Avoiding disturbances in sensorgrams

Disturbances in sensorgrams are mainly caused by improper experimental conditions like buffers that are not degassed, samples with particles and solutions, which differ too much in concentration. Most of these disturbances can be overcome by proper preparation of the samples and adjusting the method of injection.

Three types of disturbances (errors) can be distinguished.
Error Description Cause Solution
obvious error outliners clearly outside the expected range operator error, precipitates, gas bubbles proper sample preparation
systematic error every time the same conditions are run bulk/solvent effects, concentration errors calibration, adjustment method
random errors occurs irregularly contamination cleaning, adjustment method

Typical disturbances in sensorgrams

Flow change
Change in flow rate can cause a drift in the sensorgram, which will level out over time (5 - 30 minutes). The duration of this effect depends on the type of sensor surface and ligand bound to it. The effect can also be seen at the start of a sensorgram. One remedy is to start the sensorgram at the desired flow rate and add a WAIT command of 15 minutes to minimize the effect.

Drift and shift
Drift and shift are often caused by differences in flow buffer. Make sure that one batch of buffer is used and the PRIME-command is performed after buffer change. Another cause can be the build-up of small air bubbles in the flow channels. Especially at low flow rates (< 10 µl/min) the small air bubbles are not driven out fast enough, giving them time to grow until they come visible in the sensorgram. At high temperatures, e.g. 37°C air bubbles are easily formed. Thoroughly degassed buffers will help to minimize the effect. When using a method, a high flow rate step (100 μl/min) can be used to flush out small air bubbles between cycles. In addition, sufficient equilibration after immobilization is necessary because of the chemicals used. It is recommended to equilibrate the system over night after immobilization.

'Wave' curve
When upon injection of buffer (e.g. during start-up cycles or blanks) the curve has some 'wave'-form this can indicate that system needs some cleaning. First prime the system to look if a better equilibrated system has a better and more stable injection curve. If the waviness is still there, then clean the system with a desorb and sanitize solution. Let the system equilibrate and make sure that the pumps are full with liquid without any air bubbles.

Carry over
In spite of the built-in washing procedures used during method runs, some carry over does occur when liquids of high viscosity or high molarity (e.g. some regeneration solutions) are used. Extra washes or special cleaning may be necessary to minimize carry-over effects. However, this procedure is sometimes not sufficient to wash the needle and sample line. A sequence of wash commands is given in the table below.

One sequence of wash is given in the table below:

typical wash sequence.
Command Volume Remark
Inject 20 µl Injection of sample with high viscosity
Extraclean - Standard cleaning
Transfer 450 µl Transfer flow buffer, maximal volume to clean the needle and tubing
Wash i - Wash all IFC channels

Sample dispersion
During the injection, the sample is separated from the flow buffer by air plugs. The air plugs will minimize the mixing of the sample with the flow buffer minimizing concentration differences at the beginning and the end of the analyte injection. To check for a uniform sample plug, inject running buffer with an extra 25 mM NaCl. A non-uniform sample plug may indicate that the system needs cleaning.
Use the best injection command to minimize dispersion at the beginning and end of the analyte injection. Minimizing needle and auto sampler movement will also give a better dissociation curve, especially with low response levels. Long injection times can suffer from some dispersion at the end of the injection. Especially when steady state is reached, some downward response can be seen.

Spikes after reference subtraction
After reference subtraction, large spikes can be seen at the beginning and end of an injection. The overall association and dissociation curves are correct except for at the beginning (1-4 seconds) and at the end of the injection. This is because the flow channels are in series and the sample moment deviates slightly. The effect is that the reference is a little "out of phase" with the other channels. Using the inline reference subtraction function of the control software when available can minimize this effect. In addition, changing all the samples to the flow buffer can minimize the bulk refractive index problems further.