Chemical Physics Letters 393 (2004) 5157

www.elsevier.com/locate/cplett

A new hybrid exchangecorrelation functional using the

Coulomb-attenuating method (CAM-B3LYP)
a,*
Takeshi Yanai , David P. Tew b, Nicholas C. Handy b

a
Oak Ridge National Laboratory, P.O. Box 2008, MS6367, Oak Ridge, TN 37831, USA
b
Department of Chemistry, University of Cambridge, Cambridge, CB2 1EW, UK
Received 15 March 2004; in nal form 4 June 2004

Abstract

A new hybrid exchangecorrelation functional named CAM-B3LYP is proposed. It combines the hybrid qualities of B3LYP and
the long-range correction presented by Tawada et al. [J. Chem. Phys., in press]. We demonstrate that CAM-B3LYP yields atom-
ization energies of similar quality to those from B3LYP, while also performing well for charge transfer excitations in a dipeptide
model, which B3LYP underestimates enormously. The CAM-B3LYP functional comprises of 0.19 HartreeFock (HF) plus 0.81
Becke 1988 (B88) exchange interaction at short-range, and 0.65 HF plus 0.35 B88 at long-range. The intermediate region is smoothly
described through the standard error function with parameter 0.33.
2004 Elsevier B.V. All rights reserved.

1. Introduction interaction. The key in the long-range-corrected (LC)

exchange functional scheme for DFT [1013] is that the
In density functional theory (DFT), as it is used for DFT exchange interaction is included using the rst
computational chemistry, the hybrid functional B3LYP term (short-range), and the long-range orbitalorbital
[2,3] appears to oer the greatest contribution (if mea- exchange interaction is described with the Hartree
sured by the number of applications which have been Fock (HF) exchange integral via the complementary
published). However it is unsuccessful in a number of term. Specically, the short-range part of the exchange
important applications: (i) the polarizability of long interaction is incorporated by modifying P the
R usual ex-
3
chains, (ii) excitations using time dependent theory change functional form, Ex
1=2 r q4=3 r Kr d r,
(TDDFT) [46] for Rydberg states, and perhaps most into
important (iii) charge transfer (CT) excitations [79]. Z
1X
The reason for these failures is understood, at long- Exsr
q4=3
r Kr
range the exchange potential behaves as
0:2r
1 , in- 2 r

   
stead of the exact value
r
1 . Even so, the potential is 8 p 1
 1
ar perf 2ar br
cr d3 r;
an improvement over that LDA and BLYP, where there 3 2ar
is no r
1 dependence in the potential. 2
Recently Tsuneda and co-workers [1] have overcome

1 where ar , br and cr are:
this deciency through an Ewald split of r12 into
1 1
erf lr12 erf lr12 lK 1=2
: 1 ar pr 1=3 ; 3
r12 r12 r12 6 pqr
The rst term accounts for the short-range interac-  
1
tion, and the second term accounts for the long-range br exp
2
1; 4
4ar
*
Corresponding author. Fax: +1-865-574-0680.
E-mail address: yanait@ornl.gov (T. Yanai). cr 2a2r br 12: 5

0009-2614/$ - see front matter 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.cplett.2004.06.011
52 T. Yanai et al. / Chemical Physics Letters 393 (2004) 5157

To reach this form, erfclr12 has been multiplied by

the square of the one-particle density matrix for the
uniform electron gas, and then integrated [12]. Use of Kr
allows the incorporation of generalized gradient ap-
proximation functionals [10]. The long-range part of the
exchange interaction is expressed with the HF exchange
integral [12],
occ Z Z
1XX occ X
lr
Ex
wir r1 wjr r1
2 r i j

