Professional Documents
Culture Documents
1 s2.0 S0039914009007061 Main
1 s2.0 S0039914009007061 Main
Talanta
journal homepage: www.elsevier.com/locate/talanta
a r t i c l e i n f o a b s t r a c t
Article history: Interest in tungsten occurrence and geochemistry is increasing due to increased use of tungsten com-
Received 10 July 2009 pounds and its unknown biochemical effects. Tungsten has a complex geochemistry, existing in most
Received in revised form 9 September 2009 environmental matrices as the soluble and mobile tungstate anion, as well as poly- and heteropoly-
Accepted 10 September 2009
tungstates. Because the geochemistry of tungsten is substantially different than most trace metals,
Available online 12 September 2009
including the formation of insoluble species under acidic conditions, it is not extracted from soil matrices
using standard acid digestion procedures. Therefore, the current work describes a modication to a com-
Keywords:
monly used acid digestion procedure to facilitate quantication of tungsten in soil matrices. Traditional
Tungsten
Acid digestion
soil digestion procedures, using nitric and hydrochloric acids with hydrogen peroxide yield <1 up to 50%
Extraction procedures recovery on soil matrix spike samples, whereas the modied method reported here, which includes the
addition of phosphoric acid, yields spike recoveries in the 7698% range. Comparison of the standard and
modied digestion procedures on National Institute of Standards and Technology Standard Reference
Materials yielded signicantly improved tungsten recoveries for the phosphoric acid modied method.
The modied method also produces comparable results for other acid extractable metals as the standard
methods, and therefore can be used simultaneously for tungsten and other metals of interest.
Published by Elsevier B.V.
effective with sodium versus ammonium hydroxide, which can with a mortar and pestle, and sieved through a #40 ASTM sieve to
lead to plasma instabilities and interferences necessitating sample improve sample homogeneity similar to the NIST reference soils
dilution which raises analyte reporting limits [14,17]. [20].
The modied procedure described in the current work capi-
talizes on tungstens ability to polymerize with oxyanion acids,
particularly phosphoric acid, to yield acid soluble tungsten species 2.3. Digestion procedures
for measurement by ICP-AES or ICP-MS, while simultaneously
allowing determination of other metals extracted during the Three digestion procedures were compared for the present
digestion procedure. The modication involves the addition of work, two published versions of US EPA method 3050B [12], and a
phosphoric acid to a standard nitric acid and hydrogen peroxide third modied from the standard procedures by adding phosphoric
digestion procedure to promote tungsten solublization [1214]. acid to the acid cocktail, as described below and outlined in Table 1.
The phosphoric acid modied procedure was tested on standard All acid digestions were conducted on a DigiPrep hotblock digestion
reference materials and test eld soils. Results were compared for system (SCP Science, Champlain, NY), using uoropolymer diges-
tungsten and other metals routinely measured by ICP-AES follow- tion vessels to prevent tungsten polymerization onto borosilicate
ing acid digestion. Tungsten recoveries are greatly improved using glass vessel walls. Each digestion batch included a matrix blank,
the modied procedure, with only minor differences observed for blank spike, and NIST standard reference soils, as well as the three
some metals between the modied and standard digestion proce- test soils (GL, FS-1, FS-2), digested in triplicate.
dures.
Table 1
Procedures for the three acid digestions compared in the current work.
