Infrared Spectroscopic and Theoretical Studies of Group 3 Metal

Abstract: A series of group 3 metal isocyanide complexes were prepared via the ... form M(NC)2 (M=Sc, Y, La) when the samples were subjected to λ>220...
1 downloads 0 Views 1MB Size
Download PDF
Article Cite This: J. Phys. Chem. A XXXX, XXX, XXX−XXX

pubs.acs.org/JPCA

Infrared Spectroscopic and Theoretical Studies of Group 3 Metal Isocyanide Molecules Xiuting Chen,†,‡ Qingnuan Li,† Lester Andrews,§ and Yu Gong*,†,§ †

Department of Radiochemistry, Shanghai Institute of Applied Physics, Shanghai 201800, China University of Chinese Academy of Sciences, Beijing 100049, China § Department of Chemistry, University of Virginia, Charlottesville, Virginia 22904-4319, United States ‡

J. Phys. Chem. A Downloaded from pubs.acs.org by UNIV OF SOUTH DAKOTA on 08/25/18. For personal use only.

S Supporting Information *

ABSTRACT: A series of group 3 metal isocyanide complexes were prepared via the reactions of laser ablated scandium, yttrium, and lanthanum atoms with (CN)2 in an argon matrix. The product structures were identified on the basis of their characteristic infrared absorptions from isotopically labeled (CN)2 samples as well as the calculated frequencies and isotopic frequency ratios. Group 3 metal atoms reacted with (CN)2 to form M(NC)2 (M = Sc, Y, La) when the samples were subjected to λ > 220 nm irradiation. Other products such as M(NC)3 and MNC were produced together with M(NC)2 through either the reactions of M(NC)2 and (CN)2 or the loss of one CN ligand from M(NC)2. CCSD(T)// B3LYP calculations reveal that ScNC possesses a 3Δ ground state, while 1Σ+ is most stable for YNC and LaNC. All of the M(NC)2 and M(NC)3 complexes were predicted to have doublet and singlet ground states, respectively. Group 3 metal cyanides are less stable than the isocyanides by at least 4 kcal/mol at the CCSD(T) level, and their C−N stretches are much weaker than the N−C stretches of the isocyanides. No absorption can be assigned to the M(CN)x complex, which would appear between 2100 and 2250 cm−1.

■

INTRODUCTION As the polyatomic analogue of halogens, cyanogen is composed of two CN groups connected by a C−C bond. Different from the transition metal halides with metal−halogen bonds, CN can be coordinated to the metal center with either carbon or nitrogen atom, which results in the formation of two linkage isomers, namely, cyanides and isocyanides.1 Compared with the well-known transition metal halides,2 neutral binary metal cyanides and isocyanides are far less understood and characterized. Theoretical calculations have indicated that the isocyanide isomers are more stable for early transition metals such as scandium and titanium while late transition metals such as cobalt, nickel, and copper prefer the cyanide arrangement.3 From the experimental point of view, several transition metal (Cr, Fe−Zn) monocyanides and isocyanides have been observed in the gas phase and their molecular structures as well as spectroscopic constants were determined by rotational spectroscopy.4−9 In a cryogenic matrix, mononuclear thorium, uranium, manganese, and iron isocyanides with different coordination numbers have been prepared via the reactions of laser ablated metal atoms and (CN)2, and their structures and formation mechanism were investigated by infrared spectroscopy and theoretical calculations.10−13 Dinuclear aluminum complexes in the form of Al2(NC)x (x = 1−3) containing rhombic ring core (AlN)2 were observed in a very recent study on the reactions of aluminum atoms and (CN)2.14 Compared with the metal cyanide/isocyanide systems studied before, no experimental result is available for the © XXXX American Chemical Society

simplest group 3 metal cyanide/isocyanide complexes. Only ScCN and ScNC have been the subject of density functional theory (DFT) and ab inito computations.3,15 In this paper, we report the formation of a series of scandium, yttrium, and lanthanum isocyanide molecules in argon matrix, and the structures of the new products were characterized by infrared spectroscopy. Theoretical calculations were employed to verify the experimental assignments, and understand the geometries and electronic structures of the new products as well as the unobserved group 3 metal cyanides.

■

EXPERIMENTAL AND THEORETICAL METHODS The reactions of laser-ablated group 3 metal atoms with (CN)2 in excess argon at 4 K were investigated by using matrix isolation infrared spectroscopy, and details of the experimental apparatus have been described previously.16,17 The Nd:YAG laser fundamental (Continuum II, 1064 nm, 10 Hz repetition rate with 10 ns pulse width) was focused onto a freshly cleaned scandium, yttrium, or lanthanum target mounted on a rotating rod. Laser-ablated metal atoms were codeposited with argon (research grade) containing 1% (CN)2 gas prepared in this laboratory. We employed thermal decomposition of AgCN at 360−380 °C until a constant pressure was reached in a stainless steel vacuum line following earlier work.10−12 The product gas was condensed at 77 K and evacuated before use. Received: July 16, 2018 Revised: August 11, 2018 Published: August 13, 2018 A

DOI: 10.1021/acs.jpca.8b06810 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

Figure 2 shows the infrared spectra from the reactions of laser ablated yttrium atoms and (CN)2 in solid argon. A broad

Isotopic reagents were synthesized by thermal decomposition of Ag13CN and AgC15N, which were prepared via the reactions of K13CN and KC15N (99% enriched, Cambridge Isotopic Laboratories) and silver nitrate. FTIR spectra were recorded on a Nicolet iS50 FTIR instrument at 0.5 cm−1 resolution with a HgCdTe range A detector. Matrix samples were annealed at different temperatures and cooled back to 4 K for spectral acquisition. Selected samples were subjected to full arc (λ > 220 nm) irradiation by a medium-pressure mercury arc lamp (Philips, 175W) with the outer globe removed. DFT calculations were performed using the Gaussian 09 program.18 The hybrid B3LYP functional was employed,19,20 and the 6-311+G(d) basis set was used for C, N, Sc.21−24 The 28-electron core SDD pseudopotential was used for Y and La.25,26 Harmonic vibrational frequencies at the B3LYP level were obtained analytically at the optimized structures, and zero-point vibrational energies were derived. The single-point energies of all the structures optimized at the B3LYP level of theory were obtained using the CCSD(T) method with the same basis sets.27 The T1 values for all the molecules are less than 0.035. The natural bond orbital (NBO) analysis was performed on the structures optimized at the B3LYP level of theory.28

Figure 2. Infrared spectra in the product absorption region from the reactions of laser-ablated yttrium atoms with (CN)2 in solid argon at 4 K. (a) Y and 1% (CN)2 codeposited for 1 h, (b) after annealing to 25 K, (c) after λ>220 nm irradiation, (d) after annealing to 30 K, (e) after annealing to 35 K. s denotes the site absorption of YNC.

