ADVANCED ION CHROMATOGRAPH DETECTS LOW MEA LEVELS

Jan. 22, 1990
Scott Lindsay Texaco Refining & Marketing Inc. Anacortes, Wash. David Palladino Beckmann Instruments Inc. Westbury, N.Y. Texaco Refining & Marketing Inc. is using an advanced method of ion chromatography at its Puget Sound refinery in Anacortes, Wash., to detect and measure monoethanolamine (MEA) in process effluent water at low-ppm levels. The method is electronically suppressed, single-column, ion chromatography (SCIC).
Scott Lindsay
Texaco Refining & Marketing Inc.
Anacortes, Wash.
David Palladino
Beckmann Instruments Inc.
Westbury, N.Y.

Texaco Refining & Marketing Inc. is using an advanced method of ion chromatography at its Puget Sound refinery in Anacortes, Wash., to detect and measure monoethanolamine (MEA) in process effluent water at low-ppm levels. The method is electronically suppressed, single-column, ion chromatography (SCIC).

The method was selected for use in this service subsequent to successful use to improve titrimetric analysis of wash water, where low-ppm ranges of chlorides needed to be measured. SCIC was also able to detect halides of other constituents, such as bromides and iodides.

The advanced method can measure precisely very low levels of both monoethanolamine and ammonia, and it assists in locating the sources that allow the contaminants to escape into the wastewater stream.

In these analyses, the SCIC unit operated in less than one third the time of automatic titrimetry and other wet chemistry methods available in the laboratory. It also proved to be more economical than conventional chemically suppressed ion chromatography (CSIC).

WASH WATER ANALYSIS

Texaco's Puget Sound plant makes a variety of fuels. For this purpose, several million gallons of crude oil are processed each day.

The process begins with crude oil passing through a series of heat exchangers and two desalters where wash water is injected to remove chlorides present in the crude oil.

The chlorides are in various forms, but mainly magnesium and calcium chlorides.

These chlorides, if not removed, would hydrolyze in the crude atmospheric tower and form acids in the overhead condensers, which would lead to corrosion and possible failure.

To control corrosion, chlorides levels are measured in the crude oil before and after the desalters. Also, a sodium-ion caustic is added to the crude oil to further reduce the chloride content after the desalters. The reaction of caustic with calcium and magnesium chlorides forms a more heat-stable sodium-chloride salt which will not hydrolyze in the crude tower.

As a control parameter, chlorides are monitored in an overhead wash-water accumulator. Chloride levels of 10-20 ppm in the atmospheric tower overhead accumulator indicate good chloride control. However, lower levels are sometimes present, in the ppb range and must be measured to signal a possible problem.

In the past, to test for chlorides in the wash water, an automatic titrator with silver nitrate titrant and a silver electrode was used. Four potentiometric anion analyses were performed each day.

In using the titrator, it was found that hydrogen sulfide (and various mercaptan compounds) in the wash water, by reacting with the silver nitrate titrant, caused interference in the chloride measurements. This prohibited clear detection of the chloride end-point. Approximately 15% of the analyses were inaccurate because of this interference.

Also, information on trace amounts (less than 0.1 ppm) of other halides in the wash water, such as bromide and iodide-requested by Texaco corrosion engineers for halogenated hydrocarbon studies-was unobtainable by titrimetry. The automatic titrator could only detect concentrations above 1 ppm.

To eliminate hydrogen sulfide and mercaptan compound interference in chloride measurements, and to simultaneously measure bromide and iodide traces in the wash water, the Texaco laboratory at the plant considered more sensitive analytical techniques.

A CSIC system was investigated and found adequate on both counts. However, it was considered costly to purchase and operate because it requires two columns, two pumps, and two eluents.

Seeking economy, the laboratory explored the comparatively new SCIC technique which employs only one column, one pump, and one eluent (Fig. 1). In a laboratory demonstration, the unit performed successfully down to the ppb range in analyzing multiple ions.

Texaco purchased and installed a unit of this type in October 1987. Precise measurements of the halides with the instrument led to the development of a new wash water sample preparation step. Each wash water sample was boiled and mixed with 1 ml of nitric acid to remove hydrogen sulfide and mercaptan compounds prior to analysis.