Sample mass and digestion vessel 0.5 g in PTFE vessels 0.5 g in PTFE vessels 0.5 g in PTFE vessels
1st acid addition 5 mL of 1:1 nitric acid 5 mL of 1:1 nitric acid 5 mL 1:1 nitric acid + 1 mL concentrated
phosphoric acid
Heating 30 min at 95 C 30 min at 95 C 30 min at 95 C
Acid additions 42.5 mL concentrated nitric acid each 42.5 mL concentrated nitric acid each 42.5 mL concentrated nitric acid each
with 30 min heating with 30 min heating with 30 min heating
Heating 120 min at 95 C 120 min at 95 C 120 min at 95 C
1st hydrogen peroxide addition 3 mL 30% hydrogen peroxide 3 mL 30% hydrogen peroxide 3 mL 30% hydrogen peroxide
Heating 15 min at 95 C 15 min at 95 C 15 min at 95 C
2nd hydrogen peroxide addition 2 mL 30% hydrogen peroxide 2 mL 30% hydrogen peroxide 2 mL 30% hydrogen peroxide
Heating 120 min at 95 C 120 min at 95 C 120 min at 95 C
Hydrochloric acid addition 5 mL concentrated hydrochloric acid
Heating 15 min at 95 C
Dilution to volume Filter and dilute to 50 mL with 1% nitric Filter and dilute to 50 mL with 1% nitric Filter and dilute to 50 mL with 1% nitric
acid acid acid
A.J. Bednar et al. / Talanta 80 (2010) 12571263 1259
2.3.3. Phosphoric and nitric acids digestion were calibrated using mixed analyte standards containing 0, 0.100,
The third digestion procedure was developed specically to 1.000, 10.00, and 100.0 mg/L of each analyte. All other analytes,
increase tungsten extraction. A third 0.5 g aliquot of each of the including tungsten, were calibrated using mixed analyte stan-
solid samples was weighed into uoropolymer vessels and digested dards containing 0, 0.100, 1.000, and 10.00 mg/L of each element.
following the above described nitric acid procedure, with the mod- The ICP-MS was calibrated using standards containing 0, 0.001,
ication of adding 1 mL of concentrated phosphoric acid with the 0.010, and 0.100 mg/L tungsten, and used holmium as the internal
initial 5 mL of 1:1 nitric acid:deionized water. The digestate was standard.
then ltered and diluted as described above. All calibration responses were linear and had correlation coef-
Aliquots of the 50 mL digestate solutions were further diluted cients of at least 0.9999. A second source calibration verication
with 1% nitric acid prior to analysis as needed such that analyte standard was analyzed immediately after calibration following
concentrations were within the calibration range of the analytical standard US EPA Method 6010C procedures [21]; analyte recoveries
instrumentation (generally 1:4 dilution). were within 10% of the nominal value (4.000 and 0.040 mg/L for
ICP-AES and ICP-MS, respectively). Continuing calibration verica-
2.4. Instrumentation tion standards and blanks were analyzed periodically throughout
the analytical batch, with recoveries within 10% of the nominal
Metal concentrations in the diluted digestates, including tung- value (5.000 and 0.050 mg/L for ICP-AES and ICP-MS, respectively)
sten, were determined by Inductively Coupled Plasma Atomic and subsequent blanks below the instrument reporting limits
Emission Spectroscopy (ICPAES, following modications of EPA (<0.020 and <0.001 mg/L for ICP-AES and ICP-MS, respectively).
Method 6010C [21]), using a PerkinElmer (Wellesley, MA) Optima
5300DV equipped with a MiraMist nebulizer and cyclonic spray 3. Results and discussion
chamber. The plasma was operated at 1400 W and nebulizer
gas ow rate of 0.65 mL/min. Tungsten was quantied using the 3.1. Tungsten
207.912 nm emission line with analyte conrmation at 224.876 nm
using axial plasma viewing. Due to the low concentration of The tungsten concentrations measured in the two standard
tungsten in the standard reference soil 2709, tungsten was also reference materials, using the three digestion procedures, and
measured by Inductively Coupled Plasma Mass Spectrometry (ICP- associated quality control samples are listed in Table 2. The diges-
MS following modications of EPA Method 6020A [21]), using a tion blanks for all three procedures were below the instrument
PerkinElmer Elan DRC-II equipped with a MiraMist nebulizer and reporting limit (<0.040 mg/L including dilution factor), indicat-
Scott spray chamber. The ICP-MS plasma was operated at 1250 W ing that contamination was not present in the acids or vessels
and nebulizer gas ow of 0.85 mL/min. Tungsten was quantied at used. However, the blank spike sample, consisting of aque-
m/z 182 with conrmation using m/z 184 after correction for 184 Os. ous 10 mg/L tungsten from a NIST-traceable standard shows a
marked difference among the three digestion procedures. The
2.5. Calibration and quality control nitric acid method yielded a blank spike recovery of 39%, while
the nitrichydrochloric acid procedure recovery was 43%, both
The ICP-AES was calibrated using NIST-traceable calibration were well below the phosphoricnitric procedure recovery of 98%.
