Image metrology and system controls for scanning beam...

Image metrology and system controls for scanning beaminterference lithography

Carl G. Chen,a) Paul T. Konkola, Ralf K. Heilmann, G. S. Pati, and Mark L. SchattenburgMassachusetts Institute of Technology, Cambridge, Massachusetts 02139

~Received 1 June 2001; accepted 16 August 2001!

We are developing scanning beam interference lithography~SBIL! for writing and reading largegratings with nanometer level distortion. Our distortion goals require fringe locking to a movingsubstrate with subnanometer spatial phase error while measuring and controlling the fringe periodto approximately one part per million. In this article, we describe the SBIL optical system designalong with some major subsystems. The design incorporates measurements and controls of theparameters that limit the accuracy of our system. We describe in detail a novel image metrologyscheme, which uses interferometry to measurein situ both the period and the phase of the gratingimage formed by the interference of two laser beams. For a grating period of approximately 2mm,experiments demonstrate a period measurement repeatability of three parts per ten thousand, onesigma. Phase measurement indicates a slow fringe drift at 0.25 mrad/s. Both the repeatability errorand the phase drift are expected to improve by about three orders of magnitude after severalimprovements including the installation of an environmental enclosure and thermally stablemetrology frames. ©2001 American Vacuum Society.@DOI: 10.1116/1.1409379#

gry

ri

t-

ubhP

onThthth

siaatttetiav

u-c-diagnrioyhun

ticsel-

ub-exhealtor,

rate.-

li-theec-of

dontof

o-igh

hym-ter.ntd

I. INTRODUCTION

Interference lithography~IL ! is the process of recordininterference fringes. A good description of IL and its histois available elsewhere.1 Figure 1 depicts our ‘‘traditional’’ ILsystem. Here, the pattern is defined by the interferencespherical waves. Shapes of the fringes produced by sphewaves have been studied in detail1–4 and they contain aninherent hyperbolic distortion. The distortions limit the patern size that can be considered linear.

In the traditional IL system~Fig. 1!, the split beams areconditioned by several optics before interfering on the sstrate. The variable attenuator equalizes the power of eacthe interfered beams and thus maximizes fringe contrast.larizers may be included to ensures-polarized light exposingthe substrate. Spatial filters attenuate wavefront distortiby blocking undesired angular components of the beam.focal length of the lens in the spatial filter is chosen to setdivergence of the beams, thereby defining the size ofregion of interference. The beams typically have a Gausintensity distribution and the spot size on the substrshould be much larger than the desired interference pasize for good dose uniformity. The distance from the spafilter to the substrate defines the radius of the spherical wafront. It is desirable for this distance to be large~typically.1 m! for reduced hyperbolic distortions. However, turblence, vibration, and thermal drift limit the maximum pratical propagation distance. Even with 1 m wavefront raand the assumption of perfectly aligned beams, the imradius with subnanometer accumulated phase distortioonly 1.4 mm. This assumes a 400 nm nominal grating peproduced with a 351.1 nm wavelength laser. Lenses maused to collimate the beam after the spatial filter and teliminate the hyperbolic distortion. However, it is questio

a!Electronic mail: [email protected]

2335 J. Vac. Sci. Technol. B 19 „6…, Nov ÕDec 2001 1071-1023Õ200

ofcal

-ofo-

seeenernle-

ieisdbes

-

able as to whether it is practical to fabricate and align opcapable of producing large gratings with subnanometer fidity.

A phase error sensor located near the plane of the sstrate measures fringe drift, which is mainly due to air indchanges, vibration, and thermal drift of the optical setup. Tdifferential signal from two photodiodes is the error signthat drives the controller for a phase displacement actuawhich actuates so as to stabilize the fringes at the substThe nominal period,L, of the interference fringes is controlled by the beam incident angleu and is given by

L5l

2 sinu. ~1!

