Metal Ions, Element Speciation Forms Retained on Wet Chitin: Quantitative Aspects of Adsorption and Implications for Biomonitoring and Environmental Technology

Abstract

Analyses of mosses and lichens provide some information on the contents of both particulate and dissolved (from hydrometeors including snow and flooding) metal ions and other elements like As and Sb in the local environment. However, this information is compromised by rarity (and thus duly legal protection) of suitable species (particularly lichens) for regular sampling and also by poorly understood mechanisms of binding. Hence, it is crucial to find an alternative that does not harm or kill rare and/or protected organisms for sampling purposes while providing data that can be traced to environmental levels (e.g., metal ions in water) in a comprehensible way. Studying the coordination of aq. metal ions on some biogenic surface which can form ligating bonds to these ions provides such information. The most abundant and thus cheap such biopolymer acting as both a possible ligand and a water- (or environment-)biomass interface is chitin. Data from chitin exposed in either water, common sandy sediments, and ferric gels delivered by Fe-oxidizing bacteria are processed to understand adsorption in quantitative terms depending on local conditions, accounting for observed BCFs >> 1 for certain elements (Bi, V, LREEs). Slopes of functions that describe the increase of retention of some element upon increasing aq. concentrations allow us to construct (a) some function giving BCF by numerical integration, (b) predict the behavior of other elements for which certain parameters guiding complex formation are known as well. It turns out that top sensitivities (maximum BCF- or partition factor) values are reached with different elements depending on the environment the chitin sample was exposed to. PF can extend the detection and determination of many elements below levels directly observable in water or sediments. The detection of fallout radionuclides on chitin is even more sensitive (by a factor of 20–25) because of omitting dilution in workup by direct observation of γ radiation.

1. Introduction

Previous work on chitin application dealt with either metal ion removal from wastewater [1,2], including anthropogenic radionuclides [3,4], or strived to outline an alternative to older methods of biomonitoring [5,6]. While the latter gave strong evidence for the response of chitin adsorption amounts to diffusion gradients in sediment [5,7,8], a formally rigorous treatment of partition factors PF aiming at their first-principles calculation and prediction still was missing. For a proper understanding, one should use all the available information from the performance of different chitin sorts at the same spot to extract thermodynamic information rather than trying to fit data to one or another of different empirical adsorption isotherms (without using electrochemical data proving pieces of information on adsorption to chitin previously [8]. Differences among metal ions in terms of binding to chitin from water samples cover a range of typical environmental concentrations. Methods and parameters previously used [5,6] to predict complex formation constants also will apply to polymeric ligands like long-chain polycarboxylates, polyacrylonitrile, proteins, or chitin. Numerical integration of the said function should eventually reproduce information on possible bioaccumulation on chitin in both water and different types of sediment. Therefore, elements that are likely or provably not involved in biogeochemical processes, biomethylation, or precipitation in local sediments are selected to avoid perturbations of the partitioned signal PF’. An equation describing the relationship between the variation of (log) PF and that of certain chemical parameters among studied elements is derived and used to compare the measured behavior of elements that are actually involved in biogeochemistry to the PF’ predictions thus calculated. Chemical parameters are related to oxidation state and similar ion size but describe and predict different complex-forming behavior, which is taken to influence PF, too. Exceptions can be accepted for those elements having no essential biological function for whichever organism is present in the respective ecosystem.

Thus, incomplete data from an element detected in one or few of the sites only (e.g., in chitin immersed into the water but not that in sediment or from some water directly) can be extended by using the PFs to calculate missing data for some sites. Efficient biomonitoring means bioconcentration of analytes, that is, partition factors water- or soil→organ, organism >> 1, and thus DNA of an organism alters the distribution function of some analyte around this very organism. This distribution function, including partition onto some introduced organism, is described by the Schrödinger equation using a pseudopotential of adsorption (Figure 1):

Metabolic transformations may increase adsorption because the adsorbed chemical species is constantly removed from equilibrium (cp. uptake of nitrate by roots rendered irreversible by Mo-based nitrate reductase).

