Supporting Information

© Wiley-VCH 2008

69451 Weinheim, Germany

S1

Zipper Assembly of Vectorial Rigid-Rod -Stack Architectures with Red

and Blue Naphthalenediimides: Toward Supramolecular Cascade n/p-

Heterojunctions

Adam L. Sisson,† Naomi Sakai,

† Natalie Banerji,

‡ Alexandre Fürstenberg,

‡ Eric Vauthey,*

,‡ and Stefan

Matile*,†

†Department of Organic Chemistry and

‡Department of Physical Chemistry, University of Geneva,

Geneva, Switzerland. *To whom correspondence should be addressed. E-mail:

eric.vauthey@chiphy.unige.ch, stefan.matile@chiorg.unige.ch

Supplementary Information

Table of Content

1. Materials and methods S2

2. Supplementary text S4

2.1 Synthesis of cationic N,Cl-NDI octamer 2 and N,Cl-NDI monomer 2a S4

2.2 Electrochemistry S8

2.3 Steady-state photophysics S9

2.4 Fluorescence dynamics S10

2.5 Transient absorption spectroscopy S11

2.6 Zipper assembly on gold electrodes S12

3. Supplementary schemes and figures S14

4. Supplementary tables S22

5. Supplementary references S24

S2

1. Materials and methods

As in ref. [S1], Supplementary Information. Briefly, reagents for synthesis were purchased

from Fluka, amino acid derivatives from Novabiochem and Bachem, HATU from Applied

Biosystems, buffers, and salts from Sigma or Fluka-Aldrich. All reactions were performed under N2

or argon atmosphere. Unless stated otherwise, column chromatography was carried out on silica gel

60 (Fluka, 40-63 m). Analytical (TLC) and preparative thin layer chromatography (PTLC) were

performed in silica gel 60 (Fluka, 0.2 mm) and silica gel GF (Analtech, 1000 m), respectively.

HPLC was performed using either Jasco HPLC system (PU-980, UV-970, FP-920) or an Agilent

1100 series apparatus with a photo diode array detector. [ ]20

D values were recorded on a Jasco P-

1030 Polarimeter, melting points (m.p.) on a heating table from Reichert (Austria), IR spectra were

recorded on a Perkin Elmer Spectrum One FT-IR spectrometer (ATR, Golden Gate, unless stated)

and are reported as wavenumbers in cm-1

with band intensities indicated as s (strong), m

(medium), w (weak). ESI-MS were performed on a Finnigan MAT SSQ 7000 instrument or a ESI

API 150EX, HR ESI-MS on a Sciex QSTAR Pulsar mass spectrometer, MALDI-TOF on a Axima

CFR+ (Shimadzu).

1H and

13C spectra were recorded (as indicated) either on a Bruker 300 MHz,

400 MHz or 500 MHz spectrometer and are reported as chemical shifts ( ) in ppm relative to TMS

( = 0). Spin multiplicities are reported as a singlet (s), doublet (d), triplet (t), quartet (q) and quintet

(quint) with coupling constants (J) given in Hz, or multiplet (m). Broad peaks are marked as br. 1H

and 13

C resonances were assigned with the aid of additional information from 1D & 2D NMR

spectra (H,H-COSY, DEPT 135, HSQC and HMBC). UV-Vis spectra were measured either on a

Varian Cary 1 Bio spectrophotometer (solution samples) or Agilent 8453 spectrophotometer (solid

samples) and are reported as maximal absorption wavelength in nm (extinction coefficient in

mM-1

cm-1

). Electrochemical measurements were done on an Electrochemical Analyzer with

S3

Picoamp booster and Faraday cage (CH Instruments 660C). Eppendorf Minispin centrifuge was

used for the precipitation of gold nano-particles. Photocurrents measurements were performed

using a 150 W Xe lamp (Hamamatsu), a monochromator (Instrument SA, H-10 1200UV), and an

Electrochemical Analyzer (CH Instruments 660C). The power of light was measured using a

portable laser power meter (Spectra Physics Model 407A).

Abbreviations. ACN: acetonitrile; Alloc: Allyloxycarbonyl; Au-nps: gold nanoparticles; Boc:

t-Butoxycarbonyl; calcd: Calculated; CS: charge-separated; DMA: N,N-Dimethylaniline; DMF:

N,N-Dimethylformamide; en: Ethylenediamine; ESA: excited state absorption; Fc: Ferrocene; FF:

Fill factor; FWHM: full width at half maximum; Gla: Glycolic acid; Glu: L-Glutamic acid; HATU:

N-[(Dimethylamino)-1H-1,2,3 -triazolo[4,5-b]pyridin-1-ylmethylene]-N-methylmethanammonium

hexafluorophosphate N-oxide; HMB: hexamethylbenzene; HRMS: High resolution mass

spectrometry; Lys: L-Lysine; NDI: Naphthalenediimide; RPHPLC: Reverse phase high

performance liquid chromatography; rt: Room temperature; SE: stimulated emission; sh: Shoulder;

TCNE: Tetracyanoethylene; TCSPC: time-correlated single photon counting; TEA: Triethylamine;

TEOA: Triethanolamine; TFA: Trifluoroacetic acid; TFE: 2,2,2-Trifluoroethanol; Z:

(Benzyloxy)carbonyl.

