CN100510967C - Phase shift mask alignment method and device - Google Patents

Abstract

The present invention relates to alignment of a writing system and a workpiece. In particular, it relates to alignment to write a second layer pattern on a workpiece that has a first layer pattern, using an SLM. It extends to producing a mask or a reticle, and to producing a layer of a device using the mask or reticle. Particular aspects of the present invention are described in the claims, specification and drawings.

Description

Phase shifting mask alignment methods and device

Technical field

The present invention relates to the aligning of writing system and workpiece.Specifically, it relates to the aligning that adopts spatial light modulator (SLM) to write second layer pattern on the workpiece with ground floor pattern.It extends to makes mask or scale sheet, or adopts mask or scale sheet to make the layer of device.

Background technology

The mask that various employings have passed through twice composition or the lithography technique of scale sheet have been developed.Composition has produced phase-shift structure for the second time.

The composition second time to mask or scale sheet becomes more and more important with accurate aligning of composition for the first time.With regard to the alternating phase-shift structure, this point is particularly important, and described alternating phase-shift structure may produce in plural pattern step.

One is improved the chance of aiming at and manifests.Better aim at and produce better mask or scale sheet, perhaps on substrate, produce better device architecture.Adopt better mask or scale sheet will produce better device layer.

Summary of the invention

The present invention relates to the aligning of writing system and workpiece.Specifically, it relates to the aligning that adopts SLM to write second layer pattern on the workpiece with ground floor pattern.It extends to makes mask or scale sheet, or adopts mask or scale sheet to make device layer.

Thereby the invention provides a kind of aligning spatial light modulator and on the ground floor pattern, writing second layer method of patterning on the workpiece, wherein, described workpiece comprises at least one coarse alignment mark, and adopt to comprise that the stand of at least one reference marker supports described workpiece, described method comprises: adopt and shine described coarse alignment mark from the illumination of spatial light modulator transmission; Adopt video camera to detect described coarse alignment mark; Detect described reference marker; And adopt the described workpiece of information calculations that the detection by described reference marker and coarse alignment mark obtains and the aligning of described spatial light modulator at least.

Described workpiece can further comprise at least one accurate alignment mark, and described method further comprises: adopt and shine described accurate alignment mark from the illumination of spatial light modulator transmission; And adopt video camera to detect described accurate alignment mark; Wherein, the calculating of the aligning of described workpiece and described spatial light modulator is further comprised the information that utilization is obtained by the detection of accurate alignment mark.Accurately the detection of alignment mark can further comprise the Luminance Distribution match with small echo and coarse alignment mark.

Description of drawings

In Fig. 1, alignment system has been carried out graphic extension.

Fig. 2 shows an Analysis of Nested Design rule, and it can be applied in the alignment mark location of the second layer;

Fig. 3 shows the details of accurate alignment mark;

Fig. 4 shows the example layout of different types on the mask;

How Fig. 5 shows by a plurality of laser pulses are carried out quadrature, and obtains the mean value that any substrate/stand moves, and creates the original image through reflection thus;

Fig. 6 has reflected the research that compares and measures the measuring accuracy that the integral dose that adopted obtains;

Fig. 7 shows the preanalysis of doing on resulting image;

Fig. 8 shows how to arrange measurement window according to the central point of calculating;

Fig. 9 shows twice measurement to the arm of targeted graphical;

In Figure 10, intensity distributions is carried out differential, so that carry out rim detection;

In Figure 11,, will have the second order polynomial and the edge peak fitting of deriving of the point of predetermined quantity in order to reach sub-pixel precision;

Figure 12 shows the strategy that calculates ground floor alignment mark position, wherein, described calculating location be vector and;

Figure 13 shows by having about 300nm FEP 171 resists, is divided into the SEM image of the grating acquisition of 3 μ m in the ranks;

Figure 14 shows the emulation setting, and it has described the 50nm resist depression on the etched pattern;

Emulation among Figure 15 shows reflection differences quartzy and that black chromium is expected under NA0.82;

Figure 16 shows a series of images and corresponding Luminance Distribution thereof under various resist thickness to Figure 19;

Figure 20 illustrates the relation between measurement of reflectivity and the emulation reflectivity;

Figure 21 shows the test result that adopts the DX1100P resist to carry out;

Figure 22 shows when measuring the position of ground floor aligning cross mark, a series of images that the same position place obtains on substrate;

Why not together, the design of Simulation that Figure 23 shows BARC emulation setting and front has;

The simulation result that adopts different B ARC type has been shown among Figure 24;

Figure 25 has repeated the calculating of emulation among Figure 24, is made as 0 emulation bleaching by the k value with BARC;

Figure 26 shows chromium and quartzy reflectivity after the DUV exposure.

Embodiment

Be elaborated below with reference to accompanying drawings.Preferred embodiment is given purpose of description be the present invention is illustrated, rather than for the scope of the present invention that defines at claims is limited.Those of ordinary skills will find various of equal value changes on the basis of following explanation.

Being used for the second layer alignment system that phase shifting mask writes is one of subfunction of Micronic Sigma 7300 mask write devices.The strategy that selected execution PSM aims at is DUV to be write laser (spatial light modulator SLM) is used in combination, to be created in the light stamp image (light stamp image) that reflects on the ground floor alignment mark with spatial light modulator.Adopt the responsive ccd video camera of DUV to capture and measure this reflected image.Employing writes laser and has avoided the position deviation that caused by the dislocation of a plurality of lasing light emitters.

(anti-reflection AR) has all reduced 248nm reflection of incident light rate anti-reflection function in chemistry amplification resist (CAR), bottom antireflective coating (BARC) and the reflection coating provided (TARC).Carry out on time at the laser that adopts 248nm, this will reduce the signal intensity and the accuracy of alignment system.Although described discussion has illustrated the observed phenomenon of employing 248nm light source, also can consider to adopt the 193nm light source.