erflr12
 wir r2 wjr r2 d3 r1 d3 r2 ; 6
r12
where wir is the ith r-spin molecular orbital. The pa- Fig. 1. Plots for f r 1
erflr and f r 1
a berflr.
rameter l determines the balance of DFT to HF ex-
change at intermediate r12 . If l 0, the LC DFT where the relations 0 6 a b 6 1, 0 6 a 6 1, and 0 6 b 6 1
calculation corresponds to the pure (non-LC) DFT should be satised. We term this the Coulomb-attenu-
calculation, and conversely l 1 corresponds to the ating method (CAM) approach. Fig. 1 illustrates the
standard HF calculation. Tsuneda and co-workers have schematic plots of two functions, Eqs. (1) and (7). The
demonstrated that their LC method for the Becke 1988 parameter a allows us to incorporate the HF exchange
[14] or PBE [15] generalized-gradient approximation contribution over the whole range by a factor of a, and
(GGA) exchange functionals in conjunction with their the parameter b allows us to incorporate the DFT
one-parameter progressive (OP) correlation functional counterpart over the whole range by a factor of
[16] can address the above notorious problems in DFT. 1
a b. We note that the widely-used hybrid B3LYP
The incorrect long-range exchange interaction delivered functional [2,3] takes CAM potential partitioning of Eq.
by the standard DFT exchange functionals seems to lead (7) with a 0:2 and b 0:0 for the mixing of Slater ex-
to the underestimation of 4s3d interaction energies of change EXSlater and the HF exchange EXHF as follows:
the rst-row transition metals, the overestimation of the
EXB3 1
aEXSlater aEXHF cB88 DEXB88 ; 8
longitudinal polarizabilities of p-conjugated polyenes,
the poor description of the weak interaction of van der where the additional term DEXB88
is Beckes 1988 gradient
Waals bonding of rare-gas dimers, and the underesti- correction for exchange [14] with the semiempirical pa-
mations of Rydberg excitation energies, oscillator rameter cB88 0:72, which Becke obtained by a linear
strengths, and charge-transfer excitation energies. least-square t to experimental data [2]. Also, the ori-
Tsuneda and co-workers have shown that their LC ginal LC corresponds to the CAM with a 0:0 and
method with l 0:33 gives greatly improved results for b 1:0. Figs. 2ac show the contributions to exchange

1
all of these phenomena [1,10,17]. from r12 , apportioned into DFT and HF, for B3LYP,
Unfortunately, we nd that LC-BOP does not work LC and CAM methodologies.
well for the more standard energy calculations on which The extra exibility arising from two extra parame-
the parameters for B3LYP were derived. Indeed, the ters a and b allows us to look at how important the HF
mean absolute error in the atomization energies of 53 exchange contribution is for the short-range region and
molecules with the high quality basis sets, augmented cc- the DFT counterpart is for the long-range region. In the
pVQZ, increases from 2.5 kcal mol
1 (B3LYP) to original form of the LC decomposition (Eq. (1)), either
9.5 kcal mol
1 (LC-BOP), which seriously detracts from HF or DFT exchange vanishes at r 0 and r 1.
the quality of LC-BOP. The purpose of this Letter is to We note that CAM with the Gaussian-type basis
show how it is possible to combine the ideas behind implementation requires the same types of the Cou-
B3LYP and LC-BOP to deliver a functional which has lomb-attenuated and non-attenuated (standard) two-
the energetic qualities of B3LYP and the asymptotic electron integrals [12] as the original LC approach of
qualities of LC-BOP. Eq. (1). The details about modifying the DFT exchange
functionals and the HF exchange integral to involve the
error function are described in [1,10].
2. Coulomb-attenuating method with three parameters

Now, we generalize the form of Eq. (1) using two 3. Procedure: trial CAM exchangecorrelation functionals
extra parameters a and b as,
In this Letter, we investigate the performance of the
1 1
a b  erf lr12  a b  erf lr12 Coulomb-attenuating method with several existing
; 7
r12 r12 r12 GGA functionals. Table 1 summarizes the exchange
 T. Yanai et al. / Chemical Physics Letters 393 (2004) 5157 53

(a) B3LYP (=0.2, +=0.2) (b) LC ( =0.0, +=1.0) (c) CAM (=0.2, +=0.6)
0 0 0
HF
0.2 0.2 0.2

+
HF
0.4 0.4 HF 0.4

0.6 DFT 0.6 0.6

0.8 DFT 0.8 DFT

0.8

1 1 1
0 r12
0 r12 0 r12

1
Fig. 2. Schematic plots of the contributions to exchange from r12 , apportioned into DFT and HF, for: (a) B3LYP, (b) LC, and (c) CAM.

