Cadmium(II) Complex Formation with Selenourea and Thiourea in Solution: An XAS and ¹¹³Cd NMR Study

Abstract

The complexes formed in methanol solutions of Cd(CF₃SO₃)₂ with selenourea (SeU) or thiourea (TU), for thiourea also in aqueous solution, were studied by combining ¹¹³Cd NMR and X-ray absorption spectroscopy. At low temperature (~200 K) distinct ¹¹³Cd NMR signals were observed, corresponding to CdL_n²⁺ species (n = 0 - 4, L = TU or SeU) in slow ligand exchange. Peak integrals were used to obtain the speciation in the methanol solutions, allowing stability constants to be estimated. For cadmium(II) complexes with thione (C=S) or selone (C=Se) groups coordinated in Cd(S/Se)O₅ or Cd(S/Se)₂O₄ (O from MeOH or CF₃SO₃^-) environments the ¹¹³Cd chemical shifts were quite similar, within 93-97 ppm and 189 – 193 ppm, respectively. However, the difference in the chemical shift for the Cd(SeU)₄²⁺ (578 pm) and Cd(TU)₄²⁺ (526 ppm) species, with CdSe₄ and CdS₄ coordination, respectively, shows less chemical shielding for the coordinated Se atoms than for S, in contrast to the common trend with increasing shielding in the order: O > N > Se >S. In solutions dominated by mono- and tetra-thiourea / selenourea complexes, their coordination and bond distances could be evaluated by Cd K-edge extended X-ray absorption fine structure (EXAFS) spectroscopy. At ~200 K and high excess of thiourea a minor amount (up to ~30%) of [Cd(TU)_5-6]²⁺ species was detected by an upfield shift of the ¹¹³Cd NMR signal (up to 423 ppm) and an amplitude reduction of the EXAFS oscillation. The amount was estimated by fitting linear combinations of simulated EXAFS spectra for [Cd(TU)₄]²⁺ and [Cd(TU)₆]²⁺ complexes. At room temperature, [Cd(TU)₄]²⁺ was the highest complex formed, also in aqueous solution. Cd L3-edge X-ray absorption near edge structure (XANES) spectra of cadmium(II) thiourea solutions in methanol were used to follow changes in the CdS_xO_y coordination at room temperature. The correlations found from the current and previous studies between ¹¹³Cd NMR chemical shifts and different Cd(II) coordination environments are generally useful for evaluating cadmium coordination to thione-containing or Se-donor ligands in biochemical systems or for monitoring speciation in solution.

Introduction

Thione-containing molecules have important biochemical functions and also industrial applications. For example, L-ergothioneine (a thiohistidine) is a naturally occurring amino acid with antioxidant properties that can be made by few organisms (bacteria and fungi), taken up by plants and ingested by animals and human.¹ It has two tautomeric forms: thiol and thione;² the latter predominates at physiological pH in aqueous solution,³ and can form complexes with many metal ions: Cu(II) > Hg(II) > Zn(II) > Cd(II) > Ni(II) ~ Co(II), in order of increasing stability.^4,5 Some diabetic patients obtain high levels of ergothioneine, which due to the high stability of its zinc complexes (logβ₂ ~10¹²),⁴ leads to zinc deficiency and glucose intolerance.⁶ Also, it has been shown that ergothioneine has protective effects toward cadmium-induced teratogenesis.⁷

Increasing our knowledge about how interactions related to group 12 zinc(II) and cadmium(II) ions affect biological systems is essential. Although its ionic radius is larger (~0.2 Å),⁸ cadmium(II) is known to attach to the binding sites of Zn-enzymes / proteins with higher affinity than zinc(II) and thereby inhibit important functions.^9,10 Thiourea (TU, Scheme 1) provides a structurally simple model for studies of thione-containing ligands. The structures of many metal-thiourea complexes, in which the thiourea molecule is coordinated to the metal center via its S atom, are available in the Cambridge structural database (CSD),¹¹ as e.g. for the four-coordinated [Cd(TU)₂Cl₂],¹² [Cd(TU)₄](NO₃)₂,¹³ and six-coordinated [Cd(TU)₂(NO₃)₂],¹³ [Cd(TU)₂(HCOO)₂],¹⁴ and [Cd(TU)₆](ReO₄)₂ complexes.¹⁵ In aqueous solution cadmium(II) complexes with up to four thiourea ligands have been reported with the formation constants β₁ = 20, β₂ = 140, β₃ = 260 and β₄ = 1200 (Figure S-1), based on a polarographic study.¹⁶

Scheme 1.

Thiourea (TU) and selenourea (SeU)

CdS semiconducting nanoparticles (NPs) and thin films have promising photo-electrochemical applications, e.g. in solar cells.¹⁷ Highly luminescent water-soluble CdS NPs can be prepared by thermal decomposition of Cd(II) and thiourea mixtures. Their size influences the optical activity,¹⁸ and the photoluminescence quantum yield depends on pH, and the concentration and ratio of thiourea to cadmium(II).^19,20 O’Brien and Saeed deposited thin films of CdS from alkaline solutions of Cd(CH₃COO)₂, ethylenediamine and thiourea, and correlated the quality and morphology of CdS thin films to the conditions used for growth, e.g. by following speciation changes in the solution by ¹¹³Cd NMR spectroscopy.²¹ The observed ¹¹³Cd NMR chemical shifts, however, were not connected to specific Cd(II) species. ¹¹³Cd NMR spectroscopy has also been used to investigate the formation mechanism of Cd NPs through oxidative aggregation of [Cd₁₀S₄Br₄(SR)₁₂]^4- (SR = p-tolylthiolate) by elemental sulfur.²²

CdSe NPs and thin films have been prepared from mixtures of selenourea and cadmium salts in similar ways as for cadmium sulfide. Selenourea (SeU, Scheme 1) is a resonance-stabilized selenoketone (selone) with π-conjugation between the selenocarbonyl and amino groups.²³ It is sensitive to heat, light and moisture and is easily oxidized in acidic aqueous solution forming a Se-Se bond in a diselenobis(formamidinium) cation,²⁴ or decomposed by atmospheric oxygen to red elemental selenium in neutral / alkaline aqueous solution,²⁵ and even in methanol and ethanol.²⁶ Hydrolytic decomposition of selenourea generates cyanamide (H₂CN₂), Se^2- and HSe^-, and is accelerated by increasing pH.^25,27,28 Hydrochemical decomposition of metal selenourea compounds to metal selenide nanoparticles is a common procedure for generating thin films for new semiconductor materials.²⁷

While several reports on syntheses of simple selenourea metal complexes all conclude that the Se atom is the only coordination site for metal ions,^26,29,31 very few crystal structures are available^32,33 (apart from a few with selenourea derivatives). There is only one structurally known complex where a selenourea derivative, N,N’-diphenylselenourea, acts as a bidentate ligand with (Se, N) coordination to a Pt(II) ion.³⁴ A series of selenourea complexes with different Co(II), Zn(II), Cd(II) and Hg(II) salts including Cd(SeU)₂X₂ (X = halide), were prepared and carefully analyzed using electronic (UV-vis.) and vibrational (IR) spectroscopic techniques, reporting metal-selenium stretching vibrations within the range 167 – 245 cm^-1.³⁵ A solid state ¹¹³Cd NMR spectrum of [Cd(SeU)₂Cl₂] showed an isotropic chemical shift of δ_iso(¹¹³Cd) = 458 ppm.³⁶ Recently, spray hydrolysis of [Cd(SeU)₂Cl₂] in aqueous solution, prepared by reacting CdCl₂ hydrate with 4-fold selenourea, was used to generate CdSe films.³⁷ CdSe nanocrystals have also been prepared by fast injection of a base to a solution of Cd(NO₃)₂·4H₂O and selenourea at room temperature. The mechanism of this reaction seems to be via an unidentified intermediate Cd-selenourea complex.³⁸ A polarographic study showed that Cd(II) ions form a selenourea complex with limited solubility in de-aerated water with [Cd²⁺] = 0.01 mM.³⁹ The formation constant for the 1:1 complex between cadmium(II) and SeU was recently determined to be log β₁ = 2.11 ± 0.04 in 0.5 mol dm^-3 HClO₄ and ionic strength I = 1.0 (NaClO₄) at 298 K using UV-vis. spectroscopy,⁴⁰ which differs from the previously reported value log β₁ = 0.9 (log β₂ = 3.7) using a cadmium amalgam electrode.⁴¹

In the current study, we compared the complex formation of cadmium(II) ions with selenourea and thiourea as ligands, keeping C_Cd(II) = 0.1 mol dm^-3 and varying the mole ratios SeU / Cd(II) from 1.0 to 5.0 and TU / Cd(II) from 1.0 to 20.0. The speciation and structure of the complexes were investigated with ¹¹³Cd NMR spectroscopy as the main tool, complemented by Cd K-edge EXAFS and Cd L₃-edge XANES spectra. Methanol was chosen as solvent because selenourea reacts with water,²⁵ and it also allowed low temperature studies (at ~200 K). In addition, the cadmium(II)- thiourea complexes formed in aqueous solution were studied at room temperature.

