JUNE, 1973 THE ANALYST Vol. 98, No. I 167 Pyridylazonaphthols (PANS) and Pyridylazophenols (PAPS) as Analytical Reagents Part I.* Synthesis and Spectroscopic Examination of Reagents and Some Chelates BY D. BETTERIDGE AND D. JOHN? (Chemistry Department, University College of Swansea, Swansea, Glarnorgan, SA 2 8PP) 2-(2-Pyridylazo)- l-naphthol (o-a-PAN), 2- (2-pyridylazo)phenol (o-PAP) and 4-(2-pyridylazo)phenol ($-PAP) have been prepared. They and 1-(2- pyridylazo)-2-naphthol (O-p-PAN) and their chelates with various transition metals have been examined and characterised by infrared spectroscopy and mass and nuclear magnetic resonance spectrometry. The purity and structures of the reagents have been established and it was confirmed that they are terdentate ligands. A REVIEW by Anderson and Nicklessl has shown that analysts have not been slow to realise the importance of 1-(2-pyridylaz0)-2-naphthol (o-P-PAN).Since Cheng and Bray’s original work,2 there have been many developments with respect to both the use of o-P-PAN and the development of related reagents. These reagents react with many cations to form intensely coloured complexes that are eminently suitable for spectrophotometric determinations, chelatometric end-point detection and solvent extraction. The sensitivities are comparable with those obtained with dithizone (diphenylthiocarbazone) , and the versatility is comparable with that of 8-hydroxyquinoline. The great advantage of o-P-PAN is that its solutions and also solutions of its complexes are unusually stable for such a sensitive reagent.Reagent solutions can be kept for several months without change in the absorbance (Galik,3 and D. Betteridge, unpublished work), which should make o-P-PAN well suited for use in auto- mated determinations. The disadvantage of using PAN-type compounds at present is that their solution chemistry is very complex. Slight alterations of conditions, when modifying a published procedure, can result in irreproducible results or even in no results being obtained. For example, Galik3 has shown that copper is not extracted from a sulphate medium at pH 4 although it is extracted with other common anions from the bulk of the electrolyte. It has also been found that care has to be taken when extracting manganese, although a useful procedure for determining manganese in high-purity zirconium, etc., has been pub- l i ~ h e d .~ We shall show later that this care is necessary because of the formation of a hydroxy species, of the form MnR,OH, which is very dependent upon pH. These difficulties can be easily overcome if the basic chemistry of the system being used is understood. Accordingly, we have undertaken the study of four related compounds and their reactions with metal ions. Two prime considerations have been borne in mind: the chemistry of the system considered should be examined in considerable detail, and the information obtained and the methods used to obtain it should be of value to analysts. The reagents examined are insoluble in water and they form water-insoluble complexes that can often be extracted.Hence the work of Sommer and co-workerss-9 on water-soluble complexes of 4- (2-pyridylazo) resorcinol (PAR) and 4-(2-thiazolylazo)resorcinol (TAR) is to some extent paralleled and the possibilities of solvent extraction are considered more fully. The reagents are 1- (2-pyridylazo)-2-naphthol (o-p-PAN), 2-(2-pyridylazo)-l-naphthol (o-a-PAN), 2-(2-pyridylazo)phenol (o-PAP) and 4- (2-pyridy1azo)phenol ($-PAP). This paper deals with the nature of the reagents and some chelates. Subsequent papers in this series will deal with methods for investigating the equilibria of complex formation and analytical procedures. * For Part I1 of this series, see p. 390. t Present address: BP Chemicals (U.K.) Ltd., Llandarcy, Swansea. @ SAC and the authors. 377378 BETTERIDGE AND JOHN : PYRIDYLAZONAPHTHOLS AND [Analyst, Vol.98 EXPERIMENTAL REAGENTS- PAN-type compounds have been synthesised by the Chichibabinlo reaction of coupling 2-pyridyldiazotate with an appropriate naphthol or phenol under an inert atmosphere. Low yields and ill defined products usually result. Pollard, Nickless and Anderson,ll Anderson and Nickless12 and Andersonl3 have shown that the coupling of 2-hydrazinopyridine, syn- thesised by the procedure of Fargher and Furness,l4 with a suitable quinone results in relatively pure products of definite composition being produced in over-all yields of 40 to 60 per cent. We confirm these findings with the reservation that 2-hydrazinopyridine and o-quinones are usually unstable so that some experimental skill is required in order to obtain the desired results. The final purification stages were very difficult, and the products were subjected to repeated recrystallisation and, when appropriate, sublimation until the microanalyses and mass spectrum indicated that the compound was pure or that further attempts at purification would be valueless.2- (2-PyridyZaxo)$kenoZ (o-PAP)-Aqueous solutions of 2-hydrazinopyridine and 1,2- benzoquinone acidified with perchloric acid were allowed to react together to give o-PAP. The 1,2-benzoquinone was prepared by reaction of catechol in anhydrous diethyl ether with tetra- chloro-o-quinone, which had been synthesised by chlorination of catechol. The 1,2-benzo- quinone is unstable and was used immediately. 