Paper

5/20/2018 Paper

1/14

Ultramicroscopy 50 (1993) 157-170

North-Holland

ultramicroscoo

.

Effect of probe force on the resolution

of atomic force microscopy of DNA

Jie Yang and Zhifeng Shao *

Bio-SPM Laboratory and Department of Physiology Uni versit y of Krgini a Box 449 Charlott esvil le VA 22908 USA

Received 15 December 1992

Experimental results are presented to show that the adhesion force is the single most important limiting factor in

high-resolution atomic force microscopy of DNA in air, prepared by the cytochrome-C-assisted spreading method. It is also

shown that humidity plays a minor role in the control of probe force. Using a pure carbon film as the substrate to clean the

APM tip prior to imaging, it is demonstrated that 4-6 nm resolution on DNA can be routinely obtained by the atomic force

microscope with commercial Si,N, pyramid cantilevers. We also show that in organic solvents a resolution of up to 3 nm can

be obtained under optimal conditions

1 Introduction

The atomic force microscope (AFM) has been

applied to various biological materials in air and

in fluid. Despite the soft nature of these speci-

mens and the relatively large probe force re-

quired with the instruments available today, some

remarkable results have been published, such as

AFM imaging of native membrane fragments [ 1,2],

synthetic lipid bilayers [3-81, purified membrane

proteins [61, and both double-stranded and

single-stranded DNA [9-171. For imaging DNA,

several specimen preparatory methods have been

developed [9-153 that yielded highly reproducible

results with the best resolution comparable or

exceeding those achieved by electron microscopy

of stained or heavy-metal-shadowed specimens

[18,191. Our preparatory method used the con-

ventional cytochrome-C-assisted spreading of

DNA on carbon-coated mica [20]. With this

method without further fixation or staining or

shadowing, the DNA molecules can be directly

visualized by AFM either in air or in some or-

ganic solvents, with a consistent lateral resolution

of 4-6 nm [9] (fig. 1). A molecular width of 3 nm

* To whom correspondence should be addressed.

was observable in fluids, such as isopropanol and

n-butanol. The specimens prepared by this

method are extremely stable that, when stored in

air for over 6 months, the same resolution can

still be achieved without apparent sign of degra-

dation. Repeated scanning with controlled probe

force usually does not damage the specimen.

Other methods have not used the protein-assisted

DNA spreading [lo-151. To facilitate DNA adhe-

sion to the substrate, mica was treated by various

chemicals, such as 3-aminopropyltriethoxy silane

[15,17] and magnesium acetate [11,12,16]. With

these methods, the specimen can also be imaged

in air or in fluid. Using a micro-fabricated su-

pertip 3 nm resolution was achievable in some

organic solvents [12,161. With the silane-treated

mica, DNA was shown to adhere rather strongly

to the substrate [15,171. Furthermore, the silane-

treated method also achieved DNA imaging in

water [17]. These recent developments firmly es-

tablished the validity of AFM results on imaging

DNA molecules, which is in sharp contrast to the

earlier reports of STM images of DNAs on

HOPG [21,221. These results indicated that the

AFM might be a useful tool in the study of

DNA-protein interactions, DNA packing, as well

as gene sequencing.

0304-3991/93/%06.00 0 1993 - Elsevier Science Publishers B.V. All rights reserved

5/20/2018 Paper

2/14

15X

J. Yrrng, Z. Shao / Effkt of probe orce on the resolution f AFM of DNA

Fig. I. Typical AFM image of a 10 kbp plasmid DNA ob-

tained in air, using a commercially available Si,N, pyramid

tip with a spring constant of k = 0 06 N/m and a scanning

speed of 4.73 Hz. The probe force was about 2 nN. Along the

DNA molecules. the minimum measured width in many places

was between 3 and 4 nm, when scanned at higher magnifica-

tion (200-400 nm). The image was obtained under a humidity

of 35%.

So far, the limitation on the resolution achieved

on DNA molecules has been largely attributed to

the tip geometry [lo-121. The commercially avail-

able Si,N, pyramid tip cantilevers have an apical

angle of N 55 and a nominal tip radius of curva-

ture of

- 30 nm [23,24]. By assuming a rigid

cylindrical chain for the DNA molecules, the tip

broadening effect was indicated [lO,ll]. It was

reported that,

supertips having an apical angle

of less than lo and a reduced radius of curvature

at the tip yielded improved lateral resolution

[

11,121 at otherwise similar conditions.

In this article, we present a comparative study

on the effect of the probe force and the environ-

ment on the lateral resolution of DNA molecules

by AFM using commercially available cantilevers.

We found that the lateral resolution is closely

related to the so-called adhesion force, and to a

much lesser degree to the probe geometry. This

result seems to indicate that some sort of contam-

ination on the tip, which should be directly re-

lated to the adhesion force, may be responsible

for the frequently observed degradation of reso-

lution. We will present a procedure for cleaning

AFM tips to reduce the adhesion force to as low

as 1 nN. This method can yield a resolution of

4-6 nm for most specimens in air, comparable to

that obtained by supertips in organic solvents

112,161. A practical approach using the adhesion

force measured on a distance-versus-force curve

as a diagnostic procedure was found quite suc-

cessful in the assessment of the quality of the tip,

as well as the specimen. All these results are

directly from the experimental findings, and a

complete theoretical analysis has not been devel-

oped.

