Brønsted and Lewis Solid Acid Catalysts in the Valorization of Citronellal

catalysts

Article Brønsted and Lewis Solid Acid Catalysts in the Valorization of Citronellal

Federica Zaccheria 1, Federica Santoro 1, Elvina Dhiaul Iftitah 2 and Nicoletta Ravasio 1,*

1 CNR Institute of Molecular Science and Technology, Via Golgi 19, 20133 Milano, Italy; [email protected] (F.Z.); [email protected] (F.S.) 2 Chemistry Department Faculty of Science, Brawijaya University Jl. Veteran, Malang 65145, Indonesia; [email protected] * Correspondence: [email protected]; Tel.: +39-02-5031-4382

Received: 30 July 2018; Accepted: 12 September 2018; Published: 22 September 2018

Abstract: Terpenes are valuable starting materials for the synthesis of molecules that are of interest to the ﬂavor, fragrance, and pharmaceutical industries. However, most processes involve the use of mineral acids or homogeneous Lewis acid catalysts. Here, we report results obtained in the liquid-phase reaction of citronellal with anilines under heterogeneous catalysis conditions to give tricyclic compounds with interesting pharmacological activity. The terpenic aldehyde could be converted into octahydroacridines with a 92% yield through an intramolecular imino Diels–Alder reaction of the imine initially formed in the presence of an acidic clay such as Montmorillonite KSF. Selectivity to the desired product strongly depended on the acid sites distribution, with Brønsted acids favoring selectivity to octahydroacridine and formation of the cis isomer. Pure Lewis acids such as silica–alumina with a very low amount of alumina gave excellent results with electron-rich anilines like toluidine and p-anisidine. This protocol can be applied starting directly from essential oils such as kafﬁr lime oil, which has a high citronellal content.

Keywords: solid acids; acidic clays; terpenes; citronellal; octahydroacridines; heterogeneous catalysis

1. Introduction Terpenes represent a class of natural compounds suited for the synthesis of several types of molecules useful for the industrial production of intermediates for fragrances, flavors, and pharmaceuticals [1]. α-pinene, e.g., one of the main components of turpentine, can be used as a starting material for the synthesis of β-santalol and sandalwood fragrances. These are valuable alternatives to toxic nitro-musks and low-biodegradability polycyclic musks that are among the commonly used fragrances in European laundry detergents, fabric softeners, cleaning agents, and cosmetic products, therefore ubiquitously present in the aquatic environment [2]. In previous years, the use of terpenes has been widely investigated in the polymer industry because of the strong need for renewable and biodegradable materials in this sector [3,4]. Among terpenic molecules, citronellal, citral, limonene, carene, and pinene are common because of their large occurrence in essential oils and in oils derived from agro-industrial residues. Citronellal and citral offer noteworthy opportunities in synthetic strategies because of their condensation and addition reactions by virtue of their unsaturated aldehydic structure. The abundance of these two molecules in essential oils such as kaffir lime, citronella, lemongrass, and krangean oil, promotes their use as raw materials for the sustainable synthesis of different chemicals, specifically N-containing ones. A Schiff base with antibacterial activity was recently synthesized through acid catalysis from citronellal, [5] while benzimidazole derivatives of both aldehydes were obtained by using microwave irradiation [6]. In this case, the aldehyde reacts with o-phenylenediamine to give the benzimidazol

Catalysts 2018, 8, 410; doi:10.3390/catal8100410 www.mdpi.com/journal/catalysts Catalysts 2018,, 8,, xx FORFOR PEERPEER REVIEWREVIEW 2 of 10

Catalysts 2018, 8, 410 2 of 10 obtained by using microwave irradiation [6]. In this case,, the aldehyde reacts with o-phenylenediamine to give the benzimidazol moiety, but when an aromatic monoamine reacts with moiety,an unsaturated but when aldehyde an aromatic it gives monoamine an N-aryl imine reacts that with can an further unsaturated react with aldehyde the electron it gives-rich an N-arylalkene iminemoiety that through can further the Povarov react with reaction, the electron-rich which isis alkene an an intramolecular intramolecular moiety through imino imino the Povarov Diels–Ald reaction,er reaction which [7 is] an(Scheme intramolecular 1). imino Diels–Alder reaction [7] (Scheme1).

HN N HN N

SchemeScheme 1.1. General intramolecular PovarovPovarov reaction.reaction.

ThisThis classclass ofof organicorganic reactionsreactions hashas attractedattracted interestinterest asas aa usefuluseful routeroute ofof accessaccess toto N-containingN-containing polycyclicpolycyclic structures, structures, such as as substituted tetrahydroquinoline and octahydroacridine octahydroacridine (OHAs) derivatives.derivatives. OctahydroacridineOctahydroacridine systemssystems are are an an interesting interesting class class of of biologically biologically active active molecules molecules in the in fieldthe field of drugs of drugs and pharmaceuticals, and pharmaceuticals, acting as acting gastric as acid gastric secretion acid inhibitors secretion [ 8 inhibitors] in the prevention [8] inin the the of senileprevention dementia of senile [9]. dementia [9]. DespiteDespite severalseveral varyingvarying methodsmethods available offering access to thethe octahydroacridineoctahydroacridine skeleton,skeleton,, suchsuch as as the the Beckmann Beckmann rearrangement rearrangement of oxime of oxime sulfonat sulfonate,e,, [10] the [10 catalytic] the catalytic hydrogenation hydrogenation of acridine of,, acridine,[11] and Friedel [11] and–Crafts Friedel–Crafts acylation,, acylation,[12] the simplest [12] the one simplest remains one the remains acid catalyzed the acid imino catalyzed Diels– iminoAlder Diels–Alderreaction of 2-azadienes. reaction of Because 2-azadienes. of their Becausepoor reactivity of their in poorthe Povarov reactivity reaction, in the 2- Povarovazadienes reaction, need to 2-azadienesbe activated need through to be coordination activated through with acidic coordination catalysts with that acidic enhance catalysts their that electron enhance-defici theirent electron-deficientcharacter. Both Lewis character. and Brønsted Both Lewis acids have and Brønstedbeen used acids,, but,, havedespite been their used, effectiveness but, despite,, many their of effectiveness,these catalysts many show of disadvantages these catalysts,, showsuch as disadvantages, multistep procedures, such as multistep long reaction procedures, times long,, useuse reaction ofof inertinert times,atmosphere, use of inert expensiveness, atmosphere, and expensiveness, tedious work and- tediousup. There work-up.fore, developing Therefore, developing green and green efficient and efficientcatalysts catalysts for this reaction for this reactionis an important is an important challenge. challenge. AnAn efficient efficient synthesis synthesis of of OHAs OHAs starts starts from from citronellal citronellal and and N-arylamines N-arylamines [13,14 [13] (Scheme,14] (Scheme 2). This2). ThisLewis Lewis-acid-acid catalyzed catalyzed imino imino Diels– Diels–AlderAlder reaction reaction is the is most the most atom- atom-economiceconomic way to way OHAs, to OHAs, with withhigh highyields yields and, and,in some in some cases, cases, high highstereoselectivity. stereoselectivity.

