1708
J. E’. COETZEE AND G. R. PADMARAHWAN
Vol. 66
PROPERTIES OF BASES I N ACETONITRILE AS SOLVEXT. 11. THE AUTOPROTOLYSIS COSSTANT OF ACETONITRILE BY J. F. COETZEE~ AXD G. R. PADMAXABHAX Department of Chemistry, University of Pittsburgh, Pittsburgh 13, Pennsylvania Received April 6, 1962
A conventional general purpose glass electrode responds reversibly to hydrogen ion activity in picric acid-tetraethylammonium picrate and 1,3-diphenylguanidine-diphenylguanidinium perchlorate buffers in acetonitrile as solvent. From glass electrode measurements in these buffers, and from the dissociation constants of picric acid and 1,3-diphenylguanidine in acetonitrile, determined before by conductometric and spectrophotometric methods, the autoprotolysis constant of acetonitrilc is found to be equal to 3 X It is shown that a conventional agar-potassium chloride salt bridge is not suitable for precise potential measurements in acetonitrile.
Introduction
fresh phosphorus pentoxide. Method B consisted of preliminary drying, first with silica gel, and then wit,h calcium hydride, followed by two fractional dist,illations, first from differentiating solvent for acid-base titrations. phosphorus pentoxide, and then from calcium hydride. The main cause of the strongly differentiating Method C was developed to remove traces of unsaturated from the solvent, and involved, successively, renature of acetonitrile is the fact that the basic and nitriles fluxing with a small amount of aqueous 1% potassium hyparticularly the acidic2 properties of this solvent droxide solution (1m1.b. of solvent), fractional distillation, are very weak. The effect of the magnitude of its drying with calcium hydride, and then two fractional distildielectric constant (36.0) will be considered later. lations, one from phosphorus pentoxide and finally one calcium hydride. In view of the numerous empirical and several from The water content of acetonitrile prepared by method -4 theoretical acid-base studies that have already is generally between 1 and 2 mM, and that from methods been carried out in acetonitrile, the value of the B and C below 1 mM. The relative advantages of theae autoprotolysis constant of this solvent mould be three methods will be discussed elsewhere. In the present, batches of solvent prepared by these three methods of considerable interest. However, a reliable study, gave virtually identical results. value of this constant has not yet been reported. Buffer Components and Other Chemicals.-Baker and In a previous communication2 it was pointed Adamson picric acid was recrystallized twice from acetone out that several investigators had encountered the and dried a t 80” in a vacuum oven. Eastmari White 1,3-diphenylguanidine was recrgst,allized twice from complication that conventional hydrogen ion indi- Label toluene and dried a t 40’ in a vacuum oven. Tet,raet~hylcator electrodes did not respond reversibly in solu- ammonium picrate was prepared by titrating a 0.05-M tions of acids in “anhydrous” acetonitrile as sol- (saturated) aqueous solution of picric acid with a 10% vent. This limitation constituted a serious draw- aqueous solution of tetraethylamnionium hydroxide (Eastto just beyond the equivalence point (detected with back for exact, quantitative acid-base studies in aman) glass electrode), followed by evapodion until crystalacetonitrile, and such studies had to be carried out lization occurred. The product was recrystallized twice, almost entirely by means of c o n d u c t ~ m e t r i c ~ ,first ~ from water and then from 95% ethanol, and finally and spectrophotometric3 methods. However, was dried a t room temperature over phosphorus pentoxide a vacuum desiccator. Tetraethylammonium perchlorate several authors (notably Hall4) obtained results of in was prepared as described else~here.5 Ferchloric acid a semiquantitative nature by using the glass elec- Eolutions were prepared by dissolving Baker and Adamson trode, usually in conjunction with the aquzous ( 0 % aqueous perchloric acid in acetonitrilc. Perchloric saturated calomel electrode as reference electrode. and picric acid and diphcnylguanidinc solutionti always used iinniediatcly aftcr preparation. I n this communication we report that the re- mere Potential Measurements.