Determination, synthesis and survey of iodinated trihalomethanes in water treatment processes

Article abstract of DOI:10.1016/S0043-1354(00)00079-8

Trihalomethanes (THMs) are formed as a result of the interaction of free aqueous chlorine, used as a disinfectant in drinking water, with the organic matter in raw water. Although chlorinated, brominated and chlorobromotrihalomethanes are the most common disinfection by-products reported, iodinated trihalomethanes (ITHMs) can be formed when iodide is present in raw water. ITHMs have been usually associated with several medicinal or pharmaceutical taste and odor events in drinking water. For instance, the odor and taste threshold concentrations of iodoform are 0.02 and 5 μg/l, respectively. Different analytical techniques have been studied to identify these compounds but their quantitative determination has not been performed due to the lack of commercial standards. In this study ITHMs (CHCl₂I, CHClI₂, CHBr₂I, CHBrI₂ and CHBrClI) have been synthesized in order to evaluate headspace (HS), purge and trap (P&T), closed loop stripping analysis (CLSA) and liquid-liquid extraction (LLE) as analytical methods for determination. HS and LLE were followed by gas chromatography and electron capture detector (GC/ECD); whereas P&T and CLSA with gas chromatography and mass spectrometry (GC/MS). The most appropriate method, LLE/GC/ECD, was applied to evaluate the stability of ITHMs in water (ultrapure, raw and treated water) in order to confirm their presence in tap water. Ascorbic acid was the quenching reagent chosen to avoid free chlorine at the time of sample collection. Finally, samples from the different stages of the treatment plant in Barcelona (NE Spain) were analyzed. Only three (CHCl₂I, CHBrClI and CHBr₂I) of the six ITHMs were identified and determined, at average levels lower than 1μg/l, in sand filtered and ozonated waters. No ITHMs were identified in distribution system water. Trihalomethanes (THMs) are formed as a result of the interaction of free aqueous chlorine, used as a disinfectant in drinking water, with the organic matter in raw water. Although chlorinated, brominated and chlorobromotrihalomethanes are the most common disinfection by-products reported, iodinated trihalomethanes (ITHMs) can be formed when iodide is present in raw water. ITHMs have been usually associated with several medicinal or pharmaceutical taste and odor events in drinking water. For instance, the odor and taste threshold concentrations of iodoform are 0.02 and 5 μg/l, respectively. Different analytical techniques have been studied to identify these compounds but their quantitative determination has not been performed due to the lack of commercial standards. In this study ITHMs (CHCl₂I, CHClI₂, CHBr₂I, CHBrI₂ and CHBrClI) have been synthesized in order to evaluate headspace (HS), purge and trap (P and T), closed loop stripping analysis (CLSA) and liquid-liquid extraction (LLE) as analytical methods for determination. HS and LLE were followed by gas chromatography and electron capture detector (GC/ECD); whereas P and T and CLSA with gas chromatography and mass spectrometry (GC/MS). The most appropriate method, LLE/GC/ECD, was applied to evaluate the stability of ITHMs in water (ultrapure, raw and treated water) in order to confirm their presence in tap water. Ascorbic acid was the quenching reagent chosen to avoid free chlorine at the time of sample collection. Finally, samples from the different stages of the treatment plant in Barcelona (NE Spain) were analyzed. Only three (CHCl₂I, CHBrClI and CHBr₂I) of the six ITHMs were identified and determined, at average levels lower than 1μg/l, in sand filtered and ozonated waters. No ITHMs were identified in distribution system water. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0043-1354(00)00079-8

Wat. Res. Vol. 34, No. 13, pp. 3380±3390, 2000
7
2000 Elsevier Science Ltd. All rights reserved
Printed in Great Britain
0043-1354/00/$ - see front matter
PII: S0043-1354(00)00079-8
www.elsevier.com/locate/watres
DETERMINATION, SYNTHESIS AND SURVEY OF
IODINATED TRIHALOMETHANES IN WATER
TREATMENT PROCESSES
1
BEATRIZ CANCHO , FRANCESC VENTURA *, MATERESA GALCERAN ,
1
2
3
3
ALFREDO DIAZ and SUSAGNA RICART
AGBAR, Societat General d'Aigues de Barcelona, S.A., Passeig de Sant Joan, 39, 08009 Barcelona,
Spain; Department of Analytical Chemistry, University of Barcelona, Avda Diagonal, 647, 08028
Barcelona, Spain and Department of Organic Chemistry, CID-CSIC, Jordi Girona, 18-26, 08034
Barcelona, Spain
1
È
2
3
(
First received 1 July 1999; accepted in revised form 1 December 1999)
AbstractÐTrihalomethanes (THMs) are formed as a result of the interaction of free aqueous chlorine,
used as a disinfectant in drinking water, with the organic matter in raw water. Although chlorinated,
brominated and chlorobromotrihalomethanes are the most common disinfection by-products reported,
iodinated trihalomethanes (ITHMs) can be formed when iodide is present in raw water. ITHMs have
been usually associated with several medicinal or pharmaceutical taste and odor events in drinking
water. For instance, the odor and taste threshold concentrations of iodoform are 0.02 and 5 mg/l,
respectively. Di€erent analytical techniques have been studied to identify these compounds but their
quantitative determination has not been performed due to the lack of commercial standards. In this
2 2 2 2
study ITHMs (CHCl I, CHClI , CHBr I, CHBrI and CHBrClI) have been synthesized in order to
evaluate headspace (HS), purge and trap (P&T), closed loop stripping analysis (CLSA) and liquid±
liquid extraction (LLE) as analytical methods for determination. HS and LLE were followed by gas
chromatography and electron capture detector (GC/ECD); whereas P&T and CLSA with gas
chromatography and mass spectrometry (GC/MS). The most appropriate method, LLE/GC/ECD, was
applied to evaluate the stability of ITHMs in water (ultrapure, raw and treated water) in order to
con®rm their presence in tap water. Ascorbic acid was the quenching reagent chosen to avoid free
chlorine at the time of sample collection. Finally, samples from the di€erent stages of the treatment
2 2
plant in Barcelona (NE Spain) were analyzed. Only three (CHCl I, CHBrClI and CHBr I) of the six
ITHMs were identi®ed and determined, at average levels lower than 1mg/l, in sand ®ltered and
ozonated waters. No ITHMs were identi®ed in distribution system water. 7 2000 Elsevier Science Ltd.
All rights reserved
Key wordsÐclosed loop stripping analysis, drinking water, gas chromatography, headspace, iodinated
trihalomethanes, liquid±liquid microextraction, purge and trap, stability, synthesis
INTRODUCTION
and N'Guyen, 1986; Coleman et al., 1976; Glaze et
al., 1975; Hanai et al., 1996; Hansson et al., 1987;
Kleopfer, 1976). For instance, dichloroiodomethane
Trihalomethanes (THMs) are formed as a conse-
quence of the reaction between natural organic mat-
has been found in 85 out of 111 USA water
ter in raw water and chlorine used in the treatment
supplies and is reported more frequently than bro-
moform (Brass et al., 1977). The formation and
process (Bellar et al., 1974; Rook, 1974; Bunn et
al., 1975). Although chlorinated, brominated and
mixed chlorobromoderivatives are the most fre-
stability of ITHMs has been studied (Gould et al.,
1
985; Gould, 1993; Rickabaugh and Kinman, 1978;
quent disinfection by-products (DBPs) reported,
iodinated trihalomethanes (ITHMs) are also formed
when iodide (from natural sources, sea-water intru-
sion or brines) is present. Iodide levels in ground-
water usually vary from 0.01 to 20 mg/l (Ali-
Mohamed and Jamali, 1989). ITHMs have been
identi®ed in drinking waters worldwide (Anselme
Thomas et al., 1980; Vel Leitner et al., 1998a, b)
and the detection of several iodinated DBPs in re-
cycled water treated with iodine, the alternative dis-
infectant chosen in the USA space program, has
been reported (Dietrich et al., 1997; Frazey et al.,
1
998). In addition iodinated substances have often
been studied in marine environments; methyl iodide
is the most abundant compound identi®ed,
although other minor species such as chloroiodo-
methane, diiodomethane and alkyl iodides have
*
Author to whom all correspondence should be adressed.
