Citation
Effects of secondary reactions on the measurements of the rate of OH radical reaction with n-butane

Material Information

Title:
Effects of secondary reactions on the measurements of the rate of OH radical reaction with n-butane
Creator:
Bonilla, Juan Luis
Publication Date:
Language:
English
Physical Description:
viii, 140 leaves : ; 28 cm

Subjects

Subjects / Keywords:
Chemical kinetics ( lcsh )
Chemical kinetics ( fast )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaves 101-102).
General Note:
Submitted in partial fulfillment of the requirements for the degree, Master of Science, Department of Chemistry.
Statement of Responsibility:
by Juan Luis Bonilla.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
19967758 ( OCLC )
ocm19967758
Classification:
LD1190.L46 1987m .B654 ( lcc )

Downloads

This item has the following downloads:


Full Text
EFFECTS OF SECONDARY REACTIONS ON THE MEASUREMENTS
OF THE RATE OF OH' RADICAL REACTION WITH N-BUTANE
by
Juan Luis Bonilla
B.S., University of Puerto Rico, 1976
A thesis submitted to the
Faculty of the Graduate School of the
University of Colorado at Denver in partial fulfillment
of the requirements for the degree of
l
Master of Science
Department of Chemistry
1987


This thesis for the Master of Science degree by
Juan Luis Bonilla
has been approved for the
Department of
Chemistry
by
Robert Damrauer
Sandra Eaton
Date
-.r-'i',-
^7"


Bonilla, Juan Luis (M.S., Chemistry)
Effects of Secondary Reactions on the Measurement of the Rate
of HO Radical Reaction with n-Butane
Thesis directed by Professor Larry G. Anderson
Hydroxyl radicals are considered to play an important
role in combustion and atmospheric chemistry. They are known
to be one of the most reactive species in the atmosphere. The
reactions of hydroxyl radicals with hydrocarbons in the tropo-
sphere are believed to be the key initiating step of the
oxidation of hydrocarbon compounds. There are considerable
amounts of small chain hydrocarbons in the troposphere. The
reaction of HO + n-butane has been used as a representative
alkane in models o*' photochemical air pollution. This reaction
has also been used as the reference reaction in the study of
rate constants for other reactions using relative methods.
The first step in the reaction of HO radicals with n-
butane involve the abstraction of a hydrogen atom from n-
butane. In a system of reactions in which HO radicals are
present, the reaction of these HO radicals and the product
butyl of the main reaction can be enhanced as the initial
concentration of HO is increased. This secondary process is of
some significance in the measurement of the main reaction and
should be taken into consideration. The study of the reaction
of n-butane and HO radicals showed that there is a significant
influence of the secondary reaction to account for differences
of up to 30$ in the measurement of the blmolecular rate con-
stant.


iv
Experiments performed at initial HO concentrations of
11 3
approximately 2x10 particles/cm resulted in an apparent
-12 3
bimolecular rate constant of 2.23.0x 10 cm /molec-sec,
12
whereas initial HO concentrations of 2 x 10. and 8 to 16 x
12 3
10. particles/cm gave apparent bimolecular rate constants of
-12 -12 3
2.85.05 x 10 and 3.07.15_3.13.23 x 10 cm /molec-sec
respectively. Modeling of the experimental conditions for the
results obtained helped to elucidate the approximate magnitude
of the rate constant for the secondary process in the HO + n-
butane system, which ranges from 1.2 X 10 10 to 5 x 10 10
3
cm /particle^sec. The best value for the reaction of HO + n-
-12
butane at room temperature is 2.1.25 x 10 .
cm^/particle-sec. The simulations also demonstrated that the
bimolecular rates at two different HO concentrations are going
to be more nearly the same when the decay of HO is followed by
a variable time method, which is the method most commonly used
with flash photolysis systems.
Trapping and analysis of products in the flow system
for the reaction of HO and n-butane showed the formation of a
product which contains butyl as its main component. Kinetic
data demonstrated that the butyl was reacting with HO to form
the product. A clear analysis of the product was not possible
due to the interference of NO or N02 in it. This interference
was shown in the kinetic results and in the mass spectrum of
the product.


CONTENTS
CHAPTER
1. INTRODUCTION............................................ 1
General considerations............................... 1
Calculation of a bimolecular reaction rate........... ^
Experimental techniques.............................. 6
Previous n-butane and hydroxyl reaction
rate results......................................... 9
Purpose of this research............................ 16
2. EXPERIMENTAL........................................... 19
Flow tube components................................ 19
Reactants........................................... 2^
Detection of HO radicals............................ 27
Experimental procedure.............................. 29
3. COMPUTERIZED SYSTEM FOR STUDYING GAS PHASE KINETICS. 37
Advantages.......................................... 37
Hardware interfacing................................ 39
Software description................................ ^9
4. RESULTS................................................ 58
Experimental........................................ 58
Error Analysis...................................... 72
Modeling............................................ 76
Discussion.......................................... 88
Conclusions
98


vi
BIBLIOGRAPHY................................................ 101
APPENDIX
1. ACTUATE PROGRAM...................................... 103
2. KINETIC RESULTS FOR HO AND n-BUTANE REACTIONS......... 137
3. REACTIONS OF SYSTEM TO BE SIMULATED................... 1 UO


vii
TABLES
Table
1. Previous studies of HO + n-butane at room
temperature (295-302 K).................................. 10
2. Transducers used in gas flow kinetic experiments.... 39
3. Signal conditioner devices............................... 41
H. Summary of input/output modes............................ ^6
5. Features of the Lab Master board......................... ^7
6. Progamming of the Lab Master board....................... 50
7. Simulations with minimum k2.............................. 79
8. Effect of kj and k2 on model values for
apparent rates........................................ 80
9. Maximum k2 value and effect of k2 on
apparent bimolecular rate................................ 81
10. Flash Photolysis simulations............................. 87
11. Previous direct studies of HO and propane................ 97


viii
FIGURES
Figure
1. Schematic of discharge flow-resonance
fluorescence apparatus.................................. 20
2. Detection area's six way fitting....................... 22
3. Optical set up......................................... 28
4. High current relay driver connections.................. 43
5. Control of multiport valve............................. 45
6. Program flow......................................... 52
7. Titration curve....................................... 59
8. Pseudo-first-order decay.............................. 61
9. Relative signal decay of HO for High and Low
HO concentrations....................................... 64
10. Kinetic results for Low, High and Very High
HO concentrations....................................... 66
11.1 Gas chromatogram.................................... 68
11.2 Mass spectrum.......................................... 69
12. Modeling and experimental results...................... 83
13.1 Simulated concentration profile for Low HO........... 84
13.2 Simulated concentration profile for High HO...... 85


CHAPTER 1
INTRODUCTION
General Considerations
Science is considered to be the state of knowledge
derived from observation, study and experimentation carried on
in order to determine the nature or principles of the subject
being studied. The Scientist in all of us has the curiosity or
drive to carry on the task and to obtain such knowledge in
order to satisfy the very nature of its existence. Chemistry,
as a science that studies composition and relations of the
most simple forms of existent bodies, strives for the absolute
values of those principles that affect its own nature.
Within the science of the chemical nature of things,
kinetics is the study of the rates of the chemical reactions
including all the factors that influence them and the
mechanisms involved. Kinetics is concerned with the details of
the process in which a system gets from one state to another
and the time required for that transformation. The details
include concentrations of initial reactants, product forma-
tion, temperature and pressure effects on rates of the
processes studied.
The kinetics of the reactions of hydroxyl radicals
(HO*) with different species have been largely studied in the
gas phase. Its interest results from the fact that this radi-


2
cal is present in considerable amounts in our atmosphere and
in combustion systems. It is very reactive since it has an odd
number of electrons. The reactivity and quantity of this
radical make it one of the most important species in our
atmosphere and in combustion systems.
As modern civilization continues to emit hydrocarbon
pollutants into the atmosphere and use combustion systems for
different purposes, it becomes more and more important to
characterize the kinetics of HO radicals with hydrocarbons.
Hydrocarbons are involved in chain reactions that oxidize NO
to NO^ in the atmosphere. These reactions are initiated by the
reactions between HO radicals and the hydrocarbons.^ Con-
sidering the concentrations of hydrocarbons in the atmosphere,
it is estimated that alkanes can account for up to 33? in
(2) (33)
non-rural and 20-25? in rural areas of the HO reactions
with hydrocarbons. The other reactions of HO and hydrocarbons
involve the reactions with olefins, and aromatics.
The principal mechanism of fuel consumption in lean
and stoichiometric alkane/air flames is considered to be the
(3)
abstraction of hydrogen atoms from alkanes by HO radicals.
This reaction can be described by:
(1) R-H + HO- -* R- + H20
in which the hydroxyl radical abstracts a hydrogen atom from


3
the alkane reactant forming an alkyl radical and water. The
sequence of reactions that could follow from this main reac-
tion are numerous making the study of the rate of this main
reaction suceptible to undesirable effects if not properly
considered. The undesirable effects could result from a com-
petitive reaction between the very reactive HO radical and
reactive products that are formed from the main reaction. If
these secondary reactions are fast, they will tend to affect
the measurement of the first reaction. The reaction of HO
radicals with alkyl radicals can be of great importance in the
study of the first reaction if a sufficient concentration of
alkyl radicals are formed in the system of study.
(2) R* + HO* -* Products
Due to the importance of the reactions of HO with
alkanes it is necessary that highly accurate measurements be
done for the rates of these reactions in order for these
results to be used in modeling and estimates of complex sys-
tems like the atmosphere or combustion. The rate of the secon-
dary reaction between alkyl radicals and HO radicals have
never been studied. On the other hand the rate of reaction
between alkanes and HO radicals have been studied by numerous
researchers using different experimental techniques. The
possible effect of the secondary reaction on the measurements


of the reaction of alkanes and HO radicals obtained by these
researchers was considered to be negligible. Some secondary
effects for alkanes and HO radical reactions have been ob-
served at extreme experimental conditions used by few of these
researchers^^12)(1^ ) ^ ^ addition to an accurate knowledge
of the rate of the primary reaction, it is necessary to
measure the rate of the reaction of alkyl and HO radicals.
Calculation of a Bimolecular Reaction Rate
The study of bimolecular reactions is generally done
under pseudo-first-order conditions in which one of the reac-
tants is in excess. This will ensure that the reactant in
lesser amount will react with the one that is at higher con-
centrations and not with itself. Also the higher concentration
reactant will stay relatively unchanged during the reaction.
The kinetics of this system is followed by measuring the
change in concentration of the reactant in smaller quantity
with reaction time. The rate of a bimolecular reaction can be
expressed as;
Rate of Reaction = k [A][B]
where k is the reaction rate constant of species A reacting
with B and [a] is the concentration of species A and [b] is
the concentration of species B. The rate of the reaction can


5
be considered to be the change of concentration of one of the
reactants in time. Considering the change of concentration of
species A in time we can write a differential equation in
which:
-d[A]/dt = k [a][b]
and if species B is at higher concentration than species A, we
can consider the concentration of species B to be constant
during the reaction. The concentration of species B should be
at least 40 times that of species A for pseudo-first-order
(4)
conditions to exist. If this is the case, then the con-
centration of species B can be included as part of the rate
constant, making this a pseudo-first-order reaction for A with
a pseudo-first-order rate constant of k1. Rearrangement and
integration of the previous equation,
/dA/[A]=k1/dt
results in:
ln[A]=klt + c
which is a linear equation. Measurements of the variable t
(time) and the logarithm of the concentration of A will enable


\
6
us to calculate the pseudo-first-order rate constant for the
reaction assuming a constant concentration of B. The ratio of
k*/[B] will give the value for the bimolecular rate constant
for the reaction of A + B, or the slope of a plot of k1 vs [b]
will also give us the same results. The second method is
preferred because it will correct for loss of reactants to the
walls as well as losses due to other reactions.
If reactant A is also reacting with the walls or
itself at a significant rate, then the observed pseudo-first-
rate constants will be higher than that for the reaction of A
with B. This higher pseudo-first-rate constant will be inde-
pendent of the amount of species B that is used. Its higher
value will result in a higher calculated bimolecular rate
constant when the ratio of k'/[B] is used. If this higher
pseudo-first-rate constant is independent of the amount of
species B, then it will appear as part of all the values of
the .pseudo-first-order rate constant at the different con-
centrations of B. Therefore the slope of the pseudo-first-
order rate constant versus the concentration of B will not
include the influence of these losses of species A.
Experimental Techniques
In the study of the kinetics of the reactions of HO +
alkanes, the hydroxyl radical is usually selected as the
reactant to be monitored, since it can be easily measured by


7
fluorescence in the 308 nm wavelength region.
The study of the kinetics of HO radicals with alkanes
can be separated into absolute or relative techniques. The
absolute techniques can be separated into discharge flow (DF)
or flash photolysis (FP) methods. Also included in these
absolute techniques are pulse radiolysis and modulation-phase
shift, which have been infrequently used for small chain
alkane studies. The discharge flow method will be discussed in
detail in the experimental section. In general it can be
considered as a steady state method in which the concentration
of the species remains constant in the detection area. The
reaction time of the species is changed by varying the loca-
tion of addition of one of the reactants in the flow tube
system. The reactive HO* species are produced by passing the
appropriate precursors through an electrical discharge prior
to entering the main flow area.
The flash-photolysis method involves the use of a
reaction cell in which the reactants are added to the precur-
sors (H^O, ^2^2 or ot*ier mixtures) of HO radicals. HO
radicals are formed by the photolysis of these precursors. The
radiation for the photolysis is obtained by the use of flash
lamps or pulsed lasers, which are separated from the reaction
cell by the use of transmitting windows (LiF, MgF^, CaF^,
sapphire and Suprasil) in the vacuum ultraviolet region. The
loss of HO radicals can be monitored as a function of time


8
after photolysis of the precursors by the use of one of the
following detection methods.
The detection of HO radicals can be accomplished by
resonance absorption (RA), resonance fluorescence (RF),
electron paramagnetic resonance (EPR), mass spectroscopy (MS),
laser magnetic resonance (LMR), or laser induced fluorescence
(LIF). The concentration of HO radicals used varies depending
on the detection method used. For resonance fluorescence,
laser-induced flourescence and laser magnetic resonance the
9 12 3
concentrations are typically low, 10 -10. molecule/cm .
Resonance absorption is usually used for higher concentration
12 1JJ 3
detection ranges, 10. -10. molecules/cm Electron paramag-
netic resonance detects ranges from lo]] to 10^ molecules/cm^
12 13 3
and mass spectrometry from 10. to 10. molecules/cm .
Relative measurement techniques are accomplished by
monitoring the relative rate of disappearance of two or more
organic compounds in the presence of HO radicals in a chemical
system. Using a known rate constant for the reaction of HO and
one of the organic compounds the absolute rate constant for
the reaction of HO and the other compound can be calculated.
Relative measurements are made in Teflon chambers coupled to
Gas Chromatograph, FT-IR absorption spectrometer or differen-
tial optical absorption spectrometer for detection of the
organic reactants.


9
Previous n-butane and Hydroxyl Reaction Rate Results
In the study of hydroxyl radical and short chain
alkanes reactions, the scatter of the room temperature data is
varied. The biggest scatter occurs for n-butane and HO in
which there is a factor of almost 1.9 in the reported data
(5)
(See table 1.) by different researchers These dis-
crepancies in the evaluation of the rate constants suggest the
possible complication of undesirable effects occurring during
the experimental process. Such undesirable effects include the
effects of secondary reactions between HO radicals and butyl
radicals. The discrepancies may also be due to the use of
erroneous calibration factors and other systematic experi-
mental errors.


10
Table 1. Previous studies of HO + n-butane
at room temperature (295~302 K).
xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx#
(10 icnT>/molec-s) Method Reference
2:570.12 Flash Photolysis- Kinetic Spectroscopy Greiner 1970
4.1 Discharge Flow- Mass Spectroscopy Morris & Niki 1971
2.350.35 Flash Photolysis- Resonance Fluorescence Stuhl 1973
4.22+0.17 Pulse Radiolysis- Gordon & Mulac
Resonance Absorption 1975
2.720.27 Flash Photolysis- Resonance Fluorescence Perry, Atkinson & Pitts 1976
2.670.22 Flash Photolysis- Resonance Absorption Paraskevopoulos & Nip 1980
2.230.23 Discharge Flow- Resonance Fluorescence Anderson & Stephens 1982
2.41+0.10 Laser Photolysis- Tully & Droege
Laser Induced Fluorescence 1985
2.300.3 Laser Photolysis- Laser Induced Fluorescence Schmidt, Zhu, Becker & Fink 1985
Relative technique Reference reaction
2.90.7 HO + CO Gorse & Volman 1974
2.370.14 HO + CO Campbell et al. 1975
2.520.25 HO + propene Atkinson et al. 1981
2.500.24 HO + n-pentane Behnke et al. 1984
2.710.32 HO + propene Atkinson & Aschmann 1984
XXXXXXXXXKXXKKXXXXXXXXXXXXXXXXXXXXXXXXXKXXXXXXKXXXXXXXXXXXXKXXXX


11
N. R. Greiner was the first to study the reaction of
HO with a series of alkanes by a direct method. The purpose of
his study was the characterization of the rates of abstraction
for primary, secondary and tertiary hydrogen atoms found in
hydrocarbons, and to enable further prediction of reaction
rates for more complex alkanes. In this study, the effect of a
side reaction for alkanes larger than CH^ was observed. The
rates show a 15? decrease as the alkane concentration is
(6)
increased by tenfold at a given temperature This behavior
results from the competitive reaction between HO and alkyl
radicals at an assumed rate of 1.66 x 10 ^ cm^molec ^sec 1.
The data were corrected taking into consideration this effect
by the use of a computer simulation that included the main
reaction, the proposed competitive reaction, an assumed ini-
13 3
tial HO concentration of 3.0 x 10 molec./cm produced in
each flash, and the experimental concentration of the alkane
used. The uncorrected experimental results for n-butane at
room temperature of 298 K and a concentration of the alkane of
15 3
4.35 x 10 molecules/cm resulted in a rate constant of 3.12
-12 3
x 10 cm /molec.-sec. This value was corrected by a factor
of 22$ using the simulation to obtain the rate of the reaction
-12 3
of HO and n-butane of 2.57 x 10 cm /molec-sec.
The experimental results of Morris and Niki are ap-
proximately 40$ higher than those of other researchers. Their
experimental results were obtained using a high initial HO


12
13 3
concentration (1 x 10. molecules/cm ) relative to the alkanes
11 3 (7)
(5 x 10 molecules/cm ) studied Using a discharge flow
system coupled to a time-of-flight mass spectrometer, they
observed reactant disappearance for the reactions of olefins
and other compounds (including n-butane) with HO. These reac-
tions were studied under pseudo-first-order conditions (excess
HO) and the hydrocarbons were the species monitored. They
considered the effect of secondary reactions to be insig-
nificant due to the large excess of HO radicals present,
making it unlikely for any intermediate like the alkyl radical
to compete with the reactions of HO and hydrocarbons. They
estimated an error of less than 5? for the effects of the
reaction of oxygen atoms (which are products from the reaction
of HO + HO) with the hydrocarbons studied. They also con-
sidered the reactions of hydrocarbons with H atoms, which are
used in the formation of their HO radicals, to be too slow to
have an effect on their results.
Stuhl working with alkane concentration ranges of 0.5
11| 3
to 2.5 x 10 molecules/cm monitored HO concentration by
resonance fluorescence. HO was made by flash photolysis of a
/ Q \
H20 (0.25$ in He) mixture at a total pressure of 2.67 kPav
The initial HO concentration used was approximately 3 x 1011
3
molecules/cm The effect of secondary reactions in this study
was considered to be unlikely. Experiments performed at two
different nrbutane concentrations and at an initial HO con-


13
12
centration of approximately 2x10 showed no effect on the
pseudo-first-order rate constant results as compared to ex-
periments done at an HO concentration of 3 x 1011.
Gordon and Mulac did pulse radiolysis studies of H20
(1.39100%)/Ar mixtures at total pressures of 96.0-101.3 kPa
in the presence of n-butane at 66.7-328 Pa pressure. The
hydroxyl radical was monitored by U.V. absorption spectroscopy
at 308.7 nm. The run at 298 K had a diluted mixture of H20 and
(9)
Ar It is very likely that possible effects of secondary
reactions could have occured due to the relative high con-
centration of HO needed to be monitored by absorption
spectroscopy.
Perry, Atkinson and Pitts worked in an n-butane con-
13 1H 3
centration range of 1 x 10 to 1J x 10. molecules/cm ,
using a flash photolysis-resonance fluorescence method. In
their study^10^ there is no indication of the HO concentration
used. Because pseudo-first-order rate constants for the HO
radicals were determined, it is assumed that n-butane con-
centrations were at least forty times bigger than HO. They
concluded that secondary reactions were negligible because a
variation of a factor of two in the flash energy showed no
effect in the rate constants. The variation on the flash
energy should have doubled the photolysis of water which
produced the initial HO concentration. HO radical concentra-
tion was monitored over at least three half-lives.


The study of Paraskevopoulos and Nip^V, using a
flash photolysis-resonance absorption technique, also indi-
cates that there was no significant interference from secon-
dary reactions. They varied the concentration of HO radicals
by a factor of 5 to 10 by varying both the pressure of H^O by
a factor of 10, and the flash energy (30 to 350 Joules) needed
to photolyze the H20. These variations did not change the
measured rates within the experimental errors. The concentra-
15 3
tion of n-butane varied from 0.6 to 9 x 10 molecules/cm and
13 3
HO was estimated at 0.6 to 5 x 10 molecules/cm .
Anderson and Stephens used a discharge flow-resonance
(12)
fluorescence system in which the HO concentration of 1-3 x
11 3
10 molecules/cm was estimated based on titrations of H
atoms by N0£. Their n-butane concentration varied from 1.17 to
13 3
8.18 x 10. molecules/cm In some experiments done at 250 K
in which they increased the HO concentration by a factor of Jl-
5 times, they were able to observe a non-linear decay in the
HO fluorescence when compared to experiments done with low
initial HO concentrations. This curvature appeared to be less
important at higher temperatures in their experiments and was
considered to be the result of secondary reactions, mainly
between HO and butyl radicals. The rates observed for these
particular experiments in which HO initially was high were
consistent with model calculations done in a further
(13)
study There the rate constant for the secondary process


15
-10 3
was estimated at 2.0 x 10 cm /molecule-sec. All of their
experiments were done at low initial HO concentration where
the effect of secondary reactions was considered to be least.
Tully and Droege determined site-specific rate coeffi-
cients for H- and D-atom abstraction by HO from methyl and
methylene sites in n-butane using the laser photolysis-laser
(1 H)
induced fluorescence technique. Their typical reaction
-3
mixtures concentrations in molecules cm consisted of sNgO =
114 15 1H
1.3 x 10 , H2O = 3*3 x 10 , n-butanes = 0-3.8 x 10 , and
18
He = 6.4-13.2 x 10 . They varied the photolysis flux density
2
over the range of 1-4 mJ/cm which would increase their
initial HO concentrations, and observed no significant changes
in the kinetics. At 599 K they were able to observe secondary-
reaction kinetic interference for n-butane and HO reactions.
10
HO concentration was considered to be 7.5 x 10 molecules
cm ^ and time dependence of HO concentration in the reaction
with n-butane was taken over at least a factor-of-ten varia-
tion from its initial concentration.
Schmidt et al. performed experiments using excimer
laser photolysis of the precursors of HO radicals. The HO
concentration decay as a function of time was monitored using
(15)
laser induced fluorescence. The experiments were performed
at 1 atmosphere total pressure, but there is no indication of
the approximate HO or n-butane concentration used. The experi-
ments for the reaction of HO and n-butane, propane, and ethane


16
were done to check the method and apparatus. The study was
mainly directed at the study of the effects of oxygen on the
reaction of HO and C^H2 and with some sulfur containing com-
pounds.
Purpose of this Research
As we can see the effects of secondary reactions were
considered to be negligible in the previous n-butane and HO
radical reaction studies. A variation of the initial HO con~
centration was used to see if there was any effect of secon-
dary processes. There was no effect by the use of a change in
the initial HO concentration with the exception of the studies
of Greiner^, Anderson^12\ and Tully^1^ in which the change
(12)
of the initial HO concentration was done at very low or
w. (14)
high temperature or by changing the ratio of n-butane
concentration to that of HO^.
There seems to be the indication that secondary reac-
tions exist in the reaction of n-butane and HO radicals. These
could only be apparent in these previous studies if the rate
of reaction for butyl and HO radicals was significantly high
in order to show its effect. It is important to evaluate the
rate of the reaction of butyl and HO radical to be able to use
its value in complicated modeling schemes and to correct for
the influence of it in the study of the reaction of n-butane
and HO radicals. It is necessary to evaluate the conditions at


17
which the effect of secondary reactions are apparent, so the
effect of systematic errors due to these are minimum. The lack
of effect of secondary reactions found by some researchers in
their studies tend to suggest that there was not enough of a
change in the initial HO concentration in order to see the
effect. This does not necessarily means that the effect of the
secondary reactions was not present in their studies or was
r
negligible.
The purpose of this research is to investigate the
effect of the secondary reactions of HO radicals with butyl
radicals on the measurement of the rate for the reaction of HO
with n-butane. This is performed by varying the initial con-
centrations of HO radicals in the different experiments which
should give different apparent rate results for the
bimolecular reaction of n-butane and HO radicals, if secondary
reactions are present. A computer simulation and analysis of
the kinetic data will serve to place meaningful limits on the
rate of the reaction of HO radicals with butyl radicals. Using
the kinetic results obtained for the reactions of HO with n-
butane and HO with n-butyl radicals, it is intended to model
the conditions used by other researchers to see if any secon-
dary reactions could have been affecting their measurements.
An analysis of the products formed will be performed
to help elucidate the mechanism of HO radical and n-butyl
radical reactions. This will be performed by trapping products


directly from the flow system and analyzing these product
using gas chromatographic-mass spectral methods.


CHAPTER 2
EXPERIMENTAL
Flow Tube Components
The kinetics experiments were performed using a disn
charge flow-resonance fluorescence apparatus (fig 1). The flow
system was built using 1.25 in. i.d. copper pipes and fittings
through which the gases flowed at a constant rate. The pipes
lengths varied depending on the fittings needed at the dif-
ferent positions for introduction of gases, readout devices,
and detection. The pipe and fittings were soldered together
into five different sections. These permit the system to be
disassembled for maintenance purposes.


20
MCROVAVE
CAVITY
Figure 1. Schematic of discharge flow-resonance
fluorescence apparatus.
The first section consists of three gas inlets. The
first inlet is used for the introduction of the carrier or
diluent gas (Helium) at a distance of 92.4 cm from the detec-
tion area. The second inlet is used for the introduction of
hydrogen atoms created by passing helium gas that contains
small traces of hydrogen impurities, hydrogen that is absorbed
from the teflon tubing or a known mixture of hydrogen/helium
through a 11 mm i.d. glass tubing coupled to a microwave
discharge cavity prior to entering the flow area. The
microwave discharge system consist of an Opthos microwave


21
power supply and cavity. The forward power was maintained at
40-50 watts and the reflected power at 0-2 watts. Hydrogen
atoms were introduced at a distance of 82.6 cm from the detec-
tion area. The third inlet of the first section is used for
the introduction of a known mixture of N02 and helium at a
distance of 75.2 cm from the detection area. The N02 will
-10 3
rapidly (1.3 x 10 cm /molecules-sec) react with hydrogen
atoms to make HO radicals and NO.
The second section has a fitting through which a
copper constantan thermocouple can be introduced to measure
temperature. It also has a valve at a distance of ap-
proximately 52 cm. before the detection area. Through the
valve one can measure the pressure by a 10 torr range Baratron
capacitance manometer (pressure gauge). This pressure gauge
also measures pressures after the detection area through means
of another valve. This configuration allows pressure measure-
ments at two positions before or after the reaction area, thus
quantifying for pressure drops along the axis of the flow
system. The two valves were connected to the pressure gauge by
means of 1/4 in. teflon tubing.
The third section, contains eight different inlet
ports. These are made of eight 1 in. to 1/4 in. NPT Tee fit-
tings and are separated a distance of 5 cm between the cen-
ters. The first port is 45 cm from the detection area. The
eighth port is a distance of 10 cm from the detection area.


22
The inlet ports allow hydrocarbons to be introduced through
1/16 in. stainless steel tubing which is inserted to the
center of the main flow tube. The port selection is made by
using a Carle model 2026 multiport valve, to which the eight
different outlet tubes are connected. The inlet to the multi-
port valve is connected to the source of the hydrocarbon. A
Carle Valve Actuator (Model H203) changes the selection of the
outlet port and is controlled by means of an IBM personal
computer. Details of the computer interfacing and control will
be discussed in Chapter 3-
The fourth section is the detection area, and was made
with two 1 in. Tee fittings. These two fittings were modified
by making apertures in their walls and welding them to each
other in such a way as to form a six way fitting (Fig 2).
Figure 2. Detection area's six way fitting.


23
Two of the apertures enable the main flow of gases to
pass through the detection area. A quartz window was connected
with epoxy to a third aperture through which the light from
the fluorescence lamp enters the middle of the gas flow.
Included in this aperture is a 1 in. to 1 cm. reducer at
approximately 1 in. from the window and next to the main flow
tube. This 1 cm. hole reduced the scattered light from the
outside. The fourth and fifth apertures had Wood's horns
connected to them. These were also made of copper fittings
soldered together. One of these Wood's horns was in line with
the light coming from the light source. It will trap the light
after passing through the gas flow. The other was perpen-
dicular to the fluorescence excitation axis and it further
helped to trap the scattered light. Welded to the sixth aper-
ture is a 1 /4 in. thick and 6 in. diameter round copper plate
with a 1 cm. hole in the middle. This serves as the base for
the photomultiplier which sits perpendicular to the light
source entering the flow system. A groove made in the plate
allows a 1 in. diameter quartz window to be mounted on the
plate.
The fifth section of the flow system contains a valve
which allows pressure measurements immediately after the
detection area. It also contains fittings to enable the intro-
duction of another thermocouple. This section is connected to
the pump by a 1 1/4 in. i.d. gum rubber hose with an inner


24
steel spring reinforcer. A glass trap was inserted ap-
proximately halfway between this section and the pump's fit-
tings.
The walls of the flow tube were coated with a halocar-
bon wax (Series No. 15-00, Halocarbon Products Corp., Hacken-
sack, New Jersey) to reduce the wall loss of HO radicals in
the system. The five sections of the flow tube were assembled
and the interconnections were sealed using a sealing tape from
3M and a Dux Seal plaster compound from Manville Corp. placed
over the sealing tape. This proved to be an effective way to
maintain the low pressure conditions needed. Two large bellows
sealed valves were inserted after the fifth section of the
flow tube and before the pump. The glass trap is located
between these two valves. The valves were used to control
velocity of the flow.
Reactants
The helium, hydrogen/helium (8.5%), nitrogen
dioxide/helium (0.55$) and n-butane (99.5$) were used without
further purification. Gas chromatographic-mass spectometric
analyses of n-butane indicated that the alkane was within the
99.5$ purity specified by the manufacturer. Due to the nature
of the experiments performed with low HO, the nitrogen
dioxide/helium gas was diluted by a factor of almost 14 times.
These dilutions were performed using a commercial propane


25
tank, an absolute 1001 in. of water range pressure gauge and
a 30.5 psi relative to atmospheric pressure gauge. The
hydrogen source for these low HO experiments was obtained by
using helium instead of a hydrogen helium mixture. This helium
had small traces of hydrogen containing impurities which were
either part of the helium gas, collected from the 1/4 in.
plastic tubing or both. The gases were connected from the gas
cylinders to the flow controllers using 1/4 in. polyethylene
tubing with the exception of the NO^ cylinder which requires
1/4 in. teflon tubing due to the reactive nature of N02 with
the polyethylene tubing.
All gas flows were controlled and measured by Unit
Instruments (Model UFC-1000) electronic mass flow controllers.
These were calibrated using soap bubble flowmeters. The flow
controller used for n-butane was also calibrated by measuring
the time required for the pressure to increase by a few torr
when filling a container of known volume. The results for the
different volumes flowed versus the set voltages gives a slope
for the relation between these two parameters. The results of
the slope for the calibration of the flow controller using the
pressure increase technique were 12$ lower than those found
using the soap bubble method. A significant fluctuation using
the pressure increase technique was noted in the high flow
region from 4-5 volts during the calibration runs, which gave
higher flows than expected. This suggested an instability of


26
the sensor in the flow controller for n-butane at high flows.
The calibration results used in the experiments for the set-
ting of flows were for the low voltage (0-3 volts) range of
the calibration of the pressure increase technique. The fluc-
tuation was not observed during the calibration using the soap
bubble flowmeter for n-butane which could have resulted in the
higher calibration results for the soap bubble method compared
to the pressure increase method. This fluctuation in the
sensor could be the result of the low thermal conductivity and
high molecular weight of n-butane. A higher negative intercept
resulted in the soap bubble method, which suggests a loss of
n-butane possibly by reaction with the soap bubbles.
The use of a calibration method using a pressure
increase in a known volume for n-butane was performed to
evaluate the possibility of n-butane reacting with the soap of
the bubbles. The other flow controllers were calibrated for
helium using the soap bubble flowmeter, reproducible results
were obtained. The flow controllers were then connected to the
flow system using 1/H in. Teflon tubing. The flow controllers
were controlled by the IBM personal computer and power was
furnished by a power supply with +15, -15 and +5 volts lines.
These flow controllers feature a shut-off valve which stops
the flow of gas when required.


27
Detection of HO Radicals
The components for the detection of HO radicals con-
sist of a lamp or light source for excitation and a photomul-
tiplier for detection of fluorescence. The light source was
obtained by passing a mixture of H£0 and helium through 11 mm.
i.d. glass tubing coupled to a microwave cavity. The mixture
of H^O and helium was made by passing helium gas through a
vessel which contained liquid H2O. The microwave power supply
was of the same type as that used for the generation of
hydrogen atoms. The H20 is dissociated as it passes through
the microwave cavity into H and electronically excited HO,
thus emitting light with a strong component near 310 nm.
wavelength. This is used for fluorescence excitation of HO
radicals inside the flow system.. This light source is directed
through one of the apertures of the six way fitting in the
detection area of the flow system by means of two quartz
lenses. These lenses have been set up to obtain the focal
point of the light source near the center of the flow system,
thus reducing the amount of scattered light that could other-
wise be detected by the photomultiplier (Fig 3).


