| Introduction | | | | contamination. For instance, the cadmium reduction |
| Think of your old manual Spectronic 20, or your | | | | nitrate test contains significant amounts of ammonia |
| direct reading spectrophotometer that you use in | | | | in the buffer reagent and phosphate in the color |
| your lab. You line up your samples in a row. In front | | | | reagent. Using individual disposable cuvettes ensures |
| of them, you place some small sample cups or | | | | that there is no contamination. Washing cuvettes, or |
| maybe even a series of cuvettes, and you pipette a | | | | using a flow cell, means you can never be sure. |
| known amount of sample into each cup. You then | | | | Using disposable optical cuvettes is the only way you |
| add a reagent and somehow mix the reagent and | | | | can guarantee no carryover between tests or |
| sample. You do this for each sample. You may have | | | | samples. The concept is similar to use of disposable |
| more reagents to add so you repeat the whole | | | | petri dishes, disposable pipette tips, and disposable |
| process until all reagents are added. Then you start a | | | | hypodermic needles. The discrete analyzer easily and |
| timer. When the timer beeps you know you have a | | | | rapidly analyzes multiple tests on single sample |
| certain "time window" to read the absorbance (or | | | | solutions. Only disposable individually contained |
| concentration) of your samples. You read by manually | | | | reactions ensure that there is no interaction between |
| transferring the color-developed sample to a | | | | samples or tests. |
| spectrometer cuvette, by using a peristaltic pump to | | | | Let the robot do your pipetting. |
| transfer the sample to a flow cell already in the | | | | When you manually pipette samples you, hopefully, |
| spectrometer, or by inserting the tube or cuvette | | | | use a different pipette per sample. If not, you will at |
| that you used to develop the sample color in. Then, | | | | least rinse it in between samples, and possibly with |
| you press a button to send the reading to a printer, | | | | sample prior to transferring your sample aliquot to |
| a computer program, or you manually record the | | | | the sample container. This is to avoid carryover |
| reading onto a laboratory worksheet. | | | | between samples. A flow analyzer uses an auto |
| Did you shake and mix every sample exactly the | | | | sampler. The sampling probe immerses in the wash |
| same way every time? Will you mix them the same | | | | station rinsing the outside of the probe, and pulls |
| way every day? Will every analyst run them exactly | | | | wash solution from the station and into the analytical |
| the same way you have? | | | | cartridge. |
| Is there color or turbidity in the samples? Should you | | | | A discrete analyzer also uses a probe; however, it |
| zero your instrument with each sample, or only with | | | | operates differently than flow analyzers. A discrete |
| reagent water blanks? | | | | analyzer's level detect mechanism ensures that the |
| Is the exact time you read the final absorbance | | | | probe immerses into the sample or reagents no |
| critical? | | | | further than necessary to withdraw the required |
| The process described is what you are automating | | | | sample aliquot. The probe then washes itself on the |
| by using a discrete analyzer. Instead of lining up | | | | outside at the wash station and pushes the sample |
| samples, you are pouring aliquots into sample cups | | | | or reagent out into the sample cuvette. Between |
| that are placed on an auto sampler tray. Instead of | | | | dispenses, the probe pushes excess wash water out |
| transferring a known amount of sample to a cuvette, | | | | ensuring no carryover. In other words, unlike a flow |
| the discrete analyzer does. Instead of adding | | | | system that only pulls sample in one direction, the |
| reagents and mixing, the discrete analyzer does. | | | | sampling probe on a discrete analyzer is bidirectional |
| Instead of starting a timer, the discrete analyzer | | | | pulling reagent and sample into its internal tubing only |
| does. Instead of reading the absorbance, recording | | | | far enough to withdraw the correct volume and then |
| the reading, and calculating a result the discrete | | | | dispensing it by pushing it out the other way. |
| analyzer does. | | | | The machine can think. |
| The analyzer has automated almost all the simple | | | | When doing a manual test you know if you ran out |
| colorimetric methods for you. Sample volume is | | | | of reagent or sample. A flow analyzer does not |
| measured and dispensed exactly the same way, | | | | know. A flow analyzer could end up aspirating from |
| every time. Reagents are added and mixed exactly | | | | empty sample cups or empty reagent bottles all night |
| the same way every time. The timer is set and | | | | long and think it is still running samples. A discrete |
| absorbance is measured exactly the same way | | | | analyzer with level detection prevents this. The level |
| every time. Results are calculated exactly the same | | | | detect mechanism is a capacitance detector that |
| way every time. | | | | senses the difference between liquid and air. The |
