Reactive groups in organic reagents and their application in inorganic

Without such a systematization, the search for new compounds suitable for use as analytical reagents might well appear to be a very baffling, if not h...
4 downloads 0 Views 4MB Size
REACTIVE GROUPS in ORGANIC REAGENTS and THEIR APPLICATION in INORGANIC ANALYSIS* LANDON A. SARVER The University of Minnesota, Minneapolis, Minnesota

A

LTHOUGH only a beginning has been made in the use of organic reagents ininorganic analysis, enough has been done t o . demonstrate their great importance and to indicate in a general way what types of new compounds ought to be prepared and tested in order to advance our knowledge in this direction. Without such a systematization, the search for new compounds suitable for use as analytical reagents might well appear to be a very baffling, if not hopeless, task. About one thousand research papers have been publisbed on the subject during the past fifty years, a goodly portion of them during the last ten, and more than three hundred such reagents have been proposed. As a result of this work, it is possible to predict with a fair degree of certainty that certain compounds not previously employed will produce precipitates or colored salts with a given metal. It is of course much more difficult to say that a given new reagent will be specific and react with only one metal. In addition to the salt-forming property, the reagent may also have other functions; e. g., i t may act as an oxidizing or reducing agent, or it may serve as an indicator. In the present discussion, these latter properties will not be considered. From the standpoint of structure, there are two fundamental requirements for a good organic analytical reagent. First, with the exception of a few coordinating compounds, they must contain an acidic or salt-forming group in order to read with a metallic ion (7). The most important of these are: (1) the imino, or =NH group; (2) the oxime, or =NOH gronp; (3) the hydroxyl, or -OH group; and (4) the mercapto, or -SH gronp. The carboxyl, or -COOH group, and the sulfonic acid, or -SOsH group, have been of little importance so far, although they have bden employed. In addition, any compound which is capable of forming one of the above groups by molecular rearrangement must be considered. In each case, the hydrogen of the acidic group is replaced by one equivalent of metal whenever other conditions are favorable to t h e formation of an insoluble or undissociated salt. The second fundamental requirement is that the molecular structure of the reagent must be such as to

make possible the formation of a ring containing the metal (IS), usually by means of one electrovalent link and one coiirdiuate valence, althongh this is not absolutely essential, and either two electrovalences or two coordinate valences may be employed. I t is true that organometallic compounds are known in which there appears to be no chelate ring containing the metal, but from the analytical point of view ring formation seems to be an absolute requirement. In those cases where coordinate valences are required to complete the ring, the metal is most often attached in this way to nitrogen, but also frequently to sulfur or oxygen. In the design and synthesis of a new reagent, Baeyer's strain theory is frequently of great assistance, althongh the laws concerning strain are not yet known with sufficient accuracy to make predictions with absolute certainty. It canbe said, in a qualitative way a t least, that the most stable rings are those containing six members and two double bonds, and those with five members and no double bonds. A six-membered ring with one double bon' should be somewhat more stable than a five-membered ring with one double bond and considerably more stable than a six-membered ring with no double bond. In accordance with the theory, the great majority of known reagents give chelate rings containing five or six atoms; a smaller number give salts with four-medered rings, while three-, seven- and eight-membered rings are very rare. When two or more of the atoms of a chelate ring also form part of an unsaturated aromatic ring, it can be assumed that a double bond is or is not present, according to the requirements for the formation of a stable compound. THE IMINO GROUP

Numerous substances containing the imino, or =NH gronp, are capable of reacting with certain metals, especially silver, mercury, gold, platinum, palladium, and copper. Examples of these-are rhodanine and its derivatives. By the introduction of suitable chromophoric groups into the molecule the resulting salts may be given a color, and the usefulness of the reagents thereby greatly increased. Thus, 5-(9-dimethylamino* Cantrihution to the s ~ m ~ o s i u mon Organic Analytical benza1)rhodanine (8) gives a reddish violet silver salt Reagents conducted by the Division of Physical and Inorganic which is insoluble in dilute acid solutions (Figure 1). Chemistry at the ninety.frrSt meeting of the American Chemical The precipitation of mercury, gold, etc., may be preSociety, Kansas City, April 16. 1936. 511

vented by the presence of cyanide, so that under these conditions the reagent serves as a specific precipitant for silver. It will be noted that the metal becomes part of a four-membered ring (along with nitrogen, carbon, and sulfur) with one double bond; i t is attached to the nitrogen by an ordinary bond, and to the sulfur by a coordinate valence.

