1
Introduction

1.1 A Short Definition of Laboratory Automation

The term "automation" first appeared in 1936. Harder described automation as "the transfer of work tasks to machines in a production process without human intervention" [1]. In 1946, he while working as Vice President founded the Automation Department of Ford Motor Company. After World War II, two books by Diebold (1926-2005) appeared in 1952, describing automation as "automatic operation or a process for the automatic production of material goods." Diebold defined two main meanings of automation. On the one hand, he defined automation as an automatic control through feedback. On the other hand, automation for him was also the integration of a different number of machines [2]. The Diepold concept was further developed by Bright, who described the various stages of mechanization and automation [3], and Drucker, who recognized automation as "a conceptual system beyond technology." These three theories form the basis for understanding the concept and importance of automation [4]. "Automation" can be seen as an abbreviation for "automation technology" or "automatic operation." Alternatively, automation is also a combination of the Greek "automotos" (means "to move yourself") and the Latin "-ion" (means "a state"). "Mechanization" is the replacement of physical labor with machines; however, machine operation is controlled by human operators. "Automation" also replaces these control measures with machines, i.e. it replaces the physical and mental activities of humans with machines.

Laboratory automation is part of automation technology and aims to develop and optimize technologies for the automation of classic laboratories. This includes a wide variety of laboratories in the fields of medical diagnostics, environmental analysis, or quality control, for example, in the pharmaceutical industry, food monitoring, or industrial production. Laboratory automation is a strongly multidisciplinary field. The main goal of automating laboratory processes has not changed since the first steps in this area and consists of increasing the number of processed samples (and thus productivity), reducing the processing times required per sample, and improving the quality of those obtained experimental data or the creation of opportunities for examinations that would not be possible without suitable laboratory automation.

Laboratory automation can today be defined as a highly complex integration of robotics, liquid handling systems, sample processing, and analyzing devices and computers for process control. The most important part of laboratory automation is laboratory robotics, which develops robots and robotic solutions adapted to the specificity of laboratory processes. Since the robots in laboratory automation systems generally only take on transport tasks, the development of suitable devices and components for the automatic execution of laboratory processes (e.g. dosing, shaking, incubating, etc.) is of immense importance. Suitable software algorithms are required to control the individual systems and to evaluate the data collected.

1.2 Short History of Laboratory Automation

The main drivers of the development of automation solutions are often the development of special branches of industry as well as new and more complex requirements for specific analytical processes. Very often, the impulses for the development of new solutions result from the end-users who are confronted with several problems and inadequacies in their everyday laboratory work. For a long time, the requirements of industrial process control drove the development of automated systems.

1.2.1 Early Developments in Laboratory Automation

The first reports on the use of automated devices can be traced back to 1875 [5]. The first steps that have been made accessible to automation seem very simple from today's perspective: washing filtration residues on filter paper or liquid extractions. In 1875, Stevens described a device that made it possible to wash filter residues with water at a controlled flow rate. The wash solution was in a closed reservoir, through which air was passed through an opening. The flow rate could be controlled by the size of the opening [6]. This concept was further developed by Mitchel [7] and Lathrop [8]. In the analysis of fertilizers, the samples were washed successively with 10?ml water each until a total volume of 2500?ml was reached in order to wash out the soluble components. For this purpose, Horne developed a device for the automatic washing of the samples [9]. The first automatic burette for laboratories with recurring titrations was described by Squibb in 1894 [10]. In the same year, Greiner presented an automatic pipette, which was used for the Babcock milk test [11]. The previous developments were not suitable for slow extractions over several hours; therefore, Hibbard developed a suitable system with which flow rates of approximately 40 drops/minute were possible. A further reduction in the dripping speed could be achieved by installing a splitter [12]. The first liquid-liquid extractors were used for botanical studies. By spraying the extraction solvent into the aqueous phase, the efficiency of the extraction could be increased considerably by increasing the surface area [13]. The first devices were developed by different scientists, who were faced with different problems in the laboratory. They were very fragile systems that could be easily broken and very difficult to clean; therefore, the solutions were proprietary and did not find widespread use.

The better understanding of combustion processes and the steadily growing production of electrical energy at the end of the nineteenth century revolutionized power generation. The development of automation was therefore decisively driven by the coal and power generation industry since at the beginning of the twentieth century, there was an increasing need for more precise knowledge of the quality of coal (calorific value). The first commercial laboratory automation device was therefore a device for grinding coal samples. The Sturtevant Automatic Coal Crasher was operated by an external motor, and it made it possible to provide representative samples [14]. Another important parameter in industrial production was the determination of carbon dioxide in flue gases for the optimization of combustion processes. A commercial system was introduced to the market by Simmance and Abady. The system could be operated unattended for longer periods of time, but only provided intermittent values. A continuous variant was proposed by Stache et al. with the development of the autolyser [15]. Taylor and Hugh developed a system for the automated determination of carbon monoxide, which was based on a change in conductivity of a solution when the gas was passed through [16]. Conductivity measurements have also been reported for the control of sulfuric acid content in papermaking. Edelmann developed a device that enabled the automatic supply of sulfuric acid based on the measured values. This had previously been done manually and therefore represented an enormous source of errors [17].

The first commercial automated laboratory devices were developed during the First World War due to an increased need for rapid gas analysis. Such systems could now be used for the detection of chemical warfare agents in armed conflicts. The first systems were based on the measurement of changes in the conductivity of a heating wire. Since there was no chromatographic separation of the components prior to the measurement, clear identification of substances was not possible. Commercial variants were sold by the Cambridge Instrument Company and others [18].

In the 1920s, new requirements came from the sugar and paper industries, where there was an increasing need for pH determinations in different production steps. An essential step is the liming of sugar cane juice to remove non-sugars, for which an automated system was first developed in 1928 [19]. This system marked the beginning of the era of the development of electrodes for pH control. The electrodes available at the time required too long equilibration times, were too complicated for use in an industrial environment and were too susceptible to poisoning from sulfur dioxide, which was used in the process. Balch and Kane used tungsten-calomel electrodes for their developments, but it turned out that these exhibited variabilities in calibration, did not last long, and were also susceptible to poisoning [20].

In 1929, the first automated titration systems were introduced, which used a photocell to detect the color change in the solution. After the color change was detected, a valve was automatically closed so that no further titrant was dosed. The authors reported that "the device was 165 times more sensitive than the human eye" [21]. Hickman and Sanford developed a much more sophisticated titration device at Eastman Kodak. The device had an option to empty the previous sample to avoid contamination. In addition, the indicator was automatically supplied [22].

With the beginning of World War II, there was a further boost in the development of automation solutions in process control. This resulted from increased demands on the production of war-relevant goods and a lack of qualified workers....