1

"This document is the Accepted Manuscript version of a Published Work that appeared in final form in ULTRAMICROSCOPY after peer review and technical editing by the publisher. To

access the final edited and published work see Ultramicroscopy, 97 (2003) 47-53.

Scanning Tunneling Microscopy of 1, 2, and 3 Layers of

Electroactive Compounds.

Dan Barlow, Louis Scudiero, and K.W. Hipps*

Department of Chemistry and Materials Science Program Washington State University Pullman, WA 99164-4630

hipps@wsu.edu

ABSTRACT

Bi-layer and tri-layer organic films grown on Au(111) were studied by scanning tunneling

microscopy (STM). Studies were carried out under UHV conditions with the sample cooled to

either 80 or 100 K. Cobalt(II)phthalocyanine [CoPc] was deposited from vapor onto Au(111),

followed by vanadyl phthalocyanine, VOPc. CoPc coverages studied were 0.5 and 1 monolayer,

while VOPc coverages were about 0.5 monolayer. Constant current images were acquired at

high tunneling gap resistance, of the order of 30 G. Two different physical structures were

observed for VOPc on CoPc, and each had a characteristic I(V) curve indicating significantly

different unoccupied state density. It is also demonstrated that the transmission factor for two

layers of VOPc is not simply the product of the transmission factors for each layer.

2

INTRODUCTION

The presence of metal, ring, and peripheral group centered electron transport processes makes

porphyrins and phthalocyanines exciting model systems for scanning tunneling microscopy

(STM) and orbital-mediated tunneling spectroscopy (OMTS) studies.1-20 In these previous

studies, monolayer, and sub-monolayer films have been investigated in UHV, under an inert

atmosphere, and in solution. These studies have produced clear molecular images and tunneling

spectra that show a clear connection between orbital mediated electron transfer processes and the

expected HOMO and LUMO energies. In contrast, study of the growth of films of these

compounds in the vertical direction has not received much attention, primarily because of the

difficulty in obtaining stable images from systems showing partial bilayer and trilayer structure.

Thus, most STM studies of multilayer organic films that have been reported so far have not

focused on high resolution imaging.21 While stable images of phthalocyanines on alkane layers

have been reported, the alkane layer does not constitute an electroactive layer.22

This report provides a first look at the structure of, and electron transport through,

multilayers of metal phthalocyanines. In particular, one and two layers of vanadyl

phthalocyanine, VOPc, on a near monolayer of cobalt phthalocyanine, CoPc, on Au(111) is

examined. A bilayer, therefore, is a two layer system composed of a CoPc island on Au(111)

followed by a single layer of VOPc. A trilayer is two layer of VOPc covering a single layer of

CoPc on Au(111). This system was chosen because the component parts have been studied by

our group,6,7 and because the non-planar structure of VOPc might pack in an unusual way onto

the planar CoPc monolayer. Also, VOPc and CoPc appear to form stable bilayer islands at low

CoPc coverage, rather than the expected mixed monolayers as formed by other mixed

phthalocyanines5,6. The primary result of this work is a straightforward demonstration that the

3

transmission coefficient through two layers of VOPc is not simply the square of the transmission

coefficient for one layer. Surprisingly, to the accuracy of this measurement, there is no

discernable difference between the conductivity of one layer of VOPc on CoPc and that of 2

layers.

EXPERIMENTAL

Scanning tunneling microscopy, STM, analysis was done in UHV with a variable

temperature STM (model UHV300) and control electronics (model SPM100) from RHK

Technology.23 Metal and organic thin films were prepared with a vapor deposition system that

was custom designed in collaboration with RHK Technology. This system was designed to work

with RHK sample holders and was in a chamber that was attached to the STM chamber so that

samples could be prepared and analyzed without exposure to air. A typical RGA of the STM

chamber showed the principle residual gas was H2 at mass 2 (~2x10-10 torr), with very low levels

of water (18), CO (28), and CO2 (44), all below 2x10-11 torr. These pressures were reached with

a 410 L/s Starcell ion pump from Varian Vacuum Technologies.24 The instrument was housed in

a vibration isolation laboratory at Washington State University.