Picture contrast when this instructions has especially been discussed the picture contrast under the different resist thickness and adopted the AR coating.In addition the algorithm of measuring ground floor alignment mark position is illustrated.

Result to the research of this point and final PSM alignment system discloses useful method and device.

1. foreword

(Phase Shift Mask PSM) has become the resolution enhance technology of widespread use, is used to overcome the contraction k1 factor for decay and alternating phase-shift mask.In the middle of more advanced PSM technology is being developed.A common ground of this technology is, makes PSM at least two write steps, between double exposure to mask develop, etching and apply new resist once more.For the resist that double exposes, the mask write device must be able to be aimed at second level exposure, thereby makes pattern and first order coupling.Usually adopt the electron beam mask write device to finish first order exposure.

For attenuating PSM, desired layer is generally hundreds of nm with layer alignment accuracy.The desired pattern degree of accuracy of the second layer (resolution, CD degree of accuracy, arrange degree of accuracy) also is non-constant width pine, therefore, its can be printed on (for example) low side i laser line generator pattern generator (pattern generator, PG) on.The desired layer of alternative P SM is much then strict with layer alignment accuracy.Requirement to the pattern degree of accuracy is also strict more, and its purpose is to avoid to produce phase error on PSM.Use for initial alternative P SM, still adopt i line laser PG usually.But, along with feature size at 90nm, 65nm and low technical node and dwindling more, the pattern degree of accuracy of i line laser PG will no longer satisfy the requirement of second layer alternative P SM or other advanced PSM technology.

At the second layer, a kind of alternative of i line laser PG be adopt with ground floor in identical electron beam mask write device, R.Plontke, L.Bettin, D.Beyer, J.Butschke, M.Irmscher, C.Koepernik, B.Leibold, A.Vix and P.Voehringer are at " Avoidance/Reduction of Charging Effects in Case of Partially InsufficientSubstrate Conductivity when using ESPACER 300Z " (20th European MaskConference on Mask Technology for Integrated Circuits and Micro-Components, GMMFB 43, pp 233-240,2004) a kind of scheme is disclosed in.If the mask write device can correctly be aimed at the second layer, will guarantee the matched well between two layer patterns.Adopt the shortcoming of electron beam system to be to produce electric charge in fluctuating place of ground floor pattern at the second layer, described electric charge will reduce the degree of accuracy of electron beam.A kind of method that proposes for fear of producing electric charge be under the second layer resist or on adopt conducting film, this is an additional processing step that possible cause more defects count and reduce fabrication yield.

Another attractive alternative of i line laser PG is the laser PG that adopts the DUV wavelength.As H.Martinsson, J.Hellgren, N.Eriksson, M.Bjuggren and T.

At " Transparent corner enhancement scheme for a DUV pattern generator " (Photomask Japan 2003, Proceedings SPIE vol.5130, pp297-308,2003) described in, the pattern degree of accuracy that the short wavelength is provided together with the pattern fidelity enhancement techniques can be fully and the ground floor of electron beam printing be complementary.Because the second layer has been adopted optical exposure, therefore can not produce charge effects, there is no need to use any conducting film.The second layer still must be aimed at ground floor, and DUV PG must have point-device alignment system.Sigma 7300DUV PG from Micronic Laser Systems adopts 248nm wavelength and the image processing techniques that is similar to 248nm photoetching scanner.It provides required resolution and pattern degree of accuracy, thereby avoids the phase defects between first and second layers, and it has point-device second layer alignment system.

Sigma7300 second layer alignment system adopts and writes the pattern generation spatial light modulator (SLM) of laser illumination [referring to WO99/45439 with 248nm DUV, be introduced into for your guidance at this], and adopt ccd video camera to observe from the image of the substrate reflection that comprises the ground floor aligned pattern.Referring to T.

, P.Askebjer, J Sellander, R.Zerne, A.Karawajczyk " PatternGeneration with SLM Imaging " (21st Annual BACUS Symposium on PhotomaskTechnology, Proceedings SPIE vol.4562, pp38-44,2001).Optical alignment difference between same laser instrument when measuring in the employing ablation process has been avoided measuring and writing.Only expose appointed area little on the mask, reduced the risk of the true pattern that influences PSM.Because image-capture is fast, even has the redundancy that a lot of robustnesss constantly increase, so the yield rate of system is very high.

Alignment system shown in Fig. 1 has shown remarkable robustness, yield rate and degree of accuracy, and the time of always aiming at, the average+3 σ results of degree of accuracy were between 13-30nm less than 6 minutes.But this paper will can not pay close attention to the performance and the strategy of total system, but how an early stage difficult problem of only paying close attention to measurement technology is handled.

Even this technology is seemingly implemented the big good approach that the second layer is aimed at, but still exist some other problemses to be worth before starting development in an all-round way, being resolved.First problem is how system responds different resist thickness.The early stage film emulation of the chemistry on the black chromium being amplified the reflectivity of resist (CAR) shows, can be expected at and obtain the polarity that changes in the reflected image.Also have a kind of worry to be, in the process of the transfer between polarity, existence is made the resist thickness of the contrast total loss in the image.First studies show that this point without any relation, together with the description to the graphical analysis position measurement technique, has illustrated that for all resist thickness, measuring accuracy all is splendid.

Second problem is whether the AR coating of arbitrary 248nm light absorbing zone will bring hell and high water to graphical analysis.A kind of method of how handling these problems is disclosed below, and actual test and emulation.