¹¹³Cd NMR spectroscopy has in the past few decades developed into a powerful tool for studying the coordination of cadmium ions in compounds and biological systems.^42,43 The ¹¹³Cd NMR chemical shift depends on several factors including: number and type of the donor atoms (with the general trend O > N > Se > S, in order of increasing shielding) and the coordination geometry around cadmium, the nature and charge of the ligand (e.g. different shielding of sulfur in thiolate S^- and thiocyanate SCN^-) and its coordination mode (bridging vs. terminal).^43,44 This is evident from the ¹¹³Cd NMR isotropic chemical shifts (δ_iso) for crystalline Cd(II) complexes with S-thione containing ligands in Table 1, and the ¹¹³Cd NMR chemical shifts for Cd(II) complexes with Se-donor ligands in Table 2. In solution ligand-exchange dynamics affects the peak shape and peak position, which becomes a weighted average of the chemical shifts of all Cd(II) species involved in fast ligand-exchange equilibria.⁴⁵ The temperature and the nature of the solvent also influence the ¹¹³Cd NMR chemical shift.

Table 1.

Solid state ¹¹³Cd NMR isotropic chemical shifts (δ_iso) for crystalline Cd(II) complexes with S-thione containing ligands; see Figures S-3 and S-4 (a, b)

Cd(II)-thione compounds a	Coordination	113Cd NMR chemical shift (δiso, ppm)	Reference
[Cd(tztH)4](CF3SO3)2	CdS4	605 b	46
[Cd(TU)4](NO3)2	CdS4	577 (this work)	13
[Cd(TU)2Cl2]	CdS2Cl2	462 c	36
[Cd(tzt)2]	CdS2N2	430 b	47
[Cd(DMTF)6](ClO4)2	CdS6	402 (this work)	48
(HIm)[Cd(tsac)3(H2O)]	CdS3O	378 (this work)	49
[Cd(TU)3(SO4)]	CdS3O	346 (this work)	50
[Cd2(6-MP)4(H2O)2](NO3)4	CdS3N2O	313	51
trans-[Cd(DMTF)4(CF3SO3)2]	CdS4O2	263 (this work)	48
[Cd2(6-MP)4(NO3)2](NO3)2	CdS3N2O2	233	51
cis-[Cd(TU)2(H2O)4](SO4)	CdS2O4	148 (this work)	52

^atztH = 1,3-thiazolidine-2-thione, DMTF = dimethylthioformamide, HIm = imidazolium, tsac = thiosaccharinato (with negatively charged thionate-S)⁴⁹, 6-MP = 6-mercaptopurine;

^bthe calibration is shifted (see the text) by changing δ_iso for the solid reference compound 3CdSO₄·8H₂O from -77.4 to -58 ppm;

^cthe calibration is shifted by changing δ_iso for the solid reference compound Cd(NO₃)₂·4H₂O from -132.2 to -100 ppm.

Table 2.

¹¹³Cd NMR chemical shifts for mononuclear Cd(II) complexes with Se-donor ligands

Cd(II) compounds a	Coordination	113Cd NMR chemical shift (δ, ppm)	Reference
[Cd(dmpSe)2](ClO4)2	CdSe2P2	670 (CH2Cl2, 193 K)	55
Cd[N(iPr2PSe)2]2	CdSe4	629 (solid)	56
CdSe (bulk)	CdSe4	558 (solid) b	57,58
[Cd(SePh)4](NO3)2	CdSe4	533 (CH3OH, 213 K) c	59
Cd(L-ImSe)2Cl2	CdSe2Cl2	506 – 591 (solid) d	60
Cd(SeU)2Cl2	CdSe2Cl2	490 (solid) d	36
[Cd(Se-aryl)2(bpy)]	CdSe2N2	460, 472 (solid), 411 (CHCl3)	61
[Cd(SeC(O)R’)2(tmeda)]	CdSe2N2O2	388 - 391 (CDCl3) e	62
[Cd(SeC5H4N)2(tmeda)]	CdSe2N4	300 (CDCl3) e	63
(Ph4P)[Cd(SeC(O)Tol)3]	CdSe3O2	250 (CH2Cl2, 193 K)	64

^admpSe = Ph₂PCH₂P(Se)Ph₂; N(ⁱPr₂PSe)₂^- = Tetraisopropyldiselenoimidodiphosphoinato; ^-SePh = phenylselenolate; L-ImSe = imidazolidine-2-selone, L = H, CH₃, C₂H₅, n-C₃H₇, i-C₃H₇; Se-aryl = ^-Se-2,4,6-ⁱPr₃-C₆H₂; R’ = Ph, Tol, Tol = C₆H₄-p-CH₃; tmeda = N,N,N’,N’-tetramethylethylenediamine;

^bThe value reported here is referenced relative to Cd(ClO₄)₂ (aq) by changing the chemical shift of solid Cd(NO₃)₂·4H₂O from 0.0 to -100 ppm⁴³;

^cthe calibration for 4.5 mol·dm^-3 Cd(NO₃)₂ (aq) was changed from -65 to -73 ppm;

^dthe calibration was adjusted by changing δ_iso for Cd(NO₃)₂.4H₂O from -132.2 to - 100 ppm;

^ereported relative to Cd(CH₃COO)₂·2H₂O in CH₃OD (0.0 ppm).

¹¹³Cd NMR chemical shifts from different sources are sometimes calibrated with different standards.⁵³ To keep consistency in Tables 1 and 2 between our ¹¹³Cd chemical shifts calibrated as 0 ppm for solid Cd(ClO₄)₂·6H₂O and those previously reported, we have used -100 ppm as calibration point for solid Cd(NO₃)₂·4H₂O, and -45 and -58 ppm for the two signals for solid 3CdSO₄·8H₂O.⁵⁴ For the solution spectra our calibration 0 ppm for 0.1 mol·dm^-3 Cd(ClO₄)₂(aq) corresponds to -73 ppm for 4.5 mol·dm^-3 Cd(NO₃)₂(aq) (Table 2).⁴³

The low temperature measurements presented here show that at ~200 K the Cd(II)-selenourea and most Cd(II)-thiourea complexes are in slow ligand-exchange, resulting in a separate ¹¹³Cd NMR peak for each Cd(II) complex (Figure 1). Their peak positions and areas allowed quantitative speciation of the Cd(II)-thione / selone complexes, and also an estimate of their stability constants. The EXAFS studies of the solutions with a dominating Cd(II)-thione / selone complex corroborated the assignments of the ¹¹³Cd NMR resonances and provided additional structural information.

Figure 1.

¹¹³Cd NMR spectra of 0.1 mol dm^-3 Cd(CF₃SO₃)₂ methanol solutions with increasing mole ratios L/Cd(II) of selenourea (S1 – S5; left, at 203 K) or thiourea (T1 – T7; middle at 203 K; right at 295 K). For notations see Table 3.

The results and the methodology used in this research provide benchmarks for evaluating coordination environments involving Cd(II) ions and thione/ selone groups in complex biological systems, or for monitoring speciation in solution, e.g. to control the formation of CdS/ CdSe nanoparticles via thiourea/ selenourea complexes.

Experimental Section

Sample preparation

Selenourea, thiourea and methanol-d₄ (CD₃OD) from Sigma-Aldrich were used without further purification. Methanol (MeOH) was dried by refluxing twice over magnesium turnings under argon atmosphere. Dehydrated cadmium triflate Cd(CF₃SO₃)₂ was prepared by dissolving freshly prepared Cd(OH)₂ in triflic acid, in a procedure described previously.⁶⁵

Cadmium(II) selenourea / thiourea solutions

Cadmium(II) triflate solutions with [Cd²⁺]_tot = 0.1 mol dm^-3 and varying amounts of SeU or TU in O₂-free, dried methanol were prepared in a glove box with dry inert atmosphere. For the ¹¹³Cd NMR measurements the solvent contained 30% CD₃OD. Five solutions with SeU / Cd(II) mole ratios of 1.0 (S1), 2.0 (S2), 3.0 (S3), 4.0 (S4) and 5.0 (S5) were prepared. Higher SeU / Cd(II) mole ratios were prevented by the limited solubility of SeU in methanol at ~200 K. Similar cadmium(II) thiourea solutions in methanol were prepared with TU / Cd(II) mole ratios 1.0 (T1), 2.0 (T2), 3.0 (T3), 4.0 (T4), 5.0 (T5), 8.0 (T6) and 10.0 (T7). To check the solvent effect, a parallel series of Cd(II)-thiourea solutions (W1 to W9) were prepared in distilled water containing 10% D₂O. Table 3 presents the composition of the solutions.

