Performance of a corona ion source for measurement of sulfuric acid by chemical ionization mass spectrometry

Introduction Conclusions References


Introduction
The measurement of gaseous sulfuric acid is important since H 2 SO 4 is one of the key compounds responsible for atmospheric new particle formation (Curtius, 2006;Kulmala and Kerminen, 2008).The nucleation of particles has been observed in many places around the world on ground-based measurement sites as well as in the free troposphere (Kulmala et al., 2004).In most cases the formation rate of new particles correlates with the concentration of sulfuric acid (Yu and Turco, 2001;Fiedler et al., 2005;Kuang et al., 2008).The concentration of H 2 SO 4 during atmospheric nucleation events is usually between 10 6 and 10 7 molecules per cm 3 (Sipil ä et al., 2010), i.e. in the sub-ppt range under standard conditions.Therefore, the precise and accurate measurement of sulfuric acid is essential for studying new particle formation under atmospheric conditions as well as during chamber experiments.Introduction

Conclusions References
Tables Figures

Back Close
Full For the real-time measurement of sulfuric acid, chemical ionization mass spectrometry (CIMS) is generally deployed.CIMS is a very sensitive and selective method and detection limits around 10 5 molecules of H 2 SO 4 per cm 3 and one minute integration time can be reached (Eisele and Tanner, 1993;Young et al., 2008).While most instruments initiate the production of the NO − 3 primary ions -which are used for H 2 SO 4 -CIMS -through the decay of a radioactive substance, these ions can also be generated by a corona discharge.The reason why radioactive ion sources are generally used (usually alpha emitters like Polonium 210 or Americium 241) is probably because these sources are known to be more stable over time and produce cleaner mass spectra, i.e. create lower background concentrations and less interference with other substances like SO 2 .However, with respect to health risk, cost and meeting safety regulations for shipment, storage and operation, corona ion sources have a clear advantage over their radioactive counterparts.
Corona ion sources for CIMS instruments have been described and used by several groups for the measurement of OH and peroxy radicals (Kukui et al., 2008), other atmospheric trace gases like SO 2 , acetonitrile and acetone (Jost et al., 2003) as well as HNO 3 (Furutani and Akimoto, 2002).Using drift-chemical ionization mass spectrometry, corona ion sources have also been deployed for the detection of HNO 3 , N 2 O 5 and isoprene (Zheng et al., 2008;Fortner et al., 2004).Some of the corona ion sourcebased measurements show a higher complexity as compared to the use of radioactive ion sources (Kukui et al., 2008) while they suffer at the same time from higher detection limits due to radicals produced by the corona discharge (Kukui et al., 2008;Jost et al., 2003).Our findings, however, do not support these observations, something which we relate to the special design of the ion source region and the ion drift tube of the instrument being used in this study.
Here, we describe in detail the set-up of a corona-type ion source and report on its performance when used with a chemical ionization mass spectrometer from THS Instruments (THS Instruments LLC, USA).The results obtained with this ion source are compared to the ones from an americium ion source, a device generally used Introduction

Conclusions References
Tables Figures

Back Close
Full for this type of CIMS measurement.It is demonstrated that the corona ion source works stable and reliably and shows a negligible cross-sensitivity to SO 2 for sulfuric acid measurements.Most importantly the corona ion source yields almost identical quantitative results as an americium ion source for the same instrument.

Instrumental description
The corona ion source has been developed for a chemical ionization mass spectrometer from THS Instruments (THS Instruments LLC, USA) for the detection of sulfuric acid.The measurement relies on the reaction between NO − 3 primary ions generated in the ion source and sulfuric acid in the sample gas (see Eisele and Tanner, 1993;Berresheim et al., 2000).Originally, we had the instrument equipped with an americium containing thin gold-plated foil (NRD LLC, USA) for providing the primary ions.A schematical drawing of this ion source is shown in Fig. 1.This figure, showing the old set-up, is used to illustrate the working principle of the instrument while the new corona ion source will be introduced and explained in detail further below.

CIMS and americium ion source
The sample gas containing the sulfuric acid is pulled into the ion drift region through a stainless steel tube with an outer diameter of 12.7 mm.It is then exposed to NO − 3 (HNO 3 ) x with x = 0-2 primary ions which can react with H 2 SO 4 to form HSO − 4 (HNO 3 ) x ions (Eisele and Tanner, 1993;Viggiano et al., 1997).The primary ions in the old set-up originate from the interaction of alpha particles from the radioactive decay of americium ( 241 Am) with the sheath gas.This sheath gas consists of room air cleaned by an activated charcoal as well as by a HEPA filter and has a total flow rate of approximately 22 standard liters per minute (slm).A small amount of HNO 3 (∼0.005slm of N 2 saturated with HNO 3 at room temperature) is added to the sheath gas which leads to the generation of NO − 3 (HNO 3 ) x primary ions through a series of Introduction

