scieee Open visual document viewer

Silica-supported Fe/Fe–O nanoparticles for the catalytic hydrogenation of nitriles to amines in the presence of aluminium additives

Chandrashekhar, Vishwas G.

Abstract

The hydrogenation of nitriles to amines represents an important and frequently used industrial process due to the broad applicability of the resulting products in chemistry and life sciences. Despite the existing portfolio of catalysts reported for the hydrogenation of nitriles, the development of iron-based heterogeneous catalysts for this process is still a challenge. Here, we show that the impregnation and pyrolysis of iron(II) acetate on commercial silica produces a reusable Fe/Fe-O@SiO2 catalyst with a well-defined structure comprising the fayalite phase at the Si-Fe interface and alpha-Fe nanoparticles, covered by an ultrathin amorphous iron(III) oxide layer, growing from the silica matrix. These Fe/Fe-O core-shell nanoparticles, in the presence of catalytic amounts of aluminium additives, promote the hydrogenation of all kinds of nitriles, including structurally challenging and functionally diverse aromatic, heterocyclic, aliphatic and fatty nitriles, to produce primary amines under scalable and industrially viable conditions.

Full text

A icles h ps://doi.o g/10.1038/s41929-021-00722-x 1Leibniz-Ins i u ü Ka alyse e.V., Ros ock, Ge many. 2Regional Cen e o Ad anced Technologies and Ma e ials, Czech Ad anced Technology and Resea ch Ins i u e, Palacky Uni e si y Olomouc, Olomouc, Czech Republic. 3Nano echnology Cen e, Cen e o Ene gy and En i onmen al Technologies, VŠB– Technical Uni e si y o Os a a, Os a a-Po uba, Czech Republic. 4Depa men o Indus ial and Enginee ing Chemis y, Ins i u e o Chemical Technology Mumbai-Ma a hwada Campus, Jalna, Maha ash a, India. ✉e-mail: [email p o ec ed]; mb.g[email p o ec ed].edu.in; jagadeesh. [email p o ec ed]; [email p o ec ed] Ca alysis plays a decisi e ole in many basic and applied chemical p ocesses and is in ol ed in he indus ial p o- duc ion o mo e han 90% o ine and bulk chemicals as well as polyme ic ma e ials and many o he e e yday p oduc s1–5. In addi ion, he syn hesis o bioac i e compounds o li e sciences and he success o sus ainable ene gy echnologies, including g een uels, depend on e icien ca aly ic p ocesses1–5. Indeed, he suc- cess o o ganic syn hesis in he las 100 yea s elied o a la ge ex en on he disco e y and applica ion o sui able ca alys s1–5. In his espec , speci ically, he de elopmen o p ac ical and cos -e icien hyd ogena ion s a egies cons i u es a key achie emen o he las cen u y6. Nowadays, a ple ho a o molecula ly de ined me al com- plexes as well as he e ogeneous ma e ials a e equen ly applied o he selec i e hyd ogena ion o alkynes7,8, ole ins7,9,10, ca bonyl com- pounds11,12, ni oa enes13,14 as well as (he e o)a enes7,15–17. Among hese hyd ogena ion eac ions, he ans o ma ion o ni iles o p i- ma y amines is pa icula ly aluable because he la e p oduc s a e p i ileged compounds in chemis y, medicine and biology18–22. In gene al, p ima y amines cons i u e impo an p ecu so s and cen- al in e media es in he chemical, pha maceu ical, ag ochemical and ma e ials indus ies18–22. A e he o iginal disco e y o benzo- ni ile hyd ogena ion in he p esence o nickel in 1905 ( e . 23), many ca alys s we e de eloped o his and ela ed eac ions (Fig. 1)24–42. These achie emen s we e mainly possible due o he design o p e- cious me al sys ems, which allow eac ions o be pe o med a low empe a u e and p essu e24,30,31,33–38. Howe e , despi e hei emen- dous success, hei limi ed a ailabili y and highe p ice cons i u e majo d awbacks. Thus, s a e-o - he-a ca alys s o ni ile hyd o- gena ion in indus y con inue o be Raney nickel26,27,31,33 and coppe ch omi e29, which demand ha sh condi ions and su e om oxici y issues. To sol e hese p oblems, al e na i e nickel- and cobal -based he e ogeneous ca alys s ha e been epo ed in ecen yea s35,39–41. F om a sus ainabili y poin o iew, i on would be an ideal ca a- lys sys em o ni ile hyd ogena ion because o i s abundance (a 4.7% i is he second-mos abundan me al in he ea h c us ), low p ice and low oxici y14,17,42–44. Indeed, in he las decade, using sophis ica ed/syn he ically demanding PNP pince ligands, ac i e i on complexes ha e been de eloped o hyd ogena e ni iles44. In gene al, howe e , hese me al complexes a e less s able and di icul o use. In con as , he e ogeneous ca alys s show imp o ed s abili y and can be easily eused14,17,22,45–47. Un o una ely, o he bes o ou knowledge, analogous suppo ed ac i e i on cen es o he hyd o- gena ion o ni iles ha e no been c ea ed ye . He e, we epo a s able and con enien i on-based nanoca alys ob ained by he py olysis o i on ace a e on comme cial silica ha shows excellen ac i i y and selec i i y o he hyd ogena ion o a la ge numbe o ni iles o p oduce p ima y amines in he p esence o ca aly ic amoun s o aluminium addi i es. Resul s Syn hesis and cha ac e iza ion o Fe-based nanoca alys s. In he las decade, he excellen ca aly ic pe o mance o ca bon-encapsula ed co e–shell 3d me al nanopa icles has been demons a ed in a a ie y o hyd ogena ion eac ions14,16,17,22,39,40,46. These ca alys s we e p epa ed by he immobiliza ion and py oly- sis o me al complexes o me al o ganic amewo ks on ino ganic suppo s ( o example, ca bon, SiO2 and Al2O3)14,16,17,22,39,40,46. In his con ex , we de eloped ac i a ed Fe2O3 nanopa icles, su ounded by Silica-suppo ed Fe/Fe–O nanopa icles o he ca aly ic hyd ogena ion o ni iles o amines in he p esence o aluminium addi i es Vishwas G. Chand ashekha    1, Thi usangumu ugan Sen hama ai1, Ra ishanka G. Kadam   2, Ondřej Malina2, Jose Kašlík   2, Radek Zbořil   2,3 ✉ , Manoj B. Gawande   2,4 ✉ , Rajenahally V. Jagadeesh   1 ✉ and Ma hias Belle    1 ✉ The hyd ogena ion o ni iles o amines ep esen s an impo an and equen ly used indus ial p ocess due o he b oad appli- cabili y o he esul ing p oduc s in chemis y and li e sciences. Despi e he exis