Here,l is the wavelength of the laser.Despite the large hyperbolic distortions, interference

thography has many inherent advantages. First of all,interference pattern produces highly coherent gratings. Sondly, the fringes have high contrast over a large depthfocus. Additionally, the topology of a spatial filter followeby no subsequent optics provides extremely low wavefrdistortions. Other advantages include: built-in metrologythe interfered pattern,1,3 a diffraction-limited resolution thatis approximately twice that of traditional on-axis optical prjection lithography, and excellent image contrast even at hnumerical apertures.

II. SCANNING BEAM INTERFERENCELITHOGRAPHY CONCEPT

The goal of the scanning beam interference lithograp~SBIL! system is to write gratings and grids with subnanoeter distortions over substrates up to 300 mm in diameThese ultralow distortion gratings would enable importaadvances in metrology,5 electro-optics, spectroscopy, anmany other applications.

23351Õ19„6…Õ2335Õ7Õ$18.00 ©2001 American Vacuum Society

elereinreb-, tan

in-rat-

ithan-

d-the

redthet-

rmra-nghe

jor

canideso,of

ting

2336 Chen et al. :Image metrology and system controls 2336

Figure 2 depicts the SBIL concept. The optics closresemble those of the traditional IL system but the imagmuch smaller than the total desired patterning area. The ging image diameter is typically designed to be betwe200mm and 2 mm. Large gratings are fabricated by scannthe substrate at a constant velocity under the image. Figuillustrates how SBIL achieves a uniform exposure doseoverlapping scans. Figure 3~a! shows the grating image being scanned along the substrate. At the end of the scanstage steps over by an integer number of grating periodsreverses direction for a new scan. The system has the sigcant added complexity over previous interfer-

FIG. 1. Traditional IL system.

FIG. 2. SBIL concept.

J. Vac. Sci. Technol. B, Vol. 19, No. 6, Nov ÕDec 2001

yisat-ng3

y

hendifi-

ence lithography systems of accurately synchronizing theterference image to the moving substrate. The scanning ging image is illustrated in Fig. 3~b!. The interference patternhas a Gaussian intensity envelope since we interfere wGaussian beams. The effective number of fringes in the scning grating image may be many thousands~i.e., 10 000fringes for a 2 mm 1/e2 image diameter and a 200 nm periograting!. Figure 3~c! shows the individual scan intensity envelopes in dashed lines and the sum of the envelopes insolid line. A maximum step size is constrained by the desidose uniformity. For instance, a step size of 0.9 timesGaussian beam 1/e2 radius produces dose uniformity of beter than 1%.

Beam pick-offs direct a fraction of the power of each ato a fringe locking system. The stage error and the lithogphy interferometer’s signals are fed to the fringe lockicontroller, which in turn stabilizes the grating image in treference frame of the moving substrate.

The relatively small beams used in SBIL provide a mabenefit in ease of obtaining small wavefront distortions.4 Fur-thermore, by scanning the image, distortions along the sdirection can be averaged out. The overlap of scans proveven further averaging of the wavefront distortions. Alscritical alignments, such as lens positioning and angleinterference, are much relaxed for the small beams.

In our approach, we aim to demonstrate repeatable gra

FIG. 3. Grating scan method.

thtei

ril

beg

veee

eroodmtheltrrassu

ysuousuaw

eh

hit

depr

em-

-en-so-alhaseptice-n-heatesata,

ur-arecali-er-beand

to

encyeingbet--c-

ista-ein

rrect

2337 Chen et al. : Image metrology and system controls 2337

writing first. Then, self-calibration methods6 will be imple-mented to correct systematic errors and finally achieveaccuracy goals. We are designing and implementing a syswith a CW 351.1 nm wavelength argon–ion laser thatintended to achieve nanometer level distortions for arbitraselected grating periods in the range of 200 nm to 2mm.Realizing our performance goals is a major challengecause of the many error sources in lithography. We caterize the errors for SBIL into five sources: non-plane waperiod control, substrate and metrology frame, displacemmeasuring interferometer, and lithography interferometerrors. The non-plane wave errors arise from the departurthe interfering wavefronts from planar. The period conterrors are due to the writing period instability and pericalibration inaccuracy. The substrate and metrology fracategory refers to errors from substrate distortion andinability of the metrology reference surfaces to accuratmeasure displacement between the fringes and the subsThe interferometer categories refer to errors in the lithogphy and displacement interferometers. We describe thenals and controls in our system that are designed to meaand correct major error sources.