2. position measurement strategy

2.1. the second layer is aimed at general introduction

The second layer is aimed at the introduction that provides accompanying drawing described below is provided.In one embodiment, the ground floor alignment mark that is adopted comprises two coarse alignment marks, 16 accurate alignment marks and 2 reserve areas.Can easily adopt the mark of varying number.Reserve area is used for the reference background standardization and CCD aims at.The coarse alignment strategy is to improve overall performance and efficient.Accurately alignment mark is the cross figure of simple 2 μ m live widths.Define pattern position among the Sigma 7300 by the pattern position on stand position and the SLM.For the absolute position of the ccd image of the ground floor pattern that obtains reflection, coordinate system must be transferred to CCD from SLM.This point is aimed at by CCD and is realized, is about to pattern displaying on SLM, and the reservation chromium zone on substrate is with described pattern reflected back CCD.This technology makes this system need compensation hardly.Repeat coarse alignment measurement and the measurement of accurate aligning cross figure and created instantaneous surperficial map conversion, it is included in translation, convergent-divergent, quadrature and the rotation of adopting in the second layer printing process.

Fig. 2 shows an Analysis of Nested Design rule, and it can be applied in the alignment mark location of the second layer.One of ordinary skill in the art will appreciate that, can be to keeping for alignment mark or assigned region is adjusted, translation, bending or size change.There is shown four ranges of distribution 211.These regional 4mm are wide, with edge offset 6mm.They depart from turning 212a-212d15mm.These ranges of distribution can hold coarse alignment mark, accurate alignment mark and blank mark.For example, two coarse alignment marks are preferably placed at opposite side in any position that can be in the range of distribution.The coarse alignment mark can be the square of length of side 1.5mm, and every side keeps the frame of 0.5mm.In addition, can place two blank marks in any position in the range of distribution, described blank is labeled as the square of length of side 1.5mm or other sizes.Can hold 16 accurate alignment marks in the zone of 10cm, preferred per minute is joined four of area arrangements, makes the minimum interval that has 1cm between the mark, makes between the head and the tail mark to have the minimum interval.These marks can be that 1mm is square.

Fig. 3 shows the details of accurate alignment mark 310.Accurately alignment mark comprises several figures 320.In alternative configurations, can adopt 16 or 25 figures.For aligning, only need analyze a figure of each mark, therefore, other figures provide redundancy.Therefore, also can adopt the figure of other right quantities in the mark.As shown in the figure, the figure that is repeated can be a cross curve, perhaps can be grid or other regular figures at interval of supporting translation and rotating calculating.Each transposition arm of described cross curve has the length of 2 μ m live widths and 40 μ m.

Fig. 4 shows the example layout of different types on the mask 440.This figure shows two coarse alignment marks 410, two blank marks 420 and 15 accurate alignment marks 430.Described blank mark can assist CCD to aim at.A coarse alignment mark 410 comprises the level and the vertical raster pattern of superimposition.The line of described grating has constant live width, for example 1 μ m.They have since 3 μ m is the spacing (i.e. 3.0 μ m, 3.2,3.4......24.4 μ m) that step-length increases gradually with 200nm, perhaps, if measure in opposite direction, has the spacing of successively decreasing.As shown in the figure, the coarse alignment mark comprises 108 lines along each direction.Can use different live widths, initial spacing, step sizes and line number.

An aligning is placed on the appropriate location in proper order from aiming at SLM and ccd video camera with workpiece, adopts the coarse alignment mark to carry out coarse alignment.The pattern of these marks has reduced being positioned at the required search in accurate alignment mark location of other positions, range of distribution.Next step measures accurate alignment mark, thereby is provided for calculating the starting point, convergent-divergent of mask, the data of positive direct sum rotation.Rely on some or all in these counting statistics, can verify described measurement, and write second mask layer.

2.2. graphical analysis position measurement algorithm

Show whole order by Fig. 5 to the measuring sequence shown in a series of figures of Figure 13 from the primary reflection image of capturing the ground floor alignment mark to measured position.

How Fig. 5 shows by a plurality of laser pulses are carried out quadrature, and obtains the mean value that any substrate/stand moves, and creates the original image through reflection thus.Control reference background image 512 makes reflected image 511 standardization, to extract targeted graphical 513.

Fig. 6 has reflected the research that compares and measures the resulting measuring accuracy of integral dose that is adopted.Vertical pivot is a measuring accuracy.Transverse axis is the logarithmically calibrated scale of dosage.In this research, the flash of light that increases gradually by quantity changes integral dose, increases by 10 standard dose at every turn.The order of accumulation test flash of light is 2,5,10,20,40,80,160...... in the figure, 611 same focus dosage (single pass isofocal dose) corresponding to the one way process.At 612 places, the dosage that standard dose is 16 times begins to produce in resist and changes.Passing through with focus dosage in scope, to have the wide dosage range that under the condition of not destroying resist, produces good signal and measuring accuracy less than 16 times of standard doses from one way.

Fig. 7 shows the preanalysis of carrying out on resulting image 513, it discerns effective ground floor alignment mark and authentication image quality.This step has also been measured the rough position of alignment mark central point 714.

Fig. 8 shows how to place measurement window 815 according to the central point of calculating, to solve the problem of rotation sensitivity.For figure with 1 μ m live width, this centre of figure of measuring center preferred distance 2 μ m.The rotation problem that the repeatedly measurement of same figure is helped to solve figure.In the bottom windows 821 of figure, show measuring-signal 822 and 823 above threshold value.From Fig. 9, can more clearly see these measuring-signals.In this figure, show twice measurement 915a-b of the arm of figure 513.The reflective graphics of figure 513 passes resist up to ccd video camera through reflection.Along providing the edge flank for the figure arm perpendicular to the intensity distributions on the one-row pixels of direction of measurement.The number affects consequential signal of measured pixel thick line.Two signature tune line charts are provided, and the longitudinal axis is represented intensity, transverse axis remarked pixel quantity.Be regarded as the signal 922 that the noise of the signal that narrower pixel line produced 921 of 915b is produced greater than broad pixel line 915a.Note, minimizing signal 922 occurs corresponding to the trap signal 822 and 823 in the window 821 of Fig. 8 about 140 pixels and 250 pixel places greatly.

