pH of these depends upon the skill of the preparer and, since they are unbuffered, their pH's tend to be easily disturbed. We have found that dissolved carbon dioxide in equilibrium with one atmosphere of carbon dioxide gas makes a nearly ideal test solution. It has a defined pH (3.91), appreciable buffering capacity, and a low conductivity (50 micromho/cm). And, since no volumetric or qravimetric steps are necessary, the solution is very easy to make. CARBON DIOXIDE BUFFER PREPARATION All that is needed to make up the buffer is a bottle or tank of carbon dioxide gas (with regulator) and a suitable volume of deionized water (l liter, say). Using a length of tubing, bubble the carbon dioxide through the water at a few liters per minute until the pH reaches its equilibrium value of 3.91 at standard temperature and pressure. Twenty minutes or more is generally required for equilibration. Be sure that bubbling of the solution is continued as long as the solution is being used. The performance of the buffer is predicated upon continuous contact with the gas phase. CARBON DIOXIDE BUFFER PROPERTIES The carbon dioxide buffer depends upon the two reactions (l) and (2) Combine (l) and (2) to give (3) and, since (4) (5) Because ionic strength is low, we can have pH = - log H so from (5) (6) (7) The C02 (gas) factor in brackets is the partial pressure of C02 in atmospheres, so at standard pressure and temperature, the log term vanishes giving 3.91 for the standard pH of the buffer. Temperature Effects: The temperature coefficient of the carbon dioxide buffer is respectably small because of two offsetting effects: equilibrium (1) shifts to the left (Ks gets smaller) with increasing temperature while equilibrium (2) shifts to the right (Kl gets larger). The table shows the standard pH at several temperatures. !C pH 0 3.85 5 3.86 20 3.87 15 3.88 20 3.89 25 3.91 30 3.92 Atmospheric Pressure: The effect of altitude upon the buffer pH, while significant, is relatively modest. For example, the pH of a buffer prepared in Denver would differ from the standard by just about + 0.04 units (contribution of log term of (7)). Absol. Atmos. Pressure pH mm Hg 760 3.91 725 3.92 700 3.93 675 3.94 650 3.94 625 3.95 Dilution: In reasonable quantities, dilution causes only a transient pH disturbance since the continued contact with the gas phase will restore the original equilibria and the original pH. A 20 percent (vol) addition of water should elevate the pH to perhaps 4, but only temporarily. Conductivity: A freshly prepared carbon dioxide buffer should have a conductivity in the neighborhood of 50 micromhos/cm. If the conductivity is to be kept as lcw as possible, some care must be taken since the addition of any electrolyte will add conductivity. The diffusion of salt from a pH reference electrode can, depending upon electrode type and buffer volume, cause a noticeable upward creep in conductivity. To avoid the carryover of buffer salts and other contaminants, always wash all electrode assemblies in deionized water before they are placed in the carbon dioxide solution. Acids and BaseY: Although not a strong buffer, the carbon dioxide solution does oppose the tendency of small amounts of acid or base to change its pH. If base is added, depleting H plus, (2) shifts to the right, forming more H+. If H+ (acid) is added, (2) shifts to the left, reducing H+ . Quantitatively, adding 0.025 millimole per liter of base ~acid) would raise (lower) the buffer pH by something llke 0.1 pH. Salt Effect: In some situations it may be desirable to deliberately increase the conductivity (ionic strength) of the carbon dioxide buffer -- to observe the behavior of some instrument as a function of that parameter, perhaps. Adding potassium chloride to tne buffer has virtually no effect upon pH at concentrations up to 0.1 mole per liter. At higher salt concentrations, there is reason to believe that pH decreases, but probably by less than 0.1 unit.