28
Figure 3. Optical set up.
The fluorescence of the HO radicals was detected using
a Thorn EMI photomultiplier for photometry and
spectrophotometry applications. The necessary voltage for the
tube was furnished with an independent 15 volts power supply
which was then amplified to a low current and high voltage
(800 volts) line connected via coaxial cable to the photomul-
tiplier housing. A narrow band interference filter was placed
between the previously mentioned quartz window in the sixth
aperture of the six-way fitting and the photomultiplier. This
filter transmitted only light in the 280-325 nm. wavelength
region which allowed the fluorescence of the HO radicals to be
transmitted to the photomultiplier.


29
The photomultiplier output was connected via a BNC
cable to a EMI Gencom model AD-100 amplifier discriminator
which reduced the noise and amplified the electron bursts into
regular pulses. The amplifier discriminator was then connected
via a 9 pin connector cable to a EMI Gencom model C-10 photon
counter which displayed the number of photons per set time
interval. The photon counter has different interfacing
capabilities which output the photon counts either through
digital, analog or TTL outputs. The last was used to interface
to the IBM Personal Computer via a BNC cable. The fluorescence
signal was determined after subtracting the background light
levels. The background was measured by the photon counts
observed when the HO was not present (during the time the NO^
reactant was turned off). These counts were measured before
and after each signal reading.
Experimental Procedure
The experiments were performed after conditioning of
the flow system with only HO radicals in helium flowing
through the flow tube for approximately 30 to 60 minutes. The
HO concentration was determined by two techniques. The first
one involved the titration of H atoms produced in the
microwave discharge by the addition of a measured amount of
N02> while the HO fluorescence was monitored. The hydrogen
atoms will react with NO^ with a rate constant of 1.3 x 10 10


30
3
cm /particles-sec, which is considered to be a fast reaction.
N02 + H- -* HO* + NO*
This reaction produces one HO radical for each N02 molecule
added to the system. The signal observed is related to the
amount of HO produced. There will be a proportional increase
in the signal as the N02 is added to the system. After all the
hydrogen atoms produced in the discharge react with the added
N02 then any further addition of N02 will not produce any more
HO radicals, therefore the signal will stay the same. These
two trends observed in the signal as the N02 molecules are
added will produce a titration curve. The endpoint of this
curve occurs at the point at which the curve changes slopes.
From the determination of the endpoint an equivalence between
the N02 concentration and H atom concentration was determined.
The second technique was based upon a calibration
curve for HO fluorescence signal versus N02 concentration.
This curve was generated by monitoring the HO fluorescence
signal resulting from the addition of a measured amount of N02
to an excess of H atoms. The HO concentration is assumed to be
equal to the concentration of added N02 and proportional to
the signal. Using this calibration curve, the HO concentration
can be determined by the amount of signal observed. The two
techniques were in reasonable agreement, within a factor of


31
two for the different HO concentrations used.
A titration curve was generated daily before the
kinetics runs or when the HO concentration was changed and
sometimes at the end of the day. The signal levels were
measured before and after the kinetic runs for a given HO
concentration. A further note should be made to the observa-
tion that the efficiency of the production of H atoms in the
microwave discharge is inversely proportional to the amount of
gas flowing through the discharge. As the flow of
hydrogen/helium passing through the discharge increases then a
smaller hydrogen atom production efficiency in the
hydrogen/helium mixture results from the experimental titra-
tions. This could be seen from the results of the titrations
of different flows of hydrogen/helium through the discharge,
in which a flow of 300 seem showed a hydrogen atom concentra-
-ll 3
tion of 1.56 x 10 atoms/cm a flow of 200 seem showed a
-iJ 3
hydrogen atom concentration of 1.65 x 10 atoms/cm a flow
of 100 seem resulted in a hydrogen atom concentration of 1.77
-H 3
x 10 atoms/cm and a flow of 50 seem resulted in a con-
-H 3
centration of 2.^6 x 10 atoms/cm If the efficiency of the
discharge would have been the same in all flows it would have
resulted in the observation of the same concentration of
hydrogen atoms in the mixture, due to the fact that it is the
same hydrogen/helium gas source. The previous could be ex-
plained from the fact that as the flow is reduced, so is the


32
velocity in the discharge area. A slower velocity of gas in
the discharge area will increase the residence time of the
hydrogen molecules in it and subsequently increase efficiency
of the production of hydrogen atoms by the discharge.
Three different HO concentrations were used in the
experiments. The first one is described as Low HO concentra-
11 -3
tion and it ranges from 1 to 3 x 10 particles-cm the
second one is described as High HO and ranges from 1 to 3 x
12 -3
10 particles-cm and the third HO concentration is
1 2
described as Very High HO and it ranged from 8 to 16 x 10
-3
particles-cm The approximate HO fluorescence signal levels
, it
for these three concentrations were respectively 2 X 10 2 X
5 6
10 and 1 X 10 counts per second. The standard deviation for
the signal at low HO was less than 2%, and for high and very
high HO was less than 1$. These errors in the signal are
without any n-butane added. During the kinetics runs the
standard deviation of a particular data point increases as the
relative signal decreases, especially at low HO concentrations
and high n-butane. These standard deviations were used in the
weighted-least-squares analysis of the kinetic runs, and the
errors in the pseudo-first-order rate constants were usually
less than 10% for a particular n-butane concentration.
3
The background counts were from 1.5 to 3*0 x 10
counts per second, with a standard deviation of less than 1?.
The range for the background counts depended on factors which


33
included: the fluorescence lamp stability, any change in the
adjustment of the lenses, the conditioning of the flow tube,
the cleanliness of the window through which the light entered
the system, the concentration used and others. The background
was basically stable throughout a day of experiments, with a
total drift of a few (1-^) thousand counts. The difference
between the background before and after any data point was
less than 3%, which was controlled by the computer program.
The variation in the background from day to day was due to the
factors mentioned. The signal levels for a particular con-
centration were quite reproducible after adjusting for back-
ground levels. Wall loss rates determined from the intercepts
of the plots of pseudo-first-order rate constants versus n-
butane were generally less than 10 s 1. A study was conducted
in which the point of NO2 injection was varied and the slope
of the In (HO fluorescence) vs. wall exposure time was
measured. This resulted in a wall loss rate for HO of 6 s 1.
The use of fixed injector ports eliminates the variation in
surface area that is otherwise encountered in other discharge
flow systems that use movable injectors.
The rate of reaction for a particular concentration of
n-butane and HO was measured by monitoring the signal levels
after introduction of the n-butane through each reaction port
(i.e. different reaction times). The concentration range for
13 3
n-butane was varied from 1 to 9 x 10 molecules/cm and


3^
pseudo-first-order rates for six or more different concentra-
tions of n-butane were measured for a particular HO concentra-
tion. Repetition runs of the same concentration of n-butane
were done at times to verify reproducibility. Daily experi-
ments were performed for both Low HO and High HO to account
for any systematic fluctuations. At least ten readings for
each of the transducers were done for each particular data
point. The measurements of the transducers included voltage
readings for the pressure gauge, total counts of the fluores-
cence signal, voltage reading for each flow setting in the
flow controllers and voltage reading for the position of the
reaction port. The set of measurements for each transducer was
averaged and if the standard deviation of the total photon
counts was more than 1 $ of the counts reading a new set of
measurements for all transducers was taken. The background
photon count levels were also monitored before and after each
signal reading. If these two were different by more than 3%
another new set of data was collected. Approximately 30
seconds were allowed before readings were done for a par-
ticular reaction port in order to allow for conditioning of n-
butane at the new position and more than one minute elapsed
before readings were done when a different n-butane concentra-
tion was used. The control and monitoring of the transducers
and experimental parameters will be discussed in the com-
puterized system section of Chapter 3.


35
Measurements of bimolecular rate constants for the
different HO concentrations were reproducible from day to day.
The experiments were performed at room temperature which
fluctuated between 297 to 301 K. Initial problems with the
thermocouple amplifier circuitry, which was susceptible to the
high voltage discharge of the Tesla coil used to turn on the
microwave cavities, inhibited the use of these for the
measurement of temperatures in the flow system. Although most
of the experiments were performed at 1 torr of pressure, some
were performed at H torr to try to determine if there was any
pressure dependence.
The trapping of reaction products was performed under
conditions which would maximize yields of the products of
interest, mainly Very High HO concentrations. At these con-
centrations we would expect to have more HO and butyl radical
reactions, thus more butene or butanol formation as possible
products. The n-butane concentration used was in the low range
13 3
studied (1 x 10 molecules/cm ). The system was stabilized to
these conditions and the trap was submerged in liquid nitrogen
for approximately four to five hours. At the conclusion of
product trapping the glass trap was then pressurized to room
pressure and the bath was changed to a bath of CCl^ and dry
ice. The contents of the trap were then Injected to a Hewlett
Packard Gas Chromatograph-Mass Spectrometer (Models 5890A,
5970) for analysis. The head pressure in the GC-MS was kept at


36
6 psi and the temperature in the capillary column was main-
tained at 3^7 K. Previous runs with n-butane, sec-butanol,
trans-2-butene and air were used to measure the corresponding
retention times.


CHAPTER 3
COMPUTERIZED SYSTEM FOR STUDYING
GAS PHASE KINETICS
Advantages
The increased power of personal computers (PC) and the
lower cost and better performance of data acquisition equip-
ment facilitates computer control of experimental conditions,
data acquisition, manipulation, and storage. A PC interfaced
with experimental transducers enhances the ability to perform
experiments and make measurements.
The advantages for interfacing are numerous. First the
control of initial and ongoing experimental conditions and
monitoring of these conditions can be relegated to an
electromechanical servant. This automation facilitates the
tasks involving large numbers of repetitive operations. The
continual need to change experimental conditions and the large
amounts of data to be monitored can be successfully automated.
Secondly the quantity per unit time and quality of the data
acquired are increased. The data acquisition equipment has the
capability to access and read numerous devices in a short time
as compared to a human investigator. This capability will
markedly increase the amount of information obtained from each
experiment. This quantity of data and the ability to determine
statistical significance of data at the time of acquisition
helps to discriminate against data of inadequate quality.


38
Third, the computational power of a computer gives the re-
searcher a flexible tool to adjust the way in which the infor-
mation retrieved from the experiments will be analyzed.
Fourth, the use of additional accessories like printing,
plotting and disk storage devices facilitates the display and
storage of data obtained. Graphic displays can be presented
for further evaluation of experimental results. Fifth, and not
least, is the convenience of time savings given to the re-
searcher which can be used to immediately study the results
obtained and design new experiments.
Computerized systems to monitor and control experi-
ments are commonly found in classical laboratory equipment,
for example Gas Chromatographs, Mass Spectrometers, Nuclear
Magnetic Resonance instruments and others. The discharge flow-
resonance fluorescence technique is one in which the in-
strumentation is custom designed to meet the research needs.
The use of a computer system with such non-traditional ex-
perimental equipment requires that a series of signal
modifiers be incorporated in order to communicate with the
experimental transducers. This process is known as hardware
interfacing of the computer with the experimental transducers.
It is also necessary to write the instructions or program that
will incorporate all the needed algorithms. This is the
software development necessary for the computer to be able to
perform the. experiment.


39
Hardware Interfacing
The interfacing of the computer and experimental
equipment requires that the transducers used in the experi-
ments generate a signal or will respond to a signal generated
by the computer. The signal is considered to be an electrical
current at a given voltage level or a pulse that can be
counted. For the transducers used in the flow system we can
summarize the signals generated or used by those listed in the
following table 2
Table 2. Transducers used in gas flow kinetic experiments.
***tt**Ktt***ft****tt***tttt***tt
TRANSDUCER_________________________
PHOTOMULTIPLIER
CAPACITANCE MANOMETER
THERMOCOUPLE TYPE J(I/C)
MASS FLOW CONTROLLER
MULTIPORT VALVE ACTUATOR
*************tt***ttftft******
a**#######*##########*####*##***
:SIGNAL
: PULSES OUTPUT
:0-10 VOLTS ANALOG OUTPUT
:millivolts ANALOG OUTPUT
:0-5 VOLTS ANALOG INPUT
:0-5 VOLTS ANALOG OUTPUT
.CONTACT CLOSURE CIRCUIT
:SWITCH OF 115 VOLTS AC
**tt***ft****ftft****Xtt*******
The photomultiplier generates pulses or bursts of
electrons. The number of pulses is related to the amount of
light that hits the photocathode of the photomultiplier tube.
The signal of interest is due mainly to fluorescence of HO
radicals in the system, and scattered light from the source.
These pulses are bigger than 1 millivolt.
The capacitance manometer for pressure measurements
consists of a flexible metal diaphragm placed between two


40
fixed electrodes. When different pressures are applied to each
side of the diaphragm it causes a deflection which results in
a change in capacitance between the electrodes, thus a change
in output voltage. One side of the diaphragm is vacuum sealed
-7
to less than 1 0 torr and the other side receives the pres-
sure which causes the deflection. The output voltage is linear
in pressures from 0 to 10 torr in the flow tube.
The thermocouples transmit a different millivolt
signal related to the difference in thermal conductivity
between an iron and a constantan metal junction. The dif-
ference for the thermal conductivity in the junction is re-
lated to the temperature of the environment. The thermocouples
output low voltage signals.
The mass flow controllers are controlled with a 0 to 5
volt signal. This input command signal into the flow control-
ler opens the valve to obtain the required flow of gases with
a response of 6 seconds to within 2? of set point. An output
of 0 to 5 volts DC is proportional to the measured mass flow
rate in the sensor of the flow controller. A contact closure
circuit opens or closes the valve on the flow controller. It
is necessary to close the contacts if it is required to shut
the valve of the flow controller.
The multiport valve actuator is a motor that moves
unidirectionally for an 8-position Stream Selector Valve. The
movement is activated by 110-130 volts AC signal that starts


the motor. The setting of the position in the valve is ob-
tained through means of a cam and a contact closure switch. As
the cam moves so does the switch position, which deactivates
the movement of the motor at a new valve position. The intro-
duction of the 110-130 volts signal through another line
reactivates the motor and changes the valve to a new position.
Some of the signals to be read from the previous
transducers or needed to control them are not necessarily
compatible with the signals generated or read by the data
acquisition equipment of a computer. It is necessary to condi-
tion some of the signals between the transducers and the
computer in order to obtain compatibility. The signal con-
ditioning devices used for these transducers are presented in
table 3.
Table 3- Signal conditioner devices.
ft****************************************************************
TRANSDUCER :DEVICE :SIGNAL
PHOTOMULTIPLIER :AMPLIFIER DISCRIMINATOR DIGITAL PHOTON COUNTER :ECL PULSE :TTL PULSE OUTPUT OUTPUT
THERMOCOUPLES :MONOLITHIC THERMOCOUPLE
AMPLIFIER AD59^ : (10mV/C) OUTPUT
FLOW CONTROLLERS: RELAYS 5 :5 VOLTS INPUT
:RELAY DRIVER :DIGITAL INPUT
VALVE ACTUATOR :POTENTIOMETER :0~5 VOLTS OUTPUT
:MULTIPORT VALVE RELAY :5 VOLTS INPUT-115 VOLTS
**************************ft**************************************
The amplifier discriminator amplifies the pulses
generated by the photomultiplier. The presence of a photomul-


42
tiplier signal pulse of 20 microamps (1 mV threshold voltage)
or greater will cause the amplifier discriminator to output a
50 nanosecond differential ECL (electron coupled logic) pulse.
This ECL pulse is read by the digital photon counter which in
turn can output a TTL signal of more than 2.4 volts for each
ECL pulse received. The TTL signal can be read by a computer
logic system. The millivolt outputs of the thermocouples are
amplified by a monolithic thermocouple amplifier chip that
outputs 10 millivolts per each degree of temperature relative
to an ice point reference. These amplifiers have been
precalibrated to match the characteristics of type J (iron-
constantan) thermocouples.
Five Gordos 8312 relays were used to open or close the
circuit that controls the shut off valve of each of the five
flow controllers. These are activated by 5 volts passing
through a coil in the relay which in turn creates a magnetic
field that closes a switch, thus closing the circuit. Due to
the low current levels generated by the computer it was neces-
sary to use a Sprague model ULN-2003A high current relay
driver made of transistorized circuitry. This driver was
activated by the TTL output of the computer's digital lines
and used to drive an external 5 volts line through the coil of
a selected relay. The following figure 4 represents the way in
which these were connected.


U3
Figure High current relay driver connections.
The potentiometer was used to monitor the position of
the actuator for the multiport valve. This potentiometer was
attached to the multiport valve with two gears perpendicular
to each other. One of the gears followed the movement of the
actuator valve selector axis and concurrently moved the second
gear that was coupled to the axis of the potentiometer. A five
volt line from an external power source was connected to the


potentiometer. As the potentiometer axis rotated it changed
the internal resistance, outputting a particular voltage
related to the position of the multiport valve. A Magnecraft
relay model W88CPK-5 was used to control 115 volts AC which
drives the actuator motor circuit of the multiport valve. This
relay was controlled by a single digital TTL output from the
computer. A change on or off of TTL output from the computer
activated the relay which then changed the path taken by the
a.c. power used to drive the motor. A series of switches were
used to set up either manual control of the actuator or com-
puter control of it. Figure 5 represents their configuration.


Figure 5. Control of multiport valve.
The following table 4 represents a summary of all the
necessary lines that were Interfaced with the computer system.
These are described according to their input-output mode in
relation to the computer system. The A/D mode represents an
analog signal from a transducer that needs to be read as a
digital signal by the computer. The D/A mode represents a
digital signal from the computer that needs to be converted to


an analog signal in order to control a given transducer. The
D/D mode represents a digital signal from the computer to be
interpreted by a transducer as an ON/OFF condition.
Table Summary of input/output modes.
****#***********************a#*###*#*#*##*####*##*
A/D D/A D/D PULSES
PHOTOMULTIPLIER 5 FLOW CONTROLLERS 5 5 5 1
PRESSURE GAUGE 1
2 TEMPERATURE 2
MULTIPORT VALVE 1 1
TOTAL 9 5 6 1
a#*###*#*##*####****###*##*#####*#*###**####*#**##**##*####*#*##
The selection of the data acquisition equipment was
based on having the capabilities to perform the input/output
functions previously described and fast counting of pulses
generated by the photomultiplier. An IBM XT Personal Computer
was used to interface with the flow tube system. It consisted
of a hard disk, a 360K floppy disk drive, 640K of memory,
color monitor and graphics printer. Two boards from Tecmar
Incorporated were used to perform the data acquisition task.
These boards were added to the bus system of the PC. The first
board known as Lab Master Board can perform the four modes of
input/output for the different signals. Some of the main
features of this board are outlined in the following table 5.


Table 5. Features of the Lab Master board.
***ttffft*ft**N*ft********#******ft*****ft**ttft*ft**
ANALOG/DIGITAL:
16 single-ended or 8 true differential inputs
12 bit resolution
30 KHz A/D conversion rate
Voltage range selection 10V, 0-10V
DIGITAL/ANALOG:
2 independent D/A output channels
12 bit resolution
Voltage range 2.5V, 5V, 10V, 0-5V, 0-10V
DIGITAL/DIGITAL I/O (INTEL 8255)
24 parallel input/output lines programable in
groups of 8 or 12
TIMER/COUNTER (AM9513)
5 independent 16*bit counters(cascadable)
Programable gating and count source selection
#**#*#a#****#################*##**#*##*#**#*##*#*#*######**##***
Due to the limited number of digital to analog output
channels for the Lab Master board it was necessary to add an
additional board that would perform D/A conversions for
another three channels. This was performed with a second board
from Tecmar Incorporated, known as a Dadio Board. This second
board features an addition of four digital/analog channels and
24 parallel digital I/O lines.
The analog to digital convertor channel's were set at a
range of -10 to +10 volts to read the signals generated by the
nine transducers lines. The two digital to analog convertor
lines from the Lab Master and three of the same type of lines
from the Dadio board were set to output a voltage in a range
of 0 to +10 volts. These settings were done by switches found
on the boards. Of the 24 parallel digital to digital
input/output lines in the Lab Master only six were used to


48
control the five contact closure circuits of the flow control-
lers and the additional line that controls the relay for the
actuator, only requiring the use of one eight bit D/D port.
The five independent counters can be set up to inter-
act with each other through hardware and software control. Of
the five independent 16 bits counters only three were used to
read the pulses. The first counter was set up as a timer to
count for a requested time interval. It uses an internal
oscillator in the board with a 1 MHz crystal as the source of
counts for timing. The maximum counts in BCD for 16 bits in
counter one is 9999. The reading of counts from the internal
oscillator can be done in different scales, allowing for
flexibility in the range of time to be set to count. The
second counter reads the signal from the source (photomul-
tiplier) and when its contents is full it overflows into the
third counter, thus allowing a maximum count of more than 9.9
7
x 10 pulses of signal to be counted. Counter one was
programmed to output a high level signal when it completes the
counting of a given time interval. This signal goes through a
hardware connection to the third counter. The second counter
was also programmed to sense this signal in the third counter,
and to deactivate itself when it was sensed, thus stopping the
counting process.


Software Description
The program used to control and monitor the experi-
ments was written in the BASIC Language. The selection of this
high level language was based first on its feasibility to
program the data acquisition equipment and secondly because
the kinetics experiments did not require control by a fast low
level language. This second consideration can be seen from the
fact that there is a need to wait until proper conditioning of
the system is achieved when a new set of conditions is
programmed.
Programming of the data acquisition equipment is
obtained by writing or reading a byte of information at the
different ports (addresses) of the boards associated with the
different signal modes. The different ports (addresses) in the
boards are determined relative to a set memory location
through means of an external switch. For example if the board
is located in decimal memory location 1808 and the port to
write the first 8 bits of a byte for a digital to analog
conversion is located 2 relative to the set memory location,
then a digital to analog signal output can be achieved by
writing the byte to decimal memory location 1810 of the com-
puter. The use of the function INP or OUT of the BASIC lan-
guage will read or write a byte from or to a particular port
of the computer memory. The relative memory locations for the
different ports (addresses) of the Lab Master Board are


50
presented in the following table 6.
Table 6. Programming of the Lab Master board.
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
RELATIVE READ (INP)
MEMORY OR
LOCATION WRITE (OUT) FUNCTION
0 W D/A#1 (LOW 8 BITS)
1 W D/A#1 (HIGH H BITS)
2 w D/A#2 (LOW 8 BITS)
3 w D/A#2 (HIGH BITS)
i w CONTROL (INTERR.,INCRE.,ETC)
5 W/R A/D CHANNEL (SELECT,LOW BITS)
6 W/R A/D START (CONVER..HIGH BITS)
7 w TIMER INTERRUPT
8 W/R DATA PORT OF TIMER
9 W/R CONTROL PORT OF TIMER
10-11 NOT USED
12.13,14 W/R DATA OF D/D PARALLEL PORTS
15 W CONTROL OF D/D PORTS
XXKXXXXXXXXXXXXXXXXXXXXXXXKXXXXXXXXXXXXXXKXXKXXXXXXXXXXXXXXXXXX
The programming of the Lab Master board as well as the
Dadio board is straight forward in terms of reading and writ-
ing to the ports that are associated with the A/D and D/A
lines. The counters are each programmed individually by writ-
ing to the control port of the counters. This individual
programming of each counter is done taking into consideration
the functions to be perform by each counter. There is also an
individual loading of initial data on each counter prior to
activating them, which is performed by writing to the data
port of the counters. The reading or writing to each 16 bit
counter is done by reading/writing 8 bits at a time, due to
the fact that the PC is an eight bit computer. Therefore the
reading of the contents of a counter is done by assigning the


51
first half of the sixteen bit readout as the low values and
the second half as the high values. The final resulting number
will be a combination of these two numbers. If binary coded
decimal units are selected, then the high values will repre-
sent 100 times its contents as compared to the low values. If
an additional counter is used to count for overflows in the
first counter then the final resulting number will be a com-
bination of the two numbers (high and low) of each counter in
which the high values of the second counter represents 1000000
times its contents as compared to the low values of the first
counter.
The individual programs to perform the data acquisi-
tion task were incorporated as subroutines in a main program
named Actuator, see Appendix 1. The flow of tasks performed by
this program is outlined in the following figure 6.


52
a*##*#***#*#**##**########*###*####*##**#**########*#*####**##*#
SET UP GASES/FLOW 4-4-4--
CONDITIONS +
NON-FLOW t
TURN 4-
ON | D/D ,D/A I +
FLOWS t
DATA #

ACQUISITION * OFF D/D COUNTER ON
+ + + PORT D/D A/D READ A/D
+ INCRE.D/A READ COUNTER
+ * ON D/D AVERAGES
* * OFF D/D 4-4-4-4- CRITERIA
CRITERIA +
+ SIGNAL *
NEXT PT. r
+
OUTPUT READINGS

CONCEN.
COMPUTATIONAL VELOCITY
TASK RXNS.TIME
WLSQ
OUTPUT
+
RESULTS
TITRATION
OR
RATE GRAPHICS

+
+



-*-*+-------
Figure 6. Program flow


.53
The first part of the program goes through an interac-
tive section in which the user sets the experiments and condi-
tions to be performed by the computer. This section starts by
opening a file that contains the previous conditions used and
loads these as default values. The previously used conditions
include calibration parameters of the flow controllers for a
particular gas, name of reagents, concentration of these
reagents in the cylinders, flow of gases previously used,
increments in the flow of gases, and other parameters not
related to the flow of gases. The user is asked if he desires
to change any of the previous conditions, if so, he can go
through each flow controller known as Alpha, Beta, etc. and
change any of the previously mentioned parameters. The last
part of this interactive section takes the user to the setting
of non-flow parameters. Here the user can select the amount of
time to be loaded to the timer in the counters, the upper
value in the Y axis of the output graphs, and the mode of
counting. In mode one the computer reads all the transducers
while each counting interval is performed or in mode two the
computer counts signals for each reading of each transducer.
The user can also select which flow controller is used with
the N02 and hydrogen source. This section allows you to set
the number of readings to be done per data point by setting
the number of times for the counting cycle and also to set the
stabilization criteria used to accept a point of data. This


54
stabilization criteria sets the limit in the value to be
accepted for the standard deviation. If the standard deviation
of the total of readings of the fluorescence signal is bigger
than a percent value of the average of this signal, then a new
set of readings is obtained. The user can also select the
waiting time before readings are done when either N02 is
turned on or off. The last part of the setting of non-flow
parameters sets the number of points to be collected (maximum
of 10) and if the multiport valve (actuator) will be active or
not.
The program sets the flow of gases by turning on
(opening the contact closure circuit) the flow controller
valves through means of the D/D lines and setting the required
voltage in the digital to analog lines that sets the flows in
the flow controllers. The required voltage is calculated from
the set flow of the gases stated in the interactive section of
the program and the calibration parameters for the flow con-
troller. The program adjusts the total flow of gases to be
equal to the set flow of the helium diluent that flows through
the Epsilon flow controller. Therefore the total flow in the
system depends on the initial set flow of the helium diluent.
The data acquisition task is mainly composed of two
sections. In the first section it turns off the N02 gas by
means of a subroutine that interacts with the D/D port. Then
it either sets a port location for the multiport valve if the


56
After all the data have been collected for a par-
ticular point, which includes data for OFF, ON and OFF cycles,
it verifies that the baseline (background signal in the OFF
cycles) does not differ more than 3? between the two OFF
signals. The baseline is the average of these two OFF signals.
If the criteria for the baseline is not met it collects
another ON and OFF set of data. If the new point is acceptable
it calculates the HO signal as the difference between the
signal in the ON cycle and the baseline. Thereafter it goes to
the collection of other points by alternating between ON and
OFF cycles of the data acquisition sections and setting new
flows or port positions.
The first output task prints the values of the read-
ings of voltage for the different flow controllers, pressure
gauge, thermocouples, port positions, total counts, baseline,
and signal. Thereafter it goes into the computational task
subroutines which first calculates the concentration of the
reagents in the flow tube based on the individual flows, total
flow, pressure, and concentration in the reagent cylinders.
Another subroutine calculates the velocity of the gases in the
flow tube by taking into consideration the standardized total
flow rate, the approximate inside diameter of the flow tube
and the pressure. After this, a computation is performed if
the actuator is active, in which the reaction time at the
different ports is calculated based on the velocity at which


57
the gases travel inside the flow tube and the distance between
each port and the detection area. If the multiport valve is
active then a weighted least squares analysis subroutine is
used to calculate the slope of the natural log of the HO
signal versus the reaction time at each port.
The last output includes the previously calculated
parameters of concentration, velocity, reaction time at ports
and slope (pseudo-first-order rate constant) and a graphical
display of the signal versus flow of NC^ for a titration or
natural log of signal versus reaction time for a kinetics run
will be displayed. After these outputs the program can be
started again with previously used conditions or setting new
ones in the set up conditions section.
The calibration of the flow controllers using the soap
bubble method was monitored by the use of another software
program named Comascal. Comascal was written by Paul M.
Gates, and includes a least-squares-analysis method to
compute the slope of the flow and set voltage. The least-
squares-analysis method is based on a publication from Irving,
(17)
J. A. and Quikenden, T. I. The program calculates the
slope of flow versus voltage, the flow intercept, errors in
the slope and intercept, the covariance, and correlation
coefficients.


CHAPTER H
RESULTS
Experimental
The experimental results for the kinetic runs were
obtained from the data output of the computer with the cor-
responding experimental conditions. As mentioned previously
the experiments were started with the titration of the
hydrogen atom content in the flow system with NO^, thus an
approximate HO concentration was determined. An example of
titration of one of the experiments is presented in the fol
lowing figure.


59
N02 TITRATION 14:59:59 04-28-1966
150000
134990
$
I 119980
G
N 104970
A
L 89960
C 74950
0
U 59940
N
T 44930
S
29920
14910
,0
1 i i 1 T i r i
0 v 8
0
0
0
0 *
0
0
0
1 1 I I 1 i i i 1
15.6 31.2 46.8 62.4 78.0 93.6 109.2 124.8 140.4 156
FLOW N02
Figure 7. Titration curve.
This shows the relation of the HO fluorescence versus
the flow in standard cubic centimeters per minute of NO^. As
the NOg flow is increased, the production of HO radicals in
the system increases, until a maximum production of signal or
HO radicals is reached. Above this N02 flow the signal stays
relatively constant or could decrease slightly due to the


60
reaction of HO radicals with the excess of NO2. This last
point is insignificant at the conditions of pressure and
velocity used in the experiments. The end point of the titra-
tion can be calculated by finding the intercept of the two
lines that would follow the two different trends in the graph.
It is assumed that every NO2 molecule reacts with a hydrogen
atom to form HO and NO.
N02 + H* -* HO* + NO*
This last reaction is considered to be fast at 1.3 x
-10 3
10 cm /particle-sec. The concentration of hydrogen atoms in
the system can be calculated from the results of the point
where the signal changes slopes. By knowledge of the con-
centration and flow of N02 at the point of change in the slope
and the flow of the hydrogen/helium mixture we can compute the
concentration of hydrogen atoms or HO radicals.
The pseudo-first-order decay of HO is shown for a
particular n-butane concentration in figure 8.


61
0.00280.00560.06840.(31120.01410.01690.01970.02250.02530.0281
TIME IN SECONDS
Figure 8. Pseudo-first-order decay.
The figure shows a plot of the natural log of the
signal versus the eight reaction times in seconds. The reac-
tion time corresponds to different reaction distances of the
hydrocarbon reactant introduced through eight reaction ports
at a fixed total flow velocity. The slope of the line cor-
responds to the pseudo-first-order rate constant for the


62
reaction of HO radicals at the set conditions.
The experimental results for all the kinetic runs are
presented in appendix 2. The date, n-butane concentration,
pseudo-first-order rate constant, and standard deviation are
included. Also included is the apparent bimoleeular rate
constant for a particular set of data and for a whole set of
comparable HO conditions and corresponding intercepts. Kinetic
results for higher pressure are also included. It should be
pointed out that not all of the decay kinetics results were
(JO ,
obtained under the previously recommended conditions of a
forty-fold excess of n-butane over HO radicals, especially for
the Very High initial HO concentration experiments. This is
due to the nature of the experiments performed, in which the
effects of secondary reactions are bigger when there is less
excess n-butane. The second point to notice is the use of the
term "apparent bimoleeular rate constant" which is the calcu-
lated rate constant for the bimoleeular reaction of HO radi-
cals and n-butane, assuming no other reaction process is
affecting this result, which is not what it was found in this
study.
The calculation of the apparent bimoleeular rate
constant for a particular set of data was done by performing a
weighted linear least-squares-analysis of the pseudo-first-
order rate constants versus the n-butane concentrations,
weighted by the error in the rate constant. As we can see from


63
the experimental results in appendix 2 there is a significant
difference in the results obtained for pseudo-first-order rate
constants at different initial HO concentrations. With the
exception of the very high intercept for the last data set in
the Very High HO results, the results are quite comparable
between each set collected at the same HO concentration. As
mentioned before, the intercepts are related to loss processes
for HO radicals which are independent of reactions with n-
butane, such as the loss of HO by interaction with the walls.
This is believed to be the case for the Very High HO results
due to deconditioning of the wall prior to these experiments.
It can be seen that there is a bigger difference
between the pseudo-first-order rate constants measured at high
n-butane concentrations compared to low n-butane concentra-
tions, for the different HO concentrations. This difference
can be readily explained from the fact that even though there
is a larger percent of secondary reactions occurring at lower
n-butane concentrations, the effect is less noticeable in the
small rate constants in this lower n-butane concentration
range than is a smaller effect of secondary reactions affect-
ing a bigger pseudo-first-order rate constant. An example of
this would be a 25? effect on a given rate constant of ^0
sec ^ would be a net effect of 10 sec ^ in the rate as com-
pared to a 10? percent effect on a rate constant of 200 sec
which would be a net effect of 20 sec ^.


In (Relative H0>
The following graph compares the HO signal decay for
low and high initial HO concentrations at a n-butane con-
13 3
centration of 7.5 x 10. molecules/cm done on May 1H, 19B6.
The experimental signals were normalized relative to the
signal at the smallest reaction time within each set of points
and the natural log of this relative signal was plotted versus
the reaction time.
T I M£ (sec. )
Figure 9. Relative signal decay of HO for High and Low HO
concentrations.


65
The non-linear effect of the signal decay at high HO
concentration is quite noticeable when compared to low HO
conditions. This effect is more apparent at longer reaction
times. This effect could be explained by the fact that at
longer reaction times there will be more product n-butyl
radical formed, therefore, the HO radical will be more likely
to react with this product. A higher rate of reaction for a
second process compared to the rate of the reaction for HO and
n-butane is necessary to account for the deviation observed.
Figure 10 shows the values of all the pseudo-first-
order rate constants plotted versus the n-butane concentra-
11 3
tions, for initial Low HO (2 x 10 particles/cm ), High HO (2
12 3 13 3
x 10 particles/cm ) and Very High HO (1 x 10 particles/cm ).


