上个月，一家做电动工具齿轮箱壳体的客户寄了两只样条过来。
都是玻纤增强尼龙，都标着 GF30，一只我们出的，一只他从别处买的。
同一副模、同一台机、同一套参数打出来。
标签上就三个同样的字符。
测试报告摆在桌上，两只的拉伸强度差了将近两成。
他在电话里问得很直接："同样都是 GF30，凭什么差这么多？"
这个问题的答案，八成不在玻纤含量上，而在玻纤和尼龙之间那一层——尼龙偶联剂管的就是这一层。
我回了三个问题。
"玻纤含量你们复测过没有？"
"断口是齐口的，还是发毛的？"
"这两只样在打之前，料烘了没有？"
他停了一下，说含量复测过、两只确实一样；断口他没看过；料是上午开的包，下午就上机了。
三个问题里已经有一个指向了答案，剩下两个，这篇后面会一个个说清。
一、同样的 GF30，差的不是玻纤含量
先把方向定住：玻纤含量的差异，是最容易查、也最容易被背锅的一个变量。
因为它一测就知道，而且便宜。
灼烧称重，二十分钟出结果。含量一样，很多人的排查就到此为止，然后开始怀疑"料不稳定"。
但玻纤含量一样，件里的玻纤状态可以完全是两回事。
它有四个变量：玻纤有多长、在件里怎么排、含量分布匀不匀、和尼龙结合得好不好。
前三个上一篇文章讲过，这一篇只讲第四个——界面结合。
如果把玻纤想成混凝土里的钢筋，界面就是钢筋外面的那层螺纹。 没有螺纹，钢筋再粗也拽不住混凝土。
这一层有多薄？几十纳米到几百纳米，比一根头发丝的千分之一还细。
但它决定了外力是从尼龙"传进"玻纤，还是让玻纤整根被抽出来。
这就是为什么两只 GF30 可以差两成。
二、偶联剂在两分钟里干了什么
玻纤是无机的东西，尼龙的分子链是有机的东西。
这两类东西天生不亲。
玻纤表面带的是硅羟基，尼龙链上带的是酰胺基团，两者之间没有抓手，只能靠物理缠绕勉强贴在一起。
单靠这点接触力，受力时会发生什么很清楚：玻纤从尼龙里被整根拔出来。
这就是界面失效，它的断口是发毛的，不是齐口的。
尼龙偶联剂要解决的，正是这个"握手"的问题。
拿最常见的硅烷类来说，它的分子长得像一根两头都带胶的短绳：
一头水解成硅醇，能和玻纤表面的羟基缩合，形成硅氧键。
另一头是氨基或环氧基，和尼龙链上的基团有亲和力。
一根玻纤，就靠无数根这样的短绳，被钉在尼龙基体里。
这里有一个容易被忽略的量级。
玻纤的单丝直径，常见在十微米上下。
一公斤玻纤摊开来的表面积，是几十到上百平方米——相当于把一间小房间的四面墙，塞进一公斤沙子里。
所以界面的工作量，比很多人想象的大得多；偶联剂的用量看着只有千分之几，铺开的却是一层看不见的"网"。
一句大白话总结机理：偶联剂就是两头带胶的短绳，一头拴在玻纤上，一头拴在尼龙上。绳子够多够牢，力才传得过去。
为什么太短的玻纤没用、玻纤在螺杆里怎么一路被打碎，上一篇讲玻纤长度时已经说过。
这一篇补的是另一半：长度保住了，界面没搭好，强度照样上不去。
而且界面这一级的失效，比玻纤长度更难看出来——显微下玻纤还是那根玻纤，问题出在它旁边那几十纳米的壳层里。
三、几种主流偶联剂，各是什么脾气
下面这张表是品类层面的常识区间，不是配方。区间是"按玻纤或填料计"，不是按总配方计，这一点后面会再说。
| 品种 | 主要功能 | 适配体系 | 公开添加区间（按玻纤/填料计） | 脾气 |
|---|
| KH-550（氨基硅烷） | 与玻纤、矿物表面缩合，同时与尼龙端基亲和 | 玻纤增强、矿物填充的 PA6 / PA66 | 0.5%–1.0% | 反应活性高，氨基偏碱性；耐温余量有限 |
| KH-560（环氧基硅烷） | 与玻纤表面缩合，环氧基与端基反应 | 玻纤增强 PA、部分 PA 合金 | 0.3%–0.8% | 与偏酸体系相容性更好，反应温和 |
| 钛酸酯 / 铝酸酯类 | 改善无机填料与树脂的浸润 | 高填充矿物体系、部分阻燃体系 | 0.5%–1.5% | 对水分敏感，储存要防潮 |