2. Materials and methods

2 1 Specimen preparation

Our specimen preparation method is a simpli-

fied version of the Kleinschmidt method for elec-

tron microscopy [20]. Double-stranded plasmid

DNA (10 kbp), consisting of a cDNA encoding

the (Nat+ K+)-ATPase a-subunit and a mam-

malian expression factor, was used [9]. DNA solu-

tion (_ 2 pg/ml), containing 100 Fg/ml cy-

tochrome C, 1OmM Tris, 1mM EDTA, pH 8,

distilled and deionized water and 50% for-

mamide, was well mixed prior to spreading. A

glass rod of 3.3 mm diameter, cleaned by soaking

in chromic acid and rinsed thoroughly with dis-

tilled and deionized water, was positioned - 45

with respect to the horizontal level with one end

immersed in a teflon dent (made by l/2 drill)

filled with distilled and deionized water (- 0.5

ml>. A droplet of DNA solution (11 ~1) was

pipetted slowly onto the glass rod 2-3 cm above

the subphase water surface. The droplet should

flow down the rod freely, so that a protein film

was formed at the air-water interface. At certain

viewing angles, this protein-DNA film should be

observable to the naked eye. The glass rod was

slowly removed to minimize disturbance to the

protein film. A minute or so later a piece of

carbon-coated mica was used to pick up the

monolayer by touching the water surface. This

operation is performed in a humidity-controlled

environment with the humidity below 35%. The

5/20/2018 Paper

3/14

J. Yang 2. Shao / Effect of probe force on the resolur ion of AFM of DNA

159

specimen was left to dry in air. For carbon coat-

ing, freshly cleaved mica was placed in a vacuum

evaporator (Denton Vacuum, DV502A). The

evaporation was carried out at low current (35-45

A) and long period (30-40 min) to ensure fine

carbon granules. The carbon-coated mica was

then placed in an oven baked overnight at 200C.

This procedure led the carbon film strongly bound

to the mica surface such that it cannot float off

the water surface when used to pick up the pro-

tein-DNA film. The baking is not mandatory, but

useful for more reproducible preparations.

Cleanness is critical in this method for high-qual-

ity specimens.

2.2. A FM imaging

The AFM images were obtained by a

NanoScope II AFM (Digital Instruments Inc.,

Santa Barbara, CA) with commercially available

cantilevers (k = 0.06 N/m, also from Digital In-

struments Inc.). For AFM in air, we found that

the probe force remained stable over a period

over 10 h for humidities below 40%. For imaging

in an organic solvent, such as isopropanol or

n-butanol (Sigma Chemicals, St Louis, MO), a

home-made fluid cell retrofitted to the Nano-

Scope II was used [6]. Different organic solvents

did not affect the quality of AFM imaging appre-

ciably, only that the evaporation rate is lower

with n-butanol. It is highly advisable to use a very

small scan area for initial approach (usually 5-10

nm). After the tip is engaged, the distance-

versus-force curve should be examined immedi-

ately to minimize the probe force applied. This

procedure will prevent specimen/tip damage by

the very large, initial probe force. Only if the

adhesion force is small, a larger area scan can be

started to search for DNA molecules. The force

curve should be constantly checked to correct for

drifting that can cause either accidental disen-

gagement or gradual increase of the probe force.

3. Experimental results

It was reported that, in the Kleinschmidt

method, the denatured cytochrome C molecules

Fig. 2. AFM image showing a defect on the cytochrome C

film, with the depth of - 1.3 nm. The defect was made by

scanning an area of 100x 100 nm2 at 78.13 Hz with a probe

force about 10 nN. The image was obtained after reducing the

probe force to - 3 nN using a scanning speed of 4.73 Hz, at

45 rotation to avoid any possible drag by the tip from the

build-up at the edge of the defect to influence the area with

the DNA. All above operation was carried out at a humidity

of 37%.

formed a thin layer at the substrate surface with

DNA adsorbed on top of it [251. Presumably the

DNA also binds to the denatured cytochrome C

molecules with details completely unknown [25].

This assertion of a denatured protein film on the

substrate is confirmed by AFM as shown in fig. 2,

where not only a DNA molecule is shown, but

also a 100 nm square surface defect, which was

made by using a large probe force that pushed

the denatured proteins to the edge of the scanned

area. This protein film, measured from fig. 2, is

about 1.3 nm thick in the dehydrated form, al-

though the degree of compression is not clear.

From the nature of this preparation method, we

can expect that the image quality, both the con-

trast and the resolution, should be quite sensitive

not only to the geometry of the tip, but also to a

range of other parameters, such as the probe

force, humidity etc. Due to the soft nature of

both the DNA and the protein film, it is clear

that only under a probe force that does not

exceed a certain limit, imaging can be successful.