Since we deal with mesoscopic (plankton, cell walls of bacteria) to macroscopic (lichens, arthropods, mosses, larger animals or fungi) test organisms and the environment to be measured has dimensions of meters or more, the Schrödinger equation to predict the chance to encounter the analyte or some other molecule, ion at the organism or chitin surface rather than somewhere else in the “box” ([9], and Figure 1) can be replaced with a simple partition function from empirical thermodynamics. Fitting to classical adsorption isotherms is not wise except for a few cases (mainly adsorption of Hg atoms from vapor to lichens which occurs without oxidation of sorbate). Rather cases must be selected where secondary biochemistry will not alter adsorption (i.e., elements not involved in biochemistry at the site at least) to construct an empirical retention function and apply this to analytes intriguing for toxicity- or other reasons. F.e., Cd-based carboanhydrases mainly exist in certain kinds of marine phytoplankton [10], thus, may be encountered when doing such studies in mangrove regions, while W is seen to work in thermophilic and obligately anaerobic archaeons only [11]. As does accumulate in marine arthropods without having a function [12], while I and Se mainly “operate” in vertebrates. Neither I nor Se nor As is a metal, however, and thus they are not capable of forming comparable complexes; in fact, Se (and As) PF′ on chitin were not seen to respond to different conditions so far.

This function can be integrated to obtain theoretical partition factors (see [13] and Section 2). If there is metabolism, measured partition factors should be higher. This does hold for grafted chitin in sediment and parts of organisms; however, grafted chitin obtained from shrimp Pandalus borealis provides more tractable and reproducible data than living arthropods (which yet can survive the procedure as shown before [5,14]. The method using grafted chitin additionally saves time, material- and logistic expenditures. We shall now study the effects of different aquatic environments on the partition function. For technical reasons, so far, the results are limited to di- and trivalent metals and Sb, Bi, but this does, e.g., suffice to cover most problems associated with nuclear fallout ad with common toxic metals. For oxidation states +IV and higher, the a and b terms that describe complex formation properties could not yet be derived with sufficient precision, unfortunately, precluding a similar analysis for partitions of V, Mo, U, Zr, or Pu.

2. Materials and Methods

The exact procedure of calculating expected partition factors from empirical data on water/chitin concentrations and a, b values for the respective ions are described elsewhere [13], so they should not be fully repeated here to avoid self-plagiarism. Chitin samples obtained by either exposing grafted chitin (from Pandalus borealis) or representing parts of crayfish (Faxonius limosus) cover are treated with a defined amount of a solution of 15 g LiClO₄ in 100 mL dimethylformamide (DMF; 0.5 mL of solution/cm² of surface). After 20 s, the solution is absorbed by 350 mg of cation exchanger resin (acid form; Amberlite H-120) and the resin is subjected to back exchange using 1% HNO₃. These data are also used to identify non-equilibrium states due to the vertical diffusion of elements by biological uptake, precipitation, or biomethylation.

The arrow indicates the spring. Water- and sediment samples using grafted chitin were taken from the front part of this ditch, as was the open water sample. Anions include substantial sulfate (299 mg/L ≡ 3.11 mM/L) and little chloride (7.40 mg/L ≡ 208 µM/L) but no other ions. P is fully retained by gel, but As is not, indicating the local reduction of As(V). Photo SF, December 2022.

Pertinent data of this ditch and chitin retention are [7] (

Table 1. Analytical data for ferric gels and overhead water in the ditch shown in Figure 2.

Element	Level in Water [µg/L; nMol/L]	Level on Chitin [µg/L; nMol/L]	Log PF	Level on Chitin in Sediment [µg/L; nMol/L]	Log Ratio PF′
M3+
Al	21.82	16.25	−0.128	109.68	0.829
Cr	2.78	42.72	1.187	62.13	0.163
Bi	0.18	0.04	−0.653	0.04	0
M2+
Ca	89,200	70	−3.1053	60	−0.067
Sr	255.7	0.78	−2.5156	0.32	−0.387
Ba	104.09	0.99	−2.022	0.36	−0.439
Mn	645.57	1.62	−2.600	1.02	−0.201
Co	3.04	0.18	−1.228	0.14	−0.109
Ni	8.23	2.57	−0.5055	3.77	+0.166
Cu	2.27	1.18	−0.284	1.03	−0.059
Zn	10.43	12.51	+0.079	11.57	−0.034
Cd	0.01	0.11	1.041	0.09	−0.087
Pb	0.03	0.16	0.727	0.14	−0.058
Other or unknown oxidation states
V	0.69	5.11	0.870	8.42	0.217
Mo	0.08	0.12	0.176	0.10	−0.079
U	0.02	0	-	0	-
Ag	Not detected (<0.005)	0.18		0.30	0.222

):

