Breath Test Defenses

11.1 Breath Tests

For a Per Se limit DWI charge, or for an ALS suspension to kick in, the test must be done with the Intoxilyzer 5000 (Not a PBT). In New Hampshire, the Intoxilyzer 5000en[1] is the only approved instrument for breath tests for the Per Se law, or ALS purposes. Technically, the machine reports a BrAC(Breath alcohol concentration), and then converts to a BAC (Blood alcohol concentration). However, throughout this chapter and this book, I mean BAC to include BrAC.


11.1.1 Intoxilyzer 5000 Overview

According to the manufacturer (CMI), the Intoxilyzer 5000 is used by more than 20 States. With a one year, up to 3 year manufacturer warranty. (I can almost guarantee you the model used by the department is more than 3 years old, but if you want to check for yourself, there is a date on the back of the Intoxilyzer. How the Intoxilyzer Works

The Intoxilyzer works using infrared (IR) technology. Meaning, it calculates BAC by light.  Different chemical compoundsabsorb light energy at specific wavelengths.

At one end of the Intoxilyzer sample chamber is a Quartz-Iodide lamp, which directs light through the chamber. The light then passes through a filter wheel onto a Lead Selenide infrared energy detector.

The filter wheel allows only specific wavelengths of light to pass through.  The wheel rotates so that each filter is inserted one at a time. It rotates at around 1,800 revolutions per minute[2]. There is both a 3 filter version, and a 5 filter version[3].  The five filter one checks for microns at 3.80 (reference); 3.52 (toluene); 3.47(ETOH / Ethanol); 3.40 (acetone); and 3.36 (acetaldehyde)[4].  The machine measures a reference point by measuring the amount of infrared energy striking the Lead Selenide. During a breath test, the alcohol vapor in the chamber rises, and the amount of light reaching the detector decreases[5]. A BrAC is obtained by comparing the ratio of microns at the reference, to the ones at the alcohol wavelength. The result is displayed in grams of alcohol per 210 liters of breath (this is also referred to as a partition ratio of 2100:1)[6]. Mouth Piece

The user first blows through a mouth piece (spit trap). There are a few different variants available.

[7] .

Some versions are more accurate than others. The purpose behind the spit trap is to prevent saliva/ other foreign substances from entering the breath tube, and then the machine.Saliva has a different alcoholpartition ratio to blood than breath does.

The officer must replace the mouthpiece between each sample, as well as replacing it if the person attempts to give a sample that the machine does not recognize. The officer has 60 seconds to replace the spit trap between the breath samples[8]. If the officer fails to replace the mouthpiece, alcohol might have been built up from the previous test. Breath Tube

The air being blown then passes through a heated breath tube. The tube is heated to prevent condensation from forming[9]. The tube has an antenna that is supposed to check for RFI (Radio frequency interference). [10] Types of Checks by the Machine

The machine checks ambient air, the subject’s breath, and the simulator solution.

The ambient air check determines how much alcohol is in the air surrounding the mouthpiece. Basically, the machine pulls in the air without anyone blowing into it. The air blank result will always be zero, even if the machine detects alcohol. The machine uses this reference as a starting point. It will only produce an ambient failure result if the result is a 0.10 reading or higher[11]. During the ambient air phase, the screen of the Intoxilyzer does not show the present result as it does while someone is giving a breath sample.

The simulator solution is a predetermined amount of BAC (.10, .20, and .30). It is checked at internal 1,2, and 3 respectively.  After the subject blows and the machine runs another blank, it captures a sample from an external reference (A Guth Model 2100 simulator)[12] Slope Detector

The machine also checks the slope of the result. This means, if the BAC value is changing too rapidly, the machine will consider it mouth alcohol, or some other problem, and not accept the sample. One can check this by putting alcohol in their mouth right before blowing. The result will start off incredibly high, but fall rather quickly. If it falls, or even increase too fast, the machine will not accept the sample.

Generally, the machine will give a continuous tone while it is collecting a sample. However, once there is a 0 to the left of the decimal place in the Dot Matrix display, the machine will consider it a valid result if the person stops blowing. Sometimes, when the slope is low enough at a given time, the 0 both comes and goes away. Accordingly, it is possible to collect a result while the tone is still going, and while the machine is still obtaining a sample and doing calculations. Required Sample Amount

The machine requires the following for a successful sample:

A blow for minimum for one second;

Minimum sustained rate of .15 L/sec;

Minimum volume of 1.1 liters of sample;

Blow until the BrAC change is minimal[13].