5/20/2018 Paper

4/14

IhO

.I. Yang Z. Shao / Effect uf prdw e on the solution of AFM of DNA

.42

Fig. 3. Two force curves with corresponding AFM images are presented to show the effect of electrostatic interaction. In (a), the

electrostatic interaction became effective when the tip was about 1.8 pm away from the specimen surface. Since the magnitude of

this force was so large that it was out of the range of the instrument, only part of it is shown to demonstrate the long-range

characteristic of the electrostatic interaction. Two arrows indicate the tip approaching and leaving the specimen surface, which is to

the left of the display. The curvature of the curves is due to the electrostatic attraction. The set point marks the setting of the

cantilever during AFM imaging. The magnitude of this force ( - 17 nN) was estimated from the set point voltage change when the

tip was engaged. The same specimen, when grounded, showed absence of any long-range interaction. In (b), the force curve after

grounding is shown (notice the different scale). The force curve shows a straight line when the tip approaches the specimen until it

touches the surface (indicated by the arrow). When the tip withdraws, it disengages from the specimen at a position where the

restoring force of the cantilever overcomes the adhesion force. Such a hysteresis was well observed with specimens with a finite

adhesion force. In (b), the adhesion force is only

- 1.2 nN. (c) AFM image with the force curve shown in (a); (d) AFM image with

the force curve in (b). Both (c) and (d) were obtained with the same tip on the same specimen. The humidity was 30%-35%.

Indeed, the control of the probe force is found

very important for AFM of DNA. But our study

also indicated that other factors arc equally im-

portant, if not more so. In the following, we

present experimental results which indicate what

factors must be carefully controlled in order to

obtain AFM images of DNA.

3. I. Electrostatic forces

Static charge often accumulates on the surface

of an insulator, with the amount sensitively de-

pendent on location, environment and surface

geometry. Therefore, it is virtually impossible to

even estimate its exact effect in a given experi-

ment. Mica, often used as the substrate in AFM

experiments, is known to carry some negative

surface charges after cleavage 1261. If a grounded

conductor is placed in front of such a charged

Fig. 4. Sketch to illustrate the grounding of the specimen. In

the NanoScope II AFM, the steel plate is grounded, so is the

metallic cantilever holder. The carbon paste will provide

grounding for the carbon film on the mica surface.

5/20/2018 Paper

5/14

surfac

ze, an attractive force will arise [27]. As we

founi %,with the carbon-coated mica, a long-range

force often occurs, as shown in fig. 3a. The de-

cline of the base line could be attributed to this

elect] :ostatic interaction [28], because the electro-

static

charge on the surface can induce an oppo-

J. Yang 2. Shao / Effect of probe force on the resohrt ion of AFM of

DNA 6

site charge in the gold coating on the cantilever

(grounded in the NanoScope II AFM) which

would result in an attractive force. This attractive

force can aIso be confirmed from the fact that, as

the cantilever was approaching the substrate sur-

face, the differential signal (A-B voltage output

19.94 nm div

Fig. 5, AFM images of two DNA molecules at two different adhesion forces. Images (a) and (b) were obtained with a probe force of

- 1.8 nN and the adhesion force of

- 1.2 nN shown in Cc); images Cd) and (e) show the same molecules under a probe force and

the adhesion force of

- 17 nN. The change of lateral resolution by a factor of 3-5 on the same molecule is clearly shown. Scanning

speed for all 4 images was 4.73 Hz. Humidity was at 32%.

5/20/2018 Paper

6/14

162

J .

Yang Z. Shao / Effect of probe force on the resolut ion of AFM of DNA

from the NanoScope II AFM) decreased to a very

large negative value before tip engagement. Fig.

3a shows that this force becomes effective when

the tip is 1.8 pm away from the surface. The

magnitude of this electrostatic force varies for

different preparations, but is almost always pre-

sent. With the electrostatic force shown in fig. 3a,

no stable images of DNA were obtained (fig. 3c),

possibly because the initial engagement has so

large a probe force that the tip punched through

the protein film right after engagement; it is very

likely to have some pieces from the specimen

attached to the tip or to have some damage on

the tip. Furthermore, with the force curve shown

in fig. 3a, the smallest probe force one could use

is no less than 15-20 nN (estimated from the

change of set point voltage during tip engage-

ment), otherwise the tip would disengage from

the specimen quite easily, due to the non-uni-

formity of the specimen. However, since the car-

bon film on the mica surface is conductive, any

surface charge can be totally shielded if the car-

bon film is electrically short-circuited to the

ground, as shown in fig. 4. By doing so, charges

on the carbon film are also removed, and charges

on the protein film will be balanced by the in-

duced charge in the carbon film (thus its effect is

very much reduced). Fig. 3b is the force curve

after grounding with fig. 3d as an example of the

effect of grounding for AFM imaging of DNA.

The long-range interaction is totally eliminated.

This procedure, as we found, is crucial for obtain-

ing any images of DNA with this kind of speci-

men. We should stress that the images in fig. 3

were obtained from the same specimen and with

the same cantilever under identical conditions,

with the only difference being the grounding. The

effect of grounding on image quality is quite

apparent when operated in air.