K means the empirical slope of the function log PF vs. log [M]_aq for common levels in water; that is, about 1 µM/L for Ba, Mn, and Al, 5–30 nM/L for Ni, Cu, and Zn, and ≤ 2 nM/L for Cd, Pb, La, Ce, and other REEs, and 10 pM/L for Bi [13]. Except for 12 µM/l Mn and somewhat elevated levels of Sr (3 µM), Cr, Ni, Zn, and As (50–300 M/L each) values for a ditch filled with biogenic ferric gels (Figure 2) are similar to those mentioned before. The As level in water over ferric gels indicates that the element is not in its fully oxidized pentavalent state.

k_{2sedim→grafted chitin} = 0.2394 a − 1.761

(1)

k_{1water→grafted chitin} = −0.0831 a − 1.166

(2)

which combine the on-chitin partition factor PF′ to yield

Log PF′ = 0.006743 a² − 0.1359 a + 0.2724 for M²⁺

(3)

Table 2. Chitin desorption timescales and partition factors calculated from the slope of adsorption functions for M²⁺, M³⁺ in water, sediment, ferric gels and water/sediment partition (lower limit excluding activity of organisms in or on the sediment).

Scenario, Conditions, Examples of Metal Ions	Oxidation State, Formula, Range of Corresponding Parameters	Picture	Results
water→chitin
In-water M transport on chitin; sufficiently short desorption timescale (24 h or less), PFaq ≥ 1 alkaline earth, Mn	M2+ k = −0.0831 a − 1.166, that is, k = −1.34…−0.86 (Cu, Pb) → PF = 2.8…1.8		M release from chitin (lower water column) into open water does escape M biomagnification in zooplankton predators but enables M-dependent biochemical activities at sediment interface and above; Mg and Ca desorb so fast that they are not found in prewashed shrimp chitin, whereas some Sr, Ba, and particularly REEs are retained
Same, ferric ditch (FOBs) near Lake Olbersdorf	M2+ k = −0.3914 a − 0.697		M = Sr, Ba, Co, Ni, and Cu
REEs ≠ Y	M3+ K = −0.0113 a − 0.894, that is, k = − 1.04 (Al, Ti(III))…−0.81 (Ga) → PF ≈ 2.05…1.69	same	Methanol oxidation by bacteria operating in oxidizing top sediment layers or aq. slurries
Same, ferric ditch (FOBs) near Lake Olbersdorf	M3+ K = −0.0875 a + 0.960
	Higher oxid. State, e.g., VO2+ for which a ≈ +3	Possibly same	No detailed statement is possible as yet, e.g., N2 assimilation or phenol arene ring halogenations using V
Long τdesorb (>>24 h); heavy 3d ions (Co…Zn), and Cd, Pb, Y			Transfer within water takes place only by predation (M biomagnification in zooplankton predators [fishes, dragonfly nymphs, certain birds]), otherwise deposition to sediment
sediment→chitin
M transfer to sediment upon single or repeated contact (leg tips, partly benthic or digging chitin-clad organisms), (or) action of chitinases in sediment after chitin was discarded (molting, ant wings, …) or deposited from dead organisms; effective if PFchitin,sedim./chitin,water > 1	M2+ for grafted chitin: K = 0.2394 a − 1.761, that is, k = −2.65 (Cu)…−1.30; PF ≈ 22.85…2.6		Includes leg tips (pronounced transport of Mn) Range: grafted chitin: PF′ = 8.2 (Cu, Pb)…1.45 (Ba) Leg tips: PF = 0.65 (Cu, Pb)…1.45Non-equilibrium increasing transfer occurs when there is associated biological activity (Ni)
Same, ferric ditch (FOBs) near Lake Olbersdorf	M2+: k = −0.4799 a − 1.946 for grafted chitin		M = Sr, Ba, Co, Ni, Cu; including Mn: M2+: k = −0.4954 a − 2.042 for grafted chitin
	M3+ for grafted chitin: K = −0.0422 a − 0.778 that is, k = −1.30 (Al)…−0.47 (Ga): PF ≈ 2.61…1.32	same	Range grafted chitin: PF′ = 1.27 (Al)…0.78 (Ga) Leg tips: PF ≈ 0.9 for all M(III) ionsnon-equilibrium increasing transfer to chitin in sediment occurs when there is associated biological activity (V, La, other LREEs)
Same, ferric ditch (FOBs) near Lake Olbersdorf	M3+ for grafted chitin: K = −0.0431 a + 1.238

gives equations and conditions for the sites [13]:

Vanadium has its biochemical roles in haloperoxidases of bacteria, algae, and fungi (usually Hal = Br is introduced, but also Cl, I, NCS, CN, NC, and NO₂ can become bound to benzenoid and naphthenoid aromatics) and in Mo-independent nitrogenases [15,16].

