Chemical generation of atomic iodine for the chemical oxygen-iodine laser. II. Experimental results

Article abstract of DOI:10.1016/S0301-0104(02)00629-8

A new method for the chemical generation of atomic iodine intended for use in a chemical oxygen-iodine laser (COIL) was investigated experimentally. The method is based on the fast reaction of hydrogen iodide with chemically produced chlorine atoms. Effects of the initial ratio of reactants and their mixing in a flow of nitrogen were investigated experimentally and interpreted by means of a computational model for the reaction system. The yield of iodine atoms in the nitrogen flow reached 70-100% under optimum experimental conditions. Gain was observed in preliminary experiments on the chemical generation of atomic iodine in a flow of singlet oxygen.

Full text of DOI:10.1016/S0301-0104(02)00629-8

Chemical Physics 282 (2002) 147–157
www.elsevier.com/locate/chemphys
Chemical generation of atomic iodine for the chemical
oxygen–iodine laser. II. Experimental results
a,b,
a,b
a
a
ꢀ
Otomar Spalek
*
ꢀ
, V ꢁı t Jir ꢁa sek , Miroslav Censk ꢁy , Jarmila Kodymov ꢁa ,
Ivo Jakubec , Gordon D. Hager
b
c
a
Institute of Physics, Academy of Sciences, Na Slovance 2, Prague 182 21, Czech Republic
b
ꢀ
Institute of Inorganic Chemistry, Academy of Sciences, Re ꢀz 250 68, Czech Republic
Air Force Research Laboratory/DE, Kirtland Air Force Base, NM 87117, USA
c
Received 16 April 2002
Abstract
A new method for the chemical generation of atomic iodine intended for use in a chemical oxygen–iodine laser (COIL)
was investigated experimentally. The method is based on the fast reaction of hydrogen iodide with chemically produced
chlorine atoms. Effects of the initial ratio of reactants and their mixing in a flow of nitrogen were investigated experi-
mentally and interpreted by means of a computational model for the reaction system. The yield of iodine atoms in the
nitrogen flow reached 70–100% under optimum experimental conditions. Gain was observed in preliminary experiments
on the chemical generation of atomic iodine in a flow of singlet oxygen. Ó 2002 Elsevier Science B.V. All rights reserved.
Keywords: Atomic iodine; Atomic chlorine; Chemical oxygen–iodine laser, COIL
1
. Introduction
Population inversion on this transition is main-
tained by near resonant energy transfer from
1
The process of chemical generation of atomic
metastable excited oxygen, O ð D Þ, to iodine at-
2
g
iodine was studied for use in the chemical oxygen–
iodine laser (COIL). This laser is the shortest
wavelength high-energy chemical laser operating
on the electronic transition of the P ground term
of atomic iodine at 1:315 lm
oms in the ground state
2
1
2
3
À
Ið P3=2Þ þ O
k
2
ð D
g
Þ ! Ið P1=2Þ þ O
2
ð R Þ
g
ð2Þ
2
À11
3
cm molecule s :
À1 À1
2
¼ 7:6 Â 10
In conventional COIL, atomic iodine is produced
2
2
Ið P1=2Þ ! Ið P3=2Þ þ hm
ð1Þ
by dissociation of molecular iodine through pooling
1
processes which consume 4–6 O
2
ð D
g
Þ molecules
ð3Þ
1
3
À
g
I þ nO ð D Þ ! 2I þ nO ð R Þ n ¼ 4–6:
2
2
g
2
1
*
The energy of O
process can be utilised for laser pumping and an
increase in laser power up to (25%) [1] can be
2
ð D
g
Þ saved by avoiding this
Corresponding author. Tel.: +420-2-66052603; fax: +420-2-
6890265.
E-mail address: spaleko@fzu.cz (O. Spalek).
8
0
ꢀ
301-0104/02/$ - see front matter Ó 2002 Elsevier Science B.V. All rights reserved.
PII: S0301-0104(02)00629-8

ꢀ
148
O. Spalek et al. / Chemical Physics 282 (2002) 147–157
achieved. Another drawback of using molecular
iodine as a precursor of iodine atoms in COIL is
the fast quenching reaction
ClO þ 2NO ! Cl þ 2NO
ð5Þ
2
2
This reaction represents a branched chain reaction
system consisting of the following steps [6]:
2
2
Ið P1 Þ þ I ! Ið P Þ þ I2
=2
2
3=2
ð4Þ
À11
3
cm molecule s :
À1 À1
ClO þ NO ! ClO þ NO
k6 ¼ 3:4 Â 10
2
2
k
4
¼ 3:5 Â 10
ð6Þ
ð7Þ
ð8Þ
À13
3
cm molecule s ;
À1 À1
Iodine vapour for COIL operation is produced by
evaporation of solid or liquid iodine, which is
difficult to control due to a time-varying interfacial
surface area and high temperature gradients in the
bed of solid iodine. In addition, the overall iodine
management system must be heated to 80–100 °C
to prevent solid iodine precipitation. To avoid or
suppress these problems, and to increase the COIL
efficiency, other methods of generating atomic io-
dine have been investigated. Endo et al. [2] pro-
ClO þ NO ! Cl þ NO
2
À11 ðÀ120=TÞ
3
cm molecule s ;
À1 À1
k7 ¼ 1:7 Â 10
e
Cl þ ClO ! 2ClO
2
À11
3
cm molecule s :
À1 À1
k
8
¼ 5:9 Â 10
Reaction (6) initiates the chain. Reaction (8) is
the chain-branching step with Cl and ClO as chain
carriers. For the fast and efficient course of the
whole process, other processes must not exhaust
Cl atoms and ClO radicals. Atomic chlorine is the
main reaction product for the initial ratio
duced iodine atoms by dissociating I
2
in a
microwave discharge. Pazyuk et al. [3] produced
iodine atoms with electrical discharges in alkyl
iodine vapours.
