Proposal for Sets of 77Se NMR Chemical Shifts in Planar and Perpendicular Orientations of Aryl Group and the Applications

The orientational effect of p-YC6H4 (Ar) on δ(Se) is elucidated for ArSeR, based on experimental and theoretical investigations. Sets of δ(Se) are proposed for pl and pd employing 9-(arylselanyl)anthracenes (1) and 1-(arylselanyl)anthraquinones (2), respectively, where Se–C R in ArSeR is on the Ar plane in pl and perpendicular to the plane in pd. Absolute magnetic shielding tensors of Se (σ(Se)) are calculated for ArSeR (R = H, Me, and Ph), assuming pl and pd, with the DFT-GIAO method. Observed characters are well reproduced by the total shielding tensors (σ t(Se)). The paramagnetic terms (σ P(Se)) are governed by σ P(Se)xx + σ P(Se)yy, where the direction of nP(Se) is set to the z-axis. The mechanisms of the orientational effect are established both for pl and pd. Sets of δ(Se: 1) and δ(Se: 2) act as the standards for pl and pd, respectively, when δ(Se) of ArSeR are analyzed based on the orientational effect.

After the establishment of the orientational effect of aryl group in p-YC 6 H 4 SeR, together with the mechanism, δ(Se) of some aryl selenides are plotted versus δ(Se: 1) and/or δ(Se: 2). The treatment shows how δ(Se) of aryl selenides are in- 1 The contribution of relativistic terms has been pointed out for heavier atoms, but the perturbation would be small for the selenium nucleus. 2 This decomposition includes small arbitrariness due to the coordinate origin dependence, though it does not damage our chemical analyses and insights into the 77 Se NMR spectroscopy. terpreted based on the orientational effect. And it is demonstrated that the sets of δ(Se: 1) and δ(Se: 2) give a reliable guideline to analyze the structures of p-YC 6 H 4 SeR based on δ(Se).

