Featured Article
Automation solutions have reached many areas in the life sciences. In bioscreening and drug development, they have become standard tools, without which today’s research and development would be unimaginable. The success of automation in these areas is due to two main factors: 1) biological processes are very simple processes under mild ambient conditions, and 2) establishment of the microtiter plate has created a standard format that has positively advanced the development of suitable systems such as liquid handlers and readers.
Increasingly, automation solutions are becoming necessary in other laboratory sectors as well. Due to continuously rising costs and increased regulations for quality control in, for example, the pharmaceutical and food industries, a larger number of samples must be processed. Simple use of the methods and technologies known from the field of bioautomation is usually not possible.
Standardization of vessels has not been successful in classical laboratory analysis, in which a variety of vessels with different shapes and volumes is used. Analytical measurement technologies usually also require preanalytical sample preparation that goes far beyond simple liquid handling. Often, derivatizations, extractions, and purifications are used to convert the samples into a measurable form, which enables selective and sensitive determination of the desired analytes. Due to the often strong inhomogeneities of the sample material as well as the extremely low concentration in the ppb and ppt range, larger amounts of sample are required to ensure representative analysis results.
The different formats and larger sample volumes necessitate development of new automation solutions in this area. For general application in analytical laboratories, they must overcome the status of proprietary solutions and become standard systems.
Automated filtration
Environmental, medical, or quality control samples are often inhomogeneous. Therefore, prior to a workup and analytical measurement of the relevant analytes, separation of solid and liquid components is often required. In addition to centrifugation, filtration is a standard technology. For small amounts of sample, automated methods are now available, such as filter tips or microtiter plates with integrated filters. Use of these methods for larger volumes (>10 mL) is not meaningful and would be very time-consuming due to multiple process steps. For the processing of larger sample volumes, therefore, a system is required that allows automated pouring of the samples from a starting vessel into a target vessel with simultaneous filtration of the samples.
The proprietary system described here includes pneumatic elements, servo drives, and sensor components. Three filter positions on the system take up the target vessels and the required filters. Figure 1 shows a 3-D CAD representation of the system.
Figure 1 – CAD of automated filtration system.
To enable a programmable decanting process, the actuator is implemented as a direct rotating container that holds the sample container. After inserting the sample vessel, a pneumatic two-finger parallel gripper is closed to fix the starting vessel during the pouring process. Decantation is carried out by tilting the sample vessel by rotating the entire sample vessel holding unit. The implemented servo drive allows the targeted control of acceleration, deceleration, start, end, and stop speed as well as rotation angle. For feeding the samples to the filter positions, the decanter unit of the filter station is mounted on a horizontal linear servo rail. Position control is via a closed loop with incremental encoder. A pneumatic axis raises the decanter unit together with the linear servo axis. This is necessary to pass the current filter position when accessing one of the other two positions.
Automated monitoring of filtration processes
A problem in many filtration processes is clogging of the filters. While in manual processing it is easy to react, a corresponding sensor system has to be integrated in an automated system. For this reason, a gradual decantation of the solutions is done in the filter, so that the maximum capacity of the filter is not exceeded. The maximum volumes per dosing step are dependent on the filters used in each case. Due to the additional integration of an ultrasonic sensor, a level measurement is carried out in the filters. Normally, when the entire solution has been successfully filtered into the target vessel, filtration of the solution is continued with the next dosing steps. If residual liquid in the filter is detected via ultrasonic measurement and thus blockage of the filter is identified, further filtration of the sample is suspended. The system software generates an error message which, in the case of the use of a higher-level control system, enables the entire workflow to be rescheduled and the faulty samples to be eliminated from the process. Alternatively, it is possible to continue filtration on the other two positions.
Application areas
The system developed for the filtration of larger sample quantities can be used for several applications. For this purpose, adaptation to the required starting and destination vessels and the filter to be used is needed in each case. Depending on the vessel, the algorithms of the pouring and filtration process must be modified to the respective volumes.
Environmental samples
One field of application is environmental samples. Aqueous samples can be filtered directly to remove small particles. Solid soil samples can be processed after slurrying in water or organic solvents to give a pure filtrate.
Biological samples
In addition, the system can be used for biological samples. Organic bile, for example, contains many solids that must be filtered before further processing of the samples and determination of organic constituents such as cholesterol or determination of the elemental analytical composition. Urine samples are frequently hydrolyzed; impurities and matrix components can be centrifuged or filtered.
Cell analysis
Another application is cell-containing solutions, in which either separation of the cells from the liquid phase or purification of the cells is intended. The focus is on medical-diagnostic procedures based on primary human cells derived directly from the patient.
All of the above applications require the use of different tubes and vessels with varying volumes. Thus, the developed system is flexible in handling different source and destination vessels as well as different filters (see Figure 2).
Figure 2 – Filtration station for different source and destination vessels.
Future work
The system can be operated as a standalone unit or can be combined with robotic solutions. In the most basic configuration, the latter takes place with a simple robot arm, which carries out loading of the system with the samples, target vessels, and filters to be used.
In addition, integration into complex automation systems is possible, in which the filtration system represents only one substation. For this purpose, integration of the device into a superordinate software component is required, which initiates control of the system.
In addition to expanding the above application areas, further developments will focus on two applications in particular. The individual steps of the pouring process are currently still determined empirically. Future developments will include implementation of suitable algorithms in the software, which also take into account different vessels and filling heights. The liquid level in the sample vessel should be calculated dynamically, depending on the actual tilt angle of the respective vessel. On the other hand, constructive adaptation of the system is required, making it possible to use different vessels and vessel sizes in as simple a way as possible.
Prof. Dr. Kerstin Thurow is with the Center for Life Science Automation, University of Rostock, F.-Barnewitz-Str. 8, 18119 Rostock, Germany; tel.: +49 381 498 7800; e-mail: [email protected]; http://www.celisca.de. Steffen Junginger and Sebastian Neubert are with the Institute of Automation, University Rostock. The authors thank Dipl.-Ing. Lars Woinar and Heiko Engelhardt for their support in the device construction.