xvEPA
United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-451/R-92-001
May 1992
Air/Superfund
AIR/SUPERFUND
NATIONAL TECHNICAL
GUIDANCE STUDY SERIES
Applicability of
Open Path Monitors
for Superfund Site Cleanup
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing}
1. REPORT NO.
EPA-451/R-92-001
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Air/Superfund National Technical Guidance Study Series
Applicability of Open Path Monitors for Superfund
Site Cleanup
S. REPORT DATE
May 1992
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Jeff Draves
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
10. PROGRAM ELEMENT NO.
Radian-Corporation
8501 Mo-Pac Boulevard
Austin, Texas 78759
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This document provides guidance on the applicability of Open Path Monitors
(OPMs) for ambient air monitoring at Superfund sites. The relatively limited use of
these devices for ambient air monitoring has led to many question concerning their
applicability, reliability, and sensitivity for these applications.
Information is provided as to the general capabilities of OPMs. The compounds
that each method is capable of detecting are tabulated. The detection limits ef
each type of instrument are discussed and compared to the detection limits achievable
using conventional techniques such as EPA Method TO-14 as well as to short-term and
long-term health-based action levels. The data turnaround time and the ability to
resolve temporal variations in emissions are also discussed.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Air Pollution
Superfund
18. DISTRIBUTION STATEMENT
19. SECURITY CLASS (This Report I
21. NO. OF PAGES
20. SECURITY CLASS (This page)
22. PRICE
EPA Form 2220-1 (R«v. 4-77) PREVIOUS COITION is OBSOLETE
-------
APPLICABILITY OF OPEN PATH MONITORS
FOR SUPERFUND SITE CLEAN-UP
EPA Contract No. 68-DO-0125
Work Assignment 1-31
FINAL REPORT
Prepared for:
Joseph Padgett (MD-10)
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
MavSl. 1992
-------
TABLE OF CONTENTS
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
EXECUTIVE SUMMARY v
ACRONYMS vi
INTRODUCTION 1-1
SUMMARY OF RESULTS 2-1
THE SUPERFUND PROGRAM 3-1
3.1 Pre-Remediation Phase 3-1
3.2 Emergency Response 3-2
3.3 Remediation Phase 3-2
3.4 Post-Remediation Phase 3-3
3.5 General Types of Air and Emission Monitoring Data 3-3
DESCRIPTION OF AMBIENT AIR MONITORING AND DATA
REDUCTION METHODS 4-1
4.1 Open Path Technologies 4-1
4.2 Conventional Point Monitoring Methods 4-14
4.3 Identification and Quantitation of Species Using OPM Systems 4-21
4.4 Application of Path Weighted Concentration 4-24
SELECTION CRITERIA FOR AMBIENT AIR MONITORING
METHODS 5-1
5.1 Detection Limits 5-1
5.2 Response Time of Conventional and Open Path Monitors ... 5-12
5.3 Physical Site Limitations 5-14
5.4 Summary of Advantages and Disadvantages 5-15
5.5 Monitoring During Site Activities 5-15
OPEN PATH MONITORING DURING PRE-REMEDIATION ... 6-1
6.1 Monitoring Needs During the SI 6-1
6.2 Comparison of OPM and Point Monitoring Systems 6-2
6.3 Selection of a Specific Open Path Monitoring 6-4
6.4 Combination of Monitoring Systems 6-5
OPEN PATH MONITORING DURING REMEDIATION 7-1
7.1 Monitoring Needs During the Remedial Investigation 7-3
7.2 Monitoring Needs During Pilot Scale Activities 7-4
7.3 . Monitoring Needs During the Remedial Action 7-6
OPEN PATH MONITORING DURING POST-REMEDIATION . . 8-1
11
-------
9.0
TABLE OF CONTENTS (Continued)
FUTURE DIRECTION OF OPM RESEARCH AND USE OF OPM
FOR SUPERFUND APPLICATIONS 9-1
10.0
11.0
9.1 Overview . . . .
9.2 Field Test and Operational Protocol Developments . . . .
9.3 OPM Hardware and Software Development
9.4 Technology Transfer and User Education
TECHNOLOGY UPDATES
REFERENCES
9-2
9-3
9-6
9-10
10-1
11-1
APPENDIX A:
APPENDIX B:
APPENDIX C:
APPENDIX D:
APPENDIX E:
APPENDDC F:
THE LAMBERT-BEER LAW
POTENTIAL OPTICAL REMOTE SENSING METHODS
GLOSSARY OF TERMS
TRANSECT METHOD
CASE EXAMPLES OF THE USE OF OPM SYSTEMS AT
SUPERFUND SITES
OPEN PATH MONITOR VENDORS
in
-------
LIST OF TABLES
2-1 Data Quality Needs for Air Monitoring During Various
Superfund Actions 2-4
4-1 Optical Remote Sensing 4-3
4-2 Methods for Monitoring Specific Compounds 4-15
5-1 Detection Limits for OPM's 5-5
5-2 Comparison of Limits for Field Measurements vs. Vendor
Specifications 5-6
5-3 Comparison of Detection Limits and Action Levels 5-8
5-4 Comparison of Conventional Point Monitoring and
Open Path Monitors 5-16
LIST OF FIGURES
2-1 Phases of the Superfund Process where Air Monitoring is Conducted 2-2
*+
4-1 Bistatic Configuration of an OPM System 4-5
4-2 Monostatic Configuration of an OPM System 4-6
4-3 Electromagnetic Spectrum 4-7
4-4 Infrared Spectra in the 3,000 cm"1 Region 4-10
4-5 Light Emission Process 4-11
4-6 Mie Light Scattering 4-12
4-7 Compounds Measureable by Various Methods 4-18
4-8 Dispersion in the "xy" plane 4-26
5-1 Extension of the Light Path 5-9
5-2 Multi-passing the Light Beam 5-10
5-3 Source Term Variability 5-13
•
6-1 Monitoring System Setup 6-3
IV
-------
EXECUTIVE SUMMARY
This document provides guidance on the applicability of Open Path Monitors
(OPMs) for ambient air monitoring at Superfund sites. OPMs are a subgroup of Optical
Remote Sensors (ORSs) which are a group of spectroscopic techniques that may have
application to environmental monitoring. The relatively limited use of these devices for
ambient air monitoring has led to many question concerning their applicability,
reliability, and sensitivity for these applications. OPMs are still undergoing development
and the instrumentation and associated software are being refined, but there is a current
need for guidance to help determine when an OPM may be suitable for Superfund
applications.
The actions undertaken during the Superfund clean-up process are divided into
three phases: pre-remediation, remediation, and post-remediation. Each phase may
consist of several steps. This document discusses the specific ambient air monitoring
(AAM) needs during each step. Guidance is provided for the suitability of OPMs for
use during each of these steps based on the typical goals, emission levels, data quality
objectives, etc.
Information is provided as to the general capabilities of OPMs. More specific
information is presented for the three most widely used OPMs for AAM applications:
Fourier transform infrared (FTIR) spectrometry, ultraviolet-differential absorbance
spectrometry (UV-DOAS), and gas filter correlation (GFC) spectrometry. The
compounds that each method is .capable of detecting are. tabulated. The detection limits
of each type of instrument are discussed and compared to the detection limits achievable
using conventional AAM techniques such as EPA Method TO-14 as well as to short-term
and long-term health-based action levels. The data turnaround time and the ability to
resolve temporal variations in emissions are also discussed.
The steps of the Superfund process where OPMs are generally applicable are
identified. OPMs are not considered generally applicable for use during the Site
Investigation, Remedial Investigation, and Operation and Maintenance steps of the
Superfund process. The typical monitoring goals and the baseline emission levels during
these steps dictate that more sensitive monitoring methods be employed. OPMs,
however, may be well-suited for use during the Emergency Response, Feasibility Study
and Remedial Action steps. Emissions are likely to be much higher than baseline levels
during these activities. Also, the variability in the emission rate and potential exposure
of downwind receptors are usually the dominant concerns and OPMs offer certain
advantages over conventional AAM methods with regards to addressing these concerns.
Conventional point monitoring methods and OPMs each have relative advantages
and limitations that must be considered when selecting a monitoring approach. No
single approach is always going to be the best approach for a given type of site or for
activities over the lifetime of a given site. The use of a combination of conventional and
open path monitoring techniques may offer the best potential of meeting the monitoring
needs for various Superfund applications.
-------
ACRONYMS LIST
AAM
APA
BTXE
CERCLA
CF4
CPM
DQO
DIAL
DL
ECD
EM
EPA
ER
FGC
FID
FS
FTIR
GC
GC/MD
GC/MD-MS
GC/MS
GFC
HRS
IR
LC
LIDAR
LIF
LOQ
MDL
MS
NPL
NTG
O&M
OAQPS
OP-FTIR
OPMs
ORSs
OVA
PA
PAH's
PCB's .
PICs
PID
PM
Ambient Air Monitoring
Air Pathway Assessment
Benzene/Toluene/Xylene/Ethylbenzene
Comprehensive Environmental Response, Compensation and Liability Act
Carbon Tetrafluoride
Conventional Path Monitor
Data Quality Objective
Differential Absorbance LIDAR
Detection Limit
Electroconductivity Detector
Electromagnetic Spectrum
Environmental Protection Agency
Emergency Response
Field Gas Chromatographs
Flame lonization Detector
Feasibility Study
Fourier Transform Infrared Spectrometry
Gas Chromatography
Gas Chromatography-multiple Detectors
Gas Chromatograph/multiple Detectors-mass Spectroscopy
Gas Chromatography-mass Spectroscopy
Gas Filter Correlation Spectrometry
Hazardous Ranking System
Infrared
Liquid Chromatography
Light Detection and Ranging
Laser Induced Fluorescence
Limit of Quantitation
Method Detection Limit
Mass Spectrometry
National Priorities List
National Technical Guidance
Operations and Maintenance
Office of Air Quality Planning and Standards
Open Path-FTIR
Open Paths Monitors
Optical Remote Sensors
Organic Vapor Analyzers
Preliminary Assessment
Polycyclic Aromatic Hydrocarbons
Polychlorinated Biphenyls
Products of Incomplete Combustion
Photoionization detector
Paniculate matter
VI
-------
ACRONYMS LIST
(Cominued)
RA Remedial Action
RD Remedial Design
RI Remedial Investigation
ROD Record of Decision
RPMs Remedial Project Managers
SARA Superfund Amendment and Reauthorization Act
SI Site Investigation
SITE Superfund Innovative Technology Evaluation
SODAR Sound Detection and Ranging
SVOC Semi-Volatile Organic Compounds
THC Total Hydrocarbon
UV Ultraviolet
UV-DOAS Ultraviolet Differential Absorbance Spectrometry
VOC Volatile Organic Compounds
Vll
-------
SECTION 1
INTRODUCTION
The Office of Air Quality Planning and Standards (OAQPS) directs a national
Air/Superfund Coordination Program to help EPA Headquarters and the Regional
Superfund Offices evaluate Superfund sites and determine appropriate remedial actions
to mitigate their impacts on air quality. Each Regional Air Program Office has an
Air/Superfund Coordinator who coordinates activities at the regional level.
OAQPS has a number of responsibilities that relate to the Air/Superfund
program, including preparation of national technical guidance (NTG) documents. Since
the Air/Superfund program was initiated in 1987, a number of NTG documents have
been prepared, including a four-volume set of guidance documents1"*. Volume IV of this
set provides information about air monitoring at Superfund sites. Since preparation of
that document, however, the use of spectroscopic methods for air monitoring has
undergone rapid development, and these methods are not adequately described in the
guidance documents.
Open Path Monitors (OPMs), are a group of spectroscopic techniques that apply
to environmental monitoring. Specifically, OPMs are spectroscopic instruments
configured to monitor the open air over extended paths (e.g., paths on the order of
hundreds of meters). They rely on the interaction of light with matter to obtain
information about that matter. Since most OPMs are incapable of determining the
location of species along the path, the data obtained are of a path-weighted variety. For
example, measurements are reported as concentration*path length (or ppm'meter).
While these units may at first appear odd, they provide information that is at least as
useful as data from conventional point samplers or monitors.
The relatively limited use of these devices in the past for monitoring ambient air
has led to many questions about their applicability, reliability, and sensitivity for these
applications. Open path monitoring is still undergoing development and the
1-1
-------
instrumentation and associated software are being refined, but there is a current need for
guidance to help Superfund Remedial Project Managers (RPMs) and other decision-
makers determine when an OPM may be suitable for Superfund applications5'6.
For the Superfund Program, the potential advantages of OPMs compared with
more conventional air monitoring approaches are several: 1) There is rapid, essentially
real-time data analysis that can be used to determine action-level exceedences on site;
2) no sample collection is required in the normal sense of the term; and 3) data are
path-weighted concentrations rather than concentrations for specific sampling points,
which allows source terms to be directly determined. These advantages will be discussed
in detail later in this document.
This document discusses the capabilities of OPMs and compares them with
conventional point monitoring approaches, such as sorbent tube samplers and evacuated
canisters. Continuous, total hydrocarbon (THC) ambient air monitoring (AAM) methods
are considered to be less applicable to Superfund sites; therefore, their relative
advantages and limitations compared to OPMs are not discussed. The use of OPM and
conventional AAM methods for the simultaneous detection of multiple VOC and SVOC
compounds is stressed, since information about the ambient concentrations of individual
compounds is needed for risk assessment purposes.
The goals of this document are twofold. The first is to provide guidance on the
appropriate use of OPMs. This is accomplished by explaining their operational
characteristics and describing the relative strengths and weaknesses of various OPMs.
The second goal is to provide information on the types of OPMs (e.g., FTIR, UV-
DOAS) that are best suited for a specific application. This is primarily accomplished by
presenting expected field detection limits for various OPMs and compounds of potential
interest.
1-2
-------
A related document is being prepared that provides guidance on structuring an
Open Path-FTlR (OP-FlIR) monitoring program7. Similar documents for other types of
OPMs will be prepared in the future if needed.
V
It is assumed that the reader has only a limited knowledge of spectroscopy in
general, and open path monitoring in particular. It is also assumed that the reader is
conversant with the major issues related to ambient air monitoring (AAM) using
conventional approaches, meteorological monitoring, dispersion modeling, and the
assessment of AAM and meteorological data. Volume IV of the NTG series4 provides
an introduction to these topics for readers requiring further information. Furthermore, it
is assumed that the reader is familiar with emission rate measurements using
conventional approaches. Volume II of the NTG series2 provides an introduction to this
topic.
1-3
-------
SECTION 2
SUMMARY OF RESULTS
The Superfund process consists of three phases: pre-remediation, remediation,
and post-remediation. Each of these phases involve several actions or steps that are
undertaken to ensure that the site is cleaned up in a controlled manner. Monitoring or
other air pathway assessment activities may be needed during the course of these actions.
Figure 2-1 shows the various phases of the Superfund process and indicates those steps
where the use of OPMs are most likely.
The past several years have seen an increased application of Open Path Monitors
(OPMs) to both Superfund and industrial sites. However, this use has been limited and
a number of questions remain about the applicability, reliability, and sensitivity of OPM
systems. Although OPMs are still undergoing development, there is a current need for
guidance to answer these questions and to help users determine when an OPM may be
suitable for Superfund applications. This document contains information on the current
state of development of various open path monitoring methods, the air monitoring data
quality objectives associated with the remediation of Superfund sites, and when an OPM
may be a viable option to meet those air monitoring needs.
Since all Superfund sites are unique, this document presents a process for
selecting the appropriate AAM tool rather than absolute recommendations. This process
considers the following questions:
1) What are the goals of the monitoring during this action? For example, is
the goal to determine an absolute concentration or is the goal to determine
emission rates for use in determining receptor exposure?
2) What are the data quality needs? Is screening (qualitative) data adequate
or does quantitative information need to be obtained?
3) What are the requirements for data turnaround? Is there a concern that
short-term action levels may be exceeded?
2-1
-------
Phases of the Superfund Process Where Air Monitoring is Conducted
Air Monitoring
0PM Use Unlikely
0PM Use Possible
0PM Use Possible
0PM Use Possible
0PM Use Possible
0PM Use Unlikely
Site Discovery
Preliminary Assessments
Site Inspections
National Priorities List
Remedial Investigation Feasibility
Study
Records of Decision
Remedial Design
Remedial Action
Operation and Maintenance
Pre-Remediation
Emergency Response
Remediation
Post Remediation
Figure 2-1. Phases of the Superfund Process where Air Monitoring is Conducted.
2-2
-------
4) What are the time resolution needs? Are the sources highly variable,
needing information on the order of minutes to hours, or are they relatively
constant?
5) What are the ambient concentration levels expected to be during the
action?
The answers to these questions define the data quality objectives (DQOs) of a specific
application. These DQOs are then compared with the performance capabilities of the
various monitoring options (OPMs and conventional approaches). The comparison
should help the user decide what type of monitoring would be most appropriate or
indicate whether the DQOs should be modified.
Table 2-1 summarizes the information presented for specific steps in the
Superfund process. Each action is classified by the type of monitoring needed, the
concentration levels likely to be present, the goal of the monitoring, the need for data
turnaround, and the need for high-time resolution. The last column contains a decision
about what type of monitoring is generally most appropriate. In general, when an
absolute ambient concentration level is needed, conventional monitoring is the best
option. When emission rate information is needed, OPMs are often the best option. As
shown in Table 2-1, OPMs should be considered for application to the FS, ER, and RA
since emission rate information is needed, but are not a likely option for the SI, RI, or
the Operation and Maintenance (O&M) actions since absolute concentrations at low,
ambient levels are needed.
As previously mentioned, OPMs are undergoing rapid development and this may
lead to even wider applications of the technology. Some near-term developments and
possible field tests that may further the development of OPMs are discussed later in this
document. These topics include emission rate measurement tests, monitoring during the
RA, quality assurance test, user education, standard operating procedure development,
and complex source flux measurements with plume detection. A discussion of
improvements that the OPM instrumentation needs, such as lowering of detection limits,
expansion of capabilities to the measurement of new compounds, and development of
full stand-alone operation, is also included.
2-3
-------
Cfl
e
^o
«5
. 3
T1 Q
^ ea
~ .5
5 -2
*c
o
i
1
(A
•a
U
"Z.
^^
*j
13
§
*_>
a
Q
§
a.
O
Monitoring
Rcsolulion of Temporal
Variability
0
C
|
§
«
13
Q
,.
1 ^ > §
1 = 1^
ilil
S °s < ®
2^5
e —
.0 =
•i — "p
|1
^
^5
l|
2 Z
Q
c
_o
~
<
«S
*C <« °
3 C ^
^ •— M
115
§t-«
_„ L
2 8 | a
"3 •£ Q. o
S ••= *-S
o jf"2 -
alii
« e 2 2
111!
13
u
1
15
3
vt
I
U
i
0
Z
1!
" M C
Hi
5 5 s
< u Ji
.3
i:
o
VI
II
i §
5 w
oA
^
|
0 «
•5 c
1 8 £
1. 0 §
2 -5 3
o « "i
i & ±
u u 'o
-i*-
« i> >.
i "|f
i-a|
III
Imporlant to estimate
impaci al receptors;
temporal rcsolulion must
be reasonably high
•* _._
c O j«
ill
I'll
OCR
if 1
58"
•- t:
&A S.
E E
J- ~
S « = =•
ills
„ 2 5 a
e c .£ o
"y •— e tc
= -a 5 .§
,u E u 3
— u *3 u
.a
s
Al
1
.—
1
o
fO
= s „
3 C ^*
^ .2 ws
115
3 C r^
^ c —
2 S £ u-
>. u jr "
"C -S a. §
.•:£<- °--=
- '- o u
sin
1 c 22
ilil
1
5
"n
3
5
0
8
5
0
z
m
S
"• ^
111
5 .S "e5
|I|
U -o —
j;
S;
o
VI
>
5
c
tc
3
I
CO
^/
.52
O .t2
CJ -=
°O 1=
O ttT
t_ "
o '5 u
u .C C
° ° .2
.— •- '«
|i|
J= — "3
5*8:
ill
|5|
ill
Key for assessing source
variability temporal
resolution must be
reasonably high.
1
5
o
Z
g
i •
2 3
0 C S
_ 2 ^
o •= 5
l||
— •31-
"5.
0
§
8 F
,s u
-i 5,
U w
0 .0
'e 2
§0
«»
systems may be the
is present will diclau
s-g
ol-
ImporlanI lo estimate
impaci al receptors;
temporal resolution must
be reasonably high.
'o ,
1 S
s -5 *i
3 js "3
1|I
3 "^ s
511
a i
<* s.
= E
i: T
2 o J3
g M c Q.