RATE k obs. (1/s)
66
+L0W *HI GH xVERY
Figure 10. Kinetic results for Low, High and Very High HO
concentrations.
The lines in figure 10 were obtained by plotting the
results of an unweighted least-squares-analysis. The analysis
by the least-squares techniques gives slopes of 2.25, 2.91 and
2.96 for Low, High and first set of Very High HO concentra-
tions which is even less dramatic between High and Very High


67
HO than the weighted least square analysis. The least-square
method is used in the previous plotted figure. The graph gives
us a visual representation of the experimental results. These
results are not significantly different from the weighted
analysis shown in appendix 2. The difference in appendix 2
-12
between the apparent Low HO bimolecular rate (2.231.0*1 x 10
cm^molec !*sec and the apparent High HO bimolecular rate
-12 3 -1 -1
(2.85.05 x 10 cm molec sec ) is quite significant. On
the other hand, the difference between apparent High HO
(2.85+.05 x 10 cm^molec 1sec 1) and Very High HO
bimolecular rate (3-08.15 3.13.23 x 10 ^ cm^molec ^sec 1)
is less dramatic. A contribution to these differences is the
fact that there is a factor of 10 difference between Low HO
and High HO as compared to a factor of 5 difference between
High HO and Very High HO. Also to be considered is the fact
that the experiments for Very High HO were mostly done on the
lower end of n-butane concentration range where the change of
the pseudo-first-order rate constants is going to be smaller,
as previously mentioned. Even so, the values obtained at Very
High HO are smaller than expected, based on preliminary model
simulations. At Very High HO there is a larger quantity of
secondary reactions occurring, thus a higher apparent
bimolecular rate constant should have resulted.
Results for experiments performed at higher pressures
showed a reduction in the apparent bimolecular rate constants


68
when compared to the low pressure studies. This result is
different than expected. If the HO radicals add to butyl
radicals to form an alcohol, then the increase in pressure
would stabilize the activated complex. Therefore, the rate for
the secondary process would be higher, thus one would expect
an increase in the apparent bimolecular rate constant. If the
HO reaction with butyl radical proceeds by abstraction of
another H to make an alkene, one would expect no significant
effect of increasing pressure.
Trapping results of the products of the reaction of HO
radicals and n-butane are presented in the following
chromatogram and mass spectrum.
Figure 11.1 Gas Chromatogram.


Figure 11.2 Mass Spectrum.
The chromatogram shows three main peaks. Analysis of
the first peak at 1.388 minutes shows it to be air in the
injected sample which is expected due to the sampling tech-
nique used. The samples were collected from the trap submerged
in a bath of CCl^ and dry ice at atmospheric pressure. At
these conditions the gas sample of the products is in a liquid
state which immediately becomes a gas when it is collected
with the liquid sample syringe. The syringe will aquire air
from the environment when it is withdrawn from the trap.
Analysis of the mass spectrum for the peak at 1.388 showed a
high abundance of m/e 28, 32 and Mo, which are due to
nitrogen, oxygen and argon.
The second peak at 1.533 minutes' is n-butane. The last
main peak emerged at approximately the same retention time
(2.107 minutes) as the reference samples of sec-butanol. There


70
was no indication of a chromatogram peak or mass spectra for
butene as a product. Further analysis of the mass spectrum of
the peak at 1.533 minutes of n-butane was done to see if there
was any contribution of butene within the n-butane mass
spectrum. Previous studies of the chromatogram for butene
showed that this olefin emerged immediately before or at the
same time as the n-butane at the conditions used. The analysis
of the n-butane mass spectra to check for any contribution
from butene within it was done by checking the ratio of the
abundance of peaks 56, 55, 53. 51 and 40 relative to peak 29.
Peaks 56, 55, 53, 51 and 40 relative to 29 have been shown to
be significantly larger for trans-butene (2.5/1, 1.24/1, .6/1,
.46/1 and .53/1) as compared to the same peaks relative to 29
in the n-butane spectra (.026/1, .029/1, .026/1, .031/1 and
0/1). If any olefin would have appeared together with n-butane
product peak it would have been observable in a higher ratio
of abundance of peak 56, 55, 53, 51 and 40 relative to 29. The
ratio of these mass peaks in the n-butane product peak of the
flow system were 0.024/1, 0.027/1, 0.020/1, 0.032/1 and 0/1
respectively. There was no indication that any butene appeared
before or within the butane peak.
The mass spectrum in figure 11.2 is presented for the
peak with a retention time of 2.105 minutes contains charac-
teristic mass fragments of n-butane and an alcohol (sec-
butanol). It also contains peaks that do not belong to either


71
species. The fragmentation pattern shows a 14 mass unit
repetition characteristic of hydrocarbons up to a mass unit of
57. The spectra also contains a mass peak at 30 that is not
characteristic of either n-butane or an alcohol. A m/e = 30 is
characteristic of fragments NH2CH2, CH20, or NO. It contains a
relatively large peak at ^5 which is characteristic of sec-
butanol, and a relatively small peak at 31 characteristic of
alcohols. The peak at 31 is relatively small compared to the
abundance of it for alcohols. A large peak at 57 is found
which is not as big in n-butane or sec-butanol, but is the
obvious result of the breakage of a parent compound that
contains butyl. Another relatively large peak at 7^ mass units
is found, which appears to be that of a possible parent peak.
This peak at Ik mass units is not as big as in the reference
sample of sec-butanol but is the mass of a butanol molecule.
These results suggest that the product contains some combina-
tion of a butyl and OH radical. It also suggest the possible
influence of NO or N02 molecules in the product based on the
large intensity at m/e of 30.
A further study was done to see if n-butane was react-
ing with NO, N02, HO or H20 after been trapped. This was
performed by trapping only n-butane for a given interval of
time. After n-butane was trapped it was turned off and the
hydroxyl radicals were made and also trapped together with NO,
N02 and H2Q. The results showed that none of these react with


72
n-butane in the trap at the conditions used. Therefore the
previous observed product at a retention time of 2.105 is
either the product of butyl radical and HO which further
reacted in the trap or in the gas phase with NO or NO^f or the
product is the result of butyl radical and NO or N0£ reacting
in the gas phase prior to being trapped or any other reaction
involving butyl radicals.
Error Analysis
The calculation of the uncertainty in the bimolecular
rate constant was based on the evaluation of the systematic
errors and random errors in the different parameters that go
into its different phases of its calculation. The propagation
of the systematic errors were done unidirectionally using the
absolute values obtained at the different phases of the cal-
culations.
The sources of systematic errors used in the calculations
were:
1) Propagation error results for calibration of flow control-
06)
lers using the results from the comascal program. These
varied for the different flow controllers and were less than
2% for the slopes of the calibrations.
2) A voltage setting resolution of 1 bit in ^096 bits for the
D/A, which for the 10 V range corresponds to .002 V.
3) A pressure error of 50 millitorrs, the main source of


73
uncertainty being in the zero offset value of the gauge.
The uncertainty of .5% in the concentration of n-butane in
the cylinder.
5) The uncertainty of 5? in the reaction system temperature
used to determine concentrations from flow and pressure
measurements.
6) An error of 1/62500 in the timer and 1/10000 in the read-
ing of counts by the counter to calculate the signals.
7) An uncertainty of .2 cm in the correct distance between
the different ports and the detector.
The results of the error analysis showed that the
biggest systematic uncertainties resulted in the calculation
of the flow of n-butane, which was about 3$ for the highest
flow of n-butane. For the lowest range of flow of n-butane
this error was 13%. The systematic error in the flow of n-
butane resulted in the calculation of systematic errors in the
concentration of n-butane that ranged from 9% at high n-butane
concentrations to 19% at low n-butane concentrations.
The calculation of random errors were mainly based on
the standard deviation of the set of readings obtained for the
different transducers and included:
1) The statistical error obtained for the calibration of the
flow controllers.
2) The experimental readings of the voltages of the flow
controllers for a particular set voltage. The variability


74
observed in these readings were approximately .005 volts for
a 0 to 5 volts range.
3) The readings of the voltage for the pressure which was
approximately .005 volts at 1 volt (1 torr).
4) The readings for the counts to estimate the signals. This
particular error was controlled by the software to be less
than 1 % for each data set used.
The biggest random errors were also related to the
flow of n-butane which ranged from about 3% for high flows to
11% for low flows of n-butane. The calculation of the random
errors in the concentration of n-butane ranged from 7? for the
high n-butane concentrations to 13% at the low concentrations
of n-butane.
The error analysis determined that the errors for the
total flow were 4% for both systematic and random analysis at
all ranges of n-butane concentration used. The major con-
tributor to the error in the signal was the random error and
was not higher than 1.2%. The errors in velocity and reaction
times were 2% from the systematic analysis and 4% from the
random analysis at all ranges of n-butane concentrations.
The error analysis of the calculation of the pseudo-
first-order rate constant for the total reaction time showed
the uncertainties were heavily influenced by the total reac-
tion time studied and the amount of signal change during that
reaction time. At high n-butane the systematic and random


75
error in the calculation of the pseudo-first-order rates were
6? and 10? respectively.
The use of weighted least-squares-analysis for the
calculation of the pseudo-first-order rates (weighted in Y
axis for signal error) and bimolecular rates (weighted in Y
axis for errors in the pseudo-first-order rate constants)
minimized the sum of squares of the deviations in the Y values
for the set of X values used. The weighted-least-squares error
results are the errors that appear in appendix 2. In general
the errors for the pseudo-first-order rates constants were
less than 5?. For the precision of the bimolecular rate con-
stants it can be seen that the errors weighted in the Y axis
also resulted in less than 5?. The randomness in the n-butane
concentration was not taken into consideration when the
bimolecular rate constants were calculated using the weighted-
least-squares. For similar experiments performed on different
days the random error in the bimolecular rate constant in the
results was found to be less than 5?.
The error in the determination of the bimolecular rate
constants based upon the propagation of systematic errors were
about 11?. The overall accuracy of the results for the
bimolecular rate constant were 12?. This value is near the 10-
15? accuracy range which is believed to be the best one can do
in flow tube measurements of rate constants as Indicated by
(31 )m
Howard


76
Modeling
Simulations of the kinetics involved for the reactions
of HO and n-butane were performed to elucidate the possible
mechanism and rates which affect it. The simulations were
performed in an IBM personal computer using the integration
techniques described by C. W. Gear. ^ 9) (20) pates use(j
for the reactions in the simulation were as recommended
(21)(22)
for the experimental conditions involved. The selec-
tion of reactions used in the simulations was based on which
of them could have a significant effect in the model. This was
done by taking into consideration the concentrations of the
species in the system and the rates at which these react with
other species. The rates for the reaction of n-butane and HO
radicals and the postulated reaction of HO and butyl radicals
were adjusted to best fit the experimental results.
Simulations were done which involved the modeling for
the formation and lifetime of HO radicals in the system for
the three different HO concentrations. The reaction for the
formation of HO,
H* + N02 -* HO* + NO*
and subsequent reactions were included. These simulations were
done omitting the reactions that involve the hydrocarbon in


77
order to obtain a description of the formation and disap-
pearance of HO radicals in the system. The results show that
the expected HO concentration was smaller than the previously
calculated concentrations from the titration and signal
monitoring technique. For Low HO the simulated HO concentra-
tion was 3? smaller, for High HO concentration it was 9$
smaller and for Very High HO it was 22% smaller. The calcu-
lated HO concentration from the model indicates a lower HO in
the system as compared to the experimental results from the
titrations. The results from the model HO lifetime are more
accurate and were used in the modeling of the reactions of n-
butane and HO. It was also noticed that the HO concentration
for Very High HO varied almost 25% between the first port of
reaction and the last port. The variation for High and Low HO
was b% and ^% respectively. Further studies of how the ap-
parent bimolecular rate of HO and n-butane was affected by
this variation in the HO concentration at Very High HO showed
no effect on the rate, but it reduced the intercept by ap-
proximately 50%. This reduction in the intercept from the
simulation is in accord with the low intercept observed ex-
perimentally. The simulations performed in this study were
done without varying the concentration of HO at the different
ports.
The main simulation model included 28 different reac-
tions and 19 species that were calculated. The list of reac-


78
tions and rates used appears in appendix 3. The reactants
involved in the simulation included H*, N02, NO, HO,
H20, H02, H2, HNO^, 0*, 02> NO^, n-butane, butyl, R-OH, R-0*,
octane, butene, alkoxy, R-NO. The value of the rate of com-
bination and disproportionation for butyl radicals was es-
timated from the rate data for similar reactions of propyl
/ pc N / pf. \
radicals'' and t-butyl radicals. The rate for the reac-
tions of NO and butyl radicals was estimated as the same as
(27)
the rate for the reaction of NO and methyl radicals. The
rate for the reaction of N02 and butyl radicals used was the
same as the rate of the reaction of N02 and methyl
(28)
radicals. It is believed that these last two reactions
could be faster than the values here used. This is from the
fact that radical reactions of HO and hydrocarbons increase in
their rate as the hydrocarbon chain increases. The simulations
were performed at eight different n-butane concentrations. The
apparent bimolecular rate constant was calculated via a least-
squares-analysis of the obtained pseudo-first-order rate
constants and n-butane concentration. The reactions were
simulated for 25 milliseconds, which is approximately the
maximum reaction time of the experiments. The simulations were
done with a tolerance error of 1$ in the calculations.
The first series of simulations was done to estimate
the minimum rate constant for the secondary reaction of HO and
butyl radicals. The simulations were done excluding all the


79
reactions that deplete the butyl radical, with the exception
of the reaction of HO radical and butyl radical. The exclusion
of these reactions maximized the concentration of butyl radi-
cals, hence calculation of a lower limit for the rate constant
of the secondary reaction of butyl and HO radical, in order to
best match the observed effect of the change of initial HO
concentration on the apparent bimolecular rate constant. The
simulations were done adjusting the values of the rate of the
primary reaction kj between HO and n-butane and the secondary
reaction rate k2 between HO and butyl radicals. The values
obtained are presented in the following table.
Table 7. Simulations with minimum k2.
a*#*#*#*##########*##*####****####*##*#*##*#*#####***##*###**##*
-12 3
kx = 2.1 x 10 cm /molec-sec
k2= 1.2 x 10 10cm3/molec-sec
a###*##*##**#*###*###*####*#**###*#*#*#####*##***#*##**#*#*####
APPARENT BIMOLECULAR RATE
INITIAL HO (x 10~12cm3/molec-sec)
CONCENTRATION_____________MODEL___________EXPERIMENTAL
"LOW 05 TTFh
HIGH 2:78 2.85
VERY HIGH 3.60 3.08-3.13
a#**#***###*#*###*#*#***############*##*#*###***###**##**#**###
As we can see from the results the values of rates
obtained with the model are similar to those obtained ex-
perimentally, with the exception of the values at very high


80
HO. The values obtained with the model at Low and High HO are
4$ and 2% different from the experimental values as compared
to a 16$ difference at Very High HO. The sensitivity of the
secondary reaction between HO and butyl radicals on the ap-
(29)
parent rate was previously studied in a similar reaction
scheme for the Low and High HO concentration range. This study
was conducted to show how k2 influences the apparent rate. The
(29)
results of the previous study are presented in the follow-
ing table.
Table 8. Effect of kj and k2 on model values for apparent rates.
**tt*tt*****K*******tt*****tt***********************
Apparent
- Bimolecular Rate for
(cnr/molec-sec) HO + n-Butane
LOW HO HIGH HO
,k> -12, ,k2 -10 (x io-12) Difference
(x 10 ) (x 10 .) ##****#*####*##*#*##*###**##***
1 .9 1.2 1.96 2.58 .62 *
2.1 ,U * 2.05 2.27 .22 *
2.1 .8 * 2.10 2:58 .U8 *
2.1 1.2 * 2..15 2.80 .65 *
2.1 1.6 * 2.20 2.97 .77 *
2.1 3:6 * 2.in 3.36 .95 *
Experimental - 2.23 2.85 .62 *
a#**#*****#*####*#***###**#**#**######**#####*#*##**#*#*#***#*#*
The magnitude of the difference between apparent
bimolecular rate constants measured at Low HO and High HO is
highly influenced by the value of k2 used in the model, and
not as much by the value of k,.


81
A second series of simulations was done to estimate a
maximum value of the rate constant for the secondary reaction.
This series of simulations included all the reactions that
involved the butyl radicals. Including these reactions, which
are loss processes for the butyl radical, reduced the con-
centration of butyl radicals in the model. Having a smaller
amount of butyl radicals will require that the rate of reac-
tion between HO and butyl be increased to be able to simulate
the experimental results. The results of these simulations are
presented in the following table.
Table 9. Maximum k2 value and effect of k2 on apparent
bimolecular rate.
a****#***#*#***#**#*#*####*****##**######*##*##*#####**########*
APPARENT BIMOLECULAR RATE
(x 10~^cm^/molec-sec)
1 k2 VERY
(xio"12) (x10"1) LOW HIGH HIGH
2.1 8 2.34 3.02 3.09
2.1 5 2.24 2.93 3-07
EXPERIMENTAL -> 2.23 2.85 3.07-3.13
a#####*#*#**##***#*#*######**#*#*####*##***#########****##*##**#
-10
The simulation with k2 = 5 x 10 gives values near
the experimental apparent bimolecular rate constant. The


82
series of simulations for maximum k2 included the reactions of
NO and NO2 with butyl radicals. These two reactions become
important at Very High HO concentrations, because of the high
concentration of NO and NO^ in the experiments. As we can see
the simulations for maximum k2 predict a relatively low value
for the apparent bimolecular rate constant at Very High HO,
which was observed experimentally.
The values of the experimental pseudo-first-order rate
constants and modeled pseudo-first-order rates constants
versus n-butane concentrations are presented in the following
figure 12 with adjustments made to all the rates in order to
make the graph easy to see. The successive curves were offset
on the y axis. The offset was obtained by adding or substract-
ing a given amount to a set of obtained pseudo-first-order
rates. This addition or subtraction will not change the slopes
of the curves. The intercepts of the experimental results for
the bimolecular rate constants for Low, High and Very High HO
were respectively set at 1, 10 and 20 s-1. The intercepts for
the model results were also set accordingly.


RATE k
+L *H xV #MODEL
83
The results for the concentration profiles of the main
species in the simulations for Low HO and High HO are
presented in figures 13.1 and 13.2. The profiles are presented
for the concentration of the species of HO radicals, butanol
and butyl radicals versus reaction time. The values of k,-2.1
-12 3 -10 3.
x 10 cm /molec-sec and ka1.2 x 10 cm /molec-sec were
used.


8U
[n-BUTANE]= 4E13
Figure 13.1. Simulated concentration profile for Low HO.


85
[n-BUTANE]=4E13
Figure 13.2. Simulated concentration profile for High HO.
The profiles in figures 13.1 and 13.2 for HO, butyl
and alcohol indicate a higher concentration of the product
alcohol, resulting from the secondary reaction between HO
radicals and butyl radicals, at High HO compared to Low HO
conditions. These profiles of the concentration were done at a
13 3
n-butane concentration of ^ x 10 molecules/cm The profile


86
for High HO indicate a loss of approximately 30? of the
original HO radicals to be due to the formation of alcohol. At
Low HO a loss of approximately 10? of the HO radicals is due
to the secondary reaction.
A third series of simulations was performed to model
the experimental conditions of the flash photolysis experi-
ments performed by Stuhl^ and Atkinson et al.^^ The reac-
tions and concentration of reactants used for these simula-
tions were based on the experimental conditions of these
systems. The rate for the reactions used were the same as
those rates presented in appendix 3, in which k for the secon-
dary process is assumed to be maximum. The simulation results
for these are presented in the following table 10.


87
Table 10. Flash Photolysis Simulations.
***#**#*#*#*a*****##**##****##*#************* a#*#######*#*#**###
STUHL SIMULATED APPARENT BIMOLECULAR RATE
HO RADICAL
CONCENTRATION
-12 3
( x 10 cm /molec-sec)
FIXED TWO
TIME LIFETIMES
3
(molec/cm )
3 x 10
2 x 10
EXPERIMENTAL
3 X 10
2 x 10
11 1 2 2.36 2.12
2.911 2.60
A .58 A .48
11 12 2.35
NO CHANGE
ATKINSON
1
2
11
8
16
10
10
10
10
10
11
11
11
11
11
EXPERIMENTAL
OH CONCENTRATION
NOT MENTIONED
2.11
2.25
2.50
2.83
3.30
Experimental--> 2.72
##*###****#****#****#####**###**####**##*####*#**#**#**#*###*
The simulations performed for Stuhl's system were
first done using a fixed time approach in which the total
reaction time was the same for all the different n-butane
concentrations used. The same set of n-butane concentrations
were simulated but the pseudo-first-order rate constants were
calculated over the time range required for 8991 % of initial
HO concentration to react. These two methods are compared
because most of the experiments done in flash photolysis use
this last method, where in the flow systems the fixed time
approach is used. The results for the apparent bimolecular
rate constants determined over the time range of 89-91$ appear


88
in the column labeled "two lifetimes". As we can see from
these two sets, the use of a fixed time approach gives higher
apparent bimolecular rate constant values compared to the
method in which the HO concentration is monitored for two
lifetimes. The difference between the apparent bimolecular
rate constant for the Low HO and High HO in the simulations
done is 18? smaller when the method of two lifetimes is used
as compared with the fixed time method. In the experiments
performed by Stuhl no significant difference was observed in
the results of the experiments performed at these two HO
concentrations. The reason for not observing a significant
difference could be due to the use of the two lifetime tech-
nique to analyze the time dependence of the HO concentration.
The simulations performed for Atkinson's system were
done by doubling the initial HO concentration for each set of
n-butane concentrations. The time interval was fixed in the
simulations, compared to the experiments performed in which
the pseudo-first-order rate constants were measured until at
least three HO lifetimes had passed in the reaction with the
different n-butane concentrations.^10^
Discussion
The results obtained in this study demonstrate the
influence of the secondary reaction.


89
HO* + butyl -* R-OH
This reaction is important in systems with high and very high
initial HO concentrations and less important in low HO sys-
tems. The influence that these secondary reactions have on the
measured rates is observed when the initial concentrations of
HO radicals is changed significantly. The changes in pseudo-
first-order rate constants observed indicate that these rates
increase as the initial HO concentration increases, thus the
results of the measurement of the rate constant for the reac-
tion,
HO* + n-butane -* butyl + H20
is going to appear higher as the initial HO concentration gets
higher. The significant change in the pseudo-first-order rates
is the result of fast secondary reactions which even influence
the results at Low HO concentration. Thus the measured results
for the reaction of HO and n-butane should be corrected ac-
cordingly.
The value of the rate constant for the secondary
process in which HO radicals react with butyl radicals appears
to be in the range from 1.2 to 5 x 10 1cm^/raolec-sec. The
maximum value for this reaction is dependent on the value of
the rates of the reactions of


90
NO^ + butyl -* alkoxy + NO
and
NO* + butyl + (M) -> butyl-NO + (M)
which have been assumed in the simulation studies to be
similar in magnitude to the reactions of nitric oxide and
nitrogen dioxide with methyl radicals. 1
The reduction in the apparent bimolecular rate con-
stant for Low and High HO when the pressure was increased
could be the result of either of two processes. The first
could result from an increase of butyl radical loss through
any pressure dependent reaction, reducing the importance of
the secondary reaction of HO and butyl. The reaction of NO and
butyl could be such a pressure dependent process. A higher
wall loss rate for the butyl radical could also account for
the observed effect. If this*first proposal is true then the
value for the bimolecular rate constant of the primary reac-
tion of HO and n-butane would be significantly lower than the
apparent bimolecular rate constant used in the simulations and
Ihe secondary reaction of HO and butyl radical would be ever
higher than that previously used as a maximum in the simula-
tions.


91
The second possible process that could account for the
reduction of the apparent bimolecular rates at high pressures
could be the result of the influence of a radial concentration
gradient.^0)(31)(32) change to a higher pressure and same
velocity of flow will create a bigger velocity differential
across the flow tube. The flow at the center of the flow tube
is going to be faster than the flow near the walls. The radial
concentration gradient is the distribution of concentratioon of
molecules across the flow tube. The diffusion of these
molecules is going to be less efficient at the higher pressure
conditions. Therefore the n-butane molecules are going to be
more concentrated towards the center of the flow tube because
these molecules are added directly to the center of flow. The
pressures are higher and the velocity of flow is faster at the
center as compared to lower pressure conditions. On the other
hand, the HO radicals will be more diffuse. The assumption
that the HO is more diffuse comes from the fact that HO radi-
cals have a longer time to diffuse as compared to the n-butane
because HO is added a longer period of time before the addi-
tion of n-butane. These two effects will reduce the possible
reactions between n-butane, butyl and the HO radicals. This
second possible process will also result in lower measured
reaction rates.
The data for the trapping of products demonstrated the
formation of a product made by the reaction of butyl radicals


92
and other species. The mass spectrum at 2.105 showed that the
product contained a hydrocarbon unit of mass 57, which is
butyl. The kinetic results demonstrated that HO reacts with
the butyl due to the fact that as the initial HO concentration
increased the rates also increased. It is also demonstrated
that butyl reacts with NO and N02 based in the results of the
apparent bimolecular rate constants at Very High HO. There is
still the need to study the possible reaction of NO and N02
with an alcohol in the trap. This will help to elucidate the
spectrum obtained for the mentioned peak in figure 11.2 at a
retention time of 2.105 minutes.
From the simulations of the flash photolysis systems
it can be seen that the influence of secondary reactions
should have been observed if a fixed time interval was used
for measuring the HO time decay. The reason for not observing
this influence in most of the experiments could be due to
various factors. First, the modeling results for the flash
photolysis systems demonstrated that the use of a calculation
method that calculates the pseudo-first-order rate constants
for a variable time results in lower values for the apparent
bimolecular rate constant and in a smaller difference between
the apparent bimolecular rate constant for two different
initial HO concentrations as shown in table 10. The variable
time scale method is typically used in the flash photolysis
systems that studied this reaction. ^1 1 ^ This method will


93
make it more difficult to observe the effect of secondary
reactions in the flash system as compared to the use of a fix
time interval in the flow method. This does not necessary
diminish the influence of secondary reactions in flash
photolysis systems. The reason for smaller than expected
apparent bimolecular rate constant at high HO results from the
fact that there is going to be a longer time of measurement of
HO decay at the low n-butane range, thus a longer time for
secondary reactions to occur. Therefore, the pseudo-first-
order rates at low n-butane are going to be bigger than in a
fixed time method. On the other hand, the measurement time at
high butane concentrations is going to be smaller in the
variable method as compared to the fixed time interval. This
smaller measurement time will reduce the influence of secon-
dary reactions in the high end of the n-butane range. There-
fore the pseudo-first-order rates at high butane are smaller
in the variable time method as compared to the fixed time
method. These two opposite effects will tend to account for a
smaller apparent bimolecular rate and smaller differences
between bimolecular rates for High and Low HO in Flash
Photolysis.
The second consideration is the fact that the simula-
tions performed for the flash photolysis systems were done
using a maximum value for the rate of the secondary reaction
of HO and butyl radicals. A lower value for this secondary


Full Text

PAGE 1

EFFECTS OF SECONDARY REACTIONS ON THE MEASUREMENTS OF THE RATE OF OH' RADICAL REACTION WITH N-BUTANE by Juan Luis Bonilla B.S., University of Puerto Rico, 1976 A thesis submitted to the Faculty of the Graduate School of the University of Colorado at Denver partial fulfillment of the requirements for the degree of Master of Science Department of Chemistry 1987

PAGE 2

Date. ____

PAGE 3

Bonilla, Juan Luis (M.S., Chemistry) Effects of Secondary Reactions on the Measurement of the Rate of HO Radical Reaction with n-Butane Thesis directed by Professor Larry G. Anderson Hydroxyl radicals are considered to play an important role in combustion and atmospheric chemistry. They are known to be one of the most reactive species in the atmosphere. The reactions of hydroxyl radicals with hydrocarbons in the troposphere are believed to be the key initiating step of the oxidation of hydrocarbon compounds. There are amounts of small chain hydrocarbons in the troposphere. The reaction of HO n-butane has been used as a representative alkane in models photochemical air pollution. This reaction has also been used as the reference reaction in the study of rate constants for other reactions using relative methods. The first step in the reaction of HO radicals with nbutane involve the abstraction of a hydrogen atom from nbutane. In a system of reactions in which HO radicals are present, the reaction of these HO radicals and the product butyl of the main reaction can be enhanced as the initial concentration of HO is increased. This secondary process is of some significance in the measurement of the main reaction and should be taken into consideration. The study of the reaction of n-butane and HO radicals showed that there 1s a significant influence of the secondary reaction to account for differences of up to 30% in the measurement of the bimolecular rate constant.

PAGE 4

iv Experiments performed at initial HO of approximately 2 x 10'1 particles/cm3 in an apparent -12 3 bimolecular rate constant of 2.23.OQ x 10 cm /molec-sec, whereas initial HO concentrations of 2 x 1012 and 8 to 16 x particles/cm3 gave apparent bimolecular rate constants of 2.85.05 x 10-12 and 3.07.15-3.13.23 x cm3/molec-sec respectively. Modeling of the experimental conditions for the results obtained helped to elucidate the approximate magnitude of the rate constant for the secondary the HO n-butane system, which ranges from 1.2 X to 5 x 10-10 The best value for the reaction of HO n--12 butane at room temperature is 2.'.25 x 10 cm3/particle-sec. The also demonstrated that the bimolecular rates at two different HO concentrations are going to be more nearly the same when the decay of HO followed a variable time method, which is the method most commonly used with flash photolysis systems. Trapping and analysis of products in the flow system for the reaction of HO and n-butane showed the formation of a product which contains butyl as its main component. -Kinetic data demonstrated that the butyl was reacting with HO to form the product. A clear analysis of the product was not possible due to the interference of NO or N02 in it. This interference was shown in the kinetic results and in the mass spectrum of the product.

PAGE 5

CONTENTS CHAPTER General considerations ............ Calculation of a bimolecular reaction rate........ 4 Experimental techniques........................... 6 Previous n-butane and hydroxyl reaction rate results ............... 9 Purpose of this research .... 6 2. EXPERIMENTAL 19 Flow tube components 19 Reactants .... 24 Detection of HO radicals.......................... 27 Experimental procedure............................ 29 3. COMPUTERIZED SYSTEM FOR STUDYING GAS PHASE KINETICS. 37 Advantages........................................ 37 Hardware interfacing.............................. Software description.............................. 49 4. 58 Experimental. 58 Error Analysis. 72 Discussion........................................ 88

PAGE 6

vi APPENDIX 1 ACTUATE PROGRAM..................................... 103 2. KINETIC RESULTS FOR HO AND n-BUTANE REACTIONS 137 3. REACTIONS OF SYSTEM TO BE SIMULATED ..