ing po olio o ca alys s epo ed o he hyd ogena ion o ni iles, he de elopmen o i on-based he e ogeneous ca alys s o his p ocess is s ill a challenge. He e, we show ha he imp egna ion and py olysis o i on(II) ace a e on comme cial silica p oduces a eusable Fe/Fe–O@SiO2 ca alys wi h a well-de ined s uc u e comp ising he ayali e phase a he Si–Fe in e ace and α-Fe nanopa icles, co e ed by an ul a- hin amo phous i on(III) oxide laye , g owing om he silica ma ix. These Fe/Fe–O co e–shell nanopa icles, in he p esence o ca aly ic amoun s o aluminium addi i es, p omo e he hyd ogena ion o all kinds o ni iles, including s uc u ally challeng- ing and unc ionally di e se a oma ic, he e ocyclic, alipha ic and a y ni iles, o p oduce p ima y amines unde scalable and indus ially iable condi ions. NaTuRe CaTalySiS | VOL 5 | JANUARY 2022 | 20–29 | www.na u e.com/na ca al 20 A icles NaTu e CaTalySIS ni ogen-doped g aphene, as selec i e ca alys s o he hyd ogena- ion o ni oa enes o anilines14. In addi ion, suppo ed Fe-based nanopa icles ha e also been ound ac i e o he hyd ogena ion o quinolines17. Howe e , hese and ela ed Fe ma e ials showed no ac i i y o mo e challenging subs a es, including ni iles (Supplemen a y Table 1, en ies 1 and 2). To iden i y po en ial i on-based he e ogeneous ca alys s o ni ile hyd ogena ion, we p epa ed a se ies o i on nanopa icles suppo ed on a ious sup- po s. Speci ically, comme cially a ailable neu al, acidic and basic ino ganic suppo s, o example, Vulcan XC72R ca bon powde , Ae osil silica (SiO2), γ-Al2O3 and MgO, we e imp egna ed wi h i on(II) ace a e. Subsequen ly, hese ma e ials we e py olysed a 800 °C unde educ i e (H2) condi ions. A schema ic illus a ion o he syn he ic p ocedu e wi h he SiO2 suppo is p esen ed in Fig. 2. He ea e , hese ma e ials a e deno ed as Fe(OAc)2-suppo -x, whe e x deno es he py olysis empe a u e. As a benchma k eac ion, he hyd ogena ion o 4-chlo obenzoni ile (1) o 4-chlo obenzylamine (2) was chosen (Fig. 3), no only o iden i y an ac i e ca alys sys em, bu also a selec i e one. No ably, 1 easily unde goes educ i e dehalogena- ion in he p esence o many known hyd ogena ion ca alys s. To ou su p ise, du ing ini ial con ol expe imen s, we obse ed some ac i i y (26% yield o 2) and high selec i i y (>90%) o he p ima y amine in he p esence o Fe(OAc)2-SiO2-800 (Fig. 3). To imp o e he con e sion and yield, we a ied he eac ion condi ions ( empe a u e, sol en , ca alys loading) and in es iga ed he in lu- ence o addi i es (Supplemen a y Tables 1–3). Applying highe ca alys loadings (up o 12.8 mol%), he p oduc yield inc eased up o 50%, and he selec i i y o he p ima y amine emained e y good. G a i yingly, in he p esence o aluminium iisop opoxide, he yield o 4-chlo obenzylamine (2) d ama ically inc eased o 96%. Following his excellen esul , se e al o he me al alkoxides, alu- minium compounds as well as Lewis acids and bases we e es ed as addi i es (Supplemen a y Table 2). Su p isingly, only a ew addi- i es, o example, p- oluenesul onic acid, showed a posi i e e ec on he eac ion, while mos , o example, bases, had a nega i e impac . Op imal esul s, wi h an almos quan i a i e yield o 2, we e achie ed in he p esence o inexpensi e aluminium oil (Fig. 3 and Supplemen a y Fig. 1). In his case, he aluminium oil comple ely dissol ed in he sol en (isop opanol), which explains he simila posi i e e ec o aluminium oil and aluminium iisop opoxide. Con ol expe imen s p o ed ha his dissolu ion only akes place in he p esence o ammonia (Supplemen a y Table 4). To elucida e he c ucial ole o aluminium addi i es, we pe o med kine ic in es iga- ions o he model eac ion in he absence and p esence o aluminium oil and aluminium iisop opoxide (Supplemen a y Figs. 2 and 3). Su p isingly, all he eac ions needed a p eac i a ion ime (3–9 h) o s a . Only in he p esence o aluminium iisop opoxide was his p eac i a ion d as ically educed. Based on hese esul s and he es ing o he di e en addi i es, we assume ha bo h aluminium (2016 onwa ds) Suppo ed Co and Ni NPs de i ed om me al complexes and MOFs 1905 (1905) Ni (1909) Colloidal Pd (1923) Raney Ni (1925) P O2 Cu2C 2O5 (1937) (1937) Raney Co Suppo ed ca alys s (Rh, Pd, P , Ru, Ni, Co) (1946 onwa ds) (A ound 1969) Me al bo ides (RhB, P B, RuB, NiB, CoB) Me al alloys (PdNi, NiAl, NiCo) (A ound 1970) His o y o ni ile hyd ogena ion using he e ogeneous ca alys s o e a cen u y 2021 Fe/FeÐO@SiO2 This wo k Fig. 1 | His o ical o e iew o he e ogeneous ni ile hyd ogena ion ca alys s. The hyd ogena ion o ni iles using he e ogeneous ca alys s has a long his o y ha co e s mo e han a cen u y. Fe(OAc)2 Fe FeO FeII immobilized on SiO2 Py olysis 1. S i , DMF, 150 ¡C 2. Addi ion o SiO2 3. S i , DMF, 150 ¡C, 4 h 4. Slow e apo a ion o DMF and d ying 800 ¡C, 4 h, H2 Fe/FeÐO co eÐshell nanopa icles suppo ed on SiO2 Fig. 2 | Ca alys p epa a ion. Syn hesis o Fe/Fe–O co e–shell nanopa icles by he imp egna ion and py olysis o i on(II) ace a e on SiO2. DMF, N,N-dime hyl o mamide. NaTuRe CaTalySiS | VOL 5 | JANUARY 2022 | 20–29 | www.na u e.com/na ca al 21 A icles NaTu e CaTalySIS iisop opoxide and aluminium oil a e con e ed unde he eac- ion condi ions o an ac i e Lewis acid co-ca alys ha ac i a es he ni ile g oup. These Lewis acidic cen es can p obably also be gene - a ed on he silica suppo close o he nanopa icles by eac ion wi h Si–OH si es on he su ace. No ably, ca aly ic (subs oichiome ic) amoun s (20 mol%) o he aluminium addi i es we e su icien o achie e imp o ed yields. Unde he op imized condi ions, o he suppo ed ca alys s, such as Fe(OAc)2-C-800, Fe(OAc)2-Al2O3-800 and Fe(OAc)2-MgO-800 (Supplemen a y Figs. 4–6), did no show any ac i i y (Fig. 3). In hese samples we did no obse e needle-like well-de eloped α-Fe nanopa icles g owing om he ma ix, as we did in he case o he op imal ca alys , Fe(OAc)2-SiO2-800. In con as , he i on nanopa - icles we e highly agg ega ed and/o encapsula ed wi hin he ma ix (Supplemen a y Figs. 4–6). Simila ly, Fe(OAc)2 on SiO2 py olysed