Conclusions References
Tables Figures

Back Close
Full ion/molecule interactions.In order to shield the sample gas from radicals produced in the region around the americium foil there is an annular space between the 12.7 mm tube and the cylinder which holds the americium foil.Sheath gas flowing through this region avoids that radicals (especially OH) mixes with the sample gas and can interact e.g. with SO 2 to produce spurious amounts of H 2 SO 4 .The generated NO − 3 (HNO 3 ) x ions, however, are mixed with the sample gas.This is achieved by applying electrostatic voltages to different sections of the ion source, the sample tube and the ion drift tube, respectively, thereby focusing the primary ions towards the centerline of the ion/sample gas interaction region.While the sample and sheath gas mixture containing the neutral molecules is pumped away, the ions are accelerated towards the pinhole plate which separates the vacuum region of the mass spectrometer from the ion drift region.A small flow of dry nitrogen is used to evaporate excess water from the ions.
Behind the pinhole plate the ions are guided through an octopole where most of them are declustered by removing excess HNO 3 and remaining H 2 O molecules.The primary ions and the analyte ions are then detected by a quadrupole mass spectrometer and a channeltron at m/z 62 (NO − 3 ) or m/z 64 (NO 18 2 O − ) and m/z 97 (HSO − 4 ), respectively.The ratio of the measured count rates, together with the reaction time and the reaction rate yields the concentration of sulfuric acid in the sample gas (Berresheim et al., 2000).

Corona ion source
The same measurement principle has been used for the set-up with the corona ion source and only a few adjustments were made to the outer stainless steel cylinder which houses the ion source (the part where −220 V are applied to, see Fig. 1).The overall length of the outer cylinder has been increased from 40 to 60 mm in order to eliminate any contributions to the electric field in the ion drift region from the electric field of the corona needle.If the corona needle is too close to the edge where the inner cylinder terminates then the high voltage applied to the needle can distort the electric field and destroy its cylindrical symmetry.Therefore, the needle has been 5299 Introduction

Conclusions References
Tables Figures

Back Close
Full moved backwards (in the direction of flows) in order to shield it with respect to the ion/sample gas mixing zone (see Fig. 2).For the corona needle we have chosen a needle from acupuncture supply (Moxom SP-X Gold, Moxom Acupuncture GmbH, Germany).These needles have the advantage that they are industrially machined with high precision and therefore differ very little from each other.Another advantage is that they are gold-plated and are therefore not dissolved by the HNO 3 in the sheath gas.Additionally, a comparison of different corona needle materials has shown that needles made of gold show a preferable combination of both durability and relatively low production of NO x and O 3 (Asbach et al., 2005).The type of needles used here has a cylindrical shape at its blunt end with an outer diameter of ∼1.3 mm.Although this size is still rather small it allows handling and aligning the needle rather easily compared to other thinner needles or wires.An adapter piece made of PEEK (polyether ether ketone) was machined which allows installing and exchanging a needle easily.
Due to these advantages we haven't observed any change in the performance of the ion source after a needle has been exchanged meaning that results are very well reproducible.The overall length of the needle from its sharp tip to the cylindrical end is ∼3.1 mm.The needle is held by a little ring made of stainless steel which centers it inside a cylindrical bore of the PEEK part.This part has a tiny through-hole which allows the tip of the needle to penetrate but not its cylindrical end.The other end of the PEEK piece has an internal thread to fit a standard o-ring-sealed high voltage feedthrough (SHV type).The electrical contact between the SHV connector and the needle is established by means of a small stainless steel compression spring (Century Spring Corp., USA) and a stainless steel disc.The PEEK part has also an outer thread and can be attached and sealed to the stainless steel ion source housing with an o-ring.
Instead of aligning the needle tip in the center of the annular space between the inner and the outer cylinder, we have decided to move the needle slightly further towards the outside wall.The idea behind this is that radicals produced close to the needle surface are further away from the centerline of the ion source and are therefore less likely to Introduction

Conclusions References
Tables Figures

Back Close
Full interact with the sample gas.Instead of using just one corona needle we have also tried-out a set-up with four corona needles aligned 90 • apart from each other when viewed perpendicularly to the gas flow direction.However, this set-up did not improve the overall functionality.We think that space charge effects might suffice distributing the ions homogenously even when only one point source for the ion generation is used (see also discussion in Sect.3.1).
The electronics driving the corona needle voltage is relatively simple.A negative high voltage supply (model 4100N, EMCO High Voltage Corporation, USA) is used which is set to a fixed voltage (adjusted between −5 kV and −7.5 kV).The high voltage is connected in series with a resistor (500 MOhm) and the corona needle.The corona onset voltage is around −3 kV (with respect to the −220 V applied to the ion source housing).Therefore, the resistor limits the current between ∼4 µA and 9 µA.Having the resistor in series with the corona needle both limits and stabilizes the corona current since random changes in the corona onset voltage (e.g.due to changes in the gas humidity) then translate only into small changes of the corona current.Therefore, the ion current is quite stable even without actively controlling for a defined corona emission current (e.g. with a PID controller).