III. SBIL SYSTEM ARCHITECTURE

Figure 4 shows a simplified schematic of our optical stem design along with some major subsystems. Details sas optical mounts are omitted. The system employs fcolumn-referencing heterodyne interferometers to meastagex- and y-axis displacement and yaw. Redundant yinterferometry enables mapping of the stage systematicrors including stage mirror flatness and orthogonality. Tdesign replaces the reflective beamsplitter of Fig. 2 wit11/21 order beamsplitter because it provides insensitivto the spatial7 and temporal coherence of the laser. Thesign includes a Class 10 environmental enclosure that

FIG. 4. SBIL system design.

JVST B - Microelectronics and Nanometer Structures

em

sy

-o-,ntr-ofl

eeyate.-

ig-re

-chrre

r-eay-o-

vides acoustic attenuation and as well as controls over tperature (60.005 K), relative humidity (,60.8%), andpressure gradient (,15.5 Pa/m). A differential-plane-mirrorinterferometer-based refractometer corrects for environmtally induced index changes, which are mainly due to ablute pressure variations. Fringe drift and jitter in the opticbench are measured and corrected with a heterodyne pmeasurement system and high-bandwidth acousto-ofringe locking.8 The fiber, column reference block, and mtrology optics shown are part of this system. During scaning, a controller locks fringes in the reference frame of tmoving substrate based on an error signal that incorporstage position inaccuracy, index corrections, calibration dand fringe drift measured in the metrology optics.

Figure 5 shows the chuck with metrology reference sfaces attached. Again, details such as optical mountsomitted. The grating reference enables repeatable scalebration of the stage interferometer that is substantially ppendicular to the grating lines. The grating is required toabout the same period that the system is set up to writeis read with a heterodyne fringe reading scheme.8 The imagemetrology grating is a very high accuracy grating usedcombine the reflected 0 and diffracted21 order beams. Acamera images the moire´ pattern between the grating and thinterference image, thus revealing small spatial frequedistortions. The moire´ image also allows us to correct thradius of curvature of the interfering beams by repositionthe collimating lenses. The image metrology grating maya small low-distortion patch of grating cut from a large graing fabricated by traditional IL. Alternatively, SBIL can produce this grating by overlapping many scans tightly, effetively averaging out the grating image nonlinearities.

The alignment and period calibration beamsplitter cubeused forin situ measurement of the period, phase, and rotion of the grating image, the first two of which we havdemonstrated experimentally. Details will be presentedSec. IV. An accurate period measurement enables the co

FIG. 5. Chuck with metrology references.

gh

ticrynsn-aerin

sigth

bythedkine

d

ecs.tha

ntmthD

lint

nginos

ionfer-rift.stemdedtheorstoe to

theheure-thein-

ten

there-

derof

al-al

lrmthethpe-f theare

eter:the-1%

theieswrit-is

ra-b-ub-se.en-to

ac-w

usheper-


stitching of adjacent scans and makes a uniform, hicontrast exposure dose possible. Figure 6~a! shows thescheme. As the beamsplitter is laterally displaced, the oppath lengths for the left- and right-hand side beams will valeading to a sinusoidal intensity variation at a position seing detector~PSD!, whose power readout duplicates the itensity variation. The spatial period of the grating image cthen be derived directly from this intensity readout. Furthmore, by moving the beamsplitter to a fixed location withthe grating image and continuously sampling the powernal from the PSD, we can measure the grating phase atlocation. During the writing or reading of a substrate,periodically referencing the phase of the grating image tobeamsplitter, any slow drift in the optical system is detectthen compensated by adjusting the heterodyne fringe locsystem. In principle, this phase referencing scheme is vsimilar to the automatic write scan correction method usea MEBES electron beam writer, where the electron beamperiodically strobed over a reference grid in order to corrthe scan length drift and the resulting stripe butting error9