a 400 °C (Fe(OAc)2-SiO2-400) was comple ely inac i e (Fig. 3). This is explained by a no ully de eloped ac i e Fe nanos uc u e a low py olysis empe a u e, which is e iden om he powde X- ay di ac ion (PXRD) pa e n (Supplemen a y Fig. 7) and ansmis- sion elec on mic oscopy (TEM) image (Supplemen a y Fig. 8) o he Fe(OAc)2-SiO2-400 sample. By con as , Fe(OAc)2-SiO2-600 and Fe(OAc)2-SiO2-1,000 exhibi ed compa able ac i i ies o ha o Fe(OAc)2-SiO2-800, p o iding 93 and 94% yields o he desi ed p oduc , espec i ely. This co ela es well wi h he simila size and well-de eloped co e–shell s uc u e o he Fe(OAc)2-SiO2-600 and Fe(OAc)2-SiO2-1,000 samples (see he TEM images in Supplemen a y Figs. 9 and 10) compa ed wi h Fe(OAc)2-SiO2-800 (Fig. 4c). As expec ed, i on(II) ace a e, unpy olysed Fe(OAc)2-SiO2 and Al addi i es alone we e comple ely inac i e in he eac ion (Fig. 3). Addi ionally, we p epa ed con ol samples, including pu e amo phous Fe2O3 nanopa icles (NPs), ayali e (Fe2SiO4) NPs and ma ix- ee Fe–Fe2O3 co e–shell NPs wi h a e y hin oxidic shell (Supplemen a y Figs. 11–13), and in es iga ed hei pe o mance in he model eac ion. No ably, he ayali e and Fe2O3 NPs we e comple ely inac i e, whe eas he Fe–Fe2O3 co e–shell NPs ga e 30% yield (Supplemen a y Table 5, en ies 1–3). This con i med he c ucial ole o he Fe–Fe2O3 co e–shell supe s uc u e in igge ing he ca aly ic p ocess. We belie e ha he ac i e ma e ial in ol es Fe cen es and/o he Fe–O a omic in e ace48. The high ac i i y o he ca alys inco po a ing he SiO2 ma ix (Fe(OAc)2-SiO2-800) s ongly indica es ha he ma ix egula es he size o he i on oxide c ys alli es49,50. Indeed, i has al eady been epo ed ha he Cu–O– SiOx in e ace in a silica-suppo ed coppe (Cu@SiO2) ca alys plays a key ole in H2 dissocia ion o o m Cu–Hδ− and SiO–Hδ+ species51. Thus, we belie e ha he silica in Fe(OAc)2-SiO2-800 would con- ibu e o he ca aly ic ac i i y by o ming such an ac i e me al–sup- po (Fe–O–SiOx) in e ace. Nex , we conduc ed a de ailed cha ac e iza ion o he mos ac i e ca alys Fe(OAc)2-SiO2-800. TEM analysis e ealed he o ma ion o co e–shell s uc u es wi h globula and od-shape mo phologies, wi h he needle diame e s anging om 10 o 30 nm and leng hs up o 100 nm (Fig. 4a–c). Ene gy-dispe si e X- ay spec oscopy (EDS) o his ma e ial showed he p esence o Si, O and Fe ele- men s (Supplemen a y Fig. 14). The high- esolu ion TEM image (HRTEM; Fig. 4d) con i ms ha he me allic pa o he ca alys is composed o an α-Fe co e g owing om he SiO2 ma ix. Indeed, he high-angle annula da k- ield scanning ansmission elec- on mic oscopy (HAADF-STEM) and elemen al mapping images clea ly e i y ha he Fe co e nanopa icles a e g owing om he SiO2 ma ix and a e co e ed by a laye o ul a hin i on oxide wi h a hickness o a ew nanome es (Fig. 4 –i). Based on his assignmen , he mos ac i e Fe(OAc)2-SiO2-800 ca alys is abb e ia ed o Fe/ Fe–O@SiO2 in he ollowing ex . A ep esen a i e HAADF-STEM image o a globula pa icle and ypical dep h p o ile plo showing he in ensi y dis ibu ion o he Si, O and Fe elemen s a a ious dis ances om he su ace a e shown in Fig. 5a,b, espec i ely. The dep h p o iles con i m ha he hickness o he oxidic Fe-O shell is less han 5 nm. Clea ly, he ca alys su ace is composed o i on nanopa icles, which g ow om he SiO2 ma ix, s abilized by an ex emely hin i on oxide shell. Fu he mo e, we pe o med e y de ailed chemical mapping wi h a ocus on he i on-con aining su ace componen s ha a e esponsible o he ca aly ic ac i i y. All he iden i ied Fe-bea ing su ace-ac i e phase was composed o Fe nanopa icles co e ed wi h a e y hin shell o i on oxide, i espec i e o he size and mo phology (globula , needle-like) o he Fe NPs (Supplemen a y Fig. 15). To iden i y he chemical and s uc u al cha ac e o he ca a- lys , we analysed he Fe/Fe–O@SiO2 sample by PXRD, Mössbaue spec oscopy, X- ay pho oelec on spec oscopy (XPS) and elec on pa amagne ic esonance (EPR) spec oscopy. The PXRD pa e n o Fe/Fe–O@SiO2 (Supplemen a y Fig. 16) shows s ong me allic α-Fe e lec ions a 2θ = 52.33, 77.16 and 99.60°, co esponding o c ys alline ace s o he Fe (110), (200) and (211) planes, espec- i ely (Join Commi ee on Powde Di ac ion S anda ds (JCPDS) ca d numbe 04-012-6482). Thus, α-Fe is he dominan c ys alline phase in ol ed in he ca alys supe s uc u e. The low-c ys alline SiO2 ma ix is ep esen ed by a b oad peak a 2θ ≈ 26°, indica ing he p esence o poo ly c ys alline c is obali e (JCPDS ca d num- be 04-008-7643). The ul a hin i on oxide laye is, in acco d wi h expec a ion, no iden i iable in he PXRD pa e n due o i s mos ly amo phous na u e. Howe e , de ailed PXRD analysis clea ly showed addi ional low-in ensi y di ac ion peaks co esponding o ayali e (Fe2SiO4, JCPDS ca d numbe 04-002-3681) and c ys alline silicon (Si(0), JCPDS ca d numbe 04-014-8844). In summa y, PXRD p o- ided a complex pic u e o he high- empe a u e chemis y o he Fe–Si–O sys em. These obse a ions a e in line wi h he XPS analysis, which con i med he p esence o jus Fe, Si and O elemen s in he su ey spec um (Supplemen a y Fig. 17a). The high- esolu ion O1s spec- um o Fe/Fe–O@SiO2 (Supplemen a y Fig. 17b) iden i ies peaks 26% 50% 96% 97% <1% <1% <1% <1% 94% 93% <1% <1% <1% 0 10 20 30 40 50 60 70 80 90 100 Yield o 4-chlo obenzylamine (%) Ca alys Fe(OAc)2-SiO2-800 Fe(OAc)2-SiO2-800 + AI(i-OP )3 b Fe(OAc)2-SiO2-800 + AIc Fe(OAc)2-C-800 + AIc Fe(OAc)2-Al2O3-800 + AIc Fe(OAc)2-MgO-800 + AIc Fe(OAc)2-SiO2-400 + AIc Fe(OAc)2-SiO2-600 + AIc Fe(OAc)2-SiO2-1000 + AIc Fe(OAc)2-SiO2 + AIc Fe(OAc)2 + AIc AIc Fe(OAc)2-SiO2-800a Fe ca alys N Cl Cl NH2 50 ba H2, 5–7 ba NH3 i-P OH, 120 °C, 24 h Fig. 3 | e alua ion o Fe ca alys s. Hyd ogena ion o 4-chlo obenzoni ile. Reac ion condi ions: 0.5 mmol 4-chlo obenzoni ile, 40 mg ca alys (8.5 mol% Fe), 50 ba H2, 5–7 ba NH3, 3 ml i-P OH, 120 °C, 24 h. Gas ch oma og aphy (GC) yields a e gi en using n-hexadecane as s anda d. aWi h 60 mg ca alys . bWi h 20 mol% Al(i-OP )3. cWi h 20 mol% (3 mg) Al oil. NaTuRe CaTalySiS | VOL 5 | JANUARY 2022 | 20–29 | www.na u e.com/na ca al 22 A icles NaTu e CaTalySIS a 530.72 and 533.14 eV, co esponding o Fe–O and Si–O bonds, espec i ely. Mos impo an ly, he high- esolu ion Fe2p spec- um (Supplemen a y Fig. 17d) e eals Fe 2p3/2 peaks a 710.27 and 712.73 eV, and Fe2p1/2 peaks a 723.37 and 725.83 eV, which can be asc ibed o Fe3+ species52, along wi h wo sa elli e peaks a 717.41 and 730.51 eV. This is s ong p oo ha he amo phous i on oxide phase co e ing he Fe(0) co e co esponds o amo phous Fe2O3. The o ma ion o i on(III) oxide was con i med by he hype ine pa am- e e s de i ed om oom- empe a u e Mössbaue spec oscopy (Fig. 5c). The 57Fe Mössbaue spec um o he sample shows a dominan sex e (68% o he ela i e spec um a ea) and wo double compo- nen s. Acco ding o he Mössbaue hype ine pa ame e s, he sex e wi h ze o isome shi and a hype ine ield o 32.9 T can be unam- biguously assigned o e omagne ic α-Fe. The double wi h high isome shi (1.18 mm s–1) and quad upole spli ing (2.64 mm s–1) clea ly belongs o Fe(II) ions in he ayali e s uc u e o med a he Si–Fe in e ace53, in pe ec ag eemen wi h he esul s o he PXRD measu emen s. Finally, he double wi h an isome shi o 0.35 mm s–1 is ypical o high-spin Fe(III) in amo phous i on(III) oxide wi h diso de ed symme y o he i on en i onmen , as p o ed by he ela i ely high quad upole spli ing (1.31 mm s–1)54. I is wo h men ioning ha possible aces o Fe(III) ions usually in ol ed in he ayali e s uc u e would o e lap he double o he i on(III) oxide phase. Finally, he EPR spec um o Fe/Fe–O@SiO2 shows b oad aniso opic signals wi h g ac o alues o gx = 2.72, gy = 2.04 and gz = 1.8 (ga e = 2.19) a 77 K (Fig. 5d), which indica es he p es- ence o e omagne ic pa icles co esponding o Fe(0) wi h dis inc size and mo phology. In summa y, HRTEM, HAADF-STEM, XPS, PXRD, EPR and Mössbaue spec oscopy allowed us o explo e he chemical and s uc u al cha ac e o he Fe/Fe–O@SiO2 ca alys , being composed o a SiO2 ma ix, a ayali e in e ace (Fe2SiO4) and α-Fe–amo phous Fe2O3 co e–shell nanopa icles g owing om he silica ma ix and ep esen ing he su ace-ac i e phase pa icipa ing in he ca aly ic p ocess. The EPR and Mössbaue da a con i m he e omagne ic cha ac e o he sample, p ede e mining he ca alys o simple magne ic sepa a ion. Hyd ogena ion o benzoni iles and he e ocyclic ni iles. Wi h an ac i e Fe-based ca alys (Fe/Fe–O@SiO2) in hand, we demons a ed i s gene al applicabili y o he selec i e hyd ogena ion o all kinds o ni iles. Al hough in mos o he eac ions aluminium oil was used as an inexpensi e addi i e, expe imen s pe o med o com- pa ison in he p esence o aluminium iisop opoxide ga e simila p oduc yields. Fi s , we ca ied ou he hyd ogena ion o a se ies o a oma ic ni iles (Fig. 6). Simple benzoni iles as well as subs i- u ed ones bea ing a oma ic o alkyl g oups ga e he co espond- ing p ima y amines in yields o up o 96% (Fig. 6, p oduc s 3–7). Fo he gene al applicabili y o any new ca alys , i s chemoselec i - i y is an impo an aspec . Thus, om a syn he ic poin o iew, i is impo an o no e ha his i on-based ca alys sys em is highly selec i e o he hyd ogena ion o he ni ile g oup in unc ionalized and mul isubs i u ed subs a es. As an example, amino-subs i u ed and halogena ed benzylic amines we e p epa ed, which a e e sa ile in e media es in o ganic syn hesis as well as o pha maceu icals and ag ochemicals. Such p oduc s, including he mo e sensi i e 4-iodobenzylamine, we e easily p oduced om he co esponding 200 nm 20 nm 20 nm 10 nm 10 nm 10 nm 10 nm 10 nm 5 nm abc de ghi Fe Fe Fe Fe O Si OSi HAADF Fig. 4 | TeM and HRTeM imaging o Fe(Oac)2-SiO2-800. a–i, TEM (a–c), HRTEM (d) and HAADF-STEM (e) images and elemen al mapping o i on ( ), i on and oxygen (g), i on and silicon (h), and i on, silicon and oxygen (i) o he Fe(OAc)2-SiO2-800 ca alys . NaTuRe CaTalySiS | VOL 5 | JANUARY 2022 | 20–29 | www.na u e.com/na ca al 23 A icles NaTu e CaTalySIS benzoni iles in yields o up o 97% (Fig. 6, p oduc s 8–16). In addi ion, i luo ome hyl-subs i u ed benzylamines we e ob ained in yields o 94 and 95% (Fig. 6, p oduc s 17 and 18, espec i ely). We we e also pleased o ind ha he ni ile g oup was also selec- i ely hyd ogena ed in he p esence o he mo e challenging C≡C, es e , bo onic es e , amide, e he , i luo ome hoxy and hioe he g oups (Fig. 6, p oduc s 19–30). Fu he mo e, when mul isubs i- u ed ni iles we e subjec ed o hyd ogena ion, educ ion o he CN g oup again ook place highly selec i ely owa ds he co esponding benzylic amines in yields o up o 95% (Fig. 6, p oduc s 31–46). He e ocyclic amines cons i u e e sa ile in e media es in he pha maceu ical and ag ochemical indus ies. In his espec , he selec i e hyd ogena ion o cyano-subs i u ed he e ocycles, o example, quinolines, indoles, py oles, benzodioxoles, benzodiox- anes, u ans, mo pholines and ph halanes, is o pa icula in e es . The co esponding he e ocyclic amines we e ob ained (excep o 3-cyano u an) in yields o 85–94% (Fig. 6). Hyd ogena ion o alipha ic ni iles. Compa ed wi h a oma ic ni iles, he hyd ogena ion o alipha ic ni iles is in gene al mo e challenging. Impo an ly, Fe/Fe–O@SiO2 exhibi ed high ac i i y and selec i i y o hese subs a es, including dini iles, unde iden ical condi ions (Fig. 7). Ini ially, se e al benzylic cyanides we e hyd o- gena ed o he co esponding p ima y amines in excellen yields (Fig. 7, p oduc s 57–72). In e es ingly, he 2-a yle hylamino mo i is a common sca old in many cen al ne ous sys em-ac i e com- pounds. He e, a a ie y o subs i u ed de i a i es we e smoo hly hyd ogena ed and u nished he co esponding p ima y amines in yields o up o 98% (Fig. 7, p oduc s 57–68). Phenylp opylamines a e ano he impo an class o pha maceu ically ele an amines. Fo example, he pa en compound (phenylp opylamine) is used in he syn hesis o ca boxypep idase B- ype enzyme inhibi o s, mus- ca inic ecep o an agonis s and po en ial an icance agen s. He e, i was p epa ed in 94% yield om he co esponding ni ile (Fig. 7, p oduc 69). Al hough 