~--E:lectromotivc Eorw ni(murcsponse of a conventional, “general purpose” type ments with low resistance cells were made with a I,t>edsanti glass electrode is reversible both in picric acid- Sorthrup volt potentiometer (Cat. No. 8G87), and Kith a tetraethylammonium picrate and in 1,3-diphenyl- Beckman Model G pH meter for cells containing glass clectrodes. Several Beckman “general purpose” No. 1190-80 guanidine-diphenylguanidinium perchlorate buf- glass electrodes were used, both with and Kithout “condifers in anhydrous acetonitrile as sol\-ent. From tioning” in acetonitrile for several months, which seemed glass electrode measurements in these two types to have little effect. A silver elect’rode was prepared by of buffer solutions, in conjunction with the results silver plating a platinum wire electrode in 0.05 aqueous argentocyanide solution (free of excess cyanide). of conductometric studies of solutions of 1,3- potassium An H-type cell was constructed, with two IO-mm. didiphenylguanidine2 and conductometric as m7ell ameter fritted glass disks of fine porosity inserted 5 em. as spectrophotometric studies of picric acid apart in the horizontal (salt bridge) sect,ion of the cell. s o l ~ t i o n s ,the ~ autoprotolysis constant of aceto- Between the two disks a vertical inlet-outlet, tube was provided for the introduction or removal of salt bridge solution. nitrile is calculated. All three sections of the cell were stoppered Lo avoid abeorption of moisture and carbon dioxide. Experimental Resistance Measurements.-Resistance measurements Purification of the Solvent.-Sohio acetonitrile was puri- were carried out with an Industrial Instruments Inc. Model fied by 3 different methods. Method A involved prelimi- RC-16B2 conductance bridge operated at a frequency 01 nary drying, first with silica gel, and then with phosphorus 1000 cycles. pentoxide, followed by two fractional distillations from
It is well known that acetonitrile is an cxccllent
(1) Addres- all correspondence t o J. F Coetzee. (2) Pait I of the series. W S hluney and J. F. Coetzee, J . Phgs. Chcm 66, 89 (1062). ( 3 ) I. 11. Holthoff, S. Bluekenstein, and R I IC. Cbantooni, I t , .J 4m Chem. SOC 83, 3927 (1961). (4) 15. I< Hall, Jr , .I. Phus. Chem., 60, 63 (19%).
Results and Discussion The Reference Electrode.-In preliminary experiments, an aqueous saturated calomel electrode (5)
I. 31. Kolthoff and J. F. Coeteee, J . Am. Chem.
(1957).
SOC.,7 9 , 870
AUTOPROTOLYSIS CONSTAXT OF ACETONITRILE
Sept., 1962
(s.c.e.) was used as the working reference electrode, together with a device (described before for voltammetric measurements in acetonitrile)5 intended to prevent, accidental introduction of water into the buffer solutions. This device consisted of a conventional aqueous potassium chloride-agar salt bridge dipping into a 0.1-M solution of tetraethylammonium perchlorate in acetonitrile contained in a fritted sealing tube, which in turn dipped into the buffer solution under investigation. However, i’n cells in which this assembly served as salt bridge, the glass electrode appeared to give drifting potentials. The cause of the drift was traced to the interface between the agar salt bridge and the acetonitrile solution, as shown by the follolving two experiments. For the first experiment, one working (vertical) compartment of the H-cell described above contained an Ag,’(0.01 AgXOs in acetonitrile) electrode as reference (described before by Pleskov6). The remaining two compartments contained a 0.1-X solution of tetraethylammonium perchlorate in acetonitrile, with the agar-potassium chloride salt bridge of an s.c.e. dipping into the second working compartment. The e.m.f. of this cell increased with time. Representative e.m.f. and resistance data for several series of experiments are presented in Table I. It is evident that TABLE I POTEhTIAL OF AQUEOUS SATURATED CALOMEL ELECTRODE vs. PLEGKOV REFERESCE ELECTRODE IK ACETOh’PAREsT
SITRILE Expt. Rb
Expt A‘ Time, niin.
--E, my.
R, ohms X 0.1
-E, mv.