■

RESULTS AND DISCUSSION Figure 1 shows the infrared spectra from the reactions of laserablated scandium atoms and (CN)2 in excess argon. Two

band centered at 2031 cm−1 was observed after sample deposition. Sample annealing to 25 K resulted in the appearance of three bands at 2027.9, 2033.3, and 2035.5 cm−1, which increased during λ > 220 nm irradiation. At the same time, an intense band at 1995.1 cm−1 and a weak band at 2015.4 cm−1 were produced. All of these bands sharpened after sample annealing to 30 K and decreased when the sample was annealed to 35 K. The infrared spectra from the reactions of lanthanum and (CN)2 are shown in Figure 3. New product absorptions at 1981.7, 1996.2, 2027.3, and 2042.8 cm−1 appeared upon λ > 220 nm irradiation, and their behaviors during sample annealing are similar to those observed in the reactions of scandium and yttrium with (CN)2. All of the lanthanum product absorptions exhibit multiple splittings due to different

Figure 1. Infrared spectra in the product absorption region from the reactions of laser-ablated scandium atoms with (CN)2 in solid argon at 4 K. (a) Sc and 1% (CN)2 codeposited for 1 h, (b) after λ > 220 nm irradiation, (c) after annealing to 25 K, (d) after annealing to 30 K, (e) after annealing to 35 K.

absorptions were observed right after sample deposition. The weaker band at 2054.2 cm−1 arises from CNCN, which is a common product band in the reactions of laser ablated metal atoms and (CN)2.29 The more intense band at 2051.1 cm−1 increased dramatically when the sample was subjected to full arc irradiation. New product absorptions at 2055.1 cm−1 and a series of broad bands appeared between 2010 and 2040 cm−1 upon λ > 220 nm irradiation as well. Subsequent sample annealing to 25 and 30 K sharpened these new absorptions, and new bands at 2031.0 and 2043.2 cm−1 appeared at the expense of the broad bands. All these absorptions decreased when the sample was further annealed to 35 K.

Figure 3. Infrared spectra in the product absorption region from the reactions of laser-ablated lanthanum atoms with (CN)2 in solid argon at 4 K. (a) La and 1% (CN)2 codeposited for 1 h, (b) after annealing to 25 K, (c) after λ > 220 nm irradiation, (d) after annealing to 30 K, (e) after annealing to 35 K. B

DOI: 10.1021/acs.jpca.8b06810 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A Table 1. Infrared Absorptions (cm−1) Observed for the Group 3 Metal Isocyanide Molecules in Solid Argon (CN)2 ScNC Sc(NC)2 Sc(NC)3 YNCb Y(NC)2 Y(NC)3 LaNCe La(NC)2e La(NC)3e

2055.1 2031.0 2043.2 2051.1 2033.3 1995.1 2015.4 2035.5 2027.3 1981.7 1996.2 2042.8

(13CN)2 2013.9 1990.6 2002.8 2010.0 1992.5 1955.5 1976.2 1994.8 1986.3 1941.6 1956.7 2001.5

(CN)2 + NC13CN + (13CN)2 a

2055.1, 2013.6 2031.0, 1996.0, 1990.6 2043.2, 2038.0, 2002.8 2051.0, 2017.4a, 2013.6a, 2009.9 2033.3, 1992.5 1994.9a, 1963.9, 1955.5 2010.0, 1976.2d 2035.5, 2005.1, 1999.2, 1994.9a 2027.5, 1986.3 1981.5, 1947.6, 1941.6 1996.2, 1983.8, 1956.5 2042.2, 2009.7, 2006.3, 2001.0

(C15N)2

(CN)2 + NCC15N + (C15N)2

2022.5 1998.6 2010.2 2018.6 2001.1c 1963.4c 1982.3c 2003.2c 1994.6a,c 1950.7c 1964.2c 2010.4c

2055.1, 2022.5 2031.0, 2003.2, 1998.6 2043.2, 2038.1, 2010.2 2051.1, 2025.4, 2021.8, 2018.6 2033.3, 2001.1 1995.1, 1970.7, 1963.4 2009.8, 1982.3d 2035.5,d, 2007.0, 2003.2 2027.5, 1994.6a 1981.5, 1955.8, 1950.7 1995.9a, 1991.9, 1964.2 2042.2,d, 2015.0, 2010.4

Overlap with other absorptions. bMatrix site absorptions observed at 2027.9 (CN)2, 1987.1 (13CN)2, and 1996.0 (C15N)2 cm−1. cObtained from the experiments using (CN)2 + NCC15N + (C15N)2 . dToo weak to be observed. eMajor absorption of multiple matrix site splittings. a

matrix trapping sites, and the major band of each multiplet is listed in Table 1. To help assign the absorptions of the new reaction products, experiments on the reactions of group 3 metal atoms and isotopically labeled (CN)2 samples were carried out, and the corresponding infrared spectra are shown in Figures 4−6.

