Tuesday, January 23, 2018

IAM Search

Continuous Analytical Measurement - Chromatography

Imagine a major marathon race, where hundreds of runners gather in one place to compete. When the starting gun is fired, all the runners begin running the race, starting from the same location (the starting line) at the same time. As the race progresses, the faster runners distance themselves from the slower runners, resulting in a dispersion of runners along the race course over time. Now imagine a marathon race where certain runners share the exact same running speeds. Suppose a group of runners in this marathon all run at exactly 8 miles per hour (MPH), while another group of runners in the race run at exactly 6 miles per hour, and another group of runners plod along at exactly 5 miles per hour. What would happen to these three groups of runners over time, supposing they all begin the race at the same location and at the exact same time?

As you can probably imagine, the runners within each speed group will stay with each other throughout the race, with the three groups becoming further spread apart over time. The first of these three groups to cross the finish line will be the 8 MPH runners, followed by the 6 MPH runners a bit later, and then followed by the 5 MPH runners after that. To an observer at the very start of the race, it would be difficult to tell exactly how many 6 MPH runners there were in the crowd, but to an observer at the finish line with a stop watch, it would be very easy to tell how many 6 MPH runners competed in the race (by counting how many runners crossed the finish line at the exact time corresponding to a speed of 6 MPH).

Now imagine a mixture of chemicals in a fluid state traveling through a very small-diameter “capillary” tube filled with an inert, porous material such as sand. Some of those fluid molecules will find it easier to progress down the length of the tube than others, with similar molecules sharing similar propagation speeds. Thus, a small sample of that chemical mixture injected into such a capillary tube, and carried along the tube by a continuous flow of solvent (gas or liquid), will tend to separate into its constituent components over time just like the crowd of marathon runners separate over time according to running speed. Slower-moving molecules will experience greater retention time inside the capillary tube, while faster-moving molecules experience less. A detector placed at the outlet of the capillary tube, configured to detect any chemical different from the solvent, will indicate the different components exiting the tube at different times. If the retention time of each chemical component is known from prior tests, this device may be used to identify the composition of the original chemical mix (and even how much of each component was present in the injected sample).

This is the essence of chromatography: the technique of chemical separation by time-delayed travel down the length of a stationary medium (called a column). In chromatography, the chemical solution traveling down the column is called the mobile phase, while the solid and/or liquid substance residing within the column is called the stationary phase. Chromatography was first applied to chemical analysis by a Russian botanist named Tswett, who was interested in separating mixtures of plant pigments. The colorful bands left behind in the stationary phase by the separated pigments gave rise to the name “chromatography,” which literally means “color writing.”

Modern chemists often apply chromatographic techniques in the laboratory to purify chemical samples, and/or to measure the concentrations of different chemical substances within mixtures. Some of these techniques are manual (such as in the case of thin-layer chromatography, where liquid solvents carry liquid chemical components along a flat plate covered with an inert coating such as alumina, and the positions of the chemical drops after time distinguishes one component from another). Other techniques are automated, with machines called chromatographs performing the timed analysis of chemical travel through tightly-packed tubular columns.

An illustrated sequence showing thin-layer chromatography appears here:



The simplest forms of chromatography reveal the chemical composition of the analyzed mixture as residue retained by the stationary phase. In the case of thin-layer chromatography, the different liquid components of the mobile phase remain embedded in the stationary phase at distinct locations after sufficient “developing” time. The same is true in paper-strip chromatography where a simple strip of filter paper serves as the stationary phase through which the mobile phase (liquid sample and solvent) travels: the different components of the sample remain in the paper as residue, their relative positions along the paper’s length indicating their extent of travel during the test period. If the components have different colors, the result will be a stratified pattern of colors on the paper strip1.

Most chromatography techniques, however, allow the sample to completely wash through a packed column, relying on a detector at the end of the column to indicate when each component has exited the column. A simplified schematic of a process gas chromatograph (GC) shows how this type of analyzer functions:


The sample valve periodically injects a very precise quantity of sample into the entrance of the column tube and then shuts off to allow the constant-flow carrier gas to wash this sample through the length of the column tube. Each component of the sample travels through the column at different rates, exiting the column at different times. All the detector needs to do is be able to tell the difference between pure carrier gas and carrier gas mixed with anything else (components of the sample).

