electronic timer can cause excessive error in the triggering. If excessive filtering is used with low viscosity oils, the triggering pulses are reduced in amplitude. This reduces the maximum rise time of the triggering pulses, and precise timing is not possible. The ol)t,imumsettings are determined esperimentally for a new instrument when put into service. Calibration. Departure from Stokes' straight-line relationship between fall times and the viscosities of liquids occ,urs a t onset of turbulence. Under present conditions, this is observed at viscosities below 40 to 50 cp., when the drag caused by turbulence tends to increase the fall times beyond those due to viscosity alone. The fall times, howeoer, remain repeatable, and the viscosity us. fall-time relationship can be defined satisfactorily by calibration down to shout 2 cp. The function (7.4487 - d J t against q , where 7.4487 is t'he density of the steel balls, dl is the density of liquid, t is the fall time, and q is the absolute viscosity, was plotted for liquids covering a considerable range of viscosities and densities. A single straight line was obtained down to the onset of turbulence; this and the
curvature arising beyond that point are shown in Figure 5. I n routine operation it may be desired to plot fall times directly against viscosity. Because of the relation espressed by Stokes' law, this function will be valid only for liquids of a single density. If the liquids to be tested cover a significant range of densities, the calibrations can be made available on a single sheet by plotting a family of curves representing the viscosity-fall time correlation obtained for different densities. Each curve then represents a calibration for liquids having a corresponding density. Alternatively, it is possible to construct a three-branch nomogram for the purpose. Performance. T h e falling ball viscometer was calibrated with standardizing oils of known viscosities a n d densities and with mixtures of these oils. The fall times were measured with a group of 10 balls with closely matching diameters and densities, each dropped once in each oil. The average relative standard deviation of the fall times for the set was O . l O l ~ o under standard testing conditions. By contrast, the corresponding standard deviation for a single ball dropped ten times was only 0.05%.
The set of balls evidently showed small variations in density, size, or sphericity, which account for the larger standard deviations. The variation can be further reduced by actually dropping the balls through a suitable long column of liquid and collecting those bracketed within a specified narrow range of fall times. ACKNOWLEDGMENl
We acknowledge the contributions made by Gordon O'Donnell and F. A. Olson. LITERATURE CITED
(1) Fidleris, V., Whitmore, R. L., J . Sci. Jnstr.. 36, 35 (1959). ( 2 ) Lana G. L., U. S. Patent 2.252.572 I
,
(August 1941). (3) McDowell, C. A., Walker, B. Y., Chem. I n d . ( L o n d o n ) 64,323 (1945). (4) Moore, L. P., Cuthbertson, A. C., IND. ENG. CHEM.,ANAL.ED. 2, 419 (1930). (5) Symmes, E. &I., Lantz, E. A,, Zbid., 1. 35 (19291. (6) 'Thompson, A . M., J. Sci. Instr. 26, 75 (1949). (7) Wertheim, R. A . P., Patent Oflice, London, Patent Spec. 758,199 (1956). RECEIVED for review August 18, 1964. Accepted October 1, 1964.
Structure and Behavior of Organic Analytical Reagents Formation Constants of Transition Metal Complexes of 2-Hyd roxypyrid ine-1-Oxide and 2-Merca ptopyridine-1-Oxide PENG-JOUNG SUN,' QUINTUS FERNANDOl and HENRY FREISER Department of Chemistry, University o f Arizona, Tucson, Ariz. The solution equilibra of 2-hydroxypyridine-1 -oxide and 2-mercaptopyridine-1 -oxide, heterocyclic analogs of hydroxamic acids, were studied to evaluate the effect of the nature of the donor atoms, the nondonor atoms in the chelate ring, and auxilliary ring systems in the metal chelate, on the analytical behavior of these reagents. The acid dissociation constants and the chelate formation constants of these two compounds with Mn(ll), Co(ll), Ni(ll), Cu(ll), and Zn(ll) were determined potentiometrically and spectrophotometrically. The significance of these results in the design of new analytical reagents is discussed.
