Electrochemical determinations of diffusion coefficients of some polynuclear copper(II) and iron(iii)complexes

Anulyricu Clrhlcci Acta Elsevier Publishing Company. Printed in The Netherlands

ELECTROCHEMICAL OF COEFFICIENTS COMPLEXES EDGAR Department

483

Amstcrdnm

N. DRAKE

DETERMINATIONS SOME POLYNUCLEAR

II* ANU JERRY

of’ Chemistry,

(Received 20th November

Texas

OF DIFFUSION COPPER(H) AND

IRON(II1)

L. JONES**

A & M University.

College

Station.

Te.ras

(U.S.d

.)

1970)

The Stokes-Einstein equation’ has been shown to’be a useful tool for estimating sizes of diffusing particles which are significantly larger than the solvent molecules in which they are dissolved. Use of the relationship does depend, however, on a knowledge of the diffusion coefficient of the particle. Electrochemical techniques afford ‘one means of measuring diffusion coefficients provided that the diffusing species can be identified. Often, by adjustment of pH or ligand concentration, a single complex ion can be made to predominate in solution and a measured diffusion coeffi,cient can be reliably attributed to a particular ion. While ligand substitution reactions seldom produce dramatic changes in the diffusion coefficient of a complex ion2, dimerization (or more generally polymerization) reactions’can change diffusion coefficients by 100°A or more 3. Changes of this magnitude’can be observed easily by using chronopotenticmetry or conventional d.c. polarography. Recent studies of the citrate, malate, and tartrate complexes of copper(I1) indicate that, these .ions polymerize in alkaline solutions4. The work of Timberlake’ indicates that the monomeric iron(II1) tartrate complex polymerizes in aqueous solution as the pH is increased. The four species reported to exist in acidic solutions with pH values less than 5.0 are: the monomer, FeC4H406+ ; the dimer, Fe2(0H)JC4H&O&; the d imer, .Fe,(OH),(C4H,06)(C4H406); and the trimer, Fe3(OH),(C4H,0&:. By adjustment of the pH, one of these species can be made to predominate ‘in solution. The present investigation was undertaken to measure the diffusion coeflicients of some of the polynuclear copper(I1) and iron(II1) complexes and to estimate the sizes of the particles in solution. EXPERIMENTAL

Apparatus A conventional chronopotentiometric apparatus similar to that described previously” was employed. Chronopotentiograms were obtained with a Hewlett-

* Present address’: Department of Chemistry, Angelo State College, San Angelo, Texas (U.S.A.). **. Present address: Department of Chemistry, Central Washington State Collcgc, Ellensburg, Wash. (U.S.A.). Correspondence should be sent to Dr. Jones. An& C/tint. Acta,

