Design of LabVIEW®-based software for the control of sequential injection analysis instrumentation for the determination of morphine

LabVIEW®-based software for the automation of a sequential injection analysis instrument for the determination of morphine is presented. Detection was based on its chemiluminescence reaction with acidic potassium permanganate in the presence of sodium polyphosphate. The calibration function approximated linearity (range 5 × 10-10 to 5 × 10-6 M) with a line of best fit of y=1.05x+8.9164 (R2 =0.9959), where y is the log10 signal (mV) and x is the log10 morphine concentration (M). Precision, as measured by relative standard deviation, was 0.7% for five replicate analyses of morphine standard (5 × 10-8 M). The limit of detection (3σ) was determined as 5 × 10-11 M morphine.


Introduction
Optimization and control of modern chemical processes requires high-quality chemical information [1,2]. Such information is ideally provided in real time using process analytical chemistry. Flow injection analysis (FIA) is a powerful, well-established, sample-handling technique well suited to this type of chemistry and it has been applied to online process analysis in industrial [1], fermentation [3] and environmental [4,5] analysis. The application of FIA to online process analysis can be hampered by the requirement for complex manifolds. In addition, peristaltic pump tubing is generally not compatible with harsh sample matrices [6]. To overcome these limitations, Ruzicka and Marshall introduced sequential injection analysis (SIA) [7]. In contrast to FIA, SIA uses computer-controlled¯ow programming to aå ord the application of diå erent chemistries without recon® guration of the¯ow manifold [8,9]. SIA manifolds comprise a multiposition valve operating in synchronization with a pump and a suitable detector. The manifold is robust, easily maintained and ideally suited to online process analysis [8]. It is essential that the entire instrument is computer controlled, as precise timing of the pump and valve is required to achieve controlled partial dispersion of the reagent and sample [7,10]. The con® guration of SIA instrumentation for the determination of morphine in process streams necessitated the writing of suitable software to control the system and perform data acquisition.
This paper describes the design and application of software written within the National Instruments Lab-VIEW 1 graphical programming environment [11± 13] to automate fully SIA instrumentation and data acquisition for the determination of morphine in process liquors. The LabVIEW 1 software facilitated the design of virtual instruments, which allowed synchronized control and data acquisition for the entire analytical system.

Hardware
All experiments were performed using a purpose-built sequential injection analysis instrument (® gure 1). Control of the pump (Cavro XP-3000, Global FIA, Gig Harbour, Washington, USA) and the 10-port multiposition valve (Valco C25Z, SGE, Melbourne, Australia) was achieved using a desktop computer (Pentium 133 MHz, 32 MByte RAM, Posicom, Geelong, Victoria, Australia) equipped with a data acquisition board (LabPC 1200, National Instruments, Ringwood, Victoria, Australia) running software written in-house using LabVIEW 1 v.6.0 (National Instruments). Detection was accomplished using a custom-built¯ow-through chemiluminometer, the details of which are as follows. A glass spiral¯ow cell (Embell Scienti® c, Murwillimbah, New South Wales, Australia) was mounted¯ush against a photomultiplier tube (Thorn EMI Type 9828, ETP Ltd, Ermington, New South Wales, Australia) operating at a constant 800 V supplied by a stable power supply (Thorn EMI Model PM28BN) via a voltage divider supply (Thorn EMI Model C611). All tubing was 0.8 mm i.d. PTFE (ProTECH Pty Ltd, Coolum Beach, Queensland, Australia).

Reagents and analytes
Deionized water and analytical grade reagents were used unless otherwise stated. Stock morphine solutions, working standards and acidic permanganat e reagents were all prepared by dissolution in an acidic solution (0.05 m sulphuric acid; Ajax Chemicals, Aurburn, New South Wales, Australia) of sodium polyphosphat e (Aldrich, Castle Hill, New South Wales, Australia). Stock morphine solutions were prepared by dissolution of the free base (Glaxo Smith Kline, Port Fairy, Victoria, Australia) with working standards prepared by serial dilution. Acidic permanganat e solutions were prepared by dissolution of potassium permanganate (Merck Ltd, Poole, UK).