PAGE 7

vii TABLES Table 1. Previous studies of HO + n-butane at room temperature (295-302 K)............................. 10 2. Transducers used in gas flow kinetic experiments.... 39 3. Signal conditioner devices.......................... Summary of input/output modes....................... 5. Features of the Lab Master board.................... 47 6. Progamming of the Lab Master board.................. 50 7. Simulations with minimum k2 79 8. Effect of kl and k2 on model values for 9. Maximum k2 value and effect of k2 on apparen t bimolecular rate........................... 81 10. Flash Photolysis simulations........................ 87 11. Previous direct studies of HO and propane........... 97

PAGE 8

viii FIGURES Figure Schematic of discharge flow-resonance fluorescence appara t us. . 20 2. Detection area's six way fitting.................... 22 4. High current relay driver 43 5. Control of multiport valve.......................... 45 7. Titration curve 59 8. Pseudo-first-order decay 61 9. Relative signal decay of HO for High and Low HO concentrations................................... 64 10. Kinetic results for Low, High and Very High HO concentrations................................... 66 12. Modeling and experimental results.................. 83 13.1 Simulated concentration profile for Low HO......... 84 13.2 Simulated concentration profile for High HO........ 85

PAGE 9

CHAPTER 1 INTRODUCTION General Considerations Science is considered to be the state of knowledge derived from observation, study and experimentation carried on in order to determine the nature or principles of the subject being studied. The Scientist in all of us has the curiosity or drive to carryon the task and to obtain such knowledge in order to satisfy the very nature of its existence. Chemistry, as a scierice that studies composition and relationS of the most simple forms of existent bodies, strives for the absolute values of those principles that affect its own nature. Within the science of the chemical nature of things, kinetics is the study of the rates of the chemical reactions including all the factors that influence them and the mechanisms involved. Kinetics is concerned with the details of the process in which a system gets from one state to another and the time required for that transformation. The details include concentrations of initial reactants, product formation., temperature and pressure effects on rates of the processes studied. The kinetics of the reactions of hydroxyl radicals (HO) with different species have been largely studied in the gas phase. Its interest results from the fact that this radi-

PAGE 10

2 cal is present in considerable amounts in our atmosphere and in combustion systems. It is very reactive since it has an odd number of electrons. The reactivity and quantity of thi8 radical make it one of the most important species in our atmosphere and in combustion systems. As modern civilization continues to emit hydrocarbon pollutants into the atmosphere and use combustion systems for different purposes, it becomes more and more important to characterize the kinetics of HO radicals with hydrocarbons. Hydrocarbons are involved in chain reactions that oxidize NO to N02 in the atmosphere. These reactions are initiated by the reactions between HO radlcals and the hydrocarbons.. Considering the concentrations of hydrocarbons in the atmosphere, it is estimated that alkanes can account for up to 33% in non-rural(2) and 20-25% in rural areas(33) of the HO reaction8 with hydrocarbons. The other reactions of HO and hydrocarbons involve the reactions with olefins, and aromatiCS. The principal mechanism of consumption in lean and stoichiometric alkane/air flames is considered to be the abstraction of hydrogen atoms from alkanes by HO radicals. (3) This reaction can be described by: (1) R-H + HO + R + H 0 2 in which the hydroxyl radical abstracts a hydrogen atom from

PAGE 11

3 the alkane reactant forming an alkyl radical and water. The sequence of reactions that could follow from this main reaction are numerous making the study of the rate of this main reaction Ruceptible to undesirable effects if not properly considered. The undesirable effects could result from a com-petitive reaction between the very reactive HO radical and reactive products that are formed from the main reaction. If these secondary reactions are fast, they will tend to affect the measurement of the first reaction. The reaction of HO radicals with alkyl radicals can be of great importance in the study of the first reaction if a sufficient concentrati0n of alkyl radicals are formed in the system of study. (2) R + HO Due to the importance of the reactions of HO with alkanes it necessary that highly measurements be done for the rates of these reactions in order for these results to be used in modeling and estimates of complex systerns like the atmosphere or combustion. The rate of the secon-dary reaction between alkyl radicals and HO radicals have never been studied. On the other hand the rate of reaction between alkanes and HO radicals have been studied by numerous researchers using different experimental techniques. The possible effect of the secondary reaction on the measurements

PAGE 12

of the reaction of and HO radicals obtained by these researchers was considered to be negligible. Some secondary effects for alkanes and HO radical reactions have been ob-served at extreme experimental conditions used by few of researchers In addltion to an accurate knowledge of the rate of the primary reaction, it necessary to measure the rate of the reaction of alkyl and HO radicals. Calculation of a Bimolecular Reaction Rate The study of bimolecular reactions is generally done under pseudo-first-order conditions in which one of the reac-tants is in excess. This will ensure that the reactant in lesser amount will react with the one that is at higher con-centrations and not with itself. Also the higher concentration reactant will stay relatively unchanged during the reaction. The kinetics of this system is followed by measuring the change in concentration of the reactant in smaller quantity with reaction time. The rate of a bimolecular reaction can be expressed as; Rate of Reaction k [A][B] where k is the reaction rate constant of species A reacting .with B and [A] is the concentration of species A and [B] is the concentration of species B. The rate of the reaction can

PAGE 13

5 be considered to be the change of concentration of one of the reactants in time. Considering the change of concentration of species A in time we can write a differential equation in which: -d[A]/dt k [A][B] and if B is at higher concentration than A, we can consider the concentration of species B to be constant during the reaction. The concentration of B should be at least times that of species A for pseudo-first-order conditions to eXlst. If this is the case, then the con-centration of species B can be included as part of the rate constant, making this a pseudo-first-order reaction for A with a pseudo-first-order rate constant of kl. Rearrangement and integration of the previous equation, results in: which is a linear equation. Measurements of the variable t (time) and the logarithm of the concentration of A will enable

PAGE 14

6 us to calculate the pseudo-first-order rate constant for the reaction a constant concentration of a. The ratio of kl/[a] will give the value for the bimolecular rate constant for the reaction of A + a, or the slope of a plot of kl vs [a] will also give us the same results. The second method is preferred it will correct for reactants to the walls as well losses due to other reactions. If reactant A is also reacting with the or itself at a significant rate, then the observed rate constants will be higher than that for the reaction of A with a. This higher pseudo-first-rate constant will be independent of the amount of species a that is used. Its higher value will result in a higher calculated bimolecular rate constant when the ratio of kl/[a] is used. If higher constant is independent of the amount of species a, then it will appear as part of all the values of the pseudo-first-order rate constant at the different concentrations of a. Therefore the slope of the pseudo-firstorder rate constant versus the concentration of a will not include the influence of these losses of species A. Experimental Technigues In the study of the kinetics of the reactions of HO + alkanes, the hydroxyl radical is usually selected as the reactant to be monitored, since it can be easily measured by

PAGE 15

7 fluorescence in the 308 nm wavelength region. The study of the kinetics of HO radicals with alkanes can be separated into absolute or relative techniques. The absolute techniques can be separated into discharge flow (DF) or flash-photolysis (FP) methods. Also included in these absolute techniques are pulse radiolysis and modulation-phase shift, which have been infrequently used for small chain alkane studies. The discharge flow method will be discussed in detail in the experimental section. In general it can be considered as a steady state method in which the concentration of the species remains constant in the detection area. The reaction time of the species is changed by varying the location of addition of one of the reactants in the flow tube system. The reactive HO species are produced by passing the appropriate precursors through an electrical discharge prior to entering the main flow area. The flash-photolysis method the use of a reaction cell in which the reactants are added to the precursors (H20, H 2 0 2 HN03 or other mixtures) of HO radicals. HO radicals are formed by the photolysis of these precursors. The radiation for the photolysis is obtained by the use of flash lamps or pulsed lasers, which are separated from the react ibn cell by the use of transmitting windows (LiF, MgF2 CaF2 sapphire and Suprasil) in the vacuum uitraviolet region. The loss of HO radicals can be monitored as a function of time

PAGE 16

8 after photolysis of the precursors by the use of one of the following detection methods. The detection of HO can be accomplished by resonance absorption (RA), resonance fluorescence (RF), electron paramagnetic resonance (EPR), mass spectroscopy (MS), laser magnetic resonance (LMR), or laser induced fluorescence (LIF). The concentration of HO radicals used varies depending on the detection method used. For resonance fluorescence, laser-induced flourescence and laser magnetic resonance the concentrations are typically low, molecule/cm3 Resonance absorption is usually used for higher concentration 12 14 3 detection ranges, 10. -10. molecules/cm. Electron paramag-netic resonance detects ranges from to molecules/cm3 12 13 3 and mass spectrometry from 10 to 10 molecules/cm. Relative measurement techniques are accomplished by monitoring the relative rate of disappearance of two or more organic compounds in the presence of HO radicals in a chemical system. Using a known rate constant for the reaction of HO and one of the organic compounds the absolute rate constant for the reaction of HO and the other compound can be calculated. Relative measurements are made in Teflon chambers coupled to Gas Chromatograph, FT-IR absorption spectrometer or differen-tial optical absorption spectrometer for detection of the organic reactants.

PAGE 17

Previous n-butane and Hydroxyl Reaction Rate Results In the study of hydroxyl radical and short chain alkanes reactions, the scatter of the room temperature data is varied. The biggest occurs for n-butane and HO in which there is a factor of almost 1.9 in the data (See table 1.) by different researchers(5). These dis-crepancies in the evaluation of the rate constants suggest the possible complication of undesirable effects occurring during the experimental process. Such undesirable effects include the effects of secondary reactions between HO radicals and butyl radicals. The discrepancies may also be due to the use of erroneous calibration factors and other systematic experi-mental errors.

PAGE 18

10 Table 1. Previous studies of HO + n-butane at room temperature (295-302 K). 3 (10 cm /molec-s): Method 2;57.12 :Flash Photolysis:Kinetic Spectroscopy 1I 1 2.35.35 2.72.27 2.67.22 2.23.23 2.1I1.10 2.30.3 Relative 2.9.7 2.52O.25 2.71O.32 :Discharge Flow :Mass Spectroscopy :Flash Photolysis:Resonance Fluorescence :Pulse Radiolysis:Resonance Absorption :Flash Photolysis:Resonance Fluorescence :Flash Photolysis:Resonance Absorption :Discharge Flow :Resonance Fluorescence :Laser Photolysis:Laser Induced Fluorescence :Laser Photolysis-:Laser Induced Fluorescence technigue: Reference reaction HO CO HO CO HO propene HO n-pentane HO + propene Reference Greiner 1970 Morris & Niki 1971 Stuhl 1973 Gordon & Mulac 1975 Perry, Atkinson & Pitts 1976 Paraskevopoul')s & Nip 1980 Anders.on Stephenf\ 1982 Tully & Droege 1985 Schmidt, Zhu, Becker & Fink 1985 Gorse & Volman 1974 Campbell et al. 1975 Atkinson et al. 1981 Behnke et ale 1984 Atkinson & Aschmann 1984

PAGE 19

11 N. R. Greiner was the first to study the reaction of HO with a series of alkanes by a direct method. The purpose of his study was the characterization of the rates of abstraction for primary, secondary and tertiary hydrogen found hydrocarbons, and to enable further prediction of reaction rates for more complex alkanes. In this study, the effect of a side reaction for alkanes larger than CH4 was observed. The rates show a 15% decrease as the alkane concentration is ThO bh i increased by tenfold at a given temperature IS e av or results from the competitive reaction between HO and alkyl radicals at an assumed rate of 1.66 x 10-10 cm3molec1sec-1 The data were corrected taking into consideration this effect by the use of a computer simulation that included the main reaction, the proposed competitive reaction, an assumed initial HO concentration of 3.0 x 1013 molec./cm 3 produced in each flash, and the experimental concentration of the alkane used. The uncorrected experimental results for n-butane at room temperature of 298 K and a concentration of the alkane of 15 3 4.35 x 10 moleculeslcm resulted in a rate constant of 3.12 -12 x 10 cm Imolec.-sec. This value was corrected by a factor of 22% using the simulation to obtain the rate of the reaction -12 3 of HO and n-butane of 2.57 x 10 cm /molec-sec. The experimental results of Morris and Niki are ap-proximately 40% higher than those of other Their experimental results were obtained using a high initial HO

PAGE 20

12 concentration (1 x mOlecules/cm3 ) relative to the alkaneg (5 x 1011 studied(7). Using a discharge flow system coupled to a time-of-flight mass spectrometer, they observed reactant disappearance for the reactions of olefins and other compounds (including n-butane) with HO. These reac-tions were studied under pseudo-first-order conditions (excess HO) and the hydrocarbons were the species monitored. They considered the effect of secondary reactions to be insig-nificant due to the large excess of HO radicals present, making it unlikely for any intermediate like the alkyl radical to compete with of HO and hydrocarbons. They estimated an error of less than 5% for the effects of the reaction of oxygen (which are products from the reaction of HO + HO) with the hydrocarbons studied. They also con-sidered the reactions of hydrocarbons with H atoms, which are used in the formation of their HO radicals, to be too slow to have an effect on their results. Stuhl working with alkane concentration ranges of 0.5 to 2.5 x 10 molecules/cm, monitored HO concentration by resonance fluorescence. HO was made by flash photolysis of a (8) H 2 0 (0.25% in He) mixture at a total pressure of 2.67 kPa The initial HO concentration used was approximately 3 x 10 mOlecules/cm3 The effect of secondary reactions in this study was considered to be unlikely. Experiments performed at two different concentrations and at an initial HO con-

PAGE 21

12 centration of approximately 2 x 10 showed no effect on the pseudo-first-order rate constant results as compared to experiments done at an HO concentration of 3 x 1011 Gordon and Mulac did pulse radiolysis studies of H 2 0 (1.39-100%)/Ar mixtures at total pressures of 96.0-101.3 kPa in the presence of n-butane at 66.7-328 Pa pressure. The 13 hydroxyl radical was monitored by U.V. absorption spectroscopy at 308.7 nm. The run at 298 K had a diluted mixture of H 2 0 and Ar(9). It is very likely that possible effects of secondary reactions could have occured due to the relative high con-centration of HO needed to be monitored by absorption spectroscopy. Perry, Atkinson and Pitts worked in an n-butane con-13 14 3 centration range of 1 x 10 to 1.4 x 10. molecules/cm, using a flash photolysis-resonance fluorescence method. In their stUdy(10) there is no indication of the HO concentration used. Because pseudo-first-order rate constants for the HO radicals were determined, it is assumed that n-butane con-centrations were at least forty times bigger than HO. They concluded that secondary reactions were negligible because a variation of a factor of two in the flash energy showed no effect in the rate constants. The variation on the flash energy should have doubled the photolysis of water which produced the initial HO concentration. HO radical concentra-tion was monitored over at least three half-lives.

PAGE 22

14 The study of and Nip .. using a flash photolysis-resonance absorption technique, also indi-cates that there was no significant interference from sec ondary reactions. They varied the concentration of HO radicals by a factor of 5 to 10 by varying both the pressure of H 2 0 by a factor of 10, and the flash energy (30 to 350 Joules) needed to photolyze the H 20. These did not change the measured rates within the experimental errors. The concentration of n-butane varied from 0.6 to 9 x 1015 molecules/cm3 and HO was estimated at 0.6 to 5 x 1013 molecules/cm 3 Anderson and Stephens used a discharge flow-resonance fluorescence system(12) in which the HO concentration of 1-3 x 1011 molecules/cm3 was estimated based on titrations of H atoms by N02 Their n-butane concentration varied from 1.17 to 8.18 x molecules/cm3 In some experiments done at 250 K in which they increased the HO concentration by a factor of 4-5 times, they were able to observe a non-linear decay in the HO fluorescence when compared to experiments done with low initial HO concentrations. This curvature appeared to be less important at higher temperatures in their experiments and was considered to be the result of secondary reactions, mainly between HO and butyl radicals. The rates observed for these particular experiments in which HO initially was high were consistent with model calculations done in a further stUd/13) There the rate for the secondary process

PAGE 23

15 was estimated at 2.0 x 10-10 cm3/molecule-sec. All of their experiments were done at low initial HO concentration where the effect of secondary reactions was considered to be least. Tully and Droege determined site-specific rate coeffi-cients for H-and D-atom abstraction by HO from methyl and methylene sites in n-butane using the laser photolysis-laser lnduced fluorescence technlQue. Thelr typical reaction mixtures concentrations in molecules cm-3 consisted of :N2 0 1.3 x 1014, H 2 0 3.3 x 1015, n-butanes = 0-3.8 x 1014 and 18 He = 6.4-13.2 x 10 They varied the photolysis flux density 2 over the range of 1-4 mJ/cm ,which would increase their initial HO concentrations, and observed no significant changes in the kinetics. At 599 K they were able to observe secondary-reaction kinetic interference for n-butane and HO reactions. HO concentration was considered to be 7.5 x 1010 molecules cm-3 and time dependence of HO concentration in the reaction with n-butane was taken over at least a factor-of-ten varia-tion from its initial concentration. Schmidt et al. performed experiments using excimer laser photolysis of the precursors of HO radicals. The HO concentration decay as a function of time was monitored using laser induced fluorescence. (15) The experiments were performed at 1 atmosphere total pressure, but there is no indication of the approximate HO or n-butane concentration used. The experi-ments for the reaction of HO and n-butane, propane, and ethane

PAGE 24

were done to check the method and apparatus. The study was mainly directed at the study of the effects of oxygen on the reaction of HO and C 2 H 2 and with some sulfur containing compounds. Purpose of this Research 16 As we can see the effects of secondary reactions were considered to be negligible in the previous n-butane and HO radical reaction studies. A variation of the initial HO centration was used to see if there was any effect of secon-dary processes. There was no effect by the use of a change in the initial HO concentration with the exception of the studies of Greiner(6), Anderson(12), and TUlly(14) in which the change 1 (12) of the initial HO concentration was done at very ow or (14) high temperature by changing the ratio of n-butane concentration to that of HO(6). There seems to be the indication that secondary tions exist in the reaction of n-butane and HO radicals. These could only be apparent in these previous studies if the rate of reaction butyl and HO radicals was significantly high in order to show its effect. is important to evaluate the rate of the reaction of butyl and HO radical to be able to use its value in complicated modeling schemes and to correct for the influence of in the study of the reaction of n-butane and HO radicals. It is necessary to evaluate the conditions at

PAGE 25

17 which the effect of secondary reactions are apparent, so the effect of systematic errors due to these are minimum. The lack of effect of secondary reactions found by some researchers in their studies tend to suggest that there was not enough of a change in the initial HO concentration in order to see the effect. This does not necessarily means that the effect of the secondary reactions was not present in theirrstudies or was negligible. The purpose of this research is to investigate the effect of the secondary reactions of HO radicals with butyl radicals on the measurement of the rate for the reaction of HO with n-butane. This is performed by varying the initial concentrations of HO radicals in the different experiments which should give different apparent rate results for the bimolecular reaction of n-butane and HO radicals, if secondary reactions are present. A computer simulation and analysis of the kinetic data will serve to place meaningful limits on the rate of the reaction of HO radicals butyl radicals. Using the kinetic results obtained for the reactions of HO with butane and HO with n-butyl radicals, it is intended to model the conditions used by other researchers to see if any secondary reactions could have been affecting their measurements. An analysis of the products formed will be performed to help elucidate the mechanism of HO radical and n-butyl radical reactions. This will be performed by trapping products

PAGE 26

directly from the flow and analyzing these using gas chromatographic-mass spectral methods.

PAGE 27

CHAPTER 2 EXPERIMENTAL Flow Tube Components The kinetics experiments were performed using a charge flow-resonance fluorescence apparatus (fig 1). The flow system was built using 1.25 in. i.d. copper pipes and fittings through which the gases flowed at a constant rate. The pipes lengths varied depending on the needed at the different positions for introduction of gases, readout devices, and detection. The pipe and fittings were soldered together into five different sections. These permit the system to be disassembled for maintenance purposes.

PAGE 28

20 /lICP,OIlAv[ CAYITY PI1T VAl.I'E 1 DILUEIIT H IiOOD'S Figure 1. Schematic of discharge flow-resonance fluorescence apparatus. THER.'IOCOUPLE The first section consists of three gas inlets. The first inlet is used for the introduction of the carrier or diluent gas (Helium) at a distance of 92.4 cm from the detec-tion area. The second inlet is used for the introduction of hydrogen atoms created by passing helium gas that contains small traces of hydrogen impurities, hydrogen that is absorbed from the teflon tubing or a known mixture of hydrogen/helium through a 11 mm i.d. glass tubing coupled to a microwave discharge cavity prior to entering the flow area. The microwave discharge system consist of an Opthos microwave

PAGE 29

21 power supply and cavity. The forward power was maintained at watts and the reflected power at 0-2 watts. Hydrogen atoms were introduced at a distance of 82.6 cm from the detec-tion area. The third inlet of the first section is used for the introduction of a known mixture of N02 and helium at a distance of 75.2 cm from the detection area. The N02 will rapidly (1.3 x 10-10 react with hydrogen atoms to make HO radicals and NO. The second section has a fitting through which a copper constantan thermocouple can be introduced to measure temperature. It also has a valve at a distance of ap-proximately 52 cm." before the detection area. Through the valve one can measure the pressure by a 10 torr range Baratron capacitance manometer (pressure gauge). This pressure gauge also measures pressures after the detection area through means of another valve. This configuration allows pressure measure-ments at two positions before or after the reaction area, thu5 quantifying for pressure drops along the axis of the flow system. The two valves were connected to the pressure gauge by means of in. teflon tubing. The third section contains eight different inlet ports. These are made of eight 1 in. to in. NPT Tee fit-tings and are separated a distance of 5 cm between the centerse The first port is from the detection area. The eighth port is a distance of 10 cm from the detection area.

PAGE 30

22 The inlet ports allow hydrocarbons to be introduced through in. stainless s.teel tubing which is inserted to the center of the main flow tube. The port selection is made by using a Carle model 2026 multiport valve, to which the eight different outlet tubes are connected. The inlet to the multi-port valve is 'connected to the source of the hydrocarbon. A Carle Valve Actuator (Model 4203) changes the selection of the outlet port and is controlled by means of an IBM personal computer. Details of the computer interfacing and control will be discussed in Chapter 3. The fourth section is the detection area, and was made with two 1 in. Tee fittings. These two fittings were modified by making apertures in their walls and welding them to each other in such a way as to form a six way fitting (Fig 2). Figure 2. Detection area's six way fitting.

PAGE 31

23 Two of the apertures enable the main flow of gases to pass through the detection area. A quartz window was connected with epoxy to a third aperture through which the light from the fluorescence lamp enters the middle of the gas flow. Included in this aperture is a 1 in. to 1 cm. reducer at approximately in. from the window and next to the main flow tube. This cm. hole reduced the scattered light from the outside. The fourth and fifth apertures had Wood's connected to them. These were a180 made of copper fitting8 soldered together. One of these Wood's horns was in line with the light coming from the light source. It will trap the light after passing through the gas flow. The other was perpendicular to the fluorescence excitation axis and it further helped to trap the scattered light. Welded to the sixth aperture is a in. thick and 6 in. diameter round copper plate with a cm. hole in the middle. This serves as the base for the photomultiplier which sits perpendicular to the light source entering the flow system. A groove made in the plate allows a in. diameter quartz window to be mounted on the plate. The fifth section of the flow system contains a valve which allows pressure measurements immediately after the detection area. It also contains fittings to enable the introduction of another thermocouple. This section Is connected to the pump by a in. i.d. gum rubber hose with an inner

PAGE 32

24 steel spring reinforcer. A glass trap was inserted ap-proximately halfway between this section and the pump's fit-tings. The walls of the flow tube were coated with a halo carbon wax (Series No. 15-00, Halocarbon Products Corp., Hacken-sack, New Jersey) to reduce the wall loss of HO radicals in the system. The five sections of the flow tube were assembled and the interconnections were sealed using a sealing tape from 3M and a Dux Seal plaster compound from Manville Corp. placed over the sealing tape. This proved to be an effective way to maintain the low pressure conditions needed. Two large bellows sealed valves were inserted after the fifth section of the flow tube and before the pump. The glass trap is located between these two valves. The valves were used to control velocity of the flow. Reactants The helium, hydrogen/helium (8.5%), nitrogen dioxide/helium and n-butane (99.5%) were used without further purification. Gas chromatographic-mass spectometric analyses of n-butane indicated that the alkane was within the 99.5% purity specified by the manufacturer. Due to the nature of the experiments performed with low HO, the nitrogen dioxide/helium gas was diluted by a factor of almost times. These dilutions were performed using a commercial propane

PAGE 33

25 tank, an absolute 100 in. of water range pressure gauge and a 30.5 psi relative to atmospheric pressure gauge. The hydrogen source for these low HO experiments was obtained by using helium instead of a hydrogen helium mixture. This helium had small traces of hydrogen containing impurities which were either part of the helium gas, collected from the in. plastic tubing or both. The gases were connected from the gas cylinders to the flow controllers using in. polyethylene tubing with the exception of the N02 cylinder which requires in. teflon tubing due to the reactive nature of NOt with the polyethylene tubing. All gas flows were controlled and measured by Unit Instruments (Model UFC-1000) electronic mass flow controllers. These were calibrated using soap bubble flowmeters. The flow controller used for n-butane was also calibrated by measuring the time required for the pressure to increase by a few torr when filling a container of known volume. The results for the different volumes flowed versus the set voltages gives a slope for the relation between these two parameters. The results of the slope for the calibration of the flow controller using the pressure increase technique were 12% lower than those found using the soap bubble method. A significant fluctuation using the pressure increase technique was noted in the flow region from 4-5 volts during the calibration runs, which gave higher flows than expected. This suggested an instability of

PAGE 34

26 the sensor in the flow controller for n-butane at high flows. The calibration results used in the experiments for the setting of flows were for the low voltage (0-3 volts) range of the calibration of the pressure increase technique. The fluctuation was not observed during the calibration using the soap bubble flowmeter for n-butane which could have resulted in the higher calibration results for the soap bubble method compared to the pressure increase method. This fluctuation in the sensor could be the result of the low thermal conductivity and high molecular weight of n-butane. A higher negative intercept resulted in the soap bubble method, which suggests a loss of n-butane possibly by reaction with the soap bubbles. The use of a calibration method using a pressure increase in a known volume for n-butane was performed to evaluate the possibility of n-butane reacting with the soap of the bubbles. The other flow controllers were calibrated for helium using the soap bubble flowmeter, reproducible results were obtained The flow controllers were then connected to the flow system using in. Teflon tubing. The flow controllers were controlled by the IBM personal computer and power was furnished by a power supply with +15, -15 and +5 volts lines. These flow controllers feature a shut-off valve which stops the flow of gas when required.

PAGE 35

27 Detection of HO The components for the detection of HO radicals consist of a lamp or light source for excitation and a photomultiplier for detection of fluorescence. The light source was obtained by passing a mixture of H 2 0 and helium through 11 mm. i.d. glass tubing coupled to a microwave cavity. The mixture of H 2 0 and helium was made by passing helium gas through a vessel which contained liquid H 20. The microwave power supply was of the same type as that used for the generation of hydrogen atoms. The H 2 0 is dissociated as it passes through the microwave cavity into H and electronically excited HO, thus emitting light with a strong component near 310 nm. wavelength. This is used for fluorescence excitation of HO radicals inside the flow system. This light source directed through one of the apertures of the six way fitting in the detection area of the flow system by means of two quartz lenses. These lenses have been set up to obtain the focal point of the light source near the center of the flow system, thus reducing the amount of scattered light that could otherwise be detected by the photomultiplier (Fig 3).

PAGE 36

28 Figure 3. Optical set up. The fluorescence of the HO radicals was detected using a Thorn EMI photomultiplier for photometry and spectrophotometry applications. The necessary voltage for the tube was furnished with an independent 15 volts power supply which was then amplified toa low current and high voltage (800 volts) line connected via coaxial cable to the photomul-tiplierhousing. A narrow band interference filter was placed between the previously Quartz window in the sixth aperture of the six-way fitting and the photomultiplier. This filter transmitted only light in the 280-325 nm. wavelength region which allowed the fluorescence of the HO radicals to be transmitted to the photomultiplier.

PAGE 37

29 The photomultiplier output was connected via a BNC cable to a EMI Gencom model AD-100 amplifier discriminator which reduced the noise and amplified the electron bursts into regular pulses. The amplifier discriminator was then connected via a 9 pin connector cable to a EMI Gencom model C-10 photon counter which displayed the number of photons per set time interval. The photon counter has different interfacing capabilities which output the photon counts either through digital, analog or TTL outputs. The last was used to interface to the IBM Personal Computer via a BNC cable. The fluorescence signal was determined after subtracting the background light levels. The background was measured by the photon counts observed when the HO was not present (during the time the N02 reactant was turned off). These counts were measured before and after each signal reading. Experimental Procedure The experiments were performed after conditioning of the flow system with only HO radicals in helium flowing through the flow tube for approximately 30 to 60 minutes. The HO concentration was determined by two techniques. The first one involved the titration of H atoms produced in the microwave discharge by the addition of a measured amount of N02 while the HO fluor.escence was monitored. The hydrogen atoms will react with N02 with a rate constant of 1.3 x 10-10

PAGE 38

30 cm3/particles-sec, which is considered to be a fast reaction. HO This reaction produces one HO radical for each N02 molecule added to the system. The signal observed is related to the amount of HO produced. There will be a proportional increase in the signal as the N02 is added to the After all the hydrogen atoms produced in the discharge react with the added N02 then any further addition of N02 will not produce any more HO radicals, therefore the signal will stay the same. These two trends observed in the signal the N02 molecules are added will produce a titration curve. The endpoint of this curve occurs at the pOint at which the curve changes slopes. From the determination of the an equivalence between the N02 concentration and H atom concentration was determined. The second technique was based upon a calibration curve for HO fluorescence signal versus N02 concentration. This curve was generated by monitoring the HO fluor"escence signal resulting from the addition of a measured amount of N02 to an excess of H atoms. The HO concentration is assumed to be equal to the concentration of added N02 and proportional to the signal. Using this calibration curve, the HO concentration can be determined by the amount of signal observed. The two techniques were in reasonable agreement, within a factor of

PAGE 39

31 two for the different HO concentrations used. A titration curve was generated daily before the kinetics runs or when the HO concentration was changed and sometimes at the end of the day. The signal levels were measured before and after the kinetic runs for a given HO concentration. A further note should be made to the observa-tion that the efficiency of the production of H atoms in the microwave discharge is inversely proportional to the amount of gas flowing through the discharge. As the flow of hydrogen/helium passing through the discharge increases then a smaller hydrogen atom production efficiency in the hydrogen/helium mixture results from the experimental titra-tions. This could be seen from the results of the titrations of different flows of hydrogen/helium through the discharge, in which a flow of 300 sccm showed a hydrogen atom concentra--4 3 tion of 1.56 x 10 atoms/cm, a flow of 200 sccm showed a hydrogen atom concentration of 1.65 x 10-4 atoms/cm3 a flow of 100 sccm resulted in a hydrogen atom concentration of 1.77 x 10-4 atoms/cm3 and a flow of 50 sccm resulted in a concentration of 2.46 x 10-4 atoms/cm3 If the efficiency of the discharge would have been the same in all flows it would have resulted in the observation of the same concentration of hydrogen atoms in the mixture, due to the fact that it is the same hydrogen/helium gas source. The previous could be explained from the fact that as the flow is reduced, so is the

PAGE 40

32 velocity in the discharge area. A slower velocity of gas in the discharge area will increase the residence time of the hydrogen molecules in it and subsequently increase efficiency of the production of hydrogen atoms by the discharge. Three different HO concentrations were used in the experiments. The first one is described as Low HO concentra-11 -3 h tion and it ranges from 1 to 3 x 10 partlc es-cm ,t e second one is described as High HO and ranges from 1 to 3 x 10 12 particles-cm-3 and the third HO concentration is described as Very High HO and it ranged from 8 to 16 x 1012 particles-cm-3 The approximate HO fluorescence signal levels for these three concentrations were respectively 2 X 10 2 X 10 5 and 1 X 106 counts per second. The standard deviation for the signal at low HO was less than 2%, and for high and very high HO was less than 1%. These errors in the signal are without any n-butane added. During the kinetics runs the standard deviation of a particular data point increases as the relative signal decreases, especially at low HO concentrations and high n-butane. These standard deviations were used in the weighted-least-squares analysis of the kinetic runs, and the errors in the pseudo-first-order rate constants were usually less than 10% for a particular n-butane concentration. The background counts were from 1.5 to 3.0 x 103 counts per second, with a standard devi"ation of less than 1%. The range for the background counts depended on factors which

PAGE 41

33 included: the fluorescence lamp stability, any change in the adjustment of the lenses, the conditioning of the flow tube, the cleanliness of the window through which the light entered the system, the concentration used and others. The background was basically stable throughout a day of experiments, with a total drift of a few (1-4) thousand The difference between the background before and after any data point was less than 3%, which was controlled by the computer program. The variation in the background from day to day was due to the factors mentioned. The signal levels for a particular con-centration were qui te reproducible after adjusting for background levels. Wall loss rates determined from the intercepts of the plots of pseudo-first-order rate constants versus nbutane were generally less than 10 s ."A study was conducted in which the point of N02 injection was varied and the slope of the In (HO fluorescence) vs. wall exposure time was measured. This resulted in a wall loss rate for HO of 6 s The use of fixed injector ports eliminates the variation in surface area that is otherwise encountered in other discharge flow systems that use movable injectors. The rate of reaction for a particular concentration of n-butane and HO was measured by monitoring the signal levels after introduction of the n-butane through each reaction port (i.e. different reaction times). The concentration range for n-butane was varied from 1 to 9 x 1013 molecules/cm3 and

PAGE 42

34 pseudo-first-order for six or more different concentrations of n-butane were measured for a particular HO concentration. Repetition runs of the same concentration of n-butane were done at times to verify reproducibility. Daily experiments were performed for both Low HO and High HO to account for any systematic fluctuations. At least ten readings for each of the transducers were done for each particular data point. The measurements of the transducers included voltage readings for the pressure gauge, total counts of the fluorescence signal, voltage reading for each flow setting in the flow controllers and voltage reading for the position of the reaction port. The set of measurements for each transducer" was averaged and if the standard deviation of the total photon counts was more than 1% of the counts reading a new set of measurements for all transducers was taken. The background photon count levels were also monitored before and after each signal reading. If these two were different by more than 3% another new set of data was collected. Approximately 30 seconds were allowed before readings were done for a particular reaction port in order to allow for conditioning of nbutane at the new position and more than one minute elapsed before readings were done when a different n-butane concentration was used. The control and monitoring of the transducers and experimental parameters will be discussed in the computerized system section of Chapter 3.

PAGE 43

35 Measurements of bimolecular rate constants for the different HO concentrations were reproducible from day to day. The experiments were performed at room temperature which fluctuated between 297 to 301 K. Initial problems with the thermocouple amplifier circuitry, which was susceptible to the high voltage discharge of the Tesla colI used to turn on the microwave cavities, inhibited the use of these for the measurement of in the flow system. Although most of the experiments were performed at 1 torr of pressure, some were performed at 4 torr to try to determine if there was any pressure dependence. The trapping of reaction products was performed under conditions which would maximize yields of the products of interest, mainly Very High HO concentrations. At these concentrations we would expect to have more HO and butyl radical reactions, thus more butene or butanol formation as possible products. The n-butane concentration used was in the low range studied (1 x 1013 mOlecules/cm 3). The system was stabilized to these conditions and the trap was submerged in liquid nitrogen for approximately four to five hours. At the conclusion of product trapping the glass trap was then pressurized to room pressure and the bath was changed to a bath of CCl4 and dry ice. The contents of the trap were then Injected to a Hewlett Packard Gas Chromatograph-Mass Spectrometer (Models 5890A, 5970) for analysis. The head pressure in the GC-MS was kept at

PAGE 44

36 6 psi and the temperature in the capillary column was maintained at 341 K. Previous runs with n-butane, sec-butanol, trans-2-butene and air were used to measure the corresponding retention times.

PAGE 45

CHAPTER 3 COMPUTERIZED SYSTEM FOR STUDYING GAS PHASE KINETICS Advantages The increased power of personal computers (PC) and the lower cost and better performance of data acquisition equip-ment facilitates computer control of experimental conditions, data acquisition, manipulation, and storage. A PC interfaced with experimental transducers enhances the ability to perform experiments and make measurements. The advantages for interfacing are numerous. First the control of initial and ongoing experimental conditions and monitoring of these conditions can be relegated to an electromechanical servant. This automation facilitates the tasks involving large numbers of repetitive operations. The continual rieed to change experimental conditions and the large amounts of data to be monitored can be successfully automated. Secondly the quantity per unit time and quality of the data acquired are increased. The data acquisition equipmerit has the capability to access and read numerous devices in a short time as compared to a human investigator. This capability will markedly increase the amount of information obtained from each experiment. This quantity of data and the ability to determine statistical significance of data at the time of acquisition helps to discriminate data of inadequate quality.

PAGE 46

38 Third, the computational power of a computer gives the searcher a flexible tool to adjust the way in which the information retrieved from the experiments will be analyzed. Fourth, the use of additional accessories like printing, plotting and disk storage devices facilitates the display and storage of data obtained. Graphic displays can be presented for further evaluation of experimental results. Fifth, and not least, is the convenience of time savings gjven to the researcher which can be used to immediately study the results obtained and design new experiments. Computerized systems to monitor and control experiments are commonly found in classical laboratory equipment, for example Gas Chromatographs, Mass Spectrometers, Nuclear Magnetic Resonance instruments and others. The discharge flow resonance fluorescence technique is one in which the instrumentation is custom designed to meet the research needs. The use of a computer system with such non-traditional experimental equipment requires that a series of signal modifiers be incorporated in order to communicate with the experimental transducers. This process is known as hardware interfacing of the computer with the experimental transducers. It is also necessary to write the instructions or program that will incorporate all the needed algorithms. This is the software development necessary for the computer to be able to perform the. experiment.