The initial setup requires manual rough alignment ofinterference angleu. The period measurement system mesures the actual image period and tunes the angles of iference by adjusting the appropriate picomotor-actuatedror mounts. The optical design for the angle PSD andposition PSD decouples angle and position on the PS~Fig. 4!. The PSD on the chuck~Fig. 5! is used to verify thebeam overlap in the write plane and calibrate the decoupmatrices of the optical bench PSD’s. It would be possibleuse only the beam overlap PSD on the chuck and the aPSD on the optical bench, but, this would require referencthe beamsplitter and the chuck PSD sequentially. The p

FIG. 6. ~a! Image metrology via interferometry.~b! Fringe counting.


-

al,-

n-

-at

e,g

ryinist

e-er-ir-e’s

golegi-

tion PSD on the optical bench allows faster recalibratsince only the beamsplitter on the chuck needs to be reenced. The faster calibration time reduces sensitivity to dOnce the beam overlap and angles are calibrated, the symeasures the optical power from the power reading proviby the overlap PSD. The beams are blankable by nullingamplitudes of the rf signals to the acousto-optic modulat~AOM!. The rf signal amplitudes to each AOM is adjustedprovide the same power from each beam to the substratthe desired exposure dose.

Knowledge of the spot size on the substrate allowsaccurate prediction of the dose and dose uniformity. Tcamera on the optical bench is also used for this measment. Furthermore, the camera allows the imaging ofspot intensity distribution to ensure the beams have thetended Gaussian shape. Monitor wafers are typically writto establish optimal dose settings for a given resist.

During the grating reading, we can directly measurefringe-to-grating motion with a phase interference measument of the reflected zero-order and the diffracted first-orbeams. This is important for determining the repeatabilitythe system and later for self-calibration procedures.

The electronic architecture is shown in Fig. 7. The retime control platform contains a multiprocessor digital signprocessing~DSP! board, D/A, A/D, interferometer, digitaI/O, and digital change of state input cards. This platfocontrols the stage and fringe locking system as well asoverall patterning conditions. It communicates digitally wia LabView based computer that performs functions forriod and phase measurement and alignment. The use otwo separate platforms allows parallel software and hardwdevelopment.

The system is designed to pattern a 300 mm diamwafer in less than 7 min for the following conditions300 mm/s constant scan velocity, 0.25 g acceleration atends of the scans, a 2 mm beam diameter, and a subdose uniformity requirement. This is rapid compared tostate of the art in other direct write patterning technologsuch as electron beam, ruling engine, or scanned lasering. An active vibration isolation system with feed-forwardincorporated to allow fast settling times.

IV. IMAGE METROLOGY

SBIL is designed to create large area gratings with ulthigh phase linearity. Our ultimate design goal calls for sunanometer overall phase distortion across a 300 mm sstrate when compared to a grating with an ideal linear phaTo enable the precise stitching of adjacent scans and tosure a uniform, high-contrast exposure dose, our abilitydetermine the period of the grating image with extremecuracy is critical. Formed by the interference of two narrolaser beams (l5351.1 nm), the grating image has a radiof 100mm to 1 mm. Hence, to achieve 1 nm distortion at tedge of a 1 mm radius spot, we need to measure theiod ~p! to 1 nm/1 mm51026, or one part per million,which translates to 2 pm forp52 mm and 0.2 pm forp5200 nm.


FIG. 7. SBIL system architecture.

rinro-

ntngteglit,ainad

theth

othsShopueecnate

hgisth

xiheeo

alsnd

the

forblely

ftsing

totertherre-ular

toringe-the

the

testy aere

er.L2the

e then-

na-

Exposing the interference pattern in resist and measuthe period of the developed grating with a scanning electmicroscope~SEM! is an extremely slow, laborious, and inaccurate process. Moreover, it yields period measuremeunits traceable to the SEM used, whereas we desire leunits defined by our particular displacement measuring inferometer~DMI !, which is calibrated to a reference gratin~Fig. 5!. A direct aerial image measurement utilizing a sartifact10 is difficult to implement for small grating periodsdue to energy throughput and fabrication reasons. We hthus developed a simple image metrology method usingterferometry, which is capable of measuring the period ofaerial grating image in calibrated units within a few seconand with high accuracy.