E ee *S y
o w 2 w
c c c o
!|!l
fu E o «
— O T3 U
.£
51
o
Al
.1
.^S
SB
0
^^
1
_
fl
"S c
S re
3 ta
Ji «-
o .-S
>> s
u c^
«H
sji
|2J
2 |y
o z 3
.1
5
Z
1
5
•z.
u
'§
* 3 .
c -a e»
^" ~ C ij
1 1 1 •=
£ c F c1
< u i-i S
J2
S
0
vt
1
cS
1?
0- =
T «
1 "
2
«^J
0
2-4
-------
SECTION 3
THE SUPERFUND PROGRAM
The U.S. Environmental Protection Agency (EPA), under the Comprehensive
Environmental Response, Compensation, and Liability Act (CERCLA) and the
Superfund Amendments and Reauthorization Act (SARA), is required to develop and
implement measures to clean up hazardous and uncontrolled waste sites. Under
CERCLA and SARA, the U.S. EPA is also responsible for ranking these sites, based on
relative risk to the public health, to determine the order of site cleanup. The Superfund
process generally consists of three phases: pre-remediation, remediation, and post-
remediation. However, if the contamination poses an immediate risk to human health,
an emergency removal or stabilization may be undertaken. The following sections
describe the actions that make up each of these phases.
Ambient air monitoring and emission measurements may be needed during a
number of the steps in the overall Superfund remediation process. Open path
t
monitoring is potentially a cost-effective method for some, but not all, of these steps.
Therefore, an understanding of when an OPM may be appropriate requires an
understanding of the data needs of each step of the Superfund remediation process., An
overview of this process as it relates to Air Pathway Assessments (APA) is given below,
followed by a discussion of the four types of air monitoring data that may be needed
during the course of the remediation process8.
3.1 PRE-REMEDIATION PHASE
The pre-remediation phase is concerned with determining the potential risk to
public health and the environment posed by emissions from the site. The pre-
remediation phase begins with site discovery. From there, a Preliminary Assessment
(PA) is conducted to collect as mugh information as possible about the pollutants present
and their physical state. This is meant to be a low-cost operation and involves collecting
3-1
-------
all relevant documentation about the site. EPA uses the information gathered in the PA
to determine whether further investigation or action is warranted.
If further investigation is warranted, a site investigation (SI) is conducted. The SI
is the first action which employs some form of sample collection and is concerned with
determining the immediacy of the health risk that the site poses. Samples are collected
from the various media present and analyzed and the results are used to rank the ske
within the Hazardous Ranking System (HRS) Model. The HRS model ranks the relative
contamination the site poses over five pathways. These pathways are: air, direct contact,
groundwater, surface water, and fire or explosion. (The direct contact and fire or
explosion pathways are evaluated but not currently included in the ranking.) If the site
scores higher than some predetermined amount, the site is placed on the National
Priorities List (NPL).
32 EMERGENCY RESPONSE
i
Once on the NPL, the necessity of emergency removal is evaluated. If the nature
of the site shows a significant health risk, action is taken to remove and stabilize the
hazardous waste. The emergency response (ER) action is similar to the Remedial
Action (RA) but without the benefit of the information gained during the Feasibility
Study (FS). An ER action is usually shorter in duration and more limited in scope than
an RA. If no emergency removal action is deemed necessary, the remediation phase
begins.
33 REMEDIATION PHASE
The remediation phase consists of the Remedial Investigation (RI) and Feasibility
Study (FS), producing a Record of Decision (ROD), Remedial Design (RD), and
Remedial Action (RA)., This phase requires more time than the pre-remediation phase
and results in reducing health risks from the site to acceptable levels.
3-2
-------
The RI and FS are separate steps but typically they are conducted simultaneously
and interactively. During the RI, data are collected to determine the precise nature of
the compounds present and the extent of contamination. The data gathered during the
RI are used for any risk assessment and to identify cleanup procedures and remedial
alternatives. The FS helps to identify the preferred cleanup alternative. Several
alternative cleanup methods may be considered and, when warranted, developed. After
the FS is completed, a ROD is issued that states EPA's official decision about the
preferred approach to cleaning up the site.
Next are the design and implementation of remedial actions. The RD is a
detailed plan for remediating a site and the RA can take a variety of forms - from
short-term activities when there is a concern for protecting workers to long-term
activities that can take several years.
3.4 POST-REMEDIATION PHASE
Once the remedial activity has ended, a brief monitoring period is initiated to
evaluate the effectiveness of the cleanup. This is called the post-remediation or
operation and maintenance (O&M) phase. If the ambient air monitoring and other post-
remediation testing show that the site meets the site-specific clean-up criteria, the site
may then be taken off the NPL.
3.5 GENERAL TYPES OF AIR AND EMISSION MONITORING DATA
Four categories of ambient air measurements may be required at a site:
1) Qualitative (screening) assessment of site emissions under baseline or
undisturbed conditions;
2) Quantitative assessment of site emissions under baseline or undisturbed
conditions;
3) Quantitative assessment of emissions from pilot-scale remediation
activities; and
-
-------
4) Quantitative assessment of the effects on air quality of full-scale
remediation activities.
The monitoring goal of these measurements is to determine ambient concentrations at
the site fenceline. This is typically done by making measurements both upwind and
downwind of the site contamination. The monitoring goal of categories two and three
may also be to measure ambient concentrations immediately downwind of the emission
source. These data can be combined with tracer gas measurements or atmospheric
dispersion model results to back calculate an emission rate from the source. This
emission rate can then be used with an atmospheric dispersion model to estimate
ambient concentrations at various receptor locations under various meteorological
conditions.
Category 1 monitoring is most likely to occur during the SI, early RI, or the O&M
steps of the Superfund process. Categories 2, 3, and 4 monitoring are most likely to
occur during the RI and O&M, ER, FS, and RA phases.
3-4
-------
SECTION 4
DESCRIPTION OF AMBIENT AIR MONITORING
AND DATA REDUCTION METHODS
All analytical methods rely on some physical or chemical property of a species for
identification and quantitation. Methods such as mass spectroscopy base the
identification of species on their mass. Chromatographic methods rely on the efficiency
with which a compound binds to a substrate to impede the progress of the species, thus
making the transit times through the substrate different for different species. Open path
monitoring techniques are structurally specific, i.e., the measurements are based on how
the atoms are connected or, for an atomic species, how the electrons are configured.
One advantage of OPMs over most conventional methods is their ability to monitor in-
situ, i.e., monitoring without the need to collect samples.
This section describes the basics of selected OPMs and of some of the more
conventional sampling and analytical methods. There is a brief discussion of the
interactions of light and matter, followed by a description of how qualitative and
quantitative measurements are made. Section 4.13 identifies compounds that are
amenable to detection by the three OPMs widely used for environmental applications.
4.1 OPEN PATH TECHNOLOGIES
This section discusses the general principles of optical spectroscopic
measurements. The topics covered include general spectroscopic principles9'10 and light-
matter interactions, which are discussed in terms of their application to ambient air
monitoring.
4.1.1 Spectroscopy
Open path monitoring as used in this document refers to the interaction of light
with matter (i.e., molecules, atoms, or aerosols and paniculate matter) over an open air
path (i.e., no sample cell) to yield qualitative and quantitative information about that
4-1
-------
matter. This type of measurement falls under the term spectroscopy. A device used to
measure the interaction of light and matter is called a spectrometer. A spectrometer
consists of four fundamental parts:
• Source: an element that produces light. This will be referred to as either
broad band because it produces several colors of light (called frequencies)
or laser, which produces a very narrow band of light around a single
frequency (referred to as monochromatic or single color);
• Sample: this is the substance being measured. Here, pollutants in the
ambient air;
• Dispersive element: an element which separates the light into its
component frequencies for analysis. A familiar example of a dispersive
element is a prism; and
• Detector: a light sensitive device.
For open air use, no cell is used. Instead, the species of interest are moved through the
light beam by the wind. This type of monitor is referred to as an Open Path Monitor
(OPM).
The spectrometer described above falls into the class of "active" monitors.
Another class, "passive" monitors is also available. The distinction between active and
passive is that an active spectrometer has an internal light source.
A conceptualized version of a spectrometer used as an OPM is shown in Figure
4-1. Light emitted from the source travels through the sample, where it strikes the
dispersive element and then falls on the detector for analysis. Different spectrometers
will accomplish these four steps in different ways, but the necessity of each step is
universal. Table 4-1 shows several techniques currently used in the optical remote
sensing area11. While several of these methods are appropriate for Superfund sites,
others are in the developmental phase but have the potential to ultimately play a
significant role in several monitoring scenarios.
4-2
-------
Table 4-1.
Optical Remote Sensing
System
Principle of Operation
FITR
Collects full IR spectrum using broadband source and
analyzes spectrum for constituents and concentrations
using library spectra.
UV-DOAS
Collects UV spectrum over limited spectral region and
measures differential absorption of line centers relative
to line wings to deduce gases present and their
concentrations.
GFC
Uses a sample of the gas(es) to be detected as a
spectral filter and measures the broadband correlation
between its spectrum and that of the measurement
path to evaluate gas(es) and their concentrations.
Filtered Band-Pass Absorption
Simple band-pass filtered absorption measurement
using in- and out-band channels to measure absorption
in the band of the gas(es) of interest and then-
concentrations.
Laser Absorption
Straightforward laser transmissometry using one or
more lasers to look in and out of absorption bands to
deduce total monochromatic absorption and,
consequently, gas concentration.
Photoacoustic Spectroscopy
Measures the pressure rise in a closed chamber arising
from collisional deactivation of excited molecules, and
is typically carried out in an acoustic chamber to
measure response acoustically and after excitation with
laser source.
LJDAR
Measures molecular or aerosol backscatter using either
differential absorption with two wavelengths or Raman
scattering to identify and measure gas concentration;
unlike other methods, LIDAR provides ranging
information on measurements.
Diode-laser Spectroscopy
Developing technology for open-air use. Measures
modulation caused by spectrally scanning across a line
feature of the gas(es) of interest to identify and deduce
concentrations by line-depth absorption.
SOURCE: From Reference 11.
4-3
-------
The OPM configuration shown in Figure 4-1 is called a bistatic configuration
because the transmitter and receiver are at opposite ends of the monitored path.
Another configuration, called the monostatic configuration, is depicted in Figure 4-2.
This has both the transmitter and receiver at the same end of the path. The
monostatic configuration requires the use of some type of reflecting optic such as a
corner cube (often called a retroreflector) at the other end of the beam path. The
advantage of a monostatic configuration over a bistatic configuration is that power is
needed at only one end of the path. This is an advantage for rural or inaccessible sites.
4.1.2 Light Interactions with Matter12
Light comes in a variety of frequencies, all of which, taken as a whole, make up
the electromagnetic (EM) spectrum (Figure 4-3). OPM systems typically use either the
infrared (IR) or the ultraviolet (UV) region of the EM spectrum. The various types of
light differ only in their frequency and therefore in their energy content. The energy of
a particular frequency of light is given by the following equation:
E = hv (Eq. 4-1)
where: E = energy (joules);
h = Planck's constant (h=6.6262 x 10"34 Joule'seconds); and
v = frequency in Hertz (sec"1).
Another useful measure of light is its wavelength, which may be derived from the
frequency via the following equation:
r (Eq. 4-2)
A. = -
where: A. = wavelength (m); and
C = speed of light (m/sec, C = 2.9979 x 108 m/sec).
4-4
-------
RECEIVER
SAMPLE
DISPERSIVE ELEMENT
and DETECTOR
COMPUTER
SOURCE
TRANSMITTED
Figure 4-1. Bistatic Configuration of an OPM System.
4-5
-------
REFLECTING OPTIC
RECEIVER
TRANSMITTER
DISPERSIVE ELEMENT
and DETECTOR
COMPUTER
Figure 4-2. Monostatic Configuration of an OPM System.
4-6
-------
N.M.R.
nudear spin
E.S.R.
electron spin
Microwave
rotational
motion
Infrared
vibrational
motion
Visible and
ultraviolet
electronic
distribution
X-ray
I electronic
{ distribution
Gamma-ray
nuclear
configuration
10m
100cm
1 cm
100 urn
REMOTE SENSING
REGION
1 urn
10rvn
100pm
frequency „
^6 3x108
3x10B
3x10
10
3x10
12
3x10
14
3x10
16
3x10
18
wavcnumbars
cm-1
100
10'
10'
10
Figure 4-3. Electromagnetic Spectrum.
4-7
-------
UV light, because of its higher frequency (higher energy) or shorter wavelength,
produces changes in the electron distribution about a molecule. Infrared light produces
changes in the motion of the atoms relative to the center of mass of the molecule. UV
light is therefore said to produce electronic transitions while infrared light produces
vibrational and/or rotational transitions. A transition is a change in the internal energy
content of the species.
When the light beam passes through the sample, three processes may occur that
change the intensity of the light. All of these processes can provide useful information
about the identification, quantitation, or spacial location of the species. They are:
• Absorption, which is the attenuation of the light beam due to the removal
of a photon that matches exactly one of the internal motions of the species.
The internal energy content of the species is increased.
• Emission, which is the loss of light from the species. The internal energy
content of the species is decreased. This light is also at an exact frequency
determined by the species. The ability to emit a photon indicates that the
species was previously excited by some other type of source.
• Scattering, which is the physical deflection of the photon off of its original
course. There are three types of scattering: Mie, Rayleigh, and Raman.
On a molecular level, molecules and atoms can only absorb specific frequencies of
light These frequencies depend on the structure of the molecule or atom. While many
molecules may absorb at the same frequency, no two molecules will absorb at all of the
same frequencies. Therefore, by using several frequencies, a specific pattern or "finger
print" will be obtained that will correspond to a single molecule or atom. Figure 4-4
shows examples of three molecules which absorb in the same frequency region. While
the absorptions may be close to one another, the patterns of the molecules are very
different which allows separate species to be identified even in the midst of the other
species.
4-8
-------
Emission spectra, for the most part, contain the same spectral features as
absorption spectra. Some excitation process (either light or thermal energy) can be used
to produce a species in an excited energy state. The emission of light from this species is
then monitored. Figure 4-5 shows the emission process diagramatically. The excited
state £2 is higher in energy than the other states. The species can lower its energy by
emitting light of exactly the energy Ej-Eo, or it can lose the energy in a step-wise fashion.
The energy loss may either be as thermal energy (energy of motions imparted through
collisions) or as light. If light is emitted it will only occur at the frequency given in the
figure. The emission process, however, is not likely to be used in an open path
configuration because the sensitivity is generally less than that of absorption.
The scattering process can be divided into two classes, elastic and inelastic.
Elastic scattering is a no-energy loss process typical of Mie and Rayleigh scattering.
Inelastic scattering causes the photon to change in energy and is typical of Raman
scattering. Raman and Rayleigh scattering operate by exciting internal energy states of
the molecules. These are species-specific states for the most part and therefore they
have the potential to reveal species-specific information. Mie scattering13, on the other
hand, is the physical deflection of a photon by a particle (see Figure 4-6) that is larger
than or roughly equal to the wavelength of the photon.
4.1.3 Techniques for Superfund Applications
Of the eight techniques listed in Table 4-1, three currently can meet certain
monitoring needs of the Superfund program. These three are Fourier Transform
Infrared (FTIR), Ultraviolet-Differential Optical Absorbance (UV-DOAS), and Gas
Filter Correlation (GFC). They offer the highest sensitivity, specificity, and total analysis
capability of the techniques listed in Table 4-1. Following is a general description of
these systems.
4-9
-------
Methane IR
Spectrum
•ZMO MSO WOO MSO 3000 3050 3100 31SO 3MO 32SO
Ethane IR
Spectrum
woo wso ttoo tasa toao soso 3100 Jiso 3200 szso
Propane IR
Spectrum
•290Q 2950 2900Z§SO 3OOO 3OSO 3100 31SO 32OO 32SO
Figure 4-4. Infrared Spectra in the 3,000 cm'1 Region.
SOURCE: Supplied by W.F. Herget, Nicolet Instruments.
4-10
-------
STATE 1
STATE 2
STATE 0
THERMAL ENERGY LOSS
LIGHT
FREQUENCY = *' 0
MOLECULE
Figure 4-5. Light Emission Process.
4-11
-------
AEROSOL/PARTICULATE
INCIDENT LIGHT
SCATTERED LIGHT
Figure 4-6. Mie Light Scattering.
4-12
-------
Gas Filter Correlation (GFC) Spectroscopy
This technique, which is a special case of spectral correlation techniques, is a
nondispersive infrared method. It relies on the detection of the correlation between
spectral features in the measurement beam and that produced by a reference cell that
contains a sample of the species of interest. This technique, since it is non-dispersive, is
easily applied, is applicable even in the presence of interferences, and has a very fast
(approximately one second) response time. Ten compounds can be monitored with a
single system in its current configuration.
Fourier Transform Infrared (FTIR) Spectroscopy
This technique allows collection of the entire infrared spectrum from about 4000
to 200 wave numbers in a few seconds. The advantage of taking the entire spectrum is
that if interfering species are present they will not be present at all of the same
frequencies throughout the IR. Therefore, careful choice of analysis regions will allow
for interference-free detection. FTIR offers the ability to look at numerous species
which have infrared active modes. Unfortunately, FTIR is also frequently limited in
sensitivity because of the strengths of the IR bands available and the extent of
interferents such as water vapor and CO,. This method would be used in the absorbance
mode for Superfund applications.
Ultraviolet-Differential Optical Absorbance (UV-DOAS) Spectroscopy
This is a special case of conventional absorption UV Spectroscopy with the
appropriate broad-band excitation source such as a Xenon lamp. UV-DOAS looks at
the differential absorption of UV lines relative to the baseline, which is usually a relative
baseline, not an absolute one. The technique is well developed and requires about 1
minute to identify the concentration of a single species. The UV typically has better
sensitivity than the infrared because of the larger UV band strengths, but not all
'compounds are observable in this region. One advantage, however, is that UV systems
can monitor some atomic metal species.
4-13
-------
Table 4-2 contains a list of compounds and the ourent monitoring capability of
the three OPMs just discussed. The capability of canisters analyzed by GC/MD-MS is
included for comparison. Table 4-2 indicates that infrared methods are incapable of
monitoring atomic (e.g., metal) species and that they are also incapable of monitoring a
class of compounds called homonuclear diatomics (i.e., Clj, F2, O^ N2 ). Ultraviolet
systems are currently incapable of monitoring alkanes, alkenes, and many oxygenated
species. While the GFC system can detect most of the same compounds as the FTTR,
there is also a question of the stability of the compounds hi the sample cell. An "X" hi
the GFC column refers only to the ability to detect the compound and not its stability.
The same information is depicted graphically hi Figure 4-7. The solid arrows indicate
the classes of compounds that the open path monitoring techniques can detect. Dotted
arrows indicate the potential to detect some of the compounds in the class. The dashed
arrows indicate that the method may have some problems with these compounds.
This table is not an all-inclusive list of the compounds. Also, it does not attempt
to convey how well these systems will detect these compounds. It is simply offered as a
guide to users to show the compounds that can potentially be monitored by the systems.
4.2 CONVENTIONAL POINT MONITORING METHODS
Point monitors are, as the name implies, used to monitor ambient pollutant
concentrations at a specific point in space for a predetermined time. They are capable
of reasonably low detection limits. They generally require the use of some sampling
method (i.e., canisters, sorbent tubes, etc.). Unless the analyses are performed at the
site, it may take several hours to days before data is available for review.
This section provides a brief description of canister, sorbent tube, and organic
vapor analysis (OVA), and the analytical methods used for separation and detection.
The important differences between point monitors and path monitors are discussed.
However, a more complete discussion is deferred until Section 4.3, where the meaning of
a path-weighted concentration and an effective concentration is discussed.
4-14
-------
Table 4-2.
Methods for Monitoring Specific Compounds
Compounds
Canister
UV-DOAS
FTIR
GFC
ALKANES
ethane
propane
n-butane
n-hexane
isopentane
isooctane
cyclopentane
X
X
X
X
X
X
X
X
X
a
a
a
a
X
X
X
a
a
a
a
X
ALKENES
ethene
propene
1-butene
1-hexene
trans-2-butene
1,3-butadiene
isoprene
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
ALKYNES
acetylene
X
X
X
AROMATICS
benzene
toluene
ethyl benzene
o-xylene
p-xylene
m-xylene
1,3,5-trimethyl benzene
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
HALOGENATED
trichloroethylene
chloroform
tetracholoroethylene
1,1, 1-tricholorethane
methylene chloride
t-l,2-dichloroethane
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
4-15
-------
Table 4-2. (Continued)
Compounds
vinyl chloride
1,2-dichloroethane
chlorobenzene
1, 1-dichloroethane
Carbon Tetrachloride
PCB's
Canister
X
X
X
X
X
UV-DOAS
b
FTIR
X
X
X
X
X
b
GFC
X
X
X
X
X
b
METALS
Arsenic
Cadmium
Chromium
Copper
Lead
Mercury
Zinc
X
OXYGENATES
methanol
ethanol
n-propyl alcohol
dimethyl ether
methyl t-butyl ether
acetone
formaldehyde
acetaldehyde
c
c
c
c
c
c
c
c
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
OTHERS
hydrogen chloride
hydrogen fluoride
hydrogen cyanide
chlorine
bromine
fluorine
carbon dioxide
•
X
X
X
X
X
X
X
X
X
X
X
4-16
-------
Table 4-2. (Continued)
Compounds
Canister
UV-DOAS
FTIR
GFC
CRITERIA POLLUTANTS
ozone
nitrogen dioxide
sulfur dioxide
carbon monoxide
X
X
X
X
X
X
X
X
X
X
X
X = A compound is detectable by this method.
a
b =
These compounds are generally reported as pan of the hydrocarbon continuum
and not as individual compounds.
