hazardous gas monitors

1

Chapter 1 Introduction

Chapter 1

Introduction

T he detection of hazardous gases has always beena complex subject and makes choosing an appro-priate gas monitoring instrument a difficult task.

To address this problem, this book aims to providethe following essential tools:

• A simple guide to the various sensor technolo-gies available

• Information to help you intelligently select theproper instruments for specific applications

• Information engineers can use to design a com-plete monitoring system

• Technical data and practical procedures thattechnicians can use to check and maintain a gasmonitoring system

The main emphasis of the book is on gas detec-tion technology that is used in the field of area airquality and safety. This field primarily involves theprotection of personnel and property against toxicand combustible gases. The discussion includes thetypes of sensors used, the various instruments avail-able, and the applications that incorporate these in-struments.

Analy tical Instruments and Monitoring Systems

To date, no gas sensors exist that are 100% selec-tive to a single gas. Achieving such selectivity requiresthe use of instruments that employ analytical tech-niques to identify gases.

Examples of such instruments include Fourier trans-form infrared (FTIR) instruments that use the infraredspectral characteristics of gases, gas chromatographs thatuse analytical columns, and mass spectrometers that iden-tify molecules through characteristic variable deflec-tions from a magnetic field.

These instruments provide fairly accurate and se-lective gas readings. Some typical applications forthese kinds of instruments include airport bombdetection, drug abuse screening, and analyzing airpollutants. However, these analytical instruments re-quire skilled and knowledgeable operators, and aregenerally very expensive and designed for laboratorytabletops or specific on-line applications for in-plantinstallations.

In addition, many suffer from limitations such ashigh maintenance, slow response time, and large size,making them impractical monitors for area air qual-ity and safety. Thus, they are typically used only as alast resort for applications in which a suitable sensoris not available.

For work area air quality and safety applications,monitoring systems must meet a number of practicalcriteria. These monitoring systems must be:

• rugged and corrosion-resistant

• weather- and dust-proof

• capable of being installed in hazardous areas

• durable and long-term

• operationally stable

• easy to maintain

• operated by a minimally skilled person

An Analytical Instrument. Theexample shown above automatesgas chromatography with the helpof its built-in robotic technology.(Courtesy of CE Instruments)

A Gas Monitor. IST’s MP-204is a wall-mounted unit housedin weatherproof enclosure withfour sensor channels.

2

Hazardous Gas Monitors

3


1 Any device that converts input en-ergy of one form to output energyof another.

• suitable for multisensor systems that, for ex-ample, can be used for an entire chemical plant

• low cost

This book deals with gas monitors for work areaair quality and safety applications. For practical pur-poses, we will not delve into the realm of the muchmore complex analytical instruments which, for themost part, do not meet these criteria.

Gas Sensors

A gas sensor is a transducer1 that detects gas mol-ecules and which produces an electrical signal with amagnitude proportional to the concentration of the gas.

Unlike other types of measurement, types that arerelatively straightforward and deal with voltage, tem-perature, and humidity, the measurement of gases ismuch more complicated. Because there are literallyhundreds of different gases, and there is a wide ar-ray of diverse applications in which these gases arepresent, each application must implement a uniqueset of requirements. For example, some applicationsmay require the detection of one specific gas, whileeliminating readings from other background gases.Conversely, other applications may require a quan-titative value of the concentration of every gas presentin the area.

Types of Gas Sensors. There are many differenttechnologies currently available for the detection ofgases, each with certain advantages and disadvantages.The following sensing methods are the focus of ourdiscussion and are the five types most suitable andwidely used as gas monitors for area air quality andsafety applications:

• electrochemical • infrared

• catalytic bead • photoionization

• solid state

Gas Molecules

GAS TRANSDUCER ElectricalSignal

Some Examples of Gas Sensors1. Catalytic Bead. 2. Infrared.3. Solid State. 4. Electrochemical.5. Photoionization.

1

2

34

5

4


All of these sensors are commonly used for detection oftoxic and combustible gases in the work area for hu-man and property protection, or for process control.

One common characteristic of these sensors,despite what is often claimed or implied, is that theyare not specialized to detect any one specific gas.Each sensor is sensitive to a group or a family ofgases. In other words, the sensor is non-specific andis subject to interference by other gases much likea smoke detector in a house cannot distinguish be-tween the smoke caused by a furniture fire and thesmoke caused by food burning in the stove or oven.

In limited cases, a chemical filter can be installedto filter out interference chemicals while permittingthe target gas to pass through to the sensor. Alterna-tively, an analytical column can be installed to iden-tify chemicals qualitatively and quantitatively.

For gas monitoring applications, a proper sensoris usually selected to match the specific applicationrequirements and circumstances, with the user inter-preting the readings based on an awareness of thesensor’s limitations.

Terms, Definitions, and AbbreviationsUnits of Measure for Gas Concentration

ppm: parts per million by volume (see Table 1, oppo-site page)

ppb: parts per billion by volumemg/m3: milligrams per cubic metermg/cc: milligrams per cubic centimeterg/m3: grams per cubic meterg/cc: grams per cubic centimeter

For Combustible Gases

Flash Point (Fl.P): The temperature at which a com-bustible liquid gives off enough vapor to form

Fuel

Oxygen Ignition Source

The Combustion Triangle

5


Table 1. Equations for Deriving Units of Gas Concentration

Gas concentration is commonly expressed as percent (%), ppm, or ppb. Mathematically, these areunitless expressions since they do not carry a unit for volume or weight but simply express the ratio ofgases in relation to background air.

For instance, one ppm of CH4 simply means one part of methane amongst 999,999 parts of back-ground air. It is expressed as

Vg =Vg

Va + Vg VT

Vg = volume of gas; Va = volume of air; VT = total volume of air and gas

Multiply the fraction derived from the formula above:a) by 102 % to obtain the percentageb) by 106 ppm to obtain the ppmc) by 109 ppb to obtain the ppb

For example, if you mix 1 cc of gas with 99 cc of air, the calculation is as follows:1 cc

= 0.011 cc + 99 cc

Thus, 0.01 x 102 % = 1%0.01 x 106 ppm = 10,000 ppm0.01 x 109 ppb = 10,000,000 ppb = 107 ppb

In cases like this, one would normally not cumbersomely express the units as ten million parts perbillion. Instead, the simpler expression, 1%, is preferred.

This volumetric expression of concentration is straightforward. Additionally, the volume ratio is equalto the pressure ratio according to Dalton’s law of partial pressures. It is expressed as

Pg =Pg =

Vg

Pa + Pg PT VT

Pg is the partial pressure of the gas within the total pressure PT and Pa is the partial pressure of air.As an example, 1 psi of gas within 99 psi of air with total pressure of 100 psi has a concentration of 1%as in the volume expression.

Another unit, commonly used in the medical and metallurgical industries, is mg/m3 or milligrams percubic meter, in situations wherein the chemical is either in liquid or solid state at room temperature. Toconvert mg/m3 to percent or ppm, the ideal gas law must be used. Chemical conversion factors areincluded in the Gas Data section in Appendix II, page 199, and the conversion formula is discussed onpage 173 in Chapter 11, Gas Sensor Calibration.

6


TOO RICH FOR COMBUSTION

TOO LEAN FOR COMBUSTION

0%

100%

COMBUSTIBLE MIXTURE

GA

S CO

NCE

NTR

ATIO

N(M

IXTU

RE IN

AIR

)

UEL

LEL

Fig. 1. A Window of Combustibility

an ignitable and combustible mixture when airis present near the liquid’s surface.

That is, it is the temperature at which a com-bustible liquid chemical has sufficient partialpressure2 in the air to be ignited. The partialpressure curves of chemicals are available fromchemical libraries or manufacturers of thechemicals.

Lower Explosive Limit (LEL) or Lower FlammableLimit (LFL) : The minimum concentration of gas or

vapor mixed with air (percent by volume, at roomtemperature) that will cause the propagation offlames when it comes in contact with a source ofignition. In common terminology, mixtures be-low the LEL or LFL are too lean to ignite.

Upper Explosive Limit (UEL) or Upper FlammableLimit (UFL) : The maximum concentration of gas

or vapor mixed with air (percent by volume, atroom temperature) that will cause the propaga-tion of flames when it comes in contact withan ignition source. In common terminology,mixtures above the UEL or UFL are too rich tosupport combustion. The combustible range is,therefore, between the LEL and the UEL (seeFigure 1 below).

2 The law of partial pressures, firstformulated by James Dalton in1802, states that the pressure ofa mixture of gases, P, which do notreact chemically, is the sum of theindependent pressures (partialpressures) that each gas exerts:

P = P1 + P2 + . . . + Pn

P = PO2 + PN2 + PH2O+

P = 3.087 psi + 10.29 psi + 1.323 psiP = 14.7 psi

N2 = 70%= 10.29 psi

O2 = 21%= 3.087 psi

H2O+= 9%= 1.323 psi

7


Specific Gravity (Sp.Gr.): The ratio of the weightper unit volume or mass of a substance at 68°F(20°C) to the mass of an equal volume of dis-tilled water at 39.2°F (4°C).3

Vapor Density: The weight per volume of gas orvapor compared to dry air; both componentshaving the same temperature and pressure. Forexample, air has a vapor density of 1; carbondioxide, 1.52; hydrogen, 0.07; methane, 0.55;and propane, 1.52.

For Toxic Gases

The National Institute for Occupational Safety andHealth (NIOSH) is a branch of the U.S. Departmentof Health and Human Services, Public Health Service,Centers for Disease Control and Prevention. Actingunder the authority of the Occupational Safety andHealth Act of 1970 and the Federal Mine Safety andHealth Act of 1977, it publishes recommended exposurelimits (RELs) for hazardous substances or conditionsin the work place.

To formulate these recommendations, NIOSH col-lects and evaluates data from the fields of industrialhygiene, toxicology, occupational medicine, and ana-lytical chemistry. These recommendations are thenpublished and transmitted to the Occupational Safetyand Health Administration (OSHA) and the MineSafety and Health Administration (MSHA) for use inpromulgating legal standards.

OSHA published the permissible exposure limits(PELs) that are known as the General Industry AirContaminants Standard.

The American Conference of Governmental In-dustrial Hygienists (ACGIH) is a professional society,not an official government agency. Membership is lim-ited to professionals in governmental agencies or edu-cational institutions engaged in occupational safetyand health programs in the United States and around

3 Water at 4°C has the LOWESTvolume per gram. Ice expandswhen its temperature goes below0°C and will crack a rigid encase-ment of it even if it was made ofcement.

8


the world. ACGIH publishes the exposure standard,threshold limit values (TLV).

These three different standards—NIOSH’s RELs,OSHA’s PELs and ACGIH’s TLVs—are similar to eachother, yet in some instances there are variations amongthem. All are based on time-weighted-average (TWA),short-term exposure limit (STEL), ceiling (C) and immedi-ately dangerous to life or health (IDLH) concepts. Thefollowing definition of these terms are for informa-tional purposes only. It is beyond the scope of this bookto include detailed discussions of each.

Time-Weighted Average (TWA) is the average con-centration of contaminants over a specified time

period. Mathematically,TWA is the integrated areaunder the concentrationcurve over time divided bytime period. To illustratethe concept of TWA (Figure2), let us assume that dur-ing a three-hour period, COconcentration is constant at50 ppm during the firsthour; then the CO concen-tration increases steadily to100 ppm at the end of thethird hour. The TWA at anygiven moment is repre-sented by the red line. Many microprocessor-based instruments and datalogging programs are ca-pable of performing theTWA calculation, and thetime interval used by theinstruments to calculate theTWA is much shorter thanone hour.

56.25 ppm

GA

S CO

NCE

NTR

ATIO

N(%

MIX

TURE

IN A

IR)

1 2 3HOURS

0

TWA

0

50

100

66.7 ppm

100 ppm

Fig. 2 Simplified Three-Hour TWA. PELs are based on an eight-hour TWA. This is a simplified version to illustrate the TWA concept.

During the first hour, the TWA is:(50 ppm x 1 hr.) / 1 hr. = 50 ppm

Up to the end of the second hour, the TWA is: (50 ppm x 1 hr.) + (50 ppm x 1 hr.) +

(25 ppm x 1 hr.)

2 = 56.25 ppm 2

Up to the end of the third hour, the TWA is:

(50 ppm x 3 hrs.) + (50 ppm x 2 hrs.)

2 = 66.7 ppm3

The numbers within the parentheses represent areas under the curve.

9


Recommended Exposure Limit (REL) is TWA con-centration permissible for up to ten-hour work-days during a forty-hour work week.

Permissible Exposure Limit (PEL) and Thresh-old Limit Value (TLV) are TWA permissibleconcentrations, to which workers may be ex-posed continuously, day after day, without ad-verse effects, for a normal eight-hour workdayand a forty-hour work week.

Short-Term Exposure Limit (STEL) is defined as afifteen-minute TWA exposure which should notbe exceeded at any time during a work day evenif the eight-hour TWA is withinlimits. Exposures at the STELshould not be longer than fif-teen minutes and should not berepeated more than four timesper day. There should be at least60 minutes between successiveexposures at the STEL. The con-cept of STEL is illustrated in Fig-ure 3.

Ceiling Limit : The concentrationwhich should not be exceededat any time.

Immediately Dangerous to Life or Health (IDLH)concentration is the maximum concentrationabove which only a highly reliable breathingapparatus providing maximum protection forworkers is permitted. The IDLH value is basedon the ability of a worker to escape without lossof life, irreversible health effects, or otherhealth effects such as disorientation or incoor-dination that could prevent escape.

The preceding information is a simplified versioninterpreted from the NIOSH Pocket Guide to ChemicalHazards. Appendix II at the end of this book lists data

GA

S CO

NCE

NTR

ATIO

N(M

IXTU

RE IN

AIR

)STEL

PEL

5 10 15MINUTES

0

CEILING

TWA (red line)

Fig. 3. A graphical example of permitted TWA excursionsabove STEL, provided the 15-minute TWA does not exceedthe STEL.

10


extracted from the same pocket guidethat includes REL, PEL and IDLHconcentrations.

Figure 4 illustrates the overall sum-mary of the toxic gas safety concept.

Combinations of Substances. Whentwo or more hazardous substancesare present in the area, the additiveeffects should be considered. Thatis, if the sum of the fractions of haz-ardous substance concentrations di-vided by the respective PEL exceeds

unity, then the threshold limit of the mixture mustbe considered excessive. This case is illustrated inthe following formula:

C1 + C2 + C3 + . . . + Cn < 1 PEL1 PEL2 PEL3 PELn

If it is greater than one, then the PEL is exceeded.C1,C2 and C3 represent the TWA of various hazardoussubstances.

For example, if the ambient air contains 35 ppmof carbon monoxide (PEL 50 ppm) and 350 ppm ofcarbon dioxide (PEL 5000 ppm), the calculation is asfollows:

35 + 350 = 0.77 50 5000

The threshold limit is not exceeded.This discussion is for conceptual purposes only. It does

not take into account the effect of the combining ofchemicals that can react with each other, resulting ina final mixture that can be more toxic to humans thanthe individual toxicities of each gas.

Performance Specifications

Accuracy: Webster’s dictionary defines accuracy as

GA

S CO

NCE

NTR

ATIO

N(M

IXTU

RE IN

AIR

)

STEL

PEL

HOURS

CEILING

IDLH

90 1 2 3 4 5 6 7 8

TWA

Fig. 4. Basic Rules for Compliance: TWA for eightworking hours does not exceed PEL/TLV. Observe thedefinition of STEL and never exceed the ceiling limit.

“the quality or state of being accurate or exact;precision; exactness.” Accurate is defined as “freefrom mistake or error; precise; adhering closelyto a standard.” Precise is defined as “strictly de-fined; minutely exact; low tolerance; etc.” Accord-ingly, a measurement can be precise but not nec-essarily accurate. The accuracy can only be deter-mined when compared to a standard.

Accuracy is the most important definition ofthe quality of performance for most of the ob-jects we deal with every day; for example, yourwatch, the weather thermometer, the bathroomscale, and the measuring tape, to name a few. Thedefinition of these measurement standards arewell-defined.

At the National Institute of Science and Tech-nology (NIST; formerly National Bureau of Stan-dards), standards for weight, length, temperature,etc. are kept. Internationally, there is a total agree-ment about the “absoluteness” of those standards.

In real life, the most accurate instrument maynot necessarily be the best. For example, the mea-suring tape which is used by the tailor is not veryaccurate, but it is practical for the task. Calipersused in a machine shop are more accurate thanthe tailor’s tape but would not be suitable for useby the tailor. Thus, each instrument serves a dif-ferent objective. The gas monitors serve more likea tailor’s tape than a machine shop’s caliper.

The Challenge of Accuracy. With gas moni-toring systems, there is no standard by which tocompare accuracy. There are hundreds of differ-ent chemicals, each having its own unique chemi-cal and physical properties.

As an example, what is 100 ppm of carbon mon-oxide (CO) in air? Mathematically, this equates to0.01% of CO and 99.99% air. After the mixture ismade, its accuracy is difficult to determine since

MeasuredWeight

StandardWeight

ACCURACY

The International Vocabulary of Basic andGeneral Terms in Metrology defines theterms for result of the measurement asfollows:

1. Accuracy: Closeness of the agreementbetween the result of a measurementand a true value of the measurand.*

2. Repeatability: Closeness of the agree-ment between the results of successivemeasurements of the same measurandcarried out under the same conditionsof measurement.

3. Reproducibility: Closeness of the agree-ment between the results of measure-ments of the same measurand carriedout under changed conditions of mea-surement.

4. Linear scale: Scale in which each spac-ing is related to the corresponding scaleinterval by a coefficient of proportional-ity that is constant throughout the scale.

* measurand–a particular quantity subjectto measurement; e.g., vapor pressureof a given sample of water at 20°C.

11


12


there is no standard 100 ppm of CO to comparewith and there is no common agreement that de-fines a 100 ppm CO mixture.

Because calibration standards are difficult todefine in practice, accuracy is the most misunder-stood and abused term in gas monitoring. Thereare few agreements amongst manufacturers of in-struments, and there is no common understand-ing in general.

Realistically, it is best to establish a calibrationmethod that can yield consistent and precise calibra-tion data. The accuracy of this calibration methodcan be compared to an accepted standard whenchallenged.

As long as the calibrations are done with highprecision, the accuracy of your calibration can beestablished when an accepted standard is available.

Repeatability: Repeatability is the ability of sensors torepeat the measurements of gas concentrations whenthe sensors are subjected to precisely calibrated gassamples.

Zero Air : “Zero air” is available on the market in theform of a mixture of oxygen and nitrogen in ahigh pressure tank. In normal applications, how-ever, gas monitoring sensors are put to work in“non-ideal” environments and, consequently,there are many other components in the ambientair besides O2 and N2 , such as water vapor, car-bon dioxide, carbon monoxide, and other tracegases.

Therefore, it is not practical to zero a sensorto a simple mixture of oxygen and nitrogen. Somesensors can be zeroed with dry air or N2 but somecannot. For instance, most infrared (IR) detectorscan be zeroed with dry air or N2 as long as thewavelength being detected has a minimum watervapor effect. However, with solid-state sensors orphotoionization detectors (PIDs), very different

N2

O2

N2

O2

H2O

CO2 & others

A. Artificial Air

or

B. Environmental Air

Zero Air vs. Ambient Air

13


readings may result from dry air compared withwet air.

In many applications, sensors may be satisfac-torily zeroed by exposing the sensor to a bag ofair collected from a location where the air is “nor-mal.” In fact, this is the easiest way to verifywhether a sensor is giving a real alarmor a false alarm.

Linearity : Quantitatively, linearity refersto the output signal in relation to thegas concentration: If 1 volt equals 10ppm and a full scale 5 volts represent50 ppm, then the output will be lin-ear. With most sensors, the initialoutput of the sensor is linear or closeto it, but as the gas concentration in-creases, the output signal is graduallyreduced. Figure 5 shows a typical sen-sor response curve.

Specificity or Selectivity: This is the ability of an in-strument to detect a target gas without being af-fected by the presence of other interfering gases.

Most sensors are sensitive to a family of gases,and there are no sensors specific to only one gas.Among the more specific sensors is an electro-chemical sensor for the detection of oxygen.

Different techniques are employed in orderto achieve some degree of selectivity to suit prac-tical applications. For example, a charcoal filteris used to filter out most hydrocarbons while let-ting only CO, H2, and CH4 pass through.

In solid-state sensors, the surface temperatureof the sensor can be set differently in order tomake it more sensitive to one gas and less sensi-tive to other gases.

The most common practice is to use an ana-lytical column, in which the gas sample stream isintroduced into the column and the chemical

0Gas Concentration

Nonlinear Saturation(Poor Resolution)

NearLinear

Sensor Outputin volts

5

10

Fig. 5. Typical Sensor Output Curve.As the gas concentration increases,the output signal becomes smaller inrelation to the increase in gasconcentration, resulting in poorresolution. Most sensors providebetter accuracy at lower concen-trations than at very high concentra-tions. Thus, instruments on themarket today commonly have outputsignals that are digitally linearized.

14


components are separated and come out at theend of the column at different times, where theyare detected by a sensor.

This method works well for laboratory use buthas proven to be impractical for gas monitoringbecause it requires a high degree of user knowl-edge and a high degree of maintenance. Addi-tionally, since the sample must be drawn throughthe column, the time necessary to take readingscan be relatively excessive. Sample times of 15 to30 minutes are not unusual.