standards. Scandium, added on-line using a mixing-T, was used as The nitric and nitrichydrochloric procedures produced a yel-
an internal standard to correct for instrumental drift during analy- low precipitate consisting of insoluble tungstic and polytungstic
sis. Calcium, magnesium, sodium, potassium, aluminum, and iron acids that was captured by the ltration process resulting in poor
Table 2
Comparison of the modied digestion procedure to two standard methods for extraction of tungsten from two standard reference materials and three test soils. Concentrations
are in mg/kg standard deviation of triplicate digestions for soils and mg/L for liquid quality control samples determined by ICP-AES, values in parentheses were determined
by ICP-MS.
Tungsten concentration (mg/kg) Tungsten concentration (mg/L) Reported value recovery (%) Matrix spike recovery (%)
spike recovery [6,22]; the phosphoricnitric digestion produced no respectively, further suggesting that the two standard methods
precipitate, and therefore the soluble phosphotungstates formed have a low bias for tungsten extraction.
during the modied phosphoricnitric digestion procedure were The lower concentrations determined by the standard methods
not removed by ltration. Improved recoveries for the nitric and explains the poor matrix spike recovery values listed in Table 2.
nitrichydrochloric acid procedures were not obtained by rinsing The nitric acid digestion yielded a spike recovery of 39% for the GL
the ltrate with 2% phosphoric acid, indicating that the precipitates soil, but less than 1% for FS-1, and 47% for the FS-2 sample. The
are stable once formed. nitrichydrochloric acid procedure yielded slightly higher tung-
The NIST standard reference soil material results listed in Table 2 sten recoveries than the nitric acid procedure, with the GL spike
show a similar trend as the blank spike samples. The reference recovery being 56%, but only 4 and 25% for the FS-1 and FS-2
soil 2709 has a total tungsten concentration determined by Neu- samples, respectively. Finally, the matrix spike recoveries for the
tron Activation Analysis reported by NIST of 2 mg/kg, which is phosphoricnitric procedure were 98, 76, and 79% for the GL, FS-1,
below the ICP-AES reporting limit for the current study, although, and FS-2 samples, respectively, substantially higher than the stan-
this demonstrates that matrix interferences are not present, which dard methods. FS-1 contains a high amount of iron, and therefore
could yield a false positive. However, ICP-MS analysis indicated it is possible that an irontungstate phase is causing exceptionally
that increased tungsten recoveries in reference soil 2709 were low extraction of tungsten, even lowering the phosphoricnitric
obtained using the phosphoricnitric digestion procedure com- digestion, which was 76%. The FS-2 sample is a sandy soil, which
pared to the standard techniques. The phosphoricnitric procedure could facilitate the formation of polytungstates on the silicon diox-
yielded 0.5 mg/kg tungsten, which is approximately 25% of the ide particles, resulting in low tungsten recovery, although the
reported value, however, this is a higher recovery than either the phosphoricnitric procedure yielded 79% recovery. These experi-
nitric or nitrichydrochloric digestions. ments demonstrate that the phosphoricnitric modied procedure
The reference soil 2710 has a total tungsten concentration greatly improves tungsten extraction from soils compared to the
determined by Neutron Activation Analysis reported by NIST unmodied standard procedures.
of 93 mg/kg. The nitric acid digestion procedure yielded a con-
centration of 27.4 mg/kg, whereas the nitrichydrochloric acid 3.2. Other metals
digestion yielded 50 mg/kg, which represent 29 and 54% recov-
ery, respectively. The phosphoricnitric acid procedure yielded a In addition to tungsten, other metal analyses in soils may also
concentration of approximately 80 mg/kg, or 86% recovery. These be required simultaneously, therefore, the digestate solutions were
recoveries are in line with values reported by others when standard analyzed for 20 other metals to determine the comparability of
acid digestion procedures are used for tungsten determinations the three procedures for determination of other metals. The addi-
[13,17]. tion of phosphoric acid could potentially lower recoveries of metals
It should be noted that all three acid digestion procedures tested that form insoluble phosphate salts, such as lead, aluminum, iron,
are not intended to be total dissolutions, but rather are strong nickel, and cobalt, therefore, extraction efciencies and matrix
extractions, therefore, 86% tungsten recovery for the reference soil spike recoveries were tested and compared.
