A portable, robust, stable, and tunable calibration source for gas-phase nitrous acid (HONO)

Abstract. Atmospheric HONO mixing ratios in indoor and outdoor environments span a
range of less than a few parts per trillion by volume (pptv) up to tens of
parts per billion by volume (ppbv) in combustion plumes. Previous HONO
calibration sources have utilized proton transfer acid displacement from
nitrite salts or solutions, with output that ranges from tens to thousands
of ppbv. Instrument calibrations have thus required large dilution flows to
obtain atmospherically relevant mixing ratios. Here we present a simple
universal source to reach very low HONO calibration mixing ratios using a
nitrite-coated reaction device with the addition of humid air and/or HCl
from a permeation device. The calibration source developed in this work can
generate HONO across the atmospherically relevant range and has high purity
(> 90 %), reproducibility, and tunability. Mixing ratios at
the tens of pptv level are easily reached with reasonable dilution flows.
The calibration source can be assembled to start producing stable HONO
mixing ratios (relative standard error, RSE ≤ 2 %) within 2 h, with output
concentrations varying ≤ 25 % following simulated transport or
complete disassembly of the instrument and with ≤ 10 % under ideal
conditions. The simplicity of this source makes it highly versatile for
field and lab experiments. The platform facilitates a new level of accuracy
in established instrumentation, as well as intercomparison studies to
identify systematic HONO measurement bias and interferences.


The thermal system is composed of three critical components: a temperature controller, temperature sensor and a heat source. All components were purchase from either Omega TM Environmental (St. Eustache, QC) or Allied Electronics, Inc. (Fort Worth, TX). The temperature controller regulates the temperature of the system by receiving an input signal (measurement) from the thermocouple to compare to its setpoint value. An output signal will be generated if the input value is below the temperature-control setpoint. The output signal is sent to a solid-state relay which enables the heater function and increases the temperature. A proportional-integral-differential (PID) temperature controller (Omega TM ; CN 7823) monitors and sets the temperature of the Al block shown in Fig. S4. Using PID mode regulates the temperature of the system automatically by using feedback from the thermocouple and associated heating element to compensate for any changes relative to the set point. The PID controller is programmed to match the appropriate thermocouple (Type K used here). The PID parameters are adjusted by using the auto tune function to give the precise control (±0.1 °C) without overshooting the set point. The solid-state relay (Allied; SSR, D1210, Crydom) switches the heating cycle by converting the 5 V DC output from the temperature controller to 120 V AC used to operate the heating element. To dissipate the energy required for switching, the SSR was mounted on an aluminium heat sink (Allied; HS172, Crydom) and protected by a Bussman rectifier fuse (Omega TM ; Tron, KAX-10 A) against excess current flow that could result from a latching failure. The fuse holder (Omega TM ; FB-1) allows easy replacement of a blown fuse in the circuit. A cartridge heater with an integrated K-type thermocouple housed in a high-temperature incoloy sheath (Omega TM ; CIR, 300 W, 120 V) is inserted into a 20.3 cm (8") hole with a 3/8" internal diameter, on the long axis of the Al-block. The heater is a high watt density cartridge with maximum working temperature of 760 ºC and F-type leads with fiberglass insulation. The thermocouple measures temperature by the thermoelectric effect which results in a voltage. A K-type thermocouple consists of chromel (Ni-Cr) and alumel (Ni-Al) alloys that generates an accurate voltage in the temperature range of 0 to 1250 ºC. An upper limit on the temperature controller should be set at 150 °C to prevent thermal degradation of PFA tubing in the Al-block during use. Figure S1. Three (a) and two (b) dimension layout schematics of permeation-oven components on the bent aluminium plate (1), with mounted temperature controller (2), solid state relay (3) and its heat sink (4), electrical fuse (5), cartridge heater with integrated thermocouple (6), aluminum block (7), source of dry compressed air (8), 2-way gas valve (9), critical orifice (10), PFA oven (11), and gaseous output to external system. Black lines represent PFA lines guiding gas flows throughout the system. Figure S2. Dimensions of bent aluminium plate and cut-out measurements to mount the temperature controller. Further holes for valves or a holder for the water impinger were created with an electric drill on an as-needed basis.

S1.2. Electrical connections for temperature control feedback
Power is supplied to the entire setup from 120 V AC outlet capable of providing up to 15 A of current (Fig.  S4). The temperature controller monitors the voltage signal from the thermocouple (here: yellow/+ and red/-; terminals 4 and 5, respectively; Fig. S4). When the thermocouple signal falls below the control setpoint it sends a 5 V signal (terminals 1(-) and 2(+)) to the solid-state relay (3-32 V DC input D(-) and C(+)). The solid-state relay closes the switch to deliver current to the cartridge heater leads from the AC terminals (120 V AC output A(+) and B(-); Fig. S4). The relay and cartridge heater are protected by a fuse. The entire case is grounded, with a ground wire of the power supply fixed to the bent aluminium plate. Figure S4. Schematic of the wiring and connections of the custom-built permeation oven. The power supply is distributed from the power inlet into the system through different wires consisting of live wires (black), neutral wires (grey) and a ground wire (green). The temperature controller receives signals by inputs 4 and 5 (yellow and red) connected to the thermocouple. To control the function of the heater, the temperature controller sends a signal from output 1 and 2 to input C and D of the solid-state relay to reach the temperature setpoint.

