Note 75: An Apparatus for Sampling Volatile Organics From LivePlant Material Using Short Path Thermal Desorption

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 EAS, Somerset, NJ., November 1998


Most headspace analysis of plant material is performed on culled or harvested plant parts.1,2 While adequate for most flavor and fragrance work (as these parts are the source of commercial additives and extracts), the act of harvesting the material can introduce artifacts that interfere with the study of normal biological activity. Ideally, the examination volatiles should be undertaken on living undisturbed plants because stresses, as physical damage or temperature fluctuations, can cause dramatic changes in the quality and quantity of volatile emissions.3

This work describes the design and operation of an apparatus for the collection of volatile organics from living plants. Attention was focused on minimizing physical damage to the plant being sampled, and eliminating artifacts arising from materials included in the apparatus itself. The equipment consists mainly of a specially made split glass flask that can surround the plant or plant part to be sampled and provide a known volume for quantitative calculations. Associated equipment includes a vacuum pump, flow control devices, a reservoir of clean air to replace the sampled headspace, adsorbent traps for retaining the sample, and a means of sealing the apparatus. Data were gathered from several tomato plants at different stages of development, and rates of emission for volatile organics are shown. Analysis was by Thermal Desorption/Gas Chromatography (TD/GC) with qualitative confirmation using retention indices and Mass Spectrometry. The overwhelming majority of the compounds detected from live tomato plants were terpenoids, and rates of total organic emissions varied with plant age.


Static and dynamic headspace sampling have been applied routinely to the analysis of volatile compounds from plant tissue. These methods represent the most direct way of qualitatively and quantitatively measuring the constituents of particular plant aromas as well as non odor-active components of volatile plant emissions. Flavor and fragrance researchers may be the most frequent users of these techniques, however they are also used more and more often by those investigating biologically active principals as well as individuals interested in the environmental impact of volatiles from plants.1,2 An example of the latter type would include the study of the contribution of plant volatiles to changes in both in- and outdoor air quality, while the former includes measurement of plant hormones and signaling agents, as ethylene and methyl salicylate.4,5

The focus of this work is to develop a sampling system that can accurately reflect the volatile material emitted by living plants, without introducing sampling artifacts by damaging them. Plants can react to stress by altering the type and amount of volatile material they emit, as in the case of Solanacea releasing volatile terpenes when attacked by insects, or Tobacco plants emitting Methyl Salicylate after being infected with Tobacco Mosaic Virus.6,5 Most plant tissues begin to break down enzymatically soon after harvest. Some enzymes are particularly effective at liberating volatile material from lipids, for example. Retaining the integrity of the live plant is therefore crucial to characterizing its normal volatile profile. Another objective of the work is refining the sampling methodology to eliminate background contamination and obtain cleaner samples. Many of the systems in use for collecting plant volatiles are inadequate not only because they require the vegetation to be harvested prior to sampling, but they are also constructed of materials that are incompatible with artifact-free analysis. Plastics, in particular, are often used in purging vessels and for tubing to direct airflow, etc. Stabilizers and other volatile additives are common artifacts in samples from such systems. By sampling living plants directly with a system made of inert materials, better data can be obtained for environmental as well as biological research. Direct measurement of the rate of volatile emissions, for example, is of interest to the space program, where plants will eventually be included in self-sustaining life support systems.7


Figure 1

Figure 1

Custom manufactured glass enclosures (Kimble, Inc. Vineland,NJ) are used to surround the plant part to be sampled. The enclosure is made of two identical hemispherical parts with a ground glass sealing surface where they mate. An opening along the edge of each hemisphere allows the stem of the sample to protrude and remain attached to the rest of the plant throughout the sampling procedure. A size 24/40 tapered ground glass sampling port is located on each half of the enclosure and when the unit is assembled, these ports are directly opposite one another. The nominal volume of the enclosures used in most of the experiments was 5000 ml. This value was corrected for the volume taken up by the sample (see Calculations, below). The enclosure is supported by ringstand clamps, and is sealed around the plant stem with PTFE tape or Parafilm. These materials have been evaluated and found to be free of volatile components that might contaminate the sample at the temperatures normally encountered. Figures 1&2 illustrate the apparatus empty, and while sampling respectively.

Figure 2

Figure 2

Air samples are drawn from the enclosure through one of the sampling ports. The air is drawn by a vacuum pump (Fisher Scientific) and is regulated by a fine metering valve (Nupro Inc.). The sample rate is 100 ml/minute and is verified before and after sampling with a digital flowmeter (Alltech Inc.). Sampled air is replaced from a 10 liter Tedlar reservoir (SKC Eighty-four, PA) filled with compressed air (JWS Piscataway, NJ) conditioned with a hydrocarbon trap (Supelco Bellefonte, PA). The samples are drawn through a glass-lined stainless steel desorption tube filled with 100 mg Tenax® TA (SIS Inc. Ringoes, NJ). The tubes are tightly capped immediately after sampling, and the plant part is severed at a point even with the edge of the enclosure and dried at 70 ºC for 48 hours before being weighed.

