Theory and Practical Applications of Automation Solutions in Analytical Measurements

Automation is ubiquitous in daily life, and systematic classification depends strongly on the application area and process structure. In contrast to highly automated life science laboratories, classic analytical labs still show a relatively low degree of automation. However, increasing sample numbers and escalating costs have resulted in a demand for automation focused on high flexibility and multiple applications for these specialized systems.

Beginning of automation

Beside myths and legends, automated machines originated in ancient Greece. The developers of the first automats investigated physics and tried to create a copy of nature using technical tools. While a number of mythological creatures and attendants can be found in Greek mythology, usefulness was not a priority.

The first real automatic devices are known from the era of the Alexandrian Schoolwith excellent natural philosophers such as Heron of Alexandria, Pythagoras, Euclid, and Archimedes.¹ The Alexandrian researchers combined simple tools such as screws, wedges, and levers for the execution of complex movements, using water, vacuum, or air pressure as the driving forces. Programmed simulators and automated devices as well as tools for feedback were developed, which were close to today’s understanding of automation.

Laboratory and life science automation

To facilitate automated processes in the chemical, biological, pharmaceutical, food, and medical sectors, knowledge of natural and laboratory sciences combined with process engineering was essential. Automated devices for scientific investigation have been reported since 1875. The first of these were specialized solutions built by scientists to fulfill particular needs in their laboratories.² An automatic zero burette and an automated pipette were described in 1894, and the first automated titrator was developed in 1929.^2-4 Companies started providing automated equipment after the Second World War, since the use of automatic control devices had become routine in chemical laboratories. An analog computer was the first computer applied in laboratory automation, allowing chemical researchers to generate electronic simulations of their processes.² The first use of a digital computer in combination with a mass spectrometer for the determination of hydrocarbon mixtures was reported in 1952.⁵

The first real automated systems appeared in medical laboratories in the mid-1950s. An example of this is the AutoAnalyzer⁶ for the determination of urea, sugar, and calcium in blood.⁷ Later designs offered the possibility of simultaneous determination of over 20 analytes with 150 samples per hour. Other batch analyzers were developed that could test up to 100 samples in continuous mode. The introduction of the photodiode array for spectrometers in the early 1980s allowed the simultaneous detection of multiple analytes using different wavelengths.⁸ Another approach appeared in 1959 with the Robot Chemist (Research Specialty Company) to automate all the manual steps of automatic pipetting and mixing using conventional cuvettes, but was too complex to be practical.^9,10 The introduction of robotics and informatics to clinical laboratory automation in the 1970s led to the development of total laboratory automation.

Developments in the automation of life science processes have mainly been driven by the requirements of the medical and pharmaceutical industries. Within the last 20–30 years, the number of samples to be screened increased significantly. Until the end of the 1980s, approximately 10,000 compounds have been tested per year and were the object of investigation (target). In the early 1990s, sample throughput increased up to 10,000 samples per month and target and, only five years later, grew to 10,000 samples per week.¹¹ A notable enhancement was ultrahigh-throughput screening, which enables the processing of more than 100,000 samples per day. Conventional transport devices such as turntables, conveyors, or single-arm articulated robots are integrated into automation cells for labware transport and positioning. Automated liquid-dispensing processes are typically carried out using Cartesian robots with xyz drives.¹² Robotic arms with multiple degrees of freedom act as system integrators to connect the individual stations on the automation system.^13,14

Need for automation in analytical measurement processes

Today, there is an increasing demand for flexible and universally applicable automation solutions in the field of analytical measurements, with the aim to determine qualitative and quantitative information related to elemental composition and/or obtain structurally relevant information. Up to now, this area has been dominated by automated workstations, partially automated systems, or proprietary fully automated systems for specific applications. In contrast to high-throughput screening systems, there is still a need in suitable flexible automation solutions for analytical measurement applications due to complex structured and frequently changing processes with a multitude of subprocesses. Single vessels with various shapes and volumes are used; there is no standardization compared to the well-known microplate footprint. Furthermore, sample preparation processes in some cases require environmental conditions such as high pressures and temperatures for derivatization procedures, organic solvents instead of aqueous liquids, and special subprocesses such as filtration or solid-phase extraction, to name a few.

Automation concepts for analytical measurements

The development of automation systems involving sample preparation, analytical measurement instruments, and complex sensor systems requires systematic analysis of the processes to be automated with the goal toward producing suitable structures and allocate them to these processes. In industrial applications, eight different structures can be distinguished according to their centralized or decentralized process, local, and functional structure. In analytical measurement technology, there are limitations, and adaptation is thus required. Analytical processes are always characterized by a decentralized process structure. This enables a distinction according to their local and applicative structure. Depending on the robotic technology used, two basic automation concepts can be applied to processes in analytical measurements: central system integrators and flexible robots.