2710 by the phosphoricnitric method compared to Neutron Acti- The digestion blanks listed in Table 3 show only 2 analyte detec-
vation Analysis is considered acceptable. Furthermore, some of the tions, 0.081 mg/L iron in the phosphoricnitric digestion solution,
tungsten may be present in these samples in silicate phases, which and 0.104 mg/L calcium in the nitrichydrochloric acid digestion
would normally only be extracted by a total dissolution technique solution. However, these major cations are present at several orders
(e.g. hydrouoric acid, [15,16]), or a solid phase analysis technique, of magnitude higher concentrations in soil matrices, suggesting
such as Neutron Activation or X-ray Fluorescence Spectroscopy. these blank levels are acceptable. The blank spike recoveries listed
This likely explains the low recovery observed in the reference in Table 3 for all elements measured from the three digestion
soil 2709, even with the phosphoricnitric digestion procedure. procedures are within 10% of the nominal spike concentration
However, in light of the two standard acid digestion procedures indicating no bias among the three digestion procedures for the
extracting less tungsten in all matrices, the phosphoricnitric acid elements determined.
procedure performs well. Additionally, the tungsten concentra- Table 3 lists other metal concentrations for the two NIST
tions reported by NIST are not certied values, but presented for standard reference soils determined using the three digestion pro-
information purposes only, and therefore may have some inherent cedures. The acid leachable metals concentrations reported by NIST
error associated with them. A total dissolution procedure using a are not certied values (only the total metals concentrations are
four acid cocktail, including hydrouoric acid [16], was performed certied by NIST), but are listed for information purposes only.
on the reference soils 2709 and 2710 to verify the NIST reported However, the recoveries listed generally agree among the three
concentrations. Reference soil 2709 yielded a tungsten concentra- procedures, indicating that low recoveries are not an issue for these
tion of 2.2 0.05 mg/kg determined by ICP-MS, while reference soil soil matrices and that the phosphoricnitric procedure can be used
2710 yielded a tungsten concentration of 97.8 0.4 mg/kg deter- for simultaneous determination of other metals. For example, Co
mined by ICP-AES. These values are 110 and 105% of the NIST determined by the nitric acid procedure yielded 90 and 101% recov-
reported concentrations, respectively. ery for the standard reference soils 2709 and 2710, respectively,
The three digestion procedures are also compared in Table 2 on in agreement with recoveries obtained for the nitrichydrochloric
the three test soils, with similar trends in tungsten extraction and procedures (93 and 100%, respectively). In comparison, the
matrix spike recovery observed. The nitric acid procedure yielded phosphoricnitric procedure produced cobalt recoveries of 105
tungsten concentrations of 210 30, 90.2 11, and 440 28 mg/kg and 102% for the reference materials 2709 and 2710, respectively,
for the GL, FS-1, and FS-2 samples, respectively. Tungsten con- indicating similar extraction efciencies for the three digestion pro-
centrations in the nitrichydrochloric digestion procedure samples cedures. Similar agreement is observed for other analytes listed in
were approximately the same for the GL soil (220 25 mg/kg) as the Table 3 (e.g. arsenic, barium, calcium, cadmium, lead, strontium,
nitric acid procedure, although the concentrations measured for the and zinc). Elements such as chromium, copper, iron, potassium,
FS-1 and FS-2 samples were about a factor of two higher than those magnesium, molybdenum, and nickel appear to have a slightly
measured with the nitric acid procedure. The phosphoricnitric higher extraction efciency by the phosphoricnitric procedure,
procedure yielded the highest tungsten concentrations, 762 44, compared to the standard digestion procedures, although the dif-
483 191, and 1695 203 mg/kg for the GL, FS-1, and FS-2 soils, ference is generally less than a 10% increase.
Table 3
Comparison of the modied digestion procedure to standard methods for extraction of other metals from two standard reference materials.
