Intermetallics

3.1.1 Intermetallics dealloyed in alkaline medium

A common observation is that all catalysts which will feature one or more active sites from the iron group 3d metals are always and almost dealloyed in alkaline solutions as they form highly active layered metal oxyhydroxide under OER conditions and Volmer step promoting metal hydroxides under HER conditions in alkaline solutions. An example is Hausmann and co-workers[95] reported an interesting finding that Fe₆Ge₅ intermetallics continue to leach Ge out under OER conditions creating more space for the active FeOOH. Howeverthis phenomenon was found to be surface centered and the Ge well inside the Fe₆Ge₅ particles was fairly retained after 1 h of CA at an overpotential at which it delivered 10 mA cm⁻². This resulted in an interesting Fe₆Ge₅ metallic core with FeOOH active surface. It is known that FeOOH when present alone is only moderately active because of its poor electronic conductivity.

Howevertheir core-shell structure featuring the metallic Fe₆Ge₅ helped them overcome this issue. Nonethelessa closer examination of the 25 h CA response reveals that the activity degrades with respect to time and achieves a stable performance after 20 h. This implies that the initial boost given by the creation of new active sites because of the leaching of Ge is later brought down by the structural reconstruction-induced agglomeration. This could potentially be an issue of concern for intermetallics. The same group also reported the FeSn intermetallic alloy[97] that was found to show similar activity enhancement because of the leaching of Sn. With thisthey also extended the scope of the catalyst beyond OER to hexamethyl furfural (HMF)ethanoland acetaldehyde oxidation (Fig. 4). Extending their interest to another active metal Ni and the leachable component Sithey made Ni₂Si and applied the same for OER and dehydrogenation of primary amines very recently [98]. In another workthe same group showed that these intermetallics when activated can also be used for alkaline HER alongside OER with their CoSn₂ intermetallic precatalyst [99]. Menezess and co-workers[100] prepared a nickel germanide intermetallic precatalyst that was able to form highly active NiOOH in situ while letting the Ge component be leached under OER conditions. Their post-OER characterizations revealed extensive oxidation of Ge (Fig. 5a-f). All recent similar studies were reviewed and benchmarked against their activity in Table 1.

Fig. 4. : Illustration depicting the formation of FeSn₂ intermetallics by the Kirkendall process and the in-situ activation of the same during OER and its extended scope in multi-organic compounds oxidation in alkaline waters.

Reproduced with permission from ref. 96 (Copyright 2020Wiley).

Fig. 5. : (a) OER CVs of NiGe intermetallics with controls in 1.0 M KOH. (b) CA response for over 20 h. (c-d) post-CA TEM images showing surface reconstruction and NiOOH formation. (e-f) post-CA XPS analysis showing a little change in the oxidation state of Ni and complete oxidation of Ge.

Reproduced with permission from ref. 100 (Copyright 2020Wiley).

Table 1. Benchmarking intermetallic precatalysts containing a dealloyable component based on their activity [101–105].

Note: ‘N/A′ implies ‘Not Available’. There could be some minor inconsistencies with the activity numbers as some of them have been manually calculated from the given polarization curves. Rows featuring catalysts screened in different pH are colored to distinguish them easily.

3.1.2 Intermetallics dealloyed in acidic medium

In generalintermetallics that are dealloyed in acid a have main catalytically active component which is usually either Ir or Ru because these are the two well-known OER active stable electrocatalysts in acidic solutions. An example of this kind is the electrochemical dealloying of Ga from IrGa intermetallic alloy reported by Chen and co-workers [64]. In acid solutionsthe dealloying of Ga from IrGa resulted in the formation of IrO_x surface layer that delivered a very lower OER overpotential and outperformed the commercial Ir/C by delivering threefold higher activity. Similarlyintermetallics of Pt can also be dealloyed from an acid-unstable component to deliver better HER performance. This was shown by Lim and co-workers who made popcorn-like GaPt₃ intermetallic alloy which was dealloyed for Ga in acid and the Pt left behind with increased surface area outperformed the commercial Pt/C. Howeverthe existence of a non-precious OER catalysts cannot be excluded based on this fact. There are few Co and Fe-based electrocatalysts which undergo electrochemical dealloying in acidic solutions that form relatively stable spinel Co₃O₄ and Fe₃O₄ OER electrocatalysts. An example for this type of study is the work of Shen and co-workers[66] who showed that Fe₅Si₃ can be activated in acidic solutions and applied for OER in the same. Though it is contradicting to see a Fe-based OER catalyst shown to be stable in acid solutionsthe authors have backed their observation with the corrosion resistance provided by the amorphous SiO₂ shell protecting the actual OER catalyst (i.e.FeOOH/Fe₃O₄). The results from these studies are encouraging in the context of bringing out exceptional activity enhancement.