These techniques have the potential to measure this compound but no detection
limit information is available.
c = "Polar" compounds, specifically oxygenates, may not be measured accurately by
canisters.
4-17
-------
Chemical Class
Alkanes
Alkenes
Alkynes
Halogenates
Oxygenates
Sulfur Compounds
Nitrogen Compounds
Aromatics
Criteria
Methals
Example Compounds
Capability of OPMs
Methane, Propane, Hexane, Octane
Ethene. Butene, 1,3-Butadiene
Ethyne
Chloroform, 1,1,1-trichloroethane
Methanol. Formaldehyde
Hydrogen Sulfide
ammonia, trimethylamine
Benzene, Toluene. Xylenes
Ozone. Sulfur Dioxide Nitrogen Dioxide
Mercury, Lead, Chromium
FT1R GFC
Canister/GC
Detects all compounds in class
Detects some compounds in class
Method suspect for these compounds -
UV-DOAS
' i
Figure 4-7. Compounds Measurable by Various Methods.
4-18
-------
42.1 Canister Sampling and Analysis
A canister is an evacuated stainless steel container with a passivated surface.
Sampling with a canister consists of collecting a controlled volume or mass of
atmosphere over some time period. The atmosphere collected can then be identified
and quantified by gas chromatography (GC), which is a way to separate the compounds
by some physical property, such as size, binding strength, or a host of other methods.
Once separated, the compounds are then identified and quantified.
Typical detection methods use flame ionization (FID), photoionization (PID),
electroconductivity (ECD), and/or mass spectrometry (MS). The former three are
collectively known as multiple detection, because they are frequently applied to a single
sample analysis. Each detector has its particular strengths and weaknesses for various
compounds.
The sampling and analysis methods are therefore referred to as gas chroma-
b
tography-multiple detection (GC/MD) and gas chromatography-mass spectroscopic
(GC/MS). The GC/MD method relies on the chromatographic separation in time to
identify the species; the MD detectors are not species-specific. ,The GC/MS relies to
some extent on chromatographic separation in time, but has the advantage of being mass
specific and therefore, in many cases offers species-specific detection. Table 4-2 contains
a list of compounds that can be monitored by GC/MD and GC/MS. These methods are
typically incapable of monitoring criteria pollutants. (The criteria pollutants are ozone,
nitrogen dioxide, sulfur dioxide, carbon monoxide, lead, and paniculate matter under 10
microns.) Also, the ability of GC/MD or GC/MS to measure most oxygenates and other
"polar" compounds is suspect due in pan to the sampling methods employed.
4-19
-------
4.2.2 Sorbent Tube Sampling and Monitoring
Sorbent tubes contain material that traps contaminants from ambient air via
adsorption. Ambient air is drawn through the sorbent tubes at a known rate and,
depending on the sorbent material, a specific set of compounds are adsorbed onto the
surface of the sorbent. Commonly used sorbents include charcoal, Tenax, XAD-2 resin,
and silica gel.
The tubes are desorbed and the pollutants are then analyzed by either GC or
liquid chromatography (LC). Again, the chromatography only separates the compounds,
detection is accomplished by either: 1) spectroscopic; or 2) a host of other species-
specific methods. Sorbent tubes are better suited than canisters for measuring
concentrations of polar compounds, but each sorbent material is applicable for fewer
compounds than canisters. Background contamination of the sampling media is also a
frequent concern.
6 ' "*
Sorbent tubes are prone to the same limitations of data turnaround time as
canisters are. The type of analysis method needed to detect the compounds listed in
Table 4-2 varies from compound to compound. Sorbent tubes are not amenable to
sampling for metals or criteria pollutants.
4.2.3 Organic Vapor Analyzers and Field Gas Chromatographs
Organic Vapor Analyzers (OVA) are point monitors that analyze total
hydrocarbon emissions using a PID or FID detector. These are portable instruments
that can serve as continuous monitors. They only yield total hydrocarbon data and no
information on specific compounds. Many similar portable (e.g., HNu 101A) and
continuous (e.g., Beckman 400A) analyzers are on the market that, like the OVA,
provide only limited information about the specific compounds present in the
atmosphere.
4-20
-------
Field gas chromatographs (FGC) are identical to the GCs described in Section
.1, but they are more portable. They do not require a sampling method per se, but do
require concentration of the atmosphere. They greatly speed data turnaround time (to
about one hour) but do not have detection limits as low as laboratory-based GCs.
Both OVAs and FGCs offer an advantage in data turnaround time over their
laboratory counterparts. However, they are limited by their speciating capacity and their
detection limits. The FGCs also are incapable of monitoring the criteria pollutants and
metals.
4.3 IDENTIFICATION AND QUANTITATION OF SPECIES USING OPM
SYSTEMS
OPMs are spectroscopic instruments and are therefore able to directly discern
information about the structure of species. As stated earlier, each molecule or atom has
a "finger print" that can be used to identify the compound. In many OPM systems, the
identification is performed by some form of a least squares fit of library spectra to the
open air spectra. The least squares method will fit several (sometimes on the order of
thousands) spectral points; the goodness of fit indicates how well the species have been
identified. Once the species have been identified, the concentration must be determined.
This is done by using the Lambert-Beer Law10-14 more commonly referred to as Beer's
Law. A discussion of the origin of the Lambert-Beer Law appears in Appendix A. The
Lambert-Beer Law simply states that the concentration of a particular component in a
sample is related to the transmittance of a light beam through that sample and to the
length traveled through the sample by that light beam. The transmittance is defined by
the equation:
where: I0(v ) = the intensity in the absence of the sample;
I(v ) = the intensity of the light having passed through the sample.
4-21
-------
Taking the negative base ten logarithm of the transmittance defines the absorbance
which is given as:
I0(v)
A(v) = -log,0 T = log,, -i-j.
where: A(v ) = the absorbance.
The absorbance is a useful quantity because it is directly related to the concentration and
path length by:
A(v) = a(v) • b • c
where: b = path length (m);
c = concentration (ppm); and
a(v) = molecular absorption coefficient
The term a(v ) is a known function for each molecular species. This equation is correct
for a concentration that does not vary with path length. In the ambient atmosphere, the
concentration across the path may not be homogeneous. The equation must then be
written as (see Appendix A):
L (Eq. 4-6)
A(v) = a(v) f c(b) db = a(v) - ceff • L
o
where: L = length of the path (m); and
ceff = effective or path-weighted concentration (ppm-m).
In Equation 4-6, a(v ) is taken out of the integral under the assumption that temperature
and total pressure do not vary significantly over the path.
This equation shows how the concentration depends on the path length. The
integral on the left can be rewritten as though an effective concentration across the
pathlength is being measured. This effective concentration is the quantity obtained from
an OPM and is equivalent to that concentration which, if it is homogeneously distributed
4-22
-------
over the path length, gives the same absorbance. This is often referred to as a path-
averaged concentration. Dividing each side by a(v) gives:
A(v) _ c .L
a(v) **
For OPMs, the concentration-pathlength product is called the path-weighted
concentration. The path-weighted concentration is the invariant parameter associated
with these types of measurements. It is a useful quantity because path lengths will vary
from site to site. One can estimate what effective concentrations will be necessary on
different path lengths to produce the same absorption.
Equation 4-5 shows that the absorbance will increase as the path length increases.
Thus, if one is near the detection limit of a particular species, increasing the path length
(at least in theory) will decrease the detection limits. This is true only if the lengthening
of the path incorporates more and more of the emissions plume. This concept is
discussed in greater detail in Section 5.
The response of an OPM can be used to determine a molecular concentration in
molecules/cm3. The effective concentration can be converted to ppb or ppm if pressure
and temperature are known. The volume monitored by the OPM which is probing a
path L is:
(Eq. 4-8)
V = -2- d2 • L
4
where: d = diameter of the light beam (m).
Using this volume, a total mass in the beam can be determined. Knowledge of the wind
speed, or more precisely, the component of wind speed driving the mass through the
beam allows calculation of a flux by:
J = p ' U (Eq. 4-9)
4-23
-------
where: J = Emission flux across the beam path in (wg/nr-sec);
p = The density of the species in air (ug/m ); and
U = Trie wind speed across the path (m/sec).
An OPM flux measurement is a measure of the amount of mass per second passing
through a plane, parallel to the beam path, having a dimension of d x L. Knowing d x L
and multiplying the flux by this area gives a measure of the rate of mass crossing the
beam.
4.4 APPLICATION OF THE PATH-WEIGHTED CONCENTRATION
A dispersion model can be used to relate path-weighted concentrations to
concentrations in ppb or ppm. The units of path-weighted concentration are, for typical
air samples, expressed as ppm*m, ppb'm, ^g/m2, molecules/m2, etc. Measurements of
quantities with these units can be used to derive input parameters for dispersion
modeling. For example, knowledge of a path-weighted concentration and the wind speed
allows the determination of a source term that can then be used in a dispersion model to
calculate downwind receptor exposure. Two ways of measuring a path-weighted
concentration are by OPMs and by point monitors using the transect method4. The
applications of path-weighted concentrations are described below.
This model, used for illustration, will only consider dispersion of a plume under
the Gaussian dispersion assumption15. The following equation (Turners Equation)
governs the gaussian dispersion of a plume traveling in the x direction.
Q(t)
where: x(x»y>z) = the concentration at a point x,y,z;
Q(t) = the time-dependent source term and has the units of g/sec;
U = the wind speed; and
o j = the spread of the plume in the i direction (i = y, z).
4-24
-------
A one-dimensional view of plume spread is shown in Figure 4-8. If one can obtain path-
weighted concentrations in either the y or z direction (or both), calculation of the source
term is simplified. For a path-weighted concentration in the y direction, Equation 4-10
becomes:
/y
-v Xv^yAjuy fvf\
y _ vw
ry j « u
f1 exp
°z
f-z'll
K
i zJJ
f7
J-y
Y
1
°y
/
-y
expl^-
K
7/y
dy
(Eq. 4-11)
and, replacing the integral on the left with the path-weighted concentration and canceling
the normalization integrals gives:
(Eq. 4-12)
X CW) -
ttU
— exp
dy
If the plume is contained fully within the beam path, the integration yields:
fy — exp - dy = v/2u
J -V rr M 0 > J V
(Eq. 4-13)
leaving:
x(x,y,z). j
TC U
exp
|2o
(Eq. 4-14)
If the source is at ground level, z = 0 and the equation becomes:
X(x,y,z) =
2_
7T U
— exp
2 Q(t)
(Eq. 4-15)
4-25
-------
SOURCE
Plume Center Line
Figure 4-8. Dispersion in the "xy" plane.
4-26
-------
The usefulness of the approach is apparent when the equation is used to solve the source
term Q(t). This is then:
(Eq. 4-16)
X(x,y,o) • U • oz
Q(t) - —;
_2
N «
Now, one only needs to determine az to calculate the source term.
The vertical dispersion, oz, can be determined in a variety of ways. Ideally, oz
could be measured using a lidar system which gives a vertical profile of the plume
directly. It may also be evaluated with a vertical array of point samplers, or it may be
extrapolated from measurement of the wind direction standard deviation (sigma theta)
by using the Pasquill-Gifford stability classes and the associated dispersion curves16.
e»
Another method is to use a tracer gas method on site. The tracer gas (such as
SF6 if using an FTIR) is released (emitted) at a controlled rate. The path-weighted
concentration is measured at a downwind point and this allows calculation of az using the
following equation:
(Eq. 4-17)
°z =
-Q(0,
tracer
X (x,y,o) U
where: Q(t)traccr = The emission rate of the tracer gas.
Field measurements of oz are preferred to extrapolated or estimated values. Recent
studies have shown that even with the measurement of sigma theta, az values determined
by stability class estimates can be in error17.
4-27
-------
Another use of the tracer gas method is to directly determine the source term of
the emission source. Measuring the path-weighted concentrations of both the tracer and
the emissions at some downwind location and taking their ratio will in theory give some
knowledge of the ratio of the source term. This is shown in Equation 4-1818. Since the
Q(t)tracer is known, Q(t)emissions can be determined. While not all species are likely to
behave in the same fashion, the major uncertainty lies in the degree to which the
emission source of the tracer gas emulates the characteristics of the true source. This
will be especially true of large, heterogeneous area sources.
(Eq. 4-18)
trace
CKOemi
emission *. emission
where: QWtrace = tne source term for the tracer gas;
Q(t)emission = the source term for the emission source;
3£ irace = tne tracer gas path-weighted concentration; and
X emission = tne emissions path-weighted concentration.
4-28
-------
SECTION 5
SELECTION CRITERIA FOR AMBIENT AIR MONITORING METHODS
During the various activities undertaken over the lifetime of a Superfund site,
three types of air monitoring may be necessary. These are:
• Worker or personal monitoring (all activities);
• Emissions monitoring (SI, RI/FS, RA); and
Fenceline monitoring (RI/FS, RA, O&M).
Personal monitoring is conducted to determine the exposure of workers to
hazardous emissions and is therefore only conducted by transportable point monitoring
systems. Emission and fenceline monitoring may be conducted by either point or path
monitors.
This section discusses topics that directly affect the type of ambient air monitoring
(AAM) system used. These are detection limits, time response/data turnaround, and the
*
physical characteristics of the site. The discussion covers some of the misconceptions
about detection limits (Section 5.1) and some of the advantages related to fast response
times (Section 5.2).
5.1 DETECTION LIMITS
The term "detection limits" can he defined in several ways. EPA has defined
some of the commonly used terms, including:
• Method Detection Limit (MDL) — the concentration level at which a
method detects or responds; and
• Limit of Quantitation (LOQ) -- the concentration level at which a method
can quantitate a species within some preset precision limits.
•
Currently there exists no specific protocol for determining of detection limits for OPM
systems. Vendors contacted during a telephone survey typically quote the theoretical
5-1
-------
detection limits of their instruments, which are analogous to the best-case detection limit
for a single analyte. Detection limits under field conditions will be higher (worse). This
will be discussed further in Subsection 5.1.2.
5.1.1 Interferants and Misconceptions
The path length parameter associated with the data output of OPMs has created
several misconceptions. These misconceptions arise from failure to take into account the
limitations of the Lambert-Beer Law given in Equation 4-5. This equation is valid only
in certain situations but can be (erroneously) interpreted to imply that lengthening the
light path to several kilometers will allow one to attain the desired detection limit. This
fails to take into account two very important effects:
• Interferants by other compounds; and
• Signal-to-noise ratios.
When monitoring a multi-component system, it is generally assumed that .the absorbance
is additive for most concentrations present in the atmosphere. This means that the
absorbance at a given frequency is a summation of terms caused by all the molecules and
atoms that can absorb at this frequency. Mathematically, this is:-
(Eq. 5-1)
A(v) = b • £ aj(v) • Cj(b)
i=l
where i represents a specific species in the presence of n species.
Equation 5-1 indicates that subtracting the absorbance due to background species
and interfering species will leave the spectra of the compound of interest. This
assumption is valid when the absorbance of each species can be accurately accounted for;
however, problems in cancellation arise when this is not true (i.e., when the absorbance
cannot be accurately subtracted). For example, the bands of water and carbon dioxide
are numerous and intense in the infrared. In addition, they have a very narrow
bandwidth and most FTIRs used as OPMs have insufficient resolution to obtain the exact
5-2
-------
absorbance intensities for these narrow bands. Therefore, when detecting VOCs "and
other species it is important to avoid regions where water and carbon dioxide cannot be
entirely accounted for. Unfortunately, this usually eliminates the stronger infrared
features of the VOCs of interest and limits the analysis to spectral features that have
weak absorption strengths thus increasing the limits of detection. Ultraviolet systems, on
the other hand, have less interference from background atmospheric constituents. The
absorbencies due to water and carbon dioxide in the UV are weak or nonexistent in the
spectral regions of interest.
Some studies have been performed to ascertain detection limits in the presence of
interferants under a variety of operating conditions11'18'19"0. However, there is a need to
standardize the methods used for ascertaining detection limits. Currently, several
methods are used to fit spectra and determine concentrations. Protocols that consider
the application of these specific analysis (data reduction) methods must be generally
applicable and not limited in scope to a single design or approach.
In addition to interferants, one needs to consider the signal-to-noise ratio that an
instrument can obtain. Keeping a signal-noise ratio high enough to detect compounds is
not explicitly written into the equations of Section 4. Equation 4-5 can be (erroneously)
interpreted as implying that low detection limits are attainable at very long distances;
however, the path lengths necessary to achieve these detection limits may be on the
order of several kilometers. Such long path lengths are not practical or feasible in many
cases. While several kilometer path lengths are theoretically possible even for very
confined spaces (via multipassing), the intensity of the light source decreases as the path
length increases. This decrease adversely effects the signal-to-noise ratio. In other
words, there is point of diminishing returns. Currently, FTIR sources are capable of
reasonable source strengths up to approximately 1,000 meters. One should therefore
take care in extrapolating detection limits beyond this point and any path length
significantly over 1 kilometer for FTIRs is unreasonable. Likewise, UV-DOAS systems
are capable of path lengths on the order of 2.5 to 3.0 kilometers, and GFC systems are
capable of path lengths up to about 500 meters to 1 km.
5-3
-------
The user is cautioned to be skeptical when obtaining detection limit information.
While some vendors are aware of the major concerns related to DLs, several vendors
were contacted during the survey conducted for this project who were not. The best
place to obtain valid detection limit information is generally from previous studies that
have used that same instrument. Failing that, check the literature for applications to
specific sites of similar devices.
The following subsection gives the, best estimates of a reasonable field-measured
detection limit for a given OPM and compounds of interest. In some cases, a range of
detection limits is presented because of the variety of vendor responses. When a
vendor-quoted detection limit is used, the limit is identified as such. The user should
remember that the instrument configuration, (i.e., separate transmitter and receiver pairs
versus transceivers), and analysis software may have a significant effect on the field
detection limit. Again, the best solution is to know or obtain field-measured detection
limits of the specific instrument one is planning to use. Even when using the same
instrument, however, changes in the amount of water vapor, as well as the specific mix of
compounds from day to day and location-to-location, will also affect the detection limits
of infrared systems.
5.1.2 Comparison of Detection Limits
Table 5-1 lists the FTIR, UV-DOAS, and GFC detection limits for 25
compounds21 commonly found at Superfund sites. These detection limits, which are
reported in ppm*meters (assuming normal temperature and pressure conditions), were
obtained from a literature search and a vendor survey. A list of OPM vendors is given
in Appendix F. When possible, field detection limits are used; however, owing to the
limited information available, many values are taken directly from vendor literature. It
should be noted that detection limits are dependent on water vapor, path length, and the
spectral analysis region. As Table 5-2 shows, the detection limits quoted by vendors are
typically lower than those obtained in the field.
5-4
-------
Table 5-1.
Detection Limits for OPMs1
Compound2
Trichloroethyiene
Chloroform
Tetrachloroethylene
1,1,1-Trichloroe thane
Methylene chloride
M,2-dichloroethylene
Vinyl chloride
1,2-dichloroethane
Chlorobenzene
1,1-dichloroethane
Carbon Tetrachloride
PCBs
Benzene
Toluene
m-Xylene
o-Xylene
p-Xylene
Ethylbenzene
Phenol
Cyanides
Arsenic Compound
Cadmium Compound
Chromium Compound
Copper Compound
Lead Compound —
Mercury Compound
Zinc Compound
FTIR ppm'm
2.5
3.0
0.7
2-9*
30
2-9*
2*-24
72
10.6
2-9*
2.0
a
26
24
10
15
16
50
a
d
d
d
d
d
d
d
d
GFG ppm'rn
NA v
36*
NA
NA
NA
NA
42*
NA
NA
20*
NA
a
60*
60*
60*
60*
60*
60*
a
d
d
d
d
d
d
d
d
UV-DOAS ppm*m
ND
" ND
ND
ND
ND
ND
ND
ND
NA
ND
ND
a
0.75
0.75
0.75
3.00
0.75
1.50*
0.75*
d
d
d
d
d
d
0.0005C
d
a = potential to detect these compounds
b = compound specific
c = for atomic mercury
d = potential to detect certain compounds
* = provided by vendor
'From References 11, 18, 19, 20, and 30.
