Palladium-catalyzed hydroamination of farnesene—Long chain amines as building blocks for surfactants based on a renewable feedstock

Accepted Manuscript Title: Palladium-Catalyzed Hydroamination of Farnesene – Long Chain Amines as Building Blocks for Surfactants Based on a Renewable Feedstock Authors: Thiemo A. Faßbach, Nadine G¨osser, Fridolin O. Sommer, Arno Behr, Xiaoqiang Guo, Steffen Romanski, Dirk Leinweber, Andreas J. Vorholt PII: DOI: Reference:

S0926-860X(17)30263-6 http://dx.doi.org/doi:10.1016/j.apcata.2017.06.014 APCATA 16276

To appear in:

Applied Catalysis A: General

Received date: Revised date: Accepted date:

20-4-2017 12-6-2017 13-6-2017

Please cite this article as: Thiemo A.Faßbach, Nadine G¨osser, Fridolin O.Sommer, Arno Behr, Xiaoqiang Guo, Steffen Romanski, Dirk Leinweber, Andreas J.Vorholt, Palladium-Catalyzed Hydroamination of Farnesene – Long Chain Amines as Building Blocks for Surfactants Based on a Renewable Feedstock, Applied Catalysis A, Generalhttp://dx.doi.org/10.1016/j.apcata.2017.06.014 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Palladium-Catalyzed Hydroamination of Farnesene – Long Chain Amines as Building Blocks for Surfactants Based on a Renewable Feedstock Thiemo A. Faßbach[a], Nadine Gösser[a], Fridolin O. Sommer[a], Arno Behr[a], Xiaoqiang Guo[b], Steffen Romanski[b], Dirk Leinweber[b], Andreas J. Vorholt[a]* [a]

Fakultät für Bio- und Chemieingenieurwesen, Lehrstuhl für Technische Chemie, Technische Universität Dortmund, Emil-Figge-Straße 66, 44227 Dortmund (Germany)

[b]

Clariant Produkte (Deutschland) GmbH, D-65929 Frankfurt am Main, Germany

Graphical Abstract

Highlights   

First hydroamination of -farnesene using homogeneous catalysts Wide substrate scope for amines, compatible with this reaction Synthesis of surfactants, based on farnesylamines and evaluation of their surface activity

Abstract: Long chain amines are of great importance for industrial chemistry as they are precursors for surfactants like amine oxides or quaternary ammonia compounds. The atom efficient, homogeneously catalyzed hydroamination using 1,3-dienes offers linear linkage of the amine group to renewables like -farnesene, offering a C15 skeletal structure, which is a desired size for surfactants, the so called laurics. The presented paper describes the development of a catalytic system for the hydroamination of the industrially available terpene -farnesene in good to excellent yields. The reaction works with a broad range of amines, aliphatic and aromatic ones. Furthermore, functionalities, like alcohol or ether groups, are tolerated, yielding functionalized farnesylamines. With two model nucleophiles, a scale-up to a 5,000 mL reactor was accomplished; the obtained products were functionalized to surfactants and afterwards characterized by their surface activity. Keywords: Hydroamination; Palladium; Homogeneous Catalysis; Renewables; Terpenes; Surfactants

1. Introduction Amines are widely used within the chemical industry.[1] Besides their use as a building block for polymers (e.g. polyamides) or fine chemicals, like pharmaceuticals, they play an important role as

precursors for surfactants. Typical nitrogen based surfactants are amine oxides and quaternary ammonium compounds. They usually are composed of long chain moieties and small substituents, sometimes bearing functional groups at the nitrogen atom, which is positively charged.[2] There are many routes to synthesize amines, e.g. the hydrogenation of nitriles or condensation reactions with ammonia. On a laboratory scale, there are also aminations using aryl halides, e.g. the BuchwaldHartwig amination.[3,4] In order to prevent waste formation and offer compact syntheses to obtain building blocks for surfactants, new synthetic pathways are of great interest. Especially the valorization of renewable feedstocks plays an important role, although predominantly triglycerides and their derivatives are exploited. Terpenes are not yet utilized for surfactant purposes to such an extent. Recently, we could show that the hydroaminomethylation of terpenes is a valuable tool for the synthesis of surfactants starting from terpenes.[5] However, the reaction leads to byproducts and needs high pressures. In this context the hydroamination is a promising tool to produce long chain amines that can be further functionalized to surfactants. The term hydroamination of alkenes in general refers to the addition of a non-tertiary amine to a C/C double bond (Scheme 1).

Scheme 1. General hydroamination reaction of alkenes.