Figure 5. Infrared spectra in the product absorption region from the reactions of laser-ablated yttrium atoms with isotopically substituted (CN)2 in solid argon at 4 K. Spectra were taken after λ>220 nm irradiation followed by annealing to 30 K (a) 0.25% (CN)2 + 0.5% NCC15N + 0.25% (C15N)2, (b) 1% (CN)2, (c) 1% (13CN)2, (d) 0.25% (CN)2 + 0.5% NC13CN + 0.25% (13CN)2. The asterisks denote the absorptions of CNCN and its isotopomers. s denotes the site absorption of YNC isotopomers. The dotted lines indicate where the absorptions would appear.

Figure 4. Infrared spectra in the product absorption region from the reactions of laser-ablated scandium atoms with isotopically substituted (CN)2 in solid argon at 4 K. Spectra were taken after λ > 220 nm irradiation followed by annealing to 30 K (a) 0.25% (CN)2 + 0.5% NCC15N + 0.25% (C15N)2, (b) 1% (C15N)2, (c) 1% (CN)2, (d) 1% (13CN)2, (e) 0.25% (CN)2 + 0.5% NC13CN + 0.25% (13CN)2. The asterisks denote the absorptions of CNCN and its isotopomers.

experiments with scrambled (CN)2 + NC13CN + (13CN)2 and (CN)2 + NCC15N + (C15N)2 samples, suggesting that only one CN moiety is involved in this mode. Because no other absorption was found to track this band, we assign the 2055.1 cm−1 absorption to the triatomic ScNC molecule. In the reactions of yttrium and (CN)2, the 2027.9, 2033.3 cm−1 doublet exhibits no intermediate absorption with scrambled isotopically labeled (CN)2 samples (Figure 5). The 12C/13C and 14N/15N isotopic ratios are almost the same for both bands, which are characteristic of a N−C stretch. All these experimental results indicate the 2027.9, 2033.3 cm−1 doublet is due to the N−C stretching mode of YNC trapped in different matrix sites. Following the cases of ScNC and YNC, the 2027.3 cm−1 band observed during the lanthanum and (CN)2 reactions is assigned to the LaNC molecule. Our assignments of the product structures are further supported by DFT calculations. Geometry optimization at the B3LYP level of theory reveals that ScNC possesses a linear 3Δ ground state with the 1Σ+ being 7.1 kcal/mol higher in energy,

Isotopic frequency ratios of the product bands can be obtained from the experiments using pure (13CN)2 and (C15N)2 samples. Scrambled isotopic samples such as (CN)2 + NC13CN + (13CN)2 and (CN)2 + NCC15N + (C15N)2 were employed to identify the stoichiometries of the products. The new product band positions observed in the experiments with isotopically labeled (CN)2 samples are listed in Table 1. MNC (M = Sc, Y, La). The new band at 2055.1 cm−1 shifted to 2013.9 and 2022.5 cm−1 when scandium reacted with (13CN)2 and (C15N)2, respectively (Figure 4). The experimental 12C/13C and 14N/15N isotopic frequency ratios (1.0205 and 1.0161, Table 2) are consistent with values for the N−C stretching mode of metal isocyanide molecules.10−14 No intermediate absorption was observed for this band in the C

DOI: 10.1021/acs.jpca.8b06810 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

consistent with previous CISD(+Q)/RCCSD(T) calculations.15 An intense N−C stretching band at 2103.2 cm−1 was predicted by theory which agrees well with the observed frequency at 2055.1 cm−1. The Sc−N stretch was predicted at 463.2 cm−1 with moderate intensity, but it is below the detection limit of our spectrometer and not observed in the experiment. Additional calculations were carried out on ScCN which is a stable isomer of ScNC. Based on the calculation results, ScCN also possesses a 3Δ ground state, but it is 5.6 kcal/mol less stable than ScNC at the B3LYP level. The relative stability of ScCN and ScNC remains the same with the latter being more stable by 4.1 kcal/mol according to the single-point CCSD(T) calculations, which are in line with the B3LYP and high level ab initio computational results reported before.3,15 The C−N stretch of ScCN was computed at 2227.7 cm−1, too high to match the experimental value at 2055.1 cm−1. The assignment of ScNC was further confirmed by the excellent agreement between calculated 12C/13C and 14N/15N isotopic frequency ratios (1.0204 and 1.0165) and those observed experimentally (1.0205 and 1.0161). Instead, the computed 12C/13C and 14N/15N ratios of ScCN are quite different, 1.0223 and 1.0149, since the central atom moves more in the vibration (Table 2).

Figure 6. Infrared spectra in the product absorption region from the reactions of laser-ablated lanthanum atoms with isotopically substituted (CN)2 in solid argon at 4 K. Spectra were taken after λ > 220 nm irradiation followed by annealing to 30 K (a) 0.25% (CN)2 + 0.5% NCC15N + 0.25% (C15N)2, (b) 1% (CN)2, (c) 1% (13CN)2, (d) 0.25% (CN)2 + 0.5% NC13CN + 0.25% (13CN)2. The asterisks denote the absorptions of CNCN and its isotopomers. The dotted line indicates where the absorption would appear.

Table 2. Comparisons between Calculated (B3LYP) and Experimental Absorptions (cm−1) and Isotopic Frequency Ratios of Ground State Group 3 Metal Cyanide and Isocyanide Molecules 12

C/13C

freq calcd ScCNa ScNCb Sc(CN)2 Sc(NC)2 Sc(CN)3 Sc(NC)3 YCN YNC Y(CN)2 Y(NC)2 Y(CN)3 Y(NC)3 LaCN LaNC La(CN)2 La(NC)2 La(CN)3 La(NC)3