Silver salt of 5-(p-dimethylaminobenza1)rhodanine FIGURE 1

It should be remarked in this connection that the nitrogen of pyridine behaves like an imino group, and forms coordination compounds with certain metals (Figure 2). Thus, ferrous iron reacts in a specific way with 2,2'-bipyridine, and 1,lO-phenanthroline to give intensely red complex ions which are very stable in the presence of mineral acids, and against oxidation by the air (4). The very similar 2-(2'-pyridy1)pyrrole and 2-pyridylhydrazine also give colored complex ions with ferrous iron, but these are less stable in the presence of acids and'against oxidation by the air (5).

lmkage both electrons come from the same atom, in this case the nitrogen. THE OXIME GROUP

The oximes have been the most fully investigated of all the organic analytical reagents, and more than sixty are known. They react generally with iron, cobalt, nickel, copper, platinum, palladium, and a few other metals. Indeed, some oximes precipitate almost all the metals, and thus lose their chief value as analytical reagents, since highly specific precipitants are the most to be desired. But certain ones react with a single metal under suitable conditions. Dimethylglyoxime is probably the best known organic analytical reagent, and has been used for separating nickel from cobalt since 1905 (20). Not only this, but other a-dioximes of the same general type (Figure 3) yield colored precipitates, usually red, with nickel salts (15). Either alkyl or aryl groups may replace the R, R', and R" in formulas I, 11, and 111, and the compound may be either symmetrical or unsymmetrical.

H.

H HCJ~\O

II

C He=C/ \CNN\O

1

I

I 1

H*=C\C/C\N7Me II

HI

I

I

I

H-C //C-C\Ni7Me

\c=c/ I I

H H

OH

Fe Ferrous 2-(2'-Pyridy1)pyrrole Ferrous QPyridylhydrazine Ion Metallic Salts of Dioximes and Related Compounds momE 3 C

In each of these examples, the iron forms part of a five-memberedring with no double bondsvery stable configuration, according to the strain theory of Baeyer. They emphasize two fads: first, nitrogen and carbon atoms are interchangeable (cf. 2-pyridylhydrazine), both possessing a tetrahedral symmetry; second, coordinate lmks are no less genuine than ordinary valences, and they are interchangeable with normal ones with regard to the strain produced in the ring. The two kinds of bonds diier only in the f a d that in an ordinary linkage one of the shared electrons is contributed by each of the atoms, while in a coordinate

The hydrogen of one of the oxime groups may be replaced by an organic radical (14), or the entire group may be replaced by an imino group without loss of its characteristic nickel-precipitating properties, as in formulas I1 and 111, respectively; indeed the nitrogen of a pyridine ring may serve just as well as the second group, as in formula V (5, 20). Two aliphatic carbon atoms are present in most of the reagents thus far investigated, but this is not a necessary requirement, and one of them may form part of an unsaturated ring along with one of the nitrogens, as in 2-pyridylformaldoxime (formula V) (5); or both of them may occur

in a saturated ring, as in l,2-cyclohexadione dioxime (formula IV) (22). The interchangeability of carbon and nitrogen atoms is shown again in the nickel salt of nitrosoguanidiue (Formula VI) (21), where there are three nitrogens and one carbou instead of the usual ring with two nitrogens and two carbons. All of these compounds act as monobasic acids and react with only one equivalent of metal, although many of them might be expected to be dibasic. Thus, we see that the salts of all compounds of this class possess a six-membered chelate ring consisting of the metal, one oxygen, and four nitrogen or carbon atoms, with two double bonds. According to PfeiRer (Id), the oxygen does not form part of the ring, in which case there would be only a five-memberedring with two double bonds; however, if we accept the strain theory of Baeyer, we should expect a six-membered ring, since it is much more stable. Compounds which have an oxime and a keto group attached to the same carbon atom react generally with ferrous iron to give deep-blue water-soluble chelate compounds which are soluble in organic solvents (23). In addition, they usually give green colors with ferric iron, and red precipitates with cobalt. The best-known example of this class of substances is l-nitroso-2naphthol, proposed by Ilinsky and von Knorre in 1885 for the separation of cobalt from iron and nickel (Figure 4) (13). These compounds exist in equilibrium with the corresponding oximes, and their salts have chelate rings similar to those obtained with the dioximes, except that one of the nitrogens has been replaced by an oxygen. The carbou atoms may be aliphatic in nature, as in a-benzil monoxime; or they may form part of an aromatic ring, provided that the latter is not completely saturated, as in 1-nitroso-% naphthol, 2-nitroso-1-naphthol, or 2-isonitroso-1-ketotetralin (19). In each case, there is a six-membered chelate ring, with two double bonds. These substances are used principally as reagents for cobalt and iron.