Practice Note: If the driver was considered a refusal by the arresting officer but attempted to give a breath test, see if the officer gave any incorrect instructions, such as blow hard quickly, or blow low for a longer amount of time.


11.1.2 Defenses Related to the Accuracy of the Intoxylizer

The State, and even the manufacturer, likes to tout the Intoxilyzer 5000 as very accurate when it comes to trial. However, a simple question to ask is:“if the Intoxilyzer 5000 is so accurate and reliable, why is there an Intoxilyzer 8000?” (The reason the Intoxilyzer 5000en must be used in New Hampshire, is because it is presently the only model of any breath machine that can retain capture sampled tubes). Notice, when the State wanted to upgrade the machines, they noted how there are better machines (presumably more reliable) on the market[14].

"You'd be adding confidence and validity to the test," said Nancy Mobile, supervisor of the lab's breath-testing unit.

"And you'd have greater assurance of the specificity of the alcohol level," said Pifer, Director of the State Police lab.  "But here's a situation where there is a more scientific way to do the testing.

Ann Rice of the state attorney general's office noted there is new, improved technology.”


See Chapter 3.5.9 dealing with defenses related to not complying with statutory guidelines.

Other defenses Include:

“List of physiological variables potentially important in connection with breath-alcohol measurement for legal purposes. Failure to adequately control these biological factors will contribute to the variations in the blood/breath alcohol relationship reported in the literature.

Phase of ethanol metabolism

Arterio-venous differences

Source of blood analyzed; arterial, venous, capillary

Blood hematology; hematocrit value, salt, fat, and protein content

Intra-pulmonary gas pressure

Ambient temperature and humidity

Expired-breath temperature

Breathing pattern; hypo-and hyper-ventilation

Bronchopulmonary disease

Presence of mouth alcohol

Regurgitation of stomach fluids

Phase of exhalation; end expiratory or top-lung air

Breath specimen; rebreathed, end-expired, mixed expired

Body temperature; hyperthermia-hypothermia[15]


Many of these defenses proposed will require an expert witness. Margin of Error

No machine is 100% accurate. Officers and personnel from the Department of Safety will often testify the margin of error is as low as .005 or 5%, and as high as .02. Because the margin of error is at least .000001, the State should not be able to obtain a Per Se conviction on .08, or an aggravated conviction on .16. The reason being that any result is as likely .07999999 as it is .08000001. In fact, the DMV will return to officers without action any test indicating a breath sample of .08. If the DMV finds the State cannot meet the burden by preponderance of the evidence, certainly the State at trial should not meet the burden beyond a reasonable doubt.

By statutory rules, the simulator results reported by the machine must be with 10% of the actual value[16]. Additionally, each breath sample must be within .02 of each other[17]. This makes for a good argument that the Department of Safety considers a .02 margin of error. The error rate to confirm an independent sample is even higher; as much as .05[18]. Partition Ratio

The Intoxilyzer converts breath to blood using a formula called a partition ratio. It is defined by statute as "Alcohol concentration'' shall mean either grams of alcohol per 100 milliliters of blood, grams of alcohol per 67 milliliters of urine, or grams of alcohol per 210 liters of breath[19]. This 2100:1 formula is based upon studies determining what an “average” person will be. However, there is a broad range.  One Court has found the range to be between 1142:1 and 3478:1[20]. One of the leading blood scientists has held: “Obviously, there cannot exist a universal fixed blood/breath ratio of ethanol which applies for all subjects under all conditions of testing.[21]

Besides everyone having a different ratio, decreased lung capacity/shallow breaths can affect the results by as much as 50%[22]. Temperature

Mouth/Body temperature affects the results of the Intoxilyzer. The reason for this is based upon how the Intoxilyzer works. The Intoxilyzer works using Henry’s Law (concentration of a solute gas in a solution is directly proportional to the partial pressure of that gas above the solution[23]). “The value of the Henry's law constant is found to be temperature dependent. The value generally increases with increasing temperature. As a consequence, the solubility of gases generally decreases with increasing temperature.”[24]

The Intoxilyzer assumes a sample of 34 degrees Celsius. In fact, the simulator temperature must be within +/- .2[25].  You can see what the actual temperature was (or at least that the officer remembered to write 34.0) by looking at the bottom right of the Intoxilyzer printout card.

