light energy conversion

Coordination Chemistry Reviews 375 (2018) 514–557

Contents lists available at ScienceDirect

Coordination Chemistry Reviews journal homepage: www.elsevier.com/locate/ccr

Review

Recent advances in copper complexes for electrical/light energy conversion Yurong Liu b, Sze-Chun Yiu a,b, Cheuk-Lam Ho a,⇑, Wai-Yeung Wong a,b,⇑ a b

Department of Applied Biology and Chemical Technology, The Hong Kong Polytechnic University, Hong Hom, Hong Kong, China Institute of Molecular Functional Materials and Department of Chemistry, Hong Kong Baptist University, Waterloo Road, Kowloon Tong, Hong Kong, China

a r t i c l e

i n f o

Article history: Received 6 November 2017 Accepted 10 May 2018 Available online 28 June 2018

a b s t r a c t A great deal of research effort has been put in green energy applications in the past few decades based on organic optoelectronics. Compared with conventional inorganic semiconductors, organic counterparts offer a much simpler strategy for low-cost mass production and structural modification. Hence, continuous and intensive academic and industrial research works have been done in these areas. In terms of the materials used, transition-metal complexes with the unique features of the transition metal centers represent a large group of candidates, showing high performance in energy conversion technologies. However, the commonly used transition metals, like Pt(II), Ir(III) and Ru(II), are expensive and of relatively low abundance. Concerning elemental sustainability and marketability, some abundant and cheaper metals should be investigated and further developed to replace these precious metals. Cu(I) complexes have shown their potentiality in solar energy harvesting and light emitting applications, due to their well-studied photophysics and structural diversity. In addition, copper is one of the earth-

Abbreviations: Alq3, tris(8-hydroxyquinolinato) aluminum(III); BCP, bathocuproine; BHJSCs, bulk heterojunction solar cells; BAlq-13, bis(2-methyl-quinolin-8-olato)(2,6diphenylphenolato)aluminum(III); Bphen, bathophenanthroline; bpy, 2,20 -bipyridine; bq, 2,20 -biquinoline; CAAC, cyclic alkyl(amino)carbene; CBP, 4,40 -bis(carbazol-9-yl) biphenyl; CDCA, chenodeoxycholic acid; CEmax, maximum current efficiency; CIE, Commission internationale de l’éclairage; CPPyC, 3-(carbazol-9-yl)-5-((3-carbazol-9-yl) phenyl)pyridine; CPy, 3-(carbazol-9-yl)pyridine; CPzPC, 9-(3-(6-(carbazol-9-yl)pyrazin-2-yl)phenyl)carbazole; CRI, color rendering index; CT, charge transfer; CuPc, copper (II)phthalocyanine; CzBPDCb, 5-(30 -(9H-carbazol-9-yl)-(1,10 -biphenyl)-3-yl)-5H-pyrido[3,2-b]indole; czpzpy, 2-(9H-carbazolyl)-6-(1H-pyrazolyl)pyridine; CzSi, 9(4-tert-butylphenyl)-3,6-bis(triphenylsilyl)-9H-carbazole; DFT, density functional theory; dmp, 2,9-dimethyl-1,10-phenanthroline; dpbq, 4,40 -diphenyl-2,20 -biquinoline; DPEphos, bis[2-(diphenylphosphino)phenyl]ether; DPEPO, bis[2-(diphenylphosphino)phenyl]ether oxide; dppb, 1,2-bis(diphenylphosphino)benzene; dppm, bis (diphenylphosphino)methane; DPVBi, 4,40 -bis(2,20 -diphenylvinyl)-1,10 -biphenyl; DSSCs, dye-sensitized solar cells; D-p-A, donor-p-acceptor; EBL, electron-blocking layer; EL, electroluminescent/electroluminescence; EQEmax, maximum external quantum efficiency; FB9ox, 1-methyl-2-[2,4,6-tris(9-oxiranyl-nonyloxy)phenyl]fulleropyrrolidine; FF, fill factor; FIrpic, bis[2-(4,6-difluorophenyl)pyridinato-C2,N](picolinato)iridium(III); HETPHEN, heteroleptic phenanthroline; HAT-CN, dipyrazino[2,3-f:20 ,30 -h]quinoxaline2,3,6,7,10,11-hexacarbonitrile; HLCT, hybridized local and charge transfer; HOMO, highest occupied molecular orbital; ILCT, intraligand charge transfer; Ir(ppy)3, tris[2phenylpyridinato-C2,N]iridium(III); ISC, intersystem crossing; ITO, indium tin oxide; Jsc, short-circuit photocurrent density; knr, non-radiative decay rate constant; kr, radiative decay rate constant; LC, ligand-centered; LED, light emitting diode; Liq, 8-hydroxyquinolinato lithium; LLCT, interligand charge transfer; Lmax, maximum luminance; LMMCT, ligand-to-metal-to-metal charge transfer; LUMO, lowest unoccupied molecular orbital; MC, metal-centered; mCP, 1,3-bis(N-carbazolyl)benzene; mdpbq, 3,30 -methylene4,40 -diphenyl-2,20 -biquinoline; MLCT, metal-to-ligand charge transfer; (M + X)LCT, (metal + halide)-to-ligand charge transfer; m-MTDATA, 4,40 ,400 -tris[phenyl(m-tolyl)amino] triphenylamine; NHC, N-heterocyclic carbene; NHetPHOS, N-heterocyclic phosphine; NIR, near-infrared; NPB, N,N0 -diphenyl-N,N0 -bis(1-naphthyl)-1,10 -benzidine-4,40 -diam ine; N^N, diimine; OFETs, organic field-effect transistors; OLEDs, organic light emitting diodes; OPVs, organic photovoltaic cells; OXD7, 1,3-bis[2-(4-tert-butylphenyl)-1,3,4oxadiazo-5-yl]benzene; Pc, phthalocyanine; PCE, power conversion efficiency; PCEref, power conversion efficiency of standard dye; PC71BM, [6,6]-phenyl C71 butyric acid methyl ester; PEDOT:PSS, poly(3,4-ethylenedioxythiophene):polystyrene sulfonate; PolyTPD, poly(4-butylphenyldiphenylamine); PEmax, maximum power efficiency; phen, 1,10-phenanthroline; PPh3, triphenylphosphine; PL, photoluminescent/photoluminescence; PLQY, photoluminescence quantum yield; PMMA, poly(methyl methacrylate); PVK, poly(9-vinylcarbazole); PYD2, 2,6-dicarbazolo-1,5-pyridine; pz2Bph2, diphenyl-bis(pyrazol-1-yl)borate; P^P, diphosphine; p-6P, para-sexiphenyl; RISC, reverse intersystem crossing; SOC, spin-orbit coupling; SPPO1, 2-(diphenylphosphoryl)spirofluorene; SQ2, 2,4-bis[4-(N,N-dibutylamino)-2,6-dihydroxyphenyl]squaraine; SSL, solidstate lighting device; TADF, thermally-activated delayed fluorescence; TAPC, 4,40 -cyclohexylidenebis[N,N-bis(4-methylphenyl)benzenamine]; TAZ, 3-(biphenyl-4-yl)-5-(4tert-butylphenyl)-4-phenyl-4H-1,2,4-triazole; TCCz, N-(4-(carbazol-9-yl)phenyl)-3,6-bis(carbazol-9-yl)carbazole; TCIQ, 4-[3,6-di(carbazol-9-yl)carbazol-9-yl]isoquinoline; TCO, transparent conductive oxide; TCPy, 3-[3,6-di(carbazol-9-yl)carbazol-9-yl]pyridine; TCTA, tris(4-carbazol-9-ylphenyl)amine; Td, decomposition temperature; TFB, poly [(9,9-dioctylfluorenyl-2,7-diyl)-co-(4,40 -(N-(4-sec-butylphenyl)diphenylamine); TmPyPb, 1,3,5-tri(m-pyridin-3-ylphenyl)benzene; TPBI, 2,20 ,200 -(1,3,5-benzinetriyl)-tris (1-phenyl-1H-benzimidazole); TPD, N,N0 -bis(3-methylphenyl)-N,N0 -diphenylbenzidine; TTA, triplet-triplet annihilation; Voc, open-circuit photovoltage; VON, turn-on voltage; Xantphos, 4,5-bis(diphenylphosphino)-9,9-dimethylxanthene; XLCT, halide-to-ligand charge transfer; XMCT, halide-to-metal charge transfer; 26mCPy, 2,6-bis(N-carbazolyl) pyridine; 3TPYMB, tris(2,4,6-trimethyl-3-(pyridin-3-yl)phenyl)borane; 4CIQ, 4-(carbazol-9-yl)isoquinoline; e, molar absorptivity; DE(S1–T1), band gap between the lowest singlet and triplet excited states; kem, emission peak wavelength; kmax, maximum peak wavelength; kEL, electroluminescent peak wavelength; kEL,max, maximum electroluminescent peak wavelength; kPL,max, maximum photoluminescent peak wavelength; gext, device efficiency; s, excited-state lifetime. ⇑ Corresponding authors at: Department of Applied Biology and Chemical Technology, The Hong Kong Polytechnic University, Hong Hom, Hong Kong, China (W.-Y. Wong). E-mail addresses: [email protected] (C.-L. Ho), [email protected] (W.-Y. Wong). https://doi.org/10.1016/j.ccr.2018.05.010 0010-8545/Ó 2018 Elsevier B.V. All rights reserved.

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

515

abundant metals with less toxicity, which makes it competitive to precious transition metals. As a result, a series of rational molecular engineering has been developed to boost the device performance of copper complexes. In this review, the recent progress of copper complexes in the fields of organic light emitting devices (OLEDs), photovoltaic cells (dye-sensitized solar cells (DSSCs) and bulk heterojunction solar cells (BHJSCs)) in the past two decades will be presented. Representative examples are chosen for discussion. Ó 2018 Elsevier B.V. All rights reserved.

Contents 1. 2. 3.

4.

5.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fundamental properties of copper(I) complexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Copper(I) complexes as emitters for OLEDs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Mononuclear four-coordinate compounds. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Mononuclear three- and two-coordinate compounds. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Multinuclear compounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Copper complexes as the active absorption layer in solar cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. Copper complexes for DSSCs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1.1. Homoleptic copper(I) complexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1.2. Heteroleptic copper(I) complexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1.3. Dinuclear and polymeric copper complexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. Copper complexes for organic solar cells. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.1. Copper(II) phthalocyanine and its derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.2. Other copper(I) complexes for BHJSCs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1. Introduction Transition metal complexes have aroused tremendous attention for different energy-related applications due to the unique characters of the metal centers. One of the unique characteristics is the strong spin-orbit coupling (SOC) of organometallic complexes, which enables fast intersystem crossing (ISC) and long-lived phosphorescent decay, as compared with the pure organic compounds. As a result, most of the heavy transition-metal complexes show phosphorescence [1] and are of particular interest for use in organic light emitting devices. By excitation of the lowest triplet excited state, a strong phosphorescence is produced. In view of the molecular orbitals, the emission is largely determined by the metal-to-ligand charge transfer (MLCT) state between the d orbitals of the metal center and the p orbitals of the organic ligand. It turns out that the electronic transfer between different energy states (color tuning) can be carefully controlled by modifying the chelating ligands with appropriate frontier molecular orbitals [2]. As a result, organometallic complexes are being actively studied in order to boost the efficiency of optoelectronic devices. A great deal of research effort has been paid to green energy applications using organic optoelectronics since the 1980s [3,4]. Compared to conventional inorganic semiconductors, organic small molecules and polymers can be synthesized and fabricated by utilizing different strategies to offer low-cost mass-production over a span of applications, including organic light emitting diodes (OLEDs), organic field-effect transistors (OFETs) and organic photovoltaic cells (OPVs). This triggered the fast-growing development of the field that involves the synthesis of new organometallic/ organic materials and device performance engineering. Regarding light emitting devices, the materials used were dominated by the phosphorescent emitters, especially Pt(II) and Ir(III) organometallic complexes, due to their long-lived triplet excited states [5]. On the other hand, there is a growing trend to use less expensive Cu(I) complexes, due to their thermally-activated delayed fluorescent (TADF) nature to harvest both singlet and triplet excitons. These

515 515 524 524 534 535 539 539 541 543 552 553 553 554 554 555 555

new findings redefine the traditional views of Cu(I) complexes, which were originally regarded as poor candidates for OLED active materials. In yet another important area of organic photovoltaic devices, Cu(I) complexes have attracted much attention as the active layer for the conversion of sunlight into electricity. This is because the intrinsic long-lived triplet excited state prolongs the exciton diffusion length in a solar cell and reduces the chance of charge recombination [6]. This benefits the transfer and migration of charge carriers and boosts the photocurrent within the device (Tables 1–4). Herein, we summarize the recent progress made using Cu(I) complexes in light emitting and light harvesting applications. The photophysical and electrochemical features are discussed in Section 2, which allow the readers to gain a general background of Cu(I) materials. Subsequently, representative Cu(I) complexes used as emissive materials in OLEDs are reviewed together with their device performance in Section 3, followed by a survey of their light harvesting materials for use in dye-sensitized solar cells (DSSCs) and bulk heterojunction solar cells (BHJSCs) in Section 4. Of particular emphasis is the correlation between the structures and photophysical properties, which aims at improving the photoluminescence and photosensitization performance through precise structural modifications. Finally, a short summary of the design strategies will be given in the conclusion. 2. Fundamental properties of copper(I) complexes In this section, both of the photophysical and electrochemical characters of Cu(I) complexes will be illustrated. The scientific interest in investigating the structure-property relationship of Cu(I) complexes has increased significantly since the discovery of the room temperature luminescence of [Cu(dmp)2]+ (dmp = 2,9dimethyl-1,10-phenanthroline) by Sauvage and McMillin [7–9]. However, the development of Cu(I) complexes was rather limited for light harvesting and light emitting applications in the 90s due

516

Table 1 Key performance characteristics of OLEDs using Cu(I) emitters. Device structure

Fabrication method [V]a

kPL,maxb [nm] (PLQY, %)

VON [V]

kEL,max [nm]

EQEmax [%]

CEmax [cd A1] (mA cm2)

Lmax [cd m2] (V)

References

O1d

ITO/PEDOT:PSS/1 wt% O1d:PVK/BCP/Alq3/LiF/Al

Sol

700 (0.18)

13

555

–

2.6 (1), 1.2 (100)

1704 (26)

[59]

O1f

ITO/PEDOT:PSS/16 wt% O1f:PVK/BCP/Alq3/LiF/Al

Sol

560 (16)

14

524

–

10.5 (1), 1.7 (100)

1663 (28)

[59]

O1f

ITO/PEDOT/10 wt% O1f:CBP/SPPO1/LiF/Al

Sol

560 (16)

8.0

539

3.4

11.6

5598 (19)

[61]

O1f

ITO/PEDOT/10 wt% O1f:PYD2/SPPO1/LiF/Al

Sol

560 (16)

6.3

524

8.7

28.6

13,820 (18)

[61]

O1f

ITO/PEDOT/10 wt% O1f:PYD2/DPEPO/LiF/Al

Sol

560 (16)

5.6

507

15.0

49.5

3272 (23)

[61]

O2c

ITO/PEDOT/10 wt% O2c:PVK/BCP/Alq3/LiF/Al

Sol

625 (10)

9

612

0..7

1.4 (1), 0.2 (100)

182

[62]

O2f

ITO/PEDOT/10 wt% O2f:PVK/BCP/Alq3/LiF/Al

Sol

631 (0.76)

11

620

0.9

1.4 (1), 0.2 (100)

183

[62]

O2c

ITO/PEDOT/10 wt% O2c:TCCz/TPBI/LiF/Al

Sol

625 (10)

11

606

1.7

3.4 (1), 0.8 (100)

888

[62]

O2f

ITO/PEDOT/10 wt% O2f:TCCz//TPBI/LiF/Al

Sol

631 (0.76)

11

618

4.9

8.0 (1), 1.0 (100)

967

[62]

O3

ITO/2-TNATA/NPB/10 wt% O3:CBP/TPBI/LiF/Al

Vac

–

4

585

–

5.8 (1), 1.6(100)

2322

[63]

O4

ITO/m-MTDATA/NPB/6 wt% O4:CBP/Bphen/Alq3/LiF/Al

Vac

–

–

565

–

2.94 (1), 1.41 (100)

4483 (15)

[68]

O6a

ITO/PEDOT:PSS/TCTA/10 wt% O6a:mCP/TmPyPb/LiF/Al

Vac

616

5.2

544

14.81

47.03 (9.0)

11,010 (11.0)

[72]

O6b

ITO/PEDOT:PSS/TCTA/10 wt% O6:mCP/TmPyPb/LiF/Al

Vac

616

5.6

544

11.71

35.61 (8.0)

5152 (11.0)

[72]

O6c

ITO/PEDOT:PSS/TCTA/10 wt% O6c:mCP/TmPyPb/LiF/Al

Vac

616

5.3

544

6.67

21.33 (6.0)

5242 (9.0)

[72]

O7a

ITO/m-MTDATA/NPB/1 wt% O7a:CBP/Bphen/Alq3/LiF/Al

Vac

552

–

568

–

1.82 (10), 1.46 (100)

5543

[74]

O7b

ITO/m-MTDATA/NPB/1 wt% O7b:CBP/Bphen/Alq3/LiF/Al

Vac

521

–

448

–

1.27 (10), 1.33 (100)

8669

[74] [75]

O7d

ITO/MoO3/NPB/8 wt% O7d:CBP/BCP/LiF/Al

Vac

638

–

571

–

3.04 (1.65)

4758 (12.3)

O8a

ITO/m-MTDATA/NPB/9 wt% O8a:CBP/Bphen/Alq3/LiF/Al

Vac

525 (25)

–

525

–

1.71

1500 (12)

[76]

O8b

ITO/m-MTDATA/NPB/23 wt% O8b:CBP/Bphen/Alq3/LiF/Al

Vac

–

–

480

–

1.47

2850 (15)

[77]

O8c

ITO/m-MTDATA/NPB/18 wt% O8c:CBP/Bphen/Alq3/LiF/Al

Vac

–

–

532

–

2.35

2320 (12)

[77]

O9a

ITO/PEDOT:PSS/5 wt% O9a:PYD2/DPEPO/LiF/Al

Sol

–

–

535

2.0

–

–

[78]

O9b

ITO/PEDOT:PSS/5 wt% O9b:PYD2/DPEPO/LiF/Al

Sol

–

–

549

6.1

–

–

[78]

O9c

ITO/PEDOT:PSS/5 wt% O9c:PYD2/DPEPO/LiF/Al

Sol

–

–

548

7.4

–

–

[78]

O10a

ITO/PEDOT:PSS/20 wt% O10a:26mCPy/DPEPO/LiF/Al

Sol

590 (2.1)

6.9

516

8.47

23.68

2033

[80]

O10b

ITO/PEDOT:PSS/20 wt% O10b:26mCPy/DPEPO/LiF/Al

Sol

536 (45)

6.9

504

1.59

4.07

502

[80]

O10c

ITO/PEDOT:PSS/20 wt% O10c:26mCPy/DPEPO/LiF/Al

Sol

540 (30)

6.1

508

3.18

8.82

1116

[80]

O11a

ITO/PEDOT:PSS/10 wt% O11a:PYD2/DPEPO/LiF/Al

Sol

570

6.9

508

6.4

17.8

1378

[81]

O11b

ITO/PEDOT:PSS/10 wt% O11b:PYD2/DPEPO/LiF/Al

Sol

556

6.3

498

4.7

13.0

1286

[81]

O11c

ITO/PEDOT:PSS/10 wt% O11c:PYD2/DPEPO/LiF/Al

Sol

580

5.9

498

5.5

14.5

1366

[81]

O11d

ITO/PEDOT:PSS/10 wt% O11d:PYD2/DPEPO/LiF/Al

Sol

576

6.9

506

6.1

17.4

1366

[81]

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

Complex

O12

ITO/PEDOT:PSS/20 wt% O12:czpzpy/DPEPO/TPBI/LiF/Al (direct spin-coating of ligand and Cu(I) ion)

Sol

569

5.6

516

6.36

17.53

3251 (14.3)

[82]

O12

ITO/PEDOT:PSS/20 wt% O12:czpzpy/DPEPO/TPBI/LiF/Al

Sol

569

5.6

514

6.34

17.34

2939 (14.1)

[82]

O13a

ITO/PEDOT:PSS/20 wt% O13a:mCP/3TPYMB/LiF/Al

Sol

518 (16)

5.3

490

5.83

14.01

6563

[83]

O13b

ITO/PEDOT:PSS/20 wt% O13b:mCP/3TPYMB/LiF/Al

Sol

518 (19)

7.4

501

7.42

20.24

5579

[83]

O14a

ITO/PEDOT:PSS/5 wt% O14a:PYD2/DPEPO/LiF/Al

Sol

601 (3)

7.1

537

3.5

10.7 (10)

1750

[84]

O14b

ITO/PEDOT:PSS/5 wt% O14b:PYD2/DPEPO/LiF/Al

Sol

601 (5)

7.8

546

4.6

14.1 (10)

1850

[84]

O14c

ITO/PEDOT:PSS/5 wt% O14c:PYD2/DPEPO/LiF/Al

Sol

567 (19)

7.3

526

6.7

21.4 (10)

2339

[84]

O14d

ITO/PEDOT:PSS/5 wt% O14d:PYD2/DPEPO/LiF/Al

Sol

569 (16)

7.3

516

8.7

27.1 (10)

2893

[84]

O15a

ITO/PEDOT/10 wt% O15a:TCCz/BCP/Alq3/LiF/Al

Sol

638 (0.2)

7.5

562

–

5.58, 4.11 (1)

2820

[85]

O15b

ITO/PEDOT/10 wt% O5b:TCCz/BCP/Alq3/LiF/Al

Sol

612 (1.5)

7.5

555

–

8.87, 6.00 (1)

465

[85]

ITO/TAPC/8 wt% O16a:mCP/3TPyMB/TmPyPB/LiF/Al

–

559 (34)

4.2

–

6.6

20

598 (12.6)

[87]

ITO/PEDOT:PSS/23 wt% O18:PVK/TPBI/Ba/Al

Vac

562 (4.54)

–

550

1.8

5.34

842 (35)

[88]

O20a

ITO/TAPC/10 wt% O20a:mCP/3TPYMB/LiF/Al

Vac

609c (0.1)

–

552

11.9

34.6

–

[90]

O20b

ITO/TAPC/10 wt% O20b:mCP/3TPYMB/LiF/Al

Vac

614c (0.5)

–

545

16.0

46.7

–

[90]

O20c

ITO/TAPC/10 wt% O20c:mCP/3TPYMB/LiF/Al

Vac

616c (2)

–

528

17.7

54.1

–

[90]

O21

ITO/PEDOT:PSS/PVK/10 wt% O21:30 wt% TAPC:mCP/ 3TPYMB/Al

Vac

521c (52)

5

550

7.8

21.3 (0.25)

–

[92]

O22a

ITO/PEDOT:PSS/5 wt% O22a:PYD2/DPEPO/TPBI/LiF/Al

Sol

592 (5)

–

558

16.57

49.8

3480

[93]

O22b

ITO/PEDOT:PSS/5 wt% O22b:PYD2/DPEPO/TPBI/LiF/Al

Sol

602 (4.9)

–

562

15.64

43.33

5580

[93]

O22c

ITO/PEDOT:PSS/2.5 wt% O22c:PYD2/DPEPO/TPBI/LiF/Al

Sol

–

–

573

5.10

14.75

3030

[93]

O22d

ITO/PEDOT:PSS/12 wt% O22d:PYD2/DPEPO/TPBI/LiF/Al

Sol

547 (15.9)

–

535

18.46

58.50

8650

[95]

O22e

ITO/PEDOT:PSS/12 wt% O22e:PYD2/DPEPO/TPBI/LiF/Al

Sol

602 (5)

–

582

14.30

35.27

17,600

[95]

O22f

ITO/PEDOT:PSS/4 wt% O22f:CBP/DPEPO/TPBI/LiF/Al.

