908-788-5550

Note 65: Determination of Ethylene by Adsorbent Trapping and Thermal Desorption - Gas Chromatography

Home

By Eric D. Butrym* and Thomas G. Hartman

* Scientific Instrument Services, Inc. 1027 Old York Rd. Ringoes, NJ 08551

Department of Food Science, Cook College, Rutgers University, New Brunswick, NJ 08903

Center for Advanced Food Technology, Cook College, Rutgers University, New Brunswick, NJ 08903

Presented at PittCon 98, New Orleans, LA, March, 1998.

Abstract

Ethylene is of great economic and practical importance due to its use in polymers and its pronounced biological activity in plants. Low levels of ethylene (<1ppm) can produce changes in plant growth and senescence through hormonal activity. In the past, methods for detecting low levels of ethylene relied on trapping the gas with Mercuric salts or the application of specialized photoacoustic detectors. Direct Headspace methods have also been used for higher concentrations. Major limitations of these methods include the generation of toxic by-products, low sensitivity, low trapping efficiency, and high cost.

Thermal Desorption is a reagent-free analytical method that is applied here to ethylene detection. Carbosieve SIII is used to adsorb ethylene from prepared standards and humid air. Air samples are drawn through glass-lined steel tubes filled with the trapping material. Ethylene and other light hydrocarbons trapped from the air are then desorbed in a single step directly into a GC inlet for analysis. A 2 meter micropacked column (Carboxen 1004 Supelco, Inc.) is used for analysis with flame ionization detection.

Large concentration factors in the samples make routine detection of less than 1 ppb ethylene possible without major modification of the chromatograph. Sample size is limited only by the breakthrough volume of the adsorbent. This method represents a major improvement in the availability of precise, low-level ethylene measurement without expensive extra equipment or unwanted toxic by-products. Additionally, the method is useful for other light hydrocarbons such as propylene and isoprene that must be accurately measured in very low concentration.

INTRODUCTION

Detection of ethylene at trace (< 10 ppm) levels is necessary to study the effects of this hormone on plant development as well as monitor concentrations of the gas in commercial processes. In the past, Gas Chromatography (GC) with Flame Ionization Detection (FID) combined with static headspace analysis has been the most popular method for ethylene measurement.(1,2) However, this method is limited by the amount of gas that can be reliably introduced into the GC inlet. Even recent advances in column technology have kept detection limits for ethylene in the low ppm range without the use of expensive specialized detectors. One serious deficiency is the tendency for linearity to fall off dramatically at the lower limits of detection.

An alternative to static headspace measurement is the dynamic enrichment, or trapping technique. This has the advantage of concentrating analytes to detectable levels, thereby increasing detection limits. Solutions of Mercuric Acetate or Mercuric Perchlorate have been used as a non-specific trap for olefins due to the complexation reactions they undergo. Addition of chloride salts to these solutions degrades the complex and releases ethylene for headspace analysis.(3) These methods suffer from low trapping efficiency, considerable sample preparation time and the necessity of using toxic reagents.

The use of adsorbent materials for trapping ethylene has been largely overlooked, presumably due to the difficulty of finding materials which are aggressive enough to trap the highly volatile gas while not allowing water to interfere with the subsequent analysis. Due to the large sample volume needed for trace level analysis and because of the nature of the systems being studied (plant material, biofilter effluent, etc.), significant amounts of water are expected to be trapped along with ethylene. Thermal desorption technology combined with adsorbent trapping for light hydrocarbon analysis have in the past only been available in the form of specialized analyzers costing many thousands of dollars.

Materials and Methods

Carbosieve SIII was selected as the adsorbent material for the traps due to its relatively high affinity for light hydrocarbons. Porous polymer adsorbent materials have little affinity for gaseous hydrocarbons, and polymer type molecular sieves are unstable at the elevated temperatures needed to desorb ethylene as a discrete peak into the chromatograph. Carbosieve SIII displayes the highest affinity for ethylene as measured by breakthrough volume (L/g of adsorbent) of all the materials commonly used for thermal desorption (Figure 1). The breakthrough volume for ethylene on Carbosieve SIII was determined experimentally to be approximately 4 L/g at 20°C. (4)

Figure 1 - Hydrocarbon Brakthrough volumes

Figure 1

Water is problematic with respect to normal thermal desorption analysis, because cryofocussing is generally used to immobilize analytes during the desorption step, thereby improving chromatography dramatically. Excess water in the sample freezes under cryotrapping conditions and results in decreased sensitivity and in some cases may extinguish the detector flame. Here the problem has been circumvented by the use of a chromatography column with a high loading capacity and a packing which enables resolution of ethylene from other light hydrocarbons and gasses as well as water in the sample. For this work, a 2 meter micropacked (1/16 inch O.D. stainless steel) Carboxen 1004 (Supelco Inc. Bellefonte, PA) was chosen. This choice of column combined with careful temperature programming allows samples to be injected without cryotrapping and provides excellent resolution of ethylene from water and other compounds present.

