Complexes of Ammonia and Ethylenediamine with CuI' on Zeolite A BY ROBERT A. SCHOONHEYDT," PAUL PEIGNEUR AND JAN B. UYTTERHOEVEN Centrum voor Oppervlaktescheikunde en Colloidale Scheikunde, Katholieke Universiteit Leuven, De Croylaan 42, B-3030 Heverlee, Belgium Received 15th December, 1977 Adsorption of NH3 on a dehydrated CuA zeolite gives Cu(NH3)$+ and a small amount of tetrahedrally coordinated CuII ions. These complexes are unstable and decompose by vacuum heating in the range 295-373 K to yield lattice-bonded CuII ions characterized by an e.p.r. signal with gll = 2.3087, All = 0.015 09 cm-l, g1 = 2.067 and A L = 0.000 96 cm-'. Above 373 K this signal transforms to gll = 2.3715, All = 0.013 63 cm-l, g1 = 2.062 and A 1 = 0.001 92 cm-'. Adsorption of ethylenediamine on a dehydrated CuA zeolite gives only Cu(en)", stable up to 473 K.However, in the presence of water, Cu(en):+ and Cu(en)$+ can also be synthesized on the surface. The latter two complexes are unstable and decompose by vacuum heating to the mono-complex. Cu(en)'+ can also be loaded on NaA by ion-exchange from aqueous solution, but Cu(en)a+ cannot, due to space requirements and the instability of the bis-complex in the supercages of zeolite A. First it is necessary to exchange the mono-complex and to synthesize the bis-complex in situ by adding large excesses of ethylenediamine to the aqueous suspension. The e.p.r. and reflectance spectra of the various complexes depend slightly but consistently on the environment in the supercages of zeolite A. Their analysis, however, shows that this does not induce major changes in the CuII-N bonding characteristics. The complex formation of Cu" with ethylenediamine (= en) on montmorillonite is strongly favoured with respect to aqueous s01ution.l'~ The absorption band maximum of Cu(en)i+ on these minerals is situated around 20 000 cm-l, 2000 cm-l above its band maximum in aqueous solution.The corresponding gain in crystal field stabilization energy (C.F.S.E.) is 17-25 kJ mol-l .4 On faujasite-type zeolites Cu(en)l+ is destabilized. The degree of destabilization depends on the cation exchange capacity (C.E.C.) of the mineral and the loading with Cu(en)g+.l* Thus, Y zeolites can be loaded with Cu(en)g+ up to -70 % of the C.E.C. without decom- position of the complex. On X-type zeolites the exchange with Cu(en)f+ gives a mixture of Cu(en)i+, Cu(en)2+ and aqueous Cu" ions on the solid.Clearly the negative charge density of the surface plays a major role in these phenomena. To substantiate this idea we extended these studies to the complexation of Cu" with ethylenediamine and with ammonia on A-type zeolites. In this case the Si : A1 ratio equals 1, whereas it is 1.25 and 2.50 for X- and Y-type zeolites respectively. There- fore, A-type zeolites have the highest charge density of this series of high surface area solids. EXPERIMENTAL PREPARATION OF CUII-EXCHANGED A-TYPE ZEOLITES kg NaA, as obtained from Union Carbide's Linde division, were exchanged for 24 h in 1 mol dm-3 NaCl solution at room temperature, washed 4 times and subsequently exchanged overnight in 0.9 dm3 0.003 mol dm-3 CuCI2 at room temperature.The pH of 2550 5 xR . A . SCHOONHEYDT, P . PEIGNEUR AND J . B . UYTTERHOEVEN 2551 the exchange solution before mixing with NaA was -6. The pH of the supernatant after exchange with CuIr was 7.90. The sample was washed Cl--free (AgN0,-test), dried in air at room temperature and stored in a polyethylene bottle. Its water content was 23.4 :$ and its exchangeable cation content was 6.32 meq g-l Naf and 0.99 meq g-l Cu". This sample is indicated by the symbol CuA-10. A second sample, CuA-1, was prepared in identical conditions but with 10 times less CuII in the exchange solution in order to obtain -1 % exchange, suitable for e.p.r. work. These samples were used for adsorption of ethylenediamine