R, ohms X 0.1
Expt. Cc --E, R,ohms mv. X 0.1
291 0 465 293 5 435 297 7 425 291 0 467 293 5 440 299 5 428 294 0 474 293 3 442 300 0 435 302 3 489 293 3 450 302 0 450 311 5 550 297 0 468 310 0 473 10 319 0 598 304 4 483 323 5 525 IS 321 3 ti35 309 7 505 329 1 550 15 S22 5 660 329 1 550 22 d24 3 ti65 30 8235 (i95 35 328 5 710 40 333 2 710 For description of cell, see text. bAgar-KC1 salt bridge removed, clipped into aqueous saturated KCl solution for 5 min., rinsed first with water, then with acetonitrile, and then returned to the cell for a repetition of experiment A. Same as for experiment B, but in addition 0.1 34 HeO added to cell compartment receiving agar bridge. 0 2 4 6 8
the increase in e.m.f. is caused neither by a change in the Pleskov reference electrode (since all three experiments were carried out without changing the silver nitrate solution), nor by the effect of water extracted by the acetonitrile (since addition of a large amount of mater had a comparatively slight effect). I t is seen that the increase in e.m.f. is accompanied by an increase in cell resistance. \Ye attribute the drift in e.m.f. to gradual deposition of a plug of solid potassium chloride and/or dehydrated agar at the tip of the salt bridge. Formation of such a plug, creating a constrained diffusion (6) V. A. It’leskoi , Zh. Fiz K h m , 2 2 , 351 (1948).
1709
boundary, could change the liquid junction potential appreciably. A second experiment was carried out, in which all three compartments of the H cell contained a 0.1-,lf solution of tetraethylammonium perchlorate in acetonitrile, with the agar-potassium chloride salt bridges of two saturated calomel electrodes dipping into the two working compartments. After measuring the cell e.m.f. (which was virtually zero), one salt bridge was left in position, while the other was reconditioned by the series of operations described for experimeiit B in Table I. After returning the reconditioned salt bridge, the cell e.m.f. was remeasured. The above sequence was repeated a number of times. Once again, it was found that the potential of the s.c.e. with the stationary salt bridge gradually became more negative, with a total drift of approximately 40 mv. in 30 min. The aqueous saturated calomel electrode has been used extensively as a coiiveiiient working reference electrode for non-aqueous voltammetry, and its more general use in such studies has been a d ~ o c a t e d . ~The use of the s.c.e. for such purposes is based on the implicit assumption that the liquid junction potential introduced will not change significantly during a given experiment, and will be reproducible (at least to within =t10 mv.) from one experiment to another in the same solvent. It mas recognized before6 that an aqueous agar-potassium chloride salt bridge should not be kept immersed unduly long in a solvent such as acetonitrile, because formation of a plug would increase the cell resistance and therefore the “iR-drop” in voltammetric measurements. To this must now be added that unless such measurements are completed w r y quickly (within 5 or 6 min.), potential values may be in error by as much as 0.04 v. or more, even after proper correction for the increased iR-drop in the ce1L8 It can be concluded that eveii for those potential measurements for which an accuracy of & l o mv. will suffice (e.q., the majority of voltammetric studies), an aqueous agar-potassium chloride salt bridge should be used in acetonitrile with extreme caution. For more accurate potentiometric studies, such a salt bridge definitely should not be used in acetonitrile. All further measurements reported in this communication were made in an all-acetonitrile cell l in acetowith the Pleskov Ag/(O.Ol ~ l Ag;S03 nitrile) electrode as reference, and with a 0.1-M solution of tetraethylammonium perchlorate in acetonitrile serving as salt bridge between the two working compartments of the H-type cell described before. It was realized that this salt bridge mould not eliminate the liquid junction potential between the t n o working compartments, because the mobilities of tetraethylammonium and perchlorate (7) R C . Larson, R. T I v a m o t o a n d R. N. Adams, Anal. Cham. Acta 2 5 , 371 (1961). ( 8 ) I n the voltammetric studies in acetonitrile as solvent reported b y Kolthoff and Coetzee J Am Chem Soc , 7 9 , 870, 1852, 6110 (1Q57), all measuienients wero made as quickly as possible, generally a i t h i n 5 inin (9) An aqueous agar salt biidge containing sodium 01 tetraethylammoniuin peichloiate, rathei t h a n potas-ium chloride, m a l be more satiafactorj i n acetonitrile.
J. F. COETZEE AND G. R. PADMANABHAN
1710 + 1401
+.