The problem with this temperature is that everyone’s breath is different, and the average tends to be closer to 35 degrees [26]. For every one degree difference, there is an 8.6% overestimation of BAC.

A fever, or a woman’s menstrual cycle/menopause, can give higher than average body temperatures. Arguably, if someone is stressed, such as due to an arrest, their body temperature will also rise. When the body’s fight/ flight system is activated the muscles tense, heart rate and thevital organs speed up and as a result, blood flow is shunted from the extremities and directed to the vital organs to facilitate the increased level of arousal. As a result, changes of 5°, 10°or 15° can occur within just a few minutes[27]. If the officer says the person has a flushed face, you may also be able to argue an increase in body temperature. Breathing Pattern

Similarly, breathing pattern can affect the BAC calculations. The subject's manner and mode of breathing just prior to providing a breath sample for analysis can significantly alter the concentration of alcohol in the resulting exhalation. A driver holding his breath for a short time (20 seconds) before exhalation can increase the alcohol concentration in exhaled air by 15%. Alternatively, hyperventaliting, lowers it up by to 20%[28].

As breath is expired from the lungs, the longer the person breaths, the higher the temperature of breath is. “The temperature of breath as it leaves the mouth rises from about 33.3 to 34.4°C as the volume of breath exhaled rises from 500 to 4500 ml”[29]. Accordingly, the person’s BAC should always be going up as the person continues to blow into the machine. If the officer says “blow, blow, blow”, he can artificially inflate the BAC, because the Intoxilyzer does not compensate for breath temperature, or use it as part of its calculations. Forced Agreement

Similarly, the officer can sometimes “force” the persons BAC to be a certain number. A knowledgeable officer knows that the second sample must be within .02 of the first one. As the output on the machine is either rising, or falling, if it gets to .02 difference, the officer may tell the person to stop. Even if the tone is still going, as long as there is a 0 to the left of the decimal in the Displayed reported BAC value, the machine will accept the sample. Some officers cover the display with a license to prevent the person from seeing the first sample. If the officer did this, the officer could not have cheated and forced a certain number. However, we also then do not know how long a 0 was up for the machine to accept the sample. Again, usually, the longer the blow, the higher the BAC. Hematocrit

Hematocrit is a ratio of the volume of red cells/packed cells in relation to whole blood.

The composition of whole blood varies from person to person and is exaggerated in various disease states.[30]” The water content of plasma is greater than red cells.  “This means that a blood specimen with an abnormally low hematocrit value will have more water per unit volume than a blood with high hematocrit value..” Less water equals a higher concentration of alcohol. Men have a higher value than women, so accordingly, men will tend to have a falsely higher reading (Or, women are falsely low. Without the Intoxilyzer source code, it is impossible to know what hematocrit value is used). Mouth Alcohol / Slope Detector

To determine someone’s BAC (BrAC) by obtaining a breath sample, the alcohol must be alveolar air[31] / “deep lung air.” Alveolar air is expired air which is thought to be in equilibrium with the blood (again based upon Henry’s law). In order to obtain a proper result, the Officer must ensure the subject does not put anything in his mouth, burp, vomit, or regurgitate for 20 minutes prior to beginning the test[32].  The purpose of this observation period is so that the machine does not detect mouth alcohol. If any alcohol is in either in someone’s mouth, or was there shortly before testing, the result will be high. In certain instances the result can be so high the machine will not register it.

To further prevent the Intoxilyzer from calculating based upon mouth alcohol, it has built in a “slope detector.” What this does, is if the BAC changes too rapidly, the machine is not supposed to accept the value, and will have the person continue to blow until the number levels off.  The basis of this logic is that mouth alcohol will quickly dissipate and the BAC will drop too fast as the alcohol leaves the mouth. However, how can one be sure the slope detector is working?  To my knowledge, it is not tested for at the Department of Safety. Without the source code, we do not even know how it is calculated. GERD

Similar to Mouth alcohol is Gastroesophageal reflux disease (GERD). Reflux means the stomach acids and other contents go back up through the esophagus into the throat and mouth.  In healthy people the lower esophageal sphincter closes. In patients with GERD it does not. Alcohol, beer, wine, coffee, and drinks with carbonation can make the symptoms of GERD worse because they dilate the stomach and therefore dilate the lower esophageal sphincter[33].