In Figure 10, intensity distributions is carried out differential 1010, so that carry out rim detection.Detect maximum external margin peak value 1011 and 1012, it makes algorithm polarity independent.That is to say that this method is equally applicable to the light black to be the black graphics of background or to be the light figure of background with black.

In Figure 11,, will have the second order polynomial and the edge peak fitting of deriving of the point 1111,1112 of predetermined quantity in order to reach sub-pixel precision.Between positive negative peak, calculate line central point 1113.Can pass through to create x and the match of y first order curve, and the point of crossing of calculating between them provides mark center point.Of this sort calculating pattern center provides extra robustness with respect to the site error that is caused by image rotating.Accurately preliminary central point 714 shown in Figure 7.

Figure 12 shows the strategy that calculates ground floor alignment mark position, wherein, described calculating location be vector and.Can be with the position of described alignment mark as position vector and calculating:

Alignment mark position (1234) on alignment mark position (1220)=stand position (1231)+half SLM size (1232)+CCD alignment offset (1233)+CCD

Adopt this vector and, will be by a plurality of quadrature image averaging stand site errors, thus degree of accuracy less than 1nm 3 is provided.The measurement of CCD alignment offset has the degree of accuracy less than 2nm 3.To shown in Figure 9, the alignment mark position on the CCD has the measuring accuracy less than 2nm 3 as Fig. 2.Described system adopts nearly, and 16 alignment marks calculate translation, rotation, orthogonality and scale transformation.This is very useful degree of accuracy.

Some improvement can separately or mutually combine and improve the degree of accuracy of position measurement, and these improvement comprise:

1. rough calculation mark center position when arranging measurement window.This makes this measurement insensitive to rotating, and this point is that pinpoint accuracy is necessary, because ccd image and substrate all may rotate.

2. seek the mark external margin and make that this measurement polarization is independent.

3. the curve fitting of the edge peak value of Dao Chuing is fit to concrete resist situation.

4. this algorithm up to determining, has guaranteed that the same position in image carries out the measurement of 16 all alignment marks just by repeated measurement stand position.Trend towards eliminating any influence that causes by distortion in the ccd image like this.

5. relatively reference background image is proofreaied and correct all images, thereby realizes the signal quality of optimization in measurement.

6. by low-down a plurality of (〉 500 of dosage) the laser pulse quadrature forms image, thereby in measurement to any substrate/stand error equalization.

3. different resist thickness cause the image polarity that changes

3.1. interference effect

Typically, writing of the second layer comprises the resist that individual layer swing (swing) is optimized, and for example FEP171 is coated on the mask.For example, referring to Product brouchure (product manual), FUJIFILMArch CO.Ltd., 15th Arai Bldg.19-20, Jinguhmae6-chome, Shibuya-Ku, Tokyo, 150-0001Japan.Purpose for analyzing applies FEP171 under the common chromium thickness of 90-100nm, provide optical path difference in the inboard and the outside of ground floor chromium pattern, and it has caused the difference under certain situation to be interfered.In the position that has chromium, destruction interference is big, and image is dark.In the position that does not have chromium, reflected light brightness is higher.The scattered light that transmits between comfortable chromium of brighter edge source and the glass.The Measurement Algorithm that is adopted 2.2 joints as mentioned in the above are described.

3.2. black chromium that in the emulation that changes resist thickness, demonstrates and quartzy reflectivity

With regard to the input of emulation, the problem that we solved is in not having the ground floor pattern of chromium resist to be carried out shaping.Figure 13 shows by having about 300nm FEP 171 resists, is divided into the SEM image of the grating acquisition of 3 μ m in the ranks.Its clearly shown resist in the position of having removed chromium about chromium thickness 1312 half dangle or recessed 1313, this is an important input of reflectivity emulation.Substrate 1311 is positioned under chromium and the resist layer.This SEM image shows the 3 μ m between-line spacing chromium gratings that cover 300nm FEP171 resist.Resist depression 1313 is positioned on the quartz areas, is the only about half of of chromium thickness, i.e. 50nm.

Emulation among Figure 15 shows and moves through different resist thickness, NA 0.82 time, and the expectation difference in reflectivity between quartz and the black chromium.The longitudinal axis is a reflectivity, and transverse axis is a resist thickness.The reflectivity 1511 of black chromium is a solid line, and the reflectivity 1512 of silicon dioxide is a dotted line.Because minimum chromium reflectivity, therefore concentrates on research near this zone of resist thickness so be measured as normal condition under the situation near chromium reflectivity minimum value corresponding to optimizing the swing point.

To feed back in the middle of the emulation from the resist depression that described SEM image is extrapolated.Figure 14 illustrates the xsect that is provided with corresponding to emulation, it has described the 50nm resist depression 1421 that is positioned on the described etched pattern.The resist 1411 of institute's emulation is the FEP171 resist.

In first step black chromium 1412 is carried out after the composition, resist has covered chromium and mask blank material 1413, for example fused quartz.Chromium has thickness 1422.

The digital parameters that is used for emulation comprises wavelength 248nm, resist index n=1.82, substrate index 1.508, chromium layer thickness 100nm (67nmCr+33nmCr ₂O ₃, reflection index is respectively 1.55-i1.53 and 1.88-i0.89).Finish calculating by the ZEMAX optical design procedure, see http://www.zemax.com. for details

Desired as us, for film, reflectivity is cyclical variation between maximal value and minimum value.Following formula has provided the distance between two adjacent minimum value (or maximal value):

In this example, Δ t=68nm.

Minimum value and peaked accurate location depend on the reflectivity and the thickness of chromium layer.By changing the chromium layer thickness, two curves move in opposite directions, because resist is thicker in described mark.To be illustrated the relation of the reflectivity between consequent black chromium shown in Figure 15 and the quartz hereinafter, thereby make it to get in touch effectively with the actual measurement generation.

