PB85-UH046
EPA-600/4-85-002
January 1985
EVALUATION OF CRV08EN1C TRAPPING
AS A MEANS FOR C0U.ECT1NQ OROANIC
COMPOUNDS IN AMBIENT AIR
by
Mlchaol Holdran, Stavan Ruit,
Richard Salth and John Koats
Battalia
ColtMbut Laboratories
505 King Avanuo
Colwbus. Ohio 43201
Contract No. 66-02-3447
Work Asslgnaant No. 22
Projact Offlctr
U1U1M A. MeClamuf
Cnvlronaantal MonltorlM Division
Environmental Monitoring Syttaas Laboratory
Research Trlanglo Park, North Carolina 277U
ENVIRONMENTAL N0N1T0RINI SYSTEMS LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION ASENCV
RESEARCH TRIAMLE PARK, NORTH CAROLINA 27711

-------
	TienmCAL RVOMT frjTA . .
EPA-600/4-85-002 1
I. MteifttNTI ACCI«4i4n MO.
PB85-144046
4. tlTWi AMD tUrriTkl " ¦
Evaluation of Cryogenic Trapping as a Meant for
Collecting Organic Compounds In Ambient A1r.
ft. MWRT OATt
January 1985
1. PClMOAMINa ORGANIZATION COOC
7. AuTHOWU " """ ' 			
M. Holdren, S. Rust, R. Smith and J. Koeti
ft. WHOMMNO OHOAMIXATIOM AtPOftT NO.
».P|*FO*MIMOOWOANI*ATIONItAMa ANOAOOfliift
Battel1e Columbus Laboratories
SOS King Avenue
Columbus, Oil 43201
ISl PMOOHaU iklMINT NO.
68-02-3467 US EPA
»J. SPO4MOMIN0 AMNCY na*m amo aoomim
US EPA
Environmental Monitoring Systems laboratory
Research Triangle Park, KC 27711
is. tvm o* naroAT and ramoD covtnao
i«. fowomwq AttlNCV coot
EPA/600/08
1». tU^LtMflMTAAV NOTM
The methodology used 1n reduced temperature preconcentratlon of volatile organic
compounds has been tested using a prototype automated gas chromatographic system.
Mixtures of sixteen volatile oraanlc compounds 1n humidified zero air were passed
through a Nafion tube dryer and the organic compounds Here collected on a reduced-
temperature trap. The dryer reduced the water concentration without significantly
affecting the Integrity of the trace organic species. The selective reduction of
water vapor Improves the chromatography of the trace organlcs and likewise permits
processing larger simple volumes.
Collection and recovery efficiencies from the trap (I to 10 ppbv compound) were 100
~ S percent with this preconcentratlon technique. The Integrity of sample
components was unaffected by co-collection of osone and nitrogen dioxide at typical
Mblent concentrations. Two nominally Identical automated gas chromatographic
Instruments gave nearly Identical results during simultaneous monitoring of
calibration mixtures and laboratory i1r. Calibration mixtures at ppbv levels were
also collected In passlvated stainless steel canisters. Statistical analysis of
the data Indicated that the canisters would serve as acceptable storage vessels for
tttV WOftM AMO MCUtfCNT ANAV Wt
I. NMMIttOM
INOCO TIMMt
l. OOtATI FteM/Qraip

mamma
in mwrnu




66


•M ?«• MM Ht* 4-W NIVIMt ¦•»?»•# IIMMIMI ^

-------
DISCLAIMER
Tha Information 1n this docimnt has baan fundad wholly or In part by
the U. S. Envlronaantal Protection Agancy undar Contract No. 68-02-3487 to
Battalia ColiMbus Laboratories. It hat baan tubjact to tha Agency's paar and
administrative review, and It hat baan approved for publication as an EPA
document. Nantlon of trada nines or conaarclal products doas not constltuta
endorseeent or recwandatlon for usa.
11

-------
FOREWORD
Measurement and Monitoring research efforts art designed to *nt1c1p«t«
potential environmental probleMS, to support regulatory actions by developing
an in-depth understanding of the nature and processes that Impact health and
the ecology, to provide Innovative Mans of Monitoring compllance with
regulations, and to evaluate the effectiveness of health and envlronMntal
protection efforts through the Monitoring of long-tens trends. The
Environmental Monitoring System Laboratory. Research Triangle Park, North
Carolina, has responsbfllty fort assessment of environmental Monitoring
technology and systems for air; Implamentation of agency-wide quality assurance
programs for air pollution measurement systems; and supplying technical
support to other groups 1n the Agency including the Office of Air and
Radiation, the Office of Toxic Substances, and the Office of Solid Haste.
New monitoring procedures for volatile organlcs In Mblent air have
recently evolved using modern capillary column gas chromatography. The high
resolution capability combined with s«*1e preconcentratlon procedures and gas
chromatographic detectors responding to specific subgroups of volatile
organlcs has significantly advanced our monitoring capabilities. This study
addresses the design and characterisation of an automated monitoring system.
Emphasis 1s placed on the evaluation of system features that can affect the
SMiple Integrity during reduced temperature preconcentratlon.
Thomas R. Hauler, Ph.D
01rector
Environmental Monitoring Systems Laboratory
Research Triangle Park, North Carolina 27711
111

-------
ABSTRACT
Th« methodology used In reduced temperature preconcentratlon of volatile
organic compounds has been tested using a prototype automated gas
chromatographic system. Mixtures of sixteen volatile organic compounds In
humidified zero t1r were pass«td through a Naflon tube dnrer and the organic
conpcunds were collected on a reduced-tamperature trap. The dryer reduced the
water concentration without significantly affecting the Integrity of the trace
organic species. The selective reduction of water vapor proves the
chromatography of the trace organlcs and likewise penalts processing larger
sample volumes.
Collection and recovery efficiencies froia the trap (.3 to 3 ppbv
compound) were 100 ~ S percent with this preconcentratlon technique. The
Integrity of sample components was unaffected by co-collection of ozone and
nitrogen dioxide et typical ambient concentrations. Tto nominally Identical
automated gas chromatographic Instruments gave nearly Identical results during
simultaneous monitoring of calibration mixtures and laboratory air.
Calibration mixtures at ppbv levels were also collected 1n passlvated
stainless steel canisters. Statistical analysis of the data Indicated that
the canisters would serve as acceptable storage vessels for most of the
volatile organic compounds tested.
This report was submitted In fulfillment of Contract No. 66-02-3487 by
Battalia Columbus Laboratories under the sponsorship of the U.S.
Environmental Protection Agency. This report covers a period from February
1981 to February 1984.
IV

-------
CONTENTS
For word .		 . . . .			111
Abstract	1v
Figures			vl
T«bl«i 				y1l
Acknowledgment			v111
1.	Introduction				1
2.	Conclusions . 			 			3
3.	Recommendations		4
4.	Experimental								5
Chemlcels. 					5
Instrumentation			5
Procedure		8
5.	Results and Discussion. .... 		16
Basic laboratory Studies 		16
R#f*r*nces			*			34
Append lefis
A.	Automated QC System Software.			36
B.	Raw Area Values for Canister Samples. . 		46
v

-------
FIGURES
Number	Page
1	Diagram of Automated Sampling and Analysis System 		6
2	Diagram of Flow Scheme 1n Prototype Calibrator		9
3	Experimental Set-Up for the Perma Pure Dryer Studies	10
4	Configuration of Automated GC System During Intercomparlson
Studies 			 . 		12
5	Canister Apparatus During Sample Fill Cycle 		14
6	Canister Apparatus During Sample Analysis Cycle 		15
7	Automated Chromatographic Analyses of a Standard Mixture
Using Cryogenic Preconcentratlon Techniques and Flame
Ionization Detection (100 mL sample)	17
vl

-------
TABLES
1	Statistical Analyses of Chromatograms Shown In Figure 7 (Test 1)
and Additional Analysis at Lower Concentrations (Test 2)	18
2	Retention Times for Several Compounds Cryogenlcally Collected
and Analyzed Under Dry and Humidified Conditions	18
3	Sample Integrity Results with a Perma Pure Dryer for Those
Target Compounds Responding to a Flame Ionization Detector
—Battelle Results	19
4	Sample Integrity Results with a Perma Pure Dryer for Compounds
Responding to a Flame Ionization Detector—EPA Results	20
5	Sample Integrity Results with a Perma Pure Dryer for Compounds
Responding to an Electron Capture Detector	21
6	Ozone and Nitrogen Dioxide Interference Results Using Flame
Ionization Detection*	23
7	List of Target Compounds Examined During Calibration Runs 	 26
8	Data Obtained from the Statistical Analysis of a Multipoint
Calibration Run (1,2-Olchloroethane) with GC No. 1. 	 26
9	Data Obtained from the Statistical Analysis of a Multipoint
Calibration Run (1,2-Dlchloroethane) with GC No. 2	 27
10	Comparison of Two Automated GC Units Sampling Simultaneously
from a Common Manifold			29
11	Sample Means and Standard Deviations for Canister Sample
Raw Area Values		 30
12	Point Estimates and 99 Percent Confidence Intervals
for the Percent Relative Bias on D*ys 2, 4, and 7	31
13	Summary Table for Comparison of Control Data with Combined
Canister Oata ...... 	 ......... 	 33
IB Raw Area Values for Canister Samples 	 46
vll

-------
ACKNOWLEDGMENT
The authors thank J. D. Plell and K. D. Oliver of Northrop Services,
Inc. for technical assistance 1n obtaining the canister data and 1n conducting
the gas chromatographic Intercomparlson studies. The technical advice of the
Project Officer, W. A. McClenny, throughout the course of the program 1s also
greatly appreciated.
vllt

-------
SECTION 1
INTRODUCTION
The atmosphere contains a complex mixture of organic compounds. Many of
these emitted chemicals have been found to be highly toxic substances. Data
on the Identity and concentration levels of these compounds In urban and rural
environments are continually being gathered by researchers who are attempting
to better understand the chemistry and fate of these chemicals as well as the
extent of human exposure to these compounds(1-5). From an analytical
standpoint, this data gathering task has been formidable. However, these
efforts have been dramatically Improved 1n recent years through the use of
capillary column gas chromatography coupled to compound-specific detection
systems. Furthermore, the Interfacing of the gas chromatograph to a mass
spectrometer detector has oermltted positive Identification of many of these
atmospheric const1tuents($-»).
Most of these toxic chemicals are present at ppb levels or less and
preconcentratlon 1s often necessary for accurate chemical analysis. Presently
the two primary ambient air preconcentratlon techniques employed 1n
delineating atmospheric organic burdens are cryogenic trapping and solid
adsorbents.
Cryogenic trapping can Involve either of two collection procedures.
With the first procedure, whole air samples are entirely condensed within
suitable collection devices by cryoaens such as liquid nitrogen or liquid
helium. Both cryogens are sufficiently cool to serve as the cryogenic pump 1n
this collection process. This approach Is normally used 1n studies where
remote sampling Is being conducted(^-12), once the sample 1s collected, the
coolant 1s removed and the container returned to the laboratory for subsequent
analysis.
The other cryogenic trapping procedure Involves passing air through a
reduced-temperature trap. At the appropriate temperature, trace organic
species will condense onto the trap surfaces while oxygen and nitrogen pass
through the system. Reduced-temperature trapping has several limitations
which must be considered when designing a sampling and analytical system. A
limiting factor of major Importance Is the co-collection of water In the
sampling trap. One liter of a1r at SO percent relative humidity and 25 C will
contain approximately 10 mg of water, which appears as 1ce 1n the collection
trap. The possibility of the 1ce plugging the trap and stopping sample flow
Is of concern, and water transferred to the gas chromatographic capillary
column may also cause plugging and deleterious column effects. Furthermore,
during sample preconcentratlon, chemical reactions may also occur In the
collection trap. Possible reactants could Include anwon1a/ac1ds,
ozone/oleflns, etc.
1