In our laboratory, a new method was proposed
based on fast chemical reactions of gaseous reac-
tants [4,5]. Experimental results on the chemical
generation of iodine atoms obtained on a small-
scale device are presented in this paper. The flow
conditions in this device are similar to the plenum
conditions in a supersonic COIL.
NO:ClO ¼ 2:1. ClO radicals are formed pre-
2
dominantly for the ratio 1:1 [6,7]. Chlorine atoms
can be subsequently consumed by recombination
with third molecules [6,7]. The fastest loss
processes for both Cl and ClO are the termolecular
reactions with nitrogen dioxide, which is an
unavoidable by-product of the reactions (6) and
(7).
2
. Description of the reaction system
Cl þ NO
2
þ N
2
! ClNO
cm molecule s ;
2
þ N
2
ð9Þ
ð10Þ
ð11Þ
À30
6
À2 À1
The chemical system studied experimentally is
k
9
¼ 1:5 Â 10
based on a two-step reaction process. In the first
step, atomic chlorine is formed in the reaction of
chlorine dioxide with nitrogen oxide. The
chlorine then reacts in the second step with
gaseous hydrogen iodide to produce atomic io-
dine. Most of the kinetic data for this system
could be taken from the papers concerned with
the investigation of the chemical HCl and HCl/
ClNO þ Cl ! Cl
2
2
þ NO
2
k10 ¼ 3:0 Â 10À11 cm molecule s ;
3
À1 À1
ClO þ NO þ N ! NO Cl þ N
2
2
3
2
À31
6
À2 À1
cm molecule s :
k
11 ¼ 1:4 Â 10
Because reaction (10) is much faster than (9), it can
be assumed that Cl atoms disappear with the ef-
CO
paper [7]).
2
transfer lasers [6] (see also references in our
À30
6
fective rate constant of 3 Â 10
cm mole-
which is more than 20 times higher than
À1 À1
2
.1. Generation of chlorine atoms
cule
s
the rate constant for reaction (11). Reactions
(9) and (10) can be also considered as the recom-
bination of Cl atoms catalysed by NO2 and
nitrogen.
Chlorine atoms are produced in the fast exo-
thermic reaction of gaseous chlorine dioxide with
nitrogen oxide

ꢀ
O. Spalek et al. / Chemical Physics 282 (2002) 147–157
149
2
.2. Generation of iodine atoms
atomic iodine produced with flow conditions sim-
ilar to those in COIL, but with non-reactive gas.
Experiments were then performed in the flow of
reactive singlet oxygen. A small-scale device was
designed for the flow conditions that are typical
for the region upstream in the nozzle plenum of a
supersonic COIL. The experimental setup for
measurements that were performed in nitrogen is
shown schematically in Fig. 1. The chemical flow
reactor 1 is made of stainless steel tube of 10 mm
in i.d. Three injectors for the injection of second-
ary gases into the primary gas are inserted coax-
ially in the reactor bends. They are movable
allowing a change in their position and variable
time span between injections of reactants. They are
made of stainless steel tube of 5 mm in o.d. with
one, two or three rows of holes. The diameter and
number of the holes in each injector were calcu-
lated to achieve full penetration of each secondary
flow into the primary flow. The following flow
Iodine atoms are produced by the fast exother-
mic reaction of chlorine atoms with hydrogen iodide
Ã
2
À1
Cl þ HI ! HClðHCl Þ þ Ið P3=2Þ þ 132 kJ mol
k12 ¼ 1:64 Â 10
According to Maylotte et al. [8], about 70% of the
À10
3
cm molecule s :
À1 À1
ð12Þ
reaction exothermicity is transferred into vibra-
Ã
tional energy of HCl molecules.