54 (197 1) 483-488

E. N.

484

DRAKE,

II, J. L. JO’NES

Packard Model 130A oscilloscope equipped with a Hewlett-Packard Model 196B camera. Current values were obtained by measuring the voltage drop across a precision 200-ohm resistor ( ~O.OlO/O). Conventional d.c. .polarograms ‘were obtained with a Sargent Model XXI polarograph. Mercury drops from a conventional dropping mercury electrode were caught in a-glass spoon and transferred to a recessed, amalgamated platinum wire sealed in glass tubing to form a hanging mercury drop electrode. The resulting electrode surface area, calculated from the average weight of several drops, was 0.0280 4 0.0003 cm2. Materials und methods All chemicals were reagent grade with the exception of the primary standardgrade disodium (ethylenedinitrilo)tetraacetate dihydrate (EDTA). A 0.0100 M stock solution of copper(U) was prepared with copper sulfate pentahydrate. The solution was standardized by compleximetric titration with 0.01048 M EDTA and a O-l”/, solution of 1-(2-pyridylazo)-2-naphthol (PAN) as indicator’. Solutions containing copper(H) and tartrate were prepared by adding 6.10 g of sodium perchlorate (anhydrous~ and 1.50 g of D-tartaric acid to 50.00 ml of stock copper(B) solution and diluting to 500.0 ml with distilled water. The pH was adjusted to 9.1 with 6 M sodium hydroxide. Solutions containing.copper(II) and citrate were prepared by adding 6.10 g of sodium perchlorate,and 3.00 g of sodium citrate dihydrate to 50.00 ml of stock copper(H) solution and diluting to SOO.Oml. The pH was adjusted to 9.2 with 6 M sodium hydroxide. Sdiutions of yellow copper(I1) salicylate, CuC;H,Og, were prepared by adding 138 mg of salicylic acid and 6.1. g of sodium perchlorate to 50.00 ml of stock copper(H) solution and diluting to 500.0 with distilled water. The pH was adjusted to 5.5 with 6 1Msodium hydroxide.:Precipitation occurred above pH 6. Finally, solutions of green copper(I1) 5sulfosalicylate in the form of Cu(C7H403S03)swere prepared by adding ,l.OO g of S-sulfosalicylic acid to 6.1 g of sodium perchlorate and 50.00 ml of stock (0.0211 M) copper(I1) solution and diluting to 500.0 ml with distilled water. The, pH, was adjusted to 6.4 with 6 A4 sodium hydroxide which assured solubility of the ,5-sulfosalicylic acid. Iron(II1) tartrate solutions were prepared by’adding SO&O ml of stock iron(II1) perchlorate solution to 6.,1 g of anhydrous sodium perehlorate and’2.50 g of D-tartaric acid. The solution was diluted with distilled water and the pH was adjusted to 5.0 or 3.2.with 6.N sodium,hydro~de. The stock solution was 0.0112 M iron(III),in 0.100 M perchioric acid.and was standardized by direct titration with.0.01048 M EDTA and an aqueous 1% soltition of N-(p-methoxjphenyl)-p-phenylenediamine hydrochloride (varia~ine blue) as indicators. All solutions were purged with prepurifled nitrogen (minimum‘purity’991996 “/o) prior to electrolysis. Chronopotentiograms were obtained at current densities ranging from 0.98 to 7.&n& cmB2 in the ease copper(I1) s,olutions.and,from 0.36 to 3.8 mA crnw2, in the iroln(If1) solutions. With copper 5-sulfosalicylate solutions, it was necessary to prepare a new hanging mercury drop electrode after each chronopotentiogram in order to obtain reproducibfe results. Although the currents remained constant for long peri,ods of time, the current settings used were recalibrated daily. .. Conventional d.c. ,‘polarographic reductions of all solutions were obser&d and all voltages were measured with respect to the saturated calomel electrode. r ’