Several different detector designs exist for process gas chromatographs. The two most common are the flame ionization detector (FID) and the thermal conductivity detector (TCD). Other detector types include the flame photometric detector (FPD), Nitrogen-Phosphorus Detector (NPD), and electron capture detector (ECD). All chromatograph detectors exploit some physical difference between the solutes (sample components dissolved within the carrier gas) and the carrier gas itself which acts as a gaseous solvent, so that the detector may be able to tell the difference between pure carrier and carrier mixed with solute.

Flame ionization detectors work on the principle of ions liberated in the combustion of the sample components. A permanent flame (usually fueled by hydrogen gas which produces negligible ions in combustion) serves to ionize any gas molecules exiting the chromatograph column that are not carrier gas. Common carrier gases used with FID sensors are helium and nitrogen. Gas molecules containing carbon easily ionize during combustion, which makes the FID sensor well-suited for GC analysis in the petrochemical industries, where hydrocarbon content analysis is the most common form of analytical measurement2.

Thermal conductivity detectors work on the principle of heat transfer by convection (gas cooling). Recall the dependence of a thermal mass flowmeter’s calibration on the specific heat value of the gas being measured3. This dependence upon specific heat meant that we needed to know the specific heat value of the gas whose flow we intend to measure, or else the flowmeter’s calibration would be in jeopardy. Here, in the context of chromatograph detectors, we exploit the impact specific heat value has on thermal convection, using this principle to detect compositional change for a constant flow gas rate. The temperature change of a heated RTD or thermistor caused by exposure to a gas mixture with changing specific heat value indicates when a new sample component exits the chromatograph column.

If we plot the response of the detector on a graph, we see a pattern of peaks, each one indicating the departure of a component “group” exiting the column. This graph is typically called a chromatogram:


Narrow peaks represent compact bunches of molecules all exiting the column at nearly the same time. Wide peaks represent more diffuse groupings of similar (or identical) molecules. In this chromatogram, you can see that components 4 and 5 are not clearly differentiated over time. Better separation of components may be achieved by altering the sample volume, carrier gas flow rate, carrier gas pressure, type of carrier gas, column packing material, and/or column temperature. Changes in column temperature (called temperature programming) are very commonly used to alter the retention times of different components during an analysis cycle, working on the principle of a fluid’s viscosity being dependent on temperature4. Since the flow regime of the mobile phase through a chromatograph column is definitely laminar (not turbulent), fluid viscosity plays a large role in determining flow rate.

If the relative propagation speeds of each component is known in advance, the chromatogram peaks may be used to identify the presence (and quantities of) those components. The quantity of each component present in the original sample may be determined by applying the calculus technique of integration to each chromatogram peak, calculating the area underneath each curve. The vertical axis represents detector signal, which is proportional to component concentration5 which is proportional to flow rate given a fixed carrier flow rate. This means the height of each peak represents mass flow rate of each component (W, in units of micrograms per minute, or some similar units). The horizontal axis represents time, so therefore the integral (sum of infinitesimal products) of the detector signal over the time interval for any specific peak (time t1 to t2) represents a mass quantity that has passed through the column. In simplified terms, a mass flow rate (micrograms per minute) multiplied by a time interval (minutes) equals mass in micrograms:

As is the case with all examples of integration, the unit of measurement for the totalized result is the product of the units within the integrand: flow rate (W) in units of micrograms per minute multiplied by increments of time (dt) in the unit of minutes, summed together over an interval (R t2t1), result in a mass quantity (m) expressed in the unit of micrograms. Integration is really nothing more than the sum of products, with dimensional analysis working as it does with any product of two physical quantities:



This mathematical relationship may be seen in graphical form by shading the area underneath the peak of a chromatogram:

Since process chromatographs have the ability to independently analyze the quantities of multiple components in a chemical sample, these instruments are inherently multi-variable. A single analog output signal (e.g. 4-20 mA) would only be able to transmit information about the concentration of any one component (any one peak) in the chromatogram. This is perfectly adequate if only one component concentration is worth knowing about in the process21, but some form of multi-channel digital (or multiple analog outputs) transmission is necessary to make full use of a chromatograph’s ability.

