我哋收到MP customer嘅email最常見嗰一種，係咁開頭嘅：「我特登買咗一支color-safe（又或者color-protect、sulfate-free）護髮素，因為個brand marketing話佢係vivid color嚟。我隻染髮仲係fade緊、同之前用普通護髮素一樣快。咩事？」誠實答你：customer一般check錯咗對象。將wash-out fade推快嗰啲鹽，唔喺護髮素入面，係喺洗頭水入面。坐喺典型1%至3% by weight嘅loading度，好多時就係喺啲標住「color-safe」或者「sulfate-free」嘅產品入面。
短版：sodium chloride（即係table salt）係surfactant-based洗頭水入面做load-bearing嘅viscosity modifier，typical concentration 0.5%至3% by weight。佢透過三個同步機制加速semi-permanent vivid dye退色（從cuticle拉走嘅osmotic water loss、ionic-strength打斷cationic dye同cortex之間嘅bond、cuticle swelling）。佢喺好多打住「color-safe」嘅產品入面都見到，因為sulfate-free唔等於salt-free。對HK customer嚟講有好消息：自來水嘅chloride大約11至12 mg per liter，遠遠低過任何threshold；個問題係你倒落頭嗰樽嘢、唔係龍頭嗰啲水。
兩款產品同salt嘅問題
典型嘅洗頭routine係兩款產品：洗頭水remove sebum同product buildup、跟住護髮素detangle同seal cuticle。兩款產品用嘅surfactant chemistry完全唔同，呢個difference就係解釋點解鹽坐喺其中一款入面、唔坐喺另外一款入面。
洗頭水係build around anionic surfactant（sodium lauryl sulfate、sodium laureth sulfate、decyl glucoside、sodium cocoyl isethionate、sodium lauroyl sarcosinate呢類）。呢啲分子個head group帶負電荷，搽落嗰陣個wash concentration足夠形成micelle、將油同污糟拉離頭髮。要將一個稀身嘅surfactant solution變成可以用嘅gel，formulator會加sodium chloride落去。鹽會screen住micelle嘅surface charge、令佢哋pack埋緊啲、轉化成rod-shaped嘅structure，呢啲rod entangle成一個viscous network，就係將bulk liquid整thick嗰個機制。
護髮素就係build around cationic surfactant同conditioning agent（cetrimonium chloride、behentrimonium chloride、behentrimonium methosulfate）加埋fatty alcohol（cetyl alcohol、stearyl alcohol）。呢啲分子帶正電荷，bond落damaged或者porous頭髮嗰啲negatively charged keratin site。護髮素唔需要鹽嚟整thick，因為個cationic system結構黏度嘅方式完全唔同。NaCl喺度唔係load-bearing。
Empirical INCI evidence confirm到呢個pattern。對popular嘅「color-safe」、「color-protect」、「sulfate-free」護髮素做spot-check（Aveda Color Conserve Conditioner、Matrix Total Results Color Obsessed Antioxidant、OGX Renewing+ Argan Oil of Morocco standard formula、Pureology Hydrate、Redken Color Extend Magnetics）顯示sodium chloride要嘛完全absent、要嘛喺ingredient list bottom quartile出現喺trace concentration。相比之下、同一product line嘅companion洗頭水好多時將sodium chloride擺喺ingredient list top third、坐喺load-bearing concentration。「color-safe」呢個marketing claim apply到line入面兩款產品；但係fade色嘅chemistry就唔係。
點解鹽喺洗頭水入面、唔喺護髮素入面
個chemistry喺cosmetic-chemistry trade literature入面有相關documentation。Anionic surfactant形成嘅micelle個surface charge令佢哋互相排斥開。冇鹽嘅話，個surfactant solution就keep住稀身水水。加sodium chloride落去之後，solution入面嘅Na+同Cl-離子會screen住個surface charge，micelle之間嘅electrostatic repulsion跌低，micelle就pack埋緊啲。過咗某個鹽濃度，micelle由spherical transition到rod-shaped，rod互相entangle成一個viscous network，bulk liquid就變gel。Typical loading 0.5%至3% by weight；好多formulator為咗stability同mildness會將cap擺埋喺2%附近。