Sol

659 (0.6)

–

631

10.17

11.26

4630

[95]

O23a

ITO/12.5 wt% O23a:PVK/Al

Sol

556 (0.01)

15

–

–

–

490

[96]

O23b

ITO/12.5 wt% O23b:PVK/Al

Sol

494 (2)

20

–

–

–

330

[96]

O24a

ITO/MoO3/TAPC/10 wt% O24a:mCP/TPBI/LiF/Al

Vac

–

3.5

584

9.6

24.7

1000 (6)

[97]

O24b

ITO/MoO3/TAPC/10 wt% O24b:mCP/TPBI/LiF/Al

Vac

–

3.5

584

12.4

32.7

1000 (7)

[97]

O24c

ITO/MoO3/TAPC/10 wt% O24c:mCP/TPBI/LiF/Al

Vac

–

3.5

584

16.3

40.8

1000 (6.5)

[97]

O25a

ITO/TAPC/10 wt% O25a:mCP/3TPYMB/LiF/Al

Vac

534 (43)

3.3

527

21.2

67.7

–

[100]

O25b

ITO/TAPC/10 wt% O25b:mCP/3TPYMB/LiF/Al

Vac

527 (47)

3.0

517

21.3

65.3

–

[100]

O25c

ITO/TAPC/10 wt% O25c:mCP/3TPYMB/LiF/Al

Vac

517 (60)

3.1

513

21.2

62.4

–

[100]

O25d

ITO/TAPC/10 wt% O25d:mCP/3TPYMB/LiF/Al

Vac

–

3.3

529

22.5

69.4

–

[100]

O25e

ITO/TAPC/10 wt% O25e:mCP/3TPYMB/LiF/Al

Vac

–

3.3

515

18.6

55.6

–

[100]

O26

ITO/PEDOT:PSS/TFB/20 wt% O26:PVK/Bphen/LiF/Al

Sol

–

3.4

–

9.7

30.4

7790

[101]

O27

ITO/CFx/TAPC/CBP:25 wt% TAPC/0.2 wt% O27:25 wt% TAPC: CBP/CBP/BAlq-13/LiF/Al

Vac

560d (68)

–

524

16.1

47.5

–

[106] [111]

O28a

ITO/PF01/10 wt% O28a:CBP/Bphen/KF/Al

Vac

–

–

565

4.8

10.4

1700

O28a

ITO/PEDOT/10 wt% O28a:PYD2/DPEPO/LiF/Al

Sol

–

4.7

517

9.0

30.6

14,400 (23)

[61]

O28a

ITO/MoO3/CBP/O28a/TPBI/LiF/Al

Vac

–

3.3

544

8.3

25.2

8019

[113] 517

(continued on next page)

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

O16a O18

518

Table 1 (continued) Complex

b c d

Fabrication method [V]a

kPL,maxb [nm] (PLQY, %)

VON [V]

kEL,max [nm]

EQEmax [%]

CEmax [cd A1] (mA cm2)

Lmax [cd m2] (V)

References

O30a

ITO/NPB/20 wt% O30a:mCPy/BCP/LiF/Al

Vac

–

3.6

530

4.4

13.8

9700

[114]

O30b

ITO/MoO3/CBP/4 wt% O30b:CPPyC/TPBI/LiF/Al

Vac

–

–

530

15.7

51.6

23,160

[115]

O30c O30c

ITO/MoO3/CBP/5 wt% O30c:TCIQ/TPBI/LiF/Al ITO/MoO3/CBP/0.2 wt% O30c:TCIQ/TPBI/LiF/Al

Vac Vac

– –

3.2 3.5

580 White

4.1 0.8

10.4 1.4

7100 3886

[116] [116]

O30d

ITO/MoO3/CBP/5 wt% O30d:CpzPC/TPBI/LiF/Al

Vac

–

3.6

590

6.6

15.9

8619

[117]

O30e

ITO/HAT-CN/NPB/TCTA/5 wt% Ir(bpiq)2(acac):TCTA/6 wt% O30e:CzBPCb/TPBI/Liq/Al

Vac

–

3.6

520

3.24

53.84

4450

[118]

O31

ITO/PEDOT:PSS/20 wt% O31:PYD2/3TPYMB/LiF/Al

Vac

503 (79)

7.3

522

8.3

27.1

2525

[119]

O32

ITO/PEDOT:PSS/10 wt% O32:TCTA/TPBI/LiF/Al

Sol

619 (45)

4.3

574

6.09

12.78 (6.8)

5379 (11.3)

[121]

O33a

ITO/PEDOT:PSS/polyTPD/(45:45:10) O33a:TPBI:PVK/TPBi/Ca/Ag

Sol

586 (7)

4.1

551

–

9 (6.3)

1800 (10)

[123]

O33b

ITO/PEDOT:PSS/PLEXCORE-UT-134/30 wt% O33b:PYD2/ 3TPYMB/LiF/Al

Sol

–

2.6

540

23

73

10,000 (10)

[124]

O33c

ITO/PEDOT:PSS/O33c/TPBI/LiF/Al

Sol

–

3.7

552

11.4

36.4

34,000

[125]

O34

ITO/10 wt% O34:PVK/TAZ/Al

Sol

520 (42)

12

516

0.1

–

50

[13]

O35

ITO/TPD/O35/OXD7/AlLi

Vac

–

–

493

0.2

–

400 (22)

[128]

O36

ITO/m-MTDATA/NPB/10 wt% O36:CzSi/Bphen/Alq3/LiF/Al

Vac

–

5

467

–

–

1140 (14)

[129]

O37

ITO/PEDOT:PSS/10 wt% O37:mCP/TPBI/Liq/Al

Sol

491

–

White

0.73

–

1500

[130]

Sol = solution processing, Vac = vacuum deposition. PL in CH2Cl2 at 298 K. PL in 2-MeTHF at 298 K. PL in cyclohexane at 298 K.

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

a

Device structure

519

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557 Table 2 Key performance data of DSSCs sensitized with homoleptic Cu(I) dyes.

a

Complex

kmax [nm]

Voc [mV]

Jsc [mA cm2]

FF [%]

PCE [%]

Reference standard

PCEref [%]

grel [%]a

References

D3 D4a D4b D5a D6a D6c D6d D7a D8 D9 D10a D11 D12 D13

440 482 450 470 470 475 496 437, 608 493 316, 506 564 600 560 430

600 570 630 566 556 590 570 484 490 563 515 470 340 610

0.6 1.21 3.9 5.25 5.9 7.3 4.69 0.69 0.46 3.6 0.206 0.0325 0.0338 5.95

60 65 – 64 70 69 79 63 62 70 71 66 40 70

0.1 0.45 2.5 1.9 2.3 3 2.2 0.21 0.14 1.41 0.08 0.0101 0.0046 2.57

– N719 – N719 N719 N719 N719 N719 N719 N719 – – – –

– 5 – 9.7 9.7 8.9 7.87 5 5 2.53 – – – –

– 9 – 20 24 34 28 4 3 56 – – – –

[167] [168] [169] [170] [170] [171] [172] [168] [168] [173] [174] [176] [176] [177]

Photoconversion efficiency of the device relative to the reference photosensitized N719 with efficiency set at 100%. grel = PCE/PCEref  100%.

Table 3 Key performance data of DSSCs sensitized with heteroleptic Cu(I) complexes. kmax [nm]

Voc [mV]

Jsc [mA cm2]

FF [%]

PCE [%]

Reference standard

PCEref [%]

grel [%]

References

a

552

475

2.17

69

0.71

N719

6.55

11

[181]

D15a

582

465

0.82

66

0.25

N719

6.55

4

[181]

D16a

477

525

3.76

75

1.47

N719

7.36

20

[182]

D16b

500

565

4.99

72

2.04

N719

7.36

28

[182]

D16c

504

605

10.86

71

4.66

N719

7.36

63

[182]

D16d

502

625

10.13

70

4.42

N719

7.36

60

[182]

D19b

394

347

0.233

65

0.053

N719

3.05

2

[184]

D20a

–

627

0.006

60

2.35

N719

7.29

32

[188]

D20b

–

579

0.007

61

2.33

N719

7.29

32

[188]

D21a

475

643

0.005

44

1.3

N719

7.29

18

[188]

D21b

495

627

0.006

46

1.69

N719

7.29

23

[188]

D21c

460

522

5.55

72

2.08

N719

7.45

28

[189]

D21d

475

465

1.26

64.6

0.38

N719

7.45

5

[189]

D22

455

520

5.61

71.1

2.07

N719

6.61

31

[189]

D23

475

524

3.13

71

1.16

N719

7.17

16

[190]

D24

475

548

4.18

74

1.69

N719

7.17

24

[190]

D25a

475

566

5.2

73

2.16

N719

7.17

30

[190]

D26

455

548

4.76

67

1.76

N719

7.17

25

[190]

D27a

–

451

0.001

62

0.4

N719

4.5

9

[192]

D27b

–

433

0.001

59

0.35

N719

4.5

8

[192]

D27c

–

488

0.002

57

0.64

N719

4.5

14

[192]

D27d

–

433

0.002

65

0.49

N719

4.5

11

[192]

D27e

–

451

0.002

60

0.48

N719

4.5

11

[192]

D27f

–

469

0.002

59

0.55

N719

4.5

12

[192]

D28a

–

506

0.005

54

1.25

N719

4.5

28

[192]

D28b

–

488

0.005

53

1.17

N719

4.5

26

[192]

D28c

–

524

0.005

45

1.12

N719

4.5

25

[192]

D28d

–

488

0.005

57

1.31

N719

4.5

29

[192]

D28e

–

525

0.004

49

1.07

N719

4.5

24

[192]

D28f

–

506

0.005

55

1.28

N719

4.5

28

[192]

D29a

–

506

0.001

58

0.42

N719

4.5

9

[192]

D29b

–

414

0.001

62

0.22

N719

4.5

5

[192]

D29c

–

469

0.001

57

0.25

N719

4.5

6

[192]

D29d

–

433

0.001

58

0.32

N719

4.5

7

[192]

D29e

–

506

0.001

56

0.39

N719

4.5

9

[192]

Complex D14

(continued on next page)

520

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

Table 3 (continued) Complex

kmax [nm]

Voc [mV]

Jsc [mA cm2]

FF [%]

PCE [%]

Reference standard

PCEref [%]

grel [%]

References

D29f

–

487

0.001

56

0.31

N719

4.5

7

[192]

D30a

–

561

0.004

60

1.51

N719

4.5

34

[192]

D30b

–

514

0.004

63

1.45

N719

4.5

32

[192]

D30c

–

579

0.004

57

1.34

N719

4.5

30

[192]

D30d

–

543

0.004

63

1.31

N719

4.5

29

[192]

D30e

–

561

0.004

59

1.42

N719

4.5

32

[192]

D30f

–

561

0.004

56

1.2

N719

4.5

27

[192]

D34a

443

607

7.16

72

3.12

N719

6.87

45

[195]

D34b

445

579

6.4

67

2.47

N719

6.87

36

[195]

D34c

471

558

6.42

63

2.59

N719

6.87

38

[195]

D34d

453

552

6.74

70

2.6

N719

6.87

38

[195]

D35a

–

507

4.53

72

2.3

N719

5.96

39

[197]

D35b

–

502

3.56

66

1.8

N719

5.96

30

[197]

D35c

–

549

4.44

66

1.61

N719

5.96

27

[197]

D35d

–

514

4.21

69

1.5

N719

5.96

25

[197]

D35e

–

536

3.53

68

1.29

N719

5.96

22

[197]

D36a

–

527

6.01

73

2.3

N719

5.96

39

[197]

D36b

–

525

5.3

65

1.8

N719

5.96

30

[197]

D36c

470

571

7.06

60

2.43

N719

5.96

41

[197]

D36d

–

520

6.07

60

1.9

N719

5.96

32

[197]

D36e

425

592

6.7

73

2.89

N719

5.96

48

[197]

D37a

490

520

2.54

70

0.97

N719

5.91

16

[180]

D37b

580

463

1

70

0.33

N719

5.91

6

[180]

D38

–

604

7.1

74

3.16

N719

7.63

41

[198]

D39a

470

511

1.68

73

0.63

N719

5.71

11

[199]

D39b

470

447

0.89

71

0.28

N719

5.71

5

[199]

D39c

470

508

1.57

73

0.58

N719

5.71

10

[199]

D39d

470

496

1.12

70

0.39

N719

5.71

7

[199]

D40a

470

554

4.91

73

1.99

N719

5.52

36

[199]

D40b

470

522

4.29

71

1.58

N719

5.52

29

[199]

D40c

470

583

4.96

73

2.12

N719

5.52

38

[199]

D40d

470

571

4.33

74

1.83

N719

5.52

33

[199]

D41a D41b

– 470

621 613

4.97 4.91

67 70

2.08 2.03

N719 N719

5.76 5.76

36 35

[200] [200]

D41c

–

554

4.19

69

1.6

N719

5.76

28

[200]

D41d

–

551

4.58

72

1.83

N719

5.76

32

[200]

D42b

470

535

5.9

67

2.13

N719

5.76

37

[200]

D42c

–

509

4.33

73

1.61

N719

5.76

28

[200]

D42d

–

521

5.35

74

2.07

N719

5.76

36

[200]

D43b

470

591

6.02

67

2.4

N719

5.76

42

[200]

D43c

–

544

4.17

72

1.64

N719

5.76

28

[200]

D43d

–

565

5.22

72

2.13

N719

5.76

37

[200]

D44

432

520

0.4

62

0.13

–

–

–

[201]

D45

600

510

0.91

68

0.31

–

–

–

[201]

D46

–

520

1.21

64

0.41

–

–

–

[201]

D47b

–

610

5.51

71

2.37

N719

10.93

24

[202]

D48a

–

510

6

70

2.15

N719

7.32

29

[203]

D48b

–

482

4.41

70

1.48

N719

7.32

20

[203]

D48c

–

475

4.22

70

1.39

N719

7.32

19

[203]

D48d

–

542

5.6

72

2.18

N719

7.32

30

[203]

D48e

–

468

3.22

69

1.03

N719

7.32

14

[203]

D48f

–

508

4.29

67

1.46

N719

7.32

20

[203]

521

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557 Table 3 (continued) Complex

kmax [nm]

Voc [mV]

Jsc [mA cm2]

FF [%]

PCE [%]

Reference standard

PCEref [%]

grel [%]

References

D49a

–

515

6.46

68

2.26

N719

6.9

33

[203]

D49b

–

506

6.08

71

2.18

N719

6.9

32

[203]

D49c

–

475

5.48

70

1.82

N719

6.9

26

[203]

D49d

–

488

5.25

71

1.81

N719

6.9

26

[203]

D49e

–

459

4.01

70

1.29

N719

6.9

19

[203]

D49f

–

444

3.04

69

0.93

N719

6.9

13

[203]

D50a

480

582

5.47

72

2.3

N719

8.06

29

[205]

D50b

480

562

5.41

75

2.29

N719

8.06

28

[205]

D50c

480

540

5.65

68

2.07

N719

8.06

26

[205]

D50d

480

562

5.88

72

2.38

N719

8.06

30

[205]

D51

485

557

6.81

72

2.73

N719

8.06

34

[205]

–

–

[206]

b c d e

c

–

570

0.54

69

–

N3

D55

–

101

0.024

34

0

–

–

D56a

412

240

0.121

42

0.01

–

–

D56b

420

358

0.192

46

0.03

–

–

–

D56c

412

367

0.212

55

0.04

–

–

–

[208]

D59

350

477d

0.476d

34d

0.77d

–

–

–

[211]

D60a

275

333

0.17

43

0.025

–

–

–

[212]

D60b

275

395

0.17

40

0.027

–

–

–

[212]

D61a

245

445

0.16

41

0.03

N719

3.05

1

[213]

D61b

245

296

0.19

37

0.02

N719

3.05

1

[213]

D62a

–

500

0.19

54

0.051

–

–

–

[214]

D62b

–

500

0.15

52

0.038

–

–

–

[214]

D63a

–

519

0.171

38.7

0.034

–

–

–

[215]

D63b

–

527

0.163

48.9

0.041

–

–

–

[215]

D64a

–

680

4.34

64

1.88

–

–

–

[216]

D64b

–

730

4.93

67

2.42

–

–

–

[216]

D52c

a

b

[208] [208] [208]

As in Table 2. grel = PCE/PCEref  100%.    Using I 2 radical anion/I redox mediator as the electrolyte instead of I3 /I .  Using [CoL3]2+/3+ redox mediators (where L = bpy or phen) as the electrolyte instead of I 3 /I . N3 as reference standard Voc = 0.61 V, Jsc = 1.74 mA cm2, FF = 72% under the same conditions. Using 3 layers of TiO2 and under 60 min sonication.

Table 4 Key performance data of BHJSCs incorporating Cu complexes as the active materials. Complex

Donor

Acceptor

Device structure

kmax [nm]

Voc [mV]

Jsc [mA cm2]

FF [%]

PCE [%]

Reference

B1

B1

C60

ITO/B1 (40 nm)/C60 (40 nm)/Bphen (6 nm)/Ag (100 nm)

ca. 625a

520

4.48

61

1.42

[230]

B1

m-MTDATA

B1

ITO/PEDOT: PSS (20 nm)/m-MTDATA (30 nm)/B1 (60 nm)/LiF (1 nm)/Al (100 nm)

350, 460a

1050

0.0546

30.4

1.03

[231]

B2

B2

p-6P

ITO/PEDOT:PSS/p-6P (20 nm)/B2 (40 nm)/Au (40 nm)

360, 650, 790a

420

0.96

–

0.18

[238]

B5a

B5a

PC71BM

ITO/PEDOT:PSS (30 nm)/B5a:PC71BM/BCP (7 nm)/Al (100 nm)

450b

430c

4.51c

38d

0.72d

[242]

ITO/PEDOT:PSS (30 nm)/B5b:PC71BM/BCP (7 nm)/Al (100 nm)

c

c

c

d

d

[242]

B5b a b c d

B5b

PC71BM

475

650

3.85

34

0.86

Absorption measured in thin films. Absorption measured in DMF solution. Absorption measured in CH2Cl2 solution. Measurement at 433 K.

to the following reasons. Firstly, because both the Sn ? T1 (with n  1) and T1 ? S0 transitions are highly spin-forbidden in Cu(I) complexes, especially in the mononuclear counterparts, only weak

SOC is detected in most Cu(I) complexes. Therefore, the triplet excitons formed in the charge recombination process are difficult to be harvested. In addition, the phosphorescence decay takes a

522

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

Fig. 1. a) Diagram showing electronic configuration and structural rearrangement after photoexcitation. b) The schematic illumination of the process of [Cu(dmp)2]+ from excitation into the MLCT absorption band.

longer time (i.e. several 100 ns to a few ms), which hampers the device stability through a strong saturation effect and photochemical reactions [10]. However, in multinuclear Cu(I) complexes, the enhancement in the luminescent properties is attributed to the rel-

atively strong metal-metal interaction and the corresponding ligand-to-metal-to-metal charge transfer (LMMCT) [11]. Che and Ma have employed the tetranuclear Cu(I) complex [Cu4(C„CPh)4L2] (L = 1,8-bis(diphenylphosphino)-3,6-dioxaoctane) as the first copper-based OLED emitter. Unfortunately, the performance was not satisfactory due to the poor maximum external quantum efficiency (EQEmax 0.1%) and limited brightness [12,13]. Another intrinsic problem includes the non-emissive decay induced by the Jahn-Teller flattening distortion after excitation of the MLCT state and the instability of the complexes. The electronic configuration of Cu(I) complexes has a completely filled d10 layout and gives rise to a symmetrical electron density distribution, which favors a tetrahedral geometry to minimize electrostatic repulsions [14,15]. However, during the MLCT excitation from the 3d orbital of copper to the low-lying antibonding p⁄ orbital of the ligands, tetrahedral Cu(I) ions are formally oxidized into square-planar Cu(II) species upon photoexcitation. It is expected that Cu(I) complexes would undergo Jahn-Teller distortion if the lifetime is long enough, driven by the significant difference in coordination geometry between the two valent states of copper (Fig. 1a). The flattening of the MLCT state from D2d to D2 symmetry leads to an ‘‘open” site in the perpendicular position of the Cu(II) center, which can bind to coordination molecules, such as solvent molecules or counterions, to form penta-coordinated excited compounds (exciplexes). This results in exciplex quenching and further slows down the electron transfer process [16,17]. A detailed dynamic process of such excitation has been demonstrated by the transient absorption and time-resolved fluorescent spectroscopy [18,19]. Taking the emissive model [Cu(dmp)2]+ as an example (Fig. 1b), the complex is firstly excited by external energy consistent with the MLCT absorption band, which incorporates three individual MLCT transitions, labelled as bands I, II, and III [15,20]. Band I covers the visible spectrum above 500 nm, which usually indicates aryl-substitution at the 2,9 positions of phenanthroline. It is ascribed to the p-p interaction between the phenyl rings on the two chelated ligands, which compels the distortion towards a

Fig. 2. Diagram showing the electroluminescence excitation processes for organic and organometallic emitters. (a) Processes of fluorescence in singlet organic molecules. (b) Exciton formation in a 1:3 singlet-to-triplet ratio after electron-hole recombination. (c) Triplet harvesting of phosphorescent organometallic complexes and (d) singlet harvesting of TADF organometallic complexes.

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

square-planar geometry. The absorption band II is located from 430 to 500 nm (on account of the excitation from the ground state S0 to the Franck-Condon state S2), whereas the last absorption band is present at around 390–420 nm (which usually shows a vibronic structure with weak intensities.) The absorption then leads to an electronic transition from the ground state S0 to the FranckCondon state S2 instead of to S1 due to the general transition forbiddance. Sequentially, the S2 state relaxes to S1 via non-radiative decay after a time period of around 45 fs. The process then depends on the substituent groups at the a position of the chelating ligand. If there are bulky groups, S1 would relax directly to the lowest triplet state 3MLCT through ISC. Otherwise, the S1 state will suffer from a flattening in geometry with a timescale of approximately 660 fs, reaching a flattened MLCT state and then decaying to the 3 MLCT state with a higher rate constant of <10 ps, followed by an emission process of circa 50 ns from the triplet excited state to the ground state. In addition, the degree of the MLCT state distortion is vital for the rate constant of ISC since the SOC is controlled by orbital overlap between the copper ion and the chelating ligands, which is determined by the geometry of the copper complexes. On the other hand, the rate of deactivation through radiationless relaxation aggrandizes exponentially with the reduction of the energy gap between the ground state and the excited state in Cu(I) complexes and thereby shortens the lifetime of the MLCT. Apart from the aforementioned undesired decay pathway, fast

Fig. 3. Schematic illustration showing the device and energy diagrams of a typical OLED device.

523

ligand exchange and the ease of oxidation from the Cu(I) to the Cu(II) state also hinder the further development of energyrelated applications for copper complexes. Fortunately, there are still some merits that encourage continuous study on the possibility of using Cu(I) complexes as emissive and photosensitive materials. Structural modification of the ligands prevents the active metal center from being attacked by solvent and thus stabilizes the exciplex during the flattening of the MLCT state. Incorporation of bulky groups into neighboring sites of the chelating atoms exerts a remarkable steric hindrance effect [21,22], thereby leading to the inhibition of MLCT flattening and a decrease in the energy content. Similar effects have also been found in the presence of sterically hindered diphosphine (P^P) and diimine (N^N) derivatives. In other words, it could efficiently prolong the lifetime of the MLCT state and then, in particular, facilitate electron transfer within the solar cell devices. In addition, the emission quantum yield can also be boosted in the presence of a bulky alkyl group due to the increase in rigidity. In addition, Cu(I) complexes show a 3d10 electronic configuration, which implies no unoccupied low-lying d-orbitals and thus no quenching factor of the singlet or triplet excited state caused by the presence of their energetically proximal metal-centered states (MC or d-d state). This is particularly important for the development of efficient blue emitter for OLEDs. In contrast, a delicate design of ligand is required in both Ir(III) and Pt(II) complexes [23–26] due to thermal population of the d-d⁄ state [27,28]. On the other hand, the small bandgap between the lowest singlet and triplet excited states (DE(S1–T1)) (i.e. <1000 cm1), as induced by the weak SOC, facilitates the reverse ISC (RISC) for TADF at ambient temperature [18,29–31]. The thermally re-populated singlet excited state potentially harvests both singlet and triplet excitons, as in singlet harvesting. Also, the TADF emission decay time (i.e. in the order of ls) becomes faster than that in the Cu(I) phosphorescence decay, which may alleviate the problem in the efficiency roll-off via a long-lived triplet excited state with an appropriate device structure. Finally, copper is relatively earthabundant and cost-effective, compared to the high-cost Pt(II) and Ir(III) metal complexes. All these advantages make Cu(I) complexes promising candidates for an in-depth exploration in the fields of OLEDs, DSSCs and BHJSCs. Regarding the electrochemical properties of copper complexes, the redox potentials of the complexes strongly rely on the distinction of geometry between Cu(I) and Cu(II) ions. The oxidation of the Cu(I) ion begins with the removal of an electron from the metal-centered highest occupied molecular orbital (HOMO), generating Cu(II) species with a forced distortion as described. The size of the substituents has an obvious effect on the degree of flux of the complex geometry and accordingly on the redox potentials. It has been observed that ligands with bulky substituents at the a position usually cause an anodic shift in the oxidation potentials of the Cu(I)/Cu(II) redox shuttle for Cu(I) complexes [15]. As a consequence, the oxidation potential of the copper redox couple can be used as a parameter to evaluate the rigidity of the coordination sphere surrounding the copper center. Besides, the redox process of the copper couple could be perceived as ‘‘quasi-reversible”, which depicts the occurrence of a structural transformation between the forward and backward electron transfers. In this aspect, Cu(I) complexes could be considered as excellent reductants in the excited state, which benefits electron injection from the photosensitizer to the semiconductor in DSSCs. Fundamental research has shed light on the photophysical and electrochemical processes of Cu(I) complexes, which allow scientists to further develop new materials with improved performance based on the structure–property relationship. As a result, it is worth reviewing the strategies for both structural modification of

524

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

Cu(I) complexes and device fabrication to meet the application requirements.