S4

2. Supplementary text

2.1. Synthesis of cationic N,Cl-NDI octamer 2 and N,Cl-NDI monomer 2a

Alloc-en-[Cl,Cl-NDI]-Lys(Z)-NH2 5. This compound was prepared from 6 following

previously reported procedures.[S1]

Alloc-en-[N,Cl-NDI]-Lys(Z)-NH2 7. Cl,Cl-NDI 5 (170 mg, 0.234 mmol) was dissolved in

stirred isopropylamine (15 ml). After 5 min at rt, the red solution was evaporated to dryness under

reduced pressure. Purification by column chromatography (CH2Cl2/MeOH 98:2; Rf = 0.5 with

CH2Cl2/MeOH 94:6) gave 7 (170 mg, quantitative) as a red solid. [ ]20

D = - 14.5 (c = 1.00 in

MeOH/CH2Cl2 1:1); IR: 3331 (m), 2936 (m), 1668 (s), 1637 (s), 1583 (s), 1444 (s), 1259 (s), 1214

(s), 1142 (s), 992 (m), 915 (m), 790 (m), 733 (m); 1H NMR (400 MHz, CDCl3/CD3OD 6:1, N/N =

regioisomeric equivalents): 9.94/9.85, (d, 3J(H,H) = 7.6/7.2 Hz, 1H), 8.39/8.20 (s, 1H), 8.08/7.93

(s, 1H), 7.39 - 7.15 (m, 5H), 6.16 - 6.12/6.09 - 6.05 (m, 1H), 5.89 - 5.75 (m, 1H), 5.68 - 5.60/5.60 -

5.52 (m, 1H), 5.23 - 4.95 (m, 2H), 4.94/4.93 (s, 2H), 4.44 - 4.35 (m, 2H), 4.19 - 4.09 (m, 1H), 4.10 -

4.00 (m, 2H), 3.41 - 3.25 (m, 2H), 3.12 - 3.05 (m, 2H), 2.48 - 2.25 (m, 1H), 2.21 - 2.09 (m, 1H),

1.61 – 1.35 (m, 4H), 1.49 (br.s, 6H); 13

C NMR (125 MHz, CDCl3/CD3OD 6:1, N/N = regioisomeric

equivalents): 176.7/176.3 (s), 169.3/169.2 (s), 167.1 (s), 165.9/165.7 (s), 165.7 (s), 165.1 (s),

161.0/160.9 (s), 154.8/154.7 (s), 140.5 (s), 138.8/138.5 (s), 136.6 (d), 132.3 – 131.2 (d, 5x),

132.3/132.1 (s), 132.0/131.9 (s), 131.3/131.2 (s), 130.6 (d), 125.7/125.3 (d), 125.2/125.0 (d),

124.9/124.7 (d), 121.2 (t), 103.0/102.7 (s), 70.4 (t), 69.4 (t), 58.7/58.1 (d), 57.5 (d), 44.5/44.4 (t),

44.3 (t), 43.0/42.8 (t), 33.3 (t), 31.9/31.8 (t), 27.6 (t), 26.9 (q), 26.9 (q); MS (ESI, +ve): m/z (%)

769 (25 [M + Na]+), 748 (100 [M + H]

+), 730 (60 [M – OH]

+); HR-MS (ESI, +ve): Calcd for

S5

C37H41O9N6Cl+: 747.2559, Found: 747.2539.

H-en-[N,Cl-NDI]-Lys(Z)-NH2 8. To a solution of 7 (50 mg, 67 mol) in dry CH2Cl2 (5 ml)

were added p-nitrophenol (28 mg, 0.20 mmol) and tributyltin hydride (97 mg, 0.34 mmol) followed

by Pd(PPh3)2Cl2 (1 mg, 0.001 mmol). After stirring for 3 h at rt, the reaction mixture was dried

under reduced pressure and lipophilic impurities were removed by solid-liquid extraction with

heptane (5 5 ml). Further purification by column chromatography (CH2Cl2/MeOH 90:10 then

CH2Cl2/MeOH/TEA 90:10:1; Rf = 0.4 with CH2Cl2/MeOH/TEA 90:10:1) gave 8 (44 mg,

quantitative) as a red solid. [ ]20

D = - 12.8 (c = 1.00 in MeOH); 1H NMR (400 MHz, CDCl3, N/N =

regioisomeric equivalents): 10.08/10.00 (d, 3J(H,H) = 7.6/8.0 Hz, 1H), 8.66/8.64 (s, 1H),

8.35/8.32 (s, 1H), 7.35 - 7.28 (m, 5H), 5.75/5.69 (t, 3J(H,H) = 7.6/7.2 Hz, 1H), 5.05/5.04 (s, 2H),

4.82 (br.s, 1H), 4.32 - 4.25 (m, 2H), 4.30 - 4.15 (m, 1H), 3.25 - 3.15 (m, 2H), 3.12 - 3.06 (m, 2H),

2.41 - 2.29 (m, 2H), 1.80 - 1.51 (m, 4H), 1.50 (d, 3J(H,H) = 6.0 Hz, 6H); MS (ESI, +ve): m/z (%)

664 (80 [M + H]+), 646 (100 [M - OH]

+).