3. Results and Discussion

The data for equations are then combined with empirical data for divalent and trivalent ions to outline the respective parameter spaces: With Mn(II), a is almost exactly zero (a = −0.09), thus log PF′ ≈ 0.272 and thus

PF′_Mn ≈ 1.87

(4)

while the minimum value of a for M²⁺ is −3.73 for Cu²⁺, about −3.5 for Pb²⁺, which means log PF′ = 0.8731 and

PF′_Cu ≈ 7.47

(5)

For Ni (a = −2.50; pertinent to methanogenesis) PF′_Ni should be 10⁰.⁶⁵⁴³ = 4.51 unless for local methanogenesis. Introducing the maximal a-values for divalent ions (Sr and Ba) gives

Log PF′_M2+ ≥ 0.0837

(6)

which becomes

PF′_M2+ ≥ 1.21

(7)

Thus, the range of PF′ for unperturbed M²⁺ covers a factor of about six (see below). Concerning possible release by sulfate reduction, Ba and Pb would represent opposite ends of the range.

The corresponding equations for M³⁺ are

k_2sedim._{→grafted chitin} = −0.0422 a − 0.778

(8)

and

k_{1water→grafted chitin} = −0.0113 a − 0.894,

(9)

giving

Log PF′ = −0.0002536 a² − 0.007278 a − 0.0140

(10)

for M³⁺, cp. eq. (3).

This provides the conclusion that, given the range of a for M³⁺ (maximal values about + 12.48 for Al³⁺, + 11 for some REEs like Dy and Ti³⁺ and minimum −7.2 for trivalent Ga³⁺), log PF′ should not vary too much.

We [5,7,8,9,13] did previous work on the adsorption of metal ions to both grafted chitin and crayfish in water bodies around Zittau (Germany) and in Rownia pod Sniezkam bog (Poland/Czech Republic, near Mt. Sniezka), including a small T-shaped ditch (Figure 2) below a cliff produced during mining at Olbersdorf. The former water bodies (except for the bog) did rather recently form due to flooding or spontaneous groundwater filling of former lignite pits and quarries. The ditch is filled with colonies of Fe-oxidizing bacteria, providing another kind of sediment matrix that might change patterns of element adsorption to chitin placed there (there are no macroscopic aquatic insects, crustaceans, or ostracods in this ditch, so studies had to limit to grafted chitin introduced there [7,17]. This gave a set of data concerning another solid matrix of chitin besides water and “classical” sediments (see results). For slopes describing the relationship of log PF to a, see Table 1. While slopes may be very similar, the axis intercept generally is much higher for a Fe-rich slurry or gel. An increase in k now means a decrease in log PF and, thus, also in PF. Data for different conditions and oxidation states II and III are summarized in

Table 3. k variations for grafted chitin in different conditions.

Oxidation State	Formula Water Chitin in Normal Conditions	Formula Water Chitin in Ferric Gel	Difference	Formula Sediment Chitin in Normal Conditions	Formula Sediment Chitin in Ferric Gel	Difference	Remarks, Crossover Points
+II	k = −0.0831 a − 1.166	k = −0.3914 a − 0.697	Δk = 0.469 − 0.3083 a	k = 0.2394 a − 1.761	k = −0.4954 a − 2.042	Δk = −0.281 − 0.7348 a	K gets smaller when a > 1.52 in water (<−1.29; PF ≥ 2.7) and a > −0.38 in sediment (k < −1.85; PF > 5.64)
+III	k = −0.0113 a − 0.894	k = −0.0875 a + 0.960	Δk = 1.854 − 0.0762 a	k = −0.0422 a − 0.778	k = −0.0431 a + 1.238	Δk = 2.016 − 0.0009 a	k cannot become smaller by the presence of MaFexSyOz *n H2O next to either water or sediment

. For the definition of k, see [13] and explanations below:

For M^2+, slope k is considerably decreased in both water and chitin, while c′ does not change too much. The “critical value” (column 8 of Table 3) a > −0.38 for enhanced sediment→chitin transfer of M²⁺ next to ferric gels means that 3d ion Mn²⁺, alkaline earth will respond positively (adsorb better to chitin when there are ferric gels) but Cd, Fe(II), Zn, and Pb do not. For Ni, observed values of PF′ are <<6.9 unless for local methanogenesis. For water→chitin transfer, only adsorptions of Mg, Ca, possibly Sr, and Eu(II) are increased. This must be considered when looking for highly toxic and divalent ions, thus restricted in drinking water using chitin.