2-Hydrazinopyridine was prepared by allow- ing hydrazine hydrate and 2-chloropyridine to react together.'* The solid first isolated was o-PAP perchlorate, m.p.170 "C. Successive recrystallisations of this perchlorate salt from aqueous ethanolic mixtures gave deep red, granular crystals that melted at 128 "C. The lower melting-point indicated that these granular crystals were no longer the salt, but pure o-PAP, which was verified by microanalysis, thin-layer chromatography and infrared spectro- scopy. The microanalytical results were as follows- C H N Calculated for C,,H,N,O.HClO,, per cent.. . . . 44.1 3.4 14-0 Found, per cent. .. .. .. .. . . 66.7 6.2 20.6 Calculated for C,,H,N,O, per cent. .. . . 66.0 4-6 21.0 4- (2-PyridyZazo)$kenoZ (p-PAP)-Commercially available 1,4-benzoquinone was purified by recrystallisation from light petroleum of boiling range 40 to 60 "C and allowed to react with a solution of 2-hydrazinopyridine in perchloric acid.The product was washed with water, dissolved in methanol - formic acid (10 + 6) and ammonia was added to re-precipitate the $-PAP. The microanalytical results were as follows- C H N Calculated for C,,H,N,O, per cent. .. . . 66.0 4.6 2 1.0 Found, per cent. .. .. .. .. . . 66.0 4.6 21.6 2-(2-PyridyZazo)-l-napht~oZ (o-a-PAN)-This reagent was prepared by the reaction of 1,2-naphthoquinone with 2-hydrazinopyridine under acidic conditions- 0 bH In order to obtain a pure sample of 1,2-naphthoquinone, it was synthesised by oxidising 1,2-aminonaphthol hydrochloride with iron( 111) chloride.16 The crude, commercially available 1,2-aminonaphthol hydrochloride was purified as described by Conant and Corson.16 Crude o-a-PAN was precipitated by neutralising the ice-cold acidic reaction mixture with ammonia, but repeated recrystallisations failed to yield a pure product.Purification was effected by sublimation at 122 "C. Mass-spectrometric analysis indicated that there was a slight im- purity (about 1 per cent.) of a compound similar to o-a-PAN but with a higher relative molecular mass (28 a.m.u.). The microanalytical results were as follows- C H N Calculated for C,,H,,N,O, per cent. .. . . 72-3 4.4 16.9 Found, per cent. .. .. .. .. . . 72.1 4.4 16.6June, 19731 PYRIDYLAZOPHENOLS AS ANALYTICAL REAGENTS. PART I 379 1- (2-PyridyZaxo) -2-naphthol (o-P-PA N)-Impure o-P-PAN was obtained commercially and purified by repeated recrystallisations from aqueous ethanol (m.p.140 "C). The micro- analytical results were as follows- C H N Calculated for C,I;HllNIO, per cent. .. . . 72.3 4.4 16.9 Found, per cent. . . .. .. .. . . 72.6 4-3 16.6 4-(2-PyridyZazo)-l-~aphthoZ (o-a-PAN)-This compound was prepared by reaction of 2-hydrazinopyridine with 1,4-naphthoquinone. It could not be purified by the methods outlined above as it formed tars on recrystallisaton and exploded rather than sublimed. No further work was carried out on it. MASS SPECTRA- focusing mass spectrometer. The mass spectra of reagents and chelates were obtained on an A.E.I. MS9 double- INFRARED SPECTRA- The infrared spectra of the solid reagents and chelates were obtained from caesium iodide pressed discs with a Perkin-Elmer 225 grating infrared spectrometer.The caesium iodide was spectroscopically pure. NUCLEAR MAGNETIC RESONANCE SPECTRA- Nuclear magnetic resonance spectra were obtained with a Varian 100 H.A. spectrometer. Spectra of reagents-Saturated solutions of o-a-PAN, o-P-PAN and o-PAP in deuterated chloroform were concentrated enough for spectra to be measured, but it was necessary to use dimethyl sulphoxide as a solvent for $-PAP. Spectra of chelates-Equal volumes (5 ml) of a solution of the reagent (60 mg) in carbon tetrachloride and a buffered aqueous solution of a metal ion (30 mg) were equilibrated. The organic phase was removed and dried over molecular sieves. Other experiments showed that under these conditions of excess of the metal ion at a suitable pH, all of the organic reagent in the organic phase would be converted into the chelate.RESULTS AND DISCUSSION A detailed study of the mass, infrared and nuclear magnetic resonance spectra of the reagents and of some chelates of common transition metals was carried out with several objects in mind- to confirm that no errors of identification had been made; to see if the marked difference between o-a-PAN and o-P-PAN could be explained; to check the generally held views on bonding in the chelates; to provide a reference for the chelates of less common metals to be examined later, which chelates may be anomalous. MASS SPECTRA- A detailed discussion of the mass spectra of the reagents and of their chelates with manganese(II), cobalt (11), nickel(II), copper(I1) and zinc(I1) has been presented elsewhere.17 The two steps in the main fragmentation pattern of the reagents are loss