Introduction

Analytical processes can be automated in different ways. In the simplest case, the desired parameters of a sample can be determined by direct measurement. With these inline measurement methods, the substances of interest are determined using spectroscopic methods or the use of electrochemical sensors, among other things. The measuring system is integrated into the production line, and the measurement takes place in real time.

If a direct measurement is not possible, e.g., due to the measuring method used, a sample must be taken before the actual metrological determination. The qualitative determination and quantification of the samples of interest thus take place outside the actual production process. If the measuring system is connected directly to the process via a sampling device, this is referred to as an online measuring method. Online measurement technology enables a quasi-real-time determination of the parameters of interest.

Both methods are used in process analytical technology (PAT) and are characterized by a high degree of automation.

The PAT measurement methods are often inadequate for more precise determinations of individual parameters. More complex methods are used here, which often require extensive processing of the samples before the actual metrological determination. Offline measurement technology is performed outside the process and is used to analyze the end product or the process data after production. Offline measurement technology can be more accurate and specific as the samples are analyzed under controlled conditions.

The division into inline/online methods comes from the field of PAT and is not readily applicable to many application areas of (bio)analytical measurement technology. In contrast to classic PAT, there are numerous applications that require the processing of individual samples. This applies, among other things, to medical applications in which patient-specific samples have to be examined for certain components. But also in the field of environmental monitoring or in the quality control of food, individual samples are taken at different control points and have to be examined for ingredients. Here, we are always in the area of offline analysis, unless specific biosensors are used, the use of which is limited to a few applications. Offline analysis, especially in the areas of sample preparation, has so far only been slightly automated and requires appropriate automation concepts.

Automation strategies: overview

Different parts of an automation system are required for the automation of (bio)analytical processes. In principle, a distinction can be made between data generation and data handling. In addition to the actual measuring system, the area of data generation contains all other subsystems that are required to finally generate analytical measurement data. According to the areas of sample preparation mentioned in the “Processes in (bio)analytical chemistry” section, these can be systems for handling samples (e.g., dosing) and manipulating samples (e.g., heating, shaking, incubating, grinding, purification). The scope of the required subsystems depends on the respective process. In addition, procedures are required for complete automation that take over the transport of the samples and the labware used between the individual subsystems. Conveyor belts or robots are used for this.

The second area to be automated is data handling. First, the actual processing of the measurement data collected to generate secondary data, analysis reports, etc. should be mentioned here. However, the connection of the automation systems to higher-level systems such as LIMS, ERP, or corporate-specific in-house workflow systems must also be considered. Particular attention is paid to the agreement of interfaces for data exchange and data formats between different software solutions.

Different strategies are conceivable for the automation of analytical and bioanalytical processes. Liquid handler-based systems represent the simplest form. They enable highly parallel processing of samples but are usually limited to pure liquid handling processes. The integration of peripheral devices is only possible to a limited extent. Fully automated systems based on a central robot can be used for analytical methods with extensive sub-process steps. The central robot can either only be used as a transport instance for the transport of samples and labware between the integrated substations or it can also handle the manipulation of samples itself (see Fig. 1).

Fig. 1

General strategies for automating analytical and bioanalytical laboratories

Both approaches can be designed as closed or open systems [17]. Closed systems are designed and optimized for a specific application. As a result, they can work highly efficiently, achieve high-throughput, and are quite inexpensive. Process changes or even the establishment of new processes is associated with a great deal of effort on such systems; the integration of additional or other devices is usually not possible. Open systems, on the other hand, offer greater flexibility. Depending on their equipment, they are designed more for a specific process group. The change of processes as well as the establishment of new processes are possible. New devices can be integrated into the systems and thus increase the range of functions and the type and number of processes to be automated. However, this flexibility is accompanied by significantly higher costs. In addition, the achievable throughput is usually lower than with optimized proprietary systems.

Recent developments in robotics increasingly enable the use of distributed system strategies. Here, processes are no longer processed centrally in one place, but various process-relevant substations are provided with robotic components.

