AMS column and into a second thermal conductivity detector.
A Model CR501 Chromatopac integrator was used to de-
termine peak areas.
Detection of Mn. Mixtures of AT and δ-MnO2 were
incubated at 30 °C to mimic a standard experimental run.
At 4, 12, 24, 36, 50, 75, 100, and 160 h, triplicate samples (300
mg each) of these mixtures were extracted with 3 mL of
methanol to quench the reaction. The suspension then was
filtered, and the solid was extracted again by shaking
overnight with 10 mL of 0.03 M HCl/ 0.03 M H2SO4 mixture
[modified Mehlich-1 method (20)]. Methanol was evaporated,
and the residue was combined with the acidified extract for
analysis. Untreated δ-MnO2 was used as a control. Manganese
analysis in the extracts was performed on a Perkin-Elmer
Model 4100 ZL graphite furnace atomic absorption spec-
trophotometer.
FIGURE 1. Reaction scheme for MnO -catalyzed atrazine dealhy-
2
lation. Secondary reactions with O are indicated by dotted lines.
2
UV Resonance Ram an Spectroscopy. UVRR spectra were
obtained with a 1.26 m single-stage spectrometer (SPEX 1269)
with an intensified diode array multichannel detector (21).
Laser excitation at 229 nm was provided by a frequency-
doubled Innova 300 argon ion laser (Coherent). The laser
power at the sample was about 0.35 mW. The scattered light
was collected with an f/ 1 paraboloid mirror in backscattering
geometry and focused with an f-matching lens. The reaction
extracts were contained in a 5 mm quartz EPR tube which
was spun around a stationary helical wire. Cooled N2 gas
kept the sample temperature between 10 and 15 °C. Spectral
acquisition was carried out in 5 min increments. No
photoproducts were observed under these conditions, as
verified by comparison of the UVRR spectra of the same
sample at 5 s intervals.
A least-squares method for spectral analysis (Labcalc) was
employed to quantify the products. Methanol, which was
used as the extraction solvent, has a very strong band at 1034
cm-1, well-separated from the Raman bands of the AT
degradation products. This band was used as an internal
intensity standard. After the Raman intensity was calibrated,
the UVRR spectrum of methanol was subtracted from the
experimental spectra digitally. The calibration file for the
quantification was generated from the spectral matrix of
standard solutions of pure AT, DEA, DIA, and DDA as well
as their mixtures. The least-squares fit was generated for 12
points, corresponding to peak wavenumbers: 900, 914, 930,
962, 1050, 1368, 1417, 1517, 1526, 1582, and 1614 cm-1. The
uncertainty in the resulting concentration estimates was
within (5%.
Birnessite peaks appeared at 0.720, 0.361, and 0.142 nm, in
good agreement with the results of Jones and Milne (19) for
samples of this mineral. The surface area was determined to
be 39 m2/ g similar to that reported by McKenzie (18), using
the BET single point method (Quantachrome Monosorb
analyzer). The water content of the MnO2 was determined
to be 10 wt %, from the weight loss upon heating at 110 °C
for 5 h.
Exactly 300 mg (3.5 mmol) of δ-MnO2 were suspended in
3 mL of anhydrous, analytical reagent-grade ethyl ether.
Analytical reagent-grade AT (Chem service, Westchester, PA)
was dissolved in ethyl ether to make a 10 mM stock solution.
As in our previous study (17), the basic experiment involves
deposition of AT on δ-MnO2 from diethyl ether solution,
low-temperature evaporation of the solvent, and extraction
of the sample with methanol, after appropriate incubation
intervals at 30 °C. The methanol extracts were examined by
UVRR spectroscopy to evaluate dealkylation yields. Volatile
products were determined by GC analysis of headspace gas
above the samples. The experiment was conducted under
N2 as well as in air. As a control, the same experiment was
carried out with alumina instead of δ-MnO2.
Approximately 300 µL of AT stock solution was added to
the δ-MnO2 suspension to provide initially 3.0 µmol AT for
reaction. The suspension was placed in a shaker table for 10
min, after which the solid was then placed in a vacuum
desiccator at 0 °C for 3 h. Ethyl ether was removed rapidly
under these conditions permitting subsequent investigation
of the surface kinetics at 30 °C over a 336 h period. When
anaerobic conditions were required, the δ-MnO2-AT sample
mixture was prepared in a Coy Environmental Chamber (Coy
Lab. Inc., MI) after introducing N2 gas and 5% N2/ H2 gas
mixture. Alumina was used as a desiccant to remove moisture
inside the chamber. H2 gas and a Pd catalyst were used to
remove O2 impurities caused by diffusion during normal
operating procedures.
After 0, 4, 8, 12, 24, 48, 72, and 96 h reaction at 30 °C in
the absence or presence of O2, the samples were extracted
with methanol (extracted twice in succession, with vigorous
mixing for 5 min). Quantitation of AT, DEA, DIA, and DDA
in the methanol sample extracts was by UVRR spectroscopy.
Duplicate samples of the δ-MnO2-AT mixture in 2-mL
screw top septum vials were prepared in the absence or
presence of O2, incubated at 30 °C in an oven to mimic the
UVRR experiments, and then analyzed for reaction products
other than AT, DEA, DIA, or DDA. For these reactions, the
vial headspace gas (100 µL aliquot) was analyzed after 0 (at
laboratory temperature and 30 °C), 24, 48, 72, 96, 120, 144,
168, and 336 h for products such as ethylene, propylene,
acetone, and acetaldehyde, using a Shimadzu CR501 series
gas chromatograph. Gas samples were passed through a 10
m Alltech (0.18 µm ID) QC345 fused-silica capillary column
into a thermal conductivity detector. Outflow from the
detector was then passed through a 15 m, 80-100 mesh 5
Results
Figure 2 shows UVRR spectra (excitation wavelength, 229
nm) of AT, DEA, DIA, and DDA at 1 mM concentration in
methanol. Atrazine and its dealkylation products absorb
strongly in the ultraviolet region, and the π-π* electronic
transitions resonantly enhance a series of vibrational modes
of the aromatic ring which are sensitive to alkyl substituents
(Figure 1). The enhancement pattern differs markedly among
AT and its dealkylation products, and the overall enhance-
ment decreases systematically with dealkylation: the 229
nm excitation wavelength is further from resonance as the
absorption band maximum in methanol shifts from 223 nm
for AT, to 216 nm for DIA and DEA, and to 210 nm for DDA.
The two monodealkylated products, DEA and DIA, give
similar UVRR spectra, but there are clear spectral differences
in the 900-1000 cm-1 region that permit reliable quantitation.
For example separable peaks are seen at 962, 900, 930, and
914 cm-1 for AT, DEA, DIA, and DDA, respectively. Moreover,
there are numerous differences in band intensities through-
out the spectra. The ability to resolve spectral components
made UVRR the preferred method of reaction monitoring,
because of the difficulty of resolving DEA and DIA peaks via
HPLC (17).
9
VOL. 33, NO. 18, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 3 1 6 1