Synthesis of terpene diamines based on camphor-derived dinitriles

Article abstract of DOI:10.1002/hlca.201400346

Diamines are useful as additional ligands in enantioselective hydrogenations. A new route for synthesizing such a rigid diamine from naturally available camphor via camphor oxime/camphor furoxan/deoxygenation in large scale was developed. The resulting dinitriles were subjected to hydrogenation with Raney Ni or other heterogenous catalysts. In most cases, the reduction resulted in complex mixtures. The best results were obtained with Raney Ni in THF/H₂O in strongly basic media (NaOH), which affords compound 5b only in good yield. Homogenous catalyst like Grubbs 2nd generation reduced only the less hindered CN group to yield an amino nitrile.

Full text of DOI:10.1002/hlca.201400346

Helvetica Chimica Acta – Vol. 98 (2015)
447
Synthesis of Terpene Diamines Based on Camphor-Derived Dinitriles
a
a b
by Florian R. Kinzl ) and Herbert M. Riepl* ) )
a) Weihenstephan-Triesdorf University of Applied Sciences, Schulgasse 16, DE-94315 Straubing
b
)
Fraunhofer-Group BioCat, Schulgasse 11a, DE-94315 Straubing
(
phone: þ 49-9421-187302; fax: þ 49-9421-187211; e-mail: herbert.riepl@hswt.de)
Diamines are useful as additional ligands in enantioselective hydrogenations. A new route for
synthesizing such a rigid diamine from naturally available camphor via camphor oxime/camphor furoxan/
deoxygenation in large scale was developed. The resulting dinitriles were subjected to hydrogenation
with Raney Ni or other heterogenous catalysts. In most cases, the reduction resulted in complex mixtures.
The best results were obtained with Raney Ni in THF/H O in strongly basic media (NaOH), which
2
affords compound 5b only in good yield. Homogenous catalyst like Grubbs 2nd generation reduced only
the less hindered CN group to yield an amino nitrile.
Introduction. – Among the terpenes with economic potential, pinene (1), terpinene
(2), and camphor (3) can be envisioned as potential starting materials for synthesis of
ligands with C-atom frameworks providing sterical shielding of reactivity. So, camphor-
based amines 4a and 4b had been prepared as ligands for transition metal complexes.
II
Jaramillo et al. synthesized chiral Pt complexes with these diamines, prepared from
camphoric acid [1]. Yang et al. used diamines, based on camphor, as intermediates for
the preparation of chiral Schiff-base ligands [2]. These ligands are sterically strongly
congested, and this results in low reaction rates at the respective transition-metal
centers. Nevertheless, various diamines have been used as additional chelating ligands
for enantioselective hydrogenation of more difficult substrates like tetralones [3]. To
end up in high enantioselectivity, it proved to be necessary for the diamines to have a
rigid backbone distant from the N-atoms involved in chelation. This might enable the
molecule to adopt ꢀa conformation dictated by the (...) phosphine ligandꢁ [4]. Such a
ꢂ 2015 Verlag Helvetica Chimica Acta AG, Zürich

4
48
Helvetica Chimica Acta – Vol. 98 (2015)
feature might be realized by using other backbone molecules where the N-atom is not
directly attached to the C-atom ring, e.g., our target molecules 5a/5b, where two
adjacent Me groups at C(2) and additionally one more Me group in a-position to C(2)
provides a very similar steric environment. Despite that, 5a might be able to chelate by
forming a seven-membered ring. With respect to enantioselectivity in catalysis, it will
be interesting to see if these diamines are advantageous with transition metals of the
second and third row. The bicyclic system of camphor can be cleaved by various
methods to yield a cyclopentane with a required C-atom framework. Oximes of
terpene-ketones often react anomalously to yield nitriles, which are suitable precursors
for amines [5]. Such a process is related to furoxans obtained from camphor dioxime.
We present here a new synthetic route of diamines 5a/5b in large scale from naturally
available camphor.
Results and Discussion. – Furoxans (¼1,2,5-oxadiazole 2-oxides) are a neglected
class of heterocycles, but very reactive. They are formed easily from dioximes by a fast
and clean oxidation, e.g., with NaClO. Dornow et al. tried to reduce furoxan 8 with
LiAlH and obtained an amine in poor yield (not characterized) [6]. Dicarbonitrile 9
4
could be obtained by other means and reduced with elemental Na to 5 [7]. We tried
both routes with no satisfactory results. It was advantageous then to have a suitable
method at hand for the deoxygenation of furoxan 8 with (EtO) P to 9 [8] (Scheme 1).
3
This could be reduced to the target product 5. Dicarbonitrile 9 was obtained usually as
pure stereoisomer 9a (cf. Scheme 2) by ring cleavage of 8. In need of large amounts of
nitrile, we reworked the process of exothermic phosphite deoxygenation, and found it
safe and advantageous to use trialkyl phosphates as a solvent. The furoxan is soluble to
the extent of 1.5 g furoxan per 10 ml of (EtO) PO. The final hydrogenation of the
3
nitrile proved to be a difficult process. In simple assets of hydrogenation procedures of
9
to 5, at least five products (stereoisomers not included) were found to be present,
often in equal concentrations. Several catalysts were tried (Table), including Raney Ni,
Rh on alumina [9], Pd on charcoal, Pt on charcoal, and Grubbs 2nd-generation catalyst
[
10]. The compositions of the reduced solutions are compiled in the Table. In the worst
case, compound 11 or 12, or both were the main product (Entries 9 and 10). One could
isolate 12 to reduce it in a second stage to 5, but formation of 11 was a dead end. This is
Scheme 1

