A method for manufacturing mSAP-POFV composite of ultra-high-rise arbitrary layer interconnect HDI board
By employing graded expansion and contraction compensation and Skiving partition alignment technology, combined with an improved semi-additive method and refined process control, the problems of layer misalignment, blind via reliability, and signal integrity in ultra-high-rise HDI boards have been solved, achieving high yield and low insertion loss in HDI board manufacturing.
Patent Information
- Application Number
- CN202611057691.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-07-16
- Publication Date
- 2026-11-17
AI Technical Summary
Existing technologies for producing HDI boards with arbitrary layer interconnection in ultra-high-rise buildings have problems such as excessive layer misalignment, poor reliability of ultra-small diameter blind vias, difficulty in ensuring uniformity of fine circuit fabrication, uncontrolled POFV flatness, and difficulty in systematically managing signal insertion loss consistency.
By employing graded expansion and contraction compensation process, Skiving partition alignment technology, improved semi-additive process, differentiated electroplating and grinding schemes, low-loss material treatment, and full-process on-the-job test strip tracking and control, combined with aluminum sheet vacuum plugging and refined management and control processes, we ensure interlayer alignment accuracy, blind via reliability, circuit yield, and signal integrity.
It significantly improves interlayer alignment accuracy, ensures the electrical connection reliability of ultra-small blind vias, achieves high-yield production of 40μm fine lines, meets the requirements of high-end BGA packaging, and ensures low insertion loss characteristics under high-frequency signals.
Smart Images

Figure SMS_1 
Figure SMS_2 
Figure SMS_3
Abstract
Description
Technical Field
[0001] This invention relates to the field of printed circuit board (PCB) manufacturing technology, and specifically to an mSAP-POFV composite manufacturing method for an ultra-high layer arbitrary layer interconnect (HDI) board. Background Technology
[0002] With the explosive growth of artificial intelligence, cloud computing, and big data technologies, data center core switches are evolving towards 400G / 800G and even higher bandwidths, placing unprecedentedly stringent demands on the signal transmission rate, cabling density, and reliability of internal interconnect backplanes. To meet these requirements, the design of high-end HDI boards has to integrate multiple complex technologies.
[0003] First, to achieve ultra-high density wiring, an arbitrary layer interconnect design is required, including a large number of first-order, second-order, and even multi-order laser blind vias. For example, a 28-layer 8-order HDI board includes multi-order blind vias such as L1-L2 and L1-L3, as well as a deep blind via structure of L1-L26. As the number of layers increases, the total board thickness can reach more than 3.0 mm, while the minimum mechanical drilling diameter is as small as 0.175 mm, resulting in an extremely high aspect ratio (approximately 17:1); the laser blind via diameter is as small as 0.06 mm, and the dielectric layer thickness is only about 0.04 mm, making the manufacturing process extremely difficult.
[0004] Secondly, as line width / spacing shrinks to 40μm / 40μm or even smaller, traditional subtractive etching methods cannot meet the requirements due to severe side etching. Therefore, a modified semi-additive etching process (mSAP) is necessary to achieve fine line fabrication. The mSAP process uses an ultra-thin copper base layer (approximately 2.9μm) as the starting layer. The circuit area is thickened through pattern electroplating, followed by rapid flash etching to remove the base copper, resulting in fine lines with square cross-sections and high line width accuracy.
[0005] Furthermore, to achieve the flatness requirements and high-density wiring in the BGA area, all vias and blind vias must be resin-filled and plated with POFV caps. Simultaneously, different types of low-loss prepreg and core boards must be used on the same PCB, employing M9-grade high-speed materials to meet the insertion loss requirements of high-frequency, high-speed signals.
[0006] In the current technology for producing such "all-around" HDI boards, the various process steps are mutually restrictive, leading to a geometric increase in manufacturing difficulty. Specifically, the following problems exist:
[0007] (i) Uncontrolled expansion and contraction leading to excessive layer misalignment. Current technologies lack a systematic expansion and contraction management solution for combinations of "ultra-high-layer + multi-stage lamination + mixed lamination". During multiple lamination, laser drilling, and electroplating processes, uneven stress release within the board leads to inconsistent expansion and contraction coefficients, ultimately causing excessive layer misalignment and resulting in open circuits due to blind vias and receiving pads. This problem is particularly pronounced in mSAP processes, where the precision of fine-line alignment is extremely high.
[0008] (ii) Poor reliability of ultra-small diameter blind vias. For ultra-small blind vias with a diameter ≤0.07mm and a dielectric layer thickness of only about 0.04mm, conventional laser drilling parameters cannot guarantee a round hole shape and no residual adhesive at the bottom. At the same time, horizontal copper plating and via-filling electroplating processes are difficult to achieve uniform copper layer coverage on the hole wall, which can easily lead to fatal defects such as no copper on the hole wall, cracks, or excessive dimple.
[0009] (iii) The uniformity of fine circuit fabrication is difficult to guarantee. In the mSAP process, the uniformity of flash copper thickness, the density of dry film bonding, the accuracy of exposure alignment, and the control of flash etching amount are all interrelated. The existing process lacks fine control over the flash copper thickness (target is only 5.2-6.8μm), which can easily lead to circuit gaps, short circuits, or line width deviations.
[0010] (iv) The contradiction between POFV flatness and BGA mounting yield. The BGA area has extremely high flatness requirements (Dimple≤10μm), but the depressions and protrusions after resin plugging in the POFV process are difficult to control precisely. At the same time, there is a large difference in the target copper thickness between the BGA area and the non-BGA area (20μm for BGA area vs. 28μm for large copper surface), and existing technologies lack differentiated electroplating and polishing solutions.
