A Laurin Publication

Digital Mammography UpdateDigital Mammography Update

Photonic Solutions for Biotechnology and MedicineMay 2003

MicrMicropositioningopositioningfor Micrfor Microscopyoscopy

ShrinkShrinkUltrafast LasersUltrafast Lasers

Automated stage motion control iscritical in high-performance mi-croscopy techniques that require

motion accuracy or speed that cannot bematched by human control. These tech-niques include six-dimensional imaging;montage imaging, or stitching; stereol-ogy; and automated multiwell platescreening.

To perform six-dimensional imaging,three-dimensional data (X-Y-Z) is col-lected across time at different fluorescentwavelengths or imaging modes and at dif-ferent spots on an X-Y stage. Imaging in

six dimensions is essentially multiplex-ing of 4-D multiwavelength imaging. Itallows the collection of a vast amount ofdata from living cells during a single col-lection sequence. Researchers use this tech-nique because biological conditions canbe difficult to reproduce when workingwith living cells. Multiplexing allows themto perform robust statistical analysis be-cause many examples of a condition arecollected at the same time. Revisiting thesame X and Y positions at multiple timepoints requires precise and rapid auto-mated stage motion.

Montage imaging, or image stitching, isthe seamless collection of data that islarger than the frame size of a single image(Figure 1). It is critical for constructinghigh-resolution images of tissue or ex-haustive sampling of an individual bio-logical sample. This involves revisiting thesame site again and again to look forchanges that the cell may be undergoing.Such changes may include growth in neu-rons, or changes in the physiology of thecell, such as calcium-ion concentrationsused for intercellular communications.Accurate automated stage movement isrequired to move the correct distance be-tween frames.

Stereology involves sampling, or count-ing, a specific type of cell so that an un-biased statistical analysis of an event canbe made. For example, if a scientist isstudying a neurological disorder andwants to statistically analyze the number

Reprinted from the May 2003 issue of Biophotonics International © Laurin Publishing Co. Inc.

SUMMARY

Submicron positioning stages with resolutions in the 20- to 50-nm rangeand repeatability accuracies as low as 200 nm can be combined with pow-erful imaging software to make your microscope more effective.

by John Zemek, Dr. Colin Monks and Dr. Ben Freiberg

Figure 1. Montage imagingof this fluorescence-labeledrat hippocampus alloweddata to be collected from anarea larger than the framesize of a single image.Accurate automated stagemovement is required tomove the correct distancebetween frames in thistechnique. The final imageconsists of 80 individualframes stitched to form aseamless picture. Courtesyof Karl Kilborn, IntelligentImaging Innovations Inc.

DDISCOISCOVERVERYY

Through Automation

MICROPOSITIONING

of neurons in a brain, he can take multi-ple samples from the brain, prepare themand mount them on microscope slides.The stereology program will pick a ran-dom area of the sample to examine, andthe researcher will count the number ofneurons present. This is done to a num-ber of samples, and because the exact vol-ume/area of the random samples isknown, the number of neurons withinthe complete volume of the brain can bestatistically projected. Because a computerdetermines the exact locations within asample that will be scored, it must controlthe position of the stage (Figure 2).

Automation of multiwell-plate imag-ing has become the basis of many bio-logical screening procedures. Accurate andrapid automated stage positioning allowsthe same spot in each well to be imaged.

Many applications that can benefit fromautomated micropositioning systems inmicroscopy demand similar requirementsfrom their motorized equipment. Someimportant specifications are repeatabil-ity, resolution and speed.

Repeatability is related to the accuracyof a micropositioning system. This spec-ification should be clarified as either uni-directional or bidirectional. Unidirectionalrepeatability is the ability of the stage toreturn to a given point, always comingfrom the same previously defined direc-tion. Bidirectional repeatability is the abil-ity of the stage to return to a given pointfrom any random previous point. Thevalue for each should be the standard de-viation of many moves to give a solid sta-tistical specification.

In many applications, such as multi-point capture, multiwell screening andstereology, the stage must move preciselyto a given X-Y coordinate. When imagingcells at high magnification, small shiftsin positioning can lead to huge shifts inthe location of a cell in the field. For ex-ample, if a user is trying to image manycells in multiple fields over time, it is vitalthat the motorized positioning system lo-cate the same fields repeatedly. Analysis ofimages that have poor alignment due topoor stage repeatability is difficult be-

Figure 2. Stereology is a technique thatallows counting of a specific type ofcell for statistical analysis. Becausesoftware such as this from IntelligentImaging Innovations determines thelocations for scoring, precision X-Y-Zmotorization ensures that counts arestatistically unbiased.