| The discrete analyzer pipettes, dilutes, adds reagents, | | | | discrete software calculates the volume of reagents |
| mixes, calibrates, measures, calculates, and reports all | | | | and samples based on the height of liquid. The |
| for you. You select a method by keyboard. There is | | | | software continuously monitors sample and reagent |
| no hardware to manually change, no cartridge to | | | | volumes and will not continue the test when it |
| rinse out, no baselines to monitor, no wavelength | | | | detects that reagents or samples have "run out". |
| filters to change. Sample and reagent volumes are | | | | The sampling depth on a flow analyzer is usually |
| determined by a selection in a computer program, | | | | adjustable by the user and is usually towards the |
| not by the internal diameter of a peristaltic pump | | | | bottom of the sample vial. On a discrete analyzer, |
| tube. | | | | the depth the probe immerses in a sample solution is |
| The discrete analyzer has done a lot for you but it | | | | a result of programming or instrument design. The |
| cannot control nor do everything. It cannot | | | | depth sampled on the OI Discrete analyzer is |
| accurately prepare the stock calibration standard for | | | | determined by the level detect mechanism and the |
| you, even though it can accurately dilute it. It cannot | | | | sample aliquot required for the test. For instance, if |
| guarantee the standards and samples were placed on | | | | 200 micro liters is required the probe will immerse just |
| the auto sampler tray in the right order. It cannot | | | | below 200 micro liters as determined by the volume |
| prepare the reagents for you or guarantee they | | | | of the cup and the liquid level detected and withdraw |
| were placed in the right order; however, it can | | | | a software-defined amount above 200 micro liters. In |
| monitor their purity and remind you where they are | | | | other words, the discrete analyzer samples from the |
| supposed to go. It cannot make sure you've entered | | | | top 300 micro liters of sample solution. The probe |
| the proper sample ID for each sample position, | | | | only immerses as far as it has to. This minimizes |
| however, it can guarantee that the result obtained | | | | potential carryover contamination, and speeds the |
| for that sample position is traceable to the ID you | | | | process. In this way dispensing and rinsing is fast and |
| entered. It cannot know the sample lot ID for each | | | | there is no sample or reagent carried to another on |
| standard or reagent, but if you enter those ID's into | | | | the sides of the probe. |
| the software, it can guarantee traceability of those | | | | When sampling from the top of the sample cup |
| reagents with your sample sets. | | | | there is a risk of loss of a volatile analyte from the |
| The software and built in electronics constantly | | | | top of the solution or the risk of the adsorption of |
| monitor and adjust lamp voltage so that absorbance | | | | an analyte from the laboratory air into the top of the |
| readings do not drift. Drift is common in flow | | | | solution. For instance, trace cyanide in near neutral |
| analyzers because the peristaltic pump tubing delivers | | | | solution can be slowly lost from the top layer of |
| reagents by proportion. The discrete analyzer delivers | | | | sample solution into the lab air. This is especially |
| the exact amount of sample and reagent every time. | | | | evident with lower concentrations such as 10 ppb. |
| These volumes do not change. The discrete analyzer | | | | Gain of the analyte is possible as well. Ammonia is a |
| has a fixed path length if the discrete analyzer does | | | | common laboratory contaminant. Ammonia readily |
| not transfer color-developed sample to another | | | | adsorbs into acidified solutions. It is possible for |
| cuvette, or flow cell, for measurement. In addition, if, | | | | ammonia to be "pulled" from laboratory air into the |
| the discrete analyzer reads through the walls of the | | | | sample solution. A flow analyzer would not as readily |
| cuvette the calibration curve is usually more stable | | | | detect this loss or gain because it samples from the |
| and or reproducible than your reagents and | | | | bottom of the sample cup. |
| standards. | | | | There are some drawbacks |
| Change your thoughts on calibration | | | | A discrete analyzer reacts sample in a heated cup |
| Beer's law states that the absorbance is equal to the | | | | that is open to allow the probe to dispense samples |
| absorbtivity times the path length times the | | | | and reagents. The heat increases reaction rates and |
| concentration. It seems, however, sometimes we do | | | | is especially important for chemistries such as |
| not believe that Beer's law is a law. I say this | | | | ammonia that are slow to develop color. In manual |
| because according to this law, the absorbtivity is a | | | | testing the reagents are added in open containers, |
| constant. When the path length is fixed (always the | | | | however, the container shape can vary and the |
| same), the path length is a constant as well making | | | | container can be capped during mixing, heating, and |