being replaced by an equivalent of the metal, as shown by the behavior of the two isomeric methyl ethers; thus, the hydroxyl and not the oxime is the salt-forming group in this case. Again, the chelate rings consist of six atoms, with two double bonds. The situation is somewhat diierent, however, with a benzoin oxime and its analogs, twenty-two of which have been examined (11). They act as dibasic acids, and give six-membered rings with only one double bond, the divalent metal being linked by two normal covalences instead of by one normal and one coordinate valence, as in the other classes of oximes previously mentioned. Variation of R and R' has no effect upon the insolubility of the cupric salts in water, but does have an influence upon their solubility in ammonia; they appear to be insoluble in the latter when the substituted groups are able to satisfy the coordination valences of the metal. There is no distinction between the symmetric and the unsymmetric compounds. Aromatic radicals cause insolubility in all cases, except where two of them are attached to the same carbon atom as the hydroxyl, as in a-phenylbenzoiu oxime or a-benzylbenzoin oxime. An amino group has the same effect as a phenyl; and compounds having alkyl groups of four or more carbon atoms give copper salts insoluble in ammonia, while those with smaller alkyl groups are soluble. It is interesting to note that 2-pyrrolealdoxime (5,12) also gives an insoluble copper salt having a six-membered chelate ring, with one double bond, where the oxygen atom of the usual hydroxyl group has been replaced by the nitrogen of the pyrrole ring.

Salt of Sslicylaldoxime I

Salt of Salt of an Acyloin Oxime . .%Pyrrolealdaxime .I1 I11 FIGURE 5

.

THE W D R O X Y GROUP

Some hydroxy compounds have already been discussed in connection with the oximesi there are in addition a number of other types, such as 8-quinolinol Salt of Salt of Salt of Monoxime (8-hydroxyquinoline), the hydroxyanthraquinones, the 1-Nitroso-%Naphthol 2-Isonitrosa-1of an Aliphatic ketotetralin Dicarbonyl Compound enolizable ketones and diketones, and the arsonic I I1 111 acids (Figure 6). 8-Quinolinol (1) and its derivaFIOURB 4 tives are unusually active reagents, forming salts with a large number of metals.- They have fiveMonoximes which have a hydroxyl group on the membered chelate rings, with no double bonds, which second or thud carbon atom (Figure 5) act as specific constitute one of the most stable possible configuraprecipitants for copper, giving green salts which are tions. The hydroxyauthraquinones, on the other hand, insoluble in water; most of them are also insoluble in give salts with aluminum, chromium, iron, beryllium, aqueous ammonia. Salicylaldoxime and its analogs, magnesium, calcium, strontium, barium, zirconium, fifteen of which have been examined (6),react as mono- etc., which have six-membered chelate rings, with basic adds, the hydrogen of the hydroxyl group only one double bond, and may be expected to he rather

stable, but less so than the salts of &quinolinol. Enolizable 8-diketones give a highly specific but insensitive reaction with thallium in the presence of carbon disulfide; if only one ketone group is enolizable, a monothallium salt is obtained, but no reaction with carbon disulfide (9). In each case, six-membered chelate rings, with two double bonds, are obtained. The arsonic acids are used as precipitants for zirconium (10, 17) and give four-membered rings; if the metal also links itself with the doubly bound oxygen by a coordinate valence, there is one double bond in the ring; otherwise it has only normal covalences.

