So far, we have mostly discussed imaging techniques that allow us to look inside the human body without removing tissue. X-ray, CT, MRI, and ultrasound reveal structures ranging from a few millimeters to entire organs. However, many diseases begin with changes that are much smaller than what these imaging techniques can resolve. Cancer, for example, may alter the shape of individual cell nuclei long before a tumor becomes visible in a radiological examination.
A typical mammalian cell is 10 to 30 µm wide, its nucleus a few micrometers in diameter, and many organelles are smaller still. This is where optical microscopy comes in. It can be used to count and measure cells, to follow them as they move and divide, to test how they respond to drugs, and, once tissue is stained, to diagnose disease. This page covers the physics shared by all light microscopes and the label-free techniques that image living cells without staining. Other advanced microscopy techniques, such as Fluorescence, will be discussed in the next pages.
Representative examples of phase-contrast and differential interference contrast (DIC) microscopy, two techniques that make unstained cells visible**.** (a) SH-SY5Y neuroblastoma cells imaged using phase-gradient contrast (20×). (b) Paramecium bursaria containing symbiotic green algae, imaged using DIC microscopy. (c) Chlorella algae imaged using phase-contrast microscopy. (d) Phosphatidylcholine liposomes imaged using phase-contrast microscopy. (e) Fibroblast imaged using phase-contrast microscopy. (f) Human red blood cells imaged using DIC microscopy. Image source: Wikimedia Commons.
Imaging cells poses two challenges. First, many cellular structures are a few hundred nanometers across, close to the wavelength of visible light. Second, cells are mostly water and absorb very little light, so they are nearly transparent.
This page has three parts. In the first one, we will explore how a microscope forms an image and how much detail it can resolve. The second part follows the image onto the camera: how it is divided into pixels, turned into numbers, and what can go wrong on the way. Finally, the third part focuses on transparency: how phase-contrast and DIC microscopy make unstained cells visible.
A microscope records light that has interacted with the specimen. In a transmitted-light microscope, a lamp below the stage illuminates the specimen through a condenser. An objective lens close to the specimen collects the light passing through it, and a camera records the enlarged image.
How a transmitted-light microscope forms an image. (a) Schematic of the light path (not to scale). The condenser and its aperture shape the illumination (yellow). The objective collects light from each point of the specimen within a cone of half-angle $θ$, which sets its numerical aperture (Eq. 2.8.1), and the tube lens focuses it onto the camera as an enlarged, inverted image. Blue and green trace two specimen points, on and slightly off the axis. (b) The same components on a teaching microscope, where the diaphragm lever adjusts the condenser aperture, and the image is viewed through the eyepieces rather than a camera. (b) Adapted from a photograph by Amada44, via Wikimedia Commons, CC BY-SA 3.0.
An objective lens is labeled with two numbers, for example 40×/0.95. The first is themagnification M: a structure 10 µm across becomes an image 400 µm across on the camera. Magnification determines how large structures appear, and with a fixed camera, it also determines the field of view. A 10× objective captures a region four times wider than a 40× objective, an area 16 times larger.
In modern microscopes, the objective sends parallel light into the so-called infinity space (see figure above), and a separate tube lens focuses it onto the camera. The magnification is set by the ratio of their focal lengths: $M = f_{tube} /f_{obj}$
Manufacturers use different tube-lens focal lengths: 180 mm, 200 mm, or 165 mm. A “40×” objective designed for a 180 mm tube lens therefore gives about 44× with a 200 mm tube lens. The number printed on the objective is a nominal value, which is one reason to read the pixel size from the image metadata rather than compute it from the label.
The second number is the numerical aperture (NA), which describes the cone of light the objective accepts:
$$ NA=n\sin\theta \tag{2.8.1}, $$
Here $n$ is the refractive index of the medium between the specimen and the objective, and $θ$ is the half-angle of the cone of collected light. In air, $n\approx1$, so NA stays below 1. Oil-immersion objectives replace the air gap with oil for which $n\approx1.52$, close to that of glass, so rays at steep angles are no longer lost at the glass–air interface. This is how oil objectives reach NA 1.4.
The two numbers serve different purposes. Magnification determines how large a structure appears. Numerical aperture, together with wavelength, determines how much detail the image contains. Enlarging a blurry photograph does not reveal new structures, and increasing magnification cannot recover detail the objective did not collect. In the next section, we discuss why that happens.
Visible light has a wavelength of about 400 to 700 nanometers, depending on its color; green light, for example, is around 520 nm. Now imagine a tiny source of light far smaller than that, such as a bead 20 nm across, or a single molecule. We might expect the microscope to reproduce it as an equally small point. Instead, it appears as a bright spot surrounded by faint rings. Light is a wave, and waves passing through an opening of finite size spread out: this phenomenon is known as diffraction. The objective collects light through a finite aperture, and when the collected waves meet at the camera they interfere, producing a central spot and weaker rings.