Characters in δ(Se: 1) and δ(Se: 2)
The structures of all members of 1 and 2 are confirmed to be 1 (A: pl) and (B: pd), respectively, (see Scheme 1) [25]. The nature of δ(Se: 1) must be the results of 1 (A: pl), where n p (Se) is parallel to the π(C 6 H 4 Y-p). Characteristic points in δ(Se: 1) SCS are summarized as follows.
(1) Large upfield shifts (−23 ppm to −6 ppm) are observed for Y = NMe 2 , OMe, and Me and large downfield shifts (17 ppm to 33 ppm) are for Y = COOEt, CN, and NO 2 , relative to Y = H. The characters of δ(Se: 2) SCS are very different from those of δ(Se: 1) SCS . The characteristics must be the reflection of 2 (B: pd), where n p (Se) is perpendicular to π(C 6 H 4 Y-p). Characteristic points of δ(Se: 2) SCS are as follows.
(1) Large upfield shifts (−21 to −6 ppm) are observed for Y = NMe 2 , OMe, Me, F, Cl, and Br, relative to Y = H. (2) Downfield shifts (3 ppm to 9 ppm) are brought by Y = CN and NO 2 , where the magnitude by Y = CN is larger than that by NO 2 .  Figure 1 shows the results. Indeed, it emphasizes the difference in the characters between δ(Se: 1) SCS and δ(Se: 2) SCS , but most of δ(Se: 2) SCS seem to correlate well with δ(Se: 1) SCS , as shown by a dotted line (a = 0.58). Two points corresponding to Y = H and NO 2 deviate upside and downside from the line, respectively. Namely, points for 2 with Y of non-H are more downside (upfield) than expected from δ(Se: 1a) SCS and δ(Se: 2a) SCS , especially for δ(Se: 2j) SCS .
Why are such peculiar behaviors observed in 1 and 2, caused by the orientational effect of the aryl group? The mechanism is elucidated based on the QC calculations performed on 4-6, assuming pl and pd for each.
The r values become larger in an order of 4(pl) < 5(pl) 6(pl) for 1 and in an order of 5(pd) < 4(pd) ≈ 6(pd) for 2. Namely, observed δ(Se: 1) SCS and δ(Se: 2) SCS are reproduced by σ t rel (Se: 6 (pl)) SCS and σ t rel (Se: 6 (pd)) SCS , respectively, in most successfully. Figure 2 exhibits the plots for (a) 1 versus 6 (pl) and (b) 2 versus 6 (pd). The correlations are given in Table 2 (entries 7 and 10). The results demonstrate that the characters of δ(Se) SCS observed in 1 originate from the planar structure and those in 2 from the characteristic structure, where Se−C Atq in p-YC 6 H 4 SeAtq is perpendicular to the p-YC 6  which correspond to the orientational effect caused by Ph in 4a. 7 The inverse orientational effect is predicted for 5a. σ p (Se) and σ t (Se) of 5a (pd) are smaller than those of 5a (pl) by 41 ppm and 49 ppm, respectively. While σ p (Se) and σ t (Se) of 5a (pl) are smaller than those of 4a (pl) by 90 ppm and 83 ppm, respectively, the values of 5a (pd) are smaller than those of 4a (pd) by 174 ppm and 178 ppm, respectively. The differences are −84 ppm and −95 ppm, respectively, which also correspond to the differences in the orientational effect of the Ph group between 5a and 4a, respectively. The more effective contribution to downfield shifts by the Se−C Me bond in 5a (pd), relative to 5a (pl), must be responsible for the results. The orientational effect cannot be discussed for 6a of the Cs symmetry with Y = H. 7 The DFT shieldings are deshielded in general, due to the underestimation of the orbital energy differences, which lead to the overestimation of the σ p (Se) [48]. MP2 calculations are also performed on 4a (pl), 4a (pd), 5a (pl), and 5a (pd). The geometries are optimized with the MP2/6-311+G(3d,2p) method. σ t (Se) are calculated with the MP2-GIAO method, employing the 6-311+G(2d,p) basis sets. The results are as follows (in ppm): (σ t (Se: 4a (pl)), σ t (Se: 4a (pd))) = (1827.3, 1865.5) and (σ t (Se: 5a (pl)), σ t (Se: 5a (pd))) = (1761.0, 1708.7). σ t (Se: 4a (pl)) is evaluated to be more downfield than σ t (Se: 4a (pd)) by 38 ppm, whereas σ t (Se: 5a (pl)) is evaluated to be more upfield than σ t (Se: 5a (pd)) by 52 ppm. The results support the orientational effects evaluated at the DFT level for 4a and 5a, although the basis sets are not the same. What mechanism is operating in the Y dependence? σ p (Se) of 4-6 in pl and pd are analyzed next.

Y dependence in 4-6
To get an image in the behavior of σ p (Se) xx , σ p (Se) yy , and σ p (Se) zz of 4-6, the values are plotted versus σ p (Se). Figure 3 shows the plots for 4 (pd) and 6 (pl). The correlations in 4 (pd) are linear and both σ p (Se) xx and σ p (Se) yy increase along with σ p (Se). The plot for 5 (pd) is similar to that for 4 (pd), although not shown. In the case of 6 (pl), the correlations are linear but the slope for σ p (Se) yy is inverse to that for σ p (Se) xx . The plots of σ p (Se) xx and σ p (Se) yy do not give smooth lines for 4 (pl), 5 (pl), and 6 (pd). However, the slopes for σ p (Se) zz are very smooth and the magnitudes are very close to 1.0 for all cases in 4-6.
The small Y-dependence of σ p (Se) zz is reasonably explained through the main interaction of the 4p z (Se)-π(C 6 H 4 )-p z (Y) type in pl, where 4p x (Se) and 4p y (Se) do not take part in the interaction directly. The main interaction in pd is the σ(C Ar SeX)-π(C 6 H 4 )-p x (Y) (X = H or C) type, which modifies the contributions of 4p x (Se) and 4p y (Se) in the C Ar SeX bonds. However, the results show that the interaction in pd affects on σ p (Se) xx and σ p (Se) yy but not on σ p (Se) zz .