8 t 120z 6
0 t100-
?3 w‘
+80t60-
+40-
t20
-05
I
0
105
ti0
Vol. 66
(pKI = 8.9), has been determined by Kolthoff, Bruckenstein, and Chantooni13 using conductometric and spectrophotometric methods. Not many salts are completely dissociated in acetonitrile up to moderate concentrations to 1M).lo Although incompletely dissociated salts could be used in buffer solutions, calculation of the hydrogen ion activity of such solutions would require knowledge of an additional equilibrium constant, namely the dissociation constant of the salt (in addition to the simple dissociation constant of the acid, K 1 , as well as the formation constants of any anion-undissociated acid complexes that may be formed, given by K2/Kl or Ka/K1). The use of picric acid-tetraethylammonium picrate buffers seemed most promising for the present purpose. Picric acid undergoes simple dissociation up to relatively high concentrations,8,12 and tetraethylammonium picrate is one of the strongest of the large number of electrolytes studied by Walden and Birrlo (experimental slope of A us. C‘/a plot = 373; Onsager slope = 351); this salt is virtually completely dissociated up to M concentrations. For picric acid undergoing simple dissociation, it follows that
where all symbols have their customary meaning. At constant ionic strength the activity coefficient ratio, fo/fl, should remain essentially constant. Hence, if the glass electrodc measures hydrogen ion activity reversibly, it follows from eq. 1 and the Kernst equation that a plot of the potential of the electrode as a function of the quantity log ([HPi]/[Pi-1) should be linear with a slope of 59 mv. a t 25” (provided junction potentials remain essentially constant). Three glass electrodes used in several series of buffers came to equilibrium in a relatively short; time (within 5 to 30 min.). The results obtained with one of the electrodes in two series of buffers, one at a constant concentration of tetraethylammonium picrate of M and the other a t M, are represented in Fig. 1 by lines A and B, rcspcctively. The quantities Ca and C, refer to the total (analytical) concentrations of acid and salt, respectively, and were varied only within that region where the effect of the additional complicating reactions described above would be insignificant. It can be concluded that the response of the glass electrode in picric acid-tetraethylammonium picrate buffers in acetonitrile as solvent is reversible. The hydrogen ion activity of the buffer solutions used (at least for the series a t ionic strength lo+) can be calculated from eq. 1 and the “ r e d ~ c e d ” ’ ~ Debye-Huckel equation, which becomes for 25” log f i = -3552,2D-”/”S1/z (2) For a singly-charged ion in acetonitrile (dielectric constant D = 36.0) a t an ionic strength S = 1.0 X (12) Unpub!ished results from this Laboratory ( G . Cunningham). (13) The extended form of the Dehye-Huekel equation could not be used, since ion size parameters for acetonitiile as solvent ale not yet available.
Sept., 1962
AUTOPROTOLYSIS COXSTANT OF ACETOXITRILS
- 4 3 0 y
-I 0
1
I
-0.5
0
loa
1711
,
+ 0.5
t
1.0
r w i o o - PI.
Fig. 2.--Ytesponse of glass electrode during titration of 1,3-diphenylguanidine (DPG) with perchloric acid in acetonitrile. M DPG A, 6.3 x 10-3 M DPG with 7 X 10-2 1Pf HC104* B, 2.0 X 10-2 M DPG with 2.5 x 10-1 M Hc104; C, 4.0 X with 4 X 10-2 N HClOa; closed circles, 0.1 M ht4NC104added; open circles, 0.2 M HzO added.
it follows that the activity coefficient SI = 0.89. Substituting K1 = 1.26 X lop9, and assuming that fo 1, eq. 1gives for S =
-
B
+ CHBCNH+
BH+
+ CH&N
(5)
For a base as strong as diphenyIguanidine, the equilibrium lies very far to the right. Results preaIz+= 1.41 x 1o-9c,/cs sented later indicate that for solutions as dilute (3) as those titrated, BH+C104- ion pair formation E’rom line B in Fig. 1, the potential of the glass need not be considered. The dissociation constant, electrode 2)s. the hglO.01 Ai! XgSOs reference XBHt, of the pro tona ted base is given by clectrode at Ca/C, = 1 was equal to +67 ~ n v . ~ * lIcncc tho following equation represents thc response of the particular glass clectrode used Eslnss (in mv. us. Ag ref. el.)