[0011] (v) Lack of systematic control over signal insertion loss consistency. M9 grade low-loss materials are extremely sensitive to the process. Browning, rework, inappropriate plasma treatment or grinding can all change the material properties, causing the insertion loss index to drift. The existing process lacks a full-process "no rework" control and a 100% tracking mechanism for on-the-job test coupons.
[0012] Therefore, it is necessary to provide a new mSAP-POFV composite manufacturing method to solve the above-mentioned technical problems. Summary of the Invention
[0013] The purpose of this invention is to provide an mSAP-POFV composite manufacturing method for an ultra-high-rise arbitrary layer interconnect HDI board, which solves the problems of poor alignment, low reliability, insufficient yield of fine lines, uncontrolled flatness and signal integrity in the prior art.
[0014] The technical solution of this invention is:
[0015] A method for manufacturing an mSAP-POFV composite of an ultra-high-rise arbitrary-layer interconnect HDI board includes the following steps:
[0016] Step S1, Inner layer core board fabrication: Provide N double-sided copper-clad core boards, fabricate inner layer circuits on each core board to form an inner layer core board group, where N≥2;
[0017] Step S2, Core Layer Lamination and Through-Hole Fabrication: The N inner core boards obtained in step S1 are laminated to form a core layer. The laminated core layer is then subjected to mechanical drilling, plasma desmearing, copper plating, pulse copper plating, POFV resin plugging, resin grinding and copper reduction treatment, horizontal copper plating and vertical copper plating to complete the metallization of the through holes in the core layer and the resin plugging coverage.
[0018] Step S3, Multi-layer stacking fabrication: M stacking and circuit fabrication are performed alternately on the upper and lower surfaces of the core layer, M≥2, and a conductive layer is added on each side of the core layer for each stacking; wherein, at least one stacking layer uses a conventional semi-additive process for blind via fabrication and circuit fabrication, and at least another stacking layer uses a modified semi-additive process for ultra-small blind via fabrication and fine circuit fabrication.
[0019] Step S4, Outer Layer Fabrication: The outermost layer is laminated and drilled, and the fine circuitry of the outer layer is fabricated using a modified semi-additive process.
[0020] Step S5, subsequent processes: Solder resist, lettering, surface treatment and testing are performed sequentially;
[0021] The core layer formed by laminating the N inner core boards has 2N conductive layers inside; the M-layer lamination adds a total of 2M conductive layers on both sides of the core layer; the total number of layers of the HDI board is 2N+2M layers.
[0022] Furthermore, it also includes a graded expansion and contraction compensation process: based on the actual expansion and contraction measurement values of the plates after each pressing stage, in the subsequent stacking drilling, pattern transfer and drilling processes, fixed expansion and contraction compensation, measurement fixed expansion and contraction compensation, zone alignment compensation and automatic expansion and contraction compensation strategies are adopted to carry out differentiated compensation so that the layer deviation control target is ≤60μm.
[0023] Furthermore, in the graded expansion and contraction compensation: the core layer adopts fixed expansion and contraction compensation; the stacking layer adopts measurement fixed expansion and contraction compensation, combined with Skiving partition alignment technology, the panel is divided into multiple independent physical partitions, each partition uses an independent target and expansion and contraction coefficient; the EP outer layer adopts automatic expansion and contraction compensation, which is dynamically matched by the LDI device through real-time scanning of the target by CCD.
[0024] Skiving partition alignment technology is an alignment strategy developed to address the inconsistent expansion and contraction of ultra-high-layer HDI boards caused by multiple lamination processes. Its core implementation involves physically dividing the entire panel into multiple independent virtual partitions, each with its own independent alignment target. In practice, the system uses a high-precision CCD to capture the actual coordinates of each partition's target. After algorithmic calculation, a unique expansion and contraction compensation coefficient is independently matched to each partition. The exposure machine (such as LDI) then dynamically adjusts the scaling and offset of the circuit patterns within each partition, achieving "partition-based management." This technology effectively solves the problem of excessive layer misalignment in edge areas caused by localized expansion and contraction differences in the traditional "single compensation" mode, ensuring precise alignment of blind vias and receiving pads in each area, even on large-size panels, thus meeting the stringent alignment accuracy requirements of the mSAP process.
[0025] Further, in step S2, the POFV resin is used for pore plugging using an aluminum sheet vacuum plugging method, the plugging ink is PHP900IR-10FH, the plugging fullness is 100%, and after plugging, Dimple≤20μm and Bμmp≤10μm;
[0026] After resin grinding, the surface copper thickness was reduced from 50μm to 30μm through two copper reduction processes.
[0027] Further, in step S3, the lamination process using the modified semi-additive method includes: removing the carrier copper foil to expose the ultra-thin base copper → laser drilling → plasma desmearing → brown coating removal → copper plating → flash plating → film lamination → exposure → development → VCP electroplating → film removal → flash etching; wherein:
[0028] Removing the carrier copper foil to expose the ultrathin underlying copper involves using a specialized material called "peelable ultrathin copper foil" (or carrier copper foil). After lamination, the thicker "carrier layer" is mechanically peeled off, exposing the ultrathin copper layer used in the mSAP process. Its core principle relies on the unique "sandwich" structure of this copper foil:
[0029] Carrier layer (thick copper foil): usually 18μm or 35μm conventional copper foil, as a temporary support to prevent the ultra-thin copper foil of only a few micrometers from being damaged during lamination and transportation;
[0030] Release layer (core): Located between the carrier and the ultrathin layer, this nanoscale interface layer is the core of the technology. Through chemical bonds or physical bonding forces, it allows the two to adhere together after high-temperature pressing, yet also to be reliably torn apart.