For ambient air monitoring, it is much morepractical to use a sensor directly installed at thelocation being monitored and compensated forthe different gases which may be present.

Interference Ratio: As mentioned earlier, sensors arenot selective to a single gas and will read othergases as well. Thus, a common practice for manu-facturers of gas monitoring equipment is to pro-vide data indicating the ratios that different gaseswill read on the sensor. For instance, on a 100ppm carbon monoxide sensor, hydrogen mayread at a 3-to-1 ratio. This means that 3 ppm ofH2 will read the same as 1 ppm of CO.

In many cases, even though it is stated that acertain gas will not interfere, if the concentra-tion of this gas is high enough, it may in fact ac-tually interfere. For example, while a CO sensorwith a charcoal filter has little interference fromcertain solvents at 100 ppm, when the concen-tration is increased to 1,000 ppm, they may in-terfere drastically.

Because there are so many gases, it is not pos-sible for manufacturers to present data on crosssensitivity ratios for all gases. Therefore, if inter-ference data is not provided for a gas that is ofinterest to you, you should inquire with themanufacturer if the sensor is selective to a spe-

9 ppmH2

CO Sensor 33 ppmCO Signal

30 ppmCO

9 ppm H2 shows as 3 ppm CO

15


cific gas being targeted; in which case, the manu-facturer could provide the data needed.

Response/Recovery Time: This is typically definedas the time it takes for a sensor to read a certainpercentage of full-scale reading after being ex-posed to a full-scale concentration of a given gas.For example, T80 = 30 seconds means that thesensor takes 30 seconds to reach 80% of the full-scale reading after being exposed to a full-scalegas concentration.

Temperature and Humidity: Specifications for theseparameters are easy to understand, but beware ofthe humidity specification. Relative humidity isan indication of the amount of water vapor in airas a percentage of the total amount possible at agiven temperature.

Quantitatively, the amount of actual water vaporin air is a function of temperature. For instance, at80% relative humidity and a temperature of 25°C,water vapor is present at a level of 3%. However, thesame 80% relative humidity at a temperature of 48°Cproduces a water vapor level of 10%. In the pres-ence of chemicals, combined with the changing oftemperature between day and night, the possiblewater condensation and resultant corrosive mixturescan compromise the life expectancy of a sensor.

0% Time

Scal

e

50%

100%

T100T80

Fig.6. A Typical Sensor Response

16


Hysteresis: The difference in response of the sensor whencalibrating from a zero level to mid-scale comparedto the response when calibrating from full scale tomid-scale, is known as hysteresis. This quantity is nor-mally expressed as a percentage of full scale. For ex-ample, a 100 ppm instrument, when calibrated from0 to 50 ppm and exposed to a 50 ppm calibration gas,will indicate 50 ppm. However, when the sensor is cali-brated to 100 ppm gas but is exposed to 50 ppm thesensor may indicate 55 ppm. This variation of 5 ppmis 5% full-scale hysteresis. Most infrared and photo-ionization instruments do not exhibit hysteresis, butmany other sensors, including electrochemical, solidstate, and catalytic sensors do exhibit hyteresis.

In alarm setting, the difference between the onpoint and off point alarm is also referred to as hyster-esis. For instance, if the alarm comes on at 100 ppm,the alarm will not turn off until the gas is below 90ppm. This hysteresis is needed; otherwise, an alarmcan be chattering at the set point of 100 ppm.

Zero and Span Drift: While there is no specific definitionfor these two terms, common understanding holds thatthis drift is the percentage change of the zero or spancalibration over a specified period of time, typically30 days or more.

Hazardous Locations

Gas monitoring instruments are often installed in in-dustrial process and production areas. These areas areoften classified as hazardous locations. Industrial fa-cilities in which potentially explosive gas atmospheresexist or may exist must utilize proper explosion proofprotection methods when using these types of instru-ments. It is beyond the scope of this book to providefull details; however, the following information maybe helpful. The reader is advised to consult with themanufacturer for specific needs.

North America and other parts of the world that

17


have been influenced by North American practiceshave traditionally used the National ElectricCode(NEC®) articles 500-503. They employ a classand division system: Classes identify the type of haz-ard present as gases or vapors, combustible dusts,and flammable fibers. Divisions define the condi-tion under which the hazardous materialmay be present. The devices designed andmanufactured for these hazardous loca-tions should be tested and approved foruse by a nationally recognized laboratorysuch as Under writer’s Laboratories(UL®), Factory Mutual (FM), or the Ca-nadian Standards Association (CSA®).The NEC 500 hazard classifications areas follows:

Class I: Flammable gases or vapors.Class II: Combustible dustsClass III : Easily ignitable fibers and flyingsGroups are based on flame propagation characteris-

tics, ignition temperature, and pressure generated duringexplosion of various gases and vapors. There are four dif-ferent groups in Class I. These groups are as follows:

Group A: Acetylene

Group B: Acrolein, butadiene, ethylene oxide, form-aldehyde, hydrogen, propylene oxide, and propylnitrate

Groups C and D: All other combustible gases belongto Groups C and D

Division 1: Where ignitable concentrations of gases,vapors, dusts, and fibers can exist all the time or someof the time under normal operating conditions.

Division 2: Where ignitable concentrations of gases,vapors, dusts, and fibers do not exist under normaloperating conditions. Hazardous conditions only ex-ist in the event of abnormal conditions, such as acci-

North American Certification Agencies

Underwriters Laboratories Factory Mutual

Canadian Standards Association

18


dental rupture or breakdown of a container, stor-age tank, etc.

Gas monitoring instruments are typically designedand certified for use in Class I, Division 1, Group B, Cand D hazardous locations for use in North Americanmarkets.

Zones. European countries, as well as a majority ofother nations of the world, have been influenced bythe International Electrotechnical Commission’s (IEC)three-tiered zone approach. The IEC separates thepotentially explosive atmosphere into Zones 0, 1, and2 based on the probability of occurrence and length oftime a potential explosive mixture may be present.Apparatuses designed for use in these areas are usuallytested and approved for use by the European Commit-tee for Electrotechnical Standardization (CENELEC)test authorities using Euronorm (EN) standards. Thedivision of these three zones are:

Zone 0: An area in which an explosive gas atmosphereis continuously present for long periods.

Zone 1: An area in which an explosive gas atmosphereis likely to occur in normal operation.

Zone 2: An area in which an explosive gas atmosphereis not likely to occur in normal operation, and ifit does, it will exist for a short period only.

Zones 20, 21, and 22 are a subset of Codes 0, 1, and 2that refer to ignitable dust clouds.

Definition Comparisons. In accordance with theirexplosive properties, the combustible gases and vaporsare divided into temperature classes and explosion pro-tection subgroups. There are no direct comparisonsbetween the current NEC and IEC standards. NationalFire Protection Agency (NFPA) in America adoptedarticle NEC 505 which is comparable to IEC standards.A brief comparison of IEC (world), CENELEC (Eu-rope) and NEC (USA) are as follows:

1. Condition: Hazardous conditions exist continuously

Example: Ex EEx d IIb T3

Approved mark forapparatus certifiedby an EC testauthority

Symbol for apparatusbuilt in accordancewith a Europeanstandard

Flameproof enclosure(type of protection)

Explosion group

Temperature class

CENELEC Marking

19


or for long periods of time.NEC 505: Class 1, Zone 0NEC 500: Class 1, Division 1IEC: Zone 0.CENELEC: Zone 0

2. Condition: Hazardous condition is likely to occur innormal operation.

NEC 505: Class 1, Zone 1NEC 500: Class 1, Division 1IEC: Zone 1CENELEC: Zone 1

3. Condition: Hazardous condition is not likely to oc-cur in normal operation and if it does, only infre-quently and for a short period.

NEC 505: Class 1, Zone 2NEC 500: Class 1, Division 2.IEC: Zone 2CENELEC: Zone 2

Types of ProtectionThere are several acceptable types of protection

for electrical equipment in hazardous locations. Themore common types are the following:

A. Flameproof Enclosure (d)4 : The enclosure,such as the one shown in Figure 6, will with-stand an internal explosion, without causingignition of an external explosive atmosphere.The enclosure joints and structure covers aredesigned and manufactured for such purposes.This type of protection is most commonly usedfor gas monitoring applications and can meetthe requirements of:

NEC 500 - Class 1, Division 1 & 2.NEC 505-Class 1, Zone 1 & 2, AExd.IEC-Exd.CENELEC-EExd.

Fig. 6. An Explosion-ProofEnclosure

4 The lower case letter in a CENELEC-approved marking designates thetype of protection offered by anenclosure.

20


B. Intrinsic Safety (i): The electrical energy inan intrinsically safe circuit which enters into theenclosure is not sufficient to generate a sparkand ignite a combustible mixture in the haz-ardous area, in any worst case scenario. To meetthis requirement, safety barriers or other devices

limiting the electrical energyare placed on the wires tolimit the electrical energy al-lowed to flow through the cir-cuit before the wire entersinto the hazardous location.Safety barriers are a combi-nation of zener diodes, powerresistors, and fuses which aredesigned to limit the amountof electrical energy allowed to

flow through the wires. Various approved andcertified safety barriers are available as standardelectrical components. These are limited to lowpower device applications only. This method ofprotection can meet the requirements of:

NEC 500 - Class 1, Division 1 & 2.NEC-505 - Class 1, Zones 0, 1 & 2.

AExi(a).IEC - Exi(a).CENELEC - EExi(a).

Types of Flamepaths in Flameproof Enclosures

Joint Threaded Spigot

Materials used: Aluminum, Iron (Courtesy of Cullen Associates)

ZenerDiodes

Resistor Fuse

IntrinsicallySafe Ground

Single Channel

(+) 24 v

RM

Power supply mustnot be grounded

ProcessController

SensorModule

SAFE AREAHAZARDOUSAREA

(Courtesy of Cullen Associates)

21


C. Purged and Pressurized (p): This is the pro-cess of supplying sealed electrical enclo-sures with a protective gas to prevent theentrance of flam-mable gases whilemaintaining a positiveenclosure pressure.This type of protec-tion can meet the requirements of:

NEC 500 - Type X, Y & Z.NEC 505 - Type X, Y & Z.IEC - Exp.CENELEC - EExp.

Type X pressurizing: Reduces the classificationwithin the protected enclosure from Class 1, Di-vision 1 or Class 1, Zone 1 to unclassified.

Type Y pressurizing: Reduces the classificationwithin the protected enclosure from Division 1to Division 2 or Zone 1 to Zone 2.

Type Z pressurizing: Reduces the classificationwithin the protected enclosure from Class 1,Division 2 or Class 1, Zone 2 to unclassified.

D. Increased Safety (e): This is a type of explosionprotection applied to electrical apparatus thatdoes not produce arcs or sparks in normal ser-vice, in which additional measures are appliedso as to give increased security against the pos-sibility of excessive temperature and of the oc-currence of arcs and sparks. This method ofprotection can meet the requirements of:

NEC 500 - No Standard.NEC 505 - Class 1, Zone 1 & 2, AExe.IEC - Exe.CENELEC - EExe.

ProtectiveGas Supply

OptimalAlarm orPowerCut-OffSwitchPressure

Regulator ProtectiveEnclosure

PressureIndicatorOptional Pressure

Relief Device

22


E. Other Protection Methods: Oil immersion(o),powder filling(q) and moulding(m).

Enclosure Classifications For Nonhazardous Areas

In North America, the National Electrical Manu-facturer’s Association (NEMA), as a way of standard-izing enclosure performance, classified the enclosuresin different ratings which are intended to provide in-formation for users to make proper product choices.This rating system identifies the ability of the enclo-sure to resist various possible conditions. The classifi-cations are as follows:

NEMA Type 1: For general-purpose indoor use, pro-vides protection against incidental contact withthe enclosed equipment.

NEMA Type 2: In addition to NEMA Type 1, pro-vides protection against a limited amount offalling water and dirt.

NEMA Type 3: For outdoor use. Provides protec-tion against windblown dust, rain and sleet, aswell as formation of ice on the enclosure. Pro-vides rust resistance.

NEMA Type 3R: Same as NEMA Type 3, but doesnot provide dust protection.

NEMA Type 4: Same as NEMA Type 3, except, it isfor indoors or outdoors, provides protectionagainst direct water hose down.

NEMA Type 4X: Same as NEMA Type 4, but it alsoprovides corrosion resistance for indoor use.

NEMA Type 6: Same as NEMA Type 4X, but pro-vides protection against water during tempo-rary submersion at a limited depth.

NEMA Type 7: For indoor use in hazardous loca-tions, Class 1, Groups A, B, C, and D.

NEMA Type 9: For indoor use in dust applications,Class II, Groups E, F, and G.

23


NEMA Type 12: For indoor use, provides protec-tion against dust, falling dirt, and dripping non-corrosive liquids.

NEMA Type 13: For indoor use, provides protec-tion against dust, spraying water, oil, and non-corrosive coolant.

Internationally, the protections for electrical appa-ratuses are designated with ingress protection (IP) followedby a two-digit number which defines the degree of pro-tection. The first digit (0-6) defines the protectionagainst contact and entry of foreign objects while thesecond digit (0-8) defines the protection against water.

The IP classification is set by the InternationalElectrotechnical Commission. The definitions for pro-tections are different from that of NEMA. Therefore,the IEC enclosure classification designations cannot

Protection First Digit Second Digit

0 No protection No protection

1 Large objects of more Vertically falling water than 50 mm diameter

2 Medium-sized objects of Falling water at up tomore than 12 mm diameter 15 degrees from vertical

3 Small objects of more Falling water up to than 2.5 mm 60 degrees from vertical

4 Granular objects of more Water splashes fromthan 1 mm diameter any direction

5 Dust protected, Water from a nozzlenot completely tight from any direction

6 Dust tight Powerful water jet

7 Short-term immersion

8 Continuous immersion

24


be exactly equated with NEMA enclosures.As examples, IP 66 is approximately equivalent to

NEMA 4 or 4X , IP 67 is equivalent to NEMA 6, andIP 55 is equivalent to NEMA 12.

Summary

There is no clear definition delineating a gas moni-tor from an analytical analyzer. The distinction be-tween the two is based largely upon their usage in theactual application. A gas monitor is most frequentlyused to monitor gases in toxic and combustible rangesfor area air quality and safety applications. For thistype of safety application, the concept of the units usedin the measurements and the definition of terminolo-gies is somewhat unique. It is important to understandthe terms used. However, the examples of chemicaltoxicity presented in this book do not consider thecomplexity of actual toxicology and interreactions be-tween contaminant chemicals. In critical applications,a specialist on the subject needs to be consulted.

27

Chapter 2 Electrochemical Sensors

Chapter 2

ElectrochemicalSensors

T he oldest electrochemical sensors date back tothe 1950s and were used for oxygen monitoring.More recently, as the Occupational Safety and

Health Administration (OSHA) began requiring themonitoring of toxic and combustible gases in confinedspace applications, new and better electrochemicalsensors have been developed.

By the mid-1980s, miniaturized electrochemicalsensors became available for detection of many dif-ferent toxic gases in PEL ranges, with the sensors ex-hibiting good sensitivity and selectivity. Currently, avariety of electrochemical sensors are being used ex-tensively in many stationary and portable applicationsfor personal safety. Figure 1 shows a small collectionof such electrochemical sensors.

The physical size, geometry, selection of variouscomponents, and the construction of an electrochemi-cal sensor usually depends on its intended use. Quiteoften, the final design results in a compromise betweenvarious performance parameters of the electrochemi-cal sensor. The most common misconception aboutelectrochemical sensors is that they are all the same.In fact, the appearance of the electrochemical sen-sors used to detect various gases may be similar, buttheir functions are markedly different. Consequently,

Fig. 1. Electrochemical Sensors

28


one can expect varying performance from each ofthese sensors, in terms of sensitivity, selectivity, re-sponse time, and operating life.

For example, a low concentration gas sensor withvery high sensitivity uses a coarse-porosity hydrophobicmembrane and less restricted capillary to allow moregas molecules to pass through to produce enough sig-nal for better sensitivity. However, this design also al-lows more of the electrolyte’s water molecules to es-cape out to the environment. In other words, an elec-trochemical sensor with high sensitivity would have arelatively short operating life due to evaporation ofmoisture through the porous membrane.

Similarly, the electrolyte composition and the sens-ing electrode material is selected based on the chemi-cal reactivity of the target gas. By careful selection ofthe electrolyte and/or the sensing electrode, one canachieve the selectivity towards the target gas, but thesensitivity may be reduced.

In summary, different electrochemical sensors mayappear very similar, but are constructed with differ-ent materials including such critical elements as sens-ing electrodes, electrolyte composition, and porosityof hydrophobic barriers. Additionally, some electro-chemical sensors use external electrical energy tomake them reactive to the target gas. All componentsof the sensors play a crucial role in determining theoverall characteristics of the sensors.

Principle of Operation

Electrochemical sensors operate by reacting withthe gas of interest and producing an elec-trical signal proportional to the gas concen-tration. A typical electrochemical sensor con-sists of a sensing electrode (or working elec-trode), and a counter electrode separated by athin layer of electrolyte, Figure 3.

Gas that comes in contact with the sensor

Anode Cathode

Electrolyte

+ –

Micro AmmeterGas Molecules

Fig. 2 Basic Sensor

HydrophobicMembrane

CapillaryDiffusionBarrier

Sensing Electrode

Reference Electrode

Counter Electrode

Electrolyte

Fig. 3 Typical Electrochemical Sensor Setup

29


first passes through a small capillary-type opening andthen diffuses through a hydrophobic barrier, and even-tually reaches the electrode surface. This approach isadopted to allow the proper amount of gas to react atthe sensing electrode to produce a sufficient electri-cal signal while preventing the electrolyte from leak-ing out of the sensor, Figure 4.

The gas that diffuses through the barrier reacts atthe surface of the sensing electrode involving either anoxidation or reduction mechanism. These reactions arecatalyzed by the electrode materials specifically devel-oped for the gas of interest.

With a resistor connected across the electrodes, acurrent proportional to the gas concentration flows be-tween the anode and the cathode. The current can bemeasured to determine the gas concentration. Becausea current is generated in the process, the electrochemi-cal sensor is often described as an amperometric gas sensoror a micro fuel cell.

Importance of a Reference Electrode. For a sen-sor requiring an external driving voltage, it is impor-tant to have a stable and constant potential at the sens-ing electrode. In reality, the sensing electrode potentialdoes not remain constant due to the continuous elec-trochemical reaction taking place on the surface of theelectrode. It causes deterioration of the performanceof the sensor over extended periods of time. To improvethe performance of the sensor, a reference electrode isintroduced.

This reference electrode is placed within the elec-trolyte in close proximity to the sensing electrode. Afixed stable constant potential is applied to the sens-ing electrode. The reference electrode maintains thevalue of this fixed voltage at the sensing electrode. Nocurrent flows to or from the reference electrode. Thegas molecules react at the sensing electrode and thecurrent flow between the sensing and the counter elec-trode is measured and is typically related directly to

HydrophobicMembrane

Water and Gas

Gas

ElectrolyteFig. 4 Hydrophobic Membrane:prevents liquid electrolyte fromleaking out.

30


the gas concentration. The value of the voltage appliedto the sensing electrode makes the sensor specific to thetarget gas.

The micro fuel cell-type electrochemical sensors do notrequire an external driving voltage. For example, an elec-trochemical sensor specific to oxygen has an anode, ei-ther Pb or Cd, that supplies electrons for the reduction ofoxygen at the cathode. During the oxidation of the an-ode, the electrons are released which then travel via anexternal circuit to the cathode where oxygen moleculesconsume the electrons as follows:

In acidic electrolyte

Oxidation at the anode: 2Pb + 2H2O → 2PbO + 4H+ + 4e-

Reduction at the cathode: O2 + 4H+ + 4e- → 2H2O

In basic electrolyte

Oxidation at the anode: 2Pb + 4OH- → 2PbO + 2H2O + 4e-

Reduction at the cathode: O2 + 2H2O + 4e- → 4OH-

The overall reaction in both cases is: 2Pb + O2 → 2PbO.These types of sensors do not require a reference electrode.

Major Components

An electrochemical sensor consists of the following ma-jor components:

A. Gas Permeable Membrane (also called hydrophobicmembrane): This is used to cover the sensor’s sens-ing (catalyst) electrode and, in some instances, tocontrol the amount of gas molecules reaching theelectrode surface. Such barriers are typically madeof thin, low-porosity Teflon membranes. Such sen-sors are called membrane clad sensors. Alternatively,the sensing electrode is covered with a high-poros-ity Teflon and the amount of gas molecules reach-ing the electrode surface is controlled by a capil-

31


lary. Such sensors are referred to as capillary-typesensors. Besides offering a mechanical protection tothe sensor, the membrane performs the additionalfunction of filtering out unwanted particulates. Se-lecting the correct pore size of the membrane andcapillary is necessary to transfer the proper amountof gas molecules. The pore size should be such asto allow enough gas molecules to reach the sensingelectrode. The pore size should also prevent liquidelectrolyte from leaking out or drying out the sen-sor too quickly.

B. Electrode: The selection of the electrode mate-rial is very important. It is a catalyzed materialwhich performs the half cell reaction over a longperiod of time. Typically, the electrode is madefrom a noble metal, such as platinum or gold, andcatalyzed for an effective reaction with gas mol-ecules. Depending on the design of the sensor, allthree electrodes can be made of different materi-als to complete the cell reaction.

