Tutorial 7

Downstream Processing 2
Example 7.1
Scaleup of a Protein Chromatography
A column 20 cm long, with an internal diameter of 5 cm, gives sufficient purification to
consider a scaleup. The column produces 3.2 g of purified protein per cycle, and a cycle takes
6 h, from equilibration through regeneration. You are aiming for a throughput of 10 g/h.
What are the new column's dimensions if linear velocity is held constant?
Note:
Throughput measures rate of output and, therefore, quantifies how fast you develop
products. Throughput is about cycle time.
Productivity measures how efficiently you develop them. Productivity is about cost.

Solution:
For scaleup when the linear velocity is held constant, the column diameter is increased,
and the column height is maintained the same (slide 12.33).
If the linear flow rate is held constant, then the cycle time cannot be altered.
Therefore, the cycle time is 6 h/cycle. On small scale we can produce 3.2 g/cycle or 0.53
g/h. We need to calculate the scale up factor.
Since the flow rate is proportional to the throughput of protein:
𝑄𝑄2 10 𝑔𝑔/ℎ
= = 18.87
𝑄𝑄1 0.53 𝑔𝑔/ℎ
From Eq. (12.15-16), the scaleup relationship when the gradient and the particle size are
not changed upon scale up, and since L1 = L2:
𝐷𝐷2 2
𝑄𝑄2 𝑉𝑉2 𝜋𝜋 � 2 � 𝐿𝐿2 𝐷𝐷2 2
= = = � � = 18.87
𝑄𝑄1 𝑉𝑉1 𝐷𝐷1 2 𝐷𝐷1
𝜋𝜋 � � 𝐿𝐿1
2
where D1 and D2 are the column diameters for columns 2 and 1, respectively.
Since D1 = 5.0 cm, we obtain:
𝐷𝐷2 = √18.87 × 5 = 21.7 𝑐𝑐𝑐𝑐
1
Example 7.2
Scaleup of Protein Chromatography using Standard Column Sizes
Use data from the example 7.1 and available standard column diameters of 20 and 25
cm.
What flow rates and bed depths would apply to each of these columns?

Solution:
The column volumes for both columns are still 18.87 times that used in the laboratory
column. Therefore:
𝐷𝐷1 2 5 𝑐𝑐𝑐𝑐 2
𝑉𝑉2 = 18.87 𝑉𝑉1 = 18.87𝜋𝜋 � � 𝐿𝐿 = 18.87 𝜋𝜋 � � × 20 𝑐𝑐𝑐𝑐 = 7406 𝑐𝑐𝑐𝑐3
2 2
For a column 20 cm in diameter:
𝑉𝑉2 7406 𝑐𝑐𝑐𝑐3
𝐿𝐿 = = = 23.6 𝑐𝑐𝑐𝑐
𝐷𝐷 2 20 2
𝜋𝜋 � 2 � 𝜋𝜋 � 2 �
2
For a column 25 cm in diameter:
𝑉𝑉2 7406 𝑐𝑐𝑐𝑐3
𝐿𝐿 = = = 15.1 𝑐𝑐𝑐𝑐
𝐷𝐷 2 25 2
𝜋𝜋 � 2 � 𝜋𝜋 � �
2 2

2
Example 7.3
Hormone Separation using Gel Chromatography
A pilot-scale gel-chromatography column packed with Sephacryl resin is used to separate
two hormones, A and B. The column is 5 cm in diameter and 0.3 m high. Its void volume
is 1.9x10-4 m3. The water regain value of the gel is 3x10-3 m3 kg-1 dry Sephacryl; the mass
of dry gel is 125 g. The partition coefficient for hormone A is 0.38; the partition
coefficient for hormone B is 0.15.
If the eluant flow rate is 0.7 L h-1, what is the retention time for each hormone?

