ANALYST, JANUARY 1992, VOL. 117 3 Robotic Microwave Digestion System for Dissolution of Titanium Dioxide John D. Norris, Brian Preston and Lynn M. Ross Tioxide Group Services Ltd., Central Laboratories, Portrack Lane, Stockton-on- Tees, Cleveland TS182NQ, UK A robotic microwave digestion system for the dissolution of titanium dioxide samples prior to trace element determination is described. The system weighs out samples, adds acids, carries out microwave digestion, dilutes the solutions, transfers the solutions to beakers and cleans the digestion vessels. Sixty samples can be processed in approximately 16.5 h. The robotic microwave digestion system has operated reliably for over 6 months and the accuracy and reproducibility match those which can be achieved manually. Keywords: Robotic; automation; microwave digestion; titanium dioxide With recent advances in analytical atomic spectroscopic instrumentation and associated automated sampling systems, the dissolution procedure has become by far the most time-consuming and operator-intensive stage in the analytical process for the determination of trace elements in titanium dioxide.Microwave digestion has been shown to have considerable advantages over more traditional techniques such as fusions or acid dissolution in open or closed vessels.1 The successful use of microwave digestion for the dissolution of titanium dioxide prompted the investigation and develop- ment of the total automation of the dissolution process using a robotic microwave digestion system. In this laboratory, dissolution of samples of pigmentary or high-purity titanium dioxide can be followed by the determi- nation of any combination of up to nearly 40 trace elements. In order to obtain the maximum sensitivity, these determinations are carried out by a variety of different techniques, e.g., inductively coupled plasma atomic emission, flame atomic absorption, electrothermal atomic absorption or cold vapour mercury atomic fluorescence spectrometry.It was therefore not considered desirable to connect the proposed automated sample preparation system with a single analytical instrument. This paper describes a robotic microwave digestion system capable of dissolving up to 60 samples of titanium dioxide. The system weighs out samples, adds acids, carries out microwave digestion, dilutes the solutions, transfers the solutions to beakers and cleans the digestion vessels. Although other robotic microwave digestion systems have been reported2 that carry out some of these tasks, the system described has several novel features, particularly the handling of titanium dioxide and the washing and re-use of the digestion vessels.Experimental Reagents Analytical-reagent grade hydrofluoric acid (40% m/m) and hydrochloric acid (p = 1.18 g cm-3) were used. Apparatus The components of the system shown in Figs. 1-5 are described below. The system was engineered and supplied by Peerless Systems (Gateshead, Tyne and Wear, UK). Control system Computer control is effected via an electronics interface unit. The input-output (I/O) cards in the computer interface at transistor-transistor logic (TTL) level to the electronics I/O rack mounted in the system bench.This rack converts I/O signals from TTL level to the appropriate levels for the sensors, motors, solenoids and pneumatic devices, in the bench-mounted peripherals. The two balances, the robot and the microwave oven are controlled via a single serial port on the computer, which is achieved by multiplexing the various RS232 devices. The RS232 interface allows the system to tare and read the balances, and allows complete control over the microwave oven cycle. Infrared diffuse proximity sensors are used to detect the presence of a digestion vessel, and safety interlocks prevent acid from being dispensed if the digestion vessel is not present. Computer The computer is a Viglen Vig 1 Plus, with an Epson LX 400 printer, fully IBM AT compatible, 10 MHz 80286 processor, 640K RAM, 30 Mbyte hard disk and monochrome monitor.Software The control software is written in MODULA-2 and consists of a number of discrete modules for data handling, hardware control, error handling, screen formats and process control. 1 I I H I 0 Robot The Peerless Systems robot is a cylindrical motion, track- mounted unit having five axes: waist rotate, 320"; extend, 500 mm; lift (z-axis), 345 mm; track (y-axis), 2 m; wrist, 180"; gripper, pneumatic (with sensors); positional accuracy, -to. 1 mm; and lift capacity, 2 kg. Fig. 1 Layout diagram of robotic microwave digestion system. A, Acid and water dispenser; B, sample bottle rack; C, Mettler balance; D, powder pipette standby and sample bottle holder; E, scraper; F, tip waste bin; G, tip holder; H.robot; I. robot track; J. microwave oven; K, digestive vessel storage rack; L, washing and drying station; M, beaker rack; N, cooling station; P, mixer and extension arm standby; R, capping station; S, torquing station; T. digestion vessel standby; U, watch-glass removal device; and V, watch-glass standby4 ANALYST, JANUARY 1992, VOL. 117 Fig. 2 Sampling system Fig. 3 Acid and water dispenser The multi-tasking facility of MODULA-2 allows the monitor- ing and control of a number of peripheral devices simul- taneously. An error process continuously monitors the system and flags erroneous conditions. The menu-driven software allows control of the