Using Electronic Nose for Odor Assessment


In traditional odor assessment methods, subjective human panels are used to assess ambient air samples that are collected in tedlar bags. Following this, a summary report using standard descriptors such as sweet, foul, pungent, and so on is generated. While such reports prove useful, they are not quantitative and cannot be interpreted easily.

Electronic noses, also known as eNoses, can also be used to assess odors by mimicking the olfactory system in humans. These systems use a number of dissimilar yet not specific chemical sensors. However, physical sensors can provide only limited performance due to their overlapping responses, and thus cannot quantify or isolate odor chemistry.

zNose® is an advanced type of electronic nose based on ultra-fast gas chromatography. This system can mimic countless numbers of specific virtual chemical sensors and generate olfactory images that are based on the chemistry of aroma.

This electronic nose is capable of performing near real-time analytical measurements of odors and volatile organic vapors with part-per-trillion sensitivity. Each chemical within an odor is separated and quantified all in a fraction of seconds.

The instrument has a built-in integrated vapor preconcentrator combined with the electronically variable detector to determine vapor concentrations ranging 6+ orders of magnitude. This article uses a portable zNose® system (Figure 1) to evaluate urine odors present in an urban dwelling as well as to measure chemical concentration in these odors.

Figure 1. Portable zNose® gas chromatograph

How the zNose® quantifies odor chemistry

Figure 2 shows a simplified diagram of the zNose® system that has two parts – one part or section employs a solid-state detector, a capillary tube (GC column), and helium gas, while the other part includes a heated inlet and pump that samples ambient air.

Both these section are joined through a “loop” trap, which serves as an injector when located in the inject position or helium section and as a pre-concentrator when placed in the sample position or air section. Operation of the zNose® involves two steps.

At first, ambient air (aroma) is sampled and organic vapors are collected (pre-concentrated) on the “loop” trap. After the air is sampled, the trap is placed into the helium section where the organic compounds collected are administrated into the helium gas.

Following this, separate chemicals come out from the column at characteristic times because the organic compounds travel via a capillary column with varying velocities. As the chemicals exit the column, a solid state detector is used to detect and quantify them.

The collection of sensor data is controlled by an internal high-speed gate array microprocessor that is later transferred to a computer or user interface using a USB or RS-232 connection. A single n-alkane vapor standard is used to achieve calibration.

Compound identification and machine independent measurements are achieved through a library of retention times of known chemicals that are indexed to the n-alkane response (Kovats indices).The sensor spectrum’s time derivative produces the spectrum of column flux, typically called a chromatogram.

Shown in Figure 3, the chromatogram response of n-alkane vapors (C6 to C14) offers an exact measure of retention times. The system is calibrated by graphically defined regions that are indicated as red bands.

These regions give a reference time base against which later chemical responses are indexed or compared. For instance, a response midway between C10 and C11 would possess a retention time index of 1050.

Figure 2. Simplified diagram of the zNose® showing an air section on the right and a helium section on the left. A loop trap preconcentrates organics from ambient air in the sample position and injects them into the helium section when in the inject position.

Figure 3. Chromatogram of n-alkane vapors C6 to C14

Urine odor standards

Measurement and comparison were made for odors from ‘old’ and fresh urine samples. The ‘old’ urine was acquired from bottles that were left in the attic of the house and this was found to be more than a year old.

About 10 ml of this urine was transferred to a septa-sealed 40 ml vials and then allowed to equilibrate at room temperature for a period of 10 minutes before testing the headspace vapors with the zNose® system. Figure 3 shows vertically offset chromatograms of headspace measurements. About 14 separate chemicals were detected and labeled in accordance with their Kovats indices.

The overlaid chromatograms shown in Figure 4 clearly show the variations between fresh and old urine odors. It was observed that the older urine tends to lose most of the more volatile compounds with indices less than 650.

Moreover, some of the mid range compounds such as urine 1196, urine 1057, and urine 1081 are metabolized by microorganisms like bacteria and are not present in the old urine. Despite these variations, urine odors have many of the main chemicals that do not change even after an extended period of time.

Most of the odoriferous and volatile compounds in urine, although not directly detected, are known to be amino acids. This can be detected from the comparison headspace measurements shown in Figure 5 where headspace vapors from one drop of fresh blood are compared against the headspace vapors from urine.

Figure 4. Replicate chromatograms of headspace vapors from fresh and old urine samples

Figure 5. Overlaid chromatograms and peak area counts of headspace vapors from old and fresh (shown in RED) urine

Figures 3 and 5 show the graphically defined bands that are utilized to define regions of retention indices or retention time specific to urine odors. After each region is defined, it functions as a virtual chemical sensor that is specific to an organic compound having a specific retention index. For each virtual sensor, alarm levels can be set in an array of sensors, following which the array aggregate response defines the urine odor of interest.

Description of on-site testing and findings

The aim was to use the urine’s target odor profile to test ambient air in an un-inhabited dwelling. The house in question was a single, large story dwelling with more than 14,00 square feet and included a number of rooms, as illustrated in the floor plan of Figure 6.