The mechanical system test expends time in.Therefore, the quantity of holonomic system test is restricted.Therefore, when finishing the exploitation of second layer alignment system, the process instance that is used to test be for this system the situation of difficult that can expect.After having confirmed the emulation accuracy, select preferred resist thickness between 360nm and 400nm, be used for actual second layer coverage test.

3.3. aim at the contrast of test measures from the second layer of reality

Sigma 7300 systems finish the exploitation and have under complete functional situation, based on described emulation, how the emphasis of research will react aspect signal quality and the measuring accuracy thereby understand real system particularly near the polarity transformation point at the minimum reflectivity place between 360nm and 400nm resist.Print repeatedly the test of second layer covering performance at selected resist thickness place.Figure 16 shows a series of images and corresponding Luminance Distribution under the different resist thickness to Figure 19.The order of these images is that contrast increases gradually.The control reference background image carries out standardization to described image.There is shown intensity distributions, thereby the contrast profile is illustrated perpendicular to a line in the cross pattern.Figure 16 shows the thick FEP of 390nm 171 resist layers.This example be we the situation of difficult that can anticipate, but because the diffraction effect around the chromium edge, the quadrature intensity distributions has obtained desirable quality.Figure 17 shows the thick FEP of 385nm 171 resist layers.Quartzy reflectivity increases, and still has the good signal quality.Figure 18 shows the thick PEP of 370nm 171 resist layers.Quartzy reflectivity increases, and still has the good signal quality.Figure 19 shows the thick FEP of 375nm 171 resist layers.Difference between quartz and the chromium reflectivity reaches peak value, still has the good signal quality.In Figure 16-19, under the situation that does not influence measuring accuracy, changed figure/background polarity.This point is particularly interesting, because diffraction phenomena has produced the skeleton edge in image.

These all second layers all have along x and y direction average+the covering result of 3 σ＜25nm, and as if in the skew of contrast with finally cover and do not have any relation between the result.

3.4. the comparison between actual measurement and the simulation result

If the measurement of reflectivity level is compared with the reflectivity levels of emulation, can from Figure 20, obtain a kind of strong contact.As shown in figure 20, the longitudinal axis is a reflectivity.Transverse axis is a resist thickness.The emulation reflectivity of black chromium 2011 is a solid line, and the emulation reflectivity of silicon dioxide 2012 is a dotted line.Adopt triangle and square expression measured value, thereby show the corresponding relation between measurement and the emulation reflectivity.

Also adopt the DX1100P resist to carry out similarly test, as shown in figure 21, produced identical superperformance result.

4. the absorption layer of the AR coating between chromium and the resist

The resist that new chemistry amplifies is independent of the AR layer that is used to prevent footing effect (footing effect) together with black chromium.When writing the second layer, estimate seldom BARC and TARC AR layer and new CAR/ black chromium made up and be used in combination.

Although think unlikely use AR layer, it is studied what helps to understand is that the picture contrast that the second layer is aimed at performance and reflection ccd image is expected.An influence of researching and solving the AR coating, and how to overcome hell and high water by emulation and actual measurement.Seldom be used for second layer PSM printing although estimate the AR coating, in case use, organic BARC will be most possible AR coating.

Adopted KRF17B BARC AR coating to study, how particularly looked at the reflected image contrast, and under the situation to the surface measurements exposure, any situation has taken place in image.

The result is very interesting, even because adopt very low dose of DUV laser pulse exposure, reflected image polarity also can change immediately.Therefrom can see, under the situation of poor signal quality, also can pass through the contrast of the ground floor of exposure image raising in advance alignment mark image, and can under good position resolution, measure it.

Figure 22 shows when measuring the position of ground floor aligning cross mark, a series of images that the same position place obtains on substrate.Compare with the image shown in the part 3.3, these images reference background image without comparison carry out standardization, seem that therefore some is fuzzy.Figure 22 shows the bleaching effect of BARC.This effect can be used to prevent abominable contrast situation by implementing the pre-exposure function.These images are clearly shown that, even the AR layer seems effectively between first time exposure period, and under the effect of the exposure of 248nm DUV light, reflectivity change, it has lost the AR effect.First image of Figure 22 and last image have the good signal quality, although changed polarity, and the poor contrast of second image.Because therefore second poor signal quality that image shows have reason to believe that it also may be first image, destroyed the second layer afterwards to the accuracy in the locating tab assembly.To the further emulation of discoloration, thereby whether the image among assessment Figure 22 has logicality and meaningful.

The BARC layer in the top coating of black chromium is adopted in described further research.Figure 23 shows BARC emulation setting and why not together to be provided with the emulation shown in previous Figure 14.Do not consider final resist depression, because the purpose of emulation only is to study discoloration.Layer shown in the figure comprises the substrate 2313 such as fused quartz that is covered by chromium 2312, is BARC 2320 and resist 2311 afterwards.For to chromium carry out composition the zone, directly cover described substrate by BARC.BARC is an organic film, and there are some absorptions in it to exposure wavelength.Has the optical characteristics of BARC seldom obtained open, but in order to carry out some emulation, we adopt 4 types that obtain from following source: http://www.e-insite.net/semiconductor/index.asp? layout=article﹠amp; Articleid=CA47478.Provided optical constant (at 248nm) in the table.The reflectivity attribute of these BARC types has been shown among Figure 24.

The optical constant of four kinds of different B ARC types:

BARC type N k

BARC?A 1.74 0.33

BARC?B 1.59 0.55

BARC?C 1.48 0.41

BARC?D 2.27 0.54

The simulation result that adopts different B ARC type has been shown among Figure 24.These results show similar optimum reflectivity level, but it is under the different layer thicknesses for four kinds of BARC types shown in the above table.During the measurement of determining the alignment mark position, adopt the DUV radiation that the substrate that has applied resist is exposed.Can expect that this exposure can either cause that resist layer produces bleaching and can cause that again the BARC layer produces bleaching, promptly produces optical property and changes.Its main effect may be the absorption aspect, i.e. the variation of the imaginary part of complex index of refraction (k).Shown in following emulation, this will cause the marked change of the characteristic aspect of BARC.