3.2. Adhesion force and i ts effect on resolut i on

For many substrates, such as mica, gold and

HOPG, there is always a quite large adhesion

force between the probe and the specimen sur-

face [29]. This force is often many orders of

magnitude larger than the van der Waals force,

which strongly pulls the probe towards the speci-

men surface. The origin of this force has been

attributed to the existence of a water layer at the

surface 1291. The meniscus formed between the

probe and the water surface would produce such

a force 130-321. However, for a protein film, such

as the specimens used here, it is not clear whether

the adhesion force is solely due to such a menis-

cus attraction.

We have found, through experiments, that the

success of AFM imaging of DNA is directly re-

lated to the magnitude of the adhesion force. Fig.

5 shows a pair of DNA molecules on the same

specimen and with the same tip under two differ-

ent adhesion forces. The force curves are also

shown below the DNA images for comparison.

The difference in the measured molecular width

is quite apparent: 4-5 nm averaged along the

DNA molecule for the small adhesion force and

about 15 nm for the larger one. To avoid confu-

sion, we define the lateral resolution as the half-

height full width along the direction perpendicu-

lar to the molecule throughout this article. When

the adhesion force was small, such as the one

shown in fig. 5c with a peak value of 1.2 nN, the

lateral resolution routinely achieved was 4-6 nm

when operated in air. With a large number of

specimens, we found there was a general trend

that a larger adhesion force was always related to

a lower resolution. A factor of 3-5 on the same

DNA molecules in measured lateral resolution

(as shown in fig. 5) was quite common. In order

to quantitate this effect, it would be ideal to use

the same tip on the same molecule while imaging

at various adhesion forces. Unfortunately, the

adhesion force cannot be controlled at will.

Sometimes, a reengagement of the cantilever

would cause an increase in the adhesion force,

with the increment mostly in the range between 7

and 30 nN. In addition, any given tip has a finite

usable life span. It is virtually impossible to use

the same tip for too long, because accidental

damage does occur during engagement or even

during imaging (particularly on poor specimens).

In fig. 6, we plotted the data of adhesion force

versus lateral resolution, obtained from a large

number of specimens and commercially available

cantilevers. Among the data shown, some were

from the same tip over several specimens at vari-

5/20/2018 Paper

7/14

.I. Yang Z. Shao / Effect of probe force on the resolut ion of AFM of DN A

163

1

I 10

Adhesion Force (nN)

Fig. 6. Log-log plot of the adhesion force versus lateral

resolution. Each data point in this plot was obtained by

averaging along the DNA molecule of one specimen under

one adhesion force. The error bar is the standard deviation

which indicates the non-uniformity of cytochrome C binding.

For those data points with adhesion forces below 6 nN, the

lateral resolution was measured from small scan areas (200x

200 nm to

400~400

nm2). For those data points of larger

adhesion forces, the resolution was measured directly from

images of typically 1 1 Fm2. A least-squares fitting indicated

that the lateral resolution is proportional to (adhesion

force),48. See section 4 for a discussion of this result.

~_.,., ,.. .~. ., _~ ~ _,_.. ~ .,_.., ..,. , _.,

19.94 nm/div

ous adhesion forces. These data were accumu-

lated over a period of several months with some

specimens stored in a desiccator for up to 10

months. Each data point in fig. 6 was obtained

under one adhesion force averaged along the

DNA molecule. To avoid complications from

other factors, the humidity in the imaging cham-

ber was controlled to within 30%-40% for all

experiments included in fig. 6. Despite the fact

that the specimen preparation conditions were

not identical and that the tip geometry must vary

for individual cantilevers, the correlation between

the adhesion force and the lateral resolution is

clearly recognizable in fig. 6. The variation in the

lateral resolution for the same adhesion force

may be an indication of the variation in the tip

geometry and/or in the specimen preparation.

The latter is expected to give some noticeable

deviation for the lateral resolution, because there

is no control of the amount of cytochrome C

binding on the DNA which should definitely af-

fect the molecular width. To ensure the consis-

Fig. 7. Three images of the same DNA molecule under probe forces ot (a) 2, (b) 9 and tc) 12 nN, as shown by the force curves

below. For (b) and cc), dashed lines were drawn to indicate the relative levels of the probe forces, because they were out of the

range of the display. The horizontal dashed line for (b) and (c) were positioned according to a linear extrapolation of the set point

voltage calibration. The resolution remained approximately constant. But in (cl, the molecule appeared damaged in many places.

Scanning speed for these three images was 4.73 Hz. Humidity was 38%.

5/20/2018 Paper

8/14

164

J. Yang Z. Shao / Effect of probe force on the resolut ion of AFM of DN A

tency in these measurements and to reduce the

effect just mentioned above, only those speci-

mens that showed smooth and continuous

molecules were plotted in fig. 6. For adhesion

forces larger than 30 nN, DNA images were no

longer obtainable, even when the probe force was

set below the maximum value of the adhesion

force. A very practical conclusion from fig. 6 is

that the adhesion force may be used as a diagnos-

tic method to evaluate whether the imaging con-

ditions are proper for a given experiment. If the

peak value of the adhesion force is above 30 nN,

the best solution would be to change either the

specimen or the cantilever. Another conclusion is

that the commercially available cantilevers are

quite consistent in performance. Almost all the

cantilevers we have used can yield 4-6 nm resolu-

tion when properly cleaned as indicated by a

small adhesion force (see below for details). It is

interesting to note that, when the adhesion force

is below

- 3 nN, the lateral resolution reached a

range of 4-6 nm, with many tips even capable of

a resolution of - 3 nm under optimal conditions.

