 |
Case Studies - Pd
Scavenging 1.0
Pd Scavenging 1.0
In medicinal chemistry transition metal catalysis
is one of the most interesting and under utilized classes of organic
transformations
due to the toxicity of the metal itself and its inevitable presence
in the final product. Certain transition metals are starting to see
regular use though they cause numerous headaches for the pharmaceutical
process chemist during scale-up. Palladium(Pd) is probably the most
commonly used transition metal catalyst and can be used in a variety
of synthetic transformations Heck1 and Suzuki cross-coupling reactions2,3
are two of the better known reactions.
The residual Pd is not an issue for the in-vitro screens however
if the compound proves successful a method of removal must be found
before the in-vivo work can begin. In this study we examine the effectiveness
of utilizing silica bound metal scavengers to remove Pd. To test
the robustness of the scavengers and to determine the best conditions
to bring the level of Pd down to an acceptable level for the pharmaceutical
industry we investigated the following parameters:
- Temperature
- Solvent
- Nature of the complex
- Number of equivalents
- Pore size
- Reaction time
Silica Bound Metal Scavengers
This study deals with the following functionalized silicas developed
by SiliCycle® Inc:
Si-Thiol
|
|
Si-Thiourea
|
|
Si-Triamine
|
|
Si-Triaminetetraacetatic
Acid
Si- TAA* |
|
|
| * Now available as the acid (Si – TAAcOH)
or sodium salt (Si – TAAcONa) for improved selectivity |
End-capped functionalized silica gels (figure 1) have many advantages
over their polymer counterparts; the major advantage being the inert
inorganic backbone that eliminates issues associated with swelling
and solvent compatibility. It also means they can be added directly
to the reaction mixture or used in a column to selectively remove
the metal making the process scalable all the way from the R&D
lab up to the process scale, without protocol modifications.
| Figure 1. |
The structure of functionalized end capped
silica gel. The effectiveness of these scavengers can be seen
qualitatively in Figure 2. The orange colour of the Pd(AcO)2
disappears from the solution 5 minutes after adding the silica
supported thiol. |
| Figure 2. |
On the left is a solution of Pd(AcO)2 on
the right is the same solution 5 minutes after Si – Thiol was added. |
Effect of Temperature
The effect
of the temperature has been investigated for the four scavengers
(Graph 1). It is easy to see from the graph that temperature
has a huge effect on the scavenging in four cases. We started with
an initial concentration of 1000 ppm of Pd from Pd(AcO)2 in DMF
and added 2 equivalents of scavenger. The Pd levels were measured
after one hour. It does not matter which of the scavenger is used,
as the temperature is increased the residual amount of Pd decreases
dramatically.
| Graph 1. |
| Effect of the temperature on
the Pd scavenging efficiency of four scavengers (1 hour, DMF,
2 equivalents, starting concentration of 1000 ppm Pd from Pd(AcO)2 |
Effect of Solvent
We screened 4 solvents against the 4 scavengers
at 2 different concentrations (2 and 4 equiv). From Graph 2 we can
discern that the scavengers are fairly solvent independent. They
do appear to be less effective in DMF but increasing the number of
equivalents eliminated this affect. The results could be explained
by the fact THF, toluene and DCM are non-coordinating solvents but
DMF could coordinate with Pd through the electrons of the carbonyl
group and with a p interaction of the planar OCN group. This coordination
of the solvent with Pd may be the cause of decrease in the scavenging
kinetics.