of

Anal; Chitn. Adal

54 (1971) 483488

DIFFUSION RESULTS

COEFFICIENTS AND

OF POLYNUCLEAR

485

COMPLEXES

DISCUSSION

(II) Only a single polarographic reduction corresponding to the transition Cu(II)Cu(0) was exhibited. Half-wave potentials for the copper(U)-tartrate system at pH 9.1 and the copper (II)&itrate system at pH 9.2 were - 0.29 and - 0.40 V, respectively. For the copper(I1) salicylate solution at pH 5.5 the half-wave potential was -0.05 V while the solution containing copper(I1) and S-sulfosalicylic acid at pH 6.4 exhibited a half-wave potential of -0.15 V. In the case of the last-mentioned solution, very large maxima rendered the polarographic limiting current unsuitable for use in calculating’diffusion coefficients. Meitesg reports two reductions for the copper(II)-citrate system at pH 9.2 with the Cu(II)-Cu(1) transition at -0.20 V and the Cu(I)-Cu(O) transition at - 0.40 V. The same author lo reports a single reduction corresponding to Cu(II)-Cu(0) for the copper(tartrate system at pH 9.0 with a half-wave potential of -0.285 V. In general, it appears that two reduction waves are observed for copper(I1) in ammoniacal solutions while single waves are obtained in solutions made basic with sodium hydroxide’ i. The diffusion coefficients of several copper(I1) complexes calculated from polarographic data are presented in Table I. These were obtained by the use of the Ilkovie equation and po@rographic limiting currents which had been corrected by subtraction of the residual current. Two one-electron reductions were observed for all copper(I1) solutions studied chronopotentiometrically. The potentials corresponding to one-fourth of the tran-0.20 and -0.60 V sition time were - 0.28 and - 0.52 V for the copper(tartrate, for the copper(II)-citrate, -0.16 and -0.44 V for the copper(U)-5-sulfosalicylate, systems. . and - 0.10 and -0.38 V for the copper(salicylate Before transition times may be used to calculate diffusion coeflicients for complex species, the product it* must be shown to be independent of current density to establish that the’complex species in question is not undergoing dissociation before reduction12. Such was the case for the copper(I1) systems studied and-typical chronopotentiometric data’obtained ‘for the copper(5-sulfosalicylate complex are compiled in Table’II. Transition times observed for all complexes studied were of the order of a fraction of a second. The chronopotentiogram’of the copper(tartrate complex shown in Fig. 1 illustrates the t\;vo reductions typically observed for the copper(I1) complexes. Diffusion coeflicients were ‘calculated from chronopotentiometric data by means of the Sand equation ’ 3 anh these’are presented in Table I.

Copper

‘TABLE DIFFUSION

1 COEFFICIENTS

OF SOME COPPER(H)

Species

Cu~(0%GH&b)f-

Cu,(OWdGH&‘d: CuC;H@: Cb(C,H,O,SO,)$a All diffusion

-

COMPLEXES”

Chronopotentiometric (x105)

Polurographic

0.36 -t_0.03 0.33 f0.03 0.44 f 0.03 0.30 * 0.03

0.38 f 0.02 0.33 4 0.02 0.47 f 0.U’

coefIicients are expressed

(x105)

in cm’ set- * and were measured Atd

at 25’. Chim. Acts, 54 (1971) 483-48’8

486

E. N. DRAKE

TABLE

‘II,

J. L..JONES

11

CI~IRONOI’O1’ENTIOMP.TRIC ------.

Transition (SL’C)

time

DATA

FOR THE COPIDBR(II)

Currenl

OF 5-SULFOSALICYLIC

ACID

it+ (/I sect

(A.105)

0.062 0.078 0.092 0.108

COMPLEX

3.40 3.15 2.90 2.70

* 105)

0.88 0.88 0.88 0.88

E(V) -1.6-

Fig.

1. Chronopotcntiogram

Fig. 2. Chronopot,entiogram

?f 7.03.

IO-’

of 3.7 * 10 -’

:

M Cu,(,OH),(C,H,O&M Fe3(0H),(CbH40&-

at pl-1 9.1. in 0.100

A4 NaClO_,

at pH 5.0.

Stokes-Einstein equation was used to compute approximate diameters of the electroactive species ,present in each solution. The viscosity of the solution ,;was ass,u,med to be the same as that for pure water or 8.94 - 1,0-3 poise. The particle diameter calculated for the’copper(II)+artrate complex at pH 9.9 was l.4 nm. Cu(O@(CiH506): should be a compouent of this system at @I+ 9.1. Howeyer, Rajan and Martell hive demonstrated that dimeriaation can occur, at.considerably lower pH values.and they imbly that even further polymerization most likely occurs near pH 9. Scale_models indicate that the diameter, of the Idimer,’ ‘Cuz(dH)z’(C4H406)22-I: is,’ 1.0 nm *bile that of the trimer, Cu,(OH),(C,H~O,)~‘~ , is 1.4 nm. It is therefore suggested that there exists a trimeric copper(R)-tartrate species at pH 9.1. Similar models constructed”for the copper(II)-citrate complex indicate that the dimer, CU~(OH)~(C~H~O,)$T, would be expected to have a diameter of 1.2 nm. The diameter calculated from the diffusion coefftcient of this’ion- is 1.2 nm. Thus, the dimer appears to be the predominant species present in copper(I1) solutions containing citrate at pH 9.2. Such a conclusion is ,compatible with the findings of Rajan and Martell. The salicylate and 3-sulfosalicylate complexes of: copper(I1) were studied to determine the difference in size observed when first’ one, ‘and then two,,large .oiganiti, ligands are attached to the metal ion. ,The c,opper(II)-salicylate complex containmg .dopper(II) and salicylic acid in a mole ratio of .I_: 1 was prepared by reacting copper and The'