All modern chromatographs are “smart” instruments, containing one or more digital computers which execute the calculations necessary to derive precise measurements from chromatogram data. The computational power of modern chromatographs may be used to further analyze the process sample, beyond simple determinations of concentration or quantity. Examples of more abstract analyses include approximate octane value of gasoline (based on the relative concentrations of several components), or the heating value of natural gas (based on the relative concentrations of methane, ethane, propane, butane, carbon dioxide, helium, etc. in a sample of natural gas).

The following photograph shows a gas chromatograph (GC) fulfilling precisely this purpose – the determination of heating value for natural gas7:


This particular GC is used by a natural gas distribution company as part of its pricing system. The heating value of the natural gas is used as data to calculate the selling price of the natural gas (dollars per standard cubic foot), so the customers pay only for the actual benefit of the gas (i.e. its ability to function as a fuel) and not just volumetric or mass quantity. No chromatograph can directly measure the heating value of natural gas, but the analytical process of chromatography can determine the relative concentrations of compounds within the natural gas. A computer, taking those concentration measurements and multiplying each one by the respective heating value of each compound, derives the gross heating value of the natural gas.

Although the column cannot be seen in the photograph of the GC, several high-pressure steel “bottles” may be seen in the background holding carrier gas used to wash the natural gas sample through the column.

A typical gas chromatograph column appears in the next photograph. It is nothing more than a stainless-steel tube packed with an inert, porous filling material:

This particular GC column is 28 feet long, with an outside diameter of only 1/8 inch (the tube’s inside diameter is even less than that). Column geometry and packing material vary greatly with application. The many choices intrinsic to column design are best left to specialists in the field of chromatography, not the average technician or even the average process engineer.

Arguably, the most important component of a process gas chromatograph is the sample valve. Its purpose is to inject the exact same sample quantity into the column at the beginning of each cycle. If the sample quantity is not repeatable, the measured quantities exiting the column will change from cycle to cycle even if the sample composition does not change. If the valve’s cycle time is not repeatable, component separation efficiency will vary from cycle to cycle. If the sample valve leaks such that a small flow rate of sample continuously enters the column, the result will be an altered “baseline” signal at the detector (at best) and total corruption of the analysis (at worst). Many process chromatograph problems are caused by irregularities in the sample valve(s).

A common form of sample valve uses a rotating element to switch port connections between the sample gas stream, carrier gas stream, and column:

Three slots connect three pairs of ports together. When the rotary valve actuates, the port connections switch, redirecting gas flows.

Connected to a sample stream, carrier stream, and column, the rotary sample valve operates in two different modes. The first mode is a “loading” position where the sample stream flows through a short length of tubing (called a sample loop) and exits to a waste discharge port, while the carrier gas flows through the column to wash the last sample through. The second mode is a “sampling” position where the volume of sample gas held in the sample loop tubing gets injected into the column by a flow of carrier gas behind it:

The purpose of the sample loop tube is to act as a holding reservoir for a fixed volume of sample gas. When the sample valve switches to the sample position, the carrier gas will flush the contents of the sample loop into the front of the column. This valve configuration guarantees that the injected sample volume does not vary with inevitable variations in sample valve actuation time. The sample valve need only remain in the “sampling” position long enough to completely flush the sample loop tube, and the proper volume of injected sample gas is guaranteed.

While in the loading position, the stream of gas sampled from the process continuously fills the sample loop and then exits to a waste port. This may seem unnecessary but it is in fact essential for practical sampling operation. The volume of process gas injected into the chromatograph column during each cycle is so small (typically measured in units of microliters!) that a continuous flow of sample gas to waste is necessary to purge the impulse line connecting the analyzer to the process, which in turn is necessary for the analyzer to obtain analyses of current conditions. If it were not for the continuous flow of sample to waste, it would take a very long time for a sample of process gas to make its way through the long impulse tube to the analyzer to be sampled!