2227.7 (32) 2103.2 (420) 2223.7 (53) 2227.7 (121) 2080.2 (905) 2082.0 (360) 2226.7 (225 × 2) 2233.7 (0) 2058.3 (1018 × 2) 2093.7 (0) 2224.7 (52) 2084.1 (496) 2222.4 (112) 2223.8 (18) 2077.0 (807) 2091.5 (174) 2227.5 (157 × 2) 2231.0 (0) 2072.5 (842 × 2) 2096.3 (0) 2215.3 (78) 2064.4 (608) 2209.0 (55) 2209.5 (97) 2066.3 (818) 2077.8 (302) 2215.0 (150 × 2) 2219.0 (4) 2064.7 (861 × 2) 2088.3 (0)

obsd

calcd 1.0223 1.0204 1.0222 1.0223 1.0206 1.0206 1.0225 1.0227 1.0205 1.0198 1.0223 1.0204 1.0223 1.0223 1.0206 1.0203 1.0224 1.0225 1.0206 1.0201 1.0222 1.0205 1.0222 1.0221 1.0209 1.0206 1.0222 1.0223 1.0209 1.0203

2055.1

2031.0 2043.2

2051.1

2033.3

1995.1 2015.4

2035.5

2027.3

1981.7 1996.2

2042.8

14

N/15N

obsd 1.0205

1.0203 1.0202

1.0204

1.0205

1.0203 1.0198

1.0204

1.0206

1.0207 1.0202

1.0206

calcd 1.0149 1.0165 1.0150 1.0149 1.0164 1.0164 1.0148 1.0146 1.0165 1.0171 1.0149 1.0165 1.0150 1.0149 1.0163 1.0166 1.0149 1.0148 1.0164 1.0168 1.0152 1.0164 1.0151 1.0151 1.0161 1.0164 1.0151 1.0149 1.0161 1.0166

obsd 1.0161

1.0162 1.0164

1.0161

1.0161

1.0161 1.0167

1.0161

1.0164

1.0159 1.0163

1.0161

The C−N stretching frequency and isotopic frequency ratios for the higher energy 1Σ+ ScCN: 2229.4 (73), 1.0223 (12C/13C), 1.0147 (14N/15N) The N−C stretching frequency and isotopic frequency ratios for the higher energy 1Σ+ ScNC: 2080.0 (645), 1.0203 (12C/13C), 1.0166 (14N/15N)

a

b

D

DOI: 10.1021/acs.jpca.8b06810 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

1.0203 and 1.0162 obtained from the experiment. The calculated frequency and isotopic ratios for the symmetric N−C stretch agree with the experimental values as well. For comparison, frequency calculations were also carried out on the less stable Sc(CN)2 isomer which was found to absorb at 2223.7 and 2227.7 cm−1 with 12C/13C and 14N/15N ratios of 1.0222 and 1.0150 for the first band and 1.0223 and 1.0149 for the second band. As shown in Table 2, it is clear that the computational results strongly support our assignment of the Sc(NC)2 molecule which absorbs 2031.0 and 2043.2 cm−1. In the cases of yttrium and lanthanum, the computed antisymmetric and symmetric N−C stretching vibrational frequencies and isotopic ratios of Y(NC)2 and La(NC)2 are in much better agreement with the experimentally observed values than those of Y(CN)2 and La(CN)2 (Table 2). M(NC)3 (M = Sc, Y, La). The 2051.1 cm−1 band shifted to 2010.0 and 2018.6 cm−1 in the experiments with (13CN)2 and (C15N)2 samples, respectively. The 12C/13C and 14N/15N isotopic frequency ratios (1.0204 and 1.0161) as well as the band position are characteristic of a N−C stretch as in the ScNC and Sc(NC)2 cases. In the (CN)2 + NC13CN + (13CN)2 experiment, four absorptions at 2051.0, 2017.4, 2013.6, and 2009.9 cm−1 were observed despite that fact that the two intermediate bands are overlapped with 13CN13CN and ScN13 (Figure 4). The appearance of two new intermediate absorptions using scrambled isotopic sample indicates the involvement of three equivalent CN ligands in the absorber, which is consistent with the spectroscopic characters of other highly symmetric three-coordinate molecules.30−33 Hence, we assign the 2051.1 cm−1 band to the antisymmetric N−C stretching mode of the Sc(NC)3 molecule. This assignment is further supported by the (CN)2 + NCC15N + (C15N)2 experiments. The spectroscopic characters of the yttrium and lanthanum product bands at 2035.5 and 2042.8 cm−1 are very similar to that of Sc(NC)3, and new products in the form of Y(NC)3 and La(NC)3 can be assigned accordingly. All of the group 3 M(NC)3 molecules were computed to have the singlet ground state with D3h symmetry. The corresponding cyanide isomers are less stable by 17.0 (Sc), 12.6 (Y) and 16.0 (La) kcal/mol at the CCSD(T) level. For Sc(NC)3, two N−C stretching vibrations were predicted at 2058.3 and 2093.7 cm−1. The first band corresponds to a doubly degenerate N−C stretch while the second band arises from the infrared inactive symmetric N−C stretch. As listed in Table 2, both the experimental band position and isotopic frequency ratios are in better agreement with the computed values for Sc(NC)3 than Sc(CN)3. Additional frequency calculations on the Sc(N13C)(NC)2 and Sc(NC)(N13C)2 isotopomers gave two intense N−C stretches at 2024.1 and 2033.3 cm−1, which match the observed frequencies at 2013.6 and 2017.4 cm−1. Our assignments of the Y(NC)3 and La(NC)3 molecules are also strongly supported by the calculated frequencies as well as 12C/13C and 14N/15N frequency ratios (Table 2). The optimized ground state geometries of M(NC)x and M(CN)x (M = Sc, Y, La; x = 1−3) are shown in Figure 7. The C−N bond lengths of the M(NC)x complexes are around 1.18 Å, which are 0.02 Å longer than those of the M(CN)x complexes. Such a difference is typical for metal cyanide/ isocyanide,10−14 and the C−N bond is slightly weaker when CN is nitrogen bound to the metal center. This is consistent with the fact that the C−N stretching vibrational frequencies of group 3 metal cyanides are higher than the N−C stretches of