| 马来酸酐接枝聚合物（相容剂） | 与尼龙端氨基反应，改两相界面 | PA / PP、PA / ABS 等合金 | 3%–8%（视合金比例） | 给的是树脂与树脂之间的界面，不是树脂与玻纤 |
关于这张表，有三句话要说清。
其一，区间怎么读。 表里的百分数是"相对玻纤或填料"的用量，不是"相对一吨料"。
把 1% 直接按一吨料算成十公斤，加出来往往是不够的；反过来按玻纤量算完再乘一遍总重，又会超。
这一格读错，是这类助剂最常见的浪费来源。
其二，区间来自公开资料的常见量级，具体牌号以 TDS 为准。 不同厂家有效含量、载体、粒径都不一样，拿别人的区间直接套，只能当起点。
其三，最后一行不是偶联剂。 它放在这张表里，正是为了让人一眼看到两者的位置不一样。
偶联剂是小分子，用量千分之几，管的是"无机与有机"的界面。
相容剂是高分子，用量百分之几到十几，管的是"树脂与树脂"的界面。
这两句读懂了，一半的误买就省掉了。
四、界面怎么选：需求、验证与配伍
这张表是全篇最该收藏的一页。它的用法是先在一列里找到你现在卡住的那一项，再横着往右读。
| 需求（你现在卡在哪） | 该往哪类走 | 怎么验 | 常见失效 | 与哪类助剂会打架 |
|---|
| 干态拉伸 / 弯曲要提上去 | 氨基硅烷这一类（KH-550 方向） | 干态拉伸 ISO 527 对比；灼烧后看玻纤残段表面是否带树脂 | 界面没搭上，玻纤整根拔出 | 外润滑剂、脱模剂过量会隔断界面 |
| 湿热循环后强度保留 | 环氧基硅烷这一类，或复合处理的玻纤 | 85℃ / 85% 湿度或热水浸泡后复测拉伸保留率 | 界面水解，掉得比干态快 | 与酸性阻燃组分、部分铜盐体系配伍要单独确认 |
| 件表面发白（浮纤） | 先调界面浸润，再调纤维分散 | 目视 + 断面观察 | 玻纤露头、表面粗糙 | 外润滑比例偏高时更明显 |
| 高填充矿物体系还想保韧性 | 钛酸酯 / 铝酸酯这一类 | 灰分复测 + 冲击对比 | 填料抱团，断口发白 | 与未干燥的填料一起用等于白加 |
| 两相合金的界面（PA / PP、PA / ABS） | 马来酸酐接枝相容剂（不是偶联剂） | 相形态观察 + 冲击、熔指对比 | 两相脱开、分层、力学波动 | 与偶联剂同时加会重复占位，先想清楚要哪一层 |
这张表最该记住的是最后一行。
遇到"两个树脂相合不到一起"，那是相容剂的活。
遇到"一根无机纤维和树脂拉不上手"，才是偶联剂的活。
混着买、混着加，是这一行最常见的浪费。
再补一句验证上的提醒：界面这件事，样条数据和件上数据经常对不上。
样条是规则流动、单一厚度；件上有熔接线、有尖角、有厚薄交界。
界面的问题，往往先在件上暴露，再回到样条上找证据。
所以验证的顺序建议是：样条看方向，件上看结论。
五、四种从助剂侧出的失效
界面相关的失效，很多被归到"料"头上。下面四条，全都从助剂侧解释。
失效其一：同一批件，强度忽高忽低。
根因常常不是"料不稳定"，是偶联剂分散不均。
混料段的时间、剪切、投料位置，任何一项松了，都会让一部分粒子上偶联剂偏多、另一部分偏少。
通行解法：先查混料工艺与母粒化，别急着换料。
失效其二：干态数据漂亮，湿热循环之后就垮。
这一条最典型：偶联剂只做了"一半"。
硅烷这一头的硅醇已经和玻纤缩合上了，另一头的有机基团却没能和尼龙形成足够稳的结合。
干态下靠物理缠绕还能撑住，一进水，界面上那层硅氧键开始水解，强度就往下掉。
通行解法：按件的湿热工况选品种，并在湿热老化条件下复测，不看干态就定方案。
失效其三：件表面发白，角落和流道末端最重。
先查模温当然没问题，但如果一批一批不一样，先看助剂。
外润滑剂或脱模剂过量，会在玻纤和树脂之间形成一层隔离，界面被盖住，玻纤露头。
