Technical Information

Transcription

1 CHEMCUT Techncal Informaton CORPORATION Introducton The Chemcut CC8000 etcher has many new features desgned to reduce the cost of manufacturng and, just as mportantly, the cost of ownershp. Keepng the prce of a new etcher to a mnmum and makng t as trouble free and easy to fx as possble whle runnng s all well and good but t stll has to delver good etch qualty n terms of lne wdth unformty, especally when lne wdths and spaces fall below 100 µm (4 mls). An nexpensve and easy to run etcher s stll a bad nvestment f t doesn t etch well. One of the major changes n the Chemcut CC8000 etcher s a new spray system that s desgned to gve the maxmum amount of spray coverage and etch unformty wth the least amount of soluton sprayed on the surface of the panel. Reducng the number of nozzles reduces the amount of puddlng on the panel surface, especally the top sde, and ncreases etched lne wdth unformty. The Chemcut Sgma seres etchers have gpm nozzles sprayng 63 gallons/ mnute on each sde of the panel to cover a thrty nch conveyor wdth whle the Chemcut CC8000 etcher has gpm nozzles sprayng 48 gallons/ mnute per sde to gve the same coverage wth better unformty and no loss n conveyor speed. Ths bulletn gves the results of etch tests done to demonstrate the etch capablty of the Chemcut CC8000 etcher for the standard conveyor, thn and flexble substrate conveyor and for half-etch (reducng 17 µm [ ½ oz.] fol to 9 µm [ ¼ oz.]or 5µm [ 1 / 8 oz.] fol). Etch Test Cautons It s mportant to note that the etched lne wdth varaton found after etchng measures the varaton of the entre process, not just the etcher. In addton to the varaton ntroduced by the actual etchng step, the lne wdth varaton found after etchng also ncludes varatons due to the type of fol and substrate used, cleanng, type of resst used, photo-tool manufacture and preparaton, exposure of the etch resst and developng. All of these processes must be optmzed n order to get a truer pcture of the etcher capablty. If all of the pror process steps have been optmzed and run properly then approxmately 70 to 80% of the lne wdth varaton found after etchng s due to the etchng step. If they are not done properly then only 30 to 40% of the varaton may be due to the etch step, makng t dffcult to access the etcher performance properly. The etcher, by tself, cannot produce results that are better than what goes nto the etcher n the frst place. It s safe to assume for the etch test results shown here that the pre-etch processes have been optmzed to the best of our abltes. Etch Test Background Evaluaton of etch tests can vary accordng to the test method used so some explanaton of test methods s necessary. Followng s a bref summary of the method Chemcut uses to test etch capablty. The man concern n etch qualty s the unformty of etched lne wdths over the entre panel surface. In order to evaluate the qualty of etch results properly one needs enough data to get a relable statstcal base from whch to make comparsons from one test to another. Ths requres more than just nne or ten readngs per panel. It s desrable to have at least ffty data ponts on the surface and more s better. Dong

2 ths manually through an optcal mcroscope s tme consumng so many companes use automatc optcal scanners to measure up to several hundred thousand ponts on the panel surface n a short amount of tme. Chemcut also uses automatc optcal scannng as part of ts etch evaluaton but we have found that optcal scannng has resoluton lmts and other ssues that may obscure subtle but mportant dfferences n etch qualty. In 1985 Chemcut started usng an etch test pattern that allowed us to measure the electrcal resstance of an etched lne and, from the known length of the lne and the thckness of the copper, calculate the lne wdth. By etchng many of these small test patterns n an array across the panel surface t was possble to get a relable, repeatable etched lne wdth map across the entre panel n a relatvely short tme. Ths allowed us to quckly see and dentfy the effects of any changes n the etcher desgn. In the early 1990 s a test consortum, formed by several leadng manufacturers to mprove etch qualty, ndependently developed a smlar test pattern and method of calculatng lne wdths. A company called Conductor Analyss Technologes (CATs) was formed to do the testng and analyss for the consortum and was soon offerng ther test and analyss as a pad servce to the ndustry. Snce then the CATs test and test patterns have become the de facto ndustry standard for etched lne wdth and etch capablty. Chemcut started usng the CATs test pattern n 1995 so we had a common comparson to the ndustry and have used CAT to ndependently verfy our test results. Ths pattern requres very precse etchng to acheve good results and s very senstve to any changes n the etch process. For those who are not famlar wth CATs tests the followng s a bref descrpton of the test modules and test patterns Chemcut uses n our testng. If you are already famlar wth the CATs method you may want to skp drectly to the test results. CATs Test Module A typcal CATs test module s llustrated below: CATs mult-ptch Test Module The CATs test module s mult-ptched wth four dfferent lne wdths and three dfferent spaces between lnes. They are offered n several dfferent lne wdths rangng from 1 ml (25 µm) to 4 ml (100µm) lnes wth 1 ml to 3 ml spaces up to 8 ml (200 µm) to 11 ml (275 µm) lnes wth 8 ml to 10 ml spaces. Chemcut uses the pattern startng wth a 3 ml (75 µm) lne and space to a 6 ml (150 µm) lne wth 1 oz. fol. The module s approxmately 1 nch (25mm) square and each of the lne lengths s about 25 nches (63.5cm). The two nnermost pads connect to the 3 ml lne, the next two pads out connect to the 4 ml lnes, etc. Page2of 13 12/28/2006