These catalysts are also benchmarked in Table 1 along with the ones that were studied in alkaline solutions. Howeverthese intermetallics still have some issues like poor retention of initial activitymoderate enduranceand the possibility of agglomeration and reduced ECSA upon prolonged use. These are discussed in detail in the subsequent sections. This study revealed that the leachable components do not have to get wasted in the solution but they can also provide other benefits.

14.6.1 General properties of intermetallic compounds

Like ceramicshoweverthe greatest disadvantage of intermetallics is their low ductilityparticularly at low and intermediate temperatures. The reasons for the lack of ductility vary from compound to compound but include (i) a limited number of easy deformation modes to satisfy the von Mises criterion(ii) operation of dislocations with large slip vectors(iii) restricted cross-slip(iv) difficulty of transmitting slip across grain boundaries(v) intrinsic grain boundary weakness(vi) segregation of deleterious solutes to grain boundaries(vii) covalent bonding and high Peierls–Nabarro stress and (viii) environmental susceptibility. It has been demonstratedhoweverthat some intermetallics can be ductilized by small alloying additions: Ni₃Al with boronTiAl with MnTi₃Al with Nb. This observation has encouraged recent research and development of intermetallics and the possibility of application of those materials.

14.6.2 Nickel aluminides

Ni₃Al (nickel aluminide) is the ordered fcc γ′-phase and is a major strengthening component in superalloys. Ni₃Al single crystals are reasonably ductile but in polycrystalline form are quite brittle and fail by intergranular fracture at ambient temperatures. The basic slip system is {1 1 1} ⟨1 1 0⟩ and has more than five independent slip modes but still exhibits grain boundary brittleness. Remarkablysmall additions of ~0.1 at.% boron produce elongations up to 50%. General explanations for this effect are that B segregates to grain boundaries and (i) increases the cohesive strength of the boundary and (ii) disorders the grain boundary region so that dislocation pile-up stresses can be relieved by slip across the boundary rather than by cracking. This general explanation is no doubt of significance but additionallythere are distinct microstructural changes within the grains which must lead to a reduced friction stress and ease the operation of polyslip. For examplethe addition of B reduces the occurrence of stacking fault defects. Addition of solutessuch as Bare not expected to raise the stacking fault energy and hencethis effect possibly arises from the segregation of B to dislocationspreventing the superdislocation dissociation reactions (see Section 4.9).

Microhardness measurements inside grains and away from grain boundaries indeed show that boron softens the grains. The ductilization effect is limited to nickel-rich aluminides and cannot be produced by carbon or other elementsalthough some substitutional solutes such as Pdwhich substitutes for Niand Cu produce a small improvement in elongation. Small additions of FeMn and Hf have also been claimed to improve fabricability. Grain size has been shown to influence the yield stress according to the Hall–Petch equationand B appears to lower the slope k_y and facilitate slip across grain boundaries. These alloys are also known to be environmentally sensitive. Hffor examplewhich does not segregate to grain boundaries but still improves ductilityhas a large misfit (11%) and possibly traps H from environmental reactionssuch as Al+H₂O→Al₂O₃+H. Tiwhich has a small misfitdoes not improve the ductility.

The most striking property of Ni₃Al is the increasing yield stress with increasing temperature up to the peak temperature of 600°C (Figure 14.17). This behaviour is also observed in other L1₂ intermetallicsparticularly Ni₃Si and Zr₃Al. This effect results from the thermally activated cross-slip of screw dislocations from the {1 1 1} planes to the {1 0 0} cube planes where the apb energy is somewhat lower. The glide of superdislocations is made more difficult by the formation of Kear–Wilsdorf (K–W) locks (see Chapter 11)and their frequency increases with temperature. Electron microscopy measurements of apb energies given in Table 14.6 shows that the apb energy on {1 0 0} decreases with aluminium contentand this influences the composition dependence of the strengthshown in Figure 14.17. The cross-slip of screw dislocations from the {1 1 1} planes to cube planes also gives rise to a high work hardening rate.