-------
The Advanced Analysis Techniques Branch of the Environmental Monitoring
Systems Laboratory (EMSL) 1s responsible for the development and evaluation of
state-of-tbe-art and emerging analytical techniques for the determination of
organic compounds In ambient air. Recently a priority listing of volatile
organlcs has been established and the EMSL 1s focusing on further development
of analytical methodology associated with the detection of these compounds.
Primary emphasis has been placed on developing field-compatible analytical
systems.
*
Recently a prototype automated analytical system Incorporating cryogenic
trapping for sample preconcentratlon has been developed jointly by EPA and the
Battelle laboratory. System hardware utilizes capillary column gas
chromatographic separation techniques along with flame Ionization and electron
capture detection. Software development using the basic programming
capability of the gas chromatography system permits calibration and ambient
sampling to be achieved with minimal operator Interfacing. In this report, we
shall describe Instrumentation hardware and software and shall discuss
laboratory experiments designed to test the suitability of the prototype
system for preconcentratlon of volatile organic compounds (VOCs) In ambient
air. The VOCs selected were propane, vinyl chloride, vlnylldene chloride,
trlchlorotrlfluoroethane, chloroform, 1,2-dlchloroethane, methyl chloroform,
benzene, carbon tetrachloride, trlchloroethylene, l,3~dich1oropropene(cis),
i,3-d1chloropropene(trans), toluene, 1,2-dlbromoethane, tetrachloroethylene,
chlor-obenzene, o-xylene, benzyl chloride, and hexachlorobutadlene.
The following laboratory experiments were carried out:
(1)	A Perma-Pure dryer was tested to determine 1f water vapor could be
selectively removed from the gas stream also containing the 16
target compounds without affecting the Integrity of the organlcs 1n
the gas phase.
(2)	Collection and recovery efficiencies of the organic compounds were
determined.
(3)	Studies were also conducted with the target compounds to examine
potential Interference or artifact effects from co-collected ozone
and nitrogen dioxide gases.
(4} A prototype calibration device was Interfaced to the overall system
and tested. Software was developed to permit automatic operation
and calibration of the gas chromatography (GC) system.
(5)	S1de-by-s1de comparison of two prototype automated sampling and
analysis units was carried out at the EPA facility. Calibration
mixtures and ambient air samples were analyzed.
(6)	In a Joint effort with EPA, a comparison was made between
analytical results for samples collected and temporarily stored 1n
small metal cylinders, with data collected during real-time
sampling.
2

-------
SECTION 2
CONCLUSIONS
An automated sampling and gas chromatographic analysis system employing
reduced temperature preconcentratlon methodology has been tested 1n the
laboratory. During the laboratory experiments, two Naflon tube dryers were
evaluated and found to reduce water vapor selectively 1n the gas phase without
affecting the Integrity of trace organic species also present. The selective
reduction of water vapor Improves the chromatographic resolution of the trace
organ1cs and likewise permits larger sample volumes to be processed.
Experiments showed that collection and recovery efficiencies of sixteen
volatile organic compounds at low ppbv levels (.3 to 3) were 100 + 5 percent
with this preconcentratlon technique. Ozone and nitrogen dioxide interference
studies Indicated that none of the target compounds was affected by the
additional presence of these reactive species. Furthermore, no artifact peaks
or deleterious column effects were observed during or after these tests.
In a cooperative effort with EPA, two nominally identical automated GC
systems were Intercompared with samples, both calibration mixtures and ambient
air, drawn from a comnon manifold. Statistical analysis of the calibration
data Involved regressing concentration on raw area and determining the least
squares regression line. A percent relative error of less than ten was
obtained for fourteen of the sixteen compounds listed 1n Table 7 when
comparing the estimated concentration to the actual concentration. The
remaining two compounds, benzyl chloride and hexachlorobutadlene, exhibited
relative error values of *20 percent at the lowest non-zero concentration
level (4 ppbv). Simultaneous sampling of laboratory air from a common
manifold resulted 1n reasonable agreement between the two Instruments. Nine
target compounds were Identified and quantified. The results Indicated an
overall precision of ±10 percent.
In EPA conducted experiments, sixteen target compounds (see Table 11)
were stored at low ppb concentrations (2-5 ppbv) 1n passlvated stainless steel
containers and examined over a seven day period. Statistical treatment of the
data Indicated that after four days of storage, measured concentrations for
all sixteen compounds were within -*>8 percent of the Initial values. After
seven days the average benzyl chloride concentration decreased by 
-------
SECTION 3
RECOMMENDATIONS
A gas chromatographic system employing reduced temperature
preconcentration for the collection of VOC's has been tested with respect to
sample drying procedures, co-collection of reactive ambient air species, and
collection and release efficiency. An automated GC system has been used to
facilitate the laboratory tests.
The following recommendations are suggested:
t The automated GC system should be field tested. The two automated
systems should bo evaluated s1de-by-s1de during the first phase. of
the field program. Subsequently, the sample collection cycle would
be offset so as to permit complete time coverage with both
instruments. The cryogenic sampling system should also be compared
with other preconcentration techniques such as solid adsorbents
(Tenax, Carbosleve) and passive dosimeters. These tests would
provide much needed Information regarding the advantages and
disadvantages of each technque.
e Many of the target compounds tested In this study will co-elute with
other ambient air species. Although the combination of capillary
column, and flame Ionization and electron capture detectors used In
the current program help alleviate some of these concerns, other more
selective detectors are needed. Integrating a mass-selective
detector Into the automated gas chromatographic system Is
recommended. The mass-selective detector offers both total 1on and
selected 1on monitoring capability and will thus provide both
qualitative and quantitative Information. The Increased specificity
of this detector over other detection systems will allow better
differentiation of co-elutlng GC peaks and thus Improve present
quantitative capability.
4
-------
SECTION 4
EXPERIMENTAL
CHEMICALS
For the basic laboratory studies, target compounds were obtained from
the Eastman Kodak Company and the Baker Chemical Company. Compounds were ^98
percent pure (dlchloropropene was 95 percent pure). Liquid standards were
made by diluting the neat reagents with methanol. Compound response factors
were determined by injecting diluted liquid standards and comparing peak areas
to benzene standards. Gas mixtures were prepared by Injecting predetermined
amounts of each test compound Into a cylinder that had been partially
evacuated to Insure total vaporization of the standard. Once the mixture had
equilibrated, the cylinder was pressurized to 1200 pslg with "hydrocarbon
free" air. The gas mixtures were analyzed by GC-flame Ionization detection
techniques and compared on the basis of ppb carbon response to an NBS propane
standard.
During the latter phase of the program two pressurized cylinders from
Scott Environmental Technology* Inc. were acquired. These gas mixtures
contained the target compounds at nominal concentrations of 10 ppmv 1n
nitrogen. These mixtures were utilized In conjunction with the prototype
calibrator during Intercomparlson of two automated GC systems and during the
canister analysis study.
INSTRUMENTATION
The automated GC system used for sampling and analysis was composed of
two Instruments as shown 1n Figure 1. Sample collection was accomplished with
a modified Nutech model 320-01 cryogenic unit. The Nutech Instrument
contained two subsystems. An electronics console was employed for setting and
controlling various temperature zones, while the sample handling unit
contained the six-port valve and cryogenic trapping components. Three
temperature zones were controlled by the console and Included the valve, trap,
and transfer lines. The valve and transfer lines were maintained at 120 C.
During sample collection the trap temperature was regulated by the electronics
console, which controlled the release of liquid nitrogen via a solenoid valve.
For the current study the temperature setpolnt of the trap during sample
collection was -160 C ~ 5 C. A cylindrical 2S0-W heater was used to neat the
trap during the samplT desorption cycle. Direct thermal contact between the
heater and trap provided rapid heating (-160 C to 120 C 1n 60 sec) and
efficient transfer of the sample components onto the gas chromatographic
column. The sample trap was constructed of 20-cm by 0.2-cm l.d. stainless
steel tubing packed with sllanlzad glass beads (60/80 mesh). A Selscor six-
port a1r-actuated valve (Seismograph Service Corp., Tulsa, Oklahoma) was used
5
-------
Voltage
Sample
In Vent

Six-Port Valve
Gas Chromatographic System
Figure 1. Diagraa of automted sailing and analysis system.
-------
to facilitate sample collection and Injection. Figure 1 shows the GC valve
configured for collection of VOC's (I.e., SMple air Is diverted throuah the
trap and out the vent port). Switching the valve directs carrier gas through
the trap and Into the analytical column.
Sample analysis was achieved with a Hewlett-Packard model 5880A (Level
4) gas chromatograph equipped with flame Ionization and electron capture
detectors. A 50-m by 0.32-mi i.d., OV-1 fused silica column (Hewlett-Packard)
was used to resolve the target compounds. The column flow was maintained at a
linear velocity of 30 cm/sec (flow controlled system). Zero grade helium
(Matheson) served as the carrier gas. and zero grade nitrogen (Matheson) was
used as the make-up gas to provide proper electron capture operation. Optimum
analytical results were achieved by temperature programing the gas
chromatographic column from -SO C to 150 C at 80/mln. During the initial
laboratory tests the effluent from the SC column was directed either to the
flame Ionization or the electron capture detector. During the latter phase of
the program a tee was constructed (low dead-volume union with a two-hole
ferrule) and the column exit flow was split to both detectors. The canister
sampling and GC Intercomparlson studies were completed with the column 1n this
configuration. A more detailed description of the overall automated system 1s
reported elsewhere.*")
A gas-phase dynamic dilution system from Columbia Scientific Industries
was employed to generate ppb and ppt mixtures of the target compounds during
the basic laboratory studies. This unit was also used to produce various
concentrations of ozone and nitrogen dioxide. Zero air from an Aadco, Inc.
clean-air generator was employed for sample dilution. Internal mass flow
controllers provided accurate gas flows, which were verified with a soap-film
bubble meter. An EGliG model 911 hygrometer provided data on relative
humidity. A sling psychrometer provided a calibration check of the
hygrometer.
A Perma Pure dryer (model MD-125F) with a tubular hydroscopic Ion-
exchange membrane (Naflon) was used to remove water vapor selectively from
mixed gas streams. The dryer was purchased 1n a shell and tube configuration.
The tube size was 30 cm by 0.1 cm l.d. and was Imbedded within a shell of
Teflon tubing with 0.25 cm 1.d. Sample flow through the tube was maintained
at 60 mL/m1n. A countercurrent flow of dry zero air (200 mL/m1n) was used to
purge the shell. Literature data on the drying performance of this unit under
our experimental conditions show a product dew point of less than -5 C when
the Incoming wet stream has a dew point of 24 C.U4) Experiments to verify
these data were not conducted.
In order to Implement automatic calibration 1n the sampling and analysis
system, a prototype calibrator was provided by EPA. Following receipt of the
unit several modifications were made to better Implement the calibration
procedure. The following 1s a brief description of the Instrument. More
detailed Information 1s reported elsewhere.*1"
The equipment permits a tingle stage dilution of a calibration gas mix
with a compressed "zero" gas. Two flow controllers (Tylan Corp., Carson,
California) are used to achieve dilution ratios near lOOO:1. The zero gas
7
-------
f m?nually Pr®5®t; the calibration gas fles rate is controlled fros the
HP 5880 microprocessor by selecting one of four possible "fixed" volt&ges for
Its flow controller. Each control voltage Is supplied via a reed relay that
1s closed by the execution of a keyboard command. Figure 2 shows a diagram of
the flow scheme 1n the slnole stage dilution mode of operation. Solenoid
valves 9, 11 and 12 are used for other modes of calibration but are depicted
to illustrate the actual valve positions. The vent line downstream of the
glass mixing chamber prohibits accidental over-pressurlzatlon of the apparatus
during ambient air sampling.
PROCEDURE
Perma Pure Drver
The experimental apparatus shown In figure 3 was designed to determine
the effects of the Perma Pure dryer upon the Integrity of the target
compounds. An Implnger was used to humidify the sample. Mixtures Nos. 1 and
2 contained the 15 target compounds listed In Table 3 in concentrations
ranging from 20 to 250 ppb. A CSI dilution system was used to generate ppt
concentrations for those compounds responding to the electron capture
detector. When tests were conducted on those compounds responding to the
flame Ionization detector, the dilution system was bypassed and the sample was
fed directly to the Implnger unit. This modification was necessary to deliver
sufficient material to the flame Ionization detector. A five ml sample loop
was employed for these tests, with constant flow maintained by an
orlflce/rotameter/pump assembly. Several experiments were performed under dry
conditions (I.e., no water added to the Implnger). Subsequent runs were
carried out after the Implnger was filled with distilled water. The outlet
relative humidity was measured at the vent position with the hygrometer.
Collection and Recovery Efficiencies
Cryogenic collection and recovery efficiency studies of the target
compounds were carried out by replacing the five mL sample loop with the
freeze-out trap. An apparatus to measure sample volume was used 1n place of
the or1f1ce/pump assembly to determine more accurately the amount of air
processed through the trap (3). A container with a known volume was evacuated
and the absolute pressure recorded with a pressure gauge. A four-way ball
valve controlled the flow of the air Into this "fixed-volume" container. A
measured change 1n pressure within this tank corresponded to a specific volume
of air passing through the freeze-out trap and Into the container. Equation 1
was used to calculate the actual volume of sampled air;
AP T$td Vc
V» * P.td Tc	w
where V.	¦ actual volume sampled
AP	¦ change 1n pressure within the container
Tjtd	- 298 K
Vc	¦ volume of container as measured by water displacement
Pttd	¦ 1 atmosphere
Tc	• temperature of the container.
8
-------
To Sample InUrt of
Gas Chromatograph
Ambwnt Air
SV-11
Zavo
Am
Mass Flow
Controller
Vent
Mixing
ChamlMr
( ) SV-12
SV-8
8V-9
Figure 2. Diagram of flow scheme in prototype calibrator.
-------
Air Volum*
Mmluring Apparatm
(See Loop or
CryogwticTrap}
Column