C. Electrolyte: The electrolyte must facilitate the cellreaction and carry the ionic charge across the elec-trodes efficiently. It must also form a stable refer-ence potential with the reference electrode and becompatible with materials used within the sensor.If the electrolyte evaporates too quickly, the sensor’ssignal will deteriorate.

D. Filter: Sometimes a scrubber filter is installed infront of the sensor to filter out un-wanted gases. There is a limited selec-tion of filters, each with different de-grees of effectiveness. The most com-monly used filter medium is activatedcharcoal, as shown in Figure 5. The ac-tivated charcoal filters out most chemi-cals with the exception of carbon mon-oxide and hydrogen gases. By properlyselecting the filter medium, an elec-

Charcoal Filter

CapillaryDiffusionBarrier

Sensing Electrode

Reference Electrode

Counter Electrode

Electrolyte

Fig. 5 Filtering with Activated Charcoal

32


trochemical sensor can be made more selec-tive to its target gases.

Choosing the suitable materials for the abovecomponents, and arranging the geometry of all thesecomponents to determine the optimum operating per-formance presents a challenge to scientists. Minorvariations in the details of the sensor design can havea profound influence on the sensor’s accuracy, re-sponse time, sensitivity, selectivity, and life expectancy.

Importance of Oxygen. The reactions at the sens-ing electrode(anode) for some gases are as follows:

CO + H2O → CO2 + 2H+ + 2e-

H2S + 4H2O → H2SO4 + 8H+ + 8e-

NO + 2H2O → HNO3 + 3H+ + 3e-

H2 → 2H+ + 2e-

2HCN + Au → HAu(CN)2 + H+ + e-

Simultaneously, the reactions at the counter elec-trode (cathode) need oxygen molecules to completethe process:

O2 + 4H+ + 4e- → 2H2O

An inadequate supply of oxygen to complete thereaction will shorten the life of the sensors, hence thesensors will not operate properly.

Sensors involving a reduction reaction of the tar-get gas—such as the reduction of nitrogen dioxide,chlorine, and ozone—at the cathode produce wateras a byproduct. At the anode, water is simultaneouslyoxidized. Such sensors do not require the presenceof oxygen to function properly, as shown by the fol-lowing:

NO2 + 2H+ + 2e- → NO + H2O

Cl2 + 2H+ + 2e- → 2HCl

O3 + 2H+ + 2e- → O2 + H2O

Typical Gases and theRange of Measurement ofElectrochemical Sensors

GAS NAME PPM RANGE

Ammonia, NH3 10

Arsenic Hydride, AsH3 1

Bromine, Br2 30

Carbon Monoxide, CO 300

Chlorine, Cl2 5

Chlorine Dioxide, ClO2 5

Diborane, B2H6 1

Fluorine, F2 10

Germane, GeH4 2

Hydrogen, H2 2000

Hydrogen Chloride, HCl 30

Hydrogen Cyanide, HCN 30

Hydrogen Fluoride, HF 10

Hydrogen Sulfide, H2S 30

Nitric Oxide, NO 100

Nitrogen Dioxide, NO2 50

Oxygen ppm levels to100% by vol.

Ozone, O3 3

Phosphine, PH3 1

Silane, SiH4 50

Sulfur Dioxide, SO2 100

33


Characteristics

There are many different ways that electrochemi-cal sensors are constructed, depending both on thegas to be detected as well as the manufacturer. How-ever, the main characteristics of the sensors are es-sentially very similar. Following are some of the com-mon characteristics of electrochemical sensors:

1. With a three-electrode sensor, there is nor-mally a jumper which connects the workingand reference electrodes. If it is removed dur-ing storage, it will take a long time for the sen-sor to stabilize and be ready to be used. Somesensors require a bias voltage between the elec-trodes, and in such cases, the sensors areshipped from the factory with a nine-volt bat-tery powered electronic circuit. It takes any-where from thirty minutes to twenty-four hoursfor the sensor to stabilize, and it will continueto stabilize over a three-week period. When installed in a portable or stationaryinstrument, the sensor cannot be removedfrom power for an appreciable amount of time.It is wise to double-check the instrument be-fore use if batteries or power were removed atsome point. The portable instrument’s cir-cuitry provides a small current needed tomaintain the sensor in the ready-to-use condi-tion, even if the instrument is turned off. Two-electrode sensors do not require anybias voltage. For example, oxygen sensors donot require a bias voltage.

2. Most of the toxic gas sensors require a smallamount of oxygen to function properly. Thereis a vent hole on the side or back of the sensorfor this purpose. It is wise to double-check withthe manufacturer in applications that use non-oxygen background gas.

34


3. Electrolyte within the sensor cell is an aqueoussolution separated by a hydrophobic barrierwhich will not allow the aqueous solution toleak out. However, water vapor can passthrough, just as other gas molecules can. Inhigh humidity conditions, prolonged exposurecan cause excessive water to build up and cre-ate leakage. In low humidity conditions, thesensor can dry out. Sensors that are designedto monitor high gas concentrations have lessporous barriers to limit the amount of gasmolecules that pass through, and therefore arenot affected by the humidity as much as sen-sors that are used to monitor low gas concen-trations, which have more porous barriers andallow a more free exchange of water molecules.

Pressure and Temperature

Electrochemical sensors are minimally affected bypressure changes. However, it is important to keep theentire sensor within the same pressure since differen-tial pressure within the sensor can cause sensor dam-age. Electrochemical sensors are also quite sensitiveto temperature and, therefore, the sensors are typi-cally internally temperature-compensated. However,it is better to keep the sample temperature as stableas possible.

In general, when the temperature is above 25°C,the sensor will read higher; when it is below 25°C, itwill read lower. The temperature effect is typically 0.5%to 1.0% per degree centigrade, depending on themanufacturer and type of sensor.

Selectivity

Electrochemical sensors are generally fairly selec-tive to the target gas they are designed for. The de-gree of selectivity depends on the type of sensor, thetarget gas, and the concentration of gas the sensor is

List of Typical GasInterference Ratios

for CO Sensors

Gas Without Filter With Filter

H2S 0.3:1 10:1

SO2 2:1 20:1

NO 3.3:1 10:1

NO2 1.6:1 10:1

H2 2:1 2:1

The higher the interference ratio,the less effect an interferencegas has on the sensor.

35


designed to detect. The best electrochemical sensoris for the detection of O2, which has good selectivity,is very reliable, and has a long life expectancy. Otherelectrochemical sensors are subject to interferencefrom other gases. A typical list of interference ratiosfor CO sensors is shown on page 34 as an example.

The higher the ratio, the less the effect of inter-ference gas on the sensor. The interference data aretaken using relatively low gas concentrations. In ac-tual applications, interference concentrations can bequite high, causing false readings and/or alarms.

Life Expectancy

The life expectancy of an electrochemical sensordepends on several factors, including the gas to bedetected and the environmental conditions in whichthe sensor is used.

Generally, a one- to three-year life expectancy isspecified. In reality, the life expectancy will be highlydependent on the total amount of gas exposed tothe sensor during its life, as well as other environ-mental conditions, such as temperature, pressure andhumidity.

Summary

Electrochemical sensors require very little powerto operate. In fact, their power consumption is thelowest among all sensor types available for gas moni-toring. For this reason, the sensors are widely used inportable instruments that contain multiple sensors.They are the most popular sensors in confined spaceapplications.

A sensor’s life expectancy is predicted by its manu-facturer under conditions that are considered normal.However, the life expectancy of the sensor is highlydependent on the environmental contaminants, tem-perature, and humidity to which it is exposed.

Typical Toxic GasElectrochemical

Sensor Specification

Sensor Type: 2 or 3 electrodes; mostly 3electrodes

Range: 2-10 times permissible exposure limit

Life Expectancy: 12 to 24 months normal;depends on manufacturer and sensor

Temperature Range: –40oC to +45oC

Relative Humidity: 15-95% noncondensing

Response Time T80 : < 50 sec.

Long Term Drift: drift down 2% per month

37

Chapter 3 Catalytic Combustible Gas Sensors

Chapter 3

Catalytic CombustibleGas Sensors

Catalytic bead sensors are used primarily to de-tect combustible gases. They have been in usefor more than 50 years. Initially, these sensors

were used for monitoring gas in coal mines, wherethey replaced canaries that had been used for a longperiod of time.

The sensor itself is quite simple in design and iseasy to manufacture. In its simplest form, as used inthe original design, it was comprised of a single plati-num wire. Catalytic bead sensors were produced allover the world by a large number of different manu-facturers, but the performance and reliability of thesesensors varied widely among these various manufac-turers. A catalytic bead sensor is shown in Figure 1.


Combustible gas mixtures will not burn until theyreach an ignition temperature. However, in the pres-ence of certain chemical media, the gas will start toburn or ignite at lower temperatures. This phenom-enon is known as a catalytic combustion. Most metaloxides and their compounds have these catalytic prop-erties. For instance, volcanic rock, which is comprisedof various metal oxides, is often placed in gas burningfireplaces. This is not only decorative, but it also helps

Fig. 1 A Catalytic Bead Sensor

38


the combustion process and results in cleaner andmore efficient burning in the fireplace. Platinum,palladium, and thoria compounds are also excellentcatalysts for combustion. This explains why the auto-mobile exhaust system is treated with platinum com-pounds and is called a catalytic converter. This kindof gas sensor is made on the basis of the catalytic prin-ciple, and therefore is called the catalytic gas sensor.

A gas molecule oxidizes on the catalyzed surfaceof the sensor at a much lower temperature than itsnormal ignition temperature. All electrically conduc-tive materials change their conductivity as tempera-ture changes. This is called the coefficient of temperatureresistance (Ct). It is expressed as the percentage ofchange per degree change in temperature.

Platinum has a large Ct in comparison to othermetals. In addition, its Ct is linear between 500°C to1000°C, which is the temperature range at which thesensor needs to operate. Because the signal from thesensor is linear, this means that the concentration ofgas readings are in direct proportion to the electricalsignal. This improves the accuracy and simplifies theelectronic circuitry. Also, platinum possesses excel-lent mechanical properties. It is physically strong andcan be transformed into a fine wire which can be pro-cessed into small sensor beads.

Furthermore, platinum has excellent chemicalproperties. It is corrosion resistant and can be oper-ated at elevated temperatures for a long period of timewithout changing its physical properties. It is capableof producing a constant reliable signal over an ex-tended period of time.

The electrical circuit used to measure the outputof catalytic sensors is called a Wheatstone bridge, in honorof English physicist and inventor Sir Charles Wheat-stone (1802-75). Wheatstone bridges are commonlyused in many electrical measurement circuits. Asshown in Figure 2, four circuit branches are arranged

OUTPUT

RB

R1Acti

ve Bea

d

Reference Bead

D.C.

POW

ER

RB

Fig. 2 A catalytic bead sensorWheatstone bridge–a circuit formeasuring an unknown resist-ance by comparing it with knownresistances.

39


in a square. The source of the electrical current is con-nected, and between the other pair of opposite cor-ners, the output measurement circuit is connected.

In operation, R1 is the trim resistor that keeps thebridge balanced. A balanced bridge has no output sig-nal. Resistor value RB and trim pot R1 are selected withrelatively large resistance values to ensure proper func-tion of the circuit. When the gas burns on the activesensor surface, the heat of combustion causes the tem-perature to rise, which in turn changes the resistanceof the sensor. As the bridge is unbalanced, the offsetvoltage is measured as the signal. It is important thatthe reference sensor or bead maintains a constant re-sistance during the exposure to the combustible gas;otherwise, the measured signal will be inaccurate.

Evolution of the sensor. The original catalytic sen-sor was a coil-shaped platinum wire. The coiled shape,illustrated in Figure 3, was used to obtain a compactgeometry for efficient heating and to produce a strongenough signal to function as a gas sensor. Unfortu-nately, despite the excellent physical and chemicalproperties of platinum, it is a poor catalyst for com-bustion of hydrocarbon gases.

For the proper detection of hydrocarbon gases,the sensor requires a heated surface temperature be-tween 900°C and 1000°C so that the sensor can prop-erly react with gases at a sufficiently high and stablerate. At this temperature, however, the platinum startsto evaporate. The evaporation rate increases as thegas molecules start to react with the sensor and as thesensor temperature increases. This causes a reduc-tion in the cross-section of the platinum wire, and, asa result, the resistance increases. This affects thesensor’s operating temperature, which shows up as zeroand span drifts.

The reference wire ideally should be the same asthe active wire, with the same geometry and operat-ing temperature, but should be nonreactive with the

Fig. 3 Hot Wire Sensor

40


gas. This is not practically possible, however. A com-promise is made by operating the reference wire at atemperature that is substantially lower so that no oxi-dation takes place in the presence of hydrocarbons.In addition, the reference wire is chemically treatedto reduce the catalytic property of the platinum. Thismay also be achieved by coating platinum wire with anon-catalytic metal, such as gold.

Another problem with hot platinum wire is that itbecomes very soft at a temperature of 1000°C. There-fore, it is difficult to maintain its coil shape. Also, thecoefficient of thermal resistance becomes less linearas the temperature increases. This situation also re-sults in poor zero and span quality of the sensor, aswell as a relatively short operating life.

One way to improve stability of the sensor is tocoat the platinum wire with suitable metal oxides.Thus, the final step is to treat the finished sensor orbead with a catalyst, such as platinum, palladium orthoria compounds. Figure 4, shows the sensor bead.

PLATINUM WIRE

CATALYTICALLY-TREATEDMETAL OXIDE

Fig. 4 A Catalytic Bead Sensor

The construction of the catalytic sensor bead isanalogous to constructing a building by using rein-forced concrete. The coating makes the sensor physi-cally very rugged. The sensor becomes a very small

41


mass which helps make it resistant to shock and vibra-tions. Most importantly, the catalyst coating reducesthe temperature needed to achieve a stable signal forhydrocarbons between 400°C and 600oC.

The use of fine diameter wire not only reducesthe size of the sensor, but it also increases the signal,because finer wire has a higher magnitude of resistivevalue and the signal output is the percentage changeof total wire resistance. This also reduces power con-sumption.

The reference sensor can be constructed in thesame way as the active sensor, with the exception thatthe catalyst chemical is eliminated. The bead can befurther treated with chemicals, such as potassium, toprevent the reference bead from reacting with the gas.A near perfectly compensated pair of sensors is nowpossible. The sensor is called a “catalytic” sensor be-cause the use of the catalyst is the main ingredientinvolved in the proper functioning of the sensor. Thecatalytic sensor is stable, reliable, accurate, and rug-ged, and has a long operating life. The output is lin-ear because the platinum wire has a good linear coef-ficient of thermal resistance.

Characteristics

The sensor’s output is directly in proportion tothe rate of oxidation. The maximum output of thesignal occurs at about the stoichiometric1 mixture ofthe gas, or it is based on the theoretical combustionreaction formula. Methane, for example:

CH4 + 2O2 + 8N2 → CO2 + 2H2O + 8N2

It takes 10 moles of air for one mole of methaneto complete the reaction, assuming there is one partof oxygen and four parts of nitrogen in air.

Therefore, for a theoretical combustion to takeplace, one part of methane will require 10 parts of air

1 Pertaining to substances that are in theexact proportions required for a given re-action.

42


to complete the combustion, or theoretically 9.09%of methane in a mixture of air.

For a sensor to detect methane, the signal out-put will respond linearly from 0–5% of methane(which is 100%LEL). As the concentration reachesclose to the stoichiometric value of 9%, the signalincreases very rapidly and peaks at around 10%. Thesignal starts to drop slowly as the concentration ofgas passes approximately 20%; after 20% it dropsstraight down to a level that reflects no output as theconcentration of gas reaches 100%. Figure 5, illus-trates this effect.

0

% Methane Gas Concentration

Sens

or O

utpu

tin

vol

ts 6

10

8

4

2

LEL UEL

5 10 15 20 40 60 80 100

0-3%: NEARLINEAR; 3-5%: SLIGHTLY LESS LINEAR

Fig. 5 Sensor Output vs. Gas Concentration

Consider another example, propane. The reactionformula for propane is:

C3H8 + 5O2 + 20 N2 → 3CO2 +4H2O + 20N2

or one part of propane per 25 parts of air for theoreti-cal combustion of propane. The actual theoreticalcombustion concentration for propane is 3.85%.

The LEL for methane is 5% and for propane is2.1%. This value is near half of the theoretical com-bustion value. There is a safety factor of 2 added toensure safety.

43


Sensor Operation Factors

There are several factors affecting the operationof the catalytic sensor.

1. Catalyst Poisoning: There are chemicals whichwill deactivate the sensor and cause the sensor to losesensitivity and eventually become totally nonrespon-sive to gases. The most common chemicals that canpoison catalytic sensors are those that contain silicon,such as the common oil and lubricants with siliconcompounds used as additives in machinery. Sulfurcompounds, which are often released with gases, chlo-rine, and heavy metals also cause the poisoning of thesensor.

The exact cause of this poisoning is very difficultto identify. Some chemicals, with very small concen-trations, will totally destroy the sensor. There have beeninstances in which the silicon contained in simplehand lotions has caused problems with catalytic sen-sors.

2. Sensor Inhibitors: Chemicals such as halogencompounds, which are used in fire extinguishers andFreon used in refrigerants, will inhibit the catalyticsensor and cause it to temporarily lose the ability tofunction.

Normally, after 24 or 48 hours of exposure to am-bient air, the sensor starts to function normally. Theseare just a few typical chemicals that inhibit the sensorperformance and are by no means to be consideredas the sole possible inhibitors.

3. Sensor Cracking: The sensor, when exposedto excessive concentration of gases, excessive heat,and the various oxidation processes that take placeon the sensor surface, may eventually deteriorate.Sometimes this will change the zero and span set-ting of the sensor.

4. Correction Factors: Catalytic sensors are most

44


commonly calibrated to methane for 0-100% LEL fullscale range.

The manufacturers generally provide a set of cor-rection factors that allow the user to measure differenthydrocarbons by simply multiplying the reading bythe appropriate correction factor to obtain the read-ing of a different gas. The reason for using methaneas the primary calibration gas is that methane has asaturated single bond that requires the sensor to op-erate at the highest temperature in comparison toother hydrocarbons. For instance, a typical catalyticsensor for methane gas may require a 2.5-volt bridgevoltage to obtain a good signal, while the same sensorwill only need 2.3 volts for butane gas. Therefore, ifthe sensor is set to read butane, it will not read meth-ane properly.

In addition, methane gas is a very common gasand is often encountered in many applications. Fur-thermore, it is also easy to handle and has the abilityto be mixed into different concentrations easily.However, it should be noted that the correction fac-tors are a set of numbers that should be used withgreat care. The correction factors can vary from sen-sor to sensor, and they can even change on the samesensor as the sensor ages. Therefore, the best way toobtain precise readings for a specific gas is to actuallycalibrate the sensor to the gas of interest directly.

5. Percent LEL for Mixtures of Hydrocarbons:For combustion to take place, the following require-ments must be present:

a. Combustible mixtureb. Oxygenc. Ignition source

This is sometimes referred to as the combustiontriangle. But in real life, the process of igniting a com-bustible mixture is much more complicated. The en-

Relative SensitivityAs an example for a typical sensorcalibrated for 100% LEL methanegas, the relative sensitivity toother gases is as follows:

Gas ReadingMethane 100%Propane 60%n-Butane 60%n-Pentane 50%n-Hexane 45%Methanol 100%Ethanol 70%iso-Propyl Alcohol 60%Acetone 60%Methyl Ethyl Ketone 50%Toluene 45%

45


vironmental conditions, such as pressure, tempera-ture, temperature of the ignition source, and evenhumidity can have an affect on the combustible mix-ture concentration.

If two or more chemicals are involved, it is not evenpossible to calculate and determine the combustionrange of the mixture. Therefore, it is best to considerthe worst-case scenario and calibrate the sensor ac-cordingly. Furthermore, a sensor calibrated at a per-centage LEL for one gas cannot necessarily be usedfor other gases. Many instruments on the market to-day have a scale unit as a percentage of LEL withoutindicating that the unit is calibrated on methane.Therefore, if the unit is used for some other gas ormixture of gases, the data can be totally meaningless.

For example, a catalytic sensor calibrated on meth-ane produces lower readings when exposed to hydro-carbons of higher carbon content, while infrared in-struments will produce much higher readings if ex-posed to a higher carbon content gas. This is a verycommon mistake made by many users of gas detec-tion equipment.

Summary

A catalytic sensor is relatively easy to manufacture.However, the quality of the sensor varies quite drasti-cally from one manufacturer to another.

The overall technology of making a sensor forthe market is more of an art than a predictable sci-entific event. This is particularly true in selecting,preparing and processing all the chemicals neededto make the final sensor. There are too many vari-ables in the process that inhibit the making of a pre-dictable final product. Therefore, most users of cata-lytic sensors select their sensors based on the reputa-tion of the manufacturer.

Typical Specificationsfor Catalytic Sensors

Sensor Type: Diffusion catalytic bead


Response Time: 10 to 15 sec. to 90% ofreading

Accuracy: ±5%

Repeatability: 2%

Drift: 5–10% per year

Life Expectancy: Up to 3 years; dependingon application

Sensors can be remotely mounted up to2,000-3,000 meters, depending on themanufacturer and cable size used to wirethe sensor.