3. Copper(I) complexes as emitters for OLEDs As mentioned, there is a growing interest in using Cu(I) complexes as electroluminescent materials due to their efficient TADF effect and low-cost nature. In this section, we concentrate on the development of these complexes for OLEDs and review the structure-property correlation according to the architecture of the corresponding complexes. Before going further, the mechanisms of the electroluminescence will be briefly reviewed (Fig. 2). Theoretically, after the combination of electrons and holes, excitons are generated in a ratio of 1:3 for singlet:triplet states, according to the spin statistics. If a pure fluorescent material is used, only singlet excitons are harvested via the spin-allowed S1 ? S0 transition. However, not only the triplet excitons are dissipated as heat, 80% of the singlet excitons are lost according to the out-coupling efficiency [32]. As a result, a 5% limitation for OLED devices is set by the fluorescent materials. Obviously, utilization of phosphorescent materials, such as 3rd row transition-metal complexes, provides a viable pathway to boost the device efficiency. Ir(III) and Pt(II) complexes are the two extensivelyinvestigated members due to their high SOC constants of 3909 and 4481 cm1, respectively. There are two consequences. Firstly, the rapid ISC results in a fast population of the triplet excited state from the lowest singlet excited state (S1 ? T1 transition). Secondly, the spin-forbidden T1 ? S0 transition becomes easily accessible. Thus, a nearly 100% internal quantum efficiency is reached by harvesting both singlet and triplet excitons, even at room temperature. With such an exciting result, both academic and industrial researchers have focused on the development of new materials and the engineering of the manufacturing process to achieve a better device efficiency in the past two decades [33–36]. In contrast, the common solid-state lighting (SSL) device used nowadays is based on white light emitting diode (LED) technology, which is still acquiring a growing market share for ambient illumination. LEDs are typically based on inorganic semiconductors made from In, Ga, P and N, etc. They provide high efficiency and longer lifetimes than conventional incandescent lamps, but have an energydemanding fabrication procedure. As a result, OLEDs, with a less demanding fabrication condition, are becoming the main green technology in the field of solid-state lighting (SSL) devices [37,38]. Undoubtedly, the relatively mature OLED industry will continue to increase in its market share in the display industry and in our daily life. However, it should be noted that triplet harvesters using Ir(III) and Pt(II) complexes as dopants are costly and less environmental-friendly due to their low abundance [39]. In order to maintain the sustainability of elements such as Ir and Pt, there is increasing interest in academic research to reduce and even replace the precious metals by other alternatives. Other strategies, including triplet-triplet annihilation (TTA) [40,41], hybridized local and charge transfer (HLCT) [42] and TADF [37,43–45], have proved to be promising to harvest the triplet exciton by singlet harvesting. Among them, TADF for noble-metal-free molecules and complexes has been actively studied. This redefines the traditional view of Cu(I) complexes, which were originally regarded as poor candidates for OLED active materials. For Cu(I) complexes to exhibit efficient TADF in OLEDs, there are two major photophysical requirements [46]. The complex should present an DE(S1–T1) energy separation as small as possible to facilitate the fast RISC process at ambient temperature and short radiative decay. This requirement can be accomplished by minimizing the exchange interaction between the HOMO and the LUMO, which means that

the HOMO and LUMO should be separated spatially by a larger extent. However, it should be considered that the correlation between DE(S1–T1) and the radiative rate Dkr(S1 ? S0) may restrict the decrease in the TADF decay time. Another requisite is the high photoluminescence quantum yield (PLQY). Although the intrinsic geometric distortion in Cu(I) complexes provides a pathway to non-radiative decay or quenching of luminescence, suitable molecular engineering can suppress such problems by the introduction of sterically demanding motifs. Once the conditions are fulfilled, a short radiative TADF decay could be obtained to minimize the efficiency roll-off and undesired energy transfer in the devices. Apart from the photophysical parameters, the device fabrication is also crucial to boost the OLED efficiency. OLEDs contain multilayers that are a few hundred-nanometer thick and are composed of anode/cathode, charge injection and transport layers, and a light emissive layer (Fig. 3). The operation of an OLED basically involves four steps, including the generation of charge carriers (holes and electrons), their movement towards the opposite electrodes, formation of excitons and emission of light [35]. The performance of an OLED is evaluated based on the EQEmax, electroluminescent peak wavelength (kEL), maximum current efficiency (CEmax), turnon voltage (VON), maximum power efficiency (PEmax) and maximum luminance (Lmax). Vacuum deposition is the most dominant method to fabricate OLED devices. It operates at relatively high temperatures and low pressures (5  105 mbar) to sublime the materials layer by layer. This allows a good control on the thickness of each layer, but requires a certain thermal stability of the emissive materials and is material-consuming. However, only a few reports are known in which vacuum thermal deposition was employed due to the thermal instability of most Cu(I) complexes [47]. Meanwhile, solution processing methods using spin-coating, inkjet printing or roll-to-roll fabrication have been intensively developed as alternatives for device processing. Compared to the vacuum deposition method, the film thickness is relatively difficult to control and the lifetime of the device is usually low in most cases. In addition, the VON value is usually higher than those of devices prepared using vacuum deposition techniques, because of the defects at the interface [48,49]. Defects in morphology, such as crystalline grains in functional layers, act as charge traps, whereas the aggregation, especially those in phosphorescent emitters, causes quenching in the emission. However, the processing can be done on a large scale under atmospheric conditions, with less materials being consumed. As a result, the latter method is commonly employed for Cu(I)-based OLEDs and was used to present the first OLED device of the Cu(I) family. All the above drew the attention and drove the scientific community to investigate and develop the next generation low-cost electroluminescent (EL) devices using the earth-abundant copper metal. In the following sections, Cu(I) complexes, a relatively young family of light emitting materials, will be discussed for both of their development and structural-property relationships in EL devices. Representative examples of Cu(I) complexes are presented according to their structures (coordination number and metal nuclearity). 3.1. Mononuclear four-coordinate compounds Mononuclear four-coordinate Cu(I) complexes have been studied since the late 1970s [7,50]. As mentioned, the flattening distortion in [Cu(phen)2]+ (phen = 1,10-phenanthroline) is quite severe upon photoexcitation. McMillin et al. have addressed this problem, as in the case of [Cu(dmp)2]+, by introducing alkyl groups at the 2,9 positions of phen to block the flattening of the geometry and solvent quenching of the MLCT state [7,51]. It is interesting to note that the substitution does not change the electronic transition significantly. Later on, sterically demanding ligands, such as triph-

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

525

Fig. 4. Chemical structures of the series O1a–O1f, O2a–O2f and O3–O5 for OLED devices.

enylphosphine (PPh3), were introduced. [Cu(dmp)(PPh3)2]+ showed a longer lifetime in the solid state and deoxygenated solvent [52–55]. Despite this improvement, the complex still exhibited exciplex quenching in methanol. As a result, the same research group continued their explorations by incorporating a bidentate phosphine ligand in [Cu(dmp)(DPEphos)]+ (DPEphos = bis[2-(diphenylphosphino)phenyl]ether) [56,57]. By comparing the effects of monodentate and bidentate phosphine ligands, the ether linkage in DPEphos imposes a larger bite angle and inhibits an increase in the coordination number of the Cu(I) complex. It is unexpected that the two monodentate PPh3 ligands add up to occupy a larger volume than DPEphos, due to interligand repulsion. In addition, an increase in the oxidation potential after the introduction of an alkyl chain to the diamine ligand was revealed in the cyclic voltammogram. This confirmed the ability of the alkyl

chain to resist geometric rearrangement, which was appropriate for the Cu(II) oxidation state. The benefits of using a mixed ligand system in OLEDs can be elaborated by density functional theory (DFT) calculations. The DFT results revealed that the HOMO consists of the lone pair of the P^P ligand and Cu(I) d orbitals, while the LUMO consists of the N^N p⁄ orbital [58]. The wider PACuAP angle in the mixed ligand system (with N^N and P^P) was suggested to enhance the energy of charge transfer (CT) states [56,57]. Also, the larger and more diffuse donor character of the phosphine ligand might stabilize a Cu(I) center rather than a Cu(II) one. Both resulted in the enhancement of the emission lifetime at room temperature. As a result, [Cu(dmp)(DPEphos)]+ in dichloromethane at room temperature exhibited a PLQY of 15% and an excited-state lifetime (s) of 14.3 ls. The subsequent photophysical study of [Cu(dmp)(DPEphos)]+ in poly(methyl methacrylate)

526

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

(PMMA) film (10 wt%) assigned the absorption bands from 220 to 350 nm and that from 350 to 500 nm to ligand p ? p⁄ and MLCT transitions, respectively. This series of work, especially those by McMillin et al., has opened up a new systematic strategy for building up a rigid and stable motif in mononuclear Cu(I) complexes. For the four-coordinate Cu(I)-based OLED, a series of structural modification strategies has been developed for color tuning and enhancement of device stability and efficiency. These strategies include the introduction of bulky substituents into the N^N or P^P ligands, extension of the degree of p-conjugation in the ligands, incorporation of electron donor/acceptor or electron/ hole-transporting motifs, employment of heteroatom-fused cyclometallated ligands and the use of either neutral or cationic Cu(I) complexes. On the other hand, device fabrication conditions also affect the performance of Cu(I)-based OLEDs. A number of fabrication techniques have been employed to improve device performances, such as alternation of materials in the electron-blocking layer (EBL) and host in the devices, usage of phosphor-sensitized phosphorescence and the direct spin-coating of precursors approach. Approaches from both structural and device aspects will be discussed in detail. The first literature report on a four-coordinate Cu(I)-based OLED was published in 2004. The mononuclear charged species [Cu(N^N)(PPh3)2]BF4 (O1a to O1c) and [Cu(N^N)(DPEphos)]BF4 (O1d to O1f) (Fig. 4) were reported by Wang et al. for greencolor OLED applications [59]. These Cu(I) complexes in PMMA (20 wt%) (504–555 nm) showed blue shifts in the emission wavelength and much higher PLQYs in thin film than those in dichloromethane solution (560–700 nm). These can be attributed to the more rigid matrix effect and the absence of solvent-induced exciplex quenching. In [Cu(N^N)(DPEphos)]BF4, the change in alkyl substitution at the 2,9 positions from H to Me and to n-Bu showed a gradual blue shift in both absorption and emission maxima. Increasing the alkyl chain length resulted in increases in both PLQYs (16–69%) and s (4.6–20.3 ls). Here, the introduction of an alkyl chain was again found to improve the rigidity of molecules and hence reduce the structural distortion, supported by the changes in PLQYs and s in the presence of bulky substituents on the N^N ligand. Otherwise, a narrow energy gap between the excited state and the ground state should be observed, which increases the chance of a non-radiative decay process. In contrast, a bidentate phosphine ligand can enhance the PLQY, but with a shortened lifetime. Besides, the introduction of a sterically demanding diphosphine ligand can also improve the thermal stability, showing no loss of the phosphine moiety from the Cu(I) complex at around 200 °C. Due to the promising luminescent properties and high thermal stability of O1f, it was selected for OLED fabrication under different doping conditions and the results were compared with those of O1d. ITO/PEDOT:PSS/Cu(I):PVK/BCP/Alq3/ LiF/Al was employed as the device structure (ITO = indium tin oxide; PEDOT:PSS = poly(3,4-ethylenedioxythiophene):polystyr ene sulfonate; PVK = poly(9-vinylcarbazole); BCP = 2,9-dimethyl4,7-diphenyl-1,10-phenanthroline; Alq3 = tris(8-hydroxyquinoli nato) aluminum(III); LiF = lithium fluoride; Al = aluminum). The device composed with O1f exhibited a constant EL maximum at 524 nm, which was independent of both doping concentration and driving voltage. This was not the case for O1d. Such a difference may be attributed to the higher rigidity in O1f. The best performance of O1f was achieved at a 16% of doping concentration, giving Lmax of 1663 cd m2 at 28 V, CEmax of 10.5 cd A1 at 1.0 mA cm2 and PEmax of 1.6 lm W1 at 105 cd m2. O1d exhibited a relatively poor result, with Lmax of 304 cd m2 at 30 V and CEmax of 2.8 cd A1 at 1.0 mA cm2. However, there were efficiency rolloffs in both cases, but the roll-off was less severe for O1f. In addition, the authors suggested that the substitution of an n-Bu chain in the N^N ligand resulted in a homogenous blend with PVK. Thus,

improved energy transfer efficiency and reduced concentration quenching in O1f were achieved. On the other hand, attaching a phenyl ring to the 2,9 positions was found to be ineffective in preventing a distorted, flattened tetrahedral geometry, because of the p-stacking interactions between the phenyl groups [60]. In 2012, Adachi et al. re-investigated the complex O1f and presented an in-depth study on its photophysics and the tuning of host/electron-blocking layer (EBL) materials on device performance [61]. The strategy employed was mainly focused on the device fabrication aspect. 5 wt% of O1f in PMMA film was used to study the excited state lifetime at different temperatures. The result showed that the long-lived green emission of O1f originated from two thermally equilibrated charge transfer states (3CT and 1 CT) and one non-equilibrated triplet ligand-centered excited state (3LC). The former two could be regarded as the states for ISC and RISC, whereas the transfer between 3LC and 3CT was an ISC process. From this, the 3MLCT energy level of O1f was estimated to be 2.72 eV, which played a key role in electron transfer within the device. This value is considerably higher than that of the common iridium complex tris[2-phenylpyridinato-C2,N]iridium(III), (Ir(ppy)3, 2.42 eV), so the host material and EBL were changed in this study. Increasing the T1 state of the host resulted in an increase in the PLQY of the doped film. In a series of host materials (using 2-(diphe nylphosphoryl)spirofluorene (SPPO1) as the EBL), 2,6-dicarbazolo1,5-pyridine (PYD2), with the highest T1 level (i.e. 2.93 eV), showed the best performance with an EQEmax of 8.7%, while an EQEmax of 3.4% was obtained for 4,40 -bis(carbazol-9-yl)biphenyl (CBP), with a T1 level of 2.56 eV. The author suggested that the energy was more easy to transfer from T1 of O1f to that of the host with a lower triplet energy, and the situation was more severe if the energy gap between the host and O1f was relatively small. As a result, a host with a higher triplet energy than that of the Cu(I) complex allows a better confinement of the exciton on the complex. On the other hand, the triplet energy level of the EBL is also crucial to confine the exciton. The respective T1 energy levels of SPPO1 and bis[2-(diphenylphosphino)phenyl] ether oxide (DPEPO) are 2.9 and 3.0 eV. The latter showed the best EQEmax value of 15.0% among all the O1f-based devices (using PYD2 as the host). This is not only attributed to a better triplet exciton confinement, but also to the absence of excimer formation in DPEPO by its specially twisted structure. The performance of O1f using PYD2 and DPEPO was comparable to Ir(III)-based materials prepared under the same device fabrication conditions. The study by Adachi et al. showed that the unusually high triplet energy and large bandgap in the Cu(I) complexes required specific layer materials (i.e. host/EBL with high triplet energy) in OLEDs, which may not be common in conventional Ir(III)-based devices. Wang et al. have demonstrated a color tuning strategy (from orange-red to red) by extending the p-conjugation of the N^N ligand in [Cu(N^N)(DPEphos)]BF4 (O2a to O2f) (Fig. 4), in which 2,20 -biquinoline (bq), 4,40 -diphenyl-2,20 -biquinoline (dpbq) and 3,30 -methylene-4,40 -diphenyl-2,20 -biquinoline (mdpbq) were employed as the N^N ligand for red colored OLEDs [62]. The photophysical data of O2d to O2f in the PMMA film (20 wt%) showed that the introduction of different phenyl groups (bq to dpbq) caused a slight red-shift in their emission peaks (from 623 to 628 nm), whereas further addition of a methylene bridge at the 2,20 positions blue-shifted the emission peaks (from 628 to 617 nm). Extension of the p-conjugation in the N^N ligands for O2d to O2f resulted in red-shifted emission maxima relative to those of O1a to O1f. By gradually increasing the rigidity of the N^N ligands in the [Cu(N^N)(DPEphos)]+ complexes, the decay lifetimes and PLQYs in the PMMA film were boosted from 2.9 to 20.6 ls and from 6% to 43%, respectively. Noticeably, O2c (with the PPh3 ligand) exhibited a remarkably high PLQY in deoxygenated dichloromethane, showing again the ability of the ultra-rigid mdpbq

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

527

Fig. 5. Chemical structures of the series O6 to O8 for OLED devices.

framework to avoid solvent quenching. Regarding a doping concentration study, increasing the doping level of O2f in PMMA film showed a significant red-shift in emission from 596 to 634 nm with a drop in the PLQY (56–9%), unlike the case of O1f reported by the same group. This behavior can be ascribed to a p-p stacking interaction of the highly conjugated N^N ligands. OLED devices were fabricated for all six complexes (which are not discussed in detail here), but only O2c and O2f were chosen for further device optimization. In these devices, O2c and O2f showed EQEmax values of 0.7 and 0.9%, respectively. In another device structure, replacing PVK/BCP/AlQ3 by TCCz/TPBI (BCP = bathocuproine; TCCz = N-(4-(carbazol-9-yl)phenyl)-3,6-bis(carbazol-9-yl)carbazole; TPBI = 2,20 ,200 -(1,3,5-benzinetriyl)-tris(1-phenyl-1H-benzimidazole)), higher EQEmax values of 1.7 and 4.9% in O2c and O2f were observed with kem at 606 and 618 nm at 10% dopant concentration, respectively. The improvement was attributed to a better hole injection property in TCCz (with a relatively higher HOMO level) and a better electron injection property (with a relatively higher LUMO level) of TPBI. Li et al. have demonstrated color tunability from green-yellow to orange-red in O3 by using the concentration dependent nature in thin films (Fig. 4) [63]. The electron-withdrawing 2,3dicyanopyrazine group was introduced into the phen ligand. This color tuning approach makes use of the local polarization effect (dipole-dipole interaction) between nearby Cu(I) complexes [64].

When the distance between the nearby polar molecules decreases, the local polarization field increases and red-shifts the emission spectrum of the Cu(I) complex. This is similar to the emission spectral shift by varying the polarity of solvents. In the PL study, there were no emissions observed in either the solid state or solution at room temperature, but weak PL signals in thin films and solution were detected at 77 K. Since O3 is stable toward sublimation, vacuum deposition was chosen for the device fabrication. In the device, an increase in the doping concentration from 2 to 25 wt% varied the emission peak maxima from 558 to 615 nm, which was ascribed to the strong polarization effect. The best performance was achieved at the 10 wt% doping concentration for the O3 complex, with Lmax of 4483 cd cm2 and CEmax of 11.3 cd A1 for the EL peak at 585 nm. In another study, the 2 wt% of the yellow emissive O3 doped in the CBP host layer was chosen for white OLED fabrication by combining it with the blue emissive N,N0 diphenyl-N,N0 -bis(1-naphthyl)-1,10 -benzidine-4,40 -diamine (NPB) layer [65]. A chromaticity-tuning layer was inserted between these two emissive layers. Due to the simultaneous exciton generation in both emission layers, white light can be achieved. The device showed CIE coordinates of (0.33, 0.36) at 10 V with Lmax of 2466 cd cm2 and CEmax of 676 cd A1. To improve the performance of the O3-based device, the concept of phosphor-sensitized phosphorescence was introduced by co-doping 2 wt% of O3 and 8 wt% of bis[2-(4,6-difluorophenyl)pyridinato-C2,N](picolinato)iridium(III)

528

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

(FIrpic) into the same host [66]. This phosphor-sensitized approach aims to harness both singlet and triplet excitons efficiently [67]. By the rapid ISC of FIrpic, the singlet exciton can be converted to the triplet state and further transfer the energy to the triplet state of the Cu(I) complex, hence boosting the device efficiency. Formation of excitons was possible in O3 and FIrpic. In addition, the singlet excitons generated in CPB were transferred to FIrpic via the Förster energy transfer mechanism, whereas the triplet excitons were transferred from CPB to either FIrpic or O3 through a Dexter energy transfer mechanism. Through the ISC process in FIrpic, the triplet excitons were finally moved to O3 and this was supported by the negligible emission from FIrpic and the relatively short excited state lifetime of O3 (0.41 ls) as compared to that of 1.44 ls in FIrpic. The device showed Lmax of 3883 cd cm2, CEmax of 26.2 cd A1 and PEmax of 17.8 lm W1. Compared with the mono-doped device containing 2 wt% O3, there is an enhancement by a factor of 2.6 and 2.1 in CEmax and PEmax, respectively. The improvement was attributed to the efficient energy transfer from FIrpic to O3. The authors have further demonstrated the use of phosphorsensitized phosphorescence to achieve a white light emission by combining it with the blue emission from 4,40 -bis(2,20 -diphenylvi nyl)-1,10 -biphenyl (DPVBi). In another study, the 2,3-diphenyl-pyrazine moiety was introduced into the phen ligand as in O4 (Fig. 4) [68]. The extended p-conjugation successfully shifted the emission peak to 565 nm from that in O1f in the PMMA film. Increasing the dopant concentration of O4 from 2 to 15 wt% showed a red-shifted emission, and this was attributed to the increasing interaction between the molecules of O4 at a higher dopant concentration. The configuration ITO/m-MTDATA/NPB/O4:CBP/Bphen/Alq3/LiF/Al (m-MTDATA = 4,40 ,400 -tris[phenyl(m-tolyl)amino]triphenylamine, Bphen = bathophe nanthroline) was used for the device. Due to insufficient energy transfer from the host (kem at 420 nm) at the low dopant concentration (i.e. 2 wt%), the EL spectrum showed a white emission with CIE coordinates of (0.32, 0.35). The best performance was achieved at 6% dopant concentration with Lmax of 4483 cd cm2 at 15 V and CEmax of 2.94 cd A1 at 1.0 mA cm2. Ge et al. have introduced either an electron-donating indole moiety (O5c and O5d) or an electron-withdrawing phenazine (O5a and O5b) chromophore to tune the color of the emissions (Fig. 4) [69]. This is a classical color tuning approach by directly adjusting the HOMO/LUMO energy levels of the metal complexes using ligand modifications. From the solid-state PL spectra, O5c and O5d showed almost identical emission peaks at 540 nm, indicating that the ethyl chains have a negligible effect on the emission wavelength. However, a longer lifetime (6.38 ls) was induced by the alkyl chain in O5d than that of O5c, suggesting an unfavorable interaction posed by the active hydrogen atom to the excited state. The relatively broadened emission peaks of O5a and O5b were centered at 620 and 650 nm, respectively. From the DFT study, O5b showed a lower LUMO energy level than that of O5c, indicating that energy stabilization occurred on the Cu(I) complex by introducing an electronwithdrawing unit, and hence a red-shifted PL emission was observed. The authors further suggested that the introduction of an electron donor with limited conjugation was able to blue-shift and narrow the solid-state emission. This was further attributed to a self-restriction in the geometry relaxation in the excited MLCT states. In contrast, the electron-accepting coplanar system was not effective in self-restricting the geometry relaxation, leading to a broadened and red-shifted emission band. Since O5b displayed a pure red emission at 650 nm, it was applied in the device fabrication with the structure of ITO/CuPc/NPB/40 wt% O5b:CBP/BCP/Alq3/LiF/ Al. The onset voltage was as high as 16 V, with Lmax of 680 cd m2 and an EL peak at 620 nm. The above studies illustrate the systematic ways to tune the colors of Cu(I) complexes by either extending the p-conjugation or

introducing electron acceptor/donor in the ligands, but the device performances in the red region were still relatively poor [63,69]. As mentioned, only a few Cu(I) complexes were able to be fabricated into devices through vacuum deposition, due to their poor thermal stability. To improve the sublimability of cationic Cu(I) complexes, two approaches have been reviewed, either by employing bulky negative counterions [70] or bulky ligands [71]. By learning from these guidelines, a series of [Cu(N^N)(DPEphos)]PF6 complexes containing both 4,5-diazafluorene and bis(carbazole) ligands was synthesized for realizing sublimable yellowish-green OLEDs [72]. This work presents an enhancement of thermal stability by weakening the lattice energy with bulky ligands, while charge balance in light emitting layers can be achieved by incorporating bipolar ligands. The complexes were structurally different in the substituent on the 9-N atom of carbazole, namely, ethyl (O6a), ethylhexyl (O6b) and phenyl (O6c) groups (Fig. 5). Their PL spectra showed no change in the positions of the emission peaks. There are a few interesting features in this system. Firstly, the bulky bis(carbazole) moiety made the size of complexes larger so as to reduce the lattice energies and lower the electrostatic interaction. These complexes were found to be thermally stable up to 300 °C, and hence the complexes were easily sublimed to form thin films by vacuum deposition. Another feature is the bipolar nature of the N^N ligands by introducing hole-transporting carbazole and electron-transporting 4,5-diazafluorene groups, which can further balance the charge carriers in the device and therefore improve the device efficiency. Finally, the photophysical study showed a dominant TADF nature (97% of the total emission) in these three complexes, allowing high PLQYs (77–89%) in their doped 1,3-bis (N-carbazolyl)benzene (mCP) films at 300 K. The TADF nature was also supported by the pronounced spatial separation from frontier orbital calculations, as the DFT study showed that the HOMOs were localized on the bis-carbazole p-orbitals, while the LUMOs were located on the 4,5-diazafluorene units and phenyl groups of the DPEphos ligand. The large spatial separation hence resulted in a small DE(S1–T1) energy. It should be noted that the MO distribution of O6 is very different from that of the traditional [Cu(N^N) (DPEphos)]+ complex. As a result, both singlet and triplet excitons can be harvested in these three complexes. The best device result was obtained from O6a, with EQEmax of 14.81%, Lmax of 11010 cd m2 and CEmax of 47.03 cd A1. The attractive EQEmax value was attributed to the high PLQY value as well as the TADF nature to harvest both singlet and triplet excitons. On the other hand, the relatively poor performances in O6b and O6c were ascribed to inefficient electron confinement within the emitting layers, having similar LUMO levels (2.59 to 2.61 eV) to the host (2.60 eV). Despite the good result, efficiency roll-off was still observed in all three devices. Apart from the use of bipyridine-based ligands (bpy), fivemembered heterocyclic rings, such as pyrazole, triazole, tetraazole and pyrroles, have been used to replace one of the pyridine rings [2]. The stronger r-donor property in the deprotonated state of azolate and the p-acceptor ability of pyridine can provide a highly stable, five-membered metal-chelate interaction by synergic electron delocalization. Besides, a different photoluminescent property is expected with the increasing number of nitrogen atom in a given N^N ligand [73]. In addition, the azole fragment can be further optimized by attaching substituents with different steric, electronic or degree of p-conjugation effects. In 2009, Li and Liu reported a series of [Cu(N^N)(DPEphos)]BF4 complexes based on 2-pyridinylbenzoimidazole, namely, 1-(4-50 -phenyl-1,3,4-oxadiazo lylbenzyl)-2-pyridinylbenzoimidazole (O7a), 1-(4-carbazolylbu tyl)-2-pyridinylbenzimidazole (O7b) and 1H-2pyridinylbenzimidazole (O7c) (Fig. 5) [74]. This work indicates that the introduction of either an electron- or hole-transporting motif in the N^N ligand can improve the device efficiency. The addition

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

529

Fig. 6. Chemical structures of the series O9 to O14 for OLED devices.

of either 2,5-diphenyl-1,3,4-oxadiazole (decomposition temperature Td: 322 °C) or carbazole (Td: 354 °C) gave higher thermal stability for the complexes over that of O7c (Td: 297 °C), allowing the use of vacuum deposition in the device fabrication. O7a to O7c showed PL peaks at 552, 521 and 532 nm and PLQYs of 42, 23 and 24%, respectively. The photophysical data showed that the triplet excited states of the Cu(I) complexes were raised by the addition of the hole-transporting carbazole, whereas the introduction of electron-transporting 2,5-diphenyl-1,3,4-oxadiazole lowered the triplet excited states. By increasing the dopant level from 1 to 25 wt%, the EL spectra of O7a are almost invariant, with a peak at 527 nm, whereas that of O7b displayed a red-shifted emission from 498 to 556 nm. Lmax values of 5543 and 8669 cd m2 were recorded for O7a- and O7b-based devices, respectively. The higher

brightness in O7b was attributed to the presence of the holetransporting group. The oxadiazole-based complex O7d was further developed by adding a tert-butyl group to 2,5-diphenyl1,3,4-oxadiazole [75]. The O7d-based device with 8 wt% dopant level showed yellow EL at 571 nm with Lmax of 4758 cd m2 and CEmax of 3.04 cd A1. Although the attachment of carbazole and oxadiazole ligands can improve the thermal stability of Cu(I) complexes, efficiency roll-offs were still present at high current density (101–103 mA cm2), except for the low doping concentration of the O7b-based device (1%) [74]. Alternatively, thiazole was introduced into the carbazole-pendant ligand in 1-(4-carbazolylbutyl)2-thiazol-4-yl-1H-benzoimidazole for O8a and the complex was used for the fabrication of a green OLED device (Fig. 5) [76]. O8a showed an improved PLQY (25%, at 525 nm) in dichloromethane

530

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

solution as compared to that without a carbazole-pendant ligand (0.2%, at 515 nm), as well as an increased thermal stability. The device showed Lmax of 1500 cd m2 and CEmax of 1.71 cd A1 at 525 nm. Noticeably, the device only showed a slight efficiency roll-off at high current density. This was attributed to the shielding effect of the carbazole ligand, which reduced the TTA and tripletpolar annihilations. Replacing the pendant carbazole (O8a) by a ethyl group (O8c) caused a slight red-shift in the EL spectrum (peaking at 532 nm) [77]. O8c with 18 wt% doping concentration showed an improved performance with Lmax of 2320 cd m2 and CEmax of 2.35 cd A1 at the same driving voltage. However, the device suffered from severe efficiency roll-off at high current density. 1H,10 H-[2,20 ]biimidazolyl was also employed as the N^N ligand in O8b, and the corresponding device with its dopant concentration at 23 wt% exhibited a blue color at 480 nm with Lmax of 2850 cd m2 and CEmax of 1.47 cd A1. The pyridine-imine ligands (i.e. pyridine-azole) not only provide a diversity of the color tuning spectrum, as mentioned above, but also result in less non-radiative decay via CAH vibrational

quenching in the ligand. Adachi et al. have addressed this issue by replacing imidazole with tetrazole, showing relatively low frequency NAH/NAN modes. Three ligands 2-(20 -pyridyl)imidazole (O9a), 2-(20 -quinolyl)imidazole (O9b) and 2-(5-tetrazolyl) quinoline (O9c) were used in [Cu(N^N)(DPEphos)]BF4 (Fig. 6) [78]. The PL peaks in thin films doped with the host were recorded at 530, 549 and 544 nm for O9a to O9c, respectively. In addition, the PLQYs increased from 25% to 36%, whereas the knr (nonradiative decay rate constant) decreased significantly from 4.8  104 to 2.6  103 s1. From O9a to O9b, the enhancement of PLQY and the reduction in knr can be simply attributed to the additional phenyl ring on the quinolyl unit, which increases the steric hindrance between the mixed ligands (N^N and DPEphos). The presence of excited state distortion in O9a was further supported by the large temperature dependence of PLQY and knr, and the relatively large ‘‘Stokes-like shift” [79]. On the other hand, the suppression of CAH vibrational quenching was suggested by the smaller knr value for O9c against O9b at all temperatures, and hence an improved PLQY (1.3 times to O9b) was obtained in O9c. Their OLED

Fig. 7. Chemical structures of the series O15 to O21 for OLED devices.