Results and discussion
Desired features for a H 2 SO 4 -CIMS ion source are a simple set-up and low maintenance, low background signal, especially no cross-sensitivity to SO 2 , long-term stability and most importantly the capability of performing quantitative measurements.As mentioned in the previous section the corona ion source can be set-up easily and replacement needles can be purchased inexpensively and exchanged quickly.Introduction

Conclusions References
Tables Figures

Back Close
Full

Comparison between corona and americium ion source for given H 2 SO 4 concentrations
The measurement of different amounts of H 2 SO 4 has been tested both for the americium ion source and the corona ion source with set-ups according to Figs. 1 and 2, respectively.The H 2 SO 4 has been provided by an external calibration source which provides stable and defined concentrations of H 2 SO 4 .The detailed set-up of this calibration source will be described in a forthcoming paper and is therefore only briefly explained here.A gas mixture of N 2 , O 2 , SO 2 and H 2 O is illuminated by 185 nm UV light from a mercury lamp which photolyzes water vapor and leads to the formation of OH.This OH reacts further with SO 2 , O 2 and H 2 O to form sulfuric acid (see also Young et al., 2008).Varying the amount of H 2 O yields different concentrations of H 2 SO 4 .The calibration of our CIMS has been performed once in June 2010 when we were still using the americium ion source.After switching to the corona ion source, the calibration has been repeated with the same calibration source under similar conditions.As can be seen from Fig. 3 the instrument shows almost identical responses to a given H 2 SO 4 concentration within the uncertainty range, no matter whether the corona ion source or the americium ion source is used.Figure 3 also shows the linear response of the CIMS to different concentrations provided, although it seems from this data that the linearity is only valid for larger H 2 O concentrations while for the lower moistures the response is rather quadratic or exponential.This feature, however, arises from the very low moistures within the calibration system.At low humidities the generated concentration of H 2 SO 4 is controlled by the reaction which introduces the quadratic dependence to water vapor (Lovejoy et al., 1996;Jayne et al., 1997).This feature of our calibration set-up will also be explained in more detail in a separate paper.
The observation that the correspondence between the corona and the americium ion source is nearly perfect is not necessarily expected given the fact that the ions from the corona needle are generated only by a point source whereas for the americium source they are produced over a complete circular cross section (neglecting the small gap between the edges of the americium foil when it is wrapped around the inner cylinder).However, our findings might suggest that space charge distributes the ions homogenously in the ion drift region.Another, maybe more likely explanation is that even when the primary ions are not homogenously distributed over the sample gas volume, only NO − 3 ions which have been mixed with the sulfuric acid in the sample gas enter the mass spectrometer.Therefore, the ion count rates for HSO − 4 and NO − 3 are always at a defined ratio for a given H 2 SO 4 concentration and no NO − 3 ions that were not exposed to sulfuric acid in the sample gas are counted.This behavior for the ion collection can be explained by the special design of the interface between the ion drift region and the pinhole plate through which the ions enter the vacuum of the mass spectrometer.In front of the pinhole an electrostatic lens (applied voltage of −100 V, see Fig. 1) with an internal diameter that is equal to the one of the sampling tube allows only ions to pass which are close to the centerline.Therefore, primary ions close to the outside wall of the ion drift tube are not detected and do not bias the ratio between analyte and reagent ions.

Signal stability
Figure 4 shows a time series of the signal at m/z 64 over a period of more than two hours.We use m/z 64 (NO 18 2 O − ) as our primary ion signal since the count rate at m/z 62 (NO was measured.Then a stable concentration of H 2 SO 4 was produced in our calibration set-up.The signal of the primary ion shows some fluctuations and a small drift can be identified with a slightly increasing ion count rate over time.These changes were not observed as strongly with the americium ion source and we attribute the higher "noise" of the corona ion source to changes in the corona onset voltage which might occur due to temperature changes or slight changes in the sheath gas composition (humidity).However, these changes are rather small and slow.In addition, since [H 2 SO 4 ] is proportional to the ratio of the count rates at m/z 97 and m/z 64 these drifts do not show up in the H 2 SO 4 signal.Over longer time scales of up to weeks the count rate of the primary ion signal stays within ±50%.This means that the same corona needle can be used for rather long times and that no observable limitation on the detection limits is occurring.