Solution

Given:
Hormones A & B: 𝐾𝐾𝑝𝑝𝑝𝑝 = 0.38 ;𝐾𝐾𝑝𝑝𝑝𝑝 = 0.15
Gel column: ℎ = 0.3 𝑚𝑚; 𝐷𝐷 = 5 𝑐𝑐𝑐𝑐; 𝑉𝑉0 = 1.9 × 10−4 𝑚𝑚3 ; 𝑄𝑄 = 0.7 𝐿𝐿/ℎ
Stationary phase (gel): water regain 𝑊𝑊𝑟𝑟 = 3 × 10−3 𝑚𝑚3 /𝑘𝑘𝑘𝑘;
the mass of dry gel 𝑎𝑎 = 125 𝑔𝑔

We can calculate retention time from definition of volumetric flow rate: 𝑄𝑄 = 𝑉𝑉/𝑡𝑡
Therefore, we need to determine the retention volume, which can be done from eq. (12.5)
𝑉𝑉𝑒𝑒 = 𝑉𝑉0 + 𝐾𝐾𝑝𝑝 𝑉𝑉𝑖𝑖 (12.5)
We know 𝑉𝑉0 , 𝐾𝐾𝑝𝑝 but not 𝑉𝑉𝑖𝑖 .
The water regain is the volume of water taken up per mass of dry gel:
𝑉𝑉𝑖𝑖 = 𝑎𝑎𝑊𝑊𝑟𝑟

−3
𝑚𝑚3
𝑉𝑉𝑖𝑖 = 0.125 𝑘𝑘𝑘𝑘 �3 × 10 � = 3.75 × 10−4 𝑚𝑚3
𝑘𝑘𝑘𝑘
𝑉𝑉𝑒𝑒𝑒𝑒 = 1.9 × 10−4 + 0.38(3.75 × 10−4 ) = 3.32 × 10−4 𝑚𝑚3
𝑉𝑉𝑒𝑒𝑒𝑒 = 1.9 × 10−4 + 0.15(3.75 × 10−4 ) = 2.46 × 10−4 𝑚𝑚3

3
 3.32 × 10−4 𝑚𝑚3
𝑡𝑡𝑒𝑒𝑒𝑒 = = 0.47 ℎ = 28.5 𝑚𝑚𝑚𝑚𝑚𝑚
𝐿𝐿 1𝑚𝑚3
0.7 � �
ℎ 1000𝐿𝐿
2.46 × 10−4 𝑚𝑚3
𝑡𝑡𝑒𝑒𝑒𝑒 = = 0.35 ℎ = 21.1 𝑚𝑚𝑚𝑚𝑚𝑚
𝐿𝐿 1𝑚𝑚3
0.7 � �
ℎ 1000𝐿𝐿

4
Example 7.4
Resolution and Selectivity
Albumin is being separated from IgG by chromatography using a 50 cm long column having
a voidage fraction of 0.25 and a diameter of 10 mm at a mobile phase flow rate of 10 ml/min.
The distribution coefficients for IgG and albumin are 1 and 0.1, respectively.
Determine:
7.4.1. If the albumin peak has a characteristic peak width of 0.52 minutes, predict the
selectivity and resolution.
7.4.2. When the mobile phase flow rate was increased to 20 ml/min the HETP was found to
increase by 80%. Predict the selectivity and resolution at the higher flow rate.

Solution:
7.4.1. If the albumin peak has a characteristic peak width of 0.52 minutes, predict the
selectivity and resolution.

Void volume can be represented as a voidage fraction of column volume:

𝑉𝑉0
𝜀𝜀 = (T7.1)
𝑉𝑉𝑇𝑇

Since
𝑉𝑉
𝑄𝑄 =
𝑡𝑡
Then the elution time for void:
𝑉𝑉𝑇𝑇 𝜀𝜀
𝑡𝑡0 = (T7.2)
𝑄𝑄

Then elution time for any compound can be calculated:

1−𝜀𝜀
𝑡𝑡𝑒𝑒 = 𝑡𝑡0 �1 + 𝐾𝐾� (T7.3)
𝜀𝜀

First, we need to calculate the volume of the column (volume of a cylinder):