Datactor


Figure 3. Experimental set-up for the Peraa-Pure dryer studies
-------
Interference Studies
Ozone and nitrogen dioxide Interference studies also made use of the CSI
dilution system. The unit Mas Modified so that the target mixture would enter
through one mass flow controller while nitric oxide gas simultaneously passed
through the other controller. With the unit 1n the gas-phase titration mode,
various ozone concentrations were generated and served to oxidize the incoming
nitric oxide to nitrogen dioxide. The nitrogen dioxide, excess ozone, and
target compounds were then directed through a mixing chamber and exhausted to
the gas chromatographic system for analysis. Ozone concentrations examined
were 0, 100, 200, and 300 ppb. Nitrogen dioxide concentrations remained at
250 ppb while target compound concentrations ranged from 0.9 ppb for
hexachlorobutadlene to 1/.3 ppb for trlchlorotrlfluoroethane. Gas
chromatographic sampling made use of the cryogenic trapping procedure and
sample volumes of 200 mL were processed with the volume-measuring apparatus
described earlier.
Automated System Intercomparlson
S1de-by-s1de comparisons of two essentially Identical prototype
automated sampling and analysis GC units were carried out using calibration
mixtures and ambient air samples. Calibration comparison studies made use of
the prototype calibrator to generate various concentrations of the target
compounds. During the Initial comparison, six record runs were performed for
GC No. 1 while seven record runs were carried out on GC No. 2. An additional
two record runs were completed 1n subsequent calibration tests. Each record
run produced five calibration points for each of the sixteen target compounds
listed 1n Table 7. The five calibration points include a zero concentration
and nominal concentrations of 5, 10, 20 and 50 ppb. Actual concentrations
were determined by measuring the calibration and dilution flows before and
after a series of record runs. Sample flow through the cryogenic trap was
maintained at 15 cc/mln with a mass flow controller. A collection time of 10
minutes provided sufficient sample. Compound analyses were accomplished with
flame Ionization detectors.
The second phase of the Intercomparlson focused on simultaneous sampling
from a common manifold with both GC units. During this testing period, three
successive runs of laboratory air were processed by both Instruments.
Additionally, canister samples of ambient air were analyzed with both units.
Figure 4 shows the configuration of the system during the Intercomparlson
studies. This figure shows three subunlts, the calibration unit, the sample
handling system, and the preconcentratlon and analysis system. The
calibration subunlt consists of a dynamic dilution system using pressurized
gas cylinders, electronic flow controllers (FC-L and FC-H) and valves. This
subunlt 1s explained 1n more detail in Reference 15. The sample handling
system consists of a manifold with provision for obtaining either calibration
air, spiked ambient or smblent air. The Perma-Pure dryer 1s also a part of
this subunlt. The third subunlt (preconcentratlon and analysis system) 1s
essentially the same as shown 1n Figure 1.
11
-------
Calibration Gas
Cylinder
V8. 9.10 = Solenoid Valves
NV = Needle Valve
FC-L = Flow Controfler. 0-10 teem
FC-H = Flow Coirtiolai, 0-2000 aeon
Cililn fion Marwow
Parma-Pure
N« Vent

Lam Volume Fitting


To ECO
To FID
Analytical 6C
Column Ovan
TytanFiow
Controllers (FC-260)
Figure 4. Configuration of automated GC system during
1nterconpar1son studies.
-------
Sample Integrity Study with Stainless Steel Canisters
During the program the storage characteristics were determined for
sixteen volatile organlcs (listed 1n Table 7) In stainless steel canisters
(Demaray Scientific Instrument, Ltd., Pullman, HA). The system used to fill
and sample from the canisters 1s shown 1n Figures 5 and 6. Figure 5 depicts
the configuration employed during the canister filling cycle while Figure 6
Illustrates the scheme utilized during the canister analysis cycle.
Prior to filling the canisters with the target mixture, the above
assembly was checked to Insure that the system did not contribute contaminant
peaks to the sample. Zero air was sampled directly from the manifold and
compared with zero air passing through the canister filling station.
Seven canisters were pressurized with a prepared gaseous mixture (16
organlcs) at a nominal concentration of 2 ppb. Air samples were taken from
each canister at 0, 2, 4 and 7 days after filling. During the initial filling
period nine air samples were also taken directly from the manifold system and
analyzed 1n real time. For discussion purposes, these real time samples will
be referred to as control samples.
13
-------
Manifold
S.S.
Vent
Figure 5. Canister apparatus during sample fill cycle.
-------
6-Liter
Sample
. Canister
/ PressureA
1 Gauge J
Mass Flow
Controller
Vent
Vent
s.s. Metal Bellows

Mass Flow
Pump

VOntfOMr
Carrier In
Parma-Pure Dryer
Analytical
Column
Cryogenic Trap
Figure 6. Canister apparatus during sample analysis cycle.
-------
SECTION 5
RESULTS AND DISCUSSION
BASIC LABORATORY STUDIES
Figure 7 shows several calibration mixture chromatograms that were
processed with the automated sampling and analysis system. Table 1 depicts
the mean concentration and standard deviation for each of the eight elutlng
compounds. As the data Indicate, excellent precision was obtained.
Furthermore, the sampling process was easily Initiated and terminated by
single keystroke commands. No additional operator Involvement was vieeded.
The above operations were carried out with a dry calibration mixture.
Additional experiments under humidified conditions showed that water vapor had
a significant effect upon the chromatography of the sample components. Table
2 shows the data from a test mixture that was processed under dry and
humidified conditions. Excellent retention time reproducibility was obtained
for those compounds when the sampling stream was dry. However, when water
vapor was also present, standard deviations of compound retention times were
significantly greater. When gas chromatography 1s the only Instrumentation
available for analysis, this retention time variability makes positive peak
identification extremely difficult.
The trend of Increased variability with shorter retention times 1n
Table 2 1s typical when cryogenlcally trapped compounds including water are
transferred from the trap to a capillary column held at reduced temperature.
The column 1s blocked temporarily by 1ce until the column temperature 1s
Increased. The time delay Interval 1s a smaller fraction of the retention
time for compounds elutlng last. Therefore, the erratic nature of the
blocking process causes Increased variability 1n retention time for those
compounds elutlng early 1n the chromatographic run.
Tables 3 to 5 show the results of tests with the Perma Pure dryer.
These tests were carried out with two prototype GC systems (Battelle and EPA
systems). Both units were used to test a common subset of target compounds
although not all compounds were common to both tests (Tables 3 and 4). The
Battelle results (Table 3) Indicate that for a humidified sample there Is no
significant difference between results with and without the dryer. The same
can be said for a dry sample with the exception of benzylchlorlde. However,
dry sample results were significantly lower than humid sample results for at
least three gases: l,3-d1ch1oropropene(c1s), l,2-d1bromoethane, and
benzylchlorlde. To Investigate this effect further and to obtain a measure of
variability for replicate analytical runs, additional tests were performed at
EPA. The EPA results shown 1n Table 4 verify the Battelle results but show no
16
-------
Run *1
J	Llll
Run #2
i
Run #3
Run #4
Jl
H

12 minutM
SSmtmitM
Figure 7. Automated chromatographic analyses of a standard mixture using
crvoqenlc prcconcentratlon techniques and flame Ionization detection
(100 mL sample).
17
-------
TABLE 1. STATISTICAL ANALYSES OF CHROMATOGRAMS SHOWN IN
FIGURE 7 (TEST 1) AND ADDITIONAL ANALYSIS AT
LOWER CONCENTRATIONS (TEST 2)
Compound
Test 1	Test 2
(4 runs. 200-mL samples)	(8 runs. 200-mL samples)
Mean CoeMdent	Mean Coefficient
(ppbv) of variation	(ppbv) of variation
(*)	W
Vinyl1dene chloride
144
4.4
6.1
3.9
Chloroform
84
3.8
3.5
5.8
1,2-D1chloroethane
44
3.7
1.9
5.1
Methylchloroform
63
4.5
2.7
4.9
Benzene
93
4.0
3.9
5.1
Trlchloroethylene
84
3.7
3.5
4.1
Tetr ach1oroethy1ene
G9
3.7
2.9
4.3
Chlorobenzene
46
3.3
1.9
3.2
TABLE 2. RETENTION TIMES FOR SEVERAL COMPOUNDS
CRYOGENICALLY COLLECTED AND ANALYZED
UNDER DRY AND HUMIDIFIED CONDITIONS®
Compound
Retention Time. M1nutesb
UryC— 	Aumldfled*
Chloroform
11.61
+
0.01
11.82
+
0.28
Methylchloroform
12.50
+
0.01
12.66
+
0.22
Carbon tetrachloride
13.09
+
0.00
13.22
+
0.18
Trlchloroethylene
14.12
+
0.01
14.21
+
0.12
Tetr ach1oroethy1ene
17.27
+
0.00
17.30