Table 3.

Composition of the cadmium(II)-selenourea and thiourea solutions ^a

Solutions in MeOH	L/CdII mole ratio	[Cd2+]total (mmol·dm-3)	[L]total (mmol·dm-3)	Solutions in H2O
L = SeU
S1	1.0	100	100
S2	2.0	100	200
S3	3.0	100	300
S4	4.0	100	400
S5	5.0	100	500
L = TU
T1	1.0	100	100	W1
T2	2.0	100	200	W2
T3	3.0	100	300	W3
T4	4.0	100	400	W4
T5	5.0	100	500	W5
T6	8.0	100	800	W6
T7	10.0	100	1000	W7
		100	1500	W8
		100	2000	W9

^aFor ¹¹³Cd NMR measurements, the methanol solutions contained 30% CD₃OD and aqueous solutions 10% D₂O.

¹¹³Cd NMR reference compounds

The crystalline cadmium(II) dimethylthioformamide (DMTF) compounds [Cd(DMTF)₆](ClO₄)₂ (coordination model for CdS₆) and [Cd(DMTF)₄(CF₃SO₃)₂] (CdS₄O₂ model), imidazolium tris(thiosaccharinato)aquacadmate(II) (HIm)[Cd(tsac)₃(H₂O)] (CdS₃O model), and [Cd(TU)₄](NO₃)₂ (CdS₄ model) were synthesized according to descriptions in the literature,^13,48,49 and were verified by comparing their unit cell dimensions with those of the known structures (Figure S-2).

¹¹³Cd NMR measurements

The ¹¹³Cd NMR spectra of the methanol solutions S1-S5 and T1-T7 were collected at 203 K with a Bruker DRX-400 spectrometer at 88.56 MHz, using a 5 mm broadband (BBO) probe, zgdc pulse program, a sweep width of 800 ppm, and recycle delay of 3.0 seconds. All the above solutions contained ~ 30% CD₃OD. A 0.1 mol·dm^-3 solution of Cd(ClO₄)₂·6H₂O in D₂O was used as external reference (0 ppm at 295 K).⁴³ The temperature of the instrument was calibrated using ¹H NMR of methanol.⁶⁶ The spectra of the Cd(II)-thiourea solutions T1 – T7 in methanol and W1 – W9 in water (10% D₂O) were also measured under similar conditions at room temperature (295 K). The total number of scans collected for each sample can be found in Table S-1a. The integrated areas of the peaks in the ¹¹³Cd NMR spectra of the methanol solutions T1 – T3 and S1 – S3 (203 K) are listed in Table S-1b.

Solid state ¹¹³Cd NMR spectra of the crystalline cadmium(II) reference compounds were measured using a Bruker AMX 300 spectrometer equipped with a MAS BL4 double-resonance broadband probe, at the resonance frequency 66.59 MHz. The samples were packed into 4 mm ZrO rotors with magic angle spinning (MAS) rates between 5 – 8 kHz. Solid Cd(ClO₄)₂·6H₂O was used for calibration (0 ppm at room temperature). Spectra were acquired at 300 K using a 4 μs ¹H 90° pulse, 4 ms contact time and 10 s recycle delay. The isotropic chemical shifts are listed in Table 1.

X-ray absorption spectroscopy

Cadmium K-edge EXAFS spectra of the solutions S1, S5, T1 - T7 were collected at BL 7-3, and Cd L₃-edge XANES spectra of solutions T1 – T7 were measured at room temperature at BL 4-3 at the Stanford Synchrotron Radiation Lightsource (SSRL) with the storage ring operated at 3.0 GeV and 150-200 mA. The EXAFS spectra for solutions S1 and S5 were measured at 200 K (LT) by means of a liquid-helium cryostat, while for T1 – T7 spectra were collected both at 200 K and at room temperature (RT). Pure methanol solidifies below ~ 179 K, so by keeping the temperature at 200 K the samples remained unfrozen and were kept inside the cryostat under chilled air flow (not vacuum) at all times (Note: N₂ in the air freezes below ~ 80 K). At BL 7-3, higher harmonics from the Si(220) double crystal monochromator were rejected by detuning to 50% of the maximum incident beam (I₀) intensity, while a Rh-coated mirror was used at BL 4-3 to eliminate higher harmonics from a fully tuned Si(111) double crystal monochromator. The EXAFS spectra were recorded in transmission mode, with all ion chambers filled with argon, while the Cd L₃-edge XANES spectra were measured in fluorescence mode using a PIPS solid state detector (I_f) and a helium-filled ion chamber (I₀) monitoring the X-ray fluorescence and incident beam intensities, respectively. For EXAFS measurements, the samples were held between Kapton windows in 2 mm pinhole cells sealed with vacuum grease at LT, or in 5 mm Teflon spacers at RT. Two to six scans (10 for S1) were collected for each sample. During the Cd L₃-edge XANES measurements, two scans were measured for each solution held within a 5 mm Teflon spacer between 4 μm Ultralene^© windows that were found to resist penetration of He gas that otherwise could form bubbles in the sample. Before averaging the overlapping scans using the EXAFSPAK program,⁶⁷ the energy scale was externally calibrated for each scan by assigning the first inflection point of the Cd K- and L₃-edge of a Cd foil to 26711.0 eV and 3537.6 eV, respectively.

XAS data analysis

The WinXAS 3.1 program suite was used for the data analysis.⁶⁸ The background absorption was subtracted with a first-order polynomial over the pre-edge region, followed by normalization of the edge step. For the Cd K-edge EXAFS spectra, the energy scale was converted into k-space, where k = [(8π²m_e/h²)(E-E₀)]^1/2, using the threshold energy E₀ = 26709.7 – 26711.0 (S1, S5), 26710.0 – 26713.0 eV (T1 – T7 and W1 – W7). The EXAFS oscillation was then extracted using a 7-segment cubic spline to remove the atomic background absorption above the edge. Details on extraction of structural parameters from the experimental EXAFS oscillation can be found elsewhere.⁶⁹ Theoretical EXAFS oscillations, χ(k), were constructed using the FEFF 8.1 program,^70,71 with FEFF input files generated by means of the ATOMS program⁷² using the crystal structures of Cd(SCH₂CH₂NH₂)₂,⁷³ [Cd(dmise)₄][PF₆]₂ where dmise = 1,3-dimethyl-2(3H)-imidazoleselone,⁷⁴ and [Cd(CH₂(COO)₂)(TU)₂]⁷⁵ (replacing S by Se in the input file), with CdS₃N₂, CdSe₄ and CdS₂O₃ coordination geometries, respectively. Note that two neighboring elements in the periodic table (such as oxygen and nitrogen) with very similar amplitude functions cannot be distinguished by EXAFS.

The structural parameters were refined by least-squares methods, fitting the model function χ(k) to the experimental unfiltered EXAFS oscillation using k³-weighting, generally over the k range 3 – 14 Å^-1, allowing the bond distance (R), Debye-Waller parameter (σ) and ΔE₀ (with a common value for all scattering paths) to float, while the amplitude reduction factor (S₀²) and / or coordination number (N) were fixed. The S₀² value refined to 0.99 at the EXAFS data analysis for crystalline [Cd(TU)₄](NO₃)₂ with the coordination number fixed (4 Cd-S; Figure S-7). This parameter value was then kept fixed at 1.0 during the EXAFS curve-fitting process for the solutions (Tables 4, 5 and S-2). The accuracy of the refined coordination numbers N, bond distances R and their Debye-Waller parameters (σ²) for the dominating Cd-Se and Cd-S paths are estimated to be within ± 20%, ± 0.02 Å and ±0.001 Ų, respectively, accounting for effects of systematic deviations in the extracted data, model limitations, correlated parameters, etc. The above error limits were estimated from variations in the parameter values when refining models of different complexity, systematically varying the k-range, or when trying different cubic spline removals. The corresponding structural parameters for the Cd-O path are less accurate, ± 0.04 Å for bond distances (± 0.02 Å for SeU complexes) and ± 0.002 Ų (± 0.001 Ų at LT) for Debye-Waller parameters σ², due to the difficulties associated with separating the EXAFS contribution from the oxygen (or nitrogen) atoms when their mean bond distance is rather close to that of the heavier sulfur atoms.