Figure 14.17. Effect of aluminium content on the temperature dependence of the flow stress in Ni₃Al.

After Noguchi et al. (1981).

Table 14.6. Anti-Phase Boundary Energies in Ni₃Al

Alloy	γ111 (mJ m−2)	γ100 (mJ m−2)	γ111/γ100
Ni–23.5Al	183±12	157±8	1.17
Ni–24.5Al	179±15	143±7	1.25
Ni–25.5Al	175±13	134±8	1.31
Ni–26.5Al	175±12	113±10	1.51
Ni–23.5Al+0.25B	170±13	124±8	1.37

Although the study of creep in γ′-based materials is limitedit does appear to be inferior to that of superalloys. Above 0.6T_m creep displays the characteristic primary and secondary stageswith steady-state creep having a stress exponent of approximately 4 and an activation energy of around 400 kJ mol⁻¹consistent with climb being the rate-controlling process. At intermediate temperatures (i.e. around the 600°C peak in the yield stress curve) the creep behaviour does not display the three typical stages. Insteadafter primary creepthe rate continuously increases with creep straina feature known as inverse creep. In primary creepplanar dissociation leads to an initial high creep rate which slows as the screws dissociate on {1 0 0} planes to form K–W locks. Howeverit is the mobile edge dislocations which contribute most to the primary creep strainand their immobilization by climb dissociation which brings about the exhaustion of primary creep. The inverse creep regime is still not fully researched but could well be caused by glide on the {1 0 0} planes of the cross-slipped screw components.

The fatigue life in high-cycle fatigue is related to the influence of temperature on the yield stress and is invariant with temperature up to about 800°Cbut falls off for higher temperatures with cracks propagating along slip planes. With boron doping the fatigue resistance is very sensitive to aluminium content and decreases substantially as Al increases from 24 to 26 at.%. Neverthelesscrack growth rates of Ni₃Al+B are lower than for commercial alloys.

Hyperstoichiometric Ni₃Al with boron can be prepared by either vacuum melting and casting or from powders by HIPing. Fabrication into sheets is possible with intermediate anneals at 1000°C. At presenthoweverthe application of Ni₃Al is not significant; Ni₃Al powders are used as bond coats to improve adherence of thermal spray coatings. NeverthelessNi₃Al alloys have been tested as heating elementsdiesel engine componentsglass-making moulds and hot-forging diesslurry-feed pumps in coal-fired boilershot-cutting wires and rubber extruders in the chemical industry. Ni₃Al-based alloys as matrix materials for composites are also being investigated.

Nickel aluminide (NiAl) has a caesium chloride or ordered β-brass structure and exists over a very wide range of composition either side of the stoichiometric 50/50 composition. It has a high melting point of 1600°C and exhibits a good resistance to oxidation. Even with such favourable properties it has not been commercially exploited because of its unfavourable mechanical properties. Because it is strongly orderedlow-temperature deformation occurs by an a⟨1 0 0⟩ dislocation vector and not by a/2⟨1 1 1⟩ superdislocations. {1 1 0} ⟨1 0 0⟩ slip therefore leads to insufficient slip modes to satisfy the general plasticity criterionand in the polycrystalline condition β-NiAl is extremely brittle. The ductility does improve with increasing temperature but above 500°C the strength drops off considerably as a result of extensive glide and climb. Improvements in properties are potentially possible by refinement of the grain size and by using alloying additions to promote ⟨1 1 1⟩ slipas in FeAlwhich has the same structure. In this respectadditions of FeCr or Mn appear to be of interest. For high-temperature applicationsternary additions of Nb and Ta have been shown to improve creep strength through the precipitation of second phasesand mechanical alloying with yttria or alumina is also beneficial.

A further commercial problem of this material is that conventional production by casting and fabrication is difficultbut production through a powder route followed by either HIPing or hot extrusion is more promising.

14.6.3 Titanium aluminides

Electron microscopy studies of Ti₃Al or α₂ have shown that deformation by slip occurs at room temperature by coupled pairs of dislocations with $b=1/6⟨11\overline{2}0⟩$ which glide only on $\left\{10\overline{1}0\right\}$ planes and by very limited glide on $\left\{11\overline{2}\overline{1}\right\}$ with pairs of dislocation with $b=1/6⟨11\overline{2}6⟩$. The ductility increases at higher temperatures due to climb of the $⟨11\overline{2}0⟩$ dislocations and to the increased glide mobility of $1/6⟨11\overline{2}0⟩$ and $1/6⟨11\overline{2}6⟩$ dislocations through thermal activation. Only limited activity of the {0 0 0 1} $⟨11\overline{2}0⟩$ slip systems is observedeven at high temperatures.