这条值得说得直白一点：浮纤不全是模温的事。同一副模打出两个结果的时候，方向大概率在料里，不在机器上。
通行解法：把外润滑比例压下来，用内润滑把流动补回去；同时看偶联剂有没有混匀。
失效其四：温度一提，界面反而更差。
硅烷类偶联剂有它的耐温余量。
挤出机高温段一路顶到上限，一部分偶联剂在到达玻纤表面之前就已经反应掉或分解掉了。
加了，但没起作用——这种情况从物性表上是看不出来的。
通行解法：确认助剂的耐温窗口，把高温段控制在窗口以内；必要时调整玻纤的加入位置。
上面四条里，有两条（分散不均、外润滑过量）不需要换料、不需要换助剂，只靠把工艺参数收一收就能解决。
这就是为什么排查要先从免费的动作开始。
六、加工与添加：先做什么，后做什么
这一段讲顺序。顺序错了，加进去的东西会互相消耗。
其一，先干燥。
矿物填料吸潮，玻纤表面也会带水。
带着水进挤出机，偶联剂会先去和水反应——它本来该去和玻纤表面结合的。
所以先干燥，再谈偶联。 这一条在前面讲"什么时候白加"时已经埋过一次，这里落成动作。
其二，再做预分散。
偶联剂直接一勺倒进主喂料口，是最省事、也最不匀的做法。
通行做法是先做成母粒，或者先和玻纤 / 填料做预处理，让它有机会均匀铺到无机表面上去。
其三，玻纤侧喂。
玻纤走侧喂，减少它在螺杆前半段的滞留和剪切。
这一条和界面是连着的：玻纤在螺杆里待得越久，被剪得越短，可用于承载的比表面积和有效长度都往下走。
其四，把温度分档看。
不是"整体高温"或"整体低温"，而是看偶联剂加入段与玻纤加入段的实际温度。
这两段的温度，比料筒设定值更能说明问题。
其五，真空段脱挥别省。
低分子残留留在料里，后面会以气味、析出、表面发乌的形式还回来。
顺序总结成一句话：先把水赶走，再把助剂铺匀，再让玻纤尽量晚进场，最后把挥发物抽掉。
七、什么时候不该加，加多了会怎样
这一段可能比前面几段更省事。
先说过量。这里有一句要写死：过量即失效。
多余的偶联剂在界面上形成的不是更多的"绳"，而是一层低分子的弱层。
界面从"钉住"变成"垫着"，强度和耐热反而往下走。
更麻烦的是，多余的部分会往件表面迁移，后续喷漆、印刷、超声焊接的附着力会跟着出问题。
加得越多越保险，是这类助剂最容易踩进去的一个想法。
再说四种不该加的情况。
其一，体系里没有无机表面。
纯树脂、没有玻纤、没有矿物填料——偶联剂没有可以拴的地方，加了只是多一份成本和一个残留。
其二，玻纤出厂已经带了匹配的浸润剂。
常规短切玻纤的浸润剂里本来就有偶联成分，而且是玻纤厂按自己的表面处理工艺做的。
这一条要先问清再决定加不加，不是"再加一遍更稳"。重复加、加错品种，都可能把原有的浸润体系搅乱。
其三，要解决的是"树脂与树脂"的界面。
两个树脂相合不到一起，那是相容剂的活。
偶联剂去处理合金界面，属于方向错了，剂量补不回来。
其四，件要做食品接触或医疗相关的认证。
助剂是化学品，这类场合必须回到 GB 4806.7 / ISO 10993 的语境里谈。
认证针对的是具体牌号，助剂也要一并报进去，不能等认证卡住才想起来加过什么。
把这四条写在前面，不是劝退，是省一次试模。
八、这一层界面值多少钱
先给一个总量级：助剂总成本占改性塑料吨成本的大致区间是 1%–5%。
看着不多，但缺了它们，性能往往过不了测试——这一句才是这条线的重点。
换成一个好懂的说法：一吨料里加 0.5% 的偶联剂，就是五公斤。
五公斤是什么概念？一袋料二十五公斤，它占五分之一袋。
成本上它是零头，界面上它是全部。
助剂的单价一律以 2026 年参考价、随行情波动 为准，具体按当期报价。
配方里的助剂体系按件的工况配——常规助剂常备现货，特殊型号按需配套；你报工况和牌号，料和助剂一次配齐。
读者常问的三句
问：偶联剂和相容剂，能不能只买一个？
不能互相替代。一个管无机与有机的界面，一个管树脂与树脂的界面。
