Welcome to chapter three, chemistry and molecular biology for the maker. Chapter one said that living tissue needs a suitable chemical environment and left it there. This chapter unpacks that sentence, because chemistry turns out to be where a surprising amount of device engineering actually lives. It builds on the general chemistry of chapter two and applies it to two things: the molecules of life, and the materials you build with.
Three reasons a maker needs this. First, materials are chemistry. Whatever you make a part from, moulded or machined polymer, cast silicone, stainless steel, glass or a woven strap, it has a composition, a processing history, additives and a surface. What comes out of it into tissue is a chemical question with a regulatory standard attached, and so is what a disinfectant does to it. Second, measurement is chemistry. A glucose meter is an enzyme electrode and a pregnancy test is an immunoassay on a strip. Third, the body defends its chemistry hard, and the narrowness of those ranges tells you how consequential small changes are.
Here are the bond types, ranked by what they mean for a device. Covalent bonds form polymer backbones; breaking them is degradation. Ionic bonds hold salts together, strongly when dry but much more weakly in water. Hydrogen bonds are individually weak but collectively decisive; they give water its behaviour and proteins and DNA their shapes. Van der Waals forces are tiny but scale with contact area. The last row is the one people get wrong. The hydrophobic effect is not an attraction between oily things. It is water's own hydrogen-bond network pushing nonpolar material out of the way. Every cell membrane in your body exists because of it.
Water is half to sixty percent of adult body mass, and almost every property that matters follows from one fact: the molecule is polar and forms extensive hydrogen bonds. It dissolves ions, so anything implanted is in a wet, salty, conductive environment. Its high heat capacity buffers body temperature. Its high latent heat makes sweating efficient, which is why an enclosure that traps moisture defeats it. Surface tension means a press-fit seam wicks fluid. And the last row is worth sitting with. Water is nearly incompressible, so when you compress soft tissue you are pumping fluid out of a porous matrix, which takes time. That is the physical origin of creep and of the load and duration trade-off in chapter five.
pH is the negative logarithm of hydrogen ion activity, so one pH unit is a tenfold change. Arterial blood is held between seven point three five and seven point four five. That looks like a tiny window, and in concentration terms it is about a twenty-five percent span. Fall outside roughly six point eight to seven point eight and life cannot continue. The body defends this range with buffers, and the main one is the bicarbonate system you met in chapter two.
Read the bicarbonate equation left to right and you have the link between breathing and blood chemistry. Blow off more carbon dioxide and pH rises; retain it and pH falls. The lungs compensate within minutes; the kidneys over hours to days. You will hear four names, respiratory and metabolic acidosis and alkalosis, and it is worth recognising them, but interpreting a blood gas is a clinical skill. For a device, pH matters because enzyme sensors have a working range, polyesters such as PLA hydrolyse and their acidic products speed the process, implants corrode in this environment, and drug formulation is largely about pH.
Biology builds with four classes of molecule. Amino acids link into proteins, which do most of the work: enzymes, structure, transport, signalling, antibodies and motors. Sugars link into polysaccharides for energy storage and the sugar coat on cell surfaces. Fatty acids and glycerol build lipids, which make membranes, store energy and form steroid hormones. And nucleotides link into nucleic acids, which hold and execute the genetic information. Chapter four takes the nucleic acids much further.
Proteins have one organising principle: shape determines function. The primary structure is the sequence of amino acids set by the gene. That chain then folds into helices and sheets, into an overall three-dimensional shape, and sometimes into an assembly of several chains. The folds are held by weaker bonds, so heat, extremes of pH and solvents can undo them without breaking the sequence. That is denaturation. And, crucially for devices, so can adsorption onto a foreign surface. Two proteins to know by name: collagen, the dominant tensile protein of tendon, skin and bone, and haemoglobin, whose four cooperating subunits carry oxygen.
Here is the single most useful idea in biomaterials for a newcomer. Within seconds of any material meeting blood or tissue, a layer of protein coats its surface. Everything the body does next, cell attachment, inflammation, clotting, encapsulation, responds to that adsorbed layer, not to your material directly. So the question "is this material biocompatible?" is really "what protein layer forms on it, in what shape, and what does the body do in response?" That is why surface chemistry and surface texture matter far more than bulk composition, and why a polished and a roughened sample of the same alloy can behave differently.