Even with continuous flow in the impulse line, process chromatographs exhibit substantial dead time in their analyses for the simple reason of having to wait for the next sample to progress through the entire length of the column. It is the basic nature of a chromatograph to separate components of a chemical stream over time, and so a certain amount of dead time will be inevitable. However, dead time in any measuring instrument is an undesirable quality. Dead time in a feedback control loop is especially bad, as it greatly increases the chances of instability.

One way to reduce the dead time of a chromatograph is to alter some of its operating parameters during the analysis cycle in such a way that it speeds up the progress of the mobile phase during periods of time where slowness of elution is not as important for fine separation of components. The flow rate of the mobile phase may be altered, the temperature of the column may be ramped up or down, and even different columns may be switched into the mobile phase stream. In chromatography, we refer to this on-line alteration of parameters as programming. Temperature programming is an especially popular feature of process gas chromatographs, due to the direct effect temperature has on the viscosity of a flowing gas8. Carefully altering the operating temperature of a GC column while a sample washes through it is an excellent way to optimize the separation and time delay properties of a column, effectively realizing the high separation properties of a long column with the reduced dead time of a much shorter column.



1This effect is particularly striking when paper-strip chromatography is used to analyze the composition of ink. It is really quite amazing to see how many different colors are contained in plain “black” ink!

2In fact, FID sensors are sometimes referred to as carbon counters, since their response is almost directly proportional to the number of carbon atoms passing through the flame.

3See True Mass Flowmeters, subsection Thermal Flowmeters. The greater the specific heat value of a gas, the more heat energy it can carry away from a hot object through convection, all other factors being equal.

4Liquids generally become less viscous when heated, and gases become more viscous when heated. Therefore, raising column temperature in a gas chromatograph will “slow down” the later components (raise retention time) to achieve better separation.

5Detector response also varies substantially with the type of substance being detected, and not just its concentration. A flame ionization detector (FID), for instance, yields different responses for a given mass flow rate of butane (C4H10) than it does for the same mass flow rate of methane (CH4), due to the differing carbon count per mass ratios of the two compounds. This means the same raw signal from an FID sensor generated by a concentration of butane versus a concentration of methane actually represents different concentrations of butane versus methane in the carrier. The inconsistent response of a chromatograph detector to different sampled components is not as troubling a problem as one might think, though. Since the chromatograph column does a good job separating each component from the other over time, we may program the computer to re-calibrate itself for each component at the specific time(s) each component is expected to exit the column. So long as we know in advance the characteristic detector response for each expected compound separated by the chromatograph, we may easily compensate for those variations in real time so the chromatogram consistently and accurately represents component concentrations over the entire analysis cycle.

6It is not uncommon to find chromatographs used in processes to measure the concentration of a single chemical component, even though the device is capable of measuring the concentrations of multiple components in that process stream. In those cases, chromatography is (or was at the time of installation) the most practical analytical technique to use for quantitative detection of that substance. Why else use an inherently multi-variable analyzer when you could have used a single-variable technology that was simpler? By analogy, it is possible to use a Coriolis flowmeter to measure nothing but fluid density, even though such a device is fully capable of measuring fluid density and mass flow rate and temperature.

7Since the heat of combustion is well-known for various components of natural gas (methane, ethane, propane, etc.), all the chromatograph computer needs to do is multiply the different heat values by their respective concentrations in the gas flowstream, then average the total heat value per unit volume (or mass) of natural gas.

8Whereas most liquids decrease in viscosity as temperature rises, gases increase in viscosity as they get hotter. Since the flow regime through a chromatograph column is most definitely laminar and not turbulent, viscosity has a great effect on flow rate.

Go Back to Lessons in Instrumentation Table of Contents

Comments (0)Add Comment

Write comment

security code
Write the displayed characters



  • ...more


Important: All images are copyrighted to their respective owners. All content cited is derived from their respective sources.

Contact us for information and your inquiries. IAMechatronics is open to link exchanges.

IAMechatronics Login