correlation functionals examined in this study. The We compare the present functionals with four kinds
Becke 1988 exchange functional is used in all of the LC of the widely used, well-examined exchangecorrelation
and CAM functionals, and is mixed with the HF ex- functionals, HCTH/93 [22], BLYP, B3LYP(G) (VWN1),
change according to Eq. (7). For the partner correlation and B3LYP (VWN5). We used the IN T E G R A [23] as a
functionals, we use the OP correlation functional, the part of the UT C H E M 2004 program package [24,25] to
LeeYangParr (LYP) [19], and the correlation func- carry out KohnSham self-consistent eld (KS-SCF)
tional employed in B3LYP, which is 0.19 VWN5 + 0.81 calculations with the LC and CAM methods. The KS-
LYP, where the VWN5 functional is the local correla- SCF calculations with the standard BLYP, HCTH,
tion functional of Vosko, Wilk and Nusair (VWN) [20] B3LYP(G), and B3LYP were performed using
parameterized with the data of Ceperley and Alder [21]. NW C H E M program package version 4.5 [26].
Note that this is dierent to the standard B3LYP im-
plemented in GA U S S I A N which uses VWN1 instead of
VWN5 [18], we refer to this functional as B3LYP(G). 4. Results
The possible combinations of the exchangecorrelation
functionals are termed CAM-BOP, CAM-BLYP, CAM- 4.1. Atomization energies, ionization potentials, and
B3LYP, LC-BOP, and LC-BLYP. For the parameter l, atomic energies
the same value is used as in Tawadas study [1],
l 0:33. The parameter a, which determines the con- We calculated 53 atomization energies and 22 ioni-
tribution of the HF exchange at the short-range region, zation potentials from the molecules of the G2 set
was chosen to be 0.2 for the three functionals, CAM- [27,28]. All calculations were performed with suciently
BOP, CAM-BLYP, and CAM-B3LYP. We vary the accurate correlation-consistent aug-cc-pVQZ Gaussian
parameter b so that the HF exchange could contribute basis sets. Tables 2 and 3 show the statistical data for
to the long-range region with a b 0:6; 0:8; or 1:0 for atomization energies and ionization potentials with
three functionals. comparison to the experimental data, which are taken

Table 1
Summary of the exchangecorrelation functionals
Name Exchange functional a ab Additional exchange Correlation functional
LC-BOP Becke88 0.0 1.0 OP
LC-BLYP Becke88 0.0 1.0 LYP
CAM-BOP Becke88 0.2 1.0 OP
0.8
0.6
CAM-BLYP Becke88 0.2 1.0 LYP
0.8
0.6
CAM-B3LYP Becke88 0.2 1.0 0.19 VWN5 + 0.81 LYP
0.8
0.6
B3LYP(G) Slater 0.2 0.2 0.72 DBecke88 0.19 VWN1(RPA) + 0.81 LYP
B3LYP Slater 0.2 0.2 0.72 DBecke88 0.19 VWN5 + 0.81 LYP
BLYP Becke88 0.0 0.0 LYP
HCTH xHCTH 0.0 0.0 cHCTH
54 T. Yanai et al. / Chemical Physics Letters 393 (2004) 5157