[0031] Ultra-thin copper foil layer (functional layer): a truly conductive layer with a thickness of only 1.5~5μm, which is exactly the ultra-thin bottom copper required by the mSAP process.
[0032] The process steps to remove the carrier copper foil to expose the ultra-thin copper substrate are as follows:
[0033] Lamination: The ultra-thin copper foil layer of the copper foil is laminated with a prepreg (PP) under high temperature and high pressure to firmly adhere it to the board surface;
[0034] Peeling: After lamination, mechanical force (similar to film peeling) is used to peel off the outer carrier layer along the peeling layer interface, leaving only an ultra-thin copper layer of 1.5~5μm on the board surface.
[0035] It should be noted that the specific composition of the release layer varies depending on the technical approach. For example, a release layer containing copper organic compounds or an organic / inorganic composite release layer may be used, but the core purpose is to achieve stable and controllable mechanical release.
[0036] Therefore, the material properties determine that this operation is essentially a physical separation rather than a chemical removal, which is also an important prerequisite for the mSAP process to produce fine lines.
[0037] The copper thickness after flash plating is controlled at 5.2-6.8μm;
[0038] The film used is ADP-305 dry film. No pretreatment micro-etching is performed before film application. The film is applied directly to the board after the drying section. After exposure, the board is left to stand for 45 minutes before development.
[0039] The VCP electroplating is performed using a pulse method with a current density of 4-6 A / dm², a temperature of 22-28°C, and a deep plating capability of ≥80%.
[0040] The flash etching process is as follows: a sulfuric acid-hydrogen peroxide system is used, with a sulfuric acid concentration of 75±5 g / L, a hydrogen peroxide concentration of 30±5 g / L, an operating temperature of 35±3°C, and a processing time of 30-60 seconds; the flash etching amount is controlled at 5.75±0.25 μm, and the copper thickness of the circuit surface after flash etching is 20 μm.
[0041] Furthermore, the laser drilling process employs graded differentiated laser parameters for different medium thicknesses and hole diameters, specifically:
[0042] For stacked layers with a dielectric layer thickness S > 45 μm and a target blind hole diameter of 0.0075 mm ≤ D ≤ 0.1 mm, LDD laser parameters are used for drilling. Specifically, the copper surface is first browned, and then a thicker resin is ablated with high energy to form a blind hole with a corresponding hole diameter.
[0043] For stacked layers with a dielectric layer thickness S≤45μm and a target blind hole diameter D≤50μm, drilling is performed using MSAP laser parameters. Specifically, since 1-3μm copper foil is used, no strong pretreatment is required. UV laser or low-energy CO2 laser can be used directly with special chemicals to achieve fine drilling of 50μm or even smaller.
[0044] Furthermore, the multi-stage deposition process in step S3 also includes a differentiated copper thickness control step:
[0045] Differentiated copper reduction targets were set for BGA and non-BGA regions. Through multi-stage polishing and a secondary copper reduction process, differentiated copper thicknesses were achieved: 20μm for BGA regions and 28μm for large copper areas. Furthermore, via depth (Dimple) and via height (Bμmp) in BGA regions were ≤10μm. The specific process is as follows:
[0046] Full-board electroplating thickening: First, complete the full-board electroplating to ensure that the copper thickness of the entire board surface (including BGA and large copper surfaces) reaches or exceeds a high target (e.g., ≥28μm), laying a good foundation for subsequent partitioning;
[0047] Selective BGA area coverage protection: The entire BGA area is covered with dry film or peelable blue adhesive. This process requires extremely high alignment accuracy, and there must be no leakage at the edges;
[0048] Targeted copper reduction in non-BGA areas: This involves using etching solutions to reduce copper thickness in large, uncovered copper areas. Because the BGA areas are protected, the etching solution only acts on these areas, uniformly etching the copper thickness from over 28μm to 28μm (i.e., the higher target value).
[0049] Remove the protective film and perform a second copper reduction: After removing the BGA protective film, perform a second uniform copper reduction on the entire board.
[0050] In this invention, the secondary copper reduction employs chemical micro-etching (flash etching) instead of strong etching, resulting in extremely thin, uniform, and high-precision copper areas, monitored in real-time by an online copper thickness gauge. At this stage, the BGA area decreases from 28μm to 20μm; the large copper surface area decreases from 28μm to slightly below 28μm (meeting tolerance requirements); finally, through multi-stage grinding (polishing) and etching compensation, the BGA ultimately stops precisely at 20μm, and the large copper surface stops at 28μm, meeting the flatness requirement of Dimple / Bμmp≤10μm.
[0051] Furthermore, in step S3, at least before lamination and pressing of the outermost layer, a low-loss browning agent is used for browning treatment. The low-loss browning agent is Novabond PX-S2, with an operating temperature fluctuation of ≤±0.5℃ and a processing time of 50-60 seconds. The browning, electroplating, and pressing processes corresponding to the lamination using the low-loss browning agent are prohibited from being reworked.
[0052] Furthermore, the number of layers of the ultra-high-rise arbitrary layer interconnect HDI board is 24-36, the stacking order is 6-12, the line width / line spacing is 40μm / 40μm, and the minimum laser aperture is 0.06mm.
[0053] Further, the total finished board thickness of the HDI board is 2.0-4.0mm, the warpage is ≤0.5%, the HCT test condition is 5 cycles at 220-260°C, and the resistance change rate is ≤5%. The core to achieve a resistance change rate of ≤5% lies in the following three points:
[0054] First, ensure the reliability of hole metallization. The deep plating capacity (≥80%) is improved by pulse copper plating, and with the cooperation of POFV resin hole plugging and electroplating capping, the copper layer on the blind hole wall is uniform without cracks, which provides a basis for low resistance.