system parameters, ie., the microwave digestion programme, acid volume and types, target sample weight, final volume of solution, number of samples and sample identification.Data collected during the run can be displayed on-screen, printed out and saved on disk. The software also provides a comprehensive facility for verifying the correct operation of all of the system peripherals. Sample rack A 60-position rack, constructed from Perspex, to accommo- date standard 2 02, 60 ml glass powder bottles is used. Fig. 4 cooling station Microwave oven, torquing station, capping station, mixer and Fig. 5 Washing and drying station Sampling system Titanium dioxide is a difficult material to handle automatically owing to its adherent properties, which effectively eliminated consideration of many commonly used sampling devices such as vibratory feeders because of cleaning difficulties.In order to avoid contamination problems, it was concluded that any part of the sampling device coming into contact with the sample would need to be disposable. A further disadvantage of the material to be handled is that the bulk density of different types of titanium dioxide can vary considerably. A Peerless Systems powder pipette is employed for sample transfer. It operates on the principle of sample being withdrawn into an evacuated tube and subsequently expelled under positive pressure. Disposable tips, fitted with 10 mm filters, were designed and manufactured for use with the powder pipette. The tip capacity was designed so that it would accommodate about 0.2 g of sample.Hence, depending on the bulk density of the powder, two or three portions are necessary to achieve the target sample mass of 0.5 _+ 0.1 g . The sampling system incorporates the powder pipette, a standby holder with an optical sensor, a Mettler AT400 balance, a sampling bottle position with an optical sensor, a 90-position tip rack, an optical sensor for tip location (ie., to check that the tip has been correctly attached to or removed from the powder pipette), a scraper with bin to remove excess of sample from the tips and a tip extractor with bin for used tips.ANALYST, JANUARY 1992, VOL. 117 5 Acid and water dispenser In view of the use of hydrofluoric acid in the system, it was decided that this (and all other liquids) should not be transferred under pressure.The dispensing system therefore involves gravity feed via poly(tetrafluoroethy1ene) (PTFE)- lined solenoid valves. The system consists of a stand for four acid reservoirs (500 ml) and one water reservoir (3 dm3). These reservoirs are linked, using PTFE tubing to the moveable dispenser head, mounted above a Sauter RE 1622 balance. When dispensing occurs, the head is lowered over the digestion vessel, which has been placed on the balance. The dispense operation is electronically and mechanically interlocked for maximum possible safety. The power to the solenoid valves is interlocked to proximity switches that detect the presence of a digestion vessel on the balance and check that the dispense arm is in the lowered position. Only when these two conditions have been met will there be any supply voltage available to switch on the solenoid valves.Even then, no signals are given from the computer to operate the valves, unless the computer reads a digestion vessel mass on the balance within a specified tolerance. The system can be programmed to select different dis- pensers and to specify volumes taken, although for a complete run the same conditions would be employed throughout. Micro wave digestion vessels Four Milestone SV 140-10 microwave digestion vessels are necessary for the operation of the system. These 140 ml vessels are equipped with a regulating valve designed to give a maximum pressure of 10 bar ( 3 bar = 105 Pa). The vessel caps were modified (Fig. 6) by machining off 3 mm from the body of the cap.A washer was inserted to enable the screw to be reproducibily tightened to the correct position for the valve to operate. The silicon ring seal was changed before each run. Digestion vessel storage rack The rack is fitted with pressure sensors and can store four digestion vessels and four caps. Microwave oven A CEM MDS 81D microwave oven is used. The control system was modified so that all functions could be controlled by the computer. The turntable was removed and a PTFE holder constructed to locate the digestion vessel in the oven. A forked, flat-bladed extension arm was necessary to enable the robot to reach and position the digestion vessel in the holder. A standby position is provided for this extension. Capping station A Peerless Systems capping station, which can accommodate both the sample bottles and the microwave digestion vessels, is used.This includes an optical sensor and a temporary storage Fig. 6 washer; and D , silicon ring Digestion vessel cap. A. Body of cap; B, vent screw; C, position for the sample bottle lids whilst these are detached from the bottles. Torquing station A CEM MDS 810 capping station, modified to accommodate SV 140-10 digestion vessels is used. The torque is set to 9 N m, which is the manufacturer’s recommendation for these diges- tion vessels. Optical sensors are used to determine the exact position of the digestion vessels in the torquing station. Cooling station This incorporates two standby positions for the microwave digestion vessels. One position is for