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

devices showed either green or yellow emission. The best device performance was achieved for O9c with an EQEmax value of 7.1%, which was attributed to both suppression of distortion by the quinoline unit and suppression of the high frequency CAH vibrational mode in the tetrazole ligand. A new class of functionalized 2-pyridylpyrazole ligands was introduced to elevate the LUMO energy level. [Cu(N^N)(DPEphos)]BF4 containing 1-(2-pyridyl)pyrazole (O10a), 3-methyl-1(2-pyridyl)pyrazole (O10b) and 3-trifluoromethyl-1-(2-pyridyl)pyr azole (O10c) moieties emitted colors spanning from greenish-blue to blue in their devices (Fig. 6) [80]. Tuning of the color (greenishblue to blue) was achieved by attaching either an electron donor (CH3) or an electron acceptor (CF3) group to the ligands, leading to stabilization and destabilization of the LUMO energy level, respectively. By varying the substituents from CH3 to H to CF3, their band gaps gradually shifted from 3.23 to 3.04 eV. From O10a to O10c, the PL wavelengths of 590, 536 and 540 nm and PLQYs of 2.1, 45 and 30% were recorded in dichloromethane at 298 K, respectively. The corresponding PL wavelengths of 490, 465 and 492 nm and PLQYs of 56, 87 and 75% were detected in the solid state at 298 K. Similar to other Cu(I) complexes, changing the medium from solution to solid caused blue-shifted PL peaks and increased PLQYs. This was attributed to the matrix rigidity, which reduced the freedom of the molecules upon excitation. The incorporation of CH3 and CF3 groups not only affected the complexes electronically, but also sterically due to their bulkiness. Hence, the extent of the geometry distortion was reduced in O10b and O10c. Additionally, the DE(S1–T1) values of these three complexes were measured to be around 0.17–0.18 eV, and their decay behavior confirmed their TADF emission nature at ambient temperature. The O10c-based device, with the structure ITO/PEDOT:PSS/20 wt% Cu(I):26mCPy/DPEPO/LiF/Al (26mCPy = 2,6-bis(N-carbazolyl) pyridine), exhibited a greenish-blue color at 508 nm. It also featured the best performance with EQEmax, Lmax, and CEmax values of 8.47%, 2033 cd m2 and 26.68 cd A1, respectively, which was again attributed to the good confinement of triplet excitons and charge carriers. On the other hand, the similar LUMO levels among the host (26mCPy), O10b and O10c resulted in simultaneous electron injection and exciton formation in both the host and dopants, and poor confinement of excitons, both resulting in poor device performance. A similar strategy was applied to the 3-C-linked pyrazolylpyridine, in which the N atom was substituted with phenyl (O11a), 2-(o-tolyl)-phenyl (O11b), 2-(trifluoromethyl)phenyl (O11c) and 4-(trifluoromethanyl)phenyl (O11d) groups (Fig. 6) [81]. Unexpectedly, the substituent effect was not significant and similar colors were achieved from the devices. This was supported by the similar EL wavelength (502 nm) and band gap energy (3.04 eV). In addition, the HOMO/LUMO levels only showed slight differences among the complexes despite the different substitutions. O11a showed the best performance with EQEmax, Lmax, and CEmax values of 6.4%, 1378 cd m2 and 17.8 cd A1, respectively. The results indicate that substitution of electron-donating/ accepting groups to the N^N ligands might not always be useful in boosting device efficiency. 9H-Carbazole was added to the 6 position of 2-pyridylpyrazole in O10a, yielding 2-(9H-carbazolyl)-6-(1H-pyrazolyl)pyridine as the ligand in O12 (Fig. 6) [82]. In this study, the ligand and metal precursor were directly spin-coated as the thin film inside the device structure of ITO/PEDOT:PSS/20 wt% O12:czpzpy/DPEPO/ TPBI/LiF/Al (czpzpy = 2-(9H-carbazolyl)-6-(1H-pyrazolyl)pyridine) and the EQEmax, Lmax, and CEmax values were recorded to be 6.36%, 3251 cd m2 and 17.53 cd A1, respectively. By comparing these results with the device that consisted of the directly spincoated O12 complex, there was not much difference in the efficiency. In addition, the ligand functioned as both the host and complex precursor. This work demonstrates a simplified fabrication

531

process without any purification of the Cu(I) complexes, especially when there is an easy tendency for ligand dissociation from the mononuclear four-coordinated Cu(I) complex. This makes the fabrication process much more simple and cost-effective with less material loss during the synthesis of the Cu(I) complexes. The introduction of the electron/hole-transporting moiety in the N^N ligand might sometimes result in some unexpected emission property by changing the MLCT transition to the intraligand charge transfer (ILCT) type. The N^N ligand of O10a was further extended, as in O13a, and the corresponding DPEphos ligand was also replaced by 4,5-bis(diphenylphosphino)-9,9-dimethylxan thene (Xantphos) in O13b (Fig. 6) [83]. This work presents a few approaches in improving the device performance. Firstly, an increase of the rigidity of the P^P ligand by employing Xantphos resulted in a slight increase in the PLQY in both dichloromethane and the mCP thin film when compared with that using DPEphos. To further restrain the excited state distortion, a 9,9dimethylacridan unit was introduced into the N^N ligand. Due to its strong electron-donating character, these complexes exhibited a pronounced ILCT-based emission, which was distinct from common Cu(I) complexes showing the MLCT nature. The DFT study further suggested that the HOMO was localized on the acridine unit and the LUMO was distributed on the N^N ligand. In addition, the acridine unit also facilitated the charge transfer (good holetransporting ability) and recombination process in the device. Both O13a and O13b showed a bluish green color, but O13b displayed better performance than O13a, in which the EQEmax and CEmax values were largely improved from 5.83 to 7.42% and from 14.01 to 20.24 cd A1, respectively. The authors suggested that the superior result for O13b was attributed to its better triplet exciton confinement and higher mobility of the charge carriers. Although uncommon ILCT-based Cu(I) complexes were demonstrated, the device performance was moderate and required further improvement. Another class of ligand using the more bulky and conjugated functional 2-(5-phenyl-2H-1,2,4-triazol-3-yl)pyridine was employed in Cu(I) complexes O14a to O14d (Fig. 6) [84]. In terms of structural modifications, introduction of the phenylcarbazole and methyl groups was made to improve the charge recombination and suppress the excited state geometry distortion, respectively. The introduction of the methyl group in O14c and O14d resulted in blue-shifts in both PL (601–568 nm) and EL (540–510 nm) spectra, whereas there were limited effects on the steric and electronic characters by the phenylcarbazole substitution. The PLQYs and kr of O14c and O14d were significantly larger than those of O14a and O14b, suggesting the effectiveness of the methyl group in suppressing non-radiative decay. In their devices, O14b and O14d, with the phenylcarbazole unit, showed slightly higher device efficiencies as compared to others, which was ascribed to the improved hole injection and transportation properties. By combining the merits from the methyl and phenylcarbazole moieties, the O14d-based device showed the best OLED performance among the O14 complexes (EQEmax was increased from 3.5% in O14a to 8.7% in O14d). In conventional cationic Cu(I) complexes, the counterions are believed to cause an unexpected effect on the luminescent properties, especially under high voltage operation of the OLED. Cheng and Wang et al. demonstrated a comparison between cationic (O15a) and neutral (O15b) complexes containing both 2-(20 -quino lyl)benzimidazole and DPEphos as ligands (Fig. 7) [85]. The PL peak maxima of 638 and 612 nm and PLQYs of 0.2% and 1.5% were measured in dichloromethane, respectively. From the DFT study, the lowest energy excited state in O15a was attributed mainly to the MLCT transition, whereas that in O15b was due to a mixture of MLCT and CT states. The photophysical characters are consistent with a recent study published by Bräse et al. using 2-(pyridin-2yl)benzimidazole for both cationic and neutral Cu(I) complexes [73]. The blue shift can be attributed to the negative charge in

532

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

Fig. 8. Chemical structures of the series O22a–O22f, O23a–O23b and O24a–O24c and the complex Zn1 for OLED devices.

the N^N ligand of O15b, leading to an increase in the LUMO energy level. Hence, the neutral compound exhibited a larger band gap and higher emission energy. On the other hand, the higher PLQYs in the neutral complexes might be ascribed to the stronger coordination of the deprotonated N^N ligand and less vibrational quenching due to the loss of an H atom and the counterion [73,86]. A Lmax of 2820 cd m2 and a CEmax of 5.58 cd A1 with the EL peak at 562 nm were realized for the O15a-based device, whereas the device based on O15b showed the PL maximum at 558 nm with Lmax of 465 cd m2 and CEmax of 8.87 cd A1. The lower CE in the O15a-based device was due to the drifting and accumulation of anions towards the anode upon applying the voltage. However, O15b displayed a severe efficiency roll-off at high current density, which was ascribed to the longer decay lifetime, i.e., the TTA effect. As a result, it is still questionable whether the neutral complex is superior to the cationic complex in terms of device performance. Functionalized 2-pyridylpyrrolide ligands can also be used in neutral [Cu(N^N)(DPEphos)] complexes [87]. Trifluoromethyl

groups were added at the 3,5 positions of pyrrolide ring so as to increase the steric hindrance around the Cu(I) center and avoid flattened distortion. Again, the addition of the phenyl ring to pyridine by either CAC bond linkage (O16b) or ring fusion (O17) was demonstrated (Fig. 7). By extending the pyridyl ring in O16a to a quinolinyl ring in O17, a significant red-shift in emission in the solid state was observed from 481 to 553 nm. The addition of the phenyl ring in O16b showed no significant change in the emission wavelength, indicating that the conjugation is limited within the N^N ligand. Due to the high PLQY measured for O16a, it was fabricated into an OLED device. EQEmax, Lmax and CEmax values of 6.6%, 598 cd m2 and 20.0 cd A1, respectively, were recorded. 2-(5(3,5-Difluorophenyl)-2H-1,2,4-triazol-3-yl)pyridine was used in the neutral [Cu(N^N)(DPEphos)]-type complex in O18 [88]. Its EL peak was centered at 550 nm in the yellow region. An O18-based device exhibited EQEmax, Lmax and CEmax values of 1.8%, 842 cd m2 and 5.35 cd A1, respectively. Another class of neutral Cu(I) complexes anionic pyrazol-1-ylborate (O19a to O19c) was synthesized and studied by Yersin

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

et al. for blue-light emission (Fig. 7) [89]. In the solid-state photophysical measurements at ambient temperature, high PLQYs up to 90% were recorded for O19b and O19c with short emission decay lifetimes ranging from 13 to 22 ls. The change in substitution on the borate ligand (from hydrogen, pyrazolyl to phenyl) resulted in a slight spectral shift from 436 to 464 nm in the solid-state emission. The steric demand in O19b and O19c caused a lesser red-shift in the emission wavelength on changing from a crystalline environment to solution, indicating a greater extent of distortion suppression. Replacement of the P^P ligand from DPEphos to 1,2-bis(diphenylphosphino)benzene (dppb) resulted in a redshift in emission by 70 nm. These photophysical changes were supported by DFT calculations, predicting the HOMO and LUMO being dominantly located on the 3d orbitals of the Cu ion and the p⁄ orbitals of the P^P ligand, respectively. By measuring the emission decay lifetime and emission wavelength at temperatures ranging from 30 to 300 K, the luminescence of all three complexes was assigned to the TADF mechanism. For example, O19c showed a drop in the decay lifetime by a factor of 40 to 13 ls from 40 to 300 K. Although this work only presented an in-depth photophysical study without OLED device fabrication, the harvesting of all excitons via TADF with pronounced solid-state PLQY in the blue color emission and a short emission decay lifetime represent an attractive approach to an efficient EL emission based on the earth-abundant and cheap copper metal. Most of the above-mentioned works were focused on molecular engineering of the N^N ligands so as to reinforce the metal-ligand interaction. On the other hand, by structural modification of the P^P ligand in O19 to O20, the P^P ligands not only alter the steric hindrance in the complexes, but also change both the photophysical properties and thermal stability. Osawa et al. published another class of neutral Cu(I) complexes containing diphenyl-bis(pyrazol1-yl)borate (pz2Bph2) and 1,2-bis(diphenylphosphino)benzene-ba sed ligand (dppb) (O20a to O20c) (Fig. 7) [90]. This work was partly inspired by the improved thermal stability of the Cu(I) complexes using the dppb ligand to enable fabrication of OLED devices via vacuum deposition. The frontier orbitals suggested that the HOMO was distributed on the Cu 3d orbital and the coordinating P and N atoms [46], whereas the LUMO was localized over the o-phenylene ring of the dppb ligand. This made the singlet (S1) and triplet (T1) excited states to contribute more than 90% of the HOMO-LUMO character. The thermal stability was improved by introducing F and CF3 substituents to the meta-positions of the four peripheral phenyl rings in the P^P ligands. The improved thermal stability allowed the fabrication of their OLED devices by the vacuum deposition approach. The increase in the number of F atoms not only increased the PLQY in amorphous films, but also blueshifted the EL peak (528–552 nm) within the green luminescence region. However, the spectral shifts among these three complexes were small because the fluorinated moieties stabilized both the HOMO and LUMO energy levels, leading to tiny changes in their band gaps. The best device performance was recorded for O20c with EQEmax of 17.7% and CEmax of 54.1 cd A1. Although O20a was not the best example based on its device performance among these three complexes, it showed the smallest DE(S1–T1) (370 cm1, 46 meV). This smallest value was attributed to the distinctly different spatial separation of the HOMO and LUMO energy levels. This allowed a very short decay lifetime of 3.3 ls at 300 K, which was the shortest TADF decay time so far for the reported Cu(I) complexes. However, recent in-depth studies of O20a showed that the highly allowed transition from S1 to S0 (or kr(S1–S0)) was also an important parameter to obtain a short TADF decay [46,91]. This was especially important when the involved states, S1 and T1, showed high contributions of the HOMO-LUMO character. As a result, O20a with a short TADF decay lifetime should be effective to reduce the saturation effect in OLEDs [90,91].

533

Besides pz2Bph2, 2-diphenylphosphinobenzenethiolate can also be used as an anionic ligand to give the neutral Cu(I) complex O21 (Fig. 7) [92]. The strong electron-donating nature of the thiolate ligand reduced the contribution of the Cu 3d orbitals to the HOMO, thus reducing the MLCT nature of the excited state (which only accounted for 4.6% of the luminescence). This resulted in the reduced Jahn-Teller distortion in the pseudo-tetragonal structure of O21 in the excited state, as evidenced by the small red-shift (57 nm) in the emission in 2-methyltetrahydrofuran upon warming. On the other hand, the interligand charge-transfer (LLCT) dominated the emissive excited state of O21. By optimizing the device structure, the best performance was obtained with EQEmax of 7.8% and CEmax of 21.3 cd A1. Despite the moderate device efficiency, this work showed an alternative pathway to reduce the flattened distortion in the excited state by reducing the MLCT character. Recently, Che et al. demonstrated the use of a carborane-based P^P ligand, 7,8-bis(diphenylphosphino)-7,8-dicarba-nido-undecaborate, for a series of Cu(I) complexes, O22a to O22f (Fig. 8) [93]. In this work, carborane was inserted into the P^P ligand to enhance the rigidity of complexes. Indeed, carborane has been commonly used to functionalize many phosphorescent transition-metal complexes [94]. It can impose a unique electronic property controlled by its binding site to the ligands or metal centers. On the other hand, it provides sufficient stability and rigidity to reduce intermolecular interactions, such as TTA. The complexes O22a to O22c showed no significant change in their EL wavelengths (558–573 nm), and their emissions were all MLCT in nature, as shown from both DFT calculations and photoluminescence spectra. The methyl groups (adjacent to the N atom) on the N^N ligand again proved to be effective in inhibiting the change in conformation and ligand dissociation. Their devices all emitted orange colors with EQEmax values of 16.57, 15.64 and 5.10% for O22a to O22c, respectively. The high device efficiencies of O22a and O22b were due to the TADF emission. The poor device efficiency in O22c was suggested to be the result of insufficient energy transfer from the host to O22c. To further achieve a white OLED, O22b was selected due to its suitability as a long-wavelength emitter. It showed a high EQE by doping O22b with FIrpic, but the color rendering index (CRI) was not satisfactory (61–71). Interestingly, doping O22b with Zn1 showed the best device performance with EQEmax and CEmax values of 6.88% and 14.67 cd A1, respectively. The CRI reached 81 at 300 cd m2 for the white OLED with CIE coordinates of (0.42, 0.44). The EQEmax of this device was comparable to and even the highest among all the solutionprocessed white OLEDs with fluorescent emitters. Since most of the Cu(I) complexes reported were associated with relatively low EQEs (<5%), the same group addressed such problem by adding 2-ethylhexanol (O22d), 9,9-dihexylfluorene (O22e) or

Fig. 9. Chemical structures of the series O25 and O26 for OLED devices.

534

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

9-(4-ethynylphenyl)-9H-carbazole (O22f) to the 4 and 7 positions of the 2,9-dimethylphenanthroline ligand [95]. The HOMO was delocalized on the Cu(I) center, PPh2 and the nido-carborane cage, whereas the LUMO was delocalized on the N^N ligand. As a result, by changing the substituents, the LUMO level can be easily tuned and the emission color ranges from green to yellow to red. For O22f, the neocuproine and carbazole fragments had high electron density and the p-conjugation was extended through the carbazole and alkyne moieties. As a result, O22f displays a significant redshift in its emission. At the optimal dopant concentration inside an appropriate host, the EL peaks of O22d to O22f were centered at 535, 582 and 631 nm. The green O22d, yellow O22e and red O22f-based devices showed the respective EQEmax values of 18.46, 14.30 and 10.17%. O22e gave by far the most efficient Cu(I)-based OLED in the red color region. Apart from the heteroleptic [Cu(N^N)(P^P)]+ complexes, Cu(P^P)+2 can also be used as a candidate in OLEDs. Some homoleptic and heteroleptic bis(diphosphine) Cu(I) complexes were reported using dppb [96]. Bis(diphosphine) ligands favor the formation of Cu(P^P)2 over Cu(phen)(P^P), when either dppb or bis (diphenylphosphino)methane (dppm) was added to the reaction mixture. It is because the bite angle of P^P is small enough such that the complex can accommodate two P^P ligands. Therefore Cu(dppb)2BF4 (O23a) and Cu(dppb)(DPEphos) (O23b) were prepared accordingly (Fig. 8). In the PL study, O23a and O23b were recorded with PL peaks of 556 and 494 nm and PLQYs of 0.01% and 2% in dichloromethane, respectively. The improved PLQY for O23b was attributed to the entangled structure, resulting from the large bite angle of the DPEphos ligand. This makes the structural rearrangement more difficult in solution. However, the device performance was poor in both cases and Lmax values of 490 and 330 cd m2 were recorded in O23a and O23b, respectively. A recently published work demonstrated the possible use of the tridentate phosphine ligand 2,20 -(phenylphosphinediyl)bis(2,1-ph enylene)bis(diphenylphosphine) in designing Cu(I) complexes to achieve high device efficiency [97]. In most TADF Cu(I) complexes, there is an unbalanced nature between TADF and phosphorescence at ambient temperature and usually more than 95% of the emission was attributed to TADF. Although Yersin et al. have demonstrated the possibility of a binuclear Cu(I) complex with a large phosphorescence proportion (20%) [98], it was still a challenge to achieve a balanced dual emission (1:1 TADF:phosphorescence) through limited ISC and RISC cycles. In this work, the tridentate phosphine ligand was introduced to maintain the rigid coordination geometry in the three copper complexes for the series of O24, suppressing the structural distortion in the excited state. The HOMO frontier orbitals were distributed on the phosphorus atoms, Cu center and halide, whereas the LUMO was localized on the two o-phenylene groups. By varying the halide from Cl (O24a) to Br (O24b) to I (O24c) (Fig. 8), the film at 300 K showed a blueshift in the emission from 530 to 521 nm due to the reduced ligand-field strength. However, the EL peaks were 584 nm in all cases. Besides, the structural modification utilizing the heavy atom effect of the halide was used to tune the extent of SOC in the complex. The photophysical study showed that the presence of the iodine atom significantly enhanced the SOC, hence facilitating the T1 ? S0 transition. By changing the halide to I, it was possible to harvest both the singlet and triplet excitons (61% TADF to 39% phosphorescence). As a result, the respective EQEmax values of 9.6, 12.4 and 16.3% were recorded for O24a, O24b and O24c.

around the Cu(I) center. Alternatively, three-coordinate Cu(I) complexes provide a new pathway to tackle such an issue. In terms of structure, either a bidentate N^N or P^P ligand is employed with a monodentate anionic ligand (i.e. carbene (NHC), halide or thiol) [29,31,47,99,100]. Although this kind of complex also exhibited Jahn-Teller distortion from Y-shaped to T-shaped upon photoexcitation, a sterically bulky ligand like P^P and N-heterocyclic NHC can alleviate this situation. Moreover, these neutral complexes are highly thermally stable (i.e. in the case of P^P) and can easily be vacuum-deposited for OLED fabrication. Osawa et al. have reported O25a to O25c by the combination of 1,2-bis[bis(2-methylylphenyl) phosphino]benzene ligands with either Cl, Br or I atoms to study the effect of the halide (Fig. 9) [47,100]. A frontier orbital study showed that the HOMOs were localized on the halogen atom and orbitals involving r-character between the Cu(I) and P atoms, whereas the LUMO was confined to the o-phenylene group in the P^P ligand. All three complexes showed green EL peaks which were slightly blue-shifted from Cl to I in both the PL (534–517 nm) and EL (527– 513 nm) spectra due to their different field strengths. The heavy atom effect arising from the halogen atom was found to enhance the radiative decay in the solid state. Among these three complexes, the best device was achieved for O25b, with EQEmax of 21.3% and CEmax 65.3 cd A1. Such a high device efficiency was attributed to the bulky and rigid P^P ligand, reducing the non-radiative decay. By altering the ortho-substituents from methyl to ethyl to isopropyl in O25b, O25d and O25e, respectively [100], a negligible effect of the alkyl substituents on the electronic states of the three ligands was revealed, thus causing no influence on their photophysical properties. From the X-ray structural analysis, the orthosubstituents were pointed towards the central metal atoms, giving rise to strong steric hindrance and preventing the formation of a stable binuclear Cu(I) complex. Among these three complexes, the best device structure was achieved for O25d with EQEmax of 22.5% and CEmax of 69.4 cd A1. The photophysical properties of an interesting class of twocoordinated Cu(I) complexes has been recently investigated, by either attaching an NHC or cyclic alkyl(amino)carbene (CAAC) ligand [101–104]. Linnolahti et al. recently published a report on the Cu(CAAC)(9-carbazole) complex (O26) (Fig. 9) [101], which showed a sub-microsecond emission at 300 K, and such a lifetime was considerably shorter than efficient phosphorescent emitters (>1.5 ls) and that of efficient TADF emitters (>5 ls). It was claimed that the molecular configuration can somehow achieve low exchange energies and retain appreciable oscillator strength simultaneously. As a result, a rapid emission via a triplet state occurred. Although the detailed mechanism is still under investigation, this kind of emission behavior may avoid the degradation pathway

3.2. Mononuclear three- and two-coordinate compounds As mentioned before, the most commonly used approach to reduce structural rearrangements in the excited state for the fourcoordinate compounds is by employing sterically congested ligands

Fig. 10. Chemical structures of the series O27 to O29 for OLED devices.