along with zero point energies from [2,29]. Table 4 the total atomic energies of HNe with an RMS of
summarizes the atomic energy data of the rst-row 0.090.13 Eh .
atoms H through Ne with comparison to the exact As for the CAM exchangecorrelation functionals,
energies [2]. the errors of the atomization energies are signicantly
Within the standard exchangecorrelation function- improved compared to the LC method by setting
als, B3LYP(G) gives the smallest errors for the atom- a 0:2 for CAM-BOP, CAM-BLYP and CAM-
ization energies with 2.54 kcal mol
1 for the mean B3LYP. For ionization energies, the CAM functionals
absolute error (MAE) and 3.38 kcal mol
1 for the root are also better than the LC functionals. The CAM-BOP
mean square deviation (RMS). The HCTH functional and CAM-BLYP functionals systematically overesti-
gives the best performance for the ionization potentials mate the total atomic energies of HNe with a MAE of
within the standard functionals with 0.154 eV for MAE 0.0690.030 Eh . The inclusion of the VWN5 correlation
and 0.187 eV for RMS. B3LYP performs best for the contribution into the LYP functional greatly reduces
atomization energies with 0.005Eh for both MAE and these errors to within an acceptable level. However, the
RMS. atomization energy errors deteriorate by 0.10.3
The LC-based exchangecorrelation functionals, LC- kcal mol
1 with the inclusion of VWN5. We consider
BOP and LC-BLYP, enormously overestimate the at- that due to the prevalence of hydrogen in chemical
omization energies with MAEs of 67 kcal mol
1 and systems, the good reproduction of the total atomic en-
RMS errors of 89 kcal mol
1 , while they perform ergies, hydrogen in particular, is more important than
comparably well for computing ionization potentials. the slight loss in quality of the atomization and ioniza-
Both of the LC functionals systematically overestimate tion energies.

Table 2
Statistical data for atomization energies (kcal mol
1 ) of the small G2 set (53 molecules) with aug-cc-pVQZ Gaussian basis sets
a ab MAE RMS Maximum deviation (
) Maximum deviation (+)
LC-BOP 0.0 1.0 9.19 11.22 )4.37 (Li2 ) 25.47 (CO2 )
LC-BLYP 0.0 1.0 9.48 12.07 )3.23 (H2 ) 29.64 (CO2 )
CAM-BOP 0.2 1.0 4.13 5.01 )8.98 (CN) 10.34 (N2 H4 )
0.8 3.21 4.05 )9.33 (SO2 ) 8.62 (BeH)
0.6 2.93 3.95 )13.12 (SO2 ) 8.58 (BeH)
CAM-BLYP 0.2 1.0 3.90 4.77 )7.00 (P2 ) 13.66 (N2 H4 )
0.8 2.77 3.51 )6.95 (CS) 9.96 (N2 H4 )
0.6 2.28 3.15 )10.21 (SO2 ) 7.60 (BeH)
CAM-B3LYP 0.2 1.0 3.99 4.87 )8.29 (P2 ) 13.37 (N2 H4 )
0.8 2.97 3.79 )9.03 (SO2 ) 9.68 (N2 H4 )
0.6 2.62 3.59 )12.82 (SO2 ) 8.33 (BeH)
B3LYP(G) 0.2 0.2 2.54 3.38 )12.01 (SO2 ) 8.33 (BeH)
B3LYP 0.2 0.2 2.68 3.58 )13.48 (SO2 ) 8.21 (BeH)
BLYP 0.0 0.0 4.51 5.87 )10.42 (C2 H6 ) 15.39 (O2 )
HCTH 0.0 0.0 3.50 4.76 )9.59 (C2 H6 ) 14.96 (O2 )

Table 3
Statistical data for ionization potentials (eV) of the small G2 set (22 molecules) with aug-cc-pVQZ Gaussian basis sets
a ab MAE RMS Maximum deviation (
) Maximum deviation (+)
LC-BOP 0.0 1.0 0.192 0.220 )0.312 (Be) 0.386 (F)
LC-BLYP 0.0 1.0 0.230 0.272 )0.245 (Be) 0.552 (O)
CAM-BOP 0.2 1.0 0.166 0.203 )0.337 (Be) 0.488 (N2 )
0.8 0.158 0.189 )0.358 (Be) 0.367 (N2 )
0.6 0.153 0.185 )0.375 (Be) 0.291 (O2 )
CAM-BLYP 0.2 1.0 0.195 0.245 )0.272 (Be) 0.558 (N2 )
0.8 0.185 0.225 )0.291 (Be) 0.446 (O2 )
0.6 0.178 0.213 )0.311 (Be) 0.408 (O2 )
CAM-B3LYP 0.2 1.0 0.243 0.301 )0.179 (Be) 0.658 (N2 )
0.8 0.218 0.273 )0.199 (Be) 0.537 (N2 )
0.6 0.202 0.249 )0.223 (P2 ) 0.499 (O2 )
B3LYP(G) 0.2 0.2 0.209 0.254 )0.210 (Be) 0.532 (O)
B3LYP 0.2 0.2 0.169 0.206 )0.293 (Be) 0.423 (O)
BLYP 0.0 0.0 0.208 0.250 )0.459 (Cl2 ) 0.544 (O)
HCTH 0.0 0.0 0.154 0.187 )0.359 (Cl2 ) 0.340 (N)
 T. Yanai et al. / Chemical Physics Letters 393 (2004) 5157 55