Figure 6 demonstrates our approach schematically. Inoverlapping region of the two laser beams, we introducbeamsplitter cube with its interface aligned parallel tointerference fringes. The beamsplitter is mounted on topthe air bearing stage, at a height approximately equal toof the substrate. The reflected and the transmitted beamterfere and produce a signal at the power output of the PBesides producing an intensity readout like an ordinary ptodiode, a PSD is capable of providing an analog outdirectly proportional to the position of a light spot on thdetector. As described earlier for the SBIL system architture, we exploit this useful property to correctly set up aalign the SBIL optics. As the stage moves, the optical plength difference between the two beams changes as wleading to a sinusoidal intensity oscillation at the PSD. Tintensity completes one cycle of oscillation as the statraverses one fringe period. By dividing the stage travel dtance by the total number of oscillations observed atPSD, we can derive the period of the grating image.

The contrast of the intensity fringes at the PSD is mamum when the interfering beams are on top of each otand decreases as the beams slide apart. The total numbfringes~N! is a sum of three terms: the fractional number


gn

inthr-

ve-

ns

eaefatin-D.-t

-dhll,

ee-e

-r,r off

fringesNi andNf , and the integral number of fringesNm , asillustrated in Fig. 6~b!. A precise determination ofNi andNf

depends critically upon our knowledge of the initial and finsignal levels,Vi and Vf , respectively. Furthermore, it restupon our ability to unambiguously distinguish the peaks avalleys of the fringes. For the best resolution,Vi and Vf

should be located close to points where the slopes offringes are steepest.

The stage-mounted beamsplitter can also be usedlong-term drift compensation. Even with the best availaenvironmental enclosure and metrology frames, it is likethat the SBIL system will experience minute thermal driover time. If uncorrected, these will result in the meanderof the grating lines and contribute to phase errors. Priorthe writing or reading of a grating, we place the beamsplitto a location where the interference signal detected atPSD has maximum contrast. The PSD power readout cosponds to the phase of the grating image at that particlocation. For best phase resolution, it is also desirablelocate the grating phase at or near a zero-crossing. Duthe writing or reading, the beamsplitter is periodically rturned to the same location and the phase sampled byPSD. Any detected drift can then be compensated byheterodyne fringe locking system.

V. METROLOGY RESULTS

54 sets of period measurements were carried out tothe repeatability of our measurement system, followed bphase drift study. Throughout, the interference optics wset to produce a grating image periodp'2 mm, and theimage was actively stabilized by an analog fringe lockOther hardware included two On-Trak Photonics UV2duolateral PSD’s, one to ensure the precise overlap ofbeam spots in the substrate plane and the other to sampllight intensity as described in the previous section. A/D coversion is handled by a National Instruments NI 6034E a

i,

a

Tteths

oed

ins

atm5

rio

Arif

tteSe-

rie

aan

tc

on

th az,

ea-ithg


log I/O board. The position sensing resolution of the PSDlimited by the resolution of the 16-bit I/O board to 31 nmand its voltage resolution to 305.2mV.

