183 lines
14 KiB
TeX
183 lines
14 KiB
TeX
% !TeX program = lualatex
|
|
|
|
%% ------------------------------------------------------------------%
|
|
%% SUBFILE MODE
|
|
\documentclass[../main/main.tex]{subfiles}
|
|
\graphicspath{{../graphics/}{../graphics/charts/}{../graphics/postscript_figures/}{../graphics7raster_images/}}
|
|
|
|
|
|
%% ------------------------------------------------------------------%
|
|
%% DOCUMENT
|
|
|
|
|
|
\begin{document}
|
|
|
|
\pagestyle{mainbody}
|
|
|
|
\chapter{Methods}
|
|
\thispagestyle{mainbody}
|
|
|
|
We use a sequential workflow outlined in figure \ref{methods: fig: overview over workflow}. It consists of the following steps:
|
|
|
|
\begin{figure}[h]
|
|
\centering
|
|
\includegraphics{05_Method_Overview.pdf}
|
|
\caption[Overview over Workflow]{\textbf{Overview over Workflow}. The initial step consists of producing a filter containing \gls{VOCa} in the \gls{OFRa}, which is subsequently extracted to produce a solution. }
|
|
\label{methods: fig: overview over workflow}
|
|
\end{figure}
|
|
|
|
\begin{enumerate}
|
|
\item{
|
|
\textbf{\Gls{SOAa} filter production in \gls{OFRa}}: Within the \gls{OFRa}, we produce \gls{SOAa} through the reaction of a \gls{VOCa} with \gls{O3f}. The resulting \gls{SOAa} is collected on a filter. The detailed procedure is outlined in section \ref{methods: section: SOA filter production in OFR}.}
|
|
\item{
|
|
\textbf{Filter extraction and \gls{TOCa} analysis}. The filters produced in the previous step are extracted into aqueous solution and \gls{TOCa} is measured as a proxy for sample content of organics. This is outlined in detail in section \ref{methods: section: filter extraction and TOC analysis}.}
|
|
\item{
|
|
\textbf{CWFTR experiments}. In this step, the solutions from the extracted filters are mixed with iron and \gls{I-f}. Using such solutions, flow tubes are coated and the products of the reaction of their film with \gls{O3f} are measured in a \gls{BTOF-CIMSa}. The exact methodology is described in section \ref{methods: section: CWFTR Experiments}}
|
|
\item{
|
|
\textbf{Data Analysis}. In this step, the data from the \gls{BTOF-CIMSa} are analysed using \emph{IGOR Pro} and \emph{Python}. Details are available in section \ref{methods: section: data analysis}.}
|
|
\end{enumerate}
|
|
|
|
\section{SOA filter production in OFR}\label{methods: section: SOA filter production in OFR}
|
|
|
|
An overview over the setup for filter production is displayed in figure \ref{methods: fig: overview over OFR setup}. \Gls{O2f} and \gls{N2f} from the in-house gas connection at \gls{PSIa} are first mixed at controlled ratios and guided inton the flow tube of the \gls{OFRa}. Usually, $\gls{VFLOWs}_{\ce{N2}}=\qty{3.95}{\liter\per\minute}$ and $\gls{VFLOWs}_{\ce{O2}}^{(3)}=\qty{1.05}{\liter\per\minute}$ was set, giving a relative fraction of $\glssymbol{N2f}:\glssymbol{O2f}\approx0.79:0.21$, which is close to atmospheric comoposition. This was achieved through the \glspl{MFCa} 2 (for \gls{N2f}) and 3 (for \gls{O2f}), respectively (see also figure \ref{methods: fig: overview over OFR setup}). \Gls{MFCa} 1 was used to control the flow of \gls{N2f} through a glass bottle containing \gls{VOCa}, with a value set at $\gls{VFLOWs}_{\textnormal{VOC}}=\qty{0.04}{\liter\per\minute}$. In order to inhibit diffusion of \gls{VOCa} into the \gls{OFRa}, a plug valve (PV1) was installed on the \gls{VOCa} input line just before the \gls{OFRa}. Finally, \gls{MFCa} 4 was used to control the flow of \gls{O2f} into the \glstext{O3f} generator, where \glstext{O2f} reacted with \gls{UVa} light to produce \glstext{O3f}, as in reaction below:
|
|
|
|
\begin{reaction}
|
|
\label{methods: reac: O2 + hv -> O3}
|
|
\glssymbol{O2f} + ${h\nu}$ -> 2\glssymbol{O-1D.f}
|
|
\end{reaction}
|
|
|
|
\begin{figure}[h]
|
|
\centering
|
|
\includegraphics{03_OFR_Flowchart.pdf}
|
|
\caption[Overview over OFR Flowchart]{\textbf{Overview over OFR Flowchart}. Flow Chart displaying the functionality of the OFR Setup, including various flows, instruments, diverse valves and other things.}
|
|
\label{methods: fig: overview over OFR setup}
|
|
\end{figure}
|
|
|
|
For this, \gls{MFCa} 4 was set to provide a flow of $\gls{VFLOWs}_{\ce{O2}}^{(4)}=\qty{3}{\liter\per\minute}$ to the \glstext{O3f} generator. Without any additional measures, this \gls{O2f} flow is efficiently converted into \glstext{O3f} by the generator, such that concentrations in excess $\gls{CHIs}_{\ce{O3}}\approx\qty{100}{\ppm}$ are reached within the \gls{OFRa}. In order to decrease \gls{O3f} levels within the \gls{OFRa}, a pump was installed that pumped a flow of $\gls{VFLOWs}_{\textnormal{P}}\approx\qty{2.2}{\liter\per\minute}$ of generated \gls{O3f} into exhaust 2, such that steady-state concentrations of \gls{O3f} within the \gls{OFRa} were not usually in excess of $\approx\qty{70}{\ppm}$. The flow rate of the pump was adjusted using a needle valve (NV1). This was done once before the start of experiments, and the valve setting was not changed throughout experiments. Within the \gls{OFRa}, thus, the flows of \gls{VOCa} (on its \glssymbol{N2f} carrier gas), «air» (with $\glssymbol{N2f}:\glssymbol{O2f}=0.79:0.21$) and \gls{O3f} (at $\gls{VFLOWs}_{\ce{O3}}\approx\qty{0.8}{\liter\per\minute}$) converge. Here, \gls{VOCa} is oxidised by \gls{O3f} into \glspl{OVOCa} and \glspl{LVOCa}. These \glspl{LVOCa} condense into \gls{SOAa}.
|
|
|
|
Two lines exit the \gls{OFRa}: One that takes an air sample from within the middle of the flow tube, where the flow is nearly laminar, and another one consisting of four different outlets at the flow tubes' rear end that converge into a single outlet line. The latter line is connected to a filter holder through which a flow of product gases $\gls{VFLOWs}_{\textnormal{F}}\approx\qty{4.5}{\liter\per\minute}$ is guided into exhaust 1 --- it is here that \gls{SOAa} is collected: within the filter holder, a Nylon filter with pore size \qty{0.45}{\micro\meter} is installed, on which \gls{SOAa} is collected. The other line is guided into the \gls{SMPSa} and the \gls{O3f} analyser. Downstream of those instruments, the flow through the former line is converging again, and sucked into exhaust 2 along with excess \gls{O3f} from the pump. The \gls{SMPSa} allows for an online measurement of particle size, and is described in detail in section \ref{methods: subsec: SMPS}. Using a ball valve (BV1), the line can be opened or closed towards the \gls{O3f} analyser: if BV1 is closing the \gls{OFRa} outlet flow towards the analyser, a three-way-valve (3WV1) allows for the \gls{O3f} analyser to sample ambient air. In normal operation, BV1 is open and 3WV1 closes off the system towards the ambient. In this configuration, a flow of $\gls{VFLOWs}_{\textnormal{CPC}}\approx\qty{0.3}{\liter\per\minute}$ is guided into the \gls{SMPSa} and a flow of $\gls{VFLOWs}_{\ce{O3}}^{(A)}\approx\qty{1}{\liter\per\minute}$ flows through the \gls{O3f} analyser.