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

535

Fig. 11. Chemical structures of O30a and O30b, and their pyridine-based ligands for OLED devices.

arising from bimolecular annihilation events. On the other hand, Steffen et al. have demonstrated the use of another CAAC ligand, 1-(2,6-di-iso-propylphenyl)-3,3,5,5-tetramethyl-2-pyrrolidineylidene), to synthesize [Cu(CAAC)2]+ and Cu(CAAC)X (X = halide) [102]. It was found that the CAAC ligand can act as a very potent p-chromophore. Linear [Cu(CAAC)2]+ and Cu(CAAC)X exhibited bright phosphorescence from blue to green. Finally, the trigonal [Cu(CAAC)(diimine)]+ complex was able to red-shift the emission to the red region. However, the performance of the OLED device was poor and requires further investigation and improvement. The relatively unexplored three-/two-coordinated compounds gave some attractive results in terms of their photophysical characters and device performances. They do require a more in-depth fundamental study to understand their real potential application in optoelectronics. 3.3. Multinuclear compounds Apart from the mononuclear Cu(I) complexes, multinuclear Cu(I) compounds have also been investigated for their potential application in light emitting devices. The main reason for using a multinuclear compound is to increase the structural rigidity. This reduces the chance of structural reorganization upon photoexcitation and boosts the emissive properties. Additional structural modifications include alternation of bridging halides to tune the

emission color and introduction of a long alkyl chain into a complex to improve the solubility and blend morphology. In addition, simple fabrication by direct spin-coating of the precursors is also introduced in this section. Here, dinuclear compounds will be discussed according to the nature of the bridging ligands, followed by a short discussion on polynuclear compounds. The first type of dinuclear Cu(I) compound involves nitrogen bridges. A bis(bis(di-iso-butylphenylphosphino)amido)dicopper(I) (O27) complex was reported in 2005 (Fig. 10), which had a PLQY of 68% and a long lifetime of 10.2 ls in cyclohexane at 298 K [105]. The Cu2N2 core was found to provide strong covalence, which housed a redox-active molecular orbital for transferring charge from the dicopper(I) ground state to an emissive excited state with a similar structure [106,107]. The frontier orbital study suggested that the HOMO resided on the four-membered Cu2N2 ring, whereas the LUMO was localized on the phenyl groups of the ligands. The compound was thermally stable for device fabrication using the thermal vacuum deposition approach. In the device study, the complex displayed a green color. The best device performance was achieved at a 0.2% dopant concentration of O27, with EQEmax of 16.1% and CE of 47.5 cd A1. Although the dopant concentration for high device efficiency was remarkably low, it should be reminded that there was an efficiency roll-off by increasing the current density (EQEmax of 10.9% at a current density of 1 mA cm2). This roll-off was attributed to the quenching of excited state

536

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

Fig. 12. Chemical structures of O31 and O32 for OLED devices.

by holes, or dissociation of the excited state under the effect of the electric field. Another problem on using O27 as an OLED emitter was its air and moisture sensitivity, which made the fabrication somewhat complicated. Apart from the common N/P atom as the bridging ligand, halide atoms can be used as the bridging ligand in dinuclear compounds, which are attractive for their structural diversity, rich photophysical behavior and high emission efficiency [108]. However, for those containing monodentate phosphine ligands, the PLQYs are generally low [109]. From both experimental and theoretical studies, their excited states possess halide-to-ligand charge transfer (XLCT), MLCT and/or halide-to-metal charge transfer (XMCT) states. As mentioned, upon photoexcitation, their mononuclear counterparts have a profound Cu(II) character. However, in the multinuclear complexes, the charge is delocalized and the entire square-planar Cu2X2 unit is oxidized. This leads to a smaller distortion in the excited state and thus a higher quantum efficiency [110]. Tsuboyama et al. first reported the use of [Cu(l-I)dppb]2 as an OLED emitter [111]. By varying the halide from I to Br to Cl (O28a to O28c) (Fig. 10), the PL peak in the solid state was redshifted from 502 to 533 nm, which again was in line with the order of the ligand field strength of the halogen ions (I < Br < Cl). As the ligand field strength decreased, the energy separation of the d-orbitals was reduced. Hence, a shorter wavelength on the emission peak was observed. The frontier orbital study suggested that the main contributions to the HOMOs were from the Cu, halogen and phosphorus atoms, whereas the LUMOs were localized on the P^P ligand. The emissive excited state is a mixture of (metal + halide)-to-ligand charge transfer ((M + X)LCT). Color tuning for this kind of compounds can be directly achieved by changing the bridging phenyl ring to pyridine to pyrazine (O29a to O29c) (Fig. 10), in which the p⁄ levels were gradually stabilized and the energy band gap was reduced [112]. As a result, solid-state emissions at ambient temperature ranging from green to reddishorange were achieved. The O28a-based device exhibited an EL peak at 565 nm, which is longer than the solid-state emission, but similar to that measured in 2-methyltetrahydrofuran at 140 K. It was suggested that the EL emissive state had a structure close to a flattened structure. The device showed the best performance with EQEmax of 4.8%, Lmax of 1700 cd m2 and CEmax of 10.4 cd A1 at

Fig. 13. Chemical structures of a range of N-heterocyclic ligands.

93 cd m2. The low efficiency was further attributed to the ultrahigh triplet energy of O28a and insufficient triplet exciton confinement. Similar to O1f, Adachi et al. have addressed this issue by using a host and EBL with the appropriate triplet energy [61]. The device performance was then improved to EQEmax of 9.0%, Lmax of 14400 cd m2 and CEmax of 30.6 cd A1. Although the host materials have long been investigated for OLED devices, most of them were explored and designed specifically for conventional phosphorescent emitters like Ir(III) and Pt(II) complexes. Regarding their relatively high triplet energy, another class of host materials specific for Cu(I) compounds needs to be explored. Moreover, the doping level of the emitters requires careful control for better reproducibility, while a prolonged operation might lead to a phase separation of the dopant and the host. As a result, it might be good if the fabrication of the host layer in the device could be eliminated. Liu et al. have reported a non-doped device using O28a [113]. In their study, non-doped and doped devices using CBP were compared. By increasing the dopant level from 10 to 100 wt%, a reduction in the driving voltage and better device performance were observed, which was ascribed to the insufficient triplet energy of CBP as compared with O28a. The excitons might not be formed in O28a directly, and the energy was easily transferred back to the host matrix from O28a. In addition, the neat film of O28a showed a higher PLQY than that in the doped film, which originated from the weak intermolecular interaction between the dopant molecules and hence less TTA was observed. However, due to the weak charge transport ability in O28a, the non-doped device displayed severe efficiency roll-off at high current densities. The halogen-bridged compounds O30a to O30e can also be synthesized with mono-pyridine (Fig. 11). Conventionally, they were difficult to use for OLED device fabrication, because of their poor solubility/stability in common organic solvents and poor sublimability. To utilize them, Thompson et al. have demonstrated an in situ co-deposition route from CuI and the pyridine-based ligand 3,5-bis(carbazol-9-yl)pyridine (mCPy) [114]. The pyridine-based ligand had dual functions, i.e. as a host and ligand for coordination. The film composed of [CuI(mCPy)2]2 (O30a) showed a pure green EL peak at 530 nm. The formation of O30b in the film and the

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

537

Fig. 14. Chemical structures of homoleptic and heteroleptic of NHetPHOS-type Cu(I) complexes and a range of ancillary phosphine ligands.

Fig. 15. Chemical structures of the series O33 for OLED devices.

optimal molar ratio of CuI:ligand were determined by the PL peak as compared with that of the bare ligand. The best device performance was EQEmax of 4.4%, Lmax of 9700 cd m2 and CEmax of 13.8 cd A1. The lifetime measurement demonstrated a half-life of 21 h, driven by a current density of 20 mA cm2 under vacuum (corresponding to 440 h at 100 cd m2). With this encouraging preliminary data, Thompson, Wang and Bian further developed a film of O30b using 3-(carbazol-9-yl)-5-((3-carbazol-9-yl)phenyl)pyri dine (CPPyC) [115]. The CPPyC ligand was chosen due to its triplet energy level of 2.7 eV, allowing efficient energy transfer to O30b upon co-deposition in the device. The optimal molar ratio of CuI:ligand was determined to be 1:10 (100% PLQY in the neat film). The best device performance was achieved at 4 wt% CuI dopant level with EQEmax of 15.7%, Lmax of 23160 cd m2 and CEmax of 51.6 cd A1. Unanticipatedly, by varying the CuI dopant level from 1 to 10 wt%, EQEmax values ranging from 12.7 to 15.7% were recorded. Noticeably, CPPyC has the HOMO and LUMO levels similar to those of CBP (for hole transportation) and TPBI (for electron transportation), so these two layers can be replaced by CPPyC. Although CPPyC was found to be a relatively weak charge injection/transportation layer than the two layers above, the EQEmax values up to 16.8% were still observed. This indicates that charge recombination still takes place in the emissive layer, despite a more severe efficiency roll-off. Later on, Liu, Wang and Bian further demonstrated the third role of the pyridine-based ligand as a blue

emitter for white light OLED fabrication [116]. In this study, four ligands were used, namely 4-[3,6-di(carbazol-9-yl)carbazol-9-yl]i soquinoline (TCIQ), 3-[3,6-di(carbazol-9-yl)carbazol-9-yl]pyridine (TCPy), 4-(carbazol-9-yl)isoquinoline (4CIQ) and 3-(carbazol-9-yl) pyridine (CPy). Since the LUMO was mainly concentrated on the ligand, the desired color in the co-deposited film was achieved by using different pyridine-based ligands. Either adding the carbazole moieties to the 3,7 positions of CPy/CIQ or extending the pyridine to isoquinoline led to a red-shift in the PL peaks, but the addition of carbazole moieties showed no change in the triplet excited state energy. The film co-deposited by either CuI:TCPy or CuI:TCIQ exhibited an improvement in the thermal stability, an increase in both the glass transition temperature and PLQY, due to the increase in the rigidity of the ligands. A CuI:TCIQ film (O30c) displayed a yellow color (580 nm) in the OLED device and the best performance was EQEmax of 4.1% and Lmax of 7100 cd m2. By reducing the CuI dopant level to 0.2 wt%, a white OLED was fabricated with EQEmax of 0.6% and Lmax of 2991 cd m2. Although the efficiency was not encouraging, this trifunctional pyridinebased ligand was a useful guide for future low-cost OLED engineering. 9-(3-(6-(Carbazol-9-yl)pyrazin-2-yl)phenyl)-carbazole (CPzPC) (for O30d) and 5-(30 -(9H-carbazol-9-yl)-[1,10 -biphenyl]-3-yl)5H-pyrido[3,2-b]indole (CzBPDCb) (for O30e) were also used as pyridine-based ligands for red-orange and green co-deposited OLEDs with CuI [117,118]. In the series of O30, the simplicity of the in situ

538

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

Fig. 16. Chemical structures of O34 to O37 for OLED devices.

co-deposition route has shown to be a powerful fabrication method, reducing the steps in the synthesis of Cu(I) complexes. The Cu atoms can also be bridged by multi-dentate ligands like tetraimine (O31) [119], bipyridine [120] and bipyrimidine (O32) (Fig. 12) [121], since P and N-containing ligands are soft electron donors, which favor the attachment to soft Cu(I) ions. Also, most of them are commercially available, which can simplify the synthetic procedures [109]. Until now, most of the Cu(I) OLEDs have employed a solution processing approach for device fabrication, however there are a few problems. Usually, they are not very chemically stable in solution, leading to ligand dissociation. On the other hand, some of them are poorly soluble in organic solvents, limiting their availability for solution processing. Recently, a class of dinuclear Cu(I) complexes was introduced by Bräse and Baumann employing bridging P^N-ligands, namely NHetPHOS (N-heterocyclic phosphine) [48,109,122,123]. By combining the Cu2I2 core with a bridging NHetPHOS and two ancillary phosphine ligands, a butterflyshaped Cu(I) complex is revealed, which is able to achieve a relatively high PLQY with a short emission decay (several ls). It can be either homoleptic or heteroleptic in structure. It is preferable for the heteroleptic complex to remove any undesired reactions coming from the two free coordination sites on the two ancillary ligands during the device fabrication (Fig. 13). As mentioned in the introduction part on OLEDs, morphological defects are commonly encountered in solution processing. This problem can be avoided by using materials with a low crystallization tendency, having low lattice energy and good solubility. The NHetPHOStype Cu(I) complex is definitely a promising candidate to provide a low crystallization tendency. The solubility and color tuning are achieved independently by changing the ancillary phosphine

and the bridging NHetPHOS, respectively. By changing the substituents on the phosphine ligands, the solubility of the complexes can be changed from polar (ethanol, for phosphines a–g) to nonpolar solvent (hexane, for phosphines c and g) (Fig. 14). By attaching a–g to the complexes with 2-diphenylphosphino-4-methylpyr idine, bright yellow-to-green PL bands with PLQYs ranging from 28% to 99% were obtained [48]. Most of them were thermally stable up to 250 °C. However, Cu(I) complexes with ancillary ligands bearing more than one PAO bond showed the host emission only. The authors suggested that there may be radical-induced degradation of the emissive complexes during the device operation. From a frontier orbital study, the LUMO was localized on the ligand, whereas the HOMO was centered on the Cu-halide core [122,123]. As a result, by solely changing the N-heterocycle, the PL peaks varied from deep-blue to yellow in color. For instance, electron-rich heterocycles (i.e., imidazoles, benzimidazoles and triazoles) shifted the emission peaks to the blue region (451–506 nm), whereas electron-poor heterocycles (i.e., pyridines, oxadiazoles and thiazoles) shifted the peaks to the yellow region (514– 558 nm) (Fig. 15). In addition, the introduction of an alkyl chain can also be used to fine-tune the solubility. Among them, O33a (Fig. 15), using bridging 2-diphenylphosphino-4-iso-butylpyridine and ancillary PPh3 ligands, was chosen for device fabrication and the solution-processed device showed luminance of 1800 cd m2 at 10 V and CEmax of 9 cd A1 at 6.3 V. The same group further demonstrated the use of a fourth bridging ligand, bis (diphenylphosphino)hexane, in O33b to fabricate a green OLED [124]. The compound (O33b) had a relatively high PLQY (100%, in a doped film with PYD2) and high thermal stability up to 290 °C. Also, it effectively improved the film forming ability with a reduced crystallization tendency. The triplet energy was

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

2.55 eV, and as compared to the common Cu(I) complexes (2.8 eV), O33b was able to confine excitons easily in the emission layer. Thus, the best device performance was recorded with EQEmax of 23%, Lmax of 10,000 cd m2 and CEmax of 73 cd A1. O33b gave the highest efficiency among the Cu(I)-based green-emitting OLED so far. The good performance was attributed to the suppressed structural distortion in the rigid, multinuclear motif with full bridging and chelating ligands. These ligands also contributed to minimize both the triplet energy and singlet-triplet splitting. O33c was developed for demonstrating the device fabrication process in the presence of air (i.e., spin-coating and inkjet-printing) [125]. Due to its impressive chemical stability and solubility, O33c can be dissolved in a high boiling point solvent (i.e., chlorobenzene, o-xylene, indane or toluene) and kept for several hours without degradation. The best device performance was measured with EQEmax of 11.4%, Lmax of 34,000 cd m2 and CEmax of 36.4 cd A1 at an emission wavelength of 552 nm. However, solvent selection in the device fabrication was found to be critical for the device efficiency, as the use of indane led to a relatively poor efficiency. In short, NHetPHOS-type Cu(I) complexes have been demonstrated to be possible for large scale solution processing of Cu(I) complexes, showing moderate efficiency. On moving from dinuclear complexes to nanoclusters, the rigidity and thermo-/photo-stability of Cu(I) complexes were improved [126,127]. However, they had poor processability and weak electroactivity. As mentioned above, the first OLED using Cu(I)-based emitter actually employed a tetranuclear Cu(I) complex (O34), as demonstrated by Che and Ma (Fig. 16) [12,13]. Although the PLQY was comparable to that of Alq3, the efficiency was extremely low (EQEmax of 0.1%). This was caused by the relatively long excited state lifetime. Later on, another tetranuclear compound (O35) was introduced with EQEmax of 0.2%, and the poor performance was suggested to be due to the low PLQY [128]. Later, Zhang and co-workers have shown the possibility of using a tetranuclear complex (O36) with bridging Br atoms for a blue emitting OLED (at 467 nm) [129]. With high thermal stability up to 373 °C, the O36-based device was fabricated by the vacuum deposition method. However, the resulting device performance was poor. Recently, Xie and Xu have reported a Cu4I4 cluster using the 2,9di(diphenylphosphine)-dibenzofuran ligand (O37) [130]. This P^P ligand improved the solution processability and electroactivity, and boosted the dual emission characteristics when voltage was applied. Based on theoretical calculations, the configuration in the ground state and excited states were almost the same, leading to a great suppression of structural distortion. O37 showed typical dual emissive character with a phosphorescence efficiency of 15%. The long-wavelength phosphorescence was governed by the low energy cluster-centered triplet state, whereas the TADF was due to both triplet and singlet M/XLCT with RISC. The lifetime at room temperature was 1.9 ls, which can effectively reduce collisioninduced quenching in the device operation. The solutionprocessed device provided single-molecular white EL with CIE coordinates of (0.37, 0.45) and EQEmax of 0.73% and Lmax of 1500 cd m2. Although these tetranuclear compounds showed poor efficiency, they did open a new avenue to investigate their structure-property relationship in the field of light emitting devices as compared to their mono/dinuclear counterparts.

4. Copper complexes as the active absorption layer in solar cells Issues related to the search for sufficient supplies of low-cost and carbon-free renewable energy for the future as an alternative to the limited natural resources and consequently the reduction of global warming from ejective pollution, have prompted both scientists and industrial sectors to invest a great deal of effort on

539

Fig. 17. The different coordination processes between bpy and phen ligands to form metal chelates.

solar cells [131]. Among the various renewable energy resources, the sun could afford the world’s projected energy demand in a sustainable fashion. The energy from sunlight striking the Earth in one hour is more than all that is consumed on the planet in one year. Thus, solar energy shows great promise as the next generation of green energy through efficient and cheap exploitation. Photovoltaic (solar) cells represent one of the most promising technologies employed to harness solar energy, by virtue of their low fabrication cost and good power conversion efficiency. Overall, they could be classified into inorganic and organic based devices according to their main composition [132–134]. Although inorganic solar cells perform with higher photoelectric conversion efficiency (PCE) and stability than their organic counterparts to date, their deficiencies include but are not limited to the large amount of energy required for the production of semiconductors, lack of photovoltaic grade silicon to satisfy the demand for photovoltaic modules and high manufacturing cost; all of these limit the development of solar cells based upon the presently available semiconductor technology. Therefore, the past two decades have witnessed a surge of activities for developing cheap consumables and flexible electronic products [135–138]. Among these, the development of OPV devices has played a critical role, thanks to their attractive superiorities in low cost manufacturing processes, compatibility with flexible substrates and being extremely light weight with a large size [139,140]. Given these significant advantages of OPV devices, much effort has been made from researchers to improve the performance of OPVs towards a wide and enhanced spectral response, improved charge transport and minimal recombination of electron-hole pairs. Generally, organic solar cells are divided into DSSCs [141– 144], hybrid solar cells comprised of inorganic nanoparticles [145] or nanostructured templates [146,147] and all-organic solid-state donor-acceptor based BHJSCs [148,149]. Herein, our motive lies in updating the current advances of the emerging copper complexes in organic solar cells, especially DSSCs and BHJSCs. We will pay more attention to efficient approaches for desired device performance through structurally fine-tuning the copper complexes, for instance, by attaching substituents with various steric and electronic effects and different degrees of p-conjugation. 4.1. Copper complexes for DSSCs Over the past two decades, a great deal of effort has been devoted to amending the overall efficiency of DSSCs by elaborately engineering dyes and photoanodes [150–157], and a high PCE of 14% has been achieved from a Ru(II) complex sensitized DSSC [158]. However, efficiency is not the only standard for a satisfactory photosensitizer. Usage of one of the rarest and most expensive elements, ruthenium, in photosensitizers such as N719 increases the industrial cost of the devices and thus militates against their popularization. Considering the far-reaching development of

540

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

Fig. 18. Chemical structures of the series D1 to D3 for DSSC devices.

Fig. 19. Chemical structures of the series D4 to D6 for DSSC devices.

photovoltaic technologies, it is important to replace ruthenium by other more abundant metals. Nowadays, an increasing amount of cost-oriented research focuses on using cheaper Cu complexes in the development of new photosensitizers, allowing the growth of a new generation of low-cost DSSC devices. As far as the advantages of Cu complexes are concerned, apart from the broad light absorption in the visible region, long-lived MLCT excited state, intense emission and synthetic versatility, the high photochemical and electrochemical stabilities are also quite appealing. As mentioned in Section 2, Cu(I) complexes have completely filled d10 orbitals, so they cannot suffer from non-radiative degradation from MLCT to MC as described for Ru complexes. In addition, Cu(I) complexes coordinated with bis(diimine) ligands possess similar photophysical properties to their Ru(II) counterparts. Nevertheless, there are problems for employing Cu(I) complexes in designing photosensitizers, including the rapid ligand exchange of Cu(I) complexes and lability in their excited state. Fortunately, the oxidation process of Cu(I) complexes to their Cu(II) counterparts can be dramatically constrained by introducing bulky substituents into the ligands, which block the metal and impede the formation of the preferred Cu(II) geometries [159].

The simplest methods to optimize Cu(I) photosensitizers include: (i) adding various aromatic substituents or p-conjugated segments to the chelates to gain both bathochromic shifts and hyperchromic effects, (ii) engineering ligands to maximize the photonic harvesting and to regulate the redox potential, (iii) introducing anchoring ligands bearing carboxylic acid, phosphonic acid or phenol groups, which can bond to the semiconductor and facilitate electron injection, and (iv) forming a compact hydrophobic layer on the TiO2 surface to avert corrosion [160–162]. In this approach, the superiorities of Cu(I) complexes in broad light harvesting and efficient shielding to avoid electron-hole recombination were shown to afford a satisfactory DSSC performance. As a number of breakthroughs have unveiled the strategy to fine-tune the properties of Cu(I) dyes at the molecular level with great precision in recent years, this section is aimed at underlining promising DSSCs based on Cu(I) complexes as photosensitizers. However, since there are many compositional or device-related variables that have an actual effect on the overall power conversion behavior, it brings about apparent difficulty in comparing the performance of devices fabricated in different laboratories under different conditions. A broadly acceptable practice is to set

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

541

Fig. 20. Chemical structures of the series D7 to D10 for DSSC devices.

a standard control group, i.e., to measure the light-to-current performance of the DSSCs with the designed copper complex dye together with that of the control cell assembled under similar conditions with N719 as the standard dye. Herein, N719 was taken as the reference dye and the relative PCE of the DSSC containing Cu(I) complex as the dye (grel) was reported in the text with respect to 100% for N719. The masking configuration of the fabricated photovoltaic device (i.e., top-masked, fully-masked and unmasked) is also an important consideration in the performance optimization. According to a study done by Snaith, masking of a cell could validate the active area and thus avoid overestimation of the produced photocurrent. However, the use of this technology is not widespread [163] and we are not going to discuss this in detail here as the issues on the masking of DSSCs have been specified elsewhere [164]. In this review, particular emphasis is put on the strategies for improving the performance of the fabricated DSSCs by structural modifications of the Cu(I) dyes. The exhaustive assembled circumstance of all the DSSCs involved will not be described in the text to avoid complexity and confusion to the readers. The I/I 3 redox couple was used as an electrolyte for all the representative Cu(I) dyes, unless otherwise stated. 4.1.1. Homoleptic copper(I) complexes As early as in 1977, McMillin et al. [165] carried out pioneering work for discovering the similar photophysical properties of homoleptic Cu(I) species [CuL2]+ (where L = bpy and phen) to those of the prototypical [Ru(bpy)3]2+ dye [16]. There is a stability difference in these two Cu(I) complexes. When coordinated with copper, the bpy ligand must undergo a change of configuration, while phen is pre-organized as a chelate (Fig. 17). Both bpy and phen are pop-

ular units of copper dyes. As discussed, common peripheral modifications for improving the performance of DSSCs are based on the incorporation of anchor groups and an extension of the p-conjugation of the dye. The anchor group could contribute to electron injection from the LUMO of the dye molecule to the conduction band of the semiconductor, whilst larger conjugation can stabilize p⁄ orbitals and cause a red-shift in the light absorption. In 1983, Sauvage and co-workers developed several homoleptic bis(diimine) Cu(I) complexes, D1, D2a and D2b, as photosensitizers for DSSCs (Fig. 18) [166]. The introduction of methyl or phenyl groups at the 2,9-positions efficiently screened the metal from being attacked by solvent, as well as prolonging the lifetime of the MLCT state. Further evolution was made by the same group in 1994 [167]. D2a was modified at the para positions of the phenyl rings with sodium carboxylate groups to create the new dye D3, which was then coated upon a TiO2 photoanode together with the I–3/I– electrolyte and platinum counter electrode to assemble a solar cell. Moderate open-circuit photovoltage (Voc = 600 mV) and fill factor (FF = 60%), but low short-circuit photocurrent density (Jsc = 0.6 mA cm2) were recorded (PCE = 0.1%). The bonding position as well as deprotonation of carboxylic acid groups actually militate against direct connection of the dye towards the mesoporous TiO2 surface and thereby lead to unfavorable electron injection into the conduction band of TiO2, with a consequence of a low conversion efficiency. Despite the unoptimized introduction of the anchor, it is a noteworthy step towards the development of homoleptic copper dyes. Based on the above analysis, D4a containing protonated carboxylate groups as semiconductor anchors, was designed and successfully applied in DSSCs (Fig. 19). Bpy derivatives substituted