Enormous effort has been put into this topic leading to a high number of publications regarding the linear hydroamination of olefins.[6–10] Despite this fact, many reactions have to be carried out with specific activated nitrogen compounds, such as hydrazones[11,12] or aromatic amines[13,14]. Another way of enabling a linear linkage of amines to unsaturated hydrocarbons is the use of 1,3dienes. They show a special reactivity and there are several examples for 1,3-dienes being successfully hydroaminated in a linear way, especially with palladium or nickel catalysts.[15,16] In 2010, Behr et al. showed that palladium complexes incorporating diphosphine ligands with a natural bite angle of roughly 100° are capable of catalyzing the hydroamination of myrcene with morpholine.[17–19] The hydroamination of 1,3-dienes with palladium catalysts, for example 1,3pentadiene[20] or isoprene[21,22], is the topic of several publications over the last years. The use of -myrcene yields amines with a C10-chain. Since the chain length of eight (and two additional methyl branches) is not sufficient for the most nitrogen based surfactants, there is a demand for the use of higher terpenes that are also industrially available. In 2008 Amyris Inc. developed an industrial process to produce -farnesene by yeast fermentation of sugarcane.[23] It is the next higher terpene analog to -myrcene, thus, being five carbon atoms larger, offers a chain length which is within the range of so called laurics (Figure 1).[24] The depicted product can be considered an amine bearing an unsaturated C12-chain with three methyl branches. Structurally it is very close to industrially manufactured long chain alkyl amines, but is based on a renewable feedstock. Hydroaminations of 1,3-dienes can be achieved by high amounts of bases, e.g. in the Takasago process for producing (-)-menthol starting from -myrcene using nbutyllithium.[25] This approach has been shown to work with -farnesene as well.[26,27] These methods suffer from the need for completely inert conditions and highly corrosion resistant materials. The goal we aimed for in this work was to find a catalytic system that enables a tolerant and versatile synthesis of farnesylamines with homogeneous transition metal catalysts under mild conditions. The established catalytic system was firstly evaluated with respect to different employable amines, functionalized and non-functionalized ones. We then did a scale-up to a 5-L stirred tank reactor for two nucleophiles, synthesizing some hundred grams of the desired products. The obtained products were further functionalized and their surface activity examined.

2. Experimental 2.1. Hydroamination experiments Chemicals were purchased from Acros Organics (Geel, Belgium), Sigma-Aldrich (Steinheim, Germany), ABCR (Karlsruhe, Germany) and TCI (Tokyo, Japan). All chemicals were employed without drying or any further purification. All compounds were directly weighed into the reaction vessel, which was afterwards closed and pressurized with argon (5 bar). The reactors were placed in a heated aluminum block and stirred magnetically. After 5 h the reaction was ended by cooling the reactors with an ice-bath.

2.2. Analytics Routine gas chromatographic analyses were performed on an Agilent 7890B instrument (Santa Clara, USA) equipped with a flame ionization detector (FID) and a HP-5 capillary column (30 m, diameter 0.32 mm, film thickness 0.25 µm) connected to an auto sampler (7693) and an injector (G4513A). GCMS analyses of the products were carried out on an Agilent 5977A MSD (70 eV). NMR spectra were recorded on Bruker DRX spectrometers. CDCl3 purchased from Deutero was used as solvent. High resolution mass spectra were recorded on a TSQ mass spectrometer from ThermoQuest coupled to an HPLC-System (HPLC column: Hypersyl GOLD, 50 mm × 1 mm, 1.9 μm) for HPLC-ESI-HRMS.

2.3. Surfactant property tests The synthesized surfactants were tested for their surface tension reduction properties, wetting abilities and foam abilities. The products were tested according to the methods described below.

2.3.1. CMC (critical micelle concentration) and surface tension measurements These measurements were conducted with the Krüss Tensiometer K 100 (Ring), which utilized the Du Noüy ring method. Surface tension profiles of the surfactants were measured at 25 °C.

2.3.2. Determination of wetting ability by immersion This measurement was modified from the European Standard EN 1772:2000. A 500 mL 0.1% w/w aqueous solution of N,N-dimethylfarnesylamine oxide was prepared. The test solution was kept at room temperature for 1 h. A raw cotton disc (30 mm diameter, wfk Testgewebe GmbH), was clamped in a gripper (European Standard EN 1772:2000) and immersed in the solution. Stopwatch was started at the moment when the lower part of the disc touches the solution and stopped when the disc began to sink of its own accord. The arithmetic mean of five measurements was calculated and recorded as the wetting time. The wetting ability of a surfactant is defined in six categories: 1) “very high”, when the wetting time is lower than 25 s. 2) “high”, when the wetting time is in between 25 to 50 s. 3) “medium”, when the wetting time is in between 50 to 100 s. 4) “low”, when the wetting time is in between 100 to 200 s. 5) “very low”, when the wetting time is in between 200 to 300 s. 6) “no wetting”, when the wetting time is longer than 300 s.

2.3.3. Measurement of the dynamic surface tension These measurements were conducted with the Krüss PocketDyne BP2100, which utilized the maximum bubble pressure method of surface tension analysis. Surface tension profiles of the

synthesized surfactants at the desired concentration were measured in deionized water at 25 °C. The test samples were prepared in 0.3, 1.0 and 3.0 g/L aqueous solution. The dynamic surface tension was measured at room temperature (23 °C) applying surfaces with an age from 20 to 400 ms.

