Sampling:
Aerosol samples were collected by a MICRO ORIFIFICE UNIFORM DEPOSIT IMPACTOR (MOUDI) (MSP Corporation, MN, USA) which was installed on the forward rail of Healy's flying bridge, ~23 m above sea level, to minimize the influence of sea spray. To minimize the potential for contamination from the stack exhaust, samplers were forward of the ship's stack and sampling was controlled by wind speed and direction, through a Campbell Scientific CR800 data-logger interfaced with an anemometer and wind vane set up near the samplers. Aerosol sampling was restricted to periods when in-sector conditions (defined as a relative wind direction from within ±60° of the ship's bow and a relative wind speed of >0.5 m s-1) persisted for at least five continuous minutes. The MOUDI impactor used Teflon filters for particle collection (Pall Corp., 47 mm diameter, 1 µm pore size), with a sampling flowrate of 30 L min-1. Both the MOUDI impactor and its pump were housed in enclosures to protect them from rain and sea-spray with an extension tube connected to the MOUDI inlet and extending from the enclosure. A rain shield was installed above the inlet. Due to the anticipated low dust conditions during GN01, and the relatively low frequency of in-sector wind conditions, sample collections lasted for an average of seven days.

Clean polyethylene gloves were worn for loading and unloading of sample filters, which were carried out underneath a high-efficiency particulate air (HEPA) filter blower within a plastic "bubble" clean area constructed in the ship's main laboratory. Filters were loaded onto the MOUDI impactor from labeled petri dishes using pre-cleaned Teflon tweezers and were transferred back to the same petri dishes after sample recovery. Filter holders were double-bagged for transfer between the ship's laboratory and the samplers. Deployment blanks were carried out using the same protocols, but with the pumps turned off. All sample and field blank filters were subsequently double-bagged and stored frozen until analysis.

Sampler Location: Flying deck, forward railings
Sampler Type: MICRO ORIFIFICE UNIFORM DEPOSIT IMPACTOR (MOUDI) (MSP CORP.,MN, USA)
Sampler Flow: 0.030 m3/min
Substrate Type: Pall corp. Teflon Filter, 47 mm discs (Teflon, 1um pore size)
Size Segregation Method: Size-segregated sample; 1um is used as a cut off size for COARSE and FINE particle sizes

Aerosol Sample Analyses:
Total aerosol Fe in this study was measured following the detailed procedures in Morton et al. (2013) and Gao et al. (2013).

Total Fe: Aerosol samples were analyzed for the total concentrations of atmospheric Fe by a sector field inductively coupled plasma-mass spectrometer (SF-ICPMS) in Rutgers Inorganic Analytical Laboratory, following a previously described digestion protocol (Gao et al., 2013). Briefly, a portion of each sample filter was placed in a 15 mL Teflon vial with a mixture of concentrated HNO3 (0.8 ml) and HF (0.1 ml) (Optima, Fisher Sci.) and digested for 4 hours on a hot plate at 160°C. Each digestion solution was evaporated to dryness, followed by the addition of 2 ml 3% HNO3 and 1 ppb Indium (In) for ICP-MS drift correction. Both field blanks and procedure blanks were treated in the same way as samples. All Teflon vials were acid-cleaned, and all procedures were carried out in a class-100 clean-room hood in the lab. The sample digestion procedures were assessed using Standard Reference Material (SRM) 1648a (National Institute of Standards and Technology, NIST, Gaithersburg, MD), subsamples of which were treated under the same conditions as for samples. The digest recoveries based on SRM1648a ranged between 89-99% for Fe (n=7) which was close enough to the measured quantities that no yield correction needed to be applied, and the precision determined from sample splits and duplicate digest aliquots ranged between 93-106% for Fe (n=10). The method detection limits were 0.691 pmol m-3 for Fe, which was obtained based on three times the standard deviation of a total of 14 filter blanks and a nominal 100 m3 sampling volume. A series of external calibration standards were run at the beginning and then at the end of the analyses. More details on the ICP-MS instrument settings can be found in Annett et al. (2017).

Dissolvable Fe: The concentrations of dissolvable Fe in aerosol samples were obtained using UV/Visible spectroscopy with a modified Ferrozine method (Gao et al., 2013). The leaching solution for samples was 0.5 mM ammonium acetate that was filtered through a Nuclepore® track-etch membrane filter (47 mm, 0.2 μm) and adjusted to ~pH 5.3. The leaching conditions were chosen to simulate cloud water conditions for marine aerosols. A brief description of the procedures is as follows: a portion of each sample filter was first placed into a leaching solution of ammonium acetate (0.5 mM) for 1 h, and then the leachate was split into two parts, one for Fe(II) determination and the other for total dissolvable Fe. A solution of 0.01 M hydroxylamine hydrochloride solution (1%) was added to the total dissolvable Fe filtrate portion to reduce Fe(III) to Fe(II), and the sample solution was set aside for 1 h to ensure complete reduction before adding the same ferrozine solution as for the Fe(II) filtrate portion. The Fe(II) measured in this way was considered as total dissolvable Fe. After these procedures, each sample leaching solution was filtered through a 13 mm polytetrafluorethylene syringe filter of 0.2 μm pore size. All field blanks were treated in the same way as samples. The concentrations of Fe(II) in sample solutions were determined at 562 nm using a TIDAS-1 spectrometer module with a 200 cm liquid waveguide capillary flow cell (World Precision Instruments Inc., FL, USA). The detection limit of the method for Fe(II) was 0.30 nM, calculated as three times the standard deviation of the measured blank values (n=5). A total of 8 sets of size-segregated aerosol samples were analyzed with this procedure.

Fe_Coarse.csv