We repeat the calculating of institute's emulation among Figure 24, but for emulation bleaching, and the k of BARC is made as 0.As shown in figure 25, now, radical change has taken place in the chromium reflectivity, and similar in appearance to the situation of no BARC, has only minimal value to have mobile at the effect end of the phase shift of BARC layer introducing.Figure 25 shows under DUV to before the BARC exposure, as the chromium and the quartzy reflectivity of resist thickness function.Figure 26 shows chromium and quartzy reflectivity after the DUV exposure.Emulation has indicated and observed same effect in the middle of image shown in Figure 22.

Owing to estimate that different organic BARC types can produce identical result, and the AR efficient of BARC will reduce after exposure, therefore the measure of being taked is to carry out the pre-exposure function, wherein adopts the low dosage laser pulse exposure of adjustable quantity before carrying out actual measurement exposure.Performed test has shown that also BARC bleaches the pre-exposure function and do not introduce any measuring error.

5. conclusion

Position measurement strategy: the algorithm that calculates the exact position in image is made an explanation, thereby show how edge detection algorithm is applied on this purposes.The result shows that the measuring accuracy of each alignment mark is all very high.

Different resist thickness has caused the image polarity of change: carry out emulation in ZEMAX, how to change under the influence of resist variation in thickness with the polarity of estimating image.In showing the concrete instance of best relation, this point is verified, and shown that the measurement strategies of being taked can handle different situations under the situation that does not reduce measuring accuracy.Owing under the most difficult resist thickness, carried out these tests, can show that therefore disclosed layer alignment system is effective all the time in the relevant range of resist thickness.Owing on other chemistry amplification resists except that FEP171, carried out similar test, thereby shown that described resist thickness insensitivity all is effective for all resist with similar performance.

Absorption layer in the AR coating between chromium and the resist:, lost AR validity as the KRF17BBARC layer of test process example to studies show that after the low dose of pulse exposure of adopting some 248nm light of AR absorption layer.It shows in first measurement may exist inferior signal quality, but by adopting adjustable pre-exposure function, contrast can be brought up to the good signal quality.The pre-exposure function is to resist the safety practice of first image of deterioration.

5.1. alignment tools-coarse alignment

Writing when coiling in measuring write device always will be in the face of seeking the first of pattern, for example task consuming time of the first of alignment mark V mark.This problem is by artificial or the mechanical load degree of accuracy is not high and less relatively the causing of size of the pattern wanting to find.

Basic thought is to make pattern enough big, even load degree of accuracy when being in millimeter magnitude, it can not lost yet.By design make this pattern contain expression in the whole pattern area the x position and the information of y position.Determine this pattern where the time, can carry out accurately indication residing position in pattern of single measurement.Move by adding, can extract the rotation of pattern to extra measurement.On the basis of this information, can any position positions in the pattern to be analyzed to having.This requires to know two relative positions between pattern certainly.

5.2. pattern

An embodiment of described pattern is made up of the array of V mark.It is long that described V is labeled as 4-8mm, and 70 microns wide.Along the y orientation measurement, the width of article one of described V mark is 10 microns, and the width of second (inclination) is 15 microns.

Initial separation between first mark is 80 microns.To be that step-length is linear with 1 micron increase described spacing, promptly 80,81, the 82...... quantity of adjusting the V mark makes that described pattern is square.For 4 * 4mm pattern, have 40 marks, for 8 * 8 patterns, this numerical value rises to 70.

Described V mark array is arranged in the dark territory of bright territory square (clear field square).

5.3. pattern analysis

After in pattern 11, stopping, measure flank.These flanks comprise two two continuous V marks.Because these two have different width and therefore are identified easily.After identification, calculate the spacing between the V mark, and the spacing of V mark self.(x0 is y0) to satisfy the needs of subsequent calculations for the coordinate of storage current location.

Because these measurements are subjected to the influence of contingent rotation, the rotation of therefore necessary measured pattern (j).Can calculate this rotation by on another one x coordinate, carrying out measuring the second time.In order to realize safe mobile, must at first determine to move, and not end at outside the pattern to which direction.By analyzing the measurement spacing of V mark, can get the top that has been parked in pattern center to the end or the conclusion of below.Recognize this point, can move along direction safely towards the pattern middle section.

If plate has big rotation, when moving, the x direction has the danger of losing the pattern trace so.For fear of this situation, employing one initial movable (50 microns) by a small margin provides the rough estimation to rotation.Adopt this estimation, can mobile safely 2mm, thus higher angular resolution is provided.

Utilize known rotation, the spacing that measures is adjusted to some trigonometry relations.The spacing of two V marks is that pattern " starting point " provides x compensation (Dx).Spacing between the mark is indicated its numbering (position mark) in array.This position mark provides y compensation (Dy) for pattern " starting point " again.

The compensation of being calculated is only effective in the coordinate system of pattern.By comprising pattern rotation, the pattern starting point can be transformed in the coordinate system of this stand (x_ initial point, y_ initial point).

Under situation about having, obtained the position precision of 10-20 micron up at least 5 rotations of spending.In these tests, target pattern is placed on the place apart from coarse alignment mark 3cm roughly.

5.4. further thinking

In order to make the coarse alignment pattern have generality, all devices (all ranks) all should write this pattern, can measure it again.Certainly have special other version of level, but should be target to avoid this situation.Designed pattern (coarse alignment 4 * 4 and coarse alignment 8 * 8) should satisfy these general requirements till now.They stay at least two V marks usually in the short scanning of class 60, and 10 microns width of first V mark should be differentiated by lover's rank.