Mechanism of Y dependence
The mechanism of Y dependence in 4-6 is elucidated by exemplifying 4. As shown in Scheme 2, the main interaction between Se and Y in 4 (pl) is the 4p z (Se)-π(C 6 H 4 )p z (Y) type, which modifies the contributions of 4p z (Se) in π(SeC 6 H 4 Y) and π * (SeC 6 H 4 Y). Since (σ p (Se) xx + σ p (Se) yy ) controls σ p (Se) of 4 (pl) effectively, admixtures between 4p z (Se) in modified π(SeC 6 H 4 Y) and π * (SeC 6 H 4 Y) with 4p y (Se) and 4p x (Se) in σ(C Ar SeH) and σ * (C Ar SeH) must originate the Y dependence mainly when a magnetic field is applied. 9 Since σ p zz,N contains the L z,N operator, σ p zz,N arises from admixtures between atomic p x and p y orbitals of N in various molecular orbitals. When a magnetic field is applied on a selenium compound, mixings of unoccupied molecular orbitals (MO's; ψ i ) into occupied orbital MO's (ψ i ) will occur. Such admixtures generate σ p zz,N if ψ i and ψ j contain p x and p y of N, for example. σ p xx,N and σ p yy,N are also understood similarly. Consequently, Y of both donors and acceptors are effective for the Y dependence in 4 (pl). Scheme 3(a) shows the mechanism for pl.
In the case of 4 (pd), 4p z (Se) remains in n p (Se) in the almost pure form. 10 The σ(C Ar SeH)-π(C 6 H 4 )-p x (Y) interaction occurs instead, which modifies the contributions of 4p x (Se) and 4p y (Se) in σ(C Ar SeH) and σ * (C Ar SeH) 9 σ p is exactly expressed by Ramsey's equation [49]. While σ p is evaluated accurately by the CPHF method, it is approximated as σ (a) Structures are optimized with the 6-311+G(3df) basis sets for Se and 6-311+G(3d,2p) basis sets for other nuclei at the DFT (B3LYP) level, assuming pl and pd for each of Y [47]. σ(Se) are calculated based on the DFT-GIAO method with the same methods.
(see Scheme 2). (σ p (Se) xx + σ p (Se) yy ) determines effectively σ p (Se) of 4 (pd). Therefore, Y dependence of 4 (pd) originates mainly from admixtures between 4p z (Se) in n p (Se) and 4p x (Se) and 4p y (Se) in modified σ * (C Ar SeH) since n p (Se) of 4p z (Se) is filled with electrons. Consequently, Y dependence in 4 (pd) must be more sensitive to Y of donors, which is a striking contrast to the case of 4 (pl). Scheme 3(b) summarizes the mechanism for pd. The mechanisms proposed for 4 (pl) and 4 (pd) must be applicable to 5 and 6. The expectations are just observed in δ(Se: 1) SCS and δ(Se: 2) SCS .