of azo-nitrogen with the fusion of the pyridine and phenol or naphthol groups, and rupture of the compound formed with the release of pyridine and phenol or naphthol fragments.Alternatively, the pyridine group was lost first and the azo-nitrogen second. The 1 : 2 chelates, typically, lost first a whole ligand molecule and then the 1 : 1 complex that remained fragmented to give a metal - pyridine complex and a diazophenol or naphthol group. It was found that o- and +-hydroxyl groups could be detected with certainty so that the identification of the reagents was confirmed. It was also found that the spectra of the chelates provided information that was of value in the interpretation of the results of solution studies.It was possible to check the stoicheiometry of the chelates but not to ascertain their structure.380 BETTERIDGE AND JOHN : PYRIDYLAZONAPHTHOLS AND [Analyst, v0198 INFRARED SPECTROSCOPY- of investigation of chelate compounds.18~19 INFRARED SPECTRA OF REAGENTS- The infrared spectra of the complex molecules are extremely complicated and the extensive overlap of bands makes detailed interpretation very difficult. The assignments for the spectrum of pyridine are well established,20-22 as are those for a-naphthol, p-naphthol and phen01.~3-~6 Examination of the spectra of these individual compounds shows that there are several absorption bands in similar regions. This effect can be anticipated owing to the expected similarity of the ring vibrations and the C-H stretching and bending frequencies of the compounds.It is this type of similarity that leads to the extensive overlapping of bands in the reagent spectra and to the corresponding complexity. Spectrum of 2-(2-pyridyZazo)-l-naphthoZ (o-a-PAN)-The principal bands and their assignments are shown in Table I. In contrast to mass spectrometry, infrared spectroscopy is well established as a means TABLE I INFRARED SPECTRUM OF o-a-PAN Peaklcm-l Intensity* Assignment 3050 m Aromatic C-H stretch 1620 to 1400 s (multiplet) N=N, C=N, C=C stretch 1400 to 110 s (multiplet) C-0 stretch, O-H deformation, pyridine C-H deformation 900 to 400 s (multiplet) C-H in-plane deformations 400 to 200 m (multiplet) C-H out-of-plane deformations * m, bands of medium intensity; and s, bands of strong intensity.990 S Pyridine C-H deformation The spectrum consists principally of aromatic C-H and C=C absorptions, together with CO, OH and CN absorptions. Many of the bands that arise from these absorptions will inevitably overlap and give rise to the high degree of complexity observed. The well defined band at 3050 cm-l is characteristic of aromatic C-H stretching fre- quencies [Fig. l ( a ) ] . Coupled with this band are the expected aromatic C=C ring vibration frequencies a t about 1590, 1500 and 1450 cm-1. The shoulder at 1580 cm-l could be assigned to the presence of conjugated rings, and is often taken as an indication of such a system. The spectral range from 1620 to 1400 cm-l is further complicated by the presence of the C=C and C=N stretching frequencies of the pyridine ring.According to Bassignana and Cogrossi,26 Fig. 1. Infrared spectra: (a) aromatic C-H stretching for o-a-PAN; ((I) 1600 to 1400cm-l region for o-or-PAN; and (c) 1600 to 1400cm-1 region for p-PAPJune, 19731 PYRIDYLAZOPHENOLS AS ANALYTICAL REAGENTS. PART I 381 the N=N stretching frequency could also be observed in the region of 1410 & 30 cm-1. Because of the complexity of this region, positive identification of the band, which is usually of medium intensity, is extremely difficult. Another interesting feature of this region of the spectrum is the occurrence of a well defined shoulder peak at 1610 cm-1, which is also observe4 in the spectrum of o-@-PAN [Fig.l(b)]. In view of the known tautomerism of o-hydroxyazo compounds, it is possible to assign this band to the presence of the C=O stretching frequency of the hydrazo form, it?.---- A20 Although the free carbonyl band is to be H y drazo expected iii the 1700 cm-1 region, such a shift is often observed in hydrogen-bonded systems. - Examination of the spectra of o-PAP and @PAP shows that there is no absorption of any kind at this wavelength [Fig. l(c)]. This result is in keeping with Hadzi’s observation that o-hydroxyazophenols are true azo compounds and that naphthols exist as taut~mers.~’ The strong bands observed in the region of 1100 to 1300 crn-l are probably due to the C-0 stretching and O-H deforination frequencies. Complications in this region are the expected ring vibrations and C-H deformations of the pyridine ring at about 1200 cm-1.At lower frequencies, in the range 1000 to 500 cm-l, are a large number of strong absorptions that can be attributed to the C-H deformations of both the pyridine and naphthol rings. The presence of two strong bands at 990 and 705 cm-l is indicative of the presence of the pyridine ring. Most of the other bands overlap to such an extent that assignment would be extremely speculative. The remaining region of the spectrum below 500 cm-1 consists mainly of the C-H out-of-plane deformations. A notable