This new sample-preparation step eliminated interference with the titrimeter's silver nitrate titrant and allowed the laboratory to return to more effective titrimetric analysis of wash water. Also, the SCIC instrument was used to establish ranges of halides of bromide and iodide present in the wash water.

NEW PROBLEM

Nitrogen-bearing compounds other than ammonia were found to be present in refinery effluent streams, as determined by differences between total nitrogen analysis and ion-selective electrode ammonia analysis.

The suspected cause was monoethanolamine, used in the closed-loop process for removing hydrogen sulfide from the process streams. The amine is occasionally carried over from scrubbers into the plant's sewage system along with steam, and mixes with process water and rainwater.

Effluent from the sewers passes through the plant's biological treatment system for the removal of biodegradable materials. There, some of the amine was apparently being converted by bacterial action into ammonia that remained in the effluent.

To monitor the monoethanolamine and ammonia, and ultimately locate monoethanolamine carryover sources, the laboratory needed to measure from ppm down to ppb levels. Because this could not be accomplished by titrimetry and ion-selective electrode methods, stoichiometric calculations of monoethanolamine levels were made based on the results of total nitrogen and ammonia analyses.

A weakness of this method was that the calculations included the compound error associated with the two nitrogen analyses. As a result, only estimates, not precise data, could be produced.

What the laboratory quickly needed was precise data in order to stem the monoethanolamine and ammonia contamination of the effluent. The only analytical technique in the plant that was sufficiently sensitive and fast was the SCIC unit.

Since by this time the laboratory had already discovered how to improve titrimetry results in wash water analysis, the SCIC system could be safely removed from that application and switched to monoethanolamine and ammonia analysis. In this new application, it was successful in performing the required measurements and helped to locate the carryover sources.

ION CHROMATOGRAPHY

Ion-selective electrodes (ISE's) are commonly used in qualitative and quantitative analyses of ions in solution. However, this technique suffers from interference problems because the electrode responds to ions other than the ion for which it is designed, particularly near its detection limit. Furthermore, ISE has a limited linear working range for analyte concentration.

Atomic absorption is another common method for analyzing samples containing Group I and Group 11 elements in the periodic chart. However, it cannot be used to perform multiple-element analyses, and often requires elaborate sample and standard preparation.

Inductively coupled plasma techniques offer multiple element analysis, but at high cost.

Ion chromatography (IC), on the other hand, is a relatively interference-free, low-cost method for simultaneously analyzing multiple ions over a wide concentration range.

IC is a version of liquid chromatography in which high and low-molecular weight ionic and ionizable solutes are separated according to differences in their relative affinities for an ionizable stationary phase. This ion exchange mechanism facilitates the separation, identification, and quantification of inorganic and organic anions and cations in sample mixtures.

The stationary phase generally consists of one of the following: silica particles, polymer resins, or polymercoated silica particles. These polymers are typically synthetic resins, such as styrene-divinylbenzene copolymers, chemically modified with charged functional groups such as NR3+ for anion exchange or SO3- for cation exchange (where R represents a hydrocarbon substitute).

In IC, conductivity detection is the most commonly used method for measuring analyte concentration because both the carrier phase and the target analyte are ionic. The measured change in conductivity is proportional to both the concentration of the target ion and the difference in the equivalent conductivities of the eluent and target ion.

Other detection methods, including ultraviolet-visible (UV-VIS), fluorescence, and electrochemical detection, can also be employed with IC in special cases. For example, easily oxidizable anions such as nitrite, iodide, thiocyanate, and sulfite can be detected with greater sensitivity and specificity using electrochemical detection.

Accurate conductivity detection requires a carrier phase of low ionic strength. In CSIC, the traditional form of IC, ion exchange is performed on a stationary phase or column packing of 250-750 mm diameter macroporous particles of high ion-exchange capacity, which requires a high-conductivity eluent. This eluent must then be chemically suppressed by passing it through a second suppressor column before it reaches the detector.