47

Chapter 4 Solid-State Gas Sensors

Chapter 4

Solid-State Gas Sensors

W hen scientists were doing research work re-lated to semiconductor p-n junctions,1 theydiscovered that these junctions were sensitive

to environmental background gases.At that time, such a behavior was considered a

problem. This problem, however, was solved by en-capsulating the semiconductor chip so that it was nolonger exposed to the outside environment. Subse-quently, unsuccessful attempts were made to utilizethe sensitivity of the semiconductor junction as a gasdetection device.

It wasn’t until 1968 that Mr. N. Taguchi marketeda simple semiconductor, or a solid-state sensor, for thedetection of hydrocarbons in LEL combustible ranges.The intention was to provide an alternative to thepopular catalytic bead sensor, which suffered fromseveral problems, including loss of sensitivity with timedue to poisoning and burning out when exposed tohigh gas concentrations.

In 1972, International Sensor Technology (IST) inIrvine, California introduced a solid-state sensor forthe detection of hydrogen sulfide in a range of 0-10ppm. A few years later, IST developed solid-state sen-sors for the detection of more than 100 different haz-ardous gases at low ppm levels. This was a significantdevelopment, since OSHA was being formed at about

1 The positive and negative junctionsin a semiconductor.

Fig. 1 A solid-state sensorused for detecting more than ahundred toxic gases.

48


the same time and began to regulate acceptable gasconcentration levels for safety at the workplace.

Today, solid-state sensors are available for the de-tection of more than 150 different gases, includingsensors for the gases which could otherwise only bedetected using expensive analytical instruments.

There are now several manufacturers of solid-statesensors, but each sensor has different characteristicsand different manufacturers offer different levels ofperformance and quality.

Properly manufactured, solid-state sensors offer avery long life expectancy. It is not unusual to find fullyfunctional sensors that were installed 30 years ago.

The Rise in Popular Use of Solid-State Sensors. Inthe early 1980s, Japan passed a law that required gasdetectors to be installed in residential apartmentswhere gas bottles were being used. For this huge mar-ket, the competition was between solid-state and cata-lytic bead sensors.

While there were some initial complaints aboutsolid-state sensors that produced false alarms, this wasfar outweighed by the long life the sensor provided.Because catalytic sensors burn the gas being detected,sensor material is consumed and changed in the pro-cess and the sensor eventually burns out.

With solid-state sensors, on the other hand, gassimply “adsorbs” onto the sensor surface, changingthe resistance of the sensor material. When the gasdisappears, the sensor returns to its original condi-tion. No sensor material is consumed in the process,and hence the solid-state sensors offer a long life ex-pectancy.

After a few years of usage, the catalytic sensors hadbecome less popular for such applications due to theneed for frequent sensor replacement.


A solid-state sensor consists of one or more metal

49


oxides from the transition metals, such as tin oxide,aluminum oxide, etc. These metal oxides are preparedand processed into a paste which is used to form abead-type sensor. Alternatively, thick or thin film-chipsensors are made when the metal oxides are vacuumdeposited onto a silica chip, in a fashion similar tomaking semiconductors.

A heating element is used to regulate the sen-sor temperature, since the finished sensors exhibitdifferent gas response characteristics at differenttemperature ranges. This heating element can be aplatinum or platinum alloy wire, a resistive metaloxide, or a thin layer of deposited platinum. Thesensor is then processed at a specific high tempera-ture which determines the specific characteristicsof the finished sensor.

In the presence of gas, the metal oxide causes thegas to dissociate into charged ions or complexeswhich results in the transfer of electrons. The built-in heater, which heats the metal oxide material to anoperational temperature range that is optimal for thegas to be detected, is regulated and controlled by aspecific circuit.

A pair of biased electrodes are imbedded into themetal oxide to measure its conductivity change. Thechanges in the conductivity of the sensor resultingfrom the interaction with the gas mol-ecules is measured as a signal. Typically,a solid-state sensor produces a verystrong signal, especially at high gas con-centrations.

There are different ways of makingsolid-state sensors, each arrangementmaking the sensor’s performancecharacteristics different. Two typicalstyles are the following:

1. Bead-type sensor (Figure 2)2. Chip-type sensor (Figure 3, next page)

Collector

HeaterControl

VOUT

PlatinumCoil Heater

Fig. 2. Schematic Diagram of a Bead-type Sensor

50


Most solid-state sensors have threeor four pins, depending on how theheater and bias electrodes are con-nected.

Characteristics

Solid-state sensors are among themost versatile of all sensors, as they de-tect a wide variety of gases, and can beused in many different applications. Dif-

ferent response characteristics are achieved by vary-ing the semiconductor materials, processing tech-niques, and sensor operating temperature.

Among the unique attributes of the solid-state sen-sor are the abilities of the sensor to detect both lowppm levels of gases, as well as high combustible levels.

Longevity. The main strength of the solid-statesensor is its long life expectancy, as the sensor typi-cally lasts 10 years or more in clean applications. Thisis a major advantage compared to other sensor types,such as catalytic bead or electrochemical sensors,which typically last only one to two years.

However, while solid-state sensors have a longerlife expectancy, they are also more susceptible to in-terference gases than the other types of sensors. Thus,in applications where other background gases arepresent, solid-state sensors may trigger false alarms.

In certain instances, the interferences from othergases are minimized by using appropriate filtering ma-terials that absorb all other gases except the gas to bedetected.

For example, a solid-state sensor for monitoringcarbon monoxide and hydrogen can be equipped witha charcoal filter which eliminates the majority of in-terfering gases. This way the sensor performs very welland becomes very selective for those two gases.

Versatility. The versatility of the solid-state sensor

Fig. 3 Chip-type Sensor

Heater Resistor

Silicon Substrate

Termination Termination

Collector

51


is one of its main advantages. For example, in chemi-cal plants, a gas monitoring system may involve themonitoring of many different gases and ranges, oreven the same gas with multiple ranges.

Often, the lower ranges need to be monitored forcertain gases for toxic concentrations while simulta-neously, the same gas needs to be monitored in thecombustible range for explosive concentrations.

The solid-state sensor is capable of detecting gasin both ranges. This greatly simplifies the system de-sign and maintenance required because it eliminatesor minimizes the use of multiple sensor technologieswhich must be designed and maintained differently.

A list of gases that are detectable by using IST’ssolid-state sensors is shown in Table 1 (pages 52-53).

The typical ppm ranges which are chosen arethree to five times the permissible exposure limitsfor an eight-hour work day or ranges that can bebased on the sensitivity as well as the interferencecharacteristic of the sensor, whichever is the mostpractical in an application.

Typical Specificationsfor Solid-State Sensors*

Accuracy: ±3% to 10% of full scale

Response Time: T80 ranges from 20seconds to 90 seconds


Humidity: 5–9%

Life Expectancy: 10+ years

Power Consumption: Approx. 300mW

*Actual specifications will vary depending ongas and range.

52


Acetic Acid 100 —-Acetone 100 yesAcetonitrile 100 —-Acetylene 50 yesAcrolein(Acrylaldehyde) 50 —-Acrylic Acid 100 —-Acrylonitrile 50 yesAllyl Alcohol —- yesAllyl Chloride 200 —-Ammonia x50 yesAnisole 100 —-Arsenic Pentafluoride 5 —-Arsine 1 —-Benzene 50 yesBophenyl 50 yesBoron Trichloride 500 —-Boron Triflluoride 500 —-Bromine 20 —-Butadiene 50 yesButane 400 yesButanol 1000 yesButene —- yesButyl Acetate 100 yesCarbon Disulfide 50 —-Carbon Monoxide 50 yesCarbon Tetrachloride 50 —-Cellosolve acetate 100 —-Chlorine 10 —-Chlorine Dioxide 10 —-Chlorobutadiene —- yesChloroethanol 200 —-Chloroform 50 —-Chlorotrifluoroethylene —- yesCumene —- yesCyanogen Chloride 20 —-Cychlohexane 100 yesCyclopentane 50 —-

Table 1: IST ’s Solid-State Sensors Gas List

GAS FULL-SCALE RANGE

Low ppm LELor higher



Deuterium yesDiborane 10 —-Dibromoethane 50 —-Dibutylamine —- yesDichloroethane (EDC) 50 yesDichlorofluoroethane 100 —-Dichloropentadiene 50 —-Dichlorosilane 50 —-Diesel Fuel 50 yesDiethyl Benzene —- yesDiethyl Sulfide 10 —-Difluorochloroethane —- yesDifluoroethane (152A) —- yesDimethyl Ether —- yesDimethylamine (DMA) 20 —-Epichlorohydrin 50 —-Ethane 1000 —-Ethanol 200 yesEthyl Acetate 200 yesEthyl Benzene 200 yesEthyl Chloride 100 yesEthyl Ether 100 yesEthylene 100 yesEthylene Oxide 5 yesFluorine 20 —-Formaldehyde 15 —-Freon-11 1000 —-Freon-12 100 —-Freon-22 100 —-Freon-113 100 —-Freon-114 1000 —-Freon-123 1000 —Fuel Oil or Kerosene —- yesGasoline 100 yesGermane 10 —-Heptane 1000 yesHexane 50 yes

53


Table 1 (continued) : IST ’s Solid-State Sensors Gas List





Hexane 100 yesHydrazine 5 —-Hydrogen 50 yesHydrogen Bromide 50 —-Hydrogen Chloride 50 —-Hydrogen Cyanide 20 —-Hydrogen Fluoride 20 —-Hydrogen Sulfide 5 yesIsobutane 1000 yesIsobutylene —- yesIsopentane 1000 —-Isoprene —- yesIsopropanol 200 yesJP4 1000 yesJP5 1000 yesMethane 100 yesMethanol 200 yesMethyl Acetate 30 —-Methyl Acrylate 60 —-Methyl Bromide 20 —-Methyl Butanol —- yesMethyl Cellosolve —- yesMethyl Chloride 100 yesMethyl Ethyl Ketone 100 yesMethyl Hydrazine 5 —-Methyl Isobutyl Ketone 200 yesMethyl Mercaptan 30 —-Methyl Methacrylate 100 yesMethyl-Tert Butyl Ether —- yesMethylene Chloride 20 yesMineral Spirits 200 yesMonochlorobenzene —- yesMonoethylamine 30 —-Morpholine 500 —-Naptha 1000 yesNatural Gas 1000 yes

Nitric Oxide 20 —-Nitrogen Dioxide 20 —-Nitrogen Trifluoride 50 —-Nonane 2000 —-Pentane 200 yesPerchloroethylene 200 —-Phenol 100 —-Phosgene 50 —-Phosphine 3 —-Phosphorus Oxychloride 200 —-Picoline —- yesPropane 100 yesPropylene 100 yesPropylene Oxide 100 yesSilane 10 —-Silicon Tetrachloride 1000 —-Silicon Tetrafluoride 1000 —-Styrene 200 yesSulfur Dioxide 50 —-Tetrahydrofuran 200 yesTetraline 100 —-Toluene 50 yesToluene Diisocyanate 15 —-Trichloroethane 50 —-Trichloroethylene 50 yesTriethylamine (TEA) 100 —-Trifluoroethanol 25 —-Trimethylamine (TMA) 50 —-Tungsten Hexafluoride 50 —-Turpentine —- yesVinyl Acetate 1000 yesVinyl Chloride 20 yesVinylidene Chloride 50 —-Xylene 100 —-

55

Chapter 5 Infrared Gas Sensors

Chapter 5

Infrared Gas Sensors

I nfrared (IR) gas detection is a well-developedmeasurement technology. Infrared gas analyzershave a reputation for being complicated, cumber-

some, and expensive. However, recent technical ad-vancements, including the availability of powerful am-plifiers and associated electronic components, haveopened a new frontier for infrared gas analysis. Theseadvancements have resulted from an increase in de-mand in the commercial sector, and these demandswill likely continue to nourish the advancement of thistechnology.

Gases to be detected are often corrosive and reac-tive. With most sensor types, the sensor itself is directlyexposed to the gas, often causing the sensor to driftor die prematurely.

The main advantage of IR instruments is that thedetector does not directly interact with the gas (orgases) to be detected. The major functional compo-nents of the analyzer are protected with optical parts.In other words, gas molecules interact only with a lightbeam. Only the sample cell and related componentsare directly exposed to the gas sample stream. Thesecomponents can be treated, making them resistant tocorrosion, and can be designed such that they are eas-ily removable for maintenance or replacement.

Today, many IR instruments are available for a widevariety of applications. Many of them offer simple,

Fig. 1 An example of an IRgas monitor with the gas cellassembly exposed.

Gas Cell

56


rugged, and reliable designs. In general, for toxic andcombustible gas monitoring applications, IR instrumentsare among the most user friendly and require the leastamount of maintenance. There is virtually an unlimitednumber of applications for which IR technology can beused. Gases whose molecules consist of two or more dis-similar atoms absorb infrared radiation in a unique man-ner and are detectable using infrared techniques. Infra-red sensors are highly selective and offer a wide rangeof sensitivities, from parts per million levels to 100 per-cent concentrations. This chapter provides general in-formation, with a special emphasis on instruments usedfor area air quality and safety applications.


The infrared detection principle incorporates only asmall portion of the very wide electromagnetic spectrum.The portion used is that which we can feel as heat. Thisis the region close to the visible region of the spectrumto which our eyes are sensitive. Electromagnetic radia-tion travels at close to 3 x 108 m/sec and has a wave-likeprofile. Let’s review the basic physics of electromagneticradiation by defining the terminology involved with it.

Wave: Similar to a wave in the ocean, the electro-magnetic radiation waves oscillate, one wave followedby another. There are both electromagnetic and me-chanical waves, with mechanical waves having a muchlonger wavelength. Figure 2 illustrates a mechanical wave.

1 sec 2 sec

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

1 cm 2 cm

Frequency = 10 Hz. Wavelength = 0.1 cm, Wave Number = 10 cm-1

Fig. 2 A simple mechanical wave showing 10 waves per centimeter toillustrate the concept of the wave.

57


Frequency: Number of waves per second passingthrough a point. An electromagnetic wave travels atthe speed of light which is 300 million meters per sec-ond, or 3 x 108 m/sec. Therefore, the frequency is thespeed of light divided by the wavelength, and is ex-pressed as the number of waves per second, or hertz(Hz).

Wavelength: The distance between two peaks ofthe wave, or the spacing between two waves. It is com-monly expressed in microns. It is a very popular termused in representing gas molecular absorption bandsas well as optical component characteristics.

Wave number:The number of waves in one centi-meter. It is the reciprocal of wavelength. Since 1 mi-cron = 10-6 m = 10-4 cm, the reciprocal of one micron is1/10-4 (10,000 wave numbers per cm), and 2 microns =5000 wave numbers per cm. The formula is:

Wave number = 1/wavelength

Micron: A common unit used to express wave-length in the infrared region. It is one millionth of ameter(10-6m) or a micrometer, and is abbreviated as µ.

Transmittance: The ratio of transmitted radiationenergy to the incident energy. The energy not trans-mitted is absorbed and reflected. It is used to specifyoptical parts.

Absorbance: Opposite of transmittance. Used todescribe the amount of energy absorbed by gas mol-ecules. Both percent absorption and percent transmittanceare used as the y-axis versus wave number or wavelengthas the x-axis in the infrared spectra.

Wave number and wavelength are common termsused by scientists to describe the infrared region forgas analysis because they provide a convenient methodto express radiation frequency and the mechanisms ofinteraction between infrared radiation and gas mol-ecules. Mathematically, they are the reciprocals of each

Transmittance = 100

= 65%65

65Transmitted Energy

25

AbsorbedEnergy

100Incident Energy

10Reflected Energy

58


other. For example, methane gas has the absorption wave-length of 3.4 microns, or a wave number of 2941 cm-1.(Figure 4, on page 59, shows a spectroscopic descriptionof methane gas which illustrates that methane has a strongabsorption peak at 3.4µ, or a wave number of 2941 cm-1.)

Electromagnetic waves propagate through space ormatter by oscillating electric and magnetic fields. In avacuum, they travel at the speed of light. The completerange of frequencies of these waves is called the electro-magnetic spectrum.

These frequencies range from gamma rays of 1020 Hzto radio waves of 106 Hz. They are classified from higherto lower frequencies as gamma rays, x-rays, ultravioletlight, visible light, infrared light, microwaves, and radiowaves. Figure 3 shows the electromagnetic spectrum.

Visible light, at about 4 x 1014 Hz (or 0.4 to 0.7 mi-crons), is actually only a very narrow portion of the spec-trum. Infrared is just below visible light, and this explainswhy we feel, but do not see, temperature. The infraredregion is most useful for gas analysis because absorptionby gas molecules is unique and selective in this region.

Unique Gas Absorption “Fingerprints.” The com-plexity of the gas molecules determines the number ofabsorption peaks. The more atoms that form a molecule,the more absorption bands that will occur. The regionin which this absorption occurs, the amount of absorp-tion, and the specific character of the absorption curveis unique to each gas. Gas molecules can be fingerprintedusing their absorption characteristics and archived for

Fig. 3 Location of Infrared in the Electromagnetic Spectrum

4 x 1014

Frequency (cycles/sec.)

Radio

8 x 1014

106 1020101810151012109

Television Microwave Infrared Ultra-violet X-rays Gamma rays

59


gas analysis and identification purposes. A library of thesecurves can then be stored in the memory inside an in-strument. When a given gas is scanned by the instrument,the graph is then compared with the stored curves toidentify the gas molecules. This method of gas analysis isthe most popular in analytical chemistry.

In gas monitoring applications, only one specific ab-sorption region is used to quantitatively determine thegas concentration. The wavelengths in this region arebetween 2 and 15 microns or wave numbers of 5,000 to670 cm-1. A typical spectroscopy scan of methane isshown in Figure 4.

The scan shows that methane gas has a strong absorp-tion peak at 3.4 microns, which is the wavelength usedfor methane detection. As a matter of fact, most com-mon hydrocarbon gases have strong absorption in the3.4 micron region. On the other hand, carbon dioxide isabsorbed strongly at 4.26 microns, while carbon monox-ide absorbs strongly at 4.7 microns.

Natural Frequencies of Gas Molecules. Gas moleculesare made up of a number of atoms bonded to one an-other. These interatomic bonds are similar tosprings, connecting atoms of various massestogether. Figure 5 illustrates water moleculeswith one oxygen and two hydrogen atoms. Thisbonding vibrates with a fixed frequency calledthe natural frequency.

Fig. 4 A Spectroscopy Scan of Methane Gas

Fig. 5 Water MoleculeHydrogen

HydrogenOxygen

NaturalFrequency

METHANE CH4

CH4

60


All matter has a natural frequency. The GoldenGate bridge in San Francisco has a natural frequency,and the balcony in a theater has a natural frequency.Even though these are mechanical waves, in theorythey are similar to electromagnetic waves. If the windor an earthquake shakes a bridge at the same fre-quency as the bridge’s natural or resonant frequency,this can cause a much more violent vibration of veryhigh amplitude, as shown in Figure 6, resulting insevere damage. People moving around a balcony gen-erate vibrations of certain frequencies, which cancause a similar effect. Thus, the natural frequency ofa structure is a concern for structural engineers.

Gas molecules have a number of natural frequen-cies. The larger the molecules, the more modes ofnatural frequencies they have. Natural frequencies arealso determined by the molecular structure of thechemicals. They are always the same for a given mol-ecule and bonding structure. The particular proper-ties exhibited by the chemical become its signatureand offer clues to identify the molecular structure ofthe given chemical compounds.

Two Ways of Detection. Infrared radiation con-tains a wide spectral content. When this radiation in-teracts with gas molecules, part of the energy has thesame frequency as the gas molecule’s natural fre-quency and it is absorbed while the rest of the radia-tion is transmitted. As the gas molecules absorb thisradiation, the molecules gain energy and vibratemore vigorously.

This vibration results in a rise in the temperatureof the gas molecules. The temperature increases inproportion to gas concentration, and is detected bythe detector. On the other hand, the radiation ab-sorbed by the gas molecules at the particular wave-length will cause a decrease in the original sourcestrength. This radiation energy decrease can be de-tected as a signal also.

Natural Frequency

External Resonant Wave

0

1

2

3

3

2

1

0

1

1

Resultant Wave

0

1

2

2

1

Fig. 6 Energy Absorption byGas Molecules

61


Key Components for Analysis

To get a useful signal for gas analysis, there aremany different components and parts used in the vari-ous designs found today. However, there are no spe-cific rules regarding the selection of each of the com-ponents. The configuration of these instruments de-pends on what needs to be accomplished and the dis-cretion of the designer. Following is a description ofthe major components.

1. Detector: Infrared detectors convert electro-magnetic radiation energy or temperature changesinto electrical signals. There are many types of infra-red detectors and each detector type offers a widerange of performance characteristics. Some of thetypes are described briefly below:

a. Thermoelectric: A detector that converts tem-perature into an electrical signal is commonlyknown as a thermocouple. The junction of dissimi-lar metals generates a voltage potential, which isdirectly proportional to the temperature. Thisjunction can be made into multiple junctions toimprove sensitivity. Such a configuration is calleda thermopile.

Using techniques used in the semiconductorindustry, many junctions are connected in a seriesto multiply the output of the device, yet the detec-tor package is miniaturized and compact. The sizeand mass of the device are important in determin-ing its response time and other characteristics ofthe device.