absence from the spectrum is a band in the region of 3500 cm-1 corresponding to an O-H stretching frequency. It is known that in the presence of strong intramolecular hydrogen bonding this band would become broad and very weak.The absence of this band would therefore suggest that a-PAN has a strong intramolecular hydrogen bond and this is also suggested by the nuclear magnetic resonance spectrum (see below). Spectrum of 1-(2-pyridyZaxo)-2-naphthol! (o-P-PAN)-As expected, the spectrum of o-p-PAN is very similar in nature and complexity to that of o-a-PAN. The identification of individual bands is again difficult because of the high degree of overlap, but the aromatic stretching frequency at 3060 cm-l is well defined. The presence of extensive intramolecular hydrogen bonding is again evident, due to the absence of a band in the 3500 cm-l region, corresponding to the hydroxyl stretching mode. Because of the similarity of the spectra of o-a-PAN and o-P-PAN, a table of frequencies and their assignments for o-@-PAN can be considered to be identical with that for o-a-PAN (Table I).Spectrum of 2- (2-pyridyZazo)phenoZ (o-PA P)-The principal absorption bands and their assignments are listed in Table 11. TABLE I1 INFRARED SPECTRUM OF o-PAP Peak/cm-l Intensity* Assignment 3020 m Aromatic G H stretch 1600 to 1400 s (multiplet) C==C, G N , N=N stretch 1300 to 110 s (multiplet) G O stretch, O-H deformation, pyridine C-H deformation 900 to 500 s (multiplet) C-H in-plane deformations 460 to 200 s (multiplet) C-H out-of-plane deformations 980 S Pyridine C-H deformation * rn, bands of medium intensity; and s, bands of strong intensity.382 BETTERIDGE AND JOHN PYRIDYLAZONAPHTHOLS AND [Analyst, VOl. 98 Many of the absorption bands that arise from the phenol ring will be similar to those of the a- and ,&naphthol rings, so that this spectrum is similar to those of o-a-PAN and- o-p-PAN.The aromatic C-H stretching frequency band at 3020 cm-l is well defined. The aromatic C=C ring vibrations are observed in the region 1600 to 1400 cm-1, which region is again complicated owing to the presence of C=N and N=N bands. The remaining regions of the spectrum are also similar to those of o-a-PAN and o-p-PAN, and consist of bands that arise from C-0, 0-H and C-H stretching and deformation modes. The presence of strong intramolecular hydrogen bonding is indicated by the absence of absorption in the region of 3500 cm-l. In this respect, o-PAP is again similar to o-a-PAN and o-P-PAN, and interpretation is also supported by nuclear magnetic resonance studies.Spectrum of 4-(2-py&’yZaxo)phenoZ (0-PAP)-Below 2400 cm-l, this spectrum is virtually identical with that of o-PAP, but above this frequency there is a big difference. In the region of 3400 to 2400 cm-l there is an extremely broad, very strong absorption band (Fig. 2 ) . This band must be due, in part, to aromatic C-H stretching absorptions, but is so large that some other contributions must be involved. It seems likely that the broadening is caused by intermolecular hydrogen bonds formed by polymeric association of the reagent. This type of association is known to give rise to very broad, concentration-dependent bands in the 3-pm region. Such a change from intramolecular to intermolecular bonding can also be inferred from nuclear magnetic resonance spectra.cm-’ Fig. 2. Intermolecular hydrogen-bond absorption for p-PAP Attempts to investigate the presence of this intermolecular bonding more closely were made by obtaining reagent spectra for saturated solutions in chloroform. The only absorption observed was a well defined band a t 3020 cm-l that corresponded to aromatic C-H stretching. The broad band previously observed in the spectral region of 3400 to 2400 cm-1 was no longer observed. Because of the limited solubility of the reagent in this solvent, spectra were obtained for saturated solutions in acetone. These solutions were prepared by using spectroscopically pure acetone dried over molecular sieves. The spectra were very similar to those obtained for chloroform solutions, but an additional broad, very weak band in the region of 3600 cm-l was also observed.Although acetone was used as a blank solution, this band can be assigned to the hydroxyl stretching frequency only tentatively because of the ease with which acetone picks up moisture. The solutions in both chloroform and acetone were very dilute, and the disappearance of the broad absorption band is consistent with the presence of intermolecular bonds that would be broken on dilution. Although these results tend to confirm the presence of intermolecular bonding, conclusive proof would involve a complete concentration-dependent study. Unfortunately, the relative insolubility of the reagent in non-polar, non-hydroxylic solvents would seem to preclude this.June, 19731 PYRIDYLAZOPHENOLS AS ANALYTICAL REAGENTS.PART I 383 The spectra of the chelates of each reagent are similar and resemble each other more closely than the spectrum of the parent reagent. The general similarity of all the chelate spectra suggests that each metal has a similar effect on the vibrations of the ligand. For convenience, the spectra are divided into four regions, which are discussed separately. 