[0055] Second, the fine circuit patterning is accurate. The flash etching amount is precisely controlled at 5.75±0.25μm by using the mSAP process, so as to ensure that the circuit cross-section is square without notches, and guarantee the resistance value is within the design range.
[0056] Finally, verification is carried out through high and low temperature cycle test. The finished board is subjected to 5 HCT cycle tests at 220-260°C, and it is judged as qualified when the resistance change rate is ≤5%. This is essentially a test and verification standard for the reliability of the whole process, rather than a single process parameter.
[0057] Further, it also includes a signal integrity in-process test coupon tracking step: insertion loss test coupons and reliability test coupons are set in each batch of products, which flow 100% in-process together with the product boards and undergo the same process treatment and tests.
[0058] Compared with the prior art, the mSAP-POFV composite manufacturing method for the ultra-high-layer any-layer interconnection HDI board provided by the present invention has the following beneficial effects:
[0059] First, the mSAP-POFV composite manufacturing method for the ultra-high-layer any-layer interconnection HDI board of the present invention significantly improves the interlayer alignment accuracy. Through the hierarchical expansion and contraction compensation strategy, differentiated compensation schemes such as fixed expansion and contraction, measured fixed expansion and contraction, partition alignment, and automatic expansion and contraction are adopted respectively for the expansion and contraction differences at different build-up stages (core layer, build-up layers), so that the expansion and contraction differences of the ultra-high-layer HDI board caused by multiple laminations and build-up are controlled within ±50μm. Combined with the Skiving partition alignment technology, the open circuit caused by blind hole offset is effectively avoided, and the yield of mSAP fine circuits (40μm line width) is ensured.
[0060] II. The mSAP-POFV composite manufacturing method for ultra-high-rise arbitrary-layer interconnect HDI boards of the present invention ensures the electrical connection reliability of ultra-small blind vias. By optimizing graded laser drilling parameters (differentiated energy, pulse width, and number of passes for different hole diameters of 0.06mm, 0.07mm, and 0.08mm) and plasma adhesive removal, the problems of poor hole shape and residual adhesive at the bottom of 0.06mm ultra-small blind vias are solved. Combined with an optimized via-filling electroplating process, even ultra-small blind vias with extremely high aspect ratios can obtain a uniform and continuous copper layer on the hole wall, and finally pass the rigorous HCT high-current test (5 times, 220-260°C).
[0061] III. The mSAP-POFV composite manufacturing method for ultra-high-rise arbitrary-layer interconnect HDI boards of the present invention achieves high-yield production of 40μm fine lines. Through meticulous control of the entire mSAP process, including: controlling the flash copper plating thickness to 5.2-6.8μm (fluctuation within ±0.8μm), using ADP-305 high-resolution dry film and allowing it to stand for 45 minutes after exposure, and precisely controlling the flash etching amount to 5.75±0.25μm, the yield of 40μm linewidth / spacing reaches 93.2%, meeting the high-density interconnect requirements of AI chips.
[0062] IV. The mSAP-POFV composite manufacturing method for ultra-high-layer arbitrary-layer interconnect HDI boards of the present invention can meet the requirements of high-end BGA packaging. Targeted differentiated copper reduction and multi-level polishing processes achieve the differentiated targets of BGA area flatness better than 10μm (Dimple), BGA area surface copper 20μm and large copper surface 28μm, greatly improving the mounting yield of ultra-large size BGA chips.
[0063] V. The mSAP-POFV composite manufacturing method for ultra-high-rise arbitrary-layer interconnect HDI boards of the present invention can ensure low insertion loss characteristics under high-frequency signals. The "no rework" control throughout the entire process (rework is prohibited in browning, electroplating, and lamination processes) and the low-loss browning special formula (Novabond PX-S2) maximize the protection of the low-loss characteristics of M9 grade materials, reducing insertion loss test values by approximately 8% compared to ordinary browning processes. SI Coupon's full-process tracking and HCT reliability testing ensure that the product's signal integrity is measurable, controllable, and reproducible. Detailed Implementation
[0064] To enable those skilled in the art to better understand the technical solutions in the embodiments of the present invention, and to make the above-mentioned objectives, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be further described below.
[0065] The endpoints and any values of the ranges disclosed herein are not limited to the precise ranges or values, and these ranges or values should be understood to include values close to these ranges or values. For numerical ranges, the endpoint values of the various ranges, the endpoint values of the various ranges and individual point values, and individual point values can be combined with each other to obtain one or more new numerical ranges, which should be considered as specifically disclosed herein.
[0066] Example 1
[0067] The technical solution of this invention will be described in detail using the manufacturing of a 28-layer 8-level arbitrary layer interconnect HDI board (model 28E31781A0) as an example. This product is used as a motherboard for AI accelerators, using Taikoo EM-896K3 M9 grade material, with a total board thickness of approximately 3.0mm, a minimum linewidth of 40μm, a minimum laser aperture of 0.06mm, and a BGA pitch of 0.15mm.
[0068] The specific manufacturing steps in this embodiment are as follows:
[0069] Step S1, Inner Core Board Fabrication (IP1~IP6):
[0070] (1) Material feeding: Taiguang EM-896K3 polyphenylene oxide (PPO / PPE) system board is used, with a board thickness of 0.267mm and copper foil specification of 2 / 2oz RTF2 (electrolytic copper foil with reverse processing). The feeding size is PNL 544×620mm, vertical layout, and laser marking to establish traceability information.