Trivalent ions, including fallout nuclides ²⁴¹Am and fissiogenic or neutrogenic [18] REEs ¹⁴⁴Ce, ¹⁴⁷Pm, ¹⁵¹Sm, or ^152;154;155Eu, to name just those with half-lives >1 a [18], will no longer bind efficiently to chitin from either water or sediment when there is Fe-rich gel from Leptothrix ochracea. The Fe-rich gel M_aFe_xS_yO_z*n H₂O is probably schwertmannite {M = H} or jarosite embedded in a biogenic polysaccharide, given the observed [17] prevalence of S vs. N in the gel. However, Sr (radionuclides ^89;90Sr) will bind to chitin preferentially within such ferric gels. Positive response for a < −0.38 for M²⁺ from sediment is interesting because this happens when there is an already high PF value of 5.64; from water, there is an increase from an initial PF ≥ 2.7, but a_M2+ < −0.38 holds for very few elements only.

As far as elements arising from toxic concerns are considered, chitin enrichment of Cd is little influenced by the presence of such ferric gels, permitting sensitive detection. Chitin-covered organisms inhabiting such gels while moving sometimes or regularly to other water bodies would transport metals in either direction, given the desorption timescale is short enough. Nearby ferric gels will also inhibit desorption of Eu(II) from chitin in some water bodies after photochemical reduction of Eu(III) because here, like with Mg and Ca, a ≈ +2. This conclusion needs studying the PF (k) function (11) [below] for rather positive k (+1 or so) also. Log PF can be calculated from k according to the following equation (excellent regression):

Log PF = 0.1338 k² − 0.152 k + 0.011

(11)

For a = 10 (approximately equivalent to trivalent REEs Sm, Tb, and Er), this means

Log PF′ = −0.02536 − 0.07278 − 0.0140 = −0.11214

(12)

and thus

PF′ = 0.772

(13)

while for a = −5 (approximate value applying to In(III), Cr(III), and anthropogenic Am(III)) it would be

log PF′ = 0.01605

(14)

and thus

PF′ = 1.038

Among M³⁺, that range is rather small, with PF′ covering a factor of about 1.35 as compared to more than six in M²⁺. However, except for LREEs in oxidizing top sediments, there is little biochemical activity associated with trivalent ions, whereas Fe(III) can be involved in precipitation, Bi(III) in biomethylation, and the oxidation state of V on its enzymes is unknown. However, V in Mo-independent nitrogenase likely is divalent (cp. catalytic N₂₋, RNC-, and CO reductions by mixed V/Mg hydroxides and V tren complexes [15,16]) while the speciation form absorbed from the environment is likely trivalent, rather than more oxidized, in appropriate conditions.

Eventually, log PF′ [Equations (3) and (10)] from empirical measurements is compared to calculated values to determine transport into sediment using data for different sites in the region. PF does vary with metal ions significantly (that is, maximal PF_{water→chitin} more than twice minimal value) only for M^2+;3+; grafted chitin in sediment (PF = 22.85…2.6; 2.61…1.32) M³⁺; crayfish antenna (PF = 84.0…1.4), and M²⁺; crayfish carapace in water (PF = 4.8…1.45). Accordingly, for the other cases and kinds of chitin, one can conclude there is an almost proportional relationship between aq. levels and chitin levels for a given element and oxidation state. F.e., a large enrichment of Mn (PF′ ≈ 50) is observed with leg tips in sediment [5,13], which is probably due to the fact that the leg tips repeatedly pass through some redox gradient in upper sediment. Otherwise, data allow us to pinpoint transport processes in sediment [7,13]. Eu, on the other hand, will not make it into the sediment unless for highly oligotrophic conditions, i.e., when photoreduction (and thus detachment from chitin) are blocked by missing organic substrates. Because the activity of chitinases is very low in eutrophic waters [19], there is simple equilibrium throughout the water column unless for chitin-based transport.