PAGE 47

39 Hardware Interfacing The interfacing of the computer and experimental equipment requires that the transducers used in the experi-ments generate a signal or will respond to a signal generated by the computer. The signal is considered tQ be an electrical current at a given voltage level or a pulse that can be counted. For the transducers used in the flow system we can summarize the signals generated or used by those listed in the following table 2 Table 2. Transducers used in gas flow kinetic experiments. TRANSDUCER PHOTOMULTIPLIER CAPACITANCE MANOMETER THERMOCOUPLE TYPE MASS FLOW CONTROLLER MULTIPORT VALVE ACTUATOR : SIGNAL : PULSES OUTPUT :0-10 VOLTS ANALOG OUTPUT :millivolts ANALOG OUTPUT :0-5 VOLTS ANALOG INPUT :0-5 VOLTS ANALOG OUTPUT :CONTACT CLOSURE CIRCUIT :SWITCH OF 1'5 VOLTS AC ************t***************************************** The photomultiplier generates pulses or bursts of electrons. The number of pulses is related to the amount of light that hits the photocathode of the photomultiplier tube. The signal of interest is due mainly to fluorescence of HO radicals in the system. and scattered light from the source. These pulses are bigger than 1 millivolt. The capacitance manometer for pressure measurements consists of a flexible metal diaphragm placed between two

PAGE 48

40 fixed electrodes. When different pressures are applied to each side of the diaphragm it causes a deflection which results l.n a change in capacitance between the electrodes, thus a change in output voltage. One side of the diaphragm is vacuum sealed to less than 10-7 torr and the other side receives the pres-sure which causes the deflection. The output voltage is linear pressure8 from 0 to 10 torr in the flow tube. The thermocouples transmit a different millivolt signal related to the difference in thermal conductivity between an iron and a constantan metal junction. The dif-ference for the thermal conductivity in the junction is re-lated to the temperature of the The thermocouples -output low voltage signals. The mass flow controllers are controlled with a 0 to 5 volt signal. This input command signal into the flow control-ler opens the valve to obtain the required flow of gases with a response of 6 seconds to within % of set pOint. An output of 0 to 5 volts DC is proportional to the measured mass flow rate in the sensor of the flow controller. A contact closure circuit opens or closes the valve on the flow controller. It is necessary to close the contacts if it is required to shut the valve of the flow controller. The multiport valve actuator is a motor that move8 unidirectionally for an 8-position Stream Selector Valve. The movement is activated by 110-130 volts AC signal that starts

PAGE 49

41 the motor. The setting of the position in the valve is ob-tained through means of a cam and a contact closure switch. As the cam moves so does the switch position, which deactivates the movement of the motor at a new valve position. The intro-duction of the 110-130 volts signal through another line reactivates the motor and changes the valve to a new position. Some of the signals to be read from the previous transducers or needed to control them are not necessarily compatible with the signals generated or read by the data acquisition equipment of a computer. It is necessary to condi-tion some of the signals between the transducers and the computer in order to obtain compatibility. The signal con-ditioning devices used for these transducers are presented in table 3. Table 3. Signal conditioner devices. TRANSDUCER :DEVICE :SIGNAL PHOTOMULTIPLIER :AMPLIFIER DISCRIMINATOR :ECL PULSE OUTPUT DIGITAL PHOTON COUNTER :TTL PULSE OUTPUT THERMOCOUPLES :MONOLITHIC THERMOCOUPLE AMPLIFIER AD594 : (10mV/oC) OUTPUT FLOW CONTROLLERS: RELAYS 5 :RELAY DRIVER VALVE ACTUATOR : POTENTIOMETER :MULTIPORT VALVE RELAY :5 VOLTS INPUT :DIGITAL INPUT :0-5 VOLTS OUTPUT :5 VOLTS INPUT-115 VOLTS The amplifier discriminator amplifies the pulses generated by the photomultiplier. The presence of a photomul-

PAGE 50

42 tiplier signal pulse of 20 microamps (1 mV threshold voltage) or greater will cause the amplifier discriminator to output a 50 nanosecond differential EeL (electron coupled logic) pulse. This EeL pulse is read by the digital photon counter which in turn can output a TTL signal of more than 2.4 volts for each EeL pulse received. The TTL signal can be read by a computer logic system. The millivolt outputs of the thermocouples are amplified by a monolithic thermocouple amplifier chip that outputs 10 millivolts per each degree of temperature relative to an ice point reference. These amplifiers have been precalibrated to match the characteristics of type J (ironconstantan) thermocouples. Five Gordos 8312 relays were used to open or close the circuit that controls the shut off valve of each of the five flow controllers. These are activated by 5 volts passing through a coil in the relay which in turn creates a magnetic field that closes a switch, thus closing the circuit. Due to the low current levels generated by the computer it was neces sary to use a Sprague model ULN-2003A high current relay driver made of transistorized circuitry. This driver was activated by the TTL output of the computer's digital lines and used to drive an external 5 volts line through the coil of a selected relay. The following figure represents the way in which these were connected.

PAGE 51

G S VO,I UlN-200H J (URRlHI hf1 DRIHR \ 000 GORDOS CONIACT o F.r ,)J (lOSUR[ oooN J For L.--, ( o : \ OSJ./ F.e. '3J o : \\...n 0 0 Of ..ti.J For Figure High current relay driver connections. The potentiometer was used to monitor the posItIon of the actuator for the multIport valve. ThIs potentIometer was attached to the multiport valve with two gears perpendicular to each other. One of the gears followed the movement of the actuator valve selector axis and concurrently moved the second gear that was coupled to the axis of the potentiometer. A five volt line from an external power source was connected to the

PAGE 52

potentiometer. As the potentiometer axis rotated it changed the internal resistance, outputting a particular voltage related to the position of the multiport valve. A Magnecraft relay model w88CPK-5 was used to control 115 volts AC which drives the actuator motor circuit of the multiport valve. This relay was controlled by a single digital TTL output from the computer. A change on or off of TTL output from the computer activated the relay which then changed the path taken by the a.c. power used to drive the motor. A series of switches were used to set up either manual control of the actuator or computer control of it. Figure 5 represents their configuration.

PAGE 53

HI ))SV LO\.' C-COI':PUHR OR tH'jAI
PAGE 54

an analog signal in order to control a given transducer. The D/D mode represents a digital signal from the computer to be interpreted by a transducer as an ON/OFF condition. Summary of input/output modes. ***************************************.******************.**.** PHOTOMULTIPLIER 5 FLOW CONTROLLERS PRESSURE GAUGE 2 TEMPERATURE MULTI PORT VALVE TOTAL A/D 5 2 D/A 5 D/D 5 PULSES 9 5 *.****.***.************.*********.********************.********* The selection of the data acquisition equipment was based on having the capabilities to perform the input/output functions previously described and fast counting of pulses generated by the photomultiplier. An IBM XT Personal Computer was used to interface with the flow tube system. It consisted of a hard disk, a 360K floppy disk drive, of memory, color monitor and graphics printer. Two boards from Tecmar Incorporated were used to perform the data acquisition task. These boards were added to the bus system'of the PC. The first board known as Lab Master Board can perform the four modes of input/output for the different signals. Some of the main features of this board are outlined in the following table 5.

PAGE 55

47 Table 5. Features of the Lab Master board. **************************************************************** ANALOG/DIGITAL: 16 single-ended or 8 differential inputs 12 bit resolution 30 KHz A/D conversion rate Voltage range selection V, 0-10V DIGITAL/ANALOG: 2 independent D/A output channels 12 bit resolution Voltage range .5V, V, V, 0-5V, 0-10V DIGITAL/DIGITAL I/O (INTEL 8255) 24 parallel input/output lines programable in groups of 8 or 12 TIMER/COUNTER (AM9513) 5 independent 16-bit counters(cascadable) Programable gating and count source selection **************************************************************** Due to the limited number of digital to analog output channels for the Lab Master board it was necessary to add an additional board that would perform D/A conversions for another three channels. This was performed with a second board from Tecmar Incorporated, known as a Dadio Board. This second board features an addition of four digital/analog channels and 24 parallel digital I/O lines. The analog to digital convertor were set at a range of -10 to +'0 volts to the signals generated by the nine lines. The two digital to analog convertor lines from the Lab Master and three of the same type of lines from the Dadio board were set to output a voltage in a range of 0 to +'0 volts. These settings were done by switches found on the boards. Of the 24 parallel digital to digital input/output lines in the Lab Master only six were used to

PAGE 56

48 control the five contact closure circuits of the flow control-lers and the additional line that controls the relay for the actuator, only requiring the use of one eight bit port. The five independent counters can be set up to inter-act with each other through hardware and software control. Of the five independent 16 bits counters only three were used to read the pulses. The fir8t counter was set up as a timer to count for a requested time interval. It uses an internal oscillator in the board with a 1 MHz crystal as the source of counts for timing. The maximum counts in BCD for 16 bits in counter one is 9999. The reading' of counts from the internal oscillator can be done in different scales, allowing for flexibility in the range of time to be set to count. The second counter reads the signal from the source (photomul-tiplier) and when its contents is full it overflows into the third counter, thus allowing a maximum count of more than 9.9 7 x 10 pulses of signal to be counted. Counter one was programmed to output a high level signal when it completes the counting of a given time interval. This signal goes through a hardware connection to the third counter. The second counter was also programmed to sense this signal in the third counter, and to deactivate itself when it was sensed, thus stopping the counting process.

PAGE 57

49 Software Description The program used to control monitor the experiments was written in the BASIC The selection of high level language was based first on its feasibility to program the data acquisition equipment and secondly because the kinetics experiments did not require control by a fast low level language. This second consideration can be seen from the fact that there is a need to wait until proper conditioning of the system is achieved when a new set of conditions is programmed. Programming of the data acquisition equipment is obtained by writing or reading a byte of information at the different ports (addresses) of the boards associated with the different signal modes. The different ports (addresses) in the boards are determined relative to a set memory location through means of an external switch. For example if the board is located in decimal memory location 1808 and the port to write the first 8 bits of a byte for a digital to analog conversion is located 2 relative to the set memory location, then a digital to analog signal output can be achieved by writing the byte to decimal memory location 1810 of the computer. The use of the function IN? or OUT of the BASIC language will read or write a byte from or to a particular port of the computer memory. The relative memory locations for the different ports (addresses) of the Lab Master Board are

PAGE 58

50 presented in the following table 6. Table 6. Programming of the Lab Master board. **************************************************************** RELATIVE MEMORY LOCATION o 2 3 5 6 8 9 10-11 12,13,14 READ (INP) OR WRITE (OUT) W NOT USED FUNCTION (LOW 8 BITS) (HIGH 4 BITS) (LOW 8 BITS) (HIGH 4 BITS) CONTROL (INiERR.,INCRE.,ETC) CHANNEL (SELECT,LOW BITS) START (CONVER.,HIGH BITS) TIMER INTERRUPT DATA PORT OF TIMER CONTROL PORT OF TIMER DATA OF PARALLEL PORTS CONTROL OF PORTS *************************************************************** The programming of the Lab Master board as well as the Dadio board is straight forward in of reading and writ-ing to the ports that are associated with the and lines. The counters are each programmed individually by writ-ing to the control port of the counters. This individual programming of each counter is done taking into the functions to be perform by each counter. There is also an individual loading of initial data on each counter prior to activating them, which is performed by writing to the data port of the counters. The reading or writing to each 16 bit counter is done by readinglwriting 8 bits at a time, due to the fact that the PC is an eight bit computer. Therefore the reading of the contents of a counter is done by assigning the

PAGE 59

first half of the sixteen bit readout as the low values and the second half as the high values. The final resulting number will be a combination of these two numbers. If binary coded decimal units are selected, then the high values will represent 100 times its contents as compared to the low values. If an additional counter is used to count for overflows in the first counter then the final resulting number will be a combination of the two numbers (high and low) of each counter in which the high values of the second counter represents 1000000 times its contents as compared to the low values of the first counter. The individual programs to perform the data acquisition task were incorporated as subroutines in a main program named Actuator, see Appendix 1. The flow of tasks performed by this program is outlined in the following figure 6.

PAGE 60

52 ****.* ** **.***._.**._*.*.-*-*-****.*******-****-********** SET UP GASESIFLOW +++----+++--------+++ CONDITIONS -------------TURN ON FLOWS NON-FLOW DATA ACQUISITION -OFF PORT INCRE.D/A ON OFF CRITERIA SIGNAL 1'++1 NEXT PT. OUTPUT READINGS CONCEN. COMPUTATIONAL VELOCITY 'TASK RXNS.TIME WLSQ ++++ OUTPUT RESULTS TITRATION OR RATE GRAPHICS 1 ICOUNTER ON READ READ COUNTER AVERAGES CRITERI,A Figure 6. Program flow.

PAGE 61

The first part of the program goes through an interactive section in which the user sets the experiments and condi to be performed by the computer. This section starts by opening a file that contains the previous conditions used and loads these as default values. The previously used include calibration parameters of the flow controllers for a particular gas, name of reagents, concentration of these reagents in the cylinders, flow of gases previously used, increments in the flow of gases, and other parameters not related to the flow of gases. The user is asked if he desires to change any of the previous conditions, if so, he can go through each flow controller known as Alpha, Beta, etc. and change any of the previously mentioned parameters. The last part of this interactive section takes the user to the setting of non-flow parameters. Here the user can select the amount of time to be loaded to the timer in the counters, the upper value in the Y axis of the output graphs, and the mode of counting. In mode one the computer reads all the transducers while each counting interval is performed or in mode the computer counts Signals for each reading of each transducer. The user can also select which flow controller is used with the N02 and hydrogen source. This section allows you to set the number of readings to be done per data point by setting the number of times for the counting cycle and also to set the stabilization criteria used to accept a point of data. This

PAGE 62

54 stabilization criteria sets the limit in the value to be accepted for the standard deviation. If the standard deviation of the total of readings of the fluorescence signal is bigger than a percent value of the average of this signal, then a new set of readings is obtained. The user can also select the waiting time before readings are done when either N02 is turned on or off. The last part of the setting of non-flow parameters sets the number of points to be collected (maximum of 10) and if the multiport valve (actuator) will be active or not. The program sets the flow of gases by turning on (opening the contact closure circuit) the flow controller valves through means of the lines and setting the required voltage in the digital to analog lines that sets the flows in the flow controllers. The required voltage is calculated from the set flow of the gases stated in the interactive section of the program and the calibration parameters for the flow con troller. The program adjusts the total flow of gases to be equal to the set flow of the helium diluent that flows through the Epsilon flow controller. Therefore the total flow in the system depends on the initial set flow of the helium diluent. The data acquisition task is mainly composed of two sections. In the first section it turns off the N02 gas by means of a subroutine that interacts with the port. Then it either sets a port location for the multiport valve if the

PAGE 63

56 After all the data have been collected for a particular pOint. which includes data for OFF, ON and OFF cycles, it verifies that the baseline (background signal in the OFF cycles) does not differ more than 3% between the two OFF signals. The baseline is the average of these two OFF signals. If the criteria for the baseline is not met it collects another ON and OFF set of data. If the new point is acceptable it calculates the HO signal as the difference between the signal in the ON cycle and the baseline. Thereafter it goes to the collection of other pOints by alternating between ON and OFF cycles of the data acquisition sections and setting new flows or port positions. The first output task prints the values of the readings of voltage for the different flow controllers, pressure gauge. thermocouples. port positions. total counts. baseline, and signal. Thereafter it goes into the computational task subroutines which first calculates the concentration of the reagents in the flow tube based on the individual flows. total flow. pressure. and concentration in the reagent cylinders. Another subroutine calculates the velocity of the gases in the flow tube by taking into consideration the standardized total flow rate. the approximate inside diameter of the flow tube and the pressure. After this, a computation is performed .if the actuator is active. in which the reaction time at the different ports is calculated based on the velocity at which

PAGE 64

57 the gases travel inside the flow tube and the distance between each port and the detection area. If the multiport valve is active then a weighted least squares analysis subroutine is used to calculate the slope of the natural log of the HO signal versus the reaction time at each port. The last output includes the previously calculated parameters of concentration, velocity, reaction time at ports and slope (pseudo-first-order rate constant) and a graphical display of the signal versus flow of N02 for a titration or natural log of signal versus reaction time for a kinetics run will be displayed. After these outputs the program can be started again with previously used conditions or setting new ones in the set up conditions section. The calibration of the flow controllers using the soap bubble method was monitored by the use of another software program named Comascal. Comascal was written by Paul M. Gates, and lncludes a least-squares-analysis method to compute the slope of the flow and set voltage. The least-squares-analysis method is based on a publication from Irving, J. A. and Quikenden, T. I. (17) The program calculates the slope of flow versus voltage, the flow intercept, errors in the slope and intercept, the covariance, and correlation coefficients.

PAGE 65

CHAPTER RESULTS Experimental The experimental results for the kinetic runs were obtained from the data output of the computer with the corresponding experimental conditions. As mentioned previously the experiments were started with the titration of the hydrogen atom content in the flow system with N02 thus an approximate HO concentration was determined. An example of a titration of one of the experiments is presented in the following figure.

PAGE 66

59 15Q999 14:59:59 94-28-1986 134999 r 1 119989 194979 lr" L 89969, o o o o o o o J 14959 59949 44939 29929 f l' 14919 15.6 31.2 46.8 62.4 78,9 93.6 199.2 124.8 149,4 156.9 Figure 7. Titration curve. This shows the relation of the HO fluorescence versus the in standard cubic centimeters per minute of N02 As the N02 flow is increased, the production of HO radicals in the system increases, until a maximum production of signal or HO radicals is reached. Above this N02 flow the signal stays relatively constant or could decrease slightly due to the

PAGE 67

reaction of HO radicals with the excess of N02 This last pOint is insignificant at the conditions of pressure and velocity used the experiments. The end point of the titration can be calculated by finding the intercept of the two lines that would follow the two different trends in the graph. It is assumed that every N02 molecule reacts with a hydrogen atom to form HO and NO. N02 H HO NO This last reaction is considered to be fast at 1.3 x 10-10 cm3/particle-sec. The concentration of hydrogen atoms in the system can be calculated from the results of the point where the signal changes slopes. By knowledge of the concentration and flow of N02 at the pOint of change in the slope and the flow of the hydrogen/helium mixture we can compute the concentration of hydrogen atoms or HO radicals. The pseudo-first-order decay of HO is shown for a particular n-butane concentration in figure 8.

PAGE 68

12.94 11.79 F 11.37 11.93 L 19.69 o 19.36 R 19.92 S C 9.68 N 9.35 '.91 RINETICS 61 16:24:51 94-28-1986 ..... 1l ",. .. ......... '1 j ... 0..... J 9.99289.99569.99849.91129.91419.91699.91979.92259.92539.9281 TI HE I N SECONDS Figure 8. decay. The figure shows a plot of the 'natural log of the signal versus the eight reaction times in seconds. The reac-tion time corresponds to different reaction distances of the pydrocarbon reactant introduced through eight reaction ports at a fixed total flow velocity. The slope of the line cor-responds to the pseudo-first-order rate constant for the

PAGE 69

62 reaction of HO radicals at the set conditions. The experimental results for all the kinetic runs are presented in appendix 2. The date, n-butane concentration, pseudo-first-order rate constant, and standard deviation are included. Also included is the apparent bimolecular rate constant for a particular set of data and for a whole set of comparable HO conditions and corresponding intercepts. Kinetic results for higher pressure are also included. It should be pOinted out that not all of the decay kinetics results were obtained under the previously recommended conditions of a forty-fold excess of n-butane over HO radicals, especially for the Very High initial HO concentration experiments. This is due to the nature of the experiments performed, in which the effects of secondary reactions are bigger when there is less excess The second point to notice is the use of the term "apparent bimolecular rate constant" which is the calcu-lated rate constant for the bimolecular reaction of HO radi-cals and n-butane, assuming no other reaction process is affecting this result, which is not what it was found in this study. The calculation of the apparent bimolecular rate constant for a partioular set of data was done by performing a weighted linear least-squares-analysis of the pseudo-first-order rate constants versus the n-butane concentrations, weighted by the error in the rate constant. As we can see from

PAGE 70

63 the experimental results in appendix 2 there is a significant difference in the results obtained for pseudo-firat-order rate constants at different initial HO concentrations. With the exception of the very high intercept for the last data set in the Very High HO results, the results are quite comparable between each set collected at the same HO concentration. As mentioned before, the intercepts are related to loss processes for HO radicals which are independent of reactions with n-butane, such as the loss of HO by interaction with the walls. This is believed to be the case for the Very High HO results due to deconditioning of the wall prior to these experiments. It can be seen that there is a bigger difference between the pseudo-first-order rate constants measured at high n-butane concentrations compared to low n-butane concentra-tions, for the different HO concentrations. This difference can be readily explained from the fact that even though there is a larger percent of secondary reactions occurring at lower n-butane concentrations, the effect is less noticeable in the small rate constants in this lower n-butane concentration range than is a smaller effect of secondary reactions affect-ing a bigger pseudo-first-order rate constant. An example of this would be a 25% effect on a given rate constant of sec would be a net effect of 10 sec in the rate as compared to a 10% which would be constant of 200 sec-1 percent effect on a rate a net effect of 20 sec

PAGE 71

The following graph compares the HO signal decay for low and high initial HO concentrations at a n-butane concentration of 7.5 x molecules/cm3 done on May 'll, 1986. The experimental signals were normalized relative to the signal at the smallest reaction time within each set of points and the natural log of this relative signal was plotted versus the reaction time. G/ --f,) g c -2 -3 -4 T "It:: Figure 9. Relative signal decay of HO for High and Low HO concentrations.

PAGE 72

65 The non-linear effect of the signal decay at high HO concentration is quite noticeable when compared to low HO conditions. This effect is more apparent at longer reaction times. This effect could be explained by the fact that at longer reaction times there will be more product n-butyl radical formed, therefore, the HO radical will be more likely to react with this product. A higher rate of reaction for a second process compared to the rate of the reaction for HO and n-butane is necessary to account for the deviation observed. Figure 10 shows the values of all the pseudo-first-order rate constants plotted versus the n-butane concentrations, for initial Low HO (2 x 1011particles/cm3), High HO (2 x 1012particles/cm3) and Very High HO (1 x 1013particles/cm3).

PAGE 73

66 +LOW *HIGH xVERY 260 208 ,... 156 ...... ...., U, 0 ..t: 104 0:: 52 o 20 40 60 80 100 n-BUTANECE+12) Figure 10. Kinetic results for Low, High and Very High HO concentrations. The lines in figure 10 were obtained by plotting the results of an unweighted least-sQuares-analysis. The analysis by the least-squares techniques gives slopes of 2.25, 2.91 and 2.96 for Low. High and first set of Very High HO concentra-tions which Is even less dramatic between HIgh and Very High

PAGE 74

67 HO than the weighted least square analysis. The least-square method is used in the previous plotted figure. The graph us a visual representation of the experimental results. These results are not significantly different from the weighted analysis shown in appendix 2. The difference appendix 2 between the apparent Low HO bimolecular rate (2.23.Oij x 10-12 3 cm molec .sec ) and the apparent High HO bimolecular rate -12 3 (2.85.05 x 10 cm molec sec ) is quite significant. On the other hand. the difference between apparent High HO (2.85.05 x 10-12 cm3 molec1sec-1 ) and Very High HO bimolecular rate (3.08.15 -3.13.23 x 10-12 cm3 molec1sec-1 ) is less dramatic. A contribution to these differences is the fact that there is a factor of 10 difference between Low HO and High HO as compared to a factor of 5 difference between High HO and Very High HO. Also to be considered the fact that the experiments for Very High HO were mostly done on the lower end of n-butane concentration range where the change of the pseudo-first-order rate constants is going to be smaller. as previously mentioned. Even so. the values obtained at Very High HO are smaller than expected. based on preliminary model simulations. At Very High HO there is a larger quantity of secondary reactions occurring. thus a higher apparent bimolecular rate constant should have resulted. Results for experiments performed at higher pressures Showed a reduction the apparent bimolecular rate constants

PAGE 75

68 when compared to the low pressure studies. This'result 1s different than expected. If the HO radicals add to butyl radicals to form an alcohol, then the increase in pressure would stabilize the act1vated complex. Therefore, the rate for the secondary process would be higher, thus one would expect an increase in' the apparent bimolecular rate constant. If the HO reaction with butyl radical proceeds by abstraction of another H to make an alkene, one would expect no significant effect of increasing pressure. Trapping results of the products of the reaction of HO radicals and n-butane are presented in the following chromatogram and mass spectrum. TI': a ::! 5 1 '" 5 / 1 4 Figure 11.1 Gas Chromatogram.

PAGE 76

.r69 n r .. ,1 n:; c 5u Su 813 Figure '1.2 Mass Spectrum. The chromatogram shows three main peaks. Analysis of the first peak at 1.388 minutes shows it to be air in the injected sample which is expected due to the sampling tech-nique used. The samples were collected from the trap submerged in a bath of CC14 and dry ice at atmospheric At these conditions the gas sample of the products is in a liquid state which immediately becomes a gas when it is collected with the liquid sample syringe. The syringe will aquire air from the environment when it is withdrawn from the trap. Analysis of the mass spectrum for the peak at'. 388 showed a high abundance of 28. 32 and 40. which are due to nitrogen. oxygen and argon. The second peak at 1.533 minutes is n-butane. The last main peak emerged at approximately the same retention time (2.107 minutes) as the reference samples of sec-butanol. There

PAGE 77

70 was no indication of a chromatogram peak or mass spectra for butene as a product. Further analysis of the mass spectrum of the peak at 1.533 minutes of n-butane was done to see if there was any contribution of butene within the n-butane mass spectrum. Previous studies of the chromatogram for butene showed that this olefin emerged immediately before or at the same time as the n-butane at the conditions used. The of the n-butane mass spectra to check for any contribution from butene within it was done by checking the ratio of the abundance of peaks 56, 55, 53, 51 and lJO relative to peak 29. Peaks 56, 55, 53, 51 and 40 relative to 29 have been shown to be significantly larger for trans-butene and as compared to the same peaks relative to 29 in the n-butane spectra and If any olefin would have appeared together with n-butane product peak it would have been observable in a higher ratio of abundance of peak 56, 55, 53, 51 and 40 relative to 29. The ratio of these. mass peaks in the n-butane product peak of the flow system were and respectively. There was no indication that any butene appeared before or within the butane peak. The mass spectrum in figure 11.2 is presented for the peak with a retention time of 2.105 minutes contains teristic mass fragments of n-butane and an alcohol (secbutanol). It also contains peaks that do not belong to either

PAGE 78

71 species. The fragmentation pattern shows a 14 mass unit repetition characteristic of hydrocarbons up to a mass unit of 57. The spectra also contains a mass peak at 30 that is not characteristic of either n-butane or an alcohol. 30 is characteristic of fragments NH2CH2 CH20, or NO. It contains a relatively large peak at 45 which is characteristic of secbutanol, and a relatively small peak at 31 characteristic of alcohols. The peak at 31 is relatively small compared to the abundance of it for alcohols. A large peak at 57 is found which is not as big in n-butane or sec-butanol, but is the obvious result of the breakage of a parent compound that contains butyl. Another relatively large peak at 74 mass units is found, which appears to be that of a possible parent peak. This peak at 74 mass units is not as big as in the reference sample of sec-butanol but is the mass of a butanol molecule. These results suggest that the product contains some combination of a butyl and OH radical. It also suggest the possible influence of NO or N02 molecules in the product based on the large intensity at of 30. A further study was done to see if n-butane was reacting with NO, N02 HO or H 2 0 after been trapped. This was performed by trapping only n-butane for a given interval of time. After n-butane was trapped it was turned off and the hydroxyl radicals were made and also trapped together with NO, N02 and H 2Q. The results showed that none of these react with

PAGE 79

72 n-butane in the trap at the conditions used. Therefore the previous observed product at a retention time of 2.105 is either the product of butyl radical and HO which further reacted in the trap or in the gas phase with NO or N02 or the product is the result of butyl radical and NO orN02 reacting in the gas phase prior to being trapped or any other reaction involving butyl radicals. Error Analysis The calculation of the uncertainty in the bimolecular rate constant was based on the evaluation of the systematic errors and random errors in the different parameters that go into its different phases of its calculation. The propagation of the systematic errors were done unidirectional1y using the absolute values obtained at the different phases of the ca1-cu1ations. The sources of systematic errors used in the calculations were: Propagation error results for calibration of flow controllers using the results from the comascal program. These varied for the different flow controllers and were less than 2% for the slopes of the 2) A voltage setting resolution of bit in bits for the which for the 10 V range corresponds to .002 v. 3) A pressure error of 50 millitorrs, the main source of

PAGE 80

73 uncertainty being in the zero offset value of the gauge. 4) The uncertainty of .5% in the concentration of n-butane in the cylinder. 5) The uncertainty of 5% the reaction system temperature used to determine concentrations from flow and pressure measurements. 6) An error of in the timer and the reading of counts by the counter to calculate the signals. 7) An uncertainty of .2 cm in the correct distance between the different ports and the detector. The results of the error analysis showed that the biggest systematic uncertainties resulted in the calculation of the flow of n-butane, which was about 3% for the highest flow of n-butane. For the lowest range of flow of n-butane this error was 13%. The systematic error the flow of nbutane resulted in the calculation of systematic errors in the concentration of n-butane that ranged from 9% at high n-butane concentrations to 19% at low n-butane concentrations. The calculation of random errors were mainly based on the standard deviation of the set of readings obtained for the different transducers and included: 1) The statistical error obtained for the calibration of the flow 2) The experimental readings of the voltages of the flow controllers for a particular set voltage. The variability

PAGE 81

74 observed in these readings were approximately .005 volts for a 0 to 5 volts range. 3) The readings of the voltage for the pressure which was approximately .005 volts at 1 volt (1 torr). 4) The readings for the counts to estimate the signals. This particular error was controlled by the software to be less than 1% for each data set used. The biggest random errors were also related to the flow of n-butane which ranged from about 3% for high flows to 11% for low flows of n-butane. The calculation of the random errors in the concentration of n-butane ranged from 7% for the high n-butane concentrations to 13% at the low of The error analysis determined that the errors for the total flow were 4% for both systematic and random analysis at all ranges of n-butane concentration used. The major contributor to the error in the. signal was the random error and was not higher than 1.2%. The errors in velocity and reaction times were 2% from the systematic analysis and 4% from the random analysis at all ranges of n-butane concentrations. The error analysis of the calculation of the pseudofirst-order rate constant for the total reaction time showed the uncertainties were heavily influenced by the total reaction time studied and the amount of signal change during that reaction time. At high n-butane the systematic and random

PAGE 82

75 error in the calculation of the pseudo-first-order rates were 6% and 10% respectively. The use of weighted least-squares-analysis for the calculation of the pseudo-first-order rates (weighted in Y axis for signal error) and bimolecular rates (weighted in Y axis for errors in the pseudo-first-order rate constants) minimized the sum of squares of the deviations in the Y values for the set of X values used. The weighted-least-squares error results are the errors that appear in appendix 2. In general the errors for the pseudo-first-order rates constants were less than 5%. For the precision of the bimolecular rate constants it can be seen that the errors weighted in the Y axis also resulted in less than 5%. The randomness in the n-butane concentration was not taken into consideration when the bimolecular rate constants were calculated using the weightedleast-squares. For similar experiments performed on different days the random error the bimolecular rate constant in the results was found to be less than 5%. The error in the determination of bimolecular rate constants based upon the propagation of systematic errors were about 11%. The overall accuracy of the results for the bimolecular rate constant were 12%. This value is near the 10-15% accuracy range which is believed to be the best one can do in flow tube measurements of rate constants as indicated by Howard (31 )

PAGE 83

76 Modeling Simulations of the kinetics involved for the reactions of HO and n-butane were performed to elucidate the possible mechanism and rates which affect it. The simulations were performed in an IBM personal computer using the integration techniques described by C. W. Gear. (18)(19)(20) The rates used for the reactions in the simulation were as recommended (21) (22) for the experlmental conditions involved. The selec-tion of reactions used in the simulations was based on which of them could have a significant effect in the model. This was done by taking into consideration the concentrations of the species in the system and the rates at which these react with other species. The rates for the reaction of n-butane and HO radicals and the postulated reaction of HO and butyl radicals were adjusted to best fit the experimental results. Simulations were done which involved the modeling for the formation and lifetime of HO radicals in the system for the three different HO The reaction for the formation of HO, 2 and subsequent reactions were included. These simulations were done omitting the reactions that involve the hydrocarbon in

PAGE 84

77 order to obtain a description of the formation and disappearance of HO radicals in the system. The results show that the expected HO concentration was smaller than the previously calculated concentrations from the titration and signal monitoring technique. For Low HO the simulated HO concentration was 3% smaller, for High HO concentration it was 9% smaller and for Very High HO it was 22% smaller. The calculated HO concentration from the model indicates a lower HO in the system as compared to the experimental results from the titrations. The from the model HO lifetime are more accurate and were used in the modeling of the reactions of nbutane and HO. It was also noticed that the HO concentration for Very High HO varied almost 25% between the first port of reaction and the last port. The variation for High and Low HO was 6% and 4% respectively. Further studies of how the apparent bimolecular rate of HO and n-butane was affected by this variation in the HO concentration at Very High HO showed no effect on the rate, but it reduced the intercept by approximately 50%. This reduction in the intercept from the simulation is in accord with the low intercept observed experimentally. The simulations performed in this study were done without varying the concentration of HO at the different ports. The main simulation model included 28 different reactions and 19 species that were calculated. The list ofreac-

PAGE 85

78 tions and rates used appears in appendix 3. The reactants involved in the simulation included Ho, N02 NO, HO, H 2 0 2 H 20, H02' H 2 HN03 00 02' N03 n-butane, butyl, R-OH, R-O, octane, butene, alkoxy, R-NO. The value of the rate of com-bination and disproportionation for butyl radicals was es-timated from the rate data for similar reactions of propyl radicals(25) and t-butyl radicals. (26) The rate for the reac-tions of NO and butyl radicals was estimated as the same as (27) the rate for the reaction of NO and methyl radlcals. The rate for the reaction of N02 and butyl radicals used was the same as the rate of the reaction of N02 and methyl radicals. (28) It is believed that these last two reactions could be faster than the values here used. This is from the fact that radical reactions of HO and hydrocarbons increase in their rate as the hydrocarbon chain inbpeases. The simulations were performed at eight different n-butane concentrations. The apparent bimolecular rate constant was calculated via a least-squares-analysis of the obtained pseudo-first-order rate constants and n-butane concentration. The reactions were simulated for 25 milliseconds, which is approximately the maximum reaction time of the experiments. The simulations were done with a tolerance error of 1% in the calculations. The first series of simulations was done to estimate the minimum rate constant for the secondary reaction of HO and butyl radicals. The simulations were done excluding all the

PAGE 86

79 reactions that deplete the butyl radical, with the exception of the reaction of HO radical and butyl radical. The exclusion of these reactions maximized the concentration of butyl cals, hence calculation of a lower for the rate constant of the secondary reaction of butyl and HO radical, in order to best match the observed effect of the change of initial HO concentration on the apparent bimolecular rate constant. The simulations were done adjusting the values of the rate of the primary reaction kl between HO and n-butane and the secondary reaction rate k2 between HO and butyl radicals. The values obtained are presented in the following table. Table 7. Simulations with minimum k2 *******************************-******************************** k12.1 x 10-12cm3/molec-sec k2= 1.2 x **********.*************** *.********.**********.**.**.****** INITIAL HO APPARENT BIMOLECULAR RATE (x 10-12cm3/molec-sec) CONCENTRATION MODEL EXPERIMENTAL LOW 2.15 2.23 HIGH 2;78 2.85 VERY HIGH 3.60 3.08-3.13 ****.*********************************************************** we can see from the results the values of rates obtained with the model are similar to those obtained ex-perlmentally, with the exception of the values at very high

PAGE 87

80 HO. The values obtained with the model at Low and High HO are 4% and 2% different from the experimental values as compared to a 16% difference at Very High HO. The sensitivity of the secondary reaction between HO and butyl radicals on the apparent rate was previously studied(29) in a similar reaction scheme for the Low and High HO range. This study was conducted to show how k2 influences the apparent rate. The (29) results of the previous study are presented in the follow-ing table. Table 8. Effect of kl and k2 on model values for apparent rates. Apparent Bimolecular Rate for HO + n-Butane (cm3/molec-sec) LOW HO HIGH HO k k (x Difference 1.9 1.2 1.96 2.58 .62 2.1 '.4 2.27 ;22 2.1 .8 2.10 2;58 .48 2 1 1 .2 2 5 2.80 .65 2.1 1.6 2.97 .77 2.1 2.41 3.36 .95 Experimental 2.23 2.85 .62 The magnitude of the difference between apparent bimolecular rate constants measured at Low HO and High HO is highly influenced by the value of k2 used in the model, and not as much by the value of k1

PAGE 88

A second series of simulations was done to estimate a maximum value of the rate constant for the secondary reaction. This series of simulations included all the reactions that involved the butyl radicals. Including these reactions, which are loss processes for the butyl radical, reduced the con-centration of butyl radicals in the model. Having a smaller amount of butyl radicals will require that the rate of reac-tion between HO and butyl be increased to be able to simulate the experimental results. The resulti of these simulations are presented in the following table. Table 9. Maximum k2 value and effect of k2 on apparent bimolecular rate. **************************************************************** LOW 2.1 8 2.34 2.1 5 2.24 EXPERIMENTAL 2.23 APPARENT BIMOLECULAR RATE "12 3 (x 10 cm Imolec-sec) HIGH 3.02 2.93 2.85 VERY HIGH 3.09 **************************************************************** -10 The simulation with k2 c 5 x 10. gives values near the experimental apparent bimolecular rate constant. The

PAGE 89

82 series of simulations for maximum k2 included the reactions of NO and N02 with butyl radicals. These two reactions become important at Very High HO concentrations, because of the high concentration of NO and N02 in the As we can see the simulations for maximum k2 predict a relatively low value for the apparent bimolecular rate constant at Very High HO, which was observed experimentally. The values of the experimental pseudo-first-order rate constants and modeled pseudo-first-order rates constants versus n-butane concentrations are presented the following figure 12 with adjustments made to all the rates in order to make the graph easy to see. The successive curves were offset on the y axis. The offset was obtained by adding or substracting a given amount to a set of obtained pseudo-first-order rates. This addition or subtraction will not change the slopes of the curves. The intercepts of the experimental results for the bimolecular rate constants for Low, High and Very High HO were respectively set at 1, 10 and 20 s-1. The intercepts for the model results were also set accordingly.