Table 4.

Structural Parameters Derived from EXAFS Least-Squares Curve Fitting for the Cd(II)-SeU Methanol Solutions S1 and S5 at 200 K (C_Cd(II) ~0.1 mol dm^-3; see Figure 3). ^a

Solution (SeU/Cd(II)	Cd-O			Cd-Se
	N	R (Å)	σ2 (Å2)	N	R (Å)	σ2 (Å2)	ℛb
S1 (1.0)	5 f	2.29	0.0066	1 f	2.62	0.0036	18.0
S5 (5.0)				3.8	2.63	0.0043	24.4

^aFitting k-range = 3 - 14 Å^-1; S₀² = 1.0 f (f = fixed); estimated error limits: N ± 20 %, R ± 0.02 Å, σ² ± 0.001 Ų.

^b

Residual; the residual (%) from the least-squares curve fitting is defined as:

$$\frac{\sum _{i=1}^N\left|y_{\exp }(i)­y_{\mathrm{theo}}(i)\right|}{\sum _{i=1}^N\left|y_{\exp }(i)\right|}\times 100$$

where y_exp and y_theo are k³-weighted experimental and theoretical data points, respectively.

Table 5.

Structural Parameters Derived from EXAFS Least-Squares Curve Fitting of Models for the Cd(II)-TU Methanol Solutions T1, T2, T4 - T7 at 200 K (LT) and 298 K (RT); (C_Cd(II) ~0.1 mol dm^-3; see Figure 4). ^a

Solution (TU / Cd(II) mole ratio)	113Cd δ (ppm)	Cd-S			Cd-O
		N	R (Å)	σ2 (Å2)	N	R (Å)	σ2 (Å2)	ℛb
[Cd(TU)4](NO3)2 solid	577 (δiso)	4 f	2.53	0.0063				15.4
Temp. 200 K
T1 (1.0)	-21, 93 (major), 189	1 f	2.54	0.0030	5 f	2.30	0.0047	17.4
T2 (2.0)	94, 191 (major), 337	2 f	2.55	0.0056	4 f	2.33	0.0062	14.3
T4 (4.0)	515 (broad)	3.8	2.53	0.0052				13.9
T5 (5.0)	506	3.8	2.54	0.0057				13.4
T6 (8.0)	458	3.3	2.54	0.0066				22.2
T7 (10.0)	423	3.8	2.55	0.0087				20.6
Temp. 298 K
T1 (1.0)		1 f	2.54	0.0063	5 f	2.30	0.0075	14.2
T2 (2.0)		2 f	2.51	0.0066	4 f	2.34	0.0132	18.2
T4 (4.0)	503	3.3	2.52	0.0060				8.1
T5 (5.0)	534	3.3	2.53	0.0057				10.7
T6 (8.0)	541	3.3	2.53	0.0053				13.9
T7 (10.0)	533	3.5	2.53	0.0063				10.1

^ak-range for fitting 3 - 14 Å^-1; S₀² = 1.0 f (f = fixed) for solutions; estimated error limits: N ± 20 %, for Cd-S R ± 0.02 Å, σ² ± 0.001 Ų; for Cd-O R ± 0.04 Å and σ² ± 0.002 Ų;

^bresidual.

To the experimental LT EXAFS spectra of solutions T6 and T7 linear combinations of simulated EXAFS oscillations for CdS₄ and CdS₆ coordination were fitted in the k range 3.4 – 14.4 Å^-1, following a procedure explained elsewhere,⁷⁶ with the S₀² and ΔE₀ values fixed at 1.0 and 1.0 eV, respectively. The simulations were performed for CdS₄ by varying the average Cd-S distance in the range 2.53 – 2.54 Å (σ² = 0.005 – 0.007 Ų) and for CdS₆ 2.68 – 2.74 Å (σ² = 0.006 – 0.007 Ų). The best fits were obtained with the following parameters: CdS₄ (4 Cd-S, R = 2.54 Å, σ² = 0.005 Ų) and CdS₆ (6 Cd-S, R = 2.71 Å, σ² = 0.007 Ų).

Results

¹¹³Cd NMR spectroscopy

¹¹³Cd NMR spectra were measured at 203 K of 0.1 mol dm^-3 dehydrated Cd(CF₃SO₃)₂ in methanol solutions (30% CD₃OD) containing different concentrations of thiourea (T1 – T7) or selenourea (S1 – S5) (see Figure 1 left and middle). The spectrum of Cd(CF₃SO₃)₂ in methanol without added ligands (L = TU, SeU) showed a signal at -21 ppm. In the solutions S1 and T1, both with the mole ratio L/Cd(II) = 1.0, several sharp signals with similar ¹¹³Cd chemical shifts appeared. A small peak at -21 ppm showed that at this mole ratio some Cd(II) ions still remained free, however, at increasing ligand concentration this signal disappeared. For the methanol solutions S2 – S3 and T2 – T3 with L/Cd(II) mole ratios 2.0 and 3.0 (L = TU, SeU), respectively, up to five distinct ¹¹³Cd NMR signals appeared indicating four Cd(II) species with different coordination environments in slow ligand-exchange equilibria, beside the solvated Cd²⁺ ion.

The first two sharp signals appeared with very similar chemical shifts: 94 and 191 ppm for T2 and T3, versus 97 and 194 ppm for S2 and S3, while the next two signals were broader and more shielded for the Cd(II)-thiourea solutions T2 and T3 (333 and 526 ppm) than for the Cd(II)-selenourea solutions S2 and S3 (435 and 578 ppm). For solutions S4 and S5 only one single sharp signal at 578 ppm remained, corresponding to a single dominating Cd(II)-selenourea complex. Also the Cd(II) thiourea solutions T4 to T7 displayed a single, relatively broad peak. However, the gradual increase in shielding from 515 ppm in T4 to 423 ppm in T7 with increasing thiourea concentration can be attributed to a gradual change in the coordination environment around Cd(II).

The ¹¹³Cd NMR spectra obtained for the thiourea solutions T4 – T7 at room temperature are shown in Figure 1 (right). No signal could be obtained for T1 – T3, since the ¹¹³Cd NMR signals became broader and finally disappeared at 223 K. For solutions T4 – T7 the single ¹¹³Cd NMR signal became considerably sharper and mostly less shielded at RT, which means that either a single dominating Cd(II) thiourea complex is present, or that the Cd(II)-thiourea species are in fast ligand exchange.

To investigate the influence of the solvent on the Cd(II)- thiourea complex formation, ¹¹³Cd NMR spectra were measured of a series of aqueous 0.1 mol dm^-3 Cd(CF₃SO₃)₂ solutions with different TU / Cd(II) mole ratios (Figure 2). All spectra showed a single sharp signal, which with increasing ligand concentration gradually became less shielded, from δ(¹¹³Cd) = 1 ppm for 0.1 mol dm^-3 Cd(CF₃SO₃)₂ in aqueous solution, 60 ppm with 0.1 mol dm^-3 thiourea in W1, to 446 ppm for 2.0 mol dm^-3 thiourea in W9 (mole ratio TU / Cd(II) = 20), indicating that S ligand atoms gradually replace O atoms in the Cd(II) coordination environment.

Figure 2.

¹¹³Cd NMR spectra of 0.1 mol dm^-3 Cd(CF₃SO₃)₂ aqueous solutions (W1 – W9) with thiourea mole ratios TU / Cd(II) increasing from 1.0 to 20.0, measured at 295 K; see Table 3. The spectrum at the top is for a solution without added thiourea.

X-ray absorption spectroscopy - Cd K-edge EXAFS

Cd K-edge EXAFS spectra were measured at 200 K for the methanol solutions S1 and S5 with the SeU / Cd(II) mole ratios 1.0 and 5.0, respectively (Figure 3); the curve-fitting results are shown in Table 4. Solutions S2, S3 and S4 were not measured, since the ¹¹³Cd NMR measurements showed that solutions S2 - S3 contained mixtures of Cd(II)-SeU complexes and that the speciation was similar in solution S4 as in S5. Also solution S1 contained several Cd(II)-SeU complexes, with the dominating one giving a signal at δ(¹¹³Cd) = 96 ppm (Figure 1, left). The EXAFS oscillation of this solution was well fitted with a model containing one SeU ligand and five oxygen donors from solvent (methanol) molecules and possibly also the CF₃SO₃^- anion, with average Cd-Se and Cd-O bond distances of 2.62 ± 0.02 Å and 2.29 ± 0.02 Å, respectively. This model is supported by a calculation of the average number of bound SeU per Cd(II) ion (= 1.01) from the ¹¹³Cd NMR peak integrals, which correspond to the fraction of each Cd(II) complex in solution S1 (Appendix 1, Supporting Information). The EXAFS oscillation of solution S5 with considerably higher amplitude than that of S1 was well modeled by four Cd-Se paths at 2.63 ± 0.02 Å (Table 4).