In the COIL medium, iodine atoms will un-
dergo excitation according to reaction (2). In ad-
dition, recombination of iodine atoms takes place
in termolecular reaction with molecules of carrier
À32
6
À2 À1
À30 6
gas ðkN2 ¼ 4:2 Â 10
cm molecule s Þ, and
À2
molecular iodine ðk ¼ 3:7 Â 10
cm molecule
s Þ, respectively. Due to the high concentration of
I2
À1
NO in the reaction mixture, iodine atoms are lost
2
primarily in the reactions:
I þ NO
2
þ N
2
! INO
2
þ N
2
rates of reactants were used: (0:1 mmol ClO
À1
2
þ
ð13Þ
ð14Þ
13 ¼ 3:1 Â 10À31 cm molecule s ;
6
À2 À1
2
k
5 mmol N ) s in the primary flow, (0:1 mmol
À1
NO þ 1 mmol N ) s through the first and sec-
2
ond injector, respectively, and (0:1 mmol
À1
I þ INO
k14 ¼ 8:3 Â 10
Mathematical modelling, in which instantaneous
mixing was assumed, showed that Cl atoms are
exhausted from the reaction mixture by the ex-
2
! I
2
þ NO
2
HI þ 1 mmol N
2
) s through the third injector.
À11
3
cm molecule s :
À1 À1
Twenty-four holes were made in the first injector,
20 holes in the second injector, and 16 holes in the
third injector; all holes in the 0.4 mm i.d.
The primary gas flowing into the reactor con-
tremely fast reaction (12) provided that HI, ClO
and NO are added together at a single point [7].
This results in a deceleration of the chain reaction
2
2
sisted of 4–10% ClO in nitrogen. A mixture of NO
and nitrogen (10% NO) was introduced through a
single injector only or through the first and second
(
5) and to the overall process. The formation of
injectors simultaneously. A mixture of HI and N
2
atomic iodine should therefore be faster when HI
is admixed at a certain distance downstream from
the mixing point of ClO and NO streams [7]. The
(with 8–10% HI) was introduced in most experi-
ments through the third injector. The flow rates of
secondary gases were measured by the flow meters
2
injection of NO through two injectors separated
by some distance in the flow direction was also
proposed in order to diminish the loss of chlorine
atoms [6]. Optimal conditions for the highest I
yield in this arrangement were found from the
mathematical modelling [7].
6–8 with the calibrated sonic orifices. Flow rates of
À1
reactants were 50–240 lmol ClO
2
s , 100–600
À1
À1
lmol NO s , and 0–500 lmol HI s . The pres-
sure in the reactor was 1.5–4 kPa, with a gas ve-
À1
locity 130–160 m s . The reactants entered the
reactor at room temperature.
2
Gaseous ClO was produced on site by the re-
action of chlorine with sodium chlorite [6]
3
. Experimental
Cl þ 2NaClO ! 2ClO þ 2NaCl
For safety reasons, the partial pressure of ClO
should not exceed 4 kPa [6], in special devices 13
2
2
2
ð15Þ
The experimental investigation was first per-
formed in an environment of nitrogen to study the
2

ꢀ
150
O. Spalek et al. / Chemical Physics 282 (2002) 147–157
2
Fig. 1. Schematic of small-scale device for atomic iodine generation. 1 – flow reactor, 2 – ClO generator, 3, 4 – rotameters, 5 –
pressure reducer, 6–8 – flow meters, 9 – diagnostic cell, 10 – Ar ion laser, 11 – ISD cell, 12, and 13 – VIS photometry cells.
kPa [9]. Reaction (15) took place in the generator 2
where gaseous chlorine diluted with nitrogen (in
the ratio 1:20 to 1:50) passed through the column
cell 11 (11 mm i.d., 45 mm of inner length) in a
direction parallel to the gas flow. The optical cell
was close coupled to the reactor because the life-
time of iodine atoms in the gas mixture is short
(0.2–1 ms). The longitudinal gas flow through the
cell was chosen because of the low flow rate of
atomic iodine and the need for sufficient absorption
length. The gas velocity downstream of the reactor
with NaClO at the room temperature. The gen-
2
erator (a PVC tube 7 or 11 cm in i.d., 1 m long)
was filled with either 40% (w/w) aqueous solution
2 2
or powdered NaClO . The flow rate of ClO and
residual chlorine, respectively, was determined by
chemical analysis of potassium iodide solution
À1
1 and through the ISD cell was 80–90 m s . The
(
2.5%), in which the exit gas was trapped [10].
The gas mixture from the reactor 1 first entered
gain/loss data were processed on line with a PC.