2.3.4. Measurement of foam creation and decay on a SITA foam tester This method monitors foam generation and decay at room temperature over time. A surfactant solution of 0.01 g/L was pumped into the SITA foam tester R2000. The speed of the stirring plate was set to 1200 rpm. Foam creation was recorded in a 10 second interval for 5 minutes and decay was recorded in a 30 second interval for 15 minutes at room temperature.

3. Results and Discussion 3.1. Precursor In order to establish reaction parameters for the hydroamination of -farnesene, a model reaction system is needed. We decided to use diethylamine as the nucleophile for further investigations, as it is a small, sterically not demanding and non-functionalized amine, making it a good model reagent for relevant amines. The desired hydroamination yields four different isomers, depending on where the nucleophilic attack takes place (Scheme 2).

Scheme 2. Products of the hydroamination of -farnesene with diethylamine

The 1,4-products depicted in Scheme 2 were summarized and are given in yield (Yhydroamination) as moles of products per employed mole of -farnesene, 1,2-products were not observed. It should also be mentioned that all 4 isomers were observed in equal amounts throughout these investigations. Selectivity (Shydroamination) was calculated as moles of products per converted mole of -farnesene (Xfarnesene). Starting from the already known catalytic system for the palladium catalyzed hydroamination of -myrcene, which is chemically and structurally similar to -farnesene, we started our investigations, looking for proper precursors and ligands (Table 1).

Reaction conditions: 0.4 mol% Pd based on -farnesene; ligand = 1,4-bis(diphenylphosphino)butane, metal:ligand = 1:4; n(diethylamine) = 4 mmol; n( -farnesene) = 4 mmol; solvent = DMF (5 mL); T = 100 °C, t = 5 h, 500 rpm, p(Ar) = 5 bar. X and Y in % based on -farnesene. Results determined by GC-FID using dodecane as internal standard. tfa = trifluoroacetate; acac = acetylacetonate; hfacac = 1,1,1,5,5,5-hexafluoroacetylacetonate; dba = dibenzylideneacetone.

According to previously published work[19], precursors with fluorinated ligands (entry 1.1 and 1.3) are much more active than their not-fluorinated analog (entry 1.2 and 1.4). The reason for this might be the tendency to form a Pd-H complex, which is postulated to be necessary for the hydroamination.[28] As a side reaction the telomerization of -farnesene takes place (Scheme 3).

Scheme 3. Telomerization of -farnesene with diethylamine

This is surprising because the telomerization was not yet reported with a 1,3-diene larger than -myrcene.[29]

3.2. Ligands Several publications suggests that the bite angle of diphosphine ligands should be around 100° to favor the hydroamination.[13,17] Using the same conditions as before, we tested several popular mono- and diphosphine ligand, maintaining the same metal:ligand-ratio (Table 2). Reaction conditions: 0.4 mol% Pd(tfa)2 based on -farnesene; metal:ligand = 1:4; n(diethylamine) = 4 mmol; n(-farnesene) = 4 mmol; solvent = DMF (5 mL); T = 100 °C, t = 5 h, 500 rpm, p(Ar) = 5 bar. X and Y in % based on -farnesene. Results determined by GC-FID using dodecane as internal standard. [a]: 0.8 mol% Pd(tfa)2 based on -farnesene; [b]: n(diethylamine) = 8 mmol DPPE = 1,2-bis(diphenylphosphino)ethane; DPPB = 1,4-bis(diphenylphosphino)butane; TPP = triphenylphosphine; DPEPhos = (oxybis(2,1-phenylene))bis(diphenylphosphane)

Obviously, the before mentioned requirement of ligands with a natural bite angle around 100° is present in this conversion as well (Table 2, entry 2.2 and 2.4). It is known, that the concentration of the palladium catalyst with respect to the 1,3-diene is of great importance to the selectivity.[17] High catalyst concentrations favor the hydroamination, while lower catalyst concentrations shift the reactivity towards the telomerization. Therefore the most active system (Pd/DPEPhos) from the screening before was tested with a twofold amount of catalyst (Entry 2.5). Furthermore, the amount of employed amine was raised, to increase the conversion and favor the formation of the desired hydroamination products (Entry 2.6). Higher catalyst concentrations than 0.8 mol% led to the formation of palladium black and strongly decreased conversions. Considering the high conversions reached with DPPB and DPEPhos, we reconsidered the metal:ligand-ratio with the higher catalyst concentration in order to accomplish optimal reaction conditions leading to high conversions of -farnesene and high selectivities towards the desired hydroamination products (Table 3). Reaction conditions: 0.8 mol% Pd(tfa)2 based on -farnesene; n(diethylamine) = 8 mmol; n(-farnesene) = 4 mmol; solvent = DMF (5 mL); T = 100 °C, t = 5 h, 500 rpm, p(Ar) = 5 bar. X and Y in % based on -farnesene. Results determined by GC-FID using dodecane as internal standard. DPPB = 1,4-bis(diphenylphosphino)butane; DPEPhos = (oxybis(2,1-phenylene))bis(diphenylphosphane)