3-(a ylamino)p opaneni iles a e p epa ed in a s aigh o wa d manne om anilines and ac yloni ile, he hyd o- gena ion o such subs a es is di icul because e o-Michael addi- ions can occu . Howe e , his class o compounds was smoo hly hyd ogena ed unde ou condi ions o gi e he espec i e p ima y amines in good yields o up o 85% (Fig. 7, p oduc s 70–72). Finally, a selec ion o alipha ic ni iles was es ed. G a i yingly, Fe/Fe–O@ SiO2 also showed good- o-excellen ac i i y and selec i i y o hese demanding subs a es (Fig. 7, p oduc s 73–78). Pa icula ly in e - es ing is he selec i e educ ion o 5-hexeneni ile (Fig. 7, p oduc 78). No ably, hexame hylenediamine (79), he key eeds ock o he p oduc ion o nylon 66, was p epa ed in 85% yield by di ec hyd o- gena ion o adiponi ile. Simila ly, o he diamines we e ob ained in 90–95% yield (Fig. 7, p oduc s 80 and 81). Syn hesis o a y ni iles. Wi h a wo ldwide p oduc ion o a y amines o >800,000 ons y –1, he hyd ogena ion o a y ni iles cons i u es an impo an indus ial applica ion55. Fa y amines a e aluable oleochemicals mainly used o p oduce ab ic so ene s, lo a ion agen s, emulsi ie s, co osion inhibi o s and lub ica ing addi i es55. Un il oday, he indus ial hyd ogena ion o a y ni iles o amines has elied on well-es ablished Raney Ni o Co ca alys s as well as coppe ch omi e55. These ma e ials ha e conside able ox- ici y issues o biological sys ems. Hence, al e na i e Ru-, Pd- and 100 Fe g = 2.27 Da a 300 K/0 T Fi Fe(0) Fe(II) Fe(III) g = 2.015 g = 1.80 Si O 80 60 40 20 In ensi y (a.u.) In ensi y (a.u.) 0 2 100.0 ab cd 99.9 99.8 T ansmission (%) 99.7 –10 –5 0 Veloci y (mm s–1) 5 10 0 –2 –4 100 200 300 400 Magne ic ield (mT) 500 0 10 nm 5 10 Posi ion (nm) 15 20 25 Fig. 5 | Spec al da a o Fe/Fe–O@SiO2. a–d, HAADF-STEM image (a), dep h p o iles showing he in ensi y dis ibu ion o Fe, Si and O (b), Mössbaue spec um eco ded a 300 K (c) and EPR spec um (X-band 9.090 GHz) eco ded a 77 K (d). NaTuRe CaTalySiS | VOL 5 | JANUARY 2022 | 20–29 | www.na u e.com/na ca al 24 A icles NaTu e CaTalySIS Hyd ogena ion o benzoni iles Fe/Fe–O@SiO2 AI, 50 ba H2, 5–7 ba NH3 i-P OH, 120 °C, 24 h 3: 96%a (93%)a 4: 90% 9: 92% 15: 89% (90%) 16: 80%b 21: 95% 26: 91%b 31: 80%c 27: 94% 32: 80% 37: 89% 43: 85%b 47: 91%b 53: 90% 54: 85% 55: 94%b56: 92% 48: 89% (91%) 49: 90% 50: 91% (93%) 51: 93% 52: 65% 44: 95% 45: 60%d46: 90%e 38: 93% 39: 91% 40: 94% 41: 87% (84%) 42: 89% 33: 70%b34: 95% (93%) 35: 90% 36: 95% 28: 95% 29: 92%b30: 94% (92%) 22: 90%g24: 97% 25: 97% 17: 94% 18: 95% 19: 91% 20: 85% 10: 95% (95%) 11: 97% (96%) 12: 96% 13: 97% 14: 91% 6: 96%a7: 96% (95%) 8: 90% 5: 88%b (89%)b 23: 98%a Hyd ogena ion o he e ocyclic ni iles Fig. 6 | Subs a e scope. Hyd ogena ion o (he e o)a oma ic ni iles. Reac ion condi ions: 0.5 mmol ni ile, 40 mg Fe/Fe–O@SiO2 (8.5 mol% Fe), 3 mg Al oil (20 mol%), 5–7 ba NH3, 50 ba H2, 3 ml i-P OH, 120 °C, 24 h. Isola ed yields a e gi en. aYields we e de e mined by GC using n-hexadecane as s anda d. bWi h 50 mg Fe/Fe–O@SiO2. cWi h 50 mg Fe/Fe–O@SiO2 and 5 mg Al oil. dA 135 °C. eWi h 60 mg Fe/Fe–O@SiO2 and 5 mg Al oil. Wi h 50 mg Fe/Fe–O@SiO2 a 135 °C. gMe hyl 4-cyanobenzoa e was used as subs a e. T anses e i ica ion p oduc wi h i-P OH. Yields in pa en heses e e o he eac ion pe o med in he p esence o 20 mol% Al(i-OP )3. P oduc s we e isola ed as ee amines and con e ed o hei hyd ochlo ide sal s o NMR and high- esolu ion mass spec ome y (HRMS) analysis. NaTuRe CaTalySiS | VOL 5 | JANUARY 2022 | 20–29 | www.na u e.com/na ca al 25 A icles NaTu e CaTalySIS P -based ca alys s we e de eloped55. Un o una ely, none o hese sys ems is comme cially iable due o he high p ice o he p ecious me als. G a i yingly, ou ca alys is capable o hyd ogena ing a y ni iles in a highly selec i e manne . As a esul , se en di e en a y amines we e p epa ed in excellen yields o 95–97% (Fig. 8, p od- uc s 82–88). To u he p o e he syn he ic u ili y and p ac icabili y o ou Fe ca alys , we scaled up bo h he ca alys p epa a ion (up o 12 g; Supplemen a y Table 6) and he ni ile hyd ogena ion p o ocol (up o 20 g). Rega dless o he scale o p epa a ion (1–12 g), all he Fe ma e ials exhibi ed simila ac i i y and selec i i y (Supplemen a y Table 7). Nex , he ca aly ic hyd ogena ion eac ions o ou selec ed a oma ic and alipha ic ni iles we e pe o med using quan i ies o up o 20 g ni ile. Again, simila con e sions and yields we e ob ained o hose achie ed wi h small-scale eac ions using up o 100 mg ni ile (Supplemen a y Fig. 18). Finally, ca alys ecycling was in es iga ed a ull and hal con- e sions, which is an impo an aspec o any he e ogeneous ca a- lys . Indeed, he Fe/Fe–O@SiO2 ca alys could be eused up o he ou h un. A e ha , a dec ease in he p oduc yield was obse ed. Recycling es s pe o med a hal con e sion o 14 h showed a d op in ac i i y om he hi d un onwa d (Supplemen a y Fig. 19). Conclusions We ha e p esen ed he e he de elopmen o a he e ogeneous i on-based ca alys o he hyd ogena ion o ni iles. Key o success was he use o silica-suppo ed Fe nanopa icles co e ed wi h an ul a hin shell o amo phous i on(III) oxide (Fe/Fe–O@SiO2). These co e–shell nanopa icles we e p epa ed by simple imp egna ion o i on(II) ace- a e on comme cial silica and subsequen py olysis unde educ i e condi ions. The low cos and en i onmen ally iendly cha ac e o he ca alys , easy ecycling as well as upscaling o he syn he ic p ocess ep esen key ad an ages and make he ma e ial a ac i e o many applica ions. Impo an ly, he de eloped silica-suppo ed Fe/Fe–O co e–shell ma e ial exhibi ed high chemoselec i i y o he educ ion o unc ionalized and s uc u ally di e se a oma ic, he e ocyclic and alipha ic ni iles, including indus ially ele an a y ni iles, o p o- duce he co esponding p ima y amines in good- o-excellen yields. Aluminium alkoxide species gene a ed in si u om aluminium oil o aluminium iisop opoxide p o ed o be impo an o he co-ca aly ic ac i a ion o he ni ile subs a e. Me hods Gene al conside a ions. All ni iles we e ob ained comme