Cationic surfactant就唔用同一招。護髮素嘅viscosity嚟自fatty alcohol同cationic surfactant形成嘅lamellar structure，再由guar hydroxypropyltrimonium chloride或者polyquaternium variant類嘅thickener提供body。落鹽落cationic system入面唔會produce同樣嘅micelle packing behavior；其實，高NaCl喺一個cationic護髮素system入面可能destabilize成個emulsion。當NaCl真係出現喺護髮素INCI list嗰陣，一般係坐喺最底、做raw material嘅residual electrolyte，唔係deliberately加落去做viscosity modifier。
所以當一個MP customer scan佢支「color-safe」護髮素背面、揾sodium chloride、發現NaCl近list底或者absent嗰陣，個護髮素其實做緊marketing claim suggest嘅嘢。Fade係嚟自其他地方。嗰個其他地方一般係洗頭水。
鹽用三個方法fade緊semi-permanent dye
三個同步機制係wash-out fade嘅driver。
Osmotic gradient。髮絲入面個cortex hold住一啲bound喺keratin protein上面嘅水。當一個鹹嘅洗頭水solution流過cuticle嗰陣，外面個sodium chloride concentration高過裡面嘅水concentration好多。水會由髮絲內向外diffuse、順住concentration gradient走出嚟。Cortex水入面溶咗嘅direct-dye分子會跟住水一齊走。每洗一次頭髮就係dye慢慢漏出洗頭水嘅過程，漏速同wash入面嘅鹽濃度成正比。
Ionic strength打斷dye同keratin之間嘅bond。Manic Panic嘅shade rely substantially on cationic direct dye同related charged chromophore（chromophore上面帶正電荷），同大部分semi-permanent vivid dye一樣。佢哋透過electrostatic attraction黐落cortex上面negatively charged嘅binding site。當wash water入面有高sodium chloride concentration嗰陣，Na+ ion會同cationic dye爭奪嗰啲negative binding site，加埋surrounding aqueous medium嘅ionic strength升高之後、會screen住dye-to-keratin attraction、shrink住electrostatic field嘅effective range（即係electrical double-layer thickness細咗）。Bond變弱，每洗一次就有更多dye lift走。
Cuticle swelling。Sodium chloride喺higher concentration之下會令髮絲osmotically swell起、cuticle scale lift起。Cuticle一開咗之後，cortex就expose多咗、dye分子diffuse出嚟得快啲。Cuticle最終會喺鹽俾fresh water沖走之後settle返埋，但係每一次lift-and-resettle cycle都會loss啲pigment，漂咗嘅頭髮（本身porous啲）呢個effect show得更明顯。
三個機制同一次洗頭一齊發生、effect會compound。對direct dye嘅effect最大，例如Manic Panic呢類direct dye，原因係direct dye唔係chemically bond落cortex（冇oxidative cross-link、冇peroxide reaction）。Permanent dye對個salt mechanism 大致都免疫，因為個dye分子structurally lock咗入髮絲度。同一個chemistry令到MP對頭髮gentle（冇peroxide、冇ammonia、冇PPD，篇11有講）；同一個chemistry亦係佢喺鹽底下wash走嘅原因。
點樣讀洗頭水樽嘅背面去keep色
有咗上面個chemistry，practical check係咁：
第一步。攞起你支洗頭水。反轉嚟。讀個INCI ingredient list。INCI係按concentration decreasing order排嘅，所以top third入面嗰個ingredient就係meaningful concentration（大約1%以上），近底嗰個就係trace（roughly 0.5%以下）。
第二步。揾「sodium chloride」或者「salt」。Note返佢嘅position。
Sodium chloride喺top third即係1%至3% loading、load-bearing嘅thickener、呢支洗頭水有contribute緊你嘅fade。Sodium chloride喺middle即係0.5%至1% loading，仍然meaningful但milder。Sodium chloride近底或者absent即係呢支洗頭水唔係fade driver。