rlhal. Chim. nctci,

54 (1971

j 483-488

DIFFUSION

COEFFICIENTS

OF POLYNUCLEAR

487

COMPLEXES

salicylic acid in stoichiometric amounts at pH 5.5. The particle diameter, calculated from measured diffusion coefficients, was 1.0 nm. The complex containing copper and 5-sulfosalicylate ‘in a ratio of 1 :2 was prepared by reacting ‘copper(I1) with an’ approximately four-fold ‘excess of 5-sulfosalicylic acid at pH 6.4. The particle diameter obtained from diffusion coefIicient calculations for this complex was 1.6 nm. The observed increase of 0.6 nm is very near that expected for the addition of a second ligand to the copper(M) complex. Iron(II1) tartrate solutions at pH 5.0 exhibited two reductions polarographitally with half-wave potentials of -0.13 and - 1.38 V. Reductions of iron(lII),tartrate complexes in basic solutions have been reported ’ ’ at -0.10 and - 1.38 V. These reductions correspond to the transitions Fe(III)-Fe(I1) and Fe(II)-Fe(O), respectively. ’ At pH 3.2, the one-electron reduction wasobserved at - 0.13 V but the two-electron. reduction was obscured by hydrogen evolution which began at - 1.3 V. The trimer is the only species present in appreciable amounts at pH 5.0. Its diffusion coefftcient was calculated by. means of the IlkoviLS equation and the one- and two-electron reduction waves. Limiting currents were corrected by subtraction of the residual current. The resulting diffusion coefficients are shown in Table III. ‘A typical chronopotentiogram of the iron(II1) tartrate system at pH 5.0 is shown in Fig. 2. The slight change of slope on the rising portion of the curve is attributed to the one-electron reduction, the transition time of which is small in comparison with that of the two-electron step. Both the dimer (pH 3.2) and the trimer (pH 5.0) were found to have it, products which were independent of current, indicating that these complexes are reduced without undergoing dissociation. The Sand equation was used to calculate diffusion coefficients from measured transition times. In the case of the, dimer at pH 3.2, no transition time was observed at - 1.38 V, but the much shorter transition time at -0.13 V was used to compute the diffusion coefficient. The resulting diffusion coefficient, appears in Table III.’ The Stokes-Einstein equation was used to calculate the approximate sizes of the diffusing particles. The viscosity of the solution was taken to be the same as that for pure water.‘These calculations give approximate diameters for the dimer and for the trimer of 0.8 and 1.4 nm, respectively. Timberlake’ describes the iron(II1) tartrate trimer as monomer units connected by hydroxyl bridges. For the ion to have a triple negative charge as has been proposed, six hydroxyl groups must, be present. Construction of a model of the triply negatively charged trimer on a scale of 0.1 nm per inch shows that the ion approximates a sphere TABLE DIFFUSION

III COEFFl&ENTS

01: IRON(II1) _-.

Species

~hronopotentionretric

Dimer Trimer

(0.68f0.05)~ 10-5 (0.35 i-0.05) ’ 10-S

u All diffusion

coeffkients

,TARTRATE COMPI.EXES” __----

are expressed

Polarographic

(0.35f0.02)*10-5 (-0.13 V) (0.36+0.02). 1O-5 (- 1.3&V) in cm2 set- ’ and were measured

. at 29.