A dedicated, cryopumped (Cryotorr 8, 1500 L/s, CTI-Cryogenics) chamber was used for

preparation of Au(111)/mica samples by vapor deposition. This chamber reaches a base pressure

of 8x10-10 torr without baking. Freshly cleaved mica substrates (purchased from Ted

Pella25;Pelco #54). were heated at 500 C for 24 hours to remove water from the surface and then

the temperature was reduced to 380 C for gold deposition. Gold splatters (99.999%, Cerac,

Inc.26) were used for making the Au(111) films. The gold was evaporated from resistively

heated tungsten boats (ME5-.005W, R.D. Mathis Co.27) that consisted of a .025”x.005” W strip

with a dimple in the center. With these boats, the pressure could be maintained in the low 10-9

4

torr while evaporating gold at rates of ~0.5 Å/s over a period of ~1hr. The source was heated

and the deposition rate was monitored until a constant value between 0.2 and 0.5 Å/s was

obtained. The mica was then moved over the gold source and the thickness of the film was

calculated from the amount of time the mica was exposed to the gold vapor. About 120 nm of

gold was usually deposited. The resulting Au film was allowed to cool for at least 6 hours before

the chamber was vented.

VOPc was purchased from Alfa Aesar28, and CoPc was purchased from Strem

Chemicals29. The phthalocyanines were further purified under vacuum in a quartz sublimater.

VOPc was presublimed and CoPc was doubly sublimed. STM tips were prepared from 0.25 mm

W or Pt0.8Ir0.2 wire (Alfa-Aesar28 and California Fine Wire Co.30). W tips were prepared by

electrochemical etching while Pt-Ir tips were prepared by mechanical cutting. The procedures

used were fast and simple and produced high quality tips. At least 90% of the tips were capable

of imaging Au(111) atomic steps and reconstruction corrugation, while roughly 1 in 3 were sharp

enough to obtain atomic resolution on gold. Generally, several tips were made at a time and then

loaded into the STM chamber where it could be determined which tips were best by imaging

Au(111). W tips always required additional cleaning by electron bombardment once in UHV.

Bias voltages are given relative to the sample; thus, +1.0V bias means that the sample is at +1.0

V relative to the tip.

Mixed composition samples were prepared on Au(111) in two ways: 1) deposition of

~0.5 monolayer CoPc followed by ~0.5 monolayer VOPc; 2) deposition of ~1 monolayer CoPc

followed by ~0.5 monolayer VOPc. The CoPc sample deposited at near monolayer coverage

was imaged by STM before VOPc deposition. Sample temperature was measured using a type K

thermocouple in direct contact with the sample. Data reported in this paper were obtained either

5

at 80 K or 100K. The images have been low pass filtered and some were flattened. No Fourier

filtering was applied.

RESULTS AND DISCUSSION

Deposition of 0.5 monolayer CoPc followed 0.5 monolayer VOPc appears to produce

well-ordered hetero-bilayer islands which are surrounded by scattered CoPc and VOPc

molecules [sample 0.5/0.5]. Well-ordered monolayer structures and third layer molecules were

not observed, nor were well-ordered, mixed CoPc and VOPc in the same layer. Deposition of

the full CoPc layer followed by 0.5 monolayer of VOPc produced a complete well-ordered

monolayer of CoPc on top of which were islands of second and third layer VOPc [sample 1/0.5].

Figure 1 shows constant current STM images of a dense bi-layer region at +1.3 and -1.6 bias

volts taken for the 0.5/0.5 sample. First, note that there are three types of areas in the figure.

The first, labeled A, is a dense bilayer structure that appears the same in either bias. Also seen in

Figure 1 is a region marked B that is also a bilayer, but that changes apparent height with bias

voltage, and a region marked C where scattered CoPc and VOPc on bare Au are observed.

Outside the island-like structures are scattered, single VOPc and CoPc molecules on clean gold.

Figure 2 is a higher resolution constant current image of regions A and C. While there is

blurring of some of the molecules due to thermal motion in the low density region, individual

CoPc (marked A in the figure) and VOPc (marked B in the figure) molecules can be observed.