The most successful improvements in the ductility of Ti₃Al have been produced by the addition of β-stabilizing elementsparticularly niobiumto produce α₂-alloys. An addition of 4 at.% Nb produces significant slip on $\left\{10\overline{1}0\right\}⟨11\overline{2}0⟩,\left\{0001\right\}⟨11\overline{2}0⟩$ and $\left\{11\overline{2}1\right\}⟨11\overline{2}6⟩$ as well as some slip on {1 0 1 1}⟨1 1 2 0⟩. This improvement is attributed to the decrease in covalency as Nb substitutes for Ti with a consequent reduction in the Peierls–Nabarro friction stress. Alloys based on α₂ are Ti–(23–25)Al–(8–18)Nbof which Ti–24Al–11Nb has excellent spalling resistance. Most Ti₃Al+Nb alloyssuch as super α₂also contain other β-stabilizers including Mo and Vi.e. Ti–25Al–10Nb–3V–1Mowhich exhibits about 7% room temperature elongation. Alloying Ti₃Al with β-stabilizing elements to produce two-phase alloys significantly increases the fracture strength. These Ti₃Al-based alloys can be plasma-melted and cast followed by TMP in the (α₂+β) or β-range. The improved ductility of Ti₃Al alloys has led to aerospace applications in after-burners in jet engines where it compares favourably in performance with superalloys and gives a 40% weight saving.

Developments are taking place in rapid solidification processing to include a second phase (e.g. rare-earth precipitates) and to provide powderswhich may be consolidated by HIPingto produce fully dense components with properties comparable to wrought products. There are also developments in intermetallic matrix composites by the addition of SiC or Al₂O₃ fibres (~10 µm). These have some attractive propertiesbut the fibre–intermetallic interface is still a problem.

The γ-phase Ti–(50–56)Al has an ordered fc tetragonal (L1₀) structure up to the m.p. 1460°C with c/a=1.02 (Figure 14.18). Deformation by slip occurs on {1 1 1} planes andbecause of the tetragonalitythere are two types of dislocationsnamelyordinary dislocations 1/2⟨1 1 0⟩ and superdislocations ⟨0 1 1⟩=1/2 ⟨0 1 1⟩+1/2⟨0 1 1⟩. Another superdislocation 1/2⟨1 1 2⟩ has also been reported.

Figure 14.18. Structure of (a) TiAl (L1₀) and (b) (1 1 1) plane showing slip vectors for possible dissociation reactionse.g. ordinary dislocations 1/2[110]superdislocations [0 1 1] and 1/2[1 1 2]and twin dislocations 1/6[1 1 2].

After Kim and Froes (1990).

At room temperaturedeformation occurs by both ordinary and superdislocations. However[0 1 1] and [1 0 1] superdislocations are largely immobile because segments of the trailing superpartials 1/6⟨1 1 2⟩-type form faulted dipoles. The dissociated 1/2⟨1 1 0⟩ dislocations bounding complex stacking faults are largely sessile because of the Peierls–Nabarro stress. Some limited twinning also occurs. The flow stress increases with increasing temperature up to 600°C as the superpartials become mobile and cross-slip from {1 1 1} to {1 0 0} to form K–W-type locksthe 1/2⟨1 1 0⟩ slip activity increases and twinning is promoted.

The two-phase (γ+α₂) Ti–Al alloys have better ductility than single-phase γ with a maximum at 48 at.% Al. This improvement has been attributed to the reduced c/a with decreased Alfurther promotion of twinning and the scavenging of O₂ and N₂ interstitials by α₂. The combination of high stiffness (E=175 GPa at 20°C to 150 GPa at 700°C)density-normalized strength similar to cast Ni-based alloyshigh temperature strength and reasonable oxidation resistance to 750°Clow thermal expansion coefficientand high thermal conductivityhave led to the high level of interest in TiAl-based alloys. The major limitations to their application are the intrinsic low room temperature ductility (no better than 2–3%)the low fracture toughness (between 10 and 20 MPa m^1/2 at 20°C) and the high growth rate of fatigue cracks.

