Emil W. Ciurczak, DoraMaxx Consulting11.07.22
I was just speaking with a friend and remembered that 2023 will mark my working with NIRS for forty years. So, to celebrate, I will not talk about Near-Infrared in this column. Instead, I will discuss two interesting (to me, anyway) instruments: Ion Mobility Spectroscopy (IMS) and “gated” Raman Spectroscopy.
Anyone who has flown commercially in the last decade has encountered IMS. This is used by TSA personnel to check for drugs and explosives on your person or luggage. They swab your hand or luggage handle with a small white cloth and place it on a device (about the size of a microwave oven), wait for a minute, then either wave you through or send you for a strip-search, depending on what the device reveals. I was, of course, curious about how it worked, so I enquired. (Not the TSA, they just use it, but the manufacturer.)
It turns out to simply be a “poor man’s” mass spectrometer, just easier to use and maintain. Figure 1 shows a schematic of an IMS. The sample—thermally desorbed from the sampling cloth or injected as a liquid—is vaporized and some of the sample (minus any solvent) is sent through an ionization section. Various methods are used to give the molecules a charge, normally, a simple single charge, and through a like-charged mesh. The charges repel and the ionized molecules are sped down the “drift region.”
A low velocity of air is introduced to oppose the traveling ions (in a true Mass Spectrometer, the separations happen in a vacuum). Since each has a single charge and different mass the mass/charge ration is unique for each moiety and the time it takes for them to reach the detector is proportional to their masses. When they strike the oppositely charged detector, a current occurs, proportional to the number of ions arriving at any given time. A graph of the current vs. drift time generates an equivalent to a chromatogram (see Figure 2). Note the time scale: milliseconds vs. minutes in a typical HPLC chromatogram. Yes, UPLC is faster than HPLC, but still needs more validation time, uses solvents and columns, needs more sample prep time, and still is much slower than IMS.
When you carefully read the GMP rules, there are two important conditions that are, to be honest, not strictly followed: 1) meaningful in-process tests should be run and 2) a statistically significant number of final dosage form must be run. “Meaningful” means, in my view, timely and giving useful information (otherwise, why perform the test?). Any traditional test, such as hardness, friability, or disintegration time, is neither timely, nor does it allow the process to be modified in time to save the bulk of dosage forms. Not only are these destructive (discouraging multiple tests), but take enough time that they are no longer “timely.” If disintegration is mandatory (baked into MMF from owner of the product), it could be coupled with IMS to estimate the content uniformity throughout the run.
In addition to the spectrographic approach (NIRS) I mentioned last column allowing 100% inspection of a run, we can, if management/client still insists on grab samples, why not use Raman as a tool? As much as I love my NIRS, I must admit that a Raman spectrum is far easier to interpret without sophisticated Chemometrics, making is simpler to use for a larger inventory of products, as might be found in a generic or CMO production setting.
Until several years ago, the EMA (European Medicines Agency) and U.S. FDA was hesitant to accept Raman as a means of ID or analysis. The reason was that, in those days, the only instruments available utilized a small-beam (<1mm) LASER to illuminate the sample. If the instrument were used on a blend (granulation or tablet) a small shift of position could illuminate a particle of API or talc or lactose, leading to poor results. (Drastic shifts for blend uniformity were common.) On top of this small sample area, the Raman spectrum is almost buried in fluorescence. Most samples fluoresce much more than any Raman signal, so the resultant can almost look more like a noisy baseline (see Figure 5A) than a spectrum.
Then several instrument manufacturers introduced modifications that helped with these problems. First Kaiser developed a probe with a larger spot size, allowing Raman to be used for content work (blending, moisture, residual solvents, etc.). It does not directly address the fluorescence problem however. Some more recent techniques have helped alleviate the problem.
One such technology is Spatially Resolved Raman (Figure 4B), where the fiber illuminating the sample is several millimeters away from the fiber that is gathering the resultant radiation. Since, in a condensed sample (powder, tablet, freeze-dried cake) fluorescence is rapidly re-absorbed, so locating the point of observation away from the illumination greatly reduces the background radiation and allows the Raman signal to be seen (Figure 5B). Another is software that was developed to synthesize a baseline and emphasize the Raman spectrum; relatively new, it will need to be validated completely for routine FDA/EMA approval.
Another, more sophisticated technology uses a technique called “time-gating.” The LASER is a rapidly pulsed unit. The pulse of light strikes the sample, causing both Raman and fluorescence radiation to be generated. However, since they are generated by different mechanisms, they are generated at a different rate. The fluorescence mechanism involves a photon causing an electron to be elevated to a higher energy state. When the electron returns to its ground state, it emits a photon of lower energy (everything gets taxed, it seems).
Raman, on the other hand, is based on scattering. The vast number of photons are scattered elastically and retain the same wavelength. A small number are scattered inelastically and change wavelength according to the bonds they encounter within the molecule. This causes shifts, either up or down (higher or lower wavelengths), generating a spectrum of the material from which it was emitted. But, since the scattering is nearly immediate, the Raman spectrum emerges just before the fluorescence spectrum.
The “gating-effect” comes from a rapid opacifying of the window through which both the LASER and returning light travel. So, the steps are 1) window clears, 2) LASER pulses, 3) light reacts with sample, 4) the Raman spectrum emerges and travels through the window, 5) the window opacifies, 6) the fluorescence spectrum is left looking for a home. When you couple this sharper image with the fact that Raman doesn’t “see” water—Raman-active molecules need to have a center of symmetry; water does not.
Emil W. Ciurczak
DoraMaxx Consulting
Emil W. Ciurczak has worked in the pharmaceutical industry since 1970 for companies that include Ciba-Geigy, Sandoz, Berlex, Merck, and Purdue Pharma, where he specialized in performing method development on most types of analytical equipment. In 1983, he introduced NIR spectroscopy to pharmaceutical applications, and is generally credited as one of the first to use process analytical technologies (PAT) in drug manufacturing and development.