先确定你卡住的是哪一层，再决定买哪个。 两个都买也行，但要清楚各自在做什么。
问：我们的玻纤买来就带浸润剂，还要不要再加？
先要参数，再决定。
问清玻纤的浸润剂体系是什么、有没有针对尼龙做匹配。
如果原来的界面本来就做得可以，再加一遍不是稳，是乱。
问：加了偶联剂，干态数据好看了，是不是就能过客户那边的认证？
这条路线的方向是改善界面，行不行要看件的验证结果。
认证针对具体牌号与具体助剂体系，替代要重新验证哪几项，得按客户的标准逐条对。
结语
回到开篇那两只样条。
同一副模、同一台机、同一个 GF30，差的两成，差在玻纤和尼龙之间那几十纳米。
界面这件事，判断链其实很短：
先问有没有无机表面，再问是哪一层界面，最后问验证条件对不对得上件的工况。
先接水、再铺匀、后进场——顺序对了，剩下的交给测试。
你手上要是有玻纤增强件正卡在强度或湿热保留上，把三样东西发来就能给个方向：玻纤含量与品种、件的湿热工况、现在卡住的那一项指标。
前两天接了个电话，问的就是界面这一层的事。
做改性尼龙这些年，我们更愿意先问清件在什么工况下用——干态卡住和湿热之后掉下来，是两条不同的路。
玻纤增强件的界面与试模，可以一起聊。
Last month, a customer who makes gear housings for power tools sent over two samples.
Both are fiberglass reinforced nylon, both labeled GF30, one we produced, the other he bought from elsewhere.
Produced using the same mold, the same machine, and the same set of parameters.
There are only three identical characters on the label.
The test report is on the table; the tensile strength of the two differs by nearly 20%.
He asked very directly on the phone: 'They're both GF30, so why is there such a big difference?'
The answer to this question is eighty percent likely not in the fiberglass content, but in the layer between the fiberglass and the nylon—that’s the layer that the nylon coupling agent mainly acts on.
I answered three questions.
Have you re-tested the glass fiber content?
Is the fracture smooth, or frayed?
Before these two are striking, has the material been dried?
He paused, saying that the content was retested, and the two are indeed the same; he hasn't looked at the fracture; the material was from a bag opened in the morning and used on the machine in the afternoon.