4000 to 2800 cm-l region-A common feature in the spectra of all the chelates, and of each reagent, is the presence of a well defined band in the region of 3050 cm-l. That this band is stable in position in the spectra of both ligands and chelates indicates that it is due solely to aromatic C-H stretching. INFRARED SPECTRA OF THE CHELATE COMPOUNDS- 2800 to 1600 cm-1 region-This region is largely featureless in all the spectra studied.1600 to 900 cm-1 region-This skeletal region possesses a large number of broad, very strong bands in the spectra of both chelates and ligands. It is to be expected that several of the ligand absorption bands that occur in this region will undergo shifts on chelation. Identification of these shifts, normally to lower frequencies, is very difficult because of the large degree of overlapping of bands. A number of workers have f ~ ~ n d ~ ~ s ~ ~ that one of the more prominent shifts observed in this region has been that of the C-0 stretching frequency from about 1200 cm-l to about 1100 cm-l. Although in the present spectra the positive identification of this shift is difficult, a broadening effect near 1100 cm-l in each spectrum suggests that such a shift has occurred.One of the major features of this region is the appearance of a broad, very strong band at about 1340 cm-l in the spectrum of each chelate investigated (see Table 111). TABLE 111 POSITION OF CHELATE -N=N- FREQUENCY* (cm-1) Metal o-a-PAN O- p-PAN O-PAP Ni 1330 (b, s) 1335 (b, s) 1345 (fs, s) Zn 1355 (b, s) 1335 (b, s) 1360 (fs, s) Mn 1330 (b, s) 1330 (b, s) - * b, broad; s, strong; and fs, fairly sharp. I t has been suggested by U e n ~ , ~ ~ in his study of the chelate compounds of o-hydroxy- azobenzene, that this band is due to the shift to lower frequencies of the azo stretching band. 950 to 200 cm-1 regio.n-It is in this region, and in particular the medium to far infrared region, that the metal - oxygen and metal - nitrogen frequencies are likely to be found.These bands are usually well defined, but it is possible that they may be overlapped by other strong absorptions such as those resulting from the C-H deformations of the pyridine or enol ring. METAL - NITROGEN INFRARED FREQUENCIES- The general region expected for metal - nitrogen vibrations is the spectral region from 300 to 150 cm-1.31932 Bands found in this region of the spectra of chelates, and which are not present in the ligand spectra and cannot be assigned to any form of shift, are considered to be metal - nitrogen stretching vibrations. The frequencies are summarised in Table IV. TABLE IV METAL - NITROGEN FREQUENCIES* (cm-l) o-a-PAN O- 8-PAN O-PAP Ni Zn Mn 240 (sh, m) 244 (m) ; 219 (vs, sh) 243 (sh, m); 222 (m) 245 (sh, w) ; 225 (sh) 243 (m) ; 228 (sh, m) 244 (sh, s) ; 222 (sh, s) 247 (sh, w) ; 220 (vs, sh) 243 (s, sh) ; 228 (sh) 243 (sh, m) ; 223 (sh, w) * sh, sharp; w, weak; m, medium; s, strong; and vs, very strong.METAL - OXYGEN INFRARED FREQUENCIES- The highest frequencies are associated with double bonds, while those in the region of 400 to 600 cm-1 are associated with single bonds. The lowest frequencies, below 300 cm-l, arise either with heavy oxide ligands or with co-ordinated oxy-anions. In the present study, two distinct regions of metal - ligand absorptions were observed. Metal - oxygen stretching frequencies occur between 1000 and 250 cm-l.384 BETTERIDGE AND JOHN PYRIDYLAZONAPHTHOLS AND [Analyst, VOl.98 imposed with shiftedv pyridine C-H deformations. TABLE V METAL - OXYGEN FREQUENCIES* IN THE 600 O-E-PAN O- p-PAN Ni 622 (sh, s) 622 (w) Zn 615 (sh, s) 630 (sh, m) Mn 625 (sh, s) 626 (w) 600 to 650 cm-l regi~n-Lecomte~~ suggested that the absorptions may be due solely to metal - oxygen stretching vibrations. However, it has since been shown that this suggestion is an over-simplification, and that the metal - oxygen stretching is probably coupled with ring or C-H deformations (J. M. Williams and D. Betteridge, unpublished work). The bands observed in this range are shown in Table V, but for o-PAP chelates they may be super- TO 650 Cm-’ REGION O-PAP 638 (sh, s) 640 (s) 636 (sh, s) * sh, sharp; w, weak; m, medium; and s, strong. 400 to 500 cm-l wgion-The absorption bands observed in this region are generally accepted as being solely metal - oxygen stretching frequencies.They are listed in Table VI. TABLE VI METAL - OXYGEN FREQUENCIES* IN THE 400 TO 450 cm-1 REGION O-a-PAN O- p-PAN O-PAP Ni 430 (w) 442 (m) ; 429 (m) 485 (sh, m) Zn 423 (w) 437 (sh, s) 478 (sh, m) * sh, sharp; w, weak; in, medium; and s, strong. Mn 420 (w) 438 (sh, s) - Several attempts have been made to correlate metal - ligand frequencies with stability constants or electronegativities. Such a correlation has little value because too few chelates have been examined. NUCLEAR MAGNETIC RESONANCE SPECTROSCOPY- Nuclear magnetic resonance spectroscopy has been applied to the study of chelate compounds on only a relatively few occasion^.