Two sets of plants were evaluated over a period of approximately 10 weeks. Duplicate series of five samples were taken at intervals of 7 to 10 days beginning at 25 days after planting (DAP) and continuing through approximately 75-80 DAP in each experiment.


Sampling tubes were spiked with an internal standard solution of approximately 1.0 µg/µl each d-8 toluene and d-8 naphthalene in methanol. One tube from each sample set was spiked at a higher level of approximately 10.0µg/µl to allow estimation of higher sample concentrations. The tubes were purged for approximately 15 minutes with 30 ml/min dry nitrogen to remove solvent from the standard spike and water.

Samples were analyzed with a SIS model TD-1 Short Path Thermal Desorption system on a Varian 3400 Gas Chromatograph (GC). Detection was by Flame Ionization (FID), and data were collected using a Varian Star chromatography data system version 4.0. GC parameters are summarized below:

Column: 60 m DB-1 (J&W Folsom, CA) 0.53 mm, 0.5µm film thickness

Injector : Split 25:1 or 300:1, Temperature: 250 ºC

Detector: FID, Temperature: 250 ºC

Desorption Conditions: 250 ºC for 5 minutes

Oven Program:

Initial: -20 ºC; 0 minutes

Ramp: 10 ºC /minute

Final: 280 ºC; 5 minutes

Figure 3

Figure 3

Typical chromatograms of an apparatus blank, and a tomato headspace sample are shown in Figure 3.

Gas chromatography-Mass spectrometry (GC-MS) was used for qualitative confirmation of one sample from each set. GC parameters were identical to those used for FID detection, except the column was 0.25 mm, 0.25 µm film thickness; and the sample was split 100:1 at the injector. A Finnigan MAT model 8230 mass spectrometer with a Finnigan SSX data system was used for MS detection. Electron Impact (EI) ionization (70 eV) was used, and the instrument was set to scan from 35 to 350 amu at 0.1 seconds per decade, with an interscan time of 0.8 seconds.

Standards containing hexanal, p-cymene, limonene, terpineol, and isocaryophyllene in addition to internal standards were spiked onto packed desorption tubes, purged in the same manner that the samples were, and then analyzed to create five standard curves. With the exception of p-cymene, which was saturated at the higher spike levels, all the curves had similar slopes, and all had response factors close to unity. The chromatograms were therefore calculated with a single response factor of 1.0 relative to the d-8 toluene internal standard, and the results should be considered semi-quantitative.

Table 1. Results of a Typical Tomato Headspace Analysis

Ret. Time Area RRt Ret. Index ng Compound
6.35 1990 -1.54 unknown 3 unknown
7.89 655827 0 748 991 Internal Standard
8.57 8865 0.67 771 13.4 toluene
9.65 7526 1.76 812 11.38 silicone
9.9 11636 2.01 822 17.58 unkown
10.42 1320 2.53 843 1.99 unk. terpene
10.78 2785 2.88 858 4.21 aliphatic HC
11.55 1722 3.66 891 2.61 unk. terpene
11.69 140256 3.8 897 211.94 alpha pinene
12.27 163657 4.38 923 247.29 cymene isomer
12.35 1024 4.46 927 1.55 cymene isomer
12.44 3255 4.54 930 4.92 sabinene
12.64 10222 4.75 940 15.45 2-carene
12.89 1113639 4.99 951 1682.79 alpha phellandrene
13.02 1359 5.13 957 2.05 carene isomer
13.06 1303 5.16 959 1.97 carene isomer
13.13 68050 5.24 963 102.83 alpha terpinene
13.34 *4761605 5.45 972 *7195.12 terpinene isomers + limonene
13.5 1224 5.61 980 1.85 limonene + cis-ocimene
13.62 1841 5.73 986 2.78 ocimene
13.84 5060 5.95 996 7.65 gamma terpinene
14.36 11897 6.46 1021 17.98 dimethyl styrene isomer
14.85 2414 6.96 1046 3.65 oxygenated terpene
15.33 1047 7.43 1070 1.59 2,6 dimethyl styrene
15.7 633456 7.8 1090 957.2 Naphthalene I.S.
16.33 1319 8.43 1122 1.99 1-decene
18.39 2046 10.5 1235 3.09 delta elemene
19.63 6167 11.73 1307 9.32 isocaryophyllene
19.97 1004 12.07 1327 1.52 unk sesquiterpenes
21.67 1373 13.77 1420 2.07 unknown
30.78 1067 22.88 unknown 1.62 unknown
32.48 4284 24.59 unknown 6.47 unknown
Total ng in tube: 9570.3
Total ng in chamber: 232796.5
Total ng per g Foliage: 17244.2
Emission Rate per gram: 20.7
(Rate in m g/hr/g dry weight)
mid-sample interval = 5 minutes
* Corrected


The d-8 toluene internal standard was used as a retention time reference, because it is an easily recognizable peak in the chromatograms, and its retention time was also very stable throughout the experiments. Relative retention times were calculated for all peaks with over 1000 area counts by dividing the retention time of each peak by that of the toluene internal standard. Peaks with areas under 1000 counts were below the quantitative limit of the method. A standard containing thirteen straight-chain hydrocarbons was run using the same GC oven program as the samples, and retention indices were generated using the relative retention times. The retention indices were used to help correlate the peaks in the GC chromatograms with those that could be identified in the GC-MS runs. The d-8 naphthalene internal standard was used to correct any minor spread in the retention indices by observing its relationship to the toluene internal standard.