Enzymes are protein catalysts, and their properties are design parameters for any sensor that uses one. Specificity is what makes a glucose electrode selective. Temperature raises the rate until the protein denatures, so sensors have a working range and a shelf life. Each enzyme has a pH optimum. And the one to remember is saturation. At high substrate concentration every active site is busy and the rate plateaus. Michaelis–Menten kinetics describes the curve, but all you need is the consequence: every enzyme-based sensor has a concentration above which it stops responding linearly. Inhibition is interference in a sensor and the mechanism of many drugs.
A phospholipid has a polar head and two oily tails, and in water the hydrophobic effect drives them into a bilayer, heads out and tails in. That bilayer is a selective barrier, it is fluid, and it is electrically insulating with ion channels embedded in it: a capacitor with voltage-gated resistors in parallel, which is exactly what makes the nerve impulse possible. Among carbohydrates, the one to know is the glycocalyx, the sugar coat on cells that helps keep blood vessel linings from clotting. And nucleic acids: DNA is transcribed to RNA and translated to protein, the subject of chapter four.
Cells run on ATP. Glycolysis makes a little without oxygen, but the mitochondria, with oxygen, make far more. So oxygen delivery is the body's central logistics problem, and the heart and lungs exist to solve it. Interrupt blood flow and ATP production falls within minutes, ion pumps fail and cells swell. That is the mechanism behind chapter five's insistence that sustained compression is dangerous: you are not bruising tissue, you are starving it. Brain and heart muscle tolerate interruption poorly; skin and cartilage tolerate more. And a resting adult produces sixty to seventy-five watts of heat, worth remembering inside any enclosure worn against a person.
Cells communicate chemically, over a short range or through the bloodstream, and most drugs work by interfering with that. An agonist binds a receptor and activates it; an antagonist binds and blocks it. Pharmacokinetics is what the body does to the drug; pharmacodynamics is what the drug does to the body. Bioavailability is the fraction reaching the circulation, and half-life sets the dosing interval. The phrase to remember is narrow therapeutic index. For warfarin, lithium, digoxin and many chemotherapy drugs, the gap between helpful and harmful is small. That is why infusion pumps are so heavily regulated and why a unit error in a medication app is not a cosmetic bug.
Now the engineering bridge, the section to come back to. Everything so far was background for it.
Polymer chemistry, practically. Thermoplastics soften and most deform in an autoclave, with PEEK and some polypropylene and polycarbonate grades among the exceptions. The glass transition of PLA is only fifty-five to sixty-five degrees, exceeded by a hot car or a dishwasher. Additives are the things most likely to leach out, and the datasheet names the polymer, not always the additive package. Polyesters hydrolyse; UV yellows and embrittles. Environmental stress cracking, stress plus a chemical, is a real cause of failure: polycarbonate and certain disinfectants. And processing history matters: moulding leaves residual stress, machining leaves coolant and burrs, printing leaves layers.
Extractables are what can be pulled out of a material under deliberately exaggerated conditions, the worst case. Leachables are the subset that actually comes out in normal use, the real exposure. A toxicological assessment then asks whether that exposure is tolerable for this contact type and duration. The framing to take away is that biocompatibility is not a property of a material. It is a conclusion about a finished device, in a specific contact category, for a specific time. So a medical grade feedstock, whether pellets, bar stock, silicone kit or fabric, describes what went in. Your part also carries its processing history.
Leachables come from more places than the polymer itself. Residual monomer and short chains. The additive package of plasticisers, stabilisers and colourants. Processing aids like mould release and cutting fluid. Post-processing solvents, uncured resin and polishing compounds. Textile dyes and finishes. Sterilisation residues. And degradation in service. Several ordinary workshop chemicals are also sensitisers: uncured resin, epoxy hardeners, isocyanates, rosin in solder flux fume, and nickel against the skin. For anything that touches a person, the curing, washing and ventilation steps must be specified and verified, not assumed.