3 CATs Test Pattern The test pattern s for an 18 x 24 (457 mm x 610 mm) panel. The test modules are arranged n a 16 x 22 grd (352 total modules) wth every other module turned 90 o n a checkerboard pattern as dagrammed below: P O N M K J I H G F E D C B A P O N M K J I H G F E D C B A - Vertcal Module - Horzontal Module The lower rght corner of a typcal test panel ready for etchng s llustrated below: Secton of Ready-to-Etch Test Panel Page3of 13 12/28/2006

4 Once the panel has been etched and strpped a resstance meter s used to measured the resstance of each lne and the ndvdual lne wdths, average lne wdth and standard devaton are calculated. Evaluatng Etch Test Results The etched lne wdth results represent 75 µm (3 ml) lnes wth 75 µm spaces etched n 35 µm (1 oz.) fol. The data are n mls and are presented n tables and as 3-D surface charts. Included n the tables and charts wll be nformaton on average lne wdths, standard devaton, process capablty potentals and hgh and low measurements. The average lne wdth and hgh and low values are self-explanatory but the standard devaton and process capablty potental (Cp) may need some explanaton n order to make sense of the results. The Standard Devaton s a common statstcal calculaton to express the amount of varaton found n a group of numbers. The smaller the standard devaton the less the varaton. When usng the CATs method for evaluatng etch results Chemcut uses the followng gudelnes: Standard Devaton between and Satsfactory etchng, capable of 100 µm lnes and spaces n volume. Standard Devaton between and Good etchng, capable of 75 µm lnes and spaces n volume. Standard Devaton less than Excellent etchng, capable of any lne and wdth combnaton down to the lmts of the process. The Process Capablty Potental, Cp, measures the potental of the process to meet specfcatons and s calculated by dvdng the range between the upper and lower set lmts of the lne wdth varaton by sx tmes the standard devaton. The upper and lower set lmts for these tests s ±20% of the average etched lne wdth. A process wth a Cp of 1.33 or greater s consdered to be a fully vable process that can meet and mantan specfcatons relably. To put the Cp s of the etch test results n perspectve, the IPC and CATs recently released ther PCQR 2 (Process Capablty, Qualty, and Relatve Relablty) Industry Statstcs for 1-ounce Innerlayer Conductors and Spaces. The Cp s for the 3 ml (75 µm) lnes of the 15 manufacturers who submtted panels usng the same CATs test pattern as used n our tests and the same ±20% lne wdth varaton ranged from a low of 0.33 to a hgh of 1.27 wth a medan value of 0.91 The 3-D surface chart s smply the average lne wdth of the 75 µm lnes n each test module plotted aganst ts poston on the panel. The followng sectons contan the test results for the Chemcut CC8000 etcher for the standard conveyor, thn and flexble materal conveyor, and half-etch results. Page4of 13 12/28/2006

5 Etch Results wth Standard Conveyor The standard conveyor s desgned to transport rgd panels down to a substrate thckness smlar that of 75 µm (3 ml) fberglass wth mnmum shadowng by the conveyor wheels of the lower sde of the panels. The test results are for 75µm (3 ml) lnes and spaces on 457 mm x 610 mm (18 x 24 ) panels wth 35 µm (1 oz.) fol and a 125 µm (5 ml) fberglass core. Ave. W Standard Devaton Cp Hgh ow Top Sde 2.12 mls mls 1.93 mls Bottom Sde 1.98 mls mls 1.82 mls 3-D surface charts are on the next page. Each color represents a dfference of 2.5 µm (0.1 mls) n the average lne wdth. Page5of 13 12/28/2006