Because of the large quantities of some individual compounds found in the tomato headspace, it was necessary to spike two different levels of internal standards and to run one sample from each set at a higher split to avoid overloading the largest peaks. In this way, the dynamic range of the method was extended to include compounds that are present in the highest quantity, without sacrificing sensitivity to those present at trace levels. In order to perform accurate calculations for total organic emissions, the areas of overloaded peaks must be estimated in the low-range chromatograms by using a ratio of some other peak from the sample to the peak in question. The peak to be used in the ratio must be present and on scale in both high- and low-range chromatograms. In these experiments, the one peak that was consistently overloaded was an isomer of terpinene mixed with lower levels of limonene. The peak used as a reference was identified as a -phellandrene. Care was taken to use the ratio from the same plant and sampling interval when possible.

Once all the peaks in the chromatogram had been quantified, summing the amounts of individual compounds, exclusive of internal standards and known artifacts, gave the total amount of organic material by weight that was trapped on the adsorbent tube. This value was applied to the sample volume (generally 200 ml) to yield the concentration for that sampling interval. This was considered equivalent to an instantaneous concentration at the mid-point of the sample. The time between the onset of sampling (t0) and the midpoint of a sample was termed the 'mid-sample interval' for that sample. Thus, since the samples were uniformly 2 minutes long, the mid-sample interval (in minutes) for any consecutively numbered sample 'n' was equal to 2n-1. The concentration at the mid-sample interval was then applied to the volume of the sampling enclosure after it had been corrected for the space taken up by the sample itself. This figure, when divided by the mid-sample interval, yields an emission rate in mg/min. The emission rate per gram of sample is easily calculated.


The samples from experimental Set I displayed total organic emission in the range of approximately 12 to 120 m g/hour/g, with an average value of 49.15 m g/hour/g. An apparent periodic trend, with maxima at 40 and approximately 70 days (coinciding with plant flowering and fruit ripening respectively) was not confirmed by the second set of samples. Set II showed a more or less steady decline in emission rate throughout the life of the plants, with somewhat lower emission rates between 5 and 70 m g/hour/g, and averaging 29.55 m g/hour/g. A foundation for fluctuations in the emission rate remains elusive, although environmental factors such as temperature and humidity can be investigated by plotting both graphs against the date (the experiments were not run wholly concurrently). The emission rates are plotted against plant age in Figure 4.

 Figure 4

As is apparent from the blank chromatogram, (Figure 3.) the apparatus is free from interfering contaminants. The two large peaks are the internal standards, and the three smaller peaks have been identified as siloxane compounds from the GC column. The apparatus, however, may still contribute to plant stress by inhibiting transpiration. It was noted that the relative humidity in the chamber routinely rose to near saturation during sampling. The entire sampling interval, however, lasted only 10 minutes.


A simple, rugged device has been developed that can aid in the analysis of volatile components emitted by live plants. In duplicate experiments, the apparatus allowed the collection of data that yielded emission rates for total volatile organic material from live tomatoes. The chromatographic data was largely free from interference related to the sampling apparatus. Much information remains in the data, and future work will examine relationships between plant age and qualitative sample composition as well as the potential for volatile compounds to be used as monitors of plant stress.


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


1. Bicchi, C. and Joulain, D. 1990. Headspace Gas Chromatographic Analysis of Medicinal and Aromatic Plants and Flowers. Flavor and Fragrance Journal 5, 131-145.

2. Buttery, R.G., Ling, L.C. and Light, D.M. 1987. Tomato leaf Volatile Aroma Components. J. Agric. Food Chem. 35, 1039-1040.

3.  Heath, R.R., and Manukian, A. 1994 An Automated System for use in Collecting Volatile Chemicals Released from Plants. J. Chemical Ecology 20:3,593-608.

4. Batten, J.H., Stutte, G.W. and Wheeler, R.M. 1996. Volatile Organic compounds Detected in the Atmosphere of NASA's Biomass Production Chamber. Adv. Space Res. 18:4/5, 189-192.

5.  Shulaev, V., Silverman, P.and Raskin , I. 1997. Airborne Signalling by Methyl Salicylate In Plant Pathogen Resistance. Nature. 320, 718-721.

6. Alborn, H.T., Turlings, T.C.J., Jones, T.H., Stenhagen, G., Loughrin, J.H. and Tumlinson, J.H. 1997. An Elicitor of Plant Volatiles from Beet Armyworm Oral Secretion. Science. 276, 945-949.

7. Lange, K.E., Lin, C.H. and Barnes, C. 1996 Advanced Life Support Program Requirements Definition and Technology Development Needs (Preliminary). Document no. CTSD-ADV-XXX. Crew and Thermal Systems Division, NASA Johnson Space Center, Houston, Texas. A1-A4.

Tenax® is a registered trademark of Buchem BV.