Antibacklash nut — A nut that is spring-loaded, or held

under tension, in some manner to prevent lost motion, or

backlash, when the direction of travel is reversed.

DC stepper motor — A motor that moves a given amount

when a DC pulse is applied to it. Typical stepper motors offer

200 to 400 steps per revolution, but the devices can also

be microstepped to get finer resolutions. Stepper motors are

usually used in an open-loop system that does not em-

ploy feedback encoders but, rather, relies on counting the

pulses sent to the motor for positioning information.

DC servomotor — A motor that moves a given amount

when a DC voltage is applied to it. Unlike a stepper motor,

it has an infinite resolution. However, servomotors must be

used with an encoder to provide positioning information.

Encoder — A device used to measure movement by count-

ing pulses. The encoder can be optical and count pulses, via

a grating etched in glass or plastic, or magnetic and count

a magnetic field. There are two basic types, those that mea-

sure rotary movement and those that measure linear move-

ment. Depending upon the resolution of the scale and the

interpolation used, encoders can measure movement in

the nanometer range.

Leadscrew — A precision ground screw that is machined

with a constant known pitch. If the leadscrew is machined

to exacting tolerances, a nut attached to it will travel an

exact linear amount plus or minus the tolerance of accu-

racy when the leadscrew turns. An antibacklash nut is usu-

ally used with the leadscrew to prevent lost motion when

the direction of travel is reversed. G

Glossary of Micropositioning Terms

cause objects may be in vastly differentlocations at subsequent time points. Totake this one step further, if the user wantsto measure how much the cells are mov-ing in different fields, repeatability be-comes even more critical because errors instage positioning will directly affect thedistance the object apparently travels.

The resolution factorThe resolution — the smallest step size

that the stage can move in a repeatablemanner — is also extremely important toassess when purchasing microposition-ing equipment. For example, if a scientistwanted to image a cell 100 µm in lengthand the image field captured only 50 µmof the cell at a time, a minimum of twofields would be needed to capture the en-tire cell. If the X-Y resolution of the stagewere 100 µm, it would be impossible toimage the entire cell.

However, if the resolution were 25 µm,the entire cell could easily be imaged. Atan even finer resolution, the stage couldbe positioned with even less overlap be-tween subsequent images. One thing to remember is that the resolution of mi-cropositioning equipment is usually lessthan its repeatability. Therefore, it is es-sential to look at both numbers whenevaluating devices.

The time it takes a stage to move be-tween fields can determine whether a userobserves or misses a phenomenon. Stageswith a high degree of repeatability andresolution often sacrifice speed. Movingquickly and achieving high precision isdifficult and is a major trade-off in mi-cropositioning equipment because, insome systems, the pitch of the leadscrewdetermines the resolution of the stage.The finer the pitch, the finer the resolutionand the slower the stage will move, be-cause it takes many more turns for theleadscrew to cover the same area thanwith a coarser-pitch leadscrew.

Speed is also important in many screen-ing applications. Faster movement of sam-ples leads to higher-throughput systems.When imaging multiwell plates, speed be-comes compounded between each welland each plate. At the end of the day, smalldifferences in speed can mean the differ-ence between capturing hundreds of im-ages or just a few dozen.

Other major factors to consider includefeedback, motor type, encoder resolutionand backlash.

Closed loop, open loopClosed-loop systems differ from open-

loop systems in that the former use anencoder to feed back the position of thestage or fine focus shaft of the microscope.For example, in a microscope with aclosed-loop automated Z-axis drive, thestage can be commanded to move 1.5 µmtoward the objective. The command isconverted into a DC voltage that powersa DC servomotor attached to the fine orcoarse focus shaft. A high-resolution en-coder allows the control electronics toknow when the stage or the shaft has ac-tually moved the 1.5 µm, and then thecontrol electronics can tell the motor tohold its position. However, an encodercan add substantial cost.

An open-loop system uses no feedback.Instead, a DC stepper motor’s controllerkeeps track of position by counting thepulses it sends to the motor because they

usually result in the proper linear move-ment of the Z-axis drive or X-Y stage.Stepper motors are commonly used inlow-cost positioners, but if somethingbinds the stage’s leadscrew or the steppermotor’s shaft, motion is lost and the sys-tem can lose its true position.

Systems that utilize closed-loop feed-back are usually more precise than open-loop systems because the device ensuresthat the commanded position is reached.DC stepper motor systems can be oper-ated in either open- or closed-loop mode.To obtain fine resolutions, they are usu-ally stepped in minute increments. DCservomotors, on the other hand, have avirtually infinite resolution and are al-most always used with a feedback encoderto provide a precise closed-loop design.