| the only variable the concentration. Therefore, you | | | | color reaction. When flow analyzers were first |
| prepare standards of a known concentration, | | | | introduced one of the key advantages that gained its |
| measure the absorbance and determine the | | | | acceptance over manual methods was that reactions |
| absorbtivity. Assuming you can prepare reagents | | | | occurred enclosed within the tubing limiting its |
| exactly the same way every time, measure the | | | | exposure to laboratory air. In this aspect, discrete |
| same volume every time, and incubate your samples | | | | analyzers are kind of a step backwards. |
| the same amount of time every time, there should | | | | There are significant advantages. |
| be no reason to assume that the absorbtivity would | | | | Similar to holding a color developing reaction in its |
| change. If the absorbtivity does not change, then | | | | own container till it reaches a color maximum, |
| there is no reason to calibrate every day. Moreover, | | | | discrete analyzers can also hold intermediate reactions |
| if the absorbtivity is not changing, you could actually | | | | for long periods of time without risk of carryover, |
| be introducing error every time you calibrate because | | | | dilution into a carrier reagents, or excessive |
| you may not be taking into account random errors | | | | dispersion. This can be especially useful in enzyme or |
| that occur between analysts or even with yourself | | | | reduction reactions where reaction rates are slow. A |
| as you inadvertently vary your technique on a | | | | flow analyzer would require long delay coils resulting in |
| day-to-day basis. | | | | very complex SFA chemistry manifolds. Often |
| As mentioned previously, daily calibration is required | | | | elevated temperature is used to speed reactions, but |
| for continuous flow methods because flow methods | | | | in some chemistry, there are limits to the maximum |
| proportion the reagents and sample using a peristaltic | | | | temperatures possible. Since discrete analyzer |
| pump. Those pump tubes are changing with time | | | | reactions are occurring in individually contained |
| changing the relative proportion of sample and | | | | cuvettes, the time delay between reagent additions |
| reagents. Flow analyzers are still incredibly accurate, it | | | | on discrete analyzers is limited only by software. This |
| is just you need to calibrate each time. | | | | is a significant advantage over flow chemistry. |
| Calibrating consumes time. Especially accurate ones | | | | In manual methods, obviously, the operator prepares |
| where you took great care to ensure your standards | | | | all the calibration standards from a stock solution, |
| and reagents are fresh. | | | | dilutes any QC samples from a stock solution, dilutes |
| A manual spectrometer does not necessarily require | | | | samples known to be over calibration prior to color |
| a calibration each time. Many methods written for | | | | development, and dilutes samples that were over |
| manual spectrometers merely say, "analyze a check | | | | calibration once he or she notices that they are. |
| standard with each sample set". In fact, the stability | | | | Unless you have an added auto-dilutor attached to |
| of the calibration curve is the underlying concept | | | | your flow analyzer, you will still be diluting standards |
| behind direct reading spectrophotometers and filter | | | | and over calibration samples. Auto-dilution is an |
| wheel methods. For many colorimetric tests, the | | | | integral function of a discrete analyzer. The dilutions |
| stability of the curve far exceeds the stability of the | | | | can be preset during sample table entry if you know |
| standards or the reagents. Some examples are nitrite | | | | that the samples need to be diluted. Methods can be |
| and phosphate. | | | | programmed such that they dilute every sample and |
| A discrete analyzer should not require daily | | | | standard all the time, or the instrument can be |
| calibrations and should allow us to extrapolate more | | | | programmed so that over calibration, samples are |
| the ion chromatography, gas chromatography, and | | | | diluted and re analyzed. |
| manual direct reading spectrometer concept of the | | | | An analyst changes a manual or flow method from |
| Continuing Calibration Verification, or CCV. As | | | | one to the next by memory, or by referring to the |
| mentioned, the reason the discrete analyzer curves | | | | SOP. How well this particular analyst performs the |
| are stable is that the robot exactly reproduces | | | | procedure is dependent upon his mood, the time of |
| everything every time. You cannot do this because | | | | day, his experience with the method, the availability |
| you are not a robot, the discrete analyzer, however, | | | | of equipment, and many other unquantifiable |
| is. | | | | variables. It is possible to obtain good results and bad |
| A manual method uses more reagent and sample | | | | results by the same manually performed method. A |
| volume because we, as humans, cannot work easily | | | | flow analyzer analyzes everything the same every |
| with small volumes. A flow system uses more | | | | time assuming it is set up the same every time. This |