One of the three questions already points to the answer, and the remaining two will be explained one by one later in this article.
1. The same GF30, the difference is not in the glass fiber content
First, fix the direction: the difference in fiberglass content is the easiest variable to check, and also the easiest to be blamed for.
Because you can know immediately with just one test, and it's cheap.
Burning and weighing, results come out in twenty minutes. If the content is the same, many people's investigation stops here, and then they start to suspect 'the material is unstable'.
But even if the glass fiber content is the same, the state of the glass fiber in the part can be completely different.
It has four variables: how long the fiberglass is, how it is arranged in the part, whether the content distribution is uniform, and how well it binds with the nylon.
The previous three articles covered the first three; this one only discusses the fourth — interface integration.
If you think of fiberglass as the steel bars in concrete, the interface is like the threaded layer on the outside of the steel bars. Without the threads, no matter how thick the steel bars are, they won't hold the concrete.
How thin is this layer? It is tens to hundreds of nanometers, thinner than one thousandth of a human hair.
But it determines whether the external force is 'transmitted' from the nylon to the fiberglass, or if the fiberglass is pulled out as a whole.
This is why two GF30s can differ by 20%.
2. What the coupling agent did in two minutes
Fiberglass is an inorganic material, and the molecular chains of nylon are organic materials.
These two kinds of things are naturally incompatible.
The surface of the glass fiber carries silanol groups, and the nylon chain carries amide groups. There are no grasping points between them, so they can only barely stick together through physical entanglement.
Relying solely on this small amount of contact force, it is clear what will happen under stress: the fiberglass will be pulled out of the nylon in one piece.
This is interface failure; its fracture is frayed, not smooth.
The problem that nylon coupling agents need to solve is precisely this 'handshake' issue.
Take the most common silanes, for example, their molecules look like a short rope with adhesive on both ends:
One end hydrolyzes into silanol, which can condense with the hydroxyl groups on the surface of glass fibers to form silicon-oxygen bonds.
The other end is an amino or epoxy group, which has an affinity for the groups on the nylon chain.
A single glass fiber relies on countless short cords like this, nailed into the nylon matrix.
There is a magnitude here that is easily overlooked.
The diameter of a glass fiber filament is commonly around ten micrometers.
The surface area of one kilogram of fiberglass when spread out is tens to over a hundred square meters—equivalent to cramming the four walls of a small room into one kilogram of sand.
So the workload of the interface is much greater than many people imagine; the amount of coupling agent seems to be only a few thousandths, but it spreads out into an invisible 'network'.
A straightforward way to summarize the mechanism: a coupling agent is like a short rope with glue on both ends, one end tied to the glass fiber, the other end tied to the nylon. Only if there are enough ropes and they are strong enough can the force be transmitted.
Why are fibers that are too short useless, and how are fibers being smashed all the way through in the screw? This was already mentioned in the previous article when talking about fiber length.
This part makes up the other half: the length is maintained, the interface isn’t properly connected, and the strength still can’t increase.
Moreover, failure at the interface level is even harder to detect than the length of the glass fiber—the glass fiber still appears the same under a microscope, but the problem lies in the tens-of-nanometers-thick shell layer next to it.
3. Several mainstream coupling agents, what are their characteristics
The table below shows the common range at the category level, not the formula. The range is 'calculated by glass fiber or filler,' not by the total formula, which will be explained later.
