Take a deep breath; then exhale fully

You're probably aware that law-enforcement personnel no longer rely on subjective tests, such as asking a driver to walk a straight line, focus on a moving finger, or meet his or her index fingers in front of his or her nose to assess whether the driver has been operating a vehicle under the influence of alcohol. Instead, they ask the driver to breathe into an instrument that gives a clear readout of the presumed BAC (blood-alcohol concentration). The virtue of this instrument is that it is objective and noninvasive, but there are many controversial points about its accuracy in the field with different body types and in different situations.

The instrument is commonly referred to as the Breathalyzer, and although Draeger Safety Inc (www.draeger-breathalyzer.com, also spelled Dräger) owns this name, people use the term "Breathalyzer," like Kleenex and Band-Aid, as a generic name for similar instruments made by other companies and that may use other sensing techniques. Robert Borkenstein, PhD, of the Indiana State Police and RN Harger of the Indiana University School of Medicine in 1954 invented the Breathalyzer; it has been in widespread use for decades.

The principle behind the Breathalyzer is that the concentration of alcohol (strictly speaking, it's ethyl alcohol, also called ethanol) in the air your lungs exhale, or the alveolar air, relates closely to the concentration of alcohol in your bloodstream, which feeds across the membranes of the lung's air sacs, or alveoli. This ratio, called the partition ratio, varies among and within individuals, as well as due to operating conditions, such as temperature; it has been documented as low as 1900-to-1 and as high as 2400-to-1. However, a statute sets the value at 2100- to-1, meaning that, for assessment purposes, 2100 ml of alveolar air has the same amount of alcohol as 1 ml of blood.

Therefore, if you can sample the exhaled air and measure its alcohol concentration, you can work backward to determine the BAC. Until recently, a BAC of 0.10—meaning 0.1g of alcohol per 100 ml of blood—was considered the boundary of alcohol impairment for legal purposes. Most states have lowered that level to 0.08.

In a Breathalyzer test, the person under test breathes deeply into a tube, which captures the air sample and passes it to a sensor. In the original design, this sensor uses a chemical reaction in which a reddish-orange compound changes to a greenish color (see sidebar "What color is your breath?"). The amount of color change indicates the concentration of alcohol in the exhaled air and, thus, by inference, the BAC. A photocell assembly measures this change by using parallel paths—one to measure the light's passage through the reaction results and one to directly measure the light source as a calibration baseline.

The chemical mix, housed in a sealed ampoule or vial until use, is critical. It must have the correct reactant concentration and correct absolute amount at the right temperature, and the device must control the input flow so that it acts only on the appropriate part of the breath sample when it the chemical mix is activated—that is, the machine breaks the ampoule seal. The device must also purge the entire air path—input tubes, connection tubes between input tubes and ampoule, and path to a reference ampoule used for color comparison—of previous sample residual as well as excess humidity.

Two other sample-assessment techniques now use IR (infrared) spectroscopy and fuel-cell current to measure alveolar alcohol concentration. Each method attempts to overcome the operating constraints of the chemical-reaction approach. The methods provide quicker results and suffer fewer effects of instrument temperature, aging, drift, and other factors that can affect calibrated accuracy by 10 to 20%, depending on whom you ask and whether you are the prosecutor or the defendant in a DUI case.

IR spectroscopy uses the well-established analytical technique of assessing which wavelengths of light and, thus, energy the bonds within molecules absorb as this light passes through the substance under analysis (Reference 1). The system uses a broadband IR source, such as a quartz lamp, although some designs use IR LEDs. The lamp's output passes through the breath sample, and a mechanically driven filter wheel with narrowband optical filters selects wavelengths of interest to pass on to a photocell. The photocell generates an output proportional to the amount of energy at those IR wavelengths that the sample does not absorb.

For this approach, you must calibrate the unit to look at both the outputs after the sample and the lamp's initial output energy at these wavelengths, which changes with age, temperature, and other factors. Ethanol has absorption peaks at 3.39, 3.48, 7.25, 9.18, 9.50, and 11.5 microns, as well as some secondary peaks. Manufacturers of commercial IR-based instruments do not use all these peaks, because doing so would be too expensive and complex; instead, they measure one or two of them and try to optically or electronically filter out potentially interfering peaks from substances such as acetone. The IR approach does not require wet-chemical reactions, and temperature affects it less than it does the chemical-reaction sensor.

The most recent approach, available in some commercial units from Draeger, including the handheld Alcotest 7410, and other sources, uses a fuel cell as the alcohol-to-electrical transducer—but not in the way the mass press talks about fuel cells. In this design, the sample gas passes through a tube that is coated with powdered platinum and has a porous layer that is saturated with an electrolytic solution. As the gas passes through this tube, platinum oxidizes the ethanol and produces acetic acid, protons, and electrons, with a current flow of approximately 1 μA; the device then amplifies and measures this current. The device heats the fuel cell to a controlled temperature for both consistency of results and prevention of internal condensation, which would interfere with cell operation.

If each technique has virtues and vices and different sources of potential error, a logical question is: Why not combine two or more into a single instrument and then display both results? The Draeger 7110 MKIIII-C, which the vendor claims is the most advanced unit you can take into the field, does (Figure 1). It is the size of a briefcase, weighs 16.5 lbs (7.5 kg), and combines both IR and fuel-cell sensing in one unit. The IR operates at a single wavelength of 9.5 microns and omits the filter wheel to reduce the size and weight of the system; the vendor claims that advanced signal-analysis techniques compensate for this one-wavelength approach and minimize nonethanol responses.

Note that an accurate reading involves much more than the ability of the sensor or transducer to translate ethanol concentration in the air sample to an accurate electrical signal, perform internal calibration, and purge residual samples. Among other requirements, the system and operator must make sure that the person under test provides enough air and that the air is from the end of the exhale phase (a deep-lung sample). For these reasons, the Draeger unit requires the air sample it tests to exceed a minimum flow rate, for at least a set time period, at a required volume, and with a certain slope in flow level, so that the exhale cycle is at the right point. (For a discussion of some of the sources of error, as well as legal decisions related to Breathalyzers, see Reference 2. I cannot assess the bias, if any, of the author, who is a professor of pharmacology.)

Reference 3 contains an overview of Breathalyzer technologies. Breathalyzers are not just for law enforcement or laboratory use, although the devices that law enforcers use require considerable operator training for a valid and legal reading. You can now buy personal units to carry in your pocket, as well as coin-operated ones to install in restaurants and bars at sites such as www.breathalyzer.net. However, I have been unable to find any details on their accuracy, repeatability, or overall performance.

Author Information

You can reach Executive Editor Bill Schweber at 1-617-558-4484, fax 1-617-558-4470, e-mail bschweber@edn.com.