542

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

Fig. 21. Chemical structures of D11 to D13 for DSSC devices.

at the 6,60 positions with methyl groups and 5,50 positions with two carboxylic acid groups were coordinated with copper ions to synthesize the photosensitizer. This complex furnished a low PCE of 0.45% (grel = 9% of that made with N719) with Voc = 570 mV, Jsc = 1.21 mA cm2 and FF = 65%, which may be attributed to the non-optimal positions of the anchoring groups [168]. Another significant endeavor has been made by Sakaki and co-workers, who demonstrated the possibility of employing homoleptic Cu(I) complexes in DSSC devices. The H atoms at the 4,40 positions of D4a were replaced with CH3 groups to generate D4b [169]. The resulting device showed an improved PCE of 2.5% with Voc of 630 mV and Jsc = 3.9 mA cm2. Despite the missing data of a standard dye for comparison, the primary investigation gave strong confidence of using Cu(I) complexes as photosensitizers for DSSCs and provided evidence that modifications of the ligand structures have an obvious impact on the conversion behavior. Since then, a surge of interest in the area of Cu(I)-based photosensitizers in DSSCs has

Fig. 22. The typical architecture of a heteroleptic Cu(I) complex.

emerged. Although the performance was slightly improved from D3 to D4a and D4b by directly using COOH groups instead of the deprotonated ones, the substituent positions of these dyes were still not optimal. When the COOH groups were introduced at the para-position to the chelate N atom in dyes such as D5a, D5b, D6a, D6b [170], D6c [171] and D6d [172], the formed dyes generally indicated better anchoring properties to semiconductors and enhanced performance of the DSSCs. Besides, the extended conjugation on going from the D5 series to the D6 series increased the spectral response, both in intensity and coverage area. The wavelengths of the absorption maxima of D6c and D6d (475 and 496 nm, respectively) were slightly red-shifted with respect to those of D5a, D5b and D6a at 470 nm, which may be attributed to the strong p-conjugation of the benzene rings as good spacers between the anchors and chelating units. The D6 series in particular showed an excellent capability of converting solar energy to electric current, with the best recorded data of Voc = 590 mV, Jsc = 7.3 mA cm2 and FF = 69% corresponding to PCE = 3.0% (grel = 34%) for D6c, which had the characteristics of having proper anchors and extended conjugation. Although the conjugation in D6d was further extended as compared to D6c, it displayed a slightly lower grel of 28% due to its weak optical absorption and the short lifetime of its excited state. The similar analogues D7a to D7e (Fig. 20), with a larger pconjugated periphery relative to the D6 series, were studied [168]. As stated, introducing both methyl and phenyl groups next to the metal-binding sites can stabilize the excited state, but the latter can lead to two MLCT bands rather than one for the methyl congener and so give rise to bathochromically-shifted light absorption (590–690 nm for the D7 series versus 470–500 nm for the D6 series). However, D7a showed inferior performance, despite the existence of the conjugated phenyl group. The reason is that the a-positions of both chelating bpy ligands were replaced by bulky phenyl groups, resulting in a higher steric hindrance, which could significantly reduce the rate of electron transfer and lower the photocurrent. Usage of a phosphonic acid anchor rather than carboxylic acid is also a promising approach to get the desired device performance. The peripheral furan O-donor in D8 was designed as

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

543

Fig. 23. Chemical structures of D14 to D15 for DSSC devices.

a non-classical anchor group, but this complex was not as good as expected. D8 displayed a low Jsc value and consequently a poor performance [168]. In contrast, D9 indicated a high grel of 56% relative to N719 with Voc of 563 mV, Jsc of 3.6 mA cm2 and FF of 70%, because the phosphonic acid anchoring group could facilitate the photosensitization of TiO2 in a more efficient way [173]. Consequently, many different types of anchors were incorporated into Cu(I) dyes to form D10a to D10f, and the commonly used bpy units were changed to biquinoline. The enhanced conjugation leads to a red-shift of light absorption, but dramatically increases the rigidity of the complexes and enhances the p-p interaction, resulting in the aggregation of dyes on the surface of the semiconductor. Among these, D10a showed a poor performance (Voc = 515 mV, Jsc = 0.206 mA cm2, FF = 71% and PCE = 0.08%) due to the same reason as described for D7a [174]. D10d possessed the longest maximum absorption wavelength at 659 nm, but low oscillator strength, which is related to its absorption intensity. D10f featured the closest energy gap between its LUMO and the conduction band of TiO2, indicating effective electron injection. Regardless of the low conversion ability, the various anchor groups in D10a to D10f open a novel door to develop dyes with better electron transfer behavior [175]. There are still some other Cu(I) complexes without the common bpy or phen segments. D11 and D12 show distinct skeletons as compared to the dyes mentioned above (Fig. 21) [176]. They show wide, panchromatic light harvesting, even down to the NIR region, due to the special coordination geometry around the Cu(I) center, possibly resulting from the loose interaction between the benzene rings and among the halogen atoms. They are brand-new and eminent photosensitizers in DSSCs, though the performance is not good at present. D13 has a unique structure, bearing the ferrocenyl fragments, and thus incorporates two kinds of metal elements in the complex (Fig. 21) [177]. The absorption spectrum of D13 is quite broad and intense, which depicts its good light capture ability. However, the OH anchor group results in an unsatisfactory PCE of 2.57%. Overall, the structural modifications of homoleptic complexes still has not led to remarkable improvements in their performances in DSSCs. The lesser skeletal diversities as compared with heteroleptic complexes and the absence of an electronic ‘‘push-

Fig. 24. Chemical structures of the series D16 for DSSC devices.

pull” segment, which facilitates electron transportation, all seriously limit the further development of homoleptic Cu(I) dyes. With regard to the described shortcomings, heteroleptic Cu(I) complexes have aroused increasing attention for DSSC applications. Nevertheless, homoleptic Cu(I) complex dyes are still valuable and can act as simple models in theoretical investigations for providing insight into the electronic configuration of Cu(I) dyes [178]. 4.1.2. Heteroleptic copper(I) complexes As compared to homoleptic Cu(I) complexes, the heteroleptic counterparts possess higher flexibility in ligand modification and superiority in achieving a ‘‘push-pull” architecture which can facilitate the separation of excitons and efficiently reduce charge recombination. The advantages of heteroleptic Cu(I) complexes, including i) the feasibility to synthesize donor-p-acceptor (D-pA) type complexes with enhanced charge-transporting properties, ii) the possibility to finely tune their optoelectronic properties to a larger extent than homoleptic complexes and iii) the various combination of two ligands to rapidly and easily expand the number of complexes, have stimulated intensive interest from numerous research groups. The typical structure of a heteroleptic complex is shown in Fig. 22, which mainly consist of an ancillary ligand, metal ion and anchoring moiety that is used as the binding group to the semiconductor. The ancillary ligand is employed to harvest the solar energy and acquire an electron from electrolyte, thus allowing the regeneration of the dye molecule. Electronic and structural modification can dramatically affect the HOMO level of a complex and subsequently the electron transfer process. In addition, optimization of the optical properties of a dye mostly depends on the fine-tuning of the ancillary ligand. The metal ion in the complex provides the MLCT state to produce long-lived exciton, while the spacer between the chelating atoms and anchor groups is used to improve the efficiency of the dye. On the other hand, the as-named anchor ligand is to accelerate the process of electron injection into the semiconductor, since the LUMO is usually located in this area. Additionally, the substituents at the positions next to the N,N0 -coordination site could hinder the flattening of the tetrahedral Cu(I) coordination sphere, prevent irreversible oxidation to the Cu(II) ion, and thereby stabilize the excited state of the dye. However, heteroleptic complexes suffer heavily from ligand exchange, which generally makes them difficult to isolate as pure heteroleptic complexes with two different N^N (bpy or phen) ligands. This behavior is indeed an obstacle for the further development of Cu(I) complexes in DSSCs, making the preparation of the heteroleptic complexes difficult. Fortunately, a significant advance appeared with reports on different strategies to get heteroleptic Cu(I) dyes, thus allowing for the fabrication of DSSC devices. The creative approaches are based on the HETPHEN (HETeroleptic PHENanthroline) concept [179] and the so-called ‘‘surfaces-as-ligands and surfaces-as-complexes” strategy [180]. According to the structures, the Cu(I) complexes based on the latter concept will be approximately classified as

544

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

Fig. 25. Chemical structures of the series D17 to D19 for DSSC devices.

Fig. 26. Two alternative pathways based on ‘‘surfaces-as-ligands and surfaces-as-complexes” strategy to prepare heteroleptic Cu(I) dyes [187].

six types: strong p-conjugated, heteroatom-based, halogen incorporated, dipyrrin-dominant as well as donor-containing and bulky phen-based complexes. Here, the strategies for boosting the performance of the dyes in DSSCs, such as by lengthening the conjugation, using phosphonic acid anchors and introducing halogen atoms or electron-rich groups on the peripheries, will be discussed. The HETPHEN concept was proposed by Schmittel and coworkers in [179]. This strategy relies on the usage of the very encumbered 2,9-diarylphenanthroline ligand, which prevents the

formation of a symmetrical complex owing to steric reasons. When this very bulky ligand is mixed with a non-encumbered phen molecule and one equivalent of Cu(I) ions, the exclusive generation of the heteroleptic complex is observed, which is driven by the maximum site occupancy principle and the interplay of pelectronic interactions between the two ligands. Bulky mesityl-substituted phen derivatives were used to avert ligand exchange accompanied by the coordination of the COOHmodified biquinoline to form D14 and D15 (Fig. 23) [181]. D15

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

545

Fig. 27. Chemical structures of the series D20 to D22 for DSSC devices.

was further functionalized with an electron-rich peripheral 4substitued N,N-bis(4-methyoxyphenyl)aniline with respect to D14. Both complexes showed a broad MLCT transition over a wide wavelength range with absorption peaks at around 552 and 582 nm, respectively. After the ageing process, their fabricated DSSCs afforded the performance parameters Voc = 475 mV, Jsc = 2.17 mA cm2, FF = 69% and grel = 11% for D14 and Voc = 465 mV, Jsc = 0.82 mA cm2, FF = 66% and grel = 4% for D15 relative to N719 (PCEref = 6.55%) under AM 1.5 G simulated one-sun condition. The better conversion behavior for D15 than D14 indicated that the donor group may favor the electron transport over the dye. Despite the low efficiencies of D14 and D15, which were ascribed to the low e and poor driving forces for the various interfacial processes, they are the first reported examples of stable heteroleptic bis(diimine) Cu(I) dyes prepared with the HETPHEN concept. In comparison, mesityl groups were added into the anchoring ligand, which chelate to the copper ion together with bpy-based ancillary ligand to form D16a to D16d (Fig. 24) [182]. By the replacement of methyl groups in D16a with electron-rich segments, the photosensitizers D16c and D16d showed improved performance data of Jsc = 10.86 mA cm2, grel = 63% for D16c and Jsc = 10.13 mA cm2, grel = 60% for D16d, as compared with Jsc = 3. 76 mA cm2, grel = 20% for D16a. The remarkable rise in Jsc and consequently grel of D16c and D16d was partly due to the presence of ILCT bands in the visible region. In particular, adding chenodeoxycholic acid (CDCA) as the co-adsorbent significantly increased the overall PCE of the D16 series, with broader spectral signals. It could be ascribed to the increased driving force for electron injection and impediment for aggregation of the dyes after the modification of TiO2 with CDCA. Complexes containing donor-based ligands, D17a to D17c, or bis[2-(diphenylphosphanyl)phenyl]ether (DPEphos), D18a and D18b, showed intense and broad light absorption in the visible region with high e values of 5–7  104 M1 cm1 (Fig. 25) [183], likely arising from the strong ILCT transition accompanied with the MLCT transitions. However, the similar DPEphos supported complexes D19a to D19e gave weak light absorption and thus low PCE values [184]. For instance, D19b displayed the low e = 3.3  104 M1 cm1 and grel = 2% relative to N719. Although further improvement on the photon capture ability is needed, the

D19 series sufficiently illustrates the use of the HETPHEN concept in DSSCs and also opens a new way to impede the previously mentioned MLCT flattening by DPEphos, other than using a-substituted bpy or phen ligands. Another efficient approach is to take advantage of the surface of a semiconductor to obtain heteroleptic Cu(I) complexes in a stepwise fashion, which was first demonstrated by Thummel and Hu [185], and then developed, widely applied and named as ‘‘surfaces-as-ligands, surfaces-as-complexes” by Constable and Housecroft [180,186,187]. The schematic procedure is displayed in Fig. 26. This concept encourages the development of heteroleptic Cu(I) complexes. A number of ligands were selected to chelate to the copper ion, leading to some versatile photosensitizers in DSSCs. Based on this concept, several series of photosensitizers, D20 to D26, which possess largely p-delocalized segments, were developed (Figs. 27 and 28). Comparisons were made between D20a and D21a as well as between D20b and D21b. The improvement in efficiencies (grel = 32% for D20a versus 18% for D21a, and 32% for D20b versus 23% for D21b, relative to N719) revealed that extension of the conjugation is an efficient strategy for improving performance. Besides, the D20 series benefited from the large degree of conjugation to show broad and bathochromicallyshifted solar absorption coverage [187]. Among the D21 series, D21a (grel = 18%) performed better than D21c (grel = 28%) [188], which indicated the important role of the spacer between the anchor and the binding unit. The increased PCEs on going from D21d (grel = 5%) [188] to D21b (grel = 23%) to D21c (grel = 28%) demonstrated the obvious superiority of phosphonic acid as the anchor group with respect to carboxylic acid and the phosphonate group. D22 [188] had similar design characteristics as D20a, with larger p-delocalization and a phosphonic acid anchor. It showed encouraging photovoltaic parameters, affording an overall conversion efficiency of grel = 31% relative to N719. Photosensitizers D23 to D26 all employ phosphonic acid as the anchoring group (Fig. 28) [190]. Compared with D23, one and two methyl substituents were introduced next to the binding sites of bpy to form D24 and D25a, respectively, leading to an increase in the degree of blocking on the copper center in the order D25a > D24 > D23. Upon switching the photosensitizer from D23 to D24 and then to D25a, a better set of parameters, Voc = 524 mV,

546

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

Fig. 28. Chemical structures of the series D23 to D26 for DSSC devices.

Jsc = 3.13 mA cm2 for D23 and Voc = 548 mV, Jsc = 4.18 mA cm2 for D24 and Voc = 566 mV, Jsc = 5.2 mA cm2 for D25, were achieved, affording an overall conversion efficiency of grel = 16, 24 and 30%, respectively, relative to N719 (PCEref = 7.17%). As for D25b and D25c (second- and third-generation of D25a) and D26 (the Ru(II) substituted derivative of D24), they are all multinuclear metal complexes. Among the D25 series, more Cu(I) ions are available for oxidation in D25b and D25c relative to D25a, which resulted in an increase in current along the sequence D25c > D25b > D25a. The assembled DSSCs showed a small rise in Voc (566 to 588 mV) but a dramatic decrease in Jsc (5.20 to 2.69 mA cm2), and so performed poorer in their conversion behavior on going from the first- to the second- and the third-generation dyes. This may be ascribed to the quenching of the excited state arising from aggregation of the rod-shaped complexes [191]. On the other hand, D26 displayed a wider light absorption and a higher PCE than D24, predominantly due to the presence of the Ru(II) unit. The aforementioned beneficial effect of the phosphonic acid anchor has been further validated in heteroatom based complexes. After dyeing TiO2 with the series of dyes D27 to D30 (Fig. 29), the resulting DSSCs gave the best grel values of around 30% relative to

N719 for the photosensitizers D30a to D30f, followed by D28a to D28f, with grel values of around 25% and D27a to D27f with grel values of around 11%. The worst grel values of around 7% were detected for D29a to D29f. The large increase in efficiency for the D30 series over the others is obviously due to the added phosphonic acid substituent. The D28 series performed better than the D27 series on account of the spacer between the carboxylic acid and bpy unit, as mentioned before. In addition, the type of heteroatoms introduced into the R group did not dramatically affect the performance [192]. The thiophene-based complexes D31a to D31e, D32 [193] and D33a to D33f (Fig. 30) [194] were characterized by their intense and largely covered absorptions, which also owned suitable HOMO/LUMO energy levels as the dyes in DSSCs. Fabrication of DSSCs was then attempted using the 2-(pyridin-2-yl)-3H-indole derivatives D34a to D34d, among which the best performance was recorded to be grel = 45% for D34a. The DSSCs containing those complexes, however, showed an ageing process [195]. Very recently, D34c was employed together with the commercially available organic dye 2,4-bis[4-(N,N-dibutylamino)-2,6-dihydroxy phenyl]squaraine (SQ2) as a co-sensitizer in the DSSC to signifi-

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

547

Fig. 29. Chemical structures of the series D27 to D30 for DSSC devices.

Fig. 30. Chemical structures of the series D31 to D34 for DSSC devices.

cantly enhance the performance, with a grel value up to 65.6% being achieved after 7-day ageing [196]. It represents the first example using a co-sensitization approach by a Cu(I) complex with an organic dye to create a highly efficient copper-based DSSC. A series of ionic Cu(I) photosensitizers D35a to D35e and D36a to D36e with a bromo-substituted ancillary ligand have been developed (Fig. 31). An improvement in the PCEs of the D36 series relative to the D35 series was made in a similar way (i.e., by introducing a spacer between the anchor and bpy unit) as the D27 and D28 series. The maximum efficiency climbed to 48% relative to N719 for D36e [197]. The obtained efficiency data of D37a (grel = 16%) and D37b (grel = 6%) [180] are inferior to those of the

similar structures D36a (grel = 39%) and D36e (grel = 30%), respectively. It is because the bulky phenyl groups at the a-positions of bpy could reduce the driving force for converting light into current despite a bathochromic effect displayed in the absorption. Furthermore, by the replacement of Br in D36a (Voc = 527 mV, Jsc = 6.01 mA cm2 and grel = 39%) with I to form D38 [198], enhanced performance data of Voc = 604 mV, Jsc = 7.1 mAcm2 and grel = 41% relative to N719 (PCEref = 7.63%) were obtained. Similar to D36a and D37a, the assembled DSSCs showed distinct grel values between D39a to D39d (9%) and D40a to D40d (32%), due to their different steric hindrances [199]. Overall, the conversion behavior of the halogen incorporated copper dyes is still challenging even though

548

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

Fig. 31. Chemical structures of the series D35 to D40 for DSSC devices.

Fig. 32. Chemical structures of the series D41 to D43 for DSSC devices.

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

549

Fig. 33. Chemical structures of D44 to D46 for DSSC devices.

Fig. 34. Chemical structures of the series D47 to D49 for DSSC devices.

the desired D-p-A motif is absent in this kind of complexes. Compared with other metal-organic molecules without Br atoms, the HOMO/LUMO levels were not changed when Br replacement occurred, but the energy levels were indeed different from the original ones, which promoted electron capture from the electrolyte and then boosted the performance. In addition, they all displayed ageing processes. The structures of the dyes D41a, D41b, D41c and D41d, containing 6,60 -bis(trifluoromethyl)-2,20 -bipyridine, 6-trifluoromethyl-2,2 0 -bipyridine, 6,6-bis(methyl)-2,20 -bipyridine and 6-methyl-2,20 bipyridine, respectively, are shown in Fig. 32 [200]. The performance of their DSSC devices follows the order D41a (grel = 36%) > D41b (grel = 35%) > D41d (grel = 32%) > D41c (grel = 28%). The good efficiencies of D41a and D41b can be attributed to the incorporation of the strong electron-withdrawing group CF3 either at the 6 or 6,60 positions of the bipyridine ring, because they can forcefully stabilize the HOMO levels. The same trend was also

observed on going from D42b to D42d to D42c as well as from D43b to D43d to D43c, which further confirmed the advantages of the CF3 group. Different from the D41 series, thiophene was introduced into the anchoring ligand to assemble the D42 and D43 series. For the D42 series, the phosphonic acid groups were located at the a-position of thiophene while for the D43 series, they were at the b-position. The rise in grel from 35% to 37% and then to 42% for D41b, D42b and D43b, respectively, indicated that thiophene has a positive effect on the dye in DSSC applications, especially when the anchor was located at its b-position. Similarly, the global PCE was increased along the series D41c, D42c and D43c, as well as along D41d, D42d and D43d. Generally, dipyrrin moieties exhibit excellent optical properties. This concept initiated the molecular design of D44, D45 and D46 (Fig. 33) [201], which unfortunately did not show any superiority in their photovoltaic performance. Poor efficiency data (Voc = 520 mV, Jsc = 1.21 mA cm2, FF = 64% and PCE = 0.41%) was obtained

550

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

Fig. 35. Chemical structures of the series D50 to D51 for DSSC devices.

Fig. 36. Chemical structures of the series D52 to D54 for DSSC devices.

after employing the best performing D46 as the photosensitizer on TiO2. Despite the poor device behavior, the PCE of D45 (0.31%) was higher than that of D44 (0.13%) because of the extended conjugation in the former, which can stabilize the MLCT transition. The carboxylic acid anchor also facilitates the electron injection process. Moreover, the positive effect of CF3 on the performance of

dye in DSSC was also demonstrated in D46 as compared to the other two complexes. Designing ‘‘push-pull” dyes is one of the most popular strategies for enhancing the performance of heteroleptic Cu(I) complexes in DSSCs. It mainly involves introducing electron-rich groups to functionalize the periphery as hole-transporting units. As a reference,

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

551

Fig. 37. Chemical structures of the series D55 to D59 for DSSC devices.

Fig. 38. Chemical structures of the series D60 to D61 for DSSC devices.

promising performances were achieved for D47a to D47b [202] and D48a to D49f (grel values were mostly around 30% relative to N719) (Fig. 34) [203]. In this case, the introduction of donor groups could provide electrons rapidly to the oxidized copper center and thus regenerate the dyes. Enhanced conversion efficiencies were obtained when phenyl spacers were introduced into D47a to form D48a as well as into D47b to form D49a, further evidencing the

importance of adding a spacer, as previously stated. Compared with D47a (the first-generation), D47b (the second-generation) carrying more electron-rich diphenylamine units at the 4,40 positions gave higher global efficiencies. The same trend could be observed on dyes with sterically unencumbered substituents in the D48 and D49 series, with PCE enhancement on going from the first- to the second-generation. These correspond to the

552

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

Fig. 40. Chemical structures of B1 to B2 for BHJSCs.

Fig. 39. Chemical structures of the series D62 to D64 for DSSC devices.

increase in their solid-state absorbance. In addition to an investigation on the effect of 6,60 substituents on the global efficiency, a series of R groups (Me, n-Bu, iso-Bu, hexyl, Ph and 2-naphthyl) with different steric hindrance were designed. The dyes with the methyl (i.e., D48a and D49a, grel = 29% and 33%, respectively) or n-butyl substituent (i.e., D48d and D49d, grel = 30% and 26%, respectively) performed better than those with bulky groups, such as 2naphthyl (i.e., D48f and D49f, grel = 20% and 13%, respectively). The solvents (in this case, CH2Cl2 and acetone) used in the fabrication of the DSSCs also had an impact on the performances of the dyes. However, regardless of the improved efficiencies of the D49 series relative to the D48 series, the large number of benzene rings within the second-generation dyes largely increases the risk of aggregation. To effectively address the problem, as described before, a co-adsorbent (CDCA) was added to these DSSCs, leading to a significant rise in both Jsc and Voc [204]. The above strategy of designing an electron-donating motif could also be applied to phen-based analogues such as D50a to D50d and D51 (Fig. 35) [205]. The D50 series uses the bulky phen unit containing a long alkyl chain to impede excessive loading of dye on the semiconductor. The grel values of 29, 28, 26 and 30% were measured for D50a to D50d, respectively. D50d displayed the highest PCE among the D50 series, which may arise from the good electron-donating ability of the carbazole substituent. Relative to the D50 series, D51 bearing two donor peripheries showed a better grel value of 34%. The result was consistent with those discussed for the D48 and D49 series. An improved light to electrical energy conversion of D52c (Voc = 570 mV, Jsc = 0.54 mA cm2, FF = 69%) was observed with respect to D52a and D52b (Fig. 36), which has been attained after optimization of the phen unit with donor fragments and the use of a Co(II)/Co(III) couple as the redox mediator [206]. There is also a class of Cu(I) photosensitizers bearing the phen unit that can show strong rigidity, and typical examples include those in the D53 and D54 series [207]. They display absorption profiles in the range 300–700 nm. Considering the balance between conjugation and steric hindrance, D53g and D54f showed the most intense absorptions, while D53a, D53b, D54a and D54b had the weakest ones. All the Cu(I) dyes in the series D55 to D59 showed poor efficiencies in DSSCs (Fig. 37). While the performance was slightly

improved on going from corrole-based D55 to D56, it was still awful [208]. The best result was obtained for D56c (PCE = 0.04%) among the D55 and D56 series. The diphenylamine unit in D56c is believed to donate electrons to regenerate the oxidized dye, while the long alkyl chain was designed to prevent dye aggregation. The interlocked D57a and D57b owned energetically favorable LUMO levels and thus could inject electrons into the conductive band quickly [209]. The dye D58, a novel Cu(I) complex derived from a polypyridine chelating ligand and two triphenylphosphine ligands, plays a significant role in stabilizing the supramolecular network that influences the photophysical properties [210]. D59 featuring a special three-coordinate structure, gave a global PCE of 0.77% under AM 1.5G simulated one sun condition [211].