-

-

H C-CHS

H C-CHn

0yMe"@%-A II

Salt of Thionalide I

II

Salt of Mercaptoacetanilide I1

Salt of &~uiiolinethiol I1

Salt of Rubeanic Acid IV

Salt of "Phenylthiohydantoic Acid"

v

Salt of 8-Quinolinol (8-Hydroxyquinoline) I

Salt of a Hydroxyanthraquinone I1

It is evident from the foregoing examples that a given organic compound may be expected to prove useful as an analytical reagent if it is able to unite with a metal by virtue of an acidic group (such as an imino, oxime, hydroxyl, mercapto, or carboxyl group), and if it is also possible to form a ring of four, five, or six Carbon Disullide Complex Salt of Phenylarsonic atoms, with or without double bonds. The nature of Acid with a Dithallium Salt of . the non-metallic atoms in the ring appears to be unan Enolizable @-Diketone important, but they are most frequently carbon, 111 IV nitrogen, oxygen, or sulfur. The metal is almost always F~oune6 attached to a nitrogen, oxygen, or sulfur atom by a normal covalence, very rarely directly to carbon. The THE MERCAPTO GROW ring is usually closed by means of a coordinate valence The mercapto group is extremely active, and the from the metal to a nitrogen, oxygen, or sulfur atom, analytical properties of a number of such reagents have rarely to carbon; however, it may also be closed by already been studied; e. g., thionalide (2), mercapto- another normal covalence. I n rare cases two wordinate acetanilide (thioglycollic acid anilide) (3)' S-quino- valences may complete a ring(' While the laws of linethiol (8-mercaptoquinoline), rubeanic acid (16), chemical combination are not yet known with sufficient and "phenylthiohydantoic acid"(24). All of the certainty to make accurate predictions in all cases, compounds mentioned give metallic derivatives having the application of these principles will surely be of five-membered chelate rings, with no double bonds great assistance in the search for new organic analytical reagents. (Figure 7). C

(1) (2) (3) (4) (5) (6)

(7) (8)

(9) (10) (11) (12)

LITERATUB CITED

(13) ILINSKY, M. AND YON KNORRE, G., Ber., 18,699 (1885). BERG,R., Phorm. Ztg., 85,1(1929). P.,, ihid., 63,1811 (1930). BERG,R. AND ROEBLINO, W., Angew. Chem., 48,597 (1935). (14) P P E I ~ E R (15) PONZIO, G., Gazz. chzm. ital., 51,II, 213 (1921). BERSIN, T., Z. anal. Chem., 85,428(1931). BLAU,F., Monalsh., 19,647 (1898). (16) RAY,P. AND RAY, R. M., Quart. J . I d Chem. Soc., 3 , 118 (1926). E ~ E R TB.,. el al., Be?., 60, 2011 (1927); 62, 1733 (1929); 64 1971 (1933). (17) RICE.A. C., Fooo, H.C.,ANDJAMES, C., J. Am. Chem. Soc., 48, 895 (1926). EPHRAIM, F., ibid., 63,1928 (1930); 64,1210 (1931). N. V., "The electronic theory of valence,'' OxFEIGL,.F., "Qualitative Analyse mit Hilfe von T " ~ f ~ l - (18) SIDGWICK, reaktronen," Leipzig, 1935. ford, 1927. FEIGL,F., Z. anal. Chem., 74,380 (1928); Mikrochm.. 9,165 (1905); (1931). Ber., 39,3382 (1906); 41,2219 (1908). E., Monalsh., 49,401 (1928). FEIGL,F. AND BACKER, 273,133 (1893). (21) T ~ ~ E LL., E Ann., , FEIGL,F., KnmHo~z,P., AND RAJMANN, E.,Miklikrockem., 9, (22) wALLACH, 0..;bid., 437,148 (1924). 395 (1931). 23) WHITELY, M. A., J. Chern.Sac., 83.44 (1903). FEIGL,F., SICHER,G., AND SINGER. 0.. Ber., 58,2294(1925). 24) WILLARD, H. H. AND HALL, D., 1.Am. Chem. Soc., 44,2219 (1922). FISCHER, H. AND C s m s , Z.,Ann., 508,172 (1934). .

[a:] ~",~A,";,A~~~~gw&$,",",.~$$~d61p~~ 1