Vol. 116 (2009) ACTA PHYSICA POLONICA A Supp.

Proceedings of the III National Conference on Nanotechnology NANO 2009

Ultrapure Water for Elemental Analysis Down to ppt Levels

D. Darboureta,∗ and I. Kanob
a
Research & Development, Laboratory Water Division, Millipore S.A., Saint-Quentin-en-Yvelines, France
b
Research & Development, Laboratory Water Division, Nihon Millipore Ltd., Tokyo, Japan

Recent enhancements in modern analytical instrumen- plasma — optical emission spectroscopy or mass spec-
tation have dramatically improved the sensitivity of anal- trometry (ICP-OES or ICP-MS) are the main technolo-
ysis. Trace elements are now measured at ppt and sub- gies employed by trace analysis scientists [1].
-ppt levels. These levels can be achieved if careful con-
trol of the analytical protocol is maintained. Contami- 2. Materials and methods
nation can result from anything that comes into contact
with the sample; the laboratory environment, the air, 2.1. The analytical instrument
and anything used during sample preparation. Limits
of detection down to ppt levels can be achieved if spe- ICP-MS is a choice tool for ultratrace analysis. Al-
cial care is taken with the experimental environment as though technical advances are improving the sensitivity
well as with the quality of the reagents. This includes and detection limits of most analytical techniques, in-
the ultrapure water used to run blanks, dilute samples or strumental limitations, such as difficulties in performing
prepare standards. Part of this paper looks at the most multi-elemental simultaneous studies, still remain in ad-
sensitive multi-elemental analytical tool, i.e. inductively vanced analytical technologies [2]. This is why, in many
coupled plasma mass spectrometry (ICP-MS). Some of applications, ICP-MS is often considered as the ideal in-
the most common causes of interference will be reviewed, strument to perform fast, multi trace-elemental analy-
as well as ways of reducing this interference by specific ses [3]. This technology allows fast, qualitative analysis
instrument developments and particular care taken dur- of unknown samples, and quantitative, multi-elemental
ing inductively coupled plasma mass spectrometry analy- analysis down to ppt (ng/L) and even ppq (pg/L) levels.
sis. Sub-ppt values for limits of quantification are achiev- Application fields where this instrument is a key tool are
able when the ultrapure water used is produced by a sys- quite varied: they include the medical area [4], where
tem for which all parts have been carefully selected and heavy metals have a significant impact on health; the en-
whose configuration has been specifically developed for vironmental field, with metal traceability [5]; the nuclear
ultratrace purposes. This specific water purification sys- world, with the ability to perform isotope ratio screen-
tem is described and typical values that can be achieved ings and detect speciation; and the microelectronics in-
for blanks with inductively coupled plasma mass spectro- dustry, with ultratrace analysis in various high purity
metry technology are shown. chemicals, including ultrapure water [6]. Major software
development comes with ICP-MS instruments. However,
1. Introduction this method still requires skill and specific experimental
conditions in acquiring and interpreting data. The ac-
Due to newly developed instrumentation, limits of de- tual detection limits depend on the element, the matrix,
tection reached by trace analysis are becoming lower and the sample preparation and the condition of the instru-
lower. Hyphenated techniques currently used for sample ment. Precise method protocols and experimental con-
studies and element detection can now reach ng/L lev- ditions are developed to perform some specific element
els and even pg/L levels when specific clean conditions determinations.
and careful experimental handling are ensured. Conse-
quently, instruments and reagents used for blank anal- 2.2. The water purification system
ysis, standard dilutions and sample preparations need
Blank levels of a given element are affected by such
to be of the highest quality. Depending on the ele-
factors as the purity of sample-treating solutions, con-
ments studied and the environmental conditions of the
tainer purity, and the general analytical environment.
analytical laboratory, various instrument configurations
Among the various reagents used during the key steps
are possible. Flame or electrothermal atomic absorption
of blank, standards and sample preparation, is ultrapure
spectroscopy (FAAS or ETAAS) and inductively coupled
water. Investigations are often performed on the back-
ground obtained while aspirating air, pure acid or ultra-
pure water. Ultrapure water such as that produced by
a Milli-Q◦ system, causes less spectral interference than
R
∗ Contact: Paweł Kubiński, Millipore Sp. z o.o. Bio-
science Division, Powązkowska 15, 01-797 Warszawa, Poland, high quality nitric acid. Although this acid is purified
[email protected], www.millipore.com by sub-boiling, the trace element concentrations are still