Also, it is clear that the upper layer in region A is all VOPc molecules and ordered in a manner

similar to that observed for pure VOPc on Au(111). It is easy to distinguish these molecules

because of the dramatic difference in tunneling probability through the vanadyl ion relative to the

cobalt(II) ion. Cobalt(II) complexes of porphyrins and phthalocyanines appear anomalously tall

in constant current images (leading to molecular images having filled centers).5,9 The vanadyl

6

ion, on the other hand, acts like an insulator and the constant current images of VOPc is

dominated by the four outer benzene rings. 7

Figure 3 displays the constant current contour at +1.3 volt bias along the line indicated in

Figure 1. As may be seen from figure 3, the apparent height of the bilayer in region B is about

0.15 nm lower at +1.3 volts than the bilayer in region A. This apparent change in height of

region B in the constant current images can be better understood by considering the I(V) curves

obtained over the two regions, as shown in Figure 4. Note that for a sample bias voltage less

than about +0.8 V, the two curves are almost identical. Above about 1 volt, however, they are

dramatically different, with much more current being transmitted through region A than through

region B. Thus, if the bias voltage exceeds or equals about 1 volt, to maintain constant current

while scanning from region A to region B the tip must move closer to the metal surface as it

crosses region B.

Why two regions of essentially the same chemical composition should differ so

significantly in their conductivity, is partially explained by Figure 5. In Figure 5 we show high

resolution image of the structure of the type seen in region A of figure 1 (the left hand image),

and an equal resolution image obtained from region B of Figure 1. The physical structure of the

VOPc layer in region B of Figure 1 is different than that of region A. In region B, we find

double rows of alternating rotational orientation, while in region A the VOPc molecules form

well ordered and continuous diagonal rows. The origin of this difference in structure is not

known at this time, but it may be due to either the orientation of the oxygen relative to the

surface plane (up or down), or how it is oriented in the plane of CoPc molecules. Experiments

are currently underway to clarify this issue.

7

A sample having a full CoPc monolayer and a much less homogeneous second layer of

VOPc is shown in Figure 6. Third layer VOPc growth was observed long before a complete

VOPc second layer had formed. Figure 6 is an example of such a region. The constant current

image in Figure 6 was acquired at 80 K with a sample bias of +2.0 V and gap resistance of

30G. For later reference, we here note that the LUMO in VOPc should lie well below +2 volts

sample bias.7 The selected image contrast only allows the second and third layer VOPc

molecules to be observed in the image, but there are also ordered, first layer CoPc molecules in

the dark areas. The sectional plot was made along the diagonal line in the right hand figure.

The surprising result is that the apparent height of the third layer (the second VOPc layer on a

CoPc monolayer) is equal or exceeds that of the second layer. (Note that the principal structure

in the sectional plot is not noise, but is rather the variation in conductivity resulting from

different parts of the VOPc molecule.) Conventionally, one assumes that the transmission of an

electron through a single molecule is proportional to exp(-D), where is of the order of 1 and

D is the appropriate molecular dimension. D is assumed to be the same for CoPc and for VOPc.

Taking as the exponential factor for CoPc, that of VOPc, and that for the vacuum gap

between the tip and the top molecular layer, the transmission factors for one layer of CoPc and

for one layer of VOPc are given by TCo= exp(-D) and TVO = exp(-D). Let Z2 be the distance

between the gold substrate and W tip when the tip is over the bilayer, and Z3 be the distance

between Au and tip when the tip is over the tri-layer region. If we assume that conduction

through each layer is independent of the adjacent layers and ignore the contact resistance

between molecules, we have for the overall transmission factors for the two (T2) and three (T3)

layer systems:

T2= exp(-D) exp(-D) exp(-(Z2-2D)), and

8

T3= exp(-D) exp(-2D) exp(-(Z3-3D)),

The apparent height contour in Figure 6 is one where the current remains constant. Thus, along

that line T2 = T3 or:

exp(-(Z2-2D)) = exp(-D) exp(-(Z3-3D))

However, Z3Z2+D, based on Figure 6. Thus

exp(-(Z2-2D)) = exp(-D) exp(-(Z2-2D))

or, 1 = exp(-D)

Thus, if the transmission coefficient for the composite system was simply the product of the

components, we would be forced to conclude that VOPc was perfectly transparent (=0)!

