Routes of photocatalytic destruction of chemical warfare agent simulants

Article abstract of DOI:10.1039/b109837c

Selected imitants of chemical warfare agents such as dimethyl methylphosphonate (DMMP), diethyl phosphoramidate (DEPA), pinacolyl methylphosphonate (PMP), butylaminoethanethiol (BAET) were subjected to photocatalytic and sonophotocatalytic treatment in aqueous suspensions of TiO₂. Complete conversion of the same mass of imitants to inorganic products was obtained within 600 min for DMMP, DEPA, PMP, but required a longer time for BAET. Sonolysis accelerated photodegradation of DMMP. No degradation was observed without ultraviolet illumination. Final products of degradation were PO₄^3-, CO₂ for DMMP and PMP, PO₄^3-, NO₃^- (25%), NH₄⁺ (75%), CO₂ for DEPA, and SO₄^2-, NH₄⁺, CO₂ for BAET. The number of main detected intermediate products increases in the order DMMP (7), DEPA (9), PMP (21), and exceeds 34 for BAET. Degradation of DMMP mainly proceeds through consecutive oxidation of methoxy groups and then the methyl group. Dimethyl hydroxymethylphosphonate and dimethylphosphate testify to the parallel oxidation of the methyl group. Destruction of DEPA mainly starts with cleavage of the P-NH₂ bond to form diethyl phosphate, which transforms further into ethyl phosphate. Oxidation of α and β carbons of ethoxy groups to form ethylphosphonoamidate, hydroxyethyl ethylphosphonoamidate and other products also contributes to the destruction. Photocatalytic degradation of PMP mainly starts with oxidation of the pinacolyl fragment, methylphosphonic acid and acetone being the major products. Oxidation of BAET begins with dark dimerization to disulfide, which undergoes oxidation of sulfur forming sulfinic and sulfonic acids as well as oxidation of carbons to form butanal, aminobutane, etc., and cyclic products such as 2-propylthiazole. A scheme of degradation was proposed for DMMP and DEPA, and starting routes for PMP and BAET. Quantum efficiencies of complete mineralization calculated as reaction rate to photon flux ratio approximate 10^-3%.

Full text of DOI:10.1039/b109837c

View Article Online / Journal Homepage / Table of Contents for this issue
Routes of photocatalytic destruction of chemical warfare agent
simulants
Alexandre V. Vorontsov,a Lev Davydov,b Ettireddy P. Reddy,b Claude Lion,c Eugenii N. Savinova
and Panagiotis G. Smirniotis*b
a
Boreskov Institute of Catalysis Novosibirsk 630090, Russia
Chemical Engineering Department, University of Cincinnati, Cincinnati, OH 45221-0171 USA.
b
E-mail: panagiotis.smirniotis@uc.edu
Institut de Topologie et de Dynamique des Systemes de I’Universite Paris-7 (CNRS UPRESA
c
`
7086), 1, rue Guy-de-la-Brosse, 75005 Paris, France
´
Received (in New Haven, CT, USA ) 29th October 2001, Accepted 22nd January 2002
First published as an Advance Article on the web
Selected imitants of chemical warfare agents such as dimethyl methylphosphonate (DMMP), diethyl
phosphoramidate (DEPA), pinacolyl methylphosphonate (PMP), butylaminoethanethiol (BAET) were
subjected to photocatalytic and sonophotocatalytic treatment in aqueous suspensions of TiO2 . Complete
conversion of the same mass of imitants to inorganic products was obtained within 600 min for DMMP,
DEPA, PMP, but required a longer time for BAET. Sonolysis accelerated photodegradation of DMMP. No
degradation was observed without ultraviolet illumination. Final products of degradation were PO43ꢀ, CO2 for
DMMP and PMP, PO43ꢀ, NO3ꢀ (25%), NH4+ (75%), CO2 for DEPA, and SO42ꢀ, NH4+, CO2 for BAET. The
number of main detected intermediate products increases in the order DMMP (7), DEPA (9), PMP (21), and
exceeds 34 for BAET. Degradation of DMMP mainly proceeds through consecutive oxidation of methoxy
groups and then the methyl group. Dimethyl hydroxymethylphosphonate and dimethylphosphate testify to the
parallel oxidation of the methyl group. Destruction of DEPA mainly starts with cleavage of the P–NH2 bond to
form diethyl phosphate, which transforms further into ethyl phosphate. Oxidation of a and b carbons of ethoxy
groups to form ethylphosphonoamidate, hydroxyethyl ethylphosphonoamidate and other products also
contributes to the destruction. Photocatalytic degradation of PMP mainly starts with oxidation of the pinacolyl
fragment, methylphosphonic acid and acetone being the major products. Oxidation of BAET begins with dark
dimerization to disulfide, which undergoes oxidation of sulfur forming sulfinic and sulfonic acids as well as
oxidation of carbons to form butanal, aminobutane, etc., and cyclic products such as 2-propylthiazole. A
scheme of degradation was proposed for DMMP and DEPA, and starting routes for PMP and BAET.
Quantum efficiencies of complete mineralization calculated as reaction rate to photon flux ratio
approximate 10ꢀ3%.
Chemical warfare agents (CWA) are the acting force of chemi-
cal weapons. International treaties prohibit further production
of CWA and the existing stockpiles are to be irreversibly
destroyed. Stockpiles of CWA as well as rockets, mines, bombs
and projectiles containing CWA are currently destroyed by
various techniques. In order to be destroyed, CWA and war-
heads usually have to be transported to the facilities. Trans-
portation poses additional risk for accidents, especially in
densely populated areas.
Besides incineration, which is commonly used in destruction
of organics, there are several other prospective methods of
destruction of CWA that are considered in short below.1–3
Formulas of common CWA are presented in Table 1. These
molecules can be converted into less toxic species by nucleo-
philic substitution with hydroxyl anions or monoethanolamine
(HOCH2CH2NH2), destruction with aqueous hypochlorite or
chlorine gas as well as with reagent oxone (mixture of KHSO5 ,
KHSO4 , and K2SO4), and peroxodisulfate (S2O82ꢀ). The
recently developed reagent magnesium monoperphthalate
affords complete detoxification of CWA under optimized con-
ditions.3 The common disadvantage of neutralization reactions
is that organic products have to be destructed further, either by
incineration or by other methods. Oxidative reactions lead to
complete mineralization of CWA; however, a big excess of
oxidizer is required. Biodegradation is another promising
approach. It suffers, however, from slow rates and it cannot
be applied directly to mustard gas, which destroys bacteria
cells.
Photocatalytic destruction is promising as a one-step miner-
alization process for CWA. Complete mineralization by
photocatalytic oxidation has been demonstrated for dimethyl
methylphosphonate and diethyl methylphosphonate in the
liquid4,5 and gas phases.6 Toxic pesticides such as methamido-
phos, malathion, diazinon, phorate,7 the insecticides dichlor-
vos and propoxur,8 the dye sulforhodamine B,9 which can
also be considered as imitants of CWA, were mineralized in
an illuminated suspension of TiO2 . The drawback of photoca-
talysis is the same as for biodegradation: low rate of minerali-
zation. Johnston and Hocking10 have demonstrated that
ultrasonic irradiation of TiO2 suspensions results in a marked
increase of the photocatalytic degradation rate. We apply this
approach to photocatalytic destruction of CWA simulants in
aqueous suspension of TiO2 as well.
The real CWA are deadly toxic liquids. Therefore, research
on CWA is usually conducted with their less toxic structural
analogs. We chose the least toxic compounds with chemical
bonds characteristic of CWA (Table 1). Four of these simu-
lants—dimethyl methylphosphonate (DMMP), pinacolyl
732
New J. Chem., 2002, 26, 732–744
DOI: 10.1039/b109837c
This journal is # The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2002