O NH2 O NH2 N H

Scheme 2. Reaction of citronellal and anilines into the corresponding octahydroacridine.octahydroacridine. During our recent studies on reductive amination of ketones, [15] the attempts to synthesize During our recent studies on reductive amination of ketones,, [15] the attempts to synthesize amine derivatives of C=O-containing terpenes led us to obtain OHAs structures. Herein, we report amine derivatives of C=O-containing terpenes led us to obtain OHAs structures. Herein, we report our results about the synthesis of OHAs from citronellal promoted by solid acid catalysts. our results about the synthesis of OHAs from citronellal promoted by solid acid catalysts. 2. Results and Discussion 2.. Results and Discussion Previously, we have reported on the use of pre-reduced Cu/SiO2 (Cu/Si) catalysts in the Direct Previously, we have reported on the use of pre-reduced Cu/SiO22 (Cu/Si) catalysts in the Direct Reductive Amination (DRA) of aromatic ketones [15]. This is a signiﬁcant reaction allowing one to Reductive Amination (DRA) of aromatic ketones [15]. This is a significant reaction allowing one to obtain secondary amines in one step starting from a ketone, which is still carried out in the presence obtain secondary amines in one step starting from a ketone,, which isis stillstill carriedcarried outout inin thethe presence of unfavorable reagents such as NaBH3CN or NaBH(OAc)3. The use of a heterogeneous catalyst of unfavorable reagents such as NaBH33CN or NaBH(OAc)33.. The The use use of of a a heterogeneous heterogeneous catalyst catalyst

Catalysts 2018, 8, 410 3 of 10 Catalysts 2018, 8, x FOR PEER REVIEW 3 of 10 basedbased on a non non-noble-noble metal can yield up to 98% of the amine at 100 100 °C◦C in in a a few few hours. hours. Unfortunately, Unfortunately, under the same conditions conditions,, aliphatic terpenic ketones,ketones, in particular menthone, a major component of dementholized mintmint oil,oil, reactreact slowly,slowly, requiring requiring a a longer longer time time span, span although, although selectivity selectivity to to the the amine amine is excellentis excellent (Scheme (Scheme3). 3).

O NH2

Cu/SiO2 + 100°C, toluene, H2 (1 atm) N H

Conv= 40% Sel = 98% Scheme 3. Amination of menthone with aniline (Conv(Conv == Conversion;Conversion; Sel = Selectivity) Selectivity)..

