Robots automate lab analysis.
Why employ robots in the laboratory? Like their larger counterparts that weld, paint, and assemble products on the factory floor, laboratory robots perform repetitive tasks accurately, produce results quickly, isolate workers from hazards, and document the process and the results. An analytical process consists of preparing the sample, taking a measurement, recording the result, and analyzing the data. Measurement devices with on-board microprocessors, such as tensile testers, spectrophotometers, chromatographs, and titrators, operate without user intervention. But an analyst still must tediously prepare the samples. Manual preparation of samples is the biggest bottleneck in the analytical process. Laboratory analysts are beginning to delegate this and similar routine jobs to robots, which automatically process samples and present them to other instruments for analysis.
Hundreds of different applications-from routine quality control tests to process-control procedures and specialized research techniques--have been automated using laboratory robotics systems. Several applications are described here.
The demand for accurate physical and tensile testing continues to increase in quality control departments, which must quickly test and accept product shipments and incoming materials. Manufacturing requires laboratories to prove the validity of the quality control data. This proof requires extensive testing and documenting of replicates and standards. Repetitive physical test procedures, such as density determination and tensile testing, can be automated.
Tensile Testing. To develop new products, engineers at Goodyear Tire and Rubber Co. (Akron, Ohio) searching for new rubber compounds and material scientists at E.I. du Pont de Nemours & Co. (Wilmington, Del.) experimenting with new polymers run basic tensile and flexural tests. A Zymate laboratory robot from Zymark Corp. works with the Universal Test System from Instron Corp. (Canton, Mass.) to perform consistent and accurate automated tensile testing of rubber, plastic, and metal specimens. The test system incorporates different size test frames consisting of a strain gage load cell and multirange amplifier. Gripper jaws can be interchanged for tension, compression, and flexural testing. Recorded test results may be passed to a computer via an IEEE-488 interface or to strip-chart records and x-y plotters.
To use the Instron system, operators stack specimens in sample racks and place bar-coded batch separators, of a thickness different from the specimens, between each batch of samples. The robot, with its tactile-sensing fingertips, detects the batch separators and brings them to the bar-code reader. There, a laser scans the label for batch identification and batch test parameters for the specimens that follow.
To calculate strength and modulus values, the robot takes a sample from the rack and puts it into the micrometer station that measures the specimen's thickness. After recording the measurements, the robot places the specimen between the upper and lower jaws of the Instron test machine and the controller starts the machine automatically. A PC controls the test machine and records test results, including peak stress, peak strain, breaking stress and strain, and modulus of elasticity. When the test is complete, the robot removes the broken specimen from the holder and discards it in a waste container.
Laboratory robots also work with materials that are more flexible than metals and plastics. Proctor & Gamble Co. (Cincinnati) has automated its physical tests to determine tensile strength, burst strength, and absorbency for paper towels, toilet tissue, and facial tissue. To increase the sample capacity beyond that of a conventional flat rack, the robot uses carousel racks that rotate 90 degrees and hold four times the number of samples. Proctor & Gamble developed special fingers with long rods to manipulate their paper samples.
Density Determination. At its Free-port, Tex., plant, Dow Chemical Co. is using laboratory robots to automate the process of determining polymer density. (Measurements are made according to American Society of Testing and Materials Method D-792, Density of Plastic by Displacement.) Archimedes' principle allows a material's density to be determined by comparing its specific weight in air and its specific weight when submerged in a fluid of known density. In the automated procedure, a robot obtains a polymer plaque sample from the storage rack and places it on a holder in the balance; the robot's controller records the dry sample weight. The robot then transfers the sample to a second holder, which is submerged in a heptane solution. The controller records the weight of the submerged sample and the temperature of the heptane and then calculates the sample's density. Finally, the robot retrieves the polymer plaque from the balance holder and returns it to the storage rack.
In one year, Los Alamos National Laboratory (Los Alamos, N.M.) analyzes approximately 6000 samples consisting of plutonium, uranium, and other materials that are handled in a glovebox environment. Because they are concerned with possible exposure of laboratory personnel to radiation, they use a laboratory-automation system consisting of a Zymate robot and their own custom-designed workstations to analyze material samples for the presence of radioactive compounds. The robot dissolves metal samples in acid, extracts a portion of the solution into a small glass container, generates a bar-code label containing sample information, and caps the container. To increase reliability, a control program continually displays the status of the system.
Another application involving metal samples is elemental analysis. This is typically performed at companies that use photographic chemicals containing silver or other substances containing precious metals. Here, the objective is recycling. The analysis is performed to determine the precious metal content of the scrap material before it is sent to the recycler.
In precious metal recycling, technicians commonly use strong acids to digest materials to measure metal concentrations. To automate this procedure, the technician places a small amount of sample into a container. The robot may also pour the sample into a container that is placed on a balance to obtain a specific weight. The robot adds corrosive acids to dissolve the sample and places the container in a hot block or microwave oven, where the sample is digested and evaporated. After the solution is diluted with water, the robot returns the container to a rack for analysis by inductively coupled plasma spectroscopy or atomic absorption spectroscopy.
Environmental laboratories must manage increased workloads and assure high-quality results for the Environmental Protection Agency, while maintaining qualified personnel. The EPA develops methods to measure contamination levels in industrial and municipal treatment plants. It sets standards on acceptance levels of pesticides and polychlorinated biphenyl in water, soil, and milk. To analyze water, wastewater, soil, and sludge for metals, technicians add highly corrosive acids to digest the samples. Robots reduce the analyst's exposure to these hazardous materials and relieve the analyst of time-consuming repetitive procedures.