Digestion Digestion 2709 2710 Digestion Digestion 2709 2710 Digestion Digestion 2709 2710
blank blank spikeb (% recovery) (% recovery) blank blank spikeb (% recovery) (% recovery) blank blank spikeb (% recovery) (% recovery)
(mg/L) (% recovery) (mg/L) (% recovery) (mg/L) (% Recovery)
NAconcentration measured by one of the digestion procedures is below instrument reporting limit. NSmatrix spike solution did not contain this analyte.
a
Acid leachable metal concentrations are reported by NIST for information purposes only, they are not certied values.
b
Spike levels (mg/L): Al = 10, As = 0.5, Ba = 2.5, Cd = 0.25, Co = 1, Cr = 1, Cu = 1, Fe = 5, Mn = 1, Ni = 1, Pb = 0.5, V = 1, Zn = 2.5.
1261
1262 A.J. Bednar et al. / Talanta 80 (2010) 12571263
128
139
124
113
172
115
102
117
113
147
125
107
107
NA
NA
NA
NA
NA
89
94
a refractory vanadium and titanium phase that is sensitive to the
presence of phosphoric acid.
% of nitric
digestion
128
134
177
120
115
101
113
121
111
168
145
130
the three test soils determined using the three digestion proce-
NA
NA
NA
NA
NA
NA
85
83
dures. Because there are no reference values for the concentrations
Field soil 2 metals Matrix spike
recovery (%)
-207
nitrichydrochloric acid procedures are used as the reference
108
-67
NS
NS
NS
NS
NS
NS
NS
83
91
91
93
88
91
90
91
40
90
values. The concentrations determined by the phosphoricnitric
(mg/kg) standard
The matrix spike recoveries for the Grenada Loring soil are all
6481 222
4742 42
9438 69
996 339
deviation
583 6.9
689 2.7
197 5.9
within 20% of the nominal value except for aluminum, iron, and
337 75
43 5.5
13 0.2
47 0.6
10 0.2
11 0.2
11 0.8
40 0.1
<10
<10
<10
<10
than 10% of the sample concentration for aluminum and iron, and
less than 20% of the manganese concentration in the samples. Sim-
ilarly, the matrix spike recoveries for most elements in the FS1 and
2 samples are acceptable (20%) except where the spike concen-
Metal concentrations and matrix spike recoveries in the three test soils obtained using the modied phosphoricnitric acids digestion procedure.
digestion
227
146
261
189
114
107
105
123
111
282
105
106
NA
NA
NA
NA
NA
99
95
95
digestion
231
144
111
125
267
105
146
111
385
131
111
190
105
NA
NA
NA
NA
NA
95
96
268
294
615
105
NS
NS
NS
NS
NS
NS
NS
81
91
91
92
94
92
94
procedures is excellent for copper in the eld soils 1 and 2 that have
NAconcentration measured by one of the digestion procedures is below instrument reporting limit.
nitrichydrochloric concentration
10852 1620
8023 1250
199 13.1
135 29.7
114 1.1
56 10.0
415 92
305 50
13 4.4
23 1.5
91 3.1
45 1.1
26 4.3
21 1.4
24 1.0
<10
<10
182
143
163
145
117
263
154
162
114
105
131
143
109
408
NA
NA
NA
NA
NA
NA
digestion
173
151
158
261
149
160
101
123
135
446
145
102
110
NA
NA
NA
NA
75
47
4. Conclusions
233
NS
NS
NS
NS
NS
NS
86
94
92
91
96
94
18
91
83
92
95
50
(mg/kg) standard
15446 705
388 4.9
852 11
628 65
12 1.1
94 5.9
12 5.2
81 1.3
19 0.6
14 0.2
35 1.0
29 0.5
10 0.2
<10
<10
Na
Cu
Cd
Zn
Co
Pb
Ba
Ca
As
Ni
Fe
Cr
Al
Sr
Ti
V
K
procedures and can be used for tungsten and other metals simul- [7] A.J. Bednar, R.A. Kirgan, D.R. Johnson, A.L. Russell, C.A. Hayes, C.J. McGrath, Land
taneously. Contam. Reclam. 17 (2009) 129.
[8] J.P. Gustafsson, Chem. Geol. 200 (2003) 105.
[9] M.J. Vissenberg, L.J.M. Joosten, M.M.E.H. Heffels, A.J. van Welsenes, V.H.J. de
Acknowledgments Beer, R.A. van Santen, J.A.R. van Veen, J. Phys. Chem. B 104 (2000) 8456.