Helvetica Chimica Acta – Vol. 98 (2015)
449
Table. Hydrogenation of the Dinitrile 9 under Different Reaction Conditions
Entry
Catalyst
Reaction conditions
Yields of products [%]
9
5
10
11
12
1
2
3
4
5
6
7
8
9
Raney Ni
Raney Ni
Raney Ni
Raney Ni
Raney Ni
Raney Ni
Raney Ni
Raney Ni
Raney Ni
Raney Ni
Raney Ni
Raney Ni
Grubbs II
EtOH, 6m NaOH (aq.),
34
4
29
6
0 bar, 1108, 20 h
EtOH, 6m NaOH (aq.),
5 bar, 808, 20 h
1,4-Dioxane, 6m NaOH (aq.),
0 bar, 1108, 20 h
1,4-Dioxane, 6m NaOH (aq.),
5 bar, 808, 20 h
1,4-Dioxane, 6m NaOH (aq.),
0 bar, 1058, 20 h
1,4-Dioxane, 6m NaOH (aq.),
5 bar, 1008, 20 h
THF, 6m NaOH (aq.),
0 bar, 708, 20 h
EtOH
20 bar, 908, 20 h
EtOH
0 bar, 708, 20 h
1,4-Dioxane
5 bar, 908, 20 h
1,4-Dioxane
0 bar, 1408, 20 h
1,4-Dioxane
0 bar, 1008, 20 h
Toluene
0 bar, 508, 72 h
86
21
24
38
61
79
67
9
6
5
4
7
20
5
28
6
5
8
4
6
3
7
1
31
19
37
7
1
1
1
1
0
1
2
3
15
44
42
20
10
20
17
18
6
6
6
61
8
due to the sterical hindrance which slows down the reaction rate of hydrogenation at
CN group at the tertiary C(1) (Scheme 2). The cyclic amine 11 may be formed through
deamination of the aldimine 13, which is an intermediate during the reduction of 9a.
The difficulty encountered here was to find conditions under which hydrogenation is
rapid enough to overcome the formation of secondary amine 11. This was partly
achieved by higher pressure to make H readily available at the heterogeneous reaction
2
site. Conversely, racemization at C(3) was crucial, since complete inversion yielded 10b