Adding reductions to the transmission due to interface effects would only make the situation

more extreme (<0). An alternative approach would be to assume incoherent conduction

through each molecule with an associated molecular resistance, R, that is additive. In this

extreme,

R2 = 2R and R3 = 3R , thus

1/( R2 + exp((Z2-2D))) = 1/( R3+exp((Z3-3D)))

(R2 + exp((Z2-2D))) = R3+exp((Z2-2D))

Or, R2 = R3

Which also leads to the unlikely conclusion that the second VOPc layer is a perfect conductor

when Vbias > ELUMO. That is, the molecular resistivity becomes so small that it is completely

negligible compared to the resistance of the vacuum gap. This situation would be similar to the

case of a metal, where the tunneling from the surface layer is the same independent of the

thickness of the sample. However, VOPc has no density of states in the region below about 1.4

9

volts bias 7 and CoPc has no states between 0 and +1.4 volts bias.31 Thus, no more than about

0.5 volts of the 2.0 volt bias potential could be dropped across layers one and two.

In our opinion, all these methods have flaws. What is not accounted for in the previous

models is the possibility that electrical properties of the three layer system are not simply a

superposition of the individual layers. Rather, the three layer system as a whole has an equal or

greater conductivity than the two level system. This can happen if the electronic states of the

three level system shift or broaden in such a way as to increase the density of states near the

Fermi energy. This kind of electronic interaction resulting from what are usually assumed to be

weak packing forces is also suggested by the significant differences in LUMO’s for the two

methods of VOPc adlayer packing (Figure 5).

CONCLUSIONS

Images of two and three layers of metal phthalocyanine on Au(111) are reported at 80

and 100 K and ~30 G junction resistance. It is found that the transmission factor for three

layers is much larger than the product of the individual layer transmission factors unless one

assumes near perfect conductivity. It is also observed that the effective electronic resistance of

three layers is less than predicted by assuming a series resistance model. It is postulated that the

packed structures seen here produce modifications in (shifts and changes in width) the electronic

states near the Fermi energy and thereby significantly increase the conductivity of the three layer

system relative to that of two layers.

ACKNOWLEDGEMENT

We gratefully acknowledge support provided by the National Science foundation in the

form of Grant number CHE-0138409.

10

BIBLIOGRAPHY

1 B. Duong, R. Arechabaleta, N.J. Tao, J. Electroanal. Chem. 447 (1998) 63-69.

2 N.J. Tao, G. Cardenas, F. Cunha, Z. Shi, Langmuir 11 (1995) 4445-4448.

3 W. Han, E. N. Durantini, T. A. Moore, A. L. Moore, D. Gust, P. Rez, G. Letherman, G. Seely,

N. J. Tao, S. M. Lindsay, J. Phys. Chem. B 101 (1997) 10719-10725.

4 B. Duong, R. Arechabaleta, N.J. Tao, J. Electroanal. Chem. 447 (1998) 63-69.

5 X. Lu, K.W. Hipps, X.D. Wang, Ursula Mazur. J. Amer. Chem. Soc. 118 (1996) 7197-7202.

6 Xing Lu, K.W. Hipps. J Phys. Chem. B 101 (1997) 5391-5396.

7 D. Barlow, KW Hipps. J. Phys. Chem. B 104 (2000) 5993-6000.

8 L. Scudiero, Dan E. Barlow, K.W. Hipps. J. Phys. Chem B 104 (2000) 11899-11905.

9 L. Scudiero, D. E. Barlow, U. Mazur, K.W. Hipps. J. Am. Chem. Soc. 123 (2001) 4073-4080.

10 K. Walzer, M. Hietschold, Surface Science 471 (2001) 1-10.

11 R. Strohmaier, C. Ludwig, J. Petersen, B. Gompf, W. Eisenmenger, J. Vac. Sci. Technol. B 14

(1996) 1079-1082.

12 Francesca Moresco, Gerhard Meyer, Karl-Heinz Rieder, Hao Tang, André Gourdon, Christian

Joachim. Phys. Rev. Lett. 86 (2001) 672-675.