This prompted us to test the reaction of terpenic molecules with a greater number of reactive aldehyde groups. Table1 1 reports reports selected selected results results obtainedobtained inin thethe reactionreaction ofof citronellal citronellal underunder varyingvarying conditions conditions and in the presence of several catalytic materials. Surprisingly,Surprisingly, thethe reaction reaction of ofcitronellal citronellal with with aniline aniline under under DRA conditions, DRA conditions that is,, inthat the is presence, in the of pre-reduced Cu catalyst and molecular H , was quick and gave the corresponding OHA as the only presence of pre-reduced Cu catalyst and molecular2 H2, was quick and gave the corresponding OHA productas the only in highproduct yield in (Table high 1yield, entry (Table 3). A 1, blank entry test 3). carried A blank out test without carried a out catalyst without led toa catalyst the formation led to ofthe the formation intermediate of the imine, intermediate without imine, observing without the formationobserving ofthe the formation Diels–Alder of the product, Diels–Alder showing product, that, inshowing this specific that, reaction,in this specific a catalyst reaction is only, necessarya catalyst for is theonly second necessary step, andfor a the reductive second environment step, and a isreductive not required environment (Table1, entriesis not required 1–2). Therefore, (Table 1, both entries pre-reduction 1–2). Therefore of the, catalystboth pre and-reduction molecular of Hthe2 could be avoided, and the unreduced CuO/SiO (CuO/Si) catalyst was used at room temperature in catalyst and molecular H2 could be avoided, and2 the unreduced CuO/SiO2 (CuO/Si) catalyst was theused presence at room of temperature air, giving comparablein the presence results. of air, giving comparable results. This reactivity is ascribed to the Lewis acid character of CuO/SiO . Thus, the chemisorption– This reactivity is ascribed to the Lewis acid character 2 of CuO/SiO2. Thus, the hydrolysischemisorption technique–hydrolysis used technique in the preparation used in the of preparation this catalyst of allows this catalyst us to reach allows a high us to dispersion reach a high of CuOdispersion on the of silica CuO surface on the [ 16silica]. The surface CuO particles[16]. The are CuO small particles and defective, are small thus and giving defective, an account thus giving of the Lewisan account acidity of ofthe this Lewis material, acidity able of this to promote, material e.g.,, able the to ringpromote opening, e.g., of the epoxides ring opening with alcohols of epoxides [17]. Afterwith alcoholsthe reduction [17]. pre-treatment,After the reduction highly predefective-treatment metal, highly nanoparticles defective are metal formed nanoparticles showing Lewis are acidformed activity showing in a wide Lewis range acid of activityreactions, in particularly a wide range in the of one-pot reactions, production particularly of valeric in the esters one from-pot γ production-valerolactone of valeric (GVL) esters and alcoholsfrom -valerolactone [18]. (GVL) and alcohols [18]. Therefore, otherother solid acids can be used in the synthesis of OHA. Because of our experience in the use use of of amorphous amorphous mixed mixed oxides, oxides we tested, we some tested of some them in of this them reaction, in this namely, reaction a silica–alumina, namely, a cracking catalyst with a 13% content of Al O (SiAl 13) and a 0.6% Alumina on silica (SiAl0.6). silica–alumina cracking catalyst with a 13% content2 3 of Al2O3 (SiAl 13) and a 0.6% Alumina on silica Furthermore,(SiAl0.6). Furthermore, we also tested we also two tested commercial two commercial acid-treated acid clays,-treated namely, clays, Montmorillonitenamely, Montmorillonite K10 and KSF.K10 and Clays KSF. are Clays versatile are materialsversatile widelymaterials used widely for various used for applications, various applications including catalytic, including ones. catalytic Thus, increasinglyones. Thus, stringentincreasingly environmental stringent issues environm andental process issues optimization and process call for optimization the substitution call of for liquid the acidssubstitution by more of sustainable liquid acids solid by materials.more sustainable Clays have solid both materials. Brønsted Clays (B) and have Lewis both (L) Brønsted sites, the amount(B) and andLewis the (L) strength sites, of the which amount can be and modified the strength by acid treatment. of which Commercial can be modified acid-treated by acidclays treatment.are widely usedCommercial industrially acid-t asreated acid catalysts,clays are thereforewidely used they industrially are well-defined as acid and catalysts reliable, therefore materials. they K10 are is commonlywell-defined used and inreliable the formation materials. of K10 a C–N is commonly bond as an used alternative in the formation to HCl, e.g., of ina C the–N synthesisbond as an of alkylquinolinesalternative to HCl, [19 ].e.g., in the synthesis of alkylquinolines [19]. It shouldshould bebe emphasizedemphasized that,that, under under acidic acidic conditions, conditions citronellal, citronellal can can undergo undergo the theene enereaction reaction to giveto give isopulegol isopulegol (Scheme (Scheme4). Although 4). Although this is an this interesting is an interesting reaction, as reaction isopulegol, as can isopulegol be hydrogenated can be tohydrogenated menthol, in theto menthol, present case in the this present is a secondary case this reaction is a secondary that lowers reaction the yield that in OHAlowers and, the therefore, yield in aOHA fine-tuning and, therefore of the acidity, a fine is-tuning mandatory of the inacidity order is to mandatory reach a high in selectivity order to reach to the a desiredhigh selectivity product. to the desired product.

CatalystsCatalysts 20182018,, 88,, 410x FOR PEER REVIEW 4 ofof 1010 Catalysts 2018, 8, x FOR PEER REVIEW 4 of 10

Acid catalyst Acid catalyst O OH O OH

H H

Scheme 4. Ene reaction of citronellal into isopulegol. Scheme 4. Ene reaction of citronellal intointo isopulegol.isopulegol. Results reported in Table 1 show significant differences in selectivity. Considering the reaction Results reported in Table1 1 show show significant significant differences differences in in selectivity. selectivity. Considering Considering thethe reaction reaction between citronellal and aniline, the two amorphous silica–aluminas and Montmorillonite K10 gave a between citronellal andand aniline,aniline, thethe two amorphous silica–aluminassilica–aluminas andand MontmorilloniteMontmorillonite K10K10 gavegave a low yield in OHA due to a significant formation of isopulegols (Table 1, entries 6,11,13). This agrees lowlow yieldyield inin OHAOHA duedue toto aa significantsignificant formationformation ofof isopulegolsisopulegols (Table(Table1 1,, entries entries 6,11,13). 6,11,13). This This agrees agrees with our previous findings [20], showing that these three catalysts are highly active in the cyclization withwith ourour previousprevious findingsfindings [[2020],], showingshowing thatthat thesethese threethree catalystscatalysts areare highlyhighly activeactive inin the cyclization of citronellal to isopulegol. However, moving from aniline to electron-rich amines, the selectivity to of citronellalcitronellal toto isopulegol.isopulegol. However, moving from aniline to electron-richelectron-rich amines,amines, thethe selectivityselectivity toto OHA increased, reaching a remarkable 96% yield in the case of p-anisidine over the catalyst with a OHA increased,increased, reaching a remarkable 96%96% yield in thethe casecase ofof p-anisidinep-anisidine overover thethe catalystcatalyst withwith aa very low amount of alumina (Table 1, entry 9). veryvery lowlow amountamount ofof aluminaalumina (Table(Table1 ,1, entry entry 9). 9).

Table 1. Synthesis of octahydroacridines starting from (±)± -Citronellal and anilines. TableTable 1.1. SynthesisSynthesis ofof octahydroacridinesoctahydroacridines startingstarting fromfrom ((±))-Citronellal-Citronellal and anilines.