The atomic iodine flow rate was also determined
indirectly from the measured concentration of I2
molecules that were formed by recombination of
the optical cell 11 used for the detection of atomic I
by the iodine scan diagnostics (ISD) [11]. This di-
agnostic is based on a narrow band tunable near
infrared diode probe laser that monitors the gain or
iodine atoms. The I concentration was measured
2
at some distance from the reactor by VIS absorp-
tion photometry at 488 nm. The flow rate of ni-
trogen dioxide that was formed in the reactor was
estimated by the same method (using an absorption
cross-section of 2:7 Â 10À19 cm ), from the light
2
2
absorption for the Ið P1=2Þ–Ið P3=2Þ transition at
1
3
:315 lm. The laser frequency is scanned over the
–4 hyperfine transition in the iodine atom. The
2
probe beam was double passed through the optical

ꢀ
O. Spalek et al. / Chemical Physics 282 (2002) 147–157
151
À1
absorption change upon adding NO to ClO
2
. The
passed 90 lmol Cl
2
and 2.5 mmol N
À1
2
s . This
with about
yield
photometry employed the argon ion laser 10, a
beam splitter, two optical cells 12 and 13, two sil-
icon photodiodes, and current amplifiers. The AD
converter and PC processed signals on-line from
the photodiodes, the gas flow meters 6–8, and the
temperature and pressure sensors. The gas trans-
port times were 0.2 ms from the third injector to the
entrance of the ISD cell 11, 0.6 ms to the ISD cell
yielded 140–150 lmol ClO
2
s
À1
15 lmol s of residual chlorine. The ClO
2
(80–85%) decreased with further increasing chlo-
rine flow rate. The ClO yield also declined with
‘‘aging’’ of the filler, a yield of only 40% was ob-
tained after two months’ intermittent use, even
though Cl2 reacted with only 12–15% of the
2
NaClO
by Arnold [6], is probably caused by a layer of
NaCl formed on the surface of the NaClO crys-
tals that hinders chlorine diffusion and so slows
down ClO production. When NaClO in the form
2
. The reason for this effect, also observed
2
exit, 6 ms to the centre of the first I cell, and 93 ms
to the second I
periment, ClO
2
cell. At the beginning of each ex-
and NO were introduced into the
2
2
reactor at constant flow rates, then HI was ad-
mixed with either a constant or variable flow rate.
At the end of each experiment, the HI was shut off
2
2
of 32% solution was used, 1 liter provided
À1
200–250 lmol ClO s with a yield of 81–85%.
2
first, then ClO flow, and finally the NO flow. This
2
This production rate was constant during 5 h of
operation, which corresponds to 90% utilisation of
procedure made it possible to determine the con-
tribution of NO and I , respectively, to the light
2
2
2
the NaClO . Water contained in the gas exiting the
absorption in the detection cells 12 and 13. The
temperature was measured by a Ni–NiCr thermo-
couple (1 mm in o.d. including the Inconel sheath)
placed at the beginning of the diagnostic cell 9. The
gas exiting the diagnostic cell 9 was exhausted by
solution could cause problems in a COIL medium;
its amount, however is small by comparison with
1
water vapour produced in the O
2
ð D
g
Þ generator.
4.2. Formation of molecular iodine
3
À1
the rotary pump ð25 m h Þ after passing a liquid
nitrogen trap.
The production of atomic iodine in a flow of
Molecular iodine may be formed by the re-
combination of I atoms and also by other reactions
contributing to the VIS photometry signal. The
2
singlet oxygen was performed in the same reactor
1
1
with the difference that an O
introduced as the primary gas. The oxygen was
2
ð D
g
Þ stream was
residual chlorine in the ClO flow reacts with HI
1
Cl
2
þ 2HI ¼ 2HCl þ I
2
generated in a small-scale O
ator. The performance of this generation is de-
scribed in detail in our earlier paper [12]. The
O
1
2
ð D
g
Þ jet-type gener-
À33
6
À2 À1
cm molecule s :
k16 ¼ ð3:4 Æ 1:7Þ Â 10
ð16Þ
1
2
ð D
g
Þ
concentration was evaluated from
:27 lm emission in a detection cell placed be-
The rate of this reaction was estimated in our
earlier investigation [4]. Modelling of the reaction
system producing iodine atoms showed that reac-
2
tion (16) may affect the I concentration measured
in the downstream cell 13, while in the upstream
1
tween the O
ClO mixture was introduced in the reac-
þ N
mix-
2
ð D
g
Þ generator and the reactor. The
2
2
tor through the first injector, the NO þ N
2
ture through the second injector, and HI þ N
2
mixture through the third injector.
cell 12 it may be neglected. The overall rate of
atomic iodine production was therefore calculated
from the I concentration measured in the up-
2
4
. Results and discussion
.1. ClO production
The solid NaClO
stream cell 12. Another process that could distort
the overall rate of atomic iodine formation is the
reaction
4
2
ClO
2
þ 2HI ! I
2
þ products
The rate constant for reaction (17) was not found
ð17Þ
2
(50% w/w) was used first for
ClO2 production. The column was filled with
3
1
0 dm of powdered NaClO through which was
2
in the literature and was therefore estimated

ꢀ
152
O. Spalek et al. / Chemical Physics 282 (2002) 147–157
experimentally. Chlorine dioxide ð200–250 lmol
À1
s Þ mixed with nitrogen was introduced into the
À1
reactor 1 as the primary gas, 90–340 lmol HI s
was admixed through the third injector. No atomic
iodine was detected in the ISD cell 11, but molec-
ular iodine was measured in both I cells. A rate
2
À31
6
À2 À1
constant of k17 ffi 3:8 Â 10
was determined. It was found by modelling of the
reaction system ClO –NO–HI that reaction (17)
does not increase the I concentration in the cell 12
if the HI flow rate is lower than the rate of chlorine
atom formation, i.e., at HI=ClO
6 1, and simul-
taneously at NO=ClO > 1. These conditions de-
cm molecule
s
2
2
2
2
termine when the overall rate of atomic iodine
formation may be evaluated from the I flow rate
measured in the cell 12.