T
of metal chelate formation equilibrium data in analytical chemistry arises from its use in predicting values of solution paramHE
IMPORTANCE
eters appropriate for analytical determinations. The extent of metal chelate formation depends not only on the value of the formation constants but also on the concentration of the free ligand species which in turn depends on the acid dissociation constants of the ligand. These considerations are clearly expressed in terms of the proton displacement constant, K p d ,which is the product of the overall formation conconstant, &, and the acid dissociation constant K , raised to the nth power where n is the number of ligands in the chelate. Thus, for the formation of copper acetylacetonate : Cu2+ 2Hacac Cu(acac)z 2 H +
+
+
. Ka2
(1) I n a series of related ligands chelate stabilities increase with ligand basicK D d
= Pz
ities-Le., pn increases as K , decreases (4). Inasmuch as K , appears to a higher power than pn in K p d , it is not too surprising to find that in such families of ligands, the less stable the chelate a ligand forms with a particular metal ion, the lower the pH a t which formation occurs. For example, the halogenated 8-quinolinols form metal chelates a t lower pH values than the parent compound and thenoyltrifluoroacetone (TTA) reacts with metal ions a t lower p H values than does acetylacetone. I n designing chelating agents for analysis then, the effect of structural changes on acid dissociation behavior as well as on metal complex formation must be taken into account. It will be of advantage to have reagents in which the K , as well as ' O n leave from the Department of Chemistry, National ~~i~~~ university, Taipei, Taiwan, China. VOL. 36, NO. 13, DECEMBER 1964 e
2485
Pn are as high as possible since this will permit the chelates to form in media of great'er acidity. A case in point of increasing Pn without decreasing K,, is illustrated by the comparison of chelates of 8-diketones and substituted salicylaldehydes, both of which have the same chelate rings. At comparable K , values, @-diketonesinvariably form copper(l1) chelates of higher fin than those of the salicylaldehydes so that significantly higher K p d values are obtained with the p-diketones. hnother example is the higher K jvalues found with chelates of 8-quinolinol when compared with those of o-aminophenol, bobh .of which have the same chelate ring and essentially t,he same K , values ( 3 ) . Similarly, when a mercapto group is substituted for a phenolic group in a chelating agent, a substantial increase in K , results without any decrease in Pz values for the transition and heavy metal chelates. Indeed, an increase in p2 values is observed with some metal ions. As part of a wide-ranging study of design factors of chelating agents for analytical applicat,ion, the chelating properties of 2-liydroxypyridine-l-oxide, which has the same chelate ring as the analytically useful hydroxamic acids, were studied. The chelating properties of 2-mercaptopyridine-1-oxide were also studied to explore the effect of sulfur substitution in this chelate ring system. EXPERIMENTAL