H H H H

+ + + + + + O NH2 N N OH O H H NH2 N H N H OH H H H H trans cis trans cis Entry Catalyst Amine Conditions T h Conv. % OHAs Sel % Cis/Trans. Imine Sel % Isopulegols % 1 73 87 ◦ -- - 1 None aniline Toluene, 100 C, N2 7.5T 96Conv. OHAs Cis/Tra Imine 93 Isopulegols Entry Catalyst Amine Conditions◦ 2 None aniline Heptane, 25 C, air 1Th 96Conv.% OHAsSel - % Cis/ -ns.Tra ImineSel 83 % Isopulegols - % Entry Catalyst Amine Conditions◦ 3 Cu/Si aniline Toluene, 100 C, H2 1h >99% Sel97 % 37/63ns. Sel <1 % 1 % 4 CuO/Si aniline Toluene,Toluene, 25 ◦C, air 11 >9973 96 45/5587 - 2 1 None aniline Toluene, 1 73 - - 87 - 1 5 CuO/SiNone p-anisidineaniline Toluene,100 25°C,◦C, N air2 1 >99 59- 61/39- - 32 - 100 °C, N2 7.5 96 93 6 SiAl 0.6 aniline Heptane, 25 ◦C, air 17.5 >9996 71 41/59 <193 24 Heptane,◦ 2 7 SiAlNone 0.6 anilineaniline Heptane,Heptane, 0 C, air 11 >9996 79- 49/51- 83 - 17 - 2 None aniline 25 °C,◦ air 1 96 - - 83 - 8 SiAl 0.6 aniline Dioxane,25 °C, 25 C,air air 1 >99 82 58/42 - 14 9 SiAl 0.6 p-anisidine Heptane,Toluene 25 ◦C,, air 1 >99 96 51/49 <1 2 3 Cu/Si aniline Toluene, 1 >99 97 37/63 <1 1 100 °C,◦ H2 3 10 SiAlCu/Si 0.6 toluidineaniline Heptane, 25 C, N2 21 >99>99 9097 37/6337/63 -<1 1 1 100 °C, H2 ◦ 11CuO/S SiAl 13 aniline Heptane,Toluene 25 C, N, 2 1 >99 64 41/59 - 23 4 CuO/S aniline Toluene, 1 >99 96 45/55 - 2 4 12 SiAli 13 anilineaniline Heptane,25 °C, 0 ◦C, air air 11 >99>99 8696 58/4245/55 -- 11 2 i 25 °C, air 13 Mont K10 aniline Heptane, 25 ◦C, air 1 >99 70 55/45 - 14 CuO/S p-anisidin Toluene◦ , 5 14 MontCuO/S KSF paniline-anisidin Heptane,Toluene 25 C,, air 11 >99>99 9259 71/2961/39 -- 5 32 5 i e 25 °C, air 1 >99 59 61/39 - 32 15 Monti KSF anilinee Heptane,25 °C, 0 ◦C, air air 1 >99 85 73/27 - 10 It is worthSiAl emphasizing that onlyHeptane, Lewis acid sites could be detected on the surface of SiAl0.6, as shown by 6 SiAl aniline Heptane, 1 >99 71 41/59 ˂1 24 6FT IR spectra0.6 of adsorbedaniline pyridine,25 °C, whileair the other1 solids>99 used showed71 also41/59 Brønsted acid˂1 sites (Table242 ). OHA: octahydroacridine.0.6 25 °C, air SiAl Heptane, 7 SiAl aniline Heptane, 1 >99 79 49/51 - 17 7 0.6 aniline 0 °C, air 1 >99 79 49/51 - 17 0.6 0 °C, air Table 2.SiAlQuantitative determinationDioxane, of acidic sites versus products distribution in the reaction of 8 SiAl aniline Dioxane, 1 >99 82 58/42 - 14 8citronellal 0.6 with aniline.aniline 25 °C, air 1 >99 82 58/42 - 14 0.6 25 °C, air SiAlCatalyst p-anisidinAcidic SitesHeptane, (mmol ) % OHA % Cis % Isopulegols 9 SiAl p-anisidin Heptane, py/gcat1 >99 96 51/49 <1 2 9 0.6 e 25 °C, air 1 >99 96 51/49 <1 2 0.6KSF e 0.15725 °C, air 0.025 92 71 5 SiAlK10 0.077Heptane, 0.042 70 55 14 10 SiAl toluidine Heptane, 2 >99 90 37/63 - 1 10 0.6SiAl13 toluidine 0.039 25 °C, N2 0.1632 >9964 90 4137/63 23 - 1 0.6 25 °C, N2 SiAlSiAl 0.6 -Heptane, 0.005 71 41 24 11 SiAl aniline Heptane, 1 >99 64 41/59 - 23 11 13 aniline 25 °C, N2 1 >99 64 41/59 - 23 13 25 °C, N2 SiAl Heptane, 12 SiAl aniline Heptane, 1 >99 86 58/42 - 11 12 13 aniline 0 °C, air 1 >99 86 58/42 - 11 13 0 °C, air

Catalysts 2018, 8, x FOR PEER REVIEW 5 of 10

Mont Heptane, 13 aniline 1 >99 70 55/45 - 14 K10 25 °C, air Mont Heptane, 14 aniline 1 >99 92 71/29 - 5 KSF 25 °C, air Mont Heptane, 15 aniline 1 >99 85 73/27 - 10 KSF 0 °C, air It is worth emphasizing that only Lewis acid sites could be detected on the surface of SiAl0.6, as shown by FT IR spectra of adsorbed pyridine, while the other solids used showed also Brønsted acid sites (Table 2). OHA: octahydroacridine.

Table 2. Quantitative determination of acidic sites versus products distribution in the reaction of citronellal with aniline.

Acidic Sites Catalyst % OHA % Cis % Isopulegols (mmolpy/gcat) KSF 0.157 0.025 92 71 5 K10 0.077 0.042 70 55 14 SiAl13 0.039 0.163 64 41 23 Catalysts 2018,SiAl8, 410 0.6 - 0.005 71 41 24 5 of 10

Figure 1 shows the comparison of the two amorphous silica–alumina catalysts and the two clays:Figure it is1 evident shows thethat comparison both Brønsted of the and two Lewis amorphous sites we silica–aluminare present on catalysts the surface and theof the two catalyst clays: itcontaining is evident that13% both of alumina Brønsted and and on Lewis the two sites clays, were presentwhereas on the the Brønsted surface of ones the catalystwere absent containing on the 13%low of-loading alumina one. and on the two clays, whereas the Brønsted ones were absent on the low-loading one.

L P L L B B/L P B B+L L

SiAl 135

KSF Absorbance Absorbance (a.u.)

K10

SiAl 0,6

1650 1600 1550 1500 1450 1400 Wavenumbers (cm-1) FigureFigure 1.1. FTIR spectra spectra of of pyridine pyridine adsorbed adsorbed on on the the silica silica–alumina–alumina catalysts: catalysts: P = P physisorbed, = physisorbed, B = BBrønsted = Brønsted sites sites,, L = L Lewis = Lewis acid acid sites sites..

