Analytical Methods in Oceanography

---

2 Trace Metal Contamination by Oceanographic Samplers A Comparison of Various Niskin Samplers

Downloaded by CORNELL UNIV on September 29, 2016 | http://pubs.acs.org Publication Date: June 1, 1975 | doi: 10.1021/ba-1975-0147.ch002

and a Pumping System D O U G L A S A. SEGAR and G E O R G E A . B E R B E R I A N National Oceanic and Atmospheric Administration, Atlantic Oceanographic and Meteorological Laboratories, 15 Rickenbacker Causeway, Miami, Fla. 33149

Simultaneous collections were made of water and suspended sediments by an Inter-Ocean pump sampling system, Niskin bottles with internal rubber closures, Niskin bottles with Teflon-coated coil springs, and newly designed Niskin bottles without internal closures. The samples were examined for contamination, particularly of trace metals. All the sampling systems led to some metal contamination of the samples except the new Niskin bottles. However, each system was found acceptable for nutrient analysis. Because of its novel closure, the newly designed Niskin bottle is convenient for suspended material analysis. Samples may be filtered directly from the sampling bottle without contacting the atmosphere. ' T ' h e analysis of seawater for trace components, particularly metals, is a notoriously difficult task. Sophisticated and expensive analytical instruments are used, and many difficulties are incurred in the procedure. Considerable efforts have been expended on analytical problems ( J , 2) and problems of sample contamination from storage bottles and reagents used for analysis ( 3 ) . However, little attention has been paid to the problem of sample contamination during the process of obtaining the seawater sample from the ocean. Solving the analysis problems is of little use i f uncontaminated samples cannot be obtained. I n a separate paper i n this volume, a new analytical technique is described, which 9 Gibb; Analytical Methods in Oceanography Advances in Chemistry; American Chemical Society: Washington, DC, 1975.

10

ANALYTICAL METHODS IN OCEANOGRAPHY

eliminates the worst of the analytical problems for several elements (4). These new methods were developed for use i n the National Oceanic and Atmospheric Administration's ( N O A A ) Marine Ecosystem Analysis ( M E S A ) Project and for studies of water column chemistry over m i d ocean ridges. This paper describes a parallel effort to test sampling systems for trace metal contamination and to develop a sampling pro­ cedure which minimizes contamination.

Downloaded by CORNELL UNIV on September 29, 2016 | http://pubs.acs.org Publication Date: June 1, 1975 | doi: 10.1021/ba-1975-0147.ch002

Sampling Systems T w o basically different systems, pumps and sampling bottles, are used to obtain seawater samples. Bottles such as the Nansen, Niskin, National Institute of Oceanography ( Ν Ι Ο ) , and V a n D o r n samplers have been used extensively throughout the world ocean. Pumping systems have been used less frequently, usually only i n shallow water and when large volume samples or continuous profiles are required. A review of the literature reveals that little has been published regarding contamination of samples from sampling bottles. Cooper (5), i n reviewing the prob­ lems of seawater sampling, pointed out that contamination, particularly from metal and rubber components of the samplers, was a serious prob­ lem. Subsequently, sampling bottles made from polyvinyl chloride ( P V C ) were almost universally adopted for trace metal analysis. Unfor­ tunately, the standard versions of such sampling bottles a l l use either rubber end caps or rubber "springs" inside the bottle as a closure mecha­ nism. The most widely used sampler, the Niskin bottle, has silicone internal rubber closures that pass through the sampler and remain in contact with the sample after the bottle is closed. Such silicone rubber is reported to give off considerable quantities of zinc to solution. In addi­ tion, the rubber, particularly when new, gives off large quantities of particles to the water sample during the closing process (6). To solve this problem many researchers adopted a Teflon-coated stainless steel coil spring to replace the rubber. Unfortunately, no investigation of the contamination generated from this spring has been reported. However, Teflon is known to be porous, particularly i n thin coatings or films, and the coatings themselves tend to crack with even very limited use. Faced with these sampler problems, a new design of P V C Niskin bottle was constructed which eliminated the internal closures (7). This bottle (Figures 1 and 2) is designed to be closed from the outside. The top and bottom plugs are both loaded at the top of the bottle before lowering. Upon triggering, the bottom plug free falls within the bottle until it closes the bottom end of the bottle by resting on an O-ring seal. The top plug is pulled into place by three tensioned latex rubber bands outside the bottle and then locked i n place by three spring-loaded cams.

Gibb; Analytical Methods in Oceanography Advances in Chemistry; American Chemical Society: Washington, DC, 1975.

Downloaded by CORNELL UNIV on September 29, 2016 | http://pubs.acs.org Publication Date: June 1, 1975 | doi: 10.1021/ba-1975-0147.ch002

2.

SEGAR A N D

Figure 1.

BERBERiAN

Contamination by Océanographie

Top drop Niskin bottle

Figure 2.

Samplers

11

Top drop Niskin bottle loaded

The sample thus comes i n contact only with P V C during its time i n the sampler. The pump sampling system used i n this study was an Inter-Ocean O S E AS system (7) which features a submersible, multi-stage axial flow pump. A l l metal parts of this system are Teflon coated so that samples may be collected for trace metal analysis. Comparisons of Sampling Systems U p o n acquisition of the Inter-Ocean pump sampling system, a series of seawater samples was collected from N e w York Bight at 10 m depth simultaneously with the pump and with Niskin 10-1. samplers with internal rubber closures. The samples were filtered through 0.4/x Nucleopore filters. One aliquot was frozen for subsequent analysis of nitrate, nitrite, phosphate, and silicate by standard automated techniques. A second aliquot was acidified to p H 1 with silica-distilled, concentrated nitric acid and was subsequently analyzed b y flameless atom reservoir atomic absorption spectrophotometry, both b y direct injection (4) and after extraction of the metal pyrollidine dithiocarbamates with methyl isobutyl

Gibb; Analytical Methods in Oceanography Advances in Chemistry; American Chemical Society: Washington, DC, 1975.

12

ANALYTICAL METHODS IN OCEANOGRAPHY

Table I. Phosphate ( μς atβ) Station S

Downloaded by CORNELL UNIV on September 29, 2016 | http://pubs.acs.org Publication Date: June 1, 1975 | doi: 10.1021/ba-1975-0147.ch002

6 8 12 13 22 25 α

Nitrate ( μς at/l)

Pump

Niskin

Pump

Niskin

0.70 1.2 0.64 0.92 0.78 0.63

0.69 1.4 0.58 0.76 0.74 0.64

3.0 7.8 2.6 3.1 3.5 2.6

2.9 9.6 2.2 3.0 3.5 2.5

Comparison of Pump and Nitrite ( μ$ atβ)

Silicate ( μς at/l)

Pump Niskin 0.21 0.94 0.14 0.30 0.50 0.28

0.26 1.1 0.10 0.30 0.51 0.28

Pump