Higher metal:ligand-ratios lead to higher selectivities towards hydroamination (Entry 3.2 and 3.4). This can be explained with the high sterical demand of -farnesene to coordinate the palladium catalyst. A higher ligand concentration ensures a rigid sterical environment, making the hydroamination the more likely reaction. In summary, the catalytic system consisting of 0.8 mol% of the Pd(tfa)2 precursor along with an eightfold molar excess of the DPEPhos ligand (metal:ligand = 1:4) leads to the best results with an excellent yield of 94% of the desired hydroamination products (Entry 3.4; for a yield/time-plot under these conditions see Supporting Information).

3.3. Solvents The use of DMF as solvent ensures a homogenous phase and the solubility of both the highly-nonpolar terpene and the polar amine. In terms of product purification DMF suffers from a high boiling point and its toxicity.[32] With these reaction conditions at hand, several other common solvents were tested for their suitability, the results are shown in Table 4. Reaction conditions: 0.8 mol% Pd(tfa)2 based on -farnesene; ligand = DPEPhos, metal:ligand = 1:8; n(diethylamine) = 8 mmol; n(-farnesene) = 4 mmol; V(solvent) = 5 mL; T = 100 °C, t = 5 h, 500 rpm, p(Ar) = 5 bar. X and Y in % based on -farnesene. Results determined by GC-FID using dodecane as internal standard. [a]: 0.4

mol% Pd(tfa)2 based on -farnesene, n(diethylamine) = 16 mmol; n(-farnesene) = 8 mmol; [b]: n(diethylamine) = 8 mmol

Polar protic solvents lower the reactivity (Table 4, Entry 4.2), alkyl ethers are also not favored in the hydroamination under these conditions (Entry 4.3). Surprisingly only anisole, which can be considered a recommended solvent regarding toxicity[32], leads to high selectivities and comparable results as with DMF (Entry 4.4). Interestingly, this reaction can also be carried out without using any solvent (Entry 4.5 and 4.6). This is most favorable, considering waste production and product separation, as well as space-time-yields. Due to the relatively low solubility of the catalyst in the pure substrates, only half of the catalyst concentration was applied. This does not lead to a severe decrease of selectivity, but product yields are significantly lower. The twofold excess of the amine does not increase conversion in this case (Entry 4.5), as it does in DMF. With a stoichiometric substrate ratio in return the reaction yields slightly more products (4.6). This could be due to an excessive interaction of the amine with the catalyst metal resulting in an inhibition of the reaction.[33,34]

3.4. Substrate Scope The next step to develop a broad approach was to determine the versatility of this catalytic system in order to produce different farnesyl amine derivatives (farnesylamines). At first we employed different alkyl substituted secondary amines, to check for the compatibility of this system with bulkier nucleophiles (Table 5). Reaction conditions: 0.8 mol% Pd(tfa)2 based on -farnesene; ligand = DPEPhos, metal:ligand = 1:8; n(amine) = 8 mmol; n(-farnesene) = 4 mmol; V(DMF) = 5 mL; T = 100 °C, t = 5 h, 500 rpm, p(Ar) = 5 bar. X and Y in % based on -farnesene. Results determined by GC-FID using dodecane as internal standard. [a]: Dimethylamine was employed as dimethylammonium dimethylcarbamate [b]: 48% dialkylated amine was yielded

Obviously sterical demand has a great influence on the reaction yields. Secondary amines with substituted alkyl chains, like di-iso-propylamine or di-cyclo-hexylamine, yield only traces of the desired products. Linear alkyl chains are compatible, as shown with di-n-butylamine (Yhydroamination = 81%), longer chains on the other hand lead to decreasing yields. However, the cyclic secondary amine piperidine and dimethylamine could be converted to the desired hydroamination product in nearly quantitative yields. This underlines the need for nucleophiles with low sterical demands for a successful reaction. Regarding the established processes for hydroamination, it is of great interest to enable the conversion of functionalized amines to farnesylamines. One example for a hydroamination that cannot be conducted by the base catalyzed hydroamination of 1,3-dienes would be the hydroamination of -farnesene with diethanolamine. The latter is an interesting building block in order to synthesize promising N-oxides. Furthermore, the hydroxyl groups could be ethoxylated, yielding non-ionic surfactants. It can be seen that the hydroxyl groups of the diethanolamine cause a decrease in conversion. This is consistent with the results using protic solvents like methanol. With morpholine on the other hand, the catalytic system shows nearly quantitative yields, proofing that the oxygen atom is tolerated in general. The aromatic, primary amines are very active, but only benzylamine undergoes a second substitution with a farnesyl moiety. This is most probably due to the steric hindrance of the mono alkylated aniline. This behavior aligns with di-iso-propylamine and di-cyclo-hexylamine, being nucleophiles with substituents which are branched in -position. Interestingly it is rarely reported that a catalytic system for the hydroamination of 1,3-dienes works well with both, aryl and alkyl amines[6], as it does in the presented case.

