Product Design
Low Concentrations of Carbonyl Sulfide in Air Measured by Three Analytical Methods
Abstract
Reports on a study involving the measurement of carbonyl sulphide in mainstream and sidestream cigarette smoke using three different methods: an electron capture detector; mass spectrometry; and a Beckman carbon monoxide infrared analyzer. Notes that an electron capture detector seemed to be best suited for carbonyl sulphide measurement. Discusses material and methods, noting that measurements were taken of mainstream smoke from three different commercial cigarette brands and three experimental cigarettes. Notes that in commercial cigarette brands carbonyl sulphide concentrations were 25-30 ppm for mainstream smoke while they were 25-60 ppm for the experimental cigarettes; measurements on commercial cigarette sidestream smoke were eight to ten times lower than in mainstream smoke. Includes results and discussion. Attaches figures, tables, and references.
Fields
- Author
- Hugod, C
- Kamstrup, O.
- Larsen, E.
- Hypothesis
- Mainstream constituent yieldsModification of selected mainstream smoke constituents in response to health concerns.
- Sidestream constituent yieldsModification of selected sidestream smoke constituents in response to health concerns.
- Smoke constituent testingDevelopment of methods for measurement of gas and particulate yields in mainstream and sidestream smoke.
- Keyword
- Animal testing
- Mildness (Attribute measure)
- Sidestream smoke
- Strength attributes
- Toxicity
- Smoke Constituent
- Carbonyl sulphide
- Gas phase constituents
- Design Component
- Air-cured tobacco
- Flue-cured tobacco
- Named Organization
- Beitrage Zur Tabakforschung International
- Forschungsgellschaft
- Riso National Laboratory
- Scandinavian Tobacco Company
- Subject
- chromatography
- gas phase
- Secondhand Smoke/Constituents
- spectrometry
- Test/Smoke Constituents (Testing)
Document Images
.
LOW' CON!Cr.NTRATIONS OF' CARBONYL SULPHIDE IN AIR
M.EASUR,D BY THREE ANALYTICAL :Sr.THODS.
By
Ole Kamstrup and Carl Hugod
Department of Clinical Chemistry
Rigshospitalet, Blegdamsvej 9,
DK-2100 Copenhagen, Denmark,,
an d
Elfinn Larsen
Chemistry Department
RisoNational Laboratory,.
DK-4000 Roskilde, Denmark.
jp a
BEITRXGE ZUR TABAKFORSCH'UNG INTERNATIONAL
The attached paper has been submitted for, publication.
Please look through lt and advise us by _., 14~M M~~,1~79
If you consitler alterations necessary.
Dr. Bentley
and
Dr. Kuhn
are requested to fully review the
manuscript and submit their
comments on the enclosed form.
date:
The Editorial Office
(Harvestehuder Weg 88, 0-2000 Hamburg 13)
?I
-2-
INTRODUCTIOH
Carbonyl sulphide (COS) has been reported to be pre-
sent in the gas phase of cigarette smoke in a concentra-
tion of 20-40 ug/cigarette (5,11) or 150-500 ug/cigaret-
te (10).
We have for some years been performing selective ex-
posure experiments with rabbits for the study of the
~ -
histotoxic lung effect and the atherogenic effect of gas
phase components of tobacco smoke (6,7,8); the methods
reported here for determination of COS in air at concen-
tration levels up to 120 ppm were d'evelooed for applica-
tion in our exposure studies with COS (9). These methods
comprise: mass spectrometry (.,S) gas chromatography with
an electron capture detector (ECD)., and measurements by a
Beckman CO-meter (B-CO).
During the present study ECD appeared to be suited to
the measurement of the COS content both in main stream-
and side stream smoke from cigarettes; these measurements
are reported here.