Alloy and process development have resulted in some successful applications of these alloys over the last 10 or so years. Cast turbochargers are now manufactured in Japan for cars and wrought exhaust valves were used in formula 1 cars for some years. Major applications for these alloys are still awaited despite the success of these two applications; the high cost of processing is holding commercial developments back. In the case of thermo-mechanical processing the costs are high because the alloys are strong at normal hot working temperatures and because some sort of protection (such as canning with steel) from oxidation must be used during working. In the case of castings the efficiency of material usage is very lowboth because casting technology is not efficient and because melting and casting are difficult because of the reactivity of molten TiAl-based alloys.

The compositions of TiAl-based alloys which are of commercial interest lie within the range of about Ti45–48Al (at.%)but all alloys contain other elements in attempts to improve the properties of the binary alloy. Additions of Nb between about 5 and 8 at.% are important in improving oxidation resistance and also imparting some solid solution strengthening.

An understanding of the microstructures which can be obtained in cast or in wrought products of TiAl-based alloys requires knowledge of the phase changes that occur over the temperature range from the melting point to room temperature. The relevant part of the binary phase diagram between Ti and Al is shown in Figure 14.19 which also indicates schematically the influence of some alloying additions on phase boundaries.

Figure 14.19. Partial Ti–Al phase diagram showing the influence of ternary additions on the position of the various boundaries.

Courtesy M.H. Loretto.

The various phase transformations in the Ti–Al system offer the possibility of microstructural control both for the wrought route and for the casting route. Thus cooling samples containing less than about 44 at.% Al the solidification will take place through the formation of βwhich may or may not be removed via the peritectic reaction. Subsequent cooling of the α-phase results in precipitation of γwhich under typical cooling rates encountered with castingsresults in the formation of a lamellar structure. This ‘fully lamellar’ structure consists of parallel lamellae of γ and α and of twinned γ. These lamellae are formed on the (0 0 0 1) plane of the α-phase and thustheir length is defined by the pre-existing α-grain size. Somewhat slower cooling results in some of the γ-lamellae coarsening at colony boundaries to form γ-grains through local growth of the lamellaeto form a ‘near-fully lamellar’ structure. Hot working in the two-phase region results in the formation of equiaxed γ- and α-grains; the ratio of the amounts being defined by the average alloy composition and the hot working temperature. Subsequent cooling results in the equiaxed α-grains forming lamellae to yield a duplex microstructureor if extensive hot working is used a structure consisting of γ- and α-grains is formedtermed ‘near-γ’.

If the cooling rate is increasedas in oil or water quenchingthe α-phase transforms massively to γ if the Al content is above about 44 at.% (below this Al contentα is retained)and this transformation offers a further opportunity for microstructural control in cast samples by heat treating in the two-phase field so that α can precipitate on all four {1 1 1} planesthroughout the γ-grains. The tendency to transform massively is strongly dependent upon the composition which has important consequences upon the choice of alloy composition in cast samples.

14.6.4 Other intermetallic compounds

A number of intermetallic compounds are already used in areas which do not rely on stringent mechanical properties. Fe₃Alfor exampleis used in fossil fuel plants where resistance to both sulphur attack and oxidation is important. Ni₃Si is used where resistance to hot sulphuric acid is required. There are several compounds with rare earth elements used in magnet technology (see Chapter 8). PdIn is gold-coloured and a possible dental material. Zr₃Al has a low neutron capture cross-section and is a possible reactor material.

The β-compound NiTi (Nitinol) is an important shape memory alloy. The SME manifests itself when the alloy is deformed into a shape while in a low-temperature martensitic condition but regains its original shape when the stress is removedand it is heated above the martensitic regime. Strains of the order of 8% can be completely recovered by the reverse transformation of the deformed martensitic phase to the higher temperature parent phase. The martensite transformation in these alloys is a thermoelastic martensitic transformation in which the martensite plates form and grow continuously as the temperature is lowered and are removed reversibly as the temperature is raised. NiTi was one of the original SME alloysbut there are many copper-based alloys which undergo a martensitic transformatione.g. Cu–17Zr–7Al. Application of SME alloys relies on the characteristic that they can change shape repeatedly as a result of heating and cooling and exert a force as the shape changes. By composition control (increasing the Ni content or substitution of Cu lowers the M_s temperature of TiNi)the shape memory can be triggered by normal body temperature or any other convenient temperature to operate a device. Several biomedical applications have been developed in orthopaedic devices (e.g. pulling fractures together)in orthodonticsin intrauterine contraceptives and in artificial hearts. Industrial applications include pipe couplings for ships which shrink during heatingelectrical connectorsservo-mechanisms for driving recording pensswitchesactuators and thermostats.