| Graph 2. |
Solvent effect on the scavenging
of Pd (1 hour, starting concentration of 0 ppm Pd from Pd(AcO)2. |
Nature of the Pd Complex
The nature of the Pd complex or the ligand coordinated to the metal
is very important for the reaction itself and for the removal of
the metal since the scavenger needs to have a stronger affinity for
the Pd than the ligand to effectively remove it. Strong ligands will
hinder the scavenging of the metal much more than weak ones, strength
is determined by the size of the ligand and the level of bonding
energy that it has toward Pd. Another important factor is the oxidation
state of Pd. Pd(0) will be more difficult to scavenge because it
has tetrahedral orbitals that will react by the Sn1 mechanism which
is much slower that the Sn2 mechanism that the square planar Pd(II)
moiety will undergo. To test these assumptions, we screened the following
complexes: Pd(AcO)2, Pd2(C3H5)2Cl2,
Pd(PPh3)4 Pd2(dba)3 against our 4 scavengers and followed the scavenging
kinetics
over
time. The
results are presented in the Table 1. As predicted, the easiest Pd
complexes to remove was the Pd(II) of the acetate and the other ionic
complex Pd2(C3H5)2Cl2. The two Pd(0) complexes Pd(PPh3)4 and Pd2(dba)3 were not scavenged as effectively overall but the TAA and Thiourea
were effective over time. Further optimization could be done such
as increasing the temperature to increase the effectiveness. It is
interesting to note that with Pd2(dba)3 the lowest level of Pd was
achieved after one hour. This could be due to the reaction slowly
reaching equilibrium. It is therefore important to follow the kinetics
of any scavenger evaluation to optimize your process.
Scavenger |
Time of Rxn |
Pd(AcO)2 |
Pd2(C3H5)2Cl2 |
Pd(PPh3)4 |
Pd2(dba)3 |
Si-Thiol
|
5 min |
0.90 |
- - - |
360 |
545 |
|
60 min |
0.07 |
0.04 |
320 |
20 |
|
18 h |
0.05 |
- - - |
150 |
100 |
|
|
|
|
|
|
Si-Thiourea
|
5 min |
1.4 |
- - - |
320 |
475 |
|
60 min |
0.8 |
1.3 |
95 |
50 |
|
18 h |
0.6 |
- - - |
10 |
90 |
|
|
|
|
|
|
Si-TAA
|
5 min |
40 |
- - - |
390 |
480 |
|
60 min |
9.8 |
0.25 |
150 |
50 |
|
18 h |
0.06 |
- - - |
1.4 |
190 |
|
|
|
|
|
|
Si-Triamine
|
5 min |
20 |
- - - |
540 |
525 |
|
60 min |
1.4 |
1.3 |
370 |
83 |
|
18 h |
0.3 |
- - - |
220 |
280 |
|
| Table 1. |
| Scavenging of different Pd complexes
at RT with 4 equivalents for different reaction times in THF. |
Number of equivalents
As
with most reactions the number of equivalent is also very important
as can be seen in Graph 3 which plots the Pd level over time
when using 1, 2, and 4 equivalents of Si – Thiol to remove
Pd(AcO)2 in THF at room temperature. The initial concentration was
1000
ppm. After 1 hr there is a profound increase in the amount of
Pd removed with two equivalents versus one and slight increase with
four versus two. Overtime however the effect decreases. You must
balance the time savings and effeciency against the increased
cost
of adding additional scavenger.
| Graph 3. |
| Pd scavenging using 3 different
quantities of Si-Thiol scavenger Conditions: Pd(AcO)2, THF, Si-Thiol,
RT |
Pore size
Finally we examined
the effect of the pore size of the silica support on the less reactive
Pd complexes. Graph 4 shows the level of palladium
tetrakistriphenylphosphine after 1hour with Si –Thiol grafted
on 60, 90 and 150Å silica gel in THF, the initial concentration
was 1000 ppm. The larger pore size does appear to be more effective
this could be due to steric effect or a higher diffusion coefficient,
further study is required to evaluate the role of pore size.
|
Conditions: Si-Thiol, THF, Pd(PPh3)4,
4 eq., RT, 1h
|
|
| Graph 4. |
Effect
of the pore size on the scavenging of Pd tetrakistriphenylphosphine
complex |
Conclusion
In this study we attempted to illustrate the effectiveness and study
the influence of different parameters on the scavenging of palladium
complexes with silica bound scavengers. Silica bound scavengers are
very effective in the removal of Pd and you can easily optimize the
process by adjusting several variables including the scavenger used,
temperature, solvent, reaction time and number of equivalents. When
selecting a scavenger the most important factors to consider are
the oxidation state of the metal and the ligand. Additionally although
its was not covered in this study the starting material and final
product are also expected to influence the removal of Pd, if either
components has a strong affinity for Pd, its removal will hindered.
| |
References
1. Angew. Chem. Int. Ed. Eng., 33 (1994) 2379
2. Acta Chem. Scand., 47 (1993) 221
3. J. Org. Chem., 59 (1994) 5034
|
|