13 T. A. Jung, R. Schlittler, J. K. Gimzewski, H. Tang, C. Joachim, Science 271 (1996) 181-184.

14 J. K. Gimzewski, T. A. Jung, M. T. Cuberes, R. Schlittler, Surf. Sci. 386 (1997) 101-114.

15 M. Kuniatke, N. Batina, Kingo Itaya. Langmuir 11 (1995) 2337-2340.

16 L. Scudiero, Dan E. Barlow, K.W. Hipps. J. Phys. Chem. B 106 (2002)996-1003.

17 Masashi Kunitake, Uichi Akiba, Nikola Batina, Kingo Itaya. Langmuir 13 (1997) 1607-1615.

18 Masashi Furukawa, Hiroyuki Tanaka, Ken-ichi Sugiura, Yoshiteru Sakata, Tomoji Kawai.

Surface Science 445 ( 2000) L58–L63.

11

19 Xiaohui Qiu, Chen Wang, Qingdao Zeng, Bo Xu, Shuxia Yin, Hongna Wang, Shandong Xu,

Chunli Bai. J. Am. Chem. Soc. 122 (2000) 5550-5556.

20 K. W. Hipps, Louis Scudiero, Dan E. Barlow, Manning P. Cooke, J. Am. Chem. Soc. 124

(2002) 2126-27.

21 S. F. Alvarado, L. Rossi, P. Muller, P. F. Seidler, W. Riess, IBM J. Res. and Dev. 45 (2001)

89-100.

22 S.B. Lei, C. Wang, S. Yin, C.L. Bai, J. Phys. Chem. B 105 (2001) 12272-12277.

23 RHK Technology, Inc., 1050 E Maple Rd., Troy, MI 48083, Tel: 248-577-5426.

24 Varian Vacuum Technologies, 121 Hartwell Ave., Lexington MA, 02421, Tel: 800-882-7426. 25 Ted Pella, Inc., P.O. Box 492477, Redding, CA 96049-2477, Tel: 800-237-3526.

26 Cerac, Inc., P.O. Box 1178, Milwaukee, WI 53201-1178, Tel: 414-289-9800.

27 R.D. Mathis Co., P.O Box 92916, Long Beach, CA 90809-2916, Tel: 562-426-7049

28 Alfa Aesar, 30 Bond St., Ward Hill, MA 01835, Tel: 800-343-0660.

29 Strem Chemicals, Inc., 7 Mulliken Way, Newburyport, MA 01950-4098, Tel: (800) 647-8736.

30 California Fine Wire Co., P.O. Box 446, Grover Beach, CA 93483-0446, Tel: 805-489-5144

31 K.W. Hipps, D. Barlow, in preparation.

12

FIGURE CAPTIONS

Figure 1. Constant current STM image obtained from a Au(111)/CoPc/VoPc multilayer

sample. The CoPc was first deposited to about 0.5 monolayers followed by VOPc at the same

coverage. Both images were obtained at 100 K in UHV and 20 pA set-point current.

Figure 2: Higher resolution constant current scan of the area shown in figure 1 (regions

A and C). Crosses have been placed on a few of the individual VOPc molecules to help guide

the eye. The molecule labeled A is a CoPc, while that labeled B is VOPc.

Figure 3: Constant current contour along line indicated in left hand image of figure 1.

Data collected at 100K.

Figure 4: I-V curves obtained over regions A and B (as marked in Figure 1) and as

indicated on the graph. Data collected at 100K.

Figure 5: Constant current images acquired at +0.6 volts sample bias and 40 picoamps of

current. Image A was taken from a region similar to that seen in Figure 1 area A, while image B

was taken from the composite bilayer region B of Figure 1. Data collected at 100K.

Figure 6: Constant current image and height contour for single and double add-layer

regions of VOPc on CoPc formed when a monolayer of CoPc is covered by about 0.5 layers of

VOPc. In the dark areas are ordered, first layer CoPc molecules which are not visible with this

image contrast. Data collected at 80 K.

13

Figure 1. Barlow, Scudiero, and Hipps

14

Figure 2: Barlow, Hipps, and Scudiero

15

Figure 3: Barlow, Scudiero, and Hipps

16

Figure 4: Barlow, Scudiero, and Hipps

17

Figure 5. Barlow, Scudiero, and Hipps

18

Figure 6: Barlow, Scudiero, and Hipps