2Compounds taken from Reference 21.
ND = Not detectable
NA = Not available
-------
Table 5-2.
Comparison of Detection Limits for Field Measurements
vs. Vendor Specifications
Trichloroethylene
Chloroform
Vinyl chloride
Chlorobenzene
Benzene
Toluene
m-Xylene
FTIR (ppm'm) ,.
Vendor
1
2
2
5
8
8
6
Field
2.5
3.0
13
10.6
26
24
10
UV-DOAS
Vendor
ND
ND
ND
NA
0.75
0.75
0.75
Field
ND
ND
ND
NA
0.75
0.75
0.75
5-6
-------
For comparative purposes, the path-weighted concentration in Table 5-1 is divided
by 200 meters and tabulated by compound in Table 5-3. However, in many instances, a
direct comparison of detection limits for OPMs and action levels is inappropriate (see
Section 5.13). Table 5-3 also includes the detection limits for the GC/MD method and
the long-term (annual) and short-term (1-hour) action levels for these same compounds.
All values are given in units of ppb-V. The use of a 200-meter path is arbitrary, and
some advantages in detecting compounds may be obtained by going to longer path
lengths.
Figure 5-1A shows an emission plume whose width is equal to the path length of
the OPM. The path length for this example is 100 meters. The path-weighted
concentration of the plume along this path is 80 ppm'm, making the effective
concentration 0.80 ppm. If the detection limit is 100 ppm*m (or 1.0 ppm minimum
detectable effective concentration) no emissions are detected. However, in theory,
increasing the path length by a factor of 2 will decrease the detection limit by a factor of
2. Increasing the path length must be done correctly, however, because this only lowers
the detection limit if the path extension includes more of the plume. Figure 5-1B shows
the simple extension of the path to twice its original distance (i.e., 200 meters). This has
lowered the minimum detectable effective concentration to 0.50 ppm. However, in this
example, the path extension does not include more of the plume. This results in
decreasing the plume effective concentration by a factor of 2, making the new effective
»
concentration 0.40 ppm, which is still not detectable.
The detection limit can be decreased by multipassing. Figure 5-2 shows the
simplest setup for multipassing. Note that the extended light beam transverses a path
back through the plume nearly co-linear to that of the first beam. The path-weighted
concentration over both paths is nearly the same, e.g. for the example this would be 80
ppm'm, for a total path-weighted concentration of 160 ppm'm (or 1.60 ppm effective
concentration). The detection limit is still 100 ppm*m (or 0.50 ppm effective
concentration) so the compound is now detectable. Path configurations may be of a
variety of geometries but the potential limitations must be considered. For example, the
effects of interferants may be more pronounced for longer path lengths whether they are
folded or straight.
5-7
-------
Table 5-3.
Comparison415 of Detection Limits and Action Levels
Compound
Trichloroethylene
Chloroform
Tetrachloroethylene
1, 1, 1-Trichloroethane
Methylene chloride
t- 1,2-dichloroethylene
Vinyl chloride
1,2-dichloroethane
Chlorobenzene
1, 1-dichloroethane
Carbon Tetrachloride
PCB's
Benzene
Toluene
m-Xylene
o-Xylene
p-Xylene
Ethylbenzene
Phenol
Cyanides
Arsenic Compound
Cadmium Compound
Chromium Compound
Copper Compound
Lead Compound
Mercury Compound
Zinc Compound
Canister
GC/MD (ppb)
0.47
037
0.57
0.41*
038*
0.52*
0.42
039
0.48
0.92*
0.41
NA
0.53
0.17*
0.15*
0.15*
0.15*
0.67*
a
ND
ND
ND
ND
ND
ND
ND
ND
FT1R
ppb3
115
15
3.5
10-45
150
10-45
10-120
360
53
10-45
10
a
130
120
50
75
250
SO
a
d
d
d
d
d
d
d
d
GFG
Ppb3
NA
180
NA
NA
NA
NA
210
NA
NA
100
NA
a
300
300
300
300
300
300
a
d
d
d
d
d
d
d
d
UV-DOAS
Ppb3
ND
ND
ND
ND
ND
ND
ND
ND
NA
ND
ND
a
3.75
3.75
3.75
15
3.75.
7.5
3.75
d
d
d
d
d
d
0.025=
d
Action Levels
Short-term1
(Ppb)
498
19.6
250
1830
502
2000
10.2
9.90
100
990
20.4
b
0.940
1000
998
998
998
998
49.0
b
b
b
b
b
b
b
b
Long-term1
(ppb)
0.109
0.009
0.279
183
0.606
17.72
0.005
0.009
436
124
0.011
b
0.038
533
161
161
69.0
230
516
b
b
b
b
b
b
b
b
a = potential to detect these compounds
b = compound specific
c = for atomic mercury
d = potential to detect certain compounds
^From Reference 22.
'Compounds taken from Reference 23.
3 Assumes a 200m path length.
4 Short-term action levels are on an hourly basis and long-term action levels are on an annual basis. The
various measurement techniques may yield multiple data points over a one-hour period comparison to the
action levels.
5A direct comparison of OPM detection limits with action levels may be inappropriate, see Section 5.1.3.
* = Provided by vendor
ND = Not detectable
NA = Not available
5-8
-------
SOURCE
SOURCE
B
Figure 5-1. Extension of the Light Path.
5-9
-------
SOURCE
RECEIVER
TRANSMITTER
REFLECTING OPTIC
Figure 5-2. Multipassing the Light Beam.
5-10
-------
5.13 Detection Limit Implications
Examination of the detection limits given in Table 5-2 shows that the OPM limits
are orders of magnitude higher than the canister-GC systems for 200-m path lengths and
a factor of two or three increase in the path length by proper multi-passing will not
greatly alter the comparison. The key question, however, is whether a measured
concentration level equal to the detection limit for a particular compound is sufficient to
meet the program objectives. Even though the concentration at the fenceline may be
below the detection limit, the more relevant question, in many instances, is what is the
concentration at receptor(s) further downwind. As shown in Section 3.4, the source term
may be calculated using the path-weighted concentration. The source term can then be
used with the aid of a dispersion model to predict the concentration at the receptor. In
this way, sampling methods can be used to evaluate adherence to action levels below the
detection limit of the sampling method.
This concept is the basis for the transect method with point samplers. By placing
a row of canisters or other samplers downwind of the site and normal to the wind
direction, a path-weighted concentration can be obtained. A two-dimensional array of
samplers can be used to define the cross-sectioned plume area.
The advantages of using an OPM with dispersion models instead of point
samplers with dispersion models are:
The likelihood of 100% plume capture is increased; i.e., there is less a
chance that the plume will escape detection by either going over, under, or
through the sampling system;
A truer average plume enclusion is obtained;
There is less of a problem of placing the samplers, where physical
constraints ( such as lagoons or pits) are present, however, a clear line-of-
sight is necessary; and
5-11
-------
• There is a time response factor that allows for near real-time updating of
the source term and therefore of the receptor concentration that can aid in
the assessment of short-term action levels (see the next subsection).
The disadvantage of OPMs relative to point monitor is the obtainable detection
limits. However, as stated above the question is whether a compound present at the
detection limit at the fenceline represents an unacceptable risk to downwind receptors.
5.2 RESPONSE TIME OF CONVENTIONAL AND OPEN-PATH MONITOR
One advantage of OPMs is the potential ability to report essentially realtime
compound data. This contrasts with most conventional sampling and analysis methods
that require the collected sample to be analyzed at a laboratory. These methods will
usually have a turnaround time of at least 24-48 hours. While field GCs may be used,
these will have run times of approximately 30 minutes to one hour, with some finite
recycling time, giving an overall data turnaround time of approximately 1 to 1.5 hours.
Clearly, the OPMs are superior in this respect, reporting preliminary concentration data
on the order of minutes. As such, they are comparable to continuous analyzers. Some
OPMs, however, require manual data verification and partial reduction, which increases
the turnaround time for validated data.
Having near real-time data allows the user to correlate site operations with
emissions. For example, during many types of site activity, the emission rate is likely to
vary with amount of activity, the location of the activity, and the depth of disruption (i.e.,
excavation, drilling, etc.) required. A canister system "senses" all of this activity but time-
averages it over the course of sample collection. An OPM, on the other hand, reports
frequent concentration updates and allows better short-term resolution of emissions and
any fluctuations in emissions. This is shown graphically jn Figure 5-3 .
5-12
-------
£ 60
I
E
Q.
Q.
50
O
40
30
LU
O
O
O
h-
g
UJ 10
CAMSTER PATH-WOGHTH) CONCENTRATION
TIME [HOURS]
Figure 5-3. Emission Time Variability
5-13
-------
The curve in Figure 5-3 represents the individual path-weighted concentration
values of compound x ?>? a function of time. Note that some high emissions values ore
measured for a very short duration. The dotted line represents the time-averaged value
collected by the canister. Clearly, the canister data provides no information about the
emission spikes that occurred. Time-averaging the OPM data would give nearly the
same value as the canister. This shows how real-time data can be used to assess the risk
of short-term exposures. While the duration of canister sampling could be shortened, a
practical limit in the number of samples that can be taken is soon reached. In addition,
if a concentration reaches the point where there is a significant risk to downwind
receptors, the data turn-around time the OPM will allow for a quicker shutdown of site
activity. Sampling and analysis methods with longer turnaround times only allow for
hind-sight judgements about such matters; therefore, short duration sampling with
portable analyzers is typically used. However, the lack of specificity of such analyzers
can result in overly conservative decisions about shutdown.
53 PHYSICAL SITE LIMITATIONS
The final consideration is the physical limitations of the site to be monitored.
These limitations include size, proximity to receptors, and obstructions such as trees,
geography, debris, lagoons, etc. All of these physical limitations will affect the type and
cost of the monitoring system chosen. For example, a" site several acres in size will have
relatively long fencelines to monitor and require the placement of several point samplers.
Alternatively, the OPM, while having a finite range, can cover hundreds of meters of
fenceline. A site with a number of obstructions or complex terrain, however, may not
yield a clear line of site. This is not a problem for point monitors. It often requires
ingenuity (and risk to workers) to place point samplers near lagoons or pits, while OPMs
may be able to "shoot" across the site. Other potential physical problems associated with
OPMs and point-sampler system configurations can be easily envisioned.
5-14
-------
5.4 SUMMARY OF ADVANTAGES AND DISADVANTAGES
Table 5-4 lists the advantages and disadvantages of point and 'open path
monitoring systems. Continuous point monitors are not considered because of their
general inability to provide data for specific VOCs. There are six categories: detection
limits, data turnaround time, amenability to dispersion model input, dealing with
unknowns, availability of guidance and standard operating procedures, and physical site
layout
5.5 MONITORING DURING SITE ACTIVITIES
The individual characteristics of each Superfund site will affect the choice of the
monitoring system for each of the actions undertaken. However, there is a fundamental
set of questions and goals to address. These are:
• Goal of the Monitoring - Define the overall goal of both the remediation
activity and the air monitoring;
• Types and Amounts of Emissions Present ~ Estimate the types of
emissions expected and how well each monitoring system will detect these
emissions;
• Physical Geography of the Site — Consider the size, topography and any
obstructions present;
• Proximity to receptors -- Estimate the dispersion of emissions to assess the
effect that the emissions have on downwind receptors. The short-term
exposure monitoring may also be necessary if the receptors are close to the
site. This may affect the need for rapid data turnaround; and
• Actions Taken - Consider how much of the site will be disturbed by
sampling or pilot scale activities and how this will affect the level of
emissions and possible placement of the instrumentation.
For some activities, choice of an appropriate monitoring system will be clear,
while for others specific details of the sites will govern the choice of the system. In some
scenarios, it will probably be advantageous to use both an OPM system and a point
sampling system.
5-15
-------
Table 5-4.
Comparison of Conventional Point Monitoring and Open Path Monitors
Detection Limits
Data Turn Around
Dispersion Model Inputs
Unknowns
Guidance
Site Layout
CPM
Lower detection limits than OPMs;
usually below the long-term action
levels.
For speciated data, at least 1-hour, but
generally 24-48 hours; may not be able
to warn if short-term action levels are
exceeded.
Must be extrapolated from a line of
point samplers.
More accurately identifies the presence
of compounds near the detection limit;
accuracy doesn't necessarily improve as
concentrations increase.
Well documented and accepted.
Sensitive to areas where sampler can
not be placed on the ground;
insensitive to hilly or broken site
terrain.
OPM
Higher detection limits than CPMs on
most path lengths; may be above some
long-term action levels.
Speciated data for tens of compounds
available in minutes. Can warn of
short-term health exceedances.
Direct output in one or two
dimensions.
Not accurate at lowest concentrations
observable, but gets better as
concentration increases over reasonable
range. Easier to identify unknowns at
reasonable concentrations.
Little documentation; still not well
understood.
Sensitive to line-of-sight obstruction;
insensitive to recessed obstacles such as
lagoons and pits.
CPM = Conventional Point Monitor
5-16
-------
SECTION 6
OPEN PATH MONITORING DURING PRE-REMEDIATION
As discussed in Section 3.0 the activities during pre-remediation carry the site
from site discovery through the SI. The only actions that require the use of a monitoring
system, however, are those undertaken during the SI. The primary goal of the SI is to
conclusively demonstrate what types and general levels of emissions are emanating from
the site, as opposed to some upwind source. This monitoring is conducted at both the
upwind and downwind fencelines. The subsections that follow discuss the monitoring
needs of the SI.
6.1 MONITORING NEEDS DURING THE SI
As discussed in Section 3.1, the site is in an undisturbed or baseline state during
the SI. The undisturbed site will typically have only low levels of emissions, especially if
the waste is below surface or if the site is an older one where the volatile compounds
have already dissipated from the surface layer. These concentration levels at the
fenceline are typically in the 1- 10 ppb range. In addition, little is usually known at this
point about the type of emissions, the magnitude of potential emissions present, or the
location of the emitting sources. Therefore:
• The monitoring system must be able to detect a broad range of
compounds, since it is unlikely that the contaminants present will be well
known at this stage; and
• The monitoring system must not only have the sensitivity to detect the level
of compounds present but also the sensitivity to distinguish between
emissions due to upwind sources and those due to on-site sources.
With regard to physical layout, the monitoring system must be sure to intersect
the plume of emissions emanating from both upwind and site sources. The plume must
not evade the system by going around, under, or over the monitoring path. Thus, a
physical inspection of the site and site-specific meteorology and dispersion modeling
6-1
-------
must be used to ascertain the best placement of the system or systems. Some of the
important meteorological consideration are discussed in Section 4.4.
Again, the goal of the SI is to conclusively demonstrate what emissions, if any, are
coming from the site in question. This typically entails two monitoring sessions (though
the downwind-only monitoring may be omitted in many cases). The first is monitoring
downwind of the site to determine the presence of emissions. If emissions are found, the
site is monitored both upwind and downwind of the site simultaneously. The difference
in the emissions detected by upwind and downwind monitoring will determine the
existence of site emissions. As shown in Figure 6-1, the monitoring system is configured
normal to the wind direction, so that the emissions are carried through the monitored
path. The monitored path must be long enough so that any plumes are located
completely within the monitoring path. Also, the sensitivity of the monitoring system to
wind direction and stability class must be evaluated for each site.
6.2 COMPARISON OF OPM AND POINT MONITORING SYSTEMS
The first consideration in developing any monitoring approach is the type of
species likely to be present at the site. This is important, because a given system may
simply not be able to detect the compounds present. The user is therefore referred to
Table 4-2 for a list of compounds and applicable monitoring systems. Again, this is not
an exhaustive list but should be a useful first approximation.
During the SI, the lack of historical information on the compounds present makes
the task of identifying the appropriate monitoring system difficult without resorting to a
trial-and-error approach. It is advisable, therefore, to employ a technique that can detect
a wide range of compounds. If reliable information is available from any site records,
then a more informed choice of system can be made. In general, however, the more
classes of contaminants a monitoring system can detect, the better. Once the choice of
systems has been somewhat narrowed, the important selection criteria are identification
and quantitation of the specific compounds present.
6-2
-------
WIND DIRECTION
SOURCE
SITE
MONITORING SYSTEM
PLUME BOUNDRY
Plume Center Line
Figure 6-1. Monitoring System Setup.
6-3
-------
While it is an easy matter to monitor both upwind and downwind of a site, it is
more difficult to demonstrate that an emissions plume has not evaded the system, i.e., to
prove that a false negative does not exist. The ability of the plume to- pass around,
under, or over a point monitor can be ascertained by using site-specific meteorological
data and information on the local terrain. However, it is also possible for the plume to
pass through the monitoring system without detection. For example, point samplers will
generally be set up equidistant along a path. Thus, there are some areas, (i.e., between
the samplers) that are not monitored, and very narrow plumes may escape detection by
going between the point monitoring stations. OPMs may suffer from the same problem.
While the light beam will interact with any plume coming across the beam path, the
concentration of the constituents may be below the detection limits of the device or
situated above or below the path. In the case of a narrow plume, dispersion modeling
may be necessary to determine the probability of such events. For concentrations below
the detection limit of the method, it will be impossible to determine if an emission
plume is present.
Another consideration when selecting a monitoring approach is the size of the
emitting area. In general, the larger the area, the greater the potential of multiple
sources and the longer the fenceline to monitor. These considerations magnify the
concern about plume capture. There is also a concern about the topography of the site
(see Section 5.3) and the number of point monitors needed, as well as the path length
and detection limits of the OPMs.
Temporal variability in the emission rate is also a concern when choosing a
monitoring system. The emission rate, Q(t), of both the on-site source and any upwind
sources (Section 4.4) may vary, depending on the soil type and moisture content,
temperature, wind speed, etc. This could greatly affect the use of time-averaged
approaches of long duration or poorly timed sampling events. Figure 5-3 shows that a
time-dependent emissions source can go through periods of very high emissions that,
because of the time averaging nature of most point samplers, will not be detected using
6-4
-------
conventional sampling approaches. Secondly, periodic or one-time sampling events may
miss portions or the entirety of the emissions from intermittent sources.
The primary goals of the SI, however, are to demonstrate the presence of any
emissions from the site, and to make a preliminary identification of the emissions source.
For this, the system must be able to detect the presence of the plume distinct from any
background or upwind sources. Canister type monitoring (e.g., EPA Method TO-14) has
detection limits below or near VOC concentration at typical background levels.
Conversely, OPMs do not generally have the ability to detect species at this level.
Several recent studies25-26 using OPMs during the SI reported non-detects for their target
analytes. Therefore, the detection limits of the OPMs make them an unlikely candidate
for use during the SI, despite their advantages for detecting temporal variations in the
sources spatially variable emissions.
6.3 SELECTION OF A SPECIFIC OPEN PATH MONITORING
If the situation is suited to open path monitoring, one has at least three options of
OPMs to choose from (other options are presented in Appendix B). The compound list
should be compared with the list of compounds in Table 4-2. Next, the required
detection limits should be compared with those presented in Table 5-1, keeping in mind
the physical limitations imposed by the site. As previously discussed, longer path lengths
may decrease the detection limits if the gases are distributed over this path.
The use of OPMs for SI applications is generally not advisable. If the use of an
OPM system is warranted, an FUR system will probably be the best choice. The FTIR
is capable of monitoring the greatest range of compounds, among open path monitoring
systems, an asset when little information is available concerning the types of emissions
present. However, the detection limits of FTIR systems may not be low enough to
detect levels of emissions that may be significant from a health risk standpoint. This will
be particularly true of compounds such as benzene. For BTEX compounds, the UV-
DOAS may be more applicable, since it is vastly more sensitive than FUR systems.
6-5
-------
Longer path lengths, however, may improve the FTIR's detection limits to acceptable
levels for some of these compounds. All of the open path monitoring techniques are
capable of monitoring over very short time cycles, which allows the temporal variations
in the source to be characterized.
6.4 COMBINATION OF MONITORING SYSTEMS
In some scenarios, for both the SI and for subsequent activities, a combination
approach using both a point sampler and an OPMs may be desirable. If the detection
limits of the point monitoring systems can be used to complement the range and
temporal advantages of the OPMs, a powerful monitoring tool will be available. The
following example shows how these systems might be combined for use in the SI.
As mentioned earlier, the OPMs need not detect all of the compounds present to
identify the origin of a plume. If some compounds that the OPM can detect well are
present, those compounds may be used as surrogate or indicator compounds18. That is, a
compound that will mimic the physical or meteorological behavior of all the compounds
present The surrogates may then be monitored to obtain information about the
temporal variation of the source and could be used to demonstrate the origin of this
compound. This would allow constant updating of the source term. The addition of
point samplers would allow many compounds below the detection limits of the OPM to
be detected. Fewer point samples would have to be taken since they can now be timed
to coincide with high levels of emissions, as determined by the OPM.