=
590
+ 59 log
UII
+
(4) The same electrode then was used in buffers of the type B BH+C104-. Response of the Glass Electrode in Diphenylguanidine Buffem-Several solutions of 1,3-diphenylguanidine were titrated with perchloric acid in acetonitrile as solvent. After each addition of titrant, the electrode catnie to equilibrium very quickly. The shape of the titration curves was that expected for a typical weak base-strong acid titration. Since perchloric acid behaves as a strong (completely dissociated) electrolyte in acetonitrile, its reaction with l&diphenylguanidine (a) can be represented by the simple equation
+
(14) The difference between corresponding potential values obtained
-
a t ionic strength values of lo-* and 10-8 (67 5nd 73 m v , respectively, a t C, C,)i 9 in good agreement with the 7 mv. difference predicted by eq. 2, which can give only very approximate results for an ionic strength as high a s lo-’ in a solvent such as acetonitrile.
It follows from eq. 6 and the Nernst equation that a plot of the potential of the glass electrode us. the quantity log; [P/(lOO - P ) ] ,where P = percentage of B t,itrated, should be linear mith a slope of 59 mv., provided (a) the electrode responds to hydrogen ion activity in a reversible manner, (b) neither B nor BH+ is involved in additional reactions, (c) the ratio fo/fl does not change significantly during the titration, and (d) the algebraic sum of all liquid junction potentials involved does not change significantly during the titration. Several such plots are shown in Fig. 2 . It has been shown by conductance measurements that in acetonitrile, hydrogen bonding occurs between 1,3-diphenylguanidine and its protonated form, resulting in the species BHB+ with a formation constant of 20.2 The effect of this complexation reaction will be to increase the slope of the E us. log [P/(100 - P ) ] plot to values greater than 59 mv. However, for solutions as dilute as those titrated the effect will be small. For example,
J. F. COETZEE A N D G. R. PADUANABHAX
1712
consider the titration of 6.3 X 10-d JI B with 7 x 10+ perchloric acid. At 9.1 % Titrated.-
+ [BHB+] [B] + [BHB+]
[BH+]
=
=
[BHB+] [BI [BH+I
5.7 X
5.7 X
=
20
Solving, [BHB+] = 6 X [BH.+]= 5.1 X lov4,and [B] = 5.6 X 10W3,so that [B]/ [BH+] = 11, rather than the value of 10 which would apply if the species BHB+ were not formed. At 50% Titrated-[B]/[BH+] = 1, whether or not BEIB+is formed. At 90.9% Titrated.-[B]/ [BH+]= 1,’ll. Hence, the slope of the plot between 9 and 91% titrated = (log 11) X 59mv. = 61 mv. A second type of hydrogen bonding reaction also must be considered, namely that which stabilizes the ion pairs of the salts of incompletely substituted ammonium bases in a non-hydrogen bonding solvent such as acetonitrile, which (unlike waterlike solvents) does not mask this effect. It can be shown that such an ion-association reaction will decrease the slope of the E us. log [ P / ( l O O P)] plot. However, since addition of a common ion (as 0.1 M tetraethylammonium perchlorate) did not cause a significant change in the slope of the plot (line C in Fig. 2 ) , it seems that 1,3-diphenylguaiiidinium perchlorate is extensively dissociated in acetonitrile. Finally, it is to be expected that if the ionic strength increases significantly during the titration (lines A and B in Fig. a), the activity coefficient ratio fo/fl will increase ; hence eq. 6 predicts that the slope of the E us. log [P’(100 - P)] plot will decrease. However, the liquid junction potential may also change. In the titration represented by line C in Fig. 2, the activity coefficient ratio, as well as the liquid junction potential, should have remained constant. It is concluded that thc response of the glass electrode was reversible. Since the titrant was prepared from 70% a-queous perchloric acid, it introduced appreciable amounts of water into the systems studied.16 For example, in the titration of 6.3 X JI base, 7 x 10-3 LW water had been introduced a t the “half-neutralization” point, in addition to the approximately 1 X JI water originally