13,2

3,3

2,4

3,5

2,6

3,7

2,8

3-Octakis(Gla-OH)-p-octiphenyl 9. This compound was prepared from

Fast Blue B in overall nine steps following previously reported procedures.[S2]

13,2

3,3

2,4

3,5

2,6

3,7

2,8

3-Octakis(Gla-en-[N,Cl-NDI]-Lys(Z)-NH2)-p-octiphenyl 10. A solution

of 9 (4.5 mg, 0.0037 mmol), HATU (17 mg, 0.045 mmol) and 2,6,di-tert-butylpyridine (69 mg,

0.36 mmol) in dry DMF (1 ml) was stirred for 30 min at rt. Then a solution of 8 (30 mg, 0.045

mmol) and TEA (24 mg, 0.24 mmol) in dry DMF (2 ml) was added. After stirring for 16 h at rt, the

reaction mixture was evaporated to dryness azeotropically with toluene. A preliminary purification

by column chromatography (CH2Cl2/MeOH 85:15) was followed by two PTLCs (first

S6

CH2Cl2/MeOH 85:15, Rf = 0.5, then CH2Cl2/MeOH 94:6, Rf = 0.1) to give 10 (16 mg, 66%) as a red

solid. 1H NMR (400 MHz, CDCl3/CD3OD 1:1): 9.97 - 9.70 (m, 8H), 8.21- 7.84 (m, 16H), 7.45 -

6.95 (m, 24H), 6.81 - 6.78 (m, 2H), 7.25 - 7.15 (m, 40H), 5.61 - 5.39 (m, 8H), 4.97 - 4.81 (m, 16H),

4.49 - 4.21 (m, 16H), 4.21 - 3.79 (m, 24H), 3.57 - 3.30 (m, 16H), 3.10 - 2.96 (m, 16H), 2.30 - 1.90

(m, 16H), 1.59 - 1.37 (m, 16H), 1.35 – 1.11 (m, 64H).

13,2

3,3

2,4

3,5

2,6

3,7

2,8

3-Octakis(Gla-en-[N,Cl-NDI]-Lys-NH2)-p-octiphenyl, TFA salt 2. A

solution of 10 (2 mg, 0.3 µmol) in TFA (1 ml) was stirred for 3h at 50 °C. After this time, the red

solution was evaporated to dryness under reduced pressure. Impurities were removed by solid-

liquid extraction with heptane (3 2 ml), ether (3 2 ml) and CH2Cl2 (3 2 ml), leaving 2 (2 mg,

quantitative) as a red solid. 1H NMR (400 MHz, CD3OD): 8.41 - 7.72 (m, 16H), 7.56 - 6.72 (m,

26H), 5.78 - 5.30 (m, 8H), 4.60 - 4.25 (16H), 4.40 - 3.82 (m, 24H), 3.55 - 3.28 (m, 16H), 3.15 - 2.82

(m, 16H), 2.51 - 2.28 (m, 8H), 2.24 - 2.00 (m, 8H), 1.87 - 1.61 (m, 16H), 1.67 - 1.35 (m, 16H), 1.41

- 1.22 (m, 48H); MS (ESI, +ve): m/z (%) 1765 (10 [M + 3H]3+

), 1346 (25 [M + 4Na]4+

), 1324 (35

[M + 4H]4+

), 1077 (45 [M + H + 4Na]5+

), 1060 (70 [M + 5H]5+

), 898 (55 [M + 2H + 4Na]6+

), 883

(100 [M + 6H]6+

), 770 (40 [M + 3H + 4Na]7+

), 755 (60 [M + 7H]7+

).