4
50
Helvetica Chimica Acta – Vol. 98 (2015)
Scheme 2
which is scarcely involved in ring closure. Due to the acidity of the C(3)-atom, one can
expect isomerization/racemization to occur in further manipulation with protic
solvents/basic reaction media involved. Already when pure dinitrile 9 was heated in
such solutions, an isomerisation at C(3) took place, and two stereoisomers 5a and 5b
were formed (Scheme 2). Isomer 9b was reduced to 10b and thus to 5b properly.
Reactions in EtOH/H O/NaOH gave good results, but the yields of diamine 5 varied
2
strongly (34 and 86%; Table, Entries 1 and 2). 1,4-Dioxane/H O/NaOH as reaction
2
media produced moderate yields of diamine 5 (21 – 61%; Table, Entries 3 – 6). The best
results were obtained with Raney Ni in THF/H O/NaOH, which furnished compound 5
2
in good yield (79% of 5 and only small amounts of by-products; Table, Entry 7).
Temperatures of 708 – 808 were optimal for the reaction in THF/H O/NaOH. H
2
2
Pressures of 60 bar and above were required. Small amounts of additional compound
1 were still present, but not in the same quantities as when other means of reduction
1
were employed. Hydrogenation in solvents, without additional NaOH/H O, such as
2
EtOH (Table, Entries 8 and 9) or 1,4-dioxane (Table, Entries 10 – 12) gave only poor
yields of diamine 5 and a lot of by-products 10, 11, or 12. Even variation of the reaction
conditions like temperature or pressure did not improve yields of 5. Despite an
acceptable outcome has been obtained with Raney Ni, still a plenty of catalyst was
required. Several attempts have been made to use a homogenous reduction system, e.g.,
such as Ru complexes with NHC ligands. Reduction of the dinitrile 9 with Grubbs
catalyst (2nd generation) [10] gave only amino carbonitrile 12 in poor yield with plenty
of unreacted starting material (Table, Entry 13). Formation of compound 11 was not
observed.
Conclusions. – The aim of our work was to synthesize diamines 5 from naturally
available camphor in large scale. After several modifications, furoxan/dinitrile were
obtained in quantities of as much as 40 g. Hydrogenation of the intermediate dinitrile
was performed best with the inexpensive catalyst Raney Ni in basic media to yield 5b.
The best solvent system was THF/H O/6m NaOH (Table, Entry 7).
2