Figure 8 shows signals sampled at the PSD as the besplitter is displaced by an amountD5400 mm. The samplerate is at 16 kHz. A significant amount of noise is present.implement a robust fringe counting algorithm, we must filthe data digitally. Power spectrum analysis reveals thatdesired periodic signal is at 51 Hz. A finite impulse respon~FIR! low-pass filter~with a Kaiser window! is then appliedwith a cutoff frequency at 70 Hz, a transition band width10 Hz and> 60 dB attenuation in the stop band. The filterdata is also shown in Fig. 8~b!, with very well-defined peaksand valleys. A fringe counting algorithm was implementedLabView, following the procedure outlined in the previousection. Signal levelsVi andVf are calculated by averaginglarge number of points (.8000! both at the beginning and athe end of the scan when the stage is at rest. Repeatedsurements yield the repeatability results shown in Fig. 9.separate measurements give the mean peat pmean51.7644mm. The one-sigmaDp/p repeatability iss52.4531024, or roughly three parts per ten thousand.linear fit to the data indicates that the long term period dwas at 0.061 nm/h.

Following the period measurements, the beamsplicube was moved to a location where the signal at the Pexhibited maximum contrast. The peaks and valleys wVp55.4 V andVv50.7 V, respectively, in the highest contrast region of the scan. A phase drift study was then carout, lasting 45 min at a sample rate of 500 Hz. Figure 10~a!shows the power spectral density calculated from the dgathered. Significant noise power exists between 0.11 Hz, with a standard deviations 0.1 Hz – 1 Hz50.28 Vrms.These mid-frequency fluctuations contribute dominantlythe error in repeatability, as will be shown in the next setion. Time domain data is plotted in Fig. 10~b!, which revealsa long-term phase drift at 0.25 mrad/s. For better illustrati

FIG. 8. Raw and digitally filtered period measurement data.


s

m-

oree

f

ea-4d

t

rD

re

d

tad

o-

,

the data has been processed with a FIR low-pass filter wicutoff frequency at 1 Hz, a transition band width of 0.2 Hand>60 dB attenuation in the stop band.

VI. ERROR ANALYSIS AND DISCUSSION

We have demonstrated the use of interferometry to msurein situ the period and the phase of a grating image wp'2 mm. A simplified best case model gives the followinupper-bound error budget forDp/p51 nm/1 mm51026:

DVi

Vi5

DVf

Vf51.331023 for p52 mm, ~2!

DVi

Vi5

DVf

Vf51.331022 for p5200 nm, ~3!

FIG. 9. Results of the period measurement repeatability.

FIG. 10. ~a! Power spectrum of the dc-subtracted phase data.~b! Phase datain time domain, processed with a low-pass filter with a cutoff at 1 Hz.

uaen

edrticud

ab

ac

orreatgsfoi

nct oad0

e

idcie

nny.of

uearoeo

on

ta-note itthge

hee-

ngndre-thes onbe-

troln.

or-thatin

fer-a

ms.rcesn li-gsub-

ch-tu-ch-aredero.

logy,

0.

ac.


where we have assumed a stage displacement of 400mm, amaximum fringe contrast (n 5 (I max2Imin)/(Imax1Imin) 51)and the existence of only high frequency voltage flucttions. The required initial and final voltage resolution is ttimes more stringent for a grating periodp52 mm than forp 5 200 nm. Intuitively, this makes sense. Given a fixscan length, one can fit ten times more fringes at the shoperiod, hence, the more relaxed specification. In practbecause of imperfect fringe contrast and multiple error bgets assigned toVi , Vf , Vp , andVv , the resolution must behigher.

Period measurement repeatability suffers from three msources of error. They are best classified based on theirhavior in the frequency domain: low frequency drift (f!0.1 Hz), mid-frequency fluctuation (0.1 Hz, f ,1 Hz)and high frequency fluctuation (f @1 Hz). Low frequencydrift due to thermal causes gives rise to the long-term phdrift, which is quantitatively measured in the previous setion. At 0.25 mrad/s, the drift contributes an errs low'131026 towards the period resolution, for a measument time of 10 s. While averaging a large number of dpoints to determineVi andVf can be very effective at gettinrid of high frequency voltage fluctuations, it becomes leeffective at mid frequencies. In fact, one can show thatour particular experiment, frequencies less than 1 Hz wpass through a moving average filter with length 5000. Siour period measurement lasts roughly 10 s, the lower limithe mid-frequency interval is set at 0.1 Hz. We have alrecalculated the amount of mid-frequency noise betweenHz and 1 Hz,s0.1 Hz–1 Hz50.28 V. In other words, for ourexperiment,Vi andVf can be uncertain by an amount60.28V, which, when converted to period resolution error, givrise to smid'431024. Mid-frequency noise is most likelydue to turbulence-induced air index change. Note thats mid,as calculated, represents only a lower limit for the mfrequency noise, as we have ignored the effect of frequenbelow 0.1 Hz. Nevertheless, it should provide a good ordof-magnitude estimate. Finally, atf @1 Hz, we experiencehigh frequency fluctuations,shigh'231025, due to a combi-nation of acoustic noise, stage and mechanical vibratioand laser power fluctuations. Since the standard deviatiothe mean decreases as 1/AN, s high can be reduced further baveraging an even larger number of sampled data points