Materials. The synthesis of 2hydroxypyridine-1-oxide was carried out by a method described by Shaiv et al. ( I @ , with certain modifications. The perbenzoic acid oxidation of 2bromopyridine gave 2-broniopyridine1-oxide hydrochloride which was hydrolyoed with 10% S a O H by heating on a water bath for 11j2 hours. After cooling the solution was acidified with HC1 t'o a pH of 3. The resulting solution was evaporated to dryness under reduced pressure between 50" and 60' C. The crude compound obtained was extracted with hot ethyl acetate and the extracted material was purified by vacuum sublimation. The sublimate was further purified by recrystallizing it from a mixture of dioxane and benzene until a white crystalline solid was obtained: m.p. 150-151" C.; lit. ni.1). 151" C. (8); 149-150' C. ( 1 6 ) ; 149-151' C. ( 2 3 ) ; 148-149' C. (?, 9). The purity of the compound was checked by a potentiometric determination of its neutralization equivalent. The method employed for the synthesis of 2-mercaptopyridine-I-oxide was essentially t,hat described by Shaw et al. (26). An absolute ethanol solution containing equimolar amounts of 2-broniopyridine-1-oxideand thiourea was refluxed for 1112hours and a y e cipitate of the addition iiroduct that was forined was filtered and washed with absolute ethanol. The thiouiwi addition product (ni. p. 160-160.5" C . ) , 2486
ANALYTICAL CHEMISTRY
was hydrolyzed with SasCOs solution and 2-niercaptogyridine-1-oxide precipitated by acidification of the solution with HCl. Recrystallization from 95yc ethanol gave a white crystalline compound: m.p. 71.5-72.5" C . ; lit. m.p. 68-70' C. (26); 70-73" C:. (8);63-64' C. ( 1 2 ) . The purity of the compound was checked by a potentiometric determination of its neutralization equivalent. Solutions of the compound were light sensit,ive but they were stable for two or three days when stored in the dark. Methods for the preparation and *t'andardization of S a O H and inetal perchlorate solutions, as \vel1 as for the purification of dioxane: have been previously described (11). Potentiometric Determination of the Second Acid Dissociation Constant (pK,,) of 2-HydroxypyridineI-oxide and 2-Mercaptopyridine-loxide. The pK,, values of the two compounds were determined potentiometrically a t 25' C. i 0.1' in t,he following way. A weighed quantity of the compound was transferred to a water-jacketed vessel and dissolved in a measured volume of water containing a known amount of HClO,. A carbon dioxide-free atmosphere was maintained above the solution by passing nitrogen gas, saturated with water at 25' C., t'hrough the solution and above it during the titration with carbonate-free KaOH. All pH nieasurenients were made with a Beckman Model G pH meter equipped with a glass-saturated calomel electrode pair and calibrated with standard Beckman buffers a t pH 4.00 and 7.00 a t 25' C. The values of pK,, for the two compounds were calculated from the experimental points in the buffer regions of the two titration curves. Since in both cases the dissociation of the neutral species to give the anionic species is the only reaction of importance, =
Table I. Stepwise Metal Chelate Formation Constants of 2-Hydroxypyridine-1 -oxide Determined Potentiometrically a t 25' C. f 0.1 " and Ionic Strength 0.005 in Water Metal Metal log log log ligand ion ratio R, k', Ka% Zn(I1) 1 5 0 5 33 4 34 4 b4 1 7 6 5 51 4 40 5 00 1.9 1 5 36 4 38 4 88 Ni(I1) 1:4.6 5 . 3 5 4 1 5 4.7.5 1:7.2 5.40 4.23 4.81 1:9.3 5 . 2 5 4.26 4 . T Cu(1I) 1:5.1 . . . 5 . 8 7 6.21 l:lO.l 1:j.Z
Co(I1)
...
1:8.1 1:lO.O
Mn(I1)
1:j.l 1:g.o
1:lO.l
5.86
,..