|
|
|
|
\subsection{Procedure}
|
|
\label{methods: subsec: OFR Procedure}
|
|
A detailed methodological procedure is available in appendix \ref{chapter: appendix: procedures}. Throughout the experiments, \gls{SMPSa} data was collected within $\approx\num{60}$ (this number varied across experiments) samples, each of which constitutes of aggragating data for \qty{180}{\second}. Briefly, the following steps need to be undertaken:
|
|
|
|
\begin{enumerate}
|
|
\item{
|
|
All lines are checked and it is ensured that \gls{SMPSa} and \gls{O3f} analyser are ready for data logging;}
|
|
\item{
|
|
The \gls{O3f} analyser and \gls{CPCa} are switched on so that they can warm up.}
|
|
\item{
|
|
Flows are set at $\gls{VFLOWs}_{\ce{N2}}=\qty{3.95}{\liter\per\minute}$, $\gls{VFLOWs}_{\ce{O2}}^{(3)}=\qty{1.05}{\liter\per\minute}$ and $\gls{VFLOWs}_{\ce{O2}}^{(4)}=\qty{1.05}{\liter\per\minute}$. The \gls{O3f} excess pump is switched on.}
|
|
\item{
|
|
\Gls{O3f} data logging is started, filter holder is loaded with filter and attached to the system.}
|
|
\item{
|
|
Start logging \gls{SMPSa} data and turn on \gls{O3f} generator. Let the system equilibrate for \qty{45}{\minute} (for \num{15} samples)}
|
|
\item{
|
|
Just before ($\approx\qty{10}{\second}$) \num{15} samples have been recorded, open the \gls{VOCa} valve (PV1). At the onset of the next sample, set $\gls{VFLOWs}_{\textnormal{VOC}}=\qty{0.04}{\liter\per\minute}$. Let the system run for \qty{1}{\hour} (for \num{20} samples).}
|
|
\item{
|
|
When \qty{1}{\hour} has passed, and at the onset of the sample subsequent to the ones recorded in the last step, reset $\gls{VFLOWs}_{\textnormal{VOC}}=\qty{0}{\liter\per\minute}$. Within $\approx\qty{10}{\second}$, close the \gls{VOCa} valve (PV1). Let the system equilibrate for \qty{30}{\minute}. }
|
|
\end{enumerate}
|
|
|
|
\subsection{Oxidation Flow Reactor}
|
|
\subsection{Scanning Mobility Particle Sizer}
|
|
\label{methods: subsec: SMPS}
|
|
|
|
The aerosol concentration is measured using a \gls{SMPSa}, which in turn consists of a neutraliser, a \gls{DMAa} and a \gls{CPCa}, arranged in series. The \gls{DMAa} is ramping a time-dependent voltage profile $U(t)$ between its two electrodes. This creates a time-varying electric field of the strength $E\approx U/(d_{o}-d_{i})$ which in turn causes particles within the \gls{DMAa} to experience a drift in this field of velocity $v=Z_{p}E$, where $Z_{p}$ denotes the electrical mobility.
|
|
|
|
\begin{equation}
|
|
m_{\textnormal{F}}=\int_{t_{\textnormal{in}}}^{t_{\textnormal{out}}}\gls{MFLOWs}\,\der{t}\approx\sum_{\textnormal{Samples}}\gls{MFLOWs}_{i}\Delta t_{i}
|
|
\end{equation}
|
|
|
|
\subsection{Ozone analyser and other instrumentation}
|
|
|
|
|
|
\begin{longtblr}[
|
|
caption = {Overview over filters and flows},
|
|
entry={Overview over filters and flows},
|
|
label = {tab: OFR: Overview over filters and flows}
|
|
]{
|
|
colspec ={X[0.1, l] X[0.1, c] X[0.1, c] X[0.1, c] X[0.1, c] X[0.1, c] X[0.1, c] X[0.1, c] X[0.1, c] X[0.1, c]},
|
|
width =\textwidth,
|
|
rowsep =2pt,
|
|
rowhead =2,
|
|
rowfoot =0,
|
|
row{odd} ={font={\small\normalfont}},
|
|
row{even} ={font={\small\normalfont}}}
|
|
|
|
% Header
|
|
\toprule
|
|
Date&
|
|
\ensuremath{\gls{VFLOWs}_\textnormal{VOC}}&
|
|
\ensuremath{\gls{VFLOWs}_{\ce{N2}}}&
|
|
\ensuremath{\gls{VFLOWs}_{\ce{O2}}^{(3)}}&
|
|
\ensuremath{\gls{VFLOWs}_{\ce{O2}}^{(4)}}&
|
|
\ensuremath{\gls{VFLOWs}_\textnormal{P}}&
|
|
\ensuremath{\gls{VFLOWs}_{\ce{O3}}}&
|
|
\ensuremath{\gls{VFLOWs}_\textnormal{F}}&
|
|
\ensuremath{\gls{VFLOWs}_\textnormal{CPC}}&
|
|
\ensuremath{\gls{VFLOWs}_{\ce{O3}}^{(A)}} \\
|
|
&
|
|
\glssymbol{VFLOWs}&
|
|
\glssymbol{VFLOWs}&
|
|
\glssymbol{VFLOWs}&
|
|
\glssymbol{VFLOWs}&
|
|