and ammonia from the gas phase. Exchange levels above 10 % give partial lattice destruction.Such samples were not investigated in this study. EXCHANGE OF Cu(en)$+ (x = 1,2) ON A-TYPE ZEOLITES NaA, as obtained from Union Carbide's Linde Division, was exchanged twice in 1 mol dm-3 NaCl, washed C1--free and air-dried at room temperature. This material was mixed with solutions with Cu : en ratios equal to 1 or 2 containing enough CuII to achieve 1 and 10 % exchange levels. These materials are denoted as Cu(en)A-1, Cu(en)A-10, C~(en)~A-l and Cu(en)2A-10. Bis-complexes on A were also prepared at 1 and 10 % exchange levels from a suspension with a Cu: en ratio = 1 by adding ethylenediamine progressively. E.p.r. and reflectance spectra of these samples were recorded as such, and after degassing overnight up to 523 K with 50 K intervals.GAS PHASE ADSORPTION OF ETHYLENEDIAMINE AND NH3 ON CUII-EXCHANGED CuA-1 and CuA-10 were degassed in vaciio at 673 or 573 K for 8 . 6 4 ~ lo4 s. O2 was added at these temperatures, the samples cooled and evacuated at room temperature. NH,, purified by freezing and evacuation was adsorbed at room temperature from a calibrated volume in doses of ratio NH3/Cu11 = 1. After addition of each dose the equilibration time was 1.728 x lo5 s prior to recording the spectra. Direct saturation of the samples with 4.666 x lo4 N m-2 NH3 was also performed. Ethylenediamine, previously dried over metallic Na, was adsorbed at room temperature until saturation of the samples, previously degassed as described above, was reached.E.p.r. and reflectance spectra were recorded of these en-saturated samples and after each 5OK interval in the subsequent degassing process. The maximum temperature was 523 K. In another experiment en was allowed to adsorb overnight on the samples degassed as described above, followed by adsorption of water vapour and again ethylenediamine. After each adsorption step e.p.r. and reflectance spectra were recorded. ZEOLITES A TECHNIQUES E.P.R. E.p.r. spectra were recorded at 77 K in the X-band on a Varian E-3 spectrometer. The spectra were simulated with the SIM-13 computer program written by Lozos, Hoffman and Franz of Northwestern University, under the assumption of axial symmetry. g-Values were determined relative to a DPPH standard (g = 2.0036). REFLECTANCE SPECTROSCOPY The spectra were recorded between 2000 and 360 nm on a Cary 17 spectrometer equipped with a type I reflectance unit.The integrating sphere was coated with MgO. The reference was MgO. Both sample and reference were in matching vacuum cells.6 The spectra were digitalized and stored on paper tape with a Hewlett Packard 3480C digital voltmeter and a Hewlett Packard 3489A data punch respectively. They were processed in the IBM 370/158 computer to obtain plots of the Kubelka-Munck function against wavenumber with a Calcomp 502 off-line plotter. The baseline was processed in the same way and could therefore be substracted from the experimental spectra. 1-8 12552 Cu COMPLEXES ON ZEOLITE-A RESULTS COMPLEXES WITH AMMONIA The stepwise adsorption of NH3 on the dehydrated CuA-10 was followed by reflectance spectroscopy.The spectra are shown in fig. 1. After adsorption of 1 NH3/Cu" the 3 band spectrum of dehydrated CuA-10 reduces to a single broad band in the range 10 000-17 000 crn-l. With increasing loadings however, a shoulder n 8 0.128 0.096 0.064- 0.032- 0.000, 5000 15000 25000 wavenumber /ern-' FIG. 1 .