4.1.3. Dinuclear and polymeric copper complexes Studies on dinuclear complexes or polymer-based photosensitizers in DSSCs are less developed than those for their mononuclear counterparts. Herein, we will give a brief overview on these two systems. In recent years, a series of diphenylphosphino-based ligands was successfully synthesized and coordinated with copper to form D60a, D60b [212] and D61a, D61b (Fig. 38) [213]. Similar values of Jsc and FF for D60a and D60b were observed, while the Voc and global PCE of D60b were somewhat higher than that of D60a, revealing an improvement in performance by changing the steric hindrance of the ligand. However, the limited Jsc gave rise to the awfully low absolute PCEs of 0.025 and 0.027% for D60a and D60b, respectively. As for D61a and D61b, their performances in DSSCs were poor with respective PCEs of 0.03 and 0.02%, and both corresponded to 1% relative to N719 (100%), probably due to the less efficient electron injection into the conduction band of TiO2 without the aid of the anchors. With regard to polymeric dyes, little progress has been made due to the fact that the process of forming thin films of coordination polymers is still relatively challenging. The primary investigation was done by Okubo et al. utilizing hexamethylene dithiocarbamate chelate based linear halide-bridged Cu(I)–Cu(II) coordination polymers D62a and D62b [214] and their derivatives D63a and D63b [215] to photosensitize semiconductors in DSSCs (Fig. 39). Nevertheless, they generally showed an extremely unsatisfactory conversion capability (ca. PCE = 0.03–0.05%), partially because of the lack of anchor groups required to closely connect to the TiO2 surface. Apart from the thiocarbamate derivatives, the polymers D64a and D64b unveiled the benefits of introducing donor motifs into the side-chain in improving the device performance [216]. The absolute PCE values of D64a and D64b are 1.88 and 2.42%, respectively.

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

4.2. Copper complexes for organic solar cells Currently, several of the reported BHJSCs have achieved high PCE values in excess of 8% [217], and even to 11.7% [218]. Among those organic photovoltaic devices, a polymer-based inverted cell realized 9.2% PCE [219], while a small molecular based one led to a value of 8.94% [220]. Although high-performing BHJSCs have been dominated by those containing polymer-based active layers, considerable attention towards small molecule-based solar cells has continuously increased due to the ease of thickness control during the fabrication procedure and the prominent stability of the materials such as Cu(II) phthalocyanine (CuPc). Together with the prevailing family of fullerene compounds as n-type acceptors, metal phthalocyanines like the commonly used CuPc and its derivatives are the most popular p-type semiconductors as donor materials, which could benefit hole-transporting and control the absorption behavior of the devices. The breaking of the 5% PCE barrier confirms that Cu complexes as the donor or acceptor have great potential in BHJSCs [221]. In this section, our motive lies in unveiling the chemical strategies towards enhanced properties, such as better charge carrier transport, elevated spectral sensitivity and energetically matched HOMO/LUMO energy levels of the donor and acceptor through structural tuning of copper complex-based absorption materials, without elaborating other factors for performance in connection with the device fabrication, which has been covered elsewhere [222]. Although the number of cases that employs copper complexes as absorption materials in BHJSCs is not as many as that in DSSCs, we would still like to review some typical or meaningful structures to show the relevant potential and significance of the field for future research development. 4.2.1. Copper(II) phthalocyanine and its derivatives CuPc B1 (Fig. 40) is one of the most popular p-type semiconductors serving as an electron donor which dominates holetransporting and light absorption in BHJSCs while C60 and its derivatives are widely used as electron acceptor materials in devices thanks to their high electron mobilities, long exciton diffusion lengths and complementary absorption spectra relative to that of the donor materials [223–229]. Bruder et al. [230] employed B1 as the donor accompanied by C60 as the acceptor to fabricate a typical bilayer heterojunction solar cell, the PCE of which reached 1.42%. The recorded Jsc = 4.48 mA cm2 and FF =

553

61% were encouraging, but the obtained Voc = 520 mV remains to be improved. Another meaningful work was conducted by Li and co-workers, in which a BHJSC was fabricated with B1 as the acceptor and 4,40 ,400 -tris[phenyl(m-tolyl)amino]triphenylamine (mMTDATA) as the donor [231]. The device with the structure ITO/ PEDOT:PSS (20 nm)/m-MTDATA (30 nm)/B1 (60 nm)/LiF (1 nm)/Al (100 nm) produced a Jsc value of 0.0546 mA cm2, FF of 30.4% and relatively high Voc of 1050 mV under illumination with 365 nm UV light at 1.7 mW cm2. The moderate PCE of 1.03% was obtained because the energies of both the LUMO (1.9 eV) and HOMO (5.1 eV) levels [232] of m-MTDATA are sufficiently higher than those of B1 (LUMO = 3.1 eV and HOMO = 5.3 eV) [233], giving an energetic preference for an exciton reaching the D/A interface to separate and then leave a negative polaron on the acceptor and a positive polaron on the donor, respectively [140]. It is worth noting that its absorption spectrum, peaking at 360 nm, is located from 300 to 420 nm, which mostly corresponds to UV light absorption of both m-MTDATA and B1, according to the findings. Therefore, the copper complex-based system provides an option for expanding the spectral response of heterojunction solar cells. Further evolution based on this concept was carried out through the addition of fluorine atoms towards the periphery of the aromatic rings to enhance the n-type character of metallophthalocyanines. The resulting F16CuPc, B2 (Fig. 40), exhibited a high electron mobility of 5  103 cm2 V1 s1 [234,235]. Based on the energy level offset of 1.1 eV between the HOMO of parasexiphenyl (p-6P) [236] and the LUMO of B2 [237], the bilayer heterojunction solar cell [238] incorporating B2 as an acceptor and p-6P as a donor was successfully assembled, which showed the photovoltaic effect with Voc of 420 mV, Jsc of 0.96 mA cm2 and an overall PCE of 0.18% under light illumination. The PCE of the B2-based device is truly low, which is likely to result from the high series resistance and low mobility of the p-6P film, since B2 is an efficient electron transporter. The narrow energy bandgap of only 1.5 eV for B2 gives rise to its wide absorption spectrum from 550 to 850 nm, which covers the maximum wavelength of the solar photon flux. Besides, the excellent air-stability of B2 improved the environmental stability of the solar cells, which may be attributed to the blocking effect of the hydrophobic B2 towards the harmful diffusion of water and air contaminants into the devices [239]. Generally, CuPc derivatives display an increased conversion efficiency in comparison to metal-free phthalocyanine derivatives.

Fig. 41. The structure of B3 and the sandwich B3-based solar cell with a distinct device construction.

554

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

Fig. 42. Chemical structures of the series B4 to B5 for BHJSCs.

Given such an assumption, a copper complex of the phthalocyanine monomer B3 (Fig. 41) was prepared and used for photocurrent measurements [240]. However, the performance of the cell containing B3 as the donor and a fullerene derivative as the acceptor was slightly worse than that with the copper-free counterpart (Jsc of 13.42 lA cm2 and Voc of 370 mV). Nonetheless, the in situ polymerization of active materials is still appealing for the generation of distinctive sandwich solar cells, with a bulk heterojunction layer between the single layers, due to the improvements in current and voltage compared to either pure bilayers or pure bulk heterojunctions. 4.2.2. Other copper(I) complexes for BHJSCs Apart from the phthalocyanine analogue, the structural diversity of Cu(I) complexes used in BHJSCs has been expanded as well. Several rare examples of homoleptic Cu(I) N^N systems, B4a to B4f (Fig. 42), that absorb light up to 900 nm have been reported by Papanikolaou and co-workers [241]. This class of light-absorbing complexes possesses high e values spanning from 1  105 to 2  105 M1 cm1 with maximal absorption wavelengths of around 580 nm. Studies on the concerted stereo-electronic effects of the substituents from B4a to B4f indicate that extended pconjugation of the acenaphthenequinone skeleton induces an increasing absorption in the visible and near-infrared (NIR) regions of the spectra. B4f, with a strong electron-withdrawing CF3 group, shows the highest value of e, while B4a, bearing an electrondonating OMe group, displays the strongest bathochromic shift among the series. The bathochromic behavior, hyperchromic effect and electronic capacity of the substituents control both the intensity of light absorption and the energy of the MLCT band, which all pave the way to develop tailored absorption layers in BHJSCs. Alternatively, the complexes B5a and B5b (Fig. 42) represent the second exemplary cases which diversify the structural architectures of active materials in BHJSCs. The recorded performance data of B5a-based BHJSCs show little difference from that of B5b, with Voc values of 430 and 650 mV, Jsc of 4.51 and 3.85 mA cm2, and FF of 38% and 34%, leading to low PCEs of 0.72% and 0.86%, respectively. Moreover, B5a and B5b exhibit very similar absorption behavior, which mainly stems from the excitation of the MLCT band. Investigations of a wider variety of optically superior benzoporphyrin derivatives still make sense for attaining further improved performance in benzoporphyrin-based BHJSCs [242]. 5. Conclusion In this work, we have reviewed a number of Cu(I) complexes used in both light emitting and solar energy harvesting devices. In the former application, Cu(I) complexes were recently regarded as a promising alternative for high performance OLEDs. In general,

several approaches have been demonstrated to reduce the structural distortion upon excitation, hence boosting both PLQYs and device efficiencies. The molecular engineering approaches include increasing the bulkiness and rigidity of the ligands (i.e., alkylation at the 2,9 positions of phen and sterically hindered P^P ligands), replacing CAH by NAH/NAN bonds to reduce the vibrational frequency and using three-coordinate mononuclear/four-coordinate dinuclear complexes. On the other hand, an in-depth theoretical study on the phosphorescence and TADF pathways in Cu(I) emitters allows rational manipulation of the energy difference between singlet and triplet excited states (DE(S1–T1)) in order to harvest more triplet excitons at ambient temperature by TADF. Dual emissive materials further enhance the device efficiency using a balanced fraction between the two pathways. With regard to the device fabrication, introducing different hole/electrontransporting moieties into the ligand or using hole/electrontransporting materials with appropriate triplet energies are proved to improve the charge transport and recombination and to confine the triplet exciton in the emissive layer effectively. In addition, the co-deposition of copper precursors and ligands remove the concern of degradation in solution and simplify the processing steps, which is hardly achieved in conventional Ir(III)-based devices. In OLEDs, the best device performances with EQEmax of 23.0, 16.57 and 10.17% for yellow, green and orange-red colors were achieved, respectively. Not all of them have comparable device performances with state-of-art Ir(III)/Pt(II)-based devices, but the device efficiency can be boosted by carefully designing both the molecular and device structures. As a result, an in-depth investigation and development of Cu(I)-based TADF/dual emissive complexes will be part of the study in this field in the future. On the other hand, numerous strategies for boosting the PCE of solar cells have been devised. The simple synthetic procedure for homoleptic Cu(I) complexes allowed the first demonstration of Cu(I)-based DSSCs. The energy level can be tuned by substituents with various steric/electronic effects and different degrees of p-conjugation. Bearing a thiophene-based chelate, complex D9 indicates its high potential with PCE of 56% relative to N719. To break down the symmetry of the molecule, the introduction of a push-pull structure into complexes resulted in heteroleptic Cu(I) complexes via the HETPHEN approach and the ‘‘surfaces-asligands and surfaces-as-complexes” concept. The structural modifications of the complexes here for enhancing the performance mainly include extending the p-conjugation, using a phosphonic acid anchor and introducing halogen atoms or electron-rich units. There are also reports on the improved performance of DSSCs by the addition of a co-adsorbent or co-sensitizer. Among the diverse complexes used as dyes, D16c containing the NEt2 ligand shows an enhanced capability in converting photons to electrical current with a recorded grel = 63% relative to N719. Alternatively, the use

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

of Cu(I)-based materials in heterojunction solar cells is relatively in its infancy. The fabricated BHJ device incorporating B1 as an acceptor was studied and a maximum efficiency of 1.03% was obtained. The enhanced properties of these copper coordinated molecules, such as intense light absorption, broad spectral response, better carrier transport, energetically matched LUMO/conduction band between dye and semiconductor in DSSCs and HOMO/LUMO energy levels between donor and acceptor in BHJ can be achieved by chemical strategies, which open up an exciting option in obtaining better performing OPV devices. Acknowledgements Y. Liu and S.-C. Yiu contributed equally to this work. C.-L. Ho thanks the Hong Kong Research Grants Council (PolyU 123021/17P), National Natural Science Foundation of China (21504074) and the Hong Kong Polytechnic University for their financial support. W.-Y. Wong thanks the Hong Kong Research Grants Council (HKBU 12304715), Areas of Excellence Scheme of the University Grants Committee of HKSAR (AoE/P-03/08) and the Hong Kong Polytechnic University (1-ZE1C) for financial support, and Ms. Clarea Au for the Endowed Professorship in Energy (847S). References [1] V.W.-W. Yam, K.M.-C. Wong, Luminescent molecular rods – transition-metal alkynyl complexes, in: Molecular Wires and Electronics, Springer Berlin Heidelberg, 2005, pp. 1–32. [2] Y. Chi, B. Tong, P.-T. Chou, Coord. Chem. Rev. 281 (2014) 1. [3] C.W. Tang, Appl. Phys. Lett. 48 (1986) 183. [4] C.W. Tang, S.A. VanSlyke, Appl. Phys. Lett. 51 (1987) 913. [5] A.K. Asatkar, A. Bedi, S.S. Zade, Isr. J. Chem. 54 (2014) 467. [6] G.L. Schulz, S. Holdcroft, Chem. Mater. 20 (2008) 5351. [7] M.W. Blaskie, D.R. McMillin, Inorg. Chem. 19 (1980) 3519. [8] C.O. Dietrich-Buchecker, P.A. Marnot, J.-P. Sauvage, J.R. Kirchhoff, D.R. McMillin, J. Chem. Soc., Chem. Commun. (1983) 513. [9] A.K. Ichinaga, J.R. Kirchhoff, D.R. McMillin, C.O. Dietrich-Buchecker, P.A. Marnot, J.P. Sauvage, Inorg. Chem. 26 (1987) 4290. [10] R. Czerwieniec, M.J. Leitl, H.H.H. Homeier, H. Yersin, Coord. Chem. Rev. 325 (2016) 2. [11] V.W.-W. Yam, K.M.-C. Wong, Chem. Soc. Rev. 28 (1999) 323. [12] Y. Ma, C.-M. Che, H.-Y. Chao, X. Zhou, W.-H. Chan, J. Shen, Adv. Mater. 11 (1999) 852. [13] Y.-G. Ma, W.-H. Chan, X.-M. Zhou, C.-M. Che, New J. Chem. 23 (1999) 263. [14] M. Magni, P. Biagini, A. Colombo, C. Dragonetti, D. Roberto, A. Valore, Coord. Chem. Rev. 322 (2016) 69. [15] M. Sandroni, Y. Pellegrin, F. Odobel, C. R. Chim. 19 (2016) 79. [16] N. Armaroli, Chem. Soc. Rev. 30 (2001) 113. [17] N. Armaroli, G. Accorsi, F. Cardinali, A. Listorti, Photochemistry and photophysics of coordination compounds: copper, in: V. Balzani, S. Campagna (Eds.), Photochemistry and Photophysics of Coordination Compounds I, Springer, Berlin Heidelberg, 2007, pp. 69–115. [18] M.W. Mara, K.A. Fransted, L.X. Chen, Coord. Chem. Rev. 282–283 (2015) 2. [19] M. Iwamura, S. Takeuchi, T. Tahara, J. Am. Chem. Soc. 129 (2007) 5248. [20] N.A. Gothard, M.W. Mara, J. Huang, J.M. Szarko, B. Rolczynski, J.V. Lockard, L.X. Chen, J. Phys. Chem. A 116 (2012) 1984. [21] O. Green, B.A. Gandhi, J.N. Burstyn, Inorg. Chem. 48 (2009) 5704. [22] M. Sandroni, A. Maufroy, M. Rebarz, Y. Pellegrin, E. Blart, C. Ruckebusch, O. Poizat, M. Sliwa, F. Odobel, J. Phys. Chem. C 118 (2014) 28388. [23] S. Haneder, E. Da Como, J. Feldmann, J.M. Lupton, C. Lennartz, P. Erk, E. Fuchs, O. Molt, I. Münster, C. Schildknecht, G. Wagenblast, Adv. Mater. 20 (2008) 3325. [24] T. Sajoto, P.I. Djurovich, A. Tamayo, M. Yousufuddin, R. Bau, M.E. Thompson, R. J. Holmes, S.R. Forrest, Inorg. Chem. 44 (2005) 7992. [25] J.A.G. Williams, A. Beeby, E.S. Davies, J.A. Weinstein, C. Wilson, Inorg. Chem. 42 (2003) 8609. [26] A.F. Rausch, L. Murphy, J.A.G. Williams, H. Yersin, Inorg. Chem. 48 (2009) 11407. [27] T. Sajoto, P.I. Djurovich, A.B. Tamayo, J. Oxgaard, W.A. Goddard, M.E. Thompson, J. Am. Chem. Soc. 131 (2009) 9813. [28] A. Islam, N. Ikeda, K. Nozaki, Y. Okamoto, B. Gholamkhass, A. Yoshimura, T. Ohno, Coord. Chem. Rev. 171 (1998) 355. [29] M.J. Leitl, V.A. Krylova, P.I. Djurovich, M.E. Thompson, H. Yersin, J. Am. Chem. Soc. 136 (2014) 16032. [30] C.L. Linfoot, M.J. Leitl, P. Richardson, A.F. Rausch, O. Chepelin, F.J. White, H. Yersin, N. Robertson, Inorg. Chem. 53 (2014) 10854. [31] M. Osawa, Chem. Commun. 50 (2014) 1801.

555

[32] H. Yersin, A.F. Rausch, R. Czerwieniec, T. Hofbeck, T. Fischer, Coord. Chem. Rev. 255 (2011) 2622. [33] W.C.H. Choy, W.K. Chan, Y. Yuan, Adv. Mater. 26 (2014) 5368. [34] W.-Y. Wong, C.-L. Ho, Coord. Chem. Rev. 253 (2009) 1709. [35] X. Yang, G. Zhou, W.-Y. Wong, Chem. Soc. Rev. 44 (2015) 8484. [36] P.-T. Chou, Y. Chi, M.-W. Chung, C.-C. Lin, Coord. Chem. Rev. 255 (2011) 2653. [37] Y. Tao, K. Yuan, T. Chen, P. Xu, H. Li, R. Chen, C. Zheng, L. Zhang, W. Huang, Adv. Mater. 26 (2014) 7931. [38] R.D. Costa, E. Ortí, H.J. Bolink, F. Monti, G. Accorsi, N. Armaroli, Angew. Chem. Int. Ed. 51 (2012) 8178. [39] D. Volz, M. Wallesch, C. Flechon, M. Danz, A. Verma, J.M. Navarro, D.M. Zink, S. Brase, T. Baumann, Green Chem. 17 (2015) 1988. [40] Y. Luo, H. Aziz, Adv. Funct. Mater. 20 (2010) 1285. [41] C.-J. Chiang, A. Kimyonok, M.K. Etherington, G.C. Griffiths, V. Jankus, F. Turksoy, A.P. Monkman, Adv. Funct. Mater. 23 (2013) 739. [42] W. Li, Y. Pan, R. Xiao, Q. Peng, S. Zhang, D. Ma, F. Li, F. Shen, Y. Wang, B. Yang, Y. Ma, Adv. Funct. Mater. 24 (2014) 1609. [43] A. Endo, M. Ogasawara, A. Takahashi, D. Yokoyama, Y. Kato, C. Adachi, Adv. Mater. 21 (2009) 4802. [44] A. Endo, K. Sato, K. Yoshimura, T. Kai, A. Kawada, H. Miyazaki, C. Adachi, Appl. Phys. Lett. 98 (2011) 83302. [45] H. Uoyama, K. Goushi, K. Shizu, H. Nomura, C. Adachi, Nature 492 (2012) 234. [46] H. Yersin, R. Czerwieniec, M.Z. Shafikov, A.F. Suleymanova, ChemPhysChem 18 (2017) 3508. [47] M. Hashimoto, S. Igawa, M. Yashima, I. Kawata, M. Hoshino, M. Osawa, J. Am. Chem. Soc. 133 (2011) 10348. [48] D. Volz, D.M. Zink, T. Bocksrocker, J. Friedrichs, M. Nieger, T. Baumann, U. Lemmer, S. Bräse, Chem. Mater. 25 (2013) 3414. [49] Z. Wang, Y. Lou, S. Naka, H. Okada, A.C.S. Appl, Mater. Interfaces 3 (2011) 2496. [50] M.T. Buckner, D.R. McMillin, J. Chem. Soc., Chem. Commun. (1978) 759. [51] C.E.A. Palmer, D.R. McMillin, C. Kirmaier, D. Holten, Inorg. Chem. 26 (1987) 3167. [52] G. Blasse, D.R. McMillin, Chem. Phys. Lett. 70 (1980) 1. [53] C.E.A. Palmer, D.R. McMillin, Inorg. Chem. 26 (1987) 3837. [54] P.A. Breddels, P.A.M. Berdowski, G. Blasse, D.R. McMillin, J. Chem. Soc., Faraday Trans. 2 (78) (1982) 595. [55] R.A. Rader, D.R. McMillin, M.T. Buckner, T.G. Matthews, D.J. Casadonte, R.K. Lengel, S.B. Whittaker, L.M. Darmon, F.E. Lytle, J. Am. Chem. Soc. 103 (1981) 5906. [56] S.-M. Kuang, D.G. Cuttell, D.R. McMillin, P.E. Fanwick, R.A. Walton, Inorg. Chem. 41 (2002) 3313. [57] D.G. Cuttell, S.-M. Kuang, P.E. Fanwick, D.R. McMillin, R.A. Walton, J. Am. Chem. Soc. 124 (2002) 6. [58] K. Kubicˇek, S. Thekku Veedu, D. Storozhuk, R. Kia, S. Techert, Polyhedron 124 (2017) 166. [59] Q. Zhang, Q. Zhou, Y. Cheng, L. Wang, D. Ma, X. Jing, F. Wang, Adv. Mater. 16 (2004) 432. [60] A. Lavie-Cambot, M. Cantuel, Y. Leydet, G. Jonusauskas, D.M. Bassani, N.D. McClenaghan, Coord. Chem. Rev. 252 (2008) 2572. [61] Q. Zhang, T. Komino, S. Huang, S. Matsunami, K. Goushi, C. Adachi, Adv. Funct. Mater. 22 (2012) 2327. [62] Q. Zhang, J. Ding, Y. Cheng, L. Wang, Z. Xie, X. Jing, F. Wang, Adv. Funct. Mater. 17 (2007) 2983. [63] G. Che, Z. Su, W. Li, B. Chu, M. Li, Z. Hu, Z. Zhang, Appl. Phys. Lett. 89 (2006) 103511. [64] V. Bulovic´, A. Shoustikov, M.A. Baldo, E. Bose, V.G. Kozlov, M.E. Thompson, S.R. Forrest, Chem. Phys. Lett. 287 (1998) 455. [65] Z. Su, G. Che, W. Li, W. Su, M. Li, B. Chu, B. Li, Z. Zhang, Z. Hu, Appl. Phys. Lett. 88 (2006) 213508. [66] Z. Su, W. Li, G. Che, M. Xu, Z. Kong, D. Wang, Q. Xin, L. Han, B. Chu, D. Bi, Appl. Phys. Lett. 90 (2007) 143505. [67] M.A. Baldo, M.E. Thompson, S.R. Forrest, Nature 403 (2000) 750. [68] Z. Si, J. Li, B. Li, S. Liu, W. Li, J. Lumin. 128 (2008) 1303. [69] H. Ge, W. Wei, P. Shuai, G. Lei, S. Qing, J. Lumin. 131 (2011) 238. [70] D. Ma, Y. Qiu, L. Duan, Adv. Funct. Mater. 26 (2016) 3438. [71] W.Y. Wong, G.J. Zhou, X.M. Yu, H.S. Kwok, Z. Lin, Adv. Funct. Mater. 17 (2007) 315. [72] F. Zhang, Y. Guan, X. Chen, S. Wang, D. Liang, Y. Feng, S. Chen, S. Li, Z. Li, F. Zhang, C. Lu, G. Cao, B. Zhai, Inorg. Chem. 56 (2017) 3742. [73] L. Bergmann, C. Braun, M. Nieger, S. Brase, Dalton Trans. 47 (2018) 608. [74] Z. Si, J. Li, B. Li, S. Liu, W. Li, J. Lumin. 129 (2009) 181. [75] T. Yu, P. Liu, H. Chai, J. Kang, Y. Zhao, H. Zhang, D. Fan, J. Fluoresc. 24 (2014) 933. [76] D. Zhang, J. Lumin. 130 (2010) 1419. [77] L. Zhang, B. Li, Z. Su, J. Phys. Chem. C 113 (2009) 13968. [78] A. Wada, Q. Zhang, T. Yasuda, I. Takasu, S. Enomoto, C. Adachi, Chem. Commun. 48 (2012) 5340. [79] E.C. Riesgo, Y.-Z. Hu, F. Bouvier, R.P. Thummel, D.V. Scaltrito, G.J. Meyer, Inorg. Chem. 40 (2001) 3413. [80] X.-L. Chen, R. Yu, Q.-K. Zhang, L.-J. Zhou, X.-Y. Wu, Q. Zhang, C.-Z. Lu, Chem. Mater. 25 (2013) 3910. [81] Q. Zhang, X.-L. Chen, J. Chen, X.-Y. Wu, R. Yu, C.-Z. Lu, RSC Adv. 5 (2015) 34424. [82] X.-L. Chen, C.-S. Lin, X.-Y. Wu, R. Yu, T. Teng, Q.-K. Zhang, Q. Zhang, W.-B. Yang, C.-Z. Lu, J. Mater. Chem. C 3 (2015) 1187.