Air from the stomach, is not the same as “deep lung air.”Accordingly, the result will be inaccurate. Because a person with GERD can have a constant flow of air, the slope detector will not notice the problem, as opposed to a spike by a burp.

One can be tested for GERD with an endoscopy. The photographs can be looked at to see if there is evidence of damage caused from the acid erupting. Further, photographs can determine if the Lower Esophageal Sphincter is open[34]. Dentures (or other things in the mouth such as braces, or Invisalign®)

See: Patrick Harding, et. al., The Effect of Dentures and Denture Adhesives on Mouth Alcohol Retention, 37 Journal of Forensic Science 999, 1002 (July 1992).

One of the individuals in another study had a breath sample of 70ug/100mL (Å .15) while the blood sample showed .05[35].

Often, the officer will not check the person’s mouth very well, except maybe to see if they are chewing on gum. The Invisalign is particularly difficult to detect.



The fact that a foreign substance is in the mouth might be enough to invalidate the test based upon administrative rules, or, at a minimum, go to the weight of the evidence. Obviously, someone cannot take out braces, however, the Invisalign can be removed with no issues. When someone does have braces or dentures, there is an argument that food or alcohol is getting trapped in them, and that the machine is incorrectly detecting mouth alcohol in lieu of alveolar air. Radio Frequency Interference (RFI) –

The Intoxilyzer uses electronics. Accordingly, radio waves can interfere with these electronics. For jury trials, remind the jurors that when you take off in an airplane that the pilot has you turn off all electronic devices. Similarly, there are FCC warning labels on many electronics that state: “ This device complies with part 15 of the FCC Rules. Operation is subject to the following two conditions: (1) This device may not cause harmful interference, and (2) this device must accept any interference received, including interference that may cause undesired operation."


The RFI antenna in the Intoxilyzer is based upon radio frequencies used in the 1970s and 1980s.[36] Lab personnel check to see if it is working by detecting against known signals the Intoxilyzer is programmed to detect.[37] For an interesting video showing this phenomena in action, and not a theory, see: Also, note the RFIcan interfere with the air blank showing ambient failure. Accordingly, you can get a refusal based upon something the driver did not do. Lack of Specificity

Again, the machine works by measuring light wavelengths. It does not specifically detect for ethanol. A couple of things than can give a high reading are someone who has diabetes, or has fasted, and therefore has acetone on the breath[38]. Or, someone who works with paint, glue, or lacquer fumes[39].  Further, in diabetic ketoacidosis, isopropyl alcohol is a byproduct of acetone metabolism[40].  Numerous other chemicals can interfere with breath testing instruments[41]. Operator Error

As discussed throughout the book, operator error includes: not following all the regulations, not observing for 20 minutes, “forced agreement”, improper instructions, having the person hold the breath too long, not changing the mouthpiece, not changing the sample tubes, not properly labeling the tubes, and failing to give the driver his sample tubes. Certifications & Other Compliance with Statutes

See Chapter 3.5.9 and Saf-C 6302.02 for many of the regulations. Further, make sure the officer presents evidence that he was certified at the time of operation. Officer certification is good for one year. Also, the machine must have been certified at the time of operation. Machine certifications are valid for 6 months.  Besides seeing the certification, see what, if anything, was fixed on the machine when it was certified. Besides the time in question, you may want to see if anything was fixed when the machine was re-certified since the time in question (assuming enough time went by so that the 6 month certification was up). Truncating

Unlike blood test results which are averaged, breath tests go out 2 decimal points, and are truncated (the machine just drops the third digit). One may think this benefits the defendant, but that is not always the case. The two samples must be within .02 of each other. If sample 1 is .109, and sample 2 is .081, the real difference is .028. However, based upon regulations, the difference will be .02. Accordingly, whenever the test result is .02 higher, there is a good argument to make that it was just as likely up to .029 higher/lower. This argument may be used to try andshow the samples were outside the range of acceptability, or that the machines true margin of error is .029. Testing During the Absorptive Phase