[0071] (2) Film application and exposure: Pretreatment is performed before film application. The pretreatment solution formula is sulfuric acid concentration 140g / L, hydrogen peroxide concentration 108g / L, copper ion concentration ≤25g / L, operating temperature 28℃, treatment time 60-70s, and micro-etching depth controlled at 0.6-1.2μm. Dry film DI9630A is used, UVmark alignment is performed, and expansion and contraction compensation is fixed.
[0072] (3) Etching (DES): Minimum line width after etching is 200μm, and minimum line spacing is 100μm. AOI detection shows a median line width of 203μm and a median line spacing of 97μm.
[0073] (4) Browning: A special formula for high-speed, high-frequency M9 materials is used. The bonding force is enhanced through nano-level micro-etching and chemical bonding. The temperature fluctuation is ≤±0.5℃, the processing time is 50-60 seconds, and rework is prohibited. The median thickness of the inner core board is 0.267mm, and the median copper thickness is 70μm (corresponding to 2oz copper foil).
[0074] Step S2, Core lamination and via fabrication (CT1):
[0075] (1) Pressing and expansion / contraction management: Six inner core boards are stacked using the Pin-Lam system, fixed with 12 hot-melt points and 4 rivets. The rivet specifications are 3.175×3.5×5.5mm. Pressing parameters: heating rate 1.5-2.5°C / min, high-temperature curing temperature 205-210°C, high-temperature curing time ≥60 minutes, low-pressure stage pressure 100-150psi, high-pressure stage pressure 300-450psi, cooling rate ≤2.5°C / min, total pressing cycle 160-210 minutes. After pressing, the edges are trimmed to PNL size 541×618mm. Copper is reduced by 5μm, with parameters of sulfuric acid 130-150g / L, hydrogen peroxide 100-110g / L, tank temperature 35±2°C, and processing time 80-100 seconds.
[0076] (2) Drilling target in separate piles: Automated equipment operation is required, manual operation is strictly prohibited. The piles are divided into ±50μm piles, and the layer offset is controlled at 60μm.
[0077] (3) Mechanical drilling: minimum drill bit 0.175mm, maximum aspect ratio 11.03:1. Use segmented drilling and drill through in three sections. The lifespan of 0.175mm and 0.2mm drill bits is controlled based on 300 holes. 1 sheet / stack.
[0078] (4) Plasma desmearing and copper flash plating: The plasma desmearing gas comprises 70-80% CF4 and 20-30% O2 by volume, with a working pressure of 0.3-0.6 mbar, RF power of 5-8 kW, a processing time of 25 minutes, and a temperature of 80-120°C. A dedicated mSAP line is specified. The copper flash plating working temperature is 29-34°C, the deposition rate is 15-25 microinches / minute, formaldehyde concentration is 14-22 mL / L, pH is 12.0-12.8, Cu²⁺ concentration is 1.8-2.5 g / L, and the loading ratio is 1-3 dm² / L. The line speed is 1.5 m / min, and the jet frequency is 30 Hz. The copper thickness of the flash plating is 3.5 μm.
[0079] (5) Pulse copper plating: The electroplating solution has a copper sulfate concentration of 18-22 g / L (Cu²⁺ approximately 20 g / L), a sulfuric acid concentration of 190 g / L, a chloride ion concentration of 50 mg / L, an SPS brightener of 0.08 g / L, and a temperature of 23℃. The minimum copper hole diameter is 20 μm, the average copper hole diameter is 25 μm, and the minimum copper thickness is 42 μm. The total number of mechanical holes is 216,120 per PNL.
[0080] (6) POFV resin plugging: plugging of through holes on the TOP and BOT sides, 139,833 holes / PNL. Vacuum plugging with aluminum sheet, ink PHP900IR-10FH. 100% saturation, dimple ≤ 20 μm, bmple ≤ 10 μm. Baking at 155°C for 120 minutes.
[0081] (7) Resin grinding and copper reduction: belt speed 490-510 rpm, conveyor speed 2-3 m / min, grinding current 7.1-8.4 A, copper loss 3 μm. First copper reduction: sulfuric acid 90-110 g / L, hydrogen peroxide 8-12 g / L, temperature 30℃±2°C, processing time 40-50 seconds, copper reduction 8 μm, surface copper target 37 μm. Second copper reduction: processing time 15-25 seconds, copper reduction 4 μm, surface copper target 30 μm. 100% scanning inspection: Dimple≤20 μm, Bμmp≤10 μm.
[0082] (8) Horizontal and Vertical Copper Plating: Horizontal copper plating uses a line speed of 2.5 m / min, copper ions of 1.8-2.2 g / L, sodium hydroxide of 8-12 g / L, formaldehyde of 14-22 mL / L, a temperature of 30-33°C, a treatment time of 5-6 minutes, a deposition rate of 0.55-0.75 μm, and a load of 1.2-2.8 dm² / L. Vertical copper plating uses a stepped or pulsed current, with a copper plating thickness of 32 μm and a median copper thickness of 60 μm. The minimum copper thickness of the POFV finished product is 25 μm; the current density for vertical copper plating is 2.0-3.5 ASD, using a stepped or pulsed current, with an initial 0.8-1.5 ASD impact plating, a main plating stage of 2.0-3.5 ASD, and a final current reduction for leveling.
[0083] (9) Outer layer pattern transfer: Measurement of fixed expansion and contraction plus CCD automatic alignment compensation, alignment error ≤ ±3μm, first piece of each batch is measured. Etching factor ≥ 3.0, line tolerance ± 8μm, minimum hole ring ≥ 25.4μm.