2
4.3. Generation of atomic iodine in nitrogen envi-
ronment
Fig. 2. Overall rate of atomic iodine production (ꢀ), and the
rate estimated by the ISD (ꢁ) and its dependence on HI flow
À1
À1
, sec-
rate; primary flow: 220 lmol ClO
2
s
þ 2 mmol N
2
s
First, atomic iodine generation was studied at
À1
ond injector: 375 lmol NO s , third injector: HI; total pres-
sure 2–2.9 kPa; distance between injectors: d2–3 ¼ 3:7 mm, and
between third injector and the ISD cell: d3–ISD ¼ 11 mm.
different flow rates of reactants. An example of the
effect of HI flow rate is shown in Fig. 2. It can be
seen that the overall production of atomic iodine
(
detected by VIS photometry) was rather high and
increased with increasing HI flow rate. The atomic
iodine yield, however, declined from nearly 100%
to 80% with increasing HI flow rate. The produc-
tion rate calculated from the ISD method was
significantly lower. This was caused by a drop in
the atomic iodine concentration in the flow direc-
tion downstream of the reactor. This was con-
firmed by the results of our modelling studies
calculated for the same experimental conditions
(
see Fig. 3). The calculated curves demonstrate that
the average flow rate of atomic iodine through the
ISD cell is substantially lower than the overall rate
of I formation, which is approximately equal to
twice the value of the I flow rate through the first
2
I2 cell. The flow rate of atomic iodine measured by
the ISD increases with HI flow rate to some limit
only (see Fig. 2). This can be explained by the
higher rate of I atom formation and recombination
at the higher HI flow rates. This results in a greater
concentration gradient of atomic iodine down-
stream of the HI injection point, and a smaller
fraction of the I flow rate being detected in the ISD
cell compared to the overall production rate.
Fig. 3. Calculated temporal behaviour of I and I
À1
2
flow rates
downstream of the third injector; 220 lmol HI s , gas velocity
À1
85 m s , and other conditions as in Fig. 2.

ꢀ
O. Spalek et al. / Chemical Physics 282 (2002) 147–157
153
Table 1
Effect of the position of NO injection on atomic iodine production measured by ISD ðnI Þ, and VIS photometry ðn
ISD
VIS
I
Þ
VIS
ISD
À1
n
NO (first injector)
n
NO (second injector)
n
HI (third injector)
nI (lmol s
)
nI
À1
À1
À1
À1
(lmol s )
(
lmol s
)
(lmol s
)
(lmol s
)
3
20
00
0
0
200
420
100
100
100
0
8
18
46
75
2
20
À1
À1, distance between injectors: d1–2 ¼ 30 mm, d2–3 ¼ 3:7 mm, and third injector–
Primary flow: 220 lmol ClO
ISD cell: d3–ISD ¼ 20 mm.
2
s
þ 2 mmol N
2
s
In the next experiments, the effect of the way in
which NO was injected into the primary flow on
atomic iodine production was studied. Nitrogen
sured yield of I atom formation (see Table 1) was
higher than the above value, which may be ex-
plained by a limited gas mixing rate that extends
the concentration profiles.
2
oxide was injected into the primary ClO flow ei-
ther through the first or the second injector, or
both injectors simultaneously. The results are
summarised in Table 1.
The rate of I atom formation was much higher
when the NO was injected into the primary flow
sequentially through the first and second injectors.
In this way, ClO radicals are formed mostly
downstream of the first injector, and Cl atoms
downstream of the second injector (Fig. 5). In this
case, the overall process proceeds predominantly
by the non-chain mechanism, i.e., without partic-
ipation of the chain-branching reaction (8). It was
The lowest production of atomic iodine was
measured when NO was introduced through the
first injector only. This is because a high fraction
of chlorine atoms recombined before the HI was
injected. This was confirmed by the calculated
decrease in Cl atom concentration for the experi-
mental conditions. Fig. 4 shows that the calculated
yield of Cl atoms was only 5% at a distance of 33.7
mm between the NO and HI injectors. The mea-
Fig. 5. Time history of reactants and reaction products and I
À1
atom production calculated for 220 lmol ClO
2
s
þ
À1
À1
2
mmol N
tor), 200 lmol NO s
(third injector); total pressure 2.7 kPa; d1–2 ¼ 30 mm, and
2–3 ¼ 3:7 mm; gas velocity between first and third injector:
140 m s , downstream of third injector: 85 m s
2
s
in primary flow, 200 lmol NO s (first injec-
À1
À1
(second injector), 100 lmol HI s
Fig. 4. Time history of reactants and reaction products
À1
downstream of NO injector calculated for 220 lmol ClO
2
À1
s
þ
À1
2
mmol N
2
s
velocity 140 m s , and total pressure 2.3 kPa.
in primary flow, and 420 lmol NO s ; gas
d
À1
À1
À1
.