Table 4
Statistical data for total atomic energies (hartree) of HNe with aug-cc-pVQZ Gaussian basis sets
a ab MAE RMS Maximum deviation (
) Maximum deviation (+) Deviation H
LC-BOP 0.0 1.0 0.084 0.093 0.134 (Ne) 0.0091
LC-BLYP 0.0 1.0 0.085 0.128 0.128 (Ne) 0.0091
CAM-BOP 0.2 1.0 0.068 0.075 0.112 (Ne) 0.0075
0.8 0.049 0.054 0.080 (Ne) 0.0061
0.6 0.030 0.049 0.049 (Be) 0.0047
CAM-BLYP 0.2 1.0 0.069 0.075 0.107 (Ne) 0.0075
0.8 0.050 0.054 0.075 (Ne) 0.0061
0.6 0.031 0.034 0.043 (Ne) 0.0047
CAM-B3LYP 0.2 1.0 0.033 0.035 0.045 (Be) 0.0034
0.8 0.014 0.017 0.029 (Be) 0.0019
0.6 0.009 0.013 )0.025 (Ne) 0.013 (Be) 0.0005
B3LYP(G) 0.2 0.2 0.019 0.023 )0.040 (Ne) )0.0024
B3LYP 0.2 0.2 0.005 0.005 )0.006 (F) 0.009 (Be) 0.0010
BLYP 0.0 0.0 0.011 0.016 )0.031 (F) 0.007 (Be) 0.0021
HCTH 0.0 0.0 0.008 0.010 )0.019 (Ne) 0.004 (C) )0.0064

Increasing the ratio of the Becke 88 exchange con- a 0:19 and b 0:46; a b 0:65: 9
tribution at long-range reduces the errors of the atom-
ization, ionization and the total atomic energies. For This functional yields results comparable to those of
instance, the CAM-BLYP with a 0:2 and a b 0:6 B3LYP for the atomization energies, ionization poten-
is better by 1.62 kcal mol
1 for MAE and RMS than tials and total atomic energies (Table 5). The sum a b
that with a 0:2 and a b 1:0. The maximum devi- is rather crucial, because it gives the asymptotic coe-
ations are also reduced. While the CAM-BLYP method cient of
r
1 . If we set a b 0:8, (i.e., 0.15 more
with a 0:2 and a b 0:6 gives the smallest errors for HF exchange at the long-range) we nd optimal values
the atomization energies of the three CAM functionals to be
with a MAE of 2.28 kcal mol
1 , and CAM-BOP with a 0:23 and b 0:57; 10
a 0:2 and a b 0:6 gives the best ionization ener-
gies with an MAE of 0.153 eV, we recommend CAM- for CAM-B3LYP. In Section 4.2, we use these two sets
B3LYP due to its success in reproducing total atomic of a and b to investigate how the long-range HF ex-
energies. change interaction eects the charge transfer excitations
We nd that the optimal values for the two param- in TDDFT calculations. Table 5 summarizes the atom-
eters a and b of CAM-B3LYP which yield the smallest ization energies, ionization potentials, total atomic en-
errors for the atomization energies, are found to be ergies with the above two forms of CAM-B3LYP. Both

