(392a) Identification of Reaction Sites on Supported Metal Catalysts

Introduction.

Identification of the active sites during a catalytic reaction is difficult.  Over the years analysis has moved from ex-situ, to in-situ, to operando.  Many of the methods rely heavily on various spectroscopies to probe the adsorbed state during the reaction.  In this study we have used selective poisoning to remove specific sites as well as changing catalyst support and have followed the change in catalyst selectivity and activity.

The reaction studied was the hydrogenation of cis-2-pentenenitrile (C2PN) over Ni/alumina and Ni/silica catalysts.  The potential products are pentylamine (PA), the fully saturated molecule with both alkene and nitrile groups hydrogenated; pentane nitrile (PN), where only the alkene functionality is hydrogenated; cis-pentenylamine, where only the nitrile group is hydrogenated; and the various isomerisation products such as cis- and trans-3-pentenenitrile (C3PN & T3PN) and trans-2-pentenenitrile (T2PN).  Under the conditions studied only the cis-pentenylamine was not formed.

To delineate exactly the sites involved with the isomerisation and the hydrogenation, selective poisoning with carbon monoxide and ammonia was utilised.  Carbon monoxide was used to poison the metal sites, while the ammonia was used to poison the acid sites on the support. Experimental

The catalyst used in this study was a 16% w/w Ni/Al₂O₃.  The Ni/alumina was supplied by Johnson Matthey (HTC 400, 1.2RP, trilobes), and had been pre-reduced and pacified.  The support was q-alumina with a surface area of ~100 m²g^-1.  The catalyst was reactivated by treatment in flowing 5% H₂/N₂ at 523 K for 2h.

Reactions were performed at a temperature of 373 K in an atmospheric, flow-through reactor.  In the flow system the carrier gas was 5 % H₂/N_2, which passed through a reagent vaporiser containing cis-2-pentenenitrile and on to the catalyst bed.  The concentration of the cis-2-pentenenitrile was controlled with a p-xylene/liquid nitrogen slurry bath that kept the reagent at a constant temperature of 286 K.  This gave a cis-2-pentenenitrile vapour pressure of 0.0139 atm and hence a H₂:cis-2-pentenenitrile ratio of 3.6:1.  Vapour samples were taken via a gas sample valve attached to a Varian 3300 gas chromatograph with a TCD detector using a SGE BP20 capillary column (I.D. 0.25 mm, 30m). Results

The q-alumina support, rather than being inert, plays a significant role in the catalysis.  Catalysing the complete hydrogenation of cis-2-pentenenitrile to pentyl amine and subsequently the isomerisation to trans-2-pentenenitrile.  Although some isomerisation activity may have been expected, the total hydrogenation was a surprising result.  This initial activity for pentyl amine formation occurs on a site that deactivates rapidly; no pentyl amine was detected after ~45 min on stream.  Once this site was deactivated the major product was trans-2-pentenenitrile. 

Figure 1 shows a typical reaction profile of the hydrogenation of cis-2-pentenenitrile over Ni/Al₂O₃ at 373 K.  Initially there was a short period of 1-pentylamine formation

Figure 1. Cis-2-pentenenitrile hydrogenation over q?alumina at 373 K

followed by the production of the saturated nitrile, n-pentanenitrile.  As the reaction proceeded isomerisation was observed with the formation of trans-2- and trans-3-pentenenitrile.  Selective Poisoning with Carbon Monoxide

Carbon monoxide was adsorbed at 0.25, 0.5 and 1 monolayer coverage onto the reduced catalyst prior to the hydrogenation reaction being initiated.  With increasing CO coverage there was a decrease in overall hydrogenation activity to n-pentanenitrile, with no n-pentanenitrile being produced at 1 monolayer coverage (Figure 2).  Selective Poisoning with Ammonia

The number of acid sites found experimentally in this study, for the q-alumina support, was 360 mmol.g^-1.  When the nickel was added to the alumina the number of acid sites dropped by 42 %.  70.4, 141, and 604 mmole.g^-1 of ammonia were pre-adsorbed onto the reduced Ni/Al₂O₃ catalyst prior to the hydrogenation of cis-2-pentenenitrile.  In general, as ammonia exposure was increased, there was an increase in the time taken for isomers to be observed in

Figure 2.  Cis-2-pentenenitrile hydrogenation over Ni/Al₂O₃ at 373 K after exposure to 1 monolayer coverage of CO

the system.  In all cases 1-pentanamine was observed in the early stages of the reaction and was undetectable by 50 minutes.  Table 1 reports the time at which each isomer was observed as a function of the extent of ammonia pre-adsorption.  No effect was seen on pentanenitrile formation.