Due to the binding of cytochrome C, it is not

clear whether this broadening is caused by the tip

size or the protein binding. But it is safe to

conclude that if the adhesion force is small, the

tip broadening is unexpectedly small for most

cantilevers, considering the radius of curvature of

most tips. Based on these results, the merit of a

supertip at this resolution scale is not apparent, if

one considers that, even in liquid, the supertips

only achieved a similar resolution [12,16].

3.3. Absol ut e probe for ce and adhesion for ce

The adhesion force, regardless of its origin, is

due to an interaction between the tip and the

specimen. Therefore it is intrinsic and depends

on the specimen preparation and the tip condi-

tion rather than the instrument. Logically, one

would assume that, even for a small adhesion

force, the resolution would deteriorate in a simi-

lar fashion as in fig. 6, if the probe force (the

absolute force) is raised to a similar level as the

peak value of the adhesion force. This can be

easily examined by changing the set point voltage

on the NanoScope II AFM. Surprisingly, the ex-

periment did not comply with this expectation.

Fig. 7 shows a series of 3 images with a probe

force of 2, 9 and 12 nN, respectively. Despite the

change in probe force by a factor of 6, the resolu-

tion remained roughly the same. However, at a

probe force of 12 nN, part of the DNA molecule

was damaged as indicated by the discontinuity

along the contour of the DNA molecule. Further

increase of the probe force to 15 nN resulted in a

sudden degradation in resolution, with an in-

crease of the adhesion force to 20 nN at the same

time. For all experiments conducted, it was found

consistently that the resolution is related to the

absolute probe force in an all-or-none fashion.

In other words, the resolution would remain al-

most the same, up to a threshold force at which

both the adhesion force and the molecular width

suddenly jumped to a larger value. These results

seem to indicate that the larger probe force ap-

plied to the specimen induced some kind of con-

tamination on the tip, presumably from fragments

of the specimen, which appeared as an increase

in the adhesion force. Our experiments consis-

tently showed that, for a larger adhesion force,

the specimen could normally withstand a much

larger probe force (up to 30 nN) compared with

the case of a very small adhesion force. These

observations are consistent with the results dis-

cussed in section 3.2. Apparently, the adhesion

force should be used as an indicator for the

condition of the tip during AFM imaging.

Rubber Glove Opening

Gas Outlet

Fig. 8. Illustration of the humidity control. Valves at dry

nitrogen inlet and gas outlet on the top of the humidity

chamber were used to control the speed of nitrogen gas influx,

therefore, to regulate the humidity. A rubber glove opening at

the top of the chamber was used to adjust AFM detector

position if necessary.

5/20/2018 Paper

9/14

Fig. 9. 1

and (d)

J. Yang Z. Shao / Effect of probe

f o rce on

he

r e so l u t i o n f

AFM

of

DNA

16.5

19.94 nm/div

19.94 nm/div

I a

..z

. ..> .

:::: :

./ .i 1 (Arb Unit)

./ .i..

I .:::/Q: _ /.. : :_: ...I. / ,,y;:,

_i.

..i. ,....i,.....:,..i.,,: .. / :... .., i.

_ - ._ 7

,..../ _i. .

-0 (Set Point)

.i

,:::: 1;:. 1:: ::::/j:. .::;

* :

. i:: .j: f _;. .j: i ::I

; ..; ,. . .,..... j ,:::

19.94 nm/div

19.94 nm/div

Four AFM images of the same DNA molecule with corresponding force curves, at humidities of (a) 37%, (b) 60%

90%. Neither the lateral resolution nor the adhesion force showed any appreciable change. Since the adhesion

not change with the humidity, it is not clear whether it is from the capillary force. Scanning speed was 4.73 Hz.

f

Cc)

30%

orce did

5/20/2018 Paper

10/14

166

J. Yung Z. Shao / Effect of probe force on the molut ion of AFM of DNA

3.4. Effect of humidity

Although the effect on resolution due to the

adhesion force and absolute probe force was

studied under a controlled humidity, the effect of

humidity has been indicated to change both the

adhesion force [30-321 and the resolution of AFM

on DNA with different specimen preparatory

methods [11,33]. With the cytochrome C spread-

ing method, we carried out a series of experi-

ments in a humidity-controlled chamber where

the relative humidity can be adjusted continu-

ously (see fig. 8 for an illustration). To minimize

the effects due to other factors, such as tip con-

tamination, every effort was made to ensure that

the adhesion force was less than 3 nN before

humidity variation was started. To our surprise,

we found that within the experimental error, the

resolution of DNA molecules was unchanged over

a range from less than 5% to about 95% in

humidity. One example is shown in fig. 9 where

the images were taken at humidities of 37%,

60%, 80% and 90%, respectively, over a period of

- 7 h while the tip remained in contact with the

specimen surface. The corresponding force curves

are also shown at these conditions. At least for

this kind of specimen, the change in adhesion

force versus humidity is insignificant, which seems

to indicate that the so-called capillary force is

unimportant in this case. The reason might be the

protein film present at the surface which has

quite different properties from those materials

we are familiar with. We should also point out

that, for higher humidities (> 50%), long-time

disengagement of the tip often resulted in a much

increased adhesion force with much degraded

resolution when reengaged. However, as long as

the tip was engaged, high humidity did not show

significant degradation on image quality, even

over a long period of several hours.