present in the solvent. However, line C in Fig. 2 shows that addition of even much larger amounts of water has little influence on the slope of the line (16) There is as yet no simple and convenient method available to prepare anhydrous solutions of perchloric acid in acetonitrile uithout introducing additional (more or less undesirable) substances. Anhydrous solutions of perchloric acid in acetic acid as solvent can be used for certain studies in acetonitrile b u t are obviously not applicable Generation of perchloric acid directly in for the present puiposc acetonitrile b y treating a fiolutiori of s i h e r perchlrnate uTith hydrogen chloride pa? is not completely satisfactory, hecause the mlribility of silver chloride in acetonitrile (pal ticulails i f exces4 chloride iz present) 18 appraciable, ov,ing to the stabilitg of sill er chloride comnlexes in acetonitrile. Likewise, on addition of anhydrous sulfuric acid to a n equivalent amount of barium perchlorate in acetonitrile, a considerable amount of bisulfate is formed, since sulfate ion is a ielatil ely sliong base in acetonitrile
Vol. 66
and only a moderate effect 011 the “half-neutralization” potential. The Autoprotolysis Constant of Acetonitrile.Acetonitrile is a comparatively inert amphiprotic solvent, with relatively weak basic and very weak acidic pr0perties.l Hence its self-ionization will be very limited, and is assumed t o occur by the reaction 2CH3CN If CHsCNH+
+ CI&CK-
(7)
Its autoprotolysis constant is then given by
Ks aCHsCSH7 x aCH2Ch (8) and can now be calculated as follows. The over-all ionization constant of 1,3-diphenylguanidine, determined conductometrically,2is given by K B = a’B1Ii
x
=
aCHzCN / a B
2
x
lo--”
(9)
In the titration of 6.3 X IO-d M 1,3-diphenylguanidine with 7 X W perchloric acid, the “halfneutralization” potential is equal to - 350 mv. (Fig. 2, line A). From eq. 4 the corresponding hydrogen ion activity equals 1.3 X 10-l6. The ionic strength a t this point is equal to 3 X 10-d J f . Hence, from eq. 2 , ~ B H += fl = 0.81, and from eq. 6, assuming thatfs = fo 1
-
KBHt
aB
x
aHt/aBH+
=
1.6
x
(10)
The product of eq. 9 and 10 gives the autoprotolysis constant of acetonitrile KBKBH+ =
K,
=
3
x
10-27
(11) A value for the autoprotolysis constant of acetonitrile has been reported before by Romberg and Cruse,16from the results of the titration of various nitrogen bases with several weak acids, using mainly a glass electrode as hydrogen ion indicator electrode. The autoprotolysis constant of the solvent was calculated from the equivalence point potentials of those acid-hase pairs which gave symmetrical titration curves. Romberg and Cruse’s value of K , = 3.5 X 10-20 is 7 powers of 10 larger than The autoprotolysis constants of several solvents are listed in Table 11. It is seen that, the value for acetonitrile is much smaller than that of any other solvent listed. The value of the autoprotolysis constant of a given solvent depends on the strength of its acid as well as its basic properties, and also on the magnitude of its dielectric constant. The weaker the acid-base properties are and the lower the dielectric constant is, the smaller the autoprotolysis constant becomes. The dielectric constant of acetonitrile is somewhat higher than that of methanol, yet its autoprotolysis con(16) E Romberg and K Cruse, Z. Elektiochem , 63, 404 (1959) (17) Romberg and Cruse calibrated their glass electrode in unbuffered solutions of pirric acid in acetonitrile. However, on changing the picric acid concentration bs a factor of 10, the potential changed by 64 m.r rather than b y the 30 inv associated with reversible response to simple dissociation (but i f unilateral triple ion formation is predominant, the change uould be 59 mv 1. Furthermore, the value ured for the (simple) dirrociation consttnt of picric acid (2 5 X lo-’) u r n s much largei than t h a t accepted b y us (1 26 X 10-8) T h e dissociation of weak acids in acetomtrile incieases n i t h time. Romberg and Cruse’s value might ha1 e been determined for aged solutions.
Sept ., 196f!