Applications of δ(Se: 1) and δ(Se: 2) as the standards
Odom made a lot of effort to explain δ(Se) of 7 based on the electronic effect of Y [13]. However, the attempt was not successful: δ (Se: 7) were not correlated well with δ(Se: 5). How are δ(Se) of p-YC 6 H 4 SeR interpreted based on the orientational effect? Our explanation for the relationship between δ(Se) of p-YC 6 H 4 SeR and the structures is as follows. Figure 5 shows the plot of δ(Se: 5) SCS measured in CDCl 3 [19] versus δ(Se: 1) SCS, 213 K and the correlation is given in Table 2 (entry 17: r = 0.997). The correlation coefficient is excellent when δ(Se: 5) SCS measured in neat is plotted versus δ(Se: 1) SCS, 213 K (entry 18 in Table 2: r = 0.999). These observations must be the results of the Se−C Me bond in 5 being on the p-YC 6 H 4 plane in solutions for all Y examined, under the conditions. On the other hand, δ(Se: 7) SCS do not correlate with δ(Se: 1) SCS, 213 K . Instead, they correlate well with δ(Se: 2) SCS, 213 K (entry 19 in Table 2: r = 0.995). Figure 6 shows the plot. The results are rationally explained by assuming that the Se−C O bond in 7 is perpendicular to the p-YC 6 H 4 plane in solutions for all Y examined, under the conditions. δ(Se) SCS of 6 [19] and 8 [8] are similarly plotted versus δ(Se: 1) SCS, 213 K . They give good correlations, although the r values become poorer relative to that for 5 (entries 20 and 21 (a) Structures are optimized with the 6-311+G(3df) basis sets for Se and 6-311+G(3d,2p) basis sets for other nuclei at the DFT (B3LYP) level, assuming pl and pd for each of Y [47]. σ(Se) are calculated based on the DFT-GIAO method with the same methods.
in Table 2). The reason would be the equilibrium of pl with pd for some Y in 6 and 8, may be Y of donors. δ(Se) SCS of 9 are also plotted versus δ(Se: 1) SCS, 213 K . The correlations are excellent (entry 22 in Table 2: r = 0.999). It is worthwhile to comment that the energy lowering effect by Se 4 4c-6e in 9 fixes the conformation 9 (pl, pl) for both p-YC 6 H 4 Se groups in solutions for all Y examined, under the conditions [50].
It is demonstrated that sets of δ(Se: 1) and δ(Se: 2) proposed in this work can be the standards for pl and pd, respectively, when δ(Se) of aryl selenides are analyzed based on the orientational effect.

CONCLUSION
The orientational effect is empirically established by the Y dependence on δ(Se: 1) and δ (Se: 2). The Y dependence observed in 1 and 2 is demonstrated by σ t (Se) calculated for 4-6 with the DFT-GIAO method. While σ t (Se) of 4a (pl) is predicted to be more negative than that of 4a (pd) by 46 ppm, σ t (Se) of 5a (pl) is evaluated to be larger than that of 5a (pd) by 49 ppm, which corresponds to the orientational effect by the Ph group in 4a and 5a, respectively. Excellent to good correlations are obtained in the plots of σ p (Se) versus (σ p (Se) xx +σ p (Se) yy ) for 4-6 in pl and pd. It is demonstrated that (σ p (Se) xx + σ p (Se) yy ) effectively controls σ p (Se) of 4-6 in pl and pd.
of 4p z (Se) in n p (Se) with 4p x (Se) and 4p y (Se) in modified σ * (CSeX) since n p (Se) of 4p z (Se) is filled with electrons. Therefore, Y of both donors and acceptors are effective in pl, whereas Y of donors are more effective in pd. The expectations are just observed in 1 and 2. Sets of δ(Se: 1) and δ(Se: 2) can be used as the standards for pl and pd, respectively, when δ(Se) of aryl selenides are analyzed. performed on 4-6 in pl and pd at the density functional theory (DFT) level of the Becke three parameter hybrid functionals combined with the Lee-Yang-Parr correlation functional (B3LYP). Absolute magnetic shielding tensors of Se nuclei (σ(Se)) are calculated based on the gaugeindependent atomic orbital (GIAO) method, applying on the optimized structures with the same method.  Structures of 1a-3a in various conformers are also optimized, containing frequency analysis, with the B3LYP/6-311+G(d,p) method.

ACKNOWLEDGMENT
This work was partially supported by a Grant-in-Aid for Scientific Research (no 16550038) from the Ministry of Education, Culture, Sports, Science, and Technology, Japan.