EPA/600/A-96/085
Dynamic Spiking Studies Using the DNPH Sampling Train
Jocttc L. Steger,
Radian Corporation, P, 0. Box 13000, Research Triangle Parte, N C. 27709
Joseph E. Knoll,
National Exposure Research Laboratory, U. S. Environmental Protection Agency
Research Triangle Park, N.C. 27711
The proposed aldehyde and ketone sampling method using aqueous 2,4-dimtrophenylhydrazine (DNPH) was
evaluated in the laboratory and in the field. Hie sampling trains studied were based on the train described in SW-846
Method 0011, Nine compounds were evaluated: formaldehyde, acetaldehyde, quinone, acrolein, propionaldehyde,
methyl isobutyl ketone, metby ethyl ketone, acetophencme, and isophorone. In the laboratory, the trains were spiked
both statically and dynamically. Laboratory studies indicated that formaldehyde and isophorone are efficiently
recovered from the first impinger. Laboratory studies also investigated potential interferences to the method Based cm
their potential to hydrolyze in acidic solution to form formaldehyde, dimethylolurea, saligenin, s-trioxane,
hexamethlyenetetramine, and paraformaldehyde were investigated. Dimethylolurea, hexamethylenetetramine, and
paraformaldehyde all interfered with formaldehyde analysis. The sampling train containing 200 mL of reagent in the
first impinger followed by two impingers containing 100 mL of reagent was then evaluated at a plywood veneer
manufacturing plant Ten runs were performed using quadruplicate sampling trains. Two of the four trains were
dynamically spiked with the nine aldehydes and ketones. The test results were evaluated using the EPA Method 301
criteria for method precision (< ± 50% relative standard deviation) and bias (correction facte* of 1.00± 0.30).
Formaldehyde, acetaldehyde, propionaldehyde, and acetophenone passed the requirements for accuracy and precision.
INTRODUCTION
Radian, while assisting the Method Branch of the National Exposure Research Laboratoiy (NERL), has evaluated a
multiple pollutant sampling and analytical method for aldehydes and ketones in gaseous emissions from stationary
sources. This study is part of an EPA program to develop stationary source emission test methods for the 189
hazardous air pollutants listed in Title M of the Clean Air Act Amendments of 1990. Test methods for these analytes
are needed to determine risk to the public and to support the regulatory process.
The sampling method in the present study employs an impinger train containing a saturated solution of acidified 2,4-
dinitrophenylhydrazine (DNPH) to capture and derivatize aldehyde and ketone compounds. The proposed method
was first evaluated in the laboratory; then the method was evaluated in the field at a plywood veneer dryer vent at a
pressboard manufacturing plant
The method was evaluated for precision and bias using procedures described in EPA Method 3011 in which bias is
determined by spiking sampling trains and precision is determined by collocating sampling trains. Spiking was
carried out by a dynamic method in which measured quantities of analyte were introduced into the flue gas being
sampled for die duration of the sampling run.
INTERFERENCE STUDIES
Because of the low pH reagent used, a potential exists for some compounds to be collected in the impingers and
hydrolyzed to form formaldehyde, biasing the results. Five compounds with the potential to hydrolyze under acidic
conditions to form formaldehyde were studied using duplicate aliquots of DNPH at pH 4. Blank DNPH was used as a
control. The results are reported in Table 1. Saligenin and s-trioxane did not interfere under the conditions tested.
Dimethylolurea created a slight interference and hexamethylenetetramine and paraformaldehyde significantly interfere
with the determination of formaldehyde.
1