cially om a ious chemical companies. Be o e using, he pu i y o all he ni iles was checked. I on(II) ace a e (99.99%, ca no. 517933-25G) was ob ained om Sigma Ald ich. Silica (Ae osil OX-50) was ob ained om E onik. Ca bon powde (VULCAN XC72R, wi h code XVC72R) was ob ained om Cabo Co po a ion. γ-Al2O3 and MgO we e ob ained om Sigma-Ald ich. Al oil was ob ained om Sigma-Ald ich (Mini Bin, HS23534A). Fo compa ison pu poses, Al oil used o ood co e ing was also pu chased om a local s o e (Kau land, ALUFOLIE; ICP). The pe cen age o aluminium in Al- oil was de e mined by induc i ely coupled plasma (ICP) and i was ound o be 99.97%. DMF was ob ained om Ac os Chemicals. Py olysis expe imen s we e ca ied ou in a Len on ube u nace. PXRD pa e ns we e measu ed a oom empe a u e wi h an Ae is di ac ome e (PANaly ical) in B agg–B en ano geome y equipped wi h Fe/Fe–O@SiO2 AI, 50 ba H2, 5–7 ba NH3 i-P OH, 120 °C, 24 h 57: 95% 58: 91% (93%) 59: 90% 62: 98% 67: 94% 72: 82% 73: 95%c 79: 85%d (89%)d80: 90%e (94%)e81: 95%e 74: 96%c75: 98%c 78: 65%c 77: 95%c 76: 97%a 68: 95% 69: 94% (92%) 70: 84% 71: 85% 63: 90% 64: 80%a 65: 89%b66: 91% 60: 90% 61: 91% Fig. 7 | Subs a e scope. Hyd ogena ion o alipha ic ni iles. Reac ion condi ions: 0.5 mmol ni ile, 40 mg Fe/Fe–O@SiO2 (8.5 mol% Fe), 3 mg Al oil (20 mol%), 5–7 ba NH3, 50 ba H2, 3 ml i-P OH, 120 °C, 24 h. Isola ed yields a e gi en. aWi h 50 mg Fe/Fe–O@SiO2. bA 135 °C. cYields we e de e mined by GC using n-hexadecane as s anda d. dWi h 80 mg Fe/Fe–O@SiO2 and 6 mg Al oil a 135 °C. eWi h 80 mg Fe/Fe–O@SiO2 and 6 mg Al oil a 120 °C. Yields in pa en heses e e o he eac ion pe o med in he p esence o 20 mol% Al(i-OP )3. P oduc s we e isola ed as ee amines and con e ed o hei hyd ochlo ide sal s o NMR and HRMS analysis. NaTuRe CaTalySiS | VOL 5 | JANUARY 2022 | 20–29 | www.na u e.com/na ca al 26 A icles NaTu e CaTalySIS an i on- il e ed Co Kα adia ion sou ce (40 kV, 15 mA, λ = 0.1789 nm) and PIXcell de ec o . Some samples we e measu ed employing an X’Pe PRO MPD di ac ome e (PANaly ical) in B agg–B en ano geome y equipped wi h a Co Kα adia ion sou ce (40 kV, 30 mA, λ = 0.1789 nm), p og ammable di e gence, di ac ed beam an i-sca e sli s and X’Cele a o de ec o . The angula ange o measu emen was 5–105° 2θ (Fe/Fe–O@SiO2 was measu ed in he ange 10–105° 2θ) wi h a s ep size o 0.022 and 0.033° o Ae is and X’Pe PRO MPD di ac ome e s, espec i ely. The c ys alline phases in he expe imen al PXRD pa e ns we e iden i ied using he X’Pe High Sco e Plus so wa e56 in conjunc ion wi h he PDF-4+57 and ICSD58 da abases. The comme cially a ailable silicon s anda d e e ence ma e ial SRM 640 was used o e alua e he line posi ions. Low- esolu ion TEM imaging o he ca alys mo phology was ca ied ou wi h a JEOL mic oscope equipped wi h a LaB6 emission gun, ope a ing a 160 kV. HRTEM images we e ob ained wi h a TITAN 60-300 HRTEM mic oscope equipped wi h an X-FEG- ype emission gun, ope a ing a 80 kV. This mic oscope was equipped wi h a Cs image co ec o and a HAADF-STEM ins umen . The poin esolu ion was 0.06 nm in TEM mode. Elemen al mapping was pe o med by STEM-EDS wi h an acquisi ion ime o 20 min. Fo he HRTEM analysis, he powde samples we e dispe sed in e hanol and ul asonica ed o 5 min. One d op o his solu ion was placed on a coppe g id suppo ing a holey ca bon ilm. The XPS su ace in es iga ion was pe o med on a PHI 5000 Ve saP obe II XPS sys em (Physical Elec onics) wi h a monoch oma ic Al Kα sou ce (15 kV, 50 W) and pho on ene gy o 1,486.7 eV. Dual beam cha ge compensa ion was used o all measu emen s. All spec a we e eco ded in a acuum o 1.3 × 10−7 Pa a 21 °C. A 200-µm-diame e spo was analysed o each sample. The su ey spec a we e measu ed wi h a pass ene gy o 187.850 eV and an elec on ol s ep o 0.8 eV, whe eas he high- esolu ion spec a we e eco ded wi h a pass ene gy o 23.500 eV and an elec on ol s ep o 0.2 eV. The spec a we e e alua ed wi h he Mul iPak so wa e (ULVAC-PHI). All binding ene gies a e e e enced o he C1s ca bon peak a 284.80 eV. The ansmission 57Fe Mössbaue spec a we e collec ed employing a Mössbaue MS96 spec ome e ope a ing in cons an accele a ion mode and equipped wi h a 40 mCi 57Co(Rh) sou ce. The Mössbaue spec a we e i ed wi h he MossWinn so wa e. The isome shi s a e e e enced o α-Fe a oom empe a u e. EPR spec a we e eco ded on a JEOL JES-X-320 spec ome e , ope a ing a he X-band equency (~9.14 GHz) a 77 K, and equipped wi h a JEOL ES 13060DVT5 a iable- empe a u e con ol appa a us. All ca aly ic expe imen s we e ca ied ou in 300 o 100 ml au ocla es (PARR Ins umen ). To a oid unspeci ic eac ions, all ca aly ic eac ions we e ca ied ou ei he in glass ials, which we e placed inside he au ocla e, o in au ocla es i ed wi h a glass/Te lon essel. GC and GC-mass spec ome y (GC–MS) we e pe o med on an Agilen Technologies 6890N ins umen . GC con e sions and yields we e de e mined by GC using lame ioniza ion de ec ion (FID) on an Agilen 6890N ch oma og aph equipped wi h Agilen HP-5MS 30m column (250 mm × 0.25 μm). The mass was de e mined by GC-MS using Agilen 6890N ch oma og aph equipped wi h Agilen HP-5MS 30m column (250 mm × 0.25 μm) and Agilen 5973N Mass Selec i e De ec o (MSD). 