[0084] Step S3, Layer creation:
[0085] (a) BU02 (first-second layer)
[0086] (1) Browning before lamination: Browning solution formulation: inner layer bonding agent 25-40 mL / L, concentrated sulfuric acid 55-60 mL / L, hydrogen peroxide 35-50 mL / L; operating temperature 30-45°C, treatment time 50-60 seconds. The browned film withstands ≥6 thermal shock cycles (288±5°C×10 seconds) and has a peel strength ≥0.625 N / mm. The median thickness of the laminated plate is 2.16 mm, and the PNL size at the cut edge is 538.6×617 mm.
[0087] (2) Laser pretreatment: 0.33 / 0.33oz copper foil, SI2-VSP-AM3R, copper thickness 8-9μm after browning before laser treatment (copper foil about 11.5μm).
[0088] (3) Laser drilling (LDD process): PP specification EM-896BK3 1035 RC75%×1, median dielectric thickness 0.055mm, median laser aperture 0.08mm. Alignment method is Skiving A target alignment, single aperture. The copper thickness at the bottom of the laser finishing layer is 32μm.
[0089] (4) Plasma degumming: CF4 70-80% and O2 20-30% (volume percentage), working pressure 0.3-0.6mbar, RF power 5-8kW, processing time 35 minutes, temperature 80-120°C. De-browning line speed 2m / min.
[0090] (5) Electroless copper plating: Electroless copper plating involves depositing a thin layer of conductive copper (3.5 μm) on the hole wall to achieve hole metallization. Operating temperature: 29-34℃; deposition rate: 15-25 microinches / minute; formaldehyde (reducing agent): 14-22 mL / L; Cu²⁺ concentration: 1.8-2.5 g / L; pH: 12.0-12.8; loading ratio: 1-3 dm² / L; line speed: 0.8-1.5 m / min.
[0091] (6) Vertical hole filling, electroplating parameters: copper sulfate concentration 18-22g / L, sulfuric acid 180-200g / L, chloride ion 40-60mg / L, SPS 0.05-0.1g / L, temperature 20-25℃. Electroplated copper thickness 21μm, surface copper 37μm, laser hole diameter ≤10μm.
[0092] (7) Copper reduction: sulfuric acid 100±25g / L, hydrogen peroxide 10±8g / L, temperature 25-40°C, treatment time 50-60 seconds, copper reduction 17μm, surface copper target 20μm.
[0093] (8) Circuit fabrication (40μm linewidth): Dry film RD-1225, exposure RD1225-FQ-SkivingB4-AS, zone measurement to fix expansion and contraction. Minimum linewidth after etching: 40μm, minimum line spacing: 40μm. Median AOI linewidth: 43μm, median line spacing: 37μm.
[0094] (2) BU03-BU07 (3rd-7th level stack)
[0095] (1) Lamination: The median plate thickness increases from approximately 2.306 mm for BU03 to approximately 2.786 mm for BU07, and the copper foil is 0.0833 oz (approximately 2.9 μm).
[0096] (2) Remove the load copper foil: expose the ultra-thin bottom copper (about 2.9μm).
[0097] (3) Laser drilling (MSAP process): PP specification EM-896BK3 1027 RC75%×1, median medium thickness 0.041-0.043mm. Laser hole diameter on side A: 0.06mm, laser hole diameter on side B: 0.07mm. Skiving A target alignment, single aperture.
[0098] (4) Plasma degumming: CF4 70-80% and O2 20-30% (volume percentage), working pressure 0.3-0.6mbar, RF power 5-8kW, processing time 35 minutes, temperature 80-120°C.
[0099] (5) Debrowning: sulfuric acid 70-100g / L, hydrogen peroxide 20-40g / L, temperature 30±5°C, no brushing, linear speed 6.0m / min, micro-etching amount 0.4±0.1μm.
[0100] (6) Electroless copper flash plating: A thin layer of conductive copper (2.5 μm) is deposited on the hole wall to achieve hole metallization. Operating temperature: 29-34℃, deposition rate: 15-25 microinches / min, formaldehyde: 14-22 mL / L, Cu²⁺ concentration: 1.8-2.5 g / L, pH: 12.0-12.8, loading: 1-3 dm² / L, line speed: 0.8-1.5 m / min. The flash-plated copper thickness is 2.5 μm, and the surface copper thickness is 5.2-6.8 μm.
[0101] (7) mSAP film application and exposure: Pretreatment micro-etching target 0.4-0.6μm, operating temperature 30±5°C, processing time 30-60 seconds. Key operation: Apply film directly without pretreatment, manually place the board after the drying section. Dry film ADP-305. Exposure laser wavelength 355nm, alignment accuracy ±5μm, exposure energy 40-70mJ / cm², spot size 15-25μm, Skiving zone measurement with fixed compensation, 9-point measurement on one side. After exposure, let stand for 45 minutes before development.
[0102] (8) mSAP plating (VCP): pulse mode, current density 4-6A / dm², temperature 22-28°C, TP value ≥80%. Copper plating thickness 22μm, surface copper 28μm.
[0103] (9) Film removal: The median thickness of the board is about 2.314-2.794 mm.
[0104] (10) Flash Etching: A sulfuric acid-hydrogen peroxide system is used to achieve an anisotropic etching effect, aiming to quickly remove the seed layer copper (bottom copper) while preserving the fine line shape to the maximum extent. Main parameters: sulfuric acid 75±5g / L, hydrogen peroxide 30±5g / L, temperature 35±3°C, processing time 30-60 seconds. Key control: flash etching depth 5.75±0.25μm. 20μm of copper remains after flash etching.