[10] A.J. Bednar, W.T. Jones, R.E. Boyd, D.B. Ringelberg, S.L. Larson, J. Environ. Qual.
37 (2008) 229.
The use of trade, product, or rm names in this report is for [11] A.J. Bednar, R.E. Boyd, W.T. Jones, C.J. McGrath, D.R. Johnson, M.A. Chappell, D.B.
descriptive purposes only and does not imply endorsement by Ringelberg, Chemosphere 75 (2009) 1049.
the U.S. Government. The tests described and the resulting data [12] U.S. Environmental Protection Agency, SW-846 Test Methods for Evaluating
Solid Waste, Method 3050B, U.S. Environmental Protection Agency, 1996.
presented herein, unless otherwise noted, were obtained from [13] C. Griggs, S. Larson, C. Nestler, M. Thompson, Land Contam. Reclam. 17 (2009)
research conducted under the Environmental Quality Technology 121.
Program of the United States Army Corps of Engineers by the [14] US Occupational Safety Health Administration, Tungsten and Cobalt in work-
place atmospheres (ICP-AES Analysis), Method ID-213, US Occupational Safety
USAERDC. Permission was granted by the Chief of Engineers to and Health Administration, 2004.
publish this information. The ndings of this report are not to be [15] US Environmental Protection Agency, SW-846 SW-846 Test Methods for Eval-
construed as an ofcial Department of the Army position unless uating Solid Waste, Method 3052B, U.S. Environmental Protection Agency,
1996.
so designated by other authorized documents. The authors also
[16] P.H. Briggs and A.L. Meier, The determination of forty-two elements in geologi-
thank Jennifer Seiter and Frances Hill of the USACE for their editorial cal materials by inductively coupled plasma-mass spectrometry: in Taggart,
comments. J.E., ed., Analytical methods for chemical analysis of geologic and other
materials, U.S. Geological Survey Open-File Report 02-223, 2002, Chapter I,
14 p.
References [17] C.S. Griggs, Evaluation of alkaline digestion for the determination of total tung-
sten in soil by ICP-OES, Poster Presentation, DoD Environmental Monitoring
[1] R.L. Seiler, K.G. Stollenwerk, J.R. Garbarino, Appl. Geochem. 20 (2005) 423. and Data Quality Workshop, 30 March1 April, San Antonio, TX, 2009.
[2] P.R. Sheppard, G. Ridenour, R.J. Speakman, M.L. Whitten, Appl. Geochem. 21 [18] L.S. Inouye, R.P. Jones, A.J. Bednar, Environ. Toxicol. Chem. 25 (2006) 763.
(2006) 152. [19] J. Clausen, S. Taylor, S. Larson, A. Bednar, M. Ketterer, C. Griggs, D. Lambert, A.
[3] A. Koutsospyros, W. Braida, C. Christodoulatos, D. Dermatas, N. Strigul, J. Haz- Hewitt, C. Ramsey, S. Bigl, R. Bailey, N. Perron, ERDC-CRREL TR-07-5, US Army
ard. Mater. 136 (2006) 1. Engineer Research and Development Center, 2007.
[4] N. Strigul, A. Koutsospyros, P. Arienti, C. Christodoulatos, D. Dermatas, W. [20] D.R. Felt, A.J. Bednar, T. Georgian, Talanta 77 (2008) 380.
Braida, Chemosphere 61 (2005) 248. [21] US Environmental Protection Agency, SW-846 SW-846 Test Methods for Eval-
[5] F. Feigl, Spot Tests in Inorganic Chemistry, Elsevier Publishing Co., Amsterdam, uating Solid Waste. Method 6010C and Method 6020A, U.S. Environmental
1958, 600p. Protection Agency, 2007.
[6] S.R. Kraynov, The effect of acidity-alkalinity of ground waters on the concentra- [22] M.V. Regnault, Elements of Chemistry for the use of Colleges, Academies, and
tion and migration of rare elements, Geochemistry International (1971) 828. Schools, vol. 2, 2nd ed., Clarke and Hesser, Philadelphia, 1853, pp. P230232.