| reagent than a discrete analyzer because a flow | | | | assumption is valid with experienced flow analysis |
| instrument is continuously pumping reagent through | | | | technicians; however, if the technician does not |
| the system. | | | | understand flow or if there are multiple users results |
| Discrete analyzers that measure the sample | | | | will vary. Extensive training and documentation is |
| absorbance within the same container that the | | | | necessary to guarantee that results conform to good |
| reaction occurred generate less waste than | | | | automated lab practices. |
| instruments that wash the vessel, or use a flow cell. | | | | The discrete analyzer method is selected by mouse |
| In fact, adequately rinsing a flow cell requires | | | | click when scheduling analyses on the sample tray. |
| significant rinsing between samples making the waste | | | | The method conditions do not change. In fact, |
| volume generated essentially equivalent to that of a | | | | assuming you have accurately calibrated your method |
| micro-flow Segmented Flow Analyzer, or Low Flow | | | | the calibration is stored within the method. This |
| Flow Injection Analyzer. | | | | means that an untrained analyst that only knows |
| The discrete analyzer uses significantly less reagent, | | | | what buttons to press is able to obtain identical |
| and generates significantly less waste than manual | | | | results to even the most experienced analyst. |
| methods. This chart illustrates an unscaled down | | | | Most analytes performed in an environmental |
| manual method using the exact volumes described in | | | | compliance laboratory cannot be bench spiked. If the |
| Standard Methods. The waste generated for the | | | | analyte requires a preliminary distillation, digestion, or |
| manual method does not take into account washing | | | | extraction the spiking is done prior to the preliminary |
| of glassware. As mentioned earlier, an analyzer that | | | | sample process. I realize that many labs do not distill |
| washes cuvettes or rinses a flow cell will generate | | | | ammonia or Fluoride and I would argue that if you |
| more waste than indicated here. | | | | are reporting compliance testing for the clean water |
| Eliminate the possibility of contamination, or false | | | | act you would better seriously consider changing your |
| positives | | | | SOP. Other parameters that can't be spiked are |
| The discrete analyzer measuring the absorbance of a | | | | those that are too high to spike within the matrix |
| color reacted sample contained in individual cuvettes. | | | | without preliminary dilution, such as Ca, Mg, Cl, SO4, |
| Unlike flow analysis, there is no possibility of | | | | and analytes like alkalinity that just are not spiked. |
| interaction between samples and unlike flow analysis; | | | | This shortens the list of potential analytes for the |
| the user can visually observe the reaction product | | | | automatic spiking function to nitrite, phosphate, |
| during and after analysis. | | | | Sulfide, Chromium VI, and some others. On these, I |
| Using a discrete analyzer, the analyst can observe | | | | defer back to the previous slide and ask if the |
| the reaction during color development and after the | | | | potential error is worth the risk for so few tests. |
| test is complete. The analyst can remove the | | | | Summary |
| reaction segments and verify that dispensed volumes | | | | Benefits of discrete analyzers include decreased |
| are repeatable, that there are no bubbles or turbidity, | | | | reagent consumption, decreased waste generated, |
| and that the color looks correct. A flow analyzer | | | | and ease of use among other things. The most |
| does not give the analyst the ability to visually | | | | significant advantage of the discrete analyzer, |
| examine and qualitatively guarantee the accuracy of | | | | however, is that it can eliminate the traditional |
| his or her results. | | | | concept of routine analysis and allow you to run |
| A discrete analyzer dispenses, reacts, incubates, and | | | | samples as you receive them instead of storing them |
| measures all within the reaction cuvette without | | | | until there is enough sitting around to make a flow or |
| transferring to a flow cell. Analyzers that transfer to | | | | IC analysis worthwhile. If you take advantage of the |
| a flow cell are not "true" discrete analyzers, but | | | | calibration stability of the discrete analyzer, and |
| instead, are hybrids between flow and discrete. The | | | | accurately prepare a calibration that can then be used |
| hybridization is done to achieve lower detection limits; | | | | by almost any analyst in subsequent uses an added |
| however, the advantage of the individually contained | | | | benefit is that the results are the same regardless of |
| reaction and absence of carryover is lost. In addition, | | | | who uses the machine. |
| since these analyzers require as much rinse as a flow | | | | Think of those short holding time samples. The |
| analyzer to remove preceding samples, waste | | | | phosphate, the nitrites, the chromium VI, and residual |
| generation is as high as flow. Given this, and the | | | | chlorine. These analytes cause the environmental lab |