Cross-sensitivity to SO 2 and detection limit
The dependence of the H 2 SO 4 measurement on the SO 2 concentration is also shown in Fig. 4. At 12:30 a concentration of 1 ppmv of SO 2 was introduced into the sample gas while no SO 2 had been actively introduced before.The average H 2 SO 4 concentration increases from 2.9 × 10 4 to ∼6.1 × 10 4 cm −3 due to the higher SO 2 concentration.When taking into account the point to point fluctuations for the signals, detection limits (based on the averaged one minute values plus three times their standard deviation) of 6.4 × 10 4 cm −3 and 1.1 × 10 5 cm −3 are obtained for the periods without and with the addition of SO 2 , respectively.Even the higher value is still better than what has been reported from another group for a similar CIMS with a polonium ion source (Young et al., 2008).The concentration of sulfur dioxide which was added for the example shown here is large in comparison to concentrations usually observed at ambient conditions even in moderately or strongly polluted cities (Bari et al., 2003;Yang et al., 2009).Considering the large amount of SO 2 added to the sample gas, a twofold increase in the background H 2 SO 4 seems to be a negligible contribution and indeed, no cross-sensitivity to SO 2 5304 Introduction

Conclusions References
Tables Figures

Back Close
Full is observed when only 20 ppbv are present in the sample gas which was tested in a separate experiment.This small contribution of SO 2 to the signal is surprising as other groups have reported that an OH scavenger like propane (Berresheim et al., 2000) or C 3 F 6 (Dubey et al., 1996;Sjostedt et al., 2007) were used for similar instruments in order to reduce the background levels of OH or H 2 SO 4 to an acceptable level.From our findings this seems not to be necessary for [SO 2 ] below 20 ppbv (or even higher) when using CIMS for H 2 SO 4 detection -even in conjunction with a corona ion source, a device normally known for producing higher background levels of unwanted substances in comparison to radioactive ion sources.

Summary
The performance of a corona ion source for the detection of sulfuric acid by chemical ionization mass spectrometry has been evaluated.The corona ion source consists of a simple set-up and makes use of a commercially available gold-plated acupuncture needle which can easily be installed and exchanged.
The results show that this source yields the same quantitative results for a given H 2 SO 4 concentration as compared to an americium ion source.The detection limit for the corona ion source is ∼6 × 10 4 molecules of H 2 SO 4 per cm 3 and one minute integration time.While the fluctuations of the primary ion signal are higher for the corona ion source than for the radioactive ion source, this higher noise translates only slightly into the determined H 2 SO 4 concentration since the ratio between HSO − 4 ions and NO − 3 primary ions is stable.The corona ion source only shows a slight increase in background levels of H 2 SO 4 when SO 2 at a high concentration of 1 ppmv is added to the sample gas and the detection limit then increases to ∼1.1 × 10 5 cm −3 .For smaller concentrations (20 ppbv) of sulfur dioxide no contribution to H 2 SO 4 was observed even without adding an OH scavenger to the sheath gas.Given these characteristics of the corona ion source, the more commonly used radioactive ion sources seem to have no substantial advantage over a corona-type ion source.This is especially the case when Introduction

Conclusions References
Tables Figures

Back Close
Full considering the strict safety regulations, the potential health risk and the higher costs which apply for the radioactive sources.
In future, we would like to improve the stability of the corona ion source and try to produce higher count rates for the primary ion.Making the insulating PEEK part smaller could minimize the accumulation of charges on the insulator surface and thereby decrease losses for ions.In addition, trying out different radial positions for the corona needle tip could lead to an increase in the signal, thereby improving the detection limits.Also the effect of a PID controller could be tested.Introduction

Conclusions References
Tables Figures

Back Close
Full Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | For higher humidities (here approximately [H 2 O] > 5 × 10 15 cm −3 ) the dependence of produced H 2 SO 4 on humidity becomes linear because then its formation is controlled rather by the abundance of OH which is directly proportional to [H 2 O]. Discussion Paper | Discussion Paper | Discussion Paper |

− 3 )
can produce a strong signal which might saturate the channeltron detector.The data in the figure shows raw signals with a time resolution of 5 seconds as well as values averaged over one minute.Until 12:57 the background of H 2 SO 4 Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper |

Fig. 1 .Fig. 2 .Fig. 3 .
Fig. 1.Schematical drawing of the ion source and the inlet system of the instrument used for the H 2 SO 4 measurements.The colors represent different voltages applied to the different sections of the inlet and ion source region.Note that the voltage applied to the americium foil is −220 V as well but has been assigned a different color for clarity.