3.5. Imaging in organic solvents

In organic solvents, like isopropanol and

n-

butanol (also known as isopropyl alcohol and

n-butyl alcohol), both cytochrome C and DNA

are insoluble. Therefore, under such liquids DNA

molecules should be stable at the substrate-liquid

4.83 nmldiv

Fig. 10. AFM image of a DNA molecule taken in a fluid cell

under isopropanol, with the corresponding force curve. The

distance between two arrows in the force curve was about IO

nm, corresponding to a probe force of 0.6 nN for a cantilever

of

k =

0.06 N/m. Scanning speed was 4.73 Hz. Notice that

the adhesion force is absent.

interface. These solvents can also be used as

dehydrating agents. The role of dehydration in

AFM imaging is not clear at present. However,

there is one advantage for AFM of DNA in these

organic solvents: the capillary force, if any, is

entirely eliminated, so that the probe force can

be further decreased. For AFM of DNA in these

liquids, we found an interesting phenomenon:

usually within the first half hour or so after

engagement of the tip, no DNA molecules were

found and the probe force fluctuated a great

deal. 30-40 min later, the fluctuation gradually

damped down, and only then could the probe

force be set to sub-nN range. The adhesion force

was normally absent. Shown in fig. 10 are an

image of a DNA molecule in isopropanol and the

corresponding force curve. The resolution of

DNA molecules obtained in organic solvents is

quite similar to those obtained in air, i.c. 4-6 nm.

But, occasionally, on poorly prepared specimens

5/20/2018 Paper

11/14

.I. Yang Z. Shao / Effect of probe force on the resolut ion of AFM of DNA

167

Fig. 11. Close up of a DNA molecule taken under n-butanol.

Arrows indicate where a lateral resolution of - 3 nm was

measured.

that DNA appeared discontinuous in air, a reso-

lution around 3 nm was obtained (fig. 11, as

indicated between the arrows>. The clumps along

the DNA molecules may be the bound cy-

tochrome C molecules. Since the double helix

should have a diameter of - 2 nm (although 2.7

nm was reported by electron diffraction studies

[34]), the measured molecular width still shows

some broadening. It is not clear whether this

residue broadening is caused by the cytochrome

C molecules or the AFM tip itself. Should the

former be the primary contribution, the tip

broadening would be exceedingly small. It should

also be stressed that such a good resolution was

only obtainable with those specimens considered

rather poor for AFM imaging in air. Since the

amount of broadening is comparable to those

observed on native DNA molecules with super-

tips [12,16], whether a supertip will improve upon

this resolution with these specimens is not clear.

4.

Discussion

Based on the experimental findings presented

so far, it is plausible to conclude that, at least for

the DNA specimens using the cytochrome C

spreading method, the contamination on the

probe tip is the single most important factor in

limiting the resolution achievable by AFM, rather

than the tip geometry it is often attributed to.

This conclusion may be understood by the follow-

ing: the adhesion force is related to the contact

area and the contamination near the tip would

increase this contact area. When the contact area

becomes exceedingly large, the resolution must

be degraded. If this is correct, we would have:

resolution a \/adhesion force. In fact, the data

presented in fig. 6 are quite consistent to this

notion with the power as 0.48 after a least-squares

fit. Furthermore, the observation that specimens

can only withstand a rather small probe force

(< 15 nN1 for a clean tip provided another

support for this conclusion, because for a contam-

inated tip (large contact area), the pressure would

be much smaller than a clean tip (small contact

area) for the same probe force, while the speci-

men damage should be directly related to the

pressure, rather than the force itself.

For a tip that showed a very large adhesion

force (contaminated), we found that scanning

the substrate (carbon-coated mica) can often lead

to a dramatic reduction of the adhesion force

which, in turn, provided a much improved resolu-

tion. Details of depositing a carbon film on mica

have already been described in section 2.1 and

the thickness of the carbon film is unimportant. It

was found that a scanning size of 10 pm or larger

was more effective, and the typical scan speed

was 5.79 Hz. Only when the adhesion force re-

mained small and stable (most often it did) dur-

ing subsequent scanning for at least 2 min, we

considered the tip as clean. For such a clean

tip, the adhesion force on the carbon film often

remained small (- 2 nN) for hours with only

small variations (< 10%). Most tips can be

cleaned two to three times. Interestingly, such

a procedure required the ambient humidity below

40%, and prolonged disengagement of the tip

could render the procedure ineffective. For higher

humidities (> 45%), such a procedure did not

work well. In these cases, the scanning of the

carbon film can still reduce the adhesion force.