Body fluid is warm, oxygenated and salty, an aggressive environment for metal. Stainless steel and titanium survive because of a thin, self-repairing oxide film, but under a washer or in a thread it cannot re-form, and pitting starts. Two dissimilar metals touching in an electrolyte form a galvanic cell, so mixing alloys is a design decision, not whatever fastener was in the drawer. Textiles are polymer chemistry with a large surface and long skin contact: polyester tolerates hot laundering, elastane loses its stretch, and a coated fabric becomes uncleanable once the coating cracks. Glass is chemically inert, but it fails suddenly and its fragments are sharp.
High-energy surfaces like clean metal and glass wet easily and bond well. Low-energy plastics such as polypropylene, polyethylene and PTFE, and silicone, resist both, which is why polypropylene is hard to glue and silicone needs a primer. Contact angle is a cheap, measurable proxy. Plasma, corona and flame treatment raise surface energy, but only for a working window. Adhesives have their own chemistry: cyanoacrylate is brittle, epoxy has a pot life, and condensation-cure silicones release acetic acid that can corrode. And roughness cuts both ways: better mechanical keying, but more area for protein adsorption and bacterial biofilm.
Many clinical devices run on this chemistry. A glucose meter is the classic enzyme electrode, and a continuous monitor uses the same chemistry under the skin, measuring interstitial fluid that lags blood glucose. Blood gas analysers are sets of ion-selective and gas electrodes. Dialysis is mass transfer across a membrane. Resorbable implants release drugs by controlled hydrolysis of polyesters. Transdermal patches are limited by how good the skin barrier is. Wound dressings are a moisture management problem decided by chemistry, not shape. Even a drug vial is a chemistry decision.
On the software side, electronic prescribing with decision support and pharmacokinetic dose calculators are among the highest-value and highest-risk clinical software. Molecular visualisation tools such as PyMOL, ChimeraX and Mol star render structures from the Protein Data Bank, and AlphaFold has made predicted structures for most known proteins openly available. For teaching, molecular model kits remain excellent, virtual labs let students practise without reagents, medication trainers practise the real barcode workflow with simulated drugs, and demonstration apparatus built from glassware and cheap sensors teaches diffusion and buffering.
The realistic maker entry point here is the physical molecular model. Structures come free from the Protein Data Bank, and educators genuinely want models: a haemoglobin model showing its four subunits explains cooperative binding better than a diagram. Choose the method for what the model must teach. A printed surface suits a whole folded protein. Stacked laser-cut acrylic shows internal cavities and makes a cleanable classroom object. A bent-wire backbone shows the fold and is nearly indestructible. Turned parts suit small-molecule kits. And always state the resolution, what the model omits, and that colour is a convention, not a property.
A polymer handle has been in service six months in a warm room and has gone tacky with fine surface cracks. Candidate mechanisms: plasticiser migrating to the surface, hydrolysis of a polyester, oxidation from light, environmental stress cracking from a cleaning agent, or simple cleaning residue. Each has different evidence and a different check. Note the method. Without a retained control sample from the same batch, stored in known conditions, you are comparing a six-month-old part against your memory. And note that four of the five were predictable from the material choice at the design stage, and the fifth only by asking what it would actually be cleaned with.
For the exercises: build a bilayer from clay or card and show why a sheet closes into a vesicle. Measure the pH of household liquids and see how compressed the blood range looks on a log scale. Add vinegar to plain water and to water with bicarbonate of soda and watch the buffer work. Make one molecular model from a Protein Data Bank structure, by a method chosen for what it must show, with a written limitations statement. And start the habit of keeping a control sample now.
The limits. You understand why biocompatibility is framed around a finished device and a contact category, but the evaluation is a specialist activity. Knowing that ethylene oxide leaves residues is not managing them. You know what a blood gas or glucose meter measures, but what it means for a person is clinical. You have enough pharmacology vocabulary to follow a conversation and to see why dose software is safety-critical, not more. And a molecule model is a schematic. Next, foundations of biology: life, cells, and how biologists work.