-Reflectance spectra of the stepwise adsorption of NH3 on CuA-10 : 1, dehydrated at 550 K ; 2, 0.99 N H ~ / C U I I ; 3, 2.07 N H ~ / C U I I ; 4, 4.00 NH,/CuII ; 5, 12 NH,/CuII. at 15 500 cm-l appears, which grows continuously at the expense of the spectrum of the dehydrated sample. At saturation the latter is only present as a weak shoulder. The 6580cm-l is an overtone of the NH3 vibrations. Fig. 2 shows some typical e.p.r.spectra of CuA-1 saturated with NH3, and after various NH3 desorption steps. Simulated spectra are shown as dotted lines. Table 1 summarizes the g-values and TABLE VALUES AND HYPERFINE SPLITTING CONSTANTS OF THE COMPLEXES OF CUT' WITH NH3 ON ZEOLITE A treatment II A p m - 1 91 AYlcrn-1 A rlcrn-1 structure evacuation at 373 K 2.3087 0.015 09 2.067 0.000 96 - ( 0 1 ) 3 ~ " evacuation at 473 K 2.3715 0.013 63 2.068 0.001 92 - (Od3 cU1I saturation withNH3 2.2480 0.017 61 2.0575 0.001 05 0.001 14 Cu(NH&+ hyperfine splitting constants derived from the simulated spectra. After room temperature saturation with NH, two species are present : most of the CuI1 was in the form of a tetragonal complex with 911 = 2.271 and All = 0.017 89 cm-l, a small amount was tetrahedrally coordinated [shown by the symbol T in fig.2(1)] with approximate parameters 911 = 2.0079, A , , = 0.0132 cm-l and g1 = 2.258. The overlapping spectra prevented a more accurate determination of the parameters. Prolonged saturation with NH3 at room temperature or at 373 K did not alter this spectrum, except for a small decrease in 911 for the tetragonal complex : gI1 = 2.263 and 2.258, respectively after prolonged saturation at room temperature and at 373 K. Evacuation at room temperature and at 323 K produced a new tetragonal species,R . A. SCHOONHEYDT, P . PEIGNEUR AND J . B . UYTTERHOEVEN 2553 I I 2 1 1 DPPH FIG. 2.-Experimental and simulated e.p.r. spectra of CuA-1 in the presence of NH3 : 1, saturation at room temperature for 4 days ; 2, 1 +evacuation at 373 K ; 2', simulated spectrum of 2 ; 3, 2+ evacuation at 473 K ; 4, 3 +evacuation at 773 K in O2 +saturation with NH3 after removal of 0 2 at room temperature ; 4, simulated form of 4.T is the tetrahedral spectrum. The right hand sides of spectra 3 and 4 are enlarged perpendicular regions of spectra 3 and 4 respectively.2554 Cu COMPLEXES ON ZEOLITE-A which became the only one present after evacuation at 373 K (fig. 2, spectrum 2). Its g-values, derived from the simulated spectrum, are shown in table 1 together with the hyperfine splitting constants. A third tetragonal signal is produced by evacuation at 423 K and this in turn became the only species present after evacuation at 473 K. The spectrum is shown in fig. 2 with the perpendicular region on an enlarged scale.This signal remained unchanged upon further evacuation except for a decrease in intensity above 673 K. The sample turned white after evacuation at 773 K. The signal intensity was reduced to 3 its original value but could be restored with 0,. Saturation with NH3 reproduced the original spectrum but much less tetrahedral complex was present. If the spectrum was generated in 250 Torr NH3 all the CU" was converted to the tetragonal complex and N-superhyperfine structure was visible in the perpendicular region. This is shown in fig. 2 (spectrum 4) on an enlarged scale. The g-values and hyperfine splitting constants are shown in table 1. This spectrum did not change upon addition of water. COMPLEXES WITH ETHYLENEDIAMINE The complexation of Cu" with gaseous en was followed by e.p.r.on CuA-1 and by reflectance spectroscopy on CuA-10. Fig. 3 shows the e.p.r. spectra. After 4 days saturation with gaseous en at 298 K the sample turned blue. The e.p.r. spectrum of Cu(en)2+ was then generated, with parameters similar to those of the monocomplex TABLE 2.