Alloc-en-[N,Cl-NDI]-Lys-NH2, TFA salt 2a. A solution of 7 (30 mg, 0.040 mmol) in TFA (2

ml) was stirred for 3 h at 50 °C. After this time, the red solution was evaporated to dryness under

reduced pressure. Impurities were removed by solid-liquid extraction with petroleum ether (3 2

ml) leaving 2a (28 mg, quantitative) as a red solid. [ ]20

D = - 13.2 (c = 1 in MeOH); IR: 2970

(m), 2409 (m), 1667 (s), 1634 (s), 1441 (s), 1200 (s), 1174 (s), 1126 (s), 991 (m), 918 (m), 833 (m),

790 (m), 720 (m); 1H NMR (400 MHz, CD3OD, N/N = regioisomeric equivalents): 8.38/8.29 (s,

1H), 8.14 (s, 1H), 5.82 - 5.71 (m, 1H), 5.70/5.63 (t, 3J(H,H) = 7.6/7.2 Hz, 1H), 5.20 - 5.16 (m, 1H),

S7

5.09 - 5.05 (m, 1H), 4.38/4.37 (s, 2H), 4.29 – 4.23 (m, 1H), 4.18 - 4.05 (m, 2H), 3.37 - 3.28 (m,

2H), 3.07 - 2.98 (m, 2H), 2.55 - 2.46 (m, 1H), 2.31 - 2.18 (m, 1H), 1.91 - 1.72 (m, 2H), 1.69 - 1.52

(m, 2H), 1.59 - 1.50 (m, 6H); 13

C NMR (125 MHz, CD3OD, N/N = regioisomeric equivalents):

190.6 (s), 173.1/172.8 (s), 165.5/165.3 (s), 162.0 (s), 161.9 (s), 161.7/161.6 (s), 160.9/160.8 (s),

157.5 (s), 150.7/150.6 (s), 134.0 (d), 133.7 (d), 133.0/132.9 (d), 132.0/131.9 (s), 131.3/131.2 (s),

130.6 (d), 125.7/125.3 (d), 125.2/125.0 (d), 124.9/124.7 (d), 121.2 (t), 103.0/102.7 (s), 70.4 (t), 69.4

(t), 58.7/58.1 (d), 57.5 (d), 44.5/44.4 (t), 44.3 (t), 43.0/42.8 (t), 33.3 (t), 31.9/31.8 (t), 27.6 (t), 26.9

(q), 26.9 (q); MS (ESI, +ve): m/z (%) 635 (15 [M + Na]+), 613 (100 [M]

+), 596 (50 [M – OH]

+).

N,N-NDI Initiator 1. This compound was prepared following previously reported procedures

(Fig. S1).[S3]

Anionic N,N-NDI octamer 3. This compound was prepared following previously reported

procedures (Fig. S1).[S3]

Cationic N,N-NDI octamer 4. This compound was prepared following previously reported

procedures (Fig. S1).[S1]

N,N-NDI monomer 4a. This compound was prepared following previously reported

procedures (Table S1).[S1]

S8

2.2. Electrochemistry

Oxidation and reduction potential of 7 was determined by cyclic voltammetry vs Fc/Fc+ in

dichloromethane (Figure S2, scan rate 100 mV/s, supporting electrolyte: 100 mM Bu4NPF6,

working electrode: Pt disk, counter electrode: Pt wire, reference electrode: Ag/AgCl). Results are

shown in Table S1 in comparison with literature data for related NDIs 4b,[S4]

4c[S3]

and 2b,[S4]

13,2

3,3

2,4

3-p-quateranisole (11),

[S3] p-quaterphenyl (12) and p-octiphenyl (13).

[S5] HOMO and

LUMO energies were calculated from oxidation and reduction potentials using eq (S1)[S6]

EHOMO/LUMO = - 4.8 eV - E1/2 vs (Fc/Fc+) (S1)

The obtained bandgaps ELUMO – EHOMO were converted into absorption wavelength using eq

(S2)

calc (nm) = hc / (ELUMO – EHOMO) = 1240 / (ELUMO – EHOMO) (S2)

and compared to the measured values max (Table S1, Figure S3, ref. [S7]).

S9

2.3 Steady-state photophysics

The photophysics of rNDIs has been investigated in solution in methanol (MeOH) using steady-

state absorption and steady-state and time-resolved fluorescence techniques. Absorption spectra were

recorded on a Cary 50 spectrophotometer and fluorescence spectra on a Cary Eclipse fluorimeter in

a 1 cm quartz cell. Results are compared to those previously obtained with a bNDI derivative.

Monomer 2a and octamer 2 in MeOH display a single absorption band above 400 nm centered

around 530 nm (Figure S3, Figure 3b). The fluorescence spectrum (excitation at 460 nm, Figure S4)

of 2 and 2a peaks around 570 nm in MeOH. Fluorescence quantum yields, determined as

previously reported for bNDI derivation,[S1]

were low in protic solvents and with rNDI octamer 2

but high in aprotic solvents and with rNDI monomer 2a (Table S2).