The point samplers may also be used upwind or downwind to determine the origin
of other compounds. Synchronizing the point samplers with the OPM will allow the
origin of the on-site plume to be determined. Calculation of the source term from
dispersion modeling will allow the user information about the plume spread. This, in
turn, allows placement of the point samplers so that the likelihood of plume detection is
improved.
6-6
-------
SECTION 7
OPEN PATH MONITORING DURING REMEDIATION
As discussed in Section 3.0, activities during the remediation phase that may
require monitoring are the Remedial Investigation (RI), Feasibility Study (FS), and
Remedial Action (RA). Two types of air monitoring are of interest: 1) fenceline
ambient air monitoring, and 2) ambient air monitoring just downwind of the emission
source under prescribed conditions to develop emission rate or flux estimates. Air
monitoring will also generally be conducted during the RI/FS and RA to determine
exposure of on-site workers. Since workers tend to move around during the course of a
day, the exposure monitors must also be portable. This type of industrial hygiene
monitoring is typically performed using low-volume sampling pumps and sorbent tubes or
passive dosimeters attached to the worker. Open path monitoring is not a valid
alternative for this type of air monitoring.
The rest of this section addresses the potential uses of OPMs during the RI, FS,
•
and RA steps of the Superfund process. The RI, FS, and RA may each require different
approaches to monitoring. For the FS and RA, the site tends to be in a disturbed state
during these steps and the resulting exposed, disturbed contaminated material will have
much higher emission fluxes than for the baseline, undisturbed case.
The emergency response (ER) action will consist of steps similar to those of the
RA; therefore, the remainder of this document will treat the ER as part of the RA,
pointing out where they differ and where other actions may be appropriate.
The RI is generally an undisturbed site investigation where the goal is to obtain a
more detailed knowledge of the potential air contaminants present. Certain activities
during the RI, however, such as drilling or trenching, may increase the emissions from
the site over baseline levels. The monitoring needs are similar to those of the SI but the
speciation of compounds and the location of emission sources are studied in greater
detail. Since the monitoring needs of the RI are basically similar to those of the SI, the
7-1
-------
information presented in Section 6 for SI activities generally applies except when the site
is disturbed. Then the subsequent comments apply.
Monitoring may also be performed during the RI to determine the risk potential
of the site. This monitoring may be performed directly at specific downwind receptor
locations or indirectly by making emission rate estimations for subsequent use with a
dispersion model to estimate concentrations at downwind receptor locations. Monitoring
at specific receptor locations is likely to be conducted with point samplers (e.g., canisters,
Tenax sorbent tubes), though open path monitoring could also be employed if
concentrations are sufficiently high. Monitoring to develop emission rates could be
performed using either conventional sampling methods or OPM.
The FS develops and evaluates the possible remediation alternatives. The FS may
involve testing some options on a pilot-scale basis, as well as evaluating possible control
technologies. Air monitoring during the FS may be performed to investigate the
emission rates as a function of site, waste, and operational variables. The monitoring
goal during the FS is to determine the emissions rates that will probably be encountered
during the RA. The RA is the full implementation of the chosen remedial alternative
and can proceed over several years. Monitoring during the RA step usually addresses
the risk the site emissions pose to downwind receptors. These emissions are likely to be
significantly higher than those during any other step in the Superfund process, since, at
least for ex-situ remediation processes, greater volumes of contaminated material are
exposed and handled. The monitoring will generally take place at the fenceline to
predict worst-case downwind receptor risk. Monitoring at specific receptor locations
(e.g., the nearest downwind residence) may also be conducted. This monitoring will be
analogous to fenceline monitoring and the type of system used will depend on the
proximity of the receptors to the site.
7-2
-------
The following subsections discuss the specific monitoring needs of each of these
activities (RI, FS, and RA). Information is presented about the necessary monitoring
specifications in light of the goals of the activity. General guidance is offered for
selecting a monitoring system.
7.1 MONITORING NEEDS DURING THE REMEDIAL INVESTIGATION
Determination of emission fluxes during the RI will require identification of the
compounds present in the contaminant plume and determination of the average
concentration of each contaminant of interest in the plume. The temporal variability in
the emission flux will also generally be of interest. This information can then be used to
determine the emission potential of the site; therefore, the monitor must be capable of:
i
• Providing data of a sufficient temporal resolution to assess the emissions
rate; and
• Providing a sensitivity and selectivity equivalent to that required during the
SI.
7.1.1 Comparison of OPMs and Point Monitors
The first consideration is the list of compounds to be monitored. There should be
some measurement data available from the SI or site records to support development of
a target list. This list should be compared to the list of compounds in Table 4-2. It is
also necessary to compare the detection limits and the risk levels associated with the
compounds. As previously discussed, it may be possible to meet the monitoring
objectives, even if the chosen monitoring method has a detection limit for a given
compound higher than the health-based action level for that same compound.
The emission rates from the site and the resulting ambient concentrations are
typically low for undisturbed sites. Therefore, the monitoring system used during the RI
must be able to detect the compounds of interest at low levels (e.g., 1-10 ppbv). This
requirement for high sensitivity will severely limit the choice of OPMs. Since some
7-3
-------
knowledge of the compounds present at the site should have been developed during the
SI step, identification of detectable compounds by OFMs shouid be readily achievable.
This will then help identify surrogates (see Section 7.13). In any event, the OPM must
be able to detect at least some of the compounds present in the plume from the site at
the levels expected. If the sensitivity of the OPM is not adequate to meet this need,
conventional sampling methods will definitely be a better choice. Conventional air
monitoring methods will probably be the preferred choice for most RI applications.
7.12 Selection of a Specific OPM
The OPM needs to have a wide selectivity to be appropriate for monitoring
during the RI; therefore, the JrJLLK is the likely OPM of choice. Unless a surrogate
compound can be identified (see below), however, an OPM may not be appropriate
because of the detection limits observable with OPMs.
7.13 Combinations of OPMs and Point Monitors
The monitoring needed for an RI might be accomplished with a combination of
conventional point samplers and OPMs. This combination would be similar to that
previously described for the SI, with the addition of taking the ratio of the other
compounds detected by the conventional monitoring method to that of the surrogate
detected with the OPM. As before, the surrogate is a compound emitted from the site
than can be detected with the OPM and that can therefore be used to estimate the
behavior of other compounds emitted from the site at levels too low to be detected using
the OPM. The ratios would be used to estimate the concentration of the nonsurrogate
compounds at the receptors.
7-4
-------
7.2 MONITORING NEF.DS DURING PILOT-SCALE ACTIVITIES
Pilot-scale activities during the FS may result in substantially more emissions from
the site than was the case during earlier steps. The primary use of an OPM would be to
make measurements close to the pilot-scale unit to measure emission rates. The
measurements are generally performed as close to the emission source as feasible so that
the measured ambient concentrations will be as high as possible. These concentrations
may be in the range of 100 to 200 ppb per compound. Farther downwind, of course, the
concentrations will decrease because of dispersion, and detection is still possible to the
extent that the OPM can encompass most of the plume volume. The total mass of
contaminants released from pilot-scale activities is likely to be small, so impacts at the
fenceline can be expected to be minimal. Therefore, OPMs are generally no more
appropriate as a fenceline monitor during the FS than they were during the SI or RI
steps.
Compared to the baseline emissions seen during the SI or RI, activities
undertaken during the FS will cause significant changes in both the amount and type of
emissions present. The monitoring system, therefore, must have:
• Detection limits at or below the-concentration levels of the compounds of
interest;
• Sufficient temporal resolution to determine the time variability of
emissions; and
• Ample flexibility in the, range of compounds that can be detected.
Earlier investigatory activities should have determined most of the compounds present in
the waste at the site, so the monitoring system will not need the broad species
identification capability of the systems used in the SI and RI steps.
7-5
-------
7.2.1 Comparison of QPMs and Point Monitors
The goal of monitoring during any FS pilot-scale operations will generally be to
determine emission fluxes or rates that can be ratioed to estimate potential emission
rates during full-scale remediation. This monitoring must therefore have a reasonably
high degree of temporal resolution so that variations in emissions rates can be
determined. The detection limits are a secondary concern for the following reasons:
• The levels of the compounds emitted will be greatly elevated above
baseline levels, so the measured levels are more likely to exceed the
detection limits of the OPMs;
• The monitoring systems can be placed close to the site of the remedial
activity, which allows the plume to be monitored at points where the
diffusion in the xz plane may not be as severe as it will be further away in
the x direction.
However, if no compound is present above the detection limit of the monitoring system,
a more sensitive method must be applied.
Several recent studies17'18"25 have used OPMs in conjunction with dispersion
models to predict downwind impacts17 or measure on-site meteorological parameters.25
One study18 used an FTIR to monitor compounds from a pilot scale dewatering action
and used the path-weighted concentrations to calculate the source term. No monitoring
was conducted at downwind receptors for comparison h.owever, so the accuracy of the
method could not be determined.
7.2.2 Selection of a Specific QPM
Since the compounds present at the site should be reasonably well known at this
step of the process, it is straightforward to determine the overlap of the compound list
with that in Table 4-2. The primary consideration will be the applicability of a given
OPM to the compounds of interest. The detection limits of the OPM will be a
secondary consideration.
7-6
-------
7.2.3 Combination of OPMs and CPMs
The combination approach described in Section 6.1.3 could also be applied during
this activity.
7.3 MONITORING NEEDS DURING THE REMEDIAL ACTION
The primary goal of monitoring during the RA is to protect human health and the
environment. This monitoring is therefore employed at the downwind fenceline or point
of plume impact (see below). One key requirement of any monitoring system for use
during the RA is that it provide rapid feedback to site personnel so that they can halt or
modify the remediation activities if fenceline action levels are exceeded.
Since the primary objective is to protect human health and the environment,
dispersion models play an important role in extrapolating fenceline concentrations to
receptor concentrations. Note that the concentration at the actual site fenceline may not
be a critical piece of information. Rather, only a knowledge of the upper limit (i.e.,
effective action level) fenceline concentration may be important. If this upper limit is
known, it can be used to ascertain a source term for use as input to a dispersion model
in cases where the compound is not detected at the fenceline. The dispersion model can
then be used to calculate the maximum impact at the receptor.
The chosen monitoring system must be capable of the following, regardless of the
type of remedial or sampling action :
The system must be able to measure the compounds identified in the site
emissions during previous actions (RI, FS, etc.) or, for the ER, the system
must be able to identify a wide range of compounds, since little
information may be available; and
The system must have detection limits sufficiently low that (estimated or
measured) ambient concentrations at receptors of interest can be compared
with action levels, regardless of the placement of the monitoring system.
7-7
-------
It is also useful if the system can determine both short-term and long-term exceedances
of the health based action levels.
There must be a reasonable surety that the monitoring system will intercept the
emission plume downwind of the site. One important consideration is meteorological
conditions, such as stability, that affect dispersion of the plume. Thus, the sensitivity of
the system to deviations in the wind direction must be considered.
73.1 Applications to Specific Remediation Approaches
This section considers some of the more commonly used remediation technologies
as examples for monitoring assessment. These technologies are divided into two classes:
point sources and area sources. The example technologies are:
Air Stripping;
Incineration; and
Excavation.
While many other remediation technologies exist, these three are adequate to illustrate
the general issues related to ambient air monitoring for any remediation alternative.
7.3.2 Point Source
In atmospheric dispersion modeling, emissions from point sources are generally
easier to characterize than those from areas sources. Emissions from point sources are
adequately described by the typical Gaussian dispersion and model input parameters are
readily obtained. In addition, tracer gas releases can be used to effectively mimic the
behavior of the point sources and allow for a simple estimation of emission rates using
Equation 4-18.
7-8
-------
Two remediation options that involve point sources of emissions are air stripping
and incineration. In both of these, the emissions emanate from an elevated release point
and the emitted gas will have some velocity in the vertical (z) direction because of the
elevated temperature of the exhaust gas or the effect of a mechanical blower or fan.
The height of the emissions stack, the velocity of the exit gas, and the thermal buoyancy
of the plume will affect how far downwind of the source that plume encounters the
ground. This can be as far as several kilometers downwind of the site. Thus, placement
of the monitoring system is no longer a trivial task.
The behavior of the plume under various conditions can be estimated using an
atmospheric dispersion model and this information can be used to select the monitoring
locations or path. The dispersion model must also be used to calculate the plume spread
at the point of impact so that some estimation of the needed path length for plume
inclusion can be made.
In addition to stack emissions, other sources may exist at the site, such as material
storage and handling facilities. Uncontrolled emissions from storage piles or lagoons and
waste processing and feed units are likely to be much higher than stack emissions;
therefore, some type of fenceline monitoring may be warranted, even if the plume from
the stack(s) passes over the fenceline at some considerable height.
Emissions may also be monitored directly in any process stack, vent, or flue.
These emission rate values can then be used with an atmospheric dispersion model to
estimate downwind ambient concentrations. Cross stack monitoring23 would allow the
ambient concentrations at the receptors to be estimated without the need for monitoring
downwind plumes.
Air Stripping
An air stripper is used to remove volatile compounds from water by introducing
contaminated water at the top of a packed tower and having air flow countercurrent to
the water flow to effect mass transfer of the VOCs to the air phase. Control devices
7-9
-------
such as carbon adsorption systems or thermal oxidation units may be used to reduce the
emissions from the air stripper. The control devices may change the temperature of the
emission plume, as well as its composition.
The plume emitted from an air stripper will have a high water content. The
elevated concentrations of water in the plume may affect the use of infrared
measurement techniques. As discussed in Sections 4 and 5, the interference caused by
water can adversely affect the detection limits. In addition, control devices obviously
eliminate a significant portion of the emissions (e.g., 90-99%). The control devices will
result in lower concentrations downwind of the air stripper and this must be considered
when selecting a monitoring system; i.e., the necessary detection limits at the point of
plume impact may need to be revised.
Incineration
An incinerator is a thermal treatment device that reduces the volume and toxicity
of the contaminated material by combustion. Ideally, organic compounds are oxidized to
carbon dioxide and water vapor, and the halogenated material is converted to these same
compounds plus acid gases. No combustion process is 100% efficient, so some material
is not combusted and passes through the system unchanged. Also, products of
incomplete combustion (PICs) may be formed and emitted. No control devices for
VOCs are used with incineration. Control devices for paniculate matter and acid gases,
however, will remove some organic compounds in addition to altering the characteristics
of the emissions plume.
Since incineration is designed to change the character of the waste stream being
treated, the type of emissions that need to be monitored may be changed from what is
originally present in the waste because of changes in the relative composition of the
material and the presence of PICs. PICs are waste stream dependent but can include
polycyclic aromatic hydrocarbons (PAH's), polychlorinated biphenyls (PCBs), chlorinated
benzenes, chlorinated dioxins, and chlorinated furans. These compounds are difficult to
monitor with many of the current open path monitoring methods23. The user should
7-10
-------
obtain some information about the possible combustion by-products before selecting a
monitoring system.
73.3 Area Sources
Area sources are generally large, heterogeneous, ground-level emission sources.
They may contain multiple areas with vastly, differing emission rates. Modeling the
emissions from an area source is more difficult than modeling emissions from a point
source, since the necessary input parameters to the model are usually known with much
less certainty. The output of emission models for area sources tends to have a large
uncertainty associated with it. Tracer gas methods may also be somewhat difficult to
employ since the degree to which a single point emission or an array of point emissions
mimics the true source is uncertain.
Monitoring methods to determine emission rates from area sources have typically
used flux chambers, the transect method (an array of samplers in both the y and z
» *
directions), or OPMs. While the flux chamber isolates and measures an emissions rate
from a local area source, they are limited in size to only a few square meters. The
degree to which the sampled area represents the entire source affects the number of
points that need to be sampled and the representativeness of the overall data set. Flux
chambers are best suited for relatively small, homogeneous area sources. The other two
measurement options have been compared throughout this document. These methods
yield an emission rate that can be thought of as emanating from a theoretical point
source located some distance upwind of the site, called a virtual point source24. One
common example of an area source during remediation is material handling operations
such as excavation. The monitoring issues related to excavation are significantly different
from those encountered during air stripping and incineration.
7-11
-------
Excavation
Excavation is the physical removal of the contaminated soil or waste from its
subsurface location at the site. The excavated material may be returned to the
excavation pit, transferred to an on-site remediation process or storage area, or
transported to an off-site remediation or disposal process. The excavation process
requires the use of heavy equipment such as backhoes that emit compounds associated
with the combustion of diesel fuel. In addition, large amounts of dust may become
airborne during material handling activities.
As is true for any RA, monitoring will primarily be performed to evaluate the
exposure of downwind receptors to site emissions. For excavation, this monitoring can
take place at the fenceline of the site. The highest release point of the emissions will be
from the elevated backhoe or scoop of the heavy equipment. There will be little or no
velocity component in the z direction; therefore, the plume should quickly reach ground
level.
b " •*
The dust stirred up during site activities is a potential problem for OPMs. The
dust may coat the optics and, to a lesser extent, it may cause Mie scattering if the dust is
of a certain size. From a health stand point, paniculate matter (PM) and associated
contaminants such as heavy metals can contribute to the total risk from the air pathway.
Thus the PM is itself a candidate for monitoring. This topic is discussed further in
Section 9.0.
The air monitoring during the RA is concerned with evaluating the concentration
levels and any risk these levels pose at the downwind receptors. For the calculation of
long-term risk, action levels may be set for the acceptable fenceline concentrations.
Exceedances of these action levels are only a concern if the exceedances are large or
persistent. Another approach that may be employed is to generate a running cumulative
emissions estimate made over the lifetime of the site. This total can be used to set the
operating rate of the RA or the efficiency required for any control devices. In either
case, the data turnaround time is not as critical for this assessment as it is for assessing
7-12
-------
the risk for short-term exposure. The monitoring can be accomplished by either OPMs
or CPMs.
As previously discussed, short-term risk addresses compounds that may pose a
health risk even for very brief exposures because of their action as an asphyxiant, poison,
etc. Evaluation of short-term risk requires a timely update of the ambient air
concentrations. Activities such as excavation may result in large spikes of emissions for
short durations of time. While points samplers are capable of detecting such spikes, the
sample collection period must closely coincide with the lifetime of the emission spike or
it will be missed entirely or the time averaging nature of the sampling will obscure the
severity of the spike. Exact timing of sample start and sample finish to coincide with
emission spikes is not practical for the vast majority of situations.
The suitability of an OPM for determining both short-term and long-term adverse
effects on air quality should be considered. For long-term effects, a preliminary check is
to compare the detection limits with long-term health-based action levels. Another
possible check is to use the detection limits of the OPM as input to a dispersion model
and determine the average concentrations at receptor locations over an appropriate time
interval. This will give some idea about whether the monitoring systems will have
adequate sensitivity. For short-term effects, the emission rate data acquired during any
FS pilot scale operations should be reviewed to see what maximum emission levels can
be expected. The OPMs detection limits should also be checked against short-term
health-based action levels. ~ ~
Applications of OPMs to the RA have been limited primarily because few
Superfund sites have reached the RA. OPMs have however been applied in heavily
industrial areas11'19 where numerous types of emissions are present. The industrial
applications have shown the utility of OPMs when concentrations reach 25 to 50 ppb or
above. Since these emission levels are likely to be reached during the RA. It is
expected that OPMs will be a monitoring option.
7-13
-------
7.3.4 Selecting an OPM
The type of OPM to be used during an RA will depend on the compounds
present and the type of activity. Since FTIRs and GFCs will be affected by high levels of
CO2 and water vapor, the expected concentrations of these interferents should be
considered when estimating field detection limits. The use of UV-DOAS avoids most
problems with these interferents, but the systems may not have the desired selectivity
range needed to detect all of the compounds present. The likely monitoring option for
the ER is the FTIR.
7.3.5 Receptor Confirmatory Monitoring
As previously discussed, monitoring at receptor locations of greatest concern may
be conducted in addition to any fenceline monitoring. This confirmatory monitor is
generally best conducted with point samplers. This recommendation, however, depends
on the proximity to the source and the possibility of short-term exposure exceedances.
7-14
-------
SECTION 8
OPEN PATH MONITORING DURING POST-REMEDIATION
Monitoring, after the completion of the remedial phase, is conducted to ensure
the site is clean and that it does not emit compounds at a hazardous level. This
monitoring is conducted under undisturbed conditions and close to the area of the
remedial activity. If the remedial option did in fact clean up the vast majority of the
contaminants, then the emissions will be low. Therefore, the monitoring system must be
able to detect the emitted compounds at very low concentrations.
The previous monitoring activities at the site should have revealed a wealth of
knowledge about the amounts, location, and types of compounds present. This
knowledge should make identification of the type of system capable of monitoring these
compounds straightforward. Time response is not as critical an issue as during earlier
phases nor is determining variability in the emissions. The question therefore is
primarily one of sensitivity for assessment of risk.