(S-203)
S-204 D. Darbouret, I. Kano

above those of ultrapure water. It is clear that 18.2 MΩ resins (Quantum IX) in order to monitor the very first
cm is no longer a “quality certification” value. Studies ionic release from the first purification pack (Q-Gard B1).
concerning ultratrace analysis show that blank optimiza- Final filtration is ensured through a 0.1 µm filter, con-
tion is only achievable when ultrapure water with sub- taining a membrane specifically designed for critical ul-
-ppt level contamination for most of the elements is used. tratrace applications. This membrane is made of ultra-
Contamination risks are greatly increased when the ul- high molecular weight polyethylene (UHMWPE filter).
trapure water is stored. Results clearly indicate that the The specific membrane charged structure removes col-
quality of high purity water degrades with storage time. loidal traces. The flow schematic of this water purifica-
In addition to clean air, the laboratory needs an ultra- tion system called Milli-Q Element is shown in Fig. 1.
pure water supply, and the analytical instrument, or at To be able to collect the ultrapure water directly under a
least the sample preparation station, should be located laminar hood, with ease and limited contamination risks,
close to the water outlet. a three meter distance can separate the main system cab-
inet and the point of use. The delivery of water is ensured
2.3. The pretreatment system via an automatic footswitch solenoid valve.
Production of high quality ultrapure water requires a
combination of technologies to process tap water to ul-
trapure water. Water is first purified through a system
including reverse osmosis and electrodeionization (EDI).
This EDI technology is a key step in the production of
ion-free water. The EDI module, where a direct current
voltage is applied to the resin-containing cells, maintains
consistent water quality with no significant fluctuations
due to changes in feed water ion concentration. The
high resistivity water produced, presents a low ionic chal-
lenge to ultrapure polishing resins. The hydrolysis and
ion removal in the EDI module result in a steady-state
operation of the resin without exhaustion or the need Fig. 1. Flow schematic of the Milli-Q Element water
for regeneration. A more complete description of this purification system for ultratrace analysis.
RO/EDI treatment, performed in a system called Elix◦,
R

has been previously published. This purified water is

then processed in a polishing system where additional 3. Experimental requirements
technologies produce the ultrapure water suited to ultra-
trace analysis. Elix water is stored in an intermediate Sample and/or laboratory contamination can affect the
reservoir to provide an adequate flow rate to feed ultra- accuracy of trace metal analysis. Special care must be
pure water systems. Numerous tests were performed to taken with the experimental environment, this includes
select the right materials of construction, to define the the quality of the reagent used. Moreover, most of the
reservoir design and finally to limit water degradation contamination can result from anything that comes into
during storage. As a result of these tests, polyethylene contact with the sample, including glassware, the labo-
with low extractable levels was selected for the container; ratory environment, the air, and anything used during
a blowmolding process was chosen to ensure smooth and sample preparation. Even cleanroom gloves can present
regular inner surfaces for the conical reservoir and a vent significant metal contamination. Precise washing proto-
filter containing activated carbon and soda lime was de- cols are thus established in order to remove any contam-
veloped. ination coming from the various containers used to pre-
pare samples and standards. High quality plastic bottles,
2.4. The ultrapure water polishing system
mainly polyethylene (PE), perfluoroalkoxy (PFA) or fluo-
The ultrapure water system uses high quality ion- rinated ethylenepropylene (FEP) are used throughout all
-exchange mixed bed resin in a pure natural polypropy- sampling and analysis procedures. Several acid and ul-
lene housing selected for its low leaching characteristics. trapure water washing steps should be performed prior to
The best ultrapure water to optimize blanks and pre- running the experiment in order to avoid further leach-
pare standards is obtained by adding UV photo-oxidation ing from the vials. The impact of contamination com-
technology within the water system. A 185/254 nm UV ing from the various vials, or sample adsorption on the
lamp placed upstream of the polishing packs ensures walls of such containers, have led scientists performing
the destruction of organics, including those with trapped ultratrace analysis to develop cleaning steps for materi-
metals. The released elements can then be retained by als. The procedure followed by scientists working in the
the ion exchange resins. A resin for the specific removal glaciology field is given hereafter as an example. “Low
of boron is included in the first purification pack. Addi- density PE (LDPE) bottles for storage of samples and
tionally, an accurate resistivity meter is placed upstream plastic tools were acid cleaned in the class 100 environ-
of the final polishing cartridge that contains mixed-bed ment. Briefly, items were cleaned as follows: rough rinse
 Ultrapure Water for Elemental Analysis Down to ppt Levels S-205