This detector has a relatively slow responsetime, but offers the advantages of DC stability, re-quiring no bias, and responding to all wavelengths.It is the simplest way of converting light energyinto an electrical signal.

b. Thermistor Bolometer: A bolometer changes re-sistance when incident infrared radiation interacts

Metal 2

Metal 1

– +

T1 T2

DissimilarMetalJunction

T1 = Reference or normally ice point

T2 = Measuring probe

THERMOCOUPLE

62


with the detector. This thermally sensitive semi-conductor is made of a sintered metal oxide ma-terial. It has a high temperature coefficient of re-sistance.

c. Pyroelectric Detector: Pyroelectric materials arecrystals, such as lithium tantalate, which exhibitspontaneous polarization, or a concentrated elec-tric charge that is temperature dependent.

As infrared radiation strikes the detector sur-face, the change in temperature causes a currentto flow. This current is proportional to the inten-sity of the radiation. This detector exhibits goodsensitivity and good response to a wide range ofwavelengths, and does not require cooling of thedetector. It is the most commonly used detectorfor gas monitors.

d. Photon Detector: Photons possess energy basedon their wavelength and intensity. A photon de-tector detects the quantum interaction betweenincident photons and semiconductor material. Aphoton that strikes an electron with sufficient en-ergy can raise the electron from a nonconductingstate into a conducting state. The presence of elec-trons in the conduction band will increase the con-ductivity of the chip, and a bias voltage registersthis change as a signal.

The excitation of electrons requires photonsto have a certain amount of energy. Shorter wave-lengths have higher frequencies, and thereforemore energy. This detector functions in a limitedspectral region, which depends on the detectormaterial being used. Typically, the detector mustbe cooled with a thermoelectric cooler or evenliquid nitrogen for it to function properly. A typi-cal example of such detectors are lead sulfide(PbS) detectors used in the 1-3 micron region, andlead selenide (PbSe) detectors used in the 1-5micron region.

PYROELECTRIC

Pyroelectric Crystal

Infrared

63


e. Luft Detector: The word “luft” is a Germanword meaning “air,” and the original luft detec-tors were designed in Germany. A luft detector con-sists of two chambers, either linked by a micro flowsensor or divided by a diaphragm. The chambersare sealed with a target gas at a low pressure. IRtransparent windows are fitted to seal the cham-bers and the same intensity of pulsed infrared ra-diation is received by both chambers when no tar-get gas is present.

When the sample containing target gas flowsthrough the sample cell, a reduction in radiationenergy is received by the detector chamber, whichcauses the temperature and pressure to drop inthe detector chamber. The amount of temperatureor pressure drop is in direct proportion to the gasconcentration. In the case of linked chambers, thepressure difference between the two chamberscauses a detectable flow, which is measured as asignal. In the case where a diaphragm separatesthe two chambers, a movement of the diaphragmcauses a measurable change in capacitance.

This detector finds many useful applicationsas an analyzer, and has good potential for furtherdevelopment.

f. Photoacoustic Detector: This detector is similarto the luft detector except that the pressure changeis measured by a condenser microphone. Thesample gas is passed through a chamber at a pre-set time interval and the chamber is sealed with afixed volume of sample gas trapped inside. A spe-cific wavelength of infrared radiation is pulsed intothe chamber via an infrared transparent window.The pulsating pressure change is measured by themicrophone as a frequency change which pro-duces the signal.

2. Infrared Source: A regular incandescent lightbulb is a good infrared source. A heated wire filament,

Microphone

Infrared

PHOTOACOUSTIC

SampleOut

SampleIn

IR-TransparentWindow

SAMPLECELL

REFERENCECELL

Micro-Flow Sensoror Diaphragm

Chopper

Motor

IR Source

ReferenceChamber

DetectorChamber

LUFT

64


similar to that in a pen flashlight, radiates sufficientenergy in the 1-5 micron range for the detection ofmost hydrocarbons, carbon dioxide, and carbon mon-oxide. This simple and inexpensive light source of-fers long life and long-term stability.

Any source that can generate enough radiation atthe wavelength of interest for the purpose of detect-ing the specific target gas may be used. There are manylight sources available, ranging from specially designedheating filaments to electronically generated sources.

Modulating the light. Depending on the type ofdetector used, it may be necessary to modulate thelight source, turning it on and off at a specific fre-quency, in order for the detector and correspondingcircuitry to function properly. Typically, this is doneby passing the light through a chopper blade, whichresembles a fan blade. The blade interrupts the light,creating a pulsing frequency. The frequency is deter-mined by the speed of the motor and arrangement ofthe chopper blade.

The advantage of the chopper is that it is simpleand can provide a high chopping frequency, one whichwould otherwise be unattainable by a pulsing filamentsource. This is due to the fact that the filament is aheated wire and is limited to how fast it can be heatedup and cooled down. Although choppers do have ad-vantages, the large size of the chopper and motor as-sembly can be a liability for applications that requirea simple, rugged instrument for use in harsh locations.For this reason, pulsing filament sources are usedwhenever possible. The pulsating filament source typi-cally provides a wavelength in the range of 2-5 microns,depending on the filament temperature and bulb en-closure material. There are many different lightsources and modulating techniques available.

3. Optical filter: There are two basic types of gasanalyzers, namely dispersive and nondispersive. The dif-ference between the two is the way in which the spe-

65


cific wavelength of interest is extracted from the infra-red source.

Dispersive types utilize an optical device such as agrating or prism to spread the light spectrum over anarea containing the wavelength of interest.

Nondispersive types use discrete optical bandpassfilters, similar to sunglasses that are used for eyeprotection to filter out unwanted UV radiation. Thistype of configuration is commonly referred to asnondispersive infrared (NDIR). Almost all commercialIR instruments are of the nondispersive type. The dis-persive type instrument is typically used only for spe-cial requirements. The bandpass filter is one of themost important components in designing for the typeof target gas and selectivity of an analyzer. The filtersare generally produced by a specialty optical manu-facturer. Typical specifications for a methane filter areshown in Figure 7.

Percent transmittance specifies the ratio of transmit-

Fig. 7 Methane Filter Specifications

Peak Wavelength

Peak Transmission

Transmission Region

Bandwidth

Center Wavelength

Wavelength (µm)

Tran

smitt

ance

100%

80%

60%

40%

20%

0%3.0 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 4.0

66


ted radiation to the incident radiation. Bandwidth isdefined as the range of wavelengths that pass throughthe filter at half of the peak transmission point. Band-width determines the selectivity of the filter, and hencethe selectivity of the instrument. Center wavelength de-termines the gas that will be detected. There are cer-tain designs where filters of different wavelengths aremounted in a carousal arrangement. By exposing thegas mixture to different filters, different gases in thegas mixture can be identified.

4. Gas Cell/Light Path: Gas cells are often designedin such a way as to allow the light path to interact with

the sample gas. This is normally done by using atube that allows light to enter from one end andexit the other, where it meets the detector. Thereare “inlet” and “outlet” ports that allow the samplegas to circulate through the tube.

The path length, or the distance in which thelight passes through the gas, is called the length ofthe gas cell. This length is in direct proportion to

the amount of radiation absorbed; that is, the longerthe path length, the more radiation will be absorbed.A longer path length results in a greater signal usingthe same amount of radiation. Since the gas cellsare in direct contact with the gas sample, it is desir-able to construct them using chemically inert mate-rials to ensure the long-term stability and reliabilityof the instrument.

Configuration

There are a number of ways by which various IRcomponents can be arranged to produce a gas ana-lyzer. The design may be relatively simple, or verycomplicated, using many different optical componentsdepending on the type of analyzers for the applica-tions. For applications that require high sensitivity, se-lectivity, and stability, the design of the analyzer is morecomplex. For applications where the selectivity and

Sample Gas Entrance/Exit

IR Source Detector

SELF-SAMPLING GAS CELL

67


sensitivity can be compromised for reli-ability in a hostile, industrial environ-ment, a simpler design can be imple-mented.

Figures 8, 9, and 10 illustrate some ofthe basic features of an IR analyzer.

Figure 8 shows a basic layout of: (1)an IR source, (2) bandpass filter, and (3)the interaction with the gas sample anddetector. Depending on the detector used, thebandpass filter could be placed in front of the lightsource, instead of placing it in front of the detector.

Figure 9 shows a similarlayout except that two de-tectors are used. Modulatedflashing IR sources are re-flected back to the detec-tors. The active detectorhas a filter for the targetgas, while the reference de-tector has a filter with a different wavelength. In otherwords, the active detector is used to detect the targetgas and the reference detector is used to ignore thetarget gas. In actual operation, the reference detec-tor provides a base point value or zero point while theactive detector is used to provide the signal; with thedifferential between the two detectors providing theactual span value of the instrument.

This arrangement offers the advantage of compen-sating for the changes that occur in the detector’s sen-sitivity with time. For instance, the intensity of the lightsource can change with time due to contamination,which will create a zero drift. The two-detector arrange-ment minimizes this type of drift. Also, in this arrange-ment, the path length is doubled which leads to highersignal strength.

Figure 10 illustrates another popular design. Thisdesign uses two tubes or cells. One is a reference cell

Fig. 8 A Basic Infrared Gas Detector Layout

Sample Out

IR Source

Sample In

FilterDetector

Fig. 9 A Two-Detector Layout

ActiveFilter/Detector

MIR

ROR

ReferenceFilter/Detector

IR Source

68


that is filled with a pure target or reference gas, whilethe other is a sampling cell in which the sample gaspasses through. A chopper is used in this configura-tion, which is basically a disc with a number of slotsin it. As the chopper rotates, it alternately allows thelight beam to pass through the sample and referencecells. The single detector gets its base reading fromthe reference cell, similar to the reference detectorin Figure 9. The gas signal is acquired from thesample cell.

Characteristics

1. Temperature: An IR detector is essentially atemperature sensor and is, therefore, potentially verysensitive to changes in the ambient temperature. How-ever, a properly designed detector can be operatedbetween -40°C to 60°C without being susceptible toambient temperature fluctuations. Most detectors donot react well to sudden temperature variations. Theinstrument typically requires 10 to 20 minutes toachieve temperature equilibrium. For outdoor appli-cations, this usually is not a serious problem becausethe ambient temperature changes only slowly. Gener-ally, the detector unit is operated at a temperatureslightly above the surrounding temperature to preventcondensation. Water vapor condensation at the op-tics and/or on the detector can seriously affect the

Sample OutSample In Filter/Detector

SAMPLE CELL

REFERENCE CELL

MIR

RORM

IRRO

R

Chopper Motor

IR Source

Fig. 10 Double Beams with Chopper Layout

69


performance of the analyzer.

2. Humidity: Normal environmental humidity hasvery little effect. However, high humidity could promotecorrosion and contamination that lead to the failure ofthe analyzer. High humidity poses an even more seriousproblem in the presence of corrosive gases. The wavepath (sample chamber) can be constructed of any mate-rial that does not absorb the IR light. The most commonmaterials used are stainless steel, aluminum, or copper,plated with a corrosion-resistant coating. For some ex-tremely “wet” applications, such as a confined space, wetcontainers, or drainage sumps, the wet sample shouldbe “dried” before exposing it to the detector.

3. Sensitivity: The IR energy absorption is directlyproportional to the molecular structure of the hydrocar-bon (in addition to the concentration of the hydrocar-bon present). For example, the detector is least sensitiveto methane (CH4) with its simple, single bond. With pro-pane (C3H8) and butane (C4H10), however, the sensitivityincreases dramatically. An example of the drastic differ-ence between the sensitivity among varying hydrocarbonscan be seen by the fact that an IR detector can be cali-brated to 100% pure methane, but only a few percent byvolume of propane or butane will saturate the system. Inthe case of percent lower explosive limit (%LEL) or lowerflammable limit (LFL) applications, the volume concen-tration of each gas to reach 100% LFL or LEL varies,and the response curves are nonlinear; hence, each hy-drocarbon must have its own curve programmed intothe system. Unlike the catalytic sensor, which has nearlinear response to gases at the LFL range, the IR unitrequires some means to linearize the output.

4. Life Expectancy: The IR detector is a solid-statedevice that is sealed inside a standard electronic pack-age with a sapphire window filter. It has a long life ex-pectancy, similar to most electronic devices. The IR lightsource typically has a life expectancy on the order of 3 to

70


5 years. This lifetime can be dramatically prolonged byoperating the source at a lower-than-designed energy.Alternatively, the IR source can be designed so that itcan be easily replaced when needed.

Application

As shown in Figure 11, the IR detector responds toradiation by generating a constant signal, which is con-sidered the “zero” point for the source. Once the zeropoint is established and maintained, the span calibra-tion is automatically taken care of. This is due to thefact that the absorption of radiation by the gas is alwaysin the same proportion, regardless of its initial sourceintensity. Therefore, as long as the zero point is main-tained, the accuracy of the detector remains intact. Thisis one of the biggest advantages of IR technology. How-ever, a routine calibration check is an invaluable safetycheck and should not be eliminated from any periodicmaintenance.

For gas monitoring applications, the design of theunit should be relatively compact. Sampling should bedone by diffusion. The extractive type methods thatrequire a pump to move the sample into the detectorare more troublesome due to the limited electricalmotor life expectancy and the maintenance requiredon the pump.

The IR instruments used for this monitoring appli-cation are typically limited to the detection of higherconcentrations (1% and above) of hydrocarbons andcarbon monoxide. Carbon dioxide absorbs infrared ra-diation very strongly, and many monitors are availablethat can detect carbon dioxide in concentration rangesof 0.1% and higher.

With the filter used for the %LEL combustible gasesapplication, the center wavelength is typically at 3.4microns. This is the wavelength of most hydrocarbonsand it is also where most of the hydrocarbon derivativegases have a strong absorption.

Fig. 11 Advantage of an IRDetector. Zero point has thehighest output making it easy todetect any abnormal condition.

Gas Concentration

Span

Zero

0

Signal

71


Following is a list of common gases that are detected bythis detector:

1. Alkanes or saturated hydrocarbons such as methane,ethane, propane, butane, pentane, hexane, andheptane, etc.

2. Cycloalkanes such as cyclopropane, cyclohexane,methyl cyclohexane, etc.

3. Alkenes or unsaturated hydrocarbons such as ethylene,propylene, butene, pentene, hexene, octene, etc.Acetylene has absorption at 3.1 microns which is notdetectable.

4. Cycloalkenes such as cyclohexene and pinene.5. Aromatics such as benzene, toluene, and xylene.6. Alcohols such as methanol, ethanol, propanol, and

allyl alcohol.7. Amines such as dimethyl amine, trimethyl amine,

butanamine, cyclopropanamine, and pyridines.8. Ethers such as dimethyl ether, ethyl ether, n-propyl

ether, methylvinyl ether, vinyl ether, ethylene oxide,tetrahydrofuran, furan, and 1,4-dioxane.

9. Ketones such as acetone, methyl ethyl ketone,pentanone, methyl isobutyl ketone and heptanone.

10. Aldehydes that have a central wavelength mostly atthe 3.55 micron region and generally have a weakdetection signal at 3.4 microns.

Carbon dioxide at 4.3 microns and carbon monoxideat 4.6 microns have very little interference by other gases.

Summary

Detection of hydrocarbons using IR has been availablefor a few years. However, because of the many design op-tions, the features and implementation of the technologyvary greatly from one manufacturer to another. Neverthe-less, IR detection has been well received by many indus-tries, including the petrochemical industry.

For the purpose of area air quality and safety applica-

72


tions, some comparisons between IR, solid-state, and cata-lytic bead sensors, used for the detection of combustiblegases are given below.

1. Poisoning: This is the main problem with catalyticsensors. Various chemical compounds, such as hydrogensulfide, silicon compounds, and chlorinated or fluorinatedcompounds, among others, can poison the catalyst in thesensor and cause the sensor to lose sensitivity. IR detec-tors do not suffer from this problem.

2. Burn Out: Catalytic senors will burn out if exposedto high gas concentrations. Again, IR detectors do nothave this problem.

3. Life Expectancy: Catalytic sensors have a life expect-ancy of about 1-2 years, while solid-state sensors typicallylast more than 10 years. A well-designed IR unit also has alife expectancy of more than 10 years.

4. Calibration: Periodic calibration must be done onall sensor types. However, on IR units, as long as the zerois maintained, the IR unit is assured a good response andgood span accuracy. Because of this characteristic, ab-normal functioning of an IR unit can be easily determined.

5. Continuous Exposure to Gas: In applications that re-quire the detector to be constantly exposed to a gas streamto monitor for hydrocarbons, catalytic and solid-state sen-sors will experience a shortened life span. Continuous ex-posure to gas ultimately changes the characteristic of thesensor and leads to permanent damage. However, withIR instruments, the functional components are protectedby the optical parts, which are basically inert to mostchemicals. Only the IR radiation interacts with the gas;therefore, as long as gas samples are dry and noncorro-sive, IR instruments can be used to monitor a gas streamcontinuously over a long period of time.

By selecting one of the three available sensor tech-nologies (IR, solid-state, and catalytic), one can tailor asolution to almost any hydrocarbon detection application.

73

Chapter 6 Photoionization Detectors

Chapter 6

PhotoionizationDetectors

T he photoionization detector (PID) utilizes ultra-violet light to ionize gas molecules, and is com-monly employed in the detection of volatile or-

ganic compounds (VOCs). This technique originallyfound use in bench top laboratory instruments, butits complexity limited its use elsewhere. The heart ofthe photoionization detector is an ultraviolet source,which is essentially a lamp. Early versions of this lampused electrodes inside the lamp similar to those usedin the early days of the vacuum tube and were quitecostly to manufacture.

The lamps used today do not contain electrodesand are both less costly and have a longer life expect-ancy. In the 1980s, with the advent of integrated cir-cuit (IC) technology, the electronics that were de-signed were better able to process the small signal fromPID sensors into useful and reliable data.

Additionally, requirements for the monitoringof underground storage tanks to help preventground water contamination began to emerge,which required monitoring of VOCs.

These events led to the design of small, portablePIDs (see Figure 1) that have proven to be both prac-tical and reliable, and which offer fast response andthe ability to detect low gas concentrations. To this

Fig. 1 A Pocket PID Monitor

74


day, PID sensors are the preferred choice for the de-tection of VOCs.


Ultraviolet (UV) light is the radiation group thathas frequencies that are higher and are directly abovevisible light in the electromagnetic radiation spectrum.UV wavelengths, being in the nanometer (ηm = 10-9

meters) region, are much shorter than IR wavelengths.The wavelengths of infrared light used for gas analy-sis are in the micron region, 10- 6 meters.

Because shorter wavelengths have higher frequen-cies, and hence more energy, UV radiation has moreenergy than IR radiation. UV radiation energy levelsare commonly described in terms of electron volts, oreV.1 Wavelength is related to eV through Planck’s Con-stant2 which is 4.135 x 10-5 eV. [An electron volt equals1.2395 x 10-6/wavelength (ηm). The term eV is usedfor convenience, to give a simple numerical expres-sion of the radiant strength.]

The PID Lamp. The heart of the detector is thePID lamp which has an exterior shape resembling amedical ampule and comes in different sizes and di-mensions, depending on the manufacturer. The lampis filled with a low-pressure inert gas. When this gas isenergized with energy in resonance with the naturalfrequency of the gas molecules, an ultraviolet spec-tral radiation is produced.

The wavelength of the UV light emitted dependson the type of gas in the lamp. For instance, krypton,when excited, will emit 123.9 ηm and 116.9 ηm radia-tion, or the equivalent of 10 eV and 10.6 eV. This 10.6eV lamp happens to be the most popular lamp cur-rently used in PID instruments. Various other gasesare also used. Figure 2 on the following page showswavelengths emitted by argon, krypton, and xenon.

Many different techniques are employed to powerthe PID lamp. A high voltage is required for the lamp

1 eV = energy gained when a particle ofone electronic charge is acceleratedthrough a potential of one volt.

2 Planck’s Constant is a fundamental con-stant in quantum physics. The radiantenergy only changes in “quanta” ordiscrete steps.

75


to function. The source needs to remain stable overtime and also be efficient, so as not to consume toomuch power. A typical design would place electrodesinside the lamp. These types of lamps look similar tothe vacuum tubes used in the early days of radio andTV. There are electrical terminations on the tube, andthe gas discharge is confined to a small capillary. Whilethe voltage used is in excess of 1000 volts, the currentis very low. Although this type of lamp is used in manylaboratories, it is quite expensive. As a result, a morepopular lamp in use today is the electrodeless lamp.

The Electrodeless Lamp. This lamp is only filledwith a low-pressure inert gas and there are no elec-trodes inside the lamp. The simplicity of this designallows the lamp to be miniaturized, making the de-sign of the portable instrumentation very compact.

In the electrodeless lamp, the low-pressure inertgas is separated by the lamp wall from the outsideworld. Because there are no electrodes that providepower directly to the gas, the only way to excite the

Fig. 2 The wavelengths (ηm) emitted by argon, krypton, and xenon.

XENON

147.6ηm

129.1ηm

8.4eV 9.6eV

KRYPTON

123.9ηm

116.9ηm

10.0eV 10.6eV

ARGON

105.9ηm

11.7eV

76


gas molecules is to use radiation energy that can pen-etrate the lamp wall.

There are several different ways this can be accom-plished. One method is to use electromagnetic radia-tion. A more popular technique, however, is to placea pair of electrodes on the exterior wall of the lamp. Ahigh-voltage, low-current charge is applied to the elec-trodes, and the energy applied is sufficient to excitethe low-pressure gas inside the lamp. The design issimilar to that of a fluorescent light used in a home oroffice, except on a much smaller scale.