The CSIC approach to an analytical problem is often based on a specific column and, while many column types are available, purchasing columns can rapidly become a significant part of the system cost if the laboratory has several types of analyses to perform.

SCIC

SCIC uses a single column packed with small (5-10 Jim) particles of low ion-exchange capacity. This permits the use of low-concentration, low-ionic-strength eluents so that ions can be separated and detected efficiently and rapidly without the need for chemical suppression of the eluent.

Because single-column IC requires a low-conductivity carrier phase, weak acids and bases are used as eluents. The background conductivity of the mobile phase is as low as 100-1,000 msec/cm, enabling detection of very small changes in conductivity.

For anion chromatography, weak organic acids, such as phthalic acid and benzoic acid, with intrinsic conductivities of 30-40 msec-cm2/mol, are typical eluents.

Inorganic anions have larger equivalent conductivities (60-80 msec-cm2/mol), so that when the eluent is replaced by target anions, a positive change in conductivity is measured.

In cation chromatography, the eluent cation is generally H+, which has an intrinsic conductivity of 350 msec-cm2/mol. When inorganic cations, with equivalent conductivities of 50-70 msec-cm2/mol, replace the eluent at the detector, the conductivity decreases.

The detection limit (which is also dependent on temperature and flow rate, as well as on the nature of the target ion and the sample matrix) is more sensitive for cations (0.1-1 nano grams) than for anions (1-10 nano grams) because of the greater difference between the eluent and sample ion conductivities. For an injection volume of 100 ml. ions can be detected at concentrations as low as 10 mgl. for anions and 1 mg/l. for cations.

SCIC permits the use of a broad range of eluents, and it is the eluent rather than the column that is adjusted to obtain optimal separation.

While a choice of SCIC columns is becoming available for an increasing number of applications, one column can be used for a wide range of sample types.

A selection of eluents is also available for a variety of sample types. Each eluent can be further adjusted to obtain the desired resolution in a particular sample.

With the low ionic-strength eluents used in SCIC, the pH can be adjusted or buffered. Weakly dissociating anions, such as CN- and S2-, can be detected at high pH, whereas this would be impossible with CSIC.

The use of SCIC rather than CSIC reduces costs by decreasing the amount of eluent needed for a series of analyses. Because the conductivity of the waste eluent does not change significantly from sample to sample over short periods of time, the low-conductivity eluents can be recycled.

In contrast, it is not possible to recycle the eluent issued from a CSIC system because the chemical composition of the solution is completely changed during chemical suppression.

In SCIC, because eluents must be of low ionic strength and are thus noncorrosive, normal high-performance (or high-pressure) liquid chromatography (HPLC) components can be used. The use of additional detectors is therefore facilitated because a system fitted with HPLC components can easily tolerate the increased back pressure.

The ability to use HPLC components in the SCIC environment results in increased versatility with respect to the detection method, and choice of columns, pumps, tubing, and other wet-section components. In addition to this increased flexibility, the ability to integrate components easily in an SCIC system results in lower operational costs because the user is not restricted to one vendor for the purchase of replacement parts or for instrument refitting.

TEXACO'S UNIT

The SCIC unit at Texaco contains controls and an operating compartment with a hinged door, in a single housing. In the compartment are an injector, column, and detector. The analysis operating range (full scale) and the current conductivity value are shown on a digital display.

The conductivity detector, based on the dual-electrode principle, virtually eliminates the electronic and thermal interference of the measuring signal (noise, drift, etc).

A 1.5 ml. measuring cell and preamplifier are incorporated in a temperature-controlled, thermally insulated, heating block. The eluent is heated to the block temperature in a spiral capillary before flowing into the measuring cell.

The operating compartment is insulated against changes in ambient temperature. Heat control precautions in the system produce a measuring cell accuracy to 0.01 C., an important feature because the signal intensity can vary by approximately 2% per C. change in temperature. To achieve auto zero requires simply pressing a control key which sets the instantaneous output signal to 0 mv. Electronic background compensation operates over the entire measuring range and is independent of the selected sensitivity.