PAGE 90

83 280 224 168 112 < 56 +L *H xV #MODEL o 20 40 60 80 100 n-BUTANECE+12) Figure 12. Modeling and experimental results. The results for the concentration profiles of the main species in the simulations for Low HO and High HO are presented in figures 13.1 and 13.2. The profiles are presented for the concentration of the species of HO radicals, butanol and butyl radicals versus reaction time. The values of k-2.1 x cm3/molec-sec and k2-1.2 x 10-10 cm3/molec-sec were used.

PAGE 91

[n-BUTANEJ=4E13 8 6 .>. :. Co! CO' 2 .0055 CJ .0165 .022 [,275 Figure 13.1. Simulated concentration profile for Low HO. 84

PAGE 92

[n-BUTANEJ=4E13 Be o 60 ___ __ R ....... .0055 .011 .0155 .022 .0275 Figure 13.2. Simulated concentration profile for High HO. The profiles in figures 13.1 and 13.2 for HO, butyl and alcohol indicate a higher concentration of the product alcohol, resulting from the secondary reaction between HO radicals and butyl radicals, at High HO compared to Low HO 85 conditions. These profiles of the concentration were done at a n-butane concentration of x 1013 molecules/cm3 The profile

PAGE 93

86 for High HO indicate a loss of approximately 30% of the original HO radicals to be due to the formation of alcohol. At Low HO a loss of approximately 10% of the HO radicals is due to the secondary reaction. A third series of simulations was performed to model the experimental conditions of the flash photolysis experi-(10) ments performed by Stuhl and Atklnson et al. The reac-tions and concentration of reactants used for these simula-tions were based on the experimental conditions of these systems. The rate for the reactions used the same as those rates presented in appendix 3, in which k for the sec on-dary process is assumed to be maximum. The simulation results for these are presented in the following table 10.

PAGE 94

Table 10. Flash Photolysis Simulations. STUHL HO RADICAL CONCENTRATION (molec/ cm3 ) 11 3 1012 2 10 EXPERIMENTAL 11 3 1012 x 10 ATKINSON 11 1 x 1011 2 x 1011 4 1011 8 x 16 x 10 EXPERIMENTAL OH CONCENTRATION SIMULATED APPARENT BIMOLECULAR RATE -12 3 ( x 10 cm /molec-sec) FJXED TWO TIME 2.36 2.94 2.11 2.25 2.50 2.83 3.30 LIFETIMES 2.12 2.60 2.35 NO CHANGE NOT MENTIONED Experimental--) 2.72 The simulations performed for Stuhl's system were first done using a fixed time approach in which the total reaction time was the same for all the different n-butane concentrations used. The same set of n-butane concentrations were simulated but the pseudo-first-order rate constants were calculated over the time range required for 89-91% of initial HO concentration to react. These two methods are compared because most of the experiments done in flash photolysis use this last method. where in the flow systems the fixed time approach is used. The results for the apparent bimolecular rate constants determined over the time range of 89-91% appear

PAGE 95

88 in the column labeled "two lifetimes". As we can see from these two sets, the use of a fixed time approach gives higher apparent bimolecular rate constant values compared to the method in which the HO concentration is monitored for two lifetimes. The difference between the apparent bimolecular rate constant for the Low HO and High HO in the simulations done is 18% smaller when the method of two lifetimes is used as compared with the fixed time method. In the experiments performed by Stuhl no significant difference was observed in the results of the experiments performed at these two HO concentrations. The reason for not observing a significant difference could be due to the use of the two lifetime technique to analyze the time dependence of the HO concentration. The simulations performed for Atkinson's system were done by doubling the initial HO concentration for each set of n-butane concentrations. The time interval was fixed in the simulations, compared to the experiments performed in which the pseudo-first-order rate constants were measured until at least three HO lifetimes had passed in the reaction with the different n-butane concentrations. (10) Discussion The results obtained in this study demonstrate the influence of the secondary reaction.

PAGE 96

89 HO + butyl R-OH This reaction is important in systems with high and very high initial HO concentrations and less important in low HO sys-terns. The influence that these secondary reactions have on the measured rates is observed when the initial concentrations of HO radicals is changed significantly. The changes in pseudo-first-order rate constants observed indicate that these rates increase as the initial HO concentration increases, thus the results of the measurement of the rate constant for the reac-tion, HO n-butane butyl H2 0 is gOing to appear higher as the initial HO concentration gets higher. The significant change in the pseudo-first-order rates is the result of fast secondary reactions which even influence the results at Low HO concentration. Thus the measured results for the reaction of HO and n-butane should be corrected ac-cordingly. The value of the rate constant for the secondary process in which HO radicals react with butyl radicals appears -10 3 to be in the range from 1.2 to 5 x 10 cm /molec-sec. The maximum value for this reaction is dependent on the value of the rates of the reactions of

PAGE 97

N02 + butyl alkoxy + NO and NO + butyl + (M) butyl-NO + (M) which have been assumed in the simulation studies to be similar in magnitude to the reactions of nitric oxide and nitrogen dioxide with methyl radicals.(27) (28) The reduction in the apparent bimolecular rate constant for Low and High HO when the pressure was increased could be the result of either of two processes. The first could result from an increase of butyl radical loss through 90 any pressure dependent reaction, reducing the importance of the secondary reaction of HO and butyl. The reaction of NO and butyl could be such a pressure dependent process. A higher wall loss rate for the butyl radical could also account for the observed effect. If this-first proposal is true then the value for the bimolecular rate constant of the primary reac-tion of HO and n-butane would be significantly lower than the h;:'parent bimolecular rate constant used in the simulations and secondary reaction of ilO and butyl radical would be ever; higher than that previously used as a maximum in the simulations.

PAGE 98

The second possible process that could account for the reduction of the apparent bimolecular rates at high pressures could be the result of the influence of a radial concentration gradIent. The change to a hIgher pressure and same velocity of flow will create a bigger velocity differential across the flow tube. The flow at the center of the flow tube is going to be faster than the flow near the walls. The radial concentration gradient is the distribution of concentratioon of molecules across the flow tube. The diffusion of these molecules is going to be less efficient at the higher pressure conditions. Therefore the n-butane molecules are going to be more concentrated towards the center of the flow tube because these molecules are added directly to the center of flow. The pressures are higher and the velocity of flow is faster at the center as compared to lower pressure conditions. On the other hand, the HO radicals will be more diffuse. The assumption that the HO is more diffuse comes from the fact that HO radi-cals have a longer time to diffuse as compared to the n-butane because HO is added a longer period of time before the addi-tion of n-butane. These two effects will reduce the possible reactions between n-butane, butyl and the HO radicals. This second possible process will also result in lower measured reaction rates. The data for the trapping of products demonstrated the formation of a product made by the reaction of butyl radicals

PAGE 99

92 and other species. The mass spectrum at 2.105 showed that the product contained a hydrocarbon unit of mass 57, which is butyl. The kinetic results demonstrated that HO reacts with the butyl due to the fact that as the initial HO concentration increased the rates also increased. It is also demonstrated that butyl reacts with NO and N02 based in the results of the apparent bimolecular rate constants at Very High HO. There is still the need to study the possible reaction of NO and N02 with an alcohol in the trap. This will help to elucidate the spectrum obtained for the mentioned peak in figure 11.2 at a retention time of 2.105 minutes. From the simulations of the flash photolysis systems it can be seen that the influence of secondary reactions should have been observed if a fixed time interval was used for measuring the HO time decay. The reason for not observing this influence in most of the experiments could be due to various factors. First, the modeling results for the flash photolysis systems demonstrated that the use of a calculation method that calculates the pseudo-first-order rate constants for a variable time results in lower values for the apparent bimolecular rate constant and in a smaller difference between the apparent bimolecular rate constant for two different initial HO concentrations as shown in table 10. The variable time scale method is typically used in the flash photolysis systems that studied this reaction. This method will

PAGE 100

93 make it more difficult to observe the effect of secondary reactions in the flash system as compared to the use of a fix time interval in the flow method. This does not necessary diminish the influence of secondary reactions flash photolysis systems. The reason for smaller than expected apparent bimolecular rate constant at high HO results from the fact that there is going to be a longer time of measurement of HO decay at the low n-butane range, thus a longer time for secondary reactions to occur. Therefore, the pseudo-firstorder rates at low n-butane are going to be bigger than in a fixed time method. On the other hand, the measurement time at high butane concentrations is going to be smaller in the variable method as compared to the fixed time interval. This smaller measurement time will reduce the influence of secondary reactions in the high end of the n-butane range. Therefore the pseudo-first-order rates at high butane are smaller in the variable time method as compared to the fixed time method. These two opposite effects will tend to account for a smaller apparent bimolecular rate and smaller differences between bimolecular rates for High and Low HO in Flash Photolysis. The second consideration is the fact that the simulations performed for the flash photolysis systems were done using a maximum value for the rate of the secondary reaction of HO and butyl radicals. A lower value for this

PAGE 101

94 process would have resulted in a smaller difference between the apparent bimolecular rate constant for two different initial HO concentrations. Simulations done with a lower rate constant for the secondary reaction of HO and. butyl would have resulted in a smaller difference between the apparent bimolecular rate constant results at two different HO con-centrations. This smaller predicted difference will be dif-ficult to observe experimentally. The third factor is the small increment in the HO concentration used in the flash systems to test for the effect of secondary reactions. As an example of these we can see that Atkinson et ale (10) only incremented the initial HO by about a factor of two, and Tully by about a factor of four. Even though Stuhl and ( 11) Paraskevopoulos lncremented the initial HO concentration by about a factor of 10, in all the previously mentioned flash photolysis experiments, the increments in the initial HO concentration are based on the increment of the initial pres-sure of the precursors of HO radicals or the increment of the energy of flash. In any of these previous studies there is no mention of the effect of these changes in the fluorescence signal to properly deduce the change in the initial HO con-centration. Assuming that the previous increments were proper, then there is the remaining uncertainty of the measurement of the initial HO concentration. The concentration of HO radicals in a flash photolysis system is difficult to estimate. None of

PAGE 102

95 the previous work mentioned the technique used for the estima-tion of HO radicals in their systems. The uncertainty in the initial concentrations makes it more difficult to perform a proper modeling simulation study. I ld b d h t N R G (6) bl t t shou e note t a re1ner was a e 0 see the influence of a ten-fold increase in the alkane con-centration on the pseudo-first-order rates. This is a dif-ferent way of changing the ratios between n-butane and HO concentrations. Another point to notice is the mentioned observation of secondary-reaction kinetic interference at 599 ( 1 4) K for the reaction bf HO and butane 1n the work of Tully. This tends to indicate that even in a flash photolysis system, secondary reactions exist whose influence can be detected under extreme circumstances. In the case of Greiner's work, we can deduce a change of the initial ratio between HO and alkanes, compared to the flash systems in which the change In initial HO concentrations was done through a change in the initial concentration of the precursors of HO or change in the flash energy, thus an assumed change in the initial HO con-centration. The experimental result found in this study for the rate of the HO and n-butane reaction of 2.10.25 x 10-12cm3/molec-sec. is in agreement with the direct results found in the studies of StUhl(8) (2.35.35 x 10-12 Anderson and Stephens (12) (2.23.23 X

PAGE 103

10-12cm3/molec-sec), Tully and x 10-12cm3/molecules-sec), and SChmidt (15 ) C2.30.3 x 10-12 cm3 /molec-sec) We could average the experimentally obtained 96 bimolecular rates for the reaction of propane and HO using the d' ( 5 ) f h' h th res ts 0 the dlrect stu les, or w lC ere lS ess scatter compared to the current data for the reaction of HO and butane. (5) Using the relative rate study results for the reaction of HO and propane, and the reaction of HO and n-butane(23) we can see the relation between these two rates. Then, using the obtained value for the rate of HO and n-butane relative to reaction rate of HO and propane we can estimate the bimolecular rate constant for the reaction of HO and n-butane. The selection of this relative study was based on the fact that it shows the highest ratio between the two reactions of interest. This last would overestimate the bimolecular rate constant for the reaction of HO and n-butane. Table 11 shows the obtained results for these calculations.

PAGE 104

Table 11. Previous direct studies of HO and propane. ********************.*******************************.**.**.***** RATE -12 30 ) (cm Imolec-s): 1.20 0.83 2.02 0.93 1.05 1.20 1.10 1.0 0.87 REFERENCE Greiner 1970 Bradley, Hack, Hoyermann & Wagner 1973 Overend, Paraskevopoulos & Cvetanovic 1975 Anderson & Stephens 1982 Tully, Ravishankara & Carr 1983 Baulch, Campbell & Saunders 1985 Tully & Droege 1985 Schmidt, et al."1985 (24) Bonilla & Anderson 1986 -12 Average rate= 1.13.35 x 10_12 = 1 x 10 excluding Overend et al. Ratios of reactivity of HO with n-Butane/Propane from relative rate studies 2.11 : Reference :Atkinson, Aschmann, :Carter, Winer & Pitts (1982) Calculated rate of HO + n-butane :(cm3 : -12 :2.15x10" **********************.** *.*.**.***.** ***** ****.***.***.** Based upon the relative reactivities of HO with n-butane to propane and the average absolute rate constant for the HO + propane reaction the calculated rate constant of 2.15 -12 3 x 10 cm for the HO n-Butane reaction is smaller than that determined by all direct measurements presented in table I, and in excellent agreement with the results of the present study.

PAGE 105

98 Conclusions Experimental results consistantly indicate the effect of secondary reactions in the study of the apparent bimolecular rate constant for the reaction of HO and n-butane. These results were obtained at the three different initial HO concentrations of approximately 2 x 1011 particles/cm3 2 x 12 3 12 3 10 particles/cm, and 8-16 x 10 particles/cm. This indi-cates that it is necessary to significantly change the initial HO concentration in order to see the effect of the secondary reactions of HO and butyl radicals. The effect of these secon-dary reactions was reflected in a change of the apparent bimolecular rate for the reaction of HO and n-butane of 26% to 36% when there was a change of an order of magnitude or more in the initial HO concentration. The apparent bimolecular rate for the reaction of HO and n-butane was 2.23.04 x 10-12 3 -12 cm /molec-sec at low HO conditions, 2.B5.05 x 10 cm3/molec-sec at high HO conditions, and. 3.07.15 to 3.13.23 x 10-12 cm3/molec-sec at very high HO conditions. Simulations performed to evaluate the rate constant for the secondary reaction of HO and butyl radicals indicate it ranges from 1.2 to 5 x 10-10 cm3/molec-sec which is con-sidered to be a fast reaction. The minimum value of this reaction has resulted from the exclusion of the reactions of NO an N02 with butyl radicals in the simulations, the maximum value resulted from the inclusion of these. The rate for the

PAGE 106

99 reactions of NO and N02 with butyl used in the simulations was that of NO and N02 with methyl radlcals. The simula-tions indicate that the rate for the reaction of HO radicals and n-butane can be up to 7% less than the apparent rate of -12 3 x 10 cm /molec-sec at low HO. The best value for the reaction of HO and n-butane at room temperature is 2.10.25 x 10-12 cm-3/molecules-sec. The simulations also indicated that the effect of the secondary reactions the apparent rates was dependent on the value used in the model for the secondary reaction of HO and butyl and not as depend-ent on the value used for the reaction of HO and n-butane. The results of the product trapping experiments to help elucidate the mechanism of HO and butyl radicals indi-cated the formation of products that involved butyl radicals. The kinetic results indicated the reaction of the butyl radi-cals with HO radicals. The kinetic results and mass spectrum indicated the influence of NO or N02 with the products or with the formation of products at very high HO. It is necessary that further research be performed to estimate the rate of reaction oroNO and N02 with butyl radicals and product formation of these. The influence of secondary reactions on the measurement of the apparent bimolecular rate of HO radicals and n-butane can be more easily seen when the initial HO concentra-tion is changed significantly and a fixed time method of

PAGE 107

100 measurement of HO decay is used, as the fixed time method used in the discharge flow experimental technique. This has made it more difficult to detect secondary reactions in experiments that use the flash photolysis technique, in which the HO decay is usually followed by a variable time. Also the initial HO concentrations is difficult to estimate, and the change in the initial HO concentration is based on changes in the quantity of the precursors of the HO radicals.

PAGE 108

101 BIBLIOGRAPHY (1) K. L. Demerjian, J. A. Kerr and J. G. Calvert, Adv. Environ. Sci. Tech., 4, 1 (1974) (2) J. G. Calvert, Environ. Sci. Technol., 10, 256 (1976) (3) F. P. Tully, M. L. Koszykowski and J. Stephen Binkley, Twentieth Int. Symp. Combustion, The Combustion Inst., 715 (1984) (4) J. W. Moore and R. G. Pearson, Kinetics and Mechanism, 3rd Ed., Wiley, 16 (1981) (5) R. Atkinson, Chern. Rev., 85, 91 (1985) (6) N. R. Greiner, J. Chern. Phys., 53, 1070 (1970) (7) E. D. Morris, and H Niki, J. Phys. Chern., 75, 3640 (1971) (8) F. Stuhl, Z. Naturforsch., 28A, 1383 (1973) (9) S. Gordon and W. Mulac, Int. J. Kinet., Symp. 1, 289 (1975) (10) R. A. Perry, R. Atkinson and J. N. Pitts, J. Chern. Phys., 64, 5314 (1976) (11) G. Paraskevopoulos and W. S. Nip, Can. J. Chern., 58, 2146 (1980) (12) L. G. Anderson and R. D. Stephens, Submitted for Publication, Int. J. Chern. Kinet., (1986) (13) L. G. Anderson and J. L. Bonilla, Proceedings of the Seventeenth Int. Symp. on Free Radicals, National Bureau of Standards Special Publication 716, 19 (1986) (14) F. P. Tully, J. Phys. Chern. in press, (1986) (15) V. Schmidt, G. Y. Zhu, K. H. Becker and E. H. Fink, Ber. Bunsen-Ges. Phys. Chern., 89, 321 (1985) (16) P. Gates, Graduate student at the Univ. of Co. at Denver, Comascal Computer Program, (1986) (17) J. A. and T. I. Qui kenden J. Chern. Ed., 6, 711 (1960)

PAGE 109

.102 (18) C. W. Gear, Collected Algorithms from CACM, 407, 1 (1969) (19) C. W. Gear, Comm. of the ACM, 14, 176 (1971) (20) R. J. Gelinas, J. Computational Phys., 9, 222 (1972) (21) W. J. Pitz, C. K. Westbrook, W. M. Proscia, F. L. Dryer, Twentieth Int. Symp. on Combust., The Combustion Inst., (22) W. B. Demore, J. J. Margitan, M. J. Molina, R. T. Watson, D. M. Golden, R. F. Hampson, M. J. Kurylo, C. J. Howard and A. R. Ravishankara, Chemical Kinetics and Photochemical Data for Use in Stratospheric Modeling, J.P.L. Pub. 85-37, 17, 1135 (1985) (23) R. Atkinson, S. M. Aschmann, W. P. L. Carter, A. M. Winer and J. N. Pitts, Int. J. Chern. Kinet., 781 (1982) (24) L. G. Anderson and J. L. Bonilla, studies performed for the reaction of HO and propane, (1986) (25) H. Adachi and N. Basco, Int. J. Chern. Kinet., 13, 367 (1981) (26) D. A. Parker and C. P. Quinn, J. C. S. Faraday, 72, 1952 (1976) (27) N. Washida, J. Chern. Phys., 73, 1665 (1980) (28) F. Yamada, I. R. Slagle and D. Gutman, Chern. Phys. Letters, 83,409 (1981) (29) L. G. Anderson and J. L. Bonilla, Seventhteen Informal Conf. on Photochemistry, Univ. of Co., Wed. 26 (1986) (30) F. Kaufman, Progress in Reaction Kinetics, 1, 10 (1961) (31) C. J. Howard, J. Phys. Chern., 83, 3 (1979) (32) P. J. Ogren, J. Phys. Chern .. 79, (33) R. R. Arnts, S. A. Meeks, Atmos. Environ., 15, 1643 (1981)

PAGE 110

Appendix 1 ACTUATE PROGRAM 50 LPRINT DATE$:LPRINT TIME$:LPRINT 55 CLS:COLOR 60 KEY OFF 65 DIM TITRATION(10,36):DIM 70 DIM RXTIME(8), X(10),Y(10),REJ(10) 75 DIM VOLTRDG(16),STDSUM(18) 80 DIM SIGNAL(10),ESIGNAL(10), PARAMETER(7,16),VOLTAGESET(6) 85 DIM FLAG$(Q),ESVOLTAGESET(6),EPFLOW(6)' 103 90 DIM 95 DIM BASELINE(10),EBASELINEC10),XPTC10),YPT(10) 100 LET SIGN$-"OFF" 105 110 REM "ACTUATE" ,A PROGRAM DESIGNED TO CONTROL KINETICS EXPERIMENTS 115 REM INVOLVING FLOURESCENCE PHOTON COUNTING. 120 REM 125 REM METHOD DISCHARGE / RESONANCE FLOURESCENCE 130 REM 135 REM INPUT: TO BE DETERMINED BY OPERATOR 140 REM REM OUTPUT: DATA & GRAPHICS 120 REM PROGRAMERS:JUAN L. BONILLA & PAUL GATES U.C.D. 150 REM 155 REM HARDWARE SETTING: JUAN L. BONILLA 160 REM 165 REM 170 175 ACTUATOR$-"UNACTIVE" 180 FOR TO 6:FLAG$(I)-"OFF":NEXT I 185 LET SIGN$-"OFF" 190 LET FBORN 1808 195 LET SBORN -200 REM BEGINNING THE MAIN BODY OF "EXECUTE". 205 REM CLOSING ALL FLOWCONTROLLERS 210 1-0 215 NU-63:VALVE-0:NUM-0 220 GOSUB 4680 225 REM ADDRESS 230 WIDTH 40:CLS 235 REM GOES TO THE FLOWCONTROLLER IDENTIFICATION SUBROUTINE GOSUB 870 LOCATE 10

PAGE 111

lOll Appendix 1 ACTUATE PROGRAM PRINT" DATA ENTRY DECISION MODE" 255 LOCATE 13 PRINT"DO YOU WISH TO ENTER INITIAL PARAMETERS?" 265 LOCATE 16 270 PRINT"" Y OR N?":K-l 275 GOSUB 915 280 IF AN$-MN" THEN 300 285 REM INTERACTIVE DATA ACQUISITION SUBROUTINE GOSUB 1095 295 IF AN$_"YM THEN 310 300 REM FILE DATA AQUISITION SUBROUTINE 305 GOSUB 1720 310 WIDTH 40:CLS 315 LOCATE 7 PRINT" PARAMETER ALTERATION" 325 LOCATE 10 330 PRINT" OR VERIFICATION ,MODE" 335 LOCATE 13 3110 PRINT" DO YOU WISH TO ALTER OR" 3115 LOCATE 16 PRINT" VERIFY ANY PARAMETERS?" 355 LOCATE 19 PRINT" Y OR N?":K-2 365 GOSUB 915 IF ANS$b "N" THEN 395 375 REM GOES TO PARAMETER PERUSAL SUBROUTINE. GOSUB 1790 385 REM DIRECTS TO FILE CREATION SUBROUTINE 390 GOSUB 1020 395 FOR 1-"1 TO 6 1100 REM FLOW CONTROLLER VOLTAGE AND VOLTAGE ERROR CALCULATION SUBROUTINE 405 IF PARAMETER (1,1) 0 THEN VOLTAGESET(I)-O 1110 IF PARAMETER (1,1)-0 THEN 420 415 GOSUB 2405 LPRINT"FLOW SET "iPARAMETER(I,l)i" VOLTAGESET "iVOLTAGESET(I) 1125 NEXT I 1130 WIDTH 1l0:CLS 1135 IF SIGN$-"ON" THEN 555 4110 LOCATE 9 1l1l5 PRINT" MODE"

PAGE 112

Appendix ACTUATE PROGRAM 450 455 460 465 470 475 480 485 490 495 500 505 510 515 520 525 530 535 540 545 550 555 560 565 570 575 580 585 590 595 600 605 610 615 620 625 630 635 640 645 LOCATE 12 PRINT" ARE ALL GASES TURNED ON AND" LOCATE 15 PRINT" REGULATED TO 5 PSI AT THE CYLINDER?" LOCATE 18 PRINT" Y OR N?":K-3 GOSUB 915 IF YN$ "N" THEN 440 CLS CLS LOCATE 9 PRINT" LOCATE 12 PRINT" LOCATE 15 PRINT" LOCATE 17 PRINT" GOSUB 915 PREPARATION MODE" ARE THE HIGH VOLTAGE?" MICROWAVE POWER SOURCES ON?" Y OR N?":K-4 IF MWO$."N" THEN 495 LOCATE 13 REM FLOW CONTROLLER SETTING ROUTINE" FOR I-1 TO 6 GOSUB 4460 NEXT I LOCATE 12 PRINT" TURNING ON THE FLOW CONTROLLERS" REM THIS SECTION TURNS ON THE FLOWCONTROLLERS IF HFC-1 THEN Dc1 IF HFC-2 THEN 0-2 IF HFC-3 THEN 0-4 IF HFC-4 THEN 0-8 IF SIGN$-"ON" THEN NU-47-D REM BEGIN I-LOOP FOR I1 TO 6 REM OMITS NONFLOW FLOW CONTROLLERS IF PARAMETER(I.1)-O THEN 690 REM SETS FLAG LET FLAG$(I)-"ON" REM BYPASSES ADJUSTFLOW FOR 1-5 105

PAGE 113

Appendix 1 ACTUATE PROGRAM 650 IF (SIGN$-"ON") AND (I.5) THEN 690 655 IF THEN 670 660 REM ADJUST PRESSURE 665 GOSUB 4385 670 REM TURNS ON FLOW CONTROLLERS 675 IF (IHFC) AND (SIGN$."ON") THEN 690 680 GOSUB 4720 685 REM END I-LOOP 690 NEXT I 695 LEVEL$"DOWN":R-1 700 IF (ACTUATOR$."ACTIVE") THEN GOSUB 5840 705 FOR N 1 TO 500001:NEXT N 710 IF SIGN$*"ON" THEN 750 715 CLS 720 LOCATE 13 725 PRINT" ARE THE MICROWAVE CAVITIES TUNED?" 730 LOCATE 15 735 PRINT" Y OR N?":K-5 740 GOSUB 915 745 IF THEN 725 750 CLS 755 LOCATE'3 760 LET RO" 765 IF SIGN$-"ON" THEN 780 770 PRINT" STABILIZATION SUBROUTINE" 775 GOSUB 2550 780 LOCATE 10 785 PRINT"" TITRATION SUBROUTINE" 790 GOSUB 4920 795 LOCATE 13 800 PRINT"" OUTPUT SUBROUTINE" 805 GOSUB 5425:GOSUB 6155:GOSUB 6230" 810 IF ACTUATOR$."ACTIVE" THEN GOSUB 6345 815 GOSUB 3310 820 WIDTH 40:CLS 825 LOCATE 12 830 PRINT"" DO YOU DESIRE TO RUN" 835 LOCATE 1 4 840 PRINT"" ANOTHER" EXPERIMENT?":K" 845 GOSUB 915 106

PAGE 114

Appendix 1 850 IF ANSW$-"N" THEN 865 855 LET SIGN$-"ON" 860 IF SIGN$-"ON" THEN 300 865 END ACTUATE PROGRAM 870 REM FLOW CONTROLLER IDENTIFICATION SUBROUTINE 875 REM THIS SECTION IDENTIFIES THE FLOWCONTROLLERS BY CODENAME 880 LET CN$(1 )-"ALPHA" 885 LET CN$(2)-"BETA" 890 LET CN$(3)-"GAMMA" 895 LET CN$(4)-"DELTA" 900 LET CN$(5)-"EPSILON" 905 LET CN$(6)-"IOTA" 910 RETURN 915 REM DECISION SUBROUTINE 920 REM ANSWERS THE QUESTION 925 INPUT DEC$ 930 REM ALLOWS FOR BOTH UPPER AND LOWER CASE ENTRY 935 IF DEC$"y"THEN DEC$-"Y" 940 IF DEC$-"n" THEN DEC$-"N" 945 REM PROTECTS AGAINST INCORRECT DATA ENTRY 950 IF (DEC$><"N") AND (DEC$><"Y") THEN 925 955 REM ASSIGNATION OF ANSWER TO APPROPRIATE QUESTION 960 IF THEN AN$ .. DEC$ 965 IF K-2 THEN ANS$-DEC$ 970 IF K-3 THEN YN$-DEC$ 975 IF K-4 THEN MWO$-DEC$ 980 IF K-5 THEN MCT$-DEC$ 985 IF K-6 THEN IFLP$-DEC$ 990 IF K-7 THEN AVSR$-DEC$ 995 IF K-8 THEN ICNFLP$-DEC$ 1000 IF K-9 THEN NAVSR$-DEC$ 1005 IF K-l0 THEN SFL$-DEC$ 1010 IF THEN ANSW$-DEC$ 1015 RETURN 1020 REM "FCSR",FILE CREATION SUBROUTINE 1025 REM THIS PROGRAM CREATES THE OPERATIONS DATAFILE,OPERATE. '030 OPEN "DATAFILE" FOR OUTPUT AS (11 1035 REM THIS SECTION READS DATA IN TO THE DATA FILE. 1040 REM BEGINING I-LOOP 1045 FOR TO 7 107

PAGE 115

108 Appendix ACTUATE PROGRAM 1050 REM BEGINNING OF Q-LOOP 1055 FOR TO 16 1060 PRINT #1;PARAMETER(I,Q) 1065 REM END OF Q-LOOP 1070 NEXT Q 1075 REM END OF I-LOOP 1080 NEXT I 1085 CLOSE #1 1090 RETURN 1095 REM DIRECT ENTRY MODE SUBROUTINE 1100 LET AVSR$-"Y" 1105 REM THIS I-LOOP ASSIGNS THE FLOW CONTROLLER NAMES TO THE PROPER 1110 REM STORAGE LOCATIONS. 1115 FOR I-1 TO 6:CLS 1120 LOCATE 10 "25 PRINT" DO YOU WISH TO ENTER DATA FOR" 1130 LOCATE 13 1135 PRINT" THE ";CN$(I);" FLOW CONTROLLER?" LOCATE 16 1145 PRINT" Y OR N?":K.l0 "50 GOSUB 915 1155 IF 1500 1160 WIDTH 80:CLS 1165 PRINT"ENTER THE INITIAL FLOW FOR THE ";CN$(I);" FLOW CONTROLLER" 1170 PRINT"IN CCM." 1175 INPUT PARAMETER(I,l) 1180 IF AVSR$-"N" THEN 1860 1185 PRINT"ENTER THE INCREMENT BY WHICH THE ";CN$(I);" FLOW CONTROLLER" 1190 95 1200 1205 1 21 0 1215 1220 1225 1230 1235 1240 1245 PRINT"IS TO BE VARIED IN CCM." INPUT PARAMETER(I,2) IF AVSR$-"N" THEN 1860 PRINT"ENTER THE CALIBRATION SLOPE FOR THE ";CN$(I);" FLOW" PRINT"CONTROLLER FOR THE SET VOLTAGE." INPUT PARAMETER(I,3) IF AVSR$-"N" THEN 1860 PRINT"ENTER THE STATISTICAL ERROR IN SLOPE FOR THE ";CN$(I);" FLOW PRINT"CONTROLLER FOR THE SET VOLTAGE." INPUT PARAMETER(I,4) IF AVSR$-"N" THEN 1860 PRINT"ENTER THE PROPAGATION ERROR IN SLOPE FOR THE ";CN$(I);" FLOW

