Table IV. Comparisons of V/AI Mass Ratios for Puerto Rican and Hawaiian Aerosols, California Dust and Average Crustal Material Puerto Rican Av
Station V/AI NC 0.042 BS
RF CF
0.011 0.089 0.021
Av Hawaiian, California dust, V/AI V/AI
(Hoffman etal., 1972)
0.024
0.0025
Crustal material, V/AI
Av (Goldschmidt, 1954), 0.0018
Range (Turekian and Wedepohl, 1961), 0.0003-0.005
Table V. Multielement Ratios in Puerto Rican Aerosols, NCA II and RFA I I Sample NCA I1 RFA II
Crustal av Seawater
V/AI
0.071 0.12 0.0018 2.0
Co/AI 0.013 0.0086 0.00049
Cr/AI 0.063 0.068 0.0025
0 10
0.050
Co/Cr 0.21 0.13 0.20 2.0
V/Co 5.5 14 3.8 20
V/Cr 1.1 1.8 0.75 40
processes in Puerto Rico (Martens and Harriss, 1973). A1 and V particle size distributions out of clouds change only slightly during cloud formation indicating little incorporation of condensing water vapor. This physical evidence suggests that both elements are part of nonhygroscopic particles, probably from a soil origin. It is possible that soil dust from both local and distant sources such as the African continent (Prospero, 1972) is being sampled.
Summary and Conclusions Both soil and combustion appear to be major contributors of V in the San Francisco Bay area atmosphere, the soil providing the larger particles and the combustion the smaller particles. V sampled in Puerto Rican experiments may be primarily from a soil source. The similar behavior of V and A1 in Puerto Rican orographic clouds (Martens and Harriss, 1973) further suggests soil origin for the V, although the soil dust does not appear to have average
crustal composition, especially in the case of A1 which is depleted.
Acknowledgment The authors thank Richard Clements and the staff of the Puerto Rico Nuclear Center for the use of their facilities and for their assistance during the collection of Puerto Rican samples. We thank Milton Feldstein and Wayman Siu of the Bay Area Air Pollution Control District for the use of their stations for our sampling sites. This paper has been improved through the comments of John W. Winchester, Florida State University, and Robert A. Duce, University of Rhode Island. Literature Cited Brar, S . S., Nelson, D. M., Kline, J. R., Gustafson, P. F., Kanabrocki, E. L., Moore, C. E., Hattori, D. M., J. Geophys. Res., 75, 2939-45 (1970). Dams, R., Rahn, K . A , , Winchester, J . W., Enuiron. Sei. Technol., 6, 441-8 (1972). Dams, R., Robbins, J. A., Rahn, K . A,, Winchester, J . W., Anal. Chem., 42,861-7 (1970). Flesch, J . P., Norris, C. H., Nugent, A. E., Jr., J. Amer. Indus. Hyg. ASS.,28,507-16 (1967). Goldschmidt, V. M., “Geochemistry,” Oxford Univ. Press, London (1954). Harrison. P. R.. Rahn. K . A , . Dams. R.. Robbins. J. A , . Winchester. J. h’.,Brar, S . , Selson, D. ‘M.,’J.Air Po’lluf. dontr. Ass,, 21,563-70 (1971). Hoffman, G. L., Duce, R. A., Hoffman, E. J., J. Geophys. Res., 77,5322-9 (1972). Hoffman, G. L., Duce, R. A , , Zoller, W. H., Enuiron. Sei. Technol., 3, 1207-10 (1969). John, W., Kaifer, R., Rahn, K. A , , Wesolowski, J. J., Atmos. Enuiron., 7,107-18 (1973). Martens, C . S., PhD Thesis, Florida State University, 1972. Martens, C. S., Harriss, R. C., J. Geophys. Res., 78, 949-57 (1973). Nifong, G. D., PhD Thesis, Cniversity of Michigan, 1970. Prospero, J. M., J. Geophys. Res., 77,5255-65 (1972). Rahn, K. A , , Wesolowski, J. J., John, W., Ralston, H. R., J . Air Pollut. Contr. Ass., 21,406-9 (1971). Turekian, K. K.. Wedepohl, K . H., Geol. Soc. Amer. Bull., 72, 175-92 (1961). Zoller, W.H.. Gordon, G. E.. Anal. Chem., 42, 257-65 (1970).
s.