-------
SPIKING STUDIES
Two approaches were considered for spiking an aqueous solution of the nine compounds. Static spiking of an aqueous
solution, and dynamic spiking of an aqueous solution using a syringe pump. The dynamic and static spiking
procedures were compared in the laboratory. Two trains were spiked statically by directly adding the solution to the
first impingex. Another two trains were spiked dynamically using a syringe pump. For quality control purposes, a
reference spike and method blank sample woe also analyzed.
Table 2 compares the average results for static and dynamic spiking. Regardless of the spiking procedure used, 5% or
less of the recovered formaldehyde was found in the second impinger, indicating that the spiking procedure does not
significantly affect the results obtained for formaldehyde. Although total recoveries few acetaldehyde were equivalent
by the two spiking methods when dynamic spiking was used, 30% of the recovered acetaldehyde was present in the
second impinger versus only 4% when static spiking was used Thus, although the spiking procedure does not affect
the overall performance of the train in recovering acetaldehyde, it does affect any conclusions regarding breakthrough
of acetaldehyde.
For acrolein, propionaldehyde, MIBK, and isopboroM, the total recoveries were less with dynamic spiking than with
static spiking and significant quantities of the recovered compounds were found in the second impinger. For these
compounds, static spiking would overestimate the performance of the train and could lead to false conclusions that the
sampling procedure is adequate for these compounds when in reality significant quantities of the compound would not
be recovered. Because different results were obtained with some of the compounds when dynamic spiking was used
and dynamic spiking is more representative of what occurs in an actual sampling situation, dynamic spiking was used
for the remaining studies.
LABORATORY TRAIN STUDIES
In the initial laboratory evaluation studies, low recoveries of the spiked carbooyl compounds from the sampling trains
were obtained. The effect of the amount of reagent in the first impinger and impinger temperature were evaluated in
the laboratory before the first field test to determine if spiked compound recoveries could be improved by changing
these parameters. The effect of reagent volume was studied to determine if the low recoveries were caused by
insufficient reagent rather than by the compounds breaking through the train. The impinger temperature was studied
to see if raising the impinger temperature would increase the rate of the derivatization reaction and improve recoveries
of methyl ethyl ketone (MEK) and methyl isobutyl ketone (MIBK), which appeared to be breaking through the train.
Four trains were dynamically spiked with 14 mg of each carbonyl compound. For two of the trains, the first impinger
contained 100 mL of reagent and was kept in an ice bath during the entire sampling period. For the other two trains,
the first impinger contained 200 mL of reagent. For one of these trains the first impinger was kept at room
temperature during sampling and for the other train the impinger was maintained in an ice bath. For quality control
purposes, a reference spike and method blank sample were also analyzed.
Results
Results for comparison of the amount of reagent in the first impinger are reported in Table 3. Recoveries based on the
concentration of the spiking solution and volume of solution spiked improved for all of the compounds except
quinone when the volume ofreagent in the first impinger was increased from 100 to 200 mL. The recovery for
isophorone quadrupled. Recoveries for formaldehyde, acetaldehyde, propionaldehyde, MEK, and MIBK doubled.
The recoveries for acrolein and acetophenone increased by 40 and 30%, respectively. Thus, for sampling levels of
aldehydes and ketones above 10 ppmv, using 200 mL of reagent in the first impinger appears necessary to obtain
quantitative recoveries.
Results for comparison of the temperature of the first impinger reagent solution are presented in Table 4. Recoveries
based on the concentration of the spiking solution and volume of solution spiked were above 70% in the first impinger
for formaldehyde and acetophenone regardless of whether the impinger was kept warm or cold. Recoveries and
breakthrough into the second impinger were unaffected by impinger temperature for acetophenone, formaldehyde, and
2