第三步。對你支護髮素repeat一次。上面已經establish咗，護髮素一般唔係source。如果你支護髮素真係將鹽擺喺list上半部，呢個情況unusual；轉護髮素。但係對大部分「color-safe」護髮素嚟講，護髮素係innocent嘅。
第四步。真正要換嗰隻產品係洗頭水、唔係護髮素。揾啲透過alternative thickener achieve viscosity嘅洗頭水（xanthan gum、hydroxyethylcellulose、acrylates copolymer、glycol distearate、loading高啲嘅cocamidopropyl betaine）。呢啲thickener唔需要鹽就做到黏度，一般market成「low-salt」、「no added salt」、或者specifically for keratin-treated同Brazilian-blowout-treated hair。Keratin-treatment呢個category係最reliable嘅貨架，因為keratin-treatment嗰啲chemist多年前就figure out咗鹽會strip走keratin treatment。同一個chemistry strip keratin、亦會strip direct-dye pigment，所以MP customer可以借用呢個category。
仲有一個clarification。「Sulfate-free」呢個marketing claim解決唔到呢個問題。Sulfate-free即係冇sodium lauryl sulfate或者sodium laureth sulfate做cleansing surfactant。佢唔等於salt-free。一支sulfate-free嘅洗頭水仍然可以含2% sodium chloride做thickener；好多都係。兩個claim係independent嘅。我哋見過MP customer specifically為咗color preservation將sulfate洗頭水換做sulfate-free，發現fade rate完全冇變，原因係新支洗頭水carry緊一樣嘅salt load。
點解香港自來水唔係salt嘅來源
Customer可能形成嘅一個reasonable hypothesis係：「香港自來水鹹啲所以fade緊我隻色」。Data唔support呢個講法。香港自來水主要嚟自東江（珠江tributary）、由水務署treatment、儲喺香港啲reservoir度。水務署歷史reporting put住香港自來水嘅chloride concentration大約11至12 mg per liter。
俾你做reference：World Health Organization對食水chloride嘅aesthetic guideline係250 mg per liter（過咗呢個threshold味道就會變鹹）；香港自來水11至12 mg per liter坐喺呢個taste threshold以下大約20倍。一支typical color-safe洗頭水以2% sodium chloride loading計，produce到大約12,000 mg per liter嘅chloride，roughly比香港自來水高三個order of magnitude（大約1,000倍）on a chloride-to-chloride basis。海水大約19,000 mg per liter chloride，同一支heavily-salted洗頭水sit喺同一個order of magnitude。沖涼嗰陣你頭髮接觸到嘅鹽，dominated by你倒落頭嗰樽嘢、唔dominated by龍頭出嗰啲水。
HK customer為咗color preservation purpose、唔需要water-softening filter或者shower filter。Marginal ROI近零。啲錢花喺一支low-salt洗頭水度好過。（對於游真海水或者chlorinated pool嘅HK customer嚟講，嗰個exposure category完全唔同。Pool chlorine喺篇07有講。海泳push落嚟嘅鹽load遠遠大過自來水，但係mitigation同一個shape：用自來水沖洗頭髮、再用fresh-water護髮素saturate埋條頭髮，喺海水hit到之前個cortex已經load滿fresh water、令到osmotic gradient細啲。）
幾時send相過嚟我哋睇
如果你已經轉咗去一支low-salt洗頭水、但係你嘅MP fade rate仍然快過預期，每頁嗰粒magenta nib就係24/7 WhatsApp。第3週影張日光相send過嚟、加埋你而家用緊嗰支洗頭水同護髮素嘅brand同product name。我哋一般睇fade pattern就call得到，呢個residual fade rate係正常你shade family嘅chemistry behavior（紅色fade快過藍色，無論個洗頭routine點，篇24有講）、HK environment compound（ozone或者chlorine，篇06同篇07）、HK humidity打開緊cuticle（篇04）、application issue（uneven coverage讀做快fade）、定係仲有一款haircare product contribute緊鹽你未捉到。Advice係calibrate到MP shade chemistry specifically、HK嘅水同空氣condition、加埋東亞頭髮，narrow過一個generic嘅fade troubleshooting consultation。