Different from ScCN and ScNC, both YNC and YCN possess closed shell singlet ground states, and YNC was predicted to be more stable than YCN by 5.8 kcal/mol at the CCSD(T) level of theory. Frequency calculations show that both molecules possess three vibrational modes. The N−C stretch of YNC is predicted at 2084.1 cm−1, about 140 cm−1 lower than the C−N stretch of YCN. The experimental value of 2033.3 cm−1 agrees better with the computed frequency of YNC. The calculated 12C/13C and 14N/15N isotopic frequency ratios for the N−C stretch of YNC are also much closer to the experimental ratios of 1.0205 and 1.0161, as in the scandium case. For lanthanum, both triplet LaCN and LaNC were computed to be slightly more stable than the singlet states by 2.4 and 0.9 kcal/mol at the B3LYP level, but single-point CCSD(T) calculations reveal that both isomers have the singlet ground state with the triplet state being 2.5 and 3.6 kcal/mol higher in energy. Following the yttrium case, LaNC is more stable than LaCN by 5.9 kcal/mol at the CCSD(T) level. As shown in Table 2, both the computed N−C stretching frequency and 12C/13C and 14N/15N isotopic frequency ratios strongly support our assignment of the LaNC molecule which absorbs at 2027.3 cm−1. M(NC)2 (M = Sc, Y, La). The scandium product bands at 2031.0 and 2043.2 cm−1 appeared simultaneously after λ > 220 nm irradiation followed by sample annealing. The 12C/13C and 14 N/15N frequency ratios of both absorptions (Table 2) as well as their same behaviors throughout the experiment suggest they should be due to two N−C stretches of the same isocyanide molecule. Two sets of triplets at 2031.0, 1996.0, 1990.6 cm−1 and 2043.2, 2038.0, 2002.8 cm−1 were observed when scandium reacted with the scrambled (CN)2 + NC13CN + (13CN)2 sample (Figure 4), suggesting the absorber of the 2031.0 cm−1 band consists of two equivalent CN moieties from the same (CN)2 molecule. This is also consistent with the results from the reactions of scandium and (CN)2 + NCC15N + (C15N)2. Hence, the 2031.0 and 2043.2 cm−1 bands are assigned to the antisymmetric and symmetric stretching N−C vibrational modes of Sc(NC)2. In the yttrium case, two new absorptions at 1995.1 and 2015.4 cm−1 appeared after λ > 220 nm irradiation, and they tracked each other throughout the experiment. Both of them possess typical isotopic frequency ratios for a N−C stretch. Experiments using scrambled (CN)2 + NC13CN + (13CN)2 and (CN)2 + NCC15N + (C15N)2 samples provide solid evidence for the involvement of two equivalent CN moieties in this yttrium product (Figure 5). We therefore assign the 1995.1 and 2015.4 cm−1 absorptions to the antisymmetric and symmetric N−C stretching modes of the Y(NC)2 molecule. The antisymmetric and symmetric N−C stretches of La(NC)2 were observed at 1981.7 and 1996.2 cm−1, and their assignments can be readily made following the scandium and yttrium analogs. Calculations at the B3LYP level of theory reveal that all of the M(NC)2 and M(CN)2 (M = Sc, Y, La) molecules possess bent geometries and 2A1 ground states, and the isocyanide isomers are more stable than the cyanides by 6.4(Sc), 9.0(Y), and 10.3(La) kcal/mol at the CCSD(T) level. For Sc(NC)2, two infrared active N−C stretches were predicted at 2080.2 and 2082.0 cm−1. The first band with higher intensity corresponds to the antisymmetric N−C stretching mode, which is very close to the absorption observed at 2031.0 cm−1. The computed 12C/13C and 14N/15N frequency ratios for this mode are 1.0206 and 1.0164, almost the same as the ratios of E

DOI: 10.1021/acs.jpca.8b06810 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

Table 3. NBO Charges (B3LYP) of Ground State Group 3 Metal Cyanide and Isocyanide Molecules ScCNa Sc(CN)2 Sc(CN)3 YCN Y(CN)2 Y(CN)3 LaCN La(CN)2 La(CN)3

M

C

N

0.73 1.37 1.64 0.75 1.42 2.00 0.73 1.42 2.04

−0.34 −0.34 −0.29 −0.41 −0.38 −0.37 −0.40 −0.37 −0.36

−0.39 −0.34 −0.26 −0.34 −0.34 −0.30 −0.33 −0.34 −0.32

ScNCb Sc(NC)2 Sc(NC)3 YNC Y(NC)2 Y(NC)3 LaNC La(NC)2 La(NC)3

M

C

N

0.79 1.50 1.93 0.81 1.59 2.30 0.79 1.57 2.32

0.24 0.28 0.35 0.30 0.29 0.31 0.32 0.28 0.29

−1.03 −1.03 −1.00 −1.11 −1.08 −1.08 −1.10 −1.06 −1.06

NBO charges for the higher energy 1Σ+ ScCN: 0.66 (Sc), −0.33 (C), −0.33 (N) bNBO charges for the higher energy 1Σ+ ScNC: 0.73 (Sc), 0.31 (C), −1.04 (N) a

