Edgar R. Stephens and Monty A. Price Statewide Air Pollution Research Center University of Colifornio
Analysis of
(m
important Air
Pollutant: Peroxyacetyl Nitrate
The peroxyacyl nitrates (PANs) are a family of unstable, highly oxidized organic nitrogen compounds which are formed in polluted air by the photochemical action of sunlight on hydrocarbons and nitrogen oxides ( I ) . The general formula is 0
II
RCOON02
These compounds are extremely toxic to vegetation. Their characteristic damage symptoms on plants have been recognized in Southern California for more than twenty years although the nature of the toxic agent was not recognized until the early 1960's. The first member of the family, in which R is a methyl group, has received most of the research effort. This is called peroxyacetyl nitrate (PAN) and is capable of inducing the characteristic bronzing or silvering on plant leaves which is observed in field grown plants (2). The juxtaposition of the nitro group with the peroxide link provides many possibilities for chemical reactions, including reactions with biologically important materials such as enzymes (3).A few tens of parts per billion (by volume) of PAN are present in photochemical smog, concentrations which can injure sensitive plants in a few hours exposure. Higher homologs, with R = ethyl (peroxypropionyl nitrate, PPN), R = propyl (peroxy-n-hutyryl nitrate, PnBN), and R = phenyl (peroxyhenzoyl nitrate, PBzN) are also toxic hut they have been studied to a lesser extent. The PANs not only damage plants; they are powerful eye irritants as well. PBzN is reported to cause eye irritation a t a concentration of about ten pph (4). This makes it about two hundred times as irritating as formaldehyde and about one hundred times as irritating as PAN (5). The PANs are, as one would expect from their formula, not only highly reactive hut unstahle as well. Several serious explosions of liquid PAN have occurred during experimental work with this compound (6). Even when explosion does not occur PAN sometimes decomposes on standing either thermally or perhaps by hydrolysis catalyzed by traces of alkaline materials (7). In spite of these difficulties, methods of handling PAN have been worked out including procedures for analyzing pph levels of the compound in ambient air (8, 9). This analysis and the special procedure which have been devised for calibration with this unstable compound are the subject of this paper. Since it is so unstahle, no supply house can provide pure PAN for sale. Aside from the explosion hazard, the instability of PAN makes it unwise to ship or to store calihration standards. Gas chromatography with electron capture detection provides a very sensitive and selective method for the measurement of PAN in polluted air. At UCR an automated chromatograph has been sampling the atmosphere every fifteen minutes for several years and providing a PAN measurement. Concentrations below 10 pph are routinely measured this way. Manual operation - ~ the f chromatograph has also been adapted as a laboratory exercise in a course on instrumental analysis.
PAN is prepared by photolyzing ethyl nitrite vapor (CzH50NO) in an atmosphere of oxygen, a procedure which produces acetaldehyde and methyl and ethyl nitrate (RONOz) as major by-products. T o prepare PAN in quantity for use in plant fumigations or other toxicity tests the ethyl nitrite-oxygen mixture is photolyzed in a flow system. The product mixture is collected and chromatographed to ohtain pure PAN (10). For the calibration of a gas chromatograph this purification step is unnecessary and only small quantities of PAN are needed. They can he prepared by photolyzing trace amounts of ethyl nitrite vapor (a very volatile substance) in oxygen. This is done in a borosilicate glass infrared ahsorbtion cell of 10-cm path. A few hundred ppm of PAN are generated in this way and the concentration is calculated from the spectrum. Samples from the infrared cell must he accurately diluted by a factor of about 10,000 to ohtain samples in the pph level for calibration of the chromatograph. Electron Capture Chromatography The electron capture detector uses a source of heta rays (energetic electrons from nickel 63 or from tritium imhedded in titanium) to ionize the carrier gas.