S2. Rationale and construction of custom permeation devices (PDs)
The purpose of a permeation oven is to obtain a consistent quantity of gaseous analyte via a constant permeation rate, which results in a known mass per unit time delivered to an experimental system or an analytical instrument for calibration purposes. A permeation oven is typically coated or made with inert material to limit surface interactions experienced by the generated gas. The calibration gas permeates by diffusion through the porous polymer according to Fick's second law of diffusion (E1, m 2 s -1 ). If the analyte is in an aqueous solution it is emitted into the oven as a vapor based on its effective Henry's Law constant, which is the product of its acid dissociation constant and volatility (E2, Pa mol m -3 ; R4) (Mitchell, 2000;Scaringelli et al., 1970). The product of these two properties combine into the overall permeability term (P, Pa mol m -1 s -1 ; E3).
Where, S0, and D0 are pre-exponential constants of solubility and diffusion. In Equations 1 and 2: ED is the activation energy for diffusion (J mol -1 ) and HS is the enthalpy of solvation (J mol -1 ), while T is temperature (K) and R is the ideal gas constant (J mol -1 K -1 ). The permeation rate depends on temperature because both diffusion and vapour pressure are exponentially temperature dependent. Therefore, permeation ovens can provide a stable and pure emission rate that can be tuned by adjusting the temperature of the oven.
Clean air moves through the ½" PFA oven tubes, where the vapour emissions from PDs are accumulated, to carry them into the reaction devices, a scrubbing solution, or other instrumentation. High precision of a permeation rate can be obtained by regulating and maintaining constant temperature within ± 0.1 ºC because an increase of 1 ºC of operating temperature results in a 10 % change in sample permeation rate (Lucero, 1971). Adequate precision of our custom-built permeation oven was achieved by choosing the temperature controller, heater, and thermocouple with very strict tolerances. Long-term stability of the system can be ascertained from ion chromatography or real-time measurements (see main manuscript). Also, calibrating the permeation rate of a custom-PD can easily indicate the precision and stability of this custom-built permeation oven.
To construct a custom-PD, a length ¼" PFA tube is filled with analyte solution and the ends are plugged. One end of the tubing is initially made malleable with a heat-gun. Once the tube is sufficiently pliable, a rod of porous PTFE (0.125" diameter, P/N: 84935K64; McMaster-Carr) is carefully inserted with a twisting motion and cut to keep the tubing and rod flush. To ensure an effective seal the two materials are heated further, then compressed and rolled on a flat surface. The open-end of the PD is then filled with analyte solution until it is approximately ½" from the top. The PTFE plugging process is repeated to close the tube. Extra caution during this step should be taken to avoid having the solution boil over or undergoing spontaneous ignition during heating. It is worth noting that the length and quality of the seal around the porous PTFE rod can change the emission rate, where a longer plug may yield a more reliable emission rate. Generally, plugs one centimeter in length were used in this work. If the plug length is too short leakages may arise, resulting in rapid and unstable emission of the contained solution. An alternative that avoids PTFE plugs and associated leaks is to weld the tubing ends. This can be accomplished by heating the end of the tubing with a heat-gun until it is malleable, followed by pinching it with a pair of pliers to create the weld. Figure S5. a) Two PDs containing analyte solution fully sealed using the porous PTFE rod (left) and PFA weld (right) techniques. b) Close-up perspective of finished PD ends after sealing by the methods of PTFE rod (top) and PFA weld (bottom). Figure S6. Mixing ratio output of two 6 M HCl PDs as a function of temperature (PD-6b, blue; PD-6c, green). Calculated vapour pressure of 6 M HCl (black) solution using Henry's Law (Sander, 2015). Mean HCl mixing ratios measured for 30 minutes after stable signal was observed by CRDS (1-minute average, see Fig. S7). Error bars denote 1σ standard deviation from the mean. Figure S7: Time series of the measured HCl output from PD-6b using CRDS, as well as temperature of the oven. The blue-shaded bars indicate the region where HCl output was considered stable and this data was used to calculate the variance shown in Figure S6. Table S1. The calculated average HONO (ppbv) output (AVG), standard deviation (SD), and standard error (SE) following the field transport simulations using HCl PD-6a and the same NaNO2 coated device throughout all experiments. The AVG, RSD, and RSE were all calculated from data points collected after two hours of stabilization in each trial run. All experiments were background corrected by linear interpolation using zero air and five experiments had insertions of Na2CO3 denuder for further identification of NOx impurities (*). The NOx analyzer collected one-minute measurements using a 30 s Kalman filter.     Table S1 to demonstrate that measured signal from zero air (black square) and HONO passing through the Na2CO3 annular denuder (red square) are identical and therefore free of detectable NOx. These negative controls were combined to create a linearly interpolated background correction over the course of experiments FS1-5 to quantify HONO. Figure S10. Reproducibility in generating HONO using PD-6b and the same NaNO2 coated device. The red line represents the first trial of using PD-6b followed by a second trial (green) 15 days later. In the intervening time other experiments were performed which involved shutting down and restarting the system with these components. The lag in the second trial results from keeping the PD in the oven while shut down between experiments, resulting in additional HCl that must be reacted in the calibration system before stability is reached. The NOx analyzer mad oneminute measurements using a Kalman filter of 300 s. Figure S11. Determining reproducibility in generating HONO (ppbv) against time (min) using PD-1a with a used NaNO2 PFA device. Figure S12. Time series of measured HONO output (ppb) for three different temperatures using HCl PD-6b. zero Figure S13. Ion scan of the HONO calibration source output made with the I --CIMS. Major ions observed in addition to HONO include glycerol from the coating solution, and lactic acid from skin contact with system components. Smaller quantities of formic and trifluoroacetic acids from the plastics used in the instrument assembly were also observed.