3.5. Scale-Up Once the catalysis to obtain surfactant precursors was established, the reaction needed to be scaled up in order to synthesize sufficient amounts of products for further functionalization (Table 6). Reaction conditions: 0.8 mol% Pd(tfa)2 based on -farnesene; ligand = DPEPhos, metal:ligand = 1:8; c(diethylamine) = 0.1 g mL-1 DMF; c(-farnesene) = 0.15 g mL-1 DMF; T = 100 °C, t = 5 h, 500 rpm, p(Ar) = 5 bar. Y in % based on -farnesene. Results determined by GC-FID using dodecane as internal standard. [a]: Nucleophile = dimethylamine, employed as dimethylammonium dimethylcarbamate

The first step was to transfer the reaction of -farnesene with diethylamine into a 300 mL Parr® reactor. In this stage it was discovered that the reactions would only maintain high conversions, if the gas:liquid ratio used in the smaller (25 mL, Entry 6.1) reactors was kept the same (Entry 6.3). If a gas:liquid ratio of 3.2 is underrun, the conversion decreases significantly (Entry 6.4). Surprisingly the same applies to higher ratios (Entry 6.2). When transferring the reaction to the 5,000 mL Büchi® reactor, another issue emerged. The heating phase of the double jacketed reactor with thermal oil is significantly prolonged (up to 1.5 h), compared to the single-walled reactors placed in a heated block, which are almost immediately at reaction temperature. Due to the prolonged heating time the reaction mixture is at elevated temperature, not yet at reaction temperature. In this phase, side reactions like isomerization and dimerization of -farnesene take place reducing the selectivity of the overall reaction. By increasing the heating rate and avoiding hot-spots, since the catalyst decomposes quickly at temperatures above 100 °C, the overall yield of 77% was achieved in the 5-L reactor (Entry 6.5). The reaction was carried out with dimethylamine under the same conditions and yields of over 90% (Entry 6.6) were reached. The obtained products were distilled (N,N-dimethylfarnesylamine: bp = 126 °C; 5 mbar) or purified by reactive extraction[36] with formic acid (N,N-diethylfarnesylamine).

3.6. Surfactants The purified products, dimethyl- and diethylfarnesylamine, were further converted to corresponding amine oxides, betaines and quaternary ammonium compounds (Scheme 4) to obtain the surfactants aimed for. The syntheses of these surfactant molecules were rather straightforward.

Scheme 4. Synthetic route from N,N-dimethyl- and -diethyl-farnesylamine to their corresponding amine oxides, betaines and quats surfactant molecules. Note that only the tail-product is displayed, although a mixture of all 1,4-products have been employed. The betaines were prepared from dimethylfarnesylamine only.

Both dimethyl- and diethylfarnesylamines were reacted with dimethyl sulfate at 80 °C to acquire quaternary ammonium compounds and reacted with H2O2 under the CO2 atmosphere to obtain amine oxides. These reactions gave good conversion and purity of products, which were then tested for surfactant properties without further purification. It must be mentioned that the diethyfarnesylamine oxide, unlike the dimethylfarnesylamine oxide as a clear aqueous solution, was obtained as a very unstable emulsion that underwent fast phase separation. With these model surfactants, several property tests were conducted. The test results of CMC, surface tension and wetting time are listed in Table 7. * C20: Concentration of surfactant to reduce the surface tension of water by 20 mNm -1

In general, all synthetic surfactants, except diethylfarnesylamine oxide (not listed in the table), are fairly efficient in surface tension reduction. Only dimethylfarnesylamine oxide showed high wetting ability toward cotton, the others showed rather low or no wetting ability at all. Due to the poor solubility of diethylfarnesylamine oxide in water, no correct CMC or surface tension could be measured. In the SITA foam test, dimethylfarnesylamine betaine was able to create 120 mL of foam after 5 min and all foam collapsed within 6 min. No foam was able to be created from all the other surfactants after 5 min of stirring. In general, these surfactants are none to low foaming surfactants

with fast collapsing foam. Dimethylfarnesylamine oxide can be a good wetting agent for some relevant applications, e.g. in laundry detergents. The results from dynamic surface tension measurement are concluded in Figure 2, results for other surfactants can be found in the supporting information. It is known that the surface tensions measured on newly formed surfaces of some solutions were higher than the equilibrium values. For dilute aqueous solutions there are so few surfactant molecules present initially in the surface that one would expect the surface tension to be nearly that of water, and even for more concentrated solutions the number of solute molecules in the surface initially would be small. However, as surfactant molecules are adsorbed in the surface, the surface tension decreases and finally reaches the equilibrium value. From the results, dimethylfarnesylamine oxide reduced surface tension most efficiently of all tested surfactants over the period of time measured. It indicates that dimethylfarnesylamine oxide could be potentially useful in crop protection application. When surfactants are used in crop protection applications, usually 0.3 - 3 g/L of surfactant is present, and the average time for surfactant solution spraying from nozzle to crop leaves is between 20 to 400 ms. The faster the surface tension of the water drops is reduced, the better those drops can attach and stick to the crop leaves. This results in the longer contacting time for active ingredients to penetrate through leaves and finally reach the target.