Different values for chitin samples other than grafted may thus correspond to increased or decreased detection sensitivity depending on the a-value of the respective element. The empirical equations allow us to calculate which way of measuring may be better (more sensitive, providing higher PF) comparing grafted chitin in water and a crayfish antenna or grafted chitin in sediment and leg tips, respectively. Interestingly in the former case (compartment water), M²⁺ and M³⁺ do not differ significantly concerning PF (1.80 for M²⁺ and 2.01 for M³⁺) beyond which grafted chitin is superior, meaning a_M2+ > −3.20 (i.e., everything but Cu²⁺ and Pb²⁺) and a_M3+ > 9.64 (applies only to Al, Ti(III), Sm, and HREEs Gd…Lu). This should be considered if there is a practical choice of sample carrier. For sediment, grafted chitin should be compared to leg tips, notwithstanding the effects of steadily passing through a very steep redox gradient for this moment. Here for M²⁺ a < +1.98, that is, grafted chitin is superior to leg tips with all “classical” metal ions (there are binding issues with Mg, Ca, and Eu²⁺ where a is somewhat larger) while k < −1.287 and thus PF > 2.68. The corresponding values for M³⁺ in sediment, when grafted chitin works better, are a > 4.05 (i.e., including all REEs, Y, Al, and Ti(III) but not Cr, In, Ga, or Am [²⁴¹Am from fallout]), k < −0.949, and PF > 1.89.

The superposition of both element groups determines which elements can be studied to undergo transport independent of chitin but caused by biogeochemistry (enzymes, biomethylation) or precipitation. This does include Ni but grafted chitin performs a little poorer with respect to La and Ce than the leg tips of Faxonius limosus would do. The analysis/indirect detection of top-sediment methanol oxidation or degradation of crude oil hydrocarbons would still be feasible.

Ultratrace analysis becomes feasible by bioaccumulation on chitin (that is, if PF₁ or PF₂ or both > 1) in cases where, e.g., the level in water is between detection and determination limits where, usually, a factor of four is the quotient between detection and determination limit. It does not matter which kind of sample is analyzed; hence this factor also applies to chitin-based samples. For a significant piece of information, i.e., detection of ions which cannot be pinpointed in water directly, PF₁ should be > 4 (log PF₁ > 0.60), irrespective of the toxicological significance of such traces. Elements that are highly toxic and expected to fulfill the above condition can be identified by calculating criteria for PF > 1 from Equations (1)–(3) and (11)

Log PF = 0.1338 k² − 0.152 k + 0.011

(11)

For M²⁺ log PF = 0.6 amounts to

0.1338 k² − 0.152 k = 0.589

(15)

thus

k² − 1.136 k = 4.4020

(16)

or

(k − 0.568)² = 4.7246

(17)

which gives two solutions,

k₁ = 2.742 or k₂ = −1.606

(18)

The negative solution is now introduced into Equations (1) and (2) to obtain limiting values of a for ultratrace determinations of divalent and Equations (10) and (11) for trivalent ions:

k_2sedim._{→grafted chitin} = −0.0422 a − 0.778

(8)

k_{1water→grafted chitin} = −0.0113 a − 0.894

(9)

It is obvious that the criterion cannot be met for trivalent ions in either water or sediment using grafted chitin (see below), while with M²⁺, the negative solution is fulfilled for

0.2394 a < 0.155

(19)

that is,

a < +0.65

in sediment and

−0.0831 a < −0.44

(20)

=a > 5.3 in water, respectively. Meanwhile, 3d dications Fe, Zn, Cd, Pb, and Eu(II) fulfill the criterion for (grafted chitin in) sediments (Cd is seen there [13] but never in water). However, this limiting condition applies to no dications in water, matching the observations. Very high enrichments of M³⁺ on chitin were exposed to water briefly (grafted) or for long periods of time (antenna), for which the empirical equation is:

k_{1water→chitin} = 0.1344 a − 2.298

(21)

Accordingly, a must be >−5.15 (that is, all M³⁺ save Ga, possibly In and Am) for the “sufficient” determination on chitin when there was no detection in water are observed with trivalent ions Bi, V, Cr, and several REEs.

Accordingly, elements are seen only on chitin but not (always) in water using the analytical gear available to us, such as Eu and Cd, yet can give additional information on aq. levels by chitin adsorption. The partition factor of Eu(III) (a = 9.52) between water and chitin embedded in silica (≈sediment can be estimated from the following formula for trivalent ions (Equations (8) and (9)):

k_2sedim._{→grafted chitin} = −0.0422 a − 0.778

(8)

k_{1water→grafted chitin} = −0.0113 a − 0.894

(9)

giving with

Log PF = 0.1338 k² − 0.152 k + 0.011

(11)

The PF′ values are shown in Table 2 and Table 3.