The resolution of the encoder deter-mines the accuracy of a closed-loop sys-tem. Encoders used in microscopy appli-

Figure 3. Precision Z-motorization allowed this 3-D triple fluorescent image of amacrophage engulfing an apoptotic cell. Courtesy of Jennifer Monks, NationalJewish Hospital, Denver.

cations measure either rotary or linearmovement. Rotary encoders measure theposition of the microscope’s focusing shaftfor Z-axis movement, the position of themotor shaft or the leadscrew for X-Y stagepositioning. If properly placed on a par-ticular microscope or stage, they can pro-vide a resolution from 0.1 µm to 50 nmand a repeatability of around 1 µm rms.Rotary encoders hold the advantage ofbeing easily installed, and they provide avery good price/performance ratio.

Encoders and backlashHowever, they can introduce unac-

counted backlash — systematic error cre-ated by lost motion in the drive mecha-nism that appears when changing direc-tion. For the Z-axis, the backlash is thetotal of all of the gaps between the gearson the focusing mechanism that movethe stage. It can vary between about 0.10and 5.0 µm depending upon the designand adjustment of the microscope andthe wear on these mechanical compo-nents. For X-Y stages, the backlash error isthe total mechanical gap between themotor that drives the leadscrew to positionthe plates of the stage to the optical axis.Antibacklash gear heads in the motor andantibacklash nuts on the leadscrewsshould be used to reduce thiserror. The control electronicscan use algorithms to reduce itas well.

Linear encoders also can ac-count for backlash by measur-ing the actual linear movementof the stage. This provides amore accurate means of mea-suring stage position and al-lows the use of coarser-pitchedleadscrews that, in turn, in-crease the speed with little orno loss in resolution. However,high-quality submicron linearencoders are relatively expen-sive and must be properly in-stalled and aligned to provideoptimal performance.

These devices are availablewith resolutions down to 5 nm,but because of the mechanicallimits of cross-roller bearingsand leadscrews used in high-quality X-Y stages, realistic res-olutions at the optical axis areabout 50 nm. Higher resolu-tions require air bearings or

more sophisticated designs that offer lessfriction. Standard stage designs that usecross-roller bearings and leadscrews andwell-designed controller electronics andlinear encoders can achieve bidirectionalrepeatability accuracies on the order ofless than 200 nm rms for moves of a fewhundred microns or so, and about twicethat on larger moves. When working atthese levels, the thermal expansion of themetal from which the stage is built mustalso be considered.

There are a number of other factors toconsider when selecting a stage, and ifyou went into all of the details, a com-plete textbook could be written on thetopic. Any stage manufacturer or imag-ing system dealer should be able to helpyou with these points and aid you in se-lecting a stage that is right for your ap-plication. With the right stage and a well-designed imaging software package, youcan make your microscope a more pow-erful research tool. G

Meet the authorsJohn Zemek is executive director at Applied

Scientific Instrumentation in Eugene, Ore.Dr. Colin Monks is co-president and Dr.

Ben Freiberg, vice president, at IntelligentImaging Innovations in Denver.

Figure 4. A closed-loop DC servomotor X-Y stagefrom Applied Scientific Instrumentation ismounted on an Olympus microscope. DCservomotors are almost always used with afeedback encoder to provide a precise closed-loop design.

Anumber of manufacturers

offer automated X-Y stages,

and each publishes resolu-

tion, repeatability and accuracy val-

ues for them. Many factors affect po-

sitioning accuracy, so the actual spec-

ifications for each stage may vary

slightly from the published specifi-

cations.

Most manufacturers use DC step-

per motors in microstepping mode

to control the movement of their

stages. Because stepper motors are

usually used in an open-loop config-

uration, it is important that the stage

be well-designed and thoroughly

tested to ensure accuracy and re-

peatability. In open-loop systems,

mechanical wear or contaminants on

the bearing surfaces can affect accu-

racy. Linear encoders in combination

with DC servomotors can provide a

very accurate system.

For precision stages, quality control

testing should be performed using

high-resolution equipment such as

laser interferometers, or other test

jigs such as optical imaging systems.

In the latter, a CCD camera is con-

trolled by software to take measure-

ments from the images. One can

measure positional accuracy of less

than 20 nm.

The target can be any diffraction-

limited sample, such as a speck of

dust. The position of the center of

mass of the speck image is simply

measured after various move com-

mands, and the statistical data stored

in a file for the particular stage. This

data can be used to compare the ac-

curacy of the feedback device for a

closed-loop stage or the accuracy of

the stepper motor for an open-loop

device. The raw data and statistical

information for each stage should be

available to the customer. G

The NumberGame