Special Window Materials. At the wavelength ofradiation, meaning the wave frequency of UV light,most materials will absorb the radiation and preventthe transmission of such radiation.

Therefore, special window materials are placed atthe discharge end of the lamp which allow spectralemissions to pass through. This window material is acrystal that allows good transmission of the targetedUV wavelength. For instance, 10.6 eV lamps commonlyuse krypton gas and magnesium fluoride windows,while 11.7 eV lamps, which have more energy, use ar-gon gas and lithium fluoride windows. These windowsare soft glass and are very fragile. They are quite ex-pensive, and require special care and handling. Fig-ure 3 illustrates typical PID design.

Gas Molecules

Signal i

Electrode

Electrode Platefor Lamp Illumination

Lamp

+V-

Insulation

Window

Fig. 3 A Typical Photoionization Detector Configuration

77


A pair of electrode plates are placed in close prox-imity to the lamp window where the light is emitted.The electrodes are biased with a stable DC voltagewhich will generate signals in case any small changesin the electrical field occur. As gas molecules moveinto the radiated field in the space between the elec-trodes, they are ionized and the free electrons arecollected at the electrodes resulting in a current flowwhose magnitude is directly proportional to the gasconcentration.

Distinct Ionization Potentials. Each gas has itsown unique ionization potential (IP). Gases with IP val-ues below the eV output of the lamp will be detected.For most portable instruments, the 10.6 eV lamp ismost widely used because it detects most volatile or-ganic compounds, and the lamp is easy to clean. How-ever, other lamps are also available, including 10 eV,9.5 eV, 8.4 eV, and 11.7 eV lamps. Most manufactur-ers design instruments with a lamp chamber that al-lows easy interchange of lamps. The ionization po-tentials for many gases are included in Appendix II atthe end of this book.

Following are examples of gases detectable usingdifferent lamps:

1. Gases detected by 9.5 eV lamp such as benzene,aromatic compounds, amines.

2. Gases detected between 9.5 eV and 10.6 eV such asammonia, ethanol, acetone.

3. Gases detected between 10.6 eV and 11.7 eV such asacetylene, formaldehyde, methanol.

A 9.5 eV lamp will ionize gases with ionizationpotentials below 9.5 eV and will not ionize gases withhigher ionization potentials. On the other hand, ifone uses an 11.7 eV lamp, gases with ionization po-tentials up to 11.7 eV will be ionized. Interchanginglamps in an unknown air sample helps to separate itinto groups of gases.

78


Characteristics

PID sensors offer fast response for detection ofmany volatile organic compounds. They will respondto all gases that have ionization potentials equal to orless than the eV output of their lamps.

Correction Factor. PIDs are typically calibratedwith isobutylene, a stable gas with a slightly pungentodor. This gas is easy to handle and can be stored athigh pressure, allowing calibration bottles to providemany calibrations. PID instruments typically detectgases at low concentrations, and most of these gasesare normally liquid solvents or other gases that arenot easy to calibrate. Thus, it is much easier to cali-brate these instruments using isobutylene as the cali-bration gas. Readings for other gases are obtained bymultiplying the reading by a correction factor. For ex-ample, benzene has a correction factor of 0.5, whichmeans that a reading of 100 ppm isobutylene on thesensor indicates a concentration of 50 ppm benzene.For ammonia, which has a response factor of 10, areading of 100 ppm on the sensor would indicate anammonia concentration of 1000 ppm. Instrumentmanufacturers typically supply a list of correction fac-tors with their product.

An important point to remember is that correc-tion factors are not absolute, and data can be slightlydifferent from manufacturer to manufacturer. In fact,these factors can even vary somewhat from lamp tolamp, and the results may vary depending on the qual-ity of the UV output of the lamp at the time of mea-surement. Thus, for the most accurate readings onspecific gases, it is necessary to individually calibratethe gas of interest.

PID instruments are very sensitive to the composi-tion of gases in the sample. The results will differ vastlyif one uses gas with dry nitrogen as the balance tocalibrate, from those results obtained when one usesair as the balance. Therefore, it is important to use

79


isobutylene in air as the calibration gas when the instru-ment is used for area air quality and safety applications.Still, there will be some error because the calibrationmixture is under high pressure and is dry; therefore, itdoes not exactly duplicate environmental air.

Zeroing Instruments for Best Results. Generally,high humidity decreases the response by up to 30%,when compared to dry air. In dry nitrogen, readingsare typically 10-30% higher than in dry air.

If the sample gas constituents include gas compo-nents that are ionizable but the ionization potentialsare above the detecting lamp output potential, and al-though the lamp radiant energy will not ionize thesegas molecules, the UV rays can be scattered and ab-sorbed, resulting in a lower output reading. This isknown as the “quenching effect.” For example, in envi-ronmental air, such “quenching” gases are water vapor,carbon dioxide, methane, carbon monoxide, etc. Thisis another reason why a representative air is needed tozero the sensor and prepare the calibration gas mix-ture. The common practice of calibrating the instru-ments with isobutylene in an air mixture, without mak-ing corrections, results in an inaccurate reading.

PID sensors will exhibit different zero readingswhen exposed to nitrogen, dry, clean air, and environ-mentally clean air. Thus, it is best to zero the sensor in thesame type of conditions that are present in the application inwhich it will be used. Clean, ambient air is, therefore,the best choice for zero airas well as for mixing cali-bration gas.

PID Response Charac-teristics. The output of thePID sensor is relatively lin-ear below 200 ppm, and thedetector output becomessaturated above 2000 ppm.Figures 4, 5, and 6 illustrate

Concentration Isobutylene0 50 100 150 200

0

1

2

3

4

5

Outp

ut V

olta

ge

Fig. 4 Zero-to-200 ppm Curve

80


the response curves for 200ppm, 2000 ppm, and 4000ppm isobutylene in roomair.

You will notice that the200 ppm curve is relativelylinear, while on the 2000ppm curve, 1000 ppm pro-duces an output of 3.6 voltsand 2000 ppm produces an

output of 4.8 volts. The additional 1000ppm produces only 1.2 volts, or one thirdof the initial 1000 ppm. From the 4000 ppmcurve, you can see that the detector be-comes saturated.

Thus, this output region does not pro-vide a reading with a good resolution.

Applications

PID instruments offer very fast re-sponse, high accuracy, and good sensitivityfor detection of low ppm volatile organiccompounds (VOCs). The major shortcom-

ing of these instruments is that the PID lamp requiresfrequent cleaning. Because the lamp window is directlyexposed to the sample stream, the condition of thewindow is very critical to accurate readings, and a dirtywindow will produce much different results than aclean window. The frequency of cleaning needed forthe window depends on the sample stream conditions.

Because PID sensors require periodic cleaning andhave limited life expectancies, they are not practicalchoices for use in stationary monitors, which samplecontinuously. Their use is limited to portable modelsin which only periodic readings are required.

Lamp Life Expectancy. The life expectancy of thelamp depends on the type of lamp. Generally, 10.6 eVlamps have the longest life expectancy, approximately

Isobutylene Concentration (PPM)0 1000 2000 3000 4000

0

100

200

300

400

500

Outp

ut V

olta

ge (m

V)


Concentration Isobutylene0 500 1000 1500 2000

0

1

2

3

4

5

Outp

ut V

olta

ge


81


6000 hours.The 11.7 eV lamps have lithium fluoride win-

dows that transmit the shorter UV wavelength lightat 11.7 eV, which allows for the detection of manymore gases when compared to lower eV lamps.However, the intensity of the light they emit isweaker than the light emitted from lower eVlamps. The weaker energy emitted means less sig-nal, which results in less stable and more tempera-ture-sensitive instruments. Thus, 11.7 eV lamps arenot typically used for general applications. Further-more, lithium fluoride is hygroscopic, which meansthat lithium fluoride crystals attract and absorbmoisture from the air, causing the window to de-grade. They also cannot be exposed to cleaning sol-vents because most solvents contain small amountsof water. Therefore, a special, fine solid aluminumoxide powder is needed to clean the lamp.

Typical Specificationsfor PID Instruments

Detector: 10.6 eV electrodeless dischargelamp. 11.7 eV lamp optional.

Response Time: 3 secs. to 90% reading.

Ranges: 100 ppm, 1000 ppm, and 2000ppm isobutylene equivalent, auto-ranging.

Sensitivity: 0.1 ppm isobutylene.

Accuracy: 1-10%, depending on range.

Temperature Range: –20oC to +50oC.

Humidity: 0–95% relative humidity,noncondensing.


Electrochemical

Sensor Selection Overview

Monitors Analyzers

Application Requirement

• Safety Monitoring (LEL) • Exposure Assessment (TWA)

• Toxic Limit Detection (PEL) • Ambient Air Quality Compliance Monitoring

• Leak Detection

• Personal Safety

LIMIT DETECTION ALARM QUALITATIVE & QUANTITATIVE ANALYSIS

ToxicGases

Solid-State

Catalytic

CombustibleGases

Infrared

Solid-State

VolatileOrganic

CompoundsPhotoionization

Monitor Type:

TotalHydro-

CarbonsFlame Ionization

ThermalConductivityOther

GasesColorimetric

Analyzer Type:

UV Photometers

UV Fluorescence

Target Gas:

Multi-Gas

(Laboratory)

AirQualityMonitor

Chemiluminescence

Flame Ionization

Infrared

Paramagnetic;Zirconium Oxide;Electrochemical

UV/IR Spectrometer

Fourier Transform IR

Mass Spectrometer

Gas Chromatograph

SingleGas

(Online)

O3

SO2

NOx

THC*

CO

CO2

SO2

O2

ContinuousEmissionMonitor

Most UV/IRAbsorbing Gases

Most IRAbsorbing Gases

Most Gases

*Total hydrocarbon

103

Chapter 8 Sensor Selection Guide

Chapter 8

Sensor Selection Guide

Each of the following sensors—electrochemical,catalytic bead, solid state, infrared and photo-ionization detectors—must meet certain criteria

to be practical for use in area air quality and safetyapplications. Some of the basic requirements are:

1. The sensor should be designed for a housingthat is small and rugged. The sensor should be suit-able for use in hazardous locations and harsh envi-ronments, and, it should also be explosion-proof. Thesensor should be cost-effective, designed for installa-tion and use in industrial production areas, andinstallable at a reasonable cost.

2. For portable applications, the instrumentsshould have reasonable energy consumption and the op-tion of powering the instruments with batteries shouldbe easily available. The instruments should be smalland portable so they can be carried easily. They shouldbe safe for use in industrial environments. Preferably,the instruments should be certified as intrinsically safefor use in a hazardous area.

3. The operation and maintenance of the instru-ments should be easily performed by regular plantpersonnel with minimal special training require-ments.

4. In stationary installations, the sensors shouldbe able to function continuously and reliably for a periodof time, preferably longer than 30 days. The sensor

104


should be able to function in an industrial environ-ment for at least two years or longer and should bereplaceable or renewable at a reasonable cost. Itshould be easy to install into a multi-point system andbe managed by a controller or a computer controlleddistribution system.

5. The cost of the instruments should be reason-able so that multiple sensors can be installed to effec-tively protect the area.

Four of the five sensors discussed in this book allmeet the above criteria. The exception is the photo-ionization detector. The PID is a good detector forportable applications but is limited by the lamp be-cause it has a relatively short life expectancy and thefrequency of maintenance required may not be prac-tical for stationary applications. However, there arePID stationary instruments available that can be use-ful as long as users are aware of the limitation.

There are other types of sensors which meet theabove criteria, but most have limitations. For ex-ample, thermal conductivity sensors are mostly usedfor high concentration applications and are not widelyused as gas monitors.

Factors to Consider When Selecting Sensors

One of the most frequently asked questions regard-ing sensors is: “Which sensor is the best?” Of course,there is no simple answer to this question. Each sen-sor has certain capabilities and limitations, and thusthe suitability of a given sensor depends largely onthe application in which it is to be used. Thus, tochoose the correct sensor, one must first properly de-fine the application. The illustration on page 102shows an overview of various application requirementsand their detection technologies. It is common formanufacturers to exaggerate the capabilities of thesensors that they offer and downplay sensors that theydo not offer.

105


In determining which sensor to use for a given ap-plication, the following factors/observations shouldbe considered:

A. Realistically define what objective one is trying to ac-complish and define an instrument specificationthat meets the minimum requirements. Thespecifications should define the gases andranges of the sensors. The ranges or the con-centration of the gases to be measured shouldbe 3 to 5 times the actual monitoring concen-tration. As with a voltmeter, one should alwaysselect a range higher than the actual voltageto be measured. For example, select a 50-voltrange to measure a 12-volt battery.

B. Determine the background gases in the moni-toring area. In cases where the backgroundgases cannot be determined, a representativesample should be analyzed. A major cause ofsensor failure is the presence of backgroundgases that the instrument’s manufacturers didnot take into consideration. The selectivity orspecificity of the sensor must be acceptable forthe application.

C. The temperature ranges in which the sensor isto be installed should be within the sensorspecifications and should be suitable for thegases to be monitored. For example, jet en-gine fuels have very low vapor pressure. It isuseless to install a sensor to measure the com-bustible range in a hangar if the temperaturewill never exceed 100° F because the vaporconcentration cannot reach combustible lev-els. In this example, it is more appropriate tomeasure in ppm ranges.

The temperature changes between day andnight, and during summer and winter, shouldalso be considered. A wide temperature change

106


can cause moisture condensation. This is par-ticularly important in a confined space such asa closed container where air circulation is poor.

D. A typical specification for humidity is 95% non-condensating. The occurrence of condensationis a function of temperature change, as seenon wet windows and car windshields in themorning. Normally, there is no problem innormal industrial background environments,even during the hot summer months in coastalareas such as the Gulf of Mexico, as long as theair circulation is normal. Areas with poor aircirculation can often be the cause of conden-sation.

Both solid-state and catalytic sensors haveheated elements. In addition, their transmit-ters are designed to operate at 14-24 VDC,which generates heat. Therefore, the sensortransmitters are always a few degrees warmerthan the environment in order to minimizethe possibility of condensation. Electrochemi-cal sensors normally require relatively muchless power; therefore, the temperature of theirtransmitters is similar to the surrounding tem-perature. In this case, it is easier for conden-sation to occur.

E. In applications requiring the sensors to be con-stantly exposed to gas, special considerations arerequired and the sensor specifications and thesuppliers of the sensors may need to be con-sulted. A properly designed sampling systemmay make difficult or otherwise impossible ap-plications possible to handle.

It is difficult to mention every considerationneeded, but a carefully evaluated and studied appli-cation can yield savings in both time and money.

107


With the specifications of the sensors, one candecide which sensor best meets one’s requirements.There is no general consensus that establishes whichsensor is the best for a given application. Hence, theinformation that follows contains some guidelinesthat may be helpful in making the proper selectionof a sensor.

Toxic versus Combustible Gas Monitoring

Gas monitoring applications are generally classi-fied into toxic or combustible range monitoring. Toxicgas monitors are generally used for human healthprotection and the ranges of the monitors are 3 to 5times higher than the permissible exposure limits. Formost gases, ranges are in ppm concentrations.

For combustible applications, ranges are typically100% lower flammable limits or a fraction of theseranges, such as 50% LFL. The gas concentrations arehigh and are generally in the range of several percent.

In other words, for toxic gas applications, a sensormust be able to measure gases at low concentrationswhile, for combustible gas monitoring, a sensor mustmeasure high gas concentrations.

Summary

Electrochemical Sensors. Except for oxygen appli-cations, electrochemical cell sensors are designed tobe used as toxic gas monitors. These sensors are onlysuitable for low concentration ppm ranges. For por-table applications, the electrochemical sensor hasmany advantages: it has very low power consumption,responds quickly to gas, and is not affected by humid-ity. Also, the sensors are only exposed to gas periodi-cally, which maximizes the sensor life.

Electrochemical sensors are therefore a goodchoice for portable instruments. Electrochemical sen-sor life expectancy is two years; however, dependingon the application, it can be much shorter. The cost

108


of replacement sensors is high, especially when thenumber of instruments in use is large. The annualbudget and labor to keep the instruments function-ing need to be considered.

There are approximately 20 gases that can bemonitored by electrochemical sensors. For the restof the gases (for ppm ranges), solid-state sensors orPIDs will need to be used.

Catalytic Sensors. In portable combustible gasesin the LEL ranges, a catalytic sensor is good for nor-mal, simple applications. The sensors can last for along time because they are used sporadically in por-table applications. Catalytic sensors are relatively in-expensive but one has to make sure they are made byreputable suppliers.

Infrared and Solid-State Sensors. For gases thatcan poison the catalytic sensors and make them im-possible to use, the choice is between infrared sensorsand solid-state sensors. Depending on the gases to bedetected, infrared sensors have the better perfor-mance but detection of gases is limited. On the otherhand, solid-state sensors can detect most chemicals inthe LEL ranges.

In stationary applications, the sensors are con-stantly exposed to environmental background gases.For toxic gas applications, it is generally favorable touse solid-state sensors, especially when the number ofsensors is sizable. In applications where interferencecan be a problem, it is best to study the sensor specifi-cations and to consult with the supplier.

For stationary combustible gas applications, thechoice is among catalytic sensors, solid-states sensorsand infrared sensors, which are fully described inChapters 4, 5, and 6 respectively.

There are no standardized specifications for a gasmonitor. The following table entitled “Typical Speci-fications for Gas Monitoring Instruments” is a sum-mary of several “requests for quotation” from cus-

109


Typical Specifications for Gas Monitoring Instruments

Remote Sensor Transmitter SpecificationsGas to Be Monitored: Methyl Bromide, 0–500 ppm.Sensor: Solid-state sensorTemperature: –10o C to 45o C.Humidity: 95% RH, noncondensing.Accuracy: 5%Response Time: 50 seconds to 90% full scaleBackground: Normal environmental conditionsTransmitter shall be certified for use in Class 1, Div. 1, Groups B, C, D areas. Provides

functional adjustments without the need to open the transmitter’s cover, and pro-vides protection for outdoor applications.

Power: 14 VDC to 24 VDCOutput: 4–20 mA linear output.

Controller Specifications1. Four-channel control unit, microprocessor-based.2. Independent digital display of reading for each channel. Provides calibration func-

tions, three alarm settings, and diagnosis functions.3. Three-level alarm set points (low, mid, high) with LED indicator for each alarm.4. Relays: SPDT, 5 amp. 220 VAC max. resistive

Common relays for each alarm level (3 relays per channel)Options: Individual relays for each alarm level ( 3 relays per channel)

“Fault” relays for each channel, indicating electrical fault of the sensor5. 4-20 mA analog output signal for each channel.6. Wall-mounted weatherproof NEMA 4X enclosure.7. Provide 24 VDC, 500 mA for each sensor module, accepts and processes 4-20 mA

transmitter output signal.

tomers and is presented here as a reference.

161

Chapter 11 Gas Sensor Calibration

Chapter 11

Gas Sensor Calibration

G as sensors need to be calibrated and periodi-cally checked to ensure sensor accuracy andsystem integrity. It is important to install station-

ary sensors in locations where the calibration can beperformed easily. The intervals between calibrationcan be different from sensor to sensor. Generally, themanufacturer of the sensor will recommend a timeinterval between calibration. However, it is good gen-eral practice to check the sensor more closely duringthe first 30 days after installation. During this period,it is possible to observe how well the sensor is adapt-ing to its new environment.

Also, factors that were not accounted for in thedesign of the system might surface and can affect thesensor’s performance. If the sensor functions prop-erly for 30 continuous days, this provides a good de-gree of confidence about the installation. Any possibleproblems can be identified and corrected during thistime. Experience indicates that a sensor surviving 30days after the initial installation will have a goodchance of performing its function for the durationexpected. Most problems—such as an inappropriatesensor location, interference from other gases, or theloss of sensitivity—will surface during this time.

During the first 30 days, the sensor should bechecked weekly. Afterward, a maintenance schedule,

162


including calibration intervals, should be established.Normally, a monthly calibration is adequate to ensurethe effectiveness and sensibility of each sensor; thismonthly check will also afford you the opportunity tomaintain the system’s accuracy.

The method and procedure for calibrating thesensors should be established immediately. The cali-bration procedure should be simple, straightforward,and easily executed by regular personnel. Calibra-tion here is simply a safety check, unlike laboratoryanalyzers that require a high degree of accuracy. Forarea air quality and safety gas monitors, the require-ments need to be simple, repeatable, and economi-cal. The procedure should be consistent and trace-able. The calibration will be performed in the fieldwhere sensors are installed so it can occur in any typeenvironment.

Calibration of the gas sensor involves two steps.First the “zero” must be set and then the “span” mustbe calibrated.

Step One: Setting the “Zero” Reading

There is no established standard that defines zeroair. Many analytical procedures, including some spe-cific analyzer procedures such as EPA methods, usepure nitrogen or pure synthetic air to establish thezero point. The reason for this is that bottled nitro-gen and pure synthetic air are readily available. As aresult, it is popularly believed that using bottled nitro-gen or synthetic air is a good method to zero a sensor.