\glssymbol{VFLOWs}&
|
|
\glssymbol{VFLOWs}&
|
|
\glssymbol{VFLOWs}&
|
|
\glssymbol{VFLOWs}&
|
|
\glssymbol{VFLOWs}& \\
|
|
\midrule
|
|
a&a&a&a&a&a&a&a&a&a \\
|
|
\bottomrule
|
|
\end{longtblr}
|
|
|
|
\section{Filter extraction and TOC-analysis}\label{methods: section: filter extraction and TOC analysis}
|
|
|
|
In order to quantify the exact (organic) carbon content used in \gls{CWFTRa} samples, the organic carbon content on filters needs to be known. While an estimate is provided via the \gls{SMPSa} measurements, the exact quantity of organic carbon on the filters is determined using \gls{TOCa} analysis. For this, \gls{NPOCa} and \gls{TOCa} of filter samples are measured. While the pertaining instrument, Shimazdu TOC Analyser 200, directly provides measurements of \gls{NPOCa}, the value of \gls{TOCa} has to be computed via measurements of \gls{TCa} and \gls{ICa}, respectively:
|
|
|
|
\begin{equation}
|
|
\gls{TOCs}\coloneqq\gls{TCs}-\gls{ICs}
|
|
\end{equation}
|
|
|
|
The values of \gls{TOCa} and \gls{NPOCa} are measured in \unit{\milli\gramCarbon\per\liter}.
|
|
|
|
\subsection{Preparation}
|
|
|
|
Preparation for \gls{TOCa} measurements essentially consists in cleaning glassware required for analysis and extraction. This is rather not straightforward, and two procedures need to be distinguished: a special procedure before the campaign to obtain a standard of cleanliness, and a procedure during the campaign to retain the obtained standard.
|
|
|
|
\subsubsection{Intial Glassware Cleaning}
|
|
|
|
During initial cleaning, \gls{NPOCa} of glassware is tested. Glassware is considered clean if $\gls{NPOCs}<\qty{0.1}{\milli\gram\per\liter}$.
|
|
|
|
\subsubsection{Subsequent Glassware cleaning}
|
|
|
|
\subsection{Extraction}
|
|
|
|
\subsection{Calibration}
|
|
|
|
Calibration for \gls{TOCa} and \gls{NPOCa} measurements require that the instrument be calibrated prior to measurements. For this, stock solutions of known quantites are related to the instrument response, i.e. area under the voltage curve. For \gls{TOCa} measurements, two stock solutions are required, one containing a known quantity of \gls{TCa}, and one where the exact amount of \gls{ICa} is known. In practice, \gls{TCa} calibration can be performed using a \gls{TOCa} stock solution. This is possible because the used standard of \gls{TOCa} is practically free of any \gls{ICa} --- for the used \gls{TOCa} standard, it therefore holds that $\gls{TCs}\approx\gls{TOCs}$.
|
|
|
|
A \gls{TOCa} solution of \qty{1000}{\milli\gramCarbon\per\liter} is used to create a \gls{TOCa} calibrant stock solution of \qty{4}{\milli\gramCarbon\per\liter}. For this, exactly \qty{1}{\milli\liter} of \qty{1000}{\milli\gramCarbon\per\liter} solution is mixed with \qty{244}{\milli\liter} MQ. Similarly, an \gls{ICa} stock solution of \qty{1000}{\milli\gram\per\liter} is used to mix an \gls{ICa} calibrant stock of \qty{4}{\milli\gramCarbon\per\liter} by mixing \qty{1}{\milli\liter} of stock solution with \qty{244}{\milli\liter} MQ.
|
|
|
|
|
|
\subsection{Measurements}
|
|
|
|
\begin{figure}
|
|
\centering
|
|
\includegraphics{06_Solution_Creation.pdf}
|
|
\caption[Procedure to create solutions]{Procedure to create solutions. Beaker apapted from: Diogo Losch De Oliveira (2026). Beakers [Illustration]. CC-BY 4.0. SciDraw.io. https://scidraw.io/drawing/beakers-ba8583ed, accessed 10. August 16:16}
|
|
\label{methods: fig: solution preparation}
|
|
\end{figure}
|
|
|
|
\section{CWFTR experiments}
|
|
\label{methods: section: CWFTR Experiments}
|
|
|
|
\subsection{Preparation}
|
|
|
|
\subsection{Coating solution Preparation}
|
|
\subsection{Procedure}
|
|
|
|
\section{Data analysis}
|
|
\label{methods: section: data analysis}
|
|
|
|
\end{document} |