0.04
detection,
b Average + standard deviation,
c Average of three runs at <100 ppm water vapor,
d Average of four runs at ?7X RH and 27 C.
18
-------
TABLE 3. SAMPLE INTEGRITY RESULTS WITH A PERMA PORE DRYER FOR
THOSE TARGET COMPOUNDS RESPONDING TO A FLAME IONIZATION
DETECTOR* -BATTELLE RESULTS
Compound11
Drv Sample^
Through Dryer
Dryer Bypassed
Humidified Sampled
Through Dryer
Dryer Bypassed
Tr1chlorotr1fluoroethane	259 + 1	259 + 1	255 +	1	253 + 1
Chloroform	79 + 1	79 + 1	80 +	2	80 + 2
1.2-D1chloroethane	50 + 4	50 + 3	54 +	1	56 +
Methyl chloroform	61 + 1	62 + 1	63 +	1	60 +
Benzene	92 + 2	90 + 2	93 +	1	95 +
Carbon tetrachloride	126 + 1	124 + 1	123 +	1	122 +
Trlchloroethylene	82 + 1	81 + 1	80 + 1	82 +
1.3-D1ch1oropropene	(c1s)	175 + 5	166 + 6	203 +	3	199 +
Toluene	245 + 1	243 + 2	251 +	3	248 +
l,2-D1bromoethane	124 + 3	116 + 3	153 +	5	148 +
Tetrachloroethylene	67 + 1	67 + 1	67 +	1	67 +
Chlorobenzene	46 + 2	39 + 3	49 +	1	51 +
o-Xylene	146 + 2	145 + 2	151 +	2	149 +
Benzylchlorlde	17 + 2	10 + 2	58 +	4	56 +
Hexachlorobutadlene	26 + 1	25 + 1	29 +	1	31 +
a Each concentration reported represents the average m ppb for two
runs; a five mL sample loop was used,
b Vlnylldene chloride was	not measured	In this experiment because of
excessive peak broadening due to the	size of the sample loop,
c <100 ppm water vapor,
d BZ% RH at 23 C.
19
-------
TABLE 4. SAMPLE INTEGRITY RESULTS WITH A PF.RMA PURE DRYER FOR
COMPOUNDS RESPONDING TO A FLAME IONIZATION DETECTOR®
-EPA RESULTS
Drv Samoleb
Through 0
Compound
Dryer
ryer
Bypassed
Humidified Sample^
Through Dryerd
Vinyl chloride

1076
+
32
1106 +
8
1063
+
11
D*ch1oroethy1ene

1085
7
14
1096 +
7
1118
7
73
Tr1ch1orotr1f1uoroethane
1019
7
8
1015 ~
7
1038
7
18
Chloroform

391
7
4
391 7
4
391
+
5
l,2-D1ch1oroethane

1003
7
19
1025 7
11
1026
7
19
Methylchloroform

1368
7
6
1348 7
16
1367
+
10
Benzene

2816
7
23
2943 +
20
2748
+
21
Carbon tetrachloride

331
7
3
332 +
2
340
7
5
Trlchloroethylene

1184
7
10
1182 7
9
1192
7
5
1,3-Dlchloropropene (
!c1s)
854
7
14
870 ~
8
876
7
6
l,3-01chloropropene i
trans)
683
7
24
693 +
17
728
7
4
Toluene
3605
7
38
3657 7
22
3507
7
36
l,2-D1bromoethane

1059
7
24
1070 7
10
1083
7
6
Tetr ach1oroethy1ene

1247
7
9
U51 7
15
1260
7
10
Chlorobenzene

3393
7
31
3353 7
30
3408
7
16
Benzylchlorlde

3062
7
439
3046 7
223
3435
7
62
Hexachlorobutadlne

1992
7
162
1805 7
61
1932
7
83
o-Xylene

4136
7
36
4185 7
33
4042
+
24
samples, 12 for humidified samples. Cryogenic preconceritratlon of
938 mL, concentration (~ standard deviation) measured 1n area counts,
b Less than 0.2* water vapor In air, room temperature approximately 25 C.
c 1.9 + 0.1X water vapor 1n air, room temperature approximately 25 C.
d No humidified samples were analyzed without the dryer because water
plugs the capillary column.
20
-------
TABLE 5. SAMPLE INTEGRITY RESULTS WITH A PERMA PURE DRYER FOR
COMPOUNOS RESPONDING TO AN ELECTRON CAPTURE DETECTOR®
Dry Samoleb	Humidified Samolec

Through
Dryer
Through
Dryer
Compound
Dryer
Bypassed
Dryer
Biassed
Tr1ch1orotr1f1uoroethane
375
345
376 + 60
378 + 40
Chloroform
213
213
215 7 3
?19 7 3
Methylchloroform
290
287
251 7 13
274 + 9
Carbon tetrachloride
660
660
630 7 21
655 7 28
Trlchloroethylened
m
m


1,3-D1ch1oropropene (c1s)
900
900
933 + 35
950 + 35
l,2-D1bromoethane
640
659
640 7 13
653 7 6
T etrach1oroethy1ene
318
318
316 7 7
313 7 14
Hexachlorobutadlene
130
145
140 7 3
138 7 7





a a single analysis was made during the dry sample experiment;
duplicate analyses were performed during the humidified experiments.
A five mL sample loop was used and concentration 1s 1n pptv.
b 100 ppm water vapor,
c 80% RH at 25 C.
d Contamination 1n stream selecting valve.
ZX
-------
evidence of a difference between results for dry and humid samples. All but
one of the standard deviation values are less than ten percent of the mean and
most are less than five percent of their respective means. Benzyl chloride
shows the highest standard deviation values. One reason for this variability
was obvious when the analytical results were plotted In sequence. Significant
delays 1n establishing equilibrium concentration levels were observed for some
of the compounds 1n changing from one type of test to another, I.e., from a
dry sample to a humidified sample. These effects probably are the reason for
the differences between wet and dry sample test results 1n the data 1n
Table 3.
Table 5 presents the results of experiments 1n which the CSI dilution
system was used to generate ppt levels of the target compounds and an electron
capture detector was employed for analysis. No significant difference was
observed between the humidified and dry streams and again the overall data
Indicated the Perma Pure dryer did not affect the Integrity of the target
compounds.
Cryogenic collection and recovery efficiency studies of the target
compounds were carried out by comparing the peak area response values obtained
by the direct sar?le loop Injections to the values found by cryogenic
preconcentratlon of dilute sample streams. The target compounds that were
examined, along with their concentrations, are listed 1n Table 3. Humidified
sample air was passed through the Perma Pure dryer, the 5 mL loop and then
analyzed by gas chromatographic^lame Ionization detection techniques. The
same air was subsequently diluted 1/100 fold with humidified zero air, passed
through the dryer, directed to a reduced-temperature trap (200 mL) and
analyzed. When the dilution stream (1/100 fold) and concentration effects
from the cryogenic collection procedure (200-mL samples) were accounted for,
recovery efficiencies of 100 + 5 percent were obtained for all the compounds.
This recovery Is within the "estimated experimental error of 110 percent and
consisted primarily of errors In determining gas dilution and sample volumes.
The results from the four ozone and nitrogen dioxide Interference tests
are depicted 1n Table 6. Test 1 was conducted without the addition of ozone
and nitrogen dioxide. Ourlng tests 2 to 4 ozone concentrations were 100, 200,
or 300 ppb and the nitrogen dioxide level was maintained at 250 ppb. As the
data indicate, none of the target compounds was affected by the presence of
ozone or nitrogen dioxide. Furthermore, no artifact peaks or deleterious
column effects were observed during or subsequent to these tests. Although
the above experiments were certainly not exhaustive, the generated
concentrations were construed to be realistic atmospheric levels. The results
allowed us to conclude that the cryogenic trapping techniques and
Instrumentation employed In this study did not significantly affect the sample
Integrity of the target compounds. An effective detection limit of at least
0.1 ppb per compound was obseved during these tests, based on a 200-mL sample
and flame Ionization detection. Lower detection limits could be achieved for
many of these compounds by employing more sensitive detectors such as the
electron capture or photo1on1zat1on detection system.
22
-------
TABLE 6. OZONE AND NITROGEN DIOXIDE INTERFERENCE RESULTS USING
FLAME IONIZATION DETECTION®"
Concentration In Given Test, oobv
Compound	"T ! ? 3	4
Ozone
0
100
200
300
Nitrogen Dioxide
0
250
250
250
V1nyl1dene Chloride
8.7
8.8
8.8
9.1
Trlchlorotrlfluorethane
17.3
17.0
17.0
16.8
Chloroform
4.8
5.1
5.1
5.2
l,2-D1chloroethane
3.1
3.2
3.1
3.2
Methylchloroform
3.8
3.8
3.8
4.0
Benzene
5.7
5.7
5.7
5.9
Carbon tetrachloride
7.8
7.9
8.0
7.9
Trlchloroethylene
4.8
5.0
5.0
5.2
l,3-D1ch1oropropent
9.7
9.4
9.7
8.7
Toluene
14.2
14.1
14.5
14.2
l,2-D1bromoethane
6.7
6.9
7.1
7.2
Tetrachloroethylene
4.1
4.1
4.1
4.3
Chlorobenzene
2.9
3.0
3.1
3.2
o-Xylene
7.2
7.5
7.8
7.5
Benzylchlorlde
1.3
1.3
1.4
1.4
Hexachlorobutadlene
0.9
0.9
1.0
1.1
4 200 mL samples with cryogenic trapping.
23
-------
Software Development
Following the basic laboratory studies, research efforts primarily
focused on develop^ software to carry out the monitoring and calibration
needs of the automated GC sampling and analysis system. In addition to
facilitating zone heating and cooling, sample collactlon and Injection,
chromatographic run conditions and data processing, the basic software
permitted three essential modes of GC operation:
e calibration
e ambient air
e calibration and ambient air.
In the calibration mode of operation either single or multipoint calibrations
can be carried out. The single point subroutine results 1n one zero air run
plus one or more runs (up to 10) of a single dilution mixture. The multipoint
calibration subroutine Includes one zero air run followed by runs of four
different dilution mixtures. After the completion of either calibration
subroutine, a linear regression analysis Is performed on each compound of
Interest. A linear regression equation along with the correlation coefficient
and the number of analyses are given as output parameters for each compound.
However following the Initial use of the program, the correlation coefficient
was found to be a poor Indicator of calibration linearity. In future work, 1t
will be replaced by some other parameter such as the mean square error. A
printout of the current program 1s given In Appendix A.
Presently, the system's calibration table 1s not accessible from the
BASIC program (to do this would require knowledge of the HP operating system).
As a result, once the calibration procedure Is completed, the appropriate
compound response factors are manually Inputted to the GC's existing
calibration table In preparation for ambient air sampling.
The ambient air mode of operation can be conducted v?1th or without
calibration (operator controlled). After each run 1s completed (ambient
and/or calibration), a SC report Is printed and then transferred to a cassette
for future use. This transfer process conserves space 1n the system memory so
that an unlimited number of ambient air samples can be processed.
With the current program the operator does not have to be present to
start sample collection. When the mode of Instrument operation 1s
Initialized, the operator simply specifies the start time needed (24 hour
clock). The start time option can also be easily modified so as to delay
initiation of sample collection by one or more days. Likewise with the
current software, the primary GC functions (I.e., sample collection and
analysis) take place sequentially* In order to provide for a more extended
time coverage during real-time field sampling, the program can be modified so
that sample collection for the "next" run Is Initiated during the analysis uf
the current run* By Incorporating this modification along with a second
automated GC system, one can sample ambient air with essentially complete time
coverage.
24
-------
Automated System Intercomoarlson
S1de-by-$1de comparison of two prototype automated 6C units were carried
out using calibration mixtures and ambient air samples. During the Initial
calibration studies* results gave non-linear calibration curves from both GC
units. Since the GC units were expected to perform In a linear fashion (when
using a flame Ionization detector) we suspected that the prototype calibrator
might be malfunctioning. Several modifications were subsequently made to the
calibrator unit and closer scrutiny of the actual calibration and dilution
flow measurements were carried out during the second series of tests.
The statistical treatment of the data from the second calibration period
was as follows. The five calibration points from each record run were used to
form a calibration line. The calibration line was found by regressing con-
centration (C) on raw area (RA) and determining the least squares regression
line. For a particular record run, let a and b denote the Intercept and
slope, respectively, of the calibration line. For each of the calibration
points, an estimated concentration, C, was obtained from the formula
t ¦ a + b (RAJ
The process of forming a calibration line and calculating estimated
concentrations was repeated for each record run. Thus each calibration point
(C, RA) has associated with It an estimated concentration C and an error in
estimation
e »e - c
The error 1n estimation, E, Is the absolute error. Also of Interest 1s the
relative error 1n estimation
RE ¦ (6 - C)/C ¦ E/C .
Table 7 lists the 16 target compounds that were analyzed 1n the above
fashion. Actual data from 1,2-dlchloroethane (flame Ionization detector) are
depicted 1n Tables 8 and 9. A percent relative error of less than 10 was
obtained during both calibration runs. Similar relative errors were also
obtained with both GC units for 13 of the remaining 15 compounds. The two
exceptions, benzyl chloride and hexachlorobutadlene, exhibited RE values of
^20 percent at the lowest non-zero concentration level (*4 ppbV).
The above results Indicate good linear behavior of both GC systems
(using flame Ionization detection) and demonstrate the capability of these
instruments to accurately determine low ppb concentrations of ambient air
species. However 1t 1s felt that even better detection levels can be
achieved. From Tables 8 and 9 Intercepts of 0.54 ppb and 0.75 ppb are
observed for both calibration runs from Instrument No. 1 and Intercept values
of 0.35 ppb and 0.57 ppb are obtained for the second GC unit. Both
Instruments behave similarly during each calibration run when viewing the
percent relative errors at each concentration. Based on the actual signal to
noise ratios of *30/1 at the 4 ppb level, lower Intercept values were
expected. Flow measurement errors m*y account for the higher than expected
Intercepts; likewise partial sample adsorption within the calibrator Itself
25
-------
TABLE 7. LIST OF TARGET COMPOUNDS EXAMINED DURING
CALIBRATION RUNS
Propane
Vinyl Chloride
Vlnylldene Chloride
Tr1ch1orotr1f1uoroeth ane
Chloroform
l,2-D1chloroethane
Methyl Chloroform
Carbon Tetrachloride
Trlchloroethylene
1,3-Dlchloropropene (c1s)
l,3rD1chloropropene (trans)
1,2-Dlbromoethane
Tetrachloroethylene
Chlorobenzene
Benzyl Chloride
Hexachlorobutadlene
TABLE 8. DATA OBTAINED FROM THE STATISTICAL ANALYSIS OF A
MULTIPOINT CALIBRATION RUN (1(2-DICHLOROETHANE)
WITH GC NO. 1
Regression
Calibration
Cone
Raw
Estimated