One heta will generate millions of electrons (11). The electrons have very high mobility in the electric field set up by a polarizing voltage applied to the cell. They migrate to the positive electrode. The nitrogen molecules are neutralized at the negative electrode. This leads' to a standing current through the detector. An electron ahsorbing compound (e.g., PAN) absorbs some of these electrons
The negative PAN ion is much less mobile than the free electron so it can recombine with a positive nitrogen ion PAN-
+ Nz+ -PAN* + Nz
The presence of PAN thus leads to a decrease in the standing current through the detector. The energy rich hut neutral PAN* molecule then decomposes into neutral species. Two corollaries follo* from this: (1) Sufficient concentrations of electron absorbing compounds, for example, the oxygen in an air sample, will absorb all the electrons and reduce the standing current to zero. (2) The plot of peak height versus concentration is linear only so long as the peak height is a small portion (25-35% or less) of the standing current. This factor leads to the unusual problem that concentrations above 1 ppm even in a vapor sample may be inconveniently large! In this case a smaller sample can he used or the sample can he quantitatively diluted. PAN Generation and Infrared Measurement Of the several known synthesis routes to PAN, the photolysis of ethyl nitrite in oxygen is the most suitable for Volume 50, Number 5, May 1973 /
351
preparation of small quantities for calibration purpases. The mechanism of this reaction is not worked out in any detail b u t two features are of practical importance 1) The reaction must he conducted in a dilute mixture to obtain a good yield of product. 2) Acetaldehyde and methyl and ethyl nitrate are produced as by-products. These should he recognized among the product but they need not he removed to measure the PAN. T h e infrared spectrum of PAN has five strong a n d distinctive bands. In the pure state (in air diluent) these appear a s shown in Figure 1. The absorptivities of these bands have been measured carefully to provide the data of Table 1 (12). Each band can be used t o calculate the concentration of PAN according t o the well-known BeerLambert law.
~.oh, m
, , ,
,,,,, ,
4000 3000 2000
1500 PAN
icrn-')
VAPOR
,
,
I
,
I
1000 900 800
,
l j 700
SPECTRUM
Figure 1. infrared spectrum of pure PAN vapor [nitrogen diluent) in a 10-cm cell (blank a n empty cell)
C = logla(llT)/ab = Alab C = concentration (ppm by'volume) a = absorptivity (ppm-'m-'1 b =oath leneth (m) T = iransmiitance A = absorbance (optical density)
T h e average calculated for the spectrum in Figure 1 is 2650 ppm. When operating properly a Perkin-Elmer model 137B infrared spectrometer yields spectra in which a n absorbance a s small as 0.005 can be detected. Such a weak absorption can b e seen just below the baseline a t 8.63 p m in Figure 1. T h e trace of PAN was vaporized from the cell window and stopcock grease after t h e cell was purged and it was estimated t o he 25 ppm. Even this concentration is about three orders of magnitude too high for the electron capture detector! T h e ' infrared spectrum, once determined accurately a s in Figure 1 a n d Table 1, providesaprimary standard for PAN which can be used with any spectrometer. It bas been used for vapor samples with a 0.1 m (10 cm) p a t h a n d a t hundreds of meters p a t h in a multiple reflection cell (13).The spectrum should be run a t 1 a t m tokal pressure to be comparable to the reference spectrum. For maximum convenience the photolysis of ethyl nitrite is conducted in the absorption cell itself using ultraviolet light in the range 300-4M) n m wavelength. T h e PAN need not he separated from the by-products. Samples from the infrared cell can be diluted to concentrations in the pph range and injected into t h e chromatograph. Equipment and Materials 1) A gas chromatograph fitted with a gas sample valve with a 2 or 3-ml sample loop, an electron capture detector with suitable power supply, electrometer amplifier, and strip chart recorder (see Fig. 2). 2) A column made bv filline an 18-in. leneth of '1s-in. diameter
Table 1. Infrared Absorptivities of Pan ( A s Vapor in Air at One Atmosphere) (Micrometers1
Concentrstion(ppm1 Averape 2650 ppm Figure 2
Absorbance Concentration (ppml A&"
""-
available. The column is kept at 25°C. Other temperatures and flow rates have been used for PAN analysis, hut the temperature should not be increased very much as PAN decomposition will occur. If the temperature is reduced too much, moisture may condense on the detector and cause a short. 3) A carrier gas supply of cylinder nitrogen with a suitable pressure regulator. A flaw of 60 ml/min through the column should he established. A bubble meter and stopwatch are needed ta measure the flow. 4) A dilution system consisting of a suitable large inert container such as a stainless steel aviator's breathing oxygen tank (about 35 liters volume) and a syringe for the injection of PAN snmnles. 5) A syringe for transferring samples of ambient air into the gas sample value. A 100-ml unlubricated medical syringe is suitable.