Helvetica Chimica Acta – Vol. 98 (2015)
451
This project was funded in part by a BayernFit grant for Fraunhofer IGB.
Experimental Part
General. Yields were calculated from the area analyses of the gas chromatograms. Quantification
was accomplished for 5 only with the isolated pure compound as reference. The reductions were
performed in a 50-ml steel autoclave from Premex with glass beaker inside. Another problem was to
purify the amine from the by-products; several methods were tried, including distillation, salting out, or
chromatography on silica gel, ion-exchange resins, and alumina. The best separation results were
achieved by chromatography on basic alumina with MeOH/NH 100 :1. IR Spectra: Nicolet 380 FT-IR
3
À1
with a Smart Diamond ATR (Thermo Scientific); n˜in cm . NMR Spectra: JEOL ECS 400 spectrometer;
d in ppm rel. to Me Si as internal standard; J in Hz. GC/MS: HP 6890 with MS detector HP MS 5973 and
4
HP-5MS column (Hewlett Packard); in m/z.
(
1R,4S)-1,7,7-Trimethylbicyclo[2.2.1]heptane-2,3-dione 6 was obtained by oxidation of camphor (3)
with SeO2 [11]. (1R,4S)-1,7,7-Trimethylbicyclo[2.2.1]heptane-2,3-dione dioxime (7) was prepared by
reaction of a threefold excess of NH OH · HCl with 6 [11][12]. (4R,7S)-4,5,6,7-Tetrahydro-4,8,8-
2
trimethyl-4,7-methano-2,1,3-benzoxadiazole 1-oxide 8 was synthesized from 7 with NaClO [8].
1
,2,2-Trimethylcyclopentane-1,3-dicarbonitrile (9). (EtO) P (15 ml) was heated to 1008, then 7 g
3
(
36 mmol) furoxan 8 were added in small portions (Caution: large amounts of furoxan can decompose
explosively if heated in one portion!). After complete addition, the soln. was refluxed at 1508 for 5 h.
After cooling to r.t., the soln. was poured into 70 ml of H O, and 1 ml HCl (37%) was added. Overnight,
2
white crystals precipitated, and they were separated by vacuum filtration (4.2 g, 81%).
An alternative route was to use trialkylphosphates as solvents. Furoxan 8 (1.6 g, 8.2 mmol), dissolved
in 20 ml of (EtO) PO was slowly added via a dropping funnel to 20 ml of (EtO) P, which was preheated
3
3
to 1008. Next, the mixture was heated to 1508 for 3 h. Then, (EtO) PO and (EtO) P were distilled off in
3
3
vacuum (0.4 mbar). The distillation residue was treated with 10 ml of H O and 0.5 ml of HCl (37%) to
2
precipitate 9, which was collected by vacuum filtration (0.8 g, 60%). IR (ATR): 2885 – 2977, 2228, 1475,
1
1
454, 1397, 1385, 1370, 1300, 1243, 1206, 1180, 1149, 1128, 1093, 1041, 952, 579, 470. H-NMR (400 MHz,
CDCl ): 1.2 (s, 3 H); 1.3 (s, 3 H); 1.4 (s, 3 H); 1.8 – 1.9 (m, 1 H); 2.1 – 2.2 (m, 1 H); 2.2 – 2.3 (m, 1 H); 2.4 –
3
1
3
2
4
.5 (m, 1 H); 2.7 (dt, J ¼ 1.8, 9.6, 1 H). C-NMR (100 MHz, CDCl ): 21.0; 22.0; 23.4; 25.7; 35.7; 38.2; 44.8;
3
7.4; 119.8; 122.5. GC/EI-MS: 161, 147, 135, 119, 108, 95, 82, 68, 53, 41, 28, 15.
Hydrogenations with Raney Ni (Table, Entries 1 – 7). Dinitrile 9 (1.5 g, 9.3 mmol) was dissolved in
ml of EtOH, THF, or 1,4-dioxane (better solubility in dioxane and THF), and 8 ml of NaOH (aq., 6m)
8
were added. To this soln. was slowly added Ni/Al alloy (0.4 – 0.7 g; SigmaÀAldrich). Rapid evolution of
H occurred while Al dissolved. Next, the autoclave was purged twice with N to remove O . Then, the
2
2
2
vessel was pressurized with H (55 – 75 bar). The mixture was stirred overnight at 80 – 1408. After cooling
2
to r.t., Raney Ni was filtered off. The org. solvent was removed by rotary evaporation, und the remaining
H O phase was extracted with Et O (3 Â 20 ml). Extraction with AcOEt was unsuitable, because ethyl
2
2
amides were formed. The org. layer was separated, dried (Na SO ), and the solvent was evaporated. The
2
4
resulting mixture was purified by CC (basic alumina; with MeOH/25% NH · H O 100 :1; 10.5 cm Â
3
2
4
.5 cm; TLC (aluminium oxide 60 neutral): R 0.4).
f
Hydrogenations with Raney Ni (Table, Entries 8 – 12). Raney Ni was freshly prepared from Ni/Al
alloy (0.5 g) with NaOH (aq., 6m), filtered off, and rinsed with H O. Dinitrile 9 was dissolved in 10 ml of
2
EtOH, THF, or 1,4-dioxane, and Raney Ni was added. Next, the autoclave was purged twice with N to
2
remove O . Then, the vessel was pressurized with H (55 – 120 bar). From this point on, the procedure
2
2
was the same as in Entries 1 – 7.
(
1,2,2-Trimethylcyclopentane-1,3-diyl)bis[methanamine] (5). IR (ATR): 3288, 2940, 2868, 1567, 1453,
1
1
2
(
372, 1314. H-NMR (400 MHz, CDCl ): 0.6 (s, 3 H); 0.8 (s, 3 H); 0.9 (s, 3 H); 1.2 (m, 1 H); 1.4 – 1.5 (m,
3
13
H); 1.8 – 1.9 (m, 3 H); 2.4 (t, J ¼ 9.6, 1 H); 2.5 (s, 2 H); 2.7 (dd, J ¼ 3.7, 4.1, 11.9, 1 H). C-NMR
100 MHz, CDCl ): 18.4; 20.3; 23.2; 26.4; 34.7; 44.0; 44.3; 48.52; 49.0; 51.7. GC/EI-MS: 170, 153, 141, 126,
3
1
09, 95, 81, 69, 56, 41, 30.
(
1R,5S)-1,8,8-Trimethyl-3-azabicyclo[3.2.1]octane (11). IR: (ATR): 2923, 2870, 1567, 1557, 1454,
1
1
381, 1312, 1169, 1089, 1047, 978, 881, 815, 686, 620, 433. H-NMR (400 MHz, CD OD): 0.8 (s, 3 H); 0.9 (s,
3