There were no personal belongings or furnishings in the house. An air preconcentrator and a portable gas chromatograph (GC) were used to measure the outside air as well as ambient air inside the house. 15 ml is the preconcentrated air sample volume and 10 pg were the least detection level of the GC sensor.

Among all vapor samples tested, the column (a db624) was specifically temperature programmed to increase from 40°C to 160°C at 10°C/second with a data acquisition (chromatogram) time of 20 seconds. Also used was a detector temperature of 10°C. With all these instrument settings, the ensuing detection levels for volatile organic compounds (VOCs) (C4 to C18) in air were seen to be in the low ppb range.

Figure 6. Offset and overlaid chromatograms of headspace vapors from blood and urine (shown in RED) show many of the same peaks since both contain amino acids

Throughout the house, ambient air was assessed by testing a number of locations in all the rooms and also within air ducts where covers had been taken off. More than 150 ambient air measurements were made inside the house (on site) and grab-samples were also tested at an off-site lab.

Throughout the house, trace levels (ppt to low ppb) concentrations of organic compounds that are specific to urine were identified. Vertically offset chromatograms are used to show a portion of over 150 measurements (Figure 7). Hatched red bands define compounds that are specific to urine odors, and numbered peaks specify urine compounds above a random threshold level of 20 counts.

In bathrooms, toilets had been removed. Here, emissions from sewer pipes were analyzed as possible sources of emission. For instance, sewer pipe odor (with the cover removed) is compared with the surrounding room air, as shown in Figure 8.

Within the sewer pipe, the concentrations of the target compounds were observed to be higher when the pipe cover was taken off but did not seem to be a source of emissions when the pipe was covered.

However, localized high concentration sources of odor emission in all the rooms of the house were not detected by on-site testing but rather there seemed to a general presence of urine odor across the dwelling. For example, compared to offset chromatograms, the ambient air tested in the hallway that leads to the master bedroom shows that the same organic compounds are present in urine.

Figure 8. Sewer odors in bathroom by butler’s pantry 10° detector, 30-second sample, 10ps2a1b method, 140°C valve, and 200°C inlet

Figure 9. Typical measurements vertically offset for viewing. 10° detector, 30 second sample, 10ps2a1b method, 140°C valve, and 200°C inlet

Figure 10. Odors in hallway leading to master bedroom. 10° detector, 30-second sample, 10ps2a1b method, 140°C valve, and 200°C inlet

Fertilizers were other sources of organic compounds, which were examined at the site, and had been used on shrubbery and the surrounding landscape. Shown in Figures 10 and 11 are odors from Turf fertilizer and Kelate powder that were compared with both urine odor and the exterior air surrounding of the house.

Trace amounts of the same organic compounds present in urine were detected in the fertilizers. However, other high concentration organic compounds that were not detected in the inside or outside air of the house dominate the odor signature or profile of these fertilizers. As a result, they were not considered as the source of urine odors in the residence.

Figure 11. Turf fertilizer odor compared with urine and outside air measurements.10° detector, 30 second sample, 10ps2a1b method, 140°C valve, and 200°C inlet

Insulation and drywall odors collected at the site, i.e., living room air duct, were analyzed by placing the samples in septa-sealed vials and then directly sampling vapors via a side-ported needle. Also tested were odors from similar insulation and drywall materials from a California building (EST).

All the odors were then compared against the urine odor, which is superimposed in red as shown in Figure 12. Odors from empty vials (blanks) are shown in blue. The prime peaks applicable to urine odors are indicated as hatched bands and their indices are shown at the top of Figure 12.

In the blank runs, background peaks are visible which are attributed to sample needle carryover or ambient air. The vapor concentration from these samples was extremely low (ppt levels) and measurements were made difficult by contamination and carryover.

However, peaks 1498 and 1220 hold importance as they only appeared in urine odors. Based on this factor, the drywall and insulation samples collected from the dining room duct do not seem to be contaminated with urine.

Figure 12. Kelate powder compared with urine and outside air measurements. 10° detector, 30 second sample, 10ps2a1b method, 140°C valve, and 200°C inlet

A high flow vapor preconcentration step was applied to enhance signal to noise in ambient air vapor and sample measurements. In this method, sample vapors are initially preconcentrated in a metal tube that is filled with tenax (SKC) through a high sampling airflow, usually 450 ccm.

Once preconcentration is done, the metal tube is placed inside the GC inlet and the trapped vapors are transferred to the instrument by heating the tube to 220°C. The zNose® system can house metal desorbtion tubes, and through this method increases the ambient air measurement sensitivity by many orders of magnitude.

High airflow sampling vapors were collected from urine contaminated wood shavings, i.e., bagged samples in garage, then placed in a septa-sealed vial, and finally equilibrated for a period of 10 minutes before testing the headspace vapors.

The bottom trace of Figure 13 shows the headspace chromatogram outcomes for sample No. 68332. For comparison purposes, four vertically offset chromatograms from the urine sample are also shown. Peaks appear at indices of 1498 and 1220, strongly suggesting that this particular wood sample is indeed contaminated with urine.