X Relative
Equation
Run No.
PPb
Area
Cone ppb
Error
Error
(C)-0.5395+
1
0.00
0.00
0.54
-0.54

0.8769(RA)
1
4.04
3.61
3.71
0.33
8.2

1
9.90
10.57
9.81
0.09
0.9

1
19.80
21.68
19.55
0.25
1.3

1
49.30
55.76
49.44
-0.14
-0.3
(O-0.7470+
2
0.00
0.00
.75
-0.75
m
0.9104(RA)
2
3.88
3.72
4.13
-0.25
-6.4
2
10.10
9.36
9.27
0.83
8.2

2
20.00
20.63
19*53
0.47
2.4

2
50.40
54.87
50.70
-0.30
-0.6
26
-------
TABLE 9. DATA OBTAINED FROM THE STATISTICAL ANALYSIS OF A
MULTIPOINT CALIBRATION RUN (1.2-DICHLOROETHANE)
WITH GC NO. 2
Regression
Calibration
Cone
Raw
Estimated

£ Relative
Equation
Run No.
ppb
Area
Cone ppb
Error
Error
(C)»0.3519+
1
0.00
0.00
0.35
-0.35

0.5604(RA)
1
4.04
5.91
3.66
0.30
9.30
1
9.90
17.14
9.96
-0.06
-0.58

1
19.80
34.56
19.72
0.08
0.40

1
49.30
87.42
49.35
-0.05
-0.09
(C)"0.5663+
2
0.00
0.00
0.57
-0.57

0.5786(RA)
2
3.88
5.97
4.02
-0.14
-3.62
2
10.10
14.98
9.23
0.86
8.57

2
20.00
33.64
20.03
-0.03
-0.15

2
50.40
86.35
50.53
-0.13
-0.26
27
-------
may also contribute to the "non-zero" Intercept; In many cases, a better fit
of the calibration data was actually obtained with a quadratic curve.
Unfortunately, the SC system's calibration table only permits the entry of a
slope value (compound response factor). Efforts to modify the system's
calibration table to accept non-11near curves were beyond the scope of this
contract.
Automated System Intercomparlson Using Laboratory Air
The second phase of the Intercomparlson focused on simultaneous sampling
from a common manifold with both GC units. One of the data sets Is shown 1n
Table 10. Eight of the target compounds were Identified srsd qusr.t itated with
both GCs. The chlorinated species were measured with electron capture
detection while the aromatic species were determined by flame Ionization
detection. As the data Indicate, reasonable agreement between the two GC
units was obtained. Deviations from the mean for the 24 comparisons ranged
from 0.0 to 20.5 percent.
Statistical Analysis of Canister Data
The purpose of the statistical analysis of the canister data was to
evaluate the stability of samples stored 1n canisters and consequently to
compare the data from the samples stored In canisters to data from control
samples. Table 7 lists the sixteen compounds that were statistically treated
during the canister tests. (Appendix B lists the raw area values for the
canister samples). Sample means and standard deviations for the canister data
are reported by compound 1n Table 11. Each of the SAMEDAY, DAY2, DAY4, and
DAY7 means represent the average of seven raw areas, one from each of seven
canisters (except 1n the cases noted). Sample standard deviations are
reported 1n parentheses.
The primary quantities of Interest for the canister data are the percent
relative biases (PRB) Introduced by storing samples In the canisters. For
each of days 2, 4, and 7, the percent relative bias for day 1 1s defined as
pad, . f (True average raw area for canister samples on day 1) ,1 10Q
no< L(True average raw area for samples on day 0)	J
Estimates of the percent relative biases are reported In Table 12 with 99
percent confidence Intervals reported 1n parentheses. A confidence level of
99 percent was chosen because of the Urge number of confidence Intervals
being formed. If the confidence Interval for a PRB does not contain zero,
this Is an Indication of a bias that Is significantly different from zero.
To arrive at the confidence Intervals 1n Table 12, the data for each
compound were analyzed using a randomized complete block model (canisters
represent blocks and sampling time represents the treatments). Dunnett's
procedure was used to obtain simultaneous 99 percent confidence Intervals for
the true average raw area difference between days 2, 4, and 7 and day 0 (16).
The endpolnts of these Intervals were then divided by the SAMEDAY mean and
multiplied by 100 to arrive at the confidence Intervals In Table 12.
28
-------
TABLE 10. COMPARISON OF TWO AUTOMATED GC UNITS SAMPLING SIMULTANEOUSLY
FROM A COMMON MANIFOLD3