Fax: +34-93-342-36-66; e-mail: fventura@agbar.es
3380

Iodinated trihalomethanes in water treatment processes
3381
also been found (Schall et al., 1997). The presence ing methods: headspace (HS); the EPA method
of the four most common THMs in drinking water 502.1 purge and trap (P&T) with a Tenax/silica gel/
is regulated in the USA (Fair, 1992), whereas only charcoal trap and two other di€erent adsorbents
chloroform and dichlorobromomethane are con- (Tenax and Vocarb 3000); the CLSA (unsalted and
trolled in Europe (EEC, 1980). Since the US En- salted) (Grob, 1973; Hwang et al., 1984) followed
vironmental Protection Agency (USEPA) set by GC/MS and the EPA method 551.1 liquid±
regulatory standards for THMs, several water liquid microextraction (LLE) with tert-butyl methyl
works have replaced chlorination by chloramina- ether (MtBE) were considered. The quality par-
tion. The latter reduces the amount of THMs but it ameters, precision, linearity and limit of detection
cannot control the formation of ITHMs if iodide is for LLE/GC/ECD were calculated. Levels of
present (Hatch and Gary, 1999; Vel Leitner et al., ITHMs in di€erent stages of the water treatment
1
998a, b). plant on the Llobregat river are exemplarized. The
ITHMs are associated with characteristic phar- Llobregat river contains high concentrations of bro-
maceutical or medicinal odors and taste in water. mide (0.7 mg/l, average level in 1997) and iodide
For example, the odor threshold concentration of (3.2 mg/l) (Carnicero et al., 1998) coming from salt
iodoform, 0.02 mg/l is signi®cantly lower than that mine discharges in its upper course.
of chloroform or bromoform, 100 and 300 mg/l, re-
spectively (Hanai et al., 1996). The taste concen-
tration for iodoform has been established as 5 mg/l
EXPERIMENTAL
(
Hansson et al., 1987). This means that ITHMs at
concentrations between 0.02 and 10 mg/l or higher Reagents
are able to cause medicinal taste and odor problems
Bromochloroiodo-, bromodiodo-, chlorodiodo-, dibro-
in drinking water. Complaints from consumers re- moiodo- and dichloroiodomethane were synthesized as
indicated below. Bromoform, chloroform, chlorodibromo-
methane and iodoform were purchased from Sigma
Aldrich (USA). The chemical reagents used as internal
standards were bromochloromethane, 1,4-dichlorobutane
lated to iodoform have been reported in France
(Anselme and N'Guyen, 1986; Bruchet et al., 1989)
and after chloramination of raw water in Australia
(
Hansson et al., 1987). This last problem was elimi- and 1,2-dibromopropane from Sigma Aldrich; 1-chloroalk-
nated when the order of addition was chlorine anes (C , C , C12, C16) were purchased from Fluka (Swit-
zerland); the 1-chloroalkanes (C , C10, C14 and C18),
uorobenzene, 1-bromo-3-¯uorobenzene and 1,2-dichloro-
5
6
8
before ammonia instead of the reverse combination.
Gittelman and Yohe (1989) showed that 5 min of
chlorination before the addition of ammonia elimi-
nated the medicinal odor of iodoform.
Several analytical methods have been used to
analyze ITHMs qualitatively. The most common
are purge and trap (P&T) followed by gas chroma-
¯
4
benzene-d were from Aldrich Chemie (Germany). The
chemical reagents used as quenching agents for chlorine
were L-ascorbic acid purchased from Sigma, sodium thio-
sulfate 0.1 N and sodium sul®te 0.1 M from Carlo Erba
(Italy). Other reagents used were methanol purge and trap
grade from Sigma Aldrich; methyl-t-butyl-ether (MtBE)
suprasolv grade, carbon disul®de for spectroscopy from
tography (GC) with a Hall conductivity detector Merck (Germany) and sodium sulfate ACS-ISO for analy1-
(
Thomas et al., 1980) or a gas chromatography sis from Carlo Erba. Ultrapure water was from Milli-Q
water puri®cation system (Millipore Corporation, USA).
and mass spectrometry (GC/MS) system (Ricka-
baugh and Kinman, 1978), liquid±liquid extraction
analysis followed by thermal desorption of the
adsorbent into a GC/MS (Hansson et al., 1987),
Synthesis of ITHMs
ITHMs were synthesized basically following the pro-
È
cedures proposed by Sulzle (1987) and Hine and Procer
solid phase microextraction (SPME) (Cancho et (1958) for dibromoiodo-, dichloroiodo- and bromochlor-
oiodomethane; and by Auger (1908) for bromodiiodo- and
chlorodiodomethane.
Dibromoiodomethane. Bromoform (50 g, 0.2 mol), NaI
al., 1999a) with temperature programmed de-
sorption (Frazey et al., 1998) or concentration in
Amberlite XAD-2 and further con®rmation by
(
2 2
7.5 g, 0.05 mol) in 20 ml of CH Cl , 50% KOH aqueous
GC/MS (Bruchet et al., 1989). However, the una- solution (20 ml) and 18-crown-6 (0.44 g, 0.0015 mol) were
vailability of ITHMs reference standards has ham- mixed and stirred for 2 days at 08C. The expected product
was obtained as a yellow oil after distillation in high vac-
pered their accurate quantitative determination.
uum (b.p. 908C at 40 mmHg, 20% yield) in relatively pure
form by H-NMR (acetone-D ): d, 7.0 ppm and C-
6
Only three iodinated standards are commercially
available (CHI , CH ClI and CH I ).
1
13
3
2
2 2
NMR (acetone-D ): d, � 39.5 ppm. Purity (92%) was
6
Medicinal odor is one of the descriptors associ- determined by GC/FID. Bromoform (1%) and tetrabro-
momethane (7%) were the impurities identi®ed by GC/
MS.
Thus, the aim of this study was to evaluate di€er-
Dichloroiodomethane. Chloroform (160 g, 1.3 mol), NaI
ated with Barcelona drinking water by panellists.
ent analytical procedures for quantitative deter-
(
50 g, 0.3 mol), 50% KOH aqueous solution (75 ml) and
mination of the six possible ITHMs (CHCl2I, 18-crown-6 (1.74 g, 0.006 mol) were mixed and strirred for
CHClI
2
, CHBr
2
I, CHBrI
2
, CHBrClI and CHI
3
). 2 days at 08C. The reaction mixture was diluted with a
large excess of ice-water. The organic phase was separated,
dried and evaporated in a rotary vacuum evaporator. Dis-
tillation over molecular sieves of the residue a€orded the
Levels of ITHMs were measured in drinking water
in order to determine whether they could be corre-
lated with this descriptor. All compounds except
expected product as a yellow oil (b.p. 398C at 25 mmHg,
1
iodoform were previously synthesized. The follow- 30% yield) in pure form by H-NMR (acetone-D
6
): d, 7.6

3
382
Beatriz Cancho et al.