during which the formal oxidation state of metal increases from I to III. For a cyanide/isocyanide molecule with a given stoichiometry, the charge on the metal center of isocyanide is always higher than that of the cyanide isomer because of the higher electronegativity of nitrogen atom which tends to attract more electrons from the metal side. Consistent with this notion, the nitrogen atom of isocyanide carries much more negative charge than that of cyanide. The carbon atom of isocyanide is therefore positively charged, which is completely different from the situation of cyanide where both carbon and nitrogen carry similar negative charge. Natural population analysis (Table S2) reveals that the electronic structure of an isocyanide is almost the same as the corresponding cyanide isomer regardless of the coordination number and metal center. For scandium, the two unpaired electrons of ScNC are mainly located in the 4s and 3d orbitals with the latter being slightly more populated. A 4s0.33d0.7 configuration was found for the unpaired electron of Sc(NC)2. Sc(NC)3 has only 0.12 and 0.17 e in the 4s and 4p orbitals, while 0.8 e was found in the 3d orbital, suggesting it is mainly the Sc 3d orbitals that participate the Sc-NC bonding. The YNC molecule possesses a closed shell singlet ground state with the two valence electrons of yttrium paired in the 5s orbital. Similar to the Sc(NC)3 case, 0.51 e was found in the 4d orbital of Y(NC)3, while the population in 5s and 5p orbitals are only 0.1 and 0.12 e. For Y(NC)2, the configuration of the unpaired electron is 5s0.64d0.3, which has more contribution from the 5s orbital. The electronic structures of the lanthanum products are very similar to those of the yttrium products. The infrared spectra shown in Figures 1−3 demonstrate that all the group 3 isocyanide molecules are formed upon λ > 220 nm irradiation but not on sample annealing, indicating that ground state group 3 metal atoms cannot react with (CN)2 to form the products. Instead, UV−vis irradiation is required to excite the metal atoms and form the isocyanide products via breaking the NC−CN bond with some double bond character.36 Calculations at the CCSD(T) level reveal that it is highly favorable to form Sc(NC)2, Y(NC)2 and La(NC)2 via the reactions of metal atoms and (CN)2 (Table 4). The monoisocyanides should be formed upon photoinduced dissociation of the corresponding diisocyanide molecules as observed in the experiments. Reactions of metal atoms and CN could also contribute to the formation of isocyanides which are exothermic by 96.5, 141.4, and 114.7 kcal/mol at the CCSD(T) level. For Sc(NC)3, Y(NC)3, and La(NC)3, they should result from the reactions of (CN) 2 and the

Figure 7. Optimized structures (bond lengths in angstrom units and bond angles in degrees) for the scandium, yttrium, and lanthanum cyanide/isocyanide molecules at the B3LYP/6-311+G(d) level of theory. Relative energies (relative to isocyanide, kcal/mol) at the CCSD(T) level are in parentheses.

isocyanides. The computed M−C bond lengths of cyanides are about 0.16 Å longer than the M−N bond lengths of isocyanides regardless of the coordination number. Since the M−C single bond length is only 0.04 Å longer than the M−N bond,34 some double bond character should be involved between group 3 metal atom and nitrogen bound CN ligand. For the group 3 monocyanide and isocyanide molecules, the M−C and M−N bond lengths increase from Sc to La, which parallels the increase in metal radius down the group. The Sc− C and Sc−N bond lengths of singlet ScCN (2.143 Å) and ScNC (1.989 Å) are close to those of the triplet ground state, suggesting the change in spin state has little effect on their geometries. This agrees with the fact that the two unpaired electrons of triplet ScCN/ScNC are located in the 4s and 3d nonbonding orbitals. Similar trends are found for the M−C and M−N bond lengths of M(NC)2,3 and M(CN)2,3, which follows the geometry changes of the corresponding group 3 metal fluorides.35 In addition, the CMC and NMN bond angles of dicyanide and isocyanide molecules decrease from Sc to La, which also resembles the change on bond angles of scandium, yttrium and lanthanum difluorides. The NBO charges of group 3 metal cyanide and isocyanide molecules are listed in Table 3. It is clear that the charge on metal center increases as the number of CN ligands increases, F

DOI: 10.1021/acs.jpca.8b06810 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

upon λ > 220 nm irradiation. These diisocyanides can either react with another (CN)2 molecule to form triisocyanides and CN or lose one CN ligand to form monoisocyanides under UV−vis irradiation. No absorption can be assigned to the group 3 metal cyanide isomers which were predicted to be less stable regardless of the number of CN ligands. On the basis of the CCSD(T)//B3LYP results, ScCN/ScNC possesses a 3Δ ground state while 1Σ+ is most stable for YCN/YNC and LaCN/LaNC. All of the group 3 metal dicyanides/ diisocyanides were found to have bent geometries with 2A1 ground state, and D3h or near D3h symmetry with singlet ground state was obtained for the tricyanides and isocyanides. Frequency calculations reveal that the C−N stretches for group 3 metal cyanides are about 150 cm−1 higher than the N−C stretches of isocyanides, and both the frequencies and isotopic frequency ratios for the isocyanides agree well with the experimental values. The isotopic frequency ratios are also a diagnostic for the particular isomer formed because the central atom contributes more motion to the antisymmetric vibrational mode. The difference in C−N stretching vibrational frequencies between group 3 cyanides and isocyanides is consistent with the difference in C−N bond lengths of both isomers, and the changes of their bond lengths and angles follow those of group 3 metal fluorides which possess similar geometries with group 3 metal isocyanides.

Table 4. Calculated [CCSD(T)//B3LYP] Reaction Energies (kcal/mol) for the Formation of Group 3 Metal Isocyanide Molecules M M + (CN)2 → M(NC)2 M(NC)2 → MNC + CN M + CN → MNC M(NC)2 + (CN)2 → M(NC)3 + CN M(NC)2 + CN → M(NC)3 MNC + (CN)2 → M(NC)3