6.1 Pd-based intermetallic catalysts

6.1.1 PdGa catalysts

A supported PdGa catalyst was compared to bulk PdGa in selective hydrogenation of acetylene (98) (Table A8). The supported PdGa catalyst was obtained from a layered double hydroxide (LDH)Pd_0.025Mg_0.675Ga_0.3(OH)₂(CO₃)_0.15·mH₂Othat is decomposed in H₂ at 550 °Cand leads to the formation of PdGa nanoparticles of 6.7 nm supported on the MgO–MgGa₂O₄ mixed oxide. According to HRTEM and XPSan intermetallic Pd₂Ga compound forms. After a slow but strong activation during the first 23 h of reaction at 200 °C leading an increasing conversion from < 5% to 98% (no explanation is given for this)the activity of this catalyst in selective hydrogenation of acetylene in excess of ethylene is ∼ 5000 times higher than over bulk Pd₂Gaand the selectivity to ethylene is only slightly lower for the supported catalyst (70% at 98% acetylene conversion) than for the bulk phase (74% at 94% conversion).

In another studyAl₂O₃-supported PdGa and Pd₂Ga catalysts have been compared to a commercial 5%Pd/Al₂O₃ in hydrogenation of acetylene in excess of ethylene at 200 °Cwith an initial conversion close to 100% (99). The catalysts were prepared by deposition of NPs preformed in liquid phase and annealed in an organic solvent at 250 °C to generate the intermetallic structures according to XRD (Table A8). The PdGa and Pd₂Ga/Al₂O₃ are much more selective (70–80% at ∼ 90% conversion) than the Pd catalyst (15%). In additionthey are much more stable as they lose only a few % conversion over 20 h of reactionand the selectivity remains stable whereas Pd deactivates to 45% conversion and selectivity increases to 20%. The authors also report that the specific activity of PdGa and Pd₂Ga/Al₂O₃ is considerably higher than their respective bulk IMC (factor of 35,000 and 1300respectively)and close to that of Pd/Al₂O₃but with a much higher selectivity.

Another PdGa catalyst prepared differently exhibits the same trend. PdGa was supported on a mixed oxide after [PdCl₄]^2 − adsorption on MgGaAl-LDH supported on Al₂O₃ and decomposition in H₂ at 550 °C (100) (Table A8). The results of HRTEMXPSCO-IR and TPR show that bimetallic PdGa NPs are formed on MgO–Al₂O₃ with two compositionsPd/Ga = 2/1 and 1/5. The catalysts exhibit comparable activity as monometallic Pd/MgO–Al₂O₃but with a much higher selectivity to ethylene in hydrogenation of acetylene in excess of ethylene at 45 °C: at 90% conversionthe selectivity to ethylene is 55%75% and 82% over PdPdGa(2:1) and PdGa(1:5)respectivelybut it drops down at 100% conversion (Fig. 6). The higher selectivity of PdGa(1:5) is attributed to the higher number of Ga atoms surrounding Pd that isolates them from each otherand facilitates the desorption of ethene from the surface of catalysts. The PdGa/MgO–Al₂O₃ catalysts are stable during the 48 h reaction; the stability is attributed to the “net trap confinement effect” of the support (see Scheme 3 in Section 4.2)which inhibit the migration and aggregation of the PdGa NPs.

Fig. 6. Conversion of acetylene conversion (A) and selectivity to ethylene (B) as a function of the reaction temperature; conversion versus time-on-stream at 45 °C (C) over different catalysts Pd/MgO–Al₂O₃ (Ablack)PdGa(2:1)/MgO–Al₂O₃ (Bblue); PdGa(1:5)/MgO–Al₂O₃ (Cgreen) (in competitive conditions of reaction) (100).

Reprinted from HeY.; LiangL.; LiuY.; FengJ.; MaC.; LiD. J. Catal. 2014309166–173Copyright (2014)with permission from Elsevier.