View Article Online
Table 1 Chemical warfare agents and chosen simulants
Simulant properties
CWA formula
Sarin (GB)
Simulant formula
Molecular weight/g molꢀ1
Density/g cmꢀ3
1.161
Boiling point/ꢂC
180–181
Dimethyl methylphosphonate (DMMP)
124.08
180.19
153.12
133.26
Soman (GD)
Tabun (GA)
VX
Pinacolyl methylphosphonate (PMP)
Diethyl phosphoramidate (DEPA)
2-(Butylamino)ethanethiol (BAET)
1.032
96–106 at
pressure 8 mm
–
140 at
pressure 3 mm
0.901
112–115 at
pressure 10 mm
Mustard gas or Yperite (HD)
Diethyl sulfide (DES)
90.19
0.835
1.070
92
2-Chloroethyl ethylsulfide (CEES)
124.63
156–157
methylphosphonate
(PMP),
diethyl
phosphoramidate
Procedures and analyses
(DEPA), 2-(butylamino)ethanethiol (BAET)—are used in the
current study as simulants for sarin, soman, tabun, and VX,
respectively. The simulants for mustard gas in Table 1 were
not used in this study because they are relatively volatile and
are removed from the aqueous suspension during oxygen bub-
bling. However, their photocatalytic degradation is success-
fully performed in the gas phase.11 DMMP possesses the P–
C and P–O–C bonds met in GB and GD, whereas PMP has
a structure identical to that of GD except that the fluorine
atom of GD is substituted by an OH group. The P–F bond
of GB and GD hydrolyses in water relatively quickly and even
if we used real GD after a short period of time we would have
PMP in solution. DEPA contains the P–N bond met in tabun.
BAET possesses C–S and C–N bonds and imitates the big radi-
cal of VX attached to the phosphorus atom.
The sonophotocatalytic setup used for CWA simulants degra-
dation is shown in Fig. 1. The cylindrical Pyrex reactor body
carries a jacket for circulation of fluid from a thermostat. Oxy-
gen was bubbled through a nozzle at the bottom of the reactor
with a flow rate of 100 cm3 minꢀ1. Light from a 1000 W Xe
lamp installed in an Oriel illuminator is directed to the top part
of the reactor using a mirror. The flux of UV light entering the
Experimental
Materials
Diethyl phosphoramidate (98%), dimethyl methylphosphonate
(ꢁ97%), 2-(butylamino)ethanethiol, and pinacolyl methylpho-
sphonate (97%) were supplied by Sigma-Aldrich and used
without further purification. TiO2 Hombikat UV 100 was sup-
plied by Sachtleben Chemie GmbH. Deionized water was used
for all experiments.
Trimethylsilyl (TMS) derivatization reagent N,N-bis(tri-
methylsilyl) trifuoroacetamide (BSTFA) + 1% trimethylchlor-
osilane (TMCS) was purchased from Supelco. The carbonyl
group protection reagent O-methoxyamine was supplied from
Supelco. Buffer with pH 7 containing 0.05 M solution of
KH2PO4 and NaOH was a product of Fisher Scientific. Solid
phase microextraction (SPME) was carried out using 85 mm
Carboxen/polydimethylsiloxane Stableflex fibers installed into
a manual fiber holder, both supplied by Supelco.
Fig. 1 Sonophotocatalytic setup: (1) mirror; (2) suspension of TiO2
with CWA simulant; (3) inlet and outlet of water from a thermostat;
(4) oxygen nozzle; (5) ultrasound transducer.
New J. Chem., 2002, 26, 732–744
733

View Article Online
reactor was measured using ferrioxalate actinometry combined
with a UV cutoff filter (l < 400 nm) and was equal to
1.1 ꢃ 10ꢀ4 E sꢀ1. The suspension was sonicated using a 4.8
cm thick titanium horn powered by a 600 W GE 601 ultrasonic
processor at 20 kHz. The amplitude of ultrasound was kept at
40% of maximum power, which corresponds to approximately
120 W of electric power. The power consumption was 40 W
when the reactor was empty. Therefore, approximately 80 W
was dissipated as ultrasound in the reaction mixture. The tem-
perature inside the reactor was kept around 20 ꢂC using a cool-
ing thermostat.
a thick solution. An additional 150 ml of BSTFA + 1% TMCS
reagent provided almost complete dissolution of the thick sam-
ples. Successful analysis of DEPA-TMS derivatives required
keeping the temperature of the GC/MS injection port and
transfer line below 230 ꢂC. To ensure detection of thermally
unstable compounds, the temperature of the injection port
and transfer line was kept at 200 ꢂC.
Quantification of inorganic anions and cations was per-
formed on a Dionex chromatograph with CD25 conductivity
detector. The necessary calibrations were done for measure-
ments of ion concentrations.
For photocatalytic experiments, the ultrasound transducer
at the bottom of the reactor was substituted with a Teflon bot-
tom with a slightly different geometry. A magnetic stirrer pro-
vided agitation in this case.
Identification of products
The mass spectra of products were searched in the NIST 98
library12a and Wiley Registry of mass spectral data12b (7th edi-
tion). The spectra were assigned to the corresponding products
if they had at least 90% similarity to the corresponding library
spectra. Mass spectra of a few products were absent from the
mass spectra libraries. Their structure was suggested by using
fragmentation patterns of recorded mass spectra in electron
impact (EI) ionization mode.13 Methoxime derivatives usually
had a [Mꢀ31] peak, TMS derivatives had m/z 73, multiple
TMS derivatives with oxygen had m/z 147. Chemical ioniza-
tion (CI) with methane as the reagent was applied to ascertain
the molecular weight of all unknowns. The peak of the pseudo-
molecular ion [M + 1] of TMS derivatives often had satellites
at [M + 29], [M + 41], [M + 57], and [M + 73], which helped
to determine the correct molecular weight (Mw).
Aqueous suspension of TiO2 (250 mg lꢀ1) was prepared by
stirring the mixture of TiO2 and water for 20 min followed
by sonication in an ultrasonic bath for 20 min and stirring
again for 20 min. The appropriate amount of CWA simulant
was added during the final stirring to lead to the required con-
centration of 250 mg lꢀ1
.
After 200 ml of the suspension was poured into the reactor,
ultrasound and/or ultraviolet irradiation was turned on and
samples were taken using a syringe at various intervals. The
total volume of suspension removed as samples did not exceed
25% of the initial suspension volume. TiO2 was removed from
samples by filtering through a 0.22 mm Cameo polypropylene
syringe filter. The samples were stored in a refrigerator at
about 4 ꢂC.
The pH measurements were done using a Fisher Accumet 25
pH meter with Accumet electrode. Necessary calibration was
done before measurements using standard buffer solutions
(Fisher Scientific).
The highest number of compounds absent from the libraries
was met in MOX and TMS derivatized samples of BAET. The
EI fragmentation patterns of major assigned and unassigned
unknowns follow below.
The concentration of total carbon and inorganic carbon (IC)
was measured using a Shimadzu TOC VCSH instrument. The
concentration of total organic carbon (TOC) was determined
from the difference between the concentration of total carbon
and inorganic carbon.
Volatile products of degradation were analyzed either by
direct injection into a Shimadzu QP5050A GC/MS instrument
or using solid phase microextraction (SPME). Two and a half
milliliters of sample was vigorously stirred in a 4 ml vial during
SPME either with immersion of the fiber or using headspace
absorption. Temperature of the samples was kept at about
25 ꢂC during immersion SPME and about 60 ꢂC during head-
space absorption. Supelco Supel-Q PLOT column (30 m,
0.32 mm ID) was used for analysis of volatile products of
CWA imitants destruction.
Non-volatile organic products were analyzed by means of
substitution of active hydrogen atoms with the trimethylsilyl
(TMS) group. The derivatization of products of DMMP,
DEPA, and PMP degradation was successfully performed with
samples evaporated to dryness in air. Two and a half milliliter
samples in 4 ml vials were evaporated to dryness at approxi-
mately 60 ꢂC. Then, 50 ml of BSTFA + 1% TMCS reagent
was added, and after at least 15 min the sample was analyzed
using a Shimadzu XTI-5 column (30 m, 0.25 mm ID). Products
of BAET degradation could not be analyzed with this method
because evaporation of BAET samples, even at room tempera-
ture and in vacuum, resulted in dark residues insoluble in
BSTFA + 1% TMCS reagent. Probably some products poly-
merized when the sample was concentrated. The strong acid
H2SO4 was detected in the BAET samples and could catalyze
polymerization of products. For analyses of BAET samples,
2.5 ml of each sample was mixed with 0.25 ml of pH 7 phos-
phate buffer, 0.25 ml of 5 g lꢀ1 aqueous solution of O-methox-
yamine, and slowly evaporated in air flow at 40–50 ꢂC. In this
procedure, carbonyl groups of products were protected by
forming methoximes (MOX). The dry residue was dissolved
in 50 ml of BSTFA + 1% TMCS reagent. Some samples formed
C4H9NHCH2COOH + TMS, Mw 203: 15 (2.1%), 27 (7.3%),
28 (8%), 29 (13%), 30 (28%), 41 (8.9%), 42 (23%), 43 (7.9%), 44
(49%), 45 (11%), 47 (5%), 55 (13%), 56 (6.9%), 57 (6.7%), 61
(3.7%), 70 (3.9%), 72 (17%), 73 (13%), 75 (27%), 86 (100%),
87 (6.2%), 100 (3.1%), 103 (11%), 132 (1.9%), 156 (18%), 160
(6.6%), 188 (4%), 203 (3.3%).
C4H9NHCH2CHO + MOX + TMS, Mw 216: 15 (2%), 27
(4.8%), 29 (10%), 30 (2.5%), 31 (2.7%), 41 (6.9%), 42 (100%),
43 (11%), 44 (7.9%), 45 (20%), 56 (2.6%), 57 (6.1%), 58
(3.6%), 59 (19%), 69 (11%), 72 (4%), 73 (63%), 74 (7%), 86
(12%), 89 (14%), 100 (4.1%), 115 (3.8%), 116 (5%), 127
(2.3%), 144 (4.2%), 158 (5.8%), 168 (4.9%), 173 (22%), 174
(4.3%), 185 (19%), 186 (2.8%), 201 (1.9%), 216 (0.42%).
Unassigned Mw 305: 45 (9.3%), 59 (13%), 73 (50%), 86
(26%), 87 (2.6%), 89 (2.9%), 100 (13%), 114 (4.5%), 115
(3.8%), 130 (7.9%), 131 (2%), 144 (3.2%), 147 (2.2%), 158
(2%), 172 (2.8%), 174 (100%), 175 (19%), 176 (8.4%), 245
(2.43%), 248 (2.5%), 305 (0.59%).
C2H5C(O)CH2NHCH2COOH + MOX + 2 TMS, Mw 318:
29 (5.6%), 30 (2%), 41 (3.5%), 42 (91%), 43 (7.2%), 44
(4.5%), 45 (21%), 55 (2.9%), 56 (100%), 57 (5.9%), 58 (4.2%),
59 (15%), 69 (2.9%), 70 (7.8%), 73 (99%), 74 (9.2%), 75
(11%), 84 (2.7%), 86 (14%), 89 (3.3%), 97 (21%), 100 (15%),
101 (10%), 102 (4.9%), 103 (2.2%), 114 (18%), 115 (3%), 116
(3.3%), 120 (2.2%), 128 (4.9%), 129 (2.1%), 130 (3.1%), 133
(2.7%), 144 (2.7%), 147 (11%), 156 (6.1%), 163 (2%), 188
(1.6%), 201 (11%), 202 (8.3%), 203 (2.1%), 210 (5.7%), 216
(2%), 219 (2.4%), 232 (94%), 233 (18%), 234 (8%), 246
(3.1%), 287 (3.7%), 318 (6.7%).
C4H9NHCH2CH2SO2H + 2 TMS, Mw 309: 29 (3.4%), 41
(2.1%), 42 (4.8%), 43 (3%), 45 (12%), 56 (2.3%), 58 (2.2%),
59 (11%), 73 (100%), 74 (8.7%), 75 (8%), 86 (7.9%), 98
(4.2%), 100 (13%), 101 (8%), 104 (2.3%), 115 (15%), 116
(8.3%), 128 (9%), 129 (3.9%), 130 (3.7%), 147 (5%), 156
(39%), 157 (5.7%), 158 (20%), 160 (5.1%), 172 (73%), 173
(11%), 174 (3.5%), 209 (0.63%), 254 (1.2%), 266 (8.2%), 267
(1.7%), 268 (1.1%), 294 (1%).
734
New J. Chem., 2002, 26, 732–744

View Article Online
CH3CH(OH)CH2CH2NHCH2CH2SO3H + 2 TMS, Mw
341: 45 (15%), 59 (16%), 73 (100%), 74 (8.9%), 75 (4.6%), 86
(6.4%), 100 (14%), 114 (4.1%), 130 (11%), 131 (4.3%), 133
(10%), 147 (48%), 148 (7.5%), 149 (3.9%), 160 (2.6%), 172
(4.9%), 174 (34%), 175 (6%), 176 (2.6%), 188 (19%), 189
(4%), 225 (7.1%), 238 (3.9%), 248 (2.9%), 326 (91%), 327
(26%), 328 (18%), 329 (3.9%).
C4H9NHCH2CH2SO3H + 2 TMS, Mw 325: 29 (4.9%), 41
(2.8%), 45 (11%), 55 (2.8%), 56 (26%), 57 (2.6%), 59 (9.5%),
73 (87%), 74 (7.7%), 75 (5.7%), 86 (5.9%), 87 (3.2%), 100
(2.5%), 101 (3.5%), 116 (4.4%), 133 (3.3%), 147 (14%), 158
(22%), 159 (3.2%), 172 (11%), 194 (2.7%), 225 (2.9%), 246
(2.2%), 253 (2.7%), 282 (100%), 283 (21%), 284 (14%), 285
(2.4%), 310 (7%), 325 (1.8%).
(C4H9NHCH2CH2S)2 , Mw 264: 27 (3.4%), 29 (15%), 30
(26%), 41 (9.2%), 42 (4.6%), 43 (5.7%), 44 (34%), 56 (7.3%),
57 (9.2%), 58 (2.7%), 72 (3.5%), 73 (2.3%), 77 (2%), 84
(2.2%), 86 (100%), 87 (6.1%), 88 (6.1%), 89 (4.2%), 100
(25%), 101 (3.5%), 130 (2.2%), 132 (8.4%), 133 (20%), 134
(5.6%), 165 (25%), 166 (2.1%), 167 (2.2%).
Fig. 2 TOC concentration as a function of time in sonophotocataly-
tic mineralization of CWA simulants.
Unassigned, Mw 360: 27 (2.9%), 29 (14%), 30 (12%), 42
(3.6%), 43 (3.3%), 44 (23%), 55 (2.2%), 56 (4.9%), 57 (9.3%),
86 (100%), 87 (7.6%), 88 (5.9%), 100 (3.6%), 132 (3%), 133
(2.1%), 140 (5.1%), 196 (1.2%), 228 (2.1%), 314 (6.9%).
Unassigned, Mw 456: 27 (4.1%), 29 (24%), 41 (23%), 42
(4.8%), 45 (2.7%), 55 (6%), 56 (4.4%), 57 (31%), 58 (2.3%),
59 (3.1%), 60 (2.4%), 61 (2.8%), 69 (3.1%), 88 (2%), 93
(2.1%), 126 (4.1%), 140 (60%), 141 (2.9%), 142 (0.27%), 165
(1.1%), 182 (1.8%), 196 (100%), 197 (9.3%), 228 (7.9%).
Unassigned, Mw 314: 27 (3.4%), 29 (15%), 30 (13%), 41
(10%), 42 (7.9%), 43 (7.9%), 44 (73%), 45 (9.3%), 47 (2.9%),
55 (2.3%), 56 (23%), 57 (27%), 58 (8.8%), 59 (3.2%), 70
(3.5%), 72 (4.7%), 73 (28%), 74 (3%), 75 (14%), 84 (16%), 86
(8%), 99 (3.4%), 100 (100%), 101 (7.4%), 104 (13%), 122
(3.5%), 137 (12%), 160 (1.5%), 194 (2.4%).
Degradation of DMMP
Application of ultrasound increased the rate of TOC removal
more than twofold in the case of DMMP (Table 2). One could
suspect a synergistic effect of photocatalytic and ultrasonic
treatments, as it was observed elsewhere.10 However, sonica-
tion of a TiO2 suspension containing DMMP did not result
in changes in either TOC or DMMP concentration (Fig. 3).
O’Shea et al.5 also did not detect significant degradation of
DMMP with ultrasound alone. Therefore, sonication itself
does not destruct DMMP and the positive effect of ultrasound
is seen only in combination with UV irradiation. Such an effect
can be caused by (1) increased temperature and pressure at the
microvoid implosion, (2) enhanced mass transport of reagents/
products, (3) increase of the TiO2 dispersion. TiO2 Hombikat
UV 100 contains micrometer size aggregates of primary parti-
cles of about 8 nm. Their partial deagglomeration with ultra-
sound increases the surface area available for reaction.
Transport of reagents and products through nanometer-sized
pores of TiO2 agglomerates is slow in liquid phase and could
be enhanced by ultrasound. Since ultrasound itself does not
cause degradation of DMMP, we can expect the same pro-
ducts from sonophotocatalytic and photocatalytic degrada-
tion. The final and intermediate products residing in the
solution were detected for DMMP photocatalytic degradation
and are considered below. Fig. 4 shows changes of TOC,
PO43ꢀ concentration, and pH in DMMP photocatalytic oxida-
tion. There is about a 2 h induction period in phosphate ion
formation, which was also described in the paper by O’Shea
et al.5 The pH value of the suspension decreased to about 3
and then almost levels though the concentration of phosphoric
acid continued to increase. This may indicate that the increase
in phosphoric acid concentration is balanced by decomposi-
tion of formic acid. Formic acid and formaldehyde are well
known products of DMMP photocatalytic oxidation and were
not sought in this work. They were detected and quantified in
works by O’Shea et al.4,5 Methylphosphonic acid was also
detected and quantified in their work. The overall mineraliza-
Results and discussion
Fig. 2 shows the change in total organic carbon (TOC) con-
centration during sonophotocatalytic degradation of the four
chosen CWA simulants. The initial concentration of all the
simulants was the same (250 mg lꢀ1) and the different initial
concentrations of TOC reflects the different content of carbon
atoms in the corresponding simulant molecules. Before switch-
ing on ultrasound and ultraviolet light, a blank run was carried
out for 1 or 2 h for each simulant while bubbling O2 through
the suspension. No substantial change in TOC concentration
was measured during the blank time, indicating negligible
removal of simulants by evaporation. The initial concentra-
tions of TOC for all the imitants are within experimental error
of the theoretical concentrations: BAET, 135 mg lꢀ1; DMMP,
73 mg lꢀ1; DEPA, 79 mg lꢀ1; and PMP, 117 mg lꢀ1. This indi-
cates only weak adsorption of these molecules on the TiO2 sur-
face. The mean rates of TOC removal for all the compounds
are summarized in Table 2. PMP showed the fastest rate of
mineralization. The rates for other simulants were lower, the
lowest rate being detected with BAET.
Table 2 Comparison of mean TOC removal and calculated oxygen consumption rate for sonophotocatalytic and photocatalytic degradation of
CWA imitants
TOC removal rate/mg lꢀ1 minꢀ1
Sonophotocatalytic
Oxygen consumption rate/mg lꢀ1 minꢀ1
CWA imitant
Photocatalytic
Sonophotocatalytic
Photocatalytic
DMMP
DEPA
PMP
0.17
0.15
0.21
0.047
0.062
0.14
0.25
0.22
0.20
0.41
0.12
0.080
0.19
0.49
0.17
BAET
0.066
New J. Chem., 2002, 26, 732–744
735

View Article Online
Fig. 3 TOC and DMMP concentration as a function of time during
sonication of TiO2 suspension containing DMMP.
Fig. 5 Scheme of DMMP photocatalytic degradation. All shown
products were detected.
further oxidation, in OCHO–P groups. The HOCH2O–P
group is a hemiacetal, which easily hydrolyzes in water to for-
maldehyde and the HO–P group and the hydrolysis rate
increases when pH decreases.15 The OCHO–P group in
DMMP represents an anhydride that quickly hydrolyzes in
water to formic acid and the HO–P group. Oxidation of the
methyl group CH3–P leads to the HOCH2–P group and the
corresponding product, dimethyl hydroxymethylphosphonate,
was detected. Further reactions represent combinations of oxi-
dation of methoxy group and carbon attached directly to phos-
phorus. The possible products of the latter reaction containing
OHC–P and HOOC–P groups should be stable in water but
were not detected in the products. It is possible that they are
quickly oxidized, resulting in HO–P group. The products in
the left part of Fig. 5 prevail. Therefore, the main route of
DMMP photocatalytic oxidation is oxidation of methoxy
groups followed by oxidation of the methyl group.
Below we will consider possible mechanisms of the oxidative
reactions involved in DMMP degradation. Hydroxyl radicals
OHꢄ and photogenerated holes h+ in the valance band of
TiO2 are considered as species conducting oxidative reactions
at the surface of TiO2 particles. It is often difficult to choose
which species, OHꢄ or h+, carries out the oxidation. This is
also the case for oxidation of the methoxy groups of DMMP.
Attack by both hydroxyl radicals and holes
Fig. 4 Changes of TOC, phosphate concentration and pH during
photocatalytic degradation of DMMP.
tion of DMMP proceeds according to the equation:
C3O3H9P þ 5O2 ! 3CO2 þ H3PO4 þ 3H2O
ð1Þ
As in the works by O’Shea et al., we could not detect methanol
in the products. This contrasts to photocatalytic degradation
of DMMP in the gas phase. Methanol formed, even in the
dark, obviously via TiO2 catalyzed hydrolysis.14 To estimate
the impact of hydrolysis in our case, a TiO2 suspension with
DMMP was kept at room temperature for 23 days. The pH
of the suspension only changed from 6.7 to 6.9 indicating the
absence of hydrolysis. Non-volatile products of DMMP
photocatalytic degradation were detected with GC/MS using
TMS derivatization. The main products were methyl methyl-
phosphonic acid (detected as mono-TMS ester), methylpho-
sphonic acid (detected as di-TMS ester), and phosphoric acid
(detected as tri-TMS ester). The other products that appeared
in smaller quantities were: dimethylphosphoric acid (detected
as mono-TMS ester), methylphosphoric acid (detected as
di-TMS ester), dimethyl hydroxymethylphosphonic acid
(detected as mono-TMS ester), methyl hydroxymethylpho-
sphonic acid (detected as di-TMS ester), and hydroxymethyl-
phosphonic acid (detected as tri-TMS ester). The latter
products have not been reported previously. The observed
set of products can be explained by consecutive oxidation of
different CH3 groups of the DMMP molecule. The proposed
scheme of degradation is represented in Fig. 5. The reaction
starts with oxidation of one of the methoxy groups or the
methyl group. Oxidation of methoxy groups CH3O–P of the
DMMP molecule can result in HOCH2O–P groups or, upon
CH3O P þ OHꢄ ! ꢄCH2O P þ H2O
ð2Þ
ð3Þ
CH3O P þ hþ ! ꢄCH2O P þ Hþ
will produce the same radical that can undergo further trans-
formations discussed below.
However, oxidation of the methyl group CH3–P can be
started only by reaction with OHꢄ radicals because reaction
of DMMP with holes would pꢄroceed at methꢄoxy groups.
Further transformations of CH2O–P and CH2–P radicals
can proceed via diffusion limited reactions with OHꢄ radical
to form HOCH2O–P and HOCH2–P or with oxygen to form
peroxyl radicals ꢄOOCH2O–P and ꢄOOCH2–P. Hydroxyl radi-
cals are very active species tending to recombine or scavenge H
736
New J. Chem., 2002, 26, 732–744

View Article Online
3ꢀ
atoms from organics and therefore their concentration is very
low. The concentration of oxygen in aqueous solution is of the
order of 1 mM. Thus, reaction with oxygen is more probable.
Peroxyl radicals can undergo unimolecular and bimolecular
PO4 also corresponds to complete mineralization of all the
DEPA. This indicates that no significant quantity of unreacted
DEPA resides on the surface of TiO2 . The complete minerali-
zation of DEPA follows the reaction:
ꢄ
transformations.16 Unimolecular processes comprise HO2
13
3
elimination (if a double bond is formed), ꢄO2 elimination,
ꢀ
C4O3H12PN þ O2 ! 4 CO2 þ H3PO4 þ NH3
2
4
addition to the C=C double bond, hydrogen abstraction (both
intra- and intermolecular). The bimolecular processes are elec-
tron transfer, oxygen transfer, and recombination to form
intermediate tetroxide ROOOOR. The tetroxide is very
unstable at room temperature due to the low energy of the
middle O–O bond (21–33 kJ molꢀ1) and decomposes via the
following four competitive channels:16
1
4
13
þ
HNO3 þ H2O
ð8Þ
4
After 400 min of reaction, the concentration ofꢀNH4+ starts to
level off, whereas the concentration of NO3 continues to
+
grow. This indicates that part of the generated NH4 ions
ꢀ
ꢀ
may be converted into NO3 ions. NO2 ions are considered
ꢀ
as an intermediate in the formation of NO3 ions and their
=
RCH2OOOOCH2R ! RCH O þ RCH2OH þ O2
ð4Þ
ð5Þ
ð6Þ
ð7Þ
concentration does not exceed 1.4 mM. The higher yield of
ammonia compared to nitrate is usual for the degradation of
organic compounds containing nitrogen in oxidation state
ꢀ3.9,17,18 Piccinini et al.18 observed predominate formation
=
=
RCH2OOOOCH2R ! RCH O þ RCH O þ H2O2
RCH2OOOOCH2R ! RCH2 Oꢄ þ RCH2 Oꢄ þ O2
RCH2OOOOCH2R ! RCH2OOCH2R þ O2
+
of NH4 ions even during degradation of organic compounds
with nitrogen in oxiꢀdation state +1. They also observed inter-
+
conversion of NO3 and NH4 species in illuminated TiO2
If we follow the arguments and experimental information pre-
sented by von Sonntag and Schuchmann,16 transformation of
the peroxyl radicals ꢄOOCH2O–P and ꢄOOCH2–P should pro-
ceed mostly via tetroxide intermediates with domination of the
products of channels (4) and (5) of their decay. This means for-
mation of OCH–O–P, HOCH2O–P for peroxyl radical
ꢄOOCH2O–P, and OCH–P, HOCH2–P for peroxyl radical
ꢄOOCH2–P. As we considered above, OCH–O–P and
HOCH2O–P form the detected products HCOOH, HCHO,
and HO–P in water. Products containing HOCH2–P group
were also detected, and only OCH–P group was not detected.
Thus, photocatalytic degradation of DMMP can be success-
fully explained by reactions of peroxyl radicals. Further oxida-
tion of HOCH2–P and OHC–P groups can go via reactions
with OHꢄ radicals and/or holes h+. The possible intermediate
group HOOC–P can be transformed into CO2 and a trivalent
phosphorus group via the Kolbe reaction. The trivalent phos-
phorus group should be converted into HO–P(V) by reaction
with holes and water.
suspensions. Therefore, photocatalytic degradation in oxyge-
nated water suspensions can lead to reduction of species.
As in the case of DMMP, we did not detect alcohol as a
product of degradation. Single ion monitoring and SPME
was used to detect small amount of ethanol that could be
formed as a product. However, a very small concentration
of ethanol present initially in the suspension as impurity
became even smaller after the reaction started. Volatile pro-
ducts detected by SPME are acetaldehyde and acetic acid.
Several non-volatile products were detected after derivatiza-
tion. The detected products are listed in Table 3. The major
products are diethylphosphoric acid and ethylphosphoric
acid. Hydroxyacetic acid, 2-hydroxyethyl ethylphosphoric
acid, 2-hydroxyethyl ethylphosphoramidate, ethylphosphora-
midate, and bis(2-hydroxyethyl)phosphoric acid were detected
in smaller quantities. It was not possible to provide calibra-
tions for all the organic compounds encountered in this
study. Nevertheless, qualitative information about the con-
centration profiles of products can be useful for confirming
which compounds are intermediates and understanding
kinetics.
Photocatalytic degradation of DEPA
Fig. 7 shows qualitative profiles of concentrations of organic
products during DEPA photocatalytic degradation. The
curves were obtained by normalizing the area of the chromato-
graphic peaks for selected ions to the maximum area for each
organic compound. Concentration of DEPA in Fig. 7 is the
sum of the concentrations of mono-TMS and di-TMS deriva-
tives of DEPA. The DEPA concentration falls to zero after 420
min of reaction. Therefore, all TOC in the suspension after 420
min of reaction is associated with reaction products. One can
observe in Fig. 6 that TOC concentration decreases much fas-
ter after 420 min of reaction. The explanation is that the pro-
ducts of DEPA degradation are easier to degrade to inorganic
products than the starting molecule. The profile of phosphate
ions is plotted in Fig. 7 to check whether analysis by trimethyl-
sylilalion allows us to obtain adequate qualitative profiles of
DEPA contains the P–N bond that is present in tabun (GA)
and can provide an estimate for the ability of TiO2 photocata-
lysis to detoxify tabun. Fig. 6 shows concentration versus time
dependence for TOC and inorganic ions. The initial concentra-
tion of DEPA is 1.63 mM. Almost all the organic carbon in the
solution is converted into inorganic species after 540 min of
ꢀ
+
reaction. The concentration of NO3 and NH4 ions after
540 min of reaction totals 1.6 mM, and the concentration of
3ꢀ
the concentration. When comparing the PO4 profile in Fig.
7 to that in Fig. 6, one can see that the shape of the curves
is similar and the curves have the same induction period of
about 100 min. The profiles of almost all other products also
have an induction period. We believe that this behavior is at
least partially associated with the hydrolysis rate of intermedi-
ates. Initially, the pH of the solution is close to 7, but it
becomes acidic during the reaction. Acids produced in DEPA
degradation catalyze the hydrolysis reactions. It is interesting
to note in Fig. 7 that concentrations of diethyl phosphate
and ethyl phosphate increase and decrease almost simulta-
neously although ethyl phosphate should be formed from
diethyl phosphate.
Fig. 6 TOC and inorganic ions concentration as a function of time
during photocatalytic degradation of DEPA.
New J. Chem., 2002, 26, 732–744
737

View Article Online
Table 3 Detected products of DEPA photocatalytic degradation
Measured by ion chromatography
Detected by trimethylsilylation
Detected by immersion SPME
CH3CHO (Acetaldehyde)
3ꢀ
PO4
(Diethylphosphate)
CH3COOH (Acetic acid)
+
NH4ꢀ
NO2
(Ethylphosphate)
HOCH2COOH (Hydroxyacetic acid)
ꢀ
NO3
(2-Hydroxyethyl ethylphosphoric acid)
(2-Hydroxyethyl ethylphosphoramidate)
(Ethylphosphoramidate)
[Bis(2-hydroxyethyl)phosphoric acid]
P–C(O)CH3 , both subject to hydrolysis into detected acetal-
dehyde and acetic acid, respectively. Oxidation of the b car-
bon of ethoxy groups should produce P–OCH2CH2OH and
P–OCH2CHO, as well as P–OCH2COOH, all being stable
in water. The first group was detected in small quantities
while the latter two were not detected, probably because of
their high reactivity.
Oxidation of a carbon atoms can be started either by hydro-
xyl radicals or by valence band holes in TiO2 . Only hydroxyl
radicals can start oxidation of b carbon atoms. The predomi-
nance of oxidation of a carbon atoms, however, does not
prove that reaction with holes is the main process, because
abstraction of hydrogen from the a carbon atom of ethoxy
groups by hydroxyl radicals is much faster than that from b
carbon atoms.19 The further mechanism of alkyl radical trans-
formation is supposed to be similar to that of DMMP. Peroxyl
radicals from reaction with oxygen recombine and mainly fol-
low reactions (4) and (5) to produce corresponding hydroxyl
and carbonyl groups.
Fig. 7 Normalized concentration of products of DEPA photocataly-
tic degradation detected using trimethylsilylation and SPME.
Photocatalytic degradation of PMP
Pinacolyl methylphosphonic acid (PMP) has a structure iden-
tical to soman (GD) except that a hydroxyl group replaces the
fluorine atom of soman. PMP is the hydrolysis product of
soman. Though the toxicity of PMP is low, it is important to
destroy it since soman can be obtained from PMP, whereas
one will require irreversible destruction of all the CWA.
Fig. 9 shows the absolute concentration of TOC and phos-
phate ions as well as concentration in arbitrary units of PMP
and methylphosphonic acid (MP) during photocatalytic degra-
dation of PMP. The concentration of PMP decreases to almost
zero after 200 min of reaction but TOC concentration
approaches zero only in 480 min of reaction and the phosphate
ion concentration reaches the stoichiometric value at this time.
Therefore, the longest process is to mineralize the organic
Fig. 8 shows the scheme of DEPA photocatalytic degra-
dation. The reaction starts with cleavage of the N–P bond
and oxidation of the a and b carbon atoms of the ethoxy
groups, cleavage of the N–P bond being the dominant reac-
tion. The further dominant route is oxidation of the a carbon
of ethoxy groups. Other further transformations comprise
combinations of N–P bond cleavage, and oxidation of a and
b carbon atoms. Similarly to photocatalytic destruction of
DMMP, oxidation at carbon atoms is expected to result in
hydroxyl and carbonyl groups. Thus, oxidation of a carbon
atoms of ethoxy group would produce P–CH(OH)CH3 and
738
New J. Chem., 2002, 26, 732–744

View Article Online
Fig. 8 Scheme of DEPA photocatalytic degradation. C1 corresponds to carbon atoms in the ethoxy group that is oxidized first, C2 corresponds to
those in the ethoxy group oxidized second. Compound in brackets was not detected.
ties. Normalized profiles of their concentrations are repre-
sented in Fig. 10. The time at which peaks of concentration
profiles appear may indicate how quickly compounds appear
in the PMP degradation mechanism. Three compounds have
their peak concentrations at the earliest time of reaction: 2-
methylbutenal-2, 3,3-dimethylbutanone, and butanone. The
next is the peak of 2-butenal, and then all other compounds
follow except acetic acid, which has the latest peak.
Several non-volatile products of PMP photocatalytic degra-
dation were detected after trimethylsylilation (column 2 in
Table 4). The main product was methylphosphonic acid; the
products in smaller concentrations were 1,2-dimethyl-2-for-
mylpropyl methylphosphonic acid, 3-hydroxy-1,2,2-trimethyl-
propyl methylphosphonic acid, 2-carboxy-1,2-dimethylpropyl
methylphosphonic acid, and pinacolylphosphoric acid.
Non-phosphorous products were hydroxyacetic acid, 2-hydro-
xypropanoic acid, ethylpropandioic acid, and 2-methyl-2-
hydroxypropanoic acid. Dibasic acid, alcohol and aldehyde
derivatives of PMP not listed in Table 4 were also suggested,
mainly on the basis of the molecular weights of product chro-
matographic peaks.
Fig. 9 Profiles of TOC and phosphate concentration as well as nor-
malized concentration of PMP and methylphosphonic acid (MP) in
photocatalytic degradation of PMP.
products of PMP degradation. The complete mineralization of
PMP follows the reaction:
The pinacolyl fragment of PMP molecule is more complex
than the alkoxyl fragments of DMMP and DEPA. The variety
of volatile products of PMP degradation suggests a complex
mechanism with many possibilities for crossing and branching
pathways. Because of this complexity the complete pathways
of degradation are not suggested. However, we can draw the
scheme for how the degradation of PMP gets started
(Fig. 11). The set of detected products suggests that oxidation
starts mainly at three structurally different carbon centers:
CH3–P group, a carbon of pinacolyl, and g carbons of
C7O3H17P þ 11O2 ! 7CO2 þ H3PO4 þ 7H2O
ð9Þ
In contrast to the photocatalytic degradation of DMMP and
DEPA, which resulted in few volatile products, the photocata-
lytic degradation of PMP produced about a dozen volatile
organic products. They were detected by SPME and are listed
in the third column of Table 4. The major volatile product is
acetone; other products were detected in much smaller quanti-
New J. Chem., 2002, 26, 732–744
739

View Article Online
Table 4 Main detected products of PMP photocatalytic degradation
Measured by
ion chromatography
Detected by trimethylsilylation
Detected by immersion SPME
CH3COCH3 (Acetone)
3ꢀ
PO4
(Methylphosphonic acid)
CH3COOH (Acetic acid)
CH3CH=C(CH3)CHO (2-Methylbutenal-2)
(2-Carboxy-1,2-dimethylpropyl methylphosphonic acid)
(3-Hydroxy-1,2,2-trimethylpropyl methylphosphonic acid)
(CH3)2CHC(O)CH3CHC(O)CH3 (3-Methylbutanone)
CH3CHO (Acetaldehyde)
CH2=C(CH3)CHO (2-Methylpropenal)
CH3CH2C(O)CH3 (Butanone)
CH3C(O)C(O)CH3 (Butandione)
(1,2-Dimethyl-2-formylpropyl methylphosphonic acid)
CH3CH(CH3)CHO (2-Methylpropanal)
CH3CH=CHCHO (Propenal-2)
(Pinacolylphosphoric acid)
HOCH2COOH (Hydroxyacetic acid)
(CH3)3CC(O)CH3 (3,3-Dimethylbutanone)
CH2=CHC(O)CH3 (Butenone)
CH3CH(OH)COOH (2-Hydroxypropanoic acid)
CH3CH2CH(COOH)COOH (Ethylpropandioic acid)
CH3C(CH3)(OH)COOH (2-Methyl-2-hydroxypropanoic acid)
only start oxidation at the a carbon of pinacolyl. Oxidation
of the pinacolyl fragment proceeds most probably via forma-
tion and reactions of peroxyl radicals. The stages of this
mechanism were considered in the DMMP degradation
section.
Photocatalytic degradation of BAET
2-(Butylamino)ethanethiol (BAET) imitates the sulfur and
nitrogen containing radical of VX. It also imitates the product
of P–S bond cleavage in VX. The profiles of TOC and inor-
ganic ion concentrations are plotted in Fig. 12 for BAET
photocatalytic degradation. One can observe that a 660 min
reaction time resulted in only 29% conversion of TOC. How-
ever, the concentration of sulfate ions after 660 min reached
118 ppm, which corresponds to 66% conversion of all the sul-
fur in solution. The concentration of nitrite and nitrate ions
changed negligibly during the reaction. The main final nitrogen
species is ammonia, whose concentration at 600 min of reac-
tion is 390 mM, or 21% of initial nitrogen of BAET. The BAET
mineralization follows the reaction:
21
C6H15NS þ O2 ! 6CO2 þ NH3 þ H2SO4 þ 5H2O ð10Þ
Fig. 10 Normalized concentration of organic intermediates in PMP
photocatalytic degradation as determined by immersion SPME.
2
A wide variety of volatile and non-volatile products were
detected in samples from BAET photocatalytic degradation.
Identification of unknown compounds in this case was more
difficult than for previously considered imitants because there
were many variants and isomers that could correspond to
the same molecular weight product. Therefore, compounds
that were identified without comparing to library mass spectra
are marked as ‘‘suggested’’ in Table 5, and the molecular
weight (Mw) is given for these compounds. The main volatile
products of BAET degradation are butanal and a compound
with Mw 145, which was attributed to 2-propyldihydrothia-
zole-S-oxide. Other prominent chromatographic peaks corre-
pinacolyl. Oxidation of the CH3–P group results in a HO–P
group and the possible alcoholic intermediate was not
detected. Oxidation at a carbon of pinacolyl could result in
introduction of an a hydroxyl group, a structure equivalent
to a hemiacetal. It hydrolyzes in water to form detected
methylphosphonic acid and 3,3-dimethylbutanone. Oxidation
at g carbons of pinacolyl results in g hydroxyl, carbonyl and
carboxyl groups.
It seems the oxidation starts mostly via reaction of PMP
with hydroxyl radicals since valence band holes of TiO2 can
740
New J. Chem., 2002, 26, 732–744

View Article Online
Fig. 11 Main starting routes of photocatalytic degradation of PMP.
2-ethyl-1,4-thiazine, Mw 145 to 2-propyldihydrothiazole-S-
oxide, Mw 159 to 2-acylthiazine-3-one, and Mw 163 to 2-pro-
pylthiazolidinesulfoxide.
Fig. 14 shows profiles of normalized concentrations of sev-
eral volatile products detected during BAET photocatalytic
degradation. Thiirane has its peak of concentration at the ear-
liest time of reaction, then 2-propylthiazoline, 2-propyldihy-
drothiazole-S-oxide, butenal-2, butanal peaks follow.
Propanal, acetaldehyde, N-butylpyrrole, butenone take the
third place, and last comes the peak of acetone. The vast vari-
ety of products suggests a rather complex mechanism for
BAET degradation. We can propose with some certainty only
how the BAET oxidation starts. The suggested main starting
routes of BAET photocatalytic degradation are given in
Fig. 15. Initially, BAET is oxidized into disulfide in the dark.
Further transformations include oxidation of sulfur atoms,
oxidation of carbon atoms adjacent to nitrogen, and cycliza-
tion into sulfur and nitrogen heterocycles. Oxidation of sulfur
should generate thiosulfinate R–S–S(O)–R and thiosulfonate
R–S–S(O2)–R, where R ¼ C4H9NHCH2CH2–. The thiosulfi-
nate is not stable in aqueous solution and disproportionates
into starting disulfide and thiosulfonate. The thiosulfonate also
was not detected in the products. Cassagne et al.21 found that
N,N-diisopropyl-2-aminoethyl p-fluorobenzene thiosulfonate,
a VX imitant, hydrolyzes in the presence of oxygen to
bis(N,N-diisopropyl-2-aminoethyl)disulfide and p-fluorobenze-
nesulfinic acid at pH > 1, that is when the nitrogen atom is not
protonated. In our case, the same hydrolysis pathway for
RSS(O2)R is quite possible and results in RSSR and RSO2H;
both compounds were detected. Further oxidation of sulfinic
acid produces sulfonic acid, and finally H2SO4 and alcohol
ROH. Oxidation of carbon atoms adjacent to the nitrogen
atom gives aminals and could also produce amides. Aminals
are not stable in water and hydrolyze to amine and aldehyde.
The smallest products of this reaction are butylamine and
butanal. Finally, cyclization reactions take an important part
in the degradation of BAET, producing five- and six-mem-
bered heterocycles.
Fig. 12 Profiles of TOC and inorganic ions concentration in photo-
catalytic degradation of BAET.
spond to propanal, 2-butenal, 2-propylthiazoline, N-butyl-
pyrrole. Among non-volatile compounds, butylamine,
H2SO4 , suggested C4H9NHCH2COOH, unknown Mw 305,
suggested C2H5C(O)CH2NHCH2COOH, suggested CH3CH-
(OH)CH2CH2NHCH2CH2SO3H, suggested dimer of BAET
(C4H9NHCH2CH2S)2 , and unknowns with Mw 360 and 456
had the biggest chromatographic peaks. BAET itself was not
detected in the suspension shortly after the start of reaction.
Even without UV radiation, BAET was converted in TiO2 sus-
pension to another product with molecular weight 264. This
product corresponds to the BAET dimer with the S–S bond.
Rohrbaugh20 carried out hydrolysis of VX in an equimolar
mixture with water at 90 ꢂC and detected CH3P(O)(O-
C2H5)OH and HSCH2CH2N(I–C3H7)2 as the major products.
Only minor dimerization of the thiol product was observed. In
contrast to these results, we do observe complete and fast
dimerization. In about 30 min after the start of BAET photo-
catalytic degradation, five prominent peaks were seen in the
chromatogram. One of compounds was found in the MS
library and corresponds to 2-propylthiazolidine. The EI and
CI mass spectra of the remaining four are shown in Fig. 13.
Fragmentation and molecular weight were used to suggest
their structures. They were attributed as follows: Mw 129 to
Photocatalytic degradation vs. sonophotocatalytic degradation
Values of TOC removal rate for different CWA imitants dur-
ing photocatalytic and sonophotocatalytic treatment are com-
New J. Chem., 2002, 26, 732–744
741

View Article Online
Table 5 Main detected products of BAET photocatalytic degradation
Measured by ion
chromatography
Detected by headspace SPME
and direct injection
Detected by trimethylsilylation
2ꢀ
SO4
NH2CH2COOH (Glycine)
CH3(C2H3OH)COOH (2- or 3-Hydroxybutanoic acid)
CH3CH2CH2CHO (butanal)
(2-Propyldihydrothiazole-S-oxide)
(suggested for Mw 145)
CH2=CHC(O)CH3
(Methylvinylketone)
CH3CH=CHCHO
+
NH4
CH3CH2CH2CH2NH2 (n-Butylamine)
H2SO4 (Sulfuric acid)
(2-Butenal)
CH3CH2CH2CN
(Cyanopropane)
CH3CH2CHO (Propanal)
ꢀ
NO3
C4H9NHCH2COOH + TMS [2-(N-Butylamino)acetic acid]
(suggested for Mw 203)
CH3CHO
(Acetaldehyde)
ꢀ
NO2
C4H9NHCH2CHO + MOX + TMS
[2-(N-Butylamino)acetaldehyde] (suggested for Mw 216)
(2-Propylthiazoline)
NH2CH2CH2OH + 2TMS (Aminoethanol)
(2-Ethyl-1,4-thiazine) (suggested for Mw 129)
C2H5C(O)CH2NHCH2COOH + MOX + TMS
(2-Oxobutylaminoacetic acid) (suggested for Mw 318)
(N-Butylpyrrole)
CH3COCH3 (acetone)
C4H9NHCH2CH2SO2H + 2TMS [2-(Butylamino)-
ethanesulfinic acid] (suggested for Mw 309)
(thiirane)
(2-Propylthiazole)
CS2
CH3CH(OH)CH2CH2NHCH2CH2SO3H + 2TMS
[2-(3-Hydroxybutylamino)ethanesulfonic acid] (suggested for Mw 325)
(pyrrole)
CH2=CHCHO
(Acrolein)
CH3CH2CH2CH2NO2
(Nitrobutane)
CH3CH2CH2CH2SCN
(Butylthiocyanide)
CH3COOH
(Acetic acid)
(C4H9NHCH2CH2S)2[2-(Butylamino)ethyldisulfide]
(suggested for Mw 264)
(2-Propylthiazolidinesulfoxide)
(suggested for Mw 163)
Mw 360 (Structure not suggested)
Mw 456 (Structure not suggested)
(2-Acylthiazine-3-on) (suggested for Mw 159)
Mw 314 (Structure not suggested)
Mw 305 (Structure not suggested)
(2-Propylthiazolidine)
742
New J. Chem., 2002, 26, 732–744

View Article Online
Fig. 13 Electron impact (left) and chemical ionization (right) mass spectra of several products of BAET photocatalytic degradation.
pared in Table 2. Ultrasonic irradiation induced a significant
acceleration of photocatalytic degradation for DMMP. The
rates for other CWA imitants are approximately the same with
and without sonication. This result is not surprising because
sonochemical and sonophotocatalytic degradation are known
to be sensitive to conditions and type of substrate. For exam-
ple, the rate of sonophotocatalytic degradation of trichloro-
phenol was lower than the rate of photocatalytic degradation
at 30 ꢂC and higher at 45 ꢂC.22 Sonochemical and sonophoto-
catalytic degradation are thought to depend on the distribu-
tion of substrate between solution and the gas–solution
interphase of cavitation bubbles. We used very different sub-
strates and can expect different distributions of them between
solution and the gas–solution interface.
Conclusions
Complete mineralization of compounds imitating chemical
warfare agents, namely dimethyl methylphosphonate
(DMMP), diethylphosphoramidate (DEPA) and pinacolyl
methylphosphonate (PMP), was achieved in suspensions of
TiO2 with simultaneous ultrasonication and ultraviolet irradia-
tion as well as with ultraviolet irradiation alone. 2-Butylami-
noethanethiol (BAET) also converted to inorganic products
but required a longer time for the same mass. Comparison
of sonophotocatalytic and photocatalytic treatment showed
acceleration of mineralization by sonication for DMMP. Vola-
tile and non-volatile organic products as well as inorganic ions
were detected and monitored during the photocatalytic degra-
dation. Degradation of phosphorous-containing organics was
always finalized by phosphate ions, organic sulfur converted
into sulfate ions, but nitrogen converted to nitrate ions
(25%) and ammonia (75%) in the case of DEPA, and to ammo-
nia only in the case of BAET. Pathways of degradation were
Assuming that one photon is necessary to mineralize each
molecule of CWA imitant, the quantum efficiency for complete
photocatalytic mineralization of imitants was estimated as
0.0015% for DMMP, 0.0028% for DEPA, 0.004% for PMP,
and 0.0015% for BAET.
New J. Chem., 2002, 26, 732–744
743

View Article Online
proposed for DMMP and DEPA and starting routes for PMP
and BAET. The results demonstrate the feasability of photoca-
talysis for destruction of chemical warfare agents in aqueous
phase. Optimization of degradation conditions is necessary
to achieve practical quantum efficiencies.
Acknowledgements
We gratefully acknowledge the support of the US Department
of Army Young Investigator Program under grant no. DAAD
19-00-1-0399, and NATO Science for Peace Programme grant
SfP 974209. We also acknowledge funding from the Ohio
Board of Regents (OBR) that provided matching funds for
equipment to the NSF CTS-9619392 grant through the OBR
Action Fund #333. We would like to thank Prof. D. D. Dio-
nysiou and Mr G. Anipsitakis for the help with ion chromato-
graphic measurements.
Fig. 14 Changes in normalized concentration of main organic pro-
ducts of BAET photocatalytic degradation as detected by headspace
SPME.
Fig. 15 Main starting routes of BAET photocatalytic degradation.
12 (a) NIST/EPA/NIH Mass Spectral Database, National Institute
of Standards and Technology, Gaithersburg, MD, 1998; (b) F.
McLafferty, Wiley Registry of Mass Spectral Data, CD-ROM,
Wiley, New York, 7th edn., 2000.
13 F. W. McLafferty, Interpretation of Mass Spectra, W. A. Benja-
min, Inc., Reading, MA, 1973.
14 C. N. Rusu and J. T. Yates, J. Phys. Chem. B, 2000, 104, 12 299.
15 J. Clayden, N. Greeves, S. Warren and P. Wothers, Organic
Chemistry, Oxford University Press, Oxford, New York, 2001.
16 C. von Sonntag and H.-P. Schuchmann, in Peroxyl Radicals, ed.
Z. Alfassi, Wiley, Chichester, England, p. 173.
17 S. Horikoshi, N. Watanabe, M. Mukae, H. Hidaka and N. Ser-
pone, New J. Chem., 2001, 25, 999.
18 P. Piccinini, C. Minero, M. Vincenti and E. Pelizzetti, Catal.
Today, 1997, 39, 187.
19 P. Neeb, J. Atmos. Chem., 2000, 35, 295.
20 D. K. Rohrbaugh J. Chromatogr. A, 1998, 809, 131.
21 T. Cassagne, H.-J. Cristau, G. Delmas, M. Desgranges, C. Lion,
G. Magnaud, E. Torreilles and D. Virieux, C. R. Acad. Sci., Ser.
IIc: Chim., 2001, 4, 309.
22 I. Z. Shirgaonkar and A. B. Pandit, Ultrason. Sonochem., 1998,
5, 53.
References
1
2
Y.-C. Yang, Chem. Ind., 1995, 9, 334.
Y.-C. Yang, J. A. Baker and J. R. Ward, Chem. Rev., 1992, 92,
1729.
3
4
5
6
7
8
9
C. Lion, L. Da Conceic¸a˘o, G. Magnaud, G. Delmas and M. Des-
granges, Rev. Sci. Tech. Def., 2001, 52, 139.
K. E. O’Shea, S. Beightol, I. Garcia, M. Aguilar, D. V. Kalen and
W. J. Cooper, J. Photochem. Photobiol., A, 1997, 107, 221.
K. E. O’Shea, I. Garcia and M. Aguilar, Res. Chem. Intermed.,
1997, 23, 325.
T. N. Obee and S. Satyapal, J. Photochem. Photobiol., A, 1998,
118, 45.
R. Doong and W. Chang, J. Photochem. Photobiol., A, 1997, 107,
239.
M.-C. Lu, G.-D. Roam, J.-N. Chen and C. P. Huang, Chem. Eng.
Commun., 1995, 139, 1.
G. Liu and J. Zhao, New J. Chem., 2000, 24, 411.
10 A. J. Johnston and P. Hocking, ACS Symp. Ser., 1993, 518, 106.
11 A. V. Vorontsov, E. N. Savinov, L. Davydov and P. G. Smirnio-
tis, Appl. Catal., B, 2001, 32, 11.
744
New J. Chem., 2002, 26, 732–744