What we have just presented is only an order-magnitude calculation. However,s tot5As low

2 1smid2 1shigh

2

'431024 corresponds to the measured vals 52.4531024 very well. The error analysis suggests thmid-frequency noise is by far the most dominant ersource. It should be noted that our experiments have bperformed without the use of an environmental enclosureproperly designed low coefficient of thermal expansi


-

ere,-

ine-

se-

-a

srllefy.1

s

-esr-

s,of

-

trenr

~CTE! metrology frames. Even though an active fringe sbilization system is used, the stabilizing action doesstrictly take place at the location of the grating image sincis impractical to implement. Any relative optical path lengchange from the grating image to the point of actual frinlocking may introduce phase jitters.

Soon, the SBIL system will undergo a major upgrade. Tadditions of an environmental enclosure and low CTE mtrology frames, together with a more compact fringe lockisystem, will significantly reduce the repeatability error athe long term phase drift. Furthermore, it is possible toduce the mid-frequency noise by considerably boostingstage scan speed and the data sampling rate. This reliethe establishment of a seamless communication protocoltween our LabView based data I/O and the real-time conplatform ~Fig. 7!, a process that we have started to work o

VII. CONCLUSIONS

We have described the design for SBIL. The design incporates measurements and controls of the parameterslimit the accuracy of our system. We have also describeddetail a novel image metrology scheme, which uses interometry to measurein situ both the period and the phase ofgrating image formed by the interference of two laser beaMeasurement data is presented and different error souanalyzed. Correctly addressing the many error sources ithography should ultimately allow the patterning of gratinand grids with subnanometer distortions over a 300 mm sstrate.

ACKNOWLEDGMENTS

The authors gratefully acknowledge the outstanding tenical assistance of Robert Fleming and Edward Murphy. Sdent, staff, and facility supports from the Space Nanotenology Laboratory and the NanoStructures Laboratoryalso appreciated. This work was supported by DARPA unGrant No. DAAG55-98-1-0130 and NASA under Grant NNAG5-5271.

1J. Ferrera, Ph.D. dissertation, Massachusetts Institute of Techno2000.

2K. Hibinio and Z. Hegedus, Appl. Opt.33, 2553~1994!.3J. Ferrera, M. Schattenburg, and H. Smith, J. Vac. Sci. Technol. B14,4009 ~1996!.

4C. G. Chen, M.S. thesis, Massachusetts Institute of Technology, 2005M. Schattenburget al., J. Vac. Sci. Technol. B17, 2692~1999!.6C. J. Evans, R. J. Hocken, and W. T. Estler, CIRP Ann.45, 617 ~1996!.7P. Konkola, C. G. Chen, R. Heilmann, and M. L. Schattenburg, J. VSci. Technol. B18, 3282~2000!.

8R. K. Heilmannet al., J. Vac. Sci. Technol. B, these proceedings.9H. Pearce-Percyet al., J. Vac. Sci. Technol. B12, 3393~1994!.

10W. N. Partlo, C. H. Fields, and W. G. Oldham, J. Vac. Sci. Technol. B11,2686 ~1993!.

Image metrology and system controls for scanning beam...

Documents

Transcript of Image metrology and system controls for scanning beam...