Table 1.  Effect of ammonia exposure on isomer formation during cis-2-pentenenitrile hydrogenation

DISCUSSION

The hydrogenation of cis-2-pentenenitrile over Ni/Al₂O₃, shown in Figure 1, reveals that the principal hydrogenation product is n-pentanenitrile.  When CO was adsorbed onto the catalyst the effect was to reduce the hydrogenation activity to n-pentanenitrile.  No significant effect is seen on the isomerisation reaction.  These results confirm that hydrogenation to n-pentanenitrile is related solely to the metal function of the catalyst and that the metal does not catalyse isomerisation.  This result lends credence to the mechanism proposed [1] that the nitrile is adsorbed via the CN functionality even though it is the C=C that is hydrogenated.  This mechanism invokes a cyclic intermediate and a concerted hydrogen transfer.  If the C=C functionality was adsorbed on the nickel then it would be expected that the carbon backbone would isomerise.

With increasing amounts of ammonia pre-adsorbed an increase in the time taken to observe isomerisation was noted (Table 1).  The loss of inhibition was likely to be a result of desorption of ammonia from the surface species due to the reaction temperature of 373 K.  Hence isomerisation on the acid sites was inhibited while there was adsorbed ammonia.  The behaviour of the support is also now explainable.  Pentyl amine will inhibit isomerisation in a similar manner to ammonia hence the change over between the pentyl amine and the isomers is not a conversion of one site into another but removal of the amine from the acid sites that catalyse isomerisation.  

Conclusions.

Through selective poisoning we have been able to gain an insight to the nature of the sites associated with the hydrogenation of cis-2-pentenenitile.  Firstly we can confirm that the olefinic hydrogenation of cis-2-pentenenitrile occurs on reduced nickel metal particles, on sites that can be selectively poisoned by pre-adsorption of carbon monoxide.  Secondly, the isomerisation reactions of cis-2-pentenenitrile to trans-2- and trans-3-pentenitrile occur on alumina support sites.  The sites responsible for this activity are proposed as a Lewis acid-Lewis base pair that upon adsorption with NH₃ poisons the site's ability for isomerisation though the adsorption of NH₂^- and H⁺ [2, 3].

A third site catalyses the complete hydrogenation of cis-2-pentenitrile to 1-pentanamine in the early stages of the reaction.  It occurs on a limited amount of highly active support sites that are rapidly deactivated.  These sites were not deactivated by pre-adsorbed ammonia, suggesting that the sites are not acidic.  Hence the deactivation is more likely due to carbon laydown.  In the early stages of cis-2-pentenitrile hydrogenation there was a poor mass balance and trace amounts of ammonia were detected during the reaction [4].  It is likely that the delay in isomer production under normal reaction conditions is due to poisoning of the support sites by ammonia formed in situ by hydrogenolysis of 1-pentanamine.  No hydrocarbon fragments were detected in the reactor effluent suggesting that the catalyst retained these and that they are the cause of the rapid deactivation of the pentyl amine hydrogenation sites on the support.  References

[1]        J. L. Dallons, G. James, B. Delmon, Catal. Today, 5 (1989) 257.

[2]        Y. Amenomiya, J. Catal. 46 (1977) 326.

[3]        R.L. Burwell Jr, Catalysis Science and Technology, Eds. J.R. Anderson and M. Boudart, Vol. 9, Springer-Verlag, Berlin (1991) p 60.

[4]        A. S. Canning and S. D. Jackson, A.C.S., Catalyst Deactivation and Development Preprints, 49(1) (2004) 47.