But after the change of the specimen, the adhe-

sion force returned to the larger value. A DNA

specimen may also be used for this cleaning

5/20/2018 Paper

12/14

J. Yang, Z. Shao /

f fec t of

probe force on the resolution qf

AFM

of

DNA

.:/. 2

, 1 (Arb Unit)

j.. ;. :

i_/ -L

--_-.+- -- __/

I--L-lc- 0 (Set Point)

24.17 nm/div

j i

.I

ail::- ~~111:1:~~~r-_

2

21.75 nm/div

3.02 nrhiv

Fig. 12. Four images of DNA molecules obtained by the same tip on the same specimen. (a) At an adhesion force of 27 nN. the

DNA molecule showed a resolution about 1X nm. (b) After cleaning the tip using the procedure described in the text, a typical

AFM image of a DNA molecule is shown, with an improved lateral resolution and a much reduced adhesion force ( - 1.4 nN). (c)

The same molecule as in (b) when the adhesion force was increased to 25 nN with a probe force of 22 nN. (d) Another cleaning

recovered the better resolution and the smaller adhesion force. Scanning speed for the above images was 4.73 Hz. Humidity was

30%.

5/20/2018 Paper

13/14

.I. Yang Z. Shao / Effect of probe force on the resolut ion of AFM of DNA

69

procedure, but the results were less consistent,

presumably due to the variation of the specimen

preparation and the soft nature of the cy-

tochrome C film. Fig. 12 shows an example. In

fig. 12a, we see an image obtained with a fairly

large adhesion force at a rather poor resolution.

In fig. 12b, a high-resolution image is shown

which was obtained on the same specimen with

the same tip after the cleaning process. The ad-

hesion force now is only 6% of that in fig. 12a,

while the resolution is improved by a factor of

3-5. During prolonged imaging, sometimes the

resolution became suddenly degraded, accompa-

nied with an increased adhesion force (fig. 12~).

In this case, another cleaning process can often

restore the image quality, as shown in fig. 12d

which was the result of the second cleaning.We

found that, following this procedure, the same tip

can be repeatedly used on several specimens.

The mechanism of the cleaning procedure

cannot be determined based on these experi-

ments. Although it is possible that the scanning

on carbon film caused some structural change to

produce some sort of a supertip, several argu-

ments would favor the interpretation that con-

tamination at the tip was removed by such a

procedure. First, that at higher humidities the

cleaning procedure did not work well and that

the cleaning was more effective with a short

specimen switching time indicated that the treated

tip had a finite life time standing in air and

shorter life time at higher humidities. This would

argue against a restructured supertip, which

should be insensitive to air exposure or humidity

change. It may be further argued that, at higher

humidities, water condensation on the tip should

be more significant that could backmigrate the

contamination pushed aside by the cleaning

procedure to the tip when suspended from the

specimen surface. Secondly, when imaging in or-

ganic solvents, the tips were not treated by the

cleaning procedure. Still most of them showed

high-resolution images of DNA, although it would

take some time after initial engagement (a few

minutes). This is consistent with the assumption

that contaminations on the tip, presumably some

organic materials, were dissolved away by the

solvent. Since the adhesion force does not change

with humidity appreciably, and high-resolution

images were also obtained in organic solvents, the

contamination is most likely due to hydrocarbons

abundant in air. If this is indeed the case, some

experimentation with different tip materials,

clean-room storage and operation, as well as tip

cleaning with other solvents or laser ablation,

may prove to be a fruitful approach. Despite the

above arguments, the possibility of structural

change due to such a cleaning procedure can-

not be entirely ruled out. A direct visualization of

the treated tip would be required before we can

fully understand the mechanism of such a proce-

dure

In conclusion, experimental results have shown

that the adhesion force is the most important

factor in obtaining high-resolution AFM images

on DNA. The commercially available cantilevers

are sufficient to provide resolutions up to 3 nm

for DNA and 1.8 nm for fully hydrated mem-

brane proteins [6]. Humidity control is much de-

sirable if the ambient humidity is above 40%.

Elimination of the electrostatic interaction is the

first crucial requirement for obtaining prelimi-

nary images on these specimens. As long as these

procedures are followed, reproducible high-reso-

lution DNA images can be obtained by AFM

routinely with minimal requirements on the envi-

ronment and the operator skill. Further improve-

ment on the resolution may require the use of the

so-called supertips. However, for the same probe

force, the pressure will be drastically increased.

whether these specimens can withstand such a

pressure to give even better resolution is un-

known. But, if the specimen strength eventually

becomes the limiting factor in resolution, other

techniques might be introduced, such as limited

chemical fixation and low-temperature AFM. The

most important conclusion is that, at room tem-

perature and nN force level, small broadening on

these DNA molecules can be achieved by AFM

with commercially available instruments.

Acknowledgements

We are grateful to Drs. A.P. Somlyo and A.V.

Somlyo for helpful discussions. We would also

5/20/2018 Paper

14/14

170

J. Yang Z. Shao / Effect of probe force on t he resolution of AFM o f DNA

like to thank Drs. H. Yu and K. Takeyasu for the

gift of DNA. This work is supported by grants

from the Whitaker Foundation, US Army Re-

search Office (DAAL03-92-G-0002), the National

Institutes of Health (POl-HL-48807), and the

Jeffress Memoral Trust (J-265).

References

[l] H.-J. Butt, K.H. Downing and P.K. Hansma, Biophys. J.