Many industrial and municipal treatment plants measure contamination from organic and some oxidizable inorganic compounds using the biological oxygen demand (BOD) method. The five-day BOD dilution method, which is accepted by the EPA, has become critical for both regulatory compliance and internal monitoring of water quality. Although BOD testing is straightforward, the analyst must perform a series of tedious dilution steps. Robots are meeting the increased requirements for more frequent sampling and reproducible results.
To automate the BOD process, Shell Oil Co. (Wood River, Ill.) uses a laboratory robot to transfer a precise amount of sample to a BOD bottle. Through peristaltic pumps, a dispenser adds dilution water with a known amount of bacteria that starts the reaction with the sample. The controller takes an initial dissolved oxygen reading with an oxygen probe and stores the value in memory while the robot caps the BOD bottle and creates a water seal. After five days of incubation, the robot uncaps the bottle and the controller takes the final dissolved oxygen reading, calculates the BOD values, and stores or transfers data to a laboratory computer system. The robotic system also includes a workstation to wash and rinse the BOD bottles after each use. In automating the BOD test, laboratories eliminate operator-to-operator variability, reduce operator involvement, and accurately control all process timing steps.
Robots can analyze drinking-water samples by extracting a small amount of water mixed with hexane and directly injecting the sample into a gas chromatograph (GC), which detects the presence of individual chemical components. The robot retrieves a 40-milliliter container from the input rack, uncaps the vial, adds 6 grams of sodium chloride, recaps the vial, shakes the sample, and adds 2 milliliters of hexane. Since, the robot monitors the container's height and volume, it can extract 0.5 milliliters of the hexane layer phase. After the robot dispenses the hexane layer into a GC vial, it crimp-caps the vial and returns it to a storage rack from which it can be manually placed into a GC or, for complete automation, the robot can change to a hand specifically designed to load the GC autoinjector. The robot prepares a sample every 12 minutes, a cycle time that keeps up with the GC run times. The controller keeps track of the data during all preparation steps.
The Zymate System
The Zymate system consists of a robot, workstations, and a controller, all designed specifically for the laboratory. With an industrial robot, the articulated arm does all the work. However, a laboratory robot moves the sample from workstation to workstation to perform the actual work. Laboratory automation not only includes mechanically handling and preparing samples, but also data acquisition, testing the data against predetermined quality criteria, and documenting steps throughout the sample preparation and analysis process.
Robots come in all shapes and sizes. The type of arm--cylindrical, articulated, cartesian, or polar--determines the shape of the work area. The Zymate robot uses a cylindrical work envelope to maximize the effective area around the robot and to make the most common motions match the natural axes of the robot. Consequently, the robot accesses multiple lab stations quickly, pours from one container into another, and lifts test tubes and containers from racks in a vertical motion. To ensure accurate operation, six servo-driven axes control the robot as follows: base rotation, 360 degrees plus overlap; vertical motion, 22 inches; reach motion, 26 inches; wrist rotation, 360 degrees; and two robot hand controls for grippers, syringes, electromagnets, or solenoids.
The Zymate robot and all laboratory workstations such as balance interfaces, cappers, centrifuges, and dispensers are designed for automated operation. A user connects the robot and its workstations to the Zymate controller, which configures itself based on the workstations present. As lab workstations change or new modules are added, they use the same controller. Several robots and workstations performing the same function can be used with one controller to increase sample throughput.
As the controller manages the operations of the robot and workstations, it acquires data and prints results. For more sophisticated reporting, a lab can transfer the data to a PC as input to programs like Lotus 1-2-3, or to a larger computer as input to a statistical software package.
The Next Move
Competition in the 1990s will demand that manufacturers develop more quality products, in greater quantity, at a faster rate, and in safer environments--with fewer qualified personnel. To meet these demands, robotic applications will increase. According to Glen Taylor, director of analytical chemistry R&D at Shell Development in Houston, "The question is no longer whether a robot can work, but rather when a robot should be used."
The laboratory robot system will change in the 1990s. No longer will the robot take the lead role in sample preparation; rather, robots will join a chorus of analytical instruments. The entire analytical process--from preparing a sample to taking a measurement, recording the result, and analyzing the data--will take shape as one integrated system. Interfaces between instruments will look more generic and require less user programming. In the same laboratory, the analyst will work together with robots, analytical devices, PCs, and mainframe computers that operate large data bases.
As robots and analytical devices become more integrated, they will also become simpler. Laboratories will replace specially shaped containers, such as square bottles, with more robot-friendly containers to gain faster implementation at a lower cost.
As an early example of an integrated robotic system, Hershey Foods (Hershey, Pa.) uses a robot workstation with a 90-degree work envelope and standard 16- by 100-millimeter test tubes for sugar analysis. This workstation contains a robotic arm to prepare a sugar sample and a high-pressure liquid chromatograph for analysis.
Researchers will also use laboratory robots to develop very complex systems. Scientists at Eli Lilly (Indianapolis) currently use a Microbot II robot on a gantry to transfer biological colonies around a lab to a vision system that consists of two Hitachi cameras. Based on the colony growth pictures, a computer decides what sample to analyze further and a Puma robot transfers the samples to the next workstation for analysis. Finally, a Zymate robot prepares samples for high-pressure liquid chromatograph injection. As data pass from the workstations to the mainframe computers, an artificial intelligence system controls the whole process.
Multiple robot systems require extensive programming; however, they will become increasingly necessary as environments become more hazardous to work in and as greater precision and accuracy are required. The future of laboratory robots will also depend on the daily routines of the people who will work with them. As current applications expand and new ones develop, the capabilities of laboratory robots and robotic workstations will continue to grow.