Due to their high flexibility, robots act as transport systems to connect individual subprocesses and workstations, whereby the robot functions as a central system integrator.^15,16  Figures 1 and 2 show an automation system with 3-D CAD construction and the resulting system based on this concept. Higher flexibility can be achieved when the robot performs manipulation tasks as well. The robot then acts as a flexible device. Dual-arm robots are particularly well-suited for this automation concept.^17,18

Figure 1 – 3-D CAD design of an automation system with a central system integrator. 1) Two ORCA laboratory robots (Beckman Coulter) on orthogonal linear rails acting as central system integrator, 2) storage system, 3) capping and weighing station, 4) laboratory equipment, 5) single-vial liquid handler, 6) Biomek 2000 liquid handler (Beckman Coulter), 7) positive-pressure automated labware positioner (ALP), and 8) sample transfer station.

Figure 2 – Resulting automation system with central system integrator.

Figures 3 and 4 show an automated system based on this concept (3-D CAD construction and resulting system). Flexibility of the automated system can be further increased by using mobile robots, which perform both transport between various subsystems and manipulation. This concept is called integrated robotics.

Figure 3 – 3-D CAD design of an automation system with a flexible robot. 1) Dual-arm robot CSDA10F (Yaskawa), 2) manual pipettes, 3) electronic pipettes, 4) storage system, 5) GC/MS system, 6) LC/MS system, 7) workbench with laboratory equipment and sample transfer station, and 8) safety system with light curtain and door protection.

Figure 4 – Resulting automation system with flexible robot (dual-arm robot).

Summary

Automated systems for sample preparation and analytical measurements must follow special protocols due to their complex process structure with multiple subprocesses. In contrast to classical high-throughput screening systems, standardized labware such as microplates often cannot be integrated into the automation procedure. Process complexity and special labware such as single vessels require new concepts and approaches.

Figure 5 – Automation Solutions for Analytical Measurements—Concepts and Applications.

Automation Solutions for Analytical Measurements—Concepts and Applications is the first book dedicated specifically to automated sample preparation and analytical measurements (Figure 5).¹⁹ It provides a current systematic overview to biological, medical, and environmental applications; drug discovery; and quality assurance. A critical review of realized automation solutions discusses special requirements in analytical applications. Theoretical basics and practical examples illustrate the goal and limitations of each automation concept.

References

1. Watts, E.J. City and School in Late Antique Athens and Alexandria; Berkeley, CA: University of California Press, 2006.
2. Olsen, K. JALA 2012, 17 (6), 469–80.
3. Squibb, E.R. J. Am. Chem. Soc. 1894, 16(3), 145–8.
4. Greiner, E.A. Am. Chem. Soc. 1894, 16(9), 643–4.
5. Olsen, K. Chem. Eng. News 1952, 30(11), 1092–4.
6. Olsen, K. JALA 2012, 17(6), 469–80.
7. Plumb, R.K. New York Times 1956, 16.11, E13.
8. Durner, J. Angew. Chem. Int. Ed. 2010, 49(6), 1026–51.
9. Rosenfeld, L. Clin. Chem. 2000, 46(10), 1705–14.
10. Armbruster, D.A.; Overcash, D.R. et al. Clin. Biochem. Rev. 2014, 35(3), 143–53.
11. Major, J. J. Biomol. Screening, 1998, 3(1), 13–7.
12. Kong, F.; Yuan, L. et al. JALA 2012, 17(3), 169–85.
13. Bogue, R. Industrial Robot 2012, 39(2), 113–9.
14. Fleischer, H. and Thurow, K. Automation Solutions for Analytical Measurements—Concepts and Applications. Weinheim: Wiley-VCH, 2017.
15. Vorberg, E.; Fleischer, H. et al. JALA 2016, 21(5), 682–92.
16. Fleischer, H.; Ramani, K. et al. SLAS Technol. 2018, 23(1), 83–96.
17. Fleischer, H.; Drews, R.R. et al. JALA 2016, 21(5), 671–81.
18. Fleischer, H.; Baumann, D. et al. Energies 2018, 11(10), 1–21; article no. 2567.
19. Fleischer, H. and Thurow, K. Automation Solutions for Analytical Measurements—Concepts and Applications, 2017. Copyright Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with permission.

Heidi Fleischer, Ph.D., is university lecturer at the University of Rostock, Institute of Automation, Richard-Wagner-Strasse 31, 18119 Rostock, Germany; tel.: +49 (0) 381 498 7803; fax: +49 (0) 381 498 7802; e-mail: [email protected]; www.uni-rostock.de/en/. Kerstin Thurow, Ph.D., is CEO of the Center for Life Science Automation—celisca at the University of Rostock. The authors wish to thank Yaskawa Europe GmbH and Dr.-Ing. Michael Klos for providing the SDA10F two-arm robot and support. They also thank the Integrated Systems research group headed by Dr.-Ing. Steffen Junginger, the IT Systems @ Automations research group under the direction of Dr.-Ing. Sebastian Neubert, and the Processes & Measurement research group led by PD Dr.-Ing. habil. Heidi Fleischer. Further thanks go to Prof. Dr.-Ing. habil. Hui Liu for his support in the field of mobile robotics. The authors would also like to thank all the students and interns for their contributions as part of their bachelor and master theses.