David J. Haruch

Back to homepage

Cool Precision Semi Parts

Updated 2025-09-17

This is a collection of saved Ebay (and other sites like ebay) used parts listings for mostly semiconductor processing equipment motion stages. Annotations are speculation and are my own. All photos are obtained from public web-pages or public patent literature

Table of Contents

NIKON

NSR S202A Dry ArF Lithography Machine Reticle Stage


The distinctive cream colored long stroke part appears to be alumina ceramic (Al2O3) which has good E modulus and low density. The clear short stroke/mirror plate part appears to be a glass or glass-ceramic part

QTY 4 Voice coil motors control the X, Y, and RZ DOFs of the short stroke stage

Short-stroke position appears to be measured by laser interferometers pointing at flat glass mirrors.

There are 3 vacuum pads for clamping the reticle

The short stroke appears is constrained in Z/pitch/roll by air bearings preloaded by vacuum, but the bearings aren't visible

The long stroke is constrained in all axes other than the actuated DOF by air bearings hard-preloaded by opposing air bearings. This arrangement is very sensitive and extremely good fabrication is needed for the bearings to fly well. Japanese workmanship never dissapoints regarding the cabling

NSR S322F Lithography Machine - 2015? Vintage

Suprising amount of detail in the photos for a recent lithography machine, especially one capable of 2? nanometer overlay

The wafer chuck mover appears to be controlled by moving magnet lorentz actuators in all DOFs. Although there may be an air or magnetic gravity counterbalance mechanism for Z that is not visible. Therefore the only connection to ground vibration is the dynamic cables. See Nikon Patent US9557656B2 and US20130148094A1. This layout is expensive and requires many servo controllers, power amplifiers, etc... but likely not cost prohibitive for lithography scanner.

The two stages (alignment/measurement and scanning) appear to have a different actuators. I would assume this implies the actuators are sized exactly to meet the specific duty cycle and motion profile of each stage.

Here are a few patent images for reference

ASML

Twinscan AT1100B? Reticle Stage - 2001 Vintage

These are photos of reticle stage from what I think is an AT1100B lithography machine, which came equiped with ArF laser.

Yieldstar T200 - 2013 Vintage

Original Sources

Updated 2026-10-04 New Photos

Overlay metrology machines measure special marks made on wafers. Based on the pictures or signals from imaging the marks, the offset between special marks between layers can be measured (at sub-nanometer scale). This can be feed back into the lithography machine to correct the errors.

Balance mass video

The machines use a motion system (stage) to move the wafer with respect to the optics to take pictures of all the marks on a wafer at high throughput.

This ASML machine moves the wafer on X axis and the camera on the Y axis. This approach probably gives higher performance that a stacked XY stage layout assuming the camera is lightweight and compact.

The stages float on air bearings presumably preloaded by visible magnet arrays. Based on the porting holes the bearings appear to be an orifice or orifice/groove type. Magnetic preload allows for higher force density compared to a vacuum preloaded system. This enables stiff bearings and high servo bandwidth. The distintive mirror finished surfaces are required for air bearings to fly correctly (on a few micron film of air). There appear to be 3 horizontal bearings constraining Z, Rx, Ry DOFs and 2 vertical bearings constraining Rz and Y DOFs. Air bearings are also very clean (no particle or oil/grease droplet risks)

There is a balance mass system to eliminate optics excitation or stage crosstalk from motor reaction forces. The balance mass system also floats on air bearings. There are moving magnet trim motors and commutation encoders for the balance masses.

There are EDM cut flexures on a component perhaps for thermal relief or improved repeatability.

A pair of magnet tracks move the wafer stage along its remaining X DOF. Therefore the axis is moving coil. I would expect there to be a water cooling system to cool the motors. Visually these look like Tecnotion linear magnet tracks which have distinctive nickel plating on the back irons and black epoxy coated magnets

Overall the stage motion system appears capable of both high throughput and high repeatability due to all motion axes floating on air bearings.

Unknown Vintage, Unknown Machine ASML Reticle Stage

Unknown to what machine this reticle stage belongs to, but appears relatively old and is labeled as ASML on the web-page. Frame casting has part number "4022.480.3729" which seems to be an ASML part # format.

The long stroke appears to float on air bearings preloaded by dead mass and iron core magnetic attraction force, while the short stroke appears magnetically levitated in all DOFs with moving magnets.

Water cooled iron core motors look to be used for a long stroke. 6DOF lorentz actuators for short stroke. The Z/pitch/roll actuators are round type using 3x actuators to control 3x DOFs, perhaps on future machines 4+ coils are used for Z/P/R to avoid exciting resonances through over-actuation and improve control bandwidth.