with tap water to remove dust; degrease with chloroform this limit of detection. Values are given for ultrapure wa-
and rinsing with ultrapure water; immersion in a first ter samples, even when these are under limits of quan-
acid bath (HNO3 /ultrapure water 1/3, 50 ◦C, 2 weeks), tification. BEC corresponds to the blank equivalent con-
and rinsing with ultrapure water; immersion in a sec- centration calculated using linear calibration curves ob-
ond acid bath (HNO3 /ultrapure water 1/1000, 50 ◦C, two tained with 0.50 and 100 ppt standard solutions for each
weeks) and rinsing with ultrapure water; immersion in a element (Fig. 2).
third acid bath (HNO3 /ultrapure water 1/1000, 50 ◦C, Extrapolation to the x-axis (y = 0) gives the BEC
two weeks); finally, bottles are rinsed several times with value and is a good curve obtained for Ca, which clearly
ultrapure water, filled with a diluted ultrapure HNO3 shows the limitation for this ion (due to instrumentation
fresh solution and stored inside double polyethylene acid and experimental conditions). On the other hand, the
clean bags”. Washing steps are even more important in good Fe results confirm the interference elimination with
the microelectronics industry, where efficient rinsing pro- the chosen ICP-MS conditions as indicated in Sect. 2.
tocols should be employed after wafer-cleaning steps.
5. Conclusion and perspectives
4. Results and discussion
4.1. Preliminary study Taking the environmental field as an example, over
the past 10 years, reported background dissolved trace
ICP-MS analysis with no element-specific optimization element concentrations have declined from tens of ppb
is performed on ultrapure water and counts are recorded (µg/L) through single digit ppb, to the ppt (ng/L) range.
to get an idea of the achievable blanks. The addition This in fact does not reflect improved water quality,
of 10 ppt standards of the studied elements gives infor- but rather, reduction in contamination introduced dur-
mation on the achievable quantification limits. The high ing sampling, processing, and analysis. These improved
counts generated for 40 Ca clearly show the interference instrumental and experimental procedures highlight the
due to 40 Ar and confirm that specific conditions must be impact of minute contamination. Consequently, the use
applied to the ICP-MS instrument in order to perform of very high quality ultrapure water is required to pre-
sensitive and accurate calcium determinations. pare blanks and standards, perform critical cleaning and
4.2. Elemental analysis of Milli-Q◦ Element water
R run high sensitivity analyses. Being able to modulate
the polishing cartridge composition according to the ap-
A significant number of experiments were performed plication allows blank optimization for trace analysis of
with a Milli-Q◦ Element system fed by an Elix◦ sys- a specific element such as boron. Other developments
R R

tem. In each case, the limit of detection (DL) is taken can also be geared to specific needs, such as the focus
as 3 times the standard deviation of ten replicates of a on silica 25 for example. Additional added features, such
blank (Milli-Q SP ICP-MS water, Nihon Millipore Ltd., as the ability to operate the system with a footswitch,
Japan), the limit of quantification (QL) being 3.33 times ensure minimum cross contamination by other users and
facilitate operation under a laminar hood. The combina-
tion of these different advancements in instrumental and
purification technologies has resulted in a system that
produces ultrapure water suitable or ultratrace analysis
at the sub-ppt levels.

References

[1] K.W. Jackson, C. Guoru, Anal. Chem. 68, 231R

(1996).
[2] K. Kawabata, H. Takahashi, G. Endo, Y. Inoue, Appl.
Organomet. Chem. 8, 245 (1994).
[3] J.P. Hurley, M.M. Shafer, S.E. Cowell, J.T. Overdier,
P.E. Hughes, D.E. Armstrong, Environmental Sci.
Technol. 30, 2093 (1996).
[4] K. Kawabata, Y. Kishi, O. Kawagushi, Y. Watanabe,
Y. Inoue, Anal. Chem. 63, 2137 (1991).
[5] A. Woller, H. Garraud, F. Martin, O.F.X. Donard,
P. Fodor, J. Anal. At. Spectrom. 12, 53 (1997).
Fig. 2. Selected calibration curves. [6] M. Takenaka, M. Hayashi, I. Suzuki, Y. Yamada,
K. Takamatsu, M. Kageyama, Anal. Chem. 69, 972
(1997).