Anyone who has flown commercially in the last decade has encountered IMS. This is used by TSA personnel to check for drugs and explosives on your person or luggage. They swab your hand or luggage handle with a small white cloth and place it on a device (about the size of a microwave oven), wait for a minute, then either wave you through or send you for a strip-search, depending on what the device reveals. I was, of course, curious about how it worked, so I enquired. (Not the TSA, they just use it, but the manufacturer.)
It turns out to simply be a “poor man’s” mass spectrometer, just easier to use and maintain. Figure 1 shows a schematic of an IMS. The sample—thermally desorbed from the sampling cloth or injected as a liquid—is vaporized and some of the sample (minus any solvent) is sent through an ionization section. Various methods are used to give the molecules a charge, normally, a simple single charge, and through a like-charged mesh. The charges repel and the ionized molecules are sped down the “drift region.”
A low velocity of air is introduced to oppose the traveling ions (in a true Mass Spectrometer, the separations happen in a vacuum). Since each has a single charge and different mass the mass/charge ration is unique for each moiety and the time it takes for them to reach the detector is proportional to their masses. When they strike the oppositely charged detector, a current occurs, proportional to the number of ions arriving at any given time. A graph of the current vs. drift time generates an equivalent to a chromatogram (see Figure 2). Note the time scale: milliseconds vs. minutes in a typical HPLC chromatogram. Yes, UPLC is faster than HPLC, but still needs more validation time, uses solvents and columns, needs more sample prep time, and still is much slower than IMS.
When you carefully read the GMP rules, there are two important conditions that are, to be honest, not strictly followed: 1) meaningful in-process tests should be run and 2) a statistically significant number of final dosage form must be run. “Meaningful” means, in my view, timely and giving useful information (otherwise, why perform the test?). Any traditional test, such as hardness, friability, or disintegration time, is neither timely, nor does it allow the process to be modified in time to save the bulk of dosage forms. Not only are these destructive (discouraging multiple tests), but take enough time that they are no longer “timely.” If disintegration is mandatory (baked into MMF from owner of the product), it could be coupled with IMS to estimate the content uniformity throughout the run.
In addition to the spectrographic approach (NIRS) I mentioned last column allowing 100% inspection of a run, we can, if management/client still insists on grab samples, why not use Raman as a tool? As much as I love my NIRS, I must admit that a Raman spectrum is far easier to interpret without sophisticated Chemometrics, making is simpler to use for a larger inventory of products, as might be found in a generic or CMO production setting.
Until several years ago, the EMA (European Medicines Agency) and U.S. FDA was hesitant to accept Raman as a means of ID or analysis. The reason was that, in those days, the only instruments available utilized a small-beam (<1mm) LASER to illuminate the sample. If the instrument were used on a blend (granulation or tablet) a small shift of position could illuminate a particle of API or talc or lactose, leading to poor results. (Drastic shifts for blend uniformity were common.) On top of this small sample area, the Raman spectrum is almost buried in fluorescence. Most samples fluoresce much more than any Raman signal, so the resultant can almost look more like a noisy baseline (see Figure 5A) than a spectrum.
Then several instrument manufacturers introduced modifications that helped with these problems. First Kaiser developed a probe with a larger spot size, allowing Raman to be used for content work (blending, moisture, residual solvents, etc.). It does not directly address the fluorescence problem however. Some more recent techniques have helped alleviate the problem.
One such technology is Spatially Resolved Raman (Figure 4B), where the fiber illuminating the sample is several millimeters away from the fiber that is gathering the resultant radiation. Since, in a condensed sample (powder, tablet, freeze-dried cake) fluorescence is rapidly re-absorbed, so locating the point of observation away from the illumination greatly reduces the background radiation and allows the Raman signal to be seen (Figure 5B). Another is software that was developed to synthesize a baseline and emphasize the Raman spectrum; relatively new, it will need to be validated completely for routine FDA/EMA approval.
Another, more sophisticated technology uses a technique called “time-gating.” The LASER is a rapidly pulsed unit. The pulse of light strikes the sample, causing both Raman and fluorescence radiation to be generated. However, since they are generated by different mechanisms, they are generated at a different rate. The fluorescence mechanism involves a photon causing an electron to be elevated to a higher energy state. When the electron returns to its ground state, it emits a photon of lower energy (everything gets taxed, it seems).
Raman, on the other hand, is based on scattering. The vast number of photons are scattered elastically and retain the same wavelength. A small number are scattered inelastically and change wavelength according to the bonds they encounter within the molecule. This causes shifts, either up or down (higher or lower wavelengths), generating a spectrum of the material from which it was emitted. But, since the scattering is nearly immediate, the Raman spectrum emerges just before the fluorescence spectrum.
The “gating-effect” comes from a rapid opacifying of the window through which both the LASER and returning light travel. So, the steps are 1) window clears, 2) LASER pulses, 3) light reacts with sample, 4) the Raman spectrum emerges and travels through the window, 5) the window opacifies, 6) the fluorescence spectrum is left looking for a home. When you couple this sharper image with the fact that Raman doesn’t “see” water—Raman-active molecules need to have a center of symmetry; water does not.
Emil W. Ciurczak
DoraMaxx Consulting
Emil W. Ciurczak has worked in the pharmaceutical industry since 1970 for companies that include Ciba-Geigy, Sandoz, Berlex, Merck, and Purdue Pharma, where he specialized in performing method development on most types of analytical equipment. In 1983, he introduced NIR spectroscopy to pharmaceutical applications, and is generally credited as one of the first to use process analytical technologies (PAT) in drug manufacturing and development.
