A new supported rhodium catalyst for selective hydrogenation of nitriles to primary amines

Article abstract of DOI:10.1135/cccc20070468

Nitriles are converted to primary amines with high selectivity using a newly developed alumina-supported rhodium catalyst. The high selectivity is obtained without any additives, which are often used to prevent the formation of higher amines. The catalyst is active under mild conditions In various solvents, which makes it specifically suitable for use in pharmaceutical applications or for other substrates that can react with additives like strong acids or bases.

Full text of DOI:10.1135/cccc20070468

468
Witte:
A NEW SUPPORTED RHODIUM CATALYST FOR SELECTIVE
HYDROGENATION OF NITRILES TO PRIMARY AMINES
Peter T. WITTE
BASF Nederland B.V., Strijkviertel 67, 3454 ZG De Meern, The Netherlands;
e-mail: peter.witte@basf.com
Received Decem ber 27, 2006
Accepted March 13, 2007
Dedicated to Dr Karel Mach on the occasion of his 70th birthday in recognition of his outstanding
contributions to the area of organometallic synthesis and catalysis.
Nitriles are con verted to prim ary am in es with h igh selectivity usin g a n ewly developed
alum in a-supported rh odium catalyst. Th e h igh selectivity is obtain ed with out an y additives,
wh ich are often used to preven t th e form ation of h igh er am in es. Th e catalyst is active un der
m ild con dition s in various solven ts, wh ich m akes it specifically suitable for use in ph arm a-
ceutical application s or for oth er substrates th at can react with additives like stron g acids or
bases.
Keyw ord s: Selective h ydrogen ation ; Reduction s; Nitriles; Am in es; Heterogen eous catalysis;
Rh odium .
Hydrogen ation of n itriles is an im portan t tool for organ ic ch em ists to
prepare com poun ds con tain in g am in e groups1. Catalytic h ydrogen ation
is preferred to stoich iom etric reduction s (e.g. usin g h ydridoalum in ates or
-borates), sin ce m uch less waste is form ed in catalytic reaction s. Th e m ech a-
n ism of catalytic n itrile h ydrogen ation an d couplin g to h igh er am in es is
depicted in Sch em e 1 2. Th e n itrile 1 is h ydrogen ated to an in term ediate
im in e 2, an d subsequen tly to prim ary am in e 3. Reaction of im in e 2 with
am in e 3 can lead to th e form ation of secon dary am in e 5, an d furth er cou-
plin g can lead to tertiary am in e 7.
In order to obtain h igh selectivity for prim ary am in es, large am oun ts of
additives are often used to suppress th e form ation of h igh er am in es. Th e
m ost com m on ly used additive is am m on ia3, wh ich can react with im in e 2
to in term ediate 8, th ereby preven tin g couplin g of 2 with 3 2b. However, in
order to rem ove 2 com pletely from th e reaction m ixture, large am oun ts of
am m on ia are n ecessary4. Couplin g of 2 an d 3 can also be preven ted by
proton ation 5 or acylation 6 of am in e 3, yieldin g 9 an d 10, respectively.
Collect. Czech. Chem. Commun. 2007, Vol. 72, No. 4, pp. 468–474
© 2007 Institute of Organic Chemistry and Biochemistry
doi:10.1135/cccc20070468

New Supported Rhodium Catalyst
469
However, th is requires at least a stoich iom etric am oun t of acid or acylatin g
agen t, an d an addition al deprotection step to obtain th e desired prim ary
am in e. Aqueous solution s of an in organ ic base h ave been used to preven t
am in e couplin g to take place7. Th is effect can n ot be explain ed by th e reac-
tion sch em e in Sch em e 1. An in organ ic base can n eutralize acid sites of th e
catalyst, wh ich can easily bin d basic am in es, an d are possibly respon sible
for couplin g of im in e 2 with am in e 3.
SCHEME 1
Reaction m ech an ism of n itrile h ydrogen ation to prim ary am in es
Usin g additives to in crease th e selectivity creates a waste problem if th e
additives can n ot be recycled, wh ile recyclin g often requires special equip-
m en t. In eith er way, additives m ake th e process m ore expen sive, an d less
easy to im plem en t on in dustrial scale. Dopin g of a catalyst can lead to an in -
crease in selectivity for prim ary am in es, with out th e n eed to use additives8.
However, so far all exam ples of doped catalysts used h igh tem peratures an d
pressures to obtain good activity of th e catalyst, wh ich m akes th e processes
un suitable for use in ph arm aceutical in dustry. A selective n itrile h ydroge-
n ation catalyst th at works un der m ild con dition s is th erefore desired.
EXPERIMENTAL
Th e catalysts were obtain ed from th e En gelh ard Co., Rom e or were prepared by stan dard de-
position -precipitation m eth ods (h om em ade). Solven ts an d substrates were obtain ed from
Sigm a–Aldrich an d used as received.
All h ydrogen ation s were carried out in a 250 m l stain less steel reactor, con tain in g baffles
for optim um m ixin g. Th e reactor was ch arged with 1.2 g (dry weigh t) of catalyst, 100 m l
96% EtOH, an d 10 m l butyron itrile. Th e reactor was pressurized with H2 to 5 × 102 kPa, an d
Collect. Czech. Chem. Commun. 2007, Vol. 72, No. 4, pp. 468–474