Liquid handler-based systems

Liquid handler-based systems represent the simplest form of automation. They are based on Cartesian robots and are primarily designed for dosing liquids. Automated liquid handling systems can have a different number of channels. Depending on the equipment, three main variants can be distinguished—single-channel systems, systems with 1–8 channels and highly parallel systems with more than 8 and up to 1,536 channels (see Fig. 2). As the use of micrtiter plates increased, automated liquid handlers were developed with an appropriate number of channels for the most common plate formats 96, 384, and 1.536. The channels in the pipette heads have a permanent spacing that corresponds to the spacing of the wells in the microtiter plates. The systems thus enable the parallel processing of numerous samples. Liquid handlers are optimized for handling samples in microtiter plates and are therefore particularly suitable for processing samples that can be handled in MTP. Individual samples would therefore have to be reformatted accordingly before processing. Alternatively, the labware can be adapted to handle individual samples. The SBS external dimensions should be retained to enable easy processing on the liquid handler. If individual vessels are used that do not correspond to the standard format of MTPs (96 samples in a grid of 8 × 12), multichannel pipettors can hardly be used. The so-called Span-8 functionality is ideal here, with which up to 8 channels can be used individually and at freely definable distances from one another for the dosing processes.

Fig. 2

Complex robotic system for cell-based medical diagnostics

Liquid reservoirs are to be provided on the deck of the liquid handler for the dosing processes. Usually, reservoirs in MTP format are used here, which, depending on the pipetting head, can be used as full reservoirs (for 96 and 384 heads, one solvent possible) or as half or quarter reservoirs (for 8-channel systems, 2–4 different solutions possible). The use of self-filling reservoirs limits the number of positions that must be provided on the liquid handler deck for the provision of the solvent.

Liquid handlers can be equipped with additional devices that extend the functionality of the systems. Usually, these are shakers, heaters, and coolers but also systems for automated sample purification such as solid-phase extractions. Centrifuges, incubators, analytical measurement systems, etc. can also be integrated, provided there is sufficient space on the liquid handler deck and the devices to be integrated can be equipped by the liquid handler. In addition to the pipetting head, liquid handlers with such an extended range of functions then have a second or even a third arm with a special gripper, which transports samples and labware to and between the different substations.

Liquid handler-based systems have been described for the fully automated determination of vitamin D in blood samples, among other things [18]. The heart of the system is a liquid handler with 8 parallel, freely configurable channels for dosing the liquids and a second arm that is used with a gripper to transport the labware. The system enables the proteins to be separated using an integrated centrifuge. The purification is carried out by an integrated fully automatic solid-phase extraction. Furthermore, shakers and incubators as well as refillable reservoirs and special racks for the provision of the original samples in Eppendorf vials are integrated on the deck. The overall system enables the processing of up to 96 samples. A further increase in the number of samples (up to 288) can be achieved by using special phospholipid removal cartridges, since the centrifugation of the samples can be omitted. The system can also be used flexibly for other bioanalytical methods such as the determination of THC and its derivatives in serum, saliva, and urine [19] or the determination of benzodiazepines. Systems with automated solid-phase extraction have also been used to determine diuretics in doping control [20] or to detect beta blockers in blood [21].

Systems with central robot as transport instance

Another variant of automation is complex systems, the center of which is a central robot. The devices and systems required for processing the sub-steps are arranged in the work area of the central robot, which transports samples and labware between the different stations. Liquid handlers are usually used to implement the dosing processes, or in the case of larger volumes to be dosed, dosing pumps are used. The number of peripheral devices to be integrated depends on the range of the central robot. The peripheral devices should have suitable interfaces for system internal communication and allow a robot to access the device. For example, centrifuges should be equipped with a lid that opens automatically for the introduction and removal of samples. Autosamplers of connected devices should be designed in such a way that the gripper of a robot can also feed and remove samples. If this is not the case, adjustments to the hardware may be necessary. Depending on the design of the gripper, different formats can be handled—from microtiter plates to individual sample vessels of different sizes. If different formats are to be processed within a method, it may be necessary to change the gripper. Alternatively, universal grippers can be used that cover a large range of labware dimensions. The labware can, but does not have to, be executed in SBS-MTP format. Systems of this type allow single sample handling and are therefore suitable for processes that cannot be reformatted to the MTP format. Depending on the process, the development of special racks or even additional systems may be necessary. The rate-limiting step in such systems is the instrument with the longest processing time of the samples.