Since the levels of compounds emitted are likely to be no higher than those
encountered in the SI, the user should review any air concentration data from the SI (or
RI) steps at the site. After review, the decision can be made as to whether the
sensitivity of the OPMs is sufficient. If not, as is likely be the case, the appropriate point
monitor may be determined.
Since the rate of decrease of residual emissions from the site may also be an
issue, one useful application of the OPM is as a continuous or semi-continuous monitor.
OPM's can be left to operate in an unattended fashion11. They can be controlled
remotely via phone modems and the data monitored from a remote location. This has
the advantage of keeping the monitoring cost reasonably low and also diminishing the
number of site visits. For this type of monitoring, GFC is an option along with FTIR
and UV-DOAS.
8-1
-------
SECTION 9
FUTURE DIRECTION OF OPM RESEARCH
AND USE OF OPMs FOR SUPERFUND APPLICATIONS
The previous sections discussed OPMs and point monitoring systems and
their application to Superfund monitoring activities. The field of open path monitoring
is still "coming of age" and many of the concepts and practices associated with these
technologies are new. OPMs offer many potential opportunities for monitoring sites in a
more cost-effective but data-intensive manner than conventional AAM methods.
This section discusses what steps need to be taken and what studies need
to be performed to get the maximum benefit from OPM technology for the Superfund
program. An overview is first presented, then this section discusses some of the areas
where further efforts are needed to expand and refine the capabilities of OPMs. These
areas are:
• Quality assurance and quality control tests;
• Emission rate measurements; and
• Complex source flux measurements.
Improvements in these areas will help facilitate and encourage the use of OPMs. These
areas are discussed in Section 9.2, Field Measurements and Operation Protocol
Developments.
A second developmental area for OPMs is improvement of both hardware
and software. These improvements will probably include:
• Reduction of detection limits;
• Expansion of the capabilities of OPMs to allow measurement of
other classes of compounds;
• Development of stand alone operation; and
• Measurement of I0.
9-1
-------
These topics are discussed in Section 9.3, OPM Hardware and Software Development
Tne need for information transfer and user education is discussed in Section 9.4.
9.1 OVERVIEW
Open path monitoring has the potential to eventually be able to provide
accurate, cost-effective data for many Superfund applications. These systems would
provide site managers with timely, pertinent information about exposure via the air
pathway and would ultimately result in faster and safer remediations. Much work,
however, remains to be done before OPMs realize their full potential for AAM at
Superfund sites.
The biggest need is for EPA to develop a long-term developmental plan
for the use of optical remote sensing methods, including OPMs, for environmental
applications. This scope obviously exceeds just Superfund sites. The basic questions that
need to be addressed are:
1) Current and future capabilities of optical remote sensor (ORS)
technologies;
2) Current and future monitoring needs;
3) Current applicability of ORS systems;
4) Recommended developmental and evaluative studies; and
5) Development of standard procedures and practices.
This information is necessary before a long-term plan can be developed to guide the
development of ORS and OPM as standard compliance-type methods. The U.S. EPA's
Office of Air Quality Planning and Standards and the Office of Research and
Development have begun the process of developing such a long-term plan with input
from academia, industry, other users, and equipment vendors.
It is hoped that ultimately the capabilities of various OPM technologies will
be well-demonstrated for environmental applications, that standard operating protocols
will be available, and that minimum performance specifications or requirements will be
established.
9-2
-------
9.2 FIELD TEST AND OPERATIONAL PROTOCOL DEVELOPMENTS
The recommendations given in this document are necessarily based on the
current knowledge of the capabilities of OPMs. As mentioned previously, only a limited
number of applications have been attempted to date at Superfund and other hazardous
waste sites. The number of field applications to date is small and these efforts have
been limited in scope. Furthermore, the use of OPMs requires a high degree of
technical expertise in both spectroscopy and ambient air monitoring. These factors have
lead to many questions and some apprehension about the use of these methods. This
section discusses some of the important issues and tests that are needed to address these
concerns and to help ensure the proper use of OPMs.
92,1 Quality Assurance and Quality Control
Quality Assurance (QA) and Quality Control (QC) tests are necessary in
ambient air monitoring to provide a mechanism for the ongoing control and evaluation
of data quality throughout a measurement program. They also ensure that the data
collected are of known accuracy and precision. For conventional AAM, these QA/QC
tests generally take the form of performance and systems audits, the collection of
replicate samples, the analysis of spiked samples, the analysis of system and analytical
blanks, etc. Analogous QA/QC tests for OPMs have yet to be developed and accepted
as pan of standard operating procedures (SOPs).
r
The QA/QC tests appropriate for conventional AAM methods are not
generally applicable to OPM methods because of inherent differences in the methods.
Conventional approaches generally involve the physical collection of an air sample at a
discrete point using a pump or vacuum to draw the air into a collection device or
through a filter or sorbent, while OPMs do not involve any physical collection or
compositing of the sample. The QA/QC tests used to date for OPMs involve checks of
the instrumentation itself and checks of the sampling strategy (i.e., light path location,
frequency of sampling, when to sample, etc.). Checks of the instrumentation generally
9-3
-------
involve collecting data from a standard gas in a closed cell and using these data to
demonstrate whether the OPM system is operating in accordance with some set of
established guidelines. These cell tests, however, are not directly comparable to data
from open paths since there is no accounting for light losses due to divergence, rapidly
varying path concentrations in the atmosphere, interferents, etc.
Similar limitations exist for QA/QC tests to evaluate the sampling strategy
selected for a given application. While no standard procedures have emerged, some
potential methods have been suggested and attempted in the field. Data accuracy can be
determined by collecting background spectra and through the use of tracer gases and
confirmatory sampling.
The background tests involve collecting data under field conditions without
an active emission source. For example, to determine excavation emissions data could
be collected before the excavation equipment is turned on to get background or baseline
data. A second set of data could be collected with the excavation equipment running but
before excavation of any material to determine emissions from the excavation equipment
itself. For measurements during the RI, blank tests could be performed by setting up the
OPM upwind of the emission source under conditions (e.g., path length, number of beam
reflections) that match as closely as possible the conditions encountered downwind of the
emission source.
The accuracy of OPM data can be evaluated using a tracer gas release that
closely mimics the emission source. The validity of this approach depends on how
closely the tracer gas mimics the release, transport, and analysis of the target analytes.
For Superfund applications, the limiting factors will generally be the ability to create an
area emission source of the tracer gas and getting the release point(s) to coincide with
the actual emission source. The validity of the Gaussian dispersion model inherent to
this approach will also depend on a number of factors. In general, it will not be valid
over short distances (e.g., <50-100m). Another approach for determining the accuracy of
OPM data is to set up a series of point samplers along the line of the light beam. While
9-4
-------
an excellent check of OPM performance for evaluative studies, this approach has several
drawbacks as a routine QC test such as cost and data turnaround time.
The assessment of the precision of OPMs is not as critical an issue as it is
with conventional monitoring systems. OPMs collect large numbers of spectra and
average these spectra. Least squares fitting of reference spectra to these averaged
spectra is then done using thousands of points, each a measure of the concentration.
This large number of measurements of the concentration allows rapid assessment of the
instrument precision.
9.2.2 Measurement of Emissions Rates
One of the advantages of OPMs over most conventional AAM methods is
the ability to use the path-weighted concentration to calculate a source term or emission
rate. While theoretically this is readily achievable, the accuracy and precision of this
method have not been adequately tested. Since emission rate determination is a key
area for application of OPMs for Superfund, test should be performed to validate this
application. This would involve evaluation of the atmospheric dispersion model used to
back calculate an emission rate as well as an evaluation of the approach for collecting
the OPM data.
For example, tests similar to those conducted by EPA region VII29 where
mixtures of gases were released at a known rate and detected downwind could be used
as a starting point for these evaluations. The OPM could be placed close to the source
along with a row of canisters (transect method) for purposes of comparison and
normalization. A dispersion model would then be used to predict the concentration at a
receptor some distance downwind with OPM data as input. The receptor could be a
single canister or another row of canisters.
This test would allow prediction of the emissions rate that could then be
compared to the known emission rate. The predicted receptor concentration could then
9-5
-------
be compared to the measured concentration at the receptor. The row of canisters would
be used to normalize the path-weighted concentration and to act as a reference method.
A series of such tests are needed to assess the validity of this approach for varying source
strengths, release points, meteorological conditions, terrain and so forth.
9.23 Complex Source Flux Measurement
Many Superfund sites and industrial complexes have multiple sources of
varying emissions. The plumes from these sources are often mixed by the time they
reach the site fenceline. In addition, the extent of the plume is generally not well known
making the certainty of plume capture difficult. It would therefore be advantageous to
develop integrated electro/electro-optical, monitoring systems that would sense the extent
of a plume (with LIDAR) and measure the concentration of the species present in the
plume in both the horizontal and vertical directions (with OPMs). This would in effect,
involve monitoring everything that passes through a plane parallel to the site. Coupling
these systems with radars, sodars or meteorological towers (depending on source height
and complexity) could then allow the determination of wind fields that could in turn be
used to determine the total flux of emissions from the site.
There have been some conceptual designs for such monitoring systems but
a great deal of work would have to done before they could be ready for testing.
However, this type of emissions measurement would be very comprehensive and could
potentially be used as the standard for assessing total emission from a site or complex.
It therefore has great promise for emission inventory type work.
9.3 OPM HARDWARE AND SOFTWARE DEVELOPMENT
Currently, one of the major drawbacks to the use of OPMs is their inability
to measure very low concentrations (1-5 ppb). Sensitivity to low concentrations is a key
issue when conducting monitoring in the SI, RI and O&M. There are certain
fundamental hardware and software problems that must be overcome which will allow
9-6
-------
the detection limits to be lowered. In addition to detection limits, two other problems:
stand alone operation and expansion of OPMs to new chemical classes are discussed
below.
9.3.1 Reduction of Detection Limits
The disadvantage of OPMs, compared to point monitors is their detection
limits. Several hundred meter pathlengths are often necessary to produce detection
limits in the single digit ppb range for most compounds. However, when using infrared
methods single digit ppb detection limits are not reached even at 500 meters for
compounds such as benzene and toluene (see Table 5-3).
Improvements in detection limits will likely be based on an increase in the
signal-to-noise ratio and accurate removal of any interferants. The signal-to-noise ratio
increase may be accomplished by making the detection system more sensitive, improving
the properties of the light source, or having more efficient light collection to improve the
signal intensity27. The interferant issue must be addressed by accurate assessment of the
interfering compounds present and their spectra. In particular, the effects of water and
carbon dioxide in the infrared must be fully accounted for in all aspects including their
non-linear behavior. Accurate removal of the water and carbon dioxide from the spectra
using mathematical methods will in principle give access to new regions of the spectra,
i.e., regions where other molecules have stronger absorptions. One attempt to remove
the water problem has been to collect reference spectra at several concentrations and
remove the water using the spectra that is in closest agreement. This has not been
completely proven and requires the fitting of several different spectra in a trial-and-error
manner. Even in this case however, the large span of absorbance frequently makes
complete cancellation of water vapor and carbon dioxide impossible.
9-7
-------
9.3.2 Other Classes of Compounds
Another issue where detection limits are important is the ability to detect
"other classes of compounds." These compounds include metals in both atomic and
complexed forms, polycyclic aromatic hydrocarbons (PAH's), and the chlorinated
varieties of dioxins/furans, and compounds bound to particulate matter (PM). In
addition, the increasing interest in dust emissions makes PM monitoring a likely
requirement. These classes of compounds are found at Superfund sites but can rarely be
detected using OPMs because they are below the detection limits and OPM methods are
unable to sense their presence.
Metals in their atomic state are not capable of being monitored with
infrared technologies. They can however, be monitored by ultraviolet methods. Both
path and point variety methods need to be developed that will accurately characterize
these emissions. Metal atoms in a chemical complex are much harder to account for
since different complexes will have different structures. However, metals in different
oxidation states have very different risk potentials and spectroscopic techniques have
been shown to be sensitive to these different states28. Therefore, some investigation of
these techniques is warranted to application to just such monitoring.
Other compounds such as PAH's, dioxins, and furans have extremely low
action levels. Sensitive instruments must be used to detect these species at these levels.
ORS methods have been reviewed for application as CEMs for these compounds at
municipal solid waste incinerators23 and laser induced fluorescence (LIF) methods were
recommended for their detection. Various methods for application to these compounds
were discussed in this report. A description of these methods are included in Appendix
B.
Finally, not all emissions are expelled in the molecular or atomic form but
may undergo some form of association with other compounds such as molecules or
particulate matter. These complexes are not well detected by any of the conventional
9-8
-------
methods. OPMs may have the ability to detect these species when associated or bound
to paniculate matter.
93.3 Stand Alone Operation
All OPMs are equipped with computers that at the very minimum, collect,
process, and store data. In fact, much credit for allowing some of the ORS systems to be
used as AAM goes to improvements in computing capability. The uses of the computer
control has, however, not been fully exploited.
The use of the computer could be expanded to include stand alone
operation in which the computer collected data in an unattended fashion. This lowers
labor costs and frees the operators for other duties. This type of data collection has
been tried recently with good success11. It is not without problems, however; and efforts
need to be made to introduce artificial intelligence systems and expert systems that can
investigate and correct problems that may arise.
Another use of the artificial intelligence would be to have a system capable
of detecting the presence and location of an emissions plume. The extent of the plume
could be mapped out by the. LIDAR methods and the system could be updated to follow
the evolution of the plume. This would require the introduction of meteorological data
to assess the plume direction.
A complete package for monitoring the flux of emissions from complex
sources rriay be assembled by adding a wind field profiler. This profiler could come in
the form of a simple met tower if vertical plume dispersion is low or depending on
complexity, a sodar or radar system. The profiler would measure the local wind fields
throughout the distributed source area.
9-9
-------
Advances in this area would eliminate much of the worry over plume
capture and dispersion assumptions. The operation could be performed remotely by
using phone modems and in a very cost effective manner. In addition, the concern over
tracer gases mimicking the source would no longer be a concern.
93.4 Measurement of I.
o
Appendix A contains a derivation of the fundamental equation used in
open path monitoring to relate a measured parameter, light intensity, to concentration.
Equation A-12 shows that the absorbance, which is directly related to concentration, is
equal to the logarithm of the ratio of two intensities. The first intensity, I(v ), is the
intensity measured after the light beam passes through a sample. The second intensity,
Io(v ), is defined as the light intensity which is incident on the sample. In practice, ^(v )
is modified by the detector response, divergence losses, and reflective losses from the
optics and the combined effect of these is called the instrument response function,
R(v ,1). A measurement which includes these effects may be written as:
) = Io(v).R(v,l) (Eq.9-1)
where: ^'(v ) = the modified IQ(V );
R(v ,1) = the instrument response function; and
1 = the path length.
R(v ,1) includes the contributions of the detector, the optics and any loss due to beam
divergence over an open-path. Since only the beam divergence is dependent on path
length, R(v ,1) may be written as:
R(v,l) =R(v>r(l) (Eq.9-2)
where: R(v ) = contains the frequency dependent responses; and
r(l) = the path length dependent divergence loss.
In similar fashion, the intensity measured when the sample is present, I(v ), may be
written as:
9-10
-------
I(v) = L(v) • R(v) • r(l) • t (v) (Eq. 9-3)
where: x (v) = the frequency dependent term due to attenuation of the light
beam by ambient species.
The transmittance may be calculated taking the ratio of I(v) to V(v) •
This yields:
(Eq. 9-4)
T = J£L = W ' R(V) 'I(1) ' T(V) = t(v)
Xo (V) ^^ ' ^^ ' ^
and cancelling like terms leaves only the term due to ambient species, -c (v).
IQ'(v) may be measured by conducting a short path measurement (i.e. a
path length of approximately 1 meter )31. This measurement contains the appropriate
instrument response function plus some contribution due to ambient species. A baseline
fit may then be made to the spectra which eliminates the contribution of the ambient
species. This baseline fit is then I<,'(v) for r(l) = 1. To correct for r(l), a second
measurement must be made and baseline fit to a spectrum collected over the path length
to be measured. The ratio of the short path baseline to the long path baseline is then
Sometimes, measurements of I0'(v ) are conducted over the full
measurement path and they include absorptions of ambient species. This measurement
is more correctly termed a background measurement, Ib(v). For Ib(v), equation 9-1 can
be written as:
(Eq. 9-5)
Ib (v) = I0(v) • T^V) = I0(v) - R(v) • r(l) • T^Cv)
where t ^(v ) is the frequency dependent term due to the attenuation of the light beam
by ambient species in "clean air". This measurement is generally conducted upwind of
the monitored site (or on some sample which is assumed to contain none of the species
of interest) on a path which is of the same length as that to be monitored. Here, the
measurement contains the ambient species (which are primarily water vapor, carbon
9-11
-------
monoxide, dinitrogen oxide, carbon dioxide, etc). This spectrum is then used to ratio
against the measured I(v ) spectra taken downwind of the site.
Measurement of Ib(v ) and its use as I0'(v) has the advantage that a long
path water vapor reference spectra is not needed since the water vapor component will
be cancelled if the spectra are taken near enough in time to one another. The
disadvantage to this method is that the spectra often do not completely cancel. This
incomplete cancellation is due to frequency shifts of the spectra arising from changes in
alignment or instabilities in the interferometer as well as changes in the ambient gas
concentrations.
The nature of the incomplete cancellation can be seen if the frequency and.
concentration sensitivities of each of the terms in equation 9-3 is examined. First, Io'(v)
and R(v ) are independent of ambient gas concentration while T (v) is not. Secondly,
each term may be classified as either a slowly or rapidly varying function of v. IQ'(V ),
and R(v ) are slowly varying functions of v while T (v ) is a rapidly varying function of v.
Thus, T (v) will be effected to a much greater extend by frequency shifts than will ^'(v)
or R(v ). Measurement of Io'(v) on the other hand is not sensitive to the spectral shifts
since it does not contain any rapidly varying functions of frequency. Use of ^'(v )
however, does require that a water vapor spectrum be used for quantitation.
The measurement of Io'(v )should be undertaken when a good long path
water vapor spectrum can be obtained. This method removes the severe problems
associated with spectral shifts or gas concentration variation and also limits the number
of upwind measurements that need to be made. This method also relies on the ability of
the quantitation method to remove water vapor from the spectrum.
If a quality, long path water vapor spectrum is not available, Ib(v) can be
measured. However, the spectral shifts and gas concentration must be fully accounted
for. In addition, any other species that are present in the Ib(v ) spectrum must be
removed to ensure that they will not affect the downwind measurements.
9-12
-------
9.4 TECHNOLOGY TRANSFER AND USER EDUCATION
Perhaps the largest stumbling block limiting the use of OPMs by the
general AAM community is a lack of knowledge of the system's capabilities. These
systems are not the final solution to all monitoring problems and needs, but should be
viewed as another tool in the array of monitoring options. While these systems have
some significant advantages over the conventional AAM systems, they also have their
disadvantages. The capabilities and limitations of these systems need to be clearly
understood so that they may be applied to appropriate monitoring scenarios.
Guidance documents such as the Open Path-FTIR7 manual under
preparation need to be completed along with manuals for other appropriate OPMs.
Standard operating procedures and specific performance criteria need to developed that
will allow consistent measurement of detection limits and consistent checks of the
operational status of the systems. Finally, there is an urgent need for standardized
quality assurance and quality control methods. Some attempts have been made in this
direction and the OP-FTIR document is expected to address this topic in some detail.
Technology and information transfer will play an important part if open
path monitoring approaches are to enjoy widespread, successful application at Superfund
sites. The greatest challenge will be to get the available technical information and
guidance to the decision makers and workers at each Superfund site. This effort will
require developmental experiments/studies and the preparation of guidance documents
based on the results of these studies. Furthermore, this information will need to be
packaged and distributed via workshops, announcements, project summaries, conference
presentations, and journal publications.
9-13
-------
SECTION 10
TECHNOLOGY UPDATES
Optical remote sensing is a rapidly developing field and the monitoring systems
considered here are only a small portion of the monitors available or becoming available.
Appendix B contains a descriptions of several ORS methods that are either available or
in the developmental phase for application to environmental monitoring. Some of these
methods are modifications of the methods described in Section 4.4. Also, some of these
methods will be used strictly as point monitors, while others can be used as either OPMs
or point monitors.
Application of new ORS technologies creates a need to keep users of monitoring
technologies up to date. To accomplish this, EPA is considering regularly updating, this
document via the distribution of brief inserts. The inserts would describe the new
technology, the compounds it can detect, its detection limits, and it applicability to
Superfund work. These updates could take the form of an insert that will be sent to
registered recipients of this document.