5.15 4 . 1 8 4 . 6 7 5.22 4.17 4 . 7 0 5 . 2 5 4.20 4 . 7 2 4.45 3.27 3.85 4 . 4 5 3.42 3 . 9 0 4.48 3.32 3.90
A p , and A , the absorbance of a solution containing the neutral species, the protonated species, and a mixture of the two species, respectively. The value of pK,, obtained \vas -0.9 i 0.3. .A value of -0.8 for pKsl has been reported by Katritzky and coworkers ( 7 ) . Potentiometric Determination of Chelate Formation Constants of 2Hydroxypyridin e- 1-oxid e with M eta1 Ions in Water at 25' C i. 0.1'. The stepwise formation constants were determined by the Bjerrum potentiometric method ( 2 ) . h weighed quantity of the reagent was placed in a water-jacketed titration vessel and dissolved in 50 ml. of 0.01.11 HC104. Five milliliters of a 0.01JZ standard metal perchlorate solution was added together with 45 ml. of water. The solution mas titrated potentiometrically
A N ,
+
[H+] {["I [ S a + ] - [C104-] - [OH-]) C R - [H+] - [Na+] + [ C l o d - ] [OH-]
where all the terms in square brackets are concentrations in moles per liter and CH is the analytical concentration of the reagent. Values of pK,, for the two compounds were also determined in a siinilar manner at an ionic strength of 0.1 after the addition of a calculated quantity of xaC104 to the solution. Spectrophotometric Determination of the First Acid Dissociation Constant, pKal, of 2-Hydroxypyridine-loxide. .I 9.12 X 10-jJ1 solution of the compound was prepared in H2S04 and the variation of its absorbance with concentration of H2S04 was measured a t 298 mk in a Beckman DU spectrophotonieter in a I-cm. cell a t 27.5" C. The value of pKa, was obtained froin the following equation:
where Ho is the Hanimett acidity function of the H3SOa solution (IO, 2 4 ) ,
+
with carbonate-free NaOH. Altleast three different metal :ligand ratio5 were used for each titration. The values of the stepwise formation constants, K 1 and K? were determined from sets of ?us. i - log [ R - ]data, where %C.w = C R - {C,
+ C R + [OH-] -
and
[R-]
=
K
[H+l
(C,
+ C R + [OH-] [H'l + [ S a - ] ] (5)
and C,, C.V~, C R are the analytical concentrations of acid, inetal ion, and reagent, respectively (Table I). Spectrophotometric Determination of the Chelate Formation Constants of the 1 :I Metal Chelates of 2-Hydroxypyridine-I-oxide and 2-Mercaptopyridine-1-oxide. .A series of solutions, each consisting of 10 nil. of buffer, 10 mi. of metal perchlorate, and 10 ml.
of either 2-hydroxy or 2-mereaptol)yi,idine-l-oxide, were prepared. The ionic. strmpth was 0.1 and the metal: reagent ratio was greater than 1OO:l in cach of the solutions. Absorbance mrahurcments were carried out a t the wavc,lcngth of inaximuni absorption of the metal chelate. The chelate formation constants of the 1 : 1 nietal chelates were calculated from the measured absorbances and pH v a l u c ~of the solutions by means of the following equation:
K , . K,,
=
{ .IL--EL H R R j I"[ ~
E M R. C R - d
where Kl is the formation constant of the 1:1 complex, Ka2 the second acid dissociation constant, of the reagent, C.W and C R the analytical concentrations of the iiietal ion and reagent, respectively, [ H + ] the measured hydrogen ion concentration, A the observed absorbancc of the solution, E H Rand E M Rthe molar absorptivities of the neutral reagent and metal complex, respectively, at the wavelength used in the determination, Equation 6 can be used if the absorbance measurements are made in such a pH range and wavelength that the absorbance of the cation and anion of the reagent as well as the absorbance of the metal ion can be neglected.
DISCUSSION
Although tautomeric forms of acetohydroxamic acid ( a ) , 2-hydroxypyridine-1-oxide ( b ) , and 2-mercaptopyridine-I-oxide (c) can exist, the structures as drawn have been shown to be predominant ( 5 , 7 , 12, 13, 15, 16). Hence the pK,, values (Table 11) of these compounds all represent the loss of a hydroxy proton.
I
I
OH
PK~,
PL
- 0 . 9 =k 0 . 3
5.98 A 0 . 0 2 (Ionic strength 0.005) 5.81 i 0 . 0 2 (Ionic strength 0 . 1 ) 5.9w
-0.P
2-Mercaptopyridine-1-oxide
- 1.95b
a
4.43 i 0 . 0 3 (Ionic strength 0 . 1 ) 4.60 i 0 03 (Ionic strength ca. 0.005) 6.60 i 0.04 i50c;lc aqueous dioxane, ionic strength 0.005) 4.67b 9 . 3 > (Ionic strength 0 . 1 )
Reference ( 7 ) .