4
52
Helvetica Chimica Acta – Vol. 98 (2015)
3
1
3
H); 1.0 (s, 3 H); 1.5 – 1.7 (m, 4 H); 1.9 – 2.0 (m, 1 H); 2.3 (d, J ¼ 12.8, 1 H); 2.5 (dd, J ¼ 2.8, 3.2, 10.1,
13
H); 2.9 (d, J ¼ 12.8, 1 H); 3.1 (d, J ¼ 13.3, 1 H). C-NMR (100 MHz, CD OD): 16.2; 17.1; 23.6; 24.7;
3
3.7; 41.5; 42.3; 45.2; 46.5; 52.6. GC/EI-MS: 153, 138, 122, 107, 98, 91, 81, 67, 55, 44, 30.
(
1S,3R)-3-(Aminomethyl)-2,2,3-trimethylcyclopentane-1-carbonitrile (12). IR (ATR): 3234, 2927,
2
4
1
873, 1661, 1557, 1471, 1456, 1393, 1379, 1354, 1318, 1227, 1158, 1221, 1066, 1043, 941, 881, 813, 619, 560,
1
83, 459, 421. H-NMR (400 MHz, CD OD): 0.7 (s, 3 H); 1.2 (s, 3 H); 1.3 (s, 3 H); 1.4 (m, 1 H); 1.8 (m,
3
13
H); 2.0 (m, 1 H); 2.1 (m, 2 H); 2.4 (dd, J ¼ 10.7, 11.9, 1 H,); 2.8 (dd, J ¼ 3.6, 12.4, 1 H). C-NMR
(100 MHz, CD OD): 14.8; 17.7; 22.1; 25.6; 35.2; 42.5; 44.9; 51.2; 125.0. GC/EI-MS: 166, 149, 134, 122, 109,
3
9
7, 77, 67, 53, 41, 30.
Hydrogenations with Grubbs 2nd-Generation Catalyst (Table, Entry 13). Dinitrile 9 (0.5 g, 3 mmol),
BuOK (0.11 g, 1 mmol), and Grubbs 2nd-generation catalyst (7.3 mg, 0.009 mmol) were dissolved in
t
1
5 ml of toluene under Ar and transferred into an autoclave. Then, 80 bar pressure of H was applied, and
2
the soln. was stirred at 508 for 72 h.
Hydrogenations with Pd. Dinitrile 9 (0.25 g, 1.5 mmol) and Pd/C (10% Pd) (0.14 g) were dissolved in
5
0 ml of EtOH. The soln. was pressurized with 50 bar of H in an autoclave and stirred at 708 for 5 h. No
2
products were formed.
Hydrogenations with Pt. Dinitrile 9 (0.25 g, 1.5 mmol) and Pt/C (5% Pt) (0.06 g) were dissolved in
6
0 ml of EtOH. The soln. was pressurized with 70 bar of H in an autoclave and stirred at 708 for 5 h. No
2
products were formed.
Hydrogenations with Rh. In an autoclave, dinitrile 9 (0.5 g, 3 mmol) and Rh/Al O (5% Rh; 14 mg,
2
3
0
.1 mmol) were dissolved in 50 ml of MeOH sat. with NH ; 25/55 bar of H pressure were applied, and
3
2
the soln. was stirred for 20 h at r.t. or at 1008. No products were formed.
REFERENCES
[
[
1] D. Jaramillo, D. P. Buck, J. G. Collins, R. R. Fenton, F. H. Stootman, N. J. Wheate, J. R. Aldrich-
Wright, Eur. J. Inorg. Chem. 2006, 839.
2] Z.-H. Yang, L.-X. Wang, Z.-H. Zhou, Q.-L. Zhou, C.-C. Tang, Tetrahedron: Asymmetry 2001, 12,
1579.
[
[
[
[
[
[
3] T. Ohkuma, T. Hattori, H. Ooka, T. Inoue, R. Noyori, Org. Lett. 2004, 6, 2681.
4] W. P. Hems, M. Groarke, A. Zanotti-Gerosa, G. A. Grasa, Acc. Chem. Res. 2007, 40, 1340.
5] C. de Bellefon, P. Fouilloux, Catal. Rev.-Sci. Eng. 1994, 36, 459.
6] A. Dornow, K. J. Fust, H. D. Jordan, Chem. Ber. 1957, 90, 2124.
7] J. Bredt, M. d. Souza, J. Prakt. Chem. 1932, 133, 84.
8] J. Ackrell, M. Altaf-ur-Rahman, A. J. Boulton, R. C. Brown, J. Chem. Soc., Perkin Trans. 1 1972,
1587.
[
9] M. Freifelder, J. Am. Chem. Soc. 1960, 82, 2386.
[
10] X. Miao, C. Fischmeister, C. Bruneau, P. H. Dixneuf, J.-L. Dubois, J.-L. Couturier, ChemSusChem
012, 5, 1410.
2
[
[
11] J. D. White, D. J. Wardrop, K. F. Sundermann, Org. Synth. 2002, 79, 125.
12] M. O. Forster, J. Chem. Soc., Trans. 1903, 83, 514.
Received November 3, 2014