1H and 13C NMR spec a we e eco ded on B uke ARX 300 and ARX 400 spec ome e s using [D6]DMSO and CDCl3 sol en s. P epa a ion o Fe/Fe–O@SiO2 on he 1.5 g scale. A magne ic s i ing ba and 280.33 mg Fe(OAc)2 we e ans e ed o a 50- ml ound-bo omed lask and 30 ml DMF was added. The eac ion mix u e was s i ed a 50 °C o dissol e he i on ace a e. To his solu ion, 1.2 g SiO2 (Ae osil OX 50) was added, ollowed by 10 ml DMF. Nex , a e lux condense was i ed o he ound-bo omed lask con aining he eac ion mix u e, which was hen placed in an aluminium block p ehea ed a 150 °C and s i ed o 4 h. Nex , he e lux condense was emo ed and he ound-bo omed lask con aining he eac ion p oduc s was allowed o s and wi hou s i ing o closing o 20 h a 150 °C o he slow e apo a ion o DMF. A e e apo a ion o he sol en and ensu ing comple e d ying, he solid ma e ial was cooled o oom empe a u e and g ound o a ine powde . This powde was py olysed a a de ined empe a u e (400, 600, 800 o 1,000 °C) o 4 h in a ubula u nace unde he low o 20% H2/N2 ( amp: 5 °C min–1, o al low: 3 l h–1) and hen cooled o oom empe a u e. Elemen al analysis o Fe/Fe–O@SiO2 by ICP and CHN analysis e ealed he ollowing (w %) dis ibu ion: Fe = 6.09%; Si = 37.73%; C = 0.11%; H = 0.41%. The B unaue –Emme –Telle su ace a ea was measu ed o be 46.04 m2 g–1. The same p ocedu e was employed o he p epa a ion o Fe(OAc)2-C-800, Fe(OAc)2-γ-Al2O3-800 and Fe(OAc)2-MgO-800. P epa a ion o Fe/Fe–O@SiO2 on 6 and 12 g ba ches. The same p ocedu e was used o he p epa a ion o Fe/Fe–O@SiO2 (Fe(OAc)2-SiO2-800) on he 6 and 12 g scale wi h a sligh modi ica ion o he py olysis p ocedu e, as desc ibed in Supplemen a y Table 6. Gene al p ocedu e o he hyd ogena ion o ni iles. A magne ic s i ing ba and 0.5 mmol o he co esponding ni ile we e ans e ed o a 7-ml glass ial and hen 3 ml i-P OH was added. Nex , 40 mg Fe/Fe–O@SiO2 (8.5 mol% Fe) and 3 mg Al oil ( he Al oil was cu in o small pieces and used in he eac ions) o 20.42 mg Al(i-OP )3 (20 mol%) we e added and he ial was i ed wi h a sep um, cap and needle. Then, he eac ion ials we e placed in a 300-ml au ocla e (eigh ials con aining di e en subs a es a a ime). The au ocla e was closed, lushed wice wi h 20 ba hyd ogen and hen p essu ized wi h 5–7 ba ammonia gas and 50 ba hyd ogen. The au ocla e was placed in an aluminium block p ehea ed a 133 °C and he eac ions we e allowed o p oceed o he equi ed ime unde s i ing. Fa s and oils 82: 97% (97%) 85: 96%b86: 95%b (95%) 87: 97%b 88: 95%b (94%)b 83: 97% 84: 97%a Fa y amines Fa y ni iles P e iously used ca alys s Raney Ni o Co, Cu ch omi e, Ru, Pd, P This Fe ca alys (Oc ylamine) (My is ylamine) (S ea ylamine) (Nonadecanamine) (Hep adecylamine) (Decylamine) (Lau ylamine) Fe/Fe–O@SiO2 NH3, H2 120 °C, 24 h Fig. 8 | Subs a e scope. Syn hesis o a y amines. Reac ion condi ions: 0.5 mmol ni ile, 40 mg Fe/Fe–O@SiO2 (8.5 mol% Fe), 3 mg Al oil (20 mol%), 5–7 ba NH3, 50 ba H2, 3 ml i-P OH, 120 °C, 24 h. Isola ed yields a e gi en. aA 135 °C. bWi h 50 mg Fe/Fe–O@SiO2. Yields in pa en heses e e o he eac ion pe o med in he p esence o 20 mol% Al(i-OP )3. P oduc s we e isola ed as ee amines and con e ed o hei hyd ochlo ide sal s o NMR and HRMS analysis. NaTuRe CaTalySiS | VOL 5 | JANUARY 2022 | 20–29 | www.na u e.com/na ca al 27 A icles NaTu e CaTalySIS Du ing he eac ions, he inside empe a u e o he au ocla e was measu ed o be 120 °C, and his empe a u e was conside ed o be he eac ion empe a u e. A e comple ion o he hyd ogena ion eac ions, he au ocla e was cooled o oom empe a u e. The emaining ammonia and hyd ogen we e discha ged and he ials con aining he eac ion p oduc s we e emo ed om he au ocla e. The solid ca alys was il e ed and washed ho oughly wi h e hyl ace a e. The eac ion p oduc s we e analysed by GC–MS. The co esponding p ima y amines we e pu i ied by column ch oma og aphy (silica, me hanol–dichlo ome hane). The amines we e con e ed o hei espec i e hyd ochlo ide sal and cha ac e ized by GC–MS and NMR analysis. To con e he amines o he hyd ochlo ide sal s, 1–2 ml me hanolic HCl (0.5 M HCl in me hanol) was added o he e hyl ace a e solu ion o he espec i e amine and he mix u e s i ed a oom empe a u e o 4–5 h. Then, he sol en was emo ed and he esul ing hyd ochlo ide sal was d ied unde high acuum. Fo selec ed amines, he yields we e de e mined by GC using he ollowing p ocedu e. A e comple ing he eac ion, n-hexadecane (100 µl) was added as s anda d o he eac ion ials and he eac ion p oduc s we e dilu ed wi h e hyl ace a e ollowed by il a ion using a plug o silica and hen analysed by GC. G am-scale eac ions. A magne ic s i ing ba and he co esponding ni ile we e ans e ed o a glass- i ed 300-ml Pa au ocla e and 15–50 ml i-P OH was added. Nex , he equi ed amoun o ca alys (Fe/Fe–O@SiO2, 8.5–10 mol%) and Al oil (20 mol%; he Al oil was cu in o small pieces and used in he eac ions) we e added. Then, he au ocla e was closed, lushed wice wi h 20 ba hyd ogen and hen p essu ized wi h 5–7 ba NH3 ollowed by 50 ba hyd ogen. The au ocla e was placed in an aluminium block p ehea ed a 133–147 °C (placed 30 min be o e coun ing he eac ion ime o achie e he eac ion empe a u e) and he eac ions we e s i ed o 24 h. Du ing he eac ions, he inside empe a u e o he au ocla e was measu ed o be 120–135 °C. A e comple ion o he eac ions, he au ocla e was cooled o oom empe a u e. The emaining ammonia and hyd ogen we e discha ged, and he eac ion p oduc s we e emo ed om he au ocla e. The solid ca alys was il e ed and washed ho oughly wi h me hanol and e hyl ace a e. The eac ion p oduc s we e analysed by GC–MS and he co esponding p oduc s we e pu i ied by column ch oma og aphy (silica, dichlo ome hane–me hanol) and cha ac e ized by NMR and GC–MS analysis. Ca alys ecycling. A magne ic s i ing ba and 10 mmol benzoni ile we e ans e ed o a 100-ml au ocla e and hen 20 ml i-P OH was added. Nex , 900 mg ca alys (Fe/Fe–O@SiO2) and 408.50 mg Al(i-OP )3 we e added. The au ocla e was closed, lushed wi h 20 ba hyd ogen and hen p essu ized wi h 5–7 ba NH3 and 50 ba H2. The au ocla e was placed in a p ehea ed aluminium block a 130 °C and he eac ions we e s i ed o he equi ed ime. Du ing he eac ions, he inside empe a u e o he au ocla e was measu ed o be 120 °C. A e comple ion o he eac ions, he au ocla e was cooled o oom empe a u e. The emaining ammonia and hyd ogen we e hen discha ged, and he eac ion p oduc s we e emo ed om he au ocla e. Nex , 250 µl n-hexadecane was added as s anda d o he eac ion p oduc s. The ca alys was sepa a ed by il a ion and he il a e con aining he eac ion p oduc s was subjec ed o GC analysis o de e mine he yield o benzylamine. The sepa a ed ca alys was washed wi h e hyl ace a e, d ied unde acuum and used wi hou u he pu i ica ion o eac i a ion o he nex un. Da a a ailabili y All da a a e a ailable om he au ho s upon easonable eques . Recei ed: 31 Ma ch 2021; Accep ed: 10 No embe 2021; Published online: 30 Decembe 2021 Re e ences 1. Belle , M. & Bolm, C. T ansi ion Me als o O ganic Syn hesis (Wiley-VCH, 2008). 