6 Standard Conveyor Etch Results 3-D Surface Chart Top Sde Surface Plot - Chemcut 2002 Top ne Wdth vs Panel Poston 3 ml (75 µm) nes on 18" x 24" panel (457 mm x 610 mm) n e W d t h Standard Devaton = Average ne Wdth = 2.12 mls (53.85 µm) Max. ne Wdth = 2.35 Mls (59.69 µm) Mn. ne Wdth = 1.93 mls (49.02 µm) Column eadng Edge p n l j b d f h Row D Surface Chart Bottom Sde Surface Plot - Chemcut 2002 Bottom ne Wdth vs Panel Poston 3 ml (75 µm) nes on 18" x 24" panel (457 mm x 610 mm) n e W d t h Standard Devaton = Average ne Wdth = 1.98 mls (50.29 µm) Max. ne Wdth = 2.14 Mls (54.36 µm) Mn. ne Wdth = 1.82 mls (46.23 µm) Column p n l j h f b d Row eadng Edge Page6of 13 12/28/2006

7 CHEMCUT Techncal Informaton Etch Results wth Thn and Flexble Materal Conveyor As substrate materals have become thnner and more flexble, transportng and etchng these substrates n horzontal conveyorzed etchers wthout damage and wthout resortng to leaders s a major problem. Increasng the densty of the wheels solves the transport problem but causes major nterference wth etch speed and qualty. Chemcut has developed a Thn Materal Transport (TMT) conveyor that wll transport the most flexble substrates wthout leaders wth mnmal loss of conveyor speed and excellent etch qualty. Descrpton Chemcut s conveyor rods are spaced 1 5/8 nches (41 mm) apart. By usng 2 nch (50 mm) dameter conveyor wheels, slghtly offset from rod to rod, the wheels overlap each other from rod to rod, makng t mpossble for the leadng edge of a flexble panel to slde down between them to the bottom of the etcher. The problem has been to provde enough support between wheels on a rod wthout nterferng wth the etchng. The heart of the TMT desgn s the patented S wheel. The rm of the S wheel snakes back and forth around the crcumference of the wheel n an S pattern rather than a straght lne. The wdth of the pattern s 1 nch (25.4 mm) so that the most support wth the least maskng of the lower spray s offered. However, t s dffcult to get the S wheels close enough together to provde relable transport of very flexble materals wthout nterferng wth the wheels on the next rod. Ths problem s resolved by alternatng the S wheel on every rod wth two soft wheels wth a straght, rather broad, rm on each conveyor as n the photo below. Transport Tests Intal tests wth water n the spray modules has shown that ths system s capable of transportng a wde varety of thn and flexble crcut board materals wthout leaders and wthout skewng or damage to the boards. Materals tested ncluded 2 ml (50 µm) fberglass, 3.5 ml (90 µm) Teflon, and 1 ml (25 µm) Kapton wth both 1 oz. (35 µm) and ½ oz. (17 µm) unetched copper, wth a typcal sgnal/sgnal pattern etched on the panel, and wth no copper left on the surface at all. Even under the worst case condtons of the lower pressure 15 ps (1 bar) hgher than the top and no upper dler wheels to hold the materal down, Page7of 13 12/28/2006