Unfortunately, this is not correct. Normal air con-tains traces of different gases besides nitrogen andoxygen. Also, ambient air normally contains a smallpercentage of water vapor. Therefore, it is much morerealistic and practical to zero the sensor using theair surrounding the sensor when the area is consid-ered to be clean. This reference point can be diffi-cult to establish. Therefore, a good reference point

163


can be in the area where air is always consideredclean, such as in an office area. This will give a morerealistic representation of the zero point because itwill be representative of the local ambient air condi-tion. The lack of water vapor can cause the zero pointsetting to read lower than in ambient air making thesensor zero appear to drift. This is most noticeablein solid-state sensors and PIDs.

Calibration Methods. Taking all factors such as thetype of sensor and the conditions of the applicationinto consideration, the following are some proposedmethods of calibration:

A. In applications where the ambient air is nor-mally clean, and, based on the operator’s judgmentthat no abnormal condition exists and the instrumentis indicating a close to zero reading, the procedure tozero the sensor can be skipped. When in doubt, use aplastic bag to get a sample of what is considered to be“clean air” in the facility and expose it to the sensorfor a few minutes. This is a very quick and easy proce-dure. It is also a very effective way to differentiate areal alarm from a false alarm.

B. Compressed air has the advantage that it is easyto regulate and can be carried around in a bottle. Also,in many facilities, shop air is available throughout theplant, making it very accessible and convenient. How-ever, most shop air contains small concentrations ofhydrocarbons, carbon monoxide, carbon dioxide, andpossibly other interference gases. Also, the air is typi-cally very low in humidity. A solution to this is that theair can be filtered through activated charcoal to re-move most of the unwanted gases and water vapor canbe added into the air using a humidifier in the sam-pling system. After this conditioning, the air can beused to calibrate most types of sensors. However, it isimportant to note that carbon monoxide is not re-moved by charcoal filters.

It is therefore imperative to make sure that the

N2

O2

H2O

CO2 & others

Ambient air is the best zero air.

164


CO concentration in the shop air is the same as in theambient air. Furthermore, a soda ash filter should beused to remove carbon dioxide. This is also a very goodway to zero carbon dioxide sensors since placing a sodaash filter in-line with the sampling system will removeall carbon dioxide, thus providing an easily obtain-able zero baseline.

Although synthetic air is usually very pure, it can-not be used with solid-state sensors or PID sensorsbecause these sensors require some water vapor in thesample stream. A simple solution to this problem is toadd a wet tissue paper in the sample line. This acts asa humidifier in the sample stream and providesenough water vapor for the sensor to read properly.Another option is to use a Nafion tube, which is de-scribed thoroughly in Chapter 10, “Sampling Systemsand Designs.” Figure 1 illustrates this concept.

Fig. 1 Adding Moisture to Calibration Gases

Step Two: Span Calibration

The span calibration can be quite easy or it can bevery complicated and expensive, depending on the gastype and concentration range. In principle, to achievethe best accuracy, a mixture of the target gas balanced in thebackground environmental air is the best calibration gas.However, although this can be done, it usually requiresthat the operators be more skilled than would usuallybe required. In practice, most calibration gases are pur-chased from commercial suppliers. The following sec-tion describes a few methods of span calibration.

A. Premixed Calibration Gas

Calibration Gas

Regulator

NafionDryer Tube or

Wetting MaterialTo SensorFlow

Meter

165


This is the preferred and most popular way to cali-brate gas sensors. Premixed gas mixtures are com-pressed and stored under pressure in a gas bottle. Thebottles are available in many sizes but most field cali-brators employ smaller, lightweight bottles. Thesesmall portable bottles come in two different catego-ries: a low-pressure and a high-pressure version.

The low-pressure bottles are thin-walled, light-weight bottles that are usually nonreturnable and dis-posable. High-pressure bottles are designed to bottlepure hazardous chemicals. For calibration gases, thesebottles are normally made of thick-walled aluminumwhich has a service pressure of 2000 psi.

To get this highly pressurized gas out of the bottlein order to calibrate the sensor, a regulator assemblyis needed. This assembly consists of a pressure regu-lator, a pressure gauge, and an orifice flow restrictor.The orifice flow restrictor is a fitting with a hairlinehole that allows a constant air flow at a given pres-sure difference. In operation, the high pressure fromthe bottle is reduced to a lower pressure of only a fewpsi, which provides a constant air flow through theorifice. Flow rates between 600-1000 cc/min are most common. Models can befitted with an adjustable pressure regu-lator so that the flow rate can be adjustedaccordingly. Figure 2 illustrates typicalmodels of high- and low-pressure bottleassembly.

Many gases can be premixed withair and stored under pressure, but somegases can only be mixed in inert gasbackgrounds, such as nitrogen. Somemixtures can only be stored in bottlesthat are specially treated or condi-tioned. Each type of mixture will havea different amount of time before it ex-pires or before it can no longer be used.

Low-Pressure Assembly High-Pressure Assembly

Fig. 2 Calibration Gas Bottles

166


Detailed information about storage and shelf life canbe obtained from the manufacturer. Generally, highvapor pressure gases with low reactivity, such as meth-ane, carbon monoxide and carbon dioxide, can bemixed with air and stored under high pressure. Lowvapor pressure gases, such as liquid hydrocarbon sol-vents, can only be mixed with air and stored underlow pressure. Most highly reactive chemicals are mixedwith a nitrogen background. With certain sensors, suchas solid-state sensors, whether the mixture of the gasis in the air or in the nitrogen background will dra-matically affect the sensor reading.

During calibration, some sensors may need mois-ture to get a proper reading. Moisture can be added byfollowing the same procedure described in Step 1 forzeroing the sensor.

To estimate the volume of a pressurized gas in acylinder, take the total pressure (P) divided by the at-mospheric pressure (Pa) and multiply this ratio by thevolume of the cylinder:

Vmix = V . (P/Pa)

where

Vmix = the volume of the gas mixtureV = the volume of the cylinderP = the pressure in the cylinderPa = the atmospheric pressure

For example, let’s say a given lecture bottle has a 440cc volume (V). Assume the bottle has a 1200 psi pres-sure. The estimated volume of the premixed gas at atmo-spheric pressure is: (440 cc) x (1200/14.7) = 35,918 cc.

If the flow rate of the calibration gas is 1000 cc perminute and it takes approximately one minute per sen-sor to calibrate, a single cylinder can be used to cali-brate approximately 30 times.

167


B. Permeation DevicesA permeation device is a sealed container that con-

tains chemicals in liquid and vapor phase equilib-rium. The gas molecules are either permeatedthrough the permeable container wall or throughthe end cap. The rate in which the gas moleculespermeate depends on the permeability of the mate-rial and temperature. The rate of permeation is con-stant over long periods of time. At a known rate ofpermeation at a given temperature, a constant flowrate of air mixed with the permeated chemicals formsa constant stream of calibration gas. A calibrator withconstant temperature and flow regulation is needed.However, the permeation tube continuously emitschemicals at a constant rate thus creating a storageand safety problem. Also, the rate of permeation fora given gas of interest can be too high or too low fora given application. For example, high vapor pres-sure gases permeate too quickly while very low vaporpressure chemicals have a permeation rate that is toolow to be of any use.

Permeation devices find most of their use in labo-ratories and in applications using analytical analyzers.For gas monitoring applications, the concentrationsneeded to calibrate the sensor are typically too highfor the permeation device. Therefore, they have beenfound to be of limited use.

C. Cross CalibrationCross calibration takes advantage of the fact that

every sensor is subject to interference by other gases.For example, for a sensor calibrated to 100% LEL hex-ane, it is usually much easier to use 50% LEL meth-ane gas to calibrate the sensor instead of using an ac-tual hexane mixture. This is because hexane is a liq-uid at room temperature and it has a low vapor pres-sure. Therefore, it is more difficult to make an accu-rate mixture and to keep it under high pressure.

Examples of Permeation Tubes

168


On the other hand, methane has a very high va-por pressure and is very stable. Furthermore, it canbe mixed with air and still be kept under high pres-sure. It can be used for many more calibrations than ahexane mixture in the same size bottle and it has along shelf life. A 50% LEL methane mixture is alsoreadily available. Therefore, it is common practice formanufacturers of combustible gas instruments to rec-ommend the use of methane as a substitute to cali-brate for other gases.

There are two ways to accomplished this task. Thefirst method is to calibrate the instrument to meth-ane while other gas readings are obtained by multi-plying the methane reading by response factors thatare included in the manual. This is commonly donewith catalytic sensors. Catalytic sensors have a linearoutput and therefore the use of this response factor isapplicable to the full-scale range. For example, pen-tane has an output of only half that of methane gaswhen the sensor is calibrated to methane. Therefore,it has a response factor of 0.5. So, if the instrument iscalibrated to methane but is used to measure pentane,the reading is multiplied by 0.5 to obtain the pentanereading.

The second method is to still use methane as thecalibration gas, but double the value of the reading ofthe calibration. For instance, use 50% LEL methanecalibration gas and calibrate with this as 100% LELpentane. After the calibration, the instrument directlyindicates the pentane gas concentration although itwas calibrated using methane gas.

Many low-range toxic gas sensors can be calibratedusing cross gas calibration. Also, with infrared instru-ments, any gas within the same wavelength of absorp-tion can be used for cross calibration. The advantageof cross calibration is that it allows the sensor to becalibrated with a gas and range that is easier to obtainand handle.

169


However, there are some problems with using crosscalibration. One is that the response factors for eachsensor can be different as it is generally impossible tomake most sensors exactly alike. For example, in cata-lytic sensors, the heater voltage has to be as specifiedin the manual; otherwise, the response factor will notbe applicable. The response characteristics will varywith different heater voltage settings. Therefore, it isa good practice to periodically check the calibrationof the sensor with the actual target gas.

Mixtures of stable noncombustible and nontoxicgases with various concentrations are available frommany supply sources. Check with the instrumentmanufacturer for more detailed information.

D. Gas MixingNot all calibration gases are available. Even if they

are available, it is very possible that they would not beavailable in the right concentration or in the properbackground mixture. However, many mixtures areavailable for some process uses which can be dilutedto use in calibration of gas monitors in lower concen-tration ranges. For example, 50% LEL methane hasa concentration of 2.5% or 25,000 ppm. To make a20% LEL mixture having a volume of 2000 cc, thefollowing formula can be used:

Vb =C

. V, Va =C – Cb . V

Cb C

and

Va = V – Vb

whereCb = concentration in the bottle, 50% in this

caseC = new concentration, 20% in this caseV = total final volume, 2000 cc in this caseVb = volume of mixtureVa = volume of air or other dilutant

170


Vb = 20/50 . 2000 = 800 cc

Va = 2000 – 800 = 1200 cc

The final mixture would be made by taking 800 ccof the calibration gas and mixing it with 1200 cc of airto make the mixture equal to 20% LEL.

Another example is to dilute this 25000 ppm ofmethane calibration gas to make a 100 ppm of mix-ture.

Vb = 100/25000 . 2000 = 8 cc

therefore

Va = 2000 – 8 = 1992 cc

By mixing 8 cc of calibration gas into 1992 cc ofair, 2000 cubic centimeters of 100 ppm gas mixture isobtained.

Some Calibration Tools

To perform the above procedure, the followingtools are needed:

1. Syringe and Needle: This is themost inexpensive way to accuratelymeasure the amount of gas. A dis-posable medical syringe with a largegage needle is most practical but

there are few syringes with more than one hundredcubic centimeter volume. Hence, large volume mea-surements can be troublesome. However, it is easy tomake a syringe using any standard size pipe havingabout a 2-inch diameter. It provides an easy and con-venient means to make a mixture on a regular basis.For very small volume measurements, there are mi-cro syringes that are readily available in chemicalsupply catalogues.

2. Calibration Bag: Most of the materials used infood packaging or storage are quite inert; otherwise,food would be contaminated with odor. Therefore,

A 2”-Diameter, 1000 cc Syringe

Standard Medical Syringes:1 cc, 10 cc, and 50 cc

171


food storage bags can be used to hold most chemicalsas long as they are used for relatively short durations.This is an important point to keep in mind since gasmolecules will eventually diffuse through the manythin layers of a plastic bag. For example, potato chipscan stay fresh in their original bag for long periods oftime because the bag material is less permeable bygas molecules than normal food storage bags. This isdemonstrated by the fact that when the potato chipsare transferred into a tightly sealed food containerbag, they will lose their crispness in a very short time.There are also many commercially available samplingbags on the market. One common example is a Tedlarbag. It is made from polyvinyl fluoride and has lowabsorption of gas molecules. However, this type of bagis still permeable, so a heavy gauge material will beneeded if permeability is a major concern. Samplingbags normally come with a valve and a septum that isused as an injection port.

Pressure Formula

Earlier, we described preparing a mixture basedon a volume relationship. Based on the ideal gaslaw, the same volume formula can be used as a pres-sure formula. As an illustration, take an 800 psi mix-ture of 50%LEL methane with a 1200 psi mixtureof air. This will result in a 2000 psi mixture of20%LEL methane.

Preparing gas mixtures can be a very difficult task.It is best to consult with the instrument manufac-turer regarding the best method of calibration andavailability of gas mixtures.

Following are some examples of how gas mix-tures can be made:

For ppm gas mixtures:

Cppm = Vc /(Vc + Vd) . 106 ppm

where Vc is target gas volume and Vd is the

A 5-liter Sampling Bag

1000 cc Calibration Cans

Mixing Calibration Gas

172


dilutant volume. For example, what is the concentra-tion in parts per million when 1 cc of CO is added toa 1000 cc container?

Cppm = 1/1000 . 106 = 1000 ppm

For % range gas mixtures:

C% = Vc /(Vc + Vd) . 102 %

The Vc term in the denominator can be insignifi-cant in low ppm mixtures.

Calibrating Liquid Chemical Mixtures. To make acalibration mixture for liquid chemicals, a known vol-ume of liquid is vaporized in a known volume ofdilutant air. The ideal gas law states that one grammole of molecules will occupy 24,500 cc of volume at25 degree centigrade and at 760 mm of mercury orsea level atmospheric pressure. This temperature andpressure is also called the standard condition. At stan-dard conditions, the equation is:

Cppm = 24.5 . 109 . (V x D)/(Va . M)

where V = volume of liquid, D = density of the liq-uid, which is the same as the specific gravity, Va = thevolume of the dilutant air, and M = the molecularweight of the liquid.

Since it is easier to measure the liquid using a mi-cro syringe, the equation then becomes

V = Cppm . Va . M/(24.5 . 109 . D)

where all units are in milliliters, cubic centimeters,and grams.

For example, benzene has M = 78.1 g and D =0.88 g/cc. What is the amount of benzene needed tomake a 1000 ppm mixture in a 2000 cc bottle?

V = 1000 . 2000 . 78.1/(24.5 . 109 . 0.88)

which yields

V = 7.2 . 10-3 = 0.0072 cc = 7.2 microliters

Many containers are sized by gallons.Therefore, it is useful to know thatone gallon is equal to 3785 cc.

173


In air pollution, industrial hygiene, and medicaltoxicology work, the commonly used unit of concen-tration is milligram per cubic meter. The followingequation expresses this relation, again assuming stan-dard conditions.

Cppm = C . 24.5/M

where

C = mg/m3

M = molecular weight

In conclusion, for the calibration of gas monitors,accuracy is not extremely important because these arenot analytical devices or systems. However, it is most im-portant to keep the calibration methods standardized andeasily traceable. If procedures are standardized, datacan be normalized at a later date if necessary.

Ventilation Hood for Calibration Procedures

191

Appendix I PID Correction Factors

Compound Name Formula 9.8 C 10.6 C 11.7 C IP (eV)

Acetaldehyde C2H4O 5.5 + 10.23

Acetic Acid C2H4O2 NR + 22 + 2.6 + 10.66

Acetic Anhydride C4H6O3 NR + 6.1 + 2.0 + 10.14

Acetone C3H6O 1.2 + 1.1 + 1.4 + 9.71

Acetonitrile C2H3N 100 12.19

Acetylene C2H2 2 11.40

Acrolein C3H4O 42 + 3.9 + 1.4 + 10.10

Acrylic Acid C3H4O2 12 + 2.0 + 10.60

Acrylonitrile C3H3N NR + 1.2 + 10.91

Allyl alcohol C3H6O 2.4 + 1.7 9.67

Allyl chloride C3H5Cl 4.3 0.7 9.9

Ammonia H3N NR + 9.7 + 5.7 + 10.16

Amyl alcohol C5H12O 5 10.00

Aniline C7H7N 0.50 + 0.5 + 0.5 + 7.72

Anisole C7H8O 0.8 8.21

Benzaldehyde C7H6O 1 9.49

Benzene C6H6 0.55 + 0.5 + 0.6 + 9.25

Benzonitrile C7H5N 1.6 9.62

Benzyl chloride C7H7Cl 2 0.7

Bromobenzene C6H5Br 0.6 0.5 8.98

Bromoform CHBr3 NR + 2.5 + 0.5 + 10.48

Bromopropane,1- C3H7Br 150 + 1.5 + 0.6 + 10.18

Butadiene C4H6 1.0 + 1.1 9.07

Butadiene diepoxide,1,3- C4H6O2 25 + 3.5 + 1.2 ~10

Butane C4H10 1.2 10.53

Butanol, 1- C4H10O 70 + 4.7 + 1.4 + 9.99

Butene, 1- C4H8 0.9 9.58

Butoxyethanol, 2- C6H14O2 1.8 + 1.2 + 0.6 + <10


Notes:

1. The values indicated by a plus (+) sign inthe “C” column are confirmed values; allothers are preliminary and subject to change.

2. The correction factors in this table were mea-sured in dry air. Actual readings may varywith age and cleanliness of the lamp, rela-tive humidity, components in the sample,and other factors. For accurate work, theinstrument should be calibrated regularlyunder the operational conditions in whichthe instrument is used.

3. IP (eV) data was taken from the CRC Hand-book of Chemistry and Physics, 73rd Edition,and NIST Standard Ref. Database 19A.

(Courtesy of RAE Systems, Inc.)

Appendix I

PID Correction Factorsfor Various Gases and Lamps with Instrument Calibrated to Isobutylene

192



Butyl acetate, n- C6H12O2 2.6 + 10

Butyl acrylate, n- C7H12O2 1.6 + 0.6 +

Butylamine C4H11N 7 8.71

Butyl mercaptan C4H10S 0.5 9.14

Carbon disulfide CS2 1.2 + 0.3 10.07

Carbon tetrachloride CCl4 NR + NR + 1.7 + 11.47

Chlorine Cl2 1.0 + 11.48

Chloro-1,3-butadiene, 2- C4H5Cl 3

Chlorobenzene C6H5Cl 0.44 + 0.40 + 0.39 + 9.06

Chloro-1,1-difluoroethane, 1- (R-142B) C2H3ClF2 NR NR 12.0

Chlorodifluoromethane CHClF2 NR NR NR 12.2

Chloroethane C2H5Cl NR + NR + 1.1 + 10.97

Chloroethanol C2H5ClO 10.52

Chloroethyl methyl ether,2- C3H7ClO 3

Chloroform CHCl3 NR + NR + 3.5 + 11.37

Chlorotoluene, o- C7H7Cl 0.5 0.6 8.83

Chlorotoluene, p- C7H7Cl 0.6 8.69

Crotonaldehyde C4H6O 1.5 + 1.1 + 1.0 + 9.73

Cumene C9H12 0.58 + 0.5 + 0.4 + 8.73

Cyanogen bromide CNBr NR NR NR 11.84

Cyanogen chloride CNCl NR NR NR 12.34

Cyclohexane C6H12 1.4 + 9.86

Cyclohexanol C6H12O 1.1 9.75

Cyclohexanone C6H10O 1.0 + 0.9 + 0.7 + 9.14

Cyclohexene C6H10 0.8 + 8.95

Cyclohexylamine C6H13N 1.2 8.62

Cyclopentane C5H10 0.6 10.51

Decane C10H22 4.0 + 1.4 + 0.4 + 9.65

Diacetone alcohol C6H12O2 0.7

Dibromoethane, 1,2- C2H4Br2 NR + 1.7 + 0.6 + 10.37

Dichlorobenzene, o C6H4Cl2 0.54 + 0.47 + 0.38 + 9.08

193


Compound Name Formula 9.8 C 10.6 C 11.7 C IP (eV) Dichlorodifluoromethane CCl2F2 NR + NR + 11.75