The possibility of quenching of rNDI chromophores by the p-octiphenyl scaffolds via

photoinduced electron transfer was demonstrated by mixing 2a with hexamethylbenzene (HMB)

which has the same oxidation potential as the p-quateranisole 11. The fluorescence intensity

decreased markedly with increasing HMB concentration (Figure S5, Figure 2b). From the slope of

the linear regression and using the Stern-Volmer equation for collisional quenching,[S8] a quenching

rate constant of 4·109 s-1·M-1 is extracted, relatively close to the diffusion limit of ~2·1010 s-1·M-1 in

acetonitrile. This indirect evidence was further confirmed by mixing directly 2a with bianisole 14 (170

mM) or quateranisole 11 (29 mM) in acetonitrile: Figures S5 and 2b show that the fluorescence of

2a was efficiently quenched. On the other hand, the fluorescence of bNDI 4a was not altered by the

presence of 32 mM HMB, indicating that only rNDIs can undergo photoinduced charge separation

with the p-octiphenyl scaffold.

S10

2.4 Fluorescence dynamics

The fluorescence dynamics was recorded using the time-correlated single photon counting

(TCSPC) and the fluorescence up-conversion (400 nm excitation, 575 nm detection) techniques. For

TCSPC measurements, excitation was performed at 395 nm with a pulsed laser diode (Picoquant

model LDH-P-C-400B). The pulses had duration of about 65 ps and the average power was about

0.5 mW at 10 MHz. Fluorescence was collected at 90°, and passed through an analyzer set at the

magic angle with respect to the excitation polarization, and a 450 nm-cutoff filter located in front of

a photomultiplier tube (Hamamatsu, H5783-P-01). The detector output was connected to the input

of a TCSPC computer board module (Becker and Hickl, SPC-300-12). The full width at half

maximum (FWHM) of the instrument response was around 200 ps. Measurements were performed

in a 1 cm quartz cell. Data were reproduced by iterative reconvolution of a sum of exponentials to

the measured instrument response function.

The fluorescence up-conversion set-up uses the frequency-doubled output of a Kerr lens mode-

locked Ti:sapphire laser (Tsunami, Spectra-Physics) for excitation of the sample at 400 nm.[S9] The

output pulses centered at 800 nm had duration of 100 fs and repetition rate of 82 MHz.

Experiments were carried out in a 1 mm rotating cell. Fluorescence decays were analyzed by

iterative reconvolution of a sum of exponential functions with a Gaussian response function of 280

fs FWHM. Fluorescence decay kinetics demonstrated that photoinduced charge separation in 2

occurs to 95% within 170 ps (Table S3). A detailed discussion and analysis of the fluorescence

decay kinetics data will be published in a more specialized journal.

S11

2.5 Transient absorption spectroscopy

Transient absorption measurements were performed with 400 M solutions of rNDI monomer

2a in MeOH and with 50 M solutions of rNDI octamer 2 in MeOH in a 1 mm quartz cell.

Excitation at 530 nm was achieved with a home-built two-stage noncolinear optical parametric

amplifier, fed by the 800 nm output of a standard 1 kHz amplified Ti:Sapphire system (Spectra-

Physics). After recompression with a pair of prisms, the pulse duration was of the order of 100 fs.

The energy per pulse at the sample was around 3 µJ. Probing was achieved with a white light

continuum obtained by focussing 800 nm pulses in a H2O/D2O mixture. The probe beam was split

into a pumped signal beam and an un-pumped reference beam before the sample. The transmitted

signal and reference beams were detected by an ORIEL MultispecTM

125 spectrograph coupled to a

CCD detector (Entwicklungsbüro G. Stresing, Berlin). To improve the sensitivity, the pump light

was chopped at half the amplifier frequency, and the transmitted signal intensity was recorded shot

by shot. It was corrected for intensity fluctuations using the reference beam. The transient spectra

were averaged until the desired signal to noise ratio was achieved. The angle between the

polarization of the pump and probe beam was set to 54.7° (magic angle). Artefacts due to parasitic

light (pump light reaching the detectors, dispersed spontaneous fluorescence of the sample) were

subtracted from the transient spectra. All transient spectra were then corrected for the chirp of the

white light.

Data analysis of the transient spectra of 2 in MeOH is complicated by the complete overlap of

the S1 and CS state spectra. Additionally to this, early spectral dynamics due to vibrational and

solvent relaxation is also present. Nonetheless, without a detailed analysis already, it is possible to

see that the charge-separated state decays on a slower time scale (500 ps average time constant)

than with 4 (Figure 2d). A detailed discussion and analysis of the transient absorption data is to be

S12

published in a more specialized journal (global fit, subtraction and iterative deconvolution of

complex kinetics, etc).

2.6 Zipper assembly on gold electrodes

Gold electrodes. Gold electrodes were prepared as reported in ref [S3]: Gold-coated glass

slides (22 x 22 mm2) were purchased from Mivitec GmbH, Analytical -Systems (Germany).

Before use, the plates were cleaned using ‘piranha’ solution (H2SO4 / 30 % H2O2 3 / 1; 35 °C for 5

min).[S10]

Caution: piranha solution reacts violently with organic compounds. It should be handled

with extreme care. After the treatment with piranha solution, the plates were thoroughly rinsed with

water and EtOH, and used immediately.