Run 1


Run 2


Run 3

GC 1
GC 2
VP
tt i
GC 2
XD
GC 1
GC 2
0
Trlehlorotrlfluoroethane
0.16
0.18
5.9
0.24
0.24
0.0
0.23
0.24
2.1
ttlmofara
0.15
0.20
14.3
0.43
0.48
5.5
0.57
0.60
2.6
dethylchlorofoni
0.49
0.56
6.7
0.71
0.75
2.7
0.64
0.66
1.5
Carbon tetrachloride
0.15
0.16
3.2
0.21
0.19
5.0
0.20
0.18
S.3
Trlchloroethylene
0.29
0.44
20.5
0.64
0.75
7.9
0.76
0.68
5.6
Tetrachloroethylene
1.25
1.47
8.1
4.79
5.37
5.7
4.63
5.00
3.8
•emene
0.72
c
-
0.77
c
-
0.91
c
-
Tolneae
3.40
3.49
1.3
7.17
7.56
2.6
9.29
9.52
1.3
0-*y1eae
1.39
1.15
5.5
4.16
3.68
6.1
4.30
3.41
11.5
• 0.3 Liters of 1 Moratory air collected with cryogenic preconcentration techniques (23 C, 40 t RH).
b Percent deviation from the new.
c Indicates a co-elttlng Interferant.
-------
TMLE It. SAMPLE MEANS AND STANDARD DEVIATIONS FOR CANISTER
SAMPLE RAH AREA VALUES
Compound
(S.D.)
Day 2
Mean
(S.D.)
Day 4
Mean
(S.D.)
D*y 7
Mean
(S.D.)
Propane
Vinyl Chloride
Vlnylidene Chloride
Trichlorotrifluoroethaae
Chloroform
1.2	Dichloroethane
Methyl Chlorofom
Carbon Tetrachloride
Trlchloroettylene*
1.3	D1 chloropropene (cfcr
1,3 Dichloropropene (f
1,2 Oibroaoethane*
Tetrachloroethylene
Chlorobenzene
Benzyl Chloride
Hexachlorobutadi enea
2.73
1.83
2.07
1.93
1.93
1.86
1.93
2.01
1.78
0.82
0.64
1.36
1.58
1.39
1.01
1.25
(0.090)
(0.075)
(0.040)
(0.025
(0.086
(0.070
(0.037
(0.025
(0.077
0.035!
[0.036;
[0.090;
0.059
[0.106;
0.102:
0.319
2.75
1.85
2.16
1.97
2.01
1.89
1.97
2.04
1.83
0.77
0.63
1.37
1.62
1.43
0.94
1.25
(0.073
(0.040
(0.019
(0.051
(0.071
(0.040
(0.022
(0.075
(0.045
(0.041
(0.075
(0.091
(0.042
(0.078;
(0.049;
(0.232;
2.73
1.81
2.14
1.95
1.99
1.86
1.96
2.09
1.84
0.75
0.63
1.35
1.61
1.41
0.94
1.29
(0.042)
(0.053)
(0.034)
(0.045)
(0.046)
(0.030)
(0.016)
(0.046)
(0.049)
(0.064)
(0.061)
(0.080)
(0.047)
(0.073)
(0.376)
(0.313)
2.78
1.89
2.16
1.98
1.96
1.87
1.95
2.00
1.81
0.76
0.63
1.38
1.63
1.43
0.82
1.39
(0.014)
(0.119)
(0.008)
(0.041)
(0.047)
(0.045)
(0.020)
(0.033)
0.027)
0.030)
[0.061)
0.091)
0.035)
[0.087)
0.109)
[0.417)
a Reported values based no data fnn six canisters,
b lujKMrted values based m data from five canisters.
-------
TABLE 12. POINT ESTIMATES AND 99 PERCENT CONFIDENCE INTERVALS
FOR THE PERCENT RELATIVE BIAS ON DAYS 2, 4, AND 7
D*y 2	Oty 4	Oty 7
Percent	Percent	Percent
Relative	Relative	Relative
Coapound Bias	(C.l.)c Bias	(C.I.) Bias	(C.l.)
Prop we
Vinyl Chloride
Viaylldcue Chloride
TricJilorotriflMoroethane
Chloroform
1.2	Olchloroetliie
Nrttyl Chlorotfona
Carton Tetrachloride
Trlchloroethylene*
1.3	Dlchloroprapene (cis}b
1.3 Dlchloropropene (trans)*
1.2 Olbroaoetfime*
Tetrachloroethylene
Chlorobenzene
Benzyl Chloride
Mexaeh lorobutsdieae*
0.7 (-3.4, 4.9]
0.0
1.1 (-6.0, 8.2
-1.1
4.3
[ 1.7, 7.0]
3.4
2.1
[-1.0, 5.1
1.0
4.1
! 0.6, 7.7
3.1
1.6
-1.5. 4.7
0.0
2.1
[-0.2, 4.3
1.6
1.5
[-2.9, 5.9,
4.0
2.8
[-2.1, 7.7
3.4
-6.1 (-13.2, 1.0]
-8.5
-1.6 (-14.2,11.1
-1.6
0.7
-2.9, 4.3
-0.7
2.5
-0.3, 5.4
1.9
2.9
0.1, 5.6
1.4
-6.9 (-
-41.2,27.3
-6.9
0.0 (•
-17.8,17.8
3.2
(-4.1. 4.1)
(-8.2. 6.0)
( 0.7. 6.0)
(-2.0, 4.1)
(-0.5, 6.7)
(-3.1, 3-1)
(-0.7, 3.8)
(-0.4. 8.4)
(-1.5, 8.3)
(-15.6,-1.5)
(-14.2,11.1)
i-4.3, 2.9)
0.9, 4.7)
-1.3, 4.2)
(-41.2,27.3)
(-14,6,21.0)
1.8
3.3
4.3
2.6
1.6
0.5
1.0
-0.5
1.7
-7.3
-1.6
1.5
3.2
2.9
•18.8
11.2
i-2.3, 6.0
-3.8,10.4;
1.7, 7.0
-0.5, 5.7
(-2.0, 5.1
(-2.6, 3.6
j-1.2.
(-4.9,
(-3.2, 6.6
(-14.4, -0.3
(-14.2,11.1
(-2.1. 5.1
0.3, 6.0
0.1, 5.6
(-53.1,15.5
(-6.6,29.0!
3.3
3.9
a Reported values based on data froa six canisters,
b Reported values based on data froa five canisters,
c 99 percent confidence interval.
-------
Inspection of Table 12 shows that all the compounds, excluding vlnylidene
chloride, have no significant bias. Although the confidence intervals for
vinylidence chloride do Indicate a slight positive bias, the relative biases
are less than five percent for this compound. For all sixteen compounds
tested, concentrations after four days of storage were within +8 percent of
the Initial values. However, after storage for seven days the average benzyl
chloride concentration decreased by approximately 19 percent while
hexachlorobutadlence Increased by approximately 11 percent. If the above
results using calibration mixtures are extrapolated to ambient air samples,
the passlvated canisters should serve as acceptable temporary storage devices
for the above target compounds.
For the purpose of comparing the canister data with the control sample
data, the canister data was pooled across sampling times. The overall
canister mean, standard deviation, and relative standard deviation (standard
deviation times 100 divided by mean) and the corresponding values for the
control samples are reported 1n Table 13. Each CONTROL mean and standard
deviation 1s based on nine raw area values.
An estimated percent relative bias (PRB) Is calculated for each compound
using the formula
Canister Mean
_ r tamsier nean i * ..
m " [ Control Mean - 1 J * 100
A 99 percent confidence interval for the PRB was calculated by first obtaining
a confidence Interval for the true average raw area difference (CANISTER-
CONTROL) using standard statistical techniques. The endpolnts were then
divided by the CONTROL mean and multiplied by 100 to arrive at the confidence
Intervals In Table 13.
Examining the PRB estimates and confidence Intervals 1n Table 13 shows a
general tendency towards a positive bias. That 1s, on the average, the raw
area,values for canister samples tend to be larger than the raw area values
for the control samples. For fourteen of the sixteen compounds tested, the
positive bias ranges from 1 to 8 percent. Benzyl chloride and hexa-
chlorobutadlene exhibit relative positive biases of 36 and 72 percent
respectively. The slight positive bias observed for the fourteen compounds 1s
not significant and may have been due to an error 1n determining sample
volumes during the analysis of the control samples. The abnormally high
positive bias of benzyl chloride and hexachlorobutadlene Is most likely due to
the overall difficulty in analyzing these two compounds because of their
strong surface adsorptlve characteristics.
32
-------
TABLE 13. SUMMARY TABLE FOR COMPARISON OF CONTROL DATA WITH COMBINED CANISTER DATA
Control
Mean
(S.O.) (R.S.D.)c
Canister
Mean
(S.D.)
Percent
Relative
(R.S.O.) Bias
(C.I.)d
¥1*yl Chloride
Vlnylldene Chloride
Trictilorotrlfluoroethane
Chlorofora
1.2	Olchloroethane
Methyl Chlorofora
Carbon Tetrachloride
Trichloroethylene*
1.3	Olchloropropene (c1s)b
1,3 Olchloropropene (trans)"
1,2 Dlbraaoethane*
Tetrachloroethylene
Chlorobenzene
Benzyl Chloride
Hexachlorobutadlene*
2.63
1.76
2.03
1.81
1.87
1.77
1.88
1.96
1.72
0.76
0.62
1.27
1.50
1.31
0.68
0.75
(0.150)
(0.088)
10.105)
0.111)
0.146)
(0.148)
(0.100)
(0.098)
(0.128)
(0.092)
(0.087)
(0.142)
(0.140)
(0.142)
(0.140)
(0.121)
a	Canister data based on six canisters,
b	Canister data based on five canisters,
c	Relative standard deviation (percent),
d	99 percent confidence Interval.
(5.7)
(5.0)
(5.2
(6.1)
(7.8)
(8.4)
!5.3)
5.0)
7.4)
(12.1)
(14.0)
(11.2)
(9.3)
(10.8)
(20.6)
(16.1)
2.75
1.85
2.13
1.96
1.97
1.87
1.95
2.04
1.82
0.78
0.63
1.36
1.61
1.42
0.93
1.29
(0.062)
(0.080)
(0.046)
(0.044)
(0.068)
(0.047)
(0.028)
(0.0581
(0.054)
(0.049)
(0.055)
(0.083)
(0.048)
(0.084)
(0.204)
(0.310)
-0.6, 9.5)
0.1, 9.8)
0.5, 9.6)
2.8,13.6)
-1.5,12.4)
-1.7,12.9)
-0.8, 8.5)
-0.7, 8.4)
-1.1,12.1)
-9.0,13.2)
(-12.1,15.5)
-2.8,17.6)
-1.0,15.4)
-1.6,18.1)
13.1.59.1)
46.6.98.2)
-------
REFERENCES
1.	Singh, H. C. Atmospheric Distributions, Sources, and Sinks of Selected
Halocarbons, Hydrocarbons, SF6 and N20* Final Report Project 4487, SRI
International, Menlo Park, CA, 1979.
2.	Lonneman, W. A., Kopczynskl, S. L., Darley, P* E., and Sutterfleld, F.
Hydrocarbon Composition of Urban A1r Pollution. Environ. Sc1. Technol.
8:229, 1974.
3.	Westberg, H. M., Holdren, M. W., and H1T1, H. H. Analytical Methodology
for the Identification and Quantitation of Vapor Phase Organic
Pollutants. Final Report Project No. CAPA-U-71, Coordinating Research
Council, 1981.
4.	Bertsch, W., Chang, R. C., and Zlatkls, A. J. The Determination of
Organic Volatlles 1n A1r Pollution Studies: Characterization of
Profiles. Chromatogr. Sc1. 12:175, 1974.
5.	PelUzzarl, E. D., Bunch, J. E., Berkley, R. F., and McRae, J.
Collection and Analysis of Trace Organic Vapor Pollutants 1n Ambient
Atmospheres. Anal. Lett. 9:45, 1976.
6.	Ioffe, B. V., Isldorov, V. A., and Zenevlch, I. G. Gas Chromatographic
Mass Spectrometry Determination of Volatile Organic Compounds in an
Urban Atmosphere. J. Chromatogr. 142:787, 1977.
7.	Holzer, G., Shanfleld, H., Zlatkls, A., Bertsch, W. Juarez, P., and
Mayfleld, H. Collection and Analysis of Trace Organic Emissions from
Natural Sources. J. Chromatogr. 42:787, 1977.
8.	Holdren, M. W., Westberg, H. H., and Zimmerman, P. R. Analysis of
Monoterpene Hydrocarbons In Rural Atmospheres. J. Geophys. Res.
84:5083, 1979.
9.	PelUzzarl, E. D., Bunch, J, E., Burkley, R. E., and McRae. J.
Determination of Trace Hazardous Organic Vapor Pollutants 1n Ambient
Atmospheres by Gas Chromatography/mass Spectrometry/computer. Anal.
Chem. 48:803, 19/6.
10.	Rasmussen, R. A. A Quantitative Cryogenic Sampler, Oeslgn and Operation.
Am. Lab. 4:19, 1972.
11.	Lueb, R. A., Ehhalt, D. H., and H*1dt, L. E. 8a11oon-borne Low
Temperature Air Sampler. Rev. Sc1. Inst. 46:702, 197S.
34
-------
12.	Gallagher, C. C., Forsberg, C. A., and Plerl, R. V. Stratospheric HzQ*
CF2Cl2» and CFCI3 Composition Studies, Utilizing In Situ Cryogenic, Whole
Air Sampling Methods. <3. Geophys. Res. 88:3798, 1983.
13.	McClenny, W. A., Plell, J. 0., Holdren, M. W., and Smith, R. N.
Automated Cryogenic Preconcentratlon and Gas Chromatographic Determi-
nation of Volatile Organic Compounds 1n Air. In Press Anal. Chem. Dec.
1984.
14.	Perma Pure Products, Inc. 8ullet1n 105, Farmlngdale, K.J.
15.	McClenny, W. A. and Pleil, <3. 0. Automated Calibration and Analysis of
VOCs with a Capillary Column Gas Chromatograph Equipped for Reduced
Temperature Trapping. Paper No. 84-17.6 presented at the 77th A1r
Pollution Control Association Annual Meeting, San Francisco, CA, June,
1984.
16.	Ounnett, C. W. A Multiple Comparison Procedure for Comparing Several
Treatments with a Control. J. American Statistical Assoc. 50:1096
(1955).
35
-------
APPENDIX A
AUTOMATED GC SYSTEM SOFTWARE
Documentation for Autosampllng Program
(description of subroutines)
Valve Lister -- Lists position of system valves
Flush -- Configures valves to flush calibrator with zero gas
Gas 4, Gas 3, Gas 2, Gas 1 — Configures calibrator to produce a specific
span gas concentration
Ambient Conflg -- Configures calibration to sample ambient air
Injection Sub -- Collects and Injects a gas sample
Parameter Change Delay -- Allows a delay during which system parameters
can be altered
Valve Offer — Turns all 12 valves off
Single Point Select — Prompts for Information needed to run a single
point calibration
Single Point Run — Runs single point calibration
Sum Var Init — Sets all summation variables to zero
L1n Reg Sub — Performs a linear regression
Peak Finder— Identifies peaks associated with a cal 1b table