Dichloroethane, 1,1- C2H4Cl2 11.06

Dichloroethane, 1,2- C2H4Cl2 NR + 0.6 + 11.04

Dichloroethene, 1,1- C2H2Cl2 0.9 9.79

Dichloroethene, c-1,2- C2H2Cl2 0.8 9.66

Dichloroethene, t-1,2- C2H2Cl2 0.5 + 0.3 + 9.65

Dichloro-1-fluoroethane, 1,1- (R-141B) C2H3Cl2F NR + NR + 2.0 +

Dichloropropane, 1,2 C3H6Cl2 0.7 10.87

Dichloro-1-propene, 2,3- C3H4Cl2 1.9 + 1.3 + 0.7 + <10

Dichloro-1,1,1-trifluoroethane, 2,2- (R-123) C2HCl2F3 NR + NR + 10.1 + 11.5

Diesel Fuel #1 m.w. 226 0.9 +

Diesel Fuel #2 m.w. 216 0.7 + 0.4 +

Diethylamine C4H11N 1 + 8.01

Diethylaminopropylamine, 3- C7H18N2 1.3

Diethylmaleate C8H12O4 4

Dimethylacetamide, N,N- C4H9NO 0.87 + 0.8 + 0.8 + 8.81

Dimethylamine C2H7N 1.5 8.23

Dimethyl disulfide C2H6S2 0.2 + 0.20 + 0.2 + 7.4

Dimethylformamide, N,N- C3H7NO 0.8 9.13

Dimethylhydrazine, 1,1- C2H8N2 0.8 + 0.8 + 7.28

Dimethyl sulfate C2H6O4S ~23 ~20 + 2.3 +

Dioxane, 1,4- C4H8O2 1.1 9.19

Epichlorohydrin C2H5ClO ~200 + 8.5 + 1.4 + 10.2

Ethane C2H6 NR + 15 + 11.52

Ethanol C2H6O 12 + 8 10.47

Ethanolamine (Not Recommended) C2H7NO ~4 + ~3 + 8.96

Ethene C2H4 10 + 3 10.51

Ethoxyethanol, 2- C4H10O2 3.5 9.6

Ethyl acetate C4H8O2 4.6 + 10.01

194


Compound Name Formula 9.8 C 10.6 C 11.7 C IP (eV) Ethyl acrylate C5H8O2 2.4 + 1.0 + (<10.3)

Ethylamine C2H7N 0.8 8.86

Ethylbenzene C8H10 0.52 + 0.5 + 0.5 + 8.77

Ethylene glycol C2H6O2 16 + 6 + 10.16

Ethylene oxide C2H4O 19 + 3 + 10.57

Ethyl ether C4H10O 1.1 + 9.51

Ethyl formate C3H6O2 1.9 10.61

Ethyl hexyl acrylate, 2- C11H20O2 1.1 + 0.5 +

Ethyl (S)-(-)-lactate C5H10O3 13 + 3.2 + 1.6 + ~10

Ethyl mercaptan C2H6S 0.6 9.29

Ethyl sulfide C4H10S 0.5 + 8.43

Formaldehyde CH2O 0.6 10.87

Furfural C5H4O2 0.9 + 0.8 + 9.21

Gasoline #1 m.w. 72 0.9 +

Gasoline #2, 92 octane m.w. 93 1.3 + 1.0 + 0.5 +

Glutaraldehyde C5H8O2 1.1 + 0.8 + 0.6 +

Halothane C2HBrClF3 0.6 11.0

HCFC-123 (see 2,2-Dichloro-1,1,1-trifluoroethane, R-123)

HCFC-141B (see 1,1-Dichloro-1-fluoroethane)

HCFC-142B (see 1-Chloro-1,1-difluoroethane)

Heptane, n- C7H16 2.6 + 0.5 9.92

Hexamethyldisilazane,1,1, 1,3,3,3- HMDS C6H19NSi2 0.2 + 0.2 + ~8.6

Hexane, n C6H14 300 4.3 + 0.5 + 10.13

Hexene, 1- C6H12 0.8 9.44

Hydrazine H4N2 2.6 + 2.1 + 8.1

Hydrogen H2 NR + NR + NR + 15.43

Hydrogen peroxide H2O2 NR + NR + NR + 10.54

Hydrogen sulfide H2S NR + 3.3 + 1.5 + 10.45

Iodine I2 0.1 + 0.1 + 0.1 + 9.40

Isobutane C4H10 100 + 1.2 + 10.57

Isobutanol C4H10O 19 + 3.8 + 1.5 10.02

195



Isobutene C4H8 1.00 + 1.00 + 1.00 + 9.24

Isobutyl acrylate C7H12O2 1.5 + 0.60 +

Isoflurane C3H2ClF5O

Isooctane C8H18 1.4 9.86

Isopar G Solvent m.w. 148 0.8 +

Isopar M Solvent m.w. 191 0.7 + 0.4 +

Isophorone C9H14O 3 9.07

Isoprene C5H8 0.69 + 0.6 + 0.60 + 8.85

Isopropanol C3H8O 500 + 6.0 + 2.7 10.12

Isopropyl acetate C5H10O2 2.5 9.99

Isopropyl ether C6H14O 0.8 9.20

Jet fuel JP-4 m.w. 115 1 + 0.4 +

Jet fuel JP-5 m.w. 167 0.6 + 0.5 +

Jet fuel JP-8 m.w. 165 0.6 + 0.3 +

Kerosene (C10-C16 petro.distillate - see Jet Fuels)

Mesitylene C9H12 0.36 + 0.35 + 0.3 + 8.41

Methane CH4 NR + NR + NR + 12.51

Methanol CH4O NR + NR + 2.5 + 10.85

Methoxyethanol, 2- C3H8O2 4.8 + 2.4 + 1.4 + 10.1

Methoxyethoxyethanol, 2- C7H16O3 2.3 + 1.2 + 0.9 + <10

Methyl acetate C3H6O2 1.6 10.27

Methyl acrylate C4H6O2 3.7 + 1.2 + (9.9)

Methylamine CH5N 1.0 8.97

Methyl bromide CH3Br 110 + 1.7 + 1.3 + 10.54

Methyl t-butyl ether C5H12O 0.9 + 9.24

Methyl cellosolve (see 2-Methoxyethanol)

Methyl chloride CH3Cl NR + NR + 0.7 + 11.22

Methylcyclohexane C7H14 1.1 9.64

Methylene chloride CH2Cl2 NR + NR + 0.89 + 11.32

Methyl ethyl ketone C4H8O 0.86 + 0.9 + 1.1 + 9.51

Methylhydrazine C2H6N2 1.4 + 1.2 + 1.3 + 7.7

Methyl isobutyl ketone C6H12O 1.2 + 0.9 9.30

196


Compound Name Formula 9.8 C 10.6 C 11.7 C IP (eV) Methyl isocyanate C2H3NO NR + 4.6 + 1.5 10.67

Methyl mercaptan CH4S 0.6 9.44

Methyl methacrylate C5H8O2 1.4 + 1.4 9.7

Methyl propyl ketone C5H12O 0.9 + 0.8 + 9.38

Methyl-2-pyrrolidinone, N- C5H9NO 1.0 + 0.8 + 0.9 + 9.17

Methyl salicylate C8H8O3 2

Methylstyrene, a- C9H10 0.5 8.18

Mineral spirits (Stoddard Solvent, see also Viscor 120B) 0.7 + 0.39 +

Mineral Spirits Viscor 120B Calibration Fluid 1.0 + 0.7 + 0.3 +

Naphthalene (Mothballs) C10H8 0.45 + 0.4 + 0.40 + 8.13

Nitric oxide NO 5.2 + 2.8 + 9.26

Nitrobenzene C6H5NO2 2.6 + 1.9 + 1.6 + 9.81

Nitroethane C2H5NO2 3 10.88

Nitrogen dioxide NO2 NR + NR + 9.75

Nitromethane CH3NO2 4 11.02

Nitropropane, 2- C3H7NO2 2.6 10.71

Nonane C9H20 2 9.72

Octane, n- C8H18 13.2 + 1.8 + 9.82

Pentane C5H12 80 + 8.4 + 0.7 + 10.35

Peracetic acid C2H4O3 NR + NR + 2.3 +

Peracetic/Acetic acid mix C2H4O3/C2H4O2 50 + 2.5 +

Perchloroethene C2Cl4 0.69 + 0.57 + 0.31 + 9.32

PGME C6H12O3 2.4 + 1.5 + 1.1 +

PGMEA C6H12O3 1.65 + 1.0 + 0.8 +

Phenol C6H6O 1.0 + 1.0 + 0.9 + 8.51

Phosphine in N2 PH3 2 + 1.4 9.87

Photocopier Toner 0.5 + 0.3 +

Picoline, 3- C6H7N 0.9 9.04

Pinene, a- C10H16 0.3 + 0.5 8.07

197


Compound Name Formula 9.8 C 10.6 C 11.7 C IP (eV) Pinene, b C10H16 0.38 + 0.4 + 0.4 + ~8

Piperylene, isomer mix C5H8 0.76 + 0.7 + 0.6 + 8.6

Propane C3H8 NR + 1.8 + 10.95

Propanol, n- C3H8O 6 1.7 10.22

Propene C3H6 1.7 + 9.73

Propionaldehyde C3H6O 1.9 9.95

Propyl acetate, n- C5H10O2 3.5 10.04

Propylene oxide C3H6O 6.5 2 10.22

Propyleneimine C3H7N 1.5 + 1.3 + 1.0 + 9.0

Pyridine C5H5N 0.78 + 0.7 + 0.7 + 9.25

RR7300

(70:30 PGME/PGMEA) C4H10O2/C6H12O3

1.4 + 1.0 +

Stoddard Solvent - see Mineral Spirits

Styrene C8H8 0.45 + 0.40 + 0.4 + 8.43

Sulfur dioxide SO2 NR + NR + 12.32

Tetrachloroethane, 1,1,1,2- C2H2Cl4 1.3 ~11.1

Tetrachloroethane, 1,1,2,2- C2H2Cl4 NR + NR + 0.60 + ~11.1

Tetraethyllead C8H20Pb 0.4 0.3 0.2 ~11.1

Tetraethyl orthosilicate C8H20O4Si 0.7 + 0.2 + ~9.8

Tetrafluoroethane, 1,1,1,2- C2H2F4 NR NR

Tetrafluoromethane CF4 NR + NR + >15.3

Tetrahydrofuran C4H8O 1.9 + 1.7 + 1.0 + 9.41

Therminol 0.90 + 0.7 +

Toluene C7H8 0.54 + 0.50 + 0.51 + 8.82

Tolylene-2,4-diisocyanate C9H6N2O2 1.4 + 1.4 + 2.0 +

Trichloroethane, 1,1,1- C2H3Cl3 NR + 1 + 11

Trichloroethane, 1,1,2- C2H3Cl3 NR + NR + 0.9 + 11.0

Trichloroethene C2HCl3 0.62 + 0.5 + 0.4 + 9.47

Trichlorotrifluoroethane, 1,1,2- CFC-113 C2Cl3F3 NR NR 11.99

Triethylamine C6H15N 1.3 7.50

Trifluoroethane, 1,1,2- C2H3F3 34 12.9

198



Trimethylamine C3H9N 0.9 7.82

Trimethylbenzene, 1,3,5- - (see Mesitylene)

Turpentine C10H16 0.4

Undecane C11H24 2 9.56

Vinyl actetate C4H6O2 1.2 9.19

Vinyl bromide C2H3Br 0.4 9.80

Vinyl chloride in N2 C2H3Cl 2.0 + 0.6 + 9.99

Vinyl-2-pyrrolidinone, 1- C6H9NO 1.0 + 0.8 + 0.9 +

Viscor 120B - see Mineral Spirits - Viscor 120B Calibration Fluid

Xylene, m- C8H10 0.50 + 0.4 + 0.40 + 8.56

Xylene, o- C8H10 0.57 + 0.6 + 0.7 8.56

Xylene, p- C8H10 0.5 + 0.6 + 8.44

Thermocatalitic sensor in dynamic mode 1. General principle Schematic diagram of sensor for methane concentration in dinamic mode is shown on the figure below:

C0

C

St

Figure 1

1

2

3

4

When the coil is powered, it will start to heat up, and gas into chamber will start to combust. At the beguinning inflow of the gas through calibrated hole is smaller then the quantity of the combusted gas onto the coil, concentration of gas C will be reduced. When the quantity of combusting gas becomes equal to quantity of inflow gas, concentration stabilize with new value Ci (see diagram 3). It is easy to show theoreticaly, and experiments certify, that gas concentration into chamber reduce with exponential trend. At the moment when concentration almost reach Ci, power is turned off and the chamber starts to fill again, it fills up for about the same time. The power is turned on again and so on alternatively. Transitive process in which we measure the signal in two consecutively points is repeated, and on that base we determine concentration of the target gas, as we will see later. It is validated truth that transitial process is much more reliable and much informational than the stationar.

1. Cover made of strained metal

chippings or porous ceramic 2. Calibrated hole between prechamber

and combustion chamber 3. Combustion chamber

Thermocatalitic element, platinum coil with small temperature inertia

Co - concentration of the gas outside the sensor

C- concentration of the gas inside the combustion chamber St —overall signal (current or voltage) which is result of heathing of the coil from curent flow and of gas combustion onto the coil.

Ss

Si

Figure 3

I II III IV

One short time period after energizing the coil (t=0) there is no respond from system and that is inertial time of the system, a signal remains null (period I), after that time, coil starts to heat up rapidly caused from current and also from gas combustion until the signal reach its maximal value Ss (period II). In the period III, which is of the greatest interest for metrology, flux of outer gas inflowing through the hole is smaller than the quantity of the combusted gas and the concentration decrease ‘till reach equilibrium point when the signal becomes Si. Signal then stays constant on minimal value Si (period 4). The following trend of exponencial decreasing of the concentration in the period III is changing by following formula:

C = Cs [ r + ( 1 - r ) e-t/T] Where are : r = Ci/Cs factor of concentration reduction, and T — is time constant. The both of this factors are characteristics specific for each. Time t is calculated from the end of the period II (t=0). Because the signal is proportional with concentration S = k x C, it folows:

S = k Cs [ r + (1 - r ) e —t/T ] This is, of course, only a part fo the signal and its origin is from the gas combustion. Other part of the signal, parasitic signal, is the result in the first place of resistor heating caused by current flow but its also caused by ambient temperature. It also consists of other relevant factors as is the air pressure, humidity, resistive changes with time, etc.

St= S + Spar

Temps (Unités arbitraires)

Sign

al to

tal S

t

t = 0 t1 t2

Ss

L = St(t1)-St(t2)

Si

Spar

Figure 4

On this curve coordinate beguinning (t=0) is shifted at the end of period II. Total signal is measured in two points t1 and t2 on transition curve. The difference of the signals in those two points L (t1 , t2 ) = S (t1) - S (t2) is proportional with gas concentration. From theory stated above it can be seen that the number of parameters are before the constructor to adapt them in the best way to his specific applications. We will comment only briefly some of them. The diameter of the calibrated hole and the capacity of combustion of coil, are obviously in relation to each other. It’s stated before that it is beter to have coil with as smaller thermal inertia as it is possibile, but when it is achieved by minimizing the vire dimesions and coil itself, thermocatalitic surface layer should achieve very intensive combustion and in that way increase sensitivity of sensor itself. The diameter of calibrated hole considerably influence on the form of exponential curve, determining its gradient, level of the Ci (Si) and the time for reaching that value (time constant). Selection of the points t1 and t2 in which we measure the signal is of great importance. Point t1 should be chosen far enought from maximal value, to avoid shifting of that point to left side from the maximum, because during time coil characteristic gets deformed and maximum point is sliding slowly towards right side. The distance between points t1 and t2 is from great importance. It won’t be same for stationary devices in mines which are constantly exposed to target gas and devices in industry (stationary or portable) where the presence of target gas is mostly incidently. The proper choose of this two parameters enables construction of so called “universal” explosimeter. As a matter of fact, classical explosimeters shows accurately only concentration of gas for which they are calibrated and indicate it in percents of lower explosive limit (LEL). Any other explosive gas they only indicate, and measuring of explosivity itself can deflect up to 300% in extreme cases. This makes those exposimeters practically useless, in areas where we don’t know which gas is present, or when there is a mixture of gases. Dinamic mode enables to, with proper chose of parameters t1 and t2, realize explosimeter wich will equaly acurate measure explosivity of mostly combustile gases, and for extreme cases deflection is below few tenth of percent (referred to curently 300%). 2. Instrument realized on basis with thermocatalitic sensor in dinamic mode The stated principle is able to realize only with use of micro-controller, which is done in cooperation of French company Oldham and Russian laboratory NPC-ATB, and it is realized and specific measuring instrument is called DMS-01. Measuring range of this device is from 0-100% methane, it has beside catalitic sensor in dinamic mode, which works in range of 0-5% methane, installed in the same measuring chamber another catharometric sensor wich cover range of 5-100% methane.

Design of measuring cell (sensor) is given on following figure:

sinter

HC LC

0.5 - 0.7mm

gas

Principle of opertion is as follows: first of all, micro-controller turns on thermoconductive sensor for high methane concentrations HC with short current impulse which lasts for 0.35 sec., during that time thermocatalitic sensor is turned off (no methane combustion).

As the concentration of methane is higher, the slope is under lower angle because the wire start to cool more intensive (element is constantly energized) this is previously stated distinction of the catarometric sensors. If a methane concentration is above level of DGE (5%CH4), sensor for lower concentrations stays turned off, otherwise it turns on, but only if the concentration is less than DGE. Time line of one measuring cycle, in case that after initial turning on the DMS-01 has measured concentration below DGE, is shown on following figure:

1 — short current impulse for thermoconductive sensor, lasting for 0,35 sec 2 — sensor response HC 3 — current impulse for heating the LC sensor, lasting for 2,5 sec (turns on at t=1s and turns off at

t=3.5s ), when proper temperature is reached (about 450°C) and combustion of methane starts. 4 — sensor response LC in lack of methane 5 — sensor response in presence of methane Briefly, LC sensor is a tipical thermocatalitic sensor, it is installed in a chamber with calibrated hole 0.5 — 0.7mm. The gas pass thru this calibrated hole, sensor (LC) turns on and when wire reach proper temperature the methane combustion starts. Quantity of the gas that combust is in proportion with concentration of the gas that’s inside the chamber, and after some particular time, because the power supply of the sensor is turned off, the balance renew. Transition process from the moment of turning on ‘till stabilization is used for measuring. Measuring is conducted in two points, first point is at 1.3 sec. and the second one at 2.3 sec. (after turning on the sensor). Concentration is equal to difference of these two signals.

1s

1.3s 2.3s

After impulse of 2.5 sec., there is a pause of 5 sec. to fill the chamber again, and the whole process is cyclicly repeating. Total cycle of measurment lasts for 7,5 sec. and is cyclicly repeating, untill to moment when sensor of lower concentration exceed the DGE (5% methane) when the sensor of higher concentration HC turns on. Farther work of this device is on the same principle as after turning on (when the concentration falls below DGE, sensor HC is turned off, and sensor LC turns on). 3. Analysis of Ex protection for thermocatalitic sensor in dynamic mode

The next results are achieved in establishing a special conditions for Ex protection:

- Power supply of the sensor via electronic circuitry with intrinsically safety with maximal values of electrical parameters: Um=1.41V, Iib=116 mA.

- Limited heating of the thermoelement to safety limit (≤ 450°C). - Thanks to exceptional low power consumption enough for accurate sensor operation (low wire

diameter), a high temperature coefficient of resistive change ( K=3.9 x 10-3), and rather low thermal inertia (measuring cycle lasts for 7.5 sec., 2.5 sec for heating, and 5 sec. for cooling down), its created the effect of developing a rather low amount of thermal energy, which leads to low operating temperature of the sensor (Tmax=73°C), and which is a far lower from maximal safety temperature tbez=295°C, derived with a safety factor of 1.5. Measurements in stated example are achieved with power supply conditions U=1.41V, I=116 mA, in presence of mixture of propane 5,3% and ambient air, at the temperature of 50°C (mixture is in comply with group of gases IIA).

- Inlet of the target gas inside the sensor is thru sintered mass which eliminate possibility of intruding of coal dust inside the sensor, which lead to higher permited temperature limit inside the sensor from 1500 to 450°C.

Design of the sensor, that means the operational principle of dynamic mode, have allowed to designers to construct the measuring intrument, with intrinsically safety conditions in compy with limit curves (IEC 79-11) witout any problems, i.e. far from the limit curve. 4. Summary Thermocatalitic sensors in dynamic mode of operation has following advantages compared to “classical” thermocatalitic sensor: 1. With only one element, without reference, we determine the gas concentration, avoiding in that way any inconvenience which reference element can cause (need for selection — complying of elements, operative with the comparative, different aging of this two elements, which is inevitable, and which lead to zero drift, etc.)

2. With dynamic mode, actually we compare element with itself, it is referent to itself, comparative, if its characteristics change over time, those changes are the same in points t1 and t2. (figure No. 4). On that way problem of zero drift is solved which is one of the biggest advantages of this sensor. 3. Power consumption of sensor itself is manifold smaller in compare to classical thermocatalitic sensors, because there is no reference element, thus only one element is heating (consumption is 50% reduced). Sensor isn’t working all the time (only 2.5sec) and 5 sec is the time for new charging of the chamber (additional consumption reduction). And finally, thin wire of 10 μm heats and cools down very quickly. 4. Unlike classical thermocatalitic sensors which has to be installed in flameproof housing, i.e. realised in Ex d, Ex protection of this sensor is Ex ia, thus eases in a great way construction of detector itself. 5. Proper choose of parameters of thermocatalitic sensor in dynamic mode allows construction of so called “universal” explosimeter. Particulary, classical explosimeters shows acurate readings only for concentrations of gas for which they are calibrated for and indicate it in percents of lower explosive limit (LEL). Any other explosive gas, they only indicate, and readings itself of the explosivity could deflect up to 300% in extreme cases. This make them practically useless, in areas where we don’t know which gas is present, or when there is a mixture of gases. Dynamic mode allows us to realize explosimeter, with proper selection of parameters (t1 and t2), which will equaly acurate measure explosivity of most of common combustive gases, and for those extreme cases deflection is around few tenth of percents (referred to curently 300%). General conclusion is, that the despite of use of IR sensors, changed thermocatalitic sensor in dynamic mode of operation has its distinctive position as well in mining as so in industry.

hazardous gas monitors

Documents