Initiation. Zipper initiation was done as reported in ref [S3]: The cleaned gold plates were

immersed in the solution of the initiator 1 (~7 M) in a 1:1 mixture of TFE : 1 mM sodium

phosphate buffer (pH 7) for 3 days. The obtained Au-1 electrodes were tested for defects using

the standard procedure in which reduction-oxidation of K3Fe(CN)6 (2 mM in 1 M aqueous KNO3)

was measured by cyclic voltammetry using Au-1 as working electrode.[S3,S11]

The absence of redox

wave confirmed the absence of large uncovered areas on the Au electrode.

Blue multilayers. Au-1-4-3-4 was prepared as reported in ref [S3]: Coated gold electrodes

Au-1 were immersed in the solution of cationic bNDI octamer 4 (~10 M) in a 1 : 1 mixture of TFE

and 0.5 mM sodium phosphate, 0.5 M NaCl buffer (pH 7) for overnight. The plate was rinsed with

bidistilled water and EtOH, and the photocurrent of the resulting plate was recorded (see below).

Obtained two-layers coated plates Au-1-4 were similarly treated with anionic bNDI octamer 3 to

S13

give Au-1-4-3, which in turn was treated with cationic bNDI octamer 4 to give Au-1-4-3-4.

Mixed multilayers (Figs 3a and S7). Mixed multi-layers were built by using cationic rNDI

octamer 2 (~10 M in 1:1 TFE / 0.5 mM sodium phosphate, 0.5 M NaCl buffer, pH 7) in place of

cationic bNDI octamer 4 in the procedures described above.

Photocurrent measurements. Coated gold electrodes were used as a working electrode (Pt

wire as a counter electrode) in an aqueous solution of TEOA (50 mM) and Na2SO4 (0.1 M) and

irradiated with a Xe lamp through water to filter off IR component (area of irradiation ~0.8 cm2).

Changes in current upon on-off switching of irradiations (20 seconds each) were measured at +0.4 V

vs Ag/AgCl unless stated. The power of irradiation was ~1 W cm-2

. For each gold plate,

photocurrent obtained for Au-1 (0.5 to 1.1 A) was used as a standard (=1) and relative increase in

current was reported for the following multi-layers. For the I-V measurements, the electrolyte

solution was deaerated by bubbling N2.

Action spectra (Fig. 3b). Photocurrent densities were measured at open circuit potential in the

dark upon excitation by monochromatic light (power reaching to the sample 1.5 mW cm-2

) and

plotted against the irradiation wavelength.

S14

3. Supplementary schemes and figures

N

NO O

OO

NH2

O

NHAlloc

NHZ

Cl

Cl

O

OO O

OO

Cl

Cl

7: R = Alloc; R' = Z8: R = H; R' = Z2a: R = Alloc; R' = H

N

NO O

OO

NH2

O

NHR

NHR'

NH

Cl

O

O

O

O

O

O

O

O

HO

O

HO

O

HO

O

HO

O

OH

O

OH

O

OH

O

OH

O

6

5

N

NO O

OO

NH2

O

NHR

NHR'

Cl

HN

+

9

a) b)

c)

O

O

O

O

O

O

O

O

O

HN

10: R'' = Z2: R'' = H

e)

N N

O

OO

ONH2

O

NHR''

Cl

HN

O

HN

N N

O

OO

ONH2

O

NHR''

Cl

HNO

HN

N N

O

OO

ONH2

O

NHR''

Cl

HN

O

HN

N N

O

OO

ONH2

O

NHR''

Cl

HNO

NHNN

O

O O

OH2N

O

R''HN

Cl

NH

O

NHNN

O

O O

OH2N

O

R''HN

Cl

NHO

NH

NN

O

O O

OH2N

O

R''HN

Cl

NH

O

NH

NN

O

O O

OH2N

O

R''HN

Cl

NH

d)

f)

Scheme S1. a) see ref. [S1]; b) i-PrNH2 (quant); c) PdCl2(PPh3)4, Bu3SnH, p-nitrophenol

(quant); d) HATU, di-t-Bu-pyridine, TEA, DMF (66 %); e) TFA, CH2Cl2 (quant); f) TFA, CH2Cl2

(quant). Note, 10 and 2 contain both regioisomers (2,6- and 3,7-) of rNDIs.