Summer Sub — summation of linear regression variables 1s performed
F1d Sub — Displaced portion of Fid Tester Subroutine
Reporter Sub -- Creates calibration report on printer
ECD Sub — Controls electron capture detector calibration data handling
Start on Time — Allows run to be prepared ahead of time then begun at
a specified time of day
36
-------
FIO Tester -- Controls flame Ionization detector calibration data handling
Zero Pointer -- Zero gas run data handling utility
subroutine used by peak finder
F Values	Description of F/T Values 1n Program
1	List Valves 1 » yes
2	Trace Program 1 ¦ yes
3	# of ambient samples
4	Calibrate? 1 ¦ yes
5	Indicates which point 1s to.be run (1n single point mode)
6	# of repetitions during single point mode
7	1 • single point calibration
8	Which valve 1s actuated for the single point
9-17	Unused
18	Temporary storage for dilution factor for single point
calibration
19	Number of points for linear regression
20	Dilution factor for calculation
21	Clock time for the start on time subroutine
22	Desired start time
23	Set to 1 If Q Is used
24-26	Unused
27	Specifies which detector for placement 1n arrays
28	# of compounds found on a particular detector
29	Specifies "titles" of ECD files on tape
30	Indicates to certain subs If the run 1s a zero gas run
T Values
1	Collection time
2	Calibrator flush time
3	Not used
4	Trap flush time
5	Trap cool down time
37
-------
13	R£r» **4#4*.***»#####»»*»*«##«»###««»«»*»^»*#+AUT0 SAMPLING PROGRAM
20	SYNC ON
39	INPUT "GET TABLES?...• "»A#
40	IF A*«"N" tH£H 11®
50	PRINT "GETTING RUN TABLE"
63	GET RON TBL 1 DEVICE# 16
70	PRINT "GETTING CALIBRATION TABLE"
80	GET CALIB 1 DEVICE# 16
59	PRINT "GETTING REPORT TABLE"
100	GET REPORT TBL 1 DEVICE* 16
110	PRINT "CLOSING ALL VALVES"
120	COSUB 1360
130	PRINT "WELCOME TO THE AUTOSAHPLINC PROGRAM"
140	DIM F(30)»X.R(2»20)
130	Dirt C<20>
160	FOR >1 TO 20
170	RtfiD ca>
130	NEXT I
:90	OAT* 10000*10000,10000,10000,10000,10000,10000,10000,10000
200	DATA 3000.500®,10000,1O000»10000,10000,10000,10000,10000,0
210	FOR 1-1 TO 3
220 1(1)
230	SEXT I
240	DfiTA 0, 0.00373, 0.00234, 0.00123, 0.000485
230	DATA SROPAHE,VINYL CHLORIDE,VINYLIDENt CHLORIDE
260	DATA ICHLOROTRIFLUOROETHANE,CHLOROFORM,1-2 SICHLOROETHANE
27d	LATA METHYLCHLQROFORMt BENZENE»CARBON TETRACHLORIDE
290	DATA TRiCHL0R0ETHYLENE»l-3 DICWl0R0PR0PENE
£9v	DATA 1-3 D!CHL090PR9PENE»TOLUENE*1-2 DISROMOETHftNE
300	DATn TETRACHLOROETHYLENEi CHL0MBEN2ENE»0-XYL£NE
310	DA"A 6EN2VL CHu0R:DE»HEXACHL0R08UTA8IENE
320	Dft-A TSICHLOROTRIFLUOROETHANE*CHLOROFORM,METHYLCHLOROFORM
330	DATA CA960N T£TRACHLORIDE»TRICHLOROETHYLENE
3^0	DA-A i-2 3IBROf10ETNANE»TETRACHuOROETHYLENE»HEXACHLOROBUTA5IENE
.35?	"Oft 1-1 TO 30
360	P<. :>»0
370	SEXT I
380	INPjT "DO YOU WISH TO CALIBRATE DURING THIS RUN?  ",A#
3?0	IF A«»"-3H ThEN 1070
*30	IF fiSm'Y" THEN 430
4iO	:<»uT "HOM MANY AMBIENT SAMPLES T0 BE ANALYZED IN THIS RUN? ",F'.3>
420	•"C'O 470
430	r«.4>«l
440	I^PLfT "SINGLE OR MULTIPOINT?  ",A#
430	IF A*«"M" THEN 470
460	CO SUB 1910
470	INPUT "E*TER run IDENTIFIER "»I*
480	IHB|JT "COLLECTION TIME 18? ".Tj
490	INPjT "CAwIBRATOR FLUSH TIME 18? "»T2
5»6	>#UT "TRAP FLU5* TIKE IS? ",T*
5ie	IfPUT "TRA® COOL DOWN TIME IS?"»T5
310	IF	?HE* 340
330	IfP'JT "HON 1ANY AMBIENT SAMPLES BETWEEN CALIBRATIONS? "»F<3>
3*e	INPUT "PROGRAM TRACE? <1«YM> *fr<2>
556 IF F<2)*1 ThEN 360
E«produc«ti .Iront^jl^
•it ¦valiabU copy,
38
-------
566	»PGH ANNOTATION OF"
570	CCT0 39®
530	PRC* ANNOTATION ON
590	F<.'l>-0
600	INPUT "LIST VALVES?  "»F<1>
610	INPUT "BO YOU WISH TO SEE HUN PARAMETERS?  "»At
626	IF MO-Y* THEN 47®
630	LIST
640	INPUT -Amy CHANCES?  ".At
659	IF A#<>"Y* THEH 47®
660	GOSUB 1600
670	INPUT "90 YOU WISH TO SEE THE RUN TABLE? "Y" THEN 730
«9e list sun tbl
700 INPUT "ANY CHANCES?  "»A#
7ii IF A#<>"Y* THEN 730
720 GOSUB 1600
730 I® F(7>»1 THEN 2130
7*9 GOSUB 3330
750 IF F < 4 > < > 1 THEN 1110
760 GOSUB 2450
773 GGSUS 1260
734 GG5UB 1230
790 PAGE
90? PRINT "MULTIPOINT CALIBRATION 2ER0 OAS RUN"
810 GOSUB 1660
920 SAVE REPORT 1 DEVICE# 16
830 GOSUB 1360
340 COSUB 1230
830 PACE
960 PRINT '(1ULTIP0:HT CALIBRATION DILUTION FACTOR "ID<2>
870 GOSUB 1660
sad SAVE REPORT 2 DEVICE# 16
89S GOSUB 1410
¦700 GOSUB 1230
9;0 PACE
92* P*:*T "MULTIPOINT calibration DILUTION FACTOR "»D<3>
930 COSOB 1660
SAVE REPORT 3 DEVICE# 16
950 GOScB 1470
*60 GOSUB 1230
J70 PAGE
>30 PRI*T "MULTIPOINT CALIBRATION DILUTION FACTOR "»D<4>
9»0 GOSUB 1660
1000 SAVE REPORT 4 DEVICE# IS
.010 GOSUB 1530
1020 GC SUB 1230
1030 PAGE
1040 PRINT "MULTIPOINT CALIBRATION DILUTION FACTOR "»D
1050 GOSUB 1660
1060 SAVE REPORT 3 DEVICE# It
1070 CET REPORT TBL 100 DEVICE# 16
1000 GOSwB 3600
1090 GQSUB 3000
il00 GET REPORT TBL I DEVICE# li
1110 GOSUB 1990
39
-------
1120 GOSUB 1230
1150 FCR :•1 r0 F(3>
1140 F"F<19>*l
1150 PACE
1160 PRINT "AMBIENT SAMPLE •"III"FROM "II*
1170 COSUB 1660
1180 EXECUTE Sl»"SAVE REPORT"fcVAL«>1"DEVICE»16"
1190 NEXT I
1200 PRINT "TURNING ALL VALVES OFF"
1210 COSUB 1060
1220 STOP
1230 REM	LISTER
1240 IF F<1><>1 THEN 1260
1230 LIST VALVE
1260 RETURN
1270 COSUB 1860
1280 REM
1290 VALVE 7 ON
1300 VAlVE 8 OFF
1310 VALVE'9 ON
132(3 VALVE 10 ON
1330 VALVE 11 ON
1340 WAIT T2
1370 RETURN
1360 REM	4
1370 VALVE 8 ON
1300 VALVE 12 ON
1390 WAIT T2
1400 RETURN
1410 REM	3
1420 VALVE 8 ON
1430 VALVE 12 'OFF
1440 VALVE 4 ON
14S0 WAIT T2
1460 RETURN
1470 REM	2
1480 VALVE 9 ON
1490 VALVE 4 OFF
1500 VALVE 3 ON
1510 WAIT T2
J?l« RETURN
1530 REM #•~•~~•~~~•#~~~~~•~~•~~~~~~~~•~~••~~~~~•GAS 1
1540 VALVE 8 ON
1350 VALVE 3 OFF
1560 VAlVE 2 ON
1370 WAIT T2
1300 RETURN
1590 REM •••~•~••~••~••••••••~~••~~••~~~•••••••••AMBIENT CONFIC.
1(00 COtOB 1060
1*10 VALVE 9 ON
1020 VALVE IB ON
1*31 VALVE U OH
1640 WAIT T2
1650 RETURN
1660 REM
INJECTION SUB
40
-------
1670 VAlVE I ON
1680 VftLVE 3 OFF
1690 WAIT T4
1700 OVEN TEilP INITIAL VALUE -30
1710 VALVE 3 ON
1720 WAIT T3
1730 VftLVE 1 OFF
1740 WAIT T1
1730 VftLVE 8 OFF
1740 START VftLVE I ON
1770 OVEN TEMP INITIAL VALUE 130
1780 RETURN
1790 ENS
1800 REM *#«#*»»#**»»#»##»»»*#»*##»*»*»*»«»*»<>*#»PARAMETER CHANCE DELAY
1310 PRINT "MAKE CHANCE! NOW"
1620 UAIT 0.3
1830 INPUT "DO YOU REC'JIRE MORE TINE?  "»A«
1840 IF A««"Y" THEN 1810
1830 RETURN
1660 REM	OFFER SUB
1370 FOR !«1 TO 12
1880 EXECUTE SI»"VALVE "tVAt*li"OFP"
1890 NEXT I
1900 RETURN
1918 REM	POINT SELECT
1920 INPUT "REPEAT SINGLE POINT HO* MANY TIMES? "»F<6>
1930 IF F<6)>ie THEN 1933
1940 GOTO 1980
1)30 F<6>«1*
I960 PRINT -caxiMUH OF TEN"
1978 "RInT "TEN HAS BEEN SELECTED"
1530 F<7>¦1
1990 INPUT "WHICH POINT?... 12 3 4  "»F<3>
2600 IF P<3>•1 THEN 2060
2010 IF -<3>»2 THEN 2890
2020 IF F<3>»3 THEN 2120
2030 IF ff<3)-» THEN 2130
2040 peiNT "Invalid option"
2030 GOTO 1990
2060 F>iJ>»2
2070 Fa8)»D':3>
2080 RETURN
2090 F<8>>3
2100 F». 18>"D <4>
2110 RETUP*
2120 F'8>"4
2130 F <15 >-B < 3 >
2140 RETURN
2190 F(8>>12
2160 F<18)«D<2)
2170 RETURN
2180 REM	POINT RUN
2190 50SUB 3570
2200 G09US 2430
2210 GOS08 1280
41
-------
2220 PAGE
2230 PRINT "SINGLE. POINT CALIBRATION ZERO GAS RUN"
2246 GOSUB 1660
2230 F<20>-«
2260 GOSUB 2610
2270 SfiVE REPORT I DEVICE# 16
2280 FOR S2*2 T0+1>
2290 VALVE 8 OK
2300 EXECUTE Sl>" VALVE "*VAL»>fc"ON"
2310 MnlT T2
2323 GOSUB 1230
2330 PAGE
2340 PRINT "SINGLE PT. CALIB. GAS#"»F<3>»" DILUTION FACTOR "»F<18>
2330 53-S2-1
2360 PRINT "RUN*"IS5i"Or"IF<6>
2370 GOSUB 1660
2380 F<20>»F<18)
2390 EXECUTE Sit" SAVE REPORT **VALfiO«0
2490 NEXT K
2305 NEXT J
2310 RETURN
2320 REH	REG SUB
2330 F<19>«X<6?I>
2340 rt-X(1»I>-/F<19)
2330 H»«/I>/F<19>>-(W#(X<1»I)^F<19>>)
2370 R»-</'F<19)>>t2
2380 *»9< -<t2>/F<19>>>
23*8 R»R/>>
26O0 RETURN
2610 RE*	TINDER
2620 FOR N«1 TO #PEfiK3
2630 IF CALIKN>>0 THEN 2630
2640 GOTO 2710
2630 I"CftL#
2660 3"AREA
2670 IF F\30><>1 THEN 2700
2680 GOSUB 3920
2690 GOTO 2710
2700 GOSUB 2730
2710 NEXT N
2720 RETURN
2730 REM ~~•~~~•••~••••~••••••~~••~~•~~••••~•~~~~SUHMER SUB
2740 A»A-X<7»I>
2730 X<3fI>"X<3»I>*>
2760 X(4»I>«X<4»I)*»C
-------
2779 x»x«:i»i>+ft
2789	X(2»I>»X<2,I>-KAt2>
2790	X<3i I>»X<3» I)-»>>
2899 X<6»I>«X<6»I>*1
2619 RETURN
2629 REK	SUB
2639 Z«-"FLAME IONIZATION DETECTOR"
2849 FOR 1*1 TO 16
2859 GGSOB 2329
2669
2879 B<1»I>»B
2889 Rv 1 •I >»R
2399 NEXT I
2999 RETURN
2919 REft ~•~~~~~•~~~~••~~•~~~•~~~~~~~~~~•~~~••~~•REPORTER SU3
2929 PACE
2939 PRInT r»
2949 PRINT. "	FOR V«I1X*B"
2539 FOR K«i TO 3
29e9 PRINT
2979 NEX" <
2989 PRINT "COMPOUHD"»"SLOPE*»"OFFSET"•"R SQUARED"I"(N)"
2599 FOR L"1 TO F<23>
3999 READ l*
3919 F3INT L*
3929 PRINT*.1.L)»B»L>»R»w>»"<"»X<6»L>»">"
3939 PRINT
3949 PRIhT
3939 NEXT l
3069 PACE
3079 RETURN
3969 REM ~~~~~~•~~~•~~~~~~~~~~~~~•~~•~~•••••~••~•~ECD SUB
3999 GET C»i.I8 2 DE-.'TCE# 16
3100 IF F<7>-i THEN 31*9
3119 F\19>«3
3129 GOTO 3149
3139 F<19>«':F(6)*1)
3149 2f- 'Electron capture detector-
3139 RESTORE
3169 G0SU8 2439
3179 FOR jml to 20
3189 READ C(J)
3199 NEXT „•
3209 FOR K*1 TO 3
3219 READ 3"0
3229 NEXT K
3239	FOR K-1 TO 16
3240	READ l»
3239 NEXT <
3264 FOR $3-1 TO 0
3270 C|33)"19O00
3269 NEXT 69
325*0 FOR S3-1 TO
3309	IF F<:7>«1 TnEN 3330
3310	F<29>«D<53>
43
-------
3320 GOTO 3340
3330 F<20>«F<18>
3340 F<29>"83*100
3350 EXECUTE 81>" GET REPORT "«,vm.»"S3
3370 GOSUB 2610
3390 NEXT S3
3390 FOR 83-1 TO 0
3400 X<6»S3>"X<6» S3)*l
3410 NEXT 83
3420 F<27>«2
3430 FOR I«1 TO 0
3440 GOSUB 2320
3430 «<2» X>•«
3460 8<2>I)"B
3470 R<2»I>"P
3480 NEXT I
3«90 F<28>"0
3300 COSUB 2910
3910 GET CALIB 1 DEVICE# 16
3320 RETURN
3330 REM	ON TIME
3340 PRINT "IT IS NOW I CLOCK!"HOURS"
3350 INPUT "ENTER START TIME.., ">F<22>
3360 F<2i>"CL0Ci<
3370 IF F<22>OP«:2l> THEN 3360
3360 PRINT "STARTING RUN ATMCuOCK
1190 RETURN
3630 REM	TESTER
3610 GET CflLI?'! DEVICE# 16
3620 IF F<")«1 THE* 3630
3630 f*3
3640 GOTO 3660
3650 F<1*>«
3669	RESTORE
3670	FOR X»1 TO 20
3680 READ C
3690 NEXT X
3700 FOR X«1 TO 3
3710 REM XKX)
3720 NEXT X
3730 GOSUB 2430
3743 Z*-"FL*f1E IONIZATION DETECTOR"
3730 POR $3-1 TO FU»>
3760 IP F < 7 > ¦ 1 THE* 3790
3770 *<20>"D<83)
3730 GOTO 340O
37>0 Pi20>«Fa8>
3300 EXECUTE 81 •" GET REPORT "4.VAL*<$3>ir DEVICE# 16"
3810 P<30)*83
3020 G080B 2610
3830 MEaT 13
J840 POR 83*1 TO 16
3830 X<«»83)«XU»S3>*t
3840 NEXT 83
((•produced from
b«it «v*ll*oU copy^jgy
44
-------
3378 GGSUB 2929
3530 F'.28>«16
3690 r (2 7) • 1
3?0a GOSUB 291ft
39Id RiTURM
3920 R£H •~+«»~~
393a
3940 RETURN
ERO POINTER
45
R«produc>4 .liurn
b«tt »y»ll>bl« copy>
-------
APPENDIX B
TABLE IB. RAW AREA VALUES FOR CANISTER SAMPLES
Canister 1
Compound Sanedty D«y~Z Sly"
TTW7 Saweday
Canister 2
Day <
Propane
Vinyl Chloride
VinylIdene Chloride
Trlchlorotrl-
fluoroethan#
Chloroform
1.2	Olchloroethane
Methyl Chloroform
Carbon Tetrachloride
Trlchloroethylene
1.3	Olchloropropene
(els)
1,3 Olchloropropene
(trans)
1,2 Dlbromoethana
Tetrachloroethylane
Chlorobenzene
Benzyl Chloride
Hcxachlorobutadlene
W7
2.72
2.83
2.73
2.76
2.76
2.74
2.68
2.77
1.71
1.84
1.86
1.79
1.96
1.91
1.76
2.11
2.01
2.15
2.17
2.16
2.10
2.18
2.09
2.17
1.91
2.03
2.01
2.03
1.94
1.91
1.92
1.95
1.89
2.00
1.97
1.93
1.95
2.01
1.99
1.98
1.S3
1.85
1.83
1.88
1.90
1.88
1.86
1.87
1.91
1.95
1.98
1.94
1.97
1.95
1.95
1.98
i 2.02
2.06
2.12
2.06
2.02
1.95
2.06
1.97
1.75
1.79
1.81
1.80
1.92
1.90
1.82
1.79
0.80
0.72
0.67
0.76
0.81
0.78
0.75
0.73