MATERIAL AND~METHODS
An air-tight exposure cha^mber with~a volume of about
101m3 (1) was used to obtain air with different COS con-
centrations. COS was delivered from a;gas cyLinder, pas-
sed'.througii a rotameter, and'mixed with atmospheric air
immediatelyy before the entrance to the cPahmber. A con-
stant flow rate of COS and atmospheric air was maintained
f
0
- 3 -
for 24 hours before air-gas sampl~es .for-determiningof
the--concentration were drawn from the middle of the
4
'
4
L-
.
chamber interilor, by means of a reciprotor-type pumpVV,-N,,r dekrw#n,
~'k+4 CC1~GtwtiFrakicst..
Mass spectrometry
A Varian MAT CH 5D mass spectrometer equipped with a
gas inlet system was used'. The signals were detected by
an electron multiplier. The temperature of the gas inlet
system was 130°C and of the ion source 200 °C. Ionizing
energy was 70 eV with 100 pA trap current. The gas samp-
Les were transferred from the exposure chamber to the mass
spectrometer in 100 ml glass pipettes with stopcocks
greased with Apiezon N'. The mass spectra were obtained
on a UV recorder (Fig. 1) whereas the two masses m/e 36
and m/e 60 corresponding to 36Ar and the molecular ion
of COS respectively were registered on a Honeywell recor-
der. The stable isotope 36ar at a concentration of 31.5
ppm in dry air (3) was used as internal standard.
Gas chromatographY
A 1 m stainless steel column, 2
mm i.d., packed with
Porapak QS 80-100 mesh was connected to an electron cap-
ture detector (63Ni). High purity nitrogen was used as
carrier gas with a flow`rate of 50 ml/mdn. The system
was operated at room temperature. A gas sample of 200
U1 was injected by a Hamilton syringe through a silicone
septum. The system was calibrated with air samples con-
- 4 -
`f
taining a known amount of COS. These were prepared by in-
jecting:pure COS into 2 1 volumetric glass flasks with
rubber plugs.
Beckman CO-r..eter
A Beckman infrared analyzer, model 215 equipped for
carbon monoxide (CO) determination measured the content
of COS in a continuous gas flow sucked from the expcsure
chamber by the reciprotor pump. The instrument was cali-
brated against normal air and air with 180 ppm CO.
COS~concentration in cigarette smoke
Cigarettes were mounted in a single port smoking ma-
chine (developed by Artho and Koch, Burrus, Switzerland).
By means of a Hamilton-syringe 200 u;L samples of main-
stream smoke which~had been passed through a Cambridge
filter were drawn from the freshly generated smoke of the
sixth puff. Such samples were immediately injectedinto
gus cf~.c..,aFr~ z~~h
the~~Cr7:
To perform corresponding measurements on side-stream
smoke 20 burning,cigarettes mounted in a holder were
placed under a glass gl~obe from which smoke samples could
be drawn at the top with a syringe. Measurements were
performed in the main-stream smoke from three commercial
cigarette brands (A: strong, non-filter, B: medium, fil-
ter, and C: mild, filter cig3rette), and also from three
=experimental cigarettes produce6from pure, unflavoured
- J -
tobaccos: (D, flue-cured, E, sun cured, and F, air-cured).
COS measurements on side-stream smoke were performed on
the three commercial brands only.
RESULTS, AND DISCIISSIOY
A mass spectrum,of a gas sample from the exposure
chamber containing about 115 ppm COS is shown in Figure 1. In
7
~ this figure the upper half demonstrates the normal mass
spectrum. In the case of the 5000-times higher amplifi-
cation (lower half)~ it is seen~that the signals of the
molecular ions of COS at m/e 60 and of 36Ar (31.5 ppm) at
m/e 36 are very suited to determining COS in the range of
1-150 ppm. The response at m/e 64 is from 502, which is
a background impurity in the mass spectrometer. Gas sam-
ples of air without COS show a background response at m e
60 corresponding to an equal amount of 10 ppm COS. How-
ever, at any time when the same air sample size was used
and controll'ed by the signal of 36Ar it was possible to
correct for this background and to measure the content of
.
COS down to concentration levels of 1-2 ppm. -~
The sensitivity factor for Ar and COS was estimated O
by measuring samples of pure Ar and pure COS
The respon- 0
M
,~ ~ ses were expressed agai:~'ts the partial pressure in the gas CA
inlet system.