4. Conclusions In the presented work, a catalytic system, consisting of a fluorinated palladium precursor, Pd(tfa)2, and an excess of the diphosphine DPEPhos, was developed for the firstly shown transition metal catalyzed hydroamination of C15-terpene -farnesene. The employed amine can bear functionalized moieties (e.g. hydroxyl or ether groups) and there is no need for any co-catalyst, like acids or bases. Small substituents with only low sterical demand are clearly favored, leading to almost quantitative yields with dimethylamine or piperidine and 94% yield with diethylamine. Functional groups like ethers or alcohols are tolerated as well as aromatic amines. While the reaction was developed in the solvent DMF, investigations showed that anisole is a well-working non-toxic alternative and in the case of diethylamine the reaction can even be carried out without using any solvent. This reaction opens up a whole new range of potential surfactant building blocks, using the renewable -farnesene as the hydrophobic moiety. The reaction was scaled up to a 5,000 mL reactor with the two model substrates dimethylamine and diethylamine. The products were synthesized in a several hundred gram scale and the obtained surfactant precursors were then further functionalized to betaines, quaternary ammonium compounds and N-oxides and finally characterized by their surface activity. These surfactants show very efficient surface tension reduction behavior and produce low to none foam. Especially for dimethylfarnesylamine oxide, beside these described properties, it also shows good wetting ability toward cotton and much more efficient dynamic surface tension reduction properties than the other molecules. Thus dimethylfarnesylamine oxide can be furthermore used as non-foaming wetting agent, for instance in crop protection applications, hard surface applications or laundry applications.

Acknowledgements The financial support from Clariant AG is gratefully acknowledged. We thank Umicore for the donation of palladium precursors.

References [1]

P. Roose, K. Eller, E. Henkes, R. Rossbacher, H. Höke, Amines, Aliphatic in:, Ullmann’s Encyclopedia of Industrial Chemistry, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, Germany, 2015, pp. 1–55.

[2]

K. Kosswig, Surfactants in:, Ullmann’s Encyclopedia of Industrial Chemistry, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, Germany, 2000, pp. 431–505.

[3]

F. Paul, J. Patt, J.F. Hartwig, J. Am. Chem. Soc. 116 (1994) 5969–5970.

[4]

P. Ruiz-Castillo, S.L. Buchwald, Chem. Rev. 116 (2016) 12564–12649.

[5]

T.A. Faßbach, T. Gaide, M. Terhorst, A. Behr, A.J. Vorholt, ChemCatChem 9 (2017) 1359–1362.

[6]

L. Huang, M. Arndt, K. Gooßen, H. Heydt, L.J. Gooßen, Chem. Rev. 115 (2015) 2596–2697.

[7]

M. Beller, J. Seayad, A. Tillack, H. Jiao, Angew. Chem., Int. Ed. 43 (2004) 3368–3398.

[8]

E. Bernoud, C. Lepori, M. Mellah, E. Schulz, J. Hannedouche, Catal. Sci. Technol. 5 (2015) 2017–2037.

[9]

T.E. Müller, K.C. Hultzsch, M. Yus, F. Foubelo, M. Tada, Chem. Rev. 108 (2008) 3795–3892.

[10]

C. Lepori, J. Hannedouche, Synthesis 49 (2017) 1158–1167.

[11]

A.M. Johns, Z. Liu, J.F. Hartwig, Angew. Chem., Int. Ed. 46 (2007) 7259–7261.

[12]

A. Perrier, M. Ferreira, J.N.H. Reek, J.I. van der Vlugt, Catal. Sci. Technol. 3 (2013) 1375.

[13]

D. Banerjee, K. Junge, M. Beller, Org. Chem. Front. 1 (2014) 368.

[14]

C.S. Yi, S.Y. Yun, Org. Lett. 7 (2005) 2181–2183.