Received for reuieu: December 18, 1972. Accepted May 25, 1973. Work performed in part under the auspices of the L!S. Atomic Energy Commission.
Ligand Photooxidation in Copper(l1) Complexes of NitrilotriaceticAcid Implications for Natural Waters Cooper H. Langford,’ Michael Wingham, and Vedula S. Sastri D e p a r t m e n t of C h e m i s t r y , Carleton University, O t t a w a , Ont., C a n a d a K1S 5B6
The fate of aminopolycarboxylate compounds [all of which are potential detergent builders, components of some industrial wastes, and apparently products of secondary sewage treatment (Bender et al., 1971)] involves the reactions of their metal complexes because complex formation constants are so large (Ringbom, 1963). We are engaged in a study of the photochemical degradations that may occur under sunlight irradiation. Transformation of the organic molecules may be expected to be facile only To whom correspondence should be addressed. 820
Environmental Science & Technology
when excited states of the ligand to metal charge transfer type (Balzani and Carassiti. 1970a) are accessible in the complex. This suggests that attention be given to complexes of metals from the transition group such as Cu(II), Co(II), and Fe(II1). At first sight, such a restriction might discourage interest because of the low concentration of such metals, but there are two mitigating factors; first, the complex formation constants of aminopolycarboxylates with these metals are large, and second, the rapidity of ligand substitution processes in these systems (Langford and Sastri, 1972) ensures considerable “turnover.”
~
~~
~
Copper complexes of nitrilotriacetate have been irradiated a t 350 nm in a photoreactor based on a Rayonet unit. Despite the low absorptivity of the complexes a t this wavelength, a significant photodecomposition of the ligand occurs. The identified products are CH20 and (iminodiacetic acid). Inferentially, COz must be released. The quantum yield analog function, Y , based on ferrioxalate actinometry, decreases with increasing p H linearly over the pH range 2-12. The efficiency of photodecom-
position decreases with increasing concentration above 10-3M but approaches a limiting value below about 5 x lO-4M. The low concentration limit of the efficiency of decomposition (Y/concentration) is approximately 1.0. There is no direct evidence for the intervention of Cu(1) species although a photoredox pathway involving Cu seems the most plausible mechanism. The oxidation reaction is suggested to be general for Fe and Cu complexes of aminopolycarboxylates and a-amino acids.
Photodecarboxylation of the ethylenediaminetetraacetate (EDTA) complex of Fe(II1) has been known for some time (Jones and Long, 1952), as has photodeamination of a-aminoacid complexes of Fe(II1) (Balzani and Carassiti, 1970b), although neither has been studied systematically. We have recently reported facile photodecomposition of the nitrilotriacetate (NTA) complexes of Fe(II1) according to the following equation (Trott et al., 1972):
Fe(Cz04)33- a t a concentration absorbing all light incident of wavelength less than 450 nm (Hatchard and Parker, 1956). Actinometry involves determination of photolytically produced Fez+ using o-phenthroline. The actinometer value characterizes the quantity of light entering the solution for our standard geometry. Since our solutions are weakly absorbing, we do not measure conventional quantum yields but a quantum yield-related function Y = moles reaction/einsteins light entering solution. This will vary with concentration but the efficiency of photolysis, Y/concentration, will approach a limiting value a t low concentration where illumination is nearly homogeneous. The quantity of photodecomposition was estimated by determination of formaldehyde produced using the spectrophotometric procedure of Lappin and Clarke (1951). Solutions above 0.004M required Cu(I1) removal on a cation exchange resin prior to analysis. Addition of excess CHzO in advance in some runs established that this product was not decomposing under photolysis conditions. IDA was sought as follows: An irradiated solution was evaporated to 25% of its original volume and treated with sodium diethyldithiocarbamate to precipitate Cu(I1). The dissolved Cu-dithiocarbamate complex was extracted with CHC13. The aqueous layer was then evaporated to dryness and the products were esterified with diazomethane prepared freshly from diazold. The esters were finally dissolved in a little CHC13 and thin-layer chromatography (tlc) was carried out on silica gel plates using ether as solvent. The plates were developed with KBr03 and KI. Reference chromatograms of NTA, IDA, and the dithiocarbamate were prepared.