S15

ONH

O

ONH

O

O

HN

OO

OO

O

S S

HN

HN

1

NN

O

O O

OH2N

O

OOC

HN

NH

NN

O

O O

OH2N

O

OOC

HN

NH

N N

O

OO

ONH2

O

COO

NH

HN

N N

O

OO

ONH2

O

COO

NH

HN

O

O

O

O

O

O

O

O

O

HN

N N

O

OO

ONH2

O

COO

NH

HN

O

HN

N N

O

OO

ONH2

O

COO

NH

HNO

HN

N N

O

OO

ONH2

O

COO

NH

HN

O

HN

N N

O

OO

ONH2

O

COO

NH

HNO

NHNN

O

O O

OH2N

O

OOC

HN

NH

O

NHNN

O

O O

OH2N

O

OOC

HN

NHO

NH

NN

O

O O

OH2N

O

OOC

HN

NH

O

NH

NN

O

O O

OH2N

O

OOC

HN

NH

3

O

O

O

O

O

O

O

O

O

HN

N N

O

OO

ONH2

O

NH3

NH

HN

O

HN

N N

O

OO

ONH2

O

H3N

NH

HNO

HN

N N

O

OO

ONH2

O

NH3

NH

HN

O

HN

N N

O

OO

ONH2

O

NH3

NH

HNO

NHNN

O

O O

OH2N

O

H3N

HN

NH

O

NHNN

O

O O

OH2N

O

NH3

HN

NHO

NH

NN

O

O O

OH2N

O

H3N

HN

NH

O

NH

NN

O

O O

OH2N

O

H3N

HN

NH

4

Figure S1. Structure of zipper components 1, 3 and 4 (see Scheme S1 for 2). For design details

for zipper assembly, see ref. [S3], Figures S2 and S3 (intrastack H-bonded chains and interstack ion

pairing).

S16

Cathodic current

Anodic current

-1.5-1-0.500.511.5

V (vs Fc/Fc+

)

Fig. S2. Representative cyclic voltammogram for N,Cl-NDI 7.

Figure S3. Intensity-normalized absorption spectra of the investigated dyes in MeOH.

S17

Figure S4. Intensity-normalized fluorescence spectra of the investigated dyes in MeOH.

1114

Figure S5. Fluorescence quenching of a constant concentration of 2a by bianisole 14 and p-

quateranisole 11 in acetonitrile.

S18

Figure S6. Intensity-normalized time profiles of the transient absorption of 2 at 470 nm and of

4 at 511 nm.

A

Rel

ativ

e ph

otoc

urre

nt

B

0

10

20

30

0 1 2 3 4

layers

0.8

1

1.2

1.4

1.6

520 560 600 640 680

Ab

s

nm

0

0.02

0.04

0.06

520 560 600 640 680

Ab

s

nm

C

Fig. S7. A. Build up of mixed multi-layers, evidenced by increase in relative photocurrent.

Blue filled circles, Au-1-4-3-4; red filled diamonds, Au-1-2-3-2; pink open circles, Au-1-2-3-4;

purple triangles, Au-1-4-3-2. Error bars represent standard deviation of five independent

experimental data. B and C. As measured (B) and baseline corrected (C) absorption spectra of

multi-layers on the gold plates. Blue solid line, Au-1-4-3-4; red dotted line, Au-1-2-3-2; purple

broken line, Au-1-2-3-4, grey solid line, gold plate.

S19

Comment on Fig. S7: The huge increase in photocurrent induced by the deposition of red layer

(Fig. S7A) indicated that rNDIs are much better chromophores than bNDIs. This conclusion was

supported by the action spectra (Fig. 3b), in which photocurrent generated by irradiation on the

absorption of bNDI was nearly negligible compared to that by rNDI. Note, the mismatch between

the action spectra (Fig. 3b) and UV-vis absorption spectra (Fig. S7C), which confirmed the presence

of bNDI chromophores in all the layers. Due to the scattering caused by the gold film (Fig. S7B

grey line), it was not possible to unambiguously quantify the two chromophores in the mixed layers.

Nevertheless, the presence of higher concentration of bNDI ( max = 620 nm) in Au-1-4-3-4 (blue

solid line), or higher concentration of rNDI ( max = 532 nm) in Au-1-2-3-2 (red dotted line)

compared to the others could be seen in both absorption spectra (Fig. S7B and C).

S20

BR

POPTEOA

A H I

b) e–

a) e–

c) energy

B e–Ee–

e–e–

e–

DC

F

a) b)

c)

h hh

d) e–

e) e–

d)

e)

e– G

f)g)

g)

Jf)

Fig. S8. Possible pathways of electron transfer cascades. HOMO (solid lines) and LUMO

levels (dashed lines) of rNDI (“R” in A, red), bNDI (“B” in A, blue), p-octiphenyls (“POP” in A,

grey) and TEOA (green). Orange rectangle represents the gold electrode.

Comment on Fig. S8: Upon photo-excitation of rNDI (A), exciton (B) is generated. Thus

formed hole would be filled by the electron transfer from (or hole transfer to) oligophenyl to give

the charge separated state with a radical anion on rNDI and a radical cation on the oligophenyl (C,

via path a). Alternatively if the electron transfer takes place from bNDI (path b), radical cation

would be on the bNDI in the resulting charge separated state D. A hole on the oligophenyl in C

S21

should be eventually transferred to TEOA in the solution (F) directly (path d) or via bNDI (D, path

e), while the electron on rNDI should be injected to the gold electrode (G). Photo-excitation of

bNDI (H) or energy transfer from excited rNDI B (path c) would generate bNDI exciton (E). The

electron transfer to rNDI (giving D) and subsequent steps discussed above should give rise to G.