ꢀ
154
O. Spalek et al. / Chemical Physics 282 (2002) 147–157
Table 2
Effect of distance between injectors on atomic iodine flow rate, measured by ISD ðnI Þ, and VIS photometry ðn
ISD
VIS
I
Þ
ISD
À1
VIS
À1
)
d1–2 (mm)
d
2–3 (mm)
d3–cell 11 (mm)
nI (lmol s
)
nI (lmol s
9
23
13
3.7
3.7
20
20
20
10
0.5
3
15
35
54
55
À1
1
9
7
7
3
6
10
3
À1
À1
À1
,
Primary flow: 240 lmol ClO
00 lmol HI s , total N
2
s
7:2 mmol s
,
first injector: 220 lmol NO s
second injector: 220 lmol NO s
,
third injector:
À1
1
2
.
further found experimentally that the rate of I
production was not influenced by the distance
between the NO injectors (within 8–21 mm). This
can be explained by the relatively high stability of
ClO radicals downstream of the first NO injector
(
see Fig. 5).
The highest yield of atomic iodine (see Table 1)
was attained when all the NO was introduced
through the second injector and HI was injected
downstream (3.7 mm). This distance, between the
injectors, is close to the optimum corresponding to
the calculated maximum concentration of Cl at-
oms (see Fig. 4).
In other experiments, the NO flow was split
between two injectors. The effect of the variable
distance between the two injectors is presented in
Table 2. These data show the positive effect of a
shorter distance between the second NO injector
and the HI injector on I atom production in spite
of a simultaneously longer distance between the
first and second NO injector. This is due to a lower
recombination loss rate of Cl atoms upstream of
the HI injector. A higher concentration of I atoms
in the ISD cell was measured with a shorter dis-
tance between the HI injector and the cell, even
though the overall rate of I atom formation re-
mained unchanged. This was caused by a lower
loss of iodine atoms by recombination reactions
upstream of the ISD cell.
Fig. 6. Dependence of the overall flow rate of atomic iodine
À1
(
measured in cell 12) on HI flow rate at 220 lmol ClO
2
s
, and
À1
À1
various NO flow rate: 205 lmol s (H), 330 lmol s (ꢀ),
À1
À1
4
Fig. 2.
40 lmol s (ꢁ), 560 lmol s (M); for other conditions see
nism of Cl atom formation for the different ex-
perimental conditions. At NO:ClO P 2, chlorine
2
atoms are formed very rapidly by the chain-
branching mechanism (6)–(8), and their fast re-
combination occurs before HI admixing. This
The generation of iodine atoms is substantially
affected by the initial ratio of the reactants. In the
experiments, where all of the NO was introduced
through the second injector, the overall yield of
atomic iodine was nearly 100% (related to HI) at
takes place in the regions with an excess of ClO
and NO, and insufficient HI concentration. In the
2
case of the smaller ratio NO:ClO ffi 1, more stable
2
ClO radicals are formed being in equilibrium with
Cl atoms and ClOO radicals according to
NO:ClO
lower yield (70% and 45%, respectively) was ob-
tained with NO:ClO ratios equal to 2 and 2.5.
This effect can be explained by a different mecha-
2
ffi 1 or 1.5 (see Fig. 6). A substantially
ClO þ ClO $ Cl þ ClOO
The equilibrium is strongly shifted to the left
ð18Þ
2
(k17 ¼ 2:8 Â 10À14 cm molecule
3
À1 À1
s
[13], kÀ17
¼

ꢀ
O. Spalek et al. / Chemical Physics 282 (2002) 147–157
155
À11
3
À1 À1
1
:0 Â 10
cm molecule s ) [6]. In the pres-
injectors, no significant difference in the iodine
atom production between the ‘‘reverse order’’
(NO –HI–NO ) and the ‘‘normal order’’ (NO –
II
NO –HI) of injectors was found, if a short distance
(3.7 mm) between the NOII and HI injectors was
maintained. Both arrangements are suitable for the
effective generation of iodine atoms owing to a low
loss of chlorine atoms.
ence of HI, chlorine atoms are rapidly consumed
in the reaction (12), and further Cl atoms and
ClOO radicals are formed in the reaction (18). The
excessive ClOO radicals then provide Cl atoms in
the reaction [6]
I
II
I
ClOO þ N
2
! Cl þ O
2
þ N
2
ð19Þ
k18 ¼ 1:34 Â 10À14 cm molecule s :
3
À1 À1
In all experiments the gas temperature measured
in the diagnostic cell was higher than the inlet
temperature (20 °C) due to exothermic reactions. A
Our previous modelling results have shown [7] that
the formation of atomic chlorine by the chain
steep temperature rise was observed after ClO was
2
mechanism (6)–(8) prevails at NO=ClO ffi 2 and
mixed with NO. The temperature change was
2
proceeds much faster than the non-chain process
(Eqs. (6), (18), and (19)) which occurs predomi-
nantly at NO=ClO
smaller at the HI admixing. The gas temperature in
the diagnostic cell was between 80 and 120 °C.
