


Adsorptive Radial Column Filtration (ARCF)IntroductionGradients are used in process chromatography separations by continuously varying the liquid phase with respect to the solid phase. This allows components to be selectively desorbed from the solid phase. It is important that the variation in the gradient to be repeatable with respect to time and scale, so that the product of interest elutes with a consistent temporal separation from any impurities. If a gradient is not reproductable, impurities may not be sufficiently separated and may coelute with the collection of the product.There are three common ways of forming gradients in largescale chromatography systems (>1 L/min flow rate):
1) TwoTank Method
Each method has advantages and disadvantages which are discussed here. Theoretical output curves are provided for each method. The twotank method is the least expensive alternative, but very limited in producing multistep gradients. The switching valve method is more expensive, but limited in range and accuracy. The dual pump method is more expensive, but is the most versatile gradient making method. TwoTank MethodThe twotank method of a gradient formation involves the use of two buffer holding talks and 2 single speed pumps. Tank B is initially 100% B, and Tank A is 100% A.
This assumes that mixing of A+B is instantaneous, but this will occur if you use a sparger in the A+B tank.
If flow of B= ½ for of A+B then a continuous linear gradient from 0100% will result
Valve Switching MethodThis method uses a single pump and two switching valves, or one switching threeway valve to produce the gradient.
The valves alternate between components A and B as the mixture is pumped. Nonlinearity occurs in the gradient due to the fact that pumps pulse and have more suction during one part of their cycle. As the pump suction stroke goes in and out of phase with either component, the pump will pull too much or too little of a given component into the mixture.
The blips in the gradient curve can be reduced significantly by including a large mixing volume in the system, but this reduces the response in step gradients and greatly increases dilution of sample and other buffers.
Also, the valves cannot switch instantaneously. It takes a certain amount of time to switch. This limits the minimum time it can dwell on either component A or B, and thus the minimum % gradient that can be formed.
Two Pump MethodEach component is separately pumped through two variable speed pumps:
The twopump method is the most precise and accurate method for gradient formation over a wide range. The two pump speeds can be controlled to form multistep gradients over varying ranges depending on the choice of pumps and the gradient flow rate. All pumps have a maximum speed at which they can operate. They also have a minimum speed at which they can be accurately controlled. The ratio of maximum speed to minimum speed is called the turndown ration and is expressed as a ratio such as 10:1. Rotary lobe pumps have a small gap between the rotors and between the rotors and the housing. Fluid flows backwards past the rotors through this gap. As the flow rate decreases, or the pressure differential across the pump increases, the amount of "blowby" increases. These pumps are controlled by feedback from flow meters to keep the flow rate accurate. At flow rates below approximately 10% of the maximum rating of rotary lobe pump, it is very difficult to control the output of the pump. Diaphragm pumps have two means of control over the flow rate. The stroke length may be adjusted as well as the speed of the pump. The stroke length can typically be adjusted over a 10:1 range, from approximately 1.5 mm to 15.0 mm. The stroke frequency can also be adjusted over a range of 10:1, from 70 to 170 strokes per minute. Diaphragm pumps are true positive displacement pumps. Their flow rate does not vary significantly with varying pressure differential across the pump. They are typically accurate to ± 0.5% over a 10:1 turndown of speed alone and ± 2.5% over larger 100:1 turndown of both speed and stroke. The twopump method gradient accuracy therefore depends on the maximum flow rate of the pumps used, and their turndown ratios:
If the pumps are forming a gradient at their maximum rated speed, and they have a turndown ratio of 10:1 they can form an accurate gradient from 1090%, but at gradient flow rates below their maximum speed, the range of their gradient capability is reduced. If two pumps with 10:1 turndowns are forming a gradient at ½ their rate speed, a gradient may only be formed from 20 to 80% composition:
Below 20% of its rated maximum speed it cannot form a gradient at all. The graph below shows a "Fish Curve" for a 50L/min pump with a 10:1 turndown ratio. Gradients can be formed if they fall within the curve. For example, the chart below shows that at a flow rate of 15L/min, these pumps can form a gradient between 35 and 65%.
Forming a gradient over a wide flow range is different for a rotary lobe pump. The diaphragm pump allows the user to set the stroke length to the maximum speed of the gradient. Then the stroke frequency can be turned down by 10:1 to form a gradient over the range of 1090%. Therefore, the fish curve looks like this:
ConclusionThe twotank method is the most accurate and least expensive to product full scale (0100%) accurate gradients. If gradient accuracy is less important, or large mixing volumes are needed in the system, the switching valve method can also be used effectively. Sepragen recommends the use of diaphragm pumps and high pressure mixing to form accurate gradients over a wide range of flow rates. This method also allows versatility in forming multistep linear gradients and online blending.