Figure 3.

Cd K-edge EXAFS curve-fitting for the Cd(II)-selenourea methanol solutions S1 and S5 with SeU/ Cd(II) mole ratios 1.0 and 5.0 (C_Cd(II) = 0.1 mol dm^-3 measured at 200 K (see Table 4)

For the series of Cd(II) thiourea solutions T1 – T7 (excluding T3) in MeOH, Cd K-edge EXAFS spectra were measured at both 200 K (LT) and 298 K (RT), as shown in Figure 4. Based on the ¹¹³Cd NMR spectrum of solution T1 at 203 K, a mono(thiourea) cadmium(II) complex with δ = 93 ppm dominates in this solution, together with a minor amount of solvated Cd²⁺ ions and Cd(TU)₂²⁺ species (see below). Least squares curve-fitting to the k³-weighted EXAFS oscillation of T1 at LT with a model consisting of one Cd-S and five Cd-O scattering paths resulted in the mean distances 2.54 ± 0.02 Å and 2.30 ± 0.04 Å, respectively (Table 5). The number of bound TU / Cd(II) (= 1.06) from the peak integrals in the ¹¹³Cd NMR spectrum of solution T1 (Figure 1, middle) supports this model; see Appendix 2 in Supporting Information.

Figure 4.

Cd K-edge EXAFS (black lines) and curve-fitted model (red dashes) oscillations and corresponding Fourier-transforms for Cd(II)-thiourea complexes in methanol solutions (C_Cd(II) = 0.1 mol dm^-3) with the mole ratios TU/ Cd(II) 1.0 (T1), 2.0 (T2), 4.0 (T4), 5.0 (T5), 8.0 (T6) and 10.0 (T7) measured at 200 K (LT) and 298 K (RT); see Table 5.

When increasing the thiourea concentration to 0.2 mol dm^-3 in solution T2 (mole ratio TU / Cd(II) = 2.0), the ¹¹³Cd NMR spectrum indicated two major Cd(II) thiourea complexes with δ(¹¹³Cd) = 94 and 191 ppm beside a minor peak at 337 ppm (Figure 1). The FT peak position for the LT EXAFS spectrum for this solution shifted slightly to a longer distance and gained intensity (Figure 4). The LT EXAFS oscillation of this solution could be modeled with 2 Cd-S (2.55 ± 0.02 Å) and 4 Cd-O (2.33 ± 0.04 Å) scattering paths (Table 5). The peak integrals in the ¹¹³Cd NMR spectrum of solution T2 (Figure 1, middle) correspond to 2.17 bound TU / Cd(II) (see Appendix 2, Supporting Information). This ratio is higher than the stoichiometric TU / Cd(II) ratio (2.0), indicating an error of about 10% in the number estimated from the peak integrals. The EXAFS oscillation for solution T3 was not analyzed, as it contained several Cd(II) thiourea species beside the dominating one with δ(¹¹³Cd) = 191 ppm (Figure 1).

The highest FT amplitude in the LT EXAFS series T1 – T7 was obtained for solutions T4 and T5 with δ(¹¹³Cd) = 515 and 506 ppm, respectively. The LT EXAFS oscillations of these solutions were well modeled with four Cd-S distances at 2.53 – 2.54 Å. Attempts to include a Cd-O path resulted in too short Cd-O bond distances and/or unreasonable disorder parameters. With further increase in the thiourea concentration, from 0.5 mol dm^-3 in T5 to 1.0 mol dm^-3 in T7, the single ¹¹³Cd NMR signal shifted up-field from δ(¹¹³Cd) = 506 ppm to 423 ppm (Figure 1), with a considerable amplitude reduction of the LT EXAFS oscillations, as shown in Figure S-5 for solutions T4 and T7 with TU/Cd(II) mole ratios 4.0 and 10.0, respectively. This figure also shows that at 298 K the EXAFS spectra of these solutions (T4 and T7) overlap, which is consistent with their similar sharp ¹¹³Cd NMR peaks at 503 and 533 ppm, respectively (Figure 1, right), indicating that at room temperature the difference in speciation is not large enough to be distinguished by EXAFS spectroscopy.

The RT EXAFS oscillations for solutions T4 – T7 were well modeled with four Cd-S paths at 2.53 ± 0.02 Å (Table 5). A comparison of the EXAFS oscillations measured for solution T4 at two different temperatures (Figure S-6) shows similar frequencies, reflected in their similar bond distances (2.52 – 2.53 Å). However, the amplitude is higher at LT, resulting in slightly higher N_Cd-S (with fixed S₀²), and smaller σ² values in the refinements. For solutions T1 and T2, for which no ¹¹³Cd NMR signal could be observed at RT, the EXAFS spectra had smaller amplitude at RT than at LT; see Figure S-6. For solution T2 the LT EXAFS oscillation had higher frequency, showing a significant shift to longer Cd-S bond distances from 2.51 Å at RT to 2.55 Å at LT, while the change in temperature did not influence the mean Cd-S and Cd-O bond distances for solution T1 (Table 5).

Cd K-edge EXAFS spectra for the Cd(II) thiourea aqueous solutions W1 – W7 were also measured and analyzed assuming mixtures of hydrated Cd(TU)_x²⁺ (x = 1 - 4) species; see Figure S-1 and Table S-2.

X-ray absorption spectroscopy - Cd L₃ -edge XANES

Cd L₃-edge XANES spectroscopy has been shown to be sensitive to the type of coordinating atoms and the local coordination geometry around the cadmium(II) ion in its complexes. For example, Cd(II) complexes with CdO₆ octahedral geometry show a distinct pre-edge feature at 3539.1 eV, originating from a core electron transition from Cd 2p_3/2 to higher unoccupied molecular orbitals with s or d character.^77,78,79 Therefore, Cd L₃-edge XANES spectra were measured at room temperature of the Cd(II) thiourea solutions in methanol (T1 – T7) including Cd(CF₃SO₃)₂ (model for CdO₆), and for crystalline compounds with different CdS_xO_y coordination geometries (Figure 5). The XANES spectrum of solution T1 with a distinct pre-edge feature at 3539.1 eV resembles that of CdO₆, while the second derivative of the spectral features for T2 is similar to those of CdS₂O₄. For solutions T4 – T7 the Cd L₃-edge XANES features and corresponding 2^nd derivatives are very similar, and comparable to the CdS₄ model.

Figure 5.

Cd L₃-edge XANES spectra at RT of the Cd(CF₃SO₃)₂ -thiourea solutions in methanol T1 – T7, compared with Cd(CF₃SO₃)₂ in methanol (CdO₆) and crystalline standard compounds with CdS_xO_y coordination:⁷⁹ [Cd(tmtu)₂(NO₃)₂] (CdS₂O₄ model; tmtu = tetramethylthiourea),⁸⁰ [Cd(DMTF)₄(CF₃SO₃)₂] (CdS₄O₂ model),⁴⁸ [Cd(DMTF)₆](ClO₄)₂ (CdS₆ model),⁴⁸ cadmium adamantane cage (Et₃NH)₄[S₄Cd₁₀(SPh)₁₆] (CdS₄ model),⁵³ and (HIm)[Cd(tsac)₃(H₂O)] (CdS₃O model; tsac = thiosaccharinato).⁴⁹ The dashed lines are at 3539.1 and 3541.6 eV.

Discussion

Selenourea Cd(II) complex formation

The combined results from the ¹¹³Cd NMR and Cd K-edge EXAFS spectroscopic studies show a dominating mono-selenourea Cd(SeU)²⁺ complex with CdSeO₅ coordination in the methanol solution S1 with mole ratio SeU/ Cd(II) = 1.0 (see Figures 1, 3 and Table 4). The O-donor ligands are methanol molecules and triflate ions, as in [Cd(DMTF)₄(CF₃SO₃)₂].⁴⁸ For the mole ratios SeU / Cd(II) = 4.0 and 5.0 (solutions S4 and S5, respectively) a single cadmium(II)-selenourea complex [Cd(SeU)₄]²⁺ with δ(¹¹³Cd) = 578 ppm dominates. The mean Cd-Se bond distance of 2.63 ± 0.02 Å (Table 4) is compatible with the average Cd-Se bond distances of 2.649 Å in [N(CH₃)₄]₂[Cd(SePh)₄],⁸¹ and 2.623 Å in [Cd(dmise)₄][PF₆]₂, the only crystal structures with homoleptic CdSe₄ coordination geometry.⁷⁴ The synthesis of [Cd(SeU)₄]SO₄ was reported in 1972 without structural information.⁸² The ¹¹³Cd NMR chemical shift 541 ppm for the [Cd(SePh)₄]²⁺ complex in methanol solution at 213 K,⁵⁹ is slightly more shielded than that reported here of the [Cd(SeU)₄]²⁺ complex at 203 K (Δδ = 578 – 541 = 37 ppm).