These updates would have the following suggested format:
• An introductory section describing the principles of operation of the
system, the compounds it could detect, whether it would function as a point
monitor or a path monitor, and its detection limits;
• A section on what advantages and capabilities this technology would bring
to Superfund monitoring. This would include information on time
response, the ability to detect metal or other compounds, areas where it
may be more sensitive than other monitors, etc.;
• A section describing which activities this system would be most useful for
and a description of its implementation.; and
• A section on the application of this technology and what needs to done to
evaluate its capabilities as a monitor for the Superfund program.
10-1
-------
The information for these updates would come predominantly from the literature or
instrument manufacturers, but some hands-on investigation may be warranted. The next
18 to 24 months is likely to see an exponential increase of new applications of available
and developmental technologies and officials in charge of implementing these methods
need accurate and up-to-date information on these technologies.
Within this same 18- to 24-month period, the abilities of the available ORS
systems are likely to increase. There is considerable interest in increasing the sensitivity
of the available OPMs, particularly the FTIR. This must be watched closely since there
are certainly gains to be made that may eliminate many of the detection limit concerns.
10-2
-------
SECTION 11
REFERENCES
1. US EPA. Procedures for Conducting Air Pathway Analysis for Superfund
Applications - Volume I, Applications of Air Pathway Analysis for Superfund
Activities. EPA-450/1-89-001, (1989).
2. US EPA. Procedures for Conducting Air Pathway Analysis for Superfund
Applications - Volume II, Estimation of Baseline Air Emissions at Superfund
Sites. EPA-450/1-89-002, (1990).
3. US EPA. Procedures for Conducting Air Pathway Analysis for Superfund
Applications - Volume III, Estimation of Air Emissions From Clean-up Activities
at Superfund Sites. EPA-450/1-89-001, (1989).
4. US EPA. Procedures for Conducting Air Pathway Analysis for Superfund
Applications - Volume IV, Procedures for Dispersion Modeling and Air
Monitoring for Superfund Air Pathway Analysis. EPA-450/1-89-004, (1989).
5. Padgett, J.; Pritchett, T. H., "Application of Open Path Monitors at Superfund
sites," Proc. SPIE 1433, 352-364, (1991).
6. Minnich, T. R.; Scotto, R. L.; Kagann, R. H.; Simpson, O. A., ."Optical Remote
Sensors Ready to tackle Superfund RCRA emissions monitoring tasks," Hazmat
World 3:42 (1990).
7. Minnich, T. R.; Scotto, R. L., " Open Path FTIR" in preparation.
8. Lucero, G.; Moertl, K., "Superfund Handbook, " 3rd Ed. ENSR Corp. (1989).
9. Skoog, D. A.; West, D. M., Principles of Instrumental Analysis, 2nd Ed. Saunders
Golden Sunburst Series, (1980).
10. Willard, H. H.; Merritt, L. L., Jr.; Dean, J. A.; Settle, F. A., Jr., Instrumental
Methods of Analysis, 6th Ed. D. Van Nostrand, New York, (1981).
11. Spellicy, R. L.; Crow, W. L.; Draves, J. A.; Bucholtz, W. F.; Herget, W. F.,
"Spectroscopic Remote Sensing Addressing Requirements of the Clean Air Act,"
Spectroscopy, 6, (1991).
12. Atkins, P. W., Molecular Quantum Mechanics, 2nd Ed. Oxford University Press,
New York, (1983).
13. van de Hulst, H. C., Light Scattering by Small Particles, Dover Publications Inc.,
New York, (1981).
11-1
-------
14. Thorne, A. P., Spectrophysics, 2nd Ed. Chapman and Hall, New York, (1988).
15. Csanady, G. T., Turbulent Diffusion in the Environment, Geophysics and
Astrophysics Monographs, 3, D. Reidel Publishing Company, Boston, (1980).
16. Gifford, F. A., "Consequences of Effluent Release," Nuclear Safety, Edited by R.
L. Sharp, 17, 68-86, (1976).
17. Minnich, T. R.; Kricks, R. J.; Solinski, P. J.; Pescatore, D. E.; Leo, M. R.,
"Determination of Site Specific Vertical Dispersion Coefficients In Support of Air
Monitoring at Lipari Landfill," AWMA/EPA International Symposium on
Measurement of Toxic and Related Air Pollutants," Durham, N.C., May (1991).
18. Roy F. Weston and Blasland Bouck & Lee, "VOC Emission Rates Derived from
FTIR Measurement Data During Pilot Scale Remediation Activities," Vol. I and
II, (1991).
19. Russwurm, G. M.; Kagann, R. H.; Simpson, O. A.; McClenny, W. A.; Herget, W.
F., "Long-path FTIR Measurements of Volatile Organic Compounds in an
Industrial Setting," J. Air Waste Manage. Assoc., 41, 1062-1066, (1991).
20. Grant, W. B.; Kagann, R. H.; McClenny, W. A. "Optical Remote Measurement of
Toxic Gases," J. Air Waste Manage. Assoc., 42, 18-30, (1992) and references
therein.
21. This was taken from page 24 of reference 2.
22. Eklund, B.; Smith, S.; Hendler, A. "Estimation From Excavation Activities at
Superfund Sites. EPA Contract No. 68-D1-0031, WA 13 (1991).
23. Draves, J. A.; Dayton, D-P.; Bursey, J. "Inovative Sensing Techniques for
Monitoring and Measuring Selected Dioxins, Furans, and Polycyclic Aromatic
Hydrocarbons in Stack Gas," prepared under EPA contract 68-D1-0010, WA 10
(1990).
24. See discussion present in Reference 2.
25. Russwurm, G. M.; Kagann, R. H.; Simpson, O. A.; McClenny, W. A., "Use of a
Fourier Transform Spectrometer as a Remote Sensor at Superfund Sites,"
Northrop Services Inc., EPA/600/D-91/115, (1991).
26. Grupp, D.; Rojek, G.; Bath, R. J.; Minnich, T. R.; Naman, R. M.; Brochu, A. J.;
Spear, R. D.* "The Pre-Remedial Air Toxic Program: A Case Study Using Remote
Sensing," AWMA/EPA International Symposium on Measurement of Toxic and
Related Air Pollutants," Durham, N.C., May (1991).
11-2
-------
27. Smith, R. A.; Jones, F. E.; Chasmar, R. P., The Detection and Measurement of
Infrared Radiation, Oxford, Clarendon Press, Oxford University, (1957).
28. Huheey, J. E., Inorganic Chemistry, 3rd Ed. Harper & Row, New York, (1981).
29. Carter, R.E., Jr.; Lane, D.D.; Marotz, G.A., A Field-Based Intercomparison of the
Qualitative and Quantitative Performance of Multiple Open-Path FTIR Systems for
Measurement of Selected Toxic Air Pollutants, U.S. EPA Region VH Draft, 1991.
30. . Minnich, T., Blasland, Bouck, & Lee, Private Communication.
31. Herget, W.F., Nicolet, Private Communication.
11-3
-------
APPENDIX A
THE LAMBERT-BEER LAW
-------
APPENDIX A
What follows is a brief description of the Lambert-Beer Law10'14 commonly
referred to as Beers Law. The Lambert-Beer Law is used to relate the concentration of
absorbers in media to the light attenuation caused by those absorbers. This discussion
will walk through a derivation of the form of the law (Equation 3-5) most commonly
used.
Consider a collimated beam of light incident on an absorbing medium as
shown in Figure A-l. The power absorbed by a slab of the medium between x and
x + dx is dP. The amount of power decreased per unit length is -dP/dx. In general,
light detection equipment is not suited for power determination, but for power per unit
area measurements. Power per unit area is defined as intensity or
p (A-l)
I = _L
A
Therefore the intensity decreased per unit length is -dl/dx. The amount of intensity
decreased is proportional to the incident intensity and this may be written as:
^•«"
dx
where k is the proportionality constant. Collecting like terms gives:
_,T,^ (A-3)
and integrating both sides yields:
(A-4)
(A-5)
-On Ix (v) - In I0 (v)] = kdx
A-l
-------
1(0)
x x + dx
Figure A-l. Light Incident on an Absorbing Medium.
A-2
-------
'x
0
Removing the logarithm gives the general distance dependent Lambert Law (A-7).
Lambert found that for a given concentration of absorbers the light intensity decreased
logarithmically with distance:
(A-7)
-f kdx
Is(v) = I0(v) e Ja
The integral:
(A-8)
fLkdx
Jo
is called the optical depth.
Beer is generally given credit for finding that increasing the number of
absorbers had the same effect as Lambert had observed, namely a logarithmic decease in
the light intensity. Therefore, the proportionality constant is related to the concentration
by:
k = aC (A-9)
where C is the concentration; a the new proportionality constant which is the molecular
absorption coefficient.
Combining this with Equation A-7 gives:
Dividing both sides by I0(v ), gives the transmittance, T.
_ , % . (A-ll)
T = W = e-/:«(v,x)-c(x)-dx
" I0(v) "
Taking the negative base ten logarithm of the transmittance gives the absorbance A(v),
or:
A-3
-------
I (v) ,L (A'12)
A(v) = -log,0T = log,0 -2— = 0.4343 J~ a(v,x) • C(x) •
-------
APPENDIX B
POTENTIAL OPTICAL REMOTE SENSING METHODS
SOURCE: Reference 23
-------
APPENDIX B
POTENTIAL OPTICAL REMOTE SENSING METHODS
Twelve technologies which are either available or under development for
application to environmental monitoring are:
Gas Filter Correlation (GFC) Spectroscopy
Fourier Transform Infrared (FTTR) Spectroscopy
Ultraviolet (UV-DOAS) Absorbance Spectroscopy
Matrix Isolation FUR (MI/FTIR)
Gas chromatography-MI/FnR (GC/MI/FITR)
Photoacoustic Spectroscopy (PAS)
Laser Induced Breakdown Spectroscopy (LIBS)
Laser Absorbance (Diode and Carbon Dioxide)
Laser Induced Fluorescence (LJF)
Multiphoton lonization (MPI)
Fluorescence
Shpol'skii Spectrometry (SS)
The following sections will briefly describe the principles by which each of
these techniques operate. The methods will also be categorizes according to their
spectral region of operation i.e. ultraviolet, infrared or consequence. A "consequence"
Spectroscopy method is one in which the absorption or emission of a photon is not
monitored but a result of that absorption (such as photo-dissociation) is monitored. The
methods will also be categorizes as either OPM's or point monitors.
B.1 Ultraviolet Methods
Ultraviolet (UV) spectroscopy operates on the principle of electronic
excitation of atoms or molecules. Thus, these methods are also sensitive to atomic
species such as metals. The absorption strengths for these excitation are typically much
larger than those in the infrared region. The region of the spectrum associated with this
type of spectroscopy is about 200 to 400 nm.
B.1.1 Ultraviolet (UV) Absorbance Spectroscopy
UV absorbance spectroscopy is conventional absorption spectroscopy using
the appropriate broad-band excitation source such as a Xenon lamp. The technique is
well developed and requires about 1 minute to identify the concentration of a single
species. Typically the UV has better sensitivity than the infrared because of the larger
UV band strengths but not all compounds are observable in this region.
B-l
-------
B.1.2 Fluorescence
This technique relies on the excitation of the molecule of interest to a
known state by the absorption of a photon from a broad band light source such as a Xe
arc. The photon is either emitted at the same frequency or more likely at a lower
frequency. The emitted photons frequency is monitored by a spectrometer. The response
times will generally be on the same order as those of the UV absorbance.
B.L3 Laser Induced Fluorescence (LIF)
This technique is conceptually similar to the more general fluorescence
described above but can be more species specific. LEF however uses a carefully tuned
laser to excite the molecule of interest This careful tuning allows for very specific
excitations and allows for better compound identification. As with the fluorescence
technique above the emitted photon can then be monitored as a function of frequency
with a spectrometer of appropriate resolution. This gives the ability to both selectively
excite and selectively detect the compounds of interest.
B.1.4 Shpol'skii Spectroscopy (SS)
The Shpol'skii effect is the result of a correlation between the molecule of
interest and its neighboring solvent to narrow the spectral features of the molecule.
Here, the molecule of interest is mixed with a carefully chosen alkane and frozen. The
alkane is chosen so as to mimic the carbon bound structure of the compounds. Once
frozen the sample may be interrogated by a laser and either absorption or emission
monitored. This technique has seen some success when applied to PAHs and good
spectra are available.
B.2 Infrared Methods
Infrared Spectroscopy relies on excitation of the vibrational-rotational
modes ( and some pure rotational modes ) of molecules. These infrared techniques are
sensitive to molecules only and will not detect metals or even a class of molecules known
as homonuclear diatomics ( e.g. Oj, Nj, Q^ etc ). The excitation strengths of these
transitions are generally 10 to 100 times weaker than the electronic transitions in the
UV. The spectral range is generally from about 200 to 4000 cm-1.
B-2
-------
B.2.1 Fourier Transform Infrared Spectroscopy (FTIR)
FTIR allows for the collection of the entire infrared spectrum from about
4000 to 200 wave numbers in a few seconds. This technique offers the ability to look at
all of the ambient air species which have infrared active bands. This method can«be
used in absorbance mode with the addition of an infrared source or in a passive
emissions mode. Multipassing optics such as a white cell can also be added which can
allow the FTIR to be used as an extractive white cell point sampler. In either case the
molecular signature is obtain over a broad spectral range from which accurate
identification of the species can be made. The advantage of taking the entire spectra is
that if interfering species are present they will almost certainly not be present at all of
the same frequencies throughout the IR and careful choice of analysis regions will allow
for interference free detection. Unfortunately FTIR is also frequently limited in
sensitivity because of the strengths of the IR bands available.
B.2.2 Matrix Isolation/Fourier Transform Infrared Spectroscopy (MI/FTIR)
This technique is an extension of the conventional FTIR described above.
Here, the species are frozen (at about four degrees Kelvin) in a matrix of nitrogen gas.
This freezing narrows the inherent spectral features and eliminates all excited state lines
allowing for substantial reduction of interferences from adjacent spectra, it allows for an
extended observation time, and for long signal averaging periods.
B.2.3 Gas Chromatography/Matrix Isolation/Fourier Transform Infrared
Spectroscopy (GC/MI/FTIR)
CG/MI/FTTR is a combination of the MI/FTIR technique and an FTIR
techniques. The addition of a gas chromatograph allows for the separation of species
prior to quantitation thus removing possible spectroscopic interferences. This technique
would be expected to have the lowest detection limits of the FTIR techniques. Like
FTER however, the method still suffers from the lower absorption strengths in the
infrared but does have the advantage of long time scans to improve the signal to noise
ratio.
B.2.4 Laser Absorption (Carbon Dioxide and Diode)
These two lasers operate in the infrared region of the spectrum and their
use would be similar to that of the FTIR in absorption mode. The two advantages here
are 1) a high degree of spectral resolution which will allow the molecules to be more
accurately identified over a smaller spectral range and 2) a larger signal to noise ratio
due to the intensity of the laser light source and the possible use of harmonic detection.
The main difference between a diode and CO2 laser is that the diode produces a beam
which is continuously tunable over approximately a fraction to 1 cm"1 while the CO2 laser
produces discrete lines which cannot be easily shifted. In detecting a gas concentration
the diode is chosen to coincide with an absorption of interest and swept across a feature
for identification. For the CO2 laser, lines must be chosen which inherently fall on or off
of a spectral feature of the compound(s) of interest. The latter method clearly must be a
B-3
-------
fortuitous, match the diodes can be forced. Laser absorption can be very sensitive and if
sufficient scanning is possible (or enough lines are used) it can be very specific.
B.2.5 Gas Filter Correlation (GFC) Spectroscopy
^
This technique is a non-dispersive infrared method which relies on the
determination of a species characteristic absorbance frequencies when compared to the
correlation a reference cell which contains a sample of the species of interest. This
technique, since it is non-dispersive, easily applied, is frequently applicable even in the
presence of interferences, and has a very fast (approximately one second) response time.
Unfortunately it is usually limited in sensitivity. The general working principle also
requires a vapor phase sample of the compound of interest
B.3 Consequence Spectroscopv
In consequence spectroscopy light intensity is not monitored but rather a
consequence of absorption. An example of a consequence spectroscopic phenomena is
photodissociation. Here, a molecule absorbs a photon which causes it to dissociate, the
dissociation product is then monitored by some technique such as mass spectroscopy. In
general, these techniques have low backgrounds and are therefore very sensitive to trace
constituent quantitation.
B.3.1 Photoacoustic Spectroscopy (PAS)
This technique relies on the transfer of energy from electronic, vibrational,
or rotational motions of the molecule to translational motion during a collision. The
molecules of interest are excited with a laser at a known absorption frequency. The
absorbed energy is then transferred to heat (i.e translational energy) through collisions
with other molecules and atoms. This heat warms the gas, which if contained in a closed
cell changes the pressure in the cell. The laser is usually modulated and the modulated
pressure changes are then monitored with a sensitive microphone or density detecting
interferometer. This method is very sensitive but must be used as a point sampler.
BJ.2 Laser Induced Breakdown Spectroscopy
This technique relies on the use of a very tightly focused laser beam.
Lasers are a source of very large electric fields that when focused to a small point can
strip electrons from molecules and force dissociation and frequently the formation of a
plasma. This dissociation and plasma formation results in energy emission thereby
creating its own excitation source. Nearby molecules will absorb this energy and later
re-emit light The emission of this light is monitored much in the same way as that of
fluorescence spectroscopy.
B-4
-------
BJJ Multiphoton lonization (MPI)
MPI relies on ionizing the species of interest using laser photons. In
general this requires the absorption of more than one photon thus the prefix "multi".
Specific species can be ionized if the multiple absorption exactly coincides with specific
internal states of the molecules. Here, instead of detecting light the ions are detected.
Like fluorescence this can be a very specific and very sensitive approach if lasers of
sufficient power and of appropriate wavelength are available.
B.4 Categorization of Methods
The techniques described in the previous sections are categorized for
convenience in Table B-l according to their spectral operating region (either UV, IR or
consequence). Additionally, Table B-2 lists the techniques as either OPM's or point
monitor. That is they are divided according to their ability to directly yield path-
weighted concentrations.
B-5
-------
Table B-l.
* Spectral Regions of the Optical Remote Sensing Methods
t • :;;: Ultraviolet/ ; . ;:; > ••
UV Absorbance
Fluorescence
LJF
SS
- -••.V;''^\;M£::joInfrared . ..:.
GC/MI/FITR
MI/FTIR
FTIR
Laser Absorbance
GFC
'."':-. Consequence
LIBS
PAS
MPI
Table B-2.
Open-Path and Point Monitoring Capability
... Method
OpenrPath Monitors
Point Monitors
UV Methods
UV Absorbance
Fluorescence
LJF
SS
X
X
X
X
X
X
X
IR Methods
GC/MI/FTIR
MI/FTIR
FTIR
Laser Absorption
GFC
X
X
X
X
X
X
X
X
Consequence Methods
PAS
LIBS
MPI
X
X
X
B-6
-------
APPENDIX C
GLOSSARY OF TERMS
-------
APPENDIX C
GLOSSARY OF TERMS
Attempts have been made recently to standardize -the terminology used in the field of
optical remote sensing. These preliminary compilations of terminology, while well
intentioned, are not completely consistent with the equivalent terminology used in fields
such as spectroscopy, physics, and astronomy. Consistent terminology is essential to
avoid miscommunication and confusion. The following glossary of terms was prepared
for this document; future revisions will be necessary as a consensus terminology is
developed by EPA and other researchers.
Absorbance ( A(v) )
Active System
Angstroms
Background Spectrum
Band Pass Filter
Bandwidth
Bistatic Configuration
The negative base ten logarithm of the transmittance.
While absorbance is related to optical depth or log
base of transmittance, the difference is a constant
factor of 0.434.
An optical remote sensor that interacts in an active
way with the species being monitored. For example,
absorption spectroscopy using a laser or broadband
source propogated through the sample.
A unit of length equivalent to 10~10 meters.
Commonly used to measure light wavelength in the
ultraviolet and visible regions of the electromagnetic
spectrum.
Collection of the intensity of the light beam as a
function of wavenumber or wavelength in the absence
of the absorbing molecules. This is ratioed to the
intensity observed in the presence of absorbers to
calculate the transmittance.
Optical filter that allows a selected region or band
from a light source to pass while restricting light of
both longer and shorter wavelength.
The width of a reported spectral feature (e.q.
bandpass). It is defined in terms of either the full
width at half the peak maximum (FWHM) or half the
width of the peak at half the peak maximum
(HWHM).
A measurement configuration using a transmitter and
receiver by some distance. For example, placement of
an optical transmitter and receiver at opposite ends of
a monitored path.
C-l
-------
Broad Band Source
Conventional Point Monitor
Differential Absorbance Lidar
Differential Optical Absorbance
Spectroscopy
Effective Concentration
Electromagnetic Spectrum
Emissions Flux
Emissions Spectroscopy
Electromagnetic radiation source that emits light over
a broad range of frequencies as opposed to a.
monochromatic or single frequency source. Broad
band sources are used in UV-DOAS, FTIR's and
GFCs.