^*-^^ The technique is handicapped by the difficulty of obtaining resolvable spectra for those chelates which contain a paramagnetic or ferromagnetic central metal atom.Mass spectrometry and infrared spectroscopy have been found to be more versatile, as these techniques permit the ready investigation of both reagents and chelates. The spectra are complex and those obtained on a 60-MHz instrument could not be resolved. NUCLEAR MAGNETIC RESONANCE SPECTRA OF LIGANDS- and their tau (7) values and assignments are shown in Table VII. 1 - (2-PyridyZaxo)-2-naphthoZ (o-p-PA N)-The proton signals obtained in the spectrum For convenience of discussion the protons are labelled as follows- HO H(3) The hydroxyl proton in o-/3-naphthol is observed at r 3.90. The corresponding peak with o-p-PAN is observed at T -5.70.(The positions of these hydroxyl protons were verified by deuterium oxide exchange.) This large shift can be attributed to strong hydrogen bonding in the reagent molecule. Generally, it is accepted that the greater the extent of hydrogen bonding, the greater is the “hydrogen-bond shift.”37 The hydrogen bonding was proved to be intramolecular, as no signal shift occurred on dilution of the solvent.June, 19731 PYRIDYLAZOPHENOLS AS ANALYTICAL REAGENTS. PART I TABLE VII NUCLEAR MAGNETIC RESONANCE SPECTRUM OF o-P-PAN 385 7 1.59 2-17 2.35, 2-45 2.68 2-93 3.30, 3-40 - 6-70 Peak Integration Complex multiplet 2 Complex multiplet 2 Multiplet 3 Multiplet 1 Doublet 1 Broad multiplet 1 Doublet 1 Assignment f# c (lower is f, upper is c) e, 8 4 5, 6, 7 d 3 OH Apart from the hydroxyl proton signal, all other proton signals are observed in the range 1.0 to 4.0 (Fig.3). The resulting complex spectrum is expected, as signals for aromatic and heteroaromatic protons are usually observed in this range. In order to assign the signals in this range successfully, a scale-expanded spectrum was provided. The interpretation is summarised as follows. 1 .o L l 2.0 1 0 r Fig. 3. Nuclear magnetic resonance spectrum of 0- 8-PAN The maximum effect of the electron-donating properties of the naphthol hydroxyl group is observed for proton 3, which is in an ortho position. The position of proton 3 does not permit any long-range coupling, and as there is only one adjacent proton then the expected signal would be a distinct upfield doublet. Such a doublet is observed at T 3.30 to 3.40, which also integrates for only one proton.The ortho coupling constant for protons 3 and 4 was found to be 10 Hz. In order to locate proton 4, it was therefore required to find a doublet signal with a splitting of 10 Hz. Ideally, this doublet should be a mirror image of doublet 3, but because of the position of proton 4 some long-range coupling from proton 5 would be expected. Such a doublet is observed at r 2.35 to 2.45. The complex multiplet at an average r value of 2-93 corresponds to the proton d, which is the y-pyridine proton. The complex pattern occurs owing to the two ortho couplings of protons e and c and the meta coupling of proton f. The multiplet observed in the region of T 1.60 integrates for two protons. These protons are the cc-pyridine proton f, and the p-pyridine proton c.The low field position of these protons is typical of cc-pyridine protons,% and is due to the effect of the pyridyl nitrogen atom. The low field position of proton c is a result of the proximity of the azo group which, like the pyridyl nitrogen, causes de-shielding effects. The only other pyridine proton to be accounted for is e, which is hidden under the multiplet in the region of r 2.17.386 BETTERIDGE AND JOHN : PYRIDYLAZONAPHTHOLS AND [Analyst, vol. 98 The four remaining naphthol protons, 5, 6, 7 and 8, can be assigned as follows: three of them, 5, 6 and 7, are located in the multiplet at T 2.58, while the other is further downfield at T 2.17 owing to the effect of the azo ErouD.2 - ( 2 - P y r i ~ Z a z o ) - l - n a ~ ~ ~ t ~ o Z (o-a-PAR)LThe features of the spectrum are summarised in Table VIII. TABLE VIII NUCLEAR MAGNETIC RESONANCE SPECTRUM OF o-a-PAN 7 1.67 1.64 2.20 to 2-65 2.98 2.95 to 3.05 3.07 to 3.17 - 6.22 Peak Singlet (much fine structure) Singlet (much fine structure) Complex multiplet Multiplet Doublet Doublet Broad multiplet (weak) Integration 1 1 6 1 1 1 1 Assignment f e, 6. 