HIGH
TEMPERATURE HEATCOXTENTS OF TeOz ASD Ka:l’eO4
1713
However, attempts to prepare such lyate ion solutions in sufficiently high concentrations for practical purposes (e.g., by treatment of acetonitrile with sodium or lithium metal) only results in polymerization of the solvent. Likewise, esTABLE I1 sentially anhydrous solutions of tetraalkylamAUTOPROTOLYSIS CONSTANTS O F TrARIOUS SOLYEKTS monium hydroxides in acetonitrile are not stable, Dielectric constant a t least not a t ordinary temperatures. This lack Solvent (25‘) P& of availability of a really strong base constitutes IVater 78 5 14.0 one of the major limitations of acetonitrile as a Formic acid 58.5 6.2 medium for acid-base reactions. Acetonitrile 36.0 26 5 The over-all ionization constants of amines are Methano i 32.6 16 7 approximately 8 powers of 10 smaller in acetonitrile Ethanol 24 3 19 1 than in water.2 However, the autoprotolysis Acetic acid 6.1 14.5 constant of acetonitrile is from 12 to 13 powers a All values except for acctonitrile were taken from I. Rf. of 10 smaller than that of water. Hence, titration Molthoff and S. Bruckenstein in “Treatise on Analytical of amines with perchloric acid gives a potential Chemistry, ’ Part I, J‘ol. 1 , Interscience Publ., Inc., break a t the equivalence point which is from 250 Kew Yorli, K.Y., 1959, p. 484. to 300 mv. larger in acetonitrile than in water, The titration of a strong acid (perchloric acid) and very weak bases can be titrated. Finally, with a solution of the lyate ion of acetonitrile it should be possible to titrate virtually all anion (CH?C;CT-) should give a very large potential bases which are soluble in acetonitrile. For exbreak at the equivalence point (over 1,000 mv. for ample, even chloride ion (pK of HCl = 8: ref. 3) 0.1 J1 solutions, as compared to 350 mv. in water). is a relatively strong base in acetonitrile.
stant is smaller by 10 powers of 10. This large difference is caused by the fact that acetonitrile is a weaker base and a much weaker acid than methanol.
THEIiJ4ODYNAXLIG PROPERTIES OF IXORGANIC SUBSTAKCES. IV. THE HIGH TEXPERATURE HEAT CONTENTS OF TeQz AND Na2TeO4‘ BY REIJI MEZAXI AND J0a-i L. MARGBAVE Department of Chemistry, University of Wisconsin, Madison, Wisconsin Received April 6,1961
The heat contents of TeOz and KazTeOa contained in gold capsules have been measured in a drop-type calorimeter over the ranges 446-1146 and 421-804OK., respectively. The heat of fusion of TeOz is 6.95 & 0.10 kcal./mole.
Reliablc high temperature thermodynamic data have not been available previously for TeOz or Xa2Te04. Estimatcs for Te02 have been given hy Kubaschcwski arid Evans3 bascd on aiialogics with other osidcs. S o iuformatioii has been avai1:~hlcFor NaLTcO1. Experimental Techniques and Material The calorimeter used in this work has been described previously.‘ The apparatus has been calibrated electrically and, in addition, samples of synthetic sapphire from the National Bureau of Standards have been run for comparison purposes. The accuracy of the calorimeter is It0.5%, while the rrproducibility of a given measurement is slightly better. Several drops also were made with empty pure gold capsules at various temperatures since the heat content must be corrected for the containing capsule. The measured heat content may be compared with calculated values from Kelleys and in 43 runs the observed average deviations were zkl% over the range 400-700’K. and & 0 . 5 ~ ,ovci- the range 700-1200’K. The Tech samples studied &-ere spectroscopically pure from Johnson, Natthey, and Co., Ltd., and showed only traces (a few p.p.m.) of Ag, Ca, Na, Si, and Mn. I n the original runs with TeOz sealed in an argon atmosphere, (1) 1,’or earlier payers in this series see: I. J. L. Margrave and R. T. Grimley, J . Phus. Chem., 62, 1436 (1958); 11. &I.8. Chandrasekharaiah, R. T. Grimley, and J. L. Margrave, ibid., 63, 1505 (1959); 111. R . T. Grimley and J . L. Margrave, ibid.,64, 1763 (1960). ( 2 ) 0. I