Software design
The basic layout and requirements of the SIA system are shown in ® gure 2. The syringe pump was controlled through the RS 232 serial port whilst the multiposition valve was manipulated using digital-out signals from the data-acquisition board. The board also allowed data collection using diå erential input from two analogueinput channels. Virtual instruments, developed within LabVIEW 1 , consist of a user interface and a graphical data¯ow diagram that contains the source code. These are modular and hierarchical and, as such, can be used as stand-alone programs or as a subprogram (subvirtual instrument). Using this facility, individual virtual instruments for the control of each component and for data acquisition were developed, tested and then linked together to form the ® nal program.
A virtual instrument module was designed for the electronic actuation of the multiposition valve using TTL high/low signals on a single digital out-control line. The module could`step' the valve to the next port, send the valve to the`home' position and`reset' the valve. Control of the syringe pump was achieved using a second virtual   instrument; commands written in ASCII format were sent from the computer to the pump via an RS-232 connection (® gure 3). The front panel (® gure 3A) shows three user inputs: pump¯ow rate (velocity), direction (command) and volume (number of steps). The corresponding¯ow diagram (® gure 3B) shows the source code; user inputs are wired to icons that represent a subprogram that the virtual instrument is executing with its front panel shown in ® gure 3C.
A third virtual instrument was designed to acquire and process analogue data from the photomultiplier tube. This module displays graphically both raw and smoothed data and allows the user to choose the data-acquisition rate. The sampling period is calculated by the software and is dependent upon the¯ow rate and injection volume. The high and low limit settings for the input signals allowed accurate digital reproduction of the SIA detector response pro® le. The acquired data were digitally ® ltered (Butterworth ® lter within LabVIEW 1 ) and saved in either ASCII or Microsoft Excel formats with dynamic data exchange facilitating the latter.
The front panel of the SIA virtual instrument is shown in ® gure 4 with the data ® le-saving options, pump controls and graphical display positioned on the left, bottom and right of the screen, respectively. A selector button in the top left-hand corner gives the user control over the type of experiment to be conducted (either programme initialization, loading solutions or analysis). The virtual instrument hierarchy used to control the instrument is shown in ® gure 5. For a free copy of the executable software, contact the authors.

Determination of morphine
Flow-injection analysis determination of morphine based on its chemiluminescence reaction with acidic potassium permanganate in the presence of polyphosphate s was ® rst reported in 1986 [14]. This chemistry was adapted for the  determination of morphine in process samples in 1993 [15] and more recently was modi® ed to suit SIA for both aqueous and non-aqueous process extracts [16,17]. Consequently, we used this well-established detection chemistry to test the performance of the present automated system.
Potassium permanganat e (5 : 0 £ 10 ¡4 m ) and morphine standards were prepared in sodium polyphosphat e (1.0% m/v) and adjusted to pH 2.0 with sulphuric acid. The carrier solution was sodium polyphosphat e (1.0% m/v) in sulphuric acid (pH 2.0). The three-way valve on the pump was set to the left, and carrier solution (3.1 ml) was drawn into the syringe at 12 ml min ¡1 . The three-way valve was then set to the right and port 8 chosen on the multipositon valve. Potassium permanganat e reagent (250 m l) was aspirated into the holding coil at 1.2 ml min ¡1 , the multiposition valve was switched to port 9 and sample (50 m l) was aspirated into the holding coil at 1.2 ml min ¡1 . The multiposition valve was then switched to port 10 and the entire volume (3.4 ml) was ushed passed the detector at 12 ml min ¡1 .
The analytical ® gures of merit obtained with this system are superior to those reported previously [16] (table 1). The instrumental reproducibility is demonstrated in ® gure 6, which shows ® ve replicate injections (6 : 0£ 10 ¡6 m ) overlaid. The detection limit of 5 : 0 £ 10 ¡11 m is, to the best of our knowledge, the lowest achieved for morphine using this chemistry. The lower detection limit achieved with the current system can be attributed to the replacement of the peristaltic pump [16] by a syringe pump and signi® cant improvements in the`light tight-ness' of the instrument housing as both these changes improved the signal-to-nois e ratio. The high precision (table 1) can be attributed to the superior hydrodynamic control attainable with a syringe pump, as the volumes delivered and¯ow rates are not susceptible to variation through changes in pump tubing dimensions. This instrumentation and chemistry is presently being adapted for the determination of morphine and other related alkaloids in process samples with results to be published in due course.

Conclusions
The complete automation of an SIA instrument was achieved using software written with LabVIEW 1 . Individual modules were written for each component and linked to form the ® nal program. The instrumentation was applied to the determination of morphine, with signi® cant improvements to the analytical ® gures of merit reported previously [16].  Figure 6. Five replicate injections of a 6 : 0 £ 10 ¡6 m morphine standard, overlaid.