This system concept usually requires adjustments to the existing standard operating procedures, as other devices/systems are used for individual sub-steps. For example, pipetting is done manually with classic manual single- or multichannel pipettes, which cannot usually be handled with a robot. Instead, automatic dispensers or liquid handlers are used. Complete 1:1 automation, i.e., the identical translation of a manual process to an automation system, is therefore usually not possible.

Complex, fully automatic systems have been described for different applications. Tsina described a system for an automated HPLC method to detect mycophenolic acid in human plasma [22]. In addition to a laboratory robot, the system has various stations for sample preparation, such as stations for weighing, diluting, dispensing, and pipetting. Furthermore, two online HPLC systems with optical detectors were integrated. A fully automatic, robot-based system for sample preparation and analysis was also established for routine measurements in the quality control of active ingredients and pharmaceutical end products [23]. In addition to several robotic components for sample transport, the system also has stations for homogenization, temperature control, dispensing, and pH measurement. Another system with a central robot was developed for cell-based medical diagnostics, which enables the fully automatic processing of sputum samples for subsequent examination using cell CT (see Fig. 2). The system has an integrated centrifuge, a liquid handling system for the realization of the dosing steps in the µL to mL range, specially developed cappers, adapted vortexers, and a filtration unit [24]. In the process, the height of the resulting cell pellets must be detected several times, which is realized using image processing methods [25].

Systems with central robot as transport and manipulation instance

An extension of the concept mentioned above are systems with central robots which, in addition to transporting samples and labware, can also take on direct manipulation steps. If dual-arm robots are used here, a process analogous to manual processes is possible. Manually used devices such as pipettes or syringes can be integrated into the systems. This means that actual 1:1 automation is possible; no changes to existing standard operating procedures are required. The speed of the overall system is determined by the central robot and depends on the times required to carry out individual process steps. Automation systems with dual-arm robots were used for different applications. Chu developed a corresponding system for bioanalytical applications. The dual-arm robot used enables the complete processing of samples including pipetting steps, the opening, and closing of individual vessels, the transfer of samples to devices for sample purification (ultrasonic device, solid-phase extraction, shaker, heater, etc.), and the final positioning of the prepared samples in connected analyzers [26]. The system was used, among other things, for the determination of cholesterol in biliary endoprosthesis using gas chromatography-mass spectrometry (GC–MS) [27] and for chirality studies [28]. Dual-arm robot-based systems have also been described for automated downstream analysis of epidermal models [29] or anticancer drug compounding [30].

Distributed robotic systems

Depending on the application and level of equipment, central automation systems can become huge. The associated space requirement is sometimes not available in laboratories. In addition, devices and components are tied into complex automation systems and are not available for other processes, even if they are not required in a current process flow.

With the development of lightweight robots (also known as collaborative robots, cobots), more cost-effective variants for automation are now available. These robots are characterized by a lighter construction, low speeds, and additional sensors. As a result, there is no need to install additional safety precautions (light curtains, housings, …) and the robots can work in proximity to people and even share the same workspace [31].

The lightweight robots can be integrated into the above-mentioned central systems as central robots or used to equip individual systems and system groups. Existing device technology can be used and automated through integration with a robot. However, as with the central systems, there may be a need to adapt the devices to robotic operation. Existing spatial constellations can be used. This creates distributed automation systems (see Fig. 3). The individual stations can be combined with each other in any form in complex processes or used individually. This leads to a high degree of flexibility in the processing as well as in the utilization of the existing devices. Such a distributed system is primarily suitable for open system structures, since it offers the possibility of any expansion of the overall system. With this approach, the investment required depends on the number of devices to be equipped with a robotic component.

Fig. 3

Distributed automation system