556

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557

[83] D. Liang, X.-L. Chen, J.-Z. Liao, J.-Y. Hu, J.-H. Jia, C.-Z. Lu, Inorg. Chem. 55 (2016) 7467. [84] Q. Zhang, X.-L. Chen, J. Chen, X.-Y. Wu, R. Yu, C.-Z. Lu, Dalton Trans. 44 (2015) 10022. [85] J. Min, Q. Zhang, W. Sun, Y. Cheng, L. Wang, Dalton Trans. 40 (2011) 686. [86] L. Bergmann, J. Friedrichs, M. Mydlak, T. Baumann, M. Nieger, S. Brase, Chem. Commun. 49 (2013) 6501. [87] C.-W. Hsu, C.-C. Lin, M.-W. Chung, Y. Chi, G.-H. Lee, P.-T. Chou, C.-H. Chang, P.Y. Chen, J. Am. Chem. Soc. 133 (2011) 12085. [88] Y. Sun, V. Lemaur, J.I. Beltrán, J. Cornil, J. Huang, J. Zhu, Y. Wang, R. Fröhlich, H. Wang, L. Jiang, G. Zou, Inorg. Chem. 55 (2016) 5845. [89] R. Czerwieniec, J. Yu, H. Yersin, Inorg. Chem. 50 (2011) 8293. [90] S. Igawa, M. Hashimoto, I. Kawata, M. Yashima, M. Hoshino, M. Osawa, J. Mater. Chem. C 1 (2013) 542. [91] R. Czerwieniec, H. Yersin, Inorg. Chem. 54 (2015) 4322. [92] M. Osawa, I. Kawata, R. Ishii, S. Igawa, M. Hashimoto, M. Hoshino, J. Mater. Chem. C 1 (2013) 4375. [93] G. Cheng, G.K.-M. So, W.-P. To, Y. Chen, C.-C. Kwok, C. Ma, X. Guan, X. Chang, W.-M. Kwok, C.-M. Che, Chem. Sci. 6 (2015) 4623. [94] X. Li, H. Yan, Q. Zhao, Chem. Eur. J. 22 (2016) 1888. [95] G.K.-M. So, G. Cheng, J. Wang, X. Chang, C.-C. Kwok, H. Zhang, C.-M. Che, Chem. Asian J. 12 (2017) 1490. [96] O. Moudam, A. Kaeser, B. Delavaux-Nicot, C. Duhayon, M. Holler, G. Accorsi, N. Armaroli, I. Seguy, J. Navarro, P. Destruel, J.-F. Nierengarten, Chem. Commun. (2007) 3077. [97] J. Zhang, C. Duan, C. Han, H. Yang, Y. Wei, H. Xu, Adv. Mater. 28 (2016) 5975. [98] T. Hofbeck, U. Monkowius, H. Yersin, J. Am. Chem. Soc. 137 (2015) 399. [99] V.A. Krylova, P.I. Djurovich, B.L. Conley, R. Haiges, M.T. Whited, T.J. Williams, M.E. Thompson, Chem. Commun. 50 (2014) 7176. [100] M. Osawa, M. Hoshino, M. Hashimoto, I. Kawata, S. Igawa, M. Yashima, Dalton Trans. 44 (2015) 8369. [101] D. Di, A.S. Romanov, L. Yang, J.M. Richter, J.P.H. Rivett, S. Jones, T.H. Thomas, M. Abdi Jalebi, R.H. Friend, M. Linnolahti, M. Bochmann, D. Credgington, Science 356 (2017) 159. [102] M. Gernert, U. Müller, M. Haehnel, J. Pflaum, A. Steffen, Chem. Eur. J. 23 (2017) 2206. [103] S. Shi, L.R. Collins, M.F. Mahon, P.I. Djurovich, M.E. Thompson, M.K. Whittlesey, Dalton Trans. 46 (2017) 745. [104] A.S. Romanov, D. Di, L. Yang, J. Fernandez-Cestau, C.R. Becker, C.E. James, B. Zhu, M. Linnolahti, D. Credgington, M. Bochmann, Chem. Commun. 52 (2016) 6379. [105] S.B. Harkins, J.C. Peters, J. Am. Chem. Soc. 127 (2005) 2030. [106] J.C. Deaton, S.C. Switalski, D.Y. Kondakov, R.H. Young, T.D. Pawlik, D.J. Giesen, S.B. Harkins, A.J.M. Miller, S.F. Mickenberg, J.C. Peters, J. Am. Chem. Soc. 132 (2010) 9499. [107] S.B. Harkins, N.P. Mankad, A.J.M. Miller, R.K. Szilagyi, J.C. Peters, J. Am. Chem. Soc. 130 (2008) 3478. [108] H. Araki, K. Tsuge, Y. Sasaki, S. Ishizaka, N. Kitamura, Inorg. Chem. 44 (2005) 9667. [109] M. Wallesch, D. Volz, D.M. Zink, U. Schepers, M. Nieger, T. Baumann, S. Bräse, Chem. Eur. J. 20 (2014) 6578. [110] D. Volz, M. Wallesch, S.L. Grage, J. Göttlicher, R. Steininger, D. Batchelor, T. Vitova, A.S. Ulrich, C. Heske, L. Weinhardt, T. Baumann, S. Bräse, Inorg. Chem. 53 (2014) 7837. [111] A. Tsuboyama, K. Kuge, M. Furugori, S. Okada, M. Hoshino, K. Ueno, Inorg. Chem. 46 (2007) 1992. [112] Y. Okano, H. Ohara, A. Kobayashi, M. Yoshida, M. Kato, Inorg. Chem. 55 (2016) 5227. [113] F. Wei, T. Zhang, X. Liu, X. Li, N. Jiang, Z. Liu, Z. Bian, Y. Zhao, Z. Lu, C. Huang, Org. Electron. 15 (2014) 3292. [114] Z. Liu, M.F. Qayyum, C. Wu, M.T. Whited, P.I. Djurovich, K.O. Hodgson, B. Hedman, E.I. Solomon, M.E. Thompson, J. Am. Chem. Soc. 133 (2011) 3700. [115] Z. Liu, J. Qiu, F. Wei, J. Wang, X. Liu, M.G. Helander, S. Rodney, Z. Wang, Z. Bian, Z. Lu, M.E. Thompson, C. Huang, Chem. Mater. 26 (2014) 2368. [116] X. Liu, T. Zhang, T. Ni, N. Jiang, Z. Liu, Z. Bian, Z. Lu, C. Huang, Adv. Funct. Mater. 24 (2014) 5385. [117] F. Wei, J. Qiu, X. Liu, J. Wang, H. Wei, Z. Wang, Z. Liu, Z. Bian, Z. Lu, Y. Zhao, C. Huang, J. Mater. Chem. C 2 (2014) 6333. [118] Z. Wang, J. Zhu, Z. Liu, P. Wu, H. Wang, Z. Zhang, B. Wei, J. Mater. Chem. C 5 (2017) 6982. [119] L. Lin, D.-H. Chen, R. Yu, X.-L. Chen, W.-J. Zhu, D. Liang, J.-F. Chang, Q. Zhang, C.-Z. Lu, J. Mater. Chem. C 5 (2017) 4495. [120] Y.-M. Wang, F. Teng, Y.-B. Hou, Z. Xu, Y.-S. Wang, W.-F. Fu, Appl. Phys. Lett. 87 (2005) 233512. [121] X. Yang, X. Yan, H. Guo, B. Liu, J. Zhao, G. Zhou, Y. Wu, Z. Wu, W.-Y. Wong, Dyes Pigm. 143 (2017) 151. [122] D.M. Zink, T. Baumann, J. Friedrichs, M. Nieger, S. Bräse, Inorg. Chem. 52 (2013) 13509. [123] D.M. Zink, D. Volz, T. Baumann, M. Mydlak, H. Flügge, J. Friedrichs, M. Nieger, S. Bräse, Chem. Mater. 25 (2013) 4471. [124] D. Volz, Y. Chen, M. Wallesch, R. Liu, C. Fléchon, D.M. Zink, J. Friedrichs, H. Flügge, R. Steininger, J. Göttlicher, C. Heske, L. Weinhardt, S. Bräse, F. So, T. Baumann, Adv. Mater. 27 (2015) 2538. [125] M. Wallesch, A. Verma, C. Fléchon, H. Flügge, D.M. Zink, S.M. Seifermann, J.M. Navarro, T. Vitova, J. Göttlicher, R. Steininger, L. Weinhardt, M. Zimmer, M.

[126] [127] [128] [129] [130] [131] [132] [133] [134] [135] [136] [137] [138] [139] [140] [141] [142] [143] [144] [145] [146] [147] [148] [149] [150] [151] [152] [153]

[154] [155] [156] [157] [158] [159] [160] [161] [162] [163] [164] [165] [166] [167] [168] [169] [170]

[171] [172] [173] [174]

[175] [176] [177] [178] [179] [180]

Gerhards, C. Heske, S. Bräse, T. Baumann, D. Volz, Chem. Eur. J. 22 (2016) 16400. Y. Fang, W. Liu, S.J. Teat, G. Dey, Z. Shen, L. An, D. Yu, L. Wang, D.M. O’Carroll, J. Li, Adv. Funct. Mater. 27 (2017) 1603444. X.-L. Chen, R. Yu, X.-Y. Wu, D. Liang, J.-H. Jia, C.-Z. Lu, Chem. Commun. 52 (2016) 6288. N. Mitsuharu, G. Yasuyuki, E. Masanao, Chem. Lett. 32 (2003) 32. L. Zhang, B. Li, J. Electrochem. Soc. 156 (2009) J174. M. Xie, C. Han, J. Zhang, G. Xie, H. Xu, Chem. Mater. 29 (2017) 6606. R.F. Service, Science 309 (2005) 548. M.A. Green, K. Emery, Y. Hishikawa, W. Warta, Prog. Photovoltaics Res. Appl. 18 (2010) 144. W. Shockley, H.J. Queisser, J. Appl. Phys. 32 (1961) 510. M. Grätzel, Inorg. Chem. 44 (2005) 6841. G. Niu, X. Guo, L. Wang, J. Mater. Chem. A 3 (2015) 8970. S.D. Stranks, H.J. Snaith, Nat. Nanotechnol. 10 (2015) 391. Q. Lin, A. Armin, P.L. Burn, P. Meredith, Acc. Chem. Res. 49 (2016) 545. S.T. Williams, A. Rajagopal, C.-C. Chueh, A.K.Y. Jen, J. Phys. Chem. Lett. 7 (2016) 811. G. Yu, J. Gao, J.C. Hummelen, F. Wudl, A.J. Heeger, Science 270 (1995) 1789. J. Nelson, Mater. Today 5 (2002) 20. M. Gratzel, Nature 414 (2001) 338. M. Grätzel, Acc. Chem. Res. 42 (2009) 1788. B. O’Regan, M. Gratzel, Nature 353 (1991) 737. A. Hagfeldt, G. Boschloo, L. Sun, L. Kloo, H. Pettersson, Chem. Rev. 110 (2010) 6595. D.J. Milliron, I. Gur, A.P. Alivisatos, MRS Bull. 30 (2011) 41. I. Gonzalez-Valls, M. Lira-Cantu, Energy Environ. Sci. 2 (2009) 19. K.M. Coakley, Y. Liu, C. Goh, M.D. McGehee, MRS Bull. 30 (2011) 37. B.C. Thompson, J.M.J. Fréchet, Angew. Chem. Int. Ed. 47 (2008) 58. S. Günes, H. Neugebauer, N.S. Sariciftci, Chem. Rev. 107 (2007) 1324. A.M. Spokoyny, T.C. Li, O.K. Farha, C.W. Machan, C. She, C.L. Stern, T.J. Marks, J. T. Hupp, C.A. Mirkin, Angew. Chem. Int. Ed. 49 (2010) 5339. A.J. Nozik, J. Miller, Chem. Rev. 110 (2010) 6443. M.K. Nazeeruddin, C. Klein, P. Liska, M. Grätzel, Coord. Chem. Rev. 249 (2005) 1460. S. Caramori, C.A. Bignozzi, Recent developments in the design of dyesensitized solar cell components, in: Electrochemistry of Functional Supramolecular Systems, John Wiley & Sons, Inc., Hoboken, NJ, USA, 2010, pp. 523–579. J.-L. Bredas, J.R. Durrant, Acc. Chem. Res. 42 (2009) 1689. H.J. Snaith, Adv. Funct. Mater. 20 (2010) 13. M.K. Nazeeruddin, E. Baranoff, M. Grätzel, Sol. Energy 85 (2011) 1172. S. Caramori, V. Cristino, R. Boaretto, R. Argazzi, C.A. Bignozzi, A. Di Carlo, Int. J. Photoenergy 2010 (2010) Article ID 458614. K. Kakiage, Y. Aoyama, T. Yano, K. Oya, J.-I. Fujisawa, M. Hanaya, Chem. Commun. 51 (2015) 15894. V. Chaurin, E.C. Constable, C.E. Housecroft, New J. Chem. 30 (2006) 1740. P.G. Bomben, K.C.D. Robson, B.D. Koivisto, C.P. Berlinguette, Coord. Chem. Rev. 256 (2012) 1438. Y.-S. Yen, H.-H. Chou, Y.-C. Chen, C.-Y. Hsu, J.T. Lin, J. Mater. Chem. 22 (2012) 8734. S. Zhang, X. Yang, Y. Numata, L. Han, Energy Environ. Sci. 6 (2013) 1443. H.J. Snaith, Nat. Photonics 6 (2012) 337. H.J. Snaith, Energy Environ. Sci. 5 (2012) 6513. D.R. McMillin, K.M. McNett, Chem. Rev. 98 (1998) 1201. V.E.N. Alonso-Vante, P. Chartier, C.O. Dietrich-Buchecker, D.R. Mcmillin, P.A. Marnot, J.P. Sauvage, Nouv. J. Chim. 7 (1983) 3. N. Alonso-Vante, J.-F. Nierengarten, J.-P. Sauvage, J. Chem. Soc., Dalton Trans. (1994) 1649. E.C. Constable, A.H. Redondo, C.E. Housecroft, M. Neuburger, S. Schaffner, Dalton Trans. (2009) 6634. S. Sakaki, T. Kuroki, T. Hamada, J. Chem. Soc., Dalton Trans. (2002) 840. T. Bessho, E.C. Constable, M. Graetzel, A. Hernandez Redondo, C.E. Housecroft, W. Kylberg, M.K. Nazeeruddin, M. Neuburger, S. Schaffner, Chem. Commun. (2008) 3717. A. Colombo, C. Dragonetti, D. Roberto, A. Valore, P. Biagini, F. Melchiorre, Inorg. Chim. Acta 407 (2013) 204. Y.-J. Yuan, Z.-T. Yu, J.-Y. Zhang, Z.-G. Zou, Dalton Trans. 41 (2012) 9594. K.A. Wills, H.J. Mandujano-Ramírez, G. Merino, G. Oskam, P. Cowper, M.D. Jones, P.J. Cameron, S.E. Lewis, Dyes Pigm. 134 (2016) 419. K.A. Wills, H.J. Mandujano-Ramirez, G. Merino, D. Mattia, T. Hewat, N. Robertson, G. Oskam, M.D. Jones, S.E. Lewis, P.J. Cameron, RSC Adv. 3 (2013) 23361. J. Baldenebro-López, N. Flores-Holguín, J. Castorena-González, D. GlossmanMitnik, Polyhedron 82 (2014) 33. J.W. Kee, Y.Y. Ng, S.A. Kulkarni, S.K. Muduli, K. Xu, R. Ganguly, Y. Lu, H. Hirao, H.S. Soo, Inorg. Chem. Front. 3 (2016) 651. R. Yadav, Y. Waghadkar, G. Kociok-Köhn, A. Kumar, S.B. Rane, R. Chauhan, Opt. Mater. 62 (2016) 176. M.W. Mara, N.E. Jackson, J. Huang, A.B. Stickrath, X. Zhang, N.A. Gothard, M.A. Ratner, L.X. Chen, J. Phys. Chem. B 117 (2013) 1921. M. Schmittel, A. Ganz, Synlett 1997 (1997) 710. S.Y. Brauchli, F.J. Malzner, E.C. Constable, C.E. Housecroft, RSC Adv. 5 (2015) 48516.

Y. Liu et al. / Coordination Chemistry Reviews 375 (2018) 514–557 [181] M. Sandroni, M. Kayanuma, A. Planchat, N. Szuwarski, E. Blart, Y. Pellegrin, C. Daniel, M. Boujtita, F. Odobel, Dalton Trans. 42 (2013) 10818. [182] M. Sandroni, L. Favereau, A. Planchat, H. Akdas-Kilig, N. Szuwarski, Y. Pellegrin, E. Blart, H. Le Bozec, M. Boujtita, F. Odobel, J. Mater. Chem. A 2 (2014) 9944. [183] M. Sandroni, M. Kayanuma, M. Rebarz, H. Akdas-Kilig, Y. Pellegrin, E. Blart, H. Le Bozec, C. Daniel, F. Odobel, Dalton Trans. 42 (2013) 14628. [184] C.L. Linfoot, P. Richardson, T.E. Hewat, O. Moudam, M.M. Forde, A. Collins, F. White, N. Robertson, Dalton Trans. 39 (2010) 8945. [185] R.P. Thummel, Y.-Z. Hu, U.S. Pat. Appl. Publ. (2004), US 20040225122 A1 20041111. [186] A.H. Rendondo, E.C. Constable, C.E. Housecroft, Chimia 63 (2009) 205. [187] C.E. Housecroft, E.C. Constable, Chem. Soc. Rev. 44 (2015) 8386. [188] B. Bozic-Weber, V. Chaurin, E.C. Constable, C.E. Housecroft, M. Meuwly, M. Neuburger, J.A. Rudd, E. Schonhofer, L. Siegfried, Dalton Trans. 41 (2012) 14157. [189] F.J. Malzner, S.Y. Brauchli, E. Schönhofer, E.C. Constable, C.E. Housecroft, Polyhedron 82 (2014) 116. [190] E. Schönhofer, B. Bozic-Weber, C.J. Martin, E.C. Constable, C.E. Housecroft, J.A. Zampese, Dyes Pigm. 115 (2015) 154. [191] A. Mishra, M.K.R. Fischer, P. Bäuerle, Angew. Chem. Int. Ed. 48 (2009) 2472. [192] B. Bozic-Weber, E.C. Constable, C.E. Housecroft, P. Kopecky, M. Neuburger, J.A. Zampese, Dalton Trans. 40 (2011) 12584. [193] S. Wei, X. Shi, X. Lu, Y. Shao, D. Jin, Z. Deng, Z. Zhao, K. Li, W. Guo, Sci. Adv. Mater. 7 (2015) 1361. [194] S. Wei, K. Li, X. Lu, Z. Zhao, Y. Shao, Y. Dang, S. Li, W. Guo, Mater. Chem. Phys. 173 (2016) 139. [195] F.J. Malzner, A. Prescimone, E.C. Constable, C.E. Housecroft, M. Willgert, J. Mater. Chem. A 5 (2017) 4671. [196] F.J. Malzner, M. Willgert, E.C. Constable, C.E. Housecroft, J. Mater. Chem. A 5 (2017) 13717. [197] B. Bozic-Weber, S.Y. Brauchli, E.C. Constable, S.O. Furer, C.E. Housecroft, F.J. Malzner, I.A. Wright, J.A. Zampese, Dalton Trans. 42 (2013) 12293. [198] F.J. Malzner, S.Y. Brauchli, E.C. Constable, C.E. Housecroft, M. Neuburger, RSC Adv. 4 (2014) 48712. [199] A. Buttner, S.Y. Brauchli, R. Vogt, E.C. Constable, C.E. Housecroft, RSC Adv. 6 (2016) 5205. [200] Y.M. Klein, M. Willgert, A. Prescimone, E.C. Constable, C.E. Housecroft, Dalton Trans. 45 (2016) 4659. [201] T.E. Hewat, L.J. Yellowlees, N. Robertson, Dalton Trans. 43 (2014) 4127. [202] B. Bozic-Weber, S.Y. Brauchli, E.C. Constable, S.O. Furer, C.E. Housecroft, I.A. Wright, Phys. Chem. Chem. Phys. 15 (2013) 4500. [203] S.Y. Brauchli, B. Bozic-Weber, E.C. Constable, N. Hostettler, C.E. Housecroft, J. A. Zampese, RSC Adv. 4 (2014) 34801. [204] S.Y. Brauchli, F.J. Malzner, E.C. Constable, C.E. Housecroft, RSC Adv. 4 (2014) 62728. [205] S.O. Furer, B. Bozic-Weber, M. Neuburger, E.C. Constable, C.E. Housecroft, RSC Adv. 5 (2015) 69430. [206] L.N. Ashbrook, C.M. Elliott, J. Phys. Chem. C 117 (2013) 11883. [207] S. Wei, Y. Shao, X. Shi, X. Lu, K. Li, Z. Zhao, C. Guo, H. Zhu, W. Guo, Org. Electron. 29 (2016) 142. [208] K. Sudhakar, L. Giribabu, P. Salvatori, F.D. Angelis, Phys. Status Solidi A 212 (2015) 194. [209] M.W. Mara, D.N. Bowman, O. Buyukcakir, M.L. Shelby, K. Haldrup, J. Huang, M.R. Harpham, A.B. Stickrath, X. Zhang, J.F. Stoddart, A. Coskun, E. Jakubikova, L.X. Chen, J. Am. Chem. Soc. 137 (2015) 9670.

557

[210] A. Baez-Castro, J. Baldenebro-Lopez, L. Ceballos-Mendivil, P.P. Roman-Bravo, H. Hopfl, V. Miranda-Soto, D. Glossman-Mitnik, A. Cruz-Enriquez, J.J. CamposGaxiola, Acta Cryst., Sect. C: Cryst. 73 (2017) 280. [211] S. Vuong, H.-M. Nguyen-Dang, Q.T. Tran, T.T.T. Luong, T.T.T. Pham, T. NguyenTran, A.T. Mai, J. Electron. Mater. 46 (2017) 3639. [212] T.-H. Huang, J. Yan, Y.-F. Liu, Y.-T. Xie, C. Jia, Aust. J. Chem. 68 (2015) 1144. [213] T.-H. Huang, J. Yan, H. Yang, H.-M. Du, M.-H. Zhang, J. Coord. Chem. 68 (2015) 1514. [214] N. Tanaka, T. Okubo, H. Anma, K.H. Kim, Y. Inuzuka, M. Maekawa, T. KurodaSowa, Eur. J. Inorg. Chem. 2013 (2013) 3384. [215] T. Okubo, N. Tanaka, H. Anma, K.H. Kim, M. Maekawa, T. Kuroda-Sowa, Polymers 4 (2012) 1613. [216] J. Hu, W. Zhang, Q. Xie, D. Peng, Y. Liu, Y. Liao, C. Zhu, C. Zhong, J. Mater. Sci. Mater. Electron. 25 (2014) 5144. [217] R.F. Service, Science 332 (2011) 293. [218] http://www.m-kagaku.co.jp/english/r_td/strategy/technology/topics/opv. Announced by Mitsubishi Chemical Corporation accessed: November 2013. [219] Z. He, C. Zhong, S. Su, M. Xu, H. Wu, Y. Cao, Nat. Photon. 6 (2012) 591. [220] A.K.K. Kyaw, D.H. Wang, D. Wynands, J. Zhang, T.-Q. Nguyen, G.C. Bazan, A.J. Heeger, Nano Lett. 13 (2013) 3796. [221] M.Y. Chan, S.L. Lai, M.K. Fung, C.S. Lee, S.T. Lee, Appl. Phys. Lett. 90 (2007) 23504. [222] C.-F. Lin, M. Zhang, S.-W. Liu, T.-L. Chiu, J.-H. Lee, Int. J. Mol. Sci. 12 (2011) 476. [223] J.J.M. Halls, K. Pichler, R.H. Friend, S.C. Moratti, A.B. Holmes, Synth. Met. 77 (1996) 277. [224] L.A.A. Pettersson, L.S. Roman, O. Inganäs, J. Appl. Phys. 86 (1999) 487. [225] L.S. Roman, W. Mammo, L.A.A. Pettersson, M.R. Andersson, O. Inganäs, Adv. Mater. 10 (1998) 774. [226] J.W. Arbogast, A.P. Darmanyan, C.S. Foote, F.N. Diederich, R.L. Whetten, Y. Rubin, M.M. Alvarez, S.J. Anz, J. Phys. Chem. 95 (1991) 11. [227] J.W. Arbogast, C.S. Foote, M. Kao, J. Am. Chem. Soc. 114 (1992) 2277. [228] L. Ma, J. Ouyang, Y. Yang, Appl. Phys. Lett. 84 (2004) 4786. [229] P. Peumans, S.R. Forrest, Appl. Phys. Lett. 79 (2001) 126. [230] I. Bruder, J. Schöneboom, R. Dinnebier, A. Ojala, S. Schäfer, R. Sens, P. Erk, J. Weis, Org. Electron. 11 (2010) 377. [231] L.L. Chen, W.L. Li, H.Z. Wei, B. Chu, B. Li, Sol. Energy Mater. Sol. Cells 90 (2006) 1788. [232] K. Itano, H. Ogawa, Y. Shirota, Appl. Phys. Lett. 72 (1998) 636. [233] J.-M. Nunzi, C. R. Phys. 3 (2002) 523. [234] Z. Bao, A.J. Lovinger, J. Brown, J. Am. Chem. Soc. 120 (1998) 207. [235] S.P. Singh, A. Sellinger, A. Dodabalapur, J. Appl. Phys. 107 (2010) 44509. [236] N. Koch, A. Kahn, J. Ghijsen, J.-J. Pireaux, J. Schwartz, R.L. Johnson, A. Elschner, Appl. Phys. Lett. 82 (2003) 70. [237] C. Shen, A. Kahn, J. Appl. Phys. 90 (2001) 4549. [238] X. Jiang, J. Dai, H. Wang, Y. Geng, D. Yan, Chem. Phys. Lett. 446 (2007) 329. [239] D.Q. Gao, M.Y. Chan, S.W. Tong, F.L. Wong, S.L. Lai, C.S. Lee, S.T. Lee, Chem. Phys. Lett. 399 (2004) 337. [240] K. Lintinen, L. Storbacka, A. Efimov, A. Tolkki, N. Tkachenko, H. Lemmetyinen, Sol. Energy Mater. Sol. Cells 95 (2011) 909. [241] P. Papanikolaou, P.D. Akrivos, A. Czapik, B. Wicher, M. Gdaniec, N. Tkachenko, Eur. J. Inorg. Chem. 2013 (2013) 2418. [242] K. Takahashi, N. Yamada, D. Kumagai, D. Kuzuhara, M. Suzuki, Y. Yamaguchi, N. Aratani, K.-I. Nakayama, H. Yamada, J. Porphyrin Phthalocyanines 19 (2015) 465.