-------
quinone. For isophorone, the recoveries were unaffected by impinger temperature but the breakthrough into the
second impinger was lower when the impingers were kept cold. For acetaldehyde and propionaldehyde the recoveries
were higher and the breakthrough was less when the impingers were kept cold. For acrolein, cold impingers resulted
in slightly better recoveries. In addition, less tautomer formed in the cold impingers (6% versus 16% in the warm
impingers). For MEK and MIBK, the two cold impingers recovered more compound. Interestingly, the breakthrough
of MEK and MIBK into the second impinger was also higher when the impingers were cold. In general, for all of the
compounds, higher recoveries, with less breakthrough and greater compound stability, are achieved when the
impingers are kept cold. Based on the laboratory evaluation, the field validation study used iced impingers with 200
mL of reagent in the first impinger.
FIELD EVALUATION
For the field test, flue gas samples for carbonyl analysis were collected at a pressboard manufacturing plant from a
plywood veneer dryer stack. The sampling ports were 6-inch (152 mm) diameter pipe nipples located approximately
2.6 meters (m) above the sampling platform. The ports were located at least 4 stack diameters downstream and 1
stack diameter upstream of the nearest flow disturbances.
Gas was extracted from the source duct through a glass nozzle/probe system as shown in Figure 1. Four collocated
sampling trains, referred to as a "quad-train", were used to collect four independent samples simultaneously under
identical stack conditions. Two of the trains were dynamically spiked with an aqueous solution of aldehydes and
ketones at a level from two to five times higher than the levels measured in the source gas. Flue gas samples were
collected isokinetically from a single sampling point identified from a preliminary velocity traverse. The
aldehyde/ketone sample was recovered from the train in two fractions.
The first and second impinger contents, water and methylene chloride rinses from the nozzle/probe liner and first and
second impingers were Fraction 1. The contents and methylene chloride rinses from the third and fourth impingers
were Fraction 2. The two fractions were analyzed separately. Before statistical analysis, all compound quantities
from the analytical reports were normalized using the gas volume sampled by each train.
Results
According to Method 301, methods that have bias correction factors between 0.70 and 1.30 are acceptable when the
precision of the method at the level of the emission standard is less than or equal to 50% relative standard deviation.
Method 301 assumes that the bias and standard deviation of the data remain constant over the test period.
Formaldehyde, acetaldehyde, and acetophenone were evaluated two ways: using the data from the first two impingers
only and using the data from all four impingers. Methyl ethyl ketone, methyl isobutyl ketone, acrolein,
propionaldehyde, isophorone, and quinone were evaluated using the data from all four impingers only. Precision and
bias calculations were completed using all four impingers because of the high breakthrough that occurred during
Runs 3,4,6 and 7. Two-impinger data also was reported to demonstrate that formaldehyde and acetophenone met the
acceptance criteria for bias and precision with only two impingers. Tables 5 and 6 show results of the statistical
evaluation of the test data for the spiked compounds using Method 301 using four impingers and two impingers,
respectively.
Bias and Precision
Formaldehyde and acetophenone showed acceptable precision and bias with two impingers. There was significant
breakthrough of acetaldehyde and propionaldehyde into the third impinger in four of the ten runs. The breakthrough
on these runs prevented the method from being validated for two reagent impingers. No satisfactory explanation for
the high breakthrough in the four runs was discovered. Acrolein, quinone, MEK, and MIBK did not meet bias criteria.
In addition, MEK also did not meet precision criteria for the unspiked samples. Low levels of MEK are challenging to
identify because low levels of other four-carbon carbonyl compounds can interfere with the identification and
quantification of MEK by HPLC. Isophorone failed the precision criteria for the unspiked samples. One of the
unspiked samples contained approximately 200 jig of isophorone while all the other unspiked samples contained
40 ng or less. Because the native isophorone is very low, isophorone performed acceptably using this method.
3

-------
CONCLUSIONS AND RECOMMENDATIONS
Based on the laboratory and field evaluation of the aldehydes and ketones, the following conclusions may be drawn:
•	Based on Method 301 criteria, the sampling method with 200 mL of DNPH reagent in the first impinger and 100
mL of DNPH reagent is the second impinger performs acceptably for the determination of formaldehyde,
isophorone, and acetophenone. The method with 200 mL of reagent in the first impinger followed by two
impingers with 100 mL of reagent performs acceptably for acetaldehyde and propionaldehyde.
•	The test method is not appropriate for the measurement of quinone, acrolein, MEK, and MIBK, due either to
poor collection efficiency or analytical problems.
•	Dimethylolurea, hexamethylenetetramine, and paraformaldehyde interfere with the determination of
formaldehyde.
Based on work performed in the laboratory and in the field evaluation, the following recommendations are made:
•	The sampling and analytical method tested is recommended for adoption as a standard EPA method for the
determination of formaldehyde, acetophenone, isophorone, acetaldehyde, and propionaldehyde emissions from
stationary sources.
•	To obtain quantitative recoveries of formaldehyde, acetophenone, isophorone, acetaldehyde, and
propionaldehyde, use 200 mL of DNPH in the first impinger followed by two impingers containing 100 mL of
DNPH and keep the impingers iced.
•	Alternate pollutant-specific methods should be developed for measuring emissions of quinone, acrolein, MEK,
and MIBK from stationary sources.
DISCLAIMER
The information in this document has been funded wholly or in part by the United States Environmental Protection
Agency under Contracts 68-D1-G010 to Radian Corporation. It has been subjected to Agency review and approved
for publication. Mention of trade names or commercial products does not constitute endorsement or recommendation
for use.
ACKNOWLEDGEMENTS
The authors wish to thank the pressboard manufacturing facility and the contributions of the following individuals to
the success of these evaluations: Merrill Jackson, Mike Hartman, Joan Bursey, Danny Harrison, Mark Owens, Phyllis
O'Hara, Donna Tedder, and Linh Nguyen.
REFERENCES
1. U.S. Environmental Protection Agency. Method 301—Field Validation of Pollutant Measurement Methods
from Various Waste Media, in Code of Federal Regulations, Title 40, Part 63, Appendix A. Washington,
D.C. Office of the Federal Register, July 1,1987.
4