−1 −1 −1 Thus,Thus, in in SiAl SiAl 13, 13, the the bands bands at at 1640 1640 cm cm−1, 1547 cm−1, ,and and 1492 1492 cm cm−1 areare the the intense intense modes modes of ofpyridinium pyridinium cations cations associated associated with with a a total total proton proton transfer transfer from the Brønsted Brønsted acidic acidic surface surface OHOH −1 −1 groupgroup to to the the basic basic molecule, molecule, whereas whereas the the bands bands at at 1623 1623 cm cm−1and and 1455 1455 cm cm−1 areare duedue toto thethe pyridine pyridine 3+ molecularlymolecularly coordinated coordinated on on Al Al3+ cations,cations, actingacting asas LewisLewis acidacid sitessites [[2121].]. However,However, the the spectra spectra of of pyridine pyridine adsorbed adsorbed on on clays, clays, KSF, KSF, and and K10, K10, showed showed the the presence presence of of both both LewisLewis and and Brønsted Brønsted sites, sites, although although the the former former in in lower lower concentration. concentration. −1 −1 −1 TheThe presencepresence ofof weakweak absorptionabsorption bandsbands atat 15451545 cmcm−1,, 149 14900 cm cm−1, and, and 1450 1450 cm cm−1, indicating, indicating the thepresence presence of of both both Brønsted Brønsted and and Lewis Lewis acidity acidity in in low low concentrations, concentrations, has already been been reported, reported, althoughalthough in in the the Diffuse Diffuse Reflectance Reflectance Infrared Infra Fourierred Fourier Transform Transform (DRIFT) (DRIFT) spectra ofspectra pyridine of adsorbed pyridine onadsorbed the surface on the of montmorillonitesurface of montmorillonite K10 [22], while K10 [ the22],presence while the of presence medium-weak of medium Brønsted-weak sites Brønsted and strong Lewis sites was evidenced by a range of complementary experimental techniques by Lenarda et al. [23]. However, the number of these sites has never been quantified. Table2 shows that B sites in KSF are twice the number of those in SiAl 13, that in turn are twice the number of those in K10, in agreement with previous determinations through NH3 chemisorption of acidic sites in KSF and K10 [24]. Comparing the results obtained with the relative number of acidic sites reported in Table2, it is apparent that, for catalysts containing both B and L sites, the yield in OHA increases with the number of B sites, while the amount of isopulegols increases with the number of L sites (Figure2). The catalyst that contains only L sites showed an intermediate behavior, giving only 71% of OHA in the reaction of citronellal and aniline (Table1, entry 6). However, by increasing the electro-donor ability of the substituent, this catalyst gave excellent results, reaching a 96% yield in the reaction with anisidine (Table1, entry 9). According to our results, differences in acidity can also influence products stereochemistry. Catalysts 2018, 8, x FOR PEER REVIEW 6 of 10

sites and strong Lewis sites was evidenced by a range of complementary experimental techniques by Lenarda et al. [23]. However, the number of these sites has never been quantified. Table 2 shows that B sites in KSF are twice the number of those in SiAl 13, that in turn are twice the number of those in K10, in agreement with previous determinations through NH3 chemisorption of acidic sites in KSF and K10 [24]. Comparing the results obtained with the relative number of acidic sites reported in Table 2, it is apparent that, for catalysts containing both B and L sites, the yield in OHA increases with the

Catalystsnumber2018 of, B8, 410sites, while the amount of isopulegols increases with the number of L sites (Figure6 2). of 10

Figure 2. Trend in products distribution versus the concentration of B acidic sites (A) and L sites (B). Figure 2. Trend in products distribution versus the concentration of B acidic sites (A) and L sites (B). Laschat and Lauterwein [25] reported on the use of several homogeneous Lewis acids at −78 ◦C in CHThe2Cl2 catalystand drew that the contains conclusion only that L thesites catalyst showed plays an onlyintermediate a minor role behavior in determining, giving only the cis/trans 71% of ratio.OHA In in agreement the reaction with of this citronellal work, Kouznetsov and aniline [26 ] (Table found that1, entry the bulkiness 6). However, of the by aniline increasing derivative the haselectro a pivotal-donor role ability when of both the Lewissubstituent and Brønsted, this catalyst acids gave are used excellent as a catalyst, results,allowing reaching to a reach96% yield 97% ofin thethetrans reactionisomer with for anisidine the N-benzyl-aniline (Table 1, entry derivative 9). at room temperature. However, in the presence of TiCl3,According the cis/trans toratio our results did not, differences change significantly in acidity bycan changing also influence the amine products [9]. stereochemistry. ReportsLaschat onand heterogeneous Lauterwein [25 catalysts] reported in on this the specific use of reaction several homogeneous are rare; ZnCl 2Lewissupported acids at on − silica78 °C wasin CH used2Cl2 in and the drew presence the of conclusion microwaves, that giving the catalyst a 78% plays yield only in the a case minor of roleaniline, in determining with a cis/trans the ratiocis/trans of 1:1 ratio. [13]. In The agreement introduction with of this electron-donor work, Kouznetsov or electron-withdrawer [26] found that the substituents bulkiness onof thethe anilineaniline moiety influenced both yield and stereochemistry of the products. In the present case, differences in stereochemistry are evident. Specifically, pure Lewis solids (both CuO/silica and SiAl 0.6) promote 40/60 mixtures with a slight excess of trans isomers. When introducing B acid sites, the number of cis isomers increases, and the solid with a more prominent B acid character, namely, KSF, favors the formation of the cis-isomer (Table1, entries 14 and 15, Figure2A). This could be because of the different adsorption of the intermediate imine on the catalyst surface, with Brønsted sites binding strongly to the N atom and favoring the transition state leading to the cis isomer. Catalysts 2018, 8, x FOR PEER REVIEW 7 of 10