(CPM)- point monitor such as a canister or a sorbent
tube, or other device commonly used as a point
monitor of gas concentrations, but not an ORS system
used as a point monitor.
(DIAL)- Lidar using two pulsed lasers (or two
wavelengths of a single laser), one tuned to an
absorbing frequency and the other tuned close but off
that absorbing frequency. The difference in the
detected backscatter signal between the two lasers is a
measure of the gas concentration and common
variations allow for correction of the signal for natural
attenuation losses.
(DOAS)- technique used to remove the effects of
scattering and continuum absorption so that
concentration may be obtained. This technique
involves the measurement of the difference between
absorption within a spectra feature and the wing of
that absorbing feature taken as non-absorbing.
Quantity derived from a path-weighted measurement
by dividing by the pathlength and has the equivalent
absorbance as if that concentration was
homogeneously distributed across the path. Expressed
in ppm, ppb, /zg/m3, molecules/cm3, etc.
(EM)- the total of all the possible frequencies of
electromagnetic radiation.
(J)-the measurement of the mass of emissions per
second per unit area crossing the beam path.
Knowledge of the area of a plane parallel to the beam
path which the light monitors allows determination of
the rate of emissions across the beam path.
The monitoring of light emitted (or lost) from a
species rather than light removed from a well
characterized beam of radiation by the species.
C-2
-------
Frequency
Fourier Transform Infrared
(FTIR) Interferometer
Gas Filer Correlation (GFC)
Spectrometer
Hertz
. Homonuclear diatomics
Infrared (IR) Region
The measurement in time of the crest of a sinusoidal
wave passing a particular point in space (see Figure C-
1). Numerically equal to the propagation velocity
divided by the wavelength.
An ORS system capable of open path and point
monitoring. Uses an interferometer and
mathematically manipulates the output to generate an
infrared spectrum. Typically identifies compounds by
using reference spectra. Can quantitate tens of
compounds in minutes.
A special case of correlation spectroscopy capable of
both open path and point monitoring. Uses
correlation of the spectrum measured from an
unknown to that of a reference cell to evaluate gas
concentration. Has a response time on the order of 1
second.
(Hz)-has units of reciprocal seconds. The unit of
frequency equal to the reciprocal of period or time
per cycle.
A class of molecules containing two identical atoms
(examples include C12, F2, N2, O2). Homonuclear
diatomics can not, under normal circumstances, be
detected by infrared radiation. The molecules can
absorb under high pressure conditions but then do so
in a structureless fashion.
Region of the electromagnetic spectrum from
approximately 1 pm (10,000 crn'1) to 1 mm (100 cm"1).
The subregion of the infrared used by an FTIR for
detection of the species typically encountered in
environmental monitoring applications is from
approximately 250 pm (400 cm"1) to 2.5 ^m (4000 cm"
Intensity
Interfere gram
The amount of power per unit area delivered to the
detector at a specific wavelength or wavenumber.
Direct output of an interferometer and is a complex
function caused by the interference of light.
C-3
-------
en
c
CD
/ +Amplitude
Wavelength
\
\
^Amplitude
-Amplitude
Time
Figure C-l. Sinusoidal Wave Pattern
C-4
-------
Interferometer
In-situ monitoring
Lambert-Beer Law
Laser
Monitoring systems
Monochromatic Source
Monostatic
Nanometer
Optomechanical device at the heart of the FTIR
Provides modulates light to produce a complicated
interference pattern caused by the constructive and
destructive interference of a light beam with itself.
Monitoring without the need of a sampling method
(such as a canister of sorbent tube) with real time or
near real time data turn around.
Often called Beer's Law, it governs the relation
between absorbance and concentration. For perfectly
resolved sepctra the law states that the absorbance is
linear in concentration. This is only approximately
true and significant deviation can be noticed for non-
perfectly resolved data. This law is.approximately true
for low concentrations, and for weak absorbancies.
An acronym for Light Amplification by Stimulated
Emission of Radiation. It is a device which can
produce very coherent and nearly monochromatic
radiation with very high intensities.
Any combination of sampling and analytical methods
applied to monitoring.
Light source which produces only a single frequency
or color of light A laser is a close approximation to
the monochromatic source.
A configuration in which the source and receiver are
effectively co-located. Co-location or near co-location
of the transmitter and receiver is possible for OPM by
using a reflecting optic such as a retroreflector at the
opposite end of the path. If source and receiver are
integrated into a single unit, it is typically referred to
as a transceiver.
Measure of length equivalent to 10"9 meters.
Generally used to refer to light wavelength in the
ultraviolet and visible regions of the electromagnetic
spectrum.
C-5
-------
Open Path Monitor
Optical Remote Sensor
Passive System
Path length
Path-weighted concentration
Point monitoring
Scattering
Sinusoidal Motion
Transceiver
Ultraviolet (UV) Region
(OPM) - any optical remote sensor configured so that
it monitors in the open air. No sample cell is used
and the ouput is usually a path-weighted
concentrations or effective concentration. Open path
monitoring may be conducted either indoors or out of
doors.
(ORS) - any optical instrument which is used for
detection of pollutants or meteorological parameters.
Subsets of ORS's include open path monitors, radars,
lidars, and spectroscopic point monitors (SPM's).
Optical remote sensing system monitors without
actively interacting with the sample. An example in
environmental monitoring are FTIR's used in emission
measurements without an artificial source.
The distance that the light travels from source to
detector.
The effective concentration times the pathlength
expressed in units of ppm'meters, ppb'meters, /*g/m2
or molecule/m2, etc. This nomenclature is consistent
with that used in physics and astronomy for
measurements of concentrations that are not
necessarily homogeneously distributed.
Monitoring at a specific point in space. Subsets of this
monitoring include conventional point samplers and
spectroscopic point samplers.
An elastic or inelastic deflection of a photon off of its
linear course. This type of scattering maybe
compound or size, or orientation specific.
Periodic motion of a wave often used to represent one
form of light. Represented mathematically by x = A
sin(2rc ft) where x is the position, A is the amplitude at
the crest of the wave f is the frequency and t is the
time. (See Figure C-l.)
An optical device that shares the same optics for both
beam launch and beam reception.
Region of the electromagnetic spectrum from
approximately 200 to 400 nanometers.
C-6
-------
Ultraviolet-Differential Optical
Absorbance Spectroscopy
Visible (Vis) Region
(UV-DOAS)- extrapolation of a typical laborsrory
ultraviolet absorbance technique to account for the
processes of Mie scattering and continuum absorption
which occur in the ambient air.
Region of the electromagnetic spectrum from
approximately 400 to 750 nanometers. Often grouped
with the ultraviolet region for spectroscopic purposes.
Wavelength
Waven umber
Measure of the distance between crests in a sinusoidal
wave pattern ( see Figure C-l). Typically has the
units of length expressed in micrometers for the
infrared region, nanometers or angstroms for the
visible and ultraviolet regions.
A quantity which is proportional to energy and has
units of reciprocal centimeters. Typically used when
referring to infrared spectra.
C-7
-------
APPENDIX D
TRANSECT METHOD
-------
APPENDIX D
TRANSECT METHOD2
The transect technique, also referred to as plume .mapping, measures the
concentration of the emitted species at several downwind locations aligned perpendicular
to the anticipated plume centerline. The in-depth transect technique is an indirect
emission measurement approach that has been used to measure fugitive paniculate and
gaseous emissions from area and line sources. This technique has been successfully
tested at a variety of waste sites, including landfills. Figure D-l illustrates the transect
sampling array.
The transect technique uses horizontal and vertical arrays of samplers to
measure concentrations of species within the effective cross-section of the emission
plume. The volatile species emission flux is then obtained by spatial integration of the
measured concentrations over the assumed plume area. For volatile species, the
emission flux is calculated as:
U
1 - 7- / / AP q (h,w) dhdw
where: J; = emission flux of component i (^g/nr-sec);
U = wind speed (m/sec);
A, = surface area of emitting source (nr);
Ap = effective cross-sectional area of the plume (m2);
Q = concentration of component i at point (h,w) corrected for
upwind background (/*g/m3);
h = vertical distance coordinate (m); and
w = horizontal distance coordinate (m).
For particulates, the emission flux is calculated as:
T 1 r r . m(h,w) „ ,
J = / I A —- ' on dw
^ A J J c f\
i AS a
where: J = emission flux (ug/nr-sec);
t = sampling time (sec);
AS = surface area of emitting source (m2);
Ap = effective cross-sectional area of plume (m2);
m = mass of particulates collected after correction for background
concentration Cug);
h = vertical distance coordinate (m);
w = horizontal distance coordinate (m); and
a = intake area of sampler (m2).
D-l
-------
Vliluai
Point Soutc*
Wind Oti«ctlon
Figure D-l. Example of transect technique sampling.
D-2
-------
The cross-sectional area of the source (AJ term can be eliminated from
both equations, if only the total site emission per time is required. An alternative
equation for volatile species, based on diffusion theory and measurements is:
n _
Q(t)j = S w X£, oy oz C U
1
where: Q(t)j = emission rate of species i (gm/sec);
Xj = peak concentration of species i (Gaussian Fit Curve);
KJ = conversion factor gm/ppm for species i;
oy = lateral extent of Gaussian plume;
oz = vertical extent of Gaussian plume;
C = instrument response factor;
ic = 3.141; and
— = mean wind speed.
All parameters are obtained from field measurements. (In some instances,
az is estimated from ay). As for the C-P technique, users should see the cited literature
for further guidance on the data reduction procedures. The key point to note is that the
equations are relatively complex. Furthermore, terms such as a y and a z will require
curve-fitting of the data and will typically have a large associated uncertainty.
The sampling equipment consists of a central 3.5-meter mast supporting
three equally spaced air sampling probes, and single wind direction, wind speed, and
temperature sensors at the top; and five 1.5 meter masts with single air sampling probes.
The central mast is aligned downwind along the expected plume center-line. Two masts
are placed on each side of the central mast perpendicular to the plume centerline at
equal spacings; and one mast is used to collect air samples at an upwind location. The
spacing of the associated masts is selected to cover the expected horizontal plume cross-
section, as defined by observation and/or profiling with real-time analyzers. Additional
sampling locations, both vertically and horizontally, can be added as required to provide
sufficient coverage of the plume cross-section. Prior to sample collection, meteorological
parameters must be monitored to determine if sampling conditions meet the
predetermined meteorological criteria.
The transect technique is somewhat less susceptible to changing
meteorological conditions than the concentration profile technique, but it does not
account for the vertical dispersion of the emitted species due to their varying molecular
weights. A more complex array of samplers can be employed to overcome this short-
coming, if necessary. The transect is often the preferred technique because the
technique is applicable to a variety of some types and the resulting data can be more
useful since the data are collected across the plume area.
D-3
-------
Applicability
The transect technique is applicable to emission rate measurement from all
forms of area sources, including lagoons, landfills, open dumps, and waste piles.
The technique can be used for both volatile and paniculate matter
emission rate assessment The technique is applicable both for undisturbed and
disturbed site conditions, and for testing of emission control techniques. The technique
is applicable to emissions measurement during all phases of the RI/FS process and can
therefore, provide directly comparable data throughout the process, including post
remediation monitoring.
While the method assumes a relatively homogeneous site and a well mixed
plume, these conditions are not necessarily required to use the method. The placement
of sufficient sampling stations across the plume can allow the technique to be used at a
heterogeneous site, or where the distance downwind for equipment set up is limited.
However, data collected under these conditions should be carefully evaluated before use.
Limitations
The technique requires that meteorological conditions during sampling,
particularly wind speed and direction, match the predetermined conditions used to select
the sampling locations. The center mast should be on the approximate plume centerline.
The technique may not adequately collect emissions from point sources within an area
source, such as cracks in landfill covers or vents, unless the plume is fairly well mixed at
the sampling locations. The technique provides only limited vertical profiling of the
plume. The technique is not applicable during quiescent or unstable wind conditions; it
may produce false negative results during these conditions.
Preferred Technique
The transect technique is a preferred indirect emission assessment
technique. The technique has been used for several different types of area sources and
is documented in the literature. The applicability of the technique, the conditions
required for sampling, and the moderate level of equipment and manpower needs
suggest this technique as a preferred technique relative to other indirect approaches.
D-4
-------
APPENDIX E
CASE EXAMPLES OF THE USE OF OPM SYSTEMS AT SUPERFUND SITES
-------
APPENDIX E
CASE EXAMPLES OF THE USE OF OPM SYSTEMS AT SUPERFUND SITES.
Over the past few years, OPMs have been employed at various Superfund sites
during various remediation activities. Some of the funding for such applications has
come from the Superfund Innovative Technology Evaluation (SITE) program. This
section briefly discusses the application of OPMs to five sites. An FTIR system was used
as the sole OPM for the majority of sites. One site, however attempted use of an FTIR,
UV-DOAS, and a LIDAR. The five sites and the OPM system(s) used are:
Site
New Castle, Delaware
Shavers Farm, Georgia
Middlesex County, New Jersey
Lipari Landfill, New Jersey
Abbeville, Louisiana
OPM System
FTIR
FTIR
FTIR, LIDAR, UV-DOAS
FTIR
FTIR
The discussion below covers some of the compounds of interests, the pathlengths
available, the goals of the monitoring and the phase of the remediation process during
which the OPM was used. The information may be useful for assessing analagous
situations of similar sites. Some of the configurations are well documented and the
official report documents provide additional information.
New Castle, Delaware - FTIR19
The site at New Castle, Delaware was a heavily industrialized area in which an
FTIR was evaluated under the Superfund Innovative Technology Evaluation (SITE)
program. The FTIR (was a Nicolet 730 spectrometer modified by Tecan Remote, Inc.)
was configured in a monostatic manner using a transceiver, a retro-reflector and an un-
modulated source. The FTIR and transceiver were mounted in a van so changes in wind
direction could be accommodated.
Two types of measurements were conducted. The first was a comparison of the
FTIR and a CPM method (i.e. canisters) at the industrial complex. These measurements
were conducted with a pathlength of 250 meters; the site imposed a physical limitation
that precluded longer pathlengths without multipassing. The second set of measurements
were conducted at a landfill that was several square miles in area. The pathlengths used
were 250 meters and 500 meters, depending on the monitoring goal.
E-l
-------
The authors concluded that the FTIR as configured was capable of meeting the
needs of some monitoring activities provided the concentration levels were above 50 ppb.
The authors make the claim that changing the optical system would improve the
detection limits but no specific changes were recommended. The change, as discussed
below, would be to reconfigure the FTIR so that source modulation could be conducted
before launch of the beam.
Shavers Farm, Georgia - FTIR25
The Shavers Farm site in northwestern Georgia was undergoing excavation
remedial activities and the compounds of interest were benzaldehyde and benzonitrile.
The FTIR system was (Nicolet Model 730 FTIR modified by MDA Scientific, Inc.)
configured in a monostatic method using a retro-reflector. This system however was
upgraded so that the light was modulated before launch. The FTIR was mounted in the
back of a truck. The path lengths were on the order of 500 meters.
Measurements were made on two paths each of 500 meters. The meteorological
conditions however were such that the FTIR was predominantly upwind of the
excavation and neither benzonitrile nor benzaldehyde were detected. Laboratory
detection limits for benzonitrile were reported as 186 ppm*meters in a 15 cm cell giving
a theoretical effective concentration detection limit of 372 ppb, i.e. significant emissions
would have to take place to be detected. However, the remedial activity was in its latter
stages and there were few "hits" even using canister sampling. The authors do make the
point however that at least the upper limit of the concentration present can be assigned.
The authors credit the modulation of the light source with significantly reducing
the background radiation and thereby decreasing the detection limits. A QA program
which was implemented on site was also described. While the program was specific to
their instrument, it does offer an approach to QA that can be modified if no specific QA
guidance is available.
Middlesex County, New Jersey • FTIR - Lidar - UV-DOAS26
This site saw the application of three OPMs to the pre-remediation phase. The
three systems were the FTIR (Nicolet 740 spectrometer) a UV-DOAS from the
University of Denver, and a carbon dioxide, differential absorbance Lidar (DIAL)
system. The monitoring took place over several days and several pathlengths. The
compounds of interest included toluene, benzene, chloroform, 1,2-dichIoroethane and
1,1,1 -trichloroethane.
E-2
-------
The UV-DOAS system was of the moncstatic variety with a transceiver and a
retroreflector. The UV-DOAS was mounted on a tripod and placed external to the van.
The DIAL system used two CO, lasers (ELS-Lasersafe). The first laser was tuned to a
specific absorption wavelength of a compound and the second tuned to a wavelength just
off that absorbance. The differential absorbance of the two wavelengths was used to
determine the concentration present.
Only one monitoring event showed the presence of any of the target compounds.
Toluene was detected by the FTIR at a path-weighted concentration of 20 to 23
ppm'meters. Surprisingly, the UV-DOAS, which has a lower-limit of detection, used
over nearly the same path detected no toluene. The 20-23 ppm*meters concentration is
below the detection limits of the DIAL system. The authors used various monitoring
locations to exploit the advantages of the systems within the goals of pre-remediation
phase and some of these configurations may shed some light on how to proceed under
similar scenarios.
Lapari Landfill, New Jersey - FTIR17
FTIR monitoring (using a MDA Scientific, Inc. Model 282000) was conducted at
the Lapari Landfill during the installation of extraction wells. These wells were part of a
pump and treat remedial operation. The FTIR was configured along the perimeter of
the site and path-weighted concentration measurements fed into a program to up date
emissions rates. The emissions rates were calculated using Turner's Gaussian Equation
as given in Section 4.4. The emissions rates were then used to calculate plume
centerline concentrations.
A tracer study was conducted to measure the vertical dispersion, oz. The tracer
gas used was SF6, which has a very strong ( 1 ppm*meter) infrared absorbance and is not
normally present .in the ambient air in significant concentrations. The dispersion was
first estimated using measurements of sigma theta and textbook stability class curves.
Next, the tracer gas was released at a known rate and oz calculated using equations
similar to those in Section 4.4. The tracer gas dispersion values did not agree with those
-obtained from the stability class estimates despite the measurement of sigma theta.
E-3
-------
The vertical dispersion, oz, must be known to estimate total dispersion. It can be
measured directly using LIDAR/SODAR, inferred from tracer gas studies, or estimated
from the lateral dispersion term, ay, and the atmospheric stability. Dispersion is an
important parameter in predicting (back calculating) source terms and subsequently
estimating downwind concentrations of emitted contaminants. If the dispersion is
overestimated by a factor of 2 then the calculated source term (emission rate) is
decreased by a factor of 2. Therefore, an accurate knowledge of ar (and ay) is important
to accurately assess air impacts. The authors therefore recommend use of the tracer gas
study to determine site specific vertical dispersion.
The issue of accurate dispersion estimates is particularly critical if the compounds
are below the detection limits of the monitoring system. This was the case at the Lapari
site as no compounds on the analyte list were detected. However, it was determined that
even at the detection limit, the emission rate of the compounds of interest was not
sufficient to exceed action levels. Therefore, OPM did provide useful information at this
site.
Abbeville, Louisiana - FTIR18
The FTIR (MDA Scientific, Inc. Model 282000) was applied during a pilot scale
excavation and dewatering operation to monitor emissions of benzene, toluene,
ethylbenzene and xylenes emissions. Several pathlengths were used depending on
specific monitoring locations. Tracer studies were carried out by using carbon
tetrafluoride (CF4) which has a detection limit of .03 ppm*meters.
During the pilot-scale operation, no target compounds were detected by the FTIR.
However, octane and iso-octane were being emitted in detectable concentrations. These
compounds were then used as surrogates. The target compounds were expected to
follow the meteorological behavior of the surrogates and analysis proceeded accordingly.
Emissions of methane were also detected.
This study demonstrated the use of surrogate compound detection for emission
rate assessment. A tracer gas study was conducted to determine the emission source
strength (see Equation 4-18). This relation was found to be an acceptable approach for
determining emission rates provided the source exhibits ground-level, non^buoyant
behavior.
E-4
-------
APPENDIX F
OPEN PATH MONITOR VENDORS
-------
Appendix F
Open Path Monitor Vendors
VENDOR
Air Instruments and Measurements
ANARAD
Bomem
Denver University
Insitec Measurement Systems
Mattson Instruments, Inc.
MDA Scientific, Inc.
MIDAC Corporation
Nicolet
Opsis Inc.
Optech Inc.
Simens (U. K.)
PHONE NUMBER
(818) 813-1460
(805) 963-6583
(418) 877-2944
(510) 837-1330
(608) 831-5515
(800) 323-2000
(714) 645-4096
(608) 271-3333
(203) 698-1810
(416) 661-5904
44 202 782000
------- |