| increased possibility of environmental contamination or | | | | to stop everything just to get the analysis done on |
| analyte loss that occurs from open-air heated | | | | time. Think of the other analytes that come in |
| reactions, you may as well have a flow analyzer. | | | | periodically, but maybe not frequently. Possibly silica, |
| Chemical reactions occur in individually contained | | | | ferrous iron and sulfide. How do you guarantee these |
| segments | | | | tests followed the SOP? Instead of thinking of the |
| All discrete analyzers have reaction segments. Some | | | | discrete analyzer as something to replace a flow |
| analyzers do chemical reactions in a cuvette segment | | | | instrument, think of it as something to supplement a |
| and then transfer the reacted sample to a flow cell. | | | | flow instrument. If you have hundreds of samples for |
| This type of analyzer is a hybrid of discrete and flow, | | | | one or two tests routinely and for the same analyte |
| and not a true discrete analyzer. A true discrete | | | | you are not going to save money by switching these |
| analyzer reacts and measures the sample within the | | | | tests to a discrete analyzer. Where you will save |
| optical cuvette. Some analyzers wash the optical | | | | money and great effort is removing unnecessary |
| cuvette between tests. Washing between tests | | | | strain from the flow analyzer and your analysts by |
| enables more samples to be analyzed per cuvette; | | | | performing the non - routine or "rush" tests on a |
| however, the washing cannot guarantee that there is | | | | discrete analyzer. It is possible for the sample login |
| no residual contamination that remaining after the | | | | person to analyze samples as received for almost |
| washing process. Other discrete analyzers utilize | | | | every colorimetric test that does not require a |
| disposable optical quality cuvettes. | | | | digestion. In other words, as soon as the sample is |
| Washing between tests enables more samples to be | | | | logged in it could be immediately run for nitrite, |
| analyzed per cuvette; however, the washing cannot | | | | phosphate, chromium VI, nitrate, ammonia, chloride, |
| guarantee that there is no residual contamination not | | | | and sulfate. In this example, instead of putting |
| completely removed by the washing process. This | | | | samples in a refrigerator to be gathered for analysis |
| residual contamination can come from preceding | | | | at a later time, they end up being run by ice chest |
| samples, or more likely, from the reagents used in | | | | and by client as soon as they are received. |
| processing the preceding samples. The built in | | | | If everything is to run on the discrete analyzer, then |
| computerized checking of optical quality cannot verify | | | | collect your samples in a vial that fits on the discrete |
| absence of chemical contamination. | | | | analyzer. You no longer need to transfer liquid from |
| Analyzers that use a flow cell still react samples in | | | | container A to auto sampler vial B, the sample bottle |
| some sort of cuvette. It is the number of reaction | | | | can be the auto sampler vial. Not only does this save |
| vessels on the discrete analyzer that limit the number | | | | time, but it saves shipping as well. Instead of large ice |
| of tests that the discrete can run in a single walk | | | | chests, you use tiny mailers. |
| away operation. If the discrete analyzer has 100 | | | | To summarize, the true advantage of a discrete |
| sample positions and 200 reaction cuvettes, then the | | | | analyzer is that its built in features allow any analyst |
| analyzer can run 100 samples for 2 tests each. The | | | | to get the same results every time. Discrete |
| discrete analyzer with the flow cell must rinse the | | | | analyzers are very simple to use requiring minimal |
| flow cell between each sample, and rinse vigorously | | | | software training. Once set up for your laboratory, |
| between each test. Consider that a two-channel flow | | | | properly applied methods allow you to modify your |
| analyzer can analyze 100 samples for two tests each | | | | daily routines and analyze samples as soon as they |
| in less than half the time as a discrete analyzer with a | | | | come in. Whether you are an environmental lab, |
| flow cell. Also, consider that the flow analyzer | | | | research, process control, or municipality discrete |
| generates no more waste than the discrete analyzer | | | | analyzers can be used effectively in your operation. |
| with a flow cell. If the required testing is a lot of | | | | Currently, the full power of discrete analyzers is |
| samples for one or two tests it makes more sense | | | | limited by tradition and by regulation. Once we start |
| to use a flow analyzer. | | | | to develop methods for discrete analyzers instead of |
| Reagents can interfere as cross contamination | | | | using discrete analyzers to run methods developed |
| between samples. Using disposable individual reaction | | | | for flow we will be able to see greater throughput, |
| cuvettes completely eliminates the possibility of | | | | less variability, and lower MDL. |