58 (1990) 1473.

[2] J.H. Hoh, R. Lal, S.A. John, J.-P. Revel and M.F. Arns-

dorf, Science 253 (1991) 1405.

[3] J.A.N. Zasadzinski, C.A. Helm, M.L. Longo, A.L.

Weisenhorn, S.A.C. Gould and P.K. Hansma, Biophys. J.

59 (1991) 755.

[4] A.L. Weisenhorn, B. Drake, C.B. Prater, S.A.C. Gould,

P.K. Hansma, F. Ohnesorge, M. Egger, S.-P. Heyn and

H.E. Gaub, Biophys. J. 58 (1990) 1251.

[5] H.G. Hansma, A.L. Weisenhorn, A.B. Edmundson. E.

Gaub nd P.K. Hansma, Clin. Chem. 37 (1991) 1497.

[6] J. Yang, L.K. Tamm, T.W. Tillack and Z. Shao, J. Mol.

Biol. 229 (1993) 286.

[7] J. Yang, A.V. Somlyo, M.K. Reedy, K. Takeyasu, L.K.

Tamm, M. Allietta, T.W. Tillack and Z. Shao, in: Proc.

50th Annual EMSA Meeting, Boston, MA, 1992.

[8] J. Yang, L.K. Tamm, T.W. Tillack, K. Takeyasu and Z.

Shao, Biophys. J. 64 (1993) A222.

[9] J. Yang, K. Takeyasu and Z. Shao, FEBS Lett. 301 (1992)

173.

[IO] Y.L. Lyubchenko. A.A. Gall, L.S. Shlyakhtenko, R.B.

Harrington and S.M. Lindsay, Biophys. J. 61 (1992) A149.

[ll] C. Bustamante. J. Vesenka, C.L. Tang, W. Rees, M.

Guthod and R. Keller, Biochemistry 31 (1992) 22.

[12] H.G. Hansma, J. Vesenka, C. Siegerist, G. Kelderman.

H. Morrett, P.L. Sinsheimer, V. Elings, C. Bustamante

and P.K. Hansma, Science 256 (1992) 1180.

[13] F. Zenhausern, M. Adriah, B. Ten Heggeler-Bordier, R.

Emch, M. Jobin, M. Taborelli and P. Descouts, J. Struct.

Biol. 108 (1992) 69.

]141

]151

]161

]171

E. Henderson, Nucleic Acids Res. 20 (1992) 445.

Y.L. Lyubchenko, B.L. Jacobs and S.M. Lindsay, Nucleic

Acids Res. 20 (1992) 3983.

H.G. Hansma, R.L. Sinsheimer, M.-Q. Li and P.K.

Hansma, Nucleic Acids Res. 20 (1992) 3585.

Y.L. Lyubchenko, L.S. Shlaykhtenko, R.E. Harrington,

P.I. Oden and S.M. Lindsay, Atomic force microscopy of

long DNA: imaging in air and under water, Proc. Natl.

Acad. Sci., submitted.

]181

]191

001

]211

]221

]231

1241

]251

]261

]271

[281

]29]

B.C. Westmoreland, W. Szybalski and H. Ris, Science

163 (1969) 1343.

J. Griffith, J.A. Huberman and A. Kornberg, J. Mol.

Biol. 55 (1971) 209.

R.B. Inman and M. Schnos, J. Mol. Biol. 49 (1970) 93.

C.R. Clemmer and T.P. Beebe, Jr., Science 251 (1991)

640.

W.M. Heck1 and G. Binnig, Ultramicroscopy 42-44 (1992)

1073.

T.R. Albrecht, S. Akamine, T.E. Carver and CF. Quate,

J. Vat. Sci. Technol. A 8 (1990) 3386.

H. Ximen and P.E. Russell, Ultramicroscopy 42-44 (1992)

1526.

A.K. Kleinschmidt, Meth. Enzymol. 12 (1968) 361.

N. Kawanishi, H.K. Christenson and B.W. Ninham, J.

Phys. Chem. 94 (1990) 4611.

J.D. Jackson, Classical Electrodynamics, 2nd ed. (Wiley,

New York, 1975).

H.W. Hao, A.M. Baro and J.J. Saenz, J. Vat. Sci. Tech-

nol. B 9 (1991) 1323.

]301

[3ll

]321

]331

A.L. Weisenhorn. P.K. Hansma, T.R. Albrecht and C.T.

Quate, Appl. Phys. Lett. 54 (1989) 2651;

J. Yang and Z. Shao, unpublished results.

J. Israelachvili, Intermolecular and Surface Forces, 2nd

ed. (Academic Press, London, 1992).

J.Y. Yuan, Z. Shao and C. Gao, Phys. Rev. Lett. 67

(1991) 863.

J.Y. Yuan, Z. Shao and C. Gao. Phys. Rev. Lett. 68

(1992) 2564.

T. Thundat, R.J. Warmack, D.P. Allison. L.A. Bottom-

ley, A.J. Lourenco and T.L. Ferrell, J. Vat. Sci. Technol.

A 10 (19Y2) 630.

]341

K.H. Dowinig, Ultramicroscopy I3 (1984) 35.