For Eu(III), this gives k_sedim = −1.18 and k_water = 1.002, which means PF_sedim. = 2.38 and PF_water = 1.98, respectively. The detection limit of Eu is rather high [20] compared to other REEs or Bi due to superposition with ¹³⁵BaO and ¹³⁷BaO molecule peaks (representing 6.6 and 11.2%, respectively, of Ba′s natural isotopic mix [18]) in ICP/MS mass spectra. In addition, one might anticipate the photoreductive desorption of Eu from chitin [8,21]. Yet, samples were taken at morning dawn throughout [13,20], prompting us to assume highest Eu loads on chitin over daytime with this sampling protocol.

Biomethylations of all Tl, Pb, Cd, and dimethyl sulfide formations are both associated with chlorophyll-a formation rates/levels [22]. In our dataset, there are sizable “anomalies” with Pb retention on chitin, causing us to exclude this element from the derivation of the above chitin binding k regression equations. For Cd, a = −0.89 and thus k₁ (water) = −1.09 and k₂ (sedim.) = −1.974, thus, PF_{water→chitin} = 2.17 and PF_sedim._→chitin = 6.74, corresponding to PF′ = 3.11. For Dy, Sm, Eu, Gd, and Yb, significant uptake is observed only with parts of crayfish (antennae, carapace) and water [13]. Accordingly, slow uptake (and desorption) only provides a measurable signal here, while reaction rates of adsorbed ions in sediment are low.

4. Applications to Environmental Safeguarding

For applied biomonitoring of highly toxic elements (As, Sb, Cd, U, Pb, and other di- or trivalent ones like Be, Nd, Tb), one would start with legislation limits for drinking water and detection limits in water or chitin [20] and compare them to PF values calculated in the above manner (

Table 4. Properties of ions to be surveyed in drinking water and their expected partition factors in water and sediment.

Element/Metal	Drinking Water Limit [µg/L] (EPA, EU Regulations)	Drinking Water Limit [nMol/L]	a	K; PF in Water	K; PF in Sediment	PF′, Possible Deviations
Zn	5000	76,500	−1.42	−1.28; 2.66	−1.42; 3.14	1.18; deviation is always expected due to key biochemical roles of Zn
Cd	0.5	4	−0.89	−1.24; 2.54	−1.55; 3.70	1.46; deviations possibly occur in mangrove
Pb	10	50	−3.5	−1.46; 3.28	−0.92; 1.84	0.56; deviations possibly associated with sulfate reduction?
U	2 (EPA 30)	8.5 (EPA 130)	not yet determined	Cannot be calculated because oxid. state differs	Cannot be calculated because oxid. state differs	-
Ni	20	300	−2.50	−0.87; 1.75	−0.88; 1.78	1.02
Cu	1300	20,000	−3.73	−1.48; 3.36	−2.65; 22.74	6.76
Mn	300	5500	−0.09	−1.17; 2.36	−1.78; 5.09	2.16; likely responds to redox gradient in top sediment layers→PF′increases to ≤50!
Cr	100	1900	−2.60 (assuming Cr3+)	−0.92; 1.84	−0.89; 1.78	(0.97); total Cr limit because there is a rapid change among oxidation states, e.g., catalyzed by ambient MnO2

):

Since data for di- and trivalent ions are available, and correlations for the relationship between k—and, in turn, PF and PF′—and a were derived above, it is feasible to estimate also accumulation of nuclear-technology-related radionuclides from Sr up to Am at least. This includes all (fissile and non-fissile) n targets, i.e., REE “neutron poisons” like ¹⁴⁹Sm, ¹⁵¹Eu, or ¹⁵⁷Gd, and ⁸⁸Sr, ⁶⁴Zn, ²³⁸U and fission products other than Tc, Ru, and I (these latter three and U, Pu likely display oxidation numbers ≠ +II or +III). Adsorption of long-lived (T ≥ 1 a) nuclides from nuclear energy production or bomb tests on chitin which was put to sediment will cover all Am, Sr, Y, Ce, Pm, Sm, and Eu (Cs does not adsorb to chitin [4]) while the behavior of Tc, Ru, Pd, U, and Pu cannot be determined from the above reasoning due to their differing oxidation states or lacking information with pertinent oxidation states (Pd, Ru). Recall that the original application of chitin adsorption of M ions, -complexes dealt with fission product retention (¹⁰⁶Ru nitrosyl complex [3]) and reprocessing of nuclear fuel rods [3,4]). While ambient nuclear fallout now almost entirely consists of nuclides with half-lives > 10 y, implying that pertinent isotopes of Ru, Ba, La, Ce, and Pm (for all of which T_1/2 ≤ 2.5 a [18]) decayed away long since emissions took place, the corrected PF values indicate that even traces of these nuclides can be pinpointed on chitin, also far from recent disaster- or testing sites such as Fukushima or Punggye-ri (North Korea). In addition, when directly measuring adsorbed ions by their radioactivity (β and γ energies), there is no dilution due to workup protocols; that is, the signals are going to be some 20 times more intense than estimated from above PF values (Table 2, Table 3 and Table 4).