cooling the vessel after it has been in the oven and prior to de-capping.The vessel in this position is cooled by a fan. There is a second position for temporary storage after capping and prior to placing in the oven. Optical sensors are fitted in both positions. Mixer A Stuart Scientific Autovortex SA2 mixer is used. An optical sensor is attached to the mixer. Washing and drying station This is arranged so that the inverted digestion vessel or cap is washed internally by a jet of de-ionized water pumped from a reservoir. In a second position the vessel or cap is dried internally by a jet of air and the digestion vessel is spun at high speed for part of the drying process by a second jet. Optical sensors are fitted to both positions and the entire washing and drying station is located in a poly(propy1ene) bath.After drying, the vessels and caps are allowed to equilibrate to ambient conditions for at least 10 min on the storage rack before further use. Beaker rack A 60-position rack constructed from Perspex and contained in a tray to contain possible spillages, to accommodate 170 ml tapered polythene beakers and PTFE watch-glasses, is used. A pneumatic device to remove the watch-glasses and tempor- ary storage positions for the watch-glass and digestion vessel are incorporated. The holder for the watch-glass handling device and that for the digestion vessel are equipped with optical sensors. Lay-out The system is installed on a 3 x 2 m bench in a room with fume extraction. The computer and printer are located in an adjacent room. Operation The solid samples are presented to the system in standard bottles and the solutions are removed in standard beakers for subsequent analysis, the intermediate stages being carried out automatically.A report is produced giving the sample mass and dilution factor. The operator must ensure that certain procedures are carried out before a run can commence (i.e., the sample bottle rack is filled with the required samples, the beaker rack is filled with the corresponding number of clean beakers covered with watch-glasses, new silicon ring seals are fitted to the digestion vessel caps and the digestion vessels and caps are located on the digestion vessel storage rack, the powder pipette tip rack is filled, the relevant acid and water dispensers are filled and the balances are tared).The system is switched on. Sample identification data and operating parameters (number of samples, target sample mass, acid volumes and types, microwave digestion pro- gramme, final volume of solution) are entered. Normal operating parameters for the dissolution of titanium dioxide6 ANALYST, JANUARY 1992, VOL. 117 Table 1 Operating parameters for robotic microwave digestion of titanium dioxide Target sample mass Volume of HF 7 ml Volume of HC1 3 ml Digestion programme: 0.5 -t 0.1 g Stage 1 Stage 2 5 min at 50% power 10 min at 35% power Final solution mass 5og Table 2 Reproducibility for determination of aluminium in titanium dioxide Parameter Robotic Manual No. of replicates 30 12 Range (% Al) 0.44-0.49 0.46-0.50 Mean (Yo Al) 0.468 0.481 Standard deviation (%) 0.0127 0.0112 are given in Table 1.The robot is calibrated and a status check is run to ensure that all facilities are in position and operational. The run is commenced. The following describes the procedure undergone for the digestion of a single sample. During a run three samples, at different stages in the process, are being handled simul- taneously, and therefore the procedure as shown is not continuous for any given sample. A microwave digestion vessel is transferred from the digestion vessel storage rack to the balance; the vessel is weighed and the balance is tared. A sample bottle is transferred from the sample rack to the capping station. The bottle cap is removed and placed in the bottle cap standby position. The open sample bottle is transferred to the sampling bottle position. The powder pipette is picked up from its standby holder, a tip collected from the tip rack and the pipette passed over the optical sensor to ensure that the tip is correctly located.An aliquot of sample is removed from the sample bottle, the pipette passed over the scraper to remove any excess of sample adhering to the outside of the tip and the sample aliquot transferred to the digestion vessel. The sampling process is repeated until the target sample mass has been achieved and this mass is recorded. The tip is removed using the extractor and the pipette returned to the standby holder. The digestion vessel is transferred from the balance to the acid dispenser. The dispenser head is lowered and the acids are added until the required mass has been achieved.The dispenser head is raised and the digestion vessel is transferred to the mixer, where the contents are mixed, and then to the capping station. A digestion vessel cap is transferred from the digestion vessel storage rack and placed on the digestion vessel in the capping station. The partially capped digestion vessel is transferred to the torquing station for tightening to the correct torque and then to the standby position. The microwave oven door is opened and the extension arm is collected and used to transfer the digestion vessel into its position in the microwave oven. The oven door is closed and microwave digestion is carried out according to the pre-set programme. The sample bottle is