PAGE 116

Appendix 1250 1255 1260 1265 1270 1275 1280 1285 1290 1295 1300 1305 1310 1315 320 1325 1330 1335 1340 1345 1350 1355 1360 1365 1370 1375 1380 1385 1390 1395 1400 1405 1410 1415 1420 1425 1430 1435 1440 1445 ACTUATE PROGRAM PRINT"CONTROLLER FOR THE SET VOLTAGE." INPUT PARAMETER(I,5) IF AVSR$-"N" THEN 1860 109 PRINT"ENTER THE CALIBRATION SLOPE FOR THE ";CN$(I);" FLOW" PRINT"CONTROLLER FOR THE READ VOLTAGE." INPUT PARAMETER(I,6) IF AVSR$-"N" THEN 1860 PRINT"ENTER THE STATISTICAL ERROR IN SLOPE FOR THE "; CN$ (I) ;" FLOW PRINT"CONTROLLER FOR THE READ VOLTAGE." INPUT PARAMETER(I,7) IF AVSR$-"N" THEN 1860 PRINT"ENTER THE PROPAGATION ERROR IN SLOPE FOR THE ";CN$(I);" FLOW PRINT"CONTROLLER FOR THE READ VOLTAGE." INPUT PARAMETER(I,8) IF AVSR$-"N" THEN 1860 CLS:PRINT"ENTER THE CALIBRATION INTERCEPT FOR THE ";CN$(I);" FLOW" PRINT"CONTROLLER FOR THE SET VOLTAGE." INPUT PARAMETER(I,9) IF AVSR$-"N" THEN 1860 PRINT"ENTER THE STATISTICAL ERROR IN THE INTERCEPT FOR THE "eN$(I) PRINT"FLOW CONTROLLER FOR THE SET VOLTAGE." INPUT PARAMETER(I,10) IF AVSR$-"N" THEN ,860 PRINT"ENTER THE PROPAGATION ERROR IN THE INTERCEPT FOR THE ";CN$(I PRINT"FLOW CONTROLLER FOR THE SET VOLTAGE." INPUT PARAMETER(I,'1) IF AVSR$-"N" THEN 1860 PRINT"ENTER THE CALIBRATION INTERCEPT FOR THE ";CN$(I);" FLOW" PRINT"CONTROLLER FOR THE READ VOLTAGE." INPUT PARAMETER(I,12) IF AVSR$-"N" THEN ,860 PRINT"ENTER THE STATISTICAL ERROR IN THE INTERCEPT FOR THE ";CN$(I PRINT"FLOW CONTROLLER FOR THE READ VOLTAGE." INPUT PARAMETER(I,13) IF AVSR$-"N" THEN 1860 PRINT"ENTER THE PROPAGATION ERROR IN THE INTERCEPT FOR THE ";CN$(I PRINT"FLOW CONTROLLER FOR THE READ VOLTAGE." INPUT PARAMETER(I,14) IF AVSR$-"N" THEN 1860 REM FLOWCONTROLLER GAS IDENTIFICATION SUB ROUTINE

PAGE 117

110 Appendix 1 ACTUATE PROGRAM GOSUB IF AVSR$-"N" THEN 1860 PRINT "ENTER CONCENTRATION OF THE "iGAS$(I)i" IN THE TANK "i INPUT PARAMETER(I.16) IF AVSR$-"N" THEN 1860 PRINT "THE FLOW CONTROLER STATUS FLAG IS "iFLAG$(I) INPUT "HIT RETURN TO CONTINUE";RET$ 1485 IF AVSR$-"N" THEN 1860 1490 WIDTH REM END OF I-LOOP 1500 NEXT I 1505 REM THIS SECTION INPUTS NON FLOW PARAMETERS. 1510 WIDTH 80:CLS 1515 PRINT"ENTER THE DURATION OF COUNTING INTERVALS IN SECONDS." 1520 INPUT PARAMETER(7.1) 1525 IF NAVSR$-"N" THEN 2090 1530 PRINT"ENTER THE UPPPER LIMIT IN THE GRAPH Y AXIS." 1535 INPUT PARAMETER(7.2) IF NAVSR$-"N" THEN 2090 1545 PRINT "ENTER MODE OF COUNTING' OR 2 "i 1550 INPUT PARAMETER(7.3) 1555 IF NAVSR$-"N" THEN 2090 1560 PRINT"ENTER THE NUMBER OF niE FLOW CONTROLER OF N02." 1565 INPUT PARAMETER(7.4) 1570 IF NAVSR$-"N" THEN 2090 1575 PRINT"ENTER THE NUMBER OF THE FLOW CONTROLER FOR HYDROGEN SOURCE." 1580 INPUT PARAMETER(7.5) 1585 IF NAVSR$-"N" THEN 2090 1590 PRINT"ENTER THE NUMBER OF TIMES FOR THE COUNTING CICLE -> "i 1595 REM 1600 INPUT PARAMETER(7,6) 1605 IF NAVSR$-"N" THEN 2090 1610 PRINT"ENTER THE STABILIZATION CRITERIA IN PERCENT J --> "i 1615 INPUT PARAMETER(7.7) 1620 PARAMETER(7.7)-PARAMETER(7.7)/100 1625 IF PARAMETER(7.7) l' THEN GOTO 1610 1630 IF NAVSR$-"N" THEN 2090 1635 PRINT"ENTER THE WAITING TIME FOR THE BEFORE OFF CYCLES." INPUT PARAMETER(7.8) IF NAVSR$-"N" THEN 2090

PAGE 118

111 Appendix 1 ACTUATE PROGRAM 1650 PRINT"ENTER THE BAROMETRIC PRESSURE." 1655 INPUT PARAMETER(7,9) 1660 IF NAVSR$-"N" THEN 2090 1665 PRINT"ENTER THE DESIRED OPERATING PRESSURE FOR THE SYSTEM." 1670 INPUT PARAMETER(7,10) 1675 IF NAVSR$-"N" THEN'2090 1680 PRINT"ENTER THE NUMBER OF TITRATION REPETITIONS." 1685 INPUT PARAMETER(7,11) 1690 IF NAVSR$-"N" THEN 2090 1695 IF ACTUATOR$-"UNACTIVE" THEN ACTUATOR$-"ACTIVE" ELSE ACTUATOR$-"UNACTIVE" 1700 IF ACTUATOR$-"ACTIVE" THEN PARAMETER17,11) 8 1705 IF NAVSR$-"N" THEN 2090 1710 REM END OF THE NON-FLOW PARAMETERS 1715 RETURN 1720 REM "FRDGSR",FILE READING SUBROUTINE. 1725 REM THIS SUBROUTINE READS THE DATA FILE "OPERATE". 1730 OPEN "DATAFILE" FOR INPUT AS 1735 REM BEGIN I-LOOP FOR 1-1 TO 7 REM BEGIN Q-LOOP 1750 FOR TO 16 1755 INPUT tll,PARAMETER(I,Q) 1760 REM END' OF Q-LOOP 1765 NEXT Q 1770 REM END OF I-LOOP 1775 NEXT I 1780 CLOSE #1 1785 RETURN 1790 REM "IPLSR",INPUT PARAMETER LISTING SUBROUTINE 1795 REM THIS SUBROUTINE ALLOWS PERUSAL OF INPUT PARAMETERS 1800 REM BEGIN I-LOOP 1805 FOR 1-1 TO 6:CLS 1810 REM SECTION ALLOWS PERUSAL OF INDIVIDUAL FLOW CONTROLLER PARAMETERS." 1815 LOCATE 10 1820 PRINT" DO YOU WISH TO SEE THE" 1825 LOCATE 13 1830 PRINT" PARAMETERS FOR FLOW CONTROLLER ";CN$(I);"?" 1835 LOCATE 16 PRINT" Y OR N?" :K-6 1845 GOSUB 9

PAGE 119

112 Appendix ACTUATE PROGRAM 1850 1855 1860 1865 REM STATEMENT CONTROLLS VISUALIZATION OF FLOW CONTROLLER IF IFLP$-"N" THEN 2035 '870 1875 1880 WIDTH 80:CLS PRINT"THE INPUT PARAMETERS FOR THE ";CN$(I)j" FLOW CONTROLLER." PRINT"01,THE INITIAL FLOW IS "jPARAMETER(I,1);" CCM." PRINT"fI2, THE INCREMENTATION OF THE FLOW IS "; PARAMETER (I, 2) l" CCM PRINT"fI3, THE CALIBRATION SLOPE VERSUS THE SET VOLTAGE IS l "PARAMET ER (1,3) 1885 PRINT"Oq,THE STATISTICAL .ERROR IN SLOPE VERSUS THE SET VOLTAGE IS "jPARAMETER(I,q)j" .n 1890 PRINT"05,THE PROPAGATION ERROR IN SLOPE VERSUS THE SET VOLTAGE IS ";PARAMETER(I,S);" ." 1895 PRINT"fI6, THE CALIBRATION SLOPE VERSUS THE READ VOLTAGE IS "jPARA."1ET ER(I,6);" ." 1900 PRINT"07,THE STATISTICAL ERROR IN SLOPE VERSUS THE READ VOLTAGE IS M;PARAMETER(I,7)j" ." 1905 PRINT"08,THE PROPAGATION ERROR IN SLOPE VERSUS THE READ VOLTAGE IS l PARAMETER ( I ,8) j" ." 1910 PRINT"U9,THE CALIBRATION INTERCEPT VERSUS THE SET VOLTAGE IS "jPARA MET ER ( I 9) 1915 PRINT"fll 0, THE STATISTICAL ERROR IN INTERCEPT VERSUS THE SET VOLTAGE IS "jPARAMETER(I,10)j" ." 1920 PRINT"Ull ,THE PROPAGATION ERROR IN INTERCEPT VERSUS THE SET VOLTAGE "IS "jPARAMETER(I,11);" ." 1925 PRINT"012,THE CALIBRATION INTERCEPT VERSUS THE READ VOLTAGE IS "lPA RAMETER(I,12)j" ." 1930 PRINT"U13,THE STATISTICAL ERROR IN INTERCEPT VERSUS THE READ VOLTAG IS "jPARAMETER(I,13)j" ." 1935 PRINT""1q,THE PROPAGATION ERROR IN INTERCEPT VERSUS THE READ VOLTAG E IS ." 1940 REM THIS SECTION CONVERTS PARAMETER(I,15)INTO A STRING VARIABLE 19q5 REM WHICH IDENTIFIES THE GAS USED FOR THE ITH 'FLOW CONTROLLER. 1950 IF PARAMETER(I,1S)-1 THEN GAS$(I)-"HELIUM" 1955 IF PARAMETER(I,15)-2 THEN GAS$(I)-"HYDROGEN-HELIUM" 1960 IF PARAMETER(I,15)-3 THEN GAS$(I)-"NITROGEN DIOXIDE" 1 965 IF PARAMETER ( 1,15) -q THEN GAS$( I) -"ALKANE 1970 IF PARAMETER(I,15)-5 THEN GAS$(I)-"SULFUR DIOXIDE" 1975 IF PARAMETER(I,15)-6 THENGAS$(I)-"OXYGEN" 1980 PRINT"015,THE GAS USED BY THE FLOW CONTROLLER IS "jGAS$(I)j" ." 1985 PRINT"016, THE"CONCEN. IN THE TANK OF "jGAS$(I)j" MIXTURE IS "jPARAMETER(I 6) 1990 1995 !OOO 2005 2010 2015 2020 2025 2030 2035 2040 20q5 PRINT"tl17 THE STATUS FLAG IN THE FLOW CONTROLER IS "lFLAG$(I) REM" THIS SECTION CONTROLLS ACCESS TO,THE VERIFICATION SUBROUTINE. TFLOW-PARAMETER(5,1) PRINT"ARE THE"ABOVE PARAMETERS CORRECT? Y OR N?":K-7 GOSUB 915 REM ROUTES TO VERIFICATION SUBROUTINE IF AVSR$-"N" THEN GO TO 2210 WIDTH qO:CLS REM END OF I-LOOP NEXT I REM THIS SECTION ALLOWS PERUSAL OF NONFLOW PARAMETERS. CLS

PAGE 120

113 Appendix 1 ACTUATE PROGRAM 2050 LOCATE 10 2055 PRINT" DO YOU WISH TO CHANGE OR" 2060 LOCATE 13 2065 PRINT" VERIFY THE NON FLOW PARAMETERS?" 2070 LOCATE 16 2075 PRINT" Y OR N?":K-8 2080 GOSUB 915 2085 IF ICNFLP$-"N" THEN 2400 2090 WIDTH SO:CLS 2095 PRINT"1.THE COUNTING INTERVAL DURATION IS "iPARAMETER(7,1);" SECOND S." 2100 PRINT"tl2.THE UPPER LIMIT VALUE IN THE GRAPH Y AXIS IS ";PARAMETER(7 .2); 2105 2110 ER(7.4);" PRINT"(t3.THE MODE OF COUNTING IS "jPARAMETER(7.3) PRINT" THE N02 GAS IS FLOWI NG TROUGHT THE FLOW CONTROL i PARAMET 2115 PRINT"(t5. THE HYDROGEN SOURCE GAS IS FLOWING TROUGHT THE CONTROLER "lPARAMETER(7.5);" 0" 2120 PRINT"tl6.THE NUMBER OF TIMES FOR THE COUNTING CICLE IS "i 2125 PRINT PARAMETER(7.6);" ." 2130 PRINT"f.l7.THE STABILIZATION CRITERIA IS EQUAL TO "; 2135 PRINT PARAMETER (7.7) ;" ." 2140 PRINT"ftS.THE WAITING IS APROX. ";PARAMETER(7.S);" SECS. IN OFF." 2145 PRINT"(t9.THE BAROMETRIC PRESSURE IS ";PARAMETER(7.9)j" MM HG." 2150 PRINT"ft10.THE EXPERIMENTAL PRESSURE IS ";PARAMETER(7.10);" MM HG." 2155 PRINT"(t11 THE NUMBER OF REPETITIONS IA TITRATION IS ";PARAMETER(7. 1 i 2160 PRINT THE ACTUATOR IS "jACTUATOR$ 2165 REM" THIS SECTION PROVIDES ACCESS TO THE VERIFICATION SUBROUTINE, 2170 PRINT"ARE THE NONFLOW PARAMETERS CORRECT? Y OR N?":K-9 2175 GOSUB 915. 2180 NFC-PARAMETER(7.4) 2185 HFC-PARAMETER(7.5) 2190 REM THIS SECTION ACCESSES THE VERIFICATION SUBROUTINE 2195 IF NAVSR$-"N"THEN GOTO 2210 2200 IF NAVSR$-"Y" THEN 2400 2205 REM END OF PARAMETER PERUSAL 2210 REM DATA VERIFICATION AND ALTERATION SUBROUTINE 2215 PRINT"ENTER THE NUMBER OF THE PARAMETER TO BE ALTERED." 2220 INPUT VERIFY$ 2225 REM THIS SECTION PROTECTS THE PROGRAM FROM VERIFICATION INPUT ERROR 2230 IF (1-7) THEN GOTO 2335 2235 IF (1<7) THEN GOTO 2245 2240 REM THIS SUBROUTINE ROUTES DATA FOR CORRECTIONS 2245 IF VERIFY$-",,, THEN 1165

PAGE 121

114 Appendix 1 ACTUATE PROGRAM 2250 IF VERIFY$"2" THEN 1185 2255 IF VERIFY$"3" THEN 1205 2260 IF VERIFY$-"4" THEN 1225 2265 IF VERIFY$"5" THEN 1245 2270 IF VERIFY$"6" THEN 1265 2275 IF VERIFY$"7" THEN 1285 2280 IF VERIFY$"8" THEN 1305 2285 IF VERIFY$"9" THEN 1325 2290 IF VERIFY$-"10" THEN"1345 2295 IF VERIFY$"11" THEN '365 2300 IF VERIFY$"12" THEN 1385 2305 IF VERIFY$"13" THEN 1405 2310 IF VERIFY$-"14" THEN 1425 2315 IF VERIFY$"15" THEN 1445 2320 IF VERIFY$-"16" THEN 1Q60 2325 IF VERIFY$-"'7" THEN 1475 2330 GOTO 1860 2335 IF VERIFY$"1" THEN 1515 2340 IF VERIFY$"2" THEN 1530 2345 IF VERIFY$"3" THEN 1545 2350 IF VERIFY$"4" THEN 1560 2355 IF VERIFY$"5" THEN 1575 2360 IF VERIFY$"6" THEN 1590 2365 IF VERIFY$"7" THEN 1610 2370 IF VERIFY$"8" THEN 1635 2375 IF VERIFY$"9" THEN 1650 2380 IF VERIFY$-"10" THEN"1665 2385 IF VERIFY$-"11" THEN 1680 2390 IF VERIFY$-"12" THEN 1695" 2395 GOTO 2090 21100 RETURN 2405 REM FLOW TO VOLTAGE, ERROR INCLUSIVE SUBROUTINE 2410 REM SETS COMMON FLOW FOR VARYING FLOW SOURCE INPUT 2415 REM VOLTAGE FROM FLOW PARAMETER INPUT 2420 LET VOLTAGESET(I)(PARAMETER(I,1)-PARAMETER(I,9/PARAMETER(I,3) 2425 REM THE ERROR IN VOLTAGESET(I) 2430 REM THE PARTIAL DERIVATIVE OF THE VOLTAGE SET WITH RESPECT TO THE 2435 REM SLOPE. 24110 LET DVDM-(PARAMETER(I,1)-PARAMETER(I,9)/PARAMETER(I,3A2 2445 REM THE PARTIAL DERIVATIVE OF THE VOLTAGE SET WITH RESPECT TO THE

PAGE 122

Appendix 1 ACTUATE PROGRAM REM INTERCEPT. LET DVDB-(PARAMETER(I,1)-1)/PARAMETER(I,3) REM THE STATISTICAL ERROR' 115 2455 2465 LET ESVOLTAGESET(I)-(DVDMA 2 PARAMETER(I,4)A2 + DVDBA 2 PARAMETE R(I, 10)A2)A.5 REM THE PROPAGATIONAL ERROR 2475 LET EPVOLTAGESET(I)-(DVDMA 2 PARAMETER(I,5)A2 + DVDB2 R(I,11)A2)A.5 RETURN 2485 REM VOLTAGE TO FLOW, ERROR INCLUSIVE SUBROUTINE REM FLOW FROM VOLTAGE INSTRUMENTAL INPUT 2495 LET FLOW(I)PARAMETER(I,6) VOLTRDG(I) + PARAMETER(I,12) 2500 REM THE ERROR IN FLOW(I) REM THE PARTIAL DERIVATIVE OF FLOW(I) WITH .RESPECT TO THE SLOPE. LET DFDM-VOLTRDG(I) 2515 REM THE PARTIAL DERIVATIVE OF FLOW(I) WITH RESPECT TO THE INTERCEPT 2520 LET DVDB-1 2525 REM THE STATISTICAL ERROR LET ESFLOW(I)(DFDMA 2 PARAMETER(I,7)A2 +DVDBA2 PARAMETER(I,13 A .5 2535 2540 A .5 2545 RETURN REM THE PROPAGATIONAL ERROR LET EPFLOW(I)(DFDMA 2 PARAMETER(I,8)A2 +DVDBA2 PARAMETER(I,14 2550 REM STABILIZATION SUBROUTINE 2555 DIM SUMBAR(4,18),SIG(4,18),SIGMA(4,18),STORAGE(25,18) REM MAIN SUBROUTINE CONTROL BODY' 2565 REM ARM AND START COUNTERS A AND B 2570 GOSUB 3845 2575 REM BEGIN N-LOOP CONTROLLING 4 MINUTE REPETITIONS. 2580 FOR N -1 TO 4 2585 REM BEGIN Z-LOOP,CONTROLLING 60 SECOND REPETITIONS. 2590 'FOR Z1 TO 6 2595 REM COUNTING,READING AND STORAGE SUBROUTINE 2600 GOSUB 2860 REM ARM AND START COUNTERS A AND B SUBROUTINE GOSUB 3845 2615 REM END OF Z-LOOP 2620 CLS 2625 2630 2635 2640 2645 NEXT Z LET STABILIZATION$ "Y" REM ROUTES TO STATISTICAL SUBROUTINE GOSUB REM ROUTES AROUND THE STABILIZATION REQUIREMENTS SUBROUTINE

PAGE 123

116 Appendix ACTUATE PROGRAM 2650 IF N<4 THEN 2760 2655 REM STABILIZATION REQUIREMENTS SUB ROUTINE 2660 GOSUB 3195 2665 REM ROUTES"AROUND NONSTABILIZATION ROUTING 2670 IF STABILIZATION$-"Y" THEN 2765 2675 REM ROUTING FOR NONSTABILIZATION SUBROUTINE 2680 REM BEGIN I-LOOP 2685 FOR 1 TO 18 2690 IF 1>6 THEN'2705 2695 IF (1<7) AND (PARAMETER(I,1)-0) THEN 2730 2700 REM ELIMINATES FIRST N 2705 SUMBAR(l,I)SUMBAR(2,'I):SIGMA(1,I)SIGMA(2,I) 2710 SUMBAR(2,I)-SUMBAR(3,I):SIGMA(2,I)SIGMA(3,I) 2715 SUMBAR(3,I)-SUMBAR(4,I):SIGMA(3,I)SIGMA(4,I) 2720 SIG(1,I)-SIG(2,I):SIG(2,I)-SIG(3,I):SIG(3,I)-SIG(4,I) 2725 REM END OF I-LOOP 2730 NEXT I 2735 REM SENDS PROGRAM TO ARM AND INIITIALIZE COUNTERS A & B 2740 GOSUB 3845 2745 REM ROUTING THE PROGRAM FOR NON-STABILIZATION CONDITIONS 2750 GOTO 2590 2755 REM END OF N-LOOP 2760 NEXT N 2765 REM SUMATION LOOP 2770 REM BEGIN I-LOOP 2775 FOR 1 TO 18 2780 LET SSUM(I)-O 2785 LET ESUM(I)-O 2790 REM BEGIN N-LOOP 2795 FOR 1 TO 4 2800 LET SSUM(I)-SSUM(I)+SUMBAR(N,I) 2805 LET ESUM(I)-ESUM(I)+SIGMA(N,I)A2 2810 REM END N-LOOP 28'5 NEXT N 2820 LET SSUM(I)-SSUM(I)/4 2825 LET ESUM(I)-(ESUM(I)A.5)/4 2830 REM END I-LOOP 2835 NEXT I 2840 LPRINT "COMPLETED ";FLAG$(4);" ";SSUM(17);"+I";ESUM(17);" COUNTS" 2845 REM END OF STABILIZATION SUBROUTINE MAIN BODY

PAGE 124

Appendix ACTUATE PROGRAM ERASE SUMBAR,SIG,SIGMA,STORAGE RETURN 117 2850 2855 2860 2865 2870 2875 REM COUNTING, READING, AND STORING STABILIZATION PARAMETERS SUBROUTINE REM BEGIN I-LOOP,I IDENTIFING ONE DEVICE GROUP OF PARAMETERS 2880 2885 2890 W-2 WIDTH 40:CLS LOCATE 11 PRINT TAB(20);R LOCATE 13 PRINT TAB(1S);N;",";Z LOCATE 15 PRINT TAB(20) ;FLAG$(4) FOR I-1 TO 1.8 STORAGE(Z,I)O REM REM 2895 2900 2905 2910 2915 2920 2925 2930 2935 2940 2945 2950 REM SKIPS "FLOW CONTROLLERS WHICH HAVE NO FLOW PARAMETER. 2955 2960 2965 2970 2975 2980 2985 2990 2995 3000 3005 IF I > 6 THEN 2945 IF (1<7) AND (PARAMETER(I,1)-0) THEN 3005 REM READS PARAMETER VOLTRDG(I) IF I < 17 THEN GOSUB 4600 REM READS COUNTS IF I 16 THEN GOSUB 4235 REM REM REM STORES DATA REM IF I < 17 THEN STORAGE(Z,I)-VOLTRDG(I) IF I-17"THEN STORAGE(Z,I)-ACOUNTER IF 1-18 THEN STORAGE(Z,I)-BCOUNTER REM END I-LOOP NEXT I 3010 RETURN 3015 REM STABILIZATION STATISTICAL TREATMENT SUBROUTINE 3020 REM THIS SUBTROUTINE SEPARATES THE FLOW CONTROLERS 3025 REM REMAINING DEVICES 3030 REM BEGIN I-LOOP 3035 FOR I 1 TO 18 3040 LET STDSUM(I)-O 3045 LET SUMBAR(N,I)-O FROM THE

PAGE 125

113 Appendix 1 ACTUATE PROGRAM 3050 LET SIGMA(N,I)-O 3055 LET SUM(I)-O 3060 REM BEGINS THE Z-LOOP 3065 FOR Z -1 TO 6 3070 REM CALCULATES SUM(I) 3075 LET SUM(I) SUM(I)+ STORAGE(Z,I) 30aO REM END 3085 NEXT Z 3090 REM CALCULATES SUHBAR(N,I) 3095 LET SUMBAR(N,I) SUM(I)/6 3100 REM CALCULATES STD. DEV. 3105 REM BEGIN Z-LOOP 3110 FOR Z -1 TO 6 3115 REM CALCULATES STDSUM(I) 3120 STDSUM(I) STDSUM(I) + (SUMBAR(N,I)-STORAGE(Z,IA2 3125 REM END Z-LOOP NEXT Z 3135 SIGMA(N,I) 3140 REM SETS STD. DEV. TO A MINIMUM VALUE 3145 SIG(N,I)-SIGMA(N,I) 3150 IF (SIG(N,I) < .025) THEN SIG(N,I)-.025 3155 IF (1<6) AND (SIG(N,I) .075) THEN STABILIZATION$-"N" 3160 IF (I-17) AND SIG(N,I) (.1'SUMBAR(N,I THEN STABILIZATION$a "N" 3165 REM SETS THE COUNTERS SIG(N,I) TO OF TOTAL COUNT 3170 IF (1-17) OR (I-la) THEN SIG(N,I).02SUMBAR(N,I) 3175 IF(I<7) AND (SUMBAR(N,I)<.05) THEN SUMBAR(N,I)-O 31aO IF(1<7) AND (SUMBAR(N,I)<.05) THEN SIG(N,I)-O 3185 NEXT I RETURN 3195 REM STABILIZATION REQUIREMENTS SUBROUTINE 3200 REM I-LOOP 3205 REM OOTO 6540 3210 FOR 1 TO 5 3215 IF (1)6) AND (1<11) THEN 3290 IF (1)12) AND (1(17) THEN 3290 3225 IF I > 6 THEN 3235' 3230 IF I <6 AND PARAMETER(I,1)-0 THEN 3290 3235 REM M-LOOP FOR M1 TO 4 3245 REM J-LOOP

PAGE 126

Appendix 1 ACTUATE PROGRAM 3250 FOR TO 4 3255 IF SUMBAR(M,I)-SIG(M,I> SUMBAR(J,I STABILIZATION$="N" 3260 IF SUMBAR(M,I)+SIG(M,I
PAGE 127

120 Appendix ,1 ACTUATE PROGRAM 3450 3455 3460 3465 3470 3475 3480 3485 3490 3495 3500 3505 3510 3515 3520 3525 3530 3535 3540 3545 3550 3555 3560 3565 3570 3575 3580 3585 3590 3595 3600 3605 3610 3615 3620 3625 3630 3635 3640 IF 0+80 3645 LINE (80,43)-(85,43): LINE (555,43)-(560,43) LINE (80,59)-(85,59): LINE (555,59)-(560,59) LINE (80,75)-(85,75): LINE LINE (80,91 )-(85,91): LINE (555,91 )-(560,91) LINE (80,107)-(85,107): LINE (555;107)-(560,107) LINE (80,123)-(85,123): LINE (555,123)-(560,123) LINE (80,139)-(85,139): LINE (555,139)-(560,139) LINE (80,154)-(85,154): LINE(555,154)-(560,15 4 ) LINE (128;11)-(128;15): LINE (128;166)-(128;171) LINE (176,11 )-(176,15): LINE (176,166)-(176,171) LINE (224,11 )-(224,15): LINE (224,166)-(224,171) LINE (272,11 )-(272,15): LINE (272,166)-(272,171) LINE (320,11)-(320,15): LINE (320,166)-(320,171) LINE (368,11)-(368,15): LINE (368,166)';'(368,171) LINE (416,11)-(416,15): LINE (416,166)-(416,171) LINE (11611,1,,-(11611,15): LINE (4611,166)-(4611,111) LINE (512,11 )-(512,15): LINE (512,166)-(512,111) REM THIS SECTION LABELS THE X-AXIS' LOCATE 24,1: PRINT TAB(33+(8-LEN(X$)/2 X$; REM THIS SECTION LABELS THE Y-AXIS START 3 + (8-LEN(Y$)/2) FOR I 1 TO LEN(Y$): LOCATE START+I,l: PRINT MID$(Y$,I,l): NEXT REM THIS SECTION LABELS THE INCREMENTS OF"THE Y-AXIS FOR I 1 TO 10: LOCATE 20-(I-1)*2,4 IF ACTUATOR$-"UNACTIVE" THEN PRINT USING "######";YAXIS(I) IF ACTUATOR$-"ACTIVE" THEN PRINT USING "#.##";YAXIS(I) NEXT I REM THIS SECTION LABELS THE INCREMENTS OF THE X-AXIS FOR I-l TO 10:LOCATE 23,14+(I-1)*6 IF ACTUATOR$-" ACTIVE" THEN PRINT USING "#. ;XAXIS( I) IF ACTUATOR$-"UNACTIVE" THEN PRINT USING XAXIS(I) NEXT I REM THIS SECTION CIRCLES DATA POINTS REM BEGIN R-LOOP FOR RE 1 TO PARAMETER(1,11) IF (ACTUATOR$-"ACTIVE") AND (SIGNAL(R)
PAGE 128

121 Appendix 1 ACTUATE PROGRAM 3650 IF ACTUATOR$a"ACTIVE" THEN YPT-171-LOG(SIGNAL(R-YAXIS(0/(YAXIS(10) YAXIS(O)*160 3655 CIRCLE (XPT.YPT).2 1 3660 REM END R-LOOP 3665 NEXT R 3670 IF ACTUATOR$-"ACTIVE" THEN GOSUB 3705 3675 INPUT VVVV$ 3680 SCREEN 0.1 3685 COLOR 3690 3695 LPRINT:LPRINT:LPRINT:LPRINT 3700 RETURN 3705 REM LINE GRAPHIC 3710 FOR 1-81 TO 559 STEP 2 37'5 3720 YPT-171-M*XVAL+B)-YAXIS(0/(YAXIS(10)-YAXIS(0*160 3725 IF YPT<11 OR YPT>171 GOTO 3730 ELSE PSET (I.YPT) 3730 NEXT I .. 3735 RETURN REM PROGRAM WRITTEN FOR THE CHEMISTRY DEPARTMENT AT U.C.D ON REM BY JUAN L. BONILLA 3750 REM THIS PROGRAM COUNTS PULSES FROM AN EXTERNAL DEVICE. 3755 REM THIS PROGRAM ALLOWS YOU TO COUNT FOR A SET NUMBER OF SECONDS AT 3760 REM TWO INPUT PLACES AT THE SAME TIME. THE INPUT PLACES HAVE TO BE 3765 REM CONECTED IN PINS 10 FOR COUNT ONE(N1) AND 18 FOR COUNT TWO(N2). 3770 REM THE TIME OF COUNTING DEPENDS ON THE' SECONDS DESIRE FROM 1 TO 10. 3775 REM YOU SHOULD HAVE A CABLE CONECTED BETWEEN PINS 8 11 TO BE 3780 REM ABLE TO SEND A SIGNAL OF CURRENT THAT WILL DEACTIVATE' THE 3785 REM INTERNAL COUNTERS TWO AND FOUR BY MEANS OF THEN LOOKING AT THE 3790 REM SIGNAL IN COUNTER THREE THAT COMES FROM COUNTER ONE WHEN IT 3795 REM REACHES ZERO( COUNTER ONE COUNTS DOWN TO ZERO AND TOGLES A 3800 REM CURRENT SIGNAL TO COUNTER THREE). 3805 REM COUNTERS THREE AND FIVE ARE USED WHEN THE NUMBERS AT COUNTERS 3810 REM TWO AND FOUR ARE TOO LARGE. (THEI TAKE ONE SIGNAL EACH TIME COUNTERS 3815 REM TWO AND FOUR REACH ONE THOUSAND COUNTS) 3820 REM 3825 REM N1 IS THE SUM OF COUNTS IN INPUT 10 (COUNTERS TWO AND THREE) 3830 REM N2 IS THE SUM OF COUNTS IN INPUT 18 (COUNTERS FOUR AND FIVE) 3835 REM CLS TIM-PARAMETER(7,1)

PAGE 129

Appendix 1 ACTUATE PROGRAM 3850 IF TIM > THEN A-15.2587 3855 IF TIM > THEN B-3906.25 3860 IF TIM > THEN C-13 3865 IF TIM THEN 3870 IF TIM < THEN B-62500! 3875 IF TIM < THEN C-12 3880 TIMY-INT(TIM*A) 3885 TIMX-INTTIM*B)-(TIMY*256 122 REM PROGRAM TO READ COUNTS IN COUNTER INPUT 1 AND 2 USING HARDWARE 3895 REM COUNTERS 2&3 AND 4&5 WITH TIMER OF COUNTER 1 3900 REM COUNTER PROGRAM, LAB MASTER INTERFACE 3905 FBORN-1808 REM MASTER TIMER RESET 3915 OUT (FBORN + 9),255 REM POINT DATA POINTER TO MASTER MODE 3925 OUT (FBORN+9),23 3930 REM MASTER MODE IS BCD DIVISION, ENABLE POINTER INCREMENT, 8 BIT BUS 3935 REM FOUT ON, DIVIDE BY 1, SOURCE -Fl, COMPARE AND TOO DISABLE 3940 OUT(FBORN+8),0 OUT(FBORN+8),l REM DP TO Cl,MODE REGISTER 3955 OUT (FBORNT9),1 REM SET OUTPUT COMMAND TO HAVE COUNTER 1 OUTPUT START WITH A HIGH 3965 REM IN IT'S OUT SIGNAL THAT GOES TO GATE 3 AND CHANGE TO LOW REM WHEN IT TOGLES AT TC.(TO STOP COUNTERS 2 AND 4) 3975 OUT(FBORN+9) ,233 3980 REM COUNTER C1 TC TOGGLED OUTPUT,DIS SPECIAL GATE, RELOAD FROM LOAD, 3985 REM COUNT ONCE, BINARY COUNT,DISABLE SPECIAL GATE,F3-3,906.25P/S, REM COUNT ON RISING EDGE, NO GATING 3995 OUT (FBORN+8),2 40000UT(FBORN+8),C 4005 REM PUT LOAD REGISTER 1 WITH A DECIMAL FOR 10 SEC. COUNT 40100UT(FBORN+8),TIMX 4015 OUT (FBORN+8),TIMY 4020 REM POINT DP TO C2 4025 OUT (FBORNT9),2 REM SET C2 FOR ACTIVE HIGH TERMINAL COUNT, COUNT UP, BCD COUNT, COUNT 4035 REM REPETITAVELY, RELOAD FROM LOAD, ENABLE SPECIAL GATE, COUNT ON RISING 4040 REM EDGE, SOURCE 2, GATE ON ACTIVE LOW LEVEL OF GATE N+l (GATE 3) OUT (FBORN+8),185

