Time-based injection approach for monosegmented continuous flow systems and related techniques

A time-based injection modulefor monosegmented continuousflow systems and related techniques, which uses three independently controlled solenoid valves, is described. A timer c’cuit employing three LC. 555s and three TIP-121 transistors was constructed to control the injection module valves. The injection device was tested with non-reacting chemical systems (for example with a spectrophotometric standard and calcium flame emission) and with reacting conditions (for example the determination of Cr(VI), using diphenylcarbazide as colour reagent, and acid-base titration). The performance of this injection module demonstrates its suitabilityfor everyday use.


Introduction
Most continuous flow analysis systems require the injection of a well-defined sample zone into the moving earner stream. In contrast with the flow injection (FI) analysis technique (either usual FI [1,2] or r-FI [3]), where the sample (or the reagent is injected to a continuous liquid carrier flow, the monosegmented continuous flow analysis (MCFA) system [4] was designed so that the sample is inserted into a carrier stream between air bubbles. Air-segmentation reduces the longitudinal dispersion of the sample along the flow path, reducing sample interaction with the carrier and permitting a longer sample residence time. As a consequence, this flow procedure is able to accommodate analytical methods involving relatively slow reactions without significant loss of sensitivity.
The approaches proposed for sample introduction into flow injection systems can be classified as volume-based or as time-based injection devices.
In the former case, the solution to be injected into the carrier stream is, at least for an instant, contained within an hermetically closed container, such as a valve bore or an external loop. The first prototype was a syringe with an hypodermic needle [5], later replaced by a syringe in combination with a flap valve [6]. More recently, sliding valve commutator [7] and microprocessed devices based on three-way or six-way valves [8,9] have been employed as reliable alternatives.
Time-based injection devices are operated by pumping (or aspirating) the sample solution at a constant flow-rate into a well-defined section of a flow-through channel, for a fixed period of time, and inserting the sampled volume into a carrier stream by alternating the flow directions. This is usually done by using peristaltic pumps or rotating valves controlled by a timer [1,2]. As the operation involves a combination of hydrostatic and hydrodynamic forces, this is known as 'hydrodynamic injection'. Until now, only volume-based injection procedures using a sliding commutator or a six-way solenoid valve have been employed as sample injection devices for the MCFA systems. In this work, an alternative time-based injection device, which makes use of three three-way solenoid valves and a timer circuit based on three I.C. 555s, is presented.  Vc" 5 DC Vcz-12VDC in conjunction with three TIP 121 transistors, as shown in figure (a). The exit clock pulses from the I.e. 555s are directed to two NOR logic gates used to select the TTL level (low or high). When the TTL level is high, these pulses are allowed to reach the base of the transistors responsible for the valve activation.
According to this circuit, each I.e. 555 exit clock is controlled by a bank of resistors (R1 to R7) in figure l[a]), activated from switches SW2 and SW3. These exit clocks control the activation time of each valve. Depending on the position of switch SW4, it is possible to select an extended time range for sample injection at V3. This occurs if switch SW4 is positioned in order to permit SW3 to be part of the circuit, exclusively changing the clock pulses from the I.e. 555, responsible for the activation time of V3.
It must be noted that this increment of time on V3 will automatically increase the time acting on V1, without any alteration to the operation time of V2. If the intention is to use approximately the same time of operation for all valves, SW4 must be positioned in such a way that SW3 becomes inoperative.   [10]. Prior to its reaction with Cr(VI), this solution is mixed on-line with 0"8 mol 1-1 H2SO4.
The procedures followed for both MCFA and FI determinations are those described in the literature [4,10].
The flow titrations [1,11] were performed by injecting a small volume of HC1 solution, with concentrations ranging from 10 -2 to 10mol 1-1, into a constant flow stream of 0"998 mol 1-1 NaOH standard solution containing a few drops of a 0"001% (m/v) bromotymol blue solution. This flow stream is directed through a 730 l.tl reaction cell and then to the flow-cell for signal detection.
The fluids were pumped at a flow rate of 1"7 ml min -1 using an eight-roller Rainin Rabbit peristaltic pump and Tygon pump tubing. Polyethylene (v--600 gl) or glass (v 2500 gl) tubing was employed as mixing or reaction coils in the flow manifold, according to the experimental needs. The signals were measured by a Zeiss PM2A spectrophotometer and recorded at the appropriate maximum wavelength, using an 80 gl Zeiss flow cell with an optical pathlength of 10 mm. In the flow flame emission experiments with Ca2+, the flame photometer (Micronal) was operated as directed by the manufacturer.

Results and discussion
Screening experiments using KMnO4 solutions with concentrations ranging from 5"00 x 10 -5 mol 1-1 to 2-50 x 10 -4 mol 1-1 were done to evaluate the overall performance of the injection device under usual conditions for both MCFA and FI. Either the monosegmented system or the flow injection system can be easily implemented by simply changing the gas (air) to a reagent solution or a carrier at valve V2. No other changes in the valve arrangement or the timer circuit are required.
Both curves were obtained under the same experimental conditions, using the MCFA manifold, since a permeation cell is essential for bubble removal under MCFA operation. Thus, the FI technique showed to be less sensitive, as a result of its inherent on-line sample dilution.
This injection device presents an inherent dead volume at about 9 tl, due to the longitudinal hole at the Teflon mixing plug attached to valve V3, which is responsible for the flow connection with the reaction manifold. This dead volume needs to be as small as possible to avoid response variations on the transient flow peak heights. These variations are less sensitive for FI measurements due to bolus on-line dilution, becoming important only if consecutive samples with large differences in concentration are introduced in sequence. Although these deviations are not relevant for most cases, corrections can be easily made if needed, considering that only the first signal may be affected and that the flow techniques usually make use of at least triplicate injections. Also the use of smaller solenoid valves and drilling smaller longitudinal holes in the plug M will minimize the possible effects of the dead volume.
As the injection is time-based, the major problems are expected to be related to the propelling system, because changes in the flow rates during the injection period may cause variations in the signals. However, no problems of this type were observed using good quality peristaltic pumps, such as that used here.
The injection device was also tested under other flow system situations, such as the MCFA and FI determinations of Ca(II) by flame emission photometry and the spectrophotometric determination of Cr(VI) with diphenylcarbazide [4,10], as well as the flow injection titration of HC1 solutions with standard NaOH [1,11].
The results are summarized in table 2.
Other flow applications for this time-based injection system such as FI double or triple zone injection [12] and MCFA single bubble injection [13] can be derived and are shown in figure 2. These can be implemented by changing only the reagents (or air) and the cartier input lines of each valve. FI operation using two valves can be implemented by disabling V2.  figure 1 (b).
Another important feature for the proposed injection device is that its operation does not depend on a dedicated computer for its control, making possible its use in situations where a computer is not available for full time use. On the other hand, the use of a computer to control the interface TTL output levels will make the proposed injection system much more versatile, since it could assume any value of time for the valve operations. This could permit an alternative means of control of the valves, giving rise to possible new injection configurations and applications.

Conclusion
This paper shows that it is possible to construct an automatic time-based injection module for monosegmented continuous flow system and related techniques, using only three three-way solenoid valves and a simple timer circuit. This is a low cost module whose unique operational set-up was tested for MCFA and FI using both non-reacting and reacting chemical systems, confirming its utility, in practical situations, without changes in the electronic circuit or valve arrangements.