0.55
0.56
0.61
0.66
0.65
0.64
0.54
2.31*
1.28*
1.29*
1.27*
1.42
1.45
1.42
1.38
1.56
1.59
1.54
1.62
1.64
1.65
1.68
1.65
1.30
1.38
1.37
1.39
1.50
1.52
1.47
1.52
0.99
0.96
0.73
0.95
1.16
1.00
0.78
0.69
1.54
1.54
1.79
1.93
1.71
1.48
1.51
1.90
Canister 3
Compound SamWay Day 2 Day 4
Day 7 immaty
CanUter 4
Day z Day * Day 7
Propane
Vinyl Chloride
Vinylidene Chloride
Trlchlorotrl-
fluoroethane
Chloroform
1.2	Olchloroethane
Methyl Chloroform
Carbon Tetrachloride
Trlchloroethylene
1.3	Olchloroproptn*
(els)
1,3 Olchloropropene
(tram)
1,2 01bromoethan«
Tetrachloroethylene
Chlorobenzene
Benzyl Chloride
Hexachlorobutadiene
2.75
2.73
2.75
2.78
2.56
2.73
2.74
2.76
1.85
1.83
1.90
1.85
1.86
1.80
1.78
1.77
2.09
2.16
2.16
2.16
2.03
2.14
2.16
2.15
1.89
1.92
1.88
1.93
1.91
1.97
1.99
2.02
2.08
2.11
2.05
2.03
1.82
2.00
1.90
1.94
1.9S
1.92
1.88
1.92
1.78
1.84
1.85
1.90
1.97
1.99
1.96
1.9S
1.90
1.96
1.97
1.94
1.96
2.14
2.16
1.98
2.00
2.06
2.06
2.02
1.83
1.87
1.68
1.86
1.73
1.80
1.79
1.79
0.87
0.81
0.80
0.77
0.78
0.74
0.70
0.75
MM
1.45
0.67 a
0.67*
0.67®
****
0.58*
0.60*
0.64
1.44
1.41
1.45
1.27
1.25
1.25
1.29
1.68
1.64
1.65
1.6S
1.48
1.56
1.58
1.57
1.52
1.52
1.50
1.51
1.32
1.38
1.35
1.38
1.07
0.96
0.81
0.89
1.02
0.87
0.74
0.75
1.13
1.07
1.03
1.16
1.23
1.34
1.30
1.26
(a) Data not uitd 1n the statistical analysis (eonsldtrtd as outllars).
46
-------
TABLE IB. (Continued)
Compound saMaiiy^Sty'^ 8iy 4
Day 7 saaeoay day
Canister 6
r fcy 4
Day 7
propan#
Vinyl Chloride
Vlnylldene Chloride
Trlchlorotrl-
fluoroethane
Chloroform
1.2	Olchloroethane
Methyl Chloroform
Carbon Tetrachloride
Tr1chloro#thyl#ne
1.3	Dlchloropropene
(c1»)
1,3 Olchloropropene
(trans)
1,2 01bromo#thane
Tetrachloroethylene
Chlorobenzene
Benzyl Chlorlda
Hexachlorobutadlene
2.74
2.86
2.80
2.79
2.86
2.63
2.71
2.79
1.83
1.89
1.76
1.85
1.82
1.87
1.79
1.91
2.08
2.19
2.09
2.16
2.11
2.14
2.15
2.17
1.99
2.04
1.93
1.94
1.94
1.97
1.97
2.01
1.98
2.09
2.00
2.01
1.86
1.95
1.98
1.91
1.93
1.95
1.92
1.92
1.82
1.90
1.86
1.80
1.96
1.99
1.95
1.98
1.88
1.99
1.94
1.94
2.04
2.00 .
2.04
2.02
2.01
1.97
2.13
1.97
1.84 *
1.61 *
2.44*
1.86 •
1.75
1.80
1.92
1.81
0.84
0.81
0.82
0.81
****
0.79*
0.74*
0.77
0.69
0.7S
0.72
0.71
0.62
0.61
0.59
0.64
1.44
1.46
1.44
1.52
1.28
1.29
1.28
1.31
1.61
1.68
1.61
1.68
1.56
1.60
1.58
1.61
1.49
1.S1
1.80
1.54
1.28
1.35
1.35
1.36
1.06
0.97.
0.91.
0.95
0.88
0.88
0.77
0.73
3.93*
7.36*
7.92
9.59 •
0.94
1.03
1.08
1.01
Compound sameday
CfnlstT 7
oay z IT
Day 4 Day 7
propan#
Vinyl Chloride
Vlnylldtn* Chloride
Tr1chlorotr1-
fluoroathine
Chloroform
1.2	Dlchloroathana
Methyl Chloroform
Carbon Tetrachloride
Trlchloroethylene
1.3	Dlchloropropene
(cl«)
1,3 Dlchloropropene
(tram)
1,2 Olbromoethane
Tetrachloroethylene
Chlorobenzene
Benzyl Chloride
Hexachlorobutadlene
(a) Data not used 1n
2.70
2.77
2.68
2.79
1.80
1.82
1.81
1.98
2.09
2.16
2.15
2.15
1.96
1.94
1.97
1.99
1.94
1.91
2.01
1.92
1.78
1.86
1.84
1.83
1.92
1.94
1.98
1.93
2.01
2.13
2.06
1.99
1.72
1.82
1.82
1.79
1.32*
0.80*
0.74*
0.75
0.62
0.60
0.62
0.63
1.27
1.33
1.32
1.32
1.55
1.59
1.61
1.64
1.33
1.38
1.35
1.33
0.88
0.91
0.83
0.77
0.93
1.04
1.01
1.07
the »tat1it1cal
•ntlyslt (o
47
-------
-------