470
Witte:
subsequen tly ven ted. Th is step was repeated twice, after wh ich stirrin g was started (1500
rpm ) an d th e reaction m ixture was h eated to 50 °C. At th is tem perature th e reactor was
again pressurized with H2 to 5 × 102 kPa, wh ich was con sidered th e start of th e reaction .
Th e reaction was stopped wh en 4.0 l of H2 gas was con sum ed, wh ich correspon ds to a con -
version of ca. 80%. After open in g th e reactor, th e exact con version an d selectivity were cal-
culated from th e GC spectra9 usin g 1-dodecan ol as in tern al stan dard. All reaction s sh owed a
m ass balan ce >95%. (Agilen t 6850, HP-1 Meth yl siloxan e colum n , 30.0 m × 320 µm × 0.25
µm .)
RESULTS AND DISCUSSION
A first com parison of carbon -supported precious m etal catalysts (Ru, Rh ,
or Pd on carbon powder, CP) sh ows th at on ly with Rh /CP sm all am oun ts
of prim ary am in e are produced. Ru/CP is in active un der th e used reaction
con dition s, wh ile Pd/CP produces m ain ly tertiary am in e with sm all
TABLE I
Hydrogen ation of butyron itrile on Rh catalystsa
Reaction tim e
m in
Con version
%
Selectivity
%
En try
Supportb
Catalyst
1
2
CPc
CPc
CPc
CPc
CPc
CPc
Escat 340
90
37
70
73
72
73
74
73
79
76
86
73
11f
78
76
74
31
30
30
29
30
31
75
46
93
56
50g
77
58
49
Rom e 42753
3
Rom e 43798
45
4
Rom e 46617 egg sh ell
Rom e 43951 un iform
Rom e 43671 m ixed
Rom e 4824
h om em aded
h om em adee
h om em adee
h om em adee
h om em adee
h om em adee
35
5
38
6
38
7
AP
60
8
TSP
AP
50
9
85
10
11
12
13
14
TSP
SP
140
90
SAP1.5
SAP20
SAP40
105
125
120
h om em adee
a
b
All catalysts con tain 5% Rh -loadin g. CP, carbon powder; AP, alum in a powder; TSP, tita-
n ium -silicate powder; SP, silica powder; SAP, silica-alum in ate powder (th e n um ber in dicates
c
th e am oun t of silica in wt.%). For carbon -supported catalysts on ly 0.4 g (dry weigh t) of
d
e
catalyst was used. Preparation accordin g to th e Rom e procedure. Preparation accordin g to
f
g
an adapted procedure. Exten sive deactivation was observed. Based on com bin ed secon d-
ary am in e an d im in e form ation .
Collect. Czech. Chem. Commun. 2007, Vol. 72, No. 4, pp. 468–474