Sc

Y

La

−70.0 109.4 −96.5 −16.4 −116.7 −90.3

−108.9 103.2 −141.4 −18.9 −116.9 −84.3

−82.6 103.6 −114.7 −8.0 −127.8 −95.6

corresponding diisocyanides which are the most abundant products after UV−vis irradiation. Other possible pathways for the formation of triisocyanides include the reactions of either diisocyanides and CN or monoisocyanides and (CN)2, both of which are more exothermic as listed in Table 4. In addition to the isocyanide products identified in our experiments, the group 3 metal cyanides are stable molecules as well, although all of them are less stable than the isocyanide isomers by at least 4 kcal/mol at the CCSD(T) level (Figure 7). Experimentally, no absorption due to M(CN)1−3 was detected between 2100 and 2250 cm−1 where the cyanide absorptions would appear based on the calculated frequencies. It was concluded that the ScNC-ScCN isomerization process is kinetically hindered by an energy barrier of 7.78 kcal/mol at the CCSD(T) level.3 It should be noted that group 3 metal atoms are known to form four-coordinate anions such as MH4− and MF4− where the III oxidation states of metal centers are retained.35,37,38 Ground-state group 3 M(NC)4− anions (1A1) were predicted to be more stable than the corresponding M(CN)4− isomers (1A1) based on our CCSD(T)//B3LYP calculations (Figure S1, Supporting Information). Two N−C stretching frequencies were obtained for the M(NC)4− anions. The infrared active triply degenerate N−C stretch was predicted at 2110.1 cm−1 for Sc, 2114.3 cm−1 for Y and 2104.7 cm−1 for La. If these anions were formed during the reactions of group 3 metal atoms and (CN)2, they would be observed in the region where the M(NC)1−3 molecules appear. There are some weak absorptions in the 2020−2060 cm−1 region in Figures 1−3 (trace e), and they could be the candidates for the M(NC)4− anions. It is not possible that these bands arise from the anions with lower coordination number since they cannot survive from UV−vis irradiation. Very broad features centered around 2220 (Sc), 2215 (Y), and 2200 cm−1 (La) appeared along with the isocyanide products upon λ>220 nm irradiation. Their 12C/13C frequency ratios (∼1.028) are larger than those of group 3 metal cyanides and isocyanides. These absorptions were produced only under UV−vis irradiation, and subsequent sample annealing allows the increase of these bands. It is most likely that these absorptions arise from the association complexes of (CN)2 and group 3 metal isocyanides. Analogous broad features were also observed in the reactions of iron and manganese with (CN)2.13

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpca.8b06810. Complete citation for ref 18, calculated frequencies of M(CN)x and M(NC)x (M = Sc, Y, La; x = 1−3) and their isotopomers, Mulliken atomic spin densities and natural population analysis of ground state group 3 metal isocyanides and cyanides, calculated geometries, relative stabilities and frequencies of the M(CN)4− and M(NC)4− (M = Sc, Y, La) anions (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Lester Andrews: 0000-0001-6306-0340 Yu Gong: 0000-0002-8847-1047 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by the Strategic Priority Research Program and Frontier Science Key Program (Grant Nos. XDA02030000 and QYZDY-SSW-JSC016) (X.C., Q.L., Y.G.) of the Chinese Academy of Sciences, NSFC (21771189) (Y.G.), Young Thousand Talented Program (Y.G.), and retirement funds from T.I.A.A. (L.A.).

■

■

CONCLUSIONS The reactions of group 3 metal atoms and (CN)2 were studied by matrix isolation infrared spectroscopy and theoretical calculations. Scandium, yttrium, and lanthanum atoms produced by laser ablation of corresponding metal targets reacted with (CN)2 to form Sc(NC)2, Y(NC)2, and La(NC)2

REFERENCES

(1) Cotton, F. A.; Wilkinson, G. S.; Murillo, C. A.; Bochmann, M. Advanced Inorganic Chemistry, 6th ed.; Wiley-Interscience: New York, 1999. (2) Hargittai, M. Molecular Structure of Metal Halides. Chem. Rev. 2000, 100, 2233−2301.

G

DOI: 10.1021/acs.jpca.8b06810 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A (3) Rayón, V. M.; Redondo, P.; Valdés, H.; Barrientos, C.; Largo, A. Cyanides and Isocyanides of First-Row Transition Metals: Molecular Structure, Bonding, and Isomerization Barriers. J. Phys. Chem. A 2007, 111, 6334−6344. (4) Flory, M. A.; Field, R. W.; Ziurys, L. M. The Pure Rotational Spectrum of CrCN (X 6Σ+): an Unexpected Geometry and Unusual Spin Interactions. Mol. Phys. 2007, 105, 585−597. (5) Flory, M. A.; Ziurys, L. M. Millimeter-Wave Rotational Spectroscopy of FeCN (X 4Δi) and FeNC (X 6Δi): Determining the Lowest Energy Isomer. J. Chem. Phys. 2011, 135, 184303. (6) Sheridan, P. M.; Flory, M. A.; Ziurys, L. M. Characterizing the Later 3d Cyanides: the Submillimeter Spectrum of CoCN(X 3Φi). J. Chem. Phys. 2004, 121, 8360−8368. (7) Sheridan, P. M.; Ziurys, L. M. Further Studies of 3d Transition Metal Cyanides: the Pure Rotational Spectrum of NiCN (X 2Δi). J. Chem. Phys. 2003, 118, 6370−6379. (8) Grotjahn, D. B.; Brewster, M. A.; Ziurys, L. M. The First Precise Molecular Structure of a Monomeric Transition Metal Cyanide, Copper(I) Cyanide. J. Am. Chem. Soc. 2002, 124, 5895−5901. (9) Brewster, M. A.; Ziurys, L. M. Rotational Spectroscopy of 3d Transition-Metal Cyanides: Millimeter-Wave Studies of ZnCN (X 2 + Σ ). J. Chem. Phys. 2002, 117, 4853−4860. (10) Chen, X.; Li, Q.; Gong, Y.; Andrews, L.; Liebov, B. K.; Fang, Z.; Dixon, D. A. Formation and Characterization of Homoleptic Thorium Isocyanide Complexes. Inorg. Chem. 2017, 56, 5060−5068. (11) Gong, Y.; Andrews, L.; Liebov, B. K.; Fang, Z.; Garner, E. B., III; Dixon, D. A. Reactions of Laser-Ablated U Atoms with (CN)2: Infrared Spectra and Electronic Structure Calculations of UNC, U(NC)2, and U(NC)4 in Solid Argon. Chem. Commun. 2015, 51, 3899−3902. (12) Fang, Z.; Garner, E. B.; Dixon, D. A.; Gong, Y.; Andrews, L.; Liebov, B. Laser-Ablated U Atom Reactions with (CN)2 to Form UNC, U(NC)2, and U(NC)4: Matrix Infrared Spectra and Quantum Chemical Calculations. J. Phys. Chem. A 2018, 122, 516−528. (13) Chen, X.; Li, Q.; Andrews, L.; Gong, Y. Matrix Infrared Spectra of Manganese and Iron Isocyanide Complexes. J. Phys. Chem. A 2017, 121, 8835−8842. (14) Andrews, L.; Cho, H. G.; Gong, Y. Reactions of Laser-Ablated Aluminum Atoms with Cyanogen: Matrix Infrared Spectra and Electronic Structure Calculations for Aluminum Isocyanides Al(NC)1,2,3 and Their Novel Dimers. J. Phys. Chem. A 2018, 122, 5342− 5353. (15) Kalemos, A.; Metropoulos, A.; Mavridis, A. Structure and Bonding of ScCN and ScNC: Ground and Low-Lying States. Chem. Phys. 2012, 399, 46−49. (16) Andrews, L.; Cho, H. G. Matrix Preparation and Spectroscopic and Theoretical Investigations of Simple Methylidene and Methylidyne Complexes of Group 4−6 Transition Metals. Organometallics 2006, 25, 4040−4053. (17) Cho, H. G.; Andrews, L. Matrix Preparation and Spectroscopic and Theoretical Investigation of Small High Oxidation-State Complexes of Groups 3−12, 14, Lanthanide and Actinide Metal Atoms: Carbon-Metal Single, Double and Triple Bonds. Coord. Chem. Rev. 2017, 335, 76−102. (18) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; et al. Gaussian 09, revision A.01; Gaussian, Inc.: Wallingford, CT, 2009. (19) Becke, A. D. Density-Functional Thermochemistry. III. The Role of Exact Exchange. J. Chem. Phys. 1993, 98, 5648−5652. (20) Lee, C.; Yang, W.; Parr, R. G. Development of the ColleSalvetti Correlation-Energy Formula into a Functional of the Electron Density. Phys. Rev. B: Condens. Matter Mater. Phys. 1988, 37, 785− 789. (21) McLean, A. D.; Chandler, G. S. Contracted Gaussian Basis Sets for Molecular Calculations. I. Second Row Atoms, Z = 11−18. J. Chem. Phys. 1980, 72, 5639−5648.