derivative has a pivotal role when both Lewis and Brønsted acids are used as a catalyst, allowing to reach 97% of the trans isomer for the N-benzyl-aniline derivative at room temperature. However, in the presence of TiCl3, the cis/trans ratio did not change significantly by changing the amine [9]. Reports on heterogeneous catalysts in this specific reaction are rare; ZnCl2 supported on silica was used in the presence of microwaves, giving a 78% yield in the case of aniline, with a cis/trans ratio of 1:1 [13]. The introduction of electron-donor or electron-withdrawer substituents on the aniline moiety influenced both yield and stereochemistry of the products. In the present case, differences in stereochemistry are evident. Specifically, pure Lewis solids (both CuO/silica and SiAl 0.6) promote 40/60 mixtures with a slight excess of trans isomers. When introducing B acid sites, the number of cis isomers increases, and the solid with a more prominent B acid character, namely, KSF, favors the formation of the cis-isomer (Table 1, entries 14 and 15, Figure 2A). This could be because of the different adsorption of the intermediate imine on the catalyst Catalystssurface,2018 with, 8, 410Brønsted sites binding strongly to the N atom and favoring the transition state leading7 of 10 to the cis isomer. The solid catalyst is also reusable, allowing to obtain the desired OHA product, although in The solid catalyst is also reusable, allowing to obtain the desired OHA product, although in lower lower yield (75%) under non-optimized conditions, such as without washing the catalyst. yield (75%) under non-optimized conditions, such as without washing the catalyst. This protocol can be successfully applied directly to essential oils. Thus, the reaction of kaffir This protocol can be successfully applied directly to essential oils. Thus, the reaction of kafﬁr limelime oiloil withwith anilineaniline in heptaneheptane and in the presencepresence of clayclay KSFKSF producedproduced an oiloil containingcontaining 70% of octahydroacridine inin oneone hourhour (Scheme(Scheme5 5).).

KSF, aniline Kaffir Lime

essential Oil N 25 C, heptane H

70%

Scheme 5. Octahydroacridine formation, starting directly from kaffir lime. Scheme 5. Octahydroacridine formation, starting directly from kaffir lime. 3. Experimental Section 3. Experimental Section 3.1. Materials and Methods 3.1. Materials and Methods (±)-Citronellal (>95%), aniline (>99%), p-anisidine (>99%) and toluidine (99%) were purchased from(±) Sigma-Aldrich-Citronellal (>95%), (Milan, aniline Italy) and (>99%), used p without-anisidine pretreatment. (>99%) and Kaffirtoluidine lime (99%) (Citrus were hystrix purchased D.C) oil wasfrom obtained Sigma-Aldrich from the (Milan Essential, Italy Oil’s) and Institute used without Atsiri, pretreatment. Brawijaya University, Kaffir lime Malang (Citrus (RI). hystrix D.C) oil 2 was obtainedSiO2–Al2 fromO3 13 the (Surface Essential Area Oil’s = 485Institute m /g, Atsiri, Pore Brawijaya Volume =University, 0.79 mL/g), Malang was purchased(RI). from 2 2 Sigma-Aldrich;SiO2–Al2O3 SiO13 2 (S–Alurface2O3 0.6Area (Surface = 485 Area m /g, = P 483ore mVolume/g, Pore = Volume0.79 mL/g), = 1.43 was mL/g) purchased was kindly from providedSigma-Aldrich; by GRACE SiO2– Davison.Al2O3 0.6 Montmorillonite (Surface Area = KSF 483 and m2/g, Montmorillonite Pore Volume K10= 1.43 were mL/g) purchased was kindly from 2 Sigma-Aldrich;provided by GRACE SiO2 (Surface Davison. Area Montmorillonite = 460 m /g, Pore KSF Volumeand Montmorillonite = 0.74 mL/g) wasK10 purchasedwere purchased from Flukafrom (Milan,Sigma-Aldrich; Italy). SiO2 (Surface Area = 460 m2/g, Pore Volume = 0.74 mL/g) was purchased from Fluka (Milan, Italy). 3.2. Copper Catalyst Preparation 3.2. Copper Catalyst Preparation CuO/SiO2 (with 16% of copper loading) catalyst was prepared by the chemisorption–hydrolysis 2+ techniqueCuO/SiO as reported2 (with 16% [16 ],of by copper adding loading) the support catalyst to anwas aqueous prepared [Cu(NH by the3 chemisorption)4] solution prepared–hydrolysis by droppingtechnique NHas reported4OH into [16 a Cu(NO], by adding3)2·3H the2O solutionsupport to until an pHaqueous 9 had [Cu(NH been reached.3)4]2+ solution After 20 prepared min under by ◦ stirring,dropping the NH slurry,4OH in heldto a inCu(NO an ice3) bath2·3H2 atO 0solutionC, was until diluted pH 9 with had water. been reached. The solid After was 20 separated min under by ◦ filtrationstirring, the with slurry a Büchner, held in funnel, an ice washedbath at 0 with °C, water,was diluted dried with overnight water. at The 120 solidC, and was calcined separated in airby ◦ atfiltration 350 C. with a Buchner̈ funnel, washed with water, dried overnight at 120 °C, and calcined in air at 350 °C. 3.3. Catalyst Characterization 3.3. CatalystFTIR Spectra Characterization of adsorbed pyridine. The FT-IR studies of probe molecules (pyridine) adsorption and desorption were carried out with a BioRad FTS-60 (Segrate, Italy) spectrophotometer equipped with a mid-IR MCT detector. The experiments were performed on a sample disk (15–20 mg) after a simple calcination treatment (180 ◦C, 20 min air + 20 min under vacuum). One spectrum was collected before probe molecule adsorption as a blank experiment. Therefore, pyridine adsorption was carried out at room temperature, and the following desorption steps were performed from room temperature to 250 ◦C. The spectrum of each desorption step was acquired every 50 ◦C after cooling the sample. For quantitative analysis, the amount of adsorbed pyridine (mmolPy/gcat) was calculated on the basis of the relationship reported by Emeis [27] from the integration of diagnostic bands evaluated in the spectra registered at 150 ◦C.