New Supported Rhodium Catalyst
471
am oun ts of secon dary am in e. Table I sh ows th e h ydrogen ation results for
a series of Rh catalysts on various supports.
Rh catalysts supported on carbon were m uch m ore active, but less selec-
tive th an Rh catalysts on m in eral supports. Variation s of carbon support
(en tries 1–3) or variation s in m etal location (en tries 4–6) h ad little effect on
th e prim ary am in e selectivity.
Rh catalysts on m in eral supports were less active, but th eir selectivity for
prim ary am in es was h igh er. Th e alum in a support yielded th e m ost selective
com m ercial catalyst (en try 7). Ch an ges in th e preparation m eth od in -
creased th e selectivity even m ore to 93% at 86% con version (en try 9). Th e
selectivity did n ot decrease wh en th e reaction was allowed to run to full
con version .
All reaction s could be run to full con version , except wh en silica was used
as support (en try 11). In th is case h ydrogen ation started with a com parable
rate as in en try 9, but quickly slowed down . After 90 m in H2 con sum ption
h ad alm ost com pletely stopped.
Table II sh ows h ydrogen ation results of th e n ewly developed Rh on alu-
m in a powder catalyst un der differen t reaction con dition s. Th e h ydrogen a-
tion selectivity is n ot depen den t on th e solven t polarity (en tries 9, 15),
alth ough th e catalyst is n ot equally active in all used solven ts (en tries 16,
17). Th e CH2=CH double bon d in allyl cyan ide is preferably h ydrogen ated
over th e n itrile fun ction ality (en try 18), wh ile th e ph en yl group in
3-ph en ylpropion itrile is stable to h ydrogen ation (en try 19), even after pro-
lon ged reaction tim es. Th e selectivity in th e h ydrogen ation of th is sterically
TABLE II
Nitrile h ydrogen ation s usin g h om em ade Rh /AP catalyst
Reaction tim e
m in
Con version
%
Selectivity
%
En try
Rem ark
9
15
16
17
18
19
Stan dard reaction a,b
Cycloh exan e as solven ta
NMP as solven ta
Eth yl acetate as solven ta
Allyl cyan ide as substrateb
3-Ph en ylpropion itrile as substrateb
85
120
90
86
83
0
93
92
–
60
12
28
100
87
c
95
–
90
98d
a
b
c
Butyron itrile as substrate. 96% EtOH as solven t. CH2=CH bon d com pletely h ydroge-
d
n ated. C6H5 group n ot h ydrogen ated.
Collect. Czech. Chem. Commun. 2007, Vol. 72, No. 4, pp. 468–474

472
Witte:
m ore h in dered substrate is m uch h igh er (~98%) th an for th e lin ear-ch ain
butyron itrile.
Our results of screen in g Rh , Ru, an d Pd supported on carbon powder are
in agreem en t with earlier publication s. With out additives, Pd catalysts gen -
erally yield tertiary am in es, wh ile Rh can be used to prepare secon dary
am in es1a,10. Our attem pts to in fluen ce th e selectivity of Rh /CP by ch an gin g
th e type of carbon support or location of th e m etal were un successful.
Ru/CP was in active, wh ich is also con sisten t with earlier results10. However,
Ru h as been reported to yield prim ary am in es with a reason able selectivity
wh en supported on zeolite, wh ile a lower selectivity was foun d wh en Ru
was supported on AP 11
.
Th e role of th e support in n itrile h ydrogen ation s h as been in vestigated in
several studies, but with con tradictory results. Huan g et al. in vestigated h y-
drogen ation of n itriles usin g supported Ru catalysts, an d did n ot observe
an y effect of th e support acidity11. Th ey claim th at an y acidity of th e sup-
port is n eutralized by th e basic am in es form ed in th e reaction . Verh aak et al.
observed th at catalysts with m ore basic supports were m ore selective, claim -
in g th at acid sites catalyze th e couplin g to h igh er am in es12. Dallon s et al.
foun d th at acid sites on th e catalyst actually in crease prim ary am in e selec-
tivity. Th ey argue th at th ese sites coordin ate prim ary am in es form ed in th e
h ydrogen ation , keepin g th em away from h ydrogen ation sites th at con tain
th e im in e in term ediate an d preven tin g couplin g of th e two com poun ds13.
It is our belief th at in teraction of th e support with th e active m etal deter-
m in es th e catalyst selectivity. Th is in teraction is in fluen ced by support
properties (such as its acidity) as well as by th e m eth od used to prepare th e
catalyst. In deed, our studies sh ow th at th e catalyst selectivity depen ds n ot
on ly on th e type of support, but th at th e m eth od of preparation of th e cata-
lyst also greatly affects its selectivity. Optim ization of th e catalyst prepara-
tion procedure for Rh /AP led to an in crease in prim ary am in e selectivity
from 75 to 93% at ~80% con version .
It is in terestin g to n ote th at th e silica-supported Rh catalyst (en try 11)
started with a sim ilar activity as th e alum in a-supported catalyst, but
quickly deactivated. GC m easurem en ts of th e reaction m ixture sh owed a
differen t com position of th e reaction products (Fig. 1). For very fast h ydro-
gen ation s catalyzed by Rh /CP (en tries 1–6), on ly peaks belon gin g to am in es
3, 5 an d 7 are observed. Th e slower but m ore selective h ydrogen ation s cata-
lyzed by Rh /AP (en tries 7, 9) sh ow peaks of am in e 3, 5 an d of an un kn own
com poun d. Th e peak of th e un kn own com poun d lies close to th e peak of 5;
wh en th e reaction is allowed to run to com pletion th e peak disappears. It
was th erefore ascribed to a h ydrogen ation in term ediate, m ost probably
Collect. Czech. Chem. Commun. 2007, Vol. 72, No. 4, pp. 468–474