PAGE 130

Appendix 1 ACTUATE PROGRAM 4050 OUT (FBORN+8),66 4055 REM PUT 0 INTO LOAD REGISTER 4060 OUT (FBORN+8),0 4065 OUT (FBORN+8),0 4070 REM POINT DP TO C3 4075 OUT(FBORN+9) ,3 123 4080 REM SET C3 FOR, INACTIVE OUTPUT, COUNT UP, BCD COUNT, COUNT REPETITIVELY 4085 REM RELOAD FROM LOAD, DISABLE SPECIAL GATE, COUNT ON TISING EDGE OF TCN-1 4090 REM OF C2, NO GATING 4095 OUT (FBORN+8),56 4100 OUT (FBORN+8),0 4105 REM PUT 0 INTO LOAD REGISTER C3 4110 OUT(FBORN+8) ,0 4115 4120 REM DP TO C4, MODE REGISTER 4125 OUT (FBORN+9), 4 4130 REM SET UP C4 FOR ACTIVE HIGH TERMINAL COUNT. COUNT UP, BCD COUNT, COUNT 4135 REM REPETITAVILY, RELOAD FROM LOAD, ENABLE SPECIAL GATE, COUNT ON RISING 4140 REM EDGE, SOURCE 4, GATE ON ACTIVE LOW LEVEL OF GATE N-1 (GATE 3). 4145 OUT 4150 OUT (FBORN+8),100 4155 REM PUT 0 INTO LOAD OF REGISTER C4 4160 OUT (FBORN+8), 0 4165 OUT (FBORN+8), 0 4170 REM POINT TO C5 4175 OUT (FBORN+9),5 4180 REM SET C5 FOR, INACTIVE OUTPUT, COUNT UP, BCD COUNT, COUNT REPETITIVELY 4185 REM RELOAD FROM LOAD, DISABLE SPECIAL GATE, COUNT ON RISING EDGE OF TCN-1 4190 REM OF C4, NO GATING 4195 OUT (FBORN+8), 56 4200 OUT (FBORN+8), 0 4205 REM PUT 0 INTO LOAD REGISTER OF C5 4210 OUT (FBORN+8), 0 4215 OUT (FBORN+8), 0 4220 REM LOAD & ARM ALL SELECTED COUNTERS C1,C2,C3,C4,C5 4225 OUT (FBORN+9), 127 4230 RETURN 4235 REM READ COUNTERS A AND B SUBROUTINE 4240 REM PUT VALUE OF C1 IN HOLD, POINT TO IT AND READ 4245 REM IF THE VALUE IS NOT EQUAL TO RELOAD DO IT AGAIN

PAGE 131

Appendix 1 11250 LET L-1 4255 OUT (FBORN+9),161 4260 OUT (FBORN+9),17-11265 C1L-INP(FBORN+8) 4270 C1H-INP(FBORN+8) 11275 IF C1H-TIMY THEN GOTO 11285 4280 GOTO-1I255 ACTUATE PROGRAM 11285 IF C1LaTIMX-1 THEN GOTO 11295 11290 GOTO-1I255 11295 LET L-L+1 11300 IF L-1 THEN GOTO 4255 4305 REM DISARM AND SAVE ALL SELECTED COUNTERS 1-5 4310 OUT (FBORN+9), 159 124 4315 REM POINT TO HOLD REGISTER INCREMENT STARTING ON GROUP C1 AND READ C1L, 11320 REM C1H, C2L, C2H, C3L, C3H, CIIL, CIIH, C5L, C5H 4325 OUT (FBORN+9),25 11330 FOR M-l TO 10 4335 RRAY(M'a INP (FBORN+8) 113110 NEXT M 11345 DEF FNC(X)-X-6*INT(X/16) 11350 REM 11355 REM MAKE CONVERTIONS OF L&H 153 TO BASE 10 IN HUNDREDS 11360 REM NOTE: 153 IN COUNTER OF HARDWARE IN NUMBER, SEE 1750 11365 REM 11370 ACOUNTER-FNC(RRAY(3))*1+FNC(RRAY(II))*100+FNC(RRAY(5))*10000+FNC(RRAY(6))*lQ OOOOO! 11375 BCOUNTER-FNC(RRAY(7))*1+FNC(RRAY(8))*100+FNC(RRAY(9))*10000+FNC(RRAY(10))*1 OOOOOO! 11380 RETURN 11385 REM PRESSURE CORRECTION SUBROUTINE 11390 REM SETS FLOW PARAMETER 4395 REM THIS SECTION CALCULATES THE ADJUSTED ALPHA FLOW SETTING 111100 IF FLAG$(I)"OFF" THEN PARAMETER(5,l)-PARAMETER(5,l)+PARAMETER(I,1) 11405 IF FLAG$(I)"ON" THEN PARAMETER(5,l l-PARAMETER(5,ll-PARAMETER(I,l 111110 REM THIS SECTION SETS WHICH PARAMETER(I,l) WILL BE CALCULATED IN FVISR 41115 QQ-I 41120 LET 1-5 4425 REM ROUTES TO FVISR 4430 GOSUB 2405 4435 REM THIS SECTION SETS THE ADJUSTED EPSILON VOLTAGE 44110 GOSUB 4460 4445 LET I-QQ

PAGE 132

Appendix 1 ACTUATE PROGRAM 4450 RETURN 4455 REM CALCULATES THE ERROR IN NITROGEN DIOXIDE PARTICALS 4460 REM SUBROUTINE TO OUTPUT AN SPECIFIC VOLTAGE VALUE TO THE VALVES 4465 REM LPRINT "VOLTAGESET OUTPUT FOR 1-";1;" -";VOLTAGESET(I) 4470 IF 1>4 GO TO 4535 4475 REM ADJUST VAL SO IT WILL BE BETWEEN 1 AND 10 4480 REM START A CONVERTION 4485 DAC-(I-1)*2 4490 VOLTASET(I)-VOLTAGESET(I)*2 4495 T-INT(VOLTASET(I)*409.6) 4500 IF T>4095 THEN T-4095 4505 H-INT(T/16) 4510 L-16*(T-16*H) 4515 REM OUTPUT THE VALUE TO THE DAC 4520 OUT SBORN+DAC+1,H 4525 OUT SBORN+DAC,L 4530 IF 1<5 THEN 4595 4535 REM ADDENDUM TO 1>4 OR FOR 5 OR 6 FLOW CONTROLERS 4540 REM 4545 REM CONVERT TO A DECIMAL BETWEEN -2048 AND 2047 4550 DEClMAL-204.7*VOLTAGESET(I) 4555 REM CONVERT TO TWOS CONPLEMENT 4560 DECIMAL-INT(DECIMAL) 4565 HIGH-INT(DECIMAL/256) 4570 LOWaDECIMAL-256*HIGH 4575 IF HIGH
PAGE 133

126 Appendix 1 ACTUATE PROGRAM 4650 HIGHaINP(FBORN+6) 4655 REM CONVERT TO FROMS TWO CQNPLEMENT TO A NUMBER -10 10 4660 X-256*HIGH+LOW 4665 IF X>32767 THEN X-X-65536! 4670 VOLTRDG(I)-X/204.8 4675 RETURN 4680 REM CLOSING THE ITH FLOWCONTROLLER SUBROUTINE 4685 GOSUB 4765 4690 REM SET UP ALL PORTS IN MODE 0, PORT A,B,C OUTPUT 4695 OUT FBORN+15,128 4700 REM CLOSING VALUE FOR NUM VALVES IN PORT B 4705 NU-NU+NUM 4710 OUT FBORN+13,NU 4715 RETURN 4720 REM OPENING THE ITH FLOWCONTROLLER SUBROUTINE, 4725 GOSUB 4765 4730 REM SET UP' ALL PORTS IN MODE 0, PORT A,B,C OUTPUT 4735 OUT FBORN+15,128 4740 REM OPENING VALUE FOR NUM VALVES IN PORT B 4745 NU-NU-NUM 4750 OUT FBORN+13,NU 4755 REM TIME LOOP TO WAIT FOR THE FLOW TO OPEN 4760 RETURN 4765 REM FLOWCONTROLLER SOFTWARE PIN NUMBER ASSIGNMENT SUBROUTINE 4770 REM DO PORT 4775 REM LPRINT CN$(I) 4780 IF 1-1 THEN NUM-1 4785 IF 1-2 THEN NUM-Z 4790 IF 1-3 THEN NUM-4 4795 IF 1-4 THEN NUM-8 4800 IF 1-5 THEN NUM-16 4805 IF 1-6 THEN NUM-32 4310 IF 1-7 THEN NUM-64 4815 RETURN 4820 REM FLOWCONTROLLER GAS IDENTIFICATION SUBROUTINE 4825 REM POSES QUESTION 4830 WIDTH 80:CLS:LOCATE 7 4835 PRINT" ENTER THE APPROPRIATE GAS NUMBER FOR THE ITH FLOWe ONTROLLER." 4840 LOCATE 9 4845 PRINT" HELIUM"

PAGE 134

Appendix 1 ACTUATE PROGRAM 4850 LOCATE 11 4855 PRINT" HYDROGEN-HELIUM" 4860 LOCATE 13 4865 PRINT" NITROGEN DIOXIDE" 4870 LOCATE 15 4875 PRINT" #4 n-BUTANE" 4880 LOCATE 17 4885 PRINT" SULFUR DIOXIDE" 4890 LOCATE 19 4895 PRINT" #6 ---------OXYGEN" 4900 INPUT PARAMETER(I,15) 4905 REM THIS SECTION PROTECTS THE GAS CHOICE. 4910 IF (PARAMETER(I,15)<1) AND (PARAMETER(I,15) 4915 RETURN 4920 REM TITRATION SUBROUTINE 4922 LPRINT TIME$ 4925 LPRINT "STARTING TITRATION SUBROUTINE" 4930 REM TURNS OFF THE DELTA FLOWCONTROLLER 4935 FOR R-l TO PARAMETER(7,11) 4940 LET"GASFLOW(R)-O 4945 NEXT R 4950 LET I-NFC:Ra l 4955 FLAG$(I)-"OFF":POSITION-l 4960 GOSUB 4385 4965 GOSUB 4680 4970 LOCATE 12:PRINT "TURNING OFF, PLEASE WAIT" 4975 FOR Na1"TO 530*PARAMETER(7,8):NEXT N 4980 IF PARAMETER(7,3)-1 THEN GOSUB 5940 4985 IF PARAMETER(7,3)-2 THEN GOSUB 6585 4990 GOSUB 5240 4995 REM BEGIN INCREMENTATION LOOP 5000 REM FOR R-l TO PARAMETER(7,11) THEN 4715 REM"HANDLES THE INCREMENTATION OF ANY ONE FLOWCONTROLLER FOR 1 TO 5 IF PARAMETER(I,2)-0 THEN 5055 127 5005 5010 5015 5020 5025 5030 5035 5040 IF (PARAMETER(I,2)<6) AND (R-7) THEN PARAMETER(I,2)-2*PARAMETER(I,2) IF R-8 THEN PARAMETER(I,2)-3*PARAMETER(I,2) IF (R>1) THEN PARAMETER(I,1)-PARAMETER(I,1 )+PARAMETER(I,2) LET GASFLOW(R)-PARAMETER(I;1)

PAGE 135

Appendix 1 ACTUATE PROGRAM NEXT I LET WWmI LET I a 5 IF GASFLOW(R)-O THEN 5085 LET GOSUB 2405 GOSUB 4460 IF I>
PAGE 136

'29 Appendix 1 ACTUATE PROGRAM 5245 REM BEGIN Z-LOOP 5250 LET BKG$-"GOOD" 5255 FOR Z2 TO 36 STEP 2 5260 IF (Z<34) AND (FLAG$(NFC)-"OFF") THEN 5290 5265 IF (FLAG$(NFC)-"OFF") THEN WNE(R,Z-33,POSITION)"SSUM(Zl2) 5270 IF (FLAG$(NFC)-"OFF") THEN WNE(R,ZII32,POSITION)-ESUM(Zl2) 5275 IF (FLAG$(NFC)-"ON") THEN TITRATION(R,Z-,)-SSUM(Z/2) 5280 IF (FLAG$(NFC)-"ON") THEN TITRATION(R,Z).ESUM(Z/2) 5285 REM END Z-LOOP 5290 NEXT Z 5295 IF (POSITION-') AND (FLAG$(NFC)-"OFF") THEN 5380 5300 IF (POSITION-2) AND (FLAC$(NFC)-"ON") THEN 5380 5305 REM COMPUTES BASELINE 53'0 LET BASELINE(R) -(WNE(R,1,1)+WNE(R,l,2/2 5315 LET EBASELINE(R)-.5*(WNE(R;2,1)A2+WNE(R,2,2)A2)A. 5 5320 REM CHECKS IF BASELINE IS VALID 5325 IF .03BASELINE(R)
PAGE 137

Appendix 1 ACTUATE PROGRAM 5445 LPRINT'" R' ALPHA BETA GAMMA ." 5455 FOR R -1 TO PARAMETER(7,11) 5465 LPRINT"''';:LPRINT"USING'''''';R; REM BEGIN Z-LOOP 5475 FOR Z-2 TO 6 STEP 2 LPRINT"''';: LPRINT ; TITRATION (R, Z-1 ) ; : LPRINT" + /-" ; : LPRINT "JS ING""'"."""";TITRATION(R,Z); 5485 REM END OF Z-LOOP NEXT Z 5495 LPRINT"'" REM END OF R-LOOP NEXT R REM BEGIN R-LOOP 551 5 LPRINT'" DELTA EPSILON IOTA ." 5525 FOR R -1 TO PARAMETER(7,11) 5535 LPRINT"''';:LPRINT"USING''III1'';R; REM BEGIN Z-LOOP 5545 FOR Z-8 TO 12 STEP 2 LPRINT"";:LPRINT USING"III1".O"";TITRATION(R,Z-1);:LPRINT" +/-";:LPRINT "JS ING"O".U"";TITRATION(R,Z); 5555 REM END OF Z-LOOP NEXT Z 5565 LPRINT"'" REM END OF R-LOOP 5575 NEXT R REM BEGIN R-LOOP 5585 LPRINT'" R' ONE' TEMPERATURE TWO ." 5595 5600 FOR R -1 TO PARAMETER(7,11) LPRINT"''';: LPRINT" R; REM BEGIN Z-LOOP 5615 FOR Z-14 TO 16 STEP 2 5620 LPRINT"''';:LPRINT USING"O"n.O"";TITRATION(R,Z-1);:LPRINT" +/-";:LPRINT JS ING""'."''';TITRATION(R,Z); 5625 REM END OF Z-LOOP NEXT Z 5635 LPRINT"'" REM END OF R-LOOP

PAGE 138

Appendix 1 ACTUATE PROGRAM 5645 NEXT R 5650 REM BEGIN R-LOOP 5655 LPRINT"..... 5660 LPRINT" R' TOTAL COUNTS BASELINE ." 5665 LPRINT" 5670 FOR R -1 TO PARAMETER(7,11) 5675 LPRINT"";:LPRINT"USING"##";R; 5680 REM BEGIN Z-LOOP 5685 Z-34 131 5690 LPRINT"";:LPRINT USING"#########.##";TITRATION(R,Z-1);:LPRINT" +/-";:LPRIN T USING"######.##";TITRATION(R,Z); 5695 REM END OF Z-LOOP 5700 LPRINT "."; 5705 LPRINT USING"######','.#,";BASELINE(R);:LPRINT" +/";:LPRINT USING"#" ###.##";EBASELINE(R);:LPRINT" ." 5710 REM END OF R-LOOP 5715 NEXT R 5720 REM BEGIN R-LOOP 5725 LPRINT"... 5730 LPRINT" R PRESSURE ACTUATOR ." 5735 LPRINT" 5740 FOR R -1 TO PARAMETER(7,11) 5745 LPRINT"";:LPRINT"USING"##";R; 5750 FOR Z-22 TO 24 STEP 2 5755 LPRINT"''';:LPRINT USING",##n.###";TITRATION(R,Z-1);:LPRINT" +/-";:LPRINT US ING""#.#II";TITRATION(R,Z); 5760 NEXT Z 5765 LPRINT"" 5770 REM END OF R-LOOP 5775 NEXT R 5780 REM BEGIN R-LOOP 5785 LPRINT". 5790 LPRINT'" R SIGNAL(R)' GASFLOW(R) ." 5795 LPRINT" 5800 FOR R -1 TO PARAMETER(7, 1 5805 LPRINT"''';:LPRINT"USING''##'';R; 5810 LPRINT".";:LPRINT USING"II###I#.#";SIGNAL(R);:LPRINT" +/";:LPRINT USING"# #"; ESIGNAL( R) ; : LPRINT" '''; : LPRINT USING GASFLOW( R) ; : LPRINT 5815 LPRINT"" 5820 REM END OF R-LOOP 5825 NEXT R 5830 LPRINT" 5835 RETURN 5840 REM SUBROUTINE TO CHECK FOR POSITION OF FLOW IN ACTUATOR

PAGE 139

Appendix 1 ACTUATE PROGRAM 58115 IF LEVEL$-"DOWN" THEN LEVEL$a"UP" ELSE LEVEL$a"DOWN" 5050 I -7 5855 IF LEVEL$a"UP" THEN GOSUB 4680 5860 IF LEVEL$-"DOWN" THEN GOSUB 4720 5865 I s12 5870 FOR QQ-l TO 3000: NEXT QQ 5875 GOSUB 11600 5880 PS VOLTRDG(12) 5885 IF PS > 1.8 AND PS < 2.3 THEN PORT -1 5890 IF PS > 1.1 AND PS < 1.8 THEN PORT 2 5895 IF PS > PS < THEN PORT -3 5900 IF PS > O! AND PS (" ; 4 THEN PORT -4 5905 IF PS > 4.3 AND PS < 5! THEN PORT -5 5910 IF PS > 3.7 AND PS < 4.3 THEN PORT -6 5915 IF PS > 3! AND PS < 3.7 THEN PORT 1 5920 IF PS > 2.3 AND PS < 3! THEN PORT -8 5925 LPRINT VOLTRDG(12);" PORT ";PORT 5930 IF PORT <> R THEN 5935 RETURN 5940 REM SOUTROUTINE TO ACELERATE THE PROGRAM READING SOUBROUTINE 5945 DIM STORAGE(PARAMETER(7,6) ,13) 5950 WIDTH 40:CLS 5955 FOR 1-1 TO 12 SUM(I)-O:STDSUM(I)-O:SSUM(I)-O:ESUM(I)-O:VOLTRDG(I)-O 5965 NEXT I 5970 FOR T-l TO PARAMETER(7,6) 5975 LOCATE 12:PRINT TAB(5);FLAG$(NFC),R:LOCATE 15:PRINT TAB(18);T 5980 REM TURN ON THE COUNTER 5985 GOSUB 3845 5990 REM READ VOLTAGES 5995 FOR I -1 TO 12 6000 STORAGE(T,I)O 6005 IF 1>6 THEN GOTO 6015 6010 IF (1<7) AND (PARAMETER(I,l)-O) THEN 6060 6015 GOSUB 6020 STORAGE(T,I)-VOLTRDG(I) 6025 SUM(I)-SUM(I)+VOLTRDG(I) 6030 IF T < PARAMETER(1,6) THEN GOTO 6060 6035 SSUM(I)-SUM(I)/PARAMETER(7,6) 6040 FOR Fs1 TO PARAMETER(7,6) 132

PAGE 140

Appendix ACTUATE PROGRAM STDSUM(I)-STDSUM(I)+(SSUM(I)-STORAGE(F.IA2 6050 NEXT F 6055 ESUM(I)-(STDSUM(I)/(PARAMETER(7.6)-1A.5 6060 NEXT I 6065 IF T-l THEN SUM(17)-0 6070 IF T-l THEN STDSUM(17)-0 6075 REM READ COUNTER CONTENTS 6080 GOSUB 6085 STORAGE(T.13)-ACOUNTER 6090 SUM(17)-SUM(17)+ACOUNTER 6095 IF T (PARAMETER(7.7)SSUM(17 THEN GOTO 5950 6135 CLS 6140 LPRINT "N02 ";FLAG$(NFC);" COUNTS ";SSUM(17);" +/";ESUM(17) ERASE STORAGE 6150 RETURN 6155 REM SUBROUTINE TO CALCULATE VELOCITIES OF FLOW & REACTION TIME" 6160 VELOCITY -(TFLOW)/(60.06TITRATION(1.21 6165 LPRINT "VELOCITY OF FLOW TUBE "; 6170 LPRINT USING "####U#.U#";VELOCITY; 6175 LPRINT CM/SEC" 6180 LPRINT 6185 IF ACTUATOR$-"UNACTIVE" THEN GOTO 6225 6190 DISTANCEa 5 6195 FOR R-PARAMETER(7.11) TO 1 STEP-l 6200 DISTANCE-DISTANCE+5 6205 RXTIME(R)sDISTANCE/VELOCITY 133 6210 LPRINT "THE REACTION TIME AT PORT(";R;") IS ";RXTIME(R);" SECONDS" 6215 NEXT R 6220 LPRINT 6225 RETURN 6230 REM SUBROUTINE TO CALCULATE CONCENTRATIONS 6235 FOR I -1 TO 6240 CONCENO

PAGE 141

Appendix ACTUATE PROGRAM IF PARAMETER(I,l)mO THEN GOTO 6275 6250 IF PARAMETER(I,2 THEN GOSUB 6290 131l 6255 CONCEN-PARAMETER(I,l )TITRATION(l,21)(3.2E+16)PARAMETER(I,16)/TFLOW 6260 LPRINT "CONC. FOR ";GAS$(I);" IS,Q; 6265 LPRINT CONCEN; 6270 LPRINT PARTI/CM3" 6275 NEXT I 6280 LPRINT" 6285 RETURN 6290 FOR R-l TO PARAMETER(7,11) 6295 CONCEN .. O 6300 REM 6305 REM 6310 CONCENmGASFLOW(R)TITRATION(R,21 )'(3.2E+16)'PARAMETER(I,16)/TFLOW 6315 LPRINT "CONC. FOR ";GAS$(I);" AT POINT O";R;" "; LPRINT CONCEN; 6325 LPRINT PARTI/CM3" 6330 NEXT R 6335 LPRINT"... 63110 RETURN 63115 REM SOBROUTINE TO CALCULATE LINEAR LEAST SQUARE 6350 LPRINT" 6355 LPRINT" LINEAR LEAST SQUARE RESULTS OF KINETICS WITH DATA REJECT 6360 LPRINT"---------------------------------------------------------------" 6365 N-O:Nl"O 6370 FOR Ral TO PARAMETER(7,11) 6375 X(R)-O:Y(R)-O 6380 IF SIGNAL(R)
PAGE 142

Appendix 1 ACTUATE PROGRAM (Z2(S4-S2S2/N1 6450 FOR" 1-1 TO PARAMETER(7,11) IF SIGNAL(I)0 THEN 6430 ELSE PRINT 135 6525 PRINT N-Nl; DATA PAIRS WERE REJECTED AT ";RC; STD DEV CRITERION" SDS-SQRS6N1)/N-2)ABS(S1S1)-N1S3 6535 SDS2a(N1S4-(S2-2)-(NlS5-S1S2'/(Nl.S3-S1-2-2)*(N1.S3-S1-2-.5/N-2) -.5(NlS3-S1-2)-.5) SPI-SQRS6.S3)/N-2).ABS(S1.S1)-Nl.S3 6545 LPRINT 6550 LPRINT SLOPE -";M TAB(24) "STD DEV IN SLOPE -";SDS 6555 LPRINT "SECOND CAL. OF STD IN SLOPE -";SDS2 LPRINT INTERCEPT -";B TAB(28) "STD DEV IN INTERCEPT -";SDI 6565 LPRINT CORRELATION COEFFICIENT _"; R 6570 LPRINT STD DEV IN LINE FOR ";N ; "DATA PAIRS" "SDL:LPRINT 6575 LPRINT *" 6580 RETURN 6585 REM SUBROUTINE TO READ MORE TIMES THE COUNTER REM EACH TIME THERE IS ANY VOLTAGE READ 6595 DIM COUNTS(300), STORAGE(PARAMETER(7,6),12) 6600 WIDTH 40:CLS 6605 SUM(17)-0:STDSUM(17)-0:ZZ-0 6610 FOR"T-l TO PARAMETER(7,6) 6615 FOR 1-1 TO 12 LOCATE 12:PRINT FLAG$(NFC),R,T,I 6625 REM TURN ON COUNTER 6630 GOSUB 3845 6635 IF T-l THEN SUM(I)-O 6640 IF T-l THEN STDSUM(I)-O

PAGE 143

Appendix 1 ACTUATE PROGRAM IF T-1 THEN VOLTRDG(I)-O 6650 REM READ VOLTAGE 6655 GOSUB 4600 6660 STORAGE(T,I)-VOLTRDG(I) 6665 SUM(I)-SUM(I)+VOLTRDG(I) 6670 REM READ COUNTER VALUE 6675 GOSUB .4235 6680 ZZ-ZZ+l 6685 COUNTS(ZZ)-ACOUNTER 6690 SUM(17)-SUM(17)+ACOUNTER 6695 IF T(PARAMETER(7,6) THEN GO TO 6725 6700 SSUM(I)-SUM(I)/PARAMETER(7,6) 6705 FOR F-1 TO PARAMETER (7 ,6) 6710 STDSUM(I)-STDSUM(I)+(SSUM(I)-STORAGE(F,IA2 6715 NEXT F 6720 ESUM(I)-(STDSUM(I)/(PARAMETER(7,6)-1A. 5 6725 NEXT I 6730 IF T(PARAMETER(7,6) THEN GOTO 6760 6735 SSUM(17)-SUM(17)/(PARAMETER(7,6)) 67ijO FOR F-1TO (PARAHETER(7,6)*12) 67ij5 STDSUM(17)-STDSUM(17)+(SSUM(17)-COUNTS(FA2 6750 NEXT F 6755 ESUM(17)-(STDSUM(17)/PARAMETER(7,6)*12)-1A.5 6760 NEXT T 6765 IF ESUM(17(PARAMETER(7,7)*SSUM(17 THEN GOTO 6605 6770 LPRINT "N02 ";FLAG$(NFC);" COUNTS ";SSUM(17);" ";ESUM(17) 6775 ERASE COUNTS, STORAGE 6780 CLS 6795 RETURN 136

PAGE 144

137 Appendix 2 KINETICS RESULTS FOR HO + n-BUTANE REACTIONS PSEUDO APPARENT FIRST BIMOLECULA R ORDER .RATE HO BUTANE RATE CONSTANT INTERCEPT 'CONC. CONC. (E-12) DATE (E 12) cm3/molec.sec 16-Apr-86 HIGH 85.2 16-Apr-86 HIGH 74.6 16-Apr-86 HIGH 63.8 16-Apr-86 HIGH 53.2 16-Apr-86 HIGH 42.5 16-Apr-86 HIGH 32.0 16-Apr-86 HIGH 21.3 61.5 16-Apr-86 HIGH 10.6 16-Apr-86 LOW 85.0 192. 1 8. 1 16-Apr-B6 LOW B5.0 16-Apr-86 LOW 63.7 16-Apr-86 LOW 42.6 16-Apr-96 LOW 21.2 3 21-Apr-B6 HIGH 75.4 21-Apr-86 HIGH 64.0 21-Apr-86 HIGH 53.8 21-Apr-B6 HIGH 42.8 21-Apr-85 HIGH 32.2 21-Apr-85 HIGH 21.4 21-Apr-86 HIGH 10.7 21-Apr-86 l.OW 86.2 21-Apr-86 tow 74.8 21-Apr-86 LOW 74.9 21-Apr-86 l.OW 54.9 21-Apr-B6 l.OW 53.4 121.3 3 21-Apr-B6 LOW 42.6 21-Apr-86 LOW 32.0 21-Apr-86 LOW 32.0 76.5 1.7 21-Apr-86 LOW 21. 3 56.2 4. 21-Apr-86 LOW 21. 3 21-Apr-B6 LOW 10.6 28-Apr-B6 HIGH 90.3 258.B 28-Apr-86 HIGH 68.1 28-Apr-86 HIGH 57.5 28-Apr-86 HIGH 45.6 28-Apr-86 HIGH 22. B 1 2B-Apr-B6 HIGH 11.5 O.ll 29-Apr-86 HIGH 75.2 29-Apr-86 HIGH 64.6 29-Apr-86 HIGH 50.6 29-Apr-86 HIGH 50.9 29-Apr-86 HIGH 3B.ll 29-Apr-B6 HIGH 25.11 1.2 29-Apr-86 HIGH 12.8

PAGE 145

138 29-Apr-S6 LOW SO. II 193.3 S.2 29-Apr-86 LOW 69.8 169.6 29-Apr-86 LOW 53.2 1311.9 29-Apr-86 LOW 41.9 2 29-Apr-86 LOW 21.6 1.8 29-Apr-86 LOW 13.9 14-May-S6 HIGH 75.1 111-May-86 HIGH 61.0 6 111-MaY-86 HIGH 50.4 3 111-May-86 HIGH 33.6 ll1-May-86 HIGH 16.8 111-May-86 HIGH 8.11 30.1 14-May-86 LOW 75.1 111-May-86 LOW 65. B 111-May-86 LOW 66.9 14-MaY-86 LOW 50.2 14-May-86 LOW 33.4 14-May-86 LOW 16.8 42.1 1. 3 111-May-86 LOW B.4 O.B 31-Aug-B6 VERY 53.2 3.11 31-Aug-86 VERY 44.3 3 31-Aug-86 VERY 114.4 13B.9 31-Aug-B6 VERY 35.4 31-Aug-86 VERY 35.6 113.4 31-Aug-86 VERY 26.6 31-Aug-86 VERY 26.6 31-Aug-B6 VERY 11.8 119.4 1.4 31-Aug-86 VERY 17.8 31-Aug-86 VERY B.9 31-Aug-86 VERY 8.9 21-Sep-86 VERY 61. 5 21-Sep-86 VERY 53.2 3 21-Sep-B6 VERY 53.3 21-Sep-B6 VERY 114.4 21-Sep-86 VERY 35.5 9.1 21-Sep-86 VERY 35.5 21-Sep-86 VERY 26.6 21-Sep-86 VERY 17.8 21-Sep-86 VERY 17.8 21-Sep-86 VERY 8.9 ALL LOW -----------------> DATA HIGH -----------------> POINTS VERY -----------------> NOT CONSIDERED

PAGE 146

139 Appendix 2 cont. HIGH PRESSURE KINETICS RESULTS (3.7-3.9 TORR) PSEUDO APPARENT FIRST BIMOLECULAR ORDER RATE HO BUTANE RATE CONSTANT INTERCEPT CONC. CONC. (1 (E-12) DATE (E12) cm3/molec.sec 22-Apr-B6 HIGH 102 19B.2 3.1 2.?-Apr-B6 HIGH 93.2 lB2.0 22-Apr-g6 HIGH B1. II 1 62. 1 2. 8 22-Apr-B6 HIGH 72.1 22-Apr-86 HIGH 61.2 22-Apr-86 HIGH 51. 3 10B.l 22-Apr-86 HIGH 110.9 22-Apr-86 HIGH 30.8 22-Apr-86 HIGH 20.5 22-Apr-B6 HIGH 10.3 22-Apr-B6 LOW 911.9 4.S 22-Apr-B6 LOW 83.8 22-Apr-B6 LOW 73.4 22-Apr-S6 LOW 62.9 22-Apr-S6 LOW 52.4 22-Apr-86 LOW S1.9 2 22-Apr-S6 LOW 31.4 22-Apr-S6 LOW 20.9 22-Apr-86 LOW 10.2 23-Apr-86 HIGH 95.4 23-Apr-86 HIGH 83.3 23-Apr-B6 HIGH 74.1 23-Apr-B6 HIGH 63. 1 152.B 23-Apr-86 HIGH 52.8 123.B 23-Apr-86 HIGH 42.2 23-Apr-86 HIGH 31.6 23-Apr-86 HIGH 21.0 5S.2 23-Apr-86 HIGH 10.5 31.S 23-Apr-86 LOW 86.4 23-Apr-86 LOW 86.5 23-Apr-S6 LOW 75.9 23-Apr-B6 LOW 611.4 23-Apr-86 LOW 54 2.S 23-Apr-86 LOW 42.7 23-Apr-86 LOW 32.3 23-Apr-B6 LOW 2'.4 IIS.2 23-Apr-86 LOW 10.7 25.S

PAGE 147

REACTIONS OF SYSTEM TO BE SIMULATED Reaction 1 ) H. NO? -> NO HO 2 ) HO + HO -> H202 3 ) HO H2:)2 H20 H02 4 ) HO H2 -> H20 + H. 5 ) HO HN03 -> H20 + N03 6 ) HO NO -> 7 ) HO N02 -> HN03 + 8 ) HO + HO -> H20 O. 9 ) H02 + NO -> HO + ND2 1CJ )HO -> 1 1 )0. "HO H. 02 N02 -> NO 02 13 )H. H02 -> + 14 )0. H02 -> HO + 02 15 )HO + H02 -> H20 + 02 16 )H02 + H02 -> H202 + 02 17 )0. N03 -> 02 + N02 18 )NO + N03 -> N02 N02 19 )HO + n-Butane -> n-Butyl + H20 20 )HO + n-Butyl -> R-OH 21 )HO R-OH -> Alcohyl H20 22 )n-Butyl + n-Butyl -> n-Butane + n-Butene 23 )n-Butyl n-Butyl -> Octane 24 )HO Octane -> H20 25 )HO + n-Butene -> Alcohyl + 26 )0. + n-Butane -> n-Butyl + HO 27 )NO + n-Butyl -> ButylNO + 28 )N02 n-Butyl -> ButylNO 1/ppb-sl!c 2.B25 .00055 .0425 .0001675 .0025 .00056 .00203 ;0475 .2075 2 .825 .2325 1.85 1.475 1. 75 .0425 .25 .71199999 .0525 12.5 .175 .25 .0125 .22825 .711 .00055 .005 .625 140 1 3E-1 0 ?2E-11J 1.7-12 5.7E .. 15 3.32E-14 1.9E-12 8.3E-12 8E-11 3. 3E-1 1 7.4E-ll 5.9::-11 7E-l1 1.7E-12 1E-11 3E-11 2.1-12 5E-10 7E-12 1E-ll 5E-13 9.13E-'2 2.2E-14 2E-'3 2.5E-"