And. Chim. Actu, 54 (1971) 483-488

488

E. N. DRAKE,

II, J. L. JONES

in shape which would give ,minimal viscous drag in solution. The model measures 1.5 nm in diameter which closely agrees with the 1.4-nm diameter calculated from lneasured diffusion coefficients. The model also indicates that quadridentate chelation by the tartrate, as suggested by Timberlake, is sufficiently difficult as virtually to ,preclude the possibility. This work was conducted in partial fulfilment Ph.D. degree at Texas A&M University (E.N.D. II).

of the requirements

for the

SUMMARY

Electrochemical data obtained by simple techniques appear to provide a valid means of detecting polymerization of metal complexes in aqueous solutions. Diffusion coefficients have been determined for the dimeric and trimeric iron(III) tartrate complexes and for ssome tartrate, citrate, salicylate, and 5-sulfosalicylate complexes ,of copper(B) by :means of chronopotentiometry and conventional de. polarography. Appr,oximate particle diameters were ctilculated from measured diffusion coefficients by the Stokes-Einstein equation and these confirm the presence of polynuclear speciesreported by earlier investigators. Scale models of the complexes have diameters which agree closely with those,obtained experimentally. Evidence for the existence of a trimeric copper(B) tartrate species .at pH 9.1 is presented. ZUSAMMJiNFASSUNG :

,,

Durch einfache Verfahren erhaltene elektrochemische Daten scheinen ein beweiskr%ftiges Mittel fur den Nachweis einer Polymerisation von Metallkomplexen in wassrigen Liisungen zu sein. Durch Chronopotentiometrie und’ konventionelle Gleichstrompolarographie wurden Diffusionskoeffizientenl. :ftir die dimeren und trimeren Eisen(III)-Tartrat-Komplexe und fur einige.Tartrat-, Citrat-, Salicylat- und 5-Sulfosalicylat-Komplexe von Kupfer(I1) bestimmt. Aus den gemessenen .Diffusionskoeffiiienten wurden nach der Stokes-Einstein-Gleichung die ungefghren Teilchendurchmesser berechnet ; sie bestatigen das Vorliegen mehrkerniger Spezies, iiber die,von friiheren Autoren berichtet worden ist. Schalenmodelle der ,Komplexe haben Durchmesser, die mit den experimentell erhaltenen gut. tibereinstimmen. Es wird der Beweis fur die Existenz einer trimeren Kupfer(II)-tartrat-Spezies bei pH 9.1 vorgelegt, RtiFEdtiNCES 1 2 3 4 S 6 7 8

L. Mnnks,

Polarographic Techniques, 2nd Edn., Interscicnce, New York, 196S.*p. 141. MEITES, ref. 1, p. 144. rlJ. DRAKE, II AND J. L. JON=, unpublished results. S. R~JAN AND A. 8: MAKTELL, J. In&g. Nucl. &em.; 29 (1967) 463. F. TIMREKLAKE, J. Chetk Sot., (1964) 12.29. P. DELAHAY, Discussions Faraday SW., 17 (1954) 205. H. FLA$CHKA ANP H. ABDINE. Chemisf-Analyst, 45 (1956) 58. G, Sc~l~V~lt~ENL{r\CX, Comnplexonwrric Tirrtrtiows. Interscicncc. N& York, 1957, P. 78. 9 I.,. MUTES; ‘J. Amer. C/tern., Sot., 72 (1930) 180. 10 L. MIZITEB,J. Amer. Chem. Soc.!,72 (1949) 3269. 11 Ref:l, pp. ‘661-3. . : 12 pi DELAHAY ANDT.'BERZINS,J. Amer. C/tern. Soc.,7S (1933)2486. 13 H. J. S. SAND, Phil. Msg., 1 (1901)4S. I Anal; Chim. Acta, 54 (197 1) 483-488 ,L. E. ‘K. C.

‘. .