* Reference (fd). Reference ( 1 ) .
Table Ill. Formation Constants of the 1 : 1 Metal Chelates of 2-Hydroxypyridine1 -oxide and 2-Mercaptopyridine- 1 -oxide Determined Spectrophotometrically at Ionic Strength 0.1 in Water
2-Hyd~oxypyridine-1-oxide 2-l\lercaptopyridine-l-oxide
Wavelength, Metal ion ZnlII) Cu(I1) Pil(I1) Co(I1) SIn(I1) a
log K ,
mp
5.1
213 212 5 21 7 213 7 215
7 3
5 1 4 9 3 9
log l ' h
Wavelength, mr
_2.17 _
.i 3
>8
240 243 236 240
5
5 1
4 8 3 1
Acetohydroxarnic acid,a Ki
_ _
5 4
7 9
5 3
5 1
4 0
T'alues taken from Reference ( 1 )
d - EHRCR
gave straight lines of unit slope from ivhich K1 values were calculated.
OH
Compound 2-Hydroxypyridine-1-oxide
ilcetohydroxamic acid
CM (6)
Plots of pH us. log
Table 11. Acid Dissociation Constants of 2-Hydroxypyridine-1 -oxide and 2-Mercaptopyridine-1 -oxide Determined Potentiometrically in Water at 2 5 " C. +O.l "
I
OH
The great drop in pK, from (a) to ( b ) might well reflect the electron-withdrawing influence of the ring system of the pyridine. The further decrease in pK, in going from (b) to (e), despite the substitution of the less electronegative sulfur, probably results from the presence of strong hydrogen-bonding in compound ( b ) . Sl)ectrophotometric evidence for hydrogen-bonding in ( b ) has been obtained ( 7 ) . Two of the most important factors in determining the chelating ability of a
ligand are the nature of the donor atoms and the size of the chelate ring. As shown by a diagnostic log-log plot of the formation constants of nickel(I1) LS. zinc(I1) chelates, the behavior of ligands which form 5-membered ring chelates can be represented by a family of parallel lines of approximately unit slope and whose intercepts depend on the nature of the bonding atoms ( 6 ) . For example, the line for reagents having two oxygen atom donors has an intercept of zero-i.e., the zinc and nickel chelates have approximately the same formation constants. This correlation has so far been noted only for ligands in which the nondonor chelate ring atoms have all been carbon. In this connection, the relative stab the metal chelates of 2-hydroxypyridine1-oxide and 2-mercaptopyridine-1-oxide are of interest, since both reagents have a nitrogen atom which is not a donor in the metal chelate ring. The stabilities of the nickel(I1) and ainc(I1) complexes of each of the reagents shown in Table I11 are essentially identical indicating that these reagents behave in accord with other 5-membered ring-forming chelating agents having two oxygen donor atoms. I t would seem likely that cupferron, in which both nondonor atoms in the chelate ring are nitrogen, would also behave in this manner. More data are
required before the behavior of 5membered ring-forming chelating agents having a sulfur and an oxygen as donor atoms can be systematically described. A noteworthy difference between the 2mercapto- and the 2-hydroxypyridine-1 oxide reagents is the significantly higher stability of the copper(I1) complex of the mercapto compound. This coniplex is so stable that it does not dissociate appreciably in highly acid solutions (up to 431 HC10,). The formation constants (Table 111) of the chelates of ( b ) and (c) are approximately the same. From this it is obvious that the proton displacement constants for the chelates of these reagents increase in the order, (a) < ( b ) < ( c ) . Hence the 2-merc~iptopyridine-1-oxide will form chelates in more acidic solution than will either of the other t\To reagents. On the basis of these results, substituted 2-mereaptoand 2-hydroxypyridine-1-oxides would seem to hold promise of development as analytical reagents superior to the hydroxamic acids. LITERATURE CITED