𝐷𝐷 2 1 𝑐𝑐𝑐𝑐 2
𝑉𝑉 = 𝜋𝜋 � � 𝐿𝐿 = 𝜋𝜋 � � 50 𝑐𝑐𝑐𝑐 = 39.25 𝑐𝑐𝑐𝑐3 = 39.25 𝑚𝑚𝑙𝑙
4 4
The elution time for void (T7.2):
5
 (39.25 𝑚𝑚𝑚𝑚) × 0.25
𝑡𝑡0 = = 0.981 𝑚𝑚𝑚𝑚𝑚𝑚
10𝑚𝑚𝑚𝑚/𝑚𝑚𝑚𝑚𝑚𝑚
The retention times for IgG and albumin (T7.3):
1 − 0.25
𝑡𝑡𝑒𝑒,𝐼𝐼𝐼𝐼𝐼𝐼 = 0.981 �1 + × 1� = 3.924 𝑚𝑚𝑚𝑚𝑚𝑚
0.25
1 − 0.25
𝑡𝑡𝑒𝑒,𝑎𝑎𝑎𝑎𝑎𝑎 = 0.981 �1 + × 0.1� = 1.275 𝑚𝑚𝑚𝑚𝑚𝑚
0.25

The number of theoretical plates in the column can be calculated from albumin retention time
data using eq.
𝑡𝑡𝑒𝑒 2 1.27 2
𝑁𝑁 = 16 � � = 16 � � = 95
𝜔𝜔 0.52
The peak width of IgG can be calculated from the same equation but re-arranged:
𝑡𝑡𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 3.9
𝜔𝜔𝐼𝐼𝐼𝐼𝐼𝐼 = = = 1.6 𝑚𝑚𝑖𝑖𝑛𝑛
� 𝑁𝑁 �95
16 16
The resolution of the separation – eq. (12.13)
2(𝑡𝑡𝑒𝑒2 − 𝑡𝑡𝑒𝑒1 ) 2(3.9 − 1.27)
𝑅𝑅𝑠𝑠 = = = 2.58
𝑤𝑤1 + 𝑤𝑤2 1.6 + 0.52
The selectivity – eq. (12.2):
𝑘𝑘2 1
𝛿𝛿 = = = 10
𝑘𝑘1 0.1

7.4.2. When the mobile phase flow rate was increased to 20 ml/min the HETP was found
to increase by 80%. Predict the selectivity and resolution at the higher flow rate.

The height of a theoretical plate – re-arranged eq. (12.10):

𝐿𝐿 50
𝐻𝐻 = = = 0.526 𝑐𝑐𝑐𝑐
𝑁𝑁 95
At the higher flow rate. The height of the theoretical plate increased by 80%.
𝐻𝐻 = 1.8 × 0.526 = 0.95 𝑐𝑐𝑐𝑐
6
Therefore, the number of theoretical plate is reduced:
𝐿𝐿 50
𝑁𝑁 = = = 52
𝐻𝐻 0.95
The void elution time:
(39.25 𝑚𝑚𝑚𝑚) × 0.25
𝑡𝑡0 = = 0.491 𝑚𝑚𝑚𝑚𝑚𝑚
20𝑚𝑚𝑚𝑚/𝑚𝑚𝑚𝑚𝑚𝑚
The new retention times are:
1 − 0.25
𝑡𝑡𝑒𝑒,𝐼𝐼𝐼𝐼𝐼𝐼 = 0.491 �1 + × 1� = 1.964 𝑚𝑚𝑚𝑚𝑚𝑚
0.25
1 − 0.25
𝑡𝑡𝑒𝑒,𝑎𝑎𝑎𝑎𝑎𝑎 = 0.491 �1 + × 0.1� = 0.638 𝑚𝑚𝑚𝑚𝑚𝑚
0.25
The peak widths are:
𝑡𝑡𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 1.964
𝜔𝜔𝐼𝐼𝐼𝐼𝐼𝐼 = = = 1.09 𝑚𝑚𝑚𝑚𝑚𝑚
� 𝑁𝑁 �52
16 16
0.638
𝜔𝜔𝑎𝑎𝑎𝑎𝑎𝑎 = = 0.35 𝑚𝑚𝑚𝑚𝑚𝑚
�52
16
The selectivity is independent of the flow rate, it is still 10.
The resolution is:
2(𝑡𝑡𝑒𝑒2 − 𝑡𝑡𝑒𝑒1 ) 2(1.964 − 0.638)
𝑅𝑅𝑠𝑠 = = = 1.84
𝑤𝑤1 + 𝑤𝑤2 1.09 + 0.35

This means that despite doubling the flow rate and increase of height of a theoretical plate,
IgG and albumin can be still separated well enough and perhaps, even increase loading.