3.4. Catalytic Tests

Prior to the catalytic tests, the catalysts (100 mg) were pretreated as follows: SiO2–Al2O3 was calcined for 20 min in air at 180 ◦C and 20 min under vacuum at the same temperature; Montmorillonite Catalysts 2018, 8, 410 8 of 10

◦ KSF and Montmorillonite K10 were dried in an oven at 120 C for one night; CuO/SiO2 was calcined for 20 min in air at 270 ◦C and 20 min under vacuum at the same temperature, and a following reduction step with H2 (1 atm) was performed for Cu/SiO2-catalyzed reactions. In a glass reaction vessel containing the pretreated catalyst, a mixture of citronellal (0.55 mmol) and amine (1.66 mmol) in 5 mL of solvent was charged. Therefore, the reaction proceeded with the proper temperature conditions under magnetic stirring. At the desired reaction time, the mixture was separated by simple filtration from the catalysts, and the products were analyzed by GC–MS (5%-phenyl-methyl polysiloxane column) and NMR (Bruker 500 MHz). For the recycling tests, the solution was separated from the catalyst after the first run by simple decantation, and a fresh reaction mixture was charged in the reactor without treating the catalyst. In the experiment carried out with Kaffir lime oil, a mixture containing 100 µL of essential oil and 1.66 mmol of aniline in 5 mL of solvent was charged in a glass reaction vessel containing 100 mg of KSF dried overnight at 120◦C. 1 HNMR data (300 MHz, CDCl3) (See Supplementary Materials for NMR Spectra) A) cis isomer: δ = 0.98 (d, J = 6.4, 3 H, CH3), 1.07–1.89 (m, 2H, CH2), 1.13–1.97 (m, 2H, CH2), 1.31 (s, 3H, CH3), 1.33 (s, 1 H, CH), 1.38 (s, 1 H, CH), 1.41 (s, 3H, CH3), 1.83 (m, 1H, CH), 1.86 (m, 1H, CH), 3.69 (bs, NH), 3.92 (dt, 1H, CH), 6.51 (dd, 1H, CH), 6.68 (m, 1H, CH), 7.04 (m, 1H, CH), 7.22 (dd, 1H, CH). MS (EI): m/z (%) = 229.2 (34), 214.2 (100), 158.1 (12), 144.1 (29). B) trans isomer: 0.99–1.77 (m, 2H, CH2), (d, J = 6.5, 3H, CH3), 1.13 (m, 1H, CH), 1.19 (s, 3H, CH3), 1.33–1.83 (m, 2H, CH2), 1.40 (m, 1H, CH), 1.42 (s, 3H, CH3), 1.97 (m, 1H, CH), 3.16 (dt, 1H, CH), 3.69 (bs, NH), 6.52 (dd, 1H, CH), 6.73 (s, 1H, CH), 7.04 (m, 1H, CH), 7.04 (m, 1H, CH), 7.31 (dd, 1H, CH). MS (EI): m/z (%) = 229.2 (55), 214.2 (100), 158.1 (18), 144.1 (33).

4. Conclusions The introduction of solid acid materials in organic synthesis is an ambitious target of green chemistry. Thus, a heterogeneous catalyst could be ﬁltered off at the end of the reaction without the need of a neutralization step followed by washing and the time-consuming workup of the reaction mixture producing signiﬁcant amounts of inorganic salts. N-containing heterocycles are the most abundant class of compounds synthesized by the pharmaceutical industry [28]. Here, we show that very simple inorganic materials such as Montmorillonite, that is a natural acidic clay, can effectively promote the synthesis of N-containing octahydroacridines with a 92% yield in one hour and at room temperature, starting from anilines and citronellal, the main component of several essential oils. In some cases, yields were almost quantitative due to high selectivity with respect to the competing ene-reaction to isopulegols. Both high selectivity and the use of solid acid catalyst allowed us to obtain a negligible waste process, showing high sustainability.

Supplementary Materials: The following are available online at http://www.mdpi.com/2073-4344/8/10/410/s1, Figure S1: 1HNMR of octahydroacridine obtained from citronellal and aniline. Author Contributions: Conceptualization, F.Z. and N.R.; Investigation, F.S. and E.D.I.; Writing-Original Draft Preparation, F.Z.; Writing-Review & Editing, N.R.; Supervision, N.R. Funding: This research received no external funding. Acknowledgments: Institut Atsiri Universitas Brawijaya, Malang, Indonesia is kindly acknowledged for a sample of Kaffir Lime essential oil. Conflicts of Interest: The authors declare no conflict of interest.

References

1. Monteiro, J.L.F.; Veloso, C.O. Catalytic conversion of terpenes into ﬁne chemicals. Top. Catal. 2004, 27, 169–180. [CrossRef] 2. Ravasio, N.; Zaccheria, F.; Guidotti, M.; Psaro, R. Mono- and bifunctional heterogeneous catalytic transformation of terpenes and terpenoids. Top. Catal. 2004, 27, 157–168. [CrossRef] Catalysts 2018, 8, 410 9 of 10