.~~~~~
Synthesis of PAN Calibration Mixture 1) An infrared spectraphotometer which will accept a 10-cm gas cell. Only the 8.6 Gm (1160 cm-l) region is absolutely neces-
hut a broader region is desirable (2.5-15 rm). A PerkinElmer 137B was used for the work described here. 2) A 10-cm infrared gas cell with an ultraviolet transparent body (Borosilicate glass). This is a standard accessory. It should have two stopcocks. 3) A small syringe (50-100 GI)with a needle for injection through the stopcock bore. 4) Cylinder oxygen with suitable regulator. 5) Ethyl nitrite-The commercial product (Mallinckrodt ~5002) contains 8-1595 ethyl alcohol. This need not be removed. 6) A source of ultraviolet radiation to generate PAN in the ethyl nitrite/oxygen mixture. The radiation needs to he strong between 300 and 400 nm. Commercial blacklight fluorescent lamps are suitable. They come in various wattages and use standard sockets, ballasts, and starters. They peak at 350 nm and do not emit below 300 nm; there is no need to go to the additional expense of filtering out the visible radiation. Small, low-cost hand lamps sold for minerology can be used. Universal Minerallight Model UVSL-25, a 4-W lamp, was found suitable. The sun may be satisfactory, although this has not been verified. sary
Chromatographic Measurement
Figure I Absorbance
Figure 2. Automated electron capture chromatograph showing timer, solenoid, and gas samplevalve on the rlght.
5.44
5.76
7.68
8.60
10.74
12.61
0.25 2500
0.61 Zi90
0.28 24MI
0.37 2660
0.051 2830
0.29
.
0.10 457
~llkylnitrsteinterference.
352 /Journal of Chemical Education
.'
0.06 432
2850
0.045 437
Chromatographic Procedure After the column has been prepared and purged of volatiles it is connected to the detector and the flow rate adjusted to 60 ml/min
NITROGEN
": m Q
.6
1.0 4 0m 00
3000 2000
1500
(cm-1)
IOW 900
800
700
IMPURE PAN Figure 5, Infrared spectrum of impure PAN prepared by ultraviolet phot01ySi~oIethyl nitrite in a 10-cm cell.
1 - 1
GAS SAMPLE VALVE Figure 3. Chromatograph equipped with gas sample valve and loop and electron capture detector.
'.
2 m l SAMPLE 6 5 ppb PAN
I
1
AIR
AIR
AIR
I
I
0
RECORDER
DEFLECTION
x 32
Figure 6. Chromatogram of calibration sample showing PAN along with ethyl and methyl nitrates. Z
0_
+
u
u
[L
0
W
W
W
W
+
+
PAN
PAN
t
i
3
+ a
A
PAN
:
t
3
pppJ
Table 2. Comparison of P a n a n d N-Propyl Nitrate P e a k Heights a t Various Standing Currents Defector
Cell
voltage
Current-
Peak Heiehte
PAN (24ppb)
nPN (25ppb)
Ratio
I
I
2
0
0
I
I
2
0
M I NUTES Figure 4. Three chromatograms of polluted ambient air showing about 37 ppb of PAN (November 1.1970.5PM)
using a bubble meter and stopwatch. See Figure 3. The flow rate should be the same with the sample valve in either position. The current through the detector should then be measured for various cell voltages. An operating voltage near the inflection point in the curve should then be chosen. About 50 V are usually used. Ambient air samples can be analyzed by filling the 100-ml syringe (outdoors) and then attaching it to the gas sample valve and purging it through the valve. With the conditions given the emergence time of PAN is about 13/r min. See Figure 6. It is preceded by an air peak (oxygen) which saturates the detector. Figure 4 shows chromatograms from three succewive injections of ambient air using a 9-in. column and automatic sample injection. The identification of the PAN peak should be verified by passing the sample through a 1% solution of KOH before injecting it into the chromatograph. This may be done by placing a "U" tube containing a small amount of the base between the syringe and the injection valve and passing the sample through the "U" tube and through the valve, taking care not t o get base into the valve. The PAN in the sample will be hydrolyzed and be missing from the chromatogram. Preparation of Calibration Mixture
The first step in the preparation of a PAN calibration mixture is t o flush the 10-cm infrared cell with oxygen. Then small portions of ethyl nitrite vapor are added by means of a 50.~1syringe every 15 min with the cell being exposed to ultraviolet radiation between injections. The syringe is fitted with a long needle which is inserted into the head space of the ethyl nitrite bottle while the plunger is drawn, then the needle is inserted into the cell through the bore of the stopcock on the cell. An oblique bore stopcock would require the use of a flexible needle ar delivery tube on the