*-VALUES AND HYPERFINE SPLITTING CONSTANTS OF THE COMPLEXES BETWEEN CU" AND ETHYLENEDIAMINE ON ZEOLITE A sample treatment complex gll ATicm-1 gI AYIcm-1 CUA-1 saturated with en Cu(en)2+ 2.239 0.0182 2.047 0.002 02 CUA-1 en+ H20 Cu(en)$+ 2.205 0.0191 2.050 0.002 85 CuA-1 en+ HzO+ en Cu(en)z+ 2.214 0.0172 2.048 0.00076 Cu(en)zA-l Cu(en)$+ exchange+ en Cu(en)$+ 2.1888 0.0201 2.045 0.001 90 Cu(en),A-1 Cu(en) exchange+ en+ Cu(en)$+ 2.1924 0.0197 2.048 0.002 85 Cu(en)2A-10 Cu(en) exchange+ en+ Cu(en)i+ 2.2075 0.01 94 2.047 0.002 66 Cu(en)A-10 Cu(en) exchange evacuated Cu(en)2+ 2.2667 0.0179 2.053 0.001 43 Cu(en)A-10 Cu(en) exchange, evacuated Cu(en)*+ 2.2485 0.0172 2.053 0.002 95 Cu(en)A-10 Cu(en) exchange, evacuated Cu(en)2+ 2.2595 0.01 59 2.055 0.002 57 NaA Cu(ea)2+ exchange at Cu(H,O);+ 2.3545 0.0151 2.072 O.OO0 96 pH = 5.8 washings NaOH addition to exchange solution NaOH at r.t.during 15 min at 381 K at 439 K in dehydrated X-type zeolite^.^ The e.p.r. parameters are shown in table 2. This complex was stable up to 473 K in vacuu. The spectral intensity decreased upon evacuation at 523 K and the sample turned white. Cu(en)2+ was also generated upon saturation with en at 363 K. When, however, the sample was subsequently equilibrated overnight at room temperature with 20 Torr water vapour, it turned violet and the e.p.r.spectrum was that characteristic of Cu(en)z+ (fig. 3 and table 2). The monocomplex could be regenerated by evacuation at 323 K for 1 h. Saturation with water vapour again yielded Cu(en)$+, showing the reversibility of the complexationR. A. SCHOONHEYDT, P. PEIGNEUR AND J . B. UYTTERHOEVEN 2555 reaction. When the violet sample was saturated with en, which had not been dried over Na prior to adsorption, the sample turned blue and the e.p.r. spectrum was characteristic for the tris-complex Cu(en)$+ (fig. 3 and table 2). This complex was unstable and evacuation at 323 K resulted in the formation of the mono-complex. FIG. 3.-E.p.r. spectra of Cu(en)2+ (l), Cu(en)i+ (2) and Cu(en)<+ (3) formed by adsorption of en and water from the gas phase on dehydrated CuA-1. The dotted lines represent the simulated spectra, the full lines the experimental spectra.2556 Cu COMPLEXES ON ZEOLITE-A Qualitatively the same phenomena were observed on CuA-I0 both by e.p.r.and reflectance spectroscopy, but the conversions dehydrated -+ mono -+ bis -+ tris were only partial. Fig. 4 gives some reflectance spectra for illustration. Thus, a 4 day adsorption period of en on CuA-10 even at 338 K converts Cu" only partially to Cu(en)'+. This complex absorbs at 14 540 cm-', while the band around 11 300 cm-I is reminiscent of lattice-bonded CU".~~ * Subsequent adsorption of water produces Cu(en)$+ with a band maximum at 18 400 cm-' but the spectrum shows shoulders at 11 500 and 14 500 cm-', characteristic of lattice-bonded Cu" and Cu(en)2+.Adsorption of en on this violet sample turned it blue but both e.p.r. and reflectance spectra were ill-defined. The band maximum was at 16 700 cm-I, but the band was asymmetric towards its low frequency side. In any case, evacuation above room temperature produced the mono-complexes again, besides the originally present lattice-bonded Cur'. The former was gradually converted to the latter upon prolonged evacuation at 473 K. 0.064r 14OOO 18500 23000 wavenumber /cm-' FIG. 4.-Reflectance spectra of the complexes of CuII with ethylenediamine on CuA-10 : 1, saturated with en vapour at 338 K ; 2, 1 +water vapour ; 3,2+en. EXCHANGE OF THE CU"-ETHYLENEDIAMINE COMPLEXES If an aqueous solution with Cu" : en = 1 at pH = 5.8 was mixed with an aqueous suspension of NaA, previously brought to the same pH with HN03, and if the pH was kept in the range 5.8-6.3 during exchange, the resulting zeolite contained aqueous Cu" ions as shown by the e.p.r. parameters in table 2.Addition of en to this suspen- sion increased the pH, the zeolite became blue and both e.p.r. and reflectance spectro- scopy indicated the presence of aqueous Cu" and Cu(en)2+. At pH = 7.5 Cu(en);+ was formed on the zeolite too but it was only at pH > 10 that nearly all the Cu" was converted to the bis-complex. If this suspension was left standing in the laboratory, the supernatant became violet and the zeolite white, indicating that Cu" migrates out of the zeolite to form