6, 7, 8 d 3 4 OH c The protons are labelled as follows- OH H(8) For a-naphthol, the hydroxyl proton was observed at T 3-65, whereas for o-&-PAN the signal is observed at T -5.22. This is a clear indication that hydrogen bonding occurs, and that the hydrogen-bond shift is comparable with, but slightly less than, that observed for o-P-PAN.An interesting feature of this signal is that it is much less pronounced than the corresponding signal for o-P-PAN, which may be due to a combination of relaxation, exchange and structural effects.39 The pyridyl and naphthol protons of o-a-PAN are in an environment similar to that in which they occur in o-/I-PAN. As a consequence, the signals are observed at approximately the same positions, the variations being slight. Z-(Z-PyriayEazo)p~enoZ (o-PAP)-The signals of the spectrum obtained for this compound are listed in Table IX. TABLE IX NUCLEAR MAGNETIC RESONANCE SPECTRUM OF o-PAP 7 Peak Integration Assignment 1.26 to 1.30 Doublet with much splitting 1 f 1-94 to 2.00 Doublet (fine structure) 1 G 2.07 to 2.20 Multiplet 2 el 3 2-60 to 2.70 Multiplet 2 dl 6 2-82 to 3.00 Multiplet 2 4, 6 - 2-86 Broad singlet 1 OH The protons are labelled as follows- OH The intramolecular hydrogen bonding expected to be present in this reagent is confirmed by the large hydrogen-bond shift of the hydroxyl proton to r -2.86.June, 19731 PYRIDYLAZOPHENOLS AS ANALYTICAL REAGENTS.PART I 387 A significant feature of this spectrum is that the signal of the pyridine proton f is well removed from the others. It occurs as a doublet at T 1-26 to 1.30 with much fine structure due to ortko, meta and para coupling of the protons e, d and c, respectively. Another feature of the spectrum is the decreased de-shielding of proton c, which is located as a doublet at T 1-94 to 2.00. This decreased de-shielding is due to the increased aromaticity of the phenol ring and the consequent loss of bond fixation.The increased aromaticity is further reflected in the position of proton 3. This proton, which is in conjugation with the azo group, is in a similar environment to proton c and the corresponding signal is contained in the multiplet in the range T 2.07 to 2.20. It is likely that the signal for proton 3 contributes to the central largest peak at T 2-14. The remainder of this multiplet, which integrates for two protons, consists of the triplet of the pyridine proton e. The remaining y-pyridine proton d is observed as the usual multiplet in the range T 2-50 to 2-70. Also contained in this multiplet is the phenol proton 5 , in which the triplet peaks have coupling constants matched by protons 4 and 6. The electron-donating effect of the phenol hydroxyl group should be noticeable at both the o- and $-positions of the phenol ring.As a result, both protons 4 and 6 should be shielded, the effect being greatest for proton 6. This proton occurs as a doublet at T 2.91 to 2.99, the coupling constant being 8 Hz, which is matched for proton 5 . This doublet is part of a multiplet, the remainder of which is the expected triplet of proton 4. 4- (2-Pyridylaxo)$henol (p-PA P)-The signals observed in the spectrum are summarised in Table X. TABLE X NUCLEAR MAGNETIC RESONANCE SPECTRUM OF @-PAP 7 Peak Integration Assignment 2.95 to 3.04 Doublet 2 2, 6 2.07 to 2.16 Doublet 2 3, 6 1.91 to 2.66 Multiplet 3 c, d , e 1.30 to 1-36 Highly split doublet 1 f - 0.46 Singlet 1 OH The protons are labelled as follows- The spectrum of this reagent is less complicated than those of the reagents already discussed because of the equivalence of the phenol protons (Fig.4). The very strong upfield doublet at T 2.95 and 3.04 integrates for two protons, and can be assigned to protons 2 and 6. These protons are ortho to the hydroxyl group, and are equivalent, thus giving rise to identical signals. These two protons are split by the two equivalent protons 3 and 6, which are located by the powerful doublet at T 2.07 to 2.16. This doublet is the mirror image of the first doublet and has the same splitting of 9 Hz. The pyridyl protons have the usual type of signals in the expected regions. Proton f has the downfield multiplet expected. The other pyridine protons, e, d and c, are located in the highly complex series of signals in the range T 1.9 to 2.6.One of the noticeable features of the spectrum is the occurrence of a sharp singlet signal a t T -0.46, which can be assigned to the hydroxyl proton (Fig. 4). The position and shape of this signal suggests that the introduction of a para-hydroxyl group leads to a decrease in hydrogen bonding. This evidence, coupled with the results of the infrared studies, indicates that the decreased hydrogen-bond shift is due to the presence of intermolecular hydrogen bond- ing, in contrast with the intramolecular bonding in the other reagents. NUCLEAR MAGNETIC RESONANCE SPECTRA OF CHELATES- The spectra of the nickel(I1) and zinc(I1) chelates of o-N-PAN, o-P-PAN and o-PAP have been obtained. These spectra were found to be very different from those of the parent388 BETTERIDGE AND JOHN : PYRIDYLAZONAPHTHOLS AND [Analyst, VOl.98 ligands. The spectra of the nickel chelates were very poor, with small, ill defined signal peaks. The spectra of the zinc chelates were much clearer, with strong, well defined signals. Unlike the mass and infrared spectra, similar patterns were not observed with the different metals, but the three zinc spectra were very similar, as were the nickel spectra. Detailed interpretation of these spectra is exceedingly complicated, as the review of the nuclear magnetic resonance characteristics of paramagnetic molecules by Eaton and Phillips3s clearly indicates, and was not attempted. I 4 I 7 Fig. 4. Nuclear magnetic resonance spectrum of +-PAP GENERAL CONCLUSIONS The foregoing discussions show that these three spectroscopic techniques can contribute greatly to the understanding of the chemistry of chelates.In addition, the methods can be used to confirm many of the conclusions obtained from solution studies, and as such are useful complementary techniques. It seems that mass spectrometry and infrared spectroscopy are the most useful, as nuclear magnetic resonance spectrometry is somewhat restricted. It is clear that the compounds have now been correctly defined and that the compound previously reported4* as 4-(2-pyridylazo)-l-naphthol ($-%-PAN) is in fact 2-(2-pyridylazo)- 1-naphthol. However, these techniques do not provide a ready explanation for the large difference in pK values between o-a-PAN and o-/?