transferred to the capping station, the bottle cap replaced and the closed bottle returned to its position in the sample rack.On completion of the digestion programme, the door is opened and the digestion vessel removed and placed in the cooling station (cooling is normally carried out for 15 min). The extension arm is returned to its standby position. The cooled digestion vessel is transferred to the torquing station and partially de-capped, and then to the capping station where the cap is removed. The cap is transferred to the washing and drying station, washed and dried and placed in its position in the digestion vessel storage rack. The open digestion vessel is transferred to the water dispenser. The dispenser head is lowered and water added until the required mass has been achieved. The dispenser head is raised and the digestion vessel is transferred to the mixer and the contents are mixed.The digestion vessel is transferred to the balance and the final solution mass recorded. The watch-glass removal device is collected and used to remove the watch-glass from the appropriate beaker. The watch-glass is placed in a standby position and the removal device returned to its holder. The digestion vessel is collected from the balance, its contents are decanted into the beaker and the empty vessel is placed in the standby position on the beaker rack. The watch-glass removal device is collected, used to replace the watch-glass and returned to its holder. The digestion vessel is transferred to the washing and drying station, washed and dried and placed in its position on the digestion vessel storage rack.At the end of the run, a printout is obtained listing the sample reference, sample mass, final mass of solution, dilution factor and the number of the digestion vessel used. Results and Discussion Microwave Digestion Conditions The microwave digestion conditions were evolved from the manual microwave digestion procedure previously employed, the main differences being that in the manual procedure a turntable is used in the oven and normally several samples are digested simultaneously, whereas in the robotic system one sample is digested in a fixed location in the oven. The conditions employed (see Table 1) were found to be effective for dissolving a wide range of titanium dioxide types without causing the digestion vessel to vent.Contamination The system was designed to minimize possible cross-sample contamination, particularly the sampling process. The main possible source for any contamination is in the washing and re-use of the digestion vessels. Checks for contamination were made over several complete 60-sample runs using alternate batches of six samples of a high-purity titanium dioxide (A1 <lo pg g-1) and of a coated titanium dioxide (A1 = 5 % ) . The solutions of all of the high-purity titanium dioxide samples were analysed for aluminium by inductively coupled plasma atomic emission spectrometry (ICP-AES) and all samples were found to contain <10 pg 8-1 of aluminium. Reproducibility and Accuracy The performance of the system was examined over a large number of sample runs.For most types of titanium dioxide the target sample mass of 0.5 k 0.1 g and a final solution mass between 50.0 and 50.4 g were achieved. In all instances complete dissolution of the sample was obtained. It is difficult to provide quantitative performance data, as robotic sample preparation is only a part of the over-all analytical process, and such data would relate to the whole process and not just of the robotic system. However, the analytical procedure for the determination of aluminium was employed to compare the performance of the robotic micro- wave digestion system with manual microwave digestion. A pigmentary titanium dioxide was selected, which, owing to its physical properties, was expected to provide sample handling difficulties. The solutions of these samples were analysed for aluminium by ICP-AES. The results are given in Table 2 together with those for a batch of 12 replicates of this sample prepared by the manual microwave digestion procedure. The statistical analysis of these data (F-test) shows no significant difference between the variance of the two sets of results. Although further statistical analysis (t-test) shows the means to be different, this is not considered to be analytically significant, and could be caused by limitations in the ICP-AESANALYST, JANUARY 1992, VOL. 117 7 determination. Hence it can be inferred that the accuracy and reproducibility of the robotic microwave digestion system match those which can be achieved by a competent analyst. Reliability The robotic microwave digestion system completes a 60- sample run in about 16.5 h. The first sample is completed approximately 45 min after the start of the run, with subsequent samples being completed about every 16 min. The system has been in operation for over 6 months completing runs of up to 60 samples overnight. References 1 Kingston, H. M., and Jassie, L. B., in Zntroduction to Microwave Sample Preparation, ed. Kingston, H. M., and Jassie, L. B., American Chemical Society, Washington, DC, 1988, ch. 1. 2 Labrecque, J. M., in Introduction to Microwave Sample Preparation, ed. Kingston, H. M., and Jassie, L. B., American Chemical Society, Washington, DC, 1988, ch. 10, pp. 210-229. Paper 1/01492G Received March 27, 1991 Accepted August 26, 1991