| Variety | Main Function | Adaptation System | Publicly add interval (calculated by glass fiber/filler) | Temper |
|---|
| KH-550 (Amino Silane) | Condenses on the surfaces of glass fiber and minerals, while also being compatible with nylon end groups | Glass fiber reinforced, mineral-filled PA6 / PA66 | 0.5%–1.0% | High reactivity, amino group is slightly basic; limited temperature tolerance |
| KH-560 (epoxy silane) | Condense with the glass fiber surface, epoxy groups react with terminal groups | Glass fiber reinforced PA, some PA alloys | 0.3%–0.8% | Better compatibility with slightly acidic systems, mild reaction |
| Titanate / Aluminate | Improve the wetting of inorganic fillers and resin | High-filler mineral systems, partially flame-retardant systems | 0.5%–1.5% | Sensitive to moisture; storage should be kept dry |
| Maleic anhydride grafted polymer (compatibilizer) | React with nylon terminal amino groups to modify the two-phase interface | PA/PP, PA/ABS and other alloys | 3%–8% (depending on the alloy ratio) | It refers to the interface between resin and resin, not resin and glass fiber. |
There are three things to clarify about this table.
Firstly, how to read the range. The percentages in the table are the amounts 'relative to glass fiber or filler,' not 'relative to one ton of material.'
Calculating 1% directly as ten kilograms per ton of material often results in too little; conversely, if you calculate based on the amount of fiberglass and then multiply by the total weight, it will exceed.
Misreading this cell is the most common source of waste for this type of additive.
Secondly, the range comes from commonly available data in public sources, and the specific grade should be based on the TDS. The effective content, carrier, and particle size vary between different manufacturers, so using someone else's range directly can only serve as a starting point.
Thirdly, the last row is not a coupling agent. It is placed in this table precisely to let people see at a glance that the positions of the two are different.
Coupling agents are small molecules, used in amounts of a few thousandths, and they manage the interface between 'inorganic and organic'.
Compatibilizers are polymers, used in amounts ranging from a few percent to more than ten percent, and they manage the interface between 'resin and resin'.
If you understand these two sentences, half of the mistaken purchases can be avoided.
4. How to Choose an Interface: Requirements, Verification, and Compatibility
This table is the most collectible page in the whole text. Its usage is to first find the item you are currently stuck on in one column, and then read across to the right.
| Requirement (Where are you stuck now) | Which direction should I go? | How to verify | Common Failures | Which types of additives will conflict |
|---|
| Dry stretching/bending needs to be raised | Amino silanes of this type (KH-550 direction) | Comparison of dry-state tensile ISO 527; after ashing, check whether the surface of the glass fiber fragments has resin | The interface didn't connect, and the fiberglass was pulled out entirely. | Excess external lubricants and release agents can block the interface |
| Strength retention after wet-heat cycling | Epoxy silanes of this type, or composite-treated glass fibers | Retest tensile retention after 85℃ / 85% humidity or hot water soaking | Surface hydrolysis, falls off faster than in the dry state | Compatibility with acidic flame retardant components and certain copper salt systems needs to be confirmed separately. |
| The surface of the item appears white (floating fibers) | First adjust interface infiltration, then adjust fiber dispersion | Visual cross-section observation | Fiberglass exposure, rough surface | More pronounced when the external lubrication ratio is relatively high |
| High-filled mineral systems still want to maintain toughness | Titanate / aluminate this category | Ash Re-measurement Impact Comparison | Filler agglomeration, fracture appears white | Use with undried filler is equivalent to adding white |
| Interface of two-phase alloys (PA / PP, PA / ABS) | Maleic anhydride grafted compatibilizer (not a coupling agent) | Morphological Observation: Impact and Melt Index Comparison | Two-phase separation, delamination, mechanical fluctuations | Adding it at the same time as the coupling agent will occupy the position repeatedly, so think carefully about which layer you want first. |
The most important row to remember on this table is the last one.
When two resins do not come together, it is the activity of the compatibilizer.
Only when you encounter 'an inorganic fiber and resin that cannot be gripped,' does the coupling agent show its effectiveness.
Buying mixed and adding mixed is the most common waste in this industry.
One more reminder regarding verification: for the interface, the spline data and the item data often don't match.
The spline has a regular flow and uniform thickness; there are weld lines, sharp corners, and thickness transitions on the part.
The problem with the interface often first appears on the component, and then we go back to the spline to find evidence.
So the recommended order of verification is: look at the direction for the spline, and look at the conclusion for the piece.