Cu(en),"+ in solution.If the exchange was performed directly in a Cu(en),"+ solution, no Cu(en)f+ was found on the zeolite even at pH = 11, and only after repeated washings with en solutions was it possible to record minor quantities (-0.5 % of the C.E.C.) of Cu(en);+ on the solid. The e.p.r. para- meters of these Cu(en)2A systems, determined from the simulated spectra, are summarized in table 2. There was a small but definite dependence of these quantitiesR . A . SCHOONHEYDT, P. PEIGNEUR AND J . B . UYTTERHOEVEN 2557 on the mode of preparation. The corresponding reflectance spectra are shown in fig. 5. The suspension of Cu(en),A-10, prepared by exchange of the mono-complex and addition of en until pH = 11.2, is characterized by a broad band with maximum at 17 600 cm-I.Upon evacuation, the band becomes narrower and shifts to n 8 wavenumber Icrn-' FIG. 5.-Reflectance spectra of Cu(en)zA-fO : 1, suspension prepared by addition of en until pH = 11.2 to Cu(en)A-10 (the real spectral intensity is 10 x larger than shown) ; 2, evacuated at room temperature during 300 s ; 3, evacuated at 323 K during 8.64 x lo4 s. 18 000 cm-l. Upon heating, the band maximum moves again to smaller wave- numbers and the band becomes strongly asymmetric at its low frequency side. This indicates that a decomposition of Cu(en)i+ occurred with formation of the mono- complex and above 373 K also lattice-bonded Cu". This was confirmed by the wavenumber /cm- l FIG. 6.-Reflectance spectra of Cu(en)A-10 : 1, Cu(en)A-10, evacuated at room temperature for 900 s ; 2, evacuated at 329 K for 8.64 x lo4 s ; 3, evacuated at 381 K for 8.64 x lo4 s ; 4, evacuated at 439 K for 8.64~ lo4 s.2558 Cu COMPLEXES ON ZEOLITE-A analogous e.p.r.experiments on Cu(en),A-1 . A comparison of the decomposition of Cu(en),A-1 and Cu(en),A-10 revealed that up to 473 K the ratio mono : bis was higher for the 1 % exchange level (55 : 45) than for the 10 % exchange level (33 : 60). Above 473 K Cu" was reduced and the zeolite turned white. The decomposition of the mono-complex on Cu(en)A-10 is illustrated in fig. 6. Besides the combination band of N20 at 5270cm-l and the overtones of the NH and OH stretching vibrations in the region 6800-7000 cm-l, the spectrum contains an asymmetric band due to Cu" at 14 500 cm-l with a shoulder around 17 150 cm-l.Upon evacuation the spectrum intensifies and the band maximum shifts to 13 730 cm-l. After evacuation at 381 K it is again situated at 14 500 cm-l but is asymmetric towards the low frequency side. At 439 K a strong reduction of intensity occurred with no significant change in the band position. The e.p.r. parameters of these monocomplexes are given in table 2. In every case only one species was detected, except after room temperature evacuation. Besides a signal due to Cu(en),+ we saw a signal due to CU" with oxygen ligands. The g-values and hyperfine splitting constants of Cu(en)*+ on these ion exchanged samples are different from those of Cu(en),+ on the dry zeolite. The differences are more pronounced than for Cu(en),"+. No spectra were recorded above 473 K because Cu" was reduced.DISCUSSION COMPLEXES OF CU" WITH NH3 CU" readily coordinates dry, gaseous NH3 as shown by the disappearance of the characteristic spectrum of dehydrated CuA8 even at very small loadings. For NH3/Cu11 < 1 the broad band in the region 10 000-17 000 cm-l reflects the presence of several species Cu(NH,);+ (x = 1-4) and lattice-bonded Cu". Above 1 NH,/CU~~ the situation is clearer. There is a systeniatic growth of a band at 15 500 cm-l at the expense of the 10 800 em-' band, characteristic of lattice-bonded Cu". The position of the band maximum of the ammonia complex is identical to that of the tetraammine in X- and Y-zeolites with small Cu" contents. We therefore assign the 15 500 cm-l band to Cu(NH,)i+. This is, however, not the