-------
Table 1. Results of Interference Study at pH 4.0
Formaldehyde Measured
Sample 1	Sample 2
Interferant
Area
Bias Gig)
Area
Bias (re)
Dimethylolurea
88277
+6.4
82328
+5.6
Hexamethylenetetramine
331391
+36
382432
+42
Paraformaldehyde
315908
+34
534753
+61
SaUgenin
ND
0
ND
0
s-Trioxane
ND
0
ND
0
ND - Not Detected
Table 2. Comparison of Dynamic and Static Spike Train Recoveries Using pH 0 Reagent and
Spiking at a Nominal 1.4 mg for Each Compound
Percent of Total Percent Recovered
Percent Recovered (based on Recovered in Second (based on reference
spike amount)	Impinger	spike)
Compound
Reference
Spike
Static
Trains*
Dynamic
Trains*
Static
Trains*
Dynamic
Trains*
Static
Trains*
Dynamic
Trains*
Formaldehyde
74
84
197b
<1
5
114
266b
Acetaldehyde
82
80
76
4
30
97
92
Quinone
25
BQL
BQL
NA
NA
BQL
BQL
Acrolein
41
48
34
0
20
116
83
Propionaldehyde
70
68
56
<2
31
96
81
Methyl Ethyl Ketone
91
32
26
4
58
35
29
Acetophenone
171
136
164
0
14
80
96
Methyl Isobutyl
Ketone
67
56
18
0
63
83
27
Isopherone
86
80
68
6
16
92
78
•Average of two trials
^Results biased high due to contamination of the spiking apparatus with methanol (methanol contains traces of formaldehyde).
BQL - Below the Quantitation Limit
NA ~ Not Applicable.
5

-------
Table 3. Spike Train Recoveries Using pH 0 Reagent and Spiking at a Nominal 14 mg for Each Compound
Percent of Spike Recovered (baaed on spiking solution concentration)
200 mLIn
100 mL In Fint Impinger	 p|nt
Reference	Impinger	%
Compound
Spike"
TralnS
Train 6
Meank
(Train 12)
Difference*
Difference'
Formaldehyde
74
45.5
53.8
49.6
106
+56.4
114
Acetaldehyde
82
27.0
37.9
32.4
61.8
+29.4
90.7
Quinone
25
50.5
57.9
54.2
54.5
40.3
0.6
Acrolein
41
30.1
39.9
35.0
49.9
+14.9
42.6
Propionaldehydc
70
24.3
33.7
29.0
59.9
+30.9
107
Methyl Ethyl Ketone
91
4.57
6.88
5.72
13.0
+7.28
127
Acetophenone
171
34.4
49.4
41.9
54.7
+12.8
30.5
Methyl Isobutyl Ketone
67
5.26
8.88
7.07
14.6
+7.53
107
Isophorone
86
15.4
14.0
14.7
79.9
+65.2
444
•Reference Spike " Static liquid ipike into 200" mL of reagent
'Mean - (Train 5 + Train 6) / 2.
Difference - Train 12 - Mean.
"^/•Difference - Difference / Mean X100.
Table 4. Spike Train Recoveries Using pH 0 Reagent and Spiking at a Nominal 14 mg for Each Compound
Percent of Spike Recovered (based on spiking solution concentration)
Implngen at Room Temperature (Train 13)	 	Implngen In Ice Bath (Train 12)
Compound
Impinger
1
Impinger
2
Total
First Impinger
Breakthrough* (%)
Impinger
1
Impinger
2
Total
First Impinger
Breakthrough* (%)
Formaldehyde
95.9
2.9
98.80
2.94
106
2.5
108.5
2.3
Acetaldehyde
33.2
14.1
47.30
29.81
61,8
14.2
76.0
18.7
Quinone
55.3
2.2
57.50
3.83
54.5
1.7
56.2
3.0
Acrolein
40.3
0.2
40,50
0.49
49.9
0.6
50.5
1.2
Propionaldehyde
42.5
13.4
55.90
23.97
59.9
14.7
74.6
19.7
Methyl Ethyl Ketone
4.4
3.7
8.10
45.68
13.0
16.9
29.9
56.5
Acetophenone
52.7
13.1
65.80
19.91
54.7
11.6
66.3
17.5
Methyl Isobutyl Ketone
6.2
6.1
12.30
49.59
14.6
19.2
33.8
56.8
Isoohorone
74.5
15.7
90.20
17.41
79.9


11.3
•First Impinger Breakthrough
- Impinger 2 / Total x 100.