The absorptive phase is when alcohol is still being absorbed by the body from the stomach. All BAC levels will be inaccurate (in all likelihood higher) if the person is tested in the post-absorptive phase if he was in the absorptive phase at the time of driving. Studies by Kurt M. Dubowski indicate varying ranges of elapsed time from the end of alcohol intake to peak blood alcohol concentration of 14 to 138 minutes, 9 to 114 minutes, and 12 to 166 minutes[42]. He concludes:

“It is often impossible to determine whether the postabsorptive state has been reached at any given time. Those factors make it impossible or infeasible to convert the alcohol concentration of breath or urine to the simultaneous blood alcohol concentration with forensically acceptable certainty, especially under per se or absolute alcohol concentration laws.[43]

However, specific to breath tests, “During the absorption phase and around the peak BAC breath analysis tends to overestimate venous BAC and the differences will depend on the actual blood/breath ratio for the person tested in comparison with the 2100 calibration factor and the magnitude of arterio-venous difference[44]” One study showed a breath test result can be as much as 100 percent higher than the BAC level during the absorptive phase[45].

[1] Intoxilyzer is copyright/trade mark of CMI.

[2]A Practical Guide to Trying DWI Cases in New Hampshire, Mark Stevens et al., (2010) (Hereafter referenced as Practical Guide) Ch. 3 pg 13.


[4] Practical Guide Ch.3 pg. 11

[5]Intoxilyzer 5000 Original Operator Training, State of New Hampshire Department of Safety Division of State Police Forensic Laboratory, Nancy Mobile et al. August 2009 Revision 9, at pg 5 (Hereafter referenced as Operator Training Manual)

[6] N.H. RSA 259:3-b


[8]Operator Training Manual Pg. 15

[9]Practical Guide. at 3-12

[10]Id. at 3-12

[11]Id. at 3-15


[13]Operator Manual. at pg 6


[15] A. W. Jones, Physiological Aspects of Breath Alcohol Measurement,  Alcohol, Drugs and Driving, Volume 6 Number 2, Pg. 15 (1990). Hereby referenced as Jones.

[16]Saf-C 6302.04(a)(3)

[17]Saf-C 6302.02(d)

[18]Saf-C 6305.09(b)

[19] N.H. RSA 259:3-b

[20]State v. McCarty, 434 NW 2d 67, 68  S.D. Supreme Court (1988)

[21]Jones at 14.

[22]Booth v. Director of Revenue, 34 SW 3d 221,223 (2000).



[25]Saf-C 6302.02(h)

[26]Schonecht, et al., The Technical Concept for Evidential Breath Testing in Germany. Proceedings of the ICADTS T-95 (1995).


[28]Jones at 18

[29]Jones at 17

[30]Jones at 15.

[31]Saf-C 6302.02 (a)

[32]Saf-C 6302.02(b)

[33]People v. Bonutti, 788 NE 2d 331, 334(2003)

[34]Barone at 2-62.

[35]D.J.H. Trafford & H.L.J. Makin, Breath Alcohol Concentration May Not Always Reflect the Concentration of Alcohol in Blood, 18 (4) Journal of Analytical Toxicology 225, 225 (Jul.-Aug. 1994).

[36]Practical Guide at 3-22


[38]Attacking and Defending Drunk Driving Tests, Donald Bartell (201) (hereafter referenced as Bartell), ch. 11-15

[39] For a list of articles related to this issue see:

[40]Bartell, 11-16

[41]Bartell at 11-22

[42]Kurt M. Dubowski, Absorption, Distribution and Elimination of Alcohol:Highway Safety Aspects, 10 J. Stud. Alcohol Suppl. 98-105 (July 1985).

[43]Id. at 105

[44]Jones at 17

[45]E. Martin, W. Moll, P. Schmid, L. Dettli, The Pharmacokinetics of Alcohol in Human Breath, Venous and Arterial Blood After Oral Ingestion, 26 (5) Eur. J. Clin. Pharmacol., 619 (1984). As referenced at