6
): d, 12.5 ppm. Purity was ascorbic acid solution was added to each bottle prior to
13
ppm and C-NMR (acetone-D
con®rmed by GC/FID.
Bromochloroiodomethane. Over
dibromochloromethane (20 g, 0.1 mol) and NaI (29 g,
0
3
sampling. The ascorbic acid reduces any free chlorine and
additional generation of THMs is prevented. In order to
con®rm that no free chlorine was present, the N,N-
diethyl-p-phenylendiamine (DPD) method was used
(APHA et al., 1995).
a stirred mixture of
.2 mol) in 40 ml of MeOH (Ar saturated), 14 ml of a
0% solution of MeONa was added dropwise over 4 h in
the dark under Ar atmosphere. After extraction with pen-
tane/ice-water, drying and vacuum evaporation, the yellow
oil (b.p. 68±698C at 30 mmHg, 25% yield was shown to
Analytical procedure
1
Headspace (HS). Analyses of ITHMs were carried out
by placing 8 ml of water in a glass vial (10 ml) capped
with te¯on-faced septa and metallic seals. The vials were
incubated at 708C for 15 min. Internal standards were bro-
mochloromethane and 1,4-dichlorobutane (4 ml of
methanolic solution of 50 and 500 mg/l of each com-
pound, respectively).
Purge and trap method (P&T). Analyses were carried
out in a commercial Tekmar 2000 apparatus (USA).
Water samples (5 ml) were purged in all cases for 11 min
with helium (32 psi) and adsorbed onto three di€erent
traps: Tenax/silica gel/charcoal (EPA Method 502.1),
Tenax and Vocarb 3000 (Supelco, Bellefonte USA) at
room temperature. ITHMs were desorbed at 1808C for
4 min in the Tenax/silica gel/charcoal trap, at 1808C for
be the desired product in relatively pure form by H-
13
6
NMR (acetone-D ): d, 7.3 ppm and C-NMR (acetone-
D
6
): d, � 12.5 ppm. Its purity (86%) was determined by
GC/FID. Chlorodibromomethane (10%) and chlorodiio-
domethane (4%) were the impurities identi®ed by GC/MS.
Bromodiiodomethane. Over a stirring suspension, iodo-
a
form (50 g, 0.13 mol) in 25 ml of CCl
of Br (16 g, 0,10 mol) in 25 ml of CCl
wise over 3 h. The resulting solution was stirred overnight
and quenched with a saturated solution of Na . The
4
at 08C, a solution
2
4
was added drop-
2 2 5
S O
organic layer was separated, dried and steam distilled.
After crystallization in hexane, the yellow solid. (m.p.
6
08C in literature, 30% yield) was shown to be the desired
1
6
product in relatively pure form by H-NMR (acetone-D ):
d, 6.3 ppm. C-NMR (acetone-D
purity (84.5%) was determined by GC/FID. Dibromoio-
domethane (9%) and iodoform (6.5%) were the impurities
identi®ed by GC/MS.
13
6
): d, � 95 ppm. Its
1
min in Tenax trap or at 2508C for 6 min when the
Vocarb 3000 trap was used. The compounds adsorbed
were transferred directly to the GC/MS system in all
cases. Fluorobenzene, 1-bromo-3-¯uorobenzene and 1,4-
dichlorobenzene-d were used as internal standards (2 ml
4
of a methanolic solution of 5 mg/l of each compound was
added).
Closed-loop stripping analysis (CLSA). Analyses were
carried out in a commercial CLSA apparatus (Brechbuler,
È
Switzerland) according to the method developed by Grob
(1973). Water samples (1 l for unsalted samples and 0.9 l
for salted samples with 36 g of anhydrous sodium sulfate)
were spiked with ITHMs at two concentrations (10 and
Chlorodiiodomethane. Iodoform (32 g, 0.08 mol) and
HgCl
308C for 2 h. The compound formed was distilled o€
80±100 mmHg) and ®nally obtained by extraction of the
2
(20 g, 0.07 mol) were heated under stirring at
1
(
remaining residue with pentane and vacuum evaporation
of the organic phase (b.p. 888C at 30 mm Hg, 30% yield).
The desired product was obtained in relatively pure form
1
13
by H-NMR (acetone-D
acetone-D
): d, � 69 ppm. Its purity (94%) was con®rmed
by GC/FID. Diiodomethane (1%), dichloroiodomethane
1%) and iodoform (4%) were the impurities identi®ed by
6
): d, 6.7 ppm and C-NMR
(
6
5 6
1.5 mg/l). Several 1-chloroalkanes (C , C , C10, C12, C16
(
and C18) were added to the spiked water samples as surro-
gate standards at a concentration of 800 ng/l for each
compound. Only new 1.5 mg-activated carbon ®lters were
used. The samples were stripped for 1 h. Temperatures of
GC/MS.
Standard solutions
ITHMs are photosensitive and also decompose in 458C and 558C were used for water-bath and carbon ®lter,
speci®c organic solvents. For instance, chloro-, bromo- respectively. After stripping, the carbon ®lters were spiked
and iodoacetone can be formed in acetone solution as pre- with C and C14 1-chloroalkanes at the same concen-
8
viously reported (Bruchet et al., 1989). In addition, sol- trations as the spiked water samples in order to control
utions of iodoform in non polar solvents such as n-
pentane tend rapidly to turn violet because of liberation of
iodine (Thomas et al., 1980). Consequently, the use of
methanol, amber glassware and oxygen elimination was
required.
Stock standard solutions were prepared in high purity
methanol by weighing approximately 0.1 g of analyte into
a 10-ml volumetric ¯ask and diluting to volume. A sec-
ondary standard solution was prepared by dilution in
methanol of the primary standard to give concentrations
of 10 mg/l. When ITHMs were injected directly into the
the ®lter elution. The ®lters were then extracted with 20 ml
of CS (10 ml+10 ml).
2
Liquid±liquid microextraction (LLE). Water samples
(35 ml) were placed in 40 ml EPA glass vials (Wheaton,
USA). To each sample, 1,2-dibromopropane was used as a
surrogate standard (5 ml of a methanolic solution of 70 mg/
l); 10 g of anhydrous sodium sulfate and 2 ml of glass-
bidistilled methyl tert-butyl ether were added (EPA
Method 551.1). The vials were then sealed with te¯on-
faced septa, shaken for 2 min, placed upright and left to
stand for 3 min. 500 ml of the organic layer was trans-
column, the last dilution was made using MtBE as solvent. ferred into a 2-ml vial containing bromochloromethane as
Ultrapure water solutions were prepared by spiking with
di€erent amounts of the secondary standard and used for
recovery studies and calibration.
an internal standard (5 ml of a methanolic solution of
10 mg/l). Extracts were always analyzed within 24 h in a
DB-1 column. Peaks were identi®ed by comparison of
retention times with the synthesized standards and con-
®rmed by injection of the MtBE extracts in the DB-624
column.
Sample collection
Samples from di€erent stages of the water works plant
were collected in 100-ml glass bottles with te¯on septa and
polypropylene screw caps, ®lled without over¯ow. Pre-
viously, the glass bottles had been washed in soapy water,
rinsed in tap water, ultrapure water and ®nally reagent-
grade acetone. Then they were air-dried and baked at
Instrumental conditions
Analyses by headspace were performed using a Fisons
8130 gas chromatograph equipped with an autosampler
HS8000 (Fisons) and an electron capture detector (ECD).
1
with aluminum foil to preserve them from dust and other
contaminants.