The doubled scan direction short-stroke actuators are much larger than the single short-stroke coil for the perpendicular to scan direction as expected. These to actuators driven differentially will give Rz/Yaw control of the chuck

The moving short-stroke magnets appear to mounted with a flexure arrangement. Lorentz type motors can transfer significant heat into the magnets from the coil units (order of magnitude 1/10th to 1/100th the coil power) which makes them expand. Mounting magnets with flexures could minimize the impact of these thermal stresses on the ceramic or glass-ceramic chuck

Some aspects of ASML Patent US20060072098A1 and US20050029981A1 appear to be used on this machine

The cable loop has a balanced design and minimalistic deflectors. The cables themselves do not look like the typical white PTFE dynamic cables used in semiconductor stages, perhaps there is a particle shield for this loop to separate it from particle sensitive areas.

There appear to be a few air-bearings supporting the large granite frame. Perhaps the entire granite mass is used as a balance/reaction mass to avoid transmission of stage reaction forces to the rest of the system. The big peice of granite will not move much (large mass ratio) and is a "free mass" so saves cost.

ASML "R-CHUCK ASSY, 4022.484.85329 4022.502.70321"

E-bay listing https://www.ebay.com/itm/117167323244

This is a partial reticle chuck assembly from an unknown vintage ASML lithography machine. Listing price is $40k USD. The main body appears to be a low-expansion glass/ceramic material with prominent ribbing/lightweighting. Knowing the reticles are 6x6" gives a rough idea of the size. In general this looks like an optimized version of the older machine design

There are 6 moving-magnet motors, 3 in Z, 2 in X, 1 in Y. The 3x round Z actuators control Z, tip, tilt. The rectangular actuators control X/Y and yaw. Because this is a reticle stage which only moves fast on one axis, the actuators label "X" are relatively very large.

These actuators look more mechanically and magnetically optimized than early 2000s ASML voice coil parts I've come across. There is prominent ribbing to maximize eigen-frequencies and minimize moving mass. There is shaping of back-irons in a sine/angled pattern that is mass efficient compared to a straight wall cut seen on older stages.

The magnets appear to mount to the structural frame with flexures. Without the flextures, when magnets+back irons expand due to heat transfer from coils, the thermal stress could distort or stress the fragile structural frame.

There are golden ~250mm long pads that go to a flex PCB, perhaps these are targets for 4x cap-probes that measure Z/tip/tilt/warp

ASML Random Mirror Plate #1

This photo is from a reseller web-page, it looks like a low expansion glass ceramic "mirror plate" with a 300mm chuck that looks like SiC or SiSiC material. There is extensive ribbing/lightweighting done on the glass part to achieve high eigenfrequenices and therefore high servo bandwidth

KLA-TENCOR

Archer AIM 200 - 2007 Vintage

Source: https://caeonline.com/buy/mask-wafer-inspection/kla-tencor-archer-200-aim/9255571

KLA Archer machines are overlay metrology machines. The layout is a "C-Frame" much like a bridgeport mill but with an objective/optics/camera assembly instead of a milling spindle. Special marks made on wafers during the semiconductor manufacturing process are imaged to measure how well layers are aligned.

Stage and controller have ETEL branding. Appears to be a customized off-the-shelf stage

Stacked stage contruction using iron core linear motors and roller type linear guides. Label XYT. Stage structural material looks like aluminum. Overall nothing special and low cost (ETEL must take a healthy margin)

A high voltage electronics box and labels of "Z-PZT" likely mean there is peizo axis in the optics.

The stage and optics are mounted to large grey granite. The granite connects to steel weldment through passive isolation metal springs. There are visible magnets belonging to voice coil motors used to stop granite motion due to stage reaction forces. Compared to the 2013 ASML yieldstar machine above (with balance masses absorbing stage reaction forces) these voice coils and amplifiers likely add cost and they do not eliminate the excitation of the optics from stage reaction forces.

Kulicke and Soffa (K&S)

Beryllium Alloy Voice Coil Component

Below photo is from a "mactronix" ebay listing for a K&S wire bonder component

A copper voice coil is glued into a structural component. There is a sticker indicating the part contains Beryllium, which is a dangerous material to grind or sand. Beryllium is an expensive material, but E modulus is high and density is very low which means a servo system using this part will be very high bandwidth

There is a flex-PCB with the 2 wires needed to power the voice coil