Similarly, excited oligophenyl (J) could transfer energy to rNDI (path f) and/or bNDI (path g,

giving B and E, respectively), or transfer electron to rNDI (directly or via bNDI) to give charge

separated state C. All these intermediate states should follow the paths described above to finally

give G.

S22

4. Supplementary tables

Table S1 Electrochemical and spectroscopic data

Dye E1/2 (X/X

+)

V vs Fc/Fc+

EHOMO

EV

E1/2 (X/X-)

V vs

Fc/Fc+

ELUMO

eV

abs ( calc)

nm

4ca +0.62 -5.4 -1.33 -3.5 620 (636)

4bb +0.60 -5.4 -1.40 -3.4 620 (620)

2bb (+1.01)

c (-5.8)c -1.24 -3.6 532 (550)

d

7 +1.21 -6.0 -1.10 -3.7 536 (537)

11 +1.10e -5.9

12f +0.98 -5.8 -2.77 -2.1

13g +0.79 -5.6 -2.20 -2.6

afrom ref [S3];

bfrom ref [S4];

ccalculated from absorption (see d);

destimated from the average

wavelength of absorption and emission maxima; eirreversible, EPA is listed;

ffrom ref [S5] and E1/2

(Fc/Fc+) = 0.49 V (vs Ag/AgCl);

gestimated using eqs (1) and (2) in ref [S5] and E1/2 (Fc/Fc

+) =

0.49 V (vs Ag/AgCl).

N

NO O

OO

NH2

OR

NHAlloc

NH

HN

N

N OO

O O

C8H17

C8H17

C8H17

HN

H17C8NH

4b

N

N OO

O O

C8H17

C8H17

ClH17C8

NH

2b4a: R = (CH2)2-NH2

4c: R = COO-t-Bu

O

O

O

O

H Hn

12: n = 4

13: n = 8

n

11: n = 1

14: n = 0

S23

Table S2 Fluorescence quantum yields of the investigated dyes

Dye Solvent fl

4aa MeOH 0.25

4a MeOH 0.009

2a MeOH 0.06

2a CAN 0.43

2 MeOH 0.011

afrom ref [S1].

Table S3 Time constants and relative amplitudes (in brackets) used to reproduce the

fluorescence decay of the investigated systems.

Dye Solvent 1 (ps) 2 (ps) 3 (ps) 4 (ns) 5 (ns)

2a MeOH — — 115 (0.39) 1.6 (0.32) 6.0 (0.29)

2a CAN 1.7 (0.09) — 115 (0.13) — 9.4 (0.78)

2 MeOH 4.3 (0.24) 23 (0.44) 164 (0.26) 1.4 (0.03) 4.3 (0.02)

4aa

MeOH — — — — 8.4 (1.0)

4a

MeOH 7.1 (0.66) 51 (0.19) 160 (0.12) 1.8 (0.01) 7.3 (0.02)

afrom ref [S1].

S24

5. Supplementary references

[S1] S. Bhosale, A. L. Sisson, P. Talukdar, A. Fürstenberg, N. Banerji, E. Vauthey, G. Bollot, J.

Mareda, C. Röger, F. Würthner, N. Sakai, S. Matile, Science 2006, 313, 84-86.

[S2] B. Baumeister, N. Sakai, S. Matile, Org. Lett. 2001, 3, 4229-4232.

[S3] N. Sakai, A. L. Sisson, T. Bürgi, S. Matile, J. Am. Chem. Soc. 2007, 129, 15758-15759.

[S4] C. Thalacker, C. Röger, F. Würthner, J. Org. Chem. 2006, 71, 8098-8105.

[S5] K. Meerholz, J. Heinze, Electrochim. Acta 1996, 11/12, 1839.

[S6] Y. Yamamoto, T. Fukushima, Y. Suna, N. Ishii, A. Saeki, S. Seki, S. Tagawa, M.

Taniguchi, T. Kawai, T. Aida, Science 2006, 314, 1761-1764;

[S7] F. Würthner, S. Ahmed, C. Thalacker, T. Debaerdemaeker, Chem. Eur. J. 2002, 8, 4742-

4750.

[S8] J. R. Lakowicz, Principles of Fluorescence Spectroscopy, 2nd

ed. 1999, New York: Kluwer

Academic.

[S9] A. Morandeira, L. Engeli, E. Vauthey, J. Phys. Chem. A 2002, 106, 4833-4837.

[S10] M. Twardowski, R. Nuzzo, Langmuir 2002, 18, 5529-5538.

[S11] M. D. Porter, T. B. Bright, L. David, D. L. Allara, C. E. D. Chidsey, J. Am. Chem. Soc.

1987, 109, 3559-3568.