N
~
For control of the decomposition of COS which might be ~
- caused by the use of uncoated glass pipettes (12), two ex-
periments were performed at concentration levels of about
40 and 115 ppm, respectively, by collecting the gas sam-
ples in the glass pipettes and in 5 1 volume plastic bags
(Saran). The difference was w=thin the experimental er-
ror.
Figure 2'iLlustrates a calibration curve for COS ob-
tained on the ECD. At concentrations higher than 10 ppm
the curve flattens out. Therefore in the range of 50
ppm it is possible, without dilution of the air-gas sam-
ple, to estimate the amount of COS alone with an absolute
?
e~fdd2~~tc-- certainty of about 5 ppm.
atG~~rtc: ,
has r4i
For comparison of the three analytical methods air-
gas samples were taken from the exposure chamber in im-
mediate sequence and subsequently analysed by the B-CO,
the ECD and the MS. The results are shown in Figure 3
and Tablie 1. It can be seen that the three methods-give-
Ct~E Gla52L~4 rC.lca:c~..
.a_.r}ood-agreeinent. Due to the background effect the COS
values obtained by MS have a high uncertainty in the range
Se..s. hv4v
lower than 10 ppm, whereas the ECD has a high a~c.curacy in
this range (Figure 2). The COS values obtained by the
B-CO are not absolute, as this apparatus was calibrated
against CO and not against COS. However, measurements of
the B-CO against the MS demonstrated a linear response in
the 10-120 ppm range.
Using the ECD method the results of ineasurements of
COS content of main-stream and side-stream smoke of the
various cigarette types are shown in Table 2. In commer-
cial cigarette brands COS-concentrations of 25-30 ppm
:.y f~qf
- 7 -
were found while values of 25-60 ppm were obtained from
the experimental cigarettes. These figures are in agree-
ment with those reported earlier by Phili~ppe et al. (11),
and Horton~et al. (5), but substantially lower than those
reported by Osborne et all. (10). Measurements on side-
stream smoke were performed only on the ccmmercial ciga-
rettes; eight to ten times lower values than in main-
stream smoke were found.
Hoegg (4) has demonstrated'that for most of the im-
portant constituents of tobacco smoke the side-stream
smoke yields about 3-4' times the quantity found in main-
stream smoke. Recently Brunnemann et al. (2) demonstra-
ted that H0N is excepted from this rule, since higher
concentrations of HCN were found in main-stream smoke than
in side-stream~smoke. The present study demonstrates that
like HCN, COS is substantially more abundant in main-
stream thanii'in side-stream smoke.
Acknowledgement:
We are grateful to The Scandinavian Tobacco Company,
SOborg, Denmark for supplying laboratory facilities,
cigarettes and smoking machine for this study.
The work was supported by a grant from the Forschungs-
gesellschaft Rauchen ur.d Gesundheit, Hamburg, Germany.
i
sur=RY
Three methods are described for measuring carbonyl
sulphide at low concentration levels (1-120 ppm): mass
spectrom.etry,'gaschromatography with an electron cap-
ture detector, and a Beckman carbon monoxide infrared
analyzer. When the results obtained by mass spectrometry
were used as reference values the two other methods gave
linear responses. The practical applicability for gas
chromatography with an~electron capture detector for
-neasurements of the carbonyl sulphide content of main- and
side-stream cigarette smo}:e is demonstrated. The values
obtained revealedian approximately eight times higher
concentration of carbonyl suIphide in main-stream than in
side-stream cigarette smoke.
,
c~.
a/w
0L S9
09
S5 05
i_i_. , _L,
i
.
1 I-
.. Os .,s0:)
S9 09 SS OS
I I I.I
ko00252124
S7 07 S£ 0£ SZ 0Z
d
0005 X
-001
57 07
m
_4_ F _._,_1 _. 1 .
4jd9t
sc
L-~
z0
I
:SZ
u1~
i
i
F- 0S
<
A