'vslueain chart divisionstimer sttenuatian. Current loweredwith contaminsfed deteetoreoll
syringe. If the vapor in the ethyl nitrite bottle were pure ethyl nitrite, then 50 pl injected into a cell of 140 ml volume would produce a concentration of 350 ppm. Although ethyl nitrite is very volatile (b.p. = 17°C) not all of the sample will be ethyl nitrite since the bottle is kept cold and the reagent contains some ethyl alcohol. The production of PAN can be monitored during irradiation by periodically scanning the infrared spectrum. After about 11/~ hr of ultraviolet irradiation (with additions of ethyl nitrite every 15 min) a spectrum like that of Figure 5 should be obtained. The five strong PAN bands of Figure 1 are evident along with bands a t 6.0, 9.8, and 11.6 pm due t o methyl end ethyl nitrate by-products. These also contribute to the 7.7 pm band. The concentrations calculated from the three strongest bands are given in Table 1. The average is 442 ppm. The alkyl nitrate cancentration is estimated to be 400 ppm. This PAN sample a t 442 ppm must be diluted 10,WO fold to provide a suitable sample for electron capture chromatography. A war-surplus breathing oxygen tank bas a volume of about 35 1 50 that syringe injection of 3.5 ml of the PAN mixture from the cell will provide such a dilution. The diluent should be air of normal humidity so that the chromatogram will be a realistic one. The air far dilution can be freed of PAN and alkyl nitrates by passing it through a filter of activated carbon. This flushing of the tank and the injection of the PAN can most conveniently be carried out a t atmospheric pressure. After allowing time for mixing, samples can be withdrawn with a syringe and pushed through the sample loop of the chromatograph. Injection of this sample yields a chromatogram like that of Figure 6. Lower concentrations can be reached by pressurizing the tank to a measured absolute pressure. The new concentration can then be calculated by multiplying by the pressure ratio. As a rough calibration the response of the detector t o n-propyl nitrate (nPN) can be compared to that of PAN. This compound is much more stable than PAN and can be purchased from laboratory supply houses. By serial dilution of measured quantities of Volume 50, Number 5, May 1973 / 353
the liquid into purified sir standards in the ppb range can he prepared. A comparison of the response of a particular detector to PAN and to n.pmpyl ,,itrate at various cell voltages is in Table 2, The peak heights of PAN and nPN were nearly the same even over a considerable range of c'd voltages even when the detector was so contaminated that the current was sharply reduced. If nPN is used as a substitute for PAN as the orimarv calibration of an instrument, there is no way to verify that the is being quantitatively delivered to the detector.
PAN
Literature Cited 111 Stephens, E. R.. "Advances in Environ. Sci. and Tech.." (Editor: Pith, James N.. Jr. and Metealf, Robert L.) h h n w i b y & sons. yorh, ,969, vol. I, p. 119146.
354
/ Journal of Chemical Education
(21 Taylor, 0.C..J.AirPoll. Cont. Assoc., ~ 9 1 5 1 . w11969). I31 Mudd. J . B . ; J B i d Chem., 241,4077l19661. I41 Heus, J. M..andGlesson, W. A..Enu. Sci. andTech. 2112). l l W ( l 9 6 9 ) @ I Steohens. E. R.. Dsrley. E. F..Taylor, 0.C., and Scott, W. E., Pmr. Am. Pefml. Inst. Sarl. 111, 40,325 i19601;Inrem. J A i r WoterPoll., 4.79l1961). 161 stephens. E. R.. B U , I ~ S O ~ F. , R.and H Z ,I.C~IW, K. M..I I POLI confAsroc., 19l41,281 119691. I71 s t e p h e n s . ~ . ~ . , ~ t mnui o s .iron.. 1.19119671. I81 Darlo~,E.F..Kettner,K.A.,andStephons,E.R..AwL Chsm.. 35.589l1961). I91 smith.^. ~ . , est ~ ~ e o i t h k sbe i. . MI). ~ ( 1 9 7 1 ) . I101 Stephens. E. R., Burloson. F. R., and Cardiff E. A . J Air Poll. Cont. Assoc., 15, 87 (19651. I111 L + x k . J . E . . A n d Cham.. 35.474(19631. 1121 Stcphens.E.R.,Anol. Chsm.. 36,928119641., I131 Stephens. E. R.. InfnzmdPhvsic8. 1.181 11961)