[0105] (11) Verification: Dimple ≤ 10 μm, median line width 43 μm, median line spacing 37 μm. HCT test 5 times, peak temperature 220-260°C.
[0106] (III) BU08 (8th layer)
[0107] (1) Low-loss browning: Novabond PX-S2 solution is used, with temperature fluctuation ≤ ±0.5℃ and treatment time 50-60 seconds.
[0108] (2) The remaining processes are the same as BU03-BU07, and the total plate thickness is increased to about 2.93mm.
[0109] Step S4, Outer Layer Fabrication (EP):
[0110] The median thickness of the laminated plate is 3.026 mm. Drill targets are clustered at ±50 μm intervals. The minimum drill bit size for mechanical drilling is 0.4 mm, and the maximum aspect ratio is 7.5:1. After removing the load copper foil, laser-etching is performed on side A to 0.06 mm and on side B to 0.07 mm. The mSAP process is the same as BU03-BU07. The total number of laser-etched holes is 936,954 per PNL on side A and 427,185 per PNL on side B. HCT testing is performed 5 times at 220-260°C. AOI does not allow additional lines; the template with a cross mark is Slp-20 μm.
[0111] Step S5, Solder Mask and Characters (BE):
[0112] Pre-treatment for solder resist without abrasion or brushing; micro-etching depth 0.4-0.6μm. Solder resist ink SR-FK1020G green clear varnish, 20μm thickness. DI solder resist exposure, zoned alignment. BGA pitch 0.15mm, NSMD & SMD method, solder resist PAD tolerance ±10μm. Secondary solder resist resist frame thickness 15-30μm, width 100μm±30μm. Character white clear varnish IJR-4000LW100, including customer QR code and factory traceability QR code.
[0113] Step 6, Surface treatment and testing:
[0114] Immersed nickel-palladium-gold plating: Nickel thickness ≥3μm, palladium thickness ≥0.05μm, gold thickness 0.05-0.15μm. Full-board reflow soldering (lead-free, six-temperature zone): 170°C, 190°C, 210°C, 230°C, 245°C, 255°C; 200-220°C for 35s, 220-245°C for 65s. Impedance testing with an Agilent instrument; insertion loss coupon retained. Two-wire, four-wire, and high-voltage testing. Warpage: Bowing ≤0.88mm, twist ≤1.05mm, warpage ≤0.5%. Optical solder joints aligned to pads, copper to edge distance 0.06mm. All coupons retained and sent to FQA.
[0115] The 28-layer 8-stage HDI board manufactured according to the above method was tested, and all indicators met the design requirements. The test results are shown in Tables 1-3:
[0116] Table 1: Results of Interlayer Alignment Accuracy and POFV Flatness Testing
[0117]
[0118] Table 2: Process Capability of Key Processes in mSAP Fine Linearization
[0119]
[0120] Table 3: Overall Yield and Reliability of Final Products
[0121]
[0122] The above test results show that, in terms of interlayer alignment accuracy, the maximum layer offset of the core layer stacking drill target is 58μm, which is better than the control target of 60μm; after POFV resin plugging, the maximum dimple is 18μm and the maximum bμmp is 8μm, which meet the requirements of ≤20μm and ≤10μm respectively. Regarding mSAP fine circuit fabrication, the copper thickness after flash plating is 5.4-6.6μm, the flash etching amount is 5.55-5.95μm, and the median finished linewidth is 42.5-43.5μm, all stably controlled within the target range. The yield of 40μm linewidth / spacing fabrication reaches 93.2%. In terms of signal integrity, the temperature fluctuation of the BU08 order low-loss browning process is controlled within ±0.4°C, and the insertion loss test value is reduced by approximately 8% compared to the ordinary browning process. In terms of reliability, the resistance change rate after HCT testing (5 times, 220-260°C) is ≤4.2%, and there are no hole wall cracks or delamination defects; the warpage of the finished product is 0.42%, which is better than the requirement of 0.5%. Therefore, this invention can effectively solve the problems of alignment, reliability and signal integrity in the manufacturing of HDI boards with arbitrary layer interconnection in ultra-high-rise buildings.
[0123] The manufacturing method of this invention is applicable to the mass production of 24-36 layer HDI boards for use in data centers, cloud computing, and AI accelerators.
[0124] The embodiments of the present invention have been described in detail above, but the present invention is not limited to the described embodiments. For those skilled in the art, various changes, modifications, substitutions, and variations made to these embodiments without departing from the principles and spirit of the present invention still fall within the protection scope of the present invention.
Claims
1. A method for manufacturing an mSAP-POFV composite of an ultra-high-rise arbitrary-layer interconnect HDI board, characterized in that, Includes the following steps: Step S1, Inner layer core board fabrication: Provide N double-sided copper-clad core boards, fabricate inner layer circuits on each core board to form an inner layer core board group, where N≥2; Step S2, Core Layer Lamination and Through-Hole Fabrication: The N inner core boards obtained in step S1 are laminated to form a core layer. The laminated core layer is then subjected to mechanical drilling, plasma desmearing, copper plating, pulse copper plating, POFV resin plugging, resin grinding and copper reduction treatment, horizontal copper plating and vertical copper plating to complete the metallization of the through holes in the core layer and the resin plugging coverage. Step S3, Multi-layer stacking fabrication: M stacking and circuit fabrication are performed alternately on the upper and lower surfaces of the core layer, M≥2, and a conductive layer is added on each side of the core layer for each stacking; wherein, at least one stacking layer uses a conventional semi-additive process for blind via fabrication and circuit fabrication, and at least another stacking layer uses a modified semi-additive process for ultra-small blind via fabrication and fine circuit fabrication. Step S4, Outer Layer Fabrication: The outermost layer is laminated and drilled, and the fine circuitry of the outer layer is fabricated using a modified semi-additive process. Step S5, subsequent processes: Solder resist, lettering, surface treatment and testing are performed sequentially; The core layer formed by laminating the N inner core boards has 2N conductive layers inside; the M-layer lamination adds a total of 2M conductive layers on both sides of the core layer; the total number of layers of the HDI board is 2N+2M layers.