ZF~II'(PJTA)
-
+
+
+
+
Fe" Fe"(NTA) IDA CH,O COL (1) where IDA designates iminodiacetic acid. Since NTA is a promising detergent builder, it seems a high priority choice for investigation of aminopolycarboxylate pathways. In the earlier work on FeIII(NTA) (Trott e t al., 1972), a connection between laboratory experiments and natural systems was attempted, guided by an equilibrium model simulation developed by Childs (1971) of the present composition of Lake Ontario augmented by assumed NTA levels from 1 x lO-7M to 2 x l0-4M in the p H range 6-9. Although Fe(II1) species bound up to 34% of the total NTA (at NTA = 3 x 10-6M, p H = 6), a t all NTA concentrations below the assumed total Cull concentration of 2 x 10-6M (and a t all p H values), the dominant NTA species was the copper complex, accounting for nearly 80% of dissolved NTA a t 10-6M or less. This suggests that despite weaker absorption of actinic light by Cu(I1) complexes and the greater barrier to the reduction Cu(II)/ Cu(I), the fate of copper aminopolycarboxylates may be more significant than the fate of iron complexes in many natural waters. In initial experiments, we prepared solutions of Cu(NTA)- by combining equimolar (0.001M) solutions of C u ( N 0 3 ) ~and sodium nitrilotriacetate prepared in both distilled water and water drawn from the Rideau River in Ottawa. These were adjusted to p H values from 4-8 and irradiated in closed borosilicate flasks containing air in open summer sunlight. Within three days, all irradiated samples gave strong positive indication of the presence of aldehydes with Schiff's reagent, whereas dark controls did not. Consequently, a systematic investigation was initiated.
Experimental Reagent grade from Eastman and BDH Chemicals from Fisher Scientific, Montreal, P.Q., were used in all experiments without additional purification. For each photolysis run, 50 ml of Cu-NTA complex was prepared from stock solutions of C U ~ +and H3NTA and adjusted to the desired pH with either HCl or NaOH. The solution (10 ml) was placed in a 5-cm cylindrical spectrophotometer cell, 2 cm in diameter, and irradiated. The remainder was held as a dark control. Irradiation was carried out with 350 nm of light in a Rayonet photoreactor for 100 min. Light intensity was determined using a solution of the well-characterized chemical actimometer
Results and Discussion The results of tlc indicate the presence of esters of NTA, IDA, and the dithiocarbamate. In one chromatogram there is a trailing behind the NTA spot and before the IDA spot, but it seems unlikely that any fourth substance is chromatographically detectable. Based on the detection of CHzO and the earlier results on the iron system, we postulate the Cu(I1) complex photoinitiated reaction as:
NTA
+ %O,
-
C02
+ CHzO + IDA
(2)
We suspect this is mediated by the photoreduction of Cu(I1) to Cu(1) followed by: 2Cu(I)
+ %02+ H,O
+
2Cu(II)
+ 20H-
(3)
but obtained no direct evidence for Cu(1). We did not seek Cu(1) vigorously because Morimoto and DeGraff's (1972) results on malonate decomposition suggest it is hard to find (if it is preseht a t all). In the experimental section, Y, the parameter used by Trott et al. (1972), was defined for use as a measure of the yield of the decomposition. It was suggested that "efficiency" of decomposition be expressed as Y/concentraVolume 7, Number 9, September 1973
821
b
\
clear
t
I I
4
6
8
*
0
PH
Figure 1. Y/concentration plotted against complex concentration on logarithmic scale Experiments at 25”C, pH = 4 15 Y IS quantum yield-related function defined in text
I 2
8
6
2
PH
Figure 2. Photodecomposition yield a s function of pH at two concentrations of complex Note greater pH dependence at higher concentration. Upper curve is 0.01 M Cu: lower curve is 0.004M Cu