5. Four types of failures caused by auxiliary agents
Failures related to the interface are often attributed to the 'material'. The following four points are all explained from the perspective of additives.
Failure mode one: For the same batch, the strength fluctuates unpredictably.
The root cause is often not 'unstable material,' but uneven dispersion of the coupling agent.
If any one of the mixing stage time, shearing, or feeding position is off, it will cause some particles to have too much coupling agent while others have too little.
Common solution: First check the mixing process and masterbatching, don't rush to change the material.
Failure mode two: The data looks good in the dry state, but it collapses after wet-heat cycling.
This one is the most typical: the coupling agent only did 'half'.
The silanol group on the silane end has already condensed with the glass fiber, but the organic group on the other end has not been able to form a sufficiently strong bond with nylon.
In the dry state, it can hold up through physical entanglement, but once it gets wet, the layer of siloxane bonds on the interface begins to hydrolyze, and the strength starts to drop.
Common approach: select the variety based on the wet and hot conditions per item, and retest under wet and hot aging conditions, setting the plan without considering the dry state.
Failure Mode Three: The surface of the part turns white, with the corners and the ends of the flow channels being the most severe.
It's certainly fine to check the mold temperature first, but if each batch is different, look at the additives first.
Excess external lubricant or release agent can form a layer of isolation between the glass fibers and resin, covering the interface and exposing the glass fibers.
This point is worth stating more plainly: floating fibers are not entirely a matter of mold temperature. When the same mold produces two different results, the likely factor lies in the material, not the machine.
Common approach: Lower the proportion of external lubricant and use internal lubricant to make up for the flow; at the same time, check whether the coupling agent is well mixed.
Failure Four: When the temperature rises, the interface actually worsens.
Silane coupling agents have their temperature tolerance margin.
The high-temperature section of the extruder reached the upper limit all the way, and part of the coupling agent reacted or decomposed before reaching the surface of the glass fiber.
It has been added, but it didn't work—this kind of situation can't be seen from the physical property table.
Common solution: Confirm the temperature tolerance window of the additives and keep the high-temperature section within the window; adjust the position of fiberglass addition if necessary.
Among the four points above, two of them (uneven dispersion and excessive external lubrication) do not require changing the material or additives; they can be resolved simply by adjusting the process parameters.
This is why troubleshooting should start with the free actions.
6. Processing and Adding: What to Do First, What to Do Later
This part talks about order. If the order is wrong, the things added will consume each other.
First, dry it.
Mineral fillers absorb moisture, and the surface of the glass fibers can also carry water.
When water is brought into the extruder, the coupling agent will react with the water first—it was originally supposed to combine with the surface of the glass fiber.
So first dry, then talk about coupling. This point was already mentioned once earlier when discussing 'when to add white', and here it is being put into action.
Secondly, perform pre-dispersion again.
Directly pouring a spoonful of coupling agent into the main feed inlet is the most convenient, but also the least uniform method.
The common practice is to first make it into a masterbatch, or to first do a pretreatment with glass fiber/filler, allowing it to have the opportunity to be evenly spread on the inorganic surface.
Third, side feeding of fiberglass.
Feed the fiberglass from the side to reduce its retention and shearing in the front half of the screw.
This line is connected to the interface: the longer the fiberglass stays in the screw, the more it gets cut short, and the available specific surface area and effective length for bearing decrease.
Fourth, look at the temperature in different levels.
It's not about 'overall high temperature' or 'overall low temperature', but about the actual temperature at the stage where the coupling agent is added and the stage where the glass fiber is added.
The temperature of these two sections illustrates the problem better than the set value of the barrel.
Fifth, the vacuum section is free of volatile substances.
Low molecular residues remain in the material and will later return in the form of odor, exudation, or surface browning.
Summarized in one sentence: first drive out the water, then spread the additives evenly, then bring in the fiberglass as late as possible, and finally remove the volatiles.
7. When not to add, and what happens if you add too much
This paragraph might be easier than the previous ones.