2. The mSAP-POFV composite manufacturing method for ultra-high-rise arbitrary-layer interconnect HDI boards according to claim 1, characterized in that, It also includes a graded expansion and contraction compensation process: based on the actual expansion and contraction measurement values of the plates after each pressing stage, in the subsequent stacking drilling, pattern transfer and drilling processes, fixed expansion and contraction compensation, measurement fixed expansion and contraction compensation, zone alignment compensation and automatic expansion and contraction compensation strategies are adopted to carry out differentiated compensation so that the layer deviation control target is ≤60μm.
3. The mSAP-POFV composite manufacturing method for ultra-high-rise arbitrary-layer interconnect HDI boards according to claim 2, characterized in that, In the graded expansion and contraction compensation: the core layer adopts fixed expansion and contraction compensation; the stacking layer adopts measurement fixed expansion and contraction compensation, combined with Skiving partition alignment technology, the panel is divided into multiple independent physical partitions, each partition uses an independent target and expansion and contraction coefficient; the EP outer layer adopts automatic expansion and contraction compensation, which is dynamically matched by the LDI device through real-time CCD scanning of the target.
4. The mSAP-POFV composite manufacturing method for ultra-high-rise arbitrary layer interconnect HDI boards according to claim 1, characterized in that, In step S2, the POFV resin is used to plug the holes using an aluminum sheet vacuum plugging method. The plugging ink is PHP900IR-10FH, the plugging fullness is 100%, and the diameter is ≤20μm and the diameter is ≤10μm after plugging. After resin grinding, the surface copper thickness was reduced from 50μm to 30μm through two copper reduction processes.
5. The mSAP-POFV composite manufacturing method for ultra-high-rise arbitrary-layer interconnect HDI boards according to claim 1, characterized in that, In step S3, the deposition process using the modified semi-additive method includes: removing the carrier copper foil to expose the ultra-thin base copper → laser drilling → plasma desmearing → brown coating removal → copper plating → flash plating → film lamination → exposure → development → VCP electroplating → film stripping → flash etching; wherein: The copper thickness after flash plating is controlled at 5.2-6.8μm; The film used is ADP-305 dry film. No pretreatment micro-etching is performed before film application. The film is applied directly to the board after the drying section. After exposure, the board is left to stand for 45 minutes before development. The VCP electroplating is performed using a pulse method with a current density of 4-6 A / dm², a temperature of 22-28°C, and a deep plating capability of ≥80%. The flash etching process is as follows: a sulfuric acid-hydrogen peroxide system is used, with a sulfuric acid concentration of 75±5 g / L, a hydrogen peroxide concentration of 30±5 g / L, an operating temperature of 35±3°C, and a processing time of 30-60 seconds; the flash etching amount is controlled at 5.75±0.25 μm, and the copper thickness of the circuit surface after flash etching is 20 μm.
6. The mSAP-POFV composite manufacturing method for ultra-high-rise arbitrary-layer interconnect HDI boards according to claim 5, characterized in that, The laser drilling process employs graded differentiated laser parameters for different medium thicknesses and hole diameters, specifically: For stacked layers with a dielectric layer thickness S > 45 μm and a target blind hole diameter of 0.0075 mm ≤ D ≤ 0.1 mm, LDD laser parameters are used for drilling. For stacked layers with a dielectric layer thickness S≤45μm and a target blind hole diameter D≤50μm, drilling is performed using MSAP laser parameters.
7. The mSAP-POFV composite manufacturing method for ultra-high-rise arbitrary-layer interconnect HDI boards according to claim 1, characterized in that, Step S3, the multi-stage deposition process, also includes a differentiated copper thickness control step: Differentiated copper reduction targets were set for BGA and non-BGA regions. Through multi-stage grinding and secondary copper reduction processes, a differentiated copper thickness of 20μm was achieved for the BGA region and 28μm for the large copper surface region. Furthermore, the via depth (Dimple) and via height (Bμmp) in the BGA region were ≤10μm and ≤10μm, respectively.
8. The mSAP-POFV composite manufacturing method for ultra-high-rise arbitrary layer interconnect HDI boards according to claim 1, characterized in that, In step S3, at least the laminations near the outer layer are subjected to browning treatment using a low-loss browning agent. The low-loss browning agent is Novabond PX-S2, with an operating temperature fluctuation of ≤±0.5℃ and a treatment time of 50-60 seconds. The browning, electroplating, and lamination processes corresponding to the laminations using the low-loss browning agent are prohibited from being reworked.
9. The mSAP-POFV composite manufacturing method for ultra-high-rise arbitrary layer interconnect HDI boards according to claim 1, characterized in that, The ultra-high-rise arbitrary layer interconnect HDI board has 24-36 layers, a stacking order of 6-12, a linewidth / line spacing of 40μm / 40μm, and a minimum laser aperture of 0.06mm.
10. The mSAP-POFV composite manufacturing method for ultra-high-rise arbitrary layer interconnect HDI boards according to claim 1, characterized in that, The finished HDI board has a total thickness of 2.0-4.0 mm, a warpage of ≤0.5%, and HCT test conditions of 5 cycles at 220-260°C with a resistance change rate of ≤5%.