6.1.3 PdSn catalysts

PdSn/γ-Al₂O₃ catalysts with Pd/Sn = 1 were prepared by sequential impregnation using either SnCl₄ or tetrabutyltin (Sn(n-C₄H₉)₄) as tin precursors for a study of hydrogenation of 1,3-butadiene in the presence or not of 1-butene (103) (Table A8). In the absence of 1-butenethe two PdSn/Al₂O₃ catalysts show almost the same activity as Pd/Al₂O₃ but a higher selectivity to butenes (~ 93% and ~ 97% with PdSn versus 82% with Pdconversions not mentioned)and with almost no deactivation for 8 h on stream at 100 °C (conversion not reported) (Fig. 7A). In the presence of 1-butenethe three catalysts are initially less active and deactivate faster with a deactivation rate that evolves as follows: Pd > PdSn_inorg > PdSn_organd they are less selective as n-butane is produced and a small fraction of 1-butene of the main stream is consumed (Fig. 7B). XAFS shows that the presence of Sn modifies the electronic and geometric structures of Pd. Tin addition inhibits the formation of the Pd β-hydride phasegenerates Pd–Sn bondsand reduces the size of Pd ensembles. This contributes to the increase in 1-butene selectivity and the decrease in the n-butane formation. Finallycarbonaceous species are formed in catalysts modified with organic tinwhile the inorganic tin precursor also generates oxide-like species in addition to making Sn–Pd bonds.

Fig. 7. Rate of 1,3-butadiene consumption (moles per second per mole of total Pd) (A) and selectivity to butenes (B) in hydrogenation of 1,3-butadiene (a) in the absence of 1-butene; (b) in the presence of 1-butene over PdSn/Al₂O₃ and Pd/Al₂O₃ (103).

Reprinted from ChoiS. H.; LeeJ. S. J. Catal. 2000193176–185Copyright (2000)with permission from Elsevier.

6.2 Ni-based intermetallic catalysts

As already mentioned in Section 6.1the addition of Ga or In to Pd and the formation of intermetallic NPs improve the catalytic performance of Pd. In the following exampleIn has been added to Ni. NiIn/SiO₂ catalysts were prepared by co-impregnation with Ni/In ratios between 1 and 10and reduced at 450 °C before selective hydrogenation of acetylene (107) (Table A9). The formation of NiIn alloy is attested by XRD and EDS line-scan profiles performed on single NPs. Although the reaction initiates at 100% conversion for all the catalysts (Fig. 8)the authors report that the NiIn/SiO₂ catalysts and especially those with the Ni/In ratios of 6 and 10 show much higher acetylene conversionethylene selectivity and stability than Ni/SiO₂although deactivation also occurs. The authors ascribe these results to the isolation of the active surface Ni atoms by the inert In ones and to charge transfer from In to Ni (according to XPS). These changes lower the adsorption strength of ethylene and restrain C–C hydrogenolysis and polymerization of acetylene and intermediate compounds. The authors also propose that the better stability of the catalysts results from the inhibition of the formation of nickel carbidewhich easily forms in Ni/SiO₂. Howeveras the In content increasescarbonaceous deposit becomes the main reason for NiIn/SiO₂ deactivation due to the enhanced acidity of the catalyst.

Fig. 8. Conversion of acetylene (A) and selectivity to ethylene (B) over Ni/SiO₂ and Ni_xIn/SiO₂ with time on stream of acetylene and hydrogen at 180 °C (in not competitive conditions of reaction) (107).

Reprinted from ChenY.; ChenJ. Appl. Surf. Sci. 201638716–27Copyright (2016)with permission from Elsevier.

6.3 Conclusion for the supported intermetallic catalysts

Supported intermetallic catalysts (M₁/M₂ close to 1) exhibit an excellent selectivity to alkenes in semihydrogenation reactions and high stability during long-term reactions. In most of the studies citedthese properties are obtained at 100% conversion or close to. Howeverin some casesthe selectivity to alkene drops down when conversion increases. For most of the Pd-based catalyststhe activity is as good as the Pd counterpart. The Ni-based catalysts are much less active (higher reaction temperatures are needed). The overall improvement of the catalytic performances of supported intermetallic catalysts is assigned to active-site isolation (geometric effect) due the intrinsic crystallographic structure of the intermetallic and even in the alloy when the intermetallic does not form in the nanoparticles. It is also attributed to the alteration of the electronic structure of the hydrogenating metal with a higher electron density.