2. Negishi, E.-I. Magical powe o ansi ion me als: pas , p esen , and u u e (Nobel lec u e). Angew. Chem. In . Ed. 50, 6738–6764 (2011). 3. Hagen, J.Indus ial Ca alysis: A P ac ical App oach (Wiley‐VCH, 2006). 4. Ca low, C. R., Da idson, M., Ha dac e, C. & Hu chings, G. J. Ca alysis making he wo ld a be e place. Philos. T ans. R. Soc. A 374, 20150089 (2016). 5. Busacca, C. A., Fand ick, D. R., Song, J. J. & Senanayake, C. H. The g owing impac o ca alysis in he pha maceu ical indus y. Ad . Syn h. Ca al. 353, 1825–1864 (2011). 6. Rylande , P. N.Ca aly ic Hyd ogena ion in O ganic Syn heses (Academic P ess, 1979). 7. Alonso, D. A. e al. in Science o Syn hesis: Wa e in O ganic Syn hesis (ed. Kobayashi, S.) 95–119 (Thieme, 2012); h ps://doi.o g/10.1055/ sos-SD-206-00075 8. Oge , C., Balas, L., Du and, T. & Galano, J.-M. A e alkyne educ ions chemo-, egio-, and s e eoselec i e enough o p o ide pu e (Z)-ole ins in poly unc ionalized bioac i e molecules? Chem. Re . 113, 1313–1350 (2013). 9. Massa o, L., Zheng, J., Ma ga i a, C. & Ande sson, P. G. Enan iocon e gen and enan iodi e gen ca aly ic hyd ogena ion o isome ic ole ins. Chem. Soc. Re . 49, 2504–2522 (2020). 10. Chi ik, P. J. I on- and cobal -ca alyzed alkene hyd ogena ion: ca alysis wi h bo h edox-ac i e and s ong ield ligands. Acc. Chem. Res. 48, 1687–1695 (2015). 11. Magano, J. & Dune z, J. R. La ge-scale ca bonyl educ ions in he pha maceu ical indus y. O g. P ocess Res. De . 16, 1156–1184 (2012). 12. Zhong, R., Wie, Z., Zhang, W., Liu, S. & Liu, Q. A p ac ical and s e eoselec i e in si u NHC-cobal ca aly ic sys em o hyd ogena ion o ke ones and aldehydes. Chem 5, 1552–1566 (2019). 13. Fo men i, D., Fe e i, F., Scha nagl, F. K. & Belle , M. Reduc ion o ni o compounds using 3d-non-noble me al ca alys s. Chem. Re . 119, 2611–2680 (2019). 14. Jagadeesh, R. V. e al. Nanoscale Fe2O3-based ca alys s o selec i e hyd ogena ion o ni oa enes o anilines. Science 342, 1073–1076 (2013). 15. Wiesen eld , M. P., Nai oukh, Z., Dal on, T. & Glo ius, F. Selec i e a ene hyd ogena ion o di ec access o sa u a ed ca bo- and he e ocycles. Angew. Chem. In . Ed. 58, 10460–10476 (2019). 16. Mu ugesan, K. e al. Cobal -nanopa icles ca alyzed e icien and selec i e hyd ogena ion o a oma ic hyd oca bons. ACS Ca al. 9, 8581–8591 (2019). 17. Sahoo, B. e al. A obus i on ca alys o he selec i e hyd ogena ion o subs i u ed (iso)quinolones. Chem. Sci. 9, 8134–8141 (2018). 18. Law ence, S. A. Amines: Syn hesis, P ope ies and Applica ions (Camb idge Uni . P ess, 2004). 19. Ricci, A. Amino G oup Chemis y: F om Syn hesis o he Li e Sciences (Wiley-VCH, 2008). 20. Qu eshi, M. H. Top 200 Pha maceu icals by Re ails Sales in 2020 (Uni e si y o A izona, 2020); h ps://nja da son.lab.a izona.edu/si es/nja da son.lab. a izona.edu/ iles/Top%20200%20Pha maceu icals%20By%20Re ail%20 Sales%202020V3.pd 21. Mu ugesan, K. e al. Ca aly ic educ i e amina ions using molecula hyd ogen o syn hesis o di e en kinds o amines. Chem. Soc. Re . 49, 6273–6328 (2020). 22. Hahn, G., Kunnas, P., de Jonge, N. & Kempe, R. Gene al syn hesis o p ima y amines ia educ i e amina ion employing a eusable nickel ca alys . Na . Ca al. 2, 71–77 (2019). 23. Saba ie , P. & Sende ens, J. B. Applica ion aux ni iles de la me hode d’hyd ogena ion di ec e pa ca alyse: syn hese d’amines p imai es, secondai es e e iai es. C. R. Hebd. Séances Acad. Sci. 140, 482–484 (1905). 24. Paal, C. & Ge um, J. Übe ka aly ische Wi kungen kolloidale Me alle de Pla ing uppe. VI. Reduk ionska alysen mi kolloidalem Palladium. Be . D sch. Chem. Ges. 42, 1553–1560 (1909). 25. B aun, V. J., Blessing, G. & Zobel, F. Ka aly ische hyd ie ungen un e D uck bei Gegenwa on nickelsalzen. VI.: ni ile. Be . D sch. Chem. Ges. A 56, 1988–2001 (1923). 26. Raney, M. Me hod o p epa ing ca aly ic ma e ial. US pa en 1563587A (1925). 27. Ca o he s, W. H. & Jones, G. A. The p epa a ion o some p ima y amines by he ca aly ic educ ion o ni iles. J. Am. Chem. Soc. 47, 3051–3057 (1925). 28. Alle , B. V. Raney cobal hyd ogena ion ca alys s. I. The p epa a ion o he ca alys . J. Appl. Chem. 7, 130–134 (2007). 29. Ad kins, H. Reac ions o Hyd ogen wi h O ganic Compounds O e Coppe -Ch omium Oxide and Nickel Ca alys s 53 (Uni . Wisconsin P ess, 1937). 30. F ei elde , M. A low p essu e p ocess o he educ ion o ni iles. Use o hodium ca alys . J. Am. Chem. Soc. 82, 2386–2389 (1960). 31. F eidlin, L. K. & Sladko a, T. A. Ca aly ic educ ion o dini iles. Russ. Chem. Re . 33, 319–330 (1964). 32. Ba ne , C. Hyd ogena ion o alipha ic ni iles o e ansi ion me al bo ides. Ind. Eng. Chem. P od. Res. De . 8, 145–149 (1969). 33. de Belle on, C. & Fouilloux, P. Homogeneous and he e ogeneous hyd ogena ion o ni iles in a liquid phase: chemical, mechanis ic, and ca aly ic aspec s. Ca al. Re . 36, 459–506 (1994). 34. López-De Jesús, Y. M., Johnson, C. E., Monnie , J. R. & Williams, C. T. Selec i e hyd ogena ion o benzoni ile by alumina-suppo ed I –Pd ca alys s. Top. Ca al. 53, 1132–1137 (2010). 35. Lé ay, K. & Hegedűs, L. Recen achie emen s in he hyd ogena ion o ni iles ca alyzed by ansi ional me als. Cu . O g. Chem. 23, 1881–1900 (2019). 36. Yoshimu a, M. e al. Selec i e syn hesis o p ima y amines om ni iles unde hyd ogena ion condi ions. Ad . Syn h. Ca al. 360, 1726–1732 (2018). 37. Liu, Y. e al. Mild palladium-ca alysed highly e icien hyd ogena ion o C≡N, C–NO2, and C=O bonds using H2 o 1 a m in H2O. G een Chem. 21, 830–838 (2019). 38. Wang, H. e al. Quasi Pd1Ni single-a om su ace alloy ca alys enables hyd ogena ion o ni iles o seconda y amines. Na . Commun. 10, 4998 (2019). 39. Ryabchuk, P. e al. In e me allic nickel silicide nanoca alys —a non-noble me al–based gene al hyd ogena ion ca alys . Sci. Ad . 4, eaa 0761 (2018). NaTuRe CaTalySiS | VOL 5 | JANUARY 2022 | 20–29 | www.na u e.com/na ca al 28