3. Zhang, D.; del Rio-Chanona, E.A.; Shah, N. Screening synthesis pathways for biomass-derived sustainable polymer production. ACS Sustain. Chem. Eng. 2017, 5, 4388–4398. [CrossRef] 4. Wilbon, P.A.; Chu, F.; Tang, C. Progress in renewable polymers from natural terpenes, terpenoids, and rosin. Macromol. Rapid Commun. 2013, 34, 8–37. [CrossRef][PubMed] 5. Kinanthi, R.; Puspitasari; Dwi, F.; Warsito, S.; Farid Rahman, M. Synthesis of schiff base from citronellal in kaffir lime oil (Citrus hystrix D.C) using acid catalyst. In Proceedings of the 1st International Conference of Essential Oils (ICEO2017), Malang, Indonesia, 11–12 October 2017. 6. Warsito; Ramadhan, S.D.; Al Karoma, D.; Zulfa, A. Comparison of synthesis of some benzimidazole derivatives of cytronellal and citral. In Proceedings of the 1st International Conference of Essential Oils (ICEO2017), Malang, Indonesia, 11–12 October 2017. 7. Kouznetsov, V.V. Recent synthetic developments in a powerful imino Diels–Alder reaction (Povarov reaction): Application to the synthesis of N-polyheterocycles and related alkaloids. Tetrahedron 2009, 65, 2721–2750. [CrossRef] 8. Canas-Rodriguez, A.; Canas, R.G.; Mateo-Bernardo, A. Tricyclic inhibitors of gastric acid secretion. Part, V. Octahydroacridines. An. Quim. Ser. C. 1987, 83, 24–27. 9. Mayekar, N.V.; Nayak, S.K.; Chattopadhyay, S. Two convenient one-pot strategies for the synthesis of octahydroacridines. Synth. Commun. 2004, 34, 3111–3119. [CrossRef] 10. Sakane, S.; Matsumura, Y.; Yamamura, Y.; Ishida, Y.; Maruoka, K.; Yamamoto, H. Olefinic cyclizations promoted by beckmann rearrangement of oxime sulfonate. J. Am. Chem. Soc. 1983, 105, 672–674. [CrossRef] 11. Sakanishi, K.; Mochida, I.; Okazaki, H.; Soeda, M. Selective hydrogenation of 9-aminoacridine over supported noble metal catalysts. Chem. Lett. 1990, 19, 319–322. 12. Kouznetsov, V.; Palma, A.; Rozo, W.; Stashenko, E.; Bahsas, A.; Amaro-Luis, J. A facile Brønsted acidic-mediated cyclisation of 2-allyl-1-arylaminocyclohexanes to octahydroacridine derivatives. Tetrahedron Lett. 2000, 41, 6985–6988. [CrossRef] 13. Jacob, R.G.; Perin, G.; Botteselle, G.V.; Lenardão, E.J. Clean and atom-economic synthesis of octahydroacridines: Application to essential oil of citronella. Tetrahedron Lett. 2003, 44, 6809–6812. [CrossRef] 14. Sabitha, G.; Reddy, E.V.; Yadav, J.S. Bismuth(III) Chloride: An efficient catalyst for the one-pot stereoselective synthesis of octahydroacridines. Synthesis 2002, 3, 409–412. [CrossRef] 15. Santoro, F.; Psaro, R.; Ravasio, N.; Zaccheria, F. Reductive amination of ketones or amination of alcohols over heterogeneous Cu catalysts: Matching the catalyst support with the N-Alkylating agent. ChemCatChem 2012, 4, 1249–1254. [CrossRef] 16. Zaccheria, F.; Scotti, N.; Marelli, M.; Psaro, R.; Ravasio, N. Unravelling the properties of supported copper oxide: Can the particle size induce acidic behaviour? Dalton Trans. 2013, 42, 1319–1328. [CrossRef][PubMed]

17. Zaccheria, F.; Santoro, F.; Psaro, R.; Ravasio, N. CuO/SiO2: A simple and effective solid acid catalyst for epoxides ring opening. Green Chem. 2011, 13, 545–548. [CrossRef] 18. Scotti, N.; Dangate, M.; Gervasini, A.; Evangelisti, C.; Ravasio, N.; Zaccheria, F. Unraveling the role of low coordination sites in a Cu metal nanoparticle: A step forwards the selective synthesis of second generation biofuels. ACS Catal. 2014, 4, 2818–2826. [CrossRef] 19. Campanati, M.; Savini, P.; Tagliani, A.; Vaccari, A.; Piccolo, O. Environmentally friendly vapour phase synthesis of alkylquinolines. Catal. Lett. 1997, 47, 247–250. [CrossRef] 20. Ravasio, N.; Antenori, M.; Babudri, F.; Gargano, M. Intramolecular ene reaction promoted by mixed cogels. Stud. Surf. Sci. Catal. 1997, 108, 625–632. 21. Trombetta, M.; Busca, G.; Rossini, S.; Piccoli, V.; Cornaro, U.; Guercio, A.; Catani, R.; Willey, R. FT-IR studies on light oleﬁn skeletal isomerization catalysis: III. surface acidity and activity of amorphous and crystalline

catalysts belonging to the SiO2–Al2O3 system. J. Catal. 1998, 179, 581–596. [CrossRef] 22. Ravindra Reddy, C.; Bhat, Y.S.; Nagendrappa, G.; Jai Prakash, B.S. Brønsted and Lewis acidity of modiﬁed montmorillonite clay catalysts determined by FT-IR spectroscopy. Catal. Today 2009, 141, 157–160. [CrossRef] 23. Flessner, U.; Jones, D.J.; Rozière, J.; Zajac, J.; Storaro, L.; Lenarda, M.; Pavanc, M.; Jiménez-López, A.; Rodrìguez-Castellón, E.; Trombetta, M.; et al. A study of the surface acidity of acid-treated montmorillonite clay catalysts. J. Mol. Catal. A Chem. 2001, 168, 247–256. [CrossRef] 24. Vodnár, J.; Farkas, J.; Békássy, S. Catalytic decomposition of 1,4-diisopropylbenzene dihydroperoxide on montmorillonite-type catalysts. Appl. Catal. A 2001, 208, 329–334. [CrossRef] Catalysts 2018, 8, 410 10 of 10

25. Laschat, S.; Lauterwein, J. Intramolecular hetero-Diels-Alder reaction of N-arylimines. Applications to the synthesis of octahydroacridine derivatives. J. Org. Chem. 1993, 58, 2856–2861. [CrossRef] 26. Acelas, M.; Romero Bohórquez, A.R.; Kouznetsov, V.V. Highly diastereoselective synthesis of new trans-fused octahydroacridines via. intramolecular cationic imino diels–alder reaction of N-protected anilines and citronellal or citronella essential oil. Synthesis 2017, 49, 2153–2162. 27. Emeis, C.A. Determination of integrated molar extinction coefﬁcients for infrared absorption bands of pyridine adsorbed on solid acid catalysts. J. Catal. 1993, 141, 347–354. [CrossRef] 28. Carey, J.S.; Laffan, D.; Thomson, C.; Williams, M.T. Analysis of the reactions used for the preparation of drug candidate molecules. Org. Biomol. Chem. 2006, 4, 2337–2347. [CrossRef][PubMed]

© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).