The SCIC unit provides a measuring range from 0.1 to 1,000 msec/cm with sensitivity of 1 x to 2,000 x . Electrical drift is less than 0.001 3% of the selected measuring range/h/C., and electrical noise is typically less than 0.0003% of the set measuring range.

This system can be fully automated in combination with an HPLC pump, an auto sampler to handle a large number of routine samples, an integrator for automatic evaluation, and a line recorder using the chromatograph's analog output signal.

The system can perform in a wide range of applications including those involving analysis of several anions or cations in a single determination at markedly low analyte concentrations, complex matrices, distinct concentration differences, and extremely large concentration differences.

The unit at Texaco contains an HPLC pump, an integrator, a line recorder, and anion and cation columns.

CHLORIDE ANALYSIS AT TEXACO

The procedure for chloride/halides testing at Texaco, using the SCIC unit, is a follows:

A 4 mmol para-hydroxybenzoic acid eluent is prepared in a 4 l. container, for recyclable use. A calibration is performed at the start of each day. It uses 25.0 mg/l. chloride and 30.0 mg/l. sulfate.

For the actual tests, 1 qt. samples are extracted from wash water effluent and delivered to the quality control laboratory. Three ml of straight wash water are withdrawn from each quart sample into a syringe for injection into the ion chromatograph's sample port.

In a three-step operation, the technician switches an injection valve to the fill position, discharges the syringe into the sample port, and then switches the valve to the injection position.

Range, sensitivity, and damping are preadjusted to standard settings. Titration, dilution, and reagents are no longer required as with laboratory wet methods. And no chemical suppression of the eluent is required as with csic.

The SCIC automatically processes a 100-ml. sample through an injector loop, column, and detector for separation and conductivity measurements of the anions. An output signal is sent to the integrator where peak areas are calculated and recorded on a chromatogram.

Total test time is approximately 15 min compared to 45 min or more with conventional methods.

Chloride measurements obtained exhibit 95% accuracy with values of well-defined peaks calculated to 0.1 ppm.

MONOETHANOLAMINE ANALYSIS

The procedure in the monoethanolamine application at Texaco is as follows:

The SCIC system is used with a cation column and a 3 millimols (mmol) nitric acid eluent. Effluent tests are performed every 4 hr, and an average of five additional tests per day are performed on samples taken from valve locations.

A calibration run, performed at the start of the day, consists of 30.0 mg/I. each of sodium, ammonia, and potassium, and 50 mg/l. of monoethanolamine. The calibration chromatogram in Fig. 2 shows clearly defined peaks proportional to specific cation concentrations.

One-quart samples are extracted from a downstream point in the sewer system where effluent streams merge prior to biological treatment. Portions of a 100 l. sample are filtered to remove insoluble organics which may damage the column. As with the chloride analysis, the sample loop is primed with sample from a syringe and the valve is set to inject, which initiates automatic analysis and calculation.

The effluent sample chromatogram in Fig. 3 shows a high level of monoethanolamine at 28.8 mg/l., or 29 ppm. Monoethanolamine and ammonia are detected simultaneously, eliminating the need for ammonia analysis using the ion-selective electrode method, further simplifying wastewater analysis.

Analysis time is less than 11 min, an important feature in the demanding sample load of 4-11 analyses each day.

Precise monoethanolamine data have allowed rapid resolution of amine-carryover problems. The leaks are corrected immediately before concentrations approach the ppm range. This minimizes the amount of monoethanolamine passing into the biological system, thereby minimizing the ammonia produced and released into plant effluent.

In addition to providing accurate data on chloride and monoethanolamine, the SCIC unit at Texaco has been used to analyze samples of other waste treatment processes at the plant, including halogenated hydrocarbons in crude oil and sulfate levels in effluent storage ponds.

In essence, the laboratory has acquired a versatile instrument that is especially useful in petroleum refinery applications. It is more economical than CSIC instrumentation, it has the sensitivity to pinpoint measurements in the low ppm range, and it provides the analysis speed that is always beneficial in the laboratory.

Copyright 1990 Oil & Gas Journal. All Rights Reserved.