Table 5
Statistical data for atomization energies (kcal mol
1 ) (with 53 molecules), ionization potentials (eV) (with 22 atoms and molecules), and total atomic
energies (hartrees) (with H atom through Ne atom) from the small G2 set using CAM-B3LYP with a b 0:65; 0:8, compared to B3LYP and LC-
BLYP with aug-cc-pVQZ Gaussian basis set
Name a ab MAE RMS Maximum deviation (
) Maximum deviation (+)
Atomization energy (kcal mol
1 )
CAM-B3LYP 0.19 0.65 2.53 3.46 )10.79 (SO2 ) 8.34 (BeH)
0.23 0.8 2.91 3.88 )12.25 (SO2 ) 8.41 (BeH)
B3LYP 0.2 0.2 2.68 3.58 )13.48 (SO2 ) 8.21 (BeH)
LC-BLYP 0.0 1.0 9.48 12.07 )3.23 (H2 ) 29.64 (CO2 )
Ionization potentials (eV)
CAM-B3LYP 0.19 0.65 0.208 0.256 )0.212 (Be) 0.503 (O2 )
0.23 0.8 0.213 0.271 )0.203 (Be) 0.563 (N2 )
B3LYP 0.2 0.2 0.169 0.206 )0.293 (Be) 0.423 (O)
LC-BLYP 0.0 1.0 0.230 0.272 )0.245 (Be) 0.552 (O)
Total atomic energies (hartree)
CAM-B3LYPa 0.19 0.65 0.009 0.011 )0.017 (F) 0.018 (Be)
0.23 0.8 0.012 0.015 0.026 (Be)
B3LYP 0.2 0.2 0.005 0.005 )0.006 (F) 0.009 (Be)
LC-BLYP 0.0 1.0 0.085 0.128 0.128 (Ne)
a
Deviation of H atom from the exact atomic energy (hartrees): 0.0010 (CAM-B3LYP a 0:19, a b 0:65), 0.0017 (CAM-B3LYP a 0:23,
a b 0:80).
56 T. Yanai et al. / Chemical Physics Letters 393 (2004) 5157

forms are similar in quality to B3LYP, and are a great

improvement over LC-BLYP.

CAM-B3LYP

5.72 (0.0008)
5.95 (0.0004)
8.58 (0.0001)
7.77 (0.0816)
7.43 (0.3572)
7.24 (0.0399)
a b 0:8
a 0:23,
4.2. Charge transfer excitations in the dipeptide model

Excited state calculations with TDDFT were per-

formed on a glycine dipeptide model system, which was
previously studied by Tozer and co-workers [7]. We used
CAM-B3LYP

5.65 (0.0008)
5.88 (0.0004)
7.88 (0.0001)
7.52 (0.0869)
7.32 (0.3555)
6.94 (0.0184)
a b 0:65

the same geometry with the same basis sets TZ2P

a 0:19,

(5s4p2d on rst row; 3s2p on H) as used in Tozers

study. Table 6 lists the excitation energies and the os-
cillator strengths. It is to be noted that the n ! p ex-
citations within the same fragment are well produced by
all the DFT methods, when compared to the highly
5.56 (0.0007)
5.80 (0.0003)
8.38 (0.0003)
7.57 (0.0830)
7.26 (0.3437)
7.02 (0.0430)
accurate multireference perturbation calculation with
LC-BLYP

CASPT2 [30].
The charge transfer excitations of n ! p and p ! p
Yanai et al. (this work)

between dierent fragments are very poorly predicted by

the DFT calculations with the standard exchangecor-
relation functionals, in particular the BLYP for n ! p
5.34 (0.0009)
5.62 (0.0005)
4.60 (0.0002)

6.82 (0.2397)
5.07 (0.0041)

is much lower by 3.3 eV than CASPT2. B3LYP yields

values 1.7 eV lower than CASPT2. This means that the
BLYP

0:2r
1 contribution via HF exchange in B3LYP im-

The electronic spectrum (eV) of a dipeptide (C5 N2 O2 H10 ) with the basis sets TZ2P (5s4p2d on rst row; 3s2p on H)

proves the charge transfer excitations. Tozer and Han-

dys asymptotic correction does not work for these
states.
5.49 (0.0013)
5.73 (0.0009)
6.24 (0.0000)
7.19 (0.0645)
7.23 (0.2704)
6.06 (0.0029)