only species present, Some Cu" ions are tetrahedrally coordinated, probably in the form [(OJ,-CU-NH,]~+ where O1 is a lattice oxygen of the six-membered rings.Evidence for this comes both from the e.p.r. spectra which showed a species with gL N 2.257 and gll z 2.0079 and from the reflectance spectra in the form of a low frequency shoulder on the 15 500 cm-l band. The e.p.r. parameters of Cu(NH,);+ in A-type zeolites depend slightly but in a systematic way on the environment. Thus, during the process of exposure to NH3 glr diminishes from 2.271 to 2.258. Mter destruction of the complex and regeneration of the Cu"A sieve readsorption of NH3 gives Cu(NH,)i+ with gll = 2.248 and a clear N-superhyperfine structure in the perpendicular region. This environmental dependence of the characteristics of the Cu-ammine complexes is a general observation in zeolite^.^^ However, the tetra-ammine is less stable in A-type sieves than in X or Y.In the latter cases, degassing produced an intermediate [(O,),-CU-NH,]~+, stable in the range 356-383 lo Room temperature evacua- tion of NH,-saturated A produces an e.p.r. signal with 911 = 2.3087, reminiscent of CuI1 surrounded by oxygen-containing ligands. This is therefore lattice-bonded CuI1 and the only species present after degassing at 373 K. At high temperature a second lattice-bonded Cu" species is produced with 911 = 2.3715, which became the only species present at 473 K. The same two Cu" signals were also detected after degassing a hydrated Cu"A sieve at 673 K.8 In the configuration with the largest gll, Cu" is quantitatively converted toR.A . SCHOONHEYDT, P . PEIGNEUR AND J . B . UYTTERHOEVEN 2559 Cu(NH,)i+ upon adsorption of NH3. This was not the case after the initial NH3 adsorption, i.e. when the two e.p.r. signals of lattice-bonded Cu" were present. We conclude that the configuration giving rise to 911 = 2.3715 is fully accessible to NH3, whereas the configuration with 911 = 2.3087 corresponds to Cu" ions on such sites that a small amount of Cu" remained tetrahedrally coordinated upon NH3 adsorption. We suggest that the signal with 911 = 2.3715 is due to Cu" on the hexagonal rings in the supercages and the signal with gll = 2.3087 to Cu" on the hexagonal rings inside the sodalite cages. Two similar Cu" signals are detected in dehydrated CuY zeolites, one with 911 = 2.38 and one with 911 = 2.32.The former has been attributed to Cu" on sites 11, the latter to Cu" on sites I'.I1 This assignment is analogous to ours. COMPLEXES OF CU" WITH ETHYLENEDIAMINE The adsorption of dry, gaseous ethylenediamine on a dehydrated Cu"A sieve is peculiar with respect both to the NH3 adsorption, discussed in previous paragraphs and to the adsorption of ethylenediamine on X and Y zeolite^.^ Thus, in the latter cases tris-complexes were synthesized even at small en-loadings. In the present case only the mono-complex is formed on the surface. This preference of A-type zeolites for the mono-complex is further substantiated by our experimental observations that bis- and tris-complexes can only be made in the presence of water. Moreover they are very unstable, as simple degassing at 323 K restored the mono-complex, which itself is stable up to 473 K.At higher temperature CuI1 is reduced. The e.p.r. parameters and the reflectance spectra of these complexes are very similar to those of the same complexes on the surface of X- and Y-zeolites. Therefore, this peculiar behaviour of the Cu" ethylenediamine complexes in A-type sieves with respect to X- and Y-zeolites reflects, in the first place, differences in environment. The critical diameter of the supercage of A-type zeolites (1.14 nm) is smaller than that of X- and Y ( N 1.3 nm), but more important, the surface negative charge density is higher and so is the number of exchangeable Na+ ions per supercage.Thus, without water