-PAN, which led to the initial mis-identi- fication.The infrared and nuclear magnetic resonance spectra both indicate strong intra- molecular hydrogen bonding of comparable strength. I t is clear from the infrared spectra that the reagents act as terdentate ligands so that the basic stereochemistry of the chelates is fixed. There are, however, eight basic geometrical isomers of the reagent, depending on whether the ring systems are cis or trans to the diazo-nitrogen group and the relative positions of the pyridine-nitrogen and the hydroxyl group. With the aid of models (Prentice-Hall Framework Molecular Models), it is possible to envisage that two of these eight positions, by virtue of hydrogen bonding, will be much more favoured than the others, but that one might result in a stronger hydrogen bond than the other.It is therefore possible that the difference in the pK values is explained by the geometrical configuration of the reagent. The spectroscopic techniques applied do not give any information about this possibility, which will have to be checked by an X-ray determination or by a full equilibrium study in order to isolate the relative contributions of the enthalpy and entropy terms to the free energy change. We are grateful to the T. and E. Williams Scholarship Fund for a maintenance grant to one of us (D.J.).June, 19731 PYRIDYLAZOPHENOLS AS ANALYTICAL REAGENTS. PART I REFERENCES 389 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31.32. 33. 34. 35. 36. 37. 38. 39. 40. Anderson, R. G., and Nickless, G., Analyst, 1967, 92, 207. Cheng, K. L., and Bray, R. H., Analyt. Chem., 1955, 27, 782. Galik, A., Talanta, 1969, 16, 201. Donaldson, E. M., and Inman, N. R., Talanta, 1966, 13, 489. Chromy, V., and Sommer, L., Ibid., 1967, 14, 393. Sommer, L., and Hnilickova, M., Ibid., 1969, 16, 83. Sommer, L., and Ivanov, V. M., Ibid., 1967, 14, 171. Sommer, L., Ivanov, V. M., and Novotna, H., Ibid., 1967, 14, 329. Sommer, L., and Novotna, H., Ibid., 1967, 14, 457. Chichibabin, A. E., Zh. Russk. Fiz.-khim. Obshch., 1920, 50, 512. Pollard, F. H., Nickless, G., and Anderson, R. G., Tulanta, 1966, 13, 725. Anderson, R. G., and Nickless, G., Proc. Soc. Analyt. Chem., 1966, 3, 149. Anderson, R. G., Ph.D. Thesis, University of Bristol, 1966. Fargher, R. G., and Furness, R., J . Chem. Soc., 1915, 107, 691. Fieser, L. F., Org. Synth., 1943, Collect. Vol. 11, 430. Conant, J . B., and Corson, B. B., Ibid., 1943, Collect. Vol. 11, 33. Betteridge, D., and John, D., Talanta, 1968, 15, 1227. Nakamoto, K., “Infrared Spectra of Inorganic and Co-ordination Compounds,” John Wiley, New Adams, D. M., “Metal - Ligand and Related Vibrations,” Edward Arnold (Publishers) Ltd., Kline, C. H., and Turkevich, J., J . Chem. Phys., 1944, 12, 300. Corssin, L., Fax, B. J., and Lord, R. C., Ibid., 1953, 21, 1170. Wimshurst, J . K., and Bernstein, H. J., Can. J . Chem., 1957, 35, 1185. Hunsberger, I. M., J . Amer. Chem. SOC., 1950, 72, 5626. Friedel, R. A., Ibid., 1951, 73, 2881. Barnes, R. B., Liddell, U., and Williams, V. Z., Ind. Engng Chem., Analyt. Edn, 1943, 15, 659. Bassignana, R., and Cogrossi, C., Tetrahedron, 1964, 20, 2361. Hadzi, D., J. Chem. SOC., 1956, 2143. Ueno, K., and Martell, A. E., J . Phys. Chem.. 1955, 59, 998. Charles, R. C., Freiser, H., Friedel, R., Hilliard, L. E., and Johnston, W. D., Spectrochim. Acta, Ueno, K., J . Amer. Chem. Sac., 1957, 79, 3066. Baldwin, D. A., Lever, A. B. P., and Parish, R. V., Inorg. Chem., 1969, 8, 107. Gill, N. S., Nuttall, R. H., Scaife, D. E., and Sharp, D. W. A, J . Inorg. Nucl. Chem., 1961, 18, 79. Lecomte, J., Discuss. Furaday SOC., 1950, 9, 125. Day, R. J., and Reilley, C. N., Analyt. Chem., 1964, 36, 1073. Eaton, D. R., and Phillips, W. D., in Waugh, J. S., Editor, “Advances in Magnetic Resonance,” Sawyer, D. T., and Smith, B. B., Inorg. Chem., 1960, 8, 379. --- , “High Resolution Nuclear Magnetic Resonance,” McGraw-Hill, New York, 1959, Bernstein, H. J., Pople, J. A., and Schneider, W. G., Can. J . Chem., 1957, 35, 1487. Emsley, J. W., Feeney, J., and Sutcliffe, L. H., “High Resolution Nuclear Magnetic Resonance,” Betteridge, D., Todd, P. K., Fernando, Q., and Freiser, H., Anulyf. Chem., 1963, 35, 729. York, 1963, p. 216. London, 1967. 1956, 8, 1. Academic Press, London, 1965, p. 103. pp. 400-421. Volume 1, Pcrgamon Press, Oxford, 1965, p. 534. Rcceived A$ril2nd, 1970 Accepted November 22nd, 1972