It is very encouraging that the CAM-B3LYP with

a 0:19 and a b 0:65 predictions of charge transfer
B3LYP

excitations are in excellent agreement with the CASPT2

results. It is not surprising that the LC-BLYP method
brings the charge transfer excitations into much better
agreement with CASPT2 than the standard DFT ex-
5.43 (0.0009)
5.70 (0.0005)
4.67 (0.0000)
6.98 (0.0238)
6.87 (0.2863)
5.16 (0.0043)

changecorrelation functionals, because of the 1:0r
1

AC-HCTH

asymptotic form through the HF exchange. The CAM-

B3LYP functional with a 0:23 and a b 0:8 yields
almost comparable results to LC-BLYP. By comparison
between the two sets of results obtained with CAM-
Tozer et al. [7]

B3LYP, we see that the long-range contribution of the

5.61 (0.0006)
5.82 (0.0006)
7.92 (0.0000)
6.32 (0.2737)
6.29 (0.3205)
6.92 (0.1168)

HF exchange interaction has a large impact on the

CASPT2

charge transfer excitations as compared with the other

calculated excitations.
The predictions for the p ! p excitation within the
same fragment are overestimated by all the DFT
methods. A clue to the problem lies in the oscillator
Oscillator strengths within parentheses.

strengths. In DFT p2 ! p2 is dominant, whereas

Band

NV1
NV2

p1 ! p1 and p2 ! p2 have equal strengths in CASPT2.

CT

CT
W1
W2

If the problem is truly multicongurational, then DFT

will fail. This is the most plausible explanation.
p1 ! p1
p2 ! p2
p1 ! p2
n1 ! p1
n2 ! p2
n1 ! p2

5. Conclusions

The recent work by Tsuneda and co-workers [1,10]