8 all the materal transported through the spray modules wth no jammng or skewng, ncludng the 1 ml Kapton wth no copper. There were no mpressons from the wheels n any of the copper clad materals. These tests showed that ths conveyor desgn would run most of the ndustry s problem materals wthout leaders or damage but, as you can see from the photograph, the densty of the wheels has to have some effect on the etchng from the bottom sde. A seres of etch tests was run to fnd out what prce, n terms of etch unformty and speed, was pad n order to transport the materal through the etcher. Thn Materal Transport Etch Tests The test results are agan for 75 µm lnes and spaces but the panels for ths test measured 457 mm x 305 mm (18 x 12 ) wth 35 µm fol and a 75 µm Teflon core. Chemcut s latest etcher wth standard conveyor wheels was used for the etchng. Once the speed and pressures were set, one panel was etched and strpped wth the standard wheels n place. The standard wheels were then removed and replaced wth the TMT wheels. The conveyor speed and upper and lower spray pressures were adjusted so the etched lne wdths matched those for the standard conveyor and one more panel was etched and strpped. The lne wdths on the panels were then measured and compared usng the CATs method. The panels were then optcally scanned usng an AEI TOMM optcal scanner. Etch Test Results The comparson of the conveyor speeds and spray pressures for the two sets of conveyor wheels are shown below. The denser wheel pattern of the TMT conveyor dd cause enough shadowng to slow the conveyor speed from 25 nches per mnute (0.64 M/mn.) to 22 pm (0.56 M/mn.) n order to mantan the same lne wdths. Ths s an acceptable trade-off to ensure relable transport of thn materal. Standard Conveyor TMT Conveyor Conveyor Speed 25 nches/mn. 22 nches/mn. Upper Spray Pressure 36 ps 25 ps ower Spray Pressure 25 ps 33 ps The same shadowng also caused a change n the pressure balance between the upper and lower spray tubes. Wth the standard conveyor the upper spray pressure was 11 ps hgher than the bottom to get even etchng on both sdes of the panel. Wth the TMT conveyor the bottom pressure had to be run 8 ps hgher than the top to acheve the same results. Whle ths may seem alarmng at frst glance n terms of transport capablty, our testng has shown that the panel stays on track wth bottom pressures as hgh as 15 ps greater than the top wth no upper hold down wheels. In fact, 3.5 ml Teflon cores wth ½ oz. fol showed no sgns of wanderng durng the etch process. The weght of the cuprc chlorde soluton on the top of the panel s more than enough to hold the panel down on the conveyor untl the bottom pressure s at least 15 ps hgher on the bottom. The table below shows the etch results on the bottom sdes for the 3 ml lnes for the two sets of panels run usng the CATs measurement method. The topsde results for both conveyors were n the to Page8of 13 12/28/2006

9 0.096 range for standard devaton and ddn t change snce no changes were made that affected the topsde etchng. Etch Results for 3 ml nes Bottom Sde Standard Conveyor Ave. W Standard Devaton TMT Conveyor Cp Ave. W Standard Devaton Bottom Sde 3 ml Teflon, 35 µm Fol 2.36 mls mls Whle the TMT conveyor caused some ncrease n the standard devaton of the panels run they stll remaned well wthn the excellent etchng range. Cp 3-D surface charts are on the next page. Each color represents a dfference of 2.5 µm (0.1 mls) n the average lne wdth. Etch maps from the optcal scanner are on the page followng the 3-D scans. Each color here represents a dfference of 1.25 µm (0.05 mls) n the average lne wdth. Page9of 13 12/28/2006

10 Thn and Flexble Materal Conveyor Etch Results Bottom Sde Results - Standard Conveyor n e W d t h Standard Devaton = Ave. = 2.36 mls, Max. = 2.58 mls, Mn. = 2.21 mls Column P N J H Row Bottom Sde Results - Thn Materal Transport Conveyor n e W d t h Standard Devaton = Ave. = 2.35 mls, Max. = 2.55 mls, Mn. = 2.14 mls Column P N J H Row Page10of 13 12/28/2006

11 Thn and Flexble Materal Conveyor Etch Results Optcal Scan Etch Maps Page11of 13 12/28/2006

12 Half Etch Results As lne wdth requrements approach 50 µm (2 mls) and even 25 µm (1 ml) t would be desrable to etch them n thnner fols than 17 µm ( ½ oz.). 17 µm fol s readly avalable n quantty but 9 µm ( ¼ oz.) fol s only avalable n lmted quanttes and at a hgh prce. One way around ths s to etch 17 µm fol down to 9 µm to take advantage of the prce and avalablty of 17µm fol. Test panels were 457 mm x 610 mm. The copper thckness was measured before and after half etch usng x-ray fluorescence at 25 mm ntervals over the panel surface. Fol before Half Etch After Half Etch Thckness µm 9.30 µm Standard Devaton Maxmum Thckness µm µm Mnmum Thckness µm 7.98 µm 3-D surface charts of the fol thckness data before and after half etch are shown on the next page. Each color represents a thckness dfference of 1µm. Page12of 13 12/28/2006

13 Half Etch Test Results 3-D Surface Chart 17 µm fol thckness as receved Surface Plot of Fol Thckness 1/2 oz. (17 m) Fol as Receved 18 n. x 24 n. (457 mm x 610 mm) Panel Average = mcrons Standard Devaton = Max = mcrons Mn = mcrons D Surface Chart 9 µm fol after etchng down from 17 µm Surface Plot of Fol Thckness 1/4 oz. (9 m) Fol as Etched Down from 1/2 oz. (17 m) Fol 18 n. x 24 n. (457 mm x 610 mm) Panel Average = 9.30 mcrons Standard Devaton = Max = mcrons Mn = 7.98 mcrons Page13of 13 12/28/2006