Figure 13. Odor s from insulation and drywall taken from dining room air duct.10o detector, 30 second sample, 10ps2a1b method, 140oC valve, and 200oC inlet

While urine contamination can be indicated by screening with a fast GC, this conclusion must be supported by separate lab testing with a GC/MS and by identifying the compounds of interest. The huge number of compounds identified at ppt concentration levels, make independent validation very important.

The peak with1498 index is assumed to be phenylacetic acid — a standard metabolite released in urine and exhibits an odor commonly described as sweet urine.

Due to its high boiling point of phenylacetic acid (265°C), it readily sticks to surfaces and cannot be removed easily through ventilation. Urine also contains many other metabolic compounds and even more can be produced over time because of the presence of bacteria in the ambient environment.

Figure 14. Odors from wood sample No. 68332 compared with urine odor. 20° detector, 2 min sample 450 ccm into tenax preconcentrator, 220°C desorber, 5ps2a1b method

Odors from insulation samples collected from the dining room air duct were again tested with a high flow preconcentrator. The results of this are illustrated in Figure 14 and compared with urine odor as well as EST insulation samples. Since the two peaks 1498 and 1220 are not present, the insulation does not seem to be contaminated with urine.

Figure 15. Odors from Wool Insulation compared with urine odor. 20° detector, 2 min sample 450 ccm into tenax preconcentrator, 220°C desorber, 5ps2a1b method

Using a high flow preconcentrator, odors from drywall samples collected from the dining room air duct were also tested again. These results are illustrated in Figure 15 and compared with urine odor and EST drywall samples.

There is very low background interference in the blank chromatogram (in red) from preconcentration of vapors in a blank vial. Since the two peaks 1498 and 1220 are not present, the drywall does not seem to be contaminated with urine.

Figure 16. Odors from wallboard sample compared with urine odor. 20° detector, 2 min sample 450 ccm into tenax preconcentrator, 220°C desorber, 5ps2a1b method


Using a portable, ultra high speed gas chromatograph, on-site and near real time chemical profiling of urban indoor air can be carried out. Air quality can be quantitively measured on-site and the chemical signature of nuisance odors can be detected.

One easy way to screen target compounds is to index the retention times for compounds of interest using an n-alkane vapor standard. GC testing of ambient air combined with the human sensory data provides a more objective means of categorizing and quantifying odors.

Results of tests, which were conducted on an urban dwelling for the presence of organic compounds related to urine odors, have demonstrated that these odors were indeed present within the inside air of the dwelling.

Ambient air across the house was assessed by testing in various areas within all the rooms and also within the air ducts. More than 150 ambient air measurements were carried inside the dwelling (on site) and also grab-samples were tested at an off-site lab. Generally, the detected concentration of organic compounds had very low (ppt to low ppb) concentrations.

Chromatogram peaks that were also detected in urine vapors were likewise detected in ambient air inside the dwelling. In bathrooms, where toilets had been taken off, emissions from sewer pipes were tested as possible sources of emission.

The concentrations of compounds of interest inside the sewer pipes were found to be low and did not seem to be a source of odors inside the dwelling. Yet, the presence of huge numbers of background peaks at these low levels of concentration made it difficult to be certain whether the peaks were in fact metabolic compounds from urine without an independent confirmation.

Fertilizers were other sources of organic compounds that were analyzed at the site. These compounds had been used on shrubbery and the surrounding landscape.

While fertilizers did show some trace amounts of the same organic compounds present in urine, their chemical signature or odor profile was purely governed by other high concentration organic compounds. These organic compounds were not detected in the inside or outside air of the house.

Odors from wood were tested positive for urine contamination. This wood had been removed from the house and kept in the garage. No urine contamination was detected in drywall and insulation samples from the living room air duct.


  • Tenax absorption tubes can be used to collect and test high volume air samples from within the house
  • Independent validation can be achieved with split samples
  • On-site testing protocol should be implemented to determine the concentration of organic compounds of interest in ambient air inside the dwelling
  • Acceptable levels of target compound concentration should be determined for remediation

With the help of the zNose® tool, remediation and environmental engineers can achieve the precision, speed, portability, and accuracy required for economical on-site odor measurements.

Since such measurements are based on popular chromatographic methods, they can be easily corroborated by independent lab testing. Established by trained sensory panels, acceptable odor levels can be employed to substantiate remediation efforts by quantitative and objective on-site test.


Produced from materials originally authored by Edward J. Staples, Electronic Sensor Technology, CA, USA.

About Electronic Sensor Technology

Electronic Sensor Technology

Electronic Sensor Technology, Inc has developed and patented a breakthrough chemical vapor analysis process. This process applies gas chromatography calculations and technology toward a wide variety of industries, including Homeland Security, Life Sciences, Chemical and Petrochemical, Food & Beverage and Environmental.

How does gas chromatography work? With rapid, accurate analysis of chemical odors and vapors, this patented technology helps to provide real-time analysis for quick response solutions.

Sponsored Content Policy: publishes articles and related content that may be derived from sources where we have existing commercial relationships, provided such content adds value to the core editorial ethos of News-Medical.Net which is to educate and inform site visitors interested in medical research, science, medical devices and treatments.

Last updated: May 22, 2017 at 3:23 PM

Other White Papers by this Supplier