(1) Anderegg, G , L'Eplattenier, F , Schwarzenbach, G , He16 Chim Acta 46, 1400, (1963) (2) Bierrum, J "LIetal Ammine Fnrmation In A4queous Solution," P Haase and Son, Copenhagen, 1941 ~
VOL. 36, NO. 13, DECEMBER 1964
2487
( 3 ) Bjerrum,
J., Schwarzenbach, G., Sillen, L. G., "Shbility Constants of Metal Ion Complexes" Part I, The Chemical Society of London, 1958. (4) Calvin, JI., Wilson, K. FV., J . .4m. Chem. SOC.67, 2003 (1945). ( 5 ) Cunnineham. K. G.. Newbold. G. T.. hpring, F: S., 'Stark, 'J., J . Chem; Soc: 1949. rn. ~2091.- - ~ ( 6 ) Freiser, H., Fernando, Q., Cheney, G. E., J . Phys. Chem. 63, 250 (1959). ( 7 ) Gardner, J. N.. Katritzkv, A . R..J . Chem. SOC.1957, p. 4375. ~
1
(8) Hamana, M., Yamazaki, >I., Yakugaku Zasshi 81, 574 (1961); C d 55, 24743a (1961 j. ( 9 ) Ibid., p. 612: C 4 55, 24742f (1961). ( 1 0 ) Hammett, L. P., Deyrup, A . J., J . Am. Chem. SOC.54, 2721 (1932). ( 1 1 ) Johnston, IV. I]., Freiser, H., Ibid., 74, 5239 (1952). ( 1 2 ) Jones, R . A , , Katritzky, A . R., J . Chem. SOC.1960, p. 2937. ( 1 3 ) Newbold, G. T., Spring, F. S., Ibid., 1948, p. 1864.
( 1 4 ) Paul, Ill. A., Long, F. A,, Chem. Revs. 57, 1 (1957). ( 1 5 ) Shaw, E.. J . Am. Chem. SOC. 71. 67 (1949). ( 1 6 ) Shaw, E., Bernstein, J., Losee. K., Lott, .'!I A , , Ibid., 72, 4362 (1950).
RECEIVEDfor review July 31, 1964. Accepted September 21, 1'364. \Vork supported by the Kational Science Foundation and in part by the Sational Institutes of Health.
Quantitative Conversion of Molybdate to Mo(V) by the Stannous Chloride-Perchlorate Reaction and Spectrophotometric Determination as TricapryIyImet hy I a m mo nium Oxy t e t ra t hiocy a nat o mo I y b d at e (V) A. M. WILSON Department of Chemistry, Emory University, Atlanta, Go. 3 0 3 2 2
0.K.
McFARLAND'
Department of Chemistry, Wayne State University, Detroit 2, Mich.
b Reduction of molybdate by excess stannous chloride in 6M hydrochloric acid this reaction produces Mo(lll). Excess stannous ion i s completely oxidized by perchlorate if the initial perchlorate/stannous ion mole ratio is 20.25. Oxidation of Mo(lll) to Mo(V) by excess perchlorate is rapid and autocatalytic in 9M hydrochloric acid at 64" C. The subsequent oxidation of Mo(V) to Mo(VI) by excess perchlorate in 9M hydrochloric acid at 70" C. has a half life of 8 hours. Removal of the excess stannous ibn by excess perchlorate at all hydrochloric acid concentrations, previous to thiocyanate addition and development of the characteristic orange-red color of oxypentathiocyanatomolybdate(V), assures stable absorbance of this complex in the aqueous phase up to 1 hour and in the quaternary-chloroform extract up to 2 4 hours. Beer's law is obeyed in the quaternary-chloroform media, E = 1.64 0.02 X Results of interference studies and determinations of molybdenum in some NBS alloys are presented.
*
lo4.
D E T E R W N A T I O N a$ the ouypentathiocyanatomolybdate(V) 14 a well known method ( I d , 18). 1 2 0 1 ~bdenuni, pre