-------
Table 5. EPA Method 301 Statistical Evaluation Results Using Four Impingers
Parameter
Form-
aldehyde
Acet-
aldehyde
Quinone
Acrolein
Propion-
aldehyde
Aceto-
phenone
Methyl
Ethyl
Ketone
Methyl
Isobutyl
Ketone
Isophorone
RSD Spiked (%)
7.36
7.18
40.0
12.1
7.20
7.94
26.1
17.2
7.94
RSD Unspiked (%)
10.2
10.6
39.7
17.3
21.0
42.5
74.3
32.2
211
Bias CF
1.11
1.26
1.84
2.00
1.25
1.08
2.55
2.22
1.08
Disposition
Pass
Pass
Fail
Fail
Pass
Pass
Fail
Fail
Fail
RSD = Relative Standard Deviation
CF = Correction Factor
Table 6. EPA Method 301 Statistical Evaluation Results Using Two Impingers
Parameter
Formaldehyde
Acetaldehyde
Acetophenone
RSD Spiked (%)
RSD Unspiked (%)
Bias CF
Disposition
7.32
9.95
1.10
Pass
8.15
10.3
1.34
Fails
7.79
43.5
1.11
Pass
RSD = Relative Standard Deviation
CF = Correction Factor

-------
Vacuum
Lra
k»
Balh
200 ml 100ml 100 ml
ONPH DNPH ONPH
Pump
Figure 1. Sampling Trmin for Aldehydes and Ketones
8

-------
1. REPORT NO.
EP A/600/A-96/085
2 .
3.recipie in mi ii nun mini mm in
PB97- 1228 7 3
4. TITLE AND SUBTITLE
Dynamic Spiking Studies
Using the DNPH Sampling Train
5.REPORT DATE
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Joette L. Steger* and Joseph E. Knollb
8.PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
a.	Radian Corporation, P.O. Box 13000,
Research Triangle Park, NC 27709
b.	National Exposure Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-D4-0022
12. SPONSORING AGENCY NAME AND ADDRESS
National Exposure Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
13.TYPE OF REPORT AND PERIOD COVERED
Project Report
14. SPONSORING AGENCY CODE
EPA/600/09
TECHNICAL REPORT DATA
15. SUPPLEMENTARY NOTES
16. ABSTRACT
The proposed aldehyde and ketone sampling method using aqueous
2,4-dinitrophenylhydrazine (DNPH) was evaluated in the laboratory and in the field.
The sampling trains studied were based on the train described in SW 846 Method 0011.
Nine compounds were evaluated: formaldehyde, acetaldehyde, quinone, acrolein,
propionaldeyde, methyl isobutyl ketone, methyl ethyl ketone, acetophenone, and
isophorone. In the laboratory, the trains were spiked both statically and
dynamically. Laboratory studies indicated that formaldehyde and isophorone are
efficiently recovered from the first impinger. Laboratory studies also investigated
potential interferences to the method. Based on their potential to hydrolyze in
acid solution to form formaldehyde, dimethylolurea, saligenin, s-trioxane,
hexamethylenetetramine, and paraformaldehyde were investigated. Dimethylolurea,
hexamethylenetetramine, and paraformaldehyde all interfered with formaldehyde
analysis. The sampling train containing 200 ml of reagent in the first impinger
followed by two impingers containing 100 mL of reagent was then evaluated at a
plywood veneer manufacturing plant. Ten runs were performed using quadruplicate
sampling trains. Two of the four trains were dynamically spiked with the nine
aldehydes and ketones. The test results were evaluated using the EPA method 301
criteria for method precision (<±50% relative standard deviation) and bias
(correction factor of 1.00±0.30). Formaldehyde, acetaldehyde, propionaldehyde, and
acetophenone passed the requirements for accuracy and precision.	
17.	KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
b.IDENTIFIERS/ OPEN ENDED
TERMS
c.COSATI



18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This
Report)
UNCLASSIFIED
21.NO. OF PAGES

20. SECURITY CLASS (This
Page)
UNCLASSIFIED
22. PRICE

-------