[15]

J. Pawlas, Y. Nakao, M. Kawatsura, J.F. Hartwig, J. Am. Chem. Soc. 124 (2002) 3669–3679.

[16]

O. Löber, M. Kawatsura, J.F. Hartwig, J. Am. Chem. Soc. 123 (2001) 4366–4367.

[17]

A. Behr, L. Johnen, N. Rentmeister, Adv. Synth. Catal. 352 (2010) 2062–2072.

[18]

A. Behr, T. Färber, Chem. Eng. Trans. 43 (2015) 835–840.

[19]

T. Färber, O. Riechert, T. Zeiner, G. Sadowski, A. Behr, A.J. Vorholt, Chem. Eng. Res. Des. 112 (2016) 263–273.

[20]

P. Neubert, I. Meier, T. Gaide, A. Behr, Synthesis 48 (2016) 2287–2293.

[21]

G. Kuchenbeiser, A.R. Shaffer, N.C. Zingales, J.F. Beck, J.A.R. Schmidt, J. Organomet. Chem. 696 (2011) 179–187.

[22]

B. Tamaddoni Jahromi, A. Nemati Kharat, S. Zamanian, A. Bakhoda, K. Mashayekh, S. Khazaeli, Appl. Catal. A Gen. 433–434 (2012) 188–196.

[23]

N.S. Renninger, D.J. Mephee, Fuel Compositions Comprising Farnesane and Farnesene Derivates and Method of Making and Using Same, US 7,399,323 B2, 2008.

[24]

M.A.R. Meier, J.O. Metzger, U.S. Schubert, Chem. Soc. Rev. 36 (2007) 1788.

[25]

K. Takabe, T. Katagiri, J. Tanaka, T. Fujita, S. Watanabe, K. Suga, Org. Synth. 67 (1989) 44.

[26]

C. Chapuis, Helv. Chim. Acta 97 (2014) 197–214.

[27]

H. Ujihara, M. Fujiwara, Method for Producing Optically Active 2,3-Dihydrofarnesal, EP 2 894 143 A1, 2015.

[28]

A.M. Johns, M. Utsunomiya, C.D. Incarvito, J.F. Hartwig, J. Am. Chem. Soc. 128 (2006) 1828– 1839.

[29]

A. Behr, L. Johnen, A.J. Vorholt, ChemCatChem 2 (2010) 1271–1277.

[30]

Z. Freixa, P.W.N.M. van Leeuwen, Dalt. Trans. 77 (2003) 1890–1901.

[31]

P. Dierkes, P.W.N.M. van Leeuwen, J. Chem. Soc. Dalt. Trans. (1999) 1519–1530.

[32]

D. Prat, J. Hayler, A. Wells, Green Chem. 16 (2014) 4546–4551.

[33]

E.R. Strieter, D.G. Blackmond, S.L. Buchwald, J. Am. Chem. Soc. 125 (2003) 13978–13980.

[34]

B. Tao, D.W. Boykin, Tetrahedron Lett. 44 (2003) 7993–7996.

[35]

A. Behr, G. Henze, L. Johnen, S. Reyer, J. Mol. Catal. A Chem. 287 (2008) 95–101.

[36]

R. Schulz, R. van den Bongard, J. Islam, T. Zeiner, Ind. Eng. Chem. Res. 55 (2016) 5763–5769.

Figure 1. Classic laurics-based surfactants and the desired fanesylamine-derived structure

surface tension [mN/m]

75 70

0.3 g/L

65

1.0 g/L 3.0 g/L

60 55 50 45 40 35 30 0

50

100

150

200 time [ms]

Figure 2. Dynamic surface tension of dimethylfarnesylamine oxide

250

300

350

400

Table 1. Results using different palladium precursors Entry

Precursor

Xfarnesene

Yhydroamination

Ytelomerization

1.1

Pd(tfa)2

51

24

27

1.2

Pd(acac)2

13

6

7

1.3

Pd(hfacac)2

41

20

19

1.4

[Pd2dba3]

7

1

6

Reaction conditions: 0.4 mol% Pd based on -farnesene; ligand = 1,4-bis(diphenylphosphino)butane, metal:ligand = 1:4; n(diethylamine) = 4 mmol; n( -farnesene) = 4 mmol; solvent = DMF (5 mL); T = 100 °C, t = 5 h, 500 rpm, p(Ar) = 5 bar. X and Y in % based on -farnesene. Results determined by GC-FID using dodecane as internal standard. tfa = trifluoroacetate; acac = acetylacetonate; hfacac = 1,1,1,5,5,5-hexafluoroacetylacetonate; dba = dibenzylideneacetone.

Table 2. Results using different phosphine ligands Entry

Ligand

Bite Angle [°][30,31]

Xfarnesene

Yhydroamination

Ytelomerization

2.1

DPPE

78.1

1

<1

1

2.2

DPPB

98.6

51

24

27

2.3

TPP

-/-

3

<1

3

2.4

DPEPhos

102.7

52

4

48

2.5

DPEPhos[a]

102.7

67

66

1

2.6

DPEPhos[a],[b]

102.7

87

85

1

Reaction conditions: 0.4 mol% Pd(tfa)2 based on -farnesene; metal:ligand = 1:4; n(diethylamine) = 4 mmol; n(-farnesene) = 4 mmol; solvent = DMF (5 mL); T = 100 °C, t = 5 h, 500 rpm, p(Ar) = 5 bar. X and Y in % based on -farnesene. Results determined by GC-FID using dodecane as internal standard. [a]: 0.8 mol% Pd(tfa)2 based on -farnesene; [b]: n(diethylamine) = 8 mmol DPPE = 1,2-bis(diphenylphosphino)ethane; DPPB = 1,4-bis(diphenylphosphino)butane; TPP = triphenylphosphine; DPEPhos = (oxybis(2,1-phenylene))bis(diphenylphosphane)