Let's talk about excess first. Here's a key point: excess means failure.
The excess coupling agent doesn't form more 'strings' at the interface, but rather a weak layer of low molecular weight.
The interface shifts from 'nailing down' to 'padding', but its strength and heat resistance actually decrease.
What's even more troublesome is that the excess migrates to the surface of the part, causing adhesion issues during subsequent painting, printing, and ultrasonic welding.
The more you add, the safer it is—this is the easiest idea to step into with these additives.
Now, let's talk about four situations where additives shouldn't be added.
First, there is no inorganic surface in the system.
Pure resin, no glass fiber, no mineral filler—the coupling agent has nowhere to tie it up; adding it only adds extra cost and a residue.
Second, the glass fiber factory already comes with matching wetting agents.
Conventional chopped glass fiber wetting agents already contain coupling components, and the fiberglass factory processes it according to its own surface treatment process.
This point should be clarified before deciding whether to add it, not "adding again is more stable." Repeated addition or incorrect varieties can disrupt the original wetting system.
Third, the problem is the interface between "resins."
If two resins can't be combined, that's the compatibilizer's problem.
Coupling agents treating alloy interfaces is the wrong direction, and the dosage can't be restored.
Fourth, items need to undergo food contact or medical-related certification.
Additives are chemicals, and in such cases, we must return to the context of GB 4806.7 / ISO 10993.
Certification targets specific grades, and additives must be reported together; you can't wait until certification stalls to remember what was added.
Write these four points at the beginning—not to discourage you, but to save one mold trial.
8. How much is this interface layer worth ?
First, give a total scale: the total cost of additives to the cost per ton of modified plastic is roughly 1%–5%.
It doesn't look like much, but without them, performance often fails testing—this is the key point of this line.
To put it simply: adding 0.5% coupling agent to one ton of material equals five kilograms.
What does five kilograms mean? A bag of material weighs twenty-five kilograms, which accounts for one-fifth of the bag.
In terms of cost, it's just a fraction; in terms of interface, it's the whole bag.
The unit price of additives is always based on the 2026 reference price and market fluctuations, with the current period quoted accordingly.
The additive system in the formula is tailored according to the working conditions of the piece—regular additives are always in stock, special models are matched as needed; You state the operating conditions and grade, and the materials and additives are all matched at once.
Three Frequently Asked Questions by Readers
Q: Can you buy only one of the coupling agent and the compatibilizer?
They cannot replace each other. One tube is the interface between inorganic and organic, the other is the interface between resin and resin.
First, determine which layer you're stuck on, then decide which one to buy. You can buy both, but make sure you know what each is doing.
Question: Our fiberglass comes with a staining agent, do we need to add more?
First, get the parameters, then decide.
Ask clearly what the immersion agent system for glass fiber is, and whether there's a matching for nylon.
If the original interface was already good, adding another layer isn't stable, it's messy.
Question: After adding coupling agent, the dry data looks better, does that mean it can pass the customer's certification?
The direction of this approach is to improve the interface; whether it works depends on the verification results of the parts.
Certification depends on specific grades and additive systems, and which items need to be re-verified according to the customer's standards.
Conclusion
Back to the two samples mentioned at the beginning.
Same mold, same machine, same GF30, difference of 20%, difference between glass fiber and nylon by several dozen nanometers.
Interface matter, the judgment chain is actually very short:
First ask if there is an inorganic surface, then ask which interface layer it is, and finally ask if the verification conditions match the working conditions of the parts.
First receive water, then spread evenly, then enter the site—if the order is correct, leave the rest to testing.
If you have fiberglass reinforcement parts stuck in strength or moisture heat retention, send us three things and you can give directions: glass fiber content and type, the component's damp heat condition, and the one indicator currently stuck.
I got a call a couple of days ago asking about the interface layer.
Over the years of making modified nylon, we've preferred to first ask under what conditions the part is used — dry stuck and falling off in wet heat are two different paths.
You can talk about the interface and trial mold of fiberglass reinforced parts together