A few preliminary experiments were performed
with chemically generated atomic iodine in the
presence of a singlet oxygen flow. Gas exiting the
singlet oxygen generator (not diluted with any
buffer gas) was introduced into the iodine reactor
as the primary flow. In the first experiments, ClO2
diluted with nitrogen was introduced through the
first injector, NO through the second injector, and
HI through the third injector. Nitrogen was in-
troduced into the reactor only in mixtures with
2
ffi 1. In a real system with a
limited rate of Cl and HI mixing, the recombina-
tion loss of Cl atoms is higher in the first case,
which is characterised by a high local rate of Cl
atom production.
In some experiments, the order of NO and HI
injection into the primary ClO flow was reversed.
2
A more uniform HI concentration in the region of
NO admixing was expected and consequently, re-
gions with fast generation of Cl atoms and insuf-
ficient HI concentration might be minimised. This
could reduce the recombination loss of Cl atoms.
ClO , NO, and HI, respectively. In the experiment
2
with a short distance (10 mm) between the HI
On the other hand, some ClO and HI might be
2
injector and the ISD cell, the formation rate of
atomic iodine detected by both experimental
methods was nearly the same regardless of the
consumed by the direct formation of molecular
À1
iodine (Eq. (17)). When HI (10–220 lmol s ) was
1
injected into the primary flow with ClO
À1
2
O ð D Þ presence (see Table 3). It appears that the
2
g
À1
1
(
240 lmol s ) before NO, and NO ð280 lmol s Þ
O ð D Þ did not change the reaction kinetics. At a
2
g
was admixed 20 mm downstream, the rate of
atomic iodine formation estimated by both detec-
tion techniques was about the same as in the case
when NO was admixed 3.7 mm upstream the HI
injector. Also in the configuration with two NO
longer distance (33 mm) between the HI injector
and the ISD cell, a substantial fraction of the
iodine atoms in the nitrogen flow recombined
upstream of the ISD cell. This was gleaned from
the fact that the ISD signal was 4.4 times lower
Table 3
Generation of atomic iodine either in nitrogen or a mixture with singlet oxygen
tot
N2
ISD
VIS
d
3–ISD
n
(mmol s
nO2ðDÞ
(mmol s
nO2ðRÞ
(mmol s
n
NO
N
HI
nI
nI
À1
À1
À1
À1
À1
À1
À1
(lmol s )
(mm)
)
)
)
(lmol s
)
(lmol s
)
(lmol s
)
1
1
3
3
0
0
3
3
4.2
4.1
4.6
4.5
–
–
166
166
159
159
145
135
155
145
22
20
5
70
70
70
70
0.39
–
1.9
–
0.39
1.9
24
1
3
À
À1
À1, second injector: NO þ N
Primary flow: no or O
injector: HI þ N
2
ð D
g
Þ þ O
2
ð R Þ, first injector: 130 lmol ClO
2
s
þ 1:25 mmol N
2
s
2
, third
g
2
, d2–3 ¼ 3:7 mm.

ꢀ
156
O. Spalek et al. / Chemical Physics 282 (2002) 147–157
than in the previous case. In the presence of sin-
glet oxygen, this signal increased 4.8 times due to
1
I
of iodine was the same in all cases. The observed
2
dissociation by O
2
ð D
g
Þ. The overall flow rate
production of atomic iodine in the ground state
1
even in the presence of O
1
2
ð D
g
Þ was explained by
a low ½O ð D Þꢂ=½total O ꢂ ratio in the reactor
2
g
2
(about 15–20%) which was near the threshold
value necessary for population inversion in
atomic iodine. This low ratio was caused by a
1
high O
2
ð D
g
Þ quenching loss upstream of the io-
dine reactor due to a high pressure (3–4.5 kPa)
and the absence of gas dilution. The reason for
this high pressure was the small cross-section of
the reactor (0:59 cm ), and the high flow rates of
secondary gases.
2
1
A substantial reduction in O
2
ð D
g
Þ loss was
mixture
Þ flow close to the generator exit,
instead of into the reactor through the first injec-
Fig. 7. Absorption/emission (gain) curve for atomic iodine re-
À1
achieved by introducing the ClO
2
and N
2
1
corded for primary flow of either 30 lmol ClO
2
s
þ
into the O
2
ð D
g
À1
À1
2
:5 mmol N
À1
2
s
(curve 2), or 30 lmol ClO
2
s
þ 2:5 mmol
1
À1
3
À
À1
N
2
s
þ 0:9 mmol O
2
ð D
g
Þ s þ1:4 mmol O
2
ð R Þ s (curve
g
1
À1
À1
tor. The volume between the O
2
ð D
g
Þ generator
1); 225 lmol NO s (second injector), 30 lmol HI s (third
injector); d2–3 ¼ 3:7 mm, d3–ISD ¼ 11 mm; total pressure 1.8 kPa.
and the iodine reactor was reduced, and the flow
rates of the other reactants were reduced to mini-
mise the pressure loss inside the iodine reactor.