the mono-complex is formed with Cu" still coordinated to 3 oxygens of the six-membered rings. A second ethylenediamine molecule is unable to approach the Cu" of the mono-complex within coordinating distance, but the small water molecule can. As soon as the mono-complex is freed from the surface by water, a second ethylenedi- amine molecule coordinates to Cu" to form the bis-complex, because ethylenediamine is a stronger ligand than water. In this picture we have neglected the role of the residual Na+ ions, because their coordinating power is several orders of magnitude less than that of CuII, at least in aqueous solution. This may be different on the surface of a dehydrated sieve, with its much smaller electrical permittivity, but no experimental data are available to substantiate this idea. In any case, the bis- and tris-complexes in the supercages of CuA are unstable and water is needed to stabilize them, i.e.to shield them from the attracting power of the negatively charged surface. Therefore, removal of water destroys them and restores the mono-complex again. In an aqueous suspension not only the water molecules, but also the dimensions of the complexes in solution, with respect to the free diameter of the 8-membered rings, and the pH, determine the complex loading of zeolite A. Thus, the mono- complex can be exchanged directly onto the A-type sieve above yH = 6. This is impossible or almost impossible for the bis-complex because the dimensions of the complex (0.47 x 0.77 nm) exceed the free diameter of the 8-membered windows (-0.44 nm).Another factor is the instability of the bis-complex in the supercage of zeolite A : the bis-complex can by synthesized only in aqueous suspensions by first exchanging Cu" or the mono-complex and then adding large excesses of2560 Cu COMPLEXES ON ZEOLITE-A ethylenediamine to the suspensions (pH The instability of the bis-complex is further substantiated by our observation that Cu'I moves out of the zeolite into the en solution upon standing overnight. The e.p.r. parameters and band maxima in the reflectance spectra of these hydrated Cu(en)A and Cu(en),A samples are significantly different from those of the dry sieves, which again indicates the effect of the environment on the complexes in the supercages. This idea is further evidenced from the behaviour of Cu(en)A-10 upon evacuation (fig.6). Before evacuation some Cu(en)$+ is present on the surface, as shown by the high frequency shoulder on the 14 500 cm-l band of Cu(en)2+ in the reflectance spectrum. After evacuation at room temperature no trace of Cu(en)z+ is left but some Cu" ions are in an oxygen environment. This is shown by the shift of the 14 500 cm-l band to lower wavenumbers and the appearance of the characteristic e.p.r. signal. Any heating produces only Cu(en)2+, with e.p.r. parameters strongly dependent on the treatment (table 2). This observation could not be matched with Cu(en),A samples as the bis-complex partially decomposed to Cu(en)2+. However, this reaction was not quantitative, in contrast to the behaviour of these complexes on the dry sieves.Higher complex loadings and water seem to favour the bis- complex. 11). CHARACTERIZATION OF THE CU" AMINE COMPLEXES On the dry sieves the band maxima of Cu(NH,)i+, Cu(en),+ and Cu(en)d+ are located at 15 500,14 540 and 17 400 cm-l respectively, significantly below those of the same complexes in aqueous solution (16 600 cm-l, 15 500 and 18 200 respectively). TABLE 3.