Table 3. Results using different diphosphine ligands and metal:ligand-ratios Entry

Ligand

3.1

metal:ligand

Xfarnesene

Yhydroamination

Shydroamination

1:4

7

2

29

1:8

70

69

99

1:4

11

2

18

1:8

95

94

99

DPPB 3.2 3.3 DPEPhos 3.4

Reaction conditions: 0.8 mol% Pd(tfa)2 based on -farnesene; n(diethylamine) = 8 mmol; n(-farnesene) = 4 mmol; solvent = DMF (5 mL); T = 100 °C, t = 5 h, 500 rpm, p(Ar) = 5 bar. X and Y in % based on -farnesene. Results determined by GC-FID using dodecane as internal standard. DPPB = 1,4-bis(diphenylphosphino)butane; DPEPhos = (oxybis(2,1-phenylene))bis(diphenylphosphane)

Table 4. Results using different solvents Entry

Solvent

Xfarnesene

Yhydroamination

Shydroamination

4.1

acetonitrile

98

33

34

4.2

methanol

38

24

63

4.3

tetrahydrofuran

28

15

54

4.4

anisole

74

73

99

4.5

-/-[a]

57

51

89

4.6

-/-[a],[b]

65

60

92

Reaction conditions: 0.8 mol% Pd(tfa)2 based on -farnesene; ligand = DPEPhos, metal:ligand = 1:8; n(diethylamine) = 8 mmol; n(-farnesene) = 4 mmol; V(solvent) = 5 mL; T = 100 °C, t = 5 h, 500 rpm, p(Ar) = 5 bar. X and Y in % based on -farnesene. Results determined by GC-FID using dodecane as internal standard. [a]: 0.4 mol% Pd(tfa)2 based on -farnesene, n(diethylamine) = 16 mmol; n(-farnesene) = 8 mmol; [b]: n(diethylamine) = 8 mmol

Table 5. Results with different amines Amine

Xfarnesene

Yhydroamination

Shydroamination

>99

>99

>99

34

<1

<1

>99

81

81

32

<1

<1

>99

15

15

>99

>99

>99

>99

>99

>99

72

48

66

77

77

>99

99

16 (48)

65

Dimethylamine[a]

Di-iso-propylamine

Di-n-butylamine

Di-cyclo-hexylamine

Di-n-octylamine

Piperidine

Morpholine

Diethanolamine

Aniline

Benzylamine[b] Reaction conditions: 0.8 mol% Pd(tfa)2 based on -farnesene; ligand = DPEPhos, metal:ligand = 1:8; n(amine) = 8 mmol; n(-farnesene) = 4 mmol; V(DMF) = 5 mL; T = 100 °C, t = 5 h, 500 rpm, p(Ar) = 5 bar. X and Y in % based on -farnesene. Results determined by GC-FID using dodecane as internal standard.

[a]: Dimethylamine was employed as dimethylammonium dimethylcarbamate [b]: 48% dialkylated amine was yielded

Table 6. Scale-up of the reaction Entry

Reactor

DMF Volume [mL]

Yhydroamination

6.1

Multiplex[35] (25 mL)

5

95

6.2

Parr® (300 mL)

42.2

5

6.3

Parr® (300 mL)

72.5

95

6.4

Parr® (300 mL)

110

74

6.5

Büchi® (5,000 mL)

1,200

77

6.6

Büchi® (5,000 mL)[a]

1,200

92

Reaction conditions: 0.8 mol% Pd(tfa)2 based on -farnesene; ligand = DPEPhos, metal:ligand = 1:8; c(diethylamine) = 0.1 g mL-1 DMF; c(-farnesene) = 0.15 g mL-1 DMF; T = 100 °C, t = 5 h, 500 rpm, p(Ar) = 5 bar. Y in % based on -farnesene. Results determined by GC-FID using dodecane as internal standard.

[a]: Nucleophile = dimethylamine, employed as dimethylammonium dimethylcarbamate

Table 7. Surfactant properties and wetting ability Surfactant

CMC [gL-1]

C20* [gL-1]

Surface tension [mNm-1]

Wetting time [s]

Dimethylfarnesylamine oxide

1.09

0.018

28.7

28 (high)

Trimethylfarnesylammonium monomethylsulfate

3.5

0.31

24.1

> 300 (no wetting)

Diethylmethylfarnesylammonium monomethylsulfate

0.74

0.069

29.3

> 300 (no wetting)

Dimethylfarnesylamine betaine

1.56

0.043

32.9

145 (low)

* C20: Concentration of surfactant to reduce the surface tension of water by 20 mNm -1