Standard mixtures of ethylene and helium (10 ppm and 100 ppm w/w, Scott Specialty Gasses Inc.) were used to prepare 1L dilutions of ethylene in dry nitrogen (JWS Inc. Piscataway, NJ) at various concentrations. The samples were prepared in Tedlar® sample bags (SKC Inc. Eighty-Four, PA) immediately before use.

Carbosieve SIII adsorbent (0.8 grams) (Supelco Inc. Bellefonte, PA) was packed into 4 mm I.D. glass-lined Thermal Desorption Tubes (SIS Inc. Ringoes, NJ) and secured with glass wool. The tubes were conditioned for three hours at 310 °C under a nitrogen purge. Standard dilutions of ethylene were drawn through the tubes at a rate of 100 mL/minute. Each sample was then analyzed using a Short Path Thermal Desorption System model TD-2 (SIS Inc. Ringoes, NJ) operated in manual mode and a Varian model 3400 Gas Chromatograph with FID detection (Varian Assoc. Walnut Creek, CA). Varian Star Chromatography software (version 4.0) was used for data collection and calculation of standard curves. Identities of ethylene and other compounds in the standards were confirmed by Mass Spectrometry in separate analyses. (Data not shown)

The samples were purged with carrier for 30 seconds before injection, and were desorbed for two minutes at 300°C into the injector which was held at 220°C . Desorption and the GC run were started simultaneously. The GC was programmed to remain at 100 °C for two minutes, then ramped to 225 °C at 30 °C/ minute. The FID detector was held at 300 °C and was supplied with 30 mL/ minute makeup gas. The column head pressure was kept at 46 psi , equivalent to approximately 12 mL/minute column flow.

Four sample replicates were analyzed at five different concentration levels spanning a three-log dynamic range. As a test of ruggedness, the method was used to analyze an actual sample from the effluent of an ethylene-degrading biofilter (Fig.5.).

Results

Figure 2

Figure 2

Figure 2 contains representative chromatograms for the five concentration levels used. In each chromatogram, Ethylene can be seen as a well-resolved peak at retention time ~ 5.6 minutes.

Figure 3

Figure 3

The standard curve shown in Figure 3 was generated using all five concentrations and displays excellent linearity over a wide dynamic range (R2 = 0.997). A separate curve (Figure 4) was generated using only the four lowest concentrations, and this also demonstrated good linear characteristics (R2 = 0.972) with only a minor change in the slope. Relative standard deviations for concentrations of 50, 10, 5, 1 and 0.5 ppb were 9.1%, 14.4%, 11.3%, 22.4% and 57.9% respectively. These results demonstrate an improvement over standard methods in linearity at the lower concentrations; however, the decreased precision at the lowest level indicates that the measurements are at or near the reliable limit of detection.

Figure 4

Figure 4

The sample taken from the ethylene-degrading biofilter was analyzed in order to evaluate the method under more stringent conditions. A chromatogram is shown in Figure 5. Biofilter efflutent was passed through a tube containing Mg2SO4 dessicant into a Tedlar® sample bag ( SKC Inc. Eighty-Four, PA). A 1L sample from the bag was drawn through an adsorbent trap in the same manner as described above for the standards. Analytical methodology was identical to that used for the standards, except that a Varian model 4200 integrator was used for data collection.

Figure 5

Figure 5

Conclusions

  • A simple, reliable and accurate method has been developed for the analysis of trace levels of ethylene in air.
  • The method avoids the use of expensive specialized equipment, permanent GC modification, and generation of toxic by-products.
  • Improvements to the limit of detection are possible by merely increasing sample size.
Acknowledgments

Biofilter sample courtesy of Jyoti Tambwekar, Department of Environmental Sciences, Cook College, Rutgers University.

This work was funded by the New Jersey - NASA Specialized Center of Research and Training (NJ-NSCORT) for Bioregenerative Life Support Systems.

References

1. Ethylene in Plant Biology. 1973. Abeles, F.B. Academic Press, New York.
2. Ethylene. 1987. Saltveit, M.E. Jr., and Yang, S.F. in Principals and Practice of Plant Hormone Analysis. Vol. 2. L. Rivier and A. Crozier eds. Academic Press, New York, NY. 367-401.
3. Manometric Determination of Low Concentrations of Ethylene. 1952. Young, R.E., Pratt,H.K., and Biale, J.B. Anal. Chem. 24(3):551-555.
4. Selection and Use of Adsorbent Resins for Purge and Trap Thermal Desorption Applications. 1995. Application Note No. 32 Scientific Instrument Services, Inc. Ringoes NJ.