Different application requirements patterns are of different sizes.When adopting loader to load GP, generally to have enough good degree of accuracy, to use 4 * 4mm pattern.When loading the monitor plate by hand, can adopt bigger 8 * 8mm pattern.By adopting the V slug of two kinds of different in width ratios, can discern automatically and these two kinds of pattern versions of individual processing.

Various embodiment

Discussion and the additional appendix aimed at about PSM are supplementary document.This explanation is the information at the further refinement of appendix formation.For example, can use the existing hardware of describing in the appendix in conjunction with method described herein and program.Estimating that result that method and program as described herein produces will have is similar to the degree of accuracy that illustrates in the annex, or obtains higher degree of accuracy.

An aspect of of the present present invention is to adopt same illumination source to carry out the aligning and the printing of pattern.At aligning, can adopt SLM or adopt extra catoptron that this illumination source is projected on the workpiece.When during aiming at, adopting the SLM projection light to shine,, in the previous application of Micronic Laser, introduced as assigning by suitable city's length direction may command exposure.

One embodiment of the present of invention are that a kind of employing spatial light modulator (SLM) writes the second layer method of patterning of aiming at the ground floor pattern on workpiece.This method comprises that with workpiece loading to the stand of pattern generator, described workpiece comprises ground floor pattern, at least one prealignment mark and at least one alignment mark.Described method further comprises: aim at SLM and ccd video camera, adopt ccd video camera to detect at least one prealignment mark, adopt ccd video camera to detect at least one alignment mark, and the employing information relevant with detected at least one alignment mark write second layer pattern on workpiece.

In described method, can adopt various marks.A kind of suitable prealignment mark comprises level and vertical gridlines, and wherein, described level and vertical gridlines have spacing that increases progressively or the spatial frequency of successively decreasing.The change of spacing or frequency aspect can be that linearity, index or logarithm change.

The background appearance is controlled in relating on the other hand of described method, and reference marker or alignment mark will contrast described background appearance to be manifested.According to this on the one hand, the employing ccd video camera is captured the image from the reference zone of workpiece, and the information correction that is used to the self-reference image is utilizing ccd video camera to detect the image of collecting in the process of described at least one alignment mark.Can utilize this information by deducting reference picture from collected image.Perhaps, can utilize it to proofread and correct smoothing process.

Except above-mentioned match and smoothing process, can adopt wavelet transformation that signal and scattered light or noise are separated such as ＂ db ＂ small echo or mexican hat wavelet.MATLAB Wavelet Toolbox User ' sGuide, Version 2, by Michel Misitis, Yves Mitsit, Georges Oppenheim andJean-Micel Poggi provides the explanation that realizes that possible small echo is used, and comprises discontinuous detection, be introduced into for your guidance at this, see http://www.mathworks.com/access/helpdesk/help/pdf_doc/wavelet/w avelet_ug.pdf for details.

Another aspect of the present invention relates to alignment image is delivered to from SLM on the section of workpiece, adopts ccd video camera to detect the alignment image of projection and the employing information aligning ccd video camera relevant with detected projection alignment image.

Before adopting ccd video camera detection alignment mark, the pre-exposure of above-mentioned any aspect with the resist that covers described at least one alignment mark can be combined.Pre-exposure to resist has changed its optical characteristics.

Above-mentioned any embodiment or aspect can be combined with the translation and the alignment parameter that calculate aligning at least.

Another embodiment of the present invention is the alignment mark that adopts in above-mentioned photoetching technique.This mark comprises level and vertical gridlines, and wherein said level and vertical gridlines have the spacing of increasing progressively or the spatial frequency of successively decreasing.The change of spacing or frequency aspect can be that linearity, index or logarithm change.

Another embodiment is a kind of aligning spatial light modulator (SLM) thereby writes second layer method of patterning on the ground floor pattern of workpiece, wherein, described workpiece comprises at least one coarse alignment mark, and supports described workpiece by the stand that comprises a reference marker at least.This embodiment comprises the illumination irradiation coarse alignment mark that employing is transmitted from SLM, and adopts video camera to detect the coarse alignment mark.It further comprises the detection reference marker, and utilizes at least the information calculations workpiece that obtains from the detection of reference marker and coarse alignment mark and the aligning of SLM.Can with in conjunction with or the mode selected, with the step application of this embodiment to coarse alignment mark and/or accurate alignment mark.Usually, the either side of the foregoing description can also be applied to this embodiment or combination with it.

Another embodiment a kind ofly makes the method for phase shifting mask or scale sheet according to above-mentioned arbitrary embodiment, and it further is included in and workpiece is carried out after the composition first time radiation-sensitive material layer exposes on workpiece.This method comprises that also the employing exposure layer carries out composition to the phase-shift structure on the workpiece.And, the either side of the foregoing description can also be applied to this embodiment or combination with it.

Another embodiment is the method for making fine structure according to embodiment formerly on second workpiece, and wherein said workpiece is mask or scale sheet.This method further comprises employing mask or scale sheet to the exposure of the radiation sensitive material layer on second workpiece, and adopts the exposure layer on second workpiece that the device architecture layer on second workpiece is carried out composition.And, the either side of the foregoing description can also be applied to this embodiment or combination with it.

Certainly, as selection, can realize the present invention by comprising the logic that is suitable for carrying out said method and the device of resource.Described device can be pattern generator, the module of pattern generator or the external unit of communicating by letter with pattern generator have been installed.Perhaps, can realize the present invention by being suitable for carrying out goods said method, that be printed on machine readable code.

List of references

1.R.Plontke，L.Bettin，D.Beyer，J.Butschke，M.Irmscher，C.Koepernik，B.Leibold，A.Vix，P.Voehringer，“Avoidance/Reduction?of?Charging?Effects?inCase?of?Partially?Insufficient?Substrate?Conductivity?when?using?ESPACER300Z”，20th?European?Mask?Conference?on?Mask?Technology?for?IntegratedCircuits?and?Micro-Components，GMM-FB?43，pp?233-240，2004.