The slow ligand-exchange for the methanol solutions S1, S2 and S3 at 203 K allows the peaks observed at δ(¹¹³Cd) = 96-97, 193-194, 435 and 578 ppm (Figure 1), to be assigned to complexes with CdSeO₅, CdSe₂O₄, CdSe₃O and CdSe₄ coordination, respectively. The relatively broad peaks prevented ¹¹³Cd-⁷⁷Se coupling to be observed; such one-bond coupling normally is within 50 – 200 MHz.⁶⁴ Previously, a ¹¹³Cd chemical shift of 250 ppm was obtained for a cadmium(II) selenocarboxylate complex with trigonal bipyramidal CdSe₃O₂ coordination geometry in CH₂Cl₂ at 193 K.⁶⁴ A CSD survey did not show any crystalline compound with CdSe₂O₄ coordination, only one Cd(II) complex with CdSe₂O₃ coordination,⁸³ and one monomeric CdSe₃O complex with the Cd-Se bond distances within 2.590 – 2.603 Å and Cd-OH₂ at 2.307 Å.⁸⁴

By assuming that the integrated area of a distinct ¹¹³Cd NMR peak in Figure 1 is proportional to the concentration of a specific Cd(II) species in each cadmium(II) – selenourea (S1–S3) solution, the relative fractions (%Cd) and the formation constants of the Cd(II) species were evaluated (Table S-1b, Supporting Information). The formation constants (log β_x, x = 1 - 4) of the Cd(SeU)_x²⁺ species above were varied systematically until the distribution of complexes calculated in fraction diagrams approximately agreed with those obtained from the ¹¹³Cd NMR spectra; see Figure S-8 (top) in Appendix 1, Supporting Information. The evaluation showed that the log β_x values are correlated with an estimated error limit of ±0.5 logarithmic units (Table 6).

Table 6.

Formation constants ^a of the Cd(II) selenourea / thiourea species CdL_x²⁺ (L = SeU / TU, x = 1-4) formed in methanol solutions, derived from ¹¹³Cd NMR spectra at 203 K (see Figures 1, S-8 and S-9).

	log βx (SeU)	log βx (TU)
CdL2+	4.4	4.3
CdL22+	7.3	7.3
CdL32+	10.7	9.7
CdL42+	14.0	12.4

^aβ_x = [CdL_x²⁺]/([Cd²⁺][L]^x)

Thiourea Cd(II) complex formation in MeOH

Low temperature (~200 K)

In the ¹¹³Cd NMR spectra of the methanol solutions with TU / Cd(II) mole ratios 1.0, 2.0 and 3.0 (T1, T2 and T3), four peaks appeared at 203 K with δ = 93 – 94, 189 – 191, 333 – 337, and 526 ppm (Figure 1). A similar observation was reported for methanol solutions of Cd(ClO₄)₂ and 1,3-thiazolidine-2-thione (tztH) at 189 K with four peaks at -14, 73, 153, 453 ppm, which were tentatively assigned to solvated [Cd(tztH)_w]²⁺ complexes with w = 0 - 3, respectively.⁴⁶ For solution T1, the peak at 93 ppm has the highest intensity, and its Cd K-edge EXAFS spectrum is well modeled with one Cd-S and five Cd-O bond distances (2.54 ± 0.02 and 2.30 ± 0.04 Å, respectively; see Figure 4 and Table 5), showing that a mono(thiourea) Cd(II) complex with CdSO₅ coordination (O from MeOH or CF₃SO₃^-) dominates. For the methanol solution T2 with C_Cd(II) = 0.1 mol dm^-3 and C_TU = 0.2 mol dm^-3, the ¹¹³Cd NMR signal at 191 ppm is the most intense (Figure 1). The average Cd-S and Cd-O bond distances 2.55 ± 0.02 Å and 2.33 ± 0.04 Å (Table 5) are comparable with the average Cd-S (2.585 Å) and Cd-O (2.366 Å) distances for the cis-[Cd(TU)₂(H₂O)₄](SO₄) complex,⁵² for which an isotropic solid state ¹¹³Cd NMR chemical shift of 148 ppm was observed (Figure S-4b). We conclude that a bis(thiourea)cadmium(II) complex, Cd(TU)₂²⁺, with a CdS₂O₄ coordination environment dominates in solution T2.

In solution T3 (TU/ Cd(II) mole ratio = 3.0), several species are in equilibrium giving rise to two sharp ¹¹³Cd NMR signals at 94 ppm (CdSO₅) and 191 ppm (CdS₂O₄), and two broad peaks at 333 ppm and 526 ppm (Figure 1). The reported isotropic ¹¹³Cd NMR chemical shifts for complexes with CdS₃O₂ and CdS₃O₃ coordination (S = thiolate) vary within the range δ_iso = 390 – 410 ppm.^85,86 Considering that the (HIm)[Cd(tsac)₃(H₂O)] and [Cd(TU)₃(SO₄)] complexes with CdS₃O coordination, showed isotropic solid state ¹¹³Cd NMR chemical shifts at 378 and 346 ppm at room temperature, respectively (Figure S-4a, b), we can assign the peak at 333 ppm to tris(thiourea) Cd(II) complexes with CdS₃O_1-3 coordination in relatively fast ligand-exchange that broadens the peak. The broad peak at 526 ppm in the NMR spectrum of solution T3 can be assigned to a tetra(thiourea) Cd(II) complex mainly with CdS₄ coordination (see Room temperature section).

Similar to the Cd(II)-selenourea system, the integrated peak areas in the ¹¹³Cd NMR spectra of the methanol solutions T1 – T3 at 203 K (Figure 1) were used to estimate the concentration of Cd(TU)_y²⁺ (y = 1-4) species in each solution (Appendix 2, Supporting Information). The formation constants log β_y were varied until the calculated fractions of Cd(TU)_y²⁺ species in Figure S-9 (top) closely matched the fractions obtained from the peak areas (Table 6). The resulting β_y values obtained for the solvated Cd(TU)_y²⁺ (y = 1 – 4) species in methanol at 203 K (β₁ = 2.0·10⁴, β₂ = 2.0·10⁷, β₃ = 5.0·10⁹ and β₄ = 2.5·10¹²) differ considerably from those previously reported from a polarographic study for similar complexes in aqueous solution at 298 K (β₁ = 20, β₂ = 140, β₃ = 260 and β₄ = 1200).¹⁶

By increasing the thiourea concentration from 0.4 mol dm^-3 in T4 to 1.0 mol dm^-3 in T7 (C_Cd(II) = 0.1 mol dm^-3), the single ¹¹³Cd NMR peak became more shielded, shifting from 515 ppm in T4 to 423 ppm in T7. The latter value can be compared with δ_iso(¹¹³Cd) = 402 ppm recorded for crystalline [Cd(DMTF)₆](ClO₄)₂ with octahedral CdS₆ coordination (see Figure S-3). Therefore, the gradual shift of the NMR signal from 515 to 423 ppm is probably due to partial formation of Cd(II) complexes with a higher number of coordinated thiourea ligands, e.g. [Cd(TU)₅]²⁺ (CdS₅) and octahedral [Cd(TU)₆]²⁺ (CdS₆) species. Even though solution T7 has the highest thiourea concentration, the amplitude of its LT EXAFS oscillation is visibly smaller than that for T4 (Figure S-5 left).