5. Conclusions

A straightforward mathematical approach relying on a large set of empirical data concerning enrichment (or, more generally speaking) is applied to predict adsorption thermodynamics while avoiding ad hoc fitting to any classical adsorption isotherm. There are indications for at least two different metal ion binding sites on chitin, while both ligand exchange and electrochemistry can be accomplished with dissolved chitin in DMF/Li⁺. It is unlikely that the latter fact will compromise analytical results. Thus, any deviation from PF′ can be ascribed to either biochemical (including biomethylation), precipitation or enforced multiple passage through the water/sediment interface if the corresponding metal is likely to change its oxidation state upon this process while passing the interface (e.g., Mn). The “benchmark” calibration line thus is constructed using other elements for M²⁺ while trivalent metals Fe, V, and LREEs, as a rule, do not take part in biochemistry, meaning that it is easier to produce the corresponding function; in addition, PF values for M³⁺ tend to be higher, perhaps also due to the isoelectric point of chitin (4.6) being lower than for most ambient waters and sediments. Grafted chitin is more likely to produce reproducible results, and it makes one independent of local biota existing at the spot (lichens, crayfish, larger zooplankton like sandhoppers, terrestrial and aquatic beetles, and spiders); thus, grafted chitin is preferred and recommended to use. It is cheap and avoids both the killing of local organisms for testing purposes and the digestion of samples during workup, which is costly and sometimes risky, e.g., when digesting with HClO₄.

For practical biomonitoring as well as applications of chitin adsorption in the separation of oxidation states and the biotransport of elements, e.g., by zooplankton (crustaceans, copepods, ostracods) or wastewater treatment, desorption kinetics is more important than the actual equilibrium position which is in favor of chitin adsorption (average capacity some 45 µmols/g chitin, except for Ce, Eu, or Pb) down to levels <1 nM/L in water. For alkaline earths, desorption would take place within 1–2 h, permitting leaching from surfaces by washing (Mg, Ca). In REY elements, Y is efficiently retained over months in pure water while La and Lu undergo desorption within some 10–50 h; the rate is increased by photochemistry with Eu(III) (the predicted fractionation between Ho and Y on chitin which requires experimental corroboration yet). Only Ni, Cu, and Pb partly stick to the surface while other di- and trivalent ions “hop” from one parallel chitin biopolymer strand to the next in the solid, causing diffusion, which is fastest with Zn²⁺. This does not mean ions retrieved some weeks to months before by surface adsorption will escape detection because they will not migrate beyond the dissolution horizon in the established and optimized procedure but, on the contrary, safeguards analytical results against perturbations by washing, probably including uptake from adsorbed particles.

The sensitivity of adsorption and partition measurements is good enough for freshwater since PF (uncorrected) = 1.8–3.4 for metals to be controlled. If PF > 4, also metals can be found which cannot be detected in water due to low levels.

6. Outlook

Theoretical considerations suggest one might even determine from partition in water which oxidation of elements like V, Mn, Fe, or Ce exists in water or whether there are biomethylation products (known to prevail for Tl in the ocean [22]) because chitin as such does not undergo any redox reactions with anodes or common saccharide-cleaving reagents like lead(IV) acetate, IO₄^-, or OsO₄ (the latter two do react with chitosan, however [23,24]). Biomethylation, which reduces retention to chitin, suggests that common dimethylsulfoniopropionate precursor and the respective metal in methyl transferase enzyme could likewise be studied by adsorption to chitin while the PF′ of Co should increase and PF of Tl (likely almost zero), Pb and Cd should decrease. Though chitosan, unlike chitin, can be stained using OsO₄ [24], the latter maintains its reactivity towards alkenes and arylated olefines such as styrene if co-bound on chitin along hexacyanoferrate (III) [25] in an alkaline water/t-butanol mixture, with OsO₄ amount controlling chiral excess of obtained products (e.e. = 40–78%).

A mobile sensor will be developed to study the corresponding effects in an open field, particularly to elucidate the effects associated with degrading permafrost.