(22) Krishnan, R.; Binkley, J. S.; Seeger, R.; Pople, J. A. SelfConsistent Molecular Orbital Methods. XX. A Basis Set for Correlated Wave Functions. J. Chem. Phys. 1980, 72, 650−654. (23) Wachters, A. J. H. Gaussian Basis Set for Molecular Wavefunctions Containing Third-Row Atoms. J. Chem. Phys. 1970, 52, 1033−1036. (24) Hay, P. J. Gaussian Basis Sets for Molecular Calculations. The Representation of 3d Orbitals in Transition-Metal Atoms. J. Chem. Phys. 1977, 66, 4377−4384. (25) Dolg, M.; Stoll, H.; Preuss, H. Energy-Adjusted ab initio Pseudopotentials for the Rare Earth Elements. J. Chem. Phys. 1989, 90, 1730−1734. (26) Cao, X.; Dolg, M. J. Valence Basis Sets for Relativistic EnergyConsistent Small-Core Lanthanide Pseudopotentials. J. Chem. Phys. 2001, 115, 7348−7355. (27) Pople, J. A.; Head-Gordon, M.; Raghavachari, K. Quadratic Configuration Interaction − a General Technique for Determining Electron Correlation Energies. J. Chem. Phys. 1987, 87, 5968−5975. (28) Glendening, E. D.; Reed, A. E.; Carpenter, J. E.; Weinhold, F. NBO, version 3.1; University of Wisconsin: Madison, WI, 1988. (29) Maier, G.; Reisenauer, H. P.; Eckwert, J.; Sierakowski, C.; Stumpf, T. Matrix Isolation of Diisocyanogen, CNNC. Angew. Chem., Int. Ed. Engl. 1992, 31, 1218−1220. (30) Gong, Y.; Zhou, M. Formation and Characterization of Two FeO3 Isomers in Solid Argon. J. Phys. Chem. A 2008, 112, 10838− 10842. (31) Zhou, M.; Andrews, L. Infrared Spectra and Density Functional Calculations of the CrO2−, MoO2−, and WO2− Molecular Anions in Solid Neon. J. Chem. Phys. 1999, 111, 4230−4238. (32) Zhou, M.; Andrews, L. Matrix Infrared Spectra and Density Functional Calculations of ScCO, ScCO−, and ScCO+. J. Phys. Chem. A 1999, 103, 2964−2971. (33) Jiang, L.; Xu, Q. Reactions of Laser-Ablated La and Y Atoms with CO: Matrix Infrared Spectra and DFT Calculations of the M(CO)x and MCO+ (M = La, Y; x = 1−4) Molecules. J. Phys. Chem. A 2007, 111, 3271−3277. (34) Pyykkö, P.; Atsumi, M. Molecular Single-Bond Covalent Radii for Elements 1−118. Chem. - Eur. J. 2009, 15, 186−197. (35) Wang, X.; Andrews, L. Infrared Spectra of MF2, MF2+, MF4−, MF3, and M2F6 Molecules (M = Sc, Y, La) in Solid Argon. J. Phys. Chem. A 2010, 114, 2293−2299. (36) Corain, B. The Coordination Chemistry of Hydrogen Cyanide, Cyanogen, and Cyanogen Halides. Coord. Chem. Rev. 1982, 47, 165− 200. (37) Wang, X.; Chertihin, G. V.; Andrews, L. Matrix Infrared Spectra and DFT Calculations of the Reactive MHx (x = 1, 2, and 3), (H2)MH2, MH2+, and MH4− (M = Sc, Y, and La) Species. J. Phys. Chem. A 2002, 106, 9213−9225. (38) Wang, X.; Andrews, L. Homoleptic Tetrahydrometalate Anions MH4− (M = Sc, Y, La). Matrix Infrared Spectra and DFT Calculations. J. Am. Chem. Soc. 2002, 124, 7610−7613.

H

DOI: 10.1021/acs.jpca.8b06810 J. Phys. Chem. A XXXX, XXX, XXX−XXX