An attempt to evaluate the speciation of the higher Cd(II)-thiourea complexes was made by fitting the LT EXAFS spectra for solutions T6 and T7 (TU / Cd(II) mole ratios 8.0 and 10.0, respectively) with linear combinations of EXAFS oscillations simulated for CdS₄ and CdS₆ coordination. The best fits shown in Figure 6 were obtained with the mean Cd-S distance 2.54 Å (σ² = 0.005 Ų) for CdS₄, and 2.71 Å (σ² = 0.007 Ų) for CdS₆, which are comparable with the mean Cd-S distance in crystalline [Cd(TU)₄](NO₃)₂ with CdS₄ coordination (2.53 ± 0.02 Å from EXAFS; Table 5 and Figure S-7), and 2.711 Å in a crystalline [Cd(TU)₆]²⁺ complex with CdS₆ coordination.¹⁵ Note that while the average of the crystallographic Cd-S distances in [Cd(TU)₄](NO₃)₂ is 2.560 Å,¹³ the slightly shorter EXAFS value 2.53 ± 0.02 Å could result from a higher contribution from its short, more well-defined bond distances to the EXAFS oscillation. Ignoring a possible contribution from Cd(TU)₅²⁺ species, the linear combination fitting results in 77% contribution from CdS₄ and 23% from CdS₆ to the EXAFS oscillation for solution T6; for solution T7 the contributions are 68% CdS₄ and 32% CdS₆ (estimated error ± 10%). Figure 6 shows that the amplitude of the experimental EXAFS spectra is reduced by the opposing phases of the EXAFS oscillations for CdS₄ and CdS₆ coordination. Such a change in the Cd(II) speciation as in these solutions would have been difficult to interpret solely based on EXAFS spectroscopy, without the aid from the ¹¹³Cd NMR spectra.

Figure 6.

Cd K-edge EXAFS spectra at LT for the Cd(II) thiourea methanol solutions T6 and T7 fitted with linear combinations of simulated EXAFS oscillations for CdS₄ and CdS₆ coordination, indicating more than four thiourea ligands in ~ 20 – 30% of the cadmium(II) complexes.

Room temperature

For the cadmium(II) thiourea solutions T1 – T3, for which no ¹¹³Cd NMR signal could be observed at 295 K, the Cd L₃-edge XANES spectral features were useful for indicating the coordination of the major cadmium(II) species. The Cd L₃-edge XANES spectrum of T1 resembles that of Cd(CF₃SO₃)₂ in methanol (CdO₆ model) in Figure 5. At LT the ¹¹³Cd NMR spectrum showed that some free solvated Cd²⁺ ions were present (Figure 1). The RT EXAFS spectrum of solution T1 could be fitted with a CdSO₅ model with the average Cd-O and Cd-S distances 2.30 ± 0.02 Å and 2.54 ± 0.02 Å, respectively, which are comparable to those obtained for this solution at LT (see Table 5 and Figure S-6). The resemblance between the Cd L₃-edge XANES spectra of solution T2 and the CdS₂O₄ model in Figure 5, and also the corresponding second derivatives, implies that solution T2 mainly contains the bis(thiourea)cadmium(II) complex as the dominating species. Fitting the RT Cd K-edge EXAFS spectrum of solution T2 with a CdS₂O₄ model resulted in a somewhat shorter average Cd-S distance 2.51 ± 0.02 Å.

The ¹¹³Cd NMR spectra of solutions T4 – T7 with TU / Cd(II) mole ratios 4.0 – 10.0 showed at 295 K a single sharp peak in the 503 – 541 ppm region. The RT EXAFS spectra of these solutions almost overlap (Figure S-5, right), and were well fitted with a CdS₄ model with an average Cd-S distance of 2.52 – 2.53 Å. The second derivatives of their Cd L₃-edge absorption spectra showed two minima, rather similar to those observed for the CdS₄ model compound (Figure 5). Thus, at high thiourea concentration (TU / Cd(II) > 4.0) the [Cd(TU)₄]²⁺ complex dominates at room temperature. Reducing the temperature of these solutions (T5 – T7) to ~200 K promotes the formation of CdS_5-6 complexes, resulting in longer average Cd-S distances (up to 2.55 Å for T7) and more shielded ¹¹³Cd chemical shifts (423 ppm for T7).

For the crystalline [Cd(TU)₄](NO₃)₂ complex with an average Cd-S distance of 2.53 ± 0.02 Å (from EXAFS; see Table 5 and Figure S-7), a solid state ¹¹³Cd NMR isotropic chemical shift of δ_iso = 577 ppm was obtained (Figure S-3), while the ¹¹³Cd NMR chemical shift for [Cd(TU)₄]²⁺ ions in methanol solution is more shielded (541 ppm at 295 K). A similar up-field shift has been reported for [Cd(Se-2,4,6-ⁱPr₃-C₆H₂)₂(bpy)] complex (δ_iso = 460, 472) when dissolved in chloroform (411 ppm),⁶¹ and was explained as a “solvent effect” (see Table 2).

Based on the above discussion, the broad peak at 526 ppm in the LT ¹¹³Cd NMR spectrum of solution T3 can be assigned to [Cd(TU)₄]²⁺ species with rather slow ligand-exchange on the NMR scale. The broadness of this peak could be due to a minor amount of a Cd(II) tetra(thiourea) complex with CdS₄O coordination. Higher O-donor ligand coordination is not likely, since the chemical shift becomes more shielded, e.g. for [Cd(DMTF)₄(CF₃SO₃)₂] with CdS₄O₂ coordination, δ_iso (¹¹³Cd) = 267 ppm (Table 1). The isotropic chemical shifts for CdS₄ and CdS₄O coordination (S = thiolate) in the [Cd₁₀(SCH₂CH₂OH)₁₆](ClO₄)₄ compound were reported at 674 and 503 ppm, respectively.^87,* The ¹¹³Cd NMR chemical shifts measured at room temperature for the [Cd(TU)₄]²⁺ complex (CdS₄ coordination) in the solid state (δ_iso = 577 ppm) and in methanol solution (541 ppm) are more shielded than that for mononuclear Cd(II)-tetrathiolate complexes (~600 - 750 ppm).⁴³

Thiourea Cd(II) complex formation in H₂O

The cadmium(II) complex formation with thiourea in aqueous solution was also investigated at room temperature, keeping the same cadmium(II) concentration as for the methanol solutions (~0.1 mol dm^-3 Cd(CF₃SO₃)₂), with the TU concentration varying from 0.1 to 2.0 mol dm^-3 (W1 – W9, Table 3). The single sharp peak with a gradual downfield shift from 60 ppm (W1) to 446 ppm (W9) in the ¹¹³Cd NMR spectra shown in Figure 2 corresponds to cadmium(II) complexes with an increasing number of coordinated thiourea ligands in fast ligand exchange, as the TU concentration increases. A similar trend is also observed for the isotropic ¹¹³Cd NMR chemical shifts with δ_iso = 148, 346 and 577 ppm of the cis-[Cd(TU)₂(H₂O)₄](SO₄), [Cd(TU)₃(SO₄)] and [Cd(TU)₄](NO₃)₂ solid compounds, respectively (Table 1).

The Cd K-edge EXAFS spectra for the aqueous solutions W1 – W7 were analyzed (Table S-2), keeping the average Cd-S and Cd-O coordination numbers fixed at values derived from the calculated distribution of complexes assuming [Cd(TU)_y(H₂O)_6-y]²⁺ (y = 0 – 3) and [Cd(TU)₄]²⁺ species (Figure S-1); the results showed a mean Cd-S distance of about 2.54 - 2.57 Å for solutions W1 to W5. The mean Cd-S distance became slightly shorter, about 2.53 Å, for solutions W6 and W7 (TU / Cd mole ratios 8.0 and 10.0, respectively), while the mean of the Cd-O distances is around 2.28 ± 0.04 Å for all solutions (Table S-2). This trend corresponds to that expected for hydrated [Cd(TU)(H₂O)₅]²⁺ and [Cd(TU)₂(H₂O)₄]²⁺ complexes in a mixture, with a gradually increasing amount of [Cd(TU)₄]²⁺ complexes without coordinated water (Figure S-1). The crystal structure of a hydrated cadmium thiourea sulfate showed both cis-[Cd(TU)₂(H₂O)₄](SO₄) and fac-[Cd(TU)₃(H₂O)₃](SO₄) complexes co-crystallized in the unit cell, with the mean Cd-S (Cd-O) bond distances 2.59 (2.37) Å and 2.62 (2.41) Å, respectively.⁵² A mean Cd-S distance of 2.53 ± 0.02 Å is expected for a [Cd(TU)₄]²⁺ complex from EXAFS, as obtained for the solid [Cd(TU)₄](NO₃)₂ compound (Table 5 and Figure S-7).