A volume of 0.1 ml of a freshly prepared 0.142 M
808C for 12 h. After this procedure, they were covered Injections (0.5 ml of gas phase) were made in the split
mode (1/20) into a 30-m  320-mm i.d. DB-624 (1.8-mm
®lm thickness) fused silica column (J&W Scienti®c, USA).
The GC temperature program was 408C (5 min) to 1408C

Iodinated trihalomethanes in water treatment processes
3383
at 68C/min and to 2208C (10 min) at a rate of 158C/min.
Carrier gas was helium (80 kPa) and nitrogen (100 kPa)
as make-up. Injector and detector temperatures were
2258C and 3008C, respectively.
Purge and trap analyses were carried out on a Fisons
MD800 mass spectrometer equipped with a Fisons 8060
gas chromatograph. Extracts were transferred directly to
the chromatograph (transfer line 1808C) into a 75-m Â
530-mm i.d. DB-624 (3-mm ®lm thickness) fused silica col-
umn (J&W Scienti®c). The GC temperature program was
308C (5 min) to 2208C (10 min) at a rate of 38C/min.
Helium was used as carrier gas (60 kPa). The mass spec-
trometer was operated in electron impact mode (70 eV).
The source temperature was kept at 2508C. Mass spectra
were acquired by scanning from 35 to 450 Da with 1 s/
decade.
Analyses of CLSA extracts were carried out on a Fisons
Trio1000 mass spectrometer equipped with a Fisons 8060
gas chromatograph. Injections (1 ml of CS
made cold on column into a 50-m  320-mm i.d. (0.2-mm
lm thickness) CP-Sil 19-CB column (Chrompack, The
Netherlands). The GC temperature program was 308C
5 min) to 2858C (10 min) at a rate of 38C/min. Helium
2
extracts) were
®
(
was used as the carrier gas (60 kPa). The mass spec-
trometer was operated in electron impact mode (70 eV).
The source temperature was kept at 3008C. Mass spectra
were acquired by scanning from 35 to 450 Da with 1 s/
decade.
After liquid±liquid microextraction the analyses were
performed using
a Fisons 8130 gas chromatograph
equipped with an ECD. Injections (1 ml of MtBE extracts)
were made in splitless mode (60 s) into a 30-m  250-mm
i.d. DB-1 (1.0 mm ®lm thickness) fused silica column
(
(
(
(
J&W Scienti®c). The GC temperature program was 358C
9 min) to 408C (3 min) at 18C/min, the up to 2208C
10 min) at a rate of 68C/min. The carrier gas was helium
140 kPa) and nitrogen (110 kPa) as make-up. Injector
and detector temperatures were 1758C and 3008C, respect-
ively.
1
13
H and C NMR spectra were obtained with a Varian
1
Unity-300 spectrometer (300 MHz for H and 75 MHz
for C).
13
RESULTS AND DISCUSSION
Physico-chemical properties of ITHMs
Table 1 shows di€erent parameters calculated for
the ITHMs such as octanol±water partition coe-
cients (Kow), gas±liquid partition coecients (K )
and dimensionless Henry's law constant (H ). Ex-
perimental boiling points corrected to standard
pressure are also included.
Values of Kow for ITHMs were calculated theor-
etically by means of a commercial computer pro-
gram (Prolog) that breaks down the molecule and
assigns a value to each fragment according to data
base. The ®nal value is the sum of the partial values
with a ®nal correction after considering the struc-
ture of the molecule. The log Kow increased when
molecular weight increased (Table 1).
The values of determination of K and dimension-
less H constant by static headspace were obtained
according to the phase ratio variation (PRV)
method (Ettre et al., 1993; Kolb and Ettre, 1997).
This method is based on the relationship betweeen
the reciprocal peak area and the phase ratio of the
vial containing the sample solution. We used the

3
384
Beatriz Cancho et al.
the entrophy of vaporization at 760 mm (estimated
equation
from the graph f vs b.p. at a standard pressure for
methane) (Hass and Newton, 1993).
1=A ˆ ꢀK Á Rf †=Co ‡ ꢀRf=Co† Á b
Finally, Table 1 displays the characteristic m/z
values in the electron impact (EI) mode for these
compounds as well as the retention times obtained
in the four chromatographic columns used. The
elution order for these compounds was the same in
all cases. Coelution with the uncommon 1-bromo-1-
chloropropanone and 1,1-dibromopropanone was
where A is the peak area obtained in the individual
headspace measurements, K is the gas±liquid par-
tition coecient, Co is the initial concentration of
the analyte in the sample solution, b is the gas±
f
liquid phase ratio in the vial and R is the GC re-
sponse factor. The experimental K value for each
ITHM was obtained representing plots of 1/A vs
di€erent b. K can be calculated as a ratio of y-inter-
cept and slope of the linear plots. Results are
shown in Table 1.
observed in the DB-1 column for CHCl I and
2
CHBrClI in some water samples.
Recovery eciencies of ITHMs
In addition, all boiling points (b.p.) were cor-
rected to standard pressure (760 mm Hg). This cor-
A study of the recovery eciencies obtained by
rection was made by using the equation HS, P&T, CLSA and LLE methods was performed
in order to evaluate the most appropriate method
Dt ˆ ‰ꢀ273:1 ‡ t†ꢀ2:8808 � log p†Š=f ‡ 0:15ꢀ2:8808 �
log p†Š where Dt was the 8C degrees to be added to to determine the iodinated trihalomethanes in
the observed boiling point, t was the observed boil- water. Average ITHM recoveries and relative stan-
ing point, log p was the logarithm of the observed dard deviations obtained from ultrapure water
pressure in millimeters of mercury and ®nally f was spiked at 10 and 1.5 ppb of each compound are
Table 2. Comparison of recoveries for THMs and ITHMs using the HS, P&T, CLSA and LLE methods ꢀn ˆ 5)
THMs
Br CHClBr
ITHMs
CHCl
3
CHCl
2
2
CHBr
3
CHCl
50
6.0
25
6.5
2
I
CHBrClI
CHBr
2
I
CHClI
2
CHBrI
2
CHI
3
HS/GC/ECD
a
a
21
a
a
%
%
%
%
rec (10 mg/l)
r.s.d.
rec (1.5 mg/l)
r.s.d.
23
21
6.5
17
20
36
7.0
40
47
9.5
43
48
4.9
36
56
6.4
56
0.1
42
nr
nr
9.2
26
5.6
5.1
18
8.7
17
9.4
9.9
9.4
7.2
13.0
6.3
5.5
P&T/GC/MS
Tenax/silica/charcoal
b
b
b
b
%
%
%
%
rec (10 mg/l)
r.s.d.
rec (1.5 mg/l)
r.s.d.
117
90
99
91
66
7.2
45
5.9
63
1.1
39
42
1.4
33
49
4.1
33
23
15
14
2.8
4
13.6
nr
2.7
97
1.4
6.4
83
6.7
0.1
84
1.1
1.9
69
7.0
4.8
1.9
4.3
nr
Tenax
b
b
b
b
%
%
%
%
rec (10 mg/l)
r.s.d.
rec (1.5 mg/l)
r.s.d.
114
100
96
78
72
6.5
58
4.5
57
7.6
43
41
5.0
28
27
12
16.5
6.7
4.7
31.1
nr
nr
nr
nr
nr
4.4
112
5.6
6.0
83.5
6.6
6.3
78
5.2
2.6
60
8.8
4.3
2.1
nr
Vocarb 3000 Trap
b
b
b
b
%
%
%
%
rec (10 mg/l)
r.s.d.
rec (1.5 mg/l)
r.s.d.