2.H.Martinsson，J.Hellgren，N.Eriksson，M.Bjuggren，T.

，“Transparent?corner?enhancement?scheme?for?a?DUV?pattern?generator”，Photomask?Japan?2003，Proceedings?SPIE?vol.5130，pp?297-308，2003.

3.T.

，P.Askebjer，J?Sellander，R.Zerne，A.Karawajczyk，“Pattern?Generation?with?SLM?Imaging”，21st?Annual?BACUS?Symposium?onPhotomask?Technology，Proceedings?SPIE?vol.4562，pp?38-44，2001.

4.ZEMAX?Development?Corp.http://www.zemax.com

5.http://www.e-insite.net/semiconductor/index.asp？layout＝article&articleid＝CA47478

6.Product?brouchure，FUJIFILM?Arch?CO.Ltd.，15th?Arai?Bldg.19-20，Jinguhmae6-chome，Shibuya-Ku，Tokyo，150-0001?Japan.

Introduce for your guidance in full at these all references that will enumerate here.

Claims (25)

1. thereby an aligning spatial light modulator is writing second layer method of patterning on the ground floor pattern on the workpiece, wherein, described workpiece comprises at least one coarse alignment mark, and adopts and to comprise that the stand of at least one reference marker supports described workpiece, and described method comprises:

Employing is shone described coarse alignment mark from the illumination of spatial light modulator transmission;

Adopt video camera to detect described coarse alignment mark;

Detect described reference marker; And

At least adopt the described workpiece of information calculations that the detection by described reference marker and coarse alignment mark obtains and the aligning of described spatial light modulator.

2. the method for claim 1, the calculating of wherein said aligning comprises calculates translation and rotation parameter at least.

3. the method for claim 1, wherein detect at least one coarse alignment mark and comprise a plurality of coarse alignment marks of detection.

4. the detection of the method for claim 1, wherein described coarse alignment mark further comprises the Luminance Distribution match with small echo and coarse alignment mark.

5. the method for claim 1, wherein described workpiece further comprises at least one accurate alignment mark, and described method further comprises:

Employing is shone described accurate alignment mark from the illumination of spatial light modulator transmission; And

Adopt video camera to detect described accurate alignment mark;

Wherein, the calculating of the aligning of described workpiece and described spatial light modulator is further comprised the information that utilization is obtained by the detection of accurate alignment mark.

6. method as claimed in claim 5, the calculating of wherein said aligning comprise translation and the rotation parameter that calculates described aligning at least.

7. method as claimed in claim 5, wherein, the detection of described at least one accurate alignment mark comprises a plurality of accurate alignment marks of detection.

8. method as claimed in claim 5, wherein, the detection of described accurate alignment mark further comprises the Luminance Distribution match with small echo and coarse alignment mark.

9. the method for claim 1, it further comprises described spatial light modulator and described camera alignment.

10. the method for claim 1, wherein said video camera is a ccd video camera.

11. the method for claim 1, wherein described at least one coarse alignment mark comprises the level and vertical ruling of superimposition, wherein, described horizontal ruling has the spacing of increasing progressively with described vertical ruling.

12. method as claimed in claim 11, wherein said spacing increases along at least one direction in described level and the vertical direction is linear.

13. method as claimed in claim 11, wherein said spacing increases along at least one direction index in described level and the vertical direction.

14. method as claimed in claim 11, wherein said spacing increases along at least one the direction logarithm in described level and the vertical direction.

15. method as claimed in claim 11, wherein said spacing increases towards the edge from the central authorities of described coarse alignment mark.

16. method as claimed in claim 11, wherein said spacing increases towards central authorities from the edge of described coarse alignment mark.

17. the method for claim 1, it further comprises:

The image of the reference zone by the described workpiece of cameracapture;

During the position of at least one the coarse alignment mark on determining described workpiece, the image of described at least one the coarse alignment mark from the described workpiece deducts the image of described reference zone.

18. method as claimed in claim 5, it further comprises:

During the position of at least one the accurate alignment mark on determining described workpiece, the image of the reference zone by the described workpiece of cameracapture, and the image of at least one the accurate alignment mark from the described workpiece deducts the image of described reference zone.

19. the method for claim 1, it further comprises:

To be delivered to from the alignment image of described spatial light modulator on the cross section of described workpiece, and adopt ccd video camera to detect the alignment image of projection, and utilize the information that obtains by the alignment image that detects projection to aim at described ccd video camera.

20. the method for claim 1, it further comprises:

Before the employing video camera detects described coarse alignment mark, the resist that covers described at least one alignment mark is carried out pre-exposure.

21. method as claimed in claim 9, wherein, described spatial light modulator projects an alignment mark on the workpiece and is detected by described video camera.

22. method as claimed in claim 21, wherein said video camera are ccd video camera, and detect the alignment mark of described projection in the CCD zone identical with described coarse alignment mark.

23. method as claimed in claim 21, wherein, described workpiece further comprises at least one accurate alignment mark, and described method further comprises:

Employing is shone described accurate alignment mark from the illumination of spatial light modulator transmission; And

Adopt video camera to detect described accurate alignment mark;

Wherein, the calculating to the aligning of described workpiece and described spatial light modulator further comprises the information that described accurate alignment mark obtains that detects of utilizing; And

Wherein said video camera is a ccd video camera, and detects the alignment mark of described projection in the CCD zone identical with described accurate alignment mark.

24. the method for claim 1, wherein derive the Luminance Distribution of the detected image of described at least one alignment mark on the described workpiece.

25. method as claimed in claim 24 wherein, is derived first derivative of the Luminance Distribution of described detected image, and adopts described first derivative to detect described coarse alignment mark.

Images