tic acid (EDTA) system. In the iron case, the decreased yields were attributed to formation of oxo-bridged dimeric species; a process favored by moderately high p H and high concentration. A feature of the dimers is appearance of new absorption bands in the near uv spectrum. Suggestive evidence that the copper system has behavior similar to the iron system is the variation in absorbance a t 350 nm as a function of p H shown in Figure 3. It seems that the following conclusions can be drawn. Photodecomposition of aminocarboxylate systems has been observed for complexes of Cu(II), Fe(III), and Co(II1) (Endicott, 1972). In each case, the net reaction in a system open to air is, in effect, oxidation of the ligand by 0 2 according to the process:
RHN - CH, - COOH
+ ‘/202
+
RNHL
+ C02 + CH,O ( 4 )
All evidence points to this reaction originating in excitation of a ligand to metal “charge transfer” (CT) excitation of the metal complex of the aminocarboxylate. The range of aminocarboxylates examined includes glycine, NTA, and EDTA. We feel this is adequate to suggest generality for the reaction. It is noteworthy that the requirement for C T absorption implies t h a t complexes of Ca2+ and Mg2f will probably not react in sunlight. We have sought a reaction of NTA in the presence of Mg2f irradiated a t 350 nm without success. However, the high complex formation constants of transition metal aminopolycarboxylates coupled with the relatively rapid “turnover” in these complexes should assure the occurrence of the photooxidations in all waters where light intensities are reasonable ( a zone similar to that in which photosynthesis occurs). A recent paper (Bada, 1971) has attempted to estimate the kinetics of amino acid decomposition in natural waters with a view to understanding amino acid lifetimes in natural waters. The paper concludes that amino acids should be remarkably stable and t h a t the leading nonbiological reaction is the slow metal ion-catalyzed thermal oxidation in aerated waters. It is certain that metal induced photodecomposition would be faster to depths up to 10 meters (Duntley, 1963) and the role of this reaction in amino acid cycling deserves exploration.
Acknouledgment We thank John Carey and B. F. Scott for advice and helpful discussions. Literature Cited 0-1
10-2
10-3
COMPLEX
10-1
COhC
Figure 3. Absorbance at 350 nm a s function of pH Cu concentration I S 0.004M
tion. A plot of Y/concentration vs. concentration of the complex for pH = 4.15 is shown in Figure 1. At a concentration of 5 x lO-3M the value is unity and it remains near unity to the lower concentrations of environmental significance. This implies t h a t under given conditions of illumination and pH, the fraction of the complex photolyzed per unit time is constant. We can be no more specific than to refer to copper complexes (plural) because the exact nature of hydrolytic equilibria is not known. Figure 2 shows that the photodecomposition yield decreases with increasing pH. The effect is more pronounced a t the higher total complex concentration. The behavior is similar to that observed for the iron system (Trott et al., 1972) which we will report elsewhere for the chemistry of iron-ethylenediaminetetraace-
822
Environmental Science .8 Technology
Bada, J. L., Adcan. Chem. Ser.. 106,309 (1971). Balzani. V.,Carassiti, V., “Photochemistry of Coordination Compounds,” Chap. 11, Academic Press, New York. N.Y.. 1970a. Balzani. V.. Carassiti, V.. ibid., 192 (1970b). Bender, M . E., Matson, W . R., Jordan. R. A , . Enciron. Sci. Technol., 4,520 (1971). Childs, C . W.,14th Conf.. Intern. Ass. Great Lakes Res., Toronto, Canada, April 19-21, 1971, Duntleg. S. Q . , J Opt. Soc. Amer.. 53,214 (1963). Endicott, J. F.. private communication (1972). Hatchard. C. G.. Parker, A . C.. Proc. Ro?. Soc. (London). 235, 518 (1956). Jones, S. S., Long, F. A , , J , Ph?s. C h e m . , 56,25 (1952). Langford, C. H., Sastri, V. S., .VTP Intern. Rec. Sci., Inorg. Chem., Ser. One, 9,203 (1972). Lappin, G. R., Clarke. L. C . , A n a l . Chem.. 23,541 (1951). Morimoto, J. Ti., DeGraff. B. A , . J. Ph)s. C h e m . , 76, 1387 (1972). Ringbom, A. J.. “Complexation in Analytical Chemistry.” p 351, Interscience, ?Jew York, N.Y., 1963. Trott, T.. Henwood, R. LV.,Langford, C. H., Enciron. Sci. Technol., 6, 367 (1972). Receiced for rei.ieu December 18. 1972 Accepted Ma), 31. 1973. Work supported b\ the National Adcisor) Council for Water Resources Research.