has shown how DFT studies can now be performed to a
Table 6

A00
1 00
1 00

useful accuracy for polarizability of long chains, exci-

A0
A0
1 0
A
A

A
1

1
1

tations to Rydberg states and charge transfer excita-

 T. Yanai et al. / Chemical Physics Letters 393 (2004) 5157 57

tions. Previously DFT calculations using GGA func- [3] P.J. Stephens, J.F. Devlin, C.F. Chabalowski, M.J. Frisch,
tionals (BLYP) and hybrid functionals (B3LYP) have J. Chem. Phys. 98 (1993) 11623.
[4] M.E. Casida, in: D.P. Chong (Ed.), Recent Advances in Density
been frequently used for energetic studies and structural Functional Methods, vol. 1, World Scientic, Singapore, 1995.
studies. A combination of the two clearly goes a long [5] R. Bauernschmitt, R. Ahlrichs, Chem. Phys. Lett. 256 (1996)
way for the provision of a very useful computational 454.
chemistry methodology. [6] D.J. Tozer, N.C. Handy, J. Chem. Phys. 109 (1998) 10180.
We are suggesting that the CAM-B3LYP functional [7] D.J. Tozer, R.D. Amos, N.C. Handy, B.O. Roos, L. Serrano-
Andres, Mol. Phys. 97 (1999) 859.
presented in this Letter meets these criteria. In the [8] A. Dreuw, J.L. Weisman, M. Head-Gordon, J. Chem. Phys. 119
CAM-B3LYP functional, we have replaced the Becke (2003) 2943.
parameter a by two parameters a, b for mixing Becke [9] L. Bernasconi, M. Sprik, J. Hutter, J. Chem. Phys. 119 (2003)
1988 exchange and HF exchange, with l describing the 12417.
conversion from one to the other through Eq. (7). Our [10] H. Iikura, T. Tsuneda, T. Yanai, K. Hirao, J. Chem. Phys. 115
(2001) 3540.
best functional uses a 0:19, a b 0:65 and l 0:33 [11] A. Savin, in: J.M. Seminario (Ed.), Recent Developments and
(Tawadas value). We used Beckes VWN5/LYP mixing Applications of Modern Density Functional Theory, Elsevier,
parameter ( 0.19) for the correlation functional with- Amsterdam, 1996.
out adjustment. Our calculations have shown that this [12] P.M.W. Gill, R.D. Adamson, J.A. Pople, Mol. Phys. 88 (1996)
functional predicts energetic quantities to the accuracy 1005.
[13] T. Leiniger, H. Stoll, H.-J. Werner, A. Savin, Chem. Phys. Lett.
of B3LYP. Our rst investigations on charge transfer 275 (1997) 151.
energies suggest that it is possible to achieve chemical [14] A.D. Becke, Phys. Rev. A 38 (1988) 3098.
accuracy (0.1 eV). The cost for the implementation of [15] J.P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 77 (1996)
this functional is no more severe than that required for 3865.
B3LYP (within a factor of 2 for computing two-electron [16] T. Tsuneda, T. Suzumura, K. Hirao, J. Chem. Phys. 110 (1999)
10664.
integrals). This functional is a hybrid functional with [17] M. Kamiya, T. Tsuneda, K. Hirao, J. Chem. Phys. 117 (2002)
improved long-range properties. 6010.
[18] R.H. Hertwig, W. Koch, Chem. Phys. Lett. 268 (1997) 345.
[19] C. Lee, W. Yang, R.G. Parr, Phys. Rev. B 37 (1988) 785.
Acknowledgements [20] S.J. Vosko, L. Wilk, M. Nusair, Can. J. Phys. 58 (1980) 1200.
[21] D.M. Ceperley, B.J. Alder, Phys. Rev. Lett. 45 (1980) 566.
[22] F.A. Hamprecht, A.J. Cohen, D.J. Tozer, N.C. Handy, J. Chem.
We acknowledge Dr. R.J. Harrison for the encour- Phys. 109 (1998) 6264.
agement of this work, the use of the computer resources [23] H. Iikura, H. Inoue, M. Kamiya, Y. Kawashima, T. Nakajima, Y.
at ORNL, and arranging TYs visit to NCHs group at Nakao, K. Nakayama, H. Sekino, K. Sorakubo, T. Tawada, T.
Cambridge. We would like to thank Professor K. Hirao, Tsuneda, K. Yagi, S. Yanagisawa, T. Yanai, K. Hirao, IN T E G R A ,
Dr. T. Tsuneda, Y. Tawada and Dr. M. Sprik for Department of Applied Chemistry, School of Engineering, Uni-
versity of Tokyo, Japan.
valuable discussions, Dr. D.J. Tozer for providing us [24] T. Yanai, M. Kamiya, Y. Kawashima, T. Nakajima, H. Nakano,
with input les for the dipeptide, Dr. A.J. Cohen for Y. Nakao, H. Sekino, J. Paulovic, T. Tsuneda, S. Yanagisawa, K.
input les of G2 set molecules. TY thanks Dr. S. Hirata Hirao, UT C H E M 2004, Department of Applied Chemistry, School
for teaching TY the implementation of TDDFT, and M. of Engineering, University of Tokyo, Japan.
Kamiya for his DFT quadrature code in IN T E G R A of [25] T. Yanai, H. Nakano, T. Nakajima, T. Tsuneda, S. Hirata, Y.
Kawashima, Y. Nakao, M. Kamiya, H. Sekino, K. Hirao, in:
UT C H E M . NW C H E M Version 4.5, as developed and Computational Science ICC 2003, Lecture Notes in Computer
distributed by Pacic Northwest National Laboratory, Science, Springer, 2003, p. 84.
P.O. Box 999, Richland, Washington 99352 USA, fun- [26] High Performance Computational Chemistry Group, NW C H E M ,
ded by the US Department of Energy, was used to ob- A Computational Chemistry Package for Parallel Computers,
tain the DFT results. Version 4.5, Pacic National Laboratory, Richland, Washington
99352, USA, 2003.
[27] L.A. Curtiss, K. Raghavachari, P.C. Redfern, J.A. Pople,
J. Chem. Phys. 106 (1997) 1063.
References [28] N.C. Handy, D.J. Tozer, Mol. Phys. 94 (1998) 707, References of
molecular bond lengths and bond angles therein.
[1] Y. Tawada, T. Tsuneda, S. Yanagisawa, T. Yanai, K. Hirao, [29] D. Feller, K.A. Peterson, J. Chem. Phys. 110 (1999) 8384.
J. Chem. Phys. 120, in press. [30] K. Andersson, P.-A.  Malmqvist, B.O. Roos, A.J. Sadlej, K.
[2] A.D. Becke, J. Chem. Phys. 98 (1993) 5648. Wolinski, J. Phys. Chem. 94 (1990) 5483.