--M.O. COEFFICIENTS OF THE 3d ORBITALS OF Cur' IN THE M.O. EXPRESSIONS OF THE CUII-AMMINE COMPLEXES ON A-TYPE ZEOLITES 8 2 ( h Z , dY4 U out-of-plane n complex in-plane n c?2(dXZ-Y2) 8 ? ( h Y ) Cu(NH)&+ on A 0.89 0.92 0.88 on X, Y : low loading 0.88 0.90 0.66 high loading 0.87 0.93 0.69 &(en):+ gas phase adsorption of wet Cu(en)2A-l 0.88 0.85 0.83 wet Cu(en)2A-10 0.85 0.93 0.87 dry Y 0.85 0.89 0.73 wet Y 0.84 0.92 0.76 wet X 0.81 0.90 0.77 solution 0.85 0.89 0.82 en+H20 on A 0.84 0.92 no convergence In fully hydrated sieves the band maximum of Cu(en),+ remained around 14 500 cm-1 but that of Cu(en)j+ shifted towards 18 000 cm-l.These observations, in combina- tion with the variation of the e.p.r. parameters of these complexes with their environ- ment in the supercages, prompted us to examine these complexes in more detail. This is given in the Appendix. The general conclusion emanating from table 3 is that there is no significant effect of the various preparations of Cu(en)l+ in A-type zeolites on the Cu-N bonding characteristics. There is, however, a tendency for smaller p2 values in X- and Y-zeolites than in A.This is the coefficient of the out- of-plane n-orbital and therefore susceptible to direct interaction with surface orbitals.R . A . SCHOONHEYDT, P. PEIGNEUR AND J . B . UYTTERHOEVEN 2561 This is in agreement with our previous conclusions.12 The effect is small. The zeolite matrix should be considered as a solvent with weak coordination power, at least with respect to the coordination of the ammines considered. As a solvent it drastically influences the complex formation when compared with water. R. A. S . is indebted to the N.F.W.O. (Belgium) for a grant as '' Bevoegdverklaard Navorser ". P. P. acknowledges financial support from the Belgian Government. The technical assistance of F. Pelgrims is greatly appreciated. The authors thank Prof. Smets for the use of his e.p.r.machine. This work was made possible through the finances of the Belgian Government (Staatssekretariaat voor Wetenschapsbeleid). APPENDIX It is a reasonable assumption that the effective symmetry of Cu(NH3)$+ and Cu(en);+ is D4h. Cu(en)2+ and Cu(en);f have lower symmetries, but reflectance and e.p.r. techniques on these powdered materials do not resolve them, As a consequence we have omitted them in this calculation. The experimental band maxima of Cu(NH3)$+ and Cu(en)$+ are assumed to be the average of the 3 possible spin-allowed, parity-forbidden transitions in O4h (2B2g c 2Blg, 'Alg +- 2Blg, 2Eg c 2Blg). The m.0. coefficients were calculated with the theory of Moreno and Barrius0,~~9 l4 which was successfully applied to Cu(en)$+ on X and Y zeolites and on Camp Berteau montmorillonite. l2 Our previous data for Cu(NH3)$+ on synthetic faujasites were recalculated and incorporated in table 3 together with the present results. In view of the approximations involved in the band assignments and our estimate of the overlap integrals l2 the m.0. coefficients of table 3 can only be used for comparison purposes. P. Peigneur, Adsorption A$nity and Stability of Transition Metal Ammine Complexes on Alumina-silicates, Ph. D. Thesis (K. U. Leuven, 1976). A. Maes, P. Marijnen and A. Cremers, J.C.S. Faraday I , 1977, 73, 1297. A. Maes, P. Peigneur and A. Cremers, J.C.S. Faraday I, accepted for publication. F. Velghe, R. A. Schoonheydt, J. B. Uytterhoeven, P. Peigneur and J. H. Lunsford, J. Phys. Chern., 1977,81,1187. P. Peigneur, J. H. Lunsford, W. De Wilde and R. A. Schoonheydt, J. Phys. Chem., 1977, 81, 1179. F. Velghe, R. A. Schoonheydt and J. B. Uytterhoeven, Clays and Clay Min., 1977, 25, 375. ' P. A. Jacobs, W. De Wilde, R. A. Schoonheydt, J. B. Uytterhoeven and H. Beyer, J.C.S. Furuday I, 1976,72, 1221. R. A. Schoonheydt, P. Peigneur and J. B. Uytterhoeven, to be published. W. De Wilde, R. A. Schoonheydt and J. B. Uytterhoeven, A.C.S. Symp. Ser., 1977, 40, 132. lo E. F. Vansant and J. H. Lunsford, J. Phys. Chem., 1972, 76,2860. l 1 W. Morke, F. Vogt and H. Bremer, 2. anoug. Chem., 1976,422,273. l 2 R. A. Schoonheydt, J. Phys. Chem., 1978, 82, 497. l 3 M. Moreno and M. T. Barriuso, Anales Fisica, 1975, 71, 205. l4 M. Moreno and M. T. Barriuso, Solid State Comm., 1975, 17, 1035. (PAPER 7/2194)