The ionization constant of boric acid was determined by the use of the electromotive force of cells without liquid junctions. Hydrogen and silver-silverchloride electrodes were immersed in borax-sodium chloride solutions. To reduce tbe possibility of formation of polyborates, dilute solutions were used. The emf were measured at 5-degree intervals over the temperature range 0° to 60° C.The negative common logarithm of the ionization constant (pK) of boric acid over the temperature range 0° to 60° C may be represented by the equation pK=2237.94/T+0.016883T-3.305,where T is the absolute temperature.The data indicate that the mean s.ctivity coefficients of the ions of sodium chloride and of sodium borate do not differ appreciably. The pH values for the solutions studied and for rounded values of the concentration are tabulated as functions of temperature and ionic strength. These solutions, which range in pH values from 8.934 to 9.465, can be used as standards in the calibration of glass-calomel and other electrometric pH equipment.A discussion is given of the significance of the quantity a i (commonly called the "distance of closest approach" of the ions) and its importance in the calculation of pH value~, especially when the buffer ion is not univalent.The changes in free energy, beat content, entropy, and beat capacity that accompany the ionization of 1 mole of boric acid are listed. CONTENTS Page

By application of the equation pH=(E-Ere,-Ej) /k to solutions whose pH values were known accurately, the sum of the potentials of the reference electrode and of the liquid-junction potential, E",+ E j , was obtained at 25° C by the method of cells with liquid junction for seven solutions suitable for standards of pH 38. Silversilver-chloride electrodes immersed in saturated potassium chloride solution were used rather than the calomel electrodes customarily employed.As E", remains constant when the buffer is changed, values for the df fferencel! in the liquid-junction potentials of various buffers in contact with saturated potassium chloride solution were obtained from the data. These differences were then used to calibrate seven Type 015 and three "low-alkali error" glass electrodes of commercial manufacture. The average agreement between the true pH of the buffer-chloride solution (determined from cells without liquid junctions) and that read on various commercial pH meters when corrected for the difference in the liquid-junction potentials and the alkali error of the electrode was ±0.01 pH unit. The data also furnish a critical test of the consistency of the pH values assigned to the various buffer solutions recommended by this Bureau for the calibration of the pH scale and for checking pH meters.Recommendations are made for checking pH meters.

No abstract