1
These modifications helped to increase the O
2
ð D
g
Þ
5. Conclusions
yield to 40–45% at the iodine reactor entrance.
Under these conditions, gain on the 3–4 hyperfine
transition in the iodine atom was observed (see
curve 1 in Fig. 7). The curve 2 in Fig. 7 represents
the absorption curve recorded with similar flow
A new method for the chemical generation of
atomic iodine in COIL devices was studied ex-
perimentally. It is based on the chemical produc-
tion of chlorine atoms and their subsequent
reaction with gaseous hydrogen iodide. Effects of
the initial ratio of reactants and the way in which
mixing in a flow of nitrogen occurs were studied
and interpreted by means of a kinetic model for
the reaction system. Concentrations of atomic io-
dine were determined by two independent tech-
niques: the iodine scan diagnostics, and the
absorption photometry for molecular iodine
(formed by recombination of iodine atoms). The
yield of atomic iodine generated in the nitrogen
environment was rather high (70–100%), with op-
timum conditions. In preliminary experiments
performed in the presence of singlet oxygen, gain
on the 3–4 transition in iodine atom was measured.
A detailed investigation of this system is under
way. Finally, the chemical generation of atomic
iodine is planned directly for implementation on a
supersonic COIL.
1
rates of reactants but with no O ð D Þ in the pri-
2
g
mary flow.
In the experiments with atomic iodine in ni-
trogen the gas temperature in the ISD cell was
8
1 Æ 13 °C. The temperature was calculated from
the Gaussian full width after the deconvolution of
the Voigt profile of the absorption–frequency
curves [11]. The temperature measured by a ther-
mocouple in the I cell 12 was 58 °C. In a flow of
2
1
O ð D Þ, the temperature in the ISD cell was
2
g
1
thermocouple in the I
17 Æ 22 °C, and the temperature measured by the
2
cell was 180 °C. The tem-
perature increase in the flow direction observed in
this case may be ascribed primarily to the exo-
Ã
thermic quenching of I atoms by water molecules.
Further processes that can lead to higher temper-
1
1
atures in the O ð D Þ experiments are the O ð D Þ
2
g
2
g
self-quenching and pooling reactions.

ꢀ
O. Spalek et al. / Chemical Physics 282 (2002) 147–157
157
ꢀ
[4] O. Spalek, V. Jir
asek, J. Kodymov ꢁa , I. Jakubec, M.
ꢁ
Acknowledgements
ꢀ
Censk ꢁy , Proc. SPIE 4184 (2000) 111.
ꢀ
[
[
[
[
[
5] V. Jir ꢁa sek, O. Spalek, J. Kodymov ꢁa , Proc. SPIE 4184
2000) 103.
The USAF European Office for Research and
Development (EOARD) and the Grant Agency of
the Czech Republic supported this work.
(
6] J. Arnold, K.D. Foster, D.R. Snelling, R.D. Suart, IEEE J.
Quant. Electron. QE-14 (1978) 293.
ꢀ
ꢀ
7] V. Jir ꢁa sek, O. Spalek, J. Kodymov ꢁa , M. Censk ꢁy , Chem.
Phys. 269 (2000) 167.
8] D.M. Maylotte, J.C. Polanyi, K.B. Woodall, J. Chem.
Phys. 57 (1972) 1547.
9] G. Cowley, Pred. Bull. 113 (1993) 1.
References
[
1] M. Endo, D. Sugimoto, H. Okamoto, K. Nanri, T.
Uchiyama, S. Takeda, T. Fujioka, Jpn. J. Appl. Phys. 39
[
[
10] R.D. Suart, D.R. Snelling, K.D. Foster, R. Lambert,
DREV Rep. 4005/76, April 1976.
(2000) 468.
11] R.F. Tate, B.S. Hunt, C.A. Helms, K.A. Truesdell, G.D.
Hager, IEEE J. Quant. Electron. 31 (1995) 1632.
12] J. Kodymov ꢁa , O. Spalek, Jpn. J. Appl. Phys. 37 (1998) 117.
13] M.A.A. Clyne, I.F. White, Trans. Faraday Soc. 57 (1971)
[
[
2] M. Endo, M. Kawakami, S. Takeda, F. Wani, T. Fujioka,
Proc. SPIE 3612 (1999) 56.
3] V.S. Pazyuk, N.P. Vagin, N.N. Yuryshev, Proc. SPIE 2767
ꢀ
[
[
(1995) 206.
2
068.