Therefore, we interpret the combined results of the aqueous solutions W1 – W5 (C_Cd(II) = 0.1 mol dm^-3 and TU/ Cd(II) = 1.0 – 5.0) as corresponding to dominating hydrated [Cd(TU)(H₂O)₅]²⁺ and [Cd(TU)₂(H₂O)₄]²⁺ complexes that are in fast ligand exchange on the NMR time scale. By further increasing the thiourea concentration in solutions W6 to W9 (mole ratios TU / Cd(II) = 8.0 to 20.0, respectively) the tetrahedral [Cd(TU)₄]²⁺ complex becomes the dominant species, shifting the ¹¹³Cd NMR signal downfield. When compared to the ¹¹³Cd NMR spectra of the methanol solutions T4 – T7 measured at room temperature (Figure 1, right), it is evident that the [Cd(TU)₄]²⁺ complex dominates already at the mole ratio TU / Cd(II) = 4.0 in methanol, which is a much weaker ligand than water. We are unable to ascertain the degree of hydration m in the [Cd(TU)₃(H₂O)_m]²⁺ complex, which is a minor species in solution according to the stability constants (Figure S-1), but propose that the [Cd(TU)₃(H₂O)₃]²⁺ complex is formed, as in the crystal structure.⁵²

Comparison between Cd(II) complex formation with selenourea and thiourea in MeOH at ~200 K

The ¹¹³Cd NMR spectra of 0.1 mol dm^-3 Cd(CF₃SO₃)₂ solutions (Figure 1) in methanol containing 0.1 – 0.3 mol dm^-3 selenourea (S1 – S3) or thiourea (T1 – T3) at 203 K show similar peaks in the 93 – 97 ppm and 189 – 194 ppm range, which were assigned to complexes with Cd(Se/S)O₅ and Cd(Se/S)₂O₄ coordination geometries, respectively. In this case the chemical shifts are very similar when one or two S / Se-donor ligands replace the O-donor ligands (methanol and CF₃SO₃^- in the current case), in contrast to the generally observed trend with decreasing shielding of donor atoms in the order: O ≫ Se > S.⁴³ Also the formation constants for the CdL²⁺ and CdL₂²⁺ species are virtually the same for the SeU and TU systems, while Cd(SeU)₃²⁺ and Cd(SeU)₄²⁺ are more stable than the corresponding Cd(TU)₃²⁺ and Cd(TU)₄²⁺ complexes. The tris(thiourea)cadmium(II) complexes, Cd(TU)₃²⁺, with CdS₃O_1-3 coordination give rise to broad peaks at 333 - 337 ppm in the ¹¹³Cd NMR spectra of the T2 – T3 solutions, while the tris(selenourea)cadmium(II) complexes with tetrahedral CdSe₃O coordination correspond to the relatively sharp peak at 435 ppm, as for solutions S1 – S3. In the methanol solutions S4 and S5, containing 0.4 – 0.5 mol dm^-3 selenourea, respectively, a [Cd(SeU)₄]²⁺ complex with a CdSe₄ core is predominantly present. Its chemical shift δ(¹¹³Cd) = 578 ppm is comparable to δ_iso = 577 ppm observed for CdS₄ coordination in the solid state ¹¹³Cd NMR spectrum of [Cd(TU)₄](NO₃)₂ (Figure S-3), and is more deshielded than that of [Cd(TU)₄]²⁺ (CdS4, 526 ppm) in methanol solution at 203 K. Likewise, Isab and coworkers reported a more deshielded solid state ¹¹³Cd NMR isotropic chemical shift for Cd(SeU)₂Cl₂ (458 ppm) than that for Cd(TU)₂Cl₂ (430 ppm).³⁶ The rather broad peak at 515 ppm for solution T4, containing 0.4 mol dm^-3 thiourea, indicates that the [Cd(TU)₄]²⁺ complex dominates; however, by increasing the ligand concentration higher complexes with CdS_5-6 coordination are formed.

Conclusion

In the current study, we have made a detailed evaluation of the speciation of cadmium(II) complexes with selenourea and thiourea by varying the ligand concentration, solvent and temperature using a combination of spectroscopic methods, ¹¹³Cd NMR, Cd K-edge EXAFS and Cd L-edge XANES. In methanol solutions of 0.1 mol dm^-3 Cd(CF₃SO₃)₂ with mole ratios L / Cd(II) ≤ 3.0 (L = SeU, TU), ¹¹³Cd NMR spectroscopy shows mixtures of CdL_n²⁺ (n = 1 – 4) species in slow ligand exchange equilibria at 203 K, which allowed their stability constants to be estimated. In ligand excess tetrahedral [Cd(SeU)₄]²⁺ and [Cd(TU)₄]²⁺ complexes dominate at 203 K and 295 K, respectively, with the average bond distances Cd-Se 2.63 ± 0.02 Å and Cd-S 2.53 ± 0.02 Å, as determined by Cd K-edge EXAFS spectroscopy. When lowering the temperature from 295 K to 203 K for 0.1 mol dm^-3 Cd(CF₃SO₃)₂ methanol solutions with high mole ratios, TU / Cd(II) > 5.0, complexes with even higher thiourea coordination, [Cd(TU)_5-6]²⁺, form to some extent.

In aqueous solutions of 0.1 mol dm^-3 Cd(CF₃SO₃)₂ with increasing thiourea concentration the ¹¹³Cd NMR signal at 295 K corresponds to mixtures of [Cd(TU)_y(H₂O)_6-y]²⁺ (y = 0 – 3) species in fast ligand exchange equilibria with [Cd(TU)₄]²⁺ complexes. Higher excess of the thiourea ligand is needed than in comparable methanol solutions to form the [Cd(TU)₄]²⁺ complex, which becomes the major species for free concentrations of thiourea exceeding ~0.4 mol dm^-3.

Based on the ¹¹³Cd NMR chemical shifts observed in this study, Se (selone) and S (thione) donor atoms provide similar shielding effects in Cd(II) selenourea and thiourea complexes with similar coordination geometries. Almost identical δ(¹¹³Cd) chemical shifts were observed at 203 K for the octahedrally coordinated Cd(Se/S)O₅ (δ ~ 95 ppm) and Cd(Se/S)₂O₄ species (δ ~ 190 ppm), with MeOH and possibly CF₃SO₃^- as O-donor ligands. For the [Cd(SeU)₄]²⁺ complex (CdSe₄ coordination) the chemical shift recorded at 578 ppm (MeOH, 203 K) is somewhat more deshielded than that of [Cd(TU)₄]²⁺ (CdS₄ coordination, 526 ppm in MeOH at 203 K), although close to the isotropic chemical shift δ_iso (¹¹³Cd) = 577 ppm for crystalline [Cd(TU)₄](NO₃)₂ measured at room temperature (Table 1).

When reporting ¹¹³Cd NMR shifts the coordination geometry is well defined for crystalline Cd(II) complexes; however, in solution the structure will not necessarily stay the same. Our method using X-ray absorption spectroscopy as a supporting tool to further characterize the cadmium(II) species in solution significantly increases the insight in the systems studied. We have shown that ¹¹³Cd NMR spectroscopy can be used for quite detailed speciation at LT in methanol solution, where for example the EXAFS technique was unable on its own to ascertain the partial formation of [Cd(TU)_5-6]²⁺ complexes, as the overall increase in the Cd-S bond distances was ~ 0.02 Å, within the error limit of EXAFS spectroscopy. Conversely XAS spectroscopy was needed to structurally characterize the dominating species in the methanol solutions, especially to confirm that the Cd(II)-thiourea species with δ(¹¹³Cd) = 526 ppm (541 ppm at 295 K) indeed is the [Cd(TU)₄]²⁺ complex (CdS₄ coordination), which surprisingly has a more shielded chemical shift than [Cd(SeU)₄]²⁺ with δ(¹¹³Cd) = 578 ppm and CdSe₄ coordination.

We also show that thiourea, as a model for thione-containing ligands with neutral S-donor atoms, is more shielding than the thiolate groups with negatively charged S-donors, when comparing [Cd(TU)₄]²⁺ with δ(¹¹³Cd) = 541 ppm (in MeOH at 295 K) with the reported chemical shifts for Cd(II)-tetrathiolates (~ 600 - 750 ppm). Therefore, caution should be exercised when interpreting sulfur coordination in biological systems based on ¹¹³Cd NMR chemical shifts.

The individual chemical shifts obtained here are useful additions to the ¹¹³Cd NMR chemical shifts previously reported for Cd(II)-thione complexes (Table 1), as well as cadmium(II) complexes with Se-donor ligands (Table 2). The ¹¹³Cd NMR chemical shifts reported for cadmium(II) complexes with Se-donor ligands in general are altogether quite few,⁸⁸ as also the crystal structures when compared with S-donor ligands, probably due to the air-sensitivity of most organo-selenium compounds.