New Supported Rhodium Catalyst
473
im in e 4. In case of th e Rh /SP catalyst (en try 11), th e peak of 4 is sign ifi-
can tly h igh er th an in GC spectra of oth er h ydrogen ation s. It seem s th at
im in e 4 is very difficult to h ydrogen ate with Rh /SP an d m ay act as a poi-
son . Wh en a catalyst with a very low activity is used (Ru/AP), th e peak of
im in e 4 is even h igh er th an th at of am in e 5 an d a sm all peak n ext to th e
peak of prim ary am in e 3 is observed, wh ich is ten tatively ascribed to pri-
m ary im in e 2. In th eir work on n itrile h ydrogen ation on supported Ru cata-
lysts, Huan g et al. observed th e com bin ation of secon dary im in e form ation
an d catalyst deactivation 10. Th ey observed th at perform in g th e reaction in
th e gas ph ase gives m uch m ore deactivation th an in th e liquid ph ase. Th eir
FIG. 1
Graph ical represen tation of GC spectra for selected reaction s (see Sch em e 1 an d Table I)
con clusion th at th e deactivation probably arises from “adsorbed m olecules”
is also con sisten t with our observation s.
In con clusion , we h ave developed a Rh /AP catalyst th at can selectively
h ydrogen ate n itriles to prim ary am in es (93% or h igh er yield) un der m ild
con dition s (50 °C, 5 × 102 kPa H2) with out usin g an y additives.
REFERENCES AND NOTES
1. a) Augustine R. L.: Heterogeneous Catalysis for the Synthetic Chemist, pp. 490–495. Dekker,
New York 1996; b) Carey J. S., Laffan D., Thomson C., Williams M. T.: Org. Biomol.
Chem. 2006, 4, 2337.
2. a) Volf J., Pašek J.: Stud. Surf. Sci. Catal. 1986, 27, 105; b) Gomez S., Peters J. A.,
Maschmeyer T.: Adv. Synth. Catal. 2002, 344, 1037.
3. See e.g.: Fuchigami T., Takamizawa S., Wakasa N.: U.S. 6,476,267, 2002.
Collect. Czech. Chem. Commun. 2007, Vol. 72, No. 4, pp. 468–474

474
Witte:
4. In ref.3 the ratio ammonia/nitrile 2:1 was used, which may not seem to be a large excess.
However, the intermediate imine 2, which is the actual reactive species, is present only
in very low concentrations. Therefore, the effective ratio ammonia/imine is much higher
than 2:1.
5. See e.g.: Jones R. V. H., Brown S. M.: U.S. 6,894,195, 2005.
6. See e.g.: Coqueron P. Y., Lhermitte F., Perrin-Janet G., Dufour P.: Eur. 1 674 455, 2004.
7. See e.g.: Zuckerman M. F.: U.S. 4,739,120, 1988.
8. a) Drake C.: U.S. 4,003,933, 1977; b) Sijpkes A. H., van den Brink P. J., de Ruiter R.: W.O.
03/018532, 2003.
9. conversion = 100 × (nitrileEXP – nitrileGC)/nitrileEXP; selectivity = 100 × primaryGC
/
(primaryGC + secondaryGC + tertiaryGC); mass balance = 100 × (nitrileGC + primaryGC
+
secondaryGC + tertiaryGC)/nitrileEXP; nitrileEXP = mol of butyronitrile used in experiment;
compoundGC = mol of compound observed in GC spectra.
10. Rylander P. N., Koch J. H.: U.S. 3,177,258, 1965.
11. Huang Y., Adeeva V., Sachtler W. M. H.: Appl. Catal., A 2000, 196, 73.
12. Verhaak M. J. F. M.: Ph.D. Thesis. Utrecht University, Utrecht 1992.
13. Dallons J. L., van Gysel A., Jannes G.: Catalysis of Organic Reactions, pp. 93–104. Dekker,
New York 1992.
Collect. Czech. Chem. Commun. 2007, Vol. 72, No. 4, pp. 468–474