98
113
132
98
64
1.0
36
3.6
37
25
13
8.4
4.0
39
6.8
15
18
8.3
nr
nr
nr
nr
nr
21.5
107.5
0.5
7.8
89
7.0
5.0
102
2.0
7.5
71
1.2
1.5
19.5
3.2
5.6
nr
CLSA/GC/MS
Unsalted water
c
c
c
%
%
%
%
rec (10 mg/l)
r.s.d.
rec (1.5 mg/l)
r.s.d.
nd
12
29
52
47
6.0
51
6.1
66
3.0
71
81
6.8
92
94
8.5
89
64
12
68.5
8.0
55
5.4
38
9.4
nd
nd
nd
5.0
17
6.1
4.7
42
6.0
5.4
65
6.0
4.2
2.2
6.7
Salted water
b
b
b
%
%
%
%
rec (10 mg/l)
r.s.d.
rec (1.5 mg/l)
r.s.d.
nd
nd
nd
nd
18
26
34
47
3.4
48
3.6
63
3.1
75
79
74
60
73
3.9
82.8
16.5
7.7
18
5.3
8.7
35.5
9.8
7.3
56
6.6
8.5
102
5.8
2.1
77.5
3.2
7.7
63.5
6.4
6.0
LLE/GC/ECD
b
b
b
b
%
%
%
%
%
%
rec (10 mg/l)
r.s.d.
rec (1.5 mg/l)
r.s.d.
rec (0.5 mg/l)
r.s.d.
101
100
127
111
109
4.9
102
3.3
95
9.0
103
4.8
101
2.4
90
109
3.0
109
5.5
102
1.7
99
3.7
66.5
2.0
101
3.7
108
4.1
85
4.8
102
1.7
102
3.2
108
4.0
100
2.0
4.0
100
5.0
92
12
5.0
101
5.0
106
6.7
5.0
101
5.0
112
3.8
6.7
102
5.2
125
4.2
7.3
a
3
5
.6 mg/l.
.0 mg/l.
b
c
2
.0 mg/l.
nd, not determined; nr, not recovered.

Iodinated trihalomethanes in water treatment processes
3385
given in Table 2, as well as values for chloro-, with the Tenax trap were similar to those obtained
bromo- and mixed trihalomethanes.
for CHCl I, CHBr I and CHBrClI in the EPA rec-
Headspace. This is a simple and reproducible ommended trap. However, for the three compounds
2
2
technique that can be automated easily. In the with higher number of iodine atoms, the recoveries
Barcelona water treatment plant it is used systema- decreased dramatically, and CHBrI and CHI were
2 3
tically to determine the four common trihalo- not extracted at 1.5 mg/l from water. Finally,
methanes generated in the treatment process. Vocarb 3000 was the last trap tested. It contains a
European norm NF EN ISO 10301 (July 1997) combination of adsorbents (Carbopack B/Car-
advocates heating between 508C and 808C during boxen) that allows desorption temperatures (2508C)
at least 30 min for high volatile hydrocarbons higher than other traps (1808C), which theoretically,
determination. However, the analytical conditions provides faster transfer of analytes and improves
708C and 15 min were found to be the best after chromatography. Using Vocarb 3000 we obtained
several assays performed considering di€erent incu- quantitative results for the common THMs, but low
bation temperatures and heating times. The total recovery values were found for ITHMs. A modest
mass of ITHMs transferred to the headspace 42% recovery for CHCl
during each extraction step varied from 2.6% obtained, and iodoform was not recovered. There-
CHCl I) to 0.05% for CHI , which are in the same fore, Vocarb 3000 does not improve the perform-
I was the best result
2
(
2
3
range as other DBPs such as haloketones and ance of the Tenax/silica gel/charcoal trap.
haloacetonitriles (Cancho et al., 1999b). We calcu-
Closed loop stripping analysis. This is the method
lated the recoveries of ITHMs considering both the of choice to analyze compounds related to water
partition coecient (K) and the phase ratio of the taste and odor problems at trace levels. In this
vial (b) due to their in¯uence on the analyte con- study the chloroform was not studied because it
centration in the headspace at equilibrium con- coelutes with carbon disul®de solvent impurities. In
ditions (Ettre et al., 1993). Recovery values were contrast to the P&T method, bromoform and bro-
close to 50% at 10 mg/l for all compounds except mochlorotrihalomethanes were not quantitatively
for iodoform that was practically not recovered. recovered and a range of 12% (CHBrCl
These low recoveries are not a drawback for the de- (CHBr ) recovery values was obtained for the cur-
2
) to 65%
3
termination of THMs in water from the di€erent rent THMs (see Table 2). These results were worse
stages of the plant, where approximately 100 mg/l is than those given by the P&T method. However, the
the total concentration (TTHM) determined in the eciency of CLSA for all ITHMs was much higher.
distribution system. However, as ITHMs concen- Thus, recoveries ranged from 47% (CHCl
tration was expected to be two orders of magnitude is the least favorable case, to 94% (CHClI
I), which
), the
2
2
lower than TTHM, HS can not be used because the compound with highest recovery when the unsalted
recovery values were too low for an accurate deter- method was used. An acceptable value of 55% was
mination.
3
obtained for CHI spiked at 10 mg/l. Salt was added
Purge & trap. This is a widely used technique for to increase the ionic strength of the aqueous matrix
extracting and concentrating volatile organic com- and to improve the purge eciency (Hwang et al.,
pounds in water. EPA Methods 502.1 and 502.2 are 1984), but the recovery values remained practically
applicable to chloro-, bromo- and chlorobromotri- constant except for iodoform, which recovery was
halomethanes because they have high enough vola- signi®cantly improved to 82.8% at 1.5 mg/l level. In
tility and low enough water solubility to be spite of this result, salting can be avoided for the
eciently removed from water samples with this analysis of ITHMs.
technique. CHCl
3
, CHBrCl
2
, CHBr
2
Cl and CHBr
3
Liquid±liquid extraction. Water samples were
were totally removed from water, independently of spiked at three di€erent concentration levels (10,
the trap used. However, ITHMs (except iodoform) 1.5 and 0.5 mg/l) with ITHMs. Recovery values
gave recoveries from 5 to 70%. When the Tenax/ close to 100% were obtained for all compounds at
silica gel/charcoal trap (recommended EPA trap) the range of concentrations studied. In addition,
was tested, an acceptable recovery was obtained for this method presented good reproducibilities, inde-
CHCl
2
I and CHBrClI (close to 65 and 42% at 10 pendently of the concentration level, with relative
and 1.5 mg/l, respectively). However, the eciency standard deviations lower than 5%.
of the method decreased rapidly for the other com-
LLE/GC/ECD method
pounds. In order to achieve better recoveries, some
modi®cations in the experimental conditions were
The microextraction method was the only method
performed. Thus, by increasing the helium purge capable of extracting the six iodinated trihalo-
¯
improved for CHClI , CHBrI and CHI (50, 100 liquid extraction was found to be the method of
ow and heating the water sample, eciencies methanes quantitatively (Table 2). Once liquid±
2
2
3
and 200%, respectively with respect to values choice, other analytical parameters such as linear
shown in Table 2). However, with these modi®- detector response, detection and quanti®cation
cations, no signi®cant di€erences were observed for limits were studied.
CHCl I, CHBr I or CHBrClI. The results obtained
2
Several calibration standard solutions of each
2

3
386
Beatriz Cancho et al.
Sample preservation of ITHMs. To prevent
the linearity of the detector response. The linearity further formation of disinfection by-products
compound ꢀn ˆ 7±10† were measured to determine
2
and correlation coecients (r ) were determined by (DBPs), free chlorine must be eliminated when
plotting the curves of relative areas to bromochlor- samples are collected. Sample preservation requires
omethane standards (A/Abromochloromethane) vs the special attention. For example, NH Cl, as
4
a
concentration of each compound (Ci ). Linear quenching agent for chlorine, may not completely
ranges and the correlation coecients for ITHMs stop DBP formation and additional DBPs will be
are given in Table 3. For the six compounds, the formed at levels depending on several factors such
linear range was one order of magnitude. The as water characteristics, temperature and reaction
detection limits ꢀS=N ˆ 3† ranged between 0.01 time. On the other hand, sodium sul®te, the most
(
CHCl I, CHClI , CHBrI and CHI ) and 0.03 mg/l widely used quenching agent, can destroy DBPs
2
2
2
3
(
CHClBrI).
such as cyanogen halides, haloacetonitriles and
haloketones. This was also reported by Thomas et
al. (1980) who evaluated the stability of dichloroio-
domethane.
Applications
ITHM stabilities were studied in the presence of
chemical reductors such as ascorbic acid, sodium
thiosulfate and sodium sul®te. Four samples of
ultrapure water were spiked at 6.0 mg/l with a
methanolic standard solution which contained
chloro-, bromo-, iodo- and mixed trihalomethanes.
The ®rst contained no quenching agent. To the
other samples, 1 ml of 0.1 M ascorbic acid, 0.1 N
sodium thiosulfate and 0.1 M sodium sul®te were
added, respectively. This experiment was repeated
using four samples of raw water from Llobregat
Stability of ITHMs. In order to evaluate the stab-
ility of these compounds, three samples were pre-
pared in the laboratory. One liter of ultrapure
water was spiked with 3 mg/l of chloro-, bromo-
and mixed THMs and with 6 mg/l of ITHMs with a
methanolic standard solution that contained all the
compounds. The same procedure was performed
using raw water from the Llobregat river (TOC:
5
3
.60 mg/l, NH : 0.27 mg/l and pH: 8.50), which
was considered as an extreme case, and treated
water to which ascorbic acid solution had been
added to eliminate the free chlorine. The pH value
was not modi®ed. The three di€erent samples were
analyzed systematically by LLE/GC/ECD.
No degradation of ITHMS took place in ultra-
pure and treated water; their concentration
remained practically constant in both cases during
the period of study (three weeks). However, the
behavior observed in raw water was di€erent.
river (TOC: 5.60 mg/l, NH : 0.27 mg/l and pH:
3
8
.50). A total of eight samples were analyzed by
LLE/GC/ECD for 5 days, the maximum period of
residence time in the distribution system.
Figure 2 shows the variation of ITHM concen-
trations (%) vs time. Ascorbic acid and sodium
thiosulfate become suitable quenching agents for
the ten THMs, independently of the nature of the
water (ultrapure or raw water). Slight variations
with respect to initial concentrations were related to
the reproducibility of the method. On the other
hand, sodium sul®te is a suitable agent in ultrapure
water, but it produces a degradation of CHClI2,
Degradation of CHI
3
was observed after 1 day
whereas a decrease of the other iodinated trihalo-
methanes was noticed 7 days after the sample prep-
aration and in addition, four new peaks were
detected. In order to identify these new compounds,
the raw water was analyzed by GC/MS. Dibromo-,
chloroiodo-, bromoiodo- and diiodomethane were
identi®ed as degradation products of ITHMs.
Figure 1 shows as an example the results obtained
for treated water and river raw water, respectively,
in the initial 7 days.
2 3
CHBrI and CHI in raw water.
E€ect of di€erent matrices on ITHMs quanti-
tation. The optimized LLE method was applied for
the determination of ITHMs by spiking samples of
ultrapure water and samples of ITHM-free chlori-
As a conclusion, ITHMs would not be degraded
during the water residence time in the distribution
system and then, ITHMs will reach the tap water.
Table 4. Estimated concentrations and relative standard deviations
(%) of ITHMs spiked at 1.0 mg/l in ultrapure and treated waters
determined by LLE/GC/ECD method
a
ITHMs
Concentration (mg/l)
Ultrapure water Treated water
a
Table 3. Analytical parameters of the LLE/GC/ECD method
Compound
Linearity range
mg/l)
Correlation coecient
(r )
LOD
(mg/l)
(
Mean
2SD
Mean
2SD
CHCl
CHBrClI
2
I
0.05±22
0.07±24
0.05±25
0.04±37
0.03±36
0.05±31
0.999
0.999
0.998
0.999
0.999
0.999
0.01
0.03
0.02
0.01
0.01
0.01
CHCl
CHBrClI
2
I
1.02
1.02
1.02
1.00
1.02
1.04
0.091
0.165
0.004
0.009
0.022
0.066
1.02
1.02
1.04
0.98
0.96
1.04
0.037
0.057
0.422
0.042
0.059
0.302
CHBr
CHClI
2
I
CHBr
CHClI
2
I
2
2
CHBrI
CHI
2
CHBrI
2
CHI
3
3
a
a
ꢀ
n ˆ 4† LOD and LOQ.
(n ˆ 3).

Iodinated trihalomethanes in water treatment processes
3387
Fig. 1. Stability of ITHMs in ultrapure (right) and raw water (left).
nated water. Triplicate samples of ultrapure and values were compared using the F Fischer test
treated water, spiked at 1.0 mg/l, were analyzed. (95% probability) and the Student test (95% prob-
Prior to the analysis of the spiked treated water, ability and two sides), respectively (Miller and
ascorbic acid solution was added in order to elim- Miller, 1993), in order to evaluate the e€ect of the
inate free chlorine. Quanti®cation was performed di€erent matrices. No signi®cative di€erences were
using the calibration curve for each compound found between the concentrations determined in
relative to the internal standard (1,2-dibromopro- both samples. Results con®rmed that LLE is a
pane). Results obtained for both matrices are feasible method to quantify ITHMs independently
given in Table 4. Standard deviations and mean of the nature of the matrix.
Fig. 2. Sample preservation of ITHMs, considering di€erent quenching agents for chlorine, in ultrapure
water (left) and raw water (right).

3
388
Beatriz Cancho et al.
Iodinated trihalomethanes at the Barcelona Water process. Brominated and chlorobrominated concen-
Treatment Plant. The LLE method was applied for trations were higher than chlorinated ones, account-
the determination of ITHMs in the Sant Joan ing for 83% of the total THMs (TTHMs). TTHM
Â
Despõ Water Plant (Barcelona, NE Spain), where concentration in prechlorination step was lower
water from Llobregat river is treated. Llobregat than the value determined after sand ®ltration and
river has a Mediterranean regime, which is charac- ozonation. Samples of prechlorination step were
terized by irregular ¯ows. The quality of the raw collected several minutes after the addition of chlor-
water is a€ected by industrial dumpings and salt ine, although the oxidation reaction continued for
mines located in the upper course of the river. The 3±4 h until the water reached the sand ®lters. This
average level of TOC, inorganic bromide and iodide explains the higher values found in sand ®ltered
in raw water in 1997 was 5.0 mg/l, 0.7 mg/l and samples. When the ozonated water had passed
3
.2 mg/l, respectively. Several years ago, the average through GAC ®lters, a decrease in TTHM was
bromide concentration was even higher (ca. 7 ppm observed. The eciency of the activated carbon is
in 1982) but, after a brine collector was built to dis- always greater for compounds with high number of
charge salt from the mines to the sea, these levels bromine atoms (Ventura et al., 1982) and depends
were reduced dramatically.
on the ®lter performance. TTHM concentration
The water treatment plant carries out convention- increased again with the postchlorination dosage.
al treatment consisting of prechlorination (to break- The average TTHM level in tap water correspond-
point), ¯occulation (settling), sand ®ltration, ozona- ing to the period of time studied was 90.5 mg/l.
tion, granular activated carbon (GAC) ®ltration ITHMs were determined by LLE/GC/ECD. The
and postchlorination (with a lower dosage of chlor- MtBE extracts were injected in the DB-1 column
ine, to guarantee a 0.5±1 mg/l concentration of due to coelution of CHCl I and CHBrClI with the
2
chlorine in the distribution system).
bromochloro- and dibromopropanones. Then, DB-
Table 5 shows the results obtained for THMs in 624 column was used for con®rmatory purposes for
analyses performed from January to June of 1998. these iodinated compounds (see Fig. 3). Only three
THMs (bromo-, chloro- and bromochloroderivates) of the six possible iodinated compounds were ident-
were generated in all the stages of the treatment i®ed in the water treatment plant. CHCl I was
2
Fig. 3. GC/ECD chromatogram of a LLE extract from ozonated water (DB-624 column). Compound
identi®cation: 1, chloroform; 2, carbon tetrachloride; 3, trichloroethylene; 4, dibromomethane; 5, bro-
modichloromethane; 6, 1,1-dichloropropanone; 7, dichloroacetonitrile; 8, 1,1,2-trichloroethane; 9, tetra-
chloroethylene; 10, chlorodibromomethane; 11, bromoacetonitrile; 12, dichloroiodomethane; 13, 1-
bromo-1-chloropropanone; 14, 1,1,1-trichloropropanone; 15, bromochloroacetonitrile; 16, bromoform;
1
7, 1,1-dibromopropanone; 18, bromochloroiodomethane; 19, dibromoacetonitrile; 20, dibromoiodo-
Ã
methane. denotes internal standard (bromochloromethane) and P denotes surrogate standard (1,2-
dibromopropane).

Iodinated trihalomethanes in water treatment processes
3389
a
Table 5. Average ITHMs concentration (mg/l) in the Barcelona water treatment plant (from January to June, 1998)
Trihalomethanes
Concentration (mg/l)
Sand ®ltration Ozonation
Raw water
Prechlorination
CAG
Distribution system
CHCl
CHBrCl
3
0.2
±
0.1
0.2
±
±
±
±
±
8.7
14
21
13
0.1
0.2
0.1
±
15
24.5
39
14
23
37
27
±
0.2
0.1
±
±
10
19
29
17
±
±
±
±
±
9.5
21
34
26
±
±
±
±
±
2
CHBr
CHBr
CHCl
2
Cl
3
29
2
I
0.2
0.6
0.2
0.1
±
CHBrClI
CHBr
CHClI
2
I
2
CHBrI
CHI
TTHMs
2
±
±
57.1
3
±
0.5
±
108.6
±
101.3
±
75
±
90.5
a
ꢀ
n ˆ 21†, ±, below LOD.
determined in prechlorinated, sand ®ltered and oc- longer than the normal residence time in the distri-
casionally in ozonated waters and CHBrClI and bution system, they could be present in treated
CHBr
sum of the three compounds was less than 1 mg/l in of the Barcelona treatment plant indicated that
all the samples. They were totally adsorbed in GAC CHCl I, CHBrClI and CHBr I were formed during
lters and they were not generated again with post- the treatment process, at levels lower than 1 mg/l,
2
I were also identi®ed in ozonated water. The water. Results of analyses from the di€erent stages
2
2
®
chlorination at detectable levels.
but they were not detected in the distribution sys-
Since last year the river water quality dropped tem during the period studied.
due to the drought. This unusual event forced us to
modify the treatment process. River raw water after
AcknowledgementsÐThis work has been partially funded
by AWWARF (Project RFP365). We are indebted to
sand ®ltration was diluted with groundwater prior
to ozonation. This fact prevented the formation of Djanette Khiari from Metropolitan Water District of
the high contents of brominated and iodinated deri- Southern California (MWDSC), co-ordinator of the pro-
gram, for her technical assistance. We are also grateful to
vates usually found in the average water when it is
Sonia Peiro
Â
for her co-operation.
coming exclusively from the river. Although
ITHMs were not identi®ed in treated water, they
cannot be ruled out as potential causes of medicinal
odor due to the limited period of time studied. The
REFERENCES
estimated odor threshold concentrations are very Ali-Mohamed A. Y. and Jamali S. G. A. (1989) Potentio-
metric determination of ¯uoride and iodide in natural
and processed waters of Bahrain using ion-selective elec-
low for CHBrI
2
(0.8 mg/l) and for CHClI (1.1 mg/
2
l), whereas higher values were obtained for CHBr2I,
trodes. Water Res. 23, 659±662.
Anselme C. and N'Guyen K. (1986) Proble
gouts survenus sur le reseau du Pecq, de RPO. Internal
CHCl I and CHBrClI (6.4, 8.0 and 8.4 mg/l, re-
2
Á
mes de mauvais
spectively) (Cancho and Ventura, 1999c). These
results show that levels of some ITHMs around
Ã
Â
Report Lyonnaise des Eaux.
APHA AWWA WEF (1995) Standard Methods for the
Examination of Water and Wastewater, 19th ed. Ameri-
can Public Health Association, Washington, DC.
1
.0 mg/l could be associated with odors in water.
Auger M. V. (1908) Sur les derive
methane. Comp. Rend. Acad. Sci. 1034±1037.
Â
s trihalogenes mixtes du
  Â
CONCLUSIONS
Bellar T. A., Lichtenberg and Kroner R. C. (1974) Deter-
mining volatile organics at microgram-per-litre levels by
gas chromatography. J. Am. Water Works Assoc. 66,
703±706.
Brass H. J., Feige M. A., Halloran T., Mello J. W.,
Munch D. and Thomas R. F. (1977) The national or-
ganic monitoring survey: samplings and analyses for
purgeable organic compounds. In Drinking Water Qual-
ity Enhancement Through Source Protection, ed. R. B.
Pojasek, p. 398. Ann Arbor Science Publishers, Michi-
gan.
Iodinated trihalomethanes, related to pharmaceu-
tical or medicinal odors and taste in water, were
synthesized and di€erent analytical methods for
their determination in water were evaluated. HS
and P&T were not suitable techniques to recover all
ITHMs quantitatively and CHI was not recovered
3
at low concentration levels (1.5 mg/l). CLSA
unsalted and salted) showed acceptable recovery
(
values between 50 and 90% for all ITHMs but
liquid±liquid microextraction was the procedure
which gave the highest recoveries. As a consequence
it is proposed as a fast, inexpensive and reproduci-
ble method for the analysis of ITHMs in water
using the ascorbic acid or sodium thiosulfate as
suitable quenching reductors for sampling. As
ITHM degradation occurs in a period of time
Bruchet A., N'Guyen K., Malleviale J. and Anselme C.
(
1989) Identi®cation and behaviour of iodinated halo-
form medicinal odor. In AWWA Water Quality Technol-
ogy Conference. Denver, CO, pp. 125±141.
Bunn W. W., Deane E. R., Haas B. B. and Kleopfer R.
D. (1975) Formation of trihalomethanes by chlorination
of surface water. Environ. Letters 10, 205±213.
Cancho B., Ventura F. and Galceran M. T. (1999a) Solid-
phase microextraction for the determination of iodi-

3
390
Beatriz Cancho et al.
nated trihalomethanes in drinking water. J. Chromatogr.
A 841, 197±206.
Cancho B., Ventura F. and Galceran M. T. (1999b) Beha-
viour of halogenated disinfection by-products in the
water treatment plant of Barcelona (Spain). Bull.
Environ. Contam. Toxicol. 63, 610±617.
formation of organoiodinated compounds: the import-
ant role of ammonia. Environ. Sci. Technol. 33, 336.
Hanai Y., Saitou A. and Mio N. (1996) Determination of
iodide trihalomethanes in drinking water. Bull. Inst.
Environ. Sci. Technol. Yokohama National Univ. 22, 11±
16.
Cancho B. and Ventura F. (1999c) (in preparation).
Carnicero M., Rodrõguez J. J., Valero F., Fernandez-Tur-
 Â
iel J. L., Roig A., Antich N. and Lobo A. (1998) Appli-
Hansson R. C., Henderson M. J., Jack P. and Taylor R.
D. (1987) Iodoform taste complaints in chloramination.
Water Res. 21, 1265±1271.
cation of ICP±MS to water analysis:hydrochemical Hass H. B. and Newton R. F. (1993) Handbook of Chem-
characterization of llobregat basin (Barcelona, N.E. istry and Physics, 74th Edition, ed D. R. Lide, CRC
Spain). In 5th International Symposium on Analytical Press, Inc, Florida, 15±16.
Methodology in the Environmental Field, A CorunÄa, Hine J. and Procer P. (1958) Methylene derivatives as in-
Spain.
Coleman W. H., Lingg R. D., Melton R. G. and Klop¯er
F. C. (1976) The occurrence of volatile organics in ®ve
termediates in polar reactions. XIII. The basic hydrolisis
of bromochloroiodomethane. J. Am. Chem. Soc. 80,
4282±4284.
drinking water supplies using gas chromatography-mass Hwang C. J., Krasner S. W., Moylan R. S. and Dale M.
spectrometry. In Identi®cation and Analysis of Organic
Pollutants in Water, ed. L. H. Keith, pp. 305±327. Ann
Arbor Science Publishers, Ann Arbor, MI.
S. (1984) Determination of subnanogram per liter levels
of earthy-musty odorants in water by the salted closed-
loop stripping method. Environ. Sci. Technol. 18, 535±
539.
Dietrich A. M., Dodd J. P., Mirlohi S. and Da Costa W.
F. (1997) Flavor pro®le analysis and GC/MS detection Kleopfer R. D. (1976) Analysis of drinking water for or-
of iodinated disinfection by-products in drinking water
from the USA space program. In 5th International Sym-
posium on O€ Flavours in the Aquatic Environment.
Paris (France).
EEC (1980) Directive relating to the quality of water
intended for human consumption 80/778/ECC (OJI.
ganic compounds. In Identi®cation and Analysis of Or-
ganic Pollutants in Water, ed. L. H. Keith, pp. 399±416.
Ann Arbor Science Publishers, Michigan.
Kolb B. and Ettre L. S. (1997) Static Headspace-Gas
Chromatography: Theory and Practice. Wiley-VCH,
New York, pp. 260±272.
2
29.30.8.80).
Miller J. C. and Miller J. N. (1993) Estadõ Âs tica para QuõÂ-
mica AnalõÂtica, 2nd ed. Addison-Wesley Iberoamericana.
Prolog 4.2/D. CompuDrug Chemistry Ltd: Budapest,
Hungary.
Ettre L. S., Welter C. and Kolb B. (1993) Determination
of gas±liquid partition coecients by automatic equili-
brium headspace-gas chromatography utilizing the
phase ratio variation method. Chromatographia 35, 73± Rickabaugh J. and Kinman R. N. (1978) Trihalomethane
8
3.
formation from iodine and chlorine disinfection of Ohio
river water. In Water Chlorination: Environmental
Impact and Health E€ects, Vol. 2, eds R. L. Jolley, H.
Gorchev and D. H. Hamilton, pp. 583±591. Ann Arbor
Science Publishers, Ann Arbor, Michigan.
Fair P. S. (1992) Status report on analytical methods to
support the disinfection by-products in chlorinated
drinking water. USEPA, Oce of Ground Water and
Drinking Water, Technical Support Division, August.
Frazey P. A., Barkley R. M. and Siervers R. E. (1998) Rook J. J. (1974) Formation of haloforms during chlori-
Solid-phase microextraction with temperature-pro-
grammed desorption for the analysis of iodination disin-
fection-byproducts. Anal Chem. 70, 638±644.
Gittelman T. S. and Yohe T. L. (1989) Treatment of the
iodinated haloform medicinal odors in drinking water.
In AWWA Water Quality Technology Conference, Den-
ver, CO, pp. 105±123.
nation of natural waters. J. Water Treatment Exam. 23,
234±243.
Schall C., Heumann K. G. and Kirst G. O. (1997) Bio-
genic volatile organoiodine and organobromine hydro-
carbons in the Atlantic Ocean from 428N to 728S.
Fresenius J. Anal. Chem. 359, 298±305.
È
Sulzle D. (1987) Reactions of carbenes with anions. Tetra-
hedron Lett. 28, 2769±2770.
Glaze W. H., Henderson J. E. and Smith G. (1975) Analy-
sis of new chlorinated organic compounds formed by Thomas R. F., Weisner M. J. and Brass H. J. (1980) The
chlorination of municipal wastewater. In Water Chlori-
nation Environmental Impact and Health E€ects, Vol. 1,
ed. R. J. Jolley, pp. 139±159. Ann Arbor Science Pub-
lishers, Ann Arbor, MI.
Gould J. P., Giabbai M. and Kim J. (1985) Formation of
iodinated trihalomethanes. In Water Chlorination: En-
vironmental Impact and Health E€ects, Vol. 5, eds R. L.
Jolley, R. J. Bull, W. P. Davis, S. Katx, M. H. Roberts
and V. A. Jacobs, pp. 923±926. Ann Arbor Science Pub-
lishers, Ann Arbor, MI.
®fth trihalomethane: dichloroiodomethane, its stability
and occurrences in chlorinated drinking water. In Water
Chlorination: Environmental Impact and Health E€ects,
Vol. 3, eds R. L. Jolley, W. A. Brungs, R. B. Cummings
and V. A. Jacobs, pp. 161±168. Ann Arbor Science Pub-
lishers, Ann Arbor, Michigan.
Vel Leitner N. K., Vessella J., Dore M. and Legube M.
(1998a) Conditions of formation of organoiodinated
compounds during chlorination of aqueous solutions
containing iodide ions. Rev. Sci. Eau 11, 445±457.
Gould J. P. (1993) The impact of iodide on the formation Vel Leitner N. K., Vessella J., Dore M. and Legube M.
of trihalomethanes. In AWWA Water Quality Technol-
ogy Conference. Miami, FL, pp. 107±123.
Grob K. (1973) Organic substances in potable water and
in its precursorÐmethods for their determination by
gas liquid chromatography. J. Chromatogr. 84, 255±273.
Hatch and Gary L. (1999) Comment on chlorination and
(1998b) Chlorination and formation of organoiodinated
compounds: the important role of ammonia. Environ.
Sci. Technol. 32, 1680±1685.
Ventura